The Functions of ‘General Nouns’: Theory and Corpus Analysis

The author critically discusses the concept of ‘general nouns’, which Halliday/ Hasan introduced in their approach to lexical cohesion (1976), and she provides a comprehensive overview of these nouns from a micro- and a macro-linguistic perspective. For the empirical analysis, the author compiled a corpus, which allows statements about a medium- and genre-specific use of ‘general nouns’. For this purpose, she developed an analytical tool, which takes into account formal and semantic features. The major outcome of the corpus analysis is that ‘general nouns’ are much more flexible in form and function than Halliday/ Hasan assumed and, most importantly, that they fulfil genre-specific functions some of which have not systematically been associated with lexical cohesion.

ISBN 978-3-631-74758-2

ECL_017_274758_Benninghoven_MP_A5HC 151x214 globalL.indd 1

www.peterlang.com

english corpus linguistics Thomas Kohnen · Joybrato Mukherjee (eds.) he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit

Vera Benninghoven

Vera Benninghoven · The Functions of ‘General Nouns’ Vera Benninghoven has taught linguistics at the English department of the University of Bonn and the University of Cologne. Her research interests include morphology, lexicology, text-linguistics and corpus-linguistics.

V o l u m e 17

ecl 17

n

The Functions of ‘General Nouns’ Theory and Corpus Analysis

13.08.18 03:42

The Functions of ‘General Nouns’

ENGLISH CORPUS LINGUISTICS Thomas Kohnen / Joybrato Mukherjee (eds.)

VOLUME 17

Vera Benninghoven

The Functions of ‘General Nouns’ Theory and Corpus Analysis

Bibliographic Information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the internet at http://dnb.d-nb.de. Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Zugl.: Bonn, Univ., Diss., 2015

D5 ISSN 1610-868X ISBN 978-3-631-74758-2 (Print) E-ISBN 978-3-631-76018-5 (E-PDF) E-ISBN 978-3-631-76019-2 (EPUB) E-ISBN 978-3-631-76020-8 (MOBI) DOI 10.3726/b14310 © Peter Lang GmbH Internationaler Verlag der Wissenschaften Berlin 2018 Alle Rechte vorbehalten.

All rights reserved. Peter Lang – Berlin · Bern · Bruxelles · New York · Oxford · Warszawa · Wien All parts of this publication are protected by copyright. Any utilisation outside the strict limits of the copyright law, without the permission of the publisher, is forbidden and liable to prosecution. This applies in particular to reproductions, translations, microfilming, and storage and processing in electronic retrieval systems. This publication has been peer reviewed. www.peterlang.com

Meinen Eltern gewidmet

Danksagung Mein herzlicher Dank gilt meinem Doktorvater, Herrn Prof. Dr. Jürgen Esser, für die Betreuung meiner Dissertation. Durch seine konstruktive Kritik und seinen wertvollen fachlichen Rat gab er mir immer wieder hilfreiche Denkanstöße und hat so maßgeblich zum Gelingen dieser Arbeit beigetragen. Ich danke auch Herrn Prof. Dr. Klaus Peter Schneider für seine Zweitbegutachtung, Frau Prof. Dr. Marion Gymnich für die Übernahme des Prüfungsvorsitzes und Herrn Prof. Dr. Tilman Mayer für seine Funktion in der Prüfungskommission. Ich bedanke mich bei meinen Kolleginnen und Kollegen des Forschungskolloquiums, Sanna Engell, Laura Göttmann, Sebastian Patt, Julia Sosnizka und Sharmila Vaz, für die Lektüre dieser Arbeit und die daraus resultierenden anregenden Diskussionen. Mein Dank geht auch an Herrn Prof. Dr. Joybrato Mukherjee und Herrn Prof. Dr. Thomas Kohnen für ihre Hilfestellung bei der Veröffentlichung dieser Arbeit. Ich möchte mich ganz herzlich bei den Menschen in meinem Leben bedanken, ohne deren Unterstützung, Halt und Zuspruch diese Arbeit nicht hätte entstehen können. Mein besonderer Dank geht an Felix Sosnizka für seinen uneingeschränkten Rückhalt. Außerdem gilt mein herzlicher Dank Annika Anderle, die mich immer wieder in meinem Vorhaben bestärkt und mir stets neuen Mut gemacht hat, wenn ich ihn gebraucht habe. Dies gilt auch für meine Schwester Anne Benninghoven und meinen Bruder Axel Benninghoven, denen ich für den festen (Zusammen-)Halt danke. Mein ganz besonders tief empfundener Dank aber gilt meinen Eltern, Gisela und Helmut Benninghoven. Ohne ihre bedingungslose Liebe und Unterstützung hätte ich meinen Weg niemals gehen und schließlich auch mein Promotionsvorhaben niemals umsetzen können. Ich widme ihnen diese Arbeit. Köln, Januar 2018

Vera Benninghoven

7

Table of contents List of tables and figures .................................................................................... 15 1. Introduction .................................................................................................... 21 1.1 Aim of the study .............................................................................................. 22 1.2 Structure of the study ..................................................................................... 23

2. A linguistic discussion of ‘general nouns’ ............................................ 25 2.1 Halliday/Hasan’s approach to ‘general nouns’ ............................................ 25 2.1.1 The concept of ‘cohesion’ ...................................................................... 25 2.1.2 The concept of ‘reference’ ..................................................................... 28 2.1.3 The concept of ‘reiteration’ ................................................................... 33 2.1.4 The concept of ‘general nouns’ ............................................................. 35 2.2 Previous studies of ‘general nouns’............................................................... 38 2.3 Critical evaluation of the status of ‘general nouns’ .................................... 41

3. Approaches to general noun phrases across different linguistic disciplines.................................................................... 43 3.1 Introduction ..................................................................................................... 43 3.2 The grammatical approach to general noun phrases ................................ 44 3.2.1 Classification of nouns .......................................................................... 44 3.2.2 Modification ........................................................................................... 46 3.2.3 Classification of general noun phrase heads ...................................... 51 3.3 The text-linguistic approach to general noun phrases .............................. 54 3.3.1 Constitutive features of textual communication ............................... 54 3.3.2 Regulative features of textual communication .................................. 58 3.3.3 Restricted and elaborated style of referencing .................................... 60 3.4 The pragmatic approach to general noun phrases ..................................... 62 3.4.1 The Cooperative Principle .................................................................... 62 3.4.2 Vagueness in language .......................................................................... 66 9

3.4.2.1 Defining vagueness ................................................................. 67 3.4.2.2 Types of vagueness .................................................................. 69 3.5 The cognitive approach to general noun phrases ...................................... 73 3.5.1 Categorisation ........................................................................................ 73 3.5.2 Context-dependence of categories ...................................................... 76 3.5.3 Categorisation and general noun phrases .......................................... 77 3.5.3.1 Introduction ............................................................................. 77 3.5.3.2 General noun phrases as “empty containers”....................... 81 3.5.3.3 General noun phrases as “full containers” ........................... 82 3.6 Summary: A workable definition of general noun phrases ..................... 84

4. Framework for the analysis of general noun phrases ..................... 87 4.1 Combination of structural and semantic parameters of general noun phrases ................................................................................. 87 4.1.1 Structural parameters of general noun phrases: +/- Modification ..................................................................... 88 4.1.1.1 Non-modified general noun phrase heads: - Modification .............................................................. 88 4.1.1.2 Modified general noun phrase heads: + Modification........ 89 4.1.2 Semantic parameters of general noun phrases: +/- Linkage............ 91 4.1.2.1 Non-linked general noun phrases: - Linkage....................... 92 4.1.2.2 Linked general noun phrases: + Linkage.............................. 93 4.2 Functional matrix for the analysis of general noun phrases .................... 94 4.3 The scale of specification ........................................................................... 97 4.4 General assumptions for the analysis of general noun phrases .............. 99

5. Corpus compilation .................................................................................... 105 5.1 The corpus-linguistic approach of the present study ............................. 105 5.2 Medium-, domain- and genre-specific features of the corpus data .................................................................................... 107 5.2.1 Some features of spoken and written language ............................. 107 5.2.2 Some features of legal language ......................................................... 109 10

5.2.3 Some features of political language ................................................... 111 5.2.4 Some features of conversation ........................................................... 113 5.3 The corpus of the present study ................................................................ 114 5.3.1 Written corpus ..................................................................................... 114 5.3.1.1 Supreme Court judgments ................................................. 114 5.3.1.2 Political manifestos ............................................................. 116 5.3.2 Spoken corpus ...................................................................................... 117 5.3.2.1 Parliamentary debates ........................................................ 117 5.3.2.2 General conversations ........................................................ 120 5.3.3 Summary: Overview of corpus data.................................................. 122

6. Methodology .................................................................................................. 127 6.1 Methods for the quantitative analysis ...................................................... 127 6.1.1 From raw to relevant data................................................................... 127 6.1.2 Determining the frequency of relevant general noun phrase heads ............................................................................... 131 6.2 Methods for the qualitative analysis ......................................................... 132 6.2.1 The coding system used in the present study................................... 132 6.2.2 The parameter modification................................................................ 136 6.2.2.1 Types of premodification ................................................... 136 6.2.2.2 Types of postmodification ................................................. 138 6.2.3 The parameter linkage ......................................................................... 141 6.2.3.1 Types of endophoric reference .......................................... 141 6.2.3.1.1 Halliday/Hasan’s endophoric reference ............................ 141 6.2.3.1.2 Encapsulation ...................................................................... 142 6.2.3.1.3 Reference between appositive units .................................. 144 6.2.3.1.4 Reference between subject and complement................... 145 6.2.3.1.5 Remote reference................................................................. 147 6.2.3.2 Generic reference ................................................................ 148

7. Corpus analysis ............................................................................................. 151 7.1 Quantitative analysis: Frequencies and distribution of relevant general noun phrase heads in the corpora .............................................. 151 11

7.1.1 7.1.2 7.1.3 7.1.4 7.1.5

Quantitative results from the judgment corpus .............................. 151 Quantitative results from the manifesto corpus .............................. 155 Quantitative results from the debate corpus .................................... 159 Conversation corpus ........................................................................... 161 Summary and comparison of results ................................................ 165

7.2 Qualitative analysis: The degree of specification of relevant general noun phrases in the corpora ......................................................... 173 7.2.1 Qualitative results from the judgment corpus ................................. 173 7.2.1.1 ‘Most specific’ general noun phrases in the judgment corpus ............................................................. 179 7.2.1.2 ‘Rather specific’ general noun phrases in the judgment corpus ............................................................. 184 7.2.1.3 ‘Less specific’ general noun phrases in the judgment corpus ............................................................. 187 7.2.1.4 ‘Least specific’ general noun phrases in the judgment corpus ............................................................. 190 7.2.2 Qualitative results from the manifesto corpus................................. 191 7.2.2.1 ‘Most specific’ general noun phrases in the manifesto corpus ............................................................. 194 7.2.2.2 ‘Rather specific’ general noun phrases in the manifesto corpus ............................................................. 197 7.2.2.3 ‘Less specific’ general noun phrases in the manifesto corpus ............................................................. 198 7.2.2.4 ‘Least specific’ general noun phrases in the manifesto corpus ............................................................. 200 7.2.3 Qualitative results from the debate corpus ...................................... 201 7.2.3.1 ‘Most specific’ general noun phrases in the debate corpus ................................................................... 205 7.2.3.2 ‘Rather specific’ general noun phrases in the debate corpus ................................................................... 209 7.2.3.3 ‘Less specific’ general noun phrases in the debate corpus ................................................................... 210 7.2.3.4 ‘Least specific’ general noun phrases in the debate corpus ................................................................... 212

12

7.2.4 Qualitative results from the conversation corpus ........................... 213 7.2.4.1 ‘Most specific’ general noun phrases in the conversation corpus .............................................................. 218 7.2.4.2 ‘Rather specific’ general noun phrases in the conversation corpus .............................................................. 220 7.2.4.3 ‘Less specific’ general noun phrases in the conversation corpus .............................................................. 221 7.2.4.4 ‘Least specific’ general noun phrases in the conversation corpus .............................................................. 225 7.2.5 Summary and comparison of results ................................................ 226

8. Summary and conclusions ....................................................................... 237 References ............................................................................................................. 249

13

List of tables and figures Figure 2.1: Figure 2.2: Figure 3.1: Figure 3.2: Table 3.1: Figure 3.3: Figure 3.4: Table 3.2: Figure 3.5: Figure 3.6: Figure 3.7: Figure 3.8: Figure 3.9: Figure 4.1: Figure 4.2: Figure 4.3: Figure 4.4: Figure 4.5: Table 5.1: Figure 5.1: Table 5.2:

The concept of ‘cohesion’...................................................................... 27 Different concepts of ‘reference’ .......................................................... 30 The classification of nouns ................................................................... 45 A complex noun phrase ....................................................................... 47 The classification of general noun phrase heads ............................... 51 The classification of general noun phrase heads in terms of specification ............................................................................ 53 The concepts of ‘ambiguity’, ‘polysemy’ and ‘vagueness’ .................. 68 Tzeltal plant classification according to Berlin et al. (1973) ............ 75 Horizontal dimension of categorisation illustrated by the example ‘dog’.......................................................................................... 79 Vertical dimension of categorisation illustrated by the example ‘dog’ .............................................................................. 79 Extension of a general noun phrase category dependent on co-text and context .......................................................................... 80 Extension of the general noun phrase category ‘people’ dependent on co-text and context of example (39) .......................... 82 Extension of the general noun phrase category ‘people’ dependent on co-text and context of example (40) .......................... 83 The functional matrix for the analysis of general noun phrases ............................................................................ 95 Generalisation and specification illustrated with the functional matrix for the analysis of general noun phrases ............ 96 The scale of specification ...................................................................... 98 The assumed arrangement of the four sub-corpora on the scale of specification (depending on frequency) ................ 101 The assumed arrangement of the sub-corpora on the scale of specification (depending on modification and linkage) ............ 103 The corpus of the present study ........................................................ 106 The structure of the parliamentary government in the UK ........... 118 The judgment corpus .......................................................................... 122 15

Table 5.3: Table 5.4: Table 5.5: Figure 6.1: Table 6.1: Table 6.2: Figure 6.2: Figure 6.3: Figure 6.4: Figure 6.5: Table 7.1: Figure 7.1: Table 7.2: Figure 7.2: Table 7.3: Figure 7.3: Table 7.4: Figure 7.4: Figure 7.5: Figure 7.6: Figure 7.7:

16

The manifesto corpus ....................................................................... 123 The debate corpus ............................................................................... 124 The conversation corpus .................................................................... 125 Distinguishing relevant from irrelevant general noun phrases .......................................................................... 130 Numbering of general noun phrase heads ..................................... 132 Labelling of corpus texts .................................................................... 133 Example of the coding of a general noun phrase head in the judgment corpus ...................................................................... 134 Highlighting of general noun phrase heads in the corpus texts ................................................................................. 134 Coding of general noun phrase heads in the corpus texts ............ 135 Searching for single occurrences of general noun phrase heads within a pdf document .......................................................... 135 Frequencies of general noun phrase heads in the judgment corpus ........................................................................... 152 General noun phrase head types and tokens in the judgment corpus ......................................................................... 153 Frequencies of general noun phrase heads in the manifesto corpus ........................................................................ 155 General noun phrase head types and tokens in the manifesto corpus ........................................................................ 157 Frequencies of general noun phrase heads in the debate corpus ................................................................................ 159 General noun phrase head types and tokens in the debate corpus ................................................................................ 160 Frequencies of general noun phrase heads in the conversation corpus ..................................................................... 162 General noun phrase head types and tokens in the conversation corpus ....................................................................... 163 Comparison of the normalised overall frequencies of general noun phrase heads in the four sub-corpora ...................... 166 The four sub-corpora on the scale of specification ........................ 167 Comparison of the distribution of general noun phrase heads in the four sub-corpora ............................................. 169

Figure 7.8: Figure 7.9: Table 7.5: Table 7.6: Figure 7.10: Figure 7.11: Table 7.7: Table 7.8: Table 7.9: Table 7.10: Table 7.11: Table 7.12: Table 7.13: Table 7.14: Figure 7.12: Figure 7.13: Table 7.15: Table 7.16:

Comparison of the frequencies of the general noun phrase heads ‘people’ and ‘child’ ............................................. 172 Comparison of the frequencies of the general noun phrase heads ‘matter’ and ‘question’ ....................................... 172 Modified and non-modified general noun phrase heads in the judgment corpus ........................................................... 174 Linked and non-linked general noun phrases in the judgment corpus ...................................................................... 174 Correlation of modification and linkage of general noun phrases in the judgment corpus .............................................. 175 Distribution of general noun phrase heads across the four categories of specification in the judgment corpus ................ 176 Most frequent types of modification in ‘most specific’ general noun phrases in the judgment corpus ................................ 180 Most frequent types of linkage of ‘most specific’ general noun phrases in the judgment corpus .............................................. 180 Most frequent combinations of modification and linkage of ‘most specific’ general noun phrases in the judgment corpus ...... 181 Most frequent types of linkage of ‘rather specific’ general noun phrases in the judgment corpus ............................................ 184 Most frequent types of modification in ‘less specific’ general noun phrases in the judgment corpus ................................ 187 Most frequent combinations of modification and - linkage in ‘less specific’ general noun phrases in the judgment corpus ......... 188 Modified and non-modified general noun phrase heads in the manifesto corpus .................................................................... 191 Linked and non-linked general noun phrases in the manifesto corpus ................................................................................. 191 Correlation of modification and linkage of general noun phrases in the manifesto corpus .................................................... 192 Distribution of general noun phrase heads across the four categories of specification in the manifesto corpus ........................ 193 Most frequent types of modification in ‘most specific’ general noun phrases in the manifesto corpus ............................. 195 Most frequent types of linkage of ‘most specific’ general noun phrases in the manifesto corpus ............................................. 195 17

Table 7.17: Table 7.18: Table 7.19: Table 7.20: Figure 7.14: Figure 7.15: Table 7.21: Table 7.22: Table 7.23: Table 7.24: Table 7.25: Table 7.26: Table 7.27: Table 7.28: Figure 7.16: Figure 7.17: Table 7.29: Table 7.30:

18

Most frequent types of modification in ‘less specific’ general noun phrases in the manifesto corpus ............................. 198 Most frequent combinations of modification and - linkage of ‘less specific’ general noun phrases in the manifesto corpus ........ 199 Modified and non-modified general noun phrase heads in the debate corpus ................................................................. 202 Linked and non-linked general noun phrases in the debate corpus ................................................................................ 202 Correlation of modification and linkage of general noun phrases in the debate corpus ................................................... 203 Distribution of general noun phrase heads across the four categories of specification in the debate corpus .............................. 204 Most frequent types of modification in ‘most specific’ general noun phrases in the debate corpus ..................................... 206 Most frequent types of linkage of ‘most specific’ general noun phrases in the debate corpus ................................................... 206 Most frequent combinations of modification and linkage in ‘most specific’ general noun phrases in the debate corpus ........... 207 Most frequent types of linkage of ‘rather specific’ general noun phrases in the debate corpus ................................................... 209 Most frequent types of modification in ‘less specific’ general noun phrases in the debate corpus ................................................... 210 Most frequent combinations of modification and - linkage of ‘less specific’ general noun phrases in the debate corpus .............. 211 Modified and non-modified general noun phrase heads in the conversation corpus ..................................................................... 214 Linked and non-linked general noun phrases in the conversation corpus ............................................................................ 214 Correlation of modification and linkage of general noun phrases in the conversation corpus................................................... 215 Distribution of general noun phrase heads across the four categories of specification in the conversation corpus ................... 216 Most frequent types of modification in ‘most specific’ general noun phrases in the conversation corpus........................... 218 Most frequent types of linkage of ‘most specific’ general noun phrases in the conversation corpus ........................................ 219

Table 7.31:

Most frequent types of linkage of ‘rather specific’ general noun phrases in the conversation corpus ........................................ 220 Table 7.32: Most frequent types of modification in ‘less specific’ general noun phrases in the conversation corpus ......................... 222 Table 7.33: Most frequent combinations of modification and linkage in ‘less specific’ general noun phrases in the conversation corpus ............................................................................ 222 Table 7.34: Comparison of modified and non-modified general noun phrase heads and linked and non-linked general noun phrases in the four sub-corpora (in percentage) ............................ 227 Table 7.35: Comparison of general noun phrases across the four categories of specification in the four sub-corpora (in percentage) ..................................................................................... 228 Figure 7.18: Positioning of the four sub-corpora along the scale of specification .................................................................................. 230 Table 7.36: Comparison of the most frequent general noun phrase heads in the four categories of specification across the four sub-corpora ................................................................................. 232 Table 7.37: Comparison of most frequent realisations of general noun phrases in the four categories of specification across the four sub-corpora ............................................................... 233 Figure 8.1: The functional matrix illustrating the generalisation and specification of general noun phrases ....................................... 240 Table 8.1: Compilation of the corpus data ..................................................... 242 Figure 8.2: The four sub-corpora on the scale of specification (depending on frequency) ................................................................. 243 Figure 8.3: The categories of specification which were most frequent in the four sub-corpora ........................................... 244 Figure 8.4: The categories of specification which were least frequent in the four sub-corpora ............................................ 244 Figure 8.5: The four sub-corpora on the scale of specification (depending on modification and linkage) ....................................... 245

19

1. Introduction Abstract: Halliday/Hasan’s approach to cohesion in English (1976), which treats ‘general nouns’ as a device of lexical cohesion, is a purely theoretical approach based on carefully compiled textbook examples. This shortcoming points to the relevance of the present study, which provides a comprehensive corpus-based account of ‘general nouns’.

Halliday/Hasan (1976) introduce the class of ‘general noun’ within their textlinguistic approach to cohesion and define it as “a small set of nouns having generalized reference within the major noun classes, […]” (1976: 274). Based on perfect textbook examples, Halliday/Hasan show that ‘general nouns’ are lexically superordinate terms which anaphorically refer to a preceding item in the text and which are therefore an important means to create cohesion. See the following examples (1976: 275): (1) What shall I do with all this crockery? – Leave the stuff here; someone’ll come and put it away.1 (2) Can you tell me where to stay in Geneva? I’ve never been to the place. According to Halliday/Hasan (cf. 1976: 275), a ‘general noun’ in cohesive function is usually accompanied by the reference item the, which is anaphoric and has the effect that the whole complex ‘the + ‘general noun’’ functions like an anaphoric reference item. In the above example (1), the superordinate item the stuff, consisting of the definite article and a ‘general noun’, anaphorically refers to the subordinate item all this crockery. Similarly, in example (2), the superordinate item the place anaphorically refers to the subordinate item Geneva. As already indicated, Halliday/Hasan support their classification of ‘general nouns’ with carefully compiled textbook examples which do not tell us anything about actual language use. This shortcoming is the point at which my study becomes relevant: it tries to provide a comprehensive corpus-based account of the use of ‘general nouns’ in different spoken and written genres.

1

If not marked otherwise, emphasis in this and all other examples is added by the author of the present study: all linguistic items (e.g. noun phrases) with endophoric and exophoric reference considered in the examples are printed in italics. If noun phrases contain a ‘general noun’, it is printed in bold (this also accounts for ‘general noun phrases’ which only consist of a head). Broken underlining marks items or passages of text which are referred to. See Section 2.3 for a discussion of the notion ‘general noun phrase’.

21

1.1 Aim of the study While there are several studies of cohesion, there exist only very few studies of ‘general nouns’. Often, these items are neglected in descriptions of English, and even Halliday/Hasan (1976) only dedicate a short extract to them. To the best of my knowledge, Mahlberg (2005) was the first to conduct a corpus-linguistic study of ‘general nouns’. Based on the British National Corpus and the British English component of the Bank of English, Mahlberg (cf. 2005: 173) described ‘general nouns’ as highly frequent nouns which are characterised by local textual functions.2 The present study takes up the corpus-linguistic approach and investigates ‘general nouns’ in naturally occurring language data, more specifically in a corpus of over 300,000 words including spoken and written texts from different genres. The aim is to investigate whether those ‘general nouns’ introduced by Halliday/Hasan (1976: 274) actually occur in natural language data, and whether their frequencies differ across the media and genres represented by the corpus of the present study. This is done in a quantitative analysis. A second major aim of the present study is to investigate in a qualitative analysis how ‘general nouns’ are used in the corpus texts, in how far this use corresponds to the cohesive function described by Halliday/Hasan (1976: 274), and whether patterns not yet described by the authors can be revealed. In fact, it is assumed that ‘general nouns’, when used in naturally occurring language data, are much more flexible in form and function. Concerning form, the starting point for the present study is the assumption that ‘general nouns’, as any other kinds of concrete and abstract common nouns, occur in natural language data as heads of noun phrases with pre- and postmodifiers or a combination of both. This challenges Halliday/Hasan’s (1976: 274) classification of ‘general nouns’ as nouns with very general inherent meaning because pre- and postmodification, as we will see, affects the degree of explicitness of ‘general nouns’. Concerning function, it is assumed that ‘general nouns’ are not only used to establish anaphoric reference, but that the use of ‘general nouns’ fulfils many other functions, for example that of discourse markers which are neutral shells to encapsulate complex information or that of completely unspecific items without any kind of reference which are used intentionally as empty phrases. Apart from a detailed quantitative and qualitative analysis, the present study wants to show that a purely text-linguistic approach does not do justice to the variety of uses of ‘general nouns’ in naturally occurring language data. Therefore, 2

22

See Section 2.2 for a more detailed description of Mahlberg’s study.

also grammatical, pragmatic and cognitive aspects are taken into account in a comprehensive approach to ‘general nouns’.

1.2 Structure of the study The present study combines a theoretical and an empirical part. The theoretical part starts with a description of different concepts which form the basis for a study of ‘general nouns’. Section 2.1, which mainly relies on Halliday/Hasan (1976), describes the concepts of ‘cohesion’, ‘reference’, ‘reiteration’ and ‘general nouns’. Section 2.2 gives an overview of previous studies of ‘general nouns’ and comparable phenomena. Section 2.3 critically evaluates the status of ‘general nouns’. Section 2.3 marks the transition to Chapter 3, which takes into account grammatical, text-linguistic, pragmatic and cognitive aspects in order to provide a comprehensive description of ‘general nouns’. After a grammatical classification of these items in Section 3.2, Sections 3.3 and 3.4 show that ‘general nouns’ fulfil different communicative functions and can be used by the encoder to pursue certain communicative aims. Section 3.5 offers a cognitive approach to ‘general nouns’ and explains that these items represent broad cognitive categories and that these categories, according to whether the ‘general nouns’ are specified or not through the co-text and context, move at different levels of abstraction. Chapter 4 presents the framework for the quantitative and qualitative analysis of ‘general nouns’, which has been developed for the present study. It shows that the parameters modification and linkage have an influence on whether ‘general nouns’ are superordinate or whether they are subordinate, in other words, whether they are general or specific in meaning. This framework serves as a basic tool for the description of different uses of ‘general nouns’ across different media and genres. Chapter 5 presents the corpus of the present study, which includes over 300,000 words and consists of a written and a spoken part with each two genres: Supreme Court judgments and election manifestos for the written component, and parliamentary debates and face-to-face conversations for the spoken component. It describes according to which criteria which texts were compiled for the corpus and it discusses media- and genre-specific features concerning these corpus texts. Chapter 6 marks the beginning of the empirical part of the present study and describes the methodology for the analysis of ‘general nouns’. It explains which ‘general nouns’ are taken into account in the analysis, which are not, and on the basis of which factors this decision is made. Then it describes the methodology for the quantitative and the qualitative analysis of ‘general nouns’. As to the 23

qualitative analysis, Chapter 6 gives an overview of the different types of modification and different types of linkage, which are taken into account in the analysis of ‘general nouns’ in the present study. Chapter 7 presents the results of the corpus analysis. Starting with a quantitative analysis, Section 7.1 presents the frequencies and the distribution of the ‘general nouns’ under investigation and discusses media- and genre-specific differences and similarities across the corpus texts. Section 7.2 presents the qualitative analysis and determines the degree of specification and the different uses of ‘general nouns’ across the media and genres of the corpus. The present study concludes with a summary of the most important findings concerning ‘general nouns’ and in particular with a summary of the results of the quantitative and qualitative analysis (Chapter 8). Furthermore, the conclusion offers concrete suggestions for further research on ‘general nouns’ and generally discusses whether the findings of the present study have implications for further assumptions that go beyond the area of text-linguistics and the concept of ‘general nouns’.

24

2. A linguistic discussion of ‘general nouns’ Abstract: Chapter 2 defines cohesion and reference with a focus on the use of ‘general nouns’. Halliday/Hasan (1976) describe these items as endophoric devices with an inherently general meaning. A discussion of previous studies on ‘general nouns’ and comparable phenomena leads to a critical evaluation of these items, now called ‘general noun phrases’.

2.1 Halliday/Hasan’s approach to ‘general nouns’ 2.1.1 The concept of ‘cohesion’ Halliday/Hasan (1976) introduce the term general noun in connection with the concept of ‘cohesion’. The concept of ‘cohesion’ cannot be separated from the concept of ‘text’. According to Halliday/Hasan (cf. 1976: 1), a text is defined as a medium-independent unit of language in use. Medium-independency means that a text does not depend on any medium, but can be realised in both spoken and written form. Moreover, a text can be easily transferred from one medium to the other as “[w]e can write down what we hear and we can pronounce what we read; […]” (Esser 2009: 24). Halliday/Hasan (cf. 1976: 2) point out that the defining criterion for the notion of ‘text’ is meaning; thus, they regard text as a semantic unit of language in use, not a grammatical or structural unit of form.1 Accordingly, a text should not be defined in terms of formal criteria (such as text length), but in terms of semantic criteria. The following statement by Halliday/ Hasan (1976: 1) supports this idea: A text may be spoken or written, prose or verse, dialogue or monologue. It may be anything from a single proverb to a whole play, from a momentary cry for help to an all-day discussion on a committee.

Despite their distinct formal criteria (such as text length), the texts listed above can all be defined as texts because they constitute meaningful units which are prototypically devoted to one topic or a limited number of topics (cf. Esser 2009: 2). Moreover, Halliday/Hasan (cf. 1976: 2) emphasise that the essential criteria for the above listed texts to be defined as texts is the concept of ‘texture’. Texture can 1

When speaking of text as a unit of form, Halliday/Hasan (cf. 1976: 2) mean the entity of finite structures such as phrases and clauses because, strictly speaking, there is no structural unit above the clause. Thus, there is no structural unit of text. When speaking of a text as a semantic unit, this does however not deny that a text contains formal or structural references (e.g. through the use of pronouns).

25

be described as the property of being a text, in other words, the property of being a unified whole in contrast to a collection of unrelated sentences (cf. Halliday/ Hasan 1989: 72). Texture is provided by cohesion and coherence. According to Halliday/Hasan (1976: 4), cohesion can be described as follows: Cohesion occurs where the INTERPRETATION of some element in the discourse is dependent on that of another. The one PRESUPPOSES the other, in the sense that it cannot be effectively decoded except by recourse to it. When this happens, a relation of cohesion is set up, and the two elements, the presupposing and the presupposed, are thereby at least potentially integrated into a text.

According to Halliday/Hasan, cohesion is a semantic concept, which describes the dependencies (which can be of grammatical and of lexical nature) between elements on the text surface. Cohesion is set up when the decoder2 is required to look at the surrounding sentences for the interpretation of some textual element (cf. Hoey 1991: 3–4). For a demonstration of a cohesive tie, a single instance of a pair of cohesively related items (cf. Halliday/Hasan 1976: 3), see the following often quoted example from Halliday/Hasan (1976: 2): (1) Wash and core six cooking apples. Put them into a fireproof dish.

In the above example, the personal pronoun them in the second sentence refers back to the noun phrase six cooking apples in the first sentence. The decoder establishes a link between these two items, interprets them in reference to another and thus integrates the two sentences into a text (cf. Halliday/Hasan 1976: 2). Cohesion is expressed partly through the grammar and partly through the vocabulary. This is why Halliday/Hasan (cf. 1976: 6) distinguish grammatical cohesion from lexical cohesion. See the following two examples to illustrate this distinction (Halliday/Hasan 1976: 2–3): (2) Wash and core six cooking apples. Put them into a fireproof dish. (3) Wash and core six cooking apples. Put the apples into a fireproof dish.

In example (2), the cohesive relation is considered grammatical because a grammatical item, the pronoun them, is used to refer back to six cooking apples. In example (3), the cohesive relation is established through repetition, which is expressed by a lexical item, the noun phrase the apples in the second sentence. Note that Halliday/Hasan (cf. 1976: 6) themselves emphasise that the distinction between grammatical and lexical cohesion is not clear-cut but rather one of degree. Grammatical and lexical cohesion influence each other in a twofold way, 2

26

The present study uses the term encoder as a synonym for ‘speaker’ or ‘writer’ and decoder as a synonym for ‘hearer’ or ‘reader’.

as grammatical cohesion, in order to be effective, requires the support of lexical cohesion and vice versa. In a text, therefore, both types of cohesion typically move hand in hand (cf. Halliday/Hasan 1989: 82–83). For grammatical and lexical cohesion Halliday/Hasan each identify different sub-types. The types of grammatical cohesion are called reference3, substitution, ellipsis and conjunction. These types describe referential identity of either meaning, lexico-grammatical form or category (cf. Esser 2009: 41). The types of lexical cohesion are called reiteration (including (i) repetition, (ii) synonymy and (iii) superordination) and collocation. While reiteration describes the relation between lexical items based on (i) the sharing of all, (ii) essential or (iii) fewer features, collocation describes the relation between lexical items based on the complementarity or the transfer of features (cf. Esser 2009: 43). Note that for the purpose of the present study, only the concepts of ‘reference’ and ‘reiteration’ will be focused on more extensively (Sections 2.1.2 and 2.1.3). See the following figure for an overview of the concept of ‘cohesion’: Figure 2.1: The concept of ‘cohesion’ Cohesion

Lexical co es o cohesion

3

Supern ordination

Reiteration

Synonymyy

Collocation

Repititionn

Conjuncttion

Ellipsis

Substituttion

Referennce

Grammatical co es o cohesion

Note that the term reference is understood in different ways. Here, reference is understood as a specific type of grammatical cohesion where typically pronouns, determiners or comparative adverbs and adjectives are used to refer to a preceding or following item in the text (cf. Halliday/Hasan 1976: 43 ff., 57 ff., 76 ff.). See Section 2.1.2 for a detailed discussion of the term reference.

27

Each of these cohesive relations are established by means of linguistic devices – a variety of grammatical or lexical items, e.g. pronouns or full noun phrases. Note that these items do not have an inherent cohesive nature but only function as cohesive devices (cf. Halliday/Hasan 1989: 75). As noted above, the concept of cohesion is closely related to the concept of coherence as both concepts provide texture. Halliday/Hasan (cf. 1976: 23) point out that cohesion is a concept that is limited to the actual text, while coherence also involves the situational context. Cohesion describes the overt, linguistically-signalled relationship between textual elements and thus refers to the grammatical and lexical devices on the text surface that establish connections between text elements (cf. Tanskanen 2006: 7). Therefore, cohesion is based on the level of expression, while coherence refers to the underlying level of meaning. It is important to point out that cohesion can be regarded as a means of contributing to the coherence of a text, but it is by no means a sufficient factor. Tanskanen adds an important aspect of coherence by including the decoder-perspective. She states that the establishment of coherence depends heavily on the decoder and their understanding and interpretation of the encoder’s intentions. She concludes (Taskanen 2006: 21): Cohesion can be regarded as a property of the text, while coherence depends upon the communicators’ evaluation of the text. Cohesive devices, being on the surface of the text, can be observed, counted and analysed and are therefore more objective. Coherence, on the other hand, is more subjective, and communicators may perceive it in different ways.

The concepts of cohesion and coherence can be seen as independent yet intertwined (cf. Tanskanen 2006: 15). It is necessary for the study of cohesion and ‘general nouns’ to clarify at this point two other notions: the notion of ‘co-text’ and the notion of ‘context’. According to Sinclair (cf. 1991: 171), the notion of ‘co-text’ means, in any continuous text, the words that come on either side of a word or phrase under scrutiny, thus, the linguistic environment of that expression. The notion of ‘context’ means the general, non-linguistic environment of any language activity and includes the sociocultural background. Note that very often in linguistic literature, the notion of ‘context’ is used to mean both – the linguistic and the non-linguistic environment. The present study, however, will distinguish the notions of ‘co-text’ and ‘context’.

2.1.2 The concept of ‘reference’ The concept of ‘reference’ is dealt with in many linguistic disciplines which is why there is considerable variation in the meaning of the term reference. The two most important approaches, the semantic and the text-linguistic approach to reference, will be discussed here. 28

In semantics, the term reference is commonly understood as describing textexternal relations (cf. Esser 2009: 35). Here, the term reference describes the relation between a linguistic expression and an object of the extra-linguistic world. This can be illustrated with the often quoted example from Halliday/Hasan (1976: 2), which is repeated here as example (4): (4) Wash and core six cooking apples […].

In example (4), the noun phrase six cooking apples is used to refer to an object in the extra-linguistic world, the actual apples. This relation is text-external. A very influential approach to the notion of reference as a text-external relation has its origin in the model of the linguistic sign developed by Ogden and Richards in 1923. With their ‘semiotic triangle’, they show that there is no direct relation between a linguistic expression and the extra-linguistic object or referent. Instead, the relation is mediated by a concept or thought (cf. Ogden/ Richards 1923: 14–15). In other words: The concept of ‘reference’ describes the relation between the full linguistic sign (expression and concept) and an extra-linguistic object (cf. Lipka 2002: 57). In contrast to this rather abstract and de-contextualised understanding of the term reference, Lyons (1977) presents another widely accepted approach which describes a contextualised meaning of the term reference. He states that reference describes “[…] the relationship which holds between an expression and what that expression stands for on particular occasions of its utterance” (1977: 174). What Lyons describes here as reference is the relationship between a concrete expression and a single extralinguistic object. In text-linguistics, the term reference is also used to describe text-external relations (exophoric reference) as well as text-internal relations (endophoric reference). One very influential approach to reference is presented by Halliday/ Hasan (1976). In their work on cohesion in English, they describe three different though related phenomena that underlie the terms reference and refer: (i) text-external relations through exophoric reference, (iia) text-internal relations through endophoric reference in general, e.g. through the use of ‘general nouns’ or synonyms (lexical cohesion), and (iib) a specific type of endophoric reference through the use of mostly grammatical items (grammatical cohesion). See the following examples taken from Halliday/Hasan (1976: 33, 275, 55) for an illustration of these three understandings of reference: (5) That must have cost a lot of money. (6) Can you tell me where to stay in Geneva? I’ve never been to the place. (7) John has moved to a new house. He had it built last year.

29

In example (5), the demonstrative pronoun that exophorically refers to something that is present in the situation of the utterance; the reference relation is therefore text-external. The pronoun that signals that reference must be made to the context of situation (cf. Halliday/Hasan 1976: 33). In example (6), the noun phrase the place in the second sentence is used to refer back to Geneva in the first sentence. This link is set up on the text surface and is thus endophoric. Similarly, in example (7), the pronoun he refers back to John and the pronoun it refers back to the noun phrase a new house. This link is also endophoric; it illustrates reference as a specific type of grammatical cohesion. The following figure gives an overview of reference in semantics and text-linguistics: Figure 2.2: Different concepts of ‘reference’ Reference

Semantics

De-contextualised De contextualised understanding of the term reference: relation between a full linguistic li i i sign i andd an extra-linguistic object

Text-linguistics

Contextualised understanding of the term reference: relation between a concrete expresssion and a single extra-linguistic object

Exophoric p

Endophoric p

reference

reference

Text external relationship

Text internal relationship

(not cohesive)

(cohesive)

S ifi Specific reference Generic reference

Reference as a general means of cohesion (different devices, e.g. `general ` l nouns´, ´ synonyms etc.)

Reference as a specific means of grammatical cohesion (different devices, e.g. pronouns, comparatives etc.)

Figure 2.2 gives an overview of the approaches to reference in semantics and text-linguistics. It can be seen that in semantics, the term reference is generally understood as describing text-external relations. In text-linguistics, the concept of reference is understood as describing text-external relations (exophoric reference) and text-internal relations (endophoric reference). Exophoric reference, as understood by Esser (2009) and Halliday/Hasan (1976) is closely related to the text-external understanding of the term reference in semantics, which is why in Figure 2.2 these branches are connected with a dashed line. It will be shown that there are different types of exophoric reference (specific reference and generic reference) and that in Halliday/Hasan’s approach, exophoric reference is 30

typically established through deictic expressions which refer to the situational context of an utterance (cf. Halliday/Hasan 1976: 33). Halliday/Hasan’s understanding of endophoric reference implies either a very general concept or a specific cohesive device (cf. reference of types (iia) and (iib)). It is important to note that exophoric reference is not cohesive. Only endophoric reference is cohesive. Halliday/Hasan (1976: 37) state: Exophoric reference contributes to the CREATION of text, in that it links the language with the context of situation; but it does not contribute to the INTEGRATION of one passage with another so that the two together form part of the SAME TEXT.

Halliday/Hasan (cf. 1976: 18) note that there is no sharp line between the two concepts of ‘exophoric’ and ‘endophoric reference’, and Esser (2009: 35) concludes that “[…] endophoric and exophoric reference interact in the creation of a text world by a decoder.” It seems comprehensible that whenever a decoder hears or reads a text, their interpretation of it is always based on their knowledge of the world. We now turn to a more detailed discussion of exophoric and endophoric reference as understood in text-linguistics. It has been stated earlier that there are two types of exophoric reference, namely specific reference and generic reference. Some linguists distinguish further types of exophoric reference, e.g. nonspecific reference (cf. Biber et al. 1999). However, for the purpose of the present study, it will be sufficient to discuss specific and generic reference. For an illustration of these two types of exophoric reference, see the following examples taken from Quirk et al. (1985: 265 [original emphasis]): (8) A lion and two tigers are sleeping in the cage. (9) Tigers are dangerous animals.

The noun phrases a lion and two tigers in example (8) refer to a specific lion and to two specific tigers in the extra-linguistic world. This example shows that specific reference can be established by indefinite as well as definite noun phrases (cf. Quirk et al. 1985: 265). By contrast, in example (9), the noun phrase tigers generically refers to the class of tigers, i.e. to any possible member of this class (cf. Quirk et al. 1985: 265). Generic reference can be established with definite and indefinite noun phrases as well as with singular and plural noun phrases. For a more detailed discussion of generic reference, see Section 6.2.3.2. In Halliday/Hasan’s approach to exophoric reference (1976), the text-external relation is established through certain linguistic items which can only be interpreted with recourse to the situation of utterance. This is illustrated in the following example, which is repeated here as example (10) (Halliday/Hasan 1976: 33): 31

(10) That must have cost a lot of money.

In example (10), the information needed for the interpretation of the demonstrative pronoun that must be retrieved from the situation of utterance. Given that example (10) is from a conversation between two people, the demonstrative pronoun that probably refers to some entity in the environment of the decoder and encoder. Note that this type of exophoric reference involves the linguistic phenomenon of deixis. Deixis is the process in which an encoder establishes direct reference to the speech situation with the help of deictic or ‘pointing’ items such as pronouns, demonstratives and time and place adverbs. The interpretation of these deictic expressions depends on the shared situation (cf. Ehlich 1983: 85–86).4 In contrast to exophoric reference, endophoric reference is reference on the text surface. As mentioned before, Halliday/Hasan (1976) use the terms reference and refer to describe a general concept and to describe a certain type of grammatical cohesion. As a general concept, Halliday/Hasan (1976: 31) define endophoric reference as follows: There are certain items in every language which have the property of reference, […]; that is to say, instead of being interpreted semantically in their own right, they make reference to something else for their interpretation. […] These items are directives indicating that information is to be retrieved from elsewhere.

In the case of endophoric reference as a type of grammatical cohesion, this property of reference is limited to personal pronouns and determiners, demonstrative pronouns and determiners, and comparatives, which include adjectives and adverbs like same, other, such, likewise and differently (Halliday/Hasan 1976: 31, 38–39). In the case of endophoric reference as a general concept, this property also applies to lexical items such as noun phrases. See example (6) above for an illustration. Generally, in the case of endophoric reference, the source for the interpretation of linguistic items lies within the linguistic co-text. This kind of reference may be either anaphoric when it points backwards to the preceding text, or cataphoric when it points forward to the following text. Halliday/Hasan (1976: 49, 56) demonstrate this distinction with the help of the following examples: (11) There was a brief note from Susan. She just said ‘I am not coming home this weekend.’ (12) I would never have believed it. They’ve accepted the whole scheme.

4

32

This aspect is closely related to the distinction of two styles of referencing introduced by Bernstein (1971) and will be discussed in Section 3.3.3.

In example (11), the personal pronoun she in the second sentence refers back to Susan in the preceding sentence, thereby creating a cohesive relation between the two clauses. Example (12) is less straightforward. Here, the pronoun it refers forward to they’ve accepted the whole scheme and thereby links the two sentences. In doing so, it fulfils some kind of pointing function, as it announces that something will follow in the text which explains what it is that is hard to believe. At the same time, this example of cohesion presents a special case of ‘extended reference’ (cf. Halliday/Hasan 1976: 52). Unlike in the examples mentioned so far, the pronoun it does not refer to a particular person or thing, but instead refers to a complex process (cf. 1976: 52). Note that with this kind of reference the cohesive device can refer not only to a particular sentence, but possibly also to a larger portion of text. This aspect is important for the analysis in the present study, since ‘general nouns’ very often tend to make reference to larger stretches of text. When speaking about reference-relations, Halliday/Hasan (1976) do not make a clear distinction between the terms reference item and referring item. For the purpose of the present study, I will use the term referent for the extra-linguistic object, reference item for its verbal representation in the text and referring item for the linguistic item that refers to it. For a demonstration, see the following example taken from Halliday/Hasan (1976: 2), which is repeated here for convenience as example (13): (13) Wash and core six cooking apples. Put them into a fireproof dish.

In example (13), the personal pronoun them in the second sentence is the referring item, as it refers back to the noun phrase six cooking apples in the first sentence, which is the reference item. The referent is the object in the extralinguistic world, in this case, the apples.

2.1.3 The concept of ‘reiteration’ In contrast to grammatical cohesion, which describes the cohesive effect achieved through the choice of grammatical items, lexical cohesion describes the cohesive effect that is achieved by the selection of lexical items (cf. Halliday/Hasan 1976: 274). Halliday/Hasan (1976: 12) describe the concept of ‘lexical cohesion’ as follows: This form of cohesion is lexical […]; it consists in selecting the same lexical item twice, or selecting two that are closely related. The two instances may or may not have the same referent; but the interpretation of the second will be referable in some way to that of the first.

33

As has been shown in Figure 2.1, the types of lexical cohesion are called reiteration (including repetition, synonymy and superordination) and collocation. For the purpose of the present study, the present section will focus on the concept of reiteration which, according to Halliday/Hasan (1976: 278), can be defined as follows: Reiteration is a form of lexical cohesion which involves the repetition of a lexical item, at one end of the scale; the use of a general word to refer back to a lexical item, at the other end of the scale; and a number of things in between – the use of a synonym, nearsynonym, or superordinate.

The different types of reiteration have in common that one lexical item refers back or forth to another, to which it is related by having a common referent (cf.  Halliday/Hasan 1976: 278). In accordance to what has been discussed in terms of the threefold use of the terms reference and refer in Halliday/Hasan (1976) (cf. Figure 2.2 in Section 2.1.2), it might be argued that the cohesive effect of reiteration is based on the underlying principle of endophoric reference as a general means of cohesion, which in Section 2.1.2 has been described as reference of type (iia) of Halliday/Hasan’s approach to reference. Describing the different types of reiteration, Halliday/Hasan (cf. 1976: 279) claim that they lie on a scale with a very specific term at the one end (repeated item) and a very general term at the other end (‘general noun’) with only a pronoun being even more general. This scale shows a continuum or ‘cline’ of cohesive elements, which can be presented with the help of the following example taken from Halliday/Hasan (1976: 279) in a modified version: a) The ascent    b) The climb    (14) I turned to the ascent of the peak. c) The task  is perfectly easy. d) The thing    e) It 

Example (14) shows the range of paradigmatic choice of lexical items that can be used to establish lexical cohesion between the two sentences. Here, the noun phrase the ascent of the peak in the first sentence is the reference item, the interpretation of which is shown to be identical with that of the following lexical items (in brackets) to which it is related by (a) repetition (ascent), (b) synonymy (climb), (c) superordination (task), (d) the use of a ‘general noun’ (thing) or (e) the use of a pronoun (it). Thus, from the top of this scale to its bottom, the referring items range from specific to general. What can also be observed in the above

34

example is that a reiterated item is in most cases accompanied by the anaphoric referring item the. For the purpose of the present study, reiteration through the use of a superordinated term or a ‘general noun’ is particularly interesting. In general, the category of superordinates refers to any item “[…] whose meaning includes that of the earlier one; in technical terms, any item that dominates the earlier one in the lexical taxonomy” (Halliday/Hasan 1976: 280). As Salkie (cf. 1995: 16) states, it is in most cases the hyponym that is mentioned first and the superordinate term is then used to refer to it anaphorically. This fact also holds true for the superordinate noun phrase the task expressed in example (14) above. Here, the more specific noun phrase the ascent of the peak has the fuller, richer meaning and is followed by the superordinate noun phrase the task. According to Salkie (cf. 1995: 18), using ‘general nouns’ to establish lexical cohesion is an extreme instance of superordination. Referring back to example (14) above, the ‘general noun’ thing refers back to the more specific noun phrase the ascent of the peak and is thereby accompanied by the definite article which establishes a relation of co-reference between the two noun phrases. As Hoey (1991: 6) states, “[…] the boundary between reiteration by superordinate and reiteration by general word is extremely fuzzy; […]”, since ‘general nouns’ are themselves superordinate members in lexical hierarchies and are moreover being used in exactly the same way as superordinate terms (cf. Halliday/Hasan 1976: 279). Note that Halliday/Hasan (cf. 1976: 278–279) themselves do not clearly explain the difference between the use of superordinate terms and ‘general nouns’. They merely state that there is a difference in the level of generality with the ‘general noun’ being even more general than a superordinate term. Apart from this rather vague explanation, Halliday/Hasan use the terms general word and general noun synonymously and it is not always clear whether the concept of superordination or the concept of ‘general nouns’ or both is meant (cf. 1976: 278). When speaking of the concept of ‘general nouns’, the present study will stick to the term general noun.

2.1.4 The concept of ‘general nouns’ Halliday/Hasan give the following definition of ‘general nouns’ (1976: 274): The class of general noun is a small set of nouns having generalized reference within the major noun classes, those such as ‘human noun’, ‘place noun’, ‘fact noun’ and the like.

Although the class of ‘general noun’ is certainly fuzzy and has many possible members, Halliday/Hasan draw up a list of what they count as ‘general nouns’ 35

and give in brackets the names of the major noun classes to which they belong, such as human noun, fact noun, etc. (1976: 174): people, person, man, woman, child, boy, girl [human] creature [non-human animate] thing, object [inanimate concrete count] stuff [inanimate concrete mass] business, affair, matter [inanimate abstract] move [action] place [place] question, idea [fact]

Halliday/Hasan give the following example to illustrate the cohesive function of ‘general nouns’ (1976: 274–275): (15) Didn’t everyone make it clear they expected the minister to resign? – They did. But it seems to have made no impression on the man.

In this example, the ‘general noun’ man in combination with the definite article in the third sentence refers back to the specific item the minister in the first sentence and thus establishes a cohesive relation between the two. Like most ‘general nouns’ with a cohesive function, man is accompanied by the definite article the, which functions here as an anaphoric referring item (cf. Halliday/Hasan 1976: 275). As explained before, it establishes a relation of co-reference between minister and man, thereby making it clear that the two terms refer to the same person in the extra-linguistic world. Halliday/Hasan (cf. 1976: 274) emphasise that ‘general nouns’ as a means of establishing cohesion are items on the borderline between grammatical and lexical cohesion, as “[…] a general noun is itself a borderline case between a lexical item (member of an open set) and a grammatical item (member of a closed system)” (Halliday/Hasan 1976: 274). Traditional grammar distinguishes between two categories of words: the open-system words and the closed-system words. According to Jackson (cf. 1980: 61), members of the closed class, also called function words, are grammatical items such as numerals, determiners, prepositions, conjunctions and pronouns. As the name already implies, the closed class includes a small set of fixed members which hardly undergo any change over time. The open class, on the other hand, includes all lexical items, also called content words, such as nouns, verbs, adjectives and adverbs. This class has a large open-ended vocabulary which changes permanently as new members readily enter. The members of the open class bear the greatest load in terms of meaning, while the function of the closed-class members is oriented more towards the internal linguistic relationships (cf. Jackson 1980: 61). Halliday/Hasan 36

(cf. 1976: 275) state that a ‘general noun’ shows characteristics of a grammatical item (pronoun) as well as of a lexical item (noun). They argue that on the one hand, a ‘general noun’ is a lexical item, more specifically a superordinate member of a lexical set. It serves as a device for establishing lexical cohesion as the superordinated and more abstract ‘general noun’ refers anaphorically or cataphorically to a subordinated and more specific term within the same lexical set. On the other hand, Halliday/Hasan (cf. 1976: 275) state that the combination of a ‘general noun’ and a specific determiner, as e.g. in the man, is very similar to a pronoun which is a grammatical item. For an illustration, compare the following examples (Halliday/Hasan 1976: 275): (16) Didn’t everyone make it clear they expected the minister to resign? – They did. But it seems to have made no impression on the man. (17) Didn’t everyone make it clear they expected the minister to resign? – They did. But it seems to have made no impression on him.

As the two examples show, there is little difference in meaning between the two referring items the man and him and according to Bolinger (cf. 1977: 52), who has studied the use of pronouns and nouns, an alternation between a pronoun and a ‘general noun’ is common in the English language. Another reason why ‘general nouns’ can be described as linguistic items that have characteristics of both the open and the closed class of items is the fact that their inherent semantic load is often very low and thus similar to pronouns. Only through reference to the linguistic co-text and the extra-linguistic context is a ‘general noun’ filled with more semantic content and thus behaves more like a noun. This behaviour has been described by Ivanič (1991: 103), who states: “[s]emantically these nouns resemble pronouns in that their meaning is not self-contained. They have both a constant meaning and a variable meaning.” What Ivanič defines as the constant meaning of a word is its pure dictionary meaning which is de-contextualised. The constant meaning of a word is often very broad and unspecific and the dictionary even lists several possible meanings. See the extract from the dictionary entry for the ‘general nouns’ thing and things, taken from the online version of the Longman Dictionary of Contemporary English online (cf. http://www.ldoceonline.com/dictionary/thing, last access 09th November 2014): •   An idea, action, feeling, or fact that someone thinks, does, says, or talks about, or that  happens. •   An object that you are talking about without saying its name, or whose name you do  not know. •   Life in general and the way it is affecting people.

37

With this dictionary entry, it becomes clear that the constant meaning of ‘general nouns’ such as thing and things is very broad and unspecific. Only through reference to the co-text and the context of an utterance, a contextualised, specific meaning is attached to the ‘general noun’ (cf. Ivanič 1991: 95). Besides the special borderline status of ‘general nouns’, Halliday/Hasan (cf. 1976: 2) point out another characteristic of these items: ‘General nouns’ can sometimes be used by the speaker to bring an interpersonal element into the meaning expressing a certain attitude. Consider the following example from Halliday/Hasan (1976: 274–275), which was mentioned above and is repeated here for convenience as example (18): (18) Didn’t everyone make it clear they expected the minister to resign? – They did. But it seems to have made no impression on the man.

In example (18), the attitude conveyed through the use of the ‘general noun’ man is one of distance. Through the use of such a general and superordinate word, the speaker might want to emphasise that they dissociate themselves from the minister and probably also from what this minister stands for. In this case, the ‘general noun’ man carries a contemptuous overtone. The expression of interpersonal meaning and attitude through the use of ‘general nouns’ can also – as we shall see – be realised through modifying elements. This aspect will be further discussed in Section 7.2.4.1.

2.2 Previous studies of ‘general nouns’ As Halliday/Hasan (cf. 1976: 274) themselves remark and, for example, Partington (cf. 1998: 90) and Mahlberg (cf. 2005: 1) still confirm, ‘general nouns’ have until now received little attention in linguistic literature. They are mentioned only briefly in classic reference works such as A Comprehensive Grammar of the English Language by Quirk et al. (cf. 1985: 1056, 1067, 1260, 1321) or the Cambridge Grammar of English (cf. Carter/McCarthy 2006: 147 ff.)5, and are fully neglected, for example, in The Cambridge Grammar of the English Language by Huddleston and Pullum (2002). Furthermore, Halliday/Hasan (1976, 1989) do not pay them much attention in their later works on cohesion. Nevertheless, the term general noun is used in various contexts and it is often not clear whether or not the concept of ‘general 5

38

Note that Carter/McCarthy (cf. 2006: 147 ff.) only concentrate on two particular ‘general nouns’, namely thing and stuff, emphasising their means to express vagueness and hedge statements.

nouns’ as introduced by Halliday/Hasan (1976) is implied. Accordingly, there is terminological variance in the literature. Among those linguists who have focused on ‘general nouns’ and comparable types of nouns are: Bolinger (1977), Winter (1977), Ivanič (1991), Schmid (2000), Hunston/Francis (2000), Flowerdew (2003a, 2003b) and Mahlberg (2005). In his work “Pronouns and repeated nouns”, Bolinger (1977) discusses similarities between nouns and pronouns when they are used as referring items. This phenomenon was also addressed by Halliday/Hasan when talking about the similar position of ‘general nouns’ and pronouns on a scale of cohesive elements (cf. 1976: 279, Section 2.1.3.). In this context, Bolinger identifies “[…] a sizable vocabulary of nouns […], sharing coreferential functions with pronouns” (1977: 50). Bolinger states that an encoder, when referring to a preceding item, chooses a noun or pronoun depending on whether greater or lesser semantic complexity is appropriate (cf. 1977: 40). In accordance with Halliday/Hasan’s definition of a ‘general noun’ as “a borderline case between a lexical item (member of an open set) and a grammatical item (member of a closed system)” (1976: 274), Winter (cf. 1977: 2, 18–20) defines a class of open-system words including nouns, verbs and adjectives, which he calls ‘vocabulary 3 items’. The semantic properties of these items are similar to those of closed-system words (grammatical words) in sentence connection. Functioning as signposts, vocabulary 3 items play an important role for the expression of clause relations in English. These items directly or indirectly paraphrase the semantics of subordinators and connectors and thus explicitly name clause relations (cf. 1977: 2). Winter gives an example (1977: 21): (19) One condition for the success of the course is obvious. If the student likes the course, he will follow with enthusiasm.

Here, the vocabulary 3 item condition in the first sentence reifies the information that is given in the conditional clause in the second sentence (cf. 1977: 21). This phenomenon is taken up by Ivanič (1991), who investigates a group of nouns she calls ‘carrier nouns’. Similar to what was observed before, Ivanič states that these nouns behave like pronouns in that they can have two meanings. Besides a constant meaning, “these nouns frequently carry a specific meaning within their context in addition to their dictionary meaning” (1991: 95). Carrier nouns express an additional meaning depending on the context they refer to. Ivanič gives the following example (1991: 95 [original emphasis]): (20) Cut out the bottom and sides accurately from the plan measurements but leave sufficient margin for planning after fixing the boat. Use your two lengths of 8.0 ft. by 4.0 ft [sic] for this purpose, […]

39

In example (20), the carrier noun purpose in the second sentence is in addition to its dictionary meaning specified through the information given in the first sentence. Through reference to the preceding sentence, which is signalled by the demonstrative pronoun this, the noun purpose carries a specific meaning. While the works presented above concentrate on the open- or closed-system status of ‘general nouns’, there are several works that focus on the cohesive function of ‘general nouns’ and comparable nouns. These works focus on the signposting functions of such nouns which are only fully interpretable in reference to their co-text and context. Among these works is the discussion by Schmid (2000), who uses the term ‘shell nouns’ for any abstract noun that frequently occurs in two specific grammatical patterns (Schmid 2000: 3 [original emphasis]): (a) Determiner + (Premodifier) + Noun + postnominal that-clause, wh-clause or to-infinitive: The (deplorable) fact that I have no money. (b) Determiner + (Premodifier) + Noun + be + complementing that-clause, whclause or to-infinitive: The (big) problem was that I had no money. Schmid is interested in the linking function of shell nouns and investigates their “[…] potential for being used as conceptual shells for complex, proposition-like pieces of information” (2000: 4). At the same time, Hunston/Francis (cf.  2000: 185) are concerned with shell nouns within their grammatical approach to the classification of nouns, assuming that they are a ‘new’ class of nouns, which are identifiable on the basis of their behaviour. These nouns are unspecific and require lexical realisation in their immediate co-text. According to Hunston/Francis (cf. 2000: 185–186), shell nouns such as allegation, theory or fact are used with some kind of expansion in the surrounding text, indicating what the allegation, the theory, or the fact is. ‘General nouns’ – as we shall see in Section 7.2 – readily occur in the shell noun patterns described by Schmid (2000) and Hunston/Francis (2000). In the context of this discussion, Flowerdew (2003b) uses the term ‘signalling noun’ for “[…] potentially any abstract noun, the meaning of which can only be made specific by reference to its context” (2003b: 329). He names signalling nouns such as attitude, difficulty, process, reason, result, etc. (Flowerdew 2003b: 329–330). Flowerdew, who is particularly interested in pedagogic aspects, investigates the use of signalling nouns within different genres. He claims that a frequent use of signalling nouns in science and technology has the effect of facilitating the understanding of highly technical contexts for the learner (cf. 2003a: 41). Mahlberg (2005) takes a corpus theoretical approach to ‘general nouns’. Based on the British National Corpus and the British English component of the Bank of English, Mahlberg classifies ‘general nouns’ as highly frequent nouns which are 40

characterised by local textual functions (cf. 2005: 173). The frequencies of a certain group of ‘general nouns’ in a particular type of text suggest that these nouns share the same textual functions and that these textual functions characterise groups of texts (cf. 2005: 177). Based on their local textual functions, Mahlberg subcategorises ‘general nouns’ into three groups: ‘time nouns’, ‘people nouns’ and ‘world nouns’ (cf. 2005: 169). The overview of previous studies of ‘general nouns’ and comparable phenomena has shown that there is considerable variation in terminological use. What is common to the type of nouns discussed in the different works presented above is the fact that a semantically rather unspecific noun is being specified through reference to its co-text and context. This – as we shall see – is also true for ‘general nouns’.

2.3 Critical evaluation of the status of ‘general nouns’ As the overview of previous studies on ‘general nouns’ has shown, there are several comparable phenomena such as ‘shell nouns’, ‘carrier nouns’ or ‘signalling nouns’. All these nouns have in common that their inherent meaning is very general and that they, when used cohesively, are specified through reference to their co-text and context. In that way, all these different groups of nouns have similar textual functions. It is questionable, in what way ‘general nouns’ differ from these other groups of nouns. Halliday/Hasan (cf. 1976: 274–275, 279) describe that ‘general nouns’ are superordinate terms which are general in meaning and which are used to refer to a more specific item in the preceding or following text. That way, the use of general nouns is indeed very similar to that of shell nouns, carrier nouns or signalling nouns. Resulting from this, it is also questionable whether one should define a set list of 18 ‘general nouns’ as defined by Halliday/Hasan (cf. 1976: 274). Partington (cf. 1998: 96) suggests that ‘general nouns’ are an open set and this is also what the overview of previous studies on ‘general nouns’ and comparable phenomena suggests. Therefore, the list of ‘general nouns’ set up by Halliday/Hasan is rather an indication of the possibilities with many more potential members (cf. Partington 1998: 96). Interestingly, Halliday/Hasan (cf. 1976: 276) themselves do not consequently stick to their list of ‘general nouns’ but, in further discussions, speak of additional ‘general nouns’ such as idiot and fool. Moreover, it must be critically remarked that the selection of the 18 ‘general nouns’ made by Halliday/Hasan (1976: 274) seems rather heterogeneous and partly arbitrary. The listing of the ‘general nouns’ boy and girl next to child, accompanied by ‘general nouns’ such as matter and fact (but not e.g. issue) is not transparent. 41

Halliday/Hasan view ‘general nouns’ as a certain class of nouns that are characterised by their general meaning and their cohesive function (cf. Halliday/ Hasan 1976: 274). There are three points of criticism to this definition: firstly, it is criticised that ‘general nouns’ are described as constituting a class of their own. In my view, ‘general nouns’ should rather be seen as a concept than as a certain class of nouns. As has been illustrated when discussing comparable phenomena to ‘general nouns’ in Section 2.2, it is rather that certain nouns can be used in the ‘general noun’-sense which implies a cohesive use of a superordinate term referring to a subordinate term. This view also allows for other nouns to be used this way (see Partington 1998: 96). Secondly, it is criticised that ‘general nouns’ as described by Halliday/Hasan necessarily function cohesively by referring anaphorically or cataphorically to another item in the text. In my view, ‘general nouns’ do not necessarily occur in the ‘general noun’-sense. It is assumed that ‘general nouns’ occur in natural language data in many different forms and functions. Concerning form, it is assumed that ‘general nouns’, as any other kinds of concrete and abstract common nouns, occur in natural language data as heads of noun phrases of varying complexity (with and without modification). Therefore, it is better to speak of ‘general noun phrases’.6 Concerning function, it is assumed that ‘general nouns’ are not only used to establishing endophoric reference, but that the use of ‘general nouns’ fulfils many other functions. This is in line with Partington (cf. 1998: 104), who states that these items are more flexible in form and function than has often been suggested. Thirdly, it is criticised that ‘general nouns’ are necessarily general in meaning. Since ‘general nouns’ assumingly occur in natural language data with different kinds of modifiers and are assumingly used in different kinds of flexible patterns, we can proceed from the assumption that ‘general nouns’ are not necessarily general in meaning. Chapter 4 will discuss in detail which parameters influence the degree of specification of ‘general nouns’.

6

42

Partington (cf. 1998: 92) already proposes the term ‘general noun phrases’. However, he only refers to the fact that ‘general nouns’ are not only typically accompanied by the definite article (as stated by Halliday/Hasan 1976: 275), but also by determiners, longer determining phrases such as that sort of, or adjuncts like similar or like this. The present study proposes the term ‘general noun phrase’ because it assumes that ‘general nouns’ readily occur in natural language data with different kinds of pre- and postmodifiers in the form of phrases and clauses.

3. Approaches to general noun phrases across different linguistic disciplines Abstract: General noun phrases constitute a micro- and macro-linguistic phenomenon touching upon different linguistic disciplines. Chapter 3 gives a holistic and comprehensive approach of general noun phrases integrating grammatical, text-linguistic, pragmatic and cognitive aspects. The outcome is a workable definition of these items for the present study.

3.1 Introduction As Halliday/Hasan (cf. 1976: 274) have indicated, ‘general nouns’ are often neglected in linguistic literature. According to Mahlberg (cf. 2005: 1), this is due to the conceptual elusiveness of these items, which makes them very hard to grasp. Mahlberg also cites the borderline status of ‘general nouns’ between a noun and a pronoun (see Section 2.1.4) as a reason for their rather marginal position in the relevant linguistic literature. In my opinion, the problematic status of general noun phrases is also due to the fact that they constitute a phenomenon that touches upon different linguistic disciplines and that therefore needs to be considered from different linguistic perspectives. Accordingly, the present chapter offers an approach to general noun phrases which takes into account different linguistic disciplines relevant for a comprehensive description of these items. Starting from a micro-linguistic approach to general noun phrases, Section 3.2 grammatically classifies these items by means of word class membership as suggested by Quirk et al. (1985). Widening this micro-linguistic perspective, Section 3.3 offers a textlinguistic approach to general noun phrases, based on de Beaugrande/Dressler (1981). This approach regards text as a communicative occurrence in which the intentions of a text producer and the effects on a text receiver are of great importance. The study of general noun phrases in this context accounts for micro-structural features (such as the linking of linguistic items on the text surface) as well as macro-structural features (such as communicative functions of texts). That way, Section 3.3 represents the transition from a microlinguistic perspective to a macro-linguistic perspective on general noun phrases. Section 3.4 offers a macro-linguistic approach to general noun phrases dealing with pragmatic aspects in terms of Grice’s ‘Cooperative Principle’ (1975). It will be discussed how the use of general noun phrases can affect the participants’ communicative behaviour and thus influence the success of a communication 43

situation. In this context, the use of general noun phrases as a means to express vagueness will be discussed. Section 3.5 takes a cognitive approach to general noun phrases, assuming that these items reflect broad categories in the decoder’s mind. Depending on the degree of specification of the general noun phrase, these categories move at different levels of abstraction. As a conclusion, Section  3.6 summarises the major aspects of Chapter 3 and gives a workable definition of general noun phrases for the present study.

3.2 The grammatical approach to general noun phrases 3.2.1 Classification of nouns As the term already implies, general noun phrases can be defined as a certain type of noun phrase. According to the notional definitions of traditional grammar, the noun phrase in English is usually composed of four main parts: 1. the obligatory central part, the head, around which the other constituents cluster; 2. the determinatives preceding the noun, consisting of predeterminers, central determiners and postdeterminers; 3. the premodifiers which comprise all the items placed after the determinatives and before the head; and 4. the postmodifiers, comprising all the items placed after the head (cf. Quirk et al. 1985: 1238– 1239). Noun phrases vary greatly both in form and function. Concerning form, we must distinguish simple noun phrases from complex noun phrases. The noun phrase head is an obligatory constituent; it is the minimal requirement for the occurrence of a noun phrase (cf. Jackson 1980: 66). With the noun phrase head alone or with the noun phrase head preceded by determiners, the noun phrase is simple. With varying possibilities of adding pre- and postmodifier, the noun phrase can be very complex (cf. Biber et al. 1999: 230, 240). Concerning function, the noun phrase can be subject, object, complement of clauses and complement of prepositional phrases (cf. Quirk et al. 1985: 245). The most usual kind of head of a noun phrase is a noun. Alternatively, the head may be a pronoun of some kind, very commonly a personal pronoun, but also an indefinite pronoun, a possessive pronoun, or a demonstrative pronoun (cf. Jackson 1980: 66). Note that when a pronoun functions as head of a noun phrase, it cannot be premodified and it cannot have determiners. Generally speaking, nouns refer to ‘things’. These things include people, places or institutions, unique objects, classes of objects, feelings, and ideas (cf. Jackson 1980: 61). The class of nouns can be further categorised into subclasses, the number of which differs across linguistic theories. The classification presented in the present section is based on Quirk et al. (1985), Halliday (2004) and Sinclair 44

(2005). Nouns can be divided into six main groups: ‘proper nouns’, ‘common nouns’, ‘count nouns’, ‘non-count nouns’, ‘concrete nouns’ and ‘abstract nouns’. Figure 3.1, taken from Quirk et al. (1985: 247) in a slightly modified way, illustrates this classification: Figure 3.1: The classification of nouns

count common no n noun

non-count proper

concrete abstract concrete abstract

Proper nouns, which are written with initial capital letters, are mainly used for naming specific people, places, months, days and events. Examples are Shakespeare, Milwaukee, September, Thursday and Christmas (cf. Quirk et al. 1985: 288). These items usually have a unique denotation, which means that in the extra-linguistic world, a proper noun names a unique entity and thus refers to one single referent instead of a class of referents. It is important to distinguish the terms proper noun and proper name: while the former only refers to single words, the latter includes names which can consist of two or more words, such as Ayers Rock (cf. Halliday 2004: 325). At this point, the discussion cuts across the morphological simple/complex distinction, a phenomenon that deserves special attention. A deeper analysis of noun compounds as an example of morphological complex words, however, is not relevant for the present study and will therefore be neglected here. The second large group of nouns includes common nouns, which can be further divided into count and non-count nouns. Common nouns are generalised as a class of referents, which means that, in contrast to proper nouns, which name specific things, and unique entities, common nouns refer to general items and classes of entities. Examples are: book, table and brick (cf. Halliday 2004: 326). Count nouns, which form the largest group of nouns in English, are common nouns which refer to entities that can be counted. Accordingly, they can take the singular and the plural. Count nouns can be modified, for example by taking numerals. When used in the singular, these items can take a determiner, when used in the plural, they do not take a determiner because then they refer to something

45

in general (cf. Sinclair 2005: 6). Compare the following examples taken from Sinclair (2005: 6 [original emphasis]) in a slightly modified version: (1) He got into the car and started the motor. (2) They all live in houses.

In example (1), the count nouns car and motor take a determiner, namely the definite article. In example (2), the plural noun houses does not take a determiner because it does not refer to specific houses, but to houses in general. Non-count nouns are common nouns which denote an undifferentiated mass or continuum (cf. Quirk et al. 1985: 246). As opposed to count nouns, which refer to countable objects or events, non-count nouns refer to general things such as qualities, substances, processes and topics which cannot be counted. Non-count nouns only take the singular and are usually not used with numerals or determiners. Examples are: food, money, religion and happiness (cf. Sinclair 2005: 8). Cutting across the above-presented count/non-count distinction, there is a further subdivision of each of these noun groups into concrete and abstract nouns, which is based on semantic aspects. According to Quirk et al. (cf. 1985: 247), concrete nouns are physical entities which are accessible to the senses, observable and measurable such as gold, toy and pig. Abstract nouns refer to abstract objects, ideas or concepts that are typically not observable as well as not objectively measurable. Examples are difficulty, remark and music. Note that there is considerable overlap within the count/non-count group of nouns. A non-count noun can, in certain contexts, be used as a count noun. Consider the following example given by Quirk et al. (1985: 246 [original emphasis]): (3) The house is built of brick. (4) He used bricks to build the house.

While the singular noun brick in example (3) refers to non-countable material, the plural noun bricks in example (4) refers to a countable object.

3.2.2 Modification As has been explained in the preceding section, noun phrases are composed of the obligatory head around which the determinative and the optional pre- and postmodifying elements cluster. For a better understanding, see the following Figure 3.2 (based on examples taken from Quirk et al. (1985: 1239)) as an example of a complex noun phrase:

46

Figure 3.2: A complex noun phrase

determiner some qquantifier

premodifier p very

expensive

adjective pphrase j

head

ppostmodifier

office

furniture

in the room

noun

noun

pprepositional p pphrase

Figure 3.2 shows a complex noun phrase with the possible four noun phrase constituents (determiner, premodifer, head, postmodifer) and exemplary realisations of these constituents. The third constituent, the head, is the obligatory part of a noun phrase. It has already been discussed in the preceding section. The first constituent of the complex noun phrase is the determiner. Determiners are closed-class items such as articles, demonstrative and possessive determiners, and quantifiers (cf. Biber et al. 1999: 69–70). These items, as the name implies, determine the noun phrase, which means they specify the reference of a noun phrase (cf. Biber et al. 1999: 258).1 The second and the fourth constituent of the complex noun phrase are modifiers. With respect to form, modifying elements can be of any length, from words to phrases to clauses. With respect to function, noun phrase modification gives ‘descriptive’ information to the head (cf. Quirk et al. 1985: 65, 245). Here, we need to distinguish restrictive and non-restrictive modification. According to Quirk et al. (1985: 1239), restrictive modification indicates a limitation on the possible reference of the head that is being modified. In this case, the reference of the head is a member of a class which can be identified only through the modification that has been applied. Non-restrictive modification is additional information which is not essential for identification, because e.g. the referent of a noun phrase may be unique or has already been identified in the preceding context. For an illustration, compare the following examples taken from Quirk et al. (1985: 1239 [original emphasis]): (5) Come and meet my younger daughter. (6) Mary Smith, who is in the corner, wants to meet you.

1

While some linguists argue that determiners are premodifying elements (Halliday (2004), Jackson (1980)), others argue that determiners are not to be treated under premodification (cf. Quirk et al. (1985), Biber et al. (1999)). The present study is in line with Quirk et al. (1985) and Biber et al. (1999) and views determination as a phenomenon separate from modification. Since modification will be an essential criterion for the analysis of general noun phrases in the present study, the present section will concentrate on modification only.

47

The premodification in example (5) is restrictive as it identifies the noun phrase as one out of, for example, two daughters. The postmodification in example (6) is non-restrictive, as Mary Smith’s identity is independent of whether or not she is in the corner (cf. Quirk et al. 1985: 1239). Modifiers are classified into premodifiers and postmodifiers. Premodifiers comprise all the items placed before the head other than determiners. The major types are adjectives, or rather adjective phrases, participles, nouns and genitives. Postmodifiers comprise all the items placed after the head. The major types are prepositional phrases, non-finite clauses, relative clauses and complementations, such as complementing that-clauses (cf. Quirk et al. 1985: 1239, 1322). The discussion first turns to different types of premodifiers. Starting with adjectives, Sinclair (cf. 2005: 62) states that these items are used to identify a noun or to describe it in more detail. Some premodifying adjectives, or more precisely some premodifying adjective phrases, can take a degree adverb or intensifier such as quite, very, so, rather. Those that can take a degree adverb are called gradable adjectives, as opposed to non-gradable adjectives. Compare the following examples taken from Quirk et al. (1985: 1323, 435 [original emphasis]) in a slightly modified version: (7) (8)

Her really delightful family. An atomic scientist.

Example (7) displays the use of the premodifying gradable adjective delightful, which is itself premodified by the degree adverb really. Example (7) shows the use of an adjective phrase. The adjective in example (8) is non-gradable and can therefore not be intensified, as in *a very atomic scientist. In addition, a distinction can be made between inherent and non-inherent adjectives, or adjective phrases. Inherent adjectives characterise the referent of the head directly and describe an integral quality whereas non-inherent adjectives can be seen as an extension of the basic sense of the noun (Quirk et al. 1985: 435 [original emphasis]): (9) A firm handshake. (10) A firm friend.

In example (9), the adjective firm directly applies to the referent of the noun handshake. In example (10), it does not apply to the referent of the noun directly. Here, a firm friend is ‘a friend whose friendship is firm’. Besides the use of adjectives, a head noun can also be premodified by present participles, as for example in an approaching train, and past participles, as for example in a married man (Quirk et al. 1985: 1326–1327). 48

The third group of premodifying elements besides adjective phrases and participles is that of nouns, or noun phrases, and genitives. Similar to inherent adjectives, nouns classify the heads of noun phrases, as for example in dish cloth or, as has been shown in Figure 3.2, in office furniture. (cf. Quirk et al. 1985: 1330, 1239). Since noun premodifiers are so closely associated with their accompanying head noun, they are often regarded as compounds (cf. Quirk et al. 1985: 1330). However, as has been mentioned before in Section 3.2.1, it is not relevant for the present study to discuss noun-compounds in more detail. When using a genitive as premodifier, as for example in his fisherman’s cottage, the genitive fisherman’s describes the head noun, here, meaning that the cottage belongs or belonged to a fisherman (cf. Quirk et al. 1985: 1335–1336). We now turn to the discussion of postmodification. Different grammatical structures may be used to postmodify a noun phrase, the most common of which are: prepositional phrases, non-finite clauses such as infinitive clauses, present participle clauses, past participle clauses and appositive clauses, and finite clauses such as relative clauses and appositive clauses (cf. Quirk et al. 1985: 1239). The prepositional phrase is the most common type of postmodification in English and comprises the full range of prepositions (cf. Quirk et al. 1985: 1274 [original emphasis]). Consider the following examples: (11) This book on grammar (12) The meaning of this sentence (13) The house beyond the church

In each of these examples, the head noun is specified by the following prepositional phrase in italics. The second type of postmodification is the use of non-finite clauses such as infinitive clauses, present participle clauses, past participle clauses and appositive clauses. These clauses usually have no subject and are introduced by a nonfinite form of the verb (cf. Jackson 1980: 68, Quirk et al. 1985: 1271). See the following five examples for an illustration (Quirk et al. 1985: 1266, 1263, 1265 [original emphasis]): (14) (15) (16) (17) (18)

The man to help you is Mr Johnson. The dog barking next door sounded like a Terrier. A report written by my colleague appeared last week. Your duty to report the accident takes precedence over everything else. The freedom of holding an opinion […] is a human right.

Example (14) demonstrates the use of the non-finite infinitive clause to help you which postmodifies the noun phrase the man. Example (15) illustrates the use of a postmodifying present participle clause. The non-finite verb barking introduces 49

the post-modifying clause that refers back to the dog. Similarly, example (16) shows a non-finite past participle clause that refers back to the noun phrase a report. Example (17) illustrates the use of a postmodifying non-finite appositive clause. Non-finite appositive clauses occur with infinitive and present participle clauses (cf. Quirk et al. 1985: 1271), which is presented in examples (17) and (18). Note that the typical postmodifying function of appositive present participle clauses is as complement of a preposition, such as presented in example (18) (cf. Quirk et al. 1985: 1272). Such constructions will be discussed in more detail in Section 6.2.2.2. The third type of postmodification comprises finite clauses such as relative clauses and appositive clauses. Consider the following examples from Quirk et al. (1985: 1244, 1260 [original emphasis]): (19) The news that appeared in the papers this morning was well received. (20) The fact that he wrote a letter to her suggests that he knew her.

Example (19) displays the use of a relative clause which postmodifies the noun phrase the news. Here, the postmodifying relative clause is restrictive, which means that the applied information in the relative clause identifies the referent of the noun phrase head news (cf. Quirk et al. 1985: 1239). In example (20), the noun phrase the fact is postmodified by an appositive that-clause. In contrast to the relative clause, the appositive that-clause actually presents the complete content of the head noun instead of presenting additional information (cf. Biber et al. 1999: 645). Quirk et al. (1985: 1260, 1321) point out that appositive clauses are common with general abstract nouns such as fact, idea, proposition, reply, remark, answer, view, question, or duty. See also Schmid (cf. 2000: 3) on this subject. Post-modifying appositive clauses, whether finite or non-finite, are important for the present study of general noun phrases and will thus be further discussed in Section 6.2.2.2. Concerning the degree of explicitness associated with each type of modification introduced, Quirk et al. (cf. 1985: 1243) state that modification specifies the head noun to different degrees. According to Quirk et al. (1985: 1243, 1321), premodifiers are generally less explicit than postmodifiers. This is due to the fact that postmodifiers more explicitly express certain relations (e.g. through the use of verbs, verb tenses) which in premodifiers are left to be inferred by the reader. While the noun phrase an oil man is not very explicit, the noun phrase a man who produces oil or the noun phrase a man who advocates the use of oil is more explicit. Therefore, Quirk et al. (cf. 1985: 1243) interpret postmodification as a more explicit description of the same content of premodification. In this context, Jackson (Jackson 1980: 69), discussing the different types of postmodifiers, 50

states that “[…] the amount of detail included becomes less as one progresses from relative clauses through non-finite clauses to prepositional phrases.” The fact that different types of modifiers express different degrees of explicitness is an important aspect for the study of general noun phrases since modification, as has been explained before, is one factor that influences the degree of specification of general noun phrases. It can be assumed that different modifiers with different degrees of explicitness have different effects on the specification of general noun phrases (see Section 7.2).

3.2.3 Classification of general noun phrase heads From a grammatical perspective and with reference to the classification of nouns presented in Section 3.2.1, all of the general noun phrase heads investigated in the present study (Halliday/Hasan 1976: 274) are common nouns. Sixteen of the general noun phrase heads investigated are count nouns and two of them are non-count nouns. Seven general noun phrase heads are abstract and eleven are concrete. See the following table for a classification of general noun phrase heads: Table 3.1: The classification of general noun phrase heads General noun phrase heads common noun

Types of nouns non-count abstract count noun concrete noun noun noun x x

people

x

person

x

x

x

man

x

x

x

woman

x

x

x

child

x

x

x

boy

x

x

x

girl

x

x

x

creature

x

x

x

thing

x

x

x

object

x

x

stuff

x

business

x

affair

x

x

x

matter

x

x

x

move

x

x

x

x x x

x

51

General noun phrase heads common noun

Types of nouns non-count abstract count noun concrete noun noun noun x x

place

x

question

x

x

x

idea

x

x

x

As we can see in Table 3.1, there are ten concrete count nouns; these are people, person, man, woman, child, boy, girl, creature, thing and object. There are six abstract count nouns; these are affair, matter, move, place, question and idea. There is one concrete non-count noun, namely stuff, and there is one abstract noncount noun, namely business. In her study on nouns with both open- and closed-system characteristics, Ivanič (1991: 98) mentions the property of abstractness that is typical for nouns similar to ‘general nouns’: “Most of these nouns are recognizably abstract rather than concrete, but they differ from many other abstract nouns in that they are frequently countable.” As we have seen in Table 3.1, there are some examples of these kinds of nouns, which are affair(s), matter(s), place(s), question(s) and idea(s). However, the majority of general noun phrase heads is not countable abstract but countable concrete. The different classes of nouns presented in Section 3.2.1 differ concerning their degree of generality or specification. Sinclair (2005: 3) remarks: In the relation between language and the world, these different types of noun group show a range of choices between a very clear identification of someone or something, and a clear decision not to identify.2

According to Sinclair (cf. 2005: 2–3), a clear identification of someone or something means using a proper noun which names a particular person or unique entity. If you use a common noun with a general determiner such as a, some, or any, you are referring to a class of entities. The identification of something or someone is therefore less clear. The least specific way of referring to someone or something is by using an indefinite pronoun, such as nobody or anyone, which does not refer to any particular person or thing. Halliday (2004: 327) argues that a pronoun as the most general type of noun on a scale of possible referring items is the “limiting case of anaphoric generalization”. A salient characteristic of general noun phrase heads is certainly their semantic generality

2

52

Instead of using the term noun phrase, Sinclair (2005) uses the term noun group.

or unspecificity already mentioned by Halliday/Hasan (cf. 1976: 274). In this context, Halliday/Hasan also emphasise the borderline status of general noun phrase heads which in combination with the definite article are very similar to pronouns. According to Sinclair’s statement that the reference of different types of nouns ranges from general to specific, it can be assumed that general noun phrase heads, similar to pronouns, have general reference. See the following figure for an illustration: Figure 3.3: The classification of general noun phrase heads in terms of specification

count nouns

nouns

common nouns proper nouns

non-count nouns

conrete count nouns abstract count nouns concrete noncount nouns abstract noncount nouns

from specific to general general noun phrase heads

Figure 3.3 shows that proper nouns, denoting unique entities, are positioned at the very specific end of the scale while abstract non-count nouns, which denote undifferentiated substances or concepts, are positioned at the very general end of the scale. At the beginning of the present section, it has been explained that ten of the general noun phrase heads investigated are concrete count nouns, that six of them are abstract count nouns, that one of them is a concrete non-count noun and that one of them is an abstract non-count noun. Accordingly, Figure  3.3 shows that the general noun phrase heads investigated in the present study can be positioned at a point where the classes of concrete and abstract count nouns and concrete and abstract non-count nouns overlap. This point is located towards the general end of the scale. It is important to note that the above discussion of general noun phrase heads and their positioning presented in Figure 3.3 only refers to de-contextualised general noun phrase heads. As has been mentioned before, a treatment of general noun phrase heads within their linguistic co-text will take into account 53

modification. Accordingly, a contextualised view of general noun phrases then probably reveals a different degree of specification.

3.3 The text-linguistic approach to general noun phrases 3.3.1 Constitutive features of textual communication In their approach to text-linguistics, de Beaugrande and Dressler (1981) focus on the functions of texts (spoken and written) in human interaction. They define a text as a communicative occurrence which is characterised by seven features of textuality: cohesion, coherence,3 intentionality, informativity, acceptability,4 situationality and intertextuality (cf. de Beaugrande/Dressler 1981: 3–11). Considering the communicative function that texts intend to fulfil, the seven features of textuality entail factors such as cognition, planning and social environment. Note that de Beaugrande/Dressler (cf. 1981: 3) originally assumed that the seven features of textuality were defining criteria. The authors even speak of constitutive principles and state that a text can only successfully fulfil its communicative function and therefore be defined as text if all seven principles are satisfied. Today, this view seems too rigid and there has been some controversy on whether the seven features named by de Beaugrande/Dressler (1981) are in fact constitutive principles of textuality. However, it is generally recognised that a text has a communicative function which is determined by the encoder’s intention and the decoder’s expectation. I therefore refer to the seven constitutive principles of textuality as seven features of textual communication. The aspect of communicative functions of texts and the text producer’s intentions describe phenomena which go back to the works of J.L. Austin (1962). His theory of performative utterances states that any utterance is composed of the following three speech act components: the locutionary act (the actual uttering of words), the illocutionary act (the intention of the speaker) and the perlocutionary act (the effect of the utterance on the hearer) (cf. Austin 1962: 108). Austin’s theory (1962) was further developed by J.R. Searle (1969), who divides the locutionary act into an utterance act (the actual uttering of words) and a propositional act (conveying the content), so that he distinguishes four components of a speech act: the utterance act, the propositional act, the illocutionary

3 4

54

Note that the first two features listed, cohesion and coherence, refer to criteria of the actual text, only the remaining five features refer to the situation of communication. Note that the order of the features of textuality has been slightly changed. See footnote 14 for explanations.

act and the perlocutionary act (cf. 1969: 23–25). Later, Grice (1975) focused on the perlocutionary act and describes that the success of any communication situation heavily depends on the participant’s cooperative behaviour. Grice’s ‘Cooperative Principle’ is considered to belong to the area of pragmatics, which is why it will be dealt with in more detail in Section 3.4. The present section now turns to a detailed discussion of the seven features of textual communication. Cohesion as the internal linking of textual elements within the surface text is the first essential aspect. It is the linguistically overt relations between sentences that form the basis in the constitution of text. The use of general noun phrases as one means of establishing these relations plays an important role for the decoding of semantic links within the text and thus contributes to the creation of successful communication (see Sections 2.1.3 and 2.1.4). The second feature of textual communication is coherence (cf. de Beaugrande/ Dressler 1981: 4), which refers to the non-overt semantic linking of a text to the context of the situation. Coherence refers to the underlying level of meaning and concerns the interaction of the textual world and the decoder’s cognitive world. Depending on the decoder’s stored knowledge of the world, coherence is thus, in contrast to cohesion, decoder-dependent (or at least culture-specific), since the encoder’s intentions can be interpreted and perceived differently from decoder to decoder (or from culture to culture) (cf. Tanskanen 2006: 20–21). There can be a mismatch between the encoder’s and the decoder’s understanding and interpretation of an expression when both language users activate different chunks of knowledge (cf. de Beaugrande/Dressler 1981: 85). For example, when using the general noun phrase the man to refer back to the minister (cf. Halliday/Hasan 1976: 274– 275), both language users must have the same knowledge about mental categories and subcategories in order to correctly interpret that a minister is a man and that man can be used as a superordinate term for minister. This aspect cuts across observations of cognitive linguistics and will therefore be discussed in more detail in Section 3.5, when dealing with a cognitive perspective on general noun phrases. The third feature of textual communication, intentionality, describes the intentions the encoder follows concerning, for example, the distribution of knowledge or the attainment of a certain aim (cf. de Beaugrande/Dressler 1981: 7). The pursuing of the encoder’s intentions can influence the cohesion and coherence of a text. Coherence establishes continuity of senses among the decoder’s knowledge (cf. de Beaugrande/Dressler 1981: 84). This continuity of senses is activated by the referring item, for example a general noun phrase, so that a referent, once correctly identified, may pass longer stretches of text. The following example taken from a Supreme Court judgment demonstrates this: 55

(21) “The real question is, were these payments genuinely director’s remuneration? […] In the absence of any evidence of actual motive, the court must, I think, look at the matter objectively and apply the standard of reasonableness.” [J4-14-1]5

The decoder needs to recognise that the general noun phrase the matter refers back to the underlined portion of text and simultaneously acts as a starting point for the following development in the discourse. The general noun phrase reifies the question that has arisen at that point in the judgment and thus serves as a starting point for further elaborations on the propriety of director’s remuneration that have been paid. Ivanič (1991: 104), who observes this function in reference to ‘carrier nouns’, states that “[t]hese nouns ‘encapsulate’ preceding meaning and turn it into ‘given’ information in subsequent discourse”. This way, general noun phrases can help the encoder to encapsulate complex contents and thereby give propositional density to a text. In the above example, the decoder is able to identify the reference so that the cohesion and coherence of the text are not disturbed. However, when an encoder’s intention is to be vague and unspecific about something, this might be achieved through reference that the decoder is no longer able to identify. This can be achieved, for example, through ‘remote reference’ which means that too much text lies between the referring general noun phrase and the reference item. Then, the decoder might not be able to establish a cohesive link between both items which has the consequence that the general noun phrase remains indefinite (cf. Partington 1998: 91). See the following example from a Supreme Court judgment: (22) As this summary suggests, their Lordships appear to have been much influenced by their perception that the late Professor Halliday, whom they rightly described as “the architect of the Act”, considered […]. In that event it would also be worth bearing in mind the observation of the Earl of Halsbury LC, that the worst person to construe a statute is the person who was responsible for its drafting, […]. [J2-2-7]

In example (22), the general noun phrase the person who was responsible for its drafting remotely refers back to the noun phrase the late Professor Halliday, whom they rightly described as “the architect of the Act”. There are eleven smaller paragraphs which pass three and a half pages of text in between the referring general noun phrase and the reference item. Therefore, it is not clear whether the decoder can identify Professor Halliday as the reference item. The encoder’s intention for such remote reference might be that they do not want to be impolite – offending Professor Halliday in referring to him directly as the worst 5

56

The coding system used in the present study will be explained in detail in Section 6.2.1.

person to construe a statute. The third feature of textual communication touches upon pragmatic observations concerning communicative intentions of text producers and will thus be taken up again in Section 3.4. The fourth6 feature of textual communication, informativity (cf. de Beaugrande/Dressler 1981: 8–9), is concerned with the informational load of an utterance and thus ties in with the encoder’s aforementioned intentional use of general noun phrases to convey either much or little semantic content. Depending on the extent to which an utterance is specified, the encoder is confronted with a large amount of expected and known information as opposed to unexpected and unknown information (cf. de Beaugrande/Dressler 1981: 9). Particularly low informativity, i.e. a small amount of new information, is likely to be disturbing, causing the loss of the decoder’s attention or even their rejection of the text (cf. de Beaugrande/Dressler 1981: 9). This aspect is interesting for the study of general noun phrases as these items can be used to carry a great load of semantic content as well as very little semantic content. See the following examples for a comparison. Example (23) is from the Liberal Democrat manifesto and example (24) is from the Labour manifesto: (23) Only Liberal Democrats will break up the banks and start building things again, […]. [LDM-9-1] (24) Strong community life also depends on protecting the places in which people come together. […] we are investing £235 million to create new or refurbished play spaces and adventure playgrounds. We will protect the Post Office network, […]. The local pub and social club are also hubs of community life. [LM-16-9]

The general noun phrase things in example (23) conveys very little information, as it is not specified what things refers to, not even in the wider co-text, which cannot be presented here. By contrast, the general noun phrase the places in which people come together in example (24) conveys much more information, as it is specified through cataphoric reference to the underlined noun phrases, i.e. the kind of places that are talked about. Chapter 4 will discuss in detail the aspects that influence the load of information of general noun phrases. The fifth feature of textual communication, acceptability, concerns the decoder’s attitude towards the acceptability of the utterance. An utterance is acceptable 6

Note that the order of the features of textuality has been slightly changed. While in de Beaugrande/Dressler’s discussion (cf. 1981: 7–8) acceptability is the fourth and informativity the fifth feature, in my presentation they have changed positions. This is due to the fact that informativity (e.g. a very small amount of information) can influence the acceptability (e.g. the decoder cannot follow the argument and thus finds the utterance disturbing), and not vice versa.

57

if it constitutes a cohesive and coherent text which is relevant for the receiver (cf. de Beaugrande/Dressler 1981: 7). Concerning the aspects of relevance and text-reception, this standard of textuality is closely related to pragmatic aspects of communication, such as the ‘Cooperative Principle’ introduced by Grice (1975), and will therefore be discussed in more detail in Section 3.4. The sixth feature of textual communication, situationality, deals with the degree of relevance of a certain utterance or text to a situation of occurrence (cf. de Beaugrande/Dressler 1981: 9). A text is interpreted according to the situation in which it is used. Whether a text is suitable or not depends to a great extent on its “[…] believability and relevance to the participants’ outlook regarding the situation.” (de Beaugrande/Dressler 1981: 179). Imagine, for example, a politician giving a speech during an election campaign. The audience will probably have certain expectations regarding the content of that speech and its relevance to current political issues and the forthcoming elections. A speech that does not meet these expectations of the audience (for example if it has little relevance to current political topics), would not be acceptable and thus make successful communication difficult. It can be assumed that the use of very unspecific general noun phrases, such as the general noun phrase things in example (23), can lead to a rejection of the text by the decoder when they feel that the utterance conveys too little relevant information. The seventh feature of textual communication, intertextuality, states that the production and reception of one text depends on the decoder’s knowledge of other previously encountered texts (cf. de Beaugrande/Dressler 1981: 182). A decoder’s knowledge of previous texts, contexts and situations enables them to interpret a present text. However, this knowledge also raises expectations on a present communication. These matters again touch upon pragmatic aspects of communication situations and also cognitive aspects of text reception and will thus be taken up again in Sections 3.4 and 3.5.

3.3.2 Regulative features of textual communication Besides the seven features of textual communication mentioned before, de Beaugrande/Dressler (1981) name three further features of textual communication which influence the communicative success of texts. The regulative features of textual communication, which include the efficiency, the effectiveness, and the appropriateness of texts, can be defined as follows (de Beaugrande/Dressler 1981: 11): The efficiency of a text depends on its use in communicating with a minimum expenditure of effort by the participants. The effectiveness of a text depends on its

58

leaving a strong impression and creating favourable conditions for attaining a goal. The appropriateness of a text is the agreement between its setting and the ways in which the standards of textuality are upheld.

According to the above definition of efficiency, an encoder, but also a decoder, is meant to preferably put the least possible effort and endeavour into the production and reception of the text. The degree of effectiveness of a text determines whether a text is successful in influencing the decoder and thus attaining a communicative goal. Both regulative features of textual communication are closely related: the effectiveness of a text can depend on the degree of efficiency. There is a close relation between the efficiency-principle and the textual feature of intentionality, which states that the encoder determines the degree of cohesion, coherence and informativity depending on their communicative aims (cf. Section  3.3.1 above). The effectiveness-principle is related to the textual feature of acceptability, which has to do with the decoder’s attitude and expectancy towards the relevance of an utterance. The encoder always faces the challenge to strike a balance between the efficiency of a text and its effectiveness, as the aim is to obtain maximum reward for minimum effort (cf. de Beaugrande/Dressler 1981: 170). The criterion of efficiency is manifested in the use of cohesive devices to establish reference, such as general noun phrases. The use of a general noun phrase as a referring item can reduce the text-producer’s effort and correspondingly raise the text’s efficiency because a general noun phrase, as has been shown before, can be used to encapsulate larger portions of semantic content. There is, however, a ‘trade-off ’ between propositional density and clarity of thought (cf.  de Beaugrande/Dressler 1981: 64): If the decoder is not able to identify or locate the reference item because, for example, the referring item passes too much text, then the granted effectiveness of the text is lost again. This aspect ties in with the aforementioned intentions of an encoder. As mentioned before, this aspect will be taken up again in Section 3.4. The third regulative principle, appropriateness, regulates the above-mentioned trade-off situation between the efficiency and the effectiveness of a text. Thus, a text must provide a certain degree of clarity (achieved through effort of the producer) in order to be successful (effective). Only then is the text appropriate to the communication situation and the participants’ roles (cf. de Beaugrande/Dressler 1981: 170).

59

3.3.3 Restricted and elaborated style of referencing The above mentioned aspect of appropriateness is connected to the sociolinguistic terms ‘restricted code’, ‘elaborated code’ and ‘common ground’ that go back to Bernstein and his work Class, Codes and Control (1971). Bernstein (1971), originally interested in the correlation of social class and language use, introduces two styles of referencing which have led to the recognition of two codes: the restricted style of referencing and the elaborated style of referencing.7 According to Bernstein (cf. 1971: 76–77), the difference between the two language-uses lies in the syntactic options that speakers use to organise meaning. He explains that the elaborated style of referencing offers a wide range of alternatives while the restricted style of referencing has only a limited range of options in that respect. The restricted style of referencing is a semantic mode of communication characterised by little explicit naming of things or rather entities (cf. Halliday/Hasan 1976: 36). Instead, the restricted style of referencing heavily relies on exophoric reference in that it strongly depends on the text-external context of situation and the participants’ shared experience. Halliday/Hasan demonstrate this language use by quoting the summary of a picture story given by five-year-old children (1976: 35): (25) They’re playing football and he kicks it and it goes through there it breaks the window and they’re looking at it and he comes out and shouts at them because they’ve broken it so they run away and then she looks out and she tells them off.

In this example, the encoder frequently uses pro-forms such as they, he, it and there to refer to things in their environment. The surrounding objects are treated as mutually accessible. This summary of events is therefore closely tied to its context without which the listener could not interpret what is said (cf. Halliday/ Hasan 1976: 35). A restricted style of referencing is characterised by a large amount of exophoric reference and deictic expressions. In contrast to the limited alternatives of syntactic organisation with the restricted style of referencing, the elaborated style of referencing is characterised by greater lexical flexibility in the use of referring items. According to Esser (cf. 2009: 36), this is manifested in the use of lexically filled noun phrases which function as referring items. Halliday/Hasan (1976: 35) provide a demonstration

7

60

Note that Bernstein (1971) originally speaks of ‘restricted code’ and ‘elaborated code’. The present section will stick to Esser (2009) who prefers to speak of ‘restricted style of referencing’ and ‘elaborated style of referencing’.

of the use of the elaborated style of referencing in the equivalent version of a summary of the picture story: (26) Three boys are playing football and one boy kicks the ball and it goes through the window and the ball breaks the window and the boys are looking at it and a man comes out and shouts at them because they’ve broken the window so they run away and then that lady looks out and she tells them off.

This summary contains much more explicit naming. Instead of pro-forms the encoder makes use of a wider range of lexical alternatives (the ball, the window, that lady) in order to refer to things. Therefore, this version is independent of a shared situation between the communication partners, as the encoder properly introduces new entities in the text by using explicit noun phrases (cf. Esser 2009: 37). When distinguishing the two types of language use, note that the term ‘restricted’ might be misleading because it may imply an evaluation. According to Halliday/Hasan (cf. 1976: 35), the restricted style of referencing is no less grammatical than the elaborated style of referencing and there is no significantly greater amount of information given with the elaborated style of referencing. That is to say there is no difference as to what is said but in how it is said. As already mentioned at the very beginning of the present section, the distinction between restricted and elaborated style of referencing is rather one of appropriateness. The restricted style of referencing is typical of spoken face-to-face communication, where the participants share a common ground. This language use is efficient as a great amount of meaning is conveyed with a few words, each of which points the decoder to a lot more information which remains unsaid. In terms of appropriateness, the restricted style of referencing only becomes restricting if it is transferred to a context in which such an amount of encoding and encapsulating is inappropriate (cf. Halliday/Hasan 1976: 36). The elaborated style of referencing, on the other hand, is much more explicit in that it names things. It is thus appropriate for written communication, where the circumstances set the limit for using expressions that point to the context. In order to be understood, the encoder must elaborate on things. Note that the distinction between restricted and elaborated style of referencing was later described by Kay (1977) as non-autonomous and autonomous style. While a non-autonomous style (restricted style) depends on the encoder and decoder sharing the communication situation, an autonomous style (elaborated style) is independent of that situation (cf. Kay 1977: 21–22, 24). Concerning the different styles of referencing discussed above, the use of general noun phrases is relevant particularly in terms of a restricted style of 61

referencing typically found in conversations. See the following example taken from the conversation corpus for an illustration: (27) PS02B: So anyway Zoe’s gonna have some. And they fitted her Ange so I told you I bet she put them on first. PS029: Yeah because they were baggy. PS02B: She, I told Jenny Yeah I told you. She must have stretched them. That’s what she did. PS029: Cos the box was open wasn’t it? PS02B: Mm. PS029: And one of the things were. [C1-9-8] (28) PS02E: Oh have you got a lighter now? PS029: No. I’ve got matches behind you Sue. Up on the PS02E: Oh. Why do you PS029: Over there Sue. Up on that thing. [C1-9-11]

Examples (27) and (28) demonstrate the use of general noun phrases with a restricted style of referencing. In example (27), the simple general noun phrase one of the things is used to refer to something that is present in the situation of utterance. The same holds for the general noun phrase that thing in example (28). Here, this pointing function is overtly expressed through the use of the demonstrative pronoun that.

3.4 The pragmatic approach to general noun phrases 3.4.1 The Cooperative Principle In his article “Logic and Conversation”, Grice (1975) discusses the pragmatic aspects of communication and emphasises the participants’ roles in conversation. He states (1975: 45): Our talk exchanges do not normally consist of a succession of disconnected remarks, […]. They are characteristically, to some degree at least, cooperative efforts; and each participant recognizes in them, to some extent, a common purpose or set of purposes, or at least a mutually accepted direction.

What Grice describes here is that the success of communication heavily depends on the participants’ behaviour. According to Grice, the conversational partners must cooperate in that they both accept one another to be understood in a particular way. The cooperative efforts of the participants can partly account for ‘conversational implicatures’ which means that an encoder can imply in their utterance things that they need not verbalise because the decoder is assumed to 62

be cooperative and willing to follow what the encoder says and means (cf. Grice 1975: 43–44). The cooperative behaviour of the participants is formulated by Grice in terms of the ‘Cooperative Principle’, which he describes as follows (1975: 45): Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.

Grice emphasises that this principle is by no means a set of prescriptive rules as to how participants must behave in conversation, but rather a set of descriptive norms as to how participants actually behave and are usually expected to behave in conversation. The ‘Cooperative Principle’ is divided up into four maxims which Grice calls the maxims of ‘quantity’, ‘quality’, ‘relation’ and ‘manner’ (cf. Grice 1975: 45). The first maxim, the ‘maxim of quantity’, relates to the quantity of information that is required in an utterance and includes the following two statements: “1. Make your contribution as informative as is required (for the current purposes of the exchange)”, and “2. Do not make your contribution more informative than is required” (Grice 1975: 45). Note that this maxim can be related to the textual feature of informativity which also deals with the informational load of an utterance (de Beaugrande/Dressler (cf. 1981: 8–9; see Section 3.3.1)). The second maxim, the ‘maxim of quality’, refers to the truth-value of an utterance and states: “Try to make your contribution one that is true” (Grice 1975: 46). It can be subcategorised into two submaxims which state: “1. Do not say what you believe to be false”, and “2. Do not say that for which you lack adequate evidence” (Grice 1975: 46). The third maxim, the ‘maxim of relation’, concerns the focus of relevance of the utterance and it simply says: “Be relevant” (Grice 1975: 46). This maxim is related to the textual features of acceptability and situationality which deal with the relevance of an utterance for the decoder and the relevance of an utterance for the situation (cf. de Beaugrande/Dressler 1981: 7, 9; see Section 3.3.1). The fourth maxim, the ‘maxim of manner’, differs from the preceding three maxims in that it does not refer to what is said, but to how it is said. Its supermaxim states: “Be perspicuous” (Grice 1975: 46), including the four submaxims: “1. Avoid obscurity of expression”, “2. Avoid ambiguity”, “3. Be brief (avoid unnecessary prolixity)”, and “4. Be orderly” (Grice 1975: 46). The four maxims presented by Grice are closely connected to the purpose that talk exchange is primarily employed to serve as a maximally effective exchange of information (cf. Grice 1975: 47). In short, the ‘Cooperative Principle’ requires 63

the conversation to be precisely informative, truthful, relevant and explicit. In this context, it is important to note that conversation participants do not always follow these maxims strictly. In fact, Grice (cf. 1975: 49) states that very often one or more of the maxims are violated rather than followed, either because the participants deliberately and blatantly violate them or because the participants are faced with the dilemma of being unable to fulfil a certain maxim without disregarding another. Since conversation participants are expected to cooperate in communication, a speaker who disregards or violates a maxim can be interpreted by their communication partner as conveying implicit meaning in their utterance. Yet, Grice (cf. 1975: 50) emphasises that the conversational partners must be capable of working out this implicature, for example by reference to the co-text or context. The use of general noun phrases in terms of the ‘Cooperative Principle’ is interesting because these items, as has been mentioned before in Sections 3.3.1 and 3.3.2, can influence the degree of informativity and relevance, and thus, the degree of effectiveness in an utterance. As a result, a text, or in particular, a conversational exchange might no longer be acceptable for the conversation partners. For example, when completely unspecific general noun phrases are used in a text without reference to the preceding or following text, the general noun phrases convey very little meaning. Using such a general noun phrase, the encoder violates the maxim of ‘quantity’ because they do not give an appropriate amount of information. As a result, the decoder might not be able to identify the referent, might not be capable of working out an implicature and might not be able to understand an utterance altogether. This can be the case, for example, when the encoder is faced with the dilemma of wanting to fulfil the ‘maxim of quality’, which expects them to tell the truth. Especially in the context of political issues, telling the truth sometimes seems difficult, so encoders often decide to solve this dilemma by violating the ‘maxim of quantity’. By withholding certain information, the encoder avoids lying, which saves them from facing the problem of admitting errors or compromising themselves. Also consider the use of general noun phrases in terms of the maxims of ‘relation’ and ‘manner’. As mentioned above, very unspecific general noun phrases convey little meaning and only inexplicit and broad information. A frequent use of such general noun phrases can lead to a rejection of a text by a decoder because the utterance is no longer relevant for them. Then, the ‘maxim of relation’ has been violated. Similarly, the encoder can use unspecific general noun phrases to make their utterance vague and long-winded instead of concise and explicit, which would display a violation of the ‘maxim of manner’. See the following 64

example taken from a parliamentary debate in the House of Commons of the British Parliament: (29) Nic Dakin (Scunthorpe) (Lab): The European Food Safety Authority has concluded that the major factor causing poor welfare in dairy cows is genetic selection to produce high yields. Given proposals to intensify milk production for higher yields, such as those planned at Nocton, will the Secretary of State agree urgently to review the welfare code for dairy cows in the UK and to meet a delegation of cross-party MPs and non-governmental organisations to discuss how her Department can ensure that its code takes into account the latest scientific advice and ensures that any new dairies do not compromise cow welfare? Mr Paice: I am happy to meet the hon. Gentleman and others to discuss the matter, but I assure him that the Department puts welfare at the top of our agenda, as I hope our record to date shows. However, we must be guided by science, which is why I am looking forward to the results of the three-year study being carried out in Scotland on intensive dairy farming, and the work at Harper Adams university college on the same issue. After that, we will be in a better position to know the precise answers. I remind him that the Farm Animal Welfare Council, which the Department normally listens to, has said clearly that welfare is a function more of management that of scale. [D3-14-5]

Nic Dakin from the Labour party has heard of proposals to intensify milk production for higher yields and therefore asks the Secretary of State for Environment, Food and Rural Affairs to review the welfare code for dairy cows in the UK. He also wants the Secretary of state to discuss how the Department ensures that cow welfare is not compromised in the future. Mr Paice, who is Minister of State in the Department for Environment, Food and Rural Affairs, does not give a clear answer to the two questions by Mr Dakin. Mr Paice uses the unspecific general noun phrase the matter to refer to the problems raised by Mr Dakin, such as the important demand to review the code for cow welfare and to ensure cow welfare in the future. This might simply have economic reasons since the general noun phrase the matter helps to encapsulate complex facts and thus makes the utterance denser. However, by using such a neutral and general expression as the matter, Mr Paice avoids a critical evaluation of the problems related to cow welfare in the course of which he might have to admit the Department’s failures. The use of the general noun phrase the matter neutralises and plays down the problematic issue of cow welfare. By using the general noun phrase the matter, Mr Paice violates the maxim of ‘quantity’ as he gives very little information (for example as opposed to a complex expression such as the problem of cow welfare or the need to ensure cow welfare). Mr Paice does not give a clear answer to Mr Dakin’s questions. Instead of affirming to review the welfare code and to meet a delegation to discuss how to 65

ensure the welfare code in the future, or instead of denying it, Mr Paice answers with irrelevant facts such as that the Department puts welfare at the top of their agenda, or that cow welfare is a function more of management than of scale. Mr Paice violates the maxim of ‘relation’ because this is not a relevant answer to Mr Dakin’s questions. This is not only due to irrelevant facts listed by Mr Paice but also through the use of the general noun phrase the matter. Even if the reference to the problems of cow welfare can be understood by the communication partner, the matter is used here as a means of concealing or even downgrading because the speaker avoids spelling out displeasing issues. This makes it not appropriate in the sense of Grice’s maxims.

3.4.2 Vagueness in language As described in the preceding sections, the use of very unspecific general noun phrases touches upon aspects of vagueness in language which will be discussed in the present section. In the foreword of Channell’s work Vague Language, Sinclair comments on the status of vague language in linguistics (1994: xviii): “The art of being vague is a neglected concern for the linguist, and yet an important part of the armoury of every speaker and writer.” According to Sinclair, vagueness in language is an important and efficient device in communication that is often neglected in the relevant linguistic literature because of its rather negative image. Vagueness is commonly assumed to be “a defect to be avoided” (Jucker et al. 2003: 1738) and is thus seen as the undesirable counterpart of clear, precise and unambiguous language. Accordingly, a common view of vagueness in language is that ‘good’ language usage involves clarity and precision whereas ‘bad’ language usage involves a high degree of vague expressions. Channell comments on this fallacy and emphasises a very important aspect (1994: 3): “[V]agueness in language is neither all ‘bad’ nor all ‘good’. What matters is that vague language is used appropriately.” Vagueness constitutes an important aspect in language and can even be an intended means in communication. Sinclair notes that language in communication is in fact not always precise. Speakers of the same language share a common code and a “rough agreement about the meaning of the code” (Channell 1994: xviii). In reference to the ‘Cooperative Principle’ explained in the preceding Section 3.4.1, this agreement leaves room for vague expressions which, to a certain degree, the participants are able to decode. But an encoder may also use vague linguistic items (e.g. general noun phrases) to express a degree of vagueness which they know the decoder is no longer able to work out. Then, the encoder is deliberately vague and they follow communicative intentions that mislead the decoder. 66

3.4.2.1 Defining vagueness There is considerable variation in linguistic theories about how to define vagueness in language and about how vagueness is expressed in terms of linguistic devices. Vagueness must be marked off from ambiguity and polysemy. All three terms refer to different instances of plurality of meaning. Firstly, an expression is ambiguous if one single overt form entails two lexical units with two semantically unrelated senses8 (Cruse 1986: 51–52). Through the context9 only one of the two possible senses is selected (cf. Cruse 1986: 51). Consider the following example given by Cruse (1986: 51): (30) We finally reached the bank.

Here, the word bank entails two lexical units with two unrelated senses (bank1 = financial institute, bank2 = edge of a river) and by reference to the context, only one of the two possible senses is selected (cf. Cruse 1986: 51). Secondly, an expression is polysemous if one single overt form entails two or more lexical units with semantically related senses. The fact that the senses are semantically related can, for example, be due to metaphorical transfer of meaning (cf. Lipka 2002: 156). The following example is a simplified version of the representation of various semantically related senses of the polysemous term cup given by Lipka (2002: 94 [original emphasis]): (31) cup = ‘container usu. with handle’ ‘container with liquid’ (cup of tea) ‘ornamental vessel as prize’ (Davis Cup) ‘experience’ (her cup of sorrows)

Thirdly, vagueness10 is defined as a contrast to ‘well-defined’ (cf. Cruse 1986: 81). An expression is vague when it entails only one lexical unit with a single sense. This sense is very general and unspecific and only through reference to the

8

Cruse speaks of a lexical unit as the “union of a single sense with a lexical form” (1986: 80). 9 In his discussion about ambiguity and vagueness, Cruse leaves it open as to whether such expressions are solved by the linguistic co-text or the situational context. Unfortunately, Cruse does not give the linguistic co-text of examples (30) and (32), nor does he describe the situational context of these examples. For the discussion in the present section, I assume that it is an interaction of both the linguistic co-text and the situational context that solves ambiguous utterances. 10 Note that Cruse (1986: 51) uses the term generality to refer to vagueness.

67

context can it be modified and specified (cf. Cruse 1986: 51–52). Cruse gives the following example (1986: 51): (32) Sue is visiting her cousin.

Here, the term cousin carries a general sense which covers all the more specific possibilities, for example male/female. Only through the context of the utterance is the sense of cousin modified so that the decoder can identify the sex of the person (cf. 1986: 51). To sum up, the superordinate concept of plurality of meaning subsumes two similar yet different concepts: that of ambiguity and polysemy on the one hand and that of vagueness on the other hand. Building on that, a central position has to be granted to the notions of ‘co-text’ and ‘context’. In the case of ambiguity and polysemy, the co-text and context select one lexical unit from a group of distinct alternatives, while in the case of vagueness, the co-text and context rather modify the general sense of one lexical unit. The following figure illustrates this: Figure 3.4: The concepts of ‘ambiguity’, ‘polysemy’ and ‘vagueness’ Plurality of i meaning

CO-TEXT, C CONTEX XT

Vagueness (one lexical unit with one sense)

CO-TEXT, C CONTEX XT

Polysemy (two lexical units with ttwo related senses)

CO-TEXT, C CONTEX XT

Ambiguity (two lexical units with two unrelated senses)

Selection of one lexical unit

Selection of one lexical unit

Modification of single sense

For a pragmatic approach to general noun phrases, the focus lies on the concept of ‘vagueness’. Ullmann describes vague terms as ‘words with blurred edges’ (cf. 1962: 116). He accounts for vagueness in terms of the generic character of words which refer to classes of things instead of unique objects (cf. Ullmann 68

1962: 118). Corresponding to what Cruse states, Ullmann (cf. 1962: 124) emphasises that vague terms can only be specified through the co-text and the context of the communication situation. Similarly, Goss (cf. 1972: 287–288) defines ‘vague nouns’ as nouns that refer to a class with many members. In a dictionary, vague nouns would therefore have one definition followed by many examples. See the following entry of matter in the Longman Dictionary of Contemporary English online (cf. http://www. ldoceonline.com/dictionary/matter_1, last access 2nd January 2015): •  Subject/situation: a subject or situation that you have to think about or deal with •  Matter: a situation that you are in or have been describing •  Material: the material that everything in the universe is made of, including solids, liquids, and gases In contrast to vague nouns, ‘clear nouns’ refer to a class with only one member, for example apple or motorcycle (cf. 1972: 288). Van Rooij (cf. 2011: 123) defines vagueness in contrast to precision and also emphasises its contextual aspect: “An expression is vague, or has a tolerant meaning, if it is insensitive to small changes in the objects to which it can be meaningfully predicated.” A vague expression can thus be modified through the co-text and context so that its meaning shifts from use to use. According to van Rooij (cf. 2011: 124), members of almost any lexical category can be vague. Prototypical vague expressions include adjectives such as tall, fast and red, nouns such as heap, adverbials such as rather and probably, and quantifiers such as many and a lot. Van Rooij (2011: 124) even goes so far as to say that “[…] no linguistic expression whose meaning involves perception and categorization can be entirely free of vagueness”. In line with van Rooij’s understanding of vagueness, Channell (cf. 1994: 42 ff., 95 ff., 119 ff., 157 ff.) includes in her definition of vague language expressions such as approximators (about, around), approximating and vague quantifiers (loads of, some), tags (something like that, and whatnot), placeholder words (thingy, whatsisname) and vague nouns such as thing and stuff.

3.4.2.2 Types of vagueness It is commonly assumed in linguistics that vagueness constitutes a deficiency in language because it lacks precision and clarity. However, several linguists have criticised this view and have shown that vague language is not all “bad”. Channell says that the key issue is appropriateness (cf. 1994: 3). Jucker et al. (cf. Jucker et al. 2003: 1738) agree with this and explain that, for example, in a legal contract a high degree of precision is needed, while the same type of precision in a casual 69

chat between friends might be counterproductive or even off-putting. As a consequence, I will distinguish two types of vagueness: 1. ‘Conventional’ vagueness that occurs as a natural language phenomenon and is easily decodable. 2. ‘Rhetorical’ vagueness that is used intentionally in order to manipulate and mislead the decoder. Rhetorical vagueness is hard or even impossible to decode. The first type entails vagueness that occurs commonly in spoken as well as written language. According to what has been defined as vague language in Section 3.4.2.1, this category includes vague expressions such as approximators, placeholders and general noun phrases such as thing and stuff typically found in conversations where the speakers share the situation of utterance. Conventional vagueness is not completely unintentional, but, in contrast to rhetorical vagueness, this use of vagueness can be identified and decoded within the communication situation. Channell (1994: 159, 168 [original emphasis]) gives the following examples: (33) …they didn’t actually have a word p – er processing software thingy package with it. (34) E: I’m talking about you know sort of acceptable middle class language and (.) sort of working class language you know Bernstein and mm you know sort of elaborated code and things like that.

In example (33), the speaker uses the placeholder thingy for the proper name for the software they are talking about. According to the ‘Cooperative Principle’, this shows a violation of the maxim of quantity, as the encoder gives less information than is needed, they are not being specific. Decoders, assuming cooperative behaviour, then interpret such vague expressions as implicatures11 of the encoder which they use because they either do not know or have forgotten the actual name or the encoder does not want to refer explicitly to it by using the name because it is not relevant (cf. Chanell 1994: 33, 174–175, 181). Based on the encoder’s evaluation of the decoder’s cognitive abilities and the shared common ground, the encoder assumes that the decoder is able to discover and interpret 11 Grice uses the term implicature to describe the implying or suggesting of something that is not overtly expressed in an utterance (cf. Grice 1975: 43–44). Note that Grice (cf. 1975: 43–44, 49–50) does not expressly refer the term implicature to the use of vague expressions which, as illustrated in examples (33) and (34), are chosen by the encoder because they have difficulties finding the right words or because they do not know the actual name for something they want to say. For the purpose of the present study, however, I extend the term implicature for such uses.

70

the implications that they convey (cf. Jucker et al. 2003: 1742). In example (34), which is an extract from a discussion between student and tutor about a linguistic topic, the student uses the vague expression sort of elaborated code and things like that which shows that they are unsure what they are talking about. According to Channell (cf. 1994: 170), the use of such vague expressions enables the speaker to talk about subjects they are not very knowledgeable about. We have seen that ‘conventional’ vagueness is a form of vagueness that can be identified and solved by the decoder. In contrast to ‘conventional’ vagueness stands ‘rhetorical’ vagueness. It is characterised by an amount of vagueness that is no longer seen as appropriate in conversational contexts and therefore has a manipulating and misleading effect on the hearer. Referring to the data of the present study, vagueness can be expressed by two means: (i) either by using a very unspecific general noun phrase where a more specific term would be necessary, or (ii) by creating remote reference. The latter is the case when too much text lies in between a referring item and its reference item. The capacity of the decoder’s short-term memory is then deliberately being overwhelmed so that they can no longer establish reference between a general noun phrase and what it refers to in the text, or at least has problems doing so. For the use of a very unspecific general noun phrase were a more specific term would be needed, see the following example from the Conservatives manifesto: (35) Conservatives understand the inherent value of conserving things, and we know the importance of ensuring […] a good quality of life for future generations. [Conservative Manifesto, p. 95]12

This example, taken from the 2010 political manifesto of the British Conservative Party, demonstrates the use of the very unspecific general noun phrase things as a means of expressing vagueness in a way that aims at manipulating and persuading the reader. A political manifesto is a public declaration of a party’s political intents, aims and visions for the upcoming electoral period. With such a text, politicians have the opportunity to promote their ideas and convince the reader of the manifesto to vote for them. In the above example, the encoder uses rather strong and impassioned expressions such as ‘inherent value’ and ‘conserving things’. These expressions are supposed to convey a feeling of security: the decoder should get the impression that in times of a relentlessly changing 12 The general noun phrase things in example (35) is taken from the Conservative Manifesto. It is not coded as it is not taken into account in the quantitative and qualitative analysis because it is not clearly a concrete count noun (as defined by Halliday/Hasan 1976: 274). However, for the purpose of the present section, it serves as a good example.

71

globalised world, it is the Conservative Party that recollects the core values and conserves and protects things. When reading such profound words, it seems to fade into the background what things actually refers to. By using such a general noun phrase, the encoder reaches the decoder without committing themselves to any particular political actions and aims. Channell (cf. 1994: 178–179) states that an encoder deliberately uses vague expressions to withhold information. In that way, the encoder presents their data in a way which best supports their argument and thus helps to persuade their audience. In the above example, it would surely be beneficial for the future electorate to know what the general noun phrase things refers to. However, by using the unspecific general noun phrase, the politician protects themselves from any sort of binding commitment and, moreover, protects themselves from later being proven wrong. In that way, vague expressions help not to threaten face (cf. Channell 1994: 190), an aspect which seems to be of great importance in political contexts. For the creation of remote reference as a means to express rhetorical vagueness, see the following example: (36) Mrs Spelman: […] I might have made an error, Mr Speaker, in not responding myself to the question put by the hon. Member for Bristol East (Kerry Mc Carthy). I apologise, as I should have taken it – but, of course, I agree with everything that the Minister of State had to say. [D3-17-2]

Example (36) is from a parliamentary debate from the British House of Commons. Mrs Spelman, who is the Secretary of State for Environment, Food and Rural Affairs, uses the general noun phrase the question put by the hon. Member for Bristol East (Kerry Mc Carthy) to refer to a question concerning the Hunting Act 2004 and issues related to the protection of wildlife brought up by Mrs Mc Carthy. The question was asked 13 questions and 13 answers before the one in example (36) and therefore the reference is considered remote. The decoder can certainly not remember what the question exactly referred to, especially because Mrs Mc Carthy posed more than that one question. This might be intended by Mrs Spelman, who could have repeated the content of the question shortly. However, she wants to avoid bringing up the topic again and personally comment on the Hunting Act and related topics. In example (36) Mrs Spelman just says that she agrees with everything that her Minister of State responded and thus manages to get the topic off the agenda. The following example from the Labour manifesto illustrates an extreme case of remote reference because the decoder cannot identify a reference item altogether:

72

(37) In this election, the first of the post-crisis era, we stand as the people with the experience, values and ideas to help our country through the next phase of national renewal. [LM-18-1]

In example (37) the general noun phrase ideas to help our country through the next phase of national renewal does not refer to any other item in the co-text, which for reasons of space cannot be presented here. Example (37) is from the Labour manifesto, therefore, the general noun phrase is probably used to make general promises to the decoder. However, since the decoder cannot identify what the general noun phrase refers to, these promises are not being specified. Similar to example (35), the encoder uses rather strong and impassioned expressions which are supposed to persuade the decoder that the Labour Party is able to transform the country. When reading this, it seems to become irrelevant what Labour’s ideas involve in concrete terms. The above examples of ‘rhetorical’ vagueness have shown how encoders use general noun phrases as a strategy to express a degree of vagueness that manipulates the decoder. By using a very unspecific general noun phrase, the encoder withholds information and refrains from going into (unpleasant) detail. That way, the decoder is distracted and, in a way, blinded to specific (political) facts. As a second means of expressing vagueness, the encoder refers to facts within a discourse in a way that is not precise or definite enough for the decoder to identify. Therefore, reference relations become unclear and the encoder achieves that problems and facts may be weakened and fade away in the mind of the decoder. This is corroborated by Myers (cf. 1996: 3), who states that vagueness enables the terms and interests of one group to be translated into those of another group because “[…] [v]agueness allows mediation of apparently conflicting interests.” (Myers 1996: 3). Therefore, vagueness is to be seen as an important communicative strategy.

3.5 The cognitive approach to general noun phrases 3.5.1 Categorisation Categorisation is the mental process of classification (cf. Ungerer/Schmid 1996: 2). With its roots in ancient Greek philosophy, categorisation is a concept in psychology which deals with the segmentation of the environment or extra-linguistic world into classifications where unique particular objects or events can be treated equivalently as members of one class (cf. Rosch/Loyd 1978: 1). A classical view of categorisation since Plato and Aristotle has been that categories are discrete entities with clear-cut boundaries (cf. Rosch 2009: 41). According to the Aristologian 73

laws of logic, categories divide the universe into two sets of entities, members or non-members of those categories. There are no ambiguous cases (cf. Taylor 1995: 23). In his work Philosophische Untersuchungen, the philosopher Wittgenstein anticipated some of the inadequacies of the classical theory of categorisation. He coined the notion of ‘family resemblance’, which refers to a complicated network of overlapping similarities of category members (cf. 1953: 31e-32e). With these investigations, Wittgenstein challenged the idea of clear-cut category boundaries and laid the ground for further research in the field of categorisation. Important works in that area include the studies on the categorisation of colours conducted by Berlin and Kay (1969) and the subsequent studies on categorisation conducted by Rosch (1973, 1975 and 1978) and Rosch and Mervis (1975) and Rosch et al. (1976). Berlin and Kay (cf. 1969: 1) examined basic colour terms in 20 selected languages from a number of unrelated language families and found out that there exists a universal inventory of eleven basic perceptual colour categories in the 20 languages investigated (cf. Berlin/Kay 1969: 2). Berlin and Kay claim that the colour continuum is structured by a universal system of reference points for orientation, so-called ‘focal colours’, which the basic colour terms rely on (cf. Berlin/Kay 1969: 10). In the early 1970s, the psychologist Rosch set out to explore the psychological background of focal colours (cf. Ungerer/Schmid 1996: 6). In an experiment with the Dani, a non-westernised culture in Papua New Guinea, Rosch tested the processing of focal colours and found out that they appear to possess a perceptual cognitive salience, since the Dani learnt them more rapidly and more easily than other stimuli (cf. Rosch 1973: 341). Rosch suggested that domains of colour are structured into semantic categories which develop around ‘natural prototypes’ (foci of organisation for categories) (cf. Rosch 1973: 328). Having taken an interest in the internal structure of categories, Rosch tested the concept of ‘prototypicality’ and the categorisation of concrete objects in a series of experiments with American college students who were asked to judge the typicality of category members, for example members of the category ‘dog’ (cf. Rosch 1975: 198). In line with Wittgenstein, Rosch’s findings constitute a refutation of the classical approach to categorisation, as she showed that category membership is not, as was assumed by philosophers and linguists for a long time, a yes-or-no distinction but that categories have internal structure and are represented in terms of prototypes (best examples of the category) surrounded by other members of decreasing similarity to the prototype (cf. Rosch 1975: 225, 193).

74

After investigating the internal structure of categories and their prototypes, Rosch et al. (1976) also studied the external structure of categories, i.e. the relationship between categories. Categories are related hierarchically, which means that they occur at different levels of inclusiveness or levels of abstraction. If we consider, for example, the cognitive categories ‘creature’, ‘dog’ and ‘Scotch Terrier’, we can claim that these categories are related in the way that ‘creature’ is a superordinate category to ‘dog’, which in turn is a superordinate category to ‘Scotch Terrier’ (cf. Ungerer/Schmid 1996: 60). The principle underlying this hierarchical structure is the notion of class inclusion as defined by Lyons (1977: 156): To be distinguished from class-membership is class-inclusion. […] Inclusion is defined as follows: X ⊃ Y (X includes Y) and Y ⊂ X (Y is included in X; Y is a subclass of X) both mean that every member of Y is a member of X.

This means that a superordinate category includes all items of its subordinate categories, so that, referring to the above example, the superordinate category ‘creature’ includes all the members of the subordinate category ‘dog’, which in turn includes all the members of the subordinate category ‘Scotch Terrier’. Important insights concerning the levels of categorisation were gained from the works on folk taxonomies by Berlin et al., who studied the plant classification of the Tzeltal people, a Mayan-speaking community in Southern Mexico (cf. 1973: 219). The following table shortly summarises the results of the study by Berlin et al. The table is taken in a modified version from Ungerer/Schmid (1996: 64): Table 3.2: Tzeltal plant classification according to Berlin et al. (1973) Scientific class regnum

Level 1. unique beginner

No. of category 1

Example plant

class

2. life form

4

tree

genus

3. generic

471

pine

species

4. species

273

red pine

varietas

5. varietal

8

One major outcome of the study by Berlin et al. (cf. 1973: 240) is the primacy of the generic level, which can be supported by the high number of categories (471) on that level (see Table 3.2). This prominence of the generic level was further investigated by Rosch et al. (1976), who refer to it as the basic level. Based on attribute-listings of American college students, Rosch et al. showed that basic categories are not only most numerous, but also perceptually salient in as much 75

as they are the most inclusive categories whose members have significant numbers of attributes in common, have similar shapes and are at the same time the most differentiated from another (cf. Rosch et al. 1976: 382). This becomes clear when we again refer to the above mentioned example of the categories ‘creature’, ‘dog’ and ‘Scotch Terrier’. The subordinate category ‘Scotch Terrier’ is defined by category attributes which are so closely related to the attributes of neighbouring categories such as, e.g. ‘German shepherd’ and ‘Collie’ that differences between those specimens are indeed very small (cf. attributes such as “has four legs and a tail and barks”). The superordinate category ‘creature’, however, embraces such a disparate variety of members (imagine the difference between elephants and mice) that there are hardly any similarities. Therefore, the category ‘dog’ is considered as a basic level category. Against this background, basic level categories achieve an ideal balance between internal similarities and external distinctiveness (cf. Ungerer/Schmid 1996: 67).

3.5.2 Context-dependence of categories Prototypes have been defined by Rosch as central tendencies or “typical members” of cognitive categories (cf. Rosch 1973: 328–329). This view, however, disregards one major aspect, namely that of context. Thus, categories, and therefore also their internal and external structure, may change when they occur in a particular context. As has been explained in Section 2.1.1, the notion of ‘context’ refers to the general, non-linguistic environment of any language activity including the sociocultural background (cf. Sinclair 1991: 171). Halliday/Hasan (cf. 1989: 6–7) distinguish between the ‘context of situation’, meaning the situational environment of an utterance, and the ‘context of culture’, meaning the cultural history of the participants, and state that both are necessary for a proper understanding of an utterance.13 Thus, concerning cognitive categories, it is our social and cultural knowledge that heavily influences the categorisation of our experiences of the world (cf. Ungerer/Schmid 1996: 43). This knowledge base, called ‘cognitive model’, consists of all kinds of experienced and stored contexts that we can use for association and comparison (cf. Ungerer/Schmid 1996: 47). Cognitive models are not universal, but always depend on the culture in which a person grows up and lives, since the culture provides the background for all the situations we experience. Cognitive models thus ultimately depend on cultural 13 Halliday/Hasan (cf. 1989: 6–7) state that these terms have originally been coined by Malinowski (1923).

76

models (cf. Ungerer/Schmid 1996: 50). Consider the following example given by Ungerer and Schmid (1996: 43): (38) Right from the start of the race the dogs began chasing the rabbit.

Without reference to any co-text and context, we would rely on our cognitive category ‘dog’ and would probably think of an Alsatian as a prototypical dog and a Retriever or a Poodle as being just slightly less prototypical. Considering our social and cultural context in example (38), however, we are looking for a dog that is typically associated with dog racing. Our prototypical dog (Alsatian) would no longer fit here; instead, we would probably think of a dog like a Greyhound. Concerning the other members of the category ‘dog’, it would be quite odd in this context to think of a Poodle as a typical member of the category. We would now rather think of a Poodle as a peripheral member of the category ‘dog’. This example demonstrates that the context not only determines the choice of the category prototype, but that it also leads to an adjustment of the position of other category members (cf. Ungerer/Schmid 1996: 44). Ungerer and Schmid state that cultural models are important for the choice of the basic level perspective (cf. Ungerer/Schmid 1996: 70). Schmid (1993: 74) states: Während das Prinzip der Basisebene als universal angesehen werden muß, variiert die Höhe der Basisebene von Kultur zu Kultur, von Sprecher zu Sprecher je nach Interesse, Erfahrung und Auseinandersetzung mit dem jeweiligen Bereich.14

According to Schmid (cf. 1993: 74), the concept of the ‘basic level’ is universal, however, the position of the basic level category is context-dependent.

3.5.3 Categorisation and general noun phrases 3.5.3.1 Introduction The preceding sections have pointed out that due to the general meaning and the superordinate nature of general noun phrases,15 these items reflect broad 14 While the principle of the basic level must be considered universal, the height of the basic level varies considerably from culture to culture, from speaker to speaker depending on interest, experience and confrontation with the respective area. (Translation provided by V.B.) 15 It has been critically remarked in Section 2.3 of the present study that Halliday/Hasan (cf. 1976: 275) tend to simplify the status of general noun phrases when describing them as superordinate members of lexical sets. Moving away from perfectly constructed textbook examples, the present study has shown that the co-text and context

77

cognitive categories. As Rauh (cf. 2002: 260) explains, members of a category share essential properties or attributes. The general noun phrase people, for example, encapsulates different meanings; in cognitive terms, the general noun phrase people constitutes a very broad category with many category members. Following Wittgenstein, Anderson et al. (cf. 1976: 667) speak of a “family of potential meanings” and state that the set of criterial properties for a category reflected by a general noun phrase shifts from use to use depending on the co-text and context, so that a property which is distinguishing in one case may be unimportant or even absent in another. The different meanings of people can only be specified or singled out through reference to its co-text and context in each single utterance. In particular, only if we take into account modification and references within the text and outside the text can we identify what people means in a specific utterance, or in cognitive terms, which member of the category people is selected as prototypical. When the general noun phrase head people is specified through modification and reference, we can say that the members of this category share less common attributes than the members of the category ‘people’ reflected by a completely unspecific general noun phrase. This aspect will be explained in more detail in Sections 3.5.3.2 and 3.5.3.3. Resulting from this process, the present section suggests to describe general noun phrases as semantically empty containers that are filled with content to different extents through references to the co-text and context. Rauh (cf. 2002: 260–261) explains that the use of linguistic items (here general noun phrases) in actual discourse means that we have to take into account their co- and context and that only then we can determine the degree of extension of a category. In order to explain what Rauh means with ‘extension of category’, it is necessary to consider category-internal and category–external structures. In Rosch’s terminology (cf. 1978: 30), the internal structure of categories constitutes the horizontal dimension of categorisation and refers to the segmentation of categories and category members at one and the same level of inclusion, or in other words, at the same level of abstraction. The horizontal dimension is concerned with the distribution of category members and prototypes within one category and refers to the axis where, for example, different kinds of dogs such as Pekinese, Alsatian and Poodle are described. The following figure illustrates the horizontal distribution of category members:

of general noun phrases determine whether they have a general or specific meaning. Thus, speaking of the superordinate nature of general noun phrases here only considers de-contextualised general noun phrases.

78

Figure 3.5: Horizontal dimension of categorisation illustrated by the example ‘dog’ peripheral member of the category

core or prototypical member of the category

peripheral member of the category

Pekinese

Alsatian

Greyhound

The external structure of categories refers to the hierarchical distribution of categories and category members on a vertical axis. The vertical dimension of categorisation refers to the principle of class inclusion as described by Lyons (cf. 1977: 156) and explains how categories such as ‘creature’, ‘dog’ and ‘racing dog’ are distinguished in terms of super- or subordination. Thus, the vertical dimension refers to levels of abstraction and describes relations of hyponymy.16 The following figure illustrates this: Figure 3.6: Vertical dimension of categorisation illustrated by the example ‘dog’

abstract category

creature

basic level category

dog

specific category

racing dog

A cognitive approach to general noun phrases considers internal as well as external structure aspects in order to show whether a general noun phrase has a general or specific meaning in an utterance. Schmid emphasises that both dimensions regularly interact in the analysis of language (1993: 44):

16 The vertical dimension of category structure is based on findings of Berlin et al. (1973). See Section 3.5.1 for a discussion of folk taxonomies as developed by Berlin et al.

79

Eine allgemeine, theoretisch klar fundierte Trennung zwischen intrakategorialen und interkategorialen [horizontalen und vertikalen] Beziehungen ist nicht sinnvoll, da beide Aspekte notwendigerweise bedeutsam werden, sobald zwei verschiedene Ebenen taxonomischer Tiefe eingeführt werden.17

A cognitive approach to general noun phrases always includes both, considerations of prototypes (horizontal dimension) and considerations of abstraction and specification (vertical dimension). See the following figure for a demonstration of the processes involved in a cognitive analysis of a general noun phrase in consideration of its co-text and context: Figure 3.7: Extension of a general noun phrase category dependent on co-text and context co-text and context of a general noun phrase

selection of level of categorisation

selection of prototypical meaning

periphery

core periphery prototypical member

When analysing a general noun phrase, the consideration of its co-text and context is the starting point. Co-text and context add meaning to the general noun phrase and it is only then that it can be determined whether a general noun phrase is in fact general in meaning or whether it is specific. On the horizontal axis, this means that co-text and context determine which meaning of the general noun phrase is selected as prototypical in a given utterance. On the vertical axis, co-text and context determine the degree of specification of the selected general noun phrase meaning thus determine at which level of abstraction the general noun phrase category is positioned.

17 A general, theoretically well-founded division of intra-categorial and inter-categorial [horizontal and vertical] relations is not reasonable because both aspects necessarily become relevant as soon as two different levels of taxonomic depth are introduced. (Translation provided by V.B.)

80

3.5.3.2 General noun phrases as “empty containers” As explained before, we can speak of an “empty container” when a general noun phrase conveys only very general meaning due to a lack of specification through the co-text and context. For an illustration, see the following example: (39) The question at this election is whether people think the choice we made was the right one and whether we use the power of government to help sustain recovery, or allow it to be snuffed out. [LM-1-7]

Here, the general noun phrase head people is neither pre- nor postmodified and there is no reference to any other item in the text. Thus, no specific meaning is attached to the broad category ‘people’. The only information the decoder gets for the interpretation of the general noun phrase people is through reference to the context of the utterance: since the example is taken from the manifesto of the British Labour Party, they know that the utterance is embedded in the context of political elections in the UK. Accordingly, people probably refers to British electorate. The category ‘British electorate’ constitutes a very broad category since its members can be any male or female who is at least 18 years old on polling day, who is a national of the United Kingdom, the Republic of Ireland, a Commonwealth country or a European Union member state and who has their permanent residence in the United Kingdom. Our cultural knowledge tells us that. Strictly speaking, people could also be any person who thought the choice the politician speaking in example (39) made was right, so basically any person in the world who is interested in British politics or has at least heard of the Labour Party. Given this interpretation, the category ‘people’ is very broad. Whatever interpretation we support, there is only very little limitation as to the expansion of the category reflected by the general noun phrase people in example (39). Thus, in terms of cognitive structuring, it is quite problematic, if not impossible, to name a prototypical member of such a broad category. The category ‘people’ moves at a rather superordinate and abstract level. See the following figure to demonstrate this:

81

Figure 3.8: Extension of the general noun phrase category ‘people’ dependent on co-text and context of example (39) co-text and context of the general noun phrase people l

superordinate level of categorisation

selection of prototypical meaning of this general noun phrase h

periphery new bborn baby

core anyone interested in Labour politics

periphery US citizen

subordinate level of categorisation

Since the co-text and context of the general noun phrase do not specify its meaning, it is hardly possible to single out a specific prototype of the category ‘people’. This leads to an arrangement of the category at a superordinate level of abstraction.

3.5.3.3 General noun phrases as “full containers” When a general noun phrase is specified through its co-text and context, for example through modification and endophoric reference, it is supplied with information. In cognitive terms, we can speak of a container that is filled with semantic content. See the following example for a demonstration. Note that only the first occurrence of the general noun phrase head people is of primary interest: (40) We, Labour, are the people to carry out this next stage of national renewal of our values and our understanding of the role of government: to stand by ordinary people so they can change their lives for the better. [LM-1-1]

The first instance of the general noun phrase head people is part of a complex general noun phrase, the meaning of which is further specified through reference. The general noun phrase head people is postmodified by the infinitive clause to carry out this next stage of national renewal […]. The postmodification already gives the information that the general noun phrase refers to politicians because usually only politicians find themselves in a position in which they are able to carry out political changes that affect the country, such as the above quoted “national renewal”. Our world knowledge tells us that. The general noun phrase is 82

further specified through anaphoric reference to “We, Labour, […].”,18 so that the decoder knows that of all possible people only the members of the Labour Party are concerned. In terms of category structure, we can say that the co-text and context selects Labour politician as the prototypical member of the category ‘people’. The category ‘people’, reflected by the general noun phrase the people to carry out this next stage of national renewal of our values and our understanding of the role of government: to stand by ordinary people so they can change their lives for the better is very specific and can therefore be positioned at a subordinate level of abstraction. See the following figure for an illustration: Figure 3.9: Extension of the general noun phrase category ‘people’ dependent on co-text and context of example (40)

co-text and context of the general noun phrase the p people p p to carryy out this next stage of national renewal of our values and our understanding of the role of government: to stand by ordinary people so they can change their lives for the better

selection of prototypical meaning of this general noun phrase h

superordinate level of categorisation g

subordinate level of g categorisation

periphery new born baby

core periphery Labour US politician citizen

Figure 3.9 illustrates the cognitive steps involved in the interpretation of the complex general noun phrase the people to carry out this next stage of national renewal of our values and our understanding of the role of government: to stand by 18 The general noun phrase is the subject complement of the clause and anaphorically refers to its subject “We, Labour, […]”. Note that this type and other types of reference relevant for the analysis in the present study will be further discussed in Section 6.2.3.

83

ordinary people so they can change their lives for the better from example (40). As described in this example, the co-text specifies the general noun phrase so that Labour politician can be identified as the prototypical meaning in the utterance. The general noun phrase is positioned at a subordinate level of abstraction.

3.6 Summary: A workable definition of general noun phrases The preceding sections have shown that a description of general noun phrases within the textlingustic framework of cohesion by Halliday/Hasan (1976) is by no means sufficient for a comprehensive description of these items because it fails to consider important micro- and macrolinguistic aspects. The preceding sections have therefore taken into account different linguistic disciplines for the description of general noun phrases. Starting from a grammatical point of view, general noun phrase heads were defined as common nouns. Ten of the general noun phrase heads investigated in the present study were defined as concrete count nouns, six were defined as abstract count nouns, one was defined as concrete noun-count noun and one was defined as abstract non-count noun (see Table 3.1 in Section 3.2.3). In a classification of nouns along a scale from specific to general, the general noun phrase heads investigated in the present study were positioned at a point where the classes of concrete and abstract count nouns, and concrete and abstract noncount nouns overlap. This point is located towards the general end of the scale (see Figure 3.3 in Section 3.2.3). It has been emphasised that this arrangement only refers to de-contextualised general noun phrase heads which are taken out of their co-text and context. With the text-linguistic framework of de Beaugrande and Dressler (1981), general noun phrases were described in terms of a user-oriented approach to texts and communication. Aspects such as intentions of the encoder as well as efficiency and effectiveness of text-production and communication were considered. It was shown that general noun phrases can be a helpful tool from the point of view of the encoder and their intentions concerning the distribution of knowledge and the attainment of a certain aim. The informational load of an utterance can be strongly influenced by the use of general noun phrases: either general noun phrases are very unspecific or they encapsulate a large amount of information. This influences the degree of effectiveness and efficiency of the text. In a balanced relation, the encoder is able to gain maximum reward for minimum effort. In an unbalanced relation, however, there might be a ‘trade-off ’ between propositional density and clarity of thought (cf. de Beaugrande/Dressler 1981: 64, 170).

84

In line with the text-linguistic approach to general noun phrases was a pragmatic approach which has shown that general noun phrases have a great influence on the communicational behaviour which can be explained with the ‘Cooperative Principle’ as formulated by Grice (1975). General noun phrases can influence the communicational success in that they can be used deliberately to express vagueness. Vagueness as a linguistic phenomenon has thus been looked at and it has been stated that vague nouns refer not to single entities but rather to a class with many members (cf. Goss 1972: 287–288). Thus, general noun phrases that are used vaguely need conceptual specification through reference to their co-text and context. Schmid (1993: 183) states: Für die semantische Interpretation eines solchen “general nouns” […] hat seine Unbestimmtheit die Folge, daß Ko- und Kontext zur Fixierung der aktuellen Bedeutung einen immensen Beitrag leisten.19

This aspect has been further discussed within a cognitive approach to general noun phrases. Ivanič (cf. 1991: 103) states that general noun phrase heads are semantically rather undefined and that their meaning is not self-contained. From a cognitive point of view, general noun phrases can be viewed like empty containers that contain the least amount of substantial meaning, for example the general noun phrase thing. Only through reference to the co-text and context is a specific, context-dependent meaning attached to the general noun phrase head (cf. Ivanič 1991: 95). Then the general noun phrase can be viewed as a full container. It has been shown that general noun phrases represent categories which move at different levels of abstraction depending on the degree of specification of the general noun phrase. Without modification and without reference to the cotext and the context, a general noun phrase represents a superordinate category with members that share an indefinite number of semantic features or attributes. By contrast, a general noun phrase which is modified and refers to the co-text and context moves at a less abstract level of categorisation. This can be the generic or basic level or even a more subordinate level. It has been emphasised that the level of abstraction of a general noun phrase category and also the selection of a prototypical meaning of that general noun phrase category varies depending on the co- and context of the general noun phrase.

19 For the semantic interpretation of such a “general noun” […] the result of its indefiniteness is that co- and context make a great contribution to the fixation of the actual meaning. (Translation provided by V.B.)

85

4. Framework for the analysis of general noun phrases Abstract: Chapter 4 presents the corpus-analytical framework, which takes into account structural and semantic features of general noun phrases. The major tool is a matrix which cross-tabulates the parameters +/-modification and +/-linkage. The result is a classification of general noun phrases into four categories ranging from general to specific.

4.1 Combination of structural and semantic parameters of general noun phrases The present chapter introduces the framework for the quantitative and qualitative analysis of general noun phrases in the present study.1 This framework is based on the insights that have been gained by looking at general noun phrases from different linguistic perspectives, such as the grammatical, the text-linguistic, the pragmatic and the cognitive perspective. As an outcome of these approaches, a major role must be granted to the notions of ‘co-text’ and ‘context’. It has been shown that the degree of specification of general noun phrases can only be determined through the consideration of these notions: while de-contextualised general noun phrase heads – general noun phrase heads which are taken out of their co-text and context – have very little intrinsic meaning, contextualised general noun phrases heads – general noun phrase heads which are regarded within their co-text and context – can carry a great load of semantic information. This depends on whether they are modified and whether they are linked within the co-text and context, thus, whether for their interpretation they make reference to another item in the text or to the situation of the utterance. For the functional analysis of general noun phrases, the notions of ‘co-text’ and ‘context’ must be operationalised. As explained in Section 2.1.1, the cotext of a general noun phrase head refers to its linguistic surrounding. It reveals whether a general noun phrase head is pre- or postmodified and whether the general noun phrase refers to another item in the text. The context reveals

1

The qualitative analysis of general noun phrases determines the degree of specification of general noun phrases and discusses their functions across the different genres and media in the underlying corpus. Beforehand, the quantitative analysis of general noun phrases determines the frequencies and distribution of general noun phrases in the same corpus. For the methods of both analyses, see Chapter 6.

87

whether a general noun phrase refers to situational aspects outside the text for its interpretation. For the analysis of a general noun phrase, therefore, we rely on structural parameters and semantic parameters which determine the specification of general noun phrases. The structural parameters are + modification and - modification and determine whether the general noun phrase is simple or complex, thus whether the general noun phrase head is modified or not. The semantic parameters are + linkage and - linkage and determine whether the general noun phrase establishes semantic links within its co-text and to its context. For the analysis of general noun phrases it is assumed that only the combination of both parameters, modification and linkage, fully determines the degree of specification of general noun phrases. Therefore, the major tool for the analysis of general noun phrases in the present study is a matrix which combines the presence and absence of the two parameters modification and linkage. This functional matrix will be introduced in detail in Section 4.2.

4.1.1 Structural parameters of general noun phrases: +/- Modification As has been explained in Sections 3.2.1 and 3.2.2, general noun phrase heads can be part of a simple or complex noun phrase. As part of a complex noun phrase, general noun phrase heads can take modifiers of different lengths, from words, to phrases, to clauses. These modifying elements supply the general noun phrase head with more information and enable us to build up a more detailed and precise description of it (cf. Quirk et al. 1985: 65). While a non-modified general noun phrase head is very general in meaning, a pre- or postmodified general noun phrase head or a pre- and postmodified general noun phrase head is enriched with meaning and is thus more specific.

4.1.1.1 Non-modified general noun phrase heads: - Modification As has been explained in Section 3.2.1, we must distinguish simple general noun phrases from complex general noun phrases. The general noun phrase is simple when the general noun phrase head occurs alone or when it is preceded by determiners (cf. Biber et al. 1999: 230). Simple general noun phrases are semantically not specified and therefore have a very general meaning. See the following example: (1) Yvette Cooper: The right hon. Lady said last week […]. Safeguarding children is too important to have such loopholes. I urge her to listen to experts and think again. [D5-5-13]

88

In example (1), the general noun phrase head children is neither premodified nor postmodified. Therefore, children constitutes a simple general noun phrase. Moreover, the general noun phrase head is not preceded by any determiners, which means that the general noun phrase consists only of the head. Since the general noun phrase head children is not modified, it is semantically very unspecific and refers to a very broad category. It must therefore be assumed that in example (1), children refers to all children in general, regardless of parameters such as sex, age, and origin. See another example which shows the use of a simple general noun phrase: (2) The periods of four years retained in respect of both building operations and change of use to use as a dwelling house clearly reflect the legislator’s view that this would give adequate opportunity for enforcement steps, after the expiry of which the infringer would be entitled to repose and to arrange his affairs on the basis of the status quo. [J7-13-1]

In example (2), the general noun phrase his affairs is considered simple, because the general noun phrase head affairs is not modified. It is only preceded by the possessive determiner his. Semantically, the general noun phrase is very unspecific and it is not clear to the decoder what exactly is meant by his affairs. Note that through reference to the co-text and context of the utterance, in particular through association with utterances like ‘change of use to use of a dwelling house’ and ‘to repose and to arrange his affairs on the basis of the status quo’, the decoder can assume that his affairs refers to things that have been changed at or in the house. This, however, includes semantic parameters of linkage which will be discussed in Section 4.1.2.

4.1.1.2 Modified general noun phrase heads: + Modification As has been explained in Section 3.2.2, modification of general noun phrase heads comprises all the items placed before the head other than determiners and all the items placed after the head. It has been explained that modifiers give more information and specification to the head. They denote or rather specify properties of noun phrase heads in that they give additional information which serves to enrich the semantic content of the head (cf. Quirk et al. 1985: 65, 1239). The following examples demonstrate how pre- and postmodification and a combination of both specify a general noun phrase head: (3) We will also reform Access to Work, so disabled people can apply for jobs with funding already in place for equipment and adaptation that they need. [LDM-1-29]

89

In example (3), the general noun phrase head people is premodified by the past participle disabled. The premodifier specifies the general noun phrase head so that the decoder understands that disabled people refers to a certain group of people with certain characteristics. (4) And we will give the people who work in our public services much greater responsibility. [CM-1-13]

In example (4), the general noun phrase head people is postmodified by the relative clause who work in our public services. The relative clause specifies the general noun phrase head and supplies it with essential information. It has been explained in Section 3.2.2 that postmodifiers are generally more explicit than premodifiers and that relative clauses in particular are most explicit (cf. Quirk et al. 1985: 1243, 1321; Jackson 1980: 69). In line with this, we can see that the postmodifying relative clause in example (4) gives more semantic specification to the general noun phrase head than the premodifying past participle in example (3). While in example (3) we only know that disabled people refers to a relatively broad sub-category of people, in example (4) we know that the people who work in our public services refers to a rather specific category of British people working in specific positions. (5) 200,000 jobs through the Future Jobs Fund, with a job or training place for young people who are out of work for six months, […]. [LM-1-144]

In example (5), the general noun phrase head people is premodified by the adjective young and postmodified by the relative clause who are out of work for six month. Similar to example (4), the general noun phrase head in example (5) is postmodified by the most explicit type of postmodification. The relative clause supplies the head noun people with specific information concerning their momentary work condition. The premodifying adjective supplies the head noun people with additional information concerning the parameter age so that the whole general noun phrase young people who are out of work for six months refers to a rather specific category of people. Examples (3)-(5) illustrate an increase in the degree of specification of the general noun phrases in italic. In cognitive terms this means that the category represented by the general noun phrase disabled people is rather broad and has many possible members. It moves at a rather abstract level. The category represented by the general noun phrase the people who work in our public services is more specific and therefore has less possible members. Accordingly, it moves at a less abstract level. The category represented by the general noun phrase young people who are out of work for six months is the most specific category of the three 90

and therefore has the least possible members. It moves at the least abstract level of the three.

4.1.2 Semantic parameters of general noun phrases: +/- Linkage Besides structural parameters of modification, the specification of general noun phrases also depends on semantic parameters such as endophoric reference and exophoric reference to the context of the situation. Section 2.1.2 discussed the concept of endophoric reference as formulated by Halliday/Hasan (1976). Accordingly, endophoric reference has been described as a process that links two or more items in a text so that instead of being interpreted semantically in its own right, a linguistic item makes reference to other linguistic elements in the text for its interpretation (cf. Halliday/Hasan 1976: 31). For the full interpretation of a linguistic item, for example a general noun phrase, the decoder also makes reference to the extra-linguistic context of the utterance. Esser (cf. 2009: 137) explains that in order to interpret a linguistic item (here a general noun phrase) a text decoder then draws on previous knowledge (textual or of the world) for a proper understanding of the utterance. As has been explained in Section 4.1, the parameters modification and linkage must be operationalised for the analysis of general noun phrases. Concerning the parameter modification, this can be done without any problems as it is objectively identifiable whether a general noun phrase head is modified or not. Concerning the parameter linkage, this is more problematic. It can be determined quite clearly whether a general noun phrase is linked within its co-text in terms of endophoric reference or not because endophoric reference is detected on the text surface. However, it cannot be clearly determined whether a general noun phrase is linked to the textual or situational context of an utterance. The decoder surely considers factors such as their textual knowledge, their world knowledge and the situation of utterance for their interpretation of a general noun phrase. However, it is not objectively measurable whether this is the case or to what extent this is the case because the text-external factors that are included in the interpretation of a general noun phrase vary from decoder to decoder and are thus not objectively measurable. Therefore, the parameter linkage in the functional matrix for the analysis of general noun phrases only includes the presence and absence of endophoric reference. It should be born in mind, nonetheless, that reference to the decoder’s textual knowledge, their world knowledge and to the situational context of an utterance always plays an important role in the interpretation of a general noun phrase.

91

4.1.2.1 Non-linked general noun phrases: - Linkage When speaking of non-linked general noun phrases we speak of general noun phrases that are semantically not linked through endophoric reference within their co-text. These general noun phrases are not specified and, depending on whether they have already been supplied with information through modification, are very or at least rather unspecific. Consider the following example taken from the Conservative manifesto: (6) We have the energy, the ideas and the ambition to get Britain back on track. And that includes everyone in Britain, wherever they live and whatever their circumstances. […] We can bring about the change Britain needs. [CM-18-1]

In example (6), the general noun phrase the ideas to get Britain back on track is not specified through endophoric reference. Neither the preceding nor the following co-text specifies what exactly the Conservative’s ideas are. Note that for reasons of space the above example cannot provide the wider co-text of the utterance. The postmodifying infinitive clause to get Britain back on track supplies the general noun phrase head with additional information which, however, does not specify what the ideas exactly are. A party manifesto gives politicians the opportunity to introduce to their electorate their political concerns and aims. In such a context it would be of particular importance to address and name the party’s ideas clearly, but this is not the case in the above example of the Conservative manifesto. Instead, the general noun phrase the ideas to get Britain back on track does by no means refer to any other item or passage of text in the preceding or following text which helps to identify the Conservative’s ideas. See another example taken from a House of Common’s debate: (7) The Secretary of State for Communities and Local Government (Mr Eric Pickles): The government have published proposals to allow local retention of business rates and are seeking views by 24 October. The plans give councils a strong financial inventive to drive economic growth, as well as providing protections for places in need of additional support. [D8-16-1]

The general noun phrase places in need of additional support is not linked within its co-text through anaphoric or cataphoric reference. Note that due to reasons of space, the wider co-text of the utterance cannot be presented here. An indication for the fact that the general noun phrase is not linked through endophoric reference is the missing definite article, which according to Halliday/Hasan (cf. 1976: 275) is an indicator for endophoric reference. The general noun phrase head places is postmodified by the prepositional phrase in need of additional support which specifies that the general noun phrase refers to a sub-category of the 92

category places. However, there is no further specification through reference on the text surface. Only through textual hints can the decoder identify that places in need of additional support probably refers to businesses. However, this is not a criterion for the analysis of general noun phrases in the present study which is why the general noun phrase in example (7) is considered not linked.

4.1.2.2 Linked general noun phrases: + Linkage When a general noun phrase is semantically linked within its co-text, it anaphorically or cataphorically refers to another item or a passage of text. This way, the decoder identifies the information they need for their interpretation of the general noun phrase. See the following example taken from the manifesto of the Liberal Democrats: (8) Only Liberal Democrats have the big ideas for fundamental, structural changes in the way our country works to make it fair. Only Liberal Democrats will shake up the tax system to put 700 back in the pockets of tens of millions of low and middle-income families, […]. Only Liberal Democrats will break up the banks and start Britain building things again, creating a sustainable economy that no longer threatens our planet’s future. Only Liberal Democrats will invest in our schools to give every child, no matter their background, a fair start in life. And only Liberal Democrats will sort out our rotten political system once and for all. [LDM-18-1]

In the above example, the general noun phrase head ideas is premodified by the adjective big and postmodified by the prepositional phrase for fundamental, structural changes in the way our country works and the infinitive clause to make it fair. Through multiple modification, the general noun phrase head is already supplied with specific information. Moreover, the general noun phrase is specified through cataphoric reference to the underlined passage of text which outlines the Liberal Democrat’s ideas. The general noun phrase the big ideas for fundamental, structural changes in the way our country works to make it fair refers to improvements in the tax system, the economy and education. The above example illustrates that general noun phrases are in fact not necessarily general in meaning but can be quite specific. See another example for such a specific general noun phrase: (9) Robert Flello (Stoke-on-Trent South) (Lab.): […] What steps are the Ministry of Defence and the Secretary of State taking to ensure that when former British military personnel lay their lives on the line, like their currently serving colleagues, the terms, conditions and welfare of those very brave men and women are looked into and they are looked after and taken care of? [D2-3-1], [D2-4-6]

93

In example (9), the general noun phrase heads men and women are premodified by the adjective phrase very brave. Both general noun phrases are specified through anaphoric reference to the noun phrases former British military personnel and their currently serving colleagues. The demonstrative pronoun those which is part of the general noun phrases clearly points to the reference items, the two preceding noun phrases which are underlined. Through the anaphoric reference, the general noun phrases are specified and it becomes clear who those very brave men and women are. For a demonstration of another specific general noun phrase, see the following example taken from a Supreme Court judgment: (10) One other matter should be mentioned at this stage. […], the Secretary of State in December 2010 published the Localism Bill which, if enacted, will by section 104 amend the 1990 Act by inserting three new subsections […] expressly to deal with issues of concealment. […], I would observe only, first, that their proposed inclusion in the legislation surely indicates that […]; secondly that, pending the proposed statutory amendments, only truly egregious cases such as this very one […] should be regarded as subject to the Connor principle. [J7-14-10]

In example (10), the general noun phrase one other matter cataphorically refers to the following passage of text that is underlined. This passage of text elaborates what that one other matter is and gives detailed information which specifies the general noun phrase. Ivanič (cf. 1991: 104) states that general noun phrases can encapsulate meaning which is exposed in the following sentences. In that way, general nouns can function as signals for the organisation of the discourse. This is the case in the above example (10). The general noun phrase one other matter announces what follows and serves as an umbrella term for the explanations marked by underlining.

4.2 Functional matrix for the analysis of general noun phrases As has been explained at the beginning of Section 4.1, the degree of specification of general noun phrases can only be determined when both parameters, modification and linkage, are taken into account. Therefore, the basic tool for the analysis of general noun phrases is a functional matrix which combines the two parameters modification and linkage, or rather the absence and presence of these two parameters. See Figure 4.1, which presents the functional matrix for the analysis of general noun phrases in the present study:

94

Figure 4.1: The functional matrix for the analysis of general noun phrases

+ linkage

specification

– modification

– liinkage

+ modification

generalisation

With the functional matrix, the general noun phrases that are investigated in the present study (cf. Halliday/Hasan 1976: 274) will be classified based on the four parameters + modification, - modification, + linkage, and - linkage. As presented in Sections 4.1.1 and 4.1.2, these parameters determine whether a general noun phrase is used in its very general sense or in a more specific sense. The combination of these four parameters result in four categories of specification: The first category, which will be called the ‘most specific’ category, is characterised by the parameters + modification and + linkage. The second category, which will be called the ‘rather specific’ category, is characterised by the parameters - modification and + linkage. The third category, which will be called the ‘less specific’ category, is characterised by the parameters + modification and - linkage, and the fourth category, which will be called the ‘least specific’ category, is characterised by the parameters - modification and - linkage. Figure 4.1 shows that as we move from top left to bottom right we speak of generalisation and as we move from bottom right to top left we speak of specification. In other words, when we speak of generalisation, we move from the ‘most specific’ category of general noun phrases to the ‘least specific’ category of general noun phrases and vice versa when we speak of specification. See the following figure for an illustration:

95

Figure 4.2: Generalisation and specification illustrated with the functional matrix for the analysis of general noun phrases

‘less specific’

– linkagee

+ modification

‘rather specific’

‘most specific’

– modification

+ modification

‘least specific’

‘less specific’

+ linkage

specification

‘rather specific’

– modification

– linkagee

‘most specific’

+ linkage

generalisation

‘least specific’

As we can see in Figure 4.2, it is assumed that the parameter linkage has more influence on the degree of specification of a general noun phrase than the parameter modification. This explains why a general noun phrase which is not modified but linked within its co-text is considered ‘rather specific’ while a general noun phrase which is modified but not linked within its co-text is considered ‘less specific’. See the following examples to demonstrate this: (11) Karen Bradley (Staffordshire Moorlands) (Con): Is the Prime Minister aware of the case of my constituent Mr Edmond Arapi, who is facing extradition to Italy, having been tried in his absence? Will my right hon. Friend look into the matter urgently and accelerate the review of extradition cases […]? [D1-14-4] (12) Dr Fox: We have repeatedly made it clear that we believe that having an independent nuclear deterrent is a vital part of the United Kingdom’s sovereign capability, […]. Where we can co-operate on technical matters with the French, […], it would make sense to do so. [D2-14-3]

The general noun phrase the matter in example (11) is specified not through modification but through anaphoric reference to the preceding complex noun phrase the case of my constituent Mr Edmond Arapi, who is facing extradition to Italy, having been tried in his absence. The simple general noun phrase is interpreted in reference to the preceding noun phrase (marked by underlining). That way, the general noun phrase is filled with meaning, which makes it rather specific. By contrast, the general noun phrase technical matters in example (12) is 96

less specific. Although the general noun phrase is complex because it is modified by the adjective technical, it does not refer to any other item in the co-text for its interpretation. Therefore, the general noun phrase technical matters is not specified any further and refers to a rather broad category. The functional matrix for the analysis of general noun phrases is a basic tool for a broad classification of these items into four categories of specification. Section 3.2.2 has explained that there is an increase of explicitness from premodification to postmodification and more particularly from phrases to nonfinite clauses to finite clauses (cf. Quirk et al. 1985: 1243, 1321; Jackson 1980: 69). Therefore, the qualitative analysis of general noun phrases will also distinguish the different types of modifier within a general noun phrase. Similar to the distinction of different types of modification, there are different types of semantic links between general noun phrases and other items in the co-text. A general noun phrase can, for example, refer to a preceding noun phrase or to a larger stretch of text. Therefore, the qualitative analysis of general noun phrases will also distinguish different types of endophoric reference. This leads to a finer subdivision of general noun phrases within the ‘most specific’, ‘rather specific’, ‘less specific’ and ‘least specific’ category (see Sections 7.2.1-7.2.4).

4.3 The scale of specification Sections 4.1 and 4.2 have shown so far that the degree of specification expressed by a general noun phrase is dependent on modification and endophoric reference. A general noun phrase such as thing or matter, which is used without any specification such as modification or endophoric reference, is lexically superordinate and constitutes a semantically broad category. We can say that in such cases, general noun phrases live up to their name and are indeed very general in meaning. However, other examples have shown that general noun phrases also occur in utterances where they are specified through modification and endophoric reference and are thus very specific. Such general noun phrases are less abstract than the superordinate general noun phrases referred to above. As has been explained in Section 4.2, the two uses of general noun phrases described above do not constitute the only possible ways for general noun phrases to function in texts. They rather constitute two poles on a continuum or scale on which general noun phrases range from ‘most specific’, to ‘rather specific’, to ‘less specific’, to ‘least specific’. Therefore, the use of general noun phrases can be described along a scale of specification. The following figure illustrates this scale:

97

Figure 4.3: The scale of specification

a abstractio on

general

instan ntiation

specific

Figure 4.3 shows the scale of specification, a de-contextualised concept which can be implemented to describe general noun phrases in naturally occurring language data, thus contextualised general noun phrases. At first glance, the concept illustrated in Figure 4.3 corresponds to the concepts illustrated in Table 3.2 (Section 3.5.1) and Figure 3.6 (Section 3.5.3.1). However, the terms general and specific illustrated in Figure 4.3 do not directly correspond to the terms unique beginners and species used in Table 3.2 or abstract category and specific category illustrated in Figure 3.6. Those terms used in Table 3.2 and Figure 3.6 denote abstract, scientific classes which are defined by a de-contextualised description of features. By contrast, the terms general and specific in Figure 4.3 refer to the degree of specification of contextualised general noun phrases. This degree of specification can only be determined through the co-text and context of a general noun, in particular through the parameters modification and linkage. This has been illustrated in Figures 3.8 (Section 3.5.3.2) and 3.9 (Section 3.5.3.3). Figure 4.3 introduces the terms instantiation and abstraction which can be used to describe the interpretation of general noun phrases. When for the interpretation of general noun phrases their co-text and context is taken into account, general noun phrases are typically encoded on the basis of an instantiation (cf. Anderson et al. 1976: 667). Instantiation then refers to the process of selecting one of the particular meanings encapsulated in a general noun phrase (cf. Anderson et al. 1976: 668). In other words, instantiation here refers to a single realisation of a general noun phrase in an utterance. Instantiation, indicated by the left arrow that reaches from general to specific, is guided by the co-text and context of an utterance. Anderson (1976: 673) nicely illustrates this with the 98

following example: “Without context, a robin may be ideally bird-like, but at the Thanksgiving table a robin is not the best instantiation of bird.” If we apply this to the use of general noun phrases this means that, without reference to the co-text and context, people could be instantiated, for example, by German citizens while this would not be the best instantiation of people in the Labour manifesto, the Conservative manifesto or the Liberal Democrat manifesto. Here, people would probably be instantiated by British citizens. In contrast to the process of instantiation stands the process of abstraction, indicated by the right arrow that reaches from specific to general. Abstraction means that due to a lack of specification through the co-text and context, none of the encapsulated meanings of a general noun phrase is selected so that the general noun phrase remains unspecific.

4.4 General assumptions for the analysis of general noun phrases Partington (1998) remarks that the literature on general noun phrases is scant and that they have undoubtedly not received the attention they deserve. He criticises that general noun phrases are not mentioned in classic reference works such as Quirk et al. (1985), and states that even Halliday/Hasan (1976, 1989) themselves do not owe them much attention in their later works (cf. Partington 1998: 90). If general noun phrases are discussed, the discussion is very often based on perfect textbook examples which do not reveal much about their occurring in actual language data. To the best of my knowledge, Mahlberg (2005) was the first to conduct a comprehensive corpus study of general noun phrases.2 In a similar way, the present study wants to show how general noun phrases are used in authentic language data and how their meaning and their functions become observable through that use. Thus, the very fact that the investigation of general noun phrases and their functions is based on authentic language data can be described in terms of the first assumption underlying the analysis in the present study: 1. General noun phrases are used in naturally occurring language data.

Halliday/Hasan (cf. 1976: 275) state that general noun phrases are superordinate members of major lexical sets and are therefore general in meaning. Based upon 2

Note that Partington (1998) has also based his work on corpus linguistic methods; however, in his work Patterns and Meanings (1998) only a small section deals with general noun phrases.

99

this statement, the present study assumes that the frequency of general noun phrases in a text influences the degree of specification of that text. While a high frequency of general noun phrases is an indication for a high degree of generality of a text, a low frequency of general noun phrases is an indication for a high degree of specification of a text. Provided that, it is assumed that the frequency of general noun phrase heads differs depending on the genre. Genres that are typically characterised as having a high degree of precision are genres from the legal domain (cf. Tiersma 1999: 71). And genres that are typically associated with a low degree of precision or in other words a high degree of generality are genres from the political domain (cf. Gruber 1993: 1, Obeng 1997: 49). Politicians are interested in promoting their ideas and thus try to convince and persuade their audience (cf. Håkansson 1997: 82, 85). They tend to avoid explicit statements and make general promises (see Section 3.4.1 and Section 3.4.2.2 for the use of general noun phrases in political language). A genre that is also typically associated with a high degree of generality is conversation. As will be explained in Sections 5.2.4 and 5.3.2.2, conversation typically takes place in face-to-face interaction with others and is thus characterised by shared contexts of the conversation partners (cf. Esser 2009: 36). Since the meaning of utterances can be reconstructed through the shared contexts, conversation tends to avoid lexical and syntactic elaboration or specification of meaning (cf. Biber et al. 1999: 1044). Considering these aspects, the second assumption for the analysis of general noun phrases in the present study is as follows: 2. Judgments have a low frequency of general noun phrases, manifestos and debates have a high frequency of general noun phrases and conversation has the highest frequency of general noun phrases. This indicates that legal language has a high degree of precision or specification, political language has a high degree of generality and conversation has the highest degree of generality.

Besides a genre-based difference concerning the frequency of general noun phrases, the present study also assumes that the frequency of general noun phrases differs in the spoken and the written medium. This assumption was developed on the basis of the study of spoken and written language by Miller/ Weinert (1998) who conclude that spoken language is characterised by the simplicity of noun phrases in comparison with the complexity of noun phrases that occur in written language, particularly in the language of formal written texts (cf. Miller/Weinert 1998: 135). The simplicity of noun phrases in the spoken medium is closely related to the avoidance of lexical and syntactic elaboration in that medium (cf. Biber et al. 1999: 1044). This is also assumed to be related to a frequent use of general noun phrases which are characterised as unspecific and 100

general in meaning. In contrast, written language is assumed to be characterised by a lower frequency of general noun phrases. See therefore the following assumption: 3. Spoken language is assumed to show a higher frequency of general noun phrases than written language.

Considering the genre- and medium-dependent aspects determining the frequency of general noun phrases, the four sub-corpora of the present study can be presented along the scale of specification. The result is an arrangement from general texts with a high frequency of general noun phrases to specific texts with a low frequency of general noun phrases. Bearing in mind genre- and mediumspecific criteria as explained above, the four sub-corpora are assumed to be arranged on the scale of specification as illustrated in the following figure:3 Figure 4.4: The assumed arrangement of the four sub-corpora on the scale of specification (depending on frequency) general high g frequency q y of ggeneral noun phrases Conversation corpus D b corpus Debate Manifesto corpus Judgment corpus

low frequency of general noun phrases specific p

3

Figure 4.4 presents the four sub-corpora of the present study along the scale of specification. Sections 5.1 and 5.3 present in detail the compilation of the corpus data for these four sub-corpora (cf. Tables 5.1–5.5).

101

Note that the terms general and specific in Figure 4.4 must be restricted: while the frequency of general noun phrases is only a first indication of the generality and specification of the corpus texts, it is, in particular, the parameters modification and linkage which determine the generality and specification of the general noun phrases, and thus of the corpus texts. The arrangement of the four sub-corpora on the scale of specification as illustrated in Figure 4.4 can be formulated in terms of the fourth assumption which states: 4. The conversation corpus, the debate corpus, the manifesto corpus and the judgment corpus can be arranged on the scale of specification from general to specific. In the stated order the frequency of general noun phrases and thus the degree of generality is assumed to decrease.

Assumption four shows that the frequency of general noun phrases is assumed to be first and foremost dependent on the medium. This is why the two spoken corpora are positioned at the general end of the scale. An assumed genre-based difference in the frequency of general noun phrase heads can be seen in the positioning of the conversation corpus at the very general end of the scale followed by the debate corpus, the manifesto corpus and the judgment corpus, which is positioned at the very specific end of the scale. The frequency of general noun phrases is dealt with in the quantitative analysis of the present study in Section 7.1, whereas the form and function of general noun phrases are dealt with in the qualitative analysis of the present study in Section 7.2. Concerning the form of general noun phrases, Section 4.1.1 has explained that general noun phrase heads can be modified or not modified. While non-modified general noun phrases are indeed rather general in meaning, modified general noun phrase heads can be rather specific. Therefore, a general noun phrase is not necessarily general in meaning but its degree of specification depends on its form, more precisely, whether it is modified or not. Section 4.1.2 has explained that the degree of specification of general noun phrases is moreover dependent on whether they are endophorically linked within the co-text or whether they are endophorically not linked. Non-linked general noun phrases are rather general in meaning while linked general noun phrases can be rather specific. Section 4.2 has shown that it is the correlation of the two parameters modification and linkage that fully determines the degree of specification of a general noun phrase. While a general noun phrase without modification and endophoric reference is in fact very general in meaning, a general noun phrase with modification and endophoric reference is quite specific. In reference to what has been said before about the degree of specification in the different 102

genres and media in the corpus of the present study, this leads us to the following assumptions: 5. Judgments have a low frequency of simple general noun phrases without endophoric reference, manifestos and debates have a high frequency of simple general noun phrases without endophoric reference, and conversation has the highest frequency of simple general noun phrases without endophoric reference. 6. Judgments have a high frequency of complex general noun phrases with endophoric reference, manifestos and debates have a low frequency of complex general noun phrases with endophoric reference, and conversation has the lowest frequency of complex general noun phrases with endophoric reference. 7. Spoken language has a higher frequency of simple general noun phrases without endophoric reference than written language and written language has a higher frequency of complex general noun phrases with endophoric reference than spoken language.

Concerning the degree of specification of general noun phrases used in the four sub-corpora, the arrangement on the scale of specification corresponds to the arrangement explained in Figure 4.4 above. See the following figure: Figure 4.5: The assumed arrangement of the sub-corpora on the scale of specification (depending on modification and linkage) general simple general noun phrase structure [– modification], no endophoric reference [– linkage] Conversation corpus D b t corpus Debate Manifesto corpus Judgment corpus

noun complex general no n phrase structure [+ modification], endophoric reference [+ linkage] specific

103

Figure 4.5 shows the arrangement of the four sub-corpora on the scale of specification concerning the presence and absence of the parameters +/- modification and +/- linkage of general noun phrases in the four sub-corpora. The conversation corpus is assumed to be positioned at the very general end of the scale, which means it is assumed to have a high frequency of simple general noun phrases without linkage. The debate corpus and the manifesto corpus follow. The judgment corpus is assumed to be positioned at the very specific end of the scale with a high frequency of complex general noun phrases with endophoric reference.

104

5. Corpus compilation Abstract: The corpus, which was compiled for the present study, includes spoken and written texts from four genres: Supreme Court judgments, political manifestos, parliamentary debates, and general conversations. Chapter 5 describes features of written and spoken language as well as the sampled domains (legal language, political language and conversation).

5.1 The corpus-linguistic approach of the present study The analysis of general noun phrases conducted in the present study emphasises language use and is thus based on naturally occurring language data. As has been critically remarked in Section 2.3, the description of general noun phrases in Halliday/Hasan (1976) builds on well-formed sentences and textbook examples – the plausibility and likelihood of which are called into question. Therefore, the goal of the present study is a functional description of general noun phrases based on authentic and genuine instances of language use which is not guided by what is theoretically possible, but which is concerned with what is likely to occur. The analysis of general noun phrases in the present study is based on a corpus of written and spoken language consisting of texts from different genres and media. The analysis is ‘corpus-based’, which means that the underlying corpus is used to “expound, test or exemplify theories and descriptions that were formulated before” (Tognini-Bonelli 2001: 65). The framework for the analysis of general noun phrases (cf. Sections 4.1–4.3) and the assumptions concerning the functions of general noun phrases across different genres and media (cf. Section 4.4) will be tested against the underlying corpus of the present study. For this purpose, a corpus consisting of four smaller sub-corpora was compiled. The judgment corpus consists of eight British Supreme Court judgments; the manifesto corpus consists of the manifestos of the Labour Party, the Conservative Party and the Liberal Democrat Party; the debate corpus consists of eight parliamentary debates of the British House of Commons; and the conversation corpus consists of three everyday conversations taken from the spoken section of the British National Corpus (BNC). These four sub-corpora represent four different genres and two media. See the following table for an overview of the corpus of the present study:

105

Table 5.1: The corpus of the present study Domains Legal language Political language Conversation

Corpus compilation Genres Sub-corpora Supreme Court judgments Judgment corpus Political manifestos

Manifesto corpus

Parliamentary debates

Debate corpus

General conversation

Conversation corpus

Medium written spoken

Certain aspects of corpus design were considered for the compilation of the corpus so that it can serve the purpose of the analysis in the present study. At first, aspects of representativeness, corpus size and balance were considered. The question of corpus representativeness is problematic. While it is sometimes argued that corpora should be as large as possible (cf. Sinclair 1991: 18), corpus size is not a reliable proof for representativeness since any corpus, however large, is by definition finite and will thus never be able to capture the infinite resources of language (cf. Lipka 1972: 155). Corpus representativeness is closely associated with sampling. Given that natural language cannot be described exhaustively, sampling is unavoidable to achieve a representativeness which matches the research questions (cf. McEnery et al. 2006: 19). However, sampling cannot solve the problem of representativeness. Kennedy (cf. 1998: 62) states that corpus representativeness always remains a question of achieving the ideal. He points out that the notion of representativeness is, in the final analysis, a matter of judgment and can only be approximate. The sampling of the texts in the corpus of the present study was based on the general assumptions concerning the use of general noun phrases across different genres and media (cf. Section 4.4). In order to test whether the use of general noun phrases is dependent on the medium, written and spoken texts were compiled for the corpus. In order to test whether the use of general noun phrases depends on the genre, texts from different genres were compiled for the corpus. And finally, in order to test whether the use of general noun phrases correlates with the degree of precision and generality of a text, and with communicative aspects such as the intention of the encoder, texts from the domains of legal language, political language and conversation were compiled for the corpus of the present study. For the purpose of the analysis in the present study, the sections of the corpus were equally weighted, which means that the four sub-corpora roughly have the same size and are composed of equal parts of spoken and written texts. Given the fact that the corpus used for the present study represents four genres, namely 106

judgment, manifesto, debate and conversation, it can be defined as a specialised corpus. Another important aspect of corpus compilation is the corpus mark-up, a system of standard codes inserted into a document stored in electronic form to provide information about the text itself and govern processes such as formatting (cf. McEnery et al. 2006: 22). This information includes text features such as line breaks, line numbers, word breaks and boundaries, and hesitations (cf. Kennedy 1998: 82). This form of coding is included in the conversation corpus of the present study which is extracted from sections of the spoken British National Corpus (BNC).

5.2 Medium-, domain- and genre-specific features of the corpus data 5.2.1 Some features of spoken and written language The study of spoken and written language has a long history that goes back to the ancient Greek Aristotelian rhetoric. Since the middle of the twentieth century, linguists and anthropologists have attempted to define the similarities and differences between spoken and written language (cf. Roberts/Street 1997: 168). However, very often the terms spoken language and written language are used to describe several distinct phenomena. In this context, Esser (cf. 2006: 24) points to the polysemy of these notions, which is threefold: firstly, the terms spoken language and written language refer to the medium in which language-signals are realised. While spoken language is transmitted via sound waves, written language is transmitted via graphic marks (cf. Schäpers 2009: 3). Esser (cf. 2006: 24) refers to this difference in the substance of encoding with the terms phonic and graphic patterns of word forms. Note that the differences in realisation include medium-dependent devices, such as stress and intonation in spoken language, and punctuation in written language (cf. Schäpers 2009: 3). Secondly, the terms spoken language and written language point to the origin of the first encoding. Texts that have their origin in speech, such as conversations, must be distinguished from texts that are only presented in their spoken realisation but have their origin in writing, such as judgments (cf. Esser 2006: 24). Thirdly, the terms spoken language and written language entail a distinction of the underlying abstract grammatical and lexical forms, thus, the medium-independent grammatical and lexical word forms that can be transferred from the spoken to the written medium and vice versa (cf. Schäpers 2009: 5). In this context, Esser (1984: 1) points out a major aspect: 107

Für die Klärung des Begriffs ‘gesprochenes Englisch’ ist wichtig, daß die Unterschiede ‘gesprochen’/’geschrieben’ nicht nur auf einen Substanzunterschied (phonisch/graphisch), sondern auch auf einen (stilistischen) Formunterschied zu beziehen ist.1

This statement points to a structural and a stylistic level of distinction between spoken and written language concerning, for example, the degree of complexity, as investigated by Halliday (1987) and later by Schäpers (2009). The degree of complexity of spoken or written language can be manifested in different realisations in the two media concerning the different structural levels of phrases and clauses. Accordingly, Miller and Weinert (1998: 4–5) point out: the terms ‘spoken language’ and ‘written language’ […] relate to partially different systems of morphology, syntax, vocabulary, and the organisation of texts. Each code, the spoken and the written code, shows different degrees of complexity on the level of phrases and sentences (cf. Schäpers 2009: 8). For the purpose of the present study, the level of phrases (namely the level of general noun phrases) is of particular importance. Schäpers’ study of nominal and clausal complexity in spoken and written language (2009) builds on Halliday’s assumption that written language and spoken language show different degrees of complexity on different levels (cf. 1987: 71): while writing assumingly shows a high degree of complexity on the level of phrases, spoken language assumingly shows a high degree of complexity on the level of clauses. Thus, writing seems to be characterised by having more lexical items per clause but fewer clauses. Spoken language seems to be characterised by having more clauses within a clause complex but fewer lexical items per clause. This goes in line with Chafe (cf. 1982: 37–39) who states that differences in spoken and written language can be attributed to differences in the process of speaking and writing. Accordingly, the slow pace of writing allows for an integration of complex ‘idea units’2 (and thus complex phrases within few clauses), while the rapid pace of spoken language leads to a fragmentation which shows up in a stringing together of smaller idea units (simple phrases within many clauses). Speaking about distributional differences in the complexity of noun phrases, Schäpers (2009) refers to Miller and Weinert (1998) who analysed complex noun phrases in the spoken and the written medium. The overall conclusion that Miller and Weinert draw from their data is that spontaneous spoken language 1 2

It is important for the clarification of the term ‘spoken English’ that the difference ‘spoken’/’written’ is not only to be referred to a difference in substance (phonic/graphic) but also to a difference in form (stylistic). (Translation provided by V.B.) According to Chafe (cf. 1982: 37), idea units typically have a coherent intonation contour and each represents a single idea.

108

is characterised by the simplicity of noun phrases in comparison with the noun phrases that occur in written language, particularly in the language of formal written texts (cf. Miller/Weinert 1998: 135). It can be concluded from the above discussion that the complexity of structural units such as the noun phrase is dependent on the medium in which it occurs. This is closely related to the distinction of two stylistic language uses, the restricted style of referencing and the elaborated style of referencing introduced by Bernstein (1971), which was discussed in Section 3.3.3. Especially conversations, which typically take place in shared contexts, avoid lexical and syntactic elaboration or specification of meaning and therefore have a very low frequency of nouns, especially of complex noun phrases (cf. Biber et al. 1999: 235, 578). In written discourse, however, referencing relies more strongly on endophoric reference and the reference items must be properly introduced in forms of lexically filled, complex noun phrases (cf. Esser 2009: 36). This aspect will be taken up in Section 5.2.4 when discussing some features of conversation. In the same context, Nystrand (cf. 1987: 198) contrasts the explicitness and context-free character of written language with the implicit and context-dependent character of oral language. He states that written texts, unlike spoken, must function apart from the context of their production. Written texts must thus be ‘autonomous’ and ‘explicit’ in order to fulfill their function detached from the context (cf. 1987: 198). In accordance with Kay (cf. 1977: 21–22, 24), Nystrand (cf. 1987: 211) further points out that the explicitness of written texts and the implicitness of spoken texts is not a matter of right or wrong but merely a matter of appropriateness. Nystrand (cf. Nystrand 1987: 211) emphasises that writing and speech work differently as language systems and that the maintenance of coherence is simply a question of appropriateness. While spoken language can be more abbreviated, written language must be more elaborated. This aspect is of major importance for the present study since the use or rather the overuse of general noun phrases may lead to a degree of implicitness in a text which is no longer appropriate in the context in which the text is used.

5.2.2 Some features of legal language During the past three decades, ‘legal language’ has become a subject of linguistic analysis (cf. Tiersma 1999: 1). According to Tiersma (cf. 1999: 49), legal English is a variety of English that follows the rules that govern English in general but diverges stylistically from ordinary speech. Tiersma discusses these stylistic differences concerning first and foremost syntactic and lexical features. According to Tiersma (cf. 1999: 55–56), legal language is considered notorious for its lengthy 109

and complex sentences with many embeddings. It is typical of legal syntax to insert a huge quantity of information between the subject and the verb phrase (cf. Tiersma 1999: 57). The motivation for such a complex syntax is the desire to encompass all information on a particular topic as legal language ought to be as precise and specific as possible (cf. Tiersma 1999: 56, 71). In this context, Charnock points out that it is necessary, from a linguistic point of view, to distinguish between the language of statutes and the language of judgments (cf. 2010: 132). In his view, the often stated syntactic complexity of legal language mainly applies to statutes and contractual texts because these are essentially written legal documents which were never intended to be read from beginning to end, but are mainly used for reference (cf. 2010: 113–114). By contrast, judgments are written documents which were originally read out in court, which is why Charnock states that they show many of the characteristics of oral discourse (cf. 2010: 113).3 What Charnock disregards is that judgments recurrently make reference to and quote acts and statutes, which is why I still assume that complex sentences can also be found in judgments. Together with the aforementioned complexity of noun phrases in the written medium (cf. Schäpers 2009: 153), I assume the use of complex and thus specified general noun phrases to be quite frequent in the written Supreme Court judgments (cf. assumption 6 in Section 4.4). If this is the case, we could say that these complex general noun phrases are an indication of the specific and precise nature of Supreme Court judgments. In this context, Gibbons (2004) points to a very interesting aspect: following Halliday (1985), Gibbons (cf. 2004: 2, 4) states that the long and complex noun phrases typically used in legal language are a product of de-contextualisation. This is due to the fact that language in legal documents attempts to encompass linguistically general classes of events and participants. Because these documents deal with events that are distant in time and place, the language is distant in character. This can be manifested through the use of complex noun phrases which are a result of the abstraction of ideas and complex concepts. Besides special syntactic features of legal language, Tiersma (cf. 1999: 79) also discusses vagueness as one of the lexical characteristics of legal language. According to Tiersma (cf. 1999: 67), legal documents such as in particular statutes are meant to be of general applicability and address several audiences at once. 3

Legal documents in general can be either written or spoken. The present study is concerned with Supreme Court judgments which can be defined as written to be spoken. This means that judgments have their origin in writing and are, for the purpose of the present study, therefore defined as written legal documents (cf. Charnock 2010: 113–114).

110

This can be achieved through the use of flexible, general or vague terms that leave enough room for interpretation (cf. 1999: 79). However, vagueness, although it is typical of legal documents such as statutes, is not typical of legal judgments. According to Charnock (cf. 2010: 132), the language used in judgments is explicit and clear. While lawyers rely on statutes for their argumentation, they benefit from flexible and vague terms that leave enough room for different interpretations. On the contrary, judges settle for one interpretation in order to make their decision in a case. Their decision then does probably not include many flexible and vague terms. This again supports my aforementioned assumption that the use of specific and complex general noun phrases should be frequent in judgments while vague, simple general noun phrases should be rather rare in judgments (cf. assumption 5 and 6 in Section 4.4).

5.2.3 Some features of political language The study of political language and political communication can trace its roots back to Ancient Greek and the earliest classical studies of Aristotle and Plato and is closely related to the studies of rhetoric as the art of verbal persuasion (cf. Kaid 2004: xiii). The study of political language is an interdisciplinary field of research drawing on concepts of e.g. communication, political science, journalism, sociology, rhetoric and other fields (cf. Kaid 2004: xiii). Since the research field of language and politics is very complex, the present section can only concentrate on one major aspect of political communication, namely persuasion. Since persuasion itself constitutes a rather broad concept, the present section focuses on one aspect of persuasive communication which is relevant for the present study: the aspect of vague or indirect communication in political language. The primary goal of politicians in an election is to get elected, and in order to achieve that goal they need to communicate their views and political standpoints in a way that persuades the possible voter to do so (cf. Håkansson 1997: 82). Sjöblom (1968: 30) characterises elections as games of power and persuasion in which parties play roles determined by a strategic consideration of the actors. Especially election campaigns preceding national elections constitute crucial test points, where communication is of special importance to both politicians and citizens (cf. Håkansson 1997: 82). This holds true especially for party-political external communication, e.g. in the form of political manifestos as investigated in the present study. When communicating their political ideas and standpoints, politicians find themselves in a dilemma: on the one hand, the potential electorate needs to be provided with relevant information to be convinced of the rightness of the 111

parties policies and to convey the message that the country is in strong, competent hands (cf. Halmari 2005: 108). On the other hand, too much information can be off-putting, therefore, unpleasant and politically risky topics need to be avoided. Obeng (1997: 49) describes how politicians solve this conflict: in order to achieve certain political goals, politicians tend to communicate indirectly and vague, especially when the topic of the discourse “communicates difficulty”. The use of unspecific general noun phrases is one means of communicating indirectly (see Section 3.4.2.2). The motivations for a politician to communicate indirectly and vaguely are diverse and range from political to personal interests: indirectness is motivated first and foremost by political interests and political necessity and as such it can serve for example as a marker of ‘diplomacy’ and of politeness (cf. Obeng 1997: 49, 52). In this context, Obeng states (1997: 52): Care must be taken during speech since an improper verbalization may lead to serious political or diplomatic consequences […] and to the compromising of the politician’s face integrity. Speaking candidly is therefore oftentimes seen as confrontational, impolite and politically risky.

Closely related to the foregoing is the pursuit of personal interests of politicians: politicians communicate vaguely to assure and maintain their personal face-saving and to avoid face-threatening acts. Gruber states (1993: 3): […] every politician has a public positive face […] which claims the consistent image of himself as being a rational, trustworthy person whose political ideas and actions are better fitted to the wants and demands of the general public than those of his opponents.

Through the use of indirect communication, this image must be maintained, which constitutes a face-saving act, while the image of the opponent must be threatened. Furthermore, indirect communication can help politicians to communicate topics which are tabooed in society but nonetheless useful for their political aims (cf. Gruber 1993: 3). Finally, politicians communicate indirectly because they must address different audiences at the same time which means they must convince most different target groups with individually different needs and wants (cf. Gruber 1993: 3). Gruber (cf. Gruber 1993: 1) states that vagueness occurs most often in the area of (party)-political external communication, e.g. in political manifestos, in which politicians communicate with the general public in order to convince them of their programs or ideas. This aspect is of major importance for the analysis of general noun phrases in the present study since these items, as has been shown in Section 3.4.2.2, can be used to express vagueness. Referring to assumption 5 (cf. Section 4.4) this means that particularly simple, unspecific general noun phrases are supposed to be frequent in the political manifestos. 112

5.2.4 Some features of conversation Conversation can be defined as “any interactive spoken exchange between two or more people” (Pridham 2001: 2). It includes face-to-face exchanges such as talk between friends and family or more public and ritualised exchange forms such as classroom talk. Another example for the latter type is ‘Question Time’, the replying to parliamentary questions as part of the debate procedure in the British House of Commons. This genre is also included in the spoken corpus of the present study and will be further discussed in Section 5.3.2.1. Besides faceto-face exchanges, conversation also includes non-face-to-face exchanges such as telephone conversations, and broadcast materials such as a live radio phone-in or a television chat show (Pridham 2001: 2). There are various approaches to the study of conversation. Within a macrolinguistic framework, conversation is studied in terms of the use of language in communication. Furthermore, linguists are interested, for example, in functional aspects of conversation, speech acts and conversational processes, and the structure of conversation (cf. Tsui 1994: 3–4). One major framework for the study of conversation is called Conversation Analysis and is applied to analyse how conversations are socially organised and managed by participants. This includes aspects such as turn-taking mechanisms, sequence organisation such as the use of adjacency pairs, and repair (cf. Levinson 1983: 296 ff., 303 ff.) Within a micro-linguistic framework, conversation is studied in terms of the specific features of spoken language some of which have already been discussed in Section 5.2.1. In the context of the present study, this particularly concerns the nonnominal character of speech. Schäpers (cf. 2009: 153) found out that spoken language is characterised by a simple noun phrase structure. These results are in accordance with the findings of Biber et al. (cf. 1999: 235, 578) who showed that conversation, as one example of spoken text, typically has a very high frequency of pronouns and a very low frequency of nouns, especially complex noun phrases. The higher frequency of pronouns in conversation can be explained by the shared situation and the personal involvement of the speakers. They avoid lexical and syntactic elaboration as found in written registers, because conversation relies on context for meaning. As for noun phrase structure, which is of special interest for the present study, Biber et al. (cf. 1999: 235, 578) state that speakers use fewer nouns, and if they do, they use fewer elaborated noun phrases that contain modifiers and complements. Consequently, I assume the conversation corpus of the present study to reveal a high frequency of simple general noun phrases (cf. assumption 5 in Section 4.4). In their discussion on spoken and written language, Carter/Mc Carthy (cf. 2007: 19) state that one major characteristic of spoken language, and thus 113

also of conversation, is vagueness. Vagueness in spoken language is often not harmful as through the shared speech-situations decoders are nevertheless able to reconstruct the meaning of a vague utterance (see Section 3.4.2.2). For the analysis of the present study, it will be interesting to see whether the conversational corpus supports these statements and reveals cases of vagueness using simple, unspecific general noun phrases.

5.3 The corpus of the present study 5.3.1 Written corpus 5.3.1.1 Supreme Court judgments Established by the Constitutional Reform Act 2005, the United Kingdom Supreme Court is the final court of appeal within the United Kingdom, replacing the Appellate Committee of the House of Lords (cf. Dickson 2013: 2; cf. UKSC Annual Reports 2009–2010: 9). The Supreme Court is the highest appellate court for all civil cases within the United Kingdom and for all criminal cases from England, Wales and Northern Ireland. The cases that are heard are cases of the greatest public or constitutional significance (cf. Dickson 2013: 4; cf. UKSC Annual Reports 2012–2013: 19). The Supreme Court has twelve justices, including a president, which is currently Lord Neuberger (cf. Dickson 2013: 7; cf. UKSC Annual Reports 2012–2013: 15). However, not all twelve justices hear every case. Typically a case is heard by a panel of five justices, though sometimes the panel may consist of three, seven or nine members (cf. UKSC Annual Reports 2012–2013: 26). Supreme Court judgments have been chosen as corpus data for the analysis of the present study to represent written legal language in the United Kingdom. According to Moens (cf. 2007: 1749), Supreme Court judgments follow a conventional schematic form and typically contain five parts: the first part refers to the legal basis of the case such as the cause of action and type of complaint. The second part explains the background of the case such as the description of the underlying events and undisputed legal concepts. The third part contains the disputed legal question, as for example an alleged offense in a criminal case. The fourth part contains the justifying theory or the opinion and the fifth part then contains the decision of each of the judges involved in the judgment. According to Pérez/Rizzo (cf. Pérez/Rizzo 2012: 131), judgments are fundamental as a legal genre due to the central role they play in the common law system: judgments stand at the very core of the common law system, such as in the United Kingdom, acting as the main source of law followed by statutes and equity. Therefore, judgments hold a prominent position in legal English (cf. Pérez/ 114

Rizzo 2012: 135). Moens (cf. 2007: 1748) explains that lawyers and courts value past cases and use them in precedent reasoning. This holds true especially for Supreme Court judgments since cases heard at the Supreme Court, as the final court of appeal, follow a complex and long rout of appeal that implies much greater argumentation and case citation than a case tried at a first-tier tribunal (cf. Pérez/Rizzo 20112: 138). There are some characteristic features of judgments of common law courts: according to Mattila (cf. 2006: 85), judgments of common law courts are usually written in a ‘personal’ style. While the position taken by the court is expressed by the ‘we’ form, that of the judges as individuals is expressed by the ‘I’ form. Lashöfer (cf. 1992: 13) states that the personal style of the Supreme Court judgments can be explained with the high prestige which judges in the United Kingdom generally have. As Mattila (2006: 85) formulates it, “[…] the judges as individuals are clearly visible: their grounds are full and detailed, their language often colourful”. This aspect points at another characteristic feature of judgments of common law courts: the expression of personal opinions and feelings of the judges (cf. Lashöfer 1992: 19). Charnock (cf. 2010: 132) explains that British Supreme Court judges often make use of an idiosyncratic, sometimes even conversational style. In addition, it has been noted that House of Lords judges sometimes use idiomatic expressions and rhetorical devices (cf. Lashöfer 1992: 22, 25). See, for example, the dictionary entry in the Oxford Dictionary of Idioms (Siefring 2004: 182) for “the man on the Clapham omnibus” meaning “the average man, especially with regard to his opinion”: This expression is attributed to the English judge Lord Bowen (1835-94), who used it as a metaphor for any ordinary reasonable person – such as a juror is expected to be. Clapham is a district in South London.

The Supreme Court judgments that were chosen for the corpus of the present study have been downloaded in pdf-format from the Supreme Court website.4 Since the Supreme Court gives free access to all decided cases the day after judgments have been handed down, I was able to choose from a great number of texts. One criterion for the selection of the eight judgments was that they roughly belong to the same time period; all the texts chosen date from the period between 2010 and 2011. Another criterion was that all the judgments chosen dealt with cases that belong to the same field of law. Accordingly, eight judgments were

4

http://www.supremecourt.gov.uk/decided-cases/index.html (last access, 30th January 2013)

115

chosen that all dealt with cases of civil law. For an overview of the judgments chosen for the judgment corpus, see Section 5.3.3.

5.3.1.2 Political manifestos The 2010 general elections in Great Britain marked a transition from a Labour governed system to a more complex political geography: the Liberal Democrats joined the Conservative Party in a coalition government (cf. Driver 2011: 1). This meant a shift from a two-party system (Labour and Conservatives as relevant parties) to a multi-party system (Labour, Conservatives and Liberal Democrats as relevant parties) (cf. Driver 2011: 2). British electoral politics was becoming more competitive and a reduced number of political ‘diehards’ (committed voters unwavering for support of ‘their’ party) has had important consequences for the organisation of political parties and how they go about winning votes. Examples include the professionalisation of politics and campaign trails that were more expensive than ever before (cf. Driver 2011: 32). Labour, Liberal Democrats and Conservatives must have had a particular interest in winning votes and persuading the electorate. Therefore it seemed very promising to investigate the 2010 election manifestos of these parties. Kavanagh (2000: 1) states that “it is a boast of British parties that they are programmatic. They fight general elections on manifestos and, if elected, promise to carry them out.” Kavanagh (cf. 2013: 7) adds, however, that in their wish to win elections, the parties are claimed to heat up popular demands and expectations, often with little regard to a government´s capacity to meet them. A high degree of persuasion strategies in the 2010 political manifestos is of particular interest for the analysis of general noun phrases in the present study. It has been shown that these items are a means of indirect and vague communication which persuades the decoder. Election manifestos have been chosen as corpus data for the analysis in the present study as an example of written political language in the United Kingdom. Election manifestos are little booklets which are intended for interested voters to give them an overview of the political standpoints, aims and proposals of a party for the upcoming parliamentary term (cf. Håkansson 1997: 92). The general election manifestos that are analysed in the present study have been chosen from the three British political parties that played a major role in the general election 2010: the Labour Party, the Liberal Democrats Party and the Conservative Party. As mentioned above, in 2010, British electoral politics was becoming more competitive than before because of the reorganisation of political parties. A criterion for choosing the 2010 election manifestos of Labour, the Liberal Democrats and the Conservatives

116

was therefore a supposedly high degree of linguistic persuasion strategies, such as indirect communication through the use of general noun phrases. In order to constitute a balanced corpus in terms of size, the manifesto subcorpus is approximately as large as the judgment sub-corpus. Moreover, the manifestos chosen are all from the same election, namely the 2010 general election. Another criterion concerns the compatibility of the three political manifestos concerning content. The party manifestos chosen for the analysis in the present study are concerned with quite similar political issues, namely economics, society, education, healthcare, immigration and environment. The manifestos were downloaded in pdf-format from the corresponding websites where they are open available for inspection.5 All three party manifestos date back to the 2010 general election in the United Kingdom. For an overview of the manifestos chosen for the manifesto corpus, see Section 5.3.3.

5.3.2 Spoken corpus 5.3.2.1 Parliamentary debates Britain can be described as a ‘unitary state’, which means that power exclusively lies in a central government. The guiding principle of British government is parliamentary sovereignty which means that there is no higher authority than the parliament6 and that the legitimacy of its decisions cannot be questioned by any other national body (cf. McAnulla 2006: 13). The Parliament of the United Kingdom is the central legislative body which means it can make, amend or repeal any law. Furthermore, it can create or bring down a government. The Prime Minister and the Cabinet carry out the executive function and the courts carry out the judicial function (cf. Williams 1998: 7–10). The parliament is bicameral with a lower house, the House of Commons, and an upper house, the House of Lords. While the former is elected democratically by the people, the latter is based on a system of appointment by the Sovereign on the advice of the Prime

5

6

For the manifesto of the Labour Party, see the following link: http://www.labour.org. uk/labours-manifesto-for-a-future-fair-for-all (last access 30th January 2013). For the manifesto of the Liberal Democrats, see the following link: http://www.libdems. org.uk/our_manifesto.aspx (last access 30th January 2013). For the manifesto of the Conservative Party, see the following link: http://www.conservatives.com/News/ News_stories/2010/04/Conservatives_launch_election_manifesto.aspx (last access 30th January 2013). Theoretically, the monarch is above the parliament as they hold the power to dissolve parliament.

117

Minister and has only reduced powers (cf. Williams 1998: 7–10). The composition of the House of Commons determines the party of government and the leader of the party then becomes Prime Minister (cf. Williams 1998: 8). See the following figure taken in a slightly modified version from Williams (1998: 8) to demonstrate the structure of the parliamentary government in Britain: Figure 5.1: The structure of the parliamentary government in the UK

Prime Minister Cabinet

Government

Junior Ministers

House of Lords

House of Commons

Parliament

The British people

On a regular basis, the House of Commons and the House of Lords hold parliamentary debates in which members discuss government policy, propose new law and current political issues. These discussions are often lively with ministers intervening on each other’s speeches. Although debates follow a strict set of rules, they are still rather dynamic discussions where ministers respond to the points made by others instead of reading out formal speeches (cf. http://www. parliament.uk/about/how/role/debate/, last access 17th December 2014). During a debate, the Speaker, or Deputy Speaker, calls MPs in turn to give their opinion on an issue. Very often then, MPs intervene with opinions, expressions of approval or disapproval, or humour. It is the Speakers role on the one hand to ensure that MPs observe the courtesies of debate, and on the other hand to allow a lively debate to take place (http://www.parliament.uk/about/ how/guides/people-events-and-places/, last access 17th December 2014), (http:// www.parliament.uk/documents/commons-information-office/Debates.pdf, last access 17th December 2014). 118

One particular part of the daily parliamentary business of both Houses is the putting of questions. The replies to parliamentary questions are published daily as part of the debates and form part of the ‘Official Hansard Report’.7 According to Chilton (cf. 2004: 92), the putting of questions in the British parliament has been acknowledged to be an important sub-genre of parliamentary discourse. The specific institution of ‘Prime Minister’s Question Time’ is characteristic of the British Parliament (cf. Chilton 2004: 92). Asking a question in the British House of Commons is the culmination of a lengthy discourse process, involving several channels, speakers, and writers. Members have to give notice of questions to officials (the Clerks in the Table Office) in advance and officials from the relevant ministerial departments can prepare answers and will give them orally from a written brief. In general, a member is called by the Speaker, and the member puts their question, after which point the questioner and other members are invited to put follow-onquestions. This often leads to lengthy discussions (cf. Chilton 2004: 93). The current system for oral answer questions is substantially unchanged since 1906 and there are certain rules that ought to be followed in parliamentary debates (cf. Chilton 2004: 95).8 The most important rules concern the purpose of a question which is to obtain information or press for action and should not be framed primarily as to convey information or in effect a short speech. Furthermore, questions which seek an expression of opinion, or which contain arguments, inference or imputations, unnecessary epithets, or rhetorical, controversial ironical or offensive expressions are not in order. Chilton (cf. 2004: 95) remarks that the existence of these rules does not prevent their being broken. For the analysis in the present study, parliamentary debates have been chosen as an example of spoken political discourse. The comparison of written political manifestos and spoken political debates will show whether the use of general noun phrases differs not only across genres but also across media. As mentioned before, the lively debates in British parliaments show conversational elements. While written political discourse is supposed to show a frequent use of simple general noun phrases as a means to communicate indirectly and vaguely, it will be interesting to investigate whether spoken political discourse (with a slight conversational element) shows an even more frequent use of simple general noun phrases. 7

8

‘Hansard’ is the official record of parliamentary proceedings in the United Kingdom. It contains transcripts of parliamentary debates including votes, written statements and written answers to questions (cf. http://www.parliament.uk/about/how/publications/ hansard/, last access 11th March 2015). Chilton here quotes from “Erskine May”, one of the rule books of the UK parliament.

119

The parliamentary debates of both the House of Commons and the House of Lords are recorded in a publication called ‘Hansard’ which is published on a daily basis and is available online or in print.9 Apart from that, Parliamentary TV gives access live streams of the parliamentary debates of both houses.10 In order to constitute a balanced corpus in terms of size, the debate sub-corpus is approximately as large as the manifesto and the judgment sub-corpus. Note that it was not possible to sample entire debates since they are so long that this would have exceeded the scope of the sub-corpus. Therefore, only the section “Oral Answers to Questions” was selected from randomly chosen eight House of Commons debates. These debates all range from 2010 to 2011 and are therefore approximately from the same period of time as the manifestos and judgments chosen. The debates (Official Reports) were downloaded in pdf format from the parliament website.11 For an overview of the debates chosen for the debate corpus, see Section 5.3.3.

5.3.2.2 General conversations The general conversation corpus of the present study is sampled from the spoken section of the British National Corpus (BNC). The BNC is a sample of approximately 100 million words of present day spoken and written British English with a large majority of texts (over 93%) dating from the period 1985–1994 (cf. Leech et al. 2001: 1). The BNC consists of 4,124 different text files with 90% of the corpus representing current written, and 10% of the corpus representing current spoken British English. It is noticeable that the spoken part of the BNC is proportionally fairly small in contrast to the written part. This is due to the skilling and timeconsuming task of transcribing speech into computer readable orthographic text (cf. Leech et al. 2001: 1). The BNC was designed to be a finite, balanced, sampled corpus representing current British English as a whole including a wide range of different genres, subject fields and registers (cf. Kennedy 1998: 50). The written section of the BNC consists of about 80% ‘informative’ prose (non-fictional expository writing) and about 20% ‘imaginative’ prose (mostly fiction, but including some other literary texts such as poetry) (cf. Leech et al. 2001: 2). The spoken section of the BNC is divided into two parts: the conversational part which makes up about 40% of the spoken section and the taskoriented part which makes up about 60%. While the conversational part of the 9 http://www.parliament.uk/business/publications/ (last access 17th December 2014) 10 http://www.parliamentlive.tv/Main/Home.aspx (last access 17th December 2014) 11 http://www.parliament.uk/business/publications/hansard/commons/bydate/# session=26&year=2014& month=11&day=16 (last access 17th December 2014)

120

spoken BNC corpus is largely made up of every day spontaneous interactions, the task-oriented part of the BNC represents different types of task-oriented spoken activities that were “[…] unlikely to be recorded by the conversational volunteers during a typical day in their lives […]” (Leech et al. 2001: 3). This means that the task-oriented part of the spoken BNC corpus represents spoken British English which is typically used in public activities such as lectures, consultations, sermons, TV/radio broadcasting, etc. (cf. Leech et al. 2001: 3). The general conversations taken from the BNC represent spoken British English, more specifically spontaneous conversational English from the United Kingdom. One major criterion for the selection of conversational English as a sub-corpus of the present study is the fact that I assume spontaneous conversational English to show a frequent use of simple general noun phrases. Section 5.2.4 has discussed some linguistic aspects of conversation. It has been explained that spoken language in general and conversational language in particular is characterised by a simple noun phrase structure (cf. Biber et al. 1999: 235, 578) and also by vague utterances (cf. Carter/Mc Carthy cf. 2007: 19). These utterances can only be reconstructed by the decoder through the shared context of conversational situation. For the analysis of the present study it will be interesting to see whether the conversational corpus supports these statements and reveals cases of vagueness through the use of simple, unspecific general noun phrases (cf. assumption 5 in Section 4.4). The texts from the conversational part of the BNC were recorded by some 124 adults aged over 15. These adults were recruited systematically from 38 different parts of the United Kingdom and represent four different socio-economic classes with balanced coverage of gender and age (cf. Kennedy 1998: 51). The conversational part of the BNC corpus was intended to record all the spoken conversational interactions of the volunteers during a typical day of their lives. The volunteers were asked to carry a walkman and in many cases the only person aware that the conversation was being taped was the person carrying the recorder. That way all conversations were recorded as unobtrusively as possible and the material gathered approximated closely to natural, spontaneous speech (cf. Leech et al. 2001: 3). The volunteers were asked to record all of the conversations they were involved in over a two to seven-day period. Therefore, one BNC conversational text file (indicated by a text classification code or header) includes many separate conversations by one volunteer (cf. Burnard 2000: 13). For the sampling of the conversation sub-corpus, I was able to choose from 153 conversational text files in the spoken part of the BNC corpus. The conversations that I sampled for the conversation sub-corpus were chosen randomly. In order to get a balanced corpus in terms of size, the conversations chosen had approximately the same number of words as the texts in the other three sub-corpora. 121

Three conversations were chosen: conversations 1 and 2 were taken from the BNC conversational text file with the code name KB6. This means that both conversations were recorded from one and the same volunteer. According to the BNC User’s Reference Guide (cf. Burnard 2000: 298, 305), KB6 is a text file with eight conversations recorded by ‘Angela’ (speaker PS029) between 2nd and 6th December 1991. I randomly chose two out of the eight conversations. Conversation 3 was taken from the BNC conversational text file with the code name KCD. This text file includes 107 conversations recorded by `Helen’ (speaker PS0E8) between 31th May and 1st June 1991. I randomly chose one conversation. For an overview of the conversations in the conversation corpus, see the following section.

5.3.3 Summary: Overview of corpus data For an overview of the corpus data sampled for the analysis in the present study, see Tables 5.2–5.5 below: Table 5.2: The judgment corpus Supreme Court Judgments Title in the present study

Official title

Judgment 1 (J1)

[2010] UKSC 47

Judgment 2 (J2)

[2010] UKSC 50

Judgment 3 (J3)

[2010] UKSC 54

Judgment 4 (J4)

[2010] UKSC 55

122

Content the proper construction of an option clause in a lease of land a granted standard security over a house in favour of the Royal Bank of Scotland the Secretary of State’s right to recover certain social security benefits profits of a limited company not in liquidation which are distributed to shareholders

Year

No. of words

2010

7,814

2010

14,155

2010

7,260

2010

8,361

Total no. of words

78,473

Supreme Court Judgments Title in the present study

Official title

Content

Judgment 5 (J5)

[2010] UKSC 57

Judgment 6 (J6)

[2010] UKSC 58

Judgment 7 (J7)

[2011] UKSC 15

Judgment 8 (J8)

[2011] UKSC 22

the decision of two costs officers appointed by the president of the SC loan transaction and taxable incomes planning permissions for a hay barn which was in fact used as a dwelling house a refused application of asylum

Year

No. of words

2010

10,853

2010

8,127

2011

16,120

2011

5,783

Total no. of words

Table 5.3: The manifesto corpus Title in the present study

Political Manifestos Official title

Content

economy, The Labour Party education, health, Labour Manifesto Manifesto – A immigration, (LM) future fair for all family, community, environment The Liberal Democrat Manifesto – values, money, job, Liberal Democrat change that life, community, Manifesto (LDM) works for you, finances building a fairer Britain The Conservative Manifesto – economy, society, Conservative invitation to join environment, Manifesto (CM) the government national interest of Britain

Year

No. of words

2010

26,883

2010

21,450

2010

28,715

Total no. of words

77,048

123

Table 5.4: The debate corpus Title in the present study Debate 1 (D1)

Debate 2 (D2)

Debate 3 (D3)

Debate 4 (D4)

Debate 5 (D5)

Debate 6 (D6)

Debate 7 (D7)

Debate 8 (D8)

124

Parliamentary Debates Official title House of Commons Official Report. Vol. 510 no. 8 House of Commons Official Report. Vol. 515 no. 44 House of Commons Official Report. Vol. 517 no. 65 House of Commons Official Report. Vol. 521 no. 102 House of Commons Official Report. Vol. 524 no. 127 House of Commons Official Report. Vol. 527 no. 153 House of Commons Official Report. Vol. 529 no. 168 House of Commons Official Report. Vol. 523 no. 193

Content

Year

No. of words

international development

2010

9,829

defence

2010

10,089

environment, food and rural affairs

2010

9,660

culture, media and sport

2011

10,697

Total no. of words

80,160 home department

2011

9,779

violence against women and girls

2011

9,671

work and pensions

2011

10,522

communities and local government

2011

9,913

Table 5.5: The conversation corpus Title in the present study Conversation 1 (C1)

General Conversations Official title

Content

KB6

Conversation 2 (C2)

KB6

Conversation 3 (C3)

KCD

Everyday chats between families and friends

Year

No. of words

1991

14,593

1991

25,003

1991

31,145

Total no. of words

70,741

125

6. Methodology Abstract: Chapter 6 describes the quantitative and qualitative method used in the corpusstudy, such as the selection of relevant general noun phrases and the coding system used in the analysis. Corpus examples illustrate the different types of pre- and postmodifier as well as the types of linkage, which occurred in the data.

6.1 Methods for the quantitative analysis 6.1.1 From raw to relevant data The quantitative and qualitative analysis of the present study investigates the general noun phrase heads listed by Halliday/Hasan (1976: 274). These are people, person, man, woman, child, boy, girl, creature, thing, object, stuff, business, affair, matter, move, place, question and idea (cf. Section 2.1.4). As has been discussed in the preceding chapters, the object of investigation for the analysis of general noun phrases in the present study is the general noun phrase head in use. This means that we are not dealing with lemmas or lexemes but with concrete word forms of general noun phrase heads that occur in the corpus data. Thus, the corpus data had to be searched for concrete word forms of the intended general noun phrase heads and other word forms which were irrelevant for the analysis had to be excluded. Therefore, as part of the preliminaries of the quantitative analysis, the forms of general noun phrase heads that were found in the corpus data had to be divided into ‘irrelevant’ and ‘relevant’ forms according to defining criteria that included formal aspects, grammatical aspects such as word-class membership, and semantic aspects such as polysemy, homonoymy and idiomaticity. Since the corpus texts were in pdf format, the Acrobat Reader search function was used for automatic counting of the general noun phrase heads in the corpus texts. However, the tool had limited capacities when dividing irrelevant forms of general noun phrase heads from relevant forms, because this division is among other things based on grammatical and semantic aspects. Since the Acrobat Reader search function is no linguistic tool that is able to take into account such criteria and since the corpus texts were not tagged, supplementary manual investigation was needed. The process of identifying relevant general noun phrase heads was carried out in four major steps: in a first step, the Acrobat Reader search function was used to search for general noun phrase heads, for example move. As a result, the computer displayed sequences of characters. These sequences of characters could be 127

coextensive with the general noun phrase head move but they could also be part of a more complex unit such as a compound1 (move in prime mover), or a word form of the same lexeme (move in the plural noun moves), or a word form of a different lexeme (move in the past tense verb moved or move in mover), or it could also be a completely irrelevant sequence of characters in a word (such as man in manifesto). Consequently, in a second step the settings in the Acrobat Reader search function were adjusted to searching for ‘Ganzes Wort’ which means that the tool only searched for formally limited sequences of characters. For example, when searching for the formally limited sequence move, the Acrobat Reader search function, as a result, displayed a list of forms that belonged to the lemma MOVE. According to Kennedy (cf. 1998: 207), a lemma is the term for a set of inflectionally and derivationally related word forms. Thus, the lemma MOVE is a formal abstraction of the word forms of the noun move (move, moves, move’s and moves’) and the verb move (move, moves, moved and moving). Since a lemma is a purely formal abstraction, it does not take into account grammatical distinctions such as word class membership. However, since we are searching for the general noun phrase head move, in a third step, supplementary manual investigation was needed to distinguish the word forms of the noun move from the word forms of the verb move. The result of the manual investigation was a distinction of the word forms of the lemma MOVE into the word forms of the two lexemes move1 (verb) and move2 (noun).2 All the word forms of the verb move were excluded from the analysis and all the word forms of the noun move were further regarded as relevant. In a fourth step, which was also done manually, the word forms of the lexeme move (noun) were investigated according to the semantic criteria. This is because the remaining word forms of the noun move may also be used with ‘non-general noun phrase head’ meanings, which are in a relation of polysemy or homonomy to the intended ‘general noun phrase head’ meaning. Examples are move1 meaning ‘Umzug’ and move2 meaning ‘(Schach)zug’ or ‘Schritt’, or object1 meaning ‘Ziel’ or ‘Absicht’ and object2 meaning ‘Objekt’ or ‘Ding’.3 For the analysis of general noun 1 2 3

As explained in Section 3.2.1 the phenomenon of compounding is very complex and deserves special attention. Therefore, I have decided to exclude from the analysis cases where the general noun phrase head forms part of a compound. Note that the concept of ‘lexeme’ distinguishes word classes, which is why the verb move constitutes a different lexeme than the noun move. When a single lexeme combines several senses which are more or less related we speak of a polysemous lexeme (cf. Lipka 2002: 92, 154). For example, tree1 in the sense of ‘botanic object’ and tree2 in the sense of ‘geometrical object’ are polysemous insofar that

128

phrase heads this means that sometimes a general noun phrase head is used in a different sense than the one described by Halliday/Hasan (cf. 1976: 274), or that a word is formally identical to a general noun phrase head described by Halliday/ Hasan (cf. 1976: 274) but really is a different lexeme with a completely unrelated meaning. In both cases, we speak of irrelevant general noun phrase heads. An example is the general noun phrase head object, which according to Halliday/ Hasan (cf. 1976: 274) is an inanimate concrete noun. Instances of object used as an abstract noun in the sense of ‘purpose’ were excluded from the analysis. Besides detecting polysemy and homonymy, it was also checked whether the remaining general noun phrase heads were part of a formally and/or syntactically and/or semantically fixed unit, such as a general noun phrase head in a compound (written in two orthographic words), in a proper name, in a heading, in a title, in a figure or in an index, in an idiom or idiomatic expression, or in a set phrase. General noun phrase heads as part of compounds which were written in two orthographic words were excluded from the analysis just like compounds which were written in one orthographic word. Due to the settings in the Acrobat Reader search function (‘Ganzes Wort’), compounds that were written in one orthographic word were automatically excluded but compounds that were written in two orthographic words had to be excluded manually. General noun phrase heads that were part of proper names, as for example in the name of a fund (The Child Fund), were excluded from the analysis. According to Quirk et al. (cf. 1985: 288), proper names may or may not consist of more than one word. More importantly, these names grammatically function as single units which means they cannot be varied by the insertion of words or by change of inflection. With regard to the analysis of modification, therefore, proper names containing general noun phrase heads were excluded from the analysis. Since the qualitative analysis focuses on the reference of general noun phrases, general noun phrase heads within headings, titles, indexes and figures were excluded because they are not integrated into the body text. Therefore, it seemed problematic to describe the reference of those items to preceding and following items of the co-text. tree2 is a metaphorical extension of tree1 which is based on its branching properties (cf. Esser 2006: 52). Similarly, move1 and move2 have related senses. In this context, Cruse (1986: 80) speaks of a lexical unit as the “union of a single sense with a lexical form” and of a lexeme as “a family of lexical units”. Thus, tree1, as well as tree2, is a lexical unit while tree1 and tree2 together constitute the lexeme TREE. In contrast to polysemy, we speak of homonymy when we have several lexemes which are formally identical but have completely unrelated senses (cf. Lipka 2002: 92, 154). This is true for object1 and object2.

129

General noun phrase heads were also excluded from the analysis when they were part of a fixed unit such as an idiom, an idiomatic expression, or a set phrase. This touches upon aspects of idiomaticity, a concept which is controversially discussed among linguists. The present study will use the broad classification by Mahlberg (cf. 2005: 57) who speaks of “patterns of units of meaning with strong dependencies between the surface elements”. This rather broad definition allows me to include in this category different kinds of units of meaning with varying degrees of stable syntactic and semantic structure, varying degrees of lexicalisation and varying degrees of idiomaticity. Therefore, I will define the inventory of the following phenomena as such ‘patterns of units of meaning’ which will be excluded from the analysis: firstly, idioms or idiomatic phrases such as business as usual. Secondly, set phrases such as as things now stand, and thirdly, fixed expressions such as a matter of fact and per person. The following figure summarises how relevant general noun phrase heads are distinguished from irrelevant ones: Figure 6.1: Distinguishing relevant from irrelevant general noun phrases

Step 1

Automatic search for general noun phrase heads sequences of characters

Step p2

Automatic search for th hi word d fforms orthographic lemmata

Step 3

Manual investigation: grammatical distinction lexemes

Step 4

Manual investigation: semantic distinction lexical units

130

As we have seen, the computer cannot distinguish between relevant general noun phrase heads and irrelevant ones. Thus, supplementary manual investigation was needed to identify which general noun phrase heads were taken into account in the analysis. In this context, Mahlberg (2005: 61) states: “[t]he computer can assist the human observer by counting, selecting and displaying words; but the computer cannot yield information about meaning.” In other words, the computer can only count, select and display, it cannot interpret semantic and grammatical information. The linguist must do this.

6.1.2 Determining the frequency of relevant general noun phrase heads For the frequency count of relevant general noun phrase heads, all the word forms of a general noun phrase head were taken into account and searched for individually: for example, when searching for the general noun phrase head move, the word forms move, moves, move’s and moves’ were searched for individually. For the relevant general noun phrase heads, the total number of tokens of each type was counted for each of the four corpus samples. In a first step, this frequency count was done for each text in each of the four corpus samples individually. In a second step, the frequency of relevant general noun phrase heads was determined for each of the four sub-corpora as a whole; this means an overview was given of the total number of tokens of each type of general noun phrase head in the judgment corpus, in the manifesto corpus, in the debate corpus and in the conversation corpus. All the texts of the corpus samples were in pdf-format and the frequency count of the general noun phrase heads in question was done with the Adobe Reader search function. No lexical analysis software was used for the identification of general noun phrase heads and the calculation of their frequency in the texts. The main reason for this was that the span of context that the lexical analysis software provides, in other words the span of words to the left and the right of the general noun phrase head in question, was not large enough for further qualitative analysis. Since anaphoric and cataphoric reference between a general noun phrase and another textual item sometimes overcomes longer stretches of text, it would not have proved useful to work with a list of isolated instances of general noun phrase heads because for the qualitative analysis it was necessary to look at the instances of general noun phrases within their full co-text.

131

6.2 Methods for the qualitative analysis 6.2.1 The coding system used in the present study After the relevant general noun phrase heads had been identified and counted in the quantitative analysis, they were numbered and coded in order to facilitate further analyses. In a first step, the different types of general noun phrase heads in question were numbered from one to eighteen according to the order in Halliday/Hasan (1976: 274). See the following table: Table 6.1: Numbering of general noun phrase heads

1

General noun phrase head people

2

person

Number

3

man

4

woman

5

child

6

boy

7

girl

8

creature

9

object

10

thing

11

stuff

12

business

13

affair

14

matter

15

move

16

place

17

question

18

idea

In a second step, a coding system was developed which marks the individual occurrences (tokens) of each type of general noun phrase head in the particular sub-corpus and allows the reader of this study to retrieve these single occurrences in the attached corpus data without much effort. For this coding system, the single texts in the sub-corpora were given a label and they were numbered. In the judgment corpus the judgments were given the label J and they were numbered 132

so that J1, J2, J3 etc. stands for judgment number one, judgment number two, judgment number three etc. The manifesto corpus consists of three British party manifestos which were named after the political party they were from. The three manifestos were given a label which abbreviates the name of the party and the word manifesto. Accordingly, the label LM stands for Labour manifesto, the label LDM stands for Liberal Democrat manifesto and the label CM stands for Conservative manifesto. The labelling of the debates in the parliamentary debate corpus and the conversations in the conversation corpus corresponds to the labelling of the judgments in the judgment corpus. Thus, D1 marks debate number one and C1 marks conversation number one. See the following table for an overview: Table 6.2: Labelling of corpus texts Corpus

Judgment corpus

Manifesto corpus

Debate corpus

Conversation corpus

Official title of single texts [2010] UKSC 47

Labels J1

[2010] UKSC 50

J2

[2010] UKSC 54

J3

[2010] UKSC 55

J4

[2010] UKSC 57

J5

[2010] UKSC 58

J6

[2011] UKSC 15

J7

[2011] UKSC 22

J8

Labour manifesto

LM

Liberal Democrat manifesto

LDM

Conservative manifesto

CM

House of Commons Official Report. Vol. 510 no. 8

D1

House of Commons Official Report. Vol. 515 no. 8

D2

House of Commons Official Report. Vol. 517 no. 8

D3

House of Commons Official Report. Vol. 521 no. 8

D4

House of Commons Official Report. Vol. 524 no. 8

D5

House of Commons Official Report. Vol. 527 no. 8

D6

House of Commons Official Report. Vol. 529 no. 8

D7

House of Commons Official Report. Vol. 523 no. 8

D8

KB6

C1

KB6

C2

KCD

C3

133

With the labels of the corpus texts and with the numbers of the types of general noun phrase heads (1–18), it was possible to mark the single instances (tokens) of general noun phrase heads in the different corpus-texts. See the following figure for an illustration: Figure 6.2: Example of the coding of a general noun phrase head in the judgment corpus number three

general noun number 17

judgment

fi t occurrence first

J3-17-1 The coding displayed in Figure 6.2 marks the first occurrence of the 17th general noun phrase head (which according to Table 6.1 is the general noun phrase head question) in the third judgment from the judgment corpus. In a next step, the instances of relevant general noun phrase heads (tokens) were marked in the corpus texts so that they were retrievable for the reader and could be displayed within their co-text. The relevant general noun phrase head tokens in the corpus texts were highlighted in the pdf document and a note, which contained the particular coding, was added to each of them. This was done for each relevant general noun phrase head in the corpus: Figure 6.3: Highlighting of general noun phrase heads in the corpus texts

134

Figure 6.4: Coding of general noun phrase heads in the corpus texts

Now, every general noun phrase head could easily be accessed and displayed. Using the “Kommentar” function in the Acrobat Reader, all the highlighted and coded occurrences of general noun phrase heads were displayed in a list and particular occurrences could be searched for by codes: Figure 6.5: Searching for single occurrences of general noun phrase heads within a pdf document

135

6.2.2 The parameter modification As has been explained in Section 4.2, general noun phrases that occur in natural language data are not necessarily general in meaning but can in fact be rather specific. The degree of specification is determined by the correlation of the parameters +/- modification and +/- linkage. Concerning modification, various types of premodification, postmodifiation and also combinations of pre- and postmodification occurred in the corpus data.

6.2.2.1 Types of premodification The types of premodification that occurred within the general noun phrases in the corpus data correspond to the major types of premodifying items introduced by Quirk et al. (cf. 1985: 1322), which were discussed in detail in Section 3.2.2 and are listed below: 1. adjectives, or adjective phrases 2. present and past participles 3. nouns and genitives Additionally, there is another minor type of premodifier which occurred in the corpus data: 4. compound adjectives See the following examples for a demonstration of the above listed types of premodifier: (1) The question whether the planning assumptions can be taken into account too is the crucial question in this case. [J1-17-4] (2) It is not directed to the entirely different question whether the cost to the paying party would be prohibitively expensive, which is what the Aarhus test is concerned with. [J5-17-4]

Example (1) demonstrates the use of the adjective crucial as a premodifier of the general noun phrase head question. In example (2), the general noun phrase head question is premodified by the adjective phrase entirely different. The following two examples illustrate the use of a premodifying present participle (3) and a premodifying past participle (4): (3) Where a literal construction would seriously damage the public interest, and no deserving person would be prejudiced by a strained construction to avoid this, the court will apply such a construction. [J7-2-3]

136

(4) […] the main issue was on section 320 of the Companies Act 1985 (approval by company in general meeting of acquisition of non-cash asset by director or connected person). [J4-2-1]

For the use of a noun and a genitive as premodifier, see examples (5) and (6): (5) In a world of shifting economic power and increased threats, the UK stands to lose a great deal of its ability to shape world affairs unless we act to reverse our declining status. [CM-13-2] (6) However, it is contended that such indication was given by [him] in error and that when providing his answer to the inspector’s question [he] misunderstood what it was that was being asked of him. [J7-17-10]

In example (5), the noun world premodifies the general noun phrase head affairs. Note that all nouns that directly precede a relevant general noun phrase head and cannot be found in either the Oxford Dictionaries Online or the Cambridge Dictionary Online as part of a compound with that relevant general noun phrase head (in which case the general noun phrase head would not have been taken into account for the qualitative analysis) qualify as noun modifiers. In example (6), the genitive inspector’s premodifies the general noun phrase head question. As mentioned above, premodifying adjective compounds were also found in the corpus data: (7) We will rebuild confidence in the criminal justice system so that people know it is on the side of victims and working for law-abiding people, not criminals. [CM-1-96]

In example (7), the adjective compound law-abiding premodifies the general noun phrase head people. Quirk et al. (cf. 1985: 1567) state that compounds in general are units which consists of more than one base and function both grammatically and semantically as a single word. The authors (cf. 1985: 1576–1578) give examples of different types of adjective compounds, for example the type ‘verb and object’ consisting of object + ing participle, as illustrated in example (7) meaning ‘X abides law’. The above presented types of premodifier also occurred in the corpus data in various combinations as multiple premodifier. For an illustration, see the following two examples from the corpus: (8) Work experience and apprenticeships are central to improving the prospects of young unemployed people. [D7-1-2] (9) Creative Bursaries will support the most artistically gifted young people in their early professional careers. [LM-1-112]

In example (8), a combination of the adjective young and the past participle unemployed premodifies the general noun phrase head people. In example (9), 137

a combination of the adjective phrase most artistically gifted and the adjective young premodifies the general noun phrase head people.

6.2.2.2 Types of postmodification The types of postmodification that occurred within a general noun phrase in the corpus data correspond to the major types of postmodification introduced by Quirk et al. (cf. 1985: 1239), which are described in detail in Section 3.2.2: 1. prepositional phrase 2. non-finite clauses (past and present participle clause, infinitive clause, appositive clause) 3. finite clauses (relative clause, appositive that-clause, appositive interrogative clause) In addition, three minor types of postmodifier were found in the corpus data of the present study: 4. appositive noun phrase 5. adjective or adjective phrase 6. adverb or adverb phrase See the following examples for a demonstration of the types of postmodifier found in the corpus data, starting with the prepositional phrase: (10) Millions of people in this country are at best detached from democracy, at worst angry and disillusioned. [CM-1-108]

In example (10), the general noun phrase head people is postmodified by the prepositional phrase in this country. The following four examples demonstrate the use of non-finite postmodifying past- and present participle clauses, infinitive clauses and appositive clauses: (11) […] although a section 83 appeal was a status appeal (i.e. one that depended on the status of the person making the appeal as opposed to the species of decision appealed against) it was nevertheless restricted to a particular class of persons. [J8-2-5] (12) They help to answer some (but not all) of the questions prompted by Mummery LJ’s summary. [J8-17-3] (13) The parties are invited to make submissions in writing within 28 days on the questions to be referred to the Court of Justice. [J5-17-17]

Example (11) demonstrates the use of the non-finite present participle clause making the appeal which postmodifies the general noun phrase head person. Similarly, in example (12), the non-finite past participle clause prompted by 138

Mummery LJ’s summary postmodifies the general noun phrase head question. In example (13), the infinitive clause to be referred to the Court of Justice postmodifies the general noun phrase head questions. (14) Steve Webb: I am grateful to the hon. Gentleman for his question. The idea of measuring pensioner poverty in terms of material deprivation is supplementary to the income measures, […]. [D7-18-3]

Non-finite appositive clauses occurred in the corpus data with present participle clauses as described by Quirk et al. (cf. 1985: 1271). Note that the typical postmodifying function of appositive present participle clauses is as complement of a preposition, such as presented in example (14) and explained by Quirk et al. (cf. 1985: 1272). The following two examples illustrate the use of postmodifiying finite relative clauses: (15) This too is not a question which can sensibly be answered on a day by day basis. [J7-17-7] (16) By the stage when her appeal reached the House of Lords the question which she wished to raise had already been considered twice in the courts below. [J5-17-13]

In example (15), the restrictive relative clause which can be answered on a day by day basis postmodifies the general noun phrase head question. Here, the relative clause gives essential information as it identifies what kind of question is referred to. In example (16), the non-restrictive relative clause which she wished to raise provides additional information about the general noun phrase head question. The information is additional insofar as it is not needed to identify what question means. Besides relative clauses, appositive clauses (appositive that-clauses and appositive interrogative clauses) were used in the corpus data to postmodify a general noun phrase head. Quirk et al. (cf. 1985: 1301) state that appositive clauses involve the linking of units of the same rank. For linguistic units to be appositives they must normally be identical in reference or the reference of one must be included in the reference of the other.4 Another characteristic of appositive clauses is that the appositive units are in a copular relationship. This means that in contrast to a head-modifier construction, the two appositive units can be linked

4

Meyer (cf. 1992: 58–59) argues that this is not a necessary criterion for a construction to be an apposition. He introduces various types of referential and also nonreferential relations between the units of an apposition (cf. Meyer 1992: 59–73) (see Section 6.2.3.1.3 for further information).

139

with be or other copular verbs5 (cf. Quirk et al. 1985: 1261). Furthermore, Quirk et al. (cf. 1985: 1260, 1261) state that with appositive clauses, an antecedent noun is often a general abstract noun such as fact, idea, proposition, reply, remark, answer etc. (see also Schmid (cf. 2000: 3) on this subject)). See the following examples taken from the judgment corpus: (17) It took a purely subjective approach to the question whether a case for such an order had been made. [J5-17-14] (18) This provision must be based on the general idea that the change of use has been there for all to see for four years. [J7-18-1]

In example (17), the appositive interrogative clause whether a case for such an order has been made defines the content of the general noun phrase head question. Similarly, in example (18), the appositive that-clause that the change of use has been there for all to see for four years postmodifies the general noun phrase head idea in that it presents its complete content. In both examples, there is identity of reference between the appositive units, more precisely between the general noun phrase head and the appositive clause. Furthermore, both examples could be reformulated using a copular verb (the question is whether…, the idea is that…) and both examples nicely illustrate that the head of the noun phrase that is modified by an appositive clause is a general abstract noun. The minor types of postmodifier that occurred in the corpus data of the present study were appositive noun phrases, adjectives or adjective phrases and adverbs or adverb phrases. See the following examples: (19) This section applies where on or after the appointed day a person (the transferor) agrees to sell any securities, […]. [J6-2-1] (20) Under Labour, youth unemployment has reached over 900,000, with one in five young people unable to find a job. [CM-1-25] (21) PS01V: just happened to go in one day for some meat […]? That’s the cheapest place down there. [C2-16-1]

In example (19), the appositive noun phrase the transferor postmodifies the general noun phrase head person. Quirk et al. (cf. 1985: 1301) state that apposition primarily and typically is a relation between noun phrases. In example (19), the second noun phrase, the transferor, which is enclosed in parentheses, defines or rather characterises the first noun phrase, a person. In example (20), the adjective phrase unable to find a job postmodifies the general noun phrase head 5

Copular verbs (or linking verbs) are verbs such as be, become and appear that link the subject of a clause and the subject complement or adverbial and occur in SVC and SVA clause types (cf. Quirk et al. 1985: 54, 1172).

140

people. According to Quirk et al. (cf. 1985: 1348), noun phrase modifiers can have their own complementation, as for example an infinitive clause (cf. Quirk et al. 1985: 1226) which can be seen in example (20). In example (21), the general noun phrase head place is postmodified by the adverb phrase down there.

6.2.3 The parameter linkage Besides modification, linkage is the second crucial parameter to determine the degree of specification of general noun phrases. As explained in Section 4.1.2, the presence of the parameter linkage refers to various types of endophoric reference while its absence refers to the fact that general noun phrases show textexternal reference, which is either specific or generic (cf. Quirk et al. 1985: 265, see Section 2.1.2). Since the qualitative analysis in the present study revealed that one of the functions of general noun phrases is generic reference, this type is discussed in more detail in Section 6.2.3.2.

6.2.3.1 Types of endophoric reference 6.2.3.1.1 Halliday/Hasan’s endophoric reference ‘Halliday/Hasan’s endophoric reference’ describes cases of reference that most likely approximate those cases of ‘classic’ endophoric reference described by Halliday/Hasan (1976). Cases of classic endophoric reference show clear anaphoric or cataphoric reference of general noun phrases to preceding or following items, in particular noun phrases and other phrases, either within the same sentence or across sentence boundaries. Thus, cases of classic endophoric reference are straightforward examples of endophoric reference, such as the ministerman-example, provided by Halliday/Hasan (1976: 274–275) and repeated here for convenience: (22) Didn’t everyone make it clear they expected the minister to resign? – They did. But it seems to have made no impression on the man.

Two examples of Halliday/Hasan’s endophoric reference taken from the judgment corpus are provided below: (23) The Prime Minister: I thank the hon. Gentleman for his question […] about ensuring that we protect and help those at the front line […]. […] it is all through the lifetime of those people. [D1-1-23] (24) We will continue to press for stronger international action against terrorism […] and build stability and the rule of law in places that would otherwise shelter terrorist networks. We have shown in Afghanistan and elsewhere that […]. [LM-16-12]

141

In example (23), the general noun phrase those people refers back to the noun phrase those at the front line and establishes a relation of co-reference. In example (24), the general noun phrase places that would otherwise shelter terrorist networks refers to the following noun phrase Afghanistan and the adverb elsewhere.

6.2.3.1.2 Encapsulation ‘Encapsulation’ as a type of reference is a phenomenon that has already been dealt with in Section 2.2 when describing previous studies on and different approaches to general nouns phrase heads and comparable types of nouns. Among those linguists who have focused on the referential behaviour of such nouns are Winter (1977), Ivanič (1991), Schmid (2000), and Flowerdew (2003a and 2003b).6 In accordance with these works, Halliday/Hasan (1976) take into account the function of general noun phrases in discourse under the heading of ‘extended reference’ (see Section 2.1.2 for further explanations). In the case of extended reference, the general noun phrase does not refer to one particular person, object or thing but instead refers to a complex process (cf. Halliday/ Hasan 1976: 52). The common element shared by the nouns in the above cited works is their special potential to refer to pieces of information and to paraphrase and reify this information. For the purpose of the qualitative analysis in the present study, this type of reference will be called encapsulation. Encapsulation can be either anaphoric or cataphoric. Conte (1996), who focuses on anaphoric encapsulation, states that the anaphoric item consists of a ‘general noun’ (or an evaluative noun) as the lexical head and a clear preference for a demonstrative determiner (cf. 1996: 1). The author defines anaphoric encapsulation as follows (1996: 1): […] a cohesive device by which a noun phrase functions as a resumptive paraphrase for a preceding portion of a text. This portion of text […] may be of various length and complexity (a whole paragraph or just one sentence).

Referents of the encapsulating noun phrases are not individuals but entities like state-of-affairs, events, situations, processes, facts, propositions or utterance-acts (cf. Conte 1996: 2). The following example taken from the judgment corpus illustrates anaphoric encapsulation:

6

For a more detailed discussion of each of these works, see Section 2.2 of the present study.

142

(25) I regret to learn that your indebtedness to the Bank […] is not being repaid […] and I have therefore to advise that unless within ten days from the date of this letter you effect repayment of the whole sums due to the Bank […]. I shall have no alternative but to institute proceedings against you for recovery. […] and it is therefore in your own interest to give this matter your immediate attention. [J2-14-1]

In example (25), the general noun phrase this matter refers back to the preceding portion of text, more precisely to the two clauses marked by broken underlining, and resumptively paraphrases them. In addition, the general noun phrase this matter is used to reify what has been said, as on the basis of old information (information in the preceding portion of text) a new discourse referent is created, which has not been mentioned before (the general noun phrase this matter) and which becomes the reference item in the subsequent discourse (cf. Conte 1996: 4). In that way, anaphoric encapsulation often functions as an organising principle in discourse. The new referential expression (this matter in the above example) summarises the preceding discourse portion and simultaneously functions as a starting point for the new one (cf. Conte 1996: 5). In the corpus data of the present study, there were also cases of cataphoric encapsulation. See the following example taken from the judgment corpus: (26) The respondent appealed against the costs officer’s decision under rule 53 of the Supreme Court Rules. They asked the single Justice to refer to the following questions [J5-17-1] to a panel of Justices under rule 53(2): (1) whether it was open to the costs officers, […], to achieve that result through the detailed assessment process; and (2) if it was, whether the test indicated by the phrase “prohibitively expensive” should be focused exclusively on the actual circumstances of the parties to the litigation and not on the question what would be prohibitively expensive for the ordinary member of the public. The single Justice referred the application to a panel of five Justices and directed that these questions [J5-17-3] should be decided after an oral hearing.

Example (26) illustrates an interesting case of reference through encapsulation which is cataphoric as well as anaphoric. The first general noun phrase the following questions [J5-17-1] functions as an advance organiser and cataphorically encapsulates the clauses under (1) and (2). The second general noun phrase these questions [J5-17-3] anaphorically encapsulates these clauses and thus also functions as a discourse organiser because it resumptively paraphrases what has been explained in the preceding text. Interestingly, the two general noun phrases function as a sort of structural frame for the complex information given under (1) and (2) and therefore guide the reader: while the first general noun phrase functions as a signpost of what is about to come, the 143

second general noun phrase functions as a summary and reifies what has just been explained.

6.2.3.1.3 Reference between appositive units In the type of reference called ‘reference between appositive units’, the general noun phrase anaphorically or cataphorically refers to a preceding or following item which is in apposition to the general noun phrase.7 Meyer (cf. 1992: xiii) in his work Apposition in Contemporary English defines apposition as a grammatical relation between two units realised by constructions having particular syntactic and semantic characteristics. We have already seen in Section 6.2.2.2 that appositive units can have a variety of syntactic forms, such as the form of noun phrases, ing-clauses, that-clauses and interrogative clauses. As far as the semantic structure of apposition is concerned, Meyer (cf. 1992: 57–73) states that there are different kinds of referential and also non-referential relations between the two units in an apposition. In the focus of the analysis of general noun phrases in the present study, however, are only referential relations between appositive units. The type of reference between appositive units that typically occurred in the corpus data of the present study was cataphoric reference of the first appositive unit to the second appositive unit. According to Meyer (cf. 1992: 64), this type occurs in any apposition in which the first unit is a noun phrase and the second unit is a clause or sentence. See the following examples from the corpus: (27) It took a purely subjective approach to the question whether a case for such an order had been made. [J5-17-4]

In example (27), the first unit of the apposition (the general noun phrase the question) refers cataphorically to the second unit of the apposition (the appositive interrogative clause whether a case for such an order had been made). The second appositive unit presents the complete content of the first appositive unit. A further example, which is less straightforward, is provided below: (28) Building the Big Society means encouraging the concept of public-spirited service – the idea that everyone should play a part in making their communities stronger. [CM-18-6]

7

Note that Meyer (cf. 1992: xiii) speaks of ‘appositive units’ while Quirk et al. (1985: 1301) speaks of ‘appositives’. Concerning the relation of apposition, both authors (cf. Meyer 1992: xiii, Quirk et al. 1985: 1301) use the expression ‘to be in apposition’ in order to explain that two appositive units, or two appositives, are related through apposition. The present section will stick to Meyer’s terminology.

144

In example (28), the noun phrase the concept of public-spirited service is the first appositive unit and the general noun phrase the idea that everyone should play a part in making their communities stronger is the second appositive unit. There is a second apposition embedded in the general noun phrase; here the idea is the first unit and the appositive that-clause is the second unit. In the embedded apposition, the first appositive unit, the idea, cataphorically refers to the second appositive unit, the appositive that-clause. The reference relation in the superior apposition is less straightforward. Since both appositive units are definite noun phrases we can argue that both appositive units are referring. Note that Meyer (1992) does not discuss appositions in which both appositive units are definite noun phrases, however, I suggest that the first appositive unit, the noun phrase the concept of public-spirited service, cataphorically refers to the general noun phrase because it seems like a subordinate unit which is more specific than the first appositive unit. It can, however, also be argued that the second appositive (the general noun phrase the idea that…) unit refers anaphorically to the first appositive unit (the noun phrase the concept of…).

6.2.3.1.4 Reference between subject and complement We can classify a linguistic unit either on the basis of its form (e.g. as a noun phrase) or on the basis of its function (e.g. as subject of the clause) (cf. Quirk et al. 1985: 48). Besides modification and apposition, which are two clause functions which can be realised through different linguistic forms, there is a third clause function, that of complement (cf. Quirk et al. 728–729). For the purpose of the analysis in the present study, the present section will focus on subject complements. Subject complements apply some attribute or definition to the subject and thus semantically identify or characterise the referent of the subject. They are in a copular relationship with the subject and occur in the clause pattern SVC (cf. Quirk et al. 1985: 728). Subject complements normally have the form of noun phrases or adjective phrases such as in John is a boy or John seems restless but subject complements may also be realised by nominal clauses such as that-clauses and interrogative clauses (cf. Quirk et al. 1985: 728, 1049, 1051, 1054). The use of nominal clauses as subject complements in SVC clause patterns using a copular verb in either subject or complement position was quite frequent in the corpus data. Concerning the type of reference in these instances, there typically was a referential relation between the subject of a clause and its subject complement with a general noun phrase being either in the subject or in the subject complement position.

145

In the case where the general noun phrase is in the subject position and refers to its subject complement we can speak of cataphoric reference of the general noun phrase subject to its subject complement. In the case where the general noun phrase is in the subject complement position we can speak of anaphoric reference of the general noun phrase subject complement to the subject of the clause. To illustrate the former, see the following examples taken from the judgment corpus: (29) The question that these appeals give rise to is whether a creditor who wishes to enforce the security to obtain performance of the debt for which security was given can choose whether to proceed by way of a calling-up notice or may proceed instead on the basis that the debtor is in default under standard condition 9(1)(b). (S NP + V + Cs interrogative clause) [J2-17-3] (30) The next matter to note is that the 1992 Act was a consolidating statute. (S NP + V + Cs that-clause) [J3-14-1]

Example (29) provides a case of cataphoric reference of a general noun phrase in subject position to its subject complement. The general noun phrase the question that these appeals give rise to is the subject of the clause (S NP) and the underlined interrogative clause is the respective subject complement (Cs interrogative clause) which provides the definition of the subject and identifies what the question that these appeals give rise to refers to. Subject and subject complement in example (29) are linked with the copular verb be. Similarly, in example (30), the subject of the clause which is realised by the general noun phrase the next matter to note (S NP) is complemented by the subject complement of the clause which is realised by the that-clause that the 1992 Act was a consolidating statute (Cs that-clause). Both units are linked by the copular verb be. The subject refers cataphorically to its subject complement because the complementing that-clause identifies the referent of the general noun phrase in subject position. For a demonstration of anaphoric reference of a general noun phrase in subject complement position to its subject see the following examples from the judgment corpus: (31) The question1 whether the planning assumptions can be taken into account too is the crucial question2 in this case. (S NP + V + Cs NP) [J1-17-3] (32) Whether this conduct (and that of his father-in-law with whom he secretly constructed the house) was or was not susceptible to prosecution under the general criminal law cannot be the determining question here. (S interrogative clause + V + Cs NP) [J7-17-16]

In example (31), the general noun phrase the question1 whether the planning assumptions can be taken into account too is the subject of the clause (S NP). The 146

general noun phrase the crucial question2 in this case is the subject complement of the clause (Cs NP). Both units are linked by the copular verb be. While the subject complement is merely a characterising attribute for the subject, we can say that, vice versa, the subject defines and identifies the referent of the subject complement. Thus, the crucial question in this case anaphorically refers to the question whether the planning assumptions can be taken into account too. In example (32), the subject of the clause is the interrogative clause beginning with whether (broken underlining) (S interrogative clause) and the subject complement of the clause is the general noun phrase the determining question (Cs NP). Both units are linked with the copular verb be. Similar to the reference relation in example (31), the subject complement in example (32) refers anaphorically to its subject, which provides a definition for it and thus identifies the referent of the general noun phrase the determining question.

6.2.3.1.5 Remote reference As described under Sections 3.3.1 and 3.4.2.2 the type of reference called ‘remote reference’ describes anaphoric or cataphoric reference between a general noun phrase and another item in the text which is remote. This means that a larger passage of text lies in between the two items, which makes it difficult if not impossible for the decoder to identify the reference relation between the general noun phrase and the reference item. See the following example taken from the judgment corpus to illustrate a case of remote reference: (33) The question whether the planning assumptions can be taken into account too is the crucial question in this case. [8 pages of text in between] I do not think that the parties can have given express consideration to the question that has arisen in this case. [J1-17-5]

Example (33) shows a case of remote reference. The general noun phrase the question that has arisen in this case remotely refers to the reference item, the general noun phrase the question whether the planning assumptions can be taken into account too, which goes back 8 pages in the judgment. Since the reference is remote, the memory of the decoder might be overwhelmed and they might not be able to identify the reference item. However, it can also be argued that the decoder is nevertheless able to identify the reference relation despite its remoteness. Very often, remote reference is used, for example in judgments, for economic reasons. For example, the general noun phrase the matter remotely refers to the overall concern or the facts of the case which have been explained at the beginning of a judgment. See the following example to illustrate this: 147

(34) The two sides have now been arguing for over six years about DCC’s tax return for the relevant period […]. [5 pages of text in between] That figure of £1.8m is the only one agreed by both sides (and by the Special Commissioner and all the judges who have so far considered the matter) as an element in the tax computations. [J6-14-1]

Since the facts of the case include the essential information for the judgment in question, it can be argued that the decoder is able to memorise these facts and to identify that the matter refers to these facts even if there is a longer passage of text in between the remotely referring general noun phrase and the reference item, the explanations of the facts of the case. In this context, Löbner (cf. 1998: 13–14) introduces the notion of ‘globally defined functional concept’. This means that a particular referent (for example the referent of the matter in example (34)) is only defined once for the given situation of utterance. This means that at the very beginning of the judgment, it was defined what the matter refers to, thus, what the facts of the case are. Löbner (cf. 1998: 14) explains that since the situation of utterance is one and the same throughout the entire text (here: the judgment) the general noun phrase the matter (used to referring to the facts of the case) works as a ‘globally defined functional concept’. Löbner (1998: 14) states that “[e]ach time the expression occurs […], its anchoring is reactivated”. In this respect, the globally defined noun phrase differs from other referring items whose referents are determined by only locally defined functional concepts (cf. 1998: 14). Section 3.3.1 has shown that remote reference can be used by the encoder for different reasons, for example to be deliberately unspecific about something. This might be intended by the encoder, for example, to avoid being unpolite and offending (see Section 3.3.1 for a more detailed discussion). Additionally, the present section has shown that remote reference occurs when a general noun phrase, once globally defined for a situation of utterance, refers to a preceding item in the text. Very often, this is done for economic reasons.

6.2.3.2 Generic reference According to Quirk et al. (cf. 1985: 265), generic reference is reference to a class of entities without specific reference to a particular member of that class. This reference occurs with singular and plural noun phrases as well as with definite and indefinite noun phrases. A general noun phrase with the indefinite article, for example, can generically refer to a representative member of a class (cf. Quirk et al. 1985: 281). See the following example from the judgment corpus: 148

(35) It was well understood that a person was in principle entitled to recover at common law money paid under a mistake of fact. [J3-2-7]

Generic general noun phrases with zero article can have both plural nouns and non-count nouns as head. These general noun phrases consider a class of entities as an undifferentiated whole (cf. Quirk et al. 1985: 282). See the following example from the judgment corpus: (36) …‘subsidiary protection’ which Council Directive 2004/83/EC of 29 April 2004 (the Qualification Directive) requires to be given to certain third country nationals or stateless persons. [J8-2-1]

General noun phrases with the definite article can only express generic reference with singular heads indicating the class as represented by its typical specimen (cf. Quirk et al. 1985: 282): (37) Where any such determination […] is made, the person making the determination shall […] (b) specify the period during which that amount was paid to the person concerned. [J3-2-2-], [J3-2-3]

According to Quirk et al. (cf. 1985: 283) the definite article together with plural nouns other than nationality names (the Chinese) or phrases with an adjective head referring to a group of people (the blind) cannot be used for generic reference. However, there were cases of generic reference expressed by a combination of the definite article and a plural noun in the corpus data. See the following example taken from the judgment corpus: (38) It is worth noting that the deed contained a declaration in terms of which the expression, ‘the Obligant’, was to mean both the persons who granted the security […]. [J2-2-2]

Here, the general noun phrase the persons who granted the security is generic because it refers to all persons who grant securities hypothetically and therefore generally refers to a group of persons. The form ‘all + noun’ with no article usually has generic reference (cf. Quirk et al. 1985: 259). Similarly, the form ‘no + noun’ with no article usually has generic reference (cf. Quirk et al. 1985: 392). See the following examples from the corpus: (39) All young people should be active and engaged in their communities, […]. [LM1-124] (40) No child should have to grow up in poverty. [LM-5-51]

There were instances of generic reference in the corpus data expressed by the first person plural and a general noun phrase head. See the following example: 149

(41) The Tories want to gamble with our children’s education. [LM-5-7]

Quirk et al. (cf. 1985: 350) state that ‘we’ and ‘our’ can be used in the collective sense of ‘we, the nation’ or ‘we, the party’ expressing generic reference. This is demonstrated in example (41), where our children generically refers to ‘Britain’s children’.

150

7. Corpus analysis Abstract: Chapter 7 presents the results of a detailed quantitative and qualitative analysis of general noun phrases across the four sub-corpora. As an outcome, a number of patterns of general noun phrase use are suggested, which show that these items are much more flexible in form and function than described by Halliday/Hasan (1976).

The present chapter presents the results of the corpus analysis of general noun phrases. For the analysis, corpus occurrences of the general noun phrase heads introduced by Halliday/Hasan (1976) were investigated in the four different subcorpora and will be discussed on the basis of the general assumptions concerning the use of general noun phrases in different media and genres presented earlier (cf. Section 4.4). The analysis in the present study is based on quantitative and qualitative methods. After the raw data had been turned into relevant data, the frequency of relevant general noun phrase heads was determined and their distribution was investigated in and across the four sub-corpora. The results of this quantitative analysis are presented in Section 7.1. Then, the degree of specification of relevant general noun phrases, which is defined by the correlation of the parameters modification and linkage, was determined in each of the four sub-corpora. The results of this qualitative analysis are presented in Section 7.2.

7.1 Quantitative analysis: Frequencies and distribution of relevant general noun phrase heads in the corpora 7.1.1 Quantitative results from the judgment corpus The results of the frequency count of general noun phrase heads in the judgment corpus are presented in Table 7.1 below. The table indicates the raw, thus nonnormalised as well as the normalised frequency of each general noun phrase head in the judgment corpus.1

1

Note that the four sub-corpora differ in length. Thus, in order to compare the number of general noun phrase heads in these corpora, the results were normalised for a subcorpus with a length of 100,000 words. Note that the discussion in Section 7.1 will refer only to the normalised frequencies of general noun phrase heads.

151

Table 7.1: Frequencies of general noun phrase heads in the judgment corpus

people

count non-normalised 3

Tokens counted manually non-normalised 3

counted automatically normalised 3.82

person

54

52

66.25

man

7

4

5.1

woman

3

3

3.82

child

13

1

1.27

boy

0

0

0

girl

0

0

0

creature

2

0

0

thing

9

0

0

object

7

0

0

Type

stuff

0

0

0

business

16

0

0

affair

4

1

1.27

matter

62

36

45.86

move

10

0

0

place

15

3

3.82

question

119

78

99.37

idea

3

1

1.27

sum

327

182

231.85

In Table 7.1, the left column presents the 18 general noun phrase heads that were investigated in the analysis (types). The second column from the left presents the non-normalised frequencies of all the general noun phrase heads in the judgment corpus, both relevant and irrelevant (tokens). The third column from the left presents the final results of frequency count of the 18 general noun phrase heads after supplementary manual investigation and therefore presents only the non-normalised frequencies of relevant general noun phrase heads in the judgment corpus. In addition, the rightmost column displays the normalised frequencies of these relevant general noun phrase heads. The normalised overall frequency of all 18 general noun phrase heads in the judgment corpus is 231.85. This frequency is made up of a rather uneven distribution of general noun phrase head types: we can see that three general noun 152

phrase head types stand out in terms of their frequency, namely person, matter and question. In comparison to the 15 remaining general noun phrase head types which occur considerably less often or do not occur at all, person, matter and question show a relatively high frequency within the judgment corpus: person occurs 66.25 times, matter occurs 45.86 times and question occurs 99.37 times. Together, these three general noun phrase heads occur 211.48 times, which already accounts for almost 90% of the overall frequency of general noun phrase head tokens in the judgment corpus. The remaining 7 general noun phrase head types occur only between 1.27 times and a maximum of 5.1 times in the judgment corpus. These are people, man, woman, child, affair, place and idea. Eight general noun phrase heads do not occur at all, namely boy, girl, creature, thing, object, stuff, business and move. The distribution of general noun phrase head types and tokens in the judgment corpus is shown in the following figure:

i idea

place

m move

a affair

b business

sttuff

o object

th hing

g girl

crreature

b boy

woman

c child

man

matter

person n people

115 110 105 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0

question n

Figure 7.1: General noun phrase head types and tokens in the judgment corpus

The x-axis of Figure 7.1 displays the 18 general noun phrase heads under investigation and the y-axis represents their normalised frequencies in the judgment corpus. Mahlberg (cf. 2005: 164) states that frequent nouns characterise a group of text. These nouns have local textual functions that refer to properties of texts, 153

e.g. texts that deal with certain topics. In reference to the results of the quantitative analysis of general noun phrase heads in the judgment corpus presented in Table 7.1 and Figure 7.1, we can similarly assume that person, matter and question are frequent general noun phrase heads (in comparison to the remaining general noun phrase heads investigated). This implies that they characterise the Supreme Court judgments as a group of text.2 For the general noun phrase head person, this holds true insofar as person is a word typically used in legal English. According to the Oxford Dictionaries online, person in legal or formal contexts refers to an unspecified individual (http://www.oxforddictionaries.com/definition/english/person, last access 30th September 2014). See the following example, which illustrates the typical use of the general noun phrase head person in the judgment corpus: (1) Where any such determination as is referred to in subsection (1) above is made, the person making the determination shall in the case of the Secretary of State or the First-tier Tribunal, and may in the case of the Upper Tribunal or a court – […]. [J3-2-2]

Besides the typical use of the general noun phrase head person in Supreme Court judgments, the general noun phrase heads matter and question also typically occurred in the Supreme Court judgments investigated. The following three examples demonstrate the use of the general noun phrase heads question and matter in Supreme Court judgments: (2) All this is expressly provided for by section 71 of the Social Security Administration Act 1992 (the 1992 Act). But does section 71 provide an exclusive code for recovery? That is the question. [J3-17-2] (3) The respondents appealed against the costs officers’ decision under rule 53 of the Supreme Court Rules. They asked the single Justice to refer the following questions to a panel of Justices under rule 53(2): […] The single Justice referred the application to a panel of five Justices and directed that these questions should be decided after an oral hearing. [J5-17-1], [J5-17-3]

2

What Mahlberg (cf. 2005: 164) refers to as frequent nouns are 20 ‘general nouns’ that are among the most frequent nouns in the BNC and the Bank of English (rank 1–91). Note that these ‘general nouns’ do not correspond to the general noun phrase heads listed by Halliday/Hasan (1976) and investigated in the present study. Nevertheless, it can be argued that Mahlberg’s conception of frequent nouns with local textual functions can be transferred to the general noun phrase heads investigated in the present study. Therefore, it can be assumed that the most frequent general noun phrase heads in each sub-corpus have local textual functions that characterise a group of corpus texts.

154

(4) The council denied that the building constructed was a dwelling house, maintained that a ten year period for enforcement applied under section 171B(3) and on 30  August 2007 refused a certificate. Mr Beesley appealed and the matter came before Mr K L Williams, a planning inspector appointed by the second respondent, the Secretary of State. [J7-14-4]

Example (2) shows that the general noun phrase head question is used in Supreme Court judgments to express that facts of the case are questioned and revised. Apart from that, questions that are raised are typically referred to throughout the hearings, which can be illustrated in example (3). Here, the general noun phrase the following questions is used cataphorically, while the general noun phrase these questions is used anaphorically. Both general noun phrases thus have a textorganising function. Example (4) reveals that the general noun phrase head matter is typically used in Supreme Court judgments when referring to the facts of the case, the disputable aspects or the legal dispute as a whole. The general noun phrase head matter can then be used to refer to and encapsulate these complex points for the purpose of reduction.

7.1.2 Quantitative results from the manifesto corpus The results of the quantitative analysis of general noun phrases in the manifesto corpus are presented in Table 7.2 below: Table 7.2: Frequencies of general noun phrase heads in the manifesto corpus count non-normalised 467

Tokens counted manually non-normalised 418

person

9

5

6.49

man

12

8

10.38

Type people

counted automatically normalised 542.56

woman

21

15

19.47

child

219

154

199.89

boy

0

0

0

girl

2

2

2.60

creature

0

0

0

thing

26

1

1.30

object

0

0

0

stuff

0

0

0

business

144

0

0

155

affair

count non-normalised 7

Tokens counted manually non-normalised 3

counted automatically normalised 3.89

matter

12

1

1.30

move

21

2

2.60 22.10

Type

place

77

17

question

6

3

3.89

idea

12

10

12.98

sum

1,035

639

829.42

The normalised overall frequency of general noun phrase heads in the manifesto corpus is considerably high with 829.42 occurrences of all 18 general noun phrase heads. Similar to the results of the judgment corpus, there are two general noun phrase head types in the manifesto corpus that stand out concerning their frequency; these are people and child. People occurs 542.56 times in the manifesto corpus and child occurs 199.89 times. These frequencies show that the two general noun phrase heads people and child together account for almost 90% of the overall frequency of general noun phrase head tokens in the manifesto corpus. Following the very frequent general noun phrase head types people and child, the general noun phrase head types woman and place occur 19.47 times and 22.10 times. The remaining general noun phrase head types occur 1.30 times (thing) up to 12.98 times (idea). Five general noun phrase head types, namely boy, creature, object, stuff, business, did not occur at all in the manifesto corpus. Interestingly, these five general noun phrase heads are the very general noun phrase heads that did not occur in the judgment corpus. See the following figure, which illustrates the distribution of general noun phrase head types and tokens in the manifesto corpus:

156

idea

place

qquestion

m move

aaffair

m matter

bbusiness

sttuff

oobject

thhing

ggirl

crreature

bboy

child woman

man

pperson

600 80 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0

peop ple

Figure 7.2: General noun phrase head types and tokens in the manifesto corpus

Figure 7.2 shows that in the manifesto corpus, the general noun phrase heads people and child considerably outnumber the remaining 16 general noun phrase heads regarding their frequency. In the context of Mahlberg’s discussion of textual functions of frequent nouns (cf. 2005: 164), this suggests that people and child among other nouns textually characterise the political manifestos investigated. Considering the content and purpose of political manifestos, this is certainly true: a manifesto describes a party’s political concerns and goals for the upcoming electoral period and is thus aimed at possible future voters. In the case of the British party manifestos that were investigated in the present study, a highly frequent use of the general noun phrase head people is thus comprehensible: the manifestos are aimed at British people and address all those political issues that affect these people. See the following example: (5) Under a Liberal Democrat government, you will not have to pay any income tax on the first £10,000 you earn. This will put £700 back into the pockets of millions of people on low and middle incomes and free 3.6 million more people on low incomes from having to pay any income tax at all. [LDM-1-11], [LDM-1-12]

In example (5), the general noun phrases millions of people on low and middle incomes and 3.6 million more people on low incomes are used to refer to a particular part of the British people, namely all British people on low and middle incomes. 157

The Liberal Democrats describe a financial advantage that they want to provide for this rather broad class of the British population. The frequent use of the general noun phrase head people in the manifesto corpus can also be said to express that the politicians very often address the people of Britain (the possible electorate) in order to emphasise the integration of the voters into the party’s political concerns. See the following example: (6) We will stamp out corruption and abuse by giving people power to sack corrupt MPs, end big money politics, and make sure those who seek to sit in Parliament pay full UK taxes. [LDM-1-7]

In example (6), the general noun phrase head people is used to refer to all the British people, or at least all eligible voters. The Liberal Democrats emphasise that all those people will be integrated into political concerns and that they will gain certain powers. Similar to the frequent use of the general noun phrase head people demonstrated in examples (5) and (6), the general noun phrase head child is also used frequently and thus indicates that topics such as family life, childcare and education are dealt with in the manifestos. See the following example for a demonstration: (7) Don’t settle for low politics and broken promises: be more demanding. Set your sights on the Britain you want for your children and your grandchildren, and use your vote to make it happen. [LDM-5-2]

Here, possible voters are addressed directly with the expression the Britain you want for your children. The possible voter is asked to change the political situation so that the future of their (future) children is safe. The whole topic ‘children’ is an emotional issue for many people, which is why manifestos deal with this topic. This explains a high frequency of the general noun phrase head child. In the context of textual functions of general noun phrases, another aspect is very interesting: the two general noun phrase heads woman (19.11 times) and girl (2.55 times) are more frequent than man (10.19 times) and boy (does not occur) in the manifesto corpus. This might indicate that the rights of women and the equality of men and women are important topics in political discourse. The following example supports this assumption: (8) Introducing fair pay audits for every company with over 100 employees to combat discrimination in pay, for example against women. [LDM-4-2]

158

7.1.3 Quantitative results from the debate corpus The results of the quantitative analysis of general noun phrase heads in the debate corpus are presented in Table 7.3 below: Table 7.3: Frequencies of general noun phrase heads in the debate corpus Type people

count non-normalised 233

Tokens counted manually non-normalised 219

counted automatically normalised 273.31

person

7

6

7.49

man

12

10

12.48

woman

45

36

44.93

child

73

49

61.15

boy

0

0

0 14.98

girl

13

12

creature

0

0

0

thing

41

2

2.50

object

3

0

0

stuff

0

0

0

business

99

1

1.25

affair

37

0

0

matter

74

64

79.89

move

17

3

3.74

place

83

16

19.97

question

111

20

24.96

idea

17

17

21.22

sum

865

455

567.84

As we can see in Table 7.3, there are 567.84 normalised occurrences of general noun phrase heads in the debate corpus. Similar to the manifesto corpus, the general noun phrase head people outnumbers the remaining general noun phrase heads in the corpus concerning its frequency. People is the most frequent general noun phrase head in the debate corpus; it occurs 273.31 times. Second most frequent is a group of general noun phrase heads including matter with 79.87 occurrences, followed by child with 61.15 occurrences and woman with 44.93 occurrences. Third most frequent is a group of general noun phrase heads 159

including question with 24.96 occurrences, idea with 21.22 occurrences, place with 19.97 occurrences, girl with 14.98 occurrences, man with 12.48 occurrences, and person with 7.49 occurrences. The remaining general noun phrase head types, including move, thing, and business occur 1.27 up to 3.74 times. Five general noun phrase head types do not occur altogether, namely boy, creature, object, stuff and affair. These general noun phrase heads correspond to those that did not occur in the judgment corpus and in the manifesto corpus. For an overview of the distribution of general noun phrase head types and tokens in the debate corpus see the following figure: Figure 7.3: General noun phrase head types and tokens in the debate corpus 300 280

people

320

260 240 220 200 180 160 140

idea

questionn

place

affair

b business

sttuff

o object

t thing

girl

crreature

0

b boy

20

man

40

person

60

child

wooman

80

m move

matterr

120 100

The distribution of general noun phrase head types and tokens in the debate corpus, which is presented in Figure 7.3, shows clearly that the general noun phrase head people is the most frequent type with a considerable distance to the remaining general noun phrase heads. Figure 7.3 also shows that of the remaining general noun phrase heads, woman, child and matter occur quite frequently with 44.93 up to 79.87 tokens. Similar to what has been discussed concerning the general noun phrase head people in the manifesto corpus, it can be stated that people is also typically used in the parliamentary debates investigated and therefore has a characterising function for this group of texts. Considering the content and purpose of 160

parliamentary debates, people is often used because parliamentary debates discuss topics that concern and affect the British people. See the following example: (9)

Richard Benyon: […] I do not want to burden people with regulation – that is not the direction that the Government are going in – and I am happy to meet my hon. Friend and the MMO to see whether we can find another way forward, […]. [D3-1-1]

The following example demonstrates the second most frequent general noun phrase head matter in the debate corpus: (10) Nic Dakin (Scunthorpe) (Lab): The European Safety Authority has concluded that the major factor causing poor welfare in dairy cows is genetic selection to produce high yields. Given proposals to intensify milk production for higher yields, […], will the Secretary of State agree urgently to review the welfare code for dairy cows in the UK, and to meet a delegation of cross-party MPs and non-governmental organisations to discuss how her Department can ensure that its code takes into account the latest scientific advice and ensures that any new dairies do not compromise cow welfare? Mr Paice: I am happy to meet the hon. Gentleman and others to discuss the matter, but I assure him that the Department puts welfare at the top of our agenda, […]. [D3-14-5]

Example (10) illustrates the use of the general noun phrase the matter of one member of parliament to refer back to a preceding statement of another member of parliament (here the passage that is marked by broken underlining). In example (10), as in many other examples in the debate corpus, the general noun phrase head matter is used throughout a discussion to refer back to political issues that are on the agenda of a debate. This way, the general noun phrase can be used to encapsulate complex facts and reify them.3

7.1.4 Conversation corpus The results of the quantitative analysis of general noun phrase heads in the conversation corpus are presented in Table 7.4 below:

3

See Halliday/Hasan (cf. 1976: 52) for ‘extended reference’ (Section 2.1.2); see also Section 6.2.3.1.2 for the concept of ‘encapsulation’ through the use of general noun phrases.

161

Table 7.4: Frequencies of general noun phrase heads in the conversation corpus count non-normalised 35

Tokens counted manually non-normalised 26

counted automatically normalised 36.76

person

4

3

4.24

man

11

1

1.41

Type people

woman

7

7

9.90

child

24

14

19.80

boy

18

10

14.14

girl

41

23

32.52

creature

1

1

1.41

thing

140

58

82.01

object

0

0

0 62.22

stuff

56

44

business

6

0

0

affair

4

3

4.24

matter

22

0

0

move

12

0

0

place

31

18

25.45

question

3

2

2.83

idea

6

2

2.83

sum

421

212

299.77

As we can see in Table 7.4, there are 299.77 normalised instances of general noun phrase heads in the conversation corpus. The two most frequent general noun phrase head types are thing and stuff with a normalised frequency of 82.01 and 62.22. Second most frequent is a group of general noun phrase head types including people with 36.76 occurrences, girl with 32.52 occurrences and place with 25.45 occurrences. Third most frequent is a group of general noun phrase head types including woman with 9.90 occurrences, child with 19.80 occurrences, and boy with 14.14 occurrences. The remaining general noun phrase head types including person, man, creature, affair, question, and idea occur 1.41 up to 4.24 times. Four general noun phrase head types do not occur altogether, these are object, business, matter, and move. They partly correspond to those general noun phrase 162

head types that did not occur in the judgment corpus, the manifesto corpus and the debate corpus. Concerning the distribution of general noun phrase head types and tokens in the conversation corpus, see the following figure:

t thing

Figure 7.4: General noun phrase head types and tokens in the conversation corpus 85 80 75 stuff

70 65 60 55

girl

plaace

35 30

idea

m move

affair

m matter

b business

0

o object

5

man

10

person

15

creature

man wom

20

boy b

child

25

question

45 40

peop ple

50

Figure 7.4 shows that although the overall number of general noun phrase head tokens in the conversation corpus is rather low, the conversation corpus has the broadest distribution of general noun phrase head types; in other words, more different general noun phrase head types occur in the conversation corpus and they occur quite frequently (with more than five tokens). Note that this becomes clearer when comparing the frequencies and distribution of general noun phrase heads across the four sub-corpora in Section 7.1.5. Two general noun phrase heads, namely thing and stuff, are used more frequently than the remaining general noun phrase heads. This can be explained with the purpose and appropriateness of language use in conversations: in the context of the restricted style of referencing,4 which is typical of conversations, it was mentioned that this style of referencing is characterised by a frequent use of pronouns and deictic expressions and an infrequent use of complex noun phrases (cf. Esser 2009: 36). Partington (cf. 1998: 91) states that ‘general nouns’ 4

For a detailed discussion, see Section 3.3.3.

163

are part of the system of deixis and function as pro-forms which corresponds to Halliday/Hasan’s (cf. 1976: 274) statement that ‘general nouns’ are items on the borderline between the open and closed class system. When general noun phrase heads like thing and stuff, which are very general in meaning, are used to refer to objects that are talked about in conversations, referencing heavily relies on the shared situation. The objects that are talked about are mutually accessible to encoders and decoders and therefore it is sufficient to refer to those objects even with such a general term as thing, because the encoder assumes that the decoder sees and understands what is meant. In contrast, objects that are talked about in written discourse have to be properly introduced with lexically filled noun phrases (cf. Esser 2009: 36–37). Thus, very general noun phrases like thing and stuff would be inappropriate in such a context. The frequent use of the general noun phrase heads thing and stuff in conversations can also be explained with the theory of conversational implicature formulated by Grice (1975).5 Conversation is a cooperative venture and speakers are assumed to adhere to four maxims of conversation. According to these maxims, speakers try to send truthful, informative, relevant and unambiguous messages and hearers assume them to do so (cf. Grice 1975: 45–46). When a speaker is faced with the dilemma of being unable to fulfil a certain maxim without disregarding another, then vague expressions (e.g. general noun phrase heads such as thing and stuff) can help them to be, for example, as truthful as possible in a situation where they are not able to give an exact reply (for example because they do not know better) (cf. Channell 1994: 33). Furthermore, vague expressions enable a speaker to imply in their utterance things that they need not verbalise because the listener is assumed to be cooperative and willing to follow what the speaker says and means (cf. Grice 1975: 44). This aspect again follows up the statement that an elaborated style of referencing is not necessary and not appropriate in face-to-face situations. The following example demonstrates the use of the general noun phrase head stuff, which enables the speaker to follow the Gricean maxims of conversation: (11) PS02A: PS029: PS02A: PS029: PS02A:

5

What colour are you having the worktop? Well I thought cos I like that marble flooring. If they haven’t got any green worktops cos Ange said when she went up there You’d best price that as well. What? Your flooring.

For a detailed explanation, see Section 3.4.1.

164

PS029:

[C1-11-4]

Yeah. They didn’t have any worktops, the green Ange said. She can’t remember any but there again you were she weren’t looking for green. But they’ve got that marble. You know what we had? Over Jan’s, pity we didn’t nick that weren’t it? They got that sort of marble stuff. The same as the floor so I might pick the green marble. To go with the floor.

The extract from the conversation in example (11) shows that speaker PS029 uses the general noun phrase that sort of marble stuff to refer to all the objects that are made of marble or that are of marble colour. Speaker PS029 probably does not know the exact term for the material, which is why they use the general noun phrase head stuff, which is a very broad and general description. That way, speaker PS029 manages to be as truthful as possible according to the evidence they have and is thus able to follow the maxim of quality. At the same time, it is not relevant for the speakers to know the exact term for the worktop material. Using the general noun phrase head stuff, speaker PS029 implies that they speak of a type of stone material used for worktops and floors. Given that the other speakers share the situation and are cooperative, they understand what is meant. Example (11) has shown how the general noun phrase head stuff is typically used in conversation. This use is representative for many similar uses of the general noun phrase head stuff and also thing in the conversation corpus.

7.1.5 Summary and comparison of results The preceding sections have dealt with the results of the quantitative analysis for each of the four sub-corpora. The aim of the present section is to compare these results and discuss in how far they confirm the assumptions concerning the frequency and distribution of general noun phrases across different genres and media (cf. Section 4.4). The results of the quantitative analysis confirmed the first of my assumptions, as they show that general noun phrases are indeed used in naturally occurring language data, namely in different texts from different genres and media (see Section 4.4). However, this does not apply to all 18 general noun phrase head types investigated in the present study. The analysis of the four sub-corpora showed that the general noun phrase head object does not occur at all in the four sub-corpora. The general noun phrase heads boy, creature, stuff and business do not occur in three of the four sub-corpora, the general noun phrase head move does not occur in two of the four sub-corpora and the general noun phrase heads affair, matter, girl and thing do not occur in one of the four sub-corpora. This means that the general noun phrase heads people, person, man, woman, child, 165

place, question and idea occur in all four sub-corpora investigated (for a detailed discussion of the distribution of general noun phrase head types in the four subcorpora, see Figure 7.7). The quantitative analysis revealed that the normalised overall frequency of general noun phrase heads differs across the four sub-corpora of the present study. The following figure presents a comparison: Figure 7.5: Comparison of the normalised overall frequencies of general noun phrase heads in the four sub-corpora 800 700 600 500 400 300 200 100 0

Judgment corpus

Manifesto corpus

Debate corpus

Conversation corpus

Figure 7.5 displays the normalised frequencies of general noun phrase heads in the four sub-corpora. The manifesto corpus has the highest overall frequency of general noun phrase heads with 829.42 occurrences. It is followed by the debate corpus with a normalised overall frequency of 567.84 general noun phrase heads, and the conversation corpus with the third highest normalised overall frequency of general noun phrase heads with 299.77 occurrences. The difference between the frequencies of general noun phrase heads in these three corpora is approximately the same. The judgment corpus, which has the lowest normalised frequency of general noun phrase heads (231.85 occurrences), closely follows the conversation corpus. The results illustrated in Figure 7.5 confirm the second of my assumptions, which stated that both sub-corpora from the political domain show a higher frequency of general noun phrase heads than the legal language corpus which shows a lower frequency of general noun phrase heads (see Section 4.4). The conversation corpus, which was assumed to have the highest frequency of general noun phrase heads, surprisingly shows only the third highest frequency of general 166

noun phrase heads. The findings displayed in Figure 7.5 confirm that the overall frequency of general noun phrase heads differs across genres. The third of my assumptions, which stated that the overall frequency of general noun phrase heads also depends on the medium and that spoken language displays a higher overall frequency of general noun phrase heads than written language, could, however, not be generally confirmed. This can be explained by the fact that the overall frequency of general noun phrase heads in the written medium is both very high (in the manifesto corpus) and very low (in the judgment corpus). As explained in Section 4.4, the frequency of general noun phrase heads in a text is an indication of its generality or specification. Accordingly, the four subcorpora can be presented along the scale of specification, as illustrated in the following Figure 7.6. Note, however, that the frequency of general noun phrases is only a first indication of the generality and specification of the corpus texts under investigation and that rather the parameters modification and linkage (cf. Section 7.2) determine the generality and specification of the general noun phrases, and thus of the corpus texts. Figure 7.6: The four sub-corpora on the scale of specification general high frequency of general noun phrase heads

829.42

Judgment corpus

Manifesto corpus

567.84

Debate corpus

299.77

C Conversation i corpus

231.85

l frequency f low off generall noun phrase heads specific

167

Figure 7.6 shows that the manifesto corpus has the highest normalised overall frequency of general noun phrase heads with 829.42 and is therefore positioned at the very general end of the scale of specification. The debate corpus follows with a normalised overall frequency of general noun phrase heads of 567.84 and is thus in a relatively central position with a slight tendency towards the general end of the scale. The conversation corpus follows with a normalised overall frequency of general noun phrase heads of 299.77. Thus, the conversation corpus is positioned towards the specific end of the scale. This was not expected, but can be explained in the context of the restricted style of referencing in conversations. Biber et al. (cf. 1999: 235, 578) and Esser (cf. 2009: 36) state that conversations typically have a very high frequency of pronouns and a very low frequency of nouns (for a detailed discussion, see Sections 3.3.3 and 7.1.4). Thus, in view of the fact that conversations have a low frequency of nouns altogether, it is understandable that the normalised overall frequency of general noun phrases is accordingly low in conversations. The judgment corpus is positioned near the conversation corpus and exhibits the lowest normalised overall frequency of general noun phrase heads with 231.85. It is therefore positioned at the specific end of the scale. As we can see, the fourth of my assumptions concerning the positioning of the four sub-corpora on the scale of specification was not confirmed. According to the fourth of my assumptions (see Section 4.4), the two spoken sub-corpora (first the conversation corpus, then the debate corpus) were positioned at the general end of the scale of specification, while the two written sub-corpora (first the manifesto corpus then the judgment corpus) were positioned at the specific end of the scale. It was thus assumed that the positioning of the sub-corpora on the scale of specification would first and foremost dependent on the medium. However, as discussed above (see Figure 7.5), the overall frequency of general noun phrase heads in a sub-corpus, and thus also its positioning on the scale of specification, does not depend on the medium but on the genre. As illustrated above, this can be explained by the fact that in the two written sub-corpora, the frequency of general noun phrase heads is very high (in the manifesto corpus) and also very low (in the judgment corpus). In the discussion of the overall frequency of general noun phrase heads so far, we did not take into account the difference in the range of general noun phrase types that occurred in the four sub-corpora. Thus, another interesting finding the quantitative analysis has revealed is the distribution of general noun phrase head types: there are differences as well as similarities across the four sub-corpora. The following figure provides an overview: 168

169 0

20

40

60

80

100

120

140

160

180

200

220

240

260

280

300

320

340

360

380

400

420

440

460

480

500

520

540

person

Judgment corpus

matter question

child

Manifesto corpus

people woman child

Debate corpus

matter

Figure 7.7: Comparison of the distribution of general noun phrase heads in the four sub-corpora

people man woman child boy girl creature thing object stuff business affair move place idea

people

person man woman boy girl creature thing object stuff business affair matter move place question idea person man boy girl creature thing object stuff business affair

move place question idea

Conversation corpus

people person man woman child boy girl creature object business affair matter move place question idea

thing

stuff

Figure 7.7 compares the distribution of general noun phrase heads across the four sub-corpora that were investigated. One interesting aspect that Figure 7.7 shows is that the extent to which the whole range of 18 different general noun phrase head types are used in the sub-corpora differs. As explained at the beginning of the present section, there are certain general noun phrase heads from Halliday/Hasan’s list that do not occur at all in the sub-corpora. The conversation corpus makes use of the highest number of different general noun phrase heads (14 types), followed by the manifesto corpus and the debate corpus, which have the same number of different general noun phrase heads (13 types), followed by the judgment corpus, which makes use of the smallest number of different general noun phrase heads (10 types). Apart from the number of different general noun phrase heads in the four sub-corpora, it is interesting to see how frequently these different general noun phrase heads occur in the four sub-corpora (tokens). This can be illustrated with the red line in Figure 7.7, which marks a normalised frequency of 20. It seems a reasonable limiting mark, which serves as a good indicator to illustrate differences and similarities in the distribution of general noun phrase heads across the four sub-corpora. When comparing the general noun phrase heads above that mark, it becomes apparent that in the judgment corpus, there are only three general noun phrase head types that occur more than 20 times. These are person, matter and question. Similarly, there are only three general noun phrase head types in the manifesto corpus that occur more than 20 times, namely people, child and place. In the debate corpus, there are six general noun phrase head types that occur more than 20 times. These are people, woman, child, matter, place, question and idea. In the conversation corpus, there are five general noun phrase head types, namely people, girl, thing, stuff and place, which occur more than 20 times. We can see that in the two spoken sub-corpora, the debate corpus and the conversation corpus, more different general noun phrase head types occurred frequently (more than 20 times). Therefore, we can say that it depends on the medium how many different general noun phrase head types occur frequently (more than 20 times). The quantitative analysis revealed another interesting aspect which can be observed in Figure 7.7: In each sub-corpus, two or more general noun phrase heads occurred that were considerably more frequent than the remaining general noun phrase heads in that sub-corpus. This is in line with Mahlberg’s discussion of local textual functions of general noun phrase heads (cf. 2005: 164). Mahlberg states that frequent nouns characterise a group of text. These nouns have local textual functions that refer to properties of texts, e.g. texts that deal with certain topics (cf. 2005: 164). Accordingly, we can say that those general noun phrase 170

heads that are considerably more frequent than the remaining ones have a characterising function for the sub-corpus they occur in. In the judgment corpus, these general noun phrase heads are person, matter and question. As explained in Section 7.1.1, person is typically used in legal language to refer to an unspecific individual. Question and matter are typically used to refer to the questions and the facts of the case. In the manifesto corpus, the general noun phrase heads that considerably outnumbered the remaining ones were people and child. As explained in Section 7.1.2, people is characteristic for British political manifestos in so far as it is typically used to address the British people, the possible future voters. Moreover, the general noun phrase head people is used to emphasise the integration of the British people in political concerns. Child was typically used when referring to (emotional) topics such as family life, childcare and education. The general noun phrase heads that characterise the debate corpus in terms of frequency were people, woman, child and matter. As explained in Section 7.1.3, people, woman and child were typically used because the parliamentary debates investigated dealt with topics that affected the British people, topics such as childcare, education and the equality of and violence against women. Matter was typically used to refer to (complex) political issues that were on the agenda. In the conversation corpus, the general noun phrase heads thing and stuff typically occurred. As Section 7.1.4 explained, this is due to the restricted style of referencing typically used in conversations. This style of referencing heavily relies on the shared situations, so that thing and stuff are typically used when referring to entities that are present in the situation of utterance. As an example of characterising general noun phrase heads in the four subcorpora see the following figures:

171

Figure 7.8: Comparison of the frequencies of the general noun phrase heads ‘people’ and ‘child’ 550 500 450 400 350 300 250 200 150 100 50 0

Judgment corpus

Manifesto corpus people

Debate corpus

Conversation corpus

child

Figure 7.9: Comparison of the frequencies of the general noun phrase heads ‘matter’ and ‘question’ 110 100 90 80 70 60 50 400 30 20 10 0

Judgment corpus

Manifesto corpus matter

Debate corpus

conversation corpus

question

Figure 7.8 illustrates that the use of textually characteristic general noun phrase heads is similar in the manifesto corpus and in the debate corpus. This can be explained by the fact that both sub-corpora belong to the same domain and that the subject-matters dealt with in the text samples are therefore similar. Interestingly, 172

the judgment corpus and the debate corpus, which belong to different domains, also show similarities in the use of typically characterising general noun phrase heads (cf. Figure 7.9). Although this similarity cannot be explained with a domain-specific use of typical general noun phrase heads, and certainly not with a medium- or genre-specific use of typical general noun phrase heads, it can nevertheless be explained with a similarity in the discourse function of the general noun phrase heads matter and question. In the judgment corpus, as well as in the debate corpus, matter and question are used to refer to (complex) questions and facts that are dealt with in the Supreme Court judgments investigated, or that are on the agenda of the parliamentary debates investigated. Having discussed and interpreted the overall frequency, the range and the distribution of general noun phrase heads in the four sub-corpora, the following can be summarised: If only the overall frequency of general noun phrase heads in the four sub-corpora is considered, the difference is genre-dependent. If we consider the range of different general noun phrase head types (i.e. how many of the 18 general noun phrase head types do in fact occur in the four sub-corpora), we can see a small difference between the domains that the four sub-corpora represent. The conversation corpus has the highest amount of different general noun phrase types, followed by the two sub-corpora from the political domain (debate corpus and manifesto corpus) which each have the same number of general noun phrase head types. The corpus from the legal domain has the lowest amount of different general noun phrase types. Now, if we only take into account the range of different general noun phrase head types with a minimum token frequency of 20, we can see a medium-based difference between the four sub-corpora. Figure 7.7 has shown that the two spoken sub-corpora, the debate corpus and the conversation corpus, show a wider distribution of general noun phrase head types with a minimum frequency of 20. There are six different types in the debate corpus and five in the conversation corpus while there are only three different general noun phrase head types with a minimum frequency of 20 in the two written corpora (judgment and manifesto corpus).

7.2 Qualitative analysis: The degree of specification of relevant general noun phrases in the corpora 7.2.1 Qualitative results from the judgment corpus After all relevant general noun phrase heads in the judgment corpus had been identified, numbered and coded, they were analysed in terms of modification and reference. The purpose of this was to determine their degree of specification 173

and, finally, to make a statement about their use across the different genres and media in the four sub-corpora of the present study. In a first step, it was determined how many of the 182 relevant general noun phrases in the judgment corpus were complex and how many were simple (+/- modification). Then, it was analysed how many of the 182 relevant general noun phrases in the judgment corpus were linked to their co-text through endophoric reference and how many were not linked (+/- linkage). The results are illustrated below: Table 7.5: Modified and non-modified general noun phrase heads in the judgment corpus Total no. of general noun phrases 182

+ Modification 108

– Modification 74

100 %

59.34 %

40.66 %

Table 7.6: Linked and non-linked general noun phrases in the judgment corpus Total no. of general noun phrases 182

+ Linkage 105

– Linkage 77

100 %

57.69 %

42.31 %

As we can see in Table 7.5, there were more complex general noun phrases in the judgment corpus than simple ones. In particular, 59.34% of the 182 relevant general noun phrases had a modifier, either premodified, postmodified or both. The other 40.66% of the 182 relevant general noun phrases were simple, which means their heads were not modified. In terms of linkage, Table 7.6 shows that the number of linked general noun phrases in the judgment corpus was higher than the number of those that were not linked. In particular, 57.69% of the relevant general noun phrases enodphorically referred to some other item in their co-text. The other 42.31% of the 182 general noun phrases had no endophoric reference. Note that neither the parameter modification nor the parameter linkage alone is sufficient to determine the degree of specification of general noun phrases. It is rather the correlation of both parameters. The outcome is illustrated in the following figure:

174

Figure 7.10: Correlation of modification and linkage of general noun phrases in the judgment corpus ‘rather specific’ general noun phrases

29.57% (55)

+ linkage

‘most specific’ general noun phrases

29.12% (53)

‘less specific’ general noun phrases

– linkage

+ modification

27.47% (50)

– modification

13.19% (24)

‘least specific’ general noun phrases

As illustrated in Figure 7.10, the distribution of general noun phrases is rather even across three of the four categories of specification: 29.57% of the total of 182 relevant general noun phrases were ‘most specific’, which means they were characterised by the parameters + modification and + linkage. Similarly, 27.47% of the relevant general noun phrases were defined as ‘rather specific’ (- modification/+  linkage), and 29.12% of the relevant general noun phrases were ‘less specific’ (+  modification/- linkage). By contrast, only 13.19% of the relevant general noun phrases were neither modified nor linked, which means they were ‘least specific’. In a second step of the qualitative analysis of general noun phrases in the judgment corpus, the distribution of general noun phrase heads across the four categories of specification was determined. The results are illustrated in the following figure:

175

Figure 7.11: Distribution of general noun phrase heads across the four categories of specification in the judgment corpus (1) (1) (1)

question (27) matter (22) person (1) ‘most specific’

29.57% (55)

woman place people

(2) (2) (1)

27.47% (50)

– modification

‘less specific’

‘least specific’

(53)

(31) (10) (5) (2)

‘rather specific’

+ modification

29.12% person matter question man

+ linkage

woman place idea

– linkage

question (46) person (3) matter (3)

13.19% (24)

person people man

(16) (2) (2)

question (2) child (1) affair (1)

Figure 7.11 shows which general noun phrase heads typically occurred in the four categories of specification in the judgment corpus (raw frequencies). In the ‘most specific’ and ‘rather specific’ category, question was the most frequent general noun phrase head. In the ‘less specific’ and ‘least specific’ category, person occurred most frequently as a general noun phrase head. These results reveal a parallel, which can be observed across the two categories that are characterised by the parameter linkage: Question was the general noun phrase head that was most frequently linked through endophoric reference, whether modified or not. Person was the general noun phrase head that most frequently occurred without endophoric reference, whether modified or not. There was no such parallel for the parameter modification. The following examples (12)-(15) from the judgment corpus demonstrate the general noun phrase heads that typically represent the four categories illustrated in Figures 7.10 and 7.11. These examples illustrate a decline from ‘most specific’ general noun phrases to ‘least specific’ general noun phrases in terms of presence and absence of the two parameters modification and linkage. (12) Given that Mrs Brown, in Arun, “cheated” on her neighbours and planning authority, should she too have lost the benefit (after whatever was the relevant limitation period) of immunity from enforcement action?

176

In responding with a resounding “no” to that forensic question (posed, I should at once make clear, in my language rather than Mr Maurici’s), it is necessary to identify […]. [J7-17-15]

Example (12) illustrates the use of a ‘most specific’ general noun phrase. The general noun phrase head question is premodified by the adjective forensic and postmodified by the past participle clause posed […] in my language rather than Mr Maurici’s. Endophoric reference further specifies the general noun phrase: that forensic question (posed […] in my language rather than Mr Maurici’s) anaphorically encapsulates what is being asked in the preceding three lines of example (12), marked with broken underlining. (13) The question whether the review procedure is prohibitively expensive is a matter1 that can, and should, be addressed by the Court itself. Preferably this should be done at the outset of the proceedings. […] But the advantages of having the matter2 resolved at the outset apply just as much at that stage as they do at first instance. [J5-14-4]

In example (13), the general noun phrase the matter2, which has no modifier, does not seem very specific at first. However, Halliday/Hasan’s anaphoric reference to the preceding general noun phrase the question whether […] (marked by broken underlining) specifies what the simple general noun phrase the matter2 refers to. Thus, although the matter2 is not specified through modification, it is nevertheless specified through anaphoric reference to another item in the text, which provides the content of the general noun phrase the matter2. This way, example (13) illustrates the use of a ‘rather specific’ general noun phrase. Note that the general noun phrase the matter2 does not refer back to the noun phrase a matter1 that can, […]. The function of a matter1 that can, […] is merely to complement the general noun phrase the question whether […]; it could just as well be left out. (14) Where any such determination as is referred to in subsection (1) above is made, the person making the determination shall […] determine whether […]. [J3-2-2]

The general noun phrase the person making the determination in example (14) is specified through the postmodifying present participle clause making the determination. However, since there is no reference to any preceding or following item in the text that would clarify what the general noun phrase refers to, example (14) is ‘less specific’.6

6

Note that the general noun phrase in example (14) is generic and therefore refers to a class of entities. This will be explained in detail in section 7.2.1.3 when discussing ‘less specific’ general noun phrases.

177

(15) By 1986, the law of unjust enrichment or restitution was by no means in its infancy. It was well understood that a person was in principle entitled to recover at common law money paid under a mistake of fact. [J3-2-7]

The simple general noun phrase a person in example (15) is ‘least specific’ because the noun phrase head is neither specified through modification nor does the general noun phrase refer to any preceding or following item in the co-text. Similar to the general noun phrase in example (14), the general noun phrase in example (15) has generic reference. So far, some first results show that complex general noun phrases are more frequent in the judgment corpus than simple general noun phrases, just as linked general noun phrases are more frequent than non-linked general noun phrases. Therefore, ‘most specific’ general noun phrases, which are characterised by modification and endophoric reference, are more frequent in the judgment corpus than ‘least specific’ general noun phrases, which are characterised by the absence of both of the parameters modification and linkage. Furthermore, question is typically found as a noun phrase head in ‘most specific’ and ‘rather specific’ general noun phrases. Person is typically found as a noun phrase head in ‘less specific’ and ‘least specific’ general noun phrases. For a finer classification of general noun phrases in terms of their degree of specification, the parameter modification was subcategorised and different types and realisations of modifier were determined. According to Quirk et al. (1985: 1243, 1321), premodifiers are generally less explicit than postmodifiers. This is due to the fact that postmodifiers express certain relations more explicitly (e.g. through the use of verbs, verb tenses). In the case of premodifiers, these relations are left to be inferred by the reader. While the noun phrase an oil man is not very explicit, the noun phrases a man who produces oil or a man who advocates the use of oil are more explicit. Therefore, Quirk et al. (1985: 1243) interpret postmodification as a more explicit description of the same content of premodification. Moreover, it was necessary to determine the different realisations of modifiers, as the degree to which modification specifies the head noun varies along the different types of phrases and clauses that realise modification. In this context, Jackson (1982), who discusses the different types of postmodifiers, states: “[a]s one passes from relative clauses through non-finite clauses to prepositional phrases, […] one finds a graduation from most to least explicit” (Jackson 1982: 70). The parameter linkage must be subcategorised into the different types of endophoric reference (as described in Section 6.2.3), since the degree of specification of a general noun phrase varies depending on these different types. A general 178

noun phrase that encapsulates a complex thought or fact is more specific than a general noun phrase that remotely refers to some item in the text that possibly goes back several paragraphs or pages. Thus, in order to make reliable statements about the degree of specification of general noun phrases in terms of modification and linkage, the following sections separately discuss the degree of specification of general noun phrases in the categories ‘most specific’, ‘rather specific’, ‘less specific’ and ‘least specific’.7

7.2.1.1 ‘Most specific’ general noun phrases in the judgment corpus The first category of general noun phrases in the judgment corpus, the category of ‘most specific’ general noun phrases, is defined by the correlation of the parameters + modification and + linkage. This category includes 55 general noun phrases with different types of pre- and postmodifiers and different types of endophoric reference. For the results of the more in-depth analysis of these 55 general noun phrases, see Tables 7.7–7.9:

7

Note that some of the numbers determined in the qualitative analysis of general noun phrases in the four sub-corpora were very small. These numbers should, however, not be considered statistically significant but are to be understood in the sense of an exhaustive list. For the sake of clarity, the tables which present the results of the qualitative analysis of general noun phrases in the four sub-corpora (Sections 7.2.1.17.2.4.4), will only present the (up to) five most frequent types of modifiers, endophoric reference, and combinations of those. Note that the discussion of these results and the illustration of examples may only focus on the (up to) three most relevant numbers in detail.

179

Table 7.7: Most frequent types of modification in ‘most specific’ general noun phrases in the judgment corpus + Modification Types of modification postmodifier

11

others

15

7

5

1

30

12

4

8 12

31

sum

relative clause

3

multiple

appositive int. clause

1

prepositional phrase

2

multiple

others

9

others

combination

single

adjective

single

adjective + appos. int. clause

premodifier

55

Table 7.8: Most frequent types of linkage of ‘most specific’ general noun phrases in the judgment corpus

H/H’s anaphoric reference

anaphoric encapsulation

cataphoric ref. to 2nd appos. unit

cataphoric ref. to subj. complement

remote reference

others

sum

+ Linkage Types of endophoric reference

3

4

23

12

6

7

55

180

Table 7.9: Most frequent combinations of modification and linkage of ‘most specific’ general noun phrases in the judgment corpus + Linkage

cataphoric ref. to subj. complement

remote reference

2

0

5

0

prep. phrase

0

0

2

0

Types of modification

+ Modification

post-mod.

appos. int. clause

0

14

0

0

relative clause

0

0

2

4

combination

adj. + inf. clause

0

0

2

0

adj. + appos. int. clause

0

4

0

0

others sum

sum

cataphoric ref. to 2nd appos. unit

adjective

pre-mod.

others

anaphoric encapsulation

Types of endophoric reference

31 55

Table 7.7 shows that of the 55 general noun phrases in the ‘most specific’ category, eleven were only premodified, one had multiple premodifiers, 30 were only postmodified, one had two postmodifiers and twelve were both pre- and postmodified. In terms of explicitness of the different types of modification mentioned earlier in Section 3.2.2 and 7.2.1 (cf. Quirk et al. 1985: 1243), this means that most of the 55 general noun phrases of the ‘most specific’ category had a very explicit type of modifier (postmodifier). The most frequent type of premodifier was the adjective, which occurred nine times, and the most frequent type of postmodifier was the appositive interrogative clause, which occurred 15 times with the general noun phrase head question. 181

This can be explained by the fact that the general noun phrase head question implies that it is likely to be postmodified by the grammatical category of interrogative clauses. The second most frequent type of postmodifier was the relative clause, which occurred seven times. Concerning endophoric reference, Table 7.8 shows that the most frequent type of endophoric reference was cataphoric reference to the 2nd appositive unit. This type occurred 23 times in the ‘most specific’ category of general noun phrases in the judgment corpus. In 21 of these occurrences, question was head of the general noun phrase with reference to its appositive unit. The second most frequent type of endophoric reference was cataphoric reference of the general noun phrase to its subject complement. In ten of the twelve instances in the ‘most specific’ category of general noun phrases, this type of reference occurred with the general noun phrase head question. Interestingly, the third most frequent type of endophoric reference was remote reference with six occurrences, four of which contained the general noun phrase head question. Remote reference is not considered to be a very specific type of linkage. However, all of the occurrences of remote reference were postmodified or pre- and postmodified, which in turn adds a certain degree of explicitness to the general noun phrase. The analysis of the general noun phrases in the first category has revealed four major types of ‘most specific’ general noun phrases (cf. Table 7.9), which are characterised by typical combinations of certain modifiers and certain types of reference. (i) The general noun phrase head question, which was postmodified by an appositive interrogative clause and which referred to this appositive unit cataphorically. (14 times) (ii) The general noun phrase head question, which was premodified by an adjective and which referred to its subject complement cataphorically. (five times) (iii) The general noun phrase heads question or person, which were postmodified by a relative clause and which remotely referred to some item or expression in the preceding text. (four times) (iv) The general noun phrase head question, which was premodified by an adjective and postmodified by an appositive interrogative clause and which referred to that appositive unit cataphorically. (four times) The following examples from the judgment corpus demonstrate the abovementioned patterns of general noun phrase use in the ‘most specific’ category:

182

(16) The focus […] was initially on […] the question whether the decision of AIT deprived him of an effective judicial remedy against an adverse act of the administration, contrary to general principles of European Union law. [J6-17-1]

Example (16) illustrates the typical use of the general noun phrase head question that is postmodified by an appositive interrogative clause (broken underlining). At the same time this appositive interrogative clause is the reference item to which the question refers cataphorically and which provides its content. (17) The essential question seems to be whether the bill of costs will be, or is, excessive bearing in mind the overriding requirement of access to justice. [J5-17-7]

Example (17) illustrates the typical use of the general noun phrase head question, which is premodified by an adjective. The whole general noun phrase the essential question refers to its subject complement, in this case an interrogative clause, cataphorically. Examples (16) and (17) are semantically very similar but differ with regard to their syntax: While the interrogative clause in example (16) is part of the general noun phrase, and thus a phrase element, it is the subject complement in example (17), and thus a clause element. (18) The question1 whether the planning assumptions can be taken into account too [J1-17-3] is the crucial question in this case. [8 pages in between] I do not think that the parties can have given express consideration to the question2 that has arisen in this case. [J1-17-5]

Example (18) shows the use of the general noun phrase head question2 that is postmodified by the relative clause that has arisen in this case. The general noun phrase remotely refers to the preceding general noun phrase the question1 whether the planning assumptions can be taken into account too that occurred 8 pages before. It is therefore questionable whether the reader still has in mind what exactly that question was. However, the question that has arisen in this case refers to the essential question in that case, which has been taken up and rephrased at several points during the judgment. Therefore, the reader, although he might not have in mind the exact wording of the question referred to by the general noun phrase [J1-17-5] in example (18), might still know what it was roughly concerned with. (19) Even assuming that it could be shown that the development of a hay barn was “begun” within section 56(2), this cannot assist on the essential question whether the building as constructed and completed was a barn […]. [J7-17-2]

Example (19) shows the general noun phrase head question, which is premodified by the adjective essential and postmodified by the appositive interrogative clause whether the building as constructed and completed was a barn. This 183

appositive interrogative clause is at the same time the reference item for the essential question. To sum up, the closer analysis of the general noun phrases in the ‘most specific’ category has revealed that explicit modification, particularly through the use of postmodifiers (most frequently appositive interrogative clauses), typically specifies the general noun phrase heads in this category. The most frequent type of reference that occurred in the ‘most specific’ category was cataphoric reference of a general noun phrase to its appositive unit. Concerning the correlation of modification and reference, it has been observed that the modified general noun phrase with question as its head and with reference to the 2nd appositive unit is the most frequent type of ‘most specific’ general noun phrase in the judgment corpus.

7.2.1.2 ‘Rather specific’ general noun phrases in the judgment corpus The second category of general noun phrases in the judgment corpus, the category of ‘rather specific’ general noun phrases, is defined by the parameters modification and + linkage. This category includes 50 general noun phrases that were not modified but linked to their co-text through endophoric reference. The following table illustrates the results of the more detailed analysis of these 50 general noun: Table 7.10: Most frequent types of linkage of ‘rather specific’ general noun phrases in the judgment corpus

H/H’s anaphoric reference

anaphoric encapsulation

cataphoric ref. to subj. complement

remote reference

others

sum

+ Linkage Types of endophoric reference

10

14

14

9

3

50

Table 7.10 presents the most frequent types of endophoric reference in the ‘rather specific’ category of general noun phrases in the judgment corpus: anaphoric encapsulation and cataphoric reference to the subject complement. Both types occurred 14 times. The second most frequent type of endophoric reference was Halliday/Hasan’s anaphoric reference, which occurred ten times.

184

The following examples demonstrate the typical behaviour of ‘rather specific’ general noun phrases in the judgment corpus. Example (20) below starts with an illustration of cataphoric reference to the subject complement: (20) The question is whether the agreement dated 20 October 2003 involved a return of capital by PPC to its shareholder TUK through TUK’s subsidiary Moorgarth (it being common ground that no relevant distinction exists in this context between TUK and Moorgarth). [J4-17-5]

In example (20), the general noun phrase the question is not modified. The general noun phrase is the subject of the clause complex and refers cataphorically to its complement, which is marked by broken underlining. Note that the subject complement is usually a noun phrase or an adjective phrase, but it can also be a nominal clause, such as the interrogative clause in example (20) (cf. Quirk et al. 1985: 728, 1050–1051). This pattern typically occurred with the general noun phrase the question in the ‘rather specific’ category. The following example demonstrates the other most frequent pattern of the ‘most specific’ category of general noun phrases in the judgment corpus, namely anaphoric encapsulation: (21) So long as PPC owned the YMS freeholds, it owned property which had been overvalued (on instructions) by about £4m. On this analysis the sale negotiated between Mr Price and Mr Moore, two experienced businessmen, was not at a gross undervalue, and perhaps not at an undervalue at all. But the dismissal of this appeal means that these matters will not be the subject of any further adjudication by the court. [J4-14-4]

In example (21), the simple general noun phrase these matters anaphorically encapsulates the arguments in the preceding two orthographic sentences in the preceding five lines of the example. The simple general noun phrase these matters encapsulates complex information and reifies it for the subsequent discourse in order to give density to the text. This makes reading more economic. When comparing Tables 7.8 and 7.10, it becomes apparent that anaphoric encapsulation was found more frequently in the ‘rather specific’ category of general noun phrases than in the ‘most specific’ category. Thus, it is assumed that anaphoric encapsulation shows a tendency to occur with general noun phrases that have no modifier. This might be due to the fact that non-modified general noun phrase heads, as for example matter, are more suitable as neutral shells for complex facts and thoughts and are also more suitable for making reading more economic than general noun phrase heads that are modified and thus more specific. In cognitive terms, we could say that non-modified ‘empty containers’ are better suited for encapsulating complex content than ‘containers which are already filled with content’ through modification. Example (22) below illustrates this nicely. 185

(22) The appeal arose out of a Judicial Review, which was initially brought by a Mr David Edwards. He instructed Mr Richard Buxton, whose fees were funded by the Legal Services Commission. The claim was dismissed by Lindsay J ([2005] EWHC 657) and Mr Edwards brought an appeal to the Court of Appeal. On the third and final day of that appeal Mr Edwards withdrew his instructions from Messrs Richard Buxton, and, at that stage, Mrs Pallikaropoulos was added as an additional party in order to continue the appeal. Mrs Pallikaropoulos was not eligible for legal aid, but the Court of Appeal made a costs capping order limiting her exposure to the Respondents’ costs to the sum of £2,000. The appeal to the Court of Appeal was dismissed ([2006] EWCA Civ 1138) and Mrs Pallikaropoulos successfully petitioned the House of Lords for leave to appeal to that court. Having obtained leave she then applied for a waiver of the security sum payable on such an appeal, and also applied for a protective costs order. By letter dated 22 January 2007 the Judicial Office indicated to her that, on the basis of the information then before them, the members of the Appeal Committee were not then minded to grant either application. The appeal was heard in January 2008 and lasted three days. On 16 April 2008 the House of Lords dismissed the appeal, thereby affirming the Court of Appeal’s decision. The matter was then adjourned for the parties to make written representations on costs. [J5-14-8]

In example (22), the simple general noun phrase the matter anaphorically encapsulates the underlined description of the facts of the case, which covers ten orthographic sentences over 24 lines. The general noun phrase the matter reifies this complex description and, whenever the matter is mentioned henceforth, the reader reactivates the description of the facts of the case. See another example below: (23) In more concrete terms, if a formal notice had been given, Mrs Wilson would have been warned about the situation and about the danger of being ejected from her home, before any proceedings were started. Which seems only reasonable. Approaching the matter on this footing, I would have allowed the appeal. [J2-14-2]

In example (23), the simple general noun phrase the matter makes general reference to the facts of the case. In contrast to examples (21) or (22), however, no concrete linguistic item or items in the preceding or following text can be located as reference item(s) that the general noun phrase the matter refers to (reference here is actually remote). Instead, the matter refers to every aspect of the facts of the case that has been described up to that point. This construction was typical of the Supreme Court judgments analysed. The following example demonstrates the second most frequent type of endophoric reference in the ‘rather specific’ category: (24) […] no clear and simple answer is available to the question as to what is the right text. […] In any event it cannot be said to be so obvious as to leave no reasonable scope for doubt as to the manner in which the question would be resolved […]. [J5-17-16]

186

In example (24), the simple general noun phrase the question anaphorically refers to the preceding general noun phrase the question as to what is the right text. To sum up, in the ‘less specific’ category of general noun phrases in the judgement corpus, anaphoric encapsulation was most typically used with non-modified general nouns as neutral shells to encapsulate complex thoughts and facts.

7.2.1.3 ‘Less specific’ general noun phrases in the judgment corpus The third category of general noun phrases in the judgment corpus, the category of ‘less specific’ general noun phrases, is defined by the parameters + modification and - linkage. This category includes 53 general noun phrases that are specified through modification but not through linkage to their co-text. An important observation was made when analysing the ‘less specific’ and also the ‘least specific’ general noun phrases in the judgment corpus: the general noun phrases without endophoric reference very often generically referred to a class of entities in the extra-linguistic world. Thus, generic reference of general noun phrases seemed typical of Supreme Court judgments. The results of the more in-depth analysis of the 53 general noun phrases in the ‘less specific’ category are discussed in the following: Table 7.11: Most frequent types of modification in ‘less specific’ general noun phrases in the judgment corpus + Modification Types of modification postmodifier

premodifier

11

others

6

13

3

35

5

sum

relative clause

5

multiple

past part. clause

8

combination

8

3

present part. clause

4

prepositional phrase

4

multiple

others

single

adjective

single

2

53

40

187

postmodifier

Types of modification

+ Modification

Table 7.12: Most frequent combinations of modification and - linkage in ‘less specific’ general noun phrases in the judgment corpus – Linkage no reference 4

prep. phrase

generic reference 4

present participle clause

5

0

past participle clause

5

1

relative clause

7

6

14

7

35

18

others sum

sum

53

Table 7.11 illustrates the types of modification that were found in the ‘less specific’ category. It reveals that of the 53 general noun phrases in the ‘less specific’ category, eight were premodified only, three had multiple premodifiers, 35 were postmodified only, five had more than one postmodifier and two were both preand postmodified. This suggests that most of the 53 general noun phrases of the ‘less specific’ category were modified by a rather explicit type of modification (postmodifer). The most frequent type was the relative clause, which occurred 13 times. In seven of these instances, person was the head of the general noun phrase. The second most frequent type was the postmodifying prepositional phrase (8 instances), followed by the past participle clause (6 instances). Of the 53 general noun phrases in the ‘less specific’ category, 35 general noun phrases were used to make generic reference. This equals 66.04% of the total. In a majority of 29 of the instances with generic reference, person was the head. The remaining 18 general noun phrases in the ‘less specific’ category were nonreferring, which means they were used to make neither endophoric nor generic reference. This amounts to 33.96% of the total in the ‘less specific’ category. Here, matter was the most typical non-referring general noun phrase head. Concerning the correlation of modification and reference, which is illustrated in Table 7.12, two typical patterns of ‘less specific’ general noun phrases will be pointed out here: (i) The generic general noun phrase with the general noun phrase head person and a relative clause as postmodifier. (seven times) (ii) The non-referring general noun phrase (most typically with the head matter), which was postmodified by a relative clause. (six times) 188

The following examples demonstrate these two typical types of ‘less specific’ general noun phrases: (25) Section 83 of the Act gives a specific right of appeal against a refusal of asylum to a person who, like FA, has been granted leave to enter or remain for a period exceeding one year. [J8-2-1]

Example (25) displays the use of a postmodified general noun phrase that generically refers to a class of persons. Through the use of the indefinite article, the generic general noun phrase functions as a representative member of the class of persons who have been granted leave to enter or remain for a period exceeding one year (cf. Quirk et al. 1985: 281). The comparison like FA emphasises this function. Example (25) is representative of many other uses of the general noun phrase head person in the judgment corpus. Person is typically used in Supreme Court judgments to make generic reference, particularly when quoting or referring to acts, statutes and directives and when discussing legal questions referring to the case in question. (26) In the present case, the question1 is whether it is right to describe a dwelling house as having or being of no use as a dwelling house, when it has just been completed and its owner intends to occupy it within days. This too is not a question2 which can sensibly be answered on a day by day basis. [J7-17-7]

Example (26) shows the use of a postmodified non-referring general noun phrase. The general noun phrase head question2 is postmodified by the relative clause which can sensibly be answered on a day by day basis. The general noun phrase is non-referring: It is referring neither anaphorically nor cataphorically to another item in the text and it does not make generic reference. The general noun phrase a question2 which can sensibly be answered on a day by day basis is the complement for the subject this and therefore only has an attributive function. (27) This makes it unnecessary at this point to decide whether change of use under section 171B(2) can consist in a simple departure from permitted use, without any actual prior use. I doubt this, since the word “use”, in each place where it appears in that subsection is on its face used in a real or material sense, rather than in the legal sense of “permitted use”. [J1-16-1]

In example (27), the general noun phrase each place where it appears in that subsection does not refer endophorically to any preceding or following item in the co-text. Moreover, the general noun phrase is not used generically. Therefore, it is non-referring. To sum up, the more detailed analysis of general noun phrases in the ‘less specific’ category has shown that the explicit types of postmodification typically 189

specified the general noun phrase heads in that category. It has also been shown that ‘less specific’ general noun phrases, in particular those with the general noun phrase head person, were typically used for generic reference in parts of the Supreme Court judgments that quote acts, statutes and directives or that deal with legal questions.

7.2.1.4 ‘Least specific’ general noun phrases in the judgment corpus The fourth category of general noun phrases in the judgment corpus, the category of ‘least specific’ general noun phrases, is defined by the parameters - modification and - linkage. This category contains 24 general noun phrases that are neither modified nor linked to their co-text through endophoric reference. The closer analysis of the least specific category of general noun phrases revealed that 20 of the 24 ‘least specific’ general noun phrases had generic reference. This amounts to 83.33%. Within this group of generically referring general noun phrases, person was the most frequent head with 16 occurrences. The following example demonstrates the generic use of the general noun phrase person: (28) Lloyd LJ sought to distinguish the tax cases to which I have referred on the grounds that payments by the state to a person have nothing to do with the tax regime. [J3-2-8]

In example (28), the non-modified general noun phrase a person refers generically to the class ‘person’ of which it is a representative member. The remaining four general noun phrases in the ‘least specific’ category of the judgment corpus were non-referring. Neither of them was linked endophorically to a preceding or following item in the co-text nor were they used to make generic reference. The following example demonstrates this: (29) This question arises, for example, where a claimant has notified a change of circumstances (such as that he has begun full- time work or that his child has left the household) and by mistake the Department overlooks (or delays actioning) the notification and continues making benefit payments at the same rate; […]. [J3-5-1]

The general noun phrase his child in the above example is considered to be nonreferring. There is no endophoric reference relation to any preceding or following item in the co-text. Moreover, the general noun phrase is not generic. Since the noun phrase a claimant in the first line of example (29) is generic, the general noun phrase his child might also be considered generic. However, general noun phrases are usually not generic when they have a possessive determiner. Therefore, the general noun phrase his child in example (29) is considered to be non-referring. 190

To sum up, the frequent generic reference of general noun phrases in the ‘least specific’ category corresponds to the frequency of generic general noun phrases in the ‘less specific’ category. Of the 77 general noun phrases in the judgment corpus without endophoric reference, a majority of 55 make generic reference, which adds up to 71.43%. This seems typical of the Supreme Court judgments that were analysed.

7.2.2 Qualitative results from the manifesto corpus In a first step of the qualitative analysis of relevant general noun phrases in the manifesto corpus, it was determined how many of these 639 relevant general noun phrases were complex and how many were simple (+/- modification). Then it was determined how many of the 639 relevant general noun phrases in the manifesto corpus were linked and how many of them were not linked (+/- linkage). The results are illustrated below: Table 7.13: Modified and non-modified general noun phrase heads in the manifesto corpus Total no. of general noun phrases 639 100 %

+ Modification

– Modification

293

346

45.85 %

54.15 %

Table 7.14: Linked and non-linked general noun phrases in the manifesto corpus Total no. of general noun phrases 639 100 %

+ Linkage

– Linkage

23

616

3.60 %

96.40 %

As Table 7.13 displays, simple general noun phrases occurred more often in the manifesto corpus than complex general noun phrases. The majority of 54.15% of the relevant general noun phrases had no modifier while the remaining 45.85% of the 639 relevant general noun phrases in the manifesto corpus were either premodified, postmodified or pre- and postmodified. In terms of linkage, Table 7.14 illustrates that a majority of 96.40% of the relevant general noun phrases in the manifesto corpus were not linked to their co-text through endophoric reference. The remaining 3.6% of the relevant general noun phrases were linked to their co-text through endophoric reference.

191

Further analyses of the correlation of the parameters modification and linkage have revealed the following results: Figure 7.12: Correlation of modification and linkage of general noun phrases in the manifesto corpus ‘rather specific’ general noun phrases

3.29% (21)

+ linkage

‘most specific’ general noun phrases

42.57% (272)

‘less specific’ general noun phrases

– linkage

+ modification

0.31% (2)

– modification

53.83% (344)

‘least specific’ general noun phrases

In contrast to the judgment corpus, the manifesto corpus had a rather uneven distribution of general noun phrases across the four categories of specification. Interestingly, the categories which are characterised by the parameter - linkage have the highest numbers of general noun phrases: 272 of the 639 relevant general noun phrases in the manifesto corpus were defined as ‘less specific’ (42.57%), and there were 344 ‘least specific’ general noun phrases, which makes up 53.83% of the total. The two remaining categories of the manifesto corpus (+ linkage) were made up of 21 ‘most specific’ general noun phrases (3.29% of the total 639 relevant general noun phrases) and only two ‘rather specific’ general noun phrases (0.31% of the total). So far, the results presented above confirm the hypothesis that simple general noun phrases, which are not linked to their co-text through endophoric reference (‘least specific’ general noun phrases), are most frequent in the manifesto corpus. They represent a very low degree of specification. In a second step of the qualitative analysis of relevant general noun phrases in the manifesto corpus, the distribution of general noun phrase heads across the four categories of specification was determined. 192

Figure 7.13: Distribution of general noun phrase heads across the four categories of specification in the manifesto corpus question (2) child (1)

child (2) ‘most specific’

3.29% (21)

affair idea girl matter question

(3) (3) (1) (1) (1)

0.31% (2)

– modification

‘less specific’

‘least specific’

(272)

(197) (39) (14) (5) (5) (3)

‘rather specific’

+ modification

42.57% people child place person woman man

+ linkage

(8) (5) (5)

– linkage

people place idea

53.83% (344)

people child woman man

(213) idea (112) thing (10) girl (5)

(2) (1) (1)

Figure 7.13 shows that in three of the four categories of specification, namely the ‘most specific’ category, the ‘less specific’ category and the ‘least specific’ category, the general noun phrase head people occurred most often. In the ‘rather specific’ category, child occurred as the head in both of the general noun phrases. Child also occurs frequently in the other categories; it is second most frequent in the ‘less specific’ category and in the ‘least specific’ category. For a demonstration of the general noun phrases that typically represent the four categories of specification in the manifesto corpus illustrated, refer to the following examples (30)-(33). They nicely illustrate a decline in the degree of specification of general noun phrases: (30) But we believe Sure Start needs to work better because the people who need it most – disadvantaged and dysfunctional families – are not getting enough of the benefit. [CM-1-68]

In example (30) the general noun phrase people is ‘most specific’, as it is postmodified by a relative clause (who need it most) and also by an appositive noun phrase (disadvantaged and dysfunctional families). Furthermore, there is cataphoric reference of the first appositive unit (the people who need it most) to the 193

second appositive unit (disadvantaged and dysfunctional families) within the general noun phrase. (31) Living in a family on a low income, having special educational needs or disabilities, and being in care all remain strongly linked to children1 failing to achieve. We are determined to narrow the gap between these children2 and their peers, […]. [LM-5-27]

Example (31) shows the use of the ‘rather specific’ general noun phrase these children2, which anaphorically refers to the preceding noun phrase children1 failing to achieve in the preceding sentence and establishes a relation of co-reference between the two items. (32) We will seek to ensure that disabled people are able to lead dignified and independent lives, […]. [LM-1-40]

In example (32) the ‘less specific’ general noun phrase disabled people is specified through the premodifier, the adjective disabled. However, the general noun phrase is not linked to its co-text endophorically and is thus not further specified. Disabled people describes a category or a class and therefore has generic reference. (33) We must not let the mis-selling of financial products put people off saving. [CM-1-22]

The general noun phrase people in example (33) is ‘least specific’ because it is neither modified nor linked to its co-text through endophoric reference. Similar to example (32), people also refers generically to a category or class. To arrive at a more accurate and precise determination of the degree of specification of general noun phrases in the manifesto corpus, the following sections discuss the four categories of specification separately.

7.2.2.1 ‘Most specific’ general noun phrases in the manifesto corpus The category of ‘most specific’ general noun phrases is defined by the correlation of the parameters + modification and + linkage. This category includes 21 general noun phrases that had different premodifiers, postmodifiers or pre- and postmodifiers. Moreover, these general noun phrases were linked to their co-text through different types of endophoric reference. The following table illustrates the results of a closer analysis:

194

Table 7.15: Most frequent types of modification in ‘most specific’ general noun phrases in the manifesto corpus + Modification Types of modification postmodifier

2

11

sum

others

3

others

relative clause

4

combination

appos. that-clause

2

multiple

prep. phrase

single

5

1

21

4

15 Table 7.16: Most frequent types of linkage of ‘most specific’ general noun phrases in the manifesto corpus

cataphoric encapsulation 3

+ Linkage Types of endophoric reference cataphoric ref. to 2nd anaphoric ref. to subject/ others appos. unit object 8

4

6

sum 21

Of the 21 general noun phrases in the ‘most specific’ category, only one was premodified, which is subsumed under the category ‘others’ in Table 7.15. The majority of 15 general noun phrases were postmodified (eleven had one postmodifier, four had multiple postmodifiers) and five were pre- and postmodified. In terms of explicitness of different types of modification (cf. Quirk et al. 1985: 1243, Sections 3.2.2 and 7.2.1), this means that most of the 21 ‘most specific’ general noun phrases were modified by a rather explicit type of modification. The most frequent type of postmodifier was the appositive that-clause, which occurred four times with the general noun phrase head idea. Second most frequent was the relative clause, which occurred three times with the general noun phrase head place. Concerning endophoric reference, Table 7.16 shows that the most frequent type of endophoric reference was cataphoric reference to the second appositive 195

unit, which occurred eight times (most typically with the general noun phrase head people). The second most frequent type of endophoric reference that occurred within the ‘most specific’ category of general noun phrases was anaphoric reference to the subject or object of the clause (people and place were the two general noun phrase heads). Concerning the combination of modification and reference, there was a wide range of different types of combinations of modification and endophoric reference, which each occurred only once. Only two types could be considered most typical, although it must be emphasised that they were not significantly more frequent than the rest (cf. footnote 55 in Section 7.2.1): (i) The general noun phrase head idea, which was postmodified by an appositive that-clause displaying cataphoric reference of the first appositive unit (head) to the second appositive unit (postmodifer). (three times) (ii) The general noun phrase head place which was postmodified by a relative clause and which, together with the postmodifer, encapsulated following pieces of information cataphorically. (two times) The following examples demonstrate these two patterns of ‘most specific’ general noun phrases in the manifesto corpus: (34) It is a change from one political philosophy to another. From the idea1 that the role of the state is to direct society and micro-manage public services, to the idea2 that the role of the state is to strengthen society and make public services serve the people who use them. [CM-18-2], [CM-18-3] (35) Strong community life also depends on protecting the places in which people come together. […] we are investing £235 million to create new or refurbished play spaces and adventure playgrounds. We will protect the Post Office network, […]. The local pub and social club are also hubs of community life. [LM-16-9]

In example (34), the ‘most specific’ general noun phrases the idea1 that […] and the idea2 that […] both include a postmodification through appositive thatclauses (that the role of the state is…). In both cases, there is cataphoric reference within the general noun phrases to these appositive that-clauses. In example (35), the general noun phrase head places is postmodified by the relative clause in which people come together. The whole phrase cataphorically encapsulates the noun phrases (marked by broken underlining). To sum up, the closer analysis of general noun phrases in the ‘most specific’ category of the manifesto corpus has revealed that noun phrase heads in that category are typically postmodified by an appositive that-clause or a relative clause. Moreover, the type of endophoric reference that typically occurred in the ‘most specific’ category of general noun phrases in the manifesto corpus was 196

cataphoric reference to the second appositive unit and anaphoric reference to the subject or object of the clause. Concerning the correlation of modification and linkage, the general noun phrase head idea was typically postmodified by an appositive that-clause and cataphorically referred to this second appositive unit.

7.2.2.2 ‘Rather specific’ general noun phrases in the manifesto corpus The second category of general noun phrases in the manifesto corpus, the category of ‘rather specific’ general noun phrases, is defined by the parameters - modification and + linkage. This category includes only two general noun phrases, which were not modified but endophorically linked to their co-text. In both of the two ‘rather specific’ general noun phrases, child was the head of the phrase. One of the two general noun phrases was linked anaphorically to a preceding item, and the other was the second part of an apposition and referred anaphorically to its first appositive unit. The following two examples demonstrate the two instances of ‘rather specific’ general noun phrases found in the manifesto corpus: (36) Living in a family on a low income, having special educational needs or disabilities, and being in care all remain strongly linked to children1 failing to achieve. We are determined to narrow the gap between these children2 and their peers… [LM-5-27] (37) We will […] increase the funding of the most disadvantaged pupils, around one million children. [LDM-5-19]

In example (36), there are two instances of the general noun phrase head children. Children1 is the reference item and children2, which is considered here, is the referring item [LM-5–27]. Example (36) illustrates the use of the ‘rather specific’ general noun phrase these children2, which is not modified but which anaphorically refers to children1 failing to achieve and establishes a relation of co-reference between the two linguistic items. In example (37), the general noun phrase around one million children is not modified.8 It is the second appositive unit that stands in apposition to the first appositive unit, the noun phrase the most disadvantaged pupils. Here, the general noun phrase is the postmodifier for another preceding noun phrase. In example (37), the second appositive unit makes anaphoric reference to the first appositive unit. It is generally arguable whether second appositive units can refer back to first appositive units or whether second appositive units rather have an attributive or descriptive function only. For the purpose of the present study, we assume that second appositive units can 8

Note that for the purpose of the present study, determiners were not defined as premodifiers (cf. section 3.2.2).

197

in fact refer back to first appositive units, provided that they are definite. This means they are preceded by the definite article, other determiners with a similar function (such as demonstrative and possessive pronouns), or certain quantifiers such as many and several, or ordinal and cardinal numerals (cf. Quirk et al. 1985: 253–254, 261–264). Therefore, the second appositive unit in example (37) refers back to the first appositive unit. Moreover, in example (37) it is not clear that the second appositive unit is merely a description of the first appositive unit. The first appositive unit the most disadvantaged pupils can just as well be a description of the second appositive unit around one million children.

7.2.2.3 ‘Less specific’ general noun phrases in the manifesto corpus The third category of general noun phrases in the manifesto corpus, the ‘less specific’ category, is defined by the correlation of the parameters + modification and – linkage. This category includes 272 general noun phrases, which are specified through modification but not through linkage to their co-text. The following tables illustrates the results of the closer analysis of the general noun phrases in this category: Table 7.17: Most frequent types of modification in ‘less specific’ general noun phrases in the manifesto corpus + Modification Types of modification postmodifier

139

198

105

5

13

15 28

sum

7

others

25 100

multiple

17

others

51

relative clause

4

pres. part. clause

18

prep. phrase

13 135

multiple

others

104

combination

single

past participle

adjective

single

adjective + prep. phrase

premodifier

272

Table 7.18: Most frequent combinations of modification and - linkage of ‘less specific’ general noun phrases in the manifesto corpus – Linkage generic reference

no reference

89

15

prepositional phrase

34

17

relative clause

22

3

72

20

217

55

Types of modification

+ Modification

premodifier postmodifier

adjective

others sum

sum

272

Table 7.17 shows the frequency of types of modification that occurred within the general noun phrases of the third category of specification in the manifesto corpus. It becomes apparent that of the 272 ‘less specific’ general noun phrases in the manifesto corpus, 139 were premodified (135 had one premodifier, four had multiple premodifiers), 105 were postmodified (100 had one postmodifier, five had multiple postmodifiers) and 28 general noun phrases were pre- and postmodified. This shows that the most frequent type of modification of general noun phrases in the ‘less specific’ category was premodification, which, according to Quirk et al. (1985: 1243), is the least specific type of modification. The most frequent premodifier was the adjective, which occurred 104 times. In the majority of 85 of these cases, people was the head of the general noun phrase. The most frequent postmodifier was the prepositional phrase, which occurred 51 times. In 30 of these cases, people was head of the general noun phrase, followed by 13 instances of child as general noun phrase head. The second most frequent postmodifier was the relative clause with 25 instances. People was head of the general noun phrase in 19 of these instances. Of the 272 general noun phrases in the ‘less specific’ category, 217 were used to make generic reference, which adds up to 79.78%. In the majority of 163 of these generic general noun phrases, people was the head of the phrase. These results indicate characteristic patterns of general noun phrase use in political manifestos. The remaining 55 general noun phrases in the ‘less specific’ category had neither endophoric reference nor generic reference; they were classified as non-referring. The non-referring general noun phrases amount to 20.22% of the general noun phrases in the ‘less specific’ category. Typically, people was head of the phrase (34 times).

199

Concerning the combination of modification and linkage, Table 7.18 shows that generic general noun phrases most often had an adjective as premodifier. This pattern occurred 89 times. In a majority of 77 of these cases, people was the head of this generic general noun phrase. Non-referring general noun phrases typically had a prepositional phrase as postmodifier. This pattern occurred 17 times. In ten of these cases, people was head of the phrase. The following examples demonstrate the use of generic general noun phrases and non-referring general noun phrases in the ‘less specific’ category of general noun phrases in the manifesto corpus: (38) We will also reform Access to Work, so disabled people can apply for jobs with funding already in place for equipment and adaptation that they need. [LDM-1-29] (39) This will put £700 back into the pockets of millions of people on low and middle incomes and free 3.6 million more people on low incomes from having to pay any income tax at all. [LDM-1-12], [LDM-1-11]

In example (38), the general noun phrase head people is premodified by the adjective disabled. The general noun phrase disabled people refers generically to the class of disabled people; it is not linked endophorically within the text. In example (39), the general noun phrases millions of people on low and middle incomes and 3.6 million more people on low incomes are also not linked endophorically to any other item in the preceding or following text. Moreover, these general noun phrases are not generic, because the quantifiers millions of and 3.6 million more indicate that the general noun phrases refer to a certain number of people, and not to people in general. To sum up, the closer analysis of the general noun phrases in the ‘less specific’ category revealed that adjectival premodification was the most frequent type of modification in this category. It was also observed that a majority of 79.78% of the general noun phrases in the ‘less specific’ category was used to make generic reference, most frequently with the general noun phrase head people. Concerning the combination of modification and linkage, it was observed that the general noun phrase head people was typically used with an adjectival premodifier to make generic reference.

7.2.2.4 ‘Least specific’ general noun phrases in the manifesto corpus The fourth category of general noun phrases in the manifesto corpus, the category of ‘least specific’ general noun phrases, is defined by the parameters modification and - linkage. This category contains 344 general noun phrases that are neither modified nor linked to their co-text through any kind of endophoric reference. The majority of 76.45% of the 344 general noun phrases had 200

generic reference (263 instances). Thus, similar to the majority of general noun phrases in the ‘less specific’ category, the majority of general noun phrases in the ‘least specific’ category were used to make generic reference. In 181 of the 263 generic general noun phrases, people was the head of the phrase. This is followed by child, which was the second most frequent head of the generic general noun phrases (67 times). The remaining 81 general noun phrases in the ‘least specific’ category (23.55%) were non-referring. This means they were neither linked endophorically nor did they have generic reference. In 45 of these cases, child was the head of the non-referring general noun phrase. In 32 of these cases it was people. The following examples from the manifesto corpus demonstrate the generic use of the general noun phrase head people and the non-referring use of the general noun phrase head child: (40) This manifesto makes the case that there should be no return to business as usual. People have suffered too much with their jobs, livelihoods and confidence to allow a return to the same old ways. [LM-1-10] (41) For the next generation we will protect – not cut – the Child Trust Fund – the world’s finest universal savings policy for young people, already giving 4.8 million children a nest egg for the future. [LM-5-5]

Example (40) shows that the general noun phrase people is neither modified nor linked to its co-text through endophoric reference. People is used here to refer generically to a category or class and thus refers to people in general. Since the Labour manifesto is addressed to the British people, the general noun phrase people in the above example probably refers to the class of British people. In example (41), the non-modified general noun phrase children is non-referring. Through the use of the quantifier 4.8 million, the general noun phrase refers to a specific number of children. Therefore, it does not make generic reference to a group. Moreover, it is not linked endophorically. To sum up, a majority of 76.45% of the general noun phrases in the ‘least specific’ group of the manifesto corpus were used to make generic reference. In most of these cases, people was the head of the generic general noun phrase.

7.2.3 Qualitative results from the debate corpus In a first step of the qualitative analysis of relevant general noun phrases in the debate corpus, it was determined how many of the 455 relevant noun phrases in the debate corpus were complex and how many were simple (+/- modification). Then, it was determined how many of the 455 relevant general noun phrases

201

in the debate corpus were linked to their co-text through endophoric reference and how many were not linked (+/- linkage). The results are illustrated below: Table 7.19: Modified and non-modified general noun phrase heads in the debate corpus Total no. of general noun phrases 455 100 %

+ Modification

– Modification

271 59.56 %

184 40.44 %

Table 7.20: Linked and non-linked general noun phrases in the debate corpus Total no. of general noun phrases 455 100 %

+ Linkage

– Linkage

101 22.20 %

354 77.80 %

As we can see in Table 7.19, complex general noun phrases occurred more often in the debate corpus than simple general noun phrases. The majority of 59.56% of the 455 relevant general noun phrases in the debate corpus were either premodified, postmodified or pre- and postmodified, and the remaining 40.44% of the relevant general noun phrases in the debate corpus were not modified. In terms of linkage, Table 7.20 shows that the majority of 77.80% of relevant general noun phrases in the debate corpus are not linked to their cotext through endophoric reference. The remaining 22.2% of the 455 relevant general noun phrases in the debate corpus were linked to their co-text through endophoric reference. In order to fully determine the degree of specification of relevant general noun phrases in the debate corpus, the parameters modification and linkage were taken into account.

202

Figure 7.14: Correlation of modification and linkage of general noun phrases in the debate corpus ‘rather specific’ general noun phrases

9.23% (42)

+ linkage

‘most specific’ general noun phrases

50.33% (229)

‘less specific’ general noun phrases

– linkage

+ modification

12.97% (59)

– modification

27.47% (125)

‘least specific’ general noun phrases

Figure 7.14 displays that, similar to the distribution in the manifesto corpus, the biggest proportion of general noun phrases in the debate corpus were defined by the parameter - linkage. The majority of 50.33% were ‘less specific’, which means they were modified but not linked to their co-text through endophoric reference. Second most frequent were ‘least specific’ general noun phrases in the debate corpus, which make up 27.47% of the total. 9.23% of the total of relevant general noun phrases were ‘most specific’, and 12.97% were ‘rather specific’. So far, the results presented above confirm the hypothesis that simple general noun phrases, which are not linked to their co-text through endophoric reference (‘least specific’ general noun phrases), are frequent in the debate corpus. Furthermore, the results show that complex general noun phrases without linkage are the most frequent type of general noun phrases in the debate corpus. This confirms that general noun phrases with a rather low degree of specification are most frequent in the debate corpus. In a second step of the qualitative analysis of relevant general noun phrases in the debate corpus, it was analysed which general noun phrase heads were typically found across the four categories of specification of general noun phrases in the debate corpus.

203

Figure 7.15: Distribution of general noun phrase heads across the four categories of specification in the debate corpus child thing matter move place

(3) (2) (1) (1) (1)

matter (36) woman people (12) place question (7) Idea ‘most specific’

9.23% (42)

(131) (27) (24) (14) (11) (5)

girl person idea move business question

(5) (4) (4) (2) (1) (1)

12.97% (59)

– modification

‘less specific’

‘least specific’

(229)

(2) (1) (1)

‘rather specific’

+ modification

50.33% people matter child woman place man

+ linkage

(12) (12) (4) (3) (3)

– linkage

question idea people man woman

27.47% (125)

people child woman girl

(72) (23) (16) (7)

place person man

(3) (2) (2)

Figure 7.15 shows that in the ‘most specific’ category, the general noun phrase heads question and idea occurred most frequently. In the ‘rather specific’ category, the general noun phrase head matter was most frequent. In both categories, which are defined by the absence of the parameter linkage (the ‘less specific’ and the ‘least specific’ category), the general noun phrase head people occurred most frequently. The following examples (42)-(45) demonstrate typical general noun phrases of the four categories of specification in the debate corpus. These examples illustrate a decline from specific to general: (42) The idea of measuring pensioner poverty in terms of material deprivation is supplementary to the income measures, and we will continue to publish both. [D7-18-3]

In example (42), the general noun phrase is ‘most specific’: firstly, its head idea is postmodified by an appositive present participle clause (of measuring pensioner poverty in terms of material deprivation). Note that the structure the idea of measuring […] could also be understood as a complementing present participle clause of a preposition; however, in appositive structures, these present participle clauses have prepositions that are absent in the corresponding finite clause: X has 204

the idea that measuring […] is supplementary […] (cf. Quirk et al. 1985: 1272). Secondly, the general noun phrase in example (42) is further specified through cataphoric reference to the second appositive unit, the appositive present participle clause cited above. (43) I thank my hon. Friend for his kind words. I am well aware of his work in the past on the know how fund and I appreciate his continuing interest in the matter. [D1-14-1]

Example (43) shows the use of a ‘rather specific’ general noun phrase the matter, which is not modified but anaphorically refers to the preceding noun phrase the know how fund in the same sentence. This establishes a relation of co-reference between the two items. (44) People who have served in the armed forces need to declare that they have done so, […]. [D2-1-4]

In example (44) the ‘less specific’ general noun phrase head people is specified through the postmodifying relative clause who have served in the armed forces. However, the general noun phrase is not linked to its co-text endophorically and is thus not further specified. People who have served in the armed forces describes a category or a class of people and therefore has generic reference. (45) We have the Directgov site in place, and we are keen for people to be able to put petitions before the House at the earliest opportunity – and this provides the earliest opportunity. [D4-1-22]

In example (45), the general noun phrase head people is ‘least specific’ because it is neither modified nor linked to its co-text through endophoric reference. Here, people also refers generically to a category or class.

7.2.3.1 ‘Most specific’ general noun phrases in the debate corpus The category of ‘most specific’ general noun phrases is defined by the correlation of the parameters + modification and + linkage. This category includes 42 general noun phrases that have different premodifiers, postmodifiers or pre- and postmodifiers. Moreover, these general noun phrases were linked to their cotext through different types of endophoric reference. The following tables 7.21– 7.23 illustrate the results of the more in-depth analysis of these 42 general noun phrases:

205

Table 7.21: Most frequent types of modification in ‘most specific’ general noun phrases in the debate corpus + Modification Types of modification postmodifier

appos. int. clause

2

2

4

5

1

2

17

16

sum

appos. pres. part. clause

3

combination

prep. phrase

1

appositive noun phrase

others

2

others

genitive

2

relative clause

adjective phrase

11

multiple

appos. that-clause

adjective

single

others

premodifier

6

42

1 3

20

Table 7.22: Most frequent types of linkage of ‘most specific’ general noun phrases in the debate corpus

H/H’s anaphoric reference 14

206

+ Linkage Types of endophoric reference remote anaphoric cataphoric ref. to reference encapsulation 2nd appos. unit 7

13

5

others

sum

3

42

Table 7.23: Most frequent combinations of modification and linkage in ‘most specific’ general noun phrases in the debate corpus

remote reference

6

0

3

postmod.

appositive that-clause

0

4

0

relative clause

4

0

0

others sum

sum

cataphoric ref. to 2nd appos. unit

adjective

premod.

others

H/H’s anaphoric reference Types of modification

+ Modification

– Linkage Types of endophoric reference

25 42

Table 7.21 shows that of the 42 general noun phrases in the ‘most specific’ category, 16 were premodified, 20 were postmodified (17 were postmodified, three had multiple postmodifiers) and six were pre- and postmodified. In terms of explicitness of the different types of modification (cf. Quirk et al. 1985: 1243, Sections 3.2.2 and 7.2.1), this means that most of the 42 ‘most specific’ general noun phrases were modified by a rather explicit type of modification. The premodifier that occurred most frequently in the ‘most specific’ category of general noun phrases in the debate corpus was the adjective, which occurred eleven times. Typically, question was the head in these general noun phrases. The most frequent type of postmodifier was the relative clause, which occurred five times with a variety of general noun phrase heads. This is followed by the appositive that-clause, which occurred four times with the general noun phrase head idea. Concerning endophoric reference, Table 7.22 shows that the most frequent type of endophoric reference was Halliday/Hasan’s anaphoric reference, which occurred 14 times with a variety of general noun phrase heads. The second most frequent type of endophoric reference that occurred within the ‘most specific’ category of general noun phrases was cataphoric reference to the second appositive unit, which occurred 13 times. In the majority of these cases, idea was head of the referring general noun phrase.

207

The analysis of the general noun phrases in the ‘most specific’ category has revealed three major types of ‘most specific’ general noun phrases in the debate corpus, which were characterised by typical combinations of certain modifiers and certain types of endophoric reference (cf. Table 7.23). (i) A general noun phrase (with either man, child or people as head of the phrase), which was premodified by an adjective and anaphorically referred to a preceding item. (six times) (ii) A general noun phrase (with either people, child or matter as head of the phrase), which was postmodified by a relative clause and anaphorically referred to a preceding item. (four times) (iii) A general noun phrase (most typically with idea as its head), which was postmodified by an appositive that-clause and which referred to that second appositive unit. (four times) The following examples (46)-(48) demonstrate these typical patterns of ‘most specific’ general noun phrases in the debate corpus: (46) This morning, I met a young Chevening scholar from Iraq who is studying for an MSc in engineering and robotics at Sussex university. […] Will my hon. Friends confirm that we want as many overseas students like that young man as possible to come to the UK, […]? [D5-3-1]

In example (46), the ‘most specific’ general noun phrase that young man contains the premodifying adjective young and the phrase anaphorically refers to the complex noun phrase a young Chevening scholar from Iraq who is studying for an MSc in engineering and robotics at Sussex university. (47) However, what we have to realise as a country is that this is […] about recognising that the people who have been injured so badly in Iraq and Afghanistan will need a lifetime of help. […] I have a strong defence team and a strong health team, who are going to work together to ensure that we deliver for those people, who have done so much for us. [D1-1-26]

In example (47), the general noun phrase head people is modified by the relative clause who have done so much for us. The whole phrase anaphorically refers to the noun phrase the people who have been injured so badly in Iraq and Afghanistan in the first sentence of example (47). (48) […] I would have thought he would welcome the idea that as we move to the new benefit, we are planning to cash-protect those who are already in receipt of other benefits. [D7-18-1]

208

In example (48), the general noun phrase head idea is postmodified by the appositive that-clause that as we move to the new benefit, we are planning to cashprotect those who are already in receipt of other benefits. Within the general noun phrase, the first appositive unit (the idea) cataphorically refers to the second appositive unit (the appositive that-clause). To sum up, the closer analysis of general noun phrases in the ‘most specific’ category of the debate corpus has revealed that noun phrase heads in that category are typically premodified by an adjective or postmodified by a relative clause or an appositive that-clause. Moreover, the type of endophoric reference that typically occurred in the ‘most specific’ category of general noun phrases in the debate corpus was Halliday/Hasan’s anaphoric and cataphoric reference to the second appositive unit. Concerning the correlation of modification and linkage, it could be observed that general noun phrase heads, which were premodified by an adjective, or those, which were postmodified by a relative clause, typically displayed anaphoric reference to a preceding item in the text. General noun phrases (with idea as its head) which were postmodified by an appositive that-clause and which cataphorically referred to this second appositive unit occurred frequently, too.

7.2.3.2 ‘Rather specific’ general noun phrases in the debate corpus The second category of general noun phrases in the debate corpus, the category of ‘rather specific’ general noun phrases, is defined by the parameters - modification and + linkage. This category includes 59 general noun phrases that were not modified but were linked endophorically to their co-text. The following table illustrates the results of the closer analysis of these 59 ‘rather specific’ general noun phrases: Table 7.24: Most frequent types of linkage of ‘rather specific’ general noun phrases in the debate corpus + Linkage Types of endophoric reference H/H’s anaphoric reference anaphoric encapsulation 29

28

others

sum

2

59

Table 7.24 demonstrates that the most frequent types of endophoric reference found in the ‘rather specific’ category of general noun phrases in the debate corpus were Halliday/Hasan’s anaphoric reference with 29 instances and anaphoric encapsulation with 28 instances. In 16 of the 29 instances of Halliday/Hasan’s anaphoric reference, matter was head of the referring general noun phrase. It was also the most frequent head in 20 of the 28 instances of anaphoric encapsulation. 209

The following two examples demonstrate the typical use of general noun phrases in the ‘rather specific’ category in the debate corpus: (49) On press reports about the sharing of aircraft carriers, may I say that, despite having 500 constituents who work in the Upper Clyde shipyards, I have always seen the matter as a strategic, not primarily industrial, question? [D2-14-2] (50) He knows that my view is very clear that we would be better off with a British Bill of Rights rather than with the Human Rights Act, and that matter is being examined. [D1-14–5]

In example (49), the simple general noun phrase the matter anaphorically refers to the preceding noun phrase the sharing of aircraft carriers in the same sentence. In example (50), the simple general noun phrase that matter anaphorically encapsulates the preceding clause that we would be better off with a British Bill of Rights rather than with the Human Rights Act.

7.2.3.3 ‘Less specific’ general noun phrases in the debate corpus The third category of general noun phrases in the debate corpus, the category of ‘less specific’ general noun phrases, is defined by the correlation of the parameters + modification and - linkage. This category includes 229 general noun phrases, which were specified through modification but not through linkage to their co-text. The following tables illustrate the results of the closer analysis of the general noun phrases in this category: Table 7.25: Most frequent types of modification in ‘less specific’ general noun phrases in the debate corpus

72

210

38

13

121 134

multiple

14

others

56

relative clause

8

pres. part. clause

7

prep. phrase

12 64

multiple

others

45

single

past participle

adjective

single

13

sum

premodifier

combination

+ Modification Types of modification postmodifier

23

229

Types of modification

+ Modification

Table 7.26: Most frequent combinations of modification and - linkage of ‘less specific’ general noun phrases in the debate corpus

adjective past participle

11

1

prepositional phrase

28

28

relative clause

29

9

49

29

155

74

premodifier postmodifier

– Linkage no reference 7

generic reference 38

others sum

sum

229

We can see from Table 7.25 that of the 229 ‘less specific’ general noun phrases in the debate corpus, 72 were premodified, 134 were postmodified and 23 general noun phrases were pre- and postmodified. This shows that the most frequent type of modification of general noun phrases in the ‘less specific’ category was postmodification, which, according to Quirk et al. (cf. 1985: 1243), is a rather explicit type of modification. The most frequent premodifier was the adjective, which occurred 45 times. In 37 of these cases, people was head of the general noun phrase. The most frequent postmodifier was the prepositional phrase, which occurred 56 times. Most frequent heads in these phrases were people (25 instances) and matter (16 instances). The most frequent combination of preand postmodifiers was an adjective in combination with a prepositional phrase. This pattern occurred six times (typically with people as head of the phrase). Of the 229 general noun phrases in the ‘less specific’ category, 155 were used to express generic reference, which constitutes a majority of 67.98%. In 103 of these cases, people was head of the generic general noun phrase. This is in accordance with the results of the analysis of general noun phrases in the ‘less specific’ category of the manifesto corpus. In 74 of the 229 ‘less specific’ general noun phrases of the debate corpus were non-referring, which means they were neither linked on the text surface by endophoric reference nor did they have generic reference. Most typically, people was the head of a non-referring general noun phrase (28 times), followed by matter (21 times). In reference to Table 7.26, we can point out three major patterns of general noun phrase use in the ‘less specific’ category of the debate corpus:

211

The generic general noun phrase (typically with people as the head of the phrase), which was premodified by an adjective. (38 times) (ii) The generic general noun phrase (typically with people as the head of the phrase), which was postmodified by a relative clause. (29 times). (iii) The non-referring general noun phrase (typically with matter as the head of the phrase), which was postmodified by a prepositional phrase. (28 times) (i)

The following two examples demonstrate the typical use of a generic general noun phrase (i) and a non-referring general noun phrase (iii) in the ‘less specific’ category of general noun phrases in the debate corpus: (51) What proportion of his Department’s expenditure for 2010-11 supports access to culture for young people; and if he will make a statement. [D4-1-12] (52) Incentives for biofuel production are primarily a matter for my right hon. Friend the Secretary of State for Energy and Climate Change, but… [D3-14-3]

In example (51), the general noun phrase head people is premodified by the adjective young. The general noun phrase young people refers to a class of people and is thus generic. It is not linked endophorically. In example (52), the general noun phrase head matter is postmodified by the prepositional phrase for my right hon. Friend the Secretary of State for Energy and Climate Change. The general noun phrase is indefinite; this is why it is not referring anaphorically to the preceding noun phrase Incentives for biofuel production. It has a complementing or descriptive role and thus functions more like a predicative adjective here (cf. Quirk et al. 1985: 273). Therefore, the general noun phrase in example (52) is considered non-referring. To sum up, the closer analysis of the general noun phrases in the ‘less specific’ category has revealed that postmodification, as a rather explicit type of modification, typically occurred in this category. It also revealed that the majority of modified general noun phrases in the ‘less specific’ category (67.69%), in particular postmodified general noun phrases, were generic. The typical pattern, which occurred most often in the ‘less specific’ category of general noun phrases in the debate corpus, was the generic general noun phrase head people, which was premodified by an adjective.

7.2.3.4 ‘Least specific’ general noun phrases in the debate corpus The fourth category of general noun phrases in the debate corpus, the category of ‘least specific’ general noun phrases, is defined by the parameters - modification and - linkage. This category contains 125 general noun phrases, which are neither modified nor linked to their co-text through endophoric reference. 212

The closer analysis of the ‘least specific’ general noun phrases revealed that 86 of the 125 general noun phrases in this category (68.8% of the total) were used generically. This is similar to the use of general noun phrases in the ‘less specific’ category of the debate corpus. In 46 of the 86 generic general noun phrases, people was head of the phrase. Child was the second most frequent head of the generic general noun phrases with 19 occurrences. The remaining 39 of the 125 ‘least specific’ general noun phrases in the debate corpus were non-referring, which means they were neither linked endophorically within the text nor did they have generic reference. This amounts to 31.2%. In 26 of the 39 non-referring general noun phrases people was the head. The following examples from the debate corpus demonstrate the generic and non-referring use of the general noun phrase head people: (53) I do not want to burden people with regulation – that is not the direction that the Government are going in – and I am happy to meet my hon. Friend […]. [D3-1-1] (54) We estimate that universal credit as a static system, not even taking into account any dynamic effect, will lift 900,000 people out of poverty, […]. [D7-1-14]

Example (53) shows that the general noun phrase head people is neither modified nor linked to its co-text through endophoric reference. People is used here to make generic reference to a category or class and refers to people in general. By contrast, the simple general noun phrase people in example (54) is not generic but refers to a specific number of people. Since this general noun phrase is not linked endophorically, it is non-referring.

7.2.4 Qualitative results from the conversation corpus In a first step of the qualitative analysis of relevant general noun phrases in the conversation corpus, it was determined how many of the 212 relevant noun phrase heads in the conversation corpus were modified and how many of them were not modified (+/- modification). Then, it was determined how many of the 212 relevant general noun phrases in the conversation corpus were linked to their co-text through endophoric reference and how many were not linked (+/- linkage). The results are illustrated below:

213

Table 7.27: Modified and non-modified general noun phrase heads in the conversation corpus Total no. of general noun phrases 212 100 %

+ Modification

– Modification

98

114

46.23 %

53.77 %

Table 7.28: Linked and non-linked general noun phrases in the conversation corpus Total no. of general noun phrases 212 100 %

+ Linkage

– Linkage

35

177

16.51 %

83.49 %

As we can see in Table 7.27, simple general noun phrases occurred more often in the conversation corpus than complex general noun phrases. In particular, 98 of the 212 relevant general noun phrases were complex, which means they were either premodified, postmodified or pre- and postmodified. This amounts to 46.23% of the total. 112 of the 212 relevant general noun phrases were simple, which means they were not modified. This amounts to 53.77% of the total. In terms of linkage, Table 7.28 shows that a majority of general noun phrases in the conversation corpus were not linked to their co-text through endophoric reference. More specifically, 35 of the 212 relevant general noun phrases in the conversation corpus were linked to their co-text through endophoric reference. This makes up 16.51% of the total. The remaining 177 general noun phrases in the conversation corpus were not linked to their co-text through endophoric reference, which amounts to 83.49% of the total. The correlation of the parameters modification and linkage fully determine the degree of specification of relevant general noun phrases in the conversation corpus. For an illustration of the results, see the following figure:

214

Figure 7.16: Correlation of modification and linkage of general noun phrases in the conversation corpus ‘rather specific’ general noun phrases

7.55% (16)

+ linkage

‘most specific’ general noun phrases

38.68% (82)

‘less specific’ general noun phrases

– linkage

+ modification

8.96% (19)

– modification

44.81% (95)

‘least specific’ general noun phrases

Similar to the distribution of general noun phrases in the manifesto corpus and the debate corpus, we can see that also in the conversation corpus, general noun phrases are generally more frequent in the two categories, which are characterised by the parameter - linkage. There were 95 ‘least specific’ general noun phrases in the conversation corpus, which amount to 44.81% of the total of 212 general noun phrases. This is the largest category. Second largest is the ‘less specific’ category, which includes 82 of the 212 relevant general noun phrases in the conversation corpus. 8.96% of the relevant general noun phrases in the conversation corpus are ‘rather specific’, and the smallest category of the conversation corpus includes 7.55% of the relevant general noun phrases. So far, the results presented above confirm the hypothesis that simple general noun phrases, which are not linked to their co-text through endophoric reference (‘least specific’ general noun phrases), are most frequent in the conversation corpus. This implies that the conversations investigated are characterised by general noun phrases with a very low degree of specification. In a second step of the qualitative analysis of the conversation corpus, the distribution of general noun phrase heads across the four categories of specification was determined. The following figure shows the results:

215

Figure 7.17: Distribution of general noun phrase heads across the four categories of specification in the conversation corpus person place idea

(1) (1) (1)

stuff thing girl ‘most specific’

7.55% (16)

(21) (16) (12) (11) (7)

boy child woman person idea

(5) (4) (3) (2) (1)

place child question

(27) (16) (14) (14) (9) (5)

woman (4) place (3) man (1) creature (1) question (1)

(2) (1) (1)

8.96% (19)

– modification

‘less specific’

‘least specific’

(82)

(7) (6) (2)

‘rather specific’

+ modification

38.68% thing stuff place people girl

+ linkage

(5) (4) (3) (1)

– linkage

stuff thing affair people

44.81% (95)

thing stuff people girl child boy

Figure 7.17 shows that in the ‘most specific’ and the ‘rather specific’ category of general noun phrases in the conversation corpus, the general noun phrase head stuff was most frequent and the general noun phrase head thing was second most frequent. In the ‘less specific’ and the ‘least specific’ category of general noun phrases in the conversation corpus this order was reversed: Here, the general noun phrase head thing was most frequent and the general noun phrase head stuff was second most frequent. This is interesting, because stuff did not occur at all in the judgment corpus, the manifesto corpus and the debate corpus. For a demonstration of the general noun phrases typically representing the four categories of specification in the conversation corpus illustrated in Figure 7.17, see the following examples (55)-(58): (55) PS029: Yeah. They didn’t have any worktops, the green Ange said. She can’t remember any but there again she weren’t looking for green. But they’ve got that marble. You know what we had? Over Jan’s, pity we didn’t nick that weren’t it. They got that sort of marble stuff. [C1-11-4]

In example (55), the ‘most specific’ general noun phrase head stuff is premodified by the adjective marble. Concerning endophoric reference, we can see that there is a reference chain between the three items any worktops, that marble (marked 216

by broken underlining) and the general noun phrase that sort of marble stuff. It can be assumed that the second item in this chain, that marble, is elliptic, the noun worktop is left out. Therefore, the general noun phrase that sort of marble stuff anaphorically refers to the preceding item that marble (worktop). (56) PS01U: PS01V: PS01U: PS01V: PS01U: [C2-11-17]

We bought loads of stuff. I got Mm. erm a lovely piece of middle leg pork for Sunday there were erm a piece of turkey breast, a bit like a joint Mm. you know with all the skin put round.

Example (56) shows the use of a ‘rather specific’ general noun phrase, which is not modified but linked to its co-text through endophoric reference. The general noun phrase head stuff cataphorically encapsulates the utterance in the following two turns of the speaker PS01U, a lovely piece of middle leg pork for Sunday […] a piece of turkey breast, a bit like a joint […] with all the skin out around. (57) PS0E9: PS0E8: PS0E9: PS0E8: [C3-9-10]

Shall I ask daddy to get me a lable for my Huh, you can try. Well you can do one with the erm, with the little thing. I don’t need to take scissors or anything. No, no.

In example (57) the ‘less specific’ general noun phrase the little thing is specified through the premodifying adjective little. However, the general noun phrase is not linked to its co-text endophorically and is thus not further specified. The little thing probably refers to a kind of labelling machine that is either present in the situation or at least known by both speakers and is therefore common ground to them. (58) PS02E: PS029: PS02E: PS029: PS02E: [C1-9-11]

Oh have you got a lighter now? No. I’ve got matches behind you Sue. Up on the Oh. Why do you Over there Sue. Up on that thing Where? Oh.

In example (58), the general noun phrase that thing is ‘least specific’ because it is neither modified nor linked to its co-text through endophoric reference. Similar to the general noun phrase in example (57), that thing probably refers to something, which is present in the speaking situation, in this case probably a piece of furniture in the presence of the speakers. This is implied by the deictic expression 217

that. Examples (57) and (58) illustrate that ‘less specific’ and ‘least specific’ general noun phrases in conversations often express ‘conventional vagueness’, a type of vagueness that is defined as a natural language phenomenon (cf. Section 3.4.2.2). This type of vagueness, which can be expressed through general noun phrases, is typical of spoken face-to-face interaction. Channell (cf. 1994: 180) states that in unspecified and vague utterances (e.g. example (57) and (58)), very often the encoder does not know or has forgotten the actual name of something, e.g. the labelling machine or the piece of furniture. This lexical gap is then filled with a general noun phrase, which nevertheless conveys enough meaning and is easily identifiable for the decoder. Due to the shared situation in such conversations, the vague expressions are easily decodable and the utterance is nevertheless understandable. For a more accurate and precise determination of the degree of specification of general noun phrases in the conversation corpus, the following sections separately discuss the four categories of specification.

7.2.4.1 ‘Most specific’ general noun phrases in the conversation corpus The category of ‘most specific’ general noun phrases is defined by the correlation of the parameters + modification and + linkage. This category includes 15 general noun phrases that had different premodifiers, postmodifiers or pre- and postmodifiers. Moreover, these general noun phrases were linked to their co-text through different types of endophoric reference. Tables 7.29–7.30 below illustrate the results of a more in-depth analysis:

12

218

2

1 3

sun

2

multiple

1

single

3 10

multiple

others

6

noun

adjective

premodifier single

+ Modification Types of modification postmodifier

combination

Table 7.29: Most frequent types of modification in ‘most specific’ general noun phrases in the conversation corpus

1

16

Table 7.30: Most frequent types of linkage of ‘most specific’ general noun phrases in the conversation corpus

H/H’s anaphoric reference 10

+ Linkage Types of endophoric reference H/H’s cataphoric anaphoric reference encapsulation 2

2

others

sum

2

16

Table 7.29 shows that of 16 general noun phrases in the ‘most specific’ category, 12 were premodified, 3 were postmodified and one was pre- and postmodified. In terms of explicitness of different types of modification mentioned earlier in Sections 3.2.2 and 7.2.1 (cf. Quirk et al. 1985: 1243), this means that most of the ‘most specific’ general noun phrases were modified by a rather inexplicit type of modification, namely premodification. The premodifier that most often occurred in the ‘most specific’ category of general noun phrases in the conversation corpus was the adjective (six times). In three of these cases, stuff was head of the general noun phrase. Concerning endophoric reference, Table 7.30 shows that the most frequent type of endophoric reference was Halliday/ Hasan’s anaphoric reference, which occurred ten times with various general noun phrase heads. The analysis of the general noun phrases in the first category has revealed one major type of ‘most specific’ general noun phrases which is characterised by a typical combination of a certain modifier and certain type of endophoric reference. Note that all other combinations only occurred once (see footnote 55): (i) A general noun phrase (with various general noun phrase heads), which was typically premodified by an adjective and anaphorically referred to a preceding item in the text. For a demonstration of this typical pattern of ‘most specific’ general noun phrases in the conversation corpus, see the following examples: (59) PS01V: PS01U: PS01V: PS01U: PS01V: [C2-11-5] (60) PS02C: PS02F: […]

We went to MFI, we went to erm BQ, Furniture Factors didn’t we? And What you looking for? For bedroom stuff Mhm there’s some nice stuff, MFI beat the lot, didn’t they Alec? What’s Father Christmas bringing you Ben? Erm a bobblin

219

PS000: [C1-9-3]

Oh Ben. You stupid thing. Where’s my drink, where’s my drink?

In example (59), the general noun phrase head stuff is premodified by the adjective nice. Furthermore, the whole general noun phrase some nice stuff anaphorically refers to the preceding generic general noun phrase bedroom stuff. In example (60), the general noun phrase head thing is premodified by the adjective stupid. Here, the general noun phrase anaphorically refers to Ben. The general noun phrase carries an interpersonal element in its meaning conveying a certain attitude on the part of the speaker. According to Halliday/Hasan (1976: 276), this is an important function of general noun phrases. However, general noun phrases that conveyed interpersonal meaning, particularly through the use of attitudinal modifiers (such as in example (60)), hardly occurred in the corpus data. The instances that did occur predominantly occurred in the spoken conversation corpus.

7.2.4.2 ‘Rather specific’ general noun phrases in the conversation corpus The second category of general noun phrases in the conversation corpus, the category of ‘rather specific’ general noun phrases, is defined by the parameters - modification and + linkage. This category includes 19 general noun phrases, which were not modified but which were endophorically linked to their co-text. See the following table for the results of the closer analysis of these 19 ‘rather specific’ general noun phrases: Table 7.31: Most frequent types of linkage of ‘rather specific’ general noun phrases in the conversation corpus

H/H’s anaphoric reference 12

+ Linkage Types of endophoric reference H/H’s cataphoric cataphoric remote reference reference encapsulation 3

2

2

sum 19

Table 7.31 shows that the most frequent types of endophoric reference found in the ‘rather specific’ category of general noun phrases in the conversation corpus were Halliday/Hasan’s anaphoric reference with 12 instances and Halliday/ Hasan’s cataphoric reference with 3 instances. Of the 12 instances of Halliday/ Hasan’s anaphoric reference in the ‘rather specific category’, stuff was head in five of these general noun phrases, followed by thing in four of these phrases. The following two examples demonstrate the typical use of general noun phrases in the ‘rather specific’ category in the conversation corpus: 220

(61) PS02B: PS02F: PS029: PS029: PS02F: PS02B: PS029: PS02B: [C1-9-4]

No. What babe? And and the big bear said who’s been sleeping in my bed. Zoe wants I gotta tell her Yeah she wants I know more, I know more. And and the big bear said who been sleeping in my bed. That’s right. Very good. Thank you my love. said what’s that? I said that’s just, nothing. Oh it’s only a tape recorder. Oh that thing again!

In example (61), the simple general noun phrase that thing anaphorically refers to the preceding noun phrase a tape recorder. Note that the demonstrative pronoun that carries a certain derogatory tone, which might express that the speakers in utterance (61) are irritated by carrying the tape recorder that recorded the conversation. Similar to the use of the general noun phrase you stupid thing in example (60) in the preceding Section 7.2.4.1, the general noun phrase that thing in example (61) above also conveys interpersonal meaning. (62) PS0E8: PS0E9: PS0E8: PS0E8: [C3-9-14]

I’ve done the margarine, you you can get me the erm thing. Where’s the margarine? It’s in here, it’s melted, you’ve just watched it melt. Can you get me the flan tin down there?

In example (62), the simple general noun phrase the thing cataphorically refers to the noun phrase the flan tin in the next but two utterances. Note that similar to examples (57) and (58) in Section 7.2.4 above, example (62) shows the use of a rather vague general noun phrase which is probably used by the speaker because they cannot think of the actual name of the tin. The general noun phrase the thing is used to fill that lexical gap (cf. Channell 1994: 180).

7.2.4.3 ‘Less specific’ general noun phrases in the conversation corpus The third category of general noun phrases in the conversation corpus, the category of ‘less specific’ general noun phrases, is defined by the correlation of the parameters + modification and - linkage. This category includes 82 general noun phrases, which were specified through modification but not through linkage to their co-text. Tables 7.32 and 7.33 illustrate the results of the more in-depth analysis of these general noun phrases:

221

Table 7.32: Most frequent types of modification in ‘less specific’ general noun phrases in the conversation corpus

relative clause

others

5

5

11

2

2

23

46

multiple

infinitive clause

4

44

prep. phrase

8

32

multiple

others

single

noun

adjective

single

5

sum

premodifier

combination

+ Modification Types of modification postmodifier

8

82

28

Types of modification

+ Modification

Table 7.33: Most frequent combinations of modification and - linkage in ‘less specific’ general noun phrases in the conversation corpus

premodifier postmodifier

adjective

generic reference 3

– Linkage no reference 29

noun

1

7

infinitive clause

0

5

relative clause

5

6

5

21

14

68

others sum

sum

82

Table 7.32 gives the frequencies of the different types of modification that occurred within the general noun phrases of the third category of specification in the conversation corpus. Of a total of 82 ‘less specific’ general noun phrases, 46 were premodified, 28 were postmodified and eight general noun phrases were pre- and postmodified. This means that in the majority of ‘less specific’ general noun phrases in the conversation corpus, we find a rather inexplicit type of modification (cf. Quirk et al. 1985: 1243). The most frequent premodifier was the adjective, which occurred 32 times with a variety of different general noun phrase heads. The most frequent postmodifier was the relative clause, which occurred 222

eleven times, and neither showed a clear preference for a particular general noun phrase head. Table 7.33 illustrates that the majority of 68 general noun phrases in the ‘less specific’ category were non-referring. This amounts to 82.93% of the total. There was a range of different general noun phrase heads, e.g. thing, place, stuff, or people. The remaining 14 general noun phrases in this category (17.07%) were generic. There was no clear preference for a specific general noun phrase head. Concerning the combination of the parameters modification and linkage, Table 7.33 shows that there were three typical patterns of general noun phrases in the ‘less specific’ category: (i)

A generic general noun phrase, which was postmodified by a relative clause. (five times) (ii) A non-referring general noun phrase with an adjective as premodifier. (29 times) (iii) A non-referring general noun phrase with a noun as premodifier. (seven times) For a demonstration of these typical uses of generic and non-referring general noun phrases in the ‘less specific’ category in the conversation corpus, see the following examples: (63) PS029: […] PS029:

just say right we’ll put a pound a week in and at the end of the year

But you want people who’s gonna pay every week. You don’t want people you gonna have to keep [C1-1-4], [C1-1-5]

In example (63), the two general noun phrase heads people are both postmodified by a relative clause (who’s gonna pay every week and you gonna have to keep). Both general noun phrases refer to a class or group of people and are therefore generic. (64) PS01U: […] PS01V: […] PS01T: PS01U: PS01V: PS01U: PS01V:

Ah well I got my freezers from Millers, and I got my fridge. In fact, I looked at Currys’ prices, cos Currys’ were a, and Currys looked as if they were the cheaper But Currys were always de dearer than Millers. Different people who you speak to though Joyce Yeah. not just round here but er asking, and like, I’ve have our Yeah.

223

PS01U: PS01V: [C2-9-19]

Kim’s husband Ian th when they were shopping round for different things and Mm.

In example (64), the general noun phrase head things is premodified by an adjective. The general noun phrase different things is not linked to any other item in the text endophorically. It is also not generic, as it does not refer to a class or group of things but to specific things. The general noun phrase different things expresses conventional vagueness, a type of vagueness that is considered appropriate in conversations for two reasons: firstly, it can be easily decoded, due to the shared situation, that different things probably refers to different objects that one would want to shop. Secondly, in the context of the above conversation, it is not relevant to specify what the different things were that Kim and her husband Ian were shopping round for. (65) PS01T: PS01V: PS01T: PS01V: PS01T: PS01V: PS01T: PS01V: PS01T: PS01T: PS01V: PS01T: PS01V: PS01T: PS01V: PS01T: PS01V: PS01T: [C2-9-25]

Looks as though it had been cut out for some reason. It does, don’t it? Is it collar? Well I think it’s probably off that other pattern. Well I do, I’m waiting for two patterns aren’t I? Looks like er making for a collar don’t it? Yeah. I’ll keep that though because I don’t know if. I’ll put it in here. Mm. I’ll have that stapler. And I, if I see anybody I’m not gonna stop and talk to them! could of could of walked on. Mm. And every time, every time I’m glad she’s starting summat else! Mm mm. And then we went and had a look in that that er tape thing. Yeah. No, so it will be possible talk to them then? Yeah. It’s alright to talk to him you know i giving you, just give you that bit of extra warmth without er The weight.

In example (65), the general noun phrase head thing is premodified by the noun tape. The general noun phrase neither refers anaphorically or cataphorically nor does it have generic reference. The deictic pronoun that indicates that the general noun phrase that tape thing refers to an entity in the shared situation of the speakers. However, since exophoric reference is not a criterion for the analysis 224

in the present study, the general noun phrase that tape thing is defined nonreferring. To sum up, the closer analysis of the general noun phrases in the ‘less specific’ category has revealed that they were frequently premodified by an adjective and that they were non-referring in most of the cases (82.93%). This specific combination was also the most frequent type in the ‘less specific’ category.

7.2.4.4 ‘Least specific’ general noun phrases in the conversation corpus The fourth category of general noun phrases in the conversation corpus, the category of ‘least specific’ general noun phrases, is defined by the parameters modification and - linkage. This category contains 95 general noun phrases, which were neither modified nor linked to their co-text through any kind of endophoric reference. The closer analysis of the ‘least specific’ general noun phrases has revealed that, similar to the majority of general noun phrases in the ‘less specific’ category, the majority of general noun phrases in the ‘least specific’ category were non-referring (82 instances, which equals 86.32% of the total). The remaining 13.83% (13 instances) were generic. In eight of the 13 generic general noun phrases, people was head of the phrase. There was a greater variety of general noun phrase heads in the non-referring group, including heads such as thing (27 instances), stuff (16 instances), and girl (11 instances). For a demonstration of the generic and non-referring uses of the general noun phrase heads people, girl, thing and stuff see the following examples from the conversation corpus: (66) PS029: PS02E: PS029: [C1-1-7]

She’s a hairdresser. I’m a hairdresser, yeah. Does people’s hair like mummy’s. Makes me look tidy occasionally.

Example (66) shows that the general noun phrase people is neither modified nor linked to its co-text through endophoric reference. People is used generically to refer to the class people in general. (67) PS01V: PS01T: PS01U: PS01T: PS01U: PS01T:

Colonaise the council car park at Colonaise Pete works there don’t he? Yeah. Or just started to work there. It were thirty pence when it were, when it was er this fella owned it Yeah. the council’s took it over No, the council’s took it over, they sold it to this bloke.

225

PS01U: PS01V: PS01U: PS01V: [C2-9-13]

The council’s took it over now and it’s sixty pence! The council has because that’s who Pete works for! And that’s how he’s got that job. And what about me paying ten pound fine didn’t I? Cos I were twenty minutes over the Yeah. thing.

In example (67), the simple general noun phrase the thing is non-referring because it is neither endophorically linked within the surface text nor is there generic reference to a class. Through associations triggered by the co-text of example (67), more particularly by the noun phrase the council car park at Colonaise, it can be assumed that the general noun phrase the thing refers to a parking ticket; however, such associations are not a criterion for endophoric reference in the present study. (68) PS01V: Are you gonna have a bit of this stuff on some toast Alec? PS01T: No thanks love. [C2-11-18]

Example (68) shows a non-referring use of the simple general noun phrase this stuff. The deictic pronoun this indicates that the general noun phrase makes reference to something that is present in the situation of the conversation. However, exophoric reference is not considered in the present study.

7.2.5 Summary and comparison of results The preceding sections have dealt with the results of the qualitative analysis for each of the four sub-corpora. The aim of the present section is to compare these results and discuss in how far they confirm the assumptions concerning the use and the functions of general noun phrases across different genres and media (cf. Section 4.4). Assumptions five, six, and seven (cf. Section 4.4) were dealt with in individual steps of the qualitative analysis. In a first step, the number of simple and complex general noun phrases was determined. It was also analysed, how many of the relevant general noun phrases were linked and how many were not linked. Table 7.34 presents a comparison of these results:

226

Table 7.34: Comparison of modified and non-modified general noun phrase heads and linked and non-linked general noun phrases in the four sub-corpora (in percentage) Sub-corpora Judgment corpus

Complex General Simple general Linked general general noun noun phrases noun phrases noun phrases phrases without linkage 59.34 40.66 57.69 42.31

Manifesto corpus

45.85

54.15

3.60

96.40

Debate corpus

59.56

40.44

22.20

77.80

Conversation corpus

46.23

53.77

16.51

83.49

With regard to the parameter modification, Table 7.34 shows that in the judgment corpus and in the debate corpus, complex general noun phrase heads were more frequent than simple general noun phrase heads. Interestingly, the frequencies of modified and non-modified general noun phrase heads in both corpora were very similar, with approximately 60% modified and 40% non-modified general noun phrase heads. By contrast, in the manifesto corpus and the conversation corpus, simple general noun phrase heads were more frequent than complex ones. Here, the frequencies in both corpora were again very similar, with approximately 46% modified and 54% non-modified general noun phrase heads. Regarding the parameter linkage, Table 7.34 shows that, in the judgment corpus, general noun phrases which are linked to their co-text through endophoric reference were more frequent than general noun phrases without linkage. By contrast, in the manifesto corpus, the debate corpus and the conversation corpus, general noun phrases which are not linked to their co-text through endophoric reference were significantly more frequent than general noun phrases with linkage. Note that in the judgment corpus, there is only a small difference between the number of linked and non-linked general noun phrases (approximately 58% linked and 42% non-linked). In contrast, in the other three corpora, the difference between linked and non-linked general noun phrases is much bigger. In the debate corpus, the ratio of linked and non-linked general noun phrases is approximately 22% to 78%, in the conversation corpus, it is approximately 17% to 83%, and it is the biggest in the manifesto corpus with approximately 4% linked to 96% non-linked general noun phrases. An interesting observation, which was made in the first step of the qualitative analysis, was that the difference between the frequencies of modified and nonmodified general noun phrases is not so big in the four sub-corpora, whereas the difference between the frequencies of linked and non-linked general noun 227

phrases is much bigger in the four sub-corpora (cf. Table 7.34). This supports the assumption that linkage is a crucial parameter in the analysis of general noun phrases. In a second step of the qualitative analysis, the focus was placed on the correlation of the two parameters modification and linkage and the distribution of general noun phrases across the resulting four categories of specification: ‘most specific’, ‘rather specific’, ‘less specific’, and ‘least specific’ general noun phrases. See Table 7.35 for a comparison of the frequencies of general noun phrases across these categories: Table 7.35: Comparison of general noun phrases across the four categories of specification in the four sub-corpora (in percentage)

Sub-corpora

‘rather ‘most specific’ specific’ general noun general noun phrases phrases 29.57 27.47

‘less specific’ general noun phrases

‘least specific’ general noun phrases

29.12

13.19

Manifesto corpus

3.29

0.31

42.57

53.83

Debate corpus

9.23

12.97

50.33

27.47

Conversation corpus

7.55

8.96

39.15

44.34

Judgment corpus

Table 7.35 shows that in the judgment corpus, ‘most specific’ general noun phrases were most frequent and ‘least specific’ general noun phrases were least frequent. However, note that the distribution of general noun phrases across three of the four categories of specification is almost completely even in the judgment corpus. By contrast, in the manifesto corpus, ‘least specific’ general noun phrases were most frequent and ‘rather specific’ general noun phrases were least frequent. In the debate corpus, ‘less specific’ general noun phrases were most frequent and ‘most specific’ general noun phrases were least frequent. In the conversation corpus, ‘least specific’ general noun phrases were most frequent and ‘most specific’ general noun phrases were least frequent. Similar to the distribution of general noun phrases across the four categories of specification in the manifesto corpus, there is a great difference in frequency in the conversation corpus between ‘most specific’ and ‘rather specific’ general noun phrases on the one hand and ‘less specific’ and ‘least specific’ general noun phrases on the other hand. Again, an interesting observation was made: the distribution of the four categories of specification is rather even in the judgment corpus. In the manifesto 228

corpus, debate corpus, and conversation corpus, however, there is a great difference in the distribution of general nouns across the ‘most specific’ and ‘rather specific’ categories on the one hand and the ‘less specific’ and ‘least specific’ categories on the other hand. Similar to what has been stated above (cf. Table 7.34), this supports that linkage is a crucial parameter in the use of general noun phrases, in particular in the three mentioned sub-corpora. The results presented in Table 7.35 partly confirm the fifth, sixth and seventh of my assumptions (cf. Section 4.4). Concerning assumption five, my findings do confirm that the frequency of simple general noun phrases without endophoric reference (‘least specific’) is lowest in legal language and high in the manifestos. However, they do not confirm that the debate corpus, which together with the manifesto corpus represents the political domain, consequently has a similar frequency of ‘least specific’ general noun phrases. In fact, the debate corpus is second lowest with regard to this category. The results also do not confirm that ‘least specific’ general noun phrases are most frequent in the conversation corpus. Concerning assumption six, my findings do confirm that judgments (representing legal language) have the highest frequency of ‘most specific’ general noun phrases and that the two political corpora have a low frequency of general noun phrases from this category. However, the findings do not confirm that the conversation corpus has the lowest frequency of ‘most specific’ general noun phrases. In fact, theses general noun phrases are least frequent in the manifesto corpus. Concerning the seventh of my assumptions, the results presented in Table 7.35 do not confirm that ‘least specific’ general noun phrases are more frequent in spoken language than in written language and that ‘most specific’ general noun phrases are more frequent in written language than in spoken language. We might already conclude from this, that the use of general noun phrases, more particularly their degree of specification, is not medium-dependent. When making statements about the degree of specification of general noun phrases, we have so far only taken into account the ‘most specific’ and the ‘least specific’ category. If we want to make more reliable statements about the degree of specification of general noun phrases, all four categories (‘most specific’, ‘rather specific’, ‘less specific’, and ‘least specific’) must be considered. For this purpose, the frequencies presented in Table 7.35 were calculated as follows: first, the frequencies of general noun phrases in the ‘most specific’ category and in the ‘rather specific’ category were added, in order to get an average frequency of ‘specific’ general noun phrases. Second, the frequencies of general noun phrases in the ‘less specific’ category and in the ‘least specific’ category were added, 229

in order to get an average frequency of ‘unspecific’ general noun phrases.9 The four sub-corpora were positioned along the scale of specification accordingly: sub-corpora with a low frequency of ‘most specific’ and ‘rather specific’ general noun phrases and a high frequency of ‘less specific’ and ‘least specific’ general noun phrases are at the very general end of the scale. Sub-corpora with a high frequency of ‘most specific’ and ‘rather specific’ general noun phrases and a low frequency of ‘less specific’ and ‘least specific’ general noun phrases are at the very specific end of the scale. Figure 7.18 illustrates this: Figure 7.18: Positioning of the four sub-corpora along the scale of specification general high frequency of ‘less specific’ and ‘least specific’ general noun phrases

low frequency of ‘most specific’ and ‘rather specific’ general noun phrases

96.4

3.6

Manifesto corpus

16.51

Conversation corpus

77.8

22.2

Debate corpus

42.31

57.04

Judgment corpus

83.49

low frequency of ‘less specific’ and ‘least specific’ general noun phrases

high frequency of ‘most specific’ and ‘more specific’ general noun phrases

specific

9

The principle underlying this calculation is the assumption that for determining the degree of specification of general noun phrases, linkage is a more crucial parameter than modification.

230

Figure 7.18 shows that due to the very low frequency of ‘most specific’ and ‘rather specific’ general noun phrases and the very high frequency of ‘less specific’ and ‘least specific’ general noun phrases, the manifesto corpus is positioned at the very general end of the scale. In contrast to the assumption made in Section 4.4, the conversation corpus follows the manifesto corpus on the scale of specification. With a relatively low frequency of ‘most specific’ and ‘rather specific’ general noun phrases and a relatively high frequency of ‘less specific’ and ‘least specific’ general noun phrases, the conversation corpus is still positioned towards the general end of the scale. The debate corpus follows closely, with a similarly low frequency of ‘most specific’ and ‘rather specific’ general noun phrases and a similarly high frequency of ‘less specific’ and ‘least specific’ general noun phrases. The position of the debate corpus is relatively central, with a tendency towards the general end of the scale. The judgment corpus is in fact clearly positioned at the very specific end of the scale, with the highest frequency of ‘most specific’ and ‘rather specific’ general noun phrases and the lowest frequency of ‘less specific’ and ‘least specific’ general noun phrases. The arrangement of corpus texts on the scale of specification illustrated in Figure 7.18 deviates from the assumed arrangement illustrated in Section 4.4. The first conclusion that can be drawn from the arrangement of the sub-corpora on the scale of specification presented in Figure 7.18 is that the degree of specification of general noun phrases is indeed not medium-dependent, at least not in general. The similar position of the two spoken corpora might suggest a medium-dependency, but the opposite positions of the two written corpora, the manifesto corpus and the judgment corpus, contradicts a general mediumdependency. Nevertheless, we can suggest that general noun phrases are used similarly at least in spoken language. This raises new questions, which can only be answered on the basis of further research in a larger corpus of spoken language. Another conclusion, which can be drawn from the results presented in Figure  7.18, is that the arrangement of the four sub-corpora along the scale of specification is not influenced by the domain which the four sub-corpora represent. This can be explained by the fact that the two corpora from the political domain, the manifesto corpus and the debate corpus, have different positions on the scale. Therefore, we can conclude that the results presented in Figure 7.18 suggest a genre-specific use of general noun phrases. What seems to be a crucial factor here is the purpose of the text. Judgments, which need to be precise and clear, use more ‘most specific’ and ‘rather specific’ general noun phrases than ‘less specific’ and ‘least specific’ general noun phrases. The high frequency of ‘least specific’ general noun phrases in the manifesto 231

corpus can be explained with the need to make general election promises and generically refer to people and things. Simple general noun phrases without endophoric reference are a useful means for that purpose. The use of general noun phrases in the debate corpus is, in this respect, similar. Yet, these non-referring or generic general noun phrases are more often modified, which makes them a bit more specific than those in the manifesto corpus. Parliamentary discussions are often very lively, and ministers openly express their opinions, and their approval or disapproval. This is often done with the use of general noun phrase modifiers. In the conversation corpus, the frequent use of ‘less specific’ and ‘least specific’ general noun phrases can be explained by the fact that the shared situations in conversations do not require a degree of elaboration typical of ‘most specific’ and ‘rather specific’ general noun phrases. Conversation partners can nevertheless interpret implications such as those expressed by ‘less specific’ and ‘least specific’ general noun phrases. The third step of the qualitative analysis in the present study involved determining the distribution of general noun phrase heads across the four categories of specification. These results are compared in Table 7.36: Table 7.36: Comparison of the most frequent general noun phrase heads in the four categories of specification across the four sub-corpora

Sub-corpora Judgment corpus

Most frequent Most frequent Most frequent ‘most specific’ ‘rather specific’ ‘less specific’ general noun general noun general noun phrase heads phrase heads phrase heads question question person

Most frequent ‘least specific’ general noun phrase heads person

Manifesto corpus

people

child

people

people

Debate corpus

question, idea

matter

people

people

Conversation corpus

stuff

stuff

thing

thing

Table 7.36 shows that in the judgment corpus, the general noun phrase head question was most frequent in the ‘most specific’ and ‘rather specific’ categories, whereas person was most frequent in the ‘less specific’ and ‘least specific’ categories. In the manifesto corpus, people was most frequent in the ‘most specific’, the ‘less specific’ and the ‘least specific’ categories. Child was most frequent in the ‘rather specific’ category. In the debate corpus, question and idea were most frequent in the ‘most specific’ category. Matter was the most frequent general noun phrase head in the ‘less specific’ and ‘least specific’ categories. In the conversation corpus, the most frequent general noun phrase head in the ‘most specific’ 232

and ‘rather specific’ categories was stuff, and thing was most frequent in the ‘less specific’ and ‘least specific’ categories. Table 7.36 illustrates that the distribution of general noun phrase heads across the four categories of specification reveals parallels. In two of the four sub-corpora, the judgment corpus and the conversation corpus, the typical general noun phrase heads coincide in the ‘most specific’ category and the ‘rather specific’ category as well as in the ‘less specific’ category and the ‘least specific’ category. In the manifesto corpus and the debate corpus, this parallel can be observed in the ‘less specific’ and ‘least specific’ category. This again supports the assumption that the parameter linkage is a crucial parameter for the use of general noun phrases. In Section 7.1.5 (see the discussion in reference to Figure 7.7), it was explained that the frequent use of certain general noun phrase heads in certain sub-corpora points to a characterising function that these general noun phrase heads have in the respective sub-corpus. We can therefore assume that the frequent use of certain general noun phrase heads is genre-dependent. The results presented in Table 7.36 also support this assumption. Additionally, we can see that there are typical domain-specific general noun phrase heads in the manifesto corpus and the debate corpus. In a fourth step of the qualitative analysis, typical realisations of ‘most specific’, ‘rather specific’, ‘less specific’ and ‘least specific’ general noun phrases were analysed. These results are illustrated in Table 7.37: Table 7.37: Comparison of most frequent realisations of general noun phrases in the four categories of specification across the four sub-corpora

Sub-corpora

Judgment corpus

Most frequent realisation of ‘most specific’ general noun phrases question + appositive interrogative clause, cataphoric reference to 2nd appositive unit

Most frequent realisation of ‘rather specific’ general noun phrases

Most frequent realisation of ‘less specific’ general noun phrases

Most frequent realisation of ‘least specific’ general noun phrases

question, anaphoric person + relative encapsulation clause, generic or cataphoric reference reference to subj. complement

person, generic reference

233

Most frequent realisation of ‘most specific’ general noun phrases

Most frequent realisation of Sub-corpora ‘rather specific’ general noun phrases child, H/H’s idea + appositive anaphoric that-clause, reference or Manifesto cataphoric anaphoric corpus reference to 2nd reference to appositive unit subject Debate corpus

adjective + man/ child, H/H’s anaphoric reference

adjective + stuff, Conversation H/H’s anaphoric corpus reference

Most frequent realisation of ‘less specific’ general noun phrases

Most frequent realisation of ‘least specific’ general noun phrases

adjective + people, generic reference

people, generic reference

matter, H/H’s anaphoric reference

adjective + people, generic reference

people, generic reference

stuff, H/H’s anaphoric reference

adjective + thing, non-referring

thing, nonreferring

Table 7.37 shows the most frequent realisations of general noun phrases in the four categories of specification across the four sub-corpora. The categories, which were most frequent in the four sub-corpora are given in dashed cells. The realisations of general noun phrases in these cells are therefore considered to be characteristic of the respective sub-corpus. In the judgment corpus, the general noun phrase head question, which was postmodified by an appositive interrogative clause and referred cataphorically to this clause, was the most frequent realisation of ‘most specific’ general noun phrases. Thus, patterns like the question which… or the question whether… were typically found in the judgment corpus. Such patterns were used to refer to the underlying question of the case repeatedly and, in the case of the pattern question + ‘appositive interrogative clause’ + ‘cataphoric reference to 2nd appositive unit’, this could be done in a concise way. In the manifesto corpus, the most frequent realisation of the ‘least specific’ category was the non-modified general noun phrase head people, which was used to make generic reference. This pattern was used frequently because political manifestos address people in general – the British people in the case of the investigated manifestos. Manifestos advertise political views and aims that affect people. This explains why the use of generic general noun phrases such as in the following example occurred frequently in the manifestos:

234

(69) Labour believes we should […] reform our public services to put people in control. [LM-1-14]

Through the generic use of people in example (69), Labour addresses every possible reader of the manifesto, thereby achieving that everyone who reads it feels integrated into the political actions. Labour’s plan to put the British people in control and give them power sounds very promising. However, in the example above, it is not specified who will really be put in control. Note that the frequent generic and also non-referring general noun phrases in the manifesto corpus, also with heads other than people, are very often used in similar ways as in example (69). The following example illustrates the use of a simple non-referring general noun phrase: (70) […] the problems confronting Britain have escalated, and escalated fast. So our ideas are ambitious and radical as well as modern. [CM-18-4]

The use of the general noun phrase in example (70) is rather vague because it is not specified, through neither modification nor endophoric reference, what ideas refers to. However, such a general statement sounds promising and impresses the potential voter. Similar to the results of the manifesto corpus, the typical pattern in the debate corpus was the general noun phrase head people, which was used to make generic reference. However, in the debate corpus, the general noun phrase people was typically premodified by an adjective. The function of such generic noun phrases corresponds to that in the manifesto corpus, namely the expression of general promises, but, as mentioned above, premodifiers might additionally be used to express opinions, approval or disapproval of ministers. The typical pattern found in the conversation corpus was the general noun phrase head thing, which was not modified and which was neither linked endophorically, nor used to make generic reference, i.e. it was non-referring. The fact that this pattern was most frequent can be explained by the restricted style of referencing typically used in conversations. This style of referencing relies heavily on shared situations, so that thing is typically used when referring to entities that are present in the situation of utterance. Thus, ‘least specific’ general noun phrases in conversations often express ‘conventional vagueness’, a type of vagueness that is defined as a natural language phenomenon (cf. Section 3.4.2.2). This type of vagueness is typical of spoken face-to-face interaction where the speakers share the situation and have a common ground. Channell (cf. 1994: 180) states that in unspecified and vague utterances, the speaker very often does not know or has forgotten the actual name of an object and thus fills the lexical gap with a general noun phrase such as thing in order to be able to convey meaning all the 235

same. Due to the shared situation in such conversations, the vague expressions can be decoded easily and the utterance is nevertheless understandable. A very important aspect that the results in Table 7.37 reveal is that Halliday/ Hasan’s anaphoric or cataphoric reference was found in the corpus data in only five of the 16 typical patterns of general noun phrases illustrated in Table 7.37 (cells with a red frame). This takes up my critical remarks in Section 2.3 and confirms that general noun phrases do not at all exclusively occur in the anaphoric function illustrated in the perfect textbook examples given by Halliday/Hasan (1976). Quite the contrary, it has been demonstrated that general noun phrases occur in many different patterns and fulfil functions that have so far not been associated systematically with general noun phrases as described by Halliday/ Hasan (1976): (i) (ii) (iii) (iv) (v) (vi)

236

General noun phrases are neutral shells, which encapsulate complex information and reify it for subsequent discourse. General noun phrases encapsulate subsequent information and then function as discourse markers. General noun phrases can be used as the first of two appositive units, which cataphorically refers to the second appositive unit (very often a clause). General noun phrases can be used as the subject in a clause, referring cataphorically to the subject complement. General noun phrases can be used to make generic reference. General noun phrases can be used as completely unspecific items without any kind of endophoric or generic reference and function as empty phrases (especially in political language) or as fillers (especially in conversation).

8. Summary and conclusions Abstract: Chapter 8 summarises the most important findings of a comprehensive study of general noun phrases based on a corpus of four different genres. The results imply a genre-specific and also purposeful use of general noun phrases. This emphasises the role of these items, which so often have been neglected in descriptions of English.

The present study has examined general noun phrases in naturally occurring language data, more specifically, in a corpus of written and spoken texts from four different genres, namely Supreme Court judgments, political manifestos, parliamentary debates and general conversations. These genres represent three different domains, namely legal language, political language and conversation. The starting point for the present study was Halliday/Hasan’s text-linguistic approach to ‘general nouns’, which exclusively focuses on the cohesive function of these items. According to Halliday/Hasan (1976: 274–275), ‘general nouns’ are nouns with only very general meaning, which are used anaphorically or cataphorically in order to create cohesion. With this status, ‘general nouns’ share some characteristics with pronouns. I have critically remarked in Section 2.3 that this definition cannot sufficiently describe the form and function of ‘general nouns’ in naturally occurring language data. Concerning form, it has been suggested that ‘general nouns’ occur in natural language data as heads of noun phrases of varying complexity and should therefore be referred to as ‘general noun phrases’ or ‘general noun phrase heads’. This refutes the assumption that they are necessarily general in meaning. Furthermore, it is assumed that general noun phrases are not exclusively anaphoric but can fulfil other functions. These critical remarks were the starting point for a comprehensive description of general noun phrases, which combines grammatical, text-linguistic, pragmatic and cognitive aspects. From a grammatical perspective, all of the 18 general noun phrase heads were defined as common nouns. Sixteen of them were defined as count nouns and two of them as non-count nouns. Ten of the count nouns were defined as concrete count nouns; these were people, person, man, woman, child, boy, girl, creature, thing and object. Six of the count nouns were defined as abstract count nouns; these were affair, matter, move, place, question and idea. One of the non-count nouns was defined as concrete, namely stuff, and the other one was defined as abstract, namely business. On a scale that ranged from specific to general, the general noun phrase heads investigated in the present study were positioned at a point where the classes of concrete and abstract count nouns and concrete 237

and abstract non-count nouns overlapped. This point was located towards the general end of the scale. A text-linguistic description of general noun phrases included aspects of communication in terms of the constitutive and regulative features of textual communication1 defined by de Beaugrande/Dressler (1981)2 and the distinction of two styles of referencing: the restricted and the elaborated style of referencing described by Bernstein (1971). It was shown that general noun phrases can be used intentionally by the encoder to follow certain aims, for example to be vague and unspecific about something. One means of achieving this is to influence the informational load of an utterance either through the use of a general noun phrase, which is so unspecific that it conveys only little information, or through remote reference, which means that too much text lies in between the general noun phrase and the reference item so that the decoder cannot establish the reference relation, or through a lack of any reference relation at all. Since particularly low informativity is likely to be disturbing (cf. de Beaugrande/Dressler 1981: 9), general noun phrases have shown to influence the decoder’s attitude towards the acceptability of an utterance. In terms of the regulative features of textual communication defined by de Beaugrande/Dressler (1981), it was shown that the use of a general noun phrase as a referring item can reduce the text-producer’s effort and correspondingly raise the text’s efficiency because a general noun phrase can be used to encapsulate longer stretches of text. There is, however, a ‘trade-off ’ between propositional density and clarity of thought (cf. de Beaugrande/Dressler 1981: 64): if the decoder is not able to identify or locate the reference item, for example, because the referring item passes too much text, then the granted effectiveness of the text is lost again. This aspect ties in with the aforementioned intentions of an encoder. In line with the decoder’s attitude towards the appropriateness of an utterance is the distinction of the restricted and the elaborated style of referencing. According to Esser (cf. 2009: 36), the restricted style of referencing heavily relies on exophoric reference, which is manifested in the frequent use of pronouns and 1

2

De Beaugrande/Dressler (1981) originally speak of regulative and constitutive principles which must necessarily be fulfilled. Today, this view seems too rigid, however, it is generally recognised that a text has a communicative function which is determined by the encoder’s intention and the decoder’s expectation. Therefore, I refer to the constitutive and regulative principles as features of textual communication (cf. section 3.3.1). The aspect of communicative functions of texts and the text producer’s intentions really describe phenomena which go back to the works of J.L. Austin (1962) and J.R. Searle (1969).

238

deictic expressions. The elaborated style of referencing, on the contrary, is manifested in the use of lexically filled noun phrases. Concerning the different styles of referencing, the use of general noun phrases is relevant particularly in terms of a restricted style of referencing typically found in conversations. General noun phrases such as that thing and one of the things, which point to the situation of utterance, typically occurred in conversations where the shared situation of the speakers allowed for such expressions. Closely related to textual features of communication are pragmatic features of communication, such as the ‘Cooperative Principle’ formulated by Grice (1975). According to Grice (cf. Grice 1975: 45–47), the conversational partners must cooperate in that they both accept one another to be understood in a particular way. Grice sets up four maxims, which require the conversation to be informative, truthful, relevant and explicit. Grice (cf. 1975: 49) states that very often one or more of the maxims are violated. If, for example, completely unspecific general noun phrases are used in a text without reference to the preceding or following text, the general noun phrases in fact convey very little meaning. The encoder violates the maxim of ‘quantity’ (cf. Grice 1975: 45), and, as a result, the decoder might not be able to identify the referent and, in the end, might not be able to understand an utterance altogether. If, for example, the encoder is faced with the dilemma of wanting to tell the truth, he or she might solve this by holding back information. This is particularly relevant in the context of political issues, and general noun phrases, as the present study has shown, are a useful means for this purpose. As an outcome of the description of general noun phrases under text-linguistic and pragmatic aspects, it was shown that unspecific general noun phrases can be used to express vagueness. In Section 3.4.2.2, two types of vagueness were distinguished: (i) ‘conventional’ vagueness, which occurs as a natural language phenomenon and is easily decodable, and (ii) ‘rhetorical’ vagueness, which is used intentionally in order to manipulate and mislead the decoder and which is hard or even impossible to decode. The first type of vagueness, conventional vagueness, was expressed, for example, by unspecific general noun phrases such as thing and stuff typically found in conversations where the speakers share the situation of utterance. As explained above in the context of the restricted style of referencing, this type of vagueness is appropriate in conversations and can easily be decoded. Rhetorical vagueness, by contrast, is characterised by an amount of vagueness that is no longer seen as appropriate and therefore has a manipulating and misleading effect on the decoder. Rhetorical vagueness was expressed, for example, by using a very unspecific general noun phrase where a more specific term would be necessary, or by an unspecific general noun phrase, which remotely referred to another 239

item for its interpretation, or which did not refer to any other item at all. Especially in political contexts, it was shown that the encoder deliberately uses rhetorically vague general noun phrases to withhold information, refrain from going into (unpleasant) detail and to weaken down problems. This way, politicians protect themselves from any sort of binding commitment and face threat. Moving to a cognitive approach to general noun phrases, de-contextualised general noun phrases can be viewed as representing broad cognitive categories which, according to whether they are specified through the co-text and context, move at different levels of abstraction. A de-contextualised general noun phrase such as matter encapsulates different meanings; in cognitive terms, the general noun phrase constitutes a very broad category with many category members. The different meanings of matter can only be specified through reference to its co-text and context in each single utterance. Resulting from this process, it was suggested to describe general noun phrases as semantically empty containers that are filled with content to different degrees through reference to the co-text and context. As an outcome of the description of general noun phrases under grammatical, text-linguistc, pragmatic and cognitive aspects, a major role was granted to the parameters modification and linkage. These parameters were integrated into a functional matrix, which served as the major tool to determine the degree of specification of a general noun phrase. With this innovative framework, the present study has offered a systematic method to describe the degree of specification or generalisation of a word. Figure 8.1: The functional matrix illustrating the generalisation and specification of general noun phrases

‘less specific’

240

– linkage

+ modification

‘rather specific’

‘most specific’

– modification

+ modification

‘least specific’

‘less specific’

+ linkage

specification

‘rather specific’

– modification

– linkage

‘most specific’

+ linkage

generalisation

‘least specific’

The functional matrix classifies general noun phrases into ‘most specific’, ‘rather specific’, ‘less specific’ and ‘least specific’ general noun phrases. According to the presence and absence of the parameters modification and linkage, the degree of specification of general noun phrase increases or decreases. This is illustrated by the arrows, which show a generalisation or a specification. See the following examples for an illustration of the four categories of specification: (1) Only Liberal Democrats have the big ideas for fundamental, structural changes in the way our country works to make it fair. Only Liberal Democrats will shake up the tax system to put 700 back in the pockets of tens of millions of low and middle-income families, […]. Only Liberal Democrats will break up the banks and start Britain building things again, creating a sustainable economy that no longer threatens our planet’s future. Only Liberal Democrats will invest in our schools to give every child, no matter their background, a fair start in life. And only Liberal Democrats will sort out our rotten political system once and for all. [LDM-18-1] (2) Living in a family on a low income, having special educational needs or disabilities, and being in care all remain strongly linked to children1 failing to achieve. We are determined to narrow the gap between these children2 and their peers… [LM-5-27] (3) […] the plans give councils a strong financial inventive to drive economic growth, as well as providing protections for places in need of additional support. [D8-16-1] (4) This manifesto makes the case that there should be no return to business as usual. People have suffered too much with their jobs, livehoods and confidence to allow a return to the same old ways. [LM-1-10]

Examples (1)-(4) show a decline in the specification of general noun phrases. In example (1), the general noun phrase the big ideas for fundamental, structural changes in the way our country works to make it fair is ‘most specific’ because it is specified through modification and through cataphoric reference to the underlined part of the text. In example (2), the general noun phrase these children2 is ‘rather specific’ because it is not specified through modification but it is specified through anaphoric reference to the underlined noun phrase. In example (3), the general noun phrase places in need of additional support is ‘less specific’ because it is specified through modification but not through reference. In example (4), the general noun phrase people is ‘least specific’ because it is specified neither through modification nor through reference. The corpus of the present study was compiled from different domains and genres. See the following table for an overview of the corpus data:

241

Table 8.1: Compilation of the corpus data Corpus compilation Domains

Genres

Sub-corpora

Total no. of words

Legal language

Supreme Court judgments

Judgment corpus

78,473

Political manifestos

Manifesto corpus

77,048

Parliamentary debates

Debate corpus

Political language Conversation

General conversation Conversation corpus

80,160 70,741

Medium written spoken

The compilation of the corpus data was based on the assumptions that (i) legal language is very precise and specific, (ii) political language is rather unspecific in the sense of a rhetorical type of vagueness, and (iii) conversation is unspecific in the sense of a conventional type of vagueness. This way, the corpus of the present study provided a very suitable database for a study of general noun phrases and how they are (consciously or subconsciously) used in different genres and for different purposes. In search for insights into the use of general noun phrases, the results of a detailed quantitative and qualitative corpus analysis were presented and discussed. The findings of a qualitative analysis (Section 7.1) showed that the overall frequency of general noun phrase heads was highest in the manifesto corpus and lowest in the judgment corpus. The debate corpus had the second highest frequency of general noun phrase heads, and the conversation had the third highest (or second lowest) overall frequency of general noun phrase heads. This result reveals a difference in the overall frequency of general noun phrase heads across genres. The role that the different media or domains of the corpus data played was only marginal. See the following figure, which illustrates the results of the quantitative analysis in terms of a positioning of the four sub-corpora along the scale of specification:

242

Figure 8.2: The four sub-corpora on the scale of specification (depending on frequency)

general high frequency of general noun phrase heads

829.42

Judgment corpus

Manifesto corpus

567.84

Debate corpus

299.77

Conversation corpus

231.85

low frequency of general noun phrase heads specific

Section 7.2 presented the results of the qualitative analysis of general noun phrases in the four sub-corpora. In a first step, it was shown that in the judgment corpus and in the debate corpus, complex general noun phrase heads were more frequent than simple general noun phrase heads. By contrast, in the manifesto corpus and in the conversation corpus, simple general noun phrase heads were more frequent than complex ones. Furthermore, it was shown that in the judgment corpus, general noun phrases that were linked to their co-text through endophoric reference were more frequent than general noun phrases that were not linked. By contrast, in the manifesto corpus, the debate corpus and the conversation corpus, general noun phrases that were not linked to their co-text through endophoric reference were significantly more frequent than general noun phrases that were linked. Furthermore, the functional matrix for the classification of general noun phrases into four categories of specification was applied. It was shown that in the judgment corpus, ‘most specific’ general noun phrases (+ modification/+ linkage) were most frequent and ‘least specific’ general noun phrases (- modification/- linkage) 243

were least frequent. By contrast, in the manifesto corpus, ‘least specific’ general noun phrases were most frequent and ‘rather specific’ general noun phrases (- modification/+ linkage) were least frequent. In the debate corpus, ‘less specific’ general noun phrases (+ modification/- linkage) were most frequent and ‘most specific’ general noun phrases were least frequent. In the conversation corpus, ‘least specific’ general noun phrases were most frequent and ‘most specific’ general noun phrases were least frequent. For an overview of these results, see the following figures:

Judgment corpus

+ linkage

Figure 8.3: The categories of specification which were most frequent in the four sub-corpora

Debate corpus

– linkage

+ modification

– modification Manifesto corpus Conversation corpus

Debate corpus Conversation corpus

+ linkage

Figure 8.4: The categories of specification which were least frequent in the four sub-corpora

– linkage

+ modification

244

Manifesto corpus

– modification

Judgment corpus

For an illustration of these results on the scale of specification, an average frequency of ‘most specific’ and ‘rather specific’ general noun phrases and an average frequency of ‘less specific’ and ‘least specific’ general noun phrases was calculated. This way, the results of all the four categories of specification for each corpus were taken into account. See the following figure for an overview: Figure 8.5: The four sub-corpora on the scale of specification (depending on modification and linkage)

general high frequency of ‘less specific’ and ‘leastt specific’’ general noun phrases

low frequency of ‘most h specific’ ifi ’ specific’’ and ‘rather general noun phrases

96.4

3.6

Manifesto corpus

16.51

Conversation corpus

77. 77 8

22 2

Debate corpus

42.31

57.04

Judgment corpus

83.49

low frequency of ‘less specific’ and ‘least specific’ general noun phrases

high frequency of ‘most specific’ and ‘more specific’ general noun phrases

specific

245

The arrangement of the four sub-corpora on the scale of specification is genredependent. This can be explained with the purpose of each sub-corpus (i.e. genre): for example, Judgments, which need to be precise and clear, have many ‘most specific’ and ‘rather specific’ general noun phrases and few ‘less specific’ and ‘least specific’ general noun phrases. By contrast, political manifestos, which are used to make general election promises, have many ‘less specific’ and ‘least specific’ general noun phrases and only few ‘most specific’ and ‘rather specific’ general noun phrases. The qualitative analysis revealed typical realisations of general noun phrases across the four categories of specification. In sum, there were 16 typical patterns of general noun phrase use across the four genres of the corpus. A very important observation was that in only five of the 16 typical patterns of general noun phrases, Halliday/Hasan’s anaphoric or cataphoric reference was found in the corpus data. This outcome has convincingly demonstrated that general noun phrases are not exclusively used anaphorically or cataphorically. Instead, it was shown that these items are much more flexible in (form) and function. Based on the findings of the corpus analysis, we can claim that there are six major functions of general noun phrases. Neither of these functions have so far been associated systematically with general noun phrases as described by Halliday/ Hasan (1976): (i) (ii) (iii) (iv) (v) (vi)

General noun phrases are neutral shells, which encapsulate complex information and reify it for subsequent discourse. General noun phrases encapsulate subsequent information and then function as discourse markers. General noun phrases can be used as the first of two appositive units, which cataphorically refers to the second appositive unit (very often a clause). General noun phrases can be used as the subject in a clause, referring cataphorically to the subject complement. General noun phrases can be used to make generic reference. General noun phrases can be used as completely unspecific items without any kind of endophoric or generic reference and function as empty phrases (especially in political language) or as fillers (especially in conversation).

The present study has methodological implications. While genre-specific patterns are usually analysed in a macro-linguistic framework, the present study has integrated a micro- and macro-linguistic perspective in order to systematically study the use of a specific class of words – here general noun phrases – in a corpus of different spoken and written genres. A combination of approaches lead to the development of an innovative analytical tool (the functional matrix 246

described in Section 4.2), which convincingly illustrated that the specification of general noun phrases can only be analysed systematically by integrating grammatical aspects (parameter modification) and text-linguistic and semantic apsects (parameter linkage). Most importantly, the results of the qualitative analysis have implications for the text-strategic potential of general noun phrases in different genres. For example, it was shown that general noun phrases in judgments have helped to represent complex issues and, thus, have a major influence on discourse organisation (functions (i)-(iv)). This is especially interesting in the light of an ongoing debate in the legal field about the drafting of texts. A major concern of scholars in the so-called ‘Plain Language Movement’ is to make legal language intelligible for ordinary people, in particular with regard to lexical and syntactic complexity (cf. Trosborg 1997: 141). The encapsulation of complex issues with a common noun such as the general noun phrase head matter or question, may be said to contribute to the lexical simplification of legal language in judgments. Moreover, the use of general noun phrase heads in syntactic structures such as those described in functions (iii) and (iv) may also play a role in the light of the ‘Plain Language Movement’. These issues offer opportunities for further research. The text-strategical potential of general noun phrases is especially relevant with regard to the purposeful use of language in the political context. We have seen that ‘least specific’ general noun phrases are typically used in political manifestos for generic reference (function (v)) and as completely unspecific items or empty phrases (function (vi)). In the former, general noun phrases are a useful tool to construct (national) collectivity. In the latter function, general noun phrases are a valuable tool to keep references deliberately vague, to make general promises without too much commitment, and, to avoid responsibility. This way, general noun phrases contribute to the purposeful use of language in political discourse. As one outcome of the analysis of general noun phrases in the present study, it was stated that the overall frequency, the distribution and the degree of specification of general noun phrases is genre-dependent. Therefore, it would be interesting for further research to examine general noun phrases in a greater variety of genres. It would also be interesting for such research to extend the list of general noun phrase heads by Halliday/Hasan (1976: 274). Concerning the genre-based implications the present study offered, it would be interesting to validate the findings in a larger corpus of legal language or political language only.

247

References Anderson, R.C., J.W. Pichert, E.T. Goetz, D.L. Schaller, K.V. Stevens & S.R. Trollip (1976): “Instantiation of general terms”, Journal of Verbal Learning and Verbal Behaviour 15, 667–669. Austin, J.L. (1962): How to Do Things with Words. Cambridge: Harvard University Press. Berlin, B. & P. Kay (1969): Basic Color Terms. Berkeley, California: Center for the Study of Language and Information. Berlin, B., D. Breedlove & P. Raven (1973): “General principles of classification and nomenclature in folk biology”, American Anthropologist 75, 214–242. Bernstein, B. (1971): Class, Codes and Control. Theoretical Studies towards a Sociology of Language 1. London: Routledge & Kegan Paul. Biber, D., S. Johansson, G. Leech, S. Conrad & E. Finnegan (1999): Longman Grammar of Spoken and Written English. Harlow: Longman. Bolinger, D. (1977): Pronouns and Repeated Nouns. Indiana: Indiana University Linguistics Club. Carter, R. & M. McCarthy (2006): Cambridge Grammar of English. A Comprehensive Guide. Spoken and Written English Grammar and Usage. Cambridge: Cambridge University Press. Carter, R. & M. McCarthy (2007): Exploring Spoken English. Cambridge: Cambridge University Press. Chafe, W. (1982): “Integration and involvement in speaking, writing, and oral literature”, Spoken and Written Language: Exploring Orality and Literacy, ed. D. Tannen, Norwood: Alex Publishing Corporation, 35–53. Channell, J. (1994): Vague Language. Oxford: Oxford University Press. Charnock, R. (2010): “Traces of orality in Common Law judgments”. Researching Language and the Law: Textual Features and Translation Issues, eds. D.S. Giannoni & C. Frade. Bern: Peter Lang, 113–134. Chilton, P. (2004): Analysing Political Discourse. Theory and Practice. London: Routledge. Conte, M.-E. (1996): “Anaphoric encapsulation”, Belgian Journal of Linguistics: Coherence and Anaphora 10, 1–10. Cruse, D.A. (1986): Lexical Semantics. Cambridge: Cambridge University Press. De Beaugrande, R. & W. Dressler (1981): Introduction to Text Linguistics. London: Longman. 249

Dickson, B. (2013): Human Rights and the United Kingdom Supreme Court. Oxford: Oxford University Press. Driver, S. (2011): Understanding British Party Politics. Cambridge: Polity Press. Esser, J. (1984): Untersuchungen zum Gesprochenen Englisch: Ein Beitrag zur Strukturellen Pragmatik. Tübingen: Gunter Narr. Esser, J. (2006): Presentation in Language: Rethinking Speech and Writing. Tübingen: Gunter Narr. Esser, J. (2009): Introduction to English Text-Linguistics. Frankfurt a.M.: Peter Lang. Ehlich, K. (1983): “Deixis und Anapher”, Essays on Deixis, ed. G. Rauh. Tübingen: Gunter Narr, 79–98. Flowerdew, J. (2003a): “Register-specificity of signalling nouns in discourse”, Corpus Analysis: Language Structure and Language Use, eds. C. Meyer & P. Leistyna. Amsterdam: Rodopi Publishers, 35–46. Flowerdew, J. (2003b): “Signalling nouns in discourse”, English for Specific Purposes 22, 329–346. Gibbons, J. (2004): “Taking legal language seriously”, Language in the Law, eds. J. Gibbons, V. Prakasam, K.V. Tirumalesh & H. Nagaraja. New Delhi: Orient Longman, 1–16. Goss, B. (1972): “The effect of sentence context on associations to ambiguous, vague and clear nouns”, Speech Monographs 39, 286–289. Grice, H.P. (1975): “Logic and conversation”, Speech Acts, eds. P. Cole & J.L. Morgan. New York: Academic Press, 41–85. Gruber, H. (1993): “Political language and textual vagueness”. Pragmatics 3, 1–28. Håkansson, N. (1997): “Election rhetoric of political parties. An appraisal and a proposal”, The Public 4, 81–95. Halliday, M.A.K. (1985): Spoken and Written Language. Geelong, Victoria: Deakin University Press. Halliday, M.A.K. (1987): “Spoken and written modes of meaning”. Comprehending Oral and Written Language, ed. R. Horowitz & S. Samuels. San Diego: Academic Press, 55–82. Halliday, M.A.K. (2004): An Introduction to Functional Grammar, third edition revised by Christian M.I.M. Matthiessen. London: Arnold. Halliday, M.A.K. & R. Hasan (1976): Cohesion in English. London: Longman. Halliday, M.A.K. & R. Hasan (1989): Language, Context and Text: Aspects of Language in a Social-Semiotic Perspective, second edition. Oxford: Oxford University Press. 250

Halmari, H. (2005): “In search of ‘successful’ political persuasion: A comparison of the styles of Bill Clinton and Ronald Regan”, Persuasion Across Genres. A Linguistic Approach, eds. H. Halmari & T. Virtanen. Amsterdam: John Benjamins, 105–134. Hoey, M. (1991): Patterns of Lexis in Text. Oxford: Oxford University Press. Huddleston, R. & G.K. Pullum (2002): The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Hunston, S. & G. Francis (2000): Pattern Grammar: A Corpus-Driven Approach to the Lexical Grammar of English. Amsterdam: John Benjamins. Ivanič, R. (1991): “Nouns in search of a context: A study of nouns with both open- and closed-system characteristics”, International Review of Applied Linguistics in Language Teaching 2, Berlin: Mouton de Gruyter, 93–114. Jackson, H. (1980): Analyzing English. An Introduction to Descriptive Linguistics. Oxford: Pergamon Press. Jucker, A.H., S.W. Smith & T. Lüdge (2003): “Interactive aspects of vagueness in conversation”, Journal of Pragmatics 35, 1737–1769. Kaid, L.L. (2004): Handbook of Political Communication Research. Mawah, New Jersey: Lawrence Erlbaum Associates. Kavanagh, D. (2000): “Labour Party manifestos 1900–1997”. Labour Party General Election Manifestos, 1900–1997, ed. I. Dale. London: Routledge, 1–8. Kavanagh, D. (2013): The Politics of the Labour Party. London: Routledge. Kay, P. (1977): “Language evolution and speech style”, Sociocultural Dimensions of Language Change, eds. B.G. Blount & M. Sanches. New York: Academic Press, 21–33. Kennedy, G. (1998): An Introduction to Corpus Linguistics. London: Longman. Lashöfer, J. (1992). Zum Stilwandel in richterlichen Entscheidungen. Über stilistische Veränderungen in englischen, französischen und deutschen zivilrechtlichen Urteilen und in Entscheidungen des Gerichtshofs der Europäischen Gemeinschaften. Münster & New York: Waxmann. Leech, G., P. Rayson & A. Wilson (2001): Word Frequencies in Written and Spoken English: Based on the British National Corpus. Harlow: Longman. Levinson, S. (1983): Pragmatics. Cambridge: Cambridge University Press. Lipka, L. (1972): Semantic Structure and Word-Formation. Verb-Particle Constructions in Contemporary English. München: Wilhelm Fink. Lipka, L. (2002): English Lexicology. Lexical Structure, Word Semantics and WordFormation. Tübingen: Gunter Narr. Lyons, J. (1977): Semantics, 2 volumes, Cambridge: Cambridge University Press. 251

Mahlberg, M. (2005): English General Nouns: A Corpus Theoretical Approach. Amsterdam: John Benjamins. Malinowski, B. (1923): “The problem of meaning in primitive languages”, Supplement 1, The Meaning of Meanings, eds. C.K. Ogden & I.A. Richards. London: Keagan and Paul, 451–510. Mattila, H.E.S. (2006): Comparative Legal Linguistics. Aldershot: Ashgate. Mc Anulla, S. (2006): British Politics. A Critical Introduction. London: Bloomsbury Academic. Mc Enery, T., R. Xiao & Y. Tono (2006): Corpus-Based Language Studies. An Advanced Resource Book. New York: Routledge. Meyer, C.F. (1992): Apposition in Contemporary English. Cambridge: Cambridge University Press. Miller, J. & R. Weinert (1998): Spontaneous Spoken Language: Syntax and Discourse. Oxford: Clarendon Press. Moens, M.F. (2007): “Summarizing court decisions”, Information Processing and Management 43, 1748–1764. Myers, G. (1996): “Strategic vagueness in academic writing”, Academic Writing, eds. E. Ventola & A. Mauranen. Amsterdam: John Benjamins, 3–18. Nystrand, M. (1987): “The role of context in written communication”, Comprehending Oral and Written Language, eds. R. Horowitz & J. Samuels. San Diego: Academic Press, 197–214. Obeng, S.G. (1997): “Language and politics: Indirectness in political discourse”, Discourse & Society 8, London: Sage Publications, 49–83. Odgen, C.K. & I.A. Richards (1923): The Meaning of Meaning. A Study of the Influence of Language upon Thought and of the Science of Symbolism. London: Kegan Paul, Trench, Trubner & Co. Partington, A. (1998): Patterns and Meanings. Using Corpora for English Language Research and Teaching. Amsterdam: John Benjamins. Pérez, J.M. & C.R. (2012): “Structure and design of the British Law Report Corpus (BLRC): a legal corpus of judicial decisions from the UK”. Journal of English Studies 10, 131–145. Pridham, F. (2001): The Language of Conversation. London: Routledge. Quirk, R., J. Svartvik, G. Leech & S. Greenbaum (1985): A Comprehensive Grammar of the English Language. London: Longman. Rauh, G. (2002): “Word classes as prototypical categories”, Context and Cognition. Papers in Honour of Wolf-Dietrich Bald’s 60th Birthday, eds. S. Scholz, M. Klages, E. Hantson & U. Römer. München: Langenscheidt-Longman, 259–270. 252

Roberts, C. & B. Street (1997): “Spoken and written language”, The Handbook of Sociolinguistics, ed. F. Coulmas. Oxford: Blackwell, 168–186. Rosch, E. (1973): “Natural categories”. Cognitive Psychology 4, 328–350. Rosch, E. (1975): “Cognitive representations of semantic categories”, Journal of Experimental Psychology: General 104, 192–233. Rosch, E. (1978): “Principles of categorization”, Cognition and Categorization, eds. E. Rosch & B.B. Lloyd. Hillsdale, New Jersey: Lawrence Erlbaum Associates, 27–48. Rosch, E. (2009): “Categorization”, Cognition and Pragmatics, Handbook of Pragmatics Highlights 3, eds. D. Sandra, J.-O. Östman & J. Verschueren. Amsterdam: John Benjamins, 41–52. Rosch, E. & C.B. Mervis (1975): “Family resemblances: Studies in the internal structure of categories”, Cognitive Psychology 7, 573–605. Rosch, E., C.B. Mervis, W.D. Gray, D.M. Johnson & P. Boyes-Braem (1976): “Basic objects in natural categories”, Cognitive Psychology 8, 382–439. Rosch, E. & B.B. Loyd (1978): Cognition and Categorization. Hillsdale, New Jersey: Lawrence Erlbaum Associates. Salkie, R. (1995): Text and Discourse Analysis. London: Routledge. Schäpers, U.K.E. (2009): Nominal versus Clausal Complexity in Spoken and Written English: Theory and Description. Frankfurt: Peter Lang. Schmid, H.J. (1993): Cottage und Co., Idea, Start vs. Begin: Die Kategorisierung als Grundprinzip einer differenzierten Bedeutungsbeschreibung. Tübingen: Max Niemeyer. Schmid, H.J. (2000): English Abstract Nouns as Conceptual Shells: From Corpus to Cognition. Berlin: Mouton de Gruyter. Searle, J.R. (1969): Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Siefring, J. (2004): The Oxford Dictionary of Idioms. Oxford: Oxford University Press. Sinclair, J. (1991): Corpus Concordance Collocation. Oxford: Oxford University Press. Sinclair, J. (2005): Collins Cobuild English Grammar. Glasgow: Harper Collins. Sjöblom, G. (1968): Party Strategies in a Multiparty System. Lund: Studenttliteratur. Tanskanen, S.-K. (2006): Collaborating towards Coherence. Lexical Cohesion in English Discourse. Amsterdam: John Benjamins.

253

Taylor, J. (1995): Linguistic Categorization. Prototypes in Linguistic Theory, second edition. Oxford: Clarendon Press. Tiersma, P.M. (1999): Legal Language. Chicago: The University of Chicago Press. Tognini-Bonelli, E. (2001): Corpus Linguistics at Work. Amsterdam: John Benjamins. Trosborg, A. (1997): Rhetorical Strategies in Legal Language: Discourse Analysis of Statues and Contracts. Tübingen: Gunter Narr. Tsui, A.B.M. (1994): English Conversation. Oxford: Oxford University Press. Ullmann, S. (1962): Semantics. An Introduction to the Science of Meaning. Oxford: Basil Blackwell. Ungerer, F. & H.J. Schmid (1996): An Introduction to Cognitive Linguistics. London: Longman. Van Rooij, R. (2011): “Vagueness and linguistics”, Vagueness: A Guide, ed. G. Ronzitti. Heidelberg: Springer, 123–170. Williams, A. (1998): UK Government & Politics. Oxford: Heinemann. Winter, E.U. (1977): “A clause-relational approach to English texts: A study of some predictive lexical items in written discourse”, Instructional Science 6, Amsterdam: Elsevier, 1–92. Wittgenstein, L. (1953): Philosophische Untersuchungen. Oxford: Basil Blackwell.

Online sources Burnard, L. (2000): Reference Guide for the British National Corpus (World Edition). , last access 11th December 2014. Cambridge Dictionaries Online, , last access 18th December 2014. Conservative Manifesto. Invitation to Join the Government of Britain. , last access 30th January 2013. Liberal Democrat Manifesto. Change that works for you. Building a fairer Britain. , last access 30th January 2013. Löbner, S. (1998): “Definite Associative Anaphora”, unpublished manuscript, Düsseldorf, , last access 11th December 2014. Longman Dictionary of Contemporary English Online, , last access 18th December 2014. 254

Oxford English Dictionaries Online, , last access 18th December 2014. Parliamentary Debates, “Debates” , last access 17th December 2014. The Supreme Court Annual Reports and Accounts 2009–2010. “Establishment of the court”, London: The Stationary Office, , last access 18th December 2014. The Supreme Court Annual Reports and Accounts 2012–2013. “Serving the UK and beyond: Jurisdiction and casework”, London: The Stationary Office, , last access 18th December 2014. The Labour Manifesto. For a Future Fair for All, , last access 30th January 2013.

255

English Corpus Linguistics Edited by Thomas Kohnen and Joybrato Mukherjee Vol.

1

Roberta Facchinetti / Frank Palmer (eds.): English Modality in Perspective. Genre Analysis and Contrastive Studies. 2004.

Vol.

2

Grace Y. W. Tse: A Corpus-based Study of Proper Names in Present-day English. Aspects of Gradience and Article Usage. 2005.

Vol.

3

Sabine Braun / Kurt Kohn / Joybrato Mukherjee (eds.): Corpus Technology and Language Pedagogy. New Resources, New Tools, New Methods. 2006.

Vol.

4

Clemens W. A. Fritz: From English in Australia to Australian English. 1788–1900. 2007.

Vol.

5

Christiane Brand: Lexical Processes in Scientific Discourse Popularisation. A corpus-linguistic study of the SARS coverage. 2008.

Vol.

6

Sebastian Hoffmann / Stefan Evert / Nicholas Smith / David Lee / Ylva Berglund Prytz: Corpus Linguistics with BNCweb – a Practical Guide. 2008.

Vol.

7

Lars M. Blöhdorn: Postmodifying Attributive Adjectives in English. An Integrated CorpusBased Approach. 2009.

Vol.

8

Uta Katharina Elisabeth Schäpers: Nominal versus Clausal Complexity in Spoken and Written English. Theory and Description. 2009.

Vol.

9

Ulrike Gut: Non-native Speech. A Corpus-based Analysis of Phonological and Phonetic Properties of L2 English and German. 2009.

Vol.

10 Sandra Boggel: Metadiscourse in Middle English and Early Modern English Religious Texts. A corpus-based study. 2009.

Vol.

11 Donald Sims MacQueen: The Integration of MILLION into the English System of Number Words. A Diachronic Study. 2010.

Vol.

12 Tanja Rütten: How to Do Things with Texts. Patterns of Instruction in Religious Discourse 1350–1700. 2011.

Vol.

13 Yvonne Alexandra Breyer: Corpora in Language Teaching and Learning. Potential, Evaluation, Challenges. 2011.

Vol.

14 Kirsten Gather: Syntactic Dislocation in English Congregational Song between 1500 and 1900. A Corpus-based Study. 2014.

Vol.

15 Daphné Kerremans: A Web of New Words. A Corpus-Based Study of the Conventionalization Process of English Neologisms. 2015.

Vol.

16 Ewa Jonsson: Conversational Writing. A Multidimensional Study of Synchronous and Supersynchronous Computer-Mediated Communication. 2015.

Vol.

17 Vera Benninghoven: The Functions of ‘General Nouns’. Theory and Corpus Analysis. 2018.

www.peterlang.com

The author critically discusses the concept of ‘general nouns’, which Halliday/ Hasan introduced in their approach to lexical cohesion (1976), and she provides a comprehensive overview of these nouns from a micro- and a macro-linguistic perspective. For the empirical analysis, the author compiled a corpus, which allows statements about a medium- and genre-specific use of ‘general nouns’. For this purpose, she developed an analytical tool, which takes into account formal and semantic features. The major outcome of the corpus analysis is that ‘general nouns’ are much more flexible in form and function than Halliday/ Hasan assumed and, most importantly, that they fulfil genre-specific functions some of which have not systematically been associated with lexical cohesion.

www.peterlang.com

ECL_017_274758_Benninghoven_MP_A5HC 151x214 globalL.indd 1

english corpus linguistics Thomas Kohnen · Joybrato Mukherjee (eds.) he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit st it out on a test corpus of 50,000 words of spontan he influence of the corpus revolution on applied ling composition of this corpus shows its representativene he compilation of a corpus that is able to capture th studies, diachronic corpus linguistics is a very prom to have access to a corpus as a representative sample advances in English corpus linguistics include the fo h between a monitor corpus for lexicographical descri was observed in the corpus of Old English texts on th itative analysis of corpus data may yield interesting e kind of reference corpus is represented by the Brit

Vera Benninghoven

Vera Benninghoven · The Functions of ‘General Nouns’ Vera Benninghoven has taught linguistics at the English department of the University of Bonn and the University of Cologne. Her research interests include morphology, lexicology, text-linguistics and corpus-linguistics.

V o l u m e 17

ecl 17

n

The Functions of ‘General Nouns’ Theory and Corpus Analysis

13.08.18 03:42