Syntactic Change in French (Oxford Studies in Diachronic and Historical Linguistics) 9780198864318, 0198864310

Citation preview

Syntactic Change in French

OX F O R D S T U D I E S I N D IAC H R O N I C AND HISTORICAL LINGUISTICS General editors Adam Ledgeway and Ian Roberts, University of Cambridge Advisory editors Cynthia L. Allen, Australian National University; Ricardo Bermúdez-Otero, University of Manchester; Theresa Biberauer, University of Cambridge; Charlotte Galves, University of Campinas; Geoff Horrocks, University of Cambridge; Paul Kiparsky, Stanford University; Anthony Kroch, University of Pennsylvania; David Lightfoot, Georgetown University; Giuseppe Longobardi, University of York; George Walkden, University of Konstanz; David Willis, University of Oxford recently published in the series 40 The History of Negation in the Languages of Europe and the Mediterranean Volume II: Patterns and Processes Anne Breitbarth, Christopher Lucas, and David Willis 41 Variation and Change in Gallo-Romance Grammar Edited by Sam Wolfe and Martin Maiden 42 Phonetic Causes of Sound Change The Palatalization and Assibilation of Obstruents Daniel Recasens 43 Syntactic Features and the Limits of Syntactic Change Edited by Jóhannes Gı´sli Jónsson and Thórhallur Eythórsson 44 Romance Object Clitics Microvariation and Linguistic Change Diego Pescarini 45 The Diachrony of Differential Object Marking in Romanian Virginia Hill and Alexandru Mardale 46 Noun-Based Constructions in the History of Portuguese and Spanish Patrı´cia Amaral and Manuel Delicado Cantero 47 Syntactic Change in French Sam Wolfe For a complete list of titles published and in preparation for the series, see pp. 323–7

Syntactic Change in French SAM WOL F E

1

3 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Sam Wolfe 2021 The moral rights of the author have been asserted Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2021937951 ISBN 978–0–19–886431–8 DOI: 10.1093/oso/9780198864318.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

Contents Series Preface Acknowledgements List of Figures and Tables List of Abbreviations

1. Introduction 1.1 Background and motivation 1.2 Aims and objectives

2. Grammatical change from Latin to French 2.1 2.2 2.3 2.4 2.5

Introduction From synthetic to analytic? The evolution of negation Word-order change Summary and conclusion

3. The left periphery 3.1 Introduction 3.1.1 The left-peripheral architecture of French and Romance 3.1.1.1 The focus field 3.1.1.2 The topic field 3.1.1.3 The frame field 3.1.1.4 Fin, Force, and complementizers 3.1.1.5 Verb movement and Verb Second 3.1.2 Summary

3.2 The Latin left periphery 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.2.6

The Indo-European foundation The focus field The frame-topic field Embedded clauses in Latin Change in subliterary and late Latin Summary

3.3 The Old French left periphery 3.3.1 3.3.2 3.3.3 3.3.4

The prefield of a V2 language The focus field The frame-topic field Particle si

ix x xii xiv

1 1 4

7 7 7 18 22 28

29 29 29 29 32 33 35 36 37 37 37 39 43 45 46 49 49 49 51 53 56

vi

contents 3.3.5 Embedded clauses in Old French 3.3.6 Syntactic change from Early Old French to Later Old French

3.4 The Middle, Renaissance, and Classical French left periphery 3.4.1 3.4.2 3.4.3 3.4.4 3.4.5

Middle French and the Verb Second prefield The frame, topic, and focus fields in Middle French Particle si in Middle French The focus field in Renaissance and Classical French The frame-topic field in Renaissance and Classical French 3.4.6 Embedded clauses in Middle, Renaissance, and Classical French 3.4.7 Summary and conclusions

3.5 The left periphery in Modern French 3.5.1 3.5.2 3.5.3 3.5.4

The focus field in Modern French The frame-topic field in Modern French The embedded left periphery in Modern French Summary

3.6 Summary and conclusions 3.6.1 Summary 3.6.2 The making of the French left periphery

4. Verb placement and verb movement 4.1 Introduction: verb movement in Romance and beyond 4.1.1 The V-to-T typology 4.1.2 The V-to-C typology 4.1.3 V-in-situ

4.2 Verb movement in Latin 4.2.1 The conservative V-in-situ system 4.2.2 The triggers for upwards reanalysis 4.2.3 V-to-C in late Latin

4.3 Verb movement in Old French 4.3.1 V-to-C movement and V2 in Old French 4.3.1.1 Arguments against C-V2 4.3.1.2 Arguments for C-V2 4.3.2 From Fin-V2 to Force-V2 in Old French 4.3.3 Marked verb placement in Old French

4.4 Verb movement in Middle French 4.4.1 V2 as V-to-Fin movement in Middle French 4.4.2 Marked verb placement in Middle French

4.5 Verb movement from Renaissance to Modern French 4.5.1 The emergence of V-to-T movement 4.5.2 V-to-C relics from Renaissance French onwards 4.5.2.1 V-to-C in Renaissance and Classical French

60 64 64 64 67 70 72 76 77 81 82 82 84 86 88 89 89 90

94 94 94 96 99 100 100 101 105 107 107 107 109 113 118 120 120 125 127 127 132 132

contents 4.5.2.2 V-to-C in Modern French

4.6 Summary and conclusions 4.6.1 4.6.2 4.6.3 4.6.4

Summary The great leap from Latin to Early Old French From V2 to SVO Conclusion

5. The subject system 5.1 Introduction: the subject system of French and Romance 5.1.1 The fine structure of the subject layer 5.1.2 Null subjects in French and Romance

5.2 The Latin subject system 5.2.1 5.2.2 5.2.3 5.2.4

Preverbal subjects Postverbal subjects Null arguments in Latin Summary

5.3 The Old French subject system 5.3.1 5.3.2 5.3.3 5.3.4

Preverbal subjects Postverbal subjects Null subjects in Old French Summary

5.4 The Middle French subject system 5.4.1 5.4.2 5.4.3 5.4.4

Preverbal subjects Postverbal subjects Null subjects in Middle French Summary

5.5 The Renaissance and Classical French subject system 5.5.1 5.5.2 5.5.3 5.5.4

Preverbal subjects Postverbal subjects Null subjects Summary

5.6 The Modern French subject system 5.6.1 Preverbal subjects and the status of pronouns 5.6.2 Postverbal subjects 5.6.3 Summary

5.7 Summary and conclusions 5.7.1 The evolution of the French subject system 5.7.2 Change in subject positions 5.7.3 French and the null-argument typology

6. OV orders and the middlefield 6.1 Introduction: head-directionality and OV orders in Romance and beyond 6.1.1 Background 6.1.2 Head-finality and the Final-Over-Final Condition

vii 134 138 138 140 143 146

148 148 148 150 152 152 155 158 160 161 161 163 170 174 175 175 180 183 186 186 186 190 192 195 195 195 199 202 202 202 205 208

213 213 213 214

viii

contents 6.1.3 The vP-periphery and the Uniformity of Phases

6.2 OV orders in Latin 6.2.1 Classical Latin as an OV language 6.2.2 Evidence for change? 6.2.3 Summary

6.3 OV orders and Stylistic Fronting in Old French 6.3.1 6.3.2 6.3.3 6.3.4

OV contexts across the clause vP scrambling Stylistic Fronting Summary

6.4 Relic OV orders from Middle to Modern French 6.5 Summary and conclusions 6.5.1 Summary 6.5.2 From OV to VO

7. A new perspective on syntactic change in French 7.1 Major findings 7.1.1 Overall summary

7.2 The history of French and parametric theory 7.3 Beyond clausal word order 7.3.1 Change in the nominal expression 7.3.2 Change in the negative system

7.4 Periodization, French, and the ‘Romance Club’

References Index

215 216 216 220 222 223 223 225 233 235 235 239 239 240

243 243 243 246 258 258 262 265

273 319

Series Preface Modern diachronic linguistics has important contacts with other subdisciplines, notably first-language acquisition, learnability theory, computational linguistics, sociolinguistics, and the traditional philological study of texts. It is now recognized in the wider field that diachronic linguistics can make a novel contribution to linguistic theory, to historical linguistics, and arguably to cognitive science more widely. This series provides a forum for work in both diachronic and historical linguistics, including work on change in grammar, sound, and meaning within and across languages; synchronic studies of languages in the past; and descriptive histories of one or more languages. It is intended to reflect and encourage the links between these subjects and fields such as those mentioned above. The goal of the series is to publish high-quality monographs and collections of papers in diachronic linguistics generally, i.e. studies focussing on change in linguistic structure, and/or change in grammars, which are also intended to make a contribution to linguistic theory, by developing and adopting a current theoretical model, by raising wider questions concerning the nature of language change, or by developing theoretical connections with other areas of linguistics and cognitive science as listed above. There is no bias towards a particular language or language family, or towards a particular theoretical framework; work in all theoretical frameworks, and work based on the descriptive tradition of language typology, as well as quantitatively based work using theoretical ideas, also feature in the series. Adam Ledgeway and Ian Roberts University of Cambridge

Acknowledgements I am deeply grateful to my friends, family, and colleagues who have supported me in many different ways in the writing of this book. From the outset, I would like to thank Julia Steer at Oxford University Press for her encouragement to write Syntactic Change in French and her work, alongside Vicki Sunter, in ensuring the finished volume is the very best it can be. Working with Julia and Vicki is always a pleasure and the reason I am delighted once again to publish with Oxford University Press. I would also like to thank Wyn Shaw for their invaluable assistance in collecting some of the Renaissance and Classical French data which feature in several chapters. Since arriving in Oxford in 2016 I have had the privilege of working with some outstanding colleagues in the Faculty of Linguistics, Philology, and Phonetics and would like to extend particular thanks for their continued support to Xavier Bach, Béatrice Rea, Sandra Paoli, Ros Temple, Ian Watson, and David Willis. Both Aditi Lahiri and Martin Maiden deserve a special mention for their continuing mentorship and advice which I value more than I can say. In my original academic home, the University of Cambridge, I would like to thank Adam Ledgeway and Ian Roberts for their words of wisdom, encouragement, and support throughout my career. St Catherine’s College is a truly special academic institution and its Fellows, Lecturers, and staff have quickly come to feel like extended family. Too many colleagues to mention have contributed directly or indirectly to the writing of this book, but I would like to thank specifically for their friendship and support Tom Adams, Ben Bollig, Bill Booth, Kersti Bo¨rjars, Andrew Dickinson, Fram Dinshaw, Andrew Elliot, Jess Goodman, Ashok Handa, Philipp Koralus, Richard Parish, Duncan Robertson, Karl Sternberg, Alex Teytelboym, Colin Thompson, and David Womersley. JC Smith has gone above and beyond to make me feel at home in the College in the last four years and was also kind enough, alongside an anonymous reader for Oxford University Press, to comment on a pre-final manuscript of this book. Needless to say, all errors that remain are entirely my own responsibility. I could not have written this book without the love and practical support of my wife Louisa, whose forbearance during a significant time in our personal

acknowledgements

xi

lives ensured the manuscript was delivered ahead of schedule. Our daughter arrived just one week after the full draft of this book was delivered to Oxford University Press. The book was thus described by one friend and colleague as my ‘first magnum opus of 2021’ and, in that spirit, I would like to dedicate the first magnum opus to the – far more important – second. To Rosa

List of Figures and Tables Figures 3.1. Left dislocation in Priestley (1955)

71

3.2. A parameter hierarchy for focus movement

91

3.3. A parameter hierarchy for topic movement

91

4.1. A parameter hierarchy for unmarked verb movement

146

5.1. Romance vs. Germanic inversion in Old French

167

6.1. OV before past participles

229

6.2. OV before infinitives and gerunds

229

6.3. A parameter hierarchy for head-finality and object-movement

242

Tables 2.1. The Latin genitive and French de compared

9

2.2. cantō ‘to sing’ in Classical Latin

9

2.3. Chanter ‘to sing’ in Modern French

9

2.4. Articles in Modern French

13

3.1. Preverbal constituents in V2 clauses

51

3.2. Preverbal constituents in embedded clauses in La Queste

62

3.3. Preverbal constituents in V2 matrix clauses in Middle French

66

3.4. Discourse status of preverbal constituents in Middle French

67

3.5. Preverbal constituents in linear-V2 clauses in Renaissance and Classical

French 3.6. Preverbal constituents in embedded V2 clauses in Middle French

72 78

3.7. Preverbal constituents in embedded V2 clauses in Renaissance and

Classical French

79

3.8. Subordinators from Latin to French

93

4.1. Schifano’s typology of verb movement in Modern Romance

96

4.2. Ledgeway’s (2017) analysis of verb placement in the Peregrinatio

105

4.3. Matrix verb placement in La Queste

109

4.4. Embedded verb placement in La Queste

112

list of figures and tables

xiii

4.5. Matrix verb placement in Middle French

121

4.6. Embedded verb placement in Middle French

123

4.7. The locus of verb movement in the history of French

140

5.1. Subject placement in the Cena Trimalchionis (Polo 2005)

153

5.2. Preverbal subjects in Old French

162

5.3. Postverbal subjects in Old French

165

5.4. Inversion types in Old French

166

5.5. The information structure of Germanic inversion

168

5.6. The information structure of Romance inversion

168

5.7. Preverbal subjects in Middle French

176

5.8. Inversion types in Middle French

180

5.9. The information structure of Germanic inversion in Middle French

181

5.10. The information structure of Romance inversion in Middle French

181

5.11. Middle French null vs. overt subjects

185

5.12. Preverbal constituents in linear-V2 clauses in Renaissance and Classical

French

187

5.13. Null vs. overt subjects in Renaissance and Classical French

192

5.14. Repetition rates for coordinated subject clitics from Culbertson (2010)

198

5.15. Subjects and subject positions in the history of French

206

5.16. Null arguments from Latin to French

211

6.1. OV vs. VO in Latin from Ledgeway (2012a: 228)

218

6.2. Rates of OV in Zaring (2010)

225

6.3. Rates of VO in Marchello-Nizia and Prévost (2020)

226

6.4. OV before past participles

227

6.5. OV before infinitives and gerunds

228

6.6. OV from Proto-Indo-European to Modern French

240

7.1. Syntactic change from Latin to French

247

7.2. Parametric typology in the history of French

252

7.3. XP movement in the history of French

253

7.4. French and the ‘Romance Club’

270

List of Abbreviations = 1/2/3 ABL ACC Adj(P) Adv(P) C(P) CFr. Cl COND D(P) DAT DEF EF EPP F Foc(P) Force(P) Fr. GDV GEN I(P) IcePaHC INF Int(P) IPFV L1 L2 LD LOC LSD M MidFr. N Neg(P) NID NOM N(P)

Cliticized to First/Second/Third Person Ablative Accusative Adjective (Phrase) Adverbial (Phrase) Complementizer (Phrase) Classical French Clitic Conditional Determiner (Phrase) Dative Definite Edge Feature Extended Projection Principle Feminine Focus (Phrase) Force (Phrase) French Gerundive Genitive Inflection (Phrase) Icelandic Parsed Historical Corpus Infinitive Interrogative (Phrase) Imperfect First Language Second Language Left Dislocation Locative Leftward Stylistic Displacement Masculine Middle French Neuter Negative (Phrase) Northern Italian Dialect Nominative Noun (Phrase)

list of abbreviations NSL O OFr. PART PASS PL PLD PLUPERF POSS P(P) PROG PST PTCP φ Q(P) REFL RenFr. RM S SBJV SID SF SG Spec Subj(P) T(P) Top(P) uF v(P) V(P) V1 V2 V3 XP

Null-Subject Language Object Old French Partitive Passive Plural Primary Linguistic Data Pluperfect Possessive Preposition(al Phrase) Progressive Past Participle Phi Quantifier (Phrase) Reflexive Renaissance French Relativized Minimality Subject Subjunctive Southern Italian Dialect Stylistic Fronting Singular Specifier Subject (Phrase) Tense (Phrase) Topic (Phrase) Uninterpretable Feature Little Verb (Phrase) Verb (Phrase) Verb First Verb Second Verb Third Phrasal Category

xv

1 Introduction 1.1 Background and motivation Few languages offer the same wealth of linguistic data as French, in both the synchronic and diachronic domains: from the earliest extant examples of its parent language, Latin, to the present day, we have an attested textual history spanning well over two millennia and French today is spoken by 77.2 million people as a first language (Eberhard, Simons, and Fennig 2020). In this book I attempt to make use of this rich array of linguistic data to better understand the evolution of French syntax and the factors contributing to French’s unique typological footprint in Romance today. Studying linguistic change in French is far from an untrodden path: numerous historical and diachronic works on the language have emerged from the late 19th century onwards, ranging from handbook treatments of particular periods and general descriptive histories1, 2 to works concerned principally with sociolinguistics or the French prescriptivist tradition (Certeau, Julia, and Revel 1975; Ayres-Bennett 1987; Ayres-Bennett 2004; Lodge 1991; Lodge 2004; Klinkenberg 2001). As arguably the most widely studied of the Romance languages, French data often have a particularly prominent role in seminal comparative works on Romance linguistics and philology (Diez 1882; Wartburg 1950; Elcock 1960; Harris 1976a; Posner 1996; Harris and Vincent 1988; Maiden, Smith, and Ledgeway 2011; Maiden, Smith, and Ledgeway 2013; Ledgeway and Maiden 2016). An equally wide-ranging literature exists in the

1 For particular periods, see—amongst very many others—Foulet (1919), Anglade (1930), Moignet (1973), Skårup (1975), and Buridant (2000) on Old French, Martin and Wilmet (1973) and MarchelloNizia (1980) on Middle French, and Brunot (1927; 1939), Gougenheim (1973), Spillebout (1985), Fournier (2007a), and Lardon and Thomine (2009) on Renaissance and Classical French. 2 For influential general histories of French, see Brunot (1905–1938), Ewert (1933), Brunot and Bruneau (1954), Wartburg (1958), Picoche and Marchello-Nizia (1989), Lodge (1992), Ayres-Bennett (1996), Posner (1997), Marchello-Nizia (1999), Rickard (2003), and Marchello-Nizia, Combettes, Prévost, and Scheer (2020).

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0001

2

introduction

synchronic domain, with sociolinguistic and regional variation the object of enquiry in an ever-growing body of work.3 Despite these facts, a diachronically oriented and theoretically informed treatment of French syntax is lacking in the literature. Brunot’s (1905–1938) magnum opus, the Histoire de la langue française, des origines à nos jours, although offering a rich array of data for the historical syntactician, is now over a century out of date, with similar comments applying to early historical works such as Darmesteter (1899), Nyrop (1928), and Wartburg (1958). Although more recent and still widely cited to this day, Harris’s (1978) explicitly comparative treatment of the evolution of French syntax is now sufficiently dated that some of its core insights have been superseded by significant changes in relevant areas of theory and developments in understanding the empirical phenomena under discussion. By far the most significant work in recent years is Marchello-Nizia, Combettes, Prévost, and Scheer’s (2020) Grande Grammaire Historique du Français, which, although unparalleled in its empirical coverage of the evolution of French, makes minimal reference to much influential formal morphosyntactic research on the language that has appeared in the last four decades. Indeed, a recurring characteristic of French linguistic studies in the last two centuries has been a frequently cursory treatment of syntax when discussing the history of the language and its present state.⁴ The current book aims to combine original empirical research with a synthesis of the vast array of existing work on some of the most significant syntactic changes to have taken place from Latin to Modern French, whilst also developing a formal analysis of French’s syntactic evolution in light of recent developments in parametric theory. This intervention is timely on a number of grounds, which I now outline. Following the very earliest formal diachronic work on French in the 1980s and 1990s, which focussed principally on the evolution of the V2 constraint and null subjects (Vanelli, Renzi, and Benincà 1986; Adams 1987c; Dupuis 1988; Vance 1988; Vance 1993; Vance 1995; Vance 1997; Roberts 1993), the decades since have seen the emergence of a plethora of generative studies on syntactic phenomena, in both the historical-diachronic and synchronic 3 As we note further on in this section, regional variation in the domain of syntax has often been under-studied. However, significant works considering both regional and sociolinguistic variation in French include Ashby (1976; 1977b; 1981a; 1982; 2001), Sankoff and Vincent (1977), Sankoff and Thibault (1977), Coveney (1996), Carruthers (1996; 1999), Jones (2001; 2011), Rowlett (1998; 2007), Martineau and Mougeon (2003), Massot (2008), Beeching, Armstrong, and Gadet (2009), Mooney (2016), Gadet and Hambye (1996; 2003), and the various contributions in Wolfe and Maiden (2020), alongside works referenced throughout this book for specific phenomena. ⁴ See Jones (2011: 505–506) for the astute observation that variationist sociolinguistics has never been prominent in France, which may in part account for the lack of research on syntax in this area.

1.1 background and motivation

3

domains.⁵ This body of work sits alongside a growing range of formal scholarship on Latin syntax (cf. for recent examples Polo 2005; Devine and Stephens 2006; Devine and Stephens 2017; Ledgeway 2012a; Ledgeway 2012b; Danckaert 2012; Danckaert 2017a; Oniga 2014). Whilst some of the studies on French are relatively narrow in terms of the phenomena under investigation, or focussed on a particular period of the language’s history, our greatly improved understanding of a number of phenomena offers an opportunity to formulate both a longitudinal analysis from Latin to the present day and an understanding of a highly significant but often overlooked issue, namely how particular phenomena interact in the language’s history.⁶ Overall, the body of formal work in this area—coupled with insights from the descriptive literature—enables us to reach a more nuanced understanding of morphosyntactic changes in specific domains than was the case even two decades ago. An increasingly nuanced understanding of the empirical landscape sits well with recent developments in formal syntactic theory in general, and the ever-growing awareness of the importance of syntactic microvariation specifically. Whilst I do not offer a full review here (see Roberts and Holmberg 2010; Biberauer 2010b; Picallo 2014; Holmberg and Roberts 2014; Roberts 2019: Chapter 1), most recent Minimalist approaches have moved away from viewing syntactic variation as constrained solely by the far-reaching macroparameters of the Government and Binding Theory era (Chomsky 1981), in favour of an account where fine-grained parametric variation is conditioned by the featural properties of individual functional heads (Borer 1984; Chomsky 1995; Chomsky 2000; Kayne 2000; Kayne 2009). In the recent approach established by Roberts (2019),⁷ a taxonomy of parametric variation is developed where the presence of the relevant feature on a single functional head will yield microparametric variation, whilst the presence of this feature on a progressively larger class of heads will yield meso- and macroparametric variation; variation at the level of a particular lexical item—by contrast—would constitute nanoparametric variation. Consider in this regard the schema in (1) (Biberauer and Roberts 2012: 268):

⁵ See References for specific phenomena throughout this book, alongside Kayne (1975; 1983a; 1983b; 1989) for some of the most prominent early generative studies of French syntax. ⁶ The most noteworthy case of two seemingly unrelated phenomena interacting is the case of V2 and null subjects, which have been understood as interrelated since Foulet (1919), with this intuition developed in much subsequent formal work (amongst others Adams 1987c; Roberts 1993; Vance 1997; Wolfe 2018a; Ingham 2018). ⁷ See also Biberauer and Roberts (2014; 2015; 2017).

4

introduction

(1) For a given value vi of a parametrically variant feature F: a. Macroparameters: all heads of the relevant type, e.g. all probes, all phase-heads, etc., share vi ; b. Mesoparameters: all heads of a given natural class, e.g. [+V] or a core functional category, share vi ; c. Microparameters: a small, lexically definable subclass of functional heads, e.g. modal auxiliaries, subject clitics, share vi ; d. Nanoparameters: one or more individual lexical items is/are specified or share vi . The important point for our purposes is that the fine-grained syntactic variation found within French varieties today and the variation observable between different historical stages of the language do not solely need to be captured in terms of dichotomies concerning the ‘presence’ or ‘absence’ of a property. Rather, diachronic and synchronic variation can be systematically modelled in terms of the featural makeup of functional heads, which I assume are themselves hierarchically ordered along the clausal spine (Rizzi 1997; Rizzi 2001; Rizzi 2006a; Rizzi 2010; Cinque 1999; Cinque 2006; Poletto 2000; Poletto 2002; Benincà and Poletto 2004; Cardinaletti 2004; Cardinaletti 2021; Cinque and Rizzi 2009). A particularly important point to note is that these features include those associated with discourse-pragmatic notions such as ‘topic’ and ‘focus’, which in some frameworks are viewed as outside syntax proper. The motivation for their inclusion, aside from purely theoretical reasons (Aboh 2010; Cruschina 2012: Chapter 1), comes from the growing acceptance of the importance of pragmatics and information structure in understanding syntactic change in Romance and elsewhere (Hinterho¨lzl and Petrova 2009; Hansen and Visconti 2009a; Zaring 2011; Larrivée 2011; Larrivée 2019; Poletto 2014; Labelle and Hirschbu¨hler 2018; Wolfe 2018b; Wolfe 2020b), and the role these factors have played in the core empirical phenomena considered in this book. Finally, we should note that the emergence of large-scale digitized corpora permits certain phenomena to be investigated with a degree of systematicity which has not always been possible. In the chapters that follow I draw on data from the Base de Français Médiéval (Guillot-Barbance, Heiden, and Lavrentiev 2017) and Frantext corpora, but also profit indirectly from secondary sources using other tagged corpora for French and Latin.

1.2 Aims and objectives Against the background set out in §1.1, this book seeks to better understand the syntactic evolution of French with reference to a number of core

1.2 aims and objectives

5

word-order phenomena; these include: the structure of the left periphery, the locus of unmarked and marked verb movement, the properties of the overt and null subject system, and the placement of objects relative to finite and non-finite verbs. Both Chapter 2 and Chapter 7 also deal—in a more limited fashion—with discussion and analysis of variation and change in other morphosyntactic domains. A major area of focus is whether phenomena which might be conceived of as showing separate trajectories of change in fact show interrelated developments. Empirically, the scope of the discussion is deliberately broad, owing to wide-ranging discussion of this nature being generally absent in the syntactic literature on French.⁸ Therefore, discussion of particular phenomena begins with analysis of the Latin data and, at times, Proto-Indo-European, and continues until the present day. When discussing Modern French, data from standard and Metropolitan French varieties are considered alongside data from non-standard and non-Metropolitan varieties. Whilst I have not aimed to be exhaustive in capturing the extent of syntactic variation in the Francophone world today, the discussion does aim to right where possible an often-noted imbalance in linguistic research on French, where non-standard data occupy a far less prominent place than is the case for a number of other Romance varieties.⁹ Although, for reasons of length, a comparative Romance analysis of the relevant developments is not offered throughout the book, each chapter begins with an outline of the broad Romance context in the relevant domain. In addition, §7.4 explicitly considers the findings of the book in terms of French’s position within a wider Romance typology. This aim is particularly significant in light of the well-established fact that French shows a number of morphosyntactic characteristics which set it apart from the majority of other Romance varieties spoken today (Diez 1882; Pei 1949: 139; Bec 1970: 9–10; Coșeriu 1988; Posner 1996: 38–39; Barra Jover 2004; Bossong 2016: 69; Smith 2020). My intention throughout is to view the relationship between synchronic and diachronic syntactic tools and the French data as bidirectional; this means asking both how insights from a range of syntactic analyses can lead to a better understanding of the French data, and also how the French data can lead us

⁸ Though see Ledgeway (2011; 2012a; in prep.) and Adams (2013) for extensive comparative discussion of the Latin to Romance transition. ⁹ The notable exception to this generalization is work by Blanche-Benveniste and colleagues (Blanche-Benveniste and Jeanjean 1986; Blanche-Benveniste 1990; Blanche-Benveniste 1997a; Blanche-Benveniste 1997b).

6

introduction

to refine certain aspects of the relevant theory. This said, more formally oriented discussion is found principally at the end of each of the main empirical Chapters (3–6) and in the overall discussion and conclusion in Chapter 7. Despite the highly controversial nature of periodization labels for particular stages in French’s history (Smith 2002; Wright 2013; Ayres-Bennett and Caron 2016; Wolfe 2020d), they are used to delimit empirical discussion in the book.1⁰ Wherever relevant I draw attention to variation within the syntax of a particular period and provide the century of textual composition for all French examples cited. Numbered linguistic examples throughout the book are from my own analysis of texts, unless indicated otherwise with a secondary reference.

1⁰ For the purposes of this book ‘Early Old French’ refers to the language of the 9th to the 12th centuries, ‘Later Old French’ to the language of the 13th century, ‘Middle French’ to the language of the 14th and 15th centuries, ‘Renaissance French’ to the language of the 16th century, and ‘Classical French’ to the language of the 17th and 18th centuries. ‘Modern French’ refers to the language spoken from 1800 onwards, with appropriate qualifiers such as ‘standard’, ‘colloquial’, or ‘written’ offered where appropriate.

2 Grammatical change from Latin to French 2.1 Introduction The aim of this chapter is to provide some basic empirical background to the more focussed discussion that forms the central concern of this book, whilst also highlighting the complex and—at times—problematic nature of studying morphosyntactic change in the history of French. We will review in turn some of the diverse approaches which have been put forward to consider the history of specific phenomena in Latin and French. Our discussion begins with the purported shift from a synthetic grammatical system to an analytic one (§2.2), before moving on to consider the body of descriptive and formal scholarship which exists on the evolution of negation (§2.3). The chapter then finishes in §2.4 with a brief outline and critique of some prominent approaches to wordorder change, which constitutes the central concern of this book.

2.2 From synthetic to analytic? Many works on historical French and Romance morphosyntax stress the role played by a shift from syntheticity to analyticity in the transition from Latin to the modern languages (Schlegel 1818: 13–18; Bourciez 1910: 22–24; Jespersen 1922: 421–423; Anderson and Rochet 1979: 35, 80; Ashby 1977a: 34–35, 86; Harris 1978: 15–16; Posner 1996: 156; Posner 1997: 120, 299, 320; Rowlett 2007: 8, 11), with Posner (1996: 156) noting that ‘Romance languages, like most other Indo-European languages, do seem to show a drift from heavily inflected synthetic word morphology towards more analytic structures with free units syntactically signalling grammatical relations’. On first inspection, this appears to capture much of the significant data: Latin nouns bear case morphology alongside gender-and number-marking (1), in contrast to Modern French where no case-marking exists on nouns (2),1 and gender and number 1 For a recent overview of the literature on case morphology in Latin and Romance, see Dragomirescu and Nicolae (2016).

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0002

8

grammatical change from latin to french

often receive no overt realization (Wartburg 1958: 255–257; Posner 1997: 331; Marchello-Nizia 1999: 83; Rowlett 2007: 16–17, 48; Smith 2016: 302; Maiden 2016: 697; Loporcaro 2018: 41; Carlier et al. 2020a: 661–663). Rather, in Modern French—as in other (Gallo-) Romance varieties2—gender and number are typically encoded on a dedicated definite or indefinite article, a functional category which was not present in Latin:3 (1)

(2)

nom.sg folium acc.sg folium gen.sg foliī dat.sg foliō abl.sg foliō voc.sg folium ‘leaf, leaves’ a. une feuille a.f leaf ‘a leaf ’ b. des feuilles some leaves ‘some leaves’

nom.pl acc.pl gen.pl dat.pl abl.pl voc.pl

folia folia foliōrum foliīs foliīs folia (Latin)

(ModFr.)

(ModFr.)

Likewise, it can be argued—as many traditional works do—that the functions performed by case morphology in Latin have now been taken over in French by a dedicated functional category, namely prepositions (3). Consider in this regard Table 2.1 from Vincent and Bo¨rjars (2010: 139) showing the functional equivalence of French de ‘lit. of, from’ and the Latin genitive. Similar generalizations can be made about the verbal system: looking at inflected verbs and auxiliaries we find richer marking of person, tense, aspect, and mood in Latin forms than is the case in Modern French, where verbal paradigms have been subjected to extensive levelling. Compare in this regard the present indicative forms in Tables 2.2 and 2.3.⁴ Perhaps more significantly from a typological point of view, the Latin to French transition has seen the emergence of a range of analytic constructions marking an array of temporal and aspectual values, with Ledgeway (2019: 348) 2 On article use in the Lorrain variety of Ranrupt see Aub-Bu¨scher (1962: 37–39), in the Picard variety of Vimeu see Vasseur (1996: 65–66), and in Wallon see Hendschel (2001). 3 Cf. Bouchard (2002: 42–43) for a formal implementation of the idea that number features are encoded exclusively on D but not N in Modern French. ⁴ For an overview of the relevant French facts, see Price (1971: 168–250), Zink (1997), and BazinTacchella (2020a) and for a comparative Romance perspective, Maiden (2018: Chapter 3).

2.2 from synthetic to analytic?

9

Table 2.1 The Latin genitive and French de compared Construction

Latin

French

Gloss

Complement of N Complement of A Complement of V

rex regnum avidus gloriae memini vivorum multi civium

le roi des rois avide de gloire je me souviens des vivants beaucoup des citoyens un homme d’une grande éloquence le palais du roi

‘the king of kings’ ‘eager for glory’ ‘I remember the living’

Partitive Quality Possessive

vir magnae eloquentiae domus regis

‘many of the citizens’ ‘a man of great eloquence’ ‘the king’s palace’

Table 2.2 cantō ‘to sing’ in Classical Latin cantō cantās cantat cantāmus cantātis cantant

Table 2.3 Chanter ‘to sing’ in Modern French chante

/ʃãt/

chantes chante chantons chantez chantent

/ʃãt/ /ʃãt/ /ʃãtõ/ /ʃãte/ /ʃãt/

noting that auxiliation is ‘[a]n area of spectacular diachronic and synchronic microvariation in Romance’; in French these include the passé composé ‘compound past’, plus-que-parfait ‘pluperfect’, surcomposé ‘double compound past’, and go-future forms (3). (3)

a. J’ai chanté I-have.1sg sing.ptcp ‘I have sung, I sang’ b. J’avais chanté I-1sg.ipfv sing.ptcp ‘I had sung’

(ModFr.)

(ModFr.)

10

grammatical change from latin to french

c. J’ai eu chanté I-have.1sg have.ptcp sing.ptcp lit. ‘I have had sung’ d. Je vais chanter I go.1sg sing.inf ‘I am going to sing, I will sing’

(ModFr.)

(ModFr.)

If we add to these examples the emergence of overt finite subordinators such as qui and que ‘that, which’, subject clitics such as je, il, and elle ‘I, he, she’ in Renaissance, Classical, and Modern French,⁵ and analytic comparative markers, derived from Latin plūs, e.g. plus grande ‘bigger’, the tempting conclusion is that a shift from syntheticity to analyticity captures such a range of data from distinct domains that it offers both a convincing description and analysis of a number of major changes in French’s history.⁶ Ewert (1933: 123)—not atypically for classic grammatical treatments of the language—thus goes so far as to state that ‘[t]he history of the French language consists largely in the abandonment of flexions in favour of particles and word-order, in the passage from a synthetic to an analytic language’. As has been highlighted elsewhere—most notably from a comparative Romance perspective, in Ledgeway (2012a: Chapter 2; 2017a)—such an account has descriptive flaws and cannot be considered genuinely explanatory. Of particular importance for a general theory of morphosyntactic change from Latin to French is the fact that the terms ‘synthetic’ and ‘analytic’ may be applicable at the level of a construction or defined groups of constructions,⁷ but arguably not at the level of a language as a whole; to attempt to do so would require us to tally the relative weight of one construction against another and—even more problematically—to establish a cross-linguistically applicable typology of what counts as the total of a language’s components.⁸ Furthermore—as highlighted by Ledgeway (2012a: 24–29)—even if we could

⁵ Though see our discussion in §5.5.2 and §5.6 on the varieties in which these are genuine subjectagreement markers. For two different views on the status of Modern French and colloquial varieties see De Cat (2005) and Culbertson (2010). ⁶ For a sympathetic though not uncritical discussion of syntheticity vs. analyticity, see Schwegler (1986). ⁷ Note that this is the approach taken by Smith (2020) in discussing morphosyntactic tendencies in a range of Romance languages. Coșeriu (1988) also discusses groups of constructions which have traditionally been discussed under the synthetic/analytic labels, but rejects the terminology in favour of an internal/external dichotomy. ⁸ These methodological concerns are almost exactly parallel to those expressed by Deutscher (2009) when criticizing approaches which seek to establish the overall simplicity or complexity of an entire language.

2.2 from synthetic to analytic?

11

systematically assess analyticity at the level of a whole language, simply stating that French or another Romance language has become more analytic than Latin does not tell us in either formal or functional terms why this is the case. The issue becomes particularly acute when considering the purported outlier status of French in a comparative Romance context (cf. Chapter 7 for more detailed discussion): we might observe as Coșeriu (1988) and Smith (2020) have done that French shows greater analyticity in certain domains than other Romance languages. Whilst this statement might be valid in descriptive terms, it does not tell us why the relevant shift towards analyticity occurred in these domains in French but not elsewhere in Romance. A binary opposition between synthetic and analytic also has the potential to mask a nuanced picture of change, where we observe—as with other grammaticalization phenomena (Roberts and Roussou 2002; Hopper and Traugott 2003; Lehmann 2015)—a cline of increasing analyticity over centuries in some cases and millennia in others, rather than an abrupt shift. Furthermore, as we will see, analytic and synthetic constructions frequently coexist in particular domains, an observation that is challenging to tally with the postulation of ‘synthetic Latin’ on the one hand and ‘analytic French’ on the other. Considering first the Latin nominal system, it is unambiguously not the case that relations within the clause are solely marked by case morphology. Rather, the evidence is well established that prepositions exist as a dedicated functional category from the very earliest texts onwards (Vincent 1999; Ledgeway 2012a: 13; Pinkster 2015a: Chapter 12), and are not in free distribution relative to clauses with bare NPs alone, as Adams (2013: Chapter 13) shows on the basis of a detailed textual analysis, concluding that ‘prepositional usages have sometimes been wrongly equated with plain cases’ (Adams 2013: 159).⁹ As such, the analytic functional category of preposition is not an exclusively French or Gallo-Romance phenomenon. Nor is synthetic case-marking via inflectional morphology an exclusively Latin phenomenon: Old French shows clear evidence for a binary case system, with a distinction between nominative and oblique case (Zink 1997: 30–31; Schøsler 2000a; Schøsler 2013; Buridant 2000; Ashdowne and Smith 2007; Smith 2011: 282). Schøsler (2000a; 2013) shows that the breakdown of this system progresses in such a way that casemarking is retained longer on determiners than nouns and is also subject to regional variation. This cautions against a binary characterization of whether case-marking is present in the grammar, in favour of a nuanced analysis of the

⁹ See also Pinkster (1990a) for discussion of change within the adpositional system of Latin.

12

grammatical change from latin to french

factors conditioning marking on specific lexical or functional categories. Furthermore, whilst Old French demonstrably does show greater syntheticity in this aspect of the nominal domain than other medieval varieties which do not retain case-marking, this observation alone does not reveal why this is the case or indeed why case-marking was lost by the 15th century.1⁰ Looking at gender-and number-marking in the nominal domain, a similar observation holds that a Latin vs. French opposition—even in descriptive terms—is insufficiently granular: although overt gender-and number-marking is not systematic in Old French,11 a greater number of nouns do show such marking than in Modern French:12 (4)

a. murs, mur, mur, murs wall.m.nom.sg wall.m.nom.pl wall.m.obl.sg wall.m.obl.pl ‘wall, walls’ (OFr.) b. fille, filles girl.f.nom/obl.sg girl.f.nom/obl.pl ‘girl, girls’ (OFr.)

Even in Classical Latin, we should underscore the fact that there is not a oneto-one mapping between inflectional morpheme and gender, number, and case, with syncretism across paradigms found from the very earliest texts and increasing as a result of a number of sound changes that took place in the classical and post-classical periods (Baldi 1999: §7.2; Herman 2000: 50; Clackson 2011: 106–109; Maiden 2011: 162; Loporcaro 2018: Chapter 2). Indeed, in support of the notion that analyticity and syntheticity should be considered relative terms (Ledgeway 2012a: 16–21), we should note that Classical Latin—despite being the benchmark synthetic language in Romance studies— is hypothesized to be markedly less synthetic in the domain of gender, number, and case morphology than its reconstructed ancestor, Proto-Indo-European (Baldi 1999; Clackson 2007: Chapter 4; Ringe 2017: 25–27). The final phenomenon to consider in the domain of nominal morphosyntax is the genesis of articles, which is held in a number of histories of the language to be one of the most significant developments in the transition from Latin to French (Nyrop 1928: 168–180; Ewert 1933: 150–155; Brunot and Bruneau 1954: 214; Wartburg 1958: 40–41; Price 1971: 115–118; Harris 1978: 67–96; Posner 1997: 381–390; Marchello-Nizia 1999: 75–77; Iliescu 2009). 1⁰ Though see Ashdowne and Smith (2007), alongside Schøsler (2013), for proposals as to which factors destabilized the system. 11 See Zink (1997: 9–21) for the full range of data. 12 Note that the final /-e/ marking a number of feminine nouns was realized in Old French, in contrast to the modern language, as was the /-s/ marking certain plurals (e.g. Moignet 1973: 16; Zink 1997: 10; Buridant 2000: 36; Rickard 2003: 48; Scheer 2020: 397–398).

2.2 from synthetic to analytic?

13

Table 2.4 Articles in Modern French

Definite Indefinite Partitive

Masculine Singular

Feminine Singular

Plural

le un du

la une de la

les – des

Once again, a syntheticity/analyticity dichotomy in this area fails to capture the full range of data. Modern French features three classes of article: definite, indefinite, and partitive, as shown in Table 2.4 (cf. also Wagner and Pinchon 1962: 87–97; Harris 1976a: 81; Posner 1997: 381–387; Smith 2016: 307–308). However, whilst the emergence of these functional heads which were not present in Latin is tempting to classify as a clear example of analyticity at play in French, with gender and number no longer encoded on the noun through inflectional morphemes, there are good reasons to be cautious about framing any account solely in these terms. Firstly, whilst it is correct that no definite or indefinite articles exist in Latin, this is not the case for D-heads in general:13 Latin, for example, features overt demonstratives, hic, ille, and iste, and in a recent analysis Giusti and Iovino (2016) present evidence that these occupy a fixed position within the extended nominal expression in the unmarked case, though they can undergo movement to either precede or follow adjectives; data such as these undermine the hypothesis that the Latin nominal expression was devoid of analytic exponents of functional categories. (5)

a. illa magna commoditas that.nom.f.sg great.nom.f.sg comfort.nom.f.sg ‘that great comfort’ (Latin, Cic. Rep. 2.9, Giusti and Iovino 2016: 231) b. illam meam cladem that.acc.f.sg my.acc.f.sg misfortune.acc.f.sg ‘That misfortune of mine’ (Latin, Cic. Sest. 31, Giusti and Iovino 2016: 231)

13 Though see Giusti (2015) for an analysis of demonstratives as phrasal categories.

14

grammatical change from latin to french

(6)

a. magnam illam laetitiam great.acc.f.sg that.acc.f.sg happiness.acc.f.sg ‘that great happiness’ (Latin, Cic. Fam. 7.2.2, Giusti and Iovino 2016: 231) b. noster ille amicus our.nom.m.sg that.nom.m.sg friend.nom.m.sg ‘that friend of ours’ (Latin, Cic. Rep. 1.30, Giusti and Iovino 2016: 231)

Turning to the early French evidence, it has been observed by a variety of scholars that both indefinite and definite articles show a markedly distinct distribution from their Modern French counterparts: the indefinite article uns—derived from the Latin unus ‘one’—is found in its earliest grammaticalized uses highlighting particular referents which are discourse-new (7) and is not typically found in cases of specific indefinite reference (cf. 8), where it would be in Modern French (Foulet 1919: 51–58; Moignet 1973: 101; Harris 1978: 80; Marchello-Nizia 1999: 88; Buridant 2000: 79; Carlier 2001; Carlier 2012: §2; Ledgeway 2012a: §4.2.1). (7)

Uns almaçurs i ad de Moriane an emir loc.cl have.3sg from Moriane ‘There is a Saracen emir from Moriane’ (OFr. Roland 73, 12th century)

(8)

Si ele a enfanté aveuque home habita if she have.3sg give-birth.ptcp with man live.3sg.pst ‘If she has given birth, she has been living with a man’ (OFr. Antioche 27, 13th century, Carlier 2012: 48)

Space constraints prevent a full discussion of the grammaticalization of the definite article here (cf. Ledgeway 2012a: §4.2.2 and Adams 2013: 482–527), but comparable distinctions between Old and Modern French to those found for the indefinite article also exist:1⁴ Ledgeway’s (2012a: 96) apt comparative observation for Medieval Romance that ‘the use of the article proves anything but systematic’ is also applicable to Old and Middle French (Foulet 1919: 47; Moignet 1973: 102–106; Harris 1978: 74–78; Posner 1997: 382–384; Buridant 2000: 102–112), where bare NPs are often found in contexts where a definite article would be present today: (9)

Francs sunt mult gentilz home! French be.3pl very noble men ‘The French are very noble men!’ (OFr. Roland 29, 12th century)

1⁴ On the development of the partitive article, see in particular Carlier (2004).

2.2 from synthetic to analytic?

15

Widespread bare NPs are not an exclusively Old French phenomenon and are found in contexts where they would not be licit in Modern French until the Classical French period (Spillebout 1985: 41–52; Gougenheim 1973: 63–65; Martin and Wilmet 1973: 111–113; Combettes 1987; Fournier 2007a: 143–151). However, in the modern language a distinctive feature of French and a number of Northern Gallo-Romance varieties (e.g. Aub-Buscher 1962: 38; Manzano 1997: 65–66) relative to other Romance languages is the widespread use of an overt D-element in nominal expressions, often as the sole exponent of gender and/or number features: (10)

a. l’ami [lami] the-friend ‘friend’ b. les amis [lezami] the friends ‘friends’

(ModFr.)

(ModFr.)

Taking into account these data from the nominal system, we observe that the picture is one of increased analyticity in certain domains, but that the relevant changes have taken place gradually and synthetic marking of gender and number has coexisted alongside marking by a class of D-elements for all of the language’s recorded history. Furthermore, it is also the case that framing the genesis of articles in terms of a synthetic/analytic split does not in and of itself account for why the articles grammaticalized as they did, nor why French and Northern Gallo-Romance varieties pattern distinctly from the rest of Romance in their use of articles. In macrotypological terms there is arguably greater continuity between the Latin and French verbal systems than is the case in the nominal domain (Herman 2000: 77–80; Ledgeway 2012a: 15).1⁵ Nevertheless, the emergence of a dedicated set of auxiliaries in the history of French (cf. 3) represents a clear point of morphosyntactic difference as compared to Latin and is an area that might seem to indicate greater analyticity.1⁶ Again, however, whilst the number of overt T-heads clearly increases in French relative to Latin (Roberts 1⁵ With reference to the medieval language, note, for example, the comment by Harris (1978: 130) that ‘what clearly emerges … is that OFr. did in fact retain a largely distinctive set of person/number suffixes’. 1⁶ For further discussion of auxiliation see in particular Gougenheim (1929), Green (1982), Harris (1982), Vincent (1982; 1987), Fleischman (1982), Pinkster (1985), and Ledgeway (2011; 2012a: Chapter 4; 2016: 767–770).

16

grammatical change from latin to french

and Roussou 2002: 35–36, 48–58; Roberts 2012b: 357–361; Ledgeway 2016: 768–770), the empirical picture is more heterogeneous than might be assumed, particularly as concerns the chronology of the developments. Considering first the compound past formed with être ‘be’ or avoir ‘have’ and the past participle, we should note that the combination of auxiliary esse ‘be’ + past participle is already attested in early Latin texts (Ledgeway 2012a: §4.3.1.4; Adams 2013: 616; Danckaert 2016), whilst incipient uses of habere ‘have’ + past participle are readily attested in Classical Latin texts (Vincent 1982; Harris 1982; Vincent 1987; Ledgeway 2012a: 132). Adams (2013: 652–673) also discusses in considerable detail the precursors to the—now synthetic— French conditional and simple future, which in Latin were analytic constructions formed from the infinitive and a conjugated form of habere ‘have’:1⁷ (11)

(12)

cantare habebat > chanterait sing.inf have.3sg.pst sing.3sg.cond ‘He/she/it would sing’

(Latin, ModFr.)

cantare habeo > sing.inf have.1sg ‘I will sing’

(Latin, ModFr.)

chanterai sing.1sg.fut

To further complicate the empirical picture, we note that the Modern French analytic future formed with a conjugated form of aller and the infinitive is a substantially later development, encoding future meaning from only the 15th or 16th century onwards (Gougenheim 1929: 96; Wilmet 1970: 183–190; Marchello-Nizia 1980: 136; Fleischman 1982): (13) Tu vas perdre ta consience // tu t’en you go.2sg lose.inf your conscience you you.cl-part.cl vas au dyable servir go.2sg to-the devil serve.inf ‘You are going to lose your conscience // you are going to serve the devil’ (MidFr. Moralité de Charité 3, 392, 15th century, Gougenheim 1929: 98) In an inverse fashion, a number of analytic periphrases not present in Latin but used throughout the Old, Middle, and Renaissance French periods, such as forms of être ‘be’ or aller ‘go’ + present participle, are not generally used from

1⁷ On the Latin facts see also—amongst many others—Pinkster (1985; 1987), Adams (1991), and Va¨a¨na¨nen (1987).

2.2 from synthetic to analytic?

17

the 17th century onwards (Gougenheim 1973: 136; Martin and Wilmet 1973: 46; Fournier 2007a: 254, 264): (14)

Les diadèmes vont sur ma tête pleuvant the diadems go.3pl on my head rain.prog ‘The crowns kept raining upon my head’ (CFr. La Fontaine, Fables 7.9, 17th century, Fournier 2007a: 262)

We see, therefore, that both a global view of syntheticity and analyticity and one focussed on verbal morphosyntax alone come up against problems when faced with the French data, which do not evince a neat transition from Latin to French in a particular time period. Parallel complications are also raised by the diachronic, diatopic, and diamesic variation observed in the use of synthetic verbal forms in French varieties. The simple past, for example, is argued to encroach on some of the functions of the imperfect and compound past in the Old French period (Foulet 1919: 156–164), before undergoing semantic specialization to refer to punctual events alone in Renaissance and Classical French, after which it is near entirely lost from spoken French by the modern period, except for certain regional varieties of the south (Désirat and Hordé 1976: 147; Wagner and Pinchon 1962: 351–353; Harris 1978: 148; Posner 1997: 317–319; Smith 2016: 305). Crucially, however, the simple past can still be used in certain—typically formal—styles and registers, as highlighted by Benveniste (1959; 1966) amongst many others. Similar remarks could also be applied to the synthetic imperfect subjunctive, which, with the exception of certain Channel Island French varieties (Jones 2000; 2001), is now an exclusively written phenomenon restricted to certain styles (Barral 1980; Lagerqvist 2009; Kragh 2010) and to the simple future, which although widespread in Modern French varieties has a more restricted distribution in several Canadian varieties (Deshaies and Laforge 1981; Poplack and Turpin 1999; Wagner and Sankoff 2011; Rebotier 2015). These three cases, alongside others, raise a basic issue of what constitutes ‘French’—in terms of both synchronic and diachronic variation and change—if using this entity as an allegedly homogeneous object of study. Furthermore, as we observed for the nominal system, describing Modern French as ‘more analytic’ in the verbal domain than Latin, Old, Middle, Renaissance, and Classical French does nothing to inform us about why certain analytic constructions were employed in the latter but not the former, or vice versa. Overall, we conclude this discussion by noting that the morphosyntactic changes discussed here without doubt figure amongst the most important in the transition from Latin to French. However, it is doubtful that framing a heterogeneous set of changes, whose outcomes are each conditioned by a number

18

grammatical change from latin to french

of distinct factors, in terms of a binary split between analyticity and syntheticity, is descriptively adequate. Nor, despite certain commonalities between the changes observed, does an alleged typological change towards analytic forms offer a satisfactory account of why the nominal and verbal systems of French have taken the form that they do today.

2.3 The evolution of negation The history of negation in French has—not uncontroversially—been modelled most prominently in terms of Jespersen’s Cycle (Jespersen 1917).1⁸ Under its original conception, Jespersen proposes that, in a number of languages, originally monopartite preverbal negation (Stage 1) is perceived by speakers to be weak and thus in need of reinforcement by an originally emphatic postverbal element. This postverbal element comes to be reanalysed as non-emphatic, yielding bipartite negation (Stage 2). Eventually, the preverbal negator becomes optional and then later is not used at all, leading to use of a sole postverbal negator (Stage 3). (15)

Stage One Neg V Stage Two Neg V Neg Stage Three V Neg

At first glance, this characterization appears to capture the basic facts for the history of French: Latin predominantly makes use of preverbal negation (16) (Devine and Stephens 2006: 183; Pinkster 2015b: 675–680; Gianollo 2018: 144), as do Old and Middle French in certain cases (17) (Foulet 1919: 235–243; Moignet 1973: 274–282; Buridant 2000: 695–724), whilst bipartite negation is considered the norm in descriptive treatments of Renaissance and Classical French and still persists in formal Modern French today (18) (Wagner and Pinchon 1962: 394–417; Désirat and Hordé 1976: 155–159; Zanuttini 1997: 15–17; Rowlett 1998; Rowlett 2007: 137–144). However, in Montreal (Sankoff and Vincent 1977), Ottawa-Hull (Poplack and St-Amand 2007), and Swiss French (Fonseca-Greber 2007), as well as spoken and informal Metropolitan French (Pohl 1975; Ashby 1976; Ashby 1981a; Ashby 2001; Coveney 1996; Rowlett 1998; Armstrong 2002; Armstrong and Smith 2002; Martineau and Mougeon 2003), rates of ne-deletion often exceed 90%, suggesting that certain French varieties are clearly at Stage Three of the schema in (15): 1⁸ For overviews of the French data, see Price (1962; 1993), Posner (1985), Ayres-Bennett (1994), Rowlett (1998), Martineau and Mougeon (2003), Ingham (2011; 2014), and Hansen (2013; 2020).

2.3 the evolution of negation

(16)

(17)

(18)

19

Nōn faciō neg do.1sg ‘I don’t do’

(Latin)

jo ne fai I neg do.1sg ‘I don’t do’

(OFr.)

Je ne fais pas I neg do.1sg neg ‘I don’t do’

(ModFr.)

However, closer examination of the relevant facts reveals a number of empirical, methodological, and theoretical challenges when the evolution is viewed purely through this lens. Turning first to the basic empirical picture, it has been noted that Jespersen’s original formulation is more accurately modelled as a four-, five-, or six-stage cycle (van der Auwera 2009: 38–44, 66; Willis, Lucas, and Breitbarth 2013: 27; Meisner, Stark, and Vo¨lker 2014), given the optionality of the post- and preverbal negators in the transitional phase between Stage 1 and 2 and Stage 2 and 3 respectively. More significantly, it is not at all obvious that a one-to-one mapping exists between particular periods within the history of Latin and French’s and the individual systems suggested by a multiple-stage model of Jespersen’s Cycle. Latin, for example, had a number of postverbal reinforcers,1⁹ which were employed negating specific clause-types, so does not constitute a ‘clean’ Stage 1 language. In a similar fashion, Old French did not feature a single preverbal negator, but three—ne, nen, and non—the second a phonologically conditioned variant of ne, with non used only in so-called presuppositional contexts, where it negates a proposition which is already active in the preceding portion of text:2⁰ (19)

Honte i avrai et reproche toz tans shame loc.cl have.1sg.fut and reproaches all times –Non avrez, sire, dist Vivïens li frans neg have.2pl sir say.3sg.pst Vivien the noble ‘I will have shame and reproaches for ever. Not so, Sir, answered the noble Vivien’ (OFr. Aliscans 204–205, 12th century, Larrivée 2011: 1989)

1⁹ For recent discussion see Pinkster (2015b: Chapter 8) and Gianollo (2018: 144–150). 2⁰ On the role of this form of activation, see Hansen and Visconti (2007) alongside discussion in Larrivée (2011).

20

grammatical change from latin to french

One of the most significant changes of all, namely the point at which nedeletion becomes systematic within the grammar of French, also poses an empirical challenge around dating, which has been extensively discussed. Whilst the first attested example of ne-deletion is found in the 13th-century text, the Roman de la Rose (20) (Price 1978: 600–601; Ayres-Bennett 1994: 66; Hansen 2020: 1686), the date at which omission of the preverbal negator can be considered a widespread phenomenon has proved quite controversial, with dates being offered from the 17th (Posner 1985; Blanche-Benveniste and Jeanjean 1986), through to the 19th (Pohl 1975; Ayres-Bennett 1994: 75–81; Martineau and Mougeon 2003: 131–132) and 20th centuries (Ashby 1981a).21, 22 (20)

a. Je sui pa beau I be.1sg neg handsome ‘I am not handsome’ (RenFr. Héroard 11.01.1608, Ayres-Bennett 1994: 72) b. Suis je pas bele dame et gente be.1sg I neg beautiful lady and noble ‘Am I not a beautiful and noble lady?’ (OFr. Roman de la Rose 5768, 13th century, Price 1978: 601)

Whilst historical studies tracking this particular development are beset with classic bad-data problems centring around the reliability of the source material and how well it accurately reflects the French language as spoken at the time (for specific discussion see Ayres-Bennett 1994), the same cannot be said of Modern French, where a truly vast range of linguistic data is available. Despite this fact, there is no absolute consensus on where Modern French sits on the schema in (15), owing to the fact that negation is subject to considerable variation, conditioned by a range of sociolinguistic and morphosyntactic factors. Whilst a tempting generalization is to characterize ne-deletion as either a spoken or an informal phenomenon, this alone fails to capture—amongst other observations—the fact that deletion rates are regionally conditioned both within and outside France (Sankoff and Vincent 1977; Coveney 1996; Auger 2003; Martineau and Mougeon 2003; Poplack and St-Amand 2007; Palasis 2015), recent results that indicate deletion is encroaching on formal or monitored speech (Ashby 2001; Armstrong 2002; Armstrong and Smith 2002), 21 In the most comprehensive quantitative study to date, Combettes, Marchello-Nizia, and Prévost (2020: 1255) lend support to the hypothesis that the most significant increase in ne-deletion occurred in the 18th century. 22 The example in (20b) is included for completeness but note that Labelle (2019)—among others— questions whether interrogative examples should be considered as direct evidence for ne-deletion.

2.3 the evolution of negation

21

and analyses that suggest ne has undergone reanalysis as an emphatic particle in the grammar of certain speakers (van Compernolle 2008; van Compernolle 2010). As such, as we noted was the case with broad-brush characterizations of analyticity and syntheticity, ascribing a particular ‘stage’ of Jespersen’s Cycle to a particular period or variety of French can at best be a grossly simplified descriptive label, which masks a highly complex picture of variation and change. We should also note that Jespersen’s Cycle—as originally formulated— does not offer a true explanation for the changes that have taken place: whilst Jespersen (1917) stresses the importance of weakening of the original preverbal negator as a triggering factor behind postverbal ‘reinforcement’ (cf. also Ewert 1933: 260–262; Dauzat 1939: 197), this account is straightforwardly falsified: whilst postverbal ‘reinforcement’ of ne is found across Old French texts (21), evidence of it becoming unstressed only appears in Middle French texts (Martineau and Mougeon 2003: 123–124). Furthermore, as highlighted by Hansen (2013: 54), such an approach also fails to account for why ne alone can negate clauses for centuries after reinforcement becomes widespread, and indeed still can do today in specific structural contexts in high-register, written French (22) (Rowlett 1998: 26–27). (21)

De sa parole ne fut mie hastifs of his speech neg be.3sg.pst neg hasty ‘He was not hasty in his speech’ (OFr. Roland 140, 12th century)

(22)

a. Pierre ne savait que faire Pierre neg know.3sg.pst what do.inf ‘Pierre didn’t know what to do’ (ModFr. Rowlett 1998: 27) b. Je n’ osais venir I neg dare.1sg.pst come.inf ‘I didn’t dare come’ (ModFr. Rowlett 1998: 27)

Before concluding we should also note that a simplistic formulation of Jespersen’s Cycle as in (15) also fails to account adequately for how postverbal reinforcers such as pas, mie, and goutte—originally ‘step’, ‘crumb’, and ‘drop’— came to grammaticalize in the first instance, and by what route they subsequently developed diachronically from the earliest French texts.23 Nor does

23 On this particular point, see Price (1962; 1993), Hansen and Visconti (2007; 2009a; 2009b; 2012), and Hansen (2013: 59, 65).

22

grammatical change from latin to french

Jespersen’s Cycle in and of itself offer an account of why certain French varieties have reached Stage 3,2⁴ with postverbal negation alone, whilst other Romance varieties have not (23): (23)

a. Noun vouòli neg want.1sg ‘I don’t want’ (Niçard, Oliviéri and Sauzet 2016: 346) b. I n è ne rvinu he neg be.3sg neg come-back.ptcp ‘He hasn’t come back’ (Wallon, Bruneau 1948: 47) c. Lo film l’ ëra pa dzen the film it.cl-be.3sg.pst neg beautiful ‘The film wasn’t good’ (Francoprovençal, Valdôtain, Zanuttini 1997: 4)

Overall, we have seen that Jespersen’s Cycle, interpreted with appropriate caution, offers some useful descriptive generalizations which capture the dominant strategies for encoding negation at different stages in the language’s history. However, under the classic formulation, we have seen that it encounters descriptive and methodological problems, as well as not offering genuine explanations of the factors driving change. Whilst offering a formal account of the evolution is not a central goal of this book, we return to some of the issues raised in this section in §7.3.2.

2.4 Word-order change Few areas of research have attracted as much scholarship or indeed as much controversy as the study of word-order change in the history of Latin and French. Whilst the sheer volume of work in this area has led to enormous progress in developing a nuanced understanding of the major changes that take place, a striking characteristic of research in this domain is the diverse range of—often incompatible—claims made about the word-order properties of particular stages of the languages. Latin, for example, is often considered to stand in stark contrast to French and the other Modern Romance languages in showing allegedly ‘free’ word order (Meillet 1903; Ku¨hner and Stegmann 1955; Ramsden 1963: 42), in contrast 2⁴ Note that this is intended to stress that the cycle lacks inherent value without clear theories of its functional and formal motivation. On formal factors conditioning Jespersen’s Cycle, see in particular van Gelderen (2008), van der Auwera (2009), Biberauer and Roberts (2011), and Breitbarth, Lucas, and Willis (2020), and discussion in §7.3.1.

2.4 word-order change

23

to the ‘fixed’ word order which now obtains (cf. 24 vs. 25). This supposedly binary distinction is often—but not always—linked to Latin’s status as an inflectionally rich language (e.g. Vennemann 1974): (24)

(25)

a. Puella puerum amat girl.nom boy.acc love.3sg b. Puella amat puerum c. Puerum amat puella d. Puerum puella amat e. Amat puerum puella f. Amat puella puerum ‘The girl loves the boy’ Le garçon aime the boy love.3sg ‘The boy loves the girl’

(Latin)

la fille the girl (ModFr.)

There are a multitude of problems with this approach: even if we examine the Latin corpus on the basis of linear ordering alone, it becomes abundantly clear that many areas of clausal and nominal word order do not—in statistical terms—show an ‘anything goes’ syntax, with an equally common preference for every possible word orders.2⁵ To choose one example, despite both prepositions and postpositions being attested in Latin texts (cf. §2.1), a clear preference for prepositions is evident from even the earliest Latin (Vincent 1999; Devine and Stephens 2006: 86, 589; Ledgeway 2012a: 193; Adams 2013: §3; Spevak 2014: 212; Pinkster 2015b: Chapter 12) (cf. also §6.2).2⁶ How, then, do we account for the numerous cases where two or more orders are robustly attested in quantitative terms? The emerging consensus—particularly prevalent in more recent theoretically informed accounts2⁷—is that specific word-order alternations are not unconstrained but rather obtain under distinct syntactic or pragmatic conditions.2⁸ This, for example, is the conclusion reached by Bauer (1995; 2009), Ledgeway (2012a: 226–234), and Adams (1977; 2016: 2⁵ For a detailed discussion of statistics pertaining to particular word-order phenomena, see in particular Danckaert (2012; 2015a; 2017a) as well as references in §3.2, §4.2, §5.2, and §6.2. 2⁶ See also Adams (2013: 257–316) for clear evidence that constructions where prepositions are used in combination with case-marked DPs are not in complementary distribution with constructions where no preposition is employed (§2.1). 2⁷ See in particular Panhuis (1982; 1984), Salvi (1991; 2000; 2004; 2005), Polo (2005), Devine and Stephens (2006; 2017; 2019), Spevak (2004; 2008; 2010; 2014), Ledgeway (2011; 2012a; 2012b; 2017b; in prep.), and Danckaert (2012; 2015a; 2015b; 2016; 2017a), among many others referred to throughout this book. 2⁸ See §1.2 for references on the growing body of work on the importance of pragmatics and information structure for understanding syntactic change.

24

grammatical change from latin to french

343–344) when considering the alternation between OV and VO order, Bolkestein (1995), Spevak (2004), and Danckaert (2017b) on the VS/SV alternation, and Giusti and Oniga (2007), Giusti and Iovino (2011), and Giusti (2019) in their accounts of word-order alternations within the extended nominal expression. A further failing of ‘free’ word-order accounts is that they fail to capture the significant diaphasic, diamesic, and diachronic variation acknowledged to exist within the surviving Latin textual corpus.2⁹ Thus, whilst quite considerable word-order variation may be found in the writings of certain authors, these should not be taken as necessarily representative of the variety of syntactic systems attested within the entire period of Latinity. A simplistic characterization of Modern French as showing a fixed word order is equally problematic, unless employed with certain caveats. Whilst it is undeniable that there is less word-order flexibility in Modern French and its varieties than high-register texts by certain authors writing Classical Latin, it is nevertheless the case—as we shall see—that both pragmatic and syntactic factors can condition word-order variation in even a simple sentence such as (26). (26)

Claude, il a Claude he have.3sg ‘Claude, he’s done a lot’

beaucoup fait a-lot do.ptcp (ModFr.)

Again—as we noted for Latin—assuming a single order of constituents also obscures the established fact that word-order variants licit in certain registers, media, and regional varieties may be marginal or flatly ungrammatical in others.3⁰ As such, we conclude that although tracking the diachronic development of word-order flexibility across the clause gives us valuable insights into the history of Latin and French, assuming a binary ‘free’ Latin vs. ‘fixed’ French split is at best inaccurate and at worst actively misleading. Those who do not advocate the view that Latin word order was genuinely free have long noted a strong preference for SOV word order (27),31 which is particularly common in embedded clauses (Linde 1923; Adams 1977: 69; Adams 2016: 321; Panhuis 1982: 117; Ostafin 1986: 165–169; Wanner 1987: 380; Polo 2005: 378; Ledgeway 2012a: 185; Danckaert 2017c: 113). 2⁹ On sociolinguistic variation in Latin, see in particular Adams (2008; 2013); on diachronic variation, see the contributions in Adams and Vincent (2016) alongside discussion in Herman (2000), Clackson and Horrocks (2007), Ledgeway (2012a), and Danckaert (2012; 2017a). 3⁰ Left dislocation fits very clearly into this category (Barnes 1985; Ashby 1988; Lambrecht 1994; De Cat 2009; Riou and Hemforth 2015)—discussed in §3.5.2—as does verb–subject inversion (Rizzi and Roberts 1989; de Bakker 1997; Sportiche 1999; Rowlett 2007: 198–225), which we turn to in §4.5.2. 31 Unless stated otherwise, throughout this book I take ‘object’ in the domain of word order to refer to a range of verbal complements including, for example, prepositional and locative objects, selected VP-adverbials, and predicative adjectives.

2.4 word-order change

(27)

25

Mancinus domum revenisset Mancinus.nom house.acc return.3sg.pluperf.sbjv ‘Mancinus had returned home’ (Latin, De Orat. 1.181, Devine and Stephens 2006: 11)

Beyond this generalization, studies of Latin word order are mired in controversy, with the status of late Latin being particularly controversial: some scholars have proposed that late Latin is already an SVO language (Harris 1976b; Harris 1978) akin to Modern French, others have proposed that late Latin features a form of V2 or VSO grammar (Salvi 2004; Salvi 2011; Clackson and Horrocks 2007; Ledgeway 2012a; Ledgeway 2017b; Wolfe 2015a), and another group have cautioned against making any systematic word-order generalizations of later texts, given the often heterogeneous picture that emerges when a sufficiently large corpus is examined (Pinkster 1991; Danckaert 2015a). The status of Old French—and other Medieval Romance languages—has been equally contentious.32 Since Thurneysen (1892), a tendency has been noted for the finite verb to appear in second position in Old French texts (28), with the hypothesis since having been developed that Old French was a Theme-V-X or V2 language,33 with widespread inversion structures and strong asymmetries between matrix and embedded clauses, and SVO order obtaining in the latter case (29).3⁴ (28)

a. Mult fu bele cele estoire et riche very be.3sg.pst beautiful that story and rich ‘That story was very beautiful and magnificent’ (OFr. Villehardouin 52, 49, 13th century) b. A cest conseil s’acorderent li conte et at this council refl.cl-agree.3pl.pst the counts and li baron the barons ‘The counts and the barons were in agreement at this council’ (OFr. Villehardouin 132, 131, 13th century)

32 For a recent overview of the Medieval Romance V2 debate, see Wolfe (2018a). Benincà (1983b; 1995; 2004; 2006) offers a comparative V2 account based on Old Gallo-Romance, Italo-Romance, and Ibero-Romance data. 33 See amongst other references given in §3.3, Adams (1987a; 1987b; 1987c; 1988), Roberts (1993), and Vance (1988; 1993; 1995; 1997). 3⁴ The former TVX label has been used most widely in descriptive/typological work and V2 in more formal work. Whilst this may seem a simple matter of nomenclature, the latter term is preferable as preverbal material in Old French is not exclusively thematic, with preverbal foci attested from the very earliest French textual examples onwards (cf. §3.3.2).

26

(29)

grammatical change from latin to french

Et fu devisé que il prendroient port and be.3sg.pst decide.ptcp that they take.3pl.pst port a Corfol at Corfu ‘And it was decided that they would disembark at Corfu’ (OFr. Villehardouin 112, 110, 13th century)

As we will see in §4.3.1, the V2 hypothesis allows us to account for a wide range of characteristics which differentiate Old French markedly from its modern counterpart. Nevertheless, non-V2 accounts of Old French word order have been developed which seek to minimize the differences between Old and Modern French by analysing both as forms of SVO system (Kaiser 2002; Rinke and Meisel 2009; Rinke and Elsig 2010; Meisel, Elsig, and Rinke 2013), principally on the basis of orders where the finite verb is not second in the linear order, which are perceived as incompatible with a formal V2 analysis of the language. Whilst not generally considered controversial, the properties of Middle, Renaissance, and Classical French constitute the biggest lacuna in word-order studies. Whilst most recent studies of Middle French have converged upon the view that it was a V2 language (Roberts 1993: §2.3; Vance 1995; Muller 2009), the status of Renaissance and Classical French is far less clear, with most accounts suggesting that French was essentially SVO from the 16th century onwards (30) (Brunot 1927: 479; Gougenheim 1973: 253–255; Spillebout 1985; Fournier 2007a: 83), but with various marked word-order alternations which are not licit even in formal Modern French, such as the null subjects we find in (30). How to formally characterize the language of this period is therefore an open question and one which is addressed throughout this book.3⁵ (30)

Les herbes estoient sans verdure, les rivieres the grass be.3pl.pst without green the rivers taries les fontaines à sec dried-up the fountains at dry ‘Grass was not green, rivers were dried up, fountains were dry’ (RenFr. Rabelais 2, 31, 16th century)

(31)

Puys retourna non à Poictiers, mais volut then return.3sg.pst neg to Poitiers but want.3sg.pst visiter les aultres universitez de France visit.inf the other universities of France ‘Next he did not return to Poitiers but wanted to visit the other universities of France’ (RenFr. Rabelais 5, 54, 16th century)

3⁵ See in particular discussion in §3.4, §4.5, §5.5, and §6.4.

2.4 word-order change

27

Most formal and descriptive work has classed Modern French as SVO, both in its own right and when compared to other Romance varieties which show considerably greater word-order flexibility.3⁶ Another approach—based principally on the widespread attestation of left dislocation and subject doubling (32)—has been to view French as a discourse-configurational language (Lambrecht 1981; De Cat 2009: 94–96),3⁷ where preservation of a topic–comment structure conditions word order to a greater extent than placing subject, verb, and object in their dedicated positions within the clausal hierarchy. (32)

Jean (il) vient Jean he come.3sg ‘Jean’s coming’

(ModFr.)

In a similar vein, Harris (1976b; 1978) maintains that a combination of grammaticalization of subject clitics as agreement markers and what he views as increasing attestation of right and left dislocation has conspired in such a way that French is—or will soon become—a VS(O) system (33). (33)

[S Jean-Paul] [S Il] [V arrive], [RD Jean-Paul] ⇒[V Il arrive] he arrive.3sg Jean-Paul he arrive.3sg Jean-Paul ‘He’s arriving, Jean-Paul’ (ModFr.) / ‘Jean-Paul’s arriving’ (Colloquial French)

Although our analysis in §5.6.1 does not ultimately support this conclusion, both Lambrecht’s and Harris’s claims show that ongoing variation and change in the word order of French varieties today mean that any analysis of word order must be appropriately nuanced and that the label ‘SVO’—although ultimately correct—is a cover term for a syntactic system permitting a number of word-order permutations. Overall, we see from the outset that no stage of the history of Latin or French is free from some form of controversy in the domain of word order, although the fact that there is considerable evidence for change in this area is beyond debate. We also note—in a similar fashion to what was observed in §2.2 and §2.3—that merely stating that the language changed from a free to a fixed order, from SOV to V2, or increasingly favours a topic–comment ordering, even if descriptively accurate, does not help account for why a change took place.

3⁶ Cf. amongst others Wagner and Pinchon (1962: 511–512), Jones (1996: 463), Rowlett (2007: 43), Cruschina and Ledgeway (2016: 556), and Smith (2016: 310–311). 3⁷ On which see Lambrecht (1981), De Cat (2005), Nadasdi (1995), and Culbertson (2010) alongside references in §5.6.1.

28

grammatical change from latin to french

2.5 Summary and conclusion In this chapter we have seen a basic survey of three major areas of change in the history of French, namely the alleged rise of analytic structures, the evolution of negation, and word-order change. Despite the enormous body of research in all three areas, some of the most frequent accounts put forward are subject to empirical, methodological, or theoretical challenges. It should be stressed that this is not to say that there is no value in the approaches surveyed in this chapter, with many offering some degree of descriptive adequacy with reference to broad tendencies in the language’s history. However, if—as suggested in Chapter 1—it is desirable to not automatically equate a particular structural property with a defined period in French’s history and to assume that broad labels such as ‘free ordering’, ‘SVO’, and ‘analytic structures’ are cover terms for morphosyntactic phenomena showing extensive microvariation, certain shortcomings do come into relief. Our discussion in Chapters 3 to 7 will focus principally on refining our understanding of the development of word order in the history of French, although we return to consider the wider consequences of the findings as they pertain to some of the phenomena discussed in §2.2 and §2.3 in §7.3.

3 The left periphery 3.1 Introduction In recent years our understanding of the left-peripheral architecture of French and other Romance languages has improved considerably.1 Originally conceived as a single functional projection—CP—work in the cartographic tradition has expanded the range of functional projections assumed to form part of the extended left periphery following the seminal work of Rizzi (1997; 2001).2 Rizzi’s (1997: 297) original schema posits multiple topic projections and a focus projection sandwiched between projections associated with force and finiteness, with each of the latter able to host complementizers (1): (1)

[Force … [Top … [Foc … [Top … [Fin … [IP …]]]]]]

However, much subsequent work by Rizzi (2004; 2006a; 2006b; 2010; 2013) and others has motivated both empirically and theoretically the postulation of a far richer array of ‘fields’, each home to a fine-grained functional structure. In this introductory section I briefly review the evidence put forward for these claims before we move on to consider how the makeup of the left periphery has changed in the history of French.

3.1.1 The left-peripheral architecture of French and Romance 3.1.1.1 The focus field Much early work on the notion of left-peripheral focus was centred exclusively on the notion of contrast and a position within the C-layer for contrastive focus in Italian is postulated by Rizzi (1997: 286) on the basis of data like those in (2) 1 For overview and discussion of recent cartographic research in this area, see Rizzi (2004), Benincà and Poletto (2004), Benincà and Munaro (2010), Ledgeway (2010a), Cruschina (2012), Haegeman (2012), and Rizzi and Bocci (2017). 2 See also Haegeman (1996) and Benincà (2001).

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0003

30

the left periphery

from Modern Italian. Such patterns are not restricted to Italian and are found across a range of Romance languages,3 including examples in the secondary literature for Modern French (Rowlett 2007: 183).⁴ Reducing all left-peripheral focalization to some form of implied contrast is still a stance taken in some of the literature (e.g. López 2009: 117–143). (2)

IL TUO LIBRO ho letto, (non il suo) the your book have.1sg read.ptcp neg the his ‘YOUR BOOK, I read (not his)’ (Modern Italian)

This said, it has been known since work by Kiss (1998) that Hungarian licenses a form of focus fronting which is not inherently contrastive.⁵ Since, a range of data from standard, non-standard, and historical Romance varieties has emerged which shows clearly that a constituent can be moved to the left periphery purely to encode that the information is new (Benincà and Poletto 2004; Benincà 2006: 62–66; Cruschina 2006; Ledgeway 2007: 124–125; Wolfe 2016a: 469).⁶ Data from Sicilian have been discussed extensively by Cruschina (2006; 2008; 2011; 2012), where a very wide range of constituents can undergo information-focus fronting. Sardinian also shows similar though not identical patterns (Jones 1993: 332–333; Mensching and Remberger 2010: §2.2; Mensching and Remberger 2016: 290–291), as do other Southern Italian dialects such as Neapolitan and Cosentino (Ledgeway 2009a; Ledgeway 2009b: 784–790). Recent work on Romanian also shows that it licenses CP information focus (Zafiu 2013: 570) and Southern Peninsular Spanish may do (Jiménez-Ferna´ndez 2015b), but to a more restricted extent (7): (3)

Un libbru ci detti a book him.cl give.1sg.pst ‘I gave him a book’ (Modern Sicilian, Cruschina 2012: 54)

(4)

Unu mariane appo vistu a fox have.1sg see.ptcp ‘I have seen a fox’ (Modern Sardinian, Jones 1993: 353)

3 See Cruschina (2012: 90–96) for review and consider his observation that ‘most Romance languages employ FF [Focus Fronting—SW] as a syntactic strategy to mark contrastive focus’ (Cruschina 2016: 605). ⁴ Though see our discussion in §3.5. ⁵ Although not originally explored in any degree of depth for Romance, Rizzi (1997: 286) notes that, in contrast to Standard Italian, ‘[o]ther languages use the clause initial focus position for noncontrastive focus’. ⁶ I do not discuss mirative focus fronting in this book. See Bianchi (2012), Jiménez-Fernández (2015a: 51–53), and Cruschina (2016: 606–607).

3.1 introduction

31

(5)

Cosa difficile è la povertà thing difficult be.3sg the poverty ‘Poverty is a difficult thing’ (Modern Neapolitan, Ledgeway 2009b: 789)

(6)

Albă e doar pe margini white be.3sg only on margins ‘[The dress] is white only on the margins’ (Modern Romanian, Pană Dindelegan 2013: 570)

(7)

Pasta esta´ comiendo pasta be.3sg eat.prog ‘She is eating pasta’ (Modern Southern Peninsular Spanish, Jiménez-Ferna´ndez 2015b: Chapter 3)

Whilst these patterns are pervasive in certain modern Romance languages, it has been argued that focus fronting was far more widespread across the medieval Romance languages (Vanelli 1986; Benincà 2004; Benincà 2006; Benincà 2013; Cruschina 2011). Wolfe (2016a: 467–474) suggests that the licensing of CP information focus is a pan-Romance characteristic in the earliest medieval texts (cf. 8) and Poletto (2006a; 2006b; 2014) has linked the extensive activation of the focus field to a range of syntactic properties found in the Old Italian CP, vP, PP, and DP (see discussion on French parallels in Chapter 7). (8) a. Mortification de carn comenset a penre tantost mortification of flesh begin.3sg.pst to take.inf early ‘She began to flagellate herself early [in the morning]’ (Old Occitan, Douceline 48, Wolfe 2017: 66) b. Buona pulcella fut Eulalia good girl be.3sg.pst Eulalia ‘Eulalia was a good girl’

(OFr. Eulalie 1, 10th century)

As we will see in more detail in §3.3.1, this is also true of Early Old French (cf. 8b), where informationally new constituents in the left periphery are strikingly frequent in their distribution in the very earliest texts despite the notable absence of information-focus fronting in the modern language (§3.5.1). Given that the left periphery in Romance can only host a single instance of either contrastive or information focus (Rizzi 1997: 290–291), it could be tempting to postulate a single focus phrase, as in Rizzi’s original schema. However, Cruschina (2012) argues persuasively against this view on the grounds that contrastive and information focus have significant distinguishing properties that suggest they are not in the same position, not least the fact that

32

the left periphery

an information focus must be strictly adjacent to the finite verb, unlike contrastive focus (Cruschina 2012: 106). In addition to these two contrastive focus and information-focus projections, we can also add a dedicated projection for Quantified Phrases, which Quer (2002) convincingly argues target a left-peripheral position and show properties distinct from other types of foci. In summary, we have seen evidence for the lowest discourse-related field within the left periphery. The projections in this field have in common that they host constituents which are low in referentiality and reach their position via internal merge.⁷ We will shortly see in §3.2–§3.5 that the exact makeup of this field has been extremely variable in the history of French. 3.1.1.2 The topic field There is a very long tradition of discussing topicalization operations in French and other languages and the postulation of a dedicated topic phrase in the left edge of the clause predates Rizzi’s early work (see for example Benincà’s 1983b, 1995 proposal for Medieval Romance). Although there is microvariation between Romance languages as to the extent to which they employ topicalization and the precise characteristics of the topicalized constituents, the picture is a more uniform one than the focus field and we will see that this also holds true of the history of French. There is, at a minimum, a binary split between two classes of left-peripheral topics in Romance. The characteristics of each class have been discussed for decades (e.g. Cinque 1990), but Benincà and Poletto (2004: §4.1) distinguish left-dislocated constituents from hanging topics based on a variety of diagnostics—including uniqueness, case-matching, and resumption properties—with the result that a hanging-topic projection is postulated higher than that for left dislocation (9). (9)

[HT Giorgio, [LD ai nostri amici, […TP … non parlo mai di Giorgio to-the our friends neg speak.1sg never of lui]]] him ‘Giorgio, to our friends, I never speak of him’ (Modern Italian, Benincà and Poletto 2004: 65)

However, this picture is complicated considerably in work by Frascarelli and Hinterho¨lzl (2007) who differentiate between structural positions for aboutness shift, contrastive, and familiar topics, on pragmatic, syntactic, and ⁷ See Benincà and Poletto’s (2004: 72) discussion of the operator properties of these elements, which strongly indicate that they are moved to/internally merged in the left periphery.

3.1 introduction

33

prosodic grounds for both Germanic and Romance languages. In recent work, Meklenborg, Helland, and Lohndal (2021) apply Benincà and Poletto’s (2004) tests to left-peripheral topics in Modern French and Norwegian and suggest that the neat two-way split cannot be straightforwardly applied. Given that this matter is far from settled and terminology is remarkably inconsistent in the literature, I make reference only where absolutely relevant in this discussion to the distinguishing properties of different types of topic but distinguish minimally between topicalization involving movement and topicalization involving external merge. Hanging topics belong straightforwardly in the latter class and have thus been argued to lexicalize the frame field (see §3.2.3), whereas I leave open the possibility that left-dislocated constituents may be subject to cross-linguistic variation as to whether they are moved or merged (Meklenborg, Helland, and Lohndal 2021; Van Kemenade and Meklenborg 2021).⁸ If, as Frascarelli and Hinterho¨lzl (2007) argue, a distinct class of familiarity topic does exist, which can either be null or overt and are derived via movement (Frascarelli 2007; Walkden 2013b; Poletto 2014; Wolfe 2016b), these would occupy the lowest portion of the topic field. 3.1.1.3 The frame field An area of major interest in the past decade has been the ‘upwards’ expansion of the C-layer into what Benincà and Poletto (2004) refer to as the frame field (Giorgi 2010; Giorgi 2012; Poletto and Zanuttini 2010: 222–224; Woods 2014; Corr 2016; Frascarelli 2017; Haegeman and Greco 2018: 35–36; Poletto and Zanuttini 2010: 222–224). Many elements located in this portion of the left periphery might elsewhere be viewed as ‘clause-external’ (e.g. Boeckx and Grohmann 2005). Broadly speaking, the projections in this field are associated with the discourse-participants, locative and temporal anchoring of those participants, and speaker-attitude towards the speech event. Some of the most discussed overt elements in this field are circumstantial or scenesetting clauses, alongside adverbials, such as the adverb duman ‘tomorrow’ which has distinct distributional properties in the Raeto-Romance variety of San Leonardo when it scopes over the entire clause rather than being focalized (Benincà and Poletto 2004: §4.2). Specifically, whilst the focalized variant can be readily embedded under certain matrix verbs, the non-focalized scenesetting variant cannot (10). This can be straightforwardly accounted for if one assumes that the frame field is not typically selected by matrix predicates in ⁸ Benincà and Poletto (2004: 72), despite developing the topic vs. focus distinction as one of external merge vs. operator-movement do leave open the possibility that topics in some languages may be moved.

34

the left periphery

contrast to the lower portions of the left-peripheral structure, which can be embedded in certain conditions: (10) a. Al m a dit c DUMAN va-al he me have.3sg tell.ptcp that tomorrow.foc go.3sg-he a Venezia to Venice b. *Al m a dit c duman va-al he me have.3sg tell.ptcp that tomorrow go.3sg-he a Venezia to Venice ‘He told me that he is going to Venice tomorrow’ (San Leonardo Raeto-Romance, Benincà and Poletto 2004: 66) Turning to clauses, Munaro (2005) argues convincingly for a detailed typology of preposed clauses and their position in various left-peripheral functional projections. It seems clear that a subclass of such clauses can lexicalize projections within the frame field when they have a scene-setting function, whereas others lexicalize lower topic projections. In fact, forming this hypothesis readily accounts for an otherwise puzzling property of Old French and Medieval Romance syntax, where certain clauses ‘count’ as the first constituent of a V2 clause, triggering inversion of the verb and subject, whereas others do not (Vance, Donaldson, and Steiner 2009; Donaldson 2012). Under Munaro’s account, we can understand the absence (11a) or presence (11b) of inversion as a reflex of the structural position the clause occupies within the left periphery, be that a low position in topic ‘visible’ to the V2 constraint or a high one in frame which is not: (11) a. Endementres qu’ il parloit en tel while that he speak.3sg.pst in such maniere, il resgarde loign en la mer manner he look.3sg far in the sea ‘While he was speaking in such a way, he looked far away across the sea’ (OFr. La Queste 112, 13th century, Donaldson 2012: 1023) b. Car se Franceis te veient entrepiez, for if Frenchmen you.cl see.3pl trample.ptcp Diront Normant en nom de reprovier … say.3pl.fut Normans in name of insult ‘For if the Frenchmen see you trampled, the Normans will say, as an insult …’ (OFr. Coronemenz Looïs, v. 197–198, 12th century, Donaldson 2012: 1032)

3.1 introduction

35

3.1.1.4 Fin, Force, and complementizers We have so far seen evidence for three distinct fields of projections in the left periphery, the contents of which are associated with particular pragmaticosemantic characteristics. In addition to these, the articulated left periphery also hosts complementizers and relativizers.⁹ In Rizzi’s original schema, the fundamental distinction is between finite complementizers such as French que and Italian che ‘that’, which lexicalize a high left-peripheral position, namely Force, and non-finite complementizers such as French de and Italian di ‘of, for’, which are the very lowest boundary of the C-layer, namely Fin: (12) [Force que, che [Top … [Foc … [Top … [Fin de, di [IP …]]]]]] Going beyond the binary Force/Fin distinction, however, is necessary to account for data such as the Old Neapolitan example in (13), where, in contrast to the better-known usage of a single complementizer in Romance, we observe multiple complementizers co-occurring with topics and foci. Whilst such patterns today are particularly prominent in Southern Italian dialects (Paoli 2003; Manzini and Savoia 2003; Manzini and Savoia 2010; Damonte 2005; Damonte 2010; Ledgeway 2005; Ledgeway 2009a; Ledgeway and Lombardi 2005; Ledgeway and Lombardi 2014; Colasanti 2017; Colasanti 2018) and Romanian (Pană Dindelegan 2013: Chapter 10), in an earlier stage of French there is considerable evidence that variants of que ‘that’ could lexicalize multiple left-peripheral heads (14) (cf. in particular Salvesen 2014). (13)

bastava che la mia Aurora, suave e benegna be-enough.3sg.pst that the my Aurora sweet and kind che era verso di me, che superba e crudele that be3sg.pst towards me that proud and cruel m’è ritornata me.cl-be.3sg return.ptcp ‘it was sufficient that my Aurora, sweet and kind that she was to me, that proud and cruel she has returned to me now’ (Old Neapolitan, Galeota LXXIX 3, Ledgeway 2005: 382)

⁹ Although I touch upon them in various places, the exact makeup of relative clauses is not a major focus of this work. On French relative clauses, see in particular Kayne (1975), Rowlett (2007: 191–193), and detailed comparative discussion in Cinque (2020).

36

the left periphery

(14) Or dit li contes que quant mes sires Gauvains now say.3sg the story that when my Lord Gawain se fu partiz de ses compaignons qu’ il refl.cl be.3sg.pst leave.ptcp of his companions that he chevaucha … ride.3sg.pst ‘The story now says that when my Lord Gawain left his companions, he rode …’ (OFr. La Queste 171d, 13th century) Note that although the postulation of two complementizer positions in Fin and Force captures much of the extant data, the pan-Romance picture also indicates that complementizers can target other head positions within the topic-focus layer (Ledgeway 2005; Villa-Garcı´a 2012a; Villa-Garcı´a 2012b; Munaro 2015). 3.1.1.5 Verb movement and Verb Second The standard analysis of the V2 property since Den Besten (1983) has been that it involves two operations: finite-verb movement to C and movement/merger of an XP in Spec-CP (Vikner 1995; Cardinaletti and Roberts 2002; Holmberg 2015). This is, with minor modifications, the analysis put forward by Adams (1987a) for Old French, which was a V2 language (cf. 15).1⁰ (15) A cel consel ot Nichodemus amis at that meeting have.3sg.pst Nichodemus friends ‘Nichodemus had friends at that meeting’ (OFr. Roman du Graal 26, 13th century, Adams 1987c: 4) As discussed in Wolfe and Woods (2020), the advent of the cartographic enterprise and Rizzian approaches to the left periphery pose some complications for this classic account of V2. Pending further discussion below in §3.3–4 and Chapter 4, I adopt the emerging view here that the Fin and Force heads can both potentially be targets for verb movement in a V2 language (Poletto 2002; Rouveret 2004; Walkden 2015; Wolfe 2016b; Haegeman and Greco 2018; Greco and Haegeman 2020), in addition to hosting complementizers. Furthermore, these heads can be endowed with an Edge Feature (EF) which triggers merger of a phrasal constituent in that head’s specifier. That is to say, in contrast to the cases introduced in §3.2.1–3, where a particular discoursepragmatic feature triggers merger of a constituent in the left periphery, in V2 1⁰ This V2 stage very likely continues into Middle French. See Chapter 4, the discussion in §3.4, and Vance (1995).

3.2 the latin left periphery

37

languages certain left-peripheral heads can bear feature-blind EFs not linked to a specific discourse-pragmatic value. This, as we will see, is a major distinction between different stages of French.

3.1.2 Summary In this section, I have sketched the outline of my assumptions concerning the left periphery. Based on comparative Romance data, we have seen evidence for a range of hierarchically ordered functional projections, which can be structured in layers, namely frame, topic, and focus. In the case of the topic and focus layers, they appear ‘sandwiched’ between Force and Fin, which can host complementizers and—in V2 languages—the finite verb.11 This schema is outlined in (16) and will be modified as necessary throughout this chapter. (16) [Frame TopicHT , AdverbialScene Setting /ClauseScene Setting [Force Complementizer, VerbFin [Topic TopicLD , TopicMoved [Focus FocusContrastive , QuantifierIndefinite ,FocusInformation [Fin Complementizer, VerbFin …]]]]]

3.2 The Latin left periphery We will see in this section that despite the difficulties inherent in dealing with the Latin corpus, we can accurately develop a detailed outline of the architecture of the Latin left periphery and how that left periphery varied over time. Overall, my proposal will be that there is a greater degree of continuity between this aspect of the syntactic system observable in Latin texts than is sometimes assumed.12

3.2.1 The Indo-European foundation Our discussion starts with a brief consideration of Proto-Indo-European. Despite these data seeming somewhat distant from the central concerns of this book, they are nevertheless important in evaluating how far French may have deviated from a common Indo-European inheritance and thus how 11 For the sake of simplicity, I exclude in the schema the possibility of complementizers lexicalizing other head positions. 12 For an analysis of the extent of continuity or discontinuity between early and late Latin, see Adams and Vincent (2016).

38

the left periphery

typologically exceptional its left periphery may be from its closest relatives. Furthermore, there is a growing acceptance that certain common morphosyntactic properties found in the Medieval Romance languages may be closely related to properties of the proto-language, in particular those involving second-position phenomena (see in particular Roberts 2012a, 2021). The central observation to make is that numerous studies of Indo-European daughter languages favour a reconstruction of a parent with a richly articulated left periphery.13 Indeed, prior to the widespread adoption of Rizzi’s (1997) articulated left periphery, a number of scholars postulated additional positions to the left of the complementizer (e.g. Hale 1987: 10–12; Garrett 1994: 31; Ferraresi 1997). As we will now see, this is necessary not only to account for the fact that multiple elements can be fronted to the left edge of the clause in a fixed order across a large number of daughter languages but also to account for the behaviour of weak so-called Wackernagel elements relative to the constituents that precede them (Wackernagel 1892). As Hale (2007: 197) notes, Delbru¨ck (1888: 16) is one of the first to highlight that fronting of constituents for reasons of emphasis is a major feature of Vedic Sanskrit syntax. This observation has since been developed in detail by Hale (1987; 2007; 2017) for Sanskrit and Indo-Iranian in general, with the result that in his most recent work he posits at least two positions before the CP in Indo-Iranian, namely a topic and focus phrase, the latter hosting wh-words, relatives, and ‘fronted’ constituents (17). (17) idhma´m. ya´s te jabha´rac chas´ramān.a´h. kindling.a.sg who.n.sg you.d.sg would-bear exerting-oneself ‘who would bear the kindling to you, exerting himself ’ (Sanskrit, RV 4.12.2a, Hale 2007: 198) What is perhaps most striking is the degree of similarity between this proposal for an articulated left-peripheral structure and the two positions proposed by Garrett for Lycian and Hittite: ‘Topic’ and ‘Front’ (Garrett 1994: 31), the latter position interpreted as corresponding to contemporary understanding of the focus position by Goedegebuure (2009: 956) and Sideltsev and Molina (2015: 18–20). Similar assumptions about an articulated C-domain are put forward by Newton (2006a) for Proto-Celtic and Walkden (2014; 2015) for ProtoGermanic.1⁴ Under one of the weakest possible interpretations, these data 13 Space constraints preclude a discussion of the feasibility of reconstruction of Proto-IndoEuropean syntax here. The stance taken is that careful consideration of the data from the daughter languages can inform a reconstruction of Proto-Indo-European syntax, in line with recent methodological work by Walkden (2013a; 2014: Chapter 1). 1⁴ For discussion of the Gothic evidence, see Ferraresi (1997; 2005).

3.2 the latin left periphery

39

challenge the idea put forward by Kiparsky (1995) that Proto-Indo-European had no CP at all and call for at least two positions before the finite verb, one for merged and another for moved constituents. A stronger and more theoretically instructive hypothesis is put forward by Roberts (2012a: 414) who proposes a left-peripheral structure as in (18).1⁵ Keeping in mind the fact that Roberts’ structure could be further enriched into fields of projections,1⁶ we conclude that early Latin may have inherited from its parent language an articulated left-peripheral structure not dissimilar to that proposed in (16). (18) Topic > Focus > Clitic/Topic > TP … Such a schema takes on considerable explanatory value as it helps us better to understand the distribution of a range of weak elements traditionally referred to as ‘Wackernagel elements’, which I discuss in more detail for Latin specifically in §3.2.2. Hale (1996) shows convincingly, for example, that in Sanskrit such elements do not blindly target the second position of the clause but are sensitive to the category of preceding constituent and whether it occupies a topic position or one lower in the functional structure, most likely focus or FinP, to use the terminology in Rizzi (1997) and Benincà and Poletto (2004). The important conclusion to draw for our purposes is that far from inheriting ‘free’ word order from its parent or originally lacking a left periphery at all (Kiparsky 1995), it is likely that Latin inherited a relatively strict syntaxpragmatics mapping with left-peripheral constituents occupying dedicated positions within an articulated functional structure. We will now see that the extant textual evidence from various stages of Latin supports this conclusion.

3.2.2 The focus field Many ‘free’ or ‘(semi)-fixed’ accounts of Latin syntax (cf. §2.3) frequently draw attention to a range of constituents which occur in clause-initial position for reasons of ‘emphasis’ and/or contrast (Marouzeau 1938: 139; Panhuis 1982; Pinkster 1990b: 35; Pinkster 1991: 70; Bauer 2009: 268–275; Spevak 2008: 124; Spevak 2010: 41); in more recent years the idea that such constituents move to a focus field, low in the left periphery, has been made explicit by a number of scholars (Salvi 2004: 50; Salvi 2005: 438–439; Salvi 2011: 356; Polo 2005: 395; 1⁵ See also Clackson (2007: 166–171) who notes a number of revealing similarities concerning leftperipheral structure assumed for a number of Indo-European daughter languages. 1⁶ See Newton (2006b) and Goedegebuure (2009) on the notion of an articulated focus field and Ferraresi (2005: 133), Fortson (2008: 160), and Ringe (2017: 80) on an articulated topic field for IndoEuropean daughter languages

40

the left periphery

Devine and Stephens 2006: 231; Devine and Stephens 2019: 48; Horrocks 2011: 132; Danckaert 2012: 122; Ledgeway 2012a: 269; Ledgeway 2018: 268; Bortolussi 2017: 104). Constituents encoding new information focus can be realized in situ in Latin (Devine and Stephens 2006: 87–88; Devine and Stephens 2019: 36; Spevak 2008: 124; Danckaert 2012: 285), most likely within a dedicated projection in the vP periphery, which we will consider further in §6.2. Consider for example the focal subject in (19), which is clearly low in the functional structure:1⁷ (19)

maiores Legantur tamen in Africam maiores natu natu send.3pl.pass nevertheless to Africa.acc of-years born.abl nobiles nobiles nobleman.nom.pl ‘Nevertheless men of years and rank were sent to Africa’ (Latin, Sall. Iug. 25.4, Spevak 2008: 118)

Whilst there is some debate about which of the two positions is the unmarked one for realization of new-information focus, there is a general consensus that new-information constituents can also occur preverbally (Devine and Stephens 2006: 231; Devine and Stephens 2019: 48; Spevak 2010: 41) (20, 21). As in the contemporary Romance data we saw in §3.2.1 these constituents are not resumed by a pronominal form, in keeping with the fact that they reach their position in the left periphery via movement: (20) … Gaiumque Gallonium, equitem Romanum familiarem Gaius.acc-and Gallonius.acc equestrian Roman.acc relative.acc Domiti … oppido Gadibus praefecit … of-Domitius town.dat Gades.dat put-in-charge.3sg.pst ‘and he put in charge of the town of Gades Gaius Gallonius, a Roman equestrian, a relative of Domitius’ (Latin, Caes. B.C. 2.18.2, Spevak 2010: 41) (21) Grata mihi vehementer est memoria nostri pleasant me.dat extremely be.3sg memory.nom we.gen tua quam significasti litteris your.nom which.acc indicate.2sg.pst letter.abl ‘Your memory of us as indicated by your letter gives me extreme pleasure’ (Latin, Cic. Fam. 7.17.1, Salvi 2005: 439) 1⁷ See also de Jong (1989) for more examples of this type.

3.2 the latin left periphery

41

In addition to new-information focus, we can also identify a dedicated position for contrastive focus in Latin (Adams 1994a: 47; Spevak 2010: 46; Devine and Stephens 2019: 58–60). Witness the contrast between the unmarked subject-object-verb order of (22) with the marked order of (23), where Ledgeway (2012b: 434) argues that the object-subject-goal-verb order is a result of contrastive focalization of copias suas ‘his troops’: (22) Caesar suas copias in proximum collem Caesar.nom his.acc troops.acc in next.acc hill.acc subducit withdraw.3sg ‘Caesar leads off his forces to the next hill’ (Latin, Caes. B.G. 1.22.3) (23) copias suas Caesar in proximum collem troops.acc his.acc Caesar.nom in next.acc hill.acc subduxit equitatumque, qui sustineret withdraw.3sg.pst cavalry.acc-and who.nom sustain.3sg.pst.sbjv hostium impetum misit enemies.gen attack.acc send.3sg.pst ‘Caesar drew off his forces to the next hill, (and sent the cavalry to sustain the attack of the enemy’ (Latin, Caes. B.G. 1.24.1, Ledgeway 2012b: 434) We saw previously that QPs show idiosyncratic behaviour in the modern Romance languages, which suggests that, when they are fronted, they occupy a dedicated position low in the focus field. Although their distribution has not been extensively discussed for Latin from a formal perspective, cases such as (24, 25) are found where they occupy a clause-initial position (Salvi 2005: 439; Devine and Stephens 2006: 259, 499): (24) nihil te omnino fefellit nothing.nom you.acc absolutely escape.3sg.pst ‘Nothing whatever escaped your notice’ (Latin, Cic. Fam. 9.2.2, Salvi 2005: 439) (25) Multae et bonae et firmae sunt legiones Lepidi Many and good and strong be.3pl legion.nom.pl Lepidus.nom.pl et Asini and Asinus.nom.pl ‘The legions of Lepidus and Asinius are many, good and strong’ (Latin, Cic. Fam. 11.9.1, Devine and Stephens 2006: 259)

42

the left periphery

Danckaert (2015b: 269) also argues that bare quantifiers like nihil ‘nothing’, omnia ‘everything’, and nemo ‘nobody’ are able to undergo focalization to the embedded left periphery in Latin. Under Ross’s (1973) Penthouse Principle, that there are always as many movement operations licensed ‘upstairs’ in matrix clauses as ‘downstairs’ in embedded clauses, we can interpret this as an additional piece of evidence that Latin permitted QP-fronting into the focus field. A final piece of evidence for an active focus field comes from Wackernagel clitics. A range of elements including unstressed pronouns and forms of copula and auxiliary esse ‘be’ are noted by Wackernagel (1892) to occupy second position of the clause in Latin and a range of Indo-European languages (cf. §3.2.1). However, there are a number of exceptions to this generalization noted by Marouzeau (1938) and Wanner (1987: 67–85) amongst many others. Adams (1994a; 1994b), offering the most exhaustive discussion to date of the Latin data, proposes that in Latin Wackernagel clitics are consistently adjoined to a ‘focal host’ (Adams 1994b: 112),1⁸ noting that deviations from strict secondposition ordering are thus accounted for as ‘often a focused term comes later than first word’. (26) non PUBLICO me praesidio, sed pruiata not public.abl me.acc protection.abl but private.abl diligentia defendi diligence.abl defend.1sg.pst ‘I defended myself not by PUBLIC protection but by private diligence’ (Latin, Cic. Cat. 1.11, Ledgeway 2012a: 192) The reasons for this are obvious if we consider the left-peripheral schema exemplified in (16): foci can readily be preceded by a range of constituents in the frame and topic fields. To summarize, evidence is readily available that Latin had an active focus field, able to host contrastively focussed constituents as well as those encoding new information and QPs.

1⁸ This class includes demonstratives, deictics, adjectives of quantity and size, intensifier adverbials, temporal adverbials, and negation (Adams 1994b: 122–130). Note that the fact that QPs headed by omnis ‘all’ and maximus ‘great’ can attract clitics to the Wackernagel position further supports the suggestion that the Latin left periphery hosted a phrase for QPs in the focus field.

3.2 the latin left periphery

43

3.2.3 The frame-topic field As noted in §3.1, diagnosing distinct types of topic within the left periphery can be both methodologically challenging and controversial, something Bortolussi (2017: 108, 120) explicitly acknowledges for Latin.1⁹ Given these facts, I limit myself here to highlighting that Latin clearly makes widespread use of topicalization and, furthermore, that a subclass of such operations may involve base-generated ‘dislocated’ topics which occupy a higher position within the topic-frame field. The possibility of displacing a thematic or topical constituent from its base-generated position within the clausal core is widely discussed in the literature (Pinkster 1990b: 37; Pinkster 1991: 75–76; Polo 2005: 389; Salvi 2005: 445; Salvi 2011: 357–358; Devine and Stephens 2006: 109–110; Devine and Stephens 2019: 143; Spevak 2010: 29; Horrocks 2011: 132; Danckaert 2012: 24; Ledgeway 2012a: 269; Ledgeway 2018: 265; Halla-Aho 2016; Bortolussi 2017), and can be subsumed under the operation Ledgeway (2012a: §5.4.3; 2018) terms Edge-Fronting. Consider cases such as (27) involving a topicalized object alongside (28) which features a topicalized prepositional complement, where we note in both cases that no resumptive pronoun is used within the clausal core2⁰: (27) Quintum fratrem cotidie expectamus quintus.acc brother.acc every day expect.1pl ‘We expect brother Quintus back any day’ (Latin, Cic. Att. 1.5.8, Pinkster 1991: 75) (28) De Aufidiano nomine nihil te hortor about Aufidius.abl debt.abl nothing.acc you.acc urge.1sg ‘In the matter of Aufidius’s debt, I put no pressure on you’ (Latin, Cic. Fam. 16.19, Salvi 2005: 437) However, we also find cases in Latin which are typically analysed as having a looser connection with the clausal core. Firstly, left-dislocated constructions with resumptive pronouns are attested (Polo 2005: 388; Salvi 2005: 431; Spevak 2010: Chapter 3; Halla-Aho 2016: 368; Bortolussi 2017: 105, 109) which do not correspond exactly to the Romance cases discussed by Benincà and Poletto

1⁹ Note also that the resumption tests used for Modern Italian in Benincà and Poletto (2004) do not readily transfer to a language making widespread use of zero-anaphora (cf. Bolkestein 2000). 2⁰ Pinkster (1991: 75) notes that Quintus would have been familiar to Cicero’s correspondent.

44

the left periphery

(2004) amongst others but are similar to the Romance type.21 Although—as we will see in §3.5—viewing dislocations as a purely colloquial phenomenon in Modern French is inaccurate, they are sometimes associated with lower registers of Latin (Bortolussi 2017: 105):22 (29) Amicos domini, eos habeat sibi friends.acc master.gen them.acc consider.3sg.sbjv he.dat amicos friends.acc ‘The master’s friends, he must consider them his own friends’ (Latin, Cato Agr. 5.3, Bortolussi 2017: 109) In addition to dislocations with overt pronominal resumption, note that we also find cases of the so-called nominativus pendens or hanging topic, which is widespread in Latin (Hofmann and Szantyr 1965: 29–30; Danckaert 2012: 24–25; Halla-Aho 2016; Devine and Stephens 2019: 145). As in Romance, these lack case-matching and appear non-integrated relative to the clausal core. Recall from our discussion in §3.2.1 that constituents such as these would be located in the frame field. Aside from hanging topics, is there additional evidence for a frame field in Latin? Although many of the tests that can be applied in living languages do not fare well with Latin data, certain temporal adverbials and circumstantial expressions frequently precede all other constituents, be they left-peripheral or in the clausal core (Mo¨bitz 1924: 120–121; Panhuis 1982: 44; Bauer 1995: 95; Spevak 2010: 66; Salvi 2011: 357). Consider in this respect the placement of igitur ‘thus’ in (30): (30) Igitur in eo colle … Iugurtha consedit thus on this hill Jugurtha.nom set-down.3sg.pst ‘So, on this hill … Jugartha took up position’ (Latin, Sall. Jug. 48.3–49.1, Spevak 2010: 66)

21 See Salvi (2005: 431) in particular on the danger of viewing LDs in Latin as the ‘same’ phenomenon as that found in Old French. As we will see in §4.3–5, it is also not self-evident that the Old French dislocation constructions show the same properties as those observable in Modern French. 22 But see Halla-Aho (2016: 373) for the view that ‘[l]eft-detached constructions, though rare, are not excluded in classical Latin’.

3.2 the latin left periphery

45

3.2.4 Embedded clauses in Latin In contrast to the considerable freedom observable in matrix clauses, owing to the accessibility of the frame, topic, and focus layers for hosting a range of leftperipheral constituents, embedded clauses show a remarkably uniform SOV ordering (Linde 1923; Adams 1977: 69; Adams 2016: 321; Panhuis 1982: 117; Ostafin 1986: 165–169; Wanner 1987: 380; Polo 2005: 378). Ledgeway (2012a: 185) for example states that ‘while the flexibility of classical Latin word order is, as we have observed on several occasions in the preceding chapters, widely recognized and reported in the literature, it is less frequently remarked that such flexibility typifies root clauses. The word order of embedded clauses, by contrast, is much less flexible (pace Salvi 2004: 43), generally following a very rigid SOV arrangement’. Pending further discussion in Chapters 4–6, this can be readily captured if we assume that the finite verb and its arguments typically stay within the clausal core rather than undergoing movement into the left periphery (31): (31) [Frame [Force [Topic [Focus [Fin Complementizer/Relativizer [TP … S…O…V…O] In fact, Danckaert (2012: 107) argues convincingly that the subordinating conjunctions cum, si, and ut occupy only the very lowest left-peripheral head position, Fin, within his Latin corpus. If this is correct,23 Classical Latin could be showing a similar embedded syntax to V2 languages such as Modern German and Dutch, where the presence of a complementizer or relativizer within the left periphery is well known to block embedded topicalization and focalization (Den Besten 1983; Koster 1975; Vikner 1995: Chapter 1; Holmberg 2015; Wolfe and Woods 2020). Danckaert (2012: Chapters 5–6) additionally shows that when embedded fronting of constituents does obtain in his corpus of adverbial clauses, the relevant constituent(s), in this case eum ‘that’, precede(s) the C-element, here cum, in an operation he refers to as Left Edge-Fronting: (32) Eum cum uidero, Arpinum pergam that.acc when see.fut.1sg Arpinum.acc proceed.fut.1sg ‘When I have seen him, I’ll move on to Arpinum’ (Latin, Cic. Att. 9.15.1, Danckaert 2012: 108)

23 However, see Salvi (2005: 447) for some cases of embedded topicalization, which might suggest that Classical Latin C-elements can exceptionally lexicalize Force.

46

the left periphery

The conclusion is therefore that Latin embedded clauses featured a rich embedded left periphery but that this left periphery was not typically rendered accessible to the same operations that were displacing constituents in matrix clauses.

3.2.5 Change in subliterary and late Latin On initial examination, the transition from the underlying SOV grammar of Classical Latin to the V2 syntax of Old French and Medieval Romance (cf. Adams 1987c; Roberts 1993; Vance 1995; Vance 1997; Benincà 2004; Wolfe 2018a: Chapter 3, and Chapters 2 and 4) may seem abrupt and marked by syntactic discontinuity. However, there is a growing body of evidence that subliterary and late Latin texts embody an important transitional stage between an SOV syntax and a full V2 grammar.2⁴ Some of the relevant changes attested in the texts will be discussed in more detail in Chapters 4–6, but note for now that they involve the expansion of V-to-C movement (Salvi 2004: 47–63; Salvi 2005: 511; Ledgeway 2012a: 150– 153; Wolfe 2015a; Wolfe 2018a: 33, 143–144), an increase in VO orders (Adams 1977: 67–75; Bauer 1995: 99–102; Ledgeway 2012a: 228–229), particularly in matrix clauses, and an increasing tendency for the left-peripheral topicalization and focalization of the type we surveyed in §§3.2.2–3 above to become obligatory, as in a V2 language. We focus on the factors that may have led to reanalysis of pragmatically or syntactically marked verb movement as unmarked V-to-C movement in Chapter 4. Here, we can instead consider the second ‘ingredient’ of a V2 grammar: the postulation of an EF on Fin or Force. What would have triggered reanalysis of marked XP-fronting to become increasingly unmarked? Classical Latin structures where a moved or merged non-subject constituent appears adjacent to the finite verb are of undoubted importance for the Latinto-French transition and are viewed by Ledgeway (2012a: 153) as ‘an unmistakable precursor to the full-fledged V2 syntax of Late Latin/early Romance’ (see also Wolfe 2015a). These accidental V2 constructions within an otherwise SOV grammar involve several types of fronted constituents including adverbials, subordinate clauses, ablative absolutives, and negation (Mo¨bitz 1924: 120–122; Bauer 1995: 95–98; Bauer 2009: 275–276; Polo 2005: 400–401). 2⁴ The idea that a V2 or, in other frameworks, a T(heme)-V-X stage might be a frequent transitional grammar from SOV to SVO, is quite common in the typological literature (Vennemann 1974; Harris 1978: 18–22; Harris 1984; Fife 1993).

3.2 the latin left periphery

47

What is perhaps most significant is that incipient V2 cases such as (33) are acknowledged to be most common in authors writing in lower registers or in texts composed late in the period of Latinity (Wanner 1987: 482; Salvi 2004: 102–111; Clackson and Horrocks 2007: 291; Bauer 2009: 276): (33) Quod si resilierit, destinavi illum that if be-restless.3sg.prf.sbjv decide.1sg.pst him.acc artificium docere trade.acc teach.inf ‘if he is restless, I have determined that he will learn a trade’ (Latin, Petron. Satyricon 46, 7) Also significant, and frequently invoked as a precursor to V2 in Germanic languages (Hock 1982; Hock 1986: 330–340; Eythorsson 1995: 175; Faarlund 2010: 207), are the cases reviewed in §3.2.2, where an auxiliary or form of copula esse ‘be’ is immediately preceded by a focal host (Adams 1994a; Adams 1994b). Under this account a clause such as (34) would be reanalysed as having the structure indicated, with a fronted focus and V2-conditioned verb movement to Fin or focus, rather than the adjacency between the XP and verb being due to a phonosyntactic cliticization process: (34) a. Is est hodie locus saeptus that.nom= be.3sg today place.nom guard.ptcp.nom ‘That site is guarded to this day’ (Latin, Cic. Diu. 2.84, Ledgeway 2012a: 256) b. [Frame [Force [Topic [Focus is [FinP [Fin est ][TP … hodie…]]]]]] Taken together, these two pieces of evidence were undoubtedly amongst the most significant, leading acquirers to postulate the second V2 ingredient of an EF on a C-related head. It is important to note, however, that numerous cases of topicalization and focalization in Classical Latin do not involve adjacency between the verb and the merged constituent (cf. 24, 26–28), with Salvi (2005: 441) stating specifically that ‘anteposition of the verb never co-occurs with the anteposition of a focussed constituent’. Given the widespread attestation of V-late structures in Latin of a wide range of registers and periods, Wolfe (2016a: 486–487) specifically suggests that such clauses with multiple constituents before the finite verb, in particular SOV clauses, were reanalysed as (frame)-topic-focus-verb, with the finite verb in Fin rather than v (cf. Chapter 4). The outcome of such a reanalysis is a late Latin left periphery where frame-setters, topics, and foci readily co-occur, with both Clackson and Horrocks (2007: 291–292) and Ledgeway (2017b: 197–199) noting that such

48

the left periphery

orders, where the finite verb appears in third or greater position, are characteristic of late Latin prose, and in particular the syntactic system attested in the Peregrinatio, a 4th–6th-century Christian Latin text: (35) a. Omnis populus usque ad unum all.nom people.nom down to one.acc cum ymnis ducent episcopum in Syon with hymns.abl lead.3pl bishop.acc in Syon ‘all the people down to the very last one, lead the bishop down to Syon with hymns’ (Latin, Peregrinatio 43, 2) b. Quod cum ceperit legi, tantus that when will-have-begun read.inf.pass so-much.nom rugitus et mugitus fit omnium moaning.nom and groaning.nom do.pass.3sg all.gen.pl hominum men.gen ‘And when the reading is begun, there is so great a moaning and groaning among all’ (Latin, Peregrinatio 24, 10, Ledgeway 2017b: 174) Although Ledgeway (2017b) does not go so far as to consider constituent fronting obligatory in the Peregrinatio, cases where one or more constituents precede the finite verb in Fin account for 84.1% of his matrix-clause sample (Ledgeway 2017b: 169). Overall, it appears clear that in late Latin the richly layered left-peripheral structure found in Classical Latin persists, but that previously optional constituent fronting is becoming increasingly obligatory, leading to the development of a form of V2 grammar. Adams (1977; 2016: 321, 343, 440, 651) draws attention to the syntactic structure of embedded clauses in informal or late texts, which, as we saw above for Classical Latin (cf. §3.2.4), is often quite distinct from matrix clauses. This observation extends to Ledgeway’s (2017b: 191–199) analysis of embedded clauses in the Peregrinatio. He shows that embedded SOV is extremely rare in his corpus but that V1 (36) and V2 (37) clauses dominate to an extent not typically observed for Classical Latin: (36) uigilatur in Anastase ut legat watch.pass.3sg in Anastasis.abl so-that read.3sg.sbjv episcopus locum illum eyangelii bishop.nom place.acc that.acc gospel.gen ‘vigil is kept in the Anastasis, and the bishop reads the passage from the Gospel’ (Latin, Peregrinatio 41.1, Ledgeway 2017b: 193)

3.3 the old french left periphery

49

(37) et dicentibus ei aliis apostolis, and say.prs.ptcp.abl.pl him.dat other.abl.pl apostles.abl quia Dominum uidissent that lord.acc see.3pl.pst.pluperf.sbjv ‘and the other Apostles told him that they had seen the Lord’ (Latin, Peregrinatio 39.5, Ledgeway 2017b: 196) Indeed, in contrast to Classical Latin where embedded topicalization (Ostafin 1986: 165–169), and focalization in particular (Salvi 2004: 53), are rare, he suggests that the embedded left periphery can be accessible for constituent fronting as well as finite-verb movement to derive the orders in (36–37) (cf. on embedded verb movement §4.2). As we shall see, although this does indicate a major point of discontinuity with Classical Latin, the finding about a relatively late Latin text is unsurprising in the light of the properties of the embedded left periphery in Early Old French (see Labelle 2007 and §3.3).

3.2.6 Summary Overall, we have seen evidence that Latin of all periods makes extensive use of the articulated frame-topic-focus fields for pragmatically salient constituents, which are frequently moved to the topic-focus positions from their basegenerated positions. A brief review of the Indo-European evidence suggests that an articulated left-peripheral structure may have been inherited from Proto-Indo-European. Embedded clauses, in sharp contrast to matrix clauses, are far more restricted in terms of embedded topicalization and focalization possibilities, though we have seen evidence that this may change in late Latin texts where the embedded left periphery is accessible in certain clause-types.

3.3 The Old French left periphery 3.3.1 The prefield of a V2 language Essential background to our discussion of the Old French left periphery is the established claim that Old French was a V2 language (Thurneysen 1892; Foulet 1919; Adams 1987a; Roberts 1993; Vance 1995; Vance 1997; Benincà 1995; Benincà 2004; Benincà 2006; Salvesen 2013; Wolfe 2018a). Rather than its being an innovation in Old French, we saw in §3.2.5 that an incipient form of the V2 property may have been present in late Latin (Salvi 2004; Ledgeway 2017b),

50

the left periphery

where a requirement held that the finite verb raise to the C-layer, accompanied by optional constituent-fronting, which was not yet a systematic requirement. We discuss the controversial status of Old and Middle French V2 in §4.3, but note here that the V2 hypothesis makes very specific predictions about the structure of the left periphery: in an SVO language with V-to-T movement such as Modern French (cf. §4.4.1), constituents preceding the finite verb are typically subject expressions within the inflectional layer, occupying Spec-TP or Spec-SubjP (Cardinaletti 1997; Cardinaletti 2004; Guasti and Rizzi 2002). This is not the case in the paradigm instance of a V2 grammar, where the prefield can host a range of constituents from various grammatical categories and is not restricted to subjects (Diesing 1990: 44; Lightfoot 1995: 40; Vikner 1995: 41; Holmberg and Platzack 1995: 71; Poletto 2002: 222–223; Mu¨ller 2004: Chapter 2; Westergaard 2008: 1843; Benincà 2013; Cognola 2013: Chapter 1; Holmberg 2015: 1; Wolfe 2019: 20–23; Wolfe and Woods 2020). Scholars from a wide range of perspectives have noted that this characterization of an ‘unrestricted’ prefield clearly describes Medieval French (Foulet 1919: 306–332; Wartburg 1958: 103; Moignet 1973: 357; Skårup 1975: 9–69; Skårup 1975: 9– 69; Vanelli 1987: §.4.1; Adams 1987c: 4–5; Adams 1988; Roberts 1993: 85–87; Hulk and van Kemenade 1995: 235–236; Vance 1997: 43–47; Donaldson 2012: 1025; Mathieu 2012: 327; Salvesen 2013: 135–136; Steiner 2014: xviii; Wolfe 2018a: 67–72; Ledgeway Forthcoming). As we see in (38–42), a range of different constituents can occur in the left periphery, the discourse values of which we explore below in §3.3.2–4. (38) Un faldestoed i unt a chair loc.cl have.3pl ‘They have a (folding) chair’ (39) La cité prist par traïson the city take.3sg.pst by treason ‘He took the city through treason’

(OFr. Roland 115, 12th century)

(OFr. Eneas 5, 12th century)

(40) Par Petit Pont sont en Paris entré by Petit Pont be.3pl in Paris come.ptcp ‘They entered Paris by the Petit Pont’ (OFr. Nîmes 11, 27, 12th century, Roberts 1993: 95) (41) et ausi fist chascuns des autres and also do.3sg.pst each of-the others ‘And each of the others did the same’ (OFr. La Queste 344, 35–36, 13th century)

3.3 the old french left periphery

51

(42) A ches paroles s’en parti li dux to these words refl.cl-part.cl leave.3sg.pst the duke ‘After these words [were said], the duke left’ (OFr. Clari 60, 13th century) We note for now that the fact that the preverbal field is not a specialized subject position can be demonstrated quantitatively, and has been in a number of studies. In Table 3.1, the proportion of subjects vs. non-subjects in V2 clauses is compared from several studies, although the methodologies may not be exactly comparable. The proportion of initial objects, if provided from the figures on non-subjects, is also given. Table 3.1 Preverbal constituents in V2 clauses Study

SVO

XP-V-(S)

O-V-(S)

Total

Roberts (1993: 95). Roland. 12th-century verse. Roberts (1993: 95). Nîmes. 12th-century verse. Vance (1997). La Queste. 13th-century prose. Steiner (2014: 137). Mixed Corpus. 13th-century prose. Steiner (2014: 137). Mixed Corpus. 14th-century prose. Wolfe (2018a). La Queste. 13th-century prose. Klaevik-Pettersen (2018: 93). Tristan. 13th-century prose. Klævik-Pettersen (2018: 93). Eustace. 13th-century prose. Ledgeway (Forthcoming). Histoire Ancienne. 13th-century prose.

31

34.1%

60

65.9%

20

22%

91

100%

23

26.4%

64

73.6%

16

18.4%

87

100%

799

46.7%

939

54.3%

Not Specified

1738

100%

204

67.3%

99

32.7%

Not Specified

303

100%

187

73.6%

67

26.4%

Not Specified

254

100%

220

46.3%

255

53.7%

61

12.8%

475

100%

372

48.6%

393

51.4%

61

8.0%

765

100%

182

48.4%

194

51.6%

34

9.0%

376

100%

165

38.5%

264

61.5%

83

19.3%

429

100%

3.3.2 The focus field Information-focus fronting is considered a central characteristic of Medieval Romance word order in Benincà’s (2004; 2006; 2013) work and, as noted in

52

the left periphery

§3.2.2, is argued by Wolfe (2016a) to be characteristic of all the earliest Romance vernaculars.2⁵ It should, therefore, come as no surprise that information focus-initial V2 clauses are widespread in the earliest written attestations of French from the 10th to the late 12th centuries, a finding concurred with elsewhere in the literature (Labelle 2007: 302–305; Mathieu 2012: 341; Wolfe 2016a: 480; Steiner 2014: 170–174; Labelle and Hirschbu¨hler 2018: 275–280):2⁶ (43) a. Plein est de figure é signefiance full be.3sg of figure and meaning ‘This is full of symbolism and meaning’ (OFr. QLR 1, 4, 12th century) b. Prozdom i out pur sun seignur aider knights loc.cl have.3sg.pst for his lord help.inf ‘He had knights to help his Lord’ (OFr. Roland 3, 12th century) c. un filz lur dunet a son them.cl give.3sg.pst ‘He gave them a son’ (OFr. Alexis 28, 11th century) As in Modern Romance languages showing information-focus fronting, adjacency holds between the finite verb or auxiliary and the fronted constituent. This can be seen in cases like (44) where a topic precedes the focus, yielding a V3 order: (44) Ma lunge atente a grant duel est venude my long wait to great pain be.3sg come.ptcp ‘my long wait has ended in great pain’ (OFr. Alexis 44, 11th century) However, there is an increasing acceptance that the properties of the focus field, which are broadly comparable to Latin in the Early Old French texts, change such that fronted foci are rare by 1200 (Marchello-Nizia 1995: 99; Steiner 2014; Wolfe 2016a: 480). Marchello-Nizia (1995: 99–100) observes that fronted objects progressively encode more ‘old’ information and Labelle and Hirschbu¨hler (2018) in the most detailed quantitative study to date using the MCVF corpus show that ‘a preverbal object tended to be the I-focus of the clause before 1205, but the I-topic of the clause, after 1225’ (Labelle and Hirschbu¨hler 2018: 276). They also note a modest decline in fronted QPs, which as we saw in §3.2.2 show properties that are partially diverse from other 2⁵ See Donaldson (2012) for an account of Old French word order under which the focus field is the locus of V2. This is also the account developed by Poletto (2014) for Old Italian. 2⁶ Note that this observation makes the frequently used descriptive label of Old French’s ‘Theme-V-X’ syntax hugely problematic, as the initial constituent can be focal/rhematic in the earliest texts.

3.3 the old french left periphery

53

foci. These are, however, still found in 13th-century prose (46) alongside contrastive foci (47) and we will see that QPs retain a special status into the Middle French period and beyond: (45) .XX. eschles ad li reis anumbrees twenty columns have.3sg the king number.ptcp ‘The King has organized them in twenty divisions’ (OFr. Roland 112, 12th century, Labelle and Hirschbu¨hler 2018: 276) (46) Grant piece parlerent de ceste chose … great amount speak.3pl.pst of this thing ‘They spoke a great amount about this thing’ (OFr. La Queste 4.37, 13th century) (47)

Meïsmes la pulcele y fu even the girl loc.cl be.3sg.pst ‘Even the girl was there’ (OFr. Cassidorus 149, 13th century, Labelle and Hirschbu¨hler 2012: 17)

As we will discuss further in §3.3.6, the decline in fronted foci observable during this period is a significant change with wide-reaching ramifications for the structure of the left periphery, still observable in the syntactic system we find in Modern French.

3.3.3 The frame-topic field There is greater continuity in the licensing of constituents within the frametopic field than in the focus field during the Old French period. However, there is a notable point of contrast with the modern language, insomuch as fronted topics are widely attested without a resumptive clitic in the clausal core (Priestley 1955; Kroch 1989: 213–215; Roberts 1993: 108, 234; Vance 1997: 234; Troberg 2004; Salvesen 2013: Chapter 3). This—as we saw in §3.2.3—constitutes a point of continuity with Latin. Consider in this regard the following examples from the early 12th-century Roland and 13th-century La Queste: (48) a. Ço senefiet pais e this signify.3sg peace and ‘This signifies peace and humility’

humilitet humility (OFr. Roland 73, 12th century)

54

the left periphery

b. Ço dist li reis this say.3sg the King ‘The King says this’ (OFr. Roland 327, 12th century) c. Iceste espee porterai en Arabe this sword bring.1sg.fut in Arabia ‘I will bring this sword to Arabia’ (OFr. Roland 2282, 12th century) (49) a. Cele semblance donc l’evangile parole poons nos that appearance of-which the-gospel speak.3sg can.1pl we veoir … see.inf ‘We can see that figure about which the gospel speaks …’ (OFr. La Queste 167.5, 13th century) b. Ceste aventure veïstes vos aucune foiz this event see.2pl.pst you some time ‘You have sometimes seen this event’ (OFr. La Queste 152.22, 13th century) c. Il oste ses armes he remove.3sg his weapons ‘He removes his weapons’ (OFr. La Queste 160.34, 13th century) Indeed, given the general absence of new information focus in 13th-century French (§3.3.2), the claim has been made that the prefield in Later Old French should be considered exclusively as a topic position (Rinke and Meisel 2009). Although, as we saw, this claim is too strong when the full range of focal constituents is considered, it is true that ‘old’-information constituents such as (48, 49) constitute the majority of the extant data.2⁷ Discussion of dislocation structures where a resumptive form is present has not been prominent in the literature owing to their very limited distribution, but a number of works have discussed their characteristics in Old French. Present—albeit to a limited extent—from at least the time of the Roland onwards (de Bakker 1997: 66; Marchello-Nizia 1998: 161–162; Labelle 2007; Mathieu 2012: 342–344), the consensus is that they increase in frequency during the Old and Middle French period (Priestley 1955), a point that Kroch (1989) has demonstrated quantitatively.

2⁷ Alongside Marchello-Nizia’s (1995: 99–100) study of the Roland compared to La Queste, consider Wolfe’s (2015b: 89) finding that none of the initial direct objects in his 1000-clause sample encode ‘new’ information.

3.3 the old french left periphery

55

(50) Ceste bataille veirement la ferum this battle truly it deliver.1pl.fut ‘Truly we will wage this battle’ (OFr. Roland 882, 12th century) Buridant (2000: 756–757) considers dislocation to typically be a topicalization strategy—as it appears to be in (50)—but one that can also be employed to ‘bring a rheme into relief ’, i.e. to encode focus (Buridant 2000: 756); MarchelloNizia (1998: 176) reaches a similar conclusion. Thus, although not a core part of Old French syntax, we can conclude that the Old French left-periphery hosted left dislocates, as we saw was also the case for Latin to a limited extent. Whilst suggesting that the position targeted by left dislocates is higher than that hosting moved topics which satisfy V2, I do not locate this in either of Benincà and Poletto’s (2004) LD or HT positions, as Meklenborg, Helland, and Lohndal (2021) have recently shown that Old French dislocates do not straightforwardly pass the tests proposed for either category. Finally, note that in keeping with Latin and French of all periods, framesetters can occur sentence-initially before other topical or focal constituents (cf. in particular Donaldson 2012 on the distribution of clauses). Their high position within the left periphery can be diagnosed by the fact that they consistently fail to trigger verb–subject inversion, instead yielding V3 orders, and thus do not ‘count’ for the purposes of V2 (Roberts 1993: 144; Vance 1997: 61–62; Salvesen 2013; Wolfe 2016a):2⁸ (51) Quant cil a Rome sont ensi repairié li when those to Rome be.3pl thus left-again.ptcp the cuens Guillelmes sor un perron s’assiet count Guillaume on a rock refl.cl-sit.3sg.pst ‘When they returned to Rome, the count Guillaume sat down on a rock’ (OFr. Coronemenz Looïs vv. 1352–1353, Donaldson 2012: 1038). Crucially, these account for the vast majority of V3* cases in Later Old French; V4, unlike in Early Old French, is generally absent. We develop this analysis further in Chapter 4, but note for now that this strongly suggests that there are only two positions available before the finite verb in Later Old French, a proposal which has been made elsewhere in the literature (Skårup 1975: 179; Ferraresi and Goldbach 2002; Labelle and Hirschbu¨hler 2005; Labelle 2007:

2⁸ The situation for Old French frame-setters is thus identical to that described for certain Flemish and Dutch varieties by Haegeman and Greco (2018), who show that a similar class of adverbial clauses act as V3-triggers in West Flemish in particular.

56

the left periphery

302; Simonenko and Hirschbu¨hler 2012; Salvesen 2013). I take these two positions to be Spec-ForceP and Spec-FrameP in keeping with the V2 typology presented in §3.1: (52) [Frame XP1 [Force XP2 [Force VFin ] …]]

3.3.4 Particle si A central component of Old French left-peripheral syntax is the particle si. A range of particles can occur in the preverbal field in French (see in particular Van Reenen and Schøsler 2000), but si is by far the most controversial and one of the most frequent: Foulet (1919: 300) notes that ‘there is hardly a page where it does not appear several times’.2⁹ The basic distributional patterns of si are as follows: it can stand in initial position of the clause and thus satisfy V2; it can be preceded by a range of initial constituents, including adverbials, subjects, and clauses (Foulet 1919: 301; Moignet 1973: 288; Einhorn 1974: 103; Jensen 1990: 473; Fleischman 1991: 262; Benincà 1995: 333; Buridant 2000: 509; Salvesen 2013: 143); it frequently—as with other V2 satisfiers—triggers verb–subject inversion (Salvesen 2013: 143; Wolfe 2018b: 342–345); it is exceptionally rare in embedded clauses (Marchello-Nizia 1985: 15; Lemieux and Dupuis 1995: 96; Ferraresi and Goldbach 2002: 11; Ferraresi and Goldbach 2003: 113); and it never occurs at the beginning of a portion of text (Marchello-Nizia 1985: 25; Ménard 1988: 328; Lemieux and Dupuis 1995: 96; Vance 1997: 54; Reenen and Schøsler 2000: 86; Buridant 2000: 508; Bonnard and Régnier 2008: 209), termed ‘thematic paragraphs’ by Fleischman (1991: 256). Two accounts of si have been particularly prominent in Old French and Medieval Romance scholarship. The first is the account developed by Fleischman (1991), under which si is a marker of topic-continuity or same-subject reference.3⁰ Motivation for such an analysis comes from examples such as (53), where si clearly performs a disambiguating function in encoding that the nominal subject Li vaslés ‘the vassal’ is the intended topic of all three matrix clauses:

2⁹ On the frequency of si, see also Marchello-Nizia (1985: 2), Fleischman (1991: 261), Buridant (2000: §408), and Wolfe (2018b: 338). 3⁰ See also Van Reenen and Schøsler (1992) and Buridant (2000: 508). It is worth noting that the idea that si can encode continuity with a referent in the preceding discourse is not new (Diez 1882: 2060; Skårup 1975: 238–239; Blumenthal 1980: 121).

3.3 the old french left periphery

57

(53) Li vaslés entendi bien que li the vassal understand.3sg.pst well that the empereres li donnoit boin consel; si emperor him.cl- give.3sg.pst good counsel si s’atorna au plus belement refl.cl-prepare.3sg.pst at-the most best qu’il peut, si s’en that-he can.3sg si refl.cl-loc.clavec les messages vint come.3sg.pst with the messengers ‘The servant understood clearly that the emperor had given him good advice. He prepared the best he could and came with the messengers …’ (OFr. Clari 30, 31, 13th century) However, it is less straightforward to see how this analysis captures the full range of data, including cases where si is preceded by an initial clause including a readily identifiable subject, a pattern which is near systematic in certain 13thcentury prose texts (Wolfe 2018b: 339; Shaw 2020), or cases where si co-occurs with a subject encoding brand new information, and thus an extremely poor candidate for ‘same-subject reference’ (54a, b):31 (54) a. E si i furent e Gerin e Gerers and si loc.cl- be.3pl.pst and Gerin and Gerer ‘And Gerin and Gerer were there’ (OFr. Roland 107, 12th century) b. Quant je me fui endormiz si m’avint when I me.cl be.1sg.pst asleep si me.cl-come.3sg.pst une avision merveilleuse a vision marvellous ‘When I was asleep a marvellous vision came to me’ (OFr. La Queste 197a, 23–24, 13th century) The second account which has been prominent in recent theoretically informed studies is one under which si is a last-resort mechanism for satisfying V2 (Lemieux and Dupuis 1995: 95; Benincà 1995: 333; Benincà 2006: 64; Vance 1995: 184–185; Rouveret 2004: 193–195; Salvesen 2013: 143; Wolfe 2016a: 469–470). This account has many advantages: it captures why, in certain contexts, si is described as semantically vacuous (Foulet 1919: §300; 31 For evaluation of other analyses of si, including the simple adverbial analysis (Foulet 1919: 301; Anglade 1930: §248; Moignet 1973: 287; Jensen 1990: 472–473; Bonnard and Régnier 2008: 209–210) and the head analysis (Ferraresi and Goldbach 2002; Ferraresi and Goldbach 2003; Ledgeway 2008; Meklenborg 2020a; Meklenborg 2020b), see Wolfe (2018b).

58

the left periphery

Ménard 1968: 81; Moignet 1973: 287–288; Skårup 1975: 238–239; Jensen 1990: 472–473; Roberts 1993: 330; Vance 1995: 185; Vance 1997: 53); it explains why si is near-absent in embedded contexts (Marchello-Nizia 1985: 15; Lemieux and Dupuis 1995: 96; Ferraresi and Goldbach 2002: 11; Ferraresi and Goldbach 2003: 113), where V2 is generally not operative (see Adams 1987c and §3.3.5); and it accounts for the marked decline in the frequency of si in the 16th century (Marchello-Nizia 1985: 200; Fleischman 1991: 278), when V2 effects are no longer present (Roberts 1993: 220–227; Vance 1995; Vance 1997: 350–351; Steiner 2014: 260; §3.4). However, it is also worth noting that the exact conditions under which this ‘last-resort’ usage would obtain have never been outlined in detail. Nor is it entirely straightforward to reconcile observations of semantic vacuousness, on the one hand, with demonstrable effects in encoding topic-continuity in certain texts, on the other. Against this somewhat conflicting set of observations concerning the distribution and correct formal analysis of si, Wolfe (2018b; 2020a) adopts a hybrid account, under which si can target either Spec-FinP or Spec-ForceP depending on the period. This distributional change, from FinP to ForceP expletive, correlates with a change in the locus of the V2 property around 1200, which we discuss in §4.2.2. Within Old French texts, the idea is that at least two distinct types of si are diagnosable:32 Stage 1: This is found in the Chanson de Roland, but the relative scarcity of texts from the 10th and 11th centuries make it challenging to establish whether it is typical of the very earliest French texts. At this stage, si encodes topiccontinuity in keeping with Fleischman’s (1991) analysis. Its low position within the left periphery—Spec-FinP—is evidenced by the fact it can be preceded by frame-setters, topics, and foci, and even focal objects (55). A small piece of supporting evidence that this is not an idiosyncratic characteristic of a single text comes from the Strasbourg Oaths, which show si preceded by multiple left-peripheral constituents (56): (55) Reis Vivien si succuras en Imphe King Vivien si help.2sg.fut in Imphe ‘You will help King Viven in Imphe’ (OFr. Roland 3995–3996, 12th century)

32 See Wolfe (2018b; 2020a; Forthcoming) for a more fine-grained typology of the distribution in Old French and Medieval Romance.

3.3 the old french left periphery

59

(56) D’ist di in avant, in quant Deus savir from-this day in forward in so-far God knowledge et podir me dunat si salvari eo …. and power me.cl give.3sg si support.1sg.fut I ‘from this day forward, insofar as God gives me knowledge and power, I will support …’ (OFr. Strasbourg Oaths, 9th–10th centuries) Stage 2: This system is typical of 13th-century prose and may continue into the Middle French period (cf. §4.3–4). The evidence is that si is merged in Spec-ForceP with no clear role in encoding topic continuity.33 In distributional terms, there are major differences from the earlier system, with foci no longer compatible with si and topical arguments only to a highly limited extent. The generalization for this period appears to be that si can only occur on its own or with a single preverbal constituent lexicalizing the frame field of the left periphery, in the majority of cases a scene-setting adverbial (57) or clause (58), with clause + si accounting for the majority of the data in several 13th-century texts.3⁴ (57) puis si se departirent then si refl.cl leave.3pl.pst ‘then they departed’

(OFr. Clari 4, 2, 13th century)

(58) Quant il vint en la valee si comença a when he come.3sg.pst in the valley si begin.3sg.pst to penser mout durement think.inf very hard.adv ‘When he came to the valley, he began to think very hard’ (OFr. La Queste 194, 8, 13th century) The overall conclusion is that si is far from the homogeneous phenomenon many works on Old French syntax would assume. There is compelling evidence that the particle targets distinct positions in Early Old French and Later Old French (59) and I will suggest in §3.6 that it may have had an important role in triggering other changes observable in the history of the language. (59) [Frame [Force siLater Old French [Topic [Focus [Fin siEarly Old French [TP …]]]]]]

33 See Shaw (2020) for a recent quantitative evaluation of this claim using a corpus of 13th- and 14th-century prose. 3⁴ In Wolfe’s (2018b) sample of Villehardouin clauses account for 57.3% (122/213) of XP + si configurations, 76.7% (109/142) in Clari, and 94.1% (177/188) in La Queste.

60

the left periphery

3.3.5 Embedded clauses in Old French As with our discussion of the matrix left periphery, it is necessary to divide our discussion of the architecture of embedded clauses to differentiate between the distinct systems found in Early and Later Old French.3⁵ In a directly parallel fashion to what we have observed for matrix clauses we will see that the passage from Early Old to Later Old French is characterized by increased restrictions on access to the embedded left periphery. Two hypotheses have been proposed to account for the structure of embedded clauses in Early Old French. Under one account Early Old French is argued to have a symmetrical V2 syntax (Lemieux and Dupuis 1995; Labelle 2007); the ‘symmetrical’ V2 label has been employed in Germanic linguistics to refer to some varieties of Icelandic (Ro¨gnvaldsson and Thra´insson 1990; Santorini 1994; Ga¨rtner 2003) and Yiddish (Santorini 1989; Diesing 1990; Diesing and Santorini 2020), amongst others. In recent years, its status has been heavily criticized (Wiklund et al. 2007; Wiklund et al. 2009; Wolfe 2015c) and some have argued that this particular class of language does not exist at all (Walkden and Booth 2020). In the case of Early Old French, if V2 effects really are symmetrical, then we predict that the left periphery is able to host fronted constituents to the same degree as the matrix left periphery. However, although the consensus is that word-order flexibility in the earliest Old French embedded clauses is greater than Later Old French (Mathieu 2006; Wolfe 2018a: Chapter 3; Labelle and Hirschbu¨hler 2017), the notion that the language had a symmetrical V2 syntax seems too strong. Indeed, no quantitative study has yet been undertaken on the earliest Old French texts which shows the makeup of matrix and embedded clauses to be identical. Rather, what we find is a general pattern of embedded (S)VO complement (60) and relative (61) clauses,3⁶ with a very small proportion of relatives and wh-clauses licensing an SOV order (Skårup 1975: 502–503; Bauer 1995: 110; Buridant 2000: 748–749; Salvi 2011: 366–367). Embedded focalization and topicalization is found,3⁷ but in complement clauses only under verbs known to license main clause phenomena cross-linguistically (Hooper and Thompson 3⁵ The importance of distinguishing between an ‘archaic/early’ and an ‘innovative/later’ system is already acknowledged in Adams (1987b), Vance (1988), and Roberts (1993), among others. We will discuss this distinction relative to embedded subjects in §5.3. 3⁶ An exhaustive study of complement clauses with the verb in second position in the Roland shows 58 (74.4%) to be SVO and 20 PP, Adverb, or Object-VS (25.6%). These are clearly the inverse of the figures provided by Roberts (1993) for matrix clauses in Table 3.1 (34.1% SVO vs. 65.9% XVS). 3⁷ Labelle’s (2007) data, however, clearly show that a greater range of embedded clauses permit access to the left periphery in Early Old French, a finding that supports my line of analysis here that Early Old French is not ‘symmetrical’, but more symmetrical than Later Old French.

3.3 the old french left periphery

61

1973; Rizzi and Roberts 1989; Heycock 2006: Chapter 4; Julien 2007; Aelbrecht, Haegeman, and Nye 2012) (62): (60) il me dist que il me trova// en un bois he me.cl say.3sg that that me.cl find.3sg.pst in a wood ‘He told me that he found me in a wood’ (OFr. Thèbes1 505, 12th century) (61) et la terre qui est dedenz and the ground that be.3sg inside ‘and the ground that is inside’ (OFr. Thèbes1 640, 12th century) (62) a. Quant Carles veit que morte l’ad truvee when Charles see.3sg that death he.cl-have.3sg find.ptcp ‘When Charles sees that he has found death’ (OFr. Roland 3728, 12th century) b. Franceis veient que paiens i ad tant French see.3pl that pagans loc.cl have.3sg many ‘The French see that there are many pagans’ (OFr. Roland 1510, 12th century) In contrast to Later Old French, V1 (63) and a range of V3 (64) orders are also found in the embedded domain (Hirschbu¨hler 1990; Roberts 1993; Labelle 2007): (63) et ainz que fust le cors couverz and as-soon that be.3sg.pst the body cover.ptcp ‘and as soon as the body was covered …’ (OFr. Thèbes2 12, 12th century, Labelle 2007: 307) (64) … Que enz enfern jo demaine that in hell I undergo.1sg ‘… that I undergo in hell’ (OFr. Brandan 1301–1302, 61, 12th century, Labelle 2007: 299) Overall, the embedded left periphery of Early Old French—which I illustrate here for complement clauses—is as follows. The unmarked SVO order is indicative that in the paradigm case the finite verb targets T (see §4.5.1) with a subject optionally merged in Spec-TP. We can assume in such cases that the complementizer’s presence in Fin (65a) in most cases blocks the V2 effects observable in matrix clauses. However, a restricted class of selecting verbs known to license main-clause phenomena cross-linguistically can take as a complement a clause where the complementizer is in Force (65b); V-to-Fin movement

62

the left periphery

can thus obtain and embedded topicalization and focalization of the type we see in (62) can be licensed. (65) a. [Frame [Force [Topic [Focus [Fin [Fin QUE] [TP … (S) V O…]]]]]] b. [Frame [Force [Force QUE][Topic Topic [Focus Focus [Fin [Fin V] [TP …(S) V O]]]]]] The situation in Later Old French, which has been more widely discussed (Adams 1987c; Adams 1987b; Roberts 1993: Chapter 2; Vance 1997: Chapter 4; Wolfe 2018a: Chapter 5; Salvesen and Walkden 2017), is quite distinct. Generally, an SVO order obtains as in Early Old French, with subjects accounting for the vast majority of initial constituents in embedded ‘V2’ clauses, in contrast to matrix clauses, where we saw that this is not the case. Indicative data is given in Table 3.2, based on the discussion of the 13th-century prose text La Queste in Wolfe (2018a: 81). Table 3.2 Preverbal constituents in embedded clauses in La Queste Total SVO Total XP-V-(S) (Of which) Total O-V-(S) Total Count

119 23 14 142

83.8% 16.2% 9.9% 100.0%

In 13th-century prose, genuine cases of embedded V2 with object-fronting are rare and, when found, are restricted to complements of verbs permitting main-clause phenomena, as in Early Old French (Salvesen and Walkden 2017; Wolfe 2018a: 84–86) (66). (66) et bien i pariot a ce qu’ il avoit and well loc.cl seem.3sg.pst to that what he have.3sg.pst le jor fet que d’iluec en avant the day do.ptcp that of-here in forward porroit il legierement sormonter de proece can.3sg.pst he easily surmount.inf of prowess toz les autres chevaliers all the other knights ‘And based on what he had done that day, it seemed likely that he would easily surpass the boldness of all the other knights in the future’ (OFr. La Queste 163a, 32, 13th century, Salvesen and Walkden 2017: 177)

3.3 the old french left periphery

63

One notable exception to this relatively strict word-order generalization is recomplementation structures like the following (67), where a topic—frequently a clause—occurs between two finite complementizers (see in particular Salvesen 2014). Recall from §3.2.4 that these structures are present elsewhere in other Medieval Romance varieties but not found in Modern French: (67) Or dit li contes que quant Galaad se now say.3sg the story that when Galahad refl.cl fu partiz de ses compaignons que il be.3sg.pst leave.ptcp of his companions that he Chevaucha .iii. jors ou iiii. …. ride.3sg.pst three days or four ‘Now the story says that when Galahad left his companions, he rode for three or four days …’ (OFr. La Queste 37, 27–30, 13th century) Extending Ledgeway’s (2005) analysis of Medieval Italian Dialects to Old French, Salvesen (2014) proposes that such cases can be seen as overt spellout of a complementizer merged in Fin, which undergoes movement to Force. Recall from our discussion in §3.2 of the Latin embedded left periphery that Latin complementizers/subordinators are canonically merged in Fin and do not undergo such movement: (68) [Frame [Force [Force QUE][Topic Topic [Focus [Fin [Fin QUE] [TP …(S) V O]]]]]] The important point for our analysis of Later Old French is that cases such as (67) are the only instances where more than one constituent can precede the finite verb in the embedded domain; Wolfe (2018a: 80–81) shows that no embedded V3* is found within his embedded corpus from La Queste and Clari. This points to a major difference between matrix and embedded clauses, where two constituents can precede the finite verb to a restricted extent in 13th-century prose in the former but not the latter case (§3.3.4 and §4.3.3). I therefore conclude that by the 13th century, the range of embedded contexts licensing main-clause phenomena has diminished and that embedded V2, as in (65b), is attested, but to a lesser extent than in Early Old French and never with more than one constituent fronted to the left periphery.

64

the left periphery

3.3.6 Syntactic change from Early Old French to Later Old French Despite traditional handbooks (Foulet 1919; Skårup 1975; Ménard 1988; Buridant 2000) and more recent treatments (Mathieu 2012) often treating the Old French left periphery as a somewhat homogeneous entity, our discussion in this section has given us significant grounds to question this assumption and advocate a split periodization between Early and Later Old French. Without doubt, certain significant properties are found across a wide textual base from the earliest written attestations through to the late 13th century; these include the V2 property, topicalization through movement, dislocation structures, frame-setters occupying a high left-peripheral projection, recomplementation, and matrix/embedded asymmetries. However, certain properties are indicative of ongoing changes in the system. We have seen that focus is not uniformly licensed across Old French texts, with evidence for new information focus, a major syntactic isogloss separating Modern Romance languages today (Cruschina 2012), diminishing in the course of the late 12th and early 13th centuries. This leads to a grammar where the left periphery increasingly hosts frame-setting or topical constituents encoding ‘old’ information. Turning to the syntax of si, its original function encoding topic continuity and placement low in the left periphery also undergoes change, such that by Later Old French texts it has undergone bleaching of this original function and occupies a higher left-peripheral position. Finally, the strength of matrix/embedded asymmetries increases during this period, with access to the embedded left periphery heavily restricted in Later Old French texts. We will suggest in Chapter 4 that some of these changes in left-peripheral syntax are attributable to changes in the locus of verb movement.

3.4 The Middle, Renaissance, and Classical French left periphery 3.4.1 Middle French and the Verb Second prefield As noted by Smith (2002) amongst many others, the Middle French period is often seen as one of instability and flux; this also pertains to the left periphery where points of discontinuity relative to Later Old French have frequently been noted (Hirschbu¨hler and Junker 1988; Roberts 1993: §2.3; Hirschbu¨hler 1995; Lemieux and Dupuis 1995; Vance 1995; Vance 1997). Undoubtedly the most significant question that can be asked is if the prefield continues to show

3.4 the middle, renaissance, and classical french left periphery

65

the characteristics of a V2 language beyond the early 13th-century texts which have formed the basis for many classic studies on this property (e.g. Adams 1987c; Vance 1997: Chapter 2; Salvesen 2013; Wolfe 2018a: Chapter 5). If, as Vance (1995) and Roberts (1993) have argued, a form of V2 system persists into the 15th century, other questions then arise about the nature of the system at a more fine-grained level. Does the relatively strict V2 syntax found in 13th-century prose persist, where a small class of frame-setters alone can trigger V3 orders (cf. §3.3.3)? Furthermore, do we see a continuation of the trend whereby the left periphery is increasingly specialized in hosting non-focal constituents which encode ‘old’ information? We will now consider the relevant facts from three Middle French texts: Enguerrand de Monstrelet’s Chronique (1441–1444), the Roman de Jean de Paris (1494), and Philippe de Commynes’s Mémoires (1490–1505). We consider detailed facts on verb placement in §4.4, but as background to our discussion of the left periphery, note that second position is the most frequent one for matrix-clause finite verbs in our three texts (69), but V3 is also found (70):3⁸ (69) Le chasteau tint et ne fut point assailly the castle hold.3sg.pst and neg be.3sg.pst neg assailed ‘He held the castle and it was never attacked’ (MidFr. Mémoires 16, 15th–16th centuries) (70) Quant vint au matin, le roy se leva when come.3sg.pst to-the morning the king refl.cl get-up.3sg.pst ‘When the morning came, the king got up’ (MidFr. Roman de Jean de Paris 2, 3, 15th century) However, we saw above for Old French that finite verb placement alone is insufficient to establish the V2 status of the system. Rather, we require clear evidence that the prefield is not a specialized subject position, as in Modern French. The statistics in Table 3.3 thus lend support to the hypothesis that preverbal constituents are located in the left periphery and not the inflectional layer. Consider the examples in (71–75) which show the range of constituents that can occupy the prefield.

3⁸ Chronique (46.1%, 404/877), Roman de Jean de Paris (60.2%, 526/874), and Mémoires (55.1%, 468/849).

66

the left periphery

Table 3.3 Preverbal constituents in V2 matrix clauses in Middle French

Total SVO Total XP-V-(S) Total O-V-(S) Total Count

Chronique

Roman de Jean de Paris

Mémoires

109 295 49 453

248 278 20 526

252 216 35 468

27.0% 73.0% 12.1% 100.0%

47.1% 52.9% 3.8% 100.0%

53.8% 46.2% 7.5% 100.0%

(71) a. ce dit le roy d’Angleterre it say.3sg the king of-England ‘The King of England said it’ (MidFr. Roman de Jean de Paris 46, 15th century) b. Et ces parolles m’a compté le roy and these words me.cl=have.3sg tell.ptcp the king ‘And the King has told me these words’ (MidFr. Mémoires 9, 15th–16th centuries) (72) a. Et de ce fut le Roy content and of this be.3sg.pst the King happy ‘And the King was happy with this’ (MidFr. Chronique 3, 15th century) b. De ceste response rirent moult longuement les of this response laugh.3pl.pst very long.adv the Anglois English ‘The English laughed for a long time at this response’ (MidFr. Roman de Jean de Paris 42–43, 15th century) (73) Si fut ceste conclusion tenue si be.3sg.pst this conclusion keep.ptcp ‘This decision was maintained’ (MidFr. Chronique 7, 15th century) (74) Pour lors estoient les subjectz de ceste maison Back then be.3pl.pst the subjects of this house de Bourgongne en grande richesse … of Burgundy in great wealth ‘Back then the subjects of this house of Burgundy were very rich …’ (MidFr. Mémoires 6, 15th–16th centuries) What is perhaps most striking about these examples and the corresponding figures in Table 3.3 is that they not only match but often supersede the figures

3.4 the middle, renaissance, and classical french left periphery

67

in Table 3.1 for the left periphery in Old French texts. Taken in isolation, we see that subjects account for only something over half of V2 clauses in the latest text Mémoires and that non-subject-initial clauses are robustly attested in all three texts. Taken in comparison with the Old French data in Table 3.1, we find that the proportion of SVO is, in fact, lower in Chronique than all the Old French case studies. Looking at the other two texts, their figures are comparable to those presented in Table 3.1. I therefore take this as a piece of supporting evidence that V2 was retained in French until the end of the 15th century at least.

3.4.2 The frame, topic, and focus fields in Middle French Change in the mapping between syntax and information structure is well established for Later Old French (Marchello-Nizia 1995: 99–100; Rinke and Meisel 2009: 112; Labelle and Hirschbu¨hler 2018; Wolfe 2018a: 70–71): recall from our discussion in §3.3.3 that in 13th-century prose the left periphery is increasingly specialized in hosting topics and frame-setters that encode ‘old’ information. In order to test whether this is also true of Middle French texts, all preverbal (pro)nominal objects were extracted from the corpus used for Table 3.3, as well as the first 100 preverbal subjects.3⁹ All preverbal constituents were then tagged as to whether they were discourse-old or -new, using the decision-tree from Rahn (2016: 39–40) and Larrivée (2019: 8), which takes into account (i) explicit mention in the preceding discourse, (ii) syntactic indicators of deixis or anaphoricity, (iii) set constructions, and (iv) contextual use which refers to previous textual content or world knowledge. The results are summarized in Table 3.4. Table 3.4 Discourse status of preverbal constituents in Middle French

Old Object New Object Old Subject New Subject Total Old Total New Total Count

Chronique

Roman de Jean de Paris

Mémoires

5 1 86 14 91 15 106

7 0 86 14 93 14 107

3 5 75 25 78 30 108

4.7% 0.9% 81.1% 13.2% 85.8% 14.2% 100.0%

6.5% 0.0% 80.4% 13.1% 86.9% 13.1% 100.0%

2.8% 4.6% 69.4% 23.1% 72.2% 27.8% 100.0%

3⁹ This relatively low number is necessary to standardize the data, as there are only 109 SVO clauses in the Chronique sample as noted in §3.4.1.

68

the left periphery

Table 3.4 shows that—as we might expect if the trend observable in Later Old French continues—the majority of constituents occurring in all three texts are discourse-old, with many showing clear morphosyntactic indicators of this, as in (75): (75) a. Lesquelz nagaires avoient esté ou pays the-which sailors have.3pl.pst be.ptcp in country de Rethelois of Rethelois ‘These particular sailors had been in the country of Rethelois’ (MidFr. Chronique 35, 15th century) b. Ceulx furent prins they be.3pl.pst taken ‘They were taken’ (MidFr. Roman de Jean de Paris 11, 15th century) c. Ledict conte se mist par le Champ the-said count refl.cl put.3sg.pst on the field pour rallier gens to rally.inf people ‘The aforementioned count went on to the field to rally people’ (MidFr. Mémoires 13, 15th–16th centuries) However, as Table 3.4 indicates, there exists a very small number of initial constituents which encode new information and thus may be focal; the five new-information objects in Mémoires are suggestive that new information focus may be licensed in this text to a limited extent, despite the strong tendency for preverbal constituents to occur in the frame-topic layer. However, upon close inspection, four of the five feature initial QPs, which we saw in §3.3.2 have a special status in Later Old French due to their dedicated functional projection within the left periphery (cf. §3.2.1 alongside Quer 2002): (76) Autres en y a qui sont si bons … others part.cl loc.cl have.3sg who be.3pl so good ‘There are others who are so good … (MidFr. Mémoires 27, 15th–16th centuries) So far, we have observed considerable continuity with the Later Old French data. However, there are two areas in which this is clearly not the case. Firstly, we noted in §3.3.3 the well-established view that in Later Old French constituents triggering V3 ‘deviations’, where two constituents precede the finite verb, belong to a small class of frame-setters (Roberts 1993: 144;

3.4 the middle, renaissance, and classical french left periphery

69

Vance 1997: 61–62; Salvesen 2013; Wolfe 2016a) in contrast to Early Old French where a wider range of frame-setters, topics, and foci can co-occur, yielding V3 or V4* orders. Whilst we do find cases such as (77) which are widely attested in Later Old French (see other examples in Wolfe 2018b: 78– 79), we also find examples where multiple constituents co-occur which do not fall into this category (78), as well as examples of V4* (79): (77) Et adonc, le dessusdit seigneur de Labreth and thus the said knight of Labreth entra … enter.3sg.pst ‘And thus the aforementioned knight from Labreth entered …’ (MidFr. Chronique 25, 15th century) (78) a. A Quoy le Roy fist responce to which the King make.3sg.pst response ‘The king responded to this’ (MidFr. Chronique 29, 15th century) b. Pour Ce je suis de vostre Opinion for This I be.1sg of your opinion ‘Because of this I agree with you’ (MidFr. Roman de Jean de Paris 25, 15th century) (79) D’aultres seigneurs, quand ils ont requis pour of-other knights when they have.3pl seek.ptcp for personnes qui le valent ès aultres offices de people who it.cl be-worth.3pl the other offices of la justice du royaume, le Roy y the justice of-the realm the King loc.cl a mis … have.3sg put.ptcp ‘As for the other knights, when they sought people who were worthy of the other offices of the jurisdiction, the King put …’ (MidFr. Chronique 18, 15th century) In contrast to what we found for Later Old French, these structures evidence a more articulated left-peripheral structure. Pending discussion on the exact target of Middle French verb movement in §4.3, I suggest that the frame, the topic, and the focus field—to a limited extent for QPs—can be accessed in Middle French: (80) [Frame (Frame-Setter) [Force [Topic (Topic) [Focus (Focus) [Fin [Fin VFin ] [TP …]]]]]]

70

the left periphery

A second point of discontinuity with the earlier French data concerns leftdislocation structures, which we noted above did exist in Latin and Old French, but to a limited extent (cf. §§3.2.3, 3.3.3). The consensus from several qualitative and quantitative studies is that the 14th and 15th centuries see a marked rise in the frequency of left-dislocation structures such as those we see in (81–82) (Priestley 1955; Marchello-Nizia 1980; Kroch 1989; Kroch 2001: 711–713): (81) Et le premier homme qui mourut, ce fut and the first man that die.3sg.pst that be.3sg.pst luy et ses gens him and his men ‘And the first man that died, that was him and his men’ (MidFr. Mémoires 9, 15th–16th centuries) (82) Et celles qui faisoient grant signe de rebellion il and they who make.3pl.pst great sign of rebellion he les faisoit raser et mectre tout them.cl make.3sg.pst tear-down.inf and put.inf all a feu Et a sang to fire and to blood ‘And those [towns] which showed signs of rebelling, he tore them down, set them ablaze, and shed their blood’ (MidFr. Roman de Jean de Paris 9, 15th century, Combettes, Marchello-Nizia, and Prévost 2020: 1317) Although present in our corpus, Priestley’s (1955) data reproduced in Kroch (1989) show this uplift most clearly, which we see also continues beyond the Middle French period (Figure 3.1). Significantly, Kroch shows that the rise in left dislocation occurs concomitantly with the decline in object topicalization without resumption, of the type we saw above in (71).

3.4.3 Particle si in Middle French Despite having been most prominently discussed for Old French (MarchelloNizia 1985; Fleischman 1991; Reenen and Schøsler 2000), si is still found in Middle French, a point acknowledged in Martin and Wilmet (1973: 270) and Marchello-Nizia (1980): in Chronique 77 of 784 non-coordinated matrix V2 clauses feature si and in Roman de Jean de Paris this figure is 134 of 701. There

3.4 the middle, renaissance, and classical french left periphery

71

reprise cases per hundred sentences

2.5 2.0 1.5 1.0 0.5 0.0 1200

1350

1450

1550

1650

1750

1875

Fig. 3.1 Left dislocation in Priestley (1955)

are only five such instances in the latest text, Mémoires, which I do not discuss further here. The vast majority of cases have si in initial position of the clause: 75/77 in Chronique and 127/134 in Roman de Jean de Paris. This is a widespread pattern in both Early and Later Old French so an important point of continuity with the 13th century (Wolfe 2018b: 343): (83) Si s’en allerent moult joyeusement si refl.cl-part.cl go.3pl.pst very joyfully ‘They went very joyfully’ (MidFr. Roman de Jean de Paris 12, 15th century) However, although marginal, with only one case in Chronique and three in Roman de Jean de Paris, cases like (84) where a fronted subject co-occurs with si are not typical of Later Old French, although they are found in the very earliest Old French texts (Wolfe 2018b: 345). (84) a. La première si est qu’il voelt … the first si be.3sg that-he want.3sg ‘The first [reason] is that he wants …’ (MidFr. Chronique 15, 15th century) b. Le maistre d’ostel si y alla the master of-house si loc.cl go.3sg.pst ‘The master of the household went there …’ (MidFr. Roman de Jean de Paris 52, 15th century) We could dismiss these as a rare word-order phenomenon. However, we have already seen that Middle French allows a greater range of V3* orders than Later Old French, such that the behaviour of si is not exceptional. Rather, if

72

the left periphery

we assume minimally—in line with the proposal in (59)—that si can occur in either Spec-ForceP or Spec-FinP, these data are only compatible with the latter hypothesis: fronted subjects are quite obviously not frame-setters. The location of si would thus have undergone ‘downwards’ reanalysis from Force to Fin during the 13th and 14th centuries, permitting si to occur alone, with a frame-setter as in Later Old French, or, exceptionally, with a topic (84). We will suggest in Chapter 4 that this downstairs reanalysis is concomitant with a downstairs reanalysis of verb movement. (85) [Frame (Frame-Setter) [Force [Topic (Topic) [Focus [Fin siMiddle French [TP …]]]]]]

3.4.4 The focus field in Renaissance and Classical French Indications that a major change has taken place in the transition from the time of our Middle French texts to the 16th century comes from the makeup of the preverbal field, which, alongside evidence of the decline of verb–subject inversion (§5.3) and matrix/embedded asymmetries (§3.4.6), suggests the loss of the V2 property. As an initial indicator that neither Force nor Fin bear an EF requiring a left-peripheral constituent to be merged in all matrix clauses, consider Table 3.5, which shows the proportion of subject-initial and non-subject-initial ‘V2’ matrix declaratives in a sample of texts.⁴⁰ Table 3.5 Preverbal constituents in linear-V2 clauses in Renaissance and Classical French

Total SVO Total XP-V(S) Total O-V(S) Total Count

Rabelais (1532–1534)

Montaigne (1580)

Phèdre (1677)

Candide (1759)

416

89.8%

382

91.4%

738

47

10.2%

36

8.6%

3

0.4%

7

0.9%

1

0.2%

1

0.2%

1

0.13%

1

0.12%

463

100.0%

418

100.0%

741

100.0% 809

100.0%

99.6% 802

99.1%

⁴⁰ Note that these figures are comparable to those presented by Marchello-Nizia and Prévost (2020: 1092, 1103) for the Renaissance and Classical period, with SVO proportions regularly over 90% in their large-scale corpus.

3.4 the middle, renaissance, and classical french left periphery

73

The generalization from the table is quite clear: in none of the texts does SVO represent less than c.90% of the clauses sampled. Furthermore, we see that cases of non-subject-initial clauses are vanishingly rare in the 17th century. Therefore, we see clear evidence that the matrix prefield, when filled, is overwhelmingly a dedicated subject position (cf. 86) akin to that reported for SVO languages like Modern French and Modern Italian (Cardinaletti 1997; Cardinaletti 2004; Guasti and Rizzi 2002). This finding will be confirmed in our discussion of the subject layer in §5.5. (86) a. Il fera choses merveilleuses he do.3sg.fut things marvellous ‘He will do marvellous things’ (RenFr. Rabelais 2, 37, 16th century) b. ma force m’abandonne my strength me.cl-abandon ‘My strength is leaving me’ (CFr. Phèdre 3, 154, 17th century) However, we know that a range of Modern Romance SVO languages still license focus fronting of some form (cf. Cruschina 2012: Chapter 3; Cruschina 2016 and references in §3.2.1) and that, in some cases, the gradual restrictions on focus fronting during the loss of V2 can be tracked diachronically.⁴1 Given what we have already seen for Later Old French and Middle French, namely that focus fronting had already become heavily restricted, it should come as no surprise that this is also the case from the 16th century onwards. Such is the rarity of fronted foci that no focal DP objects are found in initial position of the clause in any of our textual samples. In the secondary literature, cases of fronted QP direct objects are, however, reported (cf. 87). Also note the single instance of a focal PP object in Phèdre in (88): (87) a. Peu de prudence eurent les pauvres gens little of prudence have.3pl.pst the poor people ‘The poor people had little caution’ (CFr. La Fontaine, Fables 7.7, 17th century, Fournier 2007a: 35) b. Autre chose ne me sceut il dire other thing neg me.cl know.1sg.pst he say.inf ‘He couldn’t tell me anything else’ (CFr. Biard 96, 17th century, Combettes 2003a: 180)

⁴1 See, for example, the discussion in Ledgeway (2009b: Chapter 21) for Old Neapolitan.

74

the left periphery

(88) A nos amis communs portons nous justes cris to our friends common bring.1pl we just cries ‘We will bring our righteous complaints to our common friends’ (CFr. Phèdre 5, 1, 1367, 17th century) We should note, however, the general claim by Marchello-Nizia (1995: 106) that OVS—although found in the 16th century—has ‘essentially disappeared’ by the 17th, with the exception of texts written by La Fontaine, who is deliberately archaizing. This claim of a diachronic split between early 16th-century texts, on the one hand, and later 16th-, 17th-, and 18th-century texts, on the other, is repeated elsewhere by various scholars who point to the decreasing attestation of fronted complements of any kind (Darmesteter 1899: 840; Fournier 2001: 103; Fournier 2007a: 34, 253; Fournier 2007b: 86; Combettes 2003a: 175–180; Lardon and Thomine 2009: 99–102; Marchello-Nizia and Prévost 2020: 1094). It is the case, however, that a small number of non-topical constituents continue to move to the left periphery in this period. Aside from the QPs mentioned immediately above, we can add to this class predicative adjectives (Gougenheim 1973: 229; Brunot 1939: 841; Combettes 2003a: 181; Fournier 2007b: 37–38; Marchello-Nizia and Prévost 2020: 1095) and adverbials (Spillebout 1985: 405; Gougenheim 1973: 230; Brunot 1939: 840; Lardon and Thomine 2009: 401; Marchello-Nizia and Prévost 2020: 1905): (89) a. Possible n’est pour le present possible neg-be.3sg for the present ‘It is not presently possible …’ (RenFr. Rabelais 4, 35, 16th century, Gougenheim 1973: 229) b. Tel a esté le renommé Membertou such have.3sg be.ptcp the renowned Membertou ‘The renowned Membertou was such …’ (CFr. Biard 54, 17th century, Combettes 2003a: 181) (90) a. Mais maintenant crioit-elle fermement que but now believe.3sg.pst resolutely that ‘But now she believed resolutely that …’ (RenFr. Marguerite de Navarre, Heptaméron 3, 1, 44, 16th century, Gougenheim 1973: 230) b. là coulent mille divers ruisseaux d’une Eau there flow.3pl thousand different streams of-a water claire clear ‘There flow a thousand different streams of clear water’ (CFr. Fénelon, Tél. 3, 164, 17th century, Fournier 2007a: 36)

3.4 the middle, renaissance, and classical french left periphery

75

c. à peine reprenais-je mes sens que … hardly re-take.1sg.pst-I my senses that ‘Hardly had I reclaimed my senses, than …’ (CFr. Candide 61, 18th century) d. Lors commença le seigneur de Humevesne then begin.3sg.pst the mister of Humevesne ‘Then Mr Humevesne began …’ (RenFr. Rabelais 12, 6, 16th century) With regard to adverbials, both Fournier (2007b: 37) and Marchello-Nizia and Prévost (2020: 1905) note a diachronic distinction in the period we consider here, where prior to the mid-16th century a wider range of adverbs can trigger inversion,⁴2 with a restricted class triggering inversion after this point.⁴3 We should note that non-inversion of adverbials, in this case peut-être ‘perhaps’, followed by an initial constituent, is also found (91): (91) Peut-être le récit d’un amour si sauvage// vous perhaps the tale of-a love so wild you.clfait, en m’écoutant rougir de votre ouvrage make.3sg in me.cl-hear.prog blush.inf of your work ‘Perhaps the tale of such wild love will make you blush in hearing of your work’ (CFr. Phèdre 2, 2, 37, 17th century) Diachronic variation is also observed with predicative adjectives such as those in (88), which are noted to persist longer than fronted objects, but with the trigger for movement becoming increasingly lexically specialized to a small class of adjectives, most notably tel ‘such’ (Marchello-Nizia and Prévost 2020: 1095). In summary, the passage from the 16th to 18th centuries, although characterized uniformly by the absence of V2, is one of change in the focus layer. Focus fronting continues, but is increasingly restricted to defined subclasses of foci, e.g. QPs, or lexically specified terms, as with adverbials and predicative adjectives.

⁴2 For example, bien ‘well, indeed’, si, et—both not straightforwardly translatable—seulement ‘only’, difficilement ‘with difficulty’, au moins ‘at least’. ⁴3 This includes spatial and temporal adverbials alongside the class triggering inversion today (cf. §3.5.1).

76

the left periphery

3.4.5 The frame-topic field in Renaissance and Classical French Fronted thematic objects with no clitic resumption are not found in any of our texts featured in Table 3.5. Cases are reported in the literature, however, but with the observation that whilst initial objects are found at the beginning of the 16th century, they become increasingly unusual thereafter. A difference from focal objects, however, is that thematic DP objects are reported to be licensed, with several scholars noting the explicitly ‘anaphoric’ nature of the constituents in these cases (92) (Combettes 2003a: 175–176; Lardon and Thomine 2009: 401; MarchelloNizia and Prévost 2020: 1094). The explicit discourse-old nature of thematic complements can also be seen in cases with an initial PP, such as (93): (92) Ceste region dit Herodian estre feconde this region say.3sg Herodian be.inf fertile ‘Herodian says that this region is fertile’ (RenFr. Thevet f5, r, 16th century, Combettes 2003a: 175) (93) A ceste destinée ne pouons nous contreuenir to this destiny neg can.1pl we contrevene.inf ‘We cannot prevent this fate’ (RenFr. Rabelais 3, 51, 16th century, Brunot 1927: 481) Recall from Figure 3.1 that left dislocation continues to rise in frequency throughout the Renaissance and Classical period. Whilst direct object topicalization without clitic resumption is not found in any of our sample in Table 3.5, left dislocation with clitic resumption is, and in other texts of the period (Gougenheim 1973: 231; Fournier 2001: 98; Fournier 2007b: 99; Combettes 2003b; Combettes 2009), and is thus indicative of the concomitant fall in topicalization/focalization of complements via movement and rise in base-generated left-dislocation structures noted by Kroch (2001). (94) a. L’amitié que nous portons à nos femmes, elle the-friendship that we bring.1pl to our women it est très légitime is very legitimate ‘The friendship that we bring to our wives, it is entirely proper’ (RenFr. Montaigne 166, 16th century)

3.4 the middle, renaissance, and classical french left periphery

b. et moi je brûle encore and me I burn.1sg again ‘and me I burn again’

77

(CFr. Phèdre 75, 17th century)

Base-generated frame-setting adverbials and clauses, of the type we saw in Latin, Old French, and Middle French, still persist in Renaissance and Classical French texts. As expected, they precede other constituents within the left periphery (95): (95) Quelque jour, je ne sçay quand, Pantagruel se some day I neg know.1sg when Pantagruel refl.cl pourmenoit walk.3sg.pst ‘One day, I don’t know when, Pantagruel went for a walk’ (RenFr. Rabelais 3, 6, 16th century) As a final point, witness the ability for multiple frame-setting or topical (96) constituents to co-occur in the left periphery; alongside other indicators considered in §5.5, the fact that multiple-fronted arguments do not trigger verb– subject inversion is one of many indicators that the V2 property has been lost: (96) a. de laquelle race peu furent qui aimassent … of the-which race few be.3pl.pst that love.3pl.pst.sbjv ‘there were few of that race who loved …’ (RenFr. Rabelais 1, 19, 16th century) b. à tes conseils je me laisse entraîner to your advice I me.cl let.1sg lead.inf ‘I let myself be led by your advice’ (CFr. Phèdre 28, 17th century) We can therefore conclude that the loss of V2, occurring at the very beginning of the 16th century, leads to a substantial split between early Renaissance French texts and later texts from the Classical French period. Alongside very heavy restrictions on focalization, contexts where topicalization obtains continually decline, with fronted topical arguments highly marked, unambiguously ‘anaphoric’, and increasingly rare within the system.

3.4.6 Embedded clauses in Middle, Renaissance, and Classical French Recall that in §3.3.5 we noted asymmetries between Old French matrix and embedded clauses, which become increasingly marked in the transition

78

the left periphery

from Early to Later Old French. Considering first the Middle French data in Table 3.6, we observe a clear asymmetry insomuch as the prefield in Middle French matrix clauses is not a dedicated subject position, with between only 27% and 53.8% of constituents being subjects. However, when we consider similar data on the prefield in the embedded domain, we observe that subjects constitute between 82.9% and 93.8% of preverbal constituents. Table 3.6 Preverbal constituents in embedded V2 clauses in Middle French

Total SVO Total XP-V-(S) Total O-V-(S) Total Count

Chronique

Roman de Jean de Paris

Mémoires

103 8 3 111

92 19 5 111

123 11 0 134

93.8% 7.2% 2.7% 100.0%

82.9% 17.1% 4.5% 100.0%

91.8% 8.2% 0.0% 100.0%

Furthermore, we find no examples in embedded clauses of the CP-particle si, which we can take as a clear indicator that access to the left periphery is not widespread in such clauses. Rather, we observe—as in Later Old French— evidence of an asymmetric V2 grammar where topicalization/focalization of non-subjects does not typically obtain in the embedded domain owing to the blocking effect of the complementizer (cf. Koster 1975: §1; Den Besten 1983: 54–64; Vikner 1995; Wolfe and Woods 2020: 1–4). Consider the SVO clauses in (97) which are typical of the Middle French data: (97) a. ains respondirent qu’ilz n’en feroient but respond.3pl.pst that-they neg-part.cl do.3pl.cond riens nothing ‘But they responded that they would do nothing’ (MidFr. Chronique 9, 15th century) b. et luy commanda le roy que il and him.cl command.3sg.pst the king that he fist extreme diligence make.3sg.pst extreme diligence ‘and the King ordered him to be extremely careful’ (MidFr. Roman de Jean de Paris 7, 15th century) The SVO nature of embedded clauses unambiguously continues into the Renaissance and Classical period, albeit to a greater extent in our latest two texts, Phèdre and Rabelais, where all clauses with the finite verb in second position are SVO. Consider Table 3.7 and the data in (98) in this regard.

3.4 the middle, renaissance, and classical french left periphery

79

Table 3.7 Preverbal constituents in embedded V2 clauses in Renaissance and Classical French Rabelais (1532–1534) Montaigne (1580) Phèdre (1677) Candide (1759) Total SVO Total XP-V-(S) Total O-V-(S) Total Count

93

88.6%

203

99.0%

12

11.4%

2

1.0%

0

0.0%

0

0.0%

1

1.0%

0

0.0%

0

0.0%

0

0.0%

105

100.0%

205 100.0%

79 100.0%

79 100.0%

110 100.0%

110 100.0%

(98) a. Je croy bien que tous les officiers de I believe.1sg well that all the officers of sa estoyent tant occupés … his be.3pl.pst so occupied ‘I firmly believe that all the officers of his court were so occupied …’ (RenFr. Rabelais 4, 47, 15th century) b. mais ayant entendu dire que tout le but have.prog hear.ptcp say.inf that all the monde était riche dans ce pays-là people be.3sg.pst rich in that country-there ‘but having heard that everybody was rich in that country’ (CFr. Candide 21, 18th century) What we also observe, however, is unambiguous evidence of the decline of embedded V2. Walkden and Booth (2020), in discussing controversial cases of embedded V2 in Germanic, highlight that embedded object-fronting is arguably the least ambiguous exponent of a V2 grammar. If we consider Tables 3.7 and 3.8, we see that cases of embedded object-fronting, although always low, are hardly represented in the Renaissance and Classical French data, with a single case in Rabelais and none in Montaigne, Phèdre, or Candide. Note, however, that the single case of object/complement-fronting in our sample of Rabelais involves the predicative adjective tel ‘such’, which we saw in §3.4.5 shows special behaviour in Classical French and later periods in triggering inversion of the verb and subject in matrix clauses (Marchello-Nizia and Prévost 2020: 1095). (99)

eussiez dict que tel estoit son esperit say.2pl.pst.sbjv say.ptcp that such be.3sg.pst his spirit ‘… you would say that such was his spirit …’ (RenFr. Rabelais 110, 16th century)

80

the left periphery

However, just as SVO systems still permit optional topicalization and/or focalization in certain contexts, this is also true of the embedded domain in SVO languages. As such, cases of embedded V3 and V4 are found in Middle, Renaissance, and Classical French, where multiple constituents in the embedded left periphery precede the subject in Spec-TP (cf. §4.5). (100) a. Et dist plusieurs raisons and say.3sg.pst several reasons bonnes et que son good and that his advis estoit que … le opinion be.3sg.pst that the matin, à l’aube du morning at the-dawn of-the jour, on assaillist le roy day one attack.3sg.pst the king ‘and he gave several good reasons and said that his view was that … at dawn they should attack the king’ (MidFr. Mémoires 15, 15th–16th centuries) b. … lesquelz afferment que veritablement the-which affirm.3pl that really ledict Hurtaly n’estoit dedans the-said Hurtaly neg.cl-be.3sg.pst inside l’Arche de Noë the-ark of Noah ‘… who affirm that the said Hurtaly really was not inside Noah’s Ark’ (RenFr. Rabelais 28, 16th century) Additional evidence that the embedded left periphery can be accessed comes from recomplementation structures, which we saw were widespread in Old French texts (cf. §3.3.5 and Salvesen 2014). Whilst these are found in our Middle French textual samples (cf. 101), they are not present in the Renaissance and Classical French texts. This finding is in keeping with Spillebout (1985: 262) who suggests that recomplementation in the classical period is restricted to ‘familiar’ texts and not found at all after 1650. (101) et and oultre over

croy believe.1sg deux two,

que, that traictz shots

s’il if-he d’arc, of-bow

eust passé have.3sg.pst.sbjv pass.ptcp qu’il that-he

3.4 the middle, renaissance, and classical french left periphery

81

eust esté prins comme aucuns autres have.3sg.pst.sbjv be.ptcp taken like some others qui chassoient devant Luy that drove.3pl.pst before him ‘and I believe that if he had passed over two rounds of arrows, he would have been captured like some others who pressed ahead of him’ (MidFr. Mémoires 13, 15th–16th centuries) Overall, we have seen greater evidence of continuity in the embedded domain than in matrix clauses: in the unmarked case Old, Middle, Renaissance, and Classical French show SVO embedded clauses, with verb movement into the T-layer (§4.5.1). The notable point of change we observe within the Middle, Renaissance, and Classical periods is the decline in the licensing of embedded V2, of which there is seemingly no evidence in our latest Classical French texts.

3.4.7 Summary and conclusions Although, as will be proposed in §3.6, certain significant syntactic changes shaping the Modern French left periphery occur in the late 12th and early 13th centuries, much of what we observe in the Middle, Renaissance, and Classical French period constitutes in turn the destabilization, loss, and decline in ‘relics’ of a V2 grammar. The point of continuity across Middle French texts, in contrast to Later Old French, is evidence of access to the full range of left-peripheral projections above Fin. Unlike Early Old French, which appears superficially similar with access to the frame-topic-focus fields above Fin, the lexicalization of the topic and focus fields becomes progressively specialized from the 15th century onwards. Arguably the most significant parametric change in the period considered here is the loss of an EF on Fin, which there is clear evidence for in Middle French texts. By contrast, in Renaissance and Classical French, we find evidence of an SVO grammar, where movement to the left periphery triggered by particular pragmatic or syntactic features progressively declines. Left-dislocation structures, derived through base-generation, however, rise during the whole period considered in this section. In the embedded domain, there is less evidence of a marked discontinuity with Later Old French texts: the general pattern is one of embedded SVO orders, with minimal access to the embedded left periphery. Residual embedded V2 is found, but to a diminishing extent, as in matrix clauses.

82

the left periphery

3.5 The left periphery in Modern French 3.5.1 The focus field in Modern French In contrast to other Modern Romance varieties discussed in §3.2.1 and all earlier stages of the language, in particular Latin and Old French, Modern French has a markedly less active focus field. Nevertheless, that should not be interpreted as a claim that the focus field is entirely absent in any varieties of the language (pace Zubizarreta 2001: 184 as quoted in Rowlett 2007: 184). Rather, I suggest here that the focus field is only rarely activated, and when it is this is typically associated with more formal varieties. An important point of departure is that no modern variety of French permits widespread licensing of left-peripheral contrastive focus or information focus. To illustrate this point, consider the contrast between the Modern Sicilian examples of information focus in (102) and their Modern French equivalents. Similarly, we see that contrastive focus, licit in a range of Modern Romance varieties, is consistently rejected by many French speakers (103), though reported to be acceptable to a subset (Belletti 2005a): (102) a. i. Un librru ci detti a book him.cl give.1sg.pst ‘I gave him a book’ (Modern Sicilian, Cruschina 2011: 54) ii. *Un livre je lui ai donné a book I him.cl have.1sg give.ptcp ‘I gave him a book’ (ModFr.) b. i. Fami avi! hunger have.3sg ‘She’s hungry’ (Modern Sicilian, Cruschina 2011: 63) ii. *Faim elle a! hunger she have.3sg ‘She’s hungry!’ (ModFr.) (103) ?? Pierre ils ont arrêté (pas Jean) Pierre they have.3pl arrest.ptcp neg Jean ‘They have arrested Pierre, not Jean’

(ModFr.)

Rather, what we observe in formal French varieties is movement of a syntactically or lexically defined subclass of foci, with a degree of optionality obtaining in many cases. The best-studied case is adverbial inversion. Recall that

3.5 the left periphery in modern french

83

in Old French, all adverbials could undergo focalization, with growing restrictions on this class emerging from the Middle French period onwards (§§3.4.1, 3.4.4). In Modern French, the class is further restricted, with somewhat conflicting accounts of its delimitation. However, the class is not semantically homogeneous, principally including left-peripheral speaker-oriented adverbials such as peut-être ‘perhaps’, sans doute ‘without doubt’, and probablement ‘probably’, but also adverbials standardly taken as base-generated in the T-layer (Cinque 1999; Schifano 2018), such as encore ‘again’ and toujours ‘still’.⁴⁴ (104) a. Peut-être viendra-t-il maybe come.3sg.fut-he ‘Maybe he’ll come’ (ModFr. Rowlett 2007: 208) b. De même continuera-t-elle à plaider la cause de Likewise continue.3sg.fut-she to plead.inf the case of l’Afrique auprès de ses créanciers the-Africa with of its creditors ‘Likewise she’ll continue to plead the case of Africa with its creditors’ (ModFr. Ouest France, 09.11.94, 2, Guimier 1997: 47) Another area of lexically conditioned variation concerns focalization of adjectives. Although Modern French, unlike earlier stages, does not permit widespread focalization of adjectives, a restricted class of monosyllabic adjectives including tel ‘such’, vrai ‘true’, rare ‘rare, few’, and seul ‘only’ can undergo focalization (Marchello-Nizia and Prévost 2020: 1167): (105) Tel est mon avis such be.3sg my view ‘Such is my view’

(ModFr. Wagner and Pinchon 1962: 519)

In formal registers, initial PPs can also undergo focalization out of their basegenerated position. Note that this is a point of continuity with Classical French, where we observed that left-peripheral PP complements are licensed later than DP complements (§3.4.4). (106) A ça vous n’avez pas pensé to that you neg-have.2pl neg think.ptcp ‘You haven’t thought of that’ (ModFr. Wagner and Pinchon 1962: 530)

⁴⁴ Cf. for discussion of the members of this class Guimier (1997: 43–55) and Grevisse and Goosee (2016).

84

the left periphery

Rowlett (2007: 183) also discusses cases such as (107), which look similar to cases of focus fronting elsewhere in Romance and have been argued to be typical of Belgian French. Following Kroch (2001: 712), however, he argues that these are not genuine instances of new information focus but rather cases of left dislocation (§3.5.2), because they show an intonational pause between the ‘fronted’ constituent and sentential core:⁴⁵ (107) 10F, ce truc m’a coûté 10F this thing me-have.3sg cost.ptcp ‘This thing cost me 10F’

(ModFr. Rowlett 2007: 183)

The general conclusion to draw is that the focus field is rarely employed in any variety of Modern French but is particularly restricted in more colloquial varieties. As we consider further in §3.6 this is the predictable extension of a diachronic trajectory where foci become increasingly restricted in the history of the language.

3.5.2 The frame-topic field in Modern French In contrast to the focus field, which we have seen has a restricted use in Modern French and is employed near exclusively in conservative, formal varieties, the frame-topic field is widely used and active in all Modern French varieties. By far the most productive topicalization strategy in Modern French is left dislocation, which we have already seen becomes increasingly prominent from the Middle French period onwards: (108) a. Le garçon il attend devant la porte the boy he wait.3sg in-front-of the door ‘The boy, he’s waiting in front of the door’ (ModFr. Lambrecht 1981: 61) b. Moi, personne ne veut m’aider me nobody neg want.3sg me.cl-help.inf ‘Me, nobody wants to help me’ (ModFr. Delais-Roussarie, Doetjes, and Sleeman 2004: 502)

⁴⁵ See also Pohl (1984) on complement-initial cases analogous to (107), such as les fleurs j’aime, lit. ‘flowers, I like’. Despite being cited by Blanche-Benveniste (1997a: 21) as an area in need of study, there has been no detailed investigation of their distribution, which appears partly to be conditioned by region.

3.5 the left periphery in modern french

85

c. aux enfants, on leur donnait très peu to-the children one them.cl give.3sg.pst very little ‘to the children, we gave them very little’ (ModFr. Ashby 1988: 208) The literature in this area is vast (see De Cat 2009 for a book-length overview and analysis),⁴⁶ but some of the most important points to note are as follows: a wide range of constituents can be left-dislocated and resumed by a pronoun within the T-layer, but subjects appear more frequently in this construction than objects and other types of constituents;⁴⁷ there is clear evidence that the specific topic projection associated with left dislocation is recursive, with cases reported in the literature of three or more left-dislocated constituents (Lambrecht 1981: 58; Delais-Roussarie, Doetjes, and Sleeman 2004: 503; Rowlett 2007: 178–179; De Cat 2009: 139); although stigmatized within formal French schooling and frequently associated with lower registers of French and spoken language (Ashby 1988: 225–226), the licensing of left dislocation in a wide range of registers and across both written and spoken media is reported, albeit to a lesser extent (Riou and Hemforth 2015). Importantly, left dislocation is consistently associated with old information and topichood. (Barnes 1985; Lambrecht 1981; De Cat 2009: 77–92). Rowlett (2007: 174–175) discusses a range of arguments in support of this view, including the unacceptability of left-dislocating the answer to a ‘what happened?’ question: (109) Q: Qu’est-ce qui s’est what-be.3sg-it that refl.cl-be.3sg ‘What happened? A: Les voisins (*ils) the neighbours they mangé mon lapin eat.ptcp my rabbit ‘The neighbours, they ate my rabbit’

passé? happen.ptcp ont have.3pl

(ModFr.)

Although a debate about whether left dislocation is derived via movement of base-generation has been ongoing for decades, the near-consensus in the French literature is now that left dislocation is a case of base-generation ⁴⁶ For further discussion and references, see Le Bidois and Le Bidois (1938: 59–64), Harris (1978: 20– 23), Lambrecht (1981: 51–75), Barnes (1985), Ashby (1988), Cinque (1990), Fuchs and Marchello-Nizia (1998), Rowlett (2007: 172–182), and Haegeman (2006). ⁴⁷ This point is made quantitatively in several studies, including De Cat (2009: 211) on the distinction between subjects and non-subjects.

86

the left periphery

(Cinque 1990; Benincà 2001; Haegeman 2012: 104; De Cat 2009: 149–155), with Rowlett (2007:178) concluding ‘that LD is a base-generated configuration rather than the result of movement’. Similar conclusions apply to hanging topics (Delais-Roussarie, Doetjes, and Sleeman 2004: 503–506; Rowlett 2007: 178–179; De Cat 2009: 134–139),⁴⁸ which have been less widely discussed in the French literature, but evidence a higher base-generated position than that occupied by the left dislocate, and are more characteristic of colloquial and spoken language: (110) Ce métier on se déplace tous les jours this job one refl.cl move.3sg all the days ‘In this job you’re on the move every day’ (ModFr. Rowlett 2007: 178) Finally, note that in addition to left dislocation and hanging topic constructions within the frame-topic field, scene-setting adverbials and clauses are licensed in a high left-peripheral position, to the left of the projection hosting left dislocates: (111) Hier, les voisins, ils ont mangé mon lapin yesterday the neighbours they have.3pl eat.ptcp my rabbit ‘Yesterday, the neighbours, they ate my rabbit’ (ModFr.) To conclude, in contrast to the focus field, the frame-topic field is active in all varieties of French, although base-generation of hanging topics and left dislocates is more frequently associated with colloquial varieties. (112) [Frame TopicHT , AdverbialScene Setting /ClauseScene Setting [Force [Topic TopicLD …]]]

3.5.3 The embedded left periphery in Modern French As in all stages of Latin and French discussed so far, left-peripheral word order in embedded clauses is more restricted than in their matrix counterparts.⁴⁹ Thus, as is the case across other Modern Romance languages, embedded focalization is generally not found in Modern French. However, the phenomenon ⁴⁸ See De Cat (2009: 135–136), however, for a critical view that ‘there is in fact no reliable empirical base for a distinction between HTLD [SW —Hanging Topic Left Dislocation] and CLLD [SW—Clitic Left Dislocation] in spoken French’. ⁴⁹ See Lahousse (2003) and Chapter 5 for discussion of embedded verb–subject inversion, which I treat separately from focalization/topicalization.

3.5 the left periphery in modern french

87

of left dislocation is found in a range of embedded environments, including— as noted by De Cat (2009: 158–159)—various types of embedded clauses where matrix phenomena are not typically taken to hold (113a) alongside those where they are (113b):⁵⁰ (113) a. Tu veux que moi, je le dessine you want.2sg that me I it.cl draw.1sg.sbjv ‘Do you want me to draw it?’ (ModFr. De Cat 2009: 158) b. Ils ont dit que les they have.3pl say.ptcp that the myrtilles, ils les avaient toutes bilberries they them.cl have.3pl.pst all cueillies aujourd’hui pick.ptcp today ‘They said that they had picked all the bilberries today’ (ModFr. De Cat 2009: 160) Three further points are worth noting with reference to the makeup of the embedded left periphery. Firstly, in all varieties of French, recomplementation structures are absent, in contrast to a number of other Romance languages and historical varieties of French until the 17th century (see §3.4.6).⁵1 Secondly, and also in contrast to a number of other Romance varieties (Wanner 1981; Cocchi and Poletto 2007; Giorgi 2010: Chapter 3), complementizerdeletion is absent in Standard Modern French (Salvesen 2014: 16), although present in certain North American varieties:⁵2 (114) *Je pense il vient I think.1sg he come.3sg ‘I think (that) he’s coming’

(ModFr.)

Under the assumption that complementizer-deletion involves incorporation via raising of a complementizer from Fin to Force, just as in recomplementation structures, this could be seen as a piece of evidence that Modern French lacks Fin-to-Force movement, which we saw was evident until the classical period (§3.4.5). Rather, the standard language features a single position for the complementizer que ‘that’ in Force, as envisaged in Rizzi’s (1997) original schema. ⁵⁰ See Salvesen and Walkden (2017), however, for the proposal that so-called ‘volitive’ verbs are added to Hooper and Thompson’s (1973) classes permitting matrix-like phenomena. ⁵1 Roehrs and Labelle (2003) do, however, show such structures to be present in Child French. ⁵2 On the deletion of que in certain contexts, see Roberge and Rosen (1999) and references therein.

88

the left periphery

(115) [Force [Force que] [Topic TopicLD [Focus [Fin … [TP …]]]]] A second phenomenon to consider, attested in earlier texts, but widespread in colloquial varieties today, is generalization of que ‘that’ to a wide range of embedded clause-types that would feature in the standard language qui ‘who, that’ (116) or another C-element, exemplified here by parce que ‘because’ (117) (Bauche 1946: 92–93; Gadet 1997: 56–57; Blanche-Benveniste 1997b: 42; Rowlett 2007: 147–148). (116) J’ai une soeur qu’elle s’appelle Samia I-have.1sg a sister that-she refl.cl-call.3sg Samia ‘I have a sister called Samia’ (Colloquial French, Blanche-Benveniste 1997b: 42) (117) Il est pas venu qu’il est malade he be.3sg neg come.ptcp that-he be.3sg ill ‘He hasn’t come because he’s ill’ (Colloquial French, Rowlett 2007: 148) This phenomenon is evidence not only that Modern French typically makes use of the Force head for a range of complementizers but, furthermore, that analogical pressure is yielding a situation where only que can spell out this position in certain varieties.

3.5.4 Summary Overall, this section has shown that access to the left periphery in Modern French is qualitatively restricted in most varieties. Although focalization remains an option in conservative styles with a lexically restricted class of adverbials, adjectives, and PPs, many other varieties of French make extensive use the left periphery, but via a restricted range of base-generated left-dislocation or hanging-topic structures, the former also attested in embedded clauses. This general lack of access to the left periphery through movement can be seen as the ultimate sign that even ‘relic’ V2 structures are being lost in declarative clauses. The schema in (118) attempts to differentiate between a conservative system and a more colloquial one, typical also of spoken language, although it is not intended to endorse a binary distinction between the two:

3.6 summary and conclusions

(118)

89

a.

[Frame TopicHT , AdverbialScene Setting /ClauseScene Setting [Force [Force que/qui] [Topic TopicLD , [Focus Adverbials/AdjectivesFocus , PPFocus [Fin …]]]]] (ModFr.) b. [Frame TopicHT , AdverbialScene Setting /ClauseScene Setting [Force [Force que] [Topic TopicLD , [Focus [Fin …]]]]] (Colloquial French)

3.6 Summary and Conclusions 3.6.1 Summary This section has shown clearly that the left periphery sees a considerable degree of change in the history of French; this change is still ongoing today. Classical Latin makes extensive use of focalization and topicalization operations derived via movement—Edge-Fronting operations in the terms of Ledgeway (2018)— and their frequency increases in colloquial and late Latin, becoming entirely systematic in Old and Middle French, where merger of a left-peripheral constituent is necessary to satisfy the V2 constraint, which entails the presence of an Edge Feature on Fin or Force. In Renaissance, Classical, and Modern French, the loss of the V2 constraint means that topicalization and focalization are never obligatory but constitute a discourse-marked word-order option. However, fronting of a pragmatically salient constituent via internal merge, i.e. movement, progressively declines from the Renaissance onwards, such that certain varieties of French today systematically lack syntactic operations where constituents reach the left periphery via internal merge, instead featuring structures such as left dislocation, which are derived via base-generation. The structure of the embedded left periphery has also varied considerably in the history of the language, with the position targeted by complementizers and relativizers systematically conditioning such variation. We saw evidence that such elements typically target Fin in Classical Latin, significantly constraining the possibilities for pragmatically conditioned word-order variation. By contrast, Ledgeway (2017b) reports that the embedded left periphery is accessible to focalization and topicalization operations in late Latin, a state of affairs which continues into Early Old French, where we suggested that complementizers and relativizers are typically merged in Fin, but can merge in Force, yielding a substantial number of embedded V2 orders. This system is quite distinct to that found in Later Old French, where embedded topicalization and focalization are subject to far greater restrictions, with embedded verb-third

90

the left periphery

orders near entirely absent. We linked this to the fact that C-elements are typically base-generated in Fin and then move to Force, blocking access to the embedded left periphery; the overt reflex of such Fin-to-Force movement is seen in recomplementation structures which are attested until the mid-17th century. Matrix/embedded asymmetries are also less strict from Renaissance French onwards, which is suggestive of a high position for complementizers in the unmarked case, namely Force.

3.6.2 The making of the French left periphery We can consider the diachronic variation surveyed in this chapter in more detail by considering each of the left-peripheral fields in more detail. Starting with focus, we saw evidence that Classical Latin had an active focus field, which could encode new and contrastive information, alongside QPs, and reviewed evidence that an active focus field was an inherited characteristic from Proto-Indo-European (Hale 2017). Whilst the licensing of all these types of focus extends into late Latin and Early Old French—where foci, as constituents that can only reach their left-peripheral position via internal merge, may have played a vital role in the genesis of V2—new information and contrastive focus become particularly restricted from Later Old French onwards (Wolfe 2016a). Whilst QPs and a small number of syntactic categories are able to undergo focalization in Renaissance and Classical French, non-quantified focalized constituents increasingly constitute lexical exceptions, such as specific adjectives or adverbials. Turning to the contemporary language, we observe that QPs cannot consistently focalize in Modern French, although a small, predefined class of adverbials and adjectives are able to do so in formal registers. This situation in Modern French contrasts with Colloquial French which has lost fronted foci entirely. We should note that this state of affairs neatly maps on to the meso- > micro- > nano parametric approach to syntactic change outlined in Roberts (2019) and introduced in §1.2, and can thus be modelled on a parameter hierarchy as in Figure 3.2, where we see that the triggers for focusrelated movement to the left periphery become progressively more specialized before being lost from the system entirely: The variation concerning moved topics is strikingly similar, albeit with a distinct diachronic trajectory: whilst Latin of all periods features widespread topicalization, as do Old and Middle French, we have seen that in Renaissance and Classical French this kind of topicalization only obtains when the constituent is unambiguously discourse-old or ‘anaphoric’ (cf. Zaring

3.6 summary and conclusions

91

Can Foc attract all XPs to the left periphery? Yes No Latin Contrastive foci, QPs, and specific classes of XP? Early Old French Yes No Later Old French QPs and specific classes of XP [AdjP, AdvP, PP]? Middle French Yes Renaissance French Classical French

No Lexically defined AdjPs, AdvPs, PPs? Yes

Modern French

No Colloquial French

Fig. 3.2 A parameter hierarchy for focus movement

2011 and Chapter 6 for parallels with the loss of OV orders). In all but the most formal Modern French and Colloquial French, left dislocation is extremely widespread, and we have seen evidence that this form of topicalization via base-generation rises whilst the possibility for topicalization via movement falls, from the Middle French period onwards (Kroch 1989: 199). This variation is modelled in Figure 3.3. Can Top attract all XPs to the left periphery via internal merge? Yes Latin Early Old French Later Old French Middle French

No Only unambiguously discourse-old XPs? Yes Renaissance French Classical French

No Topicalization only via base-generation? Modern French Colloquial French

Fig. 3.3 A parameter hierarchy for topic movement

The important generalization emerging from the two schemas is that access to the Modern French left periphery via internal merge is heavily restricted and is restricted to an even greater extent in Colloquial French. In fact, if we consider the increasing preference for wh-in-situ in the colloquial language (Cheng 1991; Cheng and Rooryck 2000; Rowlett 2007: §5.7.1),⁵3 we see that ⁵3 Note that for some speakers, a wh-in-situ question is presuppositional in nature: tu vas où, lit. ‘you go where?’, ‘where are you going?’, presupposes that the hearer is going somewhere (Rowlett 2007: 197).

92

the left periphery

evidence of operator-movement of almost any kind is lacking. Rather, since the Middle French period two alternative strategies have taken hold to encode that constituents are topical on the one hand, or focal on the other: left dislocation in the former case and clefting in the latter (cf. 119) (Lambrecht 2001; Dufter 2008; Tailleur 2013). As we discuss further in §7.4, this gives French a markedly distinct syntactic footprint from many other Romance languages spoken today, where argument-fronting operations are widespread. In this respect, Colloquial French has become a true mirror image of Latin: the property of argument fronting to edge-positions is arguably one of Latin’s most salient features typologically (Ledgeway 2018), whereas fronting of any sort is absent in many French varieties. (119) a. C’est par ty qu’il n’est fait! it-be.3sg by you that-it neg-be.3sg do.ptcp ‘It’s because of you that it’s not done!’ (MidFr. Ysaÿe 509, 57, 14th–15th centuries, Muller 2003: 118) b. C’était un weekend à Paris qu’ il a it-be.3sg.pst a weekend in Paris that he have.3sg gagné win.ptcp ‘It was a weekend in Paris that he won’ (ModFr. Dufter 2008: 33) Note finally that the increasing lack of activation of the lowest left-peripheral projections in the C-layer also extends to subordinators base-generated within it. Whilst in Latin such elements are base-generated in Fin, both late Latin and Early Old French show evidence that they can also move higher to Force. In Later Old French strict clause-type asymmetries are indicative of Fin-to-Force movement becoming increasingly systematic, which eventually yields the Modern French system, where we suggested that such elements are base-generated in Force, accounting for the lack of null complementizers in most varieties and the absence of recomplementation structures (Roehrs and Labelle 2003). The evolution of French complementizers is therefore a classic case of ‘upwards’ reanalysis mediated by an intermediate movement phase, as is typical of the grammaticalization cases discussed by Roberts and Roussou (2002). This trajectory is summarized in Table 3.8. By way of conclusion, our central claim here is that an articulated leftperipheral structure is needed to properly understand the relevant facts for

3.6 summary and conclusions

93

Table 3.8 Subordinators from Latin to French Classical Latin

Base-generated in Fin

Late Latin and Early Old French

Base-generated in Fin, optional movement to Force Base-generated in Fin, systematic movement to Force Base-generated in Force

Later Old French Middle French to Modern and Colloquial French

Latin and French. However, a wide range of data suggest that the lower portion of the CP, Benincà and Poletto’s (2004) ‘operator layer’, where constituents reach projections within the topic and focus field via movement, has become increasingly restricted diachronically, with the most ‘advanced’ French varieties lacking access to these fields entirely. Furthermore, we have suggested that a range of C-elements may have undergone upwards reanalysis in the passage from Latin to Modern French.

4 Verb placement and verb movement 4.1 Introduction: verb movement in Romance and beyond Verb placement has held a fundamental role in cross-linguistic typological studies and,1 since the seminal work by Emonds (1976), Belletti (1990), and Pollock (1989), finite-verb movement has accrued a substantial theoretical literature in Romance.2 In this short introductory section, we will see an overview of some of the variation observable in verb-movement typologies cross-linguistically before dedicated sections deal with where Latin and French sit within the typology. Although we will discuss marked verb-movement options in Latin and French our introductory section focusses on unmarked placement in the first instance.

4.1.1 The V-to-T typology Although not the first to suggest that finite verbs might undergo some form of movement in certain languages (cf. in particular Emonds 1976), Pollock’s (1989) study was highly significant in accounting for notable differences between French on the one hand and English on the other, in terms of finite-verb movement into the inflectional layer. In this regard, consider the contrasting placement of the finite verb relative to negation, adverbials, and quantifiers in (1–3) (Pollock 1989: 367): (1)

*John likes not Mary Jean (n’)aime pas Marie Jean neg-like.3sg neg Marie ‘John doesn’t like Marie’

(ModFr.)

1 For an early typological approach to Romance word order, informed by Greenberg’s (1966) comparative work, see in particular Harris (1976b; 1978; 1984). 2 For critical overview and discussion, see Schifano (2015a; 2015b; 2018) and Roberts (2019: Chapter 5).

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0004

4.1 introduction: verb movement in romance and beyond

(2)

(3)

95

*John kisses often Mary Jean embrasse souvent Marie Jean kiss.3sg often Marie ‘John often kisses Mary’

(ModFr.)

My friends love all Mary Mes amis aiment tous Marie My friends love.3pl all Marie ‘My friends all love Mary’

(ModFr.)

Alongside Pollock’s work on French, a range of studies have claimed that other Romance languages, such as Italian (Belletti 1990; Zanuttini 1997; Schifano 2015b; Ledgeway In Press), Spanish (Zagona 1988: 148; Zagona 2001: 166; Zubizarreta 1998: 100; Schifano 2018: 67–69), Portuguese (Martins 1994; Tescari Neto 2012; Schifano 2018: 69–74), and Romanian (Nicolae 2014) also feature verb movement to I/T. As such, we can conclude that verb movement to the inflectional layer is an important unifying property of most Modern Romance languages in contrast to their medieval counterparts (Salvi 2004; Benincà 2004; Benincà 2006; Wolfe 2018a). (4)

Lui non ha più sempre vinto da allora he neg has.3sg any more always win.ptcp from then ‘Since then, he hasn’t won any more’ (Modern Italian, Ledgeway in Press: §2.1.1.2)

(5)

O João vê sempre este tipo de filmes the João watch.3sg always this type of movies ‘João always watches this type of movie’ (Modern Brazilian Portuguese, Schifano 2018: 70)

(6)

Părint¸ii mei nu merg des la bunici parents.def my neg go.3pl often at grandparents ‘My parents don’t often go to my grandparents’ (Modern Romanian, Nicolae 2019a: 14)

However, recent work by Schifano (2015a; 2015b; 2018) has shown the picture to be more nuanced than is conventionally assumed. Taking as her point of departure the empirically grounded cartographic assumption that there is not a single TP but rather a range of functional projections associated with tense, aspect, and mood (Cinque 1999; Cinque 2001; Cinque 2006; Cardinaletti 1997; Guasti and Rizzi 2002; Cardinaletti and Shlonsky 2004), demarcated by adverbials which spell out particular functional projections, she highlights that the

96

verb placement and verb movement

finite verb does not target a single uniform position across Romance but rather can reach a range of heights within the inflectional field of the clause. Thus, consider the contrasting behaviour of French and Spanish in (7, 8). We observe that the French finite verb precedes all adverbials in the Higher Adverb Space, but that the Spanish verb is preceded—in some cases optionally—by adverbs in the Lower Adverb Space such as ‘already’ but precedes bien ‘well’. (7)

Antoine connaît probablement /déjà / complètement / bien Antoine know.3sg probably already completely well

(8)

Marı´a probablemente / ya / completamente conoce bien Maria probably already completely know.3sg well

On the basis of this and a broad range of comparative data, Schifano (2018: 134) develops the ‘macrotypology’ in Table 4.1, under which broad classes of Romance varieties are classed as ‘high’, ‘clause-medial’, ‘low’, or ‘very low’. Table 4.1 Schifano’s typology of verb movement in Modern Romance High

French, Romanian

Clause-Medial

Sardinian, Northern Regional Italian, Central Regional Italian, Northern Italian Dialects European Portuguese, Southern Regional Italian, Southern Italian Dialects Spanish, Valencian Catalan

Low Very Low

Therefore, we see that ‘Romance V-to-T movement’ may be a useful shorthand to indicate that finite verbs—in contrast to a number of other languages—target positions within the inflectional layer. However, in formal terms, the hypothesis can be further refined to yield a picture of microvariation between closely related varieties. As we will now see, this insight can be extended in a novel fashion to verb movement in other layers of the clause.

4.1.2 The V-to-C typology As already introduced in §3.2.5, V2 systems are arguably the best-studied class of languages where finite-verb movement consistently reaches the left periphery. It is worth noting from the outset that the typology of these languages has been developed principally on the basis of Germanic data where the phenomenon is most salient (Haider and Prinzhorn 1986; Vikner 1995; Holmberg and Platzack 1995; Haegeman 1996; Mu¨ller 2004), and that early

4.1 introduction: verb movement in romance and beyond

97

accounts focussed in particular on German and Dutch (Williams 1974; Den Besten 1983; Koster 1975; Koster 1978). In both languages, finite verbs were assumed to target a single ‘C’ or ‘Comp’ head, which simultaneously accounts for why fronting of a non-subject triggers verb–subject inversion (10), but also why V-to-C movement is absent in the embedded domain, where the complementizer sits in the position usually targeted by the finite verb (11). (9)

(10)

(11)

Johann hat das Buch Johann have.3sg the book ‘Johann has bought the book’

gekauft buy.ptcp

Das Buch hat Johann gekauft the book have.3sg Johann buy.ptcp ‘Johann has bought the book’

(Modern German)

(Modern German)

*Ich glaube, dass das Buch hat Johann gekauft I believe.1sg that the book have.3sg Johann buy.ptcp ‘I believe that Johann bought the book’ (Modern German)

An important point is that the postverbal subject in such cases can be shown to occupy a position within the T-layer, such that ‘Germanic’ inversion is typically differentiated in the French and Romance literature from Romance inversion, where postverbal subjects are demonstrably in the vP (Adams 1987a; deBakker 1997; Hulk and Pollock 2001; Salvesen and Bech 2014; Poletto 2014: Chapter 1; Wolfe 2020b). Importantly, the presence of a postverbal subject in Spec-TP is thus a clear indicator that the finite verb is in the C-layer and not a position within the Tense-Aspect-Mood field. The seminal insight holds that V2 languages feature a prefield able to host a wide range of constituents, inversion structures, and some form of matrix/embedded asymmetry.3 However, the emerging microtypology of V-to-T movement (cf. §3.1.1) raises a significant question. If V-to-T is merely a cover term for a range of languages with quite distinct verb-movement properties, should the V-to-C typology not be extended in a similar fashion? This question is particularly pertinent in the light of the cartographic data we reviewed in §3.1, where we saw that the left periphery consists of an articulated 3 This last point is highly controversial. Despite earlier claims to the contrary (Fontana 1993; Lemieux and Dupuis 1995), Wolfe (2020c) suggests that there are no symmetrical V2 systems in Romance. The status of Icelandic and Yiddish is still debated (Ga¨rtner 2003; Ga¨rtner 2016; Bentzen et al. 2007a; Bentzen et al. 2007b; Wiklund et al. 2009; Diesing and Santorini 2020), but see Walkden and Booth (2020) for a critical review of the evidence presented for a symmetrical analysis of these languages.

98

verb placement and verb movement

range of projections, each hosted in fields associated with particular pragmatic and syntactic functions (Rizzi 1997; Rizzi 2001; Poletto 2000; Benincà 2001; Benincà and Poletto 2004; Benincà and Munaro 2010; Cruschina 2011; Haegeman 2012; Rizzi and Bocci 2017). The notion that the V2 typology was richer than originally envisaged was put forward most prominently by Poletto (2000; 2002) in her work on Raeto-Romance. Poletto (2002) highlights that the common cartographic assumption that verb movement in V2 languages targets Fin predicts that,⁴ all other things being equal, the full range of projections in the frame, topic, and focus fields should be accessible to constituents preceding the finite verb. This in turn would yield widespread V3 or greater orders. By contrast, if the original insight from the Germanic literature is maintained that the finite verb targets the very highest projection of the clause in V2 languages—Force in cartographic terms—only frame-setters or hanging topics should be able to precede a V2 clause. Poletto (2002) shows that within a single language—the Raeto-Romance variety spoken in San Leonardo—distinct clause-types can have different loci of V2, on either Fin or Force. Consider in this regard the data in (12), which show that a left-dislocated constituent cannot precede a V2-satisfier in a declarative but can in an interrogative. This leads Poletto to include that the locus of V2 varies across clause-types: the finite verb targets Fin in interrogatives, permitting a wide range of V3 triggers including left dislocates, whereas the verb targets Force in declaratives, with heavy restrictions on V3. (12) a. *De Giani CUN PIERO ai bel baié of Giani with Piero have.1sg=I already speak.ptcp b. De Giani, con che bai-la pa? of Giani with whom speak.3sg.pst-she.cl prt ‘With whom did she talk about John?’ (Raeto-Romance, San Leonardo, Poletto 2002: 231) Following Poletto’s (2000; 2002) work, the Fin/Force typology has been discussed with reference to Modern Germanic (Frascarelli and Hinterho¨lzl 2007; Walkden 2017), Early Germanic (Hinterho¨lzl and Petrova 2010; Walkden 2014; Walkden 2015; Haeberli, Pintzuk, and Taylor 2020),

⁴ Note that locating V2-related V-to-C movement in Fin is in keeping with earlier Germanic research where V2 was linked to verbal finiteness (Platzack 1985; Platzack 1986; De Haan and Weerman 1986; Vikner 1995; Holmberg and Platzack 1995).

4.1 introduction: verb movement in romance and beyond

99

and Early Celtic (Meelen 2020). With regard to Medieval Romance, Wolfe (2016b; 2018a; 2020c) has put forward the proposal that certain Romance varieties showing a V2 grammar were Force-V2 systems whilst others were Fin-V2. The former class patterns similarly to Modern Germanic in having a restricted class of V1 and V3 triggers, whilst the latter is more liberal in this regard and thus corresponds to the descriptive ‘relaxed V2’ label used by Benincà (1983b; 1995) to refer to the Medieval Romance languages in general. The important insight for our purposes is that earlier stages of Romance showed microvariation in their verb-movement properties which is parallel to the variation observed today (cf. §3.1.1), even if the locus of such variation has changed, with the notable exception of V2 Raeto-Romance (Poletto 2002; Casalicchio and Cognola 2020). We should note before concluding that the V-to-C typology is not restricted to V2 systems, even though they are the best-studied cases of left-peripheral verb movement cross-linguistically. Recent research on verb-initial languages has established that V1 orders can be derived either via V-to-C movement, as in Old Irish (Carnie, Harley, and Pyatt 1994; Carnie, Harley, and Pyatt 2000), Middle Welsh (Willis 1998), Old Sardinian (Wolfe 2015d), or V-to-T movement, as in Modern Irish and Welsh (Roberts 2005).⁵

4.1.3 V-in-situ So far, we have considered cases where the finite verb targets a clause-medial or left-peripheral position via head movement. However, there is an additional possibility, namely that the finite verb stays in situ or moves to a head within the extended vP projection. To take an example from outside Romance, Chinese has been argued to show no or very little verb movement (Huang, Li, and Li 2009), a property which Roberts (2019: §5.4) links to its general lack of inflectional morphology. Returning to more familiar ground, the hypothesis is particularly appealing for certain languages showing unmarked OV orders, such as Latin (cf. references in §6.2); following Kayne’s (1994) antisymmetric approach to syntax, under which head-final orders do not involve a sisterhood relation between head and complement but rather ‘rollup’ movement of the complement to a higher specifier position, the OV order would thus be derived via movement of ⁵ Note for completeness that V1 orders can also be derived via phrasal movement. See the contributions in Carnie and Guilfoyle (2000), alongside Carnie, Dooley-Collberg, and Harley (2005) for discussion.

100

verb placement and verb movement

the object to a vP specifier, with the finite verb remaining lower in the extended verbal projection. This proposal, as we shall see in detail in §3.2, has been put forward for Classical Latin syntax by Ledgeway (2012a).

4.2 Verb movement in Latin 4.2.1 The conservative V-in-situ system Latin, as already discussed in Chapter 2, is renowned for its word-order flexibility; this flexibility has often led to empirically and theoretically problematic assertions in the literature that the placement of the finite verb and other constituents is ‘free’ (Meillet 1903; Ku¨hner and Stegmann 1955; Ramsden 1963: 42).⁶ Although widespread displacement of constituents to leftperipheral positions does constitute a complicating factor in any analysis of Latin word order,⁷ we can nevertheless diagnose the unmarked position of the finite verb in Latin as final, which in formal terms corresponds to it remaining in V (Ledgeway 2012a: 238). There are several pieces of evidence for the claim that the finite verb remains within the VP in the unmarked case yielding an unmarked SOV order. First, as is the case cross-linguistically, Latin embedded clauses show markedly less word-order flexibility than their matrix counterparts (Linde 1923; Adams 1977: 69; Adams 2016: 321; Panhuis 1982: 117; Ostafin 1986: 165–169; Wanner 1987: 380; Polo 2005: 378; Ledgeway 2012a: 185); this can be taken as indicative of heavy restrictions on embedded topicalization and focalization, alongside restrictions on marked verb movement in the presence of a complementizer or relativizer.⁸ Therefore, an embedded SOV clause can be taken—all things being equal—as an instance of constituents other than the object occupying their base-generated positions without further instances of discourse-driven movement masking the underlying position of the verb and its arguments (cf. also Salvi 2004: 54): (13)

…[Fin [Fin si/cum/ut] [TP [vP Subject [VP Object [V VFin Object]]]]]

An additional argument comes from the behaviour of objects relative to finite verbs. Although the transition from an OV to a VO grammar is most likely ⁶ For two—more informative—views of relative freedom in Latin, see the analysis in Marouzeau (1938) and Panhuis (1982), both of whom, in different terms, stress the role of pragmatic factors. ⁷ For discussion of left-peripheral constituent fronting, see Ledgeway (2018) alongside extensive references in §3.2. ⁸ Recall here Danckaert’s (2012: 107) proposal that various subordinators occupy Fin in Classical Latin (cf. §3.2.4).

4.2 verb movement in latin

101

not a neat diachronic progression when viewed through the lens of a range of corpora (Pinkster 1991: 70; Ledgeway 2012a: 127–128; Danckaert 2017a: 110–112), it is nevertheless the case that V-final placement is the most common word order for much of the archaic and classical period, and particularly frequent in the earliest texts (Linde 1923: 154–158; Bauer 1995: 89–92; Clackson and Horrocks 2007: 106; Halla-Aho 2009: 129; Cabrillana 2011: 70–71). Indeed, the majority of reconstructions for Proto-Indo-European reconstruct Latin’s ancestor as an OV language (Lehmann 1974; Roberts 2007a: 364–365; Hock 2015; Ringe 2017: 64). Likewise, the end point of the OV > VO transition is established, insomuch as no Medieval Romance language has unmarked OV orders, with the notable exception of a few residual contexts which we review in §6.3. As such, the fact that fronting of the object was—initially at least—not triggered by a specific syntactic or pragmatic feature has to be captured formally. Following Kayne’s (1994) postulation of rollup movement, Ledgeway (2012a: 270) highlights that the most straightforward way of implementing this intuition is to assume that the conservative Latin system is one where the finite verb remains in situ in V and the complement raises to Spec-VP. If we couple this with the assumption that Latin did not feature a grammaticalized Spec-TP subject position (Danckaert 2017b), we are left with a schema such as the one below in (14) for the conservative Latin system, where only the object undergoes movement in the unmarked case, whereas the subject and finite verb remain within their first-merged positions within the thematic layer: (14)

…[FinP [TP [vP Subject [VP Object [V VFin Object]]]]]

4.2.2 The triggers for upwards reanalysis In §3.2.5, we noted a range of evidence that late Latin texts instantiate a form of V2 system (Ledgeway 2017b) and that this persists into the medieval period, yielding the well-studied V2 grammar of Medieval French (Thurneysen 1892; Adams 1987c; Roberts 1993; Vance 1997; Wolfe 2018a), which we deal with in detail in §4.3. Our question here is therefore the following: how did a grammar where the finite verb remains in the very lowest head position within the clausal hierarchy come, in the course of several centuries, to be reanalysed as one where finite verbs systematically reach the left periphery, that is to say the highest target of movement on the verb-movement typology sketched in §4.1?

102

verb placement and verb movement

A key insight is that, despite systematic V-to-C movement being a far later innovation, V1 orders which target the C-layer are an archaic feature of Latin syntax which is plausibly reconstructed as a feature of Proto-Indo-European (Lehmann 1974; Miller 1975: Chapter 2; Viti 2014: 81). Thus, Luraghi (1995) suggests, on the basis of a comparative analysis of several Indo-European languages, that V1 is plausibly reconstructed for Proto-Indo-European in cases of contrastive focus, presentatives, topic continuity, and imperatives. Unsurprisingly, therefore, several scholars highlight that although verb-initial orders become increasingly frequent during the period of Latinity, they should not be viewed as an innovation (Linde 1923: 158–168; Wanner 1987: 381– 382; Bauer 1995: 93–97; Bauer 2009: 272). Considering the left-peripheral schema outlined in §3.1, we can therefore pinpoint likely targets of finite-verb movement. In contexts like the example in (15), topic continuity is clearly encoded. Note in this regard comments by Bauer (2009: 277) that V1 in Classical Latin serves to ‘push ahead the narrative’ or indicate continuity of topic or theme (Kroll 1918: 114–117; Linde 1923: 160–162; Luraghi 1995: 367–369; Bolkestein 2000: 121; Devine and Stephens 2006: 156–159; Horrocks 2011: 131). A natural analysis of these structures, in keeping with similar analyses for a range of other languages,⁹ is that the finite verb raises to a topic-related head, possibly in the presence of a null (aboutness) topic (cf. Frascarelli and Hinterho¨lzl 2007: 289) (15)

Transfigitur scutum Pulloni … Avertit hic pierce.3sg.pass shield.acc Pullo.gen turn.3sg this.nom casus vaginam … Succurrit inimicus illi case.nom scabbard.acc run.3sg enemy.nom his.dat Vorenus Vorenus.nom ‘Pullo’s shield is pierced through …This event turns his scabbard aside … His enemy Vorenus runs up to him’ (Latin, Caes. B.G. 5.44)

In addition to V-to-Top movement, we can also postulate V-to-Foc movement in cases such as the following (Panhuis 1982: 145; Ostafin 1986: 159; Luraghi 1995: Chapter 3; Spevak 2004: 385; Polo 2005: 402; Bauer 2009: 278; Horrocks 2011: 132; Cabrillana 2011: 76; Wanner 1987: 236), where the entire content of the clause is rhematic. Recall from §3.2.1 that Salvi (2005: 441)

⁹ See, for example, Huang (1984), Frascarelli (2007), and Sigurdsson (1993; 2011) on the postulation of null topics in Germanic and Romance.

4.2 verb movement in latin

103

argues that V-to-Foc movement is initially in complementary distribution with left-peripheral fronting of a focal XP in Latin. (16)

a.

Crescebat interim urbs grow.3sg.pst meanwhile city.nom ‘Meanwhile the city was growing’ (Latin, Liv. 1.8.3) b. Erat vallis inter duas acies be.3sg.pst valley.nom between two lines.acc ‘There was a valley between the two battle lines’ (Latin, Caes. B.C. 2.34)

In addition to these very frequent contexts, verb-initial matrix clauses are also found extensively with imperatives, polar questions, and hortative subjunctives (Marouzeau 1938: 81–85; Wanner 1987: 381–382; Luraghi 1995: Chapter 1; Polo 2005: 399–400; Bauer 1995: 93–95; Bauer 2009: 277–278; Devine and Stephens 2006: 146–148; Spevak 2010: Chapter 5; Pinkster 2015b: 316, 348–360). Under the original left-peripheral map proposed by Rizzi (1997; 2001; 2004), we would interpret this as V-to-Force movement, though note that specific heads associated with polarity and mood have since been proposed in the literature.1⁰ Imperatives and interrogatives are particularly noteworthy, as recent work by Biberauer (2016; 2019) has drawn attention to the role that such structures may play in triggering reanalysis during language acquisition owing to their particular salience in the PLD to which children are exposed: (17)

a.

dic ergo si me amas peristasim declamationis tell.imp thus if me love.2sg subject.acc speech.gen tuae your.gen ‘If you love me, then let me know what your discourse was’ (Latin, Petron. Satyricon 32, 27–29)

b. Infirmas igitur tu acta C.Caesaris, viri invalidate.2sg then you acts.acc G.Caesar.gen man.gen fortissimi? very-valiant ‘Do you then invalidate the acts of Gaius Caesar, that most admirable citizen?’ (Latin, Cic. Dom. 39, Spevak 2010: 202) 1⁰ See, for example, Progovac (1994), Van Craenenbroeck (2010: Chapters 12–13), Moscati (2011), and Holmberg (2016: 36), amongst others, on the postulation of Pol(arity)P, which overlaps considerably with Laka’s (1990) ΣP.

104

verb placement and verb movement

Verb-medial orders, however, are not generally considered a conservative feature and Bauer (2009: 268), in particular, highlights that they are most associated with subliterary and later Latin texts which are recognized to show particularly innovative characteristics.11 These contexts, discussed in more detail in §4.2.2, include V-to-C movement in the presence of an initial ablative absolutive, conditional, or temporal subordinate clause (Kroll 1918: 117–118; Mo¨bitz 1924; Polo 2005: 400–401). Noting this construction’s distribution in colloquial texts, Polo (2005: 400) remarks that ‘the construction is documented with a certain frequency, and the properties it manifests seem to be reminiscent of Verb Second phenomena in Germanic languages’. (18)

Quod si hoc fecerit, eripiat Norbano that if it do.3sg.fut.sbjv steal.3sg.fut Norbanus.dat totum favorem all favour.acc ‘If it comes off, he’ll put Norbanus right out of the running’ (Latin, Petron. Satyricon 45, 10, Polo 2005: 400)

We should also recall here our discussion of the fronted Wackernagel elements in §3.2.1, which, following insights by Adams (1994b; 1994a) and Salvi (2004: Chapter 4), we take to be phrasal foci in Classical Latin. Their raising into the C-layer consistently triggers verb or auxiliary fronting, which we take to target Fin in line with Ledgeway (2012a: 192).12 Importantly, none of the verb movement triggered by initial clauses or initial foci in Wackernagel configurations is argued to be itself associated with a marked pragmatic value, which may have been significant in feeding a reanalysis where V-to-C movement was systematic and thus unmarked. An additional important point for us to note is the following: far from being an insignificant part of the Classical Latin syntactic system, verb movement to Foc(us), Top(ic) and Force was a consistent feature of even the archaic Latin grammar, that would have been highly salient to the acquirer. A more innovative aspect of the system—not associated with a particular pragmatic value—is the accidental V2 structure discussed immediately above, which plausibly features V-to-Fin movement. 11 On this point, see also Wanner (1987: 382), Herman (2000: 86), and Ledgeway (2012a: 64–68). 12 Note that if Ledgeway (2012a) is correct to reinterpret verb/auxiliary movement to the edges of cola as movement to phase-edges in formal terms, this generalization would also fit with postulating V-to-Fin movement. Despite there being no fully developed cartographic implementation of phasetheory (but see Roberts 2012a; Branigan 2020), Fin is a strong candidate as head of the middle phase, with Rizzi (1997) explicitly identifying in early work its role in looking ‘upwards’ to the discourse domain and ‘downwards’ to the inflectional domain.

4.2 verb movement in latin

105

4.2.3 V-to-C in late Latin The generalization of V-to-C movement and corresponding emergence of a V2 grammar has, since Benincà’s (1983b) early work on Medieval Romance, typically been considered a medieval development. However, recent work on the syntax of late Latin has suggested that a form of V2 system is already detectable in a number of post-classical texts (Salvi 2004: 107–111; Ledgeway 2012a: 153; Ledgeway 2017b; Wolfe 2015a; Antonelli 2015; Danckaert 2017a: 254–255).13 Most recently, Ledgeway (2017b) has presented a detailed corpus analysis of the Peregrinatio, a 4th–6th-century Christian Latin text. He notes that the text shows all the major hallmarks of a V2 grammar, namely a prefield not specified for subjects, a preference for second-position placement of the finite verb (Table 4.2), verb–subject inversion structures (cf. also Bauer 2009: 275), and a clear matrix/embedded asymmetry. Table 4.2 Ledgeway’s (2017) analysis of verb placement in the Peregrinatio

Matrix Embedded

V1

V2

V3

V4

V5

V6

V7

V8

Total

192 (15.9%) 208 (41.9%)

489 (40.4%) 170 (34.3%)

274 (22.7%) 78 (15.8%)

161 (13.3%) 24 (4.8%)

67 (5.5%) 13 (2.6%)

20 (1.6%) 2 (0.4%)

4 (0.4%) 1 (0.2%)

2 (0.2%) 0 (0.0%)

1209 (100.0%) 496 (100.0%)

However, as noted in §3.2.5, Ledgeway (2017b) does not consider constituent-fronting obligatory in the text, which may suggest that the generalization of V-to-Fin movement preceded the generalization of optional phrasal movement triggers to an EF on Fin. We would thus consider a late Latin V1 clause such as (19) as having the structure in (20): (19)

Praedicant etiam omnes presbyteri preach.3pl also all.nom priests.nom ‘All the priests also preach’ (Latin, Peregrinatio 26, 1)

(20)

… [FinP [Fin Praedicant] [TP … etiam … [vP omnes presbyteri]]]

As with the earliest Old French texts (cf. §4.3.1), weak pronouns are typically enclitic in such contexts and in the presence of an initial first-merged 13 Antonelli (2015) suggests that V-to-C movement is a consistent feature of wh-interrogatives in late Latin texts.

106

verb placement and verb movement

topic (21), in contrast to cases where an unambiguously moved focal constituent occupies the first position of the V2 clause (22) (Ledgeway 2017b: 189):1⁴ (21)

a.

Et ait nobis sanctus episcopus and say.3sg us.dat holy.nom bishop.nom ‘And the holy bishop tells us …’ (Latin, Peregrinatio 20, 4, Ledgeway 2017b: 190) b. Et at ubi perdicti fuerint and to where recit.prf.ptcp be.3pl.fut.prf iuxta consuetudinem lebat se according custom.acc raise.3sg self.acc episcopus bishop.nom ‘And when all these have been recited according to custom, the bishop rises’ (Latin, Peregrinatio 24, 5, Ledgeway 2017b: 190)

(22)

et adhuc nobis superabant milia tria and still us.dat remain.3pl.pst miles.nom three.nom ‘we still had three miles to cover’ (Latin, Peregrinatio 4, 5, Ledgeway 2017b: 190)

Turning to embedded clauses, Ledgeway (2017b: 198) suggests that V-to-T movement is standard, whilst embedded V-to-Fin movement can obtain in embedded root environments (cf. 23), yielding the high proportion of verbinitial orders we see in Table 4.2. Note that this embedded access to the C-layer through verb movement strongly suggests that complementizers and relativizers are not exclusively base-generated in Fin in late Latin texts in contrast to Danckaert’s (2012: 107) proposal for Classical Latin (cf. also discussion in §3.2.4). (23)

posteaquam scripserat Aggarus rex ad after-that write.3sg.pluperf Abgar.nom king.nom to Dominum lord.acc ‘after that King Abgar had written to the Lord’ (Latin, Peregrinatio 19, 8, Ledgeway 2017b: 194)

1⁴ For detailed discussion of the Tobler-Mussafia system of Old French and Medieval Romance, see Benincà (2004; 2006) and Labelle and Hirschbu¨hler (2005) alongside the discussion in §4.3.1–2.

4.3 verb movement in old french

107

Overall, we have seen that whilst one ‘half ’ of the V2 constraint (cf. Roberts 2005: 123), namely systematic constituent fronting to the left periphery, may not have emerged in late Latin texts (see §3.2.3), the other half, namely finiteverb movement to Fin, evidently has. We will now see that this constitutes a major point of continuity in the verb-movement typology between late Latin and Early Old French.

4.3 Verb movement in Old French 4.3.1 V-to-C movement and V2 in Old French The status of Old French as a V2 language is probably the single most controversial area in historical French syntax. The V2 constraint was first described in atheoretical terms as early as Thurneysen (1892) and was a central topic of enquiry in early diachronic generative research (Adams 1987a; Adams 1987b; Adams 1987c; Adams 1988; Vance 1987; Vance 1988; Vance 1993; Vance 1995; Vance 1997; Dupuis 1988; Dupuis 1989; Hirschbu¨hler 1990; Hirschbu¨hler 1995; Roberts 1993; Lemieux and Dupuis 1995). However, in the last two decades a body of research has emerged questioning the basic idea that Old French was a form of V2 system with V-to-C movement, with Kaiser (2002) the earliest significant intervention in this area.1⁵ This debate raises a fundamental issue in the internal history of French in terms of where the medieval language sits within the verb-movement typology set out in §4.1, but also a wider cross-linguistic issue as to how the syntactic status of a textually attested language can be diagnosed. Since we have already seen some of the evidence for a V2 analysis of Old French in §3.3, we consider here some of the arguments against V2 which have been put forward in the literature, before evaluating their validity. 4.3.1.1 Arguments against C-V2 The first and most prominent line of argumentation against assuming a V2 analysis comes from allegedly linear ‘deviations’ from a V2 order in Old French. Supposedly inspired by classic works on Germanic V2 (e.g. Koster 1975; Den Besten 1983), these works view V2 as a strict linear ordering constraint, which essentially rules out V1 or V3* orders as incompatible with the V2 nature of the grammar (Kaiser 2002: 134; Kaiser 1⁵ Though see also Lemieux and Dupuis (1995) on the notion that Old French was an IP-V2 system, with V-to-I movement. I return to this issue later in this section.

108

verb placement and verb movement

2009: 140; Rinke and Meisel 2009: 101; Rinke and Elsig 2010: 2566; Zimmermann 2014). Zimmermann (2014: 56), for example, states that ‘V2 languages distinguish themselves from non-V2 languages by the strictness of the requirement that the finite verb directly follow one and only one sentence-initial constituent’. Thus, under this analysis, a V3 clause such as (24) is a clear piece of evidence against the V2 hypothesis and therefore V-to-C movement: (24)

Quant la messe fu dite, li dux when the mass be.3sg.pst say.ptcp the duke manda por les messages send.3sg.pst for the messengers ‘When the mass was performed, the Duke sent for the messengers …’ (OFr. Villehardouin 26, 13th century)

Whilst objections around verb-placement ‘deviations’ focus on what is found in the medieval texts, another argument against V2 focusses on what is allegedly not found. Considering the prevalence of verb–subject inversion structures, Kaiser (2002: 134) highlights what he views as the rarity of XPNon-Subject -V-S orders in Old French, based both on his own analysis of certain texts and a critique of the data put forward in Roberts (1993). Turning to more qualitative considerations, Rinke and Meisel (2009: 126) suggest that ‘Old French inversion resembles inversion in Romance null-subject languages with respect to the structural position of the subject and its information structural interpretation’. The central idea for these scholars is therefore that a fundamental V2 correlate—inversion—is not found either with sufficient frequency or with the correct characteristics for the overall grammar to be properly termed ‘V2’. An additional point, which has been raised in general terms and for other Medieval Romance varieties (Cruschina and Sitaridou 2009; Sitaridou 2011: 163; Rinke 2009: 312), concerns the purported diachronic implausibility of a V2 stage. The basic idea here would be that the shift from SOV, to V2, to SVO is an ‘unattested evolutionary path’ (Cruschina and Sitaridou 2009). We should also note a final hypothesis before evaluating these claims, namely that French was a form of V2 system, but one where finite-verb movement targeted not the C-layer but the T-layer; this TP or IP-V2 hypothesis was put forward by Lemieux and Dupuis (1995), inspired by similar approaches to Germanic systems which were argued to lack clause-type asymmetries in their

4.3 verb movement in old french

109

V2 licensing (Santorini 1989; Diesing 1990; Ro¨gnvaldsson and Thra´insson 1990; Iatridou and Kroch 1992; Pintzuk 1999). 4.3.1.2 Arguments for C-V2 Against this background, we now address whether these arguments undermine the V2 hypothesis whilst also briefly revisiting why the V2 hypothesis has been advanced for French in the first place. In many cases we will see that these original pro-V2 arguments still stand in spite of the criticism of the approach and have—for the most part—not been directly addressed by those advocating a V-to-T analysis. The idea that the unmarked position for the finite verb is second for much of the medieval period is not purely a generative idea and is one that has been advanced by those working in a range of descriptive, functional, and formal frameworks, where it is noted both that the verb tends to appear in second position of the clause—as we see is the case in the 13th-century text La Queste in Table 4.3—and that the preverbal field is not a specialized subject position but one that can host a range of constituents that are flatly ungrammatical in the modern language (Thurneysen 1892; Foulet 1919: 306–332; Wartburg 1958: 103; Moignet 1973: 357; Skårup 1975: 9–69; Vanelli, Renzi, and Benincà 1986: §4.1; Adams 1987c: 4–5; Jensen 1990; Roberts 1993: 85–87; Benincà 1995: 333; Benincà 2004: 267–270; Hulk and van Kemenade 1995: 235–236; Vance 1997: 43–47; Mathieu 2012: 327; Salvesen 2013: 135–136; Steiner 2014; Wolfe 2018a: Chapter 5; Ledgeway Forthcoming). Table 4.3 Matrix verb placement in La Queste

Transitive/Unergative Athematic Raising Total Percentage

V1

V2

V3

V4

Total

0 0 0 0 0.0%

250 221 4 475 75.2%

76 79 0 155 24.5%

0 2 0 2 0.3%

632 100.0%

Looking at the prefield, we saw, for example, in §3.5 that fronting of an object to initial position without clitic resumption is flatly ungrammatical in most contexts in Modern French despite this OV(S) configuration being described as a major word-order pattern in a range of handbooks of Old French (Skårup 1975: Chapter 4; Jensen 1990: 534–535; Marchello-Nizia 1995; Buridant 2000: 744–746; Marchello-Nizia and Prévost 2020: 1146–1149). Old French patterns straightforwardly with V2 systems in this aspect of its syntax (Lightfoot 1995:

110

verb placement and verb movement

40; Vikner 1995: 41; Frey 2004; Holmberg 2015; Woods and Wolfe 2020), and sets itself apart from SVO systems. Furthermore, as shown in §3.3.1, XPNon-Subject -V-(S) orders cannot be dismissed as an exceptional word-order type produced by an SVO system in the medieval period: when quantitatively analysed their occurrence either matches or surpasses similar data available for uncontroversial V2 systems studied today. Despite these facts, should we—as Kaiser (2002) and others suggest— dismiss the V2 account in light of V1 and V3* orders which appear to deviate from surface V2? Empirically the answer is straightforwardly negative; V1 and V3* orders are extensively found across a range of V2 systems in other historical Indo-European languages and a growing literature shows that—away from some of the best-studied V2 systems—V1 and V3* orders are a core feature of the grammar of a number of V2 languages spoken today.1⁶ In a recent study, for example, Haegeman and Greco (2018) show that certain varieties of West Flemish show near-identical patterns of V3-licensing to what we described in §3.3.3 for Later Old French, where a range of scene-setting clauses and adverbials in the frame field can trigger V3. In theoretical terms, this empirical finding is unsurprising; in an articulated left periphery, no independent property of Universal Grammar rules out multiple constituents occurring preverbally if the finite verb targets a low left-peripheral head. Furthermore, null left-peripheral constituents have been postulated since at least the 1980s (cf. Huang 1984 among many others), so nothing a priori prevents a null constituent satisfying the V2-related EF and yielding a V1 surface order. The structure of the prefield also has to be considered alongside the postverbal field, in particular the licensing of postverbal subjects. The fact that Old French licenses widespread inversion structures has long been noted and has held a central role in studies of the V2 property (Tobler 1875; Meyer-Lu¨bke 1889; Foulet 1928; Benincà 1983a: 195; Vanelli, Renzi, and Benincà 1986; Adams 1987c; Jensen 1990: 388–390; Roberts 1993: 56; Vance 1995: 177; de Bakker 1997; Vance, Donaldson, and Steiner 2009: 313–316). However—as we will discuss in detail in §5.3—one important point that anti-V2 studies have drawn on is the potentially ambiguous status of some inversion structures, which could entail either V-to-T movement and a vP-internal subject or V-toC movement and a Spec-TP subject. Of particular importance, therefore, are 1⁶ See, amongst many others, Den Besten (1983: 61–63), Huang (1984: 546–549), Cardinaletti (1990), Sigurdsson (1993), Zwart (1993: Chapters 2, 5), Eythorsson (1995), Tomaselli (1995), Kroch and Taylor (2000: 363–367, 378–385), Faarlund (2004: 192–193, 231–234), Ferraresi (2005), Boeckx and Grohmann (2005), Petrova (2006; 2012), Westergaard (2009: 94–98), Hinterho¨lzl and Petrova (2010), Salvesen (2013: 144–154), Walkden (2013b; 2014: Chapter 5; 2015; 2017), Holmberg (2015: 201; 2020), Hsu (2017), Casalicchio and Cognola (2020), Meelen (2020), and Greco and Haegeman (2020).

4.3 verb movement in old french

111

cases like the following—known as Germanic inversion—where V-to-C movement is unambiguous owing to the presence of a past participle or infinitive demonstrating that the postverbal subject is in Spec-TP (Adams 1987c: 4; Roberts 1993: §2.2; Vance 1997: 78–79; de Bakker 1997; Salvesen and Bech 2014; Wolfe 2020b): (25)

a.

Par tantes teres ad sun cors over so-many lands have.3sg his body traveillet suffer.ptcp ‘His body has suffered over so many lands’ (OFr. Roland 540, 12th century) b. Et quant il furent ajosté ensemble, and when they be.3pl joust.ptcp together si furent li noir vaincuz si be.3pl the blacks defeated ‘And when they met in the jousting, the blacks were defeated’ (OFr. La Queste 182, 40, 13th century) c. Et enqui ot Guillelme de Chanlite and there have.3sg Guillelme de Chanlite brisié le braz d’une pierre break.ptcp the arm of-a stone ‘and there a stone broke Guillelme de Chanlite’s arm’ (OFr. Villehardouin, 167, 13th century)

There are good reasons to treat with heavy scepticism any criticism of an alleged ‘insufficiently robust’ attestation of inversion. Firstly, any comparison put forward with Modern Germanic regarding absolute frequency of overt postverbal subjects fails to capture the fact that Old French was a form of null-subject system (Adams 1987c; Roberts 1993; Vance 1997; discussion and references in §5.3) and as such overt postverbal subjects are predicted to be less frequent than in the Modern Germanic languages which are not full nullsubject systems (Biberauer 2010a). Secondly, our analysis in §5.3 shows that in cases of verb–subject inversion, unambiguous Spec-TP subjects outnumber vP-internal subjects in all but the very latest Middle French texts. As such, the most frequent inversion structure is that which is indicative of V-to-C movement. A final argument in favour of an analysis of Old French as featuring V-toC movement is that of matrix/embedded asymmetries. Note that—all things being equal—either an SVO system with V-to-T movement or a T-V2 system

112

verb placement and verb movement

(Lemieux and Dupuis 1995) would make the prediction that the prefield and the verb-placement properties of matrix (26a) and embedded clauses (26b) are broadly similar: (26)

a. …[FinP [TP Subject [T VFin ] [vP …Object]]] b. …[FinP [Fin que][TP Subject [T VFin ] [vP …Object]]]

We have already seen extensive evidence in §3.3, however, that this is not the case. Recall that whilst the prefield in matrix clauses in Old French can host a wide range of constituents, often alongside verb–subject inversion when a nonsubject is fronted, this is not the case in the embedded domain where clauses are overwhelmingly SVO (cf. §3.3.5 and Adams 1987c; Adams 1987b; Roberts 1993: Chapter 2; Vance 1997: Chapter 4; Wolfe 2018a: Chapter 5; Salvesen and Walkden 2017; Ledgeway Forthcoming). This relatively strict SVO ordering is shown in Table 4.4, which shows complement-clause verb placement in La Queste; we see that the finite verb is only ever in second position and recall also from Table 3.2 that the preverbal constituent is a subject in 83.8% of the sample. Table 4.4 Embedded verb placement in La Queste

Transitive/Unergative Athematic Raising Total Percentage

V1

V2

V3

V4

Total

0 0 0 0 0.0%

72 70 0 142 100.0%

0 0 0 0 0.0%

0 0 0 0 0.0%

142 100.0%

This sharp clause-type asymmetry is exactly reminiscent of what we find in a variety of other V2 systems (Den Besten 1983; Koster 1975; Vikner 1995; Holmberg and Platzack 1995; Holmberg 2015) and further strengthens the V2 hypothesis. Indeed, the fact that when embedded V2 and thus embedded V-toFin movement does obtain, it does so in precisely the same environments one finds embedded V2 in Germanic (cf. §3.3.5 and Vikner 1995), and is a further piece of supporting evidence that the systems are underlyingly similar. By way of contrast, it would receive no uniform explanation if we assumed that verb movement symmetrically targeted the T-layer in both matrix and embedded clauses. Despite these arguments in favour of a V2 analysis of Old French, we have noted a final objection, specifically that the emergence of V-to-C movement

4.3 verb movement in old french

113

and V2 in the history of French and Medieval Romance is diachronically implausible. Are such objections strong enough to lead us to cast aside the evidence for V2? Following the analysis in Wolfe (2015a; 2016a; 2018a: 33– 34) I suggest not. For an acquirer to reanalyse the primary linguistic data to which they were exposed as consistent with a full V2 grammar, two innovations are necessary: the postulation of an EF on a C-related head and also a systematic trigger for verb movement. In the former case, we have already seen that a range of focalization and topicalization operations occur in Latin (see §§3.2.2–3.2.3) and that a subclass of these are ‘accidental V2’ structures, insomuch as they entail phrasal movement and V-to-C movement co-occurring, albeit in specific syntactic contexts in a non-V2 grammar. In the latter case, i.e. verb movement, we have already seen that verb movement to a range of left-peripheral heads is a Classical Latin feature which may have been inherited from Proto-Indo-European. If Ledgeway’s (2017b) analysis of late Latin is correct, as discussed in §4.2.2, evidence of the reanalysis of discoursemarked or structurally specific verb movement triggers as unmarked V-to-Fin movement is already present in the Latin corpus. As such, far from being implausible, the two necessary reanalyses instantiate the very widespread pattern of change where a previously marked syntactic structure is reanalysed as unmarked; this scenario is precisely what is assumed to be responsible for the emergence of V2 in Celtic (Meelen 2020) and Germanic (Walkden 2014: 91–92, 110) and, as we discuss further in §4.5, is an instance of Feature Economy, where the acquirer postulates the minimum number of syntactic features consistent with the input they receive (Roberts and Roussou 2002; Van Gelderen 2009a). As such, I conclude that despite criticisms of it, the V2 hypothesis for Old French has a number of strengths and accounts well for the textual evidence available from the medieval period. We will now see, however, that the precise target of verb movement within the left periphery may not remain constant within the Old French period.

4.3.2 From Fin-V2 to Force-V2 in Old French Whilst we have already seen that there are good reasons to maintain Thurneysen’s (1892) proposal that Old French was V2, it will be suggested in this section that the phenomenon is not entirely homogeneous. Rather, as was necessary when considering the general structure of the left periphery in §3.3.6, we will see evidence that unmarked verb movement originally targets Fin, but

114

verb placement and verb movement

that this is reanalysed as V-to-Force movement at the beginning of the 13th century. Drawing on Poletto’s (2000; 2002) Fin/Force-V2 typology outlined in §4.1.2, an examination of Early Old French texts offers convincing evidence that the target of verb movement was originally Fin; this is an unsurprising finding as we saw in §4.2.2 that Fin is also the target of unmarked verb movement in late Latin and V-to-Fin movement is argued to be a universal feature of Early Medieval Romance syntax in Wolfe (2016a). The main piece of evidence for the V-to-Fin hypothesis is the licensing of V4* orders in Early Old French, which frequently entail co-occurrence of a topic and focus. Such structures are found across a range of texts and are indicative of a rich left-peripheral structure to the left of the finite verb (Benincà 1995: 329; Rouveret 2004: 189–190; Labelle 2007: 296–303; Salvi 2012: 105; Wolfe 2016a: 468): (27)

E pur çó que cist lignages numéément and since that these lineages named dout si le servise Deu celebrer, besuinz should.3sg thus the service God celebrate.inf necessary fud ke be.3sg.pst that ‘Since these lineages had to therefore celebrate the service of God, it was necessary that …’ (OFr. QLR 1, 1, 12th century)

A further piece of evidence comes from the distribution of si, which we discussed in §3.3.4. Although si’s distribution in later texts is far more restricted, in Early Old French texts it can be preceded by a range of argumental XPs in the topic-focus fields (28) and can also be preceded by more than one constituent (29) (Marchello-Nizia 1985: 7, 158; Fleischman 1991: 265–266; Wolfe 2018b: 347–348). This pattern is, in fact, exemplified in the very earliest Gallo-Romance text, the Strasbourg Oaths (30). (28)

Reis Vivien si succuras en Imphe King Vivien si help.2sg.fut in Imphe ‘you will help King Vivien there in Imphe’ (OFr. Roland 3995–3996, 12th century)

(29)

E puis les cors des barons si unt pris and then the bodies of-the barons si have.3pl take.ptcp ‘Then they take the bodies of those barons’ (OFr. Roland 2967, 12th century)

4.3 verb movement in old french

(30)

115

D’ist di in avant, in quant Deus savir from-this day in forward in so-far God knowledge et podir me dunat si salvari eo … and power me.cl give.3sg si support.1sg.fut I ‘from this day forward, insofar as God gives me knowledge and power, I will support …’ (OFr. Strasbourg Oaths, 9th–10th century)

If si is merged as a last-resort strategy to satisfy the V2-related head’s EF (Salvesen 2013; Wolfe 2016a), it follows that verb movement must therefore target a low left-peripheral position in order that si can be preceded by multiple constituents in the frame, topic, and focus fields: (31)

[Frame Scene-Setter [Force [Topic Topic [Focus Focus [Fin siEarly Old French [Fin VFin ][TP …]]]]]]

Furthermore, assuming a low locus of V2 in Early Old French Accounts for another property of the language at this stage, namely the availability of V1 orders (32) (Roberts 1993: 96–97; Salvi 2012: 106; Zimmermann 2014: 164–176). In line with the intuition that V1 clauses in V2 languages feature a null topic (Huang 1984; Cardinaletti 2010; Walkden 2013b; Salvi 2012; Poletto 2014: Chapter 1; Poletto 2020), Wolfe proposes that such clauses in Early Old French feature a variant of the null pronoun, pro, which raises to the topic field and constitutes a null aboutness topic in the terms of Frascarelli and Hinterho¨lzl (2007). If V-to-Top(ic) raising is necessary to license such structures (Wolfe 2016b), the locus of V2 cannot be higher than the topic field. (32)

Vint en Bethléém come.3sg in Bethlehem ‘He came to Bethlehem’

(OFr. QLR 1, 58, 12th century)

Finally, an articulated left periphery to the left of the finite verb readily accounts for the behaviour of object clitics in Early Old French. Whilst proclisis is already the norm in V2 clauses in the earliest texts, this is not the case with a small class of left-dislocated topics (33) or in V1 clauses (34) (Skårup 1975; Rouveret 2004: 193; Labelle and Hirschbu¨hler 2005: 60–69). Under Benincà’s (2004; 2006) implementation of the Tobler-Mussafia law—adopted in §4.2.2 for late Latin—we can interpret the proclisis vs. enclisis distinction as a reflex of an XP’s position either in the high frame-topic field (enclisis) or the low topic-focus field (proclisis). Crucially this account rests on a low position for the finite verb and V2 bottleneck, as highlighted by Rouveret (2004).

116

verb placement and verb movement

(33) Cum il vint a únes loges a´ pasturs en when he come.3sg.pst to one hut of shepherd in cel chemı´n, truvad i les freres Achazı´e this path find.3sg.pst there.cl the brothers Achazie ‘When he arrived at the shepherd’s hut on this path, he found the Achazie brothers there’ (OFr. QLR 195, 13, 12th century, Rouveret 2004: 197) (34)

Vait s’en li pople go.3sg refl.cl=part.cl the people ‘The people go away’ (OFr. Alexis, 71, 1, 11th century)

To summarize, a number of crucial assumptions about the matrix-clause syntax of Early Old French rest upon the assumption that verb movement and the locus of V2 effects is low left-peripheral head, Fin, which is therefore a point of continuity with late Latin. There is an increasingly prominent analysis of Old French syntax, however, that the syntactic system instantiated in 13th-century prose texts is quite markedly distinct to what we observe in earlier centuries. Following the analysis of the clitic system in Rouveret (2004), Wolfe (2015b; 2016b; 2016a; 2018a: Chapter 5; 2018b) suggests that the changes observed are indicative of the locus of verb movement and V2 effects having been reanalysed upwards to Force. In contrast to the system described above for Early Old French (cf. also §§3.3.2–3.3.3), this is reflected in the makeup of the Later Old French left periphery, where the prefield typically hosts a single constituent, optionally accompanied by a base-generated scene-setting clause or adverbial in the frame field (35) (Roberts 1993: 144; Vance 1997: 61–2; Salvesen 2013; Wolfe 2016a). (35)

et neporec tant a alé que … and nevertheless such have.3sg go.ptcp that ‘nevertheless he has gone so far that …’ (OFr. La Queste 77, 18, 13th century)

These restrictions on what can occur in the prefield also apply to the syntax of si: in 13th-century prose the particle is typically not preceded by argumental XPs—which we take to raise to the low topic-focus fields (cf. also Wolfe 2018b: 352–354)—but rather by the same base-generated scene-setters which constitute V3-triggers elsewhere (36). (36)

puis si se departirent then si refl.cl leave.3pl.pst ‘Then they departed’

(OFr. Clari 4, 2, 13th century)

4.3 verb movement in old french

117

Thus, the proposal is put forward in Wolfe (2018b; 2020a) that the most heavily grammaticalized version of si found in 13th-century prose texts like La Queste can only be preceded by base-generated constituents in the frame field (cf. 35). This proposal follows logically from the notion that si is a mechanism of satisfying V2 (Ledgeway 2008; Ledgeway Forthcoming; Salvesen 2013), the locus of which has changed in the passage from Early Old French to Later Old French. (37)

[Frame Scene-Setter [Force siLater Old French [Force VFin ] …]]

As we discuss in §4.2.3, there is clear evidence that V1 structures become increasingly rare after the end of the 12th century and are absent entirely in certain texts (Skårup 1975: 291; Marchello-Nizia 1980: 331; Vance 1997: 32; Rouveret 2004: 193–195; Labelle and Hirschbu¨hler 2005: 66; Simonenko and Hirschbu¨hler 2012). If a significant number of V1 structures entail the licensing of proTop and verb movement to the topic field, which no longer constitutes part of the prefield in Later Old French, this is unsurprising. Note that additional evidence that this analysis is along the right lines comes from a number of Modern Germanic languages, where the finite verb plausibly targets Force (Brandner 2004; Frascarelli and Hinterho¨lzl 2007; Roberts 2012a) and preverbal null topics of the type found in Early Medieval Romance and Early Germanic are highly restricted or flatly ungrammatical (Jaeggli and Safir 1989; Sigurdsson 1993; Biberauer 2010a). Finally, note changes in the behaviour of object clitics in Later Old French. With the exception of a small number of enclitic V1-contexts we discuss in §4.3.3, proclisis is generalized in V2 declaratives in Later Old French (38) (Dardel and de Kok 1996; Rouveret 2004; Labelle and Hirschbu¨hler 2005). If we assume that clitic placement is sensitive to the position of a preverbal XP in the prefield this falls out naturally from the Force-V2 analysis as all V2 satisfiers occupy the same position, namely Spec-ForceP, and will thus trigger proclisis due to their adjacency with the finite verb in Force (39). (38)

Li doi autre s’en fuirent the two others refl.cl=part.cl flee.3pl.pst ‘The two others fled’ (OFr. Clari 21, 13th century)

(39)

[Frame (Scene-Setter) [Force XPV2 [Force Cl + VFin ] …]]

Overall, I have suggested here that the often-noted strict V2 stage of Later Old French can be reformulated as a phase in the language’s history where the finite

118

verb placement and verb movement

verb consistently targets a very high head in the left periphery, Force. This accounts for a number of changes in the structure of the prefield, si, the licensing of null subjects and topics, and the generalization of unmarked proclisis.

4.3.3 Marked verb placement in Old French Although the position of the finite verb is overwhelmingly second in Old French, particularly in later texts, a range of marked placement options also existed, motivated by syntactic or pragmatic factors. As Table 4.3 shows, V3* orders are frequent across all Old French texts. Having already considered their distribution in detail in §4.2.3, we simply note here that they can be derived either via the simultaneous merger of a framesetter, topic, and focus in Early Old French (Wolfe 2016a: 467–470) or via a syntactically restricted class of clauses and adverbials in Later Old French (Roberts 1993: 144; Vance 1997: 61–62). The most important point for us to note is that in both the Early Old French (40) and the Later Old French systems (41), there is no reason to believe that the V2/V3 alternation is the result of a difference in the height of verb movement, but rather that it falls out from a distinction in which phrasal constituents are merged in the left periphery. On this point see our discussion in §§3.3.2–3.3.4.1⁷ (40)

a.

[Frame Scene-Setter [Force [Topic Topic [Focus Focus [Fin [Fin VFin ] [TP …]]]]]] b. [Frame [Force [Topic [Focus Focus [Fin [Fin VFin ][TP …]]]]]]

(41)

a. [Frame Scene-Setter [Force XPV2 [Force VFin ] …]] b. [Frame [Force XPV2 [Force VFin ] …]]

Verb-initial orders, however, do constitute a genuinely distinct class which— at least in some cases—plausibly derives from a difference in the height of verb movement. Starting our discussion with matrix declaratives, recall that V1 orders are only a fully productive word-order pattern in Early Old French (Skårup 1975: 291; Marchello-Nizia 1980: 331; Vance 1997: 32; Rouveret 2004: 193–195; Labelle and Hirschbu¨hler 2005: 66; Simonenko and Hirschbu¨hler 2012). In keeping with a well-established analysis of V1 orders in Germanic and Romance V2 systems we can propose that alongside verb movement, the 1⁷ I exemplify with a focus-initial V2 clause in (38b) but this is not intended to indicate that all V2 clauses feature initial foci (pace Benincà 2004; Donaldson 2012; Poletto 2014: Chapter 1). As a reviewer notes, it is also the case that scholars advocating a focus-V2 analysis frequently use focus as a label for a syntactic projection but not an information-structural value.

4.3 verb movement in old french

119

other ‘half ’ of the V2 constraint is satisfied by a null phrasal constituent in such cases. Wolfe (2019) proposes a tripartite typology of V1 orders in V2 systems, under which rhematic clauses (42a) and cases of V1 under topic-continuity (42b) both feature a null topic which is a variant of pro (Huang 1984: 546; Santorini 1989: 55; Cardinaletti 1990; Sigurdsson 1993: 251; Benincà 2004: 290; Salvi 2012: 106–107; Poletto 2014: 21–23). In contrast, V1 with verba dicendi (42c)—following work by Reis (1995; 2000a; 2000b)—is interpreted as encoding a distinct kind of sentential force which merely recounts a proposition rather than asserting its truth, and thus features a null operator in Spec-ForceP. (42)

a. Aparceurent sei que l’arche fud venue appear.3pl.pst as-if that the-arch be.3sg.pst come.ptcp en l’ost on the-host ‘It appeared as if the arch was above the host’ (OFr. QLR 1, 14, 12th century) b. Vint en Bethléém come.3sg.pst in Bethlehem ‘He came to Bethlehem’ (OFr. QLR 1, 58, 12th century) c. Respundi Samuel: ‘cument? … respond.3sg.pst Samuel: ‘how?’ ‘Samuel responded: “how?”’ (OFr. QLR 1, 56, 12th century)

As such, the Early Old French system shows V1 orders in matrix declaratives which are akin to those found elsewhere in V2 systems as pragmatically marked word-order alternations. If we assume that the finite verb raises into a local configuration with the null element in such cases, they differ from unmarked V-to-Fin movement in that the verb targets the head position of either Top(ic) or Force. The reanalysis to consistent V-to-Force movement that we witness in Later Old French (§4.3.2) leads to the loss of the V-to-Top option, accounting for the notable decline of V1 orders from the late 12th century onwards, noted by Simonenko and Hirschbu¨hler (2012) amongst others. Two syntactically marked verb-placement alternatives deserve special mention. Recall from our discussion in §4.2.2 that both polar questions (43) and imperatives (44) could trigger V-to-Force movement in Latin with corresponding initial placement of the finite verb. This is also the case for both Early and Later Old French (Skårup 1975: 155; Jensen 1990: 340; Roberts 1993: 89–91; Elsig 2009: 256–257; Combettes, Marchello-Nizia, and Prévost 2020: 1221–1222). Enclisis obtains consistently in polar questions and imperatives

120

verb placement and verb movement

until the early 13th century (45, 46) (Foulet 1919: 109–110; Skårup 1975: 358–359; Hulk 1996: 102–103; Labelle and Hirschbu¨hler 2005: 66–68): (43)

est vostre sire ancor levez? be.3sg your lord yet up ‘Is your lord up yet?’ (OFr. Tristan 1, 8021, 12th century, Roberts 1993: 91)

(44)

aiez de moi merci have.2pl.sbjv of me mercy ‘Have mercy on me!’ (OFr. La Queste 218, 31, 13th century)

(45)

Conois la tu? know.2sg =her.cl you ‘Do you know her?’

(46)

(OFr. La Queste 112. 17, 13th century)

faites le vos de gred? do.2pl =it.cl you of will ‘Do you do it willingly?’

(OFr. Roland 2000, 12th century)

To summarize our discussion in this section, we have differentiated between differences in surface verb-placement and genuine differences in the target of finite-verb movement. V3* orders belong to the former class and result not in a difference in the unmarked target of verb movement but in additional merger of XPs in the left periphery. By contrast, a range of pragmatic or syntactic features can lead to the finite verb targeting Top or Force, the former option becoming unavailable in Later Old French.

4.4 Verb movement in Middle French 4.4.1 V2 as V-to-Fin movement in Middle French Whilst the verb-movement properties of Old French have been the centre of extensive attention, this is not the case for Middle French, which is often given a more marginal role in word-order studies. Nevertheless, formal and descriptive studies alike have suggested that Middle French is still a form of V2 system (Adams 1988; Roberts 1993; Vance 1995; Muller 2005). Given that we noted a number of points of discontinuity between Old and Middle French regarding the structure of the left periphery in §§3.4.1–3.4.3, this section will explore the extent to which finite-verb movement is a point of continuity with the Later Old French system. The proposal developed will be that Middle French is still a

4.4 verb movement in middle french

121

form of V2 system, albeit one with a lower locus of unmarked verb movement than what we observed in Later Old French in §4.3.2. From the outset, we can note that the central indicators of a form of V2 grammar with V-to-C movement still stand in the Middle French period generally, and more specifically in the three texts I consider here, Chronique (1441–1444), Roman de Jean de Paris (1494), and Mémoires (1490–1505). First, note that second position is still the preferred one for the finite verb (46.1– 60.2%), as Table 4.5 shows. This is therefore a clear point of continuity in the syntax not just with Later Old French but with all texts from the early medieval period: Table 4.5 Matrix verb placement in Middle French Et-V1

V1

V2

V3

V4

V5

V6*

Total

Chronique 193 22.0% 17 1.9% 404 46.1% 214 24.4% 38 4.3% 8 0.9% 3 0.3% 877 Roman 173 19.8% 0 0.0% 526 60.2% 164 18.8% 10 1.1% 1 0.1% 0 0.0% 874 de Jean de Paris Mémoires 248 29.2% 3 0.4% 468 55.1% 118 13.9% 9 1.1% 3 0.4% 0 0.0% 849

We also note that V3* orders constitute a large proportion of the data in all three texts (cf. also §3.4.2),1⁸ and will consider the qualitative characteristics of these orders below in terms of how they might refine our hypotheses around verb movement: (47)

Toutesfois, je croy qu’ … however I believe.1sg that ‘However, I believe that ...’ (MidFr. Mémoires 4, 15th–16th centuries)

An additional piece of evidence for the V2 status of the language comes from the non-specialized nature of the prefield, which strongly indicates that finiteverb movement does not target the subject field within the T-layer. Detailed data on this point were presented in Chapter 3, Table 3.4, where we observed that XPNon-Subject -V-(S) orders constitute between 46.2% and 73.0% of V2 clauses in the sample, with non-clitic-resumed object topicalization still found in all the texts: 1⁸ Note that an increase in V3* orders has been discussed by various scholars with reference to Middle French texts (Marchello-Nizia 1979; Roberts 1993: 192–204; Ayres-Bennett 1996: 92; Vance 1997: 264–279; Muller 2009: 244).

122

verb placement and verb movement

(48)

ce dit le roy d’Arragon that say.3sg the king of-Aragon ‘The King of Aragon says that’ (MidFr. Roman de Jean de Paris 46, 15th century)

(49)

pour ce vous prie que for that you.cl ask.1sg that ‘For that I ask you that’ (MidFr. Roman de Jean de Paris 22, 15th century)

The licensing of Germanic inversion across all three texts is also suggestive of V-to-C movement of some kind.1⁹ Consider the data in (50) in this regard, which show clear evidence that the finite verb does not remain within the T-layer (cf. for similar findings Roberts 1993: 190): (50)

a.

Encores n’avez vous riens ouy still neg-have.2pl you nothing hear.ptcp ‘You’ve still heard nothing’ (MidFr. Roman de Jean de Paris 47, 15th century) b. Item, pour ce que lesdiz thus for it that the-said Seigneurs se doibvent prouchainement lords refl.cl must.3pl soon assambler à Nevers, ont lesdiz assemble.inf at Nevers have.3pl the-said ambassadeurs requis au Roy que … ambassadors ask.ptcp to-the king that ‘Thus, given that the said lords must soon assemble at Nevers, the said ambassadors have asked the King that … ’ (MidFr. Chronique 17, 15th century)

A final indicator comes from matrix/embedded asymmetries. As discussed in §3.4.6, Middle French complement clauses overwhelmingly show an SVO order (82.9–93.6%) (51), with embedded V2 effects, the licensing of si, or inversion of any kind restricted and found under predicates known to license matrix phenomena in other V2 languages (Vikner 1995; Holmberg and Platzack 1995; Holmberg 2015).

1⁹ As noted in §4.1.2, Germanic inversion is linked to V2 in much work on Old French (Adams 1987c; Vance 1997; de Bakker 1997), and Poletto (2014: 62–63) uses the presence of G-inversion as a key reflex of V-to-C movement in Renaissance Italian.

4.4 verb movement in middle french

(51)

123

Encores disoit ledict Morvillier qu’il ne again say.3sg.pst the-said Morvillier that-he neg povoit penser can.3sg.pst think.inf ‘The aforementioned Morvillier said again that he could not think …’ (MidFr. Mémoires 4, 15th–16th centuries)

Under a V2 account with V-to-C movement these asymmetries are accounted for, which would not be the case if V-to-T movement were licensed symmetrically, as part of either an SVO or T-V2 system. As such, we can conclude that the overwhelming preference for embedded preverbal subjects coupled with the second-position verb placement shown in Table 4.6 are indicative of distinctions in the target of verb movement in the matrix and embedded domains. Table 4.6 Embedded verb placement in Middle French V1 Chronique 0 Roman 8 de Jean de Paris Mémoires 4

V2

V3

V4

V5

V6*

Total

0.0% 110 6.3% 111

89.4% 13 88.1% 5

10.6% 0 4.0% 2

0.0% 0 1.6% 0

0.0% 0 0.0% 0

0.0% 123 0.0% 126

2.6% 134

88.7% 13

8.6% 0

0.0% 0

0.0% 0

0.0% 151

With these indicators suggesting the presence of V-to-C movement, can we further refine this hypothesis and specify the target of verb movement within the C-layer? Several pieces of evidence suggest that Middle French—unlike Later Old French—is a Fin-V2 system. First, we see in Table 4.5 that all three texts feature V4*, which is straightforwardly ruled out under the Force-V2 analysis and very rarely found in Later Old French texts (Wolfe 2016a; Wolfe 2018a). Thus, a clause such as the one in (52) cannot be accommodated within the structure proposed in (53): (52)

Le roy d’Angleterre pour ces parolles ce print the king of-England by these words it take.3sg.pst moult fort a rire very strong to laugh.inf ‘The King of England started laughing loudly in response to these words’ (MidFr. Roman de Jean de Paris 39, 15th century)

(53)

[Frame Scene-Setter [Force XPV2 [Force VFin ] …]]

124

verb placement and verb movement

Looking at the qualitative patterns found in V3 orders, which account for a substantial portion of the data, it is also not clear that they can be accommodated under the Force-V2 model (see for more examples §3.4.2). Recall that in Later Old French—similarly to other Force-V2 systems—V3 triggers are a small class of scene-setting clauses and adverbials in the frame field. When we consider cases such as (54), however, it is far from obvious that this restricted definition applies: an indirect object selected by a finite verb is not analysable as a scene-setter: (54)

A quoy le roy fist responce to which the King make.3sg.pst response ‘The King responded to this’ (MidFr. Chronique 29, 15th century)

A final argument that we are not dealing with a Force-V2 system comes from the distribution of the particle si. In §3.4.3 we saw that Middle French si frequently occurs with fronted subjects (55), in contrast to Later Old French but in a similar fashion to Early Old French (Wolfe 2018b: 345): (55)

L’aultre point si est que the-other point si be.3sg that ‘The other point is that …’ (MidFr. Roman de Jean de Paris 24, 15th century)

Once again, a subject DP with no clitic resumption is not a straightforward candidate for a constituent in the frame field. We can therefore conclude that si is base-generated in a position lower than the topic-focus layer, which hosts fronted arguments, more specifically Spec-FinP. These points together yield an important observation: whilst we see the stepwise upwards reanalysis of the locus of verb movement from archaic Latin to Later Old French (V-v-Fin-Force), V-to-Force movement is not immediately reanalysed as V-to-T movement in the passage to Modern French. Rather, we see the return of a ‘low’ V2 system where finite-verb movement targets Fin. Although Middle French has not been analysed before in the light of the Fin/Force dichotomy, earlier approaches did implement the intuition that V2 and verb movement target a lower position in Middle French than in Later Old French, with both Vance (1993; 1995; 1997) and Roberts (1993) suggesting this was reanalysis of V-to-C movement as V-to-Agr movement.

4.4 verb movement in middle french

125

4.4.2 Marked verb placement in Middle French In line with others working on Middle French word order, we saw in Table 4.5 that between 13.9% and 24.4% of matrix clauses show a V3 word order, with some V4* also attested. As we suggested for Later Old French in §4.3.3, there is no reason to believe that these cases involve a different locus of verb movement. Rather, once V-to-Fin movement has taken place, frame-setters, topics, and foci can co-occur in the left periphery (56). There is reason to believe that such orders are, however, more qualitatively restricted than was the case in the V-to-Fin grammar of Early Old French. This is, once again, due not to a difference in verb movement but rather to the fact that the focus field is rarely activated in the language of this period (cf. §3.4.2), such that V3* orders are typically licensed when framesetters or multiple topics in the recursive topic field co-occur (57), in this case seemingly a left-dislocated constituent co-occurring with a moved topical subject DP: (56)

[Frame (Frame-Setter) [Force [Topic (Topic) [Focus (QP) [Fin [Fin VFin ] [TP …]]]]]]

(57)

De prisonniers bons, les gens du roy of prisoners good the people of-the king en eurent des meilleurs de ceulx part.cl have.3pl.pst of-the best of those qui fuyoient who flee.3pl.pst ‘Of the good prisoners, the King’s people had the best of those who fled’ (MidFr. Mémoires 14, 15th–16th centuries)

Verb-first orders, as Table 4.6 shows, are a much more marginal phenomenon. In fact, in the only text where they appear with any consistency— Chronique—they occur in a specific context where the author appears to be interpreting a preceding adverbial as part of the following clause (58). As such, the situation is very similar to that reported for Later Old French (Adams 1987c; Simonenko and Hirschbu¨hler 2012, and references above in §4.3.3) where matrix declarative verb-initial clauses are exceptionally rare.2⁰ 2⁰ But see §5.4.3 for a discussion of an exception to this generalization for Middle French, in the presence of either et ‘and’ or ne ‘not’ in initial position of a matrix clause.

126

verb placement and verb movement

(58)

a.

Item. Ont remoustré au Roy Thus have.3pl state.ptcp to-the King ‘They have thus stated to the King …’ (MidFr. Chronique 19, 15th century) b. Item. Ont dict lesdiz ambassadeurs Thus have say.ptcp the-said ambassadors ‘The aforementioned ambassadors thus said …’ (MidFr. Chronique 23, 15th century)

As in Later Old French and Modern French, imperatives still trigger V1 (59), however, which we interpreted above as V-to-Force movement (Rizzi 1997). (59)

Véez là le roy et tout ce peuple sailly see.2pl there the king and all those people rush.ptcp de la ville from the town ‘See there the King and all those people rushing [out] from the town’ (MidFr. Mémoires 31, 15th–16th century)

Turning to interrogatives, several detailed treatments have concluded that early indications of change in this part of the grammar are observable in the Middle French period (Roberts 1993; Elsig 2009; Combettes, Marchello-Nizia, and Prévost 2020). Before considering the development of these changes from the Renaissance onwards, note for now that in the mid-15th century instances of complex inversion (cf. §4.5.2) are found (Roberts 1993: 172), alongside a simple VS order in interrogatives (60), which is still the dominant pattern across the three texts considered in this section.21 (60)

viendra encores Jehan de Paris ? come.3sg.fut again Jehan de Paris? ‘… will Jehan de Paris come again?’ (MidFr. Roman de Jean de Paris 61, 15th century)

Several scholars have noted that differentiating genuine cases of complex inversion from simple inversion with a left-dislocated subject is extremely challenging,22 but if Roberts (1993) is correct, then V-to-Force movement was already being lost in a subset of interrogatives by the end of the 15th century.

21 See Elsig (2009: 133) for the proposal that pronominal inversion in interrogatives was already becoming lexically restricted by the 15th century, although such restrictions become far more marked later on. 22 For recent discussion, see Combettes, Marchello-Nizia, and Prévost (2020: 1224).

4.5 verb movement from renaissance to modern french

127

4.5 Verb movement from Renaissance to Modern French 4.5.1 The emergence of V-to-T movement Medieval French—from the earliest textual attestations through to the end of the 15th century—is characterized by a diachronically variable V2 syntax with V-to-C movement, as we have seen in §4.4. Whilst precise dating varies, there is general agreement that this is not the correct characterization of Renaissance, Classical, or Modern French, which show increasing evidence of an SVO syntax with verb movement no higher than the T-layer. What is the evidence for an SVO syntax? The linear position of the finite verb alone is insufficient to diagnose this, as second position is unmarked in both a V2 and an SVO grammar. However, the structure of the prefield is markedly distinct if we compare V2 Old French with SVO Modern French. The crucial observation for our purposes is the quite dramatic distinction between our Middle French texts, on the one hand, and our Renaissance French texts, on the other: whilst in Chronique (1441–1444), Roman de Jean de Paris (1494), and Mémoires (1490–1505) subjects are the preverbal constituents in linear Verb Second clauses in 27.0%, 47.1%, and 53.5% of cases respectively, this rises to 89.8% in Rabelais (1532–1534), 91.4% in Montaigne (1580), 99.6% in Phèdre (1677), and 99.1% in Candide (1759). Whilst attempts such as Lightfoot’s (1995) to quantify precisely what proportion of clause-initial constituents must be non-subjects to ‘trigger’ V2 should be taken with some caution,23 no account based on either acquisitional or historical data would maintain that a language with approximately 90% or above preverbal subjects in linear verbsecond clauses features a V2 grammar with V-to-C movement. Alongside the marked decline of verb–subject inversion—and in particular Germanic inversion, which we will consider in more detail in §5.3—the embedded domain provides additional evidence for a symmetrical V-to-T SVO analysis. In §3.4.6 it was shown that no significant clause-type asymmetries exist from the Renaissance onwards. Rather, embedded clauses are overwhelmingly SVO, specifically 88.6% to 100.0% of the linear verb second complement-clause sample. The evidence is quite unambiguous, therefore, that in the grammar of certain speakers, unmarked matrix V-to-Fin movement was lost at the turn of the 16th century. Note that, although the precise decade identified is subject to variation based on the textual corpus used, dating the loss of V2 as occurring 23 For recent discussion of the V2 ‘cut-off ’ and the extent to which contemporary acquisitional data can be compared with historical corpus data, see discussions in Wolfe (2015c: 135–139; 2018a: 23–27).

128

verb placement and verb movement

at the end of the 15th or beginning of the 16th century is broadly in keeping with the analysis of Roberts (1993: 166–204) and Vance (1995), which are still the most detailed theoretical studies on V2-loss to have been carried out to date. As such, we are faced with a strikingly homogeneous picture for a period of five centuries from approximately 1500 onwards; whilst the locus of unmarked verb movement went through at least three significant reanalyses in the first five centuries—Fin to Force at the beginning of the 13th century, Force to Fin at the beginning of the 15th century, and Fin to T at the beginning of the 16th century—V-to-T movement has apparently characterized French for the last half-millennium. This claim is closely in keeping with the majority of descriptive works on Renaissance (Brunot 1927: 479; Gougenheim 1973: 253–255), Classical (Spillebout 1985; Fournier 2001: 89; Fournier 2007b: 83), and Modern French syntax (Blinkenberg 1928; Désirat and Hordé 1976; Harris 1978: 118; Ashby 1988: 227; Rowlett 2007: 7–8; Smith 2016: 311–312; Salvi 2016: 1000), which view the unmarked order of constituents as SVO. The V-to-T hypothesis clearly holds true of Modern French in contrast to a lower-verb-movement language like English (e.g. Roberts 2019: 359–362). Pollock (1989: 367) shows unambiguously that the finite verb and auxiliary consistently precede a range of adverbials, negation, and quantifiers, which indicate this (61–63), and Schifano (2018: 63) suggests that there is no evidence of diatopic variation across speakers whose judgements she elicited from France, Belgium, or Quebec, regarding the placement of the finite verb relative to adverbials. (61)

(62)

(63)

a. *John likes not b. Jean n’aime pas Jean neg=like.3sg neg ‘John doesn’t like Mary’

Mary Marie Mary (ModFr.)

a. *John kisses often Mary b. Jean embrasse souvent Marie Jean kiss.3sg often Mary ‘John often kisses Mary’

(ModFr.)

a. *My friends love all Mary b. Mes amis aiment tous Marie my friends love.3pl all Mary ‘My friends all love Mary’

(ModFr.)

4.5 verb movement from renaissance to modern french

129

However, we noted in §4.1.1—following Schifano’s (2015a; 2015b; 2018) work—that V-to-T movement may be an illusory cover term for a richer typology of verb-movement phenomena. In addition, we have seen that adopting the Fin vs. Force typology, as opposed to a supposedly homogeneous phenomenon of ‘V-to-C’ movement, can yield important insights into the earlier history of the French language. We can therefore ask the following: if—as Schifano suggests—Modern French is consistently a ‘high’ V-to-T language (cf. Table 4.1), does the same hold of all stages from the Renaissance onwards? Schifano (2018: Chapter 5) makes use of very fine-grained speaker judgements for a wide range of verb classes to form her typology of verb movement. However, the central observation she makes about Modern French is that—alongside Romanian—it is the only Modern Romance variety where finite verbs and auxiliaries consistently precede adverbials lexicalizing Cinque’s (1999) Higher Adverb Space.2⁴ Using the ARTFL-FRANTEXT database,2⁵ we are thus able to ascertain—albeit to a limited extent given the limitations of the data—whether this is also true of other post-medieval stages of French. First, note that the French verb consistently raises above adverbials in the Lower Adverb Space, here exemplified by bien ‘well’ (64), toujours ‘always, still’ (65), and déjà ‘already’ (66): (64)

a.

Et sçavent bien que des païs estranges and know.3pl well that of-the countries foreign ‘And they know well that foreign countries …’ (RenFr. Marot. Des inédits, Appendice 1, 16th century) b. Je recognoy bien qu’ agiter ceste question … I recognize.1sg well that debate.inf this question ‘I know well that to debate this question …’ (CFr. Dupleix, Métaphysique 2, 247, 17th century) c. Je compris bien qu’ elle ne voudrait … I understand.1sg well that she neg want.3sg.cond ‘I understood well that she didn’t want …’ (CFr. Saint-Simon, Mémoires 9, 165, 18th century)

2⁴ For a comparative Romance analysis of the different adverb classes, see also Ledgeway (2012a: 142–143). 2⁵ https://artfl-project.uchicago.edu/content/artfl-frantext.

130

verb placement and verb movement

(65)

a.

(66)

a.

elle luy respondoit toujours qu’ il she him.cl respond.3sg.pst always that he viendroit … come.3sg.pst ‘she always replied that he would come …’ (RenFr. Marguerite de Navarre, Heptaméron 1, 741, 16th century) b. J’obérai toujours à son commandement I-obey.1sg.fut always to his commandment ‘I will always obey his commandment’ (CFr. Corneille, La Galerie du palais 9, 88, 17th century) c. Aussi la reine aima toujours also the Queen love.3sg.pst still tendrement le duc de parme deeply the Duke of Parma ‘The Queen also still loved the Duke of Parma deeply’ (CFr. Saint-Simon, Mémoires 19, 22, 18th century)

Je sens déjà un piteux souvenir I feel.1sg already a sorry memory ‘I am already aware of a sorry memory’ (RenFr. Labé, Oeuvres poétiques, 1, 97, 16th century) b. Je vois déjà Mélite I see.1sg already Mélite ‘I already see Mélite’ (CFr. Corneille, Mélite 2, 232, 17th century) c. Ils côtoyoient déjà le rivage occidental they go-coast-to-coast.3pl.pst already the shore western ‘They already went coast to coast along the western shore’ (CFr. Terrason, Histoire, 6, 11, 18th century)

However, a corpus search also reveals that from the 17th century onwards the French finite verb raises higher than Cinque’s (1999) Higher Adverb Space, illustrated here with the adverbials probablement ‘probably’, apparemment ‘apparently’, and peut-être ‘perhaps’. Owing to their comparative rarity or nonattestation in the 16th-century corpus, it is not possible to test the distribution of these adverbials relative to lexical verbs in Renaissance French. However, a 16th-century example of a finite lexical verb preceding the high adverb generalement ‘generally’ is given in (70), suggesting that Renaissance French also featured high verb movement within the T-layer:

4.5 verb movement from renaissance to modern french

131

(67)

a. et de là vient probablement la conformité and from there come.3sg probably the conformity ‘Conformity probably comes from there …’ (CFr. Bonours, Ariste et Eugène 71, 17th century) b. Philippe II parloit probablement espagnol Philip II speak.3sg.pst probably Spanish ‘Philip II probably spoke Spanish’ (CFr. Schwab, Dissertation, 206, 18th century)

(68)

a. Cela vient apparemment de … this come.3sg.pst apparently from ‘This apparently comes from …’ (CFr. Bonhours, Ariste et Eugène 75, 17th century) b. Dieu veut apparemment que je meure … God want.3sg apparently that I die.1sg.sbjv ‘God apparently wants me to die’ (CFr. Marivaux, Spectateur 237, 18th century)

(69)

a. et je ne la verray peutestre and I neg her.cl see.1sg.fut perhaps jamais never ‘And I might not ever see her’ (CFr. Scudéry, Artamène 5971, 17th century) b. Cette difference vient peutêtre des … this difference come.3sg perhaps from-the ‘This difference perhaps comes from the …’ (CFr. Rollin, Histoire 495, 18th century)

(70)

Nous les appellons generalement Scythes we them.cl call.1pl generally Scythes ‘We generally call them Scythians’ (RenFr. Roy Loys, Univers 11, 16th century)

Overall, we can conclude that from the 16th century onwards, French has featured a grammar with unmarked V-to-T movement. This is evidenced by clear syntactic indicators that the prefield is a subject-related position in the T-layer and the fact that V2 correlates attested in texts until the 15th century are no longer present to a significant degree. Adopting a nuanced typology of V-to-T movement following work by Schifano (2015a; 2018), we saw evidence from the position of the verb relative to adverbials that unmarked verb movement

132

verb placement and verb movement

to a position at the upper reaches of the Higher Adverb Space may have been a consistent feature of French since the 16th century.

4.5.2 V-to-C relics from Renaissance French onwards In parallel to phrasal movement to the left periphery (cf. §3.4.4), the period from the 16th century onwards is characterized by the progressive loss of Vto-C movement in contexts where it previously obtained in Latin, Old French, and Middle French. As we will see, despite V-to-C movement being a hallmark of Old and Middle French grammar, its loss in recent times has become so advanced that many colloquial varieties of French lack V-to-C movement in almost all contexts where it was previously found. 4.5.2.1 V-to-C in Renaissance and Classical French As we saw in §4.4.2, interrogatives in Old and Middle French feature V-to-C movement, but there are some early indications from complex inversion structures that interrogatives may have undergone reanalysis by the late Middle French period, such that they no longer feature the same locus of verb movement. Although precise dating varies, there is near-universal agreement that by the 16th century there is a marked decline in both nominal and pronominal inversion in interrogatives, with the decline in nominal inversion—sporadically attested in the 16th century—particularly prominent (Foulet 1919; Gougenheim 1973: 236; Harris 1978: 31–34; Roberts 1993: 166–177; Elsig 2009: 222; Combettes, Marchello-Nizia, and Prévost 2020: 1223–1224). Indeed, Fournier (2007a: 120) reports that cases of nominal inversion such as (71) in interrogatives are almost entirely absent from 1600 onwards. (71)

Mais à qui donnera Saffredent sa voix? but to whom give.3sg.fut Saffredent her voice ‘But who will Saffredent give her voice to?’ (RenFr. Marguerite de Navarre, Heptaméron 41, 22, 250, 16th century, Gougenheim 1973: 212)

Elsig (2009: 133, 212–213) also shows using quantitative data that pronominal inversion becomes increasingly lexically restricted from the 16th century onwards. In particular, he finds that high-frequency verbs favour inversion, in particular être ‘be’, vouloir ‘want’, savoir ‘know’, avoir ‘have’, and voir ‘see’.2⁶ 2⁶ Elsig (2009: 137) shows that person plays no significant role in favouring pronominal inversion in the 15th–17th centuries, in contrast to Modern French (§4.5.2.2).

4.5 verb movement from renaissance to modern french

133

These lexical restrictions go hand in hand with a general decline of pronominal inversion in interrogatives, which is shown to be particularly sharp between the 16th and 17th centuries (Elsig 2009: 212). (72)

a. Voulez-vous mon cousteau? want.2pl-you my knife ‘Do you want my knife?’ (RenFr. Rabelais 21, 235, 16th century) b. Veux-je la souffrir? Veux-je y répondre? want-1sg her.cl suffer.inf want-I loc.cl respond.inf ‘Do I want to suffer her? Do I want to respond?’ (CFr. Lafayette, Clèves 347, 17th century, Fournier 2007a: 120)

It is important to note that speakers at this time would have had access to three alternative structures to encode interrogatives: polar interrogatives are likely to have been licensed via intonation alone since the very earliest French texts, with Combettes, Marchello-Nizia, and Prévost (2020: 1220) suggesting that this is the case even in Early Old French. In contrast to Old and Middle French, however, the grammar of Renaissance and Classical French also features est-ce que (lit. ‘is it that’) as an interrogative marker (Foulet 1919: 172–173; Spillebout 1985: 381; Fournier 2007a: 121; Elsig 2009: 223–225). (73)

Pour-quoy est-ce que vous me faicetes une si why be.3sg-it that you me.cl do.2pl a so grande et longue harangue? great and long lecture ‘Why are you treating me to such a great long rant?’ (RenFr. Marguerite de Navarre, Heptaméron 10, 1, 113, 16th century, Gougenheim 1973: 213)

This construction—prior to reanalysis itself an inverted verb and subject, followed by a complementizer—grows in prevalence from the 16th and 17th centuries onwards. Clear evidence that it is grammaticalized by the 17th century and not an instance of inversion itself comes from the tense and number of est, which cannot be modified from the 17th century onwards, in contrast to Middle French, and Renaissance French, where both could be (74): (74)

Jusques à quand sera-ce que tu abuseras until to when be.3sg.fut-it that you abuse.2sg.fut de la jeunesse de nostre Roy? of the youth of our King? ‘Until when will you take advantage of the youth of our King?’ (RenFr. Hotman, Épitre 1, 16th century, Gougenheim 1973: 212)

134

verb placement and verb movement

Furthermore, instances of complex inversion—sporadically attested in Middle French—are increasingly common in Renaissance and Classical French (Foulet 1919: 249; Combettes, Marchello-Nizia, and Prévost 2020: 1224; Elsig 2009: §5.2.7; Fournier 2007a: 121). The important point to note is that the ability of the nominal subject to follow wh-words—typically exponents of the focus field—is strongly suggestive that these subjects are not left-dislocated, and as such we are dealing with a construction with lower verb movement than in previous stages of the grammar: (75)

a.

Mais, marquis, par quelle raison, de what reason of but marquis for grâce, cette comédie est-elle ce que grace this comedy be.3sg-it that what tu dis? you say.2sg ‘But, marquis, for what reason, please, is this comedy what you say?’ (CFr. Molière 204, 17th century, Elsig 2009: 222) b. ces deux fillettes sont elles pucelles? these two little-girls be.3pl they virgins ‘Are these two little girls virgins?’ (RenFr. Rabelais 15, 183, 16th century)

We can summarize, therefore, that by the end of the classical period in the 18th century, the foundations of the interrogative system found in the modern language have already been laid. Nominal inversion is entirely absent, pronominal inversion is increasingly restricted to a particular class of verbs, and complex inversion, intonation, and est-ce que are the dominant interrogative-marking strategies. Indeed, the only clear instance of V-to-C movement remaining in the grammar as a point of continuity with Old and Middle French is the imperative, which still entails V-to-C movement in Renaissance and Classical French: (76)

écoutez-moi! listen.2pl-me ‘Listen to me!’

(CFr. Lafayette, Clèves 396, 17th century)

4.5.2.2 V-to-C in Modern French As in Classical French, Modern French lacks so-called ‘simple’ inversion (Roberts 1993), where a nominal subject can appear postverbally after an auxiliary or finite verb in C (77).

4.5 verb movement from renaissance to modern french

(77)

a.

*Partira son leave.3sg.fut his ‘Will his son leave? b. *A Jean have.3sg Jean ‘Has John spoken?’

135

fils? son (ModFr.) parlé? speak.ptcp (ModFr.)

As such, we can view the licensing of V-to-C movement with postverbal nominal subjects as a characteristic of French from the earliest attestations through to the 16th century, and the illicit nature of this construction as characteristic of the last five centuries leading up to the present day. However, pronominal inversion is still grammatical in Modern French, albeit in a restricted number of contexts and subject to certain sociolinguistic and diatopic variation (Terry 1970; Désirat and Hordé 1976: 152–153; Ashby 1977b; Harris 1978: 31; Blanche-Benveniste 1990: 39; Zribi-Hertz 1994; Coveney 1996: 93–110; Rowlett 2007: 199–207; De Cat 2009: 113–114; Elsig 2009: 18; Tailleur 2013). Following the analysis in Kayne (1975; 1983a), I take both pronominal inversion (78) and complex inversion (79) to be underlyingly similar phenomena here, with the same target of verb movement.2⁷ (78)

(79)

Quand est-il parti? when be.3sg-he leave.ptcp ‘When did he leave?’

(ModFr.)

Quand Jean est-il parti? when Jean be.3sg-he leave.ptcp ‘When did John leave?’

(ModFr.)

It is tempting—through a parallel with similar English inversion structures (Rizzi 1986a; Rizzi 1996; Biberauer and Roberts 2012) and the V2 syntax of Old and Middle French—to treat such cases as instances of V-to-C movement and thus ‘relic’ V2. Analyses along these lines have been put forward most prominently by Kayne (1983a), Rizzi and Roberts (1989), Cardinaletti and Roberts (2002), and Roberts (2010c: 311–319). However, the V-to-C account has been challenged by Sportiche (1998: 325–337), Poletto and Pollock (2004; 2017), Rowlett (2007: 202–217), and Cardinaletti (2021) amongst others. Following Cardinaletti’s (2021) analysis in particular, I suggest that the 2⁷ Some varieties of French also feature an affixal interrogative agreement marker /-ti/ or /-tu/, which probably emerged in the 16th century (Elsig 2009: 31): ta mère est-tu là? ‘is your mother there?’. Outside Quebec—where it is still widely used (Elsig and Poplack 2006)—there is general consensus that its use is either obsolete or in rapid decline (Désirat and Hordé 1976: 153; Coveney 1996; Vinet 2004: 237–238; Rowlett 2007: 207).

136

verb placement and verb movement

V-to-C movement is not the correct analysis of either pronominal inversion or complex inversion. Instead, I suggest in line with the analysis of the subject layer outlined in Chapter 5, that Modern French lexical subjects occupy SpecSubjP (Cardinaletti 2004) whereas pronouns occupy a position lower within the TP—most likely Spec-TP in Modern French. If we assume that Modern Standard French interrogatives feature V-to-Subj movement (80), we therefore straightforwardly capture the fact that inversion structures are licit with pronominal subjects but not lexical subjects (cf. 81) and why a non-dislocated preverbal lexical subject can precede the finite verb in complex inversion (79):2⁸ (80)

… [Focus Wh [FinP [SubjP (SubjectDP ) [Subj VFin ] [TP SubjectPronominal [vP [v VFin ] … ]]]]]

(81)

a.

Quand est-il when be.3sg-he ‘When did he leave?’ b. *Quand est when be.3sg

parti? leave.ptcp (ModFr.) Jean parti? Jean leave.ptcp

Most recent studies of interrogation in French have reported intonation alone or the use of est-ce que as the dominant ways to form interrogatives in corpora (Coveney 1996; Elsig 2009; Auger and Villeneuve 2019: 214). However, even in the relatively formal styles in which it is licensed, V-to-Subj movement in interrogatives is quite restricted. Grammatical person of the subject has an effect—for example—with any form of pronominal inversion extremely marginal with first-person-singular subjects (Coveney 1996: 106– 110; Rowlett 2007: 205; Roberts 2010a: 314), and auxiliary verbs are more acceptable in inversion structures than lexical verbs.2⁹ (82)

a.

*Chante-je? sing-I ‘Do I sing?’ b. Chantez-vous? sing.2pl-you ‘Do you sing?’

(ModFr.)

(ModFr.)

2⁸ For the sake of simplicity, I suggest that subject pronominal forms are phrasal in postverbal position and thus occupy Spec-TP. In §5.5 we will revisit this intuition and suggest that they are, instead, cliticized to the finite verb when ‘inverted’. 2⁹ Note in reference to (81a) that although this example would be unacceptable in spoken language, forms such as chanté-je do form part of the formal literary standard (see for discussion Grevisse and Goosse 2016 : §794b).

4.5 verb movement from renaissance to modern french

137

Furthermore, licensing of inversion shows substantial regional variation with both Canadian and Belgian varieties showing more widespread attestation of pronominal inversion than Metropolitan French (De Cat 2009: 6; Elsig 2009). As such, we can conclude that whilst pronominal inversion is a feature of only relatively formal registers in Standard Metropolitan French (ZribiHertz 1994), this is not the case in certain parts of Belgium and Canada, which are therefore more conservative in featuring verb movement to a higher head in these particular constructions. Is verb movement out of the T-layer entirely ruled out elsewhere in Modern French? Imperatives are the clearest indicator that it is not: these have targeted a high left-peripheral position—Force—from as early as archaic Latin onwards (§4.2.2.2). Ledgeway (2020: 75) also discusses several residual V-to-C contexts licit in formal registers of Standard Modern French where non-declarative force is also at stake: (83)

(84)

(85)

Puisse-t-elle réussir! may.3sg.sbjv-she succeed.inf ‘May she succeed!’

(ModFr. Ledgeway 2020: 75)

Est-elle jolie! be.3sg=she pretty ‘Isn’t she pretty!’

(ModFr. Ledgeway 2020: 75)

M’eût-il encouragé me.cl-have.3sg.ipfv.sbjv-he encourage.ptcp ‘Had he encouraged me …’ (ModFr. Ledgeway 2020: 75)

However, as Ledgeway himself acknowledges, such structures are increasingly restricted to a defined class of auxiliaries even in very conservative registers and, as such, we can observe ‘well-advanced and ongoing morphosyntactic and lexical restrictions on a once fully productive movement operation’ (Ledgeway 2020: 76). Overall, therefore, we conclude that V-to-C movement is an increasingly restricted option across French varieties today, though V-to-Force movement is the most robust ‘survivor’ in this respect. In line with the analysis of Zribi-Hertz (1994), we can thus see inversion as an increasingly sociolinguistically marked option in most diatopic varieties, with evidence suggesting that interrogative inversion may—in fact—feature verb movement to a higher head in the inflectional layer, rather than V-to-C movement.

138

verb placement and verb movement

4.6 Summary and Conclusions 4.6.1 Summary We started this chapter in §4.1 by highlighting three broad types of syntactic system: those where the finite verb stays in situ or remains within the v–VP complex; those where the finite verb raises into the T-layer; and those where the finite verb raises into the C-layer. Note that this tripartite typology is based on the three phasal layers—discourse, inflectional, and thematic—presented in Grohmann (2003). Verb movement to each of these layers thus corresponds to the most significant meso typological stages of French’s history, with the verb staying in situ in the thematic layer during Latin’s SOV stage, moving to the discourse layer in the late Latin, Old French, and Middle French V2 stage, and moving to the inflectional layer in the SVO stage observable from Renaissance French onwards. Whilst our discussion has shown that there is value in the SOV, V2, and SVO labels we first outlined in §2.3, the precise verb-movement properties of the language at different stages have been shown to be far more complex. In Classical Latin we suggested that the finite verb remains in situ, that is to say in V. However, in a parallel fashion to phrasal constituents which could be moved to the left periphery in a range of syntactically or pragmatically marked environments, finite verbs could also target Fin, Foc(us), Top(ic), and Force. In terms of unmarked verb movement, we suggested that this may have been reanalysed upwards to v during the classical period (Ledgeway 2012a), whilst the number of marked environments triggering V-to-C movement also increases in Classical Latin, most straightforwardly evidenced by the rise in V1 and V2 surface structures (Bauer 1995; Bauer 2009; Salvi 2004; Wolfe 2016a). Throughout the classical period, embedded V-movement, however, appears to stay in situ, yielding a strict SOV syntax. A reanalysis of previously marked V-to-C movement as unmarked may have already taken place in late Latin, accounting for the substantial decline in SOV orders and the fact that V1 and V2 now constitute the dominant orders for the finite verb in both matrix and embedded clauses (Ledgeway 2017b). It was suggested in §4.2.3 that the finite verb in late Latin systematically undergoes movement to Fin in matrix clauses, and T and Fin in embedded clauses. The evidence from Early Old French is that the locus of verb movement remains constant on Fin, with early texts thus showing the properties of a ‘relaxed’ Fin-V2 system. The finite verb can therefore appear in third, fourth, or greater position when multiple constituents co-occur in the left periphery.

4.6 summary and conclusions

139

We also linked the licensing of V1 structures in Early Old French to V-toTop(ic) movement, and suggested that the potential for the finite verb to raise higher than Fin in certain environments is suggested by the presence of enclisis in certain clause-types (cf. Benincà 2004; Benincà 2006). In embedded clauses, we suggested that V-to-T movement was the norm, with embedded V-to-Fin movement obtaining in certain contexts. However, after approximately 1200, the system changes to a stricter form of V2 grammar, which we saw featured unmarked V-to-Force movement. Crucially, although a restricted class of V3 clauses are licensed, there is no evidence to suggest that Force does not attract the finite verb in all matrix clauses. Embedded clauses, by contrast, are generally SVO and thus feature the embedded V-to-T movement which still persists into Modern French, with only a small subclass showing embedded V2 effects suggestive of V-to-Fin movement. Whilst we might expect Middle French to show a continuation of the Later Old French Force-V2 system, this is not the case. Instead, we saw evidence that the finite verb targets a lower left-peripheral position, namely Fin. Whilst this yields an increase in the qualitative range of V3* orders, non-coordinated V1 orders do not re-emerge as a feature of the grammar. V-to-Force movement— we suggested—may still be retained as a marked word-order option, specifically in imperatives and polar interrogatives. In the embedded domain, the picture is similar to Later Old French, with unmarked V-to-T movement and embedded V-to-Fin movement in a small subset of environments. In Renaissance, Classical, and Modern French we see clear indicators of the loss of any kind of V2 syntax featuring V-to-C movement. Instead, the finite verb consistently targets T in both matrix and embedded clauses, as it does today. The major differences across the last five centuries concern the extent to which residual V-to-C movement is licensed: cases of residual V-to-C movement with postverbal nominal subjects are reported as marked but relatively widely attested in Renaissance French,3⁰ but become extremely rare in Classical French and—as we saw in §4.5.2—ungrammatical in Modern French. Pronominal inversion—which entails V-to-Subj movement—is widely attested in Classical French and is still retained, with categorial restrictions, in formal Metropolitan Modern French and some varieties of Belgian and Canadian French. In more colloquial varieties, however, even V-to-Subj movement is not found, with in situ alternatives employed for interrogatives. V-to-Force

3⁰ I take the clear decline in nominal inversion in both declaratives and interrogatives as the key piece of evidence that V-to-Fin/Foc/Top movement is reanalysed as V-to-Subj movement in the 16th century. Recall from our discussion in §§3.4.4–3.4.5 that the 16th century also shows a substantial decline in argument fronting derived via movement.

140

verb placement and verb movement

movement is still found in imperatives, as it also was throughout French’s history. These findings are summarized in Table 4.7, with bracketing indicating a significant pragmatically or syntactically marked option. Table 4.7 The locus of verb movement in the history of French

Archaic Latin Classical Latin Late Latin Early Old French Later Old French Middle French Renaissance French Classical French Formal Modern French Colloquial Modern French

Matrix

Embedded

V (Fin, Foc, Top, Force) v (Fin, Foc, Top, Force) Fin (Foc, Top, Force) Fin (Foc, Top, Force) Force Fin (Foc, Top, Force) T (Fin, Foc, Top, Force) T (Subj, Force) T (Subj, Force) T (Force)

V v T (Fin) T (Fin) T (Fin) T (Fin) T T T T

4.6.2 The great leap from Latin to Early Old French Considering the traditional characterization of Latin as SOV with Old and Middle French as V2, it is easy to consider the Latin > French transition as an abrupt one between two quite distinct kinds of syntactic system. Indeed, we saw in §4.3.1.1 that certain scholars have invoked this alleged discontinuity to argue against the V2 analysis of Old French. However, the data assembled so far in this chapter show quite clearly that a plausible scenario can be reconstructed as to how a V2 grammar might have been innovated. Limitations of the Latin textual records make precisely dating the genesis of unmarked V-toFin movement challenging, but data from the Peregrinatio suggest that it was not later than the 6th century. If we consider the fact that V1 and V-medial orders are also characteristic of the letters of Terentianus (86) (2nd century ad) (Adams 1977; Bauer 1995: 96; Bauer 2009: 276; Salvi 2004: 102; Clackson and Horrocks 2007: 255; Ledgeway 2012a: 257–258), we can tentatively suggest that in certain varieties of Latin the reanalysis may have been several centuries earlier. (86)

Uidit Germani liberatam see.3sg.pst Germanus.gen freedwoman.acc ‘She saw Germanus’s freedwoman’ (Latin, Terentianus, P. Mich 469)

Regardless of dating, it was established above that even conservative varieties of Latin featured verb fronting to Fin, Foc, Top, and Force, in the former

4.6 summary and conclusions

141

three cases either independently of, or alongside, merger of an XP. In order for V-to-Fin movement to feature in a new grammar, a reanalysis therefore needs to take place where previously marked structures are unmarked. In formal terms, this would be a shift from a grammar with separate verb-movementtriggering features on Fin, Foc, Top, and Force to a single [uV] feature on Fin. Far from being implausible, such a reanalysis is a straightforward case of Feature Economy driving reanalysis (Roberts and Roussou 2002: 301; Van Gelderen 2008: 297; Van Gelderen 2009a: 93; Biberauer and Roberts 2015: 301; Biberauer and Roberts 2017: 145–146) as well as Input Generalization (Roberts 2007a: 275; Biberauer and Roberts 2017: 147), with a previously diverse class of triggers being reanalysed as a single trigger at play in all matrix declaratives. As well as being plausible from a theoretical perspective, the reanalysis of marked left-peripheral verb movement as unmarked is empirically well supported, with similar accounts having been developed for Germanic (Hinterho¨lzl and Petrova 2010; Walkden 2014), Celtic (Meelen 2020), and Anatolian (Garrett 1992; Sideltsev and Molina 2015). Two further points are worth making which account for why unmarked verb movement eventually targeted Fin. First, consider Input Generalization. Roberts (2007a; 2014a) suggests that acquirers, after ascertaining that a particular formal feature is present in the system, will generalize it across as many functional heads as possible that they view as forming a natural class, providing this is compatible with the PLD to which they are exposed. In the light of changes to the syntax of rollup of objects (Kayne 1994) that we discuss further in §6.2, we suggested in §§4.2.1–4.2.2 that a small-scale reanalysis takes place in Classical Latin such that a V–in-situ syntax is reanalysed as having shortdistance V-to-v movement. What might v and Fin have in common which would lead acquirers to view them as a natural class to which similar featural properties could be extended? The proposal put forward in Wolfe (2016a: 489) is that they are both associated with finiteness, but perhaps a more fundamental property is that v, Fin, and arguably Force show phase-head properties.31 This is, therefore, a case of a particular feature—uV—being extended upwards within the clausal spine from one phase-head to another, such that the formal properties of phases become progressively uniform. This is an important factor that we have seen to be at play in the syntax of argument fronting in §3.6 and we will also see that it affects the syntax of subjects (§5.5) as well as the vPperiphery (§6.4). The second point to make regarding the proposed reanalysis

31 For a discussion of Fin and Force as phase-heads, see Roberts (2012a; 2021) and Branigan (2020). The idea that the phase-head status of v, Fin, and Force may drive the emergence of V2 in Romance is outlined in Wolfe (2015e).

142

verb placement and verb movement

is that strings that could plausibly have acted as acquisitional ‘blocks’ to a V2 grammar are integrated into the V-to-Fin grammar, albeit with a distinct underlying structure. Considering SOV orders, whilst their frequency is greatly diminished in texts like the Peregrinatio (Ledgeway 2017b), both Danckaert (2015a) and Adams (2016: 343) have suggested that there is not a neat, uniform decline in SOV orders, even in certain texts which are associated with more colloquial or innovative syntactic characteristics. Whilst this would not be true of a Force-V2 grammar, a Fin-V2 grammar can accommodate such structures, as it permits the co-occurrence of a topic and focus before the finite verb. Clear-cut cases of multiple argument-fronting V3* orders are not frequent in Early Old French but are nevertheless found: (87)

la dame un suen escrin desserre the woman a her case unlock.3sg ‘The lady unlocks a jewel-case of hers’ (OFr. Yvain 2960, 12th century)

As such, the v-to-Fin reanalysis would entail the remaining SOV orders being reanalysed upwards as instances of Topic + Focus + VFin . Whilst certain strings would not be suitable for reanalysis along these lines, many would, given the established partial overlap between the notion of ‘subject’ with that of ‘topic’ and ‘object’ with ‘focus’.32, 33 (88)

a. … [vP … Subject … Object … [v VFin ] [VP …]] b. … [Topic Subject [Focus Object [Fin [Fin VFin ] [TP …]]]]

V1 orders present a similar case. In late Latin there is evidence that finite Vto-Fin movement has become systematic, but that Fin does not yet feature an EF, meaning that merger of a phrasal constituent in the left periphery is not yet obligatory. This observation fits with other reconstructions of Proto-Romance, under which an (XP)-VSO syntax is assumed to be a precursor to Medieval Romance V2 (Dardel 1996; Salvi 2004; Wolfe 2015a). This point is important as it means that V1 orders, which are argued to grow in frequency throughout the history of Latin, also do not constitute potential blocks to reanalysis of the grammar as consistent with V-to-Fin movement. As we saw in §4.3.3, these structures then undergo further reanalysis as featuring null elements satisfying

32 See amongst others Li and Thompson (1976), Firbas (1992), Lambrecht (2000), Jacobs (2001), Rizzi (2005), and Dufter and Jacob (2009). 33 For simplicity’s sake I have not specified the exact position of the Classical Latin subjects and objects in (87a), other than to indicate that they occupy specifier positions within the extended v–VPcomplex. Note, however, that if a subclass of subjects and objects could occupy positions within the vP Topic and Focus projections postulated by Belletti (2004) and others, this could have been a factor favouring reanalysis as both the low and high topic/focus projections would form a natural class.

4.6 summary and conclusions

143

Fin’s EF, so as to be consistent with the requirement that a phrasal constituent always be merged in the left periphery alongside verb movement. These observations may have a wider cross-linguistic significance; in Germanic (Walkden 2014) and Celtic (Willis 1998; Meelen 2016; Meelen 2020) the earliest attested forms of V2 system are of the ‘relaxed’ Fin-V2 kind. If the output of an SOV grammar can plausibly be reanalysed as consistent with V-to-Fin movement but not V-to-Force movement, this may account for why the SOV > V-to-Fin pathway is widely attested in Indo-European, whereas the SOV > V-to-Force pathway is not. Overall, I have suggested in this section that the reanalysis of an SOV grammar with marked verb movement to the left periphery to a grammar with unmarked V-to-Fin movement is independently motivated by established principles of syntactic change. Furthermore, I have suggested that the fact that the innovative grammar features V-to-Fin movement, rather than verb movement to a higher left-peripheral head, is due to characteristics of the SOV grammar, which would have blocked reanalysis to a V-to-Force grammar.

4.6.3 From V2 to SVO When we compare our findings on the verb-movement properties of late Latin in §4.2.3 with those of Early Old French in §4.3.2, we observe a large degree of continuity, insomuch as both systems are characterized by V-to-Fin movement in matrix clauses. This low locus of verb movement does not rule out further movement of the verb upwards within the left periphery to Foc, Top, or Force, nor does it rule out V3* orders where multiple constituents co-occur before the finite verb, which we saw in Chapter 3 were frequent in both late Latin and Early Old French texts. The question then arises as to why the Fin-V2 system attested in the 11th and 12th centuries undergoes reanalysis. My proposal here—following ideas outlined in Wolfe (2016a)—is that the decline in initial foci (cf. §3.3.2) may have been the trigger for this change. If both Fin- and Force-V2 systems are found cross-linguistically, the acquirer will need to make use of relatively subtle judgements to reconstruct one system or the other. Given the fact that foci occupy a low position within the left periphery, and that new information foci must obligatorily be adjacent to the finite verb (Cruschina 2012: 106), the string [New Information Focus + VFin ] potentially provides a very significant cue that the grammar being acquired is a Fin-V2 and not a Force-V2 system. Note, furthermore, that a decline in the licensing of left-peripheral foci would also lead to a reduction in V3* orders, which are a hallmark of a Fin-V2 system but heavily restricted in a Force-V2 grammar. The proposal, therefore, is that declining use of the focus

144

verb placement and verb movement

field paved the way for the reanalysis towards a Force-V2 system. However, keeping in mind basic principles of Input Generalization (Roberts 2007a), the shift to a Force-V2 system is also not surprising: the acquirer is simply extending a [uV] feature present on two clausal phase-heads—v and Fin—to the highest phase-head in the clause, Force: (89)

[Frame [Force [Force ] [Topic [Focus [FinP [Fin ][TP [vP [v ][VP [V ]]]]]]]]]

What is less obvious is why acquirers at the end of the 13th century reanalysed verb movement ‘downwards’ again, such that its eventual target was Fin. The conclusion may simply be that the PLDs generated by Fin-V2 and Force-V2 grammars are sufficiently similar in certain instances that the property ends up being relatively unstable. Evidence for this proposal comes from a language like German, where changes in the class of V3* triggers have been observed throughout the language’s history, with Old High German showing a Fin-V2 grammar (Axel 2004; Axel 2007), Modern Standard German a Force-V2 system (Brandner 2004; Frascarelli and Hinterho¨lzl 2007), and evidence emerging from urban vernaculars of innovative Fin-V2 grammars, which have certain similarities to Old High German (Walkden 2015). Aside from the internal change within the V2 system, we know that the very beginning of the 16th century sees the emergence of an SVO grammar with unmarked V-to-T movement. Which factors discussed in this book could have fed this particular reanalysis? In line with Poletto’s (2019) cross-linguistic study of the loss of V2, I suggest that we should not look for a single factor given the relatively heterogeneous typology of V2 systems which is now acknowledged. Rather, consider the following observations, which may all have contributed to the paucity of evidence for a V2 analysis of the grammar:

(i) After 1180, we have seen that the attestation of foci—new information foci in particular—declines. If we adopt the standard assumption that foci, in contrast to topics, are always moved to the left periphery (cf. Benincà and Poletto 2004 and discussion in §3.1), this undermines one key piece of evidence for V-to-C movement, specifically a moved constituent in the CP preceding a finite verb. Given that non-subjects are more typically focal than subjects, the decline in initial foci will lead

4.6 summary and conclusions

145

to a decline in the proportion of non-subjects found in the prefield, which we saw in §3.4.1 is a change already observable when comparing preverbal constituents in Middle French texts with their Old French counterparts. (ii) We noted in §4.3.1.2 the consensus view that Germanic inversion provides a particularly clear piece of evidence that the finite verb targets a C-related head in Old and Middle French, and Medieval Romance in general. However, we will see in §5.3 that this position becomes qualitatively specialized in hosting discourse-active subjects in Later Old French and Middle French and, furthermore, that there is a concomitant decline in its use relative to inversion structures which are ambiguous between a V-to-C and a V-to-T/Subj analysis. A diachronic decline in a key exponent of a V2 grammar thus undermines the acquirer’s potential cues for V-to-C movement. (ii) Alongside the decline in focal constituents, the Middle French period also sees a rise in left-dislocation structures which are derived via basegeneration (cf. Kroch 1989; Kroch and Santorini 2009, and §3.4). If—as has been suggested by Holmberg (2015)—a V2-related EF must be satisfied by internal merge of a constituent rather than base-generation, a rise in these structures will decrease the evidence the acquirer has access to that internally merged constituents reach the left periphery to satisfy V2. Coupled with the decline in foci, the acquirer will thus no longer postulate an EF on Fin. Given that V1 orders are almost entirely absent in the system by this point (§4.4.2), they are not going to then reanalyse the system as Fin-VSO; a ‘downstairs’ reanalysis to the T-layer, however, is compatible with the wide range of initial subjects they are exposed to, alongside the base-generated constituents occurring preverbally. My suggestion is that none of these factors in isolation would have destabilized V2, but that taken together they conspire towards a grammar which can plausibly be reanalysed as SVO, with unmarked V-to-T movement and a small number of residual V-to-C contexts.3⁴ This is of course exactly what we find in Renaissance and Classical French texts. Given that sporadic instances of nominal inversion (§4.5.2), Germanic inversion (§5.5), and argument fronting (§3.4.4) are found in the 16th century in particular, I do not rule out the possibility that a transitional stage existed during which the finite verb could 3⁴ My proposal on V2-loss in French here is thus the inverse of Weerman’s (1989) ‘V2 Conspiracy’ on what is necessary to trigger a V2 grammar.

146

verb placement and verb movement

optionally target Fin in declaratives. However, the overwhelming attestation of pronominal inversion from the 17th century onwards strongly suggests that the finite verb raises to T in the unmarked case, and Subj in a subset of marked environments in certain varieties.3⁵ V-to-Force movement in imperatives can be considered the great ‘survivor’ of the history of French, having been attested consistently since Latin.

4.6.4 Conclusion Viewed from a broad typology of verb movement, the data we have surveyed in this chapter falls into three broad classes: an archaic V-in-situ stage, where the finite verb either stays in V or moves only as high as v; a V2 stage, where the finite verb raises either to one or both left-peripheral phase-heads; and a V-to-T stage, where the finite verb typically targets a position within the inflectional layer. Making use of a form of parameter hierarchy (Biberauer and Roberts 2014; Biberauer and Roberts 2015; Roberts 2019), we can schematize the variation in unmarked verb movement as in Figure 4.1. The term ‘phase-head’ should be understood as referring to v, Fin, and Force. Do phase-heads bear a [uV] feature? Yes No V-in-Situ Do all phase-heads bear a [uV] feature? Archaic Latin V-to-T Yes No Renaissance to Modern French V-to-Force Do v and Fin bear a [uV] feature? Later Old French Yes V-to-Fin Later Latin Early Old French Middle French

No Does v bear a [uV] feature? Yes V-to-v Classical Latin

Fig. 4.1 A parameter hierarchy for unmarked verb movement 3⁵ The notion that the Subj head could bear a [Q] feature and attract the finite verb in interrogatives may seem incompatible with the general cartographic approach pursued in this book. However, the idea that features predominantly associated with the C-system can be either ‘shared’ or ‘donated’ with T-related heads is not a new one (Ouali 2008; Miyagawa 2010; Biberauer and Roberts 2010; Jiménez-Fernández and Miyagawa 2014). With the notable exception of imperatives, the generalization would therefore be that the most ‘innovative’ varieties of French today have seen a total downstairs reanalysis of the locus of finite-verb movement, EPP-effects, and the locus of marked verb movement, due to features previously hosted on C-related heads now ‘donated’ to T-related heads.

4.6 summary and conclusions

147

Following Schifano (2018) I suggest that languages with V-to-T movement in Romance instantiate a distinct mesotypology, the details of which are not vital for our analysis of French,3⁶ as we have seen evidence which suggests the finite verb has raised to a high position within the inflectional layer since the 16th century.

3⁶ Though see some comparative remarks on this issue in §7.4.

5 The subject system 5.1 Introduction: the subject system of French and Romance 5.1.1 The fine structure of the subject layer Within formal syntactic theory a considerable amount of attention has been focussed on the syntax of subject positions. The important point to note for our purposes is that a number of refinements are necessary to the classic Government and Binding Theory view that all subjects raise to Spec-IP/TP in order to satisfy the Extended Projection Principle (Chomsky 1982). Within feature-based Minimalist approaches to parametric variation (Borer 1984; Chomsky 1995; Biberauer 2010b), a movement-triggering diacritic can in theory be associated with any head in the clause, which suggests that T is not exceptional in the way it was originally conceived of as being. Consciously simplifying, this yields a typology where a movement diacritic can be associated with v, where subjects remain in situ, T, where a language shows ‘classic EPP-effects’, and C, which is part of the featural composition of a classic V2 system (Holmberg and Platzack 1995; Cardinaletti and Roberts 2002; Holmberg 2015; Wolfe and Woods 2020). Latin—with its unmarked SOV syntax—is analysed by Ledgeway (2012a) as a language lacking EPP-effects at the TP- or CP-level, where all arguments can remain within the extended vP-complex.1 Modern French and Modern Italian, however, show robust evidence of EPPeffects at the TP-level, despite the presence of the null-subject property in the latter (Sheehan 2006; Sheehan 2010). By contrast, the movement diacritic in V2 Raeto-Romance varieties is associated with C, as was the case in Old and Middle French (Roberts 1993; Vance 1997; §4.3) and other Medieval Romance V2 languages (Benincà 2004; Benincà 2006; Wolfe 2018a). The important point for our understanding of the French subject-system and its evolution is that we do not predict finding that subjects necessarily raised to the

1 On languages featuring vP-internal subjects, see also Koopman and Sportiche (1991), McCloskey (1996), and Roberts (2005), among many others.

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0005

5.1 introduction

149

T-layer throughout the language’s history; the notion that the locus of EPPeffects can change diachronically has been explored in a number of languages, including Neapolitan within Romance (Ledgeway 2007; Ledgeway 2008) and English within Germanic (Roberts 1996; Biberauer and Roberts 2005; Biberauer 2010a). The notion that there is only a single projection associated with subjecthood in the T-layer has also been challenged, most notably in work by Cardinaletti (1997; 2004; 2010; 2021). Focussing principally on evidence from Modern Romance, she suggests that lexical subjects, strong pronouns, and other subject-like constituents constituting the Subject of Predication raise to a dedicated specifier of a SubjP projection, whereas a lower projection—her AgrSP—hosts weak pronouns and is primarily associated with ϕ-agreement (Cardinaletti 2004: 123–128). Here and elsewhere in this chapter Cardinaletti’s schema is adopted, but TP is used in lieu of AgrSP: (1)

… [FinP [SubjP Subject1 [TP Subject2 [vP …]]]]

We should also note in relation to subjects within the extended vP-complex the seminal work by Belletti (2001; 2004; 2006; 2008), which suggests that Modern Romance postverbal subjects such as the unaccusative subject in (2) are not necessarily in their base-generated position (Burzio 1986) but can also raise to a low topic-focus field at the periphery of the vP (cf. also Jayaseelan 2001). We will see that this insight can fruitfully be applied to a number of phenomena in the history of French, where the syntax of the low vP-periphery and high CP-periphery interact. (2)

… [FinP [SubjP [TP pro [T

arriva] [vP Topic [vP Focus Gianni arrive.3sg Gianni

[vP [VP …Gianni]]]]]]] ‘Gianni is arriving’

(Modern Italian)

Finally, before considering the necessary background on null subjects, we note that many Romance languages feature weak pronominal or clitic subjects which have been the centre of considerable discussion (Kayne 1975; Kayne 1983b; Kayne 1991; Rizzi 1986a; Brandi and Cordin 1989; Poletto 1995; Poletto 2000; Cardinaletti and Starke 1999; Sportiche 1999; Anderson 2006; Benincà 2013; Roberts 2014b; Poletto and Tortora 2016). In diachronic terms, a number of Northern Italian Dialects have witnessed a grammaticalization process, whereby originally strong subject pronouns undergo progressive weakening such that they see a gradual loss of phonological and syntactic autonomy (Roberts and Roussou 2002: §4.5), eventually becoming reanalysed

150

the subject system

as agreement morphology in certain cases where they obligatorily double a preverbal subject (3) (cf. Poletto 1995 on this process in Venetian). (3)

a. Al pi al mangia al pom the boy scl eat.3sg the apple ‘The boy eats the apple’ (Malonno, Eastern Lombard, Poletto 2000: 142) b. Vargu al rierà n ritardo someone scl arrive.3sg.fut in late ‘Someone will arrive late’ (Malonno, Eastern Lombard, Poletto 2000: 142)

Kayne (1975) draws an explicit parallel between Modern French subject pronouns and the subject clitics found in Northern Italian dialects, but suggests that whereas ‘true’ subject clitics occupy head positions,2 this is not true of a French subject ‘clitic’ such as il ‘he’, which constitutes a phrasal category (Rizzi 1986b; Brandi and Cordin 1989; Cardinaletti and Starke 1999). French data figure in the empirically rich analysis of pronominal typology in Cardinaletti and Starke (1999), where they ultimately conclude that French ‘clitic’ forms are, in fact, distinct from subject clitics, and constitute a specific class of weak pronouns. We will see in §§5.3–5.5 that the progressive weakening of originally strong pronominal forms can be tracked diachronically and is also an area of synchronic variation and change.

5.1.2 Null subjects in French and Romance Full null-subject Romance languages such as Modern Italian, Modern Peninsular Spanish, and Modern European Portuguese have featured prominently in the literature as paradigm instantiations of the null-subject property, linked in the research from the 1980s onwards to a positive setting of the null-subject parameter (for discussion see Roberts and Holmberg 2010). As well as featuring referential and non-referential null subjects (4, 5), these languages also show so-called free inversion (5), rich verbal agreement paradigms, clitic climbing, and complementizer-trace violations:3 (4)

a. Llueve rain.3sg ‘It’s raining’

(Modern Peninsular Spanish)

2 These head positions are within an articulated field within the C- and T-layer (Poletto 2000). 3 For the full range of Romance data, see Sheehan (2006; 2010; 2016).

5.1 introduction

b. Chove rain.3sg ‘It’s raining’ c. Piove rain.3sg ‘It’s raining’ (5)

a. Hablaba Juan speak.3sg.pst Juan ‘Juan was speaking’ b. Falava o João speak.3sg.pst the João ‘João was speaking’ c. Parlava Gianni speak.3sg.pst Gianni ‘Gianni was speaking’

151

(Modern European Portuguese)

(Modern Italian)

(Modern Peninsular Spanish)

(Modern European Portuguese)

(Modern Italian)

Following seminal work by Rizzi (1982; 1986b), these properties were linked to the availability of a null pronoun—pro—which is able to occupy Spec-IP (later TP) in the languages in question. The role of rich agreement in licensing pro has been a perennial topic of discussion in research on null subjects, but has been challenged in a number of accounts, as have other correlates originally assumed to be integral to a positive setting of the parameter. As highlighted very clearly in Biberauer (2018) and other contributions in Cognola and Casalicchio (2018), it is also becoming increasingly clear that the typological scope of variation within null-subject and null-argument systems is considerable, although this does not undermine the central intuition in Roberts and Holmberg (2010) that such variation is systematically structured. Turning to the scope of variation in Romance, in addition to the well-described full nullsubject languages, we can also observe a generalized null-argument system in Latin (Luraghi 1997; Vincent 2000), where objects and other arguments can be realized as null in addition to subjects, as well as partial null-subject systems conditioned by clause-type and person in Medieval Romance (Adams 1987c; Benincà 2004; Poletto 2014), Raeto-Romance (Haiman and Benincà 1992), Corsican (Ledgeway 2013), and Brazilian Portuguese (Barbosa 1995; Holmberg, Nayudu, and Sheehan 2009). A changing theoretical landscape alongside a much broader empirical picture of variation has led to a multitude of approaches to formal modelling of

152

the subject system

null-subject grammars.⁴ For the purposes of our discussion in this chapter 1 draw principally on the deletion account of null arguments put forward in Roberts (2010b). Under this approach, if in a probe-goal Agree relation the features of the goal form a proper subset of the features of the probe, the goal is rendered defective and deletes at Phonetic Form due to a process of Chain Reduction in the sense of Nunes (2004). More specifically, in a classic nullsubject language, pro would constitute a defective goal in light of bearing the features [ϕ, D] which are also present on the probe—T—owing to the richness of agreement in such languages.⁵ Two extensions of Roberts’ approach are worth mentioning here: firstly, Roberts (2010b: 78) explicitly suggests that generalized null arguments obtain when v bears ϕ- and D-features, as an object pro would also constitute a defective goal in such conditions; we will see in §5.2 and §5.7 that this prediction holds robustly for Latin. Secondly, following ideas developed in Wolfe (2015b; 2016b; 2018a), which in turn draw on the insight that pro-licensing in some languages involves a topic-head (Frascarelli 2007; Walkden 2013b; Poletto 2020), I suggest that when pro is also endowed with a [+Topic] feature it constitutes a defective goal in relation to the left-peripheral Top head.

5.2 The Latin subject system 5.2.1 Preverbal subjects Given the wide range of word-order variation found within Latin texts, it should come as no surprise that the analysis of the subject system has been a controversial area of Latin scholarship. However, despite the inherent difficulties in using formal diagnostics to identify the structural positions occupied by subjects in a dead language (for discussion cf. Danckaert 2017b: 126–127), we can reach an analysis which sheds light on subsequent developments in the history of French outlined in §§5.3–5.6. The consensus view is that the preverbal position for the subject is unmarked, and argued to be so in Latin of all periods (Hofmann and Szantyr 1965: 396; Lehmann 1972: 272; Cabrillana 1993: 254; Polo 2005: 384; Devine ⁴ For overview and discussion of the vast literature in this area, see the chapters in Roberts (2007b), Biberauer, Holmberg, Sheehan, and Roberts (2010), and Cognola and Casalicchio (2018), alongside Camacho (2013), Sheehan (2016), and Roberts (2019). ⁵ Note that the role of rich agreement in licensing null subjects is also maintained by Alexiadou and Anagnostopoulou (1998), albeit with distinct implementation which dispenses with pro.

5.2 the latin subject system

153

and Stephens 2006: 37; Danckaert 2012: 15). Cabrillana (1993: 254), for example, reports that 76.66% of her corpus of Cicero has preverbal subjects relative to 23.24% postverbal, with preverbal subjects constituting 80.24% of the embedded-clause sample. The notion that preverbal subject-placement is unmarked also stands irrespective of individual scholars’ takes on the debate as to whether Latin was SOV, SVO, or neither (Pinkster 1991; Polo 2005; Ledgeway 2012a: 60–61). Consider in this regard Table 5.1, which represents Polo’s (2005: 378) analysis of Petronius’s Cena Trimalchionis in the Satyricon (1st century ad), where we find preverbal subjects as the overwhelmingly preferred option in both SVO and SOV contexts. Table 5.1 Subject placement in the Cena Trimalchionis (Polo 2005) Order

All clauses

Main clauses

Subordinate clauses

(S)VO (S)OV VSO VOS OSV OVS

188 578 10 8 28 11

23.0% 70.0% 1.2% 1.0% 3.5% 1.3%

132 345 10 7 18 7

25.4% 66.5% 2.0% 1.3% 3.5% 1.3%

56 233 0 1 10 4

18.5% 77.0% 0.0% 0.3% 3.2% 1.0%

Total

823

100.0%

519

100.0%

304

100.0%

However, we if we adopt the assumption set out in §4.2 that the finite verb in Classical Latin stays within the v–VP complex in the unmarked case (Ledgeway 2012a), ‘preverbal’ placement of the subject could equate to a number of structural options, including staying in situ in Spec-vP (6a), raising to a position within the T-layer (6b), or occupying a left-peripheral topic or focus position (6c):⁶ (6)

a. [CP [TP [vP Subject [v V] ]]] b. [CP [TP Subject [vP [v V] ]]] c. [CP Subject [TP [vP [v V] ]]]

Our findings on the structure of the Latin left periphery outlined in §3.2 show quite clearly that left-peripheral topicalization and focalization of subjects—the schema in (6c)—was possible in Classical Latin and widely attested, with some—such as Devine and Stephens (2006: 36)—viewing this as the default: ‘[t]he subject is mostly placed in a left peripheral position’. Informationally focussed subjects can clearly occupy a left-peripheral ⁶ Note that the schema in (6) is simplified, and as such I do not indicate the possibility of verb movement (cf. §4.2) or mark lower copies of the moved subject in (6b, 6c).

154

the subject system

position (7), with preverbal placement often found in answer to a ‘what happened?’ question (de Jong 1989: 536–537; Devine and Stephens 2019: 48). (7)

Eo L. Caesar adulescens uenit … there.adv Lucius Caesar.nom young come.3sg ‘Thither comes the young Lucius Caesar’ (Latin, Caes. B.C. 1.8.1, Spevak 2008: 119)

However, whilst Classical Latin clearly has an active focus field within the CP (Salvi 2005: 438–439), Bolkestein (1995: 40)—on the basis of quantitative data—suggests that informationally new preverbal subjects are ‘relatively rare compared to the same constellation in VS order’. We return to alternative positions for focal subjects in §5.2.2. Turning to the ability of subjects to raise to the topic field, several scholars note that subject pronouns frequently occur in clause-initial position (Marouzeau 1938: 55; Bolkestein 2000: 118; Clackson and Horrocks 2007: 255; Bauer 2009: 276; Spevak 2010: 76–77), and Bolkestein (2000) stresses that SOV and SVO orders with both nominal and pronominal subjects overwhelmingly involve a topical subject in initial position of the clause, which either corresponds to accessible (8a) or discourse-active information, as we see in (8b) with the second occurrence of Bomilcar. (8)

a. Caesar eius dextram prendit Caesar.nom his right-hand.acc take.3sg ‘Caesar takes his right hand’ (Latin, Caes. B.G. 1.20, Devine and Stephens 2006: 38) b. Bomilcari … imperat pretio … insidiatores Bomilcar.dat command.3sg by-payment ambushers.acc Massivae paret … Bomilcar mature regis for-massive procure.3sg Bomilcar.nom quickly of-king mandata exequitur orders.acc carry-out.3sg ‘He commands Bomilcar … to produce by payment… ambushers for Massiva. Bomilcar hastens to carry out the king’s orders’ (Latin, Sall. Jug. 35.4–5, Spevak 2010: 62)

The detailed analysis put forward in Danckaert (2017b) suggests that Latin subjects could also raise to the inflectional layer, which can equate to either preverbal or postverbal placement depending on the position of the finite verb. He suggests, on the basis of data from periphrastic constructions with finite auxiliaries, that subject-placement within the T-layer represents an innovation

5.2 the latin subject system

155

in the grammar during the classical period, such that a TP-internal position for subjects such as that found in (9) cannot be considered a property of all Classical Latin texts. (9) quin se ille Marcellus tantus uir that refl.cl that.nom Marcellus.nom such.nom man.nom sic ad tolerandum aequo animo exilium so to bear.gdv.acc equal.abl mind.abl exile.acc saepe often ‘that the great man Marcellus encouraged himself in this manner, in order to bear his exile with a calm mind’ (Latin, Sen. Dial. 12.9.7, Danckaert 2017b: 156) This finding that Latin lacks a dedicated Spec-TP subject position is unsurprising; the notion that preverbal subjects always target such a position crosslinguistically is—as we saw in §5.1—probably not correct. Furthermore, other languages argued not to make use of the canonical subject position, such as Old English (Biberauer and Roberts 2005) and Old French (§5.3), also belong to a class where discourse pragmatics plays a major role in shaping word order. Overall, we report the expected finding that subjects in Latin—as in French for much of its history (§3, §§5.3–5.6)—can raise to the left periphery given an appropriate pragmatic trigger. However, the possibility for subjects to raise to a position in the T-layer can be viewed as an innovation which gains ground during the classical period. We will now see that there is also evidence for diachronic change in Latin regarding postverbal subjects.

5.2.2 Postverbal subjects In a range of Modern Romance languages, unaccusative and passive subjects frequently remain in a vP-internal position (10) (Burzio 1986), a possibility which still exists in formal Modern French subject to definiteness effects (11) (Jones 1996: §3.4). (10)

a. È arrivato be.3sg arrive.ptcp ‘Gianni has arrived’ b. Ha llegado have.3sg arrive.ptcp ‘Juan has arrived’

Gianni Gianni (Modern Italian) Juan Juan (Modern Spanish)

156

the subject system

c. Ha arribat en Joan have.3sg arrive.ptcp the Joan ‘Joan has arrived’ (11)

Il est arrivé un homme there be.3sg arrive.ptcp a man ‘A man arrived’

(Modern Catalan)

(ModFr.)

In Classical Latin unaccusative and passive subjects can also occur in this position, and thus follow the finite verb in v/V (cf. §4.2.1) (Pinkster 1991: 78; Bolkestein 1995: 35; Bauer 2009: 280; Ledgeway 2012a: 109). This leads Ledgeway (2012a: 109) to highlight ‘the distinctly unmarked nature of postverbal undergoers in this period (be they subjects or objects)’. Consider the example in (12) in this regard: (12)

erant in ea legione fortissimi viri be.3pl.pst in that.abl legion.abl very-brave men.nom ‘There were very brave men in that legion’ (Latin, Caes. B.G. 1.42, Pinkster 1991: 78)

However, although unaccusative/passive subjects favour postverbal positioning in Latin, Bolkestein (1995: 38) stresses that pragmatics plays a more significant role in determining subject position. Aside from purely syntactic conditions for postverbal subjects, presentatives and orders where the subject receives a narrow focus reading strongly favour postverbal positioning (Ostafin 1986: 156; Pinkster 1991: 78; Polo 2005: 402; Bauer 2009: 280). Additional evidence that a low subject position can encode focus comes from the fact that informationally rich or heavy/complex subjects often favour this position (13) alongside proper names (14) which often encode new information (Cabrillana 1993: 244; Bolkestein 1995; Bauer 2009: 280). (13)

Concurrebant legati, centuriones tribunique run.3pl.pst legates.nom centurions.nom tribunes.nom-and militum, ne dubitaret proelium committere of-soldiers neg doubt.3sg.ipfv.sbjv battle.acc join.inf ‘Senior commanders, centurions, and officers ran to him, urging him not to hesitate to join battle’ (Latin, Caes. B.C. 1.71.2, Spevak 2010: 176)

(14)

Non respuit condicionem Caesar neg reject.3sg condition.acc Caesar ‘Caesar did not reject the proposal’ (Latin, Caes. B.G. 1.42, Devine and Stephens 2006: 38)

5.2 the latin subject system

157

For the sake of completeness, we should note that postverbal subjects can also be topical in nature (Bolkestein 1995: 39), although such cases are not extensively discussed in the literature. Danckaert (2017b: 145) thus concludes that both Belletti’s (2001) vP topic and focus projections are active in Classical Latin and can host postverbal subjects. Overall, however, the observation stands that VS order in Classical Latin appears to canonically encode focus, either broadly or on the subject constituent specifically. This—as we will see— provides an important point of continuity in terms of the syntax–pragmatics mapping for subjects in Early Old French. Unsurprisingly in the light of the emergence of a V2 syntax in late Latin and Early Romance (Benincà 2004; Benincà 2006; Ledgeway 2017b; Wolfe 2018a, and §3.3, §4.3), there are indications that verb–subject inversion becomes a more widespread phenomenon in colloquial and late Latin texts, with Bauer (2009: 280) stating that ‘[s]ubject inversion is attested in Classical Latin, but spread in Vulgar and Late Latin’. This particular point, noted by others such as Va¨a¨na¨nen (1987), follows naturally from the verb-movement properties of colloquial and late Latin set out in §4.2, where we suggested that unmarked verb movement is eventually reanalysed upwards to Fin in late Latin. This necessarily entails that any non-fronted subjects occurring in TP or vP will appear postverbally and this is precisely what Ledgeway (2017b: 180–181) reports in his analysis of the Peregrinatio, which shows extensive verb–subject inversion accounting for 26.5% of all V1–V3 clauses (253/955): (15)

a. Illud etiam retulit sanctus episcopus: that.acc also tell.3sg.pst holy.nom bishop.nom ‘The holy bishop also told me: …’ (Latin, Peregrinatio 19.14, Ledgeway 2017b: 179) b. et sic dicet episcopus stans and thus say.3sg bishop.nom stand.ptcp.nom.sg benedictionem super cathecuminos blessing.acc over catechumens.acc ‘And the bishop stands and says the blessing over the catechumens’ (Latin, Peregrinatio 24.6, Ledgeway 2017b: 179) c. sic dicitur ymnus thus say.pass.3sg hymn.nom ‘and thus the hymn is said’ (Latin, Peregrinatio 36.3, Ledgeway 2017b: 180)

However, Ledgeway’s (2017b) formal analysis differs somewhat from Danckaert’s (2017b) in that he assumes all postverbal subjects to be in a

158

the subject system

vP-peripheral or vP-internal position both in matrix and embedded clauses, thus referring to the ‘lack of a SpecTP position about the v-VP complex’ (Ledgeway 2017b: 186). Taking the observations outlined in this section on both Classical and late Latin together, we can sketch an outline of how the system of overt subjects functioned. In the classical language, subjects could appear in the left periphery when topicalized or focalized, or in a low left periphery of the type proposed by Belletti (2001). We also assume that subjects could remain within the vP, either in Spec-vP, or within the VP, which is likely to be the case in some unaccusative structures. This gives a schema seen in (16), which confirms the intuition that Latin subject-placement is highly variable and conditioned by a range of pragmatic and syntactic factors: (16)

[Topic (Subject) [Focus (Subject) [FinP [SubjP/TP [vP Topic (Subject) [vP Focus (Subject) [vP (Subject) [VP [V VFin (SubjectUnaccusative )]]]]]]]]]

In colloquial and late Latin, however, the verb does not remain in the vP but instead systematically raises to Fin. The immediate consequence of this is that any subject occurring in a position lower than the moved verb will appear ‘inverted’. Keeping in mind the disagreement between Ledgeway (2017b) and Danckaert (2017b) on the genesis of a subject position within the inflectional layer in late Latin, we indicate it here and note that, were such a position to exist, subjects in this position would also appear postverbally: (17)

[Topic (Subject) [Focus (Subject) [FinP [Fin VFin ][SubjP/TP (Subject) [vP Topic (Subject) [vP Focus (Subject) [vP (Subject) [VP … (SubjectUnaccusative )]]]]]]]]

5.2.3 Null arguments in Latin As we will see in §5.3, Early Old French shows the properties of a full nullsubject language and we see this type of system uniformly attested across the earliest Romance vernaculars.⁷ Latin too has been classed as a full null-subject system, a characterization potentially strengthened by its rich system of verbal inflection and licensing of widespread inversion structures (cf. §5.2.2 and discussion in Devine and Stephens 2006: 152), which we saw in §5.1.2 are canonically associated with null-subject systems in Romance and elsewhere ⁷ See amongst very many others Benincà (2004: 263–266), Poletto (2014: 11–16, 17–22), Ledgeway (2009b: 751–754), and Wolfe (2015f) on Old Italo-Romance and Fontana (1993: 100–111) and Ribeiro (1995) on Old Ibero-Romance.

5.2 the latin subject system

159

(Rizzi 1982; Rizzi 1986b). In (18, 19) we see, therefore, that Latin licenses null subjects in both matrix and embedded clauses (18), as well as in referential (19a) and non-referential contexts (19b): (18)

Peto abs te, ut haec diligenter cures ask.1sg from you.abl that these.acc careful.adv care.sbjv.3sg ‘I ask you to take great care of this’ (Latin, Cic. Att. 1.9.2, Danckaert 2012: 214)

(19)

a. Epistulam tibi misi letter.acc you.dat send.1sg.pst ‘I have sent you a letter’ (Latin, Cic. Att. 15.5.1, Spevak 2010: 96) b. Erat vallis inter duas acies be.3sg.pst valley.nom between two lines ‘There was a valley between two lines’ (Latin, Caes. Civ. 2.34.1, Spevak 2010: 33)

It might thus be tempting to conclude that there is straightforward continuity between the full null-subject system of Latin and the full null-subject system observable in Early Old French. This analysis would, however, miss the important fact that Latin—along with other Early Indo-European varieties⁸— also licensed null objects and other forms of null arguments, in contrast to later stages of French when they are generally restricted (Schøsler 2000a). More specifically, Luraghi (1997) identifies a range of structural configurations favouring null objects, which include coordination (20), where an indefinite form would be expected as an overt alternative (21), or when there is clearly topic-continuity with an entity in the preceding portion of text or clause (22): (20)

Domum meam maioribus praesidiis munivi house.acc my.acc greater.abl defences.abl protect.1sg.pst atque firmavi and fortify.1sg.pst ‘I have protected and fortified my home with better defences’ (Latin, Cic. Cat. 1.4, Luraghi 1997: 242)

⁸ For discussion of null arguments in Early Germanic, see Sigurdsson (1993) and Walkden (2014), in Ancient Greek, see Luraghi (2003), and in Hittite, see Inglese, Rizzo, and Pflugmacher (2019). For a comparative discussion of Indo-European see Haug (2012). The widespread attestation of generalized null arguments across early Indo-European varieties leads both Luraghi (2004: 253) and Walkden (2014: 230) to suggest that Proto-Indo-European may have been a generalized null-argument language.

160

the subject system

(21)

Milites imperat; mittunt soldiers.acc ask.3sg send.3pl ‘He asks for soldiers; they send (some)’ (Latin, Caes. B.C. 1.15, Luraghi 1997: 245)

(22)

… ut tres legati regem that three commissioners.nom the-king.acc reducerent; … ut tu sine restore.3pl.ipfv.sbjv that you without exercitu reduceres; … ut army.abl restore.2sg.ipfv.sbjv that Pompeius reduceret Pompey.nom restore.1sg.ipfv.sbjv ‘that three commissioners should restore the king; that you should restore (him) without an army; that Pompey should restore (him)’ (Latin, Cic. Fam. 1.2.1, van der Wurff 1996: 344)

In terms of Roberts’ (2010c) typology of null arguments outlined in §5.1.2 and our assumptions on verb movement set out in §4.2.2, this finding that other arguments besides subjects can be realized as null in Latin is unsurprising. Roberts (2010c: 190) specifically states that under his assumptions on the nature of defective goals, low verb-movement languages are predicted to license generalized null arguments. As such, the null-argument status of Latin is confirmed both by a consistent set of formal assumptions, and a careful analysis of the relevant empirical picture. We therefore conclude that whilst Latin licenses null subjects in the same range of environments we would expect in a full null-subject language, the fact that non-subjects can also be realized as null suggests that classing Latin as a generalized null-argument language is a more accurate characterization of its place on the null-argument typology.

5.2.4 Summary Our findings on the Latin subject system are in keeping with the significant role pragmatics plays in constraining Latin syntax and show a number of revealing points of contrast with Modern French and Romance. We have seen that in Classical Latin subject expressions can target a range of positions within the CP and vP, with postverbal subjects becoming more frequent in late Latin, alongside the possibility of TP-internal subjects. Turning to null subjects, we have seen that Latin permits not only null subjects but also a range of null

5.3 the old french subject system

161

arguments, suggesting that it occupies a distinct point on the null-argument typology from later stages of French.

5.3 The Old French subject system 5.3.1 Preverbal subjects As was the case in Latin and as is the case in other V2 languages, overt subjects can occur preverbally in Old French and thus occupy a left-peripheral position in the topic-focus fields (23): (23)

Li reis Marsilie esteit en Sarraguce the king Marsile be.3sg.pst in Zaragoza ‘King Marsile was in Zaragoza’ (OFr. Roland 10, 12th century)

However, a question arises as to whether the conditions under which Old French subjects raise to the left periphery remain stable diachronically or show variation. The question is pertinent in the light of our discussion of the Old French left periphery in §3.3, where we saw that there are substantial differences in the types of focal constituents, in particular, that can occur preverbally and Early and Later Old French. Our observations sit against a backdrop of other studies showing diachronic variation in the structure of the left periphery, with some suggesting that preverbal complements lose a categorical discourse value (Larrivée 2019), others suggesting that the left periphery is essentially a dedicated topic position by the 13th century (Rinke and Meisel 2009; Wolfe 2016a), and another group offering a more nuanced view, where discourse-old information is increasingly encoded preverbally, albeit not categorically, from c.1180 onwards (Steiner 2014; Labelle and Hirschbu¨hler 2018; Wolfe 2020b). Although several of these studies have focussed on preverbal complements, any change in the left-peripheral syntax–pragmatics mapping should also affect subjects under the standard assumption that all types of left-peripheral arguments target the same portion of the C-layer.⁹ In order to investigate quantitatively whether information-structural status affects the distribution of subjects in Old French, a small corpus (200–247) of clauses was tagged for four texts: Roland (c.1100), Eneas (c.1155), Clari (c.1205), and La Queste (c.1225). Building on recent work on ⁹ The exception to this generalization would be if subject- and non-subject-initial V2 clauses have distinct derivations (Travis 1984; Craenenbroeck and Haegeman 2007; Greco and Haegeman 2020). I do not, however, pursue this hypothesis here or in the substantive discussion of V2 in Chapters 3–4.

162

the subject system

information-structural annotation of historic texts (Hinterho¨lzl and Petrova 2009; Labelle and Hirschbu¨hler 2018) all subjects were then tagged as quantified, pronominal, or nominal (i.e. DPs or NPs). In the latter instance, subjects were tagged using a tripartite typology under which they were assigned the status of new, active (Prince 1981: 243; Chafe 1987), or accessible in the sense of Lambrecht (1994), where referents are not present in the same thematic paragraph of text (cf. Fleischman 1990) but refer to commonly recognized entities such as ‘the King’ or ‘God’. The findings are reported in Table 5.2. Table 5.2 Preverbal subjects in Old French New Roland Eneas Clari La Queste

16 16 10 7

20.0% 25.8% 14.3% 9.9%

Accessible

Active

Pronominal Quantified

Relative

22 17 7 10

10 5 20 6

32 18 25 48

0 1 3 0

27.5% 27.4% 10.0% 14.1%

12.5% 8.1% 28.6% 8.5%

40.0% 29.0% 35.7% 67.6%

0 5 5 0

0.0% 8.1% 7.1% 0.0%

0.0% 1.6% 4.3% 0.0%

In previous work, Labelle and Hirschbu¨hler (2018) have already suggested that preverbal subjects may be less susceptible to change than preverbal complements in Old French. It is therefore expected that the data show evidence of relatively gradual change, which is indeed the case. Firstly, there is a small decline in the proportion of preverbal subjects which encode new information, as we also saw was the case for preverbal objects in §3.3.2 (cf. also MarchelloNizia 1995: 99–101), which account for only 9.9% of preverbal subjects in the latest text, La Queste: (24)

a. Mur ne citet n i est remés a wall nor city neg loc.cl- be.3sg remain.ptcp to fraindre besiege.inf ‘Not a wall or town remains to be besieged’ (OFr. Roland 5, 12th century) b. une vague li vint desore a wave it.cl come.3sg.pst on-top ‘A wave came over the top’ (OFr. Eneas 245, 12th century)

There also appears to be a split between the 12th-century and 13th-century texts as regards accessible subjects, which are less frequent in the later texts: 27.4% and 27.5% vs. 10.0% and 14.1%. Furthermore, we observe that the proportion of preverbal subjects constituted by discourse-active nominal subjects increases (25), with the notable exception of La Queste, which does however

5.3 the old french subject system

163

feature a very high proportion of preverbal pronominal subjects which are themselves highly likely to refer to an already activated entity in the text (26): (25)

Et quant li rois ot ceste parole si se and when the King hear.3sg this speech si refl.cl repent … et il dist repent.3sg and he say.3sg.pst ‘When the king heard this speech he repented … and he said …’ (OFr. La Queste 8, 13th century)

(26)

Quant les letres furent lutes when the letters be.3pl.pst read.ptcp et li dux les eut and the duke them.cl have.3sg.pst entendues, il dist qu’il understand.ptcp he say.3sg.pst that-he ‘When the letters were read and the duke had understood them, he said that he …’ (OFr. Clari 14, 13th century)

Whilst the trends in the data are relatively subtle, they do confirm and refine the broad generalizations on the left periphery outlined in §3.3. There is a decrease diachronically in the possibility for brand-new subjects to occur in preverbal position, which is expected if new information focus is declining in general, and, furthermore, there is some evidence that the proportion of discourse-active or pronominal subjects increases, whilst the proportion of accessible subjects, which are weakly inferable from the preceding text, decreases.

5.3.2 Postverbal subjects Widespread verb–subject inversion structures are noted in nearly all treatments of Old French syntax and viewed in more recent work as an important V2 correlate (Tobler 1875; Meyer-Lu¨bke 1889; Foulet 1928; Benincà 1983a: 195; Vanelli, Renzi, and Benincà 1986; Adams 1987c; Jensen 1990: 388–390; Roberts 1993: 56; Vance 1995: 177; de Bakker 1997; Vance, Donaldson, and Steiner 2009: 313–316). However, if we consider an instance of a postverbal subject such as (27), in the absence of any other diagnostics, we are unable to distinguish the exact structural position occupied by the subject in the extended TP and vP space:

164

the subject system

(27)

a.

et de cel lac issoient .ix. flum and from that lake come-out.3pl nine rivers ‘And nine rivers come out of that lake’ (OFr. La Queste 175, 192a, 13th century) b. Bels fut li vespres beautiful be.3sg.pst the evening ‘The evening was beautiful …’ (OFr. Roland 157, 12th century)

As was briefly mentioned in §4.3.1.2, it is therefore conventional in the literature to differentiate between Germanic- and Romance-inversion structures, with the former entailing a postverbal subject in a TP-internal position (28)— most likely either Spec-SubjP or Spec-TP—and the latter entailing a postverbal subject in a position within the extended vP complex (29) (Adams 1987c; Vance 1987; Vance 1997: §3.5; Roberts 1993: Chapter 2; de Bakker 1997: 33–39; Salvesen and Bech 2014; Wolfe 2020b). Witness in the following two examples cases where the subject precedes (28) and follows (29) infinitives and past participles which demarcate the vP-edge: (28)

Por ce ne doit hom desperer for this neg must.3sg one despair.inf ‘One should not despair of this …’ (OFr. Eneas 677, 12th century)

(29)

Or est descoverte l’amor now be.3sg discover.ptcp the-love ‘Now [their] love is discovered’ (OFr. Eneas 1527, 12th century)

In recent work, Salvesen and Bech (2014) suggest that the new positions do not perform the same function in terms of the information-structural properties of the subjects. Based on an analysis of 338 postverbal subjects in two early 13thcentury prose texts, they note that subjects such (28) are consistently ‘given’ or ‘inferable’ in their terms, whereas there is greater flexibility in the lower position the subject occupies in (29). They do, however, suggest that subjects in this position are often discourse-new and/or heavy (Salvesen and Bech 2014: 222).1⁰ Note that this state of affairs closely resembles the situation outlined concerning postverbal subjects in Classical Latin and we return to this point in §5.7.

1⁰ Similar characterizations of the Early English subject system are given by Biberauer and Van Kemenade (2011), Old Sardinian by Wolfe (2015d), and Medieval Romance in general by Wolfe (2018a: 147–149).

5.3 the old french subject system

(30)

165

si l’ a veincu uns chevaliers a qui ge si him.cl have.3sg defeat.ptcp a knight to whom I voudroie resembler want.1sg.cond resemble.inf ‘A knight whom I would like to resemble has defeated him’ (OFr. Mort Artu 934, 13th century)

With this background established, we can now consider the pragmatic status of, firstly, all postverbal subjects, and, secondly, the specific characteristics of those in the Germanic- and Romance-inversion positions. One hundred tokens of postverbal subjects from each of the texts examined in Table 5.1 are given in Table 5.3, with the same tagging methodology used as set out for preverbal subjects.

Table 5.3 Postverbal subjects in Old French New Roland Eneas Clari La Queste

36 31 26 19

36.0% 31.0% 26.0% 19.0%

Accessible

Active

Pronominal

Quantified

Relative

37 36 11 17

11 10 23 28

14 9 31 30

2 14 9 6

0 0 0 0

37.0% 36.0% 11.0% 17.0%

11.0% 10.0% 23.0% 28.0%

14.0% 9.0% 31.0% 30.0%

2.0% 14.0% 9.0% 6.0%

0.0% 0.0% 0.0% 0.0%

In a parallel fashion to what we observed for preverbal subjects, we see that those encoding new (31) or weakly accessible information decline in frequency, whereas subjects which are either active in the preceding portion of text, or pronominal (32), increase in frequency: (31)

a. En ceste tere n’ est remés chevaler on this land neg be.3sg remain.ptcp knight ‘Not a single knight remained in the land …’ (OFr. Roland 2798, 12th century) b. lors entra en la sale une mout then enter.3sg.pst in the room a very bele damoisele beautiful girl ‘Then a very beautiful girl entered the room’ (OFr. La Queste 1, 9–10, 13th century)

166

the subject system

(32)

a. Paris les a bien coneu¨es … Par grant Paris them.cl have.3sg well know.ptcp by great angin lo fist Paris cunning it.cl do.3sg.pst Paris ‘Paris knew them well … Paris did it through great cunning’ (OFr. Eneas 123–131, 12th century) b. Puis si manda on tous les barons … then si summon.3sg.pst one all the barons ‘Then one summoned all the barons …’ (OFr. Clari 8, 8, 13th century)

If we accept that at least a subset of postverbal subjects occur at the vP-edge and thus in Belletti’s (2001; 2004; 2005b) low left periphery, this point is important; it underscores that we are not dealing with a simple situation of focus being reassigned ‘downstairs’ from the CP to the vP but rather with a state of affairs where syntactic activation of new information-focus projections in both peripheries appears to be in decline. To investigate the factors behind this further, it is important to look more specifically at Germanic and Romance inversion so that we can separate the two classes of postverbal subjects in the Tand vP-layers.11 Using the same corpus as the one displayed in Table 5.3, Table 5.4 shows the number of instances we find of Romance inversion, Germanic inversion, and cases where the position of the postverbal subject is structurally ambiguous. The first clear finding emerging from Table 5.4 is that the proportion of ambiguous inversion structures is always substantially larger than either those of the unambiguous Romance or Germanic type. This is unsurprising on two Table 5.4 Inversion types in Old French

Roland Eneas Clari La Queste

Romance

Germanic

Ambiguous

11 4 5 7

27 18 6 14

62 78 89 79

11.0% 4.0% 5.0% 7.0%

27.0% 18.0% 6.0% 14.0%

62.0% 78.0% 89.0% 79.0%

11 Given that Germanic inversion—in particular—has been viewed as an important piece of evidence for V2 (§4.2), and that alleged non- or insufficient attestation of inversion structures has been used as evidence against the Old French or Medieval Romance V2 hypothesis (Kaiser 2002: 134; Lombardi and Middleton 2004: 571; Rinke and Meisel 2009: 126; Sitaridou 2011: 164), an understanding of the relative frequency of particular inversion structures is important.

5.3 the old french subject system

167

grounds: first, for Germanic inversion to be apparent we require an overt postverbal subject to occur in a null-subject language. Whilst postverbal subjects are far from unusual (Marchello-Nizia and Prévost 2020: 1056), we should note that they are simply one of several possibilities including a null realization of the subject, merger of the particle si (cf. Fleischman 1991; Wolfe 2018b, and §3.3.4), and fronting of an overt subject to the preverbal topic-focus fields (§5.3.1). In addition to the occurrence of a postverbal subject, we also require the presence of an unambiguous diagnostic of its structural position, specifically the presence of a vP-edge demarcating element such as an infinitive or past participle. In the latter case, recall that the compound past tense was nowhere near as common in the medieval period as it is in the contemporary language (Harris 1978; Rickard 2003: 56; Caudal 2015, and §2.2). As such, there are independent reasons why unambiguous cases are not extremely frequent in texts, as we require the coalescence of two syntactic factors: an overt postverbal subject and the presence of a diagnostic element for the position of that subject. However, moving away from absolute frequency to relative frequency, Figure 5.1 reveals an important finding, namely that Germanic inversion is always more frequent than Romance inversion in the four Old French texts. As discussed in §4.3, inversion structures have been used as evidence both for and against the V2 account (e.g. Roberts 1993 vs. Kaiser 2002), so it is highly significant that of the unambiguous data available, the subset clearly

Romance

Germanic

90.0% 81.8% 71.1%

67.5%

66.7% 54.5%

45.0%

22.5%

45.5% 33.3%

28.9% 18.2%

0.0%

Roland

Eneas

Clari

La Queste

Fig. 5.1 Romance vs. Germanic inversion in Old French

168

the subject system

indicating matrix V-to-C movement is always more robustly attested than the nevertheless V-to-C compatible Romance-inversion counterpart. Turning to the syntax-information structure mapping, recall our null hypothesis, provided by Salvesen and Bech (2014), that Germanic-inversion structures should show subjects which encode old information, whereas Romance-inversion structures can encode either new or old information. Tables 5.5 and 5.6 show the pragmatic and syntactic status of the subjects occurring in each position.12 Table 5.5 The information structure of Germanic inversion

Roland Eneas Clari La Queste

New

Accessible

Active

Pronominal

Quantified

Relative

1 3 0 0

11 16 3 5

7 7 13 7

8 2 9 17

3 2 5 1

0 0 0 0

3.3% 10.0% 0.0% 0.0%

36.7% 53.3% 10.0% 16.7%

23.3% 23.3% 43.3% 23.3%

26.7% 6.7% 30.0% 56.7%

10.0% 6.7% 16.7% 3.3%

0.0% 0.0% 0.0% 0.0%

Table 5.6 The information structure of Romance inversion New Roland Eneas Clari La Queste

14 12 7 7

70.0% 60.0% 35.0% 35.0%

Accessible

Active

Pronominal

Quantified

Relative

5 5 4 6

1 3 4 5

0 0 0 0

0 0 5 2

0 0 0 0

25.0% 25.0% 20.0% 30.0%

5.0% 15.0% 20.0% 25.0%

0.0% 0.0% 0.0% 0.0%

0.0% 0.0% 25.0% 10.0%

0.0% 0.0% 0.0% 0.0%

Whilst the sample is necessarily small, we can tentatively refine some of the existing generalizations in the literature. Clearly, Salvesen and Bech’s (2014) and Wolfe’s (2018a: 72–73, 93, 115) holding that the subject occurring in Germanic-inversion contexts encodes old information is true to some extent: new-information subjects never constitute more than 10% of the Old French sample and active or pronominal subjects as the prime exponents of old information encode 73.3% and 80% of such subjects in the two latest texts, from the period investigated by both Wolfe (2018a) and Salvesen and Bech (2014) in their previous studies. However, we also find indications of a diachronic change in progress, which we will see continues in Middle French in §5.4.2, in that new information subjects are entirely absent in this position in the 13th century and 12 Where necessary, additional examples have been added from the text to ensure that no fewer than 20 tokens were provided for Germanic and Romance inversion.

5.3 the old french subject system

169

weakly accessible subjects decline in frequency. These subtle trends in the data are suggestive that the TP-internal position targeted in such structures is associated with both weakly accessible and old information in the 12th century but unambiguously discourse-old information—as in (33, 34)—in the 13th century.13 (33)

Aprés si fist li dux crier sen ban after si make.3sg.pst the Duke cry.inf his proclamation ‘Afterwards the Duke had his proclamation called out’ (OFr. Clari 8, 7, 13th century)

(34)

ceste costume ai je toz jorz tenue this custom have.1sg I all days keep.ptcp ‘I have always upheld this custom’ (OFr. La Queste 161a, 1, 13th century)

Romance-inversion structures evince the opposite trend, in that they appear to show a broadening of the conditions under which subjects can occupy the vP-internal position. Whilst in the 12th-century texts over half of the subjects are discourse-new (35), this is not the case in the 13th century, where a more even distribution is found across the new, accessible, and active categories. (35)

a. Sur nus est venue male confusïun upon us be.3sg come.ptcp bad disaster ‘A great disaster has befallen us’ (OFr. Roland 2699, 12th century) b. La va fuiant la gent chaitive there go.3sg flee.ptcp the people wretched ‘There the wretched people flee’ (OFr. Eneas 83, 12th century)

Overall, the data are suggestive of a weakening of the pragmatic requirements on subjects in Romance-inversion structures, whereas the trigger for movement to the TP-internal position in Germanic inversion appears to become more specialized.

13 See Roberts (1993: 117–123) for a discussion of postverbal subject pronouns, in which he presents evidence that they may have undergone reanalysis as subject clitics.

170

the subject system

5.3.3 Null subjects in Old French The system Old French null subjects has—since at least Foulet (1919)—been a topic of major interest in the literature.1⁴ We should note from the outset that the null-subject systems found in Early Old French verse texts and Later Old French prose are quite distinct. Firstly, consider the Early Old French system. Early Old French patterns with canonical full null-subject languages in permitting referential (36) and non-referential (37) subjects in matrix and embedded clauses (38) (Adams 1987b; Vance 1987; Dupuis 1988; Dupuis 1989; Roberts 1993: 136–147; Wolfe 2018a: 83). (36)

Getet le a terre throw.3sg.pst it.cl to ground ‘[He] throws it to the ground’

(OFr. Roland 464, 12th century)

(37)

Set anz ad pleins que en Espaigne venimes seven years have.3sg whole that in Spain come.1pl.pst ‘[It’s] been seven whole years since we came to Spain’ (OFr. Roland 197, 12th century)

(38)

Se veïssum Rollant einz qu’il fust mort if see.1pl.impv.sbjv Roland before that-he be.3sg.pst dead ‘If [we] saw Roland before he was dead …’ (OFr. Roland 464, 12th century)

Although the data do not permit an evaluation of whether Early Old French permitted comp-trace violations, other null-subject properties such as cliticclimbing (39) (Pearce 1990; Roberts 1993) and verb–subject inversion are also found (40) (cf. extensive discussion in §5.3.2): (39)

Turnus ne la puet Turnus neg it.cl can.3sg ‘Turnus cannot have it’

pas avoir neg have.inf (OFr. Eneas 3341, 12th century)

(40)

ensi m’avïent comandé li deu del ciel thus me.cl-have.3pl.pst command.ptcp the Gods of-the sky ‘The Gods of the sky have thus commanded me’ (OFr. Eneas 2643–2644, 12th century)

1⁴ Works in this area are numerous, but for the most influential treatments of the data within Government and Binding Theory, see Adams (1987c; 1987b; 1987a; 1988), Vance (1987; 1988; 1993; 1995; 1997), and Roberts (1993). More recent Minimalist accounts are found in Roberts (2007b) and Wolfe (2018a).

5.3 the old french subject system

171

At this stage, verb-initial orders—as we saw in §4.2—are licensed (Labelle and Hirschbu¨hler 2005: 62; Labelle and Hirschbu¨hler 2018: 272; Labelle 2007: 300; Simonenko and Hirschbu¨hler 2012: 30; Zimmermann 2014: 36), as is the case elsewhere in Medieval Romance (Salvi 2012: 106–107; Wolfe 2016a: 470–471): (41)

Vint en la presse, sur les altres come.3sg.pst in the crowd upon the others s’escriet refl.cl-call.3sg.pst ‘He came into the crowd and called upon the others’ (OFr. Roland 961, 12th century)

However, whilst certain null-subject properties remain after approximately 1180–1200, there are growing restrictions on their distribution in non-root environments (Adams 1987c: 3; Roberts 1993: 139; Vance 1997: Chapter 5), as evidenced by clauses such as (42), where we see the embedded subject realized as overt, despite being co-referential with the subject of the matrix clause. (42)

Ensi corurent par mer tant que il vindrent a thus run.3pl.pst by sea such that they come.3pl.pst to Cademelee Cademelee ‘They thus raced across the water until they arrived at Cademelee’ (OFr. Villehardouin 121, 13th century)

At the same time, declarative verb-initial orders decline markedly or disappear entirely in prose texts (Skårup 1975: 291; Rouveret 2004: 193–195; Labelle and Hirschbu¨hler 2005: 66; Simonenko and Hirschbu¨hler 2012; Labelle and Hirschbu¨hler 2018). As such, it is inaccurate to refer to the ‘Old French null-subject system’. Rather, we are dealing with an original full null-subject grammar in Early Old French only, where null subjects are licensed in pre- and postverbal position— yielding V1 clauses in the former case—and matrix and embedded clauses. At the turn of the 13th century, however, we are dealing with an innovative asymmetric system, under which null subjects are licensed in a more restricted set of contexts: in particular, preverbal null subjects are not licensed, leading to the lack of V1 orders, and embedded null subjects are heavily restricted. Specifically, embedded null subjects are only licensed under predicates known to

172

the subject system

admit main-clause phenomena cross-linguistically, with dire ‘say’ being the paradigm case:1⁵ (43)

a. et disoies que ja en ceste maleurté and say.2sg.pst that already in this wretchedness ne charroies fall.2sg.cond neg ‘and you said that you would never fall into this wretchedness’ (OFr. La Queste 161, 34) b. et donna li quens bone seurté que and give.3sg.pst the count good assurance that jamés nel guerreeroit never neg-him.cl wage-war.3sg.cond ‘And the count gave assurances that he would never wage war against him’ (OFr. La Queste 188c, 27–28, 13th century)

Given that Old French at all stages shows the classic properties associated with null-subject systems elsewhere in Romance, the starting point for the analysis is the account of full null-subject languages set out in §5.1.2, namely that pro is licensed by T, and deletes at PF as a result of constituting a defective goal in the terms of Roberts (2010c). The Early Old French data (Dupuis 1988; Dupuis 1989; Hirschbu¨hler 1990), which are strikingly similar to those reported for Modern Romance full null-subject systems like Modern Spanish and Italian, are not a challenge for this account. However, the Later Old French data tell a rather different story: why should a null pronoun licensed by T— assumed to be present in both matrix and embedded clauses—not occur in an embedded environment or preverbally in a matrix clause? No answer to this question is entirely satisfactory and much previous research on the topic makes use of theoretical tools from the era of Government and Binding Theory that have been dispensed with in Minimalism,1⁶ such as the notion of Government (Chomsky 1981; Chomsky 1982). One possibility—put forward in its specific instantiation in Wolfe (2016b)—capitalizes on the long-established notion that V2 systems licensed null topics (Huang 1984; Cardinaletti 2010;

1⁵ As noted in §3.3.5 for embedded V2, lists of the specific predicates in question are found in work on main-clause phenomena such as Rizzi and Roberts (1989), Vikner (1995), Heycock (2006), and Aelbrecht, Haegeman, and Nye (2012). 1⁶ This comment applies particularly to the detailed studies on null-subject-licensing in Vance (1997), Adams (1987a, 1987b, 1987c), and Roberts (1993).

5.3 the old french subject system

173

Walkden 2013b; Salvi 2012; Poletto 2014: Chapter 1; Poletto 2020).1⁷ The core idea is that null topics in Medieval Romance have the same featural makeup as pro in the Modern Romance null-subject languages but with the addition of a topic feature, which requires them to raise to Spec-TopP in the left periphery. Simplifying somewhat, the Early Old French system would therefore feature V-to-Fin movement and a licensing site for proTop in Spec-TopP: (44)

[TopP (proTop ) [FocP [FinP [Fin VFin ][SubjP/TP [vP …]]]]]

The advantage of this system, under which Old French licenses different variants of pro, is first and foremost that it allows us to subsume the licensing of proTop within the broader class of matrix phenomena involving access to the left periphery, which we know are more restricted in Later Old French than Early Old French (cf. Labelle 2007 vs. Wolfe 2018a: 79–86). The growing restrictions on embedded null subjects are, therefore, viewed as restrictions on embedded null topics, which in a parallel fashion to overt embedded topics and foci are licensed with vastly diminishing frequency in Later Old French (cf. Wolfe 2016a and §3.3.5).1⁸ As such, proTop is not licensed as TopP does not typically constitute part of the embedded left periphery: (45)

[FinP [Fin que] [SubjP/TP [Subj/T VFin ][vP …]]]

Likewise, following the change in the locus of V2 from Early Old French to Later Old French, preverbal null topics are also predicted to be ruled out, as Spec-TopP no longer constitutes part of the matrix prefield (cf. for discussion §4.3.2). (46)

[FrameP [ForceP [Force VFin ][TopP (proTop ) [FocP [FinP …]]]]]

The basic intuition is therefore that a change in the locus of verb movement and V2 effects leads to quite marked restrictions on the nature of the null-subject system. Whilst Early Old French patterns with other Romance null-subject systems as a ‘full’ null-subject language, this is not an accurate characterization of Later Old French. Rather, Later Old French belongs to a more restrictive point on the null-subject typology, where clause-type and the pre- or postverbal status of the null pronoun play a role in determining its licensing conditions. As we will discuss further in §5.7, these changes within the system are probably further compounded by a change external to syntax, namely increased levelling in the domain of verbal morphology. 1⁷ For an account under which a topic-head plays an important role in licensing null subjects in Modern Romance null-subject languages, see Frascarelli (2007). 1⁸ Recall from our discussion in §3.6 that we viewed the growing restrictions on embedded topicalization and focalization as evidence that complementizers in Later Old French were typically base-generated in Fin and then raised to Force, restricting access to the embedded left periphery.

174

the subject system

5.3.4 Summary Our exploration of the Old French subject system has revealed—in a similar fashion to that of Latin—multiple subject positions within the clause, which are linked to the information-structural status of the subject. Furthermore, both overt and null subjects are subject to diachronic variation. Whilst this point has been made explicitly for null subjects before (cf. Roberts’ 1993: 137 discussion of the ‘innovative’ Old French null-subject system, which contrasts with the Early Old French system), evidence for diachronic variation in the overt subject system is more novel. As such, it is appropriate to differentiate between the Early Old French and Later Old French subject systems, which show different properties for overt and null subjects. More concretely, the Early Old French system works as follows. As in Latin, preverbal subjects can raise to left-peripheral topic and focus projections; pronominal subjects, and nominal subjects encoding active and accessible information can raise to Spec-TopP, whereas subjects encoding new information raise to Spec-FocP (cf. Table 5.1). Pronominal subjects, as well as those encoding active or accessible information, can occupy the TP-internal Germanic-inversion position, whereas the vP-internal Romance-inversion position is associated with subjects that encode new/focal information.1⁹ I take this as an indication that Belletti’s (2004; 2005a) low topic field is typically not active in Early Old French.2⁰ Null subjects in this period, which entail the licensing of a null pronoun—proTop— can raise to the prefield and act as V2satisfiers, accounting for the distribution of verb-initial orders in Early Old French: (47)

[Frame [Force [Topic (SubjectActive/Accessible/Pronominal /proTop ) [Focus (SubjectNew ) [FinP [Fin VFin ] [SubjP/TP (SubjectActive/Accessible/Pronominal ) [vP Topic [vP Focus (SubjectNew ) [vP [VP (SubjectUnaccusative )]]]]]]]]]]

By way of contrast, in Later Old French we see growing evidence that the preverbal position is increasingly specialized in hosting active or pronominal subjects, and this tendency is also paralleled at the TP-level in Germanic-inversion

1⁹ Unaccusative subjects were included in the information-structural analysis presented in §5.3.2 and I include the possibility in this schema that they can remain in situ—as in Latin—as well as raising to subject positions associated with particular pragmatic values. 2⁰ Note also the comments in §5.2.2 that topical subjects are typically dispreferred in postverbal position in Latin too (Ostafin 1986: 156; Pinkster 1991: 78; Polo 2005: 402; Bauer 2009: 280). Although it is challenging to isolate the specific position occupied by postverbal subjects in Classical and late Latin, this could be taken as an indicator that the non-activation of the low topic field is a point of continuity between the Latin subject system and that of Early Old French.

5.4 the middle french subject system

175

structures. In keeping with our observations in §3.3 that preverbal new information focus is progressively lost in the French of this period, we also see declining evidence for new-information subjects in preverbal position. Looking at postverbal subjects in Romance-inversion contexts, we witness a loosening of the pragmatic requirement for subjects in this position to encode new information, which I interpret as evidence that both the low topic and focus fields are active from the beginning of the 13th century onwards. Turning to the licensing of null subjects, they can no longer act as V2-satisifers, owing to the changing locus of the V2 property; they can, however, still occur in postverbal position as in (48). (48)

[Frame [Force (SubjectActive/Pronominal ) [Force VFin ] [Topic proTop ) [Focus [FinP [SubjP/TP (SubjectActive/Pronominal ) [vP Topic (SubjectActive/Accessible ) [vP Focus (SubjectNew ) [vP [VP (SubjectUnaccusative )]]]]]]]]]]

5.4 The Middle French subject system We have seen so far, in the Latin and Old French subject systems, two grammars which present quite distinct properties from those of Modern French. Whereas in the modern language, subjects are overwhelmingly preverbal—in a dedicated position in the T-layer (cf. §5.6.1)—this is not the case in either Latin or Old French, where we have seen that multiple postverbal and preverbal subject positions exist, which are sensitive to the information-structural status of the subject. In this subsection we explore the extent to which the Middle French data show continuity with the Old French subject system or— as we might expect—show changes which are indicative of a shift towards the modern SVO system.

5.4.1 Preverbal subjects Whilst we might predict a rise in preverbal subjects relative to postverbal subjects in Middle French in keeping with the fact that V2 is lost at the beginning of the 16th century, there is limited evidence to support this assertion with Vance (1995: 185; 1997: 257), Muller (2009: 246), and Wolfe (2020b: 7) suggesting that the preverbal vs. postverbal proportion of subjects remains broadly constant in the 15th century, despite certain qualitative differences which we return to.

176

the subject system

One question that then arises—stemming from our discussion of the information-structural status of Old French preverbal subjects in §5.3.1—is whether Middle French shows broad continuity with the Later Old French data, where we observed a growing tendency for preverbal subjects to be pronominal or encode unambiguously discourse-active information. In Table 5.7, data are presented from the three Middle French texts used in §4.4.1, from which 100 preverbal subjects were extracted.

Table 5.7 Preverbal subjects in Middle French New

Accessible Active

Pronominal Quantified Relative Total

Chronique 11 11.0% 9 9.0% 46 46.0% 31 31.0% Roman de Jean de Paris 8 8.0% 6 6.0% 27 27.0% 57 57.0% Mémoires 8 8.0% 14 14.0% 31 31.0% 39 39.0%

1 1.0% 1 1.0% 8 8.0%

2 2.0% 100 100.0% 1 1.0% 100 100.0% 0 0.0% 100 100.0%

Overall, we observe a continuation of the trend already observable in Later Old French for the preverbal field to host subjects which encode unambiguously old information, in this case with subjects which are already active in the preceding portion of text (49) or are realized as pronominal (50), accounting for between 70.0% and 84.0% of the data: (49)

a. Lequel seigneur de Commarcis n’estoit point the-which lord of Commarcis neg=be.3sg.pst neg en la grace du duc de Bourgongne in the grace of-the Duke of Burgundy ‘This Lord of Commarcis was not in the grace of the Duke of Burgundy’ (MidFr. Chronique 3, 15th century) b. Le roy d’Angleterre souhaicta fort a the King of-England wish.3sg.pst strong to jeune veoir le roy de France see.inf the young King of France ‘The King of England strongly wished to see the young King of France’ (MidFr. Roman de Jean de Paris 21, 15th century) c. Et ledict conte fut en très grant dangier and the-said count be.3sg.pst in very great danger ‘And the aforementioned Count was in very great danger’ (MidFr. Mémoires 13, 15th–16th centuries)

5.4 the middle french subject system

(50)

177

a. Mais ilz n’en avoient nulle volenté de ce but they neg=part.cl have.3pl.pst no wish of it faire do.inf ‘But they had no wish to do that’ (MidFr. Chronique 9, 15th century) b. Il ne sceut que faire He neg know.3sg.pst what do.inf ‘He didn’t know what to do’ (MidFr. Roman de Jean de Paris 28, 15th century) c. Je vous congnois bien I you.cl know.1sg well ‘I know you well’ (MidFr. Mémoires 13, 15th–16th centuries)

By contrast, those subjects which encode information which is discoursenew are rare in preverbal position, albeit still attested, as in (51). Once again, this is a point of continuity with the Later Old French data considered in §5.3.1: (51)

a. Et le connestable et aultres capitaines se and the constable and other governors refl.cl logèrent à l’aultre costé devant le pont stay.3pl.pst at the-other bank in-front-of the bridge ‘And the constable and other governors stayed on the other bank in front of the bridge’ (MidFr. Chronique 4, 15th century) b. Le conte de Lencastre respondit the Count of Lancaster respond.3sg.pst ‘The Count of Lancaster replied …’ (MidFr. Roman de Jean de Paris 19, 15th century) c. et division se mist entre eulx and division refl.cl put.3sg.pst between them ‘and division was sown between them’ (MidFr. Mémoires 10, 15th–16th centuries)

In §5.6.1 we will discuss in some detail the hypothesis—developed since Kayne (1975)—that Modern French subject pronouns should be analysed as forms of subject clitics (cf. §5.1.3). The core debate around the Middle French data has centred around whether subject pronouns should be analysed as maximal categories (Roberts 1993: 166; Vance 1997: 292), or whether they have already undergone reanalysis as heads by this period (Adams 1987a; Adams 1987c: 27–29). Several decades on, the core arguments advanced by

178

the subject system

Roberts (1993) and Vance (1997) against the cliticization hypothesis for preverbal subject pronouns still stand: first, it is far from obvious—as suggested by Adams (1987a; 1987c)—that the qualitatively more widespread V3 orders found in Middle French (cf. §3.4) can readily be accounted for by assuming that the subject in orders such as (52) is a clitic. In fact, we saw in §3.4 that V3 orders of a kind not found in the Later Old French system can include nominal subjects (53a) and non-subject constituents (53b): (52)

De gens d’armes il n’y avoit que of men of-arms there neg-loc.cl have.3sg.pst only ledict Joachin … the-said Joachin ‘In terms of men of arms, there was only the said Joachin …’ (MidFr. Mémoires 7, 15th–16th centuries)

(53)

a. mais Dieu a celle foiz garda Jehan de Paris but God at that time guard.3sg.pst Jehan de Paris et ses gens … and his men ‘But God guarded Jehan de Paris and his men at that time’ (MidFr. Roman de Jean de Paris 41, 15th century) b. L’aultre point si est que the-other point si be.3sg that ‘The other point is that’ (MidFr. Roman de Jean de Paris 24, 15th century)

Moreover, if we assume that preverbal pronominal subjects are clitics in Middle French, we are forced to make the ad hoc and undesirable assumption that they cannot act as V2-satisfiers, despite SubjectPronominal -V-O orders being a fully productive word-order pattern in a language with an active V2 constraint. As we noted in §3.4.6, there are alternative and more attractive ways to model the contrast between Later Old and Middle French in terms of a changing locus of the V2 property. Focussing specifically on the tests for clitichood introduced in §5.1.3, preverbal subject pronouns in Middle French do not consistently show the properties we would expect were they already cliticized. Roberts (1993: 163–164) presents the full range of tests, but we note here that preverbal subject pronouns can be stressed, separated from the verb, and coordinated in Middle French, in contrast to Modern French (54):

5.4 the middle french subject system

(54)

179

a. Et je d’autre part je sui bien de tous and I of-other part I be.1sg well of all les navieurs the sailors ‘And I, on the other hand, am well amongst the sailors’ (MidFr. Froissart, 14th century, Roberts 1993: 164, citing Foulet 1919: 308) b. Je, Pierre de Baufremont, seigneur de Chargni, I Pierre de Baufremont Lord of Chargni de Molyet et de Monfort, chevalier, consillier of Molyet and of Montfort knight advisor et chambellan de très hault, très puissant et and chamberlain of very noble very powerful and très excelent prince, mon très redoubté et souverain very excellent prince my very fear.ptcp and sovereign seigneur, monseigneur le duc de Bourgongne, lord my-lord the Duke of Burgundy fay sçavoir à tous princes … make.1sg know.inf to all princes ‘I, Pierre de Baufremont, Lord of Chargni, Molyet, and Montfort, knight, adviser, and chamberlain of the very noble, powerful, and excellent prince, my feared and sovereign lord, my lord the Duke of Burgundy, inform all princes …’ (MidFr. Chronique 32, 15th century) c. et sarons liquelz est plus fors en and know.1pl.fut the-which be.3sg more strong in ce pays, ou je ou vous this land or I or you ‘and we will know who is the stronger in this land, I or you’ (MidFr. Froissart, 14th century, Roberts 1993:164, citing Price 1971: 145)

Turning towards a conclusion regarding the nature of preverbal subjects in general, we see that all Middle French preverbal subjects can plausibly be analysed as phrasal categories. Although focal and quantified preverbal subjects are still attested, the increasingly specialized nature of the prefield in hosting discourse-active or pronominal subjects suggests that they target Spec-TopP in the majority of cases. The distribution of preverbal subjects therefore reflects our findings on the prefield in general in §3.4.1, where we saw that foci and QPs are rarely found preverbally from this period onwards.

180

the subject system

5.4.2 Postverbal subjects Postverbal subjects are still widely attested in Middle French, but two questions remain unexplored in the literature. One is whether the information-structural status of the subject is the same as in Later Old French, that is to say typically discourse-old in the case of Germanic inversion, but with no particular associated value in Romance inversion. The second, related, question is whether the relative frequencies of Germanic and Romance inversion remain stable when compared to Later Old French. On this point we should note that it has been suggested that Romance inversion increases in frequency in the Middle French period (Muller 2009; Kroch and Santorini 2009), a finding which would not be particularly striking if correct, given that it is the Romance-inversion vPsubject position which is used most extensively in Modern French (cf. Hulk and Pollock 2001 and §5.6). Table 5.8 Inversion types in Middle French

Chronique Roman de Jean de Paris Mémoires

Romance

Germanic

Ambiguous

Total

42 25 33

62 11 44

30 76 43

134 112 120

31.3% 22.3% 27.5%

46.3% 9.8% 36.7%

22.4% 67.9% 35.8%

100.0% 100.0% 100.0%

Table 5.8 gives the overall distribution of inverted subjects in all three textual samples. We see that both Germanic (55) and Romance (56) inversion are attested in all three texts, including the very latest. The fact that Germanic inversion is the most frequently attested kind in two of the texts lends credence to the V2 hypothesis for Middle French (Roberts 1993; Vance 1995; Muller 2009), as noted in §4.4. However, its relatively low attestation in Roman de Jean de Paris relative to Romance inversion and cases where the position of the postverbal subject is ambiguous could be indicative of the gradual breakdown of the system. (55)

a. Item, pour ce que lesdiz Seigneurs se thus for it that the-said lords refl.cl doibvent prouchainement assambler à Nevers, ont must.3pl soon assemble.inf at Nevers have.3pl lesdiz ambassadeurs requis au Roy que … the-said ambassadors ask.ptcp to-the king that ‘Thus, given that the said lords must soon assemble at Nevers, the said ambassadors have asked the King that …’ (MidFr. Chronique 17, 15th century)

5.4 the middle french subject system

181

b. Encores n’avez vous riens ouy still neg-have.2pl you nothing hear.ptcp ‘You’ve still heard nothing’ (MidFr. Roman de Jean de Paris 47, 15th century) (56)

Et, sur ce bruyt et cry, commança à marcher and over this noise and cry begin.3sg.pst to march.inf le conte Charroloys the Count Charroloys ‘And during this noise and clamour, the Count of Charroloys started to march’ (MidFr. Mémoires 11, 15th–16th centuries)

Turning to the information structure of these subjects, we can test the extent to which there is continuity with the Later Old French data, where subjects in Germanic-inversion structures were increasingly discourse-active or pronominal, whereas the Early Old French tendency for Romance-inversion subjects to be discourse-new appeared to have weakened considerably. Tables 5.9 and 5.10 present the relevant data. Table 5.9 The information structure of Germanic inversion in Middle French

Chronique Roman de Jean de Paris Mémoires

New

Accessible Active

Pronominal Quantified Relative Total

8 12.9% 0 0.0%

10 16.1% 1 9.1%

25 40.3% 3 27.3%

19 30.6% 3 27.3%

0 0.0% 4 36.4%

0 0.0% 0 0.0%

62 100.0% 11 100.0%

5 11.4%

5 11.4%

19 43.2%

14 31.8%

1 2.3%

0 0.0%

44 100.0%

Table 5.10 The information structure of Romance inversion in Middle French

Chronique Roman de Jean de Paris Mémoires

New

Accessible Active

Pronominal Quantified Relative Total

8 19.0% 3 12.0%

20 47.6% 8 32.0%

5 11.9% 10 40.0%

0 0.0% 0 0.0%

9 21.4% 4 16.0%

0 0.0% 0 0.0%

42 100.0% 25 100.0%

8 24.2%

14 42.4%

7 21.2%

0 0.0%

3 9.1%

1 3.0%

33 100.0%

Considering first Germanic inversion, the data are so few in Roman de Jean de Paris that it is hard to generalize on the basis of the sample. Looking to Chronique and Mémoires, however, we observe a point of continuity with the Later Old French data, with subjects in Germanic-inversion structures principally active in the preceding portion of text (57) or pronominal (58):

182

the subject system

(57)

a. Si fut ceste conclusion tenue si be.3sg.pst this agreement keep.ptcp ‘This agreement was maintained [by both parties]’ (MidFr. Chronique 7, 15th century) b. Si portoit ledict conte à tous honneur si bring.3sg.pst the-said count to all honour ‘The said Count brought honour to all’ (MidFr. Mémoires 26, 15th–16th centuries)

(58)

Et avec telz gens vueil-je avoir and with such people want.1sg-I have.inf ‘And with such people, I want to have …’ (MidFr. Mémoires 32, 15th–16th centuries)

Conversely, although there is variation between the texts, the broad generalization that the postverbal subject in Romance-inversion structures is not associated with a particular information-structural value is reinforced by the analysis of the three Middle French texts. Consider in this regard the cases of discourse-new and discourse-active subjects in the examples that follow: (59)

a. A laquelle escarmuche fut mort ung très at the-which skirmish be.3sg.pst die.ptcp a very vaillant homme nommé Guillaume du Chastel … brave man named Guillaume du Chastel ‘At this skirmish, a very brave man named Guillaume du Chasel died’ (MidFr. Chronique 8, 15th century) b. Et depuis fu prinse ycelle ville and after be.3sg.pst take.ptcp that town ‘And that town was taken afterwards’ (MidFr. Chronique 27, 15th century)

(60)

a. et ja estoit commencée une grosse et forte and already be.3sg.pst begin.ptcp a great and violent escarmouche au bout du villaige de Montlehery skirmish at-the end of-the village of Montlehery ‘And a great violent skirmish had already begun at the end of the village Montlehery’ (MidFr. Mémoires 11, 15th–16th centuries)

5.4 the middle french subject system

183

b. Et ces parolles m’a compté le roy and these remarks me.cl-have.3sg recount.ptcp the King ‘And the King made these remarks to me’ (MidFr. Mémoires 9, 15th–16th centuries) Overall, we can conclude that in Middle French texts with a productive system of Germanic and Romance inversion, both inversion structures have similar characteristics to those observable for Later Old French: Germanic inversion is reserved principally for subjects which are discourse-active or pronominal, whilst the Romance-inversion structure is non-specialized, suggesting that a wider range of subject expressions can remain within the extended vP-complex. We have also noted how the presence of inversion in general and Germanic inversion, in particular, is one piece of evidence in favour of the V2 hypothesis for Middle French.

5.4.3 Null subjects in Middle French The consensus view is that Middle French is—like Later Old French—a form of partial null-subject language; both Roberts (1993) and Vance (1993; 1997) note that the distribution of null subjects widens slightly when compared to what we observe in Later Old French texts, though certain texts of the period also appear to show increasing restrictions on null-subject distribution. As we noted in §4.4.2, genuine verb-initial orders where the finite verb occurs in absolute initial position of the clause are not a widespread productive word-order pattern in any of our Middle French texts, as they were in Early Old French. The most notable difference concerns clauses beginning with et ‘and’ or ne ‘not’, which were excluded from the main corpus analysis presented in §4.4.2, but nevertheless permit null subjects where their Later Old French counterparts would not, as described in detail by Vance (1993; 1997): (61)

a. Et monstra bien depuis qu’il ne tenoit and show.3sg.pst well after that-he neg have.3sg.pst encores pas la chose … still neg the thing ‘And [he] showed clearly afterwards that he didn’t still have the thing’ (MidFr. Mémoires 12, 15th–16th centuries) b. Ne vous chault neg you.cl care.3sg ‘[It] doesn’t matter to you’ (MidFr. Saintré 299, 25, 15th century, Vance 1997: 297)

184

the subject system

Vance also reports cases of genuine verb-initial orders featuring nonreferential null subjects, which are reproduced in (62). Note that such structures are absent from the three Middle French textual samples used so far in this chapter: (62)

a. Puet bien estre que n’en avez point can.3sg well be.inf that neg-loc.cl have.2pl neg ‘[It] may well be that you have none’ (MidFr. Saintré 8, 4, 15th century, Vance 1997: 297) b. Me semble de prime face que ensuir me.cl seem.3sg of first face that follow.inf voloit les anciennes vesves de jadis want.3sg.pst the former widows of old ‘[It] seems to me at first glance that she wanted to follow the widows of days gone by’ (MidFr. Saintré 3, 10, 15th century, Vance 1997: 297) c. Advint que celle mesme annee le happen.3sg.pst that that same year the voiage de Prusse se tint voyage of Prussia refl.cl hold.3sg.pst ‘[It] happened that that same year the voyage of Prussia took place’ (MidFr. Saintré 187, 10, 15th century, Vance 1997: 298)

Unlike the verb-initial orders that fall out of use in the transition from Early Old French to Later Old French, these do not appear to be good candidates for a null-topic analysis, with several authors suggesting that they are in no way pragmatically marked in Middle French (Marchello-Nizia 1979: 331; Vance 1997: 294; de Bakker 1997: 75). Pending further consideration in §5.7, I therefore suggest that sentences such as (61, 62) are genuine cases of null subjects, and thus feature pro, which constitutes a defective goal in the terms of Roberts (2010c), but can act as a V2-satisfier in Spec-FinP. Indeed, a further piece of evidence that the null-subject system of Middle French is of a different sort to that of Later Old French comes from embedded clauses. In all three texts, embedded null subjects are licensed, as in matrix clauses: (63)

a. ains respondirent qu’ilz n’en feroient but respond.3pl.pst that-they neg-part.cl do.3pl.cond riens, et que point ne les doubtoient nothing and that neg neg them.cl doubt.3pl.pst ‘But they responded that they wouldn’t do anything and that [they] didn’t doubt them’ (MidFr. Chronique 9, 15th century)

5.4 the middle french subject system

185

b. je vous prie que aux heraulx donnez bonne I you.cl ask.1sg that to-the heralds give.2pl good responce reply ‘I ask you that [you] give a good reply to the heralds’ (MidFr. Roman de Jean de Paris 46, 15th century) c. … sembla par ses parolles que ja en seem.3sg.pst by his words that already part.cl ennuyé fust be.3sg.pst bored ‘it seemed from his words that [he] was already bored’ (MidFr. Chronique 9, 15th–16th centuries) Table 5.11 shows the distribution of null and overt subjects in our three Middle French texts in both matrix and embedded clauses.21 Table 5.11 Middle French null vs. overt subjects

Chronique Roman de Jean de Paris Mémoires

Matrix Embedded Matrix Embedded Matrix Embedded

Overt

Null

P-Value

660 116 708 111 726 140

217 7 166 15 123 11

< 0.0001 0.0706 0.0235

What we observe is a highly statistically significant restriction on embedded null subjects in Chronique, a weakly significant result in Mémoires, and a near-significant difference in Roman de Jean de Paris. This further supports our proposal that we are dealing with genuine instances of null subjects—i.e. pro— in Middle French. If a topic head were the main licenser for null subjects, we would predict that null subjects would principally constitute a matrix-clause phenomenon, as they do in Later Old French; this is because access to the topic field is restricted in embedded clauses. However, if pro is typically licensed by a lower head which is present in embedded clauses—such as Fin or T (cf. §5.7)—we do not predict a robust clause-type asymmetry, akin to the situation observable in full null-subject languages (cf. Sheehan 2006; Sheehan 2016). Further evidence for the argument that a null-topic analysis is not correct 21 See also Hirschbu¨hler (1991; 1995) for evidence that person can affect the realization of null subjects in Middle French. I return to this issue in §5.5, as person-effects are particularly pronounced in Classical and Renaissance French.

186

the subject system

for the Middle French data comes from observations in Hirschbu¨hler (1995), Roberts (1993: 173–183), Vance (1997: Chapter 6), and Marchello-Nizia and Prévost (2020: 1071) that certain grammatical persons favour null subjects in Middle French, in particular 2nd-person-plural and 1st-person-plural subjects in embedded clauses. If this is the case, associating the licensing of pro with the absence or presence of person-related features in the inflectional domain is a priori more plausible than the presence or absence of a form of topic.22

5.4.4 Summary The main findings of this section have been that Middle French shows many of the properties also observed in the Later Old French subject-system, but with two major points of difference: the locus of verb movement and V2 effects is distinct on Fin, and this head is also able to license pro. Pulling the analysis above together, I propose that the preverbal subject position—Spec-TopP—is principally reserved for active or pronominal subjects, as is the postverbal position filled in Germanic-inversion structures, Spec-TP. By contrast, both the low topic and focus projections are active, meaning that subjects can remain in-situ within the v–VP in a subset of unaccusative constructions,23 or raise to positions at the vP-edge, where they can encode a range of informationstructural values. I assume here that pro raises to Spec-FinP to satisfy V2, but refine this analysis in §5.7: (64)

[Frame [Force [Topic (SubjectActive/Pronominal ) [Focus [FinP pro [Fin VFin ] [SubjP/TP (SubjectActive/ /Pronominal ) [vP Topic (SubjectActive/Accessible ) [vP Focus (SubjectNew ) [vP [VP (SubjectUnaccusative )]]]]]]]]]]

5.5 The Renaissance and Classical French subject system 5.5.1 Preverbal subjects Our analysis in previous chapters—alongside other work in the literature— suggests that by the 16th century French is no longer a V2 system. Clear 22 Though see Poletto (2020) and also Sigurdsson (2011) for an analysis linking different types of null arguments to logophoric operators in the left periphery. The fact that 1st-person-plural and 2ndperson-plural subjects do not form a natural semantic class makes linking those persons to particular kinds of topics unappealing. 23 If we compare (57a) with (60a) we see that unaccusative subjects can raise out of their position in the extended vP-complex to the T-layer to participate in Germanic-inversion structures.

5.5 the renaissance and classical french subject system

187

evidence for this was presented in Table 3.5, which I reproduce below as Table 5.12, where we see that preverbal placement of the subject is overwhelmingly preferred in matrix clauses, in this case in a sample of linear Verb Second clauses.

Table 5.12 Preverbal constituents in linear-V2 clauses in Renaissance and Classical French

Total SVO Total XP-V-(S) Total O-V-(S) Total Count

Rabelais(1532– Montaigne (1580) 1534)

Phèdre (1677)

Candide (1759)

416 47 1 463

738 3 1 741

802 7 1 809

89.8% 10.2% 0.2% 100.0%

382 36 1 418

91.4% 8.6% 0.2% 100.0%

99.6% 0.4% 0.13% 100.0%

99.1% 0.9% 0.12% 100.0%

Several indicators, which were set out in detail in our discussion of verb movement in Classical and Renaissance French (cf. §4.5), suggest that preverbal subjects of the type seen in (65, 66) do not obligatorily raise to the topic and focus fields—as they would in a V2 system—but instead indicate movement to a position within the T-layer. We introduced in §5.1 the cartographic conception of the subject field as consisting of two distinct projections—SubjP and TP—and thus adopt the null hypothesis that from the 16th century onwards non-dislocated preverbal subjects raise to Spec-SubjP when nominal and Spec-TP when pronominal (Cardinaletti 2004; Cardinaletti 2021): (65)

a. Le moys de mars faillit en Karesme the month of March must-be.3sg.pst in Lent ‘The month of March had to fall in Lent’ (RenFr. Rabelais 9, 1, 16, 16th century) b. Bertrand du Glesquin mourut au siege du Bertrand du Glesquin die.3sg.pst at-the siege of-the chasteau de Rancon pres du Puy en chateau de Rancon near to-the Puy in Auvergne Auvergne ‘Bertrand du Glesquin died at the siege of the chateau of Rancon, near to Puy in the Auvergne’ (RenFr. Montaigne 3, 12, 16th century) c. …[SubjP SubjectNominal [TP [T VFin ][vP …]]]

188

the subject system

(66)

a. Il vous convient doncques noter que … it you.cl suit.3sg therefore note.inf that ‘It is right for you to note therefore …’ (RenFr. Rabelais 1, 7, 16th century) b. Elle esstablit en nous peu a peu it.f establish.3sg in us little by little ‘It (truth) establishes itself in us little by little’ (RenFr. Montaigne 23, 134, 16th century) c. …[SubjP [TP SubjectPronominal [T VFin ][vP …]]]

Whilst this state of affairs is familiar when compared to the subject-system of Modern French (cf. Rowlett 2007: Chapter 4 and §5.6) and—with a limited number of possible exceptions—other Romance languages (Cardinaletti 2004; Sheehan 2006; Sheehan 2010), the importance of the typological difference between Renaissance French and preceding stages of the language’s history cannot be stressed enough: we discussed some evidence in §5.2 that the mere existence of a subject position within the inflectional layer may have been an innovation in Classical Latin (Danckaert 2017b), and whilst the position clearly exists in Old and Middle French, the data presented here suggest it was only used in a restricted set of syntactic and pragmatic contexts. As such, an unmarked position for preverbal subjects within the extended T-layer is a significant syntactic innovation in Renaissance French. We should note here that—as in Modern French (Lambrecht 1994; De Cat 2009; Rowlett 2007; §5.6)—nothing prevents Renaissance and Classical French subjects being merged in the left periphery. The crucial difference when compared to Old and Middle French is that this is not an obligatory operation triggered by an Edge Feature on a C-head but rather a pragmatically marked operation. Given that the focus field is rarely activated in Classical and Renaissance French (cf. §3.4.4), no obvious examples of informationally or contrastively focussed subjects emerge from our corpus of texts. However, leftdislocated subjects are increasingly common in texts of the Renaissance and Classical periods (Fournier 2001: 98; Fournier 2007a: 99; Combettes 2003b): (67)

L’amitié que nous portons à nos femmes, elle the-friendship that we bring.1pl to our women it est très légitime is very legitimate ‘The friendship that we bring to our wives, it is entirely proper’ (RenFr. Montaigne 166, 16th century)

5.5 the renaissance and classical french subject system

(68)

et moi je brûle encore and me I burn.1sg again ‘and me, I burn again’

189

(CFr. Phèdre 75, 17th century)

As such, we see that whilst earlier stages of the language had access to a range of strategies to mark the subject as highly topical, including use of a null subject or topic, merger of particle si in Early Old French (Fleischman 1991), or merger of a postverbal subject in Germanic-inversion contexts, the dominant strategy in use from Middle French onwards is left dislocation (Kroch 1989; Kroch 2001; Kroch and Santorini 2009). Finally, although not exclusively associated with preverbal placement, we consider here the behaviour of subject pronouns, which we turn to in more detail in §5.6. Recall that in our discussion in §5.4 we concluded that they were not yet cliticized in Middle French (Roberts 1993). The facts change somewhat in the Renaissance and Classical period: whilst preverbal subject pronouns can be separated from the verb occasionally in Renaissance French (69), this is no longer the case in Classical French texts from the 17th century, as shown by Fournier (2007a: 38) amongst others. This strongly suggests that the weakening of subject pronouns to eventually yield the system observable in the modern languages is already under way in the 17th century, though we note that—as in Modern French—subject pronouns are not obligatorily in coordination structures in Renaissance or Classical French (70) (cf. Wagner and Pinchon 1962: 173; Fournier 2007a: 22). (69)

Je, dist frère Jean, ne suis point clerc I say.3sg brother Jean neg be.1sg neg clergyman ‘I, said brother Jean, am not a clergyman’ (RenFr. Rabelais 34, 16th century, Fournier 2007a: 38)

(70)

a. Je crains Dieu, cher Abner, et n’ai I fear.1sg God dear Abner and neg-have.1sg point d’autre crainte neg of-other fear ‘I fear God, dear Abner, and have no other fear’ (CFr. Racine, Athalie, 64, 17th century, Fournier 2007a: 22) b. Et croyez, quand il dit qu’il me and believe.2pl when he say.3sg that-he me.cl quitte et vous aime leave.3sg and you.cl love.3sg ‘And believe [him], when he says that he’s leaving me and that he loves you’ (CFr. Molière, Femmes Savantes 115, 17th century, Fournier 2007a: 22)

190

the subject system

Leaving aside certain technical details, the inability of most subject pronouns to bear stress from the 17th century onwards, but the fact that they are not obligatory in coordination contexts, is indicative that they are not genuine syntactic clitics but either phonological clitics (Dufresne 1993; Dufresne and Dupuis 1994; de Bakker 1997: 81), or weak pronouns in the sense of Cardinaletti and Starke (1999). We return to this particular issue in §5.7.

5.5.2 Postverbal subjects From the 16th century onwards, postverbal subjects—previously a widespread characteristic of the language—decline in frequency.2⁴ We should, however, clarify that this does not necessarily fall out from a change in the functional makeup of the subject positions themselves. Rather, as we explored in detail in §4.5, Renaissance and Classical French lack the V2 property and do not feature unmarked V-to-C movement. As such, a V-to-T SVO syntax in the unmarked case, coupled with a concomitant decline in contexts where V-to-C or V-to-Subj movement obtains, will necessarily lead to a marked decrease in postverbal subjects. Contrast in this respect the schema in (71a) with that in (71b), where we see that the emergence of a grammar with V-to-T movement will automatically exclude TP-internal postverbal subjects which are a characteristic feature of the V-to-C grammar of Old French:2⁵ (71)

a. [FinP [Fin VFin ] [SubjP/TP (Subject) [vP Topic (Subject) [vP Focus (Subject) [vP [VP (SubjectUnaccusative )]]]]]] b. [FinP [SubjP/TP (Subject) [T VFin ] [vP Topic (Subject) [vP Focus (Subject) [vP [VP (SubjectUnaccusative )]]]]]]

It is expected, therefore, that Table 5.11 shows postverbal subjects in matrix declaratives to be considerably rarer than in Old and Middle French for both the 16th-century texts and vanishingly rare in the 17th and 18th centuries. Turning to a qualitative analysis, postverbal subjects in the Germanicinversion TP-internal position are near systematically pronominal (72c, d). The exceptions to this generalization are found in Rabelais, where one case is found from 47 postverbal subjects (72a), and in Montaigne, where one case is found from 36. Beyond these exceptions, the pattern across the sample is

2⁴ Amongst others, see Gougenheim (1973: 254–256), Roberts (1993: 200–203), Marchello-Nizia (1999: 50), Prévost (2002), Fournier (2001), and Marchello-Nizia and Prévost (2020: 1056). 2⁵ The schema in (68) indicates V-to-Fin movement as an example of a V-to-C grammar, but the core observation stands regardless of the exact locus of verb movement within the C-domain.

5.5 the renaissance and classical french subject system

191

for nominal subjects to occur within the vP-complex, in instances of Romance inversion (73).2⁶, 2⁷ (72)

a. Lors commença le pauvre Lymosin à dire thus begin.3sg.pst the poor Lymosin to say.inf ‘Thus the poor Lymousin began to say’ (RenFr. Rabelais 6, 74, 16th century) b. Et en est le nombre si infini and part.cl be.3sg the number so infinite ‘And the number of them is so infinite …’ (RenFr. Montaigne 1, 14, 16th century) c. Cela ay-je veu souvvant this have.1sg=I see.ptcp often ‘I have often seen this’ (RenFr. Montaigne 1, 14, 16th century) d. A peine fumes-nous débarquées, que … hardly be.3pl.pst=we disembark.ptcp that ‘We had hardly disembarked when …’ (CFr. Candide 11, 83, 18th century)

(73)

a. Et d’yceulx est perdue la race and from-them be.3sg lose.ptcp the race ‘And from those people [onwards], the race perished’ (RenFr. Rabelais 1, 57–58, 16th century) b. De cete consideration est née la from this consideration be.3sg born.ptcp the coustume … custom ‘The custom … is born out of this consideration’ (RenFr. Montaigne 1, 14, 16th century)

2⁶ These observations are also aligned with those in Combettes (2003a), who discusses inversion in Renaissance and Classical French (in his terms pre-Classical French) travel writers. 2⁷ The number of examples is relatively rare, but cases of inversion are also found in embedded clauses in Renaissance and Classical French: (i)

… en peu de temps, se rendit si pâle et si défait in little of time refl.cl render.3sg.pst so pale and so haggard qu’à peine l’eût-on pu reconnaître that-hardly him.cl-3sg.pst.sbjv-one can.ptcp recognize.inf ‘in a short while he had become so pale and haggard that one would have hardly recognized him’ (CFr. Urfé 118, Fournier 2007a: 36)

192

the subject system

c. Par vous aurait péri le monstre de for you have.3sg.cond perish.ptcp the monster of la Crète the Crete ‘The monster of Crete would have perished for you’ (CFr. Phèdre 2, 4, 649, 17th century) The general conclusion from our analysis is therefore that in Renaissance and Classical French we see the emergence of the modern system of subject positions, albeit with a small number of relic instances of nominal inversion in a TP-internal position, in the 16th century. This tallies with our discussion of interrogatives in §4.5.2.1, where we saw that evidence of V-to-C movement increasingly declines in the 16th and 17th centuries, but that a small number of instances of nominal inversion in interrogatives are attested in Renaissance French (Fournier 2007b:120). As such, we conclude that the broad generalization for the Renaissance and Classical periods is that postverbal subjects narrow in distribution qualitatively and quantitatively owing to emergence of an SVO grammar with V-to-T movement; where postverbal subjects persist, pronominal subjects can occupy a TP-internal position, whereas nominal subjects cannot, with a small number of 16th-century exceptions.

5.5.3 Null subjects Whilst null subjects are not numerous in either Renaissance or Classical French, they are still licensed in a sufficient range of contexts that the language at both stages is typically termed a partial null-subject language. To give an indication of frequency, Table 5.13 gives the frequency of overt vs. null subjects in a 1000-clause sample of four Renaissance and Classical French texts (950 clauses in the case of Phèdre). For comparability with our previous samples, the embedded sample is made up of complement clauses. Table 5.13 Null vs. overt subjects in Renaissance and Classical French

Rabelais (1532–1534) Montaigne (1580) Phèdre (1677) Candide (1759)

Matrix Overt

Null

673 707 868 875

127 21 3 0

84.1% 97.1% 99.7% 100.0%

15.9% 2.9% 0.3% 0.0%

Embedded Overt

Null

188 271 79 125

12 1 0 0

94.0% 99.6% 100.0% 100.0%

6.0% 0.4% 0.0% 0.0%

5.5 the renaissance and classical french subject system

193

Rabelais, as we have seen in previous sections, shows greater continuity with the Middle French texts than Montaigne, and this broad generalization appears to hold here too. Whilst a matrix null-subject rate of 15.9% is directly comparable to the 14.5% rate in Mémoires, our latest Middle French text, the evidence from Montaigne, Phèdre, and Candide suggests that null subjects in this period are far from frequent. The large-scale corpus analysis presented in Marchello-Nizia and Prévost (2020: 1056) shows that this is typical of Classical French texts, a point which should be kept in mind when assessing null subjects’ qualitative distribution in the language of the period. For example, their sample of five 17th-century and five 18th-century texts shows an average overt-subject rate across clause-types of 92.2% in the 17th century and 95% in the 18th. Once again, in big-picture terms, these data reinforce the argument that the early 16th century is a particularly significant turning point in the evolution of French clausal syntax.2⁸ Referential null subjects in matrix clauses are generally restricted to certain grammatical persons in Renaissance and Classical French, being particularly frequent with 1st-person-plural and 2nd-person-plural subjects (74) (Marchello-Nizia and Prévost 2020: 1074). In addition, impersonal null subjects (75) (Gougenheim 1973: 68; Fournier 2007a: 22), alongside null subjects in coordinated or certain embedded environments, are also found in 16th- and 17th-century texts (76), though they are generally absent in the 18th century (Roberts 1993: 215–216; Marchello-Nizia and Prévost 2020: 1076). We note in passing that although Maupas (1607) ‘tolerates’ this last class of embedded or coordinated null subject, he is the last prescriptive commentator to consider them part of the standard language. (74)

… auquel tant sommes obligez to-which such be.1pl oblige.ptcp ‘To which we are so obliged …’ (RenFr. Rabelais 1, 15, 16th century)

(75)

a. Et ne se doit attandre fiance and neg refl.cl must.3sg expect.inf confidence ‘And one must not expect confidence …’ (RenFr. Montaigne 1, 6, 16th century) b. Vray est que true be.3sg that ‘It is true that …’ (RenFr. Du Bellay 10, 16th century, Roberts 1993: 207)

2⁸ With reference to null subjects, note the observation by Marchello-Nizia and Prevost (2020: 1074) that ‘from the middle of the 16th century, non-expressed subjects in non-coordinated clauses are generally very rare, and absent in certain texts’.

194

the subject system

(76)

a. … que a toutes heures me trouveras … that at all hours me.cl find.2sg.fut ‘… that you’ll find me at all hours’ (RenFr. Rabelais 18, 69–71, 16th century) b. aussi est ce chose trop vile, et le also be.3sg it thing too vile and it.cl laisse a ces maraulx sophistes leave.1sg tothese low-life sophists ‘Also this is a vile matter, and [I] leave it to these low-life sophists’ (RenFr. Rabelais 18, 155, 16th century)

We should note furthermore that clauses such as (77)—where a null subject is licensed in what appears to be a residual V2 environment—persist until at least the 17th century, albeit rarely. In the pro-based theory of null subjects adopted here, a case such as (77b) provides evidence that pro is still licensed in certain contexts as the subject le monstre de la Crète ‘the monster of Crete’ is clearly inside the extended vP-complex, as its positioning after the past participle attests, meaning that it cannot be satisfying any form of movement diacritic associated with T or Subj: (77)

a. Et en memoire de ce n’est aujourd’huy and in memory of that neg-be.3sg today passé aulcun en la matricule de ladicte pass.ptcp nobody in the matriculation of the-said université de Poictiers university of Poitiers ‘And in memory of that, nobody today has matriculated in the said University of Poitiers …’ (RenFr. Rabelais 5, 51, 16th century) b. Par vous aurait péri le monstre de for you have.3sg.cond perish.ptcp the monster of la Crète the Crete ‘The monster of Crete would have perished for you’ (CFr. Phèdre 2, 5, 649, 17th century)

As such, our general conclusion on null subjects in the Renaissance and Classical period is that instances of pro are still licensed, as in Middle French (Vance 1997: Chapter 6) and Modern Romance null-subject systems (Roberts 2010c), but that any analysis needs to take account of the partial null-subject nature of the system, on the grounds of both absolute frequency— which is strikingly low from the mid-16th century onwards—and qualitative

5.6 the modern french subject system

195

distribution, which is characterized by increasing syntactic restrictions based on person and clause-type.

5.5.4 Summary Overall, we have seen evidence that the Renaissance French subject system is quite distinct from the Middle French system described in §5.4, and the Classical French system is so to an even greater extent. The major innovation in both cases is the emergence of a dedicated subject position within the T-layer, to which preverbal subjects raise in the unmarked case in the SVO grammar of Renaissance and Classical French, with the option of further movement to topic-related projections in the C-layer, but only as a marked word-order option. In the domain of postverbal subjects, inversion in general declines considerably; when it does obtain, postverbal pronouns occupy a position within the T-layer, which we take to be Spec-TP, and postverbal nominal subjects occupy a position within the extended vP-complex. Whilst pro is still licensed, its occurrence declines sharply in the Classical period, and from the 16th century onwards its distribution is subject to a number of restrictions based on person and clause-type: (78)

[Frame [Force [Topic (SubjectTopic ) [Focus [FinP [SubjP (SubjectNominal ) [TP (SubjectPronominal ), pro [T VFin ] [vP Topic (SubjectActive/Accessible ) [vP Focus (SubjectNew ) [vP [VP (SubjectUnaccusative )]]]]]]]]]]]

5.6 The Modern French subject system 5.6.1 Preverbal subjects and the status of pronouns The canonical order in Modern French is SVO, as in Renaissance and Classical French. Recall from our schema in §4.5—following insights by Cardinaletti (2004)—that we interpret this as meaning that subjects target projections within the T-layer, namely SubjP and TP. This SVO order, with finite verbs and subjects raising out of the vP and into the T-layer, holds in the majority of matrix and embedded clauses. Whilst the position targeted by preverbal lexical subjects is uncontroversial, the same is not true of subject clitics, which we have already seen become progressively weaker from Middle French onwards. The major debate centres around whether French subject clitics are weak pronouns in the sense

196

the subject system

of Cardinaletti and Starke (1999) and thus occupy an argument position— Spec-TP—as a phrasal category (Kayne 1975; Rizzi 1986a; De Cat 2005; De Cat 2009; Roberts 2010a)—or whether they have instead undergone such a degree of weakening that they should be analysed as agreement markers on the finite verb or auxiliary, which occupy a head position (Roberge 1990; Sportiche 1999; Culbertson 2010): (79)

a. … [FinP [SubjP [TP SubjectPronominal [T VFin ] [vP …]]]] b. … [FinP [SubjP [TP [T SubjectClitic + VFin ] [vP …]]]]

Convincing empirical and theoretical arguments have been put forward on both sides of this debate, which has quite radical implications for our overall map of Modern French clausal structure. Our contention is that the status of subject clitics in French is distinct in Modern Standard French and Colloquial French and that—furthermore—their status is subject to diatopic variation (Zribi-Hertz 1994). This is, in and of itself, not a surprising finding, given that aspects of left-peripheral syntax (§3.5), verb-movement properties (§4.5), and negation (§2.3, §7.3.2) show similarly varied patterns, which render the notion of a single ‘French’ system questionable. The Modern Standard French situation is broadly similar to the set of facts we outlined in §5.5 for Classical French: clear evidence for phonological cliticization comes from the finite verb and subject clitic, which cannot generally be separated by intervening material. However, evidence against the agreement-marker analysis comes from the fact that the negator ne can intervene between the subject and the verb (74) (Rizzi 1986a: 399; De Cat 2005: 1200–1203), and the fact that the clitic is not obligatorily repeated in all coordination structures (75) (Kayne 1975; Rizzi 1986a: 403; Poletto 2000: 18); it has been highlighted that this is a major point of distinction between Standard French and Northern Italian dialects where ‘true’ subject agreement-markers are found.2⁹ (80)

Jean n’a rien dit Jean neg-have.3sg nothing say.ptcp ‘Jean hasn’t said anything’ (ModFr., Rizzi 1986a: 397)

(81)

Il chante et (il) dance he sing.3sg and he dance.3sg ‘He sings and he dances’

(ModFr.)

2⁹ See in particular on this point Brandi and Cordin (1989), Cardinaletti and Starke (1999), and Poletto (2000: Chapter 1).

5.6 the modern french subject system

197

A further point of distinction between some Northern Italian dialects and Modern Standard French concerns doubling structures, where a strong pronoun or lexical subject co-occurs with the finite verb. In this area the two hypotheses outlined at the start of this section regarding the status of subject clitics make quite distinct predictions: if the clitics are phrasal categories, co-occurring subject expressions should be optional and left-dislocated, thus occurring in the dedicated topic field in the C-domain that we have seen is active across all varieties of Modern French (§3.5.2). However, if the subject clitics are agreement markers, we predict that they will consistently co-occur with ‘true’ subject DPs occurring within the T-layer (82a vs. 82b). (82)

a. …[Topic SubjectTopic [Focus [FinP [SubjP [TP SubjectPronominal [T VFin ] …]]]]] b. …[Topic [Focus [FinP [SubjP SubjectNominal [TP [T SubjectClitic + VFin ] …]]]]]

In Modern Standard French, the predictions made by the former hypothesis hold quite consistently, with lexical subjects in doubling structures being optional, showing characteristics associated with topicality (De Cat 2005), and doubling categorically ruled out with inherently focal quantified subjects such as (83) (Rizzi 1986a: 396–397; De Cat 2005: 1202–1211; Culbertson 2010: 105–106): (83)

Personne (*il) n’a rien dit nobody he neg-have.3sg nothing say.ptcp ‘Nobody (he) said anything’

(ModFr.)

The conclusion one can draw from this empirical picture is that the conservative Modern Standard French system is one similar to Classical French in that subject clitics are able to occupy an argument position, like full subjects, and are not yet grammaticalized as agreement markers. However, a body of evidence suggests that this generalization does not account for the full range of attested data. In a very detailed study grounded in extensive corpus research, Culbertson (2010) shows two of the most significant arguments against the agreementmarker analysis do not hold for many colloquial varieties of French. The first argument is straightforward: the possibility for ne to intervene between the subject clitic and verb should not be seen as an argument against the subject clitic acting as agreement morphology as many colloquial varieties of French no longer make productive use of the preverbal negator (cf. §2.3,

198

the subject system

§7.3.2; Ashby 1981b; Coveney 1996; Hansen 2013, but De Cat 2005: 1201– 1202 for an alternative view). The second argument is that the coordination facts sketched above do not apply to the grammar of all speakers, with many rejecting omission of the subject clitic in coordination structures (Culbertson 2010: 101–102). This shows clearly in the corpus used by Culbertson, where coordinated subject clitics are overwhelmingly realized rather than omitted (98.4% of cases). Table 5.14 Repetition rates for coordinated subject clitics from Culbertson (2010) 1st/2nd-person 3rd-person Total

126 55 181

99.2% 96.5% 98.4%

127 57 184

Further evidence that a distinct system may be used by certain speakers is provided by doubling data. Subject doubling is extremely frequent in Colloquial French (Zribi-Hertz 1994), and shows diatopic variation, with Nadasdi (1995) noting that Ontario French makes extensive use of subject doubling and Queffélec (2000: 790) suggests that subject doubling is ‘almost systematic in speech’ in Maghreb French. Crucially, for some of the speakers cited in Zribi-Hertz, inherently focal subjects can be doubled, such as the QPs in (84), which strongly indicates that the subject XP is not in a left-peripheral position but rather a position within the TP, which we take to be Spec-SubjP. (84)

a. Tout le monde il est beau, tout le everyone he be.3sg handsome everyone monde il est gentil he be.3sg nice ‘Everyone is handsome, everyone is nice’ (Colloquial French, Zribi-Hertz 1994: 137) b. Personne il fiche rien, à Toulon nobody he do.3sg nothing at Toulon ‘Nobody does anything at Toulon’ (Colloquial French, Zribi-Hertz 1994: 137)

Importantly, Culbertson highlights that certain Northern Italian dialects exist where subject clitics clearly act as agreement markers, but not all subjects can be doubled.3⁰ As such, the more permissive doubling facts for certain 3⁰ Poletto (2000: 140) draws up the following implicational hierarchy for subject doubling in subjectclitic systems (i), where we see that French varieties exist which are at least at point (b):

5.6 the modern french subject system

199

French varieties should be taken as supporting evidence for the agreementmarker hypothesis, but do not rule out the possibility that less permissive systems also feature subject-agreement markers. This leads us to conclude that present-day French varieties have two interacting but semi-separate parts of the grammar which affect the structure of the inflectional layer: the degree of weakening in the subject pronouns themselves, which affects whether they can occupy specifier or head positions, and the permissiveness of the subject-doubling structures in the variety in question. I schematize the Modern Standard French system in (85) and the colloquial and diatopically varied systems in (86a, b): (85)

[Topic SubjectTopic [Focus [FinP [SubjP [TP SubjectPronominal [T VFin ]…]]]]] (Modern Standard French)

(86)

a. …[Topic [Focus [FinP [SubjP SubjectNominal DPs [TP [T SubjectClitic + VFin ]…]]]]] b. …[Topic [Focus [FinP [SubjP SubjectNominal DPs+QPs [TP [T SubjectClitic + VFin ]…]]]]]

5.6.2 Postverbal subjects As a result of the loss of generalized V-to-C movement in the course of the first half of the 16th century (§§4.4–4.5) postverbal subjects are only licit in a restricted set of contexts in Modern French, in contrast to previous stages of the language where we saw that postverbal subjects were widespread. Furthermore, we noted in §4.5.2.2 that verb–subject inversion is near entirely absent in certain varieties of Colloquial French, termed ‘advanced’ French by Zribi-Hertz (1994) and others. Nevertheless, the literature on Modern French postverbal subjects is extensive, so we focus our attention here on notable points of continuity and discontinuity with previous stages of the language’s history. Recall from our discussion in §4.5.2.2 that varieties of French featuring inversion with pronominal subjects license V-to-Subj movement in interrogatives, in both pronominal-inversion and complex-inversion contexts:

(i) If DPs are doubled in a given dialect, tonic pronouns are also doubled. (ii) If QPs are doubled, both DPs and tonic pronouns are doubled. (iii) If variables in wh-contexts such as relatives, interrogatives, and cleft structures are doubled, doubling is always obligatory with all other types of subjects.

200

the subject system

(87)

a. Quand est-il parti? when be.3sg=he leave.ptcp ‘When did he leave?’ (ModFr.) Topic Focus b. … [Focus Quand [FinP [SubjP [Subj est] [TP il [vP [vP [vP [VP ]]] ]]]]]

(88)

a. Quand Jean est-il parti? when Jean be.3sg=he leave.ptcp ‘When did John leave?’ (ModFr.) Topic Focus b. [Focus Quand [FinP [SubjP Jean [Subj est] [TP il [vP [vP [vP [VP ] ]]]]]]]

The significant point of divergence from Old, Middle, and Renaissance French is therefore that Modern French does not permit postverbal nominal subjects within the T-layer of the clause; this property was derived independently in §4.5 by suggesting that even formal varieties of Modern French typically lack V-to-C movement (cf. Cardinaletti 2021).31 Importantly, however, as we also saw was the case in Classical French, postverbal nominal subjects are found in Modern French varieties in so-called Stylistic Inversion (Kayne and Pollock 1978; Kayne 1983b; Rowlett 2007: §5.7.2; Lahousse 2003; Lahousse 2006; Marandin 2011): (89)

a. Quand arrivent les enfants? when arrive.3pl the children ‘When do the children arrive?’ (ModFr.) b. Alors arriva Marie then arrive.3sg.pst Marie ‘Then Marie arrived’ (ModFr.) c. Je veux que soit invitée Marie I want.1sg that be.3sg.sbjv invite.ptcp Marie ‘I want Marie to be invited’ (ModFr., Lahousse 2006: 426)

A number of different analyses have been put forward for Stylistic Inversion which cannot all be reviewed here.32 There are, however, a number of points of commonality across analyses, with none assuming a combination of ‘high’ V-to-Subj or V-to-C movement and movement of the subject XP to a position within the T-layer, despite the structures being widespread in earlier stages of 31 In an earlier framework and without explicit discussion of SubjP, Barbosa (2001) also presents a similar analysis of Romance inversion, contrasting it with inversion structures in V2 Germanic languages where V-to-C movement obtains. 32 See in particular Kayne and Pollock (1978), Deprez (1990), Hulk and Pollock (2001), and Lahousse (2003; 2006).

5.6 the modern french subject system

201

French. Rather—with the notable exception of a remnant movement analysis of Stylistic Inversion offered by Kayne and Pollock (1978)—all assume that the subject occupies a position within the extended vP-complex in such structures; such subjects could either be genuinely vP-internal (Lahousse 2003; Lahousse 2006; Rowlett 2007), or occupy the vP-periphery postulated by Belletti (2001; 2004; 2006; 2008), which we assumed to be active in previous stages of French. (90)

a. … [SubjP [TP [T VFin ] [vP Topic [vP Focus [vP Subject [VP … SubjectUnaccusative ]]]]]] b. … [SubjP [TP [T VFin ] [vP Topic Subject/SubjectUnaccusative [vP Focus Subject/SubjectUnaccusative [vP [VP ]]]]]]

It is tempting, given Modern French’s status as a non-null-subject language, to assume that an overt subject is always merged in Spec-TP. However, we should note an important exception to this generalization, namely postverbal subject clitics. Having rejected the agreement-marker analysis for preverbal subject clitics in formal and standard varieties of Modern French, Roberts (2010a) nevertheless presents a number of compelling arguments that postverbal subject clitics do function as agreement markers in varieties of French which feature pronominal inversion.33 Notably, postverbal subject clitics cannot be omitted in coordinated structures in the same way that their preverbal counterparts can (§5.6.1): (91)

a. Il aime les choux, mais — ne les neg he like.3sg the cabbages but them.cl mange que cuits? eat.3sg but cooked b. *Aime-t-il les choux, mais — ne les mange like=he the cabbages but neg them.cl eat.3sg que cuits? but cooked ‘Does he like cabbage, but only eats it cooked?’ (ModFr. Cardinaletti and Starke 1999: 167)

Furthermore, Roberts notes that postverbal subject clitics show distinct forms such as t-il, t-elle, featuring an epithetic /t/, which is expected if dealing with a particular class of affixes, but not necessarily expected in a pronominal paradigm. Importantly, he suggests that in this specific environment pro is still licensed in the varieties of French which feature pronominal inversion 33 Note that Roberts’ analysis builds on previous insights in Sportiche (1999) and Cardinaletti and Starke (1999) that preverbal subject clitics and postverbal subject clitics do not pattern identically in French.

202

the subject system

(Roberts 2010a: 318). As such we can classify French as a (very) partial null-subject language. Although the behaviour of pro may seem exceptional when taken in the context of the Modern French subject system as a whole, we should note that the notion that more conservative French varieties still license pro in one specific environment is only a natural continuation of the continuing restrictions on its distribution we have observed from the Classical Latin system of null arguments onwards. We return to this particular pathway of change in §5.7.

5.6.3 Summary To summarize, we have seen a degree of continuity between the Classical French subject system and its Modern French counterpart. Whilst left dislocation of subjects to the left-peripheral topic field remains as a marked wordorder option, the vast majority of subjects occur preverbally in the T-layer, specifically in Spec-SubjP when the subjects are lexical or tonic pronouns. In the Modern Standard French subject system, weak pronominal forms occupy Spec-TP, though they are enclitic on the finite verb in Subj when inverted. Roberts (2010a) suggests that a highly restricted variant of pro occupies Spec-TP in such contexts. Inverted nominal subjects in Modern French are, however, exclusively found within the extended vP-complex. Contemporary variation suggests that, for some speakers, weak subject pronouns have been reanalysed as subject-clitic agreement markers; these more ‘advanced’ varieties further differentiate themselves from Modern French by lacking nominal and pronominal inversion. The schema for matrix declaratives appears in (92): (92)

a. [Frame [Force [Topic (SubjectTopic ) [Focus [FinP [SubjP (SubjectNominal ) (SubjectPronominal ) [T VFin ] [vP Topic (Subject) [vP Focus (Subject) [vP [VP (SubjectUnaccusative )]]]]]]]]]]] (ModFr.) b. [Frame [Force [Topic (SubjectTopic ) [Focus [FinP [SubjP (SubjectNominal ) [TP T SubjectClitic + VFin ] [vP Topic [vP Focus [vP [VP ]]]]]]]]]] (Colloquial French)

5.7 Summary and conclusions 5.7.1 The evolution of the French subject system The overall picture sketched in this chapter is one of notable variation across the span of over two millennia. Whilst the broad generalization can be drawn

5.7 summary and conclusions

203

that this period has seen successive restrictions on both the range of null arguments licensed in the grammar, and the range of subject positions which are used, the data evince a considerably more nuanced picture that these broad trends might suggest. In Classical Latin we observe a system whereby overt subjects can remain in the extended vP-complex, which includes a low left periphery of the type postulated by Belletti (2001), but frequently target high left-peripheral projections in the C-domain, for reasons of focalization or topicalization. Recall from our discussion in §3.2 that Latin of all periods makes extensive use of the topic and focus fields within the C-domain for a wide range of constituents, which include subjects. The fact that the Classical Latin finite verb can remain within the vP (§4.2 and Ledgeway 2012a) has clear implications for the licensing of null arguments within the grammar, which—as we have seen—are not restricted to subjects but can include other arguments of the verb, notably null objects, which are highly restricted in French throughout its history (Schøsler 2000b). Luraghi (1997) also suggests that there is some evidence for a diachronic decline in the licensing of null objects in the history of Latin, and the genesis of unmarked V-to-C movement in late Latin (Salvi 2004; Ledgeway 2017b) expectedly leads to an increase in overt postverbal subjects. Danckaert (2017b) also hypothesizes that a subject position within the T-domain only emerges in an innovative Latin grammar, but is not a feature of conservative Classical Latin. In Early Old French texts, we see a degree of continuity with the subject system evidenced in late Latin texts. Whilst left-peripheral preverbal subjects are found across texts, as in Latin of all periods, the increase in postverbal subjects witnessed in late Latin continues. Both preverbal and postverbal subjects can encode focal and topical information at this stage. However, the TP-internal position targeted by postverbal subjects in Germanic-inversion structures specializes in hosting discourse-active and weakly accessible information, whereas the lower vP-periphery, which is activated in Romance-inversion structures, appears specialized in hosting new/focal subjects. Early Old French is also a typical full null-subject language, with null subjects in matrix and embedded clauses. Both the overt and the null-subject system change in the transition to Later Old French. Whilst preverbal subjects can still encode both topic and focus, new/focal information is licensed only to a very limited extent. This is in fact also paralleled postverbally, where the proportion of subjects encoding new/focal information decreases. Furthermore, the TP-internal subject position becomes increasingly specialized in hosting subjects which are either

204

the subject system

pronominal or unambiguously discourse-active. Concomitantly, we observe a generalization in the information-structural properties of the Romanceinversion position. In the domain of null subjects, we observe increasing restrictions on their distribution, which we linked to a Topic-head’s role in licensing proTop : they are increasingly rare in non-root environments and can no longer satisfy V2; this means that they cannot occur preverbally, accounting for the decline in V1 structures in the French of this period (Simonenko and Hirschbu¨hler 2012). In Middle French we observe broad continuity in the overt subject system: both preverbal and postverbal subjects rarely encode new information focus, Germanic inversion remains restricted to unambiguously discourse-old subjects and subject pronouns, and Romance inversion is generalized across information-structural categories. Furthermore, we suggested that there is insufficient evidence to class subject pronouns as clitics in Middle French. Null subjects in Middle French were analysed as genuine instances of pro, rather than proTop ; the number of environments in which pro is licensed remains restricted, showing a matrix/embedded asymmetry and distribution conditioned by clause-type characteristics and grammatical person. However, pro is also found in environments where proTop is not found in Later Old French, most notably in verb-initial clauses introduced by et ‘and’ and ne ‘not’. The emergence of an SVO grammar featuring V-to-T movement changes the subject system quite fundamentally in Renaissance and Classical French. Subjects now move to the T-domain in the unmarked case, with left-peripheral topicalization a marked word-order option, as in Modern French. Postverbal subjects are more widespread in Renaissance and Classical French than in Modern French, but—aside from a small number of residual V2 contexts in Renaissance French—nominal subjects remain in the extended vP-complex. In contrast to Old and Middle French, where Germanic inversion which could apply to both nominal and pronominal subjects in appropriate informationstructural conditions, postverbal subjects in the T-layer are restricted to pronouns in Classical French. Null subjects become extremely rare as the classical period begins and are still subject to person and clause-type restrictions. The Modern French subject system is characterized by sociolinguistic and diatopic variation. In the standard language of France, preverbal subjects typically target the T-layer, pronominal inversion is licensed, and Romance inversion is also licensed in a restricted set of contexts. In the former case of pronominal inversion, Roberts (2010a) argues convincingly that pro is licensed as postverbal subjects are genuine cases of syntactic subject clitics, and thus constitute agreement morphology rather than ‘true’ subject expressions.

5.7 summary and conclusions

205

Several regional varieties of French and non-standard varieties spoken in France show distinct properties, however; these include the general absence of inversion and the reanalysis of subject pronouns as agreement markers. The result is that subject-doubling structures have a distinct structure from superficially similar constructions licensed in the standard language. The findings regarding each period surveyed in this chapter are summarized in Table 5.15.

5.7.2 Change in subject positions Whilst Table 5.15 shows that the evolution of French subject positions has seen a number of distinct stages, there is an important binary split between Latin and French prior to the 16th century on the one side, and all stages of French that follow on the other: Renaissance, Classical, and Modern French show classic EPP-effects on T, where a subject must be merged within the Tlayer. In contrast to this, all stages of Latin, Old, and Middle French do not; this means that nothing a priori prevents subjects staying within the vP and that the principal factor determining their distribution is information structure. The question arises as to whether any of the purely syntactic constraints seen in Modern French emerge as the result of reanalysis of processes previously conditioned by information structure. Our proposal here is that one major factor affecting the shape of the Modern French subject system is the change that affected the information structure of subjects in Germanic-inversion structures in the passage from Early Old to Later Old French. In this regard, we observe a change in progress throughout the Old French period where the relevant position within the T-layer becomes increasingly specialized in hosting subjects which are unambiguously discourse-active or pronominal. What might have triggered this change? I suggest that it is triggered by the need for the grammar to have a clear topicmarking strategy. The grammar of Early Old French has several ways to encode this: a verb-initial clause featuring the null topic—proTop— can be employed, as can the particle si, which we noted in §3.3.4 is associated with topic-continuity in Early Old French (Fleischman 1991). Crucially, neither of these strategies are licit in Later Old French, with verb-initial orders no longer licensed and si having grammaticalized ‘upwards’ as a Spec-ForceP V2-related expletive (Wolfe 2018b). The suggestion is therefore that the T-domain subject position is ‘repurposed’ as a position which can only host highly topical subjects, either those which are already active in the discourse or pronominal. An important consequence of this change is that acquirers only need to specialize the trigger for movement one further step to derive the Modern

206

Classical Latin Late Latin Early Old French Later Old French Middle French Renaissance French Classical French Modern French Colloquial French

Unmarked VMovement

TopP

FocP

TP/Germanic-Inversion

vP/Romance Inversion

Null

V/v Fin Fin Force Fin T T T T

Yes Yes Yes Yes Yes Yes Yes Yes Yes

Yes Yes Yes Rare Rare – – – –

Yes Accessible/Active/Pronominal Accessible/Active/Pronominal Active/Pronominal Active/Pronominal Pronominal Pronominal Pronominal –

Focal Focal Focal Unmarked Unmarked Unmarked Unmarked Unmarked –

Generalized Null Arguments Generalized Null Arguments Full Null Subject Partial Null Subject Partial Null Subject Partial Null Subject (Very) Partial Null Subject (Very) Partial Null Subject –

the subject system

Table 5.15 Subjects and subject positions in the history of French

5.7 summary and conclusions

207

French system of subject inversion, where pronouns alone can occupy this position.3⁴,3⁵ (93)

Accessible/Active/Pronominal Early Old French Pronominal Modern French

> Active/Pronominal > Later Old and Middle French

One can straightforwardly model this change in terms of changing features of the probing head. However, there is an alternative possibility. Until now, we have adopted for the sake of simplicity the maximalist cartographic assumption that the T-layer is always split along the lines outlined by Cardinaletti (1997; 2004). Suppose, however, with Giorgi and Pianesi (1996; 1997) and others, that the Subj and T heads can be syncretic in certain systems. One possibility is that during the breakdown of V2, acquirers no longer analyse the data as consistent with V-to-Fin movement and a syncretic Subj-T-head with a specialized probe for discourse-active subjects [uAct], but rather a difference in the height of verb movement. As such, in the new grammar, the fact that only pronominals can occur in T-layer-internal inversion structures is a result of the height of verb movement, not of the probing features of the subject-related head. The overall result is a grammar lacking V-to-C movement in inversion contexts. Romance inversion and the makeup of the extended vP-complex also follow a similar trajectory in that the ground is laid relatively early in French’s history for what becomes the Modern French system of postverbal subjects. Whilst we saw that subjects belonging to a range of categories could occur postverbally in Latin, it was suggested that postverbal subjects are predominantly focal (Ostafin 1986: 156; Pinkster 1991: 78; Polo 2005: 402; Bauer 2009: 280). We saw that this continues into Early Old French, where Romanceinversion structures typically feature subjects which are new/focal. However, by the beginning of the Later Old French period this is not the case and subjects occurring within the extended vP-complex can belong to a wide range of information-structural categories. The reasons for this may seem puzzling, but 3⁴, In addition to our discussion in §§4.3–4.6, this is a further instance of the specific triggers for inversion becoming specialized diachronically. See also on English inversion Biberauer and Roberts (2012). 3⁵ It is significant that pronouns play an important part in this reanalysis. Whilst ±Pronominal is clearly a purely syntactic difference—linked to the amount of functional structure present in the nominal expression (Cardinaletti and Starke 1999; Barbosa 2019)—pronouns are also significant in information-structure terms, as they are necessarily discourse-linked and typically highly topical (Ariel 1988). As such they constitute an expected ‘bridge’ between a probe associated only with information structure and one which is not.

208

the subject system

a recent analysis of Old Italian by Poletto (2006a; 2006b; 2014) may offer a possible explanation. Drawing on Chomsky’s (2001; 2008) work on phases, Poletto suggests that the syntactic makeup of phase-edges is uniform across the CP, vP, and DP, which she links to Old Italian’s licensing of widespread ‘operator movement’ at all three of these phase-edges. However, we may be able to take this hypothesis a step further: if the makeup of phase-edges is uniform and the information-focus projection in the CP is becoming progressively less active in Old French—as we saw was the case in §3.3.2—we would also predict that the same holds at the vP-edge. Our analysis of subjects in §5.3 shows this is precisely what occurs: new/focal subjects at the CP-and vP-edges decline in a uniform fashion. The proposal is therefore that weakening of the requirement for vP-peripheral subjects to be focal is triggered by a change outside the subject system, namely the loss of information focus at the CP-level. This yields the Renaissance, Classical, and Modern French system where postverbal subjects are not obligatorily focussed. Finally, we should note one important reanalysis which has taken place relatively recently in French’s history and only in certain varieties of French: the reanalysis of subject clitics as agreement markers (Culbertson 2010), which is a classic case of Spec-to-Head reanalysis in the terms of Van Gelderen (2011). Whilst this does change the structure of the subject layer for certain varieties of French, it is far from clear that the typological significance of the change is as much as suggested by Harris (1978), who proposes influentially that French is becoming a form of VSO language as a result of this reanalysis (cf. §2.3). There are two main reasons to object to this claim. Firstly, we have seen above that varieties where the subject clitic > agreement marker reanalysis have taken place also feature widespread subject doubling (Zribi-Hertz 1994; Nadasdi 1995), which does not feature a left-dislocated subject but rather a ‘genuine’ subject expression in Spec-SubjP. As such, it is not the case that such varieties lack a dedicated preverbal subject position and are not SVO languages. Furthermore, there is no clear evidence that subjects involved in right-dislocation structures are undergoing reanalysis as unmarked or becoming more widespread, as Harris suggests. Rowlett (2007: 181), for example, notes that right dislocations are ‘compatible with known topics, only’. To conclude, although the ongoing grammaticalization of subject clitics is a significant process at play in varieties of French today, its typological significance can be overstated.

5.7.3 French and the null-argument typology Throughout its history French has licensed some form of null arguments, albeit to a highly restricted extent in Modern French and—conversely—in a very

5.7 summary and conclusions

209

widespread fashion in Classical Latin. In a recent approach to modelling null arguments within parametric theory, Biberauer (2018) highlights the importance on the properties of both the licensing head and the licensee—i.e. the null pronoun—involved in null-argument structures cross-linguistically. This approach—also central to that of Roberts (2010b) adopted here, where the features of both probe and goal are relevant in determining deletion—bears fruit when considering the history of French. Considering the properties of the probing head, we see that Latin has generalized null arguments as a result of the low default locus of the finite verb in V or v. By contrast, Early and Later Old French have a distinct form of system, where a form of null subject/topic— proTop— must be licensed by a left-peripheral topic head, as appears to be the case in other Medieval Romance (Benincà 2004; Benincà 2006; Wolfe 2016b; Poletto 2014; Poletto 2020) and Early Germanic (Walkden 2013b) varieties.3⁶ In keeping with the downwards reanalysis of verb movement in the transition from Old to Middle French, the licensing head for null subjects also changes and I take this to be the locus of ϕ-agreement at each stage, namely Fin in V2 Middle French and T in Renaissance, Classical, and Modern French. Crucially, the properties of the goal also change in the history of French, with Later Old French and Middle French licensing a form of pro also endowed with an unvalued topic-feature: proTop . The overall generalization from these data is of a heterogeneous picture in French’s history, but one characterized globally by increasing restrictions on the licensing of null arguments. Precisely how these restrictions emerged has generated a substantial literature, but I suggest here that two critical factors are at play, neither of which exclusively accounts for the decline, namely ongoing levelling of inflectional morphology (cf. §2.1, references therein, and BazinTacchella 2020b) and the emergence of strict V2 with V-to-Force movement (Wolfe 2016a). Our assumption here is that in the transition from the late Latin C-VSO grammar to the Early Old French V2 grammar, acquirers needed to find a way to integrate extensive verb-initial structures into a novel full V2 grammar, where a trigger for V-to-Fin movement and an EF are present on Fin. As such, they reanalysed structures previously involving verb-fronting alone as V-to-Top movement accompanied by a probe-goal Agree relation between Top and a null pronominal, proTop . Under the cartographic and phase-theoretic 3⁶ Note the parallel here between the idea that the distribution of a null subject—pro—is intrinsically affected by information structure in the same way that overt subjects are throughout the Old French period. This observation extends to proTop ’s distribution in the left periphery, where it needs to merge in a projection within the topic-focus fields, under the standard cartographic assumption that moved constituents cannot remain in Spec-FinP (Rizzi 2010).

210

the subject system

assumption that the specifier of Fin acts as an escape hatch through which constituents move, but in which they do not typically remain (Haegeman 1996; Rizzi 2006a; Rizzi 2010; Cardinaletti 2010; Shlonsky 2014), the postulation of an initial topic feature is necessary in order to motivate pro’s movement higher into the topic-focus fields. Following Frascarelli’s (2007) analysis of Modern Italian and Poletto’s (2014) analysis of Old Italian, Wolfe (2018a) suggests that proTop is a null aboutness shift topic in the terms of Frascarelli and Hinterho¨lzl (2007). Novel empirical support for this finding comes from the quantitative analysis of Old French null subjects in Marchello-Nizia and Prévost (2020:1067–1068), where they find that third-person null subjects are considerably more frequent in Old French than their first- and second-person counterparts; this is precisely what would be predicted under an account where proTop serves as a null aboutness shift topic, which Frascarelli (2007) and Frascarelli and Hinterho¨lzl (2007)—among others—claim are particularly frequent with third-person referents. Aside from these formal motivations, typological research supports this proposal: as has been noted elsewhere in this book, whilst full null-subject systems appear not to be found in V2 languages, partial null-subject systems featuring null topics are attested extensively (Huang 1984; Axel 2005; Sigurdsson 2011; Walkden 2013b; Poletto 2020). Crucially, the upwards reanalysis of the locus of V2 to Force leads to quite substantial changes in the system. Once it reaches Spec-TopP, as a featurally inert XP with no features in need of valuation, proTop cannot raise to SpecForceP to satisfy V2; this means that preverbal null subjects are not found but that postverbal occurrences do remain. Furthermore, given the heavy restrictions on the availability of the embedded topic layer in Later Old French, embedded null subjects are only found in a very small class of matrix-like embedded clauses (cf. §5.3). The key observation is that the number of environments in which null subjects are licensed is thus massively restricted, which fundamentally destabilizes the null-subject system. The downwards reanalysis of V2 to Fin in Middle French thus takes place in an environment where acquirers would have been exposed to a diminishing range of null-subject data. My proposal is that from this point onwards, acquirers are in fact sensitive to the perceived richness and specific featural makeup of Fin or T as the licensing heads for pro, in contrast to Early and Later Old French where a finite verb in Top can always act as an appropriate probe. Although this approach—where a link exists between the licensing of null subjects and inflectional morphology of the verb—has been widely criticized (see in particular Adams 1987c), Simonenko, Crabbé, and Prévost (2019)

5.7 summary and conclusions

211

argue convincingly on the basis of a large, digitized corpus that there is a link between the richness of verbal morphology as a whole and the proportion of null subjects licensed in a given text. Furthermore, the fact that in Middle and Renaissance French null subjects are most commonly found in 1st- and 2ndperson-plural environments is expected under this approach: it is in these cells of the paradigm that distinctive verbal morphology is retained in the period in question, leading acquirers to postulate a [D] feature on Fin or T in the relevant context and deriving the partial null-subject status of Middle and Renaissance French. We should note that the observation—highlighted by Marchello-Nizia and Prévost (2020: 1077)—that first- and second-person subjects are often overtly realized in Old French is not an argument against the approach here: if proTop constitutes a null aboutness shift topic, most frequently licensed in third-person contexts (Frascarelli 2007), we predict first- and second-person null subjects/topics to be more restricted. By way of contrast, in Middle and Renaissance French the licensing head and the featural makeup of the null pronoun are distinct and we are dealing with ‘classic’ pro, where we expect the richness of verbal morphology to constrain its distribution, in line with the approach adopted in Roberts (2010b). A formal analysis is not offered here of why, in Renaissance and Classical French, et ‘and’ and ne ‘not’ exceptionally license pro, nor why specific clause-types [i.e. +Wh, +Subjunctive] permit its exceptional licensing (Roberts 1993). I note, however, that when null subjects make up as little as >1% of the available textual data, it is not surprising that acquirers postulate increasingly idiosyncratic microparametric properties of the licensing head to integrate an ever-diminishing class of null subjects into their grammar. The resulting

Table 5.16 Null arguments from Latin to French

Classical Latin Late Latin Early Old French Later Old French Middle French Renaissance French Classical French Modern French Colloquial French

Category

Licensing Head

Generalized Null Arguments Generalized Null Arguments Full Null Subject/Topic Partial Null Subject/Topic Partial Null Subject Partial Null Subject (Very) Partial Null Subject (Very) Partial Null Subject –

v Fin Top Top Fin Fin/T Fin/T Subj –

212

the subject system

microparametric variation would thus be parallel to Romance systems today where auxiliation with esse or past-participle agreement obtains only in a highly restricted class of environments, which can appear arbitrary.3⁷ Our discussion of the null-argument properties of Latin and French—above and in this section—is schematized in Table 5.16.3⁸

3⁷ For recent discussion of auxiliation, see Ledgeway (2019), and Loporcaro (1998; 2016) on pastparticiple agreement. 3⁸ I leave discussion of null objects in Modern French to future work (cf. Roberge and Cummins 2006 for recent discussion). These are arguably a distinct phenomenon from the widespread null arguments licensed in Latin.

6 OV orders and the middlefield 6.1 Introduction: head-directionality and OV orders in Romance and beyond 6.1.1 Background The ordering of heads relative to their complements has been a major topic of enquiry in functional-typological studies of Romance and other languages, and how to model the variation observable in this area has been a central area of interest within formal Principles and Parameters Theory. Latin, French, and the Romance languages offer an ideal empirical domain in which to investigate such variation as the textual records evidence the progressive loss of head-final ordering over a timespan of millennia. In this chapter, we focus in particular on OV orders, which—although widespread in Latin (1)—survive in some form throughout French’s history (2): (1) Terentia magnos articulorum dolores habet Terentia.nom heavy.acc joints.gen pains.acc have. 3sg ‘Terentia has a bad attack of rheumatism’ (Latin, Cic. Att, 1.5.8, Pinkster 1991: 75) (2) a. Qant li fos la parole antant … when the fool the speech hear.3sg ‘When the fool hears the speech’ (OFr. Perceval 2862, 12th century, Zaring 2010: 5) b. Et en ce faisant, il osta beaucoup de and in this do.prog he remove.3sg.pst many of gens de suspicion que people from suspicion that ‘And in doing this, he removed many people from suspicion that …’ (MidFr. Mémoires 18, 15th–16th centuries)

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0006

214

ov orders and the middlefield

c. Il n’a rien fait he neg–have.3sg nothing do.ptcp ‘He has done nothing’

(ModFr.)

This chapter will focus on the evolution of these OV orders, from their original status in the grammar of Classical Latin, which shows widespread head-final orders in other domains, to their increasingly exceptional status within the grammar of French from the medieval period onwards.

6.1.2 Head-finality and the Final-Over-Final Condition Within classic Principles and Parameters Theory (Chomsky 1986; Chomsky 1993), head-finality was typically modelled as contingent upon the setting of a binary head-directionality parameter, which constrained the ordering of heads and complements across categories:

(3)

XP Spec

XP X'

X Head-Initial

X' YP

YP

Spec

X Head-Final

Aside from theoretical developments within Minimalism which render this approach problematic,1 the setting of a single macroparameter also makes insufficiently nuanced empirical predictions, accounting well for so-called ‘harmonic’ head-initial or head-final languages like English or Japanese where heads consistently precede or follow their complements, but failing to account for the wide range of languages—such as Latin, Old French, and Modern German—where head-directionality is dependent on the syntactic category in question. As such, one method of modelling variation in this domain since Kayne (1994) has been to assume that all languages are underlying head-initial, with head-final orders derived via some form of movement. In a simple case, OV orders of the type we are concerned with here can be derived via movement of the complement to a position structurally higher than the head, such 1 See Biberauer (2010b) for extensive discussion, but note for our purposes the significant effect of eliminating the presence of a dedicated phrase-structure module from narrow syntax (Chomsky 1995).

6.1 introduction

215

as Spec-VP, the vP-periphery, or a position within the T-domain, to satisfy a movement-triggering diacritic. Whilst an approach to head-directionality under which the properties of each category depend on the presence or absence of movement-triggering features on particular heads may appear to permit any kind of system, the Final-Over-Final Condition is one of a number of factors showing clearly that this is not the case (Holmberg 2000a; Biberauer, Holmberg, and Roberts 2014; Sheehan et al. 2017). Informally stated, the condition can be defined as follows: (4) If a phrase α is head-initial, then the phrase β immediately dominating α is head-initial. If α is head-final, β can be head-final or head-initial. A number of case studies (e.g. Biberauer, Sheehan and Newton 2009; Biberauer, Sheehan and Newton 2010) have suggested that this condition should constrain both the shape of synchronic variation and diachronic change. As we shall see, Latin is an ideal case study to show the explanatory value of the condition (cf. in particular Ledgeway 2012a).

6.1.3 The vP-periphery and the Uniformity of Phases As has already been noted throughout the book—and utilized in particular to account for the distribution of subjects in Latin and the history of French— there is a large body of evidence that projections associated with topic and focus are not exclusively found in the C-domain cross-linguistically but also at the vP-edge (Belletti 2001; Belletti 2004; Belletti 2006; Jayaseelan 2001; Poletto 2006a; Poletto 2006b). As such, following work by Belletti (2001) in particular, our assumption throughout this chapter will be that vP-internal topic and focus projections are capable of hosting fronted objects and do so at various stages in the history of Latin and French. This is essentially the definition of scrambling given in Poletto (2014: Chapter 1), who suggests that obligatory ‘operator movement’ to vP-peripheral projections is a direct correlate of V2 in Old Italian, as a result of a requirement that the featural makeup of phase-edges be identical within a grammar. As such, an OV order derived via scrambling such as that found in (5a) is the ‘low’ counterpart to the V2 clause seen in (5b) in that both are the result of a movement diacritic on v and Fin respectively: (5) a. … quali denari avea Baldovino lasciati loro which coins have.3sg.pst Baldovino leave.ptcp them ‘… which coins Baldovino had left them’ (Old Italian, Eredità Baldovino 437, Poletto 2014: 4)

216

ov orders and the middlefield

b. i nemici avessero già il passo pigliato the enemies have.3pl.pst.sbjv already the pass take.ptcp ‘The enemies had already occupied the pass’ (Old Italian, Pagani 88, Poletto 2014: 40) Given that uniformity across phase-edges was shown to hold in the distribution of subjects in Chapter 5, it should not come as a surprise that we will see that this kind of uniformity also makes important predictions about the range of data licit in OV cases in the history of French.

6.2 OV orders in Latin 6.2.1 Classical Latin as an OV language As we have noted in other discussions of clausal structure, scholars who do not maintain the position that Latin word order was absolutely free (Meillet 1903; Ku¨hner and Stegmann 1955; Ramsden 1963: 42) typically converge upon the view that the unmarked word order of archaic and Classical Latin was SOV, with objects typically preceding finite verbs (6). (6) a. Mancinus domum revenisset Mancinus.nom house.acc return.3sg.pluperf.sbjv ‘Mancinus had returned home’ (Latin, De Orat. 1.181, Devine and Stephens 2006: 11) b. Eo tempore Iugurtha per collis sequi at-that time.abl Jugurtha.nom along hills.acc follow.prs.act.inf ‘At that time, Jugurtha would follow along the hills …’ (Latin, Sall. Jug. 55.8, Spevak 2010: 70) This finding is not unexpected when contextualized against findings for other early Indo-European languages, where OV orders are overwhelmingly more frequent than VO orders; this finding has led a number of scholars to hypothesize that Proto-Indo-European was a head-final language, where objects typically precede verbs (Lehmann 1974; Roberts 2007a: 364–365; Hock 2015; Ringe 2017: 64). From the earliest Latin texts onwards—however—the language cannot be described as harmonically head-final. In keeping with the predictions of the Final-Over-Final Condition (cf. 4) prepositions are widely found in Latin texts (7), as are complementizers in clause-initial position (8) (Vincent 1999; Ledgeway 2012a: 193; Adams 2013: pts 3, 6).

6.2 ov orders in latin

(7) in agrum into field.acc ‘into the field’

217

(Latin)

(8) legati Carteienses renuntiauerunt quod Pompeium in deputies.nom Carteian.nom report.3pl.pst that Pompey.acc in potestate haberent power.abl had.3pl.pst.sbjv ‘deputies arrived from Carteia reported that they had secured Pompey’ (Latin, B.Hisp. 36.1, Ledgeway 2012a: 105) Turning to head-directionality within the extended vP-complex and the behaviour of objects specifically, whilst quantitative data are often not directly comparable and vary considerably in terms of textual basis and sample size, a large number of studies have also demonstrated that objects precede verbs more frequently than they follow them in the majority of classical texts (Linde 1923: 154–158; Bauer 1995: 89–92; Clackson and Horrocks 2007: 106; Cabrillana 1993: 256–260; Cabrillana 2011: 70–71; Polo 2005: 378–380; Halla-Aho 2009: 129; Ledgeway 2012a: §5.2; Danckaert 2012: 15); OV ordering is, for example, extremely consistent or even absolute in some of the very earliest Latin texts (Clackson and Horrocks 2007: 106), such as the Twelve Tables (Bauer 1995: 89).2 Ledgeway (2012a: 228)—for example—offers the following synthesis of existing studies on verb and object placement, which supports the line of analysis that OV should be viewed as unmarked, albeit subject to extensive intertextual and diachronic variation. As already noted elsewhere in this book, OV is also more frequent in embedded clauses (9) than matrix clauses (Linde 1923; Adams 1977: 69; Adams 2016: 321; Panhuis 1982: 117; Ostafin 1986: 165–169; Wanner 1987: 380; Polo 2005: 378; Ledgeway 2012a: 185; Danckaert 2017c: 113), with Bauer (1995: 92) noting specifically that ‘in Classical as well as Late Latin, the tendency to use OV is most manifest in relative clauses’, a pattern which we will also see holds of various stages of French (cf. §6.3). (9) … quae ad ancoras erant deligatae which to anchors.acc be.3pl.pst fasten.ptcp ‘… which were fastened to anchors’ (Latin, Caes. B.G. 4.29, Devine and Stephens 2006: 185) 2 However, see also Adams (1976: 92) and Pinkster (1991), who both highlight that the apparently consistent OV ordering found in certain texts such as the Twelve Tables may be due to their formulaic nature.

218

ov orders and the middlefield

Table 6.1 OV vs. VO in Latin from Ledgeway (2012a: 228) Text/Author

OV/VO

S.C. Bacch. (Álvarez Pedrosa 1988) Leges 2nd-c. bc. (Álvarez Pedrosa 1988) Pl. Capt. Adams (1976a: 94–95) Pl. Amph. 1–400 (Adams 1976a: 95) Pl. Aul. 1–325 (Adams 1976a: 95) Pl. Asin. 1–380 (Adams 1976a: 95) Pl. Mil. 1–500 (Adams 1976a: 95) Ter. (Moreno Hernández 1989) Cic. Cat. (Koll 1965: 246–247) Cic. Leg. (Koll 1965: 246–247) Cic. Att. 1 (Pinkster 1991: 72) Cic. Att. 1 (Cabrillana 1993a) Cic. S. Rosc. (§§1–34) Adams (1976a: 94) Cic. Deiot. (§§1–34) Adams (1976a: 94) Cic. philosophical writings (Bolkestein 1989: 14–15) Cic. Mil. (Panchón 1986) Caes. B.G. I 1–15 (Adams 1976a: 94) Caes. B.G. 1–7 and B.C. (Pinkster 1991: 72) Caes. B.G. 1 (Panchón 1986) Vitruvius 1.1–4 (Pinkster 1991: 72) Ov. Met. (Amacker 1989) V+ 2 elements Petr. Sat. 26–68 (Polo 2004: 378–379) O = NP Petr. Sat. 26–68 (Polo 2004: 378–379) O = PP Petr. Sat. (Pinkster 1991: 72) Petr. Cena Trim. (Hinojo 1985) Celsius 1–6 (Pinkster 1991: 72) Celsius 1–4 (Pinkster 1992: 522) Pompey Inscriptions (Ramat 1984) Liv. (Amacker 1989) V+ 2 elements Terent. (Adams 1977: 68, 74–75) Terent. (Pinkster 1991: 72) Vetus, Ruth (Talavera 1981) Per. Aeth. (Va¨a¨na¨nen 1987: 106) Per. Aeth. (Cabrillana 1999: 321) O = NP Per. Aeth. 1–2 (Pinkster 1991: 72) Per. Aeth. (Hinojo 1986) Mul. Ch. (Cabrillana 1999: 321) O = NP Anon. Val. II (Adams 1976b: 136) Vulgata (100 sentences; Pinkster 1991: 72) Braulius (8th-c. ad; García Sanchidrián 1994) Lib. reg. (10th-c. ad; Carrera 1983)

100%:0% (100%:0%) 96.2%:3.8% (96%:4%) 39:45 (43:15) 33:22 (20:7) 28:20 30:15 35:45 67:33 14:7 18:4 17:0 81%:19% 71:3 60:10 79%:21% 54:45 (71:28) 75:3 393:49 73:26 (90:9) 8:4 166:125 (72:64) 352:139 (237:57) 123:27 (98:16) 52%:25% 57.6%:42.4% (82%:18%) 66:11 86.7%:13.26% 64.2%:35.8% 234:73 (149:35) 9:23 4:14 9.8%:90.2% 53:99 37.06%:62.94% 37:102 38.5%:61.5% 79.56%:20.44% 41.3%:58.7% 15:8 68.2%:31.8% (79.5%:20.5%) 61.9%:38.1% (67.6%:32.4%)

6.2 ov orders in latin

219

However, Table 6.1 shows clearly that classifying any non-archaic stage of Latin as an ‘OV language’ would be an oversimplification of the facts. Moreover, several studies have concluded that even in texts showing dominant OV ordering, specific classes of objects are frequently found postverbally: these are often characterized as heavy/complex (10) (Pinkster 1991: 73–75; Salvi 2004: 45–46; Polo 2005: 397) and focal in terms of information-structural status (11) (Panhuis 1982: 67; Pinkster 1991: 76; Spevak 2008: 119; Spevak 2010: 29, 125; Horrocks 2011: 132; Devine and Stephens 2019: 36),3 although this generalization is far from absolute (cf. the non-focal, non-heavy object, reum, in 12). Furthermore, various types of clauses are frequently found in VO configurations, as are adverbial expressions indicating a goal (13) (Bauer 1995: 98): (10) … notavimusque ostrea pectinesque e gastris notice.1pl.pst=and oysters.acc scallops.acc-and from pitcher.abl labentia, quae collecta puer lance tumbled-down which.acc gathered.acc slave.nom dish.abl circumtulit carry-round.3sg.pst ‘… and we noticed oysters and scallops tumbling from the pitchers; a slave gathered them, and carried them round in a dish’ (Latin, Cen. Trim 70, Polo 2005: 398–399) (11) secum habebat hominem χбηστoμαθη˜ himself-with have.3sg.pst man.acc scholarly (Greek) ‘He had with him a man who was scholarly’ (Latin, Cic. Att. 1.6.2, Horrocks 2011: 132) (12) Tot uadibus accusator uadatus so.many guarantees.abl prosecutor.nom admitted-to-bail.nom.m.sg est reum be.3sg accused.acc ‘With so many guarantees the prosecutor admitted the accused to bail.’ (Latin, Liv. 3.13.8, Danckaert 2017c: 437) (13) Ille aquam poposcit ad manus he.nom water.acc demand.3sg.pst for hands.acc ‘He called for water to wash his hands’ (Latin, Petr. 27, 6, Bauer 1995: 98) 3 Spevak (2010: 123), for example, notes that in the works of Cicero, postverbal objects are normally contextually ‘independent’.

220

ov orders and the middlefield

As such, a somewhat simplified generalization is that already in the classical period there is evidence that for some speakers OV ordering is not systematic but rather subject to syntactic and pragmatic constraints. Following the proposal in Ledgeway (2012a: 270), we can understand this as a transition from a grammar where all objects systematically move to Spec-VP via rollup movement (14a), to one where objects target the vP-peripheral topic and focus projections (14b), which we saw in our discussion of Classical Latin subjects in §5.2 were active in the language of the period:

(14) a.

VP Spec Object

V' V

Object

νPTopic

b.

νTopic'

Spec Object νTopic'

… νP v'

Spec ν V

VP Spec

V' V

Object

6.2.2 Evidence for change? Beyond the broad generalization that OV orders are typically more numerous than their VO counterparts and that—where an OV/VO alternation exists— this is governed by both pragmatic and purely syntactic factors, the diachronic progression of OV orders in the history of Latin is an area of controversy. Whilst the process of change is disputed, its eventual outcome is not: we

6.2 ov orders in latin

221

know that by the earliest French and Romance texts OV is a word-order variant found in a highly restricted number of contexts, albeit a productive one (cf. §6.3).⁴ Broadly speaking, similar claims have been made about the Peregrinatio, the syntactic properties of which we have discussed with reference to the left periphery (§3.2.5), verb movement (§4.2.3), and subject positions (§5.2.3). As Table 6.1 shows, this text shows a high proportion of VO (cf. 15)—between 61.5% and 72.9% depending on the study—with Ledgeway (2017b: 167) underlining that such data ‘highlight how late Latin had clearly transitioned from a (S)OV to a (S)VO language in line with a macroparametric shift (formalized in the loss of roll-up movement) from an original head-final to an innovative head-initial order’. Similar claims couched in distinct theoretical approaches are found in Adams (1976: 93), Bauer (1995: 91; 2009: 268), and Herman (2000: 86) amongst others. (15) et statim ingreditur intro spelunca and at-once enter.3sg.pass in cave.abl ‘and he immediately enters the cave’ (Latin, Peregrinatio 24, 2, Ledgeway 2017b: 173) However, scholars have challenged the notion of a neat transition from an OV to a VO grammar, ‘culminating’ in a systematically VO grammar in late texts on a number of grounds. Firstly, it is not evident that even the very earliest texts—aside from those like the Twelve Tables which are heavily formulaic— are consistently OV, which has led some scholars to date the emergence of the innovative VO grammar at a considerably earlier point than the post-classical period in which a number of ‘late’ texts are written. Secondly—as discussed in detail by Panhuis (1984: 153–156) and Ledgeway (2012a: 229)—strong evidence for an early reanalysis away from a consistently OV grammar comes from authors such as Caesar—referred to by Linde (1923: 155) as a ‘verb-final placement fanatic’—who makes extensive use of OV in military reports but not in narratives or topographical descriptions. With Adams (1976: 97), we can therefore conclude that for some authors from the time of Plautus onwards OV constituted ‘a stylistically marked variant or prestige order’, with this variant thus more marginal in subliterary texts such as the letters of Terentianus, which show a strong preference for VO ordering (Adams 1977: 68; Bauer 1995: 99–100; Clackson and Horrocks 2007: 256):

⁴ For other Romance varieties, see—amongst others—Poletto (2006a; 2006b; 2014) on OV in Old Italian, Mackenzie and Van der Wurff (2012) on Old Spanish, and Martins (2002; 2011) on Old Portuguese.

222

ov orders and the middlefield

(16) uidit Germani libertam a[dd]ente[m] see.3sg.pst Germanus.gen freedwoman.acc applying alia lineo … other.acc.pl linen.dat ‘She saw the freedwoman of Germanus applying different (decorations) to her linen’ (Latin, P. Mich. 469, Clackson and Horrocks 2007: 250) Thirdly, turning towards our latest texts, the idea that all late texts are consistently VO has also been cast into doubt (Pinkster 1991; Danckaert 2017c). Danckaert (2017c: 294), for example, notes on the basis of a corpus analysis that ‘in none of the datasets … is there any convincing evidence for there being significantly higher frequencies of VO in Late Latin than in earlier centuries’. This finding suggests that the ‘masking’ effects of normative influence identified by Adams (1976; 1977) can still be observed in the post-classical period, even if they are less prevalent in Christian Latin texts, such as the Peregrinatio and the letters of Saint Augustine (Ostafin 1986: 162; Bauer 1995: 101). As such, we conclude that even if the beginning and eventual outcome of the Latin OV > VO transition are identifiable as a grammar with obligatory rollup movement and one where a small subclass of objects moves to the vPperiphery respectively, mapping a clear diachronic trajectory throughout the period of Latinity is extremely challenging. Rather, it is easier to identify varying frequencies of VO as correlating with the genre and style of particular texts, with Adams (2016: 344) suggesting that ‘in the higher literary language conservatism and imitation of earlier models caused the preservation of a variability that included many OV features, whereas a different, VO, style was emerging in less literary texts’. Therefore, the labels ‘conservative’ and ‘innovative’ are more revealing when applied to particular texts than ‘early’ or ‘late’ in plotting how established they are relative to the loss of OV ordering.

6.2.3 Summary In this section we have suggested that as part of the progressive loss of headfinal ordering, Latin underwent a change where previously unmarked OV orders derived via systematic rollup movement to Spec-VP were reanalysed as the output of a grammar where a syntactically and pragmatically determined class of objects moved to the vP-periphery. As we will now see in §6.3, the ability of the vP-periphery to attract a specified class of objects persists into Old French.

6.3 ov orders and stylistic fronting in old french

223

6.3 OV orders and Stylistic Fronting in Old French 6.3.1 OV contexts across the clause The progressive loss of OV orders is rightly seen as one of the most significant changes to have taken place in French’s history (Brunot 1933: 255, 480–481, 492; Harris 1978: 9–11, 18–23; Buridant 1987; Buridant 2000: 747–749; Marchello-Nizia 1995: 96; Marchello-Nizia 1999: 39; Zaring 1998; Zaring 2010; Mathieu 2009: 345–349; Salvesen 2011; Labelle and Hirschbu¨hler 2017; Marchello-Nizia and Rouquier 2012; Marchello-Nizia and Prévost 2020: 1182–1210). However, we should note that a number of works treat all orders where objects precede verbs in the linear order as instances of OV, which arguably conflates a number of quite distinct syntactic properties. Marchello-Nizia (1999: 44–49), Zaring (1998; 2010; 2011), and MarchelloNizia and Rouquier (2012), for example, discuss object-initial V2 clauses like (17) alongside ‘low’ OV such as (18). (17) a. Icestui message ne ferai je … message do.1sg.fut I this neg ‘I will not deliver this message’ (OFr. Mort Artu 144, Marchello-Nizia 1999: 47) b. Ses felons gas, sa lengue male redotent tuit his wicked jokes his tongue evil fear.3pl all ‘Everyone fears his wicked jokes, his evil tongue’ (OFr. Perceval 2809, Zaring 2011: 1832) (18) a. li autre mout se traveillierent de lor oste the others much refl.cl work.3pl.pst of their guest bien aesier well ease.inf ‘The others took great pains to put their guest well at ease’ (OFr. Perceval 1929, Zaring 2010: 7) b. Quant il ço veit when he this see.3sg ‘When he sees this’ (OFr. Alexis 186, Marchello-Nizia and Rouquier 2012: 142) Parallels between the CP and vP phase aside (cf. Poletto 2014 and §6.3.2, and §6.5), there are very good reasons to treat these two constructions distinctly.

224

ov orders and the middlefield

As we saw in §3.3 and §4.3, clauses such as those found in (17) are derived through object topicalization or focalization to satisfy an Edge Feature on Fin or Force alongside verb movement to one of these heads; these two properties together form the V2 constraint (Holmberg 2015). The V2 constraint operates independently of the OV property, something we see readily attested in the Germanic languages, where Swedish, Danish, Icelandic, and Norwegian show VO ordering within the vP (19), despite being V2 systems (20) (Holmberg and Platzack 1995; Hróarsdóttir 2001; Thra´insson 2007; Westergaard 2009): (19) Han har ko¨pt boken he has buy.ptcp book-the ‘He has bought books’ (Modern Swedish, Holmberg and Platzack 1995: 7) (20) I den ha¨r skogen har jag a¨rligt talat aldrig sett in this here forest have I honestly speaking never see.ptcp huggormar adders ‘Honestly speaking, I have never seen adders in this forest’ (Modern Swedish, Holmberg 2015: 242) We note, therefore, that the number of object-initial V2 clauses declines from the earliest French textual records onwards (Kroch 1989; Marchello-Nizia 1999: 42–49; Steiner 2014; Marchello-Nizia and Prévost 2020: 1188), §§3.3.2, 3.4.1). Moreover, the information-structural characteristics change (see discussion and references in §3.3.6)—as also noted by Marchello-Nizia (1995: 99–100) among others—with left-peripheral objects in Later Old French and Middle French typically encoding discourse-old information (21) rather than focus. (21) Ceste aventure veïstes vos aucune foiz this event see.2pl.pst you some time ‘You have sometimes seen this event’ (OFr. La Queste 152.22, 13th century) It should be underlined that our contention here is that the properties of leftperipheral objects constitute an instructive parallel against which to compare the properties of objects appearing at the vP-edge; they are not, however, the ‘same’ construction. As such, ‘OV’ when used elsewhere in this chapter refers to objects appearing in a high position within the vP-complex, not within the CP. Cases such as (22) with an object appearing in a low structural position independent of V2 effects—analogous to Poletto’s (2006a; 2006b; 2014) cases of vP-scrambling in Old Italian—will be discussed in §6.3.2. I will argue that such cases should be differentiated from orders such as (23), where

6.3 ov orders and stylistic fronting in old french

225

other constituents from within the vP can undergo fronting to Spec-TP, which are traditionally labelled ‘Stylistic Fronting’ (cf. Mathieu 2006). These will be discussed briefly in §6.3.3. (22) et li rois qui tout ce ot esmeu … and the king who all this have.3sg.pst move.ptcp ‘and the king who had been moved by all this’ (OFr. La Queste 163a, 11, 13th century) (23) Qant levé furent del mangier when risen be.3pl.pst from-the eat.inf ‘When they had finished eating’ (OFr. Charrette 1043, 12th century, Labelle and Hirschbu¨hler 2017: 145)

6.3.2 vP scrambling Compared to certain other topics such as the loss of V2 and the loss of null subjects, the progressive loss of OV ordering in Old French has not generated a substantial literature. Nevertheless, several descriptive and theoretical claims have been made which are useful background to our analysis. In recent work Zaring (1998; 2010; 2011) has presented a nuanced picture of the relevant changes. Building on intuitions in Marchello-Nizia (1995: 78), she suggests that the proportion of OV found with periphrastic verb forms like (22) declines faster than before infinitives (cf. Table 6.2). Table 6.2 Rates of OV in Zaring (2010)

Perceval Conqueste

Infinitive

Periphrastic

60% 47%

43% 28%

Significantly, she also shows that the information-structural function of the preverbal object is distinct between texts. Whilst in the 12th-century Perceval, ‘all functions are attested with non-finite verbs’, in the 13th-century Conqueste pre-participial objects are not found when they ‘are accessible but only loosely tied to the discourse’ (Zaring 2010: 10). We should note that this also tallies with the description offered in Buridant (1987; 2000: 748–749), who stressed the specifically ‘anaphoric’ nature of objects in OV contexts in Old French in general.⁵ Likewise, classically focal constituents such as indefinites and ⁵ See also Moignet (1973: 356–362) and Scrivner (2014) on this particular point.

226

ov orders and the middlefield

non-idiomatic bare nouns are also not found in Zaring’s corpus. She reports that pre-infinitival objects are less restricted than their pre-participial counterparts, but that definite objects ‘loosely tied to the discourse’ are not numerous. The generalization is that syntactic environment is relevant to determining the discourse function of a preverbal object and, furthermore, that this function is subject to diachronic change. We should also note that Mackenzie and van der Wurff ’s (2012) discussion of OV orders in Medieval Spanish mirrors to a certain extent proposals made by Zaring for Old French. They show that a particular class of fronted objects favours OV orders above others: this includes quantified, negative, and discourse-active nominals (Mackenzie and van der Wurff 2012: 851–857). As a starting point for our discussion, recent analysis by Marchello-Nizia and Prévost (2020) allows us to track the loss of OV in different subordinate clause-types, although we should note that their analysis neither singles out scrambling specifically nor focusses on OV found with infinitives. Nevertheless, Table 6.3 shows clearly that OV of any kind is exceptionally rare from the 13th century onwards. Our own data are therefore focussed on this period.⁶ Table 6.3 Rates of VO in Marchello-Nizia and Prévost (2020) Text

Complement Clause

Circumstantial Clause

Relative Clauses

Passion (c.1000) St. Alexis (c.1050) Roland (c.1100) Lapidaire (c.1150) Eneas (c.1155) Troyes Yvain (1177–1181) Ch Chièvres (1194) Aucassin (Verse) (c.1200) Aucassin (Prose) (c.1200) Renart Dole (1228) La Queste (c.1225) Joinville (1305–1309) Quinze Joies (c.1400)

2/3 6/11 16/24 11/12 15/25 38/50

66% 55% 67% 92% 60% 76%

8/27 9/12 7/9 25/25 21/24 14/24

30% 75% 78% 100% 88% 58%

6/12 6/14 5/22 35/39 11/22 18/32

50% 43% 23% 90% 50% 56%

10/11 2/3

91% 67%

13/14 4/5

93% 80%

6/10 8/8

60% 100%

16/16

100%

48/50

96%

12/17

71%

23/25 14/18 46/47

92% 78% 98%

39/40 25/28 29/29

98% 89% 100%

30/37 20/24 19/25

81% 83% 76%

28/31

90%

26/26

100%

38/39

97%

⁶ Circumstantial and relative clauses are likely to be particularly instructive, owing to the pervasiveness of scrambling in these environments.

Roland (c.1100) Thèbes (c.1150) QLR (c.1190) La Queste (c.1225) Chroniques (1300)

DP/PP Active

DP/PP Accessible

DP/PP New

Pronoun

QP

Total Active

Total Accessible

Total New

Grand Total

2

5.4%

20

54.1%

12

32.4%

0

0.0%

3

8.1%

2

5.4%

20

54.1%

15

40.%

37

100.0%

12

33.3%

12

33.3%

10

27.8%

1

2.8%

1

2.8%

13

36.1%

12

33.3%

11

30.6%

36

100.0%

13 14

27.7% 20.9%

15 14

31.9% 20.9%

10 4

21.3% 6.0%

4 19

8.5% 28.4%

5 16

10.6% 23.9%

17 33

36.2% 49.3%

15 14

31.9% 20.9%

15 20

31.9% 29.9%

47 67

100.0% 100.0%

15

41.7%

8

22.2%

4

11.1%

5

13.9%

4

11.1%

20

55.6%

8

22.2%

8

22.2%

36

100.0%

6.3 ov orders and stylistic fronting in old french

Table 6.4 OV before past participles

227

228

Roland (c.1100) Thèbes (c.1150) QLR (c.1190) La Queste (c.1225) Chroniques (1300)

DP/PP Active

DP/PP Accessible

DP/PP New

Pronoun

QP

Total Active

Total Accessible

Total New

Grand Total

2

8.3%

14

58.3%

3

12.5%

2

8.3%

3

12.5%

4

16.7%

14

58.3%

3

25.0%

24

100.0%

15

16.9%

45

50.6%

12

13.5%

11

12.4%

6

6.7%

26

29.2%

45

50.6%

12

20.2%

89

100.0%

14

30.4%

19

41.3%

7

15.2%

5

10.9%

1

2.2%

19

41.3%

19

41.3%

7

17.4%

46

100.0%

5

20.8%

11

45.8%

2

8.3%

6

25.0%

0

0.0%

11

45.8%

11

45.8%

2

8.3%

24

100.0%

8

24.2%

6

18.2%

10

30.3%

9

27.3%

0

0.0%

17

51.5%

6

18.2%

10

30.3%

33

100.0%

ov orders and the middlefield

Table 6.5 OV before infinitives and gerunds

6.3 ov orders and stylistic fronting in old french Active

Accessible

229

New

60.0%

45.0%

30.0%

15.0%

0.0%

Roland

Thèbes

QLR

Queste

Chron

Fig. 6.1 OV before past participles Active

Accessible

New

60.0%

45.0%

30.0%

15.0%

0.0%

Roland

Thèbes

QLR

Queste

Chron

Fig. 6.2 OV before infinitives and gerunds

Following a similar methodology to that used for information-structural analysis of subjects in §5.3, the data for scrambled objects before participles in Table 6.4 and before infinitives and gerunds in Table 6.5 come from Wolfe (in prep.), with Figures 6.1 and 6.2 showing the proportion of active, accessible, and new objects undergoing scrambling in the respective configurations.

230

ov orders and the middlefield

Overall, the findings from our corpus broadly support Zaring’s (2010; 2011) claims regarding diachronic change in the nature of OV, but differ in certain respects. Looking first at OV at the vP-edge in pre-participial position, Table 6.4 shows that active (24), accessible (25), and new objects (26) are licensed throughout the Old French period but also attests a clear trend for such objects to be pronominal (27) or already active in the discourse—rather than merely encoding accessible information—as the 12th and 13th centuries progress. Thus, witness in Table 6.4 the fact that 2/37 objects in the Roland are active (5.4%), compared to 33/67 (49.3%) in La Queste, and 20/36 (55.6%) in Chroniques. (24) a. Quant la dame cest los oÿ when the woman this advice hear.ptcp ‘When the woman heard this advice’ (OFr. Thèbes 459, 12th century) b. Idunc entendi Hely que Deu out Samuel then understand.3sg.pst Hely that God have.3sg Samuel a´pelé call.ptcp (OFr. QLR 9, 12th century) (25) a. De sun osberc li ad les pans rumput of his armour him.cl have.3sg the flaps shatter.ptcp ‘The flaps of his armour [hauberk] have shattered’ (OFr. Roland 1601, 12th century) b. … quant il ot le glaive brisié when he have.3sg the sword smash.ptcp ‘When he has smashed his sword’ (OFr. La Queste 173a, 13th century) (26) a. puis l’ont a un chesne pendu then him.cl=have.3pl to a oak-tree hang.ptcp ‘Then they hanged him from an oak-tree’ (OFr. Thèbes 116, 12th century) b. Mes but cest this

il avint it happen.3sg.pst païs, et il country and they

quant il furent venu en when they be.3pl.pst come.ptcp in orent grant piece erré have.3pl.pst great amount travel.ptcp

6.3 ov orders and stylistic fronting in old french

231

par les estranges terres que … through the foreign lands that ‘But it came to pass when they arrived in this country and had travelled a great deal through the foreign lands that …’ (OFr. La Queste 177d, 13th century) (27) a. Quant il out çó eslı´t, nostre when he have.3sg this declare.ptcp our Sires enveiad pestilence en Israel … Lord send.3sg.pst plague in Israel ‘When he had declared this, our Lord sent a plague to Israel’ (OFr. QLR 108, 12th century) b. et quant il a ce fet si s’asieent and when he have.3sg this do.ptcp si refl.cl-sit.3pl tuit entor Galaad all around Galahad ‘and when he had done this all sat around Galahad’ (OFr. La Queste 173c, 13th century) Very similar observations apply to objects preceding infinitives and gerunds, where we observe that active (28), accessible (29), and new objects (30) are found across all texts, but that the frequency of accessible objects declines concomitantly with an increase in the frequency of objects—be they DPs, PPs, or pronouns—which are already active in the preceding portion of text and constitute between 41.3% and 51.5% of the sample after the late 12th century: (28) a. Ne vus devez pas pur çó curecier neg you must.2pl neg through this anger.inf ‘You mustn’t anger yourself through this’ (OFr. QLR 97, 12th century) b. Et li rois descent maintenant por ceste aventure and the king descend.3sg now for this event veoir see.inf ‘And the king came down at once to see the event’ (OFr. La Queste 161a, 13th century) (29) a. E Blancandrins, por la raisun cunter … and Blancandrin for the story tell.inf ‘And Blancandrin, to tell the story …’(OFr. Roland 68, 12th century)

232

ov orders and the middlefield

b. mes se tu veus le non savoir … but if you want.2sg the name know.inf ‘but if you wish to know the name …’ (OFr. Thèbes 25, 12th century)

(30) a. La veïsez tant chevaler plorer there see.2pl.pst.sbjv so-many knights weep.inf ‘there you had seen so many knights weep’ (OFr. Roland 349, 12th century)

b. car ce est examples de bone vie mener for this be.3sg example of good life lead.inf ‘for this is an example of how to lead a good life’ (OFr. Chroniques 3, 14th century) The qualitative conclusion is therefore that the conditions under which scrambling to the vP-edge obtains progressively tighten throughout the Old French period, such that unambiguously discourse-old objects and pronominals make up an increasingly large portion of the data. This provides quantitative support to Zaring’s (2010; 2011) proposals regarding the nature of the fronted object in Later Old French. Furthermore, we should note that the finding is not particularly surprising in light of our discussion of the Latin data above in §6.2: new-information/focussed objects are some of those which most commonly remain in-situ even in early classical texts with old-information objects more frequently fronted (Panhuis 1982: 67; Pinkster 1991: 76; Spevak 2008: 119; Spevak 2010: 29, 125; Horrocks 2011: 132; Devine and Stephens 2019: 36). In addition, scrambling operations elsewhere in Romance and Germanic have frequently been couched in terms of defocussing a particular constituent (Diesing 1997; Karimi and De Hoop 2003; Costa 2004: 32–42; Haider 2006: 219–223; Neeleman and van de Koot 2008: 170–179; Martins 2011: § 2), which partially parallels our claims here that explicitly non-focal constituents are most likely to be found in the scrambling position in the vP-periphery. Finally, a significant finding in the light of Poletto’s (2006a; 2006b; 2014) interpretation of the Uniformity of Phases hypothesis is that the vP-periphery increasingly specializes in hosting discourse-old objects. This exactly parallels our discussion of the ‘high’ CP-periphery in §3.3 and §3.4, where we saw that foci are increasingly restricted whereas old-information topics can occupy the prefield to a growing degree throughout the Old and Middle French periods.

6.3 ov orders and stylistic fronting in old french

233

6.3.3 Stylistic Fronting Before concluding our discussion of OV orders in Old French, we should briefly discuss a related phenomenon which has been widely discussed in the literature, namely Stylistic Fronting.⁷ Originally described and analysed on the basis of Icelandic data in Maling (1990), Stylistic Fronting is distinct from the V2 property in that it typically involves fronting of heads, such as past participles, infinitives, negation, and short adverbs, is most productive in embedded clauses, and is subject to a strict accessibility hierarchy based on principles of the Minimal Link Condition (Chomsky 1995) as in (31), where in the presence of negation, a predicative adjective, and a non-finite verb and particle, negation will preferentially satisfy Stylistic Fronting over the adjective; if neither negation nor a predicative adjective are present, either the verb or—if present—a verbal particle can move, as in (32): (31) Negation > AdjectivePredicative > Verb (/Particle) (32) a. kosningarnar sem farið hafa fram the-elections that gone have forth ‘the elections that have taken place’ (Modern Icelandic, Jónsson 1991 as cited in Holmberg 2000b: 450) b. kosningarnar sem fram hafa farið the-elections that forth have gone ‘the elections that have taken place’ (Modern Icelandic, Jónsson 1991 as cited in Holmberg 2000b: 450) Whilst apparently similar structures have been discussed in the literature for Old and Middle French (Cardinaletti and Roberts 2002; Mathieu 2006; Mathieu 2009; Labelle 2007; Labelle 2016; Labelle and Hirschbu¨hler 2017; Salvesen 2011; Salvesen 2013; Hansch 2014)—where negation, adjectives, adverbs, and non-finite verbal forms undergo some form of fronting (33)—there is no consensus on how to analyse such structures or how to delimit the phenomenon.

⁷ See—amongst many others—Maling (1990), Jónsson (1991), Holmberg (2000b; 2017), Cardinaletti and Roberts (2002), Fischer and Alexiadou (2001), Fischer (2004), Hrafnbjargarson (2004), and Poole (2007).

234

ov orders and the middlefield

(33) a. Einsi vint à un parlement à Soissons qui thus come.3sg.pst to a parliament at Soissons that només fu nominate.ptcp be.3sg.pst ‘Thus he came to a parliament in Soissons which was chosen’ (OFr. Villehardouin 17, 13th century, Mathieu 2006: 235) b. Li reis creit Deu, faire voelt sun service the king believe.3sg God do.inf want.3sg his service ‘The King believes in God, and he will be his servant’ (OFr. Roland 3666, 12th century, Salvesen 2011: 336) c. Dreit les meinet a un castel //Qui riches ert e direct them.cl lead.3sg to a castle which rich be.3sg and grant e bel large and beautiful ‘He leads them directly to a castle which is rich and large and beautiful’ (OFr. Brendan 37.159, 12th century, Labelle and Hirschbu¨hler 2017: 150). The recent debate in the Old French literature has centred on how homogeneous a phenomenon Stylistic Fronting is in the medieval period and whether it shows the same structural properties as those reported for the Modern Germanic languages (cf. for example Mathieu 2006 vs. Labelle 2007). For our purposes, it suffices to adopt the recent proposals put forward in Salvesen (2011; 2013) and Labelle and Hirschbu¨hler (2017). Salvesen (2011) argues persuasively that—in line with parallel proposals in the Germanic literature— Stylistic Fronting is phrasal movement of a (remnant) vP, although I depart from Salvesen in assuming this targets a TP-internal position. Salvesen (2011) correctly identifies that coupling this line of analysis with an account of scrambling of the type outlined above in §6.3.3 allows us to account for examples where multiple constituents appear to have undergone fronting to a clauseinternal position as in (34, 35) (cf. also Labelle and Hirschbu¨hler 2017: 161). For ease of exposition, I simplify the internal structure in (35) to convey the key points of the analysis: (34) Se trestout ce bien gardé unt if all this well keep.ptcp have.3pl ‘If they have kept all of this well’ (OFr. Boron 88.1396, 12th century, Labelle and Hirschbu¨hler 2017: 161)

6.4 relic ov orders from middle to modern french

235

(35) a. se unt [vP bien gardé trestout ce] b. se unt [vPTopic trestout ce [vP bien gardé trestout ce]] (Scrambling) c. se [TP [vPTopic trestout ce bien gardé] [T unt] [vPTopic trestout ce bien gardé] (Stylistic Fronting) Labelle and Hirschbu¨hler (2017) also adopt a number of similar assumptions, though they do suggest that movement of a phrase to a specifier position immediately below the position of the subject is just one of several constructions they subsume under the heading ‘Leftward Stylistic Displacement’. The important point for our purposes is to note that we must posit an additional clause-internal movement operation alongside scrambling; unlike scrambling, this does not target the vP-edge but rather a position within the extended T-layer.

6.3.4 Summary In this section we have seen evidence that phrasal movement of objects can target the vP-periphery in Old French in a similar fashion to what has been proposed elsewhere in Medieval Romance, i.e. Old Italian Poletto (2006a; 2006b; 2014). Whilst objects with a range of information-structural values can front to the vP-edge, a strong tendency for these to be active in the preceding portion of text and/or pronominal emerges in the 13th century. We also saw in §6.3.3 that vPs can undergo fronting to a TP-internal specifier position in Old French.⁸ (36) [CP [TP vPStylistic Fronting [vPTopic Scrambled Object [vPFocus Scrambled Object [VP … ]]]]]

6.4 Relic OV orders from Middle to Modern French With a small number of exceptions which we discuss, OV orders are overwhelmingly to be considered a Latin and Old French phenomenon, with scrambling exceptionally rare in Middle French, and Labelle and Hirschbu¨hler (2017: 166) noting that ‘in the fifteenth century … both V2 and LSD [Leftward Stylistic Displacement–SW] are disappearing’. That postverbal positioning is nearly fixed by the 15th century also finds support from Table 6.3 and in Marchello-Nizia and Prévost’s (2020: 1197) observation that Middle French offers only ‘sparse occurrences’ of embedded SOV, although they do give ⁸ For simplicity’s sake, this is simply equated with Spec-TP here. However, Labelle and Hirschbu¨hler (2017: 167) suggest that this is a position between SubjP and TP.

236

ov orders and the middlefield

the following example of scrambling from the 15th century (37), which I supplement with two from Chronique (38):⁹ (37) Et vint ceste Fortune la / Quant nous and come.3sg.pst this fortune there when we eusmes le fort conquis have.1pl.pst the fort conquer.ptcp ‘And great luck came when we had conquered the fortress’ (MidFr. ArchierBaignollet 31, 15th century, Marchello-Nizia and Prévost 2020: 1197) (38) a. Lesquelx furent de ce accusés the-which be.3pl.pst of this accue.ptcp ‘Those who were accused of this’ (MidFr. Chronique 3, 15th century) b. … où ils eurent les hateriaulx coppés where they have.3pl.pst the portions-of-meat cut.ptcp ‘… where they had cut the portions of meat’ (MidFr. Chronique 2, 15th century) Scrambling is also licensed in Renaissance and Classical French, though to a very limited degree. Gougenheim (1973: 255) notes that insertion of the object between the auxiliary and the past participle is ‘an ancient construction, which is maintained until the 17th century’; Fournier (2007a: 93) confirms this dating, giving several examples from 17th-century French, whilst noting that pre-participial objects such as (39, 40) are more common than pre-infinitival objects as in (41), which she describes as ‘very rare’ (Fournier 2007a: 94).1⁰ (39) Tu sais bien que j’avois tes desastres you know.2sg well that I-have.1sg.pst your disasters predit predict.ptcp ‘You know well that I have predicted your disasters’ (RenFr. Garnier, Les Juifves 1, 44, Gougenheim 1973: 255)

⁹ Muller (2009: 275) discussing Middle French notes—for example—that ‘in subordinate clauses, word order is not very different from that of Modern French’, a telling observation given that scrambling is most easily identifiable in the embedded domain. 1⁰ Additional evidence for the low attestation of OV orders comes from Combettes’ (2003a: 179) study of 16th-century travel writing, where SOV orders are ‘not attested’.

6.4 relic ov orders from middle to modern french

237

(40) J’avais de point en point l’entreprise tramée I-have.1sg.pst entirely the-enterprise plot.ptcp ‘I have hatched the plan entirely’ (CFr. Corneille, Clit. 429–430, 17th century, Fournier 2007a: 93) (41) il voulait son honneur réparer he want.3sg.pst his honour repair.inf ‘He wanted to repair his reputation’ (CFr. La Fontaine, Contes 2, 1, 17th century, Fournier 2001: 94) The conclusion, therefore, is that relic instances of scrambling are found until the 17th century, after which the ability for objects or verbal complements of any kind to access the vP-edge becomes extremely restricted. From Middle French onwards, certain quantifiers such as beaucoup ‘a lot’, trop ‘too much’, and rien ‘nothing’ can precede past participles and infinitives (42, 43) (Le Bidois and Le Bidois 1938: 47; Rowlett 2007: 108, 112–113; Grevisse and Goosse 2016), as part of a more general ability to front to various positions within the clause, noted since Kayne (1975: §1.11).11 Aside from quantifiers, OV with nominals is only maintained in fixed expressions such as chemin faisant (lit. route make.prog) ‘on the way’, sans mot dire (lit. without word say.inf) ‘without saying a word’, ce disant (lit. that say.prog) ‘saying that …’ (Le Bidois and Le Bidois 1938: 48; Grevisse and Goosse 2016: 411). (42) a. Assez de foys furent assembléz sans riens enough of times be.3pl.pst assemble.ptcp without nothing faire … do.inf ‘They were assembled enough times without doing anything …’ (MidFr. Mémoires 27, 15th century) que sommes icy sans b. … ja´ long temps a already long time have.3sg that be.1pl here without rien faire … nothing do.inf ‘it’s already been a long time that we’ve been here without doing anything’ (RenFr. Rabelais 10, 47, 16th century)

11 For a detailed overview of Quantifier Float in Modern French, see Cinque (1999: 119–120; 2002) and Rowlett (2007: §4.3.1)

238

ov orders and the middlefield

c. pour la servir j’ai tout fait … to her.cl serve.inf I-have.1sg all do.ptcp ‘to serve her, I’ve done everything’ (CFr. Phèdre 4, 6, 1327, 17th century) (43) a. Je n’ ai rien vu I neg have.1sg nothing see.ptcp ‘I’ve seen nothing’

(ModFr.)

b. J’ ai tout dit I have.1sg all say.ptcp ‘I’ve said everything’

(ModFr.)

c. J’ ai beaucoup fait I have.1sg a-lot do.ptcp ‘I’ve done a lot’

(ModFr.)

We should note in passing that the licensing of scrambling has been reported in Alsace French, likely maintained through influence from Alsatian, with Wolf (1983: 193–194; 2000: 698) reporting cases such as those in (44) where we see objects preceding infinitives and past participles: (44)

a. Je dois quelque chose chercher I must.1sg some thing find.inf ‘I must find something’ (Modern Alsace French, Wolf 2000: 698) b. J’ai le travail fini I-have.1sg the work finish.ptcp ‘I’ve finished the work’ (Modern Alsace French, Wolf 2000: 698)

The obvious assessment is therefore that whilst some form of scrambling persisted as a productive—albeit marginal—word-order possibility in the standard language until the 17th century, the possibility for any form of verbal complement to front to the vP-edge is extremely restricted in Modern French. Lexical exceptions and quantifier-movement aside, we therefore witness in the modern language the culmination of a change in head-directionality, which we saw was already under way in Latin (cf. §6.2), such that the order of constituents within the extended vP is increasingly fixed head-initially.12 Overall—and considered alongside the properties analysed in Chapters 3–5— this contributes to Modern French being straightforwardly characterized as SVO. 12 In contrast to the conclusion reached by Marchello-Nizia and Prévost (2020: 1149) that clauses such as (i, ii)—discussed in §3.5.1—evince the re-emergence of an OV grammar, they should not be taken as such. Instead, they evidence the widespread use of base-generated topicalization structures in

6.5 summary and conclusions

239

6.5 Summary and conclusions 6.5.1 Summary Whilst it is tempting to draw a sharp macrotypological contrast between ‘OV’ Latin and ‘VO’ French, the analysis put forward in this chapter shows quite clearly that such a stance is unsustainable. Rather—possibly from the very earliest attested stages of Latin onwards—the class of objects permitted to undergo movement to the vP-periphery or T-layer becomes increasingly qualitatively restricted; this is part of a wider set of changes in the transition from a harmonic head-final system to a harmonic head-initial system, via a long-lasting intermediate disharmonic stage. As discussed in §6.2, the study of the OV > VO transition in the history of Latin is riven with controversy. The proposal adopted here—following Ledgeway (2012a)—is that the transition from a consistent OV system to a mixed OV/VO system is best accounted for under the loss of complement-to-specifier rollup movement to Spec-VP (Kayne 1994), in favour of syntactically or pragmatically conditioned topicalization or focalization to the vP-periphery. Whilst the change may have occurred earlier—as argued by Adams (1976; 1977; 2016) amongst others—it is clear that certain speakers had access to the innovative system by the time of Plautus (254–184bc). What is also clear is that in certain grammars, the restrictions on when objects moved out of their base-generated positions to the vP-edge became sufficiently restricted that VO orders outnumber OV orders, sometimes to a significant extent, such as we find in the writings of Terentianus (Adams 1977; Bauer 1995) and in Christian Latin texts such as the Peregrinatio (Clackson and Horrocks 2007; Ledgeway 2017b). Perhaps the most striking observation regarding the Early Old French system of object-scrambling described in §6.3 is its underlying similarity to the innovative system described for certain subliterary, colloquial, and late Latin texts; that is to say, a system where objects can front to the vP-periphery with

various contemporary varieties of French rather than movement out of the vP-complex, which, as our discussion in §3.5.1 shows, is heavily restricted. (i) Le chocolat il (l’) aime the chocolate he it.cl like.3sg ‘Chocolate, he likes (it)’

(Colloquial French)

(ii) Dix euros il me doit ten euros he me.cl owe.3sg ‘Ten euros, he owes me’

(Colloquial French)

240

ov orders and the middlefield

a range of information-structural values and where OV orders are most frequent in the embedded domain. As such, we see that positing a binary split between Latin and French in the OV/VO typology is inherently problematic. Turning to variation within the Old French period, our analysis supports findings by Zaring (2010; 2011) that diachronic variation is attested in the licensing of scrambling: Later Old French texts show a greater proportion of discourse-active or pronominal objects in a scrambled position than their Early Old French counterparts. This qualitative restriction sits alongside a general quantitative decline in the frequency of scrambled objects, which decline markedly in Middle French texts and are entirely absent from the 17th century onwards. We saw in §6.4 that whilst no scrambling is attested beyond this point, nominal objects are found to the left of participles, gerunds, and infinitives in certain lexically fixed expressions and that quantifiers do— exceptionally—front to the vP-edge in Modern French unlike other verbal complements. A somewhat simplified summary of the data discussed and analysed in this chapter is presented in Table 6.6. Table 6.6 OV from Proto-Indo-European to Modern French Proto-Indo-European; Archaic Latin Classical Latin Subliterary, Colloquial, and Late Latin; Early Old French Later Old French Middle, Renaissance, and Classical French Modern French

Generalized Rollup to Spec-VP Widespread Fronting to SpecvPTopic/Focus . Restricted Class of In-Situ Objects Widespread Class of In-Situ Objects. Restricted Fronting to Spec-vPTopic/Focus Restricted Fronting to Spec-vPTopic/Focus , most commonly to Spec-vPTopic. Highly Restricted Fronting to SpecvPTopic/Focus , most commonly to Spec-vPTopic. Generalized In-Situ Objects. Lexically Restricted Quantifier Fronting to the vPperiphery and OV in fixed expressions.

6.5.2 From OV to VO The Final-Over-Final Condition (Sheehan et al. 2017) makes a clear prediction—as outlined by Ledgeway (2012a)—that in the shift from a harmonically head-final system to a harmonically head-initial one, ‘high’ heads within an extended projection, such as prepositions, determiners, and

6.5 summary and conclusions

241

complementizers, will change in directionality before ‘low’ heads such as verbs and nouns. We noted in §6.2 that the predictions of this account for Latin and French are correct, with prepositions and complementizers amongst the first categories to occur initially, whereas the ordering of verb and object—as we have seen—does not become consistently head-initial until the 17th century in the case of French. Crucially, a transitionary stage lasting nearly two millennia exists where objects could optionally front to the vP-periphery through scrambling: (45)

Stage One > Comp-to-Spec Rollup (All Categories) Stage Three > Widespread Object-Fronting to vP-Edge Restricted In-Situ Objects Stage Five Generalized In-Situ Objects

Stage Two > Comp-to-Spec Rollup (Defined Categories) Stage Four > Restricted Object-Fronting to vP-edge Restricted Fronted Objects

Viewed in this perspective, it becomes clear that the diachrony of object-verb order should not be viewed entirely in isolation but rather in the context of other clausal word-order properties, linked to the distribution of phrasal movement-triggering Edge Features across the clausal hierarchy. Roberts (2012a; 2021), for example, suggests that the innovation of scrambling to the vP-edge may be a necessary precursor to the emergence of V2; under a standard interpretation of the Phase Impenetrability Condition constituents in the lower vP-phase will need to occupy a position at its edge in order to be visible to a probing head in the higher phase: Fin in the case of late Latin and Early Old French V2 (cf. §§4.2.3, 4.3.1). As such the significant reanalysis under which Comp-to-Spec-VP movement is reanalysed as movement to the vP-periphery has ramifications beyond the mere distribution of objects in leading to the eventual emergence of an Edge Feature on Fin as part of a V2 grammar. Building on this intuition—and as set out in Wolfe (2016a: 488) and §4.6—we can see the upwards spread of movement-triggering Edge Features from V to v (Classical Latin), v to Fin (subliterary, colloquial, late Latin), and later Fin to Force (Later Old French) as a central step in the Latin to French transition. The central idea in Poletto (2006a; 2006b; 2014) that the syntactic makeup of phase-edges is essentially uniform in Old Italian, sheds new light on the data presented in this section. Considering Latin, we see that the very same texts which show the highest levels of VO, such as the writings of Terentianus

242

ov orders and the middlefield Do all heads bear an EF triggering rollup? Yes Harmonic head-final Proto-Indo-European13

No An identifiable subclass of heads? Yes V {+EF} = Object rollup Archaic Latin

No An identifiable subclass of objects?

Yes v {+EF} for+Topic/Focus Objects Old to Classical French

No Lexically defined variation Modern French

v {+EF} for fixed expression, defined quantifiers

Fig. 6.3 A parameter hierarchy for head-finality and object-movement

and the Peregrinatio—having innovated scrambling to the vP-edge—also show high levels of verb-medial orders indicative of widespread topicalization and focalization at the CP-edge (cf. §3.2.5). Looking to Old French, not only do the existence of V2 at the CP-level and scrambling at the vP-level indicate a parallelism between phase-edges, a change in the conditions under which constituent-fronting occurs also occurs in parallel at both edges: whilst in Later Old French we see a general decline in the licensing of focus (Labelle and Hirschbu¨hler 2018) and a tendency towards unambiguously discourseold constituents undergoing topicalization (cf. discussion and references in §3.3.6), at the vP-level we observe that an increasing proportion of scrambled objects are either pronominal or active in the preceding portion of text. This suggests that parallels across phase-edges can extend beyond the presence or absence of Edge Features and head movement-triggers and include the syntax–pragmatics mapping. Finally, we should note that the empirical picture regarding the diachronic progression of object placement in the history of Latin and French fits strongly with the taxonomy of parameters offered in Roberts (2019) and other work. As our proposed parametric hierarchy in Figure 6.3 shows, we observe a clear progression from the macroparametric property of harmonic head-finality across all heads, to the mesoparametric property of object rollup, to a microparametric movement trigger for topicalized and/or focalized objects, subsequently yielding the nanoparametric variation observed in Modern French where a lexically defined class of exceptions raise to the edge of the vP.

13 Although arguably the most established view, reconstructing Proto-Indo-European as harmonically head-final is not self-evident (see Lehmann 1974; Miller 1975; Viti 2014; Hock 2015; Ringe 2017).

7 A new perspective on syntactic change in French 7.1 Major findings 7.1.1 Overall summary Before considering the significance of the findings overall, we should summarize the major findings stemming from Chapters 3–6, this time considering the phenomena discussed side by side. Our hypothesis is that in the most archaic form of Latin grammar, the verb and subject remain within their base-generated position within the vP in the unmarked case, with the object, however, undergoing systematic rollup movement to Spec-VP. As predicted under the account of null arguments offered in Roberts (2010b), Latin at this stage is a generalized null-argument language. From the very earliest Latin textual records onwards, there is clear evidence that the finite verb and its arguments can undergo movement to positions within the left periphery. We saw in §3.2 that the left-peripheral space is highly articulated and can host a range of frame-setters, topics, and foci, as well as QPs. A range of Latin subordinators are base-generated in Fin, which heavily restricts the possibility of embedded focalization and topicalization and yields a strict embedded SOV order. Two major changes take place in Classical Latin, although the first may have taken place even earlier: previously systematic rollup movement of the object to Spec-VP undergoes reanalysis as discourse-marked fronting of the object to vP-peripheral topic and focus positions—scrambling—only under certain conditions; at the same time, verb movement undergoes upwards reanalysis to v. The second change in progress, noted by various authors, is the increase in verb-medial orders, which we analysed as resulting primarily from the increasing frequency of syntactically or pragmatically marked V-to-C movement.

Syntactic Change in French. Sam Wolfe, Oxford University Press. © Sam Wolfe (2021). DOI: 10.1093/oso/9780198864318.003.0007

244

a new perspective on syntactic change in french

In the most innovative late Latin texts, which undoubtedly foreshadow a number of syntactic properties seen in Early Old French, we find evidence that marked V-to-C movement has been reanalysed as unmarked V-to-Fin movement, yielding a grammar with extensive verb–subject inversion structures. Whilst scrambling is still attested, the proportion of VO orders found is also greatly increased. Our analysis in §3.2 suggests that topicalization or focalization remains a marked word-order pattern in late Latin, with the language thus having ‘half the V2 constraint’ in the terms of Roberts. We saw in §6.2.2 that there is an increasing amount of SVO in late Latin alongside embedded V2, with subordinators targeting either Fin or Force. Turning to the earliest French textual records, Early Old French is a full V2 system, where an Edge Feature and verb-movement trigger are associated with Fin. At this stage of the language’s history, a broad range of left-peripheral frame-setters, topics, and foci are licensed. Scrambling is attested for objects with a range of pragmatic values, and postverbal subjects can appear in either a TP-internal position—when pronominal or encoding accessible or active information—or a vP-peripheral position, which typically encode new information. V2 effects and inversion are more restricted in embedded clauses, though are licensed in a wider class of embedded domains than is the case in Later Old French; we analysed this as a result of the fact that subordinators can be base generated in either Fin or Force. Lastly, note that Early Old French is a full null-subject system, where proTop is licensed in both matrix and subordinate clauses, showing the same broad array of properties which are found in Modern Romance full null-subject systems. A major reanalysis takes place at the turn of the 13th century, with both V2-related features—a verb-movement trigger and Edge Feature—being reanalysed upwards on Force. As suggested in §3.3, under analogy with the innovative Fin-to-Force movement found in matrix clauses, complementizers also increasingly undergo Fin-to-Force movement, leading to greater matrix/embedded asymmetries than are found in Early Old French. As such, embedded V2, inversion, and verb-third orders become rarer, as do embedded null subjects/topics. In Later Old French the class of constituents which can occur in the left periphery is increasingly restricted to frame-setters and topics, though foci are still attested to a marginal degree. In a parallel fashion, scrambling also follows the same trajectory, with discourse-old constituents the class most frequently appearing at the vP periphery in 13th-century texts. The properties of inversion are also distinct from Early Old French in that the TP-internal Germanic-inversion position is increasingly specialized in hosting either pronouns or nominal subjects which are active in the preceding portion

7.1 major findings

245

of text. This contrasts with the behaviour of postverbal subjects occurring in the extended vP-complex which encode a range of information-structural values. A body of evidence suggests that Middle French remains a full V2 language, albeit one where the target of verb movement and the locus of the V2-associated Edge Feature is Fin—like Early Old French but unlike Later Old French. Certain points of continuity do, however, exist with Later Old French: the information-structural status of TP and vP-internal subjects is broadly similar, foci are increasingly restricted in the high CP left periphery, and scrambling to the vP periphery continues to decline in frequency. Points of discontinuity, however, include a weakening of the very sharp matrix/embedded asymmetries which are characteristic of Later Old French and a widening of the qualitative environments in which null subjects are attested, albeit also the case that their overall frequency declines. A major distinction between Renaissance and Classical French and all stages of French before them is that neither are full V2 systems, although we did suggest that finite verbs might optionally target Fin, rather than T, in some 16th- and early 17th-century texts. Inversion with nominals, even in interrogatives, is exceptionally rare from the 16th century onwards, and occurs near exclusively with pronominal subjects; we suggested that in these cases the finite verb targets Subj. Similarly to residual V2, some cases of scrambling remain in Renaissance French texts in particular, but are not a widespread word-order phenomenon. Although null subjects are near entirely absent in Classical French, residual null subjects are found in Renaissance and Early Classical French texts when licensed by particular features of the finite verb, related to person or clause-type. The general absence of sharp matrix/embedded asymmetries from the 16th century onwards led us to hypothesize that subordinators are canonically base-generated in Force in Renaissance, Classical, and Modern French, accounting for the lack of recomplementation structures and non-availability of null complementizers. Modern French is subject to extensive sociolinguistic and regional variation affecting the majority of phenomena discussed in this book. This said, one major point of continuity across the Francophone world today is systematic V-to-T movement and the presence of an EPP-feature on T, leading to a dedicated subject position within the T-layer. Verb-movement properties beyond matrix declaratives are, however, more variable with V-to-Subj movement in interrogatives characteristic of formal language and certain varieties of Quebec, with V-in-situ interrogative structures overwhelmingly preferred in colloquial language. This general absence of V-to-Subj movement also extends

246

a new perspective on syntactic change in french

to other inversion contexts, such as adverbial inversion, which is generally characteristic of more formal varieties. When nominal inversion is licensed, postverbal subjects systematically target a position at the vP-periphery. As we saw in §6.4, with the notable exception of Alsatian French, generalized scrambling is not a feature of any Modern French variety, although certain quantifiers can still target a vP-peripheral position. Table 7.1 summarizes these major findings, as set out in Chapters 3–6.

7.2 The history of French and parametric theory Modern and historical French data have had a significant role in shaping parametric theory and theories of syntactic change since the very earliest days of generative grammar.1 The data in Table 7.1 quite clearly show why this is the case, highlighting the fact that within its attested textual history French has occupied a wide range of points on the null-argument, verb-movement, and word-order typologies, whilst also showing rich variation in the syntax– pragmatics mapping and the syntax of subjects—both areas of considerable linguistic diversity in Romance and beyond. In our overview of grammatical change in French in Chapter 2, we noted from the outset the problematic nature of broad-brush labels for characterizing heterogeneous phenomena within the language’s history, even if approached from a purely descriptive perspective.2 In keeping with the general rethinking of the parametric enterprise and how to formally account for the ‘space’ of morphosyntactic variation which has occurred in recent years, it has become increasingly clear that labels such as ‘null subject’, ‘OV’, ‘V2’, and ‘flexible word order’ have the potential to mask a considerable body of variation.3 The data presented so far in this book also undermine the ultimate value of such labels when applied to the history of Latin and French without appropriate caveats. 1 See, amongst many others, Kayne (1975; 1983a; 1983b; 1989; 1991; 1993), Cinque (1982; 1982; 1990), Rizzi (1986a; 1990; 1996; 1997), Pollock (1989), Belletti (1990), and Zanuttini (1997) on Modern French, and Vanelli, Renzi, and Benincà (1986), Adams (1987a; 1987b; 1987c; 1988), Vance (1987; 1988; 1993; 1995; 1997), Roberts (1993), and Benincà (1995) on Old French. 2 Notably, some of the earliest works on the history of French, such as Foulet (1919) and Brunot (1905–1938), show a sensitivity to diachronic and intratextual variation which is not always apparent in later descriptive or formally oriented works. 3 On the null-subject property, see in particular Biberauer et al. (2010), Roberts (2010c; 2019: Chapter 3), Camacho (2013), Duguine (2014), Cognola and Casalicchio (2018), and Barbosa (2019). Recent contributions on the variation found within V2 systems are found in Lohnstein and Tsiknakis (2019), Meklenborg Salvesen (2019), and Woods and Wolfe (2020). See also Sabel and Saito (2005), Biberauer (2008), Biberauer and Sheehan (2013), and Sheehan et al. (2017) on the OV property specifically and on word-order variation more generally.

Table 7.1 Syntactic change from Latin to French Unmarked Matrix Verb Movement

Marked Matrix Verb Movement

Unmarked Locus Embedof EPP ded Verb Effects Movement

Subordinators Focus

Topics

TPSubjects

vPSubjects

Null Arguments

OV Orders

Archaic Latin

V

V

V

Fin

Yes

Yes

Yes

Focal

Generalized

Rollup

Classical Latin

v

v

v

Fin

Yes

Yes

Yes

Focal

Generalized

Scrambling to vPPeriphery

Late Latin

Fin

T

v

Fin(-toForce)

Yes

Yes

Yes

Focal

Generalized

Early Old French

Fin

Fin, Foc, Top, Force Fin, Foc, Top, Force Foc, Top, Force Foc, Top, Force

T

Fin

Fin(-toForce)

Yes

Yes

Later Old French

Force

Force

T

Force

Fin(-toForce)

CFoc, QP

Yes

Accessible, Focal Active, Pronominal Active, Yes Pronominal

Scrambling to vPPeriphery Scrambling to vPPeriphery

Full Null Subject Partial Null Subject

Scrambling to vPPeriphery (Principally vPTopic ) Continued

Table 7.1 Continued Unmarked Matrix Verb Movement

Marked Matrix Verb Movement

Unmarked Locus of EPP Embedded Verb Effects Movement

Subordinators Focus

Topics

TPSubjects

Fin

Foc, Top, Force

T

Fin

Force

CFoc, QP

Yes

Renaissance T French

Fin, Foc, Top, Force

T

T

Force

Classical French

T

Subj, Force

T

T

Formal Modern French

T

Subj, Force

T

Colloquial Modern French

T

Force

T

Middle French

vPSubjects

Null Arguments

OV Orders

Active, Yes Pronominal

Partial Null Subject

Highly Restricted Scrambling to vPPeriphery (Principally vPTopic )

QP, AdjPs, AdvPs, PPs

Active/Base- Pronominal Yes Generated

Partial Null Subject

Highly Restricted Scrambling to vPPeriphery (Principally vPTopic )

Force

QP, AdjPs, AdvPs, PPs

Active/Base- Pronominal Yes Generated

(Very) Partial Null Subject

Highly Restricted Scrambling to vPPeriphery (Principally vPTopic )

T

Force

Lexically Restricted

BaseGenerated

Pronominal Yes

(Very) Partial Null Subject

Lexically Restricted Q-Fronting

T

Force

No

BaseGenerated

No

No

Lexically Restricted Q-Fronting

No

7.2 the history of french and parametric theory

249

To consider one example, we could adopt a maximalist definition of the V2 property under which a C-head is ‘strong’ in the sense of Guardiano and Longobardi (2005) and Ledgeway (2020: 73) and obligatorily attracts a finite verb or auxiliary.⁴ Under the conservative assumption that obligatory V-to-Fin movement had taken hold by at least the time of the Peregrinatio (Salvi 2004; Ledgeway 2017b) and is a systematic feature of the grammar of French until approximately 1500 (cf. §§4.4–4.5), French would be ‘V2’ for a millennium. However, whilst there is value in this observation, it fails to capture—amongst other properties—the fact that the Edge Feature associated with ‘full’ V2 only emerges in the period between the latest Latin texts and their earliest French/Gallo-Romance counterparts, and the fact that the rise and fall of a strict Force-V2 system leads to a markedly distinct left-peripheral syntax within this period. Similar considerations apply to null arguments: some variant of pro is licensed from the very earliest Latin texts through to modern varieties of French which feature pronominal inversion. Nevertheless, the discussion in Chapter 5 shows quite clearly that the precise conditions under which null arguments are licensed varies to a considerable degree. The basic conclusion is that macro-typological generalizations must sit alongside more fine-grained meso-, micro-, and nanotypological analyses of the phenomena under consideration. The approach to parametric variation pursued in Biberauer and Roberts (2012; 2014; 2015; 2017) and Roberts (2014a; 2014b; 2019) is well suited to modelling both the big-picture generalizations emerging from our analysis of the Latin and French data as well as the more nuanced analyses required by the data (Biberauer and Roberts 2012: 268): (1)

For a given value vi of a parametrically variant feature F: a. Macroparameters: all heads of the relevant type, e.g. all probes, all phase heads, etc., share vi ; b. Mesoparameters: all heads of a given natural class, e.g. [+V] or a core functional category, share vi ; c. Microparameters: a small, lexically definable subclass of functional heads (e.g. modal auxiliaries, subject clitics) share vi ; d. Nanoparameters: one or more individual lexical items is/are specified or share vi .

⁴ This is a looser interpretation of V2 than has been adopted elsewhere in this book, where— following Cardinaletti and Roberts (2002), Holmberg (2015), and Wolfe (2018a; 2019)—it is assumed that a phrasal movement diacritic must also be present on a C-related head for a language to count as ‘full’ V2.

250

a new perspective on syntactic change in french

Much of the variation discussed in this book sits at the meso-, micro-, and nanoparametric level. However, the shift from a reconstructed harmonically head-final stage—possibly instantiated by Proto-Indo-European (Lehmann 1974)—to near-harmonic head-initiality in Modern French is a significant macroparametric shift affecting not just French but all the major Romance languages (Ledgeway 2012a). The V2 property is a particularly important mesoparametric property: on the one hand it does not characterize all probes or phase-heads but it is arguably more significant than a property affecting a single core functional category—having a bearing on the featural makeup of v, Fin, and—in Later Old French—Force, all of which are likely candidates for clausal phase-heads in a cartographic perspective, as we suggested in §4.6. It is also worth noting that although mesoparametric properties often partially characterize a genus (Roberts 2019: 77), V2 is more widespread than this in Early Indo-European, with V2 effects of some form attested in Germanic (Eythorsson 1995; Ferraresi 2005; Walkden 2014), Celtic (Carnie, Pyatt, and Harley 1994; Willis 2007; Willis 1998; Meelen 2020), and Anatolian (Garrett 1992; Luraghi 1998; Sideltsev and Molina 2015). Looking exclusively at the history of French, a grammar where Fin, or Fin and Force, attract v shows striking stability, being characteristic of late Latin grammar and French until the beginning of the 16th century, thus approximately 1000 years. With this in mind, we can consider one of the most significant splits in the history of French as between one period where clausal phase-heads probe for finite verbs/auxiliaries and are thus the main locus of ϕ-agreement, and a second period—lasting throughout the second half of the last millennium to the present day—where finite verbs are attracted to T (cf. discussion in §4.5). Note that the notion that the V-to-T typology should be kept quite distinct from the V-to-C typology is underscored in Schifano (2018: Chapter 1). Given that the locus of verb movement also correlates—with the possible exception of late Latin—with the locus of EPP-effects, triggered by the presence of an Edge Feature on the relevant head, we can informally state the most significant mesoparametric split discussed in this book, along the following lines:⁵ (2)

Mesoparametric Cluster A: uV feature and EF associated with phase-heads. Late Latin > Middle French

⁵ Note that Cluster A could be conceived of as showing greater stability if we considered the innovative Classical Latin grammar as also belonging in this category, where v alone attracts V and is the locus of EPP effects (cf. discussions in §4.2.2 and §6.2.1).

7.2 the history of french and parametric theory

251

Mesoparametric Cluster B: uV feature and EF associated with heads in the TAM layer. Renaissance French > Colloquial French The majority of other case studies considered hold uncontroversially at the meso- or microparametric level and concern the probing features of a functional head or subclass of heads: C-heads—Force, Top, Foc, Fin—in our discussion of the left periphery and T/V-related heads—Subj, T, and v—in our discussion of verb movement. Although the features associated with Top and Foc figured in our discussion, T, vTopic , vFocus , and v are central in conditioning variation and change within the subject system (cf. §5.7). The featural makeup of the lower three of these heads also determines the conditions under which rollup of objects and scrambling occurs in the history of Latin and French. We can initially characterize some of the phenomena discussed in this book as in Table 7.2. The intention in this section is not to repeat the analysis from previous chapters but rather to consider parallels that emerge across the phenomena considered. Many of the case studies here concern the evolution of movement triggers on functional heads and close analysis shows that there are parallels in how these features change over time; this is the case, for example, with the progression of XP-movement to the CP and vP peripheries, which both show a similar trajectory in that a generalized trigger for movement [+EF] is reanalysed as one which is information-structurally marked (e.g. uFoc, uTop), with this information-structurally marked trigger being reanalysed as a syntactically marked trigger (e.g. uQ), which may hold across all XPs of a particular class or a lexically defined subclass thereof. In the case of XP-movement to the CP-periphery, any form of movement is lost entirely in some French varieties (cf. §3.5), whereas quantifier-fronting appears stable at the vP-edge (cf. Table 7.3). Considering a slightly different type of operation involving Agree, null arguments also show a similar trajectory, with the ability for all kinds of arguments to delete under probe-goal agreement with v (cf. Roberts 2010b) being reanalysed as an ability for only topical subject pronouns to delete under agreement with Top (cf. §5.7.3) in Early and Later Old French. From Middle French onwards, we then see purely syntactic features—[+Wh, +2pl] etc.— determining the conditions under which pro can delete under agreement with T, before null subjects are eventually lost from the system entirely except in the varieties of French where pronominal inversion is a productive syntactic process (Roberts 2010a). Finally, subject-movement to Spec-TP from Early

252

a new perspective on syntactic change in french

Old French onwards follows a similar trajectory, with a generalized movement trigger for all subjects being reanalysed as one where only [+Active] subjects can satisfy the movement diacritic before this becomes a purely syntactic trigger for [+Pronominal] subjects, after which such movement is lost entirely in the most ‘advanced’ French varieties (Zribi-Hertz 1994). A question naturally arises looking at these data as to why previously generalized agreement and movement processes pass through an intermediate phase of being conditioned by information structure before ultimately being lost. The answer may be relatively simple, that reanalysing the output of a grammar with a generalized trigger for movement or Agree as compatible Table 7.2 Parametric typology in the history of French Meso

Micro

Nano

All Left-Peripheral Focus Heads Active (CFoc, IFoc, Q) Latin, Early Old French

Subclass of Left-Peripheral Focus Heads Active (CFoc, Q) Later Old French Single Left-Peripheral Focus Head Active (Q) Middle French

Foc Probes a Defined Syntactic Subclass of Foci (PPs, AdjPs, AdvPs) Renaissance and Classical French Foc Probes a Lexically Defined Subclass of Foci (Certain PPs, AdjPs, AdvPs) Modern French

All Classes of Objects Rollup to Spec-VP Archaic Latin

vTopic , vFocus Probe +Topic/Focus Complements Late Latin, Early and Later Old French

vFocus Probes a Defined Subclass of Complements (Certain Qs) Renaissance, Classical, Modern and Colloquial French

Unmarked V-to-C Movement Late Latin, Early Old French, Later Old French, Middle French

V-to-Force in Imperatives Latin, Old, Middle, Renaissance, Classical, Modern and Colloquial French

V-to-Subj Licensed for Subclass of Lexical Verbs Only Renaissance and Classical French

Unmarked V-to-T Movement Renaissance, Classical, and Modern French

T-to-Subj Licensed, Not V-to-Subj Modern French

Generalized Null Arguments Latin

Partial Null Subjects Later Old French, Renaissance and Classical French

Full Null Subjects Early Old French

(Very) Partial Null Subjects Modern French

7.2 the history of french and parametric theory

253

with one where the process is conditioned by information-structural factors is likely to be compatible with a greater proportion of the input than one where the movement trigger is immediately reanalysed as only being associated with a small syntactically determined class of XP. In the former case, given the frequent ambiguity that exists in encoding information-structural values between speaker and hearer, fewer pieces of primary linguistic data are likely to act as ‘blocks’ to an innovative grammar in a probabilistic learning model such as that of Yang (2003; 2004) than would be the case in the latter scenario. The hypothesis on the basis of the French and Latin data, therefore, is that movement from a mesoparametric option to a micro- or nanoparametric one Table 7.3 XP movement in the history of French

LeftPeripheral XP Movement

vPPeripheral XP Movement

Generalized

InformationStructurally Marked

Syntactically Marked

Lexically Defined

Focal/Topical XPs Obligatorily Move Early Old French

CFoc/QPs/Topical XPs Obligatorily Move Later Old French

PPs, AdjPs, and AdvPs Optionally Move Classical French

Subclass of PPs, AdjPs, and AdvPs Optionally Move Modern French

Generalized Object Rollup Archaic Latin

Focal/Topical XPs Obligatorily Move6 Classical and Late Latin, Early Old French

QPs/Topical XPs Obligatorily Move Middle French QPs/DiscourseActive XPs Optionally Move Renaissance French

Focal/DiscourseActive XPs Obligatorily Move Later Old French 6

Subclass of Qs Obligatorily Move, Subclass of Qs Optionally Move Modern and Colloquial French

Recall that certain varieties of Alsace French also show this form of object scrambling (Wolf 1983; Wolf 2000), the maintenance of which is probably due to contact.

254

a new perspective on syntactic change in french

is frequently mediated by a reanalysis in which an Agree operation is associated with information-structural features (cf. i).⁶ (i) Loss of Agree Operations on Functional Heads Generalized Probe (e.g. +EF) > Information Structure-Conditioned Probe (e.g. uTop, uFoc) > Syntactically Conditioned Probe (e.g. uQ, uD) > No Agree Operation Two factors play an especially important role in conditioning the outcomes of change in a variety of the case studies considered in this book: Feature Economy (Roberts and Roussou 2002: 301; Van Gelderen 2008a: 297; Van Gelderen 2009a: 93; Biberauer and Roberts 2015: 301) and Input Generalization (Roberts 2007a: 275). Feature Economy is pervasive in large number of reanalyses we have discussed in the history of Latin and French. Firstly, in the genesis of V2 we witness the reanalysis of a number of individual verb/auxiliary movement triggers associated with distinct syntactic or pragmatic features—i.e. Wh, Top, Foc—as a single uV feature on Fin (cf. §4.2). Considering the much later emergence of a Force-V2 grammar, an even greater range of individual verb movement-triggering features are subsumed within the single uV feature now associated with Force, which we noted in §4.2.2 could attract the finite verb in imperatives and polar questions until this point (see for discussion also Rouveret 2004; Labelle and Hirschbu¨hler 2005).⁷ (3)

a. Reanalysis 1: Top [uTop], Foc [uFoc, uWh], Fin [uT] > Fin [uV] b. Reanalysis 2: Force [uPol, uQ], Fin [uV] > Force [uV]

The same process is at play in the genesis of the V2-related Edge Feature, which replaces previous distinct movement diacritics which we saw in §3.2.2 and §3.2.3 were associated with discourse-marked topicalization and focalization in Classical and late Latin (3a).⁸ (4)

Top [+EF], Foc [+EF], Fin > Fin [+EF]

In the eventual genesis of an unmarked object position within the VP, we can also see Feature Economy at play: our proposal in §§6.2–6.4 is that movement to the vP-periphery in Classical and late Latin and Old French is conditioned ⁶ See also Biberauer and Roberts (2008) for a similar trajectory of change in the loss of OV in the history of English. ⁷ We can also add Fin here as a head which can attract the finite auxiliary of copula form of esse ‘be’ in the original Classical Latin grammar (cf. §3.2.2 and §4.2.2 and Adams 1994b; Adams 1994a). ⁸ See also Biberauer and Roberts (2015), who make precisely this observation about Feature Economy driving the innovation of V2 in Germanic.

7.2 the history of french and parametric theory

255

by whether particular information-structural features are present on the verb’s complement, with the precise makeup of the features which the probe is sensitive to subject to diachronic change (cf. also Zaring 2011). However—with only a small number of exceptions (cf. §6.4)—from the Middle French period onwards, v no longer probes for [+Focus, +Topic] complements and also no longer bears a corresponding Edge Feature which attracts them to specifiers in the vP topic-focus field. What we observe in this domain, therefore, is straightforward feature loss, where information-structural features and a corresponding movement diacritic are lost entirely. Both synchronic (Cardinaletti and Starke 1999) and diachronic approaches (Roberts and Roussou 2002: 174–186; Van Gelderen 2009a: Chapter 2; Roberts 2010c: 119) have stressed the role of Feature Economy in the evolution of subject pronouns, clitics, and agreement markers in French and Romance. In Chapter 5 we noted that French subject pronouns have seen a progressive loss of phonological and morphosyntactic autonomy since at least the 15th century, which can straightforwardly be mapped on to the pronouns’ loss of their DP-layer and associated features through reanalysis as φPs, and then further reanalysis as φ-agreement heads in a number of ‘advanced’ and colloquial French varieties of the type discussed by Zribi-Hertz (1994) and Culbertson (2010): (5)

DP >

φP >

φ

If we see the ultimate culmination of Feature Economy as being the loss of any movement-triggering feature at all, one of the most striking characteristics of the data from contemporary French varieties is the sheer extent to which movement operations have been lost across multiple areas of the clause, either in favour of a grammar where a head or XP is base-generated in a lower position in the functional structure or in favour of one where it is base-generated in the position originally targeted by movement;⁹ amassing data in support of this generalization is not challenging, as we see in the following domains which have been discussed so far in this book: (i) The loss of CP-Information Focus in all varieties of French. (ii) The loss of CP-Contrastive Focus in all but the most conservative varieties of French. (iii) The loss of QP-fronting in all varieties of French. (iv) The loss of topicalization via movement in colloquial French. ⁹ See Roberts and Roussou (2002), Van Gelderen (2008a; 2008b; 2009b; 2009a; 2011; 2016), and Roberts (2012b) for extensive case studies of this type in the domain of grammaticalization.

256

a new perspective on syntactic change in french

(v) The loss of Fin-to-Force movement of complementizers, and thus recomplementation, and C- deletion in the majority of French varieties. (vi) The loss of Wh-movement in many colloquial French varieties. (vii) The loss of declarative V-to-C movement from Renaissance French onwards. (viii) The loss of interrogative V-to-C movement from Renaissance French onwards. (ix) The loss of V-to-Subj movement in interrogatives in colloquial French. (x) The loss of subject-movement to the vP-periphery in colloquial French. (xi) The loss of pronominal subject-movement to Spec-TP in colloquial French. (xii) The loss of object-movement to the vP-periphery in all varieties of French. The overall picture, therefore, is that a generalized loss of movement operations is characteristic of all French varieties since the Renaissance, but that the most ‘advanced’ (cf. Zribi-Hertz 1994) Colloquial French varieties spoken today have lost movement-triggering diacritics from the vast majority of their functional heads, essentially maintaining in declaratives only V-to-T movement and movement of nominal subjects to Spec-SubjP. Turning to a different—at times complementary—factor, we can consider the role of Input Generalization; this is undoubtedly significant in the emergence of Mesoparametric Cluster A in (2), where we see an Edge Feature and uV feature being progressively associated with v alone (‘innovative’ Classical Latin), v and Fin (late Latin, Early Old French, Middle French), and v, Fin, and Force (Later Old French).1⁰ In the first instance, acquirers are extending the properties of one verbal head to another associated with verbal finiteness, Fin. In the second reanalysis, which we saw occurred between 1180 and 1200, these features are also assigned to Force, with the result that in the grammar of Later Old French all clausal phase-heads are [+EF, uV] (cf. 6): (6) [ForceP

Force

…

[FinP

Fin

…

[νP

1⁰ See Wolfe (2016a: 489–490) for additional discussion of this point.

ν

]]]

7.2 the history of french and parametric theory

257

Importantly, at the point at which Force-V2 is operative, French shows the most marked matrix/embedded asymmetries, which we interpreted in §3.3.5 and §3.6.2 as evidence that embedded Fin-to-Force movement occurs near systematically in this period. Given that the strictest asymmetries emerge in parallel with Force-V2, there is reason to believe that Input Generalization is at play here too: acquirers are extending the featural composition of matrix Force to embedded Force in that both heads must obligatorily probe and attract the contents of Fin; in the case of matrix Force, this leads to finite-verb movement; in the case of embedded Force, this leads to movement of the complementizer que ‘that’.11 This has the very desirable effect of making the properties of embedded clauses accessible to acquirers purely from matrix-clause primary linguistic data, thus overcoming a significant objection to the classic symmetrical/asymmetric V2 typology, put forward by Walkden and Booth (2020), that certain systems may be unlearnable under standard assumptions on the nature of degree-zero learnability (Lightfoot 1989), if evidence for parametric values associated with them is found purely in the embedded clauses themselves. Unsurprisingly, given the role of Input Generalization in conditioning featural parallels between v, Fin, and Force in other domains, we also see its role in determining the information-structural values of initial constituents in V2 clauses and scrambled complements in the low vP-periphery. Recall from our discussion in §3.3 that the preverbal field in Later Old French and Middle French becomes increasingly specialized in hosting topical, discourseactive constituents with a concomitant fall in the licensing of initial foci. As noted in §6.5, we also see a parallel change at the vP-edge, where scrambled complements—in contrast to Early Old French—show an increasing tendency to be pronominal or to be discourse-active. Combining Roberts’ (2007a) Input Generalization Principle with Poletto’s (2014) Parallelism of Phases, we see that generalization of featural values is not only restricted to features traditionally conceived of as purely syntactic—e.g. Edge Features and verbmovement triggers—but also extends to topic- and focus-related features. Furthermore, we see the effects of Input Generalization beyond phase-edges: TP-internal subjects undergo a parallel reanalysis to XPs satisfying V2 and complements undergoing scrambling in Later Old French, as they are increasingly discourse-active or pronominal (see discussion in §5.3.2); this leads to a grammar where information-structural features associated with XP-movement show striking parallels on Fin, T, and v. 11 For similar, though not identical, ideas that embedded Force probes Fin in asymmetric V2 systems, see Branigan (2011: 89–107).

258

a new perspective on syntactic change in french

Overall, this section aims to show—alongside discussion in Chapters 3–7— that the empirically fine-grained picture emerging from the analysis in this book warrants rethinking in the context of recent advances in parametric theory. Alongside the central role of Feature Economy and Input Generalization, we witness clear commonalities in the direction of change in keeping with the macro-, meso-, and nanoparametric progression along parametric hierarchies diachronically. It has also been proposed that information-structural triggers for movement can play an important role in mediating the shift between a meso- and micro- or micro- and nanoparametric point on the typology.

7.3 Beyond clausal word order In Chapter 2 we reviewed a variety of approaches to change in French morphosyntax and critiqued a number of them for failing to offer a formal account of explanation for why a particular type of change occurred in one instance, or across domains. If the proposals around parametric change in the history of Latin and French put forward in §7.2 are correct, therefore, we expect some aspects of the formal account to be able to extend to other phenomena. In this section, we will explore in a preliminary fashion the extent to which this is true for changes affecting the nominal expression and the system of negation.

7.3.1 Change in the nominal expression A priori we expect to find parallels between changes affecting clausal structure and those affecting the nominal domain: since Abney’s (1987) seminal work on the DP hypothesis parallels between the functional structure of the CP and the DP have been explored, both C and D are standardly assumed to be phase-heads in recent theorizing, and, in Medieval Romance scholarship, Poletto’s (2006b; 2014; 2015) work on Old Italian has shown that the CP-, vP-, and DP-edges show similarities in their featural makeup, as we have already noted elsewhere in this book.12 So far, however, potential diachronic parallels in these domains have not been discussed for French. From the outset, we should note that two of the most salient typological properties of Modern French in a comparative Romance context do indeed show striking parallels: we have already seen that finite-verb movement to a C-related head is restricted in Modern French and near entirely absent 12 But see Bošković (2005; 2008; 2009) for the claim that T, not C, is the counterpart of D in the clausal domain.

7.3 beyond clausal word order

259

in most colloquial varieties, with absence of inversion viewed as a defining feature of ‘advanced’ French by Zribi-Hertz (1994) and a number of others (Ashby 1977b; Blanche-Benveniste 1990: 39; Coveney 1996: 93–100; Rowlett 2007: 199–207; Elsig 2009; Auger and Villeneuve 2019: 214; Tailleur 2013). Although our discussion and analysis in Chapter 4 showed that verbmovement parameters are subject to extensive diachronic and synchronic micro- and nanovariation, the broad generalization from approximately 1500 onwards is that the decline in V-to-C movement is one of the most significant factors characterizing the evolution of French. Crucially, however, we also noted in §2.2 that one of the major changes to affect the nominal system from Latin and Old French onwards has been the decline in bare NPs, which are prevalent in Latin and Old French (Foulet 1919: 47; Moignet 1973: 102– 106; Harris 1978: 74–78; Posner 1997: 382–384; Buridant 2000: 102–112), but highly restricted in the modern language (Carlier 2004; Carlier et al. 2020a). A commonly accepted analysis put forward by Longobardi (1994)—building on insights into the nature of nominal head movement in Cinque (1993; 1994)— is that the DP layer is not necessarily absent in cases of apparently bare NPs but that the head noun itself may undergo movement to D (7). The extent to which bare NPs will be licensed in a given language, therefore, will depend on the extent to which the grammar permits N-to-D movement. Viewed in this light, there is a clear parallelism between the grammar of the Old French DP and CP, with bare NPs being one of the most salient points of Old French nominal syntax (8) and V2 clauses one of the most salient points of clausal syntax (9): (7)

[DP [D Gianni][NP [N Gianni]]] parlava Gianni speak.3sg.pst ‘Gianni was speaking’

(Modern Italian)

(8)

a.

[DP [D Miel][NP [N miel]]] … est caus et sès … honey be.3sg hot and dry ‘Honey is hot and dry …’ (OFr. Le régime du corps 159,13th century, Carlier 2001: 66) b. [DP [D Le][NP [N miel]]] … est chaut et sec … the honey be.3sg hot and dry ‘Honey is hot and dry …’ (ModFr. Le régime du corps 159, Carlier 2001: 66)

(9)

a. [ForceP Miel [Force est][TP … [vP [v est] … caus et sès]]] b. [ForceP … [SubjP Le miel [TP [T est] [vP [v est] … chaut et sec]]]]

260

a new perspective on syntactic change in french

Whilst this is a potentially appealing generalization that fits with our previous observations on parallels in the featural makeup of Force, Fin, and v, we should ask if the parallels extend further to other properties of the DP. I leave testing this hypothesis to its full extent to future work but note two important points here. Firstly, if we wish to explore parallels between the properties associated with verb movement and N-movement, we should recall that it is not the case that verb movement of any kind has been lost in Modern French. Rather, as explored in detail in §4.5, V-to-T movement has been a stable mesoparametric property of French since the 16th century and remains consistent across regional varieties today. As such, finite verbs do not remain in their base-generated position within the extended vP-complex but move to a high position within the tense-aspect-mood field, as demonstrated by Schifano (2015a; 2018) among others. Revealingly, French nouns also appear to target a position within the inflectional middlefield of the extended DP complex, which has been identified as Num(ber) by a number of scholars (Bernstein 1991: 110–113; Valois 1991: 374–375; Cinque 1994). The arguments for this analysis—set out in most detail in Cinque (1994)—are based on the assumption that adjectives preceding French nouns occupy a fixed position higher than Num in the nominal expression and that those following the noun occur in a position lower than Num such as in (10). If this line of analysis is correct,13 French nouns have not only lost N-to-D movement in tandem with V-to-C movement, they also now feature N-to-Num movement in a parallel fashion to V-to-T movement. (10)

[DP [D un] [NumP [Num livre] [AP rouge [NP [N livre]]]]] a book red ‘A red book’

(ModFr.)

Secondly, our discussion in §3.3.1 reveals that the Old French V2 property has two components: in addition to an obligatory requirement for head movement of the finite verb/auxiliary to a C-related head, a C-related head also bears an Edge Feature which requires merger of a phrasal constituent, typically a preverbal topic or focus.1⁴ As a consequence, in the loss of V2 effects the history of French attests a concomitant decline in triggers for finite V-to-C movement and triggers for topicalization or focalization via movement (cf. §3.6). In a parallel fashion to proposals for the fine structure of the clausal left periphery 13 See Laenzlinger (2005) for an alternative view on DP-internal movement in French. 1⁴ See also Vikner (1995), Cardinaletti and Roberts (2002), Holmberg (2015), and Wolfe and Woods (2020) for similar formulations of the V2 constraint in featural terms.

7.3 beyond clausal word order

261

(Rizzi 1997; Benincà 2001; Benincà and Poletto 2004; Rizzi and Bocci 2017), Giusti (1996; 2006; 2014; 2015) has suggested that the DP also has a rich leftperipheral structure, sensitive to discourse-pragmatic properties. As such, it is highly significant that in Old French the range of adjectives that could precede the noun is of a considerably larger class than exists in the modern language (cf. 11) (see among others Posner 1997: 361; Buridant 2000: 207–215; Carlier et al. 2020b: 996–997), a property that we can thus associate with the activation of the nominal left periphery, to which these APs can undergo movement (see in particular Boucher 2004; 2005): (11)

a.

Un gris a grey ‘A grey sheep’ b. Un mouton a sheep ‘A grey sheep’

motun sheep (OFr. Guillaume 398, Buridant 2000: 171) gris grey (ModFr.)

Indeed, in an exact parallel of the data that we observed for clausal leftperipheral movement in Modern French in §3.5, only a lexically defined subclass of APs can undergo such movement today, such as bon ‘good’, mauvais ‘bad’, vrai ‘true’, and grand ‘big’.1⁵, 1⁶ (12)

Une grande maison a big house ‘A big house’

(ModFr.)

Overall, therefore, there are striking parallels between the types of movement operations licensed in the DP and those licensed in the CP. These are salient in both a synchronic comparison of the two domains in Old French or Modern French and a diachronic investigation of how both domains have changed syntactically. Although a number of issues warrant exploration in further research, it appears that Poletto’s (2006b; 2014) Parallelism of Phases hypothesis also extends to the French DP.

1⁵ For a recent attempt at characterizing this class semantically, see Carlier et al. (2020a: 997). See also Smith (2016: 309). 1⁶ I leave aside here the greater range of APs which may be preposed in parts of Belgium and Normandy. Thanks to J. C. Smith for discussion on this point.

262

a new perspective on syntactic change in french

7.3.2 Change in the negative system In §2.3, it was noted that Jespersen’s Cycle has some descriptive value when applied to the history of French, but does not necessarily offer an explanatory account, even if the evolution of French negation shows striking parallels with change in the negative system of other Romance and non-Romance varieties (see for a recent overview Willis, Lucas, and Breitbarth 2013). Given the truly vast formal literature that has emerged on the evolution of negation in recent years,1⁷ I do not attempt here to offer a full motivation for the factors driving Jespersen’s Cycle in French (cf. 13) but put forward several observations on potential factors driving change which are linked to analysis already presented in this book, with these observations intended as an invitation to future research. (13)

Stage One Stage Two Stage Three Stage Four Stage Five

Neg Neg Neg (Neg)

V V (Neg) V Neg V Neg V Neg

Breitbarth, Lucas, and Willis (2020: 77–91) propose a detailed ‘pull chain’ account of the evolution of Jespersen’s Cycle under which changes in the properties of the postverbal element are the principal trigger for changes in its preverbal counterpart. As we have already noted in §2.3, such an approach, which is not contingent on weakening of the preverbal negator ne, is also desirable for French specifically, where ne only becomes unstressed centuries after the first postverbal reinforcers are attested (Martineau and Mougeon 2003: 123–124) and can also serve as monopartite negation in formal written registers to this day (Hansen 2013: 54). For Jespersen’s Cycle to take hold, therefore, two reanalyses are required to drive change: (i) a nominal minimizer must undergo reanalysis as a functional emphasizer,1⁸ and (ii) this functional emphasizer must lose its emphatic value in order to become a negative-polarity

1⁷ For detailed formal analysis of the French data, see Rowlett (1998), Roberts and Roussou (2002), Detges and Waltereit (2002), Déprez (2003), Ingham (2012; 2014), De Clercq (2017), and wide-ranging discussion in Breitbarth, Lucas, and Willis (2020). 1⁸ Old French featured a huge number of such minimizers, which are originally assumed to have co-occurred only with semantically compatible verbs, e.g. ne mangier mie ‘to not eat a crumb’. For discussion of their distribution, see in particular Price (1962), Detges (2003), and Hansen (2020: 1682– 1683). For broader cross-linguistic discussion, see Tubau (2020).

7.3 beyond clausal word order

263

item.1⁹ Although exact implementations vary, many in the literature have assumed that for an unmarked postverbal negator to emerge, the functional emphasizer must no longer be interpreted as encoding narrow focus, which we could interpret as a result of its either occupying a vP-peripheral focus projection (Breitbarth, Lucas, and Willis 2020: 87) or bearing a [+Focus] feature.2⁰ In the history of French specifically, dating varies as to when pas was no longer interpreted as emphatic, but there is general consensus that bipartite negation no longer encoded marked values in most texts by the 14th century, i.e. the Middle French period (Wartburg 1958: 121; Martin and Wilmet 1973: 14; Schwegler 1986: 155; Harris 1978: 26; Posner 1997: 372; Martineau and Mougeon 2003: 119).21 The question arises as to whether this significant step between Stage Two and Stage Three in (13) is linked to any seemingly independent change in the history of French. Reviewing our discussion in §5.3.2 and §6.3.2, we can tentatively suggest that this is indeed the case: by the 14th century, there is limited evidence that any of the texts license focalized XPs at the vP-periphery in a systematic fashion. The hypothesis to test in future research, therefore, is whether the loss of vP-information focus forces a reanalysis of postverbal negators as unmarked, due to the fact that any constituent occurring at the vP-periphery can no longer be interpreted as focal/emphatic. Potential supporting evidence for this proposal comes from a comparative overview of the negation typology in Romance (cf. for overview Poletto 2016 and Ledgeway and Schifano in press), with stage-one preverbal negation varieties such as European Portuguese, Spanish, and Italian licensing vP-information focus (Belletti 2004),22 in contrast to a number of GalloRomance and Northern Italo-Romance stage-three, -four, or -five varieties, where it is no longer licensed or heavily restricted (cf. among others Paoli 2010 on Northern Italo-Romance varieties). The next significant reanalysis, which gives French part of its unique typological footprint in Romance today is the onset of ne-deletion—discussed in §2.3—where the language moves progressively onwards from stage three as set out in (13) (Pohl 1975; Ashby 1976; Ashby 1981a; Ashby 2001; Coveney 1996; Rowlett 1998; Armstrong 2002; Armstrong and Smith 2002; Martineau and Mougeon 2003). As argued by a number of scholars (Roberts and Roussou 1⁹ For discussion of these two processes as they apply to French, see in particular Detges and Waltereit (2002). 2⁰ Although see De Clercq (2017) for an analysis of Modern French negation under which pas still occupies a focus position. 21 Note that Hansen (2020: 1680) offers a later dating, immediately after the Middle French period. 22 See—amongst others—Zubizarreta (1998) on Spanish, Belletti (2004) on Italian, and Costa (2008) for discussion of European Portuguese.

264

a new perspective on syntactic change in french

2002: 154–161; Roberts 2007a: 77–81; van Gelderen 2008b; De Clercq 2017: 69; Breitbarth, Lucas, and Willis 2020: 93–96), the ‘weakening’ of ne and its eventual deletion can be modelled in terms of Feature Economy, with ne originally bearing an interpretable negative or polarity feature, later an uninterpretable counterpart to these features, and eventually undergoing deletion (cf. in particular Van Gelderen 2008a: 204): (14)

Head (iNeg/Pol) > Head (uNeg/uPol) > Deletion

As such, negation is one of many domains where Feature Economy has had a substantial role in conditioning the direction of change (cf. discussion in §7.2). However, there are also two further factors which warrant additional discussion in future research. Firstly, if—as De Clercq (2017: 69) suggests—in cases of French bipartite negation, ne bears an [iPol] feature (cf. 14) and moves to a functional head between T and Fin, we may also be able to identify one of the factors which have destabilized ne in certain French varieties. Recall from our discussion in §4.5.2.2 that the majority of Colloquial French varieties no longer license V-to-Subj movement in polar interrogatives.23 In effect this means that since the 16th century, when polar interrogatives first became substantially restricted, acquirers are receiving consistently less evidence of polarity-related Agree operations associated with functional heads in the T-layer. The hypothesis to explore in future research, therefore, would be that Input Generalization might lead the acquirer to reanalyse the primary linguistic data such that no T-related heads can bear features associated with polarity. If this were the case, bipartite negation would no longer be licensed, as we see is the case in the majority of colloquial and non-standard French varieties today. Note that this tentative hypothesis might address an issue raised by Breitbarth, Lucas, and Willis (2020: 112), namely why French has remained at stage three of the schema in (13) longer than many other languages undergoing similar changes. If the link between polarity-sensitive inversion in interrogatives and ne-raising is a genuine one, the relative stability of ne-raising could be accounted for by the fact that the decline of interrogatives has been a gradual process, beginning in the Middle French period and still not completed today, as we saw in §4.5.2.2. In even more general terms, we should note that the loss of neraising in many varieties of French is yet another case of (head) movement being lost in the most ‘advanced’ varieties, which we saw in §7.2 is an extremely widespread pattern.

23 See Elsig (2009) and references in §4.5 for an overview of the relevant changes.

7.4 periodization, french, and the ‘romance club’

265

7.4 Periodization, French, and the ‘Romance Club’ At the very beginning of this book we noted how French has diverged morphosyntactically from a number of other Romance varieties, leaving it with a distinct typological footprint which sets it apart from what Posner (1996: 38–39) refers to as the ‘Romance Club’. The uniqueness of Modern French and its varieties has been noted by several scholars working in a number of distinct frameworks (Diez 1882; Pei 1949: 139; Bec 1970: 9–10; Coșeriu 1988; Barra Jover 2004; Bossong 2016: 69; Smith 2020). In this section, we consider how the data reviewed so far in this book can contribute to this debate: does French, in fact, have an ‘outsider’ status amongst the Romance languages? If so, at what point in the history of the language does this status originate? Starting our discussion with Modern French, we can consider a number of the most significant properties discussed in this book. Consider first the left periphery: topicalization structures derived via movement are widely licensed across the majority of Romance varieties, despite being highly restricted in all but the most formal French today (cf. §§3.4–3.5). Romance varieties licensing topicalization via movement today include many Occitan (Sauzet 1989; Faure and Oliviéri 2013), Ibero-Romance (Rivero 1980; Zubizarreta 1998: 99–118; Zagona 2001: 215–227; Vallduvı´ 2002; Kato and Raposo 2007), RaetoRomance,2⁴ Italo-Romance (Benincà 2001; Benincà and Poletto 2004; Ledgeway 2010a; Ledgeway 2010b), and Sardinian varieties (Jones 1993: 144–145; Remberger 2010), and Romanian (Zafiu 2013: 570; Giurgea 2017). In the domain of focus, there is more extensive variation: some varieties project a full focus layer, such as Romanian (Motapanyane 1998; Cornilescu 2002; Zafiu 2013), Southern Italian Dialects (Cruschina 2006; Cruschina 2008; Cruschina 2012; Ledgeway 2009a; Ledgeway 2009b: 784–790), Sardinian (Mensching and Remberger 2010), and certain Ibero-Romance varieties (Jiménez-Ferna´ndez 2015b), whilst others activate only part of the focus layer, with left-peripheral contrastive focus and QPs more widely licensed than information focus in Northern Italian dialects (Paoli 2007; Paoli 2010), Catalan (Vallduvı´ 1993; Vallduvı´ 2002; Quer 2002), and many Occitan varieties (Ronjat 1913: 249–269; Gonza´lez i Planas 2009; Oliviéri and Sauzet 2016: 341). We can note, therefore, that the most formal French varieties pattern with this second ‘restricted focus’

2⁴ Note that the underlying structure of topicalization and focalization is different in the case of V2 Raeto-Romance languages, in that topical or focal XPs obligatorily move to the left periphery as a result of the V2 constraint (Haiman and Benincà 1992; Poletto 2000; Poletto 2002; Kaiser and Hack 2009; Salvi 2016: 164; Casalicchio and Cognola 2020), as we saw in §§3.3–3.4 was the case in Old and Middle French.

266

a new perspective on syntactic change in french

group, whilst most colloquial French varieties instantiate the most innovative option, whereby no left-peripheral foci are licensed. Turning to structures which evidence multiple complementizer positions in the left periphery, we observe once again that French patterns distinctly to a wide number of other Romance varieties. As noted in §3.1.1.4, dual complementizer systems are widely licensed in Southern Italian dialects (Ledgeway 2005; Ledgeway 2009a; Damonte 2010; Colasanti 2017), Sardinian (Damonte 2005), and Romanian (Dobrovie-Sorin 1994: 93–111; Nicolae 2019b: 3.2.2.3), whilst recomplementation structures are reported elsewhere, including in Northern Italian dialects (Paoli 2007), European Portuguese (and Galician) (Barbosa 2000; Mascarenhas 2014; Gupton 2014), Spanish (Fontana 1993; Uriagereka 1995; Villa-Garcı´a 2012b; Villa-Garcı´a 2012a; Corr 2016), and Catalan (Gonza´lez i Planas 2014): (15)

Ya le dije que yo, que no voy already him.cl tell.1sg that I that neg go.1sg ‘I’ve already told him that I’m not going’ (Modern Spanish, Villa-Garcı´a 2019: 3)

However, cases of recomplementation have not been reported for RaetoRomance varieties and are absent in a number of Gallo-Romance varieties, although reported for Gascon (Ledgeway 2020) and Picard (Dagnac 2012); as we saw in Chapter 3, however, recomplementation structures are entirely absent from Modern French. As such, French is not typologically exceptional within Romance in not permitting complementizers to lexicalize distinct heads in the left periphery, but patterns distinctly from the majority of Daco, Ibero- and Italo-Romance varieties in this respect. Verb movement presents a highly varied picture across Romance, as noted in §4.1 and in Schifano’s (2015a; 2018) recent comparative survey and analysis. Given this level of variation, it is somewhat more challenging to identify ‘outlier’ varieties within Romance, but we should note that French patterns only with Romanian in Schifano’s microtypology in licensing V-to-T movement as far high as the mood field (cf. Table 4.1). A distinct verb-movement phenomenon is the structure of interrogatives where certain—principally formal—varieties of French license V-to-Subj movement (Cardinaletti 2021), whilst such movement is lacking entirely in a wide number of colloquial varieties (cf. §4.5.2.2). Within Romance, Colloquial French patterns with a number of other Gallo-Romance varieties in having lost verb–subject inversion in interrogatives (e.g. Aub-Bu¨scher 1962: 58), which contrasts markedly with the behaviour of most Ibero-Romance (Ambar 1992; Escandell-Vidal

7.4 periodization, french, and the ‘romance club’

267

1999; Zagona 2001: 50; Barbosa 2001) and Southern Italo-Romance varieties (Cruschina 2012: Chapter 5), alongside Sardinian (Jones 1993: 24–26; Mensching and Remberger 2016: 286) and Romanian (Giurgea and Remberger 2012), where such movement is still widely attested: (16)

Telefonatu at Juanne telephone.ptcp has.3sg Juanne ‘Has Juanne phoned?’ (Modern Sardinian, Jones 1993: 24)

(17)

Ha cantado Juan? have.3sg sing.ptcp Juan ‘Has Juan sung?’

(Modern Spanish)

Between these two extreme ends of the interrogative typology, we find a number of transitional Northern Italo-Romance (Munaro 1999; Parry 2003; Cardinaletti 2021), Raeto-Romance (Haiman and Benincà 1992: 170–174; Siller-Runggaldier 1993; Poletto 2002; Kaiser and Hack 2009), Catalan (PlanasMorales and Villalba 2013), and Francoprovençal varieties (De Crousaz and Shlonsky 2003; Kristol 2010), which display structural restrictions on the conditions under which V-to-Subj/C movement can obtain, and have developed a range of non-movement-related strategies for marking interrogatives, including grammaticalized clefts, interrogative particles, and right dislocation (Munaro, Poletto, and Pollock 2003; Poletto and Zanuttini 2003; Poletto and Zanuttini 2010; Prieto and Rigau 2007; Kato and Ribeiro 2009; Poletto and Pollock 2009; Planas-Morales and Villalba 2013; Mioto and Lobo 2016): (18)

a.

Que plou? that rain.3sg ‘Is it raining?’ (Modern Peninsular Catalan, Prieto and Rigau 2007: 30)

b. O vindran a Ciutadella or come.3pl.fut to Ciutadella ‘Are they coming to Ciutadella?’ (Modern Balearic Catalan, Prieto and Rigau 2007: 30) (19)

O que leu o João? what that read.3sg.pst the João ‘What did João read?’ (Modern European Portuguese)

268

a new perspective on syntactic change in french

(20)

Olà pa tu vas? where part you go.2sg ‘Where are you going?’ (Modern Raeto-Romance, Pera di Fassa, Poletto 2002: 227)

In this area, Modern French varieties featuring V-to-Subj movement in interrogatives sit apart from more ‘permissive’ systems with widespread V-to-C or V-to-Subj movement, instead patterning with the restrictive class under which such movement can obtain, but with specific constraints. Colloquial French, however, stands out in this discussion as it lacks V-to-Subj movement entirely, as we saw in §4.4. Turning to null subjects, French once again patterns distinctly from the majority of Romance varieties. In contrast to many Romance varieties which are full null-subject languages, a number show constraints on the null-subject property, most notably in the domain of person-based restrictions which are found in Brazilian Portuguese (Kato and Negrão 2000; Holmberg, Nayudu, and Sheehan 2009) and some Caribbean Spanish varieties (Camacho 2013), alongside restrictions stemming from the complex interplay of the null-subject property and subject clitics in certain Northern Italo-Romance (Rizzi 1986a; Vanelli 1987; Poletto 2000; Roberts 2014b), Raeto-Romance (Anderson 2006; Poletto 2002), Occitan (Hinzelin and Kaiser 2012), and Francoprovençal varieties (Diémoz 2007; Kristol 2010; Hinzelin and Kaiser 2012: 253–262).2⁵ Whilst—following argumentation in Roberts (2010a)—varieties of French which retain pronominal inversion can be thought of as highly restricted null-subject languages, this is not the case for the numerous French varieties which do not (cf. discussion in §§5.6.1–5.6.2), and are thus consistent non-null-subject languages. We note once again, therefore, that French falls outside the ‘Romance Club’ in this respect. A final point to note is that French also patterns distinctly from a number of Romance varieties in the activation of the vP-periphery. Whilst in formal varieties of French, focus fields in the vP field can exceptionally be activated in unaccusative structures and certain other inversion environments discussed 2⁵ Whilst Occitan varieties are typically considered full null-subject languages in the literature, Hinzelin and Kaiser (2012: §2) have recently presented data from Occitan varieties spoken in Piedmont which indicate interspeaker variation relative to the licensing of null subjects in impersonal constructions (i) and with first-person referents (ii): (i) (La) semeu qu’(el) ven it.expl seem.3sg that-he.scl come.3sg ‘It seems that he’ll come’ (Occitan, Prazzo vu Val Maira) (ii) (Me) manjo un pom I eat.1sg an apple ‘I eat an apple’ (Occitan, Prazzo vu Val Maira)

7.4 periodization, french, and the ‘romance club’

269

in §5.6.2 and §6.4, such structures are far more restricted than is the case in the majority of Ibero-, (Southern) Italo-, and Daco-Romance varieties (Belletti 2004; Gallego 2007; Gallego 2013; Martins 2011: Chapter 2; Nicolae 2019b: Chapter 3) and—perhaps more strikingly—in a number of Occitan varieties (Gonza´lez i Planas 2009: §2.3.3), where postverbal subjects and other constituents can occupy the vP-periphery in a range of constructions showing VS or VXS ordering. The following examples are from Gallego (2013: 413) (21)

a.

Comeu a sopa o Paulo eat.3sg.pst the soup the Paulo ‘Paulo ate the soup’ (Modern European Portuguese)

b. A scris o carte prietana mea have.3sg write.ptcp a book friend-the my-the ‘My friend wrote the book’ (Modern Romanian) c. Leyó la carta Marı´a read.3sg.pst the letter Marı´a ‘Marı´a read the letter’

(Modern Spanish)

d. *A qui donnera le livre ton ami? to whom give.3sg.pst the book your friend ‘To whom will your friend give the book?’ (ModFr.) As a piece of corroborating evidence that in certain French varieties the vPperiphery is only activated in highly specific conditions we should note the recent conclusion by Lahousse and Lamiroy (2012: 395) on the availability of VOS orders in Romance, who note that ‘French is the most restrictive of the Romance languages under analysis’. As such, although it is not the case that the vP-periphery is entirely inactive in Modern French, its use is incredibly restricted and licensed only in formal varieties. Once again, therefore, formal French sits with a more restrictive set of Romance languages and the colloquial language is an outlier from the point of view of syntactic typology. A simplified version of the data and varieties discussed so far is found in Table 7.4. It appears, therefore, that across the major properties considered in this book, Modern French sits apart from many other Romance varieties and Colloquial French is a genuine outlier. Upon examination, the properties appear to have in common that they represent innovations from what we might think of as a common Romance syntactic norm. Examining the very earliest Romance texts, composed before 1200, we note that (cf. amongst others Wolfe 2016a: §2):

Table 7.4 French and the ‘Romance Club’

TopicsMoved

Romanian

Sardinian

SIDs

European Portuguese

Spanish

Catalan

NIDs

RaetoRomance

Occitan

Modern French

Colloquial French

+

+

+

+

+

+

±

+

±

±

−

Focus Layer

+

+

+

±

±

±

±

±

±

±

−

C-Doubling

+

+

+

+

+

+

±

−

±

−

−

V-to-C/Subj

+

+

+

+

+

±

±

±

±

±

−

Null Subjects

+

+

+

+

+

+

±

±

±

±

−

vP Periphery

+

+

+

+

+

+

±

+

±

±

−

7.4 periodization, french, and the ‘romance club’

271

(i) They permit widespread XP-movement to the CP-periphery, to both the topic and focus layers. (ii) They licensed either dual complementizer systems or recomplementation structures indicative of complementizer movement throughout the articulated C-space. (iii) They all license verb movement to a low left-peripheral head, identifiable as Fin. (iv) They license V-to-C movement uniformly in interrogatives, a feature we noted in §4.1 is inherited from Latin. (v) They are all originally full null-subject systems. (vi) All show evidence of an active vP-periphery in the form of Romanceinversion structures and scrambling phenomena. Having identified these core properties, we can then evaluate at what point in the language’s history French departs from them. We identified in §4.3.2 that between 1180 and 1200 the syntax of French undergoes a variety of changes which differentiate it quite markedly from certain Medieval Romance languages, many of which may be linked to the changing locus of verb movement to Force, a change which itself may have been conditioned by the loss of widespread CP-information focus (Wolfe 2016a; Wolfe 2018a). At this stage, French has thus diverged from the Romance norm in its focus properties, verb-movement properties, and also—recall from §5.4.3—its null-subject properties, which are now conditioned by clause-type; Wolfe (2016a) suggests similar reanalyses may have taken place in certain Northern Italo-Romance varieties and some but not all varieties of Occitan (Wolfe 2020d). This suggests that the relevant changes occur principally—though not exclusively—in the ‘northern’ Romance varieties discussed by Ledgeway (2012a: 314). Although our analysis of Middle French indicates that it is a period of change relative to Later Old French, all the points of variation discussed in §3.4, §4.4, §5.4, and §6.4 are micro- or nanoparametric in nature: topic, focus, and QP movement to the left periphery are still licensed, albeit with increasing pragmatic restrictions, as are recomplementation structures; the language is still V2, though with a distinct locus on Fin; V-to-C movement in interrogatives still obtains, although with increasing evidence of restrictions based on person, subject-type, and verb class; the language is still a partial null-subject language, though with decreasing quantitative evidence for null subjects in general; and scrambled complements and postverbal subjects can target the vP-periphery, albeit with parallel pragmatic restrictions to those observed at the CP-level. These ever-growing microparametric restrictions on a variety

272

a new perspective on syntactic change in french

of movement and Agree-based operations—discussed in detail in §7.2—lend empirical and theoretical weight to the long-established—though potentially vacuous—notion that Middle French is a ‘transitionary’ period in the history of French, if we understand ‘transitionary’ in formal terms as referring to a period attesting a gradual build-up of small-scale microparametric change which eventually serves to undermine acquirers’ evidence for at least three interlinked larger-scale properties: (i) XP-movement to all phase-edges, (ii) verb movement to phase-heads, viz. the V2 property, and (iii) the licensing of null subjects. However, Middle, Renaissance, and Classical French are not truly outside the ‘Romance Club’ in that they still license XP-movement to the left periphery of some form, retain V-to-Subj/C movement in certain contexts, license partial null subjects, and scramble a restricted range of complements to the vP periphery; in this respect they pattern with Catalan, Northern Italo-Romance, Raeto-Romance, and Occitan varieties in Table 7.4. Given the maintenance of these characteristics in—principally formal— Modern French, it is arguably Colloquial French which sits furthest outside the synchronic ‘Romance Club’ and—as we have seen—has diverged most from the common Romance norm emerging in the earliest textual records. Notably, almost all the characteristics that give French this particular profile are linked to the loss of previously widespread movement operations.2⁶ Overall, therefore, our analysis confirms the exceptional status of French syntax in a comparative Romance typological perspective, though we can conclude that its ‘outsider’ status is a relatively recent development.

2⁶ Given the overall analysis here that Modern and Colloquial French make minimal use of internal merge in syntax, I leave it to future research to establish whether this generalization could also be extended to morphology. If—as assumed in Distributed Morphology (Halle and Marantz 1993)—the building of complex inflectional structure also involves internal merge, the hypothesis would be that French would show higher analyticity than other Romance languages. Although measuring overall analyticity is extremely complex (cf. discussion in §2.2), we should note that claims along these lines have been made in the Romance literature (e.g. Smith 2020).

References Primary Sources Clari, Robert de. 1924. Conquête de Constantinople. (ed.) Philippe Lauer. Paris: Champion. http://bfm.ens-lyon.fr. Commynes, Philippe de. 1924. Mémoires. (ed.) Joseph Calmette. Paris: Belles Lettres. http://bfm.ens-lyon.fr. Curtius, Ernst Robert. 1911. Quatre Livres des Rois. Dresden: Niemeyer. http://bfm.enslyon.fr. Guillot-Barbance, Céline, Tom Rainsford, and Alexei Lavretiev (eds). 2014. Séquence de sainte Eulalie. http://bfm.ens-lyon.fr. Longnon, Jean (ed.). 1911. Chronique de Morée. Paris: Librairie de la Société de l’Histoire de France. http://bfm.ens-lyon.fr. Marchello-Nizia (ed.). 2013. Queste del Saint Graal. http://bfm.ens-lyon.fr. Moignet, Gérard (ed.). 1972. Chanson de Roland. Paris: Bordas. http://bfm.ens-lyon.fr. Monstrelet, Enguerrand de. 1862. Chronique. (ed.) Louis Douët d’Arcq. Paris: Librairie de la Société de l’Histoire de France. http://bfm.ens-lyon.fr. Montaigne, Michel de. 1580. Essais. Bordeaux: Millanges. https://classiquesgarnier.com/montaigne-corpus.html. Rabelais, François. 1922. Pantagruel. (ed.) Abel Lefranc. Paris: Champion. Racine, Jean. 1910. Phèdre. (ed.) Irving Babbitt. Boston: D. C. Heath. Rainsford, Tom and Christiane Marchello-Nizia (eds). 2018. Vie de saint Alexis. http://bfm.ens-lyon.fr. Salverda de Grave, Jean-Jacques (ed.). 1929. Eneas. Paris: Champion. http://bfm. ens-lyon.fr. Troyes, Chrétien de. 2009. Chevalier au Lion ou Yvain. (ed.) Pierre Kunstmann. http:// bfm.ens-lyon.fr. Viard, Jules (ed.). 1920. Grandes chroniques de France. Paris: Renouard. http://bfm.enslyon.fr. Voltaire. 1759. Candide, ou l’Optimisme, traduit de l’allemand de M. le docteur Ralph. Geneva: Cramer. Wickersheimer, Édith. (ed.). 1923. Roman de Jean de Paris. Paris: Champion. http:// bfm.ens-lyon.fr.

Secondary Sources Abney, Steven. 1987. The English Noun Phrase in its Sentential Aspect. Massachusetts Institute of Technology PhD thesis. Aboh, Enoch Oladé. 2010. Information Structure Begins with the Numeration. Iberia 2(1). 12–42. Adams, James N. 1976. A typological approach to Latin word order. Indogermanische Forschungen 81. 70–100.

274

references

Adams, James N. 1977. The Vulgar Latin of the Letters of Claudius Terentianus (P. Mich. VIII, 467–72). Manchester: Manchester University Press. Adams, James N. 1991. Some neglected evidence for Latin habeo + infinitive: the order of the constituents. Transactions of the Philological Society 89. 131–196. Adams, James N. 1994a. Wackernagel’s Law and the Placement of Copula Esse in Classical Latin. Cambridge: The Cambridge Philological Society. Adams, James N. 1994b. Wackernagel’s Law and the Position of Unstressed Pronouns in Classical Latin. Transactions of the Philological Society 92. 103–178. Adams, James N. 2008. Bilingualism and the Latin Language. Cambridge: Cambridge University Press. Adams, James N. 2013. Social variation and the Latin Language. Cambridge: Cambridge University Press. Adams, James N. (ed.). 2016. An Anthology of Informal Latin, 200 BC–AD 900: Fifty Texts with Translations and Linguistic Commentary. Cambridge: Cambridge University Press. https://doi.org/10.1017/9781139626446 (28 October 2020). Adams, James N. and Nigel Vincent (eds). 2016. Early and Late Latin. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9781316450826.017. Adams, Marianne. 1987a. Old French, null subjects, and verb second phenomena. UCLA PhD dissertation. Adams, Marianne. 1987b. Embedded Pro. In James Blevins and Juli Carter (eds), Proceedings of NELS 18, vol. 1, 1–22. Adams, Marianne. 1987c. From Old French to the Theory of Pro-Drop. Natural Language and Linguistic Theory 5(1). 1–32. Adams, Marianne. 1988. Les effets V2 en ancien et en moyen français. Revue québécoise de linguistique théorique et appliquée 7. 13–39. Aelbrecht, Lobke, Liliane M. V. Haegeman, and Rachel Nye (eds). 2012. Main Clause Phenomena: New horizons. Amsterdam: John Benjamins. Alexiadou, Artemis and Elena Anagnostopoulou. 1998. Parameterizing AGR: Word Order, V-movement and EPP-checking. Natural Language and Linguistic Theory 16. 491–539. Álvarez Pedrosa, Juan Antonio. 1988. Estudio comparado del orden de palabras en inscripciones jurı´dicas arcaicas, griegas y latinas. Revista española de lingu¨´ıstica 18. 109–28. Amacker, René. 1989. Sur l’ordre des termes dans la proposition latine. In Gualtiero Calboli (ed.), Subordination and Other Topics in Latin. Proceedings of the Third Colloquium on Latin Linguistics, Bologna, 1–5 April 1985, 485–501. Amsterdam: Benjamins. Ambar, M. 1992. Para uma Sintaxe da Inversão Sujeito-Verbo em Português. Lisboa: Colibri. Anderson, James Maxwell and Bernard L. Rochet. 1979. Historical Romance Morphology. Ann Arbor: University Microfilms International. Anderson, Stephen R. 2006. Verb Second, Subject Clitics and Impersonals in Surmiran. In Proceedings of the Berkeley Linguistics Society 32, 27–39. Anglade, Joseph. 1930. Grammaire élémentaire de l’ancien français. Paris: Armand Colin. Antonelli, André. 2015. Roots of V-to-C Movement in Romance: Investigating Late Latin Grammar. Presented at the Traces of History Workshop, Oslo. Ariel, Mira. 1988. Referring and accessibility. Journal of Linguistics 24. 65–87. Armstrong, Nigel. 2002. Variable deletion of French ne: a cross-stylistic perspective. Language Sciences 24(2). 153–173. Armstrong, Nigel and Alan Smith. 2002. The influence of linguistic and social factors on the recent decline of French ne. Journal of French Language Studies 12(1). 23–41.

references

275

Ashby, William J. 1976. The loss of the negative morpheme, ne, in Parisian French. Lingua 39. 119–137. Ashby, William J. 1977a. Clitic Inflection in French: An Historical Perspective. Amsterdam: Rodopi. Ashby, William J. 1977b. Interrogative forms in Parisian French. Semasia 4. 35–52. Ashby, William J. 1981a. The Loss of the Negative Particle ne in French: A Syntactic Change in Progress. Language 57(3). 674–687. https://doi.org/10.2307/414345. Ashby, William J. 1981b. The Loss of the Negative Particle ne in French: A Syntactic Change in Progress. Language 57(3). 674. https://doi.org/10.2307/414345. Ashby, William J. 1982. The drift of French syntax. Lingua 57(1). 29–46. https://doi.org/ 10.1016/0024-3841(82)90069-9. Ashby, William J. 1988. The syntax, pragmatics, and sociolinguistics of left- and rightdislocations in French. Lingua 75(2–3). 203–229. Ashby, William J. 2001. Un nouveau regard sur la chute du ne en français parlé tourangeau: s’agit-il d’un changement en cours? Journal of French Language Studies 11(1). 1–22. Ashdowne, Richard and John Charles Smith. 2007. Some semantic and pragmatic aspects of case-loss in Old French. In Joseph Salmons and Shannon Dubenion-Smith (eds), Historical Linguistics 2005, 191–205. Amsterdam: John Benjamins. Aub-Buscher, Gertrud. 1962. Le parler rural de Ranrupt (Bas Rhin): essai de dialectologie vosgienne. Paris: Klincksieck. Auger, Julie. 2003. The development of a literary standard: the case of Picard in Vimeu- Ponthieu, France. In Brian D Joseph, Johanna DeStefano, Neil G. Jacobs, and Ilse Lehiste (eds), Languages Collide: Perspectives on Language Conflict, Language Competition, and Language Coexistence. Columbus: Ohio State University Press. Auger, Julie and Anne-José Villeneuve. 2019. Building on an old feature in langue d’Oïl: interrogatives in Vimeu Picard. Journal of French Language Studies 29(2). 209–233. https://doi.org/10.1017/S095926951900005X. Auwera, Johan van der. 2009. The Jespersen Cycles. In Elly Van Gelderen (ed.), Cyclical Change, 35–71. Amsterdam: John Benjamins. Axel, Katrin. 2004. The syntactic integration of preposed adverbial clauses on the German left periphery: A diachronic perspective. In Horst Lohnstein and Susanne Trissler (eds), The Syntax and Semantics of the Left Periphery, 23–58. Berlin; Boston: De Gruyter. Axel, Katrin. 2005. Null Subjects and Verb Placement in Old High German. In Stephan Kepser and Marga Reis (eds), Linguistic Evidence, vol. 85, 27–48. Berlin; New York: De Gruyter. Axel, Katrin. 2007. Studies on Old High German Syntax: left sentence periphery, verb placement and verb-second. Amsterdam; Philadelphia, PA: John Benjamins. Ayres-Bennett, Wendy. 1987. Vaugelas and the Development of the French Language. London: Modern Humanities Research Association. Ayres-Bennett, Wendy. 1994. Negative Evidence: Or Another Look at the Non-use of Negative ‘ne’ in Seventeenth-Century French. French Studies 48(1). 63. Ayres-Bennett, Wendy. 1996. A history of the French language through texts. London; New York: Routledge. Ayres-Bennett, Wendy. 2004. Sociolinguistic Variation in Seventeenth-Century France. Cambridge: Cambridge University Press. Ayres-Bennett, Wendy and Philippe Caron. 2016. Periodization, Translation, Prescription and the Emergence of Classical French. Transactions of the Philological Society 114(3). 339–390. https://doi.org/10.1111/1467-968X.12081.

276

references

Bakker, Cecile de. 1997. Germanic and Romance Inversion in French. The Hague: Holland Academic Graphics. Baldi, Philip. 1999. The Foundations of Latin. Berlin: De Gruyter. Barbosa, Pilar. 1995. Null Subjects. Massachusetts Institute of Technology PhD dissertation. Barbosa, Pilar. 2000. Clitics: A Window into the Null Subject Property. In João Costa (ed.), Portuguese Syntax: New Comparative Perspectives, 31–94. Oxford: Oxford University Press. Barbosa, Pilar. 2001. On inversion in wh-questions in Romance. In Aafke Hulk and JeanYves Pollock (eds), Subject Inversion in Romance and the Theory of Universal Grammar (Oxford Studies in Comparative Syntax), 20–59. Oxford: Oxford University Press. Barbosa, Pilar. 2019. ‘pro’ as a minimal NP: towards a unified theory of ‘pro’-drop. Linguistic Inquiry 50(3). 487–526. Barnes, B. K. 1985. The Pragmatics of Left Detachment in Spoken Standard. Amsterdam: John Benjamins. Barra Jover, M. 2004. The French language’s place in the classification of Romance languages. Langue Française (141). 3–13. Barral, Marcel. 1980. L’imparfait du subjonctif: étude sur l’emploi et la concordance des temps du subjonctif. Paris: Picard. Bauche, Henri. 1946. Le Langage populaire. Grammaire, syntaxe et dictionnaire du français tel qu’on le parle dans le peuple avec tous les termes d’argot usuel. Paris: Payot. Bauer, Brigitte. 1995. The Emergence and Development of SVO Patterning in Latin and French: Diachronic and Psycholinguistic Perspectives. New York: Oxford University Press. Bauer, Brigitte. 2009. Word Order. In Philip Baldi and Pierluigi Cuzzolin (eds), New Perspectives on Latin Syntax. Vol. 1. Syntax of the Sentence, 241–316. Berlin: de Gruyter. Bazin-Tacchella, Sylvie. 2020a. Catégories variables: le verbe. In Christiane MarchelloNizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande grammaire historique du français, vol. 1, 745–854. Berlin: De Gruyter. Bazin-Tacchella, Sylvie. 2020b. Catégories variables: le verbe. In Christiane MarchelloNizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande grammaire historique du français, vol. 1, 745–856. Berlin: De Gruyter. Bec, Pierre. 1970. Manuel pratique de philologie romane, vol. 2. Paris: Picard. Beeching, Kate, Nigel Armstrong, and Françoise Gadet. 2009. Sociolinguistic Variation in Contemporary French (IMPACT: Studies in Language and Society: 26). Amsterdam: John Benjamins. Belletti, Adriana. 1990. Generalised Verb Movement. Turin: Rosenberg and Sellier. Belletti, Adriana. 2001. ‘Inversion’ as Focalization. In Aafke Hulk and Jean-Yves Pollock (eds), Subject Inversion in Romance and the Theory of Universal Grammar, 60–91. Belletti, Adriana. 2004. Aspects of the Low IP Area. In Luigi Rizzi (ed.), The Structure of CP and IP, 16–51. Belletti, Adriana. 2005a. Extended Doubling and the VP Periphery. Probus 17. 1–35. Belletti, Adriana. 2005b. Answering with a cleft: The role of the Null-subject parameter and the vP periphery. In Laura Brugè, Giuliana Giusti, Nicola Munaro, Walter Schweikert, and Giuseppe Turano (eds), Proceedings of the thirtieth ‘Incontro di Grammatica Generativa’, 63–82. Venice: Editrice Cafoscarina. Belletti, Adriana. 2006. (Past) Participle Agreement. In The Blackwell Companion to Syntax, vol. 3, 493–521. Oxford: Blackwell. Belletti, Adriana. 2008. Structures and Strategies. New York; London: Routledge.

references

277

Benincà, Paola. 1983a. Osservazioni sulla sintassi dei testi di Lio Mazor. In Langue, dialecte, littérature. Études romanes à la mémoire de Hugo Plomteux, 187–197. Louvain: Leuven University Press. Benincà, Paola. 1983b. Un’ipotesi sulla sintassi delle lingue romanze medievali. Quaderni patavini di linguistica 4. 3–19. Benincà, Paola. 1995. Complement clitics in Medieval Romance: The Tobler-Mussafia law. In Ian Roberts and Adrian Battye (eds), Clause Structure and Language Change, 325–344. Oxford: Oxford University Press. Benincà, Paola. 2001. On the position of topic and focus in the left periphery. In Guglielmo Cinque and Giampaolo Salvi (eds), Current Studies in Italian Syntax. Essays Offered to Lorenzo Renzi (North-Holland Linguistic Series), 39–64. London; New York: Elsevier. Benincà, Paola. 2004. The left periphery of Medieval Romance. Studi linguistici e filologici online 2(2). 243–297. Benincà, Paola. 2006. A detailed map of the left periphery of medieval Romance. In Raffaella Zanuttini (ed.), Crosslinguistic Research in Syntax and Semantics: Negation, Tense and Clausal Architecture, 53–86. Georgetown: Georgetown University Press. Benincà, Paola. 2013. Caratteristiche del V2 Romanzo. Lingue Romanze Antiche, Ladino Dolomitico e Portoghese. In Ermenegildo Bidese and Federica Cognola (eds), Introduzione alla linguistica del mòcheno, 65–84. Torino: Rosenberg and Sellier. Benincà, Paola and Nicola Munaro (eds). 2010. Mapping the Left Periphery. Oxford: Oxford University Press. Benincà, Paola and Cecilia Poletto. 2004. Topic, focus, and V2. In Luigi Rizzi (ed.), The Structure of CP and IP, 52–75. Oxford: Oxford University Press. Bentzen, Kristine, Gunnar Hrafn Hrafnbjargarson, Þorbjo¨rg Hróarsdóttir, and Anna-Lena Wiklund. 2007a. Extracting from V2. Working Papers in Scandinavian Syntax 79. 119– 128. Lund: Christer Platzack ed. Centre for Languages and Literature. Bentzen, Kristine, Gunnar Hrafn Hrafnbjargarson, Þorbjo¨rg Hróarsdóttir, and Anna-Lena Wiklund. 2007b. The Tromsø guide to the Force behind V2. Working Papers in Scandinavian Syntax 79. 93–118. Lund: Christer Platzack ed. Centre for Languages and Literature. Benveniste, E. 1959. Les relations de temps dans le verbe français. Bulletin de la Société de linguistique de Paris 54. 237–50. Benveniste, Émile. 1966. Problèmes de linguistique générale. Paris: Gallimard. Bernstein, Judy. 1991. DPs in French and Walloon: evidence for parametric variation in nominal head movement. Probus 3. 101–126. Bianchi, Valentina. 2012. On ‘mirative’ focus fronting. Presented at the 1st GLUE workshop, University of Rome, La Sapienza. Biberauer, Theresa (ed.). 2008. The Limits of Syntactic Variation. Amsterdam: John Benjamins. Biberauer, Theresa. 2010a. Semi null-subject languages, expletives and expletive pro reconsidered. In Parametric Variation: Null Subjects in Minimalist Theory, 153–199. Cambridge: Cambridge University Press. Biberauer, Theresa. 2010b. Introduction. In The Limits of Syntactic Variation, 7–72. Amsterdam; Philadelphia: John Benjamins. Biberauer, Theresa. 2016. Going beyond the input (and UG): an emergentist generative perspective on syntactic variation, stability and change. Presented at the KNAW 2016, Amsterdam.

278

references

Biberauer, Theresa. 2018. Pro-drop and emergent parameter hierarchies. In Federica Cognola and Jan Casalicchio (eds), Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective, 94–141. Oxford: Oxford University Press. https:// doi.org/10.1093/oso/9780198815853.003.0005. Biberauer, Theresa. 2019. Factors 2 and 3: Towards a principled approach. Catalan Journal of Linguistics 45–88. Biberauer, Theresa, Anders Holmberg, and Ian Roberts. 2014. A syntactic universal and its consequences. Linguistic Inquiry 42(2). 169–225. Biberauer, et al. 2010 Theresa, Anders Holmberg, Michelel Sheehan, and Ian Roberts. 2010. Parametric Variation: Null Subjects in Minimalist Theory. Cambridge: Cambridge University Press. Biberauer, Theresa and Ian Roberts. 2005. Changing EPP parameters in the history of English: accounting for variation and change. English Language and Linguistics 9(1). 5–46. https://doi.org/10.1017/S1360674305001528. Biberauer, Theresa and Ian Roberts. 2008. Cascading parameter changes: Internally-driven change in Middle and Early Modern English. In T. Eythorsson (ed.), Grammatical Change and Linguistic Theory: The Rosendal Papers (Linguistik Aktuell = Linguistics Today 113), 79–114. Amsterdam; Philadelphia: John Benjamins. Biberauer, Theresa and Ian Roberts. 2010. Subjects, Tense and verb-movement. In Parametric Variation: Null Subjects in Minimalist Theory, 263–303. Cambridge: Cambridge University Press. Biberauer, Theresa and Ian Roberts. 2011. Negative words and related expressions: A new perspective on some familiar puzzles. In Pierre Larrivée and Richard Ingham (eds), The Evolution of Negation: Beyond the Jespersen Cycle, 23–59. Berlin: De Gruyter. Biberauer, Theresa and Ian Roberts. 2012. Towards a Parameter Hierarchy for Auxiliaries: Diachronic Considerations. Cambridge Occasional Papers in Linguistics 6. 267–294. Biberauer, Theresa and Ian Roberts. 2014. Rethinking formal hierarchies: a proposed unification. Cambridge Occasional Papers in Linguistics, 7. 1–31 Biberauer, Theresa and Ian Roberts. 2015. Rethinking formal hierarchies: a proposed unification. In Ur Shlonsky (ed), Beyond Functional Sequence: The Cartography of Syntactic Structures, vol. 10, 295–313. Biberauer, Theresa and Ian Roberts. 2017. Parameter Setting. In Adam Ledgeway and Ian Roberts (eds), The Cambridge Handbook of Historical Syntax, 134–162. Cambridge: Cambridge University Press. https://doi.org/10.1017/9781107279070.008. Biberauer, Theresa and Michelle Sheehan (eds). 2013. Theoretical Approaches to Disharmonic Word Order. Oxford: Oxford University Press. Biberauer, Theresa, Michelle Sheehan, and Glenda Newton. 2009. Limiting synchronic and diachronic variation and change: the Final-Over-Final Constraint. Language and Linguistics 10. 701–743. Biberauer, Theresa, Michelle Sheehan, and Glenda Newton. 2010. Impossible changes and impossible borrowings. In Anne Breitbarth, Christopher Lucas, Sheila Watts, and David Willis (eds), Continuity and Change in Grammar, 35–60. Amsterdam: John Benjamins. Biberauer, Theresa and Ans van Kemenade. 2011. Subject Positions and InformationStructural Diversification in the History of English. Catalan Journal of Linguistics 10. 17–69. Blanche-Benveniste, Claire. 1990. Le français parlé: études grammaticales (Collection Sciences du Langage). Paris: Éditions du Centre national de la recherche scientifique.

references

279

Blanche-Benveniste, Claire. 1997a. La notion de variation syntaxique dans la langue parlée. Langue française 115. 19–29. Blanche-Benveniste, Claire. 1997b. Approches de la langue parlée en français: L’essentiel français. Ophyrs. Blanche-Benveniste, Claire and Colette Jeanjean. 1986. Le français parlé: Transcription et édition. Paris: Didier. Blinkenberg, Andreas. 1928. L’ordre des mots en français moderne, vol. 1. Copenhagen: Andr. Fred. Høst and Søn. Blumenthal, Peter. 1980. Die Stilistik der Subjectinversion im Italienischen. Italienische Studien 3. 119–131. Boeckx, Cedric and Kleanthes K. Grohmann. 2005. Left Dislocation in Germanic. In Werner Abraham (ed.), Focus on Germanic Typology, 131–144. Berlin: Akademie Verlag. Bolkestein, Machtelt. 1989. Parameters in the Expression of Embedded Predications in Latin. In Gualtiero Calboli (ed.), Subordination and Other Topics in Latin. Proceedings of the Third Colloquium on Latin Linguistics, Bologna, 1–5 April 1985, 3–35. Amsterdam: John Benjamins. Bolkestein, Machtelt. 1995. Functions of verb-subject order in Latin. STUF – Language Typology and Universals 48(1–2). 32–43. Bolkestein, Machtelt. 2000. Discourse Organization and Anaphora in Latin. In Susan C. Herring, Pieter van Reenen, and Lene Schøsler (eds), Textual Parameters in Older Languages (Current Issues in Linguistic Theory 195), 107–137. Amsterdam: John Benjamins. Bonnard, Henri and Claude Régnier. 2008. Petite Grammaire de l’ancien français. 5th edition. Paris: Magnard. Borer, Hagit. 1984. Parametric Syntax: Case Studies in Semitic and Romance Languages. Dordrecht: Foris. Bortolussi, Bernard. 2017. Topicalizations, left dislocations and the left-periphery. Catalan Journal of Linguistics 16. 101. https://doi.org/10.5565/rev/catjl.208. Boškovic´, Željko. 2005. On the locality of left branch extraction and the structure of NP*. Studia Linguistica 59(1). 1–45. Boškovic´, Željko. 2008. Left branch extraction, structure of NP, and scrambling: The Free Word Order Phenomenon: Its Syntactic Sources and Diversity. In Joachim Sabel and Mamoru Saito (eds), The Free Word Order Phenomenon (Studies in Generative Grammar 69). Berlin; New York: De Gruyter. Boškovic´, Željko. 2009. More on the No-DP Analysis of Article-Less Languages. Studia Linguistica 63(2). 187–203. Bossong, Georg. 2016. Classifications. In Martin Maiden and Adam Ledgeway (eds), The Oxford Guide to the Romance Languages. Oxford: Oxford University Press. Bouchard, Denis. 2002. Adjectives, Number and Interfaces: Why Languages Vary. Amsterdam: North Holland. Boucher, Paul. 2004. Perfect Adjective Positions in French: a Diachronic Perspective. Unpublished manuscript. University of Nantes. Boucher, Paul. 2005. Definite Reference in Old and Modern French: The Rise and Fall of DP. In Montserrat Batllori, Maria-Lluisa Hernanz, Carme Picallo, and Francesc Roca (eds), Grammaticalization and Parametric Variation, 95–108. Oxford: Oxford University Press. Bourciez, Édouard Eugène Joseph. 1910. Éléments de linguistique romane. Paris: Klincksieck. Brandi, Luciana and Patrizia Cordin. 1989. Two Italian Dialects and the Null Subject Parameter. In Osvaldo Jaeggli and Ken Safir (eds), The Null Subject Parameter, 111–142. Dordrecht: Springer.

280

references

Brandner, Josef. 2004. Head-movement in Minimalism, and V2 as FORCE-marking. In Horst Lohnstein and Susanne Trissler (eds), The Syntax and Semantics of the Left Periphery, 97–138. Berlin: De Gruyter. Branigan, Phil. 2011. Provocative syntax (Linguistic Inquiry Monographs 61). Cambridge, MA: MIT Press. Branigan, Phil. 2020. Multiple Feature Inheritance and the phase structure of the left periphery. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 150–177. Oxford: Oxford University Press. Breitbarth, Anne, Christopher Lucas, and David Willis. 2020. The History of Negation in the Languages of Europe and the Mediterranean: Volume II: Patterns and Processes. Oxford: Oxford University Press. Bruneau, Charles. 1948. La négation en wallon namurois. In Mélanges de philologie romane et de littérature médiévale offerts à Ernest Hoepffner par ses élèves et ses amis, 45–52. Paris: Belles Lettres. Brunot, Ferdinand. 1927. Histoire de la langue française des origines à 1900. Tome II, Le seizième siècle (2e édition revue et corrigée). Paris: Colin. Brunot, Ferdinand. 1933. Histoire de la langue française des origines à nos jours. 4th edition. Vol. I, De l’époque latine à la Renaissance. Paris: Colin. Brunot, Ferdinand. 1939. Histoire de la langue française, des origines à 1900. Tome IV, La langue classique (1660–1715). Deuxième partie (2e édition). Paris: Colin. Brunot, Ferdinand and Charles Bruneau. 1954. Précis de grammaire historique de la langue française. 5th edition. Paris: Masson and Cie. Buridant, Claude. 1987. L’ancien français à la lumière de la typologie des langues: les résidus de l’ordre O V en ancien français et leur effacement en moyen français. Romania 108(429). 20–65. https://doi.org/10.3406/roma.1987.1819. Buridant, Claude. 2000. Grammaire nouvelle de l’ancien français. Paris: Sedes. Burzio, Luigi. 1986. Italian Syntax: A Government-Binding Approach. Dordrecht: Foris. Cabrillana, Concepción. 1993. Posiciones relativas en la ordenación de constituyentes (I): Estudio de la posición de Sujeto, Objeto y Verbo en latı´n. HABIS 24. 249–266. Cabrillana, Concepción. 1999. Type of Text, Pragmatic Function, and Constituent Order: A Comparative Study of the Mulomedicina Chironis and the Peregrinatio Egeriae. In Hubert Petersmann and Rudolf Ketteman (eds), Latin vulgaire – latin tardif V. Actes du Ve colloque international sur le latin vulgaire et tardif Heidelberg, 5–8 septembre 1997, 319-330. Heidelberg: Winter. Cabrillana, Concepción. 2011. Theoretical and Applied Perspectives in the Teaching of Latin Syntax: On the Particular Question of Word Order. In Renato Oniga, Rossella Iovino, and Giuliana Giusti (eds), Formal Linguistics and the Teaching of Latin, 65–85. Newcastle: Cambridge Scholars. Camacho, José. 2013. Null Subjects. Cambridge: Cambridge University Press. Cardinaletti, Anna. 1990. Subject/object asymmetries in German null topic constructions and the status of Spec, CP. In Joan Mascarò and Marina Nespor (eds), GLOW essays for Henk van Riemsdijk, 75–84. Dordrecht: Foris. Cardinaletti, Anna. 1997. Subjects and Clause Structure. In Liliane Haegeman (ed.), The New Comparative Grammar, 33–63. London: Longman. Cardinaletti, Anna. 2004. Towards a cartography of subject positions. In Luigi Rizzi (ed.), The Structure of CP and IP: The cartography of syntactic structures, vol. 2, 11–65. Oxford: Oxford University Press. Cardinaletti, Anna. 2010. On a (wh-)moved topic in Italian, compared to Germanic. In Artemis Alexiadou, Jorge Hankamer, Thomas McFadden, Justin Nuger, and Florian

references

281

Scha¨fer (eds), Advances in Comparative Germanic Syntax, 3–40. Amsterdam: John Benjamins. Cardinaletti, Anna. 2021. The position of subjects in Germanic and Romance questions. In Sam Wolfe and Christine Meklenborg (eds), Continuity and Variation in Germanic and Romance, 44–70. Oxford: Oxford University Press. Cardinaletti, Anna and Ian Roberts. 2002. Clause Structure and X-Second. In Guglielmo Cinque (ed.), Functional Structure in DP and IP: The cartography of syntactic structures, vol. 1, 123–166. Oxford: Oxford University Press. Cardinaletti, Anna and Ur Shlonsky. 2004. Clitic Positions and Restructuring in Italian. Linguistic Inquiry 35(4). 519–557. https://doi.org/10.1162/0024389042350523. Cardinaletti, Anna and Michal Starke. 1999. The typology of structural deficiency: A case study of the three classes of pronouns. In Henk Van Riemsdijk (ed.), Eurotyp. Volume 5/Part 1: Clitics in the Languages of Europe, 145–234. Berlin: De Gruyter. Carlier, Anne. 2001. La genèse de l’article. un Langue française 130(1). 65–88. https://doi.org/10.3406/lfr.2001.1027. Carlier, Anne. 2004. Sur les premiers stades de l’article partitif. Scolia 18. 115–146. Carlier, Anne. 2012. Grammaticalization in Progress in Old French: Indefinite Articles. In Deborah Arteaga (ed.), Research on Old French: The State of the Art (Studies in Natural Language and Linguistic Theory), 45–60. Dordrecht: Springer. Carlier, Anne, Céline Guillot-Barbance, Christiane Marchello-Nizia, and Lene Schøsler. 2020a. Catégories variables: noms, adjectifs, pronoms et déterminants. In Christiane Marchello-Nizia, Bernard Combettes, Sophie Prévost and Scheer Tobias (eds), Grande grammaire historique du français, vol. 1, 632–643. Berlin: De Gruyter. Carlier, Anne, Céline Guillot-Barbance, Christiane Marchello-Nizia, and Lene Schøsler. 2020b. Syntaxe interne des groupes de mots et morphèmes. In Christiane MarchelloNizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande grammaire historique du français, vol. 1, 971–1055. Berlin: De Gruyter. Carnie, Andrew, Sheila Ann Dooley-Collberg, and Heidi Harley. 2005. Verb First. Amsterdam: John Benjamins. Carnie, Andrew and Eithne Guilfoyle. 2000. The Syntax of Verb Initial Languages. New York: Oxford University Press. Carnie, Andrew, Heidi Harley, and Elizabeth Pyatt. 1994. The Resurrection: Raising to Comp, Evidence from Old Irish. Studies in the Linguistic Sciences 24(1/2). 85–100. Carnie, Andrew, Heidi Harley, and Elizabeth Pyatt. 2000. VSO Order as Raising Out of IP? Some Evidence from Old Irish. In Andrew Carnie and Eithne Guilfoyle (eds), The Syntax of Verb Initial Languages, 39–60. Oxford: Oxford University Press. Carnie, Andrew, Elizabeth Pyatt, and Heidi Harley. 1994. The Resurrection: Raising to Comp? Some Evidence from Old Irish. Studies in the Linguistic Sciences 24(2). 85–100. Carrera, Micaela. 1983. Orden de palabras en un texto latı´n alto-medieval. Revista de la sociedad española de lingu¨´ıstica 13. 63–89. Carruthers, Janice. 1996. The ‘passé surcomposé général’: On the Relationship between a Rare Tense and Discourse Organization. Romance Philology 50(2). http://search. proquest.com/docview/1296994560/citation/13FF7D3C8BD463E0CB3/1?accountid= 9851 (18 August, 2013). Carruthers, Janice. 1999. A problem in sociolinguistic methodology: investigating a rare syntactic form. Journal of French Language Studies 9(01). https://doi.org/10.1017/ S0959269500004427. Casalicchio, Jan and Federica Cognola. 2020. Parameterizing ‘lexical subject-finite verb’ inversion across Verb Second languages: On the role of Relativized Minimality at the vP

282

references

edge. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 594–622. Oxford: Oxford University Press. Caudal, Patrick. 2015. The passé composé in Old French and Modern French: evolution or revolution? In Jacqueline Guéron (ed.), Sentence and Discourse, Oxford: Oxford University Press. Certeau, Michel de, Dominique Julia, and Jacques Revel. 1975.Une politique de la langue: la Révolution française et les patois: l’enquête de Grégoire. Paris: Gallimard. Chafe, Wallace. 1987. Cognitive constraints and information flow. In R. Tomlin (ed.), Coherence and Grounding in Discourse: Outcome of a Symposium, 21–51. Amsterdam: John Benjamins. Cheng, Lisa. 1991. On the Typology of Wh-Questions. Massachusetts Institute of Technology Unpublished Doctoral thesis. Cheng, Lisa and Johan Rooryck. 2000. Licensing Wh-in-situ. Syntax 3(1). 1–19. Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris. Chomsky, Noam. 1982. Some Concepts and Consequences of the Theory of Government and Binding. Cambridge, MA: MIT Press. Chomsky, Noam. 1986. Barriers. Cambridge, MA: MIT Press. Chomsky, Noam. 1993. Lectures on Government and Binding: the Pisa Lectures. Berlin: De Gruyter. Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press. Chomsky, Noam. 2000. Minimalist Inquiries: The Framework. In R. Martin, D. Michaels, and J. Uriagereka (eds), Step by Step, Essays in Minimalist Syntax in Honor of Howard Lasnik, 89–156. Cambridge, MA: MIT Press. Chomsky, Noam. 2001. Derivation by Phase. In Michael Kentstowicz (ed.), Ken Hale: A Life in Language, 1–54. Cambridge, MA: MIT Press. Chomsky, Noam. 2008. On Phases. In Robert Freidin, Carlos P. Otero, and Maria Luisa Zubizarreta (eds), Foundational Issues in Linguistic Theory, 133–166. Cambridge, MA: MIT Press. Cinque, Guglielmo 1982a. Constructions with Left Peripheral Phrases, Connectedness, Move Alpha and ECP. Unpublished manuscript. University of Venice. Cinque, Guglielmo. 1982b. ‘Topic’ Constructions in Some European Languages and Connectedness. In K. Ehlich and Henk Van Riemsdijk (eds), Connectedness in Sentence, Text and Discourse, 7–41. Tilburg: Tilburg University. Cinque, Guglielmo. 1990. Types of A’ Dependencies. Cambridge, MA: MIT Press. Cinque, Guglielmo. 1993. On the evidence for partial N-movement in the Romance DP. University of Venice Working Papers in Linguistics 3(2). 21–40. Cinque, Guglielmo. 1994. On the Evidence for Partial N-Movement in the Romance DP. In Guglielmo Cinque, Jan Koster, Jean-Yves Pollock, Luigi Rizzi, and Raffaella Zanuttini (eds), Paths Towards Universal Grammar, 85–110. Washington DC: Georgetown University Press. Cinque, Guglielmo. 1999. Adverbs and Functional Heads: A Cross-Linguistic Perspective. Oxford: Oxford University Press. Cinque, Guglielmo. 2001. ‘Restructuring’ and the order of aspectual and root modal heads. In Guglielmo Cinque and Giampaolo Salvi (eds), Current Studies in Italian Syntax. Essays Offered to Lorenzo Renzi (North-Holland Linguistic Series 59), 137–156. Oxford : New York: Elsevier. Cinque, Guglielmo. 2002. A Note on ‘Restructuring’ and Quantifier Climbing in French. Linguistic Inquiry 33(4). 617–636. Cinque, Guglielmo. 2006. Restructuring and Functional Heads: The Cartography of syntactic sructures, vol. 4. Oxford: Oxford University Press.

references

283

Cinque, Guglielmo. 2020. The Syntax of Relative Clauses: A Unified Analysis. Cambridge: Cambridge University Press. Cinque, Gugliemo and Luigi Rizzi. 2009. The Cartography of Syntactic Structures. In Bernd Heine and Heiko Narrog (eds), The Oxford Handbook of Linguistic Analysis, 51–66. Oxford: Oxford University Press. Clackson, James. 2007. Indo-European LInguistics: An Introduction. Cambridge: Cambridge University Press. Clackson, James. 2011. The Forms of Latin: Inflectional Morphology. In James Clackson (ed.), A Companion to the Latin Language, 105–118. Malden: Blackwell. Clackson, James and Geoffrey C. Horrocks. 2007. The Blackwell History of the Latin Language. Malden, MA; Oxford: Wiley-Blackwell. Cocchi, Gloria and Cecilia Poletto. 2007. Complementizer deletion and complementizer doubling. Proceedings of the XXXII Incontro di Grammatica generativa, Florence, Dipartimento di Linguistica 1. 49–62. Cognola, Federica. 2013. Syntactic Variation and Verb Second: A German Dialect in Northern Italy (Linguistik Aktuell/Linguistics Today Volume 201). Amsterdam: John Benjamins. Cognola, Federica and Jan Casalicchio (eds). 2018. Null Subjects in Generative Grammar: A Synchronic and Diachronic Perspective. Oxford; New York: Oxford University Press. Colasanti, Valentina. 2017. Towards a microparameter C-hierarchy in Italo-Romance. In Rodica Zafiu, Gabriela Dindelegan, Adina Dragomirescu, and Alexandru Nicolae (eds), Romance Syntax. Comparative and Diachronic Perspectives, 191–227. Newcastle: Cambridge Scholars Press. Colasanti, Valentina. 2018. Romance morphosyntactic microvariation in complementizer and auxiliary systems. University of Cambridge dissertation. Combettes, Bernard. 1987. Marqueurs de généricité et ordre des mots: article défini et déterminant zéro en moyen français. In Georges Kleiber (ed.), Rencontres avec la généricité, 9–32. Paris: Klincksieck. Combettes, Bernard. 2003a. Variation dans la structure du syntagme verbal. In Bernard Combettes (ed.), Évolution et variation en français préclassique. Études de syntaxe, 171–198. Paris: Honoré Champion. Combettes, Bernard. 2003b. Aspects diachroniques des constructions à détachement. Cahiers de praxématique 40. 71–96. Combettes, Bernard. 2009. Les constructions à détachement chez Montaigne: le cas des topicalisations. In F. Giacone (ed.), La langue de Rabelais, 371–385. Geneva: Droz. Combettes, Bernard, Christiane Marchello-Nizia, and Sophie Prévost. 2020. Syntaxe de la phrase simple. In Christiane Marchello-Nizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande grammaire historique du français, vol. 2, 1055–1220. Berlin: De Gruyter. Compernolle, Rémi A. van. 2008. Morphosyntactic and phonological constraints on negative particle variation in French-language chat discourse. Language Variation and Change 20(2). 317–339. Compernolle, Rémi A. van. 2010. The (slightly more) productive use of ne in Montreal French chat. Language Sciences 32(4). 447–463. https://doi.org/10.1016/j. langsci.2009.07.004. Cornilescu, Alexandra. 2002. Rhematic focus at the left periphery: the case of Romanian. In Claire Beyssade, Reineke Bok-Bennema, Frank Drijkoningen, and Paola Monachesi (eds), Romance Languages and Linguistic Theory 2000, 77–91. Amsterdam: John Benjamins.

284

references

Corr, Alice. 2016. Ibero-Romance and the syntax of the utterance. University of Cambridge dissertation. Cos¸eriu, Eugenio. 1988. Der romanische Sprachtypus: Versuch einer neuen Typologisierung der romanischen Sprachen. In Jo¨rn Albrecht (ed.), Energeia und Ergon: Sprachliche Variation, Sprachgeschichte, Sprachtypologie (vol. 1, Schriften von Eugenio Coseriu (1965–1987)), vol. 1, 207–224. Tu¨bingen: Narr. Costa, João. 2004. Subject Positions and Interfaces: The Case of European Portuguese (Studies in Generative Grammar 73). Berlin: De Gruyter. Coveney, Aidan. 1996. Variability in Spoken French: A Sociolinguistic Study of Interrogation and Negation (Elm Bank Modern Language Studies). Exeter: Elm Bank. Craenenbroeck, Jeroen van. 2010. The Syntax of Ellipsis: Evidence from Dutch Dialects (Oxford Studies in Comparative Syntax). Oxford; New York: Oxford University Press. Craenenbroeck, Jeroen van and L. Haegeman. 2007. The Derivation of Subject-Initial V2. Linguistic Inquiry 38(1). 167–178. Cruschina, Silvio. 2006. Information focus in Sicilian and the left periphery. In Mara Frascarelli (ed.), Phases of Interpretation, 363–385. Berlin: De Gruyter. Cruschina, Silvio. 2008. Discourse-related features and the syntax of peripheral positions. A comparative study of Sicilian and other Romance languages. Cambridge: University of Cambridge PhD dissertation. Cruschina, Silvio. 2011. Focalization and word order in Old Italo-Romance. Catalan Journal of Linguistics 10. 95–132. Cruschina, Silvio. 2012. Discourse-Related Features and Functional Projections. New York: Oxford University Press. Cruschina, Silvio. 2016. Information and discourse structure. In Martin Maiden and Adam Ledgeway (eds), The Oxford Guide to the Romance Languages, 596–608. Oxford: Oxford University Press. Cruschina, Silvio and Adam Ledgeway. 2016. The Structure of the Clause. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 556–575. Oxford: Oxford University Press. Cruschina, Silvio and Ioanna Sitaridou. 2009. From modern to old Romance: The interaction of Information Structure and word order. Presented at the Diachronic Generative Syntax (DiGS), Universidade de Campinas. Culbertson, Jennifer. 2010. Convergent evidence for categorial change in French: From subject clitic to agreement marker. Language 86(1). 85–132. Dagnac, Anne. 2012. How do you double your C? Evidence from an Oïl dialect. Empirical Issues in Syntax and Semantics 9. 77–94. Damonte, Federico. 2005. Complementatori e complementi congiuntivi in alcuni dialetti Sardi. Quaderni di lavoro dell’ASIt 6. 71–95. Damonte, Federico. 2010. Matching moods: Mood concord between CP and IP in Salentino and Southern Calabrian subjunctive complements. In Paola Benincà and Nicola Munaro (eds), Mapping the Left Periphery, 228–256. Oxford: Oxford University Press. Danckaert, Lieven. 2012. Latin Embedded Clauses: The Left Periphery. Amsterdam; Philadelphia: John Benjamins. Danckaert, Lieven. 2015a. Studying word order changes in Latin: some methodological remarks. In Carlotta Viti (ed.), Perspectives on Historical Syntax (Studies in Language Companion Series 169), 233–251. Amsterdam: John Benjamins. Danckaert, Lieven. 2015b. The decline of Latin left-peripheral presentational foci: Causes and consequences. In Theresa Biberauer and George Walkden (eds), Syntax Over Time, 265–280. Oxford: Oxford University Press.

references

285

Danckaert, Lieven. 2016. Variation and change in Latin BE-periphrases: empirical and methodological considerations. In James N. Adams and Nigel Vincent (eds), Early and Late Latin. Continuity or Change?, 132–162. Cambridge: Cambridge University Press. Danckaert, Lieven. 2017a. The Development of Latin Clause Structure: A Study of the Extended Verb Phrase (Oxford Studies in Diachronic and Historical Linguistics). New York: Oxford University Press. Danckaert, Lieven. 2017b. Subject Placement in the History of Latin. Catalan Journal of Linguistics 16. 125–161. Danckaert, Lieven. 2017c. The loss of Latin OV: steps towards an analysis. In Enoch Olade Aboh, Eric Haeberli, Genoveva Puska´s, and Manuela Scho¨nenberger (eds), Elements of Comparative Syntax: Theory and Description, 401–446. The Hague: De Gruyter. Dardel, Robert de. 1996. A la recherche du protoroman. Berlin: De Gruyter. Dardel, Robert de and Ans de Kok. 1996. La position des pronoms régimes atones, personnels et adverbiaux, en protoroman. Geneva: Droz. Darmesteter, Arsène. 1899. A Historical French Grammar. London: Macmillan. Dauzat, Albert. 1939. Tableau de la langue française: origines, évolution, structure actuelle. Paris: Payot. De Cat, Cécile. 2005. French subject clitics are not agreement markers. Lingua 115(9). 1195–1219. De Cat, Cécile. 2009. French Dislocation: Interpretation, Syntax, Acquisition. New York: Oxford University Press. De Clercq, Karen. 2017. The nanosyntax of French negation: A diachronic perspective. In Silvio Cruschina, Katharina Hartmann, and Eva-Maria Remberger (eds), Studies on Negation: Syntax, Semantics, and Variation, 49–80. Go¨ttingen: Vienna University Press. De Crousaz, Isabelle and Ur Shlonsky. 2003. The Distribution of a Subject Clitic Pronoun in a Franco-Provençal Dialect and the Licensing of Pro. Linguistic Inquiry 34(3). 413–442. De Haan, G. J. and Fred Weerman. 1986. Finiteness and Verb Fronting in Frisian. In Hubert Haider and Martin Prinzhorn (eds), Verb Second Phenomena in Germanic Languages, 77–110. Foris: Dordrecht. Delais-Roussarie, Élisabeth, Jenny Doetjes, and Petra Sleeman. 2004. Dislocation. In Francis Corblin and Henriëtte de Swart (eds), Handbook of French Semantics, 501–528. Stanford, CA: CSLI Publications. Delbru¨ck, Berthold. 1888. Altindische Syntax. Halle: Verlag der Buchhandlung des Waisenhauses. Den Besten, Hans. 1983. On the Interaction of Root Transformations and Lexical Deletive Rules. In Werner Abraham (ed.), On the Formal Syntax of the Westgermania, 47–61. Amsterdam: John Benjamins. Deprez, Viviane 1990. Two Ways of Moving the Verb in French. In Lisa Cheng and H. Demirdache (eds), Papers on wh-movement: MIT Working Papers in Linguistics, Volume 13, 47–85. Cambridge, MA: MIT Press. Déprez, Viviane. 2003. Concordance négative, syntaxe des mots-N et variation dialectale. Cahiers de Linguistique Française 25. 97–118. Deshaies, Denise and Éve Laforge. 1981. Le futur simple et le futur proche dans le français parlé dans la ville de Québec. Langues et linguistiques 7. 23–37. Désirat, Claude and Tristan Hordé. 1976. La langue française au XXe siècle. Paris: Bordas. Detges, Ulrich. 2003. La grammaticalisation des constructions de négation dans une perspective onomasiologique, ou: la déconstruction d’une illusion d’optique. In Andreas

286

references

Blank and Peter Koch (eds), Kognitive romanische Onomasiologie und Semasiologie, 213–233. Tu¨bingen: Niemeyer. Detges, Ulrich and Richard Waltereit. 2002. Grammaticalization vs. reanalysis: A semanticpragmatic account of functional change in grammar. Zeitschrift fu¨r Sprachwissenschaft 21. 151–195. Deutscher, Guy. 2009. ‘Overall complexity’: a wild goose chase? In G. Sampson, D. Gil, and P. Trudgill (eds), Language Complexity as an Evolving Variable. Oxford: Oxford University Press. Devine, Andrew M. and Laurence D. Stephens. 2006. Latin Word Order: Structured Meaning and Information. Oxford; New York: Oxford University Press. Devine, Andrew M. and Laurence D. Stephens. 2017. Towards a Syntax-Semantics Interface for Latin. Catalan Journal of Linguistics 16. 79. https://doi.org/10.5565/rev/catjl.210. Devine, Andrew M. and Laurence D. Stephens. 2019. Pragmatics for Latin: from syntax to information structure. New York: Oxford University Press. Diémoz, Federica. 2007. Morphologie et syntaxe des pronoms personnels sujets dans les parlers francoprovençaux de la Vallée d’Aoste. Tu¨bingen: Francke. Diesing, Molly. 1990. Verb movement and the subject position in Yiddish. Natural Language and Linguistic Theory 8(1). 41–79. Diesing, Molly. 1997. Yiddish VP order and the typology of object movement in Germanic. Natural Language and Linguistic Theory 15(2). 369–427. Diesing, Molly and Beatrice Santorini. 2020. The scope of embedded Verb Second in modern Yiddish. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 665–682. Oxford: Oxford University Press. Diez, Friedrich. 1882. Grammatik der romanischen Sprachen. Bonn: Weber. Dobrovie-Sorin, Carmen. 1994. The Syntax of Romanian: Comparative Studies in Romance. Berlin; New York: De Gruyter. Donaldson, Bryan. 2012. Initial subordinate clauses in Old French: Syntactic variation and the clausal left periphery. Lingua 122(9). 1021–1046. https://doi.org/ 10.1016/j.lingua.2012.04.003. Dragomirescu, Adina and Alexandru Nicolae. 2016. Case. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 911–923. Oxford: Oxford University Press. Dufresne, Monique. 1993. L’Articulation syntaxique et phonologique de la cliticisation: le cas des pronoms sujets en moyen français. university of Québec in Montreal Thesis. Dufresne, Monique and Fernande Dupuis. 1994. Modularity and the renanalysis of the French subject pronoun. Probus 6. 103–123. Dufter, Andreas. 2008. On explaining the rise of c’est -clefts in French. In Ulrich Detges and Richard Waltereit (eds), Current Issues in Linguistic Theory, vol. 293, 31–56. Amsterdam: John Benjamins. https://doi.org/10.1075/cilt.293.03duf. Dufter, Andreas and Daniel Jacob. 2009. Introduction. In Andreas Dufter and Daniel (eds), Focus and Background in the Romance Languages, 1–19. Amsterdam; Philadelphia, PA: John Benjamins. Duguine, Maya. 2014. Argument ellipsis: a unitary approach to pro-drop. The Linguistic Review 31(3–4). 515–550. Dupuis, Fernande. 1988. Pro-drop dans les subordonnées en ancien français. Revue québécoise de linguistique théorique et appliquée 7. 41–62. Dupuis, Fernande. 1989. L’expression du sujet dans les subordonnées en ancien français. University of Montreal PhD dissertation.

references

287

Eberhard, David, Gary F. Simons, and Charles D. Fennig. 2020. Ethnologue: Languages of the World. 23rd edition. Dallas, TX: SIL International. http://www.ethnologue.com (18 February, 2021). Einhorn, E. 1974. Old French: A Concise Handbook. Cambridge: Cambridge University Press. Elcock, William Denis. 1960. The Romance Languages (The Great Languages). London: Faber and Faber. Elsig, Martin. 2009. Grammatical Variation across Space and Time: The French interrogative system. Amsterdam: John Benjamins. Elsig, Martin and Shana Poplack. 2006. Transplanted dialects and language change: Question formation in Québec. Penn Working Papers in Linguistics 12(2). 77–90. Emonds, Joseph. 1976. A Transformational Approach to English Syntax. New York: Academic Press. Escandell-Vidal, Marı´a Victoria. 1999. Los enunciados interrogativos. Aspectos sema´nticos y sinta´cticos. In Grama´tica Descriptiva de la Lengua Española, 3929–3991. Madrid: Espasa. Ewert, Alfred. 1933. The French Language. London: Faber and Faber. Eythorsson, T. 1995. Verbal Syntax in the Early Germanic Languages. Cornell University PhD dissertation. Faarlund, Jan-Terje. 2004. The Syntax of Old Norse. Oxford: Oxford University Press. Faarlund, Jan-Terje. 2010. Word Order. In Silvia Luraghi and Vit Bubenik (eds), The Bloomsbury Companion to Historical Linguistics, 1–38. London; New York: Bloomsbury. Faure, Richard and Michèle Oliviéri. 2013. Stratégies de topicalisation en occitan. Corpus (12). 231–270. Ferraresi, Gisella. 1997. Word order and phrase structure in Gothic: A comparative study. University of Stuttgart PhD dissertation. Ferraresi, Gisella. 2005. Word Order and Phrase Structure in Gothic. Leuven: Peeters. Ferraresi, Gisella and Maria Goldbach. 2002. V2 Syntax and Topicalization in Old French. Linguistiche Berichte 189. 2–25. Ferraresi, Gisella and Maria Goldbach. 2003. Particles and sentence structure: a historical perspective. In Syntactic Structures and Morphological Information. Berlin; Boston, MA: De Gruyter. https://doi.org/10.1515/9783110904758.101. Fife, James. 1993. Constituent-Order Frequencies in Modern Welsh. Journal of Celtic Linguistics 2. 1–33. Firbas, Jan. 1992. Functional Sentence Perspective in Written and Spoken Communication (Studies in English Language). Cambridge: Cambridge University Press. Fischer, Susann. 2004. Stylistic Fronting: A contribution to information structure. In Arthur Stepanov, Gilbert Fanselow, and Ralf Vogel (eds), Minimality Effects in Syntax, 125–147. Berlin: De Gruyter. Fischer, Susann and Artemis Alexiadou. 2001. Stylistic Fronting: Germanic vs. Romance. Working Papers in Scandinavian Syntax 68. 117–145. Fleischman, Suzanne. 1982. The Future in Thought and Language: Diachronic Evidence from Romance. Cambridge: Cambridge University Press. Fleischman, Suzanne. 1990. Philology, Linguistics, and the Discourse of the Medieval Text. Speculum 65(01). 19–37. https://doi.org/10.2307/2864470. Fleischman, Suzanne. 1991. Discourse pragmatics and the grammar of Old French: A functional reinterpretation of ‘si’ and the personal pronouns. Romance Philology 44. 251–283.

288

references

Fonseca-Greber, Bonnibeth Beale. 2007. The Emergence of Emphatic ‘ne’ in Conversational Swiss French. Journal of French Language Studies 17(3). 249–275. Fontana, Josep. 1993. Phrase structure and the syntax of clitics in the history of Spanish. University of Pennsylvania PhD dissertation. Fortson, Benjamin W.Indo-European Language and Culture. 2008. 2nd edition. Oxford: Wiley-Blackwell. Foulet, Lucien. 1919. Petite syntaxe de l’ancien français. Paris: H. Champion. Foulet, Lucien. 1928. Petite syntaxe de l’ancien français. 3rd edition. Paris: H. Champion. Fournier, Nathalie. 2001. Expression et place des constituants dans l’énoncé en français classique: la relation sujet-verbe et la relation verbe-objet. Langue française 130(1). 89–107. Fournier, Nathalie. 2007a. Grammaire du français classique. Paris: Belin. Fournier, Nathalie. 2007b. L’ordre des mots dans la prose narrative au 17e siècle. In E. Comes and F. Hrubaru (eds), Actes du XIIIe Séminaire de Didactique Universitaire, 83–104. ClujNapoca: Echinox. Frascarelli, Mara. 2007. Subjects, topics and the interpretation of referential pro.Natural Language and Linguistic Theory 25(4). 691–734. https://doi.org/10.1007/s11049-0079025-x. Frascarelli, Mara. 2017. Dislocations and framings. In Andreas Dufter and Elisabeth Stark (eds), Manual of Romance Morphosyntax and Syntax, 472–501. Berlin; Boston, MA: De Gruyter. Frascarelli, Mara and Roland Hinterho¨lzl. 2007. Types of Topics in German and Italian. In S. Winkler and K. Schwabe (eds), On Information Structure, Meaning and Form, 87–116. Amsterdam; Philadelphia, PA: John Benjamins. Frey, Werner. 2004. The grammar-pragmatics interface and the German prefield. Sprache und Pragmatik 52. 1–30. Fuchs, Christiane and Christiane Marchello-Nizia (eds). 1998. Les opérations de thématisation en français. Cahiers de praxématique 30 ‘Les opérations de thématisation en français’. Gadet, Françoise. 1996. Variabilité, variation, variété: le français d’Europe. Journal of French Language Studies 6(01). 75. https://doi.org/10.1017/S0959269500004981. Gadet, Françoise. 1997. Le français populaire (Que Sais-Je?). 2nd edition. Paris: Presses Universitaires de France. Gadet, Françoise. 2003. La variation: le français dans l’espace social, régional et international. In Marina Yaguello (ed.), Le grand livre de la langue française, 91–152. Paris: Seuil. Gallego, Ángel J. 2007. Phase Theory and Parametric Variation. Universitat Autònoma de Barcelona PhD thesis. Gallego, Ángel J. 2013. Object shift in Romance. Natural Language and Linguistic Theory 31(2). 409–451. Garcı´a Sanchidria´n, Marı´a Luisa. 1994. ‘El orden de palabras en las cartas de San Braulio’, in Actas del VIII congreso español de estudios cla´sicos, 549–53. Madrid: Sociedad española de estudios cla´sicos. Garrett, Andrew. 1992. Topics in Lycian Syntax. Historiche Sprachforschung 105. 200–212. Garrett, Andrew. 1994. Relative clause syntax in Lycian and Hittite. Die Sprache 36. 29–69. Ga¨rtner, Hans-Martin. 2003. How Icelandic Can You Be, If You Speak Icelandic B? In Lars-Olof Delsing, Cecilia Falk, Gunlo¨g Josefsson, and Halldór Ármann Sigurdsson

references

289

(eds), Grammar in Focus: Festschrift for Christer Platzack, 115–122. Lund: Department of Scandinavian Languages. Ga¨rtner, Hans-Martin. 2016. A Note on the Rich Agreement Hypothesis and Varieties of ‘Embedded V2’. Working Papers in Scandinavian Syntax 97. 1–13. Gianollo, Chiara. 2018. Indefinites between Latin and Romance. Oxford: Oxford University Press. Giorgi, Alessandra. 2010. About the Speaker: Towards a Syntax of Indexicality. Oxford: Oxford University Press. Giorgi, Alessandra. 2012. The Theory of Syntax and the Representation of Indexicality. In Laura Brugè, Anna Cardinaletti, Giuliana Giusti, Nicola Munaro, and Cecilia Poletto (eds), Functional Heads: The Cartography of syntactic structures, vol. 7, 42–54. Oxford; New York: Oxford University Press. Giorgi, Alessandra and Fabio Pianesi. 1996. Verb movement in Italian and syncretic categories. Probus 8. 137–160. Giorgi, Alessandra and Fabio Pianesi. 1997. Tense and Aspect: From Semantics to Morphosyntax (Oxford Studies in Comparative Syntax). New York: Oxford University Press. Giurgea, Ion. 2017. Non-referential topical phrases in Romanian and the theory of topicality. Bucharest Working Papers in Linguistics 19(1). 35–67. Giurgea, Ion and Eva-Maria Remberger. 2012. Verum focus and polar questions. Bucharest Working Papers in Linguistics 14. 21–40. Giusti, Giuliana. 1996. Is there a TopP and a FocP in the noun phrase structure? University of Venice Working Papers in Linguistics 6. 106–128. Giusti, Giuliana. 2006. Parallels in the clausal and nominal periphery. In Mara Frascarelli (ed.), Phases of Interpretation, 163–184. Berlin; New York: De Gruyter. Giusti, Giuliana. 2014. On Force vs. Case and Fin vs. Num. In Anna Cardinaletti, Guglielmo Cinque, and Yoshio Endoì (eds), On Peripheries. Exploring Clause Initial and Clause Final Positions, 189–207. Tokyo: Hituzi Syobo Publishing. Giusti, Giuliana. 2015. Nominal Syntax at the Interfaces: A Comparative Analysis of Languages with Articles. Newcastle: Cambridge Scholars. Giusti, Giuliana. 2019. Free not-so-free adjectival order in Latin. Bucharest Working Papers in Linguistics 21. 5–40. Giusti, Giuliana and Rossella Iovino. 2011. Evidence for a Split DP in Latin. University of Venice Working Papers in Linguistics 21. 111–129. Giusti, Giuliana and Rossella Iovino. 2016. Latin as a split-DP language. Studia Linguistica 70(3). 221–249. Giusti, Giuliana and Renato Oniga. 2007. Core and Periphery in the Latin Noun Phrase. In Gérald Purnelle and Joseph Denooz (eds), Ordre et cohérence en Latin. Communications présentées au 13 Colloque International de Linguistique Latine, 81–95. Geneva: Droz. Goedegebuure, Petra. 2009. Focus structure and Q-word questions in Hittite. Linguistics 47(4). 945–967. Gonza´lez i Planas, Francesc. 2009. Cartografia de les pocisions de subjecte en gascó. Linguistica occitana 7. 83–99. Gonza´lez i Planas, Francesc. 2014. On quotative recomplementation: Between pragmatics and morphosyntax. Lingua 146. 39–74. Gougenheim, Georges. 1929. Étude sur les périphrases verbales de la langue française. Paris: Champion. Gougenheim, Georges. 1973. Grammaire de la langue française du seizième siècle. Paris: IAC.

290

references

Greco, Ciro and Liliane Haegeman. 2020. Frame setters microvariation of subject-initial Verb Second. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 61–90. Oxford: Oxford University Press. Green, John. 1982. The Evolution of Romance Auxiliaries: Criteria and Chronology. In Martin Harris and Paolo Ramat (eds), Historical Development of Auxiliaries, 97–138. Berlin: De Gruyter. Greenberg, Joseph E. 1966. Universals of Language. 2nd edition. Cambridge, MA: MIT Press. Grevisse, Maurice and André Goosse. 2016. Le bon usage. 16th edition. Brussels: De Boeck and Duculot. Grohmann, K. 2003. Prolific Domains: on the Anti-Locality of movement dependencies. Amsterdam: John Benjamins. Guardiano, Cristina and Giuseppe Longobardi. 2005. Parametric comparison and language taxonomy. In Montserrat Batllori, Maria-Lluisa Hernanz, Carme Picallo, and Francesc Roca (eds), Grammaticalization and Parametric Variation, 149–174. Oxford: Oxford University Press. Guasti, Teresa and Luigi Rizzi. 2002. Agreement and tense as distinct syntactic positions. Evidence from acquisition. In Guglielmo Cinque (ed.), Functional Structure in DP and IP: The Cartography of Syntactic Structures, vol. 1, 167–194. Oxford: Oxford University Press. Guillot-Barbance, Céline, Serge Heiden, and Alexei Lavrentiev. 2017. Base de français médiéval: une base de référence de sources médiévales ouverte et libre au service de la communauté scientifique. Diachroniques 7. 168–184. Guimier, Claude. 1997. La place du sujet clitique dans les énoncés avec adverbe initial. In Catherine Fuchs (ed.), La place du sujet en français contemporain, 43–95. Louvain: Duculot. Gupton, Timothy. 2014. The Syntax-Information Structure Interface: Clausal Word Order and the Left Periphery in Galician. Berlin: De Gruyter. https://doi.org/ 10.1515/9781614512059. Haeberli, Eric, Susan Pintzuk, and Ann Taylor. 2020. Object pronoun fronting and the nature of Verb Second in early English. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 396–425. Oxford: Oxford University Press. Haegeman, Liliane. 1996. Verb second, the split CP and null subjects in early Dutch finite clauses. GenGenP 4(2). 135–175. Haegeman, Liliane. 2006. Argument Fronting in English, Romance CLLD, and the Left Periphery. In Raffaella Zanuttini (ed.), Crosslinguistic Research in Syntax and Semantics: Negation, Tense and Clausal Architecture, 26–52. Georgetown: Georgetown University Press. Haegeman, Liliane. 2012. Adverbial Clauses, Main Clause Phenomena, and Composition of the Left Periphery. New York: Oxford University Press. Haegeman, Liliane and Ciro Greco. 2018. West Flemish V3 and the interaction of syntax and discourse. The Journal of Comparative Germanic Linguistics 21(1). 1–56. https://doi.org/10.1007/s10828-018-9093-9. Haider, Hubert. 2006. Mittelfeld Phenomena (Scrambling in Germanic). In Martin Everaert and Henk van Riemsdijk (eds), The Blackwell Companion to Syntax, 204–274. Oxford: Blackwell. Haider, Hubert and Martin Prinzhorn (eds). 1986. Verb Second Phenomena in Germanic Languages. Dordrecht: Foris. Haiman, John and Paola Benincà. 1992. The Rhaeto-Romance languages (Romance Linguistics). London; New York: Routledge.

references

291

Hale, Mark. 1987. Studies in the comparative syntax of the oldest Indo-Iranian languages. University of Harvard thesis. Hale, Mark. 1996. Deriving Wackernagel’s Law: Prosodic and Syntactic Factors Determining Clitic Placement in the Language of the Rigveda. In Aaron L. Halpern and Arnold M. Zwicky (eds), Approaching Second: Second Position Clitics and Related Phenomena, 165–197. Stanford: CSLI Publications. Hale, Mark. 2007. Historical Linguistics: Theory and Method. Oxford: Blackwell. Hale, Mark. 2017. Preliminaries to the investigation of clitic sequencing in Greek and IndoIranian. In Claire Bowern, Laurence R. Horn, and Raffaella Zanuttini (eds), On looking into words (and beyond), 289–310. Berlin: Language Science Press. Halla-Aho, Hilla. 2009. The Non-literary Latin Letters: A Study of Their Syntax and Pragmatics. Helsinki: Societas Scientiarum Fennica. Halla-Aho, Hilla. 2016. Left-detached constructions from early to late Latin: (nominatiuus pendens and attractio inuersa). In James N. Adams and Nigel Vincent (eds), Early and Late Latin, 367–389. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9781316450826.017. Halle, Morris and Alec Marantz. 1993. Distributed Morphology and the Pieces of Inflection. In Kenneth Hale and Samuel Jay Keyser (eds), The View from Building 20. Cambridge, MA: MIT Press, 275–288. Hansch, Alexandra. 2014. Germanic Properties in the Left Periphery of Old French: V-to-CMovement. Ottawa: University of Ottawa PhD dissertation. Hansen, Maj-Britt Mosegaard. 2013. Negation in the history of French. In David Willis, Christopher Lucas, and Anne Breitbarth (eds), The History of Negation in the Languages of Europe and the Mediterranean, 51–76. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199602537.003.0002. Hansen, Maj-Britt Mosegaard. 2020. La négation de proposition. In Christiane MarchelloNizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande grammaire historique du français, vol. 2, 1679–1694. Berlin: De Gruyter. Hansen, Maj-Britt Mosegaard and Jacqueline Visconti. 2007. On the diachrony of reinforced negation in French and Italian. In Corinne Rossari, Claudia Ricci, and Adriana Spiridon (eds), Grammaticalization and Pragmatics: Facts, Approaches, Theoretical Issues, 137–171. Bingley: Emerald. Hansen, Maj-Britt Mosegaard and Jacqueline Visconti (eds). 2009a. The Grammaticalization of Negative Reinforcers in Old and Middle French: A Discourse–Functional Approach. In Current Trends in Diachronic Semantics and Pragmatics, 227–251. Leiden: Brill. https://doi.org/10.1163/9789004253216_013. Hansen, Maj-Britt Mosegaard and Jacqueline Visconti. 2009b. On the Diachrony of ‘Reinforced’ Negation in French and Italian. In Corinne Rossari, Claudia Ricci, and Adriana Spiridon (eds), Grammaticalization and Pragmatics: Facts, Approaches, Theoretical Issues, 137–171. Leiden: Brill. https://doi.org/10.1163/9789004253193_009. Hansen, Maj-Britt Mosegaard and Jacqueline Visconti. 2012. The evolution of negation in French and Italian: Similarities and differences. Folia Linguistica 46(2). 453–482. Harris, Martin (ed.). 1976a. Romance Syntax: Synchronic and Diachronic Perspectives. Salford: University of Salford. Harris, Martin. 1976b. A Typological Approach to Word-Order Change in French. In Martin Harris (ed.), Romance Syntax: Synchronic and Diachronic Perspectives, 33–68. Salford: University of Salford. Harris, Martin. 1978. The Evolution of French Syntax: A Comparative Approach. London; New York: Longman.

292

references

Harris, Martin. 1982. The ‘past simple’ and the ‘present perfect’ in Romance. In Nigel Vincent and Martin Harris (eds), Studies in the Romance Verb: Essays Offered to Joe Cremona on the Occasion of his 60th Birthday, 42–70. London: Croom Helm. Harris, Martin 1984. On the causes of word order change. Lingua 63(2). 175–204. https://doi.org/10.1016/0024-3841(84)90053-6. Harris, Martin and Nigel Vincent (eds). 1988. The Romance Languages. London: Croom Helm. Haug, Dag Trygve Truslew. 2012. Syntactic conditions on null arguments in the IndoEuropean Bible translations. Acta Linguistica Hafniensia 44(2). 129–141. Hendschel, Lorint. 2001. Grammaire wallonne en ligne. http://users.skynet.be/lorint/ croejh/. Herman, József. 2000. Vulgar Latin. University Park, PA: Pennsylvania State University Press. Heycock, Caroline. 2006. Embedded Root Phenomena. In Martin Everaert and Henk van Riemsdijk (eds), The Blackwell Companion to Syntax, 174–209. Blackwell. Hinojo, Gregorio. 1985. ‘El orden de palabras en el Satiricón’, In José Melena (ed.), Symbolae Ludovico Mitxelena Oblatae. 1, 245–54. Vitoria: Universidad Paı´s Vasco Victoriaco Vasconum. Hinterho¨lzl, Roland and Svetlana Petrova (eds). 2009. Information Structure and Language Change: New Approaches to Word Order Variation in Germanic (Trends in Linguistics. Studies and Monographs 203). Berlin; New York: De Gruyter. Hinterho¨lzl, Roland and Svetlana Petrova. 2010. From V1 to V2 in West Germanic. Lingua 120(2). 315–328. https://doi.org/10.1016/j.lingua.2008.10.007. Hinzelin, Marc-Olivier and Georg A. Kaiser. 2012. Le paramètre du sujet nul dans les variétés dialectales de l’occitan et du francoprovençal. In Mario Barra Jover (ed.), Études de linguistique gallo-romane, 247–261. Vincennes: Presses Universitaires de Vincennes. Hirschbu¨hler, Paul. 1990. La légitimation de la construction V1 à sujet nul dans la prose et le vers en ancien français. Revue québécoise de linguistique théorique et appliquée 19. 32–55. Hirschbu¨hler, Paul. 1991. Null subjects in V1 embedded clauses in Philippe de Vigneulles’ Cent Nouvelles Nouvelles. University of Ottawa, manuscript. Hirschbu¨hler, Paul. 1995. Null subjects in V1 embedded clauses in Philippe de Vigneulles’ Cent Nouvelles Nouvelles. In Ian Roberts and Adrian Battye (eds), Clause Structure and Language Change, 257–291. Oxford: Oxford University Press. Hirschbu¨hler, Paul and Marie-Odile Junker. 1988. Remarques sur les sujets nuls en subordonnées en ancien et en moyen français. Revue québécoise de linguistique théorique et appliquée 7. 63–84. Hock, Hans Heinrich. 1982. AUX-Cliticization as a Motivation for Word Order Change. Studies in the Linguistic Sciences 12(1). 91–101. Hock, Hans Heinrich. 1986. Principles of Historical Linguistics. Berlin: De Gruyter. Hock, Hans Henrich. 2015. Proto-Indo-European verb-finality: Reconstruction, typology, validation. In Leonid Kulikov and Nikolaos Lavidas (eds), Benjamins Current Topics, vol. 75, 51–78. Amsterdam: John Benjamins. https://doi.org/10.1075/ bct.75.04hoc. https://benjamins.com/catalog/bct.75.04hoc (8 October 2020). Hofmann, Johann Baptist and Anton Szantyr. 1965. Lateinische Syntax und Stilistik. Munich: C. H. Beck. Holmberg, Anders. 2000a. Deriving OV Order in Finnish. In Peter Svenonius (ed.), The Derivation of VO and OV, 123–152. Amsterdam: John Benjamins.

references

293

Holmberg, Anders. 2000b. Scandinavian Stylistic Fronting: How Any Category Can Become an Expletive. Linguistic Inquiry 31(3). 445–483. Holmberg, Anders. 2015. Verb Second. In Tibor Kiss and Alexiadou Alexiadou (eds), Syntax – Theory and Analysis, 242–283. Berlin: De Gruyter. Holmberg, Anders. 2016. The Syntax of Yes and No. Oxford: Oxford University Press. Holmberg, Anders. 2017. Stylistic fronting. In Martin Everaert and Henk Van Riemsdijk (eds), The Wiley-Blackwell Companion to Syntax, vol. VII, 4093–4138. 2nd edition. Oxford: Wiley-Blackwell. Holmberg, Anders. 2020. On the Bottleneck Hypothesis of Verb Second in Swedish. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 40–60. Oxford: Oxford University Press. Holmberg, Anders, Aarti Nayudu, and Michelle Sheehan. 2009. Three partial null-subject languages: a comparison of Brazilian Portuguese, Finnish and Marathi. Studia Linguistica 63(1). 59–97. https://doi.org/10.1111/j.1467-9582.2008.01154.x. Holmberg, Anders and Christer Platzack. 1995. The Role of Inflection in Scandinavian Syntax. New York: Oxford University Press. Holmberg, Anders and Ian Roberts. 2014. Parameters and the three factors of language design. In M. Carme Picallo (ed.), Linguistic Variation in the Minimalist Framework, 61– 81. 1st edition. Oxford: Oxford University Press. Hooper, Joan and Sandra Thompson. 1973. On the applicability of root transformations. Linguistic Inquiry 4. 465–497. Hopper, Paul J. and Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press. Horrocks, Geoffrey C. 2011. Latin Syntax. In James Clackson (ed.), A Companion to the Latin Language (Blackwell Companions to the Ancient World), 118–143. Chichester, West Sussex; Malden, MA: Wiley-Blackwell. Hrafnbjargarson, Gunnar Hrafn. 2004. Stylistic Fronting. Studia Linguistica 58(2). 88–134. https://doi.org/10.1111/j.0039-3193.2004.00111.x. Hróarsdóttir, Thorbjo¨rg. 2001. Word Order Change in Icelandic: From OV to VO. Amsterdam: John Benjamins. Hsu, Brian. 2017. Verb second and its deviations: An argument for feature scattering in the left periphery. Glossa 2(1). https://doi.org/10.5334/gjgl.132. Huang, C.-T. James. 1984. On the distribution and reference of empty pronouns. Linguistic Inquiry 15. 531–574. Huang, C.-T. James, Y.-H. Audrey Li, and Yafei Li. 2009. The Syntax of Chinese. 1st edition. Cambridge: Cambridge University Press. Hulk, Aafke. 1996. How ‘greedy’ is the French imperative? Linguistics in the Netherlands 13. 97–108. Hulk, Aafke and Jean-Yves Pollock. 2001. Subject Positions in Romance and the Theory of Universal Grammar. In Subject Inversion in Romance and the Theory of Universal Grammar, 3–19. Oxford: Oxford University Press. Hulk, Aafke and Ans van Kemenade. 1995. Verb-Second, Pro-drop, Functional Projections and Language Change. In Ian Roberts and Adrian Battye (eds), Clause Structure and Language Change, 227–256. Oxford: Oxford University Press. Iatridou, Sabine and Anthony Kroch. 1992. The Licensing of CP-recursion and its Relevance to the Germanic Verb-Second Phenomenon. Working Papers in Scandinavian Syntax 50. 1–24. Iliescu, Maria. 2009. Aspects de l’évolution de l’article défini en français et en roumain. Travaux de linguistique 59(2). 13.

294

references

Ingham, Richard. 2011. Ne-drop and indefinites in Anglo-Norman and Middle English. In Pierre Larrivée and Richard Ingham (eds), The Evolution of Negation: Beyond the Jespersen Cycle, 145–163. Berlin; Boston , MA: De Gruyter. Ingham, Richard. 2012. A Derivational Approach to Negative Polarity Licensing in Old French. In Deborah Arteaga (ed.), Research on Old French: The State of the Art (Studies in Natural Language and Linguistic Theory), 261–282. Dordrecht: Springer. Ingham, Richard. 2014. Old French negation, the Tobler/Mussafia law, and V2. Lingua (SI: Jespersen Revisited: Negation in Romance and Beyond) 147. 25–39. https://doi.org/10.1016/j.lingua.2013.11.005. Ingham, Richard. 2018. Topic, Focus and null subjects in Old French. Canadian Journal of Linguistics/Revue canadienne de linguistique 63(02). 242–263. https://doi.org/ 10.1017/cnj.2017.48. Inglese, Guglielmo, Giuseppe Rizzo, and Miriam Pflugmacher. 2019. Definite referential null objects in Old Hittite. Indogermanische Forschungen 124(1). 137–170. Jacobs, Joachim. 2001. The dimensions of topic-comment. Linguistics 39(4). 641–681. Jaeggli, Osvaldo and Ken Safir. 1989. The null-subject parameter and parametric theory. In Osvaldo Jaeggli and Ken Safir (eds), The Null Subject Parameter, 1–44. Dordrecht: Kluwer. Jayaseelan, Karattuparambil. 2001. IP-internal topic and focus phrases. Studia Linguistica 55(1). 39–75. Jensen, Frede. 1990. Old French and Comparative Gallo-Romance Syntax. Tu¨bingen: De Gruyter. Jespersen, Otto. 1917. Negation in English and Other Languages. Copenhagen: A. F. Høst and Søn. Jespersen, Otto. 1922. Language: Its Nature, Development and Origin. New York: Holt. Jiménez-Ferna´ndez, Ángel L. 2015a. Towards a typology of focus: Subject position and microvariation at the discourse–syntax interface. Ampersand 2. 49–60. https://doi.org/10.1016/j.amper.2015.03.001. Jiménez-Ferna´ndez, Ángel L. 2015b. When focus goes wild: Syntactic positions for information focus. LingBaW (Linguistics Beyond and Within) P. 1. Jiménez-Ferna´ndez, Ángel L. and Shigeru Miyagawa. 2014. A feature-inheritance approach to root phenomena and parametric variation. Lingua 145. 276–302. https://doi.org/10.1016/j.lingua.2014.04.008. Jones, Mari C. 2000. The Subjunctive in Guernsey Norman French. Journal of French Language Studies 10(2). 177–203. Jones, Mari C. 2001. Jersey Norman French: A Linguistic Study of an Obsolescent Dialect (Publications of the Philological Society 34). Oxford; Boston, MA: Blackwell. Jones, Mari C. 2011. Diatopic Variation and the Study of Regional French. French Studies 65(4). 505–514. https://doi.org/10.1093/fs/knr077. Jones, Michael Allan. 1993. Sardinian Syntax. London; New York: Routledge. Jones, Michael Allan. 1996. Foundations of French Syntax. Cambridge: Cambridge University Press. Jong, Jan R. de. 1989. The position of the Latin subject. In Gualtiero Calboli (ed.), Subordination and Other Topics in Latin: Proceedings of the Third Colloquium on Latin Linguistics, Bologna, 1–5 April 1985, 521–541. Amsterdam: John Benjamins. https://doi.org/10.1075/slcs.17.33jon. Jónsson, Jóhannes Gisli. 1991. Stylistic Fronting in Icelandic. Working Papers in Scandinavian Syntax 48. 1–43.

references

295

Julien, Marit. 2007. Embedded V2 in Norwegian and Swedish. Lund University, manuscript. Kaiser, Georg A. 2002. Verbstellung und Verbstellungswandel in den romanischen Sprachen. Tu¨bingen: Max Niemeyer Verlag. Kaiser, Georg A. 2009. Losing the null subject. A contrastive study of (Brazilian) Portuguese and (Medieval) French. In Georg A. Kaiser and Eva Maria Remberger (eds), Proceedings of the Workshop ‘Null Subjects, expletives and locatives in Romance’ (Arbeitspapier 123), 131–156. University of Konstanz: Fachbereich Sprachwissenschaft. Kaiser, Georg A. and Franziska Maria Hack. 2009. Language change in comparison: the (special) case of Raeto-Romance. In Ju¨rg Fleischer and Horst Simon (eds), Comparing Diachronies/Sprachwandelvergleich, 73–97. Berlin: De Gruyter. Karimi, Simin and Helen De Hoop (eds). 2003. Scrambling in Dutch: Optionality and Optimality. In Simin Karimi (ed.)Word Order and Scrambling, 201–216. Malden: Blackwell. Kato, Mary and Esmeralda Negrão. 2000. Brazilian Portuguese and the Null Subject Parameter. Frankfurt: Vervuert. Kato, Mary and Eduardo Raposo. 2007. Topicalization in European and Brazilian Portuguese. In José Camacho, Nydia Flores-Ferra´n, and Liliana Sa´nchez (eds), Romance linguistics 2006: Selected papers from the 36th Linguistic Symposium on Romance Languages (LSRL), New Brunswick, March 31–April 2, 2006, 205–218. Amsterdam: John Benjamins. Kato, Mary and Ilza Ribeiro. 2009. Cleft sentences from Old Portuguese to Middle Portuguese. In Andreas Dufter and Daniel Jacob (eds), Focus and Background in the Romance Languages, 123–154. Amsterdam; Philadelphia, PA: John Benjamins. Kayne, Richard and Jean-Yves Pollock. 1978. Stylistic Inversion, Successive Cyclicity and Move NP in French. Linguistic Inquiry 9. 595–621. Kayne, Richard S. 1975. French Syntax: The Transformational Cycle. Cambridge, MA: MIT Press. Kayne, Richard S. 1983a. Chains, Categories External to s, and French Complex Inversion. Natural Language and Linguistic Theory 1(1). 107–139. Kayne, Richard S. 1983b. Chaînes, catégories extérieures à S et inversion complexe en français. Langue française 58(1). 36–65. Kayne, Richard S. 1989. Facets of Romance past participle agreement. In Paola Benincà (ed.), Dialect Variation and the Theory of Grammar, 85–103. Dordrecht: Foris. Kayne, Richard S. 1991. Romance Clitics, Verb Movement, and PRO. Linguistic Inquiry 22(4). 647–686. Kayne, Richard S. 1993. Toward a Modular Theory of Auxiliary Selection. Studia Linguistica 47. 3–31. Kayne, Richard S. 1994. The Antisymmetry of Syntax. Cambridge, MA: MIT Press. Kayne, Richard S. 2000. Parameters and Universals. Oxford: Oxford University Press. Kayne, Richard S. 2009. Antisymmetry and the lexicon. Linguistic Variation Yearbook 8(1). 1–31. Kiparsky, Paul. 1995. Indo-European origins of Germanic syntax. In Ian Roberts and Adrian Battye (eds), Clause Structure and Language Change, 140–169. Oxford: Oxford University Press. Kiss, Katalin É. 1998. Identificational focus versus information focus. Language 74(2). 245– 273. Klaevik-Pettersen, Espen. 2018. Inversion, V-to-C, and verb-second: An investigation into the syntax and word order of Old French and Late Latin. University of Oslo doctoral thesis.

296

references

Klinkenberg, Jean-Marie. 2001. La langue et le citoyen: pour une autre politique de la langue française (Politique Éclatée). Paris: Presses Universitaires de France. Koll, Hans-Georg. 1965. Zur Stellung des Verbs im spa¨tantiken und fru¨hmittelalterlichen Latein. Mittellateinisches Jahrbuch 2. 241–72. Koopman, Hilda and Dominique Sportiche. 1991. The position of subjects. Lingua 85(2–3). 211–258. Koster, Jan. 1975. Dutch as an SOV language. Linguistic Analysis 1. 111–136. Koster, Jan. 1978. Locality Principles in Syntax. Dordrecht: Foris. Kragh, Kirsten Jeppesen. 2010. Le remplacement de l’imparfait du subjonctif par le présent du subjonctif considéré dans une perspective de grammaticalisation. Copenhagen: Museum Tusculanum. Kristol, Andres. 2010. Syntaxe variationnelle du clitique sujet en francoprovençal valaisan contemporain: un modèle pour la diachronie du galloroman septentrional? Travaux de linguistique (Gand) 69. 47–67. Kroch, Anthony 1989. Reflexes of grammar in patterns of language change. Language Variation and Change 1(03). 199–244. https://doi.org/10.1017/S0954394 500000168. Kroch, Anthony. 2001. Syntactic Change. In Mark Baltin and Chris Collins (eds), The Handbook of Contemporary Syntactic Theory, 699–729. Oxford: Blackwell. Kroch, Anthony and Beatrice Santorini. 2009. The comparative evolution of word order in French and English. Presented at the Diachronic Generative Syntax (DiGS) 11, University of Campinas, Brazil. Kroch, Anthony S. and Ann Taylor. 2000. The Middle English Verb-Second Constraint: A Case Study in Language Contact and Language Change. In Susan C. Herring, Pieter van Reenen, Lene Schøsler, and Donald Ringe (eds), Textual Parameters in Older Languages, 353–391. Amsterdam: John Benjamins. Kroll, Wilhelm. 1918. Anfangstellung des Verbums im Lateinischen. Glotta 9. 112–123. Ku¨hner, Raphael and Carl Stemann. 1955. Ausfu¨hrliche Grammatik der latinischen Sprache, Vol. 2. Leverkusen: Gottschalk. Labelle, Marie. 2007. Clausal architecture in Early Old French. Lingua 117(1). 289–316. https://doi.org/10.1016/j.lingua.2006.01.004. Labelle, Marie. 2016. Participle fronting and clause structure in Old and Middle French. In Christina Tortora, Marcel den Dikken, Ignacio L. Montoya, and Teresa O’Neill (eds), Romance Languages and Linguistic Theory, vol. 9, 213–232. Amsterdam: John Benjamins. https://doi.org/10.1075/rllt.9.12lab. Labelle, Marie. 2019. The French Jespersen’s Cycle and Negative Concord. In Deborah Arteaga (ed.), Contributions of Romance Languages to Current Linguistic Theory, 155–172. Dordrecht: Springer. Labelle, Marie and Paul Hirschbu¨hler. 2005. Changes in Clausal Structure and the Position of Clitics in Old French. In Montserrat Batllori, Maria-Lluisa Hernanz, Carme Picallo, and Francesc Roca (eds), Grammaticalization and Parametric Variation, 149–178. Oxford: Oxford University Press. Labelle, Marie and Paul Hirschbu¨hler. 2012. Topic and Focus in Old French V1 and V2 Structures. Presented at the DiGS 14, Ottawa. Labelle, Marie and Paul Hirschbu¨hler. 2017. Leftward Stylistic Displacement in Medieval French. In Eric Mathieu and Robert Truswell (eds), Micro-change and Macro-change in Diachronic Syntax, 145–161. Oxford: Oxford University Press. Labelle, Marie and Paul Hirschbu¨hler. 2018. Topic and focus in Old French V1 and V2 structures. Canadian Journal of Linguistics 63(2). 264–287.

references

297

Laenzlinger, Christopher. 2005. French adjective ordering: perspectives on DP-internal movement types. Lingua 115(5). 649–689. Lagerqvist, Hans. 2009. Le subjonctif en français moderne: Esquisse d’une théorie modale fondée sur des textes non littéraires. Paris: PUPS. Lahousse, Karen. 2003. The distribution of postverbal nominal subjects in French. Université de Paris PhD thesis. Lahousse, Karen. 2006. NP subject inversion in French: two types, two configurations. Lingua 116(4). 424–461. https://doi.org/10.1016/j.lingua.2004.08.020. Lahousse, Karen and Béatrice Lamiroy. 2012. Word order in French, Spanish and Italian: A grammaticalization account. Folia Linguistica 46(2). https://doi.org/ 10.1515/flin.2012.014. Laka, Itziar. 1990. Negation in syntax: on the nature of functional categories and projections. Cambridge, MA: Massachusetts Institute of Technology doctoral dissertation. Lambrecht, Knud. 1981. Topic, Antitopic, and Verb Agreement in Non-standard French. Amsterdam: John Benjamins. Lambrecht, Knud. 1994. Information structure and sentence form: topic, focus, and the mental representations of discourse referents. Cambridge: Cambridge University Press. Lambrecht, Knud. 2000. When subjects behave like objects: an analysis of the merging of S and O in sentence-focus constructions across languages. Studies in Language 24(3). 611–682. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39. 463–516. Lardon, Sabine and Marie-Claire Thomine. 2009. Grammaire du français de la Renaissance. Étude morphosyntaxique. Paris: Classiques Garnier. Larrivée, Pierre. 2011. The role of pragmatics in grammatical change: The case of French preverbal non. Journal of Pragmatics (Discursive Perspectives on News Production) 43(7). 1987–1996. https://doi.org/10.1016/j.pragma.2010.12.012. Larrivée, Pierre. 2019. To be or not to be informational: preverbal complements in Medieval French V2 configurations. Glossa 4(1). 85. Le Bidois, Georges and Robert Le Bidois. 1938. Syntaxe du français moderne. Paris: Picard. Ledgeway, Adam. 2005. Moving through the left periphery: the dual complementiser system in the dialects of Southern Italy. Transactions of the Philological Society 103(3). 339–396. https://doi.org/10.1111/j.1467-968X.2005.00157.x. Ledgeway, Adam. 2007. Old Neapolitan word order: some initial observations. In Anna Laura Lepschy and Arturo Tosi (eds), Languages of Italy: Histories and dictionaries, 121– 49. Ravenna: Longo. Ledgeway, Adam. 2008. Satisfying V2 in early Romance: Merge vs. Move. Journal of Linguistics 44(02). 437–470. Ledgeway, Adam. 2009a. Aspetti della sintassi della periferia sinistra del cosentino. (ed.) Diego Pescarini. Quaderni di lavoro ASIt 9 (Studi sui dialetti della Calabria), 3–24. Ledgeway, Adam. 2009b. Grammatica diacronica del napoletano. Tu¨bingen: Niemeyer. Ledgeway, Adam. 2010a. Introduction: The clausal domain: CP structure and the left periphery. In Roberta D’Alessandro, Adam Ledgeway, and Ian Roberts (eds), Syntactic Variation: The Dialects of Italy, 38–52. Cambridge: Cambridge University Press. Ledgeway, Adam. 2010b. Subject Licensing in CP: The Neapolitan Double-subject Construction. In Paola Benincà and Nicola Munaro (eds), Mapping the Left Periphery (Oxford Studies in Comparative Syntax), 257–296. Oxford: Oxford University Press.

298

references

Ledgeway, Adam. 2011. Syntactic and morphosyntactic typology and change in Latin and Romance. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds), The Cambridge History of the Romance Languages. Volume I, Structures, 382–471. Cambridge: Cambridge University Press. Ledgeway, Adam. 2012a. From Latin to Romance: Morphosyntactic Typology and Change. Oxford: Oxford University Press. Ledgeway, Adam. 2012b. From Latin to Romance: Configurationality, Functional Categories and Head-Marking: Transactions of the Philological Society 110(3). 422–442. Ledgeway, Adam. 2013. Testing linguistic theory and variation to their limits: The case of Romance. Corpus (12). 271–327. Ledgeway, Adam. 2016. Functional categories. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 761–771. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199677108.003.0046. Ledgeway, Adam. 2017a. Syntheticity and Analyticity. In Andreas Dufter and Elisabeth Stark (eds), Manual of Romance Morphosyntax and Syntax, 839–886. Berlin: De Gruyter. Ledgeway, Adam. 2017b. Late Latin Verb Second: The Sentential Word Order of the Itinerarium Egeriae. Catalan Journal of Linguistics (Special Issue: Generative Approaches to Latin Syntax). 163–216. Ledgeway, Adam. 2018. On the decline of edge-fronting from Latin to Romance. In Ana Maria Martins and Adriana Cardoso (eds), Word Order Change, 264–279. Oxford: Oxford University Press. https://doi.org/10.1093/oso/9780198747307.003.0014. Ledgeway, Adam. 2019. Parameters in the development of Romance perfective auxiliary selection. In Michela Cennamo and Claudia Fabrizio (eds), Current Issues in Linguistic Theory, vol. 348, 344–384. Amsterdam: John Benjamins. https://doi.org/10.1075/cilt.348.17led. Ledgeway, Adam. 2020. Variation in the Gallo-Romance left periphery: V2, complementizers, and the Gascon enunciative system. In Sam Wolfe and Martin Maiden (eds), Variation and Change in Gallo-Romance Grammar, 71–100. Oxford: Oxford University Press. Ledgeway, Adam. 2021. V2 Beyond Borders: The Histoire Ancienne jusqu’à César. Journal of Historical Syntax 5(29). Ledgeway, Adam. In Prep. From Latin to Romance Syntax: The Great Leap. In Paola Crisma and Giuseppe Longobardi (eds), The Oxford Handbook of Diachronic and Historical Linguistics. Oxford; New York: Oxford University Press. Ledgeway, Adam. In press. The Verb Phrase. In Giuseppe Longobardi (ed.), The Syntax of Italian. Cambridge: Cambridge University Press. Ledgeway, Adam and Alessandra Lombardi. 2005. Verb movement, adverbs and clitic positions in Romance. Probus 17. 79–113. Ledgeway, Adam and Alessandra Lombardi. 2014. The development of the southern subjunctive. In Paola Benincà, Adam Ledgeway, and Nigel Vincent (eds), Diachrony and Dialects, 24–47. Oxford: Oxford University Press. Ledgeway, Adam and Martin Maiden (eds). 2016. The Oxford Guide to the Romance Languages. Oxford: Oxford University Press. Ledgeway, Adam and Norma Schifano. In press. Parametric Variation. In Martin Maiden and Adam Ledgeway (eds), The Cambridge Handbook of Romance Linguistics. Cambridge: Cambridge University Press. Lehmann, Christian. 2015. Thoughts on grammaticalization (Classics in Linguistics 1). 3rd edition. Berlin: Language Science Press.

references

299

Lehmann, Winfred. 1972. Converging Theories in Linguistics. Language 48(2). 266. https://doi.org/10.2307/412134. Lehmann, Winfred. 1974. Proto-Indo-European Syntax. Austin, TX: University of Texas Press. Lemieux, Monique and Fernande Dupuis. 1995. The Locus of Verb Movement in NonAsymmetric Verb-Second Languages: The Case of Middle French. In Ian Roberts and Adrian Battye (eds), Clause Structure and Language Change, 80–110. Oxford: Oxford University Press. Li, Charles and Sandra Thompson. 1976. Subject and topic: a new theory of language. In Charles Li (ed.), Subject and Topic, 457–489. New York: Academic Press. Lightfoot, David. 1989. The child’s trigger experience: Degree-0 learnability. Behavioral and Brain Sciences 12(2). 321–375. Lightfoot, David. 1995. Why UG Needs a Learning Theory: Triggering Verb Movement. In Adrian Battye and Ian Roberts (eds), Clause Structure and Language Change, 31–52. Oxford: Oxford University Press. Linde, Paul. 1923. Die Stellung des Verbs in der lateinischen Prosa. Glotta 12. 153–178. Lodge, Anthony. 1991. Authority, prescriptivism and the French standard language. Journal of French Language Studies 1(1). 93–111. Lodge, R. Anthony. 1992. French: From Dialect to Standard. London: Routledge. Lodge, R. Anthony. 2004. A Sociolinguistic History of Parisian French. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511486685. Lohnstein, Horst and Antonios Tsiknakis (eds). 2019. Verb Second: Grammar Internal and Grammar External Interfaces. Boston, MA: De Gruyter. Lombardi, Alessandra and Roberta S. Middleton. 2004. Alcune osservazioni sull’ordine delle parole negli antichi volgari italiani. In Maurizio Dardano and Gianluca Frenguelli (eds), Sint Ant: La sintassi dell’italiano antico. Atti del convegno internazionale di studi, 553–582. Università degli Studi Roma Tre, 18–21 September 2002. Rome: Arcane. Longobardi, Giuseppe. 1994. Reference and Proper Names: A Theory of N-Movement in Syntax and Logical Form. Linguistic Inquiry 25(4). 609–665. López, Luis. 2009. A Derivational Syntax for Information Structure. Oxford; New York: Oxford University Press. Loporcaro, Michele. 1998. Sintassi comparata dellaccordo participiale romanzo. 1st Italian edition. Torino: Rosenberg and Sellier. Loporcaro, Michele. 2016. Auxiliary selection and participial agreement. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 802–818. Oxford: Oxford University Press. Loporcaro, Michele. 2018. Gender from Latin to Romance: History, Geography, Typology. Oxford: Oxford University Press. Luraghi, Silvia. 1995. The pragmatics of verb-initial sentences in some ancient IndoEuropean languages. In Pamela Downing and Michael Noonan (eds), Word Order in Discourse, 355–386. Amsterdam: John Benjamins. Luraghi, Silvia. 1997. Omission of the direct object in Classical Latin. Indogermanische Forschungen 102. 239–257. Luraghi, Silvia. 1998. The Grammaticalization of the Left Sentence Boundary in Hittite. In Anna Ramat and Paul Hopper (eds), The Limits of Grammaticalization, 189–210. Amsterdam; Philadelphia, PA: John Benjamins. Luraghi, Silvia. 2003. Definite referential null objects in Ancient Greek. Indogermanische Forschungen 108. 169–196.

300

references

Luraghi, Silvia. 2004. Null objects in Latin and Greek and the relevance of linguistic typology for language reconstruction. In Karlene Jones-Bley, Martin E. Huld, Angela Della Volpe, and Miriam Robbins Dexter (eds), Proceedings of the Fifteenth Annual UCLA Indo-European Conference. Los Angeles, November 7–8, 2003, 234–256. Washington DC: Institute for the Study of Man. Mackenzie, Ian and Wim van der Wurff. 2012. Relic syntax in Middle English and Medieval Spanish: Parameter interaction in language change. Language 88(4). 846–876. https://doi.org/10.1353/lan.2012.0084. Maiden, Martin. 2011. Morphophonological Innovation. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds), The Cambridge History of the Romance Languages. Volume I, Structures, 216–267. Cambridge: Cambridge University Press. Maiden, Martin. 2016. Number. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 697–707. Oxford: Oxford University Press. Maiden, Martin. 2018. The Romance Verb: Morphomic Structure and Diachrony. Oxford: Oxford University Press. Maiden, Martin, John Charles Smith, and Adam Ledgeway (eds). 2011. The Cambridge History of the Romance Languages. Volume I, Structures. Cambridge; New York: Cambridge University Press. Maiden, Martin, John Charles Smith, and Adam Ledgeway (eds). 2013. The Cambridge History of the Romance Languages. Volume II, Contexts. Cambridge; New York: Cambridge University Press. Maling, Joan. 1990. Inversion in embedded clauses in Modern Icelandic. In Joan Maling and Annie Zaenen (eds), The Syntax of Modern Icelandic (Syntax and Semantics 24), 71–91. San Diego, CA: Academic Press. Manzano, Francis. 1997. Le gallo à la fin du XXème siècle: mythes, réalités et perspectives. In Francis Manzano (ed.), Vitalité des parlers de l’ouest et du Canada francophone, 405–451. Rennes: Presses Universitaires de Rennes. Manzini, Maria Rita and Leonardo Maria Savoia. 2003. The nature of complementizers. Rivista di Grammatica Generativa 28. 87–110. Manzini, Maria Rita and Leonardo Maria Savoia. 2010. The structure and interpretation of (Romance) complementisers. In E. P. Panagiotidis (ed.), The Complementizer Phase, 167–199. Oxford: Oxford University Press. Marandin, Jean-Marie. 2011. Subject Inversion in French: The Limits of Information Structure. In Stefan Mu¨ller (ed.), Proceedings of the HPSG 2011 Conference. Stanford, CA: CSLI Publications. Marchello-Nizia, Christiane. 1979. Histoire de la langue française aux XIVe et XVe siècles (Collection Études: Série Langue Française). Paris: Bordas. Marchello-Nizia, Christiane. 1980. La langue française aux XIVe et XVe siècles. Paris: Nathan. Marchello-Nizia, Christiane. 1985. Dire le vrai: L’adverbe ≪si≫ en français médieval: Essai de linguistique historique. (Publications Romanes et Françaises CLXVIII). Geneva: Droz. Marchello-Nizia, Christiane. 1995. L’évolution du français: Ordre des mots, démonstratifs, accent tonique. Paris: Colin. Marchello-Nizia, Christiane. 1998. Dislocations en ancien français: thématisation ou rhématisation? Cahiers de praxématique (30). 161–178. Marchello-Nizia, Christiane. 1999. Le français en diachronie: douze siècles d’évolution. Paris: Ophrys. Marchello-Nizia, Christiane, Bernard Combettes, Sophie Prévost, and Tobias Scheer (eds). 2020. Grande grammaire historique du français. Boston, MA: De Gruyter.

references

301

Marchello-Nizia, Christiane and Sophie Prévost. 2020. Expression et position des constituants majeurs dans les divers types de propositions. In Christiane Marchello-Nizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande grammaire historique du français, vol. 2, 1055–1220. Berlin: De Gruyter. Marchello-Nizia, Christiane and Magali Rouquier. 2012. De (S)OV à SVO en français: où et quand? L’ordre des constituants propositionnels dans la Passion de Clermont et la Vie de Saint Alexis. In Monique Dufresne (ed.), Constructions en changement. Hommage à Paul Hirschbu¨hler, 111–155. Québec: Presses de l’Université Laval. Marouzeau, Jules. 1938. L’ordre des mots dans la phrase latine, vol. 2. 3 vols. Paris: Geuthner. Martin, Robert and Marc Wilmet. 1973. Manuel du français du Moyen Âge. (ed.) Yves Lefèvre. Bordeaux: Société bordelaise de diffusion des travaux de lettres et sciences humaines. Martineau, France and Raymond Mougeon. 2003. A Sociolinguistic Study of the Origins of ne Deletion in European and Quebec French. Language 79(1). 118–152. Martins, Ana Maria. 1994. Enclisis, VP-deletion and the nature of Sigma. Probus 6. 173– 205. Martins, Ana Maria. 2002. The loss of IP-scrambling in Portuguese: Clause structure, word order variation and change. In David Lightfoot (ed.), Syntactic Effects of Morphological Change, 232–248. Oxford: Oxford University Press. Martins, Ana Maria. 2011. Scrambling and information focus in old and contemporary Portuguese. Catalan Journal of Linguistics 10. 133–158. Mascarenhas, Salvador. 2014. Complementizer doubling in European Portuguese. Rivista di Grammatica Generativa 36. 105–116. Massot, Benjamin. 2008. Français et diglossie: décrire la situation linguistique française contemporaine comme une diglossie: arguments morphosyntaxiques. University of Paris doctoral thesis. Mathieu, Éric. 2006. Stylistic Fronting in Old French. Probus 18(2). 219–266. Mathieu, Éric. 2009. On the Germanic properties of Old French. In Paola Crisma and Giuseppe Longobardi (eds), Historical Syntax and Linguistic Theory, 344–357. Oxford: Oxford University Press. Mathieu, Éric. 2012. The left periphery in Old French. In Deborah Arteaga (ed.), Research in Old French: The State of the Art (Studies in Natural Language and Linguistic Theory), 327–350. Dordrecht: Springer. Maupas, Charles. 1607. Grammaire françoise. Blois. McCloskey, James. 1996. Subjects and subject positions in Irish. In Robert D. Borsley and Ian G. Roberts (eds), The Syntax of the Celtic languages: a comparative perspective, 241– 283. Cambridge: Cambridge University Press. Meelen, Marieke. 2016. Why Jesus and Job spoke bad Welsh: the origin and distribution of V2 orders in Middle Welsh. University of Leiden PhD dissertation. Meelen, Marieke. 2020. Reconstructing the rise of Verb Second in Welsh. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 426–455. Oxford: Oxford University Press. Meillet, Antoine. 1903. Introduction à l’étude comparative des langues indo-européennes. Paris: Hachette. Meisel, Ju¨rgen M., Martin Elsig, and Esther Rinke. 2013. Language Acquisition and Change: A Morphosyntactic Perspective. Edinburgh: Edinburgh University Press. Meisner, Charlotte, Elisabeth Stark, and Harald Vo¨lker. 2014. Introduction to the special issue: ‘Jespersen revisited: Negation in Romance and beyond’. Lingua 147. 1–8. Meklenborg, Christine. 2020a. Resumptive Particles and Verb Second. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 90–126. Oxford: Oxford University Press.

302

references

Meklenborg, Christine. 2020b. Resumptive structures in a Gallo-Romance perspective. In Sam Wolfe and Martin Maiden (eds), Variation and Change in Gallo-Romance 41–70. Grammar Oxford: Oxford University Press. Meklenborg, Christine, Hans Petter Helland, and Terje Lohndal. 2021. Topics in French and Norwegian. In Sam Wolfe and Christine Meklenborg (eds), Continuity and Variation in Germanic and Romance, 223–247. Oxford: Oxford University Press. Meklenborg Salvesen, Christine. 2019. Special Issue: A micro-perspective on Verb Second in Romance and Germanic. Linguistics Variation 19(1). Ménard, Philippe. 1968. Manuel d’ancien français, vol. 3. Syntaxe. Bordeaux: Sobodi. Ménard, Philippe. 1988. Syntaxe de l’ancien français. Bordeaux: Éditions Bière. Mensching, Guido and Eva Remberger. 2010. Focus fronting and the left periphery in Sardinian. In Roberta D’Alessandro, Adam Ledgeway, and Ian Roberts (eds), Syntactic Variation: The Dialects of Italy, 261–276. Cambridge: Cambridge University Press. Mensching, Guido and Eva-Maria Remberger. 2016. Sardinian. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 270–291. Oxford: Oxford University Press. Meyer-Lu¨bke, Wilhelm. 1889. Grammaire des langues romanes, tr. par E. Rabiet (A. et G. Doutrepont). Paris: Welter. Miller, D. Gary. 1975. Indo-European: VSO, SOV, SVO, or all three? Lingua 37(1). 31–52. https://doi.org/10.1016/0024-3841(75)90003-0. Mioto, Carlos and Maria Lobo. 2016. Wh-movement: Interrogatives, Relatives and Clefts. In Leo Wetzels, Sergio Menuzzi, and João Costa (eds), The Handbook of Portuguese Linguistics, 275–293. Oxford: Wiley. Miyagawa, Shigeru. 2010. Why agree? Why move?: unifying agreement-based and discourseconfigurational languages (Linguistic Inquiry Monograph 54). Cambridge, MA: MIT Press. Mo¨bitz, Otto. 1924. Die Stellung des Verbums in den Schriften des Apuleius. Glotta 13. 116–126. Moignet, Gérard. 1973. Grammaire de L’ancien français: morphologie – syntaxe. Paris: Éditions Klincksieck. Mooney, Damien. 2016. Southern regional French: a linguistic analysis of language and dialect contact (Research Monographs in French Studies 47). Cambridge: Legenda. Moreno Herna´ndez, Antonio. 1989. ‘Tipologı´a lingu¨´ıstica y orden de la palabras en el latı´n de Terencio’, Actas del VII congreso español de estudios cla´sicos. Volumen I, 523–528. Madrid: Universidad Complutense. Moscati, Vicenzo. 2011. Negation raising: logical form and linguistic variation. Newcastle: Cambridge Scholars Publishing. Motapanyane, Virginia. 1998. Focus, checking theory and fronting strategies in Romanian. Studia Linguistica.52(3). 227–243. https://doi.org/ 10.1111/1467-9582. 00035. Muller, Claude. 2003. Naissance et évolution des constructions clivées en «c’est … que …»: de la focalisation sur l’objet concret à la focalisation fonctionnelle. In Peter Blumenthal and Jean-Emmanuel Tyvaert (eds), La cognition dans le temps, 101–120. The Hague: De Gruyter. Muller, Claude. 2005. Evolution de la syntaxe sujet-verbe et réanalyse. Verbum 25. 481–493. Muller, Claude. 2009. Major constituent order, information packaging and narrative structure in two Middle French texts. In Andreas Dufter and Jacob Daniel (eds), Focus and

references

303

Background in the Romance Languages, 239–280. Amsterdam; Philadelphia, PA: John Benjamins. Mu¨ller, Gereon. 2004. Verb-second as vP first. Journal of Comparative Germanic Linguistics 7. 179–234. Munaro, Nicola. 1999. Sintagmi interrogativi nei dialetti italiani settentrionali. Padua: Unipress. Munaro, Nicola. 2005. Computational puzzles of conditional clause preposing. In Anna Maria Di Sciullo (ed.), Language, Brain and Computation, 73–94. Philadelphia, PA: John Benjamins. Munaro, Nicola. 2015. Complementizer doubling and clausal topics in (early) ItaloRomance. Paper given at the 41st Incontro di Grammatica Generativa, Perugia, Italy, 26th-28th February. Munaro, Nicola, Cecilia Poletto, and Jean-Yves Pollock. 2003. Eppur si muove!: On comparing French and Bellunese wh-movement. Linguistic Variation Yearbook 1(1). 147–180. Nadasdi, Terry. 1995. Subject NP doubling, matching, and minority French. Language Variation and Change 7(1). 1–14. https://doi.org/10.1017/S0954394500000879. Neeleman, Ad and Hans van de Koot. 2008. Dutch scrambling and the nature of discourse templates. The Journal of Comparative Germanic Linguistics 11(2). 137–189. Newton, Glenda. 2006a. The Development and Loss of the Old Irish Double System of Verbal Inflection. University of Cambridge PhD dissertation. Newton, Glenda. 2006b. The contribution of the Old Irish verbal system to the reconstruction of Proto-Indo-European syntax. (ed.) N. Ya´ñez-Bouza. Papers in Linguistics from the University of Manchester 2005. 95–112. Nicolae, Alexandru. 2014. On Verb Movement in Romanian. Paper presented at SyntaxLab, University of Cambridge. Nicolae, Alexandru. 2019. Word Order and Parameter Change in Romanian: A Comparative Romance Perspective. Oxford: Oxford University Press. Nunes, Jairo. 2004. Linearization of Chains and Sideward Movement. Cambridge, MA: MIT Press. Nyrop, Christopher. 1928. Grammaire historique de la langue française. Paris: Champion. Oliviéri, Michèle and Patrick Sauzet. 2016. Southern Gallo-Romance (Occitan). In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 319– 349. Oxford: Oxford University Press. Oniga, Renato. 2014. Latin: A Linguistic Introduction. (Trans.) Norma Schifano. Oxford: Oxford University Press. Ostafin, D. M. 1986. Studies in Latin word order: a transformational approach. University of Connecticut PhD dissertation. Ouali, Hamid. 2008. On C-toT Phi-Feature Transfer: the nature of Agreement and AntiAgreement in Berber. In Roberta D’Alessandro, Gunnar Hrafn Hrafnbjargarson, and Susann Fischer (eds), Agreement Restrictions, 159–180. Berlin; New York: De Gruyter. Palasis, Katerina. 2015. Subject clitics and preverbal negation in European French: Variation, acquisition, diatopy and diachrony. Lingua 161. 125–143. Pană Dindelegan, Gabriela (ed.). 2013. The Grammar of Romanian. 1st edition. Oxford: Oxford University Press. Panchón, Federico. 1986. Orden de palabras en latı´n (César, BG.; Cicerón Pro Milone). Studia zamorensia 7. 213–29. Panhuis, Dirk G. J. 1982. The Communicative Perspective in the Sentence: A study of Latin word order (Studies in Language Companion Series (SLCS) 11). Amsterdam: John Benjamins.

304

references

Panhuis, Dirk G. J. 1984. Is Latin an SOV language? A diachronic perspective. Indogermanische Forschungen 89. 140–159. Paoli, Sandra. 2003. Comp and the Left Periphery: Comparative Evidence from Romance. University of Manchester PhD thesis. Paoli, Sandra. 2007. The fine structure of the left periphery: COMPs and subjects: evidence from Romance. Lingua 117(6). 1057–1079. Paoli, Sandra. 2010. In focus: an investigation of information and contrastive constructions. In Roberta D’Alessandro, Adam Ledgeway, and Ian Roberts (eds), Syntactic Variation: The Dialects of Italy, 277–291. Cambridge: Cambridge University Press. Parry, Mair. 2003. The interaction of pragmatics and syntax in the development of whinterrogatives with overt complementizer in Piedmontese. In Christina Tortora (ed.), The Syntax of Italian Dialects, 152–174. New York: Oxford University Press. Pearce, Elizabeth. 1990. Parameters in Old French Syntax: Infinitival Complements. Dordrecht: Kluwer. Pei, Mario A. 1949. A New Methodology for Romance Classification. Word 5(2). 135–146. Petrova, Svetlana. 2006. A discourse-based approach to verb placement in early WestGermanic. Interdisciplinary Studies on Information Structure 5. 153–185. Petrova, Svetlana. 2012. Multiple XP-fronting in Middle Low German root clauses. The Journal of Comparative Germanic Linguistics 15(2). 157–188. https://doi.org/10.1007/s10828012-9050-y. Picallo, M. Carme (ed.). 2014. Linguistic Variation in the Minimalist Framework. 1st edition. Oxford: Oxford University Press. Picoche, J. and Christiane Marchello-Nizia. 1989. Histoire de la langue française. Paris: Nathan. Pinkster, Harm. 1985. The development of future tense auxiliaries in Latin. Glotta 63. 186–208. Pinkster, Harm. 1987. The strategy and chronology of the development of future and perfect tense auxiliaries in Latin. In Martin Harris and Paolo Ramat (eds), The Historical Development of Auxiliaries, 193–223. Berlin: De Gruyter. Pinkster, Harm. 1990a. The development of cases and adpositions in Latin. In Harm Pinkster and Inge Genee (eds), Unity in diversity: papers presented to Simon C. Dik on his 50th birthday, 195–210. Dordrecht: Foris. Pinkster, Harm. 1990b. Latin Syntax and Semantics. London: Routledge. Pinkster, Harm. 1991. Evidence for SVO in Latin? In Roger Wright (ed.), Latin and the Early Romance Languages in the Middle Ages, 69–82. London: Routledge. Pinkster, Harm. 1992. Notes on the Syntax of Celsius. Mnemosyne 45. 513–524. Pinkster, Harm. 2015. The Oxford Latin Syntax: Volume 1: The Simple Clause. Oxford; New York: Oxford University Press. https://doi.org/10.1093/acprof:oso/ 9780199283613.001.0001 Pintzuk, Susan. 1999. Phrase Structures in Competition: Variation and Change in Old English Word Order. Philadelphia, PA: Courier Corporation. Planas-Morales, Sı´lvia and Xavier Villalba. 2013. The Right Periphery of Interrogatives in Catalan and Spanish: Syntax/Prosody Interactions. Catalan Journal of Linguistics 12. 193. https://doi.org/10.5565/rev/catjl.72. Platzack, Christer. 1985. A Survey of Generative Analyses of the Verb Second Phenomenon in Germanic. Nordic Journal of Linguistics 8(01). 49. https://doi.org/10.1017/ S0332586500001256. Platzack, Christer. 1986. COMP, INFL and Germanic Word Order. In L Hellan and Christensen Koch (eds), Topics in Scandinavian Syntax, 185–234. Dordrecht: Reidel.

references

305

Pohl, Jacques. 1975. L’omission de ne dans le français parlé contemporain. Le français dans le monde 111. 17–23. Pohl, Jacques. 1984. Documents pour servir à l’étude des phrases du type “Les fleurs, j’aime”. Romanistisches Jahrbuch 35(1). https://doi.org/10.1515/roja-1984-0104. Poletto, Cecilia. 1995. The Diachronic Development of Subject Clitics in North Eastern Italian Dialects. In Adrian Battye and Ian Roberts (eds), Clause Structure and Language Change, 295–325. Oxford: Oxford University Press. Poletto, Cecilia. 2000. The Higher Functional Field: Evidence from Northern Italian Dialects. Oxford: Oxford University Press. Poletto, Cecilia. 2002. The left-periphery of V2-Rhaetoromance dialects: a new view on V2 and V3. In Sjef Barbiers, Leonie Cornips, and Susanne van der Kleij (eds), Syntactic Microvariation (Meertens Institute Electronic Publications in Linguistics 2), 214–242. Amsterdam: Meertens Institute. Poletto, Cecilia. 2006a. Old Italian Scrambling: The Low Left Periphery of the Clause. In Patrick T. Brandt and Eric Fuss (eds), Form, Structure, and Grammar: A Festschrift Presented to Gu¨nther Grewendorf on Occasion of His 60th Birthday, 209–229. Berlin: Akademie Verlag. Poletto, Cecilia. 2006b. Parallel Phases: a study of the high and low periphery of Old Italian. In Mara Frascarelli (ed.), Phases of Interpretation, 261–295. Berlin: De Gruyter. Poletto, Cecilia. 2014. Word Order in Old Italian. Oxford: Oxford University Press. Poletto, Cecilia. 2015. Word Orders in the Old Italian DP. In Ur Shlonsky (ed.), Beyond Functional Sequence, 109–128. New York: Oxford University Press. Poletto, Cecilia. 2016. Negation. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages (Oxford Guides to the World’s Languages), 833–846. Oxford: Oxford University Press. Poletto, Cecilia. 2019. More than one way out: On the factors influencing the loss of V to C movement. Linguistic Variation 19(1). 47–81. https://doi.org/10.1075/lv.16001.pol. Poletto, Cecilia. 2020. Null Subjects in Old Italian. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 325–348. Oxford: Oxford University Press. Poletto, Cecilia and Jean-Yves Pollock. 2004. On the left periphery of some Romance whquestions. In Luigi Rizzi (ed.), The Structure of CP and IP: The Cartography of Syntactic Structures, vol. 2, 251–296. Oxford: Oxford University Press. Poletto, Cecilia and Jean-Yves Pollock. 2009. Another look at wh-questions in Romance: The case of Mendrisiotto and its consequences for the analysis of French wh- in situ and embedded interrogatives. In Danièle Torck and W. Leo Wetzels (eds), Romance Languages and Linguistic Theory 2006: Selected papers from ‘Goi ng Romance’, Amsterdam, 7–9 December 2006, 199–258. Amsterdam: John Benjamins. https://doi.org/10.1075/cilt.303.12pol. Poletto, Cecilia and Jean-Yves Pollock. 2017. Subject Clitics and (Complex) Inversion. In Martin Everaert and Henk Van Riemsdijk (eds), The Companion to Syntax, 1–60. 2nd edition. Hoboken, NJ: Wiley. Poletto, Cecilia and Christina Tortora. 2016. Subject clitics: Syntax. In The Oxford Guide to the Romance Languages, 772–785. Oxford: Oxford University Press. Poletto, Cecilia and Raffaella Zanuttini. 2003. Making imperatives: evidence from central Rhaetoromance. In Christina Tortora (ed.), The Syntax of Italian Dialects, 175–207. Oxford: Oxford University Press. Poletto, Cecilia and Raffaella Zanuttini. 2010. Sentential Particles and Remnant Movement. In Paola Benincà and Nicola Munaro (eds), Mapping the Left Periphery: The Cartography of Syntactic Structures, vol. 5, 201–227. New York: Oxford University Press.

306

references

Pollock, Jean-Yves. 1989. Verb movement, UG and the structure of IP. Linguistic Inquiry 20(3). 365–425. Polo, Chiara. 2004. Word Order in Latin, Italian and Slovene between Morphology and Syntax. Padua: Unipress. Polo, Chiara. 2005. Latin word order in generative perspective: An explanatory proposal within the sentence domain. In Katalin É. Kiss (ed.), Universal Grammar in the Reconstruction of Ancient Languages, 373–428. Berlin; New York: De Gruyter. Poole, Geoffrey. 2007. Defending the subject gap requirement: Stylistic Fronting in Germanic and Romance. Working Papers in Scandinavian Syntax 79(1). 1–19. Poplack, Shana and Anne St-Amand. 2007. A real-time window on 19th-century vernacular French: The Récits du français québécois d’autrefois. Language in Society 36. 707–734. Poplack, Shana and Danielle Turpin. 1999. Does the FUTUR have a future in (Canadian) French? Probus 11(1). 134–164. Posner, Rebecca. 1985. Post-verbal negation in non-standard French: a historical and comparative view. Romance Philology 39(2). 170–197. Posner, Rebecca. 1996. The Romance Languages. Cambridge: Cambridge University Press. Posner, Rebecca. 1997. Linguistic Change in French. Oxford: Clarendon Press. Prévost, F. 2002. Postposition du sujet en français aux XVe et XVIe siècles: Analyse sémanticopragmatique. Paris: CNRS Editions. Price, Glanville. 1962. The Negative Particles Pas, Mie and Point in French. Archivum Linguisticum 14. 14–34. Price, Glanville. 1971. The French Language: Past and Present. London: Edward Arnold. Price, Glanville. 1978. L’interrogation négative sans ne. Studii ¸si cercetări linguistice 29. 599– 606. Price, Glanville. 1993. Pas (point) without ne in interrogative clauses. Journal of French Language Studies 3(2). 191–195. Priestley, Leanord. 1955. Reprise constructions in French. Archivum Linguisticum 7. 1–28. Prieto, Pilar and Gemma Rigau. 2007. The Syntax-Prosody Interface: Catalan interrogative sentences headed by que. Journal of Portuguese Linguistics 6(2). 29–59. Prince, Ellen F. 1981. Toward a taxonomy of given-new information. In Peter Cole (ed.), Radical pragmatics, 223–255. New York: Academic Press. Progovac, Liliana. 1994. Negative and positive polarity. Cambridge: Cambridge University Press. Queffélec, Ambroise. 2000. Le français au Maghreb. In Gérald Antoine and Bernard Cerquiglini (eds), Histoire de la langue française 1945–2000, 765–96. Paris: CNRS. Quer, Josep. 2002. Edging quantifiers. On QP-fronting in Western Romance. In Claire Beyssade, Reineke Rok-Bennema, Frank Drijkoningen, and Paola Monachesi (eds), Romance Languages and Linguistic Theory 2000 (Current Issues in Linguistics 232), 253–270. Amsterdam: John Benjamins. Rahn, Christine. 2016. At the edge: fronting in Middle French embedded clauses. PhD thesis. Ramat, Paolo. 1984. ‘Per una tipologia del latino pompeiano’, in Linguistica tipologica, 137142. Bologna: Il Mulino. Ramsden, Herbert. 1963. Weak-Pronoun Position in the Early Romance Languages. Manchester: Manchester University Press. Rebotier, Aude. 2015. Le futur périphrastique français avec aller: un renvoi spécifique à l’avenir ou un temps en voie de grammaticalisation? Une approche contrastive. Revue de Sémantique et Pragmatique 38. 11–34.

references

307

Reenen, Pieter van and Lene Schøsler. 1992. Ancien et Moyen Français: SI ‘thématique’, analyse exhaustive d’une série de textes. Vox Romanica 51. 101–127. Reenen, Pieter van and Lene Schøsler. 2000. The pragmatic functions of the Old French particles ainz, apres, donc, lors, or, pluis, and si. In Susan C Herring, Pieter van Reenen, and Lene Schøsler (eds), Textual Parameters in Older Languages, 59–105. Amsterdam: John Benjamins. Reis, Marga. 1995. Wer glaubst du hat recht? On So-called Extractions from Verb-Second Clauses and Verb-First Parenthetical Constructions in German. Sprache und Pragmatik 36. 27–83. Reis, Marga. 2000a. Review of O. Önnerfors: Verb-erst-Deklarativsa¨tze. Grammatik und Pragmatik. Studia Linguistica 54. 90–100. Reis, Marga. 2000b. Anmerkung zu Verb-erst-Satz-Typen im Deutschen. In Rolf Thieroff, Matthias Tamrat, Nanna Furhop, and Oliver Teuber (eds), Deutscher Grammatik in Theorie und Praxis, 215–27. Tu¨bingen: Niemeyer. Remberger, Eva-Maria. 2010. Left-peripheral interactions in Sardinian. Lingua 120(3). 555– 581. https://doi.org/10.1016/j.lingua.2008.10.010. Ribeiro, Ilza. 1995. Evidence for a verb-second phase in Old Portuguese. In Adrian Battye and Ian Roberts (eds), Clause Structure and Language Change, 110–139. Oxford: Oxford University Press. Rickard, Peter. 2003. A History of the French Language. Second Edition. Abingdon: Routledge. Ringe, Donald A. 2017. From Proto-Indo-European to Proto-Germanic (A Linguistic History of English). 2nd edition. Oxford; New York: Oxford University Press. Rinke, Esther. 2009. Verb placement in Old Portuguese. In Andreas Dufter and Daniel Jacob (eds), Focus and Background in the Romance Languages, 309–332. Amsterdam; Philadelphia, PA: John Benjamins. Rinke, Esther and Martin Elsig. 2010. Quantitative evidence and diachronic syntax. Lingua 120(11). 2557–2568. https://doi.org/10.1016/j.lingua.2010.06.012. Rinke, Esther and Ju¨rgen Meisel. 2009. Subject Inversion in Old French: Syntax and information structure. In Georg A. Kaiser and Eva Maria Remberger (eds), Proceedings of the Workshop on Null Subjects, Expletives and Locatives in Romance, 93–130. University of Konstanz: Fachbereich Sprachwissenschaft. Riou, Étienne and Barbara Hemforth. 2015. Dislocation clitique de l’objet à gauche en français écrit. Discours (16). Rivero, Marı´a-Lluisa. 1980. On Left-Dislocation and Topicalization in Spanish. Linguistic Inquiry 11(2). 363–393. Rizzi, Luigi. 1982. Issues in Italian Syntax. Dordrecht: De Gruyter. Rizzi, Luigi. 1986a. On the status of subject clitics in Romance. In Osvaldo Jaeggli and Carmen Silva-Corvala´n (eds), Studies in Romance Linguistics, 391–420. Dordrecht: Foris. Rizzi, Luigi. 1986b. Null objects in Italian and the theory of pro. Linguistic Inquiry 17(3). 501–557. Rizzi, Luigi. 1990. Relativized Minimality. Cambridge, MA: MIT Press. Rizzi, Luigi. 1996. Residual verb second and the wh-criterion. In Adriana Belletti and Luigi Rizzi (eds), Parameters and Functional Heads Essays in Comparative Syntax, 63–90. Oxford: Oxford University Press. Rizzi, Luigi. 1997. The Fine Structure of the Left Periphery. In Liliane Haegeman (ed.), Elements of Grammar: Handbook of Generative Grammar, 281–338. Dordrecht: Kluwer.

308

references

Rizzi, Luigi. 2001. On the position ‘Int(errogative)’ in the Left Periphery. In Guglielmo Cinque and Giampaolo Salvi (eds), Current Studies in Italian Syntax. Essays Offered to Lorenzo Renzi, 286–296. Amsterdam: Elsevier. Rizzi, Luigi. 2004. Locality and left periphery. In Adriana Belletti (ed.), Beyond. The Cartography of Syntactic Structures, vol. 3, 223–251. Oxford: Oxford University Press. Rizzi, Luigi. 2005. On some properties of subjects and topics. In Laura Brugè, Giuliana Giusti, Nicola Munaro, Walter Schweikert, and Giuseppe Turano (eds), Proceedings of the Thirtieth Incontro di Grammatica Generativa, 63–82. Venice: Cafoscarina. Rizzi, Luigi. 2006a. On the Form of Chains: Criterial Positions and ECP Effects. In Lisa Cheng and Norbert Corver (eds), Wh-Movement. Moving On, 97–133. Cambridge, MA: MIT Press. Rizzi, Luigi. 2006b. Phase Theory and the Privilege of the Root. In Hans Broekhuis, Norbert Corver, Riny Huybregts, Ursula Kleinhenz, and Jan Koster (eds), Organizing Grammar: Linguistic Studies in Honor of Henk van Riemsdijk, 529–537. Berlin: De Gruyter. Rizzi, Luigi. 2010. On Some Properties of Criterial Freezing. In E. Phoevos Panagiotidis (ed.), The Complementizer Phase, 17–32. Oxford: Oxford University Press. Rizzi, Luigi. 2013. The functional structure of the sentence, and cartography. In Marcel den Dikken (ed.), The Cambridge Handbook of Generative Syntax, 425–457. Cambridge: Cambridge University Press. Rizzi, Luigi and Giuliano Bocci. 2017. Left Periphery of the Clause: Primarily Illustrated for Italian. In Martin Everaert and Henk C. van Riemsdijk (eds), The Wiley Blackwell Companion to Syntax, Second Edition, 1–30. Hoboken, NJ: Wiley. Rizzi, Luigi and Ian Roberts. 1989. Complex Inversion in French. Probus 1. 1–30. Roberge, Yves. 1990. The Syntactic Recoverability of Null Arguments. Montreal: McGillQueen’s University Press. Roberge, Yves and Sarah Cummins. 2006. A modular account of null objects in French. Syntax 8 (1). 44–64. Roberge, Yves and Nicole Rosen. 1999. Preposition standing and que-deletion in varieties of North American French. Linguistica Atlantica 21. 153–168. Roberts, Ian. 1993. Verbs and diachronic syntax: a comparative history of English and French. Dordrecht: Kluwer. Roberts, Ian. 1996. Remarks on the Old English C-system and the Diachrony of V2. Language Change and Generative Grammar (Linguistiche Berichte) 7. 154–167. Roberts, Ian. 2005. Principles and Parameters in a VSO Language: A Case Study in Welsh. Oxford; New York: Oxford University Press. Roberts, Ian. 2007a. Diachronic Syntax. Oxford: Oxford University Press. Roberts, Ian. 2007b. Comparative Grammar: Critical Concepts in Linguistics, vol. 2: Null Subjects. London; New York: Routledge. Roberts, Ian. 2010a. Varieties of French and the Null Subject Parameter. In Parametric Variation: Null Subjects in Minimalist Theory, 303–328. Cambridge: Cambridge University Press. Roberts, Ian. 2010b. A deletion analysis of null subjects. In Parametric Variation: Null Subjects in Minimalist Theory, 58–87. Cambridge: Cambridge University Press. Roberts, Ian. 2010c. Agreement and Head Movement. Cambridge, MA: MIT Press. Roberts, Ian. 2012a. Phases, head movement and second-position effects. In Ángel J. Gallego (ed.), Phases: Developing the Framework, 385–440. Berlin; Boston, MA: De Gruyter.

references

309

Roberts, Ian. 2012b. Diachrony and Cartography: Paths of Grammaticalization and the Clausal Hierarchy. In Laura Brugé, Anna Cardinaletti, Giuliana Giusti, Nicola Munaro, and Cecilia Poletto (eds), Functional Heads. The Cartography of Syntactic Structures, vol. 7, 351–367. Oxford: Oxford University Press. https:// doi.org/10.1093/acprof:oso/9780199746736.001.0001. Roberts, Ian. 2014a. Syntactic Change. In Andrew Carnie, Yosuke Sato, and Daniel Siddiqi (eds), The Routledge Handbook of Syntax (Routledge Handbooks in Linguistics), 391– 408. Abingdon: Routledge. Roberts, Ian. 2014b. Subject Clitics and Macroparameters. In Paola Benincà, Adam Ledgeway, and Nigel Vincent (eds), Diachrony and Dialects, 177–200. Oxford: Oxford University Press. Roberts, Ian. 2019. Parameter Hierarchies and Universal Grammar (Rethinking Comparative Syntax). Oxford: Oxford University Press. Roberts, Ian. 2021. Second positions: a synchronic analysis and some diachronic consequences. In Sam Wolfe and Christine Meklenborg (eds), Continuity and Variation in Germanic and Romance. Oxford: Oxford University Press. Roberts, Ian and Anders Holmberg. 2010. Introduction: Parameters and Linguistic Theory. In Parametric Variation: Null Subjects in Minimalist Theory, 1–57. Cambridge: Cambridge University Press. Roberts, Ian and Anna Roussou. 2002. Syntactic Change: A Minimalist Approach to Grammaticalization. Cambridge: Cambridge University Press. Roehrs, Dorian and Marie Labelle. 2003. The Left Periphery in Child French: Evidence for a Simply-Split CP. In Joseph Quer (ed.), Romance Languages and Linguistic theory 2001: Selected Papers from ‘Going Romance’ Amsterdam, 6–8 December 2001, 279–294. Amsterdam: John Benjamins. Ro¨gnvaldsson, Eirı´kur and Ho¨skuldur Thra´insson. 1990. On Icelandic word order once more. In Joan Maling and A. Zaenen (eds), Modern Icelandic Syntax (Syntax and Semantics 24), 3–40. New York: Academic Press. Ronjat, Jules. 1913. Essai de syntaxe des parlers provençaux modernes. Mâcon: Protat. Ross, J. R. 1973. The Penthouse Principle and the order of constituents. In Claudia Corum, Thomas C. Smith-Stark, and Ann Weiser (eds), 397–422. Chicago: Chicago Linguistic Society. Rouveret, Alain. 2004. Les clitiques pronominaux et la périphérie gauche en ancien français. Bulletin de la Société de linguistique de Paris 99(1). 181–237. Rowlett, Paul. 1998. Sentential Negation in French. New York: Oxford University Press. Rowlett, Paul. 2007. The Syntax of French (Cambridge Syntax Guides). Cambridge, UK; New York: Cambridge University Press. Sabel, Joachim and Mamoru Saito (eds). 2005. The Free Word Order Phenomenon: Its Syntactic Sources and Diversity. Berlin: De Gruyter. Salvesen, Christine. 2011. Stylistic Fronting and Remnant Movement in Old French. In Janine Berns, Haike Jacobs, and Tobias Scheer (eds), Romance Languages and Linguistic Theory 2009: Selected papers from ‘Going Romance’ Nice 2009, 323–342. Amsterdam: John Benjamins. Salvesen, Christine. 2013. Topics and the Left Periphery: A comparison of Old French and Modern Germanic. In Terje Lohndal (ed.), In Search of Universal Grammar: From Old Norse to Zoque, 131–172. Amsterdam: John Benjamins. Salvesen, Christine. 2014. Le complémenteur que et la péripherie gauche: analyse diachronique. Syntaxe et Sémantique 15. 47–80.

310

references

Salvesen, Christine and Kristin Bech. 2014. Postverbal Subjects in Old English and Old French. Oslo Studies in Language 6(1). 201–228. Salvesen, Christine and George Walkden. 2017. Diagnosing embedded V2 in Old English and Old French. In Robert Truswell and Éric Mathieu (eds), Micro-change and Macro-change in Diachronic Syntax (Oxford Studies in Diachronic and Historical Linguistics), 168–181. Oxford: Oxford University Press. Salvi, Giampaolo. 1991. Defesa e illustrazione della legge di Wackernagel applicata alle lingue romanze antiche: la posizione delle forme pronominali clittiche. In Laura Vanelli and Alberto Zamboni (eds), Per Giovan Battista Pellegrini. Scritti degli allievi padovani, 439–462. Padova: Unipress. Salvi, Giampaolo. 2000. La formazione della sistema V2 delle lingue romanze antiche. Lingua e Stile 35. 665–692. Salvi, Giampaolo. 2004. La formazione della struttura di frase romanza: ordine delle parole e clitici dal latino alle lingue romanze antiche. Tu¨bingen: Max Niemeyer Verlag. Salvi, Giampaolo. 2005. Some firm points on Latin word order: The left periphery. In Katalin É. Kiss (ed.), Universal Grammar in The Reconstruction of Ancient Languages, 429–457. Berlin; New York: De Gruyter. Salvi, Giampaolo. 2011. Morphosyntactic persistence. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds), The Cambridge History of the Romance Languages, 155–215. Cambridge: Cambridge University Press. Salvi, Giampaolo. 2012. On the Nature of the V2 System of Medieval Romance. In Laura Brugè, Anna Cardinaletti, Giuliana Giusti, Nicola Munaro, and Cecilia Poletto (eds), Functional Heads The Cartography of Syntactic Structures, vol. 7, 103–111. Oxford; New York: Oxford University Press. Salvi, Giampaolo. 2016. Word order. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 997–1012. Oxford: Oxford University Press. Sankoff, Gillian and Pierrette Thibault. 1977. L’alternance entre les auxiliaires avoir et être en français parlé à Montréal. Langue Francaise 34. 81–108. Sankoff, Gillian and D. Vincent. 1977. L’emploi productif du ne dans le français parlé à Montréal. Français moderne 45. 243–256. Santorini, Beatrice. 1989. The Generalization of the Verb-Second Constraint in the History of Yiddish. University of Pennsylvania PhD dissertation. Santorini, Beatrice. 1994. Some similarities and differences between Icelandic and Yiddish. In David Lightfoot and Norbert Hornstein (eds), Verb Movement, 87–106. Cambridge: Cambridge University Press. Sauzet, Patrick. 1989. Topicalisation et prolepse en occitan. Revue des langues romanes 93. 235–273. Scheer, Tobias. 2020. Consonnes en coda (_.C). In Christiane Marchello-Nizia, Bernard Combettes, Sophie Prévost, and Scheer Tobias (eds), Grande Grammaire Historique du Français, vol. 1, 387–399. Berlin: De Gruyter. Schifano, Norma. 2015a. Verb movement: A pan-Romance investigation. University of Cambridge PhD Dissertation. Schifano, Norma. 2015b. Il posizionamento del verbo nei dialetti romanzi d’Italia. The Italianist 35(1). 121–138. https://doi.org/10.1179/0261434014Z.000000000111. Schifano, Norma. 2018. Verb Movement in Romance: A Comparative Study. Oxford: Oxford University Press. Schlegel, August W. von. 1818. Observations sur la langue et la littérature provençales. Paris: Librarie Grecque-Latine-Allemande.

references

311

Schøsler, Lene. 2000a. La déclinaison bicasuelle de l’ancien français. Études romanes 47. 105–129. Schøsler, Lene. 2000b. Le statut de la forme zéro du complément d’objet direct en français moderne. Études romanes 48. 105–129. Schøsler, Lene. 2013. The Development of the Declension System. In Deborah L. Arteaga (ed.), Research on Old French: The State of the Art (Studies in Natural Language and Linguistic Theory), 167–186. Dordrecht: Springer Netherlands. Schwegler, Armin. 1986. Analyticity and Syntheticity: A Diachronic Perspective with Special Reference to Romance Languages. Berlin: De Gruyter. Scrivner, Olga. 2014. L’ordre des mots et la place de l’objet direct dans les phrases infinitivales en bas latin et en ancien français. Presented at the Société Internationale de Diachronie du Français. L’histoire du français: nouvelles approches, nouveaux terrains, nouveaux traitements. Shaw, Wyn. 2020. Si and Information Structure in Old French. University of Oxford Masters thesis. Sheehan, Michelle. 2006. The EPP and null subjects in Romance. University of Newcastle PhD dissertation. Sheehan, Michelle. 2010. ‘Free’ Inversion in Romance and the Null Subject Parameter. In Parametric variation: Null Subjects in Minimalist Theory, 231–262. Cambridge: Cambridge University Press. Sheehan, Michelle. 2016. Subjects, null-subjects and expletives in Romance. In Susann Fischer and Christoph Gabriel (eds), Manual of Grammatical Interfaces in Romance (Manuals of Romance Linguistics 10), 362–392. Berlin: De Gruyter. Sheehan, Michelle, Theresa Biberauer, Ian Roberts, and Anders Holmberg (eds). 2017. The Final-Over-Final Condition: A Syntactic Universal. Cambridge, MA: MIT Press. Shlonsky, Ur. 2014. Subject Positions, Subject Extraction, EPP, and the Subject Criterion. In Enoch Oladé Aboh, Maria Teresa Guasti, and Ian Roberts (eds), Locality, 58–85. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199945269.003.0003. Sideltsev, Andrej and Maria Molina. 2015. Enclitic -(m)a ‘but’, clause architecture and the prosody of focus in Hittite. Indogermanische Forschungen 120(1). https:// doi.org/10.1515/if-2015-0011. Sigurdsson, Halldór Ármann. 1993. Argument-drop in old Icelandic. Lingua 89(2–3). 247– 280. https://doi.org/10.1016/0024-3841(93)90054-Z. Sigurdsson, Halldór Ármann. 2011. Conditions on Argument Drop. Linguistic Inquiry 42(2). 267–304. https://doi.org/10.1162/LING_a_00042. Siller-Runggaldier, Heidi. 1993. Caratteristiche della frase interrogativa a soggetto inverso nel Ladino Centrale. In Ramón Lorenzo Va´zquez (ed.), Actas do XIX Congreso Internacional de Lingu¨´ıstica e Filoloxı´a Roma´nicas, IV: Dialectoloxı´a e xeografı´a lingu¨´ıstica/onoma´stica, 289–295. Corunna: Fundación ‘Pedro Barrié de la Maza’, Conde de Fenosa. Simonenko, Alexandra, Benoît Crabbé, and Sophie Prévost. 2019. Agreement Syncretisation and the Loss of Null Subjects. Language Variation and Change 31(3). 275–301. Simonenko, Alexandra and Paul Hirschbu¨hler. 2012. Placement de clitiques dans les propositions V1 et évolution de la structure de la proposition en ancien français. In Monique Dufresne (ed.), Typologie, ordre des mots et groupe verbal en français médiéval, 11–53. Laval, Québec: Les Presses de l’Université Laval. Sitaridou, Ioanna. 2011. Word order and information structure in Old Spanish. Catalan Journal of Linguistics 10. 159–184.

312

references

Skårup, Povl. 1975. Les premières zones de la proposition en ancien français. Copenhagen: Akademisk Forlag. Smith, John Charles. 2002. Middle French: When? What? Why? Language Sciences 24(3–4). 423–445. https://doi.org/10.1016/S0388-0001(01)00042-0. Smith, John Charles. 2011. Change and continuity in form-function relationships. In Martin Maiden, John Charles Smith, and Adam Ledgeway (eds), The Cambridge History of the Romance Languages, 268–317. Cambridge: Cambridge University Press. Smith, John Charles. 2016. French and northern Gallo-Romance. In Adam Ledgeway and Martin Maiden (eds), The Oxford Guide to the Romance Languages, 292–318. Oxford: Oxford University Press. Smith, John Charles. 2020. Old, Middle, and Modern: temporality and typology. In Keith Allan (ed.), Dynamic Language Changes: Looking Within and Across Languages., 183–200. Singapore: Springer Nature. Spevak, Olga. 2004. Verb-subject order in Latin: The case of existential and locative sentences. Classica et Mediaevalia. Revue danoise de philologie et d’histoire 55. 381–396. Spevak, Olga. 2008. The position of Focus constituents in Latin: A comparison between Latin and Czech. Journal of Pragmatics 40(1). 114–126. https://doi.org/10.1016/ j.pragma.2007.09.003. Spevak, Olga. 2010. Constituent Order in Classical Latin Prose. Amsterdam; Philadelphia, PA: John Benjamins. Spevak, Olga. 2014. The Noun Phrase in Classical Latin Prose. Leiden: Brill. Spillebout, Gabriel. 1985. Grammaire de la langue française du XVIIe siècle. Paris: Picard. Sportiche, Dominique. 1998. Partitions and Atoms of Clause Structure: Subjects, Agreement, Case and Clitics. London: Routledge. Sportiche, Dominique. 1999. Subject Clitics in French and Romance Complex Inversion and Clitic Doubling. In Kyle Johnson and Ian Roberts (eds), Beyond Principles and Parameters: Essays in Memory of Osvaldo Jaeggli, 189–221. Dordrecht: Springer Netherlands. http://dx.doi.org/10.1007/978-94-011-4822-1_8. Steiner, B. Devan. 2014. The Role of Information Structure in the Loss of Verb-Second in the History of French. University of Indiana, Bloomington PhD thesis. Tailleur, Sandrine. 2013. The French Wh Interrogative System. Est-ce que, Clefting? University of Toronto doctoral thesis. Talavera, Francisco. 1981. Aspectos vulgares de la Vetus Latina: Ana´lisis especial del orden de palabras en el libro de Rut. Analecta malacitana 4. 211–227. Terry, Robert. 1970. Contemporary French Interrogative Structures. Montreal: Editions Cosmos. Tescari Neto, Aquiles. 2012. On Verb Movement in Brazilian Portuguese: A Cartographic Study. Venice: Ca Foscari doctoral thesis. Thra´insson, Ho¨skuldur. 2007. The Syntax of Icelandic. Cambridge: Cambridge University Press. Thurneysen, Rudolf. 1892. Die Stellung des Verbums in Altfranzo¨sischen. Zeitschrift fu¨r Romanische Philologie 16(1–4). 289–307. Tobler, Adolf. 1875. Review of J. Le Coultre. De l’ordre des mots dans Chrétien de Troyes. Vermischte Beitra¨ge zur franzo¨sischen Grammatik 5. 395–414. Tomaselli, Alessandra. 1995. Cases of verb third in Old High German. In Ian Roberts and Adrian Battye (eds), Clause Structure and Language Change, 345–369. Oxford: Oxford University Press.

references

313

Travis, Lisa. 1984. Parameters and effects of word order variation. Cambridge, MA: Massachusetts Institute of Technology PhD dissertation. Troberg, Michelle. 2004. Topic-comment resumptive pronouns in Modern French and Old and Middle French. Toronto Working Papers in Linguistics 23(0). Tubau, Susagna. 2020. Minimizers and Maximizers as Different Types of Polarity Items. In Viviane Déprez and M. Teresa Espinal (eds), The Oxford Handbook of Negation, 407–426. Oxford: Oxford University Press. Uriagereka, Juan. 1995. An F position in Western Romance. In Katalin É. Kiss (ed.), Discourse Configurational Languages, 153–175. Oxford: Oxford University Press. Va¨a¨na¨nen, Veikko. 1987. Le journal-épître d’Égérie (Itinerarium Egeriae): étude linguistique. Helsinki: Suomalainen Tiedeakatemia. Vallduvı´, E. 1993. A preverbal landing site for quantificational operators. Catalan Working Papers in Linguistics 2. 319–343. Vallduvı´, Enric. 2002. L’oració com a unitat informativa. In Joan Solà, Maria-Rosa Lloret, Joan Mascarò, and Manuel Pérez-Saldanya (eds), Gramàtica del català contemporani, 1221–1279. Barcelona: Editorial Empúries. Valois, Daniel. 1991. The internal syntax of DP and adjective placement in French and English. North East Linguistics Society 21(26). 361–381. Van Gelderen, Elly. 2008a. Where did Late Merge go? Grammaticalization as feature economy. Studia Linguistica 62(3). 287–300. Van Gelderen, Elly. 2008b. Negative cycles. Linguistic Typology 12(2). Van Gelderen, Elly. 2009a. Feature Economy in the Linguistic Cycle. In Paola Crisma and Giuseppe Longobardi (eds), Historical Syntax and Linguistic Theory, 93–109. Oxford: Oxford University Press. Van Gelderen, Elly (ed.). 2009b. Cyclical Change. Amsterdam: John Benjamins. Van Gelderen, Elly. 2011. The Linguistic Cycle: Language Change and the Language Faculty. New York: Oxford University Press. Van Gelderen, Elly (ed.). 2016. Cyclical Change Continued. Amsterdam: John Benjamins. Van Kemenade, Ans and Christine Meklenborg. 2021. Issues in the left periphery of Old French and Old English: topic types and the V2 constraint. In Sam Wolfe and Christine Meklenborg (eds), Continuity and Variation in Germanic and Romance. Oxford: Oxford University Press. Vance, Barbara. 1987. The Evolution of Prodrop in Medieval French. Manuscript, Cornell. Vance, Barbara. 1988. L’évolution de pro-drop en français médiéval. Revue québécoise de linguistique théorique et appliquée 7. 85–109. Vance, Barbara. 1993. Verb-first declaratives introduced by et and the position of pro in old and middle French. Lingua 89(2–3). 281–314. https://doi.org/10.1016/00243841(93)90055-2. Vance, Barbara. 1995. On the decline of verb movement to Comp in Old and Middle French. In Adrian Battye and Ian Roberts (eds), Clause Structure and Language Change, 173–199. Oxford: Oxford University Press. Vance, Barbara. 1997. Syntactic Change in Medieval French. Dordrecht: Kluwer. Vance, Barbara, Bryan Donaldson, and B. Devan Steiner. 2009. V2 loss in Old French and Old Occitan: The role of fronted clauses. In Sonia Colina, Antxon Olarrea, and Ana Maria Carvalho (eds), Romance Linguistics 2009. Selected papers from the 39th Linguistic Symposium on Romance Languages (LSRL), Tucson, Arizona (Current Issues in Linguistic Theory 315), 301–320. Amsterdam: John Benjamins.

314

references

Vanelli, Laura. 1986. Strutture tematiche in italiano antico. In Harro Stammerjohann (ed.), Tema-Rema in Italiano, 249–273. Tu¨bingen: Gunter Narr Verlag. Vanelli, Laura. 1987. I pronomi soggetto nei dialetti italiani settentrionali dal Medio Evo ad oggi. Medioevo Romanzo XIII, 173–211. Vanelli, Laura, Lorenzo Renzi, and Paola Benincà. 1986. Tipologia dei pronomi soggetto nelle lingue romanze medievali. Quaderni Patavini di Linguistica 5. 49–66. Vasseur, Gaston. 1996. Grammaire des parlers picards du Vimeu (Somme) avec considération spéciale du dialecte de Nibas. Abbeville: F. Paillart. Vennemann, Theo. 1974. Topics, subjects and word order: from SOV to SVO via TVX. In John M. Anderson and Charles Jones (eds), Historical Linguistics: Proceedings of the First International Conference on Historical Linguistics. Amsterdam: North Holland. Vikner, Sten. 1995. Verb Movement and Expletive Subjects in the Germanic Languages. New York: Oxford University Press. Villa-Garcı´a, Julio. 2012a. Characterizing Medial and Low Complementizers in Spanish: Recomplementation que and ‘jussive/optative’ que. In Melvin Gonza´lez-Rivera (ed.), Current Formal Aspects of Spanish Syntax and Semantics, 198–228. Newcastle upon Tyne: Cambridge Scholars Publishing. Villa-Garcı´a, Julio. 2012b. Recomplementation and locality of movement in Spanish. Probus 24(2). 257–314. Villa-Garcı´a, Julio. 2019. Recomplementation in English and Spanish: Delineating the CP space. Glossa 4(1). 56. Vincent, Nigel. 1982. The development of the auxiliaries habere and esse in Romance. In Nigel Vincent and Martin Harris (eds), Studies in the Romance Verb: Essays Offered to Joe Cremona on the Occasion of his 60th Birthday, 71–96. London: Croom Helm. Vincent, Nigel. 1987. The interaction of periphrasis and inflection: some Romance examples. In Martin Harris and Paolo Ramat (eds), The Historical Development of Auxiliaries, 237–236. Berlin: De Gruyter. Vincent, Nigel. 1999. The evolution of C-structure: prepositions and PPs from IndoEuropean to Romance. Linguistics 37. 1111–1153. Vincent, Nigel. 2000. Competition and Correspondence in Syntactic Change: Null Arguments in Latin and Romance. In Susan Pintzuk, George Tsoulas, and Anthony Warner (eds), Diachronic Syntax: Models and Mechanisms, 25–51. Oxford: Oxford University Press. Vincent, Nigel and Kersti Bo¨rjars. 2010. Complements of adjectives: a diachronic approach. In Miriam Butt and Tracy Holloway King (eds), Proceedings of the LFG2010 Conference, Carleton University, 127–148. Stanford, CA: CSLI. Vinet, Marie-Thérèse. 2004. -Tu in Quebec French as a (super)positive marker. In Reineke Bok-Bennema, Bart Hollebrandse, Brigitte Kampers-Manhe, and Petra Sleeman (eds), Romance Languages and Linguistic Theory 2002: Selected Papers from ‘Going Romance’ Groningen 28–30 November 2002, 235–253. Amsterdam: John Benjamins. Viti, Carlotta. 2014. Reconstructing syntactic variation in Proto-Indo-European. IndoEuropean Linguistics 2. 1–39. Wackernagel, Jacob. 1892. Über ein Gesetz der indogermanischen Wortstellung. Indogermanische Forschungen 1. 333–436. Wagner, René-Louis and Jacqueline Pinchon. 1962. Grammaire du français classique et moderne. Paris: Hachette. Wagner, Suzanne Evans and Gillian Sankoff. 2011. Age grading in the Montréal French inflected future. Language Variation and Change 23(3). 275–313. https://doi.org/10.1017/S0954394511000111.

references

315

Walkden, George. 2013a. The correspondence problem in syntactic reconstruction. Diachronica 30(1). 95–122. https://doi.org/10.1075/dia.30.1.04wal. Walkden, George. 2013b. Null subjects in Old English. Language Variation and Change 25(02). 155–178. https://doi.org/10.1017/S0954394513000070. Walkden, George. 2014. Syntactic Reconstruction and Proto-Germanic. New York: Oxford University Press. Walkden, George. 2015. Verb-third in early West Germanic: a comparative perspective. In Theresa Biberauer and George Walkden (eds), Syntax Over Time, 236–248. Oxford; New York: Oxford University Press. Walkden, George. 2017. Language contact and V3 in Germanic varieties new and old. Journal of Comparative Germanic Linguistics 20(1). 49–81. Walkden, George and Hannah Booth. 2020. Reassessing the historical evidence for general embedded Verb Second. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 536–555. Oxford: Oxford University Press. Wanner, Dieter. 1981. Surface Complementizer Deletion. Italian che = ø. Journal of Italian Linguistics 6(1). 45–83. Wanner, Dieter. 1987. The Development of Romance Clitic Pronouns: From Latin to Old Romance. Berlin; New York: De Gruyter. Wartburg, Walther von. 1950. Dis Ausliederung der romanischen Sprachra¨ume. Bern: Francke. Wartburg, Walther von. 1958. Évolution et structure de la langue française. 5th edition. Bern: Francke. Weerman, Fred. 1989. The V2 conspiracy: A synchronic and a diachronic analysis of verbal positions in Germanic languages. Dordrecht: Foris. Westergaard, Marit. 2008. Acquisition and change: On the robustness of the triggering experience for word order cues. Lingua 118(12). 1841–1863. https://doi.org/10.1016/j.lingua.2008.05.003. Westergaard, Marit. 2009. The Acquisition of Word Order: Micro-cues, information structure, and economy. Amsterdam: John Benjamins. Wiklund, Anna-Lena, Kristine Bentzen, Gunnar Hrafn Hrafnbjargarson, and Þorbjo¨rg Hróarsdóttir. 2009. On the distribution and illocution of V2 in Scandinavian thatclauses. Lingua (Twists of Mood: The Distribution and Interpretation of indicative and subjunctive) 119(12). 1914–1938. https://doi.org/10.1016/j.lingua.2009.03.006. Wiklund, Anna-Lena, Gunnar Hrafn Hrafnbjargarson, Kristine Bentzen, and Þorbjo¨rg Hróarsdóttir. 2007. Rethinking Scandinavian verb movement. The Journal of Comparative Germanic Linguistics 10(3). 203–233. https://doi.org/10.1007/s10828-007-9014-9. Williams, Edwin. 1974. Rule ordering in syntax. Cambridge, MA: Massachusetts Institute of Technology PhD dissertation. Willis, David, Christopher Lucas, and Anna Breitbarth (eds). 2013a. The History of Negation in the Languages of Europe and the Mediterranean: Volume I Case Studies. Oxford: Oxford University Press. Willis, David, Christopher Lucas, and Anne Breitbarth. 2013b. Comparing diachronies of negation. In David Willis, Christopher Lucas, and Anne Breitbarth (eds), The History of Negation in the Languages of Europe and the Mediterranean: Volume I Case Studies, 1–50. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199602537.003.0001. Willis, David W. E. 1998. Syntactic Change in Welsh: A Study of the Loss of Verb-Second. Oxford; New York: Clarendon Press. Willis, David W. E. 2007. Historical syntax. In Robert Borsley, Maggie Tallerman, and David Willis (eds) The Syntax of Welsh, 286–337. Cambridge: Cambridge University Press.

316

references

Wilmet, Marc. 1970. Le système de l’indicatif en moyen français. Geneva: Droz. Wolf, Lothar. 1983. Le français régional d’Alsace. Étude critique des alsacianismes. Paris: Klincksieck. Wolf, Lothar. 2000. Le français en Alsace. In Gérald Antoine and Bernard Cerquiglini (eds), Histoire de la langue française 1945–2000, 687–700. Paris: CNRS. Wolfe, Sam. 2015a. Medieval Sardinian: New Evidence for Syntactic Change from Latin to Romance. In Dag T. T. Haug (ed.), Historical Linguistics 2013: Selected papers from the 21st International Conference on Historical Linguistics, Oslo, 5–9 August 2013 (Current Issues in Linguistic Theory 334), 303–324. Amsterdam; Philadelphia, PA: John Benjamins. Wolfe, Sam. 2015b. Microvariation in Medieval Romance Syntax: A Comparative Study. University of Cambridge PhD thesis. Wolfe, Sam. 2015c. The nature of Old Spanish verb second reconsidered. Lingua 164, Part A. 132–155. https://doi.org/10.1016/j.lingua.2015.06.007. Wolfe, Sam. 2015d. The Old Sardinian Condaghes. A Syntactic Study. Transactions of the Philological Society 113(2). 177–205. https://doi.org/10.1111/1467-968X.12046. Wolfe, Sam. 2015e. Syntactic Change from Latin to Romance and Phase-Edge Properties: A New Approach. Presented at the 15th International Conference of the Department of Linguistics, University of Bucharest: Workshop on Romance Syntax in Synchrony and Diachrony, Bucharest. Wolfe, Sam. 2015f. Microvariation in Old Italo-Romance Syntax: The View from Old Sardinian and Old Sicilian. Archivio Glottologico Italiano 100(1). 3–36. Wolfe, Sam. 2016a. A comparative perspective on the evolution of Romance clausal structure. Diachronica 33(4). 461–502. https://doi.org/10.1075/dia.33.4.02wol. Wolfe, Sam. 2016b. On the Left Periphery of V2 Languages. Rivista di Grammatica Generativa: Selected Papers from the 41st Incontro di Grammatica Generativa (38), 287–310. Wolfe, Sam. 2017. Syntactic Variation in Two Sister Languages: A Study of Word Order in Old French and Old Occitan. In Gabriela Pană Dindelegan, Adina Dragomirescu, Irina Nicula, and Alexandru Nicolae (eds), Comparative and Diachronic Perspectives on Romance Syntax, 53–85. Newcastle: Cambridge Scholars Press. Wolfe, Sam. 2018a. Verb Second in Medieval Romance (Oxford Studies in Diachronic and Historical Linguistics). Oxford: Oxford University Press. Wolfe, Sam. 2018b. Probing the Syntax of a Problematic Particle: Old French ‘si’ revisited. Transactions of the Philological Society 116(3). 332–362. https://doi.org/10.1111/1467968X.12123. Wolfe, Sam. 2019. Redefining the typology of V2 languages: the view from Medieval Romance and beyond. (ed.) Christine M. Salvesen. Linguistic Variation 19(1. Special Issue: A Micro-Perspective on V2 in Germanic and Romance). 16–46. Wolfe, Sam. 2020a. Old French SI, Grammaticalization and the Interconnectedness of Change. In Bridget Drinka (ed.), Historical Linguistics 2017: Selected papers from the 23rd International Conference on Historical Linguistics, 254–271. Amsterdam: John Benjamins. Wolfe, Sam. 2020b. Reconsidering Variation and Change in the Medieval French subject system. Glossa 5(1). 1–29. Wolfe, Sam. 2020c. Rethinking Medieval Romance Verb Second. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second, 348–367. Oxford: Oxford University Press. Wolfe, Sam. 2020d. Old Gallo-Romance, Periodization and the Left Periphery. In Sam Wolfe and Martin Maiden (eds), Variation and Change in Gallo-Romance Grammar, 9–40. Oxford: Oxford University Press.

references

317

Wolfe, Sam. 2021. Microvariation in the Old Italo-Romance Left Periphery. The Case of SI. Revue roumaine de linguistique 66. Wolfe, Sam. In prep. On the loss of OV in the history of French. Unpublished manuscript. Wolfe, Sam and Martin Maiden. 2020. Variation and Change in Gallo-Romance Grammar. Oxford: Oxford University Press. Wolfe, Sam and Rebecca Woods. 2020. Introduction. In Rebecca Woods and Sam Wolfe (eds), Rethinking Verb Second. Oxford: Oxford University Press. Woods, Rebecca. 2014. The syntax of orientation shifting: Evidence from English high adverbs. In ConSOLE XXII: Proceedings of the 22nd Conference of the Student Organization of Linguistics in Europe (8–10 January 2014, Lisbon), vol. 22, 205–230. Leiden: the Leiden University Centre for Linguistics. https://www.universiteitleiden.nl/binaries/content/assets/geesteswetenschappen/lucl/ sole/console-xxii.pdf (21 April 2017). Woods, Rebecca and Sam Wolfe (eds). 2020. Rethinking Verb Second. Oxford: Oxford University Press. Wright, Roger. 2013. Periodization. In Martin Maiden, J. C. Smith, and Adam Ledgeway (eds), The Cambridge History of the Romance Languages, Volume II, Contexts, 107–124. Cambridge: Cambridge University Press. Wurff, Wim van der. 1996. Syntactic reconstruction and reconstructibility: Proto-IndoEuropean and the typology of null objects. In Jacek Fisiak (ed.), Linguistic Reconstruction and Typology, 337–356. The Hague: De Gruyter. Yang, Charles. 2003. Knowledge and Learning in Natural Language. Oxford: Oxford University Press. Yang, Charles. 2004. Universal grammar, statistics or both? Trends in Cognitive Sciences 8. 451–456. Zafiu, Rodica. 2013. Information structure. In Panâ Dindelegan (ed.), The Grammar of Romanian, 568–575. Oxford: Oxford University Press. Zagona, Karen. 1988. Verb Phrase Syntax: A Parametric Study of English and Spanish. Dordrecht: Kluwer. Zagona, Karen. 2001. The Syntax of Spanish. Cambridge University Press. Zanuttini, Raffaella. 1997a. Syntactic Properties of Sentential Negation. Oxford: Oxford University Press. Zanuttini, Raffaella. 1997b. Negation and Clausal Structure: A Comparative Study of Romance Languages. New York: Oxford University Press. Zaring, Laurie. 1998. Object shift in Old French. In Armin Schwegler, Bernard Tranel, and Myriam Uribe-Etxebarria (eds), Romance Linguistics: Theoretical Perspectives, 319–333. Amsterdam: John Benjamins. Zaring, Laurie. 2010. Changing from OV to VO: More evidence from Old French. Revista Philologica Romanica 10. 1–18. Zaring, Laurie. 2011. On the nature of OV and VO order in Old French. Lingua 121(12). 1831–1852. https://doi.org/10.1016/j.lingua.2011.07.008. Zimmermann, Michael. 2014. Expletive and Referential Subject Pronouns in Medieval French (Linguistiche Arbeiten 556). Berlin; New York: De Gruyter. Zink, Gaston. 1997. Morphologie du français médiéval. Paris: Presses Universitaires de France. Zribi-Hertz, Anne. 1994. La syntaxe des clitiques nominatifs en français standard et en français avancé. Travaux de linguistique et de philologie 32. 131–147. Zubizarreta, Maria Luisa. 1998. Prosody, Focus, and Word Order. Cambridge, MA: MIT Press.

318

references

Zubizarreta, Maria Luisa. 2001. The constraint on preverbal subjects in Romance interrogatives: a minimality effect. In Aafke Hulk and Jean-Yves Pollock (eds), Subject Inversion in Romance and the Theory of Universal Grammar, 184–205. Oxford: Oxford University Press. Zwart, C. Jan-Wouter. 1993. Dutch Syntax: A Minimalist Approach. University of Groningen PhD dissertation.

Index Adjective 13, 24, 42, 75, 83, 88, 89, 90, 233, 260–261 predicative 74–75, 79, 83 Adverbial 42, 44, 74–75, 82–83, 90, 125, 129–131, 246 scene-setting 33, 37, 59, 77, 86, 116–117 Analyticity 7–18, 272 Anatolian 141, 250 Articles 8, 12–15 Auxiliaries 8, 15, 136–137, 249, 254

Eastern Lombard 150 Edge Feature 36–37, 46, 72, 105, 109, 145, 224, 241–242, 244–245, 247, 250, 254, 257, 260 Edge-Fronting 43, 45, 92, 247, 249 Embedded clauses 24, 45–46, 49, 77–81, 86–88, 100, 106–107, 122–123, 170–171, 184–185, 193–195, 217, 243–244, 247 Enclisis 105, 115–117, 119–120, 139 Extended Projection Principle 148, 155

Base de Français Médiéval 4 Belletti, Adriana 142, 158, 215 Brunot, Ferdinand 2

Feature Economy 140–141, 252, 254–256, 264 Fin 35, 45–47, 53, 58, 72, 87, 98, 104–105, 113–118, 123–125, 140, 142–143, 184, 224, 243–244, 247, 248, 254, 256–257, 264, 271 Final-Over-Final Condition 214–216 Focus 29–32, 38–42, 49, 58–59, 73, 82–84, 91, 102–103, 105, 114, 118, 125, 142, 153, 156, 174, 177, 219, 225–232, 243–246, 247, 252, 254, 260, 263, 265–266, 271 contrastive 29–30, 32, 41, 53, 82, 247, 252, 255 information 29–32, 40, 51–52, 82, 143–144, 153, 160, 244, 252, 255, 265–266 mirative 30 Force 35, 46, 53, 56, 58, 72, 87, 98, 103, 113–119, 137, 140–145, 244–245, 247, 250, 256–257 Frame-setters 33–34, 55, 58, 86, 116–117, 125, 243–244 Francoprovençal 22 ‘Free’ word order 22, 24, 39, 100, 216 French Acadian 128. See also Canadian French. Alsace 238 Belgian 84, 128, 137, 139

Cardinaletti, Anna 135–136, 207 Cartography 4, 95, 97–98, 104, 250 Case 7–8, 11 Catalan 96, 156, 265–272 Celtic 38, 99, 113, 141, 143, 250 Chinese 99 Chomsky, Noam 3, 148, 172, 208, 214, 233 Cinque, Guglielmo 129, 130, 259–260 Classification 5, 250, 265–273 Cleft 92, 199, 267 Clitic climbing 150, 170–171 Complementizer deletion 87, 93 Complementizer doubling See Recomplementation Complementizers 35–36, 45, 61, 63, 78, 88, 93, 97, 100, 150, 257, 266, 271 Copula fronting 47 Corsican 151 Degree-zero learnability 257 Discourse pragmatics 4, 24, 243–247. See also Information Structure. Discourse-configurationality 27

320

index

Canadian 17, 87, 128, 137, 139, 196, 198, 245 Channel Island 17 Child 88 Classical 26, 72–77, 132–134, 145, 185–195, 204, 236–237, 240, 245, 247 Colloquial 88, 91–92, 140, 196–197, 251, 255–259, 266 Maghreb 198 Middle 18, 21, 26, 64–70, 77–81, 120–125, 144–145, 175–186, 204, 240, 245, 256, 264, 271 Modern 18, 24, 30, 73, 82–88, 91, 94–95, 132–134, 156, 195–202, 247, 250, 258–259, 261, 264–267 North American 87 Old 12–14, 18–21, 25, 31, 34–36, 49–64, 107–118, 143–145, 161–174, 203, 223–235, 240, 244, 247, 254, 256, 259, 261, 271 Ontario 198. See also Canadian French. Renaissance 26, 72–77, 127–132, 145–146, 186–195, 204, 236–237, 240, 245, 247, 251 Functional categories 8–11 Future 17 German Modern 97, 144 Old High 144 Germanic 36, 38, 45, 47, 60, 96–97, 99, 110–112, 117–118, 141, 143, 159, 250 Gothic 39 Grammaticalization 11, 14, 27, 92, 149, 197, 208 Hanging topic 32–33, 44, 86 Head directionality 23, 48, 138, 143, 213–215, 220–222, 238, 240–243 Hittite 38, 159 Icelandic 60, 97, 224, 233 Imperative 102–103, 119–120, 125–126, 134, 137, 139–140, 146, 252, 254 Information structure 52–56, 67, 73, 76, 85, 144–145, 154–155, 160–169,

174–177, 180–183, 203–207, 219, 225–232, 240, 242–247, 255, 257 Input Generalization 141, 144, 254, 257 Interrogative 103, 125, 132–133, 139, 192, 245, 256, 264, 266–268, 271 Inversion See also Postverbal Subjects. complex 126–127, 132, 134–136, 199 Germanic 97, 111, 122, 145, 164–169, 180–183, 203, 244, 247 Romance 113, 164–169, 180–183, 203, 245, 247 Italian Modern 29–30, 35, 73, 95, 150–151, 155, 263, 265–272 Old 31, 52, 207–208, 210, 215–216, 221, 224–225, 235, 241, 258 Jespersen’s Cycle 18–19, 21–22, 262–264 Latin 3, 9, 14, 18–19, 37–49, 100–105, 138, 152–160, 203, 216–222, 247, 254 archaic 101, 140, 216, 240, 247 Colloquial 44, 84, 139–140, 157, 239, 256 late 46–48, 105–107, 138, 157, 221, 239, 244, 247, 254, 256 subliterary 46–47, 138, 140, 221, 239 Ledgeway, Adam 10, 101, 105–107, 113, 148, 157–158, 220 Left dislocation 24, 27, 33, 43–44, 54–56, 70, 76–77, 84–87, 125, 145, 265–266 Left periphery 29–90, 110, 120, 125, 143–145, 153–154, 161, 197, 232, 243–247, 261, 271 Leftward Stylistic Displacement See Stylistic Fronting. Lycian 38 Matrix/embedded asymmetries 25, 33–34, 42, 45, 60, 78–79, 107, 111–112, 122–123, 171–173, 185, 193–195, 217, 243–245, 247, 257 Medieval Romance 14, 25, 31–32, 34, 38, 46, 51–52, 99, 101, 117, 148, 151, 269–271 Merge vs. Move 85–86, 93, 265–268 Modern Cosentino 30

index N-to-D Movement 259–260 Ne-Deletion 19–20 Neapolitan Modern 30, 35, 149 Old 35, 149 Negation 18–22, 128, 196, 198, 262–264 Niçard 22 Nominal syntax 8, 23, 255, 258–262 Northern Italian dialects 149, 196–198, 263, 265–272 Null arguments 151, 158–160, 203, 208–213, 243, 247, 249 Null subjects 108, 150–152, 170–174, 203, 208–213, 245, 247, 252, 268, 271 loss 208–212, 251–252 Object clitics See Enclisis and Proclisis Occitan Modern 265–272 Old 31 OV 24, 45–46, 48, 99–101, 138, 142–143, 153, 148, 216–239, 243, 247 Parametric theory 3–4, 90, 146, 148, 151–152, 211, 242, 246–264, 271–272 Peregrinatio 48–49, 105, 157, 221–222 Periodization 6, 21, 64, 69–70, 74, 81, 121, 174, 222, 241, 249, 265–273 Person marking 150, 255, 260 Phase theory 138, 141–142, 146, 208, 215–216, 223, 232, 240, 250, 257–259, 261 Poletto, Cecilia 98–99, 114, 144, 208, 215, 224, 241, 258, 261 Pollock, Jean-Yves 94–95, 128 Portuguese Modern Brazilian 95, 151, 268 Modern European 150–151, 263, 265–272 Old 221 Prefield 49–51, 64–67, 72, 109–110, 116, 173, 179, 232 Prepositions 11, 23 Pro See Null Subjects. Probe-goal Agree 152, 160, 172, 184, 207, 251–257 Proclisis 115–118

321

Proto-Indo-European 37–39, 101–102, 159, 216, 240, 250 Quantifers 237–238 Quantified phrases 32, 41–42, 52–53, 58, 73, 179, 243, 247, 252, 255, 265–266 Raeto–Romance 33–34, 98, 265–272 Reanalysis 92, 101–105, 132, 138, 141–144, 150, 208, 240, 243, 254–256 Recomplementation 35–36, 63, 80–81, 87, 88, 103, 244–245, 256, 257, 266 Regional variation 2, 5, 137, 196–202, 238, 245–246, 255, 258–260 Relative clauses 35, 60, 88, 217 Remnant movement 233 Right dislocation 208, 267 Rizzi, Luigi 29, 31, 36, 87, 103–104, 151–152 Roberts, Ian 99, 126, 141, 160, 172, 184, 204, 243–244, 257 Rollup movement 99–101, 220, 239–240, 243, 251 Romanian 30, 35, 95, 265–272 Sanskrit 38–39 Sardinian 30, 265–272 Schifano, Norma 146–148, 250, 266 Scrambling 215, 223–225, 235– 237, 239–240, 243, 245, 251, 257 SI 56–59, 70–72, 78, 114–116, 205 Sicilian 82, 265–272 Sociolinguistic variation 20, 24, 44, 82–83, 87–90, 137, 140, 196–202, 222, 239, 245–246, 255, 265–266, 268 Southern Italian dialects 35, 265–272 Spanish Modern Caribbean 268 Modern Peninsular 30, 96, 150–151, 155, 263, 265–272 Old 221 Stylistic Fronting 225, 233–235 Subject agreement 150, 196–197, 201, 205, 208, 255 Subject doubling 26–27, 197–198

322

index

Subject positions 146–148, 152–155, 161–163, 175–180, 186–190, 195–199, 205–208 clitic 134–136, 149, 178–180, 189, 195–202 postverbal 23, 97, 109–111, 126, 132, 134–136, 139, 145, 149–151, 155–158, 163–170, 174, 180–183, 190–192, 199–203, 247, 266, 269 preverbal 23, 72, 78–79, 121–122, 136, 139, 152–155, 145, 174–180, 185–190, 195–199, 244 pronominal 132–136, 149, 153, 162–163,167–168, 176–177, 189, 191, 195–202, 207, 244, 252, 256 SubjP 149, 164, 190, 208, 256. See also Subject Positions. Surcomposé 9 SVO 25–27, 61–62, 73, 79, 112, 122–123, 127–128, 138, 143–146, 153, 208, 219, 247 Syntheticity 7–18 T(heme)–V–X 26, 46 Tense, Aspect, Mood 8, 260–261 Tobler–Mussafia 115, 117 Topic 32–33, 38, 43–45, 49, 53–54, 76, 91, 105, 114, 121–122, 125, 142, 153, 156, 174, 179, 197, 205, 225–232, 243–247, 252, 254–255, 265–266, 271 Topic (null) 102, 115, 119, 152, 172–174, 185–186, 205, 209–211, 244, 247, 252 Topic Continuity 56–57, 102, 159–160, 189 V–in–Situ 99–101, 138, 140–143, 146, 153, 247 V–to–C movement 46, 50, 97–99, 104–105, 121–125, 127–134, 140–141, 200, 243, 252, 256, 259, 266–267

V–to–Foc movement 102–103, 138, 140 V–to–Subj movement 136, 139, 140, 145, 190, 199–202, 207, 245, 252, 256, 264, 266–268 V-to-T movement 61, 94–96, 106, 110– 111, 123, 127–132, 140, 145–146, 192, 207, 245, 250, 252, 2666 V-to-Top movement 102–103, 119, 138, 140, 252 Vance, Barbara 51, 124, 183–184 Variationist approaches 2, 20, 85, 196–198 Verb First 61, 102, 105, 108, 110, 115, 118, 125, 139, 157, 171, 183, 244 Verb Second 25–6, 36, 47, 57, 59, 65–67, 72, 97, 105–107, 113–119, 121–125, 127, 138, 140, 143–146, 161, 184, 210, 216, 244–246, 247, 249, 260 embedded 60, 79–80, 112, 123, 244, 247 genesis 46, 49, 104–107, 113, 140, 244, 254 loss 73, 88, 127–137, 139, 144–145, 175–176, 185–190, 245 symmetrical 60, 97, 123 Verb Third 48, 52, 55, 69, 107–108, 110, 114, 116–118, 121, 125–126, 157, 244 Verbal system 8–9, 158, 196–198, 209, 260 vP Periphery 142, 149, 158, 174, 186, 201, 207, 215–217, 224–233, 235, 239–247, 254, 256, 263, 268, 271 VSO 25, 99, 106–107, 138–139, 142, 208, 244 Wackernagel elements 38–39, 42, 47, 104 Wallon 22 Wh–movement 38, 91, 105, 107, 134, 254, 256 XP–movement 36–37, 140–142, 214, 241–242, 247, 251–253, 255, 260, 271 Yiddish 60

OX F O R D S T U D I E S I N D IAC H R O N I C AND HISTORICAL LINGUISTICS General editors Adam Ledgeway and Ian Roberts, University of Cambridge Advisory editors Cynthia L. Allen, Australian National University; Ricardo Bermúdez-Otero, University of Manchester; Theresa Biberauer, University of Cambridge; Charlotte Galves, University of Campinas; Geoff Horrocks, University of Cambridge; Paul Kiparsky, Stanford University; Anthony Kroch, University of Pennsylvania; David Lightfoot, Georgetown University; Giuseppe Longobardi, University of York; George Walkden, University of Konstanz; David Willis, University of Cambridge published 1 From Latin to Romance Morphosyntactic Typology and Change Adam Ledgeway 2 Parameter Theory and Linguistic Change Edited by Charlotte Galves, Sonia Cyrino, Ruth Lopes, Filomena Sandalo, and Juanito Avelar 3 Case in Semitic Roles, Relations, and Reconstruction Rebecca Hasselbach 4 The Boundaries of Pure Morphology Diachronic and Synchronic Perspectives Edited by Silvio Cruschina, Martin Maiden, and John Charles Smith 5 The History of Negation in the Languages of Europe and the Mediterranean Volume I: Case Studies Edited by David Willis, Christopher Lucas, and Anne Breitbarth 6 Constructionalization and Constructional Changes Elizabeth Closs Traugott and Graeme Trousdale 7 Word Order in Old Italian Cecilia Poletto 8 Diachrony and Dialects Grammatical Change in the Dialects of Italy Edited by Paola Benincà, Adam Ledgeway, and Nigel Vincent

9 Discourse and Pragmatic Markers from Latin to the Romance Languages Edited by Chiara Ghezzi and Piera Molinelli 10 Vowel Length from Latin to Romance Michele Loporcaro 11 The Evolution of Functional Left Peripheries in Hungarian Syntax Edited by Katalin É. Kiss 12 Syntactic Reconstruction and Proto-Germanic George Walkden 13 The History of Low German Negation Anne Breitbarth 14 Arabic Indefinites, Interrogatives, and Negators A Linguistic History of Western Dialects David Wilmsen 15 Syntax over Time Lexical, Morphological, and Information-Structural Interactions Edited by Theresa Biberauer and George Walkden 16 Syllable and Segment in Latin Ranjan Sen 17 Participles in Rigvedic Sanskrit The Syntax and Semantics of Adjectival Verb Forms John J. Lowe 18 Verb Movement and Clause Structure in Old Romanian Virginia Hill and Gabriela Alboiu 19 The Syntax of Old Romanian Edited by Gabriela Pană Dindelegan 20 Grammaticalization and the Rise of Configurationality in Indo-Aryan Uta Reino¨hl 21 The Rise and Fall of Ergativity in Aramaic Cycles of Alignment Change Eleanor Coghill

22 Portuguese Relative Clauses in Synchrony and Diachrony Adriana Cardoso 23 Micro-change and Macro-change in Diachronic Syntax Edited by Eric Mathieu and Robert Truswell 24 The Development of Latin Clause Structure A Study of the Extended Verb Phrase Lieven Danckaert 25 Transitive Nouns and Adjectives Evidence from Early Indo-Aryan John. J. Lowe 26 Quantitative Historical Linguistics A Corpus Framework Gard B. Jenset and Barbara McGillivray 27 Gender from Latin to Romance History, Geography, Typology Michele Loporcaro 28 Clause Structure and Word Order in the History of German Edited by Agnes Ja¨ger, Gisella Ferraresi, and Helmut Weiß 29 Word Order Change Edited by Ana Maria Martins and Adriana Cardoso 30 Arabic Historical Dialectology Linguistic and Sociolinguistic Approaches Edited by Clive Holes 31 Grammaticalization from a Typological Perspective Edited by Heiko Narrog and Bernd Heine 32 Negation and Nonveridicality in the History of Greek Katerina Chatzopoulou 33 Indefinites between Latin and Romance Chiara Gianollo

34 Verb Second in Medieval Romance Sam Wolfe 35 Referential Null Subjects in Early English Kristian A. Rusten 36 Word Order and Parameter Change in Romanian A Comparative Romance Perspective Alexandru Nicolae 37 Cycles in Language Change Edited by Miriam Bouzouita, Anne Breitbarth, Lieven Danckaert, and Elisabeth Witzenhausen 38 Palatal Sound Change in the Romance Languages Diachronic and Synchronic Perspectives André Zampaulo 39 Dative External Possessors in Early English Cynthia L. Allen 40 The History of Negation in the Languages of Europe and the Mediterranean Volume II: Patterns and Processes Anne Breitbarth, Christopher Lucas, and David Willis 41 Variation and Change in Gallo-Romance Grammar Edited by Sam Wolfe and Martin Maiden 42 Phonetic Causes of Sound Change The Palatalization and Assibilation of Obstruents Daniel Recasens 43 Syntactic Features and the Limits of Syntactic Change Edited by Jóhannes Gı´sli Jónsson and Thórhallur Eythórsson 44 Romance Object Clitics Microvariation and Linguistic Change Diego Pescarini 45 The Diachrony of Differential Object Marking in Romanian Virginia Hill and Alexandru Mardale

46 Noun-based Constructions in the History of Portuguese and Spanish Patrı´cia Amaral and Manuel Delicado Cantero 47 Syntactic Change in French Sam Wolfe in preparation Redevelopment of Case Systems in Indo-Aryan Miriam Butt Functional Heads Across Time Edited by Barbara Egedi and Veronika Hegedűs Classical Portuguese Grammar and History Charlotte Galves, Aroldo de Andrade, Christiane Namiuti, and Maria Clara Paixão de Sousa Morphological Borrowing Francesco Gardani Nominal Expressions and Language Change From Early Latin to Modern Romance Giuliana Giusti A Study in Grammatical Change The Modern Greek Weak Subject Pronoun τoς and its Implications for Language Change and Structure Brian D. Joseph Reconstructing Pre-Islamic Arabic Dialects Alexander Magidow Iranian Syntax in Classical Armenian Robin Meyer