The Directionality of (Inter)subjectification in the English Noun Phrase: Pathways of Change 9783110338751, 9783110318579

The book investigates pathways of (inter)subjectification followed by prenominal elements in the English Noun Phrase, by

212 58 2MB

English Pages 324 Year 2014

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Acknowledgements
Tables
Figures
Introduction
Part I: Synchrony
Chapter 1. A functional-cognitive model of the English NP
1.1. The functional make-up of the English NP
1.1.1. Halliday’s structural-functional account
1.1.2. Bache’s functional zone model
1.1.3. Langacker’s cognitive-functional account
1.1.4. A functional-cognitive model of the English NP
1.2. Categorization
1.2.1. The head of the NP
1.2.2. Subcategorization
1.3. Modification
1.3.1. Descriptive modification
1.3.1.1. Objective and subjective descriptive modifiers
1.3.1.2. Bounded, unbounded and extreme descriptive modifiers
1.3.2. Degree modification
1.3.2.1. Adjective-intensification
1.3.2.2. Noun-intensification
1.3.2.3. Structural-semantic types of degree modification
1.3.2.4. Degree modification and subjectivity
1.4. Determination
1.4.1. Primary determination
1.4.2. Secondary determination
Chapter 2. The English NP: Structure, order and the role of (inter)subjectvity
2.1. The English NP: A mere modifier-head structure?
2.2. A semantic, linear subjective-objective model?
2.2.1. Langackerian subjectivity
2.2.2. Intersubjectivity in Cogitive Grammar
2.2.3. Traugottian subjectivity and beyond
2.2.4. Traugottian intersubjectivity and beyond
2.2.5. Textual meanings and (inter)subjectivity
2.2.6. Subjectivity and intersubjectivity: Disentangling the web
2.2.7. Challenging the linear subjective-objective model
2.3. A prosodic, field-like model
Part II: Diachrony
Chapter 3. (Inter)subjectification and grammaticalization: Pathways of change in the English NP
3.1. Subjectification and intersubjectification
3.2. Grammaticalization
3.2.1. Onset contexts of grammaticalization
3.2.1.1. Ambiguity
3.2.1.2. Specialized contexts: Dialogicity and collocation
3.3. The directionality of (inter)subjectification and grammaticalization
3.4. (Inter)subjectification and grammaticalization in the English NP
3.4.1. From description to secondary determination
3.4.2. Pathways to noun-intensification
3.4.2.1. Pathway 1: From description to nounintensification
3.4.2.2. Pathway 2: From identification to nounintensification
3.4.3. Adamson’s lefward and rightward movement hypotheses
3.5. Conclusions
Part III: The case studies
Chapter 4. Data and Methods
4.1. Selection of the data
4.2. Corpora
4.2.1. Old English
4.2.2. Middle English
4.2.3. Modern English
4.2.4. Present Day English
4.3. Methodology
Chapter 5. The completeness adjectives
5.1. Complete
5.1.1. Descriptive modifier uses of complete
5.1.1.1. Bounded descriptive modifier uses of complete
5.1.1.2. Descriptive modifier uses of complete, vague between bounded and unbounded construal
5.1.2. Identifying uses of complete
5.1.3. Noun-intensifying uses of complete
5.1.3.1. Closed scale noun-intensifying uses
5.1.3.2. Noun-intensifying uses hovering between an open scale and a closed scale reading
5.1.4. Adjective-intensifying uses of complete
5.1.5. From description to identification and nounintensification
5.2. Total
5.2.1. Bounded descriptive modifier uses of total
5.2.2. Identifying uses of total
5.2.3. Noun-intensifying uses of total
5.2.4. Adjective-intensifying uses of total
5.2.5. Classifying uses of total
5.2.6. From description to identification and nounintensification
5.3. Whole
5.3.1. Descriptive modifier uses of whole
5.3.1.1. Bounded descriptive modifier uses of whole
5.3.1.2. Bounded/unbounded descriptive modifier uses of whole
5.3.2. Identifying uses of whole
5.3.3. Noun-intensifying uses of whole
5.3.3.1. Closed scale noun-intensifying uses
5.3.3.2. Open scale noun-intensifying uses
5.3.4. Adjective-intensifying uses of whole
5.3.5. From description to identification to nounintensification
5.4. Conclusions
Chapter 6. The specificity adjectives
6.1. Unbounded descriptive modifier uses of particular and specific
6.1.1. Objective descriptive modifier uses
6.1.2. Subjective descriptive modifier uses
6.2. Identifying uses of particular and specific
6.2.1. Linking secondary determiner uses
6.2.2. Individuating secondary determiner uses
6.2.3. Focusing secondary determiner uses
6.2.4. The development of secondary determiner uses
6.3. Noun-intensifying uses of particular
6.3.1. The development of noun-intensifying particular
6.3.1.1. The emphatic linking construction
6.3.1.2. The negative quantifying construction
6.3.2. Collocational evidence for the shift from identification to noun-intensification
6.4. Classifying uses of specific
6.5. From description to identification (to nounintensification)
Chapter 7. Such, zulk and what
7.1. Identifying such
7.1.1. Type-phoricity
7.1.2. Generalized instantiation
7.1.3. Definiteness and indefiniteness in conflict
7.2. Noun-intensifying such
7.2.1. Semantic-pragmatic types of noun-intensification expressed by such
7.2.2. Collocational behaviour of noun-intensifying such
7.3. The diachronic relation between identifying and noun-intensifying such
7.4. A contrastive study of English such and Dutch zulk
7.4.1. Identifying zulk
7.4.2. Noun-intensifying zulk
7.4.3. From identification to noun-intensification … and back
7.5. A comparative study of English such and what
7.5.1. Identifying what: an interrogative primary determiner
7.5.2. Noun-intensifying what
7.5.2.1. Semantic-pragmatic types of nounintensification
7.5.2.2. Collocational behaviour
7.5.2.3. Syntactic behaviour
7.5.3. Identification and noun-intensification: A fuzzy boundary
7.5.4. Such and what: A synchronic comparison
7.6. Conclusions
Chapter 8. Old and little: subjective compounds
8.1. Subjective compounds with old
8.1.1. General characterization
8.1.2. The development of subjective compouns with old
8.2. Subjective compounds with little
8.2.1. General characterization
8.2.2. The development of subjective compouns with little
8.3. Objective compounds with old and little
8.4. Subjective compounds and their repercussions for subjectivity and subjectification in the NP
Summary: Towards a reconciliation of synchrony and diachrony
References
Subject index
Recommend Papers

The Directionality of (Inter)subjectification in the English Noun Phrase: Pathways of Change
 9783110338751, 9783110318579

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

Lobke Ghesquie`re The Directionality of (Inter)subjectification in the English Noun Phrase

Trends in Linguistics Studies and Monographs 267

Editor

Volker Gast Editorial Board

Walter Bisang Jan Terje Faarlund Hans Henrich Hock Natalia Levshina Heiko Narrog Matthias Schlesewsky Amir Zeldes Niina Ning Zhang Editor responsible for this volume

Heiko Narrog

De Gruyter Mouton

The Directionality of (Inter)subjectification in the English Noun Phrase Pathways of Change by

Lobke Ghesquie`re

De Gruyter Mouton

ISBN 978-3-11-031857-9 e-ISBN (PDF) 978-3-11-033875-1 e-ISBN (EPUB) 978-3-11-039526-6 ISSN 1861-4302 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. ” 2014 Walter de Gruyter GmbH, Berlin/Boston Printing and binding: CPI books GmbH, Leck  Printed on acid-free paper Printed in Germany www.degruyter.com

Acknowledgements During my years as a doctoral and postdoctoral researcher at the KU Leuven I have been very fortunate to have been surrounded by a number of people who have all in some way contributed to this book. It is those people I wish to thank now for their support, help and simply for being there. First, I want to express my deep gratitude to Kristin Davidse who supervised my research. Her enthusiasm for the English noun phrase and its intricacies quickly proved catching and when I was offered the opportunity to prepare a dissertation on processes of change affecting the NP I eagerly accepted it. I feel privileged to have been one of Kristin’s students and greatly appreciate how she has always managed to be both an inspiring and encouraging supervisor and a caring, warm-hearted colleague. I also deeply thank my co-supervisor Hendrik De Smet for his ever precise and careful comments on my writings as well as for being a great colleague and office mate. I sincerely thank the other members of my doctoral examination board Tine Breban, Victorina González-Díaz and Heiko Narrog for their detailed and very useful comments and suggestions. My colleagues at the linguistics department I warmly thank for their encouragement, chats and friendship. For financial support, I thank the Interuniversity Attraction Poles Programme (Belgian State – Belgian Science Policy), project P6/44 Grammaticalization and (inter)subjectification, coordinated by Johan van der Auwera and the Research Fund Flanders. My research was further supported by the Spanish Ministry for Science and Innovation (grant no. HUM2007-60706/FILO) and the European Regional Development Fund, coordinated by Teresa Fanego. Finally, I thank the people dearest to me. Alwin, you have been there for me every minute of every day for the last 15 years. Thank you for your love, care and your ability to make me laugh at your silly jokes no matter what state of mind I am in. I can only hope our son and daughter will become as loving and caring as you are and that one day they will make someone as happy as you have made me. Aike, my sweet, darling little devil, your smile was all I needed to get me through another day of writing. This book is dedicated to you.

Contents Acknowledgements Tables Figures Introduction

v xii xiv 1

Part I: Synchrony Chapter 1 A functional-cognitive model of the English NP 1.1. The functional make-up of the English NP 1.1.1. Halliday’s structural-functional account 1.1.2. Bache’s functional zone model 1.1.3. Langacker’s cognitive-functional account 1.1.4. A functional-cognitive model of the English NP 1.2. Categorization 1.2.1. The head of the NP 1.2.2. Subcategorization 1.3. Modification 1.3.1. Descriptive modification 1.3.1.1. Objective and subjective descriptive modifiers 1.3.1.2. Bounded, unbounded and extreme descriptive modifiers 1.3.2. Degree modification 1.3.2.1. Adjective-intensification 1.3.2.2. Noun-intensification 1.3.2.3. Structural-semantic types of degree modification 1.3.2.4. Degree modification and subjectivity 1.4. Determination 1.4.1. Primary determination 1.4.2. Secondary determination

13 15 15 18 21 23 25 26 27 28 29 30 33 34 35 36 38 41 44 46 47

viii Contents

Chapter 2 The English NP: Structure, order and the role of (inter)subjectvity 2.1. The English NP: A mere modifier-head structure? 2.2. A semantic, linear subjective-objective model? 2.2.1. Langackerian subjectivity 2.2.2. Intersubjectivity in Cogitive Grammar 2.2.3. Traugottian subjectivity and beyond 2.2.4. Traugottian intersubjectivity and beyond 2.2.5. Textual meanings and (inter)subjectivity 2.2.6. Subjectivity and intersubjectivity: Disentangling the web 2.2.7. Challenging the linear subjective-objective model 2.3. A prosodic, field-like model

51 52 55 56 57 59 62 64 69 71 73

Part II: Diachrony Chapter 3 (Inter)subjectification and grammaticalization: Pathways of change in the English NP 3.1. Subjectification and intersubjectification 3.2. Grammaticalization 3.2.1.Onset contexts of grammaticalization 3.2.1.1. Ambiguity 3.2.1.2. Specialized contexts: Dialogicity and collocation 3.3. The directionality of (inter)subjectification and grammaticalization 3.4. (Inter)subjectification and grammaticalization in the English NP 3.4.1. From description to secondary determination 3.4.2. Pathways to noun-intensification 3.4.2.1. Pathway 1: From description to nounintensification 3.4.2.2. Pathway 2: From identification to nounintensification 3.4.3. Adamson’s lefward and rightward movement hypotheses 3.5. Conclusions

81 81 86 90 90 92 96 98 98 101 101 103 106 107

Contents ix

Part III: The case studies Chapter 4 Data and Methods 4.1. Selection of the data 4.2. Corpora 4.2.1. Old English 4.2.2. Middle English 4.2.3. Modern English 4.2.4. Present Day English 4.3. Methodology Chapter 5 The completeness adjectives 5.1. Complete 5.1.1. Descriptive modifier uses of complete 5.1.1.1. Bounded descriptive modifier uses of complete 5.1.1.2. Descriptive modifier uses of complete, vague between bounded and unbounded construal 5.1.2. Identifying uses of complete 5.1.3. Noun-intensifying uses of complete 5.1.3.1. Closed scale noun-intensifying uses 5.1.3.2. Noun-intensifying uses hovering between an open scale and a closed scale reading 5.1.4. Adjective-intensifying uses of complete 5.1.5. From description to identification and nounintensification 5.2. Total 5.2.1. Bounded descriptive modifier uses of total 5.2.2. Identifying uses of total 5.2.3. Noun-intensifying uses of total 5.2.4. Adjective-intensifying uses of total 5.2.5. Classifying uses of total 5.2.6. From description to identification and nounintensification 5.3. Whole 5.3.1. Descriptive modifier uses of whole 5.3.1.1. Bounded descriptive modifier uses of whole 5.3.1.2. Bounded/unbounded descriptive modifier uses of whole

113 113 115 116 116 116 117 119 123 125 125 125 128 130 135 135 139 141 142 145 146 148 151 156 156 157 159 160 160 162

x Contents

5.3.2. Identifying uses of whole 5.3.3. Noun-intensifying uses of whole 5.3.3.1. Closed scale noun-intensifying uses 5.3.3.2. Open scale noun-intensifying uses 5.3.4. Adjective-intensifying uses of whole 5.3.5. From description to identification to nounintensification 5.4. Conclusions Chapter 6 The specificity adjectives 6.1. Unbounded descriptive modifier uses of particular and specific 6.1.1. Objective descriptive modifier uses 6.1.2. Subjective descriptive modifier uses 6.2. Identifying uses of particular and specific 6.2.1. Linking secondary determiner uses 6.2.2. Individuating secondary determiner uses 6.2.3. Focusing secondary determiner uses 6.2.4. The development of secondary determiner uses 6.3. Noun-intensifying uses of particular 6.3.1. The development of noun-intensifying particular 6.3.1.1. The emphatic linking construction 6.3.1.2. The negative quantifying construction 6.3.2. Collocational evidence for the shift from identification to noun-intensification 6.4. Classifying uses of specific 6.5. From description to identification (to nounintensification) Chapter 7 Such, zulk and what 7.1. Identifying such 7.1.1. Type-phoricity 7.1.2. Generalized instantiation 7.1.3. Definiteness and indefiniteness in conflict 7.2. Noun-intensifying such 7.2.1. Semantic-pragmatic types of noun-intensification expressed by such 7.2.2. Collocational behaviour of noun-intensifying such

162 165 166 170 173 176 176 179 181 181 184 185 186 188 189 191 194 195 196 200 201 203 206 209 211 211 213 218 222 223 225

Contents xi

7.3. The diachronic relation between identifying and noun-intensifying such 7.4. A contrastive study of English such and Dutch zulk 7.4.1. Identifying zulk 7.4.2. Noun-intensifying zulk 7.4.3. From identification to noun-intensification … and back 7.5. A comparative study of English such and what 7.5.1. Identifying what: an interrogative primary determiner 7.5.2. Noun-intensifying what 7.5.2.1. Semantic-pragmatic types of nounintensification 7.5.2.2. Collocational behaviour 7.5.2.3. Syntactic behaviour 7.5.3. Identification and noun-intensification: A fuzzy boundary 7.5.4. Such and what: A synchronic comparison 7.6. Conclusions Chapter 8 Old and little: subjective compounds 8.1. Subjective compounds with old 8.1.1. General characterization 8.1.2. The development of subjective compouns with old 8.2. Subjective compounds with little 8.2.1. General characterization 8.2.2. The development of subjective compouns with little 8.3. Objective compounds with old and little 8.4. Subjective compounds and their repercussions for subjectivity and subjectification in the NP

231 234 235 237 238 242 243 244 244 246 246 248 250 252 253 256 256 261 266 266 268 271 273

Summary Towards a reconciliation of synchrony and diachrony

277

References Subject index

287 305

Tables Table 1. Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. Table 12. Table 13. Table 14. Table 15. Table 16. Table 17. Table 18. Table 19. Table 20. Table 21. Table 22.

The determination zone in the English NP Lehmann’s grammaticalization parameters The Helsinki Corpus and the York-Toronto-Helsinki Parsed Corpus of Old English Prose The Helsinki Corpus and the Penn-Helsinki Parsed Corpus of Middle English The Penn-Helsinki Parsed Corpus of Early Modern English The Corpus of Late Modern English Texts The COBUILD corpus Nouns co-occurring with descriptive modifier complete in the CEMET, PPCEME and CLMETEV data sets Nouns co-occurring with noun-intensifying complete in the CLMETEV data sets Quantified results of the corpus study of complete Nouns co-occurring with descriptive modifier total in the PPCEME, CEMET, CLMETEV and CB data sets Nouns co-occurring with noun-intensifying total in the CLMETEV and CB data sets Quantified results of the corpus study of total Nouns co-occurring with secondary determiner whole in the PPCEME data set Nouns co-occurring with noun-intensifying whole in the PPCEME corpus Quantified results of the corpus study of whole Nouns co-occurring with noun-intensifying particular Nouns co-occurring with classifying specific Quantified results of the corpus study of particular Quantified results of the corpus study of specific Nouns occurring with noun-intensifying such in the diachronic data Frequency lists of degree nouns combining with nounintensifying such in the pattern such (a) N(s)

44 87 116 116 117 117 118 137 138 143 147 154 158 164 172 175 202 205 207 208 227 228

Tables

Table 23. Table 24. Table 25. Table 26. Table 27. Table 28. Table 29. Table 30. Table 31. Table 32. Table 33. Table 34. Table 35. Table 36. Table 37. Table 38.

Frequency lists of modifiers intervening between noun-intensifying such and the head in the historical and contemporary data Quantified results of the corpus study of such Frequency lists of nouns modified by nounintensifying what Frequency lists of modifiers modified by nounintensifying what Schematic summary of the main differences and similarities between prenominal such and what Distribution of different functions of such and what in the CB data Distribution of different functions of such and what in the CB times data Distribution of different functions of such and what in the CB ukspok data Joint frequency, T-scores and MI-scores for old + noun in British sections of WB corpus Quantified results of the diachronic corpus study of old Quantified results of the synchronic corpus study of old Joint frequency, T-scores and MI-scores for little + noun in British sections of WB corpus Quantified results of the diachronic corpus study of little Quantified results of the synchronic corpus study of little Joint frequency, T-scores and MI-scores for objective compounds with old in British sections of WB corpus Joint frequency, T-scores and MI-scores for objective compounds with little in British sections of WB corpus

xiii

230 232 247 247 250 251 251 251 260 264 266 268 269 270 272 272

Figures Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. Figure 17. Figure 18. Figure 19. Figure 20. Figure 21. Figure 22. Figure 23. Figure 24.

The symbolic structure of a construction The experiential structure of the English NP A functional zone model of the English NP A functional-cognitive model of the English NP The recursive modification relations governing the NP elements Submodification Objective construal Subjective construal Types of subjectivity and intersubjectivity Scopal relation of expletives Integrated model of semantic change A comparison of Heine’s (2002) and Diewald’s (2002, 2006) contexts of grammaticalization Model of the prenominal string of the English NP The development of identifying and noun-intensifying uses of complete The development of identifying and closed scale nounintensifying uses of total The development of identifying and noun-intensifying uses of whole The development of identifying uses of particular Generalized instantiation Generalizing from a type-description Generalizing from concrete instances Modification patterns for noun-intensifying such Modification patterns for noun-intensifying zulk Distributional pattern for the different functions of Dutch zulk Integrated model of semantic change

14 17 19 24 53 53 57 57 70 73 85 91 109 144 159 174 193 215 216 216 226 238 239 285

Introduction This study investigates (inter)subjectification processes affecting prenominal elements in the English noun phrase (henceforth NP) and the specific pathways along which these changes proceed. In the literature on (inter)subjectification in the English NP, two main pathways in opposite directions have been identified. Functionally shifting elements which move to the left in NP structure have been predicted to acquire more subjective meaning, while shifts to the right in the prenominal structure have been said to entail de-subjectification (e.g. Adamson 2000). The case studies in the body of this book have been set up in such a way that the hypotheses about these two main paths can be verified, refined and modified, if necessary. Hypotheses about the leftward path are complicated by the fact that there are two targets of subjectification realized at the left end of NP structure: intensifying and identifying elements. If the hypothesis of leftward movement-cum-subjectification applies strictly, then identifying elements should be expected to develop from intensifying elements, as the former occupy a more leftward position in NP structure. However, the relative chronology of the development of intensifying and identifying uses by the same items has not received much attention so far. Yet, the question of how two distinct (inter)subjectification processes affect the same items is highly relevant to theory formation on (inter)subjectification. It can shed more light on the directionality issue as well as on the mechanisms of change involved in shifts both from objective to subjective functional elements, and from subjective to other subjective or intersubjective elements. In this study, the focus is thus on the development of identifying and intensifying meanings of prenominal items as they are expressed by secondary determiner elements and noun-intensifiers, illustrated in (1)–(2) and (3)–(4). (1)

Two long-lost episodes of Dad's Army from 1969 have been unearthed among a heap of rusting film cans in Hertfordshire by a man who wishes to remain anonymous. Producer and co-writer David Croft is delighted at the discovery and now has three more

2

(2)

(3)

(4)

Introduction

episodes to find for the complete set to be available. (WB brregnews)1 Throughout the rest of the ship emergency lighting was provided by hand-hung nickel-cadmium lamps: a twist of the base of such a lamp provided at least a bare minimum of illumination. (WB brbooks) All he wanted to do was get to the murder scene before the CID arrived and started treating him like a turnip. At this rate, he'd end up looking like a complete wally who hadn't even managed to keep tabs on the occupants of the house. (WB brbooks) She's such a stickler for cleanliness. Always checking up on the maids. (WB brbooks)

On the one hand, secondary determiners, as in (1) and (2), provide identifying and quantifying information supplementary to and semantically more complex than that provided by primary determiners. In (1), for instance, complete serves an identifying function, indicating, together with the definite determiner the, that reference is to the entire entity referred to by the head noun set and not just to one or more parts of it. Similarly, such in (2) makes the identification conveyed by the indefinite determiner a more specific by setting up an anaphoric relation between the NP referent and the type of lamp described in the preceding discourse. On the other hand, nounintensifiers modify the degree of gradable properties conveyed by the descriptive part of noun phrases, i.e. the head noun plus any descriptive modifiers. In (3) and (4), complete and such heighten the degree of foolishness inherent in the head noun wally and the degree of insistence implied by the noun stickler respectively. I will argue in this study that secondary determiner and nounintensifying uses typically arise through (inter)subjectification and grammaticalization processes. So far, (inter)subjectification and grammaticalization have received much less attention in the noun phrase than in the verb phrase. Yet, the NP is a particularly interesting constructional environment for these processes for a number of reasons. Synchronically, the NP forms an elaborate syntagmatic structure in which position and order tend to correlate with differences in semantics. Moreover, as I will argue in Chapters 1 1. All examples followed by (WB) have been extracted from the Wordbanks Online corpus and are reproduced here with the kind permission of HarperCollins Publishers. I also indicate the subcorpus from which the examples are taken. More information on the (sub)corpora consulted for this study can be found in §4.2.

Introduction

3

and 2, these semantics range from non-subjective to subjective and (textually) intersubjective. Diachronically, elements of the NP – both head, modifiers and determiners – have shifted from non- or less subjective to (inter)subjective and from lexical to more abstract functions, creating extensive synchronic layering (Hopper and Traugott 2003: 124; §3.2). Besides the hypothesis of the left-oriented path of subjectification, Adamson (2000) also put forward a correlation between rightward movement and de-subjectification in the English NP. An example of this path of change is the shift from descriptive modifier use, e.g. a criminal act, to classifier use, e.g. a criminal court. Adamson (2000: 60) also noted that in this case the process of change involved is lexicalization, rather than grammaticalization. In Chapter 8, I will further investigate this rightward pathway by studying the processes that led to lexicalizations such as old git and little battler, as illustrated in (5) and (6). (5) (6)

THAI BRIDE she used a British old git to get a British Passport. (www.youtube.com/watch?v=zUeIO4xqhuQ) MATE THIS IS ALL I CAN AFFORD I'm not rich I am a struggling average Australian little Battler, besides I thought every little bit counts. (http://www.liberal.org.au/Pages/No-Carbon-Tax.aspx)

The issue these lexicalizations raise is that old and little convey evaluative, subjective meaning but occur at the right end of the prenominal string and can themselves be preceded by ‘objective’ classifying adjectives such as those describing nationalities in (5) and (6). They therefore appear to challenge the view that rightward movement of adjectives (away from their central property-attributing use) always involves de-subjectification. This study will try to tie up a number of loose ends concerning the synchronic organization and diachronic development of the English NP by addressing three main points. The first, descriptive aim of this study is to develop detailed semantic and lexico-grammatical characterizations of all the functions singled out above. The secondary determiner and nounintensifying uses illustrated in (1) to (4), as well as the subjective lexicalizations in (5) and (6), will be focused on both from a synchronic and diachronic perspective and with special attention to their relation to the dimension of (inter)subjectivity. Second, a functional-cognitive model of the NP will be developed with special attention to the ordering of elements within this model. It will be suggested that the traditional view in which subjectivity is inextricably

4

Introduction

linked to the left periphery of the NP structure is unsatisfactory and that rather subjectivity is spread throughout the entire NP. Third, in addition to these more descriptive aspects, this study wants to contribute to theory formation on (inter)subjectivity and (inter)subjectification in the NP. Issues to be clarified are, on the one hand, the directionality and the underlying formal and conceptual mechanisms of the development by certain sets of elements of both identifying and nounintensifying uses and, on the other, the appearance of subjective meaning at the extreme right of the prenominal string, as in (5) and (6). Based on observations with regard to these elements in the NP, I will try to develop a more encompassing and more detailed understanding of (inter)subjectification and incorporate the different functions and functional zones of the English NP into existing hypotheses on structural-semantic changes in the NP. Also, I will present new ideas on how textual meanings fit into the traditional typology of objective, subjective and intersubjective meaning and on the the role of onset contexts in semantic change. The research reported on in this study has a strongly descriptive, dataoriented focus, while at the same time aiming to verify existing theoretical hypotheses, formulate theoretical generalizations and contribute to theoryformation. Aim 1: Description of identifying and noun-intensifying slots in the NP Synchronically, the English NP displays an intricate structural-functional make-up with a potentially long prenominal string. The NP in (7), for instance, contains a focusing adverb (almost), predeterminer (double), determiner (the), postdeterminer (usual), numerative (five), qualitative adjective (stale) and classifying noun (oatmeal). However, NP structure can become much more intricate this. For example, more than one element with the same function can appear in the same NP (e.g. stale, dry oatmeal and chocolate biscuits), and the different component elements can receive additional modification (e.g. very stale). (7)

On Easter Sunday, we were offered almost double the usual five stale oatmeal biscuits we normally get from Auntie Flo. (Davidse 1997: 11)

Most standard grammars of English feature brief discussions of all of the above slots in the NP, yet much remains to be uncovered, especially

Introduction

5

with regard to identifying and noun-intensifying meanings. The main descriptive aim of this study is the development of semantic and lexicogrammatical characterizations of these ‘layered’ constructions of elements in the NP that are engendered by grammaticalization and (inter)subjectification processes (Chapter 1). First, for the identifying secondary determiner function much work has recently been done by Breban (2002, 2006, 2010a), Breban and Davidse (2003) and Davidse, Breban, and Van linden (2008). This study will take their work as starting-point for its thinking on identifying meanings in the NP and their relation to (inter)subjectivity (§1.4; §2.4), as well as for the description of the secondary determiner uses of the items studied (Chapters 5 to 7). Second, nounintensification is a so far largely uncharted domain. The literature has mainly focused on adjective-intensification, typically expressed by adverbs (e.g. pretty good) (see for instance Peters 1994). Noun-intensification has only recently come into vogue as a research area (e.g. Paradis 2000, 2001, 2008; Athanasiadou 2006, 2007). Its incorporation in this study was motivated in the first place by the fact that items like complete and such manifested both these (noun-)intensifying and identifying uses. But the relative neglect of noun-intensification as a diachronic process was a second good reason to undertake the study of this process. Aim 2: A synchronically and diachronically dynamic model of the English NP Traditionally, the English NP is analyzed as a modifier-head structure, with the noun as head preceded and followed by dependents which modify it (e.g. Quirk et al. 1985; Huddleston and Pullum 2002). The left-right order in which prenominal modifiers and head noun appear in the NP is usually assumed to reflect a more general continuum of subjective to objective meanings (e.g. Quirk et al. 1972; Dixon 1982). The idea is that the more leftward elements in NP structure are more subjective and the more one moves to the right of the NP the less subjective the elements become. This general ordering principle has been largely undebated. To be able to evaluate this view on the synchronic ordering of the English NP, it is important to have a clear idea of what it means for an element to be objective, subjective or, by extension, intersubjective. The ordering principle and the notions it crucially rests upon are in need of critical reflection. This study hopes to contribute to theory-formation on (inter)subjectivity by discussing and elaborating on the two notions of (inter)subjectivity most influ-

6

Introduction

ential in recent cognitive-functional and constructional research, namely Langackerian and Traugottian (inter)subjectivity. Especially Traugottian (inter)subjectivity – both in its synchronic and diachronic dimensions – will feature prominently in this study as the NP proves an excellent testing ground for evaluating its validity and general applicability (§2.2.3; §2.2.4; §3.1). All of the prenominal functions as well as the head of the NP will be discussed in terms of Traugottian (inter)subjectivity (§1.2 to §1.4). Nounintensifiers will be argued to typically convey subjective meanings, indexing the speaker’s attitude toward or viewpoint on the NP referent.2 On the basis of observations made for determining elements in the NP, it will be proposed that in particular the notion of (inter)subjectivity is in need of refinement. Besides the generally acknowledged attitudinal type of intersubjectivity, indexing the speaker’s attention to the social self or face of the hearer (Traugott 2003), at least one other type has to be distinguished, i.e. textual intersubjectivity. Textual intersubjectivity deals with meanings designed to steer interpretation by the hearer (e.g. focus and backgrounding mechanisms, items establishing joint focus of attention) (§2.2.5). Further, I will argue that although there is a general tendency for elements to occur on a linear subjective-objective cline within NP structure, this tendency can be overridden. Ordering in the NP is determined by a number of competing principles, each with its own ordering preferences. Also, rather than as a rigid continuum of independent elements ordered from subjective to objective, I will claim that the English NP is better conceived of as a cline of structural-functional zones. This cline starts at the leftmost end with the determination zone (e.g. the, another, most), followed by the modification zone – comprising degree modifiers (e.g. very, greatly) and descriptive modifiers (e.g. big, beautiful) – and ends with the classifying modifiers (e.g. financial, English) and head nouns of the categorization zone at the rightmost end (§1.1.4). The relationship of these distinct zones to (inter)subjectivity is not a straightforward linear one. For instance, head nouns – both simple and compound – which are typically situated at the rightmost end of the NP, can convey either objective (e.g. man, blackbird) or subjective meaning (e.g. idiot, old fart) (see Chapter 8). I will hypothesize that a prosodic, field-like model of the NP, constrained by NP structure, can better account for the ordering of subjective and objective elements (§2.3).

2. ‘Speaker’ and ‘hearer’ will in this study be used as cover terms for the speaker and hearer in spoken communication and the writer and reader of written texts.

Introduction

7

Aim 3: Towards a more encompassing and more detailed understanding of (inter)subjectification processes and their directionality in the English NP Many linguistic items display synchronic layering in that they can convey different meanings in different contexts (Hopper and Traugott 2003: 124; §3.2), possibly covering the entire spectrum of non- or less subjective to subjective and intersubjective meanings. In example (1), complete, as a secondary determiner, conveys what is in this study referred to as textually intersubjective meaning. In (3), noun-intensifying complete expresses subjective meanings, indexing the speaker’s negative evaluation of the person talked about. Similarly, in (8), the adjective also conveys subjective speaker evaluation, but at the same time has a prominent descriptive function. In (9), finally, complete objectively describes the NP referent as containing all the necessary parts. In sum, as Sweetser (1999: 136) notes, “[i]nstead of each word always representing the same rigid and stable semantic chunk or building-block, the same word can represent very different complex meaning structures in different contexts, and may alter flexibly depending on the meanings surrounding it”. (8)

(9)

The faultless and secure expression of an exclusive worship of things formally beautiful ... The melody of colour, the symphony of form is complete: one more beautiful thing is achieved, one more delight is born into the world; and its meaning is beauty; and its reason for being is to be.' (WB brbooks) Buddhist doctrines and teachings are often presented in dualistic forms -- samsara and nirvana (the world, and freedom from the world), karma and rebirth (cause and effect), buddha and ordinary beings (awakened and deluded). Each presentation is complete in itself. (WB brbooks)

My interest lies not only with this synchronic variation per se but also and most prominently with what gave rise to it. As Partington (1993: 178) notes, “[t]he synchronic approach tells linguists how the bits and pieces of a language fit together, but a diachronic approach can help us to understand why. It can help us to understand the forces and processes which put the pieces where they are”. With the exception of the elements studied in Chapters 7 and 8, all of the adjectival prenominal elements investigated for this study, including complete, have in Present Day English objective (descriptive), subjective (descriptive and noun-intensifying), and (textually) inter-

8

Introduction

subjective (secondary determiner) uses. Quantified diachronic studies of these constructions can then be used to verify existing theoretical hypotheses about the directionality of grammaticalization and (inter)subjectification. Through in depth diachronic and synchronic corpus study of various elements of the NP (Chapters 5 to 8), I investigate whether the structuralsemantic changes affecting these elements generally proceed along the unidirectional cline of (inter)subjectification as hypothesized by Traugott (2003). Moreover, for identifying, secondary determiner adjectives, Davidse, Breban, and Van linden (2008) have proposed that they develop from lexical, descriptive adjectives through deictification (§3.4.1). For the development of noun-intensifying meanings, two main pathways have been proposed in the literature. In accordance with the diachronic objective > subjective > intersubjective hypothesis put forward by Traugott (2003), Paradis (2000, 2001) has developed a framework accounting for the development of noun-intensifying uses by lexical, descriptive adjectives which crucially relies on the boundedness construal of the adjectives. Adjectives are argued to shift from description to intensification by means of a foregrounding mechanism: the boundedness configurations of the lexical uses are foregrounded at the expense of their lexical content. Contradicting Traugott’s (2003) (inter)subjectification hypothesis, Bolinger (1972) posited the currently largely neglected pathway from more abstract identifying elements to noun-intensifiers – a shift claimed for both purely identifying (e.g. from such a man to such a mess) and quantitatively identifying items (e.g. from too much time to too much of a miser). The validity of all of the above pathways is verified by means of the quantified diachronic and synchronic studies presented in Chapters 5 to 7, making up the bulk of this study. In these chapters the different (inter)subjectification processes affecting the completeness adjectives complete, total and whole; the specificity adjectives particular and specific; and the determining elements such, zulk and what will be discussed in detail, with particular attention for the semantic, structural and collocational aspects of these processes. Importantly, specific attention will go the role and variety of trigger or onset contexts in semantic change. Besides ambiguity, dialogic and especially collocational contexts will be explored as creating a specific type of ‘specialized context’ (Traugott 2008) which may trigger the onset of grammaticalization. While these studies confirm the (inter)subjective teleology and concomitant leftward movement as a general tendency, they also reveal notable exceptions. The case studies provide empirical evidence for Bolinger’s (1972) ‘identifying to intensifying’ hypothesis as an alternative develop-

Introduction

9

mental path to that from descriptive to intensifying uses focused on in more recent work. However, as will be shown by the study of the completeness adjectives (Chapter 5), not all elements with identifying and nounintensifying uses have developed the latter from the former: complete and total developed noun-intensifying uses from descriptive uses, but with whole it was emphatic quantifying expressions that led to intensifying uses. Collocational patterning will be discussed as an important factor in all this. In addition, this study sets out to verify Adamson’s (2000) hypothesis predicting that elements in the English NP will change meaning and undergo positional shifts in the NP accordingly, either in a systematic right-toleft movement while shifting from objective to subjective meaning or in a systematic left-to-right movement while at the same time de-subjectifying. This issue will be addressed mainly in Chapter 8. In sum, this study focuses on theoretical implications and possible modifications of or additions to the general (inter)subjectification framework, based on maximally accurate and fine-grained descriptions of empirical phenomena of (inter)subjectification in the English NP. The theoretical approach adopted in this study is a constructional one of cognitive-functional affiliation such as advocated by Croft (2001). The Langacker-inspired attention to dependency structure in Croft’s construction grammar will be further supplemented with compatible insights from the functional tradition of Halliday (1994) and McGregor (1997). This approach takes constructions, viewed as functional structures in which grammar and lexis are integrated with each other, as the primary units of analysis. I will focus on the different discourse functions of the items studied as distinct form-meaning pairings, conceived of as component structures of the composite NP structure. I will investigate in the data by which different structural and lexical co-occurrence patterns these are characterized. This approach, combined with the data-driven nature of the study, will allow me to provide detailed and in some respects novel characterizations of the uses of the items studied and to see what the diachronic findings on their development entail for existing (inter)subjectification hypotheses. To conclude, I briefly introduce the organization of the discussion to follow. The study is divided into three main parts. Parts I and II together provide the necessary theoretical and conceptual background to the case studies presented in Part III. Part I consists of two chapters introducing the English NP from a synchronic perspective. Chapter 1 introduces a functional-cognitive model of the English NP as made up of three main structural-functional zones, i.e. the determination zone, the modification zone and the categorization zone. The different functional slots of each of the

10

Introduction

zones are characterized semantically and morpho-syntactically, with specific attention to their (inter)subjective nature. Chapter 2 discusses the structure of the English NP and the ordering of prenominal modifiers and head within this structure. First, the NP is described as a logical modifier-head structure. Second, the generally assumed ordering within the NP as displaying linear left-to-right progression from subjective to objective meanings is introduced. An assessment of this view on NP order requires a detailed analysis of the notion of (inter)subjectivity. Accordingly, the major part of Chapter 2 is devoted to a discussion of and elaboration on the notions of subjectivity and intersubjectivity as they are developed in the work of Langacker and Traugott. In the third and final section of Chapter 2, an alternative view on NP ordering is proposed in the form of a prosodic, field-like model governed by several competing ordering principles integrated into an overall order constrained by NP structure. Part II turns from the synchronic description of the NP to its diachrony. The processes of (inter)subjectification and grammaticalization are first introduced as general mechanisms of change, before addressing their relevance to the development of the different constituent elements of the NP, in particular of the secondary determiner and noun-intensifying functions. Different pathways of change to identifying and noun-intensifying meaning recognized in the literature are discussed. These pathways are assessed in Part III of the study, grouping together the case studies. Chapters 5 to 7 trace the development of the identifying and noun-intensifying uses of the completeness and specificity adjectives and of such, zulk and what by means of in-depth corpus studies. In Chapter 8, I will focus on the development of lexicalized expressions involving the adjectives old and little, which I will characterize as subjective compounds. The selection of the data for these studies and the methods used are detailed in Chapter 4. Finally, a summarizing chapter briefly recapitulates the findings from the case studies and attempts to fit them into the general synchronic and diachronic account of the English NP introduced in Parts I and II of the study.

Part I Synchrony Not chaos-like together crush'd and bruis'd, But, as the world, harmoniously confused: Where order in variety we see, And where tho' all things differ, all agree. Alexander Pope – Windsor Forest

Chapter 1 A functional-cognitive model of the English NP This study assumes a constructional approach such as developed in the cognitive-functional tradition (e.g. Langacker 1991, 2007; Croft 2001; Fried 2010). The notion of construction that I will work with is that of a layered organization of functions realized by structural dependency relations between the component units (Langacker 1991: 143–146). This chapter will introduce the different functions fulfilled by the constituent elements of the English NP as distinct form-meaning pairings. Chapter 2 will then focus on the structural make-up of the English NP, i.e. how the functional layers are symbolized by dependency structures of the modifier-head kind (Langacker 1991: 146). In the following sections of this chapter, the relevant functions will be discussed as distinct form-meaning pairings with special attention for their (inter)subjective character. Traditional approaches to the NP have often been class-based, taking word class categories such as verb, adjective and noun as the starting point for the description. In such a class-based approach of which – partial – examples are found in Quirk et al. (1972, 1985), Bache (2000) and Crystal (2003), the English NP is described as built up of elements from the traditional word classes determiner, adjective and noun, which are typically associated with identification, attribution and categorization respectively. Such an account of English NP structure is, however, problematic in a number of respects. First, there is not a one-to-one match between the ‘established’ grammatical classes and the different functions an item can have in discourse. The head of the NP, for instance, can be realized as a simple noun (a bird) or a compound noun (a blackbird), but also by adjectives (the British) and pronouns (you). Similarly, the secondary determiner function can be realized by adjectives (a similar story), adverbs (the then president), numerals (the third album), quantifiers (all the men), etc. (see Breban, Davidse, and Ghesquière 2011: 2691; §1.4.2). Second, linguistic elements can undergo syntactic-semantic changes as a consequence of which they may not only shift function but also word class in some of their uses, i.e. display layering in the sense of Hopper and Traugott (2003:124; §3.2) (e.g. Adamson 2000; Denison 2006; Breban 2010a). A well-known case is very which, along with a functional shift from attribution (‘true’) to degree modification (‘to a high degree’), has shifted from adjective to adverb.

14

A functional-cognitive model of the English NP

It is clear that a class-based approach to the NP cannot fully account for the synchronic variation and diachronic change associated with it. As Croft (2001: 8) notes, “[l]anguage is fundamentally DYNAMIC, at both the microlevel – language use – and the macro-level – the broad sweep of grammatical changes that take generations to work themselves out” [emphasis original]. Any valid model of the English NP should be designed in such a way as to be able to account for this synchronic and diachronic dynamicity. Linguistic analysis of specific items must therefore always take into account differences – distributional, syntactic, semantic, and pragmatic – relative to their function in the discourse (Croft 2001: 73). In this study it is assumed that a constructional approach of cognitivefunctional orientation such as advocated by Croft (2001) is best suited to account for this. In such an approach, not grammatical classes as such, but constructions are the basic units of analysis. Constructions are conceived of here as functional structures in which grammar and lexis are integrated with each other, i.e. as distinct form-meaning pairings (Croft 2001; Fried 2010), as illustrated in Figure 1. Note that this reading of ‘construction’ comprises linguistic units of variable sizes ranging from morphemes to multi-word strings. CONSTRUCTION

syntactic properties morphological properties phonological properties

FORM

symbolic correspondence (link) semantic properties pragmatic properties discourse-functional properties

(CONVENTIONAL) MEANING

Figure 1. The symbolic structure of a construction (Croft 2001: 18)

As shown in Figure 1, taking a construction or functional structure as starting point for linguistic description does not preclude attention to syntactic characteristics. On the contrary, it allows linguists to attend to the full

The functional make-up of the English NP 15

syntactic, semantic, and pragmatic diversity it displays, including scopal and collocational features.3 Grammatical categories and word classes are derived from and defined by the constructions in which they appear (Van Langendonck 1999; Croft 2001; see also Halliday [1961] 2002a). In a number of ways, cognitive-functional construction grammar incorporates, and further explicates, valuable elements from the functional tradition developed, amongst others, by Haas (1954), Firth (1959), Halliday (2002a), Matthews (1974) and linguists from the Prague school such as Danes (1964) (Fried, p.c.). Overall, the direction of linguistic description is from function to form, rather than from form to function. The remainder of this chapter will be devoted to the development of a dynamic, functional model of the English NP that can account for the synchronically and diachronically dynamic nature of NP structure and its component elements both in terms of syntax and semantics.4 Point of departure will be existing functional models such as those proposed by Halliday (1994) and Bache (2000), and the cognitive model proposed by Langacker (1991). 1.1. The functional make-up of the English NP 1.1.1. Halliday’s structural-functional account In Halliday’s (1985, 1994) view, the linguistic system encompasses three main functional-semantic components: the ideational, the interpersonal and the textual (see also Halliday and Hasan 1976: 26–27). First, the ideational component is concerned with representational semantics, with the expression of ‘content’. It offers the resources to describe referents. Second, the interpersonal component involves all speaker-hearer related meaning in language. All meanings related to the speech event and the deictic centre are interpersonal, as well as meanings reflecting the speaker’s personal attitudes toward, and evaluation of, what is being talked about. Third, the 3. According to Croft (2001: 179), collocational dependencies display both syntactic and semantic characteristics. 4. Note that the proposed functional-cognitive model of the NP is designed to account for canonical NPs, headed by a lexical noun possibly modified by prenominal (and postnominal) modifiers. It does not apply to, for instance, pronouns and proper names. The issue of canonical vs. non-canonical NPs will be looked at in more detail in § 1.1.3.

16

A functional-cognitive model of the English NP

textual component covers all discourse-related meaning (given–new, theme–rheme, topic–focus, etc.), including cohesive elements. Although interpersonal and textual meanings are to different degrees present in the NP, only the ideational component is, according to Halliday (1994: 190), necessary to develop an adequate functional model of NP structure. Importantly, the ideational component of language consists, in his view, of two subparts: the experiential, which is “more directly concerned with the representation of experience” and the logical, which “expresses the abstract logical relations which derive only indirectly from experience” (Halliday and Hasan 1976: 26). For the logical structure of the English NP, a dependency analysis is proposed in which the head and modifiers of the NP stand in a relation of recursive modification or subcategorization. In this univariate structure, all the elements of the NP are related by “the recurrence of the same function: α is modified by β, which is modified by γ, which is ...” (Halliday 1994: 193) (§2.1). The experiential structure, on the other hand, accounts for the distinct functions of the various component parts or constituents of the NP. It is conceived of not as a univariate but rather as a “multivariate structure”, i.e. “a constellation of elements each having a distinct function with respect to the whole” (Halliday 1994: 193). The remainder of this section will be devoted to a discussion of the NP on the basis of its experiential meaning, i.e. as an experiential constituency structure.5 As an experiential structure, the English NP, “taken as a whole, has the function of specifying a class of things … and some category of membership within this class” (Halliday 1994: 180). Importantly, despite the misleading term ‘Thing’, the main function of the head of the NP is to denote the class or type of entity talked about, not a distinct entity. It is typically realized by common nouns, which “are precisely what their name implies, common to a class of referents” (Halliday 1985: 168). This class can be further subcategorized by one or more of the functional elements Deictic, Numerative, Epithet and Classifier. Finally, additional subcategorization can be expressed by post-modifiers, all captured under the functional label

5. Building on and adapting Halliday’s (1994) model, McGregor (1997) has proposed an alternative analysis of the NP in which only the head fulfils an experiential role, namely that of Entity signifying a concrete physical entity. “NPs have trivial experiential structure, in the sense that they consist of a single experiential role” (McGregor 1997: 120), viz. the Thing or Entity. All the other elements of the NP are then better dealt with in other layers of analysis such as dependency.

Post-deictic

adjective

Deictic

determiner

numeral

Numerative

two Epithet1 (subjective attitude) adjective

splendid

Figure 2. The experiential structure of the English NP

same

those Epithet2 (objective property) adjective

old

noun or adjective

Classifier

electric

noun

Thing

trains

phrase or clause

Qualifier

with pantographs The functional make-up of the English NP 17

18

A functional-cognitive model of the English NP

Qualifier. Halliday’s analysis of the experiential structure of the English NP is represented in Figure 2. The Thing is “the semantic core of the nominal group” (Halliday 1994: 189) and all other elements, both preceding and following, are claimed to further characterize the Thing. First, deictic elements indicate “whether or not some specific subset of the Thing is intended” (Halliday 1994: 181). If so, they characterize it either demonstratively, i.e. in terms of proximity (this, that, those, etc.), or by possession (my, your, his, etc.) (Halliday 1994: 181). Second, the Post-deictic, “adds further to the identification of the subset” “by referring to its fame or familiarity [regular, well-known, notorious, etc.], its status in the text [above, aforementioned, etc.], or its similarity/dissimilarity to some other designated subset [similar, same, different, etc.]” (Halliday 1994: 183). Third, Numeratives, subsuming amongst others ordinal and cardinal numbers, and relative and absolute quantifiers, “indicate some numerical feature of the subset” (Halliday 1994: 183). Fourth, Epithets indicate “some quality of the subset”, either an objective, potentially defining property (e.g. old, long, blue, fast) or an expression of the speaker’s subjective attitude (e.g. splendid, silly, fantastic) (Halliday 1994: 184).6 Fifth, the Classifier “indicates a particular subclass of the thing in question, e.g. electric trains, passenger trains, wooden trains, toy trains” (Halliday 1994: 184). 1.1.2. Bache’s functional zone model A second account to which the functional-cognitive model developed in this study is greatly indebted is Bache’s (2000) functional zone model of the English NP. Rather than viewing NP structure as consisting of different constituent elements, Bache (2000) distinguishes a number of syntactic zones each associated with a particular communicative function, some of which can be further divided into a number of subfunctions. Figure 3 summarizes Bache’s (2000: 160–162, 239) model, indicating “the relationship between communicative function (shaded cells) and internal constituent structure (white cells)”.

6. Note that, although included in the experiential structure, the Epithet2 function actually conveys interpersonal meaning as it is defined in Halliday (1994), expressing the speaker’s evaluation of the entity referred to.

Figure 3. A functional zone model of the English NP (based on Bache 2000: 160–162, 239)

the no the her

determination

expression of meaning as things pre-head dependents head quantification modification categorization Mod. I Mod. II Mod. III specification description Classification very dull visit additional staffing, usual sound English stock own handsome naval officer many eager medical students other horrid psychological tricks

with the shy smile

to her parents academic or secretarial

(multi-functional)

post-head dependents

The functional make-up of the English NP 19

20

A functional-cognitive model of the English NP

Bache (2000: 159), very similarly to Halliday (1994), identifies the main function of the NP as referring to things: “[n]ominals enable speakers to code what they want to talk about as things”.7 This general function can be specified as encompassing three main subfunctions: determination, modification and categorization. The latter represents “the functional nucleus” of the NP: categorization is realized by the head noun, which “represents the referent as a member of a category of the things, persons, etc.” [emphasis original] (Bache 2000: 161). As in Halliday’s (1994: 180) and Langacker’s (1991; §1.1.3) views, the head of the NP thus identifies the type of entity referred to. Pre-head dependents further specify this type by means of determination and modification. Bache (2000) states that “[d]etermination is realized by articles, pronouns and genitive constructions, while modification is chiefly realized by adjectivals”. The modification function is further defined as encompassing specification, description and classification. Description is realized by ‘central adjectives’, i.e. gradable adjectives that “occur freely in both attributive pre-head position [e.g. these (very) funny plays] and in predicative position: these plays are funny” and “often serve as conjoints in coordination”, e.g. his ugly and fat opponent (Bache 2000: 235). Specification and classification, in contrast, are realized by ‘peripheral adjectives’, which cannot be graded (e.g. *the onlier solution, *very solar energy), used in predicative position (e.g. *The student is former, *The nomination is presidential), or coordinated with central adjectives (e.g. *interesting and primary elections) (Bache 2000: 235). Post-head dependents are argued to be multi-functional, as they can contribute both to the determination (e.g. the sudden death of my father), modification (e.g. her own handsome naval officer with the shy smile), categorization (e.g. no additional staffing, academic or secretarial) and complementation (e.g. this very dull visit to her parents) functions (§1.1.4). Interestingly, unlike in Halliday’s (1994) model, the quantification function is envisaged as spread over the entire NP structure. “The singular/plural distinction in nouns referring to countable entities (e.g. boy/boys, girl/girls, man/men, woman/ women, etc.) is at the very heart of quantification but by no means the only way of expressing this communicative subfunction” (Bache 2000: 161–162). Besides in the categorization zone, quantification is argued to also be realized by elements in the determination zone (e.g. these [plural], a [singular]), the modification zone (e.g. numerous, ten, little) and the multi-functional zone (e.g. students in great numbers). 7. Bache (2000) refers to noun phrases as ‘nominals’.

The functional make-up of the English NP 21

Despite its self-proclaimed “function-based” nature, it seems that Bache’s ‘functional’ model is in fact largely class-based. As a consequence, functionally closely related elements such as another and other are placed in distinct syntactico-functional zones – determination and specification respectively – due to the generally accepted determiner status of the former and adjective status of the latter. However, the idea of building functional generalizations into the model by the concept of ‘functional zones’ is a very useful one. Another important advantage of Bache’s (2000: 239) model is that it incorporates a notion of ‘gradience’: there is no strict separation between determination, modification and categorization but rather a continuum of values from determination to categorization: from the left determination fades into modification via specification and from the right categorization fades into modification via classification. In the middle we have modification at its purest: description.

This is reminiscent of Halliday’s (1994: 187) observation that “there is a progression in the nominal group from the element that has the greatest specifying potential to that which has the least”. However, as Bache’s functional zone model puts more emphasis on the continuity of the distinct functions instead of viewing them as distinct constituents with clear boundaries, it seems more flexible and better suited to accommodate the observed synchronic variation and diachronic changes in the NP than Halliday’s (1994) experiential model.8 1.1.3. Langacker’s cognitive-functional account Although in Halliday’s experiential model and Bache’s functional zone model there is not a one-to-one relation between the different functions and word classes (e.g. Halliday [1994] recognizes that adjectives can fulfil Post-deictic, Epithet and Classifier functions), they do still have a rather strong structuralist basis. Point of departure are the different constituent elements of the structure (i.e. form) which are only in a second step inter8. Brems (2007: 29), for instance, has argued that the diachronic and synchronic variation displayed by size nouns between head (e.g. a bunch of grapes) and quantifier uses (e.g. a bunch of drunken louts) is obscured in Halliday’s (1985: 173) analysis, which for both uses attributes Numerative function to the size noun on the experiential level and Head status to the second noun in the logical structure. Against this, Brems argues that the quantifier uses came about as the result of a diachronic shift of the size noun from head to modifier status.

22

A functional-cognitive model of the English NP

preted in terms of function. In contrast, the model of the NP proposed by Langacker (1991) is essentially built up around the functions it typically realizes. Langacker (1991), as one of the main theoreticians of Cognitive Grammar, has always stressed that it is essentially a functional grammar. His model of the English noun phrase, or nominal, is in fact more radically functional than that of Halliday.9 It is defined in terms of the four functions typically fulfilled by nominals – type specification, instantiation, quantification and grounding. In such an approach the functions, defined from a Cognitive Grammar viewpoint, have priority not only over word class, but also over the structural make-up of the NP. In Langacker’s (1991: 51, 53) words, “semantic function (rather than constituency) is the criterial factor for understanding th[e] internal organization” of the noun phrase, and as such one can “abstrac[t] away from any details of structural implementation”. According to Langacker (1991: 53, 54), “[t]he overall function of a nominal is to mention a thing”: “every nominal profiles a thing construed as an instance of some type and further incorporates some specification of quantity and grounding”. The overall function of the NP can thus be defined in terms of four more specific semantic functions – grounding, quantification, instantiation, type specification – which are realized from left to right in the NP, yet not necessarily separately. In ‘canonical’ NPs, i.e. NPs with common noun heads, the different functions can be symbolized separately. In the NP those three black cats, for instance, the modifier black and the noun cats contribute to the type specification, three fulfils quantification, and those has a grounding function (Langacker 1991: 54). It is with such canonical NPs that this study is mainly concerned. The functions of nominals can, however, also be realized non-canonically, meaning that “the semantic functions in question are not uniquely associated with distinct levels of constituency” (Langacker 1991: 54). Examples are, for instance, pronouns and proper names which incorporate the four functions in their base form. The four functions of the NP can be defined as follows. Type specification involves “an initial delimitation among the potential objects of thought, confining attention to a set of things regarded as equivalent in certain respects” (Langacker 1991: 53). In other words, type specification involves the description of a number of properties which determine a type 9. Like Bache (2000), Langacker (1991) uses the label ‘nominal’ for what is in this study referred to as a noun phrase.

The functional make-up of the English NP 23

of thing, without necessarily referring to a specific instance of that type. Common nouns, which generally realize the head of an NP, profile type specifications only. They denote a certain type of entity, not a particular instance. Any modifiers accompanying the head noun add to this type specification. “Each of them adding certain refinements and thus deriving a higher-order type specification of greater precision”, e.g. spoon, slender spoon, long slender spoon (Langacker 1991: 58–59). They do not serve to single out specific instances of a certain (sub)type. A common noun like spoon can only be associated with a specific instance when it is used in a full NP or nominal like a spoon or the two table spoons. When a specific instance or instances of a type are singled out, instantiation takes place (e.g. the spoon). Instantiation is realized through quantification – identifying the size or number of instances involved – and grounding – relating the instances to the ground, i.e. “the speech event, its participants, and its immediate circumstances” (Langacker 2002a: 7). “Grounding constitutes the final, criterial step in th[e] assembly” of noun phrases (Langacker 2002a: 7) and can roughly be equated with determination. As such, the main distinction is between type specification and the other three functions, which all work together to allow the hearer to make mental contact with the instance(s) of a general type of entity intended by the speaker. 1.1.4. A functional-cognitive model of the English NP I will now set out the eclectic functional-cognitive model of the English NP that will be assumed in this study, building on the models discussed in Sections 1.1.1 to 1.1.3. This model will serve as the theoretical-descriptive starting point for the case studies presented. Importantly, the different slots distinguished in the NP are to be understood functionally, as form-meaning pairings, and their succession in NP structure as well as their synchronic and diachronic interrelation should be viewed as a continuum. There is no one-to-one correlation between the functions distinguished and word categories, and specific linguistic elements can both synchronically and diachronically fulfil different functions in the NP, with or without correspondingly changing word class membership. Figure 4 represents the functional-cognitive model of the English NP as it will be referred to in the remainder of this study. The model is for reasons of convenience represented here as a flat structure but, as will become clear in the discussion of the different functions, it captures distinct scopal and modificational relations.

those

a

your

those

other

regular

utter

very

rather

really

bleached

lovely

nonbleached

nice

pretty

boring

long

small

little

electric

garden

legs

blackbird

trains

madness

flowers,

topic

head

categorization

objective classifier

descriptive modification

nounadjective-intensifier subjective intensifier

degree modification

modification

Figure 4. A functional-cognitive model of the English NP

such

all

the

secondary primary secondary

determination

instantiation of a type of entity

of yours

with pantographs

that linguists like to discuss annual and perennial

(multifunctional)

24 A functional-cognitive model of the English NP

The functional make-up of the English NP 25

The main function of the NP is to refer to one or more instances of a type of entity (a thing, person, place, etc.). NPs with common noun head do this by designating one or more instances of a type of entity (Langacker 1991). The type is realized by the categorization and modification zones of the NP. The instantiation of the type is realized through the determination zone. The structure of the English NP can then be envisaged as consisting from left to right of (i)

a determination zone, whose constituent elements give either identifying or quantifying information in a broad sense; (ii) a (pre)modification zone, which accommodates (a) descriptive modifiers which attribute properties and qualities to the entity referred to by the head, and (b) degree modifiers which measure the degree of gradable properties referred to by the descriptive modifiers (adjective-intensifiers) and/or the head of the NP (noun-intensifiers), and (iii) a categorization zone, grouping together (a) the head, which designates the type of which the referent of the NP is an instance, and (b) classifiers, which indicate a subtype of the type denoted by the head.

As noted in all functional approaches to the NP discussed, the type referred to by the head can be further specified by both pre- and postmodifiers. Post-head dependents, which as claimed by Bache (2000; §1.1.2) are multifunctional, fall largely outside the scope of this study. In the following sections, I will introduce the different prenominal functions, with attention for their semantic and syntactic characteristics. Attention will also be given to the relation of these functions or functional zones to subjectivity and intersubjectivity. These notions are discussed in detail in Chapter 2. To keep the discussion of the different slots as clear and accessible as possible, the distinct zones will be discussed as they occur in the NP structure from right to left. 1.2. Categorization If we do not take the multifunctional postnominal zone into account, the categorization zone is situated at the rightmost end of NP structure. In the model proposed here, the categorization zone subsumes both the lexical head noun and the classifier function. In the functional model of the NP proposed by Bache (2000), which takes into account word class criteria, the

26

A functional-cognitive model of the English NP

categorization zone consists exclusively of the head noun and is distinguished from the classification zone realized by nouns or adjectives. By contrast, I will propose to collapse Bache’s (2000) categorization and classification zones into a single categorization zone on functional grounds (see also Teyssier 1968; Breban 2010a). All the elements contributing to the delineation of the class of entities referred to are then viewed as constituting one functional-structural zone, irrespective of word class membership. 1.2.1. The head of the NP The lexical head noun designates the type of which the referent of the NP is an instance (Langacker 1991: 55–58). It may be realized as either a simple noun (topic, flowers, trains) or a compound noun, most commonly consisting either of adjective + noun (blackbird, old man ‘father’), or noun + noun (toothpaste, windmill). Although compound nouns are still recognisable as built up of two component parts, they are considered to form one lexical word. The main argument for their compound status is that the component parts cannot “enter separately into relations of coordination and modification” (e.g. *ice- and custard-creams, *ice-lollies and -creams, *two iceand 10 custard creams, *crushed ice-cream, *ice- Italian cream) (Huddleston and Pullum 2002: 449). In other words, compounds satisfy Robins’s (1980: 148–154) grammatical criteria for forming one word: they are internally inseparable but externally separable. No modifiers can intervene between the component parts and modifiers preceding the compound noun modify the entire sequence (e.g. delicious ice-cream). The semantic and syntactic unity of compounds is reflected in the fact that they are often listed as separate entries or special units in dictionaries, with mention of their specialized meanings. The compound old man, for instance, has acquired the specialized meaning of father. In the literature there is often an implicit consensus that the head of an NP conveys objective meaning. However, it has to be pointed out that the head – simple or compound – can also express more subjective meaning. First, degree nouns (Bolinger 1972) or gradable nouns (Paradis 2000: 243) like idiot, mess and splendour are clearly inherently evaluative and hence (ideationally) subjective in meaning (§2.2.3). Second, as I will argue in Chapter 8, compound nouns can also convey either objective (blackbird, toothpaste, old man) or subjective meaning (old fogey, little sod).

Categorization

27

1.2.2. Subcategorization The general type specifications provided by the head of the NP may be further subclassified by nominal or adjectival classifiers (Halliday 1994: 184–186), as in garden flowers and electric trains respectively. Such classifiers, which are generally considered to be very objective modifiers, further specify the type or class of thing denoted by the head (Sinclair 1990: 66; Bache 2000: 235). Unlike descriptive modifiers (§1.3.1), adjectival classifying modifiers typically do not allow degree modification (*very/more electric trains) and are restricted to attributive position (*the trains are electric). They do, however, allow for a marked form of ‘gradability’, pertaining to the prototypicality of class membership and delineation of the class in question. Adverbs such as almost, exclusively, largely, purely and others indicate to what degree the classifier applies (Sinclair 1990: 95–96). Largely in They worked in a largely migrant community (WB brbooks), for instance, indicates that the majority of the community are migrants. Note that classifier-noun combinations can become so entrenched in speakers’ minds that the contrast with compounds is reduced to a minimum. Although he does not view simple adjective-noun combinations containing a classifier as compounds, Halliday (1994: 185) acknowledges that “the line between a compound noun and a nominal group consisting of Classifier + Thing is very fuzzy and shifting”. However, semantically, classifiers tend to be organized in mutually exclusive and exhaustive sets (taxonomies), e.g. electric trains/steam trains, whereas compounds are not. This notional contrast is observed in e.g. Huddleston and Pullum (2000: 1645) and Vandelanotte (2002: 235), who states that [a] compound designates a self-contained, ‘self-defining’ concept, whereas a Classifier-noun sequence designates a subset of the head noun defined only in contrast with other contextually relevant subsets

So whereas the classifier-noun sequence criminal law implicitly evokes a contrast with other types of law such as civil law, the compound blackbird does not designate a subtype in contrast with other types of bird. In order for a sequence to receive compound status it has to undergo a process of semantic specialization. Other purported tests to distinguish classifier-noun sequences from compounds such as spelling as one word (e.g. physiotherapist, but also clinical therapist or family therapist) and stress have proven to be unreliable (see e.g. Giegerich 2009, 2011; §8.1.1).

28

A functional-cognitive model of the English NP

1.3. Modification In the proposed model, the modification zone is positioned following the determination zone and preceding the categorization zone and is further divided into a descriptive modification zone and a degree modification zone. The descriptive modification zone correlates quite nicely with Halliday’s (1994) Epithet1 and Epithet2 slots (§1.1.1) and corresponds closely to Bache’s (2000: 235) central adjectives of the Mod.II description zone (§1.1.2). In Langacker’s (1991: 53) model of the NP descriptive adjectives, classifying adjectives and head noun, as in excellent convention site, are all viewed as contributing to the type specification: they “specif[y] the basis for identifying various entities as being representatives of the same class”. Although I agree, I think it is important to note that the different modifiers contribute to the type specification in different ways. On the one hand, it is true that the distinction between descriptive and classifying modifiers is not a discrete one. Descriptive modifiers can be used as ad hoc classifiers, e.g. big and small in (10) and, diachronically, they can shift to classifier use. Adamson (2000: 59–60) discusses the development of the adjective criminal from a descriptive modifier, as in a criminal tyrant ‘guilty of a crime’, to a classifying modifier, as in criminal law (= not civil law). (10)

Hazel changed her mind and got out the big wine glasses instead of the small ones. (WB brbooks)

On the other hand, there are a number of arguments in support of a distinction between descriptive and classifying modifiers in a functional model of the NP. First, whereas descriptive adjectives attribute properties to the NP referent itself, classifiers modify only the type of NP referent (cf. Quirk et al. [1972: 1325] on inherent versus non-inherent modification). This distinction is reflected in their different behaviour in terms of predicative alternation. Only descriptive modifiers can be used predicatively, classifying modifiers cannot, e.g. that train is fast, but *that train is electric. Second, descriptive and classifying modifiers engage in different modification relations with the head. Whereas the former independently modify the head, e.g. beautiful little white flowers (flowers that are beautiful, flowers that are little, flowers that are white), classifiers engage in recursive modification, e.g. [American [democratic [political movement]]] (§2.1). These morpho-syntactic differences between descriptive and classifying modifiers warrant positing a non-discrete, flexible boundary between the descriptive modification and categorization zones.

Modification

29

Whereas descriptive modifiers receive considerable attention in Halliday’s (1994) and Bache’s (2000) descriptions of the NP, the degree modification zone is underdeveloped in the two accounts. Halliday (1994: 184) briefly mentions that attitudinal Epithets “tend to be reinforced by ... intensifiers”, but does not discuss the matter in more detail. In Bache’s (2000: 160, 242) model, degree modifiers like very in the very dull visit to her parents and lovely in lovely soft hands are categorized in the description zone together with the adjectives they modify. Although its position in NP structure is something of a challenge, I will argue that there are clear functional reasons for positing a distinct degree modification zone in addition to a descriptive modification zone. 1.3.1. Descriptive modification Descriptive modifiers, typically realized as adjectives, attribute a certain quality or property to the instances referred to by the NP. In other words, they designate the degree of a quality (e.g. a tall man, a beautiful flower, a closed door). Morpho-syntactically, descriptive modifiers are characterized by their ability to occur both in attributive, prenominal position, and in predicative position, e.g. the man is tall, the flower is beautiful, the door is closed. Also, descriptive modifiers typically allow for degree modification, either through degrees of comparison (e.g. the taller/tallest man, the more/most beautiful flower) or by means of degree modifiers (e.g. a very tall man, a rather beautiful flower, an almost closed door). These two grammatical tests are generally acknowledged to distinguish descriptive adjectival modifiers from determining and classifying ones (e.g. Bolinger 1967; Sinclair 1990: 65; Halliday 1994: 184). As the possibility of predicative alternation and degree modification are the features most commonly associated with adjectives, descriptive modifiers are labelled ‘core’ (Quirk et al. 1972: 231–234) or ‘central’ (Bache 2000: 235) adjectives.10 10. Note that although there is a strong tendency for descriptive modifiers to allow predicative alternation and degree modification, there are of course counterexamples. Bache (2000: 230–231), for instance, mentions the impossibility for noninherent descriptive modifiers to be used predicatively. Non-inherent adjectives “relate by way of association to the meaning of the head noun rather than ascribing a property to the referent as such. ... Thus, for example, an old friend is not necessarily a friend who is old but someone with whom one has had a friendship of long standing, and a heavy sleeper is someone who sleeps heavily, not a sleeping person

30

A functional-cognitive model of the English NP

Descriptive modifiers can convey a wide array of qualities and properties. In an attempt to bring order in this very diverse set of adjectives two important cross-categorizations have been proposed in the literature. These two classifications will be discussed in the following sections. 1.3.1.1. Objective and subjective descriptive modifiers There is a tradition of long standing to classify central adjectives in terms of their semantics, ranging from objective to subjective types. Importantly, these semantic classifications – both language-specific (Quirk et al. 1972, 1985; Huddleston and Pullum 2002) and typological (Hetzron 1978; Dixon 1982) – are always conceptualized as clines of semantic categories, leading from more subjective to more objective properties. As Hetzron (1978: 178) puts it, there is “not a discrete polar distinction but a gradual one, where ‘relatively more subjective’ is the relevant criterium”. In the literature a number of fine-grained semantic typologies of gradable adjectives have been proposed. Dixon (1977, 1982: 24), for instance, distinguishes seven semantic subtypes, i.e. value, dimension, physical property, speed, human propensity, age and colour. Tucker (1998: 215– 216) proposes a division into eight semantic categories: likelihood, affected state, affective categorization, dimension, quality, age, physical quality and colour. Such semantic categorizations, however, run the risk of continually being in need of refinement, consequently leading to an overabundance of categories. Hetzron (1978: 180), for instance, has suggested a refinement of the Dimension category originally proposed in Dixon (1977) into the subcategories Weight, Length, Height, Width and Thickness. In this study, a more restricted dual objective-subjective classification will be adhered to, acknowledging that the objective and subjective categories are not divided by a clear-cut boundary but rather represent values on a cline or continuum. Typically, objective descriptive modifiers indicate objectively recognizable, purely descriptive and potentially defining qualities (for example in the superlative or with a specific determiner, e.g. the youngest person) (e.g. long, small, green), whereas subjective descriptive modifiers express the speaker’s attitude towards the instance referred to by the NP (Halliday 1994:184). In the latter case, they are of an evaluative, often affective, nature, expressing a subjective, personal stance of the who is heavy.” Despite their descriptive function, these adjectives can only be used prenominally.

Modification

31

speaker (e.g. beautiful, terrific, lovely). Note also that one adjective can fulfil both objective and subjective descriptive modifier functions. In (11) old, for instance, receives an objective reading as it “relat[es] to properties which are (relatively) inherent in the head of the noun phrase, visually observable, objectively recognizable or assessible” (Quirk et al. 1972: 924), i.e. worn with age or use. In (12), in contrast, old functions as a subjective descriptive modifier, affectively referring to a long-standing relationship between people. As such the adjective is “concerned with what is relatively a matter of opinion, imposed on the head by the observer, not visually observed and only subjectively assessible” (Quirk et al. 1972: 926). (11) (12)

Like a Welsh terrier with an old sock, Ms Lawley refused to drop the matter. (CB times) the star must jealously guard his privacy, even if it means not returning phone calls to old friends. (CB times)

How subjective are subjective descriptive modifiers? I would like to argue that they are ideationally subjective in the sense of De Smet and Verstraete (2006) (§2.2.3). Subjective descriptive modifiers convey descriptive, propositional meanings and form possible recognition criteria for the instance in question. At the same time, these meanings are situated in the speaker’s beliefs and attitudes as they convey his/her perspective on or evaluation of the entity referred to. For instance, when someone refers to a daisy as a beautiful flower, he or she evaluates the entity in question by assigning the property of beauty to it, a property assessable only on a subjective basis. Importantly, as they describe “speaker-internal content” (De Smet and Verstraete 2006: 386), subjective descriptive modifiers are subjective on an ideational level only and not on the interpersonal level. Rather than simply involving “the enactment of speaker position” (De Smet and Verstraete 2006: 387), they convey descriptive, propositional content, albeit content “situated in the speaker’s subjective belief-state/attitude” (Traugott and Dasher 2002). De Smet and Verstraete (2006) have put forth a number of syntactic tests to distinguish between ideational and interpersonal subjectivity. All of these tests can be successfully applied to subjective descriptive modifiers: they typically allow for degree modification (13a), predicative alternation (13b) and coordination with objective descriptive modifiers (13c), and they can be focused. They can serve as the focus of a whquestion (13d) and as the focus of negation (13e).

32

(13)

A functional-cognitive model of the English NP

a. b. c. d. e.

a more/very beautiful flower that flower is beautiful a small, but beautiful flower What kind of flower is that? A beautiful flower. That is not a beautiful flower, it is an ugly one.

Note that objective descriptive modifiers allow for all of the same tests put forth in De Smet and Verstraete (2006), as is shown for tall in a tall building in (14). (14)

a. b. c. d. e.

a taller/very tall building that building is tall a narrow, but tall building What kind of building is that? A tall building. That is not a tall building, it is a short one.

That both objective and subjective descriptive modifiers allow for all of the syntactic tests is due to the fact that the tests in fact contrast propositional and non-propositional meaning (see also Visconti 2013).11 The finer distinction between subjective and objective descriptive or propositional modifiers corresponds nicely to Traugott’s (1989) characterization of externally and internally propositional meaning (§2.2.3). The propositional component, in her view, comprises all “truth-conditional relations” which are “subject to referential verification” (Traugott 1982: 248), and can be further divided into “meanings based in the external described situation”, e.g. objective descriptive meanings, and “meanings based in the internal (evaluative/perceptual/cognitive) described situation”, e.g. subjective descriptive meanings (Traugott 1989: 34). The latter are still descriptive, but the recognition criteria involved are of a more internal, evaluative nature and hence speaker-related.

11. Although it is not explicitly stated in De Smet and Verstraete (2006), I believe that ideational subjectivity is indeed understood to be inherently propositional while interpersonal subjectivity is inherently non-propositional. The former is defined as figuring “in the content of the utterance, i.e. what is described”, whereas the latter simply enacts “the way the speaker takes an interactive position with respect to that content” (De Smet and Verstraete 2006: 386) [emphasis added]. As stated by Cruse (2000: 46), propositional, descriptive and ideational are all terms used to refer to the same kind of meaning.

Modification

33

1.3.1.2. Bounded, unbounded and extreme descriptive modifiers Cross-cutting the objective-subjective distinction, descriptive modifiers can also be categorized in terms of how they are construed conceptually. In their property-describing uses, adjectives are generally accepted to be construed as either bounded or unbounded (Bolinger 1967, 1972; Paradis 2000, 2001). According to Paradis (2000, 2001), bounded adjectival meanings are conceptualized as complementaries, i.e. properties that either apply or do not apply, e.g. dead - alive, while unbounded adjectival meanings are antonyms, i.e. properties inherently conceptualized as degrees, e.g. long - short. The difference is drawn out by the modifiers with which they can be used. Bounded adjectives are modified by proportional modifiers expressing maximality, e.g. wholly dead, or approximation, e.g. almost/half dead (Kennedy and McNally 2005: 352–354). The property described by these adjectives is assessed by comparing it to the boundary they imply, i.e. as either reaching or not reaching that boundary or as going some way towards reaching it. This shows that the meaning of adjectives such as dead is inherently bounded. Unbounded adjectives cannot be modified by proportional modifiers: *wholly long, *half short. Instead they take modifiers that measure the actual degree of the quality on a scale with some form of assumed measure units, e.g. rather/very long. In other words, these properties are inherently conceptualized as degrees which are measured on a scale going up or down from a reference point, e.g. highly/far from entertaining. In addition, Paradis (2000: 52) has drawn attention to extreme adjectives such as terrible and excellent that are in some respects on the borderline between the bounded and unbounded classes as they indicate the ultimate points of a scale. As pointed out by Bolinger (1967: 4), adjectives representing extremes can be pulled off the ultimate points of the scale and be compared, i.e. “laid out on a scale”, e.g. more excellent, most excellent. Because this is their marked construal, Paradis classifies these adjectives as basically bounded. Taking the opposite track, Martin (1992: 303) focuses exclusively on the gradable nature of oppositions on a scale (Lyons 1977: 289), under which he subsumes both the ultimate and intermediate values. He characterizes all the adjectives on a scale such as excellent, good, fair, poor, awful as non-binary gradable oppositions. He thus recognizes, with reference to Lyons (1977: 289), binary scalar adjectives, e.g. big – small, as well as non-binary scalar (or unbounded) adjectives, e.g. good, fair, poor, etc. In his approach, unlike in Paradis’s, unbounded adjectives are not characterized by exclusive association with antonyms. Rather, unbounded bina-

34

A functional-cognitive model of the English NP

ry adjectives relate to antonyms, e.g. short, long, and unbounded nonbinary adjectives to contrasts, e.g. hot, warm, cool, cold. Nuancing the positions of both Paradis and Martin, Ghesquière and Davidse (2011) have argued that adjectives describing values on a scale carry in them the possibility of being construed either as points (bounded) or as ranges (unbounded) on a scale. In a similar vein, Paradis and Willners (2006: 1055, 1074) have argued that descriptive modifiers are not intrinsically either bounded or unbounded. Rather boundedness is a matter of construal. Although adjectives may have a certain bias towards one or the other reading, “it is almost always possible to coerce a bounded reading into an unbounded reading” (Paradis and Willners 2006: 1074) (see also Kamoen et al. 2011). In accordance with Paradis’s observations, it is recognized that the unbounded construal may be more marked for certain points on the scale such as the extremes. For instance, combinations of degree modifier and descriptive modifier such as very excellent and more excellent appear more marked than the scalar construal very good, with good indicating an intermediate point on the scale. However, the two construal possibilities have to be recognized for both extreme and intermediate values on a scale. Invoked as points, the adjectives are used as such (without modification) or with modifiers such as wholly, just, strictly, e.g. The work is (strictly) fair/good/excellent. Activated as ranges on a scale, they can be used in the degrees of comparison or prefaced by modifiers such as very, e.g. This new version is less weak/very good/very excellent. In sum, against the background of observations by Bolinger (1967), Lyons (1977), Martin (1992) and Paradis (2000, 2001), a two-way distinction can be proposed for descriptive modifiers into unbounded construal of adjectives, denoting a range on a scale, and bounded construal of adjectives, denoting points on a scale conceptualized in terms of 'either...or'. 1.3.2. Degree modification Besides descriptive modifiers, the NP can also contain prenominal modifiers which modify or measure the degree of qualities conveyed by a descriptive modifier (e.g. very nice) and/or a gradable head noun (e.g. a complete idiot, such a big mess) (Bolinger 1972). Depending on their scope relation, these degree modifiers can be categorized as functioning either as adjective-intensifiers or noun-intensifiers. Note that whereas in Bache’s (2000) model of the English NP, descriptive modifiers and degree modifiers are lumped together in the Mod.II description zone (Figure 3), they are as-

Modification

35

signed their own (sub)zones in the functional-cognitive model developed here (Figure 4). 1.3.2.1. Adjective-intensification Adjective-intensifiers, typically realized by adverbs, modify or indicate the extent of a quality or property denoted by a descriptive modifier they immediately precede. Adjective-intensifiers like really in a really nice book thus have scope over and modify only the following adjective, even when the NP is headed by a gradable noun, as in a really nice surprise. In NPs with a string of prenominal descriptive modifiers, some or all of them can be modified by distinct adjective-intensifiers, as in (15) and (16). Such multiple degree modification cannot in a straightforward manner be represented in the flat structure of the NP as it was proposed in Figure 4. (15) (16)

A Few Very Exciting, Humbling, Scary, and Rather Personal Details (http://mikeduran.com/?p=13944, accessed 18 August 2011) A Very Short Fairly Interesting and Reasonably Cheap Book About Studying Organizations (title of a book by Christopher John Grey)

Semantically, two types of adjective-intensifier can be distinguished, one bleached and one non- or less bleached. Bleached intensifiers have lost most of their original meaning and are reduced mainly to their intensifying function. For example, the original propositional, truth-related meaning of very has bleached completely in favour of the intensifying reading ‘to a high degree’ (e.g. Traugott 1989; Breban 2010b). Other typical bleached adjective-intensifiers are truly and fairly. The point that there are also adjective-intensifiers that are much less semantically bleached was made by Dixon (1982: 15ff). He has cross-linguistically observed the existence of a semantic category of ‘value adjectives’, which qualify some other adjective “which is taken out of its normal place in the ordering and placed immediately after the VALUE adjective” (Dixon 1982: 25). Dixon thus observes the qualifying meaning of these adjectives while at the same time recognizing the fact that they modify not the head noun but a descriptive modifier. I would like to argue that these adjectives are hence best analyzed as nonbleached or, more accurately, not-fully-bleached adjective-intensifiers. This ties in with Adamson’s (2000: 54) position. She argues that lovely in lovely legs has retained “some of the original affect” and now roughly has the meaning of ‘very + approval’”. Other examples are good in a good strong

36

A functional-cognitive model of the English NP

box (Dixon 1982: 26) and nice in nice big houses (Adamson 2000: 54). These adjectives, although they still have an inherent evaluative descriptive meaning, function as degree modifiers of the following descriptive adjectives.12 In Dixon’s view, NPs with – what he calls – value adjectives (e.g. a good new fast car) can be characterized neither as displaying recursive modification (*there’s a fast car, the fast car is a new one, the new fast car is a good one) nor as displaying independent modification (the car is fast, the car is new, the car is good). Rather the sequence is best interpreted as ‘a fast car which is new and therefore good’. Similarly, the NP lovely old trees refers to trees which are old and therefore lovely. 1.3.2.2. Noun-intensification In the literature, noun-intensifiers have been studied under many different guises, including ‘emphasizers’ (Quirk et al. 1972, 1985; Sinclair 1990) and ‘reinforcers’ (Paradis 2000; Feist 2009). This proliferation of terms hints at a terminological problem which is not easy to solve. In this study, I have opted to use the term noun-intensifiers. Whereas adjective-intensifiers have scope only over the quality described by the following adjectives, noun-intensifiers have scope over and modify the degree of all gradable qualities in the whole NP. These qualities may be conveyed by the head noun (17) or the head noun and modifiers (18). However, they may also be conveyed only by descriptive (and degree) modifiers, as in (19). Contrary to what the term may seem to imply, noun-intensifiers may thus not only have scope over the noun but over all the elements contributing to the type

12. Dixon (1982: 26) argues that value adjectives can also occur without a following descriptive modifier in the NP and then “effectively qualif[y] some implicit non-VALUE adjective, which itself qualifies [the head noun], rather than [the value adjective] directly qualifying [the head noun]”. In his view, the NP a good box is “essentially an elliptical version of a construction which includes yet another adjective following the VALUE one, which presents the reason of the ‘goodness’ of the head noun: noun is ‘good for X’ where X ranges over all those properties for which it could be considered good by speaker and hearer” (Dixon 1982: 26). With Adamson (2000: 44), I believe such an analysis to be rather far-fetched. In the theoretical framework outlined for this study, good would simply receive a subjective descriptive modifier analysis. When value adjectives are understood as adjective-intensifiers they necessarily require a following adjective.

Modification

37

specification, albeit only in terms of gradability.13 As examples (20) and (21) show, noun-intensifiers can also have scope over adjective-intensifiers. (17) (18)

(19) (20) (21)

I idolized him. Everywhere he went I followed. He thought I was a terrible nuisance. (WB brbooks) Erm and at the end erm coming back to the theme you were talking about earlier in fact where the bride-to-be goes off with him leaving behind the Liverpool background which I felt was quite a big snub at the people you're writing about and quite a lot of anger being generated at those people. (WB brspok) Robson and his team flew out to Kiev last night aware of what a huge week this will be for the club. (WB sunnow) Mexico is such a very big country (WB brbooks) Absolute Very Best of the World's Best Ever Beer Songs (http://www.amazon.com/Absolute-Very-Best-WorldsSongs/dp/B000BYAD52, accessed 22 August 2011)

Note that whereas adjective-intensifiers are typically realized by adverbs or value adjectives, noun-intensifiers are recruited from at least three grammatical source domains, each encompassing various semantic fields. First and most typically, noun-intensifiers can be realized by (originally fully, lexical descriptive) adjectives, e.g. pure nonsense, a real looker, a true success, a complete idiot, and utter darkness. Unlike ‘central’, descriptive adjectives, noun-intensifying adjectives typically cannot be used in predicative position and resist degree modification. Second, Bolinger (1972) has drawn attention to noun-intensifiers realized by more grammatical, originally identifying, elements like such in such a mess and too much in too much of a miser. Finally, a so far largely unexplored syntactic source of noun-intensifiers are (modal or other) adverbs, as in quite the gentleman, rather an idiot. In this study, attention will go to both adjectival and ‘determiner’ realizations of noun-intensifiers. Adverbial noun-intensification will have to be explored in more detail in future research.

13. Another option would be to use the label NP-intensifier, but this could create the impression that they have scope over elements from the determination zone as well which is clearly not the case. The label type specification-intensifiers would be equally unfortunate not only because of its length but also because in Langacker’s (1991) framework type specification is not exclusive to the NP but also inherent to verb phrases.

38

A functional-cognitive model of the English NP

1.3.2.3. Structural-semantic types of degree modification In the literature, the focus has been on adjective-intensification rather than noun-intensification. A number of linguists have provided (sometimes very detailed) semantic classifications of bleached adjective-intensifiers, mostly based on the semantic field from which these intensifiers are recruited (e.g. Bäcklund 1973; Ernst 1984; Allerton 1987; Lorenz 2002). Spitzbardt (1965), for instance, distinguishes no less than 19 semantic fields or categories of intensifying adverbs (e.g. approximation [almost], pettiness [barely], totality [absolutely], specialty [abnormally], fascination [arrestingly], etc.). A more workable semantic classification is arrived at when focusing on the orientation of the degree modifier function of bleached adjectiveintensifiers as expressing either positive degree, i.e. above a certain norm, or negative degree, i.e. a slight or barely perceptible degree (Traugott 2006: 338) (e.g. Quirk et al. 1972, 1985; Paradis 1997; Huddleston and Pullum 2002; Nevalainen and Rissanen 2002). Importantly, such a classification can be extended to accommodate not only adjective-intensifiers but nounintensifiers as well. Degree modifiers with a negative orientation, downtoners, have a general lowering effect and can be subdivided into approximators (almost happy, an almost lawyer), compromisers (more or less content, rather a disappointment), diminishers (partly built, a bit of a bore) and minimizers (hardly bigger, a mere child) (Quirk et al. 1972: 443). Degree modifiers with a positive orientation, amplifiers, can be further divided into maximizers, such as complete(ly) and total(ly) (e.g. completely/totally full, complete/ total failure), which “denote the upper extreme of the scale” and boosters, which “denote a high degree, a high point on the scale”, e.g. very happy, a terrible bore (Quirk et al. 1972: 444). Both positively and negatively oriented degree modification can be defined with reference to scales as “intensifiers that scale the meaning of the modified element upwards from an assumed norm, and those that scale it downwards” respectively (Nevalainen and Rissanen 2002: 361). As noted by Kennedy and McNally (2005) and Traugott (2006), degree modification or intensification is essentially scalar.14,15 The function of degree modifiers is to “measure” the degree of properties in terms of “points or intervals partially ordered along some DIMENSION” or scale (Kennedy and McNally 2005: 349). On the ba14. Kennedy and McNally’s (2005) study applies primarily to degree modification of adjectives, but allows for extension to degree modification of nouns. The same goes for the observations concerning degree modifiers made in Traugott (2006). 15. Degree modification and intensification are used interchangeably in this study.

Modification

39

sis of the exact structure of the scale involved, a dual structural classification of degree modification can be set up (Kennedy and McNally 2005; Tribushinina 2008), cross-cutting the semantic positive-negative classification. The first type of degree modification invokes “open scales” (Kennedy and McNally 2005: 349), as it involves measuring the actual degree of a quality on a scale with some form of assumed measure units. This is the mode of degree modification found in very long (of an adjective) and a mere trifle (of a noun).16 In these expressions the size implication evoked by the word being modified is further enhanced or reduced by the degree modifier. The degree modifier activates a range going up or down from a reference point on a scale defined by measuring units, not by maximum or minimum values. Therefore, it is called an open scale degree modifier. The second way of modifying the degree of properties invokes “closed scales” (Kennedy and McNally 2005: 353), as it involves comparing the degree of a quality to a boundary as either approximating or reaching it. This boundary may be either the maximum or the minimum, and closed scales may be defined by either an upper or lower boundary, or by both. Closed scale degree modifiers “calculate differences relative to minimum and maximum values on the scale” (Kennedy and McNally 2005: 353) of properties conveyed by adjectives, e.g. fully/half dead, more dead than alive, or by nouns, e.g. a complete/half failure, the completest success. We can speak of a scale here, because the various values (half, almost, fully, etc.) that can be indicated for the difference between actual degree and maximum or minimum clearly involve scalarity. The semantic positive-negative oriented classification intersects with the structural classification into open and closed scale degree modifiers. Whereas approximators, diminishers and maximizers typically invoke closed scales, the other semantic types typically invoke open scales. Degree modification, defined as the measuring of qualities invoked by the type specification components of the NP, is very closely related to quantification (see §3.4.2.2.1 for the diachronic relation between quantification and noun-intensification). The mechanisms of open and closed scale degree modification closely parallel the abstract measuring operations involved in the two generally recognized basic types of absolute and relative quantification (Milsark 1977; Langacker 1991; Davidse 1999) (see Ghesquière and Davidse 2011). Absolute quantifiers measure the size of 16. Note that semantically these intensifiers can be classified as a booster and downtoner respectively.

40

A functional-cognitive model of the English NP

some set or mass by offering a direct description of its magnitude (Milsark 1977: 23; Langacker 1991: 82–83). This direct measurement refers to an implied scale with measure units, either precise ones such as the cardinality scale (three/a hundred and one mice) or more schematic units (few/many mice). Absolute quantifiers thus imply an open quantification scale without an upper boundary. Relative quantifiers make a quantitative assessment relative to a reference mass or reference set (Milsark 1977: 23; Langacker 1991: 108). They compare the mass or set they actually designate to a reference mass/set (Langacker 1991: 83), profiling the actually predicated mass/set as a proportion (part or all) of the reference mass/set, e.g. most/all parents. It is this comparative element, not an intrinsic indication of size, that is central to their quantifying meaning. Relative quantifiers hence imply a bounded, or closed, quantification scale. Conceptually, the main difference between degree modifiers and quantifiers is then that whereas the former measure the degree of a quality attributed (adjective-intensifiers) or inherent (noun-intensifiers) to the instances referred to, the latter measure the size/quantity of the instances themselves. The relation between degree modification and quantification is also obvious in Clausner and Croft (1999), who recognize both phenomena as crucially linked to scalarity or the SCALE image schema. Scalarity, they argue, “permeate[s] the whole of human experience”, “including (i) number, such that we can have more, less or the same number of objects, (ii) an amount of substance, (iii) a degree of force, (iv) the intensity of a sensation” (Clausner and Croft 1999: 14). They explicitly mention both the degree modifier very as well as ordinal and cardinal numbers as being inherently (and solely) scalar. In addition, they argue that the degree modifier very and the ordinal numbers on the one hand and cardinal numbers on the other hand are construed as locations and configurations in the SCALE image schema respectively. Very and the ordinal numbers “indicate a specific position on the scale relative to a reference point” (Clausner and Croft 1999: 19). In the case of very, this is “a location profiled on the scale further along than the reference value”, which is determined by the gradable adjective being modified (Clausner and Croft 1999: 18). With ordinal numbers, the reference point is that “specified as first on the scale” (Clausner and Croft 1999: 19). In contrast, cardinal numbers profile configurational concepts, independent of their exact position on the scale. “Cardinal numbers such as three or eighty profile an amount (i.e. a magnitude, a number of units) that remains constant no matter where in the SCALE domain the units are counted” (Clausner and Croft 1999: 18–19).

Modification

41

1.3.2.4. Degree modification and subjectivity That degree modifiers, both of adjectives and nouns, are subjective is generally accepted (e.g. Traugott 1982, 1989, 1995, 2006; Paradis 1997, 2000; Athanasiadou 2007). As Traugott (2006: 343) notes, they are “subjective in the sense that they involve the speaker’s assessment and evaluation of intensity, position on a scale, ordering of alternatives, etc.”. The question arises, however, just how subjective degree modifiers really are. Are they subjective on the ideational level only, or are they interpersonally subjective?17 De Smet and Verstraete (2006) have proposed syntactic tests to distinguish ideationally (or propositionally) subjective items from interpersonally (or non-propositionally) subjective ones, i.e. focusability, whinterrogation and negative scope (§1.3.1.1, §2.2.3). “Focus, interrogation and negation all serve to pick out some aspect of content, either by highlighting it or denying it” (De Smet and Verstraete 2006: 386) and hence apply only to ideationally subjective and not to interpersonally subjective items. How do adjective-intensifiers and noun-intensifiers behave in terms of the proposed syntactic tests? First, both types of intensifier can function as the focus of a whquestion. Adjective-intensifiers can be questioned by how adjective?, nounintensifiers answer the question how much (of a) N?, as in (22a) and (22b) respectively. Second, both adjective-intensifiers and noun-intensifiers can fall under the scope of a negator, as in (23a, b). Third, they allow for additional modification by degree modifiers, as in (24a, b). Last, as illustrated in (25), noun-intensifiers, which are typically realized as adjectives, allow for coordination with other non-intensifying adjectives, albeit only marginally. (22) (23)

a. b. a.

b. (24)

a.

How strange is that sound? It is very strange. How much of stranger is he? He is a complete stranger. Its not very strange...based on my research, many girls all over the world fall in love with their teachers (http://www.fanpop.com/spots/love/answers/show/172772/ver y-strange-fall-love-teacher, accessed 22 August 2011) not a complete stranger, more like someone she'd known long ago but whose name she'd forgotten. (WB brbooks) a very very strange sound (WB brbooks)

17. For a discussion from a Functional Discourse Grammar perspective see Van de Velde (2007).

42

A functional-cognitive model of the English NP

Who in their right mind would want to leave a child in the care of a (virtually) complete stranger? (http://bigdaikon.org/board/viewtopic.php?f=1&t=94837&star t=0, accessed 22 August 2011) Anyone from Oregon going that doesn't mind having a complete (but harmless) stranger joining along? (http://oregonraqs.livejournal.com/, accessed 22 August 2011) b.

(25)

The syntactic tests proposed by De Smet and Verstraete (2006) thus rather unambiguously point to an analysis of degree modifiers, both of adjectives and nouns, as subjective on an ideational level only. Degree modifiers indicate the degree to which the qualities expressed by the descriptive modifiers and/or nouns they have scope over hold, thus adding descriptive information. At the same time, they express the personal opinion or evaluation of the speaker toward the entity designated by the NP as they place the qualities conveyed by the head and/or descriptive modifiers on a scale of subjectively assessed intensity (see Athanasiadou 2007: 557). Degree modifiers then pertain to the description of the NP referent while also reflecting speaker-relatedness, “in that the conceptualizers are very much involved in projecting their own perspective on an entity” (Athanasiadou 2007: 555). Although the syntactic tests clearly point out that both subjective descriptive modifiers (§1.3.1.1) and degree modifiers are ideational in nature, I would like to argue that the type of subjectivity displayed by degree modifiers is still different from that of subjective descriptive modifiers. This claim is supported by a number of observations. First, although nounintensifiers and descriptive modifiers both allow the syntactic tests proposed by De Smet and Verstraete (2006), noun-intensifiers do on a number of points display morpho-syntactic behaviour different from that of subjective descriptive modifiers. Unlike with subjective descriptive modifiers, additional degree modification is only marginally possible (e.g. a very/more complete player vs. *a very/more complete stranger) and noun-intensifiers typically resist predicative use (e.g. that player is complete vs. *the stranger is complete). Second, the semantics of degree modifiers also differs from that of subjective descriptive modifiers. Take, for example, the NPs a beautiful girl and a very beautiful boy. In both NPs, the subjective descriptive modifier beautiful attributes the property ‘beauty’ to the NP referent which is not straightforwardly objectively verifiable and is based on the subjective, aesthetic assessment by the speaker of the person referred to. Ascription of the property ‘beauty’ implicitly rests on comparison with other entities that are

Modification

43

or are not judged to be beautiful by the speaker (Hendrik De Smet, p.c.). As such, property attribution is inherently comparative and scalar: the speaker has to assess, on the basis of previous experience, whether or not the entity is such that it can be described as falling within the zone or category of items which are (prototypically) labelled beautiful, pretty, tall, etc. In the NP a very beautiful boy the scalar judgement made by the speaker of the NP referent is explicitly and semantically encoded in the adjectiveintensifier very. Similarly, use of an NP like an idiot implies a subjective evaluation by the speaker of the person referred as displaying to some degree the properties typically associated with ‘idiot’. In the NP a complete idiot the scalar evaluation is made explicit as (almost) reaching the top end of the scale. Whereas subjective descriptive modifiers only pragmatically imply scalar judgement, it is encoded in the semantics of degree modifiers. This is compatible with Paradis’s (2000) view on the development of adjectival noun-intensifiers as the foregrounding of the gradability mode of the lexical source meaning, i.e. how they are conceptualized in terms of scalarity (§3.4.2.1). Note that with adjective-intensifiers, the descriptive, ideational semantics are still more foregrounded in non- or less bleached ones than in bleached adjective-intensifiers.18 Third, support for the different or more subjective nature of degree modifiers comes from diachrony. In the Traugottian framework, subjectivity is essentially defined diachronically (§3.1). The degree of subjectivity of a linguistic item is typically established through comparison with earlier uses of the item. An item is more subjective as it becomes “increasingly based in the speaker’s subjective belief state/attitude toward the proposition” (Traugott 1989: 35). Degree modifiers are in the literature typically discussed as deriving from lexical, descriptively used adjectives. As mentioned above, the adjective-intensifier use of very, for instance, has been argued to have developed from the adjective’s original truth-related meaning, as in The written Law is but seeming justice; the Law of Nature very justice (OED s.v. very). Similarly, the noun-intensifier absolute, as in absolute nonsense, has been argued by Paradis (2000) to derive from the descriptive meaning of the adjective, as in Men set it vtterly fre and absolut from alle necessitie (Paradis 2000: 235; §3.4.2.1). In the shift to adjectiveintensifier or noun-intensifier the adjectives’ original descriptive, ideational

18. Clausner and Croft (1999: 18) even argue that a degree modifier like very “is profiled only in the domain of scale”, i.e. it is reduced solely to its scalar, degree modifying function and is devoid of any descriptive or quality attribution function.

44

A functional-cognitive model of the English NP

meaning is backgrounded and their scalar meaning is foregrounded, making explicit the speaker’s judgement or “subjective belief state/ attitude toward the proposition” (Traugott 1989: 35). As such, they become more subjective. Note that in Traugott’s (1982) analysis of language as consisting of three components, subjective descriptive modifiers would be part of the propositional component, “involv[ing] the resources of the language for making it possible to talk about something”, whereas degree modifiers would fit better into the expressive component, which “bears on the resources a language has for expressing personal attitudes to what is being talked about”. 1.4. Determination The determination zone accommodates all elements concerned with the identification and quantification of the NP referent. Its elements deictically and/or phorically anchor the instances of the type referred to by the NP in terms of such notions as givenness (the), relative quantity (most), etc. to the information the hearer has already built up from the previous discourse (Langacker 1991: 81–89; Davidse 2004). As represented in , structurally, determining elements occur either in the core or primary determiner position, or as secondary determiners in predeterminer (such a) or postdeterminer (the other) position. Table 1. The determination zone in the English NP determination zone secondary determiner most of

primary determiner

secondary determiner

head

those

problems

those

problems

the such

categorization zone

main

problem

similar

problem

a a

The determination zone as it is understood here differs from that developed in Bache (2000) in two significant respects. First, determination is understood here to encompass identification in the strict sense as well as

Determination

45

quantification. In Bache’s (2000: 162) model the quantification zone is envisaged as overarching the entire NP structure as it is “often expressed in connection with determination (These, A, Some), and categorization (singular meeting and plural books, girlfriends, students) but sometimes also in connection with modification (short, numerous, ten, little, in great numbers)” (see Figure 3; §1.1.2). However, modifiers such as short and little are in my opinion predominantly descriptive rather than quantifying in nature, albeit descriptive of size. Also, number as it is conveyed by the categorizing head of the NP is to be distinguished from quantification. Langacker (1991: 78–81) argues that number is part of the type description of the NP, involving conceptualization of one (singular nouns) or a replication (plural nouns) of type specifications. As such, number is concerned with the internal semantic structure or construal of the entity as a distinct entity or as a (non-)replicate mass and thus with type specification rather than determination. Every NP invokes “a single instance of some type, which is generally provided by the head noun together with its number specification”. In contrast, “[t]he role of a quantifier is to indicate the size of the profiled instance” [emphasis original] (Langacker 1991: 81). Rather than viewing quantification as encompassing both the determination, modification and categorization zones, I believe it is better to restrict quantification to the determination zone. As such, the determination zone is envisaged here to encompass not only identifiers in the strict sense (articles, demonstratives, and possessives) but also quantifiers. This is compatible with the fact that the core determiners are themselves inherently quantifying. As Davidse (2004: 507) states, “in all primary determiners basic conceptual mechanisms of quantification and identification are inextricably linked” (italics original). Demonstratives, for instance, pragmatically invoke notions akin to relative quantification. They can single out subsets of the instances of the relevant type in the context, e.g. These cats are bigger and broader than those cats, or, like the definite article, they can refer to all the instances of the type present in the context, e.g. What are these cats doing here? Likewise, the indefinite article is typically used with singular count nouns referring to one entity (e.g. a cat). Second, Bache (2000: 160) distinguishes the determination zone realized by identifiers in the strict sense from the specification zone realized by postdeterminers, which can be numeratives, adjectives, etc. Secondary determiners, both in pre- and postdeterminer position, have the function of providing identifying and quantifying information supplementary to that conveyed by the primary determiner to “help single out or quantify the referent of the construction in relation to some context” (Bache 2000: 235).

46

A functional-cognitive model of the English NP

As in Breban (2010a: 35), the determination and specification zones are hence better fused into one structural-functional zone comprising all the elements contributing to the identification and quantification of nominal referents (see Table 1). In the following sections, the general functions of the primary and secondary determining elements will be discussed. 1.4.1. Primary determination Primary determiners give either identifying or quantifying information in a broad sense (Bache 2000; Langacker 1991; Davidse 2004) and can be realized by four main types of determining elements: definite determiners, indefinite determiners, relative quantifiers and absolute quantifiers. Definite determiners, typically represented by the definite article, demonstratives and possessives, signal that the intended NP referent is identifiable or known to both speaker and hearer, e.g. (26). In contrast, the indefinite determiners a and the zero-determiner ø signal the nonidentifiability of the instance(s) referred to. They are not (yet) present in the discourse context as singled-out entities (Langacker 2002b: 33) and as such the hearer is required to mentally conjure up representative instances corresponding to the relevant type specifications (Langacker 1991: 103; Davidse 2004: 523). Therefore, the hearer has to be clear about the type specifications the instances have to correspond to. In this sense, indefinite reference presupposes type-identifiability (Gundel, Hedberg, and Zacharski 1993: 275). This is just as true of introducing indefinite NPs, which mention specific discourse referents for the first time (27), as of indefinite NPs with non-specific reference which designate an arbitrary instance (28) or singular indefinite generic NPs that evoke the class via a representative instance (29). In examples (27) to (29), the hearer has to know the concept ‘eligible bachelor’ and conceive of the corresponding type specifications for the cognitive operations required by these different types of indefinite reference to take place. (26)

Twenty single, attractive women are taken to a mysterious location in France. They are told they are going to be wooed by a dashing millionaire with marriage on his mind. In reality, the bachelor is actually a handsome but low-rent chap, named Evan, who works as a plumber's assistant. (WB cannews)

Determination

(27) (28)

(29)

47

A few weeks ago, I met an eligible bachelor that my parents gushed over. (http://quizilla.teennick.com/stories/17833845/letters-of-loveoneshot, accessed 29 July 2011) Fewer and fewer women are spending all their waking time looking for an “eligible bachelor.” (http://www.marginalrevolution.com/ marginalrevolution/2008/04/why-are-there-s.html, accessed 29 July 2011) An eligible bachelor is a bachelor considered to be a particularly desirable potential husband, usually due to wealth, or social status. (http://en.wikipedia.org/wiki/Eligible_bachelor, accessed 29 July 2011)

Whereas identification is traditionally discussed in terms of the definiteindefinite contrast, quantification will be defined here in terms of the absolute-relative contrast. This has been argued to be the most basic semantic contrast for quantification both within the formal semantic tradition (Milsark 1976, 1977) and the cognitive tradition (Langacker 1991). Absolute quantifiers, including cardinal numbers and more schematic quantifiers such as many, much, few and little, specify the cardinality or size of some instantiation of a type by offering “a direct description of its magnitude” (Langacker 1991: 82–83, 86). To this aim absolute quantifiers invoke an (open) scale of magnitude defined by precise and normative measure units in the case of cardinal numbers and by schematic, non-normative measure units with numeratives such as many and few (Davidse 2004: 509; §1.3.2.3). Relative quantifiers “make a quantitative assessment relative to a reference mass” (Langacker 1991: 108). In other words, they compare the mass or set actually designated by the NP to a reference mass/set containing all the contextually relevant instantiations of the available type specification. The actually predicated mass/set can be either a part or all of the reference mass/set, as in most parents and all parents respectively. Note that absolute and relative quantifiers can function either as primary or secondary determiners. 1.4.2. Secondary determination Although secondary determiners, especially in their realization as postdeterminers, have received some attention in most standard reference works on the English NP (e.g. Post-deictic in Halliday [1994: 183]; semideterminers in Biber et al. [1999: 280], specifying adjectives in Bache

48

A functional-cognitive model of the English NP

[2000: 235]), the theoretical development of the postdeterminer as a distinct function in the NP is greatly indebted to Breban and Davidse (2003), Breban (2002, 2006, 2010a) and, from a diachronic perspective, Davidse, Breban, and Van linden (2008). They have convincingly argued that the functional raison d’être of secondary determiners is to allow the speaker to provide the hearer with additional deictic and phoric instructions “that are more semantically diverse and more complex than the primary determiners they are bound by” (Davidse, Breban, and Van linden 2008: 500). Note that deixis is taken very broadly and includes the basic deictic systems of space, quantity, time and modality. On the point that secondary determiners are a sort of modifier of the determiner there is general agreement in the literature. Teyssier (1968: 227–228), for instance, characterizes the secondary determiner as an identifying ‘adjunct’. Bolinger (1967: 19) characterizes the postdeterminers in combinations such as the very, the same and the selfsame as ‘intensifiers’ of the determiner. Positionally, secondary determiners can either immediately precede (all, both, half, such) or follow (same, certain, former, last) the primary determiners with which they form a close-knit functional unit (see Figure 5). In terms of word class, secondary determiners are very diversified. They may be adjectives expressing deictic notions such as ‘known’ or ‘not known’ to the hearer, e.g. same, identical, equal, other, different, comparable, similar, related, additional, further (Breban 2002, 2006, 2010a; Breban and Davidse 2003) or ‘in a specific spatial/temporal relation to speech participants’, e.g. opposite, old (‘past’) (Davidse, Breban, and Van linden 2008). They can also be precise and fuzzy ordinal numbers such as first, second, fifth, next, final, e.g. the first day (Quirk et al. 1985: 261); precise and fuzzy cardinal numbers such as one, two, five, few, many, numerous, e.g. my three children (Quirk et al. 1985: 261), and relative quantifiers such as half, all, e.g. all the boys. Finally, secondary determiners may be adverbs such as then, sometime, once, e.g. the then practice (Brinton 2002: 67) and nominal constructions such as sort/type/kind + of, e.g. these sort of skills (Denison 2002: 3; De Smedt, Brems, and Davidse 2007). The above enumeration illustrates that although secondary determiners like primary determiners contribute to the determination of the NP referent they have typically retained more of their original lexical semantics (§3.4.1). Where do secondary determiners fit into the objective-subjectiveintersubjective continuum? With Breban (2010a), I hold that secondary determiners are typically textually intersubjective in that their specific function is to guide the hearer in his interpretational task and to help establish a joint focus of attention between speaker and hearer (§2.2.5). As noted by

Determination

49

Diessel (2006: 469), demonstratives are the linguistic elements most readily available to this function, but “there are many linguistic means that speakers can use to coordinate a joint attentional focus”. Secondary determiners are one of them. First, the joint attentional focus established by secondary determiners is typically realized discourse-internally, as with the discourse use of demonstratives. As such, they realize a type of “syntactic pointing”, “serv[ing] to direct the addressee’s attention backwards and forward along the speech stream, creating links between elements of the progressing discourse” (Diessel 2012: 45). More specifically, the discourse-organizational function of secondary determiners consists of providing the hearer with the necessary information to track and interconnect various discourse referents. As such, as argued by Breban (2010a: 115), they are “devices by which speakers take responsibility for success in communication and seek to meet the hearer’s attempts ‘to integrate new information with information that is already accessible’ (Blakemore 1990: 364)” (Traugott 1995: 45). In this sense, although of course speaker-related, secondary determiners are, in my view, primarily hearer-oriented and hence (textually) intersubjective in nature, providing instructions to the hearer to identify the NP referent thereby creating shared attentional focus.

Chapter 2 The English NP: Structure, order and the role of (inter)subjectivity The cognitive-functional constructional tradition has always insisted on the distinction between structure and order (see e.g. Langacker 1991: 157).19 In this tradition, structure is understood to refer to the syntagmatic relationships integrating component units into larger wholes. These syntagmatic relationships convey abstract grammatical meaning (McGregor 1997). For instance, if with Langacker (1987a: 309–310) one distinguishes two types of dependency structure, modification and complementation, one will also view these as differing with regard to such abstract features as which unit is the determinant of the semantic profile and what kind of relation obtains between head and dependent. In the case of complementation, a conceptually dependent head is elaborated, or ‘completed’ by its complement(s). In the case of modification, a conceptually dependent modifier is related to the nucleus formed by the head. The linear progression, or order, of elements is only the phonological or graphological manifestation of structure in spoken or written communication. As Halliday (2002a: 46) put it, “structure is the highest abstraction of patterns of syntagmatic relations”. By contrast, the whole dimension of progression in fact belongs to substance [either phonological or graphological substance, L.G.], and ... the stretches which carry grammatical patterns – or rather the members of that abstract category that we set up to account for these stretches – have to be ranged on a dimension of which linear progression is only a manifestation in substance: a dimension which we may call ‘order’ (Halliday 2002a: 43).20 19. With regard to the structure of an adjective modifying a noun, [[ADJ] [N]], Langacker (1991: 157) notes that its “temporal ordering is specified at the composite structure’s phonological pole”. 20. Halliday’s use of the term of ‘linear progression’ is reminiscent of Bolinger’s (1952: 1125) concept of ‘linear modification’, according to which “gradation of position creates gradation of meaning when there are no interfering factors”. With his notion of ‘linear modification’, Bolinger drew attention to the meaning-creating potential of order per se. Linear modification was referred to, amonst others, in the work of the Prague School on information structure. According to Firbas (1992),

52

Structure, order and (inter)subjectivity

Given that the main theoretical concern of this study is with subjectivity and the directionality of subjectification, the greater part of this chapter will be devoted to the order of elements in the NP in terms of objective– subjective meaning (§2.2). However, I will first discuss the main types of structure, or syntagmatic relations between units, in the NP recognized in the literature (§2.1). Structure constrains order in many ways, and statements about order without awareness of structure risk to remain superficial. 2.1. The English NP: A mere modifier-head structure? Traditional grammars, such as Quirk et al. (1972, 1985), Bache (2000) and Crystal (2003), have analyzed the noun phrase mainly as a modifier-head structure, with the noun as head surrounded by dependents which modify it (determiners, adjectives or adjective phrases, relative clauses). Langacker (1987a: 289, 309), similarly, describes the relation between head and modifiers in terms of the profile determinant, i.e. the part of the construction whose semantic profile the whole construction ‘inherits’. In canonical NPs, the head is the profile determinant and the modifiers are dependent predications that are not profile determining. “Red qualifies as a modifier in red cup, since the autonomous cup is the profile determinant (red cup as a whole designates a thing, not a relation)” (Langacker 2002a: 124). From a functional perspective, Halliday (1985, 1994) too views the head noun as the core of the NP from which all modification relations start. In his view, the logical relations holding between the component elements of NP structure can all be captured in terms of just one relationship: recursive modification (Halliday 1994: 191). The English NP is then defined in terms of a univariate structure, in which the head and modifiers are all related by “the recurrence of the same function”: “α is modified by β, which is modified by γ, which is ...” (Halliday 1994: 191, 193). The different elements of the NP constitute a dependency structure, with each added layer determining a further subset of the Thing. As Figure 5 shows, every added modifier describes a new subset of the general type denoted by the head noun, so that the NP is built up of progressive subsets. The modifiers do not all separately modify the head noun, but rather modify the complex consisting of the head noun and the modifiers to their right. linearity is one of the important principles operating in information structure. The unmarked ordering of theme and rheme complies with the principle of linear modification (Firbas 1992: 8).

A semantic, linear subjective-objective model? 53 [those [two Deictic Numerative

[splendid Epithet

[old [electric Epithet Classifier

trains]]]]] Thing

Figure 5. The recursive modification relations governing the NP elements

Halliday points out that, as illustrated in Figure 6, the dependency analysis of the NP as a structure showing regressive bracketing can be ‘disturbed’ through submodification, as in a rather more impressive figure where rather more functions as modifier of the ‘subhead’ impressive which is itself a modifier of the head figure (Halliday 1994: 192). a

rather

γ

more

impressive

figure

Modifier

Head

β Sub-Modifier

βγ

ββ

α SubHead

βα

Figure 6. Submodification (Halliday 1994: 192)

The modifier-head relations between the elements of the NP, which Halliday explicates in terms of recursive modification, tend to be assumed in many reference grammars of English as well. Yet, they have been subject to criticism in both older and more recent studies. One of the earliest nuances was proposed by Dixon (1982: 25). He posited that a distinction has to be made between recursive modification in the strict sense, as found in the two cleverest men, and independent modification of the head, as in a clever brave man. In his view, the difference in modification relation explains why the meaning of the two cleverest men is different from that of the cleverest two men, whereas a clever brave man is semantically identical to a brave clever man. The distinction between recursive modification and independent modification has more recently been subscribed to by Adamson (2000), Vandelanotte (2002) and Breban (2010a). Breban (2010a: 37–38) has argued that whereas classifying adjectives (§1.2.2) indeed enter into recursive modification relations with the head (e.g. [Japanese [financial markets]]), this is not necessarily the case with descriptive adjectives. She claims that descriptive adjectives – whether coordinated or non-coordinated – independently modify the NP referent. In

54

Structure, order and (inter)subjectivity

the NP a sticky red “jerk-style” sauce, for instance, both sticky and red independently modify “jerk-style” sauce. Another problem with the traditional dependency analysis pointed out by Tucker (1997) and Vandelanotte (2002) is that it does not take into account the distinction between coordinated adjectives, as in a rich, busy, jovial crowd (Vandelanotte 2002: 224), and non-coordinated adjectives, as in an ornamental eighteenth-century carved mahogany mantelpiece (cf. Halliday 1994: 192). The notion of progressively narrowing down the subsets of ‘mantelpiece’ referred to applies at least in an informal sense to the example with non-coordinated adjectives (Vandelanotte 2002: 224). However, there is no progressive narrowing of sets in NPs with coordinated adjectives, which all apply to the same subset, e.g. ‘crowd’ (Vandelanotte 2002: 222). A final problem with the traditional dependency analysis of the NP that has been remarked on in the literature pertains to syntagmatic relations between determiners and the rest of the NP. Breban (2010a: 38–39) has argued that determiners engage in a scoping relation (McGregor 1997: 210) with the NP referent, as they apply over the entire NP structure “leaving [their] mark” on the entire unit. For instance, [t]he NP this boy denotes a specific instance of the type ‘boy’. The determiner this profiles a schematic entity, the value of which can be paraphrased as ‘an entity that can be identified on the basis of a relation of proximity either in the speech situation or in the discourse’. In this way, this holds the specific entity denoted by the NP in its scope and indicates how it can be identified by the addressee (Breban 2010a: 38).

The scoping account of determiners is compatible with Langacker (1991: 146). He points out that determiners or ‘grounding predications’ have a different relation to the head noun than descriptive and classifying modifiers. The latter two are argued to add to the type specification of the ‘profile determinant’ or head noun. In a nominal such as black cats, black makes the higher-order type specification cats more specific. By contrast, grounding predications do not profile a relation to the designated entity, but profile the entity itself, linking it to the ground. In those three black cats, the demonstrative does not make the type specification more precise, but in itself designates the entity depicted by the nominal (see Brems 2007 for more detailed discussion). In this section, it has become clear that the English NP has a more complex syntagmatic structure than is traditionally assumed. Some important

A semantic, linear subjective-objective model? 55

problems with a simple modifier-head analysis have been indicated and it has become very clear that order in the English NP is structurally constrained. As such, the scene has been set for the main topic of this chapter, the ordering of the elements in the English NP in terms of objective versus subjective meaning. 2.2. A semantic, linear subjective-objective model? In the literature, the order of elements within the whole NP is traditionally believed to embody a semantic subjective-objective continuum. This has been argued for English (e.g. Adamson 2000; Bache 2000; Halliday 1985, 1994; Huddleston and Pullum 2002; Quirk et al. 1972, 1985) as well as cross-linguistically (e.g. Hetzron 1978; Seiler 1978; Dixon 1982). For English, Quirk et al. (1972: 924–926) have formulated it as follows: Modifiers relating to properties which are (relatively) inherent in the head of the noun phrase, visually observable, objectively recognizable or assessible, will tend to be placed nearer to the head and be preceded by modifiers concerned with what is relatively a matter of opinion, imposed on the head by the observer, not visually observed and only subjectively assessible.

Similarly, from a cross-linguistic perspective, Hetzron (1978: 178) noted that The major rule is to place the more objective and indisputable qualifications closer to the noun, and the more subjective, opinionlike ones farther away. This is not a discrete polar distinction but a gradual one, where ‘relatively more subjective’ is the relevant criterium [sic].

Although not explicitly stated in either of the above quotations, the subjective-objective continuum is generally considered to stretch from the far left end of the NP to the far right end, applying not only to the modifiers but also to the head of the NP. Quirk et al. (1972) and Hetzron (1978) do note that the opposition between subjective and objective uses is a relative or gradient one. There is no clear-cut division between subjective and objective uses. Rather some uses are more subjective or more objective than others. Importantly, a semantic model of the English NP built up around a distinction between subjective and objective meanings crucially hinges on well-defined and generally accepted notions of what subjectivity and objec-

56

Structure, order and (inter)subjectivity

tivity ultimately entail. Also, in the light of recent theorizing, the question arises to what extent the increasingly popular concept of intersubjectivity is relevant to the description of the English NP. In the recent cognitivefunctional literature, subjectivity and intersubjectivity are key concepts – descriptive as well as explanatory. However, despite their current ubiquity, “both concepts remain surprisingly ill-defined” (De Smet and Verstraete 2006). This is in part due to the wide range of phenomena which have been captured under the heading of subjective or intersubjective meaning, ranging from modals to discourse markers, cohesive devices and elements of the NP. The terminological picture is further blurred due to the presence of different views on (inter)subjectivity, whose interrelation is not always clear. Sections 2.2.1 to 2.2.4 will be devoted to introducing the two main viewpoints on (inter)subjectivity available in the literature, i.e. that of Langacker and that of Traugott.21 Both authors define subjectivity with reference to the centrality of the speaker in language but do so in different, perhaps even orthogonal ways. In the remainder of this study a largely Traugottian view on subjectivity and intersubjectivity (and their diachronic counterparts [§3.1]) in the NP will be developed, albeit with some reservations and alterations (§2.2.3 to 2.2.6). 2.2.1. Langackerian subjectivity A notion of subjectivity often referred to in the literature is that of Langacker (1985, 1987a, 1990, 2003, 2006). In his view, subjectivity is ultimately concerned not with the inherent meaning of a linguistic element but with the way a certain conceptual content is construed by the speaker. As such, the subjective or objective nature of a particular content is determined “relative to the asymmetry between the observer in a perceptual situation and the entity that is observed” (Langacker 1985: 120). On the one hand, an entity is construed with maximal objectivity when the speaker or conceptualizer remains implicit and interpretation does not require reference to the 21. On the basis of his research in the field of modality, Nuyts (2001a, 2001b, 2012) has developed yet another view on (inter)subjectivity. “A modal evaluation is ‘subjectiveJN’ if it is presented as being strictly the assessor’s sole responsibility. A modal evaluation is ‘intersubjectiveJN’ if it is presented as being shared between the assessor and a wider group of people, possibly (but not necessarily) including the hearer” (Nuyts 2012: 58). This more pragmatic understanding of (inter)subjectivity will not be used in this study.

A semantic, linear subjective-objective model? 57

ground, i.e. the speech event and its participants. The entity itself is the focus of attention and is placed ‘on-stage’. Figure 7 adapted from Langacker (1985), visualizes such objective construal with maximal asymmetry between the Subject S, i.e. the observer or speaker, and the Object O.

S

O

Figure 7. Objective construal

S

O

Figure 8. Subjective construal

On the other hand, subjective construal of an entity necessarily involves reference to an element of the ground for its interpretation (Figure 8). Importantly, subjectivity is seen as a gradient phenomenon, with many intermediary stages between maximally objective and maximally subjective construal, as illustrated in (30). (30)

a. Vanessa jumped across the table. b. Vanessa is sitting across the table from Veronica. c. Vanessa is sitting across the table from me. d. Vanessa is sitting across the table.

Across in (30a), for instance, is maximally objective as it profiles movement without reference to speaker-hearer position. (30b) and (30c) are more subjective in that across no longer describes objective movement of a ‘trajector’ vis-à-vis a ‘landmark’ but only the speaker’s mental scanning operation to locate the trajector with respect to a landmark or reference point related to the speech event, i.e. a participant in the speech event other than the speaker or hearer in (30b), and the first person pronoun referring to the speaker in (30c). (30d) is most subjective as the reference point is now an element of the ground – the speaker –, i.e. an off-stage element, as illustrated in Figure 8. 2.2.2. Intersubjectivity in Cognitive Grammar Although Langacker (2006) touches upon the concept of intersubjectivity, it is not developed in any detail. However, from an argumentation theory perspective but building on the Cognitive Grammar framework developed

58

Structure, order and (inter)subjectivity

by Langacker, Verhagen (2005, 2007) has proposed an account of intersubjectivity in terms of “humans’ ability to engage in deep cognitive coordination with others” (Verhagen 2005: 4). As such, the hearer is more explicitly drawn to the foreground as he is said to engage in a mutual cognitive coordination task with the speaker.22 The speaker of course is responsible for a certain utterance, but it is the hearer’s task to interpret it, taking into account the speaker’s perspective and (re)construing the object of conceptualization. In this view, the hearer is thus not only a passive addressee guided by the speaker to focus on a given conceptual content but he is crucially an active conceptualizer.23 Intersubjectivity understood as cognitive coordination of speaker and hearer is claimed to be encoded in amongst other things negation patterns with not, and causal, concessive, and contrastive connectives like but, as they all contribute to construal management and aid to ensure that the hearer makes correct inferences. In Verhagen’s (2005: 215) view, “the grammatical system of negation is used to coordinate perspectives ... not so much with respect to the descriptive contents of the utterances involved but with respect to their argumentative, rhetorical orientation”. Consider the following example: (31)

Mensen die verandering willen op één noemer brengen is altijd lastig omdat zij allicht verschillende soorten verandering willen. Maar het vinden van die noemer is niet onmogelijk. ‘It is always hard to find a common denominator for uniting people who are in favor of change, because they are likely to wish for different kinds of change. But finding such a common denominator is not impossible.’ (Verhagen 2005: 33)

The negated form niet onmogelijk ‘not impossible’ in (31) is said to explicitly invite ‘intersubjective coordination’ (Verhagen 2005: 76). “The text is understood as presenting two different mental spaces, the first with a negative attitude towards trying to unite people in favor of change, the second with a positive one. The opposition between these attitudes towards the 22. Although the wording of Verhagen’s view on intersubjectvity is reminiscent of Diessel’s idea of the creation of a joint focus attention between speaker and hearer (see §2.2.5), it captures a much wider range of phenomena and is’ in fact, understood to pervade all language use. In Verhagen’s view the cognitive coordination task of speaker and hearer is omnipresent in communication, although coded more explicitly in some constructions than others. 23. Thanks go to Elisabeth Zima for sharing her insights on Verhagen’s notion of intersubjectivity.

A semantic, linear subjective-objective model? 59

same issue is what makes the fragment coherent, and this conceptual configuration is explicitly activated by the use of sentential negation” (Verhagen 2005: 34). The Cognitive Grammar notions of subjectivity and intersubjectivity as discussed above will not figure prominently in this study. First, they are not easily applicable to (all of the component elements of) the NP (§3.1), and, second, they have not been well-developed diachronically. In contrast, the Traugottian understanding of (inter)subjectivity (§2.2.3 and 2.2.4) and (inter)subjectification (§3.1) lends itself better to account for the synchronic and diachronic description of the NP. 2.2.3. Traugottian subjectivity and beyond The Traugottian understanding of subjectivity is greatly indebted to Lyons’s (1982: 102) original characterization of subjectivity as referring “to the way in which natural languages, in their structure and their normal manner of operation, provide for the locutionary agent’s expression of himself and his own attitudes and beliefs”. Subjective meanings then “explicitly encode SP/W's [speaker/writer’s] point of view” (Traugott and Dasher 2002: 21–22). In other words, they always involve “a relationship to the speaker and the speaker’s beliefs and attitudes” (Traugott 2010a: 30). Importantly, subjectivity refers to (en)coded or semanticized meanings, rather than merely pragmatic meaning (Traugott 2010a: 35). In a general sense, all language use is inherently subjective in that it “passes through a speaker and is, as such, speaker-related” (De Smet and Verstraete 2006: 370). Such “pragmatic subjectivity” is then “independent of the semantics of a particular expression” (De Smet and Verstraete 2006: 384). In the Traugottian framework adopted in this paper, subjectivity is taken to refer to semantic or semanticized meanings, coded in conventionalized formmeaning pairings. Like Langackerian subjectivity, Traugottian subjectivity is essentially gradient. Subjectivity is very broadly defined in terms of speakerinvolvement, but of course “the speaker can be involved in an utterance in many different ways” (De Smet and Verstraete 2006: 366). As a consequence, the notion of subjectivity has been used in relation to a wide range of phenomena of grammatical, lexical or deictic nature, conveying widely divergent meanings (cf. Cuyckens, Davidse, and Vandelanotte 2010: 11– 12). The gradient nature of subjectivity is also apparent from Traugott’s (1989: 35) definition of subjectification, according to which meanings “be-

60

Structure, order and (inter)subjectivity

come increasingly based in the speaker’s subjective belief state/attitude toward the proposition” [emphasis added]. De Smet and Verstraete (2006) have attempted to typologize subjectivity on the basis of semantic and syntactico-structural criteria.24 They argue that at least two types of semantic(ized) subjectivity can be distinguished – ideational and interpersonal – each associated with its own formal reflexes. Their terminology refers to Halliday and Hasan (1976) and Halliday (1994), who refer to all representational semantics with the term ‘ideational’, while all speaker/hearer related meanings are labelled ‘interpersonal’. Ideational subjectivity “involves the description of a content ‘situated in the speaker’s subjective belief-state/attitude toward the situation’, as Traugott and Dasher (2002) put it in their definition of subjectification” (De Smet and Verstraete 2006: 385). Interpersonal subjectivity, in contrast, does not involve the “description of speaker-internal content” but rather “the enactment of speaker position [with respect to that] content” (De Smet and Verstraete 2006: 386–387). The semantic distinction between ideational and interpersonal subjectivity is argued to be reflected syntactically in distinct behaviour with respect to focusability, wh-interrogation and negative scope. “Focus, interrogation and negation all serve to pick out some aspect of content, either by highlighting it or denying it” (De Smet and Verstraete 2006: 386) and therefore only apply to ideationally subjective elements, not to interpersonally subjective ones. In Section 1.3.1.1, the syntactic criteria proposed by De Smet and Verstraete (2006) have been used to determine the type of subjectivity conveyed by the different types of modifiers in the prenominal string of the NP. An example discussed in De Smet and Verstraete (2006: 374) is the contrast between the subjective uses of the Dutch adjectives dom ‘bloody, cursed’ and leuk ‘pleasant’, as in (32) and (33). (32) Dát was een leuke verrassing! [robbers Gel. Fam.] 97. (WNT s.v. leuk) ‘That was a pleasant surprise!’ (33) Wat was er dan eigenlijk te vreezen of te duchten? t was maar dat domme geld dat er altijd te kort was, st. streuvels, Minneh. 1, 87. (WNT s.v. dom) ‘What was there to fear or be anxious about? It was always just that cursed money that ran short.’ 24. For alternative typologies of subjectivity see amongst others Smirnova (2012) and Visconti (2013).

A semantic, linear subjective-objective model? 61

De Smet and Verstraete (2006: 376–377) argue that “dom, unlike leuk, can no longer function as the focus of a wh-question or as the focus of negation”, as illustrated in (34) and (35). A: Wat voor picknickmand was de bergbeklimmer vergeten? ‘What kind of picnic basket had the mountaineer forgotten?’ b. B: De bergbeklimmer was een leuke picknickmand vergeten? ‘The mountaineer had forgotten a nice picnic basket.’ b’. B: *De bergbeklimmer was een domme picknickmand vergeten. *‘The mountaineer had forgotten a cursed picnic basket.’ (35) a. Wees gerust, mijn zus is geen domme advocaat. Ze is een beeldhouwster. ‘Don’t worry, my sister is not a cursed lawyer. She is a sculptress.’ b. Helaas, mijn zus is geen leuke advocaat. De getuigen zijn altijd als de dood voor haar. ‘Unfortunately, my sister is not a pleasant lawyer. The witnesses are always terrified of her.’

(34) a.

The syntactically distinct behaviour of the subjective uses of leuk and dom is argued to “have more general relevance as reflexes of finer semantic distinctions within the domain of subjectivity” (De Smet and Verstraete 2006: 388). Whereas leuk is subjective on an ideational level only, “describ[ing] attitudes internal to the speaker”, dom is also interpersonally subjective. “Although it also denotes an attitude, this attitude does not serve as a basis for subcategorizing the general type [denoted by the head noun]: instead, dom enacts the speaker’s negative stance towards an instance of this category” (De Smet and Verstraete 2006: 386). The distinction between ideational and interpersonal subjectivity applies both to attitudinally subjective elements, predominantly reflecting the speaker’s attitude, like the adjectives leuk and dom, as well as to linguistic elements of a more textual, discourse-organisational nature (see also §2.2.5). The latter are exemplified in De Smet and Verstraete (2006) by means of causal conjunctions such as because and since. Causal conjunctions reflect the speaker’s point of view on the causality between two events or situations and are as such subjective. A distinction can be made, however, between ideationally subjective causals like because, which merely describe a causal relation, as in (36), and interpersonally subjective causals like since in (37), which explicitly links the causal relation to the speaker.

Structure, order and (inter)subjectivity

62

(36) (37)

It is because they work too hard that many Japanese become ill. (De Smet and Verstraete 2006: 383) When my children ask, I have to look up the old formulas, since not in 25 years have I applied them in real life. (De Smet and Verstraete 2006: 382)

2.2.4. Traugottian intersubjectivity and beyond As with subjectivity (§2.2.3), a pragmatic and a semantic, coded type of intersubjectivity have to be distinguished: in a general sense the very fact of communicating with another person entails general intersubjectivity. The ‘I’ is constituted in part by conceptualizing the other member of the communicative dyad ‘you’ (Benveniste 1971; Lyons 1994) and discourse is communicatively successful only if speakers pay attention to audience needs, and if ‘mutual manifestness’ or ‘mutual management’ is worked on (Schiffrin 1990; Nuyts 2001; Verhagen 2005). (Traugott 2010a: 32)

This type of pragmatic intersubjectivity is not of great concern to the Traugottian framework. In fact, the semantic rather than pragmatic nature is, among other things, what distinguishes Traugottian intersubjectivity from other interpretations such as that of Nuyts (2001a, 2001b, 2012) and Verhagen (2005, 2007), which “assume audience interpretation and understanding” (Traugott and Dasher 2002: 23). For Traugott and Dasher (2002: 22) and Traugott (2003a, 2010a) intersubjectivity is closely connected to and cannot be defined without reference to subjectivity. Both synchronically and diachronically, subjectivity is considered a prerequisite for intersubjectivity. While expressions of subjectivity index speaker attitude or viewpoint, markers of intersubjectivity also index the speaker’s “attention to AD[dressee]/R[eader] as a participant in the speech event, not in the world talked about” (Traugott and Dasher 2002: 22). More specifically, intersubjectivity involves speaker attention to the social self of the hearer: intersubjective meanings crucially involve social deixis (attitude toward status that speakers impose on first person – second person deixis). They impact directly on the self-image or ‘face’ needs of SP/W or AD/R. (Traugott and Dasher 2002: 23)

A semantic, linear subjective-objective model? 63

This notion of intersubjectivity is thus rather restricted, limited to meanings relating “to the addressee and addressee’s face” (see recent studies of intersubjectification such as Traugott 2007a, 2007b, 2010a). According to Traugott (2012: 10) the two types of intersubjective language functions most likely to arise are related to “politeness (encoding of the Speaker’s appreciation and recognition of the Addressee’s social status)” and to particular “metadiscursive functions such as turn-giving or elicitation of response. As also argued in Ghesquière, Brems and Van de Velde (2012), these two types can be referred to as attitudinal and metadiscursive or responsive intersubjectivity respectively. First, the textbook example of attitudinally intersubjective elements are the Japanese addressee honorifics, such as Modern Japanese -mas-u-, which “started out as a morphologically complex main verb *mawi-ir-as-uru ‘let come (humilitive)’ in Old Japanese, came to be used as a [subjective] suffix verb indicating humility without lexical content, and finally became the modern [intersubjective] honorific addressee marker” (Narrog 2010: 387). Other prime examples are T/V pronouns (tu/vous), which code politeness or social distance, and hedging uses of discourse markers such as sort of or well in (38). What all of these elements have in common is that they encode the speaker’s image of (his/her relation to) the hearer, and attention to the face needs and social self of the hearer. (38)

‘And you didn’t feel guilty about it afterwards?’ ‘Well, he might have lost his girlfriend, but at least I gave him a good time!’ (WB brbooks)

Second, responsive intersubjectivity is semanticized in certain turn-taking devices and question tags soliciting a response by the hearer such as clause final right?, as in (39), and is it not? or isn’t it?, as in (40). (39) (40)

‘I’ve seen your name in the paper … that diplomat who’s been sentenced to death, right?’ ‘Yeah.’ (WB brbooks) Wei strokes his chin. `This is an imaginary cat, is it not?’ Julie nods. (WB brbooks)

The above constructions, unlike the attitudinally intersubjective ones, do not primarily code attention to the social self of the hearer or serve as facesaving strategies. Rather, they convey a type of metadiscursive or respon-

64

Structure, order and (inter)subjectivity

sive intersubjectivity, eliciting a certain action or behaviour on the part of the hearer thereby aiding discourse continuity. 2.2.5. Textual meanings and (inter)subjectivity In this section, I want to explore the place textual meanings take up in the objective-subjective-intersubjective continuum, while at the same time fleshing out the notion of intersubjectivity. In Traugott’s (1982) cline of semantic change the textual component of language occupied a prominent and independent position, as meanings were hypothesized to shift from the propositional through the textual to the expressive domain. The textual meaning component has, however, been moved somewhat to the background when Traugott and Dasher (2002) and Traugott (2003a) came to distinguish subjective and intersubjective meaning. Recently, Traugott (2010a: 31) seems to argue that the linguistic elements which used to belong to the textual domain are best classified as expressing either objective or subjective meanings: ‘Textual’, as understood [in Traugott (1982)], included various connectives such as and and therefore, as well as anaphoric and cataphoric pronouns, topicalizers, relativizers, complementizers, etc. In the nineteen-eighties it became clear that while all are essential ingredients of grammar, some of these serve more contentful (and sometimes truth-conditional) purposes of local connectivity (e.g. relativizers, complementizers), whereas others serve the procedural purposes of expressing speaker’s attitude to the text under production (topicalizers, discourse markers). Indeed, many connectives have dual functions, e.g. and, then, in fact. [emphasis added]

The potential subjectivity of textual meanings is also recognized by De Smet and Verstraete (2006), who distinguish both ideationally and interpersonally subjective discourse-organisational elements (see §2.2.3). In contrast, Narrog (2010, 2012) argues that textual meanings cannot be fully subsumed under the labels of either objectivity or (inter)subjectivity. Textual meanings are in his view best understood as occupying a distinct position in the language system, on a par with subjective and intersubjective meaning. Whereas subjective meaning then roughly corresponds with speaker-orientation and intersubjectivity with hearer-orientation, textual meanings entail “orientation towards speech, i.e. discourse or text, itself” (Narrog 2012: 30). The tripartite distinction between subjective, intersubjective and textual meanings is reminiscent of the three functions of evalua-

A semantic, linear subjective-objective model? 65

tion distinguished in Thompson and Hunston (2000: 6). They argue that evaluation can be used (i) “to express the speaker’s or writer’s opinion”, i.e. convey subjective meaning, (ii) “to construct and maintain relations between the speaker or writer and hearer or reader”, i.e. convey intersubjective meaning, or (iii) “to organize the discourse”, i.e. to express textual meaning. Importantly, in their view, all three functions are seen as subtypes of evaluation and thus essentially subjective. Yet another view on textual meanings is put forward by Breban (2010a: 115). She argues that the role of the speaker in elements with textual meaning is not simply to observe and to give an accurate representation of his observations in the form of text, as is the case with objective meaning, but he/she takes up the additional function of organizing this representation. ... In this way, the speaker ‘steps into’ the text and leaves his mark on the text. In my opinion, this is exactly what subjectivity involves: text showing the signs of the presence of the speaker.

The above quotation shows that Breban (2010a) views all textual meanings as subjective, as reflecting speaker-involvement. However, she goes yet one step further and proposes that textual meanings are not only subjective but also intersubjective. Support for this claim is found in the observation that textual meanings “also contain a prominent hearer-oriented aspect, that is, they invoke speaker-hearer negotiations and in the end all aim at facilitating the interpretation by the hearer” (Breban 2010a: 115). In this study, textual meanings are understood as meanings whose interpretation requires reference to some stretch of discourse. This broad definition captures a wide range of meanings, which cannot be put on par with subjective and intersubjective meanings nor can they be fully subsumed under either heading. Instead, I will argue that the textual component of language intersects with the entire objective-subjective-intersubjective continuum. I agree with Traugott (2010a: 31) that textual elements can convey both objective and subjective meanings. One example is the conjunction while, which has developed from the Old English phrase þa hwile þe ‘at the time that’, as in (41). In Present Day English, while can function either as a temporal connective, as in (42), or as a concessive, as in (43).

66

(41)

(42) (43)

Structure, order and (inter)subjectivity

þæs mannes sawul is belocen on his lichaman þa hwile þe he lybbende biþ. ‘Man’s soul is locked in his body while/so long as he is alive.’ (Traugott 1995: 40) While he was laughing the door opened. (OED Besant, 1882, All Sorts of Men II. xv. 15) While I quite like that kind of tiling, I don’t care enough for it to buy any. (Traugott 1982: 254)

þa hwile þe ‘at the time that’, as in (41), “refers to a temporal situation viewed as part of a verifiable state-description and in this sense is propositional” (Traugott 1989: 31). The temporal phrase has ideational, propositional meaning. As a temporal connective, as in (42), “while in the sense of ‘during’ signals a cohesive time relation not only between two events but also between two clauses, and therefore has a textual as well as a temporal function” (Traugott 1989: 31). In other words, temporal connective while serves a linking function whereby the established cohesive link reflects a link at the propositional level and, as such, conveys objective textual meaning. In contrast, concessive while, as in (43) “combines cohesion [i.e. textual meaning] with the speaker’s attitude as to the nature of the relation between the two facts expressed” (Traugott 1982: 254). As a concessive, while thus conveys subjective textual meaning, involving the speaker taking a stance with respect to a stretch of discourse. Similarly, Traugott and König (1991) and Traugott (1995) have argued that the causal but not the temporal uses of conjunctions such as since and because are rhetorical and thus a matter of the speaker’s perspective. As aptly summarized in De Smet and Verstraete (2006: 379–380), “temporal succession is an aspect of the world described in the construction, which can be assessed independently of the speaker’s perspective [i.e. objective], while causality is ... a matter of the speaker’s point of view whether two events or situations are seen as causing each other”, i.e. subjective. Likewise, the concessive relation between two propositions is determined by the speaker’s perspective on the situation. Objective textual meanings are, for instance, also conveyed by relativizers and complementizers and through subject-verb agreement. All of these phenomena rely on a textual linking relation or cross-reference between different parts of a sentence but do not require stance-taking on behalf of the speaker. On the other hand, topicalizers like as far as and so convey subjective textual meaning as they “serve to express the speaker’s

A semantic, linear subjective-objective model? 67

attitude to the textual strategy being adopted”, as in So, our speaker tonight is Bella Johnson (Traugott and Dasher 2002: 96). Whereas the textual meanings discussed above can be labelled as either objective or subjective, it is not immediately clear where the textual meanings conveyed by determining elements in the NP fit into the objectivesubjective continuum. From a cognitively-oriented perspective, Diessel (1999, 2006) has argued that demonstratives serve to coordinate the speaker and hearer's joint attention on a particular reference object. Joint attention is a complex phenomenon that involves three basic components: the actor [speaker], the addressee, and an object of reference. In order to communicate, actor and addressee must jointly focus their attention on the same entity or situation. To this end, the actor directs the addressee’s attention to a particular reference object in the surrounding situation [e.g. by means of demonstratives]. If the communicative act is successful, the communicative partners focus their attention on the same referent. (Diessel 2006: 465)

I would like to highlight two important aspects concerning the notion of joint attention. First, joint attention crucially involves attention to the hearer as “it presupposes that the speech participants recognize each other as mental agents” (Diessel 2011: 8). Second, refining the definition above, the focus of attention can be situated not only in the surrounding situation, but also in the surrounding context. “Joint attention is not only important to coordinate the interlocutors’ attentional focus in the speech situation, it also plays an important role in the organization of discourse” (Diessel 2006: 476). The distinction is evidenced in the exophoric, deictic pointing use of demonstratives versus their discourse use in which they have “text-internal reference”. Whereas exophorically used demonstratives “focus the hearer’s attention on entities in the situation surrounding the interlocutors”, i.e. the extralinguistic or extratextual world, anaphorically and discoursedeictically used demonstratives are endophoric in that they refer to elements in the ongoing discourse, namely noun phrases and propositions respectively, i.e. they have a textual function (Diessel 1999: 93–94). How can the notion of joint attention as it is associated with determining elements be fitted into the objective-subjective-intersubjective domain? Building on Diessel’s observations, I would like to propose that textual elements by which the speaker negotiates discourse referent tracking for the hearer, e.g. (complexes of) primary and secondary determiners, ensure joint attentional focus of speaker and hearer and can hence also be argued to convey intersubjective meaning (cf. Breban 2006, 2010a; Ghesquière 2009,

Structure, order and (inter)subjectivity

68

2010; Carlier and De Mulder 2010). This does, however, require a more liberal interpretation of intersubjectivity than is generally assumed. As Carlier and De Mulder (2010: 269) put it, intersubjectivity then “is not limited to the encoding of attitudinal aspects; it also concerns more globally items that materialize the strategic interaction between speaker and hearer and reflect the active role of the speaker to orient and to guide the hearer in his interpretational tasks”. Carlier and De Mulder (2010) illustrate the intersubjective textual use of determiners by means of the Late Latin demonstrative ille, as in (44). (44)

... homines illos quos Waiofarius ad defendendam ipsam ciuitatem dimiserat clementiam sue pietatis absoluit dimissisque reuersi sunt ad propria (Continuationes, § 43) ‘Of his goodness he showed mercy to the [lit. those] men that Waiofar left there to defend the city, and dismissed them to go off home.’ (Carlier and De Mulder 2010: 259)

In examples such as (12), ille has a definite first mention use, which deictically evokes ‘new’ or ‘renewed’ identification by the hearer. Rather than referring to a referent previously mentioned in the text, ille realizes first mention of referents whose identity is established by a following relative clause. In other words, by using ille the writer aims to offer to the reader the clues needed for a clear and coherent interpretation of the text and for the identification of the referents participating in the narrative. … [B]y means of the demonstrative ille, he asks the reader to make a new identification of the referent [which] requires the hearer to take into account the specific occurrence of the demonstrative and its context (Carlier and De Mulder 2010: 327).

The definite first mention use of ille can thus be viewed as conveying intersubjective textual meaning, marking a shift to the perspective of the hearer to assure joint attention. Carlier and De Mulder (2010: 266–267) suggest that the later development from demonstrative to definite article “can be described in terms of a strengthening of this intersubjective dimension”. I do want to emphasize that not all textual meanings expressed by determiners are intersubjective. Again, this is illustrated nicely in Carlier and De Mulder (2010). In (45), the Latin demonstrative ipse is used subjectively as a marker of discourse prominence, thus primarily encoding speaker attitude.

A semantic, linear subjective-objective model? 69

(45)

Eo anno uxor Anaulfi imperatores Persarum nomen Caesara uirum relinquens, cum quattuor pueris totidem puellis ad beatum Iohannem episcopum Constantinopule ueniens, se unam esse de populo dixit et baptismi gratiam ad antedictum beatum Johannem expetit. Cumque ab ipso pontifice fuisset baptizatus, agusta Maurici imperatores ea de sancto suscepit lauacro (Fredegarius, §9) ‘In this year Caesarea, wife of Arnauld, the Persian emperor, left her husband and came with four male and four female servants to the blessed John, bishop of Constantinople. She said that she was a private person and besought the blessed John to baptize her. She was baptized by the bishop in person, and the illustrious wife of the Emperor Maurice stood godmother to her.’ (Carlier and De Mulder 2010: 254)

In sum, with Breban (2006, 2010a) and Carlier and De Mulder (2010: 330), I argue for a wider view of intersubjectivity as encompassing not only an attitudinal and responsive component but also a textual component. The notion of intersubjectivity is then extended to include meanings relating not only to the social self and face of the hearer (Traugott and Dasher 2002), but also textual meanings negotiating discourse referent tracking by the speaker to the hearer (see also Brems 2007). Such an extended notion of intersubjectivity is in accordance with Traugott’s (1995: 47) broad definition of subjectification as “the tendency to recruit lexical (propositional) material for purposes of creating text and indicating attitudes in discourse situations”. As I understand it, Traugott and Dasher’s (2002: 22) original, broader definition of intersubjectivity as “crucially involv[ing] SP/W’s [speaker/writer’s] attention to AD/R [addressee/reader] as a participant in the speech event” also does not preclude a textual understanding of intersubjectivity. 2.2.6. Subjectivity and intersubjectivity: Disentangling the web As noted in the introduction to Section 2.2, a widespread view on the semantic organization of the (English) NP is that it incorporates a continuum from subjective to objective meanings. If one wants to maintain such a view, it is crucial to have a well-developed notion of subjectivity, and by extension intersubjectivity, a concept that has become and is becoming increasingly popular in cognitive-functional linguistics (e.g. contributions in Brems, Ghesquière and Van de Velde 2012). To that aim, the two main

Language takes place in a communicative dyad within which the speaker addresses the hearer  intersubjective context in which communication takes place

Figure 9. Types of subjectivity and intersubjectivity

Intersubjectivity hearer-orientation

Subjectivity speakerorientation

Pragmatic Language is produced by a speaker and is as such inherently speaker-related  subjective use of an expression

Responsive Elicit a certain action or behaviour of the hearer e.g. turn-taking devices, question tags soliciting a response by the hearer

Attitudinal Code the speaker’s image of (his/ her relation to) the hearer e.g. addressee honorifics

e.g. determiners establishing joint attention

Textual Steer interpretation by the hearer

Interpersonal Enactment of speaker position w.r.t. content Attitudinal Textual e.g. causal e.g. those stupid keys connectives involving link to the speaker Encoded relationship to the hearer as a participant in the speech event  intersubjective meaning of an expression

Ideational Description of speaker-internal content Attitudinal Textual e.g. descriptive e.g. a stupid decision causal connectives

Semantic Encoded relationship to the speaker and his/her attitudes and beliefs  subjective meaning of an expression

70 Structure, order and (inter)subjectivity

A semantic, linear subjective-objective model? 71

viewpoints on subjectivity and intersubjectivity available in the literature have been introduced in the preceding sections. Whereas Langackerian subjectivity is basically a matter of construal, Traugott’s notions of subjectivity (§2.2.3) and intersubjectivity (§2.2.4) are inherently semantic. As this study’s main concern lies with semantic change in the NP, it is this notion which interests us most. Figure 9 provides an overview of the different types of Traugottian subjectivity distinguished in this study. I do not claim this overview to be in any way complete or definitive. For instance, one might wonder to what extent the distinction between ideational and interpersonal subjectivity might be relevant to intersubjectivity and, conversely, if and how this distinction relates to the attitudinal, responsive and textual types of intersubjectivity distinguished. This certainly merits further investigation, but for this study the present typology will suffice. 2.2.7. Challenging the linear subjective-objective model In the light of the preceding discussion, an account of the ordering in the NP in terms of subjectivity versus objectivity cannot but be a simplification. Indeed, it is not hard to find examples which systematically contradict such a rigid ordering of elements. Amongst other things, idiomatic collocations, like the big bad wolf, and euphonic principles can disrupt the syntagmatic subjective-objective cline. The most well-known euphonic tendency is the ‘law of increasing members’, which points to the preference for “a short-long order of words when either order would be a priori possible” (Hetzron 1978: 176), e.g. a pure, pristine, platonic form of hospital administration and a calmer, smoother, stronger, more comfortable feeling (Vandelanotte 2002: 224–225). Another phenomenon thwarting the hypothesized correlation between subjectivity and leftness in the NP is formed by ‘degree nouns’ (Bolinger 1972) or ‘gradable nouns’ (Paradis 2000: 243) which “come with an inherent evaluation of either a positive [e.g. bliss, pleasure] or a negative character” (e.g. mess, bastard). These nouns are clearly subjective yet associated with rightmost position in the NP (not taking into account postnominal modification). Similarly, subjective compounds such as little bleeder, old chum and half victory also convey subjective meaning while typically taking up the rightmost slot in the NP. The elements immediately preceding the noun clearly convey subjective meaning and can be preceded by more objective elements, as in upstart West Coast bespectacled little twerp (WB

72

Structure, order and (inter)subjectivity

usbooks). These subjective compounds will be returned to in more detail in Chapter 8. Also challenging the subjective-objective continuum are degree modifiers which follow the elements they modify. In her study of the adjectives little and old and subjectivity in the NP González-Díaz (2009: 387) observes that old can form embedded clusters with other adjectives in the prenominal string such as good old, grand old, high old. Such clusters have been described before and analyses have been put forth, viewing them either as paratactic, balanced structures (Matthews p.c., quoted in GonzálezDíaz 2009) or as hypotactic structures, with old functioning as a ‘reinforcing’ submodifier (OED s.v. old a.III.15). The degree modifying effect of old on adjectives preceding it is particularly clear when it ‘interrupts’ a sequence in which modifier and noun collocate with each other, e.g. not a bad stick (i.e. quite a good stick) in (46). (46)

`Ah," I said, baffled. Maslow's not a bad old stick for a troop leader (WB brbooks)

Degree modifiers, like old in the above example, are generally considered to derive from (subjective) descriptive modifiers through subjectification and hence to be more strongly subjective than subjective descriptive modifiers like bad (§1.3.2.4; §3.1). Descriptive modifiers specify properties to which the instance referred to corresponds. Degree modifiers, by contrast, explicitly evoke a scale on which the speaker locates the ‘degree of the properties’, which is a matter of speaker assessment and stance. When the degree modifier follows the descriptive modifier, an element generally considered to be more subjective is to the right of one that is less subjective, rather than to the left as would be expected in a subjective-objective cline. In addition to the ordering patterns just discussed, yet another phenomenon has been noted that argues against maintaining the linear subjectiveobjective model in its current form, i.e. scopal adjectives. Following McGregor (1997), Vandelanotte (2002: 246) argues that “adjectives expressing strong speaker’s attitudes [i.e. interpersonal adjectives] involve scope”. Expletives such as damn and bloody, which “possess only expressive and no descriptive meaning” (Cruse 2000: 59), qualify as scopal adjectives.25 The scopal relationship is represented in Figure 10. 25. According to Cruse (2000: 59), damn is not a true expletive. Whereas bloody is claimed to have only expressive meaning, damn is argued to have both descriptive

A semantic, linear subjective-objective model? 73 damn ATT this

spoon

Figure 10. Scopal relation of expletives (Vandelanotte 2002: 246)

The adjective involved does not only have scope over the word it immediately precedes. It rather has scope over the entire noun phrase. More specifically, the scopal relation is attitudinal (ATT) in nature, specifying the speaker’s attitude (see McGregor 1997). As the adjectives “indicate the very private emotive response of one particular person [i.e. the speaker] at one particular occasion vis-à-vis these instances” (Vandelanotte 2002: 246), they can be considered interpersonally subjective. Also, “[i]nasmuch as the attitudinal scope involved does not really bear on the referent entity designated by the head noun, ... it should perhaps come as no surprise that the scoping elements can occur in ‘non-canonical’, non-expected places”, leading them to disrupt the subjective-objective ordering pattern in a noun phrase, like bloody in your whole condescending bloody attitude (Vandelanotte 2002: 246). Most extremely, expletives can even occur wordinternally as in abso-fuckin’-lutely, fan-fuckin’-tastic and those Aladamnbama farmers (Vandelanotte 2002: 246). The above discussion has shown that there are numerous counterexamples to the proposed subjective-objective ordering of the NP, which makes it hard to hold on to the model in its current form. 2.3. A prosodic, field-like model Section 2.2 has shown that the linear subjective-objective model of NP order has both strengths and weaknesses. At this point, I want to consider an alternative hypothesis for the ordering of elements in the NP, i.e. a prosodic, field-like model. Although Halliday (1994: 187) sees a general subjective-objective continuum in NP structure from “the kind of element that has the greatest specifying potential to that which has the least”, he also remarks that “[i]nterpersonal meanings tend to be scattered prosodically

and expressive meaning. I have to admit the difference between the two elements is not entirely clear to me.

74

Structure, order and (inter)subjectivity

throughout the unit” of the NP, in what Pike ([1959] 1972a) called a ‘field’like pattern (Halliday 1994: 190). According to Halliday and Hasan (1976), the linguistic system comprises three main semantic components or types of meaning, i.e. the ideational, the textual, and the interpersonal. (§1.1.1) First, the ideational component is concerned with representational, descriptive semantics, with the expression of ‘content’. Second, the textual component covers all discourse-related and text-building meanings. Third, the interpersonal component involves all speaker-hearer related meaning in language. Importantly, Halliday ([1979] 2002c: 200) argues that “each of these semantic components typically generates a different kind of structural mechanism as its output, or realization”. In other words, “each component of the semantic system specifies its own structures” (Halliday [1977] 2002b: 24). First, Halliday (2002c: 204) states that “[t]he basic structural mechanism [for experiential, and by extension, ideational meaning] is that of constituency, with larger units constituted out of layered clusters or bracketed strings of smaller units, each part having its function with respect to the whole.” For the clause, Halliday exemplifies this with the different functions that together can make up a clausal structure, e.g. Process, Actor, Goal, Extent and Manner. For the NP, this corresponds with the structural dependency analysis (§2.1) and the constituency analysis (§1.1.1), which constitute ideational systems that “generate part–whole structures; they are realized by organic configurations which themselves, and whose constituents, are reasonably clearly bounded, such that it can be specified where one ... element leaves off and the next one starts” (Halliday [1981] 2002d: 238, 240). Second, textual meanings are argued to generate periodic structures, which mark off units of the message as extending from one peak of prominence to the next. “[A] clause has a wave-like periodic structure created by the tension between theme–rheme (where theme is the prominent element) and given–new (where new is the prominent element); the result is a pattern of diminuendo–crescendo, with a peak of prominence at each end” (Halliday 2002d: 233). To the extent that NPs occur in successive information units, the flow of information also runs through them. Also, within NPs there is a similar progression of information as in the clause: the NP “my three brilliant articles likewise goes from a deictic my, locating the object in speaker space, to a lexical item articles, again giving the main piece of information” (Halliday 2002d: 243). Third, Halliday (2002c: 205) states that interpersonal meaning is typically realized prosodically, i.e. it is “distributed like a prosody throughout a

A prosodic, field-like model

75

continuous stretch of discourse”. ‘Prosody’ is used here in a semantic sense, as in Sinclair’s notion of semantic prosody, which applies the notion of ‘suprasegmental’ patterning, known from phonology, to semantics. For the clause-level, Halliday (2002d: 238) stated that [i]nterpersonal systems tend to generate prosodic patterns that run all the way through the clause: not only intonation contours, though these provide a clear instance, but also reiterations of various kinds like those that are typical of modality in English, e.g. surely ... can’t ... possibly ... can ... d’you think in: Surely they can’t possibly be serious about it can they d’you think?

Similarly, interpersonal or subjective meanings are spread throughout the entire NP. According to Halliday (2004: 328), “[i]nterpersonal meanings are embodied (a) in the person system, both as pronouns ... and as possessive determiners; (b) in the attitudinal type of Epithet [i.e. subjective descriptive modifiers], e.g. splendid [...] ; (c) in connotative meanings of lexical items functioning in the group [e.g. subjectively coloured heads, both simple and compound], and (d) in prosodic features such as swear-words and voice-quality” (cf. Poynton 1996: 214–215; Martin 2008). One can modify the extent of quantifiers (very many), of descriptive modifiers (very nice) and gradable nouns (complete idiot) and even of classifiers (a largely political issue). Also, degree modifiers can occur both in front of the element they modify (e.g. degree modifiers of adjectives and nominal descriptions) and following it (e.g. old reinforcing the adjective it follows, §2.6). In other words, subjective meaning is “scattered prosodically throughout the unit” of the NP (Halliday 1994: 190) and may intersperse with nonsubjective meaning. Similarly, degree modification, as a phenomenon related to but distinct from subjectivity, can also be considered a prosodically distributed device. In addition, semantic prosody is - as the term implies also realized prosodically (see Martin 2008). Partington (1998: 68) defines semantic prosody as “the spreading of connotational colouring beyond single word boundaries”. Prosodic structures, unlike constituency and periodic structures, cannot straightforwardly be translated into configurations of discrete elements. As a consequence, “they are either ignored or forced into the constituency mould”. As mentioned in Section 1.1.4, it is difficult to represent degree modification in a flat structure of the NP, envisaged as a stretch of zones following one another. Most models of the NP thus simply do not incorporate degree modification as a (separate) phenomenon. Subjectivity on the other hand is not ‘ignored’, but mapped onto the constituent structure of the

76

Structure, order and (inter)subjectivity

NP. However, as discussed in Section 2.2.7, a view of subjectivity as running through the constituency structure from left to right from strong to weak cannot be upheld without distorting the facts to a higher or lesser degree. In Halliday’s (2002c: 209–210) words, any attempt to represent interpersonal structures such as subjectivity and degree modification “in constituency terms involves idealizing them to an extent that is tantamount to a form of reductionism”. Although there might be a general tendency for more subjective elements to precede less subjective ones in NP structure, one also has to be able to account for the (minor) countercurrent of more subjective elements following less subjective elements. At this moment, it seems that a prosodic model of the NP as envisaged by Halliday is best fit for this. Halliday (e.g. 2002c: 211) relates the semantic components of language – ideational, textual and interpersonal – and the structures they generate to Pike’s (1972a) view of language as particle, wave and field. The ideational constituency and dependency structure of the NP is particulate, textual periodic structures are wave-like and interpersonal prosodic structures are field-like. Pike (1972a: 129, 1972b: 183) argues that “language as a whole, and subsystems of language”, e.g. the clause and the noun phrase, can be viewed from three distinct but complementary perspectives, namely as particle, wave and field. First, in a static understanding, language is “made up of PARTICLES – ‘things’, pieces, or parts with sharp borders” (Pike 1972a: 129). In the structural dependency and constituency analysis of the NP, for instance, the different elements in the NP are more or less viewed as “permanent bricks juxtaposed in a permanent structure” (Pike 1972a: 129). Second, from a more dynamic perspective, language can be viewed “as being made up of WAVES following one another” (Pike 1972a: 129). According to Halliday (2002d: 238, 240) the distribution of information across information units is of this wave-type nature. Third, language as a whole and the NP in specific can be viewed as a “total FIELD”, i.e. “as a system with parts and classes of parts so interrelated that no parts occur apart from their function in the total whole” (Pike 1972a: 129). As noted by Martin (2008), although Halliday adopts Pike’s notion of field, he attaches a different understanding to the concept as “a region under the influence of an agent in physics”. Interpersonal meanings are understood to function prosodically and to be strung throughout the clause. Different subjective elements in the NP thus work together and have a cumulative effect all contributing to the expression of speaker evaluation. Each subjective element functions as a kind of force field, “which in effect

A prosodic, field-like model

77

run[s] over the top of experiential meanings” (Poynton 1996: 216). However, no matter how cumulative the effects of the distinct subjective elements may be, the realization of interpersonal, subjective meaning is necessarily located discontinuously. Pike (1982: 123), in contrast, approaches ‘field’ from an inherently “relational perspective” with prime focus on “the interdependent characteristics of a system”. A field view of things highlights organization and integration into a system (Pike 1985: 449). A field view of a discipline as an intersection of various parameters of analyst viewpoints on the world outside of him has the great advantage of flexibility. One can come at the data from various theoretical or empirical directions, and therefore mine a richness and multiple interlocking cohesion of principles which is lacking if one attempts to follow a single rigid logical sequence of presentation ... apart from observer choice. (Pike 1982: 109)

For the NP, such a field view is equally advantageous. In the NP system a number of semantic and grammatical dimensions (recursive modification, subjectivity, degree modification, information structure) are meshed together, with each dimension entailing alternative ordering preferences which are ultimately all integrated into an overall order, constrained by NP structure. As mentioned above, in Halliday’s view each semantic component specifies its own structure. “It is the function of the lexicogrammatical stratum [i.e. syntax, morphology and lexis] to map the structures one on to another so as to form a single integrated structure that represents all components simultaneously” (Halliday 2002b: 24). Accordingly, the NP system is a “simultaneous networ[k] of options”, a “multi-layered ... structural compositio[n]” (Halliday 2002b: 45), or in Pike’s terminology a field. In sum, I will assume the hypothesis that a field model of the NP can better account for the ordering of elements within the NP, as constrained by the simultaneous realization of different kinds of structure, viz. the particulate constituency structure of ideational meaning, the wave-like periodic structure of textual meanings and the prosodic, ‘field’-like structure of interpersonal meanings. Through the recognition of different types of structure on each semantic level divergent types of realization can more easily be accommodated. The linear subjective-to-objective order of the NP is then rethought as a prosodic structure in which subjective, interpersonal elements appear scattered throughout – albeit centred around specific zones which may function as attractor poles for other subjective elements. In such a view, the seem-

78

Structure, order and (inter)subjectivity

ingly uncontrolled positioning of expletives and other scopal elements in the NP, the realization of degree modification both prededing and following the modified elements, as well as for the possible subjectivity of the head of the NP – either in its realization as a simple noun or as a compound noun – can easily be accounted for. In this respect, Chapter 8 on subjective compounding also provides a key argument for the prosodic, ‘field’-like model of the NP.

Part II Diachrony The art of progress is to preserve order amid change and to preserve change amid order. Alfred North Whitehead

Chapter 3 (Inter)subjectification and grammaticalization: Pathways of change in the English NP 3.1. Subjectification and intersubjectification Sections 2.2.1 to 2.2.4 discussed the two main notions of (inter)subjectivity most frequently attested in cognitive-functional literature, i.e. Langacker’s and Traugott’s. These synchronic notions can however be translated into diachronic processes of (inter)subjectification in which the construal or meaning of a linguistic element becomes increasingly subjective. As discussed in Chapter 2, subjectivity, as it is defined by Langacker (1985, 1990, 1991, 2006), pertains not to the semantics of linguistic elements but to their construal, with objective construal involving an on-stage conceptualizer and subjective construal involving an off-stage conceptualizer. Originally, Langacker (1990, 1991) conceptualized the development of subjective construal, or subjectification, as “a semantic shift or extension in which an entity originally construed objectively comes to receive a more subjective construal” (1991: 215) through simple “realignment” or replacement (1990: 17). Later, subjectification was redefined as a gradual diachronic process of “progressive attenuation”, whereby “an objective relationship fades away, leaving behind a subjective relationship that was originally immanent in it” (Langacker 1998: 75–76). This Langackerian notion of subjectification has been applied mainly to the development of modal verbs into auxiliaries (e.g. Langacker 1990, 1999 for English; Cornillie 2006 for Spanish; Mortelmans 2006 for German). Application to (elements of) the NP is less straightforward. The importance of the relation to the ground in Langacker’s view on subjectification naturally orients it toward developments involving determiners in the NP and less toward developments involving descriptive modifiers or their degree modification (but see Athanasiadou 2006, 2007; Pander Maat 2006). Langacker’s notion of subjectification has been referred to in the study of postdeterminers by Breban (2002, 2010a) and Davidse, Breban, and Van linden (2008) and will also be invoked in the discussion of secondary determiners in this study (§3.4.1). However, as the focus of this study is on both degree modification

82

(Inter)subjectification and grammaticalization

and (secondary) determination, I will mainly make use of Traugott’s more comprehensive notions of subjectification and intersubjectification.26 Traugott’s notion of subjectification originated in an attempt to grasp the semantic changes witnessed in early grammaticalization. Over the last twenty-five years Traugott has gradually developed and further focused her hypotheses on subjectification as independent of the process of grammaticalization – even though they frequently occur as each other’s concomitant – and added the notion of intersubjectification to her theoretical apparatus. Traugott (1982) developed the following cline of semantic change to account for the semantic-pragmatic facets of early grammaticalization: propositional > ((textual) > expressive))

(Traugott 1989: 31)

The elements of this cline refer to Halliday and Hasan’s (1976) three-way distinction of the linguistic system into the functional-semantic components referred to as the ideational, the textual and the interpersonal (§1.1.1). Traugott (1982) renamed the ideational and interpersonal components as ‘propositional’ and ‘expressive’ respectively and defined them as follows. The propositional component comprises all “truthconditional relations”, which are “subject to referential verification” (Traugott 1982: 248). “The textual component has to do with the resources available for creating a cohesive discourse”, and the expressive component “bears on the resources a language has for expressing personal attitudes to what is being talked about” (Traugott 1982: 248). Whereas Halliday and Hasan’s (1976) semiotics were designed as a synchronic account of the linguistic system, Traugott (1982: 256) transformed them into a unidirectional diachronic hypothesis, arguing that If there occurs a meaning-shift which, in the process of grammaticalization, entails shifts from one functional-semantic component to another, then such a shift is more likely to be from propositional through textual to expressive than in reverse direction.

Although Traugott (1982) does not yet use the term subjectification in relation to the propositional > textual > expressive cline, she does describe the changes involved as proceeding from “less personal to more personal” in the sense of “more anchored in the context of the speech act, particularly 26. The Cognitive Grammar notion of intersubjectivity put forward by Verhagen (2005, 2007) has not (yet) been developed diachronically (§2.2.2).

Subjectification and intersubjectification

83

the speaker’s orientation to the situation, text, and interpersonal relations” (Traugott 1982: 253). Note that the shifts are not obligatory in that they do not have to occur, they do not have to proceed all the way down the cline, and not every step in the cline is mandatory, e.g. meanings can shift from the propositional component directly to the expressive component. By way of illustration of the proposed pathway of change, Traugott (1982) provides a number of examples, including shifts within the structure of the English NP. The development of the definite and indefinite article is mentioned as “the evolution of a grammatical marker the prime function of which is textual” (Traugott 1982: 252), and the development of adjectiveintensifiers like awfully and horribly is said to instantiate “shifts from propositional items to grammatical markers at the expressive, interpersonal level” (Traugott 1982: 253). Ten years later, Traugott (1989: 34–35) reformulates her original cline leading from propositional through textual to expressive meanings as three separate tendencies, not restricted to grammaticalization but found in processes of semantic change in general:27 Tendency I: Meanings based in the external described situation > meanings based in the internal (evaluative/perceptual/cognitive) described situation. Tendency II: Meanings based in the external or internal described situation > meanings based in the textual and metalinguistic situation. Tendency III: Meanings tend to become increasingly based in the speaker’s subjective belief state/attitude toward the proposition

First, Tendency I can be illustrated by means of processes of “pejoration and amelioration (e.g. boor ‘farmer’ > ‘crude person’)” (Traugott 1982: 34) and by semantic shifts from objective to subjective descriptive modifier uses (e.g. Dutch leuk ‘neither hot nor cold’ > ‘pleasant, nice’ [De Smet and Verstraete 2006: 373]) (§2.2.3). As such, Tendency I actually captures changes within the propositional domain as it was defined in Traugott (1982). Second, Tendency II encompasses the development of meanings contributing to the creation of text (textual) and to meanings that reflect on language (metalinguistic or metatextual). The former type includes the 27. The three tendencies are translated in Visconti (2013) into three distinct types of subjectification: (i) lexical subjectification, as in pejoration or amelioration; (ii) textual subjectification, concerning the development of devices coding cohesion; and (iii) attitudinal subjectification, involving an increase in the coding of speaker attitude.

84

(Inter)subjectification and grammaticalization

development of cohesive devices such as connectives, the latter type can be illustrated by the development of speech-act verbs. The development of the definite and indefinite article, analyzed as a shift to the textual domain in Traugott (1982), is probably also best categorized under Tendency II. Finally, changes consistent with Tendency III entail an increase in subjectivity, as it is witnessed in, for instance, the development of epistemic meanings. For the NP, Traugott (1989: 35) gives the example of the degree modifier very: “borrowed in Middle English from Old French verai ‘true’ (a cognitive evaluation), in Early Modern English it became a scalar particle as in the very height of her career (a subjective evaluation)”. This is consistent with the (1982) analysis of degree modifiers as having shifted from the (internally) propositional to the expressive domain. The examples given show that the cline of semantic change proposed in Traugott (1982) nicely matches the three tendencies distinguished in Traugott (1989). In Traugott (1995: 47) subjectification is defined as “the tendency to recruit lexical material for purposes of creating text and indicating attitudes in discourse situations”. This definition seems to encompass both Tendency II and Tendency III, as subjectivity is understood to encompass both textually and attitudinally subjective meanings. However, Traugott’s (2003a: 126) more recent definition of subjectivity as the encoding of the speaker’s attitudes and beliefs seems to focus more exclusively on speaker-attitude and speaker-stance. Now correlating with Tendency III only, subjectification in its present definition entails that “meanings become more deeply centred on the speaker” (Traugott 2003a: 129) and are “increasingly based in the SP/W’s subjective belief state or attitude toward what is being said and how it is being said” (Traugott 2003a: 129). Irrespective of whether subjectivity is understood broadly as encompassing both attitudinal and textual meanings (Traugott 1995) or more narrowly as attitudinal only (Traugott 2003a), the direction of semantic change is typically from non- or less subjective to subjective. In line with Benveniste’s (1958) landmark paper, Traugott and Dasher (2002) and Traugott (2003a) added the notion of intersubjectivity and its diachronic counterpart intersubjectification to existing theorizing on semantic change. Accordingly, the pathway of semantic change is re-defined once more to take the present form non-/less subjective > subjective > intersubjective (Traugott and Dasher 2002: 225)

Subjectification and intersubjectification

85

Historically, non- or less subjective meaning precedes subjective meaning, which in turn precedes intersubjective meaning. In other words, “for any lexeme intersubjectivity is historically later than and arises out of subjectification” (Traugott 2003a: 130). In intersubjectification, meanings become more focused on the addressee or reader than on the speaker. Traugott (2003a: 130) defines intersubjective meanings as “encod[ing] or external[izing] implicatures regarding SP[eaker]/W[riter]’s attention to the ‘self’ of AD[dressee]/R[eader] in both an epistemic and a social sense” (Traugott 2003a: 130). As detailed in Section 2.2.5, I argue that the notion of intersubjectivity can be extended to include textual meanings, analogous to the broad definition of subjectivity as also encompassing a textual component (Traugott 1995).

Externally propositional

Textual

Internally propositional

Expressive

Figure 11. Integrated model of semantic change

Figure 11 represents an integrated model of Traugott’s successive clines of semantic change, taking the original (1982) pathway as starting-point. In line with Tendency I recognized in Traugott (1989), the propositional domain is split up into two subdomains, i.e. the externally propositional and the internally propositional domains. Whereas the textual component is

86

(Inter)subjectification and grammaticalization

understood to accommodate both objective, subjective and intersubjective meanings (§2.2.5), the expressive (or attitudinal) component is – by its very nature – taken to accommodate only subjective and intersubjective meanings. If the cline from propositional to textual to expressive meanings is understood as a diachronic continuum then of course this is problematic for any unidirectional pathway leading from objective to subjective to intersubjective meaning, as it was proposed in e.g. Traugott (2003a). 3.2. Grammaticalization Grammaticalization is in this study understood as The process whereby lexical material in highly constrained pragmatic and morphosyntactic contexts is assigned grammatical function, and once grammatical, is assigned increasingly grammatical, operator-like function. (Traugott 2003b: 645)

In this sense, grammaticalization is a diachronic process of language change whereby lexical material is made grammatical, i.e. primary grammaticalization, or already grammatical material is made increasingly grammatical, i.e. secondary grammaticalization (see Givón 1991: 305). Although crucially involving grammatical change, the grammaticalization of a linguistic element is typically accompanied by other types of language change – semantic, morphological, phonological or pragmatic – equally affecting the output of the grammaticalization process. In earlier theorizing on grammaticalization, the focus was on morphosyntactic change (e.g. Bybee, Perkins, and Pagliuca 1994; Lehmann 1985, [1982] 1995, 2002, 2004; Haspelmath 2004), typically viewed as proceeding along the following cline: discourse > syntax > morphology > morphophonemics > zero (Givón 1979: 209) A major theoretical consequence of thinking in terms of clines such as the above is that grammaticalization is taken to be a unidirectional process; a hypothesis that is still widespread and considered a firm tendency in current research. Also, grammaticalization is viewed as a gradual process with many intermediary steps or degrees of grammaticalization. Lehmann (1995) put forward six parameters – attrition, paradigmaticization, obligato-

Grammaticalization

87

rification, condensation, coalescence and fixation – which typically correlate and jointly determine the degree of grammaticalization of a linguistic item or paradigm. The parameters and their meanings are listed in Table 2. Although subject to criticism and refinement over the years (e.g. Diewald 2010), the parameters remain highly influential and will also be used in this study in relation to the structural-semantic shifts witnessed in the NP (§3.4.1; §3.4.2.1; §3.4.2.2). Table 2. Lehmann’s grammaticalization parameters (Lehmann 2002: 146) parameter integrity paradigmaticity paradigmatic variability structural scope bondedness syntagmatic variability

weak grammaticalization bundle of semantic features; possibly polysyllabic item participates loosely in semantic field free choice of items according to communicative intentions

process

strong grammaticalization attrition few semantic features; oligo- or monosegmental paradigmati- small, tightly cization integrated paradigm obligatorifica- choice systematically tion constrained, use largely obligatory condensation item modifies word or stem

item relates to constituent of arbitrary complexity item is coalescence independently juxtaposed item can be shifted fixation around freely

item is affix or even phonological feature of carrier item occupies fixed slot

The parameters put forward by Lehmann (1985, 1995 [1982]) have been nuanced in a number of respects. First, Hopper (1991: 21) argued that the parameters are “characteristic of grammaticization which has already attained a fairly advanced stage and is unambiguously recognizable as such”. As an alternative, Hopper (1991: 21) has introduced five heuristic principles “characteristic of grammaticization not only at the later, more easily identifiable stages, but also at the incipient stages”. These principles are layering, divergence, specialization, persistence and de-categorialization. Layering initially referred to “the fact that very often more than one tech-

88

(Inter)subjectification and grammaticalization

nique is available in a language to serve similar or even identical functions” (Hopper 1991: 23). Hopper and Traugott (2003: 124) redefined the notion as referring to a situation of synchronic variability where older layers of meaning of a linguistic item may coexist with newer layers of meaning. It is this use of the term layering that will be adopted in this book. In this interpretation layering is very similar to the principle of divergence. Hopper (1991: 24) noted that divergence can be seen as a special case of layering, the latter involving different grammaticalized forms within one functional domain and the former involving different grammaticalized meanings of one form in various contexts. In other words, divergence refers to the fact that “when a lexical item undergoes grammaticization ... the original form may remain as an autonomous lexical item and undergo the same changes as any other lexical items” (Hopper 1991: 24). Specialization closely corresponds to Lehmann’s parameter of obligatorification and “refers to the narrowing of choices that characterizes an emergent grammatical construction” (Hopper 1991: 25). Persistence refers to the phenomenon whereby semantic features of the lexical source uses persist in the meaning of the grammaticalizing item. Finally, de-categorialization refers to the loss of morpho-syntactic features of one category and the acquiring of features of another category. Second, the more traditional view of grammaticalization depicted above is basically one of reduction (cf. Traugott 2010b; Brems and Hoffmann 2012), defined as “an evolution whereby linguistic units lose in semantic complexity, pragmatic significance, syntactic freedom, and phonetic substance” (Heine and Reh 1984: 15). This reductionist view is challenged in later, more functionalist oriented approaches to grammaticalization. Himmelmann (2004), for instance, has posited that grammaticalization typically involves three types of context-expansion: (i) host-class expansion, i.e. expansion of the class of elements with which the grammaticalizing item co-occurs; (ii) syntactic context expansion, i.e. change in the larger syntactic context in which the construction is used; and (iii) semantic-pragmatic context-expansion, i.e. expansion of the semantic and pragmatic contexts in which the construction is used. The example given in Himmelmann (2004: 32–33) is the development of a demonstrative into a definite article. Such change is characterized by broadening of the item’s collocational set to, for instance, proper nouns (e.g. The Hague) and nouns designating unique entities (e.g. the sun) (i.e. host-class expansion) and by widening of the syntactic environments it can occur in. “Emerging article grams typically occur first in core argument positions (subject or object position) and less commonly, or not at all, in adpositional expressions. When grammaticization

Grammaticalization

89

progresses further, use of the construction with an article may also become obligatory in adpositional expressions and other syntactic environments it did not occur in before” (Himmelmann 2004: 32), i.e. the article undergoes syntactic context expansion. Pragmatically, definite articles but not demonstratives can be used for e.g. associative anaphors (e.g. a wedding – the bride) (i.e. semantic-pragmatic context expansion). Importantly, Himmelmann (2004) considers semantic-pragmatic context expansion to be the core defining feature of grammaticalization, the other types being possible but not necessary concomitants of grammaticalization. Traugott (2010a: 45), in contrast, argues that in (primary) grammaticalization “all three aspects of form-meaning pairings (lexical collocation, syntactic structure, and semantics-pragmatics) change (but not necessarily at the same time)”. Semantic-pragmatic context expansion in grammaticalization has featured prominently in the work of Traugott (1982, 1989, 1995, 2003a) and Hopper and Traugott (2003). Against the traditional reductionist view on semantic change in grammaticalization as ‘bleaching’ or a mere loss of referential, lexical meaning, they have argued that early grammaticalization correlates with an increase in pragmatic meaning, which can later be semanticized through invited inferencing or pragmatic strengthening. For instance, in addition to referring to a cluster of things fastened or growing together, the size noun construction a bunch of flowers/carrots/parsley also contains information on the typical size of a bunch. In specific discourse contexts, these size implicatures may be reinforced and, given that they are invited with considerable frequency, a quantifier sense may develop as a separate conventionalized use (Brems and Hoffmann 2012). Invited inferencing is thus a phenomenon which originates in the interaction between speaker and hearer. Speakers exploit conversational implicatures and invite addressees to infer meaning. These innovative utterance-token meanings may become conventionalized as generalized implicatures and may eventually be semanticized into an item X. (Traugott 2008: 3)

As the new, more grammatical meaning is semanticized a new polysemy arises, in which, in the case of subjectification, “the speaker’s perspective is an essential element” (Traugott 2003b: 634). Rather than a loss of meaning, there is a redistribution of meaning in grammaticalization: the original, more lexical meaning is backgrounded whereas the new, more grammatical meaning is foregrounded or enriched.

(Inter)subjectification and grammaticalization

90

3.2.1. Onset contexts of grammaticalization The view that most instances of grammaticalization originate in invited inferencing is now widespread. Recently, grammaticalization studies have, amongst other things, paid growing attention to the motivations of change, particularly to the contexts which can give rise to invited inferences and hence enable the onset of grammaticalization. The awareness of and attention to the importance of syntagmatic and discourse context goes hand in hand with a more constructional approach to grammaticalization (§1). Rather than affecting single words in isolation, grammaticalization processes affect (strings of) linguistic items or constructions in very specific environments. It will be argued in this section that at least three types of onset contexts of grammaticalization – ambiguous contexts, dialogic contexts, and specific collocational contexts –, the latter two of which can be referred to as specialized contexts. 3.2.1.1. Ambiguity Evans and Wilkins (2000: 550) argued that extended meanings typically originate in bridging contexts, i.e. regularly occurring contexts support[ing] an inference-driven contextual enrichment of [an earlier meaning] A to [a later meaning] B. In these contexts, which we call BRIDGING CONTEXTS, speech participants do not detect any problem of different assignments of meaning to the form because both speaker and addressee interpretations of the utterance in context are functionally equivalent, even if the relative contributions of lexical content and pragmatic enrichment differ. [caps original]

Like Evans and Wilkins (2000), Heine (2002) and Heine and Kuteva (2007) focus on semantic-pragmatic ambiguity in the onset contexts of grammaticalization. They distinguish four non-discrete, but necessary stages in grammaticalization (Heine 2002: 86) [italics added]: I. II.

“[T]here is an expression with a ‘normal’ or source meaning occurring in an array of different contexts.” “[T]here is a bridging context giving rise to an inference to the effect that, rather than the source meaning, there is another meaning, the target meaning, offering a more plausible interpretation of the utterance concerned.”

Grammaticalization

91

III. “[T]here is a new type of context, the switch context, that no longer allows for an interpretation in terms of the source meaning. Switch contexts may be viewed as a filtering device that rules out the source meaning.” IV. “Finally, no longer being associated with the source meaning, the target meaning is now open to further manipulation: It is freed from the contextual constraints that gave rise to it. ... I will refer to this situation as the conventionalization stage.”

Diewald (2002, 2006) similarly distinguishes three contexts or stages in grammaticalization, but defines them structurally as well as semanticpragmatically: “Untypical contexts”, in which “the new meaning ... arises as a conversational implicature, i.e. this meaning is contextually and pragmatically triggered and not explicitly encoded in the linguistic items themselves” II. “Critical contexts”, i.e. “highly marked construction[s] ... characterized by multiple structural and semantic opacity, thus inviting several alternative interpretations, among them the new grammatical meaning” III. “Isolating contexts for both the lexical and the grammaticalized readings, i.e. specific linguistic contexts that favour one reading to the exclusion of the other”. (Diewald 2006: 3–4) [italics added] I.

Traugott (2008, 2010c, 2010d) aligns Heine’s (2002) and Diewald’s (2002, 2006) models as follows. Heine I. Normal use II. Bridging contexts III. Switch contexts IV. Conventionalized contexts

Diewald 0. Normal use I. Untypical contexts II. Critical contexts III. Isolating contexts

Figure 12. A comparison of Heine’s (2002) and Diewald’s (2002, 2006) contexts of grammaticalization (Traugott 2008, 2010c, 2010d)

The two models display both convergences and divergences. Whereas Heine (2002) focuses more exclusively on the semantic changes involved, Diewald’s (2002, 2006) critical contexts are multiply opaque, i.e. they are ambiguous not only on the semantic level, but also on the formal level. In addition, whereas Diewald (2002, 2006) provides for specific context types

92

(Inter)subjectification and grammaticalization

(typically the critical contexts) to be lost during the grammaticalization process, the contexts described by Heine (2002) form an “implicational scale”. “If a given language is found to have a stage IV situation, then it can be expected to also distinguish all preceding stages” (Heine 2002: 95). Nevertheless, both Heine (2002) and Diewald (2002, 2006) highlight the importance of pragmatics as it is present in conversational implicatures or invited inferences. Also, although the ambiguous contexts – either semantic-pragmatic only or structural, semantic and pragmatic – need not lead to grammatical change, Heine (2002) and Diewald (2002, 2006) consider them to be a necessary condition for grammaticalization. However, ambiguity as a principle motivating language change is contentious. For reanalysis, for which ambiguity has long been considered a crucial motivating factor, Fischer (1988, 2007), Los (2005) and De Smet (2009: 1729) amongst others have argued that “the ambiguities that are supposed to motivate reanalysis are really the result of reanalysis, as they can only arise if the target structure of reanalysis already exists”. Similarly, semantic ambiguity, which characterizes both bridging contexts (Heine 2002) and critical contexts (Diewald 2002, 2006), can arise only when the old and new meanings are already available to the language user. Moreover, for historical data ambiguity is always constructed retroactively, with hindsight knowledge of later semantic and morpho-syntactic developments. Ambiguities perceived by a Present Day language user may but need not have been perceived by some, let alone all, past time language users (see also Traugott 2008). Although ambiguity may often to some extent accompany grammaticalization, it cannot fully explain its onset. Ambiguity may be witnessed in the early stages of grammaticalization but does not constitute a necessary stage. Also, if there is a stage of systematic ambiguity, ambiguous examples need not disappear in later stages of grammaticalization. “Opaque examples appear throughout the life of polysemous constructions, ... as points of maintenance of the link between the originally related constructions” (Traugott 2008). Following Traugott (2008), I thus view ambiguity not as an explanatory factor in language change but rather as a possible silent witness of the genetic link between two constructions. 3.2.1.2. Specialized contexts: Dialogicity and collocation According to Traugott (2008, 2010c, 2010d), ambiguous contexts are not a necessary concomitant of grammaticalization. What grammaticalized items do show are ‘specialized contexts’ in which the new meaning may develop.

Grammaticalization

93

These contexts are characterized by semantic, morpho-syntactic and/or pragmatic features and need not be restricted to just one specific context, but may take the form of a range of contexts. An example discussed in Traugott (2008, 2010c, 2010d) is the development of pseudo-clefts of the type all/what I did/said was X, as in (47) to (49). What is important for the development of the pseudo-cleft in which all means ‘only’ rather than ‘everything’ is the contesting context and the presence of linguistic elements such as but in (47), which as a focus marker “carries the meaning of exhaustiveness”. “Over time, association of the construction with exhaustiveness led to the disappearance of focus particles like only [and but], but the pseudo-clefts continue to be favored in contesting contexts”, as in (48) “where the speaker had only to listen, not do anything else such as dance”. When the pseudo-clefts are fully grammaticalized, non-contesting contexts can, however, also be found. In (49), the context is no longer dialogic but the cleft construction in itself signals dialogicity, indicating that to stand and laugh at him is the only option. (47)

(48) (49)

What need’st thou woman such a whyning keepe? Thy sonn’s as well as anie man ith’lande, Why all he did, was bidd a man but stande, And told him coyne he lackt; there’s those doe worse, Then bidd an honest man deliver’s purse. (1616 Goddard, A Mastiff VVhelp [LION: EEBO] (Traugott 2008: 18) ... classical music was just ‘music’, and therefore all one had to do was to listen to it. (COBUILD 10 Million Corpus [Bonelli 1992: 36]) (Traugott 2008: 16) When any bow’d to me with Congees (= ceremonious bow) trim, All I could do, was stand and laugh at him. (1681Baxter, Poetical fragments [LION: EEBO] (Traugott 2010d: 20)

The motivating, enabling factor for the changes discussed is dialogicity. “Dialogic orientation concerns the extent to which speakers contest, refute or build an argument toward alternative or different conclusions (e.g. but, modal in fact)” (Traugott 2010d: 15). Such dialogic contexts can enable grammaticalization and often the dialogic aspect becomes an inherent aspect of the outcome of the grammaticalization and/or subjectification process. In these cases, “the context is no longer dialogic. The construction in itself signals dialogicity. We may say that dialogicity has been semanticized into the construction” (Traugott 2010d: 20). According to Traugott (2010d: 15), linguistic expressions that index some degree of dialogicity

94

(Inter)subjectification and grammaticalization

include adversatives, concessives, negation, epistemic modal adverbs, focus particles and scalars. The latter two, which are most relevant to this study, are dialogic to the extent that they invoke comparison with alternative values (see also §1.3.2.4). Focus particles “exclude alternatives and ‘carry an implication of dissonance or incompatibility’ (König 1991: 131; also Traugott 2006), e.g. even, only” (Traugott 2010d: 15). Scalar elements similarly invoke alternative values on a scale (Traugott 2010d: 15). In the data study, special attention will go to possible dialogic contexts which may have triggered the development of noun-intensifying uses, which are inherently scalar and dialogic. Dialogic contexts are one type of specialized contexts which may trigger processes of semantic change involved in grammaticalization. I would like to suggest that collocational contexts are another. As mentioned above, specialized contexts are defined broadly as characterized by semantic, morpho-syntactic and/or pragmatic features and may take the form of a range of contexts. Collocations are, I believe, just that. Collocational preferences are tied not to lexical items as such, but to their meanings or semantics and refer to a syntagmatic “relationship of habitual co-occurrence” (Stubbs 1995: 23) with other words, i.e. “different meanings [of the same linguistic item or construction] go along with ... different collocational patterns” (Bublitz 1996: 18). Recently, studies like Brems (2007, 2010), Vandewinkel and Davidse (2008) and Ghesquière and Davidse (2011) have shown that collocational extension and collocational reclustering may serve as a mechanism of semantic change. For the noun-intensifier pure, as in pure bliss or pure paranoid fantasy, Vandewinkel and Davidse (2008) have argued that, while originating in combination with head nouns expressing strong emotions (e.g. fury, hatred, desire), its development is characterized by an increase and diversification of the collocates of the original descriptive uses of the adjective. Similarly, Brems (2010) has argued that the development of size noun-constructions, such as a bunch/heap/load of N, involved processes of semantic generalization and collocational extension, as observed for pure. In addition, the size noun constructions displayed collocational reclustering characterized by semantic prosody constraints. First, the path from the original ‘head use’ of the size noun constructions, as in a bunch of grapes/flowers, to their ‘quantifier use’, as in a bunch of suits/workshops/people, is driven by collocational extension to include not only nouns which refer to a cluster of things growing or fastened together but a larger variety of concrete and abstract nouns. Second, ‘valuing quantifiers’, as in a bunch of hamfisted dimwits, involve “interpersonal collocational reclustering, rather than extension,

Grammaticalization

95

based on the sharing of affective or emotively coloured values between the nodal SN [size noun] and its collocates” (Brems 2010: 90). Whereas the step from head use to quantifier use thus involves host-class expansion (Himmelmann 2004), traditionally associated with grammaticalization processes, the step from quantifier to valuing quantifier use involves collocational reclustering and in fact a narrowing of the collocational range to nouns with a clearly negative semantic prosody. Following Brems (2007, 2010), Vandewinkel and Davidse (2008) and Ghesquière and Davidse (2011), I believe changes in collocational preferences or collocational contexts are one way of operationalizing semantic changes. Accordingly, in the corpus studies discussed in this book attention will be given to (changes in) the preferences of the items studied for combining with specific types of head nouns or modifiers. In the literature, it has been noted that noun-intensifiers may be collocationally restricted to specific sets of head nouns (Sinclair 1990: 69). An example is raving which is typically used with nouns such as madman, lunatic, nut, etc. as well as with some other collocates such as success and fan. Moreover, nounintensifiers often display specific preferences to combine with either predominantly positive or predominantly negative collocates. The nounintensifier utter, for instance, has a distinct negative semantic prosody, preferring combinations such as utter disgust/nonsense/rubbish/madness (cf. Bublitz 1996). Importantly, one can approach changes in collocational contexts from two perspectives. Brems (2010: 103) views collocational reclustering as “a facilitating factor, promoting emergent grammaticalizing readings, viz. the valuing quantifier use”. This type of use “thrive[s] on the sharing of affective or emotively coloured values between the nodal SN and its collocates” (Brems 2010: 97–98) and it is in such collocational contexts that this new use could develop. From a different perspective, De Smet and Ghesquière (2010) have argued that collocations do not facilitate but constrain or restrict grammaticalization. In preview of the case studies presented in Chapters 5 to 7, it can be noted that the noun-intensifier uses of the items studied are collocationally constrained to occur with nouns with the required semantic properties, i.e. degree nouns inviting intensification of inherent gradable notions. For the identifying uses no such restrictions hold. As De Smet and Ghesquière (2010) argue, the fact that noun-intensifiers engage with the specificities of the lexical elements they modify constrains their generalizability. Similar observations were made for the verb particle out (wear out, clear out, fill out) which also closely interacts with the lexically specific meaning of the verb it combines with. I believe the two perspec-

96

(Inter)subjectification and grammaticalization

tives complement rather than contradict each other. On the one hand, the development of new uses is triggered in and driven by specific collocational contexts (cf. Brems 2010), but at the same time the collocational restrictions are responsible for the fact that the new uses remain fairly infrequent despite their abstract meaning (cf. De Smet and Ghesquière 2010). As such, collocational contexts both enable grammaticalization, triggering and allowing the development of alternative meanings, and at a later stage also restrain grammaticalization, as the newly developed uses are constrained to specific word-partnerships. 3.3. The directionality of (inter)subjectification and grammaticalization It is generally assumed that (inter)subjectification and grammaticalization are unidirectional processes. Linguistic elements develop from objective to (inter)subjective and from less to more grammaticalized (but see Newmeyer 1998; Haspelmath 1999; Fischer 2000a; Fitzmaurice 2000), but not vice versa. Although hardly anyone will dispute that this is the dominant situation, a number of counterexamples to these tendencies have been identified in the literature. Refuting the unidirectionality of grammaticalization, Norde (2009) has constructed a typology of degrammaticalization changes, i.e. changes whereby a function word becomes a lexical item (i.e. primary degrammaticalization) or whereby a bound morpheme becomes less bound or free (i.e. secondary degrammaticalization). Degrammaticalization is evaluated against the parameters of resemanticization, phonological strengthening, recategorialization, deparadigmaticization, deobligatorification, scope expansion, decreased bondedness (severance) and increased syntactic freedom (flexibilization), i.e. the mirror-images of Lehmann’s (1995) parameters of grammaticalization (§3.2). An example is Modern Welsh eiddo, which gradually developed from possessive pronoun to lexical noun meaning ‘property’ (resemanticization). As a noun eiddo can now be preceded by adjectives or possessives (recategorialization) and its selection is not grammatically conditioned but only dependent on lexical content (deobligatorification). Further, in the shift from pronoun to noun eiddo shifted from a closed word class to an open class (deparadigmaticization) and has gained syntactic freedom, no longer needing a following predicate marker or article (flexibilization).

Directionality 97

A case study challenging the unidirectionality of subjectification is Pakendorf and Schalley (2007). On the basis of crosslinguistic data, they show that markers of possibility may develop into negative imperative markers, thus contradicting the generally accepted pathway of change from objective, deontic to subjective, epistemic meaning. Challenging the unidirectionality of intersubjectification, Cornillie (2008) describes the process of grammaticalization and intersubjectification observed in the Spanish evidential (semi-)auxiliaries parecer and resultar. The central argument is that the evidential verbs clearly undergo grammaticalization, but at the same time the case study suggests that the less grammaticalized, intersubjective evidential constructions (parecer que ‘seem that’ and resultar que ‘turn out that’) precede the more grammaticalized epistemic constructions (parecer + infinitive and resultar + infinitive) which have a subjective reading only. The shift from less to more grammatical is paralleled by a shift from intersubjective to subjective. In contrast, Matsumoto (1988, quoted in Traugott 1995) discusses the development of dakara ‘because’ through ‘therefore, that’s why’ to an epistemic insistence particle ‘I’m telling you’ as a case of subjectification on the one hand but degrammaticalization on the other hand (from subordinating clitic to free form). The above examples show that although there is a strong tendency for elements to develop from less to more grammatical and from objective to subjective to intersubjective, they are just tendencies, not absolute rules. Grammaticalization and (inter)subjectification processes can be overridden, reversed and stopped at any given moment in their development (Fischer 1997: 265; Traugott 2010b: 275). Also, subjectification is independent of grammaticalization, despite the fact that the processes often co-occur. As Traugott (2010a: 61) notes, “since grammaticalization involves the development of markers of speaker attitude toward the ideational component and toward textual connectivity (among many other things), there is inevitably a close interaction between grammaticalization and subjectification. The interaction of grammaticalization and intersubjectification is less common”. The latter statement appears to be motivated by the restricted interpretation of intersubjectivity and intersubjectification as involving attention to the social self of the hearer only. If, as advocated in Section 2.2.5, a wider notion of intersubjectivity is adopted, including (at least) an attitudinal, a responsive and a textual component, one can plausibly expect a closer association between intersubjectification and grammaticalization as well. Such interplay is witnessed, for instance, in deictification processes affecting adjectives (§3.4.1).

98

(Inter)subjectification and grammaticalization

3.4. (Inter)subjectification and grammaticalization in the English NP For the English NP, two main types of (inter)subjectification and grammaticalization processes have been observed in the literature, namely the development of (secondary) determining meanings and the development of degree modifying or intensifying meanings. As mentioned in Section 3.1, Traugott (1982, 1988, 1989, 1995, 2006) explicitly treats the development of determiners and degree modifiers as subjectification and grammaticalization processes, with the former conveying textual meanings and the latter conveying expressive ones. The case studies presented in Part II of this book provide evidence for both types of shift, i.e. to (secondary) determination and to (noun-)intensification. In the literature, distinct pathways of change have been proposed for each development. First, the development of secondary determiner uses is traditionally described in terms of processes of grammaticalization and semantic generalization (e.g. Breban 2002 for the adjectives of comparison; Brinton 2002 for temporal adverbs; Denison 2002 and Brems and Davidse 2010 for type noun constructions with kind/sort of). Davidse, Breban, and Van linden (2008) have proposed that the development of secondary determining meanings can receive a unified account in terms of a semantic shift by which a general relation expressed by the linguistic element is given a subjective reference point in or relative to the speech event (§3.4.1). Second, for the development of noun-intensifying meanings, two distinct pathways of change have been proposed in the literature. The now prevalent view is that noun-intensifying adjective uses develop from fully lexical descriptive uses (Paradis 1997, 2000, 2001; Adamson 2000; Traugott 2006; Athanasiadou 2007) (§3.4.2.1). Earlier, however, Bolinger (1972) posited a pathway from identifying to noun-intensifying meaning (§3.4.2.2). All of the shifts observed in the literature can be described in terms of grammaticalization and subjectification processes. 3.4.1. From description to secondary determination Secondary determination can be expressed by a wide range of elements, ranging from adjectives to adverbs, cardinal and ordinal numbers, relative quantifiers and nouns (§1.4.2). Historically, most adjectival, adverbial and nominal secondary determiners were originally used lexically as descriptive modifiers or head nouns. They have developed secondary determiner uses through processes of grammaticalization and semantic generalization

Intersubjectification and grammaticalization in the English NP

99

(Breban 2002; Brinton 2002; Denison 2002; Davidse, Breban, and Van linden 2008; Brems and Davidse 2010). The secondary determiner constructions involve form-function reanalysis (Croft 2000) of the original lexical constructions. The type nouns, for instance, were re-analyzed from lexical head noun, e.g. a more serious sort of problem, to enclitic of the determiner, e.g. these sort of problems (Denison 2002: 3). In the case of the adjectives, a descriptive modifier of the head noun, e.g. other in its lexical sense of ‘different’, as in far other scene (OED s.v. other 6), was reanalyzed as a submodifier of the determiner, e.g. another day. The adjectives thus develop from autosemantic to synsemantic signs (Lehmann 1985: 308; see Breban 2002). Whereas descriptive adjectives function as largely independent modifiers in NP structure, secondary determiner adjectives form a unit with the determiner (§1.4.2). In some cases, this functional coalescence (Lehmann 1985) is reflected orthographically, as in English another and Dutch dezelfde, hetzelfde (‘the same’). In other cases, the increase in bondedness is reflected in very strong collocational ties, e.g. the same, which the OED (s.v. same 4.a) describes as functionally one word. The development of secondary determiner uses also instantiates paradigmaticization (Lehmann 1985), as the adjectives come to figure in a more restricted and more homogeneous paradigm than the very extensive and varied set of descriptively used adjectives (Breban 2002). In the functional shift from lexical descriptive modifier to more grammatical secondary determiner, the adjectives also undergo decategorialization, i.e. loss of syntactic characteristics typically associated with the adjective word class (Hopper 1991).28 They typically lose the ability to be graded by submodifiers such as far and very, as well as the ability to be used predicatively without changing their specific semantics, e.g. *the day is other/same/different. In addition, as secondary determiners the adjectives can no longer be coordinated with descriptive modifiers, e.g. *the same and beautiful day. Davidse, Breban, and Van linden (2008) have described the semantic changes involved in the development of deictic secondary determiner uses by originally lexical adjectives as deictification: a process whereby the general relation depicted by the adjective acquires a reference point in or related to the speech event. The process is described for five adjectives, invoking the main deictic notions of space (opposite), identity (identical), quantity (complete), modality (necessary), and time (old), as in (50). 28. Denison (2006) formulates decategorialization positively as gradient category change, whereby an element moves away from the prototype of one category in the direction of the prototype of another category.

100 (Inter)subjectification and grammaticalization

(50)

Anders completed the distance in 20 minutes seven-point-five-twoseconds, taking nearly nine-and-a-half seconds off the old record, set by Australia’s Kerry Saxby in January. (CB) (Davidse, Breban, and Van linden 2008: 486)

Secondary determiner old above indicates that “the status or function described by the head noun which applied to the referent in the past has been superseded, but can still be used to bring about mental contact with that referent” (Davidse, Breban, and Van linden 2008: 486). In other words, old acquires a deictic function, locating the NP referent relative to the time of utterance, i.e. an element of the ground. Deictification is a subjectification process in the sense of Traugott (1982, 1989, 2003), in that relations originally describing situations in the extralinguistic world shift to relations located in the textual situation and to meanings involving the speaker’s perspective. The interpretation of a deictic secondary determiner is crucially dependent on its relation to the speaker-hearer exchange. In (50), for instance, “the old [former] record will refer to different sports milestones in a text of 1910 than in one of 1990. ... Any precise semantic gloss of postdeterminer meanings requires reference to the speech event and ultimately to the speech participants” (Davidse, Breban, and Van linden 2008: 497). As a consequence, deictification also entails subjectification in the Langackerian sense, “in that speaker and hearer are intrinsically part of the meaning of the determiner and postdeterminer unit, but do not – and cannot – receive explicit linguistic coding”, e.g. “corresponding to the old record in [50] one cannot get *the old to now record” (Davidse, Breban, and Van linden 2008: 497). Importantly, the notion of deixis as it is used in Davidse, Breban, and Van linden (2008) includes ‘discourse-deixis’ (Tine Breban, p.c.). For instance, the determiner complex such a in (51) contributes to the identification of the intended discourse referent by setting up an anaphoric relation with a discourse referent identified in the preceding discourse, i.e. men with broad shoulders and strong chests (§7.1). (51)

At the same time I had an avid attraction for men with broad shoulders and strong chests. One day, as I was watching such a man, he said, ‘You don’t even know me - why are you so attracted?’’. (CB) (Breban, Davidse, and Ghesquière 2011: 2694)

Intersubjectification and grammaticalization in the English NP

101

Although Davidse, Breban, and Van linden (2008) would classify uses as in (51) as instantiating discourse-deixis, with the ‘moment’ in the text functioning as deictic reference point (Tine Breban, p.c.), other linguists may argue that this kind of use is referential rather than deictic. Anaphora requires recovery of a piece of previous discourse, but reference is to an extralinguistic entity, not to the text itself as with discourse deixis. The difference is clarified in Cruse (2000: 323–324). In Listen to this, it will kill you! the demonstrative this points to future discourse elements – ‘it will kill you’ – and instantiates discourse deixis. In contrast, the anaphorically used pronoun he in John entered the room? He looked tired does not strictly refer to the word John but rather to the person named John. However, as Cruse (2000: 324) notes, the distinction between discourse deixis and anaphora is not always very clear-cut. In this study, deixis and by extension deictification will be understood broadly as including (the development of) modal, spatial, temporal, identifying, quantifying and discoursal or textual meanings. 3.4.2. Pathways to noun-intensification 3.4.2.1. Pathway 1: From description to noun-intensification In recent literature (Adamson 2000; Paradis 2000; Athanasiadou 2007; Traugott 2006, 2007a, 2007b), noun-intensifying readings of NP elements are generally envisaged to have their origin in fully lexical, descriptive uses. Paradis (2000) discusses the changes affecting the ten adjectives absolute, complete, perfect, total, utter, awful, dreadful, horrible, terrible, and extreme as a semantic and grammatical shift from descriptive content items to more schematic degree modifiers or ‘reinforcing adjectives’. Absolute, for instance, originally had only a propositional reading as in (52), but could later also be used as a marker of “degree, expressiveness and subjectivity”, as in (53) (Paradis 2000: 233–234). (52) (53)

Men sen it vtterly fre and absolut from alle necessitie. (Paradis 2000: 235) The honour of its absolute sufficiency (Paradis 2000: 236)

Four of the adjectives in Paradis’s set are also discussed in Athanasiadou (2007), who reconstructs the ‘intensification’ processes leading to the degree modifying adverbs absolutely, completely, perfectly and totally.

102 (Inter)subjectification and grammaticalization

The results from this study are highly compatible with those of Paradis (2000) and the author likewise proposes a shift from the content to the function domain for the adjectives absolute, complete, perfect and total, that is, from describing a property of the head, as in (54), to indicating a more schematic type of quantification or intensification, as in (55). (54) (55)

They advised a total ban on cigarette advertising. (Athanasiadou 2007: 559) He had total contempt for her. (Athanasiadou 2007: 559)

Paradis (2000, 2001, 2008) is the only one to have proposed a clear hypothesis about the mechanism of change involved in the shift from descriptive to noun-intensifying meaning. Conceptually, this shift can be explained in terms of a foregrounding mechanism: the boundedness configurations of the lexical uses are foregrounded at the expense of their lexical content. In other words, the gradability mode of the adjective in its source use is made prominent in the noun-intensifying use, while its lexical contents are backgrounded.29 With extreme adjectives such as terrible, dreadful, horrible and awful, their “negative superlative property ... is recruited for reinforcement” (Paradis 2000: 236). With complementaries such as complete, perfect, total and absolute “a generalised property of the degree of totality has taken over completely” (Paradis 2000: 236). In Paradis’s view, only extreme and bounded adjectives are recruited for noun-intensification (§1.3.1.2). She states that as intensification requires the abstract notions of either high degree or totality, unbounded adjectives such as fair “do not have much to offer in terms of intensification” (Paradis 2000: 253). However, although admittedly rather rare, degree modification can be expressed by unbounded and non-totality adjectives, as in (56) and (57). Pretty in a pretty mess indicates that the situation is “awkward, difficult, deplorable, unwelcome” (OED s.v. pretty A.3.c.). The degree modifier half in a half chance refers to the opportunity in question as “[a]ttaining only half-way to completeness or to the actual action, quality, or character in question; falling short of the full or perfect thing” (OED s.v. half 4). 29. This is remininscent of Peters’s (1994: 270) notion of scale transfer to account for the development of adverbial intensifiers. “Scale transfer is a special case of metaphorical meaning change: a concrete meaning develops into a more abstract one on the basis of a semantic component in the concrete meaning which facilitates at first the occasional use and later the transition of an adverb with full lexical meaning into the class of degree adverbs.”

Intersubjectification and grammaticalization in the English NP

(56) (57)

103

I haven't taken a contraceptive, if I get stuffed I'd be in a pretty mess, my husband's coming back in two weeks! (WB usbooks) Those workers must have been prisoners. Lifers looking for a half chance? (WB brbooks)

The shift from description to noun-intensification involves subjectification in the Traugottian sense. First, in Traugott’s (1982) model, such a shift implies a change from the propositional content domain to the expressive domain. From describing a particular state of affairs, the adjectives come to express the speaker’s evaluation of the NP referent, measuring the degree of a certain quality or quantity inherent to it on a scale. Second, the shift from description to noun-intensification also entails subjectification in the sense of Traugott (2003a). When an objective descriptive modifier such as total in a total ban gives rise to a noun-intensifying meaning as in total contempt, there is a semantic change from objective to subjective meaning. On the other hand, when a subjective descriptive modifier like pretty in a pretty dress develops a noun-intensifying meaning as in a pretty mess, there is shift from less to more (ideationally) subjective meaning (see Chapter 1). Paradis (2000: 238) explicitly describes the development of nounintensifying uses of adjectives as an instance of grammaticalization, as the originally “lexical items in constrained contexts undergo reanalysis and come to serve increasingly pragmatic functions”. Indeed, noun-intensifiers modify the degree of qualities conveyed in the type specification component of the NP. They thus require the presence of gradable qualities, conveyed by modifiers of the head and/or the head noun itself, and they serve a more grammatical function than fully lexical descriptive adjectives. Also, the adjectives no longer directly modify the head noun, but rather have scope over all the gradable elements in the entire NP to their right. The shift to noun-intensification goes hand in hand with the rise of a number of structural constraints. Unlike typical descriptive modifiers, nounintensifiers are mostly restricted to attributive position (a complete idiot, *the idiot is complete) and typically do not allow degree modification (*a very/more complete idiot), i.e. they display decategorialization (Hopper 1991) (§3.2). 3.4.2.2. Pathway 2: From identification to noun-intensification Bolinger (1972) posited a now largely neglected pathway to nounintensification originating in determining uses of linguistic elements. He

104 (Inter)subjectification and grammaticalization

argued that “[t]he shift by which a morpheme from the determiner system ... passes from identification to intensification is typical of a kind of wholesale migration in that direction” (Bolinger 1972: 61). This ‘migration’ is envisaged to encompass shifts both from referring to 'something of X identity' to measuring the degree of a quality as ‘of X magnitude’, i.e. from pure identification to intensification; as well as from quantifying sets/masses, i.e. quantitative identification, to intensification. Interestingly, the shift is not restricted to true determiners only. Bolinger (1972) also provides examples for adjectives that have determining uses. For complete, for instance, which Paradis (2000: 245) argues to have developed from descriptive to nounintensifying meaning, e.g. from a complete sentence to complete nonsense, Bolinger (1972: 59–60) hypothesizes a shift from determining to nounintensifying meaning, e.g. from the complete period to a complete fool. The development from descriptive meanings to noun-intensifying meanings is insightfully defined by Paradis (2000, 2001, 2008) as the foregrounding of the ‘grading’ or ‘measuring’ features of the source meaning in the new noun-intensifying meaning. Ghesquière and Davidse (2011) propose that this foregrounding mechanism also holds for quantifyingidentifying sources, given some reformulation. Starting-point are the two generally recognized basic types of quantification, viz. absolute and relative quantification (Milsark 1977; Langacker 1991; Davidse 1999). On the one hand, absolute quantifiers, which measure the size of some set or mass by offering a direct description of its magnitude, imply an open quantification scale without an upper boundary. Relative quantifiers, on the other hand, make a quantitative assessment by comparing the mass or set they actually designate to a reference mass/set. They hence imply a bounded, or closed, quantification scale. In this light, the shift from quantifying-identifying to noun-intensifying meaning can be hypothesized to involve a semantic reconfiguration from the measuring of instances and sets of instances, as in (58) and (60), to the measuring of the degree of a property, as in (59) and (61) (see §1.3.2.3, where the abstract similarities between absolute and relative quantification on the one hand and open and closed scales of intensification on the other hand were pointed out). (58) (59) (60) (61)

Too much of the time he misses the point. (Bolinger 1972: 58) He’s too much of a miser to give anything. (Bolinger 1972: 58) we are a minority of the whole population. (CB times) you get all this whole range of problems stuck in there. (CB ukspok)

Intersubjectification and grammaticalization in the English NP

105

On this pathway, no bleaching of lexically specific meaning or foregrounding of the measuring mode is necessary. The quantifying source meanings are schematic and scalar to begin with. Their semantics incorporate either an open scale with conventional measuring units, as in (58), or a closed scale with different values for calculating the difference between the actual size and the reference boundary, as in (60). All that is needed for the change to happen is a shift of domain, from the size of the designated entities to the modified degree of the properties. Paradis (2000) argues that in the shift to noun-intensifying uses the boundedness construal of the source meaning is preserved. Extending this reasoning to the quantifying meanings, relative quantifying meanings can in the first place be expected to shift towards noun-intensifying uses with a closed scale and absolutely quantifying sources towards open scale noun-intensifiers. This will be confirmed by the case study of whole (§5.3). Bolinger (1972) hypothesizes that not only quantifying-identifying but also purely identifying elements can develop noun-intensifying readings. The prime example discussed is such. In (62) such is used as a secondary determiner to refer to ‘something of X identity’; in (63) such functions as a noun-intensifier, measuring the degree of a quality as ‘of X magnitude’. (62) (63)

Such a person always frightens me. (Bolinger 1972: 60) Such a blunderer always frightens me. (Bolinger 1972: 60)

Unlike quantifying-identifying elements, purely identifying elements do not have an inherent scalar component. As a consequence, the shift toward noun-intensification cannot be described in terms of a foregrounding operation as with the gradability mode or measuring component of descriptive source senses and the quantification scale inherent in quantifying source meanings. How then can purely identifying meanings develop into scalar nounintensifying meanings? It can be hypothesized at this point that such developments will crucially involve pragmatic strengthening and the foregrounding of pragmatic meanings. If this hypothesis proves to be valid, the general mechanism of change explaining the shift to scalar noun-intensifying meaning proposed by Paradis (2000, 2001, 2008) will have to be reformulated and extended: it involves not only the foregrounding of inherent semantic components but also the derivation of scalar meaning on the basis of pragmatic processes. This hypothesis will be returned to in more detail in §6.3 with regard to the noun-intensifying uses of the adjective particular.

106 (Inter)subjectification and grammaticalization

In Traugott’s (1982) model of semantic change, the development of noun-intensifying meanings out of identifying meanings – both quantifying-identifying and purely identifying – quite straightforwardly constitutes a shift from the textual component of language to the expressive component (see §2.2.3; §3.1). The place of the shift in Traugott’s (2003a) cline of subjectification is somewhat more contested. As argued in Section 2.2.5, I consider determining elements – both primary and secondary – to convey textually intersubjective meaning as they steer the hearer’s interpretation of the discourse by contributing to the establishment of a joint focus of attention between speaker and hearer. Intensifying or degree modifying elements, in contrast, are subjective, conveying the speaker’s personal evaluation of and attitude toward the NP referent. When originally identifying elements give rise to noun-intensifying meanings, there is then a shift from (textually) intersubjective meaning to subjective meaning. If one subscribes to this view on (inter)subjectivity, this development challenges the unidirectional cline of intersubjectification leading from objective to subjective to intersubjective meaning proposed in for instance Traugott (2003a). The development from identifying to noun-intensifying meanings can, to my mind, not be viewed as an instance of grammaticalization. Both secondary determiners and noun-intensifiers have a grammatical function and one cannot truly be said to be more grammatical than the other. In terms of Lehmann’s (1995) parameters of grammaticalization, both secondary determiners and noun-intensifiers display paradigmaticity and restricted syntagmatic variability. They are part of rather small, restricted paradigms and cannot be shifted around freely within the NP. Secondary determiners, together with primary determiners, are situated leftmost in the NP structure; noun-intensifiers precede all the elements they have scope over. Also, unlike for instance lexical descriptive modifiers, secondary determiners and noun-intensifiers are restricted to attributive position (*the problems are other, *the idiot is complete). Unlike noun-intensifiers, secondary determiners also display increased bondedness, as they prefer a position adjacent to the primary determiner. 3.4.3. Adamson’s (2000) leftward and rightward movement hypothesis Adamson (2000) has put forward the structural-semantic hypothesis for changes in the NP that subjectification processes are accompanied by leftward movement in the NP, while de-subjectification is characterized by rightward movement. A productive example of the latter shift is that from

Intersubjectification and grammaticalization in the English NP

107

descriptive modifier to classifier, e.g. from a criminal act to a criminal court (Adamson 2000: 59–60). The former shift is illustrated in Adamson (2000) by means of a case study of lovely, which has shifted from descriptive modifier, as in a lovely house, to value adjective, as in lovely long legs (§1.3.2.1). Although Adamson (2000) focuses on the development of degree modifying elements, her leftward movement hypothesis is designed to also apply to the development of determining meanings. The author, for instance, refers to a number of studies on the development of English possessives and quantifiers from (adjectival) modifiers to determiners. According to Adamson (2000: 59), these adjectives, which have developed identifying functions, allow her “to posit a pathway characteriser > identifier in which … leftward movement, category shift and subjectivisation are combined”. The leftward and rightward movement hypothesis is of course a natural extension of the idea of the ordering in the NP as reflecting a subjectiveobjective continuum. As discussed in Sections 2.2 and 2.3, however, the subjective-to-objective model of the NP is flawed and the NP is better thought of as a field-like model in which subjective meanings are scattered prosodically throughout the entire structure. Similarly, although the posited structural-semantic hypothesis may reflect the dominant tendency for changes in the NP, it cannot be rigidly maintained. Reinforcing old, as in a grand old building, for instance, has moved towards a position to the right of the adjective being modified (§2.2.7). Also, as will be discussed in detail in Chapter 8, the emergence of subjective compounds such as old hand and little sod involves subjectification and concomitant movement to the right end of the NP, not the left end. Originating in modifier-head structures, the development of these lexicalizations entails that the adjectives old and little clearly travel right in NP-structure. 3.5. Conclusions In this chapter, a general discussion was presented of the processes of (inter)subjectification and grammaticalization as well as of their specific realizations in the English NP. First, the Traugottian notion of (inter)subjectification were introduced, with attention for how it developed and changed over time. Second, grammaticalization was discussed as typically involving a number of distinct changes in the syntax, semantics and morphology of the affected items. Whereas invited inferencing in specific discourse contexts was argued to lie at the basis of most grammaticalization

108 (Inter)subjectification and grammaticalization

processes, specific onset contexts for grammatical changes discussed include ambiguity, dialogicity and specific collocational contexts. Third, the development of secondary determining and noun-intensifying meanings was discussed as involving both subjectification and grammaticalization and as typically proceeding along distinct pathways. For secondary determining elements a pathway was posited leading from description to identification via a process of deictification (Davidse, Breban, and Van linden 2008). For noun-intensifying elements, two pathways observed in the literature were discussed, one originating in descriptive source meanings, the other in determining source meanings. The case studies in Part III of this book will be used to assess whether all three pathways distinguished can be substantiated by corpus data. As such, this study contributes to current theorizing on structural-semantic change in the NP. Can deictification account for the development of all secondary determining meanings in a unified manner? Are both pathways to noun-intensification distinguished in the literature supported by corpus evidence or is there rather one ultimate pathway of change? And what do the structural-semantic changes observed for the English NP entail for (inter)subjectification processes in general, and their presumed unidirectionality? Figure 13 represents the functional model of the English NP as it was introduced in Section 1.1.4 and how it corresponds to Traugott’s (1982), (1989), and (2003a) thinking on (inter)subjectivity/(inter)subjectification. First, the pathway from description to identification posited for secondary determining elements fits in nicely with Traugott’s (1982) cline leading from propositional to textual (to expressive). As already noted, changes do not have to proceed all the way down the cline, and may stop at any given moment. The shift from propositional, descriptive meaning to textual, determining meaning is also compatible with Traugott’s (2003a) intersubjectification cline, corresponding to a shift from non- or less subjective meaning to (more) subjective meaning. Second, pathway 1 to noun-intensification, originating in descriptive source meanings is equally compatible with the clines of semantic change proposed in Traugott (1982) and Traugott (2003a). Along this pathway, elements shift from the propositional directly to the expressive component as they shift from expressing non- or less subjective meaning to expressing highly subjective meaning. Pathway 2 to noun-intensification, on the other hand, which originates in identifying source meanings, is more challenging. Involving a move from the textual to the expressive component, the shift is unproblematic for the cline proposed in Traugott (1982). However, it is harder to see how the shift can be captured by Traugott’s (2003a) cline of intersubjectification.

Figure 13. Model of the prenominal string of the English NP

subjective

subjective

subjective

externally propositional non-/less objective

propositional

Traugott (2003a)

expressive

descriptive modification

internally propositional

textual

degree modification

modification

Traugott (1989)

Traugott (1982)

determination

categorization

Conclusions 109

110 (Inter)subjectification and grammaticalization

In Traugott’s view, both determining and degree modifying meanings are subjective in nature. If elements shift from identification to nounintensification does this then involve a shift from less to more subjective meaning? In my opinion, it does not. As argued in Sections 1.4.2 and 2.2.5, I take secondary determining elements to typically convey textually intersubjective meanings, whereas degree modifying elements convey subjective meaning. This is, however, problematic for any unidirectional intersubjectification cline, as pathway 2 then constitutes a shift from (textually) intersubjective to subjective meaning. At this point, only Traugott’s (1982) cline of semantic change seems to be able to accommodate all the pathways of change posited in the literature for the development of identifying and noun-intensifying meanings. In contrast to the hypothesis put forward by Adamson (2000), this semantic cline is not reflected in increasing leftward movement in NP structure. Rather, the proposed developmental path is best understood as proceeding through the functional-structural zones of the NP in the order description > determination > degree modification, mirroring the propositional > textual > expressive cline.

Part III The case studies

Chapter 4 Data and methods To study historical changes in language use researchers are necessarily dependent on written records. For English, a considerable number of wellbalanced, extensive digitized text-corpora are available, which has allowed careful selection of corpus data in terms of the research questions. This chapter details the synchronic and diachronic corpora consulted for this study. The selection of linguistic items studied is motivated in terms of their aptness to realize the aims of this study. Finally, the analytical heuristics used for the qualitative study of the corpus data are discussed. 4.1. Selection of the data This study has been set up to arrive at a better understanding of (inter)subjectification processes in the prenominal string of the English NP. In particular, the focus of attention is the directionality of (inter)subjectification processes. Two important issues have been raised in the preceding chapters concerning this general question. First, for the development of noun-intensifying meanings two distinct pathways of change have been posited, one starting from identifying source meanings (Bolinger 1972) and one originating in descriptive, lexical source meanings (e.g. Adamson 2000; Paradis 2000; Traugott 2006, 2007a, 2007b) (§3.4.2.2 and §3.4.2.1). For the latter pathway, Adamson (2000) has posited that the subjectification process is accompanied by a positional shift to the left of NP structure. Conversely, de-subjectification processes are hypothesized to display concomitant rightward movement in NP structure. Second, in the literature on processes of semantic change in the NP, the notions of subjectification and intersubjectification are generally taken as given. However, different definitions are available both of subjectification and intersubjectification. The latter, for instance, is viewed as encompassing textual meanings in Breban (2010a) and Ghesquière (2010), but not in Traugott (2010a) (§2.2.5). In the light of these controversies, the focus of this study is on elements that have developed both textual (identifying) uses and straightforwardly subjective (noun-intensifying uses). Data study will then allow us to uncover possible diachronic, genetic links between the two uses and

114 Data and methods

to verify the validity of the proposed pathways toward noun-intensification. In addition, Adamson’s (2000) subjectification-cum-leftward movement and de-subjectification-cum-rightward movement hypotheses will feature prominently in Chapter 8 of this study. I will briefly return to this at the end of this section. The range of lexical elements that can develop both identifying and noun-intensifying uses is rather limited. On the one hand, for prenominal modifiers to be able to develop (secondary) determining uses they have to display either identifying or quantifying potential. In other words, “they have to express a relation which can come to be used to locate the nominal referent in the universe of the discourse” (Davidse, Breban, and Van linden 2008: 496). On the other hand, not every linguistic item can develop an intensifying meaning. For noun-intensifiers, the literature suggests that at least two syntactic source domains have to be recognized, each encompassing various semantic fields: (i) fully lexical descriptive adjectives (Paradis 2000, 2001; §3.4.2.1), and (ii) more abstract identifying elements of the NP (Bolinger 1972; §3.4.2.2). For this study, a selection of originally lexical, descriptive items and identifying items was made. A first set of elements consists of three adjectives from the semantic field of completeness, viz. complete, total and whole (Chapter 5). The quantitative notions they express makes them prone to develop both quantifying-identifying and quantifying-intensifying uses. A second adjectival set is made up of particular and specific, both from the semantic field of specificity (Chapter 6). Third, as in Bolinger (1972), the prenominal uses of such were investigated. The results of this study were compared to the results of two smaller case studies of the Dutch counterpart of English such, i.e. zulk, and of the English interrogative determiner and noun-intensifier what (Chapter 7). Syntactically such, zulk and what behave rather differently from the other items under scrutiny, but they display the same sort of functional variation between identifying and intensifying uses. Finally, the adjectives old and little were studied (Chapter 8). Neither adjective has developed a noun-intensifying use and only old has an identifying reading. The two adjectives do, however, engage in lexicalization processes resulting in both objective and subjective meaning and can thus be used to evaluate Adamson’s (2000) hypothesis on (de-)subjectification and concomitant positional change in NP structure. Two things have to be noted concerning the selected items. First, for the studies on noun-intensification I have chosen to study sets of semantically closely related adjectives, viz. the completeness adjectives complete, total and whole, and the specificity adjectives particular and specific. This will

Selection of the data

115

allow me to check whether similar source meanings give rise to similar target meanings and if their development proceeds along the same pathway. As Lehrer (1985: 283) notes, “insights into the principles of semantic change can be found by looking, not at the whole lexicon, but at words which belong to a single semantic field. A semantic field is a set of lexemes which cover a certain conceptual domain and which bear certain specifiable relations to one another”. Importantly, the author hypothesizes that “semantically related words are more likely to undergo parallel semantic changes than semantically unrelated ones precisely because of their semantic relationships” (Lehrer 1985: 286) [italics original]. Differences in the development of adjectives from the same semantic field will then have to be explained in terms of idiosyncratic syntactic, collocational or more external characteristics. Second, unlike more prototypical descriptive adjectives such as tall or nice, the descriptive source meanings of the selected adjectives show relatively abstract lexical meanings from the start. The descriptive meaning of the completeness adjectives complete, total and whole is quantifying in nature and that of particular and specific is specificational in nature, describing the detail or specificity of the NP referent. In contrast to what is assumed in current theories of (inter)subjectification, the source meaning of the adjectival noun-intensifying uses is then not purely lexical, but already has some more abstract meaning. In short, the adjectives selected allow for comparisons between adjectives that developed from (roughly) the same semantic source domain (e.g. completeness) and between adjectives that developed from different semantic source domains (e.g. completeness vs. specificity). As a result, the variations in the pathways of semantic change that are discovered are all the more significant, pertaining even to items from the same source domains, while generalizations are stronger to the extent that they apply to items from different source domains. 4.2. Corpora To trace the diachronic development of the linguistic items studied a number of corpora were consulted that span the entire period of written records, from Old English to Present Day English. This time span was split up in the customary sub-periods Old English, Middle English, Modern English, and Present Day English.

116 Data and methods

4.2.1. Old English For the Old English period two partially overlapping corpora were used: the Helsinki corpus of English Texts: Diachronic and Dialectal (HC) and the York-Toronto-Helsinki Parsed Corpus of Old English Prose (YCOE). The latter is an extensive corpus containing larger samples of the prose texts available in the Helsinki corpus. Both corpora are divided into four subperiods as shown in Table 3. Table 3. The Helsinki Corpus and the York-Toronto-Helsinki Parsed Corpus of Old English Prose Subperiod 750–850 850–950 950–1050 1050–1150 Total

HC 2190 92,050 251,630 67,380 413,250

YCOE 1,753 343,517 764,960 340,146 1,450,376

4.2.2. Middle English The Middle English data were extracted from the relevant sections of the Helsinki Corpus and the second edition of the Penn-Helsinki Parsed Corpus of Middle English (PPCME2) (Kroch and Taylor 2000) (see also http://www.ling.upenn.edu/hist-corpora/PPCME2-RELEASE-2/). The subperiods and corpus sizes are represented in Table 4. Table 4. The Helsinki Corpus and the Penn-Helsinki Parsed Corpus of Middle English Subperiod 1150–1250 1250–1350 1350–1420 1420–1500 Total

HC 113,010 97,480 184,230 213,850 608,570

PPCME 258,090 93,999 403,007 400,869 1,155,965

Corpora

117

4.2.3. Modern English The Early Modern English data were retrieved from the Penn-Helsinki Parsed Corpus of Early Modern English (PPCEME) (Kroch, Santorini and Delfs 2004) (see also http://www.ling.upenn.edu/hist-corpora/PPCEMERELEASE-1/). It consists of the Early Modern English prose texts of the Helsinki corpus, and two parallel supplements (Penn 1 and Penn 2). The three subperiods distinguished within the corpus and their word counts per subcorpus are presented in Table 5. Table 5. The Penn-Helsinki Parsed Corpus of Early Modern English Subperiod 1500–1570 1570–1640 1640–1710 Total

HC 196,754 196,742 179,477 572,973

Penn 1 194,018 223,064 197,908 614,990

Penn 2 185,423 232,993 187,631 606,047

PPCEME Total 576,195 652,799 565,016 1,794,010

Because of the limited number of occurrences of both the specificity and the completeness adjectives in the PPCEME corpus, the Corpus of Early Modern English Texts (CEMET) (De Smet 2008: 14–16) was also consulted. This corpus is compiled on the basis of electronic texts available online through archiving projects such as the Project Gutenberg and the Oxford Text Archive and is divided into two subperiods, 1570–1640 and 1640– 1710, containing 1,094,349 and 1,943,392 words respectively. For the Late Modern English period, I used the extended version of the Corpus of Late Modern English texts (CLMETEV) (De Smet 2005, 2008: 17–19, 21–29). Similar to the CEMET, this corpus is compiled on the basis of texts freely available online through the Project Gutenberg and the Oxford Text Archive. In addition, texts were also drawn from the Victorian Women Writer’s Project. Table 6 details the number of words of the three subperiods distinguished in the corpus. Table 6. The Corpus of Late Modern English Texts Subperiod 1710–1780 1780–1850 1850–1920 Total

CLMETEV 3,037,607 5,723,988 6,251,564 14,970,622

118 Data and methods

4.2.4. Present Day English The Present Day English data were taken from the COBUILD corpus the Bank of English, of which 56 million words (from the period 1986–96) were made available on Wordbanks Online via remote log-in (see Clear et al. 1996). The COBUILD corpus (CB) is divided into twelve subcorpora on the basis of geographical distribution (Australian, American and British English), register (formal and informal), and language mode (spoken and written). The build-up of the entire corpus is summarized in Table 7. Table 7. The COBUILD corpus Abbreviation oznews ukephem ukmags ukspok usephem bbc npr ukbooks usbooks times today sunnow

Full title Australian newspapers UK ephemera (leaflets, adverts, etc.) UK magazines UK spoken US ephemera (leaflets, adverts, etc.) BBC World Service Radio broadcasts National Public Radio broadcasts UK books (fiction and non-fiction) US books (fiction and non-fiction) Times newspaper Today newspaper UK Sun newspaper

Number of words 5,337,528 3,124,354 4,901,990 9,272,579 1,224,710 2,609,869 3,129,222 5,354,262 5,626,436 5,763,761 5,248,302 5,824,476

For this study, two subcorpora were systematically queried, the CB times and the CB ukspok subcorpora. The CB times section comprises newspaper articles from the Times and Sunday Times, two quality newspapers published in London. The CB ukspok section contains transcribed informal conversations between speakers from various parts of Britain. Comparison of data from the two corpora may reveal register differences between the different uses of the items studied as they represent opposed modes – written versus spoken – and different interpersonal tenors – formal versus informal (Martin 1992). In addition to the CB times and CB ukspok data, examples were also taken from dictionaries, the Internet, the Corpus of Contemporary American English (COCA, http://corpus.byu.edu/coca/), the other subcorpora of the COBUILD corpus, and the new WordbanksOnline corpus. All examples followed by (WB) have been extracted from the WordbanksOnline corpus, the successor to the COBUILD corpus. Examples extracted earlier

Corpora

119

from the COBUILD corpus are followed by (CB). With its 553 million tokens covering the period 1972–2004, the WordbanksOnline corpus is much larger than the 56 million word COBUILD corpus. The British subsections, for example, contain 259,479,077 tokens in WB versus 42,099,593 in CB. Unfortunately, with the release of the WB corpus, the CB data are no longer accessible. 4.3. Methodology For each of the items studied either exhaustive extractions were made or representative samples of some 200 prenominal examples for each 70- to 100-year period, depending on the specific make-up of the corpora consulted. The extracting and sampling was done with the aid of the Abundantia Verborum software (Speelman 1997), which was also used for the analysis of the individual attestations. All of the prenominal examples in the selected data sets were analyzed individually in terms of syntactic (e.g. definite or indefinite determiner) and collocational co-occurrence patterns and textual features (e.g. anaphoric or cataphoric). Functionally, all prenominal instances were labelled as identifying, noun-intensifying, adjective-intensifying, or descriptively modifying – either objectively or subjectively. A number of examples were given two labels as they were vague or ambiguous between two readings. The qualitative analysis of the individual attestations required careful application of analytical heuristics. The prenominal uses of adjectives are a good example of the point that corpus data per se reveal little. Usually, adjectives occur prenominally in all of their uses. NPs with multiple modifiers, for which position in the prenominal string may be revealing with respect to the function of the adjectival modifiers, are rare. As a consequence, the analysis in terms of the different functions of the adjectives looked at required careful application of recognition tests. More specifically, syntagmatic and paradigmatic recognition criteria for the different functions were systematically applied in the analysis of the corpus data. The semantic and contextual analysis was always assessed against formal criteria and vice versa. First, syntagmatically, the functional analysis took into account word order possibilities, relating NPs with single adjectives to NPs with multiple adjective strings. We can illustrate this with a straightforward example such as (64). On the basis of context, old in the NP an old bike is plausibly analyzed as an objective descriptive modifier indicating that the bike referred

120 Data and methods

to “has been relatively long in existence or use” (OED s.v. old I.3.a). Multiple adjective NPs with old as in (65) and (66) support such an analysis. In (65) old is preceded by another objective descriptive modifier – squeaky, and in (66), as in most other examples found in the WB corpus, old is followed by an objective classifying modifier. (64) (65) (66)

My friend replied that she was too old, to which her granddaughter replied. "Don't worry, we'll get you an old bike. (WB sunnow) This one, a woman, I can see her now in a old duckety-mud coat, used to pedal her squeaky old bike up the coast once a week from Rough Shop Harbor to Killick-Claw (WB usbooks) Goggle-eyed from repeat screenings of the Easy Rider movie, young daredevils were sticking apehangers on their old British bikes. (WB oznews)

Second, the paradigmatic tests used are alternation or substitution tests, which relate attestations of alternates to each other. For instance, descriptive modifiers are characterized by their ability to be modified in terms of degree and to be used in both attributive, prenominal position and predicative position. This is illustrated in (67) to (69) for old with the head noun bike. Importantly, I want to draw attention to the fact that the heuristic tests were not applied intuitively but checked in the corpus data. (67) (68) (69)

the veteran rider admitted two weeks ago at the British Grand Prix he would rather race the more reliable, but older bike (WB brregnews) Now, is it a good idea to buy a very old bike even if it is not used much? (http://www.bikeszone.com/forum/viewtopic.php?f=22&t=5495) The bike is old, and it's showing it's age. (http://uk.answers.yahoo.com/question/index?qid=2011093007175 3AA9jfTW, accessed 3 October 2011)

Despite the conscientious application of the recognition criteria, the functional analysis nevertheless proved challenging at times, not only for the adjectives studied but also for such. The identifying and nounintensifying uses of such cannot be easily distinguished in terms of overt syntactic behaviour. Both uses are for instance found in the same positions in the NP. To determine the exact function of such detailed contextual analysis is required, as for many of the attestations of the adjectives. The

Methodology

121

main contextual clues for the two readings are the presence of a phoric relation to another referent in the surrounding discourse for identifying uses and of gradable properties expressed by the descriptive part of the NP for intensifying readings. Further contextual clues for an intensifying reading that had to be taken into account are the expression of personal attitude or emotions by the speaker and the presence of other intensifying elements (Chapter 7). By adopting the methodological approach to the data discussed above, this study aims to offer detailed, well-balanced descriptive analyses of the synchronic and diachronic behaviour of the items under scrutiny, based on syntagmatic and paradigmatic criteria that are formally verifiable in the data. In addition, the qualitative analysis will also be quantified and (gradual) increases or decreases in the relative proportions of particular adjectival uses will be used as an important criterion in defining possible pathways of change. As shown in, amongst others, Ghesquière 2011, Ghesquière and Van de Velde 2011, shifts in the proportional distribution of different uses of the same item may be indicative of grammatical and/or semantic change. The planned research thus has a strongly descriptive, data-oriented focus, while at the same time aiming to verify existing theoretical hypotheses, formulate theoretical generalizations and contribute to theory-formation.

Chapter 5 The completeness adjectives This chapter will discuss the structural-semantic development of three adjectives from the semantic field of completeness, i.e. complete, total and whole. These adjectives have very similar source meanings and are consequently often listed as synonyms in dictionary entries (e.g. Oxford English Dictionary [OED]; Middle English Dictionary [MED]; Cambridge Advanced Learner's Dictionary). What the case study will show, however, is that despite the close semantic relation between their descriptive readings, the adjectives have developed noun-intensifying meanings along rather different pathways of change, influenced by their idiosyncratic collocational preferences. The noun-intensifying uses of the adjectives of completeness have so far not received a great deal of attention in the literature. Corresponding to the general association of intensification with adverbs, the intensifying adverbs completely, totally and wholly have received more scholarly attention than their adjectival counterparts (e.g. Bäcklund 1973; Lorenz 2002; Partington 2004; Aijmer 2007). Sinclair (1990: 69), Paradis (2000) and Athanasiadou (2007) do focus on adjectival intensifiers but include only complete and total in their discussion, not whole. The latter has, to my knowledge, only been treated as an intensifier in Bolinger (1972: 24, 54) and Peters (1994: 283, 285), albeit under the label of degree adverb rather than adjective. The syntactic status of whole will be returned to in Section 5.3.4 in relation to adjective-intensifying uses as in He's starting a whole new career (Bolinger 1972: 54). Paradis (2000) and Athanasiadou (2007) view the noun-intensifying uses of complete and total as resulting from a shift from property attribution to noun-intensification. Complete is claimed to have developed from the propositional reading in which it describes an NP referent as ‘having all its parts’ (Paradis 2000: 235; Athanasiadou 2007: 558), as in (70), to a ‘reinforcing’ (Paradis 2000) or ‘quantifying’ (Athanasiadou 2007: 559) intensifying meaning, as in (71). The noun-intensifying meaning of total, as in (73), is said to develop from the descriptive reading ‘relating to the whole of something’, as in (72) (Paradis 2000: 235; Athanasiadou 2007: 559). In this chapter the validity of these proposed pathways will be tested against corpus data.

124

(70) (71) (72) (73)

The completeness adjectives

It is a complete sentence (Paradis 2000: 249) It is complete nonsense (Paradis 2000: 249) They advised a total ban on cigarette advertising. (Athanasiadou 2007: 559) He had total contempt for her. (Athanasiadou 2007: 559)

Crucial to a good understanding of the development of the completeness adjectives is the boundedness construal of their different prenominal uses (§1.3.1.2). Paradis (2000: 236, 2008: 335, 337) classifies complete and total as bounded adjectives in all their uses and we can safely assume that the semantically similar adjective whole would most likely receive the same classification. All three completeness adjectives are then analyzed as complementaries: entities are always construed as being either complete or incomplete, total or not total, whole or not whole. The bounded or unbounded construal of adjectives is claimed to typically remain unaltered, both synchronically, across its diverse functions, and diachronically. In the historical shift from descriptive to noun-intensifying uses, the boundedness construal of the adjectives is simply foregrounded, while their lexical content is pushed to the background (Paradis 2000: 235–239; §3.4.2.1). In both uses complete, total and whole are thus construed as bounded, but in “their capacity as reinforcing adjectives, a generalised property of the degree of totality has taken over completely” (Paradis 2000: 236). In the following sections, I will show that although the completeness adjectives all have bounded meanings, they also display uses in which they can be construed as unbounded, denoting ranges on a scale. Moreover, the case study on complete in particular will illustrate that the boundedness construal of an adjective cannot always be clearly determined, i.e. a certain degree of vagueness in terms of boundedness will have to be allowed for. This chapter will revisit the paths to the noun-intensifying uses of complete, total and whole posited in the literature on the basis of an extensive diachronic and synchronic corpus study. Because the developmental trajectories followed by the three completeness adjectives are so distinct, they will be discussed separately. As such, the idiosyncrasies of the three items under scrutiny can be given due attention. Note that complete, total and whole have different genealogies to start with. Whereas whole is an adjective of Anglo-Saxon origin with attested uses from Old English onwards, complete and total are Romance loanwords that appear in the data only in the Early Modern English period.

Complete

125

5.1. Complete Although the corpus data largely support the general shift posited by Paradis (2000) and Athanasiadou (2007) from fully lexical quality ascribing uses to noun-intensification, a number of refinements will be proposed. It will be argued that complete has at least two distinct noun-intensifying uses which have developed from different descriptive source meanings. Moreover, the importance of the collocational patterning of complete in its different uses for the pathway of change along which it proceeds will be emphasized. 5.1.1. Descriptive modifier uses of complete 5.1.1.1. Bounded descriptive modifier uses of complete The prototypical descriptive use of complete is one in which the adjective describes the NP referent as having all its parts. This use is attested from the period 1570–1640 onwards and is the most frequent objective descriptive meaning of complete in Present Day English. In (74), for instance, the author describes his writings as not yet including all the necessary parts (OED s.v. complete 1.a). Similarly, in (75) complete indicates that the NP referent – the fossil skeleton – includes (most of) the parts that are normally associated with it. As such, the NP referents in (74) and (75) are described in relation to a definite boundary of ‘completeness’ or ‘totality’, which in these particular cases is not reached. Both Paradis (2000) and Athanasiadou (2007) have identified the descriptive reading ‘having all its parts’ as the origin of the later intensifying meanings of the adjective. However, as will be argued in Section 5.1.3, this descriptive use, although well-attested in all historical periods, does not seem to have played a key role in the development of the noun-intensifying uses. (74)

(75)

It must be likewise conceived, that in these points which I mention and set down, they are far from complete tractates of them, but only as small pieces for patterns. (CEMET Bacon, 1605, The advancement of learning) The fossil skeleton, the most complete example of its kind, was in danger of being washed out to sea by coastal erosion and winter storms (CB times)

126

The completeness adjectives

The earliest prenominal attestations of complete in the corpora date from the Early Modern English period. According to the OED and MED, predicative and postnominal uses of complete, however, date back to the fourteenth century. This is not surprising considering that complete is a Romance loanword and is used in French in postnominal position only. The adjective’s verbal origin has most likely also been an influencing factor in the late rise of prenominal uses. Although this study focuses on the prenominal uses of the adjectives under scrutiny, it is important to take the postnominal uses of complete into consideration as well as they seem to have played a role in the later rise of the noun-intensifying meanings. When used postnominally complete most often modifies nouns referring to periods of time, such as year and day, indicating that they have run their full course (OED s.v. complete 2). The verbal origin of complete is still clearly present in these uses, easily allowing the adjective to be replaced by the past participle completed. Complete in (76) and (77), for example, indicates that the years have come to completion or are completed. (76) (77)

Þis blessud virgyn ...Twolfe 3ere complete in hurre tombe lay. (MED a1450 St.Editha) ‘this blessed virgin lay in her tomb for twelve complete years’ according to this calculation, the "AEneas" takes not up above a year complete, and may be comprehended in less compass. (CEMET Dryden, 1688, Life of Saint Francis Xavier)

Postnominal uses like the above are obsolescent in the Early Modern English period, but complete continues to express the perfective ‘completed’ meaning in prenominal position. In later periods this use is extended to nouns (metonymically) referring to events or actions rather than indications of time. In (78), the NP the complete manufacture invokes a manufacturing process that is described as fully concluded or completed, producing a finished article ready for sale. (78)

It costs more labour, and therefore more money, to bring first the materials, and afterwards the complete manufacture to market. (CLMETEV Smith, 1766, An inquiry into the nature and causes of the wealth of nations)

A bounded descriptive use of complete which is very similar to the perfective use just discussed, characterizes an action, state or quality not simply as completed but as being ‘realized in its full extent’ (OED s.v. complete

Complete

127

3). Focus is now on the end-point of a process of completion rather than on the process as a whole, which explains the collocational preference of complete for nominalizations of verbs such as isolation and circumnavigation in (79) and (80). These nouns, as they imply a process, are still associated with a temporal notion and are thus a natural extension of the set of nouns with which the perfective sense originally co-occurs. In a way the ‘fully realized’ use can be considered slightly more emphatic than the perfective use, as it focuses on the thoroughness of completion rather than on the property of completeness itself. Whereas in the perfective reading complete can be paraphrased as ‘completed’, it is now semantically very close to the adverb completely. In (79), the NP the complete isolation characterizes the Jew as ‘completely isolated’ from the rest of the world. Similarly, the NP in (80) describes Cook's attempt to ‘completely’ sail around the island. (79)

(80)

the horrible cruelties of the Romans, which, during and after the war, might give some cause for the complete isolation of the Jew from the rest of the world. (CLMETEV Gibbon, 1776, The decline and fall of the Roman Empire) He explored the whole of the eastern side of this large island, which is three hundred miles in length ... Other attempts to get inside the reefs were, however, unsuccessful, and after several narrow escapes from shipwreck Cook gave up, to his regret, a complete circumnavigation of the island. (CLMETEV Cook, 1768–71, Captain Cook's journal during the first voyage round the world)

The descriptive modifier status of the perfective and ‘fully realized’ uses of complete is clear from the fact that they allow degree modification and predicative alternation. This is illustrated in (81) and (82) for the ‘fully realized’ sense. (81) (82)

After four decades of almost complete isolation, Albania faces untold difficulties in its aim to integrate with Europe. (WB brbooks) Mrs. Furze was perfectly aware that she was not deluding her daughter; but she assumed that the delusion was complete. (CLMETEV Rutherford, 1893, Catherine Furze)

128

The completeness adjectives

5.1.1.2. Descriptive modifier uses of complete, vague between bounded and unbounded construal Besides the bounded ‘having all its parts’, ‘completed’ and ‘fully realized’ senses, complete also has more subjective, evaluative descriptive uses which hover between a bounded and an unbounded construal. From the period 1570–1640 on, complete is found in the data characterizing NP referents as “perfect in nature or quality” (OED s.v. complete 4), as in (83) and (84). In the latter example, the (positive) evaluative reading of complete is triggered by the preceding coordinated adjective perfect which in this context is largely synonymous with complete. In fact, the property of perfection is attributed to the book in question despite the fact that it is actually lacking a chapter, i.e. despite its objectively measurable incompleteness. (83) (84)

Men shulden bi hooli lif of Crist trowe that his lawe is compleet. (OED Wyclif, 1380) the book is more perfect and complete by wanting the chapter, than having it (CLMETEV Sterne, 1759, The life and opinions of Tristram Shandy)

The ‘perfect’ sense of complete later gave rise to the more specialized reading ‘fully equipped or endowed’, ‘consummate’, which is particularly frequent in collocations with nouns denoting professions, arts and pursuits (OED s.v. complete 5). Examples (84) and (86) illustrate that complete functions as a gradable, descriptive modifier both in its ‘perfect’ and ‘consummate’ uses as it occurs in (predicative) comparative and superlative constructions respectively. (85) (86)

The Compleat Angler (CEMET Walton, 1653, The complete angler) if this model chances to wander abroad, I recommend it to the Roman speculativi (the most complete gentlemen of this age) for their censure (CEMET Harrington, 1656, Oceana)

The subjective descriptive uses of complete, like the more objective ‘completed’ and ‘fully realized’ senses, convey an inherent notion of quantity, but unlike the latter they are not associated with an objectively definable point of reference to determine if the NP referent really is ‘perfect’ or ‘consummate’. Whether or not this is the case is more a matter of opinion

Complete

129

and thus subjectively coloured. This distinction was already noted by Sapir (1930: 26–27) who states that whereas with the objective senses focus is on the quantitative information, with the subjective uses “the notion of quantity merely comes in as a determinant. ... In ‘the complete angler’ the emphasis is not on the enumeration of qualities required for good fishing but rather on the successful pursuit of the sport by virtue of all and sundry qualities required”. Paradis (1997: 57, 2008: 324) argues that the distinction between what we have called more objective and more subjective descriptive adjectives is reflected in the way speakers and hearers treat them in language use. Whereas with objective adjectives language users typically agree both on their meaning and their application, with subjective or “evaluativeattributive” adjectives “people are likely to agree on their meanings but may very well disagree on their application” (Paradis 2008: 324). In other words, “the speaker determines how they should be applied” (Paradis 2001: 57). These uses are thus strongly speaker-based, reflecting his or her personal attitude toward the NP referent, and are hence subjective in nature. Paradis (2001: 57, 2008: 324) associates evaluative-attributive meanings with unbounded and extreme adjectives, but not with bounded adjectives. However, in its ‘perfect’ and ‘consummate’ readings as in (84) and (86), complete conveys just such an evaluative meaning. I may find a book chapter complete and without defect, while someone else might judge that it is lacking important information or is not attractively formatted. Similarly, one speaker may well consider someone a complete and perfect gentleman, while another may find him rude. Building on this observation, I argue, following Ghesquière and Davidse (2011), that in its ‘perfect’ and ‘consummate’ sense complete does not function as a bounded adjective but as an extreme adjective. Paradis (2000: 236) states that in both its descriptive and noun-intensifying readings complete is a bounded adjective, “associated with a bounded, ‘either-or’ conceptualisation”. However, in examples (83) to (86), complete is vague between a bounded and an unbounded construal and is consequently better analyzed as an extreme adjective. Extreme adjectives can generally be considered a kind of in-between category, displaying properties of both bounded and unbounded adjectives (§1.3.1.2). On the one hand, they are associated with a scale, but on the other hand, they typically denote not ranges on but the ultimate point of a scale. This is reflected in their semantico-syntactic nature in at least two ways. First, as mentioned above, although speakers may agree on the semantics of the adjective they may disagree on its application in actual language use. Second, whereas bounded adjectives can typically only be modified by propor-

130

The completeness adjectives

tional degree modifiers such as almost and unbounded adjectives typically only allow scalar modifiers such as very, extreme adjectives occur with both types of degree modifier depending on whether they are construed as bounded or unbounded in a particular context. In both the ‘perfect’ and ‘consummate’ sense, complete hovers between a bounded and an unbounded reading, consequently allowing grading by both almost and very, as illustrated in the following examples taken from the Internet. In both (87) and (88) complete collocates with the head noun player, but, as shown by the accompanying degree modifiers, it is construed as bounded in the former example and as unbounded in the latter example. In (87), complete is modified by almost, which “relate[s] to a definite and precise value of the property expressed by the adjective and [is] bounded” (Paradis 2001: 50). In (88), in contrast, complete is intensified by very, which “indicate[s] a range on a scale of the gradable property expressed by the adjectiv[e]”, and is “in that respect unbounded” (Paradis 2001: 50). (87) (88)

he is so strong, quick, accurate and an almost complete player. (http://www.onevalefan.co.uk/forums/showthread.php?t=31248) Del Potro is a very complete player. He has an unbelievable serve right now and from the baseline, he doesn’t make any mistakes. (http://www.timesonline.co.uk/tol/sport/tennis/article6833045.ece)

5.1.2. Identifying uses of complete From the period 1570–1640 onwards, complete occurs in the corpus data as a secondary determiner adjective (cf. Halliday 1985: 162). In its determining function, complete aids the identification of the NP referent by providing quantitative information: it indicates that reference is made to the entire entity referred to by the head noun, as in (89) and (90). (89) (90)

To care for your Retractable Brushes, follow washing instructions (see Before you Begin) taking care not to immerse the complete unit in the water. (CB ukephem) However, the Unemployment Unit says that this total benefit cost accounts for only about half of the complete cost of unemployment. (CB times)

In the data, secondary determiner complete is typically found in NPs with the definite determiner the and a singular count noun, as in the above

Complete

131

examples. The quantifying-identifying function of complete is then twofold: (i) it makes explicit the implicature of inclusive reference associated with the definite article, and (ii) it stresses the boundedness construal of the count noun. First, the definite article – the only determiner with which secondary determiner complete occurs in the data set – is generally characterized as invoking ‘inclusive reference’, i.e. reference to the entire potential instantiation of the type designated by the NP (Hawkins 1978; Declerck 1991: 292; Langacker 1991: 102). A noun phrase headed by the definite article is traditionally understood as referring to all the members of a contextually salient reference set (with count nouns) or mass (with mass nouns). With singular count nouns only one instantiation is available. Noun phrases with indefinite reference, in contrast, simply refer to one or some, and not necessarily all, potential instantiations of the type referred to by the head noun. When the definite article and quantifying-identifying complete are combined a certain emphatic effect arises, as the adjective “merely makes explicit an element that is normally part of definite identification” (Davidse, Breban, and Van linden 2008: 491). Definite NPs are already associated with “an inherent notion of the ‘whole’ thing and complete, like a universal quantifier such as all, quantifies over this whole thing” (Davidse, Breban, and Van linden 2008: 489). Both the and complete in a way convey universal relative quantification. Secondary determiner complete is then to a certain extent semantically redundant and therefore mostly used in contexts where there is a need to emphasize reference to the whole or to all of the entity. In example (89), complete is added to the NP to remove any doubt as to which parts of the unit should be immersed in the water. (89) is an example of a contrastive context, in which the use of complete helps to dispel a notion of counter-expectation, viz. that it is OK to immerse the brushes completely in the water (cf. Davidse, Breban, and Van linden 2008: 490). Second, besides stressing the quantification inherent in the definite determiner, complete can in these NPs also be understood as stressing the boundedness of the count nouns it co-occurs with (Hendrik De Smet, p.c.). In this study, boundedness has been discussed in relation to adjectives, but the notion was originally developed with respect to nouns and verbs. Langacker (1987b: 58) states that count nouns designate bounded regions in one or more domains such as space, time or visual field. The designated regions are bounded in that “there is a limit to the set of participating entities (i.e. it does not extend indefinitely)” (Langacker 1987b: 62). When secondary determiner complete combines with a count noun it stresses the boundedness of the head.

132

The completeness adjectives

As secondary determiner complete in definite NPs with singular count noun mainly has a reiterative, emphasizing function, it can easily be left out of the NP without making its content unintelligible or even less precise, as is illustrated in (91). (91)

a. b.

Section II. … State of the popular mind in Christendom during the complete reign of Popery. (CLMETEV Foster, 1821, An essay on the evils of popular ignorance) Section II. … State of the popular mind in Christendom during the reign of Popery.

Although identifying complete is in the vast majority of instances used in definite singular count NPs, some instances can be found of definite NPs in which complete combines with plural or mass nouns as well as a small number of indefinite NPs. In the former case complete stresses only the inclusive reference conveyed by the definite article, in the latter case the function of the adjective is restricted to emphasizing the boundedness of the count noun. Firstly, plural and mass nouns are not bounded, i.e. they do not designate bounded sets of entities. Laffut (2006: 92) illustrates the difference between singular count nouns and mass nouns by means of the nouns pig and pork. “Pig, a physical entity, is instantiated in space and is also bounded in that domain: the ‘boundaries’ of the animal are inherent in the conception of a pig. Pork, on the other hand, also has space as its domain of instantiation, but does not show bounding there: the noun pork does not impose any limits, and in physical space, pork appears in all kinds of forms: as a hump of meat at the butcher’s, in slices in the fridge, in thick dices in carbonara-sauce, etc.” Laffut (2006: 93) further explains how plurals, like mass nouns, are unbounded. “The portion occupied by raisins, plants, dogs, etc. is not delimited, as the nominal does not impose any limits on the set of raisins, plants or dogs.” In definite NPs with a plural or mass head noun, as in (92) and (93), the function of secondary determiner complete is thus limited to stressing the inclusive reference inherent in the definite article, either to all of the entities of the contextual reference set or to the entire reference mass. Secondly, although no examples were attested in the data, identifying complete can also be found in indefinite NPs, as in (94). In these examples, complete stresses only the boundedness of the NP referent.

Complete

(92)

(93) (94)

133

At one site alone, Project Gutenberg, there are thousands of books available for browsing or copying on to your own computer. These range from the complete works of Thomas Hardy and Mark Twain to Shakespeare, Milton Julius Caesar's commentaries (in Latin) and the Communist Manifesto. (CB times) The Complete Prose of Woody Allen (http://www.amazon.com/Complete-Prose-WoodyAllen/dp/0517072297, accessed 21 September 2011) FloorCam is no way to watch a complete game of basketball -much as CatcherCam is inappropriate for nine innings -- and its distance from toes to head obscures the identities of the leaping bodies. (http://www.nytimes.com/2002/11/26/sports/tv-sportscamera-s-vertical-view-of-the-nba-is-all-legs.html, accessed 21 September 2011)30

In accordance with the case study presented in Davidse, Breban, and Van linden (2008), the present case study of complete suggests that the secondary determiner uses developed from the bounded descriptive uses of the adjective through a process of deictification, i.e. a “semantic shift by which a general relation expressed by the adjectives is given a subjective reference point in or relative to the speech event” (Davidse, Breban, and Van linden 2008: 475). In its descriptive ‘completed’ and ‘having all its parts’ senses, complete is clearly associated with a boundary, more specifically, the end-point of a particular period or process or the maximal extension of an entity respectively. As such, bounded descriptive complete intrinsically invokes an objective reference point. In the shift to secondary determiner, this objective reference point is transformed into a subjective reference point, linked to the speech event, i.e. the full contextual realization of a thing or mass of which the speaker and hearer share awareness and about which complete confirms that it coincides with the referential set. Complete has thus undergone a functional shift from attributing the gradable property of completeness to the NP referent to expressing, together with the primary determiner, the specific identifiability status of the referent. It now stresses the coincidence of the NP referent with its full contextual realization and, as such, has a textually intersubjective function, steering the interpretation by the hearer towards the intended referent and assuring a joint focus of attention. This intersubjectification process and shift from description to identication 30. I owe this example to Hendrik De Smet.

134

The completeness adjectives

entails a number of semantico-syntactic changes. First, whereas the early secondary determiner uses, like their descriptive source uses, are attested mainly in combination with indications of time, their collocational range quickly widens so that identifying complete can, for instance, also be combined with nouns referring to abstract numbers, as in (95). (95)

the late difference arisen in the Corporae-on of Our City of Dublin in that Our Kingdome, about the Election of Common Councell men out of the severall Companies, … Wee have thought fit in pursuance thereof to signify to you Our Pleasure that the last choice made by the Lord Mayor of that Our Citty of the compleate number of Common Counsell men all at once, leaving out the ten or eleven Roman Catholiques that were chosen at the first choice, shall standard be confirmed, … (PPCEME Charles, 1670, Letters)

The original Early Modern English collocates of complete, however, largely fall into two groups: measure nouns – either highly concrete (year, week, number) or more abstract (reign, circumference) – and holonyms (Davidse 2009: 272) such as range, series, and list. The collocational behaviour of secondary determiner complete thus reflects the genetic link with the descriptive source meanings ‘completed’ and ‘having all its parts’. In later stages, the use of secondary determiner complete is restricted only in the sense that it requires nouns that are in some way measurable or divisible into parts. Second, in the process of change from description to identification, focus shifted from the end-point of a particular entity or action to the whole of the referent. In (91), for example, the speaker does not draw attention to the final stage but to the entire period or reign of Popery. Third, the collocational and semantic changes are complemented by changes in the NP syntax. As mentioned in Section 1.4, secondary determiners form a functional unit with any primary determiners in the NP. Unlike descriptive adjectives they do not directly modify the head of the NP and are consequently infelicitous as predicates of the head. For instance, complete in the predicative construction ‘the reign of Popery is complete’ can only be understood as attributing ‘full realization’ to the period of papal rule, i.e. as descriptively modifying the head of the NP. A secondary determiner interpretation, focusing on the full instantiation of the NP referent, is excluded. The semantic and syntactic unity of secondary and primary determiner is also reflected in the loss of gradability of complete (*the more/most complete reign of Popery).

Complete

135

5.1.3. Noun-intensifying uses of complete In the data two distinct noun-intensifying sense strains were found for complete, one clearly invoking a closed scale, as in (96), and one allowing both closed and open scale construals, as in (97). As will be shown in the following paragraphs, these noun-intensifying uses have developed from two distinct types of descriptive meanings and have emerged in collocationally restricted contexts. (96) (97)

open up any cupboard in the flat and you’ll see complete chaos. (CB times) Look you complete turd give us a job now or I’ll nut you (CB ukspok)

5.1.3.1. Closed scale noun-intensifying uses Diachronic data study shows that the closed scale noun-intensifying use of complete, as in (96), can be traced back to the descriptive sense ‘realized to its full extent’. Complete has shifted from attributing a property to the head noun to intensifying the type specifications or characteristics inherent in it. Not surprisingly then, noun-intensifying complete, like all nounintensifying adjectives, has a preference to combine with degree nouns (Bolinger 1972; Paradis 2000). It is exactly the gradability of the head noun that allows for and creates a suitable onset context for the development of the noun-intensifying reading and hence the grammaticalization of the original descriptive reading. Noun-intensifying readings of complete are particularly prone to develop in contexts with nouns such as failure, triumph or success, which describe momentous events or final states (see Table 9). In example (98) complete stresses that the attempt to impress the inhabitants failed completely or in every respect. In other words, complete locates the failure not very high on an open-ended scale but at the extreme end of a closed scale of intensification. (98)

After the taking of Cairo, when General Buonaparte wished to produce an effect upon the inhabitants, he not only made them a speech, but supplemented it with the ascent of a fire balloon. The attempt was a complete failure, for the French alone looked up to the clouds to see what became of the balloon. (CLMETEV Bacon, 1902, The dominion of the air)

136

The completeness adjectives

The close relation between the bounded descriptive and closed scale noun-intensifying uses of complete can still be witnessed in the Present Day English data. First, complete can have either a descriptive or an intensifying reading in combination with the same head noun, depending on the context in which it occurs. Whereas in (98) the NP complete failure receives a noun-intensifying reading emphasizing the extent to which something failed, it favours a descriptive reading in (99). Here the completeness of the failure is being objectively measured and, in contrast to (98), it is possible to transform the NP into the predicative construction ‘the failure was not complete’. The considerable collocational overlap between the descriptive source meanings and the noun-intensifying target meanings is clearly shown in Tables 8 and 9. (99)

the greatest surprise is that the Paris talks did not end in complete failure. The pressure placed on the Cambodian government ... seems to had some moderating affect on Prime Minister Hun Sen (CB ukspok)

Second, the connection between the descriptive and noun-intensifying uses is reflected in the occurrence of blends (Bolinger 1961a, 1961b; Aarts 2007: 187–189), as in (100). Blends provide “formal evidence of overlapping semantic fields” (Bolinger 1961a: 19) as they combine properties of two constructions allowing speakers to convey two meanings at the same time, in this case both a descriptive and an intensifying meaning. Because the two uses are so intertwined within these constructions, they do not “pose an either-or choice” (Bolinger 1961a: 22) for speakers or hearers. Rather both readings are construed simultaneously and are difficult to recognize independently (Aarts 2007: 189). Complete in (100) can on the one hand be understood as characterizing or describing the success as fully realized ('the success is complete'), a property which can be objectively measured by the number of difficulties overcome, perils encountered and results obtained. At the same time, the NP complete success also triggers a noun-intensifying reading, emphasizing the extent of the success and its positive evaluation. (100)

The operations conducted by Colonel Parsons thus ended in complete success. Great difficulties were overcome, great perils were encountered, great results were obtained. (CLMETEV Churchill, 1899, The river war, an account of the reconquest of the Sudan)

Emotional states

Abstracts

Human types, pursuits, professions

Wholes

Actions, events, states

5

n

35.3

%

5.9

%

7.1

2

1

n

11

14.3

4

21.4

6

17.9

5

39.3

Majesty

knave, master, gentleman, man, champion

armour, steel, bosom

perfection, tractates, history

joy, happiness

idea, standard

angler, cook, gentlemen

armour, piece, way

history, victory, remission, conquest

1640-1710

%

n

6

n

29.4

29.4

%

%

5

n

1500-1640

8.6

5

6.9

4

46.6

27

37.9

22

manner, idea, contrast, character

huswife, orator, horsemen, lawyer

(suit of) armour ((of) steel), set, system, whole, picture

manufacture, work, victory, account, effect, isolation, salvation

1710-1780

13.5

5

13.5

5

2.7

1

34.4

12

37.8

14

sympathy, confidence

manner, feature

convert

knowledge, account, change, demonstration, standstill, salvation (suit of) armour ((of) steel), set, collection, whole, picture, lists

1780-1850

Table 8. Nouns co-occurring with descriptive modifier complete in the CEMET, PPCEME and CLMETEV data sets

Complete 137

Other

Emotional states, dispositions

Human types

Abstract qualities

Momentous states

Actions, events, states

3 gentleman, Hottentot, 100.0 Mystics

4 13.3

n

%

%

n

%

n

desert, dungeon

5 one-sidedness, similarity, puzzle, 16.7 groundlessness, gentleman, 8 master, devil, free thinker, Jacques Rat26.7 de-Bif 4 delirium, delusion, 13.3 confidence

n

%

16.7

%

success, victory

5

4 stillness, silence, 13.3 darkness, exile

1780–1850

n

%

n

1710–1780

Table 9. Nouns co-occurring with noun-intensifying complete in the CLMETEV data sets

moralist, despot, stranger

harmony, labyrinth

5 unconsciousness, good humour, 16.7 cock-sureness 2 disguise 6.7

10.0

3

6.7

2

1850–1920 idleness, solitude, 11 mistake, error, 36.7 silence, absence 7 success, defeat, wreck, 23.3 ruin, failure

138 The completeness adjectives

Complete

139

5.1.3.2. Noun-intensifying uses hovering between an open scale and a closed scale reading Like the bounded descriptive senses, the descriptive ‘perfect’ sense gave rise to a noun-intensifying use of complete in specific collocational contexts. In combination with head nouns describing people, such as knave, villain, and gentleman, as in (101) and (102), complete can come to intensify the negative or positive qualities inherent in these nouns. Again, these collocational contexts in which complete combines with nouns conveying inherently gradable properties serve as enabling onset contexts for the grammaticalization and concomitant subjectification of the adjective. The head nouns with which complete in this use originally co-occurs are a subset of the nouns with which it occurs in its descriptive use (Tables 8 and 9). In these combinations, a subjectification process has taken place from attributing – on a subjective basis – the property of perfection to a specific entity to intensifying or upscaling its inherent properties (Ghesquière and Davidse 2011). (101)

(102)

In genius the lowest, ill-bred and obscene; In morals most Wicked, most nasty in mien; … A dupe in each treaty, a Swiss in each vote; In manners and form, a complete Hottentot. (CLMETEV Walpole, 1742, Letters) Foregad! thou are a complete fellow (cried my uncle, still laughing) (CLMETEV Smollett, 1771, The expedition of Humphrey Clinker)

The first true noun-intensifying uses of this type appear from the period 1710–1780 onwards. In later periods, its collocational range is extended to nouns not referring to people, but nouns describing emotions or attitudes, as in complete astonishment, complete amazement and complete good humour (103), and nouns typifying states of affairs, as in complete harmony and complete disgrace (104). These head nouns are more similar to the event or action nouns with which complete combines in its closed scale noun-intensifying use. (103)

“Everything about me is a sham,” responded the ex-courier in complete good humour. (CLMETEV Chesterton, 1914, The wisdom of father Brown)

140

(104)

The completeness adjectives

All along the line the CSA has been a complete disgrace: for the fact is, since its very inception it has not been allowed to succeed (CB times)

Like its descriptive source meaning, with which it displays considerable collocational overlap (see Tables 8 and 9), this noun-intensifying use hovers between a bounded and an unbounded construal or a closed scale and an open scale reading respectively. On the one hand, complete in the NP a complete Hottentot in (101), for instance, can be understood to indicate that someone behaves like or is a typical Hottentot in every respect, thus evoking a closed scale interpretation. On the other hand, a complete Hottentot can characterize someone as being very low in culture or very uncouth, thus placing the properties inherently associated with Hottentots very high on an open-ended scale. Similarly, the NPs in (103) to (104) can be interpreted either as invoking gradable qualities which complete places high on openended scale (e.g. good humour and disgrace evoke qualities such as ‘pleasant’/’friendly’ and ‘shameful’/‘scandalous’ respectively which complete can intensify) or as invoking closed qualitative scales of which complete indicates the ultimate point (e.g. the good humour is present in all aspects of the answer and the organisation is a disgrace in every respect). The indeterminacy between bounded and unbounded construal displayed by the descriptive ‘perfect’ and ‘consummate’ senses thus continues to exist with the noun-intensifying use that developed from them, as it allows both closed and open scale interpretations. Kennedy and McNally (2005: 354n) have made a similar observation for the adverb completely, which displays the same kind of potential to convey both closed-scale and open-scale intensification. In its prototypical use, the maximality modifier completely invokes a scale of intensification that is closed both at the lower and the upper end, as in The figure was completely visible/invisible (Kennedy and McNally 2005: 355). In addition, completely also has a use that is “roughly synonymous with very” (Kennedy and McNally 2005: 354n), as in (105) and (106). In these uses, in which the adverb modifies a verb, as well as in combination with unbounded adjectives such as nice in completely nice, the bounded maximizer completely is “reinterpreted as a (high) booster” (Traugott 2007b: 527), i.e. as an unbounded, open scale degree modifier. (105)

I will tell you that things changed completely after we had our first child. And changed even more after we had our second child. (http://forums.catholic.com/showthread.php?t=253548&page=2)

Complete

(106)

141

But weapons have changed completely, as the Gulf War of 1991 dramatically demonstrated; military doctrines and concepts have changed even more drastically (Drucker. 2003. On the profession of management. Available online at http://books.google.be)

Like the closed scale ones, the open scale noun-intensifying uses are still closely linked to their descriptive source meaning. This is reflected in the occurrence of blended constructions, as in (107). Here, complete displays the syntactic behaviour of a descriptive adjective as it occurs in the superlative form. Influenced by the strongly evaluative degree noun fool, however, complete receives an intensifying reading rather than a descriptive one. The young man in question is not described as a fool who is ‘complete’ or ‘accomplished’. The function of complete is restricted to emphasizing the negative evaluation inherent in the head noun fool and hence to heightening the degree of foolishness associated with the NP referent. (107)

an intelligent young man when he was sober; but, the moment the wine began to operate, he was one of the completest fools in christendom (CLMETEV Hunt, 1820–22, Memoirs of Henry Hunt)

5.1.4. Adjective-intensifying uses of complete The intensification of adjectives is typically associated with adverbs such as really in a really expensive watch or rather in a rather peculiar situation. In the contemporary data set, however, a (small) number of adjectiveintensifying uses were found of the adjective complete.31 In these uses complete indicates that the quality denoted by the descriptive modifier it has scope over is realized to its fullest extent. As such, the adjective is fully synonymous with its adverbial counterpart completely and is best paraphrased as such. For example, the money in (108) can be said to be ‘completely new’ and the proverbial sheet in (109) is characterized as ‘completely clean’. The functional overlap with the adverbial form probably explains why the adjective-intensifying use of complete is so restricted. Interestingly, the oldest attestation of the adverb completely listed in the OED (OED

31. Four of the five adjective-intensifying uses found overall were found in the spoken subcorpus.

142

The completeness adjectives

s.v. completely) stems from the early 16th century,32 whereas the “quasiadverbial” use of complete is registered as dating back to the 14th century (OED s.v. complete 9). As a consequence, it is plausible that adjectiveintensification was first fulfilled by the adjective complete, which was later ousted from this function by the adverb completely. Further study into the development of the adverb could certainly shed more light on this matter. (108) (109)

We’ve been able to move limited resources around in [suburb name] but not produce complete any complete new money around. (CB ukspok) Erm coming onto health you've got a complete clean sheet. So not a single visit to a doctor health centre pharmacy in the whole of the six months (CB ukspok)

5.1.5. From description to identification and noun-intensification Diachronic data study of prenominal complete has provided insight into the intricate developmental paths followed by the adjective. First, it was argued that the identifying secondary determiner uses of complete developed from the bounded objective descriptive uses through a process of deictification (§3.4.1) and (textual) intersubjectification. In the process focus shifted from the maximal or end-point of a specific entity or event to the whole of the entity, while the quantifying semantics remained largely unaltered. The shift from descriptive modification to identification is supported by the quantified data represented in Table 10. From their emergence the proportion of secondary determiner uses in the data sets gradually increased – from 3.3% in the period 1640–1710 to 18.5% in the Present Day English data –, whereas the proportion of objective descriptive uses has dropped from 33.3% to 13.5%. Moreover, the secondary determiners were originally restricted to co-occur with head nouns with which the objective descriptive uses combined, namely measure nouns (e.g. nouns referring to periods of time) and nouns such as range,

32. The earliest adjective-intensifying uses listed are even more recent, dating from the early 18th century (OED s.v. completely). (i) 1712 BUDGELL Spect. No. 425 P. 3 A Man completly armed. (ii) 1766 GOLDSM. Vic. W. ii, Miss Arabella Wilmot was allowed by all to be completely pretty.

Complete

143

series, and list. However, due to the abstract semantics of the secondary determiner use of complete its collocational range quickly widened.

(un)bounded descriptive modifier

bounded descriptive modifier

adjective-intensifier

total

0

0

0

0

0

1

0

1

%

0.0

0.0

0.0

0.0

0.0

100.0

0.0

N

0

0

2

9

1

6

0

%

0.0

0.0

11.1

50.0

5.6

33.3

0.0

N

1

0

2

9

3

16

0

%

3.3

0.0

6.5

29.0

9.7

51.6

0.0

1710–1780 N (CLMETEV) %

6

3

6

22

0

36

0

8.2

4.1

8.2

30.1

0.0

49.3

0.0

9

30

4

18

0

19

0

11.3

37.5

5.0

22.5

0.0

23.8

0.0

1500–1570 (PPCEME, CEMET) 1570–1640 (PPCEME, CEMET) 1640–1710 (PPCEME, CEMET)

n 1780–1850 (CLMETEV) %

noun-intensifier noun-intensifier/ bounded descriptive modifier

N

secondary determiner

unbounded descriptive modifier

Table 10. Quantified results of the corpus study of complete

n 1850–1920 (CLMETEV) %

10

30

16

35

0

23

0

8.8

26.3

14.0

30.7

0.0

20.2

0.0

n

33

39

26

16

0

13

1

%

25.8

30.5

20.3

12.5

0.0

10.2

0.8

n

22

88

18

10

0

27

4

%

13.0

52.1

10.7

5.9

0.0

16.0

2.4

n

55

127

44

26

0

40

5

%

18.5

42.8

14.8

8.8

0.0

13.5

1.7

CB times CB ukspok CB total

18 31 73 80 114 128 200 328

Bounded Descriptive modifier ‘realized to its full extent’ the horrible cruelties of the Romans, which ... might give some cause for the complete isolation of the Jew from the rest of the world. (CLMETEV Gibbon, 1776, The decline and fall of the Roman Empire)

Figure 14. The development of identifying and noun-intensifying uses of complete

Closed scale noun-intensifier open up any cupboard in the flat and you’ll see complete chaos. (CB times)

Bounded descriptive modifier ‘completed’ It costs more labour, and therefore more money, to bring first the materials, and afterwards the complete manufacture to market. (CLMETEV Smith, 1766, An inquiry into the nature and causes of the wealth of nations)

Secondary determiner ‘reference to the entire entity’ follow washing instructions .. taking care not to immerse the complete unit in the water. (CB ukephem)

Bounded descriptive modifier ‘having all its parts’ The fossil skeleton, the most complete example of its kind, was in danger of being washed out to sea by coastal erosion and winter storms (CB times)

Open/closed scale nounintensifier Look you complete turd give us a job now or I'll nut you (CB ukspok)

Bounded/unbounded descriptive modifier ‘consummate’ The Compleat Angler (CEMET Walton, 1653, The complete angler)

Bounded/unbounded descriptive modifier ‘perfect in nature or quality’ Men shulden bi hooli lif of Crist trowe that his lawe is compleet. (OED Wyclif,

144 The completeness adjectives

Complete

145

Second, two distinct shifts were posited from descriptive meanings to noun-intensifying meanings: from the bounded ‘fully realized’ meaning, as in complete isolation, to closed scale noun-intensifying meanings, as in a complete failure; and from the ‘perfect’ and ‘consummate’ meanings, as in a complete gentleman, to noun-intensifying uses such as a complete idiot, which are both indeterminate between a bounded and an unbounded construal. In both cases, the boundedness construal of the source meaning is retained and foregrounded in the noun-intensifying target meaning (Paradis 2000, 2001, 2008; §3.4.2.1). These changes proved to be strongly collocationally driven: the noun-intensifying uses arose with a specific subset of the head nouns co-occurring with the descriptive uses conveying gradable notions, i.e. degree nouns inviting either closed scale or open scale intensification. Also, unlike with the secondary determiners, the noun-intensifying uses remain highly collocationally restricted. Whereas secondary determiner complete typically reinforces the boundedness construal of its head noun and the universal relative quantification inherent in the primary determiner, in its noun-intensifying use complete closely interacts with the lexical specificities of the head nouns. Consequently, noun-intensifying complete is in Present Day English still restricted to occur with degree nouns inviting intensification of inherent gradable notions. The development or grammaticalization of noun-intensifying complete is thus at the same time enabled as well as contrained by specific collocational contexts (§3.2.1). An overview of the different shifts toward the identifying and noun-intensifying uses of complete is given in Figure 14. 5.2. Total Like complete, total is a Romance loanword and appears in the data only from the Early Modern English period onwards. Also, as for complete, Paradis (2000) and Athanasiadou (2007) have argued that total has undergone a process of subjectification and grammaticalization as it shifted from the descriptive meaning ‘entire, relating to the whole of something’, as in a total ban, to a noun-intensifying meaning, as in total crap and total contempt. In the following paragraphs the validity of this diachronic claim will be tested by means of a historical data study of the different prenominal uses of total.

146

The completeness adjectives

5.2.1. Bounded descriptive modifier uses of total33 As for complete, Paradis (2000: 236) has argued that total is “associated with a bounded, ‘either-or’ conceptualisation (Paradis 1997: 57–58), and can be limited by, for instance, almost”, as in (41) and (42). In fact, the typical descriptive use of total is semantically very similar to the ‘fully realized’ sense of complete (§5.1.1.1), as it typically characterizes actions or states as being ‘complete in extent or degree’ (OED s.v. total 3.a). Correspondingly, total also most frequently combines with nominalizations of verbs such as defection in (110) and seclusion in (111) (see Table 11). As noted for complete, the adjective total here closely resembles its adverbial counterpart totally. The NPs in (110) and (111) can easily be paraphrased as totally defected and totally secluded respectively. (110)

(111)

Gods commaundements or prohibitions were not the originals of good and euill, but that they had other beginnings which man aspired to know, to the end, to make a totall defection from God, and to depend wholy vpon himselfe. (PPCEME Bacon, 1605, The advancement of learning) Ann Yearsley … died many years after, in a state of almost total seclusion, at Melksham. (CLMETEV Cottle, 1847, Reminiscences of Samuel Taylor Coleridge and Robert Southey)

From the period 1640–1710 onwards, the collocational range of this descriptive use of total is extended beyond nominalizations of verbs, primarily to nouns referring to states of affairs such as darkness in (112) or independence in (113). As the frequency lists in Table 11 show, the NPs in which total occurs in its bounded sense, both with head nouns from the original collocational set and with the later extensions, typically have a negative connotation (see also Bäcklund 1973: 210).

33. Although the bounded descriptive meaning expressed by total is very similar to that of complete, total does not share complete’s proneness to unbounded interpretations. No more than two (contemporary) examples were attested in the data which could receive an unbounded interpretation (e.g. His company is called Total Sports Image, presumably to distance itself from such concepts as total sports substance. [CB times]). Considering how rare they are, it is unlikely that these unbounded uses have played a role in the later development of the identifying or noun-intensifying uses of total.

1570–1710 destruction change defeat excuse rout adieu conversion darkness defection depravation

3 2 2 2 2 1 1 1 1 1

1710–1780 destruction defeat dissolution scepticism extinction separation suspense exclusion loss prohibition 9 5 4 4 3 3 3 2 2 2

1780–1850 abolition darkness silence change blindness ignorance absence deafness destruction disappearance 9 7 5 3 2 2 1 1 1 1

1850–1920 abstinence darkness destruction loss abstainer cessation condemnation difference extinction ruin 2 2 2 2 1 1 1 1 1 1

CB times ban commitment abandonment breakdown capitulation clearance control disarmament domination immersion

4 2 1 1 1 1 1 1 1 1

CB ukspok blackout recall control disorder empathy exploitation interruption objectivity permission pollution

Table 11. Nouns co-occurring with descriptive modifier total in the PPCEME, CEMET, CLMETEV and CB data sets

3 2 1 1 1 1 1 1 1 1

Total 147

148

(112)

(113)

The completeness adjectives

she approached the door that had been opened; but a sudden gust of wind that met her at the door extinguished her lamp, and left her in total darkness. (CLMETEV Walpole, 1764, The castle of Otranto) Strangely, many voted `yes" to ensure their voices were heard in the rest of Canada, but they never voted for total independence

The future is uncertain, said Mr Webber, but the separatists' narrow defeat would resonate throughout the country (CB times)

5.2.2. Identifying uses of total The secondary determiner use of total, which is attested among the earliest prenominal attestations of the adjective, is clearly quantifying in nature, indicating that reference is made to the entire entity referred to by the NP. Most frequently, identifying total collocates with head nouns indicating “a certain amount or number that is reached after everyone or everything is counted or added together” (Oxford Advanced Learner’s Dictionary s.v. total 2). The head nouns are thus typically made up of several individual entities that can be considered either independently or as a whole. Total then emphasizes that reference is being made to the totality of the different components, not to the individual parts, as in (114) and (115). Like secondary determiner complete, identifying total in the data set always combines with a definite primary determiner, most typically the definite article, and can thus be considered an emphasizing secondary determiner. As noted in Section 5.1.2., the definite article implies inclusive reference, i.e. reference to the entire potential instantiation of the type designated by the NP (Hawkins 1978; Declerck 1991: 292; Langacker 1991: 102). Secondary determiner total then reinforces the implicature of inclusive reference inherent in the primary determiner. Also, like complete, total is predominantly used in combination with singular count nouns and thus typically stresses not only the quantification inherent in the definite article but also the boundedness of the count noun construal (§5.1.2). Whereas secondary determiner complete is added to the NP mainly in contexts of doubt or counter-expectation, total is added to the NP to indicate that the focus is on the entire entity rather than on its component parts. (114)

Sinus totus, is the Semidiameter of the Circle, and is the greatest Sine that may be in the Quadrant of a Circle, which according to the first tables of Monte Regio containeth 6 000 000 and according

Total

(115)

149

to the last tables 10 000 000 parts, for the more parts that the totall Sine hath, the more true and exact shall your worke bee (PPCEME Blundeville, 1597, A briefe description of the tables of the three speciall right lines belonging to a circle, called signes, [sic] lines tangent, and lines secant.) The announcement brings the total number of jobs lost by Vernons this year to 200 and reduces the workforce to 425. (CB times)

Interestingly, in the NP the sum total, as in (116) and (117), total appears to express exactly the same meaning, albeit in postnominal rather than prenominal position.34 “The sum total of a number of things is all the things added or considered together” (Sinclair 1990: 1653). As it invokes a mathematical addition operation, Sapir (1930: 16) has labelled the phrase the sum total an “additive calculated totalizer”. The OED (s.v. total 2) provides examples of this noun-adjective combination dating back to the year 1400. In the data sample compiled for this study it is found from the period 1640–1710 up until the period 1850–1920. That the identifying function of adjectives can also be realized postnominally is consistent with Bache’s (2000) characterization of the post-head zone as ‘multi-functional’. He argues that it can accommodate not only modifying and categorizing elements but also determining elements (§1.1.2). Breban (2010a: 274–282), for instance, discusses the postnominal secondary determiner uses of the adjectives of comparison, as in causes other than Israel’s occupation of the West Bank and Gaza Strip and there is always something different to see. (116) (117)

When a man Reasoneth, hee does nothing els but conceive a summe totall, from Addition of parcels (CEMET Hobbes, 1651, Leviathan) A merchant is desirous of knowing the sum total of his accounts with any person: Why? but that he may learn what sum will have the same effects in paying his debt, and going to market, as all the particular articles taken together. (CLMETEV Hume, 1739–1740, A treatise of human nature)

34. The postnominal position of total in the sequence sum total is most likely due to the fact that the NP as a whole was borrowed from French, which may also explain why the combination is attested so early in English. Typically preceded by the definite article, (la) somme totale refers to an entity “pris dans son entier” [taken as a whole] (Le Petit Robert de la langue française, s.v. total 1).

150

The completeness adjectives

Secondary determiner uses of total occur in the data from the period 1570–1640 onwards (see Table 13). The data for this period are, however, not very reliable. All 50 examples of secondary determiner uses found for this period come from the same text in the PPCEME corpus, and they are all used in the NP the total sine which refers to one of the three fundamental trigonometrical functions (OED s.v. sine² 2). Ignoring the 1570–1640 data, a slow but steady increase in the proportion of secondary determiner uses can be observed from the period 1640–1710 up until the Present Day English period, from 4.2% to 56.6%. A reverse trend is observed for the bounded, objective descriptive modifier uses, whose proportion has gradually decreased from 95.8% to 10.2% in the periods following its occurrence in English in the period 1570–1640. These quantified results support a diachronic deictification scenario in which the bounded descriptive uses lie at the origin of the secondary determiner uses, which once established take up an increasingly large proportion of all prenominal attestations of complete. The semantics of these two uses also point in the direction of such a process of subjectification. From attributing the gradable quality of totality visà-vis some objective reference point, as in (111), total came to specify the entity referred to by the head as the whole contextual instantiation of the referent which speaker and hearer have in mind, as in (115). As with complete (§5.1.2), an objective reference point is thus transformed into a subjective reference point related to the speech event, constituting subjectification in the Langackerian sense (§3.1). As a secondary determiner, total now engages in the identification of the NP referent and serves to (emphatically) point the hearer’s attention toward the intended referent. As such, it now has a textually intersubjective function. In contrast to what we found for complete, collocational evidence for the shift from description to identification is, however, lacking. Although total combines with nouns that are in some way measurable or quantifiable in both its secondary determiner and its descriptive uses, the collocational range of the former uses does not show any overlap with the collocational set of the latter uses. Whereas the descriptive uses typically collocate with nominalizations of verbs, secondary determiner uses are found in combination with nouns denoting measurable entities that consist of distinct component parts, such as sum, amount, number, and cost. Collocational overlap is, however, more likely to be attested with shifts from descriptive to nounintensifying meanings than with shifts from descriptive to identifying meanings (§3.2.1). In contrast to noun-intensifiers, secondary determiners do not engage with the lexical specificities of the head noun. Rather secondary determiners work together with the primary determiner and, in the

Total

151

case of the completeness adjectives, reinforce the boundedness construal of the noun, largely independent of the semantics of the head noun. What links the descriptive and identifying uses of total is not collocational overlap, but boundedness. The descriptive source use of total is typically construed as bounded, describing entities as fully realized, i.e. as reaching their quantificational limit. As a secondary determiner, total reinforces the inherent boundedness of the nouns it modifies. 5.2.3. Noun-intensifying uses of total The following examples illustrate the typical noun-intensifying use of total, which is attested from the first stage of the Late Modern English period onwards and indicates that something is ‘as great in extent, degree or amount as it possibly can be’ (Sinclair 1990: 1653). (118) (119)

All you need is a little nerve and a total disregard for the truth. (CB times) `There can be no justification for gunning down a man in front of his wife in this cowardly fashion; to do so in the week before Christmas illustrates the total callousness of these vicious gunmen (CB times)

In the literature, total and complete are often discussed together and are argued to have developed along the same cline of semantic change, i.e. from property attribution to noun-intensification (Paradis 2000; Athanasiadou 2007). More specifically, for total it has been claimed that a semantic shift has taken place from describing something as ‘relating to the whole of something’ or ‘affecting or including everything’ to intensifying properties inherent in the head noun to a maximum degree (Paradis 2000: 235; Athanasiadou 2007: 558–559). These paraphrases of the descriptive use, however, easily invoke an identifying secondary determiner reading. The bounded descriptive source meaning is better paraphrased as ‘complete in extent or degree’, more clearly distinguishing it from the secondary determiner sense which refers to the entire NP referent. The shift from descriptive modification to noun-intensification can be characterized as a process of invited inferencing (Traugott and Dasher 2002; Traugott 2003a), whereby the intensifying meaning that is at first only implicitly present is later semanticized and foregrounded (Paradis 2000, 2001, 2008). In its bounded descriptive meaning, as in (120), total

152

The completeness adjectives

describes the defeat as complete in extent. In other words, it ascribes the quality of totality to the deverbal head noun defeat; a quality that can be objectively established, for instance on the basis of the number of men that died in the fight. In (121), the comparative form is indicative of the descriptive modifier status of total, yet the intensifying meaning is in this example already to some extent invited by the speaker. The noun defeat generally has a negative connotation and use of the comparative emphasizes that the defeat truly was complete or realized to its full extent. Examples (122) and (123) illustrate true noun-intensifying uses, indicating that the NP referents are ‘as great in extent, degree or amount as [they] possibly can be’ (Sinclair 1990: 1653). In these examples, the strengthening noun-intensifying meaning has now become the most prominent one and has semanticized. (120)

(121)

(122)

(123)

The troops of Licinius, though they were lately raised, ill armed, and worse disciplined, made head against their conquerors with fruitless but desperate valor, till a total defeat, and a slaughter of five and twenty thousand men, irretrievably determined the fate of their leader. (CLMETEV Gibbon, 1776, The decline and fall of the Roman Empire) In short, Pharaoh at the Red Sea, Darius at Arbela, Pompey at Pharsalia, Edward at Bannockburn, Charles at Pultoway, Burgoyne at Saratoga--no prince, potentate, or commander of ancient or modern unfortunate memory ever got a more shameful or more total defeat. (CLMETEV Burns, 1780–1796, The Letters of Robert Burns) These figures do not present our total failure, they merely show how far the less fortunate section of the community falls short of the more fortunate. (CLMETEV Wells, 1902–1903, Mankind in the making) It is the nature of despotism to abhor power held by any means but its own momentary pleasure; and to annihilate all intermediate situations between boundless strength on its own part, and total debility on the part of the people. (CLMETEV Burke, 1770, Thoughts on the present discontents)

In its noun-intensifying use total, like in its descriptive source meaning, is strongly biased towards bounded construal. As a noun-intensifying adjective total places the bounded properties invoked by the head high on a closed scale. This observation confirms Paradis’s (2000: 235–239) claim that the bounded or unbounded nature of the descriptive uses determines

Total

153

the type of noun-intensifying uses an adjective will develop. The gradability mode or boundedness construal of the adjective in its descriptive source meaning is simply made prominent in the noun-intensifying use, while its lexical contents are backgrounded (Paradis 2000, 2001, 2008; §3.4.2.1). Besides boundedness, the head nouns with which noun-intensifying total combines are also characterized by negative semantic prosodies. Total in this use typically combines with heads that have a negative, unpleasant or bad connotation (Bublitz 1996: 11; Partington 2004: 150) (see Table 12). More specifically, the semantic preference (Stubbs 2001: 65, 88; Partington 2004: 145) of total goes to nouns indicating the absence of a particular property, which is experienced as a negative thing, e.g. infirmity, want, impossibility, loss. As such, there is considerable collocational overlap with the descriptive source reading of total (see Table 11). Partington (2004: 148) and Aijmer (2007: 11) have noted a similar preference for ‘absence’ words for the adverb totally in English in general and in British English respectively. In the data used for this study, collocates with a positive connotation, as in (124) and (125), are found only in the Present Day English sample and are limited to sporadic occurrences. (124) (125)

Ireland is my idea of total heaven. (CB times) Also for me what was what I wanted to do in terms of the progression of the play was to present a room which to Rita and from somebody coming from Rita's environment with Rita's attitude the room was everything and the room was her idea of total womblike luxury. (CB ukspok)

One other interesting collocational observation deserves mention here. Whereas the earliest noun-intensifying uses of complete are found in combination with nouns referring to persons, noun-intensifying total originally exclusively collocates with nouns referring to (negative) states of affairs or events. In the period 1780–1850 total is found for the first time in combination with a head noun referring to a person, namely stranger (one example out of 35 noun-intensifying uses). This is also the only attested person-noun in the period 1850–1920 (three examples out of 27 noun-intensifying ones). In the Present Day English data the range of person nouns is widened to include the nouns fraud, slob, beginner and wimp, but still remains highly restricted. This type of noun-intensifying use also continues to take up only a minor proportion in the entire set of attested contemporary nounintensifying uses (eight out of 82 instances). Interestingly, like the nouns co-occurring with the predominant noun-intensifying use, these person

1710–1780 want neglect ruin debility deformity impossibility infirmity negligence oblivion pravity

3 2 2 1 1 1 1 1 1 1

1780–1850 want disregard loss blank debility debasement dereliction destitution discomfiture dissipation 11 2 2 1 1 1 1 1 1 1

1850–1920 want absence stranger failure misapprehension blank cataclysm collapse destruction disregard 4 3 3 2 2 1 1 1 1 1

CB times collapse callousness disregard fabrication failure frauds heaven mess non-being onslaught 4 1 1 1 1 1 1 1 1 1

Table 12. Nouns co-occurring with noun-intensifying total in the CLMETEV and CB data sets CB ukspok waste lack failure stranger crap disaster frustration beginner desperation luxury

10 5 3 3 2 2 2 1 1 1

154 The completeness adjectives

Total

155

nouns are characterized by (strongly) negative semantic prosodies, as in (126) and (127). In terms of boundedness construal, noun-intensifying total, in combination with person nouns, seems to hover between a bounded and an unbounded reading. Such indeterminacy was also observed for the corresponding type of noun-intensifying uses of complete (§5.1.3.2). The NP a total stranger can be interpreted as referring to someone who is a stranger in every respect (bounded, closed scale interpretation) or to someone to whom one is not yet well acquainted, i.e. unfamiliar to a higher or lesser degree (unbounded, open scale interpretation). Similarly, a total slob is someone who is a ‘dull, slow, or untidy person’ (OED s.v. slob n.1 3) either in every respect or to the maximum level (closed scale interpretation), or simply to a very high degree (open scale interpretation). (126)

(127)

My simplicity, arising from my being a total stranger to the intercourse of the world, was accompanied with a mind in some degree cultivated with reading, and perhaps not altogether destitute of observation and talent. (CLMETEV Godwin, 1794, The adventures of Caleb Williams) I'm a different person there. I become a total slob and live in cords and big sweaters and I can get up late or party all night things I can never do in London. (CB times)

The origin of this type of noun-intensifying use cannot be clearly determined on the basis of the data studied. As total does not occur with person nouns in any of its descriptive uses, it seems unlikely that it has developed from descriptive source meanings through invited inferencing. Possibly, this type of noun-intensifying use developed through analogy (Allan 2010; De Smet 2010) with similar uses of semantically closely related adjectives such as complete, utter, perfect and entire. These adjectives are all found as noun-intensifiers in the earliest Late Modern English subperiod in combination with the head noun stranger, as in (128) and (129). Possibly, the collocational range of the head noun stranger in the intensifying construction has widened to include total, through its resemblance with the other existing collocates. In other words, a wider variety of lexemes is attracted to the existing intensifying construction (cf. Hopper and Traugott 1993: 56). (128)

it was irrecoverably gone, and I am an utter stranger to the contents. (CLMETEV Haywood, 1744, The fortunate foundlings)

156

(129)

The completeness adjectives

she was a perfect stranger in that country (CLEMETEV Fielding, 1749, The history of Tom Jones, a foundling)

5.2.4. Adjective-intensifying uses of total As for complete, some instances of total were found in the Present Day English data which may be analyzed as adjective-intensifying uses. In these NPs, as illustrated in (130), one would expect the adverbial form totally in such constructions rather than the adjectival form. Total seems to be even more restricted as an adjective-intensifier than complete, although Internet examples can be found relatively easily, e.g. (131). Unlike for complete, the OED does not make mention of the quasi-adverbial use of total. (130)

(131)

And it's like this total Yes. like communication 'cos on the other hand there's like television an and total passive you know no non-active Things that have been bought out. I mean even stations like like P B S you know Public Broadcasting Systems (CB ukspok) Lamborghini have always made total outrageous supercars and are normally reserved for pop stars or Premiership footballers, but this is your chance to drive one on the Lamborghini Driving Thrill experience day! (http://www.lamborghinidrivingexperience.org.uk/, accessed 20 September 2011)

5.2.5. Classifying uses of total Of the three completeness adjectives studied, total is the only one with attested classifier uses, albeit in a very small number and only in the contemporary data set (Table 13). The classifier-noun combinations found in the data, e.g. (132), are not well-entrenched constructions, but rather ad hoc formations that are the product of “the speaker’s own creativity [which] leads them to use an adjective in a function that is not normally associated with it” (Breban 2010a: 31), in this case the classifier function. Although the basic semantics of total does not invite classifying readings, the adjective can occasionally be used as a classifier to meet a certain expressive or communicative need of the speaker in a specific type of context. In (132), the NP total fund-holders refers to a specific subtype of fund-holders. The

Total

157

larger context is that of a local initiative that would allow these fundholders to purchase all primary and secondary health care. (132)

it hasn’t really impinged locally with us er with G P fund-holders because er only a few of them are total fund-holders and even so it’s a paper exercise at the moment. (CB ukspok)

5.2.6. From description to identification and noun-intensification Like the data study of complete, the diachronic analysis of total has basically confirmed the validity of two pathways of change of prenominal adjectives proposed in the literature. First, the secondary determiner use of total, which was found among the adjective’s earliest attestations in the corpora consulted, was argued to have developed from the bounded descriptive meaning ‘complete in extent or degree’ through a process of deictification (Davidse, Breban, and Van linden 2008). Second, following the pathway proposed by Adamson (2000), Paradis (2000) and Athanasiadou (2007), the bounded descriptive sense ‘complete in extent or degree’ also gave rise to a closed scale noun-intensifying use, as in a total disaster, through a process of pragmatic strengthening or invited inferencing and foregrounding of the boundedness construal of the source meaning. As observed for complete, the structural-semantic shift from property attribution to nounintensification is strongly collocationally driven. Further, the data also suggest an analogical extension of the closed scale noun-intensifying use to an open scale noun-intensifying use, as in total stranger. Figure 15 gives a schematic overview of the different changes leading to the identifying and closed scale noun-intensifying uses of total. The shifts from description to identification and from description to noun-intensification are supported by the quantified results of the data study (Table 13). As discussed in Section 2.2.2, I ignore the quantitative results for the period 1570–1640, as these cannot be considered representative. In the period 1640–1710 prenominal total is used as a bounded descriptive adjective in more than 95% of all cases. In the following periods, however, the proportion of bounded descriptive uses gradually, but notably drops to just 10.2% in the Present Day English data set. At the same time, a considerable increase can be witnessed in secondary determiner uses on the one hand – from 4.2% to 56.6% – and noun-intensifying uses on the other hand – from 0.0% to 22.5% –, two uses which were argued to develop from the bounded descriptive meaning of total.

158

The completeness adjectives

noun-intensifier

noun-intensifier/ bounded descr. mod.

adjective-intensifier

unbounded descriptive modifier

bounded descr. mod.

classifier

total

n 1570–1640 (PPCEME, CEMET) %

secondary determiner

Table 13. Quantified results of the corpus study of total

50

0

0

0

0

1

0

51

98.0

0.0

0.0

0.0

0.0

2.0

0.0

0

23

0

0.0 95.8

0.0

n 1640–1710 (PPCEME, CEMET) %

1

0

0

0

4.2

0.0

0.0

0.0

n

6

16

9

0

6.9 18.4

10.3

0.0

35

12

0

20.0 30.4

10.4

0.0

27

2

0

47.0 32.5

2.4

0.0

21

9

0

70.1 11.4

4.9

0.0

61

16

2

1

15

43.0 34.1

8.9

1.1

0.6

8.4

206

82

25

2

2

37

56.7 22.6

6.9

0.6

1710–1780 (CLMETEV) 1780–1850 (CLMETEV) 1850–1920 (CLMETEV) CB times CB ukspok CB total

% n % n % n % n % n %

23 39

129 77

0

56

0

0.0 64.4

0.0

0

45

0.0 39.1 0

0.0 0

0.0 18.1

0.0

22

0.5 12.0

0.6 10.2

87

0 115

15

1

24

83

2 184 1.1 7 179 3.9 9 363 2.5

Total

159

Bounded descriptive modifier ‘complete in extent degree’ Ann Yearsley … died many years after, in a state of almost total seclusion, at Melksham. (CLMETEV Cottle, 1847, Reminiscences of Samuel Taylor Coleridge and Robert Southey)

Secondary determiner ‘reference to the entire entity’ The announcement brings the total number of jobs lost by Vernons this year to 200 and reduces the workforce to 425. (CB times)

Closed scale noun-intensifier ‘as great in extent, degree or amount as possible’ It is the nature of despotism ... to annihilate all intermediate situations between boundless strength on its own part, and total debility on the part of the people. (CLMETEV Burke, 1770, Thoughts on the present discontents)

Figure 15. The development of identifying and closed scale noun-intensifying uses of total

5.3. Whole The adjective whole is semantically similar to the adjectives complete and total discussed in the previous sections and has also developed both descriptive, identifying and noun-intensifying uses. However, unlike the Romance loanwords complete and total, whole is an adjective of Anglo-Saxon origin that has attested uses from the earliest Old English period onwards and it has, moreover, developed its noun-intensifying use along a pathway very different from those followed by complete and total. Rather than a shift from description to noun-intensification, whole has undergone a shift from quantitative identification to noun-intensification, a change which will be shown to be strongly driven by the adjective’s collocational and constructional preferences.

160

The completeness adjectives

5.3.1. Descriptive modifier uses of whole 5.3.1.1. Bounded descriptive modifier uses of whole In the data two clearly distinct types of bounded descriptive uses of whole are found, one of which has become obsolete in the contemporary data set. In Old and Middle English, whole was most often used to indicate that a certain body part of a man or animal is in good condition or uninjured (OED s.v. whole A.I.), as in (133). The near-synonyms skin and flesh are then by far the most common collocates of whole. In a number of cases the adjective refers not to body parts of persons or animals but to the beings themselves, describing them as healthy or in good condition (OED s.v. whole A.I.). Whole then functions as an antonym of ‘unhealthy’, ‘sick’, as is made explicit in (134). The health-related meaning of whole is especially clear in predicative position, where whole is often coordinated with the adjective sound whose prominent sense in Middle English was ‘unhurt, uninjured, unwounded’ (OED s.v. sound I.1.a), as in (135). (133)

(134)

(135)

Or ellus if þou wolt þou my3t schere a-wey þe fyke euene bi þe hole skyn & let þe euel blod blede out euerydel þen schalt þou strawe upon pouder of unsleked lyme or of chalke. (PPCME2 a1425, Late Middle English Treatise on Horses) ‘Or else if you want you might cut away the fig [= sarcoid] evenly along the whole skin and let the evil blood bleed out thoroughly, then you shall sprinkle upon it powder of unslacked lime or of chalk.’ Certes is not this lyke a myracle vnto a man that knowethe not, whye that swete thynges agree well to hole folke, and bytter thinges to sycke folk? (PPCEME Colville, 1556, Boethius’ Consolation of Philosophy) Tuelf fela3es wiþ him wente, Among hem aþulf þe gode, Min o3ene child, my leue fode. Ef horn child is hol and sund, And Aþulf biþute wund, He luueþ him so dere, And is him so stere, Mi3te iseon hem tueie, For ioie i scholde deie." (HC ?c1225 King Horn) ‘Twelve comrades went with him, among them Athulf the good, my own child, my loved son. If Childe Horn is whole and sound and Athulf is without wound, (he loves him so dearly and is so much a guide to him), if I could see the two of them, I would die of joy.’

Whole

161

The second bounded descriptive use of whole, which is attested from Middle English on, indicates that the NP referent is intact or not divided into parts, as in (136). Unlike the first bounded use, this use persists into Present Day English, but it is attested only rarely and, whereas it was first attested with a wide range of collocates, it is now largely restricted to nouns referring to foods, as in (137). Example (138) features this sense of whole in predicative position, confirming its descriptive modifier function. (136)

(137)

(138)

hudeð hare hale clað & doð on alre uueward. fiterokes al to torene (PPCME2 1225–30, Ancrene Riwle, II 244) ‘they hide their whole garment and put rags, all torn up, on top of everything.’ Almond brittle makes a lovely sweet. Ground to a powder it becomes praline, mainstay of patisserie and chocolate-making. Makes 200g

100g whole blanched almonds

100g caster sugar (CB times) If you cannot obtain a good supply of linseed oil , use the organically grown seeds as an alternative ... Ensure that the seeds you buy are whole and not split or lightly crushed. (WB brbooks)

Both the ‘healthy’ and ‘undivided’ uses of whole are clearly objectively descriptive and construed vis-à-vis a boundary. The properties conveyed by whole are objectively recognizable and definable, and entities are characterized as either displaying or as not displaying them, not as displaying them to a larger or lesser extent. Although the two uses are semantically quite distinct and have their own collocational and syntactic preferences, there is some overlap between them. A small number of examples in the data are vague between the two readings. In (139), for instance, the NP euery hole eye can be understood to refer to a healthy, sound eye as well as to an eye that is complete or intact, i.e. not missing any parts. The two readings are, of course, virtually interchangeable in this context. (139)

ye schal conceyue that in euery hole eye be thre partys: that ys, the qwyte or the balle off the eye; the cerkyl or the roundel off the eye, ... ; the thyrd parte off the ey ys the syte (HC Metham, c1450, Physiognomy)

162

The completeness adjectives

5.3.1.2. Bounded/unbounded descriptive modifier uses of whole Besides the two bounded uses discussed above, whole also has a more marginal subjective use which indicates that the entity designated by the NP is free from any imperfection or defect of quality. This reading is, however, only available when whole is coordinated with another subjective descriptive modifier, such as parfytte ‘perfect’ in (140). The subjective evaluation inherent in the adjective perfect is, in fact, transferred onto the adjective whole, which is normally associated with the more neutral, reading ‘no part or element wanting’ (OED s.v. whole II.6.a). In (140), the meaning of whole is very similar to that of complete in its subjective descriptive modifier use (§5.1.1.2) and similarly seems to hover between bounded and unbounded construal. (140)

For the nature of thynges toke neuer any begynnynge of thynges dymynished and vnparfitte but procedynge from hole and parfytte thynges, came downe or descended into these lower and baren thynges. (PPCEME Colville, 1556, Boethius’ Consolation of Philosophy)

5.3.2. Identifying uses of whole Like complete and total, whole has developed a secondary determiner use in which it typically aids the identification of the NP referent by expressing universal relative quantification and by reinforcing the boundedness construal of the head noun. Identifying whole indicates that reference is made to the entire entity referred to by the head noun, as in (141) and (142). Again, like the other two completeness adjectives studied, whole can be considered an emphasizing secondary determiner (§5.1.2 and §5.2.2), as it “seems to just emphasize what the semantic contribution of the NP already is, namely a term referring to a particular individual, perhaps emphasizing that reference not to a part, but to the whole has been made, putting into focus the whole of that individual” (Moltmann 2005: 603).

Whole

(141)

(142)

163

the share of the landlord seldom exceeds a third, sometimes not a fourth part of the whole produce of the land. (CLMETEV Smith, 1766, An inquiry into the nature and causes of the wealth of nations) Numerous though we of Middle England are, we are a minority of the whole population. (CB times)

Note that, as in examples (141) and (142), whole more frequently combines with mass nouns and collective nouns than the adjectives complete and total. With mass nouns, as in (141), the function of whole is restricted to stressing the inclusive reference conveyed by the primary determiner the (cf. §5.1.2 on secondary determiner complete). The quantifying-identifying semantics of whole are clear from its semantic preferences (Stubbs 2001: 88). As a secondary determiner, whole co-occurs with nouns that express a certain quantitative or measurable notion (see Table 14). From Middle English on, the nouns it combines with vary from indications of time – both direct (time, year) and more associative (sickness, dinner) –, over measures and dimensions (mile, length) to nouns referring to groups of people or animals. The latter can be straightforward collectives such as group or family, or they can be nouns designating buildings, such as house or school, which metonymically refer to the people that occupy them, e.g. (143) and (144). Roughly, secondary determiner whole can be “applied to everything of which there may be a part actually or in imagination” (Bäcklund 1973: 196) or, as Moltmann (2005: 609) argues, it “can only apply to entities that are integrated wholes (individuals or groups) and it maps such wholes to the sum of their actual parts”. (143) (144)

He said that antibiotics increased resistance by only 50 per cent and, in any case there were at first not enough tablets available for the whole school. (CB times) Canon Warren said evangelism was no longer seen as something done by a minister from a pulpit, but involved the whole church. (CB times)

Other

Buildings

Collectives

Indications of space

Indications of time

Measures and dimensions

8.3

%

1 8.3

n

%

%

n

%

n

1

n

trinity

kingdom

6 day, night, month, 50.0 year

n

%

33.3

%

diameter, motus

4

n

1350-1420

parish

year, hour

number

16.7

2

service, process

4 armies, host, suit, 33.3 legion

8.3

1

33.3

4

8.3

1

1420-1500

Table 14. Nouns co-occurring with secondary determiner whole in the PPCEME data set

37 intent, effects, 31.4 consent, thing

19

city, country, estate, land, parish, realm, 16.1 world 6 body, congregation, 5.1 army, herd 3 church, house 2.5

1500-1570 40 mile, multitude, peck, circle, line, 33.9 square, triangle 13 year, course, 11.0 day, night

164 The completeness adjectives

Whole

165

As for complete (§5.1.2) and total (§5.2.2), I argue that the secondary determiner use of whole developed from a bounded descriptive modifier use through deictification. More specifically, it is most likely the objectively descriptive sense ‘undivided’, ‘not missing any parts’, as in (136), that gave rise to the semantically strongly related secondary determiner reading of whole indicating that reference is made to the entire instance that speaker and hearer have in mind, as in (141). Secondary determiner whole expresses a type of relative quantification (Langacker 1991: 107ff): like a universal quantifier such as all, whole quantifies over the entire entity referred to. It stresses the coincidence of the NP referent with the whole reference set or mass. In the shift from description to identification the original objective reference point is replaced by a subjective, deictic reference point, i.e. the full instantiation of the NP referent. The descriptive origin of the secondary determiner whole is reflected in the quantified data. The diachronic rise in secondary determiner uses of whole is mirrored by a decrease in bounded descriptive modifier uses (see Table 16). 5.3.3. Noun-intensifying uses of whole Whole has developed two types of noun-intensifying uses, first a closed scale and later an open scale noun-intensifying use, illustrated in (145) and (146) respectively. (145) (146)

This whole impossible afternoon suddenly caved in on me, and my body broke down. (CB times) in Detroit downtown there's just not a whole lot of people there. It's just pretty ugly generally. (CB ukspok)

Unlike complete and total whose closed scale noun-intensifying uses originated in descriptive source meanings (§5.1.3.1 and §5.2.3), whole has developed this use from its quantifying-identifying secondary determiner meaning. Evidence for this shift from identification to noun-intensification is found in the occurrence of blended constructions in the data and in the similar semantic preferences of the two uses, i.e. for measure nouns in the broad sense. Moreover, for whole one type of specialized context (Traugott 2008, 2010c, 2010d) can be pinpointed that triggered the structuralsemantic shift to a closed scale noun-intensifying use, i.e. the all the whole N-construction. I will further argue that the open scale noun-intensifying use of whole did not develop independently from a descriptive or identify-

166

The completeness adjectives

ing source meaning, but arose as an extension of the closed scale nounintensifying use, through the reconfiguration of the quantification and intensification scales. I will claim that a shift has taken place from stressing the maximal extension of the NP referent by comparatively measuring the set or mass designated by the NP to a reference set or mass, viz. from relative quantification on a closed intensification scale, to intensifying the size or importance of a quantity conveyed by the NP, viz. absolute quantification on an open ended intensification scale. 5.3.3.1. Closed scale noun-intensifying uses The original noun-intensifying use of whole is typically used to emphatically stress the maximal extension of the NP referent, as in (147). Whole thus intensifies quantity on a closed scale, emphasizing that reference is made to the totality of the designated mass or entity/entities. In (147), for instance, whole stresses that the fame of Clemene resounded in each and every part of the country. In other words, noun-intensifying whole, like the secondary determiner (§5.3.2), expresses a type of relative quantification. However, instead of simply invoking, it now emphasizes the maximal extension of the reference mass by positing the NP referent at the ultimate point on a closed intensification scale. (147)

and that no man, of any nation, ever beheld her that did not fall in love with her; and that she had all the slaves perpetually at her feet; and the whole country resounded with the fame of Clemene, for so (said he) we have christen'd her: ... (PPCEME Behn, 1684, Love-Letters Between a Nobleman and His Sister)

As hinted at above, the closed scale noun-intensifying use of whole is closely related to the secondary determiner use. In fact, the data allow us to identify one particular type of specialized context (Traugott 2008, 2010c, 2010d) triggering the shift from identification in terms of quantity to intensification of quantity, viz. the construction all the whole N, as in (148) and (149). (148)

and yn especiall the ende of the seide Fyssh strete enchroched all the hole wey thurt over for a court place to the mancion of the Archideacon of Cornewaill as hyt apperyth openly (PPCME2 Shillingford, 1447-50, Letters and papers)

Whole

(149)

167

‘and especially in the end of the said Fish street [they] encroached all the whole way across for a court place to the mansion of the Archdeacon of Cornwall, as it appears openly’ Than sayd Abram vnto Lot: let there be no stryfe I praye the betwene the and me and betwene my herdmen and thyne for we be brethren. Ys not all the hole lande before the? Departe I praye the fro me. Yf thou wylt take the lefte hande I wyll take the right: Or yf thou take the right hande I wyll take the left. (CEMET Tyndale, 1526, Genesis 13, 8-9) ‘Then said Abram unto Lot: let there be no strife I pray between you and me and between my herdsmen and yours for we are brothers. Is not all the whole land before you? Depart I pray you from me. If you will take the left-hand side I will take the right: Or if you take the right side I will take the left.’

In the all the whole N-construction two quantitatively identifying secondary determiners, all and whole, are added to the primary determiner the. All and whole are semantically very similar elements that tend to be used in “parallel expressions” (Quirk et al. 1985: 259). Both elements reinforce the universal quantification inherent in the determiner and the boundedness of the singular count noun. When two such reinforcing elements are combined, the quantifying-identifying information is easily backgrounded in favour of the intensifying meaning aspect. The combination all the whole then functions to stress the size of the maximal extension of the NP referent. Later, whole came to express intensification by itself.35 The contextual emphasis that arises in the all the whole N-construction is semanticized into the adjective whole. As such, the development of noun-intensifying whole is a case of hypoanalysis. “In HYPOANALYSIS, the listener reanalyzes a contextual semantic/ functional property as an inherent property of the syntactic unit. In the reanalysis, the inherent property of the context ... is then attributed to the syntactic unit, and so the syntactic unit in question gains a new meaning or function” (Croft 2000: 126–127). The emphatic notion contextually conveyed by the all the whole N-construction is transferred onto the adjective whole by virtue of “its apparent redundancy with an overlapping ... expression”, i.e. all (Croft 2000: 130). 35. According to Buchstaller and Traugott (2006), all has been used as an intensifier of participles and adjectival heads since Old English, both as a closed scale intensifier or “top-of-scale modifier” and as an open scale intensifier or “scalar modifier”. The authors did not find instances of all as a noun-intensifier.

168

The completeness adjectives

The genetic link between the quantifying-identifying and the closed scale noun-intensifying uses of whole is supported by chronological, collocational, and structural evidence. First, the all the whole N-construction, which links the two uses, was particularly common in the Late and Modern English period, i.e. the period in which the noun-intensifying uses emerge in the data. Once the noun-intensifying use is firmly established in the language, the construction decreases in frequency and in Present Day English seems to have become obsolescent. Only one contemporary example, represented as (150), was found in the data set, not with the definite article but with a demonstrative functioning as primary determiner. Moreover, an open scale interpretation seems to be invoked here, not a closed scale one. Two contemporary closed scale noun-intensifying uses were found in the WB corpus, one of which is reproduced here as (151). The obsolescent character of the all the whole N-construction is consistent with its catalyst function. The all the whole N-construction merely served as a trigger for the development of the closed scale noun-intensifying reading of whole as an individual item and gradually becomes increasingly infrequent (cf. Diewald [2006: 5] on critical contexts, §3.2.1). (150) (151)

you get all this whole range of problems stuck in there (CB ukspok) But is she going to be out of action all the whole you know year? (WB brspok)

Second, there is considerable collocational overlap between the secondary determiner and closed scale noun-intensifying uses of whole. The head nouns with which the closed scale noun-intensifying use originally cooccurs are subsets of the range of nouns with which the determining use combines, namely measure nouns (e.g. litre) and nouns implying a certain quantity (see Tables 14 and 15). When whole is used as secondary determiner, the nouns are generally interpreted ‘literally’ as denoting clearly delineated and objectively measurable quantities (of time, of people, etc.) as in (152), whereas when whole is used as a closed scale quantity intensifier the (large) size implications of the nouns are foregrounded, as in (153). (152) (153)

Many insurers ... extend the cover for a month in any one year free of charge ... But few are keen to extend cover for a whole year. (CB times) If you’re really so smart, how come you spend years writing one single novel when, at any bookstore, you could hand over £ 6.99

Whole

169

and buy a novel that has already been written by someone else, thus saving two whole years of your own life (CB times) Interestingly, the all the whole N-construction also gave rise to a specific type of noun-intensifying use of whole which does not primarily occur with measure nouns or nouns denoting a quantity, but rather with abstract nouns referring to non-divisible entities, such as affair, idea, notion and thing. In these uses, illustrated in (154) and (155), whole conveys the strongly negative appreciation of the speaker toward the NP referent, which is typically described in the surrounding context. In (154), for instance, whole in the whole affair signals to the hearer that the event described in the preceding discourse is to be evaluated in a negative manner. In (155), the description of the NP referent is given in the following of-phrase. (156) is an example of an all the whole N-construction with a more abstract, nonmeasurable noun, which can plausibly be thought of to have served as the input for noun-intensifying uses as in (154) and (155). (154)

(155) (156)

A soldier of the guards … had seen the arrest, and heard the orders given to the coachman. This fellow, accidentally meeting Atkinson, had acquainted him with the whole affair. (CLMETEV Fielding, 1751, Amelia) He likes women who come on to him. He requires a sexual predator because he’s never asked a girl even to dance. The whole idea of being turned down is too distressing. (CB times) Thenne the Pope ressayvyd thes letters, and undyrstode alle the hoole processe, and made hys cardynallys to exampne the fryer (PPCME2 a1475, Gregory’s Chronicle)

Third, because of the similar semantics and collocational preferences, a number of instances of whole found in the data are vague between a secondary determiner and a closed scale noun-intensifying reading. In (157), for instance, both an identifying and an intensifying interpretation are plausible. Either the author may be understood to simply indicate the exact length of the rainy period or he may have added whole to the NP to emphasize the extent of the period of rain. (157)

It had been raining heavily for one whole month–raining on a camp of thirty thousand men and thousands of camels, elephants, horses, bullocks, and mules all gathered together at a place called Rawal Pindi (CLMETEV Kipling, 1894, The jungle book)

170

The completeness adjectives

5.3.3.2. Open scale noun-intensifying uses In the Late Modern English period whole has developed a second type of noun-intensifying use stressing not simply the maximal extension of the NP referent, but just how large or important it is. This use is most typically attested in contexts “where there is an implication of an unusually large quantity or number” (OED s.v. whole 7.c), as in (158) and (159). Unlike in the original noun-intensifying use, whole now emphasizes large quantity on an open intensification scale. The size of the NP referent is not compared to a particular reference mass and described as coinciding with it. Rather whole stresses the large quantity conveyed by the head noun it occurs with, activating a range on an open-ended scale of intensification. Whole consequently no longer expresses comparative, relative quantification but absolute quantification, giving a direct, non-comparative description of the remarkable size of the NP referent by placing it on an implied scale of magnitude. The semantics of this noun-intensifying use of whole is accordingly best paraphrased as ‘very many/much’ or ‘very great/big’. In (158), for instance, whole in whole years stresses the large number of years talked about and in the NP whole swarms in (159) it emphasizes the large number of Russians attending the feast. (158) (159)

One self-approving hour whole years outweighs Of stupid starers, and of loud huzzas (CLMETEV Pope, 1733–34, An essay on man) the German FLOH with his knife and fork, insatiate, not rising from table; whole swarms from all the Russias, and Asiatic hordes unnumbered - all these were there, and all rejoiced in one great international feast. (CLMETEV Kinglake, 1844, Eothen, or Traces of travel brought home from the East)

Unlike the open scale noun-intensifying uses of complete and total, the open scale noun-intensifying use of whole did not develop from a descriptive source meaning. Rather it developed from the closed scale nounintensifying use, driven by collocational extension. The absolute, open scale intensification conveyed by whole cannot be predicted on the basis of its inherent semantics, but developed through collocational extension and reclustering (§3.2.1). The shift from closed to open scale intensification, and from relative to absolute quantification accordingly, seems to have been triggered in plural NPs, as in (158) and (159), in which whole more explicitly conveys the large size of these sets (Ghesquière and Davidse 2011). Now, the open scale noun-intensifying use is most frequently attest-

Whole

171

ed in binominal constructions of the type a/Ø whole N1 of N2, in which N1 is a size noun36 in the broad sense and the entire construction serves to emphasize that there is a very large quantity of N2, as in (160) and (161) (see Table 15). Brems (2003: 229–230) already noted this use of whole in relation to the size noun bunch. She also observed that size nouns in general, such as swarms, trains, and lot, are prone to shifting from head to quantifier status. Trains in (160), for instance, displays more head characteristics than lot (of) and bunch (of) in (161) and (162) respectively, which have established quantifier uses. Sapir (1930: 23), however, considers even less entrenched combinations such as the whole swarm of and a whole pint of as “specialized totalizers”, “in which the fundamental quantitative notion is so limited as to apply only to a particular class of existents”. (160)

(161) (162)

Is it not enough that you are one of those whose passions made this cap, and force me through whole trains of years to wear it low upon my brow! (CLMETEV Dickens, 1843, A Christmas carol in prose) He is speaking up for a whole lot of people who feel that the American dream has passed them by. (CB times) There’s now a whole bunch of studies from different cities that show the same thing. (Brems 2003: 299)

The preceding paragraphs have shown that noun-intensifying whole both in its open scale and closed scale manifestations typically expresses a type of ‘quantitative degree modification’ rather than ‘qualitative degree modification’. It does not intensify evaluative notions, but modifies the extent of a notion of size or quantity conveyed by the head nouns and/or modifiers. The quantitative degree modification expressed by whole can pertain (i) to quantifiers, as in (161) and (162), (ii) to head nouns that explicitly express size or extent, like the size nouns in (158) and (159), and (iii) to nouns that more implicitly evoke a notion of size or amount, as in (160) (see also §7.2.1 on such). Interestingly, there seems to be a gradual diachronic collocational extension from type (iii) to type (ii) and (i). Size noun constructions and quantifiers are not attested with secondary determiner or closed-scale intensifying whole and are thus a clear extension of

36. Size nouns are nouns originally describing constellations of, or containers containing, things or stuff such as heaps, tons, classes, trains, lot (Langacker 1991: 88; Brems 2003).

Other

Size noun expressions

Collectives

Indications of space

Indications of time

Measures and dimensions

(peck (of N)) 1 25.0

n

%

% misliking

25.0

1

25.0

%

n

1

n

heart

(herd (of Ns))

herd (of Ns)

19.2

5

19.2

5

7.7

2

34.6

50.0

%

50.0

9 country, world

2

n

2

19.2

% town, land

peck (of N) 5

25.0

%

1570-1640

n

1

n

1500-1570

Table 15. Nouns co-occurring with noun-intensifying whole in the PPCEME corpus

burden, nose, pot

pack of Ns, train of Ns, cloud of Ns, nest of Ns, races of Ns

government, hierarchy

nation, world, earth, universe

days, years

1640-1710

172 The completeness adjectives

Whole

173

the original collocational sets as a result of the reconfiguration of the intensification scale from closed to open-ended. 5.3.4. Adjective-intensifying uses of whole Like for complete and total, a limited number of adjective-intensifying uses of whole was attested in the data, one late Late Modern English one (163), and some contemporary examples, e.g. (164). Peters (1994: 283, 285), however, mentions an adjective-intensifying use of whole, reproduced in (165), which dates back to the 18th century. According to the OED, the earliest instances even date from the 14th century. The question does arise whether we are dealing with the adjective or the adverb whole in these examples. Unlike for the other completeness adjectives studied the adverbial form of whole is not always formally distinct from the adjective as it can be realized as either whole or wholly. Peters (1994) and the OED label adjective-intensifying whole as an adverb and, as this is indeed a function typically realized by adverbs, an adverb analysis is perhaps more likely. Interestingly, all but one, illustrated in (164), of the adjective-intensifying uses of whole are found in combination with the adjective new. Not surprisingly, there is a strong, statistically significant, association between these two lemmas (T-score 18.614 and MI-score 4.511 in the CB times and CB ukspok corpus). (163)

(164) (165)

In a comparatively short time a whole new social class sprang up in the land, and a whole new public opinion. (CLMETEV Carpenter, 1915, The healing of nations and the hidden sources of their strife) it doesn't undermine what's gone before but it gives a whole different colour to the whole scene I think doesn't it. (CB ukspok) Hensius had pumpt and made you deny you had writ any such thing, and then pull'd out of his pocket a copy of your letter, wch you deny'd to be your hand, but you were whole confounded when he puU'd out an original letter of your own. (http://en.wikisource.org/wiki/Page:The_Wentworth_Papers_17151739.djvu/257)

174

The completeness adjectives

Bounded descriptive modifier ‘intact’, ‘not divided into parts’ Almond brittle makes a lovely sweet. Ground to a powder it becomes praline, mainstay of patisserie and chocolate-making. Makes 200g

100g whole blanched almonds

100g caster sugar (CB times)

Secondary determiner reference to the entire entity Numerous though we of Middle England are, we are a minority of the whole population. (CB times)

Specialized context : all the whole N-construction and yn especiall the ende of the seide Fyssh strete enchroched all the hole wey thurt over for a court place to the mancion of the Archideacon of Cornewaill as hyt apperyth openly (PPCME2 Shillingford, 1447-50, Letters and papers)

Closed scale noun-intensifier stress maximal extension of NP referent This whole impossible afternoon suddenly caved in on me, and my body broke down. (CB times)

Open scale noun-intensifier emphasize how large or important something is Is it not enough that you are one of those whose passions made this cap, and force me through whole trains of years to wear it low upon my brow! (CLMETEV Dickens, 1843, A Christmas carol in prose) Figure 16. The development of identifying and noun-intensifying uses of whole

Whole

175

1420–1500 (PPCME2) 1500–1570 (PPCEME) 1570–1640 (PPCEME) 1640–1710 (PPCEME) 1710–1780 (CLMETEV) 1780–1850 (CLMETEV) 1850–1920 (CLMETEV) CB times CB ukspok CB total

n % n % n % n % n % n % n % n % n % n % n % n

0 0.0 0 0.0 0 0.0 0 0.0 12 46.2 12 52.2 118 77.6 133 84.7 117 72.7 44 56.4 29 41.4 28

0 0.0 0 0.0 0 0.0 0 0.0 5 19.2 6 26.1 23 15.1 16 10.2 17 10.6 16 20.5 18 25.7 19

0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 4 2.6 4 2.5 26 16.1 17 21.8 23 32.9 27

0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1

1 50.0 0 0.0 0 0.0 2 33.3 0 0.0 0 0.0 1 0.7 2 1.3 0 0.0 0 0.0 0 0.0 0

1 50.0 0 0.0 7 100.0 4 66.7 9 34.6 5 21.7 6 3.9 2 1.3 1 0.6 1 1.3 0 0.0 1

% n

36.8 60

25.0 19

35.5 84

1.3 6

0.0 0

1.3 6

175

% n

34.3 62

10.9 21

48.0 91

3.4 3

0.0 0

3.4 2

179

%

34.6

11.7

50.8

1.7

0.0

1.1

n

122

40

175

9

0

8

%

34.5

11.3

49.4

2.5

0.0

2.3

total

bounded descr. mod.

1350–1420 (PPCME2)

unbounded descr. mod.

1150–1250 (PPCME2)

adjectiveintensifier

1050–1150 (YCOE)

nounintensifier

950–1050 (YCOE)

sec. det./ noun-int.

850–950 (YCOE)

secondary determiner

Table 16. Quantified results of the corpus study of whole

2 0 7 6 26 23 152 157 161 78 70 76

354

176

The completeness adjectives

5.3.5. From description to identification to noun-intensification The results of the data study on whole have shown that although whole followed the same path toward the development of secondary determiner uses as the other two completeness adjectives, i.e. deictification and (textual) intersubjectification from (bounded) description to identification, the path travelled toward noun-intensification was very different. Rather than developing its noun-intensifying uses from descriptive uses, whole shifted from identification to noun-intensification. The different shifts, visualized in Figure 16, are compatible with the quantified results of the data study (Table 16). 5.4. Conclusions In this chapter, I have shown that complete, total and whole developed both identifying and noun-intensifying uses. On the basis of diachronic data study it was argued that the secondary determiner uses of the completeness adjectives all developed through deictification of bounded descriptive modifier uses. Such a shift had already been posited for complete by Davidse, Breban, and Van linden (2008) and this study has shown that it is also valid for total and whole, which have similar quantifying lexical source meanings. All three completeness adjectives have thus undergone a semantic shift “which gives the general relation profiled by the adjective a reference point that is directly or indirectly related to the speech event” (Davidse, Breban, and Van linden 2008: 496–497). In other words, an objective reference point situated in the real world is transformed into a subjective reference point located in the speech event. More specifically, the reference point of secondary determiner complete, total and whole is always the full contextual realization of the entity referred to by the NP which speaker and hearer are aware of. For the three adjectives evidence supporting a deictification account was found in the overall decline of the descriptive uses and corresponding rise of determining uses in the corpus data. In addition, for complete, considerable collocational overlap was found between the descriptive and determining uses. For total and whole no such overlap was found between the descriptive and identifying uses. However, as argued in Sections 3.2.1 and 5.2.2, collocational overlap is of greater importance in the development of noun-intensifying uses which closely interact with the semantics of the head noun.

Conclusions

177

Contrary to the largely parallel development of the secondary determiner uses, the noun-intensifying uses of the completeness adjectives developed along clearly distinct pathways of change. The data studies have allowed us to identify at least two main pathways, one leading from description to noun-intensification and one leading from identification to noun-intensification. If one adheres to the views on subjectivity and intersubjectivity presented in this book, the first pathway constitutes a process of subjectification from objective or less subjective meanings to more subjective meanings whereas the latter pathway constitutes a shift from textually intersubjective meanings to subjective meanings, thus refuting a hypothesis of unidirectionality of (inter)subjectification. Both complete and total developed a closed scale noun-intensifying use from a bounded descriptive source meaning (e.g. from the complete isolation to a complete failure, and from total seclusion to a total defeat). In addition, complete developed a noun-intensifying use hovering between a closed and an open scale construal from a descriptive source meaning that displayed a similar indeterminacy in terms of boundedness construal (e.g. from a complete gentleman to a complete idiot). Both shifts confirm the general path of semantic change from description or property attribution to nounintensification posited by Paradis (2000) and Athanasiadou (2007), albeit that the specific descriptive source meanings claimed to give rise to the noun-intensifying readings did not correspond with the findings of this study. Moreover, the shifts from bounded description to closed scale intensification and from bounded/unbounded to closed/open scale intensification provide evidence for Paradis’s (2000: 235–239) claim that it is the bounded or unbounded nature of the descriptive source meanings which determines the type of noun-intensifying uses the adjective will develop. The gradability mode of the adjective in its descriptive use is simply made prominent in the noun-intensifying use, while its lexical meaning is backgrounded. It is plausible then that the bounded descriptive uses of complete and total give rise to closed scale noun-intensifying uses and that the indeterminacy in terms of boundedness of the descriptive ‘perfect’ sense persists in the later noun-intensifying uses that develop from it. Besides the shift from description to noun-intensification, the data studied also revealed a path from identification to noun-intensification, as predicted by Bolinger (1972). Whole was argued to have developed its closed scale noun-intensifying use from its secondary determiner use with the all the whole N- construction functioning as trigger context (e.g. the whole way > all the whole way > a whole way). The relative quantification scale inherent in the quantifyingidentifying semantics of whole is retained in the closed scale noun-

178

The completeness adjectives

intensifying use. In contrast, the later development of the absolute, openscale noun-intensifying use of whole, as in a whole lot of reading, is not predicted by the quantification mode of the source, but proved to be collocationally driven. The study of the completeness adjectives has shown that there is not one ultimate path to noun-intensification. Rather at least two distinct pathways of change have to be recognized, i.e. from description to nounintensification and from identification to noun-intensification. In contrast to what Lehrer (1985) hypothesized, the semantically related words complete, total and whole do not show parallel semantic changes (§4.1). The three adjectives have developed similar identifying and noun-intensifying meanings, but they proceeded along different pathways of change. Importantly, the observed changes – especially those toward noun-intensification – proved to be strongly collocationally restricted. Only with complete did the secondary determiner uses develop with a subset of the nouns with which its descriptive source meanings are attested. Moreover, influenced by the abstract meaning of the secondary determiner, the collocational restrictions on identifying complete are quickly loosened. In contrast, at the time of their development the noun-intensifying uses of complete, total and whole show considerable collocational overlap with their descriptive or identifying source meanings and continue to be restricted to specific sets of nouns with the required semantic properties, viz. degree nouns inviting open or closed scale intensification. The degree of collocational extension is then fairly limited.

Chapter 6 The specificity adjectives The results of the case studies of the completeness adjectives support the existence of two distinct pathways of semantic change towards nounintensifying adjective uses. On the one hand, complete and total were argued to have developed noun-intensifying uses from fully lexical descriptive uses (e.g. a complete gentleman > a complete idiot; total seclusion > total defeat), as predicted by Paradis (2000), Adamson (2000) and Athanasiadou (2007). On the other hand, for whole I have proposed a developmental path from the already grammaticalized quantifying secondary determiner use to the noun-intensifying use (e.g. the whole way > all the whole way > a whole way). This shift can be accommodated by Bolinger’s (1972) proposed pathway from identification to noun-intensification. This pathway, however, was envisaged to encompass semantic shifts not just from quantifying-identifying meaning to noun-intensifying meaning, as posited for whole, but also from purely identifying meaning to noun-intensifying meaning (§3.4.2.2). The difference between quantifying-identifying and purely identifying readings is illustrated for much and such in examples (166a) and (167a) respectively. Much in (166a) aids the identification of the NP referent by quantifying the mass referred to, whereas such in (167a) contributes to the identification by referring to something “of X identity” (Bolinger 1972: 60; §7.1). In (166b) and (167b) much and such both convey noun-intensification, measuring the degree of a quality invoked by the degree nouns miser and blunderer as “of X magnitude” (Bolinger 1972: 60). (166)

a. b.

(167)

a. b.

Too much of the time he misses the point (quantitative). (Bolinger 1972: 58) He’s too much of a miser to give anything (degree). (Bolinger 1972: 58) Such a person always frightens me. (Bolinger 1972: 60) Such a blunderer always frightens me. (Bolinger 1972: 60)

In this chapter I will argue that the specificity adjective particular has developed noun-intensifying meanings along the pathway from pure identification to noun-intensification. In the literature, particular, specific and

180

The specificity adjectives

certain are often discussed together as “adjectives of specificity”, as they are all three argued to render indefinite NPs specific (e.g. Enç 1991; von Heusinger 2007; Ionin 2010). An NP like a man is then viewed as referring to any possible instance of the type man, whereas a certain man is a specific NP, referring to one unique, yet indefinite instance. Although certain is more frequently used in English to establish specific indefinite identification, this chapter will focus on the adjectives particular and specific. Unlike certain, these adjectives can combine with other indefinite determiners than the indefinite article or zero-determiner (e.g. two particular aspects, some specific documents) as well as with definite determiners (e.g. these particular aspects, the specific documents you were looking for). Moreover, whereas particular and specific have very similar lexical source meanings, attributing the quality of specificity or detail to the NP referent (e.g. a very specific question), certain as a descriptive modifier has a truth-related meaning (e.g. A certain indication of a coming tempest [OED s.v. certain 2.a]). Unlike the completeness adjectives complete and total, the specificity adjectives are not traditionally recognized as typical noun-intensifiers. Peters (1994: 283, 285), however, does report on the adjective-intensifying use of particular, which was not attested in the data sets used for this study, but treats this use of particular – and of whole for that matter (§5.3.3) – as an instance of the zero-adverb form rather than of the adjective.37 As with the completeness adjectives, the adverbial counterparts of particular and specific have received more scholarly attention. Interestingly, as I will show for the adjectives, the adverbs particularly and specifically have also developed both what could be called an ‘identifying’ use and an intensifying use, as illustrated for particularly in (168) and (169) respectively. (168) (169)

Many people, and particularly teenagers, find it difficult to express their feelings verbally. (WB brbooks) He turned around suddenly, cheered up as if he'd just cracked a particularly difficult cryptic clue in his crossword. (WB brbooks)

37. Two examples of the intensifying use of particular with a predicative adjective were found in the WB corpus, both transcriptions of spoken language: (i) The UAE is particular attractive for Queensland firms because of the opportunities being thrown up by the country's ongoing and massive investment in tourism infrastructure. (WB oznews) (ii) There is one other issue that is particular sensitive in US-Syrian relations. (WB brspok)

Unbounded descriptive modifier uses

181

In the literature, particularly and specifically are in their identifying uses referred to as ‘particularizers’ (Jacobson 1964; Quirk et al. 1985: 604; König 1991: 97; Nevalainen 1991). Particularizers, which include such items as in particular, notably, primarily, etc., are a functional subclass of focusing subjuncts which “restrict the application of the utterance predominantly to the part focused” (Quirk et al. 1985: 604). To my knowledge, Francis (1965) is the only reference work which subsumes adjectives, including particular, certain and other, under the particularizer category. The particularizing function of particular and specific will be returned to in more detail in Section 6.2, following the discussion of the adjectives’ descriptive modifier functions (§6.1). In Section 6.3, the particularizing, identifying uses of the adjectives particular and specific will be argued to have given rise to open scale noun-intensifying uses, in a comparable manner to the development of the focusing adverbs. Because of their similar source and target meanings and largely parallel developmental histories, the corpus studies on particular and specific will be discussed together. 6.1. Unbounded descriptive modifier uses of particular and specific Unlike the completeness adjectives, which all three developed both bounded and unbounded descriptive uses, particular and specific can only convey unbounded descriptive meanings, both in the historical and the contemporary data. In their fully lexical, descriptive uses, particular and specific always denote a range on an open-ended scale. In other words, the properties they denote are inherently conceptualized as degrees and particular and specific consequently take modifiers that measure the actual degree of a quality on a scale going up or down from a reference point, e.g. very/highly/rather particular/specific (*half/almost specific/particular). As will be discussed in the remainder of this section, the unbounded, inherently scalar meanings conveyed by particular and specific do vary between more neutral, objectively verifiable and more subjective, evaluative meanings. 6.1.1. Objective descriptive modifier uses In their prototypical objective descriptive modifier use, particular and specific attribute the quality of detail or precision to the NP referents. In this

182

The specificity adjectives

use, the adjectives of specificity are near-synonyms of the adjectives exact, detailed or precise, as in (170) to (173). Although particular and specific convey roughly the same meaning in these uses, they display distinct collocational preferences. Particular has a clear preference for nouns such as description, story, and narrative, to which it then attributes the property of detail. Of this semantic set account is by far the noun with which particular co-occurs most often, as in (170). A second set of nouns with which particular combines with a certain degree of frequency are nouns referring to requests or questions. Particular then ascribes the quality of detail or precision to the solicited information, as in (171). Use of the comparative construction more particular in this example confirms the descriptive modifier status of the adjective. (170) (171)

I think fitt to add this particular accompt of what passed yesterday at Ruel betwixt Queen Mother and me. (PPCEME Rich, 1624, Letter CCCI Lord Kensington to the Duke of Buckingham) The right method had been first to consider the whole matter, and from so general a view to descend to more particular Enquiries: whereas they suffered their Minds to be forestalled with Prejudices; so that they never examined the matter impartially. (PPCEME Burnet, 1680, Some passages of the life and death of the Right Honourable John, Earl of Rochester)

Whereas particular is often used with nouns referring to the solicitation of information, specific frequently occurs with nouns such as instruction, direction, or recommendation, which solicit a certain action of the addressee. In (172), for instance, specific occurs in coordination with the nearly synonymous descriptive modifier minute. Both adjectives describe the rules in question as “precise or exact in respect of fulfilment, conditions or terms” (OED s.v. specific 4.a). Similarly, in (173), specific, modified by very, describes the instructions followed as detailed and precise. (172)

(173)

I know it would be fruitless to lay down any minute and specific rules for conversation, because none could be acted upon safely without strict reference to the object upon which they might be brought to bear. (CLMETEV Ellis, 1839, The women of England, their social duties, and domestic habits) This is cold, calculated murder for hire. He showed the jury that Perry had followed 27 very specific instructions in The Hitman almost ‘to a T’. (CB times)

Unbounded descriptive modifier uses

183

In the Late Modern English and Present Day English data sets, the range of nouns with which the objective descriptive modifier uses of particular and specific occur widens as the semantics of the adjectives becomes slightly more abstract, simply indicating that something is not or less general, indefinite or vague, as in (174) and (175). (174)

(175)

the increasing ties and daily developement of this institution renders it necessary, besides the general meeting which is destined for these halls, to have specific meetings for single branches of science. (CLMETEV Babbage, 1830, Reflections on the decline of science in England) There are two great rules of life, the one general and the other particular. The first is that every one can, in the end, get what he wants if he only tries. This is the general rule. The particular rule is that every individual is, more or less, an exception to the general rule. (CLMETEV Butler, 1912, Note-books)

In addition to the ‘exact, detailed’ meaning, a second objective descriptive meaning strain was found in the data for particular but not specific. In examples such as (176) and (177), particular indicates that something is private or personal, “concerning or known to an individual person or group of people and no other” (OED s.v. particular 4.a). The coordination of particular with the descriptive modifier outward in (176) confirms the descriptive modifier analysis. In Section 6.1.2, it will be argued that this sense gave rise to a more subjective reading in specific collocational contexts. (176)

(177)

I would not lecture on any subject for which I had to 'acquire' the main knowledge, even though a month's or three months' previous time were allowed me; on no subject that had not employed my thoughts for a large portion of my life since earliest manhood, free of all outward and particular purpose (CLMETEV Gillman, 1838, Life of Samuel Taylor Coleridge) it may be of some service to you to have general Renchild your friend: I once had the honour of a particular acquaintance with that great man, and I believe this letter ... will in part convince him of your merit (CLMETEV Haywood, 1744, The fortunate foundlings)

184

The specificity adjectives

6.1.2. Subjective descriptive modifier uses Like their typical objective descriptive meanings, the subjective descriptive meanings of the specificity adjectives are very similar. However, as Table 20 shows, only a small number of subjective descriptive uses of specific are found in the data (three Late Modern examples and five contemporary examples), which makes it hard to make any claims concerning the adjective’s collocational preferences in this function. In their typical subjective descriptive modifier uses, the adjectives of specificity indicate that the NP referent is peculiar, “worthy of notice, remarkable” or “special” (OED s.v. particular 5.a). In (178), the passages of the subject’s life are described as special, as out of the ordinary. Similarly, in (179), the two subjective descriptive modifiers particular and obliging both indicate that the return given was remarkable as it differed from the return normally given to Cavaliers. In (180), specific describes the social services’s approach as out of the ordinary. Use of the periphrastic comparative form in (178) and (179) and of the degree modifier very in (180) supports a descriptive modifier analysis of these examples. This subjective meaning is a natural extension of the typical objective descriptive modifier use of particular and specific. Something that is described as detailed or precise is, because of its property of elaborateness or minuteness, easily understood or reinterpreted as something special or remarkable. I have said nothing but what I had from his own mouth, and have avoided the mentioning of the more particular Passages of his life of which he told me not a few. (PPCEME Burnet, 1680, Some passages of the life and death of the Right Honourable John, Earl of Rochester) (179) Aurelian saw his Angel, his beautiful Incognita, and had no other way to make himself known to her, but by saluting and bowing to her after the Spanish mode; she guess'd him by it to be her new Servant Hippolito, and signified her apprehension, by making him a more particular and obliging return, than to any of the Cavaliers who had saluted her before. (CEMET Congreve, 1692, Incognita) (180) they are always interested in getting information about things you know and er So I think they they do have a different approach er obviously social services department have a very specific approach to the world. They have a language of their own. (CB ukspok)

(178)

Unbounded descriptive modifier uses

185

In addition to the ‘remarkable’, ‘peculiar’ sense, a second type of subjective descriptive meaning is attested in the data for particular. When modifying nouns such as acquaintance, friendship, and friend, as in (181), particular ascribes the quality of intimacy or closeness to the relation referred to. Such uses occur in the data from the Late Modern English period onwards and have most likely arisen as an extension of the objective descriptive meaning ‘personal’, ‘private’ (§6.1.1). (181)

She is a sweet little creature, and my very particular friend. (CLMETEV Craik, 1850, Olive)

6.2. Identifying uses of particular and specific In all the periods looked at, particular and specific are predominantly used as secondary determiners.38 In this use, particular and specific contribute to NP identification by instructing the hearer to relate the NP referent to other elements – explicitly or implicitly present in the discourse context – which are crucial to arrive at the referential identification intended by the speaker. On the basis of my analysis of the corpus data, I distinguish three main types of secondary determiner uses for the specificity adjectives, viz. a linking, an individuating, and a focusing secondary determiner use. In all these uses, identifying particular and specific contribute to the specificity of the NP identification. Enç (1991: 21) has defined (referential) specificity as “linking objects to the domain of discourse in some manner or other” and states that “one acceptable way of linking is ... by relating objects to familiar objects”. This is the type of specificity conveyed by the linking secondary determiner uses of particular and specific, in which they aid identification of the NP referent by pointing to the connection between the referent and a certain entity available in the surrounding context (§6.2.1). A second way in which the specificity adjectives contribute to NP identification is by setting up a cataphoric or anaphoric subset relation (cf. Enç 1991: 21). Such relations are conveyed by the individuating secondary determiner uses, which further identify the NP referent by invoking a part-whole rela38. An exception is the period 1640–1710 in which more classifier uses than secondary determiner uses are found for specific. There are however only 13 prenominal attestations of the adjective in total and 7 of the 10 attested classifiers stem from the same text. As a consequence, the quantified data may not be representative.

186

The specificity adjectives

tion to a contextually available reference set or mass (§6.2.2). Elaborating Enç’s (1991) study, I will argue in Section 6.2.3 that particular and specific can also contribute to the specificity of the NP by focusing on or singling out just one of a number of possible alternatives that are contextually available. The specificity adjectives then evoke not a part-whole relation but a part-part relation to other – implicit or explicit – discourse referents. Common to all of their determining uses is that particular and specific contribute to the identifying potential of the determiner complex by pointing out the necessary ‘referential anchoring’ (von Heusinger 2002, 2007) of the NP referent to another discourse referent. 6.2.1. Linking secondary determiner uses In their linking secondary determiner uses, particular and specific contribute to NP identification by pointing to the connection between the NP referent and one specific thing or person “as distinguished from another” (OED s.v. particular 3.a) which is retrievable from the discourse context. In other words, the specificity adjectives serve to further identify the NP referent by pointing to its link with a certain discourse referent and can be paraphrased by means of the adjectival phrases peculiar to or specific to. The discourse referents with which the NPs are connected are referred to by possessive pronouns preceding the adjectives, as in (183) and (185), or have to be retrieved cataphorically as they are made explicit in an of-phrase postmodifying the head noun, as in (182) and (184). The function of particular and specific is then to explicitly signal to the hearer that a referencepoint relation has to be set up (see Langacker 1993; Taylor 1996; Willemse 2005). A reference-point relation involves “the conception of one entity for purposes of establishing mental contact with another, i.e., to single it out for individual conscious awareness” (Langacker 1993: 5). In other words, to achieve full identification of the NP referent another entity – the referent of the possessive pronoun or phrase – has to be invoked first as a reference point. Speakers can use an adjective such as particular and specific not to set up reference-point relations themselves, but to signal to the hearer that he or she has to locate the entity referred to by the NP vis-à-vis another, more easily identifiable entity available in the discourse context. (182)

when we denominate him OLD, we mean that his duration is ran out almost to the end of that which men do not usually exceed. And so it is but comparing the particular age or duration of this or that

Identifying uses

(183) (184) (185)

187

man, to the idea of that duration which we have in our minds, as ordinarily belonging to that sort of animals (CEMET Locke, 1689, Essay concerning human understanding) And some affirm, that every plant has its particular fly or caterpillar, which it breeds and feeds. (CEMET Walton, 1675, Lives) The specific symptoms of poetic power elucidated in a critical analysis of Shakespeare's VENUS AND ADONIS, and RAPE of LUCRECE. (CLMETEV Coleridge, 1817, Biographia Literaria) If this, then, be "womanliness," can athletic games injure it? Do they spoil woman's usefulness as a woman? Do they damage her specific excellence? Do they tend to give her less endurance and nerve at critical times? I do not think so. Certainly lawn tennis does not. (CLMETEV Chambers, 1910, Lawn tennis for ladies)

In the above examples, the identifiability of the NP referent is assured by its connection to a certain individual or type of entity. In NPs with a possessive pronoun this connection was found to often be further stressed by the ‘emphatic’ determiner own (e.g. Quirk et al. 1985: 362–363; Saxon 1990; Huddleston and Pullum 2002: 482), intervening between the primary and secondary determiner, as in (186) and (187).39 The function of own consists of emphasizing that reference is to the discourse entities referred to by the preceding possessive as opposed to someone else (Huddleston and Pullum 2002: 482). The combination of possessive, own and particular “carries the force” of “[theirs] and nobody else’s” (Quirk et al. 1985: 362). (186)

(187)

My husband and I were with the Duke and Duchess de Medina de las Torres, in their own particular quarter in the palace, which we chose as the best place, and having the best view, whereupon we refused the balcony. (CEMET Fanshawe, 1676, Memoirs of Lady Fanshawe) Had every particular banking company always understood and attended to its own particular interest, the circulation never could have been overstocked with paper money. (CLMETEV Smith, 1766, An inquiry into the nature and causes of the wealth of nations)

39. Note that the insertion of own can favour a distributive reading, as in (187).

188

The specificity adjectives

6.2.2. Individuating secondary determiner uses In their second type of secondary determiner use, particular and specific aid NP identification by instructing the hearer to set up a contrastive relation between the NP referent and a reference set or mass. As such, reference is made to something “that is a unit or one among a number; taken or considered as an individual” (OED s.v. particular 2.a). In examples (188) to (191), particular and specific specify that reference is made to entities that are part of a larger whole or a larger number of entities. In (188), particular draws attention to individual men that together form a multitude. In (190) the NP no specific sum or portion refers to a portion of a larger sum of money available in the preceding discourse context. As such, particular and specific in their individuating secondary determiner uses set up a partwhole relation with a contextually available reference set or mass. In (188) and (190) this relation is set up anaphorically, while in (189) and (191) a cataphoric relation is construed. (188) (189)

(190)

(191)

And if this be Madnesse in the multitude, it is the same in every particular man (CEMET Hobbes, 1651, Leviathan) The Local Government Board is in a similar condition, and, although each particular Blue Book may be admirably indexed, there is no classified index of the whole series. (CLMETEV Booth, 1890, In darkest England and the way out) though there may be a considerable sum of bank money, for which there are no receipts, there is no specific sum or portion of it which may not at any time be demanded by one. (CLMETEV Smith, 1766, An inquiry into the nature and causes of the wealth of nations) He says it is too early to comment on specific changes, but he believes he has to improve its overall management. (CB times)

As illustrated by examples (188) and (189), particular in this use often occurs in NPs that are formally singular, but conceptually plural. Each particular Blue Book, for instance, although headed by a singular count noun, invokes a set consisting of a number of books. In these NPs, particular, often together with a universal relative quantifier, highlights that the NP refers to distinct, separate entities and comes to function as a type of nominal aspect marker, indicating that the NP referent is best thought of as a collective set consisting of separate individual entities rather than as a singleton, homogeneous set (Rijkhoff 2002: 101, 121). Breban (2010a: 182ff)

Identifying uses

189

has noted a similar use for secondary determiner different, illustrated in (192). In (188) and (192), the universal quantifier every aids to convey the idea of plurality and interpret the respective singular NPs every particular man and a different flavor of ‘Homestead-made’ ice cream as referring to multiple individual referents (cf. Breban 2010a: 182). In (189), it is the universal quantifier each which helps to signal the plurality of individual books. (192)

The Homestead bakery makes excellent breads and desserts. The menu also features a different flavor of ‘Homestead-made’ ice cream every night. (CB) (Breban 2010a: 182)

6.2.3. Focusing secondary determiner uses Whereas particular and specific in their individuating secondary determiner uses identify the NP referent by instructing the hearer to set up a part-whole relation, in their focusing determining uses they invite the hearer to identify the NP referent as distinguished from other entities on the same level of abstraction. More specifically, focusing particular and specific signal that a contrastive part-part relation has to be set up between the NP referent and a (set of) alternative values which are implicitly or explicitly available in the discourse context. In (193) and (194), for instance, particular and specific strengthen the pointing function of the demonstratives that and this respectively, singling out just one weekend and doctrine and not another one. In (194), the NP referent is explicitly contrasted with the referent of the demonstrative that. (193)

(194)

I arranged a weekend break for us ... When I made the booking I explained that the trip was for shopping, but the tickets arrived with a booklet listing that particular weekend as a public holiday in France. (CB times) But the good New Republican would have a wider scope for his Publishing Association than to subdue it to this specific doctrine or that. (CLMETEV Wells, 1902–3, Mankind in the making)

Importantly, secondary determiner particular and specific are found in the data both in definite and indefinite NPs. All of the examples given so far featured definite NPs. Examples of indefinite NPs are given in (195) and (196).

190

(195) (196)

The specificity adjectives

Often he would tell the dealer to collect his money from the stage door, so he felt he was dancing that evening for a particular painting or piece of furniture that he desperately wanted. (CB times) My ever wakeful reason, and the keenness of my moral feelings, will secure you from all unpleasant circumstances connected with me save only one, viz. the evasion of a specific madness. You will never ‘hear’ anything but the truth from me (CLMETEV Gillman, 1838, Life of Samuel Taylor Coleridge)

Secondary determiners, in combination with the type information provided by the head and the definiteness information provided by the determiner, typically provide the hearer with sufficient information to identify the exact instance or set of instances the speaker has in mind. This is, however, not the case in examples such as (195) and (196), which are best understood as conveying specific indefinite identification. In Cognitive Grammar, indefinite identification is understood as involving typeidentifiability (Gundel, Hedberg, and Zacharski 1993: 275). The identifying information provided by indefinite determiners is generally considered to be insufficient for the hearer to make mental contact with a unique instance of a type. Instead the hearer is required “to mentally ‘conjure up’ an instance as an instance of the type T” (Langacker 1991: 103f). In the case of specific indefinite identification, a specific instance is introduced into the discourse which is not assumed to be identifiable to the addressee. In contrast, the speaker generally does have a specific, fully identifiable instance in mind. This is reflected in the data in the frequent occurrence of phrases or clauses modifying the indefinite NP, as in (195) and (196), which add the necessary information to allow full identification of the intended NP referent by the addressee. In (195), for instance, the determiner complex a particular is phorically linked to the following relative that-clause, which specifies the identifying information provided by the determiner complex and further restricts the number of possible referents. The NP in example (195) evokes an implicit reference mass of many different instances of the types ‘painting’ and ‘piece of furniture’ which are all possible referents of the NP. Particular serves to point the hearer to the intended NP referent, but, due to its semantic generality, additional information is required to determine the exact referent. In this case, the that-clause postmodifying the head makes the description contained by the NP more specific. We are thus dealing here with a case of structural cataphora (Halliday and Hasan 1976: 56): a relation of forward reference is set up within one NP. In (196), on the other hand a cataphoric relationship is construed between the NP and the

Identifying uses

191

following sentence You will never ‘hear’ anything but the truth from me. Together, complex determiner and the phorically related sentence allow the hearer to single out just one instance from the set of possible instantiations of the type ‘madness’, viz. the inability to tell lies. The main contribution of indefinite determiner complexes with particular and specific to the identification of the NP referent then lies in preparing the addressee for further information provided by prepositional phrases, relative clauses and/or entire sentences. These add the necessary specifications to the NP and narrow down its descriptive scope. This is very similar to the situation in definite NPs. Again, the adjective signals the importance of other contextually available identifying devices or information in arriving at the intended referential identification. 6.2.4. The development of the secondary determiner uses As with the completeness adjectives (Ch.5), the grammaticalized secondary determiner uses of the specificity adjectives can plausibly be argued to have developed from fully lexical objective descriptive modifier uses through deictification. More specifically, for the specificity adjectives a change can be posited from the expression of qualitative specificity as descriptive modifiers to the expression of referential specificity as secondary determiners. The exact path of meaning change followed by particular and specific can be reconstructed as follows. In their prototypical objective descriptive modifier uses, particular and specific attribute the quality of detail to the NP referent, as in (170) to (173). When something is described as being detailed and exact, it is implicitly associated with its own special or even remarkable properties. This implicature is most prominent in subjective descriptive modifier uses, as in (178) to (180). It is but a small step then to viewing this special property as a distinguishing feature of the NP referent, marking it as distinct from other possible discourse referents. The development of the secondary determiner uses of particular and specific, which mark the NP referent as distinguished from others, can thus be viewed as the result of the pragmatic strengthening of implicatures. The semantic closeness between the descriptive modifier and secondary determiner uses of particular and specific is illustrated in examples (197) to (200), which serve as bridging contexts (Evans and Wilkins 2000: 550; §3.2) between a descriptive modifier and secondary determiner reading. Although associated with distinct structural parsings of the NP syntagm, both readings are plausible on the basis of different contextual clues. This

192

The specificity adjectives

type of contexts is attested for all three types of secondary determiner uses of the specificity adjectives found in the data. First, in (197) and (198), particular and specific can on the one hand be interpreted as descriptive modifiers attributing the quality of detail to the orders and commands respectively. On the other, they can also be interpreted as linking secondary determiners, indicating that the NP referents are to be distinguished from others on the basis of their connection to certain discourse referents, i.e. ladies in (197) and Excellency in (198). Second, example (199) allows both a descriptive modifier reading, attributing the quality of detail or specificity to the NP referent, and an individuating secondary determiner reading in which particular serves to distinguish between individual, distinct cases. The former, descriptive reading is warranted by the opposition with the descriptive modifier general. Only in this reading could particular be used in its comparative form (more particular cases). Third, in (200), specific can be interpreted either as a subjective descriptive modifier meaning ‘special, remarkable’, or as a focusing secondary determiner. The NP then refers to a certain type of levity which is further specified in the following wh-clause. (197) (198) (199)

(200)

Your ladies were so slow in giving their specific orders, that the mohairs, of which you at last sent me the patterns, were all sold. (CLMETEV Chesterfield, 1751, Letter to his son) I am running on so fast to give your Excelcie account of things here that I had almost forgot your particular commands. (PPCEME Conway, 1675, Letter to Arthur Capel) For there are so many general propositions concerning conscience, the nature and obligation of it, explained and proved, with such firm consequence and evidence of reason, that he who reads, remembers, and can with prudence pertinently apply them _hic et nunc_ to particular cases, may, by their light and help, rationally resolve a thousand particular doubts and scruples of conscience. (CEMET Walton, 1675, Lives) if their little work has hitherto floated upon the stream of time, while so many others of much greater weight and value have sunk to rise no more, it has been solely indebted for its buoyancy to that specific levity which enables feathers, straws, and similar trifles to defer their submersion until they have become thoroughly saturated with the waters of oblivion, when they quickly meet the fate which they had long before merited. (CLMETEV Smith and Smith, 1812, Rejected addresses)

Focusing secondary determiner a booklet listing that particular weekend as a public holiday in France. (CB times)

Individuating sec. determiner And if this be Madnesse in the multitude, it is the same in every particular man (CEMET Hobbes,

Linking secondary determiner every plant has its particular fly or caterpill, which it breeds and feeds. (CEMET Walton, 1675)

Figure 17. The development of identifying uses of particular

Bridging context that specific levity which enables feathers, straws, and similar trifles to defer their submersion (CLMETEV Smith & Smith, 1812)

Bridging context For there are so many general propositions concerning consience, … with prudence pertinently apply them ... to particular cases (CEMET Walton, 1675)

Bridging context I am running on so fast to give your Excelcie account of things here that I had almost forgot your particular commands. (PPCEME Conway, 1675)

Subjective descriptive modifier ’worthy of notice, remarkable, special’ making him a more particular and obliging return, than to any of the Cavaliers who had saluted her before.(CEMET Congreve, 1692)

Objective descriptive modifier ’exact, detailed, precise’ this particular accompt of what passed yesterday at Ruel (PPCEME Rich, 1624)

193 The specificity adjectives

194

The specificity adjectives

Ambiguity as it is found in the above examples, of course, may only exist in retrospect, but it does point to the close semantic ties and possible genetic link between the descriptive and identifying uses (§3.2). Unfortunately, the corpus study has not provided chronological evidence for the pathway from description to identification, as secondary determiner uses appear in the data at the same period as the descriptive modifier uses and are moreover predominant throughout all historical stages. Collocational evidence is also hard to adduce. Because of their high semantic schematicity, the range of nouns with which the secondary determiners of specificity co-occur is very wide. Unlike with complete (§5.1.2), the typical collocates of determining particular and specific do not form subsets of the collocates of the descriptive uses. Despite the lack of tangible evidence, a deictification account whereby the secondary determiner uses originate in descriptive modifier uses is intuitively plausible and compatible with existing deictification and grammaticalization hypotheses. The envisaged developmental path from description to identification can be schematically represented as in Figure 17. 6.3. Noun-intensifying uses of particular In the data sets studied, specific was found to function as a full-fledged noun-intensifier in only two instances. One example is (201) in which the NP specific liability refers to the very high risk of skull fracture run by certain species. Of course, with only two attestations, it is not possible to posit any generalizations concerning the general semantics or collocational behaviour of noun-intensifying specific. (201)

A time would no doubt come when those with a specific liability to skull fracture would all be eliminated, and the human cranium would have developed a practical immunity to damage from all sorts of falling substances. (CLMETEV Wells, 1902–1903, Mankind in the making)

In contrast, noun-intensifying uses of particular are more frequent (taking up 12.8% of all prenominal instances in the Present Day English sample) and appear in the data from the period 1570–1640 onwards. As a consequence, the discussion of the noun-intensifying uses of the specificity adjectives will mainly be concerned with particular. It will be argued that the latter adjective has developed two distinct types of noun-intensifying use –

Noun-intensifying uses of particular

195

one more qualitative (202) and one more quantitative (203) – each associated with its own collocational preferences. Note that particular always locates the degree of a gradable property associated with the NP high on an open-ended scale of intensification. In (202) the adjective places the intensity or strength with which the desire is felt high on an open ended intensification scale. In (203) particular measures the large amount of concern caused by the damaged state of Southside House. (202) (203)

My husband now returned from church … and was, by my Lord’s particular desire, ushered into the room. (CLMETEV Fielding, 1751, Amelia) Thereafter he spent his life running the various family properties ... A particular concern was Southside House which had been damaged during the war. (CB times)

6.3.1. The development of noun-intensifying particular As for whole (§5.2.3), it can be plausibly argued that the noun-intensifying uses of particular have developed from the adjective’s secondary determiner uses. Unlike for whole, however, these determining uses do not have any quantifying import but are strictly identifying. As such, the development of particular instantiates the pathway predicted by Bolinger (1972) from purely identifying to noun-intensifying meaning. The data allow us to reconstruct two distinct pathways of change towards noun-intensifying particular, one where the onset context is an emphatic linking construction and one where the shift towards nounintensification is enabled by a negative quantifying element. Common denominator in these two contexts is the presence of a comparative effect (Hendrik De Smet, p.c.). The emphatic linking constructions invoke scalar comparison of different entities possibly functioning as reference point for NP identification. With the negative quantifying construction, comparison is between possible referents of the NP with particular itself. The specialized contexts triggering the development of the noun-intensifying uses of particular are thus both dialogic in nature (Traugott 2008, 2010c, 2010d; §3.2.1).

196

The specificity adjectives

6.3.1.1. The emphatic linking construction Secondary determiners typically do not allow degree modification (§1.4.2). Accordingly, linking secondary determiner particular, which points to the connection between the NP referent and another discourse referent, cannot be modified directly in terms of degree (*his very/almost particular age). The speaker can, however, further emphasize the connection between the NP referent and another referent by inserting the emphatic determiner own between primary and secondary determiner (§6.2.1). I argue that such specialized emphatic constructions, as in (204) and (205) have served as onset contexts for the shift from linking identifying meaning to noun-intensifying meaning. (204) (205)

no consideration whatever should make me run from the engagement which I had contracted with him, at my own particular request. (CLMETEV Hunt, 1820-22, Memoirs of Henry Hunt) the hero, as it is called, of the poem, was that well known Hercules of the Persians, named Rustem; although there are several other heroes, or warriors, to each of whom their own particular glory is assigned. (CLMETEV Cary, 1846, Lives of the English poets)

Particular in (204), for instance, can receive a determining, distinctive reading in which it instructs the hearer to identify the request as made by the speaker and not by any other party. The identifying, reference-point function can be adequately fulfilled by the possessive determiner my, but is reinforced here by the emphatic determiner own. I claim that it is exactly the emphatic nature of this possessive construction which invites or triggers a possible noun-intensifying reading of particular. Emphasis on the unique relation between the referent of the pronoun my and the referent of request invites an interpretation of the request as very special or unique itself, in this case because of its strength or expressness. Rather than simply linking the NP to a specific discourse referent, the determiner complex my own particular then indicates that the request in question was made with great emphasis, leaving little room for the addressee to refuse it. Later, the emphatic, noun-intensifying meaning is semanticized into the adjective particular itself. Sentences (206) and (207) exemplify fully-fledged noun-intensifying uses of particular. In (206) identification of the NP referent is facilitated through its connection to the speaker, i.e. the source of the desire. However, this identifying function of particular, pointing to this reference-point rela-

Noun-intensifying uses of particular

197

tion, is backgrounded in favour of a noun-intensifying reading, locating the strength with which the desire is felt high on an open-ended scale. Similarly, in (207) the scorn referred to is identified as linked to the person being quoted, but focus is on the high degree of derision involved rather than on identifying the NP in terms of its connection to the discourse context. (206) (207)

Besides, if you must know it, I have a particular desire to capture the scoundrel myself. (CLMETEV Bennett, 1902, The grand Babylon Hotel) `There are too many actors today living passionless lives, who are passionless on stage. They are not getting out enough," he said from his suite at the Savoy.

He saved his particular scorn for `cerebral" actors such as Sir Derek Jacobi and Sir Ian mckellen, accusing them of leading undaring lives that resulted in poor art. (CB times)

As was the case with the path from quantitative identification to nounintensification, the path from pure identification to noun-intensification has not featured prominently in the literature.40 However, Nevalainen (1991, 1994) and Traugott (2006) have studied the semantically similar path from ‘particularizer’ to scalar intensifier for adverbs such as just, particularly, exactly, barely and even. (208) and (209) illustrate the particularizer and scalar intensifier uses of even respectively. (208)

(209)

[The disciples speak] Is not this he that sate and begged? Some sayde: this is he. Other sayd: he is lyke him. But he him selfe sayde: I am even he. ‘Is not this the man that sat and begged? Some said: This is he. Others said: He is like him. But he himself said, I am indeed he’ (1534 Tyndale, New Testament, IX, i [Helsinki]) (Traugott 2006: 349) It is a lamentable case to see how the deuill has bewitched thousands at this day to run after him: and euen to offer sacrifice vnto him. (1593 Witches, B2R [Helsinki]) (Traugott 2006: 349)

40. Vandewinkel and Davidse (2008) discuss the relation between focusing and noun-intensifying pure, but both these meanings seem to have developed at about the same time from descriptive pure.

198

The specificity adjectives

Particularizing adverbs are a subtype of focus adverbs (Nevalainen 1994: 254), i.e. adverbs used to “draw attention to a part of a sentence as wide as the predication or as narrow as a single constituent of an element” (Quirk et al. 1985: 604).41 Particularizers, more specifically, “restrict the application of the utterance predominantly to the part focused” (Quirk et al. 1985: 604) [italics original], as in (210) and (211). (210) (211)

Especially handicapped people have great difficulties in finding a job. (König 1991: 97) This table, in particular, is made of wood. (König 1991: 97)

Examples (208), (210) and (211) illustrate the scalar potential of particularizers. While the adverbs are not, in this use, inherently scalar, they can evoke pragmatic scales (see Traugott 2006: 341). Of course, in their particularizing function the adverbs heighten the “degree of precision” of identification (Nevalainen 1991: 58) and in this way convey scalarity. More importantly, particularizing adverbs are pragmatically scalar in that they evoke a scale of values, including the focus value and other alternative values. In so far as they “induce an ordering for the focus value” and the alternative values, particularizers entail comparison (König 1991: 97; Nevalainen 1991: 59). As argued by König (1991: 97), “[t]he focus value is characterised as a clear case for the predication expressed by the rest of the sentence. The alternatives under consideration do not manifest the relevant property as clearly. The ordering associated with particularisers can thus roughly be expressed by a comparative statement”. In (210), especially indicates that handicapped people, i.e. the focus value, have greater difficulties in finding a job than for instance older people, i.e. an alternative value. In (211), it is the table and not the chairs or cupboard that is made of wood. According to König (1991: 97), it is the inherent comparative nature of particularizers such as particularly and specifically that has led them to develop intensifying meanings. He illustrates the difference between the two uses by means of the following examples: (212) (213)

Especially John is intelligent. John is especially intelligent.

41. Focus adverbs have also been labelled ‘focus modifiers’ (Huddleston and Pullum 2002; Traugott 2006), ‘focus particles’ (König 1991) and ‘focusing subjuncts’ (Quirk et al. 1985).

Noun-intensifying uses of particular

199

In (212) especially simply helps to single out John among a number of other possible referents as the intended subject. In (213) the identificational comparison has led to scalar comparison. Especially no longer aids the identification of the NP referent by singling out John as the most intelligent among a number of other possible referents. Rather, especially functions as a “comparative intensifier” (Lorenz 2002: 150), heightening the degree of intelligence associated with John. More specifically, especially “achieves intensification by comparing the referent with its rivals or equals”. Calling John especially intelligent implies that the speaker finds him very intelligent and more so than other people (cf. Lorenz 2002: 150). The pragmatic process of evoking alternative agents has thus served as a basis for deriving scalar implicatures (Eckardt 2012). In contexts where there is explicit comparison of the focus and alternative values, as in (214), the identificational comparison is most easily reinterpreted in terms of scalar comparison (Nevalainen 1991: 59). (214)

If you do want to use an ellipsis, I suggest that you render it correctly, which means using exactly three dots, not two or four or twelve (http://barelybad.com/bangslashsmiley.htm, accessed 2 September 2001)

Rather than inherent in the source meaning of the intensifying element, scalar notions are in the case of focus modifiers derived purely pragmatically in specialized identifying contexts (cf. Traugott 2008, 2010c, 2010d) involving emphasis and the evoking of alternatives. The adjective particular has undergone a comparable process of change. The linking secondary determiner use of particular contains the same basic semantic components as the particularizing adverbs: it adds a notion of greater precision to the identification expressed by the NP and at the same time implies comparison. For instance, in (182) and (183), the age and insects were further identified through their connection with specific discourse referents. Both this greater degree of precision and the implied comparison carry in them the possibility of a meaning shift to intensifying scalarity. As noted above, the use of emphatic own, as in (204) and (205), can further reinforce these two elements, and in these specialized contexts (Traugott 2008, 2010c, 2010d) the new noun-intensifier meaning, which later became semanticized in particular, emerged. In (206) and (207), noun-intensifying particular evokes alternatives that are ordered on a scale. In (206), the use of particular in a particular desire implies that the first person subject also experiences other desires, none of which, however, are as strong as the desire ‘to

200

The specificity adjectives

capture the scoundrel myself’. Similarly, in (207) the NP his particular scorn indicates the presence of alternative values, all placed lower on the evoked intensification scale than that conveyed by the NP. Nounintensifying particular thus creates a downward-entailing effect (Eckardt 2012), situating the focus value higher on the implied scale than any alternative values. 6.3.1.2. The negative quantifying construction Besides the emphatic linking construction there is a second identifying use which serves as a locus of change for the shift toward noun-intensifying uses of particular. Here, the factor promoting a noun-intensifying reading is the presence of a negative quantifying element, such as any or no, which in these NPs with particular and a gradable head noun creates a kind of hedging effect, as in (215) and (216). In these specialized dialogic contexts (Traugott 2008, 2010c, 2010d), the speaker denies the existence of a special or distinctive necessity or hurry respectively, thereby avoiding a potential face-threat. By adding particular to the NP, the speaker safeguards himself by not categorically denying the existence of a certain necessity or hurry, which is more diplomatic and in (216) avoids making too explicit or direct a statement concerning Mr Boarham. At the same time, use of the negated NP with particular avoids being too imposing toward the hearer. (215)

(216)

It contains a general principle, uncontested, and established; a principle which this assembly has never denied, and from which I know not that it has ever departed. As there is, therefore, no particular necessity of confirming it by a new resolution, and as the present time seems less proper than any other, I cannot but declare my opinion, that to resume it at some other time will be more prudent, than to give the lords, who think their conduct censured, any occasion of resentment or discontent. (CLMETEV Johnson, 1740– 1741, Parliamentary debates) Mr Boarham is in no particular hurry, for he has little doubt of your acceptance; and I want to speak with you. (CLMETEV Brontë, 1848, The tenant of Wildfell Hall)

The above examples pragmatically invoke comparison with alternative types of necessity or hurry and feature negation. Both aspects – the invoking of alternatives and negation – are characterized by Traugott (2010d: 15)

Noun-intensifying uses of particular

201

as indexing dialogicity. The two types of specialized contexts triggering the development of noun-intensifying particular – the emphatic linking construction and the negative identifying construction – are then dialogic in nature. In both types it is the comparative, inherently scalar implication which has allowed the rise of truly noun-intensifying uses placing the degree of a quality inherent in the head high on an open-ended scale. The implicature leading to the noun-intensifying meaning can be summarized as ‘absence of a special or unique X’ > ‘absence of a strong/big/... X’ (Hendrik De Smet, p.c.). Later, the negation can be dropped and only the nounintensifying aspect remains. In example (217), for instance, particular functions as a noun-intensifier, emphasizing the great difficulty involved in the paper. (217)

Less than half of 1 per cent passed this test, and those who did attempt the paper found particular difficulty in the ordering of fractions, and the calculation of ratios and percentage increases. (CB times)

6.3.2. Collocational evidence for the shift from identification to noun-intensification As with the completeness adjectives, the development of noun-intensifying particular is strongly collocationally driven. Table 17 gives an overview of the main semantic sets of nouns with which noun-intensifying particular combines. The nouns with which noun-intensifying particular typically combines form a subset of the nouns with which the determining uses occur, i.e. nouns referring to concepts strongly linked to an experiencing human subject. More specifically, noun-intensifying particular frequently combines with nouns from three semantic sets, viz. nouns referring to emotions (e.g. grace, affection, kindness), requests (e.g. desire, request, wish), and observations (e.g. notice, interest, attention). The former two are mainly found with noun-intensifying uses that arose from the emphatic linking construction, the latter with those that stem from the negative quantifying construction.

Other

Observation

Request

Emotion

20

1

20

n

%

force

57

4

n

%

14

1

care, notice

desire

2 kindness, 29 respect

1640–1710

%

care

n

1

1

%

n

1570–1640 3 grace, affection, 60 fondness

attention, study

desire

7 excellence, favour, 58 genius

17

2

8

1

2 affection, 17 satisfaction

1710–1780

Table 17. Nouns co-occurring with noun-intensifying particular 1780–1850 4 fondness, pleasure, 12 spite 6 desire, request, 18 wish attention, 10 notice, care, 29 interest 14 objection, strength, 41 calamity

2 favourite, 12 importance

7

notice, attention, care, 41 interest

2 desire, 12 request

6 pleasure, 35 charm

1850–1920

27 progress, 46 blows, risk

attention, concern, interest, 36 awareness 21

11 scorn, awe, 19 sympathy

CB

202 The specificity adjectives

Noun-intensifying uses of particular

203

Interestingly, up to 1780, noun-intensifying particular has a predominantly positive semantic prosody (Stubbs 1995: 25). Only later does the adjective in this use occur in combination with nouns with a negative semantic flavour, such as dread, objection and anxiety. In Present Day English, noun-intensifying particular displays divergent collocational behaviour depending on the mode of language. In the written register it collocates mainly with positive head nouns (interest, triumph, shine), whereas it predominantly combines with negative collocates in the spoken data set (difficulty, downer, dislike). This observation might be indicative of an ongoing change in the semantic prosody of noun-intensifying particular. 6.4. Classifying uses of specific In the preceding section, I have argued that only particular has a wellestablished noun-intensifying use. In contrast, the data studied provide evidence of an established classifier use for specific only. Classifying modifiers semantically subclassify the general type denoted by the head noun, e.g. electric in electric trains (Halliday 1994: 184–186; §1.2.2). Just one such use of particular was found in the data sample, represented here as (218). In this example, particular is used in reference to persons not holding a religious position and is thus semantically similar to adjectives such as secular or private. (218)

… the Scripture plainly allows this liberty to particular and private Persons to judg for themselves … I do not think this is spoken only to the Pope or a general Council, but to Christians in general. (PPCEME Tillotson, a1679, Of the tryall of the spirits)

In contrast, a considerable number of entrenched classifier-noun combinations with specific were found in the data, both diachronic and synchronic. Some early examples are given below. (219)

(220)

… Whether by the gravity of the Atmospherical Air, surmounting the Specifick Gravity of the little and rarifi’d Atmosphere … (PPCEME Boyle, 1675-76, Experiments and notes about the mechanical origine or production of electricity) In the year 1788, M. Gioeni … published at Naples an account of a new family of Testacea, of which he described, with great minuteness, one species, the specific name of which has been taken from

204

(221)

The specificity adjectives

its habitat, and the generic he took from his own family, calling it Gioenia Sicula. (CLMETEV Babbage, 1830, Reflections on the decline of science in England) What specifick difference can be produc'd between a Knave, a Coward, and a Traytor. (CEMET Defoe, 1705, The Consolidator)

The unit Specifick Gravity, as in (219), refers to “the degree of relative heaviness characteristic of any kind or portion of matter; commonly expressed by the ratio of the weight of a given volume to that of an equal volume of some substance taken as a standard” (OED s.v. specific 2.d). In (220), specific name refers to “the second (adjectival) element in the Latin name of a species according to the binomial system, which follows the generic name and serves to distinguish a species from others in the same genus” (OED s.v. specific 5). Similarly, in (221), the NP what specifick difference refers to “a quality, mark, or characteristic, that distinguishes” a knave, a coward and a traitor from one another. In other words, it refers to a specific property “by which a species is distinguished from other species of the same genus” (OED s.v. difference 4.c). Classifier-noun sequences develop when a certain concept appears in a particular cultural context and consequently needs to be named. Most likely, the classifier-noun sequences with specific are coinages of specialized language use. Specific gravity and specific heat, for instance, are physics terms. Specific name is a term used in biology and, more specifically, zoology. Most of the classifier-noun sequences found in the data are also entrenched in other languages such as French, e.g. gravité spécifique, nom spécifique (TLFi). A defining characteristic of classifiers is that they name a subset of the head noun, not as a “self-contained, ‘self-defining’ concept”, but in contrast with other contextually relevant subsets (Vandelanotte 2002: 235). In (220), for instance, the opposition between specific name and generic name is explicitly mentioned. Classifier-noun sequences are thus bound to culturally entrenched taxonomies and consequently more likely to be accessed as entire units. Unlike descriptive modification and secondary determination, they are not the result of a particular choice of the speaker related to just one instance of language use. Rather classifier-noun combinations are stored independently as a separate word form (Langacker 1987a: 59–60). The compositionality of the distinct components lessens, as the two elements are fused together to form one lexical unit that is stored in the lexicon as one contentful item that must be learned by language users. Support for the entrenched (Langacker 1987a; Haiman 1994; Bybee 2003) nature of

n

3

2

2

2

1

noun

essences

gravity

name(s)

idea(s)

difference

1640-1710

10.0

20.0

20.0

20.0

30.0

%

identity

noun 2

n 100.0

%

1710-1780

1 1 1

description skill

1

2

2

9

n

purposes

name

heat(s)

difference(s)

gravity

noun

1780-1850

Table 18. Nouns co-occurring with classifying specific

5.9

5.9

5.9

5.6

11.8

11.8

52.9

%

immunity

weight

centres

differences

gravity

noun

1

1

2

2

4

n

1850-1920

9.1

9.1

18.2

18.2

36.4

%

gravity

taxes

performance

noun

CB

2

5

13

n

10.0

25.0

65.0

%

Classifying uses of specific 205

206

The specificity adjectives

the classifier-noun combinations with specific is found in the fact that most of them are listed in dictionaries such as the OED, accompanied by their distinct meaning (e.g. specific gravity: s.v. specific 2.d, s.v. gravity 4.c; specific difference: s.v. specific 2.b, s.v. difference 4.c; specific name: s.v. specific 5). Table 18 shows that, as time proceeds, different classifier-noun combinations crop up or become predominant in the data. However, some combinations, such as specific gravity, are found from the earliest to the most recent data sets, which corroborates their degree of entrenchment in English. 6.5. From description to identification (to noun-intensification) First, as for the three completeness adjectives, I have argued that the secondary determiner uses of particular have developed from descriptive modifier uses through deictification, i.e. from expressing qualitative specificity to expressing referential specificity. As shown in Table 19, such a shift is in line with the quantified results of the data study. The proportion of objective and subjective descriptive modifier uses of particular gradually decreases over time (from 22.0% and 6.4% respectively in the period 1570– 1640 to 0.6% in the contemporary data set). The secondary determiner uses of particular predominate already in the period 1570–1640 (54.6%), but their proportion continues to rise, reaching 82.2% in the Present Day English data sets.42 In the period 1780–1850, however, there is a momentary drop in the proportion of secondary determiner uses of particular in the data set. This drop occurs at the same time as the rise of the noun-intensifying uses. As such, the quantitative analysis also supports the development from identification to noun-intensification posited for particular. Note that, in contrast to the change posited for whole (§5.2.3), particular shifted from expressing pure rather than quantitative identification to noun-intensification. Interestingly though, particular has developed both qualitative and quantitative noun-intensifying uses. Again, as with the completeness adjectives, the first noun-intensifying uses of particular are attested in combination with a specific semantic subset of the nouns with which the identifying source uses also occur in the data, i.e. nouns referring to requests, observations and 42. The quantitative data for the period 1500–1570 cannot be taken as representative as they are based on only 5 prenominal attestations (despite consulting two different corpora).

From description to identification (to noun-intensification)

207

obj. descr. mod. 

classifier

total

5

0

0

0

0

0

0

0

0

5

% 100.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0



0

31

n

secondary determiner sec. det./  noun-intensifier

 1500–1570 (PPCEME, CEMET) 1570–1640 (PPCEME, CEMET) 1640–1710 (PPCEME, CEMET) 1710–1780 (CLMETEV)

subj. descr. mod  sec. det./  obj. descr. mod. 



noun-intensifier noun-intensifier/ subj. descr. mod. sec. det./  subj. descr. mod. 

Table 19. Quantified results of the corpus study of particular

n

77

6

5

2

6

9

%

54.6

4.3

3.5

1.4

4.3

6.4

n

137

0

0

7

20

17

%

62.6

0.0

0.0

3.2

9.1

7.8

n

167

2

12

0

0

16

%

74.9

0.9

5.4

0.0

0.0

7.2

1780–1850 n (CLMETEV) %

143

26

34

1

2

29

1

16

56.7 10.3 13.5

0.4

0.8 11.5

0.4

6.3

1850–1920 n (CLMETEV) %

172

0.0 22.0 9

23

4.1 10.5 1

25

0.4 11.2

5 141 3.5



6 219 2.7



0 223 0.0



0 252 0.0



23

17

1

0

13

0

5

0 231

74.5 10.0

7.4

0.4

0.0

5.6

0.0

2.2

0.0

7 54 3.2 24.8

0 0.0

0 0.0

2 0.9

4 1.8

2 0.9

0 218 0.0  0 251



n %

149 68.3

CB ukspok n %

237

6

6

0

0

1

0

1

94.4

2.4

2.4

0.0

0.0

0.4

0.0

0.4

0.0

n %

384 82.2

13 60 2.8 12.8

0 0.0

0 0.0

3 0.6

4 0.9

3 0.6

0 467 0.0 

CB times

CB total



208

The specificity adjectives

emotions. The quantitative profile of specific (Table 20) cannot be used as direct evidence for a deictification account. As for particular, secondary determiner uses are predominant in all historical periods looked at. One exception is 1640–1710, in which the classifier uses outnumber the secondary determiner uses. However, as 7 of the 10 attested classifiers stem from the same text, these numbers cannot be considered representative. Nevertheless, classifier uses of specific are attested from the earliest to the most recent data sets, despite a continuous gradual decrease in relative frequency (from 76.9% in 1640–1710 to 5.0% in the CB data sets). Interestingly, the proportion of descriptive modifier uses of specific, unlike that of particular increases over time, taking up almost one fifth of all prenominal uses in the contemporary data. This is surprising, seeing that prenominal particular is overall much more frequent in the data than prenominal specific.

1710–1780 (CLMETEV)

noun-intensifier

sec. det./  subj. descr. mod.

subj. descr. mod.

sec. det./  obj. descr. mod.

obj. descr. mod.

classifier

total

n 1640–1710 (PPCEME, CEMET) %

sec. det./  noun-intensifier



secondary  determiner

Table 20. Quantified results of the corpus study of specific

3

0

0

0

0

0

0

10

13

23.1

n

4

%

44.4

1780–1850 (CLMETEV)

n

36

%

53.7

1850–1920 (CLMETEV)

n

36

%

57.1

CB times

n

129

%

70.1

CB ukspok CB total

n

145

%

68.1

n

274

%

69.0

0.0 0.0 0

0

0.0 0.0 0

0

0.0 0.0 0

1

0.0 1.6 6

1

3.3 0.5 3

1

1.4 0.5 9

2

2.3 0.5

0.0 0.0 0

0

0.0 0.0 1

1

1.5 1.5 1

2

1.6 3.2 0

1

0.0 0.5 0

4

0.0 1.9 0

5

0.0 1.3

0.0 0.0 76.9 2

1

2

9

22.2 11.1 22.2 6

6

17

67

9.0 9.0 25.4 2

10

11

63

3.2 15.9 17.5 9

25

13 184

4.9 13.6 7.1 0

53

7 213

0.0 24.9 3.3 9

78



20 397

2.3 19.6 5.0



Chapter 7 Such, zulk and what In this chapter, I will scrutinize Bolinger’s (1972) hypothesized pathway leading from identifying meaning to noun-intensifying meaning. To this aim three corpus studies have been carried out. The first and most prominent case study deals with such (§7.1 to §7.3), the second and third case study focus on its Dutch equivalent zulk (§7.4) and the largely comparable English item what (§7.5). All three items have developed identifying and noun-intensifying uses, as illustrated in the examples below. (222)

(223) (224)

(225) (226) (227)

you have a real community, the sort where people make soup for sick neighbours and know whose cat has gone missing. Joining such a community, however, comes with responsibilities -making the time to take part, sit on a local committee, bake a cake for the fete, buy raffle tickets, raise funds. (CB times) She’s such a happy, friendly gregarious person and very, very responsible (CB times) ... maar zulke menschen, rijke vreemdelingen in uitheemsche kleeren, reizend voor pleizier, kijkend uit nieuwsgierigheid, hooren die in hun huis? (DBNL 1921) ‘… but such people, rich strangers in foreign clothes, travelling for fun, looking around out of curiosity, do they belong in their homes?’ (Ghesquière and Van de Velde 2011: 768) We zijn allemaal zulke afschuwelijke hypocrieten. (DBNL 1969) ‘We are all such terrible hypocrites.’ (Ghesquière and Van de Velde 2011: 768) What safety precautions do you take as a woman driver (CB times) What a mistake! (CB times)

This chapter has three main aims. First, it will provide fine-grained cognitive-functional descriptions of the two main discourse functions of prenominal such (§7.1 and §7.2). The existing literature on such focuses mainly on its classification in terms of grammatical class, often irrespective of its function in the discourse. Some studies have proposed unified analyses of such as a (pronominal) adjective (e.g. Seppänen 1978; Fischer 1992; Spinillo 2003), a determiner (e.g. Bolinger 1972; Allerton 1987; Börjars

210

Such, zulk and what

and Burridge 2001) or an adjectival predeterminer (Huddleston and Pullum 2002). Others classify identifying and noun-intensifying such in distinct word categories (e.g. Quirk et al. 1985; Altenberg 1994; de Mönnink 1996; Wood 2002). The present account will provide more detailed descriptions of the distinct characteristics of the identifying and noun-intensifying functions, both structural, semantic and collocational. Second, this chapter will look into the diachronic connection between the identifying and noun-intensifying uses of such. On the basis of synchronic observations, Bolinger in his (1972) seminal work posited that the functional variation displayed by such may be the result of a diachronic change, whereby the originally identifying element has acquired intensifying semantics over time (§7.3). Bolinger (1972: 61) hypothesized that the shift posited for such is representative of “a kind of wholesale migration in that direction”, affecting not only true determiners, but also adverbs and adjectives with identifying meanings. He explicitly mentions complete, total and particular (Bolinger 1972: 59, 149, 150) and suggests that other, semantically similar adjectives may have proceeded along the same path of change. Chapters 5 and 6, however, have shown that of the five adjectives studied only two can be argued to have shifted from identification to intensification as predicted by Bolinger (1972), viz. whole and particular. The noun-intensifying meanings of complete and total developed along the path proposed by Adamson (2000) and Paradis (2000), i.e. from description to intensification. In this chapter, the validity of Bolinger’s hypothesized shift will be evaluated for such against diachronic and synchronic data. Third, this chapter will further test the general, predictive value of the proposed pathway from identification to noun-intensification by carrying out comparative studies on Dutch zulk ‘such’ and English what, two items displaying the same kind of functional variation between identifying and noun-intensifying uses, as illustrated in (3) to (6) (§7.4 and §7.5). The comparability of such and zulk provides broader data and evidence than investigated in Chapters 5 and 6. What, together with such, is one of the key examples of the identification-to-intensification shift discussed in Bolinger (1972).

Identifying such

211

7.1. Identifying such43 In its identifying function, such, like the completeness and specificity adjectives, is realized as a secondary determiner, complementing the identifying information provided by the primary determiner. In the following paragraphs, I will argue that secondary determiner such is a means for the speaker to provide the hearer with additional phoric instructions “that are more semantically diverse and more complex than the primary determiners [it is] bound by” (Davidse, Breban, and Van linden 2008: 500). More specifically, in addition to the indefinite primary determiners it occurs with, such provides type identification information by setting up phoric relations in the discourse (§7.1.1) and hence conveys textually intersubjective meaning. In (228), for instance, the indefinite article a indicates that the hearer will not be able to identify one unique instance as the referent of the NP. The presence of such then signals to the hearer that additional information on the type of which the NP referent is an instance (i.e. ‘an idea so simple, so obvious and yet so useful that the immediate assumption is that everybody else must already know it’) can be retrieved from the discourse context, in this case the preceding sentence. (228)

Occasionally one comes across an idea so simple, so obvious and yet so useful that the immediate assumption is that everybody else must already know it. Such an idea is Mike Timms's (CB times)

7.1.1. Type-phoricity Both in the synchronic and diachronic data, identifying such combines with a range of different determiners and quantifiers, albeit all indefinite.44 Most typically, such combines with the indefinite article a in singular count NPs, as in (229), and the zero-determiner ø in uncount (230) and plural (231) NPs.45 43

Parts of the discussion of identifying and noun-intensifying such build on the diachronic research discussed in Ghesquière and Van de Velde (2011). 44. More information on the compatibility of such with other determiners is given in §7.1.3. 45. Following Davidse (2004), I am inclined to posit a zero-determiner or zeroelement in plural NPs of the type ø (such) classes. In the data such was in plural NPs also found with other indefinite determiners such as some or any (§7.1.3), and in singular count NPs an indefinite primary determiner can, of course, also be pre-

212

(229)

(230)

(231)

Such, zulk and what

German workers have taken the waters and indulged in the pleasures of herbal packs, therapy and massage since the 19th century… But the cost of such a cure has risen in recent years to an average of £ 2,000 a visit. (CB times) The teaching of mixed-ability classes becoming the norm must be avoided at all costs. Very marginal benefits come from such teaching and the system must allow for some form of streaming. (CB times) I am made very uneasy about any proposition to teach pupils in school how to become good parents. … And such classes would be inevitably vulnerable to fads and fashions. (CB times)

As discussed in Section 1.4.1, the identifying information conveyed by indefinite determiners such as a and ø is “not sufficient to put the hearer in mental contact with a uniquely determined instance of the category” (Langacker 1991: 104). A and ø form indefinite NPs which establish discourse referents, but which fail “to identify any specific individual as its referent”. As a consequence, indefinite determiners require the hearer to “conjure up” a referent as an instance corresponding to a type (Langacker 2008: 289; Davidse 2004: 522). Therefore, the hearer has to be clear about the type specifications the instances have to correspond to. In this sense, indefinite reference presupposes type-identifiability (Gundel, Hedberg, and Zacharski 1993: 275). The type-identifiability inherent in indefinite primary determiners can be further specified by such, as in (229) to (231) above. As a secondary determiner, such provides instructions for the hearer to help him determine the kind or type of which the NP referent is an instance (see e.g. Carlson 1978: 219; Duinhoven 1988: 130; Payne and Huddleston 2002: 435; Wood 2002: 98; Spinillo 2003: 198). More specifically, such contributes to the type-identification by invoking phoric relations in the discourse, which invite the hearer to connect the NP referent with the referent of a preceding or following NP or stretch of discourse.46 Hence, whereas the cognitive sent. Although not overtly realized, the zero-determiner then occupies a distinctive value within the determiner paradigm. As such, it is a true zero, i.e. a distinctive semantic value symbolized by zero within a paradigm, rather than ‘nothing’, i.e. nothing in the form and nothing in the semantics (McGregor 2003). 46. Phoric elements contribute to the identification of an NP referent by invoking a relation to another discourse referent as its antecedent (e.g. Martin 1992: 98; De Mulder 1998: 2). In other words, phoric NPs embody “directives indicating that information is to be retrieved from elsewhere” (Halliday and Hasan 1976: 31).

Identifying such

213

state corresponding to indefinite articles is type-identifiability (Gundel, Hedberg, and Zacharski 1993), that of the determiner complexes such a and such ø is type-phoricity (Breban, Davidse, and Ghesquière 2011). Halliday and Hasan (1976: 313) have further detailed the phoric relation set up by such as entailing comparative reference, i.e. the referent of the such-NP is further identified through comparison with another discourse referent. Accordingly, such is claimed to evoke not a relation of identity but one of similarity or likeness, involving “a sharing of some criterial properties but not others” (Hawkins 1978: 249). In other words, the NP referent is like the discourse referent(s) in relation to which it is interpreted, but not identical to it. Van der Auwera and Van Olmen (2013) hence refer to such and its West Germanic cognates as similative elements. In (229), for instance, the NP such a cure denotes a type of cure that can be evoked by typical instances of it such as baths, herbal packs, therapies, and massages. The complex determiner such a then refers to a type that is like, but not restricted to, the instances listed, “not point[ing] out the object itself, but something ... of the same kind or class” (Spinillo 2003: 197) (see also Quirk et al. 1972: 703; Carlson 1978: 219; Altenberg 1994: 229; Huddleston and Pullum 2002: 435, 1104; Wood 2002: 98). 7.1.2. Generalized instantiation The type-identification and phoric instructions typically provided by determiner units with such is, in my opinion, understood best in terms of Langacker’s (2005: 170ff, 2009: 9) notion of generalized instantiation.47 Depending on where this information is to be retrieved from, three main types of phoric relations can be distinguished: anaphora, i.e. the antecedent can be found in the preceding text; cataphora, i.e. the antecedent is retrievable in the following text; and exophora, i.e. the antecedent is to be found in the discourse context rather than the text itself (Halliday and Hasan 1976: 14f; Martin 1992: 14f; Willemse 2005: 91–92; Breban 2010a). The different types of phoric relation invoked by such are discussed in §7.1.2. 47. Generalized instantiation is distinct from generic reference. Generic statements refer to an arbitrary instance which is taken to be representative of the whole category of which it is an instance (Langacker 1991). Generalized statements have weaker implications in the sense that they do not necessarily apply to all the members of a category. Rather the generalization is typically a local, text-bound generalization based on a number of contingent occurrences rather than a global generalization to a class as such as part of the structure of the world, which is characteris-

214

Such, zulk and what

Generalized instantiation is the cognitive mechanism by which speakers make a certain abstraction which generalizes from concrete entities by invoking a representative instantiation capturing the commonalities of the actual instances.48 It thus allows language users to establish mental contact with a particular discourse referent “through the mediation of fictive or virtual entities conjured up for that purpose” (Langacker 2005: 170). As such, at least two distinct instantiations are involved: the ‘fictive’, generalized one and the actual antecedent, which are alike in some, but not all respects (Halliday and Hasan 1976: 77).49 Determiner units with such consequently always involve a type of dual reference (Ward and Birner 1995: 732), i.e. reference not only to a known type (retrievable from the preceding or following text) but also to a new instantiation of that type (introduced by the such NP). In other words, “[t]he addressee not only needs to access an appropriate type-representation, he must ... construct a new representation by the time the sentence has been processed” (Gundel, Hedberg, and Zacharski 1993: 276). In (232), for instance, the NP such legs is understood through its comparison to the description in the previous sentence: the legs in question are another instance of the type instantiated by Gullit’s legs which allow sportsmen to block attacks, slow the pace and instigate counter-attacks. In (233), the NP such excrement instructs the hearer to conjure up a generalized instantiation of a TV-show which is of the same type as the actual shows mentioned in the following as-clause. Examples (232) and (233) exemplify two semantically slightly different types of generalized instantiation (Kristin Davidse, p.c.; Breban, Davidse, and Ghesquière 2011). In (232) the reference to new instances via a shared type seems to be foregrounded; in (233) the generalized aspect is foregrounded. This distinction is apparent in their respective paraphrases. Such legs in (232) can be paraphrased both as ‘that type of legs’, emphasizing the reference to a known type, and as ‘similar legs’ (indefinite), emphasizing the introduction of new instance. Such excrement in (233) on the other hand does not allow the paraphrase with ‘similar’, only the 'that type of excrement' paraphrase. The link between the generalized and actual tic of generic reference (Willemse 2005: 189). See Breban (2011) for a discussion of generalized instantiation expressed by secondary determiners in English NPs. 48. Langacker (1991: 74–81) uses ‘instantiation’ as a cover term pertaining to singular count nouns (which describe concrete entity-types), uncount nouns (which describe homogeneous mass types), and plural count nouns (which designate a replication of concrete entities, or a heterogeneous mass type). 49. Langacker (1991: 78) characterizes fictive conceptualizations evoked by linguistic predications as ‘‘departures from the direct description of ACTUALITY’’.

Identifying such

215

instances is then more meronymic than comparative, e.g. such excrement as/including My Hero, The Vicar of Dibley and Dad’s Army (Hendrik De Smet, p.c.). (232)

(233)

Forest’s retort was furious, but Gullit blocked them at every turn, slowing the pace, sweeping away the danger and still finding time to instigate some flowing counter-attacks. Linford Christie should have such legs. (CB times) All we’re left with is such excrement as My Hero, The Vicar of Dibley and re-runs of Dad’s Army, which was never funny in the first place . (CB ukspok)

Visually, the generalized instantiation expressed by the NP such excrement in (233) can be represented as in Figure 18. x1, x2 and x3 correspond to the actual instances – My Hero, The Vicar of Dibley and Dad’s Army – and x denotes the instantiation evoked by the NP with such. “The filled dots indicate that [x1, x2 and x3] are actual instances, distinguished from one another by their locations in space or in time. ... The unfilled dot [for x] indicates that it is only imagined, being conjured up just to capture the generalization” (Langacker 2009: 10). space

space x1

x2

x3

actual instances

x

imagined instances (representative)

Figure 18. Generalized instantiation (Langacker 2009: 10)

The generalized instantiations evoked by such NPs can be retrieved from the discourse context in two ways: (i) through generalization from a qualitative characterization or type description, or (ii) through abstraction from specific, concrete instances in the discourse. The actual occurrences can, in other words, be present in the discourse context only implicitly or they can receive explicit mention. These two types are illustrated in (232) and (233) above. As Figures 19 and 20 visualize, their schematic Cognitive Grammar representations differ slightly. As argued above, the generalized entities evoked by such NPs are not realized in the spatial domain, but on

216

Such, zulk and what

Type specification plane

T

gi1

gi2

Generalized instantiation plane

Spatial instantiation plane

Figure 19. Generalizing from a type-description (based on Breban and Davidse 2003: 296)

Type specification plane

T

gi1

i1

i2

gi2

i3

Generalized instantiation plane

Spatial instantiation plane

Figure 20. Generalizing from concrete instances (based on Breban and Davidse 2003: 296)

Identifying such

217

the generalized instantiation plane. They instantiate a new, representative instance of a known type – available either through a following or preceding description or through an enumeration of exemplary instances –, but at the same time they generalize over spatially distinct referents. Only when a generalization is made from concrete instances explicitly mentioned in the discourse context, as in (233), is there actual instantiation on the spatial plane, as illustrated in Figure 20. Note that (232) and (233), like (234) and (235) below, represent anaphoric and cataphoric uses of identifying such, which instruct the hearer to retrieve the necessary type specifications from the preceding or following text respectively (see also Bolinger 1972; Sinclair 1990; Altenberg 1994).50 In (234), the NP such a party generalizes over and is conjured up as a representative instantiation of a type of party described by the NP a delicious evening in the preceding sentence which refers to an actual instance. In (235), the actual instances that are generalized over are to be retrieved cataphorically. The NP such simple things is understood only through abstraction from and metonymic association with the following enumeration of concrete instances, crucially typified by the property of ‘simplicity’. Note that whereas in (234) the such NP evokes just one representative instance, i.e. the absolute minimum to capture any commonalities, the plural such NP in (235) generalizes over multiple concrete instances, “reflecting more directly the multiplicity of instances all exhibiting the property in question” 50. The Late Modern English and Present Day English data also contained eight instances of the construction such and such a/ø N, as in (i) and (ii), which conjure up a highly abstract type that “has no status outside the mental space created for this purpose” (Langacker 2005: 166). Vandelanotte (2009: 121, 2011), who has studied so and so and such and such in the light of speech and thought representation, argues that these constructions create “suspensive reference”, which is “to be contextually, but only ‘schematically’ inferred”. In other words, speakers use such and such “to refer to a thing or person when [they] do not want to be exact or precise” (Sinclair 2001, s.v. such). Interestingly, in line with the results of Vandelanotte’s (2011) study, all eight instances found in the data are used in speech and thought representation contexts, either direct, as in (i), or indirect, as in (ii). Note that, even with direct speech representation, the such and such NPs “clearly do not literally repeat a prior speaker’s words” (Vandelanotte 2011). (i) Someone whispered to her: “The section must leave at such and such an hour...” (CLMETEV Bagnold, 1920, The happy foreigner) (ii) Well I remember I remember the er the teacher calling me outside and telling me not keeping company of such and such a friend. (CB ukspok)

218

Such, zulk and what

(Langacker 2005: 188). In terms of frequency, anaphoric realizations of identifying such are predominant in all the historical stages looked at. (234)

(235)

We spent a delicious evening at our next door neighbours’ and, as the dinner was coming to an end, the chef from the local traiteur was briefly introduced to the table and given a round of applause. At such a party, you could expect to meet a poissonier or fleuriste; the craftspeople of food, drink and decoration belong with the white collar professions. (CB times) Navigator’s innovations have included such simple things as changing the background colour of Web pages, the use of online forms and tables better formatting control such as the ability to centre text, and the implementation of security features. (CB times)

7.1.3. Definiteness and indefiniteness in conflict The preceding section was devoted to the discussion of the dual reference and generalized instantiation evoked by determiner complexes with such. On the one hand a generalized instantiation is introduced into the discourse, while on the other hand reference to a given type description or to a number of exemplary concrete instances is evoked, either anaphorically or cataphorically. Building on this, the identifying information provided by such a/ø can be claimed to be indefinite with regard to instance-identification and definite with regard to type-identification. Such-NPs mark a certain shift in the discourse as they introduce new instances, but at the same time they add to the textual cohesion by setting up phoric relations within the discourse. Through this dual function, such evokes both indefiniteness and definiteness, albeit on different conceptual levels: an indefinite instance gi1 on the generalized instantiation plane and a definite type T on the type specification plane (see Figure 19). The association of prenominal such with indefiniteness is clear from the fact that, introducing new instances, it has a clear preference for occurring in NPs headed by indefinite primary determiners. Such typically occurs with the indefinite article a(n) in singular NPs and with the zero-determiner ø with plural and mass head nouns, as in (229) to (231). In addition, identifying such is in the historical and contemporary data also found following the indefinite determiners any (236) and some (237), and cardinal numbers (238). Interestingly, in Present Day English identifying such is also attested in NPs of the type the + superlative adjective/ordinal numeral + such +

Identifying such

219

N(s), as in (239) and (240). These constructions, although seemingly definite, can be considered abstract or implicit partitive constructions and can be paraphrased as 'the superlative adjective/ordinal number instance of such ø Ns', in which a definite instance is picked out from an indefinite set (e.g. the first of such bodies, the biggest of such surveys). (236)

(237)

(238)

(239) (240)

And likewayes shall you iudge of this figure , which is made of two lines , and not of one onely. So that whan so euer any suche meetyng of lines doth happen, the place of their metyng is called an Angle or corner (PPCEME Record, 1551, The path-way to knowledg, containing the first principles of geometrie) Neverthelesse sometimes we have Exceedings: then we have two or three Dishes -but that is very rare- otherwise never but one: so that a Cake and a Cheese would be very welcome to me: and a Neat's tongue, or some such thing, if it would not require too much money. (PPCEME Strype, 1662-1665, Letters) Loads such as bridges boats, locomotives, bits of buildings, even whole buildings, are shifted by road. Twenty such loads go through Suffolk daily, while counties with large ports are even busier. (CB times) In the biggest such survey, 30,000 Tories across the country last autumn demanded the reform of Europe's agriculture and fisheries policies (CB times) this idea that it should be both a research and conservation body was also crucial and indeed exemplary Mm. and it was the first such body (CB ukspok)

Another observation endorsing the indefinite character of identifying such is the fact that it can occur in existential sentences of the type exemplified in (241a). Occurrence in such constructions is traditionally considered to be a clear diagnostic test for indefiniteness (Lyons 1999: 16). Definite determiners cannot occur in existential constructions, e.g. (241b). (241)

a. b.

I’m ever so grateful when anyone helps me like that ... I think it’s lovely that there are such people. (CB ukspok) *I think it’s lovely that there are these people.

While closely associated with indefiniteness, determiner complexes with such are also inextricably linked with definiteness. Denison (2002: 6) and De Smedt, Brems, and Davidse (2007: 238–243) have pointed out the func-

220

Such, zulk and what

tional similarity of such a/ø to the definite determiner complexes that/the sort/kind/type of. For (242), De Smedt, Brems, and Davidse (2007: 240) note that “[t]he referential meanings conveyed by that sort of in this example can also be expressed by an NP with ... such: It takes time to find such a man” [emphasis original]. Similarly, in (243), the NP the kind of peace Stalin and Hitler brought to Poland cataphorically refers to and generalizes from a concrete instance of peace described by the restrictive relative clause in the postmodifier. As with the anaphoric example, the NP in (243) can be paraphrased with such, e.g. “such a peace as Stalin and Hitler brought to Poland” (De Smedt, Brems, and Davidse 2007: 242). In fact, the main semantic difference between the determiner complexes such a/ø and that/the sort/kind/type of is that the kind meaning is implicit in the former, whereas it is explicit in the latter. (242)

(243)

… a man I can trust and believe in. It takes time to find that sort of man, but he is worth the wait. (www.forums.plentyoffish.com/16073117datingPostpage4aspx/) (De Smedt, Brems, and Davidse 2007: 240) … so it might as well go to Richard Holbroke. It was he who pitched together the Dayton accord, bringing to Bosnia the kind of peace Stalin and Hitler brought to Poland. (CB times) (De Smedt, Brems, and Davidse 2007: 241)

In the context of the observed tension between definiteness and indefiniteness in the meaning of such a/ø, two alternative analyses of the determiner complexes come to mind formulated from two opposite perspectives. Whereas van der Auwera and Van Olmen (2013) consider them a special type of indefinites, Lyons (1999) opts for an analysis as a demonstrative, a category typically thought of as definite. Van der Auwera and Van Olmen (2013) refer to English constructions with such and so and their Dutch, German and Afrikaans counterparts as similative indefinites, referring to “an indefinite token that is similar to a definite token which both instantiate a type, which is also definite”. As such the NPs in question are thought of as indefinite, but with a certain degree of ‘pragmatic’ definiteness. Taking the opposite standpoint, Lyons (1999: 41) proposes the at first sight somewhat paradoxical analysis of such as indefinite demonstrative. In the literature the ‘demonstrative’ or pointing characteristics of identifying such have often been observed, with the proviso that it points not to instances but to a particular class or type of instances (e.g. Altenberg 1994: 227; Mackenzie 1997: 99; Wood 2002: 92). Lyons (1999: 41, 151) considers the question if

Identifying such

221

“perhaps one can go yet further, and propose that such is an indefinite demonstrative”, as it is basically “synonymous with this/that (though lacking the deictic contrast, being therefore of the ‘general demonstrative’ type), differing only in that it is constrained to occur in indefinite noun phrases”. Whereas demonstratives “introduc[e] a referent (or referents) to the hearer; and ... instruc[t] the hearer to match this linguistic referent with some identifiable object ... known on the basis of previous mention in discourse” (Hawkins 1978: 152), such a/ø instructs the hearer to match the NP referent with a phorically identifiable type. Lyons (1999: 151) argues, however, that if “such etc. are simply indefinite variants of the demonstratives, ... we would expect to find languages in which the complementary distribution were reflected in identity of form”. Due to a lack of such typological evidence, Lyons (1999: 152) ultimately rejects the indefinite demonstrative hypothesis. Nonetheless, in English, identifying such a/ø is functionally very similar to demonstrative this/that, especially to the type or ‘variety’ interpretation of the latter (Lyons 1999: 40).51 The demonstrative in sentences like I wish I could afford to buy that car and That’s the car I want can be interpreted not as referring to a particular car, but “to a type of car ... exemplified by the entity indicated” (Lyons 1999: 40). The demonstrative is, in this respect, “very close in meaning to such a car” (Lyons 1999: 40– 41). Support for the analysis of such a/ø as a distinct determiner complex in English with its own function and place in the determiner paradigm comes from a study by Hole and Klumpp (2000). Although they label it an article rather than a demonstrative, Hole and Klumpp (2000) have identified a determiner which, like determiner complexes with such, simultaneously conveys definite type reference and indefinite token reference. In colloquial German son is used with reference to “an indefinite token of a salient type”, as in (244).52 Here the colour of the jumpers functions as the criterial feature linking the different instantiations and identifying the contextually 51. Note that, like such, the demonstratives can also be used to convey intensification, albeit adjective-intensification and not noun-intensification, as in (i) and (ii). The restricted scope of that in (ii) over the following adjective only can be illustrated by means of the predicative paraphrase ‘a singer who is that great’. (i) Not many couples could be this unsentimental, and perhaps that’s a good thing (WB times) (ii) I’m not that great a singer; I can’t carry off a performance of a song (WB times) 52. In informal Dutch a similar form zo’n is used, also grammaticalized from zo ‘so’ and the enclitic indefinite article ‘n ‘an’.

222

Such, zulk and what

salient type. The NP sonen Pullover then abstracts away from this salient type and shifts to a generalized instantiation invoking any jumper with the quality ‘green’ in the shop (cf. Hole and Klumpp 2000: 234, 237). (244)

(While talking about green jumpers in a shop) Kaufst du mir sonen Pullover? ‘Will you buy me a jumper like this?’ (Hole and Klumpp 2000: 237)

7.2. Noun-intensifying such53 This section is devoted to the description of the syntactic, semantic and collocational behaviour of noun-intensifying such. As a noun-intensifier, such heightens the degree of all gradable type specifications in the NP, either inherent in the head noun (245), in modifiers of the head noun or (246) in both the head noun and its modifiers (247) (§1.3.2.2). (245) (246) (247)

And it's such a pity. (CB ukspok) the nurses and doctors there do such terrific work (CB times) I think I fancy Australia given the fact they're giving England such a good pasting. (CB ukspok)

In the literature, noun-intensifying such is mostly argued to modify only one element in the NP: a modifier of the head in NPs of the type such an idiotic man or a property associated with the head in NPs of the type such an idiot (e.g. Allerton 1987: 15; Altenberg 1994: 227; de Mönnink 2000: 149). This type of account is very much in line with the scope relations generally defined for typical adjective-intensifiers like very and awfully in (248) and (249) respectively (§1.3.2.1).

53. Allerton (1987) and Bolinger (1972) label (some instances of) nounintensifying such as exclamatory. I have opted not to use this term as it implies a link with the exclamative clause type. As Huddleston and Pullum (2002: 293) rightly observe, noun-intensifying such is not a marker of the exclamative clause type as it “can occur in any of the major clause types – ... interrogative (Have you ever seen such chaos?), imperative (Don’t be such a wet blanket!), as well as ... declaratives”, as in (245) to (247) above.

Noun-intensifying such

(248) (249)

223

He is also a charmer, a very nice man and a national treasure. (WB brbooks) That's an awful lot of people who would kick up an awfully big fuss if they were told it was illegal. (WB times)

The adverbs here have scope over and modify only the following adjective, even when the NP is headed by a gradable noun, as in an awfully big fuss in (249). For noun-intensifying such, I argue instead that it has potential scope over the whole descriptive part of the NP it forms part of (cf. Bolinger 1972: 88–89; Spinillo 2003: 201), albeit only in terms of one dimension, viz. gradability. The difference in syntactic scope between the nounintensifier such and adjective-intensifiers like very is reflected in their distinct behaviour in predicative constructions. Whereas the NP a very nice man can easily be paraphrased as a man that is very nice, illustrating the restricted scope of the intensifier over the adjective nice only; this is not possible for NPs with noun-intensifying such (e.g. *a pasting that is such good). Moreover, the extended scope of noun-intensifying such ties in with the diachronic observation that the earliest noun-intensifying uses are attested in NPs consisting only of such and a head noun. Only in later stages could additional modifiers intervene between the noun-intensifiers and the head, a construction that later becomes the predominant one in the corpus data (see Figure 21). 7.2.1. Semantic-pragmatic types of noun-intensification expressed by such Noun-intensifying such, like all noun-intensifiers, measures the degree of certain evaluative or scalar notions inherent in the descriptive part of the NP it forms part of. More specifically, such typically functions as a booster, scaling a property upwards from an assumed norm (Quirk et al. 1985: 445). In other words, such heightens the degree of a gradable property, locating it high on an open ended intensification scale. In correlation with the type of head noun and noun modifiers it combines with, such can express different types of degree modification, viz. qualitative and quantitative degree modification, as in (250) and (251) respectively (see §5.3.3.2 for a similar distinction observed for noun-intensifying whole). In example (250), such heightens the degree of the quality ‘excitement’, caused by English cars on the continent. In (251), on the other hand, such modifies the quantifying construction a load of and the sequence can be appropriately paraphrased as ‘so much’, more directly reflecting the quantitative import.

224

(250) (251)

Such, zulk and what

It used to be such a thrill when you saw English cars on the road on the continent. You used to hoot and wave. (CB times) Even when we are not locked into a behavioural system, certain rewards can take on such a load of meaning that we live our lives trapped and fascinated by them. (WB brbooks)

With both qualitative and quantitative degree modification, the properties which such places high on an open-ended intensification scale can be either inherent in the type description provided by the head and/or modifiers of the head of the NP, or they can be more implicitly associated with the NP referent. Examples (252) and (253) illustrate the two possibilities for qualitative degree modification. In (252), such intensifies the degree of dissatisfaction or disagreement experienced by the speaker. In (253), such can be understood to intensify the property of speed naturally connected with the concept ‘pace’. (252) (253)

Well we’ve just got the garden right it seems such a shame to leave. (CB ukspok) I was a bit worried that we came racing out of the starting blocks at such a pace, because I doubted sometimes whether we would be able to keep up that kind of effort for a full 80 minutes. (CB ukspok)

When such conveys quantitative degree modification it stresses the extent of the size or amount of the NP referent referred to. As observed for noun-intensifying whole (§5.3.3.2), quantitatively intensifying such can modify (i) the extent of a quantifier, as in (254), (ii) it can modify the extent of nouns that describe size/amount/extent, as in (255), or (iii) it can modify the extent of nouns that imply a notion of extent/size/amount, as in (256). In the latter case, the boundary between qualitative and quantitative degree modification is often rather fuzzy. In (256), the NP such respect can be paraphrased as ‘so much respect’, i.e. as conveying quantitative degree modification, but the paraphrase ‘such intense respect’ (i.e. qualitative degree modification) is not precluded. Interestingly, as with whole, such is first attested with nouns implying size before co-occurring with nouns describing size and with quantifiers. Also, as will be discussed in the following section, nouns clearly implying or describing size are found with nounintensifying such only from the Early Modern English period onwards.

Noun-intensifying such

(254) (255) (256)

225

Cos there's er such a lot of old people in place name ain't there (CB ukspok) And the bees found us. Bees in such numbers you would think you were inside a forest-sized hive. (CB times) the power of the local witchdoctors (for whom his mother had such respect that she made up her own spells to counter theirs) (CB times)

Bolinger (1972: 83) already noted that the distinction between the quality and quantity readings of noun-intensifying such is not always clear-cut. Many gradable nouns can evoke both qualities and quantities and the correct interpretation is often highly context-dependent. Examples (257) and (258) both feature the NP such lies, but the discourse context is crucial to come to the correct interpretation of the NP as quantitatively intensifying in (257) (‘so many lies’) and as qualitatively intensifying in (258) (‘lies so outrageous’). (257) (258)

The talk was always good. But such lies. You can't believe the amount of lies. (COCA) The most terrible thing about AIDS is that it destroys the relationship, no matter how loving, between the two partners and eats away at the character of the person who's dying. Nobody has AIDS and is noble. That's why all these melodramatic, kitschy plays about AIDS are such lies. (COCA)

7.2.2. Collocational behaviour of noun-intensifying such It was shown above that noun-intensifying such modifies the degree of gradable qualitative type specifications in the NP or modifies the extent of the size or amount denoted either by quantifiers or head nouns with size implications. Correlating with these two types of degree modification, such displays distinct collocational behaviour. When such engages in qualitative degree modification the graded properties are typically evaluative notions, as expressed by evaluative or emotive head nouns (e.g. wonder, blessing in Old English; bitch, relief, mess in Present Day English) and/or adjectives

226

Such, zulk and what

(e.g. good, wise in Middle English54; terrific, good, bad in Present Day English). In the case of quantitative degree modification, the gradable notions have some sort of size or quantity implication, as inherent in nouns such as work and tears in Old English and number and time in Present Day English, or adjectives such as little and short in Early Modern English and long and big in Present Day English. Note that, as Figure 21 shows, nounintensifying such is in Old English found only immediately preceding the head noun it modifies. Only in later stages is noun-intensifying such found preceding additional prenominal modifiers, a construction that later becomes the predominant one in the corpus data. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

15

0

28

2

37

70

33

92

74 such + head

51 179

such + mod + head

6

Figure 21. Modification patterns for noun-intensifying such

When such immediately precedes the head, the latter is typically a degree noun (Bolinger 1972), i.e. a noun that has gradable features, either qualitative or quantitative. Some illustrative examples are given in Table 21 for each main period distinguished in the diachronic data. Note that clear

54. In the Old English data, noun-intensifying such only occurs immediately preceding the head noun. One possible exception is example (i), where at least an intensifying implicature may be present: (i) Aec ic biddo higon ðaet ge me gemynen aet ðere tide mid suilce godcunde gode suilce iow cynlic ðynce (YCOE 750–850) ‘Also I pray the religious-community that you remember me at this occasion with such divine good such seems becoming to you’

Noun-intensifying such

227

examples of nouns that describe size or extent are found only in the Early Modern English period (e.g. number, distance). Table 21. Nouns occurring with noun-intensifying such in the diachronic data

Late Middle English

blessing, light, swingel (‘scourging’), tears, unright, wonder, worship, †wulder (‘glory’) †and3it (‘intellect’), bliss, evil, holiness, lust, sorrow, worship courage, cruelty, envy, fear, power, sorrow, wrath

Early Modern English

delight, fear, fool, humbleness, number, distance, plenty

Late Modern English

clown, craving, goodness, hurry, length(s), number, extent

Old English Early Middle English

As the degree nouns are so diverse and, with only a small number of exceptions, are all attested just once in the data, it is not very instructive to provide a more systematic account of the collocational behaviour of nounintensifying such for the diachronic data. Do note that already in Old English noun-intensifying such is found with both positively (blessing, wonder, worship) and negatively (unright, swingel ‘a scourging’) connoted head nouns. For Present Day English, Table 22 provides frequency lists of the degree nouns found in combination with such for the two subcorpora studied.55 Interestingly, it is only in the (spoken) contemporary data that noun55. Some of the degree nouns with which noun-intensifying such occurs in the data can be considered true collocates (Table a). The statistical measures were obtained via the tools provided by Collins WordBanksOnline. As a rule of thumb, T-values above 2 and MI-values above 3 are taken to be linguistically interesting (Clear 1993, Stubbs 1995). Only when both statistical measures for a particular combination exceed these values, do I take the nouns to collocate strongly with such. Table a. Statistical measures for the nouns most frequently combining with nounintensifying such CB times CB ukspok

collocate commitment consternation force fun mess problem

joint frequency 13 4 38 11 19 31

T-score 3.174 1.971 5.306 3.012 4.285 4.925

MI-score 3.061 6.124 2.844 3.446 5.885 3.114

228

Such, zulk and what

intensifying such is attested with true quantifiers, as in such a load of nonsense (CB ukspok) and such a lot of old people (CB ukspok). There has thus been a gradual collocational extension for noun-intensifying such from head nouns with qualitative, evaluative notions to head nouns implying and later describing size or extent to quantifiers (cf. §5.3.3.2 on nounintensifying whole). Table 22. Frequency lists of degree nouns combining with noun-intensifying such in the pattern such (a) N(s) CB times commitment consternation force gullibility hit honeypot pace prominence respect self-belief

1 1 1 1 1 1 1 1 1 1

CB ukspok fun mess  problem  relief bitch bottleneck change danger degree extent

2 2 2 2 1 1 1 1 1 1

In the synchronic data again most degree nouns occur with nounintensifying such just once, but the frequency lists do reveal some genrerelated differences in the collocational behaviour of noun-intensifying such. In the CB ukspok corpus, such combines with more expressive degree nouns than in the CB times corpus. This corresponds to the more subjective, evaluative nature of the language represented in these corpora. Whereas in the such + head pattern the head of the NP is typically a degree noun, this is not the case with the such + modifier(s) + head construction. Although the head nouns can convey gradable concepts, like the figuratively used noun cows in (259), they more typically express nongradable, semantically more neutral concepts, as in (260). (259) (260)

They’re such Stupid people they’re such stupid cows (CB ukspok) there’s such a nice swimming pool there which nobody goes to really. (CB ukspok)

Noun-intensifying such

229

In this pattern the gradable, descriptive notions modified by such are most often conveyed by the modifiers intervening between such and the head noun. Table 23 presents frequency lists of the adjectival modifiers intervening between noun-intensifying such and the head of the NP in the diachronic and synchronic data respectively.56 Note that the modifiers attested can express both qualitative and quantitative gradable properties, e.g. terrific, good, bad and long, big, small respectively. These adjectives can be analyzed as subjective descriptive modifiers on the one hand and objective descriptive modifiers on the other.57

56. As with the head nouns, the collocational strength between noun-intensifying such and the descriptive modifiers it co-occurs with most frequently in the synchronic data samples is relatively high. Table b below details the T-scores and MIscores in the different subcorpora for the three adjectives occurring most frequently with noun-intensifying such in the samples studied. Table b. Statistical measures for the adjectives most frequently combining with noun-intensifying such in the contemporary data CB times CB ukspok

collocate huge long terrific big good

joint frequency 62 78 3 75 88

T-score 7.337 6.869 1.482 8.037 8.237

MI-score 3.874 2.170 2.791 3.883 3.036

57. In contrast to intensifying such, identifying such shows a preference for classifying modifiers, as in (i), which provide further information concerning the subtype of the general type denoted by the noun they modify; and objectively descriptive modifiers, as in (ii), which attribute an objectively assessable property to the head. Moreover, the presence of additional modification, both preceding and following identifying such, is much less frequent, both in the diachronic and synchronic data. (i) I transferred my balance to the Co-op on the same offer only to find that the rate went up to 1.1 not long afterwards. ... Such marketing tricks are hardly in keeping with the moral stance the Co-op supposedly prides itself on. (CB times) (ii) this would only yield control of the operating subsidiaries through share stakes.

Such limited control would not give the administrators the authority to enter premises and hold talks with management. (CB times)

2 1 1 1 1

wise deep good holy rude

finer grisly

1 1

Late Middle English

Early Middle English great worthy bloody chaste crooked cruel cunning dangerous extraordinary extravagant

Early Modern English 2 2 1 1 1 1 1 1 1 1

abrupt absurd agreeable bewildering bold brave brief capacious captivating critical

Late Modern English 1 1 1 1 1 1 1 1 1 1

huge long terrific arresting constructed dense dirty elegant fast-moving gorgeous

CB times 2 2 2 1 1 1 1 1 1 1

big good long bad large stupid enormous huge nice short

CB ukspok

Table 23. Frequency lists of modifiers intervening between noun-intensifying such and the head in the historical and contemporary data

5 5 5 3 3 3 2 2 2 2

230 Such, zulk and what

Identifying and noun-intensifying such: diachrony

231

7.3. The diachronic relation between identifying and noun-intensifying such The identifying and noun-intensifying uses of prenominal such each have their own specific constructional, collocational and semantic-pragmatic characteristics. Nevertheless, I argue that the noun-intensifying uses have plausibly developed from the identifying uses. First, the quantified results of the data study, given in Table 24, show that identifying such is older than noun-intensifying such. Moreover, the data suggest that, once firmly established, the two functions are in constant competition. Table 24 shows that in the earliest Old English period such is exclusively found in its identifying function. However, for this period only eight prenominal attestations are found in the data, an insufficient number to make strong descriptive or quantitative claims. What Table 24 also shows is that the identifying uses are predominant in all the periods looked at, with the exception of the contemporary data where overall there is an almost equal distribution of identifying and noun-intensifying uses. Diachronically, the proportion of identifying uses gradually decreases from 73.5% in Old English to 46.6% in Present Day English, whereas the nounintensifying uses witness a steady increase from 15.3% in Old English to 46.9% in Present Day English. There is thus a gradual, proportional shift in the data from identification to noun-intensification. Second, the quantified results of the synchronic data study may also be adduced as evidence supporting the historically prior character of identifying such. Table 24 reveals considerable differences in the distribution of the identifying and noun-intensifying functions of such in the CB times corpus and the CB ukspok corpus.58 An explanation for these specific distributional patterns can be sought in the distinct genres and modes of the two subcorpora consulted. The CB times data represent formal, written newspaper language, characterized by its informative purpose. Newspaper articles aim to inform readers about recent events and address an audience as wide as possible. As a result, the clarity of the text is of prime importance and the language used can be expected to be rather conservative. In contrast, informal spoken language, as it is represented in the CB ukspok data, is generally attributed a more innovative character (e.g. Halliday 1978; Chafe 2003; Du Bois 2003). The predominance of identifying such in the CB times data 58. A chi-square test shows that the different distribution of identifying and nounintensifying uses is statistically significant between the CB times and the CB ukspok data (² = 38.215, p < 0.0001).

232

Such, zulk and what

Table 24. Quantified results of the corpus study of such

750–850 (YCOE) 850–950 (YCOE) 950–1050 (YCOE) 1050–1150 (YCOE) 1150–1250 (PPCME2) 1250–1350 (PPCME2) 1350–1420 (PPCME2) 1420–1500 (PPCME2) 1500–1570 (PPCEME) 1570–1640 (PPCEME) 1640–1710 (PPCEME) 1710–1780 (CLMETEV) 1780–1850 (CLMETEV) 1850–1920 (CLMETEV) CB times CB ukspok CB total

n % n % n % n % n % n % n % n % n % n % n % n % n % n % n % n % n %

identifying noun-intensifying 8 0 100.0 0.0 10 2 83.3 16.7 20 7 58.8 20.6 34 6 77.3 13.6 111 19 77.6 13.3 86 11 88.7 11.3 158 9 93.5 5.3 88 48 62.4 34.0 144 32 79.1 17.6 128 16 87.1 10.9 143 22 82.2 12.6 99 53 61.9 33.1 108 55 63.5 32.4 105 54 62.5 32.1 107 41 65.6 25.2 64 106 36.0 59.6 171 147 50.1 43.1

vague 0 0.0 0 0.0 7 20.6 4 9.1 13 9.1 0 0.0 2 1.2 5 3.5 6 3.3 3 2.0 9 5.2 8 5.0 7 4.1 9 5.4 15 9.2 8 4.5 23 6.7

total 8 12 34 44 143 97 169 141 182 147 174 160 170 168 163 178 341

Identifying and noun-intensifying such: diachrony

233

(65.5%) and of noun-intensifying such in the CB ukspok data (59.6%) then fits in with the idea that noun-intensifying such is a more recent innovation. Third, the more recent character of the noun-intensifying uses is supported by the changing modification patterns of the identifying and nounintensifying uses (§7.2.2). Noun-intensifying such shifts from exclusively modifying the head noun to predominantly modifying the head of the NP and a gradable modifier of the head. This is a remarkable trend as intensification is typically associated with adjectives and noun-intensifying meanings of such might consequently have been expected to originate in the such + modifier + head construction, rather than in the such + head construction. The trend makes sense, however, if the noun-intensifying use derives from the identifying use: the rise of the noun-intensifying meaning of such is not due to the presence of the adjective, but arises naturally, i.e. through pragmatic strengthening (Hendrik De Smet, p.c.). Fourth, the claim that the noun-intensifying use of such derives from the identifying use fits in with Bolinger’s (1972) hypothesis of a diachronic pathway leading from identification to noun-intensification. More specifically, for such he argued that this development is highly plausible “because the ‘suchness’ of something is so likely to be an intensifiable characteristic. We begin by viewing it as pointed to, and end by viewing it as worthy of note, hence as enhanced”. Such a shift is compatible with the welldocumented process of meaning change through pragmatic strengthening (Traugott and Dasher 2002; Traugott 2003a). The noun-intensifying meaning is then an implicature at first, which is later fully semanticized. In its identifying function such has a phoric function, referring to a kind, characterized by a specific feature or characteristic, often a scalar, gradable notion. An NP like such a problem, for instance, can be reasonably assumed to have originally functioned as a phoric NP referring to a specific kind of problem described or illustrated in the context. The reason for picking up the NP in the ensuing discourse will often be that we are dealing with a kind of problem that is in some way remarkable, e.g. in terms of its importance or impact. Over time, this may have led to the inference that the use of such entails intensification. Support for such an account is found in the occurrence of contexts which allow a noun-intensifying interpretation as well as an identifying reading of such, as in (261) and (262). (261)

The opposition’s list of 125 Tory mps who act as part-time consultants and advisers is calculated to embarrass those named as well as the government and force the disclosure of their extramural earnings. There should be no need for such pressure. (CB times)

234

(262)

Such, zulk and what

Through wind, rain and weather, driving hell for leather, the RAC rally is unique. ... drivers face conditions ranging from dusty tracks to sodden forest roads. They get limited opportunities to become acquainted with the stages, some of which they have to complete in darkness. … They have a powerful, agile car that is likely to perform best in such hard conditions. (CB times)

In the light of these observations, the diagnostic test put forward by Altenberg (1994: 234) to distinguish between identifying and noun-intensifying uses needs to be nuanced. He argued that “what determines the interpretation is the presence or absence of two features: a defining referent in the context and a gradable element in the noun phrase. If there is a contextual referent but no gradable element, such is interpreted as identifying; in the reverse situation it is interpreted as intensifying”. This test, however, becomes problematic when both gradable features and phoric reference are contextually evoked, as in the above examples. The NP such pressure in (261) evokes identification through anaphoric reference to the implied embarrassment and disclosure. At the same time, however, a noun-intensifying reading of such can be inferred heightening the quantity or extent of pressure exerted. Similarly, in (262), such aids identification of the NP referent through backward reference to the description of the weather conditions and the poor state of the roads. On the other hand, the hardship of the conditions referred to also triggers a possible noun-intensifying reading of such. Importantly, whereas the noun-intensifying reading of such can be cancelled in these examples, this is not possible for the identifying reading. 7.4. A contrastive study of English such and Dutch zulk59 Ghesquière and Van de Velde (2011) carried out a contrastive corpus study of English such and Dutch zulk to test the cross-linguistic validity of the proposed pathway of semantic change from identification to nounintensification. Like such, Dutch zulk displays diachronic and synchronic variation between identifying and noun-intensifying uses, as in (263) and (264) respectively. In example (263), identifying zulk invites the reader to phorically link the NP zulke herhaling ‘such repetition’ to the preceding description of the repeated use of the same formula and to conjure up a representative instance of that type of repetition. Like its English counter59. This section is adapted from Ghesquière and Van de Velde (2011).

Such and zulk 235

part, identifying zulk thus realizes type-phoricity (§7.1.1) through the mechanism of generalized instantiation (§7.1.2). In example (264), zulk has a noun-intensifying rather than an identifying function, heightening the degree of the gradable properties inherent in the type description, viz. terribleness and hypocrisy. (263)

(264)

Ik zeg niet dat een schrijver nooit tweemaal of tienmaal eenzelfde formule mag hanteren, indrukwekkende oeuvres werden op zulke herhaling opgebouwd (DBNL 1969) I say not that a writer never twice or ten.times a.same formula may use impressive oeuvres became on such repetition up.built ‘I am not saying that a writer can never use the same formula two times or ten times; impressive oeuvres have been built on such repetition’ We zijn allemaal zulke afschuwelijke hypocrieten. (DBNL 1969) ‘We are all such terrible hypocrites.’

The Dutch data looked at for this study were extracted from three sources. Old Dutch data were extracted from the citation corpus of the Oudnederlands woordenboek [Old Dutch dictionary] (ONW), available online at http://gtb.inl.nl. The ONW is an exhaustive corpus based on a total number of 28,000 citations. For Middle and Modern Dutch, the text collection of the CD-ROM Klassieke literatuur; Nederlandse letterkunde van de Middeleeuwen tot en met de Tachtigers was used (2.5 million words). For Present Day Dutch, a selection of 20th century texts of the Digitale Bibliotheek voor de Nederlandse Letteren [Digital Library of Dutch Literature] (DBNL) (2.3 million words) was queried, available online at http://www.dbnl.org. For English, the same diachronic data set was used as for the study on such. For the Present Day English period, however, only the CB times data were taken into account to match as closely as possible the synchronic data available for Dutch zulk. The main results of this study will be detailed in this section. 7.4.1. Identifying zulk Like identifying such, identifying zulk instructs the hearer to (i) set up phoric relations in the discourse, as well as (ii) conjure up a representative instantiation corresponding to the type description given in the phorically related NP or sentence. As a consequence, identifying zulk displays the

236

Such, zulk and what

same sort of internal conflict between definiteness (in terms of typeidentification) and indefiniteness (in terms of instance-identification). This was already noted by Duinhoven (1988: 128): sulc is dus tegelijkertijd bepaald en onbepaald. Ook al zijn de bedoelde individuen door middel van sulc niet geïdentificeerd, via de aangewezen eigenschap of categorie zijn ze toch min of meer bekend ‘zulk is at the same time definite and indefinite. Even though the intended instances are not identified by means of zulk, they are to some degree known through the characteristic or category pointed at’

The relation of identifying such with indefiniteness was claimed to be clear from its preference for use with indefinite primary determiners and quantifiers (§7.1.3). Similarly, although in Present Day Dutch zulk is largely incompatible with other determiners, it did occur with the indefinite article een ‘a(n)’ until quite recently, as in (265). In plural NPs as in (266) there is no overtly realized determiner, but it is, in the light of the growing incompatibility of zulk with determiners, less plausible to posit the presence of a zero-determiner here than it is in English plural NPs with such. Also illustrating the affiliation with indefiniteness shown by identifying zulk is that, like such, it can be used in existential sentences of the type exemplified in (267). (265)

(266)

(267)

Het is een vrijblijvende mening, en er ligt een groot verschil tussen het geven van zulk een mening, en er ook werkelijk iets voor doen. (DBNL 1969) It is a noncommittal opinion and there lies a big difference between the giving of such an opinion and there also really something for do ‘It is a noncommittal opinion, and there is a big difference between expressing such an opinion and actually acting upon it.’ Ik heb steeds beseft dat Van Gennep een kleine uitgeverij is die zulke processen niet wezenlijk kan beïnvloeden (DBNL 1985) I have always realized that Van Gennep a small editor is that such processes not really can influence ‘I have always realized that Van Gennep is a small editor that cannot really influence such processes.’ Er zijn zulke groepen in een aantal regio’s (http://bdvereniging.nl/beroepsontwikkeling.php) There are such groups in a number regions ‘There are such groups in a number of regions’

Such and zulk 237

7.4.2. Noun-intensifying zulk Like all noun-intensifying elements, noun-intensifying zulk modifies the degree of all gradable type specifications in the NP, inherent in the head noun, in the modifier(s) of the head noun, or in both the head and its modifiers, as in the examples below respectively. (268) (269)

(270)

Zulk onweer en dan alleen in huis (DBNL 1921) Such thunderstorm and then alone at home ‘To be home alone with such a thunderstorm!’ onze maatschappij waarin publiciteit en consumptie zulk een belangrijke rol spelen. (DBNL 1969) our society wherein publicity and consumption such an important role play ‘our society in which publicity and consumption play such an important role’ Rousseau kon in het bestaan van zulk een cynisch monster eenvoudig niet gelooven (DBNL 1919) Rousseau could in the existence of such an cynical monster simply not believe ‘Rousseau could simply not believe in the existence of such a cynical monster’

As for such, the extended scope of noun-intensifying zulk ties in with the diachronic observation that the earliest noun-intensifying uses are attested in NPs consisting only of zulk and a head noun. Only from Late Middle Dutch onwards can additional modifiers intervene between zulk and the head, a construction that later becomes the predominant one in the corpus data (see Figure 22).

238

Such, zulk and what

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

2 30 37 7

zulk + head

15

23

zulk + mod + head

75 31 0

0

1

Figure 22. Modification patterns for noun-intensifying zulk

7.4.3. From identification to noun-intensification ... and back As for such, the quantified data initially support the idea of the historically prior identifying uses of zulk gradually being ousted by the nounintensifying uses. Figure 23 shows that whereas identifying uses of zulk are found already in the Old Dutch data, true noun-intensifying uses are found only from the Middle Dutch period onwards. Admittedly, the data for the Old Dutch period are too limited to draw strong conclusions. The quantified data do clearly show that once the noun-intensifying function is firmly established, it gradually takes up an increasingly large proportion of the prenominal uses of zulk, elbowing out the identifying function. Rather spectacularly, the proportion of noun-intensifying zulk rises from 0.0% in Old Dutch to 52.5% in the Late Modern Dutch period.

Such and zulk 239 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

2

4

6

7

16

0

25

22

8 25

68

105 vague

3

20

39

84

67 73

intensifying identifying

Figure 23. Distributional pattern for the different functions of Dutch zulk

The more recent character of the noun-intensifying uses of zulk is further supported by its changing modification pattern (see also §7.2.2 for such). Whereas noun-intensifying zulk, like noun-intensifying such, originally exclusively modified the head noun, it shifted to predominantly modifying the head of the NP as well as a gradable modifier of the head (Figure 22). This is compatible with a diachronic scenario in which the nounintensifying uses derive from the identifying uses. The rise of the nounintensifying meaning of zulk is not due to the presence of the adjective, but arises naturally, i.e. through pragmatic strengthening (§3.2). Unlike for such, the data show a reversal of the trend towards increasing noun-intensification in Present Day Dutch. Which factors came into play to slow down or even block the increasing frequency of noun-intensifying uses of zulk as observed for English such? A first factor that may have played a role are the different degrees to which such and zulk belong to the determiner category. Ghesquière and Van de Velde (2011) have argued that zulk is closer to the prototypical determiner than such. The authors observed that over time identifying zulk has gradually lost the ability to take modification, as in (271), and to occur as subject complement, as in (272), two characteristics normally associated with adjectives. Moreover, zulk displays increasing incompatibility with other determiners and now typical-

240

Such, zulk and what

ly only precedes secondary determiners such as andere ‘other’, and can no longer follow them as in (273). (271)

(272) (273)

Wanneer ze in den schemer van zulke dagen, dat de geluiden tot stemmen werden, in de kamer zit met stoel en hooge stoof en de aardappelen schilt en vrouw Komeijn zingt de liedjes, die ze altijd zingt ... dan schijnen ook de woorden van de liedjes veel meer dan op andere dagen te beteekenen. (DBNL 1921) When she in the twilight of such days that the noises till voices became in the room sits with chair and high stove and potatoes jacks and miss Komeijn sings the songs that she always sings ... then seem also the words of the songs much more than on other days to mean ‘When she sits in the room with her chair and high stove in the twilight of such days when the noises became voices, and Miss Komeijn sings the songs that she always used to sing ... then the words of the songs as well seem to have more meaning than on other days.’ Si mochten sulc zijn (13th century, Duinhoven 1988: 129) they might such be ‘They might be/may have been such’ ende noch andere sulke sendinge and still other such gifts ‘and such other gifts’ (MNW, s.v. sendinge) (15th century, Van de Velde 2010: 278)

As the core business of prototypical determiners is identification, and zulk is increasingly showing prototypical determiner characteristics, this may explain why zulk once more becomes more involved in identification than noun-intensification. Note that the shift of zulk to the determiner category came to completion in the 20th century (Van de Velde 2010: 293), i.e. at the time when the trend towards increasing noun-intensification was reversed (see Figure 23). According to Ghesquière and Van de Velde (2011), a second factor that may explain why Dutch zulk shows a relapse into the old identification function is that zulk loses ground to competing constructions with zo ‘so’ (see Duinhoven 1988: 131; De Rooij 1989: 199; WNT s.v. zulk; and also Demske 2004 for the idea of competition between solch and so in German). Historically, constructions with zulk such as (274) are increasingly replaced by constructions such as (275). In English, the construction with so is far less frequent than in Dutch. The construction in

Such and zulk 241

example (275), for instance, is not grammatical in English: “[t]he use of so is becoming antiquated with premodifying adjectives, and is already impossible with mass and plural nouns” (Bolinger 1972: 87). In constructions without a modifier (e.g. zo’n man ‘so a man’), the use of so is ungrammatical as well. (274) (275)

zulk mooi weêr (1887, KlasLit) such beautiful weather ‘such beautiful weather’ zo’n mooi weer so a beautiful weather ‘such beautiful weather’

Ghesquière and Van de Velde (2011) consider the gradual ousting of zulk by zo as a possible explanation for the revival of its identifier use as endangered constructions are more likely to survive longer in their older, more prototypical functions than in newer uses. Circumstantial evidence for the idea that threatened constructions revert to older uses comes from the ditransitive construction in Dutch and English, which ranges over an increasingly smaller range of lexical-semantic domains. Interestingly, the construction survives better in its prototypical “caused-reception” use than in its more peripheral “dispossession” use (Colleman and De Clerck 2011). In much the same way, identifying zulk may have endured the competition with the zo’n construction better than noun-intensifying zulk. Of course, at this stage, this explanation remains speculative and is in need of more examples to be bolstered. This case study has shown that, despite the revival of identifying zulk in Present Day Dutch, the development of zulk and such runs largely parallel. The two linguistic items have both identifying and noun-intensifying uses and the latter gradually take up larger proportions of the data and may plausibly derive from the former. In Dutch, the trend toward increasing nounintensification is reversed in Present Day Dutch, which shows that, as predicted by Fischer (1997: 265) and Traugott (2001: 3, 2010b: 275), grammatical and semantic processes of change can be overridden, stopped or reversed at any given moment in their development.

242

Such, zulk and what

7.5. A comparative study of English such and what Bolinger (1972) discusses such and what as the two most prominent elements having undergone a semantic shift from identifier to noun-intensifier. As a consequence, what, like such, can synchronically fulfil both an identifying and noun-intensifying function, as illustrated in (276) and (277). (276) (277)

This year, the boy doll, an Action Man clone says to a Barbie lookalike: `What toys do get in, doll?" And the plaintiff answer comes: `Only toys that make learning fun (CB times) If the publicist is truthful (what a crazy hypothesis!) the response would be: `Tough luck, dearie; perfect buttocks are all you've got. (CB times)

In the literature, including Bolinger (1972), what has, particularly in its noun-intensifying function, received much less attention than such. Altenberg (1994: 224) even restricts his discussion of what to stating that it “poses much the same problems” as such. Similarly, when discussing the internal variation displayed by what, contemporary grammars of English often note the “close parallels in interrogative and exclamatory forms” (Quirk et al. 1972: 927). Rather contradictorily, however, they do treat identifying what and noun-intensifying what as distinct items, belonging to different word classes and fulfilling different functions in the NP. Huddleston and Pullum (2002: 919), for instance, “take interrogative what to be a determinative functioning as a determiner; exclamative what, by contrast, is an adjective functioning as [a] modifier”. In this section, I want to explore on the basis of a synchronic data study the main similarities and differences (i) between the identifying and noun-intensifying functions of what, as well as, more generally, (ii) between identifying and noun-intensifying what and such respectively . For identifying what, its status as primary determiner will be discussed as well as the different types of identification it expresses, viz. indefinite instance-identification, indefinite type-identification, and quantifying identification (§7.5.1). For noun-intensifying what attention will go the different types of degree modification the item can express – qualitative and quantitative, and downscaling and upscaling – as well as to its collocational preferences and its syntactic realization as an intensifying unit together with the indefinite article in singular count NPs (§7.5.2). Finally, three types of contexts will be discussed which illustrate the at times fuzzy boundary between identifying and noun-intensifying realizations of pre-

Such and what 243

nominal what (§7.5.3). Based on these descriptive findings, I aim to gain more insight in the independence of or interconnection between the different uses of prenominal what and see how they relate in the spoken and written varieties of Present Day British English. Finally, the findings for what will be compared to those discussed for such in Sections 7.1 to 7.3 (§7.5.4). 7.5.1. Identifying what: an interrogative primary determiner The identifying function of what is different from that of the identifying elements discussed so far. For the adjectives of completeness and specificity and for such, I have argued that they contribute to the identification of the NP referent by providing detailed quantitative information or phoric instructions to the hearer. In contrast, the main function of what is not to provide identifying information, but rather to signal its absence and request unspecified identifying information, which characteristically has not previously been introduced into the discourse. What thus functions as an indefinite interrogative determiner, prototypically indicating a “request to single out the proper referent” (Langacker 2001: 138) which is not (yet) available in the discourse context as a singled-out entity (Langacker 2002b: 33). Interrogative what can evoke three different types or degrees of indefiniteness. First, as in (278) and (279), what has an instance-oriented use, in which it requests information on specific, concrete instances. Second, what also has a type-oriented, generalized instantiation (Langacker 2005, 2009) use, in which it asks the hearer to conjure up an arbitrary instance of a type retrievable from the discourse context, as in (280) and (281). Here what can be paraphrased as what kind/type of. In this use, identifying what is most similar to identifying such, which can be paraphrased as that/the kind/sort/type of (§7.1.3). Finally, what can also signal the absence of quantitative identifying information, as in (282) and (283). (278) (279) (280) (281)

`What time is it? Five past Kennedy," one quipped. (CB times) One thing the DVLA has done to help such bidding along is to carry out research into what initials or groups of letters occur most frequently in various parts of the country. (CB times) As yet, she doesn't know what car she'll be driving, but it is likely to be a Skoda, Ford or Rover. (CB times) This year, the boy doll, an Action Man clone says to a Barbie lookalike: `What toys do get in, doll?" And the plaintiff answer comes: `Only toys that make learning fun (CB times)

244

(282) (283)

Such, zulk and what

But if the north London venue is chosen as the site of the national stadium, what chance is there of a design of the same calibre as Manchester's? (CB times) people were emotionally attached to dicky birds and daisies and were straining the scientific evidence and didn't mind what economic damage they were doing to the country. (CB ukspok)

Note that identifying what, unlike the identifying uses of the completeness and specificity adjectives and of such, fills the primary determiner slot in the NP (e.g. Seppänen 1978: 526; Huddleston 1984: 233; Huddleston and Pullum 2002: 919). This is clear from the fact that it is mutually exclusive with the indefinite and definite articles (284a) and occurs in coordination with count singular noun phrases (284b). Also identifying what cannot be used predicatively, as illustrated in (284c).60 (284)

a. b. c.

What *the/*a car? What car? *The car is/seems what?

7.5.2. Noun-intensifying what61 7.5.2.1. Semantic-pragmatic types of noun-intensification Semantically, what, unlike the other noun-intensifiers discussed, functions not only as a booster, but also as a downtoner, as in (285) and (286) respectively (§1.3.2.3). Whereas Zanuttini and Portner (2003: 47) state that what indicates that the NP referent “lies at the extreme end of some contextually given scale”, I argue that what locates the degree of a certain property of 60. Predicative uses of interrogative what can be found, but they are infrequent and used only for specific stylistic or expressive purposes, e.g. But what does that mean exactly? The implication is what? That he's 1/2 capitalist and 1/2 socialist. (http://seekingalpha.com/article/242443-fiscally-irresponsible-friday-trading-formagic-beans) 61. Although noun-intensifying what is in the literature consistently referred to as an exclamative marker, I have chosen not to use this label. First, this allows me to be more consistent in the terminology used throughout this book. Second, as with such (fn 54), I believe the implied exclusive link with the exclamative clause type is too strong. As will be illustrated in §7.5.3, intensifying what can also occur in non-exclamative clauses.

Such and what 245

the NP referent either very high (as a booster) or very low (as a downtoner) on an open-ended scale of intensification. In (285), what scales the properties ‘big’ and ‘bad’ upwards from an assumed norm (cf. Quirk et al. 1985: 445), whereas in (286) it emphasizes the small size implications inherent in the descriptive part of the NP a small proportion. (285) (286)

Oh what big bad teeth you've got. (CB ukspok) The other I think er lesson to to draw from from these figures is what a small proportion of output some of the republic actually export or actually receive a abroad proper rather than from er other republics in the old U S S R. (CB ukspok)

As already illustrated by the examples above, noun-intensifying what, like noun-intensifying such, can express both qualitative and quantitative degree modification, as in (287)–(288) and (289)–(290) respectively. The qualitative-quantitative distinction is also noted by Rett (2008) who distinguishes a ‘gradable’ and an ‘amount’ reading. She notes that an exclamative like What languages Mimi speaks! can then be used either “to express surprise at the fact that the languages Mimi speaks are exotic to degree d”, or to express surprise at the fact that Mimi speaks an unexpected number of languages. For whole (§5.3.3.2) and such (§7.2.1), it was argued that in the pattern ‘noun-intensifier + head’ these elements can convey quantitative degree modification by (i) modifying the extent of quantifiers or (ii) by modifying the extent of nouns describing or implying size/amount/extent. With what, however, the quantitatively intensifying reading is most obvious with quantifiers, as in (289) and (290). (287)

(288) (289) (290)

This leads to a repetitious sequence of complaints by him, to her and to us about what bitches women are, and complaints by her about what a slob Harry is, and how he doesn't do what a woman wants done. (CB times) When he turns 18 in 2014, I'll become 50 and my mum will turn 80. What a party we'll have then. (CB times) What a lot of fibs (CB ukspok) But what a lot of time we're wasting, talking about something that anyway was probably just a bit of nonsense. (WB brbooks)

246

Such, zulk and what

7.5.2.2. Collocational behaviour Correlating with the type of degree modification expressed – qualitative or quantitative –, noun-intensifying what typically combines with gradable, evaluative head nouns and/or modifiers (e.g. pity, shame, bitch; wonderful, great, bad), or with quantifiers or modifiers with size or quantity implications (e.g. a lot of, a load of; small, little, big). As is clear from Table 25, when there are modifiers intervening between what and the head noun, the latter is typically non-gradable (see Section 7.2.2 for such). The gradable properties modified by what are then expressed by the intervening modifiers, as listed in Table 26. The frequency lists in Tables 25 and 26 illustrate that, in terms of semantic prosody, noun-intensifying what does not have a clear preference for either purely negatively or purely positively connoted concepts. Note that, in terms of scopal relations, I take what, like all nounintensifiers, to have scope over all gradable type specifications in the nominal description, inherent in either the head noun (291), modifiers of the head (292), or in both the head and modifiers (293). (291) (292) (293)

Champions, with their outsize egos love to rub shoulders with champions and many players have mentioned what a thrill it is to be part of such an exclusive group. (CB times) What a smashing, nimble, corrosively funny production! (CB times) Probably she's got a man out there. It's probable. I don't understand that. What a stupid cow. (CB ukspok)

7.5.2.3. Syntactic behaviour Syntactically, the behaviour of noun-intensifying what is rather noteworthy. Whereas interrogative identifying what is mutually exclusive with the articles, noun-intensifying what always occurs with the indefinite article a(n) in singular count noun phrases.62 As such, one can argue that the determiner 62. As noun-intensifying what systematically occurs with the indefinite article in singular count noun phrases, I believe this warrants positing a zero-determiner ø in uncount and plural noun phrases, as in (i) and (ii) respectively. (i) I THINK IT WAS DELACROIX who remarked on what bad taste one has as a child. (CB times) (ii) But what suggestively formidable students they were! (CB times)

Such and what 247 Table 25. Frequency lists of nouns modified by noun-intensifying what CB times

CB ukspok

pity 

5

what + modifier(s) + head idea 3

shame 

8

what + modifier(s) + head name  4

shame  hand  mistake pleasure shock  bitch

4 3 2 2 2 1

thing  man  taste  view  contrast  arse

3 2 2 2 2 1

load of N  waste  way  face  gent  shock 

4 3 3 2 2 2

idea  sound  thought  lady show time

3 3 3 1 1 1

asylum book boost

1 1 1

sucker  bugger character 

2 1 1

wheeze word accent

1 1 1

what + head

bore 1 charmer 1 chorus 1

what + head

Table 26. Frequency lists of modifiers modified by noun-intensifying what CB times wonderful big great good bad fantastic nice strong appalling arrogant

CB ukspok 6 4 4 3 2 2 2 2 1 1

good wonderful lovely great nice big beautiful brilliant interesting rotten

9 7 5 4 4 3 2 2 2 2

248

Such, zulk and what

complex what a has undergone a process of specialization to the point that it is now “a reliable indicator of an extreme rating” (Potts and Schwarz 2008: 11) or noun-intensifying reading.63 Bolinger (1972: 71) accordingly refers to the determiner a(n) in these units as “the tell-tale indefinite article”, as “[i]ts presence points to the conversion of what to an intensifier”. The establishment of the intensifying unit what a(n) is, according to the literature, a rather recent development, which may be an indication that identifying uses developed prior to the noun-intensifying uses. Rissanen (1999: 207–208) notes that “the inserted article seems to be established in Early Modern English”, but noun-intensifying uses of simple prenominal what in singular count NPs “can be found as late as the eighteenth century”. Also pointing to the identifying origin of the noun-intensifying uses is the fact that what a, like the interrogative construction, is typically restricted to occur in clause-initial position. 7.5.3. Identification and noun-intensification: a fuzzy boundary As for such (§7.3), the data study on what has revealed that the boundary between the identifying and noun-intensifying functions can be rather fuzzy. For what, three contexts can be distinguished in which a nounintensifying reading can be plausibly inferred in addition to the identifying reading. Note that, as for such, the noun-intensifying reading can still be cancelled whereas this is not possible for the identifying reading. This is another indication that noun-intensifying uses of what may diachronically derive from the identifying uses. First, what can receive a noun-intensifying interpretation in subordinate clauses with non-discriminatory licensing verbs such as remember, know, think, or wonder, as in (294) (see Huddleston 1984: 373; Quirk et al. 1985: 1055; Trotta 2000: 119). The syntax of these contexts does not reveal anything about the function of what as both interrogative and exclamative sub63. In a similar vein, Wood (2002: 108–109) has suggested that in contemporary English noun-intensifying sucha is becoming one word, as in it can be confusing with sucha range of cheese. people always know what to look for by the colour. This coalescence and specialization of sucha is, however, not confirmed by the data looked at for this study. No instances of sucha were found in the CB subcorpora, and only two in the WB corpus: (i) It is sucha good job (WB brregnews) (ii) I only worked on telly there for a week, but I got sucha great response. (WB times)

Such and what 249

clauses “have the subject positioned before the predicator” (Huddleston 1984: 373). What in (294) can, in addition to truly questioning what kind of damage has been inflicted on the children (i.e. identifying), be understood as heightening the extent or the amount of damage caused (i.e. intensifying). (294)

He said: ‘One can only wonder what long-term damage you have inflicted on these children.’ (CB times)

A second context illustrating the fuzzy boundary between identifying and noun-intensifying realizations of prenominal what is the what A-er Nconstruction, as in (295) and (296). These constructions are used in syntactically interrogative sentences, but can receive the pragmatic force of superlatives. What better way and what better tribute in (295) and (296) convey that these NPs in fact instantiate the best way and tribute possible. There thus seems to be a kind of pragmatic overriding of the clause type, resulting in a noun-intensifying reading of an identifying construction. (295) (296)

And yet, what better way to appear grown up than to engage in a bitter, pitched court battle with one of the largest entertainment conglomerates on earth? (CB times) Have you ever known anyone as fit as Gordon at 39?" Atkinson was asked. Raquel Welch," he replied. What better tribute. (CB times)

A third and last construction in which identification and nounintensification are inextricably linked is the what the hell-construction, covering combinations such as what the devil, what the deuce, what the fuck, what the dickens, etc. As with the what A-er N-construction, the NP with what is part of an interrogative clause. However, because of the combination with highly expressive words such as hell, the entire construction may receive an emphatic, intensifying reading. The earliest instances of this construction listed in the MED date from the 15th century. The OED mentions uses dating back to the 14th century, albeit without the definite article intervening between what and the head noun, as in (299). (297) (298) (299)

What the hell does he mean by it? (CB ukspok) What the bloody hell was that? (CB ukspok) Hypermestre, What devel have I with the knyfe to doo? (OED s.v. devil 20.a)

250

Such, zulk and what

7.5.4. Such and what: a synchronic comparison In the previous sections, the syntactic, semantic and collocational behaviour of prenominal what in the contemporary data from the CB times and CB ukspok corpora was discussed, both for its identifying and nounintensifying realizations. Table 27 summarizes these findings and compares them to those found for such (§7.1 to §7.3). The schematic summary shows that whereas identifying and noun-intensifying what and such are in the literature often taken to be very similar, they have many idiosyncratic properties. Further diachronic research is, in my opinion, crucial to explain their distinct present day structural-semantic behaviour. Table 27. Schematic summary of the main differences and similarities between prenominal such and what

identifying

such

what

secondary determiner

interrogative primary determiner non-phoric

phoric

nounintensifying

indefinite instance identification definite type identification

indefinite

booster

booster and downtoner specialization of what a

qualitative and quantitative degree modification positive and negative semantic prosody

qualitative and quantitative degree modification positive and negative semantic prosody

Besides their divergent structural-semantic behaviour, such and what also display considerable differences in terms of the proportional distribution of identifying and noun-intensifying uses in the data. Whereas overall there is an almost equal distribution of identifying and noun-intensifying uses of such and what in the CB data (² = 0.542; p = 0.462) (Table 28), this is not the case when the CB times and CB ukspok data are considered separately (Tables 29 and 30). The distribution of identifying and noun-intensifying such and what differs significantly in both the CB times (² = 23.119; p < 0.0001) and CB ukspok data (² = 12.011; p < 0.0001).

Such and what 251 Table 28. Distribution of different functions of such and what in the CB data

such what

n 171 297

identifying % 50.1 47.2

noun-intensifying n % 147 43.1 283 45.0

indeterminate n % 23 6.7 49 7.8

Table 29. Distribution of different functions of such and what in the CB times data

such what

identifying n % 107 65.5 106 40.2

noun-intensifying n % 41 25.2 119 45.1

indeterminate N % 15 9.2 39 14.8

Table 30. Distribution of different functions of such and what in the CB ukspok data

such what

identifying n % 64 36.0 191 52.3

noun-intensifying n % 106 59.6 164 44.9

indeterminate n % 8 4.5 10 2.7

For such the noun-intensifying uses are predominant in the spoken data, whereas in the written data the identifying uses prevail.64 This is not the case for what, which displays a more equal distribution of identifying and noun-intensifying uses both in the spoken and written data.65 Whereas for such the noun-intensifying uses are predominant in the spoken CB data, for what the identifying uses remain predominant. The persistence of identifying what in the contemporary data is due to a number of factors. First, unlike noun-intensifying such, noun-intensifying what developed only in the Middle English period. The degree modifying use may thus still be on the rise and may continue to gain ground at the expense of the identifying uses. 64. A chi-square test shows that these genre-related differences are statistically significant (² = 38,215; p < 0,001). 65. A chi-square test shows that the distribution of the two uses does not differ significantly between the CB times and the CB ukspok data (² = 2,468; p = 0.116).

252

Such, zulk and what

Second, identifying such requires careful textual, discourse-organisational planning on behalf of the speaker and suffers competition from other, more explicitly type-oriented determiner complexes such as the/that kind/type/sort of (§7.1.3). Moreover, identifying such has a preference for more formal genres (Bolinger 1972: 61; Quirk et al. 1972: 279; Altenberg 1994: 238). Identifying what, in contrast, is the default interrogative in all genres and modes and cannot easily be replaced by other interrogative markers. As such, it seems less likely that identifying what would become less frequent in English. 7.6. Conclusions In Sections 7.1 and 7.2 the identifying and noun-intensifying uses of such were discussed in detail. I argued that identifying such conveys typephoricity through generalized instantiation. Together with the indefinite article or zero-determiner, identifying such introduces a new, generalized instance into the discourse, the type specifications of which have to be retrieved from the preceding or following context. Noun-intensifying such expresses either qualitative degree modification, modifying the degree of evaluative notions inherent or implied in the head noun and/or modifiers, or quantitative degree modification. In the latter case, such modifies the extent of a quantifier or of a head and/or modifiers describing or implying a certain notion of size/amount/extent. The development of these different types of quantitative degree modification was most likely driven by collocational extension (§7.2.2). In Section 7.3, I argued that the noun-intensifying uses of such plausibly derive from the identifying uses. Arguments in support of this claim were found, among other things, in the changing distribution and modification patterns of the two uses. The corpus study of such thus confirms the pathway from identification to intensification proposed by Bolinger (1972). Moreover, the data study on Dutch zulk strengthens the cross-linguistic validity of the proposed pathway. As with such, the nounintensifying uses of zulk most likely derive from the identifying uses. Similarly, the synchronic corpus study on what upholds a possible diachronic pathway from identification to intensification. This chapter has thus provided broader evidence in support of the pathway posited by Bolinger (1972).

Chapter 8 Old and little: subjective compounds The corpus studies discussed in Chapters 5 to 7 set out to investigate existing hypotheses on grammaticalization and (inter)subjectification in the NP, as well as Adamson’s (2000) hypothesis that the latter is characterized by concomitant leftward movement of the subjectifying element. In this chapter, the focus is on the second, complementary part of Adamson’s (2000) hypothesis on category shift in the NP, i.e. that rightward movement in the NP – typically driven by lexicalization – entails de-subjectification. Accordingly, the items under scrutiny in this chapter – old and little – were selected because of their proneness to lexicalization. The completeness and specificity adjectives discussed in Chapters 5 and 6 all developed, starting from their descriptive uses, both identifying and noun-intensifying uses, which were described in detail both from a synchronic as well as diachronic perspective. This has allowed me to investigate the directionality of the (inter)subjectification processes in the English NP governing the development of these different uses. Old and little, the adjectives dealt with in this chapter, have not developed noun-intensifying uses and only old has an established identifying use (300). The two adjectives do have descriptive modifier uses – both objective (301a, b) and subjective (302a, b). (300)

(301)

(302)

The party is apparently no longer fielding `prospective parliamentary candidates" for whatever constituency; they are simply to be called `spokespersons" for their constituency. A progressive pressofficerperson explains: `We thought the old title was outdated. Spokesperson is more user-friendly (CB times) a. Like a Welsh terrier with an old sock, Ms Lawley refused to drop the matter. (CB times) b. All in all it is enough to make you feel nostalgic for the days when, instead of the vulgar Gotti chainsaw approach, mafiosi blew their enemies away using sticks of dynamite tied in little bundles with string. (CB times) a. the star must jealously guard his privacy, even if it means not returning phone calls to old friends (CB times)

254

Old and little: subjective compounds

b.

As Mr Blunden remarks in this book, published by Mr Cobden-Sanderson today, ‘memory has her little “ways””. (CB times)

The identifying or secondary determiner use of old, as in (300), appears in the Middle English period (Table 31) and has the meaning ‘former’ or ‘previous’ with respect to the speech event and contrasts with new or present. In its objective descriptive modifier use, as in (301a), attested from the earliest Old English period onwards, old typically attributes the property of “having lived or existed a relatively long time” (OED s.v. old A.I) to the entity referred to by the NP. A subjective descriptive modifier use is exemplified in (302a), in which old describes the relationship between the star and his friends referred to as longstanding. The objective and subjective descriptive modifier uses of little are exemplified in (301b) and (302b). In (301b), little describes the bundles of dynamite as “small in size” (OED s.v. little A.I.1.a). In (302b), little no longer objectively describes the size of the entity. Rather, “some feeling of amusement is involved on the part of the speaker” (OED s.v. little A.I.7.d). The descriptive and identifying uses of old and little will, in contrast with Chapters 5 and 6, not be discussed in detail. In this chapter, the focus will be on lexicalizations with old and little as in (303) and (304). (303) (304)

Didn’t take the Hezbollah long to use the use the cease fire to rearm ... or did you overlook that part, ya pommie old git? (http://scam.com/showthread.php?t=15523) no upstart West Coast bespectacled little twerp was going to take it away from them. (WB usbooks)

This study adopts the definition of lexicalization proposed by Brinton and Traugott (2005: 96): Lexicalization is the change whereby in certain linguistic contexts speakers use a syntactic construction or word formation as a new contentful form with formal and semantic properties that are not completely derivable or predictable from the constituents of the construction or the word formation pattern. Over time there may be further loss of internal constituency and the item may become more lexical.

I will argue that combinations of old/little + noun as in (303) and (304) are the result of such lexicalization processes, more specifically, of compound-

Old and little: subjective compounds 255

ing. Compounding refers to the general process of the unification or univerbation of two or more autonomous words to form a third (Bauer 1983: 11; Brinton and Traugott 2005: 34), whose component parts are no longer fully independent. “Compounds ... are most obviously cases of lexicalization because they involve processes of fusion which serve to erase or efface boundaries between independent morphemes and give rise to unified lexemes over time” (Brinton and Traugott 2005: 44). Importantly, compounding is in this study understood as including not only the formation of compounds through productive compounding processes (e.g. endocentric compounding as in blackboard, footstool; or copulative compounding as in bittersweet, sleepwalk) but also the formation of compounds through repeated use of a sequence of two (or more) originally independent words, as with the development of the old/litte + noun sequences to be discussed here. In this chapter, a number of syntactic and semantic arguments will be provided in favour of a compounding analysis of such adjective-noun combinations as in (303) and (304). The sequences of old/little + noun under scrutiny are of interest to the phenomena central to this study, viz. subjectivity and subjectification in the NP. First, as discussed in Section 2.2, the linear progression of elements in the NP is generally assumed to reflect a continuum of subjective-toobjective meaning (e.g. Quirk et al. 1972, 1985; Dixon 1982; Halliday 1985, 1994; Bache 2000; Huddleston and Pullum 2002). The combinations of old/little + noun exemplified in (4) and (5) challenge this view on the order of elements within the NP. Old in (4) and little in (5) clearly do not describe the objective properties of ‘aged’ and ‘of small stature’, but have a subjective value. They cannot, however, be labelled subjective descriptive modifiers as they do not, for instance, allow predicative alternation or degree modification, without significant changes to the semantics of the original combinations (e.g. *the twerp that is little, *the very little twerp). Also, old and little in examples (303) and (304) are preceded by an adjective whose semantics are generally considered to be less subjective. Modifiers such as pommie and West Coast preceding old and little categorize the NP referents in terms of provenance and are typically analyzed as classifiers and hence as objective. In this respect, examples (303) and (304) appear to challenge the claim that subjective and objective meanings are ordered from left to right in the English NP. Second, as discussed in Section 3.4.3, the assumed synchronic subjective-objective continuum within the NP was rethought by Adamson (2000) as a diachronic cline, predicting the directionality of subjectification in the NP. More specifically, she proposed that elements acquiring a more subjec-

256

Old and little: subjective compounds

tive function will move to the left in NP structure, while elements shifting from a more to a less subjective function will shift to the right. However, I will argue that the subjective, evaluative semantics and divergent syntax of combinations such as old git and little twerp is the result of lexicalization, and more specifically, of ‘subjective compounding’ affecting specific collocations of originally descriptive adjectives and head nouns with evaluative semantics. Although compounding is not traditionally associated with subjectification (Brinton and Traugott 2005: 109), the notion of subjective compounding recognizes the fact that lexical head nouns often contain or imply subjective semantic features which may be modified by elements immediately preceding them, forming a tight unit in a way that resembles the process of compound formation. The functional category of subjective compounds is not (yet) generally recognized in the literature and has only been developed in detail in relation to the adjectives old and little (Van linden and Davidse 2005 on old; Ghesquière 2006 and Ghesquière, Van linden, and Davidse 2013 on old and little). The structure of this chapter is as follows. In Sections 8.1 and 8.2 I will present the diachronic-synchronic case studies of the adjectives old and little. I will argue that both these adjectives are involved in the formation of subjective compounds, in which the adjectives display syntactic and collocational behaviour that is fundamentally different from all their other uses. In Section 8.3, I will briefly discuss the objective compounding in which old and little also engage. In Section 8.4, the development of the subjective compounds with old and little will be confronted with Adamson’s (2000) hypothesis that the de-subjectification of adjectives is accompanied by their rightward movement in NP structure. 8.1. Subjective compounds with old 8.1.1. General characterization The functional category of subjective compounds was first posited by Van linden and Davidse (2005) with regard to examples such as (305) to (306).66 Contemporary data contain quite a number of examples in which 66. Van linden and Davidse (2005) used the term ‘interpersonal compound’, which was later adopted in Ghesquière (2006) and referred to in González-Díaz (2009: 389). Interpersonal is used here in the sense of Halliday and Hasan (1976), i.e. as involving speaker stance. To avoid confusion with De Smet and Verstraete’s

Subjective compounds with old 257

old forms a tight combination with nouns that have either inherently positive or negative connotations, as in (305)–(306) and (307)–(308) respectively. (305) (306) (307) (308)

The contrast is poignant but, I don’t doubt, theatrically calculated by the old master with the exuberant energy and stamina of his dancers. (CB times) Still it hasn’t deterred me, for I was well aware that my opponents were old hands at such manoeuvres. (CB times) The second part of the divorce drama is my bit. Right at the start dad tried to get me involved, to back him up, the old fool, and get me to persuade mum not to do it. (CB ukephem) he was sacked two months later in January last year after allegations that she called the ladies’ captain “an old bitch” and the club officials “a load of old sods” (CB times)

There are both grammatical and lexico-semantic reasons for positing a new functional category to account for adjective-noun combinations as illustrated above. I will first discuss the grammatical arguments for analyzing the old + noun combinations in question as compounds rather than as modifier-head phrase structures. First, old and the following noun conform to Robins’s (1980: 148–154) grammatical criteria for forming one word: they are internally inseparable and function as a unit in NP-structure. Their internal inseparability is clear from the fact that no adjectives can intervene between adjective and noun. For instance, if the ladies’ captain in (308) had been called an old stupid bitch, the nature of the insult would have been quite different. Other modifiers can only precede the unit as a whole, modifying the entire sequence of old + noun, as in (309) and (310). (309)

(310)

A comic reprise of Fifteen Minutes, with elements of Beverly Hills Cop and 48 Hours, it would pair Murphy as a reckless rookie with tough old hand De Niro as stars of a ‘reality’ TV show designed to improve the force’s public image. (WB brbooks) Mr Kennet is a gossipy old woman. Mr Patterson is a gossipy old woman. Mr Jamieson is a gossipy old woman. They’re all gossipy old women. (WB brbooks)

(2006) distinction between ideational and interpersonal subjectivity, I have opted to use the more neutral term ‘subjective compound’.

258

Old and little: subjective compounds

Secondly, the typical grammatical tests for descriptive modifiers – degree modification and alternation with predicative use – do not apply (§1.3.1). It is impossible to grade the adjective old or use it in a predicative construction without changing the specific semantics it invokes. In (307), the father talked about could not be referred to as the very old fool or the older fool, and neither could De Niro in (309) be called a very old hand or an older hand. Likewise, the corresponding predicative constructions – the fool that is old and the hand that is old – are semantically very different from the original constructions with old and little in prenominal position. A subjective descriptive modifier analysis, semantically the only possible alternative, cannot on the basis of these grammatical arguments be maintained. Moreover, the adjective does not have all its systematic paradigmatic variants anymore. For instance, we do not generally speak of a young fool or a new hand. In this respect, the use of old in (303) to (309) differs fundamentally not only from descriptive but also from classifying modifiers, which are typically part of culturally entrenched taxonomies (e.g. old cheese – young cheese, old wine – new wine) (§1.2.2). A third grammatical argument against a modifier-head analysis and for a compounding account comes from the test proposed by Huddleston and Pullum (2002: 1449–1564) for distinguishing phrases from compounds, viz. the pro-one test. We cannot say an old hand and a tough one, or an old bitch and a stupid one, whereas this is possible with both descriptive and classifying modifiers (e.g. a silly man and a mean one, old wines and new ones). A final formal issue to be considered is that of stress. It has often been claimed that compounds can be distinguished from phrases on the basis of stress (e.g. Huddleston and Pullum 2002: 1644), with compounds having forestress, i.e. stress on the first component, and phrases having endstress, i.e. stress on the second component. However, Giegerich (2009, 2011) has convincingly shown that this is a tenacious myth. As first observed by Lees (1963: 120), “within the same dialect, Madison Avenue and apple pie have end-stress while Madison Street and apple cake – forms of identical syntactic behaviour and near-identical semantics – have forestress. It makes little sense to say the former are members of the category NP and the latter of N” (Giegerich 2009: 4). Also, compounds with endstress are neither anomalous nor even exceptional, e.g. town ‘crier, fruit ‘salad, world ‘leader (Giegerich 2009, 2011). These three examples are all listed as separate entries in the OED and Macmillan English Dictionary for Advanced Learners, which supports their status as compound words, stored independently

Subjective compounds with old 259

in the lexicon. The units for which I claim compound status are similar to this latter type in that they all have endstress, e.g. old ‘hag, old ‘chum. Besides their specific grammatical behaviour, the lexical semantics of these old + noun combinations are also a reason for introducing a new functional category to account for them. Old and the noun following it strongly share subjective meaning components (Sinclair 1992; Bublitz 1996). In examples such as (305) to (310), the adjective does not simply add to or restrict the meaning of the noun. “The meaning of the words chosen together is different from their independent meanings. They are at least partly delexicalized” (Sinclair 1992: 16). Besides forming a grammatical unit, adjective and noun thus also form an extended semantic unit. Importantly, the semantic features shared between old and the following noun are subjective, evaluative ones. The combination of old + noun foregrounds these specific evaluations at the expense of the original descriptive meanings, which are backgrounded or bleached. According to Sinclair (1992: 15), semantic feature sharing as it is observed in subjective compounds implies a high degree of collocational cohesion between the linguistic items involved, distinguishing them from modifier-head units with old. Collocational cohesion and, more specifically, the degree to which the occurrence of one word predicts the occurrence of another can be measured by means of mutual information scores (MIscores). “MI is a measure of the strength of association between two words” (Clear 1993: 208). As such, it is the appropriate tool to identify idioms and fixed phrases (Clear 1993: 208–282; Stubbs 1995: 35). A number of combinations of old + noun and their MI-scores in the British sections of the WB corpus are listed in Table 31, together with the MI-scores of some randomly selected combinations of clearly descriptive uses of old and noun. The latter are added to demonstrate the considerable discrepancy in MI-score between combinations like old stagers (12.578), old fogey (11.586) and old hag (10.575), and frequent combinations of old as objective descriptive modifier + head noun, like old buildings (5.875) or old house (4.507). Clear (1993: 279) uses a cut-off point of 6 for what he views as “high MI-scores”. Stubbs (1995: 53) even notes that MI-scores above 3 are “likely to be linguistically interesting”. The adjective-noun sequences studied in this chapter – with old as well as with little – have exceptional MI-scores above or around 8, which greatly supports the proposed analysis of these units as subjective compounds (see §8.2.1 for subjective compounds with little).

260

Old and little: subjective compounds

12.578 12.370 12.276 12.240 12.221 11.963 11.586 11.503 11.464 11.330

old biddies 10 old fuddy-duddies 7 old farts 41 old hag 44 old fogies 8 old chum 104 old bangers 25 old gits 8 old buildings 155 old house 383

3.161 2.644 6.400 6.629 2.827 10.190 4.996 2.826 12.238 18.710

MIscore

Tscore

4.999 2.645 5.567 4.999 5.915 4.898 4.122 7.743 4.241 3.604

Freq.

25 7 31 25 35 24 17 60 18 13

MIscore

Tscore

old stagers old dodderers old codgers old stager old codger old mucker old fogey old banger old biddy old fogeys

Freq.

Table 31. Joint frequency, T-scores and MI-scores for old + noun in British sections of WB corpus

11.300 10.971 10.833 10.575 10.563 10.249 10.206 10.189 5.875 4.507

It has been remarked that because MI-scores ignore absolute frequency, they sometimes single out relatively uncommon combinations one of whose component elements is strongly – or uniquely – associated with the other. To show that this is not the case with the subjective compounds identified Table 31 also includes T-scores, in which the main factor is the absolute frequency of joint occurrences and which measure the productivity of collocations (Stubbs 1995: 33–39). As a rule of thumb, T-scores above 2 are taken to indicate linguistically interesting phenomena. Table 31 shows that although common descriptive modifier + noun sequences generally have a higher T-score than the subjective compounds, the T-scores of the latter are all well above the 2-value. Subjective compounds are thus by no means marginal in terms of productivity. The high degree of collocational cohesion indicated by the MI-scores is a final argument for characterizing these combinations as syntactic and semantic units, i.e. as compounds. Because of the subjective, affective nature of the semantic features shared by the component elements, they are characterized as subjective compounds, distinct from ordinary compounds such as town hall and blackbird with more objective meaning (see §8.3 for objective compounds with old and little).

Subjective compounds with old 261

8.1.2. The development of subjective compounds with old Table 32 represents the chronology and relative proportion of the subjective compounds vis-à-vis the other uses of old in the diachronic dataset.67 Subjective compounds are found in the data only from the Middle English period onwards. One explanation for the relatively late development of subjective compounds is posited by Di Paolo Healey (1997: 44), who claims that “a reader of Old English has a very different concept of ‘old’ than a reader of later English texts”. In her view, this is why “Old English has nothing which might be compared to such disparaging uses” as old codger, old trot, old bag or old trout (Di Paolo Healey 1997: 44). Another reason why subjective compounds are attested only in Middle English is because they are very probably the result of processes of contextual modulation and routinization involving descriptive senses which developed only in Late Old English. First, contextual modulation has been defined by Croft and Cruse (2004: 140) as the activation of semantic features of a word, triggered by the context. The meaning of the word is enriched, as it were, by “specifying features ... contributed by the context” (Croft and Cruse 2004: 140). It is not hard to imagine how old’s subsense ‘of long standing’, as in (3a), was further enriched with notions such as ‘dear’ and ‘close’ when combined with a positively connoted and affectively coloured noun such as chum. The evaluative meaning inherent in the noun foregrounds a similar evaluative meaning in the adjective. Second, routinization is defined by Bybee (2003: 603) as the repetition of a multi-word sequence, which leads to “reanalysis of the sequence as a single processing chunk”. The recurrence of these combinations in our diachronic datasets supports the idea that they were routinized. As a consequence of their repeated co-occurrence, adjective and noun gradually lost “certain specific features of meaning ..., leaving a semantic core” (Bybee 2003: 607). This semantic core is the affectively coloured categorization designated by the subjective compound. In other words, as the result of routinization, new single lexical items with evaluative meaning were formed. An analysis of the old + noun units as subjective compounds is compatible with the fact that a considerable number of them are listed as a separate entry in dictionaries such as the OED and Macmillan English Dic-

67. In addition to subjective compounds, Table 32 also includes information on objective compounds. More details on this category will be given in §8.3.

262

Old and little: subjective compounds

subj./obj descrr.mod.

objective descr. mod.

classifier

objective compound

subjective compound

0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 1.3 1 1.7 6 9.0 5 6.7 6 7.1 5 5.5 6 6.4 5 5.7

0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 3 3.8 0 0.0 0 0.0 2 2.7 0 0.0 1 1.1 1 1.1 3 3.4

0 0.0 0 0.0 2 3.1 0 0.0 0 0.0 1 5.3 3 3.8 4 6.9 3 4.5 7 9.3 3 3.6 1 1.1 2 2.1 0 0.0

3 100.0 9 69.2 54 83.1 9 90.0 27 60.0 11 57.9 44 55.7 30 51.7 38 56.7 36 48.0 56 66.7 44 48.4 51 54.3 43 49.4

0 0.0 1 7.7 8 12.3 0 0.0 13 28.9 5 26.3 13 16.5 9 15.5 11 16.4 2 2.7 3 3.6 9 9.9 13 13.8 12 13.8

0 0.0 3 23.1 1 1.5 1 10.0 5 11.1 1 5.3 11 13.9 13 22.4 5 7.5 19 25.3 14 16.7 24 26.4 21 22.3 20 23.0

0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 1 5.3 4 5.1 1 1.7 4 6.0 4 5.3 2 2.4 7 7.7 0 0.0 4 4.6

total

subjective descr. mod.

n % n 850–950 (HC) % 950–1050 n (HC) % 1050–1150 n (HC) % 1150–1250 n (HC) % 1250–1350 n (HC) % 1350–1420 n (HC) % 1420–1500 n (HC) % 1500–1570 n (HC) % 1570–1640 n (HC) % 1640–1710 n (HC) % 1710–1780 n (CLMETEV) % 1780–1850 n (CLMETEV) % 1850–1920 n (CLMETEV) % 750–850 (HC)

secondary determiner

Table 32.Quantified results of the diachronic corpus study of old

3 13 65 10 45 19 79 58 67 75 84 91 94 87

Subjective compounds with old 263

tionary for Advanced Learners, not as an elaboration of either the noun or the adjective. Mainly three subsenses of old – both more objective and more subjective ones – seem to have fed into groups of what are listed as “special uses” in the OED (s.v. old) from about the 16th century onwards: old in the sense of ‘having the mental or physical characteristics of old age in a negative sense: old codger, old fogey, old trout, old hag, old fool, etc. (ii) old in the sense of ‘knowing, experienced’: old hand, old stager, old master, etc. (iii) old in the sense of ‘acquaintance of old standing’: old chum, old boy, old chap, etc. (i)

These subsenses were contextually modulated by the affectively coloured nouns following them, yielding combinations with strongly negative connotations, as in (i), and positive connotations, as in (ii) and (iii). The Middle English data (1250–1350) contained the first attestations of old + noun that persisted into Present Day English as subjective compounds, viz. old dotard and old lecher. Because of the restricted size of the diachronic corpora consulted and the great variation in spelling in earlier stages of the language, no reliable MI-scores could be calculated for the historical periods studied. A number of combinations of old + noun in the diachronic data can, however, be identified as subjective compounds on the basis of other criteria, such as subjective semantic feature sharing and recurrence in the different periods studied. Interestingly, all Middle English examples appear in contexts of direct speech representation, with the subjective compound used as a term of address, as in (311) and (312). (311) (312)

Treitour! þow olde dote! Þow schelt ben hanged be þe þrote. (HC Bevis, c1330 [?c1300]) ‘Traitor! You old dotard! You shall be hanged by the throat.’ Sire olde lecchour, lat thy japes be! (HC Chaucer, c1390, Canterbury Tales) ‘Sir, old lecher, let your tricks be!’

The negatively evaluative subsense of old that fed into these subjective compounds was ‘having the negative physical and mental characteristics of having lived long’. Such subjective descriptive modifier meanings were first attested in Late Old English. The processes of contextual modulation and routinization operating on such subjective modifier-head structures

264

Old and little: subjective compounds

triggered reanalysis into subjective compounds, which in the dataset were attested a few centuries after the emergence of subjective descriptive modifier uses (see Table 32). This gradual reanalysis entailed all the other changes that led to the distinctive formal and semantic characteristics of subjective compounds, viz. loss of gradability and possibility of predicative use, and reduction of systematic paradigmatic variants such as new or young.68 From the Modern English period onwards, the set of subjective compounds with old becomes more diversified. This diversification may point to the fact that the lexicalized pattern is to some extent becoming a productive word formation pattern. New subjective compounds may then arise through analogy with existing subjective compounds rather than resulting from independent lexicalization and routinization processes. In terms of referential properties, their uses also extend from terms of address, with clear second-person reference (311)–(313), to uses with generic reference (314) (first attestations in 1500–1570), and later to uses with specific third person reference (315)–(317) (first attestation in 1570–1640). (313) (314)

(315) (316) (317)

What haue I stolne fro[m] the or thine: thou ilfauored olde trot. (HC Stevenson, 1551–61, Gammer Gvrtons needle) ‘What have I stolen from you or yours, you ill-favoured old trout!’ (Mage Mumble) “I dyd nothyng but byd hir worke and holde hir peace.” (Tibet Talk) “So would I, if you coulde your clattering ceasse: But the deuill can not make olde trotte holde hir tong.” (HC Udall, a1553, Roister Doister) ‘(Mage Mumble) “I did nothing but bid her to work and hold her peace.” (Tibet Talk) “So would I, if you could cease your clattering; but the devil cannot make old trouts hold their tongue.”’ Upon my life! I believe there is actually some truth in what this old ruffian says. (CLMETEV, Smollet, 1751, The adventures of Peregrine Pickle) Met a lunatic just now. Queer old fish as ever I saw! (CLMET Carroll, 1889, Sylvie and Bruno) But the old vixen has shown her hand, so now he must fight. (CLMET Meredith, 1870, The adventures of Harry Richmond)

68. Interestingly, old and little, which do not serve as systematic paradigmatic variants of each other, have a tendency to compound with the same nouns, e.g. old/little sod, old/little bugger, old/little blighter.

Subjective compounds with old 265

Subjective compounds with a positive connotation, such as old chum and old hand, appear only in the Present Day English data and even then they play a minor part. This also transpired from Table 31, which lists the subjective compounds with the highest MI-scores in the WB corpus: the ones with negative connotations clearly predominate. In the contemporary data, we can witness softening and affective melioration of compounds with negative connotations in cases as (318) to (320). Intrinsically positive compounds, as in (321) to (322), are a minority. (318) (319) (320) (321) (322) (323)

“Shut up maundering, you daft old twat,” he said, almost affectionately. (WB brbooks) “What started your mind moving along those lines?” “Richard, primarily.” “The interfering old sod!” (WB brbooks) Suppose she lives to eighty: does she really want to come back as a wrinkled old hag? (WB brbooks) Gradually the friendship had developed. “You know, Toddy old boy,” Leo would say. “Katrina deserves someone like you. Solid, dependable, reliable (WB brbooks) “You were inexperienced. Voss ... yes ... he should have known better. A terrible risk he took. Madness, really, for such an old hand.” (WB brbooks) You weren't exactly in the Casanova league, old chum , let's face it. (WB brbooks)

The synchronic corpus results are presented in Table 33, in terms of two sets. The data in set (i) are instances of the pattern old + noun, without further elements intervening between old and the head noun. The data in set (ii) are exhaustive samples of instances of the pattern old + one or more modifiers + noun. Of course, subjective compounds were only found in the first dataset in which old immediately precedes the noun. However, the second dataset was added to have a full picture of the relative frequencies of the various prenominal uses of old. When we relate the relative frequency of old in subjective compounds to the proportions of the other prenominal uses of old in the full synchronic sample, we can see that with 2.6% the compounds account for a much smaller proportion of uses than the freely variable combinations in modifier-head structures.

266

Old and little: subjective compounds

subjective compound

total

11

36

14

21

7

118

23.7

0.8

9.3

30.5

11.9

17.8

5.9

37

4

13

43

7

48

4

23.7

2.6

8.3

27.6

4.5

30.8

2.6

objective compound

subj. descr. mod.

1

classifier

sec. det./ descr. mod.

28

obj. descr. mod.

sec. det.

Table 33. Quantified results of the synchronic corpus study of old

(i) old + noun CB times

n

CB ukspok

n

% %

156

(ii) old + adj (+adj) + noun CB times CB ukspok CB total

n

29

4

8

39

1

0

0

%

35.8

4.9

9.9

48.1

1.2

0.0

0.0

n

33

0

15

28

0

0

0

%

43.4

0.0

19.7

36.8

0.0

0.0

0.0

n

127

9

47

146

22

69

11

%

29.5

2.1

10.9

33.9

5.1

16.0

2.6

81 76 431

8.2. Subjective compounds with little 8.2.1. General characterization In contemporary English, we find combinations of little + noun, illustrated in (324) to (328), which are grammatically and semantically very similar to those discussed in Section 8.1 for old + noun and which can on the basis of the same arguments be analyzed as compounds rather than modifier-head sequences. (324) (325)

“I know the way, Turon. I hardly need an obsequious little toad like you to guide me” (WB brbooks) Like the backstage essays of David Mamet, these are little gems of practical experience: no luvvie gossip, just a brief guide from one of the great masters, and even a diagram or two. (CB ukspok)

Subjective compounds with little

(326)

(327)

(328)

267

How touching it is to watch the mothers collect their little ones at the end of another day’s hard learning

Come here you little bleeder, before I fg kill yer

What child worthy of the name could resist such an invitation? (CB ukspok) The hilarious sequel to ‘How to be a “Little Sod”’, follows our miniature despot’s development into tyrannical toddlerhood – complete with tantrums, potty training and unidentifiable rashes, abundantly illustrated with cartoons. (WB brephem) “He was a little bugger as a baby, screaming all the time” Clarke Sr once said. “That was until he was circumcised. That quietened him down a bit.” (CB times)

Firstly, unlike modifier-head structures, the little + noun combinations in (324) to (328) are internally inseparable: no qualitative or classifying modifiers can come in between them without fundamentally changing their meaning, e.g. little diamond/ informative gems of practical experience (325), little criminal/exhausting bugger (328). It can also be noted that in, for instance, a little exhausting bugger, little no longer shares the specific affective colouring of bugger, and bugger on its own has a different affective value than little bugger. The pro-one test (Huddleston and Pullum 2002: 1449–1564) likewise identifies the units as compounds rather than phrases. One cannot, for instance, speak of a little sod and a cranky one. Secondly, the syntactic unity of the little + noun sequences is also confirmed by the fact that they can be prefaced by descriptive modifiers that modify the sequence as a whole, as in the NP an obsequious little toad in (324). Moreover, these modifiers can be coordinated with other modifiers, e.g. a tired and cranky little bugger, or be stacked with other recursive modifiers, e.g. poetic literary little gems. Thirdly, in these combinations, the adjective little has lost its ability to be graded and to be used predicatively. For example, a very little bugger and a bugger that is little do not have the specific semantics of the subjective compound little bugger. Fourthly, the little + noun units do not contrast with the systematic paradigmatic variants of little. For instance, corresponding to little bleeder we do not find great bleeder or big bleeder. Semantically, finally, adjective and noun clearly share subjective meaning components (Sinclair 1992; Bublitz 1996). The evaluative features inherent in the noun foreground corresponding features in little. The sequence little bugger, for instance, consists of the evaluative noun bugger, which is used in informal language “to describe a person who has done something

268

Old and little: subjective compounds

annoying or stupid” (Sinclair 1990) and little, whose negative affective features of irritation, contempt, etc. are pulled to the fore as well in this combination. As was the case with the subjective compounds with old, these fixed collocational units are characterized by much higher MI-scores than even very common modifier-head structures. Table 34 shows the MIscores, T-scores and joint frequencies of a number of little + noun sequences as they occur in the British sections of the WB corpus. For the unit little bleeder, for instance, the MI-score is 10.33, indicating very strong internal cohesion, but its T-score is 2.234, reflecting moderate productivity. By contrast, a common objective attribute-head combination like little pieces with T-score 8.359 has an MI-score of only 4.85.

11.081 10.965 10.702 10.692 10.506 10.487 10.380 10.333 10.158 9.763

         

MIscore

MIscore 

3.604 3.740 3.160 4.896 1.413 3.739 3.314 2.234 2.826 6.626

Tscore

Tscore

13 14 10 24 2 14 11 5 8 44

Freq.

Freq.

little Hitlers little blighters little blighter little minx little bleeders little tyke little twerp little bleeder little tykes little darlings





Table 34. Joint frequency, T-scores and MI-scores for little + noun in British sections of WB corpus

little pipsqueak little scamp little buggers little stirrer little bugger little smasher little twat little fucker little pieces little shop

3 4 20 3 46 3 4 9 75 74

1.730 1.997 4.464 1.729 6.765 1.728 1.994 2.990 8.359 8.182

9.421 9.265 9.148 8.965 8.614 8.573 8.365 8.271 4.847 4.357

8.2.2. The development of subjective compounds with little Subjective compounds with little start to appear with a certain degree of regularity only from the Present Day English period onwards. This is shown by Table 35, which gives the absolute and relative frequencies with which little is attested in its various prenominal functions in the different historical periods.

Subjective compounds with little

269

subj./obj descr.mod.

objective descr.mod.

classifier

objective compound

subjective compound

total

n

0

0

45

0

2

0

0

47

%

0.0

0.0

95.7

0.0

4.3

0.0

0.0

1050–1250 n (HC) %

0

0

36

1

1

0

1

0.0

0.0

92.3

2.6

2.6

0.0

2.6

3

12

38

1

3

1

0

5.2

20.7

65.5

1.7

5.2

1.7

0.0

2

4

36

1

2

4

0

4.1

8.2

73.5

2.0

4.1

8.2

0.0

2

6

34

0

0

3

1

4.3

13.0

73.9

0.0

0.0

6.5

2.2

6

5

48

0

0

8

2

8.7

7.2

69.6

0.0

0.0

11.6

2.9

7

3

38

0

2

6

1

12.3

5.3

66.7

0.0

3.5

10.5

1.8

750–1050 (HC)

1250–1500 n (HC) % 1500–1710 n (HC) % 1710–1780 n (CLMETEV) % 1780–1850 n (CLMETEV) % 1850–1920 n (CLMETEV) %

obj. descr. modifier/ classifier

subjective descr.mod.

Table 35. Quantified results of the diachronic corpus study of little

39 58 49 46 69 57

An individual example of what looks like a subjective compound is attested as early as Early Middle English and the odd subjective compound also occurs in Late Modern English: (329) (330)

How go you on with the amiable little blot? (CLMETEV Chesterfield, 1751, Letters to his son) I saw they had never laid down, though it was past midnight; but they were calmer, and did not need me to console them. The little souls were comforting each other with better thoughts than I could have hit on. (CLMETEV Brontë, 1847, Wuthering Heights)

Table 36 gives the quantified results of the synchronic data study of little, again in terms of the two syntactic patterns as selected for old. It shows that, in the contemporary data, subjective compounds with little, like those with old, form only a small fraction of the prenominal uses (1.8%).

270

Old and little: subjective compounds

(i) little + noun CB times CB ukspok







CB ukspok CB total

total

subjective compound

objective compound 



n

48

6

223

21

73

%

12.4

1.6

57.8

5.4

18.9

n

27

20

313

10

55

%

6.3

4.7

73.0

2.3

12.8

(ii) little + adj (+ adj) + noun CB times

classifier

obj. descr. mod.



subj. descr. modifier



subj./ obj. descr. mod.

Table 36. Quantified results of the synchronic corpus study of little

15 386 3.9



4 429 0.9





n

96

14

106

2

11

1 230

%

41.7

6.1

46.1

0.9

4.8

0.4



n

15

2

72

4

1

0

94

%

16.0

2.1

76.6

4.3

1.1

0.0



n

186

42

714

37

140

20 1139

%

16.3

3.7

62.7

3.2

12.3

1.8



Diachronically, the subjective compounds with little, like those with old, are most likely the result of routinization processes, whereby the modifierhead sequences are through repeated use reanalyzed as ‘single processing chunks’ (Bybee 2003: 603). Likewise, the eventual subjective meanings of the compound units arise through contextual modulation (Croft and Cruse 2004: 140): the negative evaluative meaning inherent in the noun foregrounds a similar evaluative meaning in the adjective. Although in the Present Day English data some subjective compounds with little are found that express positive feelings of the speaker (e.g. little gem, little darlings), they predominantly display a rather negative semantic prosody, most often featuring nouns like bugger, creep, monster, bleeder, blighter and sod. Note that, as with the subjective compounds with old, the subjective compounds with little in the large majority of cases refer to human beings rather than material objects. Also noteworthy is the fact that in some cases, the adjec-

Subjective compounds with little

271

tive mitigates the negative connotations of the head noun, making the units less face-threatening. Bolinger (1967: 59) and González-Díaz (2009: 383) have rightfully observed that with some of the compounds little has an effect similar to that of the diminutive suffix. As noted by González-Díaz (2009: 383), “‘[r]ightmost’ little conveys nuances of affection (as opposed to dimension)” which are “translated into other Germanic languages, like Dutch, by an affective diminutive suffix”. Thus, little gems in (325) would be translated in Dutch as juweeltjes (lit. juweel ‘jewel’ + diminutive suffix + plural), and little sod in (327) as ettertje (lit. etter ‘sod’ + diminutive suffix). The mitigating, affective meaning of little foregrounded in the subjective compounds plausibly derives from the subjective descriptive meaning illustrated in (302b), in which the adjective is “used to convey an implication of endearment or depreciation, or of tender feeling on the part of the speaker” (OED s.v. little A.I.3). 8.3. Objective compounds with old and little Although for a study on (inter)subjectification processes in the NP subjective compounding is, of course, more relevant and interesting, I briefly want to draw attention to the fact that old and little engage not only in subjective compounding but also in objective compounding (see also González-Díaz 2009: 388–389). Interestingly, these compounds are again predominantly used with reference to people. Some examples are given in (331) to (334). (see Tables 32, 33, 35 and 36 for information on the relative frequencies of these objective compounds with old and little) (331) (332) (333) (334)

I can’t do it, old man; or I would, I presume, if I’d been made that way. (CLMETEV Kipling, 1897, Captains courageous) A youth of frolics, an old age of cards. (CLMETEV Pope, 1733– 1734, An essay on man) We haven’t got anything for you, little girl. Be off! (CLMETEV Dickens, 1848, Dombey and son) In the will of Nicholas Gimcrack the virtuoso, recorded in the Tatler, we learn, among other items, that his eldest son is cut off with a single cockleshell for his undutiful behaviour in laughing at his little sister whom his father kept preserved in spirits of wine. (CLMETEV Hazlitt, 1821–1822, Table talk)

272

Old and little: subjective compounds

As with the subjective compounds, these sequences of old/little + noun are firmly established in English and processed as single units. However, unlike the subjective compounds, objective compounds are not always characterized by high MI-scores. In contrast, as shown in Table 37 for old and in Table 38 for little, they do tend to have high T-scores, reflecting their frequency and strong collocational ties in English. Table 37. Joint frequency, T-scores and MI-scores for objective compounds with old in British sections of WB corpus  old man old age old lady old woman old days old people old men

Frequency 4074 1784 1093 1091 1049 1157 570

T-score 63.295 42.014 32.999 32.576 31.771 31.686 22.893

MI-score 6.903 7.560 9.068 6.184 5.713 3.869 4.605

Table 38. Joint frequency, T-scores and MI-scores for objective compounds with little in British sections of WB corpus  little girl little boy little girls little sister little ones little brother little boys

Frequency 2310 1497 514 306 302 276 274

T-score 47.887 38.510 22.418 17.272 17.068 16.284 16.229

MI-score 8.098 7.737 6.485 6.308 5.808 5.656 5.676

Further evidence for the compound status of the adjective-noun sequences is found in the fact that they are often listed as separate entries or special units in dictionaries, with mention of their specialized meanings. Objective compounds such as little boy and little girl can refer not only to young people in general but, more specifically, to someone’s son or daughter, as in (335). Similarly, the compound old man has acquired the specialized meaning of father, as in (336). (335)

She allows her little girl to have tantrums in public, and she is amazingly rude to the family. (WB times)

Objective compoundswith old and little

(336)

273

But I’m put in mind of what my old man said about Britain during and after the Second World War. “For as long as it lasted, everyone put aside their differences and helped one another.” (WB times)

Translations provide additional evidence for the single word status of the objective compounds. As noted for subjective compounds such as little sod, a number of objective compounds with little can also be translated by means of a diminutive suffix. Little girl and little boy, for instance, are best translated as meisje and jongetje in Dutch. The objective compounds found for old also have one-word counterparts in other languages. As noted by Wierzbicka (1986: 368), the objective compound old man is translated as vieillard in French and starik in Russian. Similarly, old age is best translated as vieillesse in French and ouderdom in Dutch. Note that, as with old, it is the age rather than the size meaning of little which is foregrounded in the objective compounds. That exactly this age meaning is prone to objective compounding might be due to the fact that “for human beings age tends to be treated as a crucial determinant ... rather than as one feature among many” (Wierzbicka 1986: 368). Additional grammatical arguments for treating the adjective-noun sequences as compounds are their unit status and internal inseparability. One can talk about the good old days or a pretty little girl, but not the old good days or a little pretty girl without changing the meaning of the sequence. Similarly, as with the subjective compounds, the pro-one test (Huddleston and Pullum 2002: 1449–1564) is infelicitous with objective compounds (*a little girl and a beautiful one, *an old man and a strange one). 8.4. Subjective compounds and their repercussions for subjectivity and subjectification in the NP On the basis of the specific grammatical and semantic characteristics of the old/little + noun combinations discussed in Sections 8.1 and 8.2, I argue that they are best analyzed as the outcome of processes of lexicalization and compounding. The original descriptive modifier-head phrase is reanalyzed as a compound lexeme, which is “no longer ‘freshly’ assembled from its constituent parts on each occasion of its use” (Himmelmann 2004: 37). Rather, the adjective-noun combinations are stored independently in speakers’ lexicons as new, contentful lexical items. This is reflected in the fact that many of them are listed as separate entries in dictionaries.

274

Old and little: subjective compounds

Importantly, as noted by Bauer (1983: 55–61), different types of lexicalization have to be distinguished, amongst others semantic and syntactic lexicalization. They can occur independently, but “it rather seems that once a form is lexicalized in one way it is easier for it to become lexicalized in others” (Bauer 1983: 61). First, semantic lexicalization is “frequently discussed in terms of lack of semantic compositionality, i.e. the meaning of the whole is not predictable from the meanings of the part” (Bauer 1983: 58). For the adjective-noun combinations with old and little we noted that their semantics is different from that of the corresponding descriptive modifier uses. Nevertheless, it seems that some subjective compounds display a higher degree of semantic lexicalization than others. The semantic compositionality of old hand or old woman (with reference to men, as in [310]) is less clear than that of old chum or little bugger. This need not, however, be problematic if lexicalization is seen as a gradual process with intermediary stages (cf. Bauer 1983: 45–50; Brinton and Traugott 2005: 45–47). Second, syntactic lexicalization refers to “lexicalization which shows up in the way the complex form interacts with other items in the sentence” (Bauer 1983: 59). This type of lexicalization is very prominent in compounding. As discussed in the preceding two sections, the syntactic behaviour of the subjective compounds with old and little is very different from that of ordinary modifier-head phrases. No modifiers can intervene between old/little and the noun and the adjective-noun sequence can only be modified in its entirety. Also, the subjective compounds do not allow for the pro-one test proposed by Huddleston and Pullum (2002: 1449–1564) to distinguish phrases from compounds. Moreover, the adjectives old and little have lost the ability to be modified in terms of degree and to be used predicatively. What does the notion of subjective compounding entail for existing hypotheses on subjectivity and subjectification in the English NP? Synchronically, subjective compounds show that an account of ordering within the NP in terms of a linear progression from subjective to objective meaning is untenable. Other phenomena challenging the linear subjective-objective model were discussed in Section 2.2.7. They include idiomatic collocations such as the big bad wolf, gradable nouns or degree nouns with inherent evaluative semantics (e.g. idiot, bliss), degree modifiers following the elements they modify (e.g. not a bad old stick) and scopal adjectives (e.g. those Aladamnbama farmers). The other main issue to consider is the challenge the development of subjective compounds poses for Adamson’s (2002) hypotheses that subjectification processes are accompanied by leftward movement in the NP, while de-subjectification is characterized by rightward movement. In Chap-

Subjective compounds, subjectivity and subjectification

275

ters 5, 6 and 7 empirical evidence was given for Bolinger’s (1972) proposed pathway leading from identifying to noun-intensifying meaning. These shifts witnessed for whole, particular, such and zulk contradict the subjectification-cum-leftward movement hypothesis, asidentifying secondary determiner meanings are situated at the leftmost determination zone of the NP, whereas noun-intensifying uses are situated more to the right in the (degree) modification zone (§1.1.4). The processes of change by which subjective compounds emerge show that Adamson’s claim of desubjectification always entailing rightward movement also cannot be maintained. Historically, subjective compounds originate in descriptive modifier-head structures. The adjectives old and little clearly travel right in NPstructure to form subjective compounds. A prosodic, field-like model of the NP as proposed in Section 2.3, I believe, captures the diachronic subjectification trends more accurately than the strictly linear left-right model. Pike (1972a: 129) argued that language as a whole and its subsystems can be viewed as ‘fields’, i.e. “as a system with parts and classes of parts so interrelated that no parts occur apart from their function in the total whole”. Applying Pike’s insight to the NP, Halliday (1994: 190) has argued that, in a field-like model, subjectivity is viewed as prosodically running all the way through the NP, possibly interspersed with objective meaning. The degree of subjectivity does not increase linearly in the NP. Rather, there are various subjective elements spread over the whole of the NP, which may attract items into new subjective word formations, or into new (sub)modification clusters in which the more subjective element may either precede or follow the less subjective one. A precise model of subjectification in the English NP has to account for the multiplicity of mechanisms that can accompany subjectification such as leftward, and occasionally rightward, movement, reanalysis of parataxis into hypotaxis, new word formations, etc.

Summary Towards a reconciliation of synchrony and diachrony It’s not just a question of conquering a summit previously unknown, but of tracing, step by step, a new pathway to it. Gustav Mahler The focus of this study has been on the directionality of change and the underlying formal and conceptual mechanisms of the development of identifying and noun-intensifying uses of different sets of linguistic elements. From the same perspective, subjective compounds were studied. Obviously, when studying structural-semantic changes in the NP, it is crucial to have a good understanding of the synchronic and diachronic build-up of the NP. Building on the accounts developed by Langacker (1991), Halliday (1994) and Bache (2000), I proposed in Chapter 1 a functional-cognitive model of the English NP which can account for the synchronic and diachronic dynamicity of its component elements. The proposed model consists of different functional-structural zones each comprising a number of functional slots. Importantly, these slots are defined functionally, as form-meaning pairings, and their succession in NP structure as well as their synchronic and diachronic interrelation is understood as a gradient or continuum without strict or discrete boundaries. The three main zones distinguished are from left to right the determination zone, the modification zone and the categorization zone. The determination zone accommodates all elements concerned with the identification and quantification of the NP referent, whether realized as primary or secondary determiners. The modification zone comprises two subzones, i.e. the degree modification zone and the descriptive modification zone. Descriptive modifiers attribute a certain quality or property to the instance referred to by the NP, either on an objectively verifiable basis or based on the speaker’s subjective evaluation of the NP referent. The degree modification zone comprises adjectiveintensifiers, which modify the extent of properties denoted by descriptive modifiers, and noun-intensifiers, which modify the extent of all gradable

278

Summary

type specifications in the NP expressed by the head and/or modifiers. The categorization zone, finally, groups together the head noun – simple or compound – and classifying modifiers, which (sub)categorize the type of which the NP referent is an instance. The main concerns of this study were with the secondary determiner function of the determination zone and the noun-intensifying uses of the degree modification zone. Discussion of the different slots in terms of their relation to (inter)subjectivity revealed that the component elements of the NP display a range of different types of subjectivity and intersubjectivity. Secondary determiners were argued to convey (textual) intersubjective meaning as they explicitly steer the hearer in his interpretational task. Descriptive modifiers and degree modifiers were argued to convey ideational subjective meaning, as they convey descriptive content but at the same time express the speaker’s subjective attitude toward the NP referent (De Smet and Verstraete 2006). However, I argued that although basically ideational, descriptive and degree modifiers convey still different subtypes of subjectivity. The variety of subjective and intersubjective meanings was discussed in more detail in Chapter 2. A number of different views on the ordering of elements within NP structure circulate in the literature, but the most widely accepted one is that there is a linear progression from subjective to objective meanings. I argued that although this view certainly has its merits and captures general ordering tendencies, it also has its shortcomings. First, the envisaged subjective-objective continuum is too general to account for the semantic variation displayed by prenominal elements in the NP. Following De Smet and Verstraete (2006), I distinguish between ideational and interpersonal subjectivity, whereby the former conveys both descriptive content and speaker attitude and the latter is restricted to the enactment of speaker position. Second, confrontation with identifying meanings as they are conveyed by determining elements in the NP indicated that the notion of intersubjectivity as it was put forward in, for instance, Traugott (2003) was in need of refinement. Secondary determiners are a means for the speaker to direct the hearer in his/her interpretational task and to coordinate a joint focus of attention (Diessel 1999, 2006). In line with Breban (2010a) and Carlier and De Mulder (2010), I argued that such elements code textually intersubjective meaning. In Traugott’s work intersubjectivity is restricted to meanings encoding the speaker’s attention to the self-image or face needs of the hearer. I advocate a broader understanding of intersubjectivity as encompassing not only such attitudinally intersubjective meanings but also textually intersubjective meanings which steer interpretation by the hearer. The distinction between textual and attitudinal meaning was already present

Reconciliation of synchrony and diachrony

279

in Traugott’s (1995) definition of subjectification as “the tendency to recruit lexical material for purposes of creating text and indicating attitudes in discourse situations” and is now extended to the domain of intersubjectivity. Accordingly, in contrast to Traugott (2010a), Narrog (2010, 2012) and Breban (2010a), I view textual elements as intersecting with the entire objective-subjective-intersubjective spectrum. Third, the subjective-toobjective ordering principle cannot account for the (minor) counter-current of more subjective elements following less subjective ones. Examples are degree modifiers which follow the elements they modify (González-Díaz 2009), degree nouns with inherent evaluative meanings (Bolinger 1972; Paradis 2000) and the subjective compounds discussed in Chapter 8. In the last section of Chapter 2, I proposed a possible alternative view on ordering within the English NP, i.e. a prosodic field-like model as posited by Pike (1972a, 1972b) and further developed by Halliday (2002d). In such a model, subjective meanings are naturally spread over the entire NP and may intersperse with non-subjective meanings. The ultimate overall ordering of elements within the NP is then the result of different semantic and grammatical dimensions (e.g. subjectivity, degree modification, information structure) all meshed together into the NP system and constrained by NP structure. In Chapter 3, I turned from a synchronic approach to the NP to a diachronic perspective looking into how structural-semantic changes proceed within the NP. In particular, I directed my attention to the development of identifying secondary determiner uses and noun-intensifying uses. For secondary determiner adjectives, Davidse, Breban, and Van linden (2008) have posited a deictification process whereby the general relation depicted by the originally lexical adjective acquires a subjective reference point in or related to the speech event. For noun-intensifying elements, two distinct pathways of change have been put forward in the literature. In recent literature, intensifying meanings are generally envisaged to originate in fully lexical, descriptive uses (e.g. Adamson 2000; Paradis 2000; Athanasiadou 2007; Traugott 2006, 2007a, 2007b). Earlier, Bolinger (1972) had posited a now largely neglected pathway from identifying meanings – either purely identifying or quantitatively identifying – to noun-intensifying meanings. In Part III of this study, the validity of the proposed pathways was assessed against corpus data. First, the case studies discussed in Chapters 5, 6 and 7 have provided evidence in support of the deictification account of the development of secondary determiner adjectives (Davidse, Breban, and Van linden 2008). The identifying uses of all three completeness adjectives were argued to have plausibly derived from descriptive modifier uses. In

280

Summary

this shift, the objective reference point of the descriptive modifier uses, i.e. the boundary or maximal extension of the NP referent was transformed into a subjective reference point, i.e. the full contextual realization of the NP referent. For particular and specific, I posited that a deictification account is also plausible, involving a shift from the expression of qualitative specificity as descriptive modifiers to the expression of referential specificity as secondary determiners. The subjective reference point is then the unfolding discourse itself. Second, the corpus studies of the completeness and specificity adjectives and the study of such, zulk and what have shown that the development of noun-intensifying meaning can proceed not only along the pathway of change now prevalent in the literature, i.e. from description to nounintensification, but also along the pathway put forward by Bolinger (1972) some 40 years ago, i.e. from identification to noun-intensification. The noun-intensifying uses of complete and total were argued to originate in descriptive modifier uses. Importantly, their development confirms Paradis’s (2000) hypothesis about the foregrounding mechanism of change involved: the boundedness construals of the lexical uses are foregrounded in the noun-intensifying uses. For complete, the bounded descriptive use ‘realized to its full extent’, as in the complete establishment, developed into a closed scale noun-intensifying use, as in a complete failure. The bounded descriptive use of total ‘complete in extent or degree’, as in almost total seclusion, also lies at the origin of closed scale noun-intensifying uses as in total debility. In parallel fashion, the descriptive ‘accomplished’ sense of complete, which hovers between bounded and unbounded construal, gave rise to noun-intensifying uses of the type a complete idiot, which similarly hover between a closed scale and an open scale reading. Whereas complete and total followed the path leading from description to nounintensification, for whole, particular, such and zulk I argued that their noun-intensifying uses most likely derive from their identifying secondary determiner uses. The same pathway was hypothesized for what on the basis of synchronic data. Interestingly, the case studies evidenced shifts both from purely identifying meaning to noun-intensification (particular, such and zulk) as well as from quantifying-identifying meaning to nounintensification (whole). As hypothesized in Chapter 3, shifts from identifying to noun-intensifying meaning can also be captured by the foregrounding mechanism to which Paradis (2000) called attention. However, her original formulation, which dealt only with the foregrounding in the intensifying use of the (un)bounded nature of the lexical source, has to be modified and extended considerably to account for quantifying-identifying and

Reconciliation of synchrony and diachrony

281

purely identifying sources of noun-intensifying uses. In the case of quantifying-identifying meaning the inherent quantification scales are foregrounded and reinterpreted as intensification scales. The universal relative quantification inherent in the secondary determiner use of whole, for instance, was reconfigured as a closed intensification scale resulting in nounintensifying uses of the type the whole world. Driven by collocational extension, these closed scale noun-intensifying uses later gave rise to open scale noun-intensifying uses as in a whole lot of people. In the case of purely identifying meanings there are no inherent scalar notions which can be foregrounded and semanticized. Rather the intensifying, scalar semantics are derived pragmatically. For noun-intensifying particular the data study showed that what triggered the intensifying reading was the presence of a comparative effect, either with respect to different entities possibly functioning as reference point for NP identification (e.g. at my own particular request) or with respect to different possible referents of the NP with particular itself (e.g. no particular hurry). For such, pragmatic strengthening was also seen as lying at the basis of the development of noun-intensifying uses from identifying uses. The corpus studies have shown that there are two valid pathways leading to noun-intensifying meanings, one starting from descriptive meanings and one starting from identifying meaning. The question does arise, however, why elements develop along the one or the other pathway. Such and zulk have no descriptive uses so the first pathway obviously never was an option for them. In contrast, the adjectives complete, total, whole and particular all have both descriptive as well as identifying uses. Yet, whereas the former two developed noun-intensifying uses from their descriptive meanings, the noun-intensifying meanings of the latter two derive from their identifying meanings. One could argue that Bolinger’s pathway from identification to intensification applies primarily to grammatical elements, like such and zulk, and less so to originally lexical elements like complete and total. This does not, however, explain why the development of whole and particular did proceed along the pathway envisaged by Bolinger (1972). Another factor distinguishing between the items studied is their origin. Whereas zulk, such and whole are originally Germanic and are attested from Old Dutch and Old English onwards, complete and total are Romance loanwords. Particular, however, is also a Romance loanword and appears in the data at around the same time as complete and total, yet it followed a different developmental pathway. At this point, I do not have an adequate explanation of why a particular linguistic item travels along either one or the other pathway toward noun-intensification.

282

Summary

What is important is that despite the specific sense strains, functions and collocational preferences of each individual item which influence the path to be followed, the corpus studies charting the development of identifying and noun-intensifying uses of prenominal elements in the NP have revealed recurrent pathways of change. As Traugott (2010a: 42) noted, “each construction has its own history, but conforms to general schematic changetypes in ways that are partly constrained by the particularities of the original meaning-form relationship”. At this point, I want to draw attention to some striking observations concerning the intersubjectification and grammaticalization processes studied. Firstly, as discussed in Section 3.2, traditional theorizing on these processes typically assumes that their onset is retraceable in diachronic data in the form of ambiguous contexts, either purely semantic or semantic and structural (e.g. Evans and Wilkins 2000; Heine 2002; Diewald 2002, 2006). In line with recent observations made by Traugott (2008, 2010c, 2010d), no such systematic ambiguous contexts were found triggering the development of the noun-intensifying uses of the items studied. Rather, their development was argued to typically be triggered in ‘specialized contexts’ characterized by semantic, morpho-syntactic and or pragmatic features. For the development of the closed scale noun-intensifying use of whole, for instance, I identified the all the whole N-construction as an important trigger. For the shift from identifying to noun-intensifying particular, NPs with a possessive determiner and the emphatic possessive own, as in at my own particular request, were argued to have served as trigger contexts. Secondly, collocational contexts or collocational patterning was shown to play an important part in the development of all noun-intensifying uses. The deictification processes posited for the completeness and specificity adjectives did not display any notable collocational overlap between the descriptive source meanings and the identifying target meanings. One exception is the development of secondary determiner complete, but here the original collocational range quickly widened and the use of identifying complete is now almost unrestricted in terms of collocation. In contrast, the development of the noun-intensifying uses of the completeness and specificity adjectives proved to be strongly collocationally driven, with considerable collocational overlap between source and target meanings. The noun-intensifying uses developed with subsets of the head nouns with which their source meanings typically occurred. Although some degree of collocational extension was observed for all of the noun-intensifying meanings studied, these uses continue to be collocationally restricted: they occur only with degree nouns inviting either closed scale or open scale intensifi-

Reconciliation of synchrony and diachrony

283

cation and/or modifiers that allow degree modification. For such, no collocational overlap between the identifying and noun-intensifying uses was found, but the latter display the same restrictions in terms of collocational behaviour as the adjectives studied. As argued by De Smet and Ghesquière (2010), the constrained generalizability of the noun-intensifying uses is due to the fact that they engage with the lexical specificities of the head nouns and/ or modifiers with which they occur, i.e. they modify inherent or pragmatically evoked scalar notions. In contrast, identifying uses do not engage with the lexical semantics of the head noun they combine with. For the completeness adjectives, I argued that in their secondary determiner uses they typically stress the boundedness of the count nouns they occur with as well as the inclusive reference of the definite primary determiner. The secondary determiner uses of the specificity adjectives typically contribute to the identification of the NP referent by instructing the hearer to look for clues in the discourse context which are crucial to the intended referent identification. The largely independent nature of secondary determiners with respect to the specific semantics of the head noun explains why there are no or hardly any constraints on their generalization. Summarizing the findings of the corpus studies discussed in Chapters 5, 6 and 7, we can state that they revealed a number of recurrent pathways of change in the prenominal string of the NP: (i)

description > identification (Davidse, Breban, and Van linden 2008) e.g. complete, total, whole, particular, specific (ii) description > noun-intensification (Adamson 2000; Paradis 2000; Athanasiadou 2007) e.g. complete, total (iii) identification > noun-intensification (Bolinger 1972) a. pure identification > noun-intensification e.g. particular, such, zulk, what b. quantifying identification > noun-intensification e.g. whole

What do these findings entail for the (inter)subjective teleology and concomitant leftward movement generally assumed to govern structuralsemantic changes in the NP? First, the observed deictification processes as well as the shifts from descriptive to noun-intensifying meaning confirm the general tendency towards increasing (inter)subjectivity and leftward position in the NP. On the one hand, I argued that the shift from descriptive modifier to secondary determiner is best viewed as a shift from nonsubjective, externally propositional meaning to textually intersubjective

284

Summary

meaning. On the other hand, the development from description to nounintensification entails a shift from non-subjective, externally propositional meaning or from less subjective, internally propositional meaning to expressive, subjective meaning. As such, deictification processes and changes from descriptive to noun-intensifying meaning fit in with Traugott’s (1982) cline of semantic change leading from the propositional to the textual (to the expressive) as well as with Traugott’s (2003) intersubjectification cline, i.e. non-/less subjective > subjective > intersubjective. Moreover, as predicted by Adamson (2000), both pathways of subjectification entail positional shifts toward the left end of the NP. Second, Bolinger’s (1972) hypothesis that identifying uses of elements in the NP may serve as source constructions of noun-intensifying uses is a notable pattern of exception not only to the presumed (inter)subjective teleology but also to the leftward movement hypothesis. I argued that, in accordance with Bolinger’s hypothesis, the secondary determiner uses of the adjectives whole and particular, and of such and zulk lie at the origin of the later noun-intensifying uses of these elements (§5.3.3; §6.3.1 and §7.3). In terms of (inter)subjectivity, the pathway from identification to nounintensification constitutes a shift from textually intersubjective meaning to subjective meaning, consequently contradicting Traugott’s (2003) unidirectional cline of intersubjectification. Also, elements following this developmental path shift from a position at the leftmost determination zone of the NP to the more rightward degree modification zone. As such, this pathway also contradicts the leftward movement hypothesis. Third, in Chapter 8, I showed that the subjectification of descriptive modifier-head structures into subjective compounds such as old fogey and little sod similarly entails rightward movement in the NP. Accordingly, Adamson’s (2000) claim of subjectification always entailing leftward movement and de-subjectification entailing rightward movement should be considered a tendency rather than a hard-and-fast rule. Although the existing hypotheses on (inter)subjectification and their proclaimed unidirectionality and concomitant leftward movement in the NP may be retained as general tendencies governing process of structuralsemantic change in the NP, this study has shown that there are a number of exceptions to this pattern, most notably changes along Bolinger’s neglected pathway from identification to noun-intensification and the development of subjective compounds. It now seems that the preliminary model of semantic change proposed in Section 3.1, reproduced here as Figure 24, can account for all of the patterns of change observed in the data studies.

Reconciliation of synchrony and diachrony

285

Externally propositional

Textual

Internally propositional

Expressive

Figure 24. Integrated model of semantic change

This study has contributed to the theorizing on the status and development of secondary determiner uses and to the mapping of the so far largely uncharted domain of noun-intensification. The fine-grained characterizations of these uses were considered in the light of existing grammaticalization and (inter)subjectification processes, which were assessed by means of in-depth corpus studies. The observations made in the case studies fed back into a more dynamic functional-cognitive approach to the English noun phrase, refinement of the notions of subjectivity and intersubjectivity and to a focus on the importance of collocational patterning as a type of specialized context which can trigger intersubjectification and grammaticalization processes.

References Aaron, Jessi Elana & Rena Torres Cacaullos. 2005. Quantitative measures of subjectification: A variationist study of Spanish salier(se). Cognitive Linguistics 16. 607–633. Aarts, Bas. 2007. Syntactic gradience: The nature of grammatical indeterminacy. Oxford: Oxford University Press. Adamson, Sylvia. 2000. A lovely little example: word order options and category shift in the premodifying string. In Olga Fischer, Anette Rosenbach & Dieter Stein (eds.), Pathways of Change: Grammaticalization in English, 39–66. Amsterdam & Philadelphia: John Benjamins. Aijmer, Karin. 2007. That's like totally cool - A new intensifier in American English. Paper presented at the University of Leuven, 19 December. Allan, Kathryn. 2010. A dull paper. An example of proportional analogy in semantic change. Paper presented at the 16th International Conference on English Historical Linguistics, University of, Pécs, 23-27 August. Allerton, D.J. 1987. English intensifiers and their idiosyncrasies. In Ross Steele & Terry Treadgold (eds.), Language Topics: Essays in honour of Michael Halliday, 15–31. Amsterdam & Philadelphia: John Benjamins. Altenberg, Bengt. 1994. On the functions of such in spoken and written English. In Nelleke Oostdijk, Pieter De Haan & Jan Aarts (eds.), Corpus-based research into language, 223–239. Amsterdam: Rodopi. Athanasiadou, Angeliki. 2006. Adjectives and subjectivity. In Angeliki Athanasiadou, Costas Canakis & Bert Cornillie (eds.), Subjectification: Various Paths to Subjectivity, 209–239. Berlin & New York: Mouton de Gruyter. Athanasiadou, Angeliki. 2007. On the subjectivity of intensifiers. Language Sciences 29. 554–565. Bache, Carl. 2000. Essentials of Mastering English: A Concise Grammar. Berlin & New York: Mouton de Gruyter. Bäcklund, Ulf. 1973. The Collocation of Adverbs of Degree in English. Uppsala: Uppsala Reprotryck AB. Bauer, Laurie. 1983. English word-formation. Cambridge: Cambridge University Press. Benveniste, Emile. 1966. De la subjectivité dans le langage. In Emile Benveniste (ed.), Problèmes de Linguistique Générale, 258–266. Paris: Gallimard. Benveniste, Emile. 1971. Subjectivity in language. In Problems in General Linguistics, 223–230. Trans. by Mary Elizabeth Meek. Coral Gables : FL : University of Miami Press. Biber, Douglas, Stig Johansson, Geoffrey Leech & Randolph Quirk. 1999. Longman grammar of spoken and written English. London: Longman.

288

References

Blakemore, Diane. 1990. Constraints on Interpretation. Proceedings of the sixteenth annual meeting of the Berkeley linguistics society. 363–370. Bolinger, Dwight Le Merton. 1952. Linear modification. Publications of the Modern Language Association of America 67 (7). 1117–1144. Bolinger, Dwight Le Merton. 1961a. Generality, Gradience, and the All-or-None. The Hague: Mouton. Bolinger, Dwight Le Merton. 1961b. Syntactic blends and other matters. Language 37. 366–381. Bolinger, Dwight Le Merton. 1967. Adjectives in English: Attribution and predication. Lingua 18. 1–34. Bolinger, Dwight Le Merton. 1972. Degree Words. The Hague: Mouton. Börjars, Kersti & Kate Burridge. 2001. Introducing English grammar. London: Arnold. Breban, Tine. 2002. Adjectives of comparison: Postdeterminer, epithet and classifier uses. Leuven: University of Leuven MA thesis. Breban, Tine. 2006. English adjectives of comparison: Lexical and grammaticalized uses. Leuven: University of Leuven dissertation. Breban, Tine. 2010a. English adjectives of comparison: Lexical and grammaticalized uses. Berlin & New York: de Gruyter Mouton. Breban, Tine. 2010b. The historical development of very: paths and mechanisms of change. Paper presented at the International Conference on Grammaticalization and (Inter)Subjectification, Brussels, 11–13 November. Breban, Tine. 2011. Secondary determiners as markers of generalized instantiation in English noun phrases. Cognitive Linguistics 22 (3). 511–533. Breban, Tine & Kristin Davidse. 2003. Adjectives of comparison: The grammaticalization of their attribute uses into postdeterminer and classifier uses. Folia Linguistica 37 (3/4). 269–317. Breban, Tine, Kristin Davidse & Lobke Ghesquière. 2011. A typology of anaphoric and cataphoric relations expressed by English complex determiners. Journal of Pragmatics 43 (11). 2689–2703. Brems, Lieselotte. 2003. Measure Noun Constructions: an instance of semantically-driven grammaticalization. International Journal of Corpus Linguistics 8. 283–312. Brems, Lieselotte. 2007. The synchronic layering of size noun and type noun constructions in English. Leuven: University of Leuven dissertation. Brems, Lieselotte. 2010. Size noun constructions as collocationally constrained constructions: lexical and grammaticalized uses. English Language and Linguistics 14 (1). 83–109. Brems, Lieselotte & Kristin Davidse. 2010 The Grammaticalisation of Nominal Type Noun Constructions with kind/sort of: Chronology and Paths of Change. English Studies 91 (2). 180–202. Brems, Lieselotte, Lobke Ghesquière & Freek Van de Velde (eds.). 2012. Intersections of intersubjectivity. [special issue] English Text Construction 5 (1).

References

289

Brems, Lieselotte & Sebastian Hoffmann. 2012. New perspectives, theories and methods: Grammaticalization. In Alexander Bergs & Laurel Brinton (eds.), Historical linguistics of English, 1558–1576. Berlin & New York: de Gruyter Mouton. Brinton, Laurel J. 2002. Grammaticalization versus lexicalization reconsidered: On the late use of temporal adverbs. In Teresa Fanego, Mariá José López-Couso & Javier Pérez-Guerra (eds.), English Historical Syntax and Morphology, 67–97. Amsterdam & Philadelphia: John Benjamins. Brinton, Laurel J. & Elizabeth Closs Traugott. 2005. Lexicalization and language change. Cambridge: Cambridge University Press. Bublitz, Wolfram. 1996. Semantic prosody and cohesive company: 'somewhat predictable'. Leuven Contributions in Linguistics and Philology 85. 1–32. Buchstaller, Isabelle & Elizabeth Closs Traugott. 2006. The lady was al demonyak: historical aspects of Adverb all. English Language and Linguistics 10. 345– 370. Bybee, Joan. 2003. Mechanisms of Change in Grammaticization: The Role of Frequency. In Brian D. Joseph & Richard D. Janda (eds.), The Handbook of Historical Linguistics, 602–623. London: Blackwell. Bybee, Joan, Revere D. Perkins & William Pagliuca. 1994. The evolution of grammar: Tense, aspect and modality in the languages of the world. Chicago: University of Chicago Press. Cambridge Advanced Learners’ Dictionary: Woodford, Kate (ed.). 2005. Cambridge Advanced Learners’ Dictionary. Cambridge: Cambridge University Press. Carlier, Anne & Walter De Mulder. 2010. The emergence of the definite article: ille in competition with ipse in Late Latin. In Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens (eds.), Subjectification, intersubjectification, and grammaticalization, 241–275. Berlin & New York: de Gruyter Mouton. Carlson, Gregory Norman. 1978. Reference to kinds in English. Bloomington: Indiana University. Chafe, Wallace. 2003. Language and the flow of thought. In Michael Tomasello (ed.), The New Psychology of Language: Cognitive and Functional approaches to Language Structure, 93–111. New Jersey: Lawrence Erlbaum. Clausner, Timothy C. & William Croft. 1999 Domains and image schemas. Cognitive Linguistics 10 (1). 1–31. Clear, Jeremy. 1993. From Firth principles: computational tools for the study of collocation. In Mona Baker, Gill Francis & Elena Tognini-Bonelli (eds.), Text and technology, 271–292. Amsterdam & Philadelphia: John Benjamins. Clear, Jeremy, Gwyneth Fox, Gill Francis, Ramesh Krishnamurthy & Rosamund Moon. 1996. COBUILD: The State of the Art. International Journal of Corpus Linguistics 1 (2). 303–314. Colleman, Timothy & Bernard De Clerck. 2011. Constructional semantics on the move: A diachronic view on the semantics of the English double object construction. Cognitive Linguistics 22 (1). 183–210.

290

References

Cornillie, Bert. 2006. Constructional and Conceptual considerations on the subjectivity of English and Spanish modals. In Angeliki Athanasiadou, Costas Canakis & Bert Cornillie (eds.), Subjectification: Various Paths to Subjectivity, 177– 205. Berlin & New York: Mouton de Gruyter. Cornillie, Bert. 2008. On the grammaticalization and (inter)subjectivity of Spanish evidential (semi-)auxiliaries. In Elena Seoane & María José López Couso (eds.), Theoretical and Empirical Issues in Grammaticalization, 55–76. Amsterdam & Philadelphia: John Benjamins. Croft, William. 2000. Explaining language change. An evolutionary approach. London: Longman. Croft, William. 2001. Radical construction grammar. Syntactic theory in typological perspective. Oxford: Oxford University Press. Croft, William & Alan Cruse. 2004. Cognitive Linguistics. Cambridge: Cambridge University Press. Cruse, Alan. 2000. Meaning in language: An introduction to semantics and pragmatics. Oxford: Oxford University Press. Crystal, David. 2003. The Cambridge Encyclopedia of the English Language, 2nd edn. Cambridge: Cambridge University Press. Cuyckens, Hubert, Kristin Davidse & Lieven Vandelanotte. 2010. Introduction. In Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens (eds.), Subjectification, intersubjectification and grammaticalization (Topics in English Linguistics 66), 1–26. Berlin & New York: de Gruyter Mouton. Danes, Frantisek. 1964. A three-level approach to syntax. Travaux linguistiques de Prague 1: 225–240. Davidse, Kristin. 1997. Introduction to English Grammar. Course notes. Department of Linguistics, University of Leuven. Davidse, Kristin. 1999. The semantics of cardinal versus enumerative existential constructions. Cognitive Linguistics 10 (3). 203–250. Davidse, Kristin. 2004. The Interaction of Quantification and Identification in English Determiners. In Michel Achard & Suzanne Kemmer (eds.), Language, Culture, and Mind, 507–533. Stanford: CSLI Publications. Davidse, Kristin. 2009. Complete and sort of: from identifying to intensifying? Transactions of the Philological Society 107 (3). 262–292. Davidse, Kristin, Tine Breban & An Van linden. 2008. Deictification: The development of secondary deictic meanings by adjectives in the English NP. English Language and Linguistics 12. 475–503. De Mönnink, Inge. 1996. A first approach to the mobility of noun phrase constituents. In Carol Percy, Charles Meyer & Ian Lancashire (eds.), Synchronic corpus linguistics: Papers from the sixteenth International Conference on English Language Research on Computerized Corpora, 143–157. Amsterdam: Rodopi. De Mulder, Walter. 1998. Anaphora. In Jef Verschueren, Jan-Ola Ostman, Jan Blommaert & Chris Bulcaen (eds.), Handbook of Pragmatics, 1–28. Amsterdam & Philadelphia: John Benjamins.

References

291

De Rooij, Jaap. 1989. Zo’n dingen zegggen ze hier (niet): regionale verschillen in het gebruik van zo’n en zulke. In Siegfried Theissen & Jef Vromans (eds.), Album Moors, 181–201. Liège: CIPL. De Smedt, Liesbeth, Lieselotte Brems & Kristin Davidse. 2007. NP-internal functions and extended uses of the 'type' nouns kind, sort, and type: Towards a comprehensive corpus-based description. In Roberta Facchinetti (ed.), Corpus linguistics 25 years on, 227–257. Amsterdam: Rodopi. De Smet, Hendrik. 2005. A Corpus of Late Modern English Texts. ICAME Journal 29. 69–82. De Smet, Hendrik. 2008. Diffusional change in the English system of complementation: Gerunds, participles and for...to-infinitives. Leuven: University of Leuven dissertation. De Smet, Hendrik. 2009. Analysing reanalysis. Lingua 119. 1728–1755. De Smet, Hendrik. 2010. Grammatical interference: subject marker for and phrasal verb particle out. In Elizabeth Closs Traugott & Graeme Trousdale (eds.), Gradualness, gradience and grammaticalization, 75–104. Amsterdam & Philadelphia: John Benjamins. De Smet, Hendrik & Lobke Ghesquière. 2010. When grammaticalization hesitates. Paper presented at the International Conference on Grammaticalization and (Inter)Subjectification, Brussels, 11–13 November. De Smet, Hendrik & Jean-Christophe Verstraete. 2006. Coming to terms with subjectivity. Cognitive Linguistics 17. 365–392. Declerck, Renaat. 1991. A Comprehensive Descriptive Grammar of English. Tokyo: Kaitakusha. Demske, Ulrike. 2004. Weshalb Demonstrativpronomina nicht immer Determinantien sind. In Franz-Josef d’Avis (ed.), Deutsche Syntax: Empire und Theorie, 53–80. Göteborg: Acta Universitatis Gothoburgensis. Denison, David. 2002. History of the sort of construction family. Paper presented at the Second International Conference on Construction Grammar, Helsinki, 7 September. Available online at http://lings.In.man.ac.uk/staff/dd/papers/sortof_iccg2.pdf. Denison, David. 2006. Category change and gradience in the determiner system. In Ans van Kemenade & Bettelou Los (eds.), The handbook of the history of English, 279–304. Oxford: Blackwell. Di Paolo Healey, Antonette. 1997. Words, story, history: The mapping of meaning and Toronto's Dictionary of Old English. In Terttu Nevalainen & Leena Kahlas-Tarkka (eds.), To explain the present: Studies in the changing English language in honour of Matti Rissanen, 35–49. Helsinki: Société Néophilologique. Diessel, Holger. 1999. Demonstratives: Form, function, and grammaticalization. Amsterdam & Philadelphia: John Benjamins. Diessel, Holger. 2006. Demonstratives, joint attention, and the emergence of grammar. Cognitive Linguistics 17. 463–489.

292

References

Diessel, Holger. 2011. Where do grammatical morphemes come from? On the development of grammatical markers from lexical expressions, demonstratives, and question words. Jena: University of Jena manuscript. Diessel, Holger. 2012. Bühler’s two-field theory of pointing and naming and the deictic origins of grammatical morphemes. In Kristin Davidse, Tine Breban, Lieselotte Brems & Tanja Mortelmans (eds.), Grammaticalization and Language Change: New reflections, 37–50. Amsterdam & Philadelphia: John Benjamins. Diewald, Gabriele. 2002. A model for relevant types of contexts in grammaticalization. In Ilse Wischer & Gabriele Diewald (eds.), New reflections on grammaticalization, 103–120. Amsterdam & Philadelphia: John Benjamins. Diewald, Gabriele. 2006. Context types in grammaticalization as constructions. Constructions (Special Volume 1: Constructions all over - case studies and theoretical implications.) Diewald, Gabriele. 2010. On some problem areas in grammaticalization theory. In Katerina Stathi, Elke Gehweiler & Ekkehard König (eds.), Grammaticalization: Current views and issues, 17–50. Amsterdam & Philadelphia: John Benjamins. Dixon, Robert. 1977. Where have all the adjectives gone? Studies in Language 1. 19–80. Dixon, Robert. 1982. Where have all the adjectives gone? And other essays in semantics and syntax. Berlin & New York: Mouton de Gruyter. Du Bois, John W. 2003. Discourse and grammar. In Michael Tomasello (ed.), The New Psychology of Language: Cognitive and Functional approaches to Language Structure, 47–87. New Jersey: Lawrence Erlbaum. Duinhoven, Antonius Maria. 1988. Middelnederlandse syntaxis, synchroon en diachroon. Deel 1: De naamwoordgroep. Leiden: Nijhoff. Enç, Mürvet. 1991. The semantics of specificity. Linguistic Inquiry 22 (1). 1–25. Ernst, Thomas Boyden. 1984. Towards an integrated theory of adverb position in English. Bloomington: Indiana University Linguistics Club. Evans, Nicholas & David Wilkins. 2000. In the mind's ear: the semantic extensions of perception verbs in Australian languages. Language 76. 546–592. Eckardt, Regine. 2012. The many careers of negative polarity items. In Kristin Davidse, Tine Breban, Lieselotte Brems & Tanja Mortelmans (eds.), Grammaticalization and Language Change: New reflections, 299–326. Amsterdam & Philadelphia: John Benjamins. Feist, Jim. 2009. Premodifier order in English nominal phrases: A semantic account. Cognitive Linguistics 20 (2). 301–340. Firbas, Jan. 1992. Functional sentence perspective in written and spoken communication. Cambridge: Cambridge University Press Firth, John Rupert. 1957. Papers in Linguistics 1934-1951. London: Oxford University Press. Fischer, Olga. 1988. The rise of the for NP to V construction: an explanation. In Graham Nixon & John Honey (eds.), An historic tongue: Studies in English linguistics in memory of Barbara Strang, 67–88. London: Routledge.

References

293

Fischer, Olga. 1992. The noun phrase. In Norman Blake (ed.), The Cambridge history of the English language. Volume 2: 1066–1476, 210–233. Cambridge: Cambridge University Press. Fischer, Olga. 1997. The grammaticalisation of infinitival to in English compared with German and Dutch. In Raymond Hickey & Stanislaw Puppel (eds.), Language history and linguistic modelling. A Festschrift for Jacek Fisiak on his 60th birthday. Volume I, 265–280. Berlin & New York: Mouton de Gruyter. Fischer, Olga. 2000. Grammaticalization: unidirectional, non-reversable? The case of to before the infinitive in English. In Olga Fischer, Anette Rosenbach & Dieter Stein (eds.), Pathways of change: Grammaticalization in English, 149– 169. Amsterdam & Philadelphia: John Benjamins. Fischer, Olga. 2007. Morphosyntactic change: Functional and formal perspectives. Oxford: Oxford University Press. Fitzmaurice, Susan. 2000. Remarks on the degrammaticalization of infinitival to in present-day American English. In Olga Fischer, Anette Rosenbach & Dieter Stein (eds.), Pathways of change: Grammaticalization in English, 171–186. Amsterdam & Philadelphia: John Benjamins. Francis, W. Nelson. 1965. The English language: An introduction. Background for writing. New York: Norton & Company. Fried, Miriam. 2010. Grammaticalization and lexicalization effects in participial morphology: a Construction Grammar approach to language change. In An Van linden, Jean-Christophe Verstraete & Kristin Davidse (eds.), Formal evidence in grammaticalization research (Typological studies in language), 191–224. Amsterdam & Philadelphia: John Benjamins. Ghesquière, Lobke. 2006. The complete fool and the little old lady: A synchronic and diachronic study of the prenominal adjectives old, little and complete. Leuven: University of Leuven MA thesis. Ghesquière, Lobke. 2009. From determining to emphasizing meanings: the adjectives of specificity. Folia Linguistica 43 (2). 311–343. Ghesquière, Lobke. 2010. On the subjectification and intersubjectification paths followed by the adjectives of completeness. In Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens (eds.), Subjectification, intersubjectification and grammaticalization (Topics in English Linguistics 66), 277–314. Berlin & New York: Mouton de Gruyter. Ghesquière, Lobke & Kristin Davidse. 2011. The development of intensification scales in noun-intensifying uses of adjectives: sources, paths and mechanisms of change. English Language and Linguistics 15 (2). 251–277. Ghesquière, Lobke & Freek Van de Velde. 2011. A corpus-based account of the development of English such and Dutch zulk: identification, intensification and (inter)subjectification. Cognitive Linguistics 22 (4). 765–797. Ghesquière, Lobke, Lieselotte Brems & Freek Van de Velde. 2012. Intersubjectivity and intersubjectification: Typology and operationalization. English Text Construction 5 (1). 128–152.

294

References

Ghesquière, Lobke, An Van linden & Kristin Davidse. 2013. Subjective Compounds and subjectivity/subjectification in the English noun phrase. English Studies 94 (1). 90–117. Giegerich, Heinz. 2009. The English compound stress myth. Word Structure 2 (1). 1–17. Giegerich, Heinz. 2011. Noun compounds and noun phrases: stress and some other similarities. Paper presented at the Second International Workshop on the Structure of the Noun Phrase in English – Synchronic and Diachronic Explorations, Newcastle University, 15–16 September. Givón, Talmy. 1979. On Understanding Grammar. New York: Academic Press. Givón, Talmy. 1991. The evolution of dependent clause morpho-syntax in Biblical Hebrew. In Elizabeth Closs Traugott & Bernd Heine (eds.), Approaches to Grammaticalization. Volume II: Types of Grammatical Markers, 257–310. Amsterdam & Philadelphia: John Benjamins. González-Díaz, Victorina. 2009. Little old problems: Adjectives and subjectivity in the English NP. Transactions of the Philological Society 107 (3). 376–402. Gundel, Jeanette K., Nancy Hedberg & Ron Zacharski. 1993. Cognitive status and the form of referring expressions in discourse. Language 69. 274–307. Haas, William. 1954. On defining linguistics units. Transactions of the Philological Society 63 (1). 54–84. Haiman, John. 1994. Ritualization and the development of language. In William Pagliuca (ed.), Perspectives on grammaticalization, 3–28. Amsterdam & Philadelphia: John Benjamins. Halliday, Michael A.K. 1978. Language as social semiotic. The social interpretation of language and meaning. London: Arnold. Halliday, Michael A.K. 1985. An Introduction to Functional Grammar. London: Arnold. Halliday, Michael A.K. 1994. An Introduction to Functional Grammar, 2nd edn. London: Arnold. Halliday, Michael A.K. 2002a [1961]. Categories of the theory of grammar. In Jonathan Webster (ed.), Collected works of M.A.K. Halliday. Volume 1: On Grammar, 37–94. London: Continuum. [First published in Word 17. 241–292.] Halliday, Michael A.K. 2002b [1977]. Text as Semantic Choice in Social Contexts. In Jonathan Webster (ed.). Collected works of M.A.K. Halliday. Volume 2: Linguistic studies of text and discourse. London: Continuum. 23–81. [First published in Teun A. van Dijk and János Petöfi (eds.). Grammars and descriptions. Berlin: de Gruyter. 176–226.] Halliday, Michael A.K. 2002c [1979]. Modes of meaning and modes of expression: Types of grammatical structure, and their determination by different semantic functions. In Jonathan Webster (ed.), Collected works of M.A.K. Halliday. Volume 1: On Grammar, 196–218. London: Continuum. First published in D.J Allerton, Edward Carney & David Holdcroft, Function and Context in Linguistic Analysis, 57–79. London: Cambridge University Press.]

References

295

Halliday, Michael A.K. 2002d [1981]. Text semantics and clause grammar: How is a text like a clause? In Jonathan Webster (ed.), Collected works of M.A.K. Halliday. Volume 1: On Grammar, 219–260. London: Continuum. [First published as Text Semantics and clause grammar: some patterns of realization. In James E. Copeland & Philip W. Davies (eds.), Seventh LACUS Forum 1980, 31–59.; How is a text like a clause? In Sture Allen (ed.), 1982, Text processing: Text analysis and generation, Text typology and attrition. Stockholm: Almqvist and Wiksell.] Halliday, Michael A.K. 2004. An Introduction to Functional Grammar, 3rd edn. Revised by Christian M. Matthiesen. London: Arnold. Halliday, Michael A.K. & Ruqaia Hasan. 1976. Cohesion in English. London: Longman. Haspelmath, Martin. 1999. Why is grammaticalization irreversible? Linguistics 37 (6). 1043–1086. Haspelmath, Martin. 2004. On directionality in language change with particular reference to grammaticalization. In Olga Fischer, Muriel Norde & Harry Peridon (eds.), Up and down the cline: The nature of grammaticalization, 17–44. Amsterdam & Philadelphia: John Benjamins. Hawkins, John. 1978. Definiteness and indefiniteness: a study in reference and grammaticality prediction. London: Croom Helm. Heine, Bernd. 2002. On the role of context in grammaticalization. In Ilse Wischer & Gabriele Diewald (eds.), New reflections on grammaticalization, 83–101. Amsterdam & Philadelphia: John Benjamins. Heine, Bernd & Tania Kuteva. 2007. The genesis of grammar: A reconstruction. Oxford: Oxford University Press. Heine, Bernd & Mechthild Reh. 1984. Grammaticalization and Reanalysis in African Languages. Hamburg: Helmut Buske. Hetzron, Robert. 1978. On the relative order of adjectives. In Hansjakob Seiler (ed.), Language Universals, 165–184. Tübingen: Narr. Himmelmann, Nikolaus P. 2004. Lexicalization and grammaticization: Opposite or orthogonal? In Björn Wiemer, Walter Bisang & Nikolaus P. Himmelmann (eds.), What makes grammaticalization? A look from its fringes and its components, 21–42. Berlin & New York: Mouton de Gruyter. Hole, Daniel & Gerson Klumpp. 2000. Definite type and indefinite token: the article son in colloquial German. Linguistische Berichte 182. 231–244. Hopper, Paul J. 1991. On some principles of grammaticization. In Elizabeth Closs Traugott & Bernd Heine (eds.), Approaches to grammaticalization, 17–35. Amsterdam & Philadelphia: John Benjamins. Hopper, Paul J. & Elizabeth Closs Traugott. 1993. Grammaticalization. Cambridge: Cambridge University Press. Hopper, Paul J. & Elizabeth Closs Traugott. 2003. Grammaticalization, 2nd edn. Cambridge: Cambridge University Press. Huddleston, Rodney. 1984. Introduction to the grammar of English. Cambridge: Cambridge University Press.

296

References

Huddleston, Rodney & Geoffrey Pullum. 2002. The Cambridge grammar of the English language. Cambridge: Cambridge University Press. Ionin, Tania. 2010. The scope of indefinites: an experimental investigation. Natural Language Semantics 18. 295–350. Jacobson, Sven. 1964. Adverbial positions in English. Stockholm: AB Studentenbok. Kamoen, Naomi, Bregje Holleman, Rick Nouwen, Ted Sanders & Huub van den Bergh. 2011. Absolutely relative or relatively absolute? The linguistic behavior of gradable adjectives and degree modifiers. Journal of Pragmatics 43(13). 3139-3151. Kennedy, Christopher & Louise McNally. 2005. Scale structure, degree modification, and the semantics of gradable predicates. Language 81 (2). 345–381. König, Ekkehard. 1991. The meaning of focus particles: A comparative perspective. London: Routledge. Kroch, Anthony & Ann Taylor. 2000. Penn-Helsinki Parsed Corpus of Middle English, 2nd edn. Available online at http://www.ling.upenn.edu/histcorpora/PPCME2-RELEASE-2/. Kroch, Anthony, Beatrice Santorini & Lauren Delfs. 2004. Penn-Helsinki Parsed Corpus of Early Modern English. Available online at http://www.ling.upenn.edu./hist-corpora/PPCEME-RELEASE-1/. Laffut, An. 2006. Three-participant constructions in English: a functionalcognitive approach to caused relations. Amsterdam & Philadelphia: John Benjamins. Langacker, Ronald W. 1985. Observations and speculations on subjectivity. In John Haiman (ed.), Iconicity in syntax, 109–150. Amsterdam & Philadelphia: John Benjamins. Langacker, Ronald W. 1987a. Foundations of Cognitive Grammar: Volume I: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1987b. Nouns and verbs. Language 63 (1). 53–94. Langacker, Ronald W. 1990. Subjectification. Cognitive Linguistics 1. 5–38. Langacker, Ronald W. 1991. Foundations of Cognitive Grammar: Volume II: Descriptive Application. Stanford: Stanford University Press. Langacker, Ronald W. 1993. Reference-point constructions. Cognitive Linguistics 4 (1). 1–38. Langacker, Ronald W. 1998. On subjectification and grammaticization. In JeanPierre Koenig (ed.), Discourse and cognition: Bridging the gap, 71–89. Stanford: CSLI Publications. Langacker, Ronald W. 2001. What WH means. In Alan Cienki, Barbara J. Luka & Michael B. Smith (eds.), Conceptual and discourse factors in linguistic structure, 137–152. California: CSLI Publications. Langacker, Ronald W. 2002a. Deixis and subjectivity. In Frank Brisard (ed.), Grounding: The Epistemic Footing of Deixis and Reference, 1–28. Berlin & New York: Mouton de Gruyter.

References

297

Langacker, Ronald W. 2002b. Remarks on the English grounding systems. In Frank Brisard (ed.), Grounding: The Epistemic Footing of Deixis and Reference, 29–38. Berlin & New York: Mouton de Gruyter. Langacker, Ronald W. 2003. Extreme subjectification: English tense and modals. In Hubert Cuyckens, Thomas Berg, René Dirven & Klaus-Uwe Panther (eds.), Motivations in language: Studies in honor of Günter Radden, 3–26. Amsterdam & Philadelphia: John Benjamins. Langacker, Ronald W. 2005. Dynamicity, Fictivity, and Scanning: The Imaginative Basis of Logic and Linguistic Meaning. In Diane Pecher & Rolf A. Zwaan (eds.), Grounding Cognition: The Role of Perception and Action in Memory, Language and Thinking, 164–197. Cambridge: Cambridge University Press. Langacker, Ronald W. 2006. Subjectification, grammaticization, and conceptual archetypes. In Angeliki Athanasiadou, Costas Canakis & Bert Cornillie (eds.), Subjectification: Various Paths to Subjectivity, 17–40. Berlin & New York: Mouton de Gruyter. Langacker, Ronald W. 2007. Cognitive Grammar. In Dirk Geeraerts & Hubert Cuyckens (eds.), The Oxford Handbook of Cognitive Linguistics, 421–462. Oxford: Oxford University Press. Langacker, Ronald W. 2008. Cognitive grammar: A basic introduction. Oxford: Oxford University Press. Langacker, Ronald W. 2009. From basic experience to abstract reason: Embodied cognition and its linguistic manifestation. Paper presented at Stanford University, 26 January 2009. Le Petit Robert de la langue française: Rey-Debove, Josette & Alain Rey (eds.), 2012, Le nouveau Petit Robert: Dictionnaire alphabétique et analogique de la langue française. Available online at http://pr.bvdep.com/. Lees, Robert. 1963. The grammar of English nominalizations. The Hague: Mouton. Lehmann, Christian. 1985. Grammaticalization: Synchronic Variation and Diachronic Change. Lingua e Stile 20. 303–318. Lehmann, Christian. 1995 [1982]. Thoughts on Grammaticalization. Newcastle: LINCOM Europe. [First published as Thoughts on Grammaticalization: A Programmatic Sketch. 1982. (Arbeiten des Kölner Universalien-Projektes.) University of Cologne, Institut für Sprachwissenschaft.] Lehmann, Christian. 2002. New reflections on grammaticalization and lexicalization. In Ilse Wischer & Gabriele Diewald (eds.), New reflections on grammaticalization, 1–18. Amsterdam & Philadelphia: John Benjamins. Lehmann, Christian. 2004. Theory and method on grammaticalization. Zeitschrift für germanistische Linguistik 32 (2). 152–187. Lehrer, Adrienne. 1985. The influence of semantic fields on semantic change. In Jacek Fisiak (ed.), Historical semantics, historical word-formation, 283–296. Berlin & New York: Mouton de Gruyter. Lorenz, Gunter. 2002. Really worthwhile or not really significant? In Ilse Wischer & Gabriele Diewald (eds.), New Reflections on Grammaticalization, 143–161. Amsterdam & Philadelphia: John Benjamins.

298

References

Los, Bettelou. 2005. The rise of the to-infinitive. Oxford: Oxford University Press. Lyons, John. 1977. Semantics. Vol. 2. Cambridge: Cambridge University Press. Lyons, John. 1982. Deixis and subjectivity: loquor, ergo sum? In Robert J. Jarvella & Wolfgang Klein (eds.), Speech, Place, and Action: Studies in Deixis and Related Topics, 101–124. New York: Wiley and Sons. Lyons, John. 1994. Subjecthood and subjectivity. In Marina Yaguello (ed.), Subjecthood and subjectivity: The status of the subject in linguistic theory, 9–17. Paris: Ophrys. Lyons, John. 1999. Definiteness. Cambridge: Cambridge University Press. Mackenzie, Lachlan J. 1997. Grammar, discourse and knowledge: the use of such. In Jan Aarts, Inge De Mönnink & Herman Wekker (eds.), Studies in English language and teaching. In honour of Flor Aarts, 85–105. Amsterdam: Rodopi. Macmillan English Dictionary for Advanced Learners: Rundell, Michael & Gwyneth Fox (eds.), 2007, Macmillan English Dictionary for Advanced Learners. Oxford: Macmillan. Martin, James R. 1992. English Text: System and Structure. Amsterdam & Philadelphia: John Benjamins. Martin, James R. 2008. What kind of structure? Interpersonal meaning and prosodic realisation across strata. WORD 58 (3). Matthews, Peter. 1974. Morphology. London: Cambridge University Press. McGregor, William. 1997. Semiotic Grammar. Oxford: Clarendon Press. McGregor, William. 2003. The nothing that is, the zero that isn’t. Studia Linguistica 57 (2). 75–119. MED: Kurath, Hans, Sherman M. Kuhn, John Reidy & Robert E. Lewis (eds.), 1952–2001, Middle English Dictionary. Ann Arbor: University of Michigan Press. Available online at http://quod.lib.umich.edu/m/med/. Milsark, Gary. 1976. Existential sentences in English. Bloomington: Indiana University Linguistics Club. Milsark, Gary. 1977. Toward an explanation of certain peculiarities of the existential construction in English. Linguistic Analysis 3. 1–29. Moltmann, Friederike. 2005. Part structures in situations: the semantics of individual and whole. Linguistics and Philosophy 28. 599–641. Mortelmans, Tanja. 2006. Langacker's subjectification and grounding: a more gradual view. In Angeliki Athanasiadou, Costas Canakis & Bert Cornillie (eds.), Subjectification: Various Paths to Subjectivity, 151–176. Berlin & New York: Mouton de Gruyter. Narrog, Heiko. 2010. (Inter)subjectification in the domain of modality and mood – Concepts and cross-linguistic realities. In Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens (eds.), Subjectification, intersubjectification and grammaticalization (Topics in English Linguistics 66), 385–429. Berlin & New York: Mouton de Gruyter. Narrog, Heiko. 2012. Beyond intersubjectification: textual uses of modality and mood in subordinate clauses as part of speaker-orientation. English Text Construction 5 (1). 29–52.

References

299

Nevalainen, Terttu & Matti Rissanen. 2002. Fairly pretty or pretty fair? On the development and grammaticalization of English downtoners. Language Sciences 24. 359–380. Nevalainen, Terttu. 1991. BUT, ONLY, JUST. Focusing adverbial change in Modern English 1500–1900. Helsinki: Société Neophilologique. Nevalainen, Terttu. 1994. Aspects of adverbial change in Early Modern English. In Dieter Kastovsky (ed.), Studies in Early Modern English, 243–259. Berlin & New York: Mouton de Gruyter. Newmeyer, Frederick J. 1998. Language form and language function. Cambridge, Mass.: MIT Press. Norde, Muriel. 2009. Degrammaticalization. Oxford: Oxford University Press. Nuyts, Jan. 2001a. Epistemic Modality, Language, and Conceptualization. Amsterdam & Philadelphia: John Benjamins. Nuyts, Jan. 2001b. Subjectivity as an evidential dimension in epistemic modal expressions. Journal of Pragmatics 33. 383–400. Nuyts, Jan. 2012. Notions of (inter)subjectivity. English Text Construction 5 (1). 53–76. Oxford Advanced Learner’s Dictionary: Hornby, Albert Sydney & Sally Wehmeier (eds.). 2005. Oxford Advanced Learner’s Dictionary. Oxford: Oxford University Press. OED: Murray, James A.H., Henry Brodly, W.A. Craigie & C.T. Onions. 1993. The Oxford English Dictionary. Oxford: Oxford University Press. Available online at http://www.oed.com. Pakendorf, Brigitte & Ewa Schalley. 2007. From possibility to prohibition: A rare grammaticalization pathway. Linguistic Typology 11. 515–540. Pander Maat, Henk. 2006. Subjectification in gradable adjectives. In Angeliki Athanasiadou, Costas Canakis & Bert Cornillie (eds.), Subjectification: Various Paths to Subjectivity, 279–320. Berlin & New York: Mouton de Gruyter. Paradis, Carita. 1997. Degree Modifiers of Adjectives in Spoken British English. Lund: Lund University Press. Paradis, Carita. 2000. Reinforcing adjectives: A cognitive semantic perspective on grammaticalization. In Ricardo Bermudez-Otero, David Denison, Richard Hogg & Christopher McCully (eds.), Generative theory and corpus studies: a dialogue from 10 ICEHL, 233–258. Berlin & New York: Mouton de Gruyter. Paradis, Carita. 2001. Adjectives and boundedness. Cognitive Linguistics 12. 47– 65. Paradis, Carita. 2003. Between epistemic modality and degree: the case of really. In Roberta Facchinetti, Manfred Krug & Frank Palmer (eds.), Modality in Contemporary English, 197–220. Berlin & New York: Mouton de Gruyter. Paradis, Carita. 2008. Configurations, construals and change: expressions of DEGREE. English Language and Linguistics 12. 317–343. Paradis, Carita & Caroline Willners. 2006. Antonymy and negation – The boundedness hypothesis. Journal of Pragmatics 38. 1051–1080.

300

References

Partington, Alan. 1993. Corpus evidence of language change: The case of the intensifier. In Mona Baker, Gill Francis & Elena Tognini-Bonelli (eds.), Text and technology: in honour of John Sinclair, 177–192. Amsterdam & Philadelphia: John Benjamins. Partington, Alan. 2004. Utterly content in each other's company: Semantic prosody and semantic preference. International Journal of Corpus Linguistics 9 (1). 131–156. Peters, Hans. 1994. Degree adverbs in Early Modern English. In Dieter Kastovsky (ed.), Studies in Early Modern English, 269–288. Berlin & New York: Mouton de Gruyter. Pike, Kenneth Lee. 1972a [1959]. Language as particle, wave, and field. In Ruth M. Brend (ed.), Kenneth L. Pike: Selected writings to commemorate the 60th birthday of Kenneth Lee Pike, 129–143. The Hague: Mouton. [First published as Language as particle, wave, and field. The Texas Quarterly 2 (2). 37–54.] Pike, Kenneth Lee. 1972b [1962]. Dimensions of grammatical constructions. In Ruth M. Brend (ed.), Kenneth L. Pike: Selected writings to commemorate the 60th birthday of Kenneth Lee Pike, 129–143. The Hague: Mouton. [First published as Dimensions of grammatical constructions. Language 38 (3). 221– 244.] Pike, Kenneth Lee. 1982. Linguistic concepts: An introduction to tagmemics. London: University of Nebraska Press. Pike, Kenneth Lee. 1985. Static, dynamic, and relational perspectives suggested in words and phrases. In Kurt R. Jankowsky (ed.), Scientific and humanistic dimensions of language. Festschrift for Robert Lado on the occasion of his 70th birthday on May 31, 1985, 447–452. Amsterdam & Philadelphia: John Benjamins. Potts, Christopher & Florian Schwarz. 2008. Exclamatives and heightened emotion: Extracting pragmatic generalizations from large corpora. Unpublished paper. Poynton, Cate. 1996. Amplification as a grammatical prosody: Attitudinal modification in the nominal group. In Margaret Berry, Christopher Butler, Robin Fawcett & Guowen Huang (eds.), Meaning and form: Systemic functional interpretations. Meaning and choice in language: Studies for Michael Halliday, 211–227. Norwood: Ablex Publishing Corporation. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1972. A Grammar of Contemporary English. London: Longman. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. A Grammar of Contemporary English. London: Longman. Rett, Jessica. 2008. A degree account of exclamatives. In Masayuki Gibson & Tova Friedman (eds.), Proceedings of Semantics and Linguistics Theory XVIII. Ithaca, NY: CLC Publications. Rijkhoff, Jan. 2002. The noun phrase. Oxford: Oxford University Press.

References

301

Rissanen, Matti. 1999. The noun phrase. In Roger Lass (ed.), The Cambridge history of the English language. Volume 3: 1476-1776, 191–209. Cambridge: Cambridge University Press. Robins, Robert H. 1980. General Linguistics. An Introductory Survey, 3rd edn. London: Longman. Sapir, Edward. 1930. Totality. Language 6 (3). 7–28. Saxon, Leslie. 1990. On one's own: The semantics and pragmatics of reflexives. In Carol P. Georgopoulos & Roberta L. Ishihara (eds.), Interdisciplinary Approaches to Language: Essays in Honor of S.-Y. Kuroda, 501–517. Dordrecht: Kluwer. Schiffrin, Deborah. 1990. The management of a co-operative self during argument: The role of opinions and stories. In Allen D. Grimshaw (ed.), Conflict talk: Sociolinguistic investigations of arguments in conversations, 241–259. Cambridge: Cambridge University Press. Seiler, Hansjakob. 1978. Determination: A functional dimension for inter-language comparison. In Hansjakob Seiler (ed.), Language Universals, 301–328. Tübingen: Narr. Seppänen, Aimo. 1978. Some notes on the construction 'adjective + a + noun'. English Studies 59. 523–537. Sinclair, John. 1990. Collins COBUILD English Grammar. London: HarperCollins. Sinclair, John. 1992 Trust the text. In Martin Davies & Louise Ravelli (eds.), Advances in Systemic Linguistics. Recent Theory and Practice, 5–19. London: Pinter. Sinclair, John (ed.). 2001. Collins COBUILD English dictionary for advanced learners. London: Harper Collins. Smirnova, Elena. 2012. On some problematic aspects of subjectification. Language Dynamics and Change 2. 34–58. Speelman, Dirk. 1997. Abundantia Verborum: A corpus-tool for carrying out corpus based linguistic case studies. Leuven: University of Leuven dissertation. Spinillo, Mariangela Galvão. 2003. On such. English Language and Linguistics 7. 195–210. Spitzbardt, Harry. 1965. English adverbs of degree and their semantic fields. Philologica Pragensia 8 (3). 349–359. Stubbs, Michael. 1995. Collocations and semantic profiles: On the cause of the trouble with quantitative studies. Functions of Language 2. 23–55. Stubbs, Michael. 2001. Words and phrases: Corpus studies of lexical semantics. Oxford: Blackwell. Sweetser, Eve. 1999. Compositionality and blending: semantic composition in a cognitively realistic framework. In Theo Janssen & Gisela Redeker (eds.), Cognitive linguistics: Foundations, scope and methodology, 129–162. Berlin & New York: Mouton de Gruyter. Taylor, John R. 1996. Possessives in English. Oxford: Clarendon Press.

302

References

Teyssier, J. 1968. Notes on the syntax of the adjective in modern English. Lingua 20. 225–249. Thompson, Geoff & Susan Hunston. 2000. Evaluation: an introduction. In Susan Hunston & Geoff Thompson (eds.), Evaluation in text: Authorial stance and the construction of discourse, 1–27. Oxford: Oxford University Press. Torres Cacoullos, Rena & Scott A. Schwenter. 2006. Towards an operational notion of subjectification. In Rebecca T. Cover & Yuni Kim (eds.), Proceedings of the 31st Annual Meeting of the Berkeley Linguistics Society: General session and parasession on prosodic variation and change, 347–358. Berkeley, CA: Berkeley Linguistics Society. Traugott, Elizabeth Closs. 1982. From propositional to textual and expressive meanings: Some semantic-pragmatic aspects of grammaticalization. In Winfred P. Lehmann & Yakov Malkiel (eds.), Perspectives on Historical Linguistics, 245–271. Amsterdam & Philadelphia: John Benjamins. Traugott, Elizabeth Closs. 1988. Pragmatic strengthening and grammaticalization. In Proceedings of the Fourteenth Annual Meeting of the Berkeley Linguistic Society, 406–416. Traugott, Elizabeth Closs. 1989. On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65. 31–55. Traugott, Elizabeth Closs. 1995. Subjectification in grammaticalization. In Dieter Stein & Susan Wright (eds.), Subjectivity and subjectivisation: Linguistic perspectives, 31–54. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs. 2001. Legitimate counterexamples to unidirectionality. Paper presented at Freiburg University, 17 October. Available online at http://www.stanford.edu/~traugott/papers/Freiburg.Unidirect.pdf. Traugott, Elizabeth Closs. 2003a. From subjectification to intersubjectification. In Raymond Hickey (ed.), Motives for Language Change, 124–139. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs. 2003b. Constructions in grammaticalization. In Brian D. Joseph & Richard D. Janda (eds.), The Handbook of Historical Linguistics, 624–647. Oxford: Blackwell. Traugott, Elizabeth Closs. 2006. The semantic development of scalar focus modifiers. In Ans van Kemenade & Bettelou Los (eds.), The handbook of the history of English, 335–359. Oxford: Blackwell. Traugott, Elizabeth Closs. 2007a. (Inter)subjectification and unidirectionality. Journal of Historical Pragmatics 8. 295–309. Traugott, Elizabeth Closs. 2007b. The concepts of constructional mismatch and type-shifting from the perspective of grammaticalization. Cognitive Linguistics 18 (4). 523–557. Traugott, Elizabeth Closs. 2008. The status of onset contexts in analysis of microchanges. Ms. for Merja Kytö (ed.), English corpus linguistics: Crossing paths. Amsterdam: Rodopi. Traugott, Elizabeth Closs. 2010a. (Inter)subjectivity and (inter)subjectification: a reassessment. In Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens

References

303

(eds.), Subjectification, intersubjectification and grammaticalization (Topics in English Linguistics 66), 29–71. Berlin & New York: de Gruyter Mouton. Traugott, Elizabeth Closs. 2010b. Grammaticalization. In Silvia Luraghi & Vit Bubenik (eds.), Continuum Companion to Historical Linguistics, 271–285. London: Continuum. Traugott, Elizabeth Closs. 2010c. The persistence of linguistic contexts over time: Implications for corpus research. Paper presented at ICAME 31, University of Giessen, 26–30 May. Traugott, Elizabeth Closs. 2010d. Dialogic contexts as motivations for syntactic change. In Robert A. Cloutier, Anne Marie Hamilton-Brehm & William A. Kretzschmar (eds.), Studies in the history of the English language V: Variation and change in English grammar and lexicon: contemporary approaches (Topics in English Linguistics 68), 11–27. Berlin & New York: de Gruyter Mouton. Traugott, Elizabeth Closs. 2012. Intersubjectification and clause periphery. English Text Construction 5 (1). 7–28. Traugott, Elizabeth Closs & Richard Dasher. 2002. Regularity in semantic change. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs & Ekkehard König. 1991. The semantics-pragmatics of grammaticalization revisited. In Elizabeth Closs Traugott & Bernd Heine (eds.), Approaches to Grammaticalization, 189–218. Amsterdam & Philadelphia: John Benjamins. TLFi: Le Trésor de la Langue Française informatisé. Available online at http://atilf.atilf.fr/tlf.htm. Tribushinina, Elena. 2008. Cognitive reference points: Semantics beyond the prototypes in adjectives of space and colour. LOT Dissertation Series 192. Trotta, Joe. 2000. Wh-clauses in English: Aspects of theory and description. Amsterdam: Rodopi. Tucker, Gordon. 1997. A functional lexicogrammar of adjectives. Functions of Language 4. 215–250. Tucker, Gordon. 1998. The Lexicogrammar of Adjectives: A Systemic Functional Approach to Lexis. London: Cassell. Van de Velde, Freek. 2007. Interpersonal modification in the English noun phrase. Functions of Language 14. 203–230. Van de Velde, Freek. 2010. The emergence of the determiner in the Dutch NP. Linguistics 48 (2). 263–299. van der Auwera, Johan & Daniel Van Olmen. 2013. On the West Germanic similative indefinites. Paper presented at A Germanic Sandwich 2013. Dutch between English and German, University of Leuven, 11-12 January. Van Langendonck, Willy. 1999. Neurolinguistic and syntactic evidence for basic level meaning in proper names. Functions of Language 6 (1). 95–138. Van linden, An & Kristin Davidse. 2005. Degrees of subjectivity of prenominal adjectives: diachronic change, linear order, and constructional relations as determining factors. Paper presented at From Ideational to Interpersonal: Perspectives from Grammaticalization, University of Leuven, 10–12 February.

304

References

Vandelanotte, Lieven. 2002. Prenominal adjectives in English: Structures and ordering. Folia Linguistica 36. 219–259. Vandelanotte, Lieven. 2009. Speech and thought representation in English: a cognitive-functional approach. Berlin & New York: Mouton de Gruyter. Vandelanotte, Lieven. 2011. Call so and so and tell him such and such: Suspensive reference in contemporary English. Paper presented at the 2nd Conference of the International Society for the Linguistics of English, Boston University, 17– 21 June. Vandewinkel, Sigi & Kristin Davidse. 2008. The interlocking paths of development to emphasizer adjective pure. Journal of Historical Pragmatics 9 (2). 255–287. Verhagen, Arie. 2005. Constructions of Intersubjectivity. Discourse, Syntax and Cognition. Oxford: Oxford University Press. Verhagen, Arie. 2007. Construal and Perspectivation. In Dirk Geeraerts & Hubert Cuyckens (eds.), The Oxford Handbook of Cognitive Linguistics, 48–82. Oxford: Oxford University Press. Visconti, Jacqueline. 2013. Facets of subjectification. Language Sciences 36. 7–17. von Heusinger, Klaus. 2002. Specificity and definiteness in sentence and discourse structure. Journal of Semantics 19. 245–274. von Heusinger, Klaus. 2007. Referentially anchored indefinites. In Ileana Comorovski & Klaus von Heusinger (eds.), Existence: Semantics and syntax, 273– 292. Dordrecht: Springer. Ward, Gregory & Betty Birner. 1995. Definiteness and the English existential. Language 71 (4). 722–742. Wierzbicka, Anna. 1986. What's in a noun? (or: How do nouns differ in meaning from adjectives?). Studies in Language 10. 353–389. Willemse, Peter. 2005. Nominal reference-point constructions: Possessive and esphoric NPs in English. Leuven: University of Leuven dissertation. Wood, Johanna L. 2002. Much about such. Studia Linguistica 56. 91–115. Zanuttini, Raffaella & Paul Portner. 2003. Exclamative clauses: at the syntaxsemantics interface. Language 79 (1). 39–81.

Subject index absolute 43, 101, 102 addressee honorifics 63 adjective-intensification 35, 221n51 adjective-intensifier 25, 35, 40, 43, 83, 141-142, 156, 173, 222 agreement 66 all 165, 167, 167n35 ambiguity 8, 90-92, 194, 282 amplifiers 38 analogy 155, 157, 264 anaphora 101, 213 anaphoric 185, 188, 217, 220 antonyms 33, 34 approximators 38, 39 attitudinal 84 see also intersubjectivity, attitudinal binominal 171 bleached intensifiers 35, 43 blended construction 136, 141, 165 bondedness 106 booster 38, 140, 223, 244 boundary 39 bounded 39, 133, 176 adjectives 33, 102 descriptive modifier 125-127, 146-148, 160-161, 165 construal 34, 128, 162 see also unbounded boundedness construal of adjectives 8, 34, 124, 145, 153, 155, 177, 280 of nouns 131, 148, 151, 162, 167, 283 bridging context 90, 92, 191

cardinal numbers 40, 47, 48 catalyst 168 cataphora 190, 213 cataphoric 185, 188, 217, 220 categorization 6, 20, 25-27, 278 see also subcategorization central adjective 28, 29, 30 class 14, 209 class-based 13, 21 classification 20 classifier 16, 18, 25, 27 ad hoc 28, 156 classifying 53, 156-157, 203-206, 229n57, 258 coalescence 99, 248 cognitive coordination 57 cognitive-functional 9, 13, 14, 51 Cognitive Grammar 22, 57, 190, 215 cohesive 16, 84 collocational 15, 92, 94, 99, 125, 178, 225, 246, 259 context 8, 94-96, 139, 145, 183, 282 extension 170, 178, 228, 281, 282 overlap 136, 140, 153, 168, 176, 178 preference 127, 182, 184, 195 range 134, 139, 146 collocationally driven 145, 157, 201, 282 comparative 43, 195, 198, 213, 281 comparison 200 complementaries 33, 124 complementation 20, 51, 66 complete 99, 104, 125-145, 210 compound 13, 26, 27

306

Subject index

see also subjective compound compounding 255, 271, 273 see also subjective compounding compromisers 38 conceptualizer 58, 81 concessives 65, 94 conjunctions 66 connectives 58, 64, 65, 84 connotation 227, 246, 261, 263, 265 constituency 16, 74, 76 construal 57, 71, 81 construction 9, 13, 14, 90 constructional 13, 14, 51 contesting context 93 context-expansion 88 contextual modulation 261, 270 contrastive 131, 188, 189 conventionalization 91 coordination 31, 54, 267 critical context 91, 92 decategorialization 88, 99, 99n28, 103 definite 45, 46, 83, 88, 131 definiteness 218-222, 236 degrammaticalization 96 degree modification 34-44, 75, 77, 277 qualitative 223, 242, 245 quantitative 171, 223, 242, 245 degree modifier 25, 72 degree noun 26, 71, 95, 141, 145, 226, 228 deictic 15, 16, 18, 48 deictification 8, 99, 133, 142, 157, 165, 176, 191, 194, 206, 279, 282, 284 deixis 48, 62, 100 delexicalized 259 demonstrative 45, 49, 67-68, 88, 189, 220

dependency 9, 13, 16, 51, 52, 74, 76 descriptive modification 28, 29-34, 277 descriptive modifier 25, 53, 83, 258 objective 30, 181-183, 229n57 subjective 30, 184-185 de-subjectification 1, 3, 9, 107, 113, 253, 274, 284 determination zone 6, 25, 44-49, 277 determiner 46, 54, 239, 244 see also secondary determiner dialogic context 8, 195, 200 dialogicity 92, 93, 201 diminishers 38, 39 directionality 96, 113, 255 see also unidirectional discourse markers 63, 64 divergence 88 downscaling 242 downtoner 38, 244 dual reference 214 dynamic 14, 15 emphasizers 36 entrenchment 206 Epithet 16, 18, 18n6, 28 existential 219 experiential 16, 17 expressive 82 extreme adjectives 33, 102, 129 field-like model 6, 73-78, 275, 279 focus 6, 31, 60 adverb 4, 198 particles 93, 94 focusability 41 focusing subjuncts 181 foregrounded 124, 145, 281 foregrounding 8, 43, 102, 104, 105, 151, 157, 280 form-meaning pairing 9, 13, 14, 277 functional 9, 15, 18, 22

Subject index functional-structural 277 generic 46, 213 generalized instantiation 213, 218, 235, 243 givenness 44 gradability 27, 37, 43, 177, 223 gradable noun 26, 71 see also degree noun gradience 21 gradient 55, 57, 59, 277 grammaticalization 8, 82, 86-96, 103, 106, 139, 145 see also degrammaticalization ground 56, 57, 100 grounding 23, 54 head 16, 25, 26 hedging 63 host-class expansion 88 hypoanalysis 167 ideational 15, 16, 66, 74, 76, 82, 278 see also subjectivity, ideational identification 44 see also instance-identification see also type identification see also quantifying identification identifying 8, 130-134, 148-151, 162-165, 185-194, 211-222, 235236, 243-244 implicature 89, 91, 92, 191, 199, 201, 226n54, 233 inclusive reference 131, 132, 148, 163, 283 indefinite 45, 46, 83, 180, 190, 212 indefiniteness 218-222, 236 independent modification 36, 53 information structure 74, 77 instance-identification 218, 236, 242 instantiation 23, 47, 191, 214n48

307

intensification scale 166, 195, 223,245 interpersonal 15, 18n6, 73, 74, 76, 82 see also subjectivity, interpersonal interrogation 43, 60 interrogative 222n53, 242, 243-244 intersubjectification 81, 88, 106, 113, 133, 142, 284 see also textual intersubjectification intersubjective 67, 85, 278 intersubjectivity 55, 57-59 attitudinal 63, 69 metadiscursive see intersubjectivety, responsive pragmatic 62, 69 responsive 63 see also textual intersubjectivity invited inferencing 89, 90, 151, 155, 157 isolating context 91 joint (focus of) attention 6, 48, 67, 106, 133, 278 layering 3, 13, 87 leftward movement 1, 8, 9, 106-107, 274, 284 lexicalization 3, 253, 254, 256, 273 logical 16 lovely 29, 36-36, 107 metadiscursive 63 see also intersubjectivity, metadiscursive minimizers 38, 39 MI-score 173, 227n55, 229n56, 259, 260, 263, 265, 268, 271, 272 mass noun 132, 163 maximality modifier 140 maximizers 38, 140 modifier-head structure 5, 13, 52 modification 20, 28, 51

308

Subject index

zone 6, 25, 28-44 see also independent modification see also recursive modification multivariate 16 negation 58, 60, 200 nominal 20 non-inherent adjective 28, 29n10 noun-intensification 36-44 noun-intensifier 2, 6, 25, 40, 95, 135-141, 151-156, 165-173, 194203, 222-230, 237-238 closed scale 135-139, 165, 166 open scale 170, 178 noun phrase (NP) 2, 76, 77, 81, 83 canonical 52 non-canonical 15n4 number 45 numerative 4, 16, 18 objectivity 27, 55, 56 obligatorification 88 old 99, 107, 256-266, 271-273 onset context 8, 90-96, 139, 195 order 51, 55, 77 ordinal numbers 40, 48 own 187, 187n39, 196, 199, 282 paradigmaticization 99 parameters 87 particular 179-203, 206-208, 210 particularizer 181, 197, 198 partitive 219 pathway (of change) 1, 83, 84, 98107, 110, 113, 157, 177, 178, 179, 281, 282, 283 periodic 74, 76 persistence 88 phoric 48, 211, 212, 212n46, 218, 235 see also anaphoric see also cataphoric

politeness 63 possessive 46, 75, 107, 186, 187, 196, 282 Post-deictic 18, 47 postdeterminer 4, 44 pragmatic strengthening 89, 105, 157, 191, 233, 239, 281 predeterminer 4, 44 predicative alternation 28, 29, 31 profile 22, 51 proportional modifiers 33 propositional 31, 32, 32n11, 46, 82, 85, 83, 283 prosodic 6, 76, 77, 275, 279 prosody 74 see also semantic prosody pseudocleft 93 Qualifier 18 quantification 20, 23, 39, 44 absolute 170 relative 165, 166 scale 40, 177 universal 131, 162, 167, 281 quantifier 40, 45 absolute 18, 39, 47, 104 relative 18, 40, 44, 47, 48, 104 universal 188, 189 quantifying identification 8, 165, 242 really 35, 141 reanalysis 92, 99, 103, 167, 261, 264, 273 reconfiguration (of scales) 166, 173 recursive modification 16, 28, 36, 52, 77, 267 reference point 196 subjective 133, 165, 176 relation 186, 196 reinforcers 36 reinforcing 101, 123 relativizers 66 representational meaning 15

Subject index representative instance 46, 217, 234, 235 rightward movement 3, 9, 106-107, 253, 274, 284 routinization 261, 270 scalar 38, 43, 96, 105, 197 scalarity 39, 40, 198, 199 scale 34, 38, 40, 42, 166, 198 closed 39, 140, 152, 155, 177 open 39, 47, 140, 155, 177, 181 see also intensification scale see also quantification scale scope 34, 43, 54, 103, 223, 246 scopal adjectives 72 secondary determination 97, 114 secondary determiner 2, 13, 45, 4749, 134, 211 emphasizing 131, 132, 148, 150, 162 focusing 189-191, 192 individuating 185, 188-189, 192 linking 185, 186-187, 192 semantic change 82, 83, 151 semantic feature sharing 259 semantic field 38, 115, 123 semantic prosody 75, 94, 153, 155, 203, 246, 270 semanticized 59, 63, 89, 93, 151, 167, 196, 281 semi-determiner 47 similative 213 size noun 21n8, 89, 94, 171, 171n36 specialization 88, 248, 248n63 specialized context 8, 92-96, 165, 166, 195, 199, 200, 282 specific 179-194, 203-208 specification 20, 45 specificity 185, 191 specifying adjective 47 speech event 15, 23, 176 stress 258 structure 51, 74

309

structuralist 21 subcategorization 16, 27 subjectification 59, 81-88, 83n27, 100, 103, 106, 113, 139, 145, 150, 255, 274, 284 see also de-subjectification subjective 75, 128, 129, 162, 278 subjective compound 71, 253-271, 284 subjective compounding 78, 256 subjective-objective continuum 5, 6, 30, 55, 77, 255, 274, 278 subjectivity 43, 55, 56, 59-62, 77, 81 ideational 31, 32n11, 43, 60, 278 interpersonal 41, 60, 73, 278 pragmatic 59 submodification 53 see also recursive modification such 105, 209-252 superlative 218, 249 switch context 91 syntagmatic variability 106 T-score 173, 227n55, 229n56, 260, 268, 272 textual 64-69, 76, 113, 278, 279 component 16, 64, 74, 82, 106 intersubjectification 142, 176 intersubjectivity 6, 48, 69, 133, 150, 211, 278, 283, 284 subjectivity 84 token reference 221 total 145-159, 210 trigger context 177 truth-conditional 32 type-identification 211, 212, 218, 236, 242 type-identifiability 46, 190, 212 type nouns 99, 220 type phoricity 211-213, 235 type reference 221 type specification 22, 28, 45, 46, 54, 222

310

Subject index

unbounded 33, 34, 102, 162, 181185 unidirectional 82, 86, 96, 106, 177, 284 see also directionality univariate 16, 52 univerbation 255 upscaling 242 utter 95

vague 128 VALUE adjective 35, 36n12 very 13, 43, 48, 84, 140, 222 what 242-252 whole 159-176 zero-determiner 211n45, 246n62 zulk 234-241