Argument and Rhetoric: Adverbial Connectors in the History of English 9783110216066, 9783110205589

Citation preview

Argument and Rhetoric. Adverbial Connectors in the History of English

Topics in English Linguistics 64

Editors

Bernd Kortmann Elizabeth Closs Traugott

De Gruyter Mouton

Argument and Rhetoric. Adverbial Connectors in the History of English by

Ursula Lenker

De Gruyter Mouton

ISBN 978-3-11-020558-9 e-ISBN 978-3-11-021606-6 ISSN 1434-3452 Library of Congress Cataloging-in-Publication Data Lenker Ursula. Argument and rhetoric : adverbial connectors in the history of English / by Ursula Lenker. p. cm. ⫺ (Topics in English linguistics ; 64) Includes bibliographical references and index. ISBN 978-3-11-020558-9 (alk. paper) 1. English language ⫺ Adverbials. 2. English language ⫺ Connectives. 3. English language ⫺ History. 4. English language ⫺ Grammar, Historical. 5. English language ⫺ Grammar. I. Title. PE1326.L46 2010 4251.76⫺dc22 2009051326

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. ” 2010 Walter de Gruyter GmbH & Co. KG, 10785 Berlin/New York Cover image: Brian Stablyk/Photographer’s Choice RF/Getty Images Typesetting: OLD-Media OHG, Neckarsteinach Printing: Hubert & Co. GmbH & Co. KG, Göttingen ⬁ Printed on acid-free paper Printed in Germany www.degruyter.com

Acknowledgements “Adverbs are an oppressed part of speech, and probably bitter about their ill treatment. If they were ever organized the upper class of nouns and verbs had better look out” (Landau 1984: 78–79). The present book, which was accepted as a Habilitationsschrift at the Ludwigs-Maximilians-Universität Munich in the winter term of 2006/2007 (this date basically also marks the end of the full coverage of literature quoted), tries to remedy the ill treatment of at least one class of adverbs: adverbial connectors. I am deeply indebted to the mentors and referees of this thesis, Professor Hans Sauer, Professor Ulrich Detges and Professor Richard Janney, who provided invaluable information, advice and encouragement while I was writing this book. I have to thank Dick Janney in particular for drawing my attention to the relevance of “(your) little pointing words”, a remark which acted as a catalyst for my ideas on the importance of different means of discourse deixis for the subject of adverbial connectors. I am also very grateful to Hans Sauer and my father, Willibald Lenker, without whose unyielding perseverance this monograph might never have been written. I am very deeply indebted to Helmut Gneuss, who gave me as good a start in academic research as anyone could wish for: from my early student days, he has always supported me in my various strands of academic interest (even in fields he would never have researched himself), unfailingly providing a plethora of references and critical comments. I would also like to thank Alfred Bammesberger, Manfred Markus, Judith Huber and Friedrich Heberlein for their very helpful critical comments on an earlier version of this book, and the series editors, Bernd Kortmann and Elizabeth Traugott, for accepting the book for the Topics in English Linguistics. A number of readers and reviewers of earlier versions of selected chapters, in particular Cynthia Allen, Walter Hofstetter, Lucia Kornexl, Andreas Mahler, Anneli Meurman-Solin and Elizabeth Traugott have substantially improved sections in this book with their pertinent remarks. At the very beginning of my research on this topic, conversations with Otto Gsell and Klaus Schubert on trains and in the language centre corridor, which seemed casual at the time, were decisive in encouraging me to focus on the “mixed bag” of the word class “adverb”. In the final stages of completing and proof-reading the manuscript, Stefanie Beckstein, Veronika Bischofberger, Christina Grießl, Christine Haunz and, at Mouton de Gruyter, Wolfgang Konwitschny and Birgit Sievert gave wonderful support. Steffi Beckstein made these tedious weeks bearable with her infinite patience and, in particular, her upbeat personality.

vi   Acknowledgements “Only connect”, from Howards End, was the slogan of a telephone company some years ago – and so, above all, I would like to thank all my friends and family who helped me to stay connected to a world outside the realm of adverbial connectors during all these years. In particular, of course, I have to thank Michael Burghart, who provided so much more than the music. Munich/Eichstätt, December 2009

Contents

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

List of abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xv

1.

The framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 1.7. 1.8. 1.9.

Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Earlier research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aims of the study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rhetoric and stylistic aspects: a sample study . . . . . . . . . . . . . . . Early Modern and Late Modern English . . . . . . . . . . . . . . . . . . . The inventory of adverbial connectors . . . . . . . . . . . . . . . . . . . . The text corpus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Presentation of corpus findings . . . . . . . . . . . . . . . . . . . . . . . . . . Outline of the present study . . . . . . . . . . . . . . . . . . . . . . . . . . . . .



1 2 4 6 9 9 10 16 20

2.

Clausal connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

2.1. 2.2. 2.3. 2.4.

Clausal connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Connectors – connects: definitions . . . . . . . . . . . . . . . . . . . . . . . Connectors: coordinating conjunctions, subordinating conjunctions, adverbial connectors . . . . . . . . . . . . . . . . . . . . . . . Connectors and information processing . . . . . . . . . . . . . . . . . . .



22 23



25 28

3.

The category “adverb” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

3.1. Adverbs and adverbials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Classification of adverbials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. “Circumstance adverbials” or “adjuncts” . . . . . . . . . . . . 3.2.2. “Stance adverbials” or “disjuncts” (content/attitudinal and style disjuncts) . . . . . . . . . . . . . . 3.2.3. Adverbial connectors (”linking adverbials” or “conjuncts”) . . . . . . . . . . . . . . . 3.3. Semantic categories of adverbial connectors . . . . . . . . . . . . . . . . 3.4. “Pure” and “impure” connectives . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Adverbial connectors in Present Day English: corpus findings . . 3.5.1. The corpus of the Longman Grammar . . . . . . . . . . . . . .

1



33 35 36



36



37 39 41 42 42

viii   Contents 3.6.

3.5.2. Different types of adverbials over the core registers . . . . 3.5.3. Syntactic realizations of adverbials . . . . . . . . . . . . . . . . 3.5.4. Positions of adverbials . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.5. Linking adverbials: distribution of semantic categories . 3.5.6. Summary: Present Day English corpus findings . . . . . . Adverbial connectors – discourse markers . . . . . . . . . . . . . . . . .



42 43 43 44 45 46

4.

Adverbs and conjunctions in earlier metalinguistic thought .

49

4.1. 4.2. 4.3. 4.4. 4.5. 4.6.

Introductory remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Greek and Latin tradition . . . . . . . . . . . . . . . . . . . . . . . . . . . Ælfric’s Old English Grammar . . . . . . . . . . . . . . . . . . . . . . . . . . Grammatical treatises of the Middle English period . . . . . . . . . . The prologue to the revision of the Wycliffite Bible . . . . . . . . . . Early Modern English dictionaries and grammars . . . . . . . . . . .



49 49 51 53 54 56

5.

Connectors in Old English . . . . . . . . . . . . . . . . . . . . . . . . . . . .

58

5.1. Semantic and syntactic polyfunctionality . . . . . . . . . . . . . . . . . . 5.2. Circumstance adverbs with connective force (Group B‑a): temporal and spatial adverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1. OE eft ‘again’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2. OE nu ‘now’ (Group B‑a, Group C) . . . . . . . . . . . . . . . . 5.2.2.1. Temporal nu . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2.2. Text-deictic nu . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2.3. Adverbial connector nu – transition/result 5.2.2.4. Adverb/conjunction nu in correlative constructions . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.2.5. Conjunction (subordinator) nu . . . . . . . . . . . . . 5.3. Ambiguous adverbs/conjunctions (Group C) . . . . . . . . . . . . . . . . 5.3.1. OE þa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2. OE þeah, swaþeah, (swa)þeahhwæþere – contrast/concession . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Position of adverbial connectors . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.1. Nacherstposition – ‘post-first-position’ . . . . . . . . . . . . . . 5.4.2. Present Day German adverbial connectors in post-first-position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.3. Post-first-position of adverbial connectors in Old English 5.5. Pronominal connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.1. Pronominal connectors in Old English . . . . . . . . . . . . . .



58



59 60 61 61 61 62



62 63 64 64



66 67 67



68 70 72 72

Contents   ix

5.5.2. Pronominal connectors in the history of the Romance languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.3. Connectors in Louisiana French and French-based creoles

73 74

6.

Adverbial connectors in the history of English . . . . . . . . . . . .

76

6.1. 6.2. 6.3. 6.4. 6.5. 6.6.

Introductory remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The diachrony of coordinating and subordinating conjunctions: general tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The diachrony of adverbial connectors: general tendencies . . . . New adverbial connectors in the sub-periods of English: methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The inventory of Present Day English connectors: donor periods . Middle English: a period of experiment . . . . . . . . . . . . . . . . . . .



76



76 77



78 80 85

7.

Adverbial connectors: morphology . . . . . . . . . . . . . . . . . . . . .

88

7.1. 7.2. 7.3. 7.4. 7.5. 7.6.

The expansion of the English lexicon . . . . . . . . . . . . . . . . . . . . . Major morphological changes . . . . . . . . . . . . . . . . . . . . . . . . . . . Simple adverbs, compounds and derivations . . . . . . . . . . . . . . . . 7.3.1. Simple adverbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.2. Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.3. Derivations: adverbs in ‑lice, ‑es and ‑ways/-wise . . . . . Pronominal connectors – lexicalized phrases . . . . . . . . . . . . . . . Pronominal connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.1. Pronominal connectors in Old English . . . . . . . . . . . . . . 7.5.2. Pronominal connectors in the history of English . . . . . . Lexicalized (prepositional) phrases . . . . . . . . . . . . . . . . . . . . . . . 7.6.1. General tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.2. Lexicalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.3. Verbal and nominal phrases . . . . . . . . . . . . . . . . . . . . . .

8.

Cognitive source domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

8.1. 8.2. 8.3. 8.4.

General tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Source domain time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Source domain place/space . . . . . . . . . . . . . . . . . . . . . . . . . . . . Source domain truth/fact . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.1. General tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .



88 92 94 94 94 95 96 98 98 100 102 102 103 104

106 108 110 114 114

x   Contents

8.4.2. 8.4.3. 8.4.4. 8.4.5. 8.4.6. 8.4.7. 8.4.8. 8.4.9.



114 116 117 122 124 127 128 128

9.

Shifting deictics in the history of English causal connection .

131

9.1. Earlier research on causals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2. Causal connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.1. The relation cause: cause – result vs. result – cause . 9.2.2. Present Day English causal connectors: corpus findings 9.2.3. Causal connectors: word classes and topology . . . . . . . . 9.2.4. Semantic and pragmatic parameters . . . . . . . . . . . . . . . . 9.2.5. Information processing . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3. OE forþæm, forþon, forþy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.1. Forms and functions of forþæm, forþon, forþy . . . . . . . . 9.3.2. Expressions for causal relations in Early and Late West Saxon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4. Discourse deixis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.1. Forþæm: morphological make-up and discourse deixis . 9.4.2. Pronominal connectors in Present Day German . . . . . . . 9.4.3. Deictic elements in English causal connectors . . . . . . . . 9.5. Causal connectors in the history of English . . . . . . . . . . . . . . . . 9.5.1. Causal connectors in English translations of Boethius’ De Consolatione Philosophiae . . . . . . . . . . . . . . . . . . . . 9.5.2. Causal adverbial connectors in the history of English . . . 9.5.3. Deixis in new adverbial connectors . . . . . . . . . . . . . . . . 9.5.4. ME for that . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.5. “Recursive” for – Latin nam/enim . . . . . . . . . . . . . . . . . 9.5.6. Subordinators: for as much as, since, because . . . . . . . . 9.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

132 133 133 134 134 136 138 140 140

10.

Conversational implicatures: concession/contrast . . . transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OE soþlice – ME soothly . . . . . . . . . . . . . . . . . . . . . . . . ME forsooth(e) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OE treowlice – ME/EModE/PDE truly . . . . . . . . . . . . . PDE indeed and in fact . . . . . . . . . . . . . . . . . . . . . . . . . . Regularities in change . . . . . . . . . . . . . . . . . . . . . . . . . . . Truth, facts and communicative principles . . . . . . . . . . .

143 147 147 149 151 152 152 154 157 158 161 164 166

contrast/concession . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

10.1. cause – condition – contrast: affinities . . . . . . . . . . . . . . . . . 168 10.2. concession – contrast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

Contents   xi

10.2.1. concession . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.2. contrast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.3. concession/contrast: information structure . . . . . . . . 10.2.4. Subordinators – adverbial connectors: mood distinctions . 10.3. Cross-linguistic patterns in the origin of concessive connectors . 10.4. The diachrony of English contrastive/concessive adverbial connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4.1. General tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4.2. Antithetic/reformulatory adverbial connectors . . . . . . . . 10.4.3. Contrastive/concessive connectors in English: general tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4.4. Shifting deictics in English contrastive/concessive connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4.5. Patterns in the origin of English concessive connectors . 10.5. OE þeah – PDE though . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5.1. Long-term developments: grammaticalization . . . . . . . . 10.5.2. OE þeah . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5.3. Contrastive adverbial connectors from Middle to Present Day English . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5.3.1. Though . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5.3.2. Yet and still . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5.3.3. However . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6. Sentence-final connectors in Present Day English . . . . . . . . . . . 10.6.1. Corpus findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6.2. Information structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6.3. Sentence-final connectors in Present Day German . . . . . 10.6.4. Sentence-final connectors in the history of English . . . . 10.7. Present Day English sentence-final though . . . . . . . . . . . . . . . . . 10.7.1. Earlier research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7.2. PDG obwohl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7.3. PDE but . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7.4. Although and though in the LLC . . . . . . . . . . . . . . . . . . . 10.7.4.1. Quantitative findings . . . . . . . . . . . . . . . . . . . . 10.7.4.2. Functions of the subordinator (al)though . . . . . 10.7.4.3. Functions of sentence-final though . . . . . . . . .

169 169 171 173 173

190 190 191 194 197 197 198 198 200 200 200 202 203 207 207 208 210

11.

addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

214

11.1. General tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2. The diachrony of reinforcing adverbial connectors . . . . . . . . . . .

214 214

175 175 175 178 178 181 184 184 186

xii   Contents 11.3. 11.4. 11.5. 11.6. 11.7.

11.2.1. General tendencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2.2. Also . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2.3. New coinages: item, plus, too . . . . . . . . . . . . . . . . . . . . . 11.2.4. Iconic principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Equative connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summative connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appositive connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enumerative (listing) connectors . . . . . . . . . . . . . . . . . . . . . . . . . Genre-dependency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .



214 215 219 219 220 221 222 222 224

12.

transition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

12.1. 12.2. 12.3. 12.4.

Preliminary considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Source domain uncertainty/doubt: peradventure, perchance, perhaps . . . . . . . . . . . . . . . . . . . . . . . Interrogatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The diachrony of transitional connectors . . . . . . . . . . . . . . . . . . 12.4.1. Paradigm shift: truth – fact . . . . . . . . . . . . . . . . . . . . 12.4.2. Regularities in semantic change . . . . . . . . . . . . . . . . . . .

13.

Perspicuity and the “New Rhetoric” . . . . . . . . . . . . . . . . . . . . 233

13.1. 13.2. 13.3. 13.4. 13.5.

Collocations vs. medial position of adverbial connectors . . . . . . Sentence-initial collocations . . . . . . . . . . . . . . . . . . . . . . . . . . . . Medial positions of adverbial connectors . . . . . . . . . . . . . . . . . . Corpus findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4.1. Present Day English . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4.2. Old English to Late Modern English . . . . . . . . . . . . . . . Copia – perspicuitas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5.1. Perspicuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5.2. Punctuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5.3. The Scottish Rhetoricians: “The New Rhetoric” . . . . . . 13.5.4. George Campbell’s The Philosophy of Rhetoric (1776) .

14.

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

227



228 229 230 230 231

233 234 235 237 237 238 241 241 242 243 243

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 1. 2.

Corpora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Contents   xiii

3. 4.

Dictionaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 Secondary sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

Appendix Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 A.1. Adverbial connectors: items . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 A.2. Alphabetical index of adverbial connectors . . . . . . . . . . . . . . . . 282 Word index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 Subject and name index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 Appendices on CD-ROM Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.1. Listing/additive adverbial connectors . . . . . . . . . . . . . . . . . . . . . B.2. Summative adverbial connectors . . . . . . . . . . . . . . . . . . . . . . . . . B.3. Causal/resultive adverbial connectors . . . . . . . . . . . . . . . . . . . . . B.4. Contrastive and concessive adverbial connectors . . . . . . . . . . . . B.5. Transitional adverbial connectors – interjections . . . . . . . . . . . .



295 295 307 308 317 329

Appendix C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C.1. Corpus texts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C.2. Selected corpus texts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C.2.1. “Treatises and Homilies” . . . . . . . . . . . . . . . . . . . . . . . . C.2.2. Translation of Boethius’ De Consolatione Philosophiae . . . . . . . . . . . . . . . . . . . .



343 343 353 353

358

List of abbreviations A Adverbial BT Bosworth & Toller (1882–1898) ByrM Byrhtferth’s Enchiridion (ed. Baker & Lapidge 1995) CE- Initial Letters of Corpus Texts from Early Modern English CH Clark Hall 1984 CL- Initial Letters of Corpus Texts from Late Modern English CLMET A Corpus of Late Modern English Texts (de Smet 2005) CM- Initial Letters of Corpus Texts from Middle English CME Corpus of Middle English Prose and Verse CO- Initial Letters of Corpus Texts from Old English Conj. Conjunction DOE Dictionary of Old English DOEC Dictionary of Old English Corpus EETS Early English Text Society EMEDD Early Modern English Dictionary Database EModE Early Modern English EModE1 (1500–1570), EModE2 (1570–1640), EModE3 (1640–1710) Eustace Life of St Eustace; ed. Skeat 1900: 190–219 HC Helsinki Corpus Ind. Indicative Lat. Latin LLC London Lund Corpus of Spoken English LModE Late Modern English LModE1 (1710–1780), LModE2 (1780–1850), LModE3 (1850–1920) LOB Lancaster-Oslo/Bergen Corpus of British English M Merkmal = ‘property’ ME Middle English ME1 (1150–1250), ME2 (1250–1350), ME3 (1350–1420), ME4 (1420–1500) MED Middle English Dictionary O Object OE Old English OE1 ( –850), OE2 (850–950), OE3 (950–1050), OE4 (1050–1150)

xvi   List of abbreviations OED ON O. S. PDE PDG PP Prs. S Sg. Subj. V V2

Oxford English Dictionary Old Norse Old Series Present Day English Present Day German Prepositional Phrase Person Subject Singular Subjunctive Verb (Predicate) Verb-second

Symbols ° [ ]

ambiguous adverb/conjunction in Appendix A.1: not attested in the corpus texts

1.  The framework

1.1.  Particles In his Essay Concerning Human Understanding, one of the first linguistic accounts explicitly addressing what we now call textual organization or cohesion,1 John Locke asserts that it is the right use of “Particles” which is crucial for the clearness and beauty of good style: These Words, whereby it signifies what connection it gives to the several Affirmations and Negations, that it unites in one continued Reasoning or Narration, are generally call’d Particles: and ’tis the right use of these, that more particularly consists the clearness and beauty of good Stile. … And to express well such methodical and rational Thoughts, he must have words to shew what Connexion, Restriction, Distinction, Opposition, Emphasis, etc. he gives to each respective part of his Dis‑ course (Locke [1690] 1975: 471; emphasis by Locke).

As is evident from this description, Locke’s “Particles” are co-referential with the linguistic elements we now call “clausal connectives”.2 Locke highlights that he regards these elements not merely as an adornment or frequent supplementary device to uncover textual organization,3 but as indispensable for the understanding of an utterance: To mistake in any of these, is to puzzle, instead of informing, his Hearer: and therefore it is, that those words, which are not truly, by themselves, the names of any Ideas, are of such constant and indispensable use in Language … (Locke [1690] 1975: 471).

The novelty and originality of this view that “Particles” play a central role in language is made explicit when Locke proceeds by criticizing previous grammars and grammarians for their ill-treatment or even neglect of these linguistic elements: 1 For details on the treatment of connectives in European and English language scholarship, see Lenker (2003) and below, Chapter 4. 2 The term “connective” is used as an umbrella term for all kinds of linguistic items signalling a linkage of sentences or chunks of discourse. The term “connector” more specifically refers to paratactic connectives, in particular adverbial connectors. 3 For the repeatedly expressed view of adverbial connectors and so‑called discourse markers as supplementary or even superfluous features of language, and for a discussion of the overlap between coordinators, subordinators, adverbial connectors and “discourse particles” as well as cases of polyfunctionality, see Chapters 2 and 3.

2   The framework This part of Grammar has been, perhaps, as much neglected, as some others overdiligently cultivated. ’Tis easy for Men to write, one after another, of Cases and Genders, Moods and Tenses, Gerunds and Supines: … But though Prepositions and Conjunctions, etc. are names well known in Grammar, and the Particles constrained under them carefully ranked into their distinct subdivisions; yet he who would shew the right use of Particles, and what significancy and force they have, must take a little more pains, enter into his own Thoughts, and observe nicely the several Postures of his Mind in discoursing (Locke [1690] 1975: 471–472).

1.2.  Earlier research More than 300 years after Locke’s attempt to draw attention to the importance of connectives and to foster their attention with grammarians, we find that this field is still neglected in linguistics.4 In 2003, the compendium on German connectives, the Handbuch der deutschen Konnektoren, has to state in an almost identical way: Mit der Wahl dieses Gegenstandsbereiches behandelt das Handbuch Einheiten des deutschen Wortschatzes [= Konnektoren “connectives”], die weder in Grammatiken noch Wörterbüchern noch in den texttheoretischen und konversationsanalytischen Arbeiten befriedigend beschrieben sind. Wie nicht anders zu erwarten ist, werden in Grammatiken nur die systematischen Eigenschaften der betreffenden Einheiten beschrieben. Dabei stehen dort jedoch traditionell in der Regel semantische Klassenbildungen im Vordergrund (Pasch et al. 2003: xv). ‘The subject area chosen by the handbook are items of the German vocabulary [= connectors, U. L.] which have neither been satisfactorily described in dictionaries, nor in studies on text theory and conversation analysis. As expected, it is only the systemic properties of the respective items which are described in grammars. Traditionally, however, these descriptions generally focus on the distinction of various semantic classes’.

This summary testifies to a lack of synchronic in‑depth studies in the field of connectives.5 It is, however, in particular the history of connectives that has received only very little attention. With regard to the history of the English language and the establishment of modern English prose, there are only two studies which explicitly approach this field diachronically in a wider and systematic perspective:

4 For the first fuller treatment of connectives in Campbell ([1776] 1963), see below, Chapter 13.5.4. 5 For English, notable exceptions are the very comprehensive studies by Altenberg, in particular Altenberg (1984, 1986).

Earlier research   3

In a study situated at the interface between literary studies6 and linguistics, Sylvia Adamson (1999) relates the emergence of new forms of sentential connection to major changes in literary style and text production in Early Modern England, namely to the evolution of the “plain style” as an indicator of perspicuitas – an idea connected with the stylistic ideals of the Royal Society – in contrast to earlier, very different stylistic ideals such as copia (for details, see below, Chapter 13.5). In the field of linguistics proper, Bernd Kortmann’s study of adverbial subordinators (1997), though predominantly concerned with typological and crosslinguistic data (see the material collected in 1997b), also gives a short outline of the history of adverbial subordinators, one of the most important groups of connectives in English. The results of Kortmann’s study of the history of subordinators will here be used as a contrastive plane for the comparison with the findings for adverbial connectors.7 In addition to these more general accounts, there is a small number of studies examining individual items, concepts and relations in the field, such as Jucker 1997 (well), Enkvist & Wårvik 1987 (OE þa; and many further studies on OE þa such as, e. g., Kim 1992), Brinton 1996 (on various “pragmatic markers”), Lenker 2000 (OE soþlice, witodlice), Lenker 2003, 2007a (forsooth), Markus 2000 (ME wherefore, therefore, etc.), Österman 1997 (there‑compounds), Rissanen 1999a (rather), Schleburg 2002 (OE swa), Stanley 2000 (OE hwæt), Traugott 1997 (after all), Traugott and Dasher 2002 (indeed, in fact) and Fischer 2007 (instead; indeed, in fact, soþlice and witodlice). The semantic relation analysed most thoroughly is that of “concessives” which features prominently in studies of the semantic‑pragmatic approach to grammaticalization (see König and Traugott 1982; König 1985a, 1985b, 1988; Barth-­ Weingarten and Couper-Kuhlen 2002). In addition to the still central study on cohesion in Present Day English by Halliday and Hasan (1976), there are a number of studies on Present Day English discourse markers, which refer to the history of some of the adverbial connectors in question (e. g., Schiffrin 1987; Lenk 1998). With the exception of Halliday and Hasan’s systematic analysis of conjunctive elements in Present Day English (1976: 226–273), these studies concentrate on single connectives or patterns of language change, such as grammaticalization, and do thus not endeavour to provide a comprehensive treatment of connectives. A more integrated view of clausal connection in the history of English is ­offered in most of the papers presented at the workshop “Clausal Connection in 6 We also find some relevant information on connectives scattered over more general works on the evolution and establishment of English prose (see, for instance, Mueller 1984, McIntosh 1998 and Robinson 1998). 7 The present study is very much indebted to Kortmann’s study and was set off by my review of his book. For the details of the comparison, see in particular Chapter 6.

4   The framework the History of English” at the 13th International Conference on English Histori‑ cal Linguistics at Vienna in 2003. This workshop was initiated because the convenors Anneli Meurman‑Solin and Ursula Lenker regarded it as a drawback that the typologies of clause‑combining devices in English as well as other European languages are widely discussed in recent literature (Devriendt, Goossens, and van der Auwera 1996; Kortmann 1997; van der Auwera 1998), but have chiefly been construed by using secondary sources such as dictionaries and grammars (for this “grammar‑cum‑dictionary‑method”, see, e. g., Kortmann 1997: 53). Most of the workshop papers are therefore corpus‑based studies of various connectives in the history of English, focusing on subordinators (while, lest, since, albeit) or specific semantic domains, such as concessives or conditionals (see Lenker and Meurman‑Solin 2007). There is as yet, however, no comprehensive, corpus‑based treatment of adverbial connectors in the history of English. 1.3.  Aims of the study The present study tries to fill at least a segment of this large gap in historical (English) linguistics by corpus‑based analyses of the development of a particular word class in connector function – namely “adverbial connectors”, which are also called “conjuncts” (Quirk et al. 1985) or “linking adverbials” (Biber et al. 1999).8 The focus of the study therefore rests on the inventory and use of linguistic elements which explicitly mark textual cohesion on a level higher than the phrase.9 Even more specifically, the focus is on connectives which signal textual organization on a more global level in discourse, i. e. a level higher than the sentence.10 In Lehmann’s universal typology of clause linkage (Lehmann 1988), adverbial connectors belong – together with subordinators and coordinators (i. e. coordinating and subordinating conjunctions) – to the little integrated, explicit markers of clausal connection (in contrast to embedded constructions, relative connectives, non‑finite verb forms, absolute constructions, etc.).11 While 8 For the terminology employed in various grammars, see below, Chapter 3. 9 Some of these items can also mark relations on the level of the phrase; these are only included, however, if they are polyfunctional and if they can also function on the higher level of textual organization. 10 This position at the interface between the sentence and the text (paragraph) may be one of the reasons why they are a neglected subject in linguistics. Theories of syntax, such as all kinds of generativist approaches and also many functional perspectives, ignore them because they are above the level of the core sentence. 11 Most of these other strategies fulfilling a “connective” function, such as linking by non‑finite constructions (present participles, infinitives etc.) extra‑textual links and non‑linguistic structuring devices, which are not in the focus of this study, do not play

Aims of the study   5

the distinction between coordinators, subordinators and adverbial connectors is rather clear‑cut for Modern English,12 it has to be stressed here right at the beginning of this study that there was no such clear distinction in Old English, which predominantly uses so‑called “ambiguous adverbs/conjunctions” (for details, see below, Chapter 5). This first of all means that the whole system of clausal connection has been re‑structured in the history of English. While other, in particular typologically related, languages such as German, show much more stability in the morphological make‑up and use of subordinating conjunctions and adverbial connectors, English has only very few remnants of the older Germanic system present in Old English. The Old English inventory mainly consisted of polyfunctional items comprising an explicitly deictic element. In addition to forms comprising swa ‘so’ (cf. swaþeah or swaþeahhwæðere ‘although; nevertheless’), we very frequently find morphologically complex connectors comprising a pronominal element, such as OE forþæm, forþon or forþy ‘for, because; therefore’ (< for [preposition] ‘because of’ + þæm/þon/þy [demonstrative; dative þæm, instrumental þon/þy]).13 This scenario suggests that the first essential motivating force for the dramatic changes in the inventory of adverbial connectors starting in the early Middle English period are structural constraints and therefore typological properties of English.14 The break‑down of the case and gender inflections of both of the Old English demonstratives and, in particular, the use of that as a demonstrative, complementizer and general subordinating particle (replacing OE þe) triggered new structures which signal anaphoric relations on the surface, disambiguating subordinators from adverbial connectors. The complex problems involved in drafting this new inventory marking discourse deixis are reflected an important role in the history of the development, because they are either rather rare (e. g. full prepositional phrases) or emerge very late in the history of English, i. e. in the Early Modern English period when non‑finite constructions become increasingly grammaticalized. Only from LModE2 onwards, infinitives (e. g. to begin with, to conclude, to proceed) become more popular. In Present Day English, they do not amount to more than four per cent of linkage constructions (Altenberg 1984: 47). In the present sample from LOB‑D and FROWN‑D (68,000 words each), the token numbers are: infinitives (3), present participles (9), and past participles (7). 12 A notable exception is the status of PDE for; see Chapter 9. 13 The changes in the morphological make‑up of these forms and their relevance are discussed in Chapter 7. In Raible (1992: inserted foldout), these pronominal connectors form a typologically separate group: “Junktion durch Wiederaufnahme (eines Teils) des vorhergehenden Satzes (II)”. 14 For similar typological changes involving the loss of deictic elements in French creoles see Raible (1992: 172–177) and below, Chapter 5.5.

6   The framework in the manifold forms in Early Middle English (ME 1/2). The innovations in these periods, however, still show symptomatic regular, pronominal patterns, comprising the new demonstratives this and that (cf. additive, reinforcing over that / over this or resultive for that / this). Early Middle English thus emerges as a period of experiment and transition, and hence as the first decisive period for the re‑structuring of the system of adverbial connectors in the history of English. This re‑structuring also led to the loss of polyfunctionality in connectives and a distinction between coordinators and subordinators, which will, as one exemplary case, be followed in the development of the system of causal connectives in the history of English in Chapter 9 (i. e. the development of because, since and as in certain positional variants and contexts); the other sections of the book focus primarily on adverbial connectors. 1.4.  Rhetoric and stylistic aspects: a sample study The following introductory example of an (unfortunately rather small) experiment conducted by Mauranen (1993: 159–168) shows that a study of adverbial connectors cannot be confined to morphological, semantic and syntactic analyses. Mauranen presented the following text (Text 1) to a group of English native speakers with professional linguistic training. All of the participants worked as English lecturers at universities or as lectors or revisers for technological research reports and were selected because they were considered to be extremely sensitive to language and textual organization. The subjects were first invited to read the text and were then, in a second step, placed in groups of two and three and asked to come to a group decision on the following issues: whether the text was correct as it stood, and, secondly and, more specifically, if they felt that connectives should be added in any of the sentences. Text 1 (Mauranen 1993: 164) In a recent study on 5‑ to 6‑year‑old children, Astington confirmed that they see a strong link between promising something and actually doing it: “To promise” means “you do it”. This link is much stronger for children than for adults, which leads children to assert that an unfulfilled promise was not a promise, but a “lie”. For young children promising is not simply a speech act but something that includes execution of the promised action as well. It seems likely that children of 5 years or so understand that the use of the word promise entails commitment, but this understanding may be based on a simple rule “If you have said ‘I promise’ then you must do what you said you would.” Although the use of the word promise may often be a sufficient condition for becoming committed, it is certainly not a necessary condition. It is seldom used in everyday ex-

Rhetoric and stylistic aspects: a sample study   7 changes between adults, who tend to say “I will meet you at 6 o’clock”, “I’ll return your book tomorrow”, and so forth. The commitment, although informal, is binding. What distinguishes a commitment (I will return your book) from a statement of intention (I will stay in tonight) or a prediction of a future event (I will get wet) is the knowledge that someone else is relying on one to carry out the commitment and the knowledge that the other person is aware that one has made a commitment. The interesting question is whether young children can recognize commitment without the help of the word promise. Our aim in this study was to investigate whether child­ ren of between 5 and 10 years are aware of reliance as the essential and necessary condition for commitment.

After having read and discussed Text 1, the groups unanimously concluded that the text “was good as it stood”. They reported that they felt it was “clear”, “easy to read” and that it “required no additions” (Mauranen 1993: 166). They also agreed that no connectives needed to be added “at all”. In a second phase, however, the groups were shown the authentic version of the text, which had been altered by Mauranen for the first test phase. For Text 1, Mauranen had deleted connectives which were dispensable, i. e. which are not crucial for the grammatical correctness of the text because they are peri­ pheral to the sentence structure and act on a level higher than the local phrase, marking textual cohesion globally. The authentic text (Text 2) with the restored connectors reads as follows: Text 2 (Mauranen 1993: 164) In a recent study on 5‑ to 6‑year‑old children, Astington confirmed that they see a strong link between promising something and actually doing it: “To promise” means “you do it”. However, this link is much stronger for children than for adults, which leads children to assert that an unfulfilled promise was not a promise in the first place, but, rather, a “lie”. In other words, for young children promising is not simply a speech act but something that includes execution of the promised action as well. It seems likely, then, that children of 5 years or so understand that the use of the word promise entails commitment, but this understanding may be based on a simple rule, such as “If you have said ‘I promise’ then you must do what you said you would.” Although the use of the word promise may often be a sufficient condition for becoming committed, it is certainly not a necessary condition. Indeed, it is seldom used in everyday exchanges between adults, who tend to say “I will meet you at 6 o’clock”, “I’ll return your book tomorrow”, and so forth. The commitment, although informal, is nonetheless binding. What distinguishes a commitment (I will return your book) from a statement of intention (I will stay in tonight) or a prediction of a future event (I will get wet) is the knowledge that someone else is relying on one to carry out the commitment and, furthermore, the knowledge that the other person is aware that one has made a commitment. The interesting question is, therefore, whether young children can

8   The framework recognize commitment without the help of the word promise. Our aim in this study was to investigate whether children of between 5 and 10 years are aware of reliance as the essential and necessary condition for commitment.

The linguistic elements which had been removed in Text 1 are underlined here for reasons of transparency. They are spread over the whole text with a ratio of about one deletion per sentence. No deletions were undertaken in the first and the last sentence of the text, as these sentences set the frame for the textual cohesion of the passage and therefore do not show any employment of connectives on the level above the sentence. After having read the original version with all the connectives restored (Text 2), Mauranen’s subjects were asked to comment on the text again and evaluate it. The immediate reaction of the subjects was that they felt a “dramatic difference” between the two versions.15 More specifically, the emphasis was seen to have changed, and the text was said to be not only “easier to read”, but also to be “more logical” and “more convincing”. It was also perceived to have “more authority” (Mauranen 1993: 167–168). The fact that the test groups registered a “dramatic” difference between the two versions of the text – with the connectives having an overall effect of making Text 2 “more authoritative, logical and convincing” – shows that a study of the history of connectives in English must certainly not be confined to their etymology, morphology and semantics, but has to be situated at the interface between syntax, semantics, pragmatics, text linguistics and rhetoric. The central role or even indispensability of these connectives in the construction of a text, as highlighted by Locke, is corroborated by Mauranen’s experiment, which also shows that it is necessary to view the employment of these items in the wider perspective of the evolution of an English prose style. Yet, the preference of certain means of textual organization and the employment of linking adverbs as connectives is certainly not the only crucial driving force in the developments of adverbial connectors in the history of English. In Present Day English, for example, an author may choose among the adverbials therefore, thus and hence for marking the semantic relation result, and their use indeed exhibits some variability by author: most academic texts show a clear preference for either thus or therefore, usually using one item at least twice as often as the other (Biber et al. 1999: 889). This choice, however, is merely a choice in the lexical material employed. In both cases, the authors do not choose to mark the organization of the text by coordinating or subordinating conjunctions (be‑ cause, since, as or for), but by adverbial connectors. Although some changes 15 For the importance of stylistic considerations, see also Lehmann (1988: 210–213). For an early discussion, see Campbell (1776: 384–415) and below, Chapter 13.5.4.

Early Modern and Late Modern English   9

and choices in the use of adverbial connectors may indeed be due to stylistic predilections of a period or of individual authors (see below, Chapter 13), the more crucial changes in the history of English connectors thus seem to have been triggered by the typological and structural changes which set English – also in this respect – apart from other Germanic languages. 1.5.  Early Modern and Late Modern English After the period of experiment and variation in early Middle English, there are two further periods which are decisive for shaping the system and the use of adverbial connectors in English. In Early Modern English (EModE 1/2), when English develops into an Ausbausprache gradually being used as a national language also in the written medium, a number of new connectors are formed. In contrast to those coined in the Early Middle English period, however, these are not replaced again, but for the most part remain in the language until Present Day English (see Chapter 6.5). The English of the late eighteenth and nineteenth centuries (Late Modern English 2 and 3; 1780–1850 and 1850–1920) then sees crucial changes in (a) the preference of adverbial connectors over coordinating conjunctions and their collocations, and (b) the sentence position of adverbial connectors (see Chapter 13.1–13.3). Instead of sentence‑initial position – very often in collocation with a conjunction such as and or but or a connective such as for –, adverbial connectors become increasingly used in medial position. This medial position, how­ ever, is virtually only found in the written medium and is attested in less than 2.5 per cent of instances in the spoken medium (see Biber at al. 1999: 891): it thus sets the English of the written medium apart from that of the spoken medium (see Chapter 13.4). In spoken interaction, however, we also find a new position for adverbial connectors: adverbial connectors such as however and, in particular, though are increasingly placed at the very end of a sentence from the middle of the twentieth century onwards. The motivating forces and consequences of the sentence‑final positioning of adverbial connectors will be exemplified by a contrastive analysis of the use of PDE although, though, and sentence‑final though and German obwohl with main clause word order in Chapter 10.5. 1.6.  The inventory of adverbial connectors The present study covers many of the issues suggested by these introductory examples. It gives an – as complete as possible – account of the single‑word or lexicalized linguistic items which have been used as adverbial connectors in the

10   The framework history of English. Furthermore, it describes their employment as well as their meaning and their text structuring and information processing functions in relation to other coordinators, such as coordinating and subordinating conjunctions. The approach employed is therefore a decidedly diachronic and functional one. The study is mainly focussed on single‑word adverbs because corpus findings for Present Day English show that it is single adverbs which are predominantly employed in this function (Biber et al. 1999: 887, and Figure 10.26), although this adverbial linking function may theoretically be realized not only by single adverbs, but also by prepositional phrases, such as for that reason, in ­other words, or on the contrary. The distinction between prepositional phrases and single adverbs, however, is, even in Present Day English, by no means clear‑cut; in Present Day American English, for example, the lexicalization of the phrase after all is indicated in its spelling afterall as one word (see OED, s. v. after; 10.): (1)

Afterall, the movement of people, not vehicles, is what counts. (1976 Billings [Montana] Gazette 1 July)

Similarly, in Middle English, the phrase at /on /of last ‘finally’ is attested as alast (see MED, s. v. a‑last). Even longer phrases such as over and above ‘furthermore’, a popular reinforcing connector in Early Modern English, could be spelt as a single word hoverendebuv (see OED, s. v. over and above). Furthermore, since many of today’s single word items are univerbated and lexicalized prepositional (e. g. indeed) or verbal (e. g. albeit, howbeit, notwithstanding) phrases and since a distinction, especially in periods without a fixed orthography, is not unambiguously possible, lexicalized prepositional phrases were included in the detailed analysis of this book. Only fully transparent, i. e. non‑lexicalized phrases, such as in other words, for that cause, or the sum is, have been excluded (for the respective morphological and syntactic criteria, see Chapter 7.4). 1.7.  The text corpus The inventories of adverbial connectors of the respective periods were first gathered by a search in the relevant dictionaries (OED for the history of English; CH and BT for Old English, MED for Middle English and the EMEDD for Early Modern English) as well as in the grammars of the earlier periods of English (Mitchell 1985; Mustanoja 1960; Kerkhof 1982; Franz 1939) and of Present Day English (Quirk et al. 1985; Biber et al. 1999; Huddleston and Pullum 2002; Halliday and Hasan 1976). While it was thus possible to gain a good overview over the system of adverbial connectors in Old, Middle, Early Mod-

The text corpus   11

ern and Present Day English, the Late Modern English Period, i. e. the English of the eighteenth and the nineteenth century, remained a neglected period. The OED, for example, whose first edition was compiled in the nineteenth century, often does not have special entries for phrases which were at that time not yet lexicalized or had only recently been lexicalized, such as after all, above all, in all events, at any rate (see OED, s.vv. after, all, event, rate). Generally, the varieties of the late eighteenth to the early twentieth century had until recently to be described as “the Cinderellas of English historical linguistic study” (Jones 1989: 279). Although the general interest in the language of an increasingly more distant past and the change of emphasis within historical linguistics to socio‑historical and corpus‑based approaches has led to a surge of interest in Late Modern English, the more recent publications on the field (Bailey 1996; Romaine 1998; Lass 1999; Görlach 1999; Görlach 2001; Beal 2004) do not treat textual cohesion in any detail. This lack of material on Late Modern English, together with the plan to study not only the inventory, but also the changes in use and position of adverbial connectors in the history of English, created the need for a diachronic corpus‑based approach, including texts from Late Modern English, which are not covered by the Helsinki Corpus of English Texts. In order to study not only changes in the inventory, but more generally developments in the use of adverbial connectors (also in contrast to conjunctions) from Old English to Present Day English, the findings of the present study are based on a corpus of texts from all periods of English (see Appendix C.1) and, for the quantitative analyses, on two smaller corpora of selected comparable texts (see Appendix C.2.1. “Treatises and Homilies” and C.2.2. “Translations of Boethius’ De Consolatione Philosophiae”). The complete corpus is modelled on the Helsinki Corpus, which comprises texts from the earliest Old English period until 1710 (see Kytö 1996). For the periods no longer covered by the Helsinki Corpus, I have compiled a corpus in 70‑year sub‑periods (equivalent to the periods from ME3 to EModE3 of the Helsinki Corpus) until 1920 (LModE1 from 1710 to 1780, LModE2 from 1780 to 1850, LModE3 from 1850 to 1920). For each of these periods, I extracted 5,000‑word passages from texts of the Project Gutenberg and the Literature Online (LION) collections.16 In order to provide a comparable text basis, the 16 Most of the Late Modern English texts I had selected are now publicly available in the CLMET (Corpus of Late Modern English Texts), also modelled on the Helsinki Corpus in its 70‑year sub‑periods (see De Smet 2005). In Appendix C.1, I only list the exact sources of these corpus texts which are not part of the CLMET (these texts are marked as Project Gutenberg with text number or LION in appendix C.1). While the CLMET usually gives the full texts, I, as a rule, took the first ca. 5,000

12   The framework selection of texts was guided by the principles of the Helsinki Corpus with respect to the length of the passages and, as far as possible, the text types (for the problems concerning text type comparability, see Kohnen 2004: 81–128 and below). The target number was 100,000 words per period; furthermore, the respective corpus texts should comprise as many complete texts (e. g. homilies) by as many different authors as possible.17 For a first survey of the sub‑periods and the design of the corpus, see the following table (for details on the texts, see Appendix C.1): Period OE1 OE2 OE3 OE4 ME1 ME2 ME3 ME4 EModE1 EModE2 EModE3 LModE1 LModE2 LModE3 PDE TOTAL

Dates

Number of Texts

Words

–850 850–950 950–1050 1050–1150 1150–1250 1250–1350 1350–1420 1420–1500 1500–1570 1570–1640 1640–1710 1710–1780 1780–1850 1850–1920 1920–1990

1 69 13 12 10 3 13 22 13 16 20 13 18 18 49

1860 48780 76350 34390 73370 1440 100540 136630 97310 108770 136040 102890 99840 102410 103900 1,224,520

This table shows that the texts which have come down to us do not allow a detailed quantitative analysis on a large scale (for the periods from 1150 to 1700, see the detailed account in Kohnen 2004: 81–128). Apart from the general words from these full texts to create a corpus comparable to the texts chosen for the Helsinki Corpus. Page numbers are thus only given when the passages are not taken from the beginning of the respective texts. 17 A selection of text as in the BROWN, LOB, F‑LOB and FROWN corpora, which choose 2,000 words from the beginnings of texts, would not have been suitable for the present study because certain kinds of adverbial connectors, i. e. listing or, in particular summative ones, only occur at certain passages of the text as a whole (in the middle or the end, respectively).

The text corpus   13

problems of the lack of documents for the periods OE1, OE4, ME1 and ME2, which are a hazard to all corpus studies,18 the present study is confronted with a more specific problem concerning conjunctions and adverbial connectors. First of all, only prose texts could be chosen because poetry employs very different means of text structuring (see, e. g., Brinton 1996: 68–79 on text structuring devices such as ME gan + infinitive in Chaucer’s Troilus & Criseyde). More importantly, Present Day English corpus findings also demonstrate that adverbial connectors are extremely rare in narrative texts belonging to the Longman Grammar’s registers fiction and news (Biber et al. 1999: 882; see also Chapter 3.5). In narrative texts, authors often choose to leave the relationships unmarked, since relations such as cause and result may be inferred from a chronological sequence. Authors of narrative fiction even avoid being too explicit about the relations because they want to keep readers in suspense. Thus the only adverbials which are frequently used in narrative prose are temporal circumstance adverbials (Biber et al. 1999: 822), which – in particular in the categories “History”, “Biography”, “Travelogue” or “Diaries” of the Hel‑ sinki Corpus – are used as chronological signposts. In the following passage from the Diaries of Samuel Pepys, for example, the majority of sentences start with a temporal adverbial: (2)

15–16 DECEMBER 1666. Lords day. Lay long, talking with my wife in bed. Then up with great content, and to my chamber to set right a picture or two – Lovett having sent me yesterday Santa Clara’s head varnished, which is very fine. And now my closet is so full stored and so fine, as I would never desire to have it better. Dined without any strangers with me – which I do not like on Sundays. Then after dinner by water to Westminster to see Mrs. Martin, whom I found up in her chamber and ready to go abroad. I sat there with her and her husband and others a pretty while; and then away to White‑hall and there walked up and down to the Queen’s side, and there saw my dear Lady Castlemayne, who continues admirable methinks – and I do not hear but that the King is the same to her still as ever. Anon to chapel, by the King’s closet,

18 There are almost no texts surviving from the earliest period of English (OE1) and the late thirteenth century (ME2). The texts collected in OE4 are mainly copies of texts by authors whom we also find in OE3 (Ælfric in COTEMPO, COAELHOM, COAELET3, COAEPREF, COAEPREG for OE3 and COAELET4 for OE4; Wulfstan with COWULF3 for OE3 and COWULF4 and COINSPOL for OE4). The first Middle English period (ME1) has some texts which are copies of Old English manuscripts and should definitely be re‑allocated to OE3 or OE4 (in particular CMBODLEY, which is typical of Late Old English).

14   The framework and heard a very good Anthemne. Then with Lord Brouncker to Sir W. Coventry’s chamber, and there we sat with him and talked. He is weary of anything to do, he says, in the Navy (CEDIAR3A, p. VII, 409). There is, to take another example, not a single adverbial connector in the 6070 word passage taken from John Evelyn’s Diary (CEDIAR3B). With respect to the text type categorization of the Helsinki Corpus, we are furthermore confronted with the problem that many texts which are labelled as different categories are in fact predominantly narrative in style. See, for example, the following passage from the Book of Margery Kempe (which is categorized as a “Religious Treatise” in the Helsinki Corpus), in which every single sentence starts with a temporal adverbial (þan ‘then’, whan ‘when’, so long ‘for such a long time’, sithyn ‘then’):19 (3)

þan sche, hauyng trust of hys a‑mendyng & compassyon of hys infirmyte, wyth scharp wordys of correpcyon promysyd to fulfillyn hys entent gyf God wolde grawntyn it. Whan sche cam to hir meditacyon, not forgetyng þe frute of hir wombe, sche askyd forgeuenes of hys synne & relesyng of þe sekenes þat owr Lord had gouyn hym gyf it wer hys plesawns & profite to hys sowle. So longe sche preyid þat he was clene delyueryd of þe sekenes and leuyd many gerys aftyr & had a wife & a childe, blissyd mote God ben, for he weddyd hys wife in Pruce in Dewchelonde. Whan tydyngys cam to hys modyr fro ouyr þe see þat hir sone had weddyd, sche was ryth glad & thankyd God wyth al hir hert, supposyng & trustyng he xulde leuyn clene & chast as þe lawe of matrimony askith. Sithyn, whan God wolde, hys wife had a childe, a fayr mayde‑child. þan he sent tydingys to hys modyr in‑to Inglond how gracyowsly God had visityd hym & hys wife … (CMKEMPE p. I, 223).

Similarly, long narrative passages – in which primarily the coordinator and and temporal adverbials are employed – are found in many of the private and official letters across all periods; for example, in the following passage from a letter by Margaret Paston to her husband John: (4)

And Jamys Gloys come with his hatte on his hede betwen bothe his men, as he was wont of custome to do. And whanne Gloys was a‑yenst Wy-

19 Similarly, in her Revelations of Divine Love – also categorized as a “Religious Treatise” – Julian of Norwich virtually only uses the coordinators and, for and but (CMJULNOR). Only the following adverbial connectors are employed (token counts): therefore (4), wherefore (1), furthermore (1), nevertheless (1).

The text corpus   15

mondham he seid þus, ‘Couere thy heed!’ and Gloys seid ageyn, ‘So I shall for the.’ And whanne Gloys was forther passed by þe space of iij or iiij strede, Wymondham drew owt his dagger and seid, ‘Shalt þow so, knave?’ And þerwith Gloys turned hym and drewe owt his dagger … And þanne Haweys ran into Wymondhams place and feched a spere and a swerd, and toke his maister his swerd. And with þe noise of þis a‑saut and affray my modir and I come owt of þe chirche from þe sakeryng; and I bad Gloys go in to my moderis place ageyn, and so he dede. And thanne Wymondham called my moder and me strong hores, and seid þe Pastons and alle her kyn were … (1448, 19 MAY; CMPRIV, p. 223). These test analyses for the respective periods show that the paucity of adverbial connectors found for the Present Day English category fiction is also attested for the earlier periods of English, mainly in texts labelled “History”, “Bio­ graphy, Lives”, “Fiction”, “Travelogue”, or “Diaries”. Apart from a few sample texts (see above, and Appendix C.1), these were not analysed because they only use very few, if any, tokens of adverbial connectors. This means that a large number of the texts which have survived from the Old and Middle English period (in particular passages from Bible translations, “Chronicles” and “Saints’ Lives”) had to be excluded. Furthermore, texts which obviously did not fit their respective categories were excluded from the quantitative parts of the present study. Generally, text type consistency and comparability is one of the most problematic issues in diachronic corpus linguistics. In his study of English participle and gerund constructions from 1100 to 1700, which had set out to correlate the respective occurrences with text‑type patterns, Kohnen shows – in a detailed analysis of the texts and their contexts – that only a certain number of texts of the Helsinki Corpus can be employed for comparable analyses of patterns according to text types (see Kohnen 2004: 81–128; as a result, his corpus consists of “Chronicles”, “Religious Treatises”, “Homilies”, “Laws/Documents”, “Narrative Prose” and “Private Letters”). The main text type chosen for the present analysis are thus argumentative texts with their focus of attention on what is commonly called “academic” or “scientific language” (Biber et al. 1999), i. e. homilies or religious, philosophical, educational, and literary treatises, some of which also allow more detailed quantitative analyses (for the principles of selection, see the introduction to Appendix C.2). This choice of texts causes another well-known problem for the historical analysis: there are not only few texts of this type extant from earlier periods of English, but most of them are translations from Latin or French. For the earliest periods of English, there are thus solely some “Pre­faces” (COPREFCP, COAEPREF, COAEPREG, CMPURVEY, CMCAXPRO) and

16   The framework some of the “Letters” (COAELET3, COAELET4, CMPRIV, CMOFFIC3, CMOFFIC4) which are relatively free from foreign influence.20 1.8.  Presentation of corpus findings The findings of the study are summarized in Appendices A.1 (“Adverbial Connectors: Items”) and A.2 (“Alphabetical Index of Adverbial Connectors”) and Appendix B (to be found on the accompanying CD-ROM). Appendix A1 provides a synopsis of all the linguistic items which have been used as adverbial connectors in the history of English in alphabetical order. The data collected in this appendix have served as the basis for analysing the developments in the inventories of adverbial connectors from Old English to Modern English in Chapters 6 to 12. Appendix B more specifically lists the occurrences of adverbial connectors in the corpus texts. For each semantic category – B.1 addition, B.2 summation, B.3 result/cause, B.4 contrast/concession, B.5 transition – the relevant adverbial connectors are listed alphabetically. The respective entries for the individual connectors then give the precise text and the sub‑periods in which the items occur in adverbial connector function. An idea about quantities, i. e. of the number of the occurrences in the respective text, is signalled by different fonts: underlining indicates that the item occurs more than five times per 5,000 words in the text, bold type shows that it occurs more than 20 times. This presentation of data records the first occurrence of the respective item in a corpus text, but, more importantly, allows insights into the use and frequency of the individual adverbial connectors over time as well as in specific sub‑periods. It furthermore highlights genre‑specific uses of individual connectors. To allow a more detailed analysis of their use, the entries also list information on attested collocations with coordinators or other adverbial connectors (e. g. and furthermore or but yet nevertheless) and on the position of the adverbial connector in the sentence; for each item, initial, post‑first, medial and final position are recorded (for the relevance of these positions, see Chapters 5.5 [post‑first‑position], 10.5 [final position] and 13 [medial position and collocations]. Since the study will obviously not discuss the developments of all of the individual connectors in detail (most of which would merit a monograph), some illustrative examples will show how this appendix is used in the present study and how it could be exploited for further research. 20 All corpus findings have been cross‑checked and supplemented by analyses of the machine‑readable corpora, the Dictionary of Old English Corpus (DOEC) and the Corpus of Middle English Prose and Verse (CME).

Presentation of corpus findings   17

In the section for additive (reinforcing) adverbial connectors (Appendix B.1.2), we see in the entry for overmore ‘furthermore’ that it was only used in texts of the two later sub‑periods of Middle English. It is only found clause‑initially; in one text, it is attested in collocation with the conjunction and, i. e. and overmore. overmore Additive (Reinforcing) ME3 CMDOCU3, CMAELR3 ME4 CMOFFIC4 Collocations: and overmore ME3  CMAELR3

Eft and eftsona ‘also, furthermore’, on the other hand, were only used in the Old and Early Middle English period. eft Additive (Reinforcing) ‘also; furthermore’ OE2 COLAECE OE3 COQUADRU, COAELHOM, COBENRUL, COAE LET3, COAEPREG OE4 COLAW4, COLACNU, COAELET4 ME1 CMPERIDI, CMBODLEY ME2 CMAYENBI Collocations:

and eft OE2  COCURA ME1  CMPERIDI, CMVESHOM, CMTRINIT, CMLAMBET, CMBODLEY, CMVICES1



ac eft OE2  COCURA

eftsona

Additive (Equative – Reinforcing) ‘moreover, likewise’ ME1 CMPERIDI, CMTRINIT

Collocations:

and eftsona ME1  CMPERIDI

Eft and eftsona are also found in collocations with the coordinator and, but also with ac ‘but’. In two texts – COLAECE and CMPERIDI – they occur more than twenty times per 5,000 words, as indicated by the bold font. Both these texts are medical handbooks (the Laeceboc and a Middle English handbook entitled Peri Didaxeon) in which sentence‑initial eft and eftsona are used to list alternative medications for specific illnesses: (5)

Eft sona wið gif þeo ylca adle cilde eggelic. on geogeþe; Nim garluces … And eft sona gif þa wunda toðindað. Nim fyrs & cnuca hine. & lege uppa þat geswollene … (CMPERIDI, p. 89).

18   The framework

‘Furthermore against if the same illness harms a child in its youth. Take garlic … And furthermore, if the wounds swell up. Take bramble and pound it and put it on the swelling …’.

This also shows that a large‑scale, strictly quantitative comparison is not sensible for the subject in question. The extract from the Laeceboc, for example, which consists of 10,420 words, only uses six different types of connective items: the conjunctions and (2 tokens) and ac (4 tokens), the adverbial connectors eac ‘furthermore’ (5 tokens), ðonne ‘therefore, then’ and ðeah ‘nevertheless’ (1 token each) compared to 6 tokens of and eft and 46 tokens of eft. The numbers are similar for the Peri Didaxeon (22 tokens of eft sona vs. 13 tokens of and and 2 of forðan ‘therefore’). A similar restriction to genres and uneven distribution is, for example, attested for the loan item ‘furthermore’: item

Additive (Reinforcing) ME3 CMEQUATO ME4 CMDOCU4, CMREYNES, CMPRIV EModE1 CELAW1

This reinforcing adverbial connector is – as its entry illustrates – only attested in “Handbooks”, “Laws”, “Statutes and Documents” (mainly wills) in the Late Middle English and Early Modern English sub‑periods, often in lists. See, for instance, the following passage taken from “An Act against deceyptfull making of wollen cloth”: (6)

Fyrst That the Wolle whiche shalbe delyvered for or by the Clothier to any persone or persones for brekyng kembyng cardyng or spyernyng of the same the delyvere therof shalbe … And that the breker or kember to delyver agayn to the seid Clothier … Item that the Wever whiche shall have the wevyng of eny wollen yerne to be webbed into cloth shall weve werk … Item that no maner persone bye eny coloured Wolle or coloured wollen Yerne of eny Carder Spynner or Wever … Item that the Walker and Fuller shall truely walke fulle thikke … Item that the Clothier nor other persone whatsoever he be after the fest of Midsomer next cuerm­ yng shall not put eny cloth to sale … Item that the byer of Wollen clothes denysen or alyen after the byeng therof shall not drawe nor cause to be drawen in lenght … (CELAW1, p. III, 28).

Since texts like these would distort a quantitative investigation based on the whole of the texts analyzed for this study, only a sub‑set of comparative texts

Presentation of corpus findings   19

was used for the more detailed analysis concerning collocations and word order frequencies in Chapter 13 (for these texts, see Appendix C.2). Apart from information on genre patterns, the survey in Appendix B also allows findings on the popularity of individual adverbial connectors in specific sub‑periods. Reinforcing also, which has been attested since the Old English period (OE eall‑swa ‘all so’) was not found – as an adverbial connector – in texts of the Late Modern English period: also Additive (Reinforcing) OE3 COBYRHTF, COTEMPO OE4 COAELET4 ME1 CMTRINIT, CMLAMBET, CMANCR, CMVICES1 ME2 CMAYENBI ME3 CMDOCU3, CMASTRO, CMHORSES, CMPHLEBO, CMCTPROS, CMWYCSER, CMAELR, CMPURVEY, CMOFFIC3 ME4 CMLAW, CMDOCU4, CMREYNES, CMCHAULI, CMROYAL, CMINNOCE, CMAELR4, CMKEMPE, CMHILTON, CMROLLTR, CMPRIV, CMOFFIC4 EModE1 CELAW1, CESCIE1A, CESCIE1B, CEEDUC1A, CEEDUC1B, CEBOETH1, CEPRIV1, CETRI1 EModE2 CELAW2, CEHAND2B, CESCIE2A, CESCIE2B, CEEDUC2A, CETRI2A EModE3 CESCIE3B, CEBOETH3, CEPRIV3, CEAUTO3 PDE FROWN‑D

Interestingly, the OED states that also generally, i. e. also in other functions, was “[n]ot common in 16th c.; Shakes., according to Schmidt, has it only 22 times” (OED, s. v. also, Etymology; for a more detailed analysis of this restriction and its causes, see below, Chapter 11). A further example illustrating the design and, more importantly, the results from the data analysis provided in Appendix B, is given by the entry for the concessive/contrastive adverbial connector though, which illustrates the discontinuous path of this connective (Appendix B.4). though Contrastive/Concessive (OE þeah/ ME þeh) OE2 COLAECE, COBOETH OE3 COQUADRU OE4 CODICTS, COINSPOL ME1 CMVESHOM, CMBODLEY, CMLAMBET, CMANCR, CMHALI ME3 CMASTRO EModE1 CESCIE3A

20   The framework

þeah/þeh, second position OE2 COBOETH, COCURA OE3 COTEMPO, COWULF3, COBENRUL, COAELET3 OE4 COLAW4, CODICTS, COINSPOL, COPREFSO ME1 CMSAWLES, CMHALI



though, second position in questions EModE1 CEHAND2A, CEBOETH2



though, medial position PDE FROWN‑D

This entry shows that though – as an adverbial connector – was used in a wide variety of texts in the Old and Early Middle English periods (OE, ME1). It is attested as an adverb in later Middle and Early Modern English, but only in a limited number of texts and in very specific contexts (questions). In the corpus texts, a use as an adverbial connector is not recorded in any of the texts from Late Modern English. By contrast, al(though) has been common as a subordinator in all periods of English (see OED, s. v. though, II). This shows that there is no unidirectional path of though from adverb to conjunction (as grammaticalization theories would suggest), but rather a zigzag path from ambiguous adverb/conjunction to conjunction to (a) initially placed conjunction and (b) medially or finally placed adverb (see below, Chapter 10.5). This path is neither evident from the entry in the OED nor from other meta‑linguistic texts, but only emerges as a result of the analysis of the corpus material. 1.9.  Outline of the present study Chapters 2 to 4 prepare the ground for the ensuing analyses by summarizing the main findings of previous and contemporary scholarship on clause linkage and the classification of connectives (Chapter 2), and on the problems of the classification and hierarchy of the word class “adverb” (Chapter 3). In this context, Chapter 3.5 summarizes those corpus findings for contemporary English adverbial connectors in the Longman Grammar which are relevant for their history. In an excursus on earlier meta‑linguistic thought, Chapter 4 gives an account of the treatment of conjunctions and adverbial connectors and their status until the times of Locke and the Scottish Rhetoricians, who, as is mentioned above, are the first to explicitly accept adverbial connectors as a separate class. The main part of the study (Chapters 5 to 13) then provides an in‑depth treatment of adverbial connectors in the respective historical periods of English. For an understanding of the distinctive developments English has undergone in its history, it is important to lay a foundation by describing the system of

Outline of the present study   21

clausal connection in Old English (Chapter 5). Chapter 6 – the heart of the present study – summarizes the general patterns in the diachrony of adverbial connectors and contrasts their long‑term developments with the changes in other connectives, i. e. coordinators and subordinators. The divergent histories of adverbial connectors and subordinators are then examined in more detail in Chapters 7 to 10. Chapter 7 analyses the morphology of adverbial connectors (in particular the innovations coined after the loss of pronominal connectors). Chapter 8 then summarizes the findings on innovations from the cognitive source domains time and place/space which have been crucial in the coinage of English adverbial connectors from Middle English onwards. In a more detailed investigation, Chapter 8.4 establishes the source domain truth/fact for adverbial connectors signalling the relation transition. Chapters 9 to 12 then concentrate on the individual semantic relations: these investigations begin with an analysis of developments in the ccc‑relations (cause, contrast/ concession) because they – in contrast to the more superficial changes in the relations addition and transition (Chapters 11–12) – also mirror structural and typological changes. Chapter 13 then resumes the topic of the influences of different schools of rhetoric, in particular the impact of the ideas of John Locke and the Scottish Rhetoricians. In all of the chapters, overarching principles of long‑term developments in language change will be addressed along the following key questions: Are there recurrent patterns of change which affect the individual connectors? Are there different patterns for the different semantic relations? Do the changes in the system of adverbial connectors correspond to more general cross‑linguistic patterns in language change, such as principles of iconicity or grammaticalization clines? Do the changes, more particularly, correspond to general tendencies in the typological history of English and the expansion of the English lexicon? And finally, to what extent are it stylistic or rhetorical preferences which have directed the paths of change rather than more deeply‑rooted typological properties of English?

2.  Clausal connection 2.1.  Clausal connection Clausal connection – which can basically be defined as “a relation of dependency or sociation between clauses” (Lehmann 1988: 182) – is one of the key building blocks of language in use. It is thus a subject which relates not only to syntactic, but also to a wide range of semantic, pragmatic and, most importantly, textlinguistic phenomena. Traditionally, a first distinction is drawn between asyndetic (or unlinked) and syndetic (or linked) clausal connection.1 Asyndetic connection is not overtly marked (the standard quoted example is Caesar’s veni, vidi, vici ‘I came, saw, conquered’). It is much rarer than syndetic linkage, usually stylistically marked and used, for instance, for dramatic intensification (Quirk et al. 1985: 918; category I in Raible 1992). Thus, cross‑linguistic accounts base their classification of clause linkage on the absence or presence of an explicit linking device; see, for instance, Lehmann’s criterion “isolation vs. linkage” (Lehmann 1988: 210–213; subsection “explicitness of linking”). In Raible’s scheme of the dimension of linkage as realisation from “aggregation to integration” (cf. the title of Raible 1992: Junk‑ tion. Eine Dimension der Sprache und ihre Realisierungsformen zwischen Aggregation und Integration), explicit linkers are used in categories II to IV (diagram on the inserted foldout): category II “Junktion durch Wiederaufnahme (eines Teils) des vorhergehenden Satzes” (‘Linkage by resumption of (a part of) the preceding sentence’), category III “Explizit verknüpfte Hauptsätze” (‘Explicitly linked main clauses’) and category IV “Verknüpfung durch subordinierende Konjunktionen” (‘Linkage by subordinating conjunctions’). Syndetic connection is thus characterized by the use of overt signals, such as clause‑integrated linkage (That’s why; The result was …) or prepositional phrases (because of …). Most frequently, however, connectives such as conjunctions or adverbial connectors are used.2 1 The following summary rests on Quirk et al. 1984: chapters 13–15, Altenberg 1984, Biber et al. 1999: chapters 10 and 11. For overall conceptions of clausal linkage comprising all kinds of non‑explicit linking options, such as non‑finite constructions, the ablative absolute, etc., see Lehmann 1988 and Raible 1992. 2 For corpus findings for Present Day English, see Altenberg 1984 and 1986 (for causal and contrastive/concessive connection) and Biber 1999: chapter 10.4 (for all kinds of linking adverbials). These studies are summarized and discussed below in Chapters 3.5, 9.2.2 and 10.5.

Connectors – connects: definitions   23

2.2.  Connectors – connects: definitions The most recent, comprehensive study of connectors in Present Day German by a project group at the Institut für deutsche Sprache at Mannheim, the 800‑page Handbuch der deutschen Konnektoren, defines “connectors” by the following five properties (M = Merkmal ‘property’; Pasch et al. 2003: 1–6 and a refined version on 331–334): M1

x ist nicht flektierbar (‘x is not inflectable’)

M2 x vergibt keine Kasusmerkmale an seine syntaktische Umgebung (‘x does not govern case’) M3

die Bedeutung von x ist eine zweistellige Relation (‘x signals a two‑place semantic relation’)

M4

die Argumente der Bedeutung von x sind propositionale Strukturen (‘the arguments of the meaning of x are propositional structures’)

M5

die Relate der Bedeutung von x müssen Satzstrukturen sein können (‘the elements linked by x may be expressed by sentential structures’).

Basically, these characteristics of connectors are generally agreed on in the literature (see, e. g., also Raible 1992: diagram on inserted foldout). Instead of classifications focussing on Present Day English (such as Quirk et al. 1985: chapter 13 or Biber et al. 1999: chapters 10 and 11), these specific criteria were chosen as the basis of the present investigation because the diachronic account requires a system which is transferable to earlier periods of the English language. Since Old English is typologically more similar to Modern German than to Modern English, the criteria of the Handbuch der deutschen Konnek‑ toren provide a solid basis for comparison. In the following section, I will briefly introduce those points of differentiation from other parts of speech such as prepositions, disjuncts or modal particles which are essential for an understanding of the diachrony of English connectors. M1 is a morphological criterion which groups connectors together with other indeclinable items such as prepositions.3 Connectors may, of course, have their origin in case forms, such as in a genitival ‑es, which is attested in PDE else (< OE elles) or – by way of analogy – in besides (OE be sidan) and hence (OE heonan); 3 In the Greek and Latin tradition, all indeclinable items were called “particles”; see the quote from Locke above, p. 1, and below, Chapter 3.5.

24   Clausal connection these forms, however, were petrified and are now indeclinable in their uses as connectors. In earlier stages of English, we regularly also find “pronominal connectors”, similar to PDG deshalb or demnach ‘therefore’. In Old English, these consist of, for instance, a preposition and the case form (mainly dative or instrumental) of a demonstrative, such as OE forþæm (< preposition for + dative þæm) or OE forþon and forþy (< preposition for + instrumental þon or þy). Analyzing these forms, it is important to differentiate full prepositional phrases, such as PDE because of this, from genuine connectors, such as PDE because or therefore. For this differentiation, M2 is formulated as the syntactic criterion distinguishing prepositions – which govern a particular case – from connectors, which do not govern case. This criterion is, however, only rarely applicable to English after the Old English period, when English had given up case (and gender) distinctions in articles and demonstratives. M3 is a semantic property which differentiates connectors from other sentence adverbials, such as the epistemic probably, perhaps or stance adverbials such as frankly. These also have a scope extending over the whole sentence, but they modify this sentence only and – unlike connectors – do not relate two clauses or even larger portions of text (for details on different classes of adverbs, see below, Chapter 3). They do thus not qualify as organizers of textual structure. Connectors have to signal a two‑place relation. M4 and M5 are properties augmenting M3. They contrast connectors, again, from prepositions and also from coordinators on the phrase level. Prepositions may also be seen to relate two items, such as tree and house in a sentence like PDE the tree in front of the house, but they do not relate propositions or clauses. M5 in particular distinguishes prototypical connectors from coordinators, such as and and or. Coordinators are not always connectors in the sense of the present study because they may also combine elements on the phrase level, as in Peter and / or Mary (were sitting in front of the house). There are, however, also classifications, which propose a gradient of connectors, ranging from coordinators to subordinators. The Comprehensive Grammar (Quirk et al. 1985: 923‑924) for example, explicitly states that “coordinators can link clause constituents”; such a gradient, however, is not suitable for the present account, which only deals with connection on the clausal or textual level.4 Whenever the present study wants to stress this fact of a two‑place relation on the level above the phrase, the clauses or parts of text connected will be referred to as 4 According to M5 of the Handbuch, the “sentences” connected have to comprise a finite verb or have at least to be able to comprise a finite verb. The restriction to clauses containing a finite verb is only applicable to earlier periods of English, but not to Present Day English, which allows non‑finite clauses.

Connectors   25

“connect(s)”, a term (PDG Konnekt) coined for the Handbuch der deutschen Konnektoren (Pasch et al. 2003). 2.3. Connectors: coordinating conjunctions, subordinating conjunctions, adverbial connectors By these five criteria we can establish three different kinds of connectors, a differentiation which is also basically agreed on in the literature: “coordinating conjunctions” (working on the clausal/textual level, not on the phrasal level; see M4 and M5), “subordinating conjunctions” (“subordinators”), and “adverbial connectors” (Quirk et al. 1985 “conjuncts”, Biber et al. 1999 “linking adverbials”). The present study focuses on adverbial connectors. As the system of adverbial connection has changed so dramatically since the Old English period, it is, however, necessary to summarize the main syntactic and text‑linguistic properties of the other types here in some detail. Virtually all of today’s adverbial connectors are new coinages or have undergone some morphological and structural change. One of the natural paths was the transfer from one of the other classes of connectors: compare, for example, the use of the Present Day English subordinators albeit and howbeit as adverbial connectors in the Early Modern English period. Since Old English is typologically much closer to Modern German than to Modern English, the following section will exemplify the relevant issues not only by material taken from Present Day English (contrastive/concessive connectors; Table 2.1) but also by Present Day German examples (causal connectors; Table 2.2).55 Table 2.1.:  Contrastive/concessive connectors in Present Day English Parataxis Coordinating ­Conjunction: Adverbial Connector: Adverbial Connector:

Hypotaxis Subordinating ­Conjunction:

(7) He tried hard, but he failed. (8) He tried hard, and yet he failed. (9a) He tried hard. However, he failed. (9b) He tried hard. He, however, failed. (9c) He tried hard; he failed, however.

[post‑position] [post‑position; and + AdvCo] [initial] [medial] [final]

(10a) (Al)though he tried hard, he failed. (10b) He failed, though he tried hard.

[pre‑posed] or [post‑posed]

5 The contrastive/concessive and causal semantic relations are chosen because they will be in the focus of the detailed diachronic studies in Chapters 10 and 9 below.

26   Clausal connection Table 2.2.:  Causal connectors in Present Day German Parataxis Adverbial Connector: (11a) Sie wird gewinnen. Sie ist nämlich stärker. (11b) Sie ist stärker. Deswegen wird sie gewinnen. Hypotaxis Subordinating ­Conjunction:

[V2; post‑posed] [V2; post‑posed]

(12a) Weil sie stärker ist, wird sie gewinnen. [V‑final; pre‑posed] or (12b) Sie wird gewinnen, weil sie stärker ist. [post‑posed]

Correlatives (13) Sie wird deswegen gewinnen, weil sie stärker ist.

[V2 + V‑final]

The criteria traditionally employed for distinguishing the different types of connectors are [1] the position of the connector, [2] the sequence and position of the respective connects, [3] the possibility of collocations of connectors, and [4] in languages such as German, constituent order. I will here first of all summarize these traditional views. It has to be stressed, however, that the constructions are by no means functionally equivalent, but are different with respect to their coding of the information structure (for details, see below, Chapter 2.4).6 [1]  While coordinating and subordinating conjunctions are only found clause‑initially (see English but, and and German weil in examples 7, 8, 12), adverbial connectors (English however, yet; German nämlich, deswegen) are, like most kinds of adverbs, more free in their position in the sentence (see examples 9 for however).7 Pasch et al. (2003) capture this difference by the terms 6 It is the major drawback of the novel classification and terminology coined for the Handbuch der deutschen Konnektoren that connectors are distinguished only by their position in the clause or connect, i. e. by topological criteria only. The Hand‑ buch thus differentiates “konnektintegrierbare vs. nicht‑konnektintegrierbare Konnektoren” and further “Subjunktoren – Postponierer – Verbzweitsatz‑Einbetter – Konjunktoren”. 7 Pasch et al. 2003 distinguish the following positions for adverbial connectors: “nicht positionsbeschränkte Adverbkonnektoren” such as allerdings, dagegen, frei‑

Connectors   27

“nicht‑konnektintegrierbare” (initial position) vs. “konnektintegrierbare” connectors.8 [2]  Parataxis usually requires a fixed order of connects. Accordingly, clauses linked by coordinating conjunctions and adverbial connectors only allow one sequence of connects. Thus there are, for example, two different German adverbial connectors for the causal/resultive relation, namely nämlich for the sequence result → cause, and deswegen for the sequence cause → result, which accordingly require reverse sequences of the respective connects. The interesting fact about Present Day English clausal connection is that it lacks an adverbial connector coding the relation result → cause (cf. Latin nam, enim; PDG denn, nämlich); it only has adverbial connectors expressing the sequence cause → result, such as therefore, hence, thus (cf. Latin igitur, PDG deshalb; for a full analysis, see below Chapter 9). Subordinate clauses, on the other hand, may be placed before or after the superordinate clause (see English although or German weil in examples 10 and 12). In contrast to parataxis, where the units are constituents at the same level of constituent structure, they form a hierarchy in subordination (or hypotaxis), the subordinate unit being a constituent of the superordinate unit (Quirk et al. 1985: 13.2). Since the present study concentrates on adverbial connectors, it does not consider all kinds of subordinate structures and excludes cases of embedding (e. g. obligatory nominal sentences functioning as subjects, objects, subject/object complements) or relative clauses, which post‑modify a noun phrase only. Neither embedding nor relative clauses connect propositions (thus failing properties M4 und M5).9

lich, also, “nicht nacherstfähige Adverbkonnektoren” such as demnach, daraufhin, trotzdem, deshalb, “stark positionsbeschränkte Adverbkonnektoren” such as aber, auch, nur and “nicht vorfeldfähige Adverbkonnektoren” such as nämlich. For the relevance of the distinction of the post‑first‑position from other medial positions, see below, Chapter 13. 8 These compounds are so typically German that they are virtually untranslatable into other languages. 9 Obviously, there are some polyfunctional elements, such as the Old English general subordinator þe, which may function as a relative particle but can also be part of a complex connector (cf. OE for þæm þe ‘because; therefore’). Similarly, the general subordinator in Middle English, that, can introduce nominal clauses, but may also be part of a complex relativizer (cf. ME which that) or of a lexicalized adverbial connector (cf. Middle English causal for that or additive over that).

28   Clausal connection [3]  Coordinating conjunctions may – in contrast to the other types – collocate with subordinators and also with adverbial connectors (see English and yet in example 8).10 [4]  In contrast to Present Day English, which employs a fixed subject – verb word order in main and subordinate sentences, Present Day German distinguishes these sentence types by employing verb‑second word order for main clauses and verb‑final for subordinate clauses (see the underlined verbs in examples 12 and 13; for a discussion of word order patterns in main and subordinate clauses in Old English, see below, Chapters 5 and 11). The morphological make‑up of German connectors also allows more easi­ ly for so‑called “correlative constructions”, which mark the relation of the ­sentences with an adverbial connector in the first connect and a subordinating conjunction in the second one (cf. German deswegen … weil in example 13 and the much rarer English because … therefore). 2.4.  Connectors and information processing The traditional accounts of connectors summarized above seem to imply that these constructions are functionally equivalent. This is mainly due to the fact that most grammars concentrate on the make‑up of a single sentence (which may also be a compound sentence). Complex sentences, however, and larger chunks of discourse are, if at all, only treated marginally. Starting with Thompson’s analysis of initial versus final purpose clauses (Thompson 1985), however, there has been an increasing and continuing interest in the discourse factors determining the position of subordinate clauses (see the literature from Haiman and Thompson 1988 to Diessel 2005). Thompson found that there is no free choice between these two positions, so that initial and final purpose clauses should not be regarded as a single construction potentially occupying two different positions. More appropriately, they should be viewed as two quite different constructions, which share the same morphology, but behave radically different in the organization of discourse. Initial clauses state a problem within the context of expectations raised by the preceding discourse, to which the following material (often many clauses) provides a solution.

10 For the popularity of these collocations in different phases of the history of English, see below, Chapter 13.

Connectors and information processing   29

(14) To cool, place the loaf on a wire rack. Final purpose clauses, on the other hand, play a much more local role of stating the purpose for the action named in the immediately preceding clause. (15) Place the loaf on a wire rack to cool. As in the ordering of other constituents, the ordering of the linguistic elements is here seen to be primarily determined by information structure: given information tends to precede new information because new information needs to be grounded in information that is already in the hearer’s knowledge store. This topic‑forming capacity of pre‑posed adverbial clauses is now commonly agreed on (see, for example, Givón 2001: chapter 18 and Meurman‑Solin 2004). In a more recent study, Diessel argues that both processing and information structure are relevant to constituent order. He shows that the positional patterns of conditional, temporal and causal clauses in Present Day English are motivated by competing functional and cognitive forces, namely economy and explicitness (Diessel 2005: 449). In an analysis based on Hawkins’ processing theory, he shows that, with regard to parsing or utterance planning, complex sentences are easier to process, and thus highly preferred, if the adverbial clause follows the main clause. The discourse pragmatic motivations sketched above, however, which favor initial occurrence, may override the processing motivation for final occurrence: the clauses are positioned sentence‑initially if they provide a thematic ground or orientation for subsequent clauses. Since the discourse pragmatic function of adverbial clauses interacts with their meaning, different semantic types of adverbial clauses differ in their distribution: Conditional clauses tend to occur sentence‑initially, because they establish a specific framework for interpreting the following connect. Temporal clauses tend to precede the main clause if they denote a situation prior to the one in the main clause so that initial occurrence results in an iconic order. Causal clauses mostly follow the main clause. According to Diessel, only scientific articles exhibit a substantial number of pre‑posed causal clauses because in this type of discourse causal clauses are often used to provide a common ground for a subsequent conclusion (Diessel 2005: 465).11 Although the present study is not primarily concerned with hypotaxis, these issues of processing and information structure are also crucial for an 11 For similar corpus findings, see Altenberg (1984 – causal clauses) and Quirk et al. (1985: 1107 – causal clauses).

30   Clausal connection understanding of adverbial connectors, and in particular their history.12 The functional relation between clause combining by conjunction vs. adverbial connectors is not as widely researched as the criteria for the different positions of subordinate clauses. Generally, Quirk et al. (1985: 13.3) see a major difference between coordination and subordination of clauses: the information in a subordinate clause is often placed in the background with respect to the superordinate clause. In this view, the syntactic inequality of subordination tends to bring with it a semantic inequality which is realized by syntactic hierarchization. Thus the information given in these clauses is usually not pursued in the following discourse. As examples they give (16) He has quarrelled with the chairman and resigned. (17) Since he has quarrelled with the chairman, he resigned. (18) He resigned because he has quarrelled with the chairman. In this cause – result relationship, the hearer is assumed to have heard about the quarrel already in (17) (for this topic‑forming capacity, see above). The information given in the subordinate clauses, however, will not be pursued in the following discourse, irrelevant of whether it is pre‑posed (17) or post‑posed (18). An analysis by Wegener (2000) for German, which – like Old English – differentiates verb‑second and verb‑final word order, yields similar results, which are supported by syntactic and prosodic criteria. In German and English, causal and concessive relations may be coded by either coordination or subordination. While concessive subordinators mainly code the concessive relation proper (“counter‑expectancy”), adverbial connectors more often code contrast (adversativity without any “counter‑expectancy”; for details, see below, Chapter 10). For conditionals, however, we never find parataxis – neither in Present Day German nor in any periods of English. Table 2.3.:  Semantic relations: parataxis and hypotaxis Semantic Relation Parataxis

Hypotaxis

cause

PDG da, weil (+ Verb‑final), zumal da PDE since, as, because

PDG denn, (weil + V2) PDE therefore

12 In pre‑posed causal clauses, for instance, polyfunctional forþæm, which may also function as an adverbial connector, is first replaced by unambiguous subordinators for as moche as or since that. For a detailed analysis, see below, Chapter 9.

Connectors and information processing   31 Semantic Relation Parataxis concession / ­contrast condition

Hypotaxis

PDG aber, zwar … aber PDG obwohl, obgleich, auch wenn PDE however, though, PDE (al)though … (yet) still – PDG wenn … (dann), falls, sofern PDE if … (then)

The importance of textlinguistic and pragmatic features is nicely illustrated by the functions of German weil with main clause word order (verb‑second), such as (19) Er ist nach Hause gefahren – weil er hatte Kopfweh. ‘He went home – for he had a headache’. instead of verb‑final, as in (20) Er ist nicht nach Hause gefahren, weil er Kopfweh hatte, (sondern weil er Besuch erwartet). ‘He didn’t go home because he had a headache (but because he is expecting visitors)’ The fairly new use of weil with main clause word order (mainly in spoken German) is certainly not as “bad” as some purist language users think (see also Keller 1993; or – for that matter – the Microsoft Word (version 2003) grammar check, which marks it as wrong in example 19). It is a common – though still informal – structural device in German, in which the lexical item weil replaces the adverbial connectors denn or nämlich. Weil is not used in a wrong way, but has acquired a second function, different to its use as a subordinator. Since this is the path which depicts the history of many of today’s English adverbial connectors, it is important to summarize the main points of difference and change here. The examples first of all show that (19) provides two information entities, whereas the hypotaxis in (20) gives only one focus; see the focus of the negator nicht and the possible questions for (19) Ist Peter noch da? ‘Is Peter still here?’ and Warum sagst du das? ‘Why do you say that?’13 in contrast to the single Warum ist Peter nach Hause gefahren? ‘Why did Peter go home?’ for (20).

13 Weil with verb‑second order does not give a reason for the proposition of the sentence but a reason for the speech‑act of the first connect. Weil therefore does not relate to the proposition of the first clause but to the speech act as a whole. For these

32   Clausal connection As illocutionary signals, German employs syntactic and prosodic means, i. e. different word orders (V2 vs. verb‑final) and, for parataxis, an intonation pattern which requires a pause, indicating that the second connect is not integrated in the first one. In subordination, the subordinate clause is syntactically and prosodically integrated: the whole complex sentence only has one intonation contour and one thematic (theme – rheme) structure (see Wegener 2000: 36 and Givón 2001: 327 for English). In parataxis, the second connect also has more assertive power, a power which may be highlighted by means of strongly deictic pronominal adverbial connectors such as PDG deshalb or dementspre‑ chend or transparent lexical adverbial connectors such as consequently. The greater illocutionary weight of the denn‑ or weil‑sentence – indicated syntactically by V2 and prosodically by pause and main stress also on the second connect – gives this second connect an illocutionary force of its own. This explains why we do not find both options of conditional relations – there the two connects cannot carry independent illocutionary forces. This summary shows that the path from subordinator to adverbial connector (as attested for however in Early Modern English and though in Present Day English) or the other way round (albeit in Early Modern English) is by no means a superficial transition. These changes are, conversely, induced by factors of information structure and may, on the other hand, also lead to changes in information structure. In this view, adverbial connectors are – in contrast to subordinators – very strong indicators of a great illocutionary weight of the second connect. Its proposition may then be pursued in the following discourse.

uses, see Keller (1993), Wegener (2000) and below, chapter 10.5, on PDG obwohl with verb-second order.

3.  The category “adverb” Adverb is not simply a lexical category, nor a functional one: rather, it emerges from the coalescence of both viewpoints (Ramat and Ricca 1998: 188).

3.1.  Adverbs and adverbials Although “adverbs” are among the oldest categories in grammatical thought (see below, Chapter 4), this category of words is most puzzling. The diversity of items collected under the heading “adverb” even seems to evoke the impression that if a linguistic item cannot be integrated into any of the other word classes, it is called an “adverb”: the category “adverb” is the “mixed bag” among the word classes. In the title of one of his articles, Guimier (1991) even asks whether it is at all possible to define the term “adverb” – “Peut‑on définir l’adverbe?”. First of all, it is necessary to distinguish between the formal category “adverb” as a word class and the functional category “adverbial”, which may be realized by a wide range of syntactic forms in Present Day English: adverbs, prepositional phrases, finite and non‑finite clauses, and noun phrases. In contrast to other functional categories such as “subject” or “direct object”, multiple adverbials can occur in a clause or sentence and they can be placed in a variety of positions, sentence‑initially, ‑medially or ‑finally. This syntactic variability is due to the fact that they perform a variety of functions and fulfil a variety of semantic roles, such as location, time, agency, attitude or, most important for the present study, a wide variety of linking functions. Most adverbials are, however, optional. An analysis of adverbs thus has to consider formal and – at the same time – functional aspects. This position at the interface between the lexicon and grammar has led to a long tradition of discussion whether the Present Day English adverbial suffix ‑ly is a lexical (derivational) or grammatical (inflectional) suffix (see Nevalainen 1997; van der Auwera 1998: 3). More important for the present study, there have also been various views on whether the coinage of new adverbs by univerbation involving fusion and coalescence is an instance of lexicalization or grammaticalization (see Brinton and Traugott 2005: 132–136, 152). Adverbs are – like conjunctions, prepositions and other smaller categories such as interjections – indeclinable and are therefore not to be distinguished

34   The category “adverb” from other indeclinable items on purely formal grounds.1 For their cross‑linguistic account of sentence adverbs in the languages of Europe, Ramat and Ricca (1998: 187) provide a very clear and workable first definition of adverbs which takes both the formal and the functional perspective into account. Since it also allows an integration of data from earlier periods of English, it will here be adopted as a basic definition: (i) Formally, adverbs are invariable and syntactically dispensable lexemes (which may have derivational status, e. g., Lat. simil‑i‑s ‘similar’, simil‑iter ‘similarly’, or even originate from inflectional status: Lat. merito(d) ‘rightly’). (ii) Functionally, adverbs are modifiers of predicates, other modifiers or higher syntactic units. In other words they add information to other linguistic elements which can stand on their own, semantically as well as syntactically.

This definition first of all highlights that the term “adverb” itself is misleading. The functional property of adverbs as being a modifier is shared by adjectives and adverbs. Yet, while adjectives are essentially noun modifiers, adverbs basically modify non‑nominal constituents. An “adverb” thus need not necessarily modify a “verb”, but can also modify an adjective (22), or another adverb (23): (21) He loudly at the door. (22) He wrote an extremely book. (23) He started smoking very . In addition to these functions on the phrase level, adverbs may also have a wider scope which extends over the whole sentence. These adverbs are accordingly labelled sentence or sentential adverbials. Sentence adverbials may be defined as “a class of syntactically dispensable lexemes which affect/modify in various ways the content of a sentence in which they occur” (Ramat and Ricca 1998: 189), as in (24) probably . (25) Fortunately, . 1 For recent accounts of the word class or category adverb and its problems, see Bellert (1977), Biber et al. (1999: chapter 10), Greenbaum (1969a), Guimier (1991), Quirk et al. (1985: 7.46–8.153; 438–653), and, in particular, the chapter in the ­EUROTYPE volume by Ramat and Ricca (1998).

Classification of adverbials   35

Their scope may then also extend further to the level of the paragraph or text, such as in (26) . Likewise, . 3.2.  Classification of adverbials To differentiate this “mixed bag” of adverbials, grammars usually suggest a tri‑partite classification based on syntactic and semantic criteria. Thus the Longman Grammar distinguishes circumstance adverbials from stance adverbials and linking adverbials (Biber et al. 1999: 762–892): – “Circumstance adverbials” add circumstantial information about the proposition in the clause, or add information about the circumstances of an activity or state described in the clause. – “Stance adverbials” express the speaker’s/writer’s stance towards the clause, and either give the speaker’s comment on the proposition of the clause or convey the speakers’ comments on what they are saying (the content of the message) or how they are saying it (the style). – “Linking adverbials” link the clause to some other unit of discourse and thus serve connective functions. These differences leading to a three‑partite classification are widely agreed on in contemporary grammars.2 One of the most influential monographs on English adverbs by Greenbaum (1969a) basically distinguishes between the same three major classes as the Longman Grammar (Biber et al.  1999), but calls them “adjuncts”, “disjuncts” and “conjuncts”. This classification is in essence3 2 The Cambridge Grammar of the English Language, however, now suggests another system which mainly affects the boundary between adverb and preposition (Huddle­ ston and Pullum 2002: 264). In this view, prepositions may also take other kinds of complements than noun phrases, so that their preposition category also includes some words without a complement. Thus outside in The basket is outside is classified as a preposition. This new categorization, however, conflicts with the properties of connectors chosen for the present study (see above Chapter 2.2). – For a cross‑linguistic approach, see the synopsis of classifications in Ramat and Ricca (1998). 3 The 1972 edition corresponds fully to the one described here. The 1985 edition distinguished a fourth category, namely “subjuncts” (examples: We haven’t yet fin‑ ished. Would you kindly wait for me? Quirk et al. 1985: 8.88–8.120, pp. 566–612). This fourth category is not adopted here because the whole category of “subjunct” has been repeatedly criticised (for a recent assessment, see Valera 1998: 267–270).

36   The category “adverb” taken over by the leading grammar of English, the Comprehensive Grammar of the English Language (Quirk et al. 1985), one of whose authors was Greenbaum: Table 3.1.: Classification of adverbials in the Comprehensive Grammar (Quirk et al. 1985) and the Longman Grammar (Biber et al. 1999)

scope: phrase scope: sentence scope: sentence/text

Quirk et al. 1985 adjunct content/style disjunct conjunct

Biber et al. 1999 circumstance adverbial stance adverbial linking adverbial

3.2.1.  “Circumstance adverbials” or “adjuncts” Circumstance adverbials are the ones which are best integrated into the clause structure. They are also the most varied class and answer questions such as “how, when, where, how much, to what extent?”: (27) Slowly, they back home. (28) He to me about it briefly.

3.2.2. “Stance adverbials” or “disjuncts” (content/attitudinal and style disjuncts) Stance adverbials (Quirk’s “content or attitudinal disjuncts”) are epistemic in that they focus on the truth value of the proposition, commenting on such factors as certainty, reality, sources, limitations or precision of the proposition. They have a scope over the entire clause and are always optional. As “content disjuncts”, they comment on the content of the proposition: (29) wisely . Paraphrase: ‘(I, the speaker, think that) it was wise of her that she didn’t attempt to apologise’.

Also, subjuncts are of no significance to the present study, because they work – like adjuncts – on the level of the phrase, and not on the level of the sentence or above.

Classification of adverbials   37

(30) The Yard’s wonder boy, appropriately, descended from the clouds. Paraphrase: ‘(I, the speaker, think that) it was appropriate that the Yard’s wonder boy descended from the clouds’. The other kind of stance adverbials (Quirk’s “style disjuncts”) convey a speaker’s comment on the style or form of the utterance, often clarifying how the speaker is speaking or how the utterance should be understood: (31) Frankly, . Paraphrases: (a) ‘Frankly speaking, I’m tired’. (b) ‘I’m frank when I say that I’m tired’. 3.2.3.  Adverbial connectors (“linking adverbials”, “conjuncts”) The adverbs which are in the centre of the present study are termed “linking adverbials” by Biber et al. (1999) and “conjuncts”, a term which explicitly stresses their proximity to conjunctions, by Greenbaum (1969a) and the Com‑ prehensive Grammar (Quirk et al. 1985). In the present study, these items will generally be called adverbial connectors. The more specific terms will only be used when necessary for contrast with other adverbial types. Linking adverbials are similar to stance adverbials in two respects: they are optional and their scope – at their lowest level – is the sentence. They also add the voice of the speaker to the proposition. In their case, however, the voice of the speaker does not affect the propositional content of the respective proposition, but reveals the structure of the text. Rather than adding information about the hows and whens of the proposition (circumstance adverbials) or about the truth value of the proposition as far as the speaker is concerned (stance adverbials), they serve a connective function making the relationship between two units of discourse explicit. Linking adverbials have a peripheral relationship with the rest of the clause, and may extend the level of the sentence, so that they can connect units of discourse of different sizes, from sentences to large chunks of discourse. Much more than with the other types of adverbials, this linking function entails a conjunct‑specific set of semantic relations, such as addition and enumeration, summation, result/inference/cause or, as in the following examples, contrast/concession: (32) All our friends are going to Paris this summer. We, however, are going to London.

38   The category “adverb” (33) He didn’t invite her. She wouldn’t have come, anyway. As signals of two‑place relations, they deal with various inter‑sentential relationships and are not involved in the semantic structure of a single sentence. That they do not work on the representational, but on the interpersonal or textual levels is evident from a number of syntactic restrictions (see Quirk et al. 1985: 631–634 and the full account in Greenbaum 1969a: 41–44): A linking adverbial cannot be the focus of a cleft sentence (34a), it cannot be the basis of contrast in alternative interrogation or negation (34b), and it cannot be focused by a focalizer such as only (34c): (34) She may be unable to attend the meeting. You should nonetheless send her the agenda. (34a)  *It is nonetheless that you should send her the agenda. (34b)  *Should you send her the agenda nonetheless or therefore? (34c)  *You should only send her the agenda. In agreement with their peripheral role in the sentence, they may – similar to stance adverbials – be paraphrased by a separate clause with a verb of communication plus a circumstance adverbial: (35) One can say + circumstance adverbial then, now (time) I tell you + circumstance adverbial here, further (space) I will say + circumstance adverbial consequently, thus (result) These paraphrases also show that there is a natural path from circumstance to linking adverbials. Indeed, many of the new adverbial connectors which have developed in the history of English started their life as circumstance adverbials on the phrasal level, modifying a verbal phrase containing a verbum dicendi. In the following passage from Ælfric’s De Temporibus Anni (COTEMPO), for example, furðor ‘further’ and nu ‘now’, which may also function as linking adverbials when on their own, are used as circumstance adverbials modifying the verbs sprecan ‘speak’ and secgan ‘say’. (36) We willað furðor ymbe ðas emnihte swiðor sprecan on gedafenlicere stowe; & we secgað nu sceortlice þæt se forma dæg ðyssere worulde is … (COTEMPO, p. 19). ‘We will further and more correctly speak about the equinox in the suitable place; and we say now briefly that the first day of this world is …’.

Semantic categories of adverbial connectors   39

This use of circumstance adverbials with a verbum dicendi is widely attested for all periods of English: (37) Also it is said that (Theriaca Athanasia) doe both resolue, breake and digest humours, being compact and gathered together in the profundity of the body. (Mercurialis) saith moreouer, that … (CESCIE2A, p. 11–12). These contexts of verb of communication plus circumstance adverb of time and space, etc. are prototypical “bridging contexts”, which allow the re-interpretation and use of circumstance adverbials as linking adverbials, most often by metaphorical extension. The source domains of circumstance adverbials which are used as linking adverbials are predominantly time (cf. causal then), space (cf. causal hence) and truth/fact (OE soþlice, ME forsooth; EModE indeed, PDE in fact; for a detailed account, see below, Chapter 8).

3.3.  Semantic categories of adverbial connectors As has been mentioned above, the linking function of adverbial connectors entails, much more than with the other types of adverbials, a conjunct(ion)‑specific set of semantic relations (see Quirk et al. 1985: 634–640; Biber et al. 1999: 875–879). Although the grammarians choose different labels, the categories distinguished are basically the same for different languages and have also changed only slightly since the first detailed account by Campbell ([1776] 1963: 404). Campbell distinguishes the following relations: Table 3.2.: Semantic relations marked by adverbial connectors according to Campbell ([1776] 1963: 404) – copulative (when facts are related in continuation, or when one argument, remark, or illustration, is with the same view produced after another) And, now, also, too, likewise, again, besides, further, moreover, yea, nay, nor – disjunctive a) adversative (if the sentiment in the second sentence is in any way opposed to that which immediately precedes) But, or, however, whereas b) exceptive (if it is produced as an exception): yet, nevertheless – causal (if the latter sentence includes the reason of what had been affirmed in the first): for – illative (if it contain an inference): therefore, then

40   The category “adverb” Similarly, the Comprehensive Grammar (Quirk et al. 1985: 634) distinguishes six conjunctive roles,4 in some cases with subdivisions. This system is adopted for the present study, so that Appendix B – providing the corpus findings for the diachrony of adverbial connectors – is structured according to these sub‑categories of semantic relations. The following survey exemplifies the semantic relations by Present Day English conjuncts (as realized not only by adverbs, but also by prepositional phrases or clauses): Table 3.3.:  Semantic relations of adverbial connectors – Enumeration and addition (see Appendix B.1) enumeration: firstly, secondly, thirdly (etc.); to begin with addition: equative correspondingly, equally, likewise, similarly reinforcing again, also, further, furthermore, moreover, in particular, then, too, (above all) – Summation (see Appendix B.2) in sum, to conclude, all in all, in conclusion, overall, to sum‑ marize – Result/Inference/Cause (see Appendix B.3) therefore, consequently, thus, so; hence, in consequence – Contrast/Concession (see Appendix B.4) on the other hand, in contrast, alternatively; though, anyway, however  focus on contrast: conversely, instead, on the contrary, in contrast  focus on concession: anyhow, besides, nevertheless, still, in any case – Transition (see Appendix B.5) now, meanwhile; incidentally, by the way

The only category which is not analysed in detail in the present study is apposition (in other words, i. e., that is, for instance; namely, specifically), since these forms mainly work on the constituent or phrase level. They only rarely mark a two‑place relation on the sentential or discourse level (for details, see below, Chapter 11). Synchronically, all semantic relations apart from enumeration/addition form a closed class, and thus have a fixed number of adverbial connectors coding the respective category. New items cannot be coined instantaneously but 4 Almost identically, the Longman Grammar distinguishes the semantic categories (1) Enumeration and Addition, (2) Summation, (3) Apposition, (4) Result/Inference, (5) Contrast/Concession and (6) Transition (Biber et al. 1999: 875–879).

Semantic categories of adverbial connectors   41

have to undergo the various common steps in language change – in the case of prepositional phrases, e. g., fusion, coalescence – to become full, i. e. coded members of the group.5 The only exception is the category enumeration and addition, which forms an open class, i. e. new items can be coined at all times as nonce‑formations from all kinds of lexical material (see Appendix B.1 for forms such as, e. g., fourthly, sixthly or first, firstly, in the first place, first of all). Since they form an open class, they are not listed in Appendix A and are not analysed in the quantitative sections of this study. 3.4.  “Pure” and “impure” connectives The Cambridge Grammar proceeds one step further and draws a distinction between “pure” and “impure” connectives (Huddleston and Pullum 2002: 775). In this view, “pure connectives” like moreover and also have no other function than that of connecting their clause to the surrounding text, while “impure connectives” such as nevertheless, then or consequently combine this connecting function with a function of concession, condition or reason/result. This distinction is also mirrored in the changes affecting the system of “pure” and “impure connectives” over time. Changes in the ccc‑relations (cause, concession/contrast) are indicative of more general patterns in language structure and long‑term developments than are the changes of the mono‑dimensional adverbial connectors expressing addition or transition. These do not primari­ ly testify to structural changes, but to changes induced by attempts at more expressivity or novelty.6 5 In the analyses, I follow the “Invited Inferencing Theory of Semantic Change” (Traugott and Dasher 2002) to account for conventionalizing of pragmatic meanings, i. e. implicatures or invited inferences, and their reanalysis as semantic meanings (cf. Levinson 2000). Following Levinson, this dynamic theory of language change builds on the distinction of three levels of meaning: coded meanings (convention of a language at a given time), utterance‑type meanings (generalized invited inferences; GIINs) and utterance‑token meanings (invited inferences which have yet not been crystallized into commonly used implicatures; IINs). Historically, there is a path from coded meanings to utterance‑token meanings (IINs) to utterance‑type, pragmatically polysemous meanings (GIINs), to new semantically polysemous (coded) meanings. Only when certain adverbs or original phrases have one of these semantic categories as their coded meaning, they will be analysed as “adverbial connector”. 6 condition does not play a role here, since it is only coded by subordinators, not by adverbial connectors; see above, Chapter 2.4.

42   The category “adverb” 3.5.  Adverbial connectors in Present Day English: corpus findings 3.5.1.  The corpus of the Longman Grammar As a basis for the following analyses of historical material, I will now summarize the main corpus findings for Present Day English as provided by the comprehensive account of the Longman Grammar of Spoken and Written English (Biber et al. 1999: 765–776 and 880–892). For the analyses of the Longman Grammar, the team compiled a core corpus of four main registers comprising ca. 5 million words each: conversation, fiction, newspaper language and academic prose (see the summary and tables in Biber et al. 1999: 24–28). 3.5.2.  Different types of adverbials over the core registers Although optional, adverbials in general are comparatively common – a fact which is, however, hardly surprising in view of the great variety of functions and meanings they may code. While circumstance adverbials are by far the most common class of adverbials in all four registers, linking adverbials are most common in academic prose, but they still only account for less than 10 per cent of all adverbials in that register. While fiction makes great use of many different circumstance adverbials for emphasis or description, conversation has a slightly higher number of stance adverbials than the other registers to convey judgements and abilities or to mark exactly how speakers mean their utterances to be understood. Linking adverbials are more frequent in academic prose, which puts an emphasis on conveying logical coherence. The most important situational properties for academic prose, the text type primarily chosen for the diachronic study here, are that it is (a) in a written mode, (b) does not show any signs of interactiveness or online production and (c) has no shared immediate situation between writer and reader. Its main communicative purpose is information, argumentation and explanation for a specialist audience. Linking adverbials are a very apt means to these ends, because they allow the writers to mark the development of their arguments overtly by relating one proposition to another and by explicitly showing contrasts, restatements and conclusions. This is immensely important in this written, non‑interactive mode which demands precision.

Adverbial connectors in Present Day English: corpus findings   43

3.5.3.  Syntactic realizations of adverbials As has been stated above, adverbials in Present Day English may not only be realized by single adverbs or adverb phrases, but also by noun phrases, prepositional phrases, finite and non‑finite clauses, such as participial clauses (Now adding/added to that …), to‑infinitive clauses (To conclude, …) or verbless clauses. Apart from non‑finite and verbless clauses, these possibilities have also been available in all of the earlier periods of English. In terms of overall frequency (Biber et al. 1999: 767–770; numbers per one million words), prepositional phrases (ca. 50,000) and single adverbs (ca. 30,000) are the most common syntactic realizations, while finite clauses (ca. 10,000), non‑finite clauses (ca. 5,000), noun phrases (ca. 5,000) and adverb phrases (ca. 2,500) are relatively rare. Closer analysis shows, however, that only circumstance adverbials exhibit a strong preference for prepositional phrases. Stance and especially linking adverbials are realized more commonly by single adverbs. Linking adverbials “display the strongest association with a single syntactic form: almost 80 per cent of the linking adverbials are realized by single adverbs” (Biber et al. 1999: 768). The Longman Grammar explains these differing realizations by the fact that, unlike circumstance adverbials realized by a prepositional phrase, many stance and linking adverbials have a more fixed meaning. Individual adverbials of these types are therefore used in a much higher token frequency – a frequency which we would expect from linguistic items forming a closed class (though the Longman Grammar does not argue along those lines). For the diachrony of adverbial connectors, it is also crucial that prepositional phrases which are used as linking adverbials (e. g. as a result, for that cause) tend to be less versatile than those used as circumstance adverbials. These frequent, fixed collocations are a prerequisite for many of these prepositional phrases to be fused and eventually univerbated and lexicalized (see below, Chapter 7). 3.5.4.  Positions of adverbials Adverbials are comparatively free as to their position in the sentence. While subject, verb and object(s) are sequentially fixed in Present Day English, adverbials can take three major positions: initial, medial, and final.7 The overall 7 The Longman Grammar does not distinguish the position after the subject (“Nacherstposition”). In the present study, this position has, however, emerged to be very

44   The category “adverb” frequencies show that all of the three positions distinguished in the Longman Grammar are common (Biber et al. 1999: 770–774). Each class of adverbial, however, has a strong preference for a different position: while circumstance adverbials are most frequently found in final position, stance adverbials are commonly found medially. Linking adverbials favour initial position. This distribution is obviously related to or even dependent on the different meanings, and, in particular, scopes of these adverbials. Circumstance adverbials often complete the meaning of the verb and must thus follow the verb (and therefore, in Present Day English, also the subject). Stance adverbials, which typically have an extended scope over the proposition of the entire clause, may be placed rather freely. Linking adverbials are most often used in initial position, so that the connection between two clauses is clearly signalled as the reader or hearer moves from the first to the second connect. Thus the readers’ or hearers’ processing of the discourse is facilitated by the initial linking adverbials explicitly marking the logical relationship between the connects. The quantitative findings, however, also show another interesting feature, namely that linking adverbials may also be placed at the end of a sentence (for example, PDE causal then, concessive however and though, though predominantly in the spoken medium). My diachronic analysis has shown this to be a relatively recent phenomenon. Final positioning of linking adverbials seems to run counter to the arguments relating to information processing given above, because one wonders why clausal connection should be marked retrospectively at the end of a sentence, i. e. not between the first and the second connect, but after the second connect. Since this position is specific to English adverbial connectors and is only rarely allowed with German connectors (“Nachsatzposition”; see Pasch et al. 2003: 572), it will be discussed in more detail below (Chapter 10.5). 3.5.5.  Linking adverbials: distribution of semantic categories Linking adverbials are, as has been pointed out above, considerably more common in conversation and academic prose than in fiction and news. For academic prose, this has already been related to register‑specific properties. It is a very important purpose of academic prose to present and support arguments in a precise and unambiguous way. The Longman Grammar relates this higher frequency of linking adverbials to these communicative needs and, furthermore, points out that it is the characteristic choice of this register to mark, as the arguments are developed, the links between ideas overtly. different from other “medial” positions with respect to focussing strategies (for details, see below, Chapter 5.4.3).

Adverbial connectors in Present Day English: corpus findings   45

In the other registers, in particular news and fiction, it is much more common to leave the relationships – apart from contrast/concession (see above) – between ideas implicit. In news, for example, sequences of events are reported with respect to their relationship in time, and relations of cause and result are presented so as to be inferred from a chronological sequence or common knowledge and experience. In news and fiction, an overuse of linking adverbials might, according to the Longman Grammar, sound too academic and patronizing. This is also reflected by the fact that texts of academic prose in particular use the “pure” linking adverbials expressing the relations enumeration, addition and summation more commonly than the other registers. Unlike concession, for instance, which highlights a contrasting information and is therefore most commonly marked in all registers, these “pure” semantic relations in particular do not have to be obligatorily marked, but are explicitly coded in academic prose, as this genre specifically tries to mark all logical relations overtly. These connectors help to structure the information in often rather complicated academic prose and give the readers clear signposts of where they are in the text. It may at first glance be surprising that linking adverbials are more common in conversation than fiction or news. Yet, this higher frequency is mainly due to the high frequency of two adverbial connectors, namely so and then as resultive/inferential linking adverbials, and to an increasingly frequent use of anyway and though. 3.5.6.  Summary: Present Day English corpus findings These corpus findings for adverbial connectors in Present Day English first of all suggested restricting the historical analysis to single adverbs, since they constitute 80 per cent of the realizations of linking adverbials in Present Day English. In view of the origin of many of the items, such as, e. g., nonethe‑ less or indeed, in (prepositional) phrases, however, the inclusion of lexicalized prepositional phrases such as in fact, after all proved to be necessary (for the problems of distinguishing lexicalized phrases in earlier stages of English, see Chapter 7). The positional frequencies further indicate that the sentence‑final position of linking adverbs awaits a closer analysis, since it is a recent phenomenon running counter to the expectations of information processing. The predominance of adverbial connectors in the register academic prose then explains the choice of corpus texts (see Appendix C.2). Fiction, for which we would have had a substantial amount of material from all periods of English

46   The category “adverb” in the form of saints’ lives or other narrative prose, is excluded here because the Present Day corpora attest only very few numbers of adverbial connectors for fictional texts (see also the examples, above, Chapter 1.7). News also shows very few instances of adverbial connectors, and is furthermore a register which may only be analysed with the emergence of the first newspapers, i. e. from the middle or end of the eighteenth century at the earliest. For conversation, the register which shows the second highest score for adverbial connectors, we have too little evidence from earlier periods of the language to allow substantial results. In order to get at least an impression of the spoken mode, all relevant texts from the Helsinki Corpus, such as Trials or Plays, were analysed. In contrast to Present Day English, however, adverbial connectors are not very frequent in these interactive, “conceptionally oral” texts. What we find are a high number of interrogatives used as interjections, such as why, how (see Appendix B.5.4); these, however, are excluded from the present investigation of adverbial connectors because they do not mark a two‑place relation, but are rather to be seen as pragmatic idioms, constituting independent (pragmatic) messages (Fraser 1999: 943; see below, 3.6). The analysis of these texts has, nonetheless, proved to be highly beneficial because they record the first instances of sentence‑final adverbial connectors (see below, Chapter 10.5). The focus on academic prose is, however, not only due to the high frequency of adverbial connectors in this register, but also to the fact that the Longman Grammar suggests that it is the characteristic stylistic choice of this register to mark the links between ideas overtly. 3.6.  Adverbial connectors – discourse markers In various research contexts, some of the items discussed in the present study have not been classified as “adverbs”, but as “discourse markers” (OE þa in Enkvist and Wårvik 1987 and Kim 1992; OE soþlice, OE witodlice in Lenker 2000; EModE indeed, in fact in Traugott and Dasher 2002; PDE after all in Traugott 1997; for a discussion, see Fischer 2007: 280–281). In the present study, they are categorized as “transitional”. This label highlights that they do not encode any kind of propositional meaning, but only work on the interpersonal and textual levels, signalling the connection of two segments of discourse, i. e. their status as connects. The term “discourse marker” is notoriously difficult to define and has seen manifold definitions in the last years, when the study of discourse markers “has turned into a growth industry in linguistics” (Fraser 1999: 932). Most of the research on discourse particles for English is, sometimes critically, based

Adverbial connectors – discourse markers   47

on the first systematic and detailed study in the field by Schiffrin (1987).8 In spite of all the ensuing work on the field, however, it has not been possible to define the term in a generally accepted way, let alone to establish a fixed group of linguistic items which would essentially be separated from other classes by being “discourse markers”. For the issues discussed in the present study, it is interesting to see that there are very wide definitions of the concept “discourse marker”. Fraser (1999), for example, arrives at the following definition. Discourse markers are lexical expressions drawn primarily from the syntactic classes of conjunctions, adverbs, and prepositional phrases. With certain exceptions, they signal a relationship between the segment they introduce, S2, and the prior segment, S1. They have a core meaning which is procedural, not conceptual (Fraser 1999: 950).

The properties proposed in this definition are basically the same as those posited for connectives in the present study (following the Handbuch der deut‑ schen Konnektoren, Pasch et al. 2003; see above, Chapter 2.2). Their central property is that they are procedural rather than conceptual: They “relate discourse segments”, i. e. mark a two‑place relation (Fraser 1999: 931) and do not contribute to the propositional meaning of either segment. To capture these properties, the present study uses the term “connector”. It excludes – in line with Fraser – modal particles and style as well as content disjuncts, since these do not signal a two‑place relation (see the exclusion of peradventure, perhaps, perchance; Chapter 12.2). It also excludes interjections such as oh, yes/yea, or no/nay (see Appendix B.5.5), which do not mark two‑place relations either, but are “pragmatic idioms” in that they constitute entire, separate messages (Fraser 1999: 943). Yet some of the connectives, in particular those coined from the source domain truth (see Chapter 8.4), show a development which has been labelled “Adverbial Cline” in a semantic‑pragmatic approach to grammaticalization (Traugott 1997, 1999). This cline from clause‑internal adverbial to sentence adverbial to discourse particle is established as a “regularity in semantic change” (Traugott and Dasher 2002) in order to distinguish it from grammaticalization 8 Schiffrin (1987: 328) gives the following “tentative suggestions as to what specific conditions allow an expression to be used as a marker. A discourse marker has to be syntactically detachable from a sentence, it has to be commonly used in initial position of an utterance, it has to have a range of prosodic contours (e. g. tonic stress and followed by a pause, phonological reduction), it has to be able to operate at both local and global levels of discourse … this means that it either has to have no meaning, a vague meaning (…) ”. Cf. Kroon (1995: 7–57) and Brinton (1996: 29–65) for the literature on discourse markers and its terminology.

48   The category “adverb” processes of the morpho‑syntactic kind which involve, for example, increased bonding and a loss of syntactic scope. The diachrony of many epistemic modal adverbials expressing the speaker’s commitment to the truth of the proposition (OE soþlice, ME forsoothe, EModE truly) shows that they – at some late stage of their development – do no longer primarily signal a two‑place relation, but carry an independent pragmatic meaning, i. e. are “discourse markers” in a more restricted sense. On the global level of textual organization, OE soþlice and ME soothly, for example, often serve as episode boundary markers (see below, Chapter 8.4), a function which may at the textual level be considered “connective”. Furthermore, these items also exhibit other patterns of recurrent change than the other adverbial connectors (see below, Chapter 12). The discrete, peripheral status of these connectors is indicated by their label “transitional” (which implies their borderline status). Yet, these considerations do not affect the general relevance of these items for the present study: Before they develop these primarily pragmatic functions, they unambiguously signal two‑place relations, mainly on the local level of discourse (Traugott’s stage “sentential adverb”; for details, see Chapter 8.4). Furthermore, their status as connectives is also corroborated by the information gathered from meta‑linguistic texts of the respective periods (see the following Chapter 4).

4. Adverbs and conjunctions in earlier metalinguistic thought 4.1.  Introductory remarks One source from which to collect relevant linguistic material which allows for a better understanding of the development of adverbial connectors – and their relation to conjunction – are metalinguistic texts. While it is not difficult to gather data from Present Day English sources such as grammars or dictionaries, contemporaneous metalinguistic sources are unfortunately of rather limited value before the middle or end of the Early Modern English period: the sources before Campbell ([1776] 1963) do not distinguish adverbial connectors as a separate, distinct category. The items are usually listed among the “adverbs”, which are commonly sub‑grouped semantically (see the quotes from Locke in Chapter 1.1), but are not further distinguished into functional subclasses. For a long time, the elements in question were included in an even larger group of indeclinable items and are classified together with prepositions and conjunctions as “particles”. The increasing sophistication in the classification of these linguistic elements is not only interesting because it illustrates the developments in English language scholarship, but also because it mirrors the growing importance of these elements, both in frequency, diversity and stability of use in and after the Early Modern English period. 4.2.  The Greek and Latin tradition For a close analysis of adverbial connectors and conjunctions in earlier periods of English, it is obviously also crucial to look at how they were conceptualized by medieval grammars of Latin and the vernacular. Following the categorization of Dionysius Thrax, there are eight parts of speech that turn up again and again in the Latin grammatical tradition, namely (1) noun, (2) pronoun, (3) article, (4) verb, (5) participle, (6) adverb, (7) conjunction, and (8) preposition (see Vorlat 1975: 43; Law 1997: 264–269). Medieval grammarians take on Thrax’s classification, which they were acquainted with via transmission by Latin authors such as Donatus, with only two adjustments: articles are omitted because they do not exist in Latin, and the interjection is detached from the adverb and becomes a separate word class (Vorlat 1975: 43). The standard sequence in medieval grammars thus is nomen, pronomen, uerbum, aduerbium, participium, coniunctio, praepositio, and interjectio.

50   Adverbs and conjunctions in earlier metalinguistic thought The word classes which are important for our study – adverb, interjection, and conjunction – together with the preposition form the “minor parts of speech”, which generally receive little attention from the grammarians (Vorlat 1975: 366–419). These minor parts of speech are separated from the other word classes by their indeclinable character and their dependency on other word classes (“particles”). The classification of adverbs into the syntactically determined subgroups outlined above is a fairly modern one: in traditional grammatical treatises, adverbs are not usually categorized according to their syntactic properties. The most influential early grammarians (Thrax, Varro, Donatus, and Priscian) take the term adverbium in the strict sense of a word added to a verb whose meaning it modifies or explains, and the authors indulge in elaborate, if not to say tedious, semantic classifications of adverbs.1 This exclusively semantic approach is taken on by most of the English grammarians – also those following the traditions of Ramus or the Port‑Royal Grammar (see Michael 1970: 72–75).2 The subcategories we now call stance adverbials (disjuncts) and linking adverbials (conjuncts) are not mentioned in any of the grammars of the Medieval and Early Modern Period (Vorlat 1975: 366–387). This is not really surprising because these categories are, as shown above, mainly based on syntactical properties. In the early grammars, however, syntax is mostly restricted to questions of concord and government (cf. Gneuss 1996: 15). Yet, the early grammars do not completely lack value for the issue in question. Although they do not explicitly refer to specific syntactical properties, the grammarians repeatedly point out that the distinction between the various word classes of the group of indeclinables is not always clear-cut.3 1 Donatus, for example, has adverbia finita et infinita, and furthermore the categories adverbia loci, temporis, numeri, negandi, adfirmandi, demonstrandi, optandi, hortandi, ordinis, interrogandi, similitudinis, qualitatis, quantitatis, dubitandi, personalia, vocandi, respondendi, separandi, iurandi, eligendi, congregandi, pro‑ hibendi, eventus, comparandi (Ars Minor, “De Adverbio”, II, XIII). Priscian has temporalia (with further subclassification), locorum, dehortativa, abnegativa, con‑ firmativa, iurativa, optativa, hortativa, remissiva, qualitatis, quantitatis, dubita‑ tiva, congregativa, discretiva, similitudinis, ordinativa, intentiva, comparativa, superlativa, diminutiva (XV, 28–37; cf. Vorlat 1975: 367). 2 The Rameian tradition, for example, contrasts adverbs as word connectors with conjunctions as clause and sentence connectors (see Vorlat 1975: 368). 3 Various grammarians further remark that a number of prepositions – in Latin as well as in English (cf. up) – may also function as adverbs (Vorlat 1975: 376). Adverbs are furthermore often considered to be of an abbreviating character, because they may be paraphrased by a prepositional phrase consisting of a preposition and a noun. Thus sapienter is seen as short for cum sapientia or English wisely short for with

Ælfric’s Old English Grammar   51

What is even more important for the present study – in particular the situation in Old and Middle English – is that already Priscian and Donatus point out that they see no clear distinction between adverbs and conjunctions. In a discussion of the adverb, Donatus, for instance, explains that there is sometimes no way of determining whether a specific item is an adverb, a conjunction or a preposition, unless its function in the sentence is taken into account: Sunt etiam dictiones, quas incertum est, utrum coniunctiones, an praepositiones, an adverbia nominemus […], quae tamen omnes sensu facile dinoscuntur, […] horum quaedam accentu discernimus, quaedam sensu (II, XV; cf. Vorlat 1975: 367).

English grammarians adopt this remark time and again, i. e. that word class “is determined by function” (Vorlat 1975: 367), which means that the grammarians do indeed see the disparity in the scope of different adverbs, but do not use this criterion as a property for classification. For the Early Modern grammarians, it does not seem to be an important issue; Newton, for example, in an imitation of Wallis, states: “I shall reckon some of these Words as Adverbs, and some of the Adverbs as Conjunctions, they being often used in both Senses, there will be no great harm done” (Vorlat 1975: 376). 4.3.  Ælfric’s Old English Grammar With respect to metalinguistic thought on English, the earliest extensive grammatical account is Ælfric’s Latin‑Old English Grammar (Zupitza [1880] 2001). It is first of all important to note that Ælfric’s account, which is based on a Latin excerpt of Priscian (see Porter 2002), basically follows the Latin tradition in that he distinguishes adverbs and conjunctions in principle (see the two sections entitled “Incipit Aduerbium” and “De Coniunctione”). This distinction is, however, overridden by functional considerations. This may be exemplified by the case of the adverb soþlice ‘truly, verily’, which is morphologically clearly marked as an adverb by the adverbial suffix {‑e} added to its base, the adjective soþlic ‘true’ (for details, see Lenker 2000, 2003 and below, Chapter 8.4.4). Ælfric, however, does not list it in his section on the adverbs (“Incipit Aduerbium”; Zupitza 2001: 222–242), but explicitly refers to this adverb (!) in his section “De Coniunctione” (Zupitza 2001: 257–266, at 261). Conjunctions are defined as follows: wisdom (see the examples of the tradition of the Port‑Royal Grammar given in Vorlat 1975: 374, 380). For a recent grammar which revives this classification of adverbs in one group with preposition, see Huddleston and Pullum (2002: 264).

52   Adverbs and conjunctions in earlier metalinguistic thought Coniunctio est pars orationis indeclinabilis adnectens ordinansque sententiam. … swaswa lim gefæstnað fel to sumum brede, swa getigð seo coniunctio þa word togædere. þes dæl gefæstnað and and gefrætwað ledenspræce and hwilon toscæt and hwilon geendebyrt. … næfð þes dæl nane mihte ne nan angit, gif he ana stent, ac on endebyrdnysse ledenspræce he gelimaþ þa word … (Zupitza 2001: 257–258). ‘A Conjunction is an indeclinable part of speech connecting and ordering the sense …; in the same way as lime fixes a hide to any surface, so conjunction draws the words together. These parts of speech fix and adorn Latin and sometimes they divide and sometimes they order. … This part of speech does not have any power and no meaning, if it stands on its own, but in the order of Latin it links the words’.

By discussing adverbs like soþlice ‘truly’ in the section “De Coniunctione”, Ælfric clearly shows that he considers the scope of the adverb to be not the verb phrase but the sentence: he thus implicitly classifies soþlice as an adverbial connector, a “conjunct”. Ælfric even further distinguishes between two uses of soþlice as a “conjunction”. First, soþlice (and also witodlice and gewislice) are found as translations of Latin items, which in our terminology are not conjunctions, but adverbial connectors, namely autem, enim, uero and nam etc., which belong to the group of the Expletivae or Completivae “þa gefyllað and gefægeriað þa ledenspræce, and, þeah ðe hig forlætene beon, ne byð swa ðeah þære spræce andgit forlæten” ‘which fill and adorn the Latin, and, even if they are left out, the sense of the utterance will not be lost’ (Zupitza 2001: 261): [H]er synd þa: autem, enim, uero, quidem, equidem, quoque, nam, namque, uide‑ licet. tu autem, domine, miserere mei et resuscita me. ðu, soðlice, drihten, miltsa me and arær me. ego enim sum dominus, deus tuus ic, soðlice, eom drihten ðin god. […] doctus sum. nam legi ic eom gelæred; soðlice, ic rædde. erat namque in sermone uerax. he wæs, soðlice, on spræce soðfæst […].

Ælfric explicitly refers to a loss of the propositional meaning (“ne byð swa ðeah þære spræce andgit forlæten”), which is essential for the use of soþlice ‘truly’ as a sentential adverb with text-organizing function, i. e. as a “pure”, transitional connector.4 This is most obvious in the last example cited above (erat namque in sermone uerax “he wæs, soðlice, on spræce soðfæst”, ‘He was, therefore, true‑speaking/reliable in his speech’): Ælfric would certainly not have chosen soþlice to render namque (but one of the alternatives such as gewislice or witodlice) if its propositional meaning ‘truly’ had still been principal to him, since this ambiguity could have obstructed the understanding of the 4 For the loss of propositional meaning (“bleaching”) and pragmatic strengthening as prototypical features of grammaticalization and similar kinds of language change, see Hopper and Traugott (1993: 87–93).

Grammatical treatises of the Middle English period   53

proposition of the sentence, which has to do with “true speaking” (Lat. uerax, OE soþfæst ‘true, sincere’). The adverbs used in the function of a conjunction are, however, not only stylistically important for the adornment of a text. Ælfric also lists witodlice ‘certainly, truly’ among the Rationales, “þas sind for sumon gesceade gesette on endebyrdnysse ledenspræce” ‘which are set for an (understanding of the) argument in the text organisation of Latin’ (Zupitza 2001: 263). Ælfric thus stresses that their function, translating Latin items such as ergo, igitur, ita, itaque and utique, is to add reasoning to the discourse. In sum, Ælfric’s Grammar does indeed prove to be helpful for an analysis of the syntactic scope and functions of Old English adverbs. For Ælfric, the defining element for a coniunctio is not the word class (neither for Latin – cf. enim, nam, etc. – nor for English – cf. soþlice), but their scope over the sentence and their function as organizers of textual structure. He thus uses the same criteria which are today chosen for the categorization of an item as a “conjunct” or “linking adverbial”.

4.4.  Grammatical treatises of the Middle English period With respect to Middle English, grammatical treatises dealing with the vernacular only survive from the end of the fourteenth century onwards, and their production reaches its peak as late as 1460–1480. These Middle English grammatical texts, which are now conveniently collected in Thomson (1984), were not designed as abstract grammars but as “working tools” for learning Latin (Thomson 1984: xiv) and show an immense degree of interdependence. Because of their repetitive character, it suffices to quote exemplary samples from one of the sources here (Aberystwyth, N. L. W., MS Peniarth 356B, fols. 54v– 57v and 48r; Text A in Thomson 1984). The adverb, in accordance with the Latin school tradition, is seen as modifying the verb only: How knos þu aduerbe? A party of speech þat ys vndeclynyt, þe wych ys cast to a verbe to declare and fulfyll þe sygnific(ac)ion of þe verbe […] (Thomson 1984: 6).

The accounts – because they are grammars of Latin and not English – only provide information on the different patterns of Latin adverb formation from the different classes of adjectives, but virtually no English examples. This method is also applied in the descriptions of the “conjunctions”: How knos a coniunccion? A party of speech þat ys vndeclynet and ionys oþer partys of spechys togedyr. How mony thyngus longon to a coniunccion? III. Wech iij? Powere, fygur and ordyr. y powers of coniunccion byn þer? V. Wech v?

54   Adverbs and conjunctions in earlier metalinguistic thought Sum be copulatyuis, sum byn byn explatyuis, sum casuels and sum racionels (Thomson 1985: 7).

The only account of the conjunction which is not taken from Priscian or Donatus and which supplies some information on the vernacular is found in Thomson’s “Accedence Text D” (Thomson 1984: 32–43; Cambridge, Trinity College, O. 5. 4, fols. 4v–6v). This source provides an exposition of the various categories of conjunctions which is unusually full; more importantly, it gives sample sentences and their English translations (cf. Thomson 1984: 244): How knowest a coniunccion? A party of reson that is not declynyd […]. How knowest a coniunccion expletyf? That at fulfylleth the sentence of a reson that is folwyng, as ‘I forsothe haue souped, thu forsothe not’, Ego quidem cenaui tu vero non (Thomson 1984: 42).

Interestingly, the adverb forsothe ‘truly’ is here – like in Ælfric’s Grammar – considered a “conjunction” of the subgroup Expletivae. 4.5.  The prologue to the revision of the Wycliffite Bible Apart from these very scarce remarks in grammatical treatises, there is some information to be gained from the “General Prologue” to the revision of the Wycliffite Bible (1388), which discusses several translation problems. Among them are two passages which overtly deal with the translation of Latin adverbial connectors, such as autem or enim. And whanne oo word is oonis set in a reesoun, it mai be set forth as ofte as it is vndurstonden, either as ofte as reesoun and nede axen; and this word autem, either vero, mai stonde for forsothe, either for but, and thus I vse comounli; and sumtyme it mai stonde for and, as elde gramariens seyn (Forshall and Madden 1850: 57).

The author here explicitly asks for a repetition of certain conjunctions to make the reasoning of the text, the textual organisation, transparent. He includes the adverb forsothe among the words rendering Latin autem and vero and marks their similiarity, but also difference, to unambiguous conjunctions such as adversative but and connective and. In a second instance, the reviser deals more generally with the problems of translating items belonging to the “minor parts of speech”, such as “aduerbis, coniuncciouns, and prepositions”, in a section which shows a high degree of awareness of the problem of textual organization: [A] translatour hath greet need to studie wel the sentence, both bifore and aftir, and loke that suche equiuok wordis acorde with the sentence […]. Also this word ex signifieth sumtyme of, and sumtyme it signifieth bi, as Jerom seith; and this word

The prologue to the revision of the Wycliffite Bible   55 enim signifieth comynli forsothe, and, as Jerom seith, it signifieth cause thus, for‑ whi; […]. Manie such aduerbis, coniuncciouns, and preposiciouns ben set ofte oon for a nother, and at fre choice of autouris sumtyme; and now tho shulen be taken as it acordith best to the sentence […] (Forshall and Madden 1850: 59–60).

Here then, ME forsothe is given as a translation of enim and thus classified as a causal connector. Yet, both passages agree on the description of forsothe as a rendering of Latin conjunctions and thus testify to the analysis of forsothe as an adverbial connector with scope over the whole sentence. The revision of the Bible translation itself, however, points towards a change of attitude on the side of the reviser(s). In the Old Testament, autem, vero and enim are almost always translated by forsothe and sothely in the earlier version, and also very frequently in the second version (the alternative being but). In the New Testament, the earlier text again renders them by forsothe or sothely. The revised text, however, almost exclusively uses the conjunctions but or and, or leaves autem, vero etc. unrendered (Forshall and Madden 1850: xxiii, note a). This becomes clear through a comparison of some verses of the Gospel according to Matthew, chapter XXVI (“The Plot against Jesus” in the Earlier and Later Version): Table 4.1.: Translations of Lat. autem and enim in the Wycliffe Bible (Matthew XXVI) Verse

Latin

Earlier Version

Later Version

XXVI, 8 XXVI, 9 XXVI, 10 XXVI, 12 XXVI, 17 XXVI, 24 XXVI, 26 XXVI, 29 XXVI, 32 XXVI, 33 XXVI, 41 XXVI, 43

[uidentes] autem [potuit] enim [sciens] autem [mittens] enim [prima] autem [die] [filius] quidem [cenantibus] autem [dico] autem [uobis] [postquam] autem [respondens] autem quidem [promptus est] [erant] enim

sothely forsothe sothely forsothe forsothe forsothe forsothe forsothe forsothe sothely forsothe forsothe

and for But – And Forsothe And And But – for for

This comparison shows that the reviser, although he argues strongly in favour of forsothe in the General Prologue where he refers to it as the “common” translation of enim, has almost completely abandoned the word by the time it comes to translating the New Testament. This first of all allows us to infer that contemporary writers and grammarians regarded forsothe as a conjunct with

56   Adverbs and conjunctions in earlier metalinguistic thought text‑organizing function which could and should be employed in prose texts to translate Latin conjunctions, but it also seems to indicate that it was more and more avoided in genuine Middle English prose. In the slightly later translation of the Latin Technical Phlebotomy (­CLPHLEBO; ca. 1400–1420), for example, there is a striking difference between the Middle English translation of the Latin text which regularly uses forsoþe in post‑first‑position, i. e., the sentence position of enim or autem, as in: (38) Alexandire forsoþe commandiþ a pacient leucoflamcie to be fleubotomyed; euel humourus forsoþe if þey be in veynes ar competenly brougt out be flebotomye. Aftir, electuary frigidum confortatiue of þe lyuere; þe lyuere forsoþe it is moche wonte to be febled. Laboryng of ictercian citrina, i. e., gelow iaundeyes, & also of agriacape gasiliontes, i. e., grene jawnes, mynusche of bacilica of þe rygt arme (CMPHLEBO, p. 51). In a later, originally Middle English part, forsoþe is not used at all. Instead, items such as sentence‑initial wherefore, for, also or for hwi are chosen as connectors, as in: (39) Wherefor som vnwise men after þe consayle of oþer onwise men will þann mynusche hem, & of þat folwiþ a werse errore þann þe firste. For þe materie þat was witoute þe weynes be flebotomye is drawen to wiþinne þe veynes; þann falliþ & continewe febre & so oftyn time deþ. Also som men when þei have a febre interpolat where þe materie is wiþout þe veynes, makiþ hem to be mynusch & falliþ into continell febre; for whi: as … (CLPHLEB, p. 53). 4.6.  Early Modern English dictionaries and grammars For the Early Modern English period until Locke, i. e. from 1560 to the middle of the seventeenth century, our main relevant metalinguistic sources are dictionaries (conveniently accessible in The Early Modern English Dictionar‑ ies Database, ed. Ian Lancashire EMEDD, now part of the database Lexi‑ cons of Early Modern English (LEME)), which includes 16 works from 1530 to 1657: six bilingual dictionaries,5 five English hard‑word dictionaries,6 the 5 John Palsgrave (1530; English – French), William Thomas (1550; Italian – English), Thomas Thomas (1587; Latin – English), John Florio (1598; Italian – English), John Minsheu (1599; Spanish – English), and Randle Cotgrave (1611; French – English). 6 Edmund Coote (1596), Robert Cawdrey (1604, based on the transcription by Raymond Siemens; and 1617), John Bullokar (1616), and Henry Cockeram (1623).

Early Modern English dictionaries and grammars   57

first full English‑only dictionary by Thomas Blount (1656), three specialized lexicons,7 and the first full English word‑list by Richard Mulcaster in his The first part of the Elementarie (1582). Although these dictionaries – in line with the grammars for their time – do not separate adverbial connectors from other kinds of adverbs or conjunctions, they are a valuable source because they do not only give lexical, but also functional synonyms (for examples, see below, Chapter 11). As far as grammars are concerned, there was no change in the treatment of adverbs and conjunctions. To cite just one example, Lowth ([1762] 1969) has the same limited definition of adverbs as the Middle English grammatical treatises, restricting them basically to modifiers of verbs and adjectives: Adverbs are added to Verbs and Adjectives to denote some modification or circumstance of an action or quality: as, the manner, order, time, place, distance, quality … (Lowth [1762] 1969: 90).

A conjunction is said to “connect or join together Sentences”, but the example Lowth gives deals with conjunctions on the phrasal level: … so as out of two to make one Sentence. Thus, “You, and I, and Peter rode to London,” is one Sentence made up of three by the conjunction and twice employed (Lowth 1969: 92–93).

This brief summary shows that Locke’s intervention for the Particles in his Essay Concerning Human Understanding ([1690] 1975) quoted at the very beginning of this study did not have any major impact on the grammarians of that period. It had, however, an impact on the very influential school of rhetoricians, which promoted a re‑orientation of rhetoric in line with the ideas of the Enlightenment, notably the so‑called Scottish Rhetoricians advocating a “New Rhetoric”. In Chapter 13, the ideas promoted by one of these rhetoricians, George Campbell, in his The Philosophy of Rhetoric, will be shown to have been decisive for the increasing sentence‑medial positioning of adverbial connectors in the written mode. The increasing sophistication in linguistic terminology thus nicely mirrors the development of the linguistic category itself.

7 Bartholomew Traheron’s translation of Vigon (1543), William Turner on herbal names (1548), and John Garfield on scientific terms in J. Renou’s Dispensatory (1657).

5.  Connectors in Old English

5.1.  Semantic and syntactic polyfunctionality To understand the history of adverbial connectors and, more generally, clause linkage in English over the centuries, it is necessary to first describe the situation in Old English. In Old English, there are only very few linguistic items which function exclusively as adverbial connectors (Group A),1 namely the complex formations naþylæs, swaþeah and swaþeahhwæðere ‘nevertheless’, all of which mark the semantic relation contrast/concession. All of the other connectors are semantically and/or syntactically polyfunctional (Groups B to E). In essence, we can differentiate four major groups: (B) circumstance or stance adverbs with a connective force, (C) so‑called “ambiguous adverbs/conjunctions”, (D) pronominal connectors, and (E) case forms of the demonstrative. (B)

Circumstance and stance adverbs with connector force

(a)

Circumstance adverbs – adverbial connectors addition/ transition



ærest ‘first’ eac ‘also’ eft ‘also’ eftsona ‘also’ eftsones ‘also’ elles ‘else’ her ‘here’ nu ‘now’ (þær)toeacan ‘in addition, besides’ þa ‘then’ þonne (mainly in post‑first‑position) ‘then’ (transition – cause/result)

(b)

Epistemic adverbs (stance adverbials) – adverbial connectors transition: eornostlice ‘earnestly’ huru ‘indeed, truly’

1 Mitchell (1985: § 3430–3433) shows that an adverbial reading is possible or preferable for all instances of OE hwæþere, swaþeah and swaþeahhwæþere for which conjunction uses have been suggested.

Circumstance adverbs with connective force (Group B‑a)   59

huruþinga ‘at least’ soþlice ‘truly’ gewislice ‘certainly’ witodlice ‘truly’ (C)

Ambiguous adverbs/conjunctions (symbol °):2 °eall‑swa ‘adv. also, moreover – conj. as, so’ °nu ‘adv. now, transition – conj. now that …’ °swa ‘adv. so – conj. so that …; as …’ °eac (swylce)/ °swylce eac ‘adv. also, moreover – conj. all such …’ °þa ‘adv. then, transition – conj. then … when’ °þeah ‘adv. nevertheless – conj. although’ °þonne ‘adv. then, transition/cause – conj. then … when’

(D)

Pronominal connectors: °forþæm (°forþon, °forþy) ‘adv. therefore – conj. because’ naþylæs ‘nevertheless’

(E)

Case forms of the demonstrative – adverbial connector cause/result: (a) genitive þæs ‘therefore’ (b) instrumental °þy (ambiguous adverb/conjunction)    ‘adv. therefore – conj. because/since’

5.2. Circumstance adverbs with connective force (Group B‑a): temporal and spatial adverbs For some temporal and spatial circumstance adverbs, a conjunctive force is widely established: frequently used connectors in Old English are OE eac, eac swylce/swylce eac, (þær)toeacan ‘also’ (addition), eft ‘again’ (addition), her ‘here’ (transition), nu ‘now’ (transition), and þa ‘then’ (transition). All of these thus code the semantic relations Addition or Transition. While some of them are exclusively adverbs and are only semantically polyfunctional, a case here exemplified by eft (see also eac, eac swylce/swylce eac, 2 For this term see Mitchell (1985: § 2536). The adverbial meaning is given first; this is followed by the conjunction meaning after the dash. In each case, only one – the central – meaning is given.

60   Connectors in Old English (þær)toeacan and her), nu and þa may also be used in correlative constructions and – on their own – as subordinators (here exemplified by nu). They are semantically and syntactically polyfunctional, and are therefore termed “ambiguous adverbs/conjunctions”. 5.2.1.  OE eft ‘again’ In origin, adverbial connectors of group B follow the path from circumstance adverbial modifying a verb of communication on the phrase level, as sketched above in Chapter 3.2.3. Eft ‘again’, for example, can be used as a regular circumstance adverb of time ‘again, another time, once more’ (DOE, s. v. eft, 1.), as in: (40) …. and se fehð on Ianuario and þær eft geendað. ‘… and it begins in January and ends there again’ (ByrM, 1.1.229)3 As such a circumstance adverb, it can modify a verbum dicendi (DOE, s. v. eft 4.a.iii), such as cyþan ‘speak; announce’ in (41) Nu we wyllað eft fullum mode her gecyðan hu oft se Easterlica monð onginð on Martio. ‘Now (time) we will again (time) make known here (space) with thorough thought how often the Easter month begins in March’ (ByrM, 3.2.96) Already in Old English, however, the co‑ and context in (41) can also be deleted. In contexts such as (42), eft functions as an additive/reinforcing adverbial connector ‘again, as another point of fact (indicating sequence or transition in discourse)’ (DOE, s. v. eft, 4.). In these functions, eft is used quite typically in a 3 To illustrate the situation in Old English, examples are primarily chosen from texts which are comparatively independent from Latin originals. Whenever possible, they are taken from one of the Prefaces (COPREFCP, COPREFSO), the rather free adaption of Boethius’ De Consolatione Philosophiae from King Alfred’s Circle (COBOETH) or Byrhtferth’s Enchiridion (ed. Baker and Lapidge 1991 = ByrM; not all of the examples given here are in the extract chosen for the Helsinki Corpus, i. e. COBYRHTF). Quotations from dictionaries or grammars are only used if there was no instance in one of my selected texts because it is hard to judge from the single sentences in dictionaries how large the piece of discourse (clause, paragraph, text) the elements connect is, i. e. whether they work on the local or global level.

Circumstance adverbs with connective force (Group B‑a)   61

collocation with another connector, the coordinating conjunction and (for these collocations, see below Chapter 13). (42) : And eft, gif he byð ahwar gesett þær he standan ne mæg, þonne sceal man hine þus genyðerian … ‘: And moreover, if it is written anywhere it cannot be, then one must remove it thus: …’ (ByrM, 3.3.184) The same distribution is also attested similarly for all the other items listed above. In example (41), for instance, we also find spatial, text‑deictic her ‘here’ (for its transitional use see Clemoes 1985; Fries 1994: 530–535) and the temporal, text‑deictic nu ‘now’.

5.2.2.  OE nu ‘now’ (Group B‑a, Group C) 5.2.2.1.  Temporal nu Nu ‘now’, of course, may be used as a circumstance adverb of time referring to the actual, present date in the world (‘now, at this time’; BT, s. v. I.). Thus nu in (43c) and (43d) definitely refers to an actual date in February in the year Byrhtferth is writing his Enchiridion, as shown by the obligatory circumstance adverb in the relative clause þe nu is ‘which is now’ in example (43c) and the collocation nu todæg, þonne ic þis write ‘now today, that I am writing this’ in (43d). It is also a temporal adverb, but with an extended metaphorical meaning in (43a) and (43b). 5.2.2.2.  Text‑deictic nu In (43a) and (43b), by contrast, nu does not refer to a fixed date, but to the “now” of reading/or writing of the text, which may be at various dates in the future, since the deictic centre here is not the real world, but the text itself (for a collection of Old English examples, see Fries 1994). Nu has become text deictic, a function which is often highlighted by verbs of communication which refer to the foregoing (cf. handlian in 43b) or ensuing text production (cf. gecyþan in 41 and ætywan in 43a): (43) Nu [43a] wylle ic bysne ætywan ymbe þa þing þe we nu [43b] handledon and fæste ymbe wæron. (\Kl Febris\) , þe nu [43c] is, eode on tun

62   Connectors in Old English



on þunresdæg, & nu to dæg, þonne ic þis write [43d], ys se fifta dæg (ByrM, 1.2.234). ‘Now [43a] I will give an example of the things we have just now [43b] treated and were intent upon. The month of February, which is now [43c], came to town on Thursday, and now today, when I write this [43d], is the fifth day’.

5.2.2.3.  Adverbial connector nu – transition/result In (44), nu functions solely as an adverbial connector – see the collocation with the subordinating conjunction gif ‘if’; here, it indicates a semantic relation of transition verging on the relation result. This use derives from the textdeictic uses sketched above, namely a full phrase such as ‘if we now proceeded to argue that two goods existed …’ where nu again refers to the text as the deictic centre: (44) Gif nu tu good wæren þe ne meahton ætsomne bion, & wæren þeah buto goode, hu ne wære hit þonne genog sweotol þæt hiora nære nauðer þæt oðer? (COBOETH, 34.85.12) ‘If, moreover?/therefore?, two goods existed, which might not be together, and were nevertheless both good, would it not then be sufficiently evident that neither of them was the other?’ 5.2.2.4.  Adverb/conjunction nu in correlative constructions As an adverbial connector, nu can also – just like þa ‘then’ and þonne ‘then’ – be used in correlative constructions, in which connectors are placed in each of the connects, one functioning as a subordinator, the other as an adverbial connector. In the correlative construction in (45), the first nu (45a) functions as a causal subordinator introducing a pre-posed topic-forming subordinate clause (BT, s. v. II “Conj. now; since; when”; for which Present Day English prefers unambiguous now that), while the second (45b) may be classified as a circumstance adverb of time, a text-deictic nu (‘now it seems fitting …’): (45) Nu [45a] we habbað sceortlice amearcod þæra hiwa gefeg þe boceras gymað, nu [45b] þingð hyt us gedafenlic þæt we heom gecyðon þæt we ær geheton … (ByrM 3.3.234) ‘Now that we have briefly written the series of figures the writers use, now it seems fitting to us to tell them what we promised before …’.

Circumstance adverbs with connective force (Group B‑a)   63

This example also shows that – in contrast to V2 vs. V-final word order in Present Day German – constituent order cannot be used as a solid criterion for differentiating Old English main and subordinating clauses: both clauses employ NuV (NuSVAO in the pre-posed subordinate clause; nuVSC in the main clause). In the Old English translation of Boethius, for example, there are many instances of correlative constructions with identical constituent order in both the subordinate and main clause (in 46: Nu SVO): (46) ða cwæð he: Nu [46a] ðu þonne wast hwæt ða leasan gesælða sint & hwæt þa soþan gesælða sint, nu [46b] ic wolde þæt þu leornodest hu þu … (COBOETH, 33.78.27). ‘Then he said: Now that [46a; subordinator] you know (then) what the false goods are, and what the true goods are, now [46b; circumstance adverb] I wish that thou should learn how you …’. 5.2.2.5.  Conjunction (subordinator) nu The use of nu as a subordinator is, however, not restricted to correlative constructions. It may also be found on its own; pre-posed, topic-forming (47a) as well as in regular post-posed subordinate clauses (48): (47) Nu [47a] we ealles heron habbað gefangen, hyt gerist cyrtenlice þæt we ne wandion naðor … Vton awendan nu [47b] ure gesetnysse to þam rihtingum þe rimcræftige preostas cweðað lunares (ByrM, 1.2.251–257) ‘Now that [47a; subordinator] we have fully begun in this matter, it is splendidly suitable that we not desist either … Let us now [47b; circumstance adverb + verb of communication] turn our tract to those regulars which priests skilled in computus call lunar’. (48) Se ilca God is, swa swa we ær sædon, þæt hehste good & þa selestan gesælða, nu hit is openlice cuð þæt ða selestan gesælða on nanum oðrum gesceaftum ne sint buton on Gode. (COBOETH, 34.84.3) ‘The same God is, as we before said, the highest good, and the best happiness since it is evidently known that the best felicities are in no other things but in God’. This account of nu can be taken as representative for most of the ambiguous adverbs/conjunctions in Old English (for forþæm, see below, Chapter 9; for the exception of þeah triggering the subjunctive in subordinate clauses, see below

64   Connectors in Old English Chapter 10.5.2). Word order is only distinctive when verb‑final order is employed: it marks a clause as subordinate. In sum, this means that … when standing first in a sentence, any word which in its written form may be either an adverb or a conjunction, (i) is in prose invariably a conjunction when followed by a conjunctive order, (ii) is ambiguous, i. e. may be either part of speech, both in prose and verse, when followed by the common order (Andrew 1940: 34).

These exemplary analyses of eft and nu also show that it is not possible to undertake the kind of investigation envisaged in the present study by a machine‑triggered count of items in the respective language corpora. The semantic and syntactic polyfunctionality of the items requires a close reading of the respective texts and the classification of highly frequent linguistic items in view of their specific co‑ and contexts. This analysis obviously also had to be restricted to selected corpus texts, since most of the items in question are high‑frequency items in their respective periods: for Old English, the DOE lists, for instance, 6,800 occurrences for eac, 5,500 for eft and 15,500 for the various forms of forþæm. 5.3.  Ambiguous adverbs/conjunctions (Group C) 5.3.1.  OE þa The most important group of connectors in Old English are those high‑frequency items that are, for lack of a better term, called “ambiguous adverb/ conjunctions” (see Mitchell 1985: § 3010). Old English þa ‘then’, for example, is notorious for presenting translation problems (see BT, s. v. þa, Mitchell and Robinson 2001: chapter 4), because it may function as –  a circumstance adverb of time ‘then’, –  a circumstance adverb of space ‘there’, –  an adverbial connector expressing transition, –  a discourse marker ‘what’s more’, –  a simple subordinating conjunction indicating a temporal relation ‘when’ or may be employed as one (or both of) the items in a correlative construction –  þa … þa ‘when … then; then … when’ (for examples, see (49) and (50)).4 4 Obviously, there are still other homonyms, such as the Plural Nom./Acc. of the personal pronoun in the plural. This, however, is not a case of polyfunctionality, but of homonymy.

Ambiguous adverbs/conjunctions (Group C)   65

In Old English narrative texts, þa is often employed, as has been extensively argued, as a pragmatic marker denoting foregrounded action, narrative segmentation, or discourse‑level shifts (see e. g. Enkvist and Wårvik 1987; Kim 1992 and the research summary in Brinton 2006: 314–315). In some narrative texts, þa indeed seems to be a rather elusive marker of the narrative sequence, i. e. a loose connector solely indicating that the narrative is going on. This recurrent use of þa is a shibboleth of Old English narrative prose style, in particular in the sequence þa followed by a verb, which is, by contrast, virtually absent in Old English poetry (Andrew 1940: 3–18). See, for example, the following passages from the Vision of Leofric (COLEOFRI, which is classified as a “religious treatise” in the Helsinki Corpus, but is clearly not argumentative but narrative), in which every single sentence starts with þa (all of them with þaV word order): (49) ða [49a] geseah he ofer þa rode ane hand swylce heo bletsode; þa [49b] wende he ærost þæt sum man hine bletsode, for þam seo cyrce wæs eall folces afylled; þa [49c] næs þæt na swa. ða [49d] beheold he hit þa [49e] gyt geornor, þa [49f] geseah he ealle þa rode swa swutole, swylce þær nan þing beforan nære, & wæs seo bletsiende hand styriende & wendende upward. þa [49g] forhtode he & tweonode him hweðer hit swa wære, swa him þuhte. ða [49h] mid þæs modes tweonunge þa [49i; adverb] æteowde heo him swa swutole …(COLEOFR, p. 84) ‘Then (When?) [49a; adverb ‘then’ or subordinator ‘when, as soon as’ in correlative construction] he saw over the cross a hand as if it blessed; (then) [49b; adverb ‘then’ (?second connector in the correlative construction 49a–b] he immediately went in order to find someone to bless him, because the church was full with people; then/there [49c; adverb] was none there. Then, when [49d; adverb ‘then’ or subordinator ‘when’ in correlative construction] he studied it [the cross] then [49e; temporal adverb] even more intensely, he then [49f; adverb ‘then’ (?second connector in the correlative construction 49d–e] saw the complete cross so clearly as if there had never been anything before it, and the blessing hand was directing and going upwards. Then [49g; adverb] he was frightened and he doubted whether it was so as it seemed to him. Then [49h; adverb] against the doubt of the mind then [49i; adverb] it [the cross] appeared so clearly … ’. Similar text passages are found in almost all pieces of narrative prose in Late Old English (for the implications of this finding for quantitative studies, see below), just as in typologically Old English texts which were copied in

66   Connectors in Old English the Early Middle English period (ca. 1175), such as the History of the Holy Rood Tree. Again, all of the instances have sentence‑initial þa with þaV word order: (50) ða andswarede him dauid. & cwæð … þa cwæð þe engel to him … þa dyde he swa ðe engel him bed. ða andswærode him moyses. … ða andswarede him dauid & cwæð … (CMROOD, p. 6) ‘Then David answered him and said … then the angel said to him … then he did as the angel had commanded; Then Moses answered him … Then David answered him and said …’. The collection of corpus instances in Appendix B.5.1 (s. v. þa) shows that transi­tional þa is frequent in a wide variety of texts in Old English, but is only attested until the first half of the Middle English period (ME1 and ME2; in ME2 only in one text, the Kentish Sermons, ca.  1275). This means that the most obvious, recurrent pattern of Old English narrative structure more or less suddenly disappeared in the middle of the thirteenth century. 5.3.2. OE þeah, swaþeah, (swa)þeahhwæþere – contrast/concession While there are no reliable criteria, such as, for instance, constituent order, for differentiating the adverbial and the subordinator use in the cases of additive, transitional and causal adverbial connectors or subordinators (for details on the causals, see Chapter 9), this is different for connectors marking the relation contrast/concession. Here, the conjunction þeah categorically governs the subjunctive mood, whereas in the adverbial use of þeah the verb takes the indicative (see Quirk 1954; Mitchell 1985: § 3386–3429). This differentiation remains stable until the eventual loss of the subjective mood in Early Modern English (see Molencki 1997b and Moessner 2006). Mitchell (1985: § 3399) says that “there is rarely any difficulty”. Example (51) illustrates the uses of þeah in the Old English adaptation of Boethius’ Consolatio in King Alfred’s Circle. (51a) and (51b) form a correlative construction where the subordinator þeah (51a; subjunctive ongite) introduces the topic‑forming pre‑posed clause, and the adverbial connector þeah (51b; indicative wolde) strengthens the concessive interpretation in the second connect. In (51c), the subordinator þeah (subjunctive todælen) introduces a post‑posed concessive clause (for further examples, see below, Chapter 10.5.2):

Position of adverbial connectors   67

(51) ða andsworede se Wisdom & cwæð : Hwæþer þu nu fullice ongite for­ hwy hit þonne swa sie? ða andswarede ic & cwæð: þeah [51a] ic his nu hwæthwugu ongite [1 Prs. Sg. Subj.], ic wolde hit þeah [51b] fullicor & openlicor of ðe ongitan. ða andsworode se Wisdom & cwæð: Genog sweotol hit is þætte God is anfeald & untodælendlic, þeah [51c] hine dysige men on mænig todælen [3 Prs. Pl. Subj.], þonne hi dwoliende secað þæt hehste god on ða sæmran gesceafta (COBOETH, 33.74.26–75.1) ‘Wisdom then answered and said: Do you now fully understand why it is so? Then I answered and said: Although [51a] I understand it now in some measure, I would nevertheless [51b] learn it more fully, and more distinctly from you. Then Wisdom answered and said: It is sufficiently clear that God is single and inseparable, though [51c] foolish men divide it into many, when/if they, erring, seek the highest good in the worse creatures’. As is common for connectors expressing the semantic relation contrast/concession, the adverbial connector þeah expresses a contrastive rather than a purely concessive meaning of counter‑expectancy. This is induced by the equal weight of the propositions of the two connects in such a paratactic construction (see below, Chapter 10). 5.4.  Position of adverbial connectors 5.4.1.  Nacherstposition ‘post‑first‑position’ Examples such as (51) show that so far we have not discussed one essential distinguishing criterion of adverbial connectors and conjunctions, namely their position in the sentence. One of the defining topological properties of conjunctions in contrast to adverbial connectors is that they are obligatorily placed at the beginning of their respective connects (though sometimes in collocations following coordinating conjunctions, such as and (g)if or and though); adverbial connectors, on the other hand, may be placed in a variety of positions (see above, Chapter 2.3). This criterion establishes þeah in (51b) clearly as an adverbial connector. With respect to adverbials in general (i. e. not only adverbial connectors, but also circumstance and linking adverbials), there has not been much change in the history of English. Old English grammars generally state that “the free variation available to Ælfric in the position of the adverbs is available today likewise. So much depends on the writer’s purpose” (Mitchell

68   Connectors in Old English 1985: § 1592–1593). This means that the distribution of adverbials found for Present Day English (see Chapter 3.5.4) is similar in Old English, at least as far as the position of circumstance adverbials is concerned. There have, however, been drastic changes in the position of linking adverbials, since these are – from the Late Modern English period onwards – increasingly found in connect‑medial and connect‑final positions (for the details, see below, Chapters 10.5 and 13). To understand these changes, it is necessary to comment on one specific clause‑internal position of adverbial connectors, namely the so‑called “Nacherstposition” (Pasch et al. 2003: 71). For lack of a better expression, the present study uses the loan‑translation ‘post‑first‑position’ for this position of an item after a sentence constituent which may fill the pre‑field on its own.5 This position of adverbial connectors has, as far as I am aware, not been discussed in the literature on English (neither historically nor synchronically) in any detail until recently. In 2005 and 2006, van Kemenade has (with various collaborators) presented a number of papers on one aspect of this position, namely the use of the adverbs þa/þonne as what these linguists call “focus particles” (the only publication available so far is van Kemenade and Los 2006b). While I agree with the central claim for Old English, namely that in its clause‑internal use þa and þonne (but not only these!) are morpho‑syntactic markers which separate the topic domain (i. e. the given information) of the clause from the focus domain (i. e. the new information), my findings differ with respect to the implications for the later periods of English, in which this possibility of position has certainly not, as van Kemenade claims, disappeared, but is still very much alive in the same functions as in Old English (it is only OE þa which is no longer used after the middle of the thirteenth century; see above, Chapter 5.3.1). Because of these different views, I will first of all introduce the discourse function of this position of adverbial connectors with Present Day German examples. 5.4.2.  Present Day German adverbial connectors in post‑first‑position While most of the adverbial connectors in German may be placed rather freely between the two connects (“Nullposition”) or in the pre‑field and 5 See also the refined definition in grammis (s. v. Nacherstposition): “Als Nacherstposition bezeichnen wir eine topologische Position im Vorfeld nach der ersten Konstituente und vor der linken Satzklammer, die von bestimmten Adverbkonnektoren besetzt werden kann”.

Position of adverbial connectors   69

middle field, their use in post‑first‑position is highly dependent on the context. In grammis (s. v. “Zur Nacherstposition bei nicht positionsbeschränkten Adverbkonnektoren”), the discourse restrictions are illustrated by the following examples: (52) Rehmann will sich nicht länger den Schwarzen Peter zuschieben lassen. Er packte also aus. [Mannheimer Morgen, 30.5.1995] ‘Rehmann does no longer want to be blamed. So / therefore he blew the whistle’. (52a) Rehmann will sich nicht länger den Schwarzen Peter zuschieben lassen. Also packte er aus. [Nullposition; ‘zero position’] (52b) Rehmann will sich nicht länger den Schwarzen Peter zuschieben lassen. *Er also packte aus. [post‑first‑position impossible] (52c) Rehmann war der einzige, der zu einer Aussage bereit war. Er also packte aus [post‑first‑position possible]. The comparison of examples (52b) and (52c) shows that a post‑first‑position of the adverbial connector predominantly encodes discourse relations: it is only possible if the element in the first position is in the (contrastive) focus (52c). In co‑texts such as (52b), where the pronoun er is purely anaphoric, but not in the contrastive focus, the post‑first‑position of the adverbial connector is not allowed. This contrastive function of adverbial connectors in post‑first‑position also explains why most of the adverbial connectors which may take this position express the semantic relations contrast (e. g. PDG aber, allerdings, dagegen, hingegen) or result (e. g. PDG also, schließlich). (53) Die weiblichen Selbstoffenbarungen, so mutmaßt Kotthoff, fördern den Gemeinsinn. Männer hingegen meiden solche Entblößungen. ‘Female self‑relevations, Kotthoff surmises, endorse the public spirit. Men, on the other hand, avoid such exposures’. (54) Allein im Großraum Hannover hat sich die Zahl der Szene‑Projekte innerhalb von 3 Jahren verdoppelt. In der Nürnberger Region allerdings konnte sie nicht zulegen. ‘Only in the Greater Hanover area the number of scene‑projects has doubled in three years. In the Nuremberg area, on the other hand, there has been no increase’.

70   Connectors in Old English 5.4.3.  Post‑first‑position of adverbial connectors in Old English As has been mentioned above, the functions and implications of adverbial connectors in post‑first‑position have only rarely been discussed in the literature. In his chapter on “þær‑ and þonne‑sentences”, Andrew, for instance, says: “Here we also simplify our problem by excluding sentences in which þonne (adv.) is not the head word” (Andrew 1940: 21). In some more detail, Mitchell (1985) comments on the diverse uses of þa and þonne as conjunctions and adverbs: It is a syntactical commonplace that, whereas þa conjunction is the equivalent of Modern German als, being used only with the preterite indicative of a completed act in the past or of a series now regarded as a single act, þonne is the equivalent of wenn, being used frequentatively in the past, present and future, and of a single act yet to be completed at some indefinite time … (Mitchell 1985: § 2562).

In his chapters on the adverbs, he then differentiates between the purely temporal and the resultive use of þonne in particular. … . The distinction between the conjunctions þa and þonne set out in § 2562 seems to hold generally for the adverbs when they are used in a purely temporal sense … When used in other than purely temporal senses, þonne adv. can be accompanied by a past tense …. where it is not only correlative with frequentative þonne, but also marks a stage in the narrative … where it implies a qualification or contrast (Mitchell 1985: § 1116–1117).

Mitchell (1985) does not comment on the constituent order of the respective uses, but it becomes clear from my corpus texts that there is a high correlation of the use of resultive þonne in post‑first‑position (not necessarily with the verb in the past tense) and its inferential use: (55) Ac þæt is openlice cuð þæt sio godcunde foretiohhung is anfeald & unandwendlic, & welt ælces þinges endebyrdlice, … Sumu þing þonne on þisse weorulde sint underðied þære wyrde, sume hire nanwuht underðied ne sint (COBOETH, 39.129.7–12) ‘But it is evidently known that the divine predestination is simple, and unchangeable, and governs everything according to order … Some things, therefore, in this world are subject to fate, others are not at all subject to it’. Often, þonne is found with imperatives, such as in (56) Sumra wyrta oððe sumes wuda eard bið on dunum, sumra on merscum, sumra on morum, sumra on cludum, sumra on barum sondum. Nim

Position of adverbial connectors   71



ðonne swa wuda swa wyrt, swa hwæðer swa þu wille, of þære stowe þe his eard & æþelo bið … (COBOETH, 34.91.16–19). ‘Of some herbs, or some wood, the native soil is on hills, of some in marshes, of some on moors, of some on rocks, of some on bare sands. Take, therefore, tree or herb, whichever you will, from the place which is its native soil …’.

This use of þonne is not only found in translations from Latin, where it may have been triggered by Latin enim and nam, which are commonly used in this position, but also in more independent texts such as the Blickling Homilies: (57) þa halgan ær Cristes cyme on hine gelyfdon, & hine lufodan, … & mid his þrowunga hie wurdan alesde of helle wite, & mid his æriste gehælde. We þonne synt þe þær æfter fylgeaþ; … (COBLICK, Homily 2, 285–288). ‘The saints believed in him before he had come and loved him … and through his sufferings they were saved from hell and (he) healed them by his resurrections. We, then, are those who will follow afterwards …’. In Old English, this position is widely attested for a number of adverbial connectors (fourteen out of thirty‑two adverbial connectors; see the sub‑entries “post‑first‑position” in Appendix B for the respective connectors). As in Present Day German, it is predominantly found with resultive (see þonne) or contrastive adverbial connectors, such as swa ðeah in (58) or þeah in (59b): (58) ætforan ælcum cwyde we setton ða swutelunge on leden. mæg swa ðeah se ðe wile þa capitulas æfter ðære forespræce geendebyrdian (COAEPREF, R 2.41) ‘In front of each homily we have put the argument in Latin; everyone who wants to, however, may order the rubrics after the preface’. (59) & ðeah [59a; Conj.] þæt cild for geogoðe sprecan ne mage þonne hit man fullað, his freonda forespræc forstent him eal þæt ylce þe hit sylf spræce. Is þeah [59b] ma manna þonne þearf wære þe þises behates gescad ne cunnan (COWULF3, p. 227–228). ‘And although [59a] the child cannot speak because of its youth when it is baptized, his friends’ intercession is equivalent to all that it himself speaks. There are, however [59b], more men than would be necessary who do not know the meaning of this vow’.

72   Connectors in Old English There can be no doubt that this post‑first‑position of adverbial connectors is triggered by discourse considerations. As has been mentioned above, however, van Kemenade (2006b) goes one step further and argues the cue status of adverbs such as þa and þonne in clause structure (see the summary in van Kemenade and Los 2006b: 231). Her main observation is that subject personal pronouns appear on the left of þa or þonne (object pronouns do so optionally), whereas nominal elements, including subjects, occur on the right. Other restrictions are that definite subjects are used on the left of þa or þonne, and that indefinite subjects are quite rare there (van Kemenade and Los 2006b: 232). In Old English, the crucial clues are thus definiteness/anaphoricity, contrastive topic status and prominence of discourse referent. In van Kemenade’s view, the transition to Middle English marks a sharp contrast because the use of þa and þonne in clause‑internal position becomes a good deal less frequent in absolute numbers and this, according to van Kemenade and Los (2006b), indicates that þa/þonne were in the process of losing their discourse‑marking properties, in particular because definiteness/anaphoricity and contrastive topic status are thought to no longer provide the crucial clue for appearance in this area of discourse. As the central implication, they find that the syntactic organization of the clause, at least in Old English, is interwoven with discourse organization much more closely than has been thought so far, and that the transition to Middle English is one that results in a more strictly syntactic organization of the clause (van Kemenade and Los 2006b: 224).

In view of the findings of the present study, however, we see that this pattern did not die out at the end of the Old English period (see, as a first indicator, its existence in the translations of the example sentences above): post‑first positions have been common in all periods of English in this specific, focussing discourse function, in particular for resultive then (see Appendix B.3.2; see, for example, also the uses of therefore, which was only coined after the Old English period). Considering all kinds of adverbial connectors, major changes in frequency and variability have, by contrast, occurred with respect to the medial and final position of adverbial connectors from the Late Modern English period onwards (see below, Chapter 13). 5.5.  Pronominal connectors 5.5.1.  Pronominal connectors in Old English The connectors discussed so far have mainly been single lexical items which are used as explicit markers of sentence or discourse connection, i. e. are situ-

Pronominal connectors   73

ated on levels III (“Explizit verknüpfte Hauptsätze” ‘Explicitly linked main clauses’) or IV (“Verknüpfung durch subordinierende Konjunktionen” ‘Linkage by subordinating conjunctions’) of Raible’s typological continuum of connectors (Raible 1992). In Raible’s continuum, which is based on comparative synchronic data of the Present Day Romance languages but also takes the diachronic perspective from Latin to today’s French‑based creoles into account, there is a phase II – situated between asyndesis (I) and the explicit paratactic connection (III) – which is important for the history of English connectors. This phase is called “Junktion durch Wiederaufnahme (eines Teils) der vorhergehenden Sätze” and singles out anaphoric connectors which comprise an explicitly text deictic, pronominal element, such as OE dative þæm in forþæm ‘therefore – because’ and instrumental þon or þy in forþon or forþy ‘therefore – because’ (cf. Present Day German forms with genitive des of the article in des‑wegen ‘therefore; lit. this-for’, des‑halb ‘therefore; lit. this-for’ or dative dem of the article in dem‑nach ‘therefore; lit. this-after’). These Old English pronominal connectors (Group D), like the use of case forms of the demonstratives in resultive þæs and þy ‘therefore’, will be discussed in more detail below in Chapter 9. For a better understanding of the loss of polyfunctionality and the diachrony of adverbial connectors in general, it is, however, helpful to follow the developments in the Romance languages, from Latin to French‑based creoles. 5.5.2.  Pronominal connectors in the history of the Romance languages While in Present Day French these anaphoric connectors mostly comprise a form of the article or demonstrative in a full nominal phrase – as, for instance, à cause de cela ‘because of this’, c’est pourquoi ‘this is why’ or pour cette raison ‘for this reason’ – Raible shows in his diachronic chapter (Chapter IV; Raible 1992: 154–190) that many originally anaphoric forms have developed into opaque conjunctions (“2. Von der pronominalen Reprise oder Vorwegnahme und vom relativen Anschluss zu Konjunktionen (Latein, romanische Sprachen)”; Raible 1992: 160–170). In Latin, these forms are manifold and allow anaphoric as well as cataphoric relations. For the comparison with the development of the causal connectors in English, I will here illustrate the system by causal pronominal connectors in Latin (the respective pronominal element is underlined): for the semantic relation cause – result, there are anaphoric adverbial connectors such as propterea, eapropter, propter id/hoc, quapropter, ob id/hoc, pro eo, or hac gratia ‘therefore’ (for the full list, see Raible 1992: 163–164). These forms can

74   Connectors in Old English also be used cataphorically in pre‑posed clauses in collocations with quod, i. e. propter ea quod, eapropter quod ‘because’. The forms comprising quod are then differentiated and yield the widely used forms of conjunctions with a second element que in the Romance languages, such as French puisque, parce que, avant que or Spanish pues que, porque; there, que functions as a general subordinating particle similar to PDG dass, OE þe or ME and PDE that (cf. now vs. now that). Two subtypes are to be distinguished: (a) correlative constructions such as French par ce que (with demonstrative ce originally in the first connect), and (b) pour que without a correlative (for a detailed analysis see Stempel 1964, Krefeld 1989). With loss of transparency, conjunctions may move from type (a) to type (b) when the originally anaphoric element is fused, as in Rhaeto‑Romanic cura cha ‘as’ which has developed out of qua hora cha ‘the hour that’. 5.5.3.  Connectors in Louisiana French and French‑based creoles For a better understanding of the processes described in the following chapters for the history of English adverbial connectors, it is illuminating to follow the path of these transparent conjunctions, which are structurally identical to, for example, OE for þæm þe (cf. French pour ce que), in français cadien (Louisiana French) and in French‑based creoles of the Caribbean. With respect to the subordinating particle, French has a tendency to obligatorily use the subordinating particle que, whereas Louisiana French shows exactly the diverging tendency: the subordinating particle que is optionally deleted so that parce que appears as parce (Raible 1992: 167–169, 200–202). This has drastic consequences for the hierarchy of connectors. Da es im français cadien nicht, wie im Deutschen, eine besondere “Nebensatz‑Stellung” gibt, und auch der Modus “Konjunktiv” als Signal der Integration fast ganz verschwunden ist, bedeutet dies im Grunde, auf die kontinental‑franzö­sischen (schriftsprachlichen) Integrations‑Techniken bezogen, ein Zusammenfallen der Techniken III und IV zu einer einzigen (Raible 1992: 200). ‘Since there is, in contrast to German, no specific word order for subordinate clauses in français cadien and since also the “subjunctive” mood as a signal of integration has been lost almost completely, this basically means, when compared to the integration patterns of (written) continental French, that levels III and IV have merged into a single level’.

This tendency is taken even further in some Caribbean French‑based creoles. In contrast to Louisiana French, where the deletion is optional, puisque has there developed into non‑transparent pis; similarly, parce que appears as non‑trans-

Pronominal connectors   75

parent pas (Raible 1992: 201). This means that parce que has lost both of its deictic elements ce and que, a process identical to the development of forþæm þe into for in the history of English. Yet this does not mean that these creoles have only simplified their inventory of connectors, since they at the same time coin new connectors from lexical material or from recurrent syntactic phrases, though only rarely as pronominal connectors (Raible 1992: 201–202).

6.  Adverbial connectors in the history of English

6.1.  Introductory remarks Based on this summary of connectors in Old English and on the corpus findings for contemporary English (Chapter 3.5), the history of the adverbial connectors in general will now be analysed from various perspectives, starting with a survey of the connectors attested in the periods from Old English onwards. In all of these sections, the findings will be compared with the relations to and the developments of the other sentence and discourse connectors, i. e. of coordinating and subordinating conjunctions. 6.2. The diachrony of coordinating and subordinating conjunctions: general tendencies In contrast to subordinators and adverbial connectors, the number and also the functions of the coordinating conjunctions relevant for the present study, i. e. those working on the sentence or discourse level, have remained fairly stable in the history of English. With the exception of the emergence and disappearance of “recursive” for as a rather loose additive connector (discussed below in Chapter 9), the only major change is the replacement of lexical material from the Early Middle English to the Late Middle period in the cases of ac by but and oþþe by the reduced form or:1 – additive relation (Appendix B.1.5): – contrastive relation (Appendix B.4.4): – reformulatory relation (Appendix B.4.4)

OE to PDE: and ME to LModE: for OE to ME2: ac; from ME3: but OE oþþe; from ME3 reduced form or

The scenario is rather different for subordinators and adverbial connectors. The diachrony of English subordinators has been analyzed in a “grammar‑cum‑dictionary‑method” (Kortmann 1997: 53) by Kortmann in the wider context of the typologies of adverbial subordinators (Kortmann 1997: 291–335). This study has yielded the following general tendencies, which are used as a contrastive 1 The OED sees øþer, the full form of reduced or, as a variant of oþþe, oþþa; see OED, s. v. other.

Introductory remarks   77

plane for the present investigation of the development of adverbial connectors. Over the periods, Kortmann finds – – – –

an increase in the number of subordinators, a decrease in their syntactic polyfunctionality, a decrease in their semantic polyfunctionality and an increase in their morphological complexity.

With regard to the individual semantic relations he finds that – subordinators for the relation contrast/concession are coined much later than for other relations and – that the inventory of concessive subordinators is least stable over time. As for the respective periods (Kortmann distinguishes between OE, ME, ­EModE and PDE only, i. e. does not have a separate category for Late Modern English), Middle English emerges as the crucial period for shaping the inventory of Present Day English subordinators. A huge number of subordinators were also coined in the Early Modern English period, but most of these innovations were ephemeral. Since most of these Early Modern English coinages have not survived into Present Day English, this period – despite its high number of new subordinators – has had only a weak impact on Present Day English (for the details, see Kortmann 1997: 334–335 and below, Figure 6.3). 6.3.  The diachrony of adverbial connectors: general tendencies In a first step, these general tendencies found by Kortmann (1997) for subordinators will now be tested for the adverbial connectors. In spite of the fuzziness of the category of adverb, the polyfunctionality or heterosemy of lexemes and the gradient character of adverbial subclasses, it is still possible to compile a rather full list of adverbial connectors for the respective sub‑periods in the history of English (with, of course, the exception of the enumerative adverbs, such as fourthly, fifthly, etc.; for this open class, see above, Chapter 3.3). This inventory of English adverbial connectors is summarized in the synopsis of the present study – Appendix A – which lists all single adverbs as well as lexicalized verbal and prepositional phrases which have been used as adverbial connectors in the various periods of English (i. e. only fully transparent phrases such as for that reason or on the contrary are excluded).2 2 For the principles of exclusion, i. e. the properties a prepositional phrase has to fulfil in order to be included, see below, Chapter 7.4.

78   Adverbial connectors in the history of English The present qualitative and quantitative analysis starts with a very general description and evaluation of the developments in the field of English adverbial connectors over the periods, which allows a first comparison of the similarities and divergences in the development of adverbial connectors and subordinators. In the following chapters, these findings are then examined from various perspectives in order to demonstrate and explain the systematic changes in the morphological make‑up and the use of cognitive source domains for the coinage of new adverbial connectors (Chapters 7 and 8). The differing developments in the individual semantic relations then show that changes in the “impure” connectors – the so‑called ccc‑relations – reflect typological changes of English (Chapter 10 on causal and resultive connectors, Chapter 11 on concessive and contrastive connectors), while the changes in the “pure” connectors – various additive and transitional relations – are mainly triggered by rhetorical and pragmatic motivations (see Chapters 11 and 12 on additive and listing connectors and Chapter 8.4 on truth‑intensifiers which are used as transitional connectors). Rhetorical and pragmatic considerations also elicit changes in constituent order, i. e. the increasing use of adverbial connectors in medial position in the written medium and in final position in the spoken medium (discussed in an exemplary analysis of PDE though in Chapter 10.5). These will in turn allow more detailed assessments on the hows and whys of the diachrony of adverbial connectors in the history of English. 6.4. New adverbial connectors in the sub‑periods of English: methodology The diachrony of English adverbial connectors is summarized in Table 6.1, which first of all lists the repertoire of the Old English items which could function as adverbial connectors. As has been shown above, most of these could also be used as subordinating conjunctions, without any change in their form or, very often, also without any change in the form of their connects (for constituent order or the use of the subjunctive mood as distinguishing criteria, see above, Chapter 5.3). Following Mitchell (1985: §  2536, 3010), these items are called “ambiguous adverb/conjunction” and are marked by the symbol °. In chronological order, Table 6.1 then presents all the items which have emerged as new adverbial connectors in the respective sub‑periods of English. In the left column, only those innovations are listed which have survived until Present Day English. The adverbial connectors in the right column died out

New adverbial connectors in the sub‑periods of English: methodology   79

in some period before Present Day English (e. g. herefore, forwhi, overmore) or have lost their capacity to function as adverbial connectors (e. g. albeit or wherefore).3 The sections after the Old English period thus only list the new adverbial connectors, i. e. those which emerged in the respective period in this function. The period ME1/2 (1150–1350) shall serve to illustrate this methodo­ logy: The innovations attested in this period are certain, certainly, certes, for, forsoothe, herefore, for that, °forhwi, more, neverthelatter, on oðer half, oðer side, overall, therefore, yet and wherefore. Some of these are already recorded for Old English (e. g. giet ‘yet’, °forhwi ‘forwhy’), but had not been used as adverbial connectors before. OE giet ‘yet’, for example, was predominantly used as a temporal adverb in Old English (‘still, yet’; see BT, s.vv. git, gita, and CH, s. v. gīet), a use which has survived into Present Day English (“implying continuance from a previous time up to and at the present time”: see OED, s. v. yet, 2.5).4 From Early Middle English onwards (see the date 1205 for its first occurrence in this function in Appendix A), it has – by a conventionalization of conversational implicatures common for temporal adverbs expressing concomitance (cf. PDE still and see below, Chapter 10.5.3.2) – acquired contrastive/concessive meanings and is used as an adverbial connector ‘nevertheless’. Since it could also be used as a subordinator yet (that) ‘notwithstanding that, although’ during the Middle English period (OED; s. v. yet, III.9b), it is classified as an “ambiguous adverb/conjunction” (indicated by the symbol °). Yet has survived in its adverbial connector function until the Present Day English period, as is attested by the wide variety of texts and collocations with other concessive/contrastive connectors such as but, nevertheless, or still (see in Appendix B.4.1). As a second illustration, compare OE for‑hwy, which was only used as an interrogative adverb “introducing a direct question: for what reason?, why?”, or as a conjunction “introducing a clause in apposition to or exemplifying the statement in the principal clause: (as to) why …, in that, in the fact that” in Old English (DOE, s. v. for‑hwåm, for‑hwon, for‑hwy¯, A.1, B.1). In Early Middle English, it began to be used as a resultive adverbial connector ‘therefore’ (see the date 1225 in Appendix A.1 and the first occurrence in the corpus texts in 3 For the exact periods, see Appendix A.1 and the Index of Adverbial Connectors (Appendix A.2), and also the discussion of the respective semantic relations. 4 Only in the collocation þa giet ‘furthermore’ is it attested as a reinforcing adverbial connector. This meaning “in addition, or in continuation; besides, also; further, furthermore, moreover; with a numeral or the like = ‘more’” is present in, for example, PDE yet another (see OED, s. v. yet, 1.).

80   Adverbial connectors in the history of English Appendix B.3.1 in CMHALI). Appendices A.1 and B also provide evidence that it was used in this adverbial connector function only during the Middle English period, i. e. it is not attested after ME4 (in CMHILTON). The lexicalization of collocations and prepositional phrases in a specific period can be exemplified by ME on oðer half and ME oðer side. While the nouns healf ‘half’ and side ‘side, part’ are found in Old English, they are not recorded in the lexicalized phrases on oðer half or oðer side ‘on the other hand’ as adverbial connectors marking the contrastive/antithetic relation (see Appendix B.4.2).5 For another kind of change in function see ME forsooth, which is employed as a transitional adverbial connector in Middle English (in ME3 and ME4; see Appendix B.1.5), but is only sporadically and jocularly used as an interjection in contemporary English (for a full analysis, see Lenker 2003 and below, Chapter 8.4.5). 6.5.  The inventory of Present Day English connectors: donor periods The data collected in Table 6.1 allow a survey of the long‑time developments of adverbial connection in English and thus a comparison with Kortmann’s data for subordinators (Kortmann 1997). At first glance, two findings are fairly obvious: First, very few (seven out of thirty‑two) of the Old English items have survived. Among them are the still very frequent additive/reinforcing also and the resultive/transitional so and then, but also the fairly marginal else (reformulatory), here and now (both transitional). There is no adverbial connector with an uninterrupted history from Old English for the ccc‑relations cause/result and contrast/concession (for the meandering path of though, which is not attested from Middle English to Late Modern English, see Appendix B.4.1 and below, Chapter 10.5). This means that the system of clause linkage by adverbial connectors was almost completely re‑organized from the Middle English period onwards. Secondly, new adverbial connectors have been coined in all of the periods of English until Present Day English and so – as a general tendency – the absolute number of types has increased over the periods. Table 6.2, which lists the number of adverbial connectors per sub‑period, shows that the number of types has increased from thirty‑two adverbial connectors in Old English (including the ambiguous adverbs/conjunctions) to fifty‑nine in Present Day English. 5 On oðer half is only attested in two texts from the earliest Middle English period (ME1; all of them are texts from the so‑called “Katherine Group” found in Oxford, Bodleian Library, Bodley 34); oðer side re‑appears in Early Modern English in the full phrase on the other side.

Lost on their way to Present Day English OE ærest ‘first’, eac ‘also’, °eac swylce/swylce eac ‘also’, cuþlice ‘therefore’ (lit. ‘certainly’), eft ‘also’, eftsona/eftsones ‘also’, eornostlice (lit. ‘earnestly’; transition), °forþæm ‘therefore’; °forþon ‘therefore’, °forþy ‘therefore’, huru ‘nevertheless’, huruþinga ‘nevertheless’, naþylæs ‘nevertheless’, °hwæt þa ‘what then’ (transition), °hwæðere ‘nevertheless’, soþlice (lit. ‘truly’; transition), swaþeah ‘nevertheless’, swaþeahhwæðre ‘nevertheless’, °þa ‘then’ (transition), þa giet ‘also’, (þær)toeacan ‘also’, þæs (demonstrative; genitive) ‘therefore’, þeahhwæðere / hwæðreþeah ‘nevertheless’, °þy (demonstrative; instrumental) ‘therefore’, gewislice (lit. ‘certainly’; transition), witodlice (lit. ‘truly’; transition) ME 1/2 ME 1/2 certainly, first, °for, at (the) last, more, over‑ certain, certes, forsooth, herefore, iwis, for that, °forhwi, neverthelatter, nought‑for‑than ‘nevertheless’, on oðer half, all, therefore, °yet oðer side, °wherefore ME 3 ME 3 further, farthermore/furthermore, moreover, °algates, ergo, for which, fartherover/furtherover, item, °nought‑for‑that ‘nevertheless’, °nought‑for‑thi ‘nevertheless’, nevertheless, otherwise, thus overmore, over this/that, otherways, sekirly, suingly ‘consequently’, truly, verily ME 4 ME 4 contrariwise, finally after ‘moreover’, °howbeit, notwithstanding EModE1 EModE1 again, consequently, contrarily, likewise, °albeit, in fine, °howsomever, likeways, over and besides, semblably nonetheless, surely, well EModE2 EModE2 above all, besides, however, indeed, at least, beside, contrary, contrariways, °howsoever, lastly, over and above (in the) meantime, (in the) meanwhile, in sum, sure EModE3 EModE3 after all, on the other hand, hence, still thence, videlicet, whence LModE1 LModE1 accordingly, at the same time, too in truth LModE2 altogether, anyhow, conversely, of course, at all events, in fact LModE3 all the same, in any case, anyway, at any rate, in the same way PDE all in all, correspondingly, incidentally, instead, plus, rather, similarly

Survivors to Present Day English OE °eall‑swa ‘also’, elles ‘else’, her ‘here’, °nu ‘now’, °swa ‘so’, °þeah ‘though’, °þonne ‘then’

Table 6.1.:  Adverbial connectors in the history of English

The inventory of Present Day English connectors: donor periods   81

82   Adverbial connectors in the history of English Table 6.2.:  Absolute number of adverbial connectors per sub‑period (types) OE

ME1/2

ME3

ME4

EModE1

EModE2

EModE3

LModE1

LModE2

LModE3

PDE

32

27

36

29

34

45

46

48

52

56

59

A steady increase of adverbial connectors is, however, only attested from the Early Modern English period onwards. During the Middle English period, the absolute number of adverbial connectors did not change considerably, when we compare the beginning (M1/2 twenty‑seven connectors) with the end of the period (ME4 twenty‑nine connectors). There was, however, a sharp increase in the middle of the period (with thirty‑six connectors in ME3). This general tendency of a constant increase of adverbial connectors – though only after the Early Modern English Period – is supported by the analysis of the inventory of adverbial connectors in the selected corpus texts which were chosen for the comparative quantitative analysis (see Appendix C.2). Table 6.3.: Absolute number of adverbial connectors as attested in the corpus texts (types) OE

ME1/2

ME3

ME4

EModE1

EModE2

EModE3

LModE1

LModE2

LModE3

17

18

30

24

33

33

26

31

38

35

This table shows that the authors of the homilies and religious treatises which are investigated as corpus texts use a wider variety of types of adverbial connectors from the Early Modern English period onwards. Yet, the differences between authors of ME3 (e. g. Chaucer, Purvey, the authors of the Wycliffite Sermons) and LModE3 (e. g. Swift, Hume) are not as striking as we might have expected (thirty vs. thirty‑one different connectors). This reflects the fact that adverbial connectors – in particular the so‑called “impure” connectors coding the ccc‑relations – belong to a closed class of items serving a function at the grammatical‑lexical interface. It may, however, also be indicative of the fact that the authors of the respective periods – though there is an increasing variety of connectors to choose from – just like today’s authors prefer certain adverbial connectors over others (in Present Day English, for instance, authors of academic prose prefer either thus or hence for the causal relation; see Biber et al. 1999: 889).

The inventory of Present Day English connectors: donor periods   83

Figure 6.1.: Number of adverbial connectors per sub‑period: absolute number of types in sub‑period and in corpus texts C.2.

The diagram in Figure 6.1, which summarizes the findings sketched above, suggests a continuous, steady path of innovations in the increase of adverbial connectors. When we have a closer look at the donor periods of the inventory of Present Day English connectors, we see that there is indeed no single period which could be said to have supplied the bulk of the adverbial connectors of today’s inventory.

Figure 6.2.:  Donor periods of Present Day English adverbial connectors

These findings are intriguing when we compare the development of adverbial connectors with the diachronies of coordinators and subordinators. As has been pointed out above (Chapter 6.2), there has been – apart from the replacement of lemmas – virtually no change in the inventory of coordinators.

84   Adverbial connectors in the history of English With respect to the subordinators, my findings for adverbial connectors and those by Kortmann (1997) for subordinators are not fully comparable because Kortmann does not distinguish a separate category for Late Modern English. Yet a comparison of the percentages of the donor periods for the connectors forming the Present Day English inventory nonetheless yields some interesting results: Kortmann could show, as has been summarized above, that the crucial period for subordinators is the Middle English period, contributing ca. forty‑seven per cent of the subordinators of Present Day English, as contrasted with ca. twenty‑five per cent from Old English and ca. thirteen per cent from Early Modern English (Kortmann 1997: 334–335). For the adverbial connectors, it is not possible to establish such a single crucial period. The influx of new, surviving connectors seems to be much more stable over time (see the graph in Figure 6.3).

Figure 6.3.: Percentages of donor periods for Present Day English adverbial connectors and subordinators (in per cent of PDE connectors; numbers for subordinators according to Kortmann 1997)

The general line of development is, as was to be expected, parallel for adverbial connectors and subordinators in the Old and Middle English periods, when the originally “ambiguous adverbs/conjunctions” lost their syntactic polyfunctionality and a new set of the respective connectors was required. In particular, there was a need for new kinds of subordinators as well as adverbial connectors when the Old English pronominal connectors comprising a case form of the demonstrative (e. g., the dative þæm) had become opaque (cf., for example, forms such as ME forthen; see MED, s. v. for‑than). The changes in this period are, therefore, mainly qualitative in nature and are thus triggered by the typological

Middle English: a period of experiment   85

changes English underwent after the Old English period (for details, see below, Chapters 7 and 9). Yet, after this time of parallel development lasting until the end of the ­Middle English period, there is a stark difference in the stability of the innovations in the category of subordinators on the one hand and adverbial connectors on the other: the inventory for subordinators was – in addition to the repertoire of coordinating conjunctions (see above, 6.1) – set to almost three quarters (ca. seventy‑three per cent) by the end of the Middle English period. For the relations cause/result, for example, Kortmann states that “[b]y and large, then, the inventory of causal connectors in Present Day English resembles that of (Late) Middle English most closely” (Kortmann 1997: 331). The findings are strikingly different for adverbial connectors: more than sixty per cent of today’s adverbial connectors were only coined in the Early Modern, Late Modern and Present Day English period.

6.6.  Middle English: a period of experiment This picture of a rather steady path in the increase of adverbial connectors, however, changes if we have a closer look at the number of new adverbial connectors coined in the various sub‑periods, in particular those which have survived into Modern English. We then see that there is no steady increase of new coinages, but that there are three notable peak periods before Present Day English: ME 1/2 (1150– 1350), EModE2 (1570–1640) and LModE2 (1780–1850).

Figure 6.4.: Innovations: surviving adverbial connectors per sub‑period (absolute number of innovations)

86   Adverbial connectors in the history of English This finding, i. e. that the history of adverbial connectors is not as steady as the graphs for the donor periods suggested, is supported by a closer investigation of the innovations (irrespective of whether they have survived into Modern English or not) and disappearances per period as presented graphically in Figure 6.5:6

Figure 6.5.: Adverbial connectors: absolute number of new coinages vs. number of last occurrences per sub‑period

This shows that the inventory of adverbial connectors was extremely unstable until the Early Modern English Period (EMod2; 1570–1640). Only from that period onwards do we see a more constant development of new coinages and losses in the respective periods. The Middle English period in particular stands out as period of “trial and error”, so that even without a qualitative analysis (which will be provided in the following chapters) we see a struggle for a new system of connectors. For a last diagram illustrating the instabilities in the Middle English, see Figure 6.6, which compares the absolute number of new coinages per period with those adverbial connectors which have not survived into Present Day English. This diagram again shows that the majority of those adverbial connectors which were coined during the Middle English period have not survived. In sum, we find that there are striking differences in the diachrony of the three main types of connectors used to explicitly signal the textual relations of sentences or discourse: While the inventory of coordinating conjunctions has, in essence, remained stable, there have been drastic changes in the categories of both subordinators and adverbial connectors. 6 The last two periods LModE3 (1850–1920) and PDE (1920– ) have been excluded because it is impossible to make sure that items last attested in corpus texts of LModE3 will not be used again.

Middle English: a period of experiment   87

Figure 6.6.:  Adverbial connectors: absolute number of innovations vs. non‑survivors

As regards subordinators, Middle English emerges as the crucial period: the inventory of subordinators has remained unchanged to a large extent (i. e. almost three quarters) since the end of the Middle English period. Most of the subordinators coined in the Early Modern English period were ephemeral and have not survived into Present Day English, so that this period emerges as a period of “experiment and transition”, which eventually has not had any major impact on the Present Day inventory of subordinators. This scenario and time frame is different from that of adverbial connectors. There is no single period which can be established as crucial for the Present Day English inventory adverbial connectors. There has been a constant coinage of new adverbial connectors in all periods of English, which finally resulted in the Present Day English inventory. A more detailed investigation of new coinages and losses, however, shows that with respect to adverbial connectors it is the Middle English period which emerges as a period of experiment and transition. The next chapters will investigate various factors which help to explain these divergent histories. Chapter 7 will summarize the main developments in the morphology of adverbial connectors from various perspectives: How are the new linking adverbs coined? Which means of word formation are employed and which language material – Germanic or Romance – is used? Is there an underlying, recurrent pattern in these new elements?

7.  Adverbial connectors: morphology

7.1.  The expansion of the English lexicon Since the diagrams showing the donor periods of Present Day English connectors seem to suggest a more or less constant flow of innovations of adverbial connectors over the periods, one might, at first glance, attribute this to the expansion of the English lexicon in general, especially since we are dealing with linguistic elements which are said to be situated at the interface between the lexicon and grammar. It is standard knowledge that – with about 252,000 words or 415,000 words including the complex words (numbers according to the New Shorter Oxford English Dictionary; see Leisi and Mair 1999: 46) – the vocabulary of Modern English is the largest of all European languages. English lost much of its native Germanic vocabulary after the Old English period and borrowed an enormous number of words from French in the Middle English period, in particular in the fourteenth century (my period ME3; see Baugh and Cable 2003: 167–187; Scheler 1977: 52). It then further enriched its vocabulary significantly in the Early Modern English period, mainly from Latin and Greek sources (Baugh and Cable 2003: 214–234). In one of the recently published studies on Late Modern English, Beal shows that in this period – with a peak at the beginning of the nineteenth century (my period LModE2) – new words were coined at a rate not seen since the sixteenth century because of trade, exploration and colonization and in particular technological and scientific innovations (see the large number of, for example, neo‑classical compounds or eponyms; Beal 2004: 14–29). In view of these facts, the increase of linking adverbs after the Old English period seems to agree with what we expect from the development of the English language. At first glance, the innovations in the field of adverbial connectors could be thought to be another instance of just this development of the enlargement of the English lexicon, with peaks in the Middle English, Early Modern English and Late Modern English periods. A closer investigation of the innovations, however, shows that this purely numerical parallelism does not get to the heart of the matter. With the increase in vocabulary sketched above, the English vocabulary became a “mixed vocabulary” (Leisi and Mair 1999: Chapter 3); even in its basic vocabulary as documented in the General Service List, only about half of the types are Germanic (Scheler 1977: 72). The summary account of innovations in the field of adverbial connectors in Table 6.1 above, however, shows that there are – in all of the sub‑periods – only very few loans among the new connectors:

The expansion of the English lexicon   89 Table 7.1.:  New adverbial connectors: loans ME 1/2 ME3 EModE3 PDE

French: certain, certes Latin: ergo, item Latin: videlicet Latin: plus

This means that the scenario sketched above has to be discarded: there are almost no unambiguous loans among the new adverbial connectors. The use of most of the Latin loans is furthermore limited to certain legal or religious (liturgical) registers, as has been shown above for item ‘furthermore’. Item is only used as a reinforcing connector in some handbooks or law texts of the Late Middle and Early Modern English period (CMEQUATO, CMDOCU4, CMREYNES, CMPRIV, CELAW1; cf. Chapter 1.7 and the list of occurrences in Appendix B.1.2). Similarly, additive/equative videlicet ‘to wit’ is only attested in Early Modern English law texts. Moreover, it is often not clear if it is indeed employed on the sentence level signalling a two‑place relation or if it does not only provide an amplification on the phrasal level:1 (60) in the said first recited Act mentioned shall alsoe before such Justice or Justices of the Peace by whom such Oath is to bee given and administred likewise take an Oath (which such Justice or Justices of the Peace is and are hereby impowered to administer) to this Effect, videlicet I A. B. doe sweare That on the Five and twentieth day of December in the Yeare of our Lord One thousand six hundred ninety & five I was actually a Prisoner in the Custody of the Goaler of Keeper of the Prison of C. in the County of D. … (CELAW3, An Act for Relief of Poor Prisoners for Debt or Damages, p. VII, 76). Another reinforcing connector with Latin roots is plus, which was first borrowed as a preposition in Early Modern English and has only acquired a connective function since the 1960s. It is still considered colloquial by the

1 Phrases such as for example, for instance, in other words, to wit have been excluded from the quantitative analyses of the present study because they are predominantly used as appositions on the phrase level (i. e. provide an amplification of one sentence constituent) rather than as connectors on the sentence or discourse level (see also below, Chapter 11).

90   Adverbial connectors: morphology OED (s. v. plus C.2: “colloq. Introducing a clause: and furthermore; and moreover”). Plus ‘furthermore’ and ergo are not attested at all in the corpus texts. Ergo is – according to the OED – used from ca. 1400 to 1780; the four quotes given in the OED are also exclusively from legal and liturgical texts. This implies that item, videlicet and ergo are predominantly used – often in very formulaic contexts – by authors who would also compose similar kinds of texts in Latin (OED, s. v. ergo). With regard to loans from French, we only find transitional certain and certes as unambiguous items borrowed in ME1/2. Both of them are, however, predominantly found in texts translated from a French exemplar (see Appendix B.5.2). Moreover, in most of the occurrences of certain, a connective reading is also not fully unambiguous in examples such as (61), i. e. whether certeyn functions as a content disjunct, modifying the propositional content of the connect, or a non‑propositional adverbial connector: (61) The seconde the semeth is light ynowgh, for thou maist hurte no man but with thy tonge. Certeyn the first is as light as thys, for ther may no matere be of yuel wille wher no couetise is … (CMAELR4, p. 17). In addition to these, we find a number of hybrids, i. e. adverbs formed from French or Latin bases by addition of the native suffix ‑ly, such as correspond‑ ingly, consequently, conversely, equally, finally, incidentally, semblably or similarly (see Table 7.2). All of these are commonly used as circumstance adverbs and acquired their connector value in collocations with a verb of communication (see above, Chapter 3.2.3); with the exception of consequently (see Altenberg 1984: 45)2 and finally (see Appendix B.1.1), most of these are used very sparingly. Both semblably (CEEDUC1A) and correspondingly (FROWN‑D), for instance, are only used once (see Appendix B1.3). The only hybrids which can be said to belong to the core group of adverbial connectors are the contrastive, antithetic contrariwise and contraryways (first appearing in the phrase in contrary wise in ME4; for the lexicalization of the suffix ‑wise/‑ways, see OED, s. v. ‑ways and Lenker 2002). Contariwise is frequent in Early Modern English (see Appendix B.4.2).

2 In the material collected from the LOB and LLC corpora, Altenberg (1984: 40) finds altogether nine instances of consequently and two of accordingly (in contrast to 265 of so and sixty‑two of therefore).

The expansion of the English lexicon   91 Table 7.2.:  New adverbial connectors: hybrids ME3 ME4 EModE1 EModE2 LModE1 LModE2 PDE

verily finally, contrariwise consequently, semablably, surely contrariways accordingly conversely correspondingly, incidentally, similarly

In sum, we see that the borrowed and hybrid items certainly do not form the core group of adverbial connectors of Present Day English. They also stand out structurally, mainly because of their length (correspondingly, incidentally and similarly). Most of the new elements which emerge in a new function as adverbial connectors in the Early and Late Modern English period are formed with Germanic word material. The main categorial sources are prepositional or other phrases (cf. lexicalized prepositional phrases such as above all or after all or coinages such as all in all, anyhow or anyway; see Table 7.3). This means that the increase in the number of items is not primarily part of the general expansion of the English vocabulary in the Middle English, Early Modern English or Late Modern English periods, which is predominantly attributable to loans from French, Latin or Greek. This is particularly intriguing when we compare these findings to those for the subordinators. In the field of subordinators, borrowed items are much more frequent from the Middle English period onwards, in particular in Early Modern English, the “period of experiment” for the subordinators (see Kortmann 1997: 300, 304–305). The higher number of loans in Early Modern English subordinators is mainly due to the use of present or past participles, often in complex subordinators with the general subordinating particle that. Most of these are borrowed from or at least modelled on the Romance patterns:3 considering (that), conditioned (that), excepting (that), provided (that), providing (that), supposing (that) during (that), granted (that).

Although many of these subordinators are ephemeral and not very frequent (Kortmann 1997: 309), it is still striking that the subordinators, which are 3 This pattern was also transferred to native material, such as the calques outtaken (that), notwithstanding (that) or seeing (that).

92   Adverbial connectors: morphology commonly classified as grammatical elements, mirror the general tendencies of the expansion of the English lexicon, i. e. the high number of borrowed material from Latin in the Early Modern English period, to a larger extent than the adverbial connectors, which are situated at the interface between the lexicon and grammar. This is even more intriguing in view of the fact that the inventory of Present Day English adverbial connectors was expanded considerably in the Early and Late Modern English period (in contrast to the repertoire of subordinators, which has remained unchanged to almost three quarters since the end of Middle English) and that adverbial connectors are particularly frequent in the very formal registers of academic prose (see above, Chapter 3.5.5), where we would have expected a substantial impact of the classical languages. 7.2.  Major morphological changes To gain a better understanding of these processes, the various source categories of adverbial connectors are summarized in Tables 7.3 and 7.4, which list all Old English adverbial connectors as well as the new coinages in the respective periods. Table 7.3 provides a survey of the adverbial connectors according to their patterns of word formation and distinguishes simple adverbs, compound adverbs and derivations. Table 7.4 then lists the complex formations and differentiates between pronominal connectors – which may be realized as case forms of a pronoun or may comprise such a case form (see above Chapter 5.5) – and lexicalized prepositional (or other) phrases. Table 7.3.:  Adverbial connectors: source categories Simple Adverbs

Case Forms

Compounds

Derivations

elles eftsones

eftsona huruþinga þærto‑eacan

cuþlice eornostlice soþlice gewislice witodlice

OE eac eft her huru nu swa þa þeah þonne Superlative: ærest

Major morphological changes   93 Simple Adverbs

Case Forms

Compounds

Derivations

ME 1/2 certainly

certain certes for iwis yet Comparative: more ME 3 ergo item Comparative: further

furthermore furtherover moreover overmore

algates sekirly truly verily

ME 4 Comparative:

finally

EModE 1

well

?howsomever

consequently semblably surely likeways likewise

however ?howsoever

besides contraryways

EModE 2

sure EModE3 still videlicet

hence thence whence

LModE1 too

accordingly

LModE2 conversely LModE3 PDE plus Comparative: rather

correspondingly incidentally similarly

94   Adverbial connectors: morphology 7.3.  Simple adverbs, compounds and derivations 7.3.1.  Simple adverbs With respect to the adverbs proper, the inventory of Old English and early Middle English adverbial connectors comprises a fairly large number of morphologically simple, i. e. so‑called “original” adverbs. The items listed in the left column of Table 7.3 show that this is not a recurrent pattern in the later periods of English: it is only attested in contrastive/concessive still and additive too (which is, however, an exception because it cannot be placed sentence‑initially). We might also add the univerbated additive/contrastive again (OE on‑gean from on-gegn ‘lit. on-straight’; cf. PDG ent‑gegen ‘facing locally’; see OED, s. v. again) and beside (OE be sidan [dative singular] ‘by the side’). The origin of these items in a prefixation and a prepositional phrase are, however, no longer fully transparent by the time they are first used as adverbial connectors in EModE1 and EModE2 respectively (signalled by < > in Table 7.3). 7.3.2.  Compounds Only a very limited use has been made of composition. In Old English, there are only three compounds, two of which express the relation of reinforcement ‘furthermore’: eftsona/-es (cf. eft ‘again, furthermore’ + sona ‘immediately’) and þærtoeacan (cf. þærto ‘thereto’ + eacan ‘as well’). The only period which shows a marked impact of a specific kind of compounds is ME3, where we find the new coinages furthermore, furtherover, more‑ over and overmore (all of them additive/reinforcing connectors). These forms are peculiar in that they use a double comparative: the comparatives more (corresponding to senses of the positive adjectives great, much, and many) or over (local ‘upper, higher in position’) are added to bases which are already comparative themselves. Farthermore and furthermore, for instance, are formed from the spatial comparative further/farther, which may, however, also be used as an adverbial connector by itself (see OED, s. v. further 1. “more forward; to or at a more advanced point. a. in space, or in a course of procedure or development …”). This – in this case double – pleonasm is particularly evident when we compare the simple comparative more (as used from ME1/2) with the forms moreover and overmore (both coined in ME3) which ‘swap over’ their bases. This pattern of forming adverbs and adjectives of place (rarely of time) in the comparative degree with more is fully transparent but no longer productive in later periods of English. The OED considers an analogy with or calque on

Simple adverbs, compounds and derivations   95

North Germanic formations probable, since the earliest instances in English occur in thirteenth century sources and since similar formations consisting of a comparative and meir occur in North Germanic (see OED, s. v. ‑more suffix). The basic pattern, however, i. e. the use of temporal or spatial adverbs in the comparative, is a native pattern of Old English, too (cf. after, rather), and is also attested widely cross‑linguistically (see Chapter 8.3 on cognitive source domains). The high number of compound adverbs in the “pure” semantic relation of addition is, however, certainly no coincidence, since they iconically – by their length – highlight that the author wants to make the connection between the connects explicit (see below, Chapter 11). 7.3.3.  Derivations: adverbs in ‑lice, ‑es and ‑ways/-wise The other morphological forms of adverbs functioning as adverbial connectors reflect the increasing tendency of English to distinguish adverbs from adjectives by various suffixes, in particular by the suffixes – -ly (from OE ‑lice) by reanalysis of adjectival ‑lic and adverbial ‑e (see McIntonsh 1991, Nevalainen 1997; this reanalysis must have taken place in the pre‑literary period, since suffixation in ‑lice is already attested in very early glossaries, such as Épinal‑Erfurt; see Sauer 2006: 265), – -es; originally a nominal or adjectival case form – genitive singular masculine/neuter – which is already in Old English transferred to feminine nouns, as in OE nihtes (cf. OE feminine niht; see also PDG ‘nachts’), and then re‑analysed as an adverbial suffix (in besides, hence, thence, whence), – -ways/‑wise from original prepositional phrases such as in a contrary/like wise yielding contrariwise/‑ways or likewise/‑ways (see OED, s. v. ‑ways and Lenker 2002 on ‑wise). In all periods, new adverbial connectors are coined from derived circumstance adverbs in OE ‑lice (added to adjectives which do not end in ‑lic) and later ‑ly. Yet in Old English the semantic range of adverbial connectors in ‑lice is very restricted. Virtually all of the adverbs are “truth‑intensifying adverbs” which may be used in connective function for the expression of the marginal connector relation of transition (cuþlice ‘certainly’, eornostlice ‘earnestly’, soþlice ‘truly’, gewislice ‘certainly’ and witodlice ‘truly’; see below, Chapter 8.5); only cuþlice may also code the resultive relation ‘therefore’ (see OED, s. v. 3). This semantic restriction of adverbial connectors in ‑ly remained fairly unchanged until the Early Modern English period (cf. certainly, sekirly, truly, verily, sure‑ ly). Only from that time onwards are original circumstance adverbs in ‑ly also

96   Adverbial connectors: morphology used for the relations of result (e. g. consequently, accordingly) or contrast (conversely). In sum, we can say that derived adverbs in -ly have never played a central part as adverbial connectors. They are predominantly found in the field of “pure” (i. e. additive or transitional) connectors, but only rarely in the more complex “impure” connectors marking the ccc‑relations cause, concession and contrast. This is different with adverbs formed in ‑es, which are attested – as a case form – from OE reformulatory elles ‘else’ (genitive singular neuter). In the thirteenth century, the ending ‑es is by analogy appended to the ambiguous adverb/ conjunction algate ‘however; nevertheless’, a loan from Old Norse (ON alla götu; cf. OED, s. v. algate). Similarly, it is used in besides (cf. OE be sidan) and the originally spatial/directional and then resultive hence (OE heonan/heonon with the directional suffix ‑an ‘hence; from there’) as well as the parallel thence and whence (which may also be used as adverbial connectors). Adverbs in ‑es are thus used predominantly for the “impure” ccc‑relations: they are more specific because they are not derived from adjectives, but have their origins in, originally spatial, adverbs or (prepositional) phrases (cf. besides; algates). 7.4.  Pronominal connectors – lexicalized phrases Simple, compound and derived adverbs have thus been employed as adverbial connectors in all periods of English. Yet, other formations such as pronominal connectors and lexicalized phrases (listed in Table 7.4) will be shown to have been much more important since they are more indicative of the structural changes in the system of adverbial connectors and their causes. The following section deals with complex expressions serving as adverbial connectors. To be included, such complex expressions have to be lexicalized. This means they have to be morphologically and semantically fixed (at least to a certain extent) and must fulfill the following conditions: (i) They must have lost at least some properties of the original phrase (e. g. modifiability of the head noun of a noun phrase). (ii) They must have at least one adverbial reading which is not fully re‑construable from the meaning of their parts. This, for instance, means that on the contrary is not included because its meaning is fully re‑construable from its parts, while on the other hand, which requires an understanding of the metaphorical transfer of hand, is included. The

Pronominal connectors – lexicalized phrases   97

section here only introduces the relevant terms and issues and sketches the general lines of development which will then be discussed in more detail in the chapters on the development in the causal/resultive and contrastive/concessive relations (see Chapters 9 and 10). Table 7.4.:  Pronominal connectors – lexicalized phrases Pronominal ­Connectors: Demonstratives OE case forms: þæs þy forþæm forþon forþy naþylæs

ME 1/2 for that forhwi ?neverthelatter nought‑for‑than ME 3 for which ?nevertheless nought‑for‑that ‘nevertheless’ nought‑for‑thi ‘nevertheless’ over this over that ME 4

Pronominal ­Connectors: Other Pronouns

Lexicalized Prepositional Phrases

hwæðere swaþeah‑hwæðere þeahhwæðre hwæt hwæt þa swa eall‑swa swaþeah eac swylce ?therefore wherefore

Lexicalized Phrases

þa giet

forsooth on oðer half overall

herefore oðer side

otherways otherwise

contrariwise howbeit notwithstanding EModE1 ?nonetheless

albeit nonetheless over and besides

98   Adverbial connectors: morphology Pronominal ­Connectors: Demonstratives

Pronominal ­Connectors: Other Pronouns

Lexicalized Prepositional Phrases

Lexicalized Phrases

above all indeed in fine at least in the meantime in the meanwhile in sum

over and above

EModE2

EModE3 after all on the other hand LModE1 at the same time in truth LModE2 anyhow

of course at all events in fact

altogether

anyway

at any rate in the same way

all the same

LModE3

PDE all in all instead

7.5.  Pronominal connectors 7.5.1.  Pronominal connectors in Old English The notion as well as the main properties of “pronominal connectors” have been introduced in Chapter 5.5 above. In Raible’s (1992) continuum, they are situated between asyndesis (I) and the explicit paratactic connection (III). This phase II, then, singles out connectors comprising a pronominal element, which function as explicit signposts of discourse deixis. In Old English, we find case forms of the demonstrative, which are used to mark a resultive relation ‘therefore’, namely the genitive þæs (masculine/neuter) and the instrumental/locative þy. While þæs is very rare and not attested at all in the corpus texts of the present study, þy is regularly employed in indepen-

Pronominal connectors   99

dent and correlative constructions (for the sequence cause – result ‘therefore’ as well as for the sequence result – cause ‘for’). Consider, for example, the following example from the beginning of Wulfstan’s Sermo Lupi ad Anglos: (62) Leofan men, gecnawað þæt soð is: ðeos worold is on ofste, & hit nealæcð þam ende, & þy hit is on worolde aa swa leng swa wyrse (COWULF3, R2) ‘Beloved men, know what the truth is: this world is in haste, and it approaches the end; and therefore it is ever worse and worse in the world; …’ This pattern of pronominal connectors is the predominant one for the Old English causal/resultive connectors °forþæm, forþon and forþy ‘for; therefore’, which consist of the preposition for and forms of the simple demonstrative, namely the dative þæm (masculine/neuter) or the instrumental þon or þy. It is also attested in the contrastive/concessive connector naþylæs ‘nevertheless’. In addition to demonstratives, the concessive/contrastive relation is marked by connectors containing the originally interrogative pronoun hwæðer ‘which of the two?’, which may also be used as an indefinite pronoun “either of two” (see BT, s. v., and Campbell 1959: § 716–718). Other cases of pronominal connectors are hwæt (þa) ‘what (then)’, which uses the neuter (nom./acc.) form of the interrogative pronoun hwa ‘who’ (Campbell 1959: § 716) and, in particular, all of the forms which employ the extremely polyfunctional particle swa ‘so’, which may also on its own be used as a resultive adverbial connector (see BT, s. v. swa IV.). For the polyfunctionality of swa, see the following entry chart extracted from BT; s. v. swa (all of these uses are discussed in detail in Schleburg 2002). Table 7.5.:  OE swa I. rel. pron. As, that II. demonst. pron.; III a. swā swā such as IV. adv. (1) defined by that which precedes so, in this or that way, thus (2) defined by that which follows, (3) used indefinitely, so and so V. adverbial conjunction, (1) with indic. as (2) with indic. or subjunct. expressing an actual or possible result, so that (3) with subjunctive, as (if) (4) with optative, so (5) with a conditional force, provided that, if so be that, so :(6) marking a consequence, so, therefore, on that account :(7) local, where :-

100   Adverbial connectors: morphology (8) temporal, as, when :-(9) marking the grounds of action, as, since :(10) although, yet :VI. swā . . . swā, (1) - where swá occurs once with a demonstrative, once with a relative force, so . . . as, so . . . that, as . . . as :-(2) correlative, (a) either . . . or, as well . . . as :-- (b) whether …or :-(c) swā hwæðer swā . . . swā whether . . . or :-

7.5.2.  Pronominal connectors in the history of English The morphological make‑up and the functions of the central pronominal connectors marking the relations cause/result (OE þy, forþæm, forþon, forþy ‘because; therefore’ and concession/contrast (OE swaþeah, þeahhwæðere, swaþeahhwæðere, etc.) will be discussed in detail in the chapters on their respective semantic relations in Chapters 9 and 10. For a first survey of the history of pronominal connectors, Table 7.4 above shows that this pattern is, as far as forms of the demonstrative are concerned, only used until the Middle English period (compare, in ME3, additive over this, over that and contrastive/ concessive nought‑for‑thi ‘nevertheless’, nought‑for‑that ‘nevertheless’). Speakers and writers of Present Day English may, of course, freely use adverbial connectors which include a pronominal reference to the unit which is to be related, such as in addition to this/that (i. e. what I have mentioned), for (all) this/that (i. e. what I have said), besides this/that, as a result of that, because of this/that, despite/in spite of this/that, instead of this/that, etc. (for a fuller list, see Quirk et al. 1985: 632 and Fraser 1999: 939–940).

In contrast to Old or Early Middle English, however, these phrases are neither fully lexicalized nor are they very frequent. In the LLC and LOB corpora, for example, the full lexical phrase for this reason is used only seven times (compared to 265 instances of so, sixty‑two of therefore and thirty‑five of thus). Overall, Altenberg finds 385 instances of adverbial connectors in contrast to only 14 instances of pronominal connection (Altenberg 1984: 40). This is a striking difference to the use of pronominal forþæm, which is used in about 80 per cent of the cases of causal/resultive connection in Old English (see below, Chapter 9, Table 9.6). In contrast to German (see the bold pronominal elements in causal/resultive PDG demnach, deswegen, infolgedessen or concessive PDG nichtsdestoweni‑ ger), English thus has not only given up this very productive pattern of forming

Pronominal connectors   101

Old and Middle English connectors. Intriguingly, Present Day English has not even kept a single of these pronominal connectors which had been central in its earliest periods.4 In view of the general history of the English language, it is certainly no coincidence that Middle English is the last period which employs (or rather: tries to employ) adverbial connectors of this “pronominal” kind in larger numbers, since Early Middle English is the time when the nominal categories case and gender were lost. Much more important for the present issue of pronominal connectors, however, are the changes which affected both of the Old English demonstratives (the proto‑article se/þæt/seo and the “emphatic” demonstrative þes/þis/þeos). These forms, which were pronominal as much as they were adjective‑ or determiner‑like, are no longer inflected for case and gender after the Old English period (see Brunner 1960–62: II, 130–133; Lass 1999: 112–116). At the very beginning of the Middle English period (around 1200), we see a collapse of the old system of demonstratives, but at the same time also the emergence of two distinct categories: the invariable article the, and the demonstratives this/that. While there is – among the nineteen different forms of the demonstratives – nowhere a direct ancestor for the Middle and Present Day English article the (for its etymology, see OED, s. v. the dem. a. (def. article) and pron.), it is clear that the neuter nominative singular forms of the simple (OE þæt) and the “emphatic” demonstratives (OE þis) yielded the proximal/ distal deictics this and that.5 The collapse of both of the paradigms of the demonstratives thus first of all triggered – when the case forms OE þæm and þy lost their transparency as dative/instrumental forms of the demonstrative – the coinage of a number of new adverbial connectors in earlier sub‑periods of Middle English (ME1 to ME3). Most of these, such as the additive over this/that or the resultive for that, employ the Old English pattern of pronominal connectors formed by a preposition and the demonstrative that (for similar patterns in the Romance languages, especially Latin, see above, Chapter 5.5). These innovations, however, have 4 Nevertheless and nonetheless might be considered marginal cases because they employ the form the, which is also found in other contexts preceding an adjective or adverb in the comparative degree (cf. the sooner … the better). For speakers of Present Day English, the formation from the original instrumental of the demonstrative, þy, however, is not transparent. The OED, for example, does not give a meaning but the paraphrase “The radical meaning is ‘in or by that’, ‘in or by so much’” (OED, s. v. the adv.). 5 Loss of inflections is, as with most inflections, earliest in the East and in the North. We find remnants of inflected forms of the Old English demonstratives in some southern and western texts as late as the thirteenth century.

102   Adverbial connectors: morphology probably not survived because, at about the same time, that also developed into the general subordinator (instead of OE þe, which, in turn, was probably given up because of its homonymy with the new invariable article the). From early Middle English onwards, that may be used as a complementizer, a relativizer and – often pleonastically – as an indicator of a subordinate clause. For the conjunctions, Fischer summarizes the developments – which are analysed in detail in Kivimaa (1966) – as follows: The scenario therefore is not that þe was replaced by that; rather, that, by now the general subordinator, becomes added, as a kind of pleonastic element, to all kinds of conjunctions. Thus we find in Middle English, now that, (g)if that, when that, and after prepositions (prepositions not encountered in Old English conjunctive phrases) before that, save that, in that, etc. (Fischer 1992: 295).

As will be argued in an analysis of causal/resultive connectors below (Chapter 9), this use of that as the general subordinating particle seems to have inhibited a further extensive use of pronominal connectors such as for that. When placed sentence‑initially, for that is not only (as OE forþæm had been) ambiguous as to whether it is an adverb or a conjunction, but openly misleading, because hearers or readers of Middle English expect a pre‑posed subordinate clause with (pleonastic) that and not a paratactic construction introduced by an adverbial connector. Indeed, we notice that no items employing this pattern of “pronominal connectors” have been coined after the Middle English period and none of the ones coined in the Middle English period have survived – or are still transparent – as pronominal connectors. It is this development which also accounts for the finding (sketched above) that most of the Middle English innovations did not survive the Middle English or Early Modern English period. Middle English is a “period of transition and experiment”, mainly because the pronominal connectors coined in its sub‑periods were given up. 7.6.  Lexicalized (prepositional) phrases 7.6.1.  General tendencies As has been pointed out above, pronominal connectors have the advantage of being inherently deictic.6 They function as signposts of discourse deixis, because they explicitly signal the link between two sentences or two chunks of 6 I use “deixis” and “deictic” as the hyperonym covering all kinds of deixis, although some linguists argue for a distinction between reference and discourse deixis; for details, see below, Chapter 9.4.1.

Lexicalized (prepositional) phrases   103

discourse by using their deictic value to anaphorically (in correlative constructions also cataphorically) take up one or more constituents of the proceeding or ensuing discourse (cf. Raible’s definition “Junktion durch Wiederaufnahme (eines Teils) des vorhergehenden Satzes” ‘Linkage by resumption of (a part of) the preceding sentence’; Raible 1992). This overt marking of the anaphoric (or cataphoric) relations is – for reasons of information signalling and processing – much more important for adverbial connectors than for subordinators. Subordinators usually work on a local level of discourse and the information given in the subordinate clause is commonly not pursued in the following discourse (for details, see above, Chapter 2.4). In paratactic constructions, however, the propositions of the two connects are of equal weight: both of them may be pursued in the following discourse. It is therefore most important that the adverbial connectors, which signal the semantic relation between the two connects, are distinctive in their deixis. After pronominal connectors were given up, speakers and writers of English have thus not given up using deictic material altogether, but mainly choose time and space deictics to compensate for the loss of pronominal connectors (for details, see below, Chapters 8.2 and 8.3). Most of these are coded in lexicalized prepositional or other phrases. 7.6.2.  Lexicalization Table 7.4 nicely illustrates – compare the two columns on the left with the two columns on the right – that only in Middle English – the “period of experiment” – both the patterns of pronominal connectors and that of lexicalized phrases are used to a considerable extent. Old English, by contrast, virtually does not employ lexicalized phrases. From Early Middle English onwards, on the other hand, speakers and writers ever more rarely coin new adverbial connectors comprising pronominal material.7 As has been pointed out above, prepositional and other phrases are only included if they show a certain degree of lexicalization (for the criteria, see above, 7.4). Often, the originally complex connectors show orthographical fusion, such as fully fused indeed. For others, orthographical variants are attested which support their inclusion in the present study. In Middle English, for instance, the phrase at/on/of last ‘finally’ is attested as alast (see MED, 7 Anyhow and anyway (comprising the element any which may be used pronominally in Present Day English), for example, are not relevant for the present issue, because they are generalizing rather than explicitly deictic. For this pattern, see below, Chapter 10.

104   Adverbial connectors: morphology s. v. a‑last). This also applies to longer phrases such as over and above ‘furthermore’, a popular reinforcing connector in Early Modern English, which is also recorded as the single word hoverendebuv (see OED, s. v. over and above). Similarly, in Present Day American English writers sometimes choose to signal the lexicalization of the phrase after all in its spelling afterall as one word (see OED, s. v. after; 10.; quoted above, Chapter 1). Some of the lexicalized phrases, however, also show subsequent reduction: this is most evident in PDE of course, which – in particular in spoken language  – is commonly reduced to course (and often hard to differentiate from causal ’cause < because). Interestingly, the OED has a separate entry for ’course, but lists the highly frequent of course among the entries of course (see OED, s. v. course n. 37c, and Vandenbergen‑Simon and Aijmer 2001 for details on the high frequency and polyfunctionality of of course in Present Day English). This shortened form has been attested since the beginning of the twentieth century (see OED, s. v. ’course, course). Its first attestation as an adverbial connector is recorded for 1967:

1901 MERWIN & WEBSTER Calumet ‘K’ i. 13 Have you tried to get any of it here in Chicago? Course not. It’s all ordered and cut out up to Ledyard. 1904 G. S. PORTER Freckles xvi. 326 ‘I bet you it’s a marked tree!’ ‘Course it is!’ cried the Angel. 1967 O. NORTON Now lying Dead iv. 58 ’Course, he might go off. He does sometimes (OED, s. v. ’course, course).

7.6.3.  Verbal and nominal phrases Most of the phrases which have been lexicalized as adverbial connectors are prepositional phrases. There are only three originally verbal phrases, namely the ambiguous adverbs/conjunctions albeit, howbeit and notwithstanding (that). Albeit and notwithstanding are calques modelled on French (cf. French tout soit il que and OF non obstant). All of these verbal phrases were first used as subordinators (see Kortmann 1997: 333), but are also – though less frequently – attested as contrastive/concessive adverbial connectors during the Middle and Early Modern English periods (see Appendix B.4.1). As has been shown above, the lexicalization of verbal phrases was a regular pattern for the coinage of new subordinators, in particular from non‑finite constructions. This pattern is, by contrast, not typically employed for the formation of adverbial connectors. Most of the new adverbial connectors are lexicalized forms of prepositional phrases, usually of the simplest kind, i. e. preposition plus noun (without any other pre‑ or postmodification). This agrees with a more general tendency in

Lexicalized (prepositional) phrases   105

the structural make‑up of the English lexicon, namely the substitution of single words by multi‑unit items (cf. the Chapter “Der Wortverband als Wort” in Leisi and Mair 1999: 94–111). Although the term “Wortverband” (referring to fixed lexical bundles with a particular meaning) is – as the authors admit – a rather vague term, it still manages to characterize phenomena such as the substitution of simple, often borrowed verbs by phrasal and prepositional verbs, or by verbal operator and gerund (cf. to have a look, to give a ring, to make a move, etc.). The origin of these verbal forms (including phrasal and prepositional verbs) has commonly been sought in everyday, colloquial language. Yet, in a corpus study of phrasal verbs from 1500 to 1700, Thim (2006) shows that this claim is unsubstantiated: he suggests that it is their specific function rather than their presumed informality which determines their use in Early Modern English texts (Thim 2006: 305). Similarly, with the exception of transitional indeed and in fact, none of the lexicalized phrases discussed here shows a strong affinity to speech‑based texts (CEPLAY, CETRI, CLPLAY). Furthermore, a colloquial origin is not likely for the lexicalization of adverbial connectors, which are mainly employed in written genres, more specifically academic prose. By contrast, a closer look at the prepositional phrases reveals that they show recurrent patterns in the material they use: the two basic source domains are time and place/space. place/space in particular is, however, only explicable on the basis of the spatialization of language in literacy.

8.  Cognitive source domains

8.1.  General tendencies The analyses in Chapter 7 have shown that new adverbial connectors coined after the collapse of the system of pronominal connectors in Old English are only rarely simple transfers of purely lexical, propositional adverbs from circumstance to linking adverbials (see the few cases such as consequently or similarly). Much more frequently, the innovations comprise material from a very restricted range of cognitive source domains which help to provide the necessary discourse deictic references for the explicit linkage of sentences or chunks of discourse. In his study of subordinators, Kortmann (1997) shows that interclausal relations differ markedly with regard to their “cognitive basicness or centrality for human reasoning” and also their “cognitive complexity or specificness”. According to Kortmann, highly grammaticalized, preferably monomorphemic or single‑word connectors which are frequently used and are stable over time indicate “cognitive basicness” (“lexical primes code cognitive primes”; Kortmann 1997: 342). Most of the basic relations thus show a low degree of complexity. Yet, cognitive basicness and cognitive complexity must nevertheless be kept separate, since concession, for instance, exhibits “a high degree of cognitive complexity, and yet clearly belongs to the core set of cognitively basic relations”. On this basis, Kortmann found a core of twelve basic relations (cause, condition, concession [ccc], result, purpose, simultaneity overlap, simultaneity duration, anteriority, immediate anteriority, terminus ad quem, place and similarity) and several layers of relations of an increasingly peripheral nature. These relations can be parcelled into four networks, i.e. temporal, locative, modal and ccc‑relations. Network‑transcending changes are generally unidirectional, so that we can distinguish between source (locative, modal) and goal (ccc) domains. The cccc‑ or four‑c relations (contrast is now commonly added as a fourth domain; see Couper‑Kuhlen and Kortmann 2000) constitute the prototypical goal network, i.e. endpoints of network‑transcending semantic changes. Cross‑linguistic polysemy patterns (Kortmann 1997: 175–211) also show that affinities are strongest between the temporal and the cccc-networks, so that original temporal connectors often develop cccc‑readings.

   107

Table 8.1 provides an overview of the three prototypical domains which can be established as source domains for adverbial connectors. Among them are time and space, which have long been known to create deictic reference frames in language. motion is listed as a separate category here, but is essentially a combination of space and time: the concept of motion implies the movement in/over a certain space during a certain period of time. truth/fact is here introduced as a third category because it plays an important role in innovations in the field of concessive and transitional connectors. Table 8.1.:  Cognitive source domains: overview place/space

time

motion

truth/fact

OE her ‘here’ þærtoeacan ‘also’

eft ‘again’ eftsona ‘again‑soon’ eftsones ‘again‑soon’ nu ‘now’ ⇔ þa ‘then’ þonne ‘then’ ærest ‘earliest’ > ‘first’ ⇔ þa giet ‘also’

cuþlice ‘certainly’ eornostlice ‘earnestly’ huru ‘certainly, indeed’ huruþinga ‘certainly, indeed’ soþlice ‘truly’ gewislice ‘certainly’ witodlice ‘truly’

at (the) last finally nevertheless neverthelatter yet

certainly certain certes forsooth iwis sekirly truly verily

ME after ‘moreover’ ⇔ algates (cf. gate) further farthermore fartherover furthermore {first; cf. fore ‘in front’} furtherover herefore moreover notwithstanding on oðer half oðer side otherways overall overmore over this/that therefore wherefore

108   Cognitive source domains place/space EModE above all again (cf. OE on‑gean) beside besides hence on the other hand over and besides over and above thence whence LModE anyway in the same way PDE instead

time

motion

truth/fact

after all ⇔ again in fine however howsoever howsomever lastly (in the) meantime (in the) meanwhile

still (i.e. non‑ motion)

indeed sure surely

at the same time

{of course} in fact in truth {rather ‘faster’}

{}  transparency of source category uncertain ⇔  time and space

8.2.  Source domain time Table 8.1 shows that the two central source domains time and space have been used since the Old English period. In Old English, we predominantly – with regard to their token‑frequency – find simple temporal (OE nu ‘now’, OE þonne ‘then’), spatial (OE her ‘here’) or temporal/spatial (OE þa ‘then; there’) circumstance adverbials which are used as connectors to signal the relation transition (for the occurrences, see Appendix B.5.1). This use of text‑deictic nu, her and resultive/transitional þonne is illustrated by Old English examples in Chapter 5. Through the use of these temporal or spatial adverbs, the spoken or written text is established as the frame of reference (instead of the real world). This, for example, allows collocations such as now then, which would be incongruous in a purely temporal use, but are attested in all periods of English (see the lists in Appendix B.5.1): (63) Nu þenne on oðer half. nim … (CMSAWLES, p. 131). ‘Now, then, on the other hand, take … ’.

Source domain time   109

In addition to these transitional connectors, Old English also employs temporal and spatial adverbs for the relations of addition (OE eft, eftsona/‑es ‘again’ and þærtoeacan ‘there‑to‑also’) in a natural transfer of the temporal sequence on the sequence of the argument. The source category time is thus preferred over space in Old English, certainly no coincidence in a time when most of the language production was oral (for the spatialization of language in literacy, see below, 8.3). For a closer inspection of the transfers from source to goal domain, the following table lists the new adverbial connectors by the semantic relations they code. Table 8.2.:  Source domain time – innovations per period addition listing/ summmative result contrast/ concession

OE ME eft ‘again’ eftsona ‘again’ + ‘soon’ eftsones ‘again’ + ‘soon’ þa giet ‘also’ ærest ‘earliest’ >‘first’ þonne ‘then’

EModE again

LModE

at (the) last in fine finally lastly

nevertheless neverthe­ latter yet

after all of course (in the) meantime at the (in the) meanwhile same time still (i.e. non‑motion) however howsoever howsomever

transition nu ‘now’ þa ‘then’ þonne ‘then’

The innovations in the semantic relation of addition are fairly straightforward: new adverbial connectors are coined in the sub‑group of listing connectors, which are used to signal the sequence of arguments by elements of time deixis (e.g. at the last, finally, etc.). With respect to the ccc‑relations, then is used to explicitly code not only the transitional, but also the resultive relation, following the principle “post hoc, propter hoc” (see below, Chapter 9). Table 8.2 shows, however, that most of the innovations from the domain time are not found in the relation of cause/ result but in the relation contrast/concession: From Early Modern English onwards, the predominant pattern is the transfer of expressions marking concomitance or simultaneity, as in at the same time, (in the) meantime or in

110   Cognitive source domains the meanwhile and also still, which implies non‑movement over the respective time period (on still and yet, see König and Traugott 1982; on after all, see Traugott 1997 and below, Chapter 10). Cross‑linguistically, simultaneity and concomitance expressions have been found to be one of the major new sources of concessive connectors, subordinators (cf. PDE since, PDG während) as well as adverbial connectors (cf. PDG dabei; PDBavarian derweil). In his study on cross‑linguistic patterns in the formation of concessives (“Where do Concessives Come from?”, König 1985b: 269), König shows that simultaneity and concomitance are often augmented by concessive inferences, and that the coinage of connectives is a process in which these conversational implicatures become conventionalized. Consider, for example, the conversational implicatures Al‑ though he is poor and Although he is rich in sentences such as (64a) and (64b): (64) (a) Poor as he is, he spends a lot of money on horses. (b) Rich as he is, he spends a lot of money on horses. 8.3.  Source domain place/space With respect to the source domain place/space,1 it has long been known that spatial thinking provides humans with analogies and tools for understanding other domains, so that many of today’s metaphors of everyday language are spatial. Spatial cognition probably plays this central role because it seems to be the evolutionarily earliest domain of systematic cognition (Lakoff and Johnson 1980: 17; Levinson 2003). In addition to this cognitive centrality, we must, however, also consider another essential aspect: it has been suggested that the metaphoric extension of spatial terms into metalinguistic ones naming speech acts is a function of literacy, since it is only with literacy that language is objectified in visual space (see Ong 1982). In one of the still pivotal studies on the subject, Ong states that [v]isual presentation of verbalized material in space has its own particular economy, its own laws of motion and structure. […] Texts assimilate utterance to the human body. They introduce a feeling for ‘headings’ in accumulation of knowledge: ‘chapter’ derives from the Latin caput, meaning head (as from the human body). Pages have not only ‘heads’, but also ‘feet’, for footnotes. References are given to what is ‘above’ and ‘below’ in a text when what is meant is several pages back or farther on. (Ong 1982: 100) 1 For the distinction between place and space, see Levinson (2003). This distinction is only fundamental for universal accounts of place/space in language because Levinson finds that ideas of Western philosophy (from Aristotle onwards) are not necessarily transferable to non‑European ideas and languages.

Source domain place/space   111

Similar ideas have already been expressed by Kant, who argues that “in a written page we have first to note the difference between front and back and to distinguish the top from the bottom of the writing; only then can we proceed to determine the position of the characters from left to right or conversely” (quoted from Levinson 2003: 9). These considerations concerning the domain space and literacy clearly suggest that literacy seems to favour the widespread use of spatial terms as markers of discourse structure. Table 8.1 supports this idea because it attests a large number of new adverbial connectors employing the source category space in the Middle English and Early Modern English period. In these periods, English developed into an Ausbausprache and more and more replaced Latin and French in written documents and in registers which had, after the Norman Conquest, been reserved to Latin and French (for the re‑establishment of English from 1200 to 1500, see Baugh and Cable 2003: Chapter 6). With respect to the close interdependency of Latin and literacy in the vernaculars, it is, of course, very hard to decide whether some of these formations are indeed native formations as opposed to calques on Latin (or, for Middle English, French) models. Latin uses basically the same source domains for adverbial connectors as English does: the spatial adverb super ‘above, on top, over’, for instance, can also be used as an adverbial connector for the semantic relation addition ‘moreover’. Similar to English then, spatial Latin exinde ‘thence; after that, next in order, thereafter, then’ can be employed for the semantic relations addition ‘furthermore’ or result ‘by that cause’. Temporal interim ‘meanwhile, in the meantime; at the same time’ is, just like English expressions for simultaneity, used to code concession ‘however, ­nevertheless’. Similar parallels are found in French, as, for instance, concessive/transitional après tout which is parallel to after all. Latin or French influence can thus never be ruled out with certainty, just as it cannot be claimed with any certainty, at least with respect to the high‑frequency adverbial connectors (an exception is the morphologically “foreign” notwithstanding, comprising a non‑finite verbal form; see above, Chapter 7.6.3). Since we are dealing with very basic cognitive source categories and their transfer to texts, these transfers are of course similar in each of the individual languages and are found in all of the European languages which have an extended tradition of literacy (see the “Euroversals” in Kortmann’s cross‑linguistic analysis of subordinators; Kortmann 1997). In line with the spatialization of language in literacy as described above, many of the new coinages from Middle English onwards use the spatial extension of the text on a page as the frame of reference, mainly to signal the relation addition. Here, the spatial sequence is used to express the sequence in the line

112   Cognitive source domains of argument. The innovations show various degrees of transparency: while the early pronominal connectors over this/that and also the double comparatives such as furthermore or moreover (discussed above, Chapter 7.3.2) have lost some of their transparency, the later coinages such as beside(s), over and besides are fully transparent in their spatial origin. Table 8.3.:  Source domain place/space – innovations per period

addition

OE

ME

EModE

LModE

þærtoeacan ‘also’ þa giet ‘also’

after ‘moreover’ further farthermore fartherover furthermore furtherover moreover overall overmore over this/that first (cf. fore ‘in front’) herefore therefore wherefore algates (cf. gate) notwithstanding on oðer half oðer side otherways

above all again beside besides over and besides over and above

in the same way

result

contrast/ concession contrast/ reformulatory transition

hence thence whence anyway again (OE on‑gean) on the other hand

instead

her ‘here’

The same differences may be seen in those connectors which signal the resultive expression. The early coinages (ME1/2) from the spatial domain, herefore, therefore and wherefore, employ spatial terms, which do, however, not carry the full discourse deictic load themselves, but are formations similar to pronominal connectors. Instead of the demonstrative, a proximal (here) or distal (there) element is employed, which takes over the deictic (anaphoric or cataphoric) function of the demonstrative. This is a regular pattern for adverbial innovations at the beginning of the Middle English period, as attested by the many adverbs and conjunctions with deictic there, here and where (often in ambiguous adverbs/conjunctions; symbol °) as their first element. The follow-

Source domain place/space   113

ing lists are compiled from the OED (on the history of the there‑compounds, see Österman 1997): (a) thereabout, thereabouts, thereabove, thereafter, thereafterward, thereagain, thereagainst, thereamong, thereas (conj.), thereat, thereatour, thereaway, there‑ aways, therebefore, therebeside, thereby, theredown, thereforth, therefro, there‑ from, theregain, therehence, therein, thereinne, thereintill, thereinto, there‑mid, ther‑mid, ther(e)‑mide, ‑mydde, there‑nigh, thereof (b) hereafterward, here‑again, here‑against, hereat, hereaway, hereaways, herebe‑ fore, hereby, herefrom, hereft, here‑hence, herein, hereof, hereon, hereout, here‑right, hereto, heretofore, heretoforetime, hereunder, hereunto, hereupon, herewith (c) whereabout (interrog. and rel. adv.), whereabouts, whereafter (rel. adv.), whereas (rel. adv., conj.), whereat, whereby, wherefro, wherefrom, where‑hence, wherein, whereinne, whereinsoever, whereintill, whereinto, wheremid, whereof, whereon, whereout, °whereso, °wheresoever, °wheresome, °wheresomever, wherethorough, wherethrough, wheretill, whereto, whereunder, whereuntil, whereunto, °whereup, whereupon, °whereever, wherewith, wherewithal

With respect to the adverbial connectors, only the highly frequent therefore has survived. In Present Day English, the pronominal sense “for that (thing, act, etc.); for that, for it” is labelled by the OED as “archaic” or “colloquial” (OED, s.v. therefore, 1.). It is distinguished from the adverbial connectors by a different stress pattern (stress on the second syllable vs. initial stress for the adverbial connector) and thus it is to be doubted whether the spatial origin of therefore is recognized by speakers of Present Day English. While therefore, herefore and wherefore thus exhibit a mixed character, hence, thence and whence, which begin to be used as adverbial connectors in the Early Modern English period, are purely spatial, and take the spatial extension of the text as their sole frame of reference. They develop a resultive meaning modelled on the cognitively central “post hoc, propter hoc”. Interestingly, here the proximal hence has survived (in contrast to the distal there in therefore).2 These differences in the employment of spatial concepts for resultive connectors thus again mirror the general tendencies in the coinage of new adverbial connectors in the history of English. While speakers and writers of Middle English employ essentially pronominal elements, speakers and writers of Early Modern English refer to the spatial reference only. 2 In many of the occurrences of hence as a linking adverbial, this spatial meaning is emphasized by a verb which also comprises a spatial element, such as proceed: “For hence it proceedeth that Princes find a solitude, in regard of able men to serue them in causes of estate, because there is no education collegiate …” (CEEDUC2B, Bacon, Sample 2, p. 3V).

114   Cognitive source domains 8.4.  Source domain truth/fact 8.4.1.  General tendencies The present study distinguishes the domain truth/fact as a third important source domain. This domain is, however, restricted to the semantic relations concession/contrast (as a conversational implicature) and, in conventionalized form, transition (which is used as the general term for labelling the semantic relation marked by truth‑intensifiers, which have not been discussed in earlier literature). Table 8.4.:  Source domain truth

contrast/ concession transition

OE

ME

EModE

LModE

{huru ‘certainly, indeed’} {huruþinga ‘certainly, indeed’}3 eornostlice ‘earnestly’ soþlice ‘truly’ gewislice ‘certainly’ witodlice ‘truly’

certain certainly certainly certain certes forsooth iwis sekirly truly verily

sure surely indeed sure surely

in truth in fact in truth

8.4.2.  Conversational implicatures: concession/contrast In interactive communication, the semantic relation of concession commonly involves the acknowledgment of a claim. Discourse‑functional approaches (see the various studies by Couper‑Kuhlen, Barth‑Weingarten and Thompson) thus see concessivity as a relation of claim – acknowledgement of claim – counterclaim, which prototypically holds between two turns at a talk. They establish a “cardinal concessive pattern”, prototypically represented as A:

x (claim)

B:

X’ (acknowledgement of claim)

Y:

(counterclaim)

3 OE huru is of unclear etymology (see Holthausen 1934, s.v. huru); it has survived in the Swedish subordinator ehuru ‘although’.

Source domain truth/fact   115

It is present when a speaker (A) makes a point or claim in one turn and when a second speaker (B) acknowledges the validity of this claim in the next turn but proceeds stating that a potentially incompatible point (Y) also holds. In discourse, but (usually classified as adversative) is by far the most frequent connective signalling this relation (96 per cent of all cases with an explicit marker; see Barth‑Weingarten 2003: 80). This acknowledgement of the claim may be signalled (or emphasized) by epistemic adverbs, which in their propositional meanings denote the concepts truth or certainty. In German, the conversational implicatures of such adverbs have been conventionalized in the concessive connector PDG zwar (< ze ware PDG ‘in Wahrheit’, ‘in truth’; Kluge and Seebold 2002, s.v. zwar). Yet for an unambiguous expression of concession/contrast, both zwar (obligatorily) and the not yet fully conventionalized PDG gewiss need a correlative aber or jedoch ‘but, on the other hand’ in the second connect. This corresponds to the situation in all periods of English: In certain contexts, epistemic adverbs such as sure, surely or of course may have conversational implicatures of concession/contrast. This conceding move – expressed by it is true … but – is most evident in full sentences such as it is true used in parenthesis, as in (65): (65) This instinct, it is true, arises from past observation and experience; but can any one give the ultimate reason, why past experience and observation produces such an effect, any more than why nature alone should produce it? (CLHUM1, Part III). (66) It is true there has been all along in the world a notion of rewards and punishments in another life, but it seems to have rather served as an entertainment to poets or as a terror of children than a settled principle by which men pretended to govern any of their actions (CLSERM1A, Swift “On the Wisdom of this World”). The full phrase it is true is very common in this function in Late Modern English (periods LModE1 and 2; see Appendix B.5.2), both sentence‑medially and sentence‑initially. The concessive/contrastive function is, however, also inherent over all periods in the adverbs certain(ly) and sure(ly), from ME1 and EModE1 respectively. Example (67) shows that correlatives such as surely … but are, in particular, used as explicit clues for highlighting the relation and linkage of larger chunks of discourse. (67) And surely we may pronounce upon it, in the words of St. James, that “This wisdom descended not from above, but was earthly and sensual”.

116   Cognitive source domains What if I had produced their absurd notions about God and the soul? It would then have completed the character given it by that Apostle, and appeared to have been devilish too. But it is easy to observe from the nature of these few particulars that their defects in morals were purely the flagging and fainting of the mind for want of a support by revelation from God … (CLSERM1A, Swift “On the Wisdom of this World”). 8.4.3.  transition The adverbs soþlice (‘truly’; from soþ ‘truth, true’, soþlic ‘true’) and witod‑ lice ‘certainly’4 are employed in various functions in Old English: on the phrase level, they may be used as circumstance adverbials, mainly in direct speech with a first‑person subject (e.g. in phrases such as ic secge soþlice ‘I tell [you] truly’), or they are employed as emphasizers. Yet, as I have argued in some preliminary studies for this book (Lenker 2000, 2003, 2007b), they also have another function – which is not commonly noted in dictionaries – in Old English prose: as sentence adverbials and eventually adverbial connectors or discourse markers, they lose much of their original meaning, extend their scope from the phrase level to at least the sentence level. At the same time they develop a metatextual function – they are used as discourse markers5 demarcating episode boundaries on the global level of discourse and as highlighting devices on the local level of discourse (cf. Lenker 2000). For lack of a better category, I have classified these adverbials as transitional adverbial connectors, since they, similar to PDE now, mainly work on the textual level in their connector function. In most of their occurrences, they – in line with the properties of adverbial connectors – signal a two‑place relation and do not alter the propositional meaning of any of the connects (i.e. they are non‑propositional). In a recent study, I could show that this development from circumstance adverbial to connective is not specific to soþlice and witodlice, but that etymologically similar adverbs, i.e. those with a basic meaning ‘truly’, show the same functional diversity and development (Lenker 2007b). The findings of these studies will be briefly summarized in the following sections (using – in 4 Witod is the past participle of the verb witian ‘to order, to decide’, which is cognate with the verb witan ‘to know’. 5 I here follow a broad definition of discourse markers, similar to that of Fraser (1999), which includes all kinds of connectives which express a two‑place relation (see above, Chapter, 3.6). It excludes, however, items without such a connective function, i.e. disjuncts such as frankly or briefly or interjections such as oh or yes.

Source domain truth/fact   117

addition to the corpus of the present study – also the corpus material of these earlier studies), and will be related to one of the clines Traugott (1995, 1999; Traugott and Dasher 2002) has established as a regularity in language change,6 the so‑called “adverbial cline” from

clause‑internal adverbial > sentence adverbial > discourse particle.

In a last step, I will show that this onomasiological approach suggests that the “adverbial cline” from referential to connective function is inherent in the semantics of the lexemes investigated, if we take a pragmatic perspective and consider communicative principles, in particular Grice’s Maxims of Quality (“Do not say what you believe to be false”; “Do not say that for which you lack adequate evidence”; Grice 1989: 25–27). 8.4.4.  OE soþlice – ME soothly These adverbs with the basic meaning ‘truly’ are, to be linguistically more precise, epistemic modal adverbials expressing the speaker’s commitment to the truth of the proposition. Accordingly, they may be – in very restricted contexts – be used as circumstance adverbials, most typically with a first‑person subject and a verb of communication (OE secgan, ME seien, tellen)7 in the present indicative form, such as (68) Nacode he scrydde, and swa ic soðlice secge, ealle nyd‑behæfnysse he wæs dælende þam þe þæs behofodon (Passion of St. Eustace, 9; ed. Skeat 1900: 190). 6 This cline has been the subject of various papers by Traugott and is summarized in Traugott and Dasher (2002: 152–189). In the last decade, it was considered to be a subtype of grammaticalization in an approach which might be labelled “a semantic‑pragmatic approach to grammaticalization”, and it was argued that this “adverbial cline” should be added to the inventories of clines in grammaticalization; see, for example, the title of Traugott (1995 [1997]). Grammaticalization processes proper and the cline sketched above indeed agree in several important characteristics, in particular semantic bleaching, increased subjectification and pragmatic strengthening. These structural similarities are also stressed in a more recent publication (Traugott and Dasher 2002), but the cline is now no longer regarded as a subtype of grammaticalization, but as a “regularity in semantic change”. 7 Of the altogether twenty manner adjunct/emphasizer uses of sothli in Chaucer, fourteen modify sei(en) or tell(en). In Gower’s Confessio Amantis, four of the five direct‑speech instances are with the communicative verbs say and tell, one with lieve ‘to believe’.

118   Cognitive source domains

‘The naked he clothed; and, as I truly tell, he distributed to every necessity of them that had need thereof’.

(69) And thus I mai you sothli telle, …, I am in Tristesce al amidde … (Gower, Confessio Amantis, 4.3496–3499; ed. Macaulay 1900–1901). ‘And so I can truly tell you, …, I am completely in the middle of Despair …’. Soþlice or soothly with a verb of communication can either be used in a parenthesis, as in (68), or as a matrix clause, as in (69). These uses provide the basis for all the later uses of soþlice as a stance adverbial (style disjunct) and connector on the textual level: speakers want to stress the assertion of another clause by explicitly pointing to its truth value. First, this underlying phrase gives rise to the adverbs’ use as “emphasizers” which may be defined as enhancing “the truth value or force of a sentence” (Quirk et al. 1985: 485). (70) Apolloni, ic oncnawe soðlice þæt þu eart on eallum þingum wel gelæred (Apollonius 16, 24; ed. Goolden 1958). ‘Apollonius, I know truly [indeed] that you are well taught in all things’. (71) … and if that he / ne hadde soothly knowen therbifoore / that parfitly hir children loved she … (Chaucer, Clerk’s Tale, (4) 688–690; ed. Benson 1987). ‘… and if he / had not truly [indeed] known before / that she loved her children perfectly …’. While examples (70) and (71) might be considered ambiguous cases (“I know in a true (certain) way …” or “I [intensifier] know ….”), an emphasizer use is clearly suggested by the stative verbs in (72) and (73) and further by the semantic contrast soothly vs. as I gesse in (73): (72) Ic eom soðlice romanisc. and ic on hæftnyd hider gelæd wæs (Passion of St. Eustace, 344; ed. Skeat 1900: 210). ‘I am truly a Roman, and I was brought hither in captivity’. (73) But Venus is it soothly, as I gesse (Chaucer, Knight’s Tale, (1) 1102; ed. Benson 1987). ‘But it is indeed Venus, as I guess’.

Source domain truth/fact   119

In non‑interactional, narrative contexts (i.e. not in direct speech) and thus in the majority of their occurrences,8 however, the scope of these adverbs is not restricted to the phrase level, but extends to the whole sentence. (74) Wæs he soðlice on rihtwisnysse weorcum … swiðe gefrætwod (Passion of St. Eustace, 4; ed. Skeat 1900: 190). ‘[transition marker], he was greatly adorned … with works of right­ eousness’. (75) Soothly, the goode werkes that he dide biforn that he fil in synne been al mortefied and astoned and dulled by the ofte synnyng (Chaucer, Par‑ son’s Tale, (10) 232; ed. Benson 1987). ‘[transition marker], the good works he did before falling into sin are all nullified, rendered null and made void by his frequent sinning’. The adverbs are here originally employed as stance adverbials and replace an independent sentence, a matrix clause, with a different proposition and a specifically first‑person subject – “soþlice ic eow secge þæt …” (style disjunct) or “soþ is þæt ic secge þæt …” (content disjunct). This change to sentence adverbial involves increased scope and syntactic freedom:9 the adverb is no longer part of the core syntactic structure, becomes syntactically optional and is often found sentence‑initially. As circumstance adverbials or emphasizers, soþlice/soothly are primarily found in direct speech with a first‑person (singular) subject because their propositional meaning ‘truly, truthfully’ demands a human agent with high trustworthiness, most likely the speakers themselves. In their sentential use, on the other hand, there is no constraint on either the subject of the sentence, which can be third person or even inanimate, or the verb, which can be in all tenses and moods. The adverbs do, however, retain their epistemic meaning ‘truly’ in that they introduce the voice of the narrator or author without directly mentioning her or him and allow the speakers to mark their opinion, their attitude or even their faith or trust in the veracity or importance of the proposition. The adverbs thus show increased subjectification, if this term is understood with Traugott as a process which focuses on the subject of a discourse and emphasizes subjective valuations.10 8 In Chaucer, eighty-six of the 114 instances of soothly are found in pieces of narrative prose. 9 This is one of the main reasons why it is difficult to regard this development as a case of grammaticalization in the morpho‑syntactic sense. 10 For the terms “subjectification” and “intersubjectification”, see Traugott and Dasher (2002: 19–24, 89–99). For a completely different conception of subjectification, see Langacker (1999).

120   Cognitive source domains The fundamental difference between the various uses of soþlice/soothly and its use as an adverbial connector lies in the fact that the adverb loses almost all of its propositional meaning ‘truth’ as a connector. Soþlice and soothly can no longer be considered style or content disjuncts because a paraphrase “ic secge soþlice” or “soþ is þæt ic secge” is impossible in these contexts. In (76), you find the instances which initially made me think about a textual function of soþlice when working on the rubrics of the Old English West Saxon Gospels (Lenker 1997). There, the beginnings of the gospel lections read during the Mass, i.e. the beginnings of the episodes, are marked by sentence‑initial soþlice. (76) (Luke I.26) Soþlice on þam syxtan monðe wæs asend gabriel se engel fram drihtne on galilea ceastre … RUBRIC (Luke I.39) Soþlice on þam dagum aras maria and ferde on muntland mid ofste. on iudeisce ceastre … RUBRIC (Luke I.56) Soþlice maria wunude mid hyre swylce þry monþas. and gewende þa to hyre huse … RUBRIC (Luke II.1) Soþlice on þam dagum wæs geworden gebod fram þam casere augusto. þæt eall ymbehwyrft wære tomearcod … (Old English Version of the Gospels; ed. Liuzza 1994) Luke I.26 ‘[transition marker] in the sixth month was sent Gabriel the angel by the Lord to a Galilean town …’. Luke I.39 ‘[transition marker] in these days Mary got ready and went to the hill‑country with haste to a Iudeaen town …’. Luke I.56 ‘[transition marker] Mary lived with her such three months and went then (back) to her house. …’. Luke II.1 ‘[transition marker] in these days an order was given by the Emperor Augustus that all the world should be described …’. Soþlice does not show any of its propositional or emphasizer meanings in these examples – the notions of ‘in a true way’ (circumstance adverbial) or of enhancing the truth value or force of a sentence as an emphasizer (implied “I tell you truly”) are no longer important. Instead, the focus is on the organization of discourse in which soþlice functions as a marker of the global text structure: it signals the beginning of a new episode. It is not employed conceptually, but procedurally. This procedural function is, however, not only found on the global level of discourse but also on the more local levels, a textual function especially attest-

Source domain truth/fact   121

ed for ME soothly. In prose texts, soothly or for soothly often follow quotations, as may be seen from the two examples taken from Chaucer’s Parson’s Tale, an argumentative piece of prose on the seven deadly sins. (77) Homycide is eek by bakbitynge, of whiche bakbiteres seith Salomon that ‘they han two swerdes with whiche they sleen hire neighebores’. For soothly, as wikke is to bynyme his good name as his lyf. Homycide is eek in yevynge of wikked conseil by fraude … For whiche the wise man seith ‘Fedeth hym that almost dyeth for honger’; for soothly, but if thow feede hym, thou sleest hym (Chaucer, Parson’s Tale, (10) 565–570; ed. Benson 1987). ‘Homicide is also by backbiting, of which backbiters Salomon says that “they have two swords with which they slay their neighbours”. For [transition marker], it is as wicked to take away a man’s good name as his life. Homicide is also in giving wicked counsel by fraud, … This is why the wise man says “Feed him who is almost dead from hunger”; for [transition marker], if you do not feed him, you kill him’. (For) soothly here signals the continuation of the original argument and marks or highlights the author’s personal opinion in respect to the quotation. Other frequent collocations with conjunctions, such as but soothly, and soothly or now soothly, also testify to its similarity to conjunctions (for the re­levance of these collocations, see below, Chapter 13).11 Yet, soþlice/soothly also reveal this procedural function when on their own. ME soothly is – similar to listing or summative connectors in Present Day English – employed to introduce or to summarize an argument: (78) … as God seith be the prophete Ezechiel, ‘Ye shal remembre yow of youre weyes, and they shuln displese yow’. Soothly synnes ben the weyes that leden folk to helle (Chaucer, Parson’s Tale, (10) 141; ed. Benson 1987). ‘… as God says through the prophet Ezechiel, “You shall remember your ways, and you will not like them”. [transition marker], sins are the ways that lead people to hell’.

11 Among the prose instances of soothly, the numbers in the corpus passages chosen for the present chapter are as follows: for soothly (31), but soothly (16), and soothly (7) and now soothly (2); soothly alone is used thirty-one times. For the explanatory force of collocations, see Traugott and Dasher (2002: 168).

122   Cognitive source domains In sum, OE soþlice and ME soothly thus follow exactly the path Traugott (Traugott 1995; Traugott and Dasher 2002) has suggested as the so‑called “adverbial cline”: original manner adverbs come to be used as sentence adverbs and finally as adverbial connectors working on the interpersonal or textual level, signalling the continuation of an argument or highlighting an episode or a part of the argument. The adverbs which originally expressed the speaker’s commitment to the truth of the proposition have become non‑propositional. 8.4.5.  ME forsooth(e) A similar pragmatic and semantic history can be seen in ME forsooth(e), which enters the semantic field at the beginning of the Middle English period, even though at that time soothly is still extensively employed. The new item is – like soþlice/soothly – overtly related to the concept of truth (noun and adjective OE soþ, ME sooth), but it is not certain whether it should be analysed as a univerbation of the prepositional phrase for soþe (see OED, s.v. forsooth) ‘in truth, for a truth’ or as a prefixed form of the adjective soþ modified by the intensifying prefix for‑ ‘very true’ (cf. OE formicel ‘very great’, forwel ‘very well’; for a more detailed discussion see Lenker 2003: 276–278 and in particular Lenker 2008). The collocation is only attested seven times in Old English, but becomes very frequent from the thirteenth century onwards, mainly as an emphasizer or intensifier (cf. MED, s.v. forsooth, and Lenker 2003: 274–278). In contrast to the adverbs mentioned above, however, it is also abundantly used in an emphasizer function in both positive and negative answers to questions (see also MED, s.v. forsoth, sense 2b), as in: (79) ‘Ye/Yis, forsothe’, quod I (Chaucer, Boece 3, prosa 11.203; prose 12.57; ed. Benson 1987) (80) ‘Nay/No forsothe’, quod I (Chaucer, Boece 3, prosa 12.150; 4, prose 2.100; ed. Benson 1987). Today’s negative connotations of forsooth, which is now only employed “parenthetically with an ironical or derisive statement” (OED, s.v. forsooth), are probably a consequence of the overuse of these formulae. In Early Modern English, forsooth even came to be considered a superfluous interjection marking the speech of the “lower classes” in London (Lenker 2003: 283–286) and is, accordingly, used for such a depiction of characters in Early Modern English

Source domain truth/fact   123

plays. Consider, for example, the overuse of different emphasizers in the following passages from Middleton’s A Chaste Maid in Cheapside: (81) (Port.) Yes by my faith Mistris, there’s no true construction in that, I haue tooke a great deale of paines, and come from the Bell sweating. Let me come to’te, for I was a Schollar forty yeers ago, ’tis thus I warrant you: (\Matri\) , it makes no matter: (\ambobus parentibus\), for a paire of Boots: (\patri\) , pay the Porter: (\amantissimis charissimis\), he’s the Carryers Man, and his name is Sims, and there he sayes true, forsooth my name is Sims indeed, I haue not forgot all my learning. A Money matter, I thought I should hit on’t (Middleton, Chaste Maid, CEPLAY2B, 2). (82) (All.) Verily you are an Asse forsooth, I must fit all these times, or here’s no Musicke, … (Middleton, Chaste Maid, CEPLAY2B, 27). In Middle English, however, the scope of forsoothe could also be wider and extend over the whole sentence. Accordingly, it is – like soþlice/soothly – found in collocations with conjunctions, such as and and but and is also employed in the procedural, textual functions sketched for soþlice/soothly above. In (83), for instance, and forsothe has a summative meaning and indicates the end of the episode: (83) … And forsothe this foreseide woman bar smale bokis in hir right hand, and in hir left hand sche bar a ceptre. And whan she saughe thise poetical muses … (Chaucer, Boece 1, prose 1.41; ed. Benson 1987). ‘… And [transition marker] this abovementioned woman bore small books in her right hand, and in her left hand she bore a sceptre. [New paragraph] And when she saw the Muses of poetry …’. The polysemous character and the various functions of ME forsothe on different levels may also be inferred from a contemporary metalinguistic scource, the Catholicon Anglicum, dated 1483. This English‑Latin dictionary provides the following list of Latin translations for ME forsothe: Amen, Autem, certe, enim, enion, eciam, equidem, nempe, nimirum, profecte, quippe, reuera, siquidem, utique, vero, vere, quidem, quoque, porro, veraciter, quin, quineciam, quinimmo, quinin, veruntamen (ed. Herrtage 1881: 138).

Only a small number of the Latin equivalents (certe, vero, vere, veraciter, ver‑ untamen) exhibit – like forsoothe – a connection to the concept of truth.

124   Cognitive source domains Only some of them may be employed as emphasizers, while most of them (autem, enim, quidem …) are procedural, i.e. they belong to the group of Latin adverbs for which Kroon (1995) has established a discourse function. While vero ‘truly’, for instance, is procedural in Latin in that it functions on the interpersonal level of communicative acts and moves, autem is a connective on the level of textual organization (Kroon 1995: 371–375). This allows us to infer that a contemporary lexicographer indeed regarded forsoothe as an adverb with a text‑organising function.

8.4.6.  OE treowlice – ME/EModE/PDE truly PDE truly (OE treowlice) also follows this “adverbial cline”, but had to develop an epistemic meaning in a first step and is thus more similar to the items employing the cognitive source domain fact, such as indeed or in fact. In Old English, the adverb treowlice (cf. treowe adj. ‘faithful’) could only be used as a circumstance adverbial, meaning ‘faithfully, loyally, with steadfast allegiance’ (see OED, s.v. truly, 1–2; MED, s.v. treuli, 1–5). In Middle English, we still find a number of examples for this conceptual use, as in (84) He was also a lerned man, a clerk, / That Cristes gospel trewely wolde preche (Chaucer, General Prologue, (1) 480–481; ed. Benson 1987). ‘He was also a learned man, a clerk / who would loyally preach Christ’s gospel …’. (85) This preyere moste be trewely seyd, and in verray feith, and that men preye to God ordinatly and discreetly and devoutly (Chaucer, Parson’s Tale, (10) 1044; ed. Benson 1987). ‘This prayer must be spoken reliably, and in true faith, so that men pray to God properly and discreetly and devoutly’. In the General Prologue, Chaucer describes the manner in which the clerk would preach the gospel as ‘loyally, with steadfast allegiance’, and in (85) the manner in which the prayer is to be said is prescribed by the adverb trewely, but also by the propositional phrase in verray faith (cf. also the following circumstance adverbials ordinatly, discreetly and devoutly). In a more abstract sense, trewely can also mean ‘accurately’ (cf. MED, s.v. treuli, 9), as is evident from the rubric of Chaucer’s Astrolabe which requires the instruments to be set correctly:

Source domain truth/fact   125

(86) To knowe the degrees of longitudes of fixe sterres after that they be determynat in thin Astrelabye, yf so be that thei be trewly sette. (Chaucer, Astrolabe 2.18. Rub.; ed. Benson 1987). ‘To know the degrees of longitudes of fixed stars after they have been placed in your Astrolabe, if it is the case that they are accurately set’. Only by the thirteenth century does trewely become semantically epistemic (cf. OED, s.v. truly, senses 3 and 5; MED, s.v. treuli, sense 11). Its epistemic force is evident in (87) from the Reeve’s Tale, since trewely does not primari­ ly designate that the students are going to pay ‘faithfully’, but indicates their subjective perspective (they are not going to pay!) and marks the speech act to be an (unfulfilled) promise. (87) But specially I pray thee, hooste deere, / get us som mete and drynke, and make us cheere, / and we will payen trewely atte fulle (Chaucer, Reeve’s Tale, (1) 4131–4133; ed. Benson 1987). ‘But we ask you in particular, dear host / bring us some food and drink, and make us merry / and we’ll fully pay for it, that’s for sure’. The newly acquired epistemic meaning and its functional truth value is most obvious in the translation of the Latin Amen, (amen) dico vobis in the New Testament. These formulaic expressions are used when Jesus reinterprets the Old Testament by virtue of his authority as the Son of God and they thereby require an epistemic value. While the Old English texts translate these by soþ ic (eow) secge or soþlice/witodlice ic (eow) secge (West‑Saxon Gospels), the Earlier and Later Wycliffite versions use treuli I seie (see Matthew XXVI.13.21.34; Mark XIV.9.18.25; Luke XXII.59; John XIII.16.20.21). At this stage, the functional variability and polysemy of trewely is most clearly seen in phrases such as loven trewely, in which the adverb may function as a manner adjunct, emphasizer or sentential adverb. In the examples taken from Gower’s Confessio Amantis, only the word order and collocation with the conjunction for distinguish the uses which in modern editions are then highlighted by punctuation, i.e. the comma in (89). (88) Sithe I have trewly loved on (Gower, Confessio Amantis, 3.66; ed. Macaulay 1900–1901). ‘Since I have indeed loved one’. (89) For trewli, fader, I love oon (Gower, Confessio Amantis, 5.2536; ed. Macaulay 1900–1901). ‘For indeed, father, I love one’.

126   Cognitive source domains Finally, trewely is also found in the same collocations as soothly – namely with the conjunctions and, for and but trewely – and is also similar in its function of highlighting the connective function of the coordinators for and and which it collocates with: (90) And ye shul understonde that looke, by any wey, whan any man shal chastise another, that he be war from chidynge or reprevynge. For trewely, but he be war, he may ful lightly quyken the fir of angre and of wratthe, which that he sholde quenche … (Chaucer, Parson’s Tale, (10) 628; ed. Benson 1987). ‘Understand, in any case, that when a man has to chastise another, he should beware lest he chide or reproach him. For [transition marker], unless he be wary, he may very easily kindle the fire of anger and wrath, which he should quench’. (91) … the which a long thyng were to devyse. / And trewely, as to my juggement / me thynketh it a thyng impertinent / save that he wole conveyen his mateere; / But this his tale, which that ye may heere (Chaucer, Clerk’s Prologue, (4) 53–56; ed. Benson 1987). ‘… it would take a long time to describe this thing to you. / And [transition marker], in my mind / it seems to be an irrelevant thing / save that he wanted to introduce his material; / But here is his tale which you may hear (now)’. The adverb became very popular in the course of the Middle English period (Swan 1988: 255), as can be seen in many examples from London English, such as (92) … & I wene ther wer a‑boute a xxx craftes, & in Chepe they sholden haue sembled to go to a newe eleccion, &, truly, had noght the aldermen kome to trete, & maked that John Norhampton bad the poeple gon hoom, they wolde haue go to a Newe eleccion … (CMDOCU3, The Ap‑ peal of Thomas Usk, 28). ‘… and I guess that there were about 30 guilds, and in Cheapside they should have assembled to go to a new election, and, [transition marker], had not the aldermen come to make terms and caused that John Norhampton bade the people to go home, they would have gone to a new election’. (93) … & that al strange vitaillers sholden with thair vitailles … frelich kome to the Cite, to selle thair vitailles as wel be retaile as in other wyse,

Source domain truth/fact   127



hauyng no reward to the Franchise. And, truly, the ful entent was that al the ordinances that wer ordeyned in hys tyme … (CMDOCU3, The Appeal of Thomas Usk, 24). ‘… and that all foreign traders in food‑stuffs should come with their victuals freely to the City [London], to sell their victuals by retail‑trading as well as in other ways, having no regard of the freedom from tax in London. And, [transition marker], the full intention was that all the commissions that were commissioned in his time …’.

OE treowlice thus shows an identical line of development as the adverbs discussed above. The major difference, however, is that treowlice first had to change its propositional meaning. It is important to note, however, that – with all of these adverbs – the various meanings do not replace each other immediately – language change necessarily involves polysemy, and loss of original meaning is relatively rare (Traugott and Dasher 2002: 280–281). Thus the old and new meanings may coexist in the same text, a property commonly referred to as “layering”. 8.4.7.  PDE indeed and in fact As has been mentioned above, the adverbs analysed here are semantically and functionally similar to PDE indeed and in fact.12 These lexemes share the expression of factuality in their original propositional meaning and only acquired epistemic meaning (veracity) in the course of their history. The development of indeed from a bare prepositional phrase with propositional meaning (‘in the deed/act’) to its procedural, textual use ‘what’s more’ has been repeatedly discussed by Traugott and may be summarized as follows (cf. Traugott and Dasher 2002: 165). In stage II, indeed develops an epistemic meaning and becomes semantically similar to soothly, forsoothe and truly: Stage I: PP ‘in action / in practice’ 1300–1850 ‘in the act’ > ‘in truth’ Stage II: epistemic uses 1450‑ ‘in truth’ > ‘in addition’ Stage III: procedural, textual uses 1600‑ ‘what’s more’ The “adverbial cline” is also identifiable in the history of in fact, originally a prepositional phrase with the head fact (borrowed from Latin in the sixteenth 12 I here exclude actually (included in Lenker 2007b) because its functions as a text­ ual connective are not as clear‑cut (see also Lenk 1998: 155–188).

128   Cognitive source domains century), which develops procedural uses, first as an epistemic adversative and, finally, as a transitional connector (cf. Traugott 1999; Traugott and Dasher 2002: 165–169). Stage I: Stage II: Stage III:

adverbial of “respect in which” epistemic, adversative uses procedural, textual uses

(1670) (1680) (1815)

8.4.8.  Regularities in change Apart from the shifts from one semantic domain to another summarized above, all the lexemes analysed here share the pragmatic‑semantic tendencies which Traugott and Dasher (2002) have established for adverbs with a procedural function as a “regularity in semantic change” (Traugott and Dasher 2002: 281). Table 8.5.:  Regularities in semantic change (after Traugott and Dasher 2002: 281) non‑subjective content scope within proposition truth‑conditional

> subjective > content‑procedural > scope over proposition >

> intersubjective > procedural > scope over discourse > non‑truth‑conditional

8.4.9.  Truth, facts and communicative principles In their conclusions, Traugott and Dasher (2002) state that adverbs of an origin­ al meaning ‘in fact, deed, action’ which refer to factuality seem to provide the “appropriate semantics” of a lexeme to undergo the recurrent changes of the “adverbial cline” (Traugott and Dasher 2002: 281). I have argued elsewhere, however, that I would like to go one step further and suggest that the semantics of these adverbs are not only “appropriate” but that the original propositional meaning of these lexemes almost inevitably leads to pragmatic strengthening, to their use as an emphasizer and, eventually, as a connector (see Lenker 2007b). I rest this view on one of the seminal concepts in pragmatics, Grice’s Co‑operative Principle and in particular his Maxims of Quality (Grice 1989: 26–27):13 13 For a survey of the principle, the maxims, revisions, applications and problems see Levinson (1983: 100–166; 2000: 12–21) and Mey (2001: 68–88).

Source domain truth/fact   129

1 Do not say what you believe to be false. 2 Do not say that for which you lack adequate evidence. Obviously, these maxims are directly connected to the adverbs analysed in this chapter: while Quality (1) is related to the adverbs meaning ‘truly’, Quality (2) clearly encompasses those with an original meaning of ‘in fact, in deed, in action’. If we accept the most basic of Gricean assumptions, namely that all partners in a communicative act want to be cooperative, then the Maxim of Quality implies that we should only talk about the veracity or factuality of a propos­ ition if speakers/writers or listeners/readers doubt it, i.e. in highly emotional or negative contexts. Speakers/writers may thus want to mark their speech acts as performative (see the “promise” in (87)) or explicitly express their commitment to the truth of the proposition because they fear the non‑acceptance of the interlocutors (cf. the many instances of soþlice, soothly and verily with verbs of communication). This gives rise to the conversational implicatures of concession as shown above in Chapter 8.4.2. Further, the speakers’ reference to truth could be cooperative in negative contexts or when they want to mark a hedge, such as in I’m not sure whether it is true but … or as far as I know …. In many of the instances quoted above – apart from some manner adjunct and emphasizer uses – none of these contexts are given. Stating the veracity or factuality of a proposition by means of the adverbs in question would thus seem superfluous and therefore un‑cooperative. The recurrent use of these adverbs would hence also impair the Maxim of Quantity (“Do not make your contribution more informative than is required”) and the Maxim of Relation (“Make your contribution relevant”).14 Yet, according to Grice’s hypothesis, when a conversation does not proceed according to the maxims, hearers/readers assume that, contrary to appearances, the principles are nevertheless being adhered to at some deeper level (cf. Levinson 1983: 102). In cases of this sort, inferences arise to preserve the assumption of co‑operation. Relying on the Maxim of Quality, listeners/readers will principally assume that speakers are telling the truth. If speakers/writers choose to assert the veracity or factuality explicitly by using a sentential adverbial such as soþlice, forsoothe, trewely, indeed or in fact, the invited inference of the hearer/listener is that the meaning of these items stating the truth or factuality has to be found on another than the propositional or lexical level. 14 This Maxim of Relevance (“Be relevant”) is considered the only important one in the cognitively oriented approach by Sperber and Wilson (1996). For a survey and criticism of the concept, see Levinson (2000: 54–64).

130   Cognitive source domains The first level which is to be expected for these truth‑intensifying adverbs are their concessive/correlative uses (in the correlative patterns surely … but) or their epistemic use implying increased subjectification, when speakers and hearers agree in their conceptualization of the adverbs as emphasizers: the speaker wants to stress the assertion of another phrase or clause by explicitly pointing to its truth value. In case this level does not provide the expected meaning in accordance with the Co‑operative Principle because the propos­ ition and the grammatical context (inanimate subject, subjunctive mood) will not allow a meaning of “truth” or “factuality”, hearers will fully understand that the speaker’s commitment to truth does not relate to the proposition but again to a different, deeper level, i.e. to the organization of discourse itself. The adverbs are employed and will be understood as connectors, because otherwise the interlocutors would violate the Maxim of Quality.15 Once they are predominantly interpreted on this level of textual organization, however, they have lost much of their force as truth‑intensifiers on the manner‑adverb level and have to be replaced by new lexemes, which may – in due time – follow the same cline. In sum, I would like to suggest that – in addition to the source domains time and space – truth/fact should be added as a third domain from which adverbial connectors are recurrently coined.16

15 Levinson (1983: 100–101) deduces the notion of cooperation from “general considerations of rationality” which might consequentially mean that we are dealing with a universal here; cf., for example, Latin concessive verumtamen ‘nevertheless’. 16 For the differences which distinguish these transitional connectors from those coding semantic relations, see below, Chapter 12.

9. Shifting deictics in the history of English causal connection In the following Chapters 9 to 11, a closer inspection of the development of adverbial connectors of the different semantic relations will show that adverbials for the different relations were affected in different ways. First of all, we have to draw a distinction between what the Cambridge Grammar calls “pure” vs. “impure” connectors, i. e. those connectors which have no other function than that of connecting their clause to the surrounding text in contrast to those which combine this connecting function with a function of concession, condition or reason/result (Huddleston and Pullum 2002: 775). This distinction corresponds to the group of relations which has been established as a typical goal domain above (see Chapter 8) and which is singled out as the set of cccc‑relations in cross‑linguistic, typological, and grammaticalization studies (cause, condition, concession, contrast). For the adverbial connectors, however, only cause/result (present chapter) and concession/contrast (Chapter 11) are relevant. For reasons of information structure and hierarchy (see above, Chapter 2.4), the relation of condition is – also cross‑linguistically – not signalled by adverbial connectors, which express an equal weight of both connects (and not the subordinate‑superordinate hierarchy inherent to if … then relations). The semantic relation of addition (Chapter 11) and the marginal one of transition (Chapter 12), on the other hand, are for the same causes not expressed by subordination, since they – as pure coordinators – merely signal the additive relation of sentences or chunks of discourse explicitly. Additional connectors thus only highlight a relation already present and are not obligatory for an understanding of the texts. They are signposts facilitating the reading process in registers such as academic prose, specifically trying to mark all logical relations overtly. Adverbial connectors signalling the relations cause/result and contrast/concession, on the other hand, are much less genre‑dependent (see above, Chapter 3.5). These basic differences are mirrored in the developments of the various adverbial connectors for the specific relations: While the changes in additional and transitional connectors mainly concern replacement of lexical elements and reflect changing stylistic or rhetorical ideas (see below, Chapters 11 to 13), the developments in the ccc‑relations are much more deeply rooted and reflect the typological and structural changes pointed out above. This will first be shown by a closer inspection of the developments in the semantic register of cause/ result, which prototypically shows all of the tendencies sketched so far: the

132   Shifting deictics in the history of English causal connection path from one central pronominal connector, which is an ambiguous adverb/ conjunction (OE forþæm), to a differentiated system of conjunctions and adverbial connectors from Early Modern English onwards. It also reflects the shift of the expression of discourse deixis from pronominal to, predominantly, spatial frames of reference. 9.1.  Earlier research on causals The following exemplary investigation provides an analysis of the most frequent connective relation in texts, the relation of cause/result/inference (for the numbers in Present Day English, see Biber et al. 1999: 880; Figure 10.25).1 The present study aims to illustrate that a detailed investigation of the instabilities in the systems of Middle and Early Modern English causal connectors allows for a better understanding of two peculiarities of the system of Present Day English causal connection and its history, namely the remarkable lack of an adverbial connector marking the relation result → cause (cf. Latin nam, enim or German nämlich, deswegen) and, secondly, the late triumph of the now central causal connector because after 1750. The focus will not be on single connectors (as because in Higashiizumi 2006), or on morphological (as in Liggins 1955; Kivimaa 1966; Rissanen 1997) or genre aspects (as in Claridge and Walker 2002), but on systematic changes in the forms of clausal connectors after the Old and Early Middle English period which, in specific contexts, triggered the coinage of new causal connectors to avoid ambiguities and facilitate information processing. The focus of the chapter is therefore on a closer inspection why some of the forms emerging in early Middle English, such as the adverbial connectors for‑ whi (ME1 to ME3) or for that (ME1 to ME3) or the subordinator for as much as (ME2 to EModE2) were rejected while others, such as hence or because, have survived.2 The changes specific to English become particularly evident when we compare the various systems of causal connectors in the history of English with Present Day German and consider the implications of Modern German 1 In contrast to most other relations in clause linkage, expressions for causality have been examined quantitatively in a number of synchronic (see Altenberg 1984, Biber et al. 1999) and diachronic corpus studies (see Liggins 1955 for Old English; Rissanen 1997 for the history of causal connectors in English; Markus 2000 for Middle English and Claridge and Walker 2002 for Early Modern English). 2 This chapter is a – slightly revised and updated – version of Lenker (2007a). I would like to thank John Benjamins for granting permission to print this chapter in revised form.

Causal connectors   133

and Old English – but not Old English and Modern English – being typologically close to one another. Such an approach requires the close examination of causal connectors and their textual functions in comparable texts. In addition to the material for adverbial connectors collected from the corpus texts in Appendix B.3, this chapter thus investigates all causal connectors – i. e. not only adverbial connectors, but also coordinating and subordinating conjunctions – in the singular argumentative text extant for all periods of English, the adaptations and translations of Boethius’ De Consolatione Philosophiae in King Alfred’s Circle (OE; ed. Sedgefield 1899), by Chaucer (ME; ed. Benson 1987), Colville (EModE; ed. Bax [1556] 1897), Queen Elizabeth I (EModE; ed. Pemberton 1899) and Preston (1695) (EModE) and in Present Day English and German translations (Watts 1969 and Neitzke in Grasmück 1997), in the representative prose selections of the Helsinki Corpus (HC). These findings are cross‑checked and supplemented by an analysis of causal connection in Byrhtferth’s Old English Enchiridion (ByrM; ed. Baker and Lapidge 1995), a text relatively independent from Latin, as well as by information elicited from dictionaries (DOE, MED) and machine‑readable corpora, in particular the Dictionary of Old English Corpus (DOEC) and the Corpus of Middle English Prose and Verse (CME). 9.2.  Causal connectors 9.2.1.  The relation cause: cause – result vs. result – cause The category of cause can be further split into causal relation on the one hand (PDE because, since) and the relation of result (PDE therefore, so that) on the other hand. First, it is important to note that all connectors form complex propositions. Thus a sequence of two propositions – (94) proposition A: John is ill and proposition B: John won’t come tonight – becomes a complex proposition “if a sequence of two propositions A and B expresses a new thought on a level other than that of the isolated propositions” (Rudolph 1989: 176). In contrast to asyndetic sequences which need not, but may express a causal relation, such as (94a) John won’t come tonight. He is ill or (94b) John is ill. He won’t come tonight, causal connectors commonly have a two‑fold function: they a) connect two states of affairs, and at the same time, b) convey the speaker’s opinion on the configuration of these state of affairs, as in (95) John won’t come tonight because he is ill.

cause

(96) John is ill so that he won’t come tonight.

result

134   Shifting deictics in the history of English causal connection While the first complex sentence (95) marks a causal relation, (96) reverses the sequence of information and relates the same state of affairs by marking the relation as result. The same relation of result can also be expressed by the employment of an adverbial connector, as in (97) John is ill. Therefore he won’t come tonight.

result

In Present Day English (and also in Old English, see below, Tables 9.6 and 9.7), there is a (slight) preference for the linear sequence cause – result by both speakers and writers (55 per cent and 52 per cent; see Altenberg 1984: 52). 9.2.2.  Present Day English causal connectors: corpus findings In Present Day English, because (with its reduced forms cos/coz) is by far the most common causal connector, marking over a third of all relations of cause/ result (see Altenberg 1984: 40–45 on the basis of the LOB and the LLC, and also Biber et al. 1999: 836, 887). With 45 per cent of all instances in the LLC (Altenberg 1984: 45), because is by far the most frequent connector in the spoken medium. In contrast to this stereotyped coding in today’s spoken English, the relation cause is marked in a lexically and grammatically much more varied form in the written medium, ranging from asyndesis (Sue won’t come tonight. She is ill) to syndesis by a variety of explicit linkers. In addition to conjunctions (PDE because, for, since or as), we find adverbial links in the form of prepositional phrases (PDE for that reason) or adverbial connectors (PDE therefore, hence) as well as a number of clause‑integrated expressions such as PDE the reason is or the result is. Of the altogether ninety-eight lexical, grammatical and positional subtypes recorded in LOB and LLC (Altenberg 1984: 39), only subordination by conjunctions (53 per cent) and adverbial connectors (31 per cent) are highly frequent (in contrast to 8 per cent each for prepositional phrases and clause‑integration; Altenberg 1984: 40–44; see also Biber et al. 1999: 842, 887). For this reason, the present chapter will concentrate on the two central ways of causal connection, conjunctions and adverbial connectors, and will not further consider lexical means or non‑univerbated prepositional phrases. 9.2.3.  Causal connectors: word classes and topology The major means of expressing the relation cause → result (examples (98), (99), (101a), (102a), (103a)) and result → cause (examples (100), (101b), (102b),

Causal connectors   135

(103b)) are illustrated in Table 9.1 not only by Present Day English but also by Present Day German examples, because German is typologically much more similar to Old English than Present Day English and, more importantly, because German employs means which are no longer used or common in Present Day English (see also above, Chapter 2, Table 2.2). German uses adverbial connectors coding the relation result (R) – cause (C) (PDG nämlich; see (100)) and correlative constructions (see (103a), (103b)), which are not generally common on the sentential level in Present Day English (*Because we are stronger, we will/are going to therefore win; *We will/are going to therefore win, because we are stronger). Since all of these patterns were possible in Old English (see below, Table 9.5), Table 9.1 thus also attests to changes and subsequent gaps in the system of causal connection in the history of English. Table 9.1.:  Causal connectors in Present Day English and Present Day German A. Parataxis (98) Sie ist stärker. Deswegen wird sie gewinnen.

[V2; post‑posed]

C→R

(99) She is stronger. Therefore she is going to win.

[post‑posed]

C→R

(100) Sie wird gewinnen. Sie ist nämlich stärker.

[V2; post‑posed]

R→C

B. Hypotaxis (101a) Weil sie stärker ist, wird sie gewinnen. (102a) Because she is stronger she is going to win.

[V‑final; pre‑posed] C→R [pre‑posed] C→R or

(101b) Sie wird gewinnen, weil sie stärker ist. (102b) She is going to win because she is stronger.

[V‑final; post‑posed] R→C [post‑posed] R→C

C. Correlatives (103a) Weil sie stärker ist, deswegen wird sie gewinnen. [V‑final + V2]

C→R

(103b) Sie wird deswegen gewinnen, weil sie stärker ist. [V2 + V‑final]

R→C

The acknowledged traditional criteria for distinguishing these different types of connectors are topological: the position of the connector, the sequence and position of the respective connected elements, the possibility of collocations of connectors and, in German, word order. In contrast to Modern English, Present Day German also differentiates main clauses from subordinate ones by employing verb‑second for the former and verb‑final word order for the latter (for details, see above, Chapter 2.3). The focus on topological criteria such as word order, sequence and collocations is still persistent in most grammars and publications on the subject, as, for

136   Shifting deictics in the history of English causal connection instance, in the Handbuch der deutschen Konnektoren (Pasch et al. 2003), which considers topological criteria only (see, e. g., the terms Postponierer ‘postponers’ or Verbzweitsatzeinbetter ‘V2‑embedders’). For Present Day English, Quirk et al. (1985) establish six, predominantly topological, parameters similar to the ones illustrated in Chapter 2.3 above; these yield a “coordination – subordin­ ation gradient” (Quirk et al. 1985: § 13.18). In addition to three discrete poles – coordinators (such as and and or), conjuncts (adverbial connectors such as however and therefore) and subordinators (such as if or because) – there are also forms, among them the causal connectors for and so that, which are situated in the middle of this gradient, sharing three parameters with paratactic and three parameters with hypotactic connectors. This has led to much uncertainty or even confusion in particular about the status of PDE for, which is classified as a subordinator rather than a coordinator but is said to be “more coordinator‑like than the more typical subordinators if and because” (Quirk et al. 1985: § 13.19), mainly on the grounds that clauses with initial for – unlike those headed by because and since – are restricted to post‑position in Present Day standard English.3 Such a discussion of the niceties of the coordination‑subordination gradient turns out to be rather elusive (see also section 9.3.1) and does not really get to the heart of the matter, since a number of recent publications on causal connection have shown that an analysis of connectors should choose a wider perspective and should, in addition to topological and syntactic criteria, also consider aspects of semantics, pragmatics and text linguistics (see, for example, Thompson 1985; Diessel 2005 and in particular the studies collected in Lefèvre 2000). 9.2.4.  Semantic and pragmatic parameters For the distinction of subgroups of causal clauses, a number of semantic and pragmatic parameters are employed. One set of distinctions differentiates three major semantic‑pragmatic sub‑groups, separating external reason clauses (cause based in the external reality) from internal ones (cause based in the speaker’s world of reasoning) and rhetorical ones (cause based in the speech act) (for these distinctions see, e. g., Halliday and Hasan 1976: 238–244; Quirk et al. 1985: 15.20, 15.45–15.47; Sweetser 1990: 76–84; Traugott 1992: 252; Higashiizumi 2006). 3 Thus Altenberg (1984: 41, footnote 3) says he follows Quirk et al. (1985) in classifying for as a subordinator. Kortmann (1997: 331), on the other hand, says that for “qualifies as a coordinator in current English”. For such a highly frequent element as for, these differences in classification obviously cause major problems for the comparative value of quantitative data for coordination and subordination (see also Rissanen 1997; Claridge and Walker 2002).

Causal connectors   137

Because, the central causal connector of Present Day English, may be employed for all three subtypes: First, because may give an inherent objective connection in the real world, such as, for example, physical causes and their effects (“external reason clauses” marking real‑world causality), as in (104) The flowers are growing so well because I sprayed them. Secondly, because may also – like PDE since, for, as and now that – give the speakers’ inference of a connection and signal their way of presenting arguments (“internal reason clauses”; epistemic because), as in (105) He must be here because his bicycle is outside (meaning “The reason I think he is here is that his bicycle is outside”). Thirdly, the reason given need not be related to the situation in the matrix clause but is a motivation for the implicit speech act of the utterance (“rhetorical or speech act because”), as in (106) Percy is in Washington, because he phoned me from there (meaning “The reason why I say this is that he phoned me from there”). While because may be employed in all three functions in today’s English, other conjunctions are more restricted in their use: PDE since and as are internal “explanation causals”, as can be seen from the fact that they do not allow why‑questions or cleft‑sentence constructions (for this terminology and the criteria, see Quirk et al. 1985: § 15.20–15.22, 15.45–15.47). Similarly, for also functions internally as explanation and ground rather than assertion of a true causal relation in the external reality (Quirk et al. 1985: § 15.45). Many languages are more rigid in specifying these different functions, at least in the written standard, obligatorily by distinctive connectors (cf. Kroon 1995: 10–17). For the present investigation of causal connectors in translations of Boethius’ De Consolatione Philosophiae, the data for Latin and French (one of Chaucer’s exemplars by Jean de Meun) are of particular interest: Table 9.2.:  Cross‑linguistic taxonomy of causal relations external (sociophysical) Latin French German PDE

quia; quod parce que weil because

internal (epistemic, conversational) explanation justification quoniam nam; enim puisque car da denn; nämlich because; because since, as; for

138   Shifting deictics in the history of English causal connection 9.2.5.  Information processing With respect to the states of affairs, there is no difference whether the relation of causality is expressed by a causal relation (95) or a relation of result as in examples (96) and (97); (the examples are repeated here for the sake of clarity). (95) John won’t come tonight because he is ill.

cause

(96) John is ill so that he won’t come tonight.

result

(97) John is ill. Therefore he won’t come tonight.

result

While the causal constant A→B remains the same in all cases, the differences between the various means of marking causal connection are to be found in the speaker’s choice of information structure and therefore in the intended and/or highlighted “aim of the message”. In the sentence highlighting the causal relation (95), the speaker’s interest is focussed on the first connect John won’t come tonight, reflecting the main information and single focus of this sentence. This information is thus presented in the main clause. As for the relation of result (96), the speaker’s interest is focussed on John is ill. Therefore this information is presented in the main clause, from which the speaker looks to the result B (He won’t come tonight). These differences are highly important for the interpretation of complex sentences, but even more so for the general organization of the text (for details, see also above, Chapter 2.4). Since both these complex sentences only have one focus, subordinate clauses are not only syntactically but also semantically subordinate (see Quirk et al. 1985: § 13.3). They thus work on a local level of textual connection, but with different functions for pre‑ and post‑posed subordinate clauses. Initial clauses present “given information” – an interpretation which is often emphasized by marking cohesion also lexically; cf. hillsides and hilly in (107) – and state a problem within the context of expectations raised by the preceding discourse: (107) The houses were perched precariously up the hillsides Because it was so hilly the area seemed constantly to be in a dark blue haze. (example from Biber et al. 1999: 835, FICT). Final adverbial clauses (cf. (96)), on the other hand, play an even more local role and state the cause for the action named in the immediately preceding clause. In a recent study, Diessel (2005) shows that with regard to parsing or ut-

Causal connectors   139

terance planning, complex sentences are easier to process and thus preferred if the adverbial clause follows the main clause. Only rarely, discourse pragmatic impulses override this processing preference for final position. In Diessel’s corpus, only scientific articles exhibit a substantial number of pre‑posed causal clauses, causal clauses in this genre being often used to provide a common ground for a subsequent conclusion (Diessel 2005: 465). Across all genres, however, most of the causal clauses follow the main clause: only 1 per cent of all causal clauses in conversation are initial, 10 per cent in fiction, and 27 per cent in scientific writing (Diessel 2005: 454). As we might have expected from these considerations, the numbers are very similar for Old English: if we include correlative constructions, the Old English adaptation of Boethius’ Consolatio attests ca. 12 per cent pre‑posed clauses (thirteen out of 109 instances; see below, Table 9.6). The text chosen as an example for scientific writing in Old English, Byrhtferth’s handbook on astronomy, has about 21 per cent pre‑posed clauses (fifteen out of seventy‑one instances; see below, Table 9.7). In contrast to the single focus of complex clauses comprising a subordinate causal clause, the result construction by means of an adverbial connector in (97) consists of two main clauses (John is ill. Therefore he won’t come tonight) and hence has two “foci”, i. e. two separate information blocks. In this view, adverbial connectors are – in contrast to subordinators – strong indicators of a great illocutionary weight of the second connect, whose proposition may then be pursued in the following discourse. This independent focus may be signalled by syntactic means, such as verb‑second (main clause) word order in Present Day German and, more generally, by prosodic means (see above, Chapter 2.4). Subordinate clauses, on the other hand, are syntactically and prosodically integrated: the whole complex sentence only has one intonation contour for only one thematic structure (see Wegener 2000: 36, and Givón 2001: 327 for English). To avoid ambiguities, the connection by means of an adverb marks the relation most prominently on the surface level, and is thus highlighted by transparent lexical adverbial connectors such as PDE consequently or PDG nämlich or by means of transparent deictic pronominal adverbial connectors such as PDG deshalb ‘for; therefore’ (PDG article, genitive singular. masc. des + lexical element halb ‘because of’; cf. PDG halber ‘for the sake of’). Because of this strong independent illocutionary weight, adverbial connectors may not only be used to link clauses but also whole chunks of discourse. In this context, two patterns which may cause ambiguity are particularly relevant for the history of causal connectors in English. Firstly, pre‑posed subordinate clauses which state “given information” might – in the spoken medium – be easily mistaken for post‑posed ones or for adverbial connectors

140   Shifting deictics in the history of English causal connection introducing a causally related new main clause (and are probably for this reason rather infrequent). Similarly, adverbial connectors also have to be as unambiguous as possible to signal overtly that the following main clause is on some level connected to the preceding clause or discourse and forms a complex proposition (external reason clause) or is an explanation or justification (in an internal causal relation); adverbial connectors have to be as unambiguous metalinguistic indicators of the text structure as possible, since they explicitly state the voice of the speakers, their opinion on the state of affairs. The following account will illustrate that changes in the system of English causal connectors have been induced by exactly these factors of information structure (see the emergence of for as moche as, since and because) and may themselves, on the other hand, also lead to changes in the options of information processing (see the restricted use of for in Present Day English). For a first survey, the central causal conjunctions and adverbial subordinators of Present Day English are summarized in Table 9.3. As has been mentioned above, Present Day English remarkably – in contrast to most other languages – lacks an adverbial connector for the sequence result → cause (cf., for instance, Latin nam and enim or German denn or nämlich). Table 9.3.:  Present Day English connectors: cause/result Conjunctions cause: Conjunctions result: Adverbial Connectors result:

external true reason clauses: because internal explanation clauses: since, as; for; now that so that accordingly; consequently; hence; so; then; therefore; thus

9.3.  OE forþæm, forþon, forþy 9.3.1.  Forms and functions of forþæm, forþon, forþy When we compare the Present Day English system of causal connectors with the system in Old English, it becomes evident that all Present Day English connectors are new coinages or developments in the history of English. Old English has only one central, highly polysemous and polyfunctional connector marking the semantic relation cause or reason, which appears in the forms forþæm (preposition for + þæm dative singular masculine/neuter demonstrative pronoun), forþon (preposition for + þon instrumental singular masculine/ neuter demonstrative pronoun) and forþy (preposition for + þy instrumental singular masculine/neuter demonstrative pronoun).

OE forþæm, forþon, forþy   141

The Dictionary of Old English lists these three forms as variants in one single entry and counts altogether about 15,500 occurrences in a wide variety of spellings, which do, however, not carry distinguishing force (see DOE, s. v. for‑þǣm, for‑þon, for‑þȳ).4 The item for‑þæm (þe) is in all these orthographic variants found in slots which in Present Day English are filled by adverbs (‘therefore’; A, examples (108) and (109)) or conjunctions (B; examples (110) and (111)), but which could also be realized by so‑called “correlative constructions” (C; examples (112) and (113)). Table 9.4 (structured in A, B, C to allow comparison with Table 9.1) gives a first survey with prototypical examples taken from Byrhtferth’s Enchiridion (ByrM; ed. Baker and Lapidge 1995). Table 9.4.:  Functions of OE forþæm A: CAUSE – RESULT (cf. “PDE Adverbial Connector” therefore) (CAUSE) Forþon VSA (= RESULT) (108) (Se soþlice … byð niwe … geendað xxix on v kalendas Septembris). Forþon byð niwe mona on IIII kalendas September … . ‘Therefore is new moon on 29 August … ’ (ByrM 2.2.137). (CAUSE) Forþan SVA (= RESULT) (109) þas þing we gemetton on Ramesige þurh Godes miltsigendan gife. Forþan ic ne swigie for ðæra bocre getingnyssum ne for þæra gelæredra manna þingum þe … . ‘We found these things at Ramsey through God’s merciful grace. Therefore I shall not be silent either on account of the eloquence of the literate or for the sake of those learned men who …’ (ByrM 1.1.158). B. RESULT – CAUSE (cf. “PDE Conjunction” because/for/since/as) (RESULT) forðan SOV (= CAUSE) (110) Ðas þing we swa hwonlice her hrepiað on foreweardum worce forðan we hig þenceað oftor to hrepian and to gemunanne. ‘We discuss these things so briefly at the beginning of this work because we intend to discuss and recall them more often’ (ByrM 1.2.250). (RESULT) forðon þe SVA (= CAUSE) (111) on þam feorðan geare he hæfð nigon and twentig, forðon þe an dæg awyxst binnan feower wintrum … . ‘in the fourth year it has twenty‑nine, because one day grows over four years …’ (ByrM 2.1.17).

4 Rissanen (1997: 393) speaks of “at least eight different forms”. Since he does not really find consistent chronological or dialectal criteria for their distinction, I here follow the DOE, Mitchell (1985: §  3010–3051) and Traugott (1992: 252–254) in regarding these forms as variants of one single connector.

142   Shifting deictics in the history of English causal connection C. Correlative Construction: RESULT – CAUSE, CAUSE – RESULT forþon SOVA (=RESULT) forþam SVC (= CAUSE) (112)  Forþon Romani hine gelogodon on þissum monðe (þæt ys on Februario) forþam he ys scyrtest ealra monða … . ‘The Romans placed it [therefore] in this month (in February) because it is the shortest of all months’ (ByrM 2.1.36). forþon SVC (= CAUSE) forþon SVO (= RESULT) (113) … and Februarius, forðon he ys se læsta and he hæfð twegra daga læs þonne þa oðre, forþon he forlæt and feowertig tida. ‘… and February, since it is the shortest and has two days fewer than the others, looses therefore forty‑eight hours’ (ByrM 2.1.372).

These instances illustrate one of the most striking features of causal connection in Old English: the syntactic distinctions between coordination and subordination as well as the semantic distinctions between internal and external reason clauses are difficult, if not impossible, to establish on morphological or syntactic grounds (Mitchell 1985: § 3007–3051; Traugott 1992: 252). There are a number of studies which have tried to establish different criteria for coordination and subordination on a large textual basis (see, e. g., Liggins 1955, who analyses all Old English prose texts), but all of these finally had to conclude that, apart from a few tendencies observable in certain authors, Old English did not draw a distinction between coordinate and subordinate causal clauses (Liggins 1955: 205; Mitchell 1985: § 3013–3015; Stockwell and Minkova 1991; Donoghue and Mitchell 1992), neither by a consistent use of different connectors or different forms of only one connector, nor by differences in word order (verb‑second vs. verb‑final), nor by the appended subordinating particle þe (see, in particular, Kivimaa 1966: 157).5 Further, Anglo‑Saxon “linguists” themselves do not seem to have felt a strong need to differentiate between coordination and subordination. In his grammar, Ælfric makes no attempt to distinguish coordination from subordination and thus draws no distinction between coordinators, subordinators and adverbial connectors in his section “De Coniunctione” (Zupitza [1880] 2001: 257–266; see also above, Chapter 4). Traugott rightly summarizes this dilem5 The number of instances of forþæm þe (i. e. with addition of the particle þe) increases towards the end of the Old English period, also showing a higher correlation of forþæm þe with subordination (particularly in the writings of Ælfric). This could be seen as “a developing consciousness of subordination” (Rissanen 1997: 394), but there is still such a wide functional variety that it cannot be said to be decisive for Old English in general.

OE forþæm, forþon, forþy   143

ma of anachronistic classification of causal connectors by pointing out that “it is usually assumed that the ‘because’‑clause is subordinate in OE, largely because the equivalent clause‑type in PDE is subordinate” (Traugott 1992: 253). Forþæm and its variants thus belong to those Old English items which are traditionally called “ambiguous adverbs/conjunctions” (see above, Chapter 5.3). Except in a few cases (such as answers to questions), there is no way of distinguishing the various functions on formal grounds on the basis of our extant Anglo‑Saxon material, because this – in contrast to material from ­later periods  – does also not provide orthographic clues such as capitalization or punctuation or indications of the prosodic differences mentioned above (though there were most certainly such distinctions in spoken Old English; see Mitchell 1985: § 3015). In sum, there are no unambiguous criteria such as individual forms of the connector or word order: forþæm and its variants may be employed in all kinds of sequences for a number of different relations. In contrast to the wide variety of forms of Present Day English, Old English thus virtually only employs one form to mark the various causal relations.6 Table 9.5.:  Distribution and functions of OE forþæm result – cause

cause – result

R for þæm (þe) C for þæm R (þe) C … for R þæm … (þe) C for þæm R for þæm (þe) C

C forþæm R forþæm (þe) C R C … for R þæm

9.3.2.  Expressions for causal relations in Early and Late West Saxon For a survey of quantitative findings showing the central status of the pronominal forþæm and its variant forms, I will now briefly summarize the findings for all items marking causal relations in two Old English texts which are comparatively independent of Latin originals, namely the Early West Saxon adaptation of Boethius’ De Consolatione Philosophiae (COBOETH), thought to be written in the circle around King Alfred (end of ninth century; for serious doubts about King Alfred as the translator, see Godden 2007) and Byrhtferth’s Late West Saxon Enchiridion (ByrM), a handbook mainly on astronomy (around 1000, after 996; see Baker and Lapidge 1995: xxvi). 6 The following table is compiled from Mitchell 1985: § 3010–3051, Traugott 1992: 252–255, Wiegand 1982 and my own material.

144   Shifting deictics in the history of English causal connection In the passages of the Old English Consolatio selected by the Helsinki Cor‑ pus (ca. 10,000 words), we find altogether 109 forms of causal connectors. In addition, there are five instances of lexical expressions and eight of inferential þonne ‘then’ in post‑first‑position (see above, Chapter 5.4). These results are comparable to the findings for Present Day English (see section 9.2.2): the relation of cause is only very rarely expressed by full lexical phrases or prepositional phrases such as be þæm þinge ‘for this reason’: Table 9.6.: Causal connection in the Old English adaptation of Boethius’ De Consolatione Philosophiae cause → result (21 forms/variants) simple forms forþy forþæm (pre‑posed) forþæm þe correlatives forþæmþe … forþæm result → cause (88 forms/variants) simple forms forþæm forþæm þe forþon þe forþy nu correlatives forþy … forþæm forþy … forþy forþy … þe forþy … þy forþy … for þy þe forþæm … þe forþæm … forþæm forþæmþe .. forþæm forþæmþe … forþyþe nu … nu þy … þy þy .. þe Other means (5 forms/variants) be þy ‘for this; hence’ for þæm þinge þonne

Sum 4 14 1 1

19 1

1

1

44 23 1 1 4

69 4

2 1 2 1 1 1 2 1 1 2 1 1 4 1 8

12 2 2 5 8

OE forþæm, forþon, forþy   145

This survey shows that all of the forms of forþæm (þe) are employed for the various subtypes of the sequences cause – result and result – cause (forþæm þæt is not included here because it is only used for purpose clauses; four instances). With altogether 100 of the 109 instances, variants of simple or correlative forþæm amount to ca. 92 per cent of all causal connectors (if we include the ambiguous cases of temporal‑causal þonne and the lexical/prepositional means of clausal connection, we still get ca. 82 per cent). The only other construction which is used more than once (in ca. 5.5 per cent of all cases), is nu, also either in simple (114) or correlative constructions (115; see also above Chapter 5.2.2): (114) Se ilca God is, swa swa we ær sædon, þæt hehste good & þa selestan gesælða, nu hit is openlice cuð þæt ða selestan gesælða on nanum oðrum gesceaftum ne sint buton on Gode (COBOETH, 34.84.3). ‘The same God is, as we before said, the highest good, and the best happiness since it is evidently known that the best felicities are in no other things but in God’. In the correlative construction (115), the causal interpretation of nu is strengthened by the incongruity of a temporal reading of nu ‘now’ and þonne ‘then’. The givenness of the pre‑posed causal clause is highlighted by the deictic reference to the preceding text by a verb of communication (witan ‘know’ in ðu … wast): (115) ða cwæð he: Nu ðu þonne wast hwæt ða leasan gesælða sint & hwæt þa soþan gesælða sint, nu ic wolde þæt þu leornodest … (COBOETH, 33.78.27). ‘Now (that) you know what the false goods are, and what the true goods are, (now) I would like that you should learn …’. Other causal connectors which are commonly also listed in surveys compiled by a “dictionary-cum-grammar method” (see Kortmann 1997: 331) turn out to be fairly infrequent. For one out of three examples using the instrumental form þy of the demonstrative, see the correlative construction in (116): (116) þy hi secað anwald & eac eall oðru good þe we ær ymb spræcon, ðy hi wenaþ þæt hit sie þæt hehste good (COBOEHT, 34.88.4). ‘Therefore they seek power, and also the other goods, which we before mentioned, because they think that it is the highest good’. These findings for Early West Saxon basically agree with the results for causal connectors in an – also rather independent – Old English text from the Late West

146   Shifting deictics in the history of English causal connection Saxon period, Byrhtferth’s Enchiridion (complete Old English text examined; altogether seventy‑one causal connectors in ca. 20,000 Old English words). Table 9.7.:  Causal connectors in Byrhtferth’s Enchiridion Sum

cause → result (30 forms/variants) simple forms forþon/forþan (pre‑posed) forþon (pre‑posed) nu correlatives forþon … forþan nu … nu

15 1 3

16 3

2 9

2 9

result → cause (41 forms/variants) simple forms forþon/forþan correlatives forþon … forþan

35

35

6

6

The forms forþon (and its orthographical variant forþan) are as predominant in this text as they are in the early West Saxon adaptation of the Consolatio (59 occurrences), amounting to ca. 83 per cent of all expressions of the causal relation. The rest (12 instances) are expressed by nu, either in the simple form or in correlative expressions. In contrast to the translation of the Consolatio, however, Byrhtferth employs nu solely for the relation cause – result, most often in pre‑posed topic‑forming causal clauses: (117) Nu ic ealles ymbe þas þing spræce hæbbe, me þingð behefe þing þæt ic swa mycel ymbe þissum getæle preostum gecyðe, swa me ne gesceamige … (ByrM, 3.3.275). ‘Now that I have spoken about these things in detail, it seems to me ne­ cessary to make known to priests enough about this numbering system that I will not be ashamed …’. (118) Nu we habbað sceortlice amearcod þæra hiwa gefeg þe boceras gymað, nu þingð hyt us gedafenlic þæt we heom gecyðon þæt we ær geheton … (ByrM, 3.3.234). ‘Now that we have briefly written the series of figures that writers use, [now] it seems fitting to us to tell them what we promised before …’.

Discourse deixis   147

In most of these cases, the temporal textual deixis and givenness of the pre‑posed causal clause is, like in (115) above, further emphasized by a verb of communication, namely sprecan in (117) and mearcian in (118). Forþæm and its variant forms can certainly be called the central causal connector in Old English, since they mark eighty to ninety per cent of all instances of causal connection. In spite of the wide variety of its possible functions, however, there are only few ambiguous instances. In most of the cases, the semantics and pragmatics are clear. In fact, the instances which cause problems for the understanding of a text can be listed individually (see the discussion in Mitchell 1985: § 3011–3014). In the following chapters, it will be argued that this low number of ambiguous attestations – and also the essential patterns in the later development of causal connectors in English – is attributable to the morphological make‑up of the Old English connectors and in the deictic value they inherently contain (for similar approaches, see Wiegand 1982; Traugott 1992; Markus 2000). They belong to category II of Raible’s continuum of connectors (introduced above in Chapters 5.4 and 7.4) because they are explicitly anaphoric or cataphoric. 9.4.  Discourse deixis 9.4.1.  Forþæm: morphological make‑up and discourse deixis The morphological make‑up of the forms forþæm, forþon and forþy is unproblematic. They are prepositional phrases consisting of the preposition for governing the distal demonstrative pronoun in the dative (þæm) or instrumental (þon and þy).7 In all functions, they may, but need not, be followed by the particle þe (see Campbell 1959: § 708–709; Mitchell 1985: § 3011–3051): preposition for for

+ + +

distal demonstrative pronoun dative þæm – instrumental þon/þy þæm/þy

[+ þe] [+ þe] [+ þe]

Traditional accounts of the history of the “for‑causals” state that the original prepositional phrase was re‑analysed as a conjunction in (pre‑)Old English. A prototypical use of one of the common forms of for þæm in an external reason 7 Campbell (1959: 290, § 708, n. 4), however, remarks in his discussion of the forms þon and þy: “The classification of these forms as instrumental is traditional, but reflects neither their origin, nor their prevailing use”.

148   Shifting deictics in the history of English causal connection clause is example (119), where the form forþon – traditionally classified as a subordinating conjunction – points anaphorically to the preceding clause. (119a) Do þærto fife forþon þunresdæg hæfð fif regulares. ‘Add thereto five, because [Conj.] Thursday has five regulares’ (ByrM, 1.2.236). This construction is commonly seen as a re‑analysis of (119b) *Do þærto fife for þon: þunresdæg hæfð fif regulares. ‘Add thereto five for that [reason] [pp as a]: Thursday has five regulares’. In construction (119b), assumed to be underlying (119a), the prepositional phrase for þon functions as an adverbial in the first clause, and cataphorically refers to the following clause. The demonstrative þon has a dual function: with respect to the first clause, it is the noun phrase in the prepositional phrase, which itself functions as an adverbial. At the same time, þon indexes, i. e. points cataphorically, to the second, causal clause, identifying the adverbial relation cause which must necessarily follow here. Since demonstratives are by definition deictic (cf. Brugmann 1904), the noun phrase þæm of the construction necessarily needs a point of reference which it points to. Theoretically, the point of reference required by for þæm could be found in the extra‑linguistic reality and the speaker could identify the cause indexed by þæm (‘this one’) by the “pointing” of his finger. Commonly, however, the point of reference is present in the co‑text, i. e. the preceding or following discourse. This is most clearly seen in the so‑called “correlative constructions”, which are rather frequent in the Old English texts examined (ca. 12 per cent in COBOETH, ca. 25 per cent in ByrM). They are the most explicit surface markers of causal connection because they index one another and thus reinforce the conjoining force of the respective connectors (see the defining properties of Raible’s category II). (120) Forþon Romani hine gelogodon on þissum monðe (þæt ys on Februario) forþam he ys scyrtest ealra monða …(ByrM, 2.1.36). ‘The Romans placed it [therefore] in this month (in February) because it is the shortest of all months’. These constructions are therefore employed for stressing result – cause sequences as in (120), but can also be used for stressing the topic‑forming quality of pre‑posed causal clauses, as in

Discourse deixis   149

(121) Aprelis, Iunius, September, and Nouember habbað feower and twentig læs, and Februarius, forðon he ys se læsta and he hæfð twegra daga læs þonne þa oðre, forþon he forlæt eahta and feowertig tida (ByrM, 2.1.372). ‘April, June, September, and November have twenty‑four fewer, and February, since it is the shortest and has two days fewer than the others, [therefore] loses forty‑eight hours’. The forms forþæm, for þy etc. are intrinsically deictic, because their demonstratives þæm or þy require an element in the near co‑text to which they relate: this can either be the clause or also a much larger piece of discourse. They are thus inherently phoric, either cataphoric or anaphoric, and point to a pragmatically governed use of deixis. Wiegand (1982) even goes as far as to maintain that the phrases for þæm with variants are not yet conjunctions in Old English because the prepositional phrase is still so transparent in its deictic reference – in its respective context – that we do not have to assume a univerbated use of forþæm with variants as a conjunction (Wiegand 1982: 388). For a better understanding of the following argument and, in particular, the diachronic and cross-linguistic comparisons, it has to be stressed that the present investigation uses – in accordance with most studies on similar topics – the term “deixis” in its widest sense. With respect to discourse deixis, a clear distinction between “deixis” (e. g. reference) and “anaphora” (discourse deixis), as this is supported by, for instance, König (2009: 151), cannot easily be drawn because the written, but also the spoken, preceding and following co‑text itself can be seen as an extra‑linguistic point of reference (for a discussion of this ­issue and definitions of deixis and anaphora, see Lenz 1997: 7–108 and Consten 2004: 4–58).

9.4.2.  Pronominal connectors in Present Day German This pattern of so‑called “pronominal adverbs” (Pasch et al. 2003: 7) is frequent in Latin and, though often opaque, in many of the Romance languages (see above, Chapter 5.4.2) and in all Germanic languages. In Old English we find it for all kinds of semantic relations: see, for instance, ær þæm (þe) ‘before’, æfter þæm (þe) ‘after’, mid þæm (þe) ‘during’ or wið þæm þe ‘provided that’. In Present Day German, the pattern is not only extant but actually the predominant one for causal adverbial connectors (cf. dem‑ in demnach or des‑ in deswegen):

150   Shifting deictics in the history of English causal connection Table 9.8.:  Present Day German connectors cause/result da, weil, weshalb, weswegen so dass also, daher, darum, demnach, demzufolge, deshalb, deswegen, folglich, infolgedessen, so, somit German Adverbial Connectors cause denn, nämlich semibold: pronominal elements; underlined: lexical elements; capitals: other types of deixis (for instance, space or time deixis)

German Conjunctions cause German Conjunctions result German Adverbial Connectors result

Virtually all of the causal connectors in Present Day German, but predominantly adverbial connectors marking result – apart from folglich, which is purely lexical – contain a deictic element (see items in bold), most often a demonstrative pronoun in the dative (dative singular masculine/neuter dem) or, in younger forms, the genitive (genitive singular masculine/neuter des), which allows its analysis as a prepositional phrase with anaphoric or cataphoric reference. Another group of important deictic connectors – marked by capital letters – are those relating to time and space deixis, such as PDG weil (cf. PDG Weile ‘while, (short) period of time’; time) or PDG denn (variant of PDG dann ‘then’, see Kluge/ Seebold 2002, s. v. denn; time). Originally spatial da (‘there’ space; by metonymic transfer also ‘then’ time) is attested in PDG da, PDG daher and PDG darum. These are text deictics (see above, Chapter 5.2.2) because the point of reference is the text in its temporal and spatial extension. Da ‘there’ in PDG daher, for instance, relates the following to the preceding element of discourse which is the cause for the result mentioned in the daher‑clause (cf. PDG ‑her ‘from (there)’). In contrast to the pronominal connectors, which explicitly require a point of reference in the co‑text, this signalling of deictic reference is more subtle and thus asks for a more sophisticated cognitive process by the listener/reader (Consten 2004: 26–37). Table 9.9.:  Deictic expression in Present Day German causal connectors demnach, demzufolge, deshalb, deswegen, infolgedessen, weshalb, weswegen Time/Space deixis da, daher, darum, denn, ?weil Deictic so8 so (dass), somit, ?also Lexical means folglich ? No longer transparent also, ?weil ? of uncertain status as to their transparency Pronominal deixis

8 So is a very strong deictic indicator but is – in German as well as in English – notoriously difficult to analyse because of its polyfunctionality; see OED, s. v. so and – for a full account of the functions of so in Old English – Schleburg 2002 (see also above, Chapter 7.5.1).

Discourse deixis   151

Present Day German hence shows three different patterns for causal connectors: in addition to the pronominal connectors, such as deshalb or deswegen, it uses linguistic items which employ time and space deixis (see daher, somit); only rarely lexical elements, such as PDG folglich ‘lit. followingly’, are found. 9.4.3.  Deictic elements in English causal connectors An examination of the system of causal connectors in Present Day English shows that there is not a single fully transparent remnant of the principal pattern of Old English, i. e. pronominal connectors such as forþæm. Table 9.10.:  Deictic expression in Present Day English causal connectors9 Pronominal deixis Time/Space deixis Deictic so, thus Lexical means Not transparent

–

(that), hence, then, ?therefore so, so (that), thus because, accordingly, consequently, for as, ?therefore since; now

semibold: pronominal elements; underlined: lexical elements; capitals: other types of deixis (for instance, space or time deixis)

With respect to the conjunctions, we find the lexical because (cf. cause) and the polyfunctional since, which is transparent in its time deixis because it still has its temporal meaning. In as (< eall swa ‘all so’), on the other hand, the deictic so is no longer transparent; the same is true in the case of for, which is no longer related to the prepositional phrase for þæm, but is only semantically transparent through the preposition for ‘because of that’. This is in stark contrast to Old English, where we only find – like in Present Day German – expressions marked explicitly for deixis, either pronominal connectors such as forþæm with variants or originally instrumental þon or þy in the second elements of correlatives. Time deixis (which may originally also denote space)10 is attested in nu ‘now’, þa ‘then’ and þonne ‘then’ (and variant forms), either as simple forms or in correlatives. Like forþæm, 9 For further polyfunctional connectors, which may also be used for the causal relation, such as after all, indeed, in fact or of course, see below, Chapter 12. 10 Clark Hall (1984), s. v. heonu lists the meaning ‘therefore’ among a wide variety of concessive meanings for this word. In all of its seventy‑eight attestations in the Rushworth and Lindisfarne Gospels, however, it exclusively glosses Latin ecce and is thus not causal (see DOEC).

152   Shifting deictics in the history of English causal connection these are ambiguous adverbs/conjunctions (marked by ° in this and the following tables). Table 9.11.:  Deictic expression in Old English causal connectors Pronominal deixis Time/Space deixis

°forþæm, forþon, forþy in correlatives: þy, þe °nu, °þa, °þonne °swa?

This summary shows that Old English only employs explicitly deictic elements to mark causal relations, mainly – when the high token number of the variants of the forþæm is taken into account – pronominal connectors.

9.5.  Causal connectors in the history of English 9.5.1. Causal connectors in English translations of Boethius’ De Consolatione Philosophiae Most of the causal connectors used in Present Day English testify to dramatic changes which have occurred in the English system after the Old English period. For a first survey of the development see Table 9.12 summarizing various causal connectors as attested in translations of Boethius’ De Consolatione Phi‑ losophiae after the Old English period (for Old English, see Table 9.6). It lists the different connectors for the relations cause – result and result – cause, ordered by diminishing frequency in the respective texts. Table 9.12.: Causal connectors in English translations of Boethius’ De Consola‑ tione Philosophiae (absolute numbers in brackets; for details on the corpus texts, see Appendix C.2.2) cause → result Chaucer, ME Boece (CMBOETH)    for whiche (8), forwhy (3), therefore (3), forthy (2)    pre‑posed: for as moche as (7), for (2), syn (1), syn that (1) Colville 1556 (CEBOETH1)    therefore (21), whereby (7), so that (6), wherefore (5), so (3)    pre‑posed: for by cause (3), synce that (1)

Causal connectors in the history of English   153 Queen Elizabeth 1593 (CMBOETH2)    therefore (17), wherefore (7), so (7)    pre‑posed: because (2), since (2)    correlative: because … now (1), now that … now (1) Preston 1695 (CMBOETH3)    therefore (17), hence (7), wherefore (4), so (1)    pre‑posed: since (3), because (2), now that (2) James 1897 (CLBOETH3)    wherefore (6), therefore (5), accordingly (3), so (2), so then (2)    pre‑posed: since (12) Watts 1969    therefore (14), so (8), indeed (3), consequently (1)    pre‑posed: since (11), because (2) result → cause Chaucer, ME Boece (CMBOETH)    for (63)    post‑posed: for (15), syn (2), syn that (1)    correlative: wherefore … for as moche as (1) Colville 1556 (CEBOETH1)    for (53)    post‑posed: for 13, for by cause (2) Queen Elizabeth 1593 (CEBOETH2)    for (60)    post‑posed: for 8 Preston 1695 (CEBOETH3)    for (53)    post‑posed: since (7), because (6) James 1897 (CLBOETH3)    for (28)    post‑posed: since (6), because (2) Watts 1969    for (22)    post‑posed: because (7), since (4), for (3)

This table attests to the wide variety of forms that have been coined as replacements of OE forþæm (and its variants), which, of course, cannot be discussed in detail here. Essentially, it shows that the dramatic structural changes can best be illustrated by the changes affecting adverbial connectors in the (Early) Mid-

154   Shifting deictics in the history of English causal connection dle English period (see 9.5.2.) and by the changes of connectors in pre‑posed causal clauses (see 9.5.6. below).11 9.5.2.  Causal adverbial connectors in the history of English Adverbial connectors are the kind of connectors which have to signal clausal linkage most clearly because they present speakers’ opinions of the relation of the states of affairs explicitly in two information blocks with independent illocutionary weight (see above, Chapter 2.4). Table 9.13 provides a list of linguistic elements which have served the function of a causal adverbial connector in any period of the history of English, based on the data in Appendices A.1 and B.3. To allow a comparison, the layout of the table is modelled on the tables in Kortmann (Kortmann 1997: 342 for causal subordinators). It only lists the central causal/resultive connectors, i. e. no borderline cases such as PDE after all.12 This table illustrates that the Early Middle English period in particular was a “period of experiment and transition”: in the periods ME1 to ME3 (1150–1350), nine adverbial connectors are coined, but most of them are ephemeral. Only two of them have survived into Present Day English, namely therefore13 and thus (a late‑comer first attested in the adverbial connector function in 1380; see OED, s. v. thus).14 11 There are some differences in the relative frequencies, but not in the forms of causal connectors between translations and comparatively independent texts. In Chaucer’s Astrolabe (CMASTRO – complete text), for example, we find for the sequence cause – result: therefore (10), wherefore (3) thanne (2), forwhy (1), pre‑posed: for as much as (2), by cause that (2), for (1), sin that (1). For the sequence result – cause, we find: for (20), for (8) and prepositional by-cause that (7). 12 Since connectors for the semantic relation cause form a closed class, there are only very few other causal connectors which are not listed here because they are attested very infrequently (such as Latin ergo; see OED, s. v. ergo). Similarly, connectors which may also be used for the relation addition/transition, such as indeed, in fact, of course are not listed here (for these, see also Table 9.15). The meanings and functions of after all are still only given in the entry for the preposition / adverb after in the OED (s. v. after). 13 For the proliferation of there-compounds from the beginning of the Middle English period, see Österman 1997, and above, Chapter 8.3. The history of therefore is rather dark (cf. OED, s. v. therefore): I suggest a similar path to that which has led to Latin propter‑ea (cf. OE for‑þæm) vs. ea‑propter (cf. ME therefore). 14 The adverb þus is, of course, attested in Old English, but it was only employed as a circumstance adverb modifying a verb, in phrases such as þus sprecende ‘speaking thus’. The different uses are, however, often hard to distinguish.

1420–1500 (ME 4)

1350–1420 (ME 3)

1100–1350 (ME 1/2)

Old English thus

therefore wherefore herefore for the which

therefore wherefore [herefore] for which thus [consequently]

for

for

1570–1640 (EModE2)

thus

forþi forhwi for for that therefore wherefore [herefore] for which

1500–1570 (EModE1)

forþi forhwi for for that therefore wherefore herefore for which [ergo] thus

1640–1710 (EModE3) thus consequently hence thence whence

therefore wherefore

for

so then

whence

thus consequently hence

therefore

for

so then

thus consequently hence

therefore wherefore

for

so then

thus consequently hence

therefore

for

so then

accordingly accordingly accordingly accordingly

thus consequently hence

therefore wherefore

for

so then

1710–1780 (LModE1)

so then

1780–1850 (LModE2)

so then

1850–1920 (LModE3)

so then

1920– (PDE)

[cuþlice] swa swa þonne thane/ thenne [þæs] þi þi forþæm/ forþon forþon forþy forþi forhwi for for that therefore wherefore herefore so then

Table 9.13.:  Causal/resultive adverbial connectors in the history of English (based on Appendices A.1 and B.3)

Causal connectors in the history of English   155

156   Shifting deictics in the history of English causal connection A closer look at the various forms coined in these centuries shows that the coinages reflect the consequences of the loss of transparency in the central causal connectors at the end of the Old English period. Forþan (from instrumental for þon or from levelled forms of the dative for þæm) and the instrumental forþi (< for þy) lost their transparency when the paradigms of both of the Old English demonstratives collapsed (see above, Chapter 7.5). We find remnants of inflected forms of the Old English demonstratives in some southern and western texts as late as the thirteenth century, but most of these inflections are number concord and thus irrelevant for forþæm and forþy. Consequently, these forms lose their deictic value and are soon lost altogether. In this process, phonologically weakened forms such as forþan [‑ðǝ(n)] are given up much earlier than the stronger forþi (cf. the long vowel in OE for-þȳ), which survives as a univerbated form until the end of the Middle English period (see MED, s.vv. for‑than and for‑þī). For forþy, orthographical forms showing a reduction of the final long vowel (such as forþe) are only attested from the late thirteenth and fourteenth century. The earlier forms seem to reflect the full (long) vowel and not a reduced schwa (cf. vor‑þui or ferþi; see OED, MED, s. v. forthi, for‑þī). The forms of OE forþǣm, on the other hand, are in spite of the Old English long vowel regularly attested as ‑þam, ‑þon, or -þen (see MED, s. v. for‑than). Studies examining the history of OE forþæm in detail find a very rapid simplification of the for‑formulae, which is first attested as simple for around 1100 in manuscript F of the Anglo‑Saxon Chronicle (see the tables in Kivimaa 1966: 214 and 250). This rapid rate is particularly evident in the continuations in the Peterborough Chronicle, where the use of simple for increased from 27 per cent in Continuation I to 86 per cent in Continuation II (for the years 1132 to 1154). In the Early Middle English texts investigated by Kivimaa (1966), for is the form used almost universally, and in several texts it is the only one attested. This line of development is very similar to that which has above been summarized as the change attested in French‑based creoles, where parce que appears as non‑transparent pas (Raible 1992: 201; see above, Chapters 5.5.2 and 5.5.3). In pas, parce que has lost both of its deictic elements ce and que, a process identical to the development of forþæm þe into for in Early Middle English. This parallel development, of course, brings the Middle English creole hypothesis back to mind, i. e. the conjecture that Early Middle English underwent creolization at the time of either the Norse or Norman Conquests, or during both periods. This hypothesis, which was very popular in the 1980s, was mainly based on the assumption of an extreme reduction in inflected forms from Old English to Middle English. While a distinct process of creolization (with all its properties) is certainly to be rejected for Middle English (for a summary of the arguments, see Görlach 1986), it is nonetheless interesting to see that the reduc-

Causal connectors in the history of English   157

tion of forms has also led – in both languages discussed here – to a collapse of the system of causal adverbial connectors. At the same time, we see – in English and in the creoles – attempts at innovations in the field. Like its Old English precursor forþæm, the reduced form for was ambiguous as to its status as an adverb, a coordinating or subordinating conjunction (see also Jucker 1991). While this does not seem to have presented any problems in Old English, we increasingly find tendencies for a better mapping of form and function by specifically marking the most problematic pre‑posed subordinate clauses (by for as moche as; see below, section 9.5.6) and adverbial connectors. 9.5.3.  Deixis in new adverbial connectors With respect to the novel adverbial connectors in Early Middle English, it is obvious at first glance that there are not very many different patterns among the new coinages. Table 9.14.:  Adverbial connectors: new coinages in Middle English (ME1 to ME3) Pronominal Connectors    with demonstrative    with interrogative/relative Text‑Deictics    spatial

°for‑that °forhwi, °for which, °wherefore distal: therefore proximal: herefore

Deictic thus Non‑Deictics    no longer transparent    lexical

for ergo

These innovations show that Middle English first of all tried to use the new material for coining causal connectors similar to the make‑up of OE forþæm. The earliest coinages are pronominal connectors which employ the new form of the demonstrative that (MED, s. v. for‑that adv. & conj.; the lexicalized status of for which is not as clear). The new coinages are also similar to their Old English models in that many of them are so‑called “ambiguous adverbs/conjunctions”, which can be employed as adverbial connectors or conjunctions (cf. °for‑that, °for which, °forhwi ‘forwhy’ and °wherefore). This pattern can be exemplified by the rather frequent for whiche (found eight times in Chaucer’s Boece; see Table 9.12) and the functional extension of forms such as the relative and originally

158   Shifting deictics in the history of English causal connection interrogative forhwi. From the thirteenth century onwards, it is not only used as an interrogative or relative but also as an adverbial connector signalling a new information unit, thus carrying the meaning ‘therefore’ (see OED, s. v. forwhy). This use is only attested as a conversational implicature in Old English (see DOE, s. v. forhwǣm, forhwon, forhwȳ B.).15 Forhwi and forwhich are, however, the last coinages which follow the Old English pattern of “pronominal connectors”. The other field of experiment are new connectors employing time or space deixis, similar to OE nu. From early Middle English onwards, deixis of space in particular is becoming more important (see above, Chapters 8.2 and 8.3). The co‑text is taken as a point of reference in patterns using the relative where and the distal and proximal forms there and here (cf. therefore, wherefore), which relate the following to the preceding discourse. Many studies have shown that spatial cognition is at the heart of our thinking and that most of our fundamental concepts are organized in terms of one or more spatialization metaphors (Lakoff and Johnson 1980: 17; Levinson 2003). With respect to the use of time and space deictics as connectors, however, there is a difference between orality and literacy. While time deixis may be used for both the spoken and the written medium, the metaphoric extension of spatial terms into metalinguistic ones is a function of literacy, since it is only with literacy that language is objectified in visual space (cf. Ong 1982; for details, see above, Chapter 8.3). Though this means of establishing cohesion is not as explicit as the anaphoric or cataphoric linkage by a demonstrative, it is still comparatively transparent in signalling deixis by reference to the written text in its chronological and spatial extension. These forms become very frequent as adverbial connectors from the beginning of the thirteenth century onwards (cf. Österman 1997 and Markus 2000), i. e. at exactly the time when the double paradigm of the demonstratives was given up, yielding the indeclinable article the and the demonstrative this/ that, which are only inflected for the referential category number. 9.5.4.  ME for that Considering these variants of pronominal forms and the fact that none of them has survived into Present Day English, one of the central questions is why speakers and writers of Early Middle English did not choose for that as the central connector or conjunction. That is, “Why did speakers not simply start 15 For the importance of the difference between conversational implicature and coded meaning for a theory of regularities in semantic change, see Traugott and Dasher 2002.

Causal connectors in the history of English   159

substituting þat for þæm, þon and þy, as they generally did for the demonstrative objects of prepositions?”16 The fact that þon and þy were not replaced by the new demonstrative form that, on the other hand, is not startling, since the instrumental þy was only marginally a case category even in Old English (see Campbell 1959: § 708, n. 4). Since þy in particular was mainly used as an element of complex conjunctions, it does not come as a surprise that forþi survived as a univerbated, lexicalized resultive connector ‘therefore’ until the end of the Middle English period. Both the weak impact and the loss of for that, however, require a more detailed investigation because they testify to the collapse of the pattern which had been central in Old English. For this examination, Early Middle English texts are chosen which employ for þat as both an adverbial connector and a conjunction (CMTRINIT ca. 1225, CMKENTSE ca. 1275, and CMAYENBI ca. 1340). In these texts, for þat is used – in clear replacement and imitation of OE forþæm – in correlative constructions with forþi: (122) and eft he us wile feie; þanne we shulen arisen of deaðe. and forþi [122a; adverb] we clepeð him fader for þat [122b; conjunction] he us feide here. ‘… and he will also join [our limbs together] when we will arise from death. And therefore we call him father because he put us together here’ (CMTRINIT, p. 26). While the instances of for‑þat in examples (123b) and (123c) should be regarded as conjunctions heading genuine internal reason clauses (they are the focus of the negation; for this criterion, see above, Chapter 3.2.3), for‑þat in (123a) unambiguously functions as an adverbial connector: (123) And herodes iherde ðet o king was ibore ðet solde bi king of geus. … for ðet [123a] he was ofdred for to liese his king riche of ierusalem. … And al swo herodes iherde ðis. swo spac te ðo ðrie kinges. and hem seide. … hic wille go and anuri hit. ðet ne seide he nocht herodes for ðet [123b] he hit wolde onuri; ac for ðet [123c] he hit wolde slon (CMKENTSE, p. 214–215). ‘And Herod heard that a king had been born which should be the king of the Jews. … Therefore [123a] he was very afraid that he might lose his kingdom of Jerusalem … And as soon as Herod heard this, he spoke to the three kings and said to them … I will go and worship it [= the child]. 16 I would like to thank Cynthia Allen for raising this very important issue at an earlier presentation of this chapter.

160   Shifting deictics in the history of English causal connection This he, Herod, did not say because [123b] he wanted to worship him, but because [123c] he wanted to kill him’. In most of its attestations in Early Middle English, for þat does not function as an adverbial connector but as a subordinator ‘because’ (often rendering Latin quia). This employment is not at all unexpected, since it is in line with the development of that as the general subordinator in Middle English. Kivimaa (1966: 248) shows that the earlier conjunctive phrase employing the subordinating particle OE þe lost þe during the twelfth century and began to be found with that only in the thirteenth century. That also takes the place of OE þe as a relativizer and becomes the general subordinator in Middle English (see the quotation from Fischer 1992 cited above, Chapter 7.5). The use of that as the general subordinator helps to explain why ME for that did not replace OE forþæm as the central adverbial connector for the relations cause/result. While the subordinating particle þe was different from all of the forms of the demonstratives in Old English, the Middle English subordinator that is homonymous with the demonstrative. This means that sentence‑initial for that would be conceptualized as introducing a pre‑posed causal clause and not as an adverbial connector. For this reason, it is more and more avoided in this position and replaced by other, unambiguous connectors. For examples of this unstable situation of experiment and transition at the beginning of the thirteenth century, see the following examples from Vices and Virtues (CMVICES1; ca. 1225): (124) Forðan [124a] hie bieð godes wiðerwinen, alle ðo ðe willen hem sel­ uen heiʒin. Godd seið him self ðat hie sculen bien ineðerede. Hierfore [124b] ic am neðer and unmihti, forðan [124c] ic habbe ʒeben prud and modi … (CMVICES1, p. 5). ‘Therefore [124a] they are God’s enemies, even though they want to raise themselves. God says himself that they shall be lowered. Therefore [124b] am I low and powerless, because [124c] I was proud and conceited …’. (125) Hie is anginn of alle cristendome, hie mai michele eadinesse of‑earnin at ure lauerde gode, for ðan ðe [125a] hie iliefð ðat hie næure niseih. For ði [125b] sade Crist: “Eadi bieð ða menn ðe on me belieuen and næure me ne seiʒen!” (CMVICES1, p. 25). ‘He is the beginning of all Christianity, he may earn much blessing from our Lord, because [125a] he believes what he has never seen. Therefore [125b] Christ said: “Blessed are those, who believe in me and have never seen me”’.

Causal connectors in the history of English   161

The form forðan with phonological levelling is still the most polyfunctional item: it is used as an adverbial connector marking result in (124a), as the second part of a correlative construction marking internal cause in (124c) and – still with the particle þe – in an internal reason clause in (125a). In adverbial connector function, we also find the spatial hierfore (124b) and the by then no longer transparent form forði (125b). Another text from the Early Middle English period, Hali Meidenhad (“Kathe­ rine Group”; CMHALI, ca.  1225) only differentiates the form for denoting cause (126a, 126c, 126d) as against forr‑þi (126b) denoting result. (126) Meiðhad is þet an ʒeoue iʒettet te of heouene: do þu hit eanes awei, ne schalt tu neauer nan oðer swuch acourin; for [126a] meiðhad is heouene cwen, & worldes alesendnesse, þurh hwam we beoð iborhen … forr‑þi [126b] þu Ahest, meiden, se deorliche witen hit; for [126c] hit is se heh þing, & se swiðe leof godd, & se licwurðe. & þet an lure þet is wituten couerunge. ʒef hit is godd [leof], þet is, him seolf swa ilich, hit na wunder; for [126d] he is leoflukest þing, & buten eauer‑euch bruche … (CMHALI, p. 135). ‘Virginity is the one gift granted to you from heaven; if you once dispose of it, you will never regain another quite like it. For [126a] virginity is the queen of heaven, and the world’s redemption through which we are saved … Therefore [126b], maiden, you have to guard it carefully. For [126c] it is the high thing and so very dear to God and so acceptable, and one loss of it is without recovering. If what is so like God is dear to him, it is no wonder, because/for [126d] he is the loveliest thing and without any sin …’. 9.5.5.  “Recursive” for – Latin nam/enim Already this text from the beginning of the thirteenth century shows a proliferation of for, which has become non‑deictic after losing its demonstrative. It is almost infinitely repeatable and may be used for post‑posed causal clauses working on the local level, providing an explanation for the first connect. In most cases, however, non‑deictic for is ambiguous (126d) or employed on a more global level, sketching or justifying the line of argument of the author. For is thus a prototypical coordinator, because it marks the second connect as having an illocutionary weight of its own and, more importantly, explicitly marking the voice of the speaker who comments on his view of the relation of textual portions.

162   Shifting deictics in the history of English causal connection It is thus comparable to the Latin adverbial connectors nam and enim (see Kroon 1995: 131–203), both of which mainly work on the global level of text­ ual organization (in the present study termed “transitional”). Instead of being adverbial connectors in a prototypical, strict sense, they are rather connective particles “concerned primarily with the presentation and organization of the information conveyed in the discourse” (Kroon 1995: 203). They are thus situated at the interface between the clausal relations of cause/result and add­ ition/transition. Accordingly, they are commonly not rendered by forþæm and its variants in Old English, but by OE soþlice or witodlice ‘lit. truly, certainly’ (see Lenker 2000 and above, Chapter 8.4.3 and below, Chapter 11). From Middle English onwards, however, these adverbs are less frequently used and for is preferred to mark the transitional relation. This can be shown by a comparison of translations of Latin enim and nam in the various translations of the Consolatio in the history of English. Table 9.15.: Rendering of Latin enim and nam in English translations of Boethius’ De Consolatione Philosophiae17 Translation text

for

No translation Adverbs – cognitive Diverse domain truth 4 certes 4 but 1 forhwy 2 and thus 1 – certes 3 and so 1 surely 1 for why 1 truly 1 11

Chaucer ME (CMBOETH)

50

Colville 1556 (CEBOETH1)

53

Queen Elizabeth 1593 (CEBOETH2) Preston 1695 (CEBOETH3)

49 45

9

Watts 1969 (PDEBOETH2)

25

21

indeed 1

but 2 now 2 so then 1 thus 1 and 1 because 1 but 1

17 Old English cannot be used for such a detailed analysis, because the Old English Consolatio is not a genuine translation, but a rather free adaptation with many divergences from and additions to the Latin text. Also, the numbers for Chaucer do not fully agree with those of the other translations, because Chaucer only additionally uses the Latin text. Chaucer’s main source is an Old French version by Jean de Meun (ed. Dedeck-Héry 1952), which renders Latin nam and enim by OF car.

Causal connectors in the history of English   163

This comparison shows that for has indeed become a transitional connector. In this discourse function, it is also called “recursive for” (Mueller 1984: 135). It has lost much of its causal meaning denoting internal reason clauses and mainly works on the textual level, indicating “This is my line of argument”. In Chaucer’s translation of the Consolatio, this is evident in the many collocations with other connectors, such as for so as (127a), for yif (127b) and for certes (127c). (127) Philosophie: “… that God, …, is good. For [127a], so as nothyng mai ben thought betere than God, it mai nat ben douted thanne that he that no thinge nys betere, … For [127b] yif God nys swyche, he ne mai nat be prince of alle thinges; for [127c] certes somthing possessyng in itself parfyt good schulde be more worthy than God… For [127d] we han schewyd apertely that alle thinges that ben parfyt ben first er thyng­ es that ben inparfit; and forthy [127e] … we owe to graunte that the sovereyn God is ryght ful of sovereyn parfit good” (CMBOETH, p. 432). ‘[It is the universal understanding of the human mind] that God … is good. For [127a], since nothing can be conceived better than God, then it may not be doubted that nothing is better … For [127b], if God is not such, he cannot be the prince of all things; for [127c], certainly, there would have to be something else possessing perfect goodness over and above God … For [127d] we have shown overtly that all things that are perfect are superior to things that are imperfect; and therefore [127e], …, we must grant that the supreme God is very full of supreme and perfect goodness’. This transitional use of for highlighting the line of argument is the predominant one from the middle of the fourteenth century until the middle of the eighteenth century (see the high number of occurrences in Table 9.15). For Early Modern English, this interpretation is also supported by the increasingly more consistent punctuation: more than half (58 per cent) of the for‑clauses analyzed by Claridge and Walker are preceded by heavy punctuation (Claridge and Walker 2002: 42–44). It has to be stressed again that in these cases for should be classified as a coordinator, even if there is no way to confirm this by the syntactic or topological criteria summarized above (cf. Chapter 2, Table 2.2 and also Table 9.1). Con­sider­ ing issues of pragmatics and text‑linguistics, however, for has to be analyzed as a coordinator because it gives an independent illocutionary weight to the second connect, in many instances by explicitly indicating that the speaker thinks that the propositions of the two sentences are connected in some way or other. Much of the uncertainty about the status of for in Present Day English is probably due to the fact that for, and in particular the rhetorical uses of for,

164   Shifting deictics in the history of English causal connection have decreased rapidly since the end of the eighteenth century. In Present Day English translations of the Consolatio (Watts 1969), the number of instances of for translating Latin nam and enim is halved. In the other half of the cases, enim and nam are not translated at all. These asyndetic constructions, however, are not complex and do not present the speaker’s opinion on the state of affairs. This is entirely different in the Modern German translation (Neitzke 1959 in Grasmück 1997), which employs the regular connectors PDG nämlich (twenty‑eight instances) and PDG denn (twenty‑one instances) in about equal numbers and only occasionally uses other forms such as nun (1), jedenfalls (1), in der Tat (1), freilich (2) and wirklich (1). In only three instances, nam or enim are not translated into German. These differences are not due to individual translators or errors of the Modern English translation (the figures are almost identical for the translations by Green 1962 and Watts 1969), but to the fact that English has lost a regular adverbial connector for the sequence result – cause. 9.5.6.  Subordinators: for as much as, since, because This transition of the all‑purpose causal connector OE forþæm to a transitional connector indicating the line of argument has also yielded changes in the system of the subordinators, i. e. the emergence of for as much as, since (that) and because, first mainly in pre‑posed subordinate clauses (see Table 9.12). Because of the proliferation of instances of for as a rhetorical marker of the line of argument, non‑deictic for was less frequently used in these – usually thematic – pre‑posed clauses, since it did no longer mark these contexts unambiguously enough. The comparison of the translations of the Consolatio in the various periods of English shows that since, as and because are never used as renderings of Latin adverbial connectors such as igitur or quare (i. e. the sequence cause – result) or nam and enim (i. e. the sequence result – cause) before the Modern English period, but only for subordinators such as Latin quoniam and causal (not temporal) cum. While therefore and then (rendering Lat. igitur or quare) and for (rendering Lat. nam or enim) are rather stable from the middle or end of the Middle English period until Modern English (in the case of therefore even until today), we find more variation for quoniam and causal cum.18 The earliest expression 18 Unfortunatly, Latin quia is only attested once in the selected corpus. This, however, mirrors the relations in the full text rather precisely. The numbers for the full text: igitur (157), nam and namque (162), enim (141), quoniam (49), quare (27), quia (9). This also shows the centrality of the adverbial connectors in Latin.

Causal connectors in the history of English   165

coined to avoid ambiguities is ME for as moche as (see MED, s. v. for as moche as), which is attested throughout Middle and Early Modern English. It is most remarkable that for as much as is the only one of the subordinators coined in Middle English which has not survived into Present Day English (see Kortmann 1997: 331–332). It is most likely that this form (probably modelled on Old French causal pourtant; see also Spanish por tanto ‘therefore’) was given up because it was neither deictically nor lexically transparent and because it was too close to the then proliferating sentence‑initial rhetorical for. While the temporally deictic since is occasionally found from Middle English onwards, often rendering Latin temporal cum (with indicative) and causal cum (with subjunctive), because is first mainly attested in the collocation for because (in five instances in Colville 1556), i. e. in a combination of the causal connector for and an original prepositional phrase (see also OED, s. v. because and MED, s. v. because), but is subsequently also, though for some time rather sparingly, used on its own. Only after 1750, because finally replaces for and becomes the all‑purpose connector of Present Day English (for quantitative data, see Claridge and Walker 2002). A rather typical instance of the various renderings of a pre‑posed causal clause introduced by Latin quoniam is (128) Quoniam igitur agnovisti, quae vera illa sit, quae autem beatitudinem mentiantur, nunc superest, ut unde veram hanc petere possis agnoscas (Book III, Prose 9; Grasmück 1997: 154). (128a) Thanne, for as moche as thou hast knowen whiche … now byhoveth … (Chaucer, COBEOTH). (128b) Now for bycause thou hast knowen …, then now … (Colville 1556; CEBOETH1). (128c) Because thou knowest now, what … , now it followith … (Queen Elizabeth 1593; CEBOETH2). (128d) Seeing then thou knowest already which … it remains … (Preston 1695; CEBOETH3). (128e) Now then, since you know what true happiness is, and the things that falsely seem to offer it, what remains now is that you have to look for true happiness’. All in all, (for) because is still very rare in the Early Modern English material investigated (eighteen instances including the collocation for because). While it occasionally occurs to render pre‑posed Latin quoniam in Colville (1556), it is later increasingly used for marking post‑posed external causal connection (see the use in Preston 1695). In view of the argument presented above, it has become the central connector in Present Day English because it is lexically transparent.

166   Shifting deictics in the history of English causal connection 9.6.  Conclusions This analysis of systematic changes in the forms of clausal connectors in the history of English has shown that the system of subordinators and adverbial connectors expressing the relation cause/result was almost completely re‑structured after the Old English period. While Old English virtually only had one all‑purpose connector – the pronominal connector forþæm –, Modern English uses different linguistic items for coding the various hierarchies and relations of cause and result. Similar to processes also attested for French creoles (French parce que – creole pas; see above, Chapter 5.5.2–5.5.3), the central Old English connector forþæm (and the variants forþon and forþy) lost its deictic value and was – throughout the Middle and Early Modern English period – used in its highest token number not as a causal connector, but as the text‑structuring connector for. This excessive use of this rather unspecific “recursive for” from late Middle to Late Modern English (see Appendix B.3.3), however, finally led to its becoming more and more infrequent; for is now, if used at all, considered to be archaic (Peters 2006, s. v. for). As a result, Present Day English lacks an adverbial connector expressing the relation result – cause. Since for was – in particular in specialized contexts such as pre‑posed thematic causal clauses – not unambiguous enough, new subordinators such as for as much as, since and because were coined, which, when used in these positions, i. e. introducing pre‑posed causal clauses, helped to disambiguate alternative structures and thus facilitated information processing. The collapse of both of the paradigms of the demonstratives at the beginning of Early Middle English, yielding the homonymic form that for the demonstrative and the subordinating particle, triggered the coinage of a number of new adverbial connectors in the first two sub‑periods of Middle English (ME1 and ME2), most of which were basically modelled on the Old English pattern of pronominal connectors (cf., e. g., for that, therefore or wherefore). The most likely candidate for a new adverbial connector, for that, was only sporadically used because of the homonymy of that with the subordinating particle that, which was increasingly used as “pleonastic that” in various kinds of subordinate contexts (as a complementizer, in relative and all kinds of adverbial clauses). Early Middle English thus emerges as a period of experiment, when speakers sought to express the relations by other means, mainly by time or space deictics. Yet, of the connectors formed by this experimental pattern, only therefore has survived (though in Present Day English even therefore does not seem to be transparent in its space deixis for most speakers; see OED, s. v. therefore). This is different with younger connectors used since the end of the

Conclusions   167

Middle English period. Adverbial connectors such as hence and after all, for instance, are still transparent in their space and time deixis, because they may still be used as adjuncts with their original spatial (“Go from hence”; OED, s. v. hence, I.) or temporal meanings (“After all this had happened …”). Another group of connectors emerging from the end of the Middle English period onwards indicates, however, that English has now almost completely abandoned its original structural pattern which explicitly (by, for instance, demonstratives) or inherently (by time or space deixis) marked deixis in connectors. English now also signals causal connection by purely lexical means: this is true for adverbial connectors such as consequently or accordingly but in particular for the high‑frequency item because (cf. the noun and verb PDE cause). With this concentration on lexical means, English has again moved far away from the Germanic system which is still well alive in Present Day German. This exemplary scenario of causal/resultive connectors is most similar to the developments in the field of concessive/contrastive connectors, which will be discussed in the following chapter. This does not come as a surprise because various recent typological studies could show that the so‑called cccc‑relations or four‑c relations cause, concession, condition19 and contrast often show parallel developments (see the papers collected in Couper‑Kuhlen and Kortmann 2000).

19 Condition is not relevant for the present study because it is not expressed by adverbial connectors, but only subordinators or, rather, correlative constructions (e. g. if … then, when … then). For details, see above, Chapter 2.4.

10.  contrast/concession 10.1.  cause – condition – contrast: affinities As has been pointed out in the research report on connectors above (Chapter 1), the semantic relation of concession has received by far the most attention in research to date. It has been discussed in a considerable number of lexico‑semantic studies analysing its general logical properties (on Present Day English, see, for instance, Halliday and Hasan 1976: 250–256; Quirk et al. 1985: § 15.40–15.43). Cross-linguistic studies focus on its differences and similarities with the other ccc‑relations, i. e. the semantic relations condition, cause and contrast. Because of these – mostly still transparent – affinities and interdependencies, concession has also played a major role in studies on language universals and long-term developments in language change, such as grammaticalization (see, e. g., König and Traugott 1982; König 1985a, 1985b, 1988; Rudolph 1989, 1996; González-Cruz 2007; Sorva 2007). Recently, a number of studies have approached the topic from a discourse‑functional perspective (Couper‑Kuhlen and Thompson 2000; Barth‑Weingarten 2003; Barth‑Weingarten and Couper‑Kuhlen 2002). In spite of all this research, however, it has not been possible to define the notion concession in a more generally accepted way: the many different approaches reflect the complexity of this relation, which is – in spite of its complexity – a very basic one. Its complexity is mirrored in the fact that it is acquired very late in language acquisition (even after the already fairly late conditionals; see Bowerman 1986). Its basicness is reflected by the fact that concessives are attested in most of the languages studied in broad cross‑linguistic analyses of subordinators (see Kortmann 1997: 343). For the present study of the diachrony of adverbial connectors, the relation of concession to contrast proves to be much more essential than its relation to cause (concessivity can be said to express a causal relation that remains unfulfilled or ineffective) or, in particular, condition. For the development of concessive subordinators, on the other hand, the relation to conditionals – via concessive conditionals – has played a much more important role (see subordinators such as PDE even if, even when consisting of a focus particle and a conditional). These developments are of only marginal interest to the present study because they are – as their origin in conditionals suggests – mainly relevant for the coinage of subordinators and not adverbial connectors. These differences in origin are related to the differences in information structure of hypotactic

concession

– contrast   169

and paratactic structures: Conditional relations, which are only expressed in subordinate constructions (see above, Chapter 2.4), are very likely to give rise to subordinators and not – at least not immediately – to adverbial connectors. 10.2.  concession – contrast 10.2.1.  concession For a better understanding of the developments and, in particular, the instabilities in the field of adverbial connectors in the history of English, it is necessary to provide working definitions for the semantic relations concession and contrast. As in all structures discussed in the present study, a prototypical concessive relation consists of two propositions realized in two different (sentential) structures, here called connects (see properties M3 to M5, discussed above, Chapter 2.2). The concessive connection between the two propositions is prototypically made explicit by a subordinator such as although: (129) (Al)though he is poor, he is satisfied with his situation. In the concessive relation “although p, q” we have one proposition in each of the connects

(129a)  p: He is poor. (129b)  q: He is satisfied with his situation.

Both p and q are thus entailed. Normally, however, the propositions p and q do not go together (“in spite of p, q”). Thus the standard implication is “normally (if p, then not‑q)”, i. e. the implication is:

(129c)  If someone is poor, she or he is not satisfied with her or his situation.

10.2.2.  contrast In relations of contrast or adversativeness “q, whereas p”, on the other hand, there is no standard assumption that under normal circumstances p and q should not go together (‘as against’).1 1 All examples in these sections are taken from either Halliday and Hasan (1976: 250–255) or Quirk et al. (1985: § 15.39–15.44).

170   (130)

contrast /concession

(130a)  Peter is poor, whereas Harry is rich. (130b)  Peter is poor, but Harry is rich. (130c)  Mr. Larson teaches physics, while Mr. Corby teaches chemistry. (130d) I ignore them, whereas my husband is always worried what they think of us.

Just as in all semantic relations (for cause, see above Chapters 9.2.4–9.2.5), contrast can not only express external, socio‑physical aspects as in examples (130a–d). In (131), taken from Alice in Wonderland, but does not relate to the socio‑physical reality, but to the act of communication (i. e. expresses an “internal” aspect). Alice recognizes that her suggestion – which was made with the intention of being helpful – may in fact not be of any use. (131) … you might catch a bat, and that’s very much like a mouse, you know. But do cats eat bats, I wonder? contrast here is equivalent with internal concessivity, i. e. counter‑expectancy on the level of the communication process – cf. external “in spite of the facts” vs. internal “in spite of the communicative roles we are playing, in spite of the state of the argument”. This internal concessive meaning “as against what the current state of the communication process would lead us to expect” may also be expressed by connectors which only signal these internal relations emphatically in a kind of “avowal” (Halliday and Hasan 1976: 253–255). This term already shows that connectors such as in fact, as a matter of fact, actually, to tell you the truth are at the interface between purely contrastive/concessive connectors and transitional ones (see below, Chapter 12). Other sub‑categories of contrastive connectors which are also purely internal (i. e. relate to epistemic or conversational aspects) are those which imply a sense of correction, of either a correction of meaning (cf. “antithetic” connectors such as instead, rather or on the contrary) or a correction of the wording (cf. “reformulatory” connectors such as at least, rather or I mean). Another special case of connectors expressing predominantly internal contrastive/concessive relations are generalized forms of the adversative relation “no matter (whether … or not), still …” such as in any/either case, any/either way, any‑ how, or at any rate. Taken by themselves, these “dismissives” (for the term, see Halliday and Hasan 1976: 254–255) have nothing cohesive about them, but they always presuppose that something has gone before, i. e. “however that may have been”. Accordingly, this generalized sense is indicated by universal quantifiers such as all, any, or ‑ever. These originally non‑cohesive forms recurrently yield – by the conversational implicatures sketched above – contrastive connectors proper, the most prominent case in English being however.

concession

– contrast   171

The various sub‑categories of concession/contrast and the Present Day English items which express these relations are summarized in Table 10.1 (adapted from Halliday and Hasan 1976: 255–256): Table 10.1.:  Connectors for external and internal relations of contrast/concession (A) Concessive Relations proper ‘in spite of’ (external and internal) but; yet; though however, nevertheless, all the same, (B) Contrastive Relations (‘as against’) (external) but, and however, on the other hand, at the same time, (C) Contrastive Relations (‘as against’) (internal) , as a matter of fact, to tell the truth, actually, in point of fact (D) Corrective Relations (‘not … but’) Correction of meaning – antithetic: instead, rather, on the contrary Correction of wording – reformulatory: at least, rather, I mean (E) Dismissive (Generalized Adversative) Relations (‘no matter … still’) (external and internal) Dismissal closed: in any/either case/event, any/either way, whichever Dismissal, open‑ended: anyhow, at any rate, in any case ambiguous examples

10.2.3.  contrast/concession: information structure In introductory accounts of the relation contrast/concession, we usually find statements or charts claiming that the relations of contrast and cause may be expressed by either subordinators or by adverbial connectors. See, for instance, the account in the Comprehensive Grammar (Quirk et al. 1985: 1088):

Subordinators

Conjuncts

concessive contrastive

whereas, while whereas, while

yet on the other hand, instead

In their more detailed accounts of the hypotactic constructions, however, we find that Quirk et al. (1985: § 15.40) introduce examples such as (132) No goals were scored, although it was an exciting game. (133) It was an exciting game, although no goals were scored.

172  

contrast /concession

as illustrative of the concessive relation, but they go on saying that “they also imply contrast between the situations between the two clauses” (Quirk et al. 1985: § 15.40), in particular in examples where the propositions p and q may be placed in either of the connects. This signals an equal weight of both propositions. This “mixture of contrast and concession” (Quirk et al. 1985: § 15.43) is primarily found in post‑posed concessive clauses. Because of the defining property of external counter‑expectancy for concessive relations proper, these are typically expressed by subordination, where there is only one central information entity and the apparently contrary or dismissed information is given in the subordinate clause. In accordance with the principles of text information structure sketched above in Chapter 2.4, the information given in the subordinate clause (in the case of concession: the dismissed information) is not pursued in the further discourse. Even more specifically, concession is commonly expressed by pre‑posed subordinate clauses introduced by although (informally though)2 or while, and may be made even more explicit by a correlative construction (i. e. by an adverbial connector in the second connect): (134) (Al)though Sam had told the children a bedtime story, June told them one too. (135) While he has many friends, Peter is (nevertheless) often lonely. The preferred pre‑position of subordinate adverbial clauses can be explained in terms of processing factors (see, for instance, Diessel 1996: 77). The introduction of the concessive connector at the beginning of the sentence guides the hearers or readers in their interpretation by signalling that the expectations raised by the assertion to follow will not be fulfilled: an underlying expected chain with the then following proposition will be broken. If, however, the introduction of the concessive connector is deferred until the middle of a sentence, a reinterpretation of the preceding assertion will be forced. Speakers and, in particular, writers, who are not able to signal the non‑validity of the proposition in the subordinate clause by intonation or gestures, therefore often choose to mark the concessive relation in both the connects, i. e. by correlative constructions (i. e. although … nevertheless, although … yet, although … nonetheless, although … still), in all cases in the sequence subordinate clause –main clause. Until the eighteenth century, this double marking was by far the most common construction for (unambiguous) concessives.

2 For examples from the LLC‑Corpus, see the detailed analysis, below, Chapter 10.7.

Cross‑linguistic patterns in the origin of concessive connectors   173

Because of the equivalent illocutionary weight of both of the respective connects in paratactic constructions, adverbial connectors, on the other hand, inevitably carry a meaning of contrast (and not a pure meaning of concession). For the present analysis of adverbial connectors, this means that it has often been impossible to decide whether a construction is indeed unambiguously concessive. For this reason, the label chosen is contrast/concession, since contrast is predominant in most of the cases. 10.2.4.  Subordinators – adverbial connectors: mood distinctions While the sub‑categorization into contrast and concession is hard to make, it is – in contrast to the difficulties in differentiating causal/resultive adverbial connectors from subordinators – not at all problematic to disambiguate homo­ nymous forms functioning as subordinators or adverbial connectors in the relation of contrast/concession, at least until the time when verbal forms marked mood distinctions. As has been shown above (Chapter 5.3.2), these differences are clearly signalled by the mood of the respective verbs (i. e. subjunctive for the subordinate clause vs. indicative in the clause with an adverbial connector). In her analysis of mood distinctions in subordinate clauses, Moessner (2006) finds that the subjunctive was fairly persistent in concessive clauses in Early Modern English, when still about 65 per cent of though‑clauses exhibit subjunctive forms (Moessner 2006: 258). This close connection of subordinate clauses and the subjunctive – in all languages which distinguish subjunctive forms – is not only evident in originally verbal phrases such as albeit or howbeit, comprising the subjunctive be (most probably modelled on French soit; for details see Sorva 2007), but also in subordinate constructions without a subordinator (although they then generally require a correlative adverbial connector in the other connect): (136) Be that as it may, I (still) think …. 10.3.  Cross‑linguistic patterns in the origin of concessive connectors Cross‑linguistic studies on the diachrony of concessive connectors have shown that most of these, regardless of whether they are realized by free morphemes, prepositions (PDE despite, notwithstanding), conjunctions (although) or adverbial connectors (nevertheless, nonetheless), are more complex in nature than the connectors for the other relations. This is already attested in the inventory of Old English adverbial connectors marking the relation concession:

174  

contrast /concession

apart from the additive þærtoeacan ‘also’, only the concessives naþylæs and swaþeahhwæðere ‘nevertheless’ consist of more than two morphemes. Both of them are, however, more complex than the fairly straightforward compound adverb þærtoeacan (consisting of the space deictic þærto ‘thereto’ and the lexical eac ‘also’) because they require a processing of their morphological structure: OE naþylæs and swaþeahhwæðere are thus clear examples of the iconic principle that formal complexity corresponds to conceptual complexity (for the compound þærtoeacan see below, Chapter 11.2.4). Most of the European languages have a wide variety of concessive markers and, apparently, newcomers are constantly added to this class in all of the languages (König 1985a, 1985b). In comparison to connectors for other semantic relations, most of these concessive connectors seem to have developed relatively late in the history of the respective languages (for subordinators, see Kortmann 1997: 329, 333, 347; for the new set of French concessives, see Klare 1958). In contrast to other markers, most concessives – because of their comparative youth and also because of their morphological complexity – have a fairly transparent history. König (1985b) has shown that, in a wide variety of unrelated languages, other meanings and functions of concessives can be easily identified and concessivity can be related to other semantic domains. On the basis of such affinities, he distinguishes five major groups which share certain aspects in their origins: (I) (II)

Concessives often contain a component which is also used as a universal quantifier, such as French tout, German and English all‑: cf. French toutefois, tout …. que, German bei all ….; allerdings Originally conditional (if) or temporal (when) are used together with an additive focus particle or emphatic particle like English even, also, too, such as German ob‑gleich; ob‑wohl; ob‑schon; wenn‑gleich; wenn … auch, Latin et‑si, French quand même, lorsque même.

This pattern has been, as is shown above, important for English concessive subordinators such as even though. (III) The connectors may signal a remarkable co‑existence or co‑occurrence of two facts as part of their literal meaning, (a) asserting that one fact p does not prevent another fact q (b) asserting that fact p does not pay attention to another fact q (c) by the simultaneity of p and q (d) by the unhindered continuation of q given p; cf. German ungeachtet, unbeschadet, gleichwohl, dennoch, or French tout de même or cependant.

The diachrony of English contrastive/concessive adverbial connectors   175

In addition to these three major sources, concessives may also be formed from items which (IV) directly lexicalize notions of conflicts, obstinacy, dissonance or our reactions to such situations, such as French en dépit de, au mépris de, German trotz, trotzdem or Spanish a pesar de (‘for sorrow/regret of’). Further, they may go back to original epistemic items which (V) mark the factual character of something (for these, see also below, Chapter 12), cf. German concessive gewiss or zwar (< mhd. zware from ze ware ‘(in) truth’; Kluge and Seebold 2002, s. v. zwar). 10.4. The diachrony of English contrastive/concessive adverbial connectors 10.4.1.  General tendencies Since most of the relevant structural and lexico‑semantic features of the development of contrastive/concessive connectors in the history of English are either illustrated above (in Chapter 9 on the causal connectors) or are discussed in detail in the relevant diachronic literature on concessives (see König and Traugott 1982; König 1985a, 1985b; González‑Cruz 2007; Sorva 2007), the following sections will only summarize the main findings for the history of adverbial connectors in English (sections 10.4–10.5). A more detailed investigation will then provide an in‑depth study of the meandering paths in the history of Present Day English adverbial though and, in particular, its functions when placed sentence‑finally in spoken discourse. 10.4.2.  Antithetic/reformulatory adverbial connectors Table 10.2 (based on Appendix A.1 and Appendices B.4.2 and B.4.3) shows the general tendencies in the history of antithetic and reformulatory connectors, i. e. those connectors which signal a correction in the meaning and the wordings (“Corrective Relations”, see above, Table 10.1 (D)). As is evident at first glance, contrastive/concessive connectors are fairly atypical adverbial connectors, because they have never been realized as pronominal connectors; neither are other deictic means of linkage employed. Most

Table 10.2.:  Corrective (antithetic and reformulatory) adverbial connectors in the history of English 1570–1640 (EModE 2)

1500–1570 (EModE 1)

1420–1500 (ME 4)

1350–1420 (ME 3)

1100–1350 (ME 1/2)

Old English contrarily

[contrarily] [contrarily] contrary [contrariways] on the other hand

[contrarily]

on the other hand

[contrarily]

on the other hand

or else

[contrarily]

on the other on the other hand hand rather instead

contrarily

or else or else on the other side

[otherways] [otherways] [otherways] otherwise [otherwise] [otherwise] [otherwise] [otherwise] [contrariwise] [contrariwise] [contrariwise] contrariwise [contrariwise]

1640–1710 (EModE 3)

[otherways] [otherways] [otherwise] otherwise contariwise contrariwise

1710–1780 (LModE 1) or else on the other side

1780–1850 (LModE 2)

or else on the other side

1850–1920 (LModE 3)

or else or else on the other side

1920– (PDE)

elles or elles or elles or elles oðer side on oðer half otherways [otherways] [otherwise] [otherwise] contrariwise

176   contrast /concession

The diachrony of English contrastive/concessive adverbial connectors   177

of today’s contrastive/concessive connectors are lexicalized nominal or prepositional phrases, which are still comparatively transparent, precisely because they comprise lexical elements. One sub‑group expresses the correction by an adjectival or nominal lexeme meaning ‘contrary’, such as conversely or contrary in contrary, contrarily, contrariwise and contrariways (cf. also by/in contrast, on the contrary which have not been included in the present study because they are fully transparent; see above, Chapters 1.6 and 7.4). Another sub‑group conveys the sense of correction by phrases comprising the adjective other, for example in lexicalized nominal phrases, such as ME other side, on other half or on the other hand (with a metaphorical extension of the spatial or body noun on the written space); the adjective other is also found in original prepositional phrases, such as otherwise and otherways (< in/on other wise; for the suffix ‑wise/‑ways see above, Chapter 7.3.3). Reformulatory else is almost exclusively attested in collocations with the adversative/reformulatory coordinating conjunctions but and or and thus contrasted to additive ellis/else (which is found in collocations with and; see Appendix B.4.3). The OED (s. v. else) even bases its introductory remarks on else on its synonymy with other: “In this use else, like its synonym other, admits contextually of two different interpretations: e. g. something else may mean ‘something in addition’ to what is mentioned, or ‘something as an alternative or a substitute’”. Both of the forms which come to be used as adverbial connectors in Present Day English – instead and rather – are formed from Germanic word material and originally go back to Old English. In their original meanings, they intriguingly denote two opposite spatial‑temporal concepts: OE stede ‘standing still, as opposed to movement’ (see OED, s. v. stead n., 1.) as opposed to OE hraþor ‘faster’, the comparative of hraðe/hræðe ‘quickly, rapidly’ (see OED, s. v. rathe, adj. “Denoting rapidity in the performance or completion of an action”). Although the noun stead is no longer used in English on its own, instead (shortened from the anaphoric, pronominal form instead of this) is still transparent in its spatial origin through compounds such as homestead or place names in ‑stead such as Grinstead, Halstead or Hampstead (on the diachrony of instead, see Fischer 2007: 277–280). Rather, on the other hand, does no longer seem to be transparent in its temporal‑spatial origin for speakers of Modern English (see Rissanen 1999a). It is thus the only of these adverbial connectors marking correction which is not transparent.

178  

contrast /concession

10.4.3. Contrastive/concessive connectors in English: general tendencies Table 10.3 below lists the adverbial connectors expressing the other external and internal sub‑groups of the semantic relations contrast/concession (see Table 10.1, sub‑groups (A), (B), (C) and (E)). This chart shows that the developments in English adverbial connectors agree with what has been found for other languages and also for the history of English subordinators. English has – like most of the European languages – a wide variety of concessive markers, and newcomers have been added to this category over all periods of English. Parallel to the development in other languages, also many of the contrastive/concessive connectors have developed rather late in the history of English. The adverbial connectors emerging from the middle of the Early Modern English period onwards have shown themselves to be fairly persistent (only the variant howsoever has not survived into Present Day English), in contrast to the concessive subordinators coined in the younger periods of English, most of which were fairly ephemeral. 10.4.4.  Shifting deictics in English contrastive/concessive connectors The history of contrastive/concessive connectors reinforces most of the general tendencies found for the development of English causal connectors. Old English employed a number of explicitly deictic items,3 namely forms comprising swa ‘so’, such as swaþeah and swaþeahhwæðere, and the – though rare – pronominal connector naþylæs ‘no + the + less’ comprising a form of the demonstrative (locative or instrumental þy/þe). Of these, only nevertheless has survived into Present Day English. Its constituent the, however, is no longer transparent as a form of the demonstrative and is classified as an adverb in the OED (s.vv. the, thy). The increasing loss of transparency of þe/þy after the Old English period (see above, Chapter 7.5) is illustrated by the many variant forms of the word during the late Middle English and Early Modern English period, in particular in spellings which fuse the original demonstrative element and thus render the form fully opaque (see OED, s. v. natheless and OED, MED, s.vv.): consider, for example ME nadleez, naughtles, naþles or ME/EModE nathles, naithlesse, nathlesse, nathless, nothless, or forms which abandon this element altogether, such as ME nether‑les, never‑les(se) or neverlater. Interestingly, the MED (s. v.) only labels nether‑les an “aberrant form” (citing about fifteen instances, all of them from the fifteenth century). 3 For the use of the term “deixis” in the present study, which includes discourse dei­ xis, see above, Chapter 9.4.1.

1100–1350 (ME 1/2)

Old English [nought-forthan] nevertheless [nevertheless]

yet

1570–1640 (EModE 2)

1640–1710 (EModE 3) notwithstanding

[nevertheless]

[nevertheless]

yet

[nevertheless]

yet

°though

[nevertheless]

yet

°though

[howsomever]

[howsomever]

[howsomever]

[howsom‑ ever]

[nonetheless] [nonetheless] [nonetheless] [nonetheless] nonetheless

notwithstanding

[nevertheless]

yet

1710–1780 (LModE 1)

notwithstanding howbeit [nonetheless] albeit [howsomever]

[nevertheless]

yet

1780–1850 (LModE 2)

[nought-for-that] [nought-for‑that] [nought‑for‑thi] [nought‑for‑thi] algates [algates] [algates] notwithstanding notwithstanding howbeit howbeit [nonetheless] albeit [howsomever]

[nought-forthan] nevertheless

natheles natheles yet yet [neverthelatter] neverthelatter

1420–1500 (ME 4)

natheles yet [neverthelatter] [nought-forthan]

1350–1420 (ME 3) °tho

1500–1570 (EModE 1)

°þeah

1850–1920 (LModE 3)

°þeah þeah­ hwæþere swaþeah

1920– (PDE)

huru huruþinga °þeah þeah­ hwæþere swaþeah swaþeahhwæþre naþylæs yet

Table 10.3.:  Adverbial connectors marking contrast/concession in the history of English The diachrony of English contrastive/concessive adverbial connectors   179

however howsoever at least [in the meantime] [meanwhile]

1500–1570 (EModE 1)

1420–1500 (ME 4)

1350–1420 (ME 3)

1100–1350 (ME 1/2)

Old English

1570–1640 (EModE 2) after all still at the same time

after all still at the same time of course in fact [anyhow] in all events

at least in the meantime [meanwhile]

1710–1780 (LModE 1) at least [in the meantime] [meanwhile]

1780–1850 (LModE 2) however

however

at least [in the meantime] [meanwhile] after all after all still still at the same at the same time time of course of course in fact in fact [anyhow] [anyhow] in all events in all events [anyway] [anyway] [all the same] [all the same] in any case in any case

at least [in the meantime] [meanwhile]

however

1850–1920 (LModE 3)

however

1920– (PDE)

[after all] still

1640–1710 (EModE 3)

however howsoever at least (in the) meantime (the) meanwhile

180   contrast /concession

The diachrony of English contrastive/concessive adverbial connectors   181

As in the case of the causal connectors (see above, Chapter 9.5), speakers and writers of Middle English tried to preserve these pronominal forms or, alternatively, coined new ones with the new demonstrative that (cf. ME nought‑for‑than, nought‑for‑thi ‘nevertheless’ and nought‑for‑that ‘nevertheless’, all of them ambiguous adverbs/conjunctions). Yet, with the exception of nevertheless, none of these has survived the Middle or Early Modern English period. Middle English thus again emerges as a period of experiment and transition (see above, Chapter 6.6). After the Middle English period, we find, just as in the history of causal connectors, a number of newcomers which – instead of demonstrative deixis – employ time deixis as a frame of reference, such as (in the) meantime, (in the) meanwhile, after all, still, and at the same time (see above, Chapter 8.2). While German also employs originally time deictic elements (see the list in Rudolph 1996: 4–5), it has – again as in the case of causal and additive connectors – preserved many connectors comprising a specifically pronominal element, such as PDG dessen ungeachtet, indes, indessen, nichtsdestoweniger, nichtdestotrotz (colloquial; mixture of nichtsdestoweniger and trotzdem), trotzdem or an element of space deixis such as da ‘there’ or hin ‘hence’, such as PDG dagegen, hingegen or archaic hinwiederum. As in the semantic relation of cause, English adverbial connectors expressing contrast/concession have virtually given up the old system of pronominal connectors, which is still very much alive in German, and have sought other means for the coinage of new adverbial connectors.

10.4.5.  Patterns in the origin of English concessive connectors Following the abandonment of the coinage of pronominal connectors after the Middle English period, speakers and writers of English have employed those means which have above been introduced as cross‑linguistic patterns for new coinages in the field of new contrastive/concessive connectors (see König 1985b, and above, Chapter 10.4, patterns (I)–(V)). For the history of adverbial connectors, Patterns (I) and (III) are most important: Pattern (I): Universal quantification (OE hwæðer ‘whether’; PDE all, any, ‑ever) OE hwæðer ‘which of two, each of two, either, any’ OE hwæðer(e) þeahhwæðere swaþeahhwæðere

182  

contrast /concession

all

ME EModE1 EModE2 LModE3 LModE3

algates (lit. all gates ‘every way’; ON alla götu) albeit after all in all events all the same

any

LModE2 LModE3 LModE3

anyhow anyway in any case

‑ever

EModE1 EModE2 EModE2

howsomever however howsoever

As has been pointed out above, these forms comprising universal quantifiers are intrinsically cohesive because they always presuppose that something has preceded. Their path to becoming concessives is most clearly seen in forms such as however, howsoever and howsomever, since they are relics of original subordinate clauses “how(so)(m)ever this may be” (see OED, s. v. however), comprising in the full clause an anaphoric element such as this. By the conventionalization of these conversational implicatures, these dismissive connectors may become contrastive connectors proper (cf. however). Their generalized sense, however, recurrently leads to further semantic bleaching, so that many of these items follow the “adverbial cline” from adverb to discourse marker suggested in the semantic‑pragmatic approach to grammaticalization (see Traugott and Dasher 2002; for a detailed discussion, see above, Chapter 8.4). It is therefore probably no coincidence that neither anyhow (which is stated to be the fourth most frequent “linking adverbial” in spoken British English) nor anyway (fourth most frequent “linking adverbial” in spoken American English; Biber et al. 1999: 887) are attested in my corpus, which mainly consists of written texts, all of which are non‑interactive (on the discourse structuring function of anyhow and anyway, see Lenk 1998; on after all, see Traugott 1997 and 2004: 554–560). Pattern (III): Remarkable co‑existence or co‑occurrence of two facts Pattern (IIIa): Negation (asserting that one fact p does not prevent another fact q) As the many adverbial connectors (and subordinators) comprising an element of negation (see na‑/no, none‑, never‑) show, there is also a close correlation between counter‑expectancy as one of the properties of concession/contrast

The diachrony of English contrastive/concessive adverbial connectors   183

and negatives. Commonly, negatives are only articulated in contexts where the corresponding affirmatives have already been discussed, or where the speakers or writers assume the hearers’ or readers’ belief in – or at least their familiarity with – the corresponding affirmative. Assumptions like these have led to the concessive meanings of English items comprising a negative. These are attested over all the periods of English, in various pronominal and non‑pronominal patterns.4 no + for + demonstrative ME nought‑for‑than, nought‑for‑thi, nought‑for‑that no + þy/þe + less OE ME EModE1

naþylæs natheless, nautheless ‘naught + the + less’, netheless, not‑the‑lesse, notheless, noughttheless nonetheless

In a mixture of Pattern (I) (universal quantifiers) and (III) (negation), we find forms comprising the negative of the universal ever, never:5 never + the + less ME

nevertheless (variants: ner‑the‑les, never‑ther‑les, never‑tho‑les)

never + the + latter ME

ner‑the‑latere, ME/EModE never the later

As a further negative, we find notwithstanding (a calque on French non ob‑ stant), which is only used as a preposition in Present Day English, but is also habitually employed as a contrastive/concessive adverbial in the prepositional phrase notwithstanding this. Pattern (IIIb): Simultaneity or unhindered continuation Communicative principles have also led to the concessive meaning of certain temporal adverbs signalling simultaneity or unhindered continuation, in particular yet and still (see König and Traugott 1982; König 1985a, 1985b), but also 4 In addition to the connectors listed below, the MED also lists the coinages ‘not + the + more’ (ME no‑the‑mo, EModE nathemore, nathemo) and never‑the‑whether, for which I could find an attestation as an adverbial connector neither in the dictionaries nor in the corpus texts, i. e. they are not attested as adverbial connectors on the level of the sentence or discourse. 5 Again, the MED lists forms such as neverthemore, for which I could not find an attestation as an adverbial connector on the level of the sentence or discourse.

184  

contrast /concession

the phrases (in the) meantime and meanwhile, which are, however, only rarely used as concessives in Present Day English (see OED, s.vv. meantime, 1b, 2., meanwhile B.2). In its draft revision entry on meanwhile, the OED (apparently unaware of the research on concessives) inherently comments on these conversational implicatures by stating that “although this is now rare as a distinct meaning, the adversative force is still often present in the use of sense 1 to mean ‘at the same time’. The two senses are not always readily distinguished …”. For such ambiguous, contextually concessive uses in Late Modern English, see, for example, (137) then the consequent practice which they direct shall be explained, and rendered easy of adoption. In the meantime, let no one anticipate evil, even in the slightest degree, from these principles; they are not innoxi­ ous only, but pregnant with consequences to be wished and desired beyond all others by every individual in society (CLOW2). These conversational implicatures are, by contrast, fully conventionalized in yet and still (for examples, see below, Chapter 10.5.3). The origins for the subsequent conventionalization of conversational implicatures contained in the notion “continuation” are clearly seen in sentences such as (138), where the originally temporal meaning can always be augmented by concessive implicatures, if this is consistent with the background knowledge of speaker and hearer: (138) It is midnight and he is still working. In view of the pragmatics of negation (Pattern IIIa) discussed above, such an augmentation is even more likely if the second clause contains a negation, as in (139) It is midnight and he still has not stopped working.

10.5.  OE þeah – PDE though 10.5.1.  Long‑term developments: grammaticalization While the contrastive/concessive connectors discussed so far support earlier cross‑linguistic and English‑specific research on concessives, a more detailed investigation of the history of PDE though seems to challenge one of the prop-

OE þeah – PDE though   185

erties of grammaticalization, the principle of uni‑directionality (for a summary on the state of research, see Brinton and Traugott 2005: 69–76). The term “grammaticalization” owes its origin to work by Meillet, who characterized grammaticalization as the attribution of grammatical character to a previously autonomous word. It is now commonly understood as “that part of the study of language that is concerned with such questions as how lexical items and constructions come in certain linguistic contexts to serve grammatical functions or how grammatical items develop new grammatical functions” (Hopper and Traugott 2003: 1). In the morpho‑syntactic approach to grammaticalization, a number of parameters have been established as prototypical, such as, for instance, increased bonding or syntactic scope reduction (for a full list of criteria, see Lehmann [1982] 1995). Among the well‑known properties of grammaticalization processes is also uni‑directionality (i. e., e. g., from lexicon to grammar, from less grammatical to more grammatical and not vice versa). Most of the newly coined grammatical elements can be shown to follow regular, unidirectional clines. A common cline is the coinage of new conjunctions from – first of all – lexicalized nominal and prepositional phrases; see, for instance, the formation of because from prep. be + N. cause + subordinator (that) (for the distinctive, subsequent processes of lexicalization and grammaticalization in the coinage of new conjunctions in Spanish, see, e. g., Lehmann 2002: 8–12). In view of these facts the path of English OE þeah/PDE though merits a more detailed investigation. This is not evident at first glance, since historical dictionaries and grammars suggest that the various forms of this linguistic item have always been syntactically polyfunctional, i. e. could be used as either adverbs or conjunctions (with mood distinction in the verb of the relevant connect). A closer inspection of the actual language data, however, suggests that this scenario is too simplistic, since þeah was only occasionally used as an adverb from Middle English to the beginning or middle of the twentieth century. In the complete corpus of the present study, adverbial though is only used three times in the Early and Late Modern English period (in three different texts; see Appendix B.4.1 and examples (149)–(151) below), predominantly in post‑first position in questions. This suggests that though has not followed a strictly unidirectional path, but a meandering way from adverb/conjunction to conjunction to, again, adverb/conjunction. Yet, though has – in contrast to such connectors as nu or forþæm – never been a prototypical “ambiguous adverb/conjunction” because of mood distinctions in the verb in earlier English and because of topological distinctions (sentence‑initial vs. sentence‑medial or ‑final) in Present Day English (see below, Chapter 10.6).

186  

contrast /concession

The following account of this path will first sketch the situation in Old English and will then, passing a long period of time in one large step, analyse the re‑emergence of adverbial (sentence‑final) though in spoken Present Day English in the middle of the twentieth century by an analysis of the forms of although and though in the London Lund Corpus of Spoken English (LLC; spoken part). In contrast to the preceding chapter on causals, the section will thus not focus on general developments in the expression of concessivity, since, as has been sketched above, the concessive relation has over all periods more prototypically been expressed by subordinate or, in particular, correlative constructions (and is also discussed widely in the relevant publications; see Quirk 1954; König 1985a, 1985b; Kortmann 1997; González-Cruz 2007; Sorva 2007). For the present study, two other aspects are essential: the re‑analysis of original concessives as markers of contrast, when used as adverbial connectors, and, at the same time, their topological fixation to the middle (written mode) or final position (oral mode). 10.5.2.  OE þeah Similar to the relation of cause, there is in essence only one central contrastive/concessive connector in Old English, þeah (and its variants þeahhwæðere; swaþeah, and swaþeahhwæðere). In the corpus used for the present study, other forms are only rarely attested, and are restricted to certain texts and contexts. Thus, occurrences of hwæðere ‘whether > nevertheless’ are confined to the Blickling Homilies (OE 3), in very generalized, dismissive contexts (“however that may be”; see above, Category (E)). (140) Ac mid þære bysene, he gecyþde þæt soþfæste men habbaþ mid him þeofas & synfulle men; & hwæþere hie sceolan heora yfel geþylde arefnan on him selfum (COBLICK, 175). ‘But with this example he told us that righteous men keep the company of thieves and sinners; and yet [‘however that may be’], they have to suffer for their sins themselves with patience’. Forms comprising a negative, such as naþylæs ‘nevertheless’, are not attested at all in my corpus (similar to nonetheless, which was – according to the OED – coined in the sixteenth century, but is first attested in my corpus material in the Present Day English material of the FROWN Corpus). These findings from the corpus texts are corroborated by a survey count in the Complete Corpus of Old English in Machine Readable Form (DOEC),

OE þeah – PDE though   187

which yields the following numbers (for the spelling variants þeah/ðeah, þeh/ ðeh with smoothing of the breaking, hwæð‑/hwæþ‑ and þy/ðy/þe/ðe): Table 10.4.:  Concessives in Old English: DOEC þeah swaþeah þeahhwæðere swaþeahhwæðere hwæðere þeah naþylæs

ca. 3380 occurrences ca. 960 occurrences ca. 290 occurrences ca. 30 occurrences ca. 15 occurrences ca. 10 occurrences

þeah (and its variants) do thus not have any serious competitors as central markers of the concessive relation in Old English.6 The following survey shows that þeah – just like OE forþæm – could be used as an adverbial connector, but also as a subordinating conjunction introducing pre‑posed, medial and post‑posed subordinate clauses. (141) Oft seo brodige henn, þeah heo sarlice cloccige, heo tospræt hyre fyðera & þa briddas gewyrmð (COBYRHTF, 2.1.225). ‘Often the brooding hen, though she clucks sorrowfully, spreads her wings and warms her chicks’. Just as in Present Day English, pre‑posed concessives are preferred since they facilitate the understanding of the complex relation of “counter‑expectancy” in concessives (see above, Chapter 10.2.4). (142) We wyllað þæt hig understandon þisne cwide, (\Vasa fictilia tanto solent esse utiliora quanto et uiliora\), þeah we wace syn & þas þing leohtlice unwreon, hig magon fremian bet þonne þa þe beoð on leoðwisan fægre geglenged (COBYRHTF, 1.2.198). ‘We wish them to understand this saying: ‘Earthen vessels are more useful the cheaper they are’. Although we are weak and we explain these things cursorily, they may do more good than those that are fairly adorned in verse’.

6 The concessive instances of hwæðere have not been counted because of the polyfunctionality of this word. The comparatively few instances of the collocation hwæðere þeah (15 in comparison with the 290 for þeahhwæðere), however, corroborate my corpus findings that hwæðere was only very rarely used.

188  

contrast /concession

The subordinate status of the pre‑posed clause may be further signalled by the general subordinating particle OE þe, as in (143) ðeah ðe ealle dagas ælce geare habbon heora (\concurrentes\), synderlice se dæg þe byð on (\ix kl Aprl\) getacnað hu fela beoð on geare (COBYRHTF, 1.2.92). ‘Though all the days in each year have their concurrents, the day that falls on 24 March shows how many there are in the year’. The double function of þeah is most obvious in correlative constructions. As has been shown above (see Chapter 5.3.2), this does, in contrast to causal connectors, not present major problems because the functions are differentiated by the mood of the verbs in the respective clause (see the subjunctive scine in the subordinate clause (144a), and the indicative astihð in the main clause (144b)). (144) & ðeah ðe [144a; conjunction] seo sunne under eorðan on nihtlicere tide scine, þeah [144b; adverb] astihð hire leoht on sumere sidan þære eorðan (COTEMPO, 1.32). ‘And although the sun shines under the earth at night, her light still comes on the summer side of the earth’. In contrast to concessive subordinators, contrastive/concessive adverbial connectors tend to express a contrastive rather than the purely concessive meaning of counter‑expectancy. (145) & þeah is [3 Prs. Sg. Ind.] an God, se is stemn & staðol eallra goda; of þæm cumað eall good, & eft hi fundiað to him, & he welt ealra (­COBOETH, 34.86.1). ‘And yet there is one God, who is the stem and the foundation of all goods and from him comes all good; and again, they tend to him and he governs all …’. The adverbial function of þeah is here first illustrated by an example from the Old English Boethius adaptation, not only because this is a comparatively independent Old English text (a criterion most important for the use of the subjunctive), but also because late Old English authors increasingly distinguish the subordinator from adverbial functions by the use of different connectors and different topologies. Already in the Old English Boethius, post‑first or medial position is preferred for þeah in its use as an adverbial connector (in contrast to the obligatory clause‑initial position for the subordinator):

OE þeah – PDE though   189

(146) þæt is þeah micel syn to geðencanne be Gode, þætte ænig god sie buton on him, … (COBOETH 34.84.31). ‘It is, however, great sin to imagine concerning God, that any good be external to him … ’. In Late Old English, authors increasingly try to distinguish the different functions by the use of different lexemes. For the adverbial connector, longer, more complex and more explicitly deictic items are chosen, which signal the equal weight of the propositions of the two connects iconically. Consider the following example with adverbial swaþeah (‘so’ + þeah), which is found almost identically in Byrhtferth’s Enchiridion and Ælfric’s De Temporibus Anni: (147) Synd swa þeah ma heofena, swa swa se witega cwyð: (\Celi celorum\) (COBYRHT, 2.1.257). Sind swa ðeah ma heofenan, swa swa se witega cwæð; Celi celorum, þæt is heofena heofenan (COTEMPO, 1.8). ‘Yet, there are more heavens, as the prophet says: ‘Heaven of heavens’’. For the adverbial connector, Ælfric in particular prefers the unambiguous forms swaþeah, þeahhwæþere (in Catholic Homilies I) or the very complex and therefore highly iconic swaþeahhwæþere (in Catholic Homilies II), mainly placed in post‑first or medial position. There are very few instances of simple þeah used as an adverbial connector (see Godden 2000: Glossary, s.vv. þeah, þeahhwæþere, swaþeahhwæþere). See, for example, the adverbial connector swaþeahhwæðere in post‑first position in (148) Is swa ðeah hwæðere an ðæra eahta winda, (\Aquilo\) gehaten, se blæwð norðan & eastan …. (COTEMPO, 10.23). ‘Yet, there is one of these eight winds, which is called Aquilo, which blows from the north and the east’. As a subordinator, Ælfric almost exclusively employs the complex form þeah þe, comprising the general Old English subordinating particle þe. This shows that even in Old English writers tried to differentiate the connectors for the semantically and cognitively rather complex relation of concession and employed more explicit forms for the adverbial connectors signalling contrast. These were also distinguished topologically and were predominantly placed in post‑first or medial position.

190  

contrast /concession

10.5.3. Contrastive adverbial connectors from Middle English to Present Day English 10.5.3.1.  though In the discussion of the general tendencies in the developments of adverbial connectors in the history of English, we have seen that the inventory of contrastive/concessive connectors is considered the least stable one over all the periods of English (see Kortmann 1997: 347). Yet, most of the new coinages from the Early Modern English period – at least, in the meantime, in the meanwhile, after all, at the same time, of course, in fact, anyhow, in all events, anyway, all the same and in any case – express the “dismissive” sub‑type of concessives (Type (E)) only. The inventory of connectors is much more stable with respect to the contrastive/concessive relation proper (Types (A), (B) and (C)). The forms continuing OE þeah – the Scandinavian loan though7 and the originally emphatic form although (with the universal quantifier all stressing the concession) – which is used as a non‑emphatic form from around 1400, are predominantly used as subordinators from the Middle English period onwards. In my corpus texts, I have only found very few instances of adverbial though from the late Middle English and Early Modern English period (see Appendix B.4.1). Most of these are, however, not attested in contexts prototypical for adverbial connectors. In the example taken from Reynes’ Commonplace Book (ME3), both instances of adverbial though work on the level of the phrase (subject complement: dredfull and þowg gracious): (149) Whoso be born in Aries, he schall be dredfull and þowg gracious. Cancer: he schall ben pore and weyke, and þowg gracious (CMREYNES, p. 241). ‘Whoever is born in Aries, he will be anxious, but nonetheless genial. … Cancer: He will be poor and weak, but nonetheless genial’. 7 The form though is not a continuation of OE þeah, but a Scandinavian loan attested from the twelfth century onwards (probably from unattested ON *þōh). Samuels argues that it spread out – like its “competitor”, the plural pronoun they – from its original Danelaw area in answer to a functional need for clear communication during the fifteenth century. He finds that they was the form first adopted into the southern area, which, however, already possessed that form in the meaning ‘though’ (cf. the attested Middle English spellings þeigh, þei, þey, þeigh, thigh or thei continuing OE þeah; see OED, s. v. though). The resulting “homonymic clash” therefore is seen to have triggered the borrowing of though shortly afterwards (cf. Samuels 1972: 71–71 and 70, map 2 and OED, s. v. though).

OE þeah – PDE though   191

In the two instances from the Early Modern English period, though is used in post‑first position in questions: (150) (Dan.) It is not in these matters to be taken as wee imagine, but as the word of God teacheth. What though a man think he worshippeth not deuils, nor seeketh not help at their handes, as he is persuaded, nor hath any such intent, is he euer the neere, when as yet it shall be found by Gods word, that he doth worship them, and seek vnto them for help? (CEHAND2A, p. B2R). (151) But the vulgar cares not for this. What tho? (CEBOETH2 (Queen Elizabeth), p. 87). In example (152), it is not quite clear whether the second though (152c) – in spite of being strengthened by the disjunct methinks – is indeed used as an adverbial connector or rather as a post‑posed subordinate clause, further modifying the correlative construction although (152a) … yet (152b): (152) Now, though [152a; conjunction] I have with great diligence endeavoured to find whether there be any such thing in those (Microscopical) pores of Wood or Piths, as the (Valves) in the heart, veins, And other passages of Animals, that open, And give passage to the contain’d fluid juices one way, And shut themselves, And impede the passage of such liquors back again, yet [152b; adverb] have I not hitherto been able to say any thing positive in it; though [152c], me thinks, it seems very probable, that Nature has in these passages, as well as in those of Animal bodies, very many appropriated Instruments And contrivances, whereby to bring her designs And end to pass, which ’tis not improbable, but that some diligent Observer, if help’d with better (Microscopes) , may in time detect (CESCIE3A. p. 13.5,116). In sum, we see that though was only rarely used as an adverb in the Middle English, Early Modern and Late Modern English period (see also the paucity of quotes for these periods in the OED, s. v. though I.a). 10.5.3.2.  Yet and still Instead of adverbial though, speakers and writers choose different forms coined on the patterns introduced above, in particular yet and still, which both employ the temporal Pattern (IIIb) “Simultaneity or unhindered continuation”.

192  

contrast /concession

While the conversational implicatures are conventionalized for yet as early as the beginning of the Middle English period, still only shows unambiguous concessive meanings from the beginning of the eighteenth century onwards. These developments will be summarized in the following sections. In its earlier temporal meanings, yet implied a “continuance from a previous time up to and at the present (or some stated) time: Now as until now (or then as until then): = still 4a.” (OED, s. v. yet II.2.a). In Ælfric’s Grammar, for instance, gyt is given as an equivalent of Latin adhuc ‘up to now, as yet, still’ (Zupitza [1880] 2001: 237), but also used to illustrate forms of the Latin future or future perfect, such as “amabo ic lufige gyt to dæg oððe to merien” or “amato tu lufa ðu gyt” in contrast to “amo ic lufige …” (Zupitza [1880] 2001: 130, 131, 133). These temporal meanings are often emphasized by another temporal adverb, as in the collocation with nu in (153): (153) Ac þu ana hwilum bescylst mid oðre eagan on þa heofonlican þincg, mid oðre þu locast nu giet on þas eorðlican (COBOETH, 121.30). ‘But thou alone sometimes lookest with one eye on the heavenly things and with the other thou lookest as yet on these earthly things’. The OED (s. v. yet III.9) and, following the OED, König and Traugott (1982) see the first instances of concessive meanings of yet in the Early Middle English Period. Examples such as (154), however, are certainly ambiguous even in Old English and allow a concessive reading as a conversational implicature, a reading supported by the collocation with the adversative conjunction ac ‘but’: (154) þa cwæð ic: Nis nan þing soðre þonne þæt; ne magon we nanwuht findan betere þonne God. ða cwæð he: Ac ic wolde giet mid sumre bisne þe behwerfan utan þæt þu ne mihtst nænne weg findan ofer (COBOETH, 85.20–23). ‘Then said I, Nothing is truer than that. We are not able to discover anything better than God. Then said he: But I would now/still prepare thee by some examples, so that thou may not find any way of escaping’. A concessive reading is even more likely in instances such as (155), where giet may be seen to function as the first connector in a correlative concessive construction (giet … þeah): (155) Ac ic þe mæg giet tæcan oðer þing ðe dysegum monnum wile ðincan giet8 ungelefedlicre, & is þeah genog gelic þam spelle ðe wit æfter­ spyriað (COBOETH, 38.118.19). 8 Giet here functions as a focus particle on the phrase level, i. e. ‘even’.

OE þeah – PDE though   193



‘But I can now/still teach you another thing, which to foolish men will seem even more incredible, and is nevertheless suitable enough to the argument we are holding’.

The first unambiguous, concessive examples, showing a conventionalization of these implicatures, are found in the Early Middle English period. Yet is, however, not a prototypical adverbial connector in its concessive uses, because it is not free in its position in the sentence, but is restricted to clauseinitial position: (156) A Meiden … eode ut to bihalden uncuðe wummen. get ne seið hit nawt þet ha biheold wepmen (CMANCR, p. 32). ‘A girl went out to watch unknown women. Yet, it does not say that she watched soldiers’. Yet’s function as an adverbial connector is highlighted by its many occurrences in correlative constructions, and in particular, in collocations with conjunctions (ac giet, and yet and but yet) and other adverbial connectors (and yet again, yet nevertheless or yet still) (see Appendix B.4.1). While yet developed its additional contrastive/concessive sense as early as the beginning of the Early Middle English period, still was exclusively used as a circumstance adverb, modifying verbs such as stand in the sense of ‘quietly’, until the end of the Early Modern English period (OED, s. v. still I.). From the sixteenth century onwards, we find uses indicating the continuance of a previous action or condition, so that still acquires the temporal reading ‘now’ and is “contextually” used for ‘as yet’ (OED, s. v. still 4.). (157) Temporal still: (Arch.) Don’t mistake me, (Aimwell), for ’tis still my Maxim, that there is no Scandal like Rags, nor any Crime so shameful as Poverty (Far­ quhar, Beaux Stratagem, CEPLAY3B, p. 3). From these temporal meanings, the contrastive/concessive sense arose at the beginning of the eighteenth century (OED, s. v. still 6.). For very early ex­amples from Farquhar’s The Beaux Stratagem (1707), see (158) and (159). (158) Contrastive still in correlative though … still: (Aim.) Tho’ the whining part be out of doors in Town, ’tis still in force with the Country Ladies; – And let me tell you (^Frank^), the Fool in that Passion shall outdoe the Knave at any time (Farquhar, Beaux Strata­gem, CEPLAY3B, p. 6).

194  

contrast /concession

The beginning of the eighteenth century then also sees still first used as an adverbial connector: (159) Adverbial connector still ‘nevertheless’: Rise, and know, I am a Woman without my Sex, I can love to all the Tenderness of Wishes, Sighs and Tears – But go no farther – Still to convince you that I’m more than Woman, I can speak my Frailty, confess my Weakness even for you … (Farquhar, Beaux Stratagem, CEPLAY3B, p. 58). 10.5.3.3.  However Another recurrent pattern employed by speakers and writers for the coinage of new adverbial connectors is the use of the full sentences as “dismissives” (Type (E)), i. e. generalizing sentences which presuppose that something has gone before, such as “however that may be”. Accordingly, forms such as howbeit and albeit (see Sorva 2007), which addi­ tionally invoke a concessive meaning by the use of the subjunctive be, are used as adverbial connectors in late Middle English and Early Modern English (see Appendix B.4.1 and MED, s. v. hou‑be‑it (a)). (160) The Front or the Forhead conteyneth nothing but the Skinne and Musculus fleshe, for the panicle vnderneth it is of Pericranium, and the bone is of the Coronal bone. Howebeit there it is made broade, as yf ther were a double bone, whiche maketh the forme of the Browes … (CESCIE1A, p. 34). (161) When I was before the Lordes at Lambeth, I was the first that was called in, all beit, Maister Doctour the Vicar of Croydon was come before me, and diuers other (CEPRIV1, Corp. Morelet to Margaret Roper, l. 1). The most prominent of these is, of course, however, which has come to be the central adverbial connector marking contrast in Present Day English (see Halliday and Hasan 1976: 255; Altenberg 1999: 250; Biber et al. 1999: 851), in particular in its non‑initial uses in a separate tone group (for details, see below, Chapter 13). For a better survey, the data of my corpus findings are repeated here (see Appendix B.4.1):

OE þeah – PDE though   195 however

Contrastive/Concessive

however, initial position EModE3 CEHAND3B, CEEDUC2B, CEPRIV3, CEPLAY3A LModE1 CLSERM1A, CLSERM1B, CLREY1, CLBUR1, CLPRIES1, CLPAR1, CLPRIV1A, CLPRIV1B, CLDIA1A LModE2 CLHAZ2, CLPRIV2B, CLOFFIC2A, CLOFFIC2C LModE3 CLEWEL3, CLCARP3 PDE LOB-D, FROWN-D, FLOB-F however, medial position EModE3 CEPRIV3 LModE1 CLSERM1A, CLSERM1B, CLHUM1, CLJOHPF1, CLSMI1, CLREY1, CLPRIES1, CLPAR1, CLDIA1A LModE2 CLGOD2, CLBLAI2, CLMAL2, CLBABPF2, CLBAB2, CLOWPF2, CLOW2, CLFOS2, CLHAZ2, CLELLI2, CLSERM2A, CLSERM2B, CLPRIV2A, CLOFFIC2A, CLOFFIC2C LModE3 CLBAGPH3, CLHUX3, CLARN3, CLPAT3, CLBOW3, CLCHES3, CLHUL3, CLCARP3, CLCHEY3, CLJAM3, CLBOETH3, CLBURY3, CLHOM3A PDE LOB-D, LOB-F, FROWN-D, FLOB-F however, final position EModE3 CEPLAY3A, CEPLAY3B LModE2 CLPRIV2A Collocations:

and however EModE3 but however EModE3 LModE1

CESCIE3B CEDIAR3A CLSERM1B

This entry shows that – in the corpus texts – however first occurred as an adverbial connector marking contrast/concession in the late Early Modern English period (EModE3 = 1640–1710),9 since when it has been attested in a wide variety of texts. It is often found in initial position (also in the collocations and and but however), but also and more frequently in medial position (see the underlined texts; underlining refers to more than 5 occurrences per 5,000 words). For this topological variation, see two examples from one text in LModE3, a sermon by Joseph Butler (CLSERM1A). Both of them also very 9 Here the OED has a slightly earlier date in EModE2 (1613) for use of however as an adverbial connector (OED, s. v. however 3.), so that this date has been used for the overall analyses.

196  

contrast /concession

nicely illustrate the contrastive function of adverbial however (162a, 163a) in contrast to the purely concessive, correlative constructions though … yet (162b and c, 163b and c): (162) However [162a], though [162b] all this be allowed, as it expressly is by the inspired writers, yet [162c] it is manifest that Christians at the time of the Revelation, and immediately after, could not but insist mostly upon considerations of this latter kind (CLSERM1A, Butler, Sermon 1). (163) I must, however [163a], remind you that though [163b] benevolence and self‑love are different, though [163c] the former tends most directly to public good, and the latter to private, yet [46d] they are so perfectly coincident that the greatest satisfactions to ourselves depend upon our having benevolence in a due degree; and that self‑love is one chief security of our right behaviour towards society (CLSERM1A, Butler, Sermon 1). Already in its earliest sub‑period, however is also attested sentence‑finally, but only in texts which record oral or colloquial language, notably two plays (Vanburgh’s The Relapse and Farquhar’s The Beaux Stratagem; CEPLAY3A and 3B) and a letter by Jane Austen from the second sub‑period of Late Modern English (CLPRIV2A). (164) (Lov.) (running to him) I hope I han’t kill’d the Fool however – Bear him up! Where’s your Wound? (CEPLAY3A, Vanbrugh, The Relapse, p. I, 39). (165) (Dor.) O, Madam, had I but a Sword to help the brave Man? (Boun.) There’s three or four hanging up in the Hall; but they won’t draw. I’ll go fetch one however (CEPLAY3B, Farquhar, The Beaux Stratagem, p. 63). (166) I danced twice with Warren last night, and once with Mr. Charles Watkins, and, to my inexpressible astonishment, I entirely escaped John Lyford. I was forced to fight hard for it, however. We had a very good supper, … (CLPRIV2A, Jane Austen, Letters to her Sister Cassandra and Others, January 9, 1796). This suggests that sentence‑final uses of adverbial connectors such as how‑ ever, which are still much more common in spoken than in written Present Day English, started in the oral mode. This diachronic observation and the corpus findings for Present Day English instigated the following analysis for

Sentence‑final connectors in Present Day English   197

the re‑emergence of though as an adverbial connector in Present Day English by an analysis of the data collected in the London Lund Corpus of Spoken English. 10.6.  Sentence‑final connectors in Present Day English 10.6.1.  Corpus findings Among the commonly accepted topological criteria used for distinguishing conjunctions from adverbial connectors is the position of the connector: While conjunctions are always placed clause‑initially, adverbial connectors are flexible in their positions in the sentence (see above, Chapter 2). In most Indo‑European languages, possible positions for adverbial connectors are the initial and the medial position (for the different medial positions, among them the specific post‑first position, see above, Chapter 5.4 and below, Chapter 13.2–13.4) with, of course, differences in the respective languages.10 In Present Day English, however, we also find final adverbial connectors. These prove to be very common in the spoken mode (see Table 10.5): Table 10.5.: Positions of adverbial connectors in conversation and academic prose (adapted from Biber et al. 1999: 891, Biber et al.’s Table 10.18)

CONV ACAD

% in initial position

% in medial position

% in final position

 

 

 

 = 5 %;  = less than 2.5 %

In the corpus of the Longman Grammar, final adverbial connectors are attested in both the spoken and the written medium. academic prose clearly favours initial and medial position: only ca. 10 per cent of all adverbial connectors are found sentence‑finally, while more than 40 per cent are attested in medial position (see below, Chapter 13 for a discussion of this high proportion of medial adverbial connectors and its history). 10 In German, for example, there are systematic differences between the positions of conjunctions and adverbial connectors on the one hand and modal particles on the other hand. The latter have to be placed in the middle field. See “Sie wollte telefonieren, aber sie hatte kein Kleingeld” ‘She wanted to make a phone call but she did not have any change’ (conjunction) vs. “Das ist aber eine Überraschung” ‘It is indeed a surprise’ (modal particle) (see Diewald 1999).

198  

contrast /concession

In the sub‑corpus conversation, on the other hand, final adverbial connectors are used 40 per cent of all instances and are thus almost as frequent as initial ones. A closer corpus analysis (Biber et al. 1999: 889) shows that this high proportion of adverbials in final position is mainly due to three frequently occurring forms – then, anyway and though –, which are commonly placed in final position. 10.6.2.  Information structure The high frequency of adverbial connectors at the end of a sentence in the spoken mode asks for an explanation, not only because this position is only rarely attested in other Indo‑European languages (for sentence‑final connector in Japanese and other non‑Indo‑European languages, see Itani 1992). While circumstance and – with some restrictions – stance adverbials are regularly placed at the end of sentences, English alone allows this final position as a regular one for adverbial connectors. In terms of information structure, the introduction of a connector at the beginning of a sentence guides the hearers and readers in their interpretation, by signalling that two sentences or chunks of discourse are linked and, secondly, by explicitly highlighting the specific semantic relation. This explicit marking at an early place of the second connect is of less importance for the “pure” additive and transitional connectors, which only emphasize a relation which would already have been there in the asyndeton. It is much more essential for the “impure” connectors marking ccc‑relations: these signify that an additional (causal) or even contradictory aspect (concession/contrast) has to be considered in the relation of the two propositions. Adverbial connectors which are placed at the beginning of a sentence function as explicit signposts, guiding the readers and hearers through the text, because they facilitate the rapid processing of a passage. For speakers and writers, they help to make sure that their intentions about the semantic relations of the two propositions are understood. The initial position can therefore be considered the unmarked position for adverbial connectors (see also Biber et al. 1999: 889). Final connectors expressing relations of cause, contrast and concession, on the other hand, force a re‑processing or even reinterpretation of the preceding assertions. 10.6.3.  Sentence‑final connectors in Present Day German A comparison with other Indo‑European languages shows that English alone regularly employs sentence‑final adverbial connectors. The Handbuch der deutschen Konnektoren, which bases its classification on topological criteria

Sentence‑final connectors in Present Day English   199

only, does not recognize this position as a standard one for Present Day German, but lists a number of connectors which may – in certain contexts – be placed in final position (Pasch et al. 2003: 505–509). In the internet version of the handbook, which is constantly being revised, the grammis‑team later introduced the new category “Nachsatzposition bei Pronominaladverbien” (cf. grammis; revision of 28/10/2004, 16:22). Yet, this final position (“Nachsatzposition”) of adverbial connectors in Present Day German is restricted to pronominal connectors, in particular to additive außerdem and davon abgesehen, contrastive/concessive trotzdem and causal darum, deshalb, and deswegen. The only non‑interactive example grammis cites is (167) Ich muss heute auf dem Markt einkaufen. Bei uns gibts nichts Ge­ scheites. Deswegen. ‘I have to go shopping at the market today. Here, there isn’t anything appropriate (to buy). Therefore’. In this example, however, deswegen is not a prototypical final connector, which should by definition signal the link of the second connect to the first one. It rather gives the impression of an afterthought. This is evident from the fact that deswegen semantically belongs to the first, and not to the second connect (cf. the “Bei uns gibts nichts Gescheites. Deswegen muss ich auf dem Markt einkaufen” or “Ich muss heute auf dem Markt einkaufen, weil es bei uns nichts Gscheites gibt” and not “Ich muss heute auf dem Markt einkaufen. *Deswegen gibt es bei uns nichts Gescheites”). All the other examples given on grammis are even less convincing, since in these the speakers do not employ sentence‑final connectors, but only use a connector in elliptical constructions (i. e. instead of the full sentence here given in brackets): (168)

A: Es wäre schade, wenn das Gebäude abgerissen würde. B: Es ist aber baufällig. A: Trotzdem! (= Trotzdem wäre es schade, wenn das Gebäude abgerissen würde). ‘A: It would be a pity if the building would be pulled down. B: But it is dilapidated. A: Nonetheless!’

(169)

A: Morgen muss das Auto zur Inspektion! B: Und wir müssen uns auch um eine Garage kümmern. A: Außerdem! (= Ja, das müssen wir außerdem noch tun!).

200  

contrast /concession

‘A: Tomorrow, the car will have to be taken in for a service. B: And we also will have to see about a garage. A: Moreover!’

All in all, Present Day German does not employ final adverbial connectors as a regular pattern. 10.6.4.  Sentence‑final connectors in the history of English In English, final adverbial connectors are also a comparatively recent phenomenon: the passages quoted above for however in (164) and (165) are the earliest occurrences I have found. As has been pointed out above, all of these early instances are attested in texts which record an oral mode, such as plays or private letters. The same restriction to these genres is found with sentence‑final too, which – like though – may not be placed sentence‑initially, but only medially and finally. The first instances of sentence‑final too are also recorded in speech‑based texts, such as court trials and plays. In view of these early attestations of sentence‑final connectors in speech‑based texts and the high proportion of sentence‑final connectors in conversation in Present Day English, the following sections will examine the uses of one of the three highly frequent sentence‑final connectors of Present Day English, though. Though is a good candidate for such an investigation, not only because of its high frequency, but, more importantly, because it expresses a contrastive/ concessive relation, i. e. an “impure” relation which requires a reinterpretation of the relation of the two connects. It is furthermore interesting because – as has been sketched above – though had not been used as an adverbial connector for many centuries, but had predominantly been employed as a subordinating conjunction. 10.7.  Present Day English sentence-final though 10.7.1.  Earlier research In spite of these rather obvious developments in the newer history of though, these phenomena have, surprisingly, not been examined in detail in the literature. Sentence‑final though is mentioned, but not discussed in Halliday and Hasan (1976: 251), Quirk et al. (1985: 644) and Altenberg (1986: 23–24). Interestingly, Quirk et al. and Altenberg give the same example:

Present Day English sentence-final though   201

(170) (a) (Al)though he is poor, he is satisfied with his situation. (b) He is poor. He is satisfied with his situation, though. The fact that the use of sentence‑final though is illustrated by the construction with a pre‑posed concessive clause suggests that these authors see the two constructions as functionally identical, i. e. as concessives proper marking “counter‑expectancy” (i. e. the implication is “if someone is poor, he or she is not satisfied with his or her situation”). This interpretation, however, runs counter to the differences in illocutionary weight of hypotactic vs. paratactic construction. As has been pointed out several times above, the equal illocutionary weight of both connects implies a contrastive rather than a purely concessive reading. This contrastive function of sentence‑final though is acknowledged in the Longman Grammar. In the discussion of the findings, however, the sentence‑final use of though is related to the complexities of the semantic relation concession: Though as a linking adverbial is found primarily in conversation, where its frequency far exceeds that of though as a subordinator … A possible factor at work is that considerable forward planning is required to construct a sentence with a concessive adverbial clause. The speaker has to have two propositions in mind, together with a realization that the one runs counter to the other, before starting to speak. This is not easy to manage in the online production of speech (Biber et al. 1999: 851).

Since this explanation focuses on speech processing in speakers, the Longman Grammar characterizes instances in the written mode as “surprising”: [t]hough as a linking adverbial is also sometimes used in writing that is informal or intended to resemble speech. Surprisingly [my emphasis, U. L.], though is also sometimes used as a linking adverbial in more formal written prose (851).

This research report first of all mirrors ongoing language change: Sentence‑final though was a comparatively rare phenomenon in the 1960s and 1970s when the Comprehensive Grammar was written (for the numbers, see below Figure 10.2), but became a highly frequent connector by the 1990s (see the counts of the Longman Grammar), since when it has also been increasingly found in written, even formal genres. The only publication focussing solely on sentence‑final though is an article by Barth‑Weingarten and Couper‑Kuhlen (2002), who observe a development of sentence‑final though from concessive marker to discourse marker (topic shifter) and examine the similarities and divergencies of this process to properties of grammaticalization. They do not discuss, however, if sentence‑final though functions as an adverbial connector proper. Characterizing sentence‑final though (and also, e. g. then) as an adverbial connector would imply that we

202  

contrast /concession

are dealing with a comparatively recent syntactic change in English, i. e. a new position for adverbial connectors, specific to Present Day English. The pragmatic/textual function of though as a topic shifter would then not be seen as its key function, but as only one of the functions of the contrastive/concessive sentence‑final adverbial connector though (since such a topic shifter function is inherently connected to the contrastive/concessive origins of an item and thus to every concession marker and every concessive construction).

10.7.2.  PDG obwohl The close relationship and interdependencies of concession and contrast with differences between parataxis and hypotaxis are also evident in the case of the German concessive subordinator obwohl. In spoken Present Day German, obwohl shows ongoing changes which are in some aspects similar to the changes which have affected the Present Day English concessive subordinator although. Obwohl is one of the central subordinators for marking the relation concession in Present Day German. In accordance with the properties of parataxis and hypotaxis in German, the main and subordinate clause are distinguished by differences in word order (verb‑second vs. verb‑final). The subordinate clause (verb‑final) introduced by clause‑initial obwohl may be placed before or after the main clause. (171) (a)  Obwohl sie hart gearbeitet hat, ist sie durchgefallen. (b)  Sie ist durchgefallen, obwohl sie hart gearbeitet hat.    ‘She failed, although she had worked hard’. Yet, in colloquial German, clause‑initial obwohl is increasingly used with verb‑second, i. e. main clause word order, as in (172) Sie hat hart gearbeitet. Obwohl – sie ist durchgefallen. The hyphen in this example signals that obwohl is marked prosodically by a pause separating it from its connect. Examples (173) and (174) illustrate this use of obwohl by corpus examples collected by Günthner (2000: 444, 446): (173) A:  ich dät SO GERN mit. echt A: obwohl ZEHN stunden flug – des fänd ich doch net so gut. ‘A: I would love to come along (really) A: although a ten hour flight – I would not like that’.

Present Day English sentence-final though   203

(174)

A: A: B: A: A: ‘A: A: B: A: A:

klingel einfach; ich komm dann runter okay obwohl. ich komm doch besser bei dir vorbei des isch glaub – ich gschickter. just ring the bell; I’ll come down okay although. I’d better come to your place that’s more convenient – I think’.

These examples illustrate that obwohl, when used as an adverbial connector with main clause word order, no longer marks a concessive relation. In these interactive examples, the equal illocutionary force of the connects does not indicate a concessive relation on the propositional level, but rather a contrast on the interpersonal or speech act level. It specifically signals correction, in both of these cases self‑correction. 10.7.3.  PDE but While PDG obwohl with main clause word order has recently appeared in this function in spoken, colloquial German, (self‑)correction on the interpersonal level is much more frequently expressed by the coordinator aber.11 This agrees with the findings for Present Day English but (cf. Barth‑Weingarten 2003; Altenberg 1986). In her analysis of public spoken discourse in the discourse‑functional perspective on concession, Barth‑Weingarten (2003) finds that in about 96 per cent of the cases with an explicit linker it is the coordinator but which serves as such a connector. These findings widely correspond to the proportions found in the only quantitative study which has as yet analysed contrastive/concessive linking in both spoken and written English in any detail (Altenberg 1986). Altenberg analyzes a sub‑corpus of 100,000 words each from the London Lund Corpus of Spo‑ ken English (LLC) and the Lancaster‑Oslo/Bergen Corpus of British English 11 In accordance with other studies in the field, I here follow the three‑level approach (for conjunctions, see, for example, Sweetser 1990), which distinguish the propositional level (i. e. the semantic content proper) from the interpersonal level (i. e. the social and expressive functions of communicative acts and moves) and the textual level (i. e. the methods of organization which create a coherent discourse).

204  

contrast /concession

(LOB), and offers the figures for the coordinator but, the adverbial connectors however, anyway, yet, still, nevertheless, though and the subordinators although, though, while, whereas and even if. Table 10.6.: but – adverbial connectors – subordinators (after Altenberg 1986)

This table first of all shows that the rank order (though not the proportion) is the same in both corpora. Speakers and writers are guided by much the same structural priorities when expressing contrastive relations: but is by far the most frequent connector for the relation of contrast/concession in both spoken and written English. With 843 in contrast to 85 occurrences, its number is about ten times higher than that of subordinators. With more than 56 per cent of all instances, but is, however, also the most frequent contrastive/concessive connector in the written corpus. In percentages, these findings yield the numbers in Table 10.7. Table 10.7.: but – adverbial connectors – subordinators: percentages (after Altenberg 1986) Coordinator but

Adverbial Connectors

Subordinators

Spoken English

78 %

14 %

8 %

Written English

56 %

22 %

22 %

But obviously functions as a potential all‑purpose link that may replace other (lexical and grammatical) alternatives in most contexts, whereas subordin­­ ation is – as we have seen above – only employed under certain circumstances, triggered by information processing considerations. The spoken material exhibits a strong tendency to prefer coordination over subordination (which only adds up to 8 per cent of all constructions). These proportions reflect tendencies

Present Day English sentence-final though   205

which facilitate information processing under conditions of real‑time planning in spoken discourse, among them the preference for right‑tending constructions (which has also been observed for causal/resultive clauses and complex noun phrases) and the presentation of two assertions in loosely linked sequence (Altenberg 1986: 20). Yet, Altenberg found that these general differences do not suffice to explain the much greater frequency of but in his spoken material. Much more import­ ant than these general tendencies is the fact that obviously speakers use but to express a number of contrastive discourse functions that are rare or absent in writing (Altenberg 1986: 27–32). None of these do express the semantic relation of concession or contrast on the propositional level, but may be characterized as “contrastive moves” on the interpersonal and textual level of discourse. The following table (Table 10.8) provides a summary of the specific sub‑types of “contrastive moves” relevant for the analysis of sentence‑final though. Table 10.8.:  Sub‑types of contrastive moves expressed by but Interactive yes‑but countering (interpersonal level) used turn‑initially to introduce an objection or deviant opinion But after disarmers (interpersonal level) face‑saving device indicating that the speaker is aware of possible objec- tions to his opinion: “I know … but, I’m terribly sorry … but” Topic resumption after disruption (interpersonal/textual level) not necessary in a written text Topic shifting but (textual level) weak countering force: does not signal objection to the previous speaker’s utterance but a deviation from the current discourse topic

Passage (175) illustrates instances of but which signal contrastive moves. The only instance of but which marks contrast on the propositional level is (175d). (175) B A A A A A B A

if they ^don`t get *t\aught#* *^y\es#* but the ^p\oint ‘is# that [i @] that if we “^ch\ange the _system# to al”^l\ow for this# that they`re *^st\arting ‘up* *but we ^still* ‘don`t pro’pose to !d\o it# – ^I ‘think we *d/\o#*

(175a) yes‑but countering

(175b) turn‑initial countering

206  

contrast /concession

A ^this was B ((*^y/\es#*)) B but “^how do we do our !f\inals ex_amining# (175c) yes‑but countering A ^w\ell# A I ^think the “_f\inalists will ‘have to st/op# A but the ^others will ‘carry !\on# (175d) contrast (propositional) B ^\I see# A but the ^th\ing ‘is# (175e) turn‑initial countering A for *^\instance# …. (LLC, Sample 1.4.714–740).

In many instances, interactive yes‑but counterings are found together with disarmers such as I know (176a, 176b), which serve as face‑saving devices indicating that the speaker is aware of possible objections to his opinion. (176) D well she was the only one of the family – there who/ D could do it . B yes I kn\ow# (176a) yes‑but countering B but ^also he [?]/ought to have s\old the ‘whole B ‘thing# B for his ^hundred ‘thousand *!p\ounds#* C *^yes !I* kn\ow# (176b) yes‑but countering C but ^he {s/aw} . th/ere you ‘see# C a ca^reer for D=an# C and a ca^reer for *((:M\argaret#))* / (LLC, Sample 1.13.834–84).

(177) then illustrates the use of but as a topic shifter. In an indirect speech act, speaker C suggests opening another bottle of wine, but speaker B changes the topic, enquiring from interlocutor A whether she was in the Archaeology or English Department. (177) C *we`ve* switched to ((a)) bottom of another ** C ** bottle now ((haven`t we)) B **(laughs – )** A ^[/\m]# – – B but ^you`re ‘archae!\ology# B I mean you`re “^s\eparate from /English# (LLC, Sample 1.9, 673–679).

In sum, the various functions of but in spoken interaction can be summarized as follows:

Present Day English sentence-final though   207

but propositional level • counter-expectancy • expressing a contrast etc.

interpersonal level • self-correction • other-correction • objection, different opinion, etc.

textual level • topic management topic change • discourse management summarizing, etc.

Figure 10.1.:  Contrastive functions of PDE but: three‑level approach

10.7.4.  Although and though in the LLC 10.7.4.1.  Quantitative findings For testing the various functions of sentence‑final though in Present Day English, I chose to examine Samples 1 to 9 from the London Lund Corpus of Spoken Eng‑ lish (LLC), i. e. the spontaneous recorded conversations comprising ca. 425,000 words. This corpus was selected because the corpus data for Present Day English and the accounts in research suggested that these conversations taped in the 1960s and 1970s would reflect the origins of the now highly frequent use of sentence‑final though. The analysis was restricted to interactive texts, because only these have the potential of attesting the diverse functions on the propositional, interpersonal and textual level. To allow a comparison with the subordinator (al)though, all instances of although and though have been examined.

Figure 10.2.:  Although and though in the LLC (Samples 1 to 9)

208  

contrast /concession

This diagram shows that the subordinator (al)though was still much more frequent than the adverbial connector though in these interactive texts of the 1960 and 1970s (in contrast to Present Day English, where adverbial though is found to be more frequent; see above, Chapter 10.6.1). Adverbial though is, however, already attested fifty times. Only seven of these occurrences are sentence‑medial. In accordance with the findings for Present Day English, the majority of adverbial though is found sentence‑finally, in altogether forty‑three occurrences. Subordinate clauses are – in accordance with the rules of grammar or style – mainly introduced by although when they are pre‑posed. More regularly, they are introduced by either although or though and are post‑posed (21 per cent pre‑posed; 59 per cent post‑posed). Although this pattern is – as has been argued above – more difficult to process, it is still more frequent in spoken interaction, because of the preference for right‑tending constructions, facilitating information processing under conditions of real‑time planning in spoken discourse. With respect to their functions, we find the following distribution: Table 10.8.:  Functions of although/though in the LLC Total

Propositional level concession

Interpersonal/textual level

pre‑posed although/though

39

39

–

post‑posed although/though

129

104

25

43

–

43

final though

10.7.4.2.  Functions of the subordinator (al)though This table shows that pre‑posed subordinate clauses introduced by (al)though are exclusively used to express the semantic relation of concession on the propositional level. The counter‑expectation in (178) is that the speaker cannot put the “CSC stuff” away. (178)

B  B  B  B 

I`ll be at ^h/\ome# and al”^though I`ll be doing CS/C _stuff# and ^that kind of th=ing# I can always ‚put it on one *s/ide#* (LLC, Sample 1.1, 116–118)

In (179), a correlative construction with adverbial nonetheless in the second connect, B expresses her embarrassment that the term ends without her being involved in any exams.

Present Day English sentence-final though   209

(179)

B  ^I`ve been “!{v\ery} emb\arrassed ((by)) the f/act#/ B  that ^though I`m ‘not . ex:\/amining#/ B  ^none the l=ess# . / B ^teaching ‘seems to :come to an /end# . (LLC, Sample 1.4, 629–633).

The semantic relation of concession in its propositional sense is even more frequently (104 times) expressed by post‑posed subordinate clauses. In (180), speaker B would expect her children to go and see her mother, but they do not do so. (180)

B  B  B  D 

‚children ‚never ‚go* to ‚see my m/other#/ al^though they`re !only on the :other side of/ B\/irmingham#+ ^y/eah# (LLC, Sample 1.12, 580–590).

In (181), speaker A had been doubtful of whether her offer would be accepted. (181)

A  A  A  C 

^=and [@m]# ^I !\asked her ‘if {she would ^h\ave him#}# ^though we`re now !s\everal ‘miles *aw/ay#* ^[/m]#* (LLC, Sample 1.8, 900–930).

In twenty‑five of the 129 instances, however, post‑posed subordinate clauses introduced by although/though do not work on the propositional, but on the interpersonal or textual level in exactly the sub‑types that have been found for the coordinator but. Example (182) is a prototypical example of self‑correction (stressed by in fact), when B acts as if she suddenly remembered that she had got her PhD the week before. (182)

B  B  C  C  B  B  B  B  B 

hel^l\o Mr B/aines# ^this is _Mrs !\Edgton# ^\oh# (good ^after!n\oon#)) [@:m] . ^((though)) in ‘fact I`m :Doctor ‘Edgton n/ow# so ^you ( – laughs) you can . ‘ob:serve my ‘new !st\atus# . (LLC, Sample 8.3b, 270–300). !got my :Ph:D last !w\eek#

Similarly, speaker B corrects her statement on the price of photo-copying by excluding a certain shop which is obviously still very expensive in (183).

210  

contrast /concession

(183) A  A  B  B  B  B 

xerography is not terribly expensive after/ all ^n/\o# ^no it /\isn`t# ^although Rank :X\erox are ^doing their !best to make :sure that it :\/is#/ (LLC, Sample 2.1, 938–942).

In the next example (184), the post‑posed subordinate clause has the potential of functioning as a topic shifter on the textual level. This is evident from the fact that speaker A again corrects herself, saying that she does not want to pursue this topic (i. e. does not want to gossip about the “dud ones”). (184) A  A  A  A  B  A  A  A  A 

^and there is “!H\erman# who is ^\also ‘known# ^in that :first b\atch [@:]# who ^came to us *:f\irst#* *^[=m]#* ^these are the !tw\o# al^though there were !three [@] !d\ud ones# [@:] that I ^don`t think at :this stage [@:m] I ‘want to [@:] to :sp\/eak a’bout# (LLC, Sample 2.6, 530–534).

In sum, these selected examples from the LLC illustrate that post‑posed subordinate clauses introduced by though or although are found in all the functions which have been established for the coordinator but. These adverbial clauses may not only work on the propositional level, but also have the potential to signal contrastive moves, such as self‑correction on the personal level, or may function as topic shifters on the interpersonal level. 10.7.4.3  Functions of sentence‑final though With respect to the functions of sentence‑final though, I have found only one example which might allow an interpretation on the propositional level. (185)

A  B  B  B 

I ^l\ike ‘chests# - ^y\eah# they`re ^\useless ‘bits of ‘furniture# ^r\/eally ‘though#

Present Day English sentence-final though   211

B  be^cause they :only ‘go _sort of of – !up to B  ‘slightly ‘below :t\able *h/eight#* (LLC, Sample 4.2, 300–340). While a purely concessive meaning may be possible (implication: “one should not like useless bits of furniture”), an interpretation as a contrastive move or self‑correction is much more likely, because the speaker stresses her corrective reflections by the epistemic really. All of the other instances of sentence‑final though only allow an interpersonal or textual reading. (186)

A  C  C  C  B  C  C 

*we`re going* to the Lake District for Christmas ^y\es# ^you !s\aid# ^that ‘should be n/ice# [m] – – ^I !love “D/\orset ‘though# it`s ^so b/\eautiful# – – (LLC, Sample 2.7, 110–170)

These analyses are supported by the fact that in nineteen of the forty‑three instances though is followed by a question tag with falling intonation (symbol ^ in the LLC). In negative question tags following a positive statement, speakers signal by falling intonation that they assume that the person spoken to will agree (Quirk et al. 1985: § 10.57). The whole phrase thus functions as a self‑correction and a floor‑holder. (187)

C  A  C  C  C  A  A  A  A 

*there* ^might have been m\ore ‘varied# ^y\es# – it`s ^quite n\ice {to ^l\ook at as [@:]#}# ^just a :series of :p\/ictures though# ^\isn`t it# ^oh v\ery# ^y\es# and I ^do th\ink# (LLC; Sample 1.8, 582–590). that ^these three p\ortraits#

(188)

C  C  C  C  C 

so I ^thought ^I`d ^I`d ‘rather p/ay my ‘eighty p\ence# than ^go through :th\at# ^it`s a !!lot of :m\oney ‘though# ^\isn`t it# -

212  

contrast /concession

B  20 [m] C  so I ^think (LLC, Sample 2.7, 914–920).

Instead of tag‑clauses, we also find collocations of sentence‑final though with you know. These may be interpreted as co‑texts complementary to disarmers such as I know which are frequent in collocations with contrastive, interpersonal but (see above, example (176)). (189)

B  and ^that really *((3 sylls))* A  *but if you`re* typing it up now [@:] **((why can`t/ A  yes** B  **it`s ^going so sl\owly** though you _know# B  it`s these ‘awful these ‘awful s/ymbols# . (LLC, Sample 2.1, 160–200).

In sum, we see that sentence‑final though is predominantly employed as a typical adverbial connector: it marks the relation of contrast, in particular on the interpersonal level in contrastive moves, such as self‑correction. In its origins, contrast here is clearly emerging from the idea of “internal concessivity”, i. e. counter‑expectancy on the level of the communication process “in spite of the state of the argument” or “as against what the current state of the communication process would lead us to expect”. Since this relation of contrast can also be marked by the coordinator but, or by post‑posed subordinate clauses introduced by (al)though, it is an inherent feature of contrastive/concessive connectors in general and is thus not specific to sentence‑final though. This means that it is certainly not correct to essentially characterize sentence‑final though as a discourse marker.12 The current re‑emergence of though is more likely to be seen as a re‑analysis of the subordinator though. However, since though has been attested as an adverb in certain contexts (see above, examples (149) to (152)), it is also problematic to regard its development as a counter‑example for unidirectionality (i. e. from the more grammatical conjunction to the less grammatical adverbial connector). The sentence‑final uses of though (and also then) should rather be seen parallel to developments of Present Day German obwohl (and weil) with verb‑second, i. e. main clause, word order. The establishments of a regular, new position for adverbial connectors in Present Day English may be seen as the unintentional result of a certain rhetorical strategy on the part of the speakers (see, for 12 Barth‑Weingarten and Couper‑Kuhlen (2002) acknowledge the function of sentence‑final though as a concessive marker, but do not further comment on it.

Present Day English sentence-final though   213

a similar account on the grammaticalization of adverb markers, Detges 1998: 7), who use a subordinator with – illocutionarily heavier – main clause word order. By doing so, speakers highlight the relation contrast in the specific co‑ and contexts of contrastive moves, mainly self‑correction. PDG obwohl is used as an adversative adverbial connector (instead of as a subordinator) and thus provides an alternative to the functionally rather overloaded PDG aber. A similar path is evident in PDE though. Since English, however, does no longer have any distinctions in the constituent order, which would differentiate – predominantly concessive – hypotaxis from contrastive parataxis (and thus the relations concession and contrast), it chooses a structural alternative and places the connector in a position which clearly differentiates hypotaxis from parataxis: in contrast to the conjunctions which are placed sentence‑initially, the adverbial connector is placed sentence‑finally. This starts in spoken, interactive texts on the interpersonal level, but is increasingly found in written texts, even in rather formal registers. The final placement of adverbial connectors (PDE contrastive however, though, causal/inferential then, additive too) is thus a change in constituent order, triggered by the lack of other distinctive means (such as verb‑second vs. verb‑final word order): in the past centuries and, in particular, the last decades, English has thus established a new slot for the placement of adverbial connectors, the sentence‑final position.

11.  addition 11.1.  General tendencies In most of the grammars and metalinguistic accounts on sentence relation past and present, adverbial connectors and conjunctions have been categorized in groups defined by the semantic relations they mark (see Locke’s criticism cited at the beginning of this study). Traditionally, the different semantic relations – addition, cause, concession/contrast, and transition (see above, Chapter 3.3) – are not divided further, but are described as if they were on equal levels. Only the authors of the Cambridge Grammar have recently proposed a distinction between “pure” and “impure” connectives (Huddleston and Pullum 2002: 775–779). While “pure” connectives have no other function than that of connecting their clause to the respective surrounding text, “impure” connectives, in addition to their linking function, also express a ccc‑relation, which adds another (cause/result) or even conflicting (concession) aspect to the relation of the two propositions connected. The long‑term developments in the field of adverbial connectors clearly illustrate these differences in rank, since they mirror the hierarchy of the increasing complexity of the semantic relations addition, cause and concession/ contrast. In the developments of adverbial connectors for the ccc‑relations, we have seen recurrent patterns triggered by changes in the typology of English, such as the loss of polyfunctionality in the differentiation of subordinators from adverbial connectors, or the replacement of pronominal by temporal or spatial deictics. Because of the differences in information structure separating paratactic from hypotactic structures, the semantic relation of pure addition is never expressed by subordination. This would contradict the very idea of the pure addition of an equally important, further proposition in the second connect. Issues such as problems of differentiating subordinators from adverbial connectors and the respective paths the Old English polyfunctional items take (cf. in particular OE causal/resultive forþæm and OE concessive þeah) do thus not play any role in the diachrony of additive adverbial connectors. 11.2.  The diachrony of reinforcing adverbial connectors 11.2.1.  General tendencies This relation of pure addition, inherently present as the unmarked case in the asyndeton, may be highlighted by the coordinator and, which has – at all

The diachrony of reinforcing adverbial connectors   215

times – been by far the most common marker of the additive relation, not only on the phrasal level, but also on the level of the sentence or discourse (see Chafe 1982; Culpeper and Kytö 2000 and the summary of corpus findings in Appendix B.1.5; for mode and genre‑differences, see below, 11.7). This additive force of and can be stressed and highlighted, i. e. “reinforced”, by its use in collocations such as and moreover or and furthermore (for the relevance of these collocations, see below, Chapter 13). These so‑called “reinforcing connectors” can, however, also be used on their own. Table 11.1 lists the development in the sub‑group of these reinforcing connectors (excerpted from Appendix A.1 and Appendices B.1.2–B.1.4). 11.2.2.  Also OE eall-swa, the only Old English additive connector comprising an explicitly deictic element (OE swa ‘so’), is also the only additive adverbial connector which has survived from Old English to Modern English (see DOE, s. v. eall‑swa). In all periods, however, it is often hard to decide whether eall‑swa/also expresses a reinforcing (‘furthermore’) or an equative (‘similarly’) additive relation.3 (190) Eallswa eac nygontyne gear gefyllað þone circul þe uðwitan hatað luna‑ rem (ByRM 2.3.116). ‘Furthermore/Similarly, nineteen years also (= eac) make up the cycle that scholars call lunar’. (191) And also, certes, if I governed me by thy conseil, it sholde seme that I hadde yeve to thee over me the maistrie, and God forbede that it so weere! … And also if I wolde werke by thy conseillyng, certes, my conseil … (CMCTPROS, p. 220, C.1). ‘And furthermore/similarly, in fact, if I behaved according to your advice, it would seem as if I had given you control over me. God forbid if it were so. … And furthermore/similarly, if I were to act according to your advice, in fact, my advice …’. The history of also in particular also reflects deliberate choices of certain authors or stylistic preferences of certain periods: also, which had been frequently used in Old English (OE eall‑swa) and Middle English, is only rarely attested from the Late Modern English period onwards (see Appendix B.1.2). 3 For the development of OE eall‑swa ‘wholly so, quite so, just so’ – which has yielded the adverb also and the conjunction as, see the OED, s. v. as.

1420–1500 (ME 4)

1350–1420 (ME 3)

1100–1350 (ME 1/2)

Old English [further] [furthermore]

moreover again above all besides too

[furthermore]

moreover again above all besides too

else also

[further]

else [also]

2 This additive, reinforcing use of after ‘moreover’ is exclusively found in CM­VICES4.

further further [further] [farther] ferthermore furthermore furthermore furthermore furthermore furthermore [fartherover], [furtherover] furtherover overmore overmore over that over that over that [over this] over this over this moreover moreover moreover moreover moreover item item item 2 after again again again [over and over and besides] besides above all above all beside besides besides over and above

[further]

1500–1570 (EModE 1) else also

1570–1640 (EModE 2)

else also

1640–1710 (EModE 3)

else also

1710–1780 (LModE 1)

eek elles also

1780–1850 (LModE 2)

eft eftsona/‑es eac swlyce [toeken] eek elles alswa [more]

besides over and above too

above all

again

moreover

[furthermore]

else also [more] further

1850–1920 (LModE 3)

þa giet eft [eftsona/-es] eac swylce þærtoeacan eac elles eall‑swa

too [plus]

besides

above all

again

moreover

furthermore

further

else also

1920– (PDE)

else also

Table 11.1.:  Reinforcing adverbial connectors in the history of English

216    addition

The diachrony of reinforcing adverbial connectors   217

In the full corpus of texts analysed for the present study, it is, for example, not attested at all as an additive adverbial connector in the Late Modern English period (LModE 1 to LModE3).3 The low esteem of also as an adverbial connector in Early and Late Modern English is also corroborated by the evidence of metalinguistic texts. In the six bilingual dictionaries conveniently accessible in the Early Modern English Dic‑ tionaries Database (EMEDD; ed. Ian Lancashire),4 furthermore, moreover and besides emerge as the central additive connectors of Early Modern English.5 Table 11.2.:  Reinforcing adverbial connectors in Early Modern English dictionaries Th. Thomas

Latin – English 1587

Etiam

Also, more over, yea; yea further, yea; why

Insuper

moreouer, ouer and beside, furthermore ….

Itemque

And moreouer, and also

Praeterea

Furthermore, more ouer, beside, else

Proporro

Further more, moreover

Quinetiam

Moreouer, be side that, yea that more is

Florio

Italian – English 1598

Alsi

also, moreover, eftsoone

Altresi

also, eke, moreouer, ouer and besides, likewise, …

Dauantaggio

moreouer, besides, more …

Et, a coniunction

and, moreouer, besides

Oltraciò, Oltradiciò

moreouer, besides that, furthermore

Minsheu

Spanish – English 1599

deMás

more, besides this, moreover, ouer and aboue, that which remaineth

3 Interestingly, the OED also asserts that ‘also was “[n]ot common in 16th c.”, referring to a frequency count of Shakespeare, who “has it only 22 times”’ (OED, s. v. also, Etymology). 4 John Palsgrave (1530; English-French), William Thomas (1550; Italian-English), Thomas Thomas (1587; Latin-English), John Florio (1598; Italian-English), John Minsheu (1599; Spanish-English), and Randle Cotgrave (1611; French-English). 5 It is still not very frequent as an adverbial connector in the Present Day English sub‑corpora analysed for the present study (two instances of also, seven of further‑ more, eight of again, and fifteen of sentence‑medial too).

218   

addition

Cotgrave

French – English 1611

D’abondant Arriere D’avantage

ouer and besides, moreouer also, moreouer, furthermore, besides, ouer and besides Moreouer, further, furthermore, else, besides, ouer and besides, aboue and besides Ouer and besides, moreouer Furthermore, moreouer, besides that Moreouer; furthermore; ouer‑and‑besides Then, afer [sic], moreouer, furthermore

De creuë Outreplus Dadvantage Puis

It is also interesting to note that in most of the dictionaries full pronominal be‑ sides this/that is preferred over the simple, lexicalized beside(s). The following passage from a letter of Cuthbert Tunstall to King Henry VIII illustrates the uses and also the effects of these longer, pronominal connectors besides that and the lexicalized over this. (192) And iff your Grace shuld accepte the said Election therby ye must confesse your realme to be under subjection off th’empire to the perpetual prejudice off your successor, or ells the said Election wer voyde as made off a person not eligible. Besids that the forme off the Election con­ tenyth that first he must be Kinge of Romains and the coronation at Rome makith hym have the name off the Emperor, wher befor he is callyd but Kinge off Romains. Over thys yff th’emperor which nou is remain stil Kyng off Romains as I understond he entendeth to doo, then yff your Grace wer eligible and undir th’empire, yet ye coud not be chosen Emperor, by cause ye were never Kinge of Romains. And also he remanyng ye could not be chosen Kynge off Romains, bycause the Kingdome is not voyde, and noon can be chosen therto but when it is voyd edyr by dethe or ellys when the Kinge off Romains is crownyd Emperor, wherby undir hym may be chosen a Kinge off Romains. (CEPRIV1, Cuthbert Tunstall to King Henry VIII, P I, 136–137). The chain of arguments begins with the simple coordinator and. The following arguments are then highlighted by pronominal besides that and over this. In the last instance, we do not find simple also, but the collocation and also. This illustrates that additive (and also transitional) connectors mainly serve rhetorical functions: they are employed to explicitly highlight a relation which is already intrinsically present in the linear sequence of the propositions.

The diachrony of reinforcing adverbial connectors   219

11.2.3.  New coinages: item, plus, too The large majority of all newcomers in the field of adverbial connectors are space deictics in origin. The only exceptions are the Latin loans item (ME3) and plus (PDE), which are, however, confined to certain genres or modes (item to documents and laws; plus is labelled “colloquial” by the OED; see above, 7.1). The most successful non‑spatial reinforcing connector is too, which is restricted to sentence‑medial or sentence‑final position. Sentence‑final and sentence‑medial too are first attested at the time when we also find the first instances of other sentence‑final connectors (LModE1; on however, see above, Chapter 10.5.3.3) and when there is generally a sharp increase of medial connectors (see Chapter 13). The emergence of too is thus not specific to the additive relations, but rather to be explained by these general topological changes affecting all adverbial connectors (for examples, see below, Chapter 13). 11.2.4.  Iconic principles Most of the new coinages in the field of additive connectors make use of the source domains time and, in particular, space, even more naturally than the adverbial connectors for the other semantic relations. In the case of additive connectors, the transfer from the temporal or spatial sequence to the linear textual and therefore argumentative sequence is straightforward and does not need any additional construal – in contrast to, for example, “post hoc, propter hoc” interpretations for cause or the conventionalisation of the conversational implicatures of simultaneity or concomitance for concession. Since most of the relevant patterns of transfer from the domains time and space have been described in Chapters 8.3 and 8.4 above, the following sections will only summarize the most important findings for the semantic relation addition. The most regular pattern of newcomers in the field of reinforcing adverbial connectors are polymorphemic items coined from the source domain space. By their polysyllabic structure and their resulting extraordinary length (in comparison to other etymologically Germanic grammatical elements in English), their reinforcing function is signalled iconically. It has been pointed out above that the additive þærtoeacan ‘thereto‑also’ is among the longest connectors in Old English. Similarly, the new coinages from Middle English onwards are typically compounds, so, for example, the double (often pleonastic) comparatives, such as furthermore, furtherover, overmore or moreover (for their morphology, see above 7.3.4). Even heavier phrases are attested from the

220   

addition

Early Modern English period: most frequent among them are over and besides (which fills two dimensions of the spatial extension) and the – again pleonastic – over and above. The pleonastic expressions in particular illustrate that these polymorphemic items are not coined for lack of other processes of word formation, but are formed and used intentionally in order to give a greater weight to the second connect, or to highlight that there is a relation of addition. It is important to distinguish these compound connectors and phrases from polymorphemic complex connectors, such as Old English contrastive/concessive naþylæs and swaþeahhwæðere. These complex concessive connectors are typical examples of the iconic principle that formal complexity corresponds to conceptual complexity (i. e. the complex relation concession signalling “counter‑expectancy”; see also Kortmann 1997: 113–136). The “pure” semantic relation addition is certainly not a complex concept and is accordingly not expressed by morphologically complex connectors which require a certain amount of processing. The compounds OE þærtoeacan and ME overmore, and also the phrase ­EModE over and besides are fairly straightforward as far as their morphological make‑up is concerned: their polymorphemic structure and length, however, serve to iconically underline their reinforcing function. This attempt to create iconic additive connectors may also be one of the reasons why also comes to be predominantly employed as a circumstance adverbial on the level of the phrase. It does not highlight the additive relation expressively enough. In (193a), for instance, also is supplemented by the full sentence it is further sayd. Without the punctuation, it could also be interpreted as working on the level of the phrase, modifying the verb say; ouer and besides (193b), on the other hand, is unambiguous and used on its own. (193) Also, it is further sayd, [193a] that a small part of your potentiall Cauteryes, doth and will worke as forcibly on a soft and tender bodye, as a great quantity thereof will doe vpon a stronge and grosse obdurate person. Ouer and besides, [193b] the greater abcessions are to bee Cauterized one way, and the lesser an other way, and that with good consideration (CESCIE2A, p. 30). 11.3.  Equative connectors The different, rhetorical character of additive connectors is even more obvious in the history of the so‑called “equative” adverbial connectors, which do not reinforce or highlight the proposition of the second connect, but place it on an

Equative connectors   221

equal level. There have always only been very few of these connectors in the English language and, more importantly, all of them are formed from still fairly transparent phrases (OE eall‑swa6, eac swlyce ‘also such’, EModE like(wise), LModE in the same way) or are derived from adjectives (semblably, corres­ pondingly, similarly), a pattern which is only rarely used for the formation of adverbial connectors of the other semantic relations (see above, Chapter 7.3.4). New coinages are EModE1 likewise, likeways, semblably EModE3 videlicet (on the very few occurrences and disputable status of the Latin loan videlicet, see above, Chapter 7.1) LModE3 in the same way PDE correspondingly, similarly

Their different status is also reflected in the fact that no such adverbial connectors were coined in the Middle English period, which had been the most productive period for all kinds of other connectives (see above, Chapter 6.6). 11.4.  Summative connectors In their predominantly lexical character, these equative connectors are similar to the sub‑type of summative ones, which mark the end of an argument or text. They commonly summarize a large chunk of preceding discourse and thus have to be very explicit. Appendix B.2 shows that these are even rarer and exhibit a still lower degree of lexicalization. The only candidates which show a certain amount of lexicalization are the calque in fine (on French à fin) and the phrase in sum, which requires a transfer of a mathematical concept to the line of argument in the text. Often, adverbials which are classified as summative are also ambiguous as to their status of adverbial connectors or style disjuncts, i. e. they may be seen not to link chunks of discourse, but to comment on the style of the text (‘in short’). In order to emphasize that a following passage is going to summarize a whole chunk of discourse, authors much more frequently employ full nominal or verbal phrases: (194) (194a) The sum is, men have various appetites, passions, and particular affections, quite distinct both from self‑love and from benevolence: all of these have a tendency to promote both public and private good, and may be considered as respecting others and ourselves equally and in common (CLSERM1A, Butler). 6 On the polysemy of eall‑swa, see above, Chapter 11.2.2, examples (190) and (191).

222   

addition

(194b) The sum is, the true reason why any man is an Atheist is because he is a wicked man (CESERM3A, Tillotson, p. II,: ii, 421). (194c) … and therefore to conclude this point, I will say no more, … (CEEDUC2B; Bacon, Sample 1, p. 23r–v).

11.5.  Appositive connectors The high amount of lexical transparency and the consequently low degree of lexicalization is also very evident in all the other sub‑groups of additive adverbial connectors. In the class of the so‑called “appositive” connectors, such as in other words, to wit, for instance or for example, all of them – apart from the archaic to wit ‘namely’ – show a very low degree of lexicalization and are fully re‑construable (the only property of lexicalization they fulfil is that they commonly do not admit a pre‑modifying adjective, but are uninterrupted in their sequence; i. e. ?for another instance). These items have not been included in the quantitative analyses of the present study precisely because they are not (yet) fully lexicalized and also because they mainly work on the phrasal level, modifying a sentence constituent rather than connecting sentences or chunks of discourse. They commonly do not signal a two‑place relation (see above, Chapter 2.2). (195) And these ˆthe Horseˆ standing in a true proportion, you shall fasten to the foure straps of leather; to wit one of them to his neere forelegs, and his nere hinder leg, and the other to his farre fore leg, and his far hinder leg; which is cald amongst ˆhorsemenˆ trauelling (CEHAND2B, p. 72). 11.6.  Enumerative (listing) connectors The most peripheral category of adverbial connectors are those of the enumerative/listing sub‑type of the additive relation, i. e. numerals such as secondly, thirdly, fourthly etc. They alone form an open class and new items can always be added to the inventory of this sub‑type. The only members of this class showing a certain amount of lexicalization are those denoting the first and the last in a list, i. e. first(ly), last(ly), finally and the phrases at (the) last. At the last, for example, is recorded as alast, i. e. in univerbated forms showing phonetic attrition in Middle English (see MED, s. v. a‑last). Furthermore, they may also show some degree of semantic bleaching, because first(ly) need not necessarily be followed by second(ly).

Enumerative (listing) connectors   223

These enumerative adverbial connectors illustrate that their use is triggered by rhetorical or stylistic considerations. In his sermons, Jonathan Swift uses lists of enumerative connectors, even on various levels in one and the same passage: level 1 (196a, 197a), level 2 (196b, 197b), level 3 (196c). (196) First (196a), I shall produce several instances to show the great neglect of preaching now among us. Secondly, I shall reckon up some of the usual quarrels men have against preaching. Thirdly, I shall get forth the great evil of this neglect and contempt of preaching, and discover the real causes whence it proceedeth. Lastly, I shall offer some remedies against this great and spreading evil. First (196b), I shall produce certain instances to show the great neglect of preaching now among us. These may be reduced under two heads. First (196c), men’s absence from the service of the church; and secondly, their misbehaviour when they are here. The first instance (196d) of men’s neglect is in their frequent absence from the church … (Swift, On Sleeping in Church; CLSERM1B). (197) First (197a), I shall produce certain points wherein the wisdom and virtue of all unrevealed philosophy in general fell short and was very imperfect. Secondly, I shall show, in several instances, where some of the most renowned philosophers have been grossly defective in their lessons of morality. Thirdly, I shall prove the perfection of Christian wisdom from the proper characters and marks of it. Lastly, I shall show that the great examples of wisdom and virtue among the heathen wise men were produced by personal merit, and not influenced by the doctrine of any sect; whereas, in Christianity, it is quite the contrary. First (197b), I shall produce certain points wherein the wisdom and virtue of all unrevealed philosophy in general fell short and was very imperfect …. (Swift, On the Wisdom of this World; CLSERM1B). Since this strategy is used in an almost identical way in both (196) and (197), it is certainly not only appropriate for sermons entitled “On Sleeping in Church” (196). It is, also for other authors of that time (see Appendix 1.1), rather a stylistic preference of this period and often further highlighted by phrases comprising a spatial element, such as place (see Appendix B.1.1): (198) (198a) I proceed, therefore, in the third place, to show the perfection of Christian wisdom from above (CLSERM1B; Swift, On the Wisdom of this World). (198b) Let me now, in the last place, offer some remedies against this great evil (CLSERM1B; Swift, On Sleeping in Church).

224   

addition

It is such a clear marker of argumentative texts and their argumentative structure that Swift – a great language critic – also employs it ironically in his A Modest Proposal: (199) For first, as I have already observed, it would greatly lessen the number of papists, with whom we are yearly overrun, being the principal breeders of the nation as well as our most dangerous enemies … Sec‑ ondly, The poorer tenants will have something valuable of their own, … Thirdly, Whereas the maintenance of an hundred thousand children … Fourthly, The constant breeders, … Fifthly, This food would likewise bring great custom to taverns …Sixthly, This would be a great inducement to marriage, … (Swift, A Modest Proposal; 1729: 21–27). (200) First, as things now stand, how they will be able to find food and raiment for an hundred thousand useless mouths and backs. And secondly, there being a round million of creatures in human figure throughout this kingdom, whose whole subsistence put into a common stock would leave them in debt two millions of pounds sterling, adding those who are beggars by profession to the bulk of farmers, cottagers, and laborers, with their wives and children who are beggars in effect (Swift, A Modest Proposal; 1729: 32). 11.7.  Genre‑dependency These examples show that the use of additive connectors in particular is highly genre‑specific. The corpus findings of the Longman Grammar corroborate this view: additive linking adverbials are much more frequent in academic prose than in all the other registers. This is related to the deliberate decision of the authors of this register to mark the links as explicitly as possible, thus guiding the reader through the texts by as many – in particular dispensable – signposts as possible (Biber et al. 1999: 880, Table 10.25). Corpus studies of Present Day English have also shown that the employment of additive connectors is very much medium‑dependent, since speech‑based registers generally prefer the simple coordinator and. In a small corpus analysis of Present Day English, Chafe (1982: 39) finds that and is employed about four times more often in the spoken than in the written medium (44.2 vs. 10.1 tokens per 1,000 words). These correlations of different modes and genres and the employment of the coordinator and are not as clear‑cut in the earlier periods of English, however. Culpeper and Kytö’s (2000) study on the frequency

Genre‑dependency   225

of and in four different registers, for instance, yielded a rather vague pattern (tokens per 1,000 words): there was no significant difference between the texts chosen to exemplify the written mode – science (17 tokens) and history (20.2 tokens) – and those exemplifying the spoken mode – drama (15.5 tokens) and trials (20.5 tokens). We can only surmise that the low frequency of and in the trial and drama texts is probably due to the fact that these are not oral, but only speech‑based texts, and that many of the ands were not jotted down when transferring the spoken version to the written medium. This example again shows that it is hard to carry out data‑driven, quantitative studies on issues like the present one. In spite of these problems, we may still gain some information from a more detailed analysis of the number of different adverbial connectors used in selected texts of the various sub‑periods, at least as far as the semantic relation addition is concerned. The graphs in Figure 11.1 are based on a count of all tokens of adverbial connectors in my corpus of “Treatises and Homilies” (see Appendix C.2.1). Figure 11.1 gives the absolute numbers of tokens for the respective semantic relations per 20,000 words.

Figure 11.1.:  Adverbial connectors per 20,000 words (corpus texts C.2.1)

The unsteady, even erratic graph for the causal connectors mainly reflects metho­dological problems concerning the status of for, which had to be counted as an adverbial connector marking the relation cause until sub‑period ME3, but was then no longer included because it – as has been shown above – came to be a rather loose connective, functionally equivalent to and (or even but) from the last Middle English period onwards (see above, Chapter 9.5.5). This, however,

226   

addition

means that neither the numbers for Middle English nor those for Early and Late Modern English are statistically robust, since for, of course, still expresses the relation of cause in some contexts. Trying to differentiate these uses would, however, have also implied a very high degree of subjective interpretation. All kinds of quantitative approaches which have been tested have unfortunately failed to yield solid, explicable results (and were thus not used in the respective chapters). This is different for the relations concession/contrast, and, in particular, for the semantic relation addition. The graph for concessives/contrastives in Figure 11.1 reflects the dominance of the contrastive coordinator but and the contrastive/concessive constructions comprising although (simple although, and correlative although … yet) by showing that there was not very much change in the absolute number of contrastive/concessive connectors until the Late Modern English period, when sentence‑medial however took over many of the functions which had until then been expressed by the coordinator but (see below, Chapter 13). The graph representing the number of the additive adverbial connectors in Figure 11.1 shows that there was a steady increase of additive adverbial connectors until the Early Modern English period, followed by a slight decrease in Late Modern English. Early Modern English thus emerges as the period when additive adverbial connectors were employed most frequently. Since corpus studies on Present Day English have shown that this kind of adverbial connectors is particularly frequent in academic prose, the increase in additive connectors in Early Modern English is most probably connected to the fact that the vernacular was increasingly used in academic registers, which had before been recorded in Latin and French. It also mirrors the tendency found for academic prose in Present Day English, namely that authors writing texts of this register, for which English was emancipated in the Early Modern English period, deliberately decide to mark the links as explicitly as possible. This again illustrates that the changes in adverbial connectors marking the relation addition are mainly triggered by rhetorical or stylistic considerations (and not by typological demands), at least from Early Modern English onwards.

12.  transition

12.1.  Preliminary considerations One of the most difficult decisions which had to be taken in the earlier phases of the present investigation was to decide whether to include the linguistic elements which will be discussed in the following sections in the quantitative parts of this study. In the end, my apprehensions have proved to be negligible, at least as far as the quantitative analyses are concerned. The items discussed here show different patterns of change than those found for the connectors marking the other semantic relations: most of them follow the “adverbial cline” as proposed by Traugott (Traugott 1995, Traugott and Dasher 2002; see above, Chapter 3.6). Since they are thus recurrently replaced onomasiologically (see below, 12.2.4), their inclusion or non‑inclusion does not affect the relative numbers of adverbial connectors as analysed in Chapter 6 above (as test analyses have shown). Their different patterns of long‑term development, however, demonstrate that connectors marking the relation transition are indeed peripheral. These items, such as OE soþlice, OE witodlice, OE þa, ME certes, ME for, or PDE of course and PDE in fact, are certainly not prototypical adverbial connectors, since they mainly work on the interpersonal and textual levels (for their specific characteristics and the reasons for their inclusion, see Chapter 3.4). If discussed at all, they have recurrently been labelled “discourse markers” in various research contexts (OE þa in Enkvist and Wårvik 1987 and Kim 1992; OE soþlice and witodlice in Lenker 2000, 2007b, EModE indeed and in fact in Traugott and Dasher 2002). In the present study, they are termed “transitional”, because they are best comparable to connectors such as now and here (for the category transition, see Quirk et al. 1985: 634). This classification also highlights their peripheral position in the field of adverbial connectors: In essence, these connectors are different from most of the other adverbial connectors because they do not express a core meaning such as the reinforcing additive relation or, in particular, any of the ccc‑relations cause, concession and contrast. With regard to the origins of transitional connectors, we can establish two unambiguous source domains (see Appendix B.5.1 and 2): Source domain time and place/space (see above, Chapter 8.3): OE þa, OE (hwæt) þa; here, now (all periods of English)

228  

transition

Source domain truth/fact (see above, Chapter 8.4): OE eornostlice ‘lit. earnestly’, soþlice ‘lit. truly’, witodlice ‘lit. ­certainly, truly’ ME certainly, certain, certes, forsooth, iwis, sekirly, truly, verily EModE indeed, sure, surely LModE in fact, in truth

As has been shown above (Chapters 3.4 and 8.4), these items share the essential properties of other adverbial connectors (for these properties, see above, Chapter 2.2), i. e. they signal – at least at certain stages of their developments – a two‑place relation in that they combine two connects which may be of sentential or even higher status (M3) and they furthermore do not contribute to the propositional meaning of either segment. By contrast, one or both of these defining conditions are not (unambiguously) fulfilled by two other groups of connectors which are listed in Appendix B.1.3 and B.1.4 but were not analyzed in the quantitative parts of this study, i. e. adverbs of the source domain uncertainty/doubt (12.2) and interrogatives (see 12.3). 12.2. Source domain uncertainty/doubt: peradventure, perchance, perhaps In the draft version of the third edition of the OED (s. v. peradventure, 5.) a connector function is suggested for ME peradventure (peraunter), EModE per‑ chance and perhaps. All of these forms are lexicalized prepositional phrases – comprising the Anglo‑Norman preposition par/per ‘for’ and a noun meaning ‘chance, good luck’1 – with an epistemic meaning “expressing a hypothetical, contingent or uncertain possibility”. In their core meaning, they do commonly not indicate a two‑place relation, but contribute to the propositional meaning of the connect: they are stance adverbials (content disjuncts) with a scope over the sentence only. (201) “All‑be‑itt þat þou arte man, gitt parauntur þou arte not cristened; and giff þou be cristeynd, gitt parauntur þou leueste not as Criste biddeste þe”. … “But sir,” þou seiste parauntur, “Criste calleþ þe frend and þer‑fore he will do well with þe and shewe þe of is mercye” (­CM­ROYAL, p. 16). 1 Peradventure and perchance are loans, whereas perhaps is a hybrid formation on the pattern of peradventure and perchance (cf. hap < ON happ ‘chance, hap, good luck’; see OED, s. v. perhaps).

Interrogatives   229



‘Even though you are human, yet perhaps you may not be baptized; and if you are baptized, you do perhaps not live as Christ wanted you to. …. ‘But, Sir’, you perhaps say, “Christ calls you friend and therefore he will do you well and show you of his mercy”’.

(202) And nowe, perchance, some enuious reder wyll hereof apprehende occasion to scorne me, sayenge that I haue well hyed me … (CEEDUC1A, p. 28). Yet there are also instances of semantic bleaching of peradventure, when it is, for example, used ‘in a statement of fact’. The OED here suggests a transitional reading ‘as it happened’ (OED, s. v. peradventure, 5): (203) Ensample as thus: The xiij day of March fyl upon a Saturday, peradven‑ ture … (CMASTRO, § 12, p. 673). ‘Example like that: The 13th day of March happened to be on a Saturday, perhaps …’. Although the exact date in March seems to rule out a modal meaning, even this instance of peradventure is not unambiguous. The passage is quoted from Chaucer’s handbook on the use of the astrolabe, and the dates are thus not real dates but example dates created for pedagogical purposes (the OED, interestingly, does not quote the introductory “Ensample as thus”).

12.3.  Interrogatives The other category which has been excluded are interrogatives such as OE hwæt (but not the phrase hwæt þa), EModE why and how. Most of these are predominantly used in speech‑based, interactive genres. They usually do not signal a two‑place relation either, but are similar in function to interjections and rather encode an entire, separate message (termed “pragmatic idioms” by Fraser 1999: 942–943). (204) (Lord Fop.) Why, that’s the Fatigue I speak of, Madam: For ’tis impossible to be quiet, without thinking: Now thinking is to me, the greatest Fatigue in the World. (Aman.) Does not your Lordship love reading then? (Lord Fop.) Oh, passionately, Madam – But I never think of what I read. (Ber.) Why, can your Lordship read without thinking? (CEPLAY3A, Vanbrugh, The Relapse, p. I, 36).

230  

transition

In Old English, these uses are further distinguished from proper adverbial uses in that they are – like all particles – not followed by inversion, but by subject – verb word order (for the relevance of this criterion, see in particular Fischer 2007: 280–297). For an example illustrating the interrogative (205a) and the interjection function of hwæt (205b) see (205) ða cwæð ic: Hwæt [205a] is þæt, la? ða cwæð he: Hwæt, [205b] þu wast ðæt ic þe ær sæde þæt sio soðe gesælð wære good (COBOETH, 34.86.15–16). ‘Then I said: What [205a] is that? Then said he, What, [205b] you know that I before said to you that the true happiness was good’. 12.4.  The diachrony of transitional connectors 12.4.1.  Paradigm shifts: truth – fact time and space have been shown to be the central source domains for all kinds of adverbial connectors. Accordingly, now and – much less frequently – here have been used as transitional connectors over all periods of English (on OE þa, see above, Chapter 5.3.1). The second major source category truth/certainty/fact is specifically used for transitional and – as a conversational implicature – concessive connectors (see above, Chapter 8.2). A more detailed look at the chronology shows that connectors from the source domain fact are only attested from the Early Modern English period onwards in the fully lexicalized and univerbated indeed and the still even more transparent in fact. At the same time, linguistic items formed from the source domain truth are no longer predominantly used as transitional connectors, but are mainly employed in contexts which suggest conversational implicatures of concession. The data collected in Appendix B.5.2 – for instance – show that verily and truly are only rarely attested from the Late Modern English period onwards. Sure(ly) and the full phrase it is true, which often have conversational implicatures of concession, gain ground, as do indeed and in fact for the trans­itional relation. It is, of course, tempting to relate these findings to the larger changes in socio‑cultural paradigms or, in the terminology of Foucault (1966), epistemes (for wider implications, see Mahler 1997): the substitution of connectors from the source domain truth by those of the source domain fact might be seen to reflect a socio‑cultural shift of the leading paradigm truth to the leading para­digm fact. In such a scenario, the paradigm of truth would be connected

The diachrony of transitional connectors   231

to theo‑centricity and dominating role of religion and theology in the Middle Ages; the leading paradigm of fact, on the other hand, could be said to reflect the dominating role of the natural sciences and the ideas of the Enlightenment. The diachrony of transitional connectors in English at least suggests such a replacement of the leading paradigms also in this field, an idea, however, which will have to be substantiated by material from other languages. Table 12.1.  Innovations of transitional connectors – source domain truth/fact 1920– (PDE)

1780–1850 (LModE 2) certes truly surely sure of course in fact

1850–1920 (LModE 3)

1710–1780 (LModE 1)

certes truly truly surely surely sure sure

1640–1710 (EModE3)

1570–1640 (EModE2)

1500–1570 (EModE1)

1420–1500 (ME 4)

1350–1420 (ME 3)

1100–1350 (ME 1/2)

Old English

eornostlice (ge)wislice witodlice witodlice soþlice soothly soothly iwis [forsoothe] forsoothe [certes] certes truly

soothly forsoothe certes certes truly truly truly surely surely sure

truly surely of of course course in fact in fact

12.4.2.  Regularities in semantic change If we compare Table 12.1 with the tables summarizing the diachrony of adverbial connectors of all the other semantic relations, we at first glance note a fundamental difference. In all the other relations, we find at least one central connector which has remained stable from the middle of the Middle English period onwards (after the periods M1 and M2 at around 1300; for the relevance of this date for the question “When did English begin?”, see Lutz 2002): there‑ fore/for for the relations of cause/result, yet for the relation of concession/ contrast and furthermore/first and at last etc. for the relation addition. In the field of transitional connectors, however, the scenario is different. We find a constant exchange and replacement of linguistic material by near‑synonyms. This finding corresponds to the ideas of “regularities in semantic change” as proposed in Traugott and Dasher (2002), in particular to the “adverbial cline” (see Traugott 1995) “clause‑internal adverbial > sentence adverbial > discourse particle”. In their closer analysis, Traugott and Dasher (2002) point out that

232  

transition

it is important to note that no lexeme is required to undergo the type of change schematized [here] … The hypothesis is that if a lexeme with the appropriate semantics undergoes change, it is probable that the language change will be of the type specified. More importantly, a reverse order of development is hypothesized to be ruled out except under special circumstances such as language engineering (Traugott and Dasher 2002: 281).

The data of the present study confirm this hypothesis for the adverbs which have been analysed in detail in Chapter 8.4 above (OE soþlice, ME soothly, ME forsoothe, ME verily). All of them follow the suggested cline: on the phrase level, they may be used as manner adverbs, mainly in direct speech with a first person subject (e. g. in phrases such as ic secge soþlice ‘I tell you truly’) or they are employed as emphasizers. As sentential adverbs (intermediate stage), they function as transitional connectors, often in collocations with the conjunctions and, but and for on the local level of discourse. On the global level of discourse then, they are employed to mark episode boundaries. All of these items thus followed the suggested “adverbial cline” once they had started it, and none of them shows a reverse movement. Most interesting in this respect is the highly frequent soþlice, which died out after trewely had acquired its epistemic and pragmatic functions. Another interesting case is for‑ soothe, which, after having been forced out by trewely, has only survived in a highly intersubjective and negatively connotated function, i. e. “parenthetically with an ironical or derisive statement” (OED, s. v. forsooth; see above, Chapter 8.4.5 and Lenker 2003: 283–286). This also supports Traugott and Dasher’s view that items – after they have reached the right end of the cline – are replaced onomasiologically (Traugott and Dasher 2002: 284): A third issue is what happens to those polysemies that have developed non‑truth‑conditional, procedural, scope‑over‑discourse, and intersubjective meanings. Our hypothesis is that Ls [= lexemes] with such meanings are replaced by newly recruited polysemies from other Ls, i. e. do not continue further semasiologically, but are replaced onomasiologically.

These instabilities and recurrent replacements in the field of transitional adverbial connectors coined from the source domain truth2 again show that the transitional category is different to the other categories, and that it is peripheral.

2 It is too early to see these developments in the adverbial connectors coined from the source domain fact.

13.  Perspicuity and the “New Rhetoric” Of all the parts of speech, the conjunctions are the most unfriendly to vivacity (Campbell [1776] 1963: 395)

13.1.  Collocations vs. medial position of adverbial connectors At the very beginning of this study, it was pointed out that the use of adverbial connectors – instead of asyndetic constructions or conjunctions – was not only a matter of grammar and information structure, but also of rhetoric and style. In the concluding chapter, I will now resume this topic and show that reflections on the use of particles, such as Locke’s (see above, Chapter 1), indeed had an impact, since the middle of the eighteenth century marks a shift in the use of adverbial connectors. These changes are related to the ideas of perspicuity and propriety as argued for in the spirit of Locke and the members of the Royal Society, and were further made possible by developments in the system of punctuation. This interaction of logical and rhetorical analysis was also affected by a change which took place in the role of rhetoric as expressed in the writings of the New Rhetoricians, in particular in George Campbell’s The Philosophy of Rhetoric ([1776] 1963). Let me introduce these issues by sample passages from texts of a comparable text type, i. e. treatises, by renowned, “mature” authors of their respective periods, whom we would normally not accuse of “not knowing how to write better prose”. (206) is taken from Tale of Melibee by Geoffrey Chaucer (ME3; 1343?–1400), and (207) is from Adam Smith’s The Wealth of Nations (LModE1; 1723–1790). (206) (a) “… but certes [206a] what ende that shal therof bifalle, it is nat light to knowe. For soothly [206b], whan that werre is ones bigonne, ther is ful many a child unborn of his mooder that shal sterve yong by cause of thilke werre, or elles lyve in sorwe and dye in wrecchednesse. And therefore [206c], er that any werre bigynne, men moste have greet conseil and greet deliberacion.” And whan [206d] this olde man wende to enforcen his tale by resons, wel ny alle atones bigonne they to rise for to breken his tale, and beden hym ful ofte his wordes for to abregge. For soothly [206e], he that precheth to hem that listen nat heeren his wordes, his sermon hem anoieth (CMCTPROS, Chaucer, Tale of Me‑ libee, p. 219.C2).

234   Perspicuity and the “New Rhetoric”

(b) And therfore [206f] ye shul venge yow after the ordre of right; that is to seyn, by the lawe and noght by excesse ne by outrage. And also [206g], if ye wol venge yow of the outrage of youre adversaries in ­oother manere than right comandeth, ye synnen. And therfore [206h] seith Senec that “a man shal nevere vengen shrewednesse by shrewednesse” (CMCTPROS, Chaucer, Tale of Melibee, p. 232, C.1).

(207) (a) This portion, however [207a], may still be considered as the natural rent of land, or the rent at which it is naturally meant that land should, for the most part, be let. The rent of land, it may be thought, is frequently no more than a reasonable profit or interest for the stock laid out by the landlord upon its improvement. This, no doubt [207b], may be partly the case upon some occasions; for [207c] it can scarce ever be more than partly the case. The landlord demands a rent even for unimproved land, and the supposed interest or profit upon the expense of improvement is generally an addition to this original rent. Those improvements, besides [207d], are not always made by the stock of the landlord, but sometimes by that of the tenant. When the lease comes to be renewed, however [207e], the landlord commonly demands the same augmentation of rent … (CLSMI1, The Wealth of Nations, Chapter XI, Part 1). (b) Hence [207f] a greater rent becomes due to the landlord. It requires, too [207g], a more attentive and skilful management. Hence [207h] a greater profit becomes due to the farmer. The crop, too [207i], at least in the hop and fruit garden, is more precarious. Its price, therefore [207j], besides compensating all occasional losses, must afford something like the profit of insurance (CLSMI1, The Wealth of Nations, Chapter XI, Part 1). 13.2.  Sentence‑initial collocations The differences between these two passages, which are prototypical of their respective period, are obvious at first glance: in the passages from Chaucer’s Tale of Melibee, all of the sentences start with explicit markers of textual cohesion. In all cases, these explicit markers are not simply conjunctions or adverbial connectors, but collocations of a conjunction (which functions as a rather loose connective) and an adverbial connector. The adverbial connector may express an additional semantic relation such as cause (and therefore), addition (and also) or transition (but certes, for soothly). Collocations like these are typical of Chaucer’s prose (Kerkhof 1982: 456, 459). Unfortunately, with the exception of the introductory passages of the Trea‑

Medial positions of adverbial connectors   235

tise on the Astrolabe (CMASTRO), all of Chaucer’s prose works are translations from Latin or French (CMBOETH: Boece, CMCTPROS: Tale of Melibee; Par‑ son’s Tale), so that we cannot rule out loan influence in these collocations. Yet, the predilection for collocations is also widely attested in prose texts which were originally composed in Middle English, such as in Caxton’s Prologues: (208) For in the sayd boke they may see what this transitorie & mutable worlde is And wherto euery mann liuyng in hit / ought to entende Thenne for as moche as this sayd boke so translated is rare & not spred ne knowen as it is digne and worthy For the erudicion and lernyng of suche as ben Ignoraunt & not knowyng of it … And furthermore I desire & require you … and therfore he ought eternelly to be remembrid. of whom the body and corps lieth buried … (CMCAXPRO, p. 63). It is also the preferred position of connectors in Early Modern English. In a quantitative study of Jonathan Swift’s prose style, Milic (1967) shows that Swift starts a third of his sentences with collocations of a conjunction (as a rather loose connective) and an adverbial connector. Milic also analyzes the rhetorical and stylistic functions of Swift’s use of connectors and finds that they lead to rhetorical persuasiveness rather than clarity. Swift’s prose may not be clear, but it induces the reader to “believe in the clarity and simplicity of what he has read” (Milic 1967: 254). 13.3.  Medial positions of adverbial connectors Not a single one of these collocations is found in the passages from Adam Smith’s The Wealth of Nations (1776). There are sentence‑initial connectors, such as causal for (207c) and two instances of resultive hence (207f, 207h). All of the other adverbial connectors are placed sentence‑medially, such as reinforcing besides (207d) and too (207g, 207i), contrastive however (207a, 207e), and resultive therefore (207j). Most of these are placed in post‑first‑position (207a, 207d, 207e, 207i, 207j), i. e. the position after the first obligatory constituent. Adverbials in post‑first‑position have – since Old English – been used to indicate that the element in first position is in the (contrastive) focus (see above, Chapter 5.4.2). Post‑first‑position – in accordance with its discourse functions – is mainly found with adverbial connectors signalling the semantic relations result/inference or contrast/concession. In Middle English and Early Modern English, post‑first‑position of adverbial connectors is only rarely attested (in only ca. thirty instances in my corpus for Middle English and ca.

236   Perspicuity and the “New Rhetoric” twenty-five instances for Early Modern English). Furthermore, it is predominantly restricted to resultive/inferential therefore and then:1 (209) For in þe loue of Ihesu þere schal be þin help. Loue is soche a migt þat it makiþ alle þing comoun. Loue þerfore Ihesu, … (CMCLOUD, p. 21). ‘For in the love of Jesus will be your help. Love is such a might that it makes all things equal. Love therefore Jesus, ….’. (210) Men that hunt so, be either ignorant persones, preuie stealers, or night walkers. Learning therefore, ye wise fathers, and good bringing vp, and not blinde & dangerous experience, is the next and readiest waie, that must leede your Children, first, to wisdom, and than to worthinesse (CEEDUC1B; Ascham, The Scholemaster, p. 215). (211) þis prayoure passeþ all oþur in worthynesse, for God hym-selfe made itt; … By þis preyour, þan, late ichon of vs þonke and preye God as Crist and Seynt Poule techeþ vs (CMROYAL, p. 9). ‘This prayer surpasses all other prayers in spiritual worth, because God himself made it, …. By this prayer, then, let each one of us thank and pray to God as Christ and Saint Paul teach us’. (212) Such a Schoole then as may be fit for the education of all sorts of child­ ren … should be situated in a City or Town of great concourse and trading, whose inhabitants are generally addicted, and sufficiently accommodated to entertain Tablers … (CEEDUC3B; Hoole, p. 221). In the passages quoted from the Wealth of Nations above in (207), by contrast, we also find additive, reinforcing connectors in post‑first‑position. Moreover, medial connectors are not restricted to this position after the first obligatory sentence constituent, but may also follow the verb (207g). Another common position attested from the Late Modern English period (increasingly from LModE2) onwards is the position after an operator 1 The only exception after the period ME1 are two instances of post‑first soothly in the main part of Chaucer’s Treatise of the Astrolabe (as compared to four instances of the collocation but soothly): “The ascendent sothly, as wel in alle nativites as in questions and eleccions of tymes, is a thing which that these astrologiens gretly observen. … The ascendent sothly, to take it at the largest, is thilke degre that ascend­ ith at eny of these forseide tymes upon the est orisounte. such as …” (CMASTRO, p. 670).

Corpus findings   237

(213) The natural sciences do not, however, stand on the same footing with these instrument‑knowledges … (CLARN3, 127). (214) The natural strength of such literature will, of course, be in the line of its tendencies; in transparency, variety, and directness. (CLPAT3, 11). or between a verb and a complement clause: (215) It is time, however, to pass on to those precursors of Babism who were … (CLCHEY3). (216) There is still some chance, however, that Sufism may be a record of its activity (CLCHEY3). (217) The case is so marked, however, that illustrations easily overflow, and there is no need of forcing doors that stand wide open (CLJAM3). Medial adverbial connectors may also be doubled, so that we find collocations, such as (218) In speaking of the future of the novel we must of course, therefore, be taken as limiting the inquiry to those types that have, for criticism, a present and a past (CLJAM3; Henry James, Future of the Novel). It has to be emphasized again that the post‑first‑position has to be distinguished from the other medial positions. Sentence‑medial (just as sentence‑initial and sentence‑final) adverbial connectors serve the function of combining their connects. Adverbial connectors in post‑first‑position do not only express this function, but – in addition – single out one individual constituent of one of the connects and put it in the focus.

13.4.  Corpus findings 13.4.1.  Present Day English The corpus findings for the position of adverbial connectors in Present Day English confirm this impression of a high number of sentence‑medial connectors in the written genres (see Biber et al. 1999: 889, Table 10.18;  = 5 %;  = less than 2.5 %):

238   Perspicuity and the “New Rhetoric” Table 13.1.: Positions of linking adverbials in conversation and academic prose (see Biber et al. 1999: 889, Table 10.18) % in initial position

% in medial position

% in final position

CONV







ACAD







 = 5 %;  = less than 2.5 %

In the corpus of the Longman Grammar, medial adverbial connectors are predominantly attested for the written genres; in less then 2.5 per cent of all instances are adverbial connectors placed sentence‑medially in spoken registers. academic prose clearly favours the initial (50 per cent) and medial positions (more than 40 per cent). The placement of adverbial connectors in the middle of a sentence is thus a shibboleth of written English. 13.4.2.  Old English to Late Modern English As has been illustrated by the examples above, medial positions – apart from post‑first‑position – have only been attested in higher frequencies since the Late Modern English period. This assertion was put to the test in a corpus study, using the smaller corpus of “Treatises and Homilies” (for the principles of compilation and the texts, see Appendix C.2). The analysis focuses on the relation of the coordinators and, but (OE ac) and for (all of which obligatorily have to be placed sentence‑initially) and adverbial connectors.

Figure 13.1.:  C  onjunctions vs. adverbial connectors (including OE þa) (all occurrences are given in normalized frequencies per 10,000 words).

Corpus findings   239

Figure 13.1. shows that ME 1/3 and Late Modern English 2/3 mark shifts in the use of connectors. In many respects, Middle English has in this study been found to be a period of transition and experiment in the use of adverbial connectors. The sharp decrease of adverbial connectors in Figure 13.1 from OE to ME1 is mainly due to the fact that OE þa, the shibboleth of Old English narrative style (see above, Chapter 5.3.1), is only rarely used from that time onwards. If we exclude the instances of OE þa (see Figure 13.2), we see a more even distribution of conjunctions and adverbial connectors over most of the periods of English.

Figure 13.2.: Conjunctions vs. adverbial connectors (without OE þa)

Yet in the period Late Modern English 3, i. e. the English of the nineteenth century (1850–1920), we see a sharp decrease in the number of the conjunctions. This is mainly due to the decreasing frequencies of and and, in particular, for, and leads to the differences between the spoken and the written mode in the frequency of and. Figures 13.1 and 13.2 thus implicitly corroborate the findings on and by Culpeper and Kytö (2000), i. e. that there was not much difference between the use of and in the spoken and written medium until the end of the Early Modern English period (see above, Chapter 11.7). The differences are, as the examples above illustrate, not only due to an increasing use of different adverbial connectors, but to deliberate choices of authors not to use sentence‑initial collocations, but medial adverbial connectors.

240   Perspicuity and the “New Rhetoric”

Figure 13.3.:  C  ollocations (conjunction + adverb) vs. adverbial connector, medial position (without OE þa)

Figure 13.3 shows that the impression we have gained from the examples above is indeed supported by the quantitative analysis of the corpus texts (C.2). We again see two periods emerge as decisive. First, the beginning of Middle English saw a rapid increase in the number of sentences which are introduced by a conjunction (and, but, or for) in a collocation together with an adverbial connector (see the examples from Chaucer in (206)). In Old English, the number of medial connectors had been higher (even if we exclude OE þa) because Old English regularly allowed post‑first‑position of adverbial connectors. ME3 (1350–1420) again appears as a “period of experiment and transition” and shows the highest numbers of such collocations. This process can be explained by the attempts at the evolution of an English prose style during the fourteenth and fifteenth centuries (see Mueller 1984), when English had not been used in the written medium – in particular in official registers and prose literature – for almost two centuries, but had been supplanted by Latin and French. The second decisive period appears in the language of the late eighteenth and early nineteenth century (LModE2 from 1780–1850), when we see a reversal of the distribution of sentence‑initial collocations and medial adverbial connectors. The difference, which leads to a specific Modern English prose style, is thus not that initial conjunctions or adverbial connectors are no longer used in Present Day English academic prose: with more than 50 per cent, the initial position can, even in academic prose, still be considered the unmarked position for adverbial connectors (see Biber et al. 1999: 891). More than 40 per cent of adverbial connectors, however, are now placed in medial position. This

Copia – perspicuitas   241

differentiates the spoken from the written language, since in spoken English less then 2.5 per cent of connectors are placed sentence‑medially. The positioning of adverbial connectors has become one of the most specific properties of the English prose style of the nineteenth and twentieth centuries.

13.5.  Copia – perspicuitas 13.5.1.  Perspicuity This process did not start by chance in the eighteenth century, but was primari­ ly inspired by the fundamental changes in the perception of language in the philosophy of the Enlightenment and, in particular, in the spirit of John Locke. In her chapter on “Literary Language” in the Cambridge History of the Eng‑ lish Language covering the years 1476 to 1776, Adamson shows that there is a “profound stylistic gulf” which separates Bacon from Locke (Adamson 1999: 541). Accordingly, Adamson (1999) divides her account into two overlapping phases, which are governed by two distinct stylistic key concepts: copia ‘copiousness’ vs. perspicuitas ‘perspicuity’. While the paradigm of copia dictated the stylistic norms of the Renaissance phase from 1500 to 1667, the concept of perspicuity determined the stylistic ideas of the neo‑classical phase from the 1640s onwards. The concept of perspicuity is primarily indebted to attitudes on language found in the philosophical investigations of Locke’s Essay Concerning Human Understanding. Locke emphasized that reason fundamentally consists in finding connections between ideas, and he saw knowledge as “the Perception of the Connection or Agreement, or Disagreement and Repugnancy of any of our Ideas” (Locke [1690] 1975: IV, i, 2). With respect to language, Locke argued for a common usage and propriety in language: Men’s Intentions in speaking are, or at least should be, to be understood; which cannot be without frequent Explanations, Demands, and other the like incommodious Interruptions, where Men do not follow the common use. Propriety of Speech is that which gives our Thoughts entrance to other Men’s Minds with the greatest Ease and Advantage; and therefore deserves some part of our Care and Study (Locke [1690] 1975: III, xi, 1)

This change in perspective has often been described as a rejection of rhetoric, but it should rather be described with Adamson as a “redirection of rhetoric” (Adamson 1999: 599). Like copia, perspicuity has its roots in classical anti­ quity, in particular in the writings of Quintilian (who saw perspicuity as a prac-

242   Perspicuity and the “New Rhetoric” tical necessity for the orator), but it was turned into a key concept of style in the late seventeenth and eighteenth century. Writers divert their energies away from copia towards alternative goals, such as the clarté that the French Academy prescribed as the first virtue of literature (see Thomas and Turner 1994). Similarly, the Royal Society rather harshly obliges their members to ban copia from their writings: They have therefore been most rigorous in putting into execution, the only Remedy, that can be found for this extravagance: and that has been, a constant Resolution, to reject all the amplifications, digressions, and swellings of Style: to return back to the primitive purity, the shortness, … They have exacted from all their members, a close, naked, natural way of speaking: positive expressions, clear senses; a native easiness; bringing all things as near the mathematical plainness, as they can … (Sprat, The History of the Royal Society, [1667] 1959: 113).

13.5.2.  Punctuation The ideas of perspicuity have influenced the use of adverbial connectors in two fundamental ways. In his survey on the history of punctuation, Parkes demonstrates that the attempts at a new kind of punctuation, representing the balance between grammatical, logical and rhetorical analyses of discourse, were to a large extent affected by attitudes towards language as promoted by Locke (Parkes 1992: 89–91). Increasingly, it was felt that a consensus should be observed in the application of the diverse punctuation marks, and thus Lowth’s grammar is the first one containing a chapter on punctuation (Lowth [1762] 1969: 154–172; see also Salmon 1988). When adverbial connectors are placed in the middle of a sentence, fixed rules in punctuation are essential for an unambiguous en‑ and decoding of the semantic relations between two connects. While the placement of adverbial connectors in post‑first‑position is, because of its contrastive focus, also feasible in the spoken medium and in periods without a fixed system of punctuation, other positions in the middle of a sentence require a clear, unambiguous system of a – rhetorical – punctuation. In the majority of the instances of medial adverbial connectors in Late Modern English and Present Day English, these are distinguished from circumstance adverbials by commas. These commas are commonly explained as mirroring pauses in the spoken medium. This, however, is an argument which is not at all substantiated by the actual language data: there are almost no medial adverbial connectors in spoken discourse. In view of the ideas of perspicuity and the attempts at mathematical plainness in the Royal Society (see the quote

Copia – perspicuitas   243

from Sprat ([1667] 1959), above), they rather resemble mathematical or logical symbols, which guide the reader in the interpretation of a formula or in the case of natural language, a text.

13.5.3.  The Scottish Rhetoricians: “The New Rhetoric” It is also not accidental that example (207), which illustrates the use of medial adverbial connectors, is taken from Adam Smith’s The Wealth of Nations, since Adam Smith is one of the main representatives of a group of scholars forming the “Scottish Enlightenment” (among them also David Hume and Hugh Blair). In this atmosphere, a new kind of rhetoric was developed which replaced the old classical art of oratory (see Gneuss 1996: 32–33). The leading exponents of these “Scottish Rhetoricians” were George Campbell (The Philosophy of Rhetoric, Campbell ([1776] 1963); over forty reprints in the nineteenth century) and Hugh Blair (Lectures on Rhetoric and Belles Lettres, Blair ([1783] 1965); at least 130 editions until 1911). The high number of editions and reprints demonstrates that these books were most influential for the formation of stylistic ideas in the eighteenth and nineteenth centuries. The books propose – following Locke’s lines of thoughts – a new kind of rhetoric which places much less emphasis on tropes or rhetorical figures, but deals with syntactical, text‑linguistic and pragmatic issues, such as sentence structure and paragraphs, which had been almost completely neg­ lected before (see above, Chapter 3.5): the fundamental qualities of this new style are “precision” and “perspicuity” (Blair [1783] 1965, Lecture X).

13.5.4.  George Campbell’s The Philosophy of Rhetoric (1776) Therefore it does not come as a surprise that George Campbell’s The Philoso‑ phy of Rhetoric is the first metalinguistic study which explicitly deals with conjunctions and adverbial connectors, and also with their respective functions: He explains why he – “in order to conduce more to perspicuity” (Campbell [1776] 1963: 385) – decided to devote a whole chapter each to connectives working on the level of the phrase and sentence (Campbell’s Chapter IV) and also to “Connectives employed in combining the Sentences in a Discourse” (Campbell’s Chapter V). In the connection of sentences of larger chunks of discourse, Campbell categorically rejects asyndetic constructions:

244   Perspicuity and the “New Rhetoric” It will scarcely be doubted by any person of discernment, that as there should always be a natural connexion in the sentiments of a discourse, there should generally be corresponding to this, an artificial connexion in the signs (Campbell [1776] 1963: 403).

After a general account of the diverse categories of conjunctions and adverbial connectors, Campbell adds altogether seven “remarks” dealing with the use of the conjunction and on its own or in collocations: (Remark 1) Being a monosyllable, it [and] will …., if not used too often, serve to smooth the current of discourse (Campbell [1776] 1963: 409). (Remark 2) One of the best expedients for preventing the connexives from becoming to conspicuous, is to avoid the frequent recurrence of the same particles, especially if they consist of more than one syllable. (Campbell [1776] 1963: 409). (Remark 3) Another useful expedient for answering the same end is to vary the situation of the conjunction, wherever the idiom of the tongue and the harmony of the sentence will permit the variation (Campbell [1776] 1963: 411). (Remark 4) Though certain circumstances require that one connexive be immediately followed by another, the accumulating of these without necessity ought always to be avoided (Campbell [1776] 1963: 411).

All of these “remarks” have, as my quantitative analyses have shown, proved to be highly influential in the evolution of a genuine English prose style.2 “Remark (2)” has, as has been shown above in Chapter 6.3, led to a steady increase in the number of types of connectors used in different texts. Campbell concedes that there may be cases when the repetition of a connector fulfils an expressive function, comparable to other rhetorical repetitions. Yet he fully rejects the repetition of connectors comprising more than one syllable: The now, and, for, but, nay, nor, have this advantage from their brevity, that though often repeated they pass unnoticed. But who, that hath any taste, can endure the incessant quick returns of the also’s, and the likewise’s, and the moreover’s and however’s, and the notwithstanding’s? An excess in these is insupportable. (Campbell [1776] 1963: 409)

With respect to the dramatic changes in constituent order demonstrated above, “Remarks (3) and (4)” are clearly the most important ones, since they recom2 It is interesting to note that Campbell places English over the classical languages in the summary of his Chapter “Modern Languages compared with Greek and Latin, particularly in regard to the Composition of Sentences”: “For this reason, I should not hesitate to pronounce that English is considerably richer than Latin, and in the main fitter for all the subtle disquisitions both of philosophy and of criticism” (Campbell [1776] 1963: 402–403).

Copia – perspicuitas   245

mend variation in the constituent order (Remark 3) and explicitly disapprove of sentence‑initial collocations of conjunction and adverbial connector (Remark 4). In agreement with iconic principles, Campbell regards the sentence‑initial position as the unmarked position for connectors. The place where we should naturally expect to find it, when it connects two sentences, is doubtless the beginning of the second (Campbell [1776] 1963: 410).

Yet Campbell rejects collocations of connective elements in all cases in which they are not necessary to encode two different relations. Collocations such as and if, for example, are not presented as problematic, because and serves as an additive connector, while the subordinator if introduces the subordinate conditional clause. And and if thus work on two different planes of discourse. In view of the fact that Campbell’s book was used as one of the most import­ ant style guides in the nineteenth century, it is furthermore crucial that he explicitly condemns collocations of three connectors, because they are “inconsistent with the principles of perspicuity, of vivacity, or of elegance” (Campbell [1776] 1963: 413). By contrast, Campbell recommends variation in the position of the adverbial connectors (conjunctions, of course, cannot be moved) because a position later in the sentence serves “as a cover to render it less observable”: In the beginning it stands by itself; whereas, placed in the manner now mentioned, it may be said to stand in a crowd (Campbell [1776] 1963: 413).

He then goes on to describe the possibilities of variation in the English of his time. Interestingly, Campbell feels that the number of syllables of the respective items is decisive: With us in particular, no monosyllabic conjunction, except the illative then, can be thus transposed (Campbell [1776] 1963: 410).

Polysyllabic adverbial connectors are, on the other hand, said to be much more versatile: Our language, however, hath been abundantly indulgent (where indulgence is of greater consequence) in the power it gives us in the disposal of those which consist of more than one syllable. Thus almost all the copulatives [reinforcing connectors, U. L.] which come under this denomination, the disjunctives however and neverthe‑ less and the illative therefore, may be shifted to the second, the third, the fourth place, and even further (Campbell [1776] 1963: 410).

This clearly descriptive account is very illuminating for a synchronic analysis of the use of English adverbial connectors at the end of the eighteenth century. Campbell himself, although very self‑confident in most of his evaluations, seems somewhat at a loss in trying to give the reasons for this distribution.

246   Perspicuity and the “New Rhetoric” It would be difficult to assign a satisfactory reason for the difference that hath been made in this respect, between conjunctions of one syllable and those of more. (Campbell [1776] 1963: 410).

This is a clear indicator that there was a situation of ongoing language change in the English of the late eighteenth century. The distinctions and restrictions noted by Campbell are clearly intermediate steps, all of which have eventually been given up. Campbell, for instance, specifically comments on the position of again in a footnote, saying that the reinforcing “again cannot conveniently be transposed, as it would scarcely fail to occasion an ambiguity, and be mistaken for the adverb signifying a second time” (Campbell [1776] 1963: 410, n. 5). This, however, had changed by the latest Late Modern English period (LModE3): (219) He, again, who thirsts alone for power squanders his wealth, despises pleasure, and thinks fame and rank alike worthless without power (CLBOETH3). (220) There, again, is the constitutional shrinking, through a kind of metaphysical prejudice, from the concrete – that fear of the actual – in this case, of the Church of history; to which the admissions, which form so large a part of these volumes, naturally lead (CLPAT3, 33). Likewise, the constraints on monosyllables are no longer valid from the nineteenth or twentieth century onwards. In the period LModE3 (1850–1920), the monosyllables here, first, thus, and still are first attested in adverbial connector function. In Present Day English, all adverbial connectors may be placed sentence‑medially in the written mode. The proliferation of this position in written English since the nineteenth century may thus be a consequence of Campbell’s “remarks”, since his Philosophy of Rhetoric has seen so many reprints.

14.  Conclusions A close reasoner and a good writer in general may be known by his pertinent use of connectives. Read that page of Johnson; you cannot alter one conjunction without spoiling the sense. It is a linked strain throughout. In your modern books, for the most part, the sentences in a page have the same connection with each other that marbles have in a bag; they touch without adhering. Samuel Taylor Coleridge, “Table Talk”

The present study set out to examine the changes in English adverbial connectors, in particular in contrast to other connectors, such as coordinating and subordinating conjunctions. Connectors are here defined as linguistic items which signal a two‑place relation between segments of text (here called “connects”) above the level of the phrase, i. e. between sentences or chunks of discourse. The meaning of a connector is procedural, not conceptual: it does not change the propositional content of any of the segments it relates. Connectors thus reveal or make explicit the connections already operating in a text. They are powerful clues about what commitment the speaker or writer makes regarding the relationship between sentences or chunks of discourse. Adverbial connectors in particular work as linguistic signposts, guiding the reader through a text. Accordingly, they have – in all periods of English – been particularly frequent in genres such as academic prose, which puts an emphasis on conveying logical coherence and whose main communicative purpose is information, argumentation and explanation for a specialist audience. Adverbial connectors are a very apt means for these ends because they allow writers to mark the development of their arguments overtly by relating one proposition to another and by explicitly showing contrasts, restatements and conclusions. Traditionally, linguists differentiate three kinds of connectors which are distinguished predominantly by topological criteria: coordinating conjunctions, subordinating conjunctions and adverbial connectors. The present study, however, emphasizes that these connectors are by no means functionally equivalent, and sees issues of discourse deixis, textual information and information processing as crucial for an understanding of the different or even diverging developments in the history of coordinators, subordinators and of adverbial connectors in English. In this view, adverbial connectors in particular are – in contrast to subordinators, which work on a more local level (the

248   Conclusions information presented in the subordinate clause is commonly not pursued in the following discourse) – very strong indicators of an equal illocutionary weight of both of the connects. Since adverbial connectors are employed to explicitly highlight or specify the relation already operating in the text (which may already be signalled less explicitly or specifically by a coordinator such as and, but or or), they are themselves commonly very explicit in marking the respective anaphoric (or cataphoric) relations, by, for example, so‑called pronominal connectors such as German des‑wegen or Old English forþæm (‘because of this’). Since adverbial connectors have as yet only rarely been discussed in the literature in a more systematic way, it was first of all necessary to compile inventories of all the adverbs which have been employed as adverbial connectors in the respective periods of English (see Appendix A.1 and the Index of Adverbial Connectors, Appendix A.2). These inventories then allowed a first comparison with the diachrony of English subordinators as sketched in Kortmann (1997). The origins of both of these connectors lie in Old English forms commonly called “ambiguous adverbs/conjunctions”, i. e. linguistic items which are syntactically and semantically polyfunctional. The very general line of development is the same for subordinators and adverbial connectors: in the course of a better form‑to‑function mapping after the Old English period, most of the “ambiguous adverbs/conjunctions” are discarded, so that we see a corresponding decrease in syntactic polyfunctionality and also in semantic polyfunctionality in both – now separate – classes. Adverbial connectors and subordinators are further similar in that the number of items (i. e. types) expressing these functions has increased from Old English to Present Day English. Yet we also observe striking differences in the developments of these two types of connectors which ask for an explanation. With respect to subordinators, Middle English emerges as the crucial period in which most of today’s subordinators were coined. The Middle English inventory of subordinators resembles that of Modern English most closely. As regards adverbial connectors, by contrast, there is no single period which could be said to have supplied the bulk of the adverbial connectors of today’s inventory. On closer inspection, we find that Middle English emerges as a “period of experiment and transition”: a large number of adverbial connectors were coined in the Middle English period, but most of them were ephemeral and have not survived into Modern English. An examination of their etymological origin and morphological make‑up furthermore shows that the specific characteristics in the history of adverbial connectors are not typical of the general expansion of the English lexicon in Middle and Early Modern English: although adverbial connectors are mainly

Conclusions   249

employed in written, specialized registers which commonly show a high percentage of Romance vocabulary, we, surprisingly, very rarely find French or Latin loans among the new coinages. Most of the common innovations in the field are univerbated or at least lexicalized prepositional phrases. These prepositional phrases became frequent at exactly the time when another pattern, the pronominal connectors typical of Old English, were no longer commonly used, let alone newly coined. Furthermore, we also see that these prepositional phrases basically only employ two cognitive source domains: time and space. In the present study, the term “shifting deictics” is used to summarize these processes, which mainly affect adverbial connectors expressing the “impure” relations cause and contrast/concession. After the collapse of both of the paradigms of the demonstratives at the end of the Middle English period, pronominal connectors could only be formed by using the new demonstratives that and this, which are, however, no longer inflected for gender and case. That, however, came to be the general subordinator in Middle English and served as a complementizer, relativizer, and, more importantly, was often also used with simple subordinators in order to highlight their subordinating function. This means that an adverbial connector such as for that is not only ambiguous, but misleading: readers would expect a pre‑posed subordinate clause introduced by for that and not an adverbial connector. Accordingly, we see that pronominal connectors – which were still coined in the (early) Middle English period  – are only rarely used from Early Modern English onwards (mainly in fully transparent prepositional phrases). The overtly deictic character, which is – for the reasons of information processing sketched above – almost a prerequisite for adverbial connectors, is now expressed by lexicalized phrases which signal the anaphoric relations by time or space deixis (such as PDE after all or hence). space in particular becomes more important from the later Middle English period onwards, when we see a spatialization of language in literacy. This shows that the history of adverbial connectors indeed mirrors overarching principles of long‑term linguistic change in English: the re‑structuring of the system of adverbial connectors is – at least partly – a consequence of typological changes (see the analogous developments in OE for þæm (þe) to for and of French par ce que to pas in French‑based creoles). Similar factors of typological change may also have been the cause for the emergence of sentence‑final adverbial connectors, such as PDE sentence‑final though, since this retrospective marker of linkage fulfils similar functions as PDG obwohl with main clause word order (instead of verb‑final), a differentiation no longer made in Present Day English.

250   Conclusions The history of adverbial connectors also attests to the increasing importance of lexical bundles or clusters of words (“Wortverband”) instead of loans from the Later Modern English period onwards, a phenomenon which can in this case – since it affects linguistic items which are mainly used in genres such as academic prose – hardly be accounted for as processes of “colloquialization”. As concerns cross‑linguistic patterns of language change, we see the universal patterns in the formation of concessives in the use of originally temporal phrases denoting simultaneity or concomitance (at the same time, meanwhile) or the use of negatives (nevertheless). The peripheral group of transitional connectors also attests clear instances of the “adverbial cline” – which has been established as one of the clines in a semantic‑pragmatic approach to grammaticalization – in the regularities of semantic change in OE soþlice, ME for‑ sooth or ME verily from adverbs used as circumstance adverbials (‘lit. truly’) to non‑propositional transitional connectors. Complex connectors such as naþylæs and swaþeahhwæðere ‘nevertheless’ are typical examples of the general iconic principle that formal complexity corresponds to conceptual complexity. More characteristic for adverbial connectors are (often pleonastic) compound adverbs expressing the relation addition, such as furthermore, overmore or moreover. These do not signal cognitive complexity, but are iconic in their length, in that they not only overtly signal, but expressively highlight the semantic relation which is already operating in the linear sequence of the text. This shows that the use of adverbial connectors is also affected by rhetorical preferences and reflections. These influences are most noteworthy in changes triggered by the idea of perspicuitas (which replaces copia as the leading style paradigm in the eighteenth century). Adverbial connectors, which make the speaker’s commitment as to the relations of discourse segments explicit, are a very apt means of fostering perspicuity in language. The spirit of perspicuity also led to a more systematic punctuation, which in turn allowed the medial positioning of adverbial connectors in the written mode. The commas, which commonly signal the status of an adverb as an adverbial connector (and not a circumstance adverbial), do, however, not primarily reflect pauses in the same construction in the spoken language. Since the medial positioning of connectors is rarely attested in spoken language, these commas should rather be regarded as rhetorical or logical (almost mathematical) signs signalling an author’s attempt at perspicuity. Since adverbial connectors are such an apt means for perspicuity, they have – from the end of the eighteenth century onwards – become, at least in genres such as academic prose, the principal connectors and are employed even more frequently than conjunctions.

References

1.  Corpora CLMET CME DOEC

F-LOB FROWN grammis HC LION LLC LOB Project Gutenberg

The Corpus of Late Modern English Texts [see De Smet 2005]

Corpus of Middle English Prose and Verse

The Complete Corpus of Old English in Machine Readable Form, ed. Antonette di Paolo Healey, Dictionary of Old English Project, Toronto: Centre for Medieval Studies, WWW-accessible version, University of Michigan Humanities Text Initiative, 2000. < http://ets.umdl.umich.edu/o/oec/> The Freiburg-LOB Corpus of British English

Literature Online. The London Lund Corpus of Spoken English

The Lancaster-Oslo/Bergen-Corpus of British English

2.  Sources Baker, Peter S. and Michael Lapidge (eds.) 1995 Byrthferth’s Enchiridion. (Early English Text Society. Supplementary Series 15.) Oxford: Oxford University Press. [= ByrM] Benson, Larry D. (ed.) 1987 The Riverside Chaucer. Oxford: Oxford University Press. [= Chaucer] Blair, Hugh [1783] 1965 Lectures on Rhetoric and Belles Lettres. Edited by Harold F. Hardings. (Landmarks in Rhetoric and Public Address.) Carbondale: Southern Illinois University Press. First published London: Printed for W. Strahan, T. Cadell and W. Creech in Edinburgh 1783.

252   References Campbell, George [1776] 1963 The Philosophy of Rhetoric. Edited by Lloyd F. Bitzer. Carbondale: Southern Illinois University Press. First published London: Printed for W. Strahan and T. Cadell and W. Creech at Edinburgh 1776. Colville, George [1556] 1897 Boethius: Boethius’ Consolation of Philosophy: Translated from the Latin. Edited by Ernest B. Bax. (The Tudor Library 5.) London: David Nutt. Crotch, Walter J. B. (ed.) 1928 The Prologues and Epilogues of William Caxton. (Early English Text Society. Original Series 176.) London: Oxford University Press. Dedeck-Héry, Venceslas L. 1952 Boethius’ De Consolatione by Jean de Meun. Medieval Studies 14: 165– 275. Forshall, Josiah and Frederic Madden (eds.) 1850 The Holy Bible: Translated from the Latin Vulgate by John Wycliff and his Followers. Oxford: Oxford University Press. Reprint: New York: AMS Press, 1982. Fox, Samuel [1864] 2004 The Consolation of Philosophy: King Alfred’s Anglo-Saxon Version of Boethius’ Work. London: H. G. Bohn. Reprint: Kessinger Publishings, 2004 (Book on Demand). Furnivall, Frederick J. (ed.) 1922 Hali Meidenhad: An Alliterative Homily of the Thirteenth Century. (Early English Text Society. Original Series 18.) Oxford: Oxford University Press. Goolden, Peter (ed.) 1958 The Old English Apollonius of Tyre. (Oxford English Monographs.) Oxford: Oxford University Press. Grasmück, Ernst Ludwig (ed.) 1997 Boethius: Trost der Philosophie. Zweisprachige Ausgabe. Aus dem Lateinischen von Ernst Neitzke. Mit einem Vorwort von Ernst Ludwig Grasmück. Frankfurt: Insel. Green, Richard H. (ed. and transl.) 1962 The Consolation of Philosophy: Boethius. Mineola, New York: Dover Publications. Herrtage, Sidney J. H. (ed.) 1881 Catholicon Anglicum: An English-Latin Wordbook, dated 1483. From the Ms. no. 1968 in the Library of Lord Monson, collated with the ad‑ ditional Ms.  15,562, British Museum: With Introduction and Notes. (Early English Text Society. Original Series 75.) London: Trübner. Holthausen, Ferdinand (ed.) 1888 Vices and Virtues: Being a Soul’s Confession of its Sins with Reason’s Description of the Virtues: A Middle-English Dialogue of About 1200 A. D., Part I. Text and Translation. (Early English Text Society. Original Series 89.) London: Oxford University Press.

Sources   253 Johnson, Samuel [1747] 1970 The Plan of a Dictionary. (English Linguistics 1500–1800 223.) Menston: Scolar Press. First published London: Printed for J. and P. Knapton, T. Longman and T. Shewell, C. Hitch, A. Millar, and R. Dodsley 1747. Johnson, Samuel [1755] 1967 A Dictionary of the English Language: In which the Words are Deduced from their Originals, and Illustrated in their Different Significations by Examples from the Best Writers: To which are Prefixed a History of the Language and an English Grammar. New York: AMS Press. First published London: W. Strahan 1755. Liuzza, Roy Michael (ed.) 1994 The Old English Version of the Gospels. Volume 1: Text and Introduc‑ tion. (Early English Text Society. Original Series 304.) Oxford: Oxford University Press. Locke, John [1690] 1975 An Essay Concerning Human Understanding. Edited and introduced by Peter H. Nidditch. (The Clarendon Edition of the Works of John Locke.) Oxford: Clarendon Press. First published London: Printed for Tho. Basset and sold by Edw. Mory 1690. Lowth, Robert [1762] 1969 A Short Introduction to English Grammar. (English Linguistics 1500– 1800 18.) Menston: Scolar Press. First published London: Printed by J. Hughs for A. Millar and R. and J. Dodsley 1762. Macaulay, George C. (ed.) 1900–1901 The English Works of John Gower. (Early English Text Society. Extra Series 81, 82.) London: Oxford University Press. [= Gower] Pemberton, Caroline (ed.) 1899 Queen Elizabeth’s Englishings of Boethius, De Consolatione Philoso‑ phiae A. D. 1593, Plutarch, De Curiositate, Horace, De Arte Poetica (Part), A. D.  1598. (Early English Text Society. Original Series 113.) London: Kegan Paul, Trench, Trübner. Preston, Richard Lord Viscount 1695 Boethius: Anicius Manlius Severinus Boetius: Of the Consolation of Philosophy: In Five Books: Made English and Illustrated with Notes. London: Awnsham and John Churchill. Scragg, Donald (ed.) 1992 The Vercelli Homilies and Related Texts. (Early English Text Society. Original Series 300.) Oxford: Oxford University Press. Sedgefield, Walter J. 1899 King Alfred’s Old English Version of Boethius’ De Consolatione Philosophiae. Oxford: Clarendon Press. Skeat, Walter H. (ed.) 1900 Ælfric’s Lives of Saints: Being a Set of Sermons on Saints’ Days For‑ merly Observed by the English Church: From Manuscript Julius E. VII in the Cottonian Collection, with Various Readings from Other Manu­

254   References scripts, Volume 2. (Early English Text Society. Original Series 114.) London: Oxford University Press. Swanton, Michael (ed.) 1975 Anglo-Saxon Prose. (Everyman’s Library 809.) London: Dent. Swift, Jonathan 1729 A Modest Proposal for Preventing the Children of Poor People from Being a Burthen to their Parents or the Country, and for Making them Beneficial to the Publick. Dublin: S. Harding (Project Gutenberg, E‑Text Nr. 1080) Thomson, David (ed.) 1984 An Edition of the Middle English Grammatical Texts. (Garland Medieval Texts 8.) New York: Garland. Watts, Victor E. (transl.) 1969 Boethius: The Consolation of Philosophy. (Penguin Classics.) London: Penguin Books. Zupitza, Julius (ed.) 2001 Aelfrics Grammatik und Glossar: Text und Varianten. Third reprint with a new introduction by Helmut Gneuss. Hildesheim: Olms. First published Berlin: Weidmann 1880.

3.  Dictionaries AND = Rothwell, William, Louise W. Stone and Thomas B. W. Reid (eds.) 1992 Anglo-Norman Dictionary. (Modern Humanities Research Association. Publications 8.) London: Modern Humanities Research Association. BT = Bosworth, Joseph and Thomas Northcote Toller (eds.) 1898 An Anglo-Saxon Dictionary: Based on the Manuscript Collections of the Late Joseph Bosworth. Edited and enlarged by Thomas Northcote Toller. Oxford: Clarendon. CH = Clark Hall, John R. 1984 A Concise Anglo-Saxon Dictionary. Fourth edition with a supplement by Herbert D. Meritt. Toronto: University of Toronto Press. DOE = Dictionary of Old English in Electronic Form A – F, Toronto: Centre for Medieval Studies [CD ROM publication 2003]. EMEDD = Early Modern English Dictionaries Database, ed. Ian Lancashire. Toronto. (accessed 2006; EMEDD has been replaced by Lexicons of Early Modern English (LEME) at ) Hoad, Terry F. (ed.) 1986 The Concise Oxford Dictionary of English Etymology. Oxford: Clarendon Press. Holthausen, Ferdinand 1934 Altenglisches etymologisches Wörterbuch. (Germanische Bibliothek 4,7.) Heidelberg: Winter.

Secondary sources   255 Kluge/Seebold = Kluge, Friedrich 2002 Etymologisches Wörterbuch der deutschen Sprache. Bearbeitet von Elmar Seebold, 24., durchgesehene und erweiterte Auflage. Berlin: Mouton de Gruyter. Mayhew, Anthony L. (ed.) 1908 The Promptorium Parvulorum: The First English-Latin Dictionary. Ed. from the Manuscript in the Chapter Library at Winchester with Introduction, Notes and Glossaries. (Early English Text Society. Extra Series 102.) London: Kegan Paul, Trench, Trübner. MED = Kurath, Hans, Sherman M. Kuhn, Robert E. Lewis et al. (eds.) 1954–2001 Middle English Dictionary. Ann Arbor, Michigan: University of Michigan Press. [online version: Middle English Compendium: ] OED = A New English Dictionary on Historical Principles. Edited by James A. H. Murray, H. Bradley, W. A. Craigie and C. T. Onions (Oxford 1888–1928). Corrected Re-Reissue with an Introduction, Supplement and Bibliography: The Oxford English Dictionary (1933). Second Edition, prepared by John A. Simpson and Edmund S. C. Weiner, 20 vols. (Oxford: Oxford University Press, 1989). Online version: . Oxford DNB = Oxford Dictionary of National Biography

4.  Secondary sources Adamson, Sylvia 1999 Literary language. In: Roger Lass (ed.), The Cambridge History of the English Language. Volume III: 1476–1776, 539–653. Cambridge: Cambridge University Press. Altenberg, Bengt 1984 Causal linking in spoken and written English. Studia Linguistica 38 (1): 20–69. Altenberg, Bengt 1986 Contrastive linking in spoken and written English. In: Gunnel Tottie and Ingegerd Backland (eds.), English in Speech and Writing. A Symposium, 13–40. (Studia Anglistica Upsaliensia 60.) Stockholm: Almqvist and Wiksell. Altenberg, Bengt 1999 Adverbial connectors in English and Swedish: semantic and lexical correspondences. In: Hilde Hasselgård and Signe Oksefjell (eds.), Out of Corpora: Studies in Honour of Stig Johansson, 249–268. (Language and Computers 26.) Amsterdam: Rodopi. Andrew, Samuel O. 1940 Syntax and Style in Old English. New York: Russell & Russell. Appenzeller-Gassmann, Verena 1961 Mittelenglische Bekräftigungsformeln. Zürich: Juris-Verlag.

256   References Auwera, Johan van der (ed.) 1998 Adverbial Constructions in the Languages of Europe. (Empirical Approaches to Language Typology 20,3. Empirical Approaches to Language Typology, EUROTYP 3.) Berlin/New York: Mouton de Gruyter. Bailey, Richard W. 1996 Nineteenth‑Century English. Ann Arbor, Michigan: University of Michigan Press. Barber, Charles Laurence 1997 Early Modern English. Edinburgh: Edinburgh University Press. Barth-Weingarten, Dagmar 2003 Concession in Spoken English: On the Realisation of a Discourse-Prag‑ matic Relation. (Language in Performance 28.) Tübingen: Narr. Barth-Weingarten, Dagmar and Elizabeth Couper-Kuhlen 2002 On the development of final though: A case of grammaticalization? In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Gram‑ maticalization, 345–361. (Typological Studies in Language 49.) Amsterdam: Benjamins. Baugh, Albert Croll and Thomas Cable 2003 A History of the English Language. (fifth ed.) London: Routledge. Beal, Joan C. 2004 English in Modern Times: 1700–1945. London: Arnold. Bellert, Irena 1977 On semantic and distributional properties of sentential adverbs. Linguis‑ tic Inquiry 8: 337–351. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad and Edward Finegan 1999 Longman Grammar of Spoken and Written English. Harlow: Longman. Blake, Norman F. 1992 The Cambridge History of the English Language. Volume 2: 1066– 1476. Cambridge: Cambridge University Press. Blakemore, Diane 2002 Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers. (Cambridge Studies in Linguistics 99.) Cambridge: Cambridge University Press. Blank, Andreas and Peter Koch (eds.) 1999 Historical Semantics and Cognition. (Cognitive Linguistics Research 13.) Berlin: Mouton de Gruyter. Blockley, Mary 2001 Aspects of Old English Poetic Syntax: Where Clauses Begin. (Illinois Medieval Studies.) Urbana: University of Illinois Press. Bolkestein, Alide Machtelt 2000 Discourse organization and anaphora in Latin. In: Susan C. Herring, ­Pieter van Reenen and Lene Schøsler (eds.), Textual Parameters in Old‑ er Languages, 107–137. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 195.) Amsterdam: Benjamins.

Secondary sources   257 Bowerman, Melissa 1986 First steps in acquiring conditionals: Some cognitive, semantic, formal and pragmatic considerations. In: Elizabeth Closs Traugott, Alice ter Meulen, Judy Snitzer Reilly and Charles A. Ferguson (eds.), On Condi‑ tionals, 258–307. Cambridge: Cambridge University Press. Breul, Carsten 1997 Grammatik und Bedeutung der kausalen Satzverbände: Because, as, since und for im schriftsprachlichen Englisch. (Linguistische Arbeiten 368.) Tübingen: Niemeyer. Brinton, Laurel J. 1996 Pragmatic Markers in English: Grammaticalization and Discourse Functions. (Topics in English Linguistics 19.) Berlin/New York: Mouton de Gruyter. Brinton, Laurel J. 2006 Pathways in the development of pragmatic markers in English. In: Ans van Kemenade and Bettelou Los (eds.), The Handbook of the History of English, 307–334. (Blackwell Handbooks in Linguistics.) Malden, Massachusetts: Blackwell. Brinton, Laurel J. and Elizabeth Closs Traugott 2005 Lexicalization and Language Change. (Research Surveys in Linguistics.) Cambridge: Cambridge University Press. Brugmann, Karl 1904 Die Demonstrativpronomina der indogermanischen Sprachen: Eine bedeutungsgeschichtliche Untersuchung. (Abhandlungen der Philologisch-Historischen Klasse der Königlich Sächsischen Gesellschaft der Wissenschaften 22, 6.) Leipzig: Teubner. Brunner, Karl 1960–1962 Die englische Sprache: Ihre geschichtliche Entwicklung. (Sammlung kurzer Grammatiken germanischer Dialekte B, Ergänzungsreihe 6.) Tübingen: Niemeyer. Brunner, Karl 1965 Altenglische Grammatik: Nach der Angelsächsischen Grammatik von Eduard Sievers. (Sammlung kurzer Grammatiken germanischer Dialekte A, Hauptreihe 3.) Tübingen: Niemeyer. Campbell, Alistair 1959 Old English Grammar. Oxford: Clarendon Press. Chafe, Wallace 1982 Integration and involvement in speaking, writing and oral literature. In: Deborah Tannen (ed.), Spoken and Written Language: Exploring Oral‑ ity and Literacy, 35–53. (Advances in Discourse Processes 9.) Norwood, New Jersey: Ablex Publishing Corporation. Claridge, Claudia and Terry Walker 2002 Causal clauses in written and speech‑related genres in Early Modern English. ICAME Journal: Computers in English Linguistics 25: 31–63. Clemoes, Peter 1985 Language in context: Her in the 890 Anglo‑Saxon Chronicle. Leeds Studies in English 16: 27–36.

258   References Consten, Manfred 2004 Anaphorisch oder deiktisch? Zu einem integrativen Modell domänen‑ gebundener Referenz. (Linguistische Arbeiten 484.) Tübingen: Niemeyer. Couper-Kuhlen, Elizabeth and Bernd Kortmann (eds.) 2000 Cause – Condition – Concession – Contrast: Cognitive and Discourse Perspectives. (Topics in English Linguistics 33.) Berlin: Mouton de Gruyter. Couper-Kuhlen, Elizabeth and Sandra A. Thompson 2000 Concessive patterns in conversation. In: Elizabeth Couper-Kuhlen and Bernd Kortmann (eds.), Cause – Condition – Concession – Contrast. Cognitive and Discourse Perspectives, 381–410. (Topics in English Linguistics 33.) Berlin: Mouton de Gruyter. Croll, Morris William 1966 Style, Rhetoric, and Rhythm. Essays by Morris W. Croll. Edited by John Max Patrick. Princeton, New Jersey: Princeton University Press. Culpeper, Jonathan and Merja Kytö 2000 The conjunction and in Early Modern English: Frequencies and uses in speech-related writing and other texts. In: Ricardo Bermúdez-Otero, David Denison, Richard M. Hogg and Chris B. McCully (eds.), Genera‑ tive Theory and Corpus Studies. A Dialogue from 10 ICEHL, 299–326. (Topics in English Linguistics 31.) Berlin: Mouton de Gruyter. Davis, Norman, Douglas Gray, Patricia Ingham and Anne Wallace-Hadrill 1979 A Chaucer Glossary. Oxford: Clarendon Press. De Smet, Hendrik 2005 A corpus of late Modern English texts. ICAME Journal. Computers in English Linguistics 29: 69–82. Denison, David 1993 English Historical Syntax. Verbal Constructions. (Longman Linguistics Library.) London: Longman. Detges, Ulrich 1998 Echt die Wahrheit sagen: Überlegungen zur Grammatikalisierung von Adverbmarkern. Philologie im Netz 4: 1–25. Detges, Ulrich and Richard Waltereit 2002 Grammaticalization vs. reanalysis: A semantic-pragmatic account of functional change in grammar. Zeitschrift für Sprachwissenschaft 21 (2): 151–195. Devriendt, Betty, Louis Goossens and Johan van der Auwera (eds.) 1996 Complex Structures. A Functionalist Perspective. (Functional Grammar Series 17.) Berlin: Mouton de Gruyter. Diessel, Holger 1996 Processing factors of pre‑ and postposed adverbial clauses. In: Jan Johnson, Matthew L. Juge and Jeri L. Moxley (eds.), Proceedings of the Twenty-Second Annual Meeting of the Berkeley Linguistics Society, 71–82. Berkeley, California: Berkeley Linguistics Society.

Secondary sources   259 Diessel, Holger 2004 The Acquisition of Complex Sentences. (Cambridge Studies in Linguistics 105.) Cambridge: Cambridge University Press. Diessel, Holger 2005 Competing motivations for the ordering of main and adverbial clause. Linguistics 43 (3): 449–470. Diewald, Gabriele 1997 Grammatikalisierung. Eine Einführung in Form und Werden gramma‑ tischer Formen. (Germanistische Arbeitshefte 36.) Tübingen: Niemeyer. Diewald, Gabriele 1999 Die Entwicklung der Modalpartikel aber: Ein typischer Grammatikalisierungsweg der Modalpartikeln. In: Hans Otto Spillmann and Ingo Warnke (eds.), Internationale Tendenzen der Syntaktik, Semantik und Pragmatik. Akten des 32. Linguistischen Kolloquiums in Kassel 1997, 83–91. (Linguistik International 1.) Frankfurt am Main: Lang. Diller, Hans-Jürgen and Manfred Görlach (eds.) 2001 Towards a History of English as a History of Genres. (Anglistische Forschungen 298.) Heidelberg: Winter. Donner, Morton 1991 Adverb form in Middle English. English Studies 72 (1): 1–11. Donoghue, Daniel and Bruce Mitchell 1992 Parataxis and hypotaxis: A review of some terms used for Old English syntax. Neuphilologische Mitteilungen 93 (2): 163–183. Eitle, Hermann 1914 Die Satzverknüpfung bei Chaucer. (Anglistische Forschungen 44.) Heidelberg: Winter. Enkvist, Nils Erik and Brita Wårvik 1987 Old English þa, temporal chains, and narrative structure. In: Anna Gia­ calone Ramat, Onofrio Carruba and Giuliano Bernini (eds.), Papers from the Seventh International Conference on Historical Linguistics, 220–237. (Amsterdam Studies in the Theory and History of Linguistic Science 4, 48.) Amsterdam: Benjamins. Fettig, Adolf 1934 Die Gradadverbien im Mittelenglischen. (Anglistische Forschungen 79.) Heidelberg: Winter. Fischer, Olga 1992 Syntax. In: Norman Blake (ed.), The Cambridge History of the English Language. Volume 2: 1066–1476, 207–408. Cambridge: Cambridge University Press. Fischer, Olga, Ans van Kemenade, Willem Koopman and Wim van der Wurff 2000 The Syntax of Early English. (Cambridge Syntax Guides.) Cambridge: Cambridge University Press. Fischer, Olga 2007 Morphosyntactic Change. Functional and Formal Perspectives. (Oxford Surveys in Syntax and Morphology 2.) Oxford : Oxford University Press.

260   References Foucault, Michel 1966 Les Mots et les Choses. Une Archéologie des Sciences Humaines. (Bibliothèque des Sciences Humaines.) Paris: Gallimard [English: The Or‑ der of Things. New York: Vintage, 1973]. Franz, Wilhelm 1939 Die Sprache Shakespeares in Vers und Prosa: Unter Berücksichtigung des Amerikanischen entwicklungsgeschichtlich dargestellt. Halle/Saale: Niemeyer. Fraser, Bruce 1999 What are discourse markers? Journal of Pragmatics 31 (7): 931–952. Fries, Udo 1994 Towards a description of text deixis in Old English. In: Klaus R. Grinda and Claus-Dieter Wetzel (eds.), Anglo‑Saxonica. Beiträge zur Vor‑ und Frühgeschichte der englischen Sprache und zur altenglischen Litera‑ tur. Festschrift für Hans Schabram zum 65. Geburtstag, 527–540. München: Wilhelm Fink. Fugier, Huguette 1987 Les connecteurs de „de cause“ en latin. Modèles Linguistiques 9: 9–17. Gil, Alberto 1995 Textadverbiale in den romanischen Sprachen: Eine integrale Studie zu Konnektoren und Modalisatoren im Spanischen, Französischen und Italienischen. (Bonner Romanistische Arbeiten 53.) Frankfurt: Lang. Givón, Talmy 2001 Syntax. An Introduction, Volume 2. Amsterdam: Benjamins. Gneuss, Helmut 1996 English Language Scholarship. A Survey and Bibliography from the Beginnings to the End of the Nineteenth Century. (Medieval and Renaissance Texts and Studies 125.) Binghamton, New York: Center for Medieval and Renaissance Texts and Studies. Godden, Malcolm 2000 Ælfric’s Catholic Homilies. Introduction, Commentary and Glossary. (Early English Text Society. Supplementary Series 18.) Oxford: Oxford University Press. Godden, Malcolm 2007 Did King Alfred write anything? Medium Ævum 76: 1–23. Görlach, Manfred 1986 Middle English – a creole? In: Dieter Kastovsky and Alexander Szwedek (eds.), Linguistics across Historical and Geographical Boundaries. In Honour of Jacek Fisiak on the Occasion of his Fiftieth Birthday. I: Linguistic Theory and Historical Linguistics, 329–344. (Trends in Linguistics. Studies and Monographs 32, 1.) Berlin: Mouton de Gruyter. Görlach, Manfred 1999 English in Nineteenth‑Century England. An Introduction. Cambridge: Cambridge University Press. Görlach, Manfred 2001 Eighteenth‑Century English. (Sprachwissenschaftliche Studienbücher.) Heidelberg: Winter.

Secondary sources   261 Görlach, Manfred 2004 Text Types and the History of English. (Trends in Linguistics. Studies and Monographs 139.) Berlin: Mouton de Gruyter. González-Cruz, Ana I. 2007 On the subjectification of adverbial clause connectives: Semantic and pragmatic considerations in the development of while‑clauses. In: Ursula Lenker and Anneli Meurman-Solin (eds.), Connectives in the History of English, 145–166. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 283.) Amsterdam: Benjamins. Greenbaum, Sidney 1969a Studies in English Adverbial Usage. London: Longman. Greenbaum, Sidney 1969b The question of but. Folia Linguistica 3: 245–254. Grice, Herbert Paul 1989 Studies in the Ways of Words. Cambridge, Massachusetts: Harvard University Press. Guimier, Claude 1985 On the origin of the suffix ‑ly. In: Jacek Fisiak (ed.), Historical Seman‑ tics. Historical Word‑Formation, 155–170. (Trends in Linguistics. Studies and Monographs 29.) Berlin: Mouton de Gruyter. Guimier, Claude 1991 Peut‑on définir l’adverbe? In: Claude Guimier and Pierre Larcher (eds.), Les États de l’Adverbe. Travaux Linguistiques du CERLICO, 11–34. (Cercle Linguistique du Centre et de l’Ouest. Travaux linguistiques du CERLICO 3.) Rennes: Presses Universitaires de Rennes. Günthner, Susanne 2000 From concessive connector to discourse marker: The use of obwohl in everyday German interaction. In: Elizabeth Couper-Kuhlen and Bernd Kortmann (eds.), Cause – Condition –Concession – Contrast. Cognitive and Discourse Perspectives, 439–468. (Topics in English Linguistics 33.) Berlin: Mouton de Gruyter. Haiman, John and Sandra A. Thompson (eds.) 1988 Clause Combining in Grammar and Discourse. (Typological Studies in Language 18.) Amsterdam: Benjamins. Halliday, Michael A. K. and Ruqaiya Hasan 1976 Cohesion in English. (English Language Series 9.) London: Longman. Halliday, Michael A. K. 1993 On the language of physical science. In: Michael A. K. Halliday and James R. Martin (eds.), Writing Science: Literacy and Discursive ­Power, 54–68. (Critical Perspectives on Literacy and Education.) London: Falmer. Heine, Bernd, Ulrike Claudi and Friederike Hünnemeyer 1991 Grammaticalization. A Conceptual Framework. Chicago: University of Chicago Press.

262   References Herman, Jószef 1963 La Formation du Système Roman des Conjonctions de Subordination. (Deutsche Akademie der Wissenschaften zu Berlin. Veröffentlichungen des Instituts für Romanische Sprachwissenschaft 18.) Berlin: Akademie-Verlag. Heuer, Hermann 1932 Studien zur syntaktischen und stilistischen Funktion des Adverbs bei Chaucer und im Rosenroman. (Anglistische Forschungen 75.) Heidelberg: Winter. Higashiizumi, Yuko 2006 From a Subordinate Clause to an Independent Clause. A History of English because‑clause and Japanese kara‑clause. (Hituzi Linguistics in English 2.) Tokyo: Hituzi Syobo Publishing. Hogg, Richard M. and David Denison (eds.) 2006 A History of the English Language. Cambridge: Cambridge University Press. Hogg, Richard M. 1992 The Cambridge History of the English Language. Volume I: The Begin‑ nings to 1066. Cambridge: Cambridge University Press. Hopper, Paul J. and Elizabeth Closs Traugott 2003 Grammaticalization. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Huddleston, Rodney and Geoffrey K. Pullum 2002 The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Hüllen, Werner 1989 “Their Manner of Discourse”. Nachdenken über Sprache im Umkreis der Royal Society. Tübingen: Narr. Itani, Reiko 1992 Japanese conjunction kedo (‘but’) in utterance‑final use: A relevance‑based analysis. English Linguistics 9: 265–283. Jones, Charles 1989 A History of English Phonology. (Longman Linguistics Library.) London: Longman. Jones, Richard Foster 1951 The Seventeenth Century. Essays by Richard Foster Jones and Others Writing in his Honour. Palo Alto: Stanford University Press. Jucker, Andreas 1991 Between hypotaxis and parataxis: Clauses of reason in Ancrene Wisse. In: Dieter Kastovsky (ed.), Historical English Syntax, 203–220. (Topics in English Linguistics 2.) Berlin: Mouton de Gruyter. Jucker, Andreas 1997 The discourse marker well in the history of English. English Language and Linguistics 1 (1): 91–110. Keller, Rudi 1993 Das epistemische weil: Bedeutungswandel einer Konjunktion. In: Hans Jürgen Heringer and Georg Stötzel (eds.), Sprachgeschichte und Sprach‑

Secondary sources   263 kritik. Festschrift für Peter von Polenz zum 65. Geburtstag, 219–247. Berlin: Mouton de Gruyter. Kemenade, Ans van and Bettelou Los (eds.) 2006a The Handbook of the History of English. (Blackwell Handbooks in Linguistics.) Malden, Massachusetts/Oxford: Blackwell. Kemenade, Ans van and Bettelou Los 2006b Discourse adverbs and clausal syntax in Old and Middle English. In: Ans van Kemenade and Bettelou Los (eds.), The Handbook of the His‑ tory of English, 224–248. (Blackwell Handbooks in Linguistics.) Malden, Massachusetts/Oxford: Blackwell. Kerkhof, Jelle 1982 Studies in the Language of Geoffrey Chaucer. (Leidse germanistische en anglistische reeks 5.) Leiden: Brill. Kim, Taejin 1992 The Particle þa in the West Saxon Gospels. A Discourse Level Analysis. (Europäische Hochschulschriften 14, 249.) Bern: Lang. Kisbye, Torben 1971–72 An Historical Outline of English Syntax, 2 Volumes. Aarhus: Akademisk Boghandel. Kivimaa, Kirsti 1966 þe and þat as Clause Connectives in Early Middle English with Es‑ pecial Consideration of the Emergence of Pleonastic þat. (Commentationes Humanarum Litterarum 39, 1.) Helsinki: Societas Scientiarum Fennica. Klare, Johannes 1958 Entstehung und Entwicklung der konzessiven Konjunktionen im Fran‑ zösischen. (Veröffentlichungen, Institut für Romanische Sprachwissenschaft, Berlin 13.) Berlin: Akademie‑Verlag. Koch, Peter and Wulf Oesterreicher 1985 Sprache der Nähe – Sprache der Distanz: Mündlichkeit und Schriftlichkeit im Spannungsfeld von Sprachtheorie und Sprachgeschichte. Roma‑ nistisches Jahrbuch 36: 15–41. Kohnen, Thomas 2004 Text – Textsorte – Sprachgeschichte: Englische Partizipial‑ und Gerun‑ dialkonstruktionen 1100 bis 1700. (Buchreihe der Anglia 37.) Tübingen: Niemeyer. Kohnen, Thomas 2007 Connective profiles in the history of English texts: Aspects of orality and literacy. In: Ursula Lenker and Anneli Meurman-Solin (eds.), Con‑ nectives in the History of English, 289–306. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 283.) Amsterdam: Benjamins. König, Ekkehard 1985a On the history of concessive connectives in English: Diachronic and synchronic evidence. Lingua 66: 1–19.

264   References König, Ekkehard 1985b Where do concessives come from? On the development of concessive connectives. In: Jacek Fisiak (ed.), Historical Semantics and Historical Word Formation, 263–282. (Trends in Linguistics. Studies and Monographs 29.) Berlin: Mouton de Gruyter. König, Ekkehard 1988 Concessive connectives and concessive sentences: Cross-linguistic regu­ larities and pragmatic principles. In: John A. Hawkins (ed.), Explaining Language Universals, 145–166. Oxford: Blackwell. König, Ekkehard 2009 Review of Lenker and Meurman‑Solin 2007. English Language and Linguistics 13:147–152. König, Ekkehard and Elizabeth Traugott 1982 Divergence and apparent convergence in the development of yet and still. In: Monica Macaulay and Orin D. Gensler et al. (eds.), Proceed‑ ings of the Eighth Annual Meeting of the Berkeley Linguistics Society, 170–179. Berkeley: Berkeley Linguistics Society. Kortmann, Bernd 1997 Adverbial Subordination. A Typology and History of Adverbial Sub‑ ordinators. Based on European Languages. (Empirical Approaches to Language Typology 18.) Berlin: Mouton de Gruyter [Material in: Kortmann, Bernd  1997b  A Cross‑Linguistic Dictionary of Adverbial Subordinators. Unterschleißheim: Lincom]. Krefeld, Thomas 1989 Unterordnung – Beiordnung – Zuordnung: Was ist romanisch an französisch que? In: Wolfgang Raible (ed.), Romanistik, Sprachtypologie und Universalienforschung: Beiträge zum Freiburger Romanistentag 1987, 11–36. (Tübinger Beiträge zur Linguistik 332.) Tübingen: Narr. Krefeld, Thomas 1999 Agens mit Leib und Seele: Zur Grammatikalisierung romanischer Adverbbildungen. In: Jürgen Lang and Ingrid Neumann-Holzschuh (eds.), Reanalyse und Grammatikalisierung in den romanischen Sprachen, 111–127. (Linguistische Arbeiten 410.) Tübingen: Niemeyer. Kroon, Caroline 1995 Discourse Particles in Latin. A Study of nam, enim, autem, vero and at. (Amsterdam Studies in Classical Philology 4.) Amsterdam: Gieben. Kryk-Kastovsky, Barbara 1997 From temporal adverbs to discourse particles: An instance of cross-linguistic grammaticalization? In: Terttu Nevalainen and Leena KahlasTarkka (eds.), To Explain the Present. Studies in the Changing English Language in Honour of Matti Rissanen, 319–328. (Uusfilologinen Yhdistys: Mémoires de la Société Néophilologique de Helsinki 52.) Helsinki: Société Néophilologique. Kühner, Raphael and Carl Stegmann 1912–1914 Ausführliche Grammatik der lateinischen Sprache. II: Satzlehre, Band 2. Hannover: Hahnsche Buchhandlung.

Secondary sources   265 Kytö, Merja 1996

Manual to the Diachronic Part of the Helsinki Corpus of English Texts. Coding Conventions and Lists of Source Texts. (International Computer Archive of Modern English. ICAME Collection of English Language Corpora 5.) Helsinki: Department of English, University of Helsinki. Kytö, Merja and Terry M. Walker 2006 Guide to a Corpus of English Dialogues 1560–1760. (Acta Universitatis Upsaliensis: Studia Anglistica Upsaliensia 130.) Uppsala: Acta Universitatis Upsaliensis. Lakoff, George and Mark Johnson 1980 Metaphors we Live by. Chicago: University of Chicago Press. Landau, Sidney I. 1984 Dictionaries. The Art and Craft of Lexicography. New York: Scribner. Lang, Ewald 2000 Adversative connectors on distinct levels of discourse: A re‑examination of Eve Sweetser’s three-level approach. In: Elizabeth Couper‑Kuhlen and Bernd Kortmann (eds.), Cause – Condition – Concession – Con‑ trast: Cognitive and Discourse Perspectives, 235–256. (Topics in English Linguistics 33.) Berlin: Mouton de Gruyter. Langacker, Ronald 1999 Losing control: Grammaticalization, subjectification and transparency. In: Andreas Blank and Peter Koch (eds.), Historical Semantics and Cog‑ nition, 147–175. (Cognitive Linguistics Research 13.) Berlin: Mouton de Gruyter. Lass, Roger (ed.) 1999 The Cambridge History of the English Language. Volume 3: 1476–1776. Cambridge: Cambridge University Press. Law, Vivien 1997 Grammar and Grammarians in the Early Middle Ages. (Longman Linguistics Library.) London: Longman. Lefèvre, Michel 2000 Subordination in Syntax, Semantik und Textlinguistik. (Eurogermanistik 15.) Tübingen: Stauffenburg. Lehmann, Christian [1982] 1995 Thoughts on Grammaticalization. Unterschleißheim: LINCOM Europa (originally published as: Thoughts on Grammaticalization. A Program‑ matic Sketch, vol. I. University of Cologne: Arbeiten des Kölner Universalienprojekts). Lehmann, Christian 1988 Towards a typology of clause linkage. In: John Haiman and Sandra A. Thompson (eds.), Clause Combining in Grammar and Discourse, 181– 225. (Typological Studies in Language 18.) Amsterdam: Benjamins. Lehmann, Christian 2002 New reflections on grammaticalization and lexicalization. In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammati‑

266   References calization, 1–18. (Typological Studies in Language 49.) Amsterdam: Benjamins. Leisi, Ernst and Christian Mair 1999 Das heutige Englisch. Wesenszüge und Probleme. (8th ed.) (Sprachwissenschaftliche Studienbücher.) Heidelberg: Winter. Lenk, Uta 1998 Marking Discourse Coherence: Functions of Discourse Markers in Spoken English. (Language in Performance 15.) Tübingen: Narr. Lenker, Ursula 1997 Die westsächsische Evangelienversion und die Perikopenordnungen im angelsächsischen England. (Texte und Untersuchungen zur Englischen Philologie 20.) München: Fink. Lenker, Ursula 2000 Soþlice and witodlice: Discourse markers in Old English. In: Olga Fischer, Anette Rosenbach and Dieter Stein (eds.), Pathways of Change. Grammaticalization in English, 229–249. (Studies in Language Companion Series 53.) Amsterdam: Benjamins. Lenker, Ursula 2002 Is it, stylewise or otherwise, wise to use ‑wise? Domain adverbs and the history of English ‑wise. In: Teresa Fanego, María José López-Couso and Javier Pérez-Guerra (eds.), English Historical Syntax and Morpho­ logy. Selected Papers from the Eleventh International Conference on English Historical Linguistics, Santiago de Compostela, 7–11 Septem‑ ber 2000, 157–180. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 223.) Amsterdam/Philadelphia: Benjamins. Lenker, Ursula 2003 Forsooth, a source: Metalinguistic thought in Early English. In: Lucia Kornexl and Ursula Lenker (eds.), Bookmarks from the Past. Studies in Early English Language and Literature in Honour of Helmut Gneuss, 262–288. (Texte und Untersuchungen zur Englischen Philologie 30.) Frankfurt: Lang. Lenker, Ursula 2004 Review of Traugott and Dasher 2002. Word 55: 262–265. Lenker, Ursula 2007a Forhwi ‘because’: Shifting deictics in the history of English causal connection. In: Ursula Lenker and Anneli Meurman-Solin (eds.), Connec‑ tives in the History of English, 193–227. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 283.) Amsterdam: Benjamins. Lenker, Ursula 2007b Soþlice, forsoothe, truly – Communicative principles and invited inferences in the history of truth‑intensifying adverbs in English. In: Susan Fitzmaurice and Irma Taavitsainen (eds.), Methods in Historical Prag‑ matics, 81–105. (Topics in English Linguistics 52.) Berlin: Mouton de Gruyter.

Secondary sources   267 Lenker, Ursula 2008 Booster prefixes in Old English – an alternative view of the roots of ME forsooth. English Language and Linguistics 12: 245–265. Lenker, Ursula and Anneli Meurman-Solin (eds.) 2007 Connectives in the History of English. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 283.) Amsterdam: Benjamins. Lenz, Friedrich 1997 Diskursdeixis im Englischen: Sprachtheoretische Überlegungen und ­lexiko-grammatische Analysen. (Linguistische Arbeiten 369.) Tübingen: Niemeyer. Levinson, Stephen 1983 Pragmatics. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Levinson, Stephen 2000 Presumptive Meanings. The Theory of Generalized Conversational Im‑ plicature. (Language, Speech, and Communication.) Cambridge, Massachusetts: MIT Press. Levinson, Stephen 2003 Space in Language and Cognition: Explorations in Cognitive Diversity. (Language, Culture and Cognition 5.) Cambridge: Cambridge University Press. Liggins, Elizabeth M. 1955 The expression of causal relationship in Old English prose. Unpublished dissertation, University of London. Lutz, Angelika 2002 When did English begin? In: Teresa Fanego, Belén Méndez-Naya and Elena Seoane (eds.), Sounds, Words, Texts and Change. Selected Pa‑ pers from 11 ICEHL, Santiago de Compostela, 7–11 September 2000, 145–171. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 224.) Amsterdam: Benjamins. Lutzky, Ursula 2006 Discourse markers? Well, … delimiting the basic features of discourse markers. Vienna English Working Papers 15: 3–24. Machan, Tim William 1985 Techniques of Translation. Chaucer’s Boece. Norman, Oklahoma: Pilgrim Books. Mahler, Andreas 1996 Die Materialität der Transparenz: Sprache, Politik und Literatur in der englischen Aufklärung. In: Klaus Garber and Heinz Wismann (eds.), Europäische Sozietätsbewegung und demokratische Tradition. Die europäischen Akademien der Frühen Neuzeit zwischen Frührenais‑ sance und Spätaufklärung, 721–754. Tübingen: Niemeyer. Mahler, Andreas 1997 Nominalism and Literary Discourse. New Perspectives, 251–268. (Critical studies 10.) Amsterdam: Rodopi.

268   References Marchand, Hans 1969 The Categories and Types of Present‑Day English Word‑Formation. A Synchronic‑Diachronic Approach. (Handbücher für das Studium der Anglistik.) München: Beck. Markus, Manfred 1996 Chaucer’s prose in the Canterbury Tales as parody. In: Jürgen Klein and Dirk Vanderbeke (eds.), Anglistentag 1995 Greifswald. Proceedings, 259–272. (Anglistentag: Proceedings of the Conference of the German Association of University Teachers of English 17.) Tübingen: Niemeyer. Markus, Manfred 2000 Wherefore therefore: Causal connectives in Middle English prose as opposed to Present‑Day English. In: Christian Mair and Marianne Hundt (eds.), Corpus Linguistics and Linguistic Theory. Papers from the Twentieth International Conference on English Language Research on Computerized Corpora (ICAME 20), Freiburg im Breisgau 1999, 215–232. (Language and Computers 33.) Amsterdam: Rodopi. Mauranen, Anna 1993 Cultural Differences in Academic Rhetoric: A Textlinguistic Study. (Nordeuropäische Beiträge aus den Human‑ und Gesellschaftswissenschaften 4.) Frankfurt: Lang. McIntosh, Angus 1991 Old English adjectives with derivative ‑lic partners: Some semantic problems. Neuphilologische Mitteilungen 92 (3): 297–310. McIntosh, Carey 1998 The Evolution of English Prose, 1700–1800. Style, Politeness and Print Culture. Cambridge: Cambridge University Press. Meurman-Solin, Anneli 2004 Towards a variationist typology of clausal connectives: Methodological considerations based on the corpus of Scottish correspondence. In: Marina Dossena and Roger Lass (eds.), Methods and Data in English Historical Dialectology, 171–197. (Linguistic Insights 16.) Bern: Lang. Mey, Jacob 2001 Pragmatics. An Introduction. Oxford: Blackwell. Michael, Ian 1970 English Grammatical Categories and the Tradition to 1800. Cambridge: Cambridge University Press. Michelsen, Martina 1993 Weg vom Wort – zum Gedankenstrich. Zur stilistischen Funktion eines Satzzeichens in der englischen Literatur des 17. und 18. Jahrhunderts. (Münchner Studien zur neueren englischen Literatur 6.) München: Fink. Milic, Louis Tonko 1967 A Quantitative Approach to the Style of Jonathan Swift. (Studies in English Literature 23.) The Hague: Mouton. Mitchell, Bruce 1985 Old English Syntax, 2 Volumes. Oxford: Clarendon.

Secondary sources   269 Mitchell, Bruce and Fred C. Robinson 2001 A Guide to Old English. (6th ed.) Oxford: Blackwell. Moessner, Lilo 2006 The subjunctive in Early Modern English adverbial clauses. In: Christian Mair, Reinhard Heuberger and Josef Wallmannsberger (eds.), Cor‑ pora and the History of English. Papers Dedicated to Manfred Markus on the Occasion of his Sixty‑Fifth Birthday, 249–263. (Anglistische Forschungen 363.) Heidelberg: Winter. Molencki, Rafal 1997a Albeit a conjunction, yet it is a clause: A counter‑example to unidirectionality hypothesis? Studia Anglica Posnaniensia. An International Review of English Studies 31: 163–178. Molencki, Rafal 1997b Concessive clauses in Chaucer’s prose. In: Jacek Fisiak (ed.), Studies in Middle English Linguistics, 351–371. (Trends in Linguistics. Studies and Monographs 103.) Berlin: Mouton de Gruyter. Morrow, Philipp R. 1989 Conjunct use in business news stories and academic journal articles: A comparative study. English for Specific Purposes 8: 239–254. Mossé, Fernand 1986 Mittelenglische Kurzgrammatik. Lautlehre, Formenlehre, Syntax. (Hueber Hochschulreihe 11.) München: Hueber. Mueller, Janel L. 1984 The Native Tongue and the Word. Developments in English Prose Style 1380–1580. Chicago: University of Chicago Press. Mugglestone, Lynda (ed.) 2006 The Oxford History of English. Oxford: Oxford University Press. Mustanoja, Tauno F. 1960 A Middle English Syntax, Part I. Helsinki: Société Néophilologique. Nate, Richard 2001 Wissenschaft und Literatur im England der frühen Neuzeit. (Figuren 9.) München: Fink. Nevalainen, Terttu 1991 BUT, ONLY, JUST: Focusing Adverbial Change in Modern English 1500–1900. (Uusfilologinen Yhdistys: Mémoires de la Société Néophilologique de Helsinki 5.) Helsinki: Société Néophilologique. Nevalainen, Terttu 1997 The process of adverb derivation in late Middle and Early Modern English. In: Matti Rissanen, Merja Kytö and Kirsi Heikkonen (eds.), Gram‑ maticalization at Work: Studies of Long‑Term Developments in English, 145–189. (Topics in English Linguistics 24.) Berlin: Mouton de Gruyter. Ong, Walter 1982 Orality and Literacy. The Technologizing of the World. (New Accents.) London: Methuen. Österman, Aune 1997 There compounds in the history of English. In: Matti Rissanen, Merja Kytö and Kirsi Heikkonen (eds.), Grammaticalization at Work. Stud‑

270   References ies of Long‑Term Developments in English, 191–276. (Topics in English Linguistics 24.) Berlin: Mouton de Gruyter. Parkes, Malcolm B. 1992 Pause and Effect. An Introduction to the History of Punctuation in the West. Aldershot: Scolar Press. Pasch, Renate, Ursula Brauße, Eva Breindl and Ulrich Hermann Waßner 2003 Handbuch der deutschen Konnektoren. Linguistische Grundlagen der Beschreibung und syntaktische Merkmale der deutschen Satzverknüp‑ fer (Konjunktionen, Satzadverbien und Partikeln). (Schriften des Instituts für Deutsche Sprache 9.) Berlin: Mouton de Gruyter. Pérez Quintero, María Jesús 2002 Adverbial Subordination in English. A Functional Approach. (Language and Computers 41.) Amsterdam: Rodopi. Peters, Pam 2006 The Cambridge Guide to English Usage. Cambridge: Cambridge University Press. Pinkster, Harm 1972 On Latin Adverbs. (North‑Holland Linguistic Series 6.) Amsterdam: North‑Holland Publishing Company. Pittner, Karin 1999 Adverbiale im Deutschen. Untersuchungen zu ihrer Stellung und Interpre‑ tation. (Studien zur deutschen Grammatik 60.) Tübingen: Stauffenburg. Poldauf, Ivan 1948 On the History of Some Problems of English Grammar before 1800. (Příspěvky k dějinám řeči a literatury anglické Sv. 7. Práce z vědeckých ústavu 55.) Prague: Prague University Press. Porter, David W. (ed.) 2002 “Excerptiones de Prisciano”. The Source for Ælfric’s Latin-Old Eng‑ lish Grammar. (Anglo-Saxon Texts 4.) Woodbridge: Brewer. Quirk, Randolph 1954 The Concessive Relation in Old English Poetry. (Yale Studies in English 124.) New Haven: Yale University Press. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A Comprehensive Grammar of the English Language. London: Longman. Quirk, Randolph and Charles L. Wrenn 1957 An Old English Grammar. (Methuen’s Old English Library.) New York: Holt, Rinehart & Winston. Raible, Wolfgang 1992 Junktion. Eine Dimension der Sprache und ihre Realisierungsformen zwischen Aggregation und Integration, vorgetragen am 4. Juli 1987. (Sitzungsberichte der Heidelberger Akademie der Wissenschaften. Philosophisch-Historische Klasse 1992/2.) Heidelberg: Winter. Ramat, Paolo and Davide Ricca 1998 Sentence adverbs in the languages of Europe. In: Johan van der Auwera in Collaboration with Dónall P. Ó Baoill (eds.), Adverbial Constructions

Secondary sources   271 in the Languages of Europe, 187–275. (Empirical Approaches to Language Typology 20, 3. EUROTYP 3.) Berlin: Mouton de Gruyter. Rissanen, Matti 1997 Towards an integrated view of the development of English: Notes on causal linking. In: Jacek Fisiak and Marcin Krygier (eds.), Advances in English Historical Linguistics (1996), 389–406. (Trends in Linguistics. Studies and Monographs 112.) Berlin/New York: Mouton de Gruyter. Rissanen, Matti 1999a On the adverbialisation of rather: Surfing for historical data. In: Hilde Hasselgård and Signe Oksefjell (eds.), Out of Corpora. Studies in Hon‑ our of Stig Johansson, 49–59. (Language and Computers 26.) Amsterdam: Rodopi. Rissanen, Matti 1999b Syntax. In: Roger Lass (ed.), The Cambridge History of the English Language. Volume.  3: Early Modern English 1476–1776, 187–331. Cambridge: Cambridge University Press. Rissanen, Matti 2006 Latin influence on an Old English idiom: “to Wit”. In: John Walmsley (ed.), Inside Old English. Essays in Honour of Bruce Mitchell, 222–241. Malden, Massachusetts: Blackwell. Rissanen, Matti, Merja Kytö and Minna Palander-Collin 1993 Early English in the Computer Age. Explorations through the Helsinki Corpus. (Topics in English Linguistics 11.) Berlin: Mouton de Gruyter. Robinson, Ian 1998 The Establishment of Modern English Prose in the Reformation and the Enlightenment. Cambridge: Cambridge University Press. Robinson, Orrin W. 1997 Clause Subordination and Verb Placement in the Old High German Isidor Translation. (Germanische Bibliothek/Neue Folge 3, 26.) Heidelberg: Winter. Romaine, Suzanne 1998 The Cambridge History of the English Language. Volume 4: 1776–1997. Cambridge: Cambridge University Press. Rudolph, Elisabeth 1989 The role of conjunctions and particles for text connexity. In: Maria-Elisa­ beth Conte, János S. Petöfi and Emel Sözer (eds.), Text and Discourse Connectedness. Proceedings of the Conference on Connexity and Co‑ herence, Urbino, July 16–21, 1984, 175–191. (Studies in Language. Companion Series 16.) Amsterdam: Benjamins. Rudolph, Elisabeth 1996 Contrast. Adversative and Concessive Relations and their Expressions in English, German, Spanish, Portuguese on Sentence and Text Level. (Research in Text Theory 23.) Berlin/New York: Mouton de Gruyter. Salmon, Vivian 1988 English punctuation theory 1500–1800. Anglia 106 (3): 285–314.

272   References Samuels, Michael 1972 Linguistic Evolution with Special Reference to English. (Cambridge Studies in Linguistics 5.) Cambridge: Cambridge University Press. Sauer, Hans 2006 Adverbs and adverbials in the earliest English text (Épinal-Erfurt). In: Andrew James Johnston, Ferdinand von Mengden and Stefan Thim (eds.), Language and Text. Current Perspectives on English and Ger‑ manic Historical Linguistics and Philology, 255–268. (Anglistische Forschungen 359.) Heidelberg: Winter. Scheler, Manfred 1977 Der englische Wortschatz. (Grundlagen der Anglistik und Amerikanistik 9.) Berlin: Schmidt. Schiffrin, Deborah 1987 Discourse Markers. (Studies in Interactional Sociolinguistics 5.) Cambridge: Cambridge University Press. Schleburg, Florian 2002 Altenglisch swa: Syntax und Semantik einer polyfunktionalen Partikel. (Sprachwissenschaftliche Studienbücher.) Heidelberg: Winter. Schmid, Hans-Jörg 2005 Englische Morphologie und Wortbildung: Eine Einführung. (Grundlagen der Anglistik und Amerikanistik 25.) Berlin: Schmidt. Scragg, Donald 1974 A History of English Spelling. (Mont Follick Series 3.) Manchester: Manchester University Press. Sorva, Elina 2007 From an analytic phrase to an atomic connective: The grammaticalization of Albeit. In: Ursula Lenker and Anneli Meurman-Solin (eds.), Connectives in the History of English, 115–143. (Amsterdam Studies in the Theory and History of Linguistic Science Series IV. Current Issues in Linguistic Theory 283.) Amsterdam: Benjamins. Sprat, Thomas 1959 History of the Royal Society. Edited with Critical Apparatus by Jackson J. Cope and Harold W. Jones. London: Routledge. First published London 1667. Sperber, Dan and Deidre Wilson 1996 Relevance. Communication and Cognition. Oxford: Blackwell. Stanley, Eric G. 2000 Hwæt. In: Jane Roberts and Janet Nelson (eds.), Essays on Anglo-Saxon and Related Themes in Memory of Lynne Grundy, 525–556. (King’s College London Medieval Studies 17.) London: King’s College, Centre for Late Antique and Medieval Studies. Stempel, Wolf-Dieter 1964 Untersuchungen zur Satzverknüpfung im Altfranzösischen. (Archiv für das Studium der neueren Sprachen und Literaturen: Beihefte 1.) Braunschweig: Westermann.

Secondary sources   273 Stockwell, Robert P. and Donka Minkova 1991 Subordination and word order change in the history of English. In: Die­ ter Kastovsky (ed.), Historical English Syntax, 367–408. (Topics in English Linguistics 2.) Berlin: Mouton de Gruyter. Stoffel, Cornelis 1901 Intensives and Down‑Toners. A Study in English Adverbs. (Anglistische Forschungen 1.) Heidelberg: Winter. Swan, Toril 1988 The development of sentence adverbs in English. Studia Linguistica 42 (1): 1–17. Swan, Toril 1988 Sentence Adverbials in English: A Synchronic and Diachronic Investi‑ gation. (Tromsø-studier i språkvitenskap 10.) Oslo: Novus. Sweetser, Eve 1990 From Etymology to Pragmatics. Metaphorical and Cultural Aspects of Semantic Structure. (Cambridge Studies in Linguistics 54.) Cambridge: Cambridge University Press. Thim, Stefan 2006 Phrasal verbs in everyday English: 1500–1700. In: Andrew James Johnston, Ferdinand von Mengden and Stefan Thim (eds.), Language and Text. Current Perspectives on English and Germanic Historical Lin‑ guistics and Philology, 290–306. (Anglistische Forschungen 359.) Heidelberg: Winter. Thomas, Francis-Noël and Mark Turner 1994 Clear and Simple as the Truth. Writing Classic Prose. Princeton: Prince­ton University Press. Thompson, Sandra A. 1985 Grammar and written discourse: Initial vs. final purpose clauses in English. Text 5 (1): 55–84. Tomlin, Russell S. (ed.) 1987 Coherence and Grounding in Discourse: Outcome of a Symposium, Eu‑ gene, Oregon, June 1984. (Typological Studies in Language 11.) Amsterdam: Benjamins. Traugott, Elizabeth Closs 1992 Syntax. In: Richard M. Hogg (ed.), The Cambridge History of the Eng‑ lish Language. Volume 1: The Beginnings to 1066, 168–289. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs 1995 The role of the development of discourse markers in a theory of grammaticalization. [Paper presented at ICHL XII; Manchester, 1995; (online version accessed 12/11/1997; )]. Traugott, Elizabeth Closs 1997 The discourse connective after all: A historical pragmatic account. [Paper prepared for ICL, Paris, July 1997; (online version accessed 12/11/2003; )].

274   References Traugott, Elizabeth Closs 1999 The rhetoric of counter-expectation in semantic change: A study in subjectification. In: Andreas Blank and Peter Koch (eds.), Historical Semantics and Cognition, 61–89. (Cognitive Linguistics Research 13.) Berlin/New York: Mouton de Gruyter. Traugott, Elizabeth Closs 2004 Historical pragmatics. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 538–561. (Blackwell Handbooks in Linguistics 16.) Oxford: Blackwell. Traugott, Elizabeth and Richard B. Dasher 2002 Regularity in Semantic Change. (Cambridge Studies in Linguistics 97.) Cambridge: Cambridge University Press. Uhler, Karl 1926 Die Bedeutungsgleichheit der altenglischen Adjektiva und Adverbia mit und ohne ‑lic(lice). (Anglistische Forschungen 62.) Heidelberg: Winter. Ungerer, Friedrich 1988 Syntax der englischen Adverbialien. (Linguistische Arbeiten 215.) Tübingen: Niemeyer. Valera, Salvador 1998 On subject-orientation in English ly‑adverbs. English Language and Linguistics 2 (2): 263–282. Vandenbergen-Simon, Anne-Marie and Karin Aijmer 2001 The expectation marker of course in a cross‑linguistic perspective. Lan‑ guages in Contrast 4 (1): 13–43. Vorlat, Emma 1975 The Development of English Grammatical Theory 1586–1737. With Special Reference to the Theory of Parts of Speech. Leuven: Leuven University Press. Walker, James A. 1949 Gothic ‑leik‑ and Germanic *lik in the Light of Gothic translations of Greek originals. Philological Quarterly 28: 274–293. Wegener, Heide 2000 Koordination und Subordination – Semantische und Pragmatische Unterschiede. In: Michel Lefèvre (ed.), Subordination in Syntax, Semantik und Textlinguistik, 33–44. (Eurogermanistik 15.) Tübingen: Stauffenburg. Westin, Ingrid 2002 Language Change in English Newspaper Editorials. (Language and Computers 44.) Amsterdam: Rodopi. Wiegand, Nancy 1982 From discourse to syntax: for in English causal clauses. In: Anders Ahlqvist (ed.), Papers from the 5th International Conference on His‑ torical Linguistics, 385–393. (Amsterdam Studies in the Theory and History of Linguistic Science 4, 21.) Amsterdam: Benjamins. Wischer, Ilse and Gabriele Diewald (eds.) 2002 New Reflections on Grammaticalization. (Typological Studies in Language 49.) Amsterdam: Benjamins.

Secondary sources   275 Workman, Samuel K. 1940 Fifteenth Century Translation as an Influence on English Prose. (Prince­ton Studies in English 18.) Princeton/New York: Princeton University Press.

Appendix A

A.1.  Adverbial connectors: items The following list provides an inventory of all linguistic elements which have served the function of an adverbial connector in any period of the history of English, giving the date for their first occurrence as an adverbial connector as well as the periods in which they were used as adverbial connectors. The dates for the first occurrence refer to the first attested use of the element in question as a linking adverbial. This means that the lexeme itself can be much older, but did not have a connective function on the sentence or text level before or after this date (for the methodology and examples, see Chapter 6.4). Items in square brackets were found in the OED or the MED, but are not attested in the corpus material of the present study (for details on the corpus texts, see Appendix C.1). Italics are used to indicate that the items were found in an adverbial connector function in the corpus texts but that there is no such meaning given in the OED. Dates are given in italics if the items are attested earlier in an adverbial connector function than stated by the OED. The inventory only comprises single word adverbs or lexicalized items such as verbal (e. g. albeit, howbeit) or prepositional phrases (e. g. after all, on the other hand, at the same time). It does not record fully transparent prepositional phrases such as as a consequence or by contrast (for the criteria applied, see Chapter 7.4). The items are listed alphabetically; lexicalized, but not yet univerbated prepositional phrases such as above all or at any rate are given under their respective nominal head element, i. e. all and rate. Length marks are only given for the Old English items. Symbols: ° ambiguous adverb/conjunction (for details, see Chapter 5.3) [ ] item not attested in the corpus texts (for this period) italics item not listed as an adverbial connector in the OED date in italics pre‑dating of adverbial connector function (earlier than the first occurrence in the OED)

[cūþlīce] ēac

ǣrest ‘first’

°algates 1386

1350–1420 (ME3)

eek

eek

[certain 1330] certain [certainly 1300] certainly [certes 1300] certes

Old English 1100–1350 (OE) (ME1/2)

contrariwise 1450

certain [certainly] certes

[algates]

after 1400/1450

1420–1500 (ME4)

contrarily 1540 contariwise

[certainly] certes [consequently 1533]

again 1533 albeit 1531 [algates]

1500–1570 (EModE1)

contrary 1593 [contrarily] contrariwise [contrariways 1588]

[contrarily] [contrariwise]

[consequently]

besides

above all [after all 1713]

again

1640–1710 (EModE3)

[certainly] certes consequently

[certainly]

beside 1592 besides 1596

above all 1625

again albeit

1570–1640 (EModE2)

1780–1850 (LModE2)

[contrarily] [contrariwise]

consequently

[certainly]

besides

above all after all

again

above all after all

again

accordingly

1850–1920 (LModE3)

also, see eall swā above all after all [all in all] [all the same]

again

accordingly

1920– (PDE)

contrarily contrariwise

of course 1811

of course

conversely 1806 [conversely]

[contrarily] [contrariwise]

consequently

certainly

certainly certes consequently

[conversely] correspondingly of course

[contrarily] [contrariwise]

consequently

certainly

besides besides in any case 1920 in any case

besides

[all the same 1878] [altogether 1817] [altogether 1817] [altogether 1817] [anyhow 1825] [anyhow] [anyhow] [anyway 1859] [anyway]

above all after all

again

accordingly 1711 accordingly

1710–1780 (LModE1)

Adverbial connectors: items   277

hēr

°forþȳ

°forþǣm °forþon

°ēac swylce °eall‑swā eft [eftsōna/-es] elles eornostlīce

here herefore 1200

on oðer half

[first 1300] °for °forhwi 1225 [forsoothe 1320] °forþon °forþon °for that forþi

eac swlyce alswa eft eftsona/-es and/or elles

Old English 1100–1350 (OE) (ME1/2)

for that forþi for which

for that forþi for which [further 1405]

here herefore

here [herefore] howbeit 1465

ferthermore 1382 furthermore furthermore 1390 [fartherover], furtherover 1390

first °for forhwi forsoothe

first °for forhwi forsoothe

finally 1425

and/or else

and/or elles

[ergo 1400]

also

1420–1500 (ME4)

also

1350–1420 (ME3)

[here] [herefore] howbeit

for which further [farther 1562] furthermore

finally [in fine 1550] first °for

or else

also

1500–1570 (EModE1)

[here] herefore howbeit however 1613

[furtherover]

furthermore

for the which further

[finally] [in fine] first °for

or else

also

1570–1640 (EModE2)

however

however

however

[furthermore]

further

in all events in fact finally in fine first °for

or else

also

1850–1920 (LModE3)

however

[furthermore]

[further]

in all events 1811 in fact 1815 [finally] in fine first °for

or else

also

1780–1850 (LModE2)

on the one/other hand hence here

[furthermore]

[further]

[finally] [in fine] first °for

or else

[also]

1710–1780 (LModE1)

on the other hand on the other hand on the one/other hand hence 1660 hence hence [here] here here

furthermore

[further]

[finally] [in fine] first °for

or else

also

1640–1710 (EModE3)

however

on the one/ other hand hence here

furthermore

further

first °for

in all events in fact finally

or else

also

1920– (PDE)

278   Appendix A

nāþȳlǣs

hūru hūruþinga hwæt hwæt þā °hwæþ(e)re

1350–1420 (ME3)

natheles [neverthelatter 1225]

[more 1230]

[more] moreover natheles neverthelatter nevertheless

nevertheless 1382

at last

at last

at last

[nevertheless] [nonetheless]

[nevertheless] [nonetheless] 1533

[nonetheless]

[nevertheless]

[nonetheless]

[nevertheless]

[more] moreover

[meanwhile]

[meanwhile] [more] moreover

lastly at least likewise [likeways] [in the mean‑ time]

[at last]

indeed

lastly at least likewise [likeways] [in the mean‑ time]

[at last]

indeed

[howsomever]

°howsoever 1586 °howsoever [howsomever] [howsomever]

indeed 1600

1710–1780 (LModE1)

1640–1710 (EModE3)

1570–1640 (EModE2)

lastly 1586 at least 1605 likewise 1509 likewise [likeways 1551] [likeways] in the meantime 1631 meantime (in the) mean‑ while 1597 [more] [more] moreover moreover

item

[howsomever 1562]

1500–1570 (EModE1)

item

1420–1500 (ME4)

[more] moreover 1385 natheles [neverthelatter]

item 1398 iwis 1175 at (the) last 1275 at last alast 1230

hwæt hwæt þa

Old English 1100–1350 (OE) (ME1/2)

[lastly] at least likewise

at last

indeed

[howsomever]

1850–1920 (LModE3)

[nonetheless]

[nevertheless]

[more] moreover

[meanwhile]

[nonetheless]

[nevertheless]

[more] moreover

[meanwhile]

in the meantime [in the mean‑ time]

lastly at least likewise

[at last]

indeed

[howsomever]

1780–1850 (LModE2)

nonetheless

[nevertheless]

[more] moreover

[meanwhile]

[in the mean‑ time]

[lastly] at least likewise

at last

incidentally 1925 indeed instead

1920– (PDE)

Adverbial connectors: items   279

soothly

swa

°swā

oðer side 1225

[overall] 1225

sōþlīce

°nū

1350–1420 (ME3)

suingly

suingly 1380

so

soothly

sekirly

sekirly

soothly

overmore

so

1710–1780 (LModE1)

surely 1556 so

now [otherways] otherwise

now [otherways] [otherwise]

now [otherways] [otherwise]

1780–1850 (LModE2)

now [otherwise]

[otherwise]

in sum sure surely so

[in sum] sure surely so

still 1707

[in sum] sure surely so

still

on the other side on the other side

[in sum] sure surely so

still

1920– (PDE)

now

1850–1920 (LModE3)

[in sum] [sure] surely so

still

on the other side

[in sum] [sure] [surely] so

still

similarly so see swā

[plus 1968] at any rate 1811 [at any rate] rather

over and above [over and above] [over and above] [over and above] over and above 1588 over and besides

now [otherways] otherwise

semblably 1531 [semblably] on the other side

[over and besides] 1533

now [otherways] [otherwise] over that over this

notwithstanding notwithstanding notwithstanding notwithstanding

1640–1710 (EModE3)

notwithstanding 1440 [nought‑for‑than] [nought‑for‑that] [nought‑for‑thi] now [otherways] [otherwise] over that over this

1570–1640 (EModE2)

1500–1570 (EModE1)

1420–1500 (ME4)

overmore 1382

[nought‑for‑than] [nought‑for‑than] [nought‑for‑that] [nought‑for‑thi] °nu nu otherways 1390 [otherwise 1393] over that 1393 [over this 1393]

Old English 1100–1350 (OE) (ME1/2)

280   Appendix A

°þeah þeahhwæþere thane/thenne

[toeken]

þa

swaþeah

(ge)wīslīce witodlīce

witodlice yet (giet) 1205

yet

yet

yet

wherefore

well

well

verily

[verily 1382]

wherefore

verily

verily

truly

truly 1417

yet

wherefore

truly

truly

therefore thus

therefore thus

then

1570–1640 (EModE2)

therefore thus 1380

therefore thus

then

then

1500–1570 (EModE1)

°tho then

1420–1500 (ME4)

°þeah

1350–1420 (ME3)

°wherefore 1275 wherefore

°þonne, see then °þȳ þi therefore 1200

swāþēah swāþēahhwæþre °þā [þā gīet] þǣrtōēacan þæs °þēah þēahhwæþere °þonne

Old English 1100–1350 (OE) (ME1/2)

yet

well whence 1665 wherefore

verily videlicet

truly

therefore thus

then thence 1652

1640–1710 (EModE3)

yet

wherefore

well

therefore thus at the same time 1739 too truly in truth 1770

then

1710–1780 (LModE1)

then

°though

1850–1920 (LModE3)

then

°though

1920– (PDE)

yet

well whence

too truly

too

yet

wherefore

yet

in the same way in the same way well well

too truly in truth verily

therefore therefore therefore thus thus thus at the same time at the same time at the same time

then

1780–1850 (LModE2)

Adverbial connectors: items   281

282   Appendix A A.2.  Alphabetical index of adverbial connectors Adverbial Connector

Semantics

Listed in First Appendix Occurrence (in Corpus Texts)

Last Occurrence (in Corpus Texts)

ǣrest ‘first’ accordingly after ‘moreover’ again

listing/enumerative

B.5.1 B.3.1 B.1.2 B.1.2

OE PDE ME4 PDE

(but yet) again albeit °algates

OE LModE1 ME4 EModE1 (EModE2)

contrast

B.4.1 B.4.1

EModE1 ME3

all in all above all after all

additive /reinforcing

EModE2 EModE1 (ME3) PDE PDE PDE

all the same also (OE eall‑swā) altogether anyhow

contrast/concession

result additive /reinforcing additive /reinforcing

contrast/concession contrast/concession

additive /reinforcing contrast/concession

additive /reinforcing contrast/concession/

in any case

– transition

contrast/concession dismissive

beside besides

B.1.2

summative dismissive

anyway

B.1.2 B.4.1

– transition

additive /reinforcing additive /reinforcing

B.1.2 B.1.2

contrast/concession

B.4.1

PDE EModE2 EModE3 (LModE1) LModE3 OE LModE2 LModE2

PDE PDE PDE PDE

LModE3

PDE

EModE2 EModE2 (EModE3) LModE3

EModE2 PDE

ME1/2 (ME3) ME1/2 ME1/2 EModE1 (LModE1) EModE2 EModE1 EModE2 ME4 (EModE1) LModE2 (LModE3) PDE LModE2

ME4

PDE

dismissive

certain

?transition

B.5.2

certainly certes consequently

?transition

B.5.2 B.5.2 B.3.1

contrary contrarily contrariways contrariwise

contrast/antithetic

transition result

contrast/antithetic

B.4.2 B.4.2

contrast/antithetic

B.4.2

conversely

contrast/antithetic

B.4.2

correspondingly of course

additive /equative

B.1.3 B.4.1 B.5.2

contrast/antithetic

contrast/concession transition

cūþlīce ‘certainly’ ēac ‘also’ °ēac swylce/swylce ēac ‘also’ °eall‑swā see also eft ‘also’

result/transition

PDE LModE2 PDE EModE2 PDE LModE3 PDE PDE PDE

additive /reinforcing additive /reinforcing

B.1.2 B.1.2

OE OE OE

OE PDE ME1

additive /reinforcing

B.1.2

OE

ME2

Adverbial connectors: items   283 Adverbial Connector

Semantics

Listed in First Appendix Occurrence (in Corpus Texts)

Last Occurrence (in Corpus Texts)

eftsōna/eftsōnes ‘also’

additive /reinforcing

B.1.2

ME1

additive /reinforcing

B.4.3

OE (ME1) OE

transition

B.5.2 B.4.1

OE ME3 LModE2

OE LModE1 PDE

B.5.2

LModE2

PDE

ME4 (EModE1) EModE1 (LModE2) ME1/2 (ME3) ME1/2

PDE

else (OE elles) eornostlīce ‘earnestly’ ergo at/in all events

contrast/reformulatory result contrast/concession

PDE

dismissive

in fact

contrast/avowal

–

transition

finally

listing/enumerative

B.1.1

in fine

summative

B.2

first

listing/enumerative

B.1.1

°for

cause

B.3.3

°forhwi forsooth

result

B.3.1 B.5.2

for that °forþǣm/°forþon °forþӯ for which further

result

furthermore/furthermore furtherover/fartherover

additive /reinforcing

B.1.2 B.1.2

ME1/2 ME1/2 (ME3) ME1/2 OE OE ME3 ME3 (EModE1) ME3 ME3

on oðer half on the other hand

contrast/antithetic contrast/

B.4.2 B.4.2

ME1/2 EModE3

hence

result

B.3.1

here (OE hēr) herefore how howbeit however

transition

?transition contrast/concession contrast/concession

B.5.1 B.3.1 B.5.4 B.4.1 B.4.1

°howsoever °howsomever hūru ‘nevertheless’ hūruþinga ‘nevertheless’ °hwæþ(e)re hwæt ‘what’ °hwæt þā ‘what then’ incidentally

contrast/concession

B.4.1

contrast/concession

B.4.1 B.4.1 B.4.1 B.5.4 B.5.4 B.5.2

EModE3 (LModE1) OE ME1/2 EModE2 ME4 EModE2 (EModE3) EModE1 EModE1 OE OE OE OE OE PDE

transition

result result result additive /reinforcing

additive /reinforcing

antithetic/replacive

result

contrast/concession contrast/concession contrast/concession transition transition transition

B.3.1 B.3.1 B.3.1 B.3.1 B.1.2

LModE3 PDE PDE (LModE3) ME3/4 ME4 ME4 ME2 ME4 EModE1 PDE PDE EModE2 (ME3) ME1/2 PDE PDE PDE EModE2 EModE2 EModE2 PDE EModE3 LModE3 OE OE OE ME1/2 ME1/2 PDE

284   Appendix A Adverbial Connector

Semantics

Listed in First Appendix Occurrence (in Corpus Texts)

Last Occurrence (in Corpus Texts)

indeed instead item iwis at (the) last lastly

transition

B.5.2 B.4.2 B.1.2

PDE PDE EModE1

listing/enumerative listing/enumerative

B.1.1 B.1.1

at least

contrast/concession

B.4.1

likeways likewise (in the) meantime

additive /equative

B.1.3 B.1.3 B.4.1

contrast/antithetic

additive /reinforcing transition

additive /equative

contrast/concession time

(in the) meanwhile more moreover

contrast/concession

additive /reinforcing

B.1.2

nāþӯlǣs ‘nevertheless’

contrast/concession

B.4.1

neverthelatter

contrast/concession

B.4.1

nevertheless next nonetheless notwithstanding nought‑for‑than ‘nevertheless’ nought‑for‑that ‘nevertheless’

contrast/concession

B.4.1 B.1.1 B.4.1 B.4.1

additive /reinforcing

?listing/enumerative contrast/concession contrast/concession contrast/concession contrast/concession

nought‑for‑thi ‘nevertheless’ now (OE °nū) otherways otherwise over this/that over and above over and besides

contrast/concession

additive /reinforcing

B.4.2 B.1.2 B.1.2 B.1.2

overmore overall peradventure perchance perhaps plus at any rate

additive /reinforcing

B.1.2

?transition ?transition ?transition additive /reinforcing contrast/concession

B.5.3 B.5.3 B.5.3

transition contrast/antithetic contrast/antithetic

additive /reinforcing additive /reinforcing

additive /reinforcing

EModE2 PDE ME3 ME1/2 ME2 EModE2 EModE2 (EModE3) EModE1 EModE1 EModE2 (LModE2) EModE2 ME1/2 ME3 (ME4) OE (ME1/2) ME1/2 (ME4) ME3 ME3 EModE1 ME4 ME2 ME3

PDE PDE (LModE3) PDE LModE1 PDE PDE PDE PDE PDE ME4 ME4 PDE LModE3 PDE LModE1 ME4 ME4 ME4 PDE LModE2 PDE EModE1 LModE3 EModE2

B.4.1

ME3 OE ME3 EModE2 ME3 EModE2 EModE1 (EModE2) ME3 ME1/2 ME3 EModE1 LModE2 PDE LModE3 PDE ME3 EModE1 ME1/2

PDE ME4 EModE2 (EModE1) (LModE3)

B.5.1

ME4 PDE EModE1 EModE1 PDE PDE PDE

dismissive

rather sekirly semblably

contrast/reformulatory additive /equative

B.4.3 B.5.2 B.1.3

oðer side

contrast/antithetic

B.4.2

transition

Adverbial connectors: items   285 Adverbial Connector

Semantics

Listed in First Appendix Occurrence (in Corpus Texts)

Last Occurrence (in Corpus Texts)

similarly sōþlīce ‘truly’ still in sum suingly ‘consequently’

additive /equative

result

B.1.3 B.5.2 B.4.1 B.2 B.3.1

PDE ME4 PDE PDE ME4

sure

transition

B.5.2

surely °swā ‘so’ swāþēah ‘nevertheless’ swāþēahhwæðere ‘nevertheless’ þā ‘then’ þā gīet ‘also’ (þǣr)tōēacan ‘also’ þæs °þēah ‘though’ þēahhwæðere ‘nevertheless’ then (OE °þonne)

transition

B.5.2 B.3.1 B.4.1 B.4.1 B.5.1

transition contrast/concession summative

result contrast/concession contrast/concession transition additive /reinforcing additive /reinforcing

B.4.1 B.4.1 B.3.2

PDE OE EModE3 EModE2 ME3 (ME4) EModE2 (EModE3) EModE1 OE OE OE OE OE OE OE OE OE OE

PDE PDE ME1/2 OE ME1 OE ME1 OE PDE ME1/2 PDE

B.3.1 B.3.1 B.3.1 B.3.1 B.4.1

EModE3 ME1/2 ME3 OE LModE1

EModE3 PDE PDE ME1/2 PDE

LModE1 (LModE2) ME3 LModE1 ME3 (ME4) EModE3 LModE3 EModE1 EModE3 ME1/2 EModE1 OE OE ME1/2

PDE

B.1.2

result contrast/concession contrast/concession result/interference

PDE

time

thence therefore thus °þӯ at the same time

result result result result contrast/concession time

too

additive /reinforcing

B.1.2

truly in truth verily

transition

B.5.2 B.5.2 B.5.2

videlicet in the same way well whence °wherefore why gewīslīce ‘certainly’ witodlīce ‘certainly’ °yet

additive /equative

transition transition

additive /equative transition result result

?transition transition transition contrast/concession

B.1.3 B.1.3 B.5.5 B.3.1 B.3.1 B.5.4 B.5.4 B.5.2 B.4.1

LModE3 LModE3 EModE3 EModE3 PDE PDE EModE3 LModE3 LModE1 OE ME1/2 PDE

Word index English above all  11, 40, 81, 91, 98, 108, 112, 216, 276–277, 282; B.1.2. ac  17–18, 76, 192, 238; B.4.4. actually  127, 170–171 accordingly  81, 90–91, 93, 96, 140, 151, 153, 155, 167, 277, 282; B.3.1. ærest  58, 81, 92, 107, 109, 277, 282; B.1.1. after  81, 107, 112, 216, 282; B.1.2. after all  3, 10–11, 45–46, 81, 91, 98, 104, 108–111, 151, 154, 167, 180–182, 190, 249, 276–277, 282; B.4.1. again  39–40, 81, 93–94, 108–109, 112, 193, 216–217, 246, 277, 282; B.1.2. alas  B.5.5. albeit  4, 10, 25, 32, 79, 81, 97, 104, 173, 179, 182, 194, 276–277, 282; B.4.1. algates  81, 93, 96, 107, 112, 179, 182, 277, 282; B.4.1. all (OE eall)  11, 170, 174, 181–182, 190, 276 all in all  40, 81, 91, 98, 277, 282 all the same  81, 98, 171, 180, 182, 190, 277, 282 also (OE eall‑swa)  19, 151, 215–219, 277, 282; B.1.2. although  9, 27, 110, 169–173, 186, 190– 191, 204, 207–210, 226; see also though, þeah altogether  81, 98, 277, 282 and  9, 14, 17, 61, 76–77, 214–215, 224– 225, 238–240, 244; B.1.5. any  103, 170–171, 181–182 anyhow  40, 81, 91, 98, 103, 170–171, 180, 182, 190, 277, 282 anyway  38, 40, 45, 81, 91, 98, 103, 108, 112, 180, 182, 190, 198, 204, 277, 282 as  6, 8, 30–31, 164–165 at (the) last  81, 107, 109, 222, 231, 279, 284; B.1.1.

at any rate  11, 81, 98, 170–171, 276, 280, 284; B.4.1. at least  81, 98, 170–171, 180, 190, 279, 284; B.4.1. at the same time  81, 98, 108–109, 171, 180–181, 190, 250, 276, 281, 285; B.4.1. at/in all events  11, 81, 98, 180, 182, 190, 278, 283; B.4.1. because  6, 8, 22, 24, 28, 30–31, 100, 104, 132–140, 151, 153, 162, 164–167, 185; B.3.3. beside(s)  23, 39–40, 81, 93–97, 100, 108, 112, 216–218, 220, 235, 277, 280, 282; B.1.2. but  9, 25–26, 77, 79, 203–207, 238– 240; B.4.4. by/in consequence  B.3. by/in contrast  B.4.2. certain  79, 81, 89–90, 93, 107, 114–115, 228, 277, 282; B.5.2. certainly  79, 81, 93, 95, 107, 114–115, 228, 277, 282; B.5.2. certes  79, 81, 89–90, 93, 107, 114, 162– 163, 227–228, 231, 234, 277, 282; B.5.2. consequently  32, 38, 40–41, 81, 90–91, 93, 96, 106, 139–140, 151, 153, 155, 167, 277, 282; B.3.1. contrarily  81, 176–177, 277, 282; B.4.2. contrariways  81, 91, 95, 176–177, 277, 282 contrariwise  81, 90–91, 95, 97, 176– 177, 277, 282; B.4.2. contrary  81, 176–177, 277, 282; B.4.2. (on the) contrary  10, 40, 77, 96, 170– 171, 177 conversely  40, 81, 90–91, 93, 96, 177, 277, 282; B.4.2.

Word index   287 correspondingly  40, 81, 90–91, 93, 221, 277, 282; B.1.3. cuþlice  81, 92, 95, 107, 155, 277, 282 eac  18, 58–59, 64, 81, 92, 174, 216, 277, 282; B.1.2. eac swylce/swylce eac  59, 81, 97, 216, 278, 282; B.1.2. eala  B.5.5. eall‑swa  151, 215, 277; see also also efne  B.5.5. eft  17–18, 58–61, 64, 81, 92, 94, 107, 109, 216, 278, 282; B.1.2. eftsona/eftsones  17–18, 58, 81, 92, 94, 107, 109, 216, 278, 283; B.1.2. else  23, 58, 80–81, 92, 96, 176–177, 216–218, 278, 283; B.1.2., B.4.3. eornostlice  58, 81, 92, 95, 107, 114, 228, 231, 278, 283; B.5.2. ergo  81, 89–90, 93, 155, 157, 278, 283 ‑es  95–96 ‑ever  168, 182 farther  see further finally  81, 90–91, 93, 107, 109, 222, 278, 283; B.1.1. first(ly)  41, 81, 107, 112, 222–224, 231, 246, 278, 283; B.1.1. for  5, 8–9, 77, 152–154, 156, 161–164, 225–226, 239–240, 249, 278, 283; B.3.3. for as much as  132, 154, 164–166 for a start  B.1.1. for example  B.1.4. forhwi  79–81, 97, 155, 157–158, 162, 278, 283; B.3.1. for instance  B.1.4. forsooth  3, 39, 54–56, 80–81, 97, 107, 114, 122–124, 228, 232, 250, 283; B.5.2. for that/for this  6, 27, 79, 81, 97, 101– 102, 132, 155, 158–161, 166, 249, 278, 283; B.3.1. forþæm/forþon/forþy  5, 24, 30, 59, 63– 64, 73, 75, 81, 97, 99–102, 132,

140–157, 159–160, 162, 164, 166, 185, 187, 214, 248, 278, 283; B.3.1. for which  81, 97, 155, 157, 278, 283; B.3.1. further/farther  38–40, 81, 93–94, 107, 112, 216–218, 220, 278, 283; B.1.2. furthermore/farthermore  14, 16, 40, 81, 93–94, 107, 112, 215–219, 231, 250, 278, 283; B.1.2. furtherover/fartherover  81, 93–94, 107, 112, 216, 219, 278, 283; B.1.2. gan  13 gewislice  52, 59, 81, 92, 95, 107, 114, 285; B.5.2. giet  see yet hence  8, 23, 27, 39–40, 81–82, 93, 95, 96, 108, 112–113, 132, 134, 140, 151, 153, 155, 167, 235, 249, 278, 283; B.3.1. here  38, 58–61, 80–81, 92, 107–108, 112, 158, 227, 230, 246, 278, 283; B.5.1. herefore  79, 81, 97, 107, 112–113, 155, 157, 278, 283; B.3.1. how  46, 183, 229, 283; B.5.4. howbeit  10, 25, 81, 97, 104, 173, 179, 194, 276, 278, 283; B.4.1. however  9, 25–26, 30–32, 37, 39–40, 44, 81, 93, 108–109, 136, 170–171, 180, 182, 194–197, 200, 204, 213, 219, 226, 235, 245, 278, 283; B.4.1. howsoever  81, 93, 108–109, 178, 180, 182, 279, 283; B.4.1. howsomever  81, 93, 108–109, 179, 182, 279, 283 huru  58, 81, 92, 107, 114, 179, 279, 283; B.4.1. huruþinga  59, 81, 92, 107, 114, 179, 279, 283; B.4.1. hwæðere  58, 81, 97, 181, 186–187; B.4.1. hwæt  3, 97, 99, 227, 229–230, 279, 283; B.5.4. hwæt þa  81, 97, 99, 227, 229, 279, 283; B.5.4.

288   Word index in addition  B.1.2. in any case  40, 81, 171, 180, 182, 190, 277, 282; B.4.1. in conclusion  B.2. in fact  3, 39, 45–46, 81, 98, 105, 108, 114, 124, 127–129, 151, 154, 170–171, 180, 190, 209, 227–231, 278, 283; B.4.1.; B.5.2. in fine  81, 98, 108–109, 221, 278, 283; B.2. in other words  B.1.4. in sum  40, 81, 98, 221, 280, 285; B.2. in the same way  81, 98, 108, 112, 221, 281, 285; B.1.3. in truth  81, 98, 108, 114, 228, 281, 285; B.5.2. incidentally  40, 81, 90–91, 93, 279, 283; B.5.2. indeed  3, 10, 39, 45–46, 81, 98, 103, 105, 108, 114, 124, 127–128, 151, 153– 154, 162, 227–228, 230, 279, 284; B.5.2. instead  3, 40, 81, 98, 100, 108, 112, 170–171, 176–177, 279, 284; B.4.2. item  18, 81, 89–90, 93, 216, 219, 279, 284; B.1.2. it is true  B.5.2. iwis  81, 93, 107, 114, 228, 231, 279, 284; B.5.2. la/loo  B.5.5. lastly  81, 108–109, 223, 279, 284; B.1.1. likeways  81, 93, 221, 279, 284 likewise  35, 39–40, 81, 93, 95, 217, 221, 279, 284; B.1.3. ‑ly  33, 90–91, 93, 95–96 (in the) meantime  81, 98, 108–109, 180–181, 184, 190, 279, 284; B.4.1. (in the) meanwhile  40, 81, 98, 108–110, 180–181, 184, 190, 250, 279, 284 ‑more  93–95, 250 moreover  39–41, 81, 93–94, 107, 112, 215–217, 219, 250, 279, 284; B.1.2.

naþylæs  58–59, 81, 97, 99, 174, 178– 179, 183, 186–187, 220, 250, 279 neverthelatter  79, 81, 97, 107, 109, 179, 279, 284; B.4.1. nevertheless  14, 16, 39–41, 79, 81, 97, 101, 107, 109, 171–173, 178–179, 181, 183, 193, 245, 250, 279, 284; B.4.1. next  284; B.1.1. no/nay  B.5.5. no doubt  B.5.2. nonetheless  38, 45, 81, 97, 101, 172–173, 179, 183, 186, 208, 279, 284; B.4.1. notwithstanding  10, 81, 91, 97, 104, 107, 111–112, 173, 179, 183, 280, 284; B.4.1. nought‑for‑than  81, 97, 179, 181, 183, 280, 284 nought‑for‑that  81, 97, 100, 179, 181, 183, 280, 284 nought‑for‑thi  81, 97, 100, 181, 183, 280, 284 now  38–40, 74, 80, 102, 108, 116, 121, 137, 140, 151, 153, 162, 227, 230, 244, 280, 284; B.5.1.; see also nu nu  38, 58–64, 81, 92, 107–109, 144–146, 152, 158, 185, 192, 280, 284; B.5.1.; see also now other side  79–81, 97, 107, 112, 176, 280, 284; B.4.2. of course  81, 98, 104, 108–109, 115, 151, 154, 180, 190, 227, 231, 277, 282; B.4.1., B.5.2. oh  B.5.5. on other half  79–81, 97, 107, 112, 176, 278, 283; B.4.2. on the one hand  B.4.2. on the other hand  40, 81, 96, 98, 108, 112, 171, 176–177, 276, 278, 283; B.4.2. or  76; B.4.4. otherways  81, 97, 107, 112, 176–177, 280, 284 otherwise  81, 97, 176–177, 280, 284; B.4.2. oþþe  76; B.4.4.

Word index   289 over and above  10, 81, 98, 104, 108, 112, 216, 220, 280, 284; B.1.2. over and besides  81, 97, 108, 112, 216, 220, 280, 284; B.1.2. over that/this  6, 27, 81, 97, 100–101, 107, 112, 216, 218, 280, 284; B.1.2. overall  40, 79, 81, 97, 107, 112, 280, 284 overmore  17, 79, 81, 93–94, 107, 112, 216, 219–220, 250, 280, 284; B.1.2. peradventure  47, 228–229, 284; B.5.3. perchance  47, 228–229, 284; B.5.3. perhaps  24, 47, 228–229, 284; B.5.3. plus  81, 89–90, 93, 216, 219, 280, 284 rather  3, 81, 93, 95, 108, 170–171, 176– 177, 280, 284; B.4.3. really  B.5.2. second(ly)  B.1.1. sekirly  81, 93, 95, 107, 114, 228, 280, 284; B.5.2. semblably  81, 90, 93, 221, 280, 284; B.1.3. similarly  40, 81, 90–91, 93, 106, 221, 280, 285; B.1.3. since  4, 6, 8, 30–31, 110, 133–134, 136–137, 140, 151, 153, 164–165 soothly  48, 117–123, 126–127, 129, 231–232, 234, 236, 280 soþlice  3, 39, 46, 48, 51–53, 59, 81, 92, 95, 107, 114, 116–123, 125, 129, 216, 227– 228, 231–232, 250, 280, 285; B.5.2. still  31, 40, 79, 81, 93–94, 108–110, 171–172, 180–181, 183–184, 191–194, 204, 246, 280, 285; B.4.1. suingly  81, 280, 285; B.3.1. sure  81, 93, 108, 114–115, 228, 230– 231, 280, 285; B.5.2. surely  81, 91, 93, 95, 108, 114–115, 130, 162, 228, 231, 280, 285; B.5.2. swa  3, 5, 59, 66, 81, 92, 97, 99–100, 151– 152, 155, 178, 215, 277, 280, 285; B.3.1. swaþeah  5, 58, 66–67, 71, 81, 97, 100, 178–179, 186–187, 189, 281, 285; B.4.1.

swaþeahhwæðere  5, 58, 66–67, 100, 174, 178, 181, 186–187, 189, 220, 250, 285; B.4.1. þa  3, 46, 58–60, 62, 64–66, 68, 70–72, 81, 92, 107–109, 151–152, 227, 230, 238–239, 281, 285; B.5.1. þa giet  79, 81, 97, 107, 109, 112, 216, 281, 285 þæm  5, 24, 27, 73–74, 84–85, 99, 101, 140, 143, 147–149, 151, 156, 159, 249 (þær)toeacan  58–60, 81, 94, 107, 109, 112, 174, 216, 219–220, 281; B.1.2. þæs  59, 73, 81, 97–99, 155, 281, 285 þe  5, 27, 74–75, 102, 142–144, 147, 152, 160–161, 178, 183, 187–189 þeah  19–20, 59, 63, 66–67, 71–72, 81, 92, 179, 184–192, 214, 281, 285; see also though þeahhwæðere  see swaþeahhwæðere þon  5, 24, 73, 99, 140, 147–149, 151, 156, 159 þonne  58–59, 68, 70–72, 81, 92, 107– 109, 144–145, 151–152, 155, 281, 285 þy  5, 24, 59, 73, 81, 97–101, 140, 144– 145, 147–149, 151–152, 156, 159, 178, 183, 187, 281, 285; B.3.1. that  5, 6, 91–92, 100–102, 249 then (OE þonne)  13–14, 31, 38–41, 44– 45, 72, 80, 108–109, 111, 131, 140, 151, 153, 155, 162, 164, 167, 198, 201, 212– 213, 218, 236, 245, 281, 285; B.3.2. thence  81, 93, 95–96, 108, 112–113, 155, 281, 285; B.3.1. there  112–113, 158 therefore  3, 8, 14, 24, 27–28, 30, 38– 40, 72, 79, 81, 90, 97, 100, 107, 112– 113, 133–140, 151–155, 157–158, 164, 166, 231, 234–236, 245, 281, 285; B.3.1. this  6, 100–102, 249 though  9, 19–20, 25–26, 30–32, 40, 44–45, 67, 78, 80, 171–172, 174–175, 179, 184–213, 249, 281, 285; B.4.1.

290   Word index thus  8, 27, 38, 40, 55, 81–82, 100, 140, 151, 154–155, 157, 162, 246, 281, 285; B.3.1. too  39–40, 81, 93–94, 174, 200, 213, 216–217, 219, 235, 281, 285; B.1.2. truly  48, 81, 93, 95, 107, 114, 124–127, 162, 228, 230–231, 281, 285; B.5.2. verily  81, 91, 93, 95, 107, 114, 129, 228, 230, 232, 250, 281, 285; B.5.2. videlicet  81, 89–90, 93, 221, 281, 285; B.1.3. well  3, 81, 93, 122, 281, 285; B.5.5. what  229–230; B.5.4.; see also hwæt what is more  B.1.2. whence  81, 93, 95–96, 108, 112–113, 155, 281, 285; B.3.1. where  112–113, 158 wherefore  3, 14, 56, 79, 81, 97, 107, 112–113, 152–155, 157–158, 166, 281, 285; B.3.1. while  4, 150, 171–172, 204 why  22, 46, 229–230, 285; B.5.4. ‑wise  90, 95–96, 177 (to) wit  B.1.4. witodlice  3, 46, 52–53, 59, 81, 92, 95, 107, 114, 116–117, 125, 162, 227–228, 231, 281, 285; B.5.2. yes/yea  B.5.5. yet (OE giet)  16, 25–26, 28, 31, 39, 79, 81, 93, 107, 109–110, 171–172, 179, 183–184, 191–194, 196, 204, 226, 231, 281, 285; B.4.1. French car  137, 162 non obstant  104, 183 parce que  74–75, 137, 156, 166, 249 puisque  75, 137, 218 que  74–75, 156, 166, 174, 249 tout soit il que  104, 173

German außerdem  199 da  30–31, 137, 150, 181 daher  150, 151 darum  150, 199 dass  74, 150 demnach  24, 27, 73, 100, 149, 150 denn  27, 30–32, 137, 140, 150, 164 deshalb  24, 27, 32, 73, 139, 150–151, 199 deswegen  26–28, 73, 100, 132, 135, 149, 150, 151, 199 falls  31 infolgedessen  100, 150 nämlich  26–27, 31, 132, 135, 137, 139–140, 150, 164 obwohl  9, 31–32, 202–203, 212–213, 249 somit  150–151 während  110 weil  26–28, 30–32, 110, 135, 137, 150, 212 weshalb  150 weswegen  150 Latin autem  52–56, 123–124 cum  164–165 enim  27, 52–56, 71, 123–124, 132, 137, 140, 161–164 ergo  53, 154 igitur  27, 53, 164 itaque  53 nam  27, 52–53, 71, 132, 137, 140, 161– 164 quare  164 quia  137, 160, 164 quidem  52, 55, 123–124 quod  74, 137 vero  52–55, 123–124

Subject and name index addition  40–41, 45, 58–59, 77, 94–95, 109, 111–112, 131, 154, 162, 214–226, 231, 234, 250, 282–285 adverb, classification/definition of  33– 39, 49–57 adverb, formation of  33, 53, 90–96 adverb/conjunction, ambiguous  5, 20, 59–60, 64–66, 78–80, 84–85, 96, 99, 104, 113, 132, 143, 152, 157, 181, 185, 248 adverbial cline  47–48, 117, 122, 124, 127–130, 182, 227, 231–232, 250 adverbial connectors, semantic categories  39–41, 44–45 adverbial, linking  4, 10, 22, 25, 35–39, 42–46, 50, 53, 87, 182, 201, 224, 238 Ælfric  51–53, 142 ambiguous adverb/conjunction  see adverb/conjunction, ambiguous Blair, Hugh  243 Campbell, George  2, 8, 39, 57, 233, 243–246 cause/result  13, 16, 21, 27, 30–32, 40–41, 45, 58–59, 73, 80, 85, 96, 99, 100, 104, 106, 109, 111, 131–168, 170– 171, 181, 186, 198, 214, 219, 225–227, 234, 249, 282–285 cccc  21, 41, 78, 80, 82, 96, 106, 131, 167–169, 198, 214, 227, 282–285; see also cause; condition; contrast/ concession CLMET  11–12 collocation(s)  9, 16–17, 26–28, 43, 61– 62, 67, 74, 79–80, 90, 108, 121–123, 125–126, 135, 163, 165, 177, 187, 192– 193, 195, 212, 215, 218, 232–237, 239– 240, 244–246 complexity, cognitive  106, 168, 174, 214, 220, 250

complexity, morphological  77, 174, 220, 250 compound  3, 92–95, 113, 154, 174, 177, 219–220, 250 concession  see contrast/concession condition  29–32, 41, 106, 131, 167– 169, 174, 282–285 conjunct  4, 25, 33–39, 52–53 conjunction  9, 17, 20, 25–28, 30, 49–53, 57–59, 61–62, 64, 66, 70, 76–77, 79, 102, 125, 140–149, 159, 187, 192, 200, 212, 234–235, 238–241, 244–245; see also adverb/conjunction, ambiguous connect  23–28, 247 connector, definition of  1–2, 23–28 connector, final position of  16, 25, 43– 44, 72, 78, 185–186, 195, 197–213, 238 connector, impure  41, 78, 82, 96, 131, 198, 200, 214, 249 connector, initial position of  16, 25, 27, 43–44, 47, 136, 195, 197–198, 238– 241, 245 connector, medial position of  9, 16, 20, 25, 27, 43–44, 72, 78, 188–189, 195, 197, 219, 233–243, 250 connector, pronominal  5–6, 21, 24, 32, 58–59, 72–75, 84, 92, 96–103, 106, 112, 132, 139, 143, 147–151, 157–159, 166, 175, 177–178, 181, 199, 218, 248–249 connector, pure  41, 45, 52, 78, 95–96, 131, 198, 214, 220 constituent order  26–28, 63, 66, 70, 78, 142–143, 213, 244–245; see also word order construction, correlative  28, 60, 62, 64–66, 74, 99, 103, 135, 139, 141–142, 145, 148, 159, 161, 167, 172, 186, 188, 191, 193, 196, 208 contrast/concession  3, 30–32, 40–41, 58, 66–67, 77, 109–110, 112, 114–116, 168–213, 225, 249, 282–285

292   Subject and name index coordination  4–6, 8–10, 21, 24–32, 76–77, 83, 85–86, 133, 136, 142–143, 163, 247–248 copia  3, 241–242 creole  5, 73–75, 156–157, 166, 249 deixis  5–6, 98, 102–103, 109, 132, 147–152, 157–158, 166–167, 178–181, 247, 249 demonstrative  5–6, 24, 58–59, 73–74, 81, 84, 97–101, 112, 140, 145, 147, 149, 156–160, 166–167, 178, 181, 183, 249 discourse marker  1, 3, 46–48, 64, 116, 182, 201, 212, 227 Early Modern English  5, 9–10, 18, 20, 49, 56–57, 77, 82, 84, 86–89, 91, 102, 109, 111, 113, 132, 178, 181, 217–218, 226, 235, 239, 248–249 enumeration  37, 40–41, 45, 77, 222– 224, 282–285 fact  21, 39, 107–108, 114–130, 228, 230–232 fiction  see text type(s) French  5, 15, 56, 73–75, 88–91, 104, 111, 137, 156, 165–166, 173–175, 183, 217–218, 221, 226, 240, 242, 249 genre(s)  see text type(s) German  2, 5, 9, 23–28, 30–32, 44, 63, 68–69, 71, 73–74, 100–101, 115, 132– 133, 135, 137, 139–140, 149–151, 164, 167, 174–175, 179, 181, 198–203, 212, 248 grammaticalization  3, 5, 20–21, 33, 47, 52, 106, 117, 119, 131, 168, 182, 184– 185, 201, 213, 250 Grice, Herbert Paul  117, 128–130 Helsinki Corpus  11–15, 46, 60, 65, 133, 144 Hume, David  243

iconicity  29, 174, 189, 219–220, 245, 250 implicature, conversational  114–116, 158, 192, 230 indicative  66, 70, 117, 165, 173 information processing  10, 28–32, 45, 132, 138–140, 166, 171–173, 198, 204– 205, 207–208, 247–249 interrogatives  79, 99, 157–158, 229–230 Kortmann, Bernd  4, 76–85, 106, 111, 154, 248 Late Modern English  9–11, 68, 72, 77, 84, 91–92, 166, 191, 215, 217, 226, 230, 236, 238–246 Latin  15, 23, 49–56, 71, 73–74, 88–92, 101, 111, 123–124, 132, 137, 149, 154, 160–164, 174, 217, 219, 221, 226, 240, 249 Lehmann, Christian  4, 22, 185 lexical bundle  see Wortverband lexicalization  9–11, 27, 33, 43, 45, 77, 80, 90–91, 96–98, 100, 102–105, 157, 159, 177, 185, 218, 221–222, 228, 230, 249 literacy  105, 109–112, 158, 249 Locke, John  1–2, 8, 20–21, 57, 214, 233, 241–243 Longman Grammar  13, 20, 35–36, 40, 42–46, 197–198, 201, 224, 238 Lowth, Robert  57, 242 main clause  9, 22, 28–29, 31, 63, 73, 135–136, 138–140, 172, 202–203, 212–213, 249 Mauranen, Anna  6–8 Nacherstposition  16, 27, 43, 56, 58, 67–72, 144, 185, 188, 191, 197, 235– 238, 240, 242 New Rhetoric  57, 233–246 news  see text type(s) particle  1–2, 23, 50, 57, 68, 99

Subject and name index   293 particle, discourse  see discourse marker particle, subordinating  5, 74, 91, 102, 142, 160, 166, 188–189 perspicuitas  3, 233, 241–246, 250 phrase, lexicalized  102–105; see also lexicalization phrase, prepositional  see prepositional phrase place  see space post-first-position  see Nacherstposition pragmatic marker  see discourse marker preposition  2, 5, 23–24, 35, 49–51, 89, 101, 154, 159, 173, 183, 228 prepositional phrase  5, 10, 22, 33, 40– 41, 43, 45, 47, 77, 80, 91–92, 94, 97–98, 102–105, 122, 127, 134, 140, 144, 147–149, 151, 154, 165, 177, 183, 185, 228, 249, 276 punctuation  125, 143, 163, 220, 233, 242–243, 250 quantifier, universal  170, 174, 181–183, 190 Raible, Wolfgang  5, 22–23, 73–75, 98, 103, 147–148, 156 result  see cause/result rhetoric  6–9, 21, 57, 78, 131, 136–137, 163–165, 212, 218, 220, 223, 226, 233–246, 250 Royal Society  3, 233, 242–243 Scottish Rhetoritians  20–21, 57, 242– 246 semantic relations  see addition; cause; contrast/concession; condition; space; time; transition Smith, Adam  233–235, 243 space  21, 38–39, 60, 64, 103, 105, 107–113, 130, 150–152, 158, 166–167, 174, 181, 219, 227, 230, 249, 282–285 Sprat, Thomas  242–243 subjunctive  63, 66, 74, 78, 99, 130, 165, 173, 188, 194

subordinate clause  27–32, 62–63, 74, 102–103, 138–139, 157, 172–173, 182, 187–188, 191, 202, 207–213, 248–249 subordination  1–6, 20–21, 25–32, 76–77, 83–85, 142–143, 173, 187–189, 204–205, 207–210, 245, 247–248 text type(s)  12–16, 18, 42–46, 105, 131– 132, 139, 200–201, 215, 219, 224–226, 229, 233, 237–238, 247, 250 text type, academic prose  42–46, 82, 92, 105, 131, 197–198, 224, 226, 238, 240, 247, 250 text type, argumentative  15, 42–46, 65, 121, 133, 219, 224 text type, conversation  42–46, 139, 197–198, 200–201, 207, 238 text type, fiction  13, 15, 42–46, 139 text type, narrative  13–15, 46, 65–66, 119, 239 text type, news  13, 42–46 time  21, 33, 38–39, 45, 60–64, 94, 105, 107–110, 130, 150–152, 158, 166–167, 249, 282–285 transition  16, 21, 40–41, 46–48, 58–59, 62, 64, 81, 95, 108–109, 112, 114, 116–117, 131, 154, 162, 214, 227–232, 282–285 Traugott, Elizabeth Closs  3, 33, 41, 46–47, 52, 117, 119, 122, 127–128, 142, 158, 227, 231–232 truth  21, 39, 47–48, 107–108, 114–130, 162, 228, 230–232, 282–285 uncertainty/doubt  228–229 univerbation  10, 33, 43, 94, 122, 149, 156, 159, 222, 230, 249, 276 word order  see also constituent order; connector, initial/final/medial position word order, verb‑final  26, 28, 30–32, 63–64, 135, 142, 202, 213, 249 word order, verb‑second  26, 32, 63, 135–136 Wortverband  105, 250