Origins of a Creole: The History of Papiamentu and Its African Ties 9781614511076, 9781614511427

Citation preview

Origins of a Creole

Language Contact and Bilingualism 3 Editor

Yaron Matras

De Gruyter Mouton

Origins of a Creole The History of Papiamentu and Its African Ties

by

Bart Jacobs

De Gruyter Mouton

ISBN 978-1-61451-142-7 e-ISBN 978-1-61451-107-6 ISSN 2190-698X Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. ” 2012 Walter de Gruyter, Inc., Boston/Berlin Cover image: Anette Linnea Rasmus/Fotolia Typesetting: PTP-Berlin Protago-TeX-Production GmbH, Berlin Printing: Hubert & Co. GmbH & Co. KG, Göttingen 앝 Printed on acid-free paper 앪 Printed in Germany www.degruyter.com

Acknowledgements

This research project was carried out at the Ludwig-Maximilians-Universität (LMU) München in partnership with the Universidade de Coimbra and was facilitated by an LMU Innovativstelle as well as by the Linguistisches Internationales PromotionsProgramm (LIPP). At the LMU Munich, I would like to thank first and foremost my Doktorvater Ulrich Detges for invaluable feedback on my dissertation and for his pleasant and intelligent guidance and advice, from start to finish, in academic and nonacademic matters. At the Department of Romance Linguistics I was furthermore very fortunate to receive advices, comments, and interesting personal communications from Thomas Krefeld and Elissa Pustka, for which many thanks! I also owe a large debt of gratitude to Stephanie Hackert from the Department of English who has been a tremendous support in the final stages of my dissertation. The Linguistisches Internationales PromotionsProgramm (LIPP) generously financed my conference participations and research activities abroad and has facilitated the most advantageous academic environment one could possibly hope for. The present study has benefited maximally from these conditions. I am very grateful to Caroline Trautmann, coordinator of the LIPP, for her support in a wide range of matters throughout my stay in Munich. At the Universidade de Coimbra, my Doktorvater John Holm has been a source of inspiration and a supporting pillar without which this book would not have seen the light of day. I would not have started my research on Papiamentu and the Portuguese-based Creoles if I hadn’t visited his seminar on contact linguistics in Coimbra in 2007. John Holm has been tremendously encouraging ever since and has provided invaluable feedback on earlier drafts of this study. I thank him also for the indispensable corrections of my English, though remaining errors and Dutchisms of course are my responsibility alone. During my stays in Coimbra, I was also very privileged to work side by side with Incanha Intumbo, Liliana Inverno, Patricia Vieira and Bernardino Tavares, and I wish to thank each one of them for their helpfulness and hospitality. I am particularly indebted to Dominika Swolkien for several personal communications and comments on an earlier draft of this study. I would like to seize the opportunity to express my gratitude to Josep Quer, Dorothée te Riele and Otto Zwartjes for their support and inspiration, which was crucial when I was a student at the Universiteit van Amsterdam and which I will not easily forget.

vi

Acknowledgements

In addition, I would like to single out Mikael Parkvall and Nicolas Quint here for being sources of extraordinary inspiration in the earliest stages of this research project and well beyond. I’d furthermore wish to thoroughly thank the following linguists for the generous sharing of their exquisite knowledge and for being supportive in many different ways: Philip Baker, Peter Bakker, Marlyse Baptista, Aymeric DavalMarkussen, Stéphane Goyette, Anthony Grant, Jürgen Lang, Philippe Maurer, John McWhorter, Pieter Muysken and Armin Schwegler. My words of thanks here are futile in comparison to the support, inspiration and generosity you have bestowed upon me. Much obliged I am also to the editor of the present series, Yaron Matras, for valuable comments and proofreading of the manuscript and for giving Papiamentu and the Portuguese-based Creoles the type of forum I believe they deserve. Many thanks also to Marcia Schwartz and Wolfgang Konwitschny from De Gruyter for their friendly guidance. I would like to mention the following friends and scholars whom I owe immense debts of gratitude for reasons too many and too diverse to enumerate: Werner Abraham, Victor Barros, Margot van den Berg, Hans den Besten (who passed away on 19 July 2010), Lucille Berry-Haseth, Noël Bernard Biagui, Harm den Boer, Geert Booij, Jo-Anne Ferreira, Cándida González-López, Toby Green, Miguel Gutiérrez Maté, Philip Havik, Silvano Rigmar Haynes, V’yacheslav Hnatyuk, Tetyana Hnatyuk, Han Jordaan, Sidney Joubert, Georg Kaiser, Wim Klooster, George Lang, Bettina Lämmle, Elisabeth Leiss, Clemens van Loyen, Ana Luís, Frank Martinus Arion, Christina Märzhäuser, Benjamin Meisnitzer, Susanne Michaelis, Matthias Perl, Sebastian Postlep, Constanza Rojas-Primus and Wolfgang Schulze. Last but certainly not least, I’d like to thank Gosia Kozyra for having been such an amazing support in the final stages of my dissertation and for sharing with me her beautiful spirit. Kocham cie˛. Loving greetings also to my parents, my sisters Fien and Sari, and my niece Carmen. The book is dedicated to the memory of Peter van Looij.

Contents

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Introduction . . . . . . . . . .

. . . . . . . . . .

1 3 3 6 8 9 10 10 11 14 14

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From Schabel (1704) to Lenz (1928) . . . . . . . . . . . . . . . . Afro-Portuguese hypotheses: from Lenz (1928) to monogenesis Spanish hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . Critical discussion of the Spanish hypotheses . . . . . . . . . . . Linguistic continuity between the pre- and post-1634 period? . Linguistic evidence against Old Spanish in PA’s superstrate . . About the tendency to attribute the Portuguese to other Hispanic varieties . . . . . . . . . . . . . . . . . . . . . . . . . . . . PA birth among the Sephardim? . . . . . . . . . . . . . . . . . . . On the linguistic profile of the early Curaçaoan Sephardim . . Demographic arguments against a PA birth among the Sephardim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Where does the Portuguese come from? . . . . . . . . . . . . . . A shared origin for all Afro-Iberian creoles in the Caribbean? . Goodman’s Brazilian Creole Hypothesis . . . . . . . . . . . . . . Gulf of Guinea Portuguese-based Creole . . . . . . . . . . . . . . Upper Guinea Portuguese-based Creole . . . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

17 17 18 20 24 27 28 29

Presentation of the languages considered in the present Papiamentu (PA) . . . . . . . . . . . . . . . . . . . . . . . Cape Verdean Creole (CV) . . . . . . . . . . . . . . . . . The Creole of Guinea-Bissau and Casamance (GBC) . Hypothesis examined in the present study . . . . . . . . Methodological remarks . . . . . . . . . . . . . . . . . . . Linguistic evidence . . . . . . . . . . . . . . . . . . . . . . Negative evidence . . . . . . . . . . . . . . . . . . . . . . . Historical evidence . . . . . . . . . . . . . . . . . . . . . . Structure of the present study . . . . . . . . . . . . . . .

study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 Critical review of the literature on the origins of Papiamentu 1.1. 1.2. 1.3. 1.4. 1.4.1. 1.4.2. 1.4.3. 1.5. 1.5.1. 1.5.2. 1.6. 1.6.1. 1.6.2. 1.6.3. 1.6.4. 1.7.

. 29 . 30 . 31 . . . . . . .

36 39 39 43 44 46 48

viii

Contents

2 Phonology 2.1. 2.1.1. 2.1.2. 2.1.3. 2.1.4. 2.2. 2.2.1. 2.2.2. 2.2.3. 2.2.4. 2.2.5. 2.3. 2.3.1. 2.3.2. 2.3.3. 2.3.4. 2.4. 2.4.1. 2.4.2. 2.5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vowel features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vowel raising . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rounding of unstressed vowels . . . . . . . . . . . . . . . . . . . . . Vowel harmony . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monophthongs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consonant features . . . . . . . . . . . . . . . . . . . . . . . . . . . . The voiceless palatal fricative /S/ in PA and Upper Guinea PC . Retention of Old Portuguese voiceless affricate /tS/ in PA and Upper Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rejection of voiced fricatives in PA and Upper Guinea PC . . . . The lack of lambdacism (/r/ > /l/) in PA and Upper Guinea PC . Rhotacism (/d/ > /r/) . . . . . . . . . . . . . . . . . . . . . . . . . . . Syllabic restructuring . . . . . . . . . . . . . . . . . . . . . . . . . . . Apheresis of prefixes . . . . . . . . . . . . . . . . . . . . . . . . . . . Vowel epenthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metathesis of the /r/ . . . . . . . . . . . . . . . . . . . . . . . . . . . Negative evidence: syllabic restructuring in PLQ and Gulf of Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paroxytonic verb stress in PA and SCV . . . . . . . . . . . . . . . Verb stress in GBC . . . . . . . . . . . . . . . . . . . . . . . . . . . . On the diachrony of paroxytonic verb stress in PA and SCV . . . Final remarks on phonology . . . . . . . . . . . . . . . . . . . . . .

3 Selected parts of speech 3.1. 3.1.1. 3.1.2. 3.1.3. 3.1.4. 3.1.5. 3.1.6. 3.2. 3.2.1. 3.2.2. 3.2.3. 3.2.4. 3.2.5.

Introduction . . . . . . . . . . . . . . . . Personal pronouns . . . . . . . . . . . . . 1sg (a)mi . . . . . . . . . . . . . . . . . . Emphatic a- subject pronouns . . . . . 2pl SCV nhos . . . . . . . . . . . . . . . Digression: 2sg polite pronouns in PA PA nan . . . . . . . . . . . . . . . . . . . Final remarks on pronouns . . . . . . . Prepositions . . . . . . . . . . . . . . . . PA / Upper Guinea PC di . . . . . . . . PA / Upper Guinea PC na . . . . . . . . PA / Upper Guinea PC te . . . . . . . . PA / Upper Guinea PC riba (di) . . . . PA / Upper Guinea PC pa . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

51 51 52 52 55 56 57 62 62 68 69 72 72 73 74 75 76 77 78 80 80 81 83 83 83 86 86 88 89 90 95 95 97 98 99 100 101

ix

Contents

3.2.6. Zero preposition with motion verb + place . . . . . . . . . . 3.2.7. Reanalysis of Iberian prepositions/adverbs ‘in front of’ and ‘behind’ as nouns . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.8. Composed prepositions . . . . . . . . . . . . . . . . . . . . . . 3.2.9. A reassessment of the time-depth of prepositions in PA . . . 3.2.10. Final remarks on prepositions . . . . . . . . . . . . . . . . . . 3.3. Interrogatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1. Equally transparent interrogative paradigms . . . . . . . . . . 3.3.2. PA: Portuguese rather than Spanish etyma . . . . . . . . . . . 3.3.3. Early PA *kantu, *kal . . . . . . . . . . . . . . . . . . . . . . . 3.3.4. PA unda, SCV unde and GBC nunde . . . . . . . . . . . . . . 3.3.5. PA / Upper Guinea PC ken . . . . . . . . . . . . . . . . . . . . 3.4. Conjunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Coordinate conjunctions . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Subordinate conjunctions . . . . . . . . . . . . . . . . . . . . . 3.4.3. Final remarks on conjunctions . . . . . . . . . . . . . . . . . . 3.5. Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1. Reciprocity and reflexivity . . . . . . . . . . . . . . . . . . . . 3.5.2. The deictic marker Early PA / Upper Guinea PC es . . . . . 3.5.3. Negation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . 103 . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

104 106 113 114 114 116 116 116 117 117 118 118 120 129 129 129 134 139

. . . . . . . .

. . . . . . . .

. . . . . . . .

143 143 144 144 145 145 149 150 151

. . . . . .

. . . . . .

. . . . . .

154 155 155 156 157 162

4 Morphology 4.1. 4.1.1. 4.1.2. 4.1.3. 4.1.4. 4.2. 4.2.1. 4.2.2. 4.3. 4.3.1. 4.3.2. 4.3.3. 4.3.4. 4.3.5. 4.3.6.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Derivational morphology . . . . . . . . . . . . . . . . . . . . . PA -mentu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PA -dó . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Upper Guinea PC -mentu / -dor . . . . . . . . . . . . . . . . . The suffix -dadi in Early PA texts . . . . . . . . . . . . . . . . Inflectional morphology . . . . . . . . . . . . . . . . . . . . . . The diachrony of PA’s past participle morpheme -/Ø/ . . . . The regularization of past participle morphology in PA and Upper Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . . Passivization in (Early) PA and Upper Guinea PC . . . . . . Passivization in present-day PA . . . . . . . . . . . . . . . . . Passivization in Upper Guinea PC . . . . . . . . . . . . . . . . Auxiliary-less passives in Early PA texts . . . . . . . . . . . . Digression: On the reliability of Early PA evangelical texts Auxiliary-less passives (/passive verbs) in present-day Papiamentu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . On the incorporation of wòrdu and ser . . . . . . . . . . . . .

. . . 163 . . . 166

x

Contents

4.3.7. Digression: The presumed non-nativeness of passives in PA . . . 168 4.3.8. Final remarks on passivization in PA and Upper Guinea PC . . . 168 4.4. Final remarks on morphology . . . . . . . . . . . . . . . . . . . . . 169

5 Verbal system 5.1. 5.1.1. 5.1.2. 5.1.3. 5.2. 5.3. 5.3.1. 5.3.2. 5.3.3. 5.3.4. 5.4. 5.4.1. 5.5. 5.6. 5.6.1. 5.6.2. 5.6.3. 5.6.4. 5.6.5. 5.6.6. 5.7. 5.7.1. 5.7.2. 5.7.3. 5.7.4. 5.8.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PA / Upper Guinea PC preverbal ta . . . . . . . . . . . . . . . Analyzing CV ta as a progressive aspect marker . . . . . . Analyzing PA ta as [+imperfective], rather than [+present] Final remarks on PA / Upper Guinea PC preverbal ta . . . . The diachrony of the PA perfective past marker a . . . . . . Future tense marking in PA and Upper Guinea PC . . . . . The PA future tense marker lo vs. its absence in Upper Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . On the origin of PA lo . . . . . . . . . . . . . . . . . . . . . . . The diachrony of future tense marking in PA and Upper Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Digression: SCV al and PA lo . . . . . . . . . . . . . . . . . . PA / BaCV taba – tabata . . . . . . . . . . . . . . . . . . . . . Digression: On the diachrony of preverbal taba and postverbal -ba . . . . . . . . . . . . . . . . . . . . . . . . . . . . The issue of relative versus absolute tense marking in PA . A comparison of stative verbs in PA and SCV . . . . . . . . The stative – nonstative distinction in creoles . . . . . . . . . Strong vs. weak stative verbs . . . . . . . . . . . . . . . . . . The class of strong stative verbs . . . . . . . . . . . . . . . . The class of weak stative verbs . . . . . . . . . . . . . . . . . Contrastive analysis . . . . . . . . . . . . . . . . . . . . . . . . Digression: The case of GBC . . . . . . . . . . . . . . . . . . . Auxiliary verbs . . . . . . . . . . . . . . . . . . . . . . . . . . . Modal auxiliaries . . . . . . . . . . . . . . . . . . . . . . . . . . Copular verbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other auxiliaries . . . . . . . . . . . . . . . . . . . . . . . . . . . Final remarks on auxiliary verbs . . . . . . . . . . . . . . . . Final remarks on the verbal system . . . . . . . . . . . . . . .

6 Summary and interim analysis of the linguistic results 6.1. 6.2.

. . . . . . .

. . . . . . .

. . . . . . .

171 171 172 173 186 192 193 197

. . . 198 . . . 200 . . . 203 . . . 207 . . . 208 . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

211 214 217 217 218 219 221 223 225 226 227 230 240 254 255

257 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 Predominance of Portuguese-derived function words in PA . . . . 257 Structural overlap between PA and Upper Guinea PC . . . . . . . 259

Contents

6.3. Negative evidence from PLQ and Gulf of Guinea PC . . . 6.3.1. Digression: What sets PA and Upper Guinea PC apart from Gulf of Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . 6.4. Old Portuguese features in PA and Upper Guinea PC . . . . 6.5. The value of historical PA and Upper Guinea PC texts . . . 6.6. West-Atlantic and Mande features in PA and Upper Guinea PC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

. . . 260 . . . 261 . . . 264 . . . 265 . . . 266

7 The historical ties between Upper Guinea and Curaçao 7.1. 7.2. 7.2.1. 7.2.2. 7.2.3. 7.2.4. 7.2.5. 7.3. 7.3.1. 7.4. 7.4.1. 7.4.2. 7.4.3. 7.5. 7.6.

269 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 On the presumed insignificance of Upper Guinea to the history of Curaçao . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 The Dutch presence in Senegambia in the 17th century . . . . . . 273 The Dutch in Gorée . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 The Dutch on the Petite Côte (Rufisque, Portudal and Joal) . . . 279 The loss of Gorée and the Dutch retreat from Senegambia . . . . 281 The Dutch ties with Cacheu and the Cape Verde Islands . . . . . 285 Final remarks on the Dutch presence in Senegambia in the 17th century . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 Dutch slave trade from Upper Guinea to Curaçao . . . . . . . . . 289 Other factors relevant to the Dutch slave trade from Upper Guinea to Curaçao . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 Sephardic Jewish networks linking Upper Guinea to Curaçao . . 294 Ties between the Sephardim in Upper Guinea and Amsterdam . 296 Sephardim networks directly linking Upper Guinea to Curaçao . 297 Partnership between the Dutch WIC and the Sephardim . . . . . 299 Diffusion of Upper Guinea PC to the mainland, 16th and 17th centuries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 Summary, conclusions, and final remarks . . . . . . . . . . . . . . 304

8 Discussion: The development from Upper Guinea PC to Papiamentu

307 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 8.1. Sociolinguistic considerations . . . . . . . . . . . . . . . . . . . . . 308 8.1.1. On the choice of slaves in the early period of Curaçao’s settlement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 8.1.2. Sociolinguistic issues relevant to the consolidation of Upper Guinea PC on Curaçao and its diffusion among the (slave) population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312

xii 8.2. 8.2.1. 8.2.2. 8.2.3. 8.2.4. 8.3.

Contents

From Upper Guinea PC to PA: a case of rapid relexification towards Spanish . . . . . . . . . . . . . . . . . . . . . . . . . . . . PA, monogenesis, and the notion of relexification in creole studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From Upper Guinea PC to PA: ‘relexification’ rather than ‘heavy borrowing’ . . . . . . . . . . . . . . . . . . . . . . . . . . . Analyzing Papiamentu as a mixed language . . . . . . . . . . . The source(s) of the Spanish elements in PA’s basic content vocabulary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of the discussion . . . . . . . . . . . . . . . . . . . . .

9 Conclusions

. . 319 . . 320 . . 322 . . 327 . . 331 . . 335 337

Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385

Abbreviations

AB ANG ANT AVDB BCV BaCV CH COP CV DEF EPA FCV FUT GBC GU Gulf of Guinea PC HAB IND NEG PA PFV PK PLQ Port. PRI PROG PST REL SCV SNCV SoCV Sp. ST SVCV TMA

Annobonese Angolar anterior Authorized Version of the Dutch Bible Brava CV Barlavento CV Chabacano copula Cape Verdean Creole definite marker Early PA Fogo CV future Guinea-Bissau and Casamance Creole Guyanese Creole Gulf of Guinea Portuguese Creole (covers ANG, ST, PRI and AB) habitual indefinite marker negator Papiamentu perfective Papia Kristang Palenquero Portuguese Principense progressive past relative pronoun Santiago CV São Nicolau CV Sotavento CV Spanish Sãotomense São Vicente CV tense, mood and aspect

xiv

Abbreviations

Upper Guinea PC PL V V Vaux WIC 1sg 2sg 3sg 1pl 2pl 3pl

Upper Guinea Portuguese Creole (covers CV and GBC) plural verb vowel (NB: only in phonological sections, otherwise V verb) auxiliary verb (Dutch) West India Company first person singular second person singular third person singular first person plural second person plural third person plural

Introduction

When delving into issues such as the origin of the name Curaçao and the origins of Papiamento, one runs the risk of searching too deep for solutions. Studies on these subjects, however, have the advantage that they bring to light – or to mind – details concerning the history of our Leeward Islands.1 (Menkman 1936: 49, 50)

This study is concerned with the origins of Papiamentu (PA), the creole language of the Caribbean islands Aruba, Bonaire and Curaçao, and its linguistic and historical ties with the Upper Guinea branch of Portuguese-based creoles (Upper Guinea PC), which includes the sister creoles of the Cape Verde Islands (CV) and Guinea-Bissau and Casamance (GBC). Both linguistic and historical evidence will be presented in support of the hypothesis that these creoles are historically and genetically related. While scholars agree that PA’s lexicon is predominantly Spanish-based – hence its typical designation as a Spanish-lexifier creole (e.g. Sebba 1997: 30) – the presence of a number of Portuguese-derived items in the fundamental part of its vocabulary is also widely recognized. Since Papiamentu is spoken in a Spanish-speaking area, the presence of these Portuguese words requires an explanation – “d’où viennent les mots d’origine portugaise” (Maurer 1985: 43) – and lies at the heart of a century-long debate, which is far from being resolved: “The source(s) of the Portuguese elements in Papiamentu may never be determined with certainty” (Lipski 2008: 547). Finding out where the Portuguese comes from, in turn, has important implications for the closely related question of whether PA is and never was anything but a Spanish-based creole or whether it was imported from elsewhere as an originally Portuguese-based variety, only to be subsequently relexified towards Spanish. Although this controversy has often been shunned by classifying PA as an Iberian-based creole, Kramer (2004: 10) correctly notes that “this fortunate term does not relieve us of the search for

1

Original quote: “Zich verdiepend in kwesties, zooals die van den oorsprong van den naam Curaçao en het ontstaan van het Papiamento, loopt men gevaar de oplossingen al te diep te zoeken. Studies over deze onderwerpen hebben echter het voordeel, dat zij leiden tot het aan het daglicht – of in herinnering – brengen van bijzonderheden, de geschiedenis onzer Benedenwindsche Eilanden betreffend”. Translations of quotes are mine unless indicated otherwise.

2

Introduction

an explanation of the partially Spanish, partially Portuguese character of Papiamentu”2 . Hancock (1975: 228) once observed that although PA “is one of the most extensively studied (…) creoles, (…) the question of its origin (…) is still in dispute”. Some three decades later, scholars are still “evenly divided as to the Spanish vs. Portuguese origins of Papiamento” (thus Lipski 2005: 282). Illustrating this division, Munteanu (1996a) painted PA as an originally Spanish-based creole, while in that same year Martinus (1996) defended itsAfro-Portuguese ancestry. The scholarly disagreement came to light again recently, when Munteanu (2007: 442) asserted that the idea of PA as an originally Portuguese-based creole has lost much ground, whereas Kouwenberg & Ramos-Michel (2007: 307) stated the opposite, namely that the Spanish hypotheses now appear untenable. Although Martinus (1996) was critically received, he had actually shown that much of what is Portuguese-derived in PA can be traced back to Upper Guinea PC. Elaborating on Martinus (1996), Quint (2000b) provided new and compelling comparative data to support the claim that PA and Upper Guinea PC descend from the same Portuguese-based ancestor creole. Like Martinus (1996), however, Quint’s (2000b) claims did not find any resonance and the ties between PA and Upper Guinea PC remain fully understudied up to date (but see McWhorter 2006: 117; Green 2007: 336). In a recent assessment of the state of the research on PA’s origins, the sobering conclusion was reached that “[t]he source(s) of the Portuguese elements in Papiamentu may never be determined with certainty” (Lipski 2008: 547) The primary aim of the present study is to examine and provide evidence for the hypothesis that PA is genetically related to Upper Guinea PC. With this purpose, these creoles will be compared at all levels of the grammar, but with a special focus on the morphosyntax. A secondary aim of this study is to provide a historical framework that accounts for the linguistic transfer from Upper Guinea to Curaçao. This interdisciplinary approach is essential, since the lack of a historical framework appears to be one of the main reasons why the linguistic ties between PA and Upper Guinea PC have still received so little recognition in the literature. In the remainder of this introductory part, I will specify the hypothesis examined and supported in this study, address some methodological issues and outline the structure of the book. First, I will briefly introduce the three principal languages considered in this study: PA, CV and GBC. 2

Original quote: “Dieser glückliche Terminus enthebt einen dennoch nicht der Suche nach einer Erklärung für die teils spanische, teils portugiesische Prägung des Wortschatzes”

Presentation of the languages considered in the present study

3

Presentation of the languages considered in the present study Papiamentu (PA) PA3 is the official language of the Leeward Dutch Antilles. These comprise Aruba, Bonaire and Curaçao, situated some 60 kilometres off the coast of Venezuela and jointly referred to as the ABC-Islands. At present, PA has an estimated number of 270,000 native speakers, of which approximately 120.000 reside on Curaçao, 60,000 on Aruba and some 10,000 on Bonaire, while the Netherlands harbor the remainder. Smaller pockets of PA speakers are found throughout the Caribbean (e.g. St. Martin) and also in the United States (Haseth 1995: 29–31; Dijkhoff, Kouwenberg & Fat 2004: 2105). The Spanish discovered the ABC-islands in 1499. Not impressed by the dry soil and lack of raw materials, they did not actively colonize the islands. Hence,

Map 1. Geographic distribution of PA, CV and GBC at present 3

Two publications, Reinecke (1975: 147–209, with the cooperation of Frank Martinus Arion) and Coomans-Eustatia (2005), stand out for providing exhaustive bibliographies of the PA language, including literary, linguistic, sociolinguistic and historical publications.

4

Introduction

when the Dutch West Indian Company (WIC) took Curaçao in 1634, a negligible number of Spanish soldiers was encountered, all of whom voluntarily left the island. A similarly small number of Amerindians, whose ethno-linguistic profile is subject to debate4 , were allowed to remain (cf. Hartog 1968: 29–42; Holm 1988: 312–316; Martinus 1996: 3–6; Maurer 1998: 185–193; Kramer 2004: 11–34). The 17th century saw the Dutch arise rapidly as a colonial power with possessions in the Caribbean, Brazil, Asia and West Africa. With a growing need for slaves to work the sugar plantations in the Dutch parts of Brazil, the WIC proportionally increased its activities in West Africa. Their subsequent development as a dominant slave trading nation was furthermore facilitated by agreements (‘asientos’) with the Spanish Crown to supply parts of the Spanish New World with slaves. And not only the Spanish, but also the French and the English initially relied on Dutch slave supplies (Parkvall 2000: 126, 127; Rawley & Behrendt 2005: 74). The Dutch taking of the Portuguese stronghold Elmina on the coast of what is now Ghana in 1637 (i.e. shortly after the conquest of Curaçao) is considered the starting point of the export of considerable numbers of African slaves by the WIC to Brazil and the Caribbean (cf. Postma 1990: 13). From the 1650s onwards, owing to its strategic position and the loss of the Dutch holdings in Brazil in 1654, Curaçao could develop from a mere naval base into the WIC’s foremost slave trading center in the New World (cf. Postma 1990: 26–55). The first shipments of slaves reached Curaçao in the late 1650s, with their numbers growing exponentially in subsequent decades. Slaves purchased in West Africa were brought in and subsequently redistributed throughout the Americas. While the majority of slaves was thus resold, an increasing num4

Traditional wisdom has it that the first Amerindian inhabitants of the ABCIslands were Venezuelan Caquetios, whose first language, Caquetio, belongs to the Arawakan language family. It is not known, however, whether (and if so to what extent) PA was ever in close contact with Caquetio; historical documentation testifies to the fact that the linguistic repertoire of the Amerindians encountered by the Dutch in 1634 included (and perhaps was limited to) Spanish, which is unsurprising given the almost one and a half century of cohabitation with Spaniards (cf. Maurer 1998: 199). On the other hand, Grant (2008b: 85) notes that the ABC-islands received a continuous influx of Amerindians from different areas, both in the pre-1634 period and afterwards, not all of whom would have known Spanish. In any case, the contribution of Amerindian languages to PA appears to be of a mainly lexical nature and has been dealt with by scholars such as Gatschet (1884), van Buurt & Joubert (1997), Kramer (2005) and van Buurt (2009). Note, though, that certain PA words traditionally attributed to Amerindian influence may also have entered PA through contact with Antillean Spanish (Grant 2008b: 81).

Presentation of the languages considered in the present study

5

ber would also stay on the island to do construction work in the urban area of Willemstad (the capital of Curaçao), to work the farms in the interior, or to perform a variety of tasks in European households. This initial period of intense slave trade, economic growth and demographic expansion, extending roughly from 1650 to 1700, is generally (though certainly not unanimously, pace Barme 2008: 20) considered to be the period in which PA emerged on Curaçao5 as a vehicle of interethnic communication (Maurer 1986b: 97; 1988: 2; Portilla 2008b: 165; Bartens 1996: 247, Kouwenberg & Muysken 1995: 205; Munteanu 1996a: 84; Fouse 2002: 83; Joubert & Perl 2007: 139). That by the turn of the 18th century PA had indeed already established itself as a (if not the) dominant language of Curaçao is, though not proven, at least suggested by a reference to ‘broken Spanish’ in Father Alexius Schabel’s 1704 travel account (cf. e.g. van Wijk 1958: 169; Ferrol 1982: 29; Bartens 1996: 248) as well as by various isolated fragments of what appears to be PA, for instance, in the nomenclature of Jewish boats in the first third of the 18th century (Martinus 1996: 8, 9). In 1732, moreover, a certain Father Caysedo reported that, in addition to Spanish, Portuguese, and Dutch, the people of Curaçao speak “the language of the country” (in Hartog 1968: 153). In assessing the origins of slaves brought to Curaçao, the bankruptcy of the 1st Dutch WIC in 1674 and the foundation of the 2nd WIC in that same year mark a break. Scholars agree that slaves shipped to Curaçao by the 2nd WIC (i.e. after 1674) were primarily drawn from Lower Guinea and Angola. However, only a fraction of the 1st WIC’s records have been preserved so that very little is known about the origins of the slaves imported into Curaçao prior to 1674. In the historical chapter of this study I will return to this issue and shed light on the pre-1674 period. It is uncontroversial that PA was first spoken on Curaçao and from there spread to the islands of Aruba and Bonaire (Eckkrammer 1999: 61, 62; Holm 2001: 73). While the three insular dialects are mutually clearly intelligible, there are some conspicuous lexical and phological differences which are yet to receive systematic scholarly attention (Kouwenberg & Murray 1994: 4). With 5

Although it is difficult not to see the similarities between the toponym Curaçao and Port. coração ‘heart’, the etymology of the island on which PA developed is more complicated and not necessarily Portuguese. On maps printed in the 16th century, i.e. prior to any presence of Lusophone communities on the island, variations of its final name such as Curasote, Corazante, Coraçante, Curaçote, Curasaote and Curasaore are used on Spanish maps and in official documents in reference to the island, and by the beginning of the 17th century, the island was already known as Curaçao or Curazao (Goslinga 1979: 10; cf. Birmingham 1970: vi).

6

Introduction

the exception of Dijkhoff’s didactic grammar of Aruban PA (2000), the PA grammars and grammatical sketches available to me at the time of this writing predominantly describe the mesolectal variety of Willemstad, Curaçao. Western Curaçao and the interior of Bonaire are home to arguably the most basilectal varieties of PA (Philippe Maurer p.c.), but descriptions of these varieties are lacking.6 Cape Verdean Creole (CV) CV is the native tongue of most of the approximately 500,000 inhabitants of the Cape Verde Islands, an archipelago of nine islands off the coast of Senegal, and of several Cape Verdean diasporic communities around the world (Batalha & Carling 2008 provide am excellent overview), including in the Netherlands, France, and the United States.The CapeVerde Islands are geographically divided into the Sotavento (‘Leeward’) and Barlavento (‘Windward’) Islands.The former encompass Santiago, Maio, Fogo and Brava and are home to around 70% of all native CV speakers. The latter comprise São Antão, São Vicente, São Nicolão, Sal and Boa Vista as well as the uninhabited island of Santa Luzia (Bartens 2000; Quint 2010). 6

The only data available on the speech of Bonaire consist of sporadic lexical entries in dictionaries such as van Putte & van Putte-de Windt (2005a & 2005b), Ratzlaff (1992) and Joubert (1999). According to Frank Martinus Arion (p.c.), from a linguistic point of view, “Bonaire is a true goldmine”. This is one of the reasons why the Curaçaoan scholar Abigail Michel (now Abigail Ramos-Michel) had planned to make a linguistic description of Bonairean PA (announced as ‘forthcoming’ by Holm 2002: 59). However, such a description has not yet been produced. Generally, Aruban PA is thought to have absorbed more Spanish influence than Curaçaoan and Bonairean PA, which in turn presumably borrowed more heavily from Dutch. The more pronounced Spanish character of Aruban PA can at least in part be explained by the proximity of Aruba to the Venezuelan mainland and the continuous influx of Venezuelan laborers throughout Aruba’s history. The linguistic propensity of Aruban PA towards Spanish is visible in the vocabulary (e.g. Aruban cuatro versus Curaçaoan kuater ‘four’) and in the pronunciation. For instance, in Curaçao, the palatalization of /s/ before diphthongs was generalized (e.g. /tentaSon/ ‘temptation’), whereas Arubans have either retained or restored a more Spanish pronunciation (/tentasjon/) (Maurer 1998: 195f.n.). Note, furthermore, that Aruba has adopted an etymological orthography based on Spanish, whereas Curaçaoans and Bonaireans prefer a phonological orthography. For instance, Arubans write , and , as opposed to Curaçaoan , and (Maurer 1998: 152, Kouwenberg & Murray 1994: 4, 5). In the present study, the orthography used by the cited authors is retained.

Presentation of the languages considered in the present study

7

This geographic division coincides with the division of CV into two main dialect groups. The Sotavento varieties (SoCV) are considered the oldest and most conservative. The more acrolectal Barlavento varieties (BaCV) arose at a later date out of the contact between immigrants from the Sotavento Islands and new generations of European Portuguese settlers (Holm & Swolkien 2006: 206). An overview of features distinguishing the two varieties is provided by Quint (2000b: 71–86). Within the Sotavento branch of CV, the demographically and sociolinguistically dominant variety is that of Santiago (SCV), with an estimated 170,000 native speakers divided between the interior and the capital city, Praia. Natives of Santiago commonly designate themselves and their speech as badiu while referring to the people and speech of all other Cape Verde Islands as sanpadjudu (Lang 2002: 684). Within the Barlavento branch of CV, the dialect of São Vicente and its capital Mindelo is the most influential (Holm & Swolkien 2006: 16–18). It is generally assumed that CV emerged first on Santiago. From there, it spread to the other Sotavento Islands in the 16th and 17th centuries and to the Barlavento Islands in the 18th and 19th centuries7 (cf. Bartens 2000: 37). The island of Santiago was discovered by the Portuguese in 1456 and settlement policies were launched soon after (Holm 1988: 273). The island’s strategic position linking West Africa to the New World allowed the Portuguese to turn it into the point of reference of the early Trans-Atlantic slave trade. Over a period of more than a century, slave ships of all nations trading in West Africa were obliged to stop over and pay taxes in Santiago before crossing the Atlantic. When the island lost these exclusive rights in the 1640s, it had already been struck by several droughts, famines and epidemics, a combination of factors that led to the demise of Santiago’s slave trade economy (Carreira 1983a; Holm 1988: 273, 274; Bartens 1996: 21–23). Though reliable historical-demographic data for Santiago’s early period are scarce, the records at hand suggest that the formation of CV “was both early and rapid” (Bartens 2000: 38) and some consensus has been reached on situating the birth of CV in the late 15th, early 16th century (Carreira 1983a: 64, 65; Quint 2000a: 55; Veiga 2000: 37). Lang (2001; 2006) provides clever, combined historical and linguistic evidence in favor of that view: given that after 1500 the Portuguese no longer purchased slaves in Wolof-speaking territory, the undeniable linguistic contribution of Wolof to the grammar of CV (see Quint 2000b & 2008b; Lang 2004, 2005, 2006 & 2009) must be taken to suggest that creolization had been largely completed by that time (Lang 2006: 54). I should 7

São Nicolau (Barlavento) is an exception: it was already populated in the 17th century (Bartens 2000: 37).

8

Introduction

stress, however, that finding out when or how CV (read: proto-Upper Guinea PC) was formed is not the aim of the present study (but see Jacobs 2010). Note, finally, that in addition to Wolof (West-Atlantic), Mandinka (Mande) and to a lesser extent Temne (West-Atlantic) are recognized as principal contributors to the substrate of CV (cf. Baptista 2006; Rougé 2006; Quint 2008a; Lang 2009). The Creole of Guinea-Bissau and Casamance (GBC) In Guinea-Bissau (with a total of 1,500,000 inhabitants), GBC (alternatively referred to in the literature as Kriyol or Guineense) is among the most widespread languages (only Balanta and Pulaar having more native speakers) with about 200,000 native and an additional 600,000 L2 speakers (e.g. Holm 1988: 275)8 . Another important GBC speaker community resides in the Senegalese province of Casamance, where the number of native speakers totals around 50,000. Most of these live in the city of Ziguinchor, which was settled by traders from Cacheu (in present-day Guinea-Bissau) in the mid-17th century (cf. Holm 1988: 275; Rougé 2004b: 149). On rare occasions the kinship between CV and GBC has been disputed (e.g. Morais-Barbosa 1975: 150; d’Andrade & Kihm 2000: 108). The present study will, however, adopt the majority’s view according to which the two creoles share a common origin: Quint’s (2000b: 99–117) concise comparison leaves little room for doubt, and more recently, Baptista, Mello & Suzuki (2007) have further explored the structural similarities, establishing a 90% correspondence of the morphosyntactic features measured by Holm & Patrick’s (2007) comparative creole study. Rougé (1999: 56), furthermore, estimates that “80% of Africanderived lexemes in CV also exist in GBC”9 . Both the 90% morphosyntactic correspondences and the shared African lexicon would be inexplicable if the two creoles were not intimately related.10 There is much unresolved debate over the question whether the proto-variety from which both CV and GBC sprang was originally formed on Santiago and from there taken to the Guinea-Bissau region, or vice versa. In keeping with scholars such as Lopes da Silva (1957), Carreira (1983a), Silva (1985, 1990), Thiele (1991), Kihm (1994), Quint (2000b), Parkvall (2000), Baptista (2000, 2002) and Peck (1988), I assume that the proto-creole emerged on Santiago 8

Cf. the Ethnologue’s report on the linguistic diversity in Guinea-Bissau: http://www.ethnologue.com/show_country.asp?name=GW 9 Original quote: “80% das palavras de origem africana do crioulo de Santiago existem também na Guiné” 10 Rougé (1994) estimates the intelligibility between CV and GBC at around 80%.

Presentation of the languages considered in the present study

9

around the turn of the 16th century and from there diffused to the continent around the turn of the 17th century. The issue is dealt with in detail in Jacobs (2010) and will have no further bearing on the present study. Since the Portuguese traded the Casamance area to the French in the 1880s, the variety spoken in that area has undergone less post-formative Portuguese influence and is thus likely to be more conservative than the dialect of GuineaBissau. Although the latter is much better-studied, features specific to the Casamance variety have also been brought to light, e.g. in the works of Wilson (1962: 35, 36), Chataigner (1963), Doneux (1979), Bal (1980, 1983a, 1983b, 1983c, 1985), Peck (1988), Doneux & Rougé (1988) and Rougé (1985, 1988, 2004a, 2004b). A forthcoming description of the Casamance variety by Biagui & Quint (forthcoming), moreover, will provide new insights into the particularities of that variety vis-à-vis Guinea-Bissau. In the present study, a strict distinction between the two varieties will not yet be made. It is relevant to note, finally, that, unlike CV, GBC continues to be in contact with African languages. This means, for instance, that it is often difficult to determine whether a given GBC feature not present in CV is an original creole feature or rather results from adstrate influence (cf. Rougé 2004a; Intumbo 2007, 2008; Holm & Intumbo 2009). Hypothesis examined in the present study The present study examines and provides linguistic as well as historical evidence for the hypothesis that PA is genetically related to Upper Guinea PC, brought from Upper Guinea to Curaçao in the second half of the 17th century. The synchronic differences between the two creoles will be argued to have resulted primarily from a rapid and far-reaching process of relexification towards Spanish. This relexification, I argue, was carried out in the time span of perhaps no more than three to four decades and completed quite shortly after the separation of the creoles (i.e. after the arrival on Curaçao of speakers of Upper Guinea PC) in the second half of the 17th century.11 The relexification affected a fundamental though purely lexical part of the original vocabulary, while, as I hope to show, leaving intact a significant portion of the original function words and morpho-syntax.

11 Following the cross-linguistically marked process of relexification, I assume of course that PA has undergone (and continues to undergo) regular processes of contact-induced and internal change like any other language.

10

Introduction

Methodological remarks To examine and support the hypothesis outlined above, the present study will rely primarily on the discussion of linguistic evidence (chapters 2–6). This evidence will then be placed in a solid historical framework (chapter 7), in order to provide a coherent interdisciplinary account of the origins of PA. Linguistic evidence To linguistically scrutinize the claim of kinship between PA and Upper Guinea PC, I will apply what is arguably “the most powerful tool in the historical linguist’s toolbox” (Thomason 2007: 42): the comparative method. As a point of departure, I take the widespread assuption that when a language changes, its function words and morphology will normally be least susceptible to change, whereas content vocabulary will be most susceptible12 (cf. e.g. Thomason & Kaufmann 1988: 14; Muysken & Smith 1990: 883; Comrie 2000: 39; Bickerton 2001: 1104, etc.). This means that, if PA and Upper Guinea PC are genetically related, we can expect to find evidence for this in the correspondence of their function words and morphology, which is why I will compare the creoles primarily in these domains. In addition to derivational and inflectional morphology, the following parts of speech will be analyzed: pronouns, prepositions, question words, conjunctions, TMA markers, and auxiliary verbs. I will compare the members of these categories in terms of their syntactic function, phonological shape and semantic content. Although the comparison of morphosyntactic categories and features will thus constitute the essential part of the argumentation, I will start off with a chapter on phonology, addressing a subset of phonological features that appear to be particularly important in the present context.13 For obvious historical-linguistic reasons, to substantiate the claim of genetic relatedness, I will compare PA with those varieties of Upper Guinea PC con12 Of the 41 languages in Haspelmath & Tadmor’s (2009) loanword database sample, three constitute an exception to the rule, displaying a higher proportion of borrowed function words compared to content words (Tadmor 2009: 59). Another interesting exception are the so-called Para-Romani languages, characterized as they are by the substitution of the original Romani grammar by that of a local dominant language, while the Romani vocabulary is preserved (e.g. Boretzky & Igla 1994; Matras, ed. 1995). 13 The chapter on phonology will be organized with the secondary purpose of bringing to light several noteworthy lexical correspondences between PA and Upper Guinea PC, although I should stress that I will not look for hard evidence of genetic ties in the purely lexical domain of the grammar.

Presentation of the languages considered in the present study

11

sidered to be the most conservative and basilectal, which are SCV and GBC. Conveniently, these are also the best described varieties. Nonetheless, whenever possible and relevant, I will seek to profit from the increasing body of literature that has become available in recent years dealing with different aspects of the grammar of the Barlavento varieties of CV14 as well as the grammar of the conservative Casamance variety of GBC. The methodological advantages of bringing multiple dialectal varieties into the equation are explained by Comrie (2000: 40): [O]ther things being equal, the more independent descendants there are, the easier it will be to establish genetic relatedness (…). [E]ach descendant will retain a certain proportion of features of the ancestor language, although the precise features retained will differ from descendant language to descendant language. The more independent descendants there are, the more likely it is that more material will be retained from the ancestor language in enough descendants to make valid comparison possible. (Comrie 2000: 40)

The data for PA will be drawn mainly from contemporary grammars, grammatical outlines, dictionaries, newspapers, stories and scholarly articles. In addition, I will draw on a body of Early PA records, most of which have not yet been used for diachronic purposes. Among these records are letters, gospels, catechisms, newspaper articles, and short grammars written over the period from 1775 to 1928.15 (References to these can be found passim throughout the monograph.) Needless to say, if there is truth in the hypothesis examined in this study, we expect Early PA to have been more similar to Upper Guinea PC than present-day PA. The study of these early sources will allow us to scrutinize this premise. Negative evidence It is important to be aware that, as Comrie (1988: 81) puts it, “[c]hance can, by definition, never be excluded completely” as an explanation for correspondences between a given pair of languages. The issue of chance is of course particularly delicate in creole studies. In historical linguistics, “the existence of cognates in the basic morpheme stock” (Weinreich 1958 cited in Thomason 2008: 255) 14 On top of the grammars of different varieties of BaCV by Lopes da Silva (1957), Almada (1961) and Cardoso (1990), the Barlavento varieties are receiving renewed attention in publications such as Pereira (2000), Bartens (2000), Veiga (2000), Swolkien (2004 & 2009) and Holm & Swolkien (2004, 2006 & 2010). 15 I decided to take Lenz’s (1928) landmark PA grammar as roughly marking the transition from Early PA to modern PA. This choice is somewhat arbitrary, but it reflects the fact that, in between Lenz (1928) and the 1950s, there is a gap of roughly two decades in which relatively little was written in PA.

12

Introduction

is still often considered sufficiently strong evidence of a genetic relationship between languages, but creoles are exempt from that basic formula in as far as the commonalities between the Indo-European lexifiers as well as the effect of creole universals may in many cases account for such cognates in the basic morpheme stock and other parallel features. Consequently, providing negative evidence from the lexifiers Spanish and Portuguese as well as from other Iberianbased creoles will be particularly essential to (the linguistic part of) the present study with the purpose of reducing as much as possible the likelihood that a certain feature shared by PA and Upper Guinea PC is due to chance. As to the lexifiers, whenever relevant, contrast with both Spanish and Portuguese will be provided in order to stress the idiosyncrasy and/or originality of certain shared PA / Upper Guinea PC features. It should be noted, however, that negative contrast with Portuguese is of course not always necessary for a correspondence between PA and Upper Guinea PC to be meaningful: since PA is not in contact with Portuguese, all PA features that can be shown to be Portugueserather than Spanish-derived may potentially have been inherited from Upper Guinea PC and thus constitute a meaningful historical link. Negative evidence from other Iberian-based creoles is similarly important, although this cannot always be presented systematically. In some cases a category or feature shared by PA and Upper Guinea PC may either not exist or be differently organized in the other creoles. Prepositions, for instance, are rife in PA and Upper Guinea PC, but relatively scarce in other creoles. In yet other cases, a category may simply not have been described as thoroughly or systematically for other Iberian-based creoles as for PA and Upper Guinea PC. The ‘negative’ creole data will be drawn primarily (though certainly not exclusively) from those Iberian-based creoles that in the literature have most frequently been claimed to be genetically related to PA: Gulf of Guinea Portuguese-based Creole (henceforth Gulf of Guinea PC) and Spanish-based Palenquero. Both will be introduced briefly below. Gulf of Guinea PC consists of four genetically related varieties distributed over three islands in the Gulf of Guinea (Map 2 below). Sãotomense (ST) and Angolar (ANG) are spoken on the island of São Tomé, Principense (PRI) and Annobonese (AB) on Príncipe and Annobón respectively. The proto-creole of this branch is likely to have emerged around the turn of the 16th century on São Tomé from where it spread to the other islands (Parkvall 2000: 133). The substrate of Gulf of Guinea PC is mixed Kwa and Bantu. A lexical Bantu component is particularly patent in ANG, which is a maroon variety of ST partially relexified by Kimbundu (Parkvall 2000: 133, 134). For further reading on these creoles, see Holm (1988: 277–284) for all varieties; Schuchardt (1882), Ferraz (1979), Major (2006), Hagemeijer (2007) and Fontes (2007) for ST; Günther (1973)

Presentation of the languages considered in the present study

13

Map 2. Geographical distribution of Gulf of Guinea PC and PLQ vis-à-vis PA and Upper Guinea PC

and Maurer (2009a) for PRI; Maurer (1995) and Lorenzino (1998) for ANG. Notwithstanding some smaller articles by e.g. Ferraz (1983), Granda (1986), Post (2000) and Maurer (2002), the availability of linguistic data on the creole of Annobón is rather limited. The role of this variety in the present study will be correspondingly small. El Palenque de San Basilio, home to speakers of the Spanish-based creole Palenquero (henceforth PLQ), is a village situated some 50 kilometers south of Cartagena, Colombia (Map 2). PLQ is likely to have been a full-fledged creole already by the mid-17th century (Schwegler 1998), but the actual circumstances under which the creole was formed are rather uncertain. Some apparently AfroPortuguese features in the core of the PLQ grammar have nourished the debate about possible genetic ties between PLQ and Afro-Portuguese creoles. (I will address this matter in §1.6.1 and §5.7.3.3.2.) Key publications on (the history of) the language and/or its sociolinguistic background include Bickerton & Escalante (1971), Granda (1978), de Friedemann & Patiño Rosselli (1983), Schwegler (1996, 1998), Schwegler & Green (2007) and Moñino & Schwegler (eds. 2002). The substrate of PLQ is thought to be exclusively Bantu, with a particularly strong Kimbundu component (see Schwegler 2002).

14

Introduction

Historical evidence While a solid historical framework may of course be favorable to any study involving genetic linguistics, it is particularly essential in the context of the hypothesized ties between PA and Upper Guinea PC. The importance of Upper Guinea to the history of the Dutch WIC and thus by extension to the history of Curaçao and PA has quite consistently been overlooked and/or marginalized in the literature. And if there were no historical contacts between Upper Guinea and Curaçao, it is difficult to postulate a scenario of language transfer from one region to the other. The principal aim of the historical chapter is thus to demonstrate that these historical contacts did in fact exist, in other words, that the conditions were met for linguistic transfer from Upper Guinea to Curaçao to occur in the second half of the 17th century. With this purpose, I will demonstrate the pronounced presence of the Dutch WIC in the Senegambia region in a period of some five decades, the beginning of which is marked by their conquest of Gorée (a small island situated directly beneath the Cape Verde Peninsula) in 1621. The period ends abruptly with the loss of that same island to the French in 1677. In addition, I will present little-known data on the slave trade between Gorée and Curaçao in that pivotal period. The historical chapter closes with a discussion of Sephardic Jewish trading networks linking Upper Guinea to Curaçao in the second half of the 17th century. The decision to discuss the linguistic evidence (chapters 2–6) before the historical (chapter 7) is born out of the assumption that linguistic evidence should generally receive priority over historical in any attempt at establishing kinship between a given pair of languages. (In this respect, see Smith 1999: 252, 256 and 2009: 308, 309; Parkvall 1999b: 57, 58 and 2006: 329, 330.) Structure of the present study Following a review of the literature (chapter 1), the present study is divided into five linguistic chapters (2–6), a historical chapter (7) and a discussion (chapter 8). I start the linguistic part with a chapter on phonology, (2), addressing what, in the present context, is a particularly revealing subset of vowel (§2.1) and consonant (§2.2) changes and furthermore focussing on strategies of resyllabification (§2.3) and the issue of verbal stress in PA and Upper Guinea PC (§2.4). Then, in chapters 3–5, the morphosyntax will receive full attention: chapter 3 (Selected parts of speech) discusses pronouns (§3.1), prepositions (§3.2), interrogatives (§3.3), conjunctions (§3.4) and a set of miscellaneous features (§3.5); chapter 4 is dedicated to derivational (§4.1) and inflectional morphology (§4.2). Chapter 5 contains an elaborate, mainly diachronic, comparative analysis of the

Presentation of the languages considered in the present study

15

verbal systems of PA and Upper Guinea PC. §5.1–§5.4 of that chapter provide a comparison based on PA’s four principal TMA markers ta, a, lo, and tabata. The issue of relative vs. absolute TMA marking in the two creoles is explained and discussed in §5.5, while §5.6 and §5.7 provide comparisons of stative and auxiliary verbs respectively. Chapter 6 provides a summary and interim analysis of the linguistic data surveyed thus far. The historical chapter 7 is organized as follows: in §7.1, I will illustrate how the tendency to marginalize the role of Upper Guinea in the history of the Dutch West India Company manifests itself in the study of (the origins of) PA. §7.2 is concerned with mapping the presence of the Dutch in the 17thcentury Senegambia region. Subsequently, in §7.3, little-known data on the slave trade from the Senegambia region to Curaçao between 1659 and 1677 is presented. Finally, §7.4 highlights the underexposed but pronounced presence of a Sephardic Jewish community in Senegambia and its commercial ties with the Jewish communities in Amsterdam and Curaçao. In chapter 8, Discussion, I will relate the linguistic and historical evidence from the previous parts in order to discuss, analyze and speculate about, firstly, the sociolinguistic factors at play in facilitating the arrival, survival and consolidation of Upper Guinea PC on Curaçao (§8.1) and, secondly, the circumstances and mechanisms underlying the relexification of Upper Guinea PC towards Spanish, i.e. the process out of which I assume PA developed (§8.2). I close this study with a summary and conclusions (chapter 9).

Chapter 1 Critical review of the literature on the origins of Papiamentu

Introduction Although the majority of PA’s modern lexicon is undoubtedly derived from Spanish, a considerable portion (estimates vary from 5% to 25%16 ) of core vocabulary items appear to be of Portuguese origin. This mixed vocabulary is at the basis of a century-long controversy about the sources of the Portuguese items and about the closely related question of how to classify PA: as an originally Spanish-based creole that emerged in situ (i.e. on Curaçao) or as an originally Afro-Portuguese creole or pidgin imported from elsewhere and subsequently relexified towards Spanish. As such, the debate on PA’s origins has long reflected the wider debate within the field of creole studies opposing mono- to polygenesis, but more differentiated accounts and hypotheses, not so easily classifiable in terms of this dichotomy, have also been proposed, particularly in the latter stages of the 20th century. Below, a detailed review of the literature on PA’s origins is presented. I will sketch how, in the late 19th/early 20th century, the awareness of the presence of Portuguese elements in PA gradually grew (§1.1) and how this awareness resulted in PA’s inclusion in monogenesis frameworks from the 1950s onwards (§1.2). §1.3 describes the emergence in the 1960s of a polygenetic ‘Spanish’ countercurrent. A critical discussion of various genesis hypotheses is provided in the remaining sections: the Spanish hypotheses are reviewed in §1.4; the role of the Sephardic Jews in the formation of PA is analyzed in §1.5; finally, §1.6 discusses the four candidates that in the literature have found most recognition as possible sources of the Portuguese material in PA: an Afro-Portuguese proto16 Lenz (1928: 207–260) provided what is arguably still the most exhaustive and meticulous etymological survey of PA’s lexicon, assessing the Spanish, Portuguese and Dutch contributions in different lexical domains. Other valuable etymological studies are presented in Fokker (1914), Maduro (1953, 1966a, 1966b), Andersen (1974), Maurer (1986a, 1991), Jeuda (1990), Kowallik & Kramer (1994), Kramer (2004: 98– 100, 139–155) and Grant (2008a). It goes without saying that an exact estimate in percentages of the Portuguese and Spanish in PA’s lexicon is impossible due to the lexical resemblance of Spanish and Portuguese. Nevertheless, when comparing the items of undoubtedly Spanish with those of undoubtedly Portuguese origin, the former clearly dominate quantitatively.

18

Critical review of the literature on the origins of Papiamentu

pidgin (§1.6.1), a Brazilian Creole (§1.6.2), Gulf of Guinea PC (§1.6.3) and Upper Guinea PC (§1.6.4).

1.1. From Schabel (1704) to Lenz (1928) While as early as 1704, Father Alexius Schabel referred to PA as ‘broken Spanish’, similar depictions – e.g. “Spaansch patois” ‘Spanish patois’ (Teenstra [1836]1977: 179), “Negerspaansch” ‘negro-Spanish’ (van Ginneken 1928: 283) or “Bastaard-Spaansch” ‘hybrid Spanish’ (Fokker 1914: 68; cf. Gatschet 1884: 303; Eybers 1915: 3) – can be found in most casual descriptions of the creole up to the early 20th century. Some early references to PA attributed an important formative role to the original Amerindian inhabitants of the ABC-islands: “The Indians of Curaçao have certainly had there own language, which was intermingled with Spanish words afterwards, while in later times also some mangled Dutch words were added, so that from this mixture PA was put together”17 (from an 1825 article by Father Phoel partially reproduced in Maduro 1965: 5).18 It took some time for PA’s Portuguese-derived features to be recognized. Also in the first scholarly work on PA (Teza 1863), or in the classic creole surveys by Van Name (1869: 149–159) and Coelho (1880: 48–51), the Portuguese element is not yet noticed and PA’s classification as a Spanish-based creole therefore not questioned.19 To the seasoned creolist it may not come as a surprise that Schuchardt (1882) was (to the best of my knowledge) the first to single out Portuguese and AfroPortuguese features in PA. In his 1882 description of “das Negerportugiesische von S. Thomé” ample reference is made to “das Curazoleñische”. Without Schuchardt drawing any conclusions in terms of its genesis, his knowledge of both Spanish and Portuguese allowed him to observe that “Curaçaoan has many

17 Original quote: “De Indianen van Curaçao hebben zeker hun eigen taal gehad, die naderhand (…) met Spaanse bastaardwoorden vermengd is en waarbij in latere tijden ook eenige verdraaide Hollandsche woorden zijn gevoegd, zoodat daaruit het samengelapte Papiament ontstaan is” 18 A similar view had been expressed by Fuchs (1849: 7, in Hancock 1984: 140). Later, scholars such as van Balen (1940), Castro (1976) or de Haseth (1990) would also attribute an important role to native Amerindian languages in the formation of PA. 19 This notwithstanding, Van Name’s (1869) grammatical sketch of PA provides remarkable detail and is of great interest to the analysis of 19th-century PA.

1.1. From Schabel (1704) to Lenz (1928)

19

Portuguese elements”20 . The parallels between ST, CV and PA on the levels of syntax, phonology and lexicon observed by Schuchardt are remarkable, given the early date of his writing. In a brief grammatical sketch of PA, the Curaçaoan scholar Jesurun (1897) drew attention to the Portuguese origin of the future marker lo and the suffix -mentu and labeled PA as “a patois or vernacular with Old Spanish or Portuguese as its base”21 . Jesurun’s paper is noteworthy, since he is probably the first Antillean to make such observations and did so independently of the writings of Schuchardt.22 Corresponding to the limited means of dissemination of knowledge at that time, Schuchardt’s and Jesurun’s insights did not spread rapidly; witness, for instance, Henríquez Ureña’s (1921: 361, 362, emphasis in original) classification of PA as “the only creole language that Spanish has produced in the New World”.23 It is only with the appearance of the landmark PA grammar by Lenz (1928) that Schuchardt’s PA-related observations were first echoed and, moreover, significantly complemented.24

20 Original quote:“das Curazoleñische [enthält] zahlreiche portugiesische Elemente”. Translations of quotes by Schuchardt (1882) are taken from Hagemeijer & Holm (2008). 21 Original quote:“een patois of volkstaal, die oud-Spaansch of Portugeesch tot grondslag heeft” 22 Of less merit, but no less remarkable, is Jesurun’s next remark: “Het voortbestaan dier volkstaal, hoe betreurenswaardig het ook zijn moge, is een feit waarmede rekening dient gehouden te worden” (1897: 96) [‘The survival of their vernacular, however lamentable it may be, is a fact that has to be counted on’]. 23 Original quote: “La única lengua criolla que el castellano ha producido en el Nuevo Mundo” 24 In his introduction, Lenz (1928: 10) paid the following tribute to Schuchardt: “Naturalmente me dirijí primero a la principal autoridad científica en lenguas criollas, el profesor Hugo Schuchardt, de la Universidad de Graz, autor de los nueve folletos ‘Kreolische Studien’(…) (1882–1891). El anciano sabio tuvo la amibilidad de mandarme a vuelta de correo una lista de 31 títulos de publicaciones en Papiamento, o estudios sobre esta lengua que poseía, editados entre 1843 i 1891 (…)” [‘Of course I first turned to the principal scientific authority on creole languages, Professor Hugo Schuchardt, from the University of Graz, author of the nine articles dubbed ‘Kreolische Studien’ (1882–1891). The wise old man was so kind to send me by return of mail a list of 31 titles of publications in PA, or studies on this language that he possessed, edited between 1843 and 1891’].

20

Critical review of the literature on the origins of Papiamentu

1.2. Afro-Portuguese hypotheses: from Lenz (1928) to monogenesis Lenz (1928) brought to the fore similarities between PA and the Portuguesebased creoles of West Africa on all levels of the grammar and presented the first plea for PA’s Afro-Portuguese origins. He describes PA as “the negroPortuguese creole brought by the slaves”25 (1928: 13), a conclusion based on the claim that not so much PA’s lexicon but rather its grammar “is primarily ‘negro-Portuguese’”26 (1928: 323). The historical framework proposed by Lenz mirrored the framework that would later come to be known as monogenesis, entailing the importation to Curaçao of Afro-Portuguese pidgin- or creole-speaking slaves who had acquired this proto-variety either during their stay in the slave camps along the African West coast or on board the vessels that crossed the Atlantic. Metaphorically speaking, “the Portuguese seed falls on African ground and a tree grows (the negro-Portuguese jargon which all blacks transported in Portuguese ships have to learn)”27 (Lenz 1928: 80). Lenz furthermore alleged that, “When this negroPortuguese language was established on Curaçao, many words were assimilated to Castilian”28 (1928: 323, 326), thus formulating the process that supporters of monogenesis would come to call relexification.29 Lenz (1928) was received enthusiastically.30 According to Latour (1940), his analysis “explains the origins of Portuguese words in Papiamentu and their often 25 Original quote: “el criollo negro-portugués traído por los esclavos” 26 Original quote: “es ‘negro-portuguesa’ en primer lugar” 27 Original quote: “La semilla portuguesa (…) cae en terreno africano (…) i nace un árbol (la jerga negro-portuguesa a la cual tienen que acomodarse todos los negros transportados en buques portugueses)” 28 Original quote: “Cuando el habla negro-portuguesa (…) se fijó en Curazao (…), muchas palabras se han asimilado (…) al castellano” 29 Lenz (1928) continues to be one of the most exhaustive sources of Portuguese words and features in PA. Moreover, the linguistic data presented in it (rich both in quantity and in quality) provide valuable insight into the way early 20th-century mainstream PA differed from modern mainstream PA. Velleman (2008) is a study of Lenz’s life as a linguist and also relates how his 1928 PA grammar came about. 30 This warm reception was in no small way due to his favorable portrayal of PA as “una de las lenguas más perfectas del mundo” (Lenz 1928: 331) [‘one of the most perfect languages of the world’]. Lenz (1928) was reviewed in detail by Hesseling (1933a), who added extensive digressions on Dutch lexico-semantic influence on PA and translated and reproduced one of the short stories provided in Lenz (1928). Note that Hesseling (1933b) is a comparison of PA and Negerhollands with a particularly focus on shared Dutch influences.

1.2. Afro-Portuguese hypotheses: from Lenz (1928) to monogenesis

21

Portuguese-like pronunciation”31 (1940: 223). As a consequence, the attention of researchers concerned with PA is “nowadays strongly directed towards the Portuguese element in the language”32 (Latour 1940: 223)33 . The correlation between Lenz (1928) and the growing scholarly awareness of the Portuguese element in PA is further exemplified by Navarro Tomás’(1953) article, a “phonetic addition to Rodolfo Lenz’s work”34 (1953: 183) in which the author concluded that “the primitiveAfro-Portuguese phonetics have made way for the Spanish”35 . Also van Wijk (1958, 1968, 1969), convinced of the Afro-Portuguese origins of PA, mentioned Lenz as the intellectual father of that view. Both Navarro Tomás and van Wijk attempted to historically account for the processes of decreolization and Hispanization to which they believed PA had been exposed after its introduction into Curaçao. Navarro Tomás (1953: 188) thereby highlighted the 19th-century migration of Venezuelan and Colombian workers to Curaçao as essential, whereas van Wijk (1958: 176) thought that a first shift towards Spanish had already been initiated with the late 17th-century visitations of Venezuelan missionaries seeking to christianize the growing Curaçaoan slave population. With the whole of Schuchardt’s oeuvre at his disposal, Lenz (1928) had firmly speculated about monogenesis: “According to the European language that prevails in the place of destination, Spanish, French, English or Dutch elements are inserted into this Negro-Portuguese stock. Its uniform characteristics can be seen even today from Central America and the Antilles to Indo-China”36 (1928: 41). With Lenz’s grammar and Navarro Tomás’ and van Wijk’s papers finding a wider audience, and with creole monogenesis being explicitly articulated from the late 1950s onwards, PA soon became a prime example of the diffusion of a West African proto-pidgin in the Caribbean. Thompson (1961: 112) 31 Original quote: “verklaart (…) de herkomst van Portugeesche (…) woorden in het Papiamento en de vaak sterk aan het Portugeesch herinnerende uitspraak” 32 Original quote: “tegenwoordig (…) sterk naar het Portugeesche element in de taal gericht” 33 Latour’s point is illustrated by short articles on the Portuguese in PA by Dutch scholars such as himself (Latour 1936, 1937, 1940), Menkman (1936, 1940) and van Balen (1940). 34 Original quote: “adición fonética al trabajo de Rodolfo Lenz” 35 Original quote: “La primitiva fonética afroportuguesa ha ido cediendo el campo a la española” 36 Original quote: “Según la lengua europea que prevalece en el lugar del destino, en este tronco negro-portugués se hacen injertos españoles, franceses, ingleses u holandeses (…). Sus rasgos uniformes se ven hasta hoi desde la América Central i las Antillas hasta la Indochina.”

22

Critical review of the literature on the origins of Papiamentu

illustrates: “Papiamentu is, according to Professors Navarro Tomás and van Wijk, a development of the West African slavers’ jargon. This jargon, much influenced, no doubt, by the West African substratum, may have been the pattern for all the West Indian Creoles”. Another forerunner of monogenesis, Whinnom (1965: 527), believed that “a process of relexification, from Portuguese to Spanish, explained completely and satisfactorily the Portuguese elements in Papiamento”. In a similar vein, Cassidy (1971: 207) stated: “The Pg [Portuguese] element in JC [Jamaican Creole] could have entered as a part of PgP [Portuguese Pidgin] spoken in the Caribbean: it formed the basis of Papiamento”.37 In the 1970s, PA’s Afro-Portuguese origins were supported most passionately by Birmingham (1970, 1971, 1975, 1976, 1978).38 Birmingham (1975) compared PA with CV, GBC and Annobonese in some detail, concluding: “The fact that the four are indeed closely related is no doubt abundantly clear” (1975: 22). New in his work was a reference to the so-called Lingua de Preto (a term coined by Teyssier 1959), the foreigner talk of African immigrants in 15thto 16th-century Lisbon as portrayed in Gil Vicente’s early 16th-century plays. Birmingham (1975: 23, 24) showed that this immigrant Portuguese shares some creole-like features with PA and the West African Portuguese-based creoles, such as the emphatic 1st person singular pronoun , the lack of gender agreement and the generalization of infinitives. Interestingly, also in 1975, Baird published a little-known paper entitled “Anticipations of Papiamento in the Afro-Portuguese of Gil Vicente”. In the slipstream of monogenesis, a group of scholars to which Granda (e.g. 1968, 1973, 1974, 1978), Perl (e.g. 1982, 1989a, 1989b, 1990) and Megenney (e.g. 1983, 1984, 1985, 1993) belong, were particularly concerned with accounting for what they considered to be conspicuous commonalities among the Caribbean creoles PA, PLQ and Saramaccan and other Afro-Iberian varieties spoken in the Caribbean (see also Otheguy 1973). An integral part of their scholarship was the analysis of the above-mentioned Lingua de Preto found in Vicente’s plays and later dubbed ‘reconnaissance language’by Naro (1978). This reconnaissance language had presumably been used by the Portuguese in their contacts with the natives of West Africa. It was claimed to have been instrumen37 In addition to the works of Thompson, Whinnom and (to a lesser extent) Cassidy, publications by, for instance, Stewart (1962), Valkhoff (1966), Taylor (1971, 1977), Voorhoeve (1973) and Hancock (1975) belong to the canon of the literature on monogenesis. For a more complete overview and discussion of relevant scholarship within the school of monogenesis, see Holm (1988: 44–52). 38 Not at my disposal but possibly relevant is a publication in the same period by Lessa (1975) titled ‘The presence of Portuguese in Papiamentu’.

1.2. Afro-Portuguese hypotheses: from Lenz (1928) to monogenesis

23

tal to the formation of an expanded Portuguese-based pidgin used as a lingua franca also in the Caribbean, and was thus seen as the ultimate source of the Afro-Portuguese words and features found throughout the Atlantic. Megenney (1984: 186), for instance, compared Afro-Portuguese elements in PA, PLQ and Afro-Cuban Spanish and concluded that “all three stem from the Portuguesebased reconnaissance language, as our data, both linguistic and historical, purport to indicate” (cf. Granda 1978: 216–233; Perl 1998: 7, 8). The hypothesized link between the reconnaissance language and PA’s origins was recently revisited in the works of Martinus (1996: 122, 123) and Kramer (2004: 127–134). Conventional monogenesis lost support from the 1980s onwards, when the similarities among creoles world-wide were increasingly explained as a result of either shared substrate languages (e.g. Alleyne 1980; Boretzky 1983) or universal faculties of language creation and acquisition (e.g. Bickerton [1974]1980, 1975, 1981, 1984). Simultaneously, the various pan-creole correspondences that had led to the emergence of monogenesis now faced a growing awareness of significant typological discrepancies as a result of the increasing access to comparative creole data.39 As a consequence of the above, according to McWhorter (2005: 379), “no working creolist today subscribes to the monogenesis hypothesis”. His own Afrogenesis hypothesis (e.g. McWhorter 1995, 1997, 2000), however, may be seen as a differentiated offshoot of the old school monogenesis model: he claims that the three branches of Portuguese-, French- and English-based New World creoles each trace back to a respective proto-variety formed in WestAfrica.An intriguing argument adduced by McWhorter is the absence of Spanish-based creoles in the New World: prior to the taking of Annobon in the late 18th century, the Spanish were not significantly established in West Africa and a Spanish-based protovariety could thus not emerge there. This circumstance, he postulates, accounts for the absence of Spanish-based creoles in the New World. McWhorter’s claim evidently entails that both PA and PLQ are relexified varieties of imported AfroPortuguese: “[W]hile these creoles are clearly Spanish-based synchronically, both have been shown to have originated as Portuguese-based diachronically” (McWhorter 2000: 13, emphasis in original). (McWhorter’s afrogenesis hypothesis is critically reviewed in Bickerton [1998] and Lipski [2005: 281–304].) A special position within the literature on PA’s afrogenesis in general and its ties with Upper Guinea PC in particular is taken by Martinus (1996, etc.). In a 39 Thus, Thomason & Kaufman (1988: 195) said with respect to monogenesis: “we do not think this hypothesis is needed to account for the structural similarities among the various European-vocabulary pidgins and creoles, and monogenesis (…) fails to account for grammatical differences among pidgins and creoles”.

24

Critical review of the literature on the origins of Papiamentu

short literary article, Martinus (1998: 113) sums up his ideas on the origins of PA as follows: “Though specialists are not in total agreement about the linguistic history of Papiamento, most (and I count myself among them) consider it to be a pure case of an Afro-Portuguese Creole brought from Africa by the slaves.” More recently, PA’s afrogenesis and its ties with PLQ and Saramaccan received renewed support in the work of Portilla (2007, 2008a, 2008b).40 His work is a return to the more traditional monogenesis and attempts at reconstructing the alleged Afro-Portuguese proto-pidgin from which he assumes all West African Portuguese-based creoles descend as well as PA, PLQ, and Saramaccan in the New World. Both Martinus’ and Portilla’s work will be discussed in more detail in §1.6.1. Starting with Lenz (1928) and ending with Portilla (2007, 2008a, 2008b), this section described the emergence and maintenance of a more or less coherent monogenesis model that sees PA as the descendant of an originally AfroPortuguese proto-variety that also formed the basis of a number of other Iberianbased creoles and overseas Afro-Hispanic varieties. Meanwhile, in response to the increasing body of literature claiming an Afro-Portuguese origin for PA, a countercurrent had emerged portraying PA as an intrinsically Spanish-based creole. This school of thought is discussed next.

1.3. Spanish hypotheses In the 1950s, scholarship surfaced in direct opposition to PA’s alleged AfroPortuguese origins, instead claiming the creole descends directly from Spanish. The most outspoken adherent of this view was the Curaçaoan scholar Maduro (1953, 1965, 1966a, 1966b, 1969, 1987a, 199141 ): “In the course of time, various authors have expressed their opinion regarding the provenance and formation of our language. Many have committed the error of choosing Portuguese. The slaves that arrived here could not have had knowledge of any stable language (neither a ‘lingua franca’, nor Afro-Portuguese, nor coastal Portuguese, nor broken Portuguese), as some contend.”42 (Maduro 1965: 1, 3, 29)43 40 I am indebted to Mario Portilla for the generous sharing of his papers. 41 Under the heading Maduro, the Encyclopedie van Curaçao (http://www.Curaçaoencyclopedia.com) lists additional publications by Maduro. 42 Original quote: “Den curso di tempu varios escritor (…) a publicá nan opinión tocante procedencia i formación di nos lenga. (…) E error qu hopi a cometé tá, di a bai tras di e portugués (…). E catibunan qu tábata yega aquí no por tábata domina ningún idioma fiho (ni ‘lingua franca’, ni afro-portugués, ni portugués costeño, ni portugués quibrá), manera algún hende ta pretende”

1.3. Spanish hypotheses

25

In addition to Maduro’s oeuvre, the works of van Balen (1940), Rona (1971, 1976), Wood (1972a), DeBose (1975), Todd Dandaré (1979), Ferrol (1982)44 and particularly Munteanu (1974, 1991, 1992, 1996a, 1996b) are important contributions to the Spanish school of thought. Rona (1976: 1021) details: “Papiamentu is the direct descendant of the Spanish spoken on Curaçao during the period in which the ABC-islands pertained to the Spanish Crown. The African grammar was applied to the Spanish lexicon and, this way, Amerindians and Africans have conserved until today the Spanish linguistic tradition without interruption”45 . In Rona’s view, the creolization of Spanish began after the arrival of the Dutch and Africans. An alternative view holds that creolization had already begun prior to that (e.g. van Ginniken 1928; van Balen 1940: 373; de Haseth 1990: 556; Gomes Casseres 1990). In both cases, there is linguistic continuity between the pre- and the post-Dutch period. To explain away and/or trivialize the Portuguese elements in PA, Maduro (1965: 8) successfully demonstrated that several of the PA words typically claimed to derive from Portuguese in fact have cognates in overseas varieties of Spanish, in other Iberian languages such as Galician and Catalan, or in 16thcentury Spanish. For instance, with respect to the PA preposition na ‘in, to’ (< Port. em + a ‘in + the’ = Sp. en la), Maduro (1966b: 9) affirms: “This is 43 Maduro’s 1965 paper is an indictment of PA’s Afro-Portuguese hypotheses in general. The paper contains a valuable collection of testimonies, statements and claims in favor of both the Spanish and the Afro-Portuguese theories with Maduro clearly supporting the former and (quite fiercely) attacking the latter. The wide variety of quotes and digressions taken from a number of encyclopedias, several issues of different Aruban and Curaçaoan journals and other sources difficult to consult outside of the ABC-islands shows how lively the debate had been up till then especially among Antillean and Dutch scholars. 44 Ferrol (1982) was reviewed by Hancock (1984), who proved himself well-informed: “Munteanu (1974) and DeBose (1975), neither of which is referred to by Ferrol, both make good cases for a Spanish origin for Papiamentu. On the other hand, the many detailed studies by Granda, culminating in his 1978 book (…), make a convincing case linking Papiamentu with the Portuguese creoles of Africa. And the fact that one recent treatment of the subject refers in its title to ‘Papiamentu and other Portuguese-based creoles’ (Martinus 1980) is clear indication that the controversy is far from settled” (Hancock 1984: 141). (Martinus [1980] is a conference paper not at my disposal.) 45 Original quote: “El papiamentu (…) es el descendiente directo del español hablado en Curaçao durante el período en que las islas ABC pertenecieron a la Corona española. La gramática africana fue aplicada al léxico español y, de este manera, indios y africanos han conservado hasta hoy ininterrumpidamente, la tradición lingüística española”

26

Critical review of the literature on the origins of Papiamentu

one of the lexemes that those concerned with the etymology of our words have constantly classified as Portuguese. We clearly find that its Portuguese origin is not as certain as they usually pretend it is. In the past, the Spanish used na and in Galician na is a common preposition.”46 Maduro’s etymological findings were echoed in the works of scholars such as Munteanu (1991, 1996a) and Ferrol (1982). In the words of the latter: “Maduro has brilliantly weakened many of the assumed Portuguese etymologies, demonstrating the existence of these words either in Old Spanish or in Spanish dialects”47 (Ferrol 1982: 32). This reasoning is also invoked in a recent, littleknown but remarkably detailed study of the PA lexicon by Grant (2008a: 60): “It would only be possible to assure the Portuguese derivation of most of these words (…) if it could be shown to be impossible to derive them from archaic or dialectal (and non-Castilian) Spanish forms”.48 Also worth mentioning is Maduro’s (1987a) juxtaposition of a PA text with its translation into SCV.49 Although the author himself made no linguistic comments, Sidney Joubert’s epilogue illustrates the paper’s intentions: Mr. Maduro, who does not believe in this theory [afrogenesis], clearly demonstrates what he has already said so often, namely that there is an immense distance – or should I say unbridgeable gap? – between Papiamentu and CapeVerdean Creole. We may hope that adherents of the Afro-Portuguese proto-creole theory soon manage to present a concrete vision of what this theory actually contends. On the basis of what Maduro has presented in this work it does not seem that an inhabitant of Cape Verde and one of our Leeward Islands will ever be able to understand a single bit of what they say to each other. And you may ask yourself how these languages could ever have had a common origin.50 (In Maduro 1987a: 22) 46 Original quote: “Ta un dje vocablonan cu esnan cu a ocupanan cu etimología di nos palabranan a calificá stedi como di origen portugués. Nos ta mira bon cla cu su origen portugués no t’asina sigur manera nan sa pretende (…). Antes spañónan tabata usa ‘na’ i na gayego (…) ‘na’ ta un preposishon coriente.” 47 Original quote: “Maduro ha desvirtuado magistralmente muchas de las supuestas etimologías portuguesas, demostrando la existencia de esas palabras en el español antiguo o en los dialectos españoles” 48 Grant (2008b) is a highly informative study of the sociolinguistic history of Curaçao. Note that Grant (2008a, 2008b) takes no explicit stand in the Spanish vs. Portuguese origins debate, but rather stipulates a convergence of different scenarios. 49 Maduro (1987b) did the same with PA and PLQ. 50 Original quote: Mr. Maduro, kende no ta kere den e teoria akí [Afrogenesis], ta mustra klaramente loke mashá tempu el a bisa kaba, esta ku tin un distansha inmenso, – o mi tin di bisa un presipisio intomabel? – entre papiamentu i kabodverdiano. Ta di spera ku esnan, ku ta atherí na e teoria di un protokrioyo afroportugues, por pre-

1.4. Critical discussion of the Spanish hypotheses

27

DeBose (1975) paid much attention to the correspondences between PA and the West African Portuguese-based creoles and admitted a secondary influence by native speakers of these creoles on PA: The Portuguese element in Papiamentu may be accounted for by positing the existence, within the linguistic repertoire of the Curaçao slaves of a Portuguesebased creole related to the creoles spoken in Guinea-Bissau, Cape Verde and the Gulf of Guinea today. It could also be accounted for by the fact that Portuguese was the mother tongue of the Sephardic Jews of Curaçao, but numerous parallels between Papiamentu and the West African creoles are difficult to account for solely on this basis of contact. (1975: 65)

He nonetheless ends up claiming what the title of his dissertation suggests: “Papiamentu: a Spanish-based creole”. Also among the Spanish theories, we find Bickerton & Escalante’s (1970) proposal for the existence of a Spanish-based pan-Caribbean proto-creole as a precursor to both PA and PLQ: “Since Colombia and Curaçao were colonized simultaneously, and since slaves for both were drawn from similar sources, it is highly likely that Palenquero and Papiamento, though by now probably mutually unintelligible, are descendants of the same Proto-Creole; the similarities between them, on all levels are too numerous to be coincidental” (Bickerton & Escalante 1970: 263). This hypothesis was subscribed to by Wood (1972a: 20): “Papiamentu may be seen to descend from the Spanish-affiliated pan-Caribbean creole, as do Palenquero in northern Colombia, the former Puerto Rican Creole Spanish, and others yet to be discovered”. Bickerton himself would of course abandon the idea only a few years later as incompatible with his renowned Language Bioprogram.

1.4. Critical discussion of the Spanish hypotheses One general point of critique that goes out to the sum of hypotheses that claim a Curaçao-situated birth of PA with Spanish as its original lexifier is pointed out by Kramer (2004: 107), who notes that they fail to provide “an acceptable explanation for the fact that the Portuguese elements are found in the deepest

sentá pronto nan vishon konkreto riba loke e teoria akí ta kome i bebe (…). Pasobra a base di loke Maduro ta presentá den e obra akí no por dedusí ku un abitante di Cabo Verde i unu di nos Islanan Abou por komprondé un pátaka di loke un ta bisá otro. I bo ta puntra bo mes si un dia e dos lenganan akí por tabatin e mesun orígen.

28

Critical review of the literature on the origins of Papiamentu

layers of the core vocabulary [of PA]”51 . Indeed, the concentration of Portuguese elements in the basic vocabulary of PA suggests that these elements belong to the oldest layers of its grammar, something that the Spanish origin hypotheses cannot readily account for. With this basic objection in mind, the tenability of the remaining arguments that underlie the Spanish hypotheses will be examined closely below. 1.4.1.

Linguistic continuity between the pre- and post-1634 period?

As mentioned, the Spanish hypotheses rely heavily on the idea of linguistic continuity between the pre- and post-1634 period: the variety of Spanish spoken on Curaçao prior to the Dutch arrival presumably persisted and laid the foundations for PA. Whether this Spanish had already been restructured prior to the Dutch arrival or not is of little relevance here, since both views can be dismissed on the basis of the same argument: the socio-demographic conditions were clearly unfavorable to linguistic continuity between the pre- and the post-1634 period. This is best illustrated by Goodman (1987: 367, 368): When the Dutch captured it [Curaçao], they deported to the mainland not only the Spaniards, but also (…) all except seventy-five of the more than 400 indigenous Indians, of whom twenty-three (…) were kept as servants of the Dutch (who numbered over 400) (…).These Indians must surely have picked up enough Dutch within a year or two to communicate with the dominant and far more numerous Europeans. The rest of the Indians were sent to the village of Ascencion, some distance away (…). These fifty or so (…) may have continued to use Spanish, but their isolation, as well as their inferior numbers and subordinate political and social status, would virtually have precluded it from functioning as a widespread lingua franca for any length of time. Apparently, a few of the Indians who had been deported to the mainland returned to the island shortly thereafter (…), but it seems unlikely that they could have altered the situation to any significant extent (…). Furthermore, there was scarcely any contact other than smuggling between the island and the nearby Spanish-speaking mainland before 1648, when peace was established between Holland and Spain.

Similar historical arguments had already been put forward by Hartog (1968: 43– 48) and were later reemphasized by Martinus (1996: 22), Maurer (1998: 200) and Kramer (2004: 108).The latter summarized it thus: “A population whose Spanish could undergo a creolization process was simply lacking. The demographic as

51 Original quote: “eine akzeptable Erklärung dafür, dass portugiesische Elemente in den tiefsten Schichten des Grundwortschatzes zu finden sind”

1.4. Critical discussion of the Spanish hypotheses

29

well as the linguistic history of the ABC-islands starts in 1634 from scratch”52 (2004: 108). The force with which Spanish reentered the Curaçaoan linguistic scene in the late 17th century as language of prestige, external trade and religion is of course a different matter (cf. §8.2.4 in this regard); the crucial point here is that the Dutch taking of Curaçao in 1634 marked a clear linguistic dividing line, severely weakening hypotheses that rely on the continuity of Spanish between the pre- and the post-1634 periods. An interesting parallel can in fact be drawn with the linguistic history of Jamaica, where the arrival of the English in the mid-17th century caused a rupture in the island’s linguistic history: “Other than a few lexical items and place-names, Jamaican Creole owes little to either the indigenous Arawaks or their Spanish conquerors, who arrived with Columbus in 1494, settled the island in 1509 from Santo Domingo, and were defeated by the English in 1655–60” (Patrick 2007: 127). 1.4.2.

Linguistic evidence against Old Spanish in PA’s superstrate

A key linguistic argument of the Spanish camp was that several PA features of apparent Portuguese origin were in fact remnants of Old (16th-century) Spanish. While the historical data discussed in the previous paragraph already severely complicate that idea, Quint (2000b: 184, 185) put forward an additional compelling piece of phonological evidence, suggesting indeed that 16th-century Spanish did not define PA’s superstrate. This evidence will be discussed in detail in §2.2.1.2. 1.4.3. About the tendency to attribute the Portuguese to other Hispanic varieties A 16th-century Spanish contribution to PA is thus implausible in the light of both historical and linguistic evidence. The apparently Portuguese-derived items can however still be played down by showing that these items also appear in one or the other non-standard or overseas variety of Spanish or in one of the Romance languages of the Iberian Peninsula other than Spanish/Portuguese, typically Catalan and/or Galician (cf. Maduro 1965, 1966a, 1966b; Ferrol 1982; Wood 1972a; Munteanu 1996a). 52 Original quote: “es fehlte schlicht und einfach die Bevölkerung, deren angestammtes Spanisch einem Kreolisierungsprozess hätte ausgesetzt werden können. Die Bevölkerungs- wie die Sprachgeschichte der ABC-Inseln beginnt 1634 auf dem Nullpunkt”

30

Critical review of the literature on the origins of Papiamentu

Maurer (1998: 198) explains why this tendency is problematic: “It is unlikely that either Galician, Catalan or Valencian played any role in the formation of PA, since the presence on Curaçao of Catalans, Valencians or Galicians is not mentioned in historical sources”53 . And although Rona (1971: 81) subscribed to the Spanish origins of PA, he also criticized: “It is not possible to conceive of Papiamentu as a type of mosaic composed of words derived from all archaic and modern dialects of Spanish. Evidently, if Papiamentu originated from Spanish, it was a variety of Spanish with well-defined characteristics.”54 (cf. Martinus 1996: 30). What is most decisive, however, is the afore-mentioned fact that the vast majority of the apparently Portuguese-derived elements is found in the core of PA’s lexicon and includes pronouns, core prepositions, modal verbs and question words. Kramer (2004: 104) therefore concludes: “The primarily Portuguese character of the non-Castilian Iberian part of PA’s vocabulary should not be ignored by pointing to the occurrence of the same word in one or the other Iberian speech variety: about one sixth of PA’s vocabulary, mostly core vocabulary items and structure words, is clearly of Portuguese origin”55 .

1.5. PA birth among the Sephardim? Several scholars claiming a Spanish origin for PA have assumed that the Portuguese elements were somehow introduced sideways into the language by the Sephardic Jews (often referred to as ‘Portuguese’ Jews) who settled on Curaçao in the course of the 17th century: “There they encountered a local gibberish of partially Spanish, partially Amerindian words, which they subsequently made as

53 Original quote: “es poco probable que el gallego o el catalán / valenciano hayan tenido alguna influencia en la formación del papiamentu, pues no se menciona la presencia en Curazao de catalanes, valencianos o gallegos en las fuentes históricas” 54 Original quote: “no es posible concebir el papiamentu como una especie de mosaico compuesto de palabras procedentes de todos los dialectos antiguos y modernos del español (…). (…) evidentemente, si el papiamento se origina en el español, fue un español con características bien definidas” 55 Original quote: “Den primär portugiesischen Charakter des nichtkastilischen iberoromanischen Anteils am Papiamento-Wortschatz darf man (…) nicht durch Verweis auf das Auftauchen desselben Wortes in mal dieser, mal jener Sprachform der iberischen Halbinsel wegdiskutieren: Etwa ein Sechstel des PapiamentoWortschatzes ist eindeutig Portugiesischer Herkunft, darunter durchaus Elemente des Grundwortschatzes und Strukturwörter”

1.5. PA birth among the Sephardim?

31

Portuguese as possible”56 (van Balen 1940: 371, 372). This view too, however, fails to account for the presence of the Portuguese elements in the oldest and most basic layers of the PA grammar. I assume, therefore, that if one insists on situating the birth of PA on Curaçao and on attributing PA’s Portuguese elements to the Curaçaoan Sephardim, one would be easier off claiming that the actual creolization process (viz. the formation of PA) took place among the Curaçaoan Sephardim. Slightly surprisingly, although the hypothesis is sometimes referred to in the literature (e.g. Sanchez 2006: 156), I am not aware of any scholarship that seriously elaborates on the possibility of a Curaçao-situated PA birth among the Sephardim. Therefore, there would seem to be no need to critically discuss it here. I will nonetheless discuss it in some detail below since (a) I feel that a PA birth among the Sephardim is perhaps the single most promising argument against an Afro-Portuguese origin of PA and (b) the hypothesis was suggested to me orally at several conferences. Although it is difficult to falsify such a hypothesis, I will cast some doubt on it by arguing that Spanish, not Portuguese, is likely to have been the predominant colloquial language in the early Curaçaoan Sephardim community (§1.5.1). In addition, I will provide some demographic data that speak against a scenario of creolization among the Curaçaoan Sephardim (§1.5.2). 1.5.1.

On the linguistic profile of the early Curaçaoan Sephardim

It is generally rather uncritically held that the Curaçaoan Sephardim’s home language, before collectively shifting to PA, was Portuguese: “Until 1865 the predominant language was Portuguese. It was the mother tongue of almost all the Jews and was used in the synagogue and in the communal chancery” (Emmanuel 1957: 112). However, Emmanuel’s affirmation requires some clarification: the use of Portuguese in the synagogue and in the communal chancery is a documented fact, but the use of Portuguese as a mother tongue is not. It is beyond reasonable doubt that Portuguese, not Hebrew or Spanish, dominated in religious and ceremonial contexts (Wood 1970: 8; Joubert & Perl 2007: 47). The analysis of the epitaphs of over 2,500 gravestones of the oldest (1659) Jewish cemetary on Curaçao, Beth Haim, unambiguously confirms

56 Original quote: “de Portugeesche joden, die in den loop der 17e eeuw op Curaçao gekomen zijn. Zij hebben aldaar een plaatselijk brabbeltaaltje van deels Spaansche, deels nog Indiaansche woorden aangetroffen, waaraan zij vervolgens een zooveel mogelijk Portugeesch cachet hebben verleend”

32

Critical review of the literature on the origins of Papiamentu

this.57 However, Wood (1970: 9) specifies that “there is no proof that Portuguese was ever the vernacular among the Sephardic Jews of Curaçao”.58 Indeed, Joubert & Perl(2007: 48) confirm that it is “extremely difficult to draw conclusions regarding what language Sephardic immigrants may have spoken”. There is an obvious explanation for that: “No documents on Jewish community life from before 1775 survive” (2007: 46). Moreover, the Sephardim who settled on Curaçao did not come from one, but from several different places (especially from Amsterdam and Brazil, but also, for instance, from Livorno, Italy, and other areas in the Caribbean) and even where the place of origin is known, it is not immediately clear what the home language was. Most agree that a high level of bi- or even multilingualism among the Sephardim was the norm. Granda (1974: 7) noted that the average Sephardic Jewish community mastered at least “three Hispanic linguistic codes: Castilian, Ladino [i.e. Judaeo-Spanish59 ] and Portuguese”60 . The above-mentioned analysis of the 57 Of these gravestones, “1,668 bear inscriptions in Portuguese, 40 in Hebrew, 72 are bilingual (Hebrew, accompanied by either Spanish, Portuguese, or Dutch), one in Yiddish, 89 in English, 32 in Dutch, 3 in French, and 433 in Spanish” (Joubert & Perl 2007: 46). 58 In the late 19th century, Portuguese ceased to be actively used in the synagogues (Emmanuel & Emmanuel 1970: 359; Granda 1974: 3; Karner 1969: 25, 26). As far as I am aware, Portuguese has since largely disappeared from Curaçao. A minority of labor migrants from Madeira arrived on Curaçao in the early 20th century to work for the Dutch Shell oil company, but I have no indications to assume that their demographic and/or linguistic impact on Curacao and PA was significant. 59 Of the literature dedicated to the languages and dialects of the Sephardim, the vast majority is dedicated to the description of Judaeo-Spanish (e.g. Salas 1976) and Ladino (e.g. Busse 1999). The term Judaeo-Spanish (/ Judezmo) designates the whole of differentiated varieties of Spanish that developed in the late 16th and 17th centuries in the Sephardic communities that had fled Spain (Kerkhof 2003: 25). Scholars do not always clearly distinguish between Ladino and Judaeo-Spanish, or use the former to refer to the latter. Bossong (2005) reports on this notional confusion and provides clarity: “Bezüglich des Begriffes Ladino gibt es eine gewisse Konfusion. Vor allem in der angelsächsichen Welt bezeichnet man die Gesamtheit der umgangssprachlichen Dialekte der Sepharden oft als ‘ladino’. Diese Verwendung des Begriffes führt jedoch in die Irre, denn das Ladino ist eine präzise definierte Schriftform, eine übersetzungsgeprägte Varietät, kein judenspanischer Dialekt” [‘There is some confusion with respect to the term Ladino. Especially in the Anglo-Saxon world the whole of Sephardic vernacular dialects is often referred to as ‘ladino’. This application of the term, however, is incorrect, since Ladino is an accurately defined form of writing, a variety used for translations, not a Jewish-Spanish dialect’] (Bossong 2005: 90; cf. also Liba 2008: 33).

1.5. PA birth among the Sephardim?

33

epitaphs of the Beth Haim cemetary, however (Joubert & Perl 2007), surprisingly revealed not a single inscription in Judaeo-Spanish, so that we can discard it as an important language among the Curaçaoan Sephardim.61 The inscriptions clearly suggest that Portuguese and Spanish were the two important codes. Admittedly, as Emmanuel (1957: 113) observes, “[t]he number of inscriptions in Spanish is relatively small (…) despite the literary activity in Spanish displayed in Amsterdam toward the end of the seventeenth century by relatives and friends of the immigrants to Curaçao”. But does the numerical superiority of Portuguese inscriptions over Spanish forcibly lead to the conclusion that Portuguese was the dominant home language of the Sephardim? As I have already hinted at, I believe it does not. First, in line with Caro Baroja (cited in Granda 1974: 6), we may assume or at least hypothesize that Spanish-Portuguese bilingual Sephardic Jews who had fled the Iberian Peninsula “used Castillian with more frequency for judging it more widely known than Portuguese”62 . Especially in light of the geographic location of Curaçao and the Jewish interest in external trade, it is a plausible guess that Jews with knowledge of both Spanish and Portuguese settling on Curaçao soon opted for Spanish. This is acknowledged also by Emmanuel (1957: 113): “[T]he active commerce conducted along the Spanish coast of Venezuela, Santo Domingo, Puerto Rico, etc., compelled the Curaçaoan Jews to use (…) Spanish from a very early period”. The domains in which Spanish was preferred by the Sephardim, I assume, included – and this is ultimately of most interest – communication with their slaves. Luckily, more concrete indications are available suggesting that the early Jewish settlers may have spoken Spanish more habitually than Portuguese. The first Jew ever to set foot on Curaçao, for instance, was Samuel Cohen, who joined the Dutch conquering expedition in 1634 as a Spanish interpreter. Tellingly, Cohen is described by Goodman (1987: 367) as “a Spanish-speaking Portuguese Jew”, which is important to stress as it shows that the tag ‘Portuguese’ was primarily an ethnic, not a linguistic, indication. This ethnic label, ‘Portuguese

60 Original quote: “tres códigos lingüísticos de base hispánica: el castellano, el ladino y el portugués” 61 Joubert & Perl (2007: 47) observe in this regard: “Most surprising is the fact that no headstones in Ladino [i.e. Judeo-Spanish] have been found. Despite its being the language most characteristic of European Sephardic communities, there are no data indicating that it may ever have been used on Curaçao”. 62 Original quote: “usaron con más frecuencia la lengua castellana, por juzgarla más conocida que la portuguesa”

34

Critical review of the literature on the origins of Papiamentu

Jew’, is one reason why many have uncritically assumed that the Sephardim spoke Portuguese. Moreover, a certain Charles Longland reported in 1657 of the many Jewish families migrating to the Dutch New World colonies, including Curaçao. Böhm (1992: 110), discussing this letter, observed that “Longland makes mention of the ‘numerous privileges’ granted to the Jews, who, according to him, had to be of great use to the Dutch because of their dominion of the Spanish language.”63 (cf. Klooster 2001: 354). Perhaps the most clear-cut indication of the importance of Spanish in the linguistic profile of the Curaçaoan Jews is provided by a letter “from October 1654 (…) written in Spanish by a Curaçaoan Jew directed to the Mahamad of Curaçao’s Mikvé Israel Congregation”64 (Böhm 1992: 172). Obviously, this letter would not have been written in Spanish, if that language had not been a very common vehicle of communication among the Curaçaoan Jews at that time. It is a fact that some of the first Jews to settle on Curaçao had previously lived in Dutch Brazil and one might be tempted to argue that these are likely to have used Portuguese colloquially. This reasoning does not apply, however: firstly, in that period, Portuguese was not yet generally diffused in Brazil (Thomas Krefeld, p.c.); secondly, the Sephardic community in Amsterdam had received no less than 600 Jewish families after the Dutch Brazilian adventure had failed (Joubert & Perl 2007: 45) and nonetheless, “the Sephardic communities generally called ‘Portuguese’ used Castilian and Ladino amply and commonly until the 19th century to such an extreme that Spanish serves in Amsterdam as lengua vulgar”65 (Granda 1974: 5).66 It is important to reflect more on the vernacular of the Amsterdam Sephardic community, since not few of the first Jews to settle on Curaçao came directly 63 Original quote: “Longland (…) hace mención de los ‘numerosos privilegios’otorgados a los judíos, los que, según él, habían de servir enormemente a los holandeses por su dominio del idioma español” 64 Original quote: “de octubre de 1654 data una carta escrita en español por un judío de Curaçao y dirigida (…) al ‘Mahamad’ de la Congregación ‘Mikvé Israel’, Curaçao” 65 Original quote: “las comunidades sefarditas llamadas generalmente ‘portuguesas’ (…) hasta el siglo XIX manejaban normal y ampliamente la lengua castellana y el ladino hasta el extremo de que el español figure en Amsterdam como ‘lengua vulgar’” 66 Granda thereby based himself on the title of a translation of a book originally written in Hebrew, by Isaac Aboab: “Almenara de la luz … compuesta en lengua ebraica por el gran sabio … Traducida en lengua bulgar para beneficio común por el Haham Iahacob Hages”, Amsterdam.

1.5. PA birth among the Sephardim?

35

from Amsterdam and most of the ‘Brazilian’ Jews arriving on Curaçao had previously lived in Amsterdam as well (Joubert & Perl 2007: 46). Regarding the Amsterdam Jews, den Boer asserts: “For the Portuguese as well as for the Spanish Jews the language of re-education was Spanish”67 (1996: 36). A good indication of the validity of that remark is that one of the most influential personalities of the Amsterdam Sephardic community was Menasseh Ben Israel (1604–1657), a Rabi of Portuguese origin: “He wrote mostly in Hebrew and Spanish in order to reach his target audience, the Sephardic Jews”68 (Delgado 2003: 286). In 1650, Schmidt (2001: 102) notes, one of Menasseh’s key writings “was published simultaneously in Spanish and Latin (…), thus assuring a wide audience both within and without the Sephardic community of Amsterdam. All told, at least thirteen editions appeared through 1723 in Latin, Spanish, English, Dutch, Yiddish, and Hebrew”. Had Portuguese been the dominant vernacular within the Amsterdam Jewish community, we would surely have expected Menasseh’s work to have been translated into Portuguese. What we know with some certainty is that by the end of the 18th century, PA had become the native tongue of large segments of the Curaçaoan society, including the Sephardim. This is suggested by the famous 1775 letter by a Curaçaoan Jew to his mistress69 written in flawless PA. Interestingly, this letter provides some clues to the linguistic repertoire of its Jewish author. If among the Sephardim, as affirmed by Emmanuel (1957: 112, cited above), Portuguese had been the dominant vernacular until 1865, then we would surely expect additional Portuguese material to pop up in the 1775 letter when compared to modern standard PA. However, as correctly noted by Wood (1972a: 28), “[t]his letter by a Portuguese Jew is singularly unproductive of Portuguese linguistic material”. One exception is the appearance of non-native in line 14 of the letter, but this word is “a voz culta with strong religious overtones and used, in this letter, in a religious context, [which] might be expected in the writing of a Sephardic Jew whose religious service was regularly conducted in Portuguese, especially the sermons (…)” (Wood 1972a: 28).70 Instead of Lusitanisms, the 67 Original quote: “Tanto para los judíos portugueses como para los españoles la lengua de reeducación fue el español” 68 Original quote: “Rabbiner portugiesischer Herkunft, Menasseh Ben Israel (1604– 1657) (…). Er schrieb zumeist in hebräischer und spanischer Sprache, wohl um die Primäradressaten, die sephardischen Juden zu erreichen” 69 This letter is reproduced and discussed for instance in Maduro (1971), Wood (1972a), Salomon (1982), Martinus (1996: 9, 10) and Kramer (2004: 217–222). 70 Kramer (2004: 219) explains the appearance of aflição in a similar manner: “der Briefschreiber wollte wohl in einer hochemotionalen Passage einen Anklang an die Gottesdienstsprache Portugiesisch evozieren” [‘In this highly emotional passage,

36

Critical review of the literature on the origins of Papiamentu

letter in fact contains a remarkable Hispanism: the use of 71 ‘to say’ (probably pronounced [Habla]) < Sp. hablar ‘to say, speak’ (= Port. falar) instead of modern PA bisa or papia (cf. Maurer 1998: 199f.n.). The 1775 letter can thus be taken to hint at a situation of Spanish-Portuguese diglossia within (segments of) the 18th century Sephardic community, with Spanish being used alongside PA in daily speech (hence the appearance of instead of bisa) and Portuguese dominating in religious and formal contexts (hence the appearance of ).72 Above, I have discussed clues for the idea that before the Sephardim shifted to PA, Spanish had been their principle home language, even though Portuguese dominated in ceremonial and religious contexts. I thus do not coincide with Jeuda (1990), who believed that the many Sephardic wills and gravestone inscriptions written in Portuguese are evidence that “the leading language of the 17th and 18th century Curaçao Jews was Portuguese” (1990: 16). Rather, a comparison with Jewish communities in British America seems in place: “Through most of the 18th century, American synagogues conducted and recorded their business in Portuguese, even if their daily language was English”73 . This is confirmed by Sarna (2001: 522): “various prayers, including part of the prayer for the government, continued to be recited in Portuguese (…), even though only a minority of the members understood that language and most spoke English on a regular basis”. 1.5.2.

Demographic arguments against a PA birth among the Sephardim

To close this section, it is relevant to briefly examine the demographic composition of Curaçao in the period in which PA is thought to have emerged, to see if the available demographic data favor a PA birth scenario with the Sephardim as principle actors. The answer is: they do not. According to Maurer, “the Sephardim were numerically important only in the last decades of the 17th century, that is after the massive import of African slaves which made the emer-

the author of the letter probably wanted to evoke an association with Portuguese as the language of religion’]. 71 The grapheme here corresponds to the phoneme /h/. This apparently misled Wood (1972a: 25), who transcribed as “Gabla (a name?)”. 72 Klooster (2001: 358), moreover, refers to a diary kept by a Spaniard temporarily residing on Curaçao, who in 1743 reported of Jews celebrating in the streets of Curaçao with “loud voices in the Spanish tongue”. 73 http://en.wikipedia.org/wiki/Sephardi_Jews

1.5. PA birth among the Sephardim?

37

gence of a creole necessary”74 (1998: 200). And even though they soon formed an important community on Curaçao, “the Sephardim did not come to possess more than 25% of the slaves”75 (1998: 199; cf. Maurer 1986b: 98, drawing on Emmanuel & Emmanuel 1970: 227). Indeed, it is unlikely that the Curaçaoan Jews ever became the dominant slave owners; from the onset they were more concerned with inter-Caribbean trade than with massive agriculture – especially after the latter activity had proven unprofitable given the island’s dry soil. Consequently, a situation demographically comparable to a société de plantation under Sephardic supervision never existed on Curaçao (Holm 2000: 77). The slave camps controlled by the Dutch WIC came much closer to providing a prototypical environment for creolization. But then again, a Dutch creole never emerged on Curaçao. (In fact, the number of Dutch-derived function words can be counted on the fingers of one hand and seem not to belong to the oldest layers of the grammar.76) For similar historical-demographic reasons, Kramer also doubted the plausibility of PA emerging on Curaçao among the Sephardic Jews: “Then how could the Sephardic Portuguese get to the black slaves working under Dutch command, who undoubtedly constituted the vast majority?”77 (2004: 116). He added that, if Sephardic Portuguese contributed at all to the shaping of PA, we might find rem-

74 Original quote: “los sefardíes (…) sólo fueron importantes numéricamente en las últimas décadas del siglo XVII, es decir después de la importación masiva de esclavos africanos que hizo necesaria la emergencia de un idioma criollo” 75 Original quote: “los sefardíes no llegaron a poseer más del 25% de los esclavos” 76 Dutch-derived function words found in modern PA, such as the coordinate conjunction òf ‘or’ (< Du. of ), the composed preposition dor di (< Du. door + PA di) as used in passive clauses to express the agent, are to my knowledge not found in Early PA texts, where PA o ‘or’ and pa ‘by’ fulfilled these functions (as they still do). Both Martinus (1996: 29, 30) and Maurer (1998: 195) noted that some Dutch-derived words in PA can be traced back to dialectal Dutch spoken in the 17th century and that some of the Dutch-derived lexemes show the type of sound adaptation (e.g. vowel epenthesis) typical of the older strata of PA’s lexicon. But Maurer also stressed that the Dutch contribution to PA is primarily lexical, whereas the grammatical morphemes derive mainly from Iberian material. For more on the Dutch contribution to PA, consult Wood (1970, 1972b), Kowallik & Kramer (1994) or Kramer (2004: 139–155). 77 Original quote: “Wie sollte denn das [sephardisch] Portugiesische zu den unter niederländischem Kommando arbeitenden schwarzen Sklaven – zweifellos die erdrückende Mehrheit – gekommen sein?”

38

Critical review of the literature on the origins of Papiamentu

nants of this in the lexicon, but not in the core grammar (Kramer 2004: 117).78 Once more, Lenz’s (1928) early accuracy of judgment is noteworthy: I can not verify whether the arrival of the Jews expelled from Brazil has raised the number of Portuguese words. In any case, it would be incorrect to attribute the Portuguese elements only to them, as some Dutch authors have done. We have seen that the large majority of the fundamental elements of PA’s grammar, the prepositions, the pronouns, the verbal particles, etc., is Portuguese79 (1928: 327, 328)

A further problem of attributing PA’s Lusophone elements to the Sephardim is that we lack detailed descriptions of what their Portuguese was like. This poses an empirical problem, for instance, when we isolate a set of archaic 15th-17th century Portuguese features present in PA’s core grammar (see §6.4): it is impossible to tell whether these were present at all in the Sephardic ethnolect(s). On the other hand, as Kramer (2004) insightfully argued, we may find interesting clues in the Surinamese creole Saramaccan. While one part of this creole’s lexicon is and always has been English-based, another considerable part was probably drawn from the Portuguese presumably spoken by Brazilian Jewish refugees. Therefore, if the Brazilian Jews that settled on Curaçao were responsible for the Lusitanisms in PA, “then we should be able to find marked correspondences between Papiamentu and the Portuguese elements of Saramaccan. Now, not even with the widest stretch of imagination can one credibly claim this to be the case: characteristic parallels between Saramaccan and Papiamentu can not be detected”80 (Kramer 2004: 116).

78 Henriquez (1988, 1991) is the authority on the lexical peculiarities of Sephardic PA. The dictionary by van Putte & van Putte-de Windt (2005a, b) is a useful source as well, with the abbreviation ‘Sef.’ indicating if a word is typical of Sephardic PA. 79 Original quote: “Si la llegada de los judíos espulsados del Brasil (…) ha aumentado las voces portuguesas, no lo puedo comprobar. En todo caso, sería falso atribuir los elementos portugueses del curazoleño sólo a estos, como lo han hecho algunos autores holandeses. Hemos visto que los elementos fundamentales de la gramática del papiamento, las preposiciones, los pronombres, las partículas verbales, etc. son en gran mayoría portugueses” 80 Original quote: “dann müsste es auffällige Berührungen geben zwischen dem Papiamento und den portugiesischen Bestandteilen des Saramakkischen in Suriname (…). Davon kann nun aber beim besten Willen nicht die Rede sein: Charakteristische Parallelen zwischen dem Saramakkischen und dem Papiamento lassen sich nicht nachweisen”

1.6. Where does the Portuguese come from?

39

1.6. Where does the Portuguese come from? As I hope to have demonstrated above, both the idea that Curaçao is the birthplace of PA and the related linguistic arguments aimed at trivializing its AfroPortuguese elements are often tenuous or at odds with our knowledge of the socio-demographic and linguistic history of Curaçao. The main flaw of the scenarios that situate the birth of PA on Curaçao is the failure to account for the fact that the apparently Portuguese features are centered in PA’s core vocabulary. Hence, starting with Lenz (1928), a sizeable number of scholars have assumed the obvious, namely that PA descends from an already restructured AfroPortuguese speech variety imported from elsewhere. Four possible sources have since surfaced: an Afro-Portuguese proto-pidgin from which all Afro-Iberian varieties in the Caribbean descend; a Brazilian Creole; Gulf of Guinea PC; and Upper Guinea PC. I will discuss each of these critically below, closing with the one candidate that strikes me as by far the most plausible one: Upper Guinea PC. 1.6.1. A shared origin for all Afro-Iberian creoles in the Caribbean? While creolists were compelled to abandon extreme forms of monogenesis well before the 1990s, the idea of a proto-pidgin spreading from West Africa to the West Indies did not cease to appeal to scholars concerned with the Iberian-based and Afro-Hispanic speech varieties of the Americas. Thus, a moderate form of monogenesis, also known as Afrogenesis, survives up to present according to which the Iberian-lexifier creoles in the Western hemisphere can be traced back to a single West African Portuguese-based proto-variety, typically said to have been a stable expanded pidgin. Martinus (1996), himself a proponent of Afrogenesis, critically remarked that “[t]he main shortcoming of the Proto-Afro-Portuguese creole theory is that its supporters [have] failed so far in building a precise and consistent framework (…) and in presenting hard empirical data to fill in the combination of logical assumptions and arguments it consists of” (1996: 14). Indeed, the proposal that PA and other Iberian-based creoles spoken in the Americas all sprang from one and the same Afro-Portuguese proto-pidgin is rather conjectural, given that no unambiguous records exist of this proto-pidgin (or jargon): “Unalloyed testimonies of this Portuguese trade jargon are of course absent; after all, we are dealing with a communication vehicle intended for oral use only”81 (Kramer 2004: 126). Ladhams (2006: 97) speaks in this context of “the lack – and unreliability – of 81 Original quote: “Unverfälschte Zeugnisse dieser portugiesischen Handelssprache gibt es natürlich nicht, denn es handelte sich ja um ein nur für den mündlichen Gebrauch gedachtes Kommunikationsmittel”

40

Critical review of the literature on the origins of Papiamentu

historical texts, as well as the lack of precision in historical references to the language” (cf. e.g. Ferraz 1987 and Bickerton for more outspoken criticism). Despite the lacking documentation, however, the existence and knowledge of the alleged proto-pidgin is often taken for granted, as Portilla (2008b: 164) illustrates: “[T]he slaves brought by the Jews from Brazil to Suriname and later Curaçao originated from the same regions in Africa where an Afro-Portuguese pidgin was already used.”82 Admittedly, testimonies regarding the existence and use of some form of ‘broken’ or ‘corrupt’ Portuguese abound in travelers’ accounts. Ladhams (2006) provides an excellent overview. Many of these testimonies were, however, drawn up in regions known to have been settled by native Cape Verdeans (e.g. Rufisque, Joal, Cacheu) (cf. §7.5). Some testimonies also came from the Gulf of Guinea, where Gulf of Guinea PC is and was natively spoken. Most of these travelers’ testimonies may thus be interpreted as testifying to the presence of native speakers of either Upper Guinea PC or Gulf of Guinea PC in those regions, rather than to the use of an expanded stable pidgin. Similarly, Father Sandoval’s 1627 testimony (e.g Martinus 1996: 16) has often been invoked as evidence of the diffusion of the proto-pidgin to the New World. Here, though, it is relatively clear that we are dealing with native speakers of Gulf of Guinea PC (“criollos y naturales de San Thomé”) rather than with Portuguese pidgin-speakers. I do not wish to rule out the possibility that an expanded stable Portuguese-based pidgin was once in use along parts of the African West Coast and perhaps also in the Caribbean. The main point is rather that speculating about ties between that hypothetical pidgin and PA is unlikely to yield any new insights into the origins of the Portuguese features in PA. There is also some terminological inconsistency within the Afrogenesis literature. Not rarely, the alleged proto-variety is typified as ‘lingua franca’, in which case the relationship to the Mediterranean Lingua Franca remains unclear.83 In other cases, the terms ‘pidgin’ and ‘creole’ are simply used in free variation (sometimes on one and the same page), without any attempt at distinguishing 82 Original quote: “los esclavos llevados de Brasil a Surinam y luego a Curazao por los judíos, provenían de las mismas regiones en África en donde se utilizaba ya un pidgin afroportugués” 83 Alleyne (1980: 131) draws attention to this, when criticizing Berry (1971) for using the term ‘Afro-Portuguese’ “without assigning any real African content to the phenomenon, since in the same context he reports that ‘the Afro-Portuguese lingua franca (is) possibly sabir, the ancient lingua franca of the Mediterranean.’The idea that sabir might be the common ancestor of Afro-American dialects was suggested by Whinnom (1965) and is really the end point in speculation on the European origin of these dialects”.

1.6. Where does the Portuguese come from?

41

between these linguistic types. This distinction is however crucial in the context of PA’s origins. For instance, as opposed to creoles, pidgins generally lack complex grammatical categories.84 Any hypothesized Afro-Portuguese protopidgin, therefore, by definition does little to account for Lenz’s previously cited affirmation that PA’s grammar and function words are mainly Afro-Portuguese. (More discussion on this issue is provided in §8.1.2.1.) Alleyne (1980) excellently clarifies how the failure to reflect on the distinction between pidgins and creoles damages monogenesis frameworks: [S]ince this ‘Portuguese pidgin hypothesis’ looks at creole languages and infers a Portuguese pidgin for them, the former (creole languages) would have to be ‘expansions’and not merely ‘relexifications’of the latter (Portuguese pidgin). But if all the structural features commonly found in creoles are also to be reconstructed in their putative origin – the Portuguese ‘pidgin’ – then it is not clear what the ‘expansions’ are. Again the conclusion is inescapable that the pidgin hypothesis is an uninteresting and trivial speculation with little explanatory power. (Alleyne 1980: 130; cf. §8.1.2.1)

Portilla (2007, 2008a, 2008b) is the latest addition to the voluminous scholarship that sustains a common ancestor for PA and the other Iberian-based creoles in the Western hemisphere, particularly PLQ and Saramaccan. It is important to point out, however, that the Afro-Portuguese origins for either Saramaccan or PLQ receive little support nowadays. Starting with PLQ, the idea that it descends from an Afro-Portuguese pidgin or creole had been nourished primarily by the afore-mentioned 1627 testimony of Father Sandoval as well as by some apparently (Afro-)Portuguese features in its core grammar. Recently, however, Afrogenesis for PLQ was tackled convincingly by Bickerton (2002). Also Schwegler, the foremost authority on PLQ and once a supporter of its Afrogenesis (e.g. Schwegler 1993), has now abandoned that idea: “[T]he precise genetic relationship between Palenquero and Papiamentu (…) and between Palenquero and Sãotomense (a creole once thought to be the direct source of Palenquero) is unknown” (Schwegler & Green 2007: 273). Furthermore, Maurer (1987a) presented a comparison of PA’s and PLQ’s TMA systems that allowed for only one conclusion, namely that “the monogenetic the84 When discussing pidgins, distinctions should preferably be made between the different developmental stages. Scholars commonly distinguish between jargons, pidgins, stable pidgins, and expanded pidgins (cf. Mühlhäusler 1997: 12). Bakker (2008a), an authority in the field of pidgin studies, prefers to speak of jargons, pidgins and pidgincreoles. To the best of my knowledge, in the accounts that have proposed a proto-pidgin ancestor for PA, such distinctions have not been made. (See §8.1.2.1 for a more elaborate discussion.)

42

Critical review of the literature on the origins of Papiamentu

ory, even in its restricted version, is not capable of explaining all the differences between Papiamentu and Palenquero”85 (1987a: 67; cf. Maurer 1998: 201). In the light of these critical assessments, it seems premature at the very least to assume that PA and PLQ descend from one and the same Afro-Portuguese proto-pidgin. Saramaccan is correctly described by Bakker (2009: 156) as “somewhat odd among the creoles because its basic vocabulary is not from one language, but from two, viz. Portuguese and English”. Early on, this fact inevitably led to the portrayal of Saramaccan as a descendant of the alleged Portuguese-based protovariety, relexified towards English (e.g. Voorhoeve 1973), but nowadays there is a broad consensus that “the Portuguese elements were superimposed on a preexisting English frame” (Bakker 2009: 155; cf. Cardoso & Smith 2004: 115, McWhorter 2000: 101; Parkvall 2000: 125; Muysken 2008: 209; Good 2009), this English frame most probably being Early Sranan (Smith 1987, McWhorter 2011: 172, 173). Holm sums up: “Some have classified [Saramaccan] as a Portuguese-based creole (…), but most creolists since Schuchardt [1914] have treated it as an English-based creole”. Thus, there appears to be little basis, either historical or linguistic, for searching for a shared ancestor pidgin for PA and Saramaccan. This section will conclude with some comments on the tendency to draw parallels between PA and the reconnaissance language (Lingua de Preto) found in Gil Vicente’s plays (cf. Naro 1978). As noted in §1.2, a group of scholars (e.g. Granda, Birmingham, Perl, Megenney) claimed that this reconnaissance language was the starting point for a Portuguese-based proto-pidgin out of which PA ultimately developed. More recently, this possibility was explored by Martinus (1996: 122, 123) and Kramer (2004: 127–134).The reconnaissance language reflected in Vicente’s plays does indeed contain several features that could be attributed to pidginization such as the generalization of infinitives, tense and modality expressed by adverbs, marginal use of plural marking, or the use of Port. nunca as a simple negator (e.g. Holm 1988: 269). However, and however interesting, particularly Lipski (2005: 54–57) has shown that the features found in Vicente’s plays would not be unexpected in any restructured Afro-Iberian variety. As Lipski (2005: 55) notes, Vicente’s plays contain “traits reflecting both the imperfect acquisition of Portuguese by adult speakers of other languages, and direct interference from African areal characteristics”. Moreover, “Vicente’s examples show a rather wide range of dispersion, ranging from standard (sixteenth-century) Portuguese forms to drastically reduced or misinter85 Original quote: “la théorie monogénétique, même dans sa version restreinte, n’est pas en mesure d’expliquer toutes les différences entre le papiamento et le PLQ”

1.6. Where does the Portuguese come from?

43

preted elements. (…) There are some differences among the texts, but nothing suggests any form of linguistic evolution” (Lipski 2005: 55). A look at some individual features is not clarifying either. For instance, the use of the 1st singular pronoun /ami/ as a subject pronoun is attested not only in Vicente’s plays, but also in a number of other Afro-Iberian historical texts, and it is attested not only in PA, but also in all of the West African Portuguese-based creoles, as well as in several (non-creole) Afro-Hispanic varieties of the Americas (cf. particularly Lipski 1991). That in Vicente’s plays a temporal adverb meaning ‘immediately, later’ – Portuguese logo – has a future reference is not particularly surprising either (see §5.3). Therefore, linking PA specifically to Vicente’s plays is unlikely to yield any novel results in terms of the creole’s origins or genetic affiliation. Whether the reconnaissance language (or any Portuguese-based pidgin or foreigner talk) was ever used among, or to address the African slaves in the camps of São Tomé and Santiago is quite a different matter. An already reduced and/or partially restructured variety of Portuguese may very well have formed the starting point of two independent processes of (pidginization and subsequent) creolization, one in Upper Guinea and another in the Gulf of Guinea, but we are unlikely to ever find out. Note, also that accepting such a scenario merely means affirming what we already know, namely that both Upper Guinea PC and Gulf of Guinea PC have a Portuguese superstrate. 1.6.2.

Goodman’s Brazilian Creole Hypothesis

Goodman (1987: 361) recalled that “[i]t is indebted (…) to the views expressed by Lenz (1928), Navarro Tomás (1953) and van Wijk (1958) that Papiamento, (…) although its vocabulary is predominantly of Spanish origin, grew out of West African Pidgin Portuguese”. However, whereas these authors (and others up to the mid-1980s) had traced the Afro-Portuguese features in PA directly back to the African West Coast, Goodman (1987) reassessed and elaborated on Valkhoff’s (1960: 81) assertion that PA “must have been originally a creole Portuguese dialect brought by Negro slaves from Brazil”.86 According to Goodman, 86 My discussion of Goodman (1987) here is restricted to his observations and claims concerning PA, but his account was of course not limited to PA. Bickerton (2002: 35), for instance, praised Goodman’s paper for having shown that “todos los criollos que tenían más que una mínima cantidad de vocabulario portugués nacieron o en territorio ocupado por los portugueses en alguna época (…) o en lugares donde moraba una minoría considerable de gente de habla portuguesa (como Suriname)” (2002: 35) [‘all creoles that had more than a minimal amount of Portuguese vocabulary were formed either in territory once occupied by the Portuguese or in places where a considerable Portuguese-speaking minority was living such as Suriname’].

44

Critical review of the literature on the origins of Papiamentu

the precursor of PA had been imported into Curaçao in the 1650s “by Dutch and Jewish refugees from Brazil accompanied by Creole Portuguese-speaking servants” (1987: 368). The diffusion of the same Brazilian Creole to Surinam was claimed by Goodman to account for the Portuguese-derived lexicon in Saramaccan. Goodman’s hypothesis had some impact on PA studies. Jeuda (1990: 21), for instance, found it “the most thorough accounting of the Portuguese element in the New World creoles” and McWhorter’s (2000: 13) earlier cited assumption that PA has “been shown to have originated as Portuguese-based” was also based primarily on Goodman’s 1987 paper. Crucial parts of Goodman’s framework, however, were later rejected by, for instance, Martinus (1996: 91–93), Maurer (1998: 198, 199) and, with special respect to the case of Saramaccan, by Ladhams (1999a, 1999b) and Arends (1999). As far as PA is concerned, two principle flaws in Goodman’s account have been highlighted. A first issue is pointed out by Maurer (1998: 198): “Until now there is no direct or indirect evidence of a Brazilian creole spoken in the part of Brazil occupied by the Dutch from 1630 to 1654”87 (cf. Rougé 2008). It is therefore difficult to empirically corroborate or contradict Goodman’s claims. Joubert & Perl (2007: 46) address a second, strictly historical issue: for logistic reasons, “[t]he Sephardic exiles from Brazil (…) did not take with them many slaves due to logistic restraints”. 1.6.3.

Gulf of Guinea Portuguese-based Creole

After the appearance of Maurer (1988), a synchronic description of PA’s verbal system and rightly considered a milestone in PA studies, Maurer dedicated a series of articles (1989, 1991, 1994a, 2002, 2005) to the identification and analysis of African and Afro-Portuguese features in PA. Maurer’s approach is appealing insofar as he placed his linguistic findings against the background of the early history of Curaçao and the origins of the slaves arriving in the 17th century. According to Maurer (1998: 191), “the WIC was accustomed to buying a third of the slaves in the Congo-Angola region and the rest especially in the Ghana-Togo-Benín area”88 (cf. §7.1). This historical assumption was an important reason why Maurer focused on the languages of Lower Guinea and Angola in his attempts to get a clearer picture of PA’s ancestry and formation. 87 Original quote:“Hasta ahora no se tiene ninguna evidencia directa o indirecta de un pidgin o criollo brasileño hablado en la parte del Brasil ocupada por los holandeses de 1630 a 1654” 88 Original quote: “la WIC solía comprar la tercera parte de los esclavos en la región Congo-Angola, y el resto sobre todo en el área Ghana-Togo-Benín”

1.6. Where does the Portuguese come from?

45

He managed to identify several Benue-Kwa and Bantu features in PA’s substrate (Maurer 1987b, 1989, 1991, 1994a), thus lending support to the idea that the creole’s origins are to be sought for in the coastal region stretching from Lower Guinea to Angola. Furthermore, his profound knowledge of both PA (cf. Maurer 1988, 1998) and Gulf of Guinea PC (cf. Maurer 1994b; 1995; 2009) allowed Maurer to recognize and describe parallels between them. He expresses his view on the relatedness of these creoles as follows: One can establish a genetic relationship between the Gulf of Guinea creoles and Papiamentu. However, a more complete comparison between the creoles does not allow us to postulate that Papiamentu constitutes a relexification of a Gulf of Guinea variety, since the grammatical differences are too big. If the hypothesis of a direct influence of the Gulf of Guinea creoles in the formation of Papiamentu is correct, we would have to consider these languages as part of Papiamentu’s substrate89 (Maurer 2002: 135; cf. Maurer 2005: 58, 66)

As the excerpt reveals, on the whole, Maurer appears to support a Curaçaosituated birth scenario that included an important influence of speakers of Gulf of Guinea PC (cf. Maurer 1998: 193–202). He explains PA’s mixed Iberian-based lexicon as a result of the fact that the dominating class on Curaçao was never linguistically homogeneous (cf. Maurer 1998: 193–202). Some aspects of Maurer’s framework, such as his placing the origins of PA in the period between 1650 and 1700, have in fact found wide acceptance in PA studies. On the other hand, although Maurer’s view is in all facets more differentiated and better supported historically than most preceding Curaçaobirth hypotheses, it faces the same principal objection other Spanish hypotheses face in as far as it fails to account for the fact that the Portuguese features are found particularly in PA’s core grammar rather than being more randomly distributed. Nevertheless, Maurer’s contributions – a major reference in the present study – have been and will continue to be instrumental to the synchronic and diachronic description of PA, the analysis of the sociolinguistic history of Curaçao as well as to the identification and description of African and AfroPortuguese features in PA. 89 Original quote: “se puede establecer una relación genética entre los criollos del Golfo de Guinea y el papiamento. Sin embargo, una comparación más completa entre los criollos (…) no nos permite postular que el papiamento constituya una relexificación de una variedad de los criollos del Golfo de Guinea, pues las diferencias gramaticales son demasiado grandes (…). Si la hipótesis de una influencia directa de los criollos del Golfo de Guinea en la formación del papiamento es correcta, habría que considerar estos idiomas como lenguas de substrato del papiamento.”

46 1.6.4.

Critical review of the literature on the origins of Papiamentu

Upper Guinea Portuguese-based Creole

CV and GBC (the two sister varieties jointly referred to as Upper Guinea PC) have quite steadily entered the equation in most discussions of PA’s possible Afro-Portuguese origins and structural correspondences have not gone unnoticed (e.g. Lenz 1928; van Wijk 1958; Birmingham 1975; DeBose 1975; Goodman 1987; Maurer 1988; Martinus 1996). However, with the exception of Quint (2000b), an exclusive link between Upper Guinea PC and PA has never been postulated, at least not in a scholarly context (see APPENDIX I). Instead, CV and GBC are typically only mentioned as one of the Iberian-based creoles to which PA might owe its Afro-Portuguese character and thus come out as mere pieces of a larger monogenetic puzzle. A case in point is provided by Birmingham (1975, 1976, 1978). In Birmingham (1975: 22), he compared PA with CV, GBC and Annobonese and found that, “as regards the three WestAfrican Portuguese dialects under consideration, Cabo Verde is the one which perhaps most closely resembles Papiamentu, for there are many words (…) which have almost exact counterparts in Papiamentu”. Some years later, Birmingham (1979, cited in Megenney 1984: 180) again reported on “curious similarities between PA and the Afro-Portuguese dialect of, for example, the Cape Verde islands”. Nonetheless, Birmingham’s faith in monogenesis was not undermined: “The obvious solution (…) is that Papiamentu was formed in the slave camps along the west coast of Africa, where there was already an Afro-Portuguese lingua franca and that this creole language was imported to the New World (and other places) by the African slaves who were collected in the slave camps, pending shipment” (1979, cited in Megenney 1984: 180). A special position within the study of PA’s origins is taken by Martinus’ oeuvre (1996; 2003, 200790 ). It should above all be emphasized that his 1996 dissertation is instrumental to the description of the ties between PA, CV and GBC. Chapter 7 of his dissertation, revealingly titled ‘The Cape Verdean Creole Hypothesis’, identifies a rich amount of features shared by PA, CV and GBC in the lexicon, syntax and phonology. However, Martinus’ bent for monogenesis also shines through: “[T]he comparison is basically one between Cape Verdean creole, Guiné-Bissau creole and Papiamentu. Yet in the light of the wider ProtoAfro-Portuguese creole theory, Congolese languages where Portuguese relics are to be found will also be included on the African side and other creoles like Saramaccan, Sranan, Palenquero on the Caribbean one” (Martinus 1996: 146). Both the linguistic and historical arguments adduced by Martinus to justify the inclusion of any creole other than Upper Guinea PC turn out rather thin 90 Martinus (2007) is a concise version of his 1996 dissertation written in PA.

1.6. Where does the Portuguese come from?

47

and largely depend on an analysis of fragments of a secret linguistic code or cryptolect known as Guene (1996: 192–263 & 284–292; cf. Martinus 2008).This secret code is now extinct but was presumably part of the linguistic repertoire of segments of the Curaçaoan slave population well into the 19th century.91 Although I disagree with Martinus’ underlying monogenesis framework and with the role he attributes to the fragmentary Guene material, this criticism is made in hindsight and should not obscure the fact that Martinus planted the seeds for subsequent research into the relationships between PA and Upper Guinea PC; Martinus (1996) will accordingly be a rich source for the present study as well. Moving further away from monogenesis, Quint (2000b) linked PA exclusively to Upper Guinea PC, classifying PA, CV and GBC as members of a separate branch, thereby making a clear-cut distinction between those creoles on the one hand and Gulf of Guinea PC on the other. Quint’s (2000b) systematic exposition of linguistic correspondences between PA and the Santiago variety of CV (SCV) on the one hand, and SCV and GBC on the other, legitimize the claim that CV, GBC and PA “form a branch of closely related creoles descended from a common ancestor”92 (2000b: 237). He attributed the dominance of Spanish-derived vocabulary in PA to a “partial relexification favored by the use of Spanish as the language of religion and prestige”93 (Quint 2000b: 197). In addition, Quint devoted a section to the description of both the differences and similarities between PA and Gulf of Guinea PC, in order to strengthen his claim that PA and Upper Guinea PC form a separate branch. I will return to this in §6.3.1. What makes Quint’s hypothesis unique within the existing literature on PA’s origins is the fact that it supports the often postulated view that PA was formed on the African West Coast and also relies crucially on the concept of relexification, but, unlike earlier scholarship, does not embed these claims in a monogenesis framework. The firm inclusion of PA in monogenesis frameworks (cf. §1.2) had as a side effect that PA’s relexification hypothesis became inextricably associated with monogenesis (more to this in §8.2.1). Quint’s (2000b) work, however, shows 91 The difficult to analyze Guene material contained in Martinus (1996) is drawn from the works of van Meeteren (1947) and Brenneker (1959, 1986). As noted by Bickerton (1999: 98) in an otherwise too critical review of Martinus (1996), this Guene material is largely limited to “fragmentary songs, riddles, proverbs, and isolated words cropping up in what is otherwise standard Papiamentu”. 92 Original quote: “Le badiais, le papiamento des îles ABC et le créole de GuinéeCasamance forment un groupe de créoles étroitement apparentés, tous issus d’un stade commun” 93 Original quote: “relexification partielle, favorisée par l’usage de l’espagnol comme langue religieuse et de prestige”

48

Critical review of the literature on the origins of Papiamentu

that PA’s relexification hypothesis and monogenesis must be kept apart and that arguing for the former does not equal supporting the latter. It is also important that, unlike monogenesis, Quint’s hypothesis can be tested by means of traditional language comparison. However, any recent attempts to do that are not known to the present writer and references to Quint (2000b) are absent in subsequent writings concerned with defining PA’s position within the family of creole languages. Quite recently, Lipski (2008: 547) summarized the state of the research as follows: Theories on the origin and formation of Papiamentu cluster around three possibilities. First, Papiamentu may be the relexification of an Afro-Portuguese protocreole which also underlies other Atlantic creoles (…). Another view is that Papiamentu was originally a Portuguese-based creole, formed on Curaçao through the infusion of Portuguese-speaking Sephardic Jews from northwestern Brazil, when this region was recaptured from the Dutch by the Portuguese in the 17th century (…). Finally, another school of thought considers Papiamentu to be in essence a Spanish-related creole, whose Portuguese elements were introduced by Sephardic Jews and Portuguese slavetraders (…).

Indeed, a reference to the hypothesis formulated by Quint and examined in detail in the present study is missing.

1.7. Summary The aim of this chapter was to sketch the development of the debate on the origins of PA and to critically discuss the hypotheses that have been central in this debate. What stands out in this overview is the overwhelming attention PA has received within monogenesis models. This tendency started with Lenz (1928) and was recently revitalized by Portilla (e.g. 2008b). The popular view of PA as a descendant of an alleged Afro-Portuguese protopidgin was challenged early on by a group of scholars led by Maduro, who defended the hypothesis that PA is no more and no less than a creolized descendant of the Spanish spoken prior to the arrival of the Dutch. The case of PA was now emblematic of the larger mono- versus polygenesis debate that in the 60s and 70s dominated the field of creole studies. With the demise of traditional monogenesis, however, more differentiated accounts of PA’s formation were presented in the shape of Goodman’s Brazilian Creole Hypothesis, Maurer’s oeuvre, Martinus (1996) and Quint (2000b). The critical discussion of each of the hypotheses involved allows for the following preliminary conclusions to be drawn:

1.7. Summary

49

– The Spanish Curaçao-situated birth hypotheses account neither for the fact that the apparent Portuguese-derived words and features are centered in PA’s core vocabulary nor for the fact that PA is an Iberian- rather than a Dutchlexified creole. – Monogenesis frameworks and derived hypotheses (such as the ‘reconnaissance language hypothesis’) are not testable and thus add little to the understanding of the origins of PA and of the source(s) of its Portuguese features. – There is no reason (either linguistic or historical) to assume that Saramaccan, PLQ and PA have shared origins.

Chapter 2 Phonology

Introduction If we assume, for the sake of argumentation, that PA and Upper Guinea PC descend from a common source, the time-depth separating the two creoles must be slightly over 300 years (cf. Chapter 7 for historical details). Though that is considerable, it would surely be short enough for characteristic shared phonological features to have stood the test of time. Identifying the most striking of these features is the aim of this chapter. These features will include shared sound changes with respect to the Iberian etyma as well as the retention of (Old) Portuguese sounds and will, whenever possible, be contrasted to data from other Iberian-based creoles, particularly PLQ and Gulf of Guinea PC. To exemplify the sound changes in question, I will draw as much as possible on PA lexemes of probable Portuguese origin and PA lexemes of which the etymon can be either Spanish or Portuguese. These offer themselves as the most plausible objects of comparison, given that Spanish did not contribute to the lexicon of Upper Guinea PC. This tertium comparationis will have the additional advantage of bringing to light a significant part of the lexicon shared between PA and Upper Guinea PC. Under the heading “Observations”, I will provide relevant phonological as well as lexico-semantic comments. Whenever the phonetic transcription of an individual phoneme is indispensible to the understanding of the discussed sound change, the phoneme will be transcribed using IPA. In all other cases, the orthography of the consulted sources will be adopted and no phonetic transcription will be provided. In general, it is important to remember that the voiceless palatal fricative /S/ corresponds to the grapheme in Portuguese and Upper Guinea PC, but to the grapheme in PA; the palatal nasal /ñ/ corresponds to in Upper Guinea PC and Portuguese, but to in PA and Spanish. Unless indicated otherwise, I have drawn the lexical data from the following dictionaries: Mendes et al. (2002), Lang (2002), Rougé (2004a) for SCV; Rougé (1988 & 2004a), Scantamburlo (2002) for GBC; Ratzlaff (1992), Joubert (1999), van Putte & van Putte-de Windt (2005a & 2005b) for PA. In §2.1, vowel changes will be discussed; I will then concentrate on consonant changes in §2.2. In §2.3, I will deal with resyllabification patterns and then close the chapter with a digression on verb stress in PA and SCV.

52

Phonology

2.1. Vowel features Below, I will first address the systematic raising of the etymological unstressed mid-vowels /e, o/ to /i, u/, then focus on some shared cases of vowel rounding and vowel harmony, and furthermore highlight the preference for monophthongs that characterizes PA and Upper Guinea PC. 2.1.1. Vowel raising 2.1.1.1. Unstressed /e/ > /i/ As exemplified in Table 1, both in PA and in Upper Guinea PC, the etymological unstressed /e/ was systematically raised to /i/ in the fundamental part of the lexicon (cf. Martinus 1996: 153; Quint 2000b: 126, 130). Table 1. Unstressed /e/ > /i/ ETYMON

PA

Upper Guinea PC

GLOSS

Sp./Port. mole Sp./Port. senti Sp./Port. mentira Sp./Port. pescar

moli sinti mintira piska

moli sinti mintira piska

‘soft, weak’ ‘to feel’ ‘lie’ ‘to fish’

Observations: – In at least one shared case, the etymological unstressed /e/ was not raised but rather incorporated as /a/: Port. então ‘then’ (= Sp. entonces) > PA anto, SCV anton, GBC anta. – It is interesting to note that Sp./Port. pensa was incorporated in both PA and Upper Guinea PC as pensa, i.e. without the raising of the /e/, which can be taken to suggest that the verb was integrated more recently in both creoles. Indeed, in Upper Guinea PC, a more common and more conservative verb for ‘to think’ is kuda < Port. cuidar ‘to take care’ (Nicolas Quint p.c.). The retention of the /e/ in PA pensa suggests there was once another verb for ‘to think’ in PA as well, which was then replaced by pensa. – In a small series of lexemes, PA differs from Upper Guinea PC in that the former has given priority to vowel harmony over the raising of the /e/. Thus, Upper Guinea PC berdi ‘green’, ntendi ‘understand’ and perdi ‘lose’ correspond to PA bèrdè, tende and pèrdè. Here, a parallel with Gulf of Guinea PC is visible, which has the following cognates: ST vêdê, têndê and plêndê, PRI vêdê, têndê and pêdê and ANG vêrê, têndê, pêndê (Quint 2000b: 169, 170).

2.1. Vowel features

53

2.1.1.2. Unstressed /o/ > /u/ In the conservative part of the vocabulary, both PA and Upper Guinea PC systematically raised the unstressed etymological /o/ to /u/ (cf. Birmingham 1970: 156; Martinus 1996: 153, 154; Quint 2000b: 126, 130). Table 2 illustrates this feature. Table 2. Unstressed /o/ > /u/ 94 ETYMON Port. formiga (= Sp. hormiga) Port. costumar (= Sp. soler, acostumbrar) Port. borboleta (= Sp. mariposa) Port. começar (= Sp. comenzar) Old Port. dia domingo (see observations) Sp./Port. probar/provar Sp./Port. pros(s)eguir

PA

Upper Guinea PC

GLOSS

fruminga

‘ant’

kustumá 96

kustuma

barbulète kumisá

SCV barbuleta GBC burbuleta kumesa 97

‘to be accustomed’ ‘butterfly’

djadumingu

dia dumingu

‘Sunday’

purba pursiguí

purba pursigi

‘to taste, try’ ‘to persue’

fruminga

95

‘to begin’

Observations: – PA/SCV purba (< Sp./Port. probar/provar) can mean either ‘to taste’ or ‘to try’, as the context demands. To express the less essential concept ‘to prove’, on the other hand, the creoles have the forms PA proba and SCV prova, whose reduced age is suggested by the lack of the vowel raising. – PA djadumingu and Upper Guinea PC dia dumingu derive from the composition dia + domingo ‘day + Sunday’, which Quint (2000b: 220) claims was typical of Old (15th/16th-century) Portuguese. Gulf of Guinea PC also has composed (though phonetically clearly distinct) forms for ‘Sunday’: ANG diaringu and ST djadjingu (Quint 2000b: 220). 94 Portuguese-derived words with a word-final /u/ such as PA / Upper Guinea PC tempu ‘time’ ( /i/, here too, PA has preferred vowel harmony over the raising of the /u/ in a small number of conservative lexemes. For instance, we find PA nobo ‘new’ and porko ‘pig’ (< Port. novo, porco) versus Upper Guinea PC nobu and porku (Quint 2000b: 170). Again, the same vowel harmony is found in Gulf of Guinea PC: ST nôbô, plôkô; PRI/ANG pôkô (Quint 2000b: 170). These cases of vowel harmony shared between PA and Gulf of Guinea PC can be accounted for by the fact that Kwa and Bantu contributed to the development of both creoles, although a direct, postformative influence from Gulf of Guinea PC on PA cannot be discarded either. Note that, in line with the hypothesis defended in the present study, I believe that any contribution made by either Kwa/Bantu or Gulf of Guinea PC to PA was made postformatively. In this respect, it is interesting to note the occurrence in early 19th-century PA texts (e.g. Putman 1852) of the form /mortu/ ‘death, dead’ (cf. Upper Guinea PC mortu), suggesting that the vowel harmony of the modern PA variant morto (and perhaps of the other forms just mentioned) may well be a more recent innovation. – In two isolated shared cases, shown in Table 3, an etymologically stressed /o/ was raised to /u/. Table 3. Stressed /o/ > /u/ ETYMON Port. corpo (= Sp. cuerpo) Port. onde (= Sp. donde)

PA kurpa unda

98

Upper Guinea PC

GLOSS

GBC kurpu SCV unde/GBC nunde

‘body’ ‘where’

Contrast with the forms in Table 3 is provided by ST klôpô < Port. corpo (Rougé 2004a: 120) and PLQ (n)kueppo < Sp. cuerpo (Schwegler 1998: 233), 98 Spanish cuerpo ‘body’ was incorporated in PA as kuerpo with the specialized meaning ‘organization’.

2.1. Vowel features

55

as well as ST/ANG andji < Port. onde (Fontes 2007; Maurer 1995: 64), PLQ a(d)onde < Sp. donde (Schwegler 1996: 421) and Chabacano donde < Sp. donde (Lipski & Santoro 2007: 394). To account for the /u/ in PA unda and Upper Guinea PC unde/nunde, we should consider the possibility of an archaic Portuguese form *unde instead of modern onde, an idea strengthened by Korlai un (Clements 2007: 170) and Papia Kristang undi (Baxter 2004: 89), both ‘where’. 2.1.2.

Rounding of unstressed vowels

Table 4 shows two lexemes shared between PA and Upper Guinea PC which stand out for showing the rounding of an etymological unstressed unrounded vowel (Quint 2000b: 129; Martinus 1996: 149). Table 4. Rounding of unstressed vowels ETYMON Sp./Port. cavar Sp./Port. primero/primeiro

PA 99

koba promé

Upper Guinea PC

GLOSS

koba (Early SCV) promeru

‘to dig’ ‘first’

Observations: – Martinus (1996: 150, drawing on Veiga 1987: 35) reveals an interesting lexico-semantic parallel between PA and Upper Guinea PC koba, as both mean not only ‘to dig’ but also ‘to curse’. This is confirmed for SCV by Lang (2002: 318), for GBC by Scantamburlo (2002: 279), and for PA by van Putte & van Putte-de Windt (2005a: 219). – From the creole verb PA / Upper Guinea PC koba ‘to dig’, the noun kobador ‘digger’ (Dijkhoff 1993: 292; Scantamburlo 2002: 280; Lang 2002: 318) was productively derived (cf. §4.1 on the productivity of the PA / Upper Guinea PC suffix -dor). – The early SCV form promeru is attested in Brito (1887) and is absent from modern dictionaries such as Lang (2002), Mendes et al. (2002) and Rougé (2004a), all of whom give SCV purmeru∼prumeru. The early SCV form promeru versus modern prumeru suggests that etymological unrounded vowels /e, i/ passed through an /o/ before passing to /u/. This idea is further strengthened by data found in Schuchardt (1882: 141), who provides SCV borgonha ‘shame’, which in modern SCV is realized as burgónha (Lang 2002: 98). This, in turn, allows for the hypothesis that, in an earlier phase, 99 Lenz (1928: 252) suggests PA koba ‘to dig’ results from the verbalization of the Portuguese noun cova ‘cave’.

56

Phonology

modern SCV forms such as purdon ‘pardon’and ruspondi ‘to answer’(< Port. perdão, responder) were realized as *pordon and *rospondi, remarkably close to their PA cognates pordon ‘pardon’ and rospondé ‘to answer’. 2.1.3. Vowel harmony Though vowel harmony is not productive in either modern-day PA or Upper Guinea PC, it may once have been so, as suggested by the shared lexemes in Table 5 (cf. Martinus 1996: 153; Quint 2000b: 130). Table 5. Vowel harmony ETYMON

PA

Upper Guinea PC

GLOSS

Sp./Port. lamber 100 Sp./Port. negocio/negócio Sp./Port. molestar Sp./Port. levantar(se) Sp./Port. destino Sp./Port. tribunal Sp./Port. crucificar

lembe

lembe

‘to lick’

nogoshi

‘business’

molostiá lanta∼lamantá

SCV nogosi/ GBC nogos GBC molosta GBC lanta∼labanta101

distinu trubunal (Early PA) krusufiká (Early PA)

distinu SCV trubunal SCV krusufika

‘destiny’ ‘tribunal’ ‘to crucify’

‘to disturb’ ‘to stand up’

Observations: – The harmonized forms PA trubunal and krusufiká are found in Early PA texts from 1803 (in Martinus 1996: 33–37) and 1844 (Conradi’s Gospel of Matthew) respectively, but have not survived into modern PA, clearly suggesting vowel harmony was once much more widespread in PA. Modern PA dictionaries give tribunal and krusifiká. – Another possibly interesting case of vowel harmony is PA fròkofròko ‘unstable’, at least if, in line with Martinus (1996: 149), we analyze it as a derivation of Port. fraco ‘weak’.

100 Standard European Spanish has lamer, but lamber is found in several non-standard varieties of Spanish (Penny 2000: 87, 188, 232), so that a Spanish etymology cannot be discarded here. 101 SCV only has the longer form labanta.

2.1. Vowel features

2.1.4.

57

Monophthongs

PA and Upper Guinea PC show a strong preference for monophthongs, which they have acquired either through preservation of the original Portuguese monophthong or through processes of monophthongization of original diphthongs, as the following overview will substantiate. 2.1.4.1. Retention of Portuguese monophthongs Table 6 demonstrates the retention of Portuguese monophthongs in PA. Table 6. Retention of Portuguese monophthongs ETYMON

PA

Upper Guinea PC

GLOSS

Port. tempo (= Sp. tiempo) Port. preto (= Sp. prieto) Port. quem (= Sp. quien)

tempu pretu ken

‘time’ ‘black’ ‘who’

Port. terra (= Sp. tierra) Port. qualquer (= Sp. cualquier) Port. sempre (= Sp. siempre) Port. ferro (= Sp. hierro) Port. festa (= Sp. fiesta) Port. valente (= Sp. valiente) Port. -mentu (= Sp. -miento) Port. porco (= Sp. puerco) Port. porta (= Sp. puerta) Port. forte (= Sp. fuerte) Port. novo (= Sp. nuevo) Port. força (= Sp. fuerza) Port. bom (= Sp. bueno, bien) Port. fora (= Sp. fuera)

tera kalke semper heru festa balente -mentu porko porta forti nobo forsa bon for (di)

tempu pretu SCV ken/ GBC kin tera kalker sempri feru festa balenti -mentu porku porta forti nobu forsa bon fora (di)

Port. logo (= Sp. luego)

lo

logu

Port. sonho (= Sp. sueño)

soño

sonu/sonhu

‘earth’ ‘any’ ‘always’ ‘iron’ ‘party’ ‘courageous’ ‘-ment’ ‘pig’ ‘door’ ‘fort’ ‘new’ ‘strength’ ‘good, well’ ‘out(side) (of), from’ FUT (PA); ‘soon, later’ (Upper Guinea PC) ‘sleep, dream’

58

Phonology

Table 6. (continued) ETYMON

PA

Upper Guinea PC

GLOSS

Port. dono (= Sp. dueño) Port. morto (= Sp. muerto) Port. torto (= Sp. tuerto) Port. corpo (= Sp. cuerpo) Port. arriscar (= Sp. arriesgar)

doño102 morto torto kurpa riska

donu mortu tortu GBC kurpu riska

‘owner’ ‘dead’ ‘cross-eyed’ ‘body’ ‘risk’

Observations: – In PLQ, by contrast, the Spanish diphthong has been retained even in the fundamental part of the lexicon: PLQ tiembo ‘time’, kiene ‘who’, tiela ‘earth’, prieto ‘black’, siempre ‘always’, tambié ‘also’, kueppo ‘body’, puetta ‘door’ < Sp. tiempo, quién, tierra, siempre, también, cuerpo, puerta (de Friedemann & Patiño Rosselli 1983: 97, 168; Dieck 2002: 150; Schwegler 2007: 211). 2.1.4.2. Monophthongization The preference for monophthongs in PA and Upper Guinea PC is further illustrated in Tables 7–13, all showing shared cases of monophthongization (cf. Martinus 1996: 148, 155–158; Quint 2000b: 134). Table 7. /aj, ei/ > /e/∼/i/ ETYMON

PA

Upper Guinea PC

GLOSS

Port. tambem (= Sp. también) Port. tem (= Sp. tiene) Port. deixar (= Sp. dejar)

tambe tin desha103

tambe ten SCV dexa/GBC disa

‘also’ ‘to have, exist’ ‘to leave, let’

102 PA doño appears to result from the conflation of a Portuguese (dono) and Spanish (dueño) etymon. Van Wijk (1958: 177) and Kramer (2004: 99) list several other PA lexemes that can be analyzed in a similar manner, i.e. as the conflation of a Portuguese and Spanish etymon. 103 PA desha has lost its original Portuguese meaning (Quint 2000b: 191). Its modernday use is restricted to the meaning ‘to discourage’ and to the idiomatic expression kome te desha eat-until-leave ‘to stuff oneself’ (van Putte & van Putte-de Windt 2005a: 93). In the sense of ‘to leave, let’, PA uses deha < Sp. dejar and laga < Sp./Port. largar.

2.1. Vowel features

59

Table 8. /je/ > /e/ ETYMON

PA

Upper Guinea PC

GLOSS

Sp./Port. quieto Sp./Port. consciencia/consciência

ketu konsenshi

ketu SCV konsensya

‘quiet’ ‘conscience’

ETYMON

PA

Upper Guinea PC

GLOSS

Port. qualquer (= Sp. qualquier) Sp./Port. lengua/língua

kalke

kalke

‘any’

lenga

linga

‘language’

Etymon

PA

Upper Guinea PC

Gloss

Port. prisão (= Sp. prisión) Port. ração (= Sp. ración)

prison (dja)rason104

prison rason

‘prison’ ‘ration’

Etymon

PA

Upper Guinea PC

Gloss

Port. vassoura (= Sp. escoba) Sp./Port. aumentar

basora

basora

‘broom’

omentá

omenta

Sp./Port. Europa

Oropa

Oropa

‘to increase, raise’ ‘Europe’

Table 9. /ua/ > /a/

Table 10. /ão/ > /on/

Table 11. /ou, au, eu/ > /o/

104 In present-day PA, a derivative of Port. ração is found only in the composition PA djarason ‘Wednesday’. Birmingham (1970: 131f.n.) comments on the origin of this word: “Legend says that Wednesday was a ‘day of ration’for the [Curaçaoan] slaves, the one day of the week on which they were given their allotted supplies of food and clothing” (cf. Fouse 2007: 61).

60

Phonology

Table 12. /ai/ > /a/ Etymon

PA

Upper Guinea PC

Gloss

Port. baixar (= Sp. bajar) Port. demais (= Sp. demasiado) Sp./Port. bailar

basha 105 dimas

SCV baxa/ GBC basa dimas

‘to go down/ bring down’ ‘too much’

balia

balia 106

‘to dance’

Etymon

PA

Upper Guinea PC

Gloss

Sp./Port. negocio/negócio Sp./Port. remedio/remédio Sp./Port. premio/prémio Sp./Port. principio/princípio

nogoshi

SCV nogosi

‘business’

remedi 107

SCV ramedi/ GBC remedi SCV premi

‘remedy’ ‘prize’

SCV prinsipi

‘beginning’

Table 13. -/iV/ > -/i/

premi prinsipi (Early PA)

Observations: – A historically revealing case of monophthongization is found in the PA item fitó ‘plantation overseer’< Port. feitor (= Sp. factor/hacedor). Maduro (1965: 17) claimed that PA fitó derives from Spanish veedor ‘supplier’, thereby overlooking the much more plausible Portuguese etymology. In fact, the word PA fitó is reminiscent of colonial times when the Portuguese word feitor ‘factor’ was commonly used to refer to owners of trading factories (feitorias) along the Upper Guinea Coast (cf. e.g. Brooks 2003: 140). Interestingly, Portuguese feitor was also integrated in Wolof as fitor (Boulègue 2006: 47), but I did not find a cognate of it in Upper Guinea PC.

105 PA basha means ‘to go/bring down (e.g. a building)’ only when used in combination with the adverb abou ‘down’. When it occurs alone, PA basha typically means ‘to pour’. Philippe Maurer (p.c.) comments that when used in the latter sense, the etymon of PA basha could also be Sp. vaciar / Port. esvaziar ‘to empty’. 106 The monophthongization here is achieved by means of metathesis of the /l/. 107 Convergence with Dutch remedie /remedi/ is of course possible.

2.1. Vowel features

61

– Consider by contrast the diphthong in PLQ kieto < Sp. quieto ‘quiet’ (de Friedemann & Patiño Rosselli 1983: 140), where PA and Upper Guinea PC have ketu (Table 8). – PA kalke and Upper Guinea PC kalker (Table 9) alternate with the diphthongized forms PA kualke and Upper Guinea PC kualker. PA still has a third allomorph with two diphthongs, PA kualkier, derived from Sp. qualquier. It is revealing to note that Spanish-derived PA kualkier bears regular stress on the last syllable (as does the etymon), whereas the Portuguese-derived variants PA kalke∼kualke have undergone a stress shift from the last to the first syllable (van Putte & van Putte-de Windt 2005a). Such a stress shift is typical of a set of lexical items belonging to the fundamental part of the PA lexicon (cf. §2.4). In other words, the stress shift on PA kalke∼kualke suggests that these two Portuguese-based variants belong to the oldest layer of the PA vocabulary and that Spanish-derived PA kualkier was acquired more recently. – The monophthong in PA lenga and Upper Guinea PC linga (Table 9) gains relevance in contrast with PLQ lengua and ST lungwa. Note that Lengua and Lungwa are also the popular glossonyms of PLQ and ST (Bartens 2002: 16, 17), which testifies to the originality of /ua/ diphthongs in these two creoles. – To be sure, PA dimas ‘too much’ (cf. Table 12) should be attributed to Port. demais ‘too much’ rather than to Spanish demás, which means something different (‘other, rest’) and is used differently. PA dimas resembles Portuguese demais in that it follows the word it modifies: e.g. PA dushi dimas ‘too sweet’ = Port. doce demais (= Sp. demasiado dulce) (Birmingham 1970: 132f.n.; cf. van Putte & van Putte-de Windt 2002: 71). Unsurprisingly, Upper Guinea PC dimas is used in the exact same way (Lang 2002: 118, 119; Scantamburlo 2002: 112). – The reduction of word-final -/iV/ to /i/ (Table 13) also characterizes words such as PA sirbishi ‘service’ (< Sp. servicio). In some cases, an old and new form coexist, e.g. PA fishi∼ofisio ‘job, occupation’ (< Sp./Port. ofício/oficio). The feature must once have been more widespread in PA, as suggested by the Early PA forms prinsipi ‘beginning’ and hustishi ‘justice’ (< Sp./Port. principio/princípio and Sp. justicia) found in Conradi (1944) and Eybers (1916: 494) respectively. Modern-day PA has restored the etymological diphthong: PA prinsipio and hustisia. (Cf. Quint 2000a: 116 for additional examples of this feature in SCV.)

62

Phonology

2.2. Consonant features A selection of consonant features and changes which suggest a link between PA and Upper Guinea PC will be discussed below. I will first address the distribution of the voiceless palatal fricative /S/ and the implications for the Spanish vs. Portuguese origins of PA, then highlight the preservation of the Old Portuguese phonemic opposition between /S/ and /tS/. Subsequently, the shared PA / Upper Guinea PC tendency to avoid voiced fricatives is highlighted. Finally, I will briefly focus on two ways in which PA and Upper Guinea PC differ from PLQ. 2.2.1. The voiceless palatal fricative /S/ in PA and Upper Guinea PC The presence of the voiceless palatal fricative /S/ in the fundamental parts of the vocabularies of Portuguese, Upper Guinea PC and PA versus its absence in Spanish is reason to discuss this phoneme at length below. As will become clear, the phoneme indeed has far-reaching consequences for the debate on the origins of PA, as it constitutes evidence against the hypothesis (addressed in §1.3) that Old (16th-century) Spanish participated in the formation of PA and furthermore provides an exclusive link to SCV. 2.2.1.1. Originality of the /S/ in PA For PA, Kouwenberg & Murray (1994: 11) treat /S/ as an allophone of /s/, but minimal pairs such as the ones provided in Table 14 show that this is untenable: Table 14. Minimal pairs of /s/ vs. /S/ in PA kasi ‘almost’ masa ‘mass’ kurason ‘heart’ pasa ‘to pass’ so ‘only’ kaso ‘case’ sopa ‘soup’ desa ‘insect’

kashi /kaSi/ ‘closet’ masha /maSa/ ‘much’ kurashon /kuraSon/ ‘curación’ pasha /paSa/ ‘to spend the night’ sho /So/ ‘show’ kasho /kaSo/ ‘anacardo’ shopa /Sopa/ ‘dovecot’ desha /deSa/ ‘discourage’

The originality of the /S/ in PA is suggested not only by its phonemic status, but also by the fact that it occurs in a set of lexemes whose etyma lack the phoneme, such as PA shusha /SuSa/ < Sp./Port. suciar/sujar ‘to make dirty’. Moreover, the /S/ appears in essential parts of the lexicon: e.g. PA mishi /miSi/ < Port. mexer ‘to agitate, stirr’. A further indication is that the phoneme occurs before back and front vowels, open and closed vowels (e.g. PA shushi /SuSi/ ‘dirty’ shen /Sen/

2.2. Consonant features

63

‘hundred’, shon /Son/ ‘mister, master’, desha /deSa/ ‘to discourage’. It appears, then, that the occurrence of the /S/ is in no way restricted by the quality of the following vowel. Aruban PA provides yet another clue for the originality of the phoneme /S/. As was noted previously, Aruban PA generally shows a stronger inclination towards Spanish phonology, which can be seen, for instance, in the restauration of the world-final unstressed /o/ where Curaçaoan PA has /u/ (cf. §2.1). Since Spanish lacks the /S/ all together, one would expect the phoneme to be less strong on Aruba. However, several allomorphic pairs, listed in Table 15, suggest otherwise, showing an Aruban palatalized form versus an unpalatalized Curaçaoan variant. Table 15. Aruban PA /S/ vs. Curaçaoan PA /s/ or /dZ/ Aruban PA

Curaçaoan PA

Gloss

kashon /kaSon/ fosha /foSa/ mangashina /mangaSina/ fishi /fiSi/ bashaka /baSaka/

karson forsa mangasina ofisio (also fishi) badjaga

‘pants, trousers’ ‘strength’ ‘store’ ‘job, occupation’ ‘type of ant’

A final indication of the originality of the /S/ in PA is that it was preserved in all those PA words whose – indeed, Portuguese – etymon also contains the phoneme (more regarding this below). 2.2.1.2. /S/ as evidence against 16th-century Spanish in PA’s superstrate The previous paragraph provided several arguments the sum of which firmly suggests that the voiceless palatal fricative /S/ belonged to the original phonemic inventory of PA. This has important theoretical implications in light of the fact that modern Spanish lacks this phoneme. The phoneme’s presence in PA’s original phonemic inventory can thus plausibly be taken to suggest that Portuguese, not Spanish, was directly involved in the creole’s formation. As noted in §1.3, however, scholars such as Rona (1971), Ferrol (1982) and Munteanu (1996a), i.e. supporters of PA’s Spanish origins, have stressed the fact that the /S/ existed in Old Spanish at least until the early 17th century. Therefore, a word such as PA pusha /puSa/ ‘to push’, according to e.g. Munteanu (1996a: 97), may not derive from Portuguese but from Old Spanish puxar /puSar/ instead.108 Ferrol (1982: 33) summarizes: “The existence of /S/ in PA, considered by the 108 Quint (2000b: 186) pointed out that, for obvious semantic reasons, PA pusha ‘to push’ is unlikely to derive from Port. puxar ‘to pull’ in the first place, English to push being a more plausible etymon. Indeed, English influence on PA is well attested

64

Phonology

pro-Portuguese school a basic piece of evidence of PA’s Portuguese origins, thus loses much of its evidential value”109 . Quint (2000b: 184, 185), however, convincingly countered this line of thought. Although the phoneme /S/ was certainly still present in Spanish up until the early 17th century, as in Old Sp. muxer /muSer/ ‘woman’ or biaxe /biaSe/ ‘journey, voyage’, the phoneme is absent from the corresponding PA lexical items muhé ‘woman’ and biaha ‘journey, voyage, time’. Evidently, if Old Spanish had been the lexifier of PA, one would expect to find forms such as PA *mushé /muSé/ or *biasha /biaSa/ instead of muhé and biaha. This line of argumentation is crucially supported by cognates found in Chabacano, a creole that is actually based on 16th-century Spanish (Holm 2001: 71), and in which the archaic Spanish phoneme /S/ was indeed preserved. Thus we find, for instance, Chabacano forms such as /muSel/ ‘woman’, /vieSo/ ‘old’ and /oreSa/ ‘ear’ (Hancock 1975: 224–227) where PA has bieu, muhé and orea. The conclusion of the foregoing is of course that the phoneme /S/ in PA was inherited neither from Old nor from modern-day Spanish, but from (Old) Portuguese. 2.2.1.3. /S/ in Upper Guinea PC and comparison with PA A noticeable difference between SCV and GBC is that the palatal fricative /S/ is part of the fundamental phonemic inventory of the former but is absent in the latter. There is a plausible explanation for this: unlike SCV, GBC has been in uninterrupted contact with several West-Atlantic languages such as Balanta and Mande languages such as Mandinka, which typically lack the palatal fricatives /S/ and /Z/ (Parkvall 2000: 44). In discussing phonological features shared by GBC and Mandinka, Rougé (2006) stresses the contrast with SCV: A clear distinction exists between Mandinka and GBC on the one hand and SCV on the other: Mandinka and the continental creole, like the majority of the region’s languages, do not have the voiced fricatives or voiceless palatal fricatives,

and dates back to the second period of English occupation of Curaçao from 1807 to 1816 (Maurer 1998: 194f.n.). English left other traces in PA’s core vocabulary, e.g. PA bèk < Eng. back (adv.), used in verbal expressions of the type bai bèk ‘to go back, to return’, or pleonastic bolbe bèk (lit. ‘to return back’) ‘to return’. Wood (1971) deals with English borrowings in PA. 109 Original quote: “La existencia de /S/ en papiamento, considerado por los portuguesistas como una de las pruebas básicas del origen portugués del papiamento pierde pues mucho de su valor probatorio”

2.2. Consonant features

65

present in Cape Verdean (as well as in Portuguese)110 (Rougé 2006: 65; cf. Quint 2000a: 69)

In other words, rather than suggesting a different genesis for GBC and SCV, the absence of the palatal fricative /S/ in GBC can be explained plausibly as the result of ongoing contact with an African adstrate. Given the absence of the phoneme in GBC, one might be tempted to assume that the phoneme was absent in the shared ancestor creole and that it was acquired later in SCV due to pressure from Portuguese. However, for several reasons it is more plausible to assume that the /S/ was present in SCV from the very beginning. Just as in PA: – /S/ in SCV occurs before all vowels, /i/, /e/, /a/, /o/ and /u/; – the phoneme is found in the essential part of the lexicon, e.g. SCV igrexa /igreSa/ ‘church’; – /S/ is found in a set of lexemes whose etymon lack the phoneme, e.g. SCV xinti /Sinti/ < Port. sentir; – the etymological /S/ is always retained (Quint 2000a: 112), i.e. whenever a Portuguese etymon has the phoneme, SCV will have it as well (see Table 16). Table 16. Retention of the Portuguese /S/111 ETYMON

PA

SCV

GLOSS

Port. baixar (= Sp. bajar) Port. mexer (= Sp. mover, mezclar, agitar) Port. deixar (= Sp. dejar) Port. peixe (= Sp. pez) Port. lagartixa (= Sp. lagartija)

basha /baSa/ mishi /miSi/

baxa /baSa/ mexi /meSi/

‘to go/come down’ ‘to stir’

desha/deSa/ *pishi /piSi/112 lagadishi /lagadiSi/113

dexa /deSa/ pexi /peSi/ lagartisha /lagartiSa/

‘to leave, let’ ‘fish’ ‘lizard’

110 Original quote: “Il existe une ligne de partage trés nette entre le mandinka et le créole de Guinée et de Casamance d’une part et le créole capverdien de l’autre. Alors que le mandinka et le créole continental, comme la plupart des langues de la región, ne connaissent pas (…) de fricatives sonores ainsi que la fricative palatal sourde, présentes en capverdien (comme en portugais)” 111 Recall that the orthographic realization of the phoneme /S/ is in PA and in CV and Portuguese. 112 This Early PA form can be hypothesized on the basis of the modern PA compound noun pishiporko /piSiporko/ < Port. peixe-porco ‘pig fish’. Modern PA has piská (< Sp./Port. pescado) for both animate and inanimate fish (Quint 2000: 317). 113 Convergence with Du. hagedisje ‘little lizard’ is possible.

66

Phonology

In other words, the synchronic distribution of /S/ in both PA and SCV suggests that this phoneme was part of the original phonemic inventories of both creoles. Table 16 shows the above-mentioned fact that whenever a Portuguese etymon has the phoneme, it is retained in PA and SCV. Also noteworthy are the cases of palatalization of the /s/ before /u/ in both PA and SCV exhibited in Table 17. Table 17. Palatalization of /s/ > /S/ before /u/ ETYMON

PA

SCV

GLOSS

Sp./Port. suciar/sujar Sp./Port. sucio/sujo Sp./Port. suciedad/sujidade

shusha /SuSa/

xuxa/SuSa/

‘to (make) dirty’

shushi /SuSi/ shushidat /SuSidat/

xuxu/SuSu/ xuxidadi /SuSidadi/

‘dirty’ ‘dirt, dirtiness’

Although both creoles have preserved the /S/ before /u/ whenever the etymon has it (e.g. SCV baxu /baSu/ < Port. baixo ‘low’; PA shut /Sut/< Eng. shoot ), the change from /s/ > /S/ before /u/ demonstrated in Table 17 is quite exceptional. In PA, I encountered no other examples of this change; in SCV only one other.114 Note that the /S/ in PA shushi and SCV xuxu is retained in the derived nouns PA shushidat /SCV xuxidadi, which strengthens the idea of a historical link between the forms in Table 17. Moreover, there is evidence from 19th-century PA texts that the suffix -dat used to be -dadi (see §4.1.4), so that we may hypothesize that Early PA once had *shushidadi, homophonous with SCV xuxidadi. Table 18 displays two rare, shared cases of the palatalization of /s/ > /S/ before /e/. Table 18. Palatalization of /s/ > /S/ before /e/ ETYMON

PA

SCV

GLOSS

Sp./Port. becerro/bezerro Sp./Port. dulce/doce

bishé /biSé/ (Aruba) dushi /duSi/

bixeru/biSeru/ doxi /doSi/

‘calf’115 ‘sweet’

114 The other example in SCV is xubinhu < Port. sobrinho ‘nephew’ (Rougé 2004a: 264). 115 The loss of etymological /ro∼ru/ in unstressed word-final position is completely regular in modern PA (cf. PA sapaté < Sp. zapatero/Port. sapateiro ‘shoe maker’, etc.). This fact allows us to hypothesize the Early PA form *bisheru, homophonous to SCV bixeru. This correspondence is meaningful also from a strictly lexical per-

2.2. Consonant features

67

Another change that has occasioned the presence of /S/ in PA and SCV is the devoicing of the original Portuguese /Z/116 , a feature shown in Table 19. In SCV, the voiced fricative /Z/ does exist, but typically in recent loan words and in the urban areas where Portuguese has more influence. When a voiced and devoiced variant of a word coexist, the latter is most likely to be heard in the interior of Santiago, while the former will likely be found in the urban areas. A case in point is the incorporation of Port. queijo as kexu /keSu/ in rural SCV and keju /keZu/ in urban SCV (Mendes et al. 2002: 339). Specialists such as Quint (2001a: 272, 273) therefore claim that the phoneme was not an original part of the phonemic inventory. Similar observations apply to PA, where nonnative borrowings from Portuguese in the Sephardic Jewish religious-cultural domain have typically retained the voiced fricative /Z/: PA zjanta /Zanta/ ‘meal’ (< Port. jantar); PA zjuzjum /ZuZum/ ‘fasting’ (< Port. jejum), etc. (cf. Henriquez 1988, 1991). Table 19. Devoicing of /Z/ > /S/ ETYMON

PA

SCV

GLOSS

Port. couragem (= Sp. coraje) Port. botija (= Sp. botija)

kurashi /kuraSi/ butishi /butiSi/

korashi /koraSi/ butisha /butiSa/

‘courage’ ‘jar’

2.2.1.4. Negative evidence from Palenquero and Gulf of Guinea PC The distribution of /S/ in PA and SCV stands in strong contrast to its distribition in Gulf of Guinea PC and PLQ. In PLQ, as to be expected of a Spanish-based creole, the phoneme /S/ does not exist (de Friedemann & Patiño Rosselli 1983: 92, 93). In Gulf of Guinea PC, below exemplified by ST117 , the phoneme /S/ exists, but its distribution is quite different from that in SCV and PA. For instance, the etymological /s/ palatalizes systematically before etymologically unstressed /e, i/. In PA and SCV, it does not. One of many examples: Port. servir ‘to serve’

> ST /Stlivi/∼/Slivi/

=

PA and SCV /sirbi/

spective: modern mainstream Portuguese and Spanish generally use ternero/terneiro to denote ‘calf’. 116 Note that the orthographic realization of the phoneme /Z/ in Portuguese is before /e/ and /i/ and before all other vowels. 117 The data on ST are drawn from Ferraz (1979), Rougé (2004a), Major (2006) and Fontes (2007).

68

Phonology

Before /a, o, u/, on the other hand, the /s/ never palatalizes in ST – it did before /u/ in a few cases shared between PA and SCV (Table 17) – and, more importantly, the etymological /S/ and /Z/ depalatalize as a rule before those same three vowels, whereas they never depalatalize in PA and SCV (Tables 16 & 19). Compare, for instance: Port. deixar ‘to leave, let’

> ST /desa/

=

PA and SCV /deSa/

Port. sujo ‘dirty’

> ST /suzu/

=

PA /SuSi/ ∼ SCV /SuSu/

Moreover, as the example of ST /suzu/ shows, the etymological /Z/ always remains voiced, whereas it is typically devoiced in PA and SCV (and GBC, for that matter). Thus, we find: Port. coragem ‘courage’ 2.2.2.

> ST /kolaZi/

=

PA /kuraSi/ ∼ SCV /koraSi/

Retention of Old Portuguese voiceless affricate /tS/ in PA and Upper Guinea PC

As is well-known, Old Portuguese still distinguished phonemically between /tS/ (written ) and /S/ (written ). According to Teyssier (1983: 53), this distinction disappeared from mainstream Portuguese in the course of the 17th century, the two phonemes merging into one: /S/. It has also been noted that this Old Portuguese phonemic distinction has been retained in SCV and GBC, which, as Rougé (1994: 140) points out, is evidence that “these creoles came into existence in an epoch in which Portuguese still distinguished phonologically between and ”118 (cf. Kihm 1994: 5; Quint 2000a: 112–114). Less well-known is that the distinction has been preserved also in PA (Quint 2000b: 133). While we have already seen that the /S/ survived in PA words of Portuguese origin (cf. Table 16), at least three PA lexemes reveal that the Old Portuguese /tS/ (written ), and thus its phonemic contrast with /S/, was preserved as well. These are displayed in Table 20. By means of contrast, this distinction was not preserved in Gulf of Guinea PC (exemplified below with ST). In fact, all the (de-)palatalization rules that apply to the integration of the etymological /S/ in Gulf of Guinea PC (as discussed in §2.2.1.4) also apply to the integration of the Old Portuguese /tS/. That means that etymological /tS/ is systematically depalatalized before /a, o, u/: 118 Original quote:“ces créoles sont nés à une époque où le portugais distinguait encore phonologiquement et ”

2.2. Consonant features

69

Table 20. Retention of the Old Portuguese voiceless affricate /tS/ in PA and Upper Guinea PC ETYMON

PA

SCV

GLOSS

Old Port. chumbo (= Sp. plomo) Old Port. murchar (= Sp. marchitar) Old Port. rachar (= Sp. rajar) Old Port. cachorro119

chumbu /tSumbu/

txumbu /tSumbu/

‘lead’

morcha /mortSa/

murtxa /murtSa/

‘to wither’

racha /ratSa/

ratxa /ratSa/

kachó /katSo:/

katxor /katSor/

‘to crack, tear’ ‘dog’

Old Port. cachorro ‘dog’

> ST kaso ‘dog’

=

PA and SCV /katSo:/ ∼ /katSor/

Old Port. chumbo ‘lead’

> ST sunbu ‘lead’

=

PA and SCV /tSumbu/

Note also that (just like the /s/) the etymological /t/ in Gulf of Guinea PC palatalizes without exception before etymologically unstressed /e, i/, whereas it normally does not in Upper Guinea PC and PA: Port. forte ‘fort’ 2.2.3.

> ST /fortSi/

=

PA / Upper Guinea PC /forti/

Rejection of voiced fricatives in PA and Upper Guinea PC

The devoicing of the etymological voiced palatal fricative /Z/ to /S/ in PA and SCV mentioned in §2.2.1.3 (Table 19) is revealing of a general tendency to avoid voiced fricatives in PA, SCV and GBC. In addition to the devoicing of /Z/, they systematically devoiced etymological /z/ to /s/ and plosivized /v/ to /b/, at least in the fundamental, most conservative parts of their vocabulary (cf. Parkvall 2000: 32; Quint 2000a: 113 & 2001b: 74), which is what concerns us here. The lack of voiced fricatives is of course typical of Spanish; however, as shown below, their rejection in PA is clearly visible in the Portuguese-derived part of the PA lexicon. 119 Note that Portuguese cachorro means ‘puppy’, not ‘dog’. Though Spanish cachorro ‘puppy’cannot be discarded as an etymon for PA kachó, the (Old) Portuguese etymon is more likely in light of the fact that a cognate of Port. cachorro with the meaning of ‘dog’ is found also in Brazilian Portuguese, Upper Guinea PC and Gulf of Guinea PC, whereas I know of no variety of Spanish (be it creolized or not) where cachorro means ‘dog’. Spanish-based PLQ, for instance, has pelo ‘dog’ < Sp. perro ‘dog’, rather than a cognate of Sp. cachorro (de Friedemann & Patiño Rosselli 1983: 97).

70

Phonology

2.2.3.1. /z/ > /s/ While some Upper Guinea PC lexemes do show the voiced palatal fricative /z/ of the Portuguese etyma, these generally do not belong to the fundamental part of the lexicon. Core vocabulary items typically show devoicing, as in Upper Guinea PC fasi ‘to make, do’, bisinhu ‘neighbor’ < Port. fazer, vezinho. In cases where a lexeme has both a voiced and a devoiced allomorph, the voiced variant will be typical of the meso- and acrolect. For instance, Port. trazer ‘to bring’ has given traze in urban SCV, but tarsi in the basilectal rural variety (Mendes et al. 2002), clearly suggesting that the devoicing of etymological /z/ is an original Upper Guinea PC feature (cf. Quint 2001a: 272, 273). Examples of the devoicing of etymological /z/ are harder to find in PA due to the lack of this phoneme in Spanish. Still, however, some PA lexemes testifying to the originality of this feature can be identified in the fundamental part of the vocabulary: PA trese ‘to bring’ PA bisiña ‘neighbor’ PA santu ‘sand’

< <
/b/ The plosivization of the etymological /v/ to /b/ (cf. Martinus 1996: 159) is observable in a series of Portuguese-derived lexemes displayed in Table 21. Table 21. /v/ > /b/ ETYMON

PA

Upper Guinea PC

GLOSS

Port. vizinho (= Sp. vecino) Port. vinho (= Sp. vino) Port. novo (= Sp. nuevo) Port. vai (= Sp. va)

bisiña biña nobo bai

bisinhu binhu nobu bai

‘neighbor’ ‘wine’ ‘new’ ‘to go’

Although the /v/ is increasingly used on the Cape Verde Islands, Quint (2000a: 132) shows that the basilectal varieties systematically plosivize the etymological /v/ and notes that this is also the case in the early work of Brito (1887).

120 Though the Dutch etymon is the more likely option, English sand cannot be discarded either, for reasons explained in footnote 108.

2.2. Consonant features

71

2.2.3.3. Contrast with Gulf of Guinea PC The rejection of voiced fricatives in PA and Upper Guinea PC contrasts significantly with Gulf of Guinea PC. In all varieties of Gulf of Guinea PC, the voiced fricatives /Z/ and /z/ appear in phonemic contrast with their voiceless counterparts. These contrasts, according to Ferraz (1979: 21), constitute “some of the more important contrasts in the consonantal system” of ST. The following minimal pairs give an idea: ST kasa ‘to hunt’

vs.

kaza ‘to marry’

ST maSi ‘more’

vs.

maZi ‘but’

(Ferraz 1979: 21, 22; cf. Maurer 2009a: 10 for similar minimal pairs in PRI) Furthermore, unlike in PA and Upper Guinea PC, the voiced labiodental fricative /v/ is a fully productive phoneme in Gulf of Guinea PC, as shown by minimal pairs in the essential part of the vocabulary such as the following: ST bé ‘to go’

vs.

vé ‘to see’ (Ferraz 1979: 21)

PRI gá ‘grain’

vs.

vá ‘boar’ (Maurer 2009a: 10)

PRI fwà ‘to perforate’

vs.

vwà ‘to fly’ (Maurer 2009a: 10)

Consider also the retention of the etymological /v/ in the following Gulf of Guinea PC lexemes and compare these with the PA / Upper Guinea PC cognates listed above in Table 21: Port. vizinho

> ST vidjan∼vijan, PRI vijin, ANG vizyan

Port. vinho

> ST vin, PRI ivin, ANG vi

To summarize, the most conservative layers of the vocabularies of PA, SCV and GBC have systematically rejected the Portuguese voiced fricatives /z, Z, v/, be it through devoicing (/z, Z/ > /s, S/) or by making them plosive (/v/ > /b/) (cf. Grant 2008a: 53 for PA and Quint 2001a: 272, 273 for SCV), whereas in Gulf of Guinea PC, each of these three fricatives (still) has full phonemic status. The difference may very well reflect the fact that Upper Guinea PC and Gulf of Guinea PC have strongly distinct substrates. As noted, Gulf of Guinea PC’s substrate is characterized by Kwa and Bantu, whereas the substrate languages of Upper Guinea PC are known to be Mande and West-Atlantic. Mandinka (Mande) and Wolof (West-Atlantic), the two main substrate sources for Upper Guinea PC, for instance, do not have voiced fricatives (Quint 2000a: 113). If the lack of voiced fricatives in the conservative part of PA’s vocabulary traces back to that same Mande/West-Atlantic substrate, it is interesting to note that in

72

Phonology

the literature on PA, these Upper Guinean languages are traditionally ignored as possible contributors to PA’s substrate (but see Intumbo 2006). Instead, it is rather uncritically assumed that its substrate consisted of Kwa and Bantu languages (e.g. Lefebvre 2011: 128; see further discussion in §7.1). 2.2.4. The lack of lambdacism (/r/ > /l/) in PA and Upper Guinea PC In both PA and Upper Guinea PC, /l/ and /r/ are in phonemic contrast (Kouwenberg & Murray 1994: 8; Quint 2000a: 27). Accordingly, the cross-linguistically common process of lambdacism is attested only sporadically (cf. Bartens 1996: 252 for PA; Quint 2000a: 121 for SCV). The lack of lambdacism may seem trivial, but it gains relevance in light of the fact that lambdacism is in fact quite typical of Bozal Spanish and PLQ, varieties to which PA has been claimed to be related on more than one occasion: “Some of the most highlighted characteristics attributed to bozal Spanish, such as the change of /r/ to bozal /l/, are rarely if ever found in Papiamentu, although such a change is widespread in Palenquero” (Grant 2008: 51f.n.).121 2.2.5.

Rhotacism (/d/ > /r/)

In addition to lambdacism, the etymological word-initial /d/ has typically been incorporated as /r/ in PLQ, as in Sp. Dios ‘God’ > PLQ Rioso, Sp. derecha > PLQ rerecha (de Friedemann & Patiño Rosselli 1983: 95; Schwegler 1998: 267). There is no trace of this word-initial rhotacism in PA and Upper Guinea PC. What we do find in PA is that the intervocalic /d/ of final unstressed syllables changed to /r/. Interestingly, this change is visible in two Portuguese-derived lexemes: Port. pode > PA por ‘to can, be able’, and Port. tudo > PA tur ‘all’. This may not be a coincidence: the development suggests that the original intervocalic /d/ of the final syllable was pronounced as a plosive, which is indeed typical of mainstream Portuguese (e.g. tudo /tud’/. In colloquial varieties of Spanish, on the other hand, this /d/ is typically pronounced as a weak approximant (e.g. todo /to:o/). This widespread Spanish tendency is reflected in PLQ – “El Palenquero elimina por lo regular la /d/ intervocálica de la sílaba final” (de Friedemann &

121 Examples of lambdacism in PLQ include kelé, ndulo, tiela, lupé < Sp. querer ‘want’, duro ‘hard’, tierra ‘earth’, romper ‘break’, etc. (Dieck 2002: 150). Lambdacism is highly characteristic also of ST (cf. Ferraz 1979: 36; Rougé & Schang 2006). However, ANG and PRI do have a phonemic contrast between /r/ and /l/ (Maurer 1995: 32; Maurer 2009a: 10), suggesting the proto-variety of Gulf of Guinea PC also did.

2.3. Syllabic restructuring

73

Patiño Rosselli 1983: 93, 136)122 – and also accounts, for instance, for the loss of the /d/ in recent PA loans from Spanish, such as PA kuidou! ‘watch out!’ < Sp. cuidado!. The hypothesized Portuguese origin of the change of -/dV/ > -/r(V)/ in PA is supported by a parallel development in SCV. For instance, although SCV has tudu ‘all’, the combination tudu + modi ‘way’ occasions the change of -/dV/ to -/r/: SCV tirmodi ‘in all ways’(Lang 2002: 786). Moreover, in colloquial speech, SCV tudu is sometimes realized as /tur/ (indeed, homophonous with PA tur), an example of which can be found in the recordings of Baptista (2002: sound sample ‘SA’: /nu bai tur mininu/ we-go-all-child ‘we all went as children’). In addition, both in PA and SCV, a series of antepenultimately stressed etyma ending in -/dV/ have typically lost the middle vowel and changed -/dV/ to -/rV/: Sp./Port. sábado ‘saturday’

>

PA (dja)sabra and SCV sabru

Sp. hígado ‘kidney’

>

PA higra

Port. dívida ‘debt’

>

SCV dibra

Port. bêbedo ‘drunk’

>

SCV bebru (cf. Martinus 1996: 161, 162; 1999: 233)

In an attempt to account for the change of /d/ > /r/ in words such as PA sabra and hígra, Maurer (1998: 195f.n.) draws a parallel with PLQ: “The change /d/ > /r/ is very frequent in PLQ and in other Afro-Hispanic varieties, such as the speech of Chocó, Colombia”123 . It should be stressed, however, that, as noted above, the change from /d/ > /r/ in PLQ has imposed itself systematically in word-initial position, of which there is no trace in PA, or in Upper Guinea PC. Note, finally, that the change in PA of word-final -/dV/ to -/r/ played a crucial role in the development of PA’s past participle morpheme (Sp./Port. -DO > -/r/ > -/Ø/), which will be discussed in more detail in §4.2.1.

2.3. Syllabic restructuring What follows is a discussion of sound changes which in various ways have affected the syllable structure of PA and Upper Guinea PC lexemes. 122 Some examples given by de Friedemann & Patiño Roselli (1983) are PLQ oganisao, kojío, pekao and kasao < Sp. organisado, cogido, pescado and casado. 123 Original quote: “El cambio d > r es muy frecuente en Palenquero y en otras variedades afrohispanas, como en el habla del Chocó, Colombia”

74

Phonology

2.3.1. Apheresis of prefixes The apheresis of the Iberian prefixes /a/-, /e/-, /es/- and /en/- is one of the more typical strategies of reducing syllable length in PA and Upper Guinea PC (cf. Martinus 1996: 159, 169; Quint 2000b: 129). Table 22 provides examples. Table 22. Apheresis of /a/-, /e/-, /es/- and /en/- in PA and Upper Guinea PC Etymon

PA

Upper Guinea PC

Gloss

Port. apertado (= Sp. apretado) Port. afogar (= Sp. ahogar) Port. até (= Sp. hasta) Port. aquentar (= Sp. calentar) Port. arriscar (= Sp. arriesgar) Sp./Port. América Port. esquecer (= Sp. olvidar) Port. espinha (= Sp. espina) Sp./Port. escribir/escrever Port. esgravatar (= Sp. escarbar) Port. enfastiar (= Sp. fastidiar) Sp./Port. encontrar

pertá

pertadu

‘tight’

foga

foga

‘to drown’

te kenta

te kenta

‘until’ ‘to heat’

riska

riska

‘to risk’

Merka skeze

SCV Merka SCV skese

‘America’ ‘to forget’

spiña

spinha

‘spine’

skibi∼skirbi

GBC skirbi

‘to write’

grawatá

‘to dig’

fastia

SCV grabata/ GBC garbata fastia

‘to bother’

kontra 124

kontra

‘to meet’

Observations: – Kramer (2004: 137), who in his otherwise thorough historical and linguistic study of PA makes no mention of possible ties with Upper Guinea PC, analyzes the tolerance for word-initial consonant clusters of the type /sC/ in PA (as in skirbi) as the result of influence from Dutch. The data contained in Table 22 show that influence from Dutch is not needed to explain this tolerance. – The Portuguese-derived form PA foga is given by Martinus (1996: 169). The hispanized variant PA hoga appears to be more common, at least nowadays. 124 A more common verb for ‘to meet’ in informal registers of present-day PA is topa ku (< Sp. toparse con) (Philippe Maurer, p.c.).

2.3. Syllabic restructuring

75

The substitution of a Portuguese etymological /f/ by a Spanish /h/ is also observable in PA heru ‘iron’ < Port. ferro (= Sp. hierro), PA hariña ‘flour’ < Port. farinha (= Sp. harina), and PA herebé ‘to boil’ < Port. ferver (= Sp. hervir). – In addition to the lexical coincidence of PA spiña with Upper Guinea PC spinha < Port. espinha (= Sp. espina), the creoles derived a word for ‘porcupine’ from Port. porco-espinho (= Sp. puerco-espín): PA porkispiña, GBC purku-spinha, SCV porku-spinhu. – The significance of the correspondence between PA Merka and SCV Merka increases considerably in view of the following remark by Rougé (2004a: 57), regarding the incorporation of toponyms in the Portuguese-based creoles of West Africa (Upper Guinea PC and Gulf of Guinea PC): “Creoles typically borrow the toponym directly from Portuguese. However, in Cape Verde, America is a frequent topic of conversation and a typical Capeverdean term has developed: Merka”125 . Note in this respect also the homophony between PA and SCV Oropa (§2.1.5.2). – Note that the apheresis of the etymological prefix /en/- (exemplified in the bottom two rows of Table 22) is applied less rigorously in Upper Guinea PC than in PA and is often reduced to a nasal in the former. For instance, Port. engolir ‘to swallow’ (= Sp. engullir) and enfadar ‘to be bored’ (= Sp. aburrirse) have given PA guli and fada versus SCV nguli and nfada. 2.3.2. Vowel epenthesis In spite of the broad tolerance for complex consonant clusters in PA and Upper Guinea PC, some of these have been dissolved by means of vowel epenthesis (see Table 23) (cf. Martinus 1996: 148; Quint 2000b: 128). Table 23. Two shared cases of vowel epenthesis in PA and Upper Guinea PC Etymon

PA

Upper Guinea PC

Gloss

Sp./Port. oscuro/escuro Sp./Port. brotar

sukú borotá 126

sukuru SCV borota

‘dark’ ‘to sping, burst out’

125 Original quote: “Les créoles empruntent généralement tel quel le terme portugais. Cependent, au Cap–Vert (…) l’Amérique est un thème de conversation fréquent et un terme typiquement capverdien s’est développé: Merka” 126 The etymology proposed here for PA borotá is somewhat speculative, as the verb’s actual meaning is not ‘to burst’ but ‘to make noise, kick up a row’. In SCV, borota means ‘to flood, flow over’ (Lang 2002: 89).

76

Phonology

Observations: – Consider also the dissimilation of /lg/ clusters by means of vowel epenthesis in PA delegá ‘thin, skinny’ and kologá ‘to hang’ (< Sp./Port. delgado, colgar) and SCV alugudon (< Port. algodão ‘cotton’) as well as of /kr/ clusters in PA kurú ‘raw’ (< Sp. crudo) and SCV kiria ‘to raise’ (< Port. criar). – The link between PA sukú and Upper Guinea PC sukuru is discussed in detail by Quint (2001). The link is reinforced by contrasting forms in Gulf of Guinea PC which are characterized by the apheresis of the prefix /es/-: Port. escuro > ST kulu and PRI ukuru (Rougé 2004a: 148). 2.3.3.

Metathesis of the /r/

The metathesis of the etymological /r/ is responsible for the close resemblence between the PA and Upper Guinea PC lexemes listed in Table 24 (cf. Martinus 1996: 174; Quint 2000b: 132). Table 24. Metathesis of the /r/ Etymon

PA

Port. formiga (= Sp. hormiga) Port. rolar 128 Sp./Port. dormir 129 Sp./Port. estorbar/estorvar Sp./Port. probar/provar Sp./Port. madrugada Sp./Port. escribir/escrever Sp./Port. torcer

fruminga lora drumi stroba purba mardugá skirbi trose

Sp./Port. atormentar Sp./Port. tormenta Sp./Port. procurar

tromentá tromenta perkurá

Upper Guinea PC 127

SCV fruminga SCV lora SCV drumi SCV stroba purba SCV mardugáda GBC skirbi SCV troxi/ BaCV truse 130 SCV tromenta SCV tromenta purkura

Gloss ‘ant’ ‘to roll’ ‘to sleep’ ‘to obstruct’ ‘to try, taste’ ‘dawn’ ‘to write’ ‘to twist’ ‘to torment’ ‘storm’ ‘endeavor’

127 Modern dictionaries such as Lang (2002) and Mendes et al. (2002) give SCV furminga. The form fruminga is given by Lopes da Silva (1957: 284). 128 Sp. rolar exists, but only as a nautical term. The meaning of the cognate verb in PA and SCV is reflected only by Port. rolar ‘to roll’. 129 The metathesis of the /r/ has affected the incorporation of the Romance verb dormir identically in several creoles of different lexifiers: besides PA/SCV drumi, we find Papia Kristang drumí, Macanés drumí, Java Creole droomi, Ceylan and Haitian French Creole dromí, PLQ drumí and ST dlumi (Hancock 1975: 226, 227). 130 Lopes da Silva (1957: 378) is the source for the BaCV form truse.

2.3. Syllabic restructuring

77

Observations: – In SCV, variation of the type durmi∼drumi and purba∼pruba is quite common. In GBC, the constraint on complex syllable onsets of the type /Cr/ appears to be slightly stronger than in PA and SCV. For instance, PA/SCV fruminga, drumi and tromenta correspond to GBC furminga, durmi and turmenta. – While metathesis might be too common a phenomenon (cf. footnote 129) for us to draw any far-reaching conclusions from Table 24, the corresponding metathesis in PA lora, fruminga, and mardugá (of which the first two are of probable Portuguese origin) and the SCV cognates lora, fruminga and mardugada is nonetheless striking and not attested in any of the varieties of Gulf of Guinea PC let alone in PLQ. 2.3.4.

Negative evidence: syllabic restructuring in PLQ and Gulf of Guinea PC

The relevance of the processes of apheresis, epenthesis and metathesis discussed above would perhaps be limited if not for the contrastive way in which PLQ and Gulf of Guinea PC deal with etymologically complex syllable structure. For PLQ, Schwegler (1998: 265) explains that Spanish clusters of the type [liquid + consonant] have been simplified by means of regressive assimilation, which means that the liquid adopted the shape of the following consonant, resulting in long and tensely articulated double consonants: Sp. cerdo ‘pig’

>

PLQ seddo

Sp. calvo ‘bold’

>

PLQ kabbo

Sp. algo,

>

PLQ aggo

Sp. árbol

>

PLQ ábbol

Sp. arde

>

PLQ adde (Schwegler 1998: 265)

Also in Gulf of Guinea PC, the restructuring of syllables follows a pattern quite different from that observed in PA and Upper Guinea PC. Characteristic for the whole of Gulf of Guinea PC is the tendency – much stronger than in PA and Upper Guinea PC – towards /CV/ syllable structure and disyllabic words (cf. Maurer 2009a: 12–14; Rougé & Schang 2006: 25–27), which is, at least in part, compensated for by a relatively complex tonal system (Maurer 2009a: 14–27). The tolerance for complex consonant clusters is generally low in Gulf of Guinea

78

Phonology

PC but remarkably high in PA and Upper Guinea PC, even in the essential part of the vocabulary, as shown for instance by some of the entries in Tables 20, 22 and 24. Furthermore, all consonants other than /n/ are avoided in word-final position in Gulf of Guinea PC. Such a constraint does not exist in either Upper Guinea PC or PA (Holm 1988: 113). Table 25 provides some examples from all members of Gulf of Guinea PC (taken from Rougé & Schang 2006: 25 and Rougé 2004a) and the corresponding PA / SCV cognates, which should suffice to illustrate the contrast in terms of resyllabification. Table 25. Contrastive examples of resyllabification in Gulf of Guinea PC, PA and Upper Guinea PC Etymon

ST

Annobonese

PRI

ANG

PA

SCV

Gloss

Port. grande Port. coração Port. barriga

glandji

gain

gani

ngai

grandi

grandi

‘big’

kloson

kusan

kosan

koso

kurason

korason

‘heart’

bega

?

bwega

beega

barika

bariga

‘belly’

2.4. Paroxytonic verb stress in PA and SCV To close, I briefly address the intriguing issue of verb stress in PA and SCV. Note that the tonal phenomena characteristic of PA will not play a role here.131 The incorporation of Iberian infinitives into the PA lexicon is characterized by a stress shift from the final to the first syllable, as in  kanta ‘to sing’,  piska ‘to fish’ < Sp./Port. cantar, piscar, etc.132 This feature is quite rare: Martinus (1996: 153) even claims it is uniquely found in PA. Indeed, in most creoles with 131 For data on tone in PA, the interested reader is referred to Roemer (1983, 1991), Murray (1994), Devonish (1995), Kouwenberg (2004), Remijsen & van Heuven (2005), Rivera-Castillo (1994, 2009), Rivera-Castillo & Pickering (2004) and Rivera-Castillo & Faraclas (2006). Interestingly, Pires & Hutchinson (1983) have described the existence of a two-tone system in SCV very similar to that of PA (see the discussion in Bartens 1996: 30). 132 The only exception known to me is PA mesté. The oxytonic stress of mesté can be accounted for by diachronically analyzing it as the past participle of its synchronic allomorph PA meste, which does have paroxytonic stress. On the verb meste-mesté ‘to need, must, be necessary’, see also §4.3.5.2 and 5.7.1.3.

2.4. Paroxytonic verb stress in PA and SCV

79

an Iberian component, verbs tend to retain the original Iberian stress on the final syllable (cf. e.g. de Friedemann & Patiño Rosselli 1983: 91 for PLQ; Ferraz 1979: 20 for ST; Cardoso & Smith 2004: 138, 140 for Saramaccan). By the time of Martinus’ (1996) writing, though, it was not yet widely recognized (as it is now) that SCV, like PA, stresses its verbs on the penultimate syllable: SCV  kanta,  piska < Port. cantar, piscar (Quint 2000b: 142). One difference between PA and SCV is that in PA, verbs with more than two syllables (henceforth ‘longer verbs’) are oxytonic (e.g. PA pursiguí ‘to prosecute’), whereas in SCV, these are also paroxytonic (e.g. SCV pur sigi). However, evidence from Early PA texts suggests that not only disyllabic but also longer verbs were paryxitonic in Early PA, as in SCV, well into the latter stages of the 19th-century. This evidence is provided by the Dutch-based orthography used in these Early PA texts (mostly written by Dutch priests and clergymen). According to this orthography, doubly written vowels are long vowels and thus provide an indication of stress. Such double vowels occur in penultimate syllables not only of disyllabic verbs, e.g. ‘(to) sing’, but also of plurisyllabic verbs, e.g. ‘(to) submit’. Since, unlike Dutch, PA does not have a length opposition in its vowel system, the orthographic practice of the texts is entirely based on the Dutch preference for long vowels in stressed positions.133 The change of stress on longer verbs from paroxytonic to oxytonic thus appears to be a relatively recent development in PA, and the correspondence with SCV (all verbs paroxytonic) stronger diachronically than synchronically. PA and SCV furthermore overlap in that the paroxytonic verbs undergo a stress shift from the first to the last syllable when followed by a pronoun (Quint 2000b: 138). The following examples show this feature as well as the fact the phoneme /o/ of the 2sg pronoun is raised to [u] in object position: (1)

ta

 mira

a. PA

mi

b. SCV

 odja m ta I IMP see ‘I see you’

+

bo

>

mi

ta

mi  ra

bu

+

bo you

> >

m I

ta IMP

o dja see

bu you

(adapted from Quint 2000b: 138) 133 Though the Dutch orthography provides the principal clues, the Spanish-oriented orthography used by Evertsz (1898) and Sintiago (1898) points in the same direction (i.e. all verbs carrying paroxytonic stress). Evertsz used grave accents on the final syllables of all multisyllabic infinitives (e.g. ‘to eat’; ‘to know’) versus acute accents on all corresponding past participles (e.g. ‘eaten’; ‘known’). Sintiago used orthographic accents on final syllables of clear past participles, versus none on final syllables of the corresponding infinitive, as in ‘to pronounce’ vs. ‘pronounced’.

80

Phonology

2.4.1. Verb stress in GBC For GBC, Wilson (1962: 19; cf. Birmingham 1975: 67) asserted that, “[w]hile the stress in disyllabic verbs is normally on the first syllable, when an object pronoun is suffixed, this stress is shifted to the last syllable”. Indeed, this is exactly what was described above for PA and SCV. Bal (1983a), in particular reference to the conservative variety of Casamance, shares the view on verb stress in GBC expressed by Wilson: “Paroxytonic are verbs like debe, more, fala, fasi, etc. The only exception, also noted consistently by Chataigner, corresponds exactly to a statement made by Wilson: when an object pronoun is postposed to the verb, the stress moves to the final syllable”134 (Bal 1983a: 19, drawing on Chataigner 1963). To be sure, this is the same pattern we find in PA and SCV (and not in any other Iberian-based creole). Moreover, in GBC too, the phoneme /o/ of the 2sg pronoun bo is raised to [u] in object position (Wilson 1962: 35), as exemplified above (1a,b) for PA and SCV. The issue of verb stress appears to be more complicated for GBC, however, given that neither Kihm (1994) nor Quint (2000a: 88) subscribe to Bal’s and Wilson’s account. Kihm (1994: 14), for instance, asserts that, contrary to PA and SCV, verbs in GBC “are stressed on the final syllable”, even though he admits that “[d]istortions arise when the verb is preceded by an aspect auxiliary or the negation (…) or the subject pronoun n ‘I’”. Research on the Casamance variety of GBC currently conducted by the native speaker and linguist Noël Bernard Biagui (forthcoming) is likely to provide a more accurate picture. 2.4.2.

On the diachrony of paroxytonic verb stress in PA and SCV

The paroxytonic verb stress in PA and SCV could be indicative of a general preference for paroxytonic disyllabic words in both creoles (cf. Quint 2001a: 274f.n.; 2001b: 76). In this respect, it is interesting to note that the shift from etymologically oxytonic to paroxytonic stress also affected words other than verbs in both creoles, as is shown by the case of PA / SCV  tambe ‘also’ < Port. também. By means of contrast, de Friedemann & Patiño Rosselli (1983: 91f.n.) draw attention to the opposite stress shift (from paroxytonic to oxytonic) in PLQ: PLQ ku sa ‘thing’ u to ‘other’ < Sp.  cosa,  otro. Interestingly, the paroxytonic verb stress found in PA and SCV has led some scholars to assume that their verbs are derived not from the lexifiers’ infinitives 134 Original quote: “Sont donc paroxytoniques des verbes comme debe (…), more (…), fala (…), fasi (…), etc. La seule exception, d’ailleurs notée de façon systématique par Chataigner, correspond exactement à un fait relevé par Wilson: quand un pronom objet est posposé au verbe, l’accent se déplace sur la syllabe finale”

2.5. Final remarks on phonology

81

but rather from the paroxytonic 3rd person singular forms. More recently, Quint (2000b: 143; 2001b: 74–76) has shown convincingly that this assumption is incorrect and that, as in most Atlantic creoles, the lexifiers’ infinitives provided the etyma for (the large majoritiy of) verbs in PA and SCV. For SCV, one such indication is that verbs in all other Sotavento varieties (Fogo, Brava, Maio) as well as in the Barlavento varieties are still oxytonic (Quint 2000b: 60; 2001b: 75)135 , suggesting they were oxytonic also in the proto-variety. Another indication is provided by verbs such as SCV fla ‘to speak’, konxi ‘to know’ and parsi ‘to seem’ < Port. falar, conhecer, parecer, all of which show the loss of the penultimate etymological vowel, suggesting this vowel was initially unstressed (Quint 2000a: 89). A third and final indication is provided by the raising of vowels in verbs such as SCV piska and kumpra < Port. pescar, comprar, again suggesting that the penultimate vowels were initially unstressed (Quint 2001b: 75). The latter two arguments apply to PA as well. First, we find forms such as PA parse, kana < Sp./Port. parecer, Sp. caminar, showing the syncope of the penultimate vowel (Quint 2000b: 143). Secondly, the penultimate vowels were systematically raised in verbs such PA piska, kumpra < Sp./Port. pescar, comprar (cf. §2.1.1 & 2.1.2). This evidence of course not only shows that Spanish and Portuguese infinitives provided the etyma for most verbs in PA and SCV; it also suggests that verbs in PA and SCV were originally oxytonic and obtained paroxytonic stress some time after creolization. It is implied that we cannot be sure whether this stress shift had already occurred prior to, or after the hypothesized separation of PA and Upper Guinea PC (in the latter stages of the 17th century; cf. chapters 7 and 8). We might thus be dealing with a parallel but independent development in PA and SCV, which would still be noteworthy, considering the idiosyncrasy of the feature within the branch of Iberian-based creoles.

2.5. Final remarks on phonology In §2.1, vowel features were discussed of which we may particularly recall the preference observable in PA and Upper Guinea PC for monophthongs, which were either retained from the Portuguese etymon or otherwise achieved through monophthongization of original diphthongs. In §2.2, a series of consonant changes was addressed.Among other things, we highlighted the distribution of the phoneme /S/ in PA. Not only is this distribution 135 But see Meintel (1975), who systematically indicates paroxytonic stress on verbs for Brava CV.

82

Phonology

remarkably close to its distribution in SCV, it also challenges the hypothesis that Old Spanish played any role of significance in the formation of PA. Instead, Old Portuguese appears to have been involved in this formation, as is suggested by the retention of the Old Portuguese phonemic contrast between /S/ and /tS/. Interestingly, it was shown that the two creoles also share a sound change that distances the lexemes from the Portuguese etyma, in as far as voiced fricatives are rejected in the more conservative layers of the vocabularies. In addition, both this feature and the distribution of /S/ and /tS/ were shown to be quite different from what we find in Gulf of Guinea PC and PLQ. Section 2.3 provided an overview of strategies of syllabic restructuring, in an attempt to show that these largely coincide in PA and Upper Guinea PC while again contrasting significantly with strategies of syllable restructuring in PLQ and Gulf of Guinea PC. The chapter closed with a reflection on the paroxytonic stress pattern found on disyllabic verbs in PA and Upper Guinea PC, a topic that warrants future research (particularly also on GBC), but in any case brings to light a remarkable synchronic parallel between PA and SCV. Needless to say, the hypothesized three centuries separating PA and Upper Guinea PC and their different histories of subsequent language contact have resulted in some conspicuous differences between these creoles in the domain of phonology. These consist, for instance, in the relatively recent incorporation in Upper Guinea PC of nonnative phonemes such as the voiced fricatives /v, z, Z/ from Portuguese or, in PA, the borrowing of the vowels /ø/, /y/ as well as the voiced fricatives /v, z/ from Dutch in addition to the /x/ from Spanish as well as Dutch (cf. Holm 1988: 316). For both PA and Upper Guinea PC, however, these phonemes can and have been shown to be typical of meso- and acrolectal varieties or otherwise to predominate in the periphery of the lexicon and recent loans (cf. Quint 2000b: 125, 127). If we filter out these nonnative features, an image of strong overlap between the phonological make-up of PA and that of Upper Guinea PC is what remains.

Chapter 3 Selected parts of speech

Introduction While the view that in principle all parts of speech can be borrowed is wellestablished in present-day contact linguistics (owing to, amongst others, Thomason & Kaufman 1988), it is uncontroversial that content words are typically more easily replaced than function words (cf. the empirical evidence provided in Tadmor 2009: 59). This is due in part to the relative semantic transparency of the former versus the semantic opacity of the latter (Matras 2009: 153–165). As a logical consequence, if PA indeed results from the relexification of an early variety of Upper Guinea PC, we can expect to find evidence for this precisely in the correspondence of their function words. To scrutinize this idea, the present chapter focuses on the following four closed classes: personal pronouns (§3.1), prepositions (§3.2), interrogatives (§3.3) and conjunctions (§3.4). The closing section (§3.5) discusses miscellaneous features which are less clearly paradigmatically organized but which are nonetheless revealing of some interesting parallels between PA and Upper Guinea PC.

3.1. Personal pronouns Pronouns are among the parts of speech considered to be most resistant to change diachronically (Thomason & Kaufman 1988: 75, 76; Parkvall 2008: 238; Muysken 2008: 95; Matras 2009: 203–208; Heine & Song 2010: 117). The stability of pronouns correlates not only with their being among the most frequent elements in most languages (e.g. Salmons 1994: 64), but also with the fact that they are generally tightly organized in paradigms (Muysken 2008: 244). This is especially true for closed sets of pronouns as found in the Indo-European language family (Muysken 2008: 90; Matras 2009: 203) as well as in most Atlantic creoles including PA and Upper Guinea PC. Their diachronic stability led Ross (2005) to use pronouns as a diagnostic for grouping Papuan languages. Similarly, Greenberg’s (1987) hypothesized language family Amerind was based primarily on the analysis of personal pronouns (cf. Muysken 2008: 95). Muysken (2008: 89) concludes that, “[w]hen a group of

84

Selected parts of speech

geographically contiguous languages shares a number of pronominal elements, there is a great likelihood that they are related”. In addition, scholars such as Arlotto (1972: 188), Schwegler (1999b: 238) and Clements & Koontz-Garboden (2002: 208) have emphasized the usefulness of pronouns in establishing genetic relationships. The relevance of a comparison of the pronominal paradigms to testing the hypothesis that PA and Upper Guinea PC are genetically related is therefore evident. Table 26 shows the paradigm of oblique pronouns in PA and Upper Guinea PC, while Table 27 contains the corresponding emphatic a- pronouns. Table 26. Oblique pronouns136

1sg 2sg 3sg 1pl 2pl 3pl

PA

GBC

SCV

mi bo e∼el nos boso nan

mi bo el nos bos elis

mi bo el nos nhos es

Table 27. Emphatic a- pronouns137

1sg 2sg 3sg 1pl 2pl 3pl

PA

GBC

SCV

ami abo – anos aboso anan

ami abo – anos abos –

ami abo ael anos anhos aes

136 Birmingham (1970: 147) and Martinus (1996: 183) compared pronouns in PA and GBC; Quint (2000b: 136–138) did the same for PA and SCV. Both GBC and SCV developed sets of distinctive clitic subject and object pronouns (cf. e.g. Rougé (2004a: 22, 24; Baptista, Mello & Suzuki 2007: 77). Although PA does not have a clitic pronominal system as extensive as Upper Guinea PC, two salient clitic features reminiscent of Upper Guinea PC were discussed in §2.4. For more general details on pronouns in the respective creoles, see, for instance, Veiga (2000: 172, 173) and Lang (2002: xxxiii) for SCV, Scantamburlo (1999: 180–182) and Kihm (1994: 141, 142, 150–152) for GBC, and Kouwenberg & Murray (1994: 38–40) and Munteanu (1991: 142–144) for PA. 137 Note that the emphatic plural pronouns anos, aboso and anan are not attested on Curaçao, but only in the Aruban variety of PA, which can therefore be said to be

3.1. Personal pronouns

85

The singular pronouns of PLQ (1sg yo∼i∼mi, 2sg bo, 3sg ele∼e) and Gulf of Guinea PC, e.g. ST 1sg ami∼mu, 2sg bo, 3sg e∼ele, are quite similar in form to those of PA and Upper Guinea PC.138 The plural pronouns provide more contrast: ST 1pl no, 2pl (i)nase, 3pl i(ne); PLQ 1pl suto, 2pl enu∼utere, 3pl ané (Bickerton 2002: 39; cf. Rougé 2004a: 27 for Gulf of Guinea PC; Schwegler & Green 2007: 298–300 for PLQ).139 Typologically, PA and Upper Guinea PC differ from Gulf of Guinea PC in that the latter has incorporated an impersonal pronoun a (Hagemeijer 2007: 17, 18; Maurer 2009a: 61). Below, additional arguments are presented in favor of grouping PA and Upper Guinea PC apart from other Iberian-based creoles in the pronominal domain. First, the form (a)mi is shown to set the Afro-Portuguese creoles and PA apart from the Asian Portuguese creoles. Then, I highlight the paradigm of emphatic a- pronouns as setting PA and Upper Guinea PC apart from Gulf of Guinea PC and other creoles. Following this, the fact that SCV has the 2pl nhos where PA and GBC have boso/bos is accounted for. I then briefly digress to focus on the more conservative in this respect (Martinus 1996: 183; cf. Holm 1988: 203; Maurer 1988: 37; Dijkhoff 2000: 66, 67). 138 It is interesting to note that the 2sg pronoun Sp. tú/Port. tu left no cognates in any of the Iberian-based creoles discussed here or in Brazilian Portuguese (cf. Kihm 1994: 282; Barme 2008: 21f.n.). One exception is Chabacano, which has 2sg ébos, tú and usté in complementary distribution (Lipski & Santoro 2007: 393), but tú in Chabacano might well be a recent acquisition. Schwegler (2002: 297f.n., 298f.n.) provides an overview of the diachronic development and diachronic and synchronic use of vos as a 2sg and a 2pl pronoun in European and Brazilian Portuguese. The Spanish 2sg/2pl vos is not widespread in the wider area around Cartagena or in Venezuela (Schwegler 2002: 297; cf. Quesada Pacheco 2000: 86), i.e. the area where PA and PLQ are spoken. But this is of course no evidence for an Afro-Portuguese origin for PA and PLQ bo: it is perfectly possible that vos was once more widespread in Caribbean/Colombian Spanish, only to disappear later due to pressure from European Spanish, with which there was ongoing contact (Schwegler 2002: 297f.n.). Gutiérrez Maté (2010: 870) draws new evidence from 17th-century archival texts showing that, indeed, “no resulta difícil aún encontrar ejemplos de voseo en el español norcolombiano de la época”, allowing for the possibility “que vos hubiera entrado en el palenquero desde el español” (Gutiérrez Maté 2010: 870). 139 Schwegler (2002) is an exhaustive study of the (origins of) the pronominal system of PLQ; for ST pronouns, Hagemeijer (2007: 9–72) offers a detailed discussion; for PRI, see Maurer (2009a: 56–63); Schmidt-Riese (2009), furthermore, provides a comparison of pronouns in GBC, ANG and PA on the one hand with those in Portuguese on the other. For overviews of the pronominal paradigms of other (Iberianbased) creoles, the reader is referred to the sections on pronouns in Holm & Patrick (2007).

86

Selected parts of speech

class of polite pronouns in PA. To close, I elaborately discuss the diachrony of the 3pl pronoun and nominal pluralizer PA nan. While this feature drastically separates PA from Upper Guinea PC on a synchronic level, I will provide a number of clues suggesting that the feature constitutes a non-native acquisition and does therefore not challenge the hypothesized kinship between PA and Upper Guinea PC. 3.1.1.

1sg (a)mi

According to Lipski, scholars agree that “the Asian [Portuguese] creoles in general are typologically different enough from Afro-Lusitanian creoles, and similar enough amongst themselves, to warrant a separate subclassification” (Lipski 2005: 30). One of the distinctive classifying features is that “all AsianPortuguese creoles use eu for the first person singular subject pronoun, while all Afro-Iberian creoles use a derivative of (a) mim” (Lipski 2005: 30). The Spanish-based creoles PLQ and Chabacano both have a derivative of Sp. yo as a subject pronoun (Schwegler & Green 2007: 298; Lipski & Santoro 2007: 393). Though PLQ has mi as an object pronoun, 1sg yo is considered “native to the Creole” (cf. PLQ ku yo ‘with me’ and PLQ *ku mi, versus PA and Upper Guinea PC ku mi ‘with me’; Schwegler & Green 2007: 298). In sum, the 1sg (a)mi seems to set PA apart from the Spanish-based creoles as well as from the Asian Portuguese creoles and provides a straightforward argument for grouping PA with the Afro-Portuguese creoles. Within the latter branch, as shown below, the presence of five emphatic a- pronouns allows grouping PA with Upper Guinea PC rather than with Gulf of Guinea PC. 3.1.2.

Emphatic a- subject pronouns

Shown in Table 27 is a characteristic set of emphatic a- subject pronouns found in PA, SCV and GBC. These are used for focusing, emphasis and insistence (Intumbo 2008: 273) and consist of the oblique pronoun and a prefixed a-. (2)

a. PA

abo, bo ta kere ku abo por pidi Dios you you IMP believe that you can ask God ‘you think you can ask God’ (Valerianus-Fermina 1992: 52) b. GBC Abo, bo ta papia Kriol diritu you you IMP speak Kriol right ‘You speak excellent Kriol’ (Einarsdóttir 2004: 188) c. SCV abo bu ta bira ómi bédju you you IMP turn man old ‘You will turn into an old man’ (Lang 2002: 90)

3.1. Personal pronouns

(3)

87

a. PA

Anos ta puntra nos mes we IMP ask our self ‘We ask ourselves’ (Solo, 15-07-2009) b. GBC anos no misti pa no bai pa votason we we want for we go for election ‘we want to go to the elections’ (Scantamburlo 2002: 634) c. SCV anos nu ta kredita na nos mé we we IMP believe in we self ‘we believe in ourselves’ (Lang 2002: 140)140

The emphatic subject pronouns may well owe their phonetic shape to analogy with the lexifiers’ dative/accusative constructions (I will return to this below). Their syntactic behavior, however, does not conform to Spanish or Portuguese (pace Munteanu 1996a: 296 and Kramer 2004: 184): as opposed to the Iberian dative/accusative paradigm (Sp. a mí / Port. a mim, etc.), the creoles’ emphatic a- pronouns are subject pronouns and thus do not occur in direct or indirect object position (Intumbo 2007: 48). Kihm (1989: 360f.n.) is right to stress the idiosyncrasy of having a paradigm of emphatic pronouns: “Such formal contrasts of two pronominal sets [are] not so clear in other creole languages”. Indeed, although the emphatic 1sg subject pronoun ami is found in a wide range of Afro-Iberian speech varieties141 , having a paradigm of four or more emphatic a- subject pronouns like PA and Upper 140 Although the emphatic pronouns are followed by a non-emphatic pronoun in examples (2a–c), they can also occur autonomously in PA (cf. 3a). Peck (1988: 132, 166) and Intumbo (2008: 273) maintain that emphatic subject pronouns do not occur autonomously in GBC, but counter-examples are found in plenty, e.g. GBC abo fala kuma… 1sg-speak-COMP ‘you said that…’ (Scantamburlo 2002: 152); Ami i vokalista 1sg-singer ‘I’m a singer’ (Peck 1988: 114); Abo na konta mintida 1sg-PROG-tell-lie ‘you’re telling lies’ (Montenegro & de Morais 1995: 175); te Guiné ami na bay until-Guinea-1sg-PROG-go ‘I am going to / till Guinea’ (Birmingham 1975: 21). The (un)grammaticality of autonomous emphatic pronouns may reflect dialectal variation within GBC. For SCV, Quint (2000a: 162) affirms – in line with Intumbo and Peck for GBC – that the emphatic subject pronouns do not occur autonomously. Here too, counterexamples exist (e.g. Lang 2002: 177), but autonomously occurring emphatic pronouns do appear to be much rarer in SCV than in PA and GBC. 141 Lipski (1991) provides an exhaustive overview of the spread of ami (and variants) among restructured Iberian varieties including e.g. Cuban Bozal Spanish, Gulf of Guinea PC, and Gil Vicente’s Lingua de Preto. The 1sg ami also appears in Media Lengua, though as an object pronoun like its Spanish etymon (Muysken 2008: 217).

88

Selected parts of speech

Guinea PC seems to be quite unique at least among Iberian-based creoles. PRI, for instance, has no more than two (ami, atxi; Rougé 2004a: 27); ST has only one (ami; Hagemeijer 2007: 21). In ANG and PLQ emphatic pronouns are absent as a category (Rougé 2004a: 27; Schwegler 1998: 260). The forms abo in examples (2a–c) and anos (3a–c) are exclusively shared between PA and Upper Guinea PC. Several authors have proposed that the category of emphatic subject pronouns should be analyzed as a substrate-induced feature (e.g. Lenz 1928: 136; Kihm 1994: 152, 360f.n.). Indeed, phonetically distinct emphatic subject pronouns are absent from Spanish and Portuguese, while present in the West-Atlantic branch of African languages spoken in Upper Guinea (Peck 1988: 117–121; Kihm 1994: 152; Intumbo 2006: 113 & 2007: 48, 49; Baptista, Mello & Suzuki 2007: 76; 2008: 273, 274; Holm & Intumbo 2009: 259–261). On the other hand, it may well be that the emphatic subject pronouns result from the substrateindependent reanalysis of the superstrate dative a- forms in focus constructions such as Sp. a mí (lo que) me interesa (es), a mí me gusta (Detges 2001: 298, 299; cf. Schwegler 1990: 108). A similar process of reanalysis probably occasioned the grammaticalization of the oblique object pronouns moi and toi as emphatic subject pronouns in (Old) French (Detges 2001: 358–360). Of course a convergence of sub- and superstrate patterns is not unthinkable either. The fact remains that a series of four or more emphatic a- subject pronouns such as is found in PA and Upper Guinea PC is attested in no other Iberian-based creole or in any (non-standard) variety of Spanish or Portuguese. 3.1.3.

2pl SCV nhos

The etymon of SCV nhos is Port. senhores ‘gentlemen’ (Rougé 2004a: 259). Table 26 makes it tempting to assume that an early 2pl SCV *bos once existed alongside, or instead of, SCV nhos, an assumption also made by Quint (2000b: 137, 204). More precisely, SCV nhos probably originated as a polite form alongside *bos only to replace it at a later stage. This idea is further supported by the fact that 2sg SCV bo has the polite variants nho/nha (< Port. senhor(a) ‘mister, madam’) (Rougé 2004a: 24) and that in GBC, both the 2sg bo and the 2pl bos have polite variants: nhu/nha and nhus/nhas (Peck 1988: 113). Note, furthermore, that the 2pl pronouns in São Vicente CV are bzot∼bezote (< Port. vos + otro) and bosês (< Port. vocês) (Swolkien 2009: 1; Veiga 2000: 172), suggesting that the CV 2pl pronoun was rather unstable in the early period. The differentiated CV forms of address are not all too surprising. Borrowing or contact-induced changes in the pronominal domain are relatively rare but, if they occur, often concern forms of address (Matras 2007: 53). Also, judging from the fact that in certain phonetic surroundings the 1pl CV nos is realized as

3.1. Personal pronouns

89

/no/ and 2pl GBC bos as /bo/, it is by analogy easy to conceive that an earlier CV 2pl *bos was, in certain contexts, also realized as /bo/ and thus homophonous with the CV 2sg bo. This homophony may have triggered the substitution of the proto-2pl *bos by its polite variants nhos and bzot∼bezote in CV. 3.1.4.

Digression: 2sg polite pronouns in PA

The idea that pronouns of address are more easily borrowed than other pronouns is supported by the fact that PA has a whole array of polite 2sg pronouns. Among these we find mener/yùfrou (< Dutch meneer ‘mister’, juffrouw ‘madam’), señor(a) (< Sp. idem) and the more archaic forms shon (< Sp. señor/Port. senhor) and mosa (< Sp. moza/Port. moça ‘lady’). They are used in indirect formal address constructions, a phenomenon typical of European and Brazilian Portuguese rather than of Spanish. Barme (2008), however, convincingly argues that Dutch should not be overlooked as a possible source by showing parallel constructions in 17th- and 18th-century Dutch as well as in Afrikaans. (Coffen [2002: 272–275] provides a discourse analysis of polite 2sg pronoun usage in PA.) Interestingly, a (polite?) 2nd singular pronoun boste is found, for instance, in the second oldest written PA sample, a 1776 PA dialogue, transcribed in the declaration of the Jew Semuel Costa Andrade (e.g. Maurer 1998: 205, 206). According to van Putte & van Putte-de Windt (2005a: 72), boste is an archaic polite 2sg pronoun (viz. a variant of bo) in standard PA, but a 2pl pronoun (viz. a variant of boso) in Sephardic PA, as confirmed by Henriquez (1988: 8). Henriquez furthermore notes that she heard it being used as late as in the 1930s, but not afterwards: “Nowadays, we only use it once in a while just for fun”142 (1988: 8). PA boste may well constitute a loan from Galician: as pointed out by Bossong (2005: 73, 74), the pronoun bostede is used in Galician as a general form of respect and politeness. This is interesting in light of the fact that, as noted in §1.4.3, scholars such as Maduro (1966a, 1966b) have on several occasions pointed at possible Galician etymologies of PA words. Ultimately, the etymon of PA boste ‘you (sg. & pl.)’ seems to be Spanish Vuestra Merced ‘Your Mercy’ or a variety thereof (an insight I owe to Miguel Gutiérrez Maté, p.c.). Before evolving into usted, this compound 2sg passed through the intermediate forms vosasted, vosted and vusted (among others), some of which survived in non-standard varieties and indeed in Galician. One can thus hypothesize the etymology from Sp. Vuestra Merced, through dialectal Sp. vosted or Galician bostede, to PA boste. (Pla Cárceles [1923] and Lapesa 142 Original quote: “Awendia nos ta uz’é únikamente un ora un ora, na chansa”

90

Selected parts of speech

[1970] discuss the development ofVuestra Merced > usted in Spanish; cf. Gutiérrez Maté [to appear a] on variation betweenVuestra Merced and usted in colonial Caribbean Spanish.) 3.1.5.

PA nan

One reason why the relatively straightforward resemblance between PA and Upper Guinea PC in the pronominal system has been largely neglected in recent studies on the origins of PA, I assume, is the 3pl pronoun PA nan, which also functions as a nominal pluralizer (a feature often abbreviated as 3pl = PL). Its marked character and strong resemblance in form and use to the Gulf of Guinea PC 3pl pronouns (also nominal pluralizers) seems to have overshadowed the paradigmatic correspondence between PA and Upper Guinea PC in the pronominal domain: “In attemps to demonstrate the interrelatedness of Papiamentu and the Gulf of Guinea PC, parallels have often been drawn (…) between their respective 3pl pronouns” (Parkvall 2000: 104). Maurer (1998: 196), for instance, states: “Papiamentu definitely contains various Afro-Portuguese elements, among which the plural marker nan, related to the plural markers of Gulf of Guinea PC, stands out”143 . Maurer (2002) is dedicated specifically to the correspondences between PA nan and Gulf of Guinea PC’s 3pl pronoun/pluralizer. Another case in point is Birmingham’s (1971: 301; cf. Valkhoff 1966: 96) discussion of PA nan: nan is one of the few African morphemes in Papiamentu. As a point of fact, the plural morpheme nan appears as well in at least one Portuguese dialect still spoken along the west coast of Africa. Part of my dissertation attempts to prove that Papiamentu originated in and around the Gulf of Guinea and that its basis is a Portuguese lingua franca spoken in that area in the days of the slave trade.

In none of the above-cited literature mention is made of the fact that the remainder of PA’s pronominal system shows a one-to-one correspondence with Upper Guinea PC.144 143 Original quote: “Es cierto que el papiamentu contiene varios elementos afroportugueses, entre los cuales el más destacado es el marcador del plural nan, que está emparentado con los marcadores del plural de los criollos afroportugueses del Golfo de Guinea” 144 PA nan also played a role in Lipski’s (1987a) paper on the origin of the item lan∼nan, “a portmanteau definite article, masculine or feminine, singular or plural” (1987a: 294), attested in written representations of 19th-century varieties of Afro-Caribbean Bozal Spanish. Cf. Álvarez Nazario (1974: 185, 186), who proposed a connection between the PA and Gulf of Guinea PC plural morphemes on the one hand and the

3.1. Personal pronouns

91

It is of course quite possible (though by no means inescapable) that PA nan is a cognate of (one of) the near-homophonous forms found in Gulf of Guinea PC (cf. ST nen; ANG ane; PRI ina; Annobonese nan; Maurer 2002: 131). However, as will be argued below, the origin of PA nan has little bearing on the origins of the remainder of PA’s pronominal paradigm: whatever the language(s) PA nan was drawn from, several clues suggest it became integrated postformatively145 into a pre-existing grammatical system containing neither a distinct 3pl pronoun nor a (productive) plural marker. 3.1.5.1. Absence of a distinct 3pl pronoun in Early Upper Guinea PC First of all, indications are available that Early Upper Guinea PC did not have a distinct 3pl pronoun but rather one pronoun for 3sg and 3pl. If Early PA inherited this pronominal paradigm, there was scope for the distinct 3pl nan to be incorporated. One such indication is that the oblique 3pl pronoun is not identical in SCV (es)146 and GBC (elis) either, suggesting a postformative development for both. Moreover, from a semantic point of view, as Quint (2000b: 204) notes, “one can easily imagine a stage in which *EL PISKA would have meant either ‘he has fished’ or ‘they have fished’ depending on the context”147 . A look at GBC substantiates the hypothesis of homophony between the 3sg and 3pl pro-

Bozal Spanish definite article lan∼nan on the other. Lispki (1987a) convincingly challenges such a connection. 145 To be sure, whenever I claim a PA feature to have been acquired ‘postformatively’ or ‘after creolization’, what I claim is that the feature was added to PA’s grammar after its separation from its Upper Guinea PC ancestor in the latter stages of the 17th century. Reversely, when I claim a PA feature to be ‘native’ or ‘part of the original grammar’, I claim it was present in the ancestor creole prior to the separation of PA. 146 The SCV 3pl es may well have its origin in the SCV demonstrative es. The development of demonstratives into 3rd person pronouns is described for several languages world wide (Heine & Kuteva 2002: 109, 112, 113). Interestingly, PA proverbs collected by Brenneker (1963) provide several instances suggestive of personal pronominal usage of the Early PA definite marker es (which has developed into the modern-day PA article e, cf. §3.5.2). Consider, for instance: Es cu hunga cu candela ta kima su man DEF-REL-play-with-fire-IMP-burn-hand ‘he who plays with fire, burns his hands’ (1963: 26); Es cu mata di heru, di heru e ta muri DEFREL-kill-of-iron-of-iron-he-IMP-die ‘he who kills with iron, will die of iron’(1963: 45). 147 Original quote: “on peut très bien imaginer un stade ou *EL PISKÁ (…) aurait signifié il a péché ou ils ont péché en fonction du contexte”

92

Selected parts of speech

nouns in Early Upper Guinea PC: GBC’s clitic 3pl subject pronoun is e148 ; the clitic 3sg subject pronoun is i ∼ el 149 (cf. PA & SCV e∼el). A clarifying parallel can be drawn between PA and PLQ. Just as in PA, an African 3pl pronoun (PLQ ané < Kikongo ba ne ‘they (over there)’) managed to establish itself in PLQ amongst a set of pronouns of Iberian origin150 (Schwegler & Green 2007: 299, 300). Moreover, in keeping with the hypothesized homophony between the 3sg and 3pl pronouns in proto-Upper Guinea PC, the PLQ 3sg pronoun ele can occasionally take on the function of 3pl pronoun (Schwegler 1999a in Parkvall 2000: 106; Schwegler & Green 2007: 300).151 These facts suggest that the African-derived 3pl pronouns PLQ ané and PA nan followed a similar trajectory in that both were integrated into a paradigm without a distinct 3pl pronoun. If the early variety of Upper Guinea PC from which I assume PA descends did indeed lack a distinct 3pl pronoun, the introduction of nan in the empty 3pl slot of the original paradigm can be explained as a rather unmarked case of ‘gap filling’– “Gap filling may simply mean that the replica language acquires another item enriching an existing grammatical paradigm” (Heine & Kuteva 2005: 128) – with the aim of resolving ambiguity between the 3sg and 3pl pronouns. Interestingly, Schuchardt (in Hagemeijer & Holm 2008: 145) had already analyzed PA nan in a very similar way, assuming it must have been a borrowing (rather than an inherited feature) motivated by original homophony between PA’s 3sg and 3pl pronouns: “This borrowing is not strange if one considers that, in consequence of the loss of the plural -s, Port. elles [‘they’] must have conflated with Port. elle [‘he, she, it’]”. 3.1.5.2. Large absence of plural morphology in Early Upper Guinea PC In addition to the absence of a distinct 3pl pronoun, the plural marking morphology is also likely to have been ‘vacant’in Early Upper Guinea PC. Maurer (2002: 131f.n.), for instance, remarks: “It is not clear when the [plural] marker -s was

148 E.g. GBC fal élus é b˜e [speak-them-they-come] ‘tell them they (should) come’ (Bal 1983b: 22). 149 E.g. GBC í labá mõ [he-wash-hand] ‘he washed his hands’ (Bal 1983b: 23). 150 Note also PLQ’s 2pl enu, which is found in addition to 2pl utere (< Sp. ustedes ‘you (pl.)’) and whose Kikongo origin is discussed in Schwegler (2002: 309–312). 151 Cross-linguistically, homophony between the 3sg and 3pl pronouns is not unique and is attested, for instance, in Lithuanian (Comrie 2000: 36) and French (Parkvall 2000: 106). On the other hand, Parkvall (2000: 106) also notes that “no major West African language uses the same form for 3sg and 3pl other than marginally”.

3.1. Personal pronouns

93

brought into use in CV and GBC”152 , raising the possibility that Upper Guinea PC initially did without plural morphology. As noted by Bartens (1996: 33), the grammatical sketches of 19th-century CV provided by Brito (1887: 348) and Costa & Duarte (1886: 258) unambiguously support this idea, as they show that by that time the Portuguese plural -s was still only attached to the polite (and non-native) personal pronouns nho, nha, nhara (< Port. senhor, senhora) and to diminutive forms. The overt plural marking in present-day varieties of CV, therefore, is likely to constitute what Bartens calls a “Dekreolisationsphänomen”, and, in fact, the rural, basilectal variety of SCV is still characterized by the large absence of overt plural marking (Quint 2000a: 321–326). The lacking of plural morphology in Early Upper Guinea PC should be no surprise, given that many (creole) languages still do without. In the IndoPortuguese creole Korlai, for instance, an overt plural marker is absent (Clements 2007: 167). Parkvall (2000), when giving estimates of the time span needed for certain syntactic features to be fully integrated into an (Atlantic) creole’s syntax, summarizes: [N]ominal plural marking might potentially be a case of late substrate influence. Its absence in many of the world’s languages proves it to be a feature that human language, regardless of whether it has gone through pidginization or not, can manage without. Indeed, overt plural marking has been shown to be a late development in the Indian Ocean FCs (…) and it was moderately grammaticalised in Negerhollands DC, which could be taken to suggest a relatively late development. (Parkvall 2000: 158, emphasis mine)

Note, furthermore, that (as already hinted at by Parkvall) the post-formative integration of a plural marker represents a relatively unmarked case of contactinduced change: “Nominal morphology is most frequently replicated in the case of plural markers” (Matras 2007: 43). Heine & Kuteva (2005: 133) confirm that processes of grammatical replication (or what they refer to as ‘equivalence’) are indeed common in the domain of nominal number marking. Several Amerindian languages (e.g. Maya, Nahuatl, Quechua) in contact with Spanish, for instance, have replicated nominal pluralization techniques and/or morphemes from Spanish and, occasionally, vice versa (cf. Muysken 2008: 183; Karttunen & Lockhart 1976: 23, 24). In fact, when confronted with the absence of the feature 3pl = PL in Fongbe, Lefebvre (1998: 86) explained Haitian yo [3pl = PL] as “a simple case of dialect levelling”.153 152 Original quote: “No está claro desde cuándo se empezó a emplear el marcador –s en el criollo de Cabo Verde y en el kriôl” 153 According to Matras (2007: 43), the main reason for the relative ease with which plural morphology is borrowed is that “[m]orphological plural marking (…) meets

94

Selected parts of speech

3.1.5.3. Final remarks on PA nan To summarize, it can plausibly be hypothesized that Early Upper Guinea PC lacked a distinct 3pl pronoun and had, if any, no pronounced plural marking morphology either. As a consequence, if Early PA did indeed develop out of a 17th-century variety of Upper Guinea PC, there seems to have been scope for PA nan (3pl = PL) to develop or be introduced by speakers of Kwa/Bantu languages (and/or Gulf of Guinea PC) once these became numerically dominant on Curaçao. It should still be pointed out that the pluralizer is prefixed to the noun in Gulf of Guinea PC, whereas PA nan is suffixed. This feature (“nominal plural suffixed”) is listed third in a ranking of cross-linguistically stable features composed by Parkvall (2008a: 238). These observations of course do not nullify the chances that PA nan was inherited from Gulf of Guinea PC, but if it was, this apparently happened some time after PA’s formation, when its typological make-up (including word order) had already stabilized.154 Regarding the hypothesized Sãotomense origins of the PLQ 3pl pronoun ané, Bickerton (2002: 40) – at his best when arguing against the afrogenesis of New World creoles – once commented: “But this is an African form and it cannot be discarded that ané derives directly from some African language. And even if this form does derive from ST, here we are merely dealing with a single form, not with a system”155 . It is revealing to note that PA and Upper Guinea the criteria for semantic transparency which is so often noted as a factor facilitating morphological borrowing”. 154 Boretzky (1983: 87–89), Holm (1988: 193–195), Parkvall (2000: 93–96) and Holm & Patrick (2007: feature 15.4) provide thorough overviews of (creole) languages with the feature 3pl = PL. The fact that the feature is absent from the European donor languages as well as from Asian and Pacific creoles, is an argument in favor of African substrate influence (Hancock 1975: 221; Parkvall 2000: 93, 94). In identifying possible African donor languages, Parkvall finds that the feature is “virtually confined to Lower Guinea”. This is supported by the fact that PLQ, whose substrate is mostly Bantu, lacks the feature (Hancock 1975: 221). For PA, Parkvall mentions the Kwa languages Twi, Ewe, Yoruba and Igbo as well as Atlantic Fulfulde as possible sources, since all these postpose their pluralizer. Interestingly, Parkvall (2000: 104) goes on to propose a Wolof etymology for PA nan. Wolof does indeed have a 3pl pronoun nañ(u) (Ndiaye 2004: 28). According to Jürgen Lang (p.c.), however, it is not used as a nominal pluralizer. 155 Original quote: “Pero ésta es una forma africana, y no se puede eliminar la possibilidad de que ané venga directamente de alguna lengua africana (…). Y aun, si de veras viene del Sãotomense este vocablo, aquí (…) se trata meramente de un vocablo suelto, no de un sistema”

3.2. Prepositions

95

PC do in fact share a system, indeed with the exception of one single form, PA nan. 3.1.6.

Final remarks on pronouns

Muysken (2008: 209) observed that “[c]reole pronouns tend to be based on the lexical shapes of the pronoun systems of their lexifiers, but the paradigm is often substantially restructured”, and so also in the case of PA and Upper Guinea PC. One is hard-pressed to attribute to chance the fact that such ‘substantial restructuring’ rendered two typologically and formally identical pronominal paradigms (PA nan excluded) in two such distant places. After all, the languages traditionally thought to have been present on Curaçao in the period in which PA emerged (i.e. second half of the 17th century) – Dutch, Spanish and perhaps Sephardic Portuguese on the one hand, and Kwa and Bantu on the other hand – were drastically different from the languages present on Santiago in the late 15th and early 16th centuries – basically only Portuguese and Mande and West-Atlantic. While some of the pronominal material discussed above may well be shown to also exist in one or another Iberian variety or creole (e.g. ami), it is important to stress, in line with Parkvall (2000: 101), that, for being a closed class with a high level of paradigmatic organization, “[a] pronoun paradigm is simply not assembled from bits and pieces in five different languages”. As a paradigm, PA’s pronouns resemble that of one other language only: Upper Guinea PC. The comparison presented above allows us to hypothesize a shared proto-paradigm of five oblique pronouns and a set of at least four emphatic a- subject pronouns (cf. Quint 2000b: 204156 ).

3.2. Prepositions Kouwenberg & Murray (1994: 52) correctly note that “Papiamentu is fairly rich in prepositions, in contrast with other Caribbean creole languages”. But while PA indeed stands isolated within the Caribbean, two other Atlantic creoles can in fact compete easily with PA in terms of prepositional richness: CV and GBC. The contrast between PA and Upper Guinea PC on the one hand and other Atlantic creoles on the other hand is remarkable. Günther (1973: 79) and Lipski (1989: 76), for instance, describe the scarcity of prepositions as typical of 156 Quint (2000b: 204) proposed a proto-Upper Guinea PC pronominal paradigm consisting of *MI, BO, EL, NOS, BOS, and four emphatic forms *AMI, ABO, ANOS, ABOS.

96

Selected parts of speech

PRI and Saramaccan157 respectively and Bruyn (2009: 323, 324), furthermore, explains that “in Sranan, (…) from the earliest sources onwards until the middle of the 20th century, simple locative prepositions with semantic content do not occur”. The opposite is true for PA and Upper Guinea PC, where simple locative prepositions with semantic content belong to the core of the morphosyntax. In creoles with a reduced set of prepositions, other syntactic devices exist to express prepositional relations (Endruschat 2004: 47; Lipski 2005: 274). For ST, for instance, Hagemeijer (2007: 86) points out that “most items that exhibit a ‘prepositional function’ cannot be considered prepositions proper. In fact, nouns and verbs fill in this function to a significant extent”. This applies to the other members of Gulf of Guinea PC as well (Günther 1973: 80; Lorenzino 2007: 22; Maurer 2009a: 50, 51), and a basic reading of glossed PRI texts provided in Maurer (2009a: 207–210) confirms that prepositions proper are largely lacking in story telling. In PLQ, moreover, mostly the locative adverbs akí, aká, aí and ayá take on a contextually determined prepositional function. Thus, PLQ akí kasa, according to the context, can be translated as any of the following: ‘in∼on∼around∼at∼to(ward) the house’ (Schwegler & Green 2007: 302). While the contrast with other Iberian-based creoles is thus quite patent, there are also some compelling dissimilarities with respect to Spanish and Portuguese. For instance, despite the relative wealth of prepositions in PA and Upper Guinea PC, both creoles lack reflexes of the Iberian high-frequency prepositions por ‘by, for, through’, a ‘to’ and sobre ‘on, over’. Throughout the remainder of this section, it will be pointed out that PA and Upper Guinea PC coincide in having grammaticalized other etyma to cover the scope of these three core Iberian prepositions. Table 28 exhibits correspondences between PA and Upper Guinea PC in the paradigm of simple prepositions. Those listed here do not require combining with di in PA (Kouwenberg & Murray 1994: 52). With the exception of SCV dentu158 , the same prepositions are singled out by Quint (2000a: 201) as representing the most basic and original set of SCV prepositions.

157 According to Muysken (2008: 66), Saramaccan has three basic prepositions each with a wide semantic range: “ku ‘instrumental, comitative, NP-coordination’, a ‘locative, directional’, and u ‘possessive, benefactive’”. 158 The inclusion of Upper Guinea PC dentu in Table 3 is justifiable, however, since it does occasionally occur as a simple preposition, as in SCV dentu-loja ‘in(side) the store’ (Mendes et al. 2002: 168), which equals PA den pakus (with pakus < Du. pakhuis ‘warehouse’).

3.2. Prepositions

97

Table 28. Simple prepositions PA

Upper Guinea PC

GLOSS

di ku na pa riba te den ∼ denter

di ku na pa riba, ruba te, ti dentu ∼ dentru ∼ dentur

‘from’ ‘with’159 ‘in, to’ ‘for, by’ ‘on, over’ ‘until’ ‘in’

While the formal resemblance between the two paradigms presented in Table 28 hardly needs any comment, some shared semantic and syntactic prepositional features and properties are discussed in the remainder of this section, with the aim, firstly, of further exploring the idiosyncrasy of the two prepositional paradigms in contrast with other Iberian-based creoles and, additionally, of weakening attempts to hold decreolization towards Spanish and Portuguese responsible for the correspondences between the prepositional paradigms of PA and Upper Guinea PC (e.g. Endruschat 2004: 66; 2005: 193). 3.2.1.

PA / Upper Guinea PC di

The genitive marker PA / Upper Guinea PC di ‘of, from’ covers largely the same semantic-syntactic ground as its etymon, Iberian de, including, unsurprisingly, the expression of possession. Moreover, word formation with di is fully productive in PA (Dijkhoff 1993: 104–108) and Upper Guinea PC (Kihm 1994: 127; Lang 2002: 111–113) and has given rise to a considerable number of coinages, such as PA pididó di limosna ‘beggar’ and GBC pididor di sumola ‘beggar’ (= Sp./Port. mendigo, pedinte). (For PA, Dijkhoff [1993] is an excellent case study of complex nouns.) Gulf of Guinea PC, by contrast, seems to prefer a zero connector for word formation (e.g. PRI môrada kaxi wall-house ‘the wall of the house’; Maurer 2009a: 29) as well as to express possession (e.g. ANG mingu tata m friendfather-my ‘the friend of my father, my father’s friend’; Lorenzino 2007: 19). Though compounds with di do exist in the varieties of Gulf of Guinea PC, Maurer (2009a: 50f.n.) believes that these are relatively recent innovations rather than original constructions. Furthermore, Lipski (2005: 297) presents the loss of Spanish de as a characteristic feature of bozal varieties of Spanish. 159 The use of comitative PA ku ‘with’ as a coordinate conjunction ‘and’ is discussed in §3.4.1.1.

98

Selected parts of speech

The particular use of genitive di ‘of’ in the formation of ordinal numbers in PA (e.g. PA e di dos les the-of-two-lesson ‘the second lesson / lesson two’) and superlatives (e.g. PA e makaku di mas grandi the-monkey-of-more-big ‘the biggest monkey’ (Maurer 1987b: 243f.n.)) has convincingly been shown to correspond to patterns found in Bantu languages (Maurer 1987b). Interestingly, however, Kihm (1994: 143) registered a similar di-based strategy for making ordinal numbers in GBC: “In ordinary situations, people will prefer to resort to the cardinal number preceded by di ‘of’ (e.g. omi ku na bin di kwatru ‘the man who is coming fourth’; also di kwatru ‘fourthly’)”. Note that Bantu played no role in the formation of GBC. 3.2.2.

PA / Upper Guinea PC na

This preposition derives from Portuguese na < em ‘in’ + a (feminine definite article), but has lost its definiteness, as shown by the fact that it combines freely with markers of definiteness (4) and generic nouns (5): (4)

(5)

a. PA

na es tera in DEF land ‘in this land’ b. GBC na es tera li in DEF land here ‘In this land’ (Scantamburlo 2002: 134)

(Jansen 1911: 27)

a. PA

Kere na Dioos i kere tambe na mi. Believe in God and believe also in me ‘Believe in God and believe in me as well’(Putman 1852: 26)160 b. SCV bu ta kre na Dés? you IMP believe in Dios ‘Do you believe in God?’ (Mendes et al. 2002: 157)

Owing to its presence in an impressive number of both Iberian-based and other creoles (Wagner 1949: 155f.n.; Parkvall 2000: 108; Holm 2009: 19), the preposition na often appeared as linguistic evidence in monogenesis frameworks (e.g. Thompson 1961: 108; Taylor 1971: 294, 295).161 Note, however, that in most

160 In modern PA, the verb kere ‘to believe’ is more typically construed with den, rather than with na. The Early PA examples of kere na ‘to believe in’ show that the distribution of na in Early PA used to resemble Upper Guinea PC more closely than it does nowadays. Still in Ambrosius (1905: 5), an example of PA kere na is found.

3.2. Prepositions

99

creoles that have na, this preposition functions as a portmanteau spatial preposition by the large absence of other locative prepositions. PA and Upper Guinea PC, on the other hand, are equipped with a relatively rich set of additional spatial prepositions (Table 28), including a zero marker between motion verbs and place indications (see further below). In other words, na is not a portmanteau spatial preposition in either PA or Upper Guinea PC. As for Gulf of Guinea PC, na appears as a portmanteau preposition in PRI, while ST has ni. Examples such as (4) are, however, impossible in Gulf of Guinea PC due to word order constraints: determiners occur postnominally (Maurer 2009a: 34–47). 3.2.3.

PA / Upper Guinea PC te

Besides its use as a simple preposition similar to the etymon (Port. até), PA / Upper Guinea PC te (∼ti in some varieties of SCV) typically heads the complex subordinating conjunction ‘until + hour + that’ (see §3.4.2.1). Thus, PA / Upper Guinea PC te ‘until’ (prep.) is in complementary distribution with te ora ki/ku ‘until’ (sub. conj.), unlike the etymon Port. até ‘until’ (prep.), which is in complementary distribution with até que ‘until’ (sub. conj.). Another indication of the level of integration and grammaticalization of PA / Upper Guinea PC te is that, besides the phonological reduction (apheresis of the /a/-), it has been decategorialized (cf. Heine & Kuteva 2005: 15): while the etymon Port. até is both a preposition (‘until’) and an adverb (‘even’), the creole form te functions as a preposition only, not as an adverb. To cover the adverbial meaning ‘even’, PA and Upper Guinea PC incorporated the clearly more recent loans PA hasta (< Sp. hasta)162 and Upper Guinea PC até (< Port. até) respectively. Noteworthy is furthermore the frequent use of the lexicalized compound preposition te na in both PA and Upper Guinea PC: (6)

a. PA

te na tempu di Huan until in time of John ‘(up) until the times of John’

(Matthew 11: 13163 )

161 In some cases at least, the homophony is mere coincidence. For instance, Dutch naar ‘towards’ was the source for the preposition na in Berbice Dutch Creole (cf. Holm 1988: 208). 162 Cf. PA luna i streanan, hende i animal, si, hasta paloe i piedra moon-and-starPLpeople-and-animal-yes-even-tree-and-stone ‘the moon and stars, people and animals, yes, even trees and stones’ (Jansen 1911: 20). 163 www.bendishon.com

100

Selected parts of speech

b. GBC te na tempu di Davi until in times of David ‘(up) until the times of David’ c. PA te na mitar until in middle ‘up to / up until the middle’ d. SCV ti na metádi until in middle ‘up to / up until the middle’

(Scantamburlo 2002: 591)

(Èxtra 25-10-2003: 4)

(Lang 2002: 10).

Needless to say, etymologically, there is very little that is Spanish about the composed PA preposition te na.164 In sum, Portuguese até has been integrated and lexicalized in identical ways in PA and Upper Guinea PC. It must be noted that reflexes of Portuguese até ‘until’ are found also in the Surinamese creoles, including Sranan (Huttar 1989: 272; Muysken 2008: 199), as well as in Gulf of Guinea PC: PRI t˜e∼ten∼ton, ST antê (Boretzky 1983: 195; Maurer 2009a: 237; Fontes 2007: 8). However, the fact that the PRI and ST forms differ considerably from one another in shape suggests that this item does not belong to the core grammar of Gulf of Guinea PC; Boretzky (1983: 195) indeed notes that t˜e is rarely used in PRI. In Papia Kristang, Port. até yielded ati ‘until’ (Baxter 2004: 147). 3.2.4.

PA / Upper Guinea PC riba (di)

Lipski (2005: 274) observed that the use of riba (derived from the adverb Sp. arriba / Port. (ar)riba ‘above’) as a bare preposition in PA “represents a significant deviation from monolingual Spanish usage”. Indeed, predicates such as PA / Upper Guinea PC riba (di) mesa ‘on the table’ (wherein di is optional in both creoles) are translated in mainstream Spanish as encima de / sobre la mesa, in Portuguese as em cima de / sobre a mesa165 . Neither Iberian encima

164 In modern mainstream Portuguese, sentences such as até na metade or até no tempo (de) are not unheard of, though less common than até a metade and até o tempo de (i.e. without em), and, in any case, surely not as regularized as the PA / Upper Guinea PC variant te na. 165 For the Saramaccan postposition liba ‘above’, Muysken (2008: 199) provides the etymon Port. riba de ‘above’. I have not been able to verify the existence of this preposition in Old or dialectal Portuguese. In any case, as far as I am aware, riba de is not in use in mainstream European Portuguese.

3.2. Prepositions

101

(de) / em cima (de) nor sobre provided entries in the prepositional paradigm of the creoles.166 In PLQ, we also find riba∼liba (< Sp. arriba; Schwegler & Green 2007: 302; Armin Schwegler p.c.), and Port. (ar)riba was the source for the postposition liba in Saramaccan (Muysken 2008: 199). Moreover, a derivative of Spanish arriba developed into a preposition in Cuban Bozal Spanish (Lipski 2005: 274). However, contrast is provided by Gulf of Guinea PC, where Port. (ar)riba was not integrated (cf. Fontes 2007; Maurer 2009a: 213–260). In PRI, for instance, (na) ixima means ‘on’, as in (na) ixima kama (in/on-)top-bed ‘on the bed’, with ixima ‘top’ < Port. (em) cima ‘(on) top’ (Maurer 2009a: 209, 251). 3.2.5.

PA / Upper Guinea PC pa

PA / Upper Guinea PC pa ‘to, for’ covers much of the semantic range that in Spanish and Portuguese is covered not only by para but also by por and a. As noted previously, the latter two produced no entries in PA and Upper Guinea PC. In various other Iberian-based creoles – although on the whole endowed with a less extensive prepositional system – a reflex of Spanish / Portuguese por has in fact been retained, as is the case, for example, in ST (plô; Rougé 2004: 236), PRI (pô; Maurer 2009a: 233), Korlai (por; Silva Rêgo 1998: 68) and PLQ (po; Schwegler & Green 2007: 302). Not only is PA / Upper Guinea PC pa semantically broader than Iberian para, it also serves as a complementizer after verbs of volition, where mainstream Spanish and Portuguese use que (cf. Quint 2000b: 141 for PA and SCV; Muller 1989: 495, 496 for PA; Doneux & Rougé 1988: 57 for GBC). Moreover, pa introduces finite (with an overt object) as well as non-finite subordinate clauses in both PA and Upper Guinea PC (Kouwenberg & Ramos-Michel 2007: 315; Baptista, Mello & Suzuki 2007: 62, 63). By way of contrast, the derivatives of Iberian para found in Korlai and PLQ do not introduce subordinate clauses, whether finite or non-finite (Clements 2007: 160; Schwegler & Green 2007: 284). In ST, pa is used as a complementizer, but unlike PA and Upper Guinea PC, speakers of ST make a distinction between the complementizer pa and the preposition pla (Kihm 1989: 371 drawing on Ferraz 1979: 80). Though several of the properties of PA / Upper Guinea PC pa (e.g. introducing subordinate clauses) can also be found in non-standard varieties of Spanish or 166 Although in SCV we find sobri, there is little doubt that this preposition is a relatively recent innovation (Quint 2000a: 205, 206). Note furthermore that in GBC, riba can also be a verb meaning ‘to return’ and in that role derives not from the Portuguese adverb arriba but from the maritime Portuguese verb arribar ‘to return’ (Rougé 2004a: 66).

102

Selected parts of speech

(Brazilian) Portuguese (cf. Lipski 2005: 298), the exact combination of functions described above for PA / Upper Guinea PC pa together with the lack of a porderivative might very well be unique to these two languages. But there is more to PA / Upper Guinea PC pa that is worth singling out. For instance, it can be at the head of a series of lexicalized spatial compound adverbs: PA (bai) paden / pafó / patras / pariba ‘(to go) inside / outside / upstairs’ (Dijkhoff 1993: 218) = Upper Guinea PC (bai) padentu / pafora / patras / pariba (see for instance Lang 2002 for SCV; Scantamburlo 2002 for GBC). Admittedly, adverbial constructions headed by pa may not be all that surprising, given that para fora / fuera de is also commonly heard in several (nonstandard) varieties of Portuguese and Spanish.167 It gains significance, however, when we note that in both PA and SCV, these adverbs headed by pa may still be preceded by di (with a subtle shift in meaning): (7) a. PA mi ta pasa di paden pa mi yega promé I IMP pass of (for) in(side) for I arrive first ‘I’ll take a shortcut so that I arrive first’ (van Putte & van Putte-de Wind 2005a: 323) b. PA mi ta pasa di pafó pa mi no shusha I IMP pass of (for) out(side) for I NEG dirty kas house ‘I’ll go (a)round so that I won’t dirty the house’ (van Putte-van Putte-de Wind 2005a: 323) c. SCV nen di pa déntu, nen di pa fora neither of for in(side) nor of for out(side) ‘neither from within, nor from without’ (Lang 2002: 204; cf. p. 25, 109, 204, 732) Iberian equivalents of (7a-c) are generally of the type Sp. por detrás, por fuera / Port. por trás, por fora, etc.168 The level of idiosyncrasy further increases with shared lexicalized adverbs composed of the noun parti∼pa ‘part’, the preposition di ‘of’ and a directional adverb (8), as well as related compound nouns built up of parti ‘part’ and a directional adverb (9). 167 Lipski (2008: 226–229) describes the use of patrás in the Spanish of the United States. Especially there were contact with English is evident, constructions of the type llamar patras ‘to call back’have evolved. Also in the São Vicente variety of CV, the construction txmá patras ‘to call back’ is increasingly used (Adira Ferreira p.c.). 168 See Kouwenberg (1990), Kouwenberg & Lefebvre (2007) and Lefebvre & Therrien (2007b) for analyses of the functions of PA pa in the generative tradition.

3.2. Prepositions

103

(8)

a. PA

Parti di pafó angua tin un kaska duru part of outside pipefish have a shell hard ‘from the outside the pipefish has a hard shell’ (Nagelkerken 1980: 18) b. SCV pa di pafora stába bunitu, má pa di padentu part of outside was pretty but part of inside stába tudu podri was all rotten ‘From the outside she was pretty, but the inside was all rotten’ (Mendes et al. 2002: 226)

(9)

a. PA

E porta su parti pafora ta tur kaská the door his part outside IMP all chipped ‘The outside of the door is all chipped’ (Henriquez 1988: 22) b. SCV párti-fóra di algum kusa part outside of some thing ‘the outside of something’ (Mendes et al. 2002: 211)

One may legitimately wonder if compounds such as the ones discussed here were in use in 17th-century Upper Guinea PC and thus whether they are meaningful in the context of establishing genetic links with PA. But perhaps more than anything else, these examples serve to show how similar PA and Upper Guinea PC operate morphosyntactically, in terms of both word order and word formation. In any case, Spanish and Portuguese can hardly be considered as models for the constructions exemplified in (7–9). 3.2.6.

Zero preposition with motion verb + place

One of the reasons why the Iberian preposition a was not integrated in PA and Upper Guinea PC is that in both creoles, verbs of movement occur without a preposition (or with a so-called zero preposition; see Holm & Patrick 2007: x, xi) when followed by a place indication: (10) a. PA Majan mi ta bai Ø Suriname tomorrow I IMP go Suriname ‘Tomorrow I’m going to Suriname’ (Lenz 1928: 175) b. GBC i ta bai Ø ospital he IMP go hospital ‘he goes to the hospital’ (Scantamburlo 2002: 141) c. SCV Sabadu nu ta bai Ø Cidadi Velha Saturday we IMP go Cidade Velha ‘On Saturday we (will) go to Cidade Velha’ (Fanha 1987: 301)

104

Selected parts of speech

In most varieties of Spanish and Portuguese either a or para would be employed.169 Zero prepositions with verbs of movement are quite common in Atlantic creoles and found also in Gulf of Guinea PC (Boretzky 1983: 195; Lorenzino 2007: 22; Holm 1988: 208). The demonstration of this feature does, however, allow one to contrast PA and Upper Guinea PC with the Spanish-based creoles PLQ and Chabacano, both of which require the presence of locative particles (prepositions or adverbs) where PA and Upper Guinea PC use a zero (Schwegler & Green 2007: 303 for PLQ; Lipski & Santoro 2007: 395 for Chabacano). 3.2.7.

Reanalysis of Iberian prepositions/adverbs ‘in front of’ and ‘behind’ as nouns

In PA and Upper Guinea PC, a seemingly idiosyncratic construction built on a possessive pronoun and one of the spatial prepositions/adverbs ‘before, in front of’ and ‘after, behind’ serves to indicate spatial relations: (11)

a. PA

E muchanan tin ku kore su tras The children have REL run his after ‘The children had to run after him’ (Allen 2007: 78) b. GBC Garandi multidon na bai si tras big multitude PROG go his after ‘A big crowd is going after him’ (Scantamburlo 2002: 417) c. SCV Tudu algen dretu ta kóre si tras all person right IMP run his after ‘All the right people run after him’ (Lang 2002: 692)

(12)

a. PA

henter mundoe ta bo dilanti entire world be your front ‘the whole world is in front of you’ (Jansen 1911: 21) b. GBC Kamiñu sta bu dianti road be your front ‘The road is ahead of you’ (Doneux & Rougé 1988: 19)

169 In modern PA, the use of prepositions in the zero slot between the movement verb and the place indication is increasing. This was already remarked upon in 1898 by Pijpers (1898: 22), who translated ‘I’m going home’ as PA Mi ta bai Ø kas, noting that Mi ta bai na cas is also heard. Convergence with Dutch naar ‘to(wards)’ plausibly accounts for the increasing use of PA na in this type of clauses (Kramer 2004: 179).

3.2. Prepositions

(13)

105

a. PA

korda di sera poorta bo tras remember of close door your behind ‘remember to close the door behind you’ (Putman 1849: 70) b. GBC n na bin na bo tras I PROG come in/to your behind ‘I am coming after you’ (Montenegro & de Morais 1995: 170)

To translate (11–13), speakers of mainstream Spanish and Portuguese would take recourse to prepositional phrases such as Sp./Port. tras de ti behind-of-you ‘behind you’ (Kihm 1994: 68; Lang 2009: 145, 145f.n.). In the construction exemplified here, PA tras/dilanti and Upper Guinea PC tras/dianti must clearly be analyzed as nouns. Kihm (1994: 68) described the phenomenon for GBC, noting that GBC diyanti and tras “are shown to be nouns by the fact that (…) we do not find, e.g., *diyanti di mi ‘in front of me’ or *tras di mi ‘behind me’, but nya diyanti and nya tras, i.e. literally ‘my front’ and ‘my behind’(…)” (cf. do Couto 2007: 495 and particularly do Couto 2009a: 290 for a similar analysis). It is noteworthy that, independently from Kihm, Birmingham (1970: 125) had described the same phenomenon for PA involving the exact same two items: “the prepositions dilanti ‘in front of’ and tras ‘behind’ may function as nouns”. To my knowledge, the feature is not dealt with in (recent) PA grammars170 (barring Birmingham 1970, cited above), although it is quite widely spread in PA and found frequently also in Early PA texts (cf. 12a). Note, however, that Boretzky (1983: 196, 197) and Holm (1988: 209, 210) have discussed the phenomenon for PA. Lang (2009: 135, 136, 144–146; cf. 2002: XXXVII) singled out the feature at issue (“l’emploi des adjectifs possessifs devant certaines prépositions”) for SCV and stressed that this construction is not found in the superstrate: “In the Romance languages, this procedure is found only at the other end of the Roman empire, in Romanian”171 (Lang 2009: 145). Indeed, to the best of my knowledge, neither Sp./Port. tras nor Sp./Port. delante/diante occur as nouns, be it in mainstream or non-standard varieties. Moreover, while the reanalysis of nouns as adverbs and/or prepositions is a common path of grammaticalization (e.g. Heine & Kuteva 2002: 49, 53, 63, etc.), the reverse path (adverb/preposition > 170 Van Putte & van Putte-de Wind (2005a: 101) do mention the nominal use of dilanti meaning ‘front (side)’, as in PA su dilanti a keda bunita his-front-PFV-AUX-lovely ‘its front side has become lovely’. 171 Original quote: “Dans les langues romanes, on ne trouve ce procédé qu’à l’autre bout de la Romania, en roumain”

106

Selected parts of speech

noun), as described above for PA and Upper Guinea PC, is rather untypical cross-linguistically, as far as I am aware. Lang seeks a substratal explanation arguing that the feature was inherited from Wolof. However, the fact that a similar feature has been attested in PRI (Boretzky 1983: 196; Holm 1988: 209) speaks against a Wolof origin, as Wolof is not part of the substrate of PRI. In PRI, we find phrases such as ubásu sé underhis ‘under him’ (Holm 1988: 209, drawing on Günther 1973: 80), but since the available sources on PRI only provide the noun phrase, not the predicate as a whole, the degree of similarity between the PRI construction and that exemplified in (11–13) remains unclear. Note, in any case, that PA and Upper Guinea PC depart crucially from Gulf of Guinea PC in terms of word order, with possessive pronouns following the noun in Gulf of Guinea PC, while preceding it in PA and Upper Guinea PC.172 3.2.8.

Composed prepositions

3.2.8.1. PA / Upper Guinea PC banda di To introduce this composed preposition, I cite Rougé (2004a: 74): “Both in Guinea and in Santiago, the locution banda di, which derives from an archaic Portuguese locution banda de, means ‘around’ and is used with both locative and temporal reference”173. Also in PA, the noun banda ‘edge, side’ (< Sp. / Port. banda idem) is at the basis of the composed preposition banda di ‘besides, next to, around’, which indeed also has both a spatial (14) and a temporal (15) use174 : (14)

di Barber banda di misa side of church of Barber ‘close to the church of Barber’ (Allen 2007: 189) b. SCV N ta mora banda di Tira Txapeu I IMP reside side of Tira Txapeu ‘I live next/close to Tira Txapeu’ (Rougé 2004a: 75).

a. PA

172 Carlos Fontes (p.c.), a native speaker of ST, finds that constructions of this type do not occur in ST and provided a translation of (11) according to the Iberian model: ST kolê tlaxi d-ê run-behind-of-3s ‘run after him’ (with ST tlaxi d-ê < Port. tras dele). 173 Original quote: “En Guinée (…) comme à Santiago, la locution banda di, qui vient d’une ancienne locution portugaise banda de, (…) signifie ‘aux environs vers’ et est employée aussi bien pour le lieu que pour le temps” 174 Both in PA and SCV the variant form pa banda di is quite common (Goilo 1974: 131; Lang 2002: XXXVII).

3.2. Prepositions

(15)

banda di ocho or side of eight hour ‘around eight o’clock’ b. FCV banda di cinc’ hor side of five hour ‘around five o’clock’ c. GBC banda di seti ora side of seven hour ‘around seven o’clock’

107

a. PA

(Allen 2007: 138)

(Parsons 1923: 42)

(Rougé 2004: 75)

The name of the Curaçaoan neighborhood Otrobanda, which lies at the ‘other side’ of the Punda district in Willemstad, shows the considerable time depth of the noun banda in PA and has a parallel in the GBC idiom utru banda ‘other side’ (e.g. Scantamburlo 2002: 66). The use of banda as the typical word for ‘side’ in PA and Upper Guinea PC contrasts with their lexifiers. Modern mainstream Portuguese has words such as lado, face and beira; modern mainstream Spanish takes recourse to nouns such as lado, borde and costado. The use of a word for ‘side’ as a locative adposition (typically ‘besides, next to’) is remarkably common among (Atlantic) creoles (Parkvall 2000: 108). The particular choice of both PA and Upper Guinea PC for the noun banda is nonetheless striking. As noted, banda is not used as a preposition in any modern variety of Spanish or Portuguese, but was typical of Old Portuguese roughly up to the 17th century (cf. examples of its use e.g. in Lemos Coelho 1669/1684[Peres 1953]). In modern mainstream Portuguese, on the other hand, locutions such as Port. ao lado de, perto de, sobre (temporal), por volta de, or arredor de fulfil the role of PA / Upper Guinea PC banda di. Modern mainstream Spanish takes recourse to composed prepositions such as al lado de, cerca de, sobre (temporal), alrededor de, etc. The archaic character of the etymon of PA / Upper Guinea PC banda di can be taken as additional evidence that the tendency towards the use of prepositions in PA and Upper Guinea PC is an early, not a recent, characteristic of their morphosyntax. No prepositional reflex of Sp. banda is attested in either PLQ or Chabacano, which is unsurprising, since, different from Portuguese, the noun never functioned as a preposition in Old Spanish in the first place, at least, as far as I am aware. In ST, the noun banda ‘side, place’ exists, but performs no prepositional functions (Fontes 2007: 55; Fontes & Holm 2008; Rougé 2004a: 75).175 175 In PRI, a reflex of banda is not found; instead, we find PRI uladu ‘side’ < Port. o lado (Maurer 2009a: 240). To express ‘next to’, PRI disposes of monomorphemic prepostions such as petu (< Port. pertu ‘close’), bwega (< ?), uku (< ?) or zuntu

108

Selected parts of speech

In Saramaccan, bándja ‘next to’ exists, but as a postposition. Moreover, a reflex of Port. de (cf. PA / Upper Guinea PC banda di) is not found in Saramaccan. In Papia Kristang, banda is the word for ‘side’ and has a grammatical usage as interrogative: ki banda what-side ‘where’, but Baxter (2004: 11) makes no mention of banda being used as preposition. In Korlai, banda is not attested; instead, Port. lado ‘side’ became lad, as in lad kadz ‘beside the house’ (Clements 1996: 144). In sum, the composed preposition PA / Upper Guinea PC banda di appears to be quite unique within the branch of Iberian-based creoles and creoles with an Iberian component. 3.2.8.2. PA for di, Upper Guinea PC fora di The preposition PA for di merits a more elaborate discussion, as it has been put forward as an argument for linguistic ties between PA and Gulf of Guinea PC (Maurer 2005). Different from its etymon, the Portuguese locative preposition fora de, PA for di has acquired the temporal meaning ‘since, ago’: PA for di dos luna out-of-twomonth ‘two months ago’. In Portuguese and Spanish, this function is covered by desde, in Upper Guinea PC by desdi: Port. desde dois meses, SCV désdi dos més. Maurer (2005; also 1995: 121) has shown that this temporal use is reminiscent of the Gulf of Guinea PC verb fô ‘to leave’176 , which at least in the variety of Annobon (AB) can take on a temporal function expressing ‘since, ago’: (16)

a. PA

b. AB

for di ayera out of yesterday ‘since yesterday’ (Maurer 2005: 59) fo dexi bi tokha osexi leave day+DEF come touch today ‘since that day until today’(Marike Post, p.c. in Maurer 2005: 61)

It should be pointed out, however, that the bimorphemic preposition PA for di ∼ fó di ∼ fo’i (< Port. fora de ‘out(side) of’ = Sp. fuera de) is first and foremost reminiscent of its Upper Guinea PC equivalent fora di: the PA and Upper Guinea PC preposition overlap not only in phonetic shape – the conservative PA ethnolect of the Sephardic Jews described by Henriquez (1988, 1991) still (< Port. junto ‘together’) (Maurer 2009a: 251), although these items are unlikely to be used very frequently. 176 Rougé (2004a: 164) speculates that convergence of Port. fora ‘out(side)’and Port. foi ‘he/she went’ might be at the basis of the Gulf of Guinea PC form.

3.2. Prepositions

109

knows the full form PA fora di (e.g. Henriquez 1988: 22) – but also in semantic content and syntactic behavior. From a syntactic point of view, an obvious but significant observation is that both PA for di and Upper Guinea PC fora di are true prepositions, whereas Gulf of Guinea PC fô is a verb. On a semantic level, it is interesting to observe that both PA for di (Maurer 2005: 58; Kriegel 2005: 93) and Upper Guinea PC fora di (e.g. Silva 2005: 241; Viaro 2005: 16; do Couto 2009a: 289) combine with verbs of movement unlike the etymon Port. fora di. Compare: (17)

a. PA

El a sali fo ’i laman he PFV leave out of sea ‘He left (/ got out of) the sea’ (Maurer 2005: 58) b. GBC no ta sai fora di matu we IMP leave out of forest ‘we leave the forest’ (Scantamburlo 2002: 412)

(18)

Port.

sair (*fora) do mar/bosque leave of sea/forest ‘to leave the sea/forest’

Note, furthermore, that PA for(a) di and Upper Guinea PC fora di can be preceded by pa, providing some remarkably similar entries: (19)

a. PA

M’ a topa kuné pafora di kas I PFV bump with+him outside of house ‘I bumped into him outside the house’ (Henriquez 1988: 22) a. FCV Tres gaselinh’ bem pa’ fóra di chique’r’ three gazelle come outside of pigsty ‘three gazelles came out of the pigsty’ (Parsons 1923: 20; cf. Mendes et al. 2002: 226)

It appears that PA for di only differs from Upper Guinea PC fora di in that the former has acquired a temporal value ‘since, ago’. But that is hardly remarkable: the semantic extension from [+spatial, −temporal] to [+spatial, +temporal] is referred to by Bybee et al. (1994: 25) as the “well-known TIME is SPACE metaphor”, characterized by “a metaphorical jump from the spatial domain to the temporal”. Similarly, Heine & Kuteva (2005: 242) speak of the “fairly well attested unidirectional development from spatial to temporal grammatical forms” (see also Bruyn 2009: 328). GBC provides a case in point, where the locative preposition tras di ‘behind’ (= PA / SCV tras di) underwent a temporal extension similar to PA for di: “[T]here is a temporal use of tras meaning ‘ago’,

110

Selected parts of speech

as in tras di tris mis ‘three months ago’” (Kihm 1994: 68). In other words, there is no need to ascribe the temporal extension of PA for di to contact with speakers of Gulf of Guinea PC, although that possibility is of course not disproven either. The crucial point is that, on the whole, PA for di resembles Upper Guinea PC fora di much more closely than Gulf of Guinea PC fô. Finally, it deserves to be stressed that PA desdi ‘since, ago’ (cf. Upper Guinea PC desdi idem) in fact still exists (van Putte & van Putte-de Windt 2005a: 92). Although, as far as I am aware, its diffusion and distribution alongside for di ‘since, ago’ is close to negligible in modern PA, this may once have been different: in various Early PA (in examples abbreviated as EPA) texts – e.g. Niewindt (1833: 3); Putman (1850: 6, 19; 1852: 12; 1853: 69); the 1803 Aruban letter (in Martinus 1996: 33, 34), Eybers (1916: 4, 494, 499) – regular use is made of PA desdi ‘since’ often instead of, or in variation with, for di: (20)

principi177

a. EPA

desdi

b. SCV

desdi prinsipi since beginning ‘from/since the beginning’

(Eybers 1916: 494, 1 John 2: 24) (Silva [ed.] 1987: 639)

One might be inclined to think that the use of PA desdi in these early texts reflects a Spanish bias on the part of the authors, but the fact that it was included as a preposition and synonym of PA for di in the PA grammars of Sintiago (1898)178 and van de Veen Zeppenfeldt (1928) speaks against this, and, moreover, suggests that desdi was still in use at the beginning of the 20th century. Interestingly, van de Veen Zeppenfeldt analyzed desdi as a contraction of desde di (1928: 72). A reduction from Early PA *desdi di > *des di is also suggested by the orthography found in other Early PA sources: ∼ . Also in GBC, we find multiple variants (des∼desdi∼dedi and desdi di) in free variation (Scantamburlo 2002: 105). The observations made thus far suggest that Early PA (like Upper Guinea PC) disposed of the spatial preposition for(a) di alongside the temporal preposition 177 Note here the change from -/io/ to /i/, a phonetic trait typical of both PA and Upper Guinea PC (see §2.1.4.2, Table 13). 178 Sintiago (1898: 30, 31) provided the following entry: “DESDE [for di, foi] ta determina tempoe i lugar, koe un cos ta socedé: ‘desde creacion di mundoe’, ‘for di (foi) Punda té coenucoe’; ta usa es preposicion aki també den es locucion adverbial: ‘desde awé’, ‘for di awor’” [‘DESDE (for di, foi) determines when something happens, also in the adverbial locution desde awé ‘since today’, for di awor ‘since now”]. This quite clearly suggests that at the time, desdi and for di were freely interchangeable.

3.2. Prepositions

111

desdi and that the former slowly but surely came to cover the temporal meaning previously assigned to the latter. 3.2.8.3. PA / Upper Guinea PC pa bia (di) Both in PA and in Upper Guinea PC, the Iberian composed preposition Port. por via de / Sp. por vía de ‘by way of, by means of’ was integrated as pa bia (di) with a meaning – ‘because (of)’ – that is subtly distinct from the etymon, while coinciding in the creoles (for PA, see Lenz 1928: 140; for GBC, see Scantamburlo 1981: 61, 1999: 184, 2002: 458; do Couto 1994: 112–114, 2007; Doneux & Rougé 1988: 56; Kihm 1994: 288; for SCV, see Schuchardt 1888: 315; Lang 2002: 529; Mendes et al. 2002: 392; Rougé 2004a: 284). (21)

a. PA

nan ta muri pa via di hende manera abo they IMP die for way of people manner you ‘they die because of people like you’ (Laclé, Rosenstand & Todd Dandaré 1983: 11) b. SCV pabia di Nhôr Dés by way of Lord God ‘because of God, the Lord’ (Brito [1887]1967: 382) c. GBC Dus pe serka kuatru pe pa bia di u˜ pe two foot approach four foot for way of one foot ‘Two feet approach four feet because of one foot’ (Bull 1989: 177)

For SCV, both Mendes et al. (2002: 392) and Lang (2002: 529) provide the preposition pa biâ (di) ‘because’ with the significant specifications “fundu” (‘deep’) and “crioulo muito fundo” (‘very basilectal/deep SCV’) respectively. Interestingly, Kihm (1994: 208) states the same for GBC, when commenting that GBC purkë ‘because’is “an acrolectal item whose ‘deep’variant is pabya”. In addition to being basilectal, pa bia (di) is an archaic feature in SCV (Rougé 2004a: 284). In PA, the conjunction appears to be archaic as well: it was mentioned by Lenz (1928) but no longer appears in modern dictionaries, which provide pa kousa di and/or pa motibu di instead. Richard Hooi179 (p.c.) assumes that pa bia di has largely disappeared from the vocabulary of modern mainstream spoken PA. Note, however, that was still actively used in written PA discourse at least up to the 1970s, as suggested by examples of its use in e.g. Martinus (a 1968 poem published in Munteanu 1991: 208), Lauffer (1971: 81), Debrot 179 Richard Hooi is a native speaker of PA and teaches linguistics at the University of the Netherlands Antilles.

112

Selected parts of speech

(1973: 1, 23), Henriquez (1988: 13, 81) and in contemporary texts gathered by Maurer (1988: 130, 376, 433).180 In contrast with PA and SCV, pabia (di) is quite frequent still in present-day GBC. I have no ready explanation for the fact that orthographic is found only in Lenz (1928: 140), whereas is the form written in the remaining PA sources. One cannot, however, take the orthographic as indication that pa via di is a recent innovation in PA: if it were, we would expect it to be a wholesale borrowing, i.e. something like PA *por via di, just as we find, for instance, PA porloménos ‘at least’ < Sp. por lo menos idem (van Putte & van Putte-de Windt 2005a: 357; cf. Birmingham 1978: 53, 54). Also suggestive of the originality of the entry discussed here is the aforementioned fact that its meaning, ‘because of’, slightly differs from its Sp./Port. etymon por via de/ por vía de ‘by way of ’.Two scholars of great importance to the historical study of PA and SCV show remarkable awareness of this semantic difference: Lenz (1928: 140) and Brito ([1887]1967: 382) independently provided identical entries: PA and SCV respectively, accompanied by the Spanish (Lenz) and Portuguese (Brito) translation por causa de ‘because of’, rather than por via de ‘by way of’. It is interesting to note, furthermore, that reflexes of Sp./Port. via as an independent noun are not attested in either Upper Guinea PC (Rougé 2004a: 284) or PA (van Putte & van Putte-de Windt 2005a). In Gulf of Guinea PC, Port. por via de provided cognates which are phonetically quite distant from the PA / Upper Guinea PC cognates: ST plô via ∼ plô vya ∼ pô vya, PRI pô vya (Rougé 2004a: 284; Maurer 2009b: 246). The Gulf of Guinea PC cognates do, however, have the same meaning (‘because (of)’) as the PA / Upper Guinea PC cognates, suggesting that Port. por via de may once have had that meaning as well. In PLQ, no derivative of Sp. por via de is attested. Note, finally, that particularly in GBC, pa bia (ku) frequently serves as a subordinate conjunction ‘because’ (22, 23) (Scantamburlo 2002: 458; 1999: 184; do Couto 1994: 112–114; Doneux & Rougé 1988: 56). In PA, we find quite comparable examples of pa via ku: (22)

a. PA

pa via ku kasi no tin hudiu aki for way that almost NEG have Jew here ‘because there are hardly any Jews here’ (Henriquez 1988: 77) b. GBC pabia abo e catchur grandi for way 2sg be dog big ‘because you are an absolute dog’ (Filho 2000: 11)

180 Future investigation must clarify whether (and, if so, in which domains) pa bia∼via (di) is still in use in modern PA.

3.2. Prepositions

(23)

113

a. PA

pa via qu é no tawatin reiz, el a seca for way that he NEG had root he PFV dry ‘because it didn’t have roots, it dried out’ (Eybers 1916: 78) b. GBC pabia ku es kusa ku Deus da-n, because REL DEF thing REL God give+me n na guarda ña ermons garandi I PROG protect my brothers big ‘because of what God gives me, I protect my big brothers’ (Montenegro & de Morais 1995: 174)

3.2.9. A reassessment of the time-depth of prepositions in PA As previously mentioned, the prepositional abundance seems to isolate PA and Upper Guinea PC typologically within the branch of Atlantic creoles. Especially for PA, this abundance has on occasion been analyzed as the result of processes of decreolization towards Spanish. Several arguments clearly disprove such claims. Firstly, in Table 28, two items of probable Portuguese origin have been identified (PA na < Port. na (em + a) = Sp. en la; PA te < Port. até = Sp. hasta), versus none of probable Spanish or Dutch origin. In addition, Portuguese provided the etyma for the composed prepositions PA for di and banda di. Since PA is in contact with Spanish, Dutch and English, but not with Portuguese (pace Endruschat 2004: 66), the relative dominance of Portuguese etyma can hardly be due to decreolization. Secondly, the prepositional behavior in Early PA documents does not differ quantitatively from that of modern PA. Synchronically, moreover, the heavy use of prepositions cuts across the dialectal continua and is typical of basilectal PA just as much as of acrolectal PA (an observation that also holds for Upper Guinea PC), which is a third argument. A fourth and final indication of the originality of PA’s core prepositional paradigm is that, as this section aimed to demonstrate, these core prepositions differ neatly from their Iberian etyma either in form, syntactic behavior or semantic content, while corresponding in these respects to their Upper Guinea PC equivalents. Again, this would be unexpected, had these prepositions been incorporated in the creole more recently due to contact with Spanish.

114 3.2.10.

Selected parts of speech

Final remarks on prepositions

As with the pronouns, it is not so much the correspondence between PA and Upper Guinea PC of single prepositions – some are, as mentioned, indeed also found in other creoles – as it is the overlap of the paradigm as a whole that merits our attention. This structural overlap (in terms of phonetic shape, semantic scope and syntactic behavior of each of the constituents) is, as far as I can see, not found between PA and any language other than Upper Guinea PC. It remains unclear why, with the exception of Martinus (1996) and Quint (2000b), this overlap has gone unnoticed in recent scholarship concerned with (the origins of) PA’s prepositions. For instance, ignoring the fact that the PA preposition pa has a direct cognate in Upper Guinea PC, Lefebvre & Therrien (2007b: 242) come to the conclusion that “if the label of Papiamentu pa most probably comes from Portuguese para through syllable truncation, its other properties generally come from those of the closest Fongbe lexical items, nú/ní”. Note that Fongbe (a Kwa language) could not have had any influence on the shaping of Upper Guinea PC pa, which, however, behaves identical to PA pa. Note, furthermore, that the predominance of Portuguese-derived prepositions clearly render untenable Munteanu’s (1996a: 387) affirmation that “the most frequently used prepositions in PA represent, in our opinion, a slightly simplified continuation of the prepositional system of Spanish”181 . Some prepositions of clear Spanish origin can no doubt be identified in PA, but these concern more peripheral items such as PA dilanti di (< Sp. delante de = Port. diante de) and PA bau di (< Sp. bajo de = Port. baixo de). Parallels can be drawn here with other intensive contact situations, such as the borrowing of peripheral prepositions from Spanish into Bolivian Quechua or from Italian into Maltese: also in these cases, the basic prepositions have remained unaffected and thus continue to clearly reflect the genetic lineage of the respective borrowing language (Muysken 2008: 179, 185).

3.3. Interrogatives The generally low borrowability rate and resultant cross-linguistic stability of interrogatives is received wisdom in contact linguistics (Muysken & Smith 1990; Matras 2003: 159; Muysken 2008: 90). Therefore, if PA and Upper Guinea PC have a common ancestor, we would expect to find commonalities in this domain 181 Original quote:“Las preposiciones más usadas en papiamento representan, en nuestra opinión, una continuación un tanto simplificada del sistema preposicional español”

3.3. Interrogatives

115

of the grammar. The most salient ones are listed in Table 29 and subsequently discussed. Table 29. Commonalities in the interrogative paradigm (cf. Quint 2000b: 141)182 PA ken (na) unda ki ki ora ki dia ki tempu

SCV

GBC

GLOSS

ken (na) unde ki ki ora ki dia ki tempu

ken ∼ nunde (na + unde) ki kal ora kal dia kal tempu kin183

‘who?’ ‘where?’ ‘what?’ ‘when?’ (lit. ‘what hour ∼ day ∼ time?’)

The question words for ‘how?’ and ‘why?’ are different in all three creoles: the forms most commonly used for ‘how?’ are SCV modi, GBC kuma and PA kon – but note the fossilized expression PA kumbai? ‘how are you?’ as well as the occurrence of as a comparative conjunction in Early PA texts (cf. §3.4.2.4) – and SCV pa modi, GBC purke and PA pa kiko for ‘why?’. Furthermore, besides the corresponding monomorphemic forms for ‘what?’ and ‘who?’ (Table 29), the three creoles have distinct bimorphemic forms: PA kiko184 , SCV kusé (< kusa ‘thing’ + e ‘to be’) and GBC ke ku (< ke ‘what?’ + ku [REL] (?)) for ‘what?’; PA kende (< ki + hende lit. ‘what person?’, with hende ‘person’ < Sp. gente ‘people’) and SCV kenha∼kenhi (< Port. quem +?) for ‘who?’.

182 The Upper Guinea PC question words are optionally accompanied by a relative pronoun (ki∼ke∼ku) (Baptista, Mello & Suzuki 2007: 78). In PA, this is rare for question words, but more common for relative pronouns (Kouwenberg & RamosMichel 2007: 328). 183 GBC ken is the form found in the more conservative dialects of Cacheu and Casamance, while kin is typical of the Guinea-Bissau variety (Wilson 1962: 35). 184 PA kiko ‘what?’ probably derives from ki + kos lit. ‘what thing?’, with PA kos ‘thing’ < Sp. cosa (Lenz 1928: 114). The compound ‘what thing?’ for ‘what?’ is quite common in creoles and attested also for instance in Gulf of Guinea PC kê kwa what-thing ‘what?’ (Rougé 2004a: 242). In various Early PA texts (e.g. Civilisadó 1872/23/11; Conradi 1844: 72, etc.), the form PA is found, making a Portuguese etymology (cf. Old Port. couza > Upper Guinea PC kusa) possible, with later phonetic alignment to Spanish cosa. The resemblence between Early PA and GBC ke ku ‘what?’ is interesting, but may be due to chance, as GBC ke ku can alternatively be analyzed as the composition of the interrogative ke ‘what?’ + relative marker ku.

116 3.3.1.

Selected parts of speech

Equally transparent interrogative paradigms

As Table 29 shows, for ‘when?’, only bimorphemic185 interrogatives are available in PA and Upper Guinea PC; the monomorphemic Iberian interrogative quando∼cuándo ‘when?’ was not integrated in either one of the creoles.186 It was, for instance, in PLQ, which has kuando for ‘when?’(Schwegler 1996: 722). Although bimorphemic question words for ‘when?’ are common to many (creole) languages, the same is true for bimorphemic question words for ‘where?’; to express the latter, however, both PA and Upper Guinea PC only have a monomorphemic question word (a derivative of Port. onde ‘where?’). These data lead to the interesting observation that, although both PA and Upper Guinea PC have a semi-transparent (mixed bi- and monomorphemic) interrogative paradigm, they are either transparent (bimorphemic ‘when?’) or non-transparent (monomorphemic ‘who?’ and ‘where?’) in exactly the same cases. 3.3.2.

PA: Portuguese rather than Spanish etyma

It is revealing to note that for both of PA’s monomorphemic question words, Portuguese, not Spanish, provided the etymon: PA ken < Port. quem (= Sp. quien), unda < Port. onde (= Sp. dónde187 ). In addition, the abstract noun PA tempu (in ki tempu? ‘when?’) is homophonous with Port. tempu, rather than with Sp. tiempo. Munteanu’s (1996a: 313) claim that the origin of PA’s interrogative paradigm is “indudablemente español” thus appears to be problematic at the very least. (The overlap of PA / Upper Guinea PC with the Portuguese-based creole Papia Kristang should not go unnoticed either: there we find keng ‘who?’, ki ora ‘when?’ and undi ‘where?’ (Baxter 2004).) 3.3.3.

Early PA *kantu, *kal

In a previous stage, the PA and Upper Guinea PC interrogative paradigms are likely to have corresponded even more closely. This can be deduced from the monophthong in the adverb PA kalke ‘any’ (= Upper Guinea PC kalker) which occurs in free variation with diphthongized PA kualke. By analogy, we may 185 See e.g. Bickerton (1981: 70), Holm (2000: 226), Parkvall (2000: 101; 2001: 199– 201), Lipski (2005: 260) or Baker (2009) for data on mono- versus bimorphemic question words in creoles. 186 In GBC, kuando exists, but is considered a neologism (Scantamburlo 2002: 308; Rougé 2004a: 241). 187 Roughly up to the 16th century, Spanish also had onde, which then fused with de, giving donde. However, for reasons explained in §2.2.1.2, 16th-century Spanish is an unlikely source for lexical material in PA.

3.3. Interrogatives

117

hypothesize the existence of Early PA forms *kantu ‘how much?’ and *kal ‘which?’ instead of the modern PA forms kuantu and kual. In Upper Guinea PC, indeed, we find kantu and kal. 3.3.4.

PA unda, SCV unde and GBC nunde

It is worth recalling the stressed /u/ instead of etymological /o/ in PA unda and Upper Guinea PC (n)unde and the contrast with ST/ANG andji, PLQ a(d)onde and Chabacano donde (cf. §2.1.1.2).188 Consider also the striking compositional resemblance between PA na unda, SCV na unde and GBC nunde (all ‘where?’) and PA pa unda and SCV punde (both ‘where to, which way?’) (Kouwenberg & Ramos-Michel 2007: 327; Lang 2002: 611). Such lexicalized question words are not attested in any other Iberianbased creole that I know of, although Papia Kristang comes close with per undi ‘which way?’ (Baxter 2004: 89). Note, finally, that several 19th-century authors writing in PA consistently used the form (e.g. Putman 1849 and particularly Niewindt 1837), not yet showing the lowering of the etymological unstressed /e/ to /a/ and thus further supporting the idea of a historical link between PA unda and Upper Guinea PC unde. 3.3.5.

PA / Upper Guinea PC ken

The homophony between PA and Upper Guinea PC ken is significant in view of the fact that, within interrogative paradigms, particularly ‘who’ is known for its stability (Matras 2003: 159). It gains importance if we consider that no reflex of Port. quem ‘who’ is attested in Gulf of Guinea PC (Rougé 2004a: 243). Instead, a reflex of Port. ninguem ‘nobody’ was integrated there with the meaning of ‘person’, which then grammaticalized into interrogative ‘who?’ (cf. Maurer 2009a: 146; Lorenzino 2007: 21). In PLQ, we find kiene < Sp. quién (Schwegler & Green 2007: 301).

188 PRI has kumin ‘where?’ from Port. caminho ‘road’ (Maurer 2009a: 147). Similarly, we find PA kaminda ‘where’ from Sp. camino ‘road’, though only as a relative pronoun.

118

Selected parts of speech

3.4. Conjunctions This section highlights corresponding conjunctions, both coordinate (§3.4.1) and subordinate (§3.4.2), as both types reveal salient similarities between PA and Upper Guinea PC in the domain of clause combining. 3.4.1.

Coordinate conjunctions

3.4.1.1. ‘with’ = ‘and’ Both PA and Upper Guinea PC make use of two distinct coordinate conjunctions, i and ku. The first coordinates independent clauses and verb phrases189 , the latter is employed in case of grammatical equality of the joined clauses. This section focuses solely on the latter, the comitative preposition PA / Upper Guinea PC ku ‘with’, in its function as a coordinate conjunction ‘and’ (cf. Richardson 1977 for PA; Baptista, Mello & Suzuki 2007: 60 for Upper Guinea PC; Quint 2000b: 140 for both PA and SCV): (24)

a. PA

Ami ku Dolly a keda na kas I with Dolly PFV stay in house ‘I and Dolly stayed home’ (Maurer 1988: 374) b. GBC ami ku el i ka kazadu I with he and NEG married ‘I and he are not married’ (Scantamburlo 2002: 271) c. SCV Lobu ku Xibinhu mister Wolf with Tubinhu ‘Wolf and Little Goat’190 (Lang 2002: 271)

(25)

a. PA

Nos ta kanta ku balia we IMP sing with dance ‘We sing and we dance’ (Kouwenberg & Ramos-Michel 2007: 328)

189 Quint (2008a: 44), however, provides field data showing that in basilectal SCV the combining of predicates is largely done without the coordinating conjunction i. Interestingly, also in the recordings of basilectal and/or elderly speakers of PA, such as transcribed by Allen (2007), the relative paucity of coordinating conjunctions draws the attention. 190 These are the two chief characters of a famous Cape Verdean fairy tale as chronicled, for instance, by Lima (2000).

3.4. Conjunctions

b. GBC nha ome sibi ku disi my man climb with descend ‘my man/husband went up and down’ (26)

119

(Peck 1988: 450)

a. PA

e kachó ta blanku ku pretu the dog be white with black ‘the dog is white and black’ (Richardson 1977: 59; cf. Lefebvre & Therrien 2007c: 173; Maurer 1988: 42) b. GBC Taxis di kor pretu ku verdi taxi.PL of color black with green ‘Taxis of the colour black and green’(Scantamburlo 2002: 298)

Languages with the feature ‘with’ = ‘and’ have been dubbed WITH-languages (e.g. Stassen 2005: 258). PA and Upper Guinea PC thus belong to the family of WITH-languages. This is not remarkable per se, as this family is sizeable (Heine & Kuteva 2002: 80–83191 ; Stassen 2008) and includes a significant number of (Atlantic) creoles (Holm 1988: 206; Parkvall 2001: 199; Holm & Patrick 2007: feature 19). However, none of the lexifiers of PA and Upper Guinea PC is a WITH-language, and neither are, for instance, the Iberian-based creoles Chabacano and Korlai (Lipski & Santoro 2007: 395; Clements 2007: 170). Moreover, some subtle characteristics seem to set PA and Upper Guinea PC apart from other Iberian-based WITH-creoles. In PLQ, for instance, ku ‘with’ conjoins only elements that are [+human], as in PLQ ele ku yo ‘he and I’ (Schwegler & Green 2007: 302). This implies that the use of ku in phrases (25, 26) is considered ungrammatical in PLQ. Also, in ANG, conjunctive ki cannot conjoin two adjectives (Maurer 1995: 65), so that (26) would be ungrammatical in that creole.192 191 Heine & Kuteva (2002: 81, 82) provided evidence from Chinese in support of the directionality (‘with’ > ‘and’) of this grammaticalization path. Evidence from the creoles of course unambiguously confirms this, as it is (to my knowledge without exception) the superstrate’s comitative preposition that develops into a conjoining conjunction in the creoles, never the other way around. This example once more illustrates that creoles do indeed provide a significant window on grammaticalization (cf. e.g. Bakker 2008b: 46). 192 Although I lack a sound explanation, it is interesting to note that both in PA and in Upper Guinea PC, ku can conjoin spatial prepositional phrases, but not prepositional phrases with purposive pa ‘for’, which must be conjoined by i instead. Thus, PA na kas ku na kaya in-house-with-in-street ‘in the house and on the street’(Maurer 1998: 185) and SCV na kasa ku na rua ‘idem’ are generally accepted, while, for instance, PA *pa Wan ku pa Maria for-John-with-for-Maria or SCV *pa Djom ku pa Maria

120

Selected parts of speech

Quint (2000b), finally, revealed an interesting idiomatic similarity: (27)

a. PA

nan

ta

biba

nan

b. SCV

es ta bibi es ku they IMP live they with ‘They live among themselves’

ku

nan es they (Quint 2000b: 157)

3.4.1.2. PA ma, o, si = Upper Guinea PC ma, o, si The homophonicity between the conjunctions ma ‘but’, o ‘or’and si ‘if ’in PA and Upper Guinea PC may have limited evidential power, considering that, firstly, the etyma (Sp./Port. mas, o∼ou and si∼se) are rather similar in Spanish and Portuguese and, secondly, coordinate conjunctions are universally quite easily borrowed (Matras 2009: 194). They do, however, add to the image of overall correspondence of clause combining in PA and Upper Guinea PC, and they contrast phonetically with, for instance, the Gulf of Guinea PC forms maji ∼ madji ∼ meji ‘but’(< Port. mas) and xi ‘if’(< Port. se) (Rougé 2004a: 201, 258). In modern PA, pero ‘but’ (< Sp. pero) and òf ‘or’ (< Du. of ) are gaining ground, but ma and o clearly predominate in 18th- and 19th-century Early PA texts. 3.4.2.

Subordinate conjunctions

Subordinate conjunctions are considered to be among the least borrowable items (as are interrogatives, to which they are often related) (cf. Muysken 2008: 177). Data from Spanish borrowings in Quechua, for instance, had led Muysken (1981) to classify subordinating conjunctions as the least borrowable parts of speech, topping even pronouns. And if “TMA and the subordinating systems provide the basic skeleton for a language” (Muysken 2008: 209), then a comparison of the category of subordinate conjunctions is of obvious significance to establishing a possible genetic relationship between PA and Upper Guinea PC. Below, the most salient correspondences are pointed out, including data from Early PA texts.

‘idem’ are considered ungrammatical (cf. Lefebvre & Therrien 2007a: 6). For PA, Lefebvre & Therrien (2007a: 6) claim that, “as a conjunction, ku relates only phrases that have a nominal component”. However, this does not explain the grammaticality of prepositional phrases headed by the preposition PA / Upper Guinea PC na, which, as far as I am aware, has no nominal component.

3.4. Conjunctions

121

3.4.2.1. ‘hour’ = ‘when’ In both PA and Upper Guinea PC, ‘hour’ equals prospective ‘when’: PA ora (ku), GBC ora ku ∼ ora ki ∼ oki and SCV óras ki ∼ ora ki ∼ óki (van Putte & van Putte-de Windt 2005a: 217; Kihm 1994: 205; Mendes et al. 2002: 337; Lopes da Silva 1957: 156): (28)

a. PA

ora ku e kiè tene su kabei ariba hour that she want keep her hair up ‘when she wants to keep her hair up’ (Kouwenberg & Murray 1994: 43) b. SCV Ora ki m da grandi (…) hour that I give big ‘When I grow up (…)’ (Mendes et al. 2002: 337) c. GBC Ora ku bo sinti nha tcheru hour that you feel my smell ‘When you feel my smell’ (do Couto 2009b: 53, 54)

In similar contexts, speakers of mainstream Spanish and Portuguese take recourse to cuando/quando. Of course, one finds strings such as Sp./Port. la hora que/a hora que ‘when, as soon as’ conjoining phrases, but few would argue that these strings are highly grammaticalized conjunctions, which PA / Upper Guinea PC ora (ki/ku) clearly is. In PA, the degree of grammaticalization is shown by the fact that ora frequently appears without the complementizer ku, as do other highly grammaticalized conjunctions such as PA promé (ku) or manera (ku) (discussed further below); for Upper Guinea PC, the degree of grammaticalization is revealed by contracted forms such as SCV óki, GBC oki. A difference between PA and Upper Guinea PC lies in the fact that Upper Guinea PC ora (ki/ku) is used only as a prospective conjunction (Nicolas Quint, p.c.), i.e. to denote events that lie in the future with respect to speech time (Kihm 1994: 205). If the reference time is past, SCV takes recourse to kantu ∼ kandu (< Port. quando) ‘when’ (Lang 2002: 291), while speakers of GBC may choose between kontra (ku) (related either to the verb kontra ‘meet’ or to the preposition kontra ‘against’) and oca (ku) (possibly related to the GBC verb oca ‘find’; Kihm 1994: 205). As far as I am aware, PA does not make such a distinction.193

193 Van Putte & van Putte-de Windt (2005a: 317, 318), however, seem to suggest that monomorphemic PA ora can be used with both prospective and past reference and that bimorphemic PA ora ku only occurs in past contexts. Examples such as (28a) seem to contradict this, but more research is needed to clarify this.

122

Selected parts of speech

The use of ‘hour’ as ‘when’ is not found in either PLQ or Chabacano; both instead take recourse to derivatives of Spanish cuando ‘when’ (Schwegler 1996: viii; Schwegler & Green 2007: 285; Lipski & Santoro 2007: 383). On the other hand, the feature is characteristic not only of PA and Upper Guinea PC but of all Afro-Portuguese creoles: ST ola ku ‘when’ (Fontes 2007: 74), PRI ora (ki) (Maurer 2009a: 255) and ANG ola kutxi ‘when’ (Maurer 1995: 64) are all variants of ‘hour + that’. Thus, PA again clearly aligns with the Portugueserather than with the Spanish-based creoles.194 Although the feature ‘hour’ = ‘when’ is thus found also in Gulf of Guinea PC, the following strings, in which PA / Upper Guinea PC ora (ki/ku) is preceded by the prepositions pa (29) and na (30) with subtle semantic effects, are not: (29)

a. PA

[e kouchi] ta sirbi pa ora ku e galiñanan the cage IMP serve for hour that the chicken+PL tin webu have egg ‘the cage serves for when the chickens have eggs’ (Kleinmoedig-Eustatia 1982: 17) b. GBC N sobra restu pa ora ku n riba I keep rest for hour that I return ‘I kept the rest for when I return’ (Kihm 1994: 205) c. SCV Marâ dretu, pa óra ki béntu ben tie up right for hour that wind come ‘Tie it up properly, for when the wind comes’ (Silva 1987: 548)

(30)

a. PA

na ora ku el a hisa su kara na lária, in hour that he PFV lift his face in air tabata un lareina be+PST IND queen ‘At the moment that he looked up in the air, there was a queen’ (Lenz 1928: 264) b. GBC na ora ki bô papia nos obi in hour that you speak we hear ‘when you spoke, we listened’195

194 The grammaticalization of ‘hour’ as ‘when’ took place also, for instance, in Haitian French Creole (lè < Fr. l’heure ‘the hour’; Kihm 1994: 205). A survey of languages with the feature ‘hour’ = ‘when’ is provided by Heine & Kuteva (2002: 176). 195 source: http://pensaodbertabissau.wordpress.com/amigos-da-avo/

3.4. Conjunctions

123

c. SCV na ora ki barku ta ben tuma-l, el po-l (…) in hour that boat IMP come take+3sg he put+3sg dentru-l bolsu in+3sg bag ‘when the boat came to pick him up, he put it [his handkerchief] in the bag’ (Silva 1987: 181) More frequent still is the combination PA / Upper Guinea PC te ora (ki/ku) (31) (cf. Lang 2002: 520; Kihm 1994: 206, 207; Maurer 1988a: 223): (31)

a. PA

the ora Ky boso abiny until hour that you PFV+come ‘until you came’ (1775 letter, in Salomon 1982: 370) b. SCV ti ora k’ e kebra until hour that it break ‘until it broke’ (Lang 2002: 520) c. GBC te ora ki alguin misti kusinhal until hour that someone want cook+3sg ‘until someone wants to cook it’ (Scantamburlo 2002: 538)

Though Spanish hasta la hora que and Portuguese até a hora que occur, modern mainstream varieties of Spanish and Portuguese would normally take recourse to the conjunctions hasta que/até que (lit. ‘until that’) to express (31). Unsurprisingly, both in PA and in Upper Guinea PC the variant ‘until + that’ is attested, but ‘until + hour + that’ is clearly the ‘deeper’ variant (cf. Kihm 1994: 206).196 As for the idiosyncrasy of ‘until + hour + that’ among Iberian-based creoles: I found something similar only in ST, which has antê ku ola until-that-hour ‘until’197 , which is analytically comparable to PA / Upper Guinea PC te ora (ki/ku). However, they clearly differ in terms of phonetic shape as well as word order. Consider, finally, the conjunctions of the type ‘until + in + hour (+ that)’ in (32), which once more remind us of how identically processes of word ordering and compounding function in PA and Upper Guinea PC and how similarly prepositions are put to use to create new lexical and syntactic elements: 196 The string ‘until + hour + that’ in Upper Guinea PC is optionally realized in phonetically reduced forms such as SCV tóki (Lang 2002: 790) and GBC tok(i) (Kihm 1994: 206 for GBC), clearly suggesting a high degree of integration. 197 The variant ST antê (ku) until-that < Port. até que ‘until’ is more common in ST (Tjerk Hagemeijer p.c.), as in ST antê (ku) ê bi da tudu kwa se di 1953 until-3sgcome-give-all-thing-SP-of-1953 ‘until all those things of 1953 occurred’ (with se = SP = Specific marker) (Hagemeijer 2007: 184).

124

Selected parts of speech

(32)

a. PA

te na ora (cu) é bini until in hour (that) he come ‘even if he comes’ (Goilo [1951]1975: 71) b. GBC te na ora ki n na sai until in hour that 1sg PROG leave ‘until I am leaving’ (Montenegro & de Morais 1995: 159) c. SCV te na ora di bem until in hour of good ‘until the very last moment’ (Mendes et al. 2002: 236)

3.4.2.2. ‘first’ = ‘before’ In both PA and Upper Guinea PC, ‘first’ equals ‘before’ (33, 34). From a phonological point of view, it is important to recall (cf. §2.1.2) that Brito (1887) still used the Early SCV form , a variant no longer attested in modern dictionaries. The resemblance between that Early SCV form and PA promé (still in Early PA texts, cf. ex. [33a, 34a]) is evident (cf. Martinus 1996: 151).198 (33)

a. PA

A pasa basta anja promer el por a drenta PFV pass many year first he could PFV enter ‘Many years passed before he could enter’ (La Union 1889, 19–04) b. SCV djobi káru purmeru bu kunsa pása ! watch car first you commence pass ‘watch the cars before you cross!’ (Quint 2000a: 271)

(34)

a. PA

Promer cu Shon Arey por a habri su boca first that Shon Arey can PFV open his mouth ‘Before Shon Arey could open his mouth’ (Baart 1983: 92) b. SCV Purmeru ku bo na ba fala manteña first that 2sg PROG go speak greeting ‘Before you are going to say the greeting’ (Montenegro & de Morais 1995: 95)

In present-day Portuguese and Spanish, primeiro (que) and primero (que) can also occur as conjunctive ‘before’, but this appears to be rare. Rather, antes 198 Martinus (1996: 151) established the same link, though on the basis of the CV form promero documented by Cardoso (1983). Unfortunately, Cardoso (1983), which is in fact a reissue of Capeverdean folklore collected as early as 1933, is not at my disposal.

3.4. Conjunctions

125

de (que) is usually preferred in both languages. Cross-linguistically, the feature ‘first’ = ‘before’ is not uncommon (see examples in Heine & Kuteva 2002: 138). While it is not attested in PLQ (Armin Schwegler p.c.), something similar appears to be found in ST (35): plumê zo pa first-then-for ‘before’: (35)

ST

Zon na kume plumê zo pa bêbê fa Zon NEG eat first then for drink NEG ‘Zon didn’t eat before drinking’ (Hagemeijer 2007: 80)

3.4.2.3. ‘manner’ = ‘how’ In both PA and Upper Guinea PC, the equivalent of ‘manner’ serves as the conjunctive comparison marker ‘how, like, as’ (36).199 Within Upper Guinea PC, this conjunctive use of manera is particularly patent in GBC. In SCV, on the other hand, modi (ki)∼mo (ki) and sima∼suma200 are used instead (cf. Mendes et al. 2002: 134). In GBC, just as in PA, manera can receive a nominal complement (37). To translate examples (36–37), mainstream Spanish and Portuguese would use conjunctive como ‘how’. (36)

(37)

a. PA

Manera hende ta mira manner people IMP watch ‘How people watch (…)’ b. GBC manera ku omi ta bibi manner that man IMP drink ‘how men drink’

a. PA

(Lauffer 1971: 125)

(Kihm 1994: 114)

Pa hende manera abo no tin lugá den es for people manner you NEG have place in the mundo aki world here ‘For people like you there is no place in this world’ (Debrot 1986: 417)

199 Extension of the noun ‘manner’ as conjunctive ‘how’ is described for Thai and Kenya Pidgin Swahili by Heine & Kuteva (2002: 212). They add that “More crosslinguistic data is required to substantiate this process, including its directionality” (2002: 212). PA and Upper Guinea PC appear to provide such cross-linguistic data, exemplifying how the noun ‘manner’ developed into the conjunction ‘as, if, like’, not the other way around. 200 In GBC, sima ∼ suma is also used as conjunctive ‘how’ (Scantamburlo 2002: 550). The etymon is likely to be the Old Portuguese conjunction assim coma (Rougé 2004a: 114).

126

Selected parts of speech

b. GBC Lubu manera ki cobarde, i bai tras di kabra hyena manner that coward he go behind of goat ‘the hyena, like a coward, (he) went after the goat’ (Scantamburlo 2002: 166; cf. 2002: 395) Interestingly, in the Barlavento varieties of CV, Port. maneira developed into the interrogative manera (ki) ∼ mané (ki) rather than into a comparative conjunction as it did in PA and GBC. Still, the parallel between the use of BaCV manera in (38) and PA / GBC manera (36–37) is obvious: (38)

a. BaCV Manera k bo ta? Manner that you be ‘How are you?’ (Dominika Swolkien, p.c.) b. BaCV Manê k povo ta fazê pá vive? manner that people IMP do for live ‘How do people manage to live?’ (Octavio Fontes, p.c.)

By way of contrast, note that there are no reflexes of Port. maneira attested in Gulf of Guinea PC (Rougé 2004a: 198), nor derivatives of Sp. manera in PLQ (Schwegler 1996: x), which has conjunctive kumo ‘how’ instead (Schwegler & Green 2007: 291). 3.4.2.4. Early PA / SCV modi As noted above, where PA uses conjunctive manera, SCV often takes recourse to the comparative conjunction modi (ki). An analysis of 19th-century PA texts, however, suggests that modi was once also part of the PA morphosyntax as a variant of manera (39–40). In fact, Schuchardt (1882: 140) still provided the conjunction PA mo di ‘as’ as well as the noun PA modi ‘way’. Both are no longer found in modern-day PA dictionaries. Again, to translate the following examples, Spanish and Portuguese would use como ‘as, like, how’: (39)

a. PA

Jioe di heende ta bai moeri, modi ta para Child of people IMP go die, manner IMP stop skierbier di eel written of he ‘The child of the people will die, as (it) has been written about him’ (Putman 1852: 25)

3.4. Conjunctions

127

b. SCV pa mostra modi es ta prununsiadu to show manner they IMP pronounced ‘to show how they are pronounced’201 (40)

a. PA

modi Bo manner you b. SCV mode nhu manner you ‘as you wish’

kiëer (Putman 1852: 26) want cré (Coelho 1880: 5) want

At some stage in Early PA, the comparative conjunctions modi and manera appear to have been used interchangeably. This is clearly suggested by parallel examples such as PA blanku modi sneu white-manner-snow ‘as white as snow’ (Conradi 1844: 73) and PA blanku manera sneeu white-manner-snow ‘as white as snow’ (van Dissel 1865: 33) or PA modi un wantomba manner-a-ghost ‘like a ghost’ (Frederiks & Putman 1959: 140) and PA manéra oen pieda paloe manner-a-stone-pole ‘like a streetlamp’ (1959: 141). By the early 20th century, PA modi seems to have disappeared fully. This is at least suggested by the comparison of Conradi’s (1844) Gospel of Matthew with that of Eybers (1916): in 1844, modi was used a total of six times; in 1916, modi appears not once. Three occurrences of Early PA modi in the 1844 Gospel of Matthew were realized in 1916 by PA meskos ku () same-thing-that ‘the same as, just as’ (which still exists in PA, alongside, but less frequent than, manera); two occurrences were substituted by PA manera; one by PA segun. With regard to the etymon of modi, it is noted that fluctuations between /u/ and /i/ are quite common in rural SCV, e.g. sapáti < Port. sapato ‘shoe’, as well as in the conservative Casamance dialect of GBC, e.g. medi (< Port. medo ‘fear’) or sedi (< Port. cedo ‘early’), etc. (Nicolas Quint, p.c.). Consequently, Port. modo ‘manner, way’ is a plausible etymon for modi (Quint 2009: 264).202 201 source: http://portugesnacrioulo.blogspot.com/2007/11/daki-nada.html 202 The noun moda ‘fashion, manner, way’is used as a conjunction in both PA (van Putte & van Putte-de Windt 2005a: 295) and SCV. Though the time-depth and originality of this conjunction is difficult to assess, it is a fact that the Spanish noun moda ‘fashion’ has no conjunctive functions, making a Portuguese etymon for PA moda more likely. For SCV moda, Quint (to appear: 264) provides Port. à moda de ‘like, as’ as the etymon, which is thus also the most plausible etymon for PA moda. PA moda is used in idiomatic expressions such as na moda di papia in-manner-of-speak ‘as a figure of speech’, or no tin moda di hasi NEG-have-manner-of-do ‘there is nothing to do about it’ and as a conjunction in (na) moda ku ‘so that’ (van Putte & van Putte-de Windt 2005a: 295): PA moda ku nan por aparesé den bida públiko atrobe fashion-that-they-can-appear-in-life-public-again ‘so that they can appear in

128

Selected parts of speech

The Portuguese noun modo was integrated with comparative qualities not only into PA and Upper Guinea PC, but also into Gulf of Guinea PC: PRI mo∼modi (ki), ST mo∼modu (ku) (Schuchardt 1882: 140). Rougé (2004a: 208) proposes convergence with Portuguese como ‘as, like, how’ to account for this feature. 3.4.2.5. Early PA relative ki = CV ki A superficial reading of Early PA texts puts beyond doubt that the primary Early PA relative marker was not ku, as it is now, but ki, as it still is in CV. Lenz (1928: 115) already observed this and drew a parallel: “It is possible that relative ke or ki was initially used on Curaçao, as on the Cape Verde Islands. Some old texts use ki instead of ku”203 (see also Martinus 1996: 36). Thus, the use of ki in (31a) provided previously or in (41) provided below would be ungrammatical in present-day PA, where ku is used instead. (41)

a. EPA

Bo marido ki tanto ta stima bo Your husband who so much IMP love you ‘Your husband, who loves you so much’ (1775 letter, in Maurer 1998: 204) b. EPA Na tempo ki lo yega in time that FUT arrive ‘In the time ahead’ (1803 letter, in Martinus 1996: 33)

The knowledge that the PA relativizer ku used to be ki strengthens the hypothesized historical link between modern conjunctions such as PA ora ku and Upper Guinea PC ora ki, which must have been homophonous (*ora ki) in an earlier stage.

public life again’ (Boletin di Notisia, 29-1-2010). In SCV, conjunctive moda means ‘as, like’, rather than ‘so that’: SCV Ti inda sa ta uza móda ki uzadu na sékulu pasádu until-still-PROG+IMP-use-fashion-that-used-in-century-past ‘It is still used as it was in the previous century’ (Lang 2002: 837). In Barlavento, moda (cu/se) can function as a subordinate conjunction, as shown by the following examples from the São Vicente variety: SVCV Môda cu el ca tava ta bem fashion-that-I-PFV+PSTIMP-come ‘It’s as though he wouldn’t be coming’ (Octavio Fontes, p.c.); SVCV Moda se el ka ti ta bem fashion-if-he-not-PST+IMP-come ‘It’s as though he’s not coming’ (Adira Ferreira, p.c.). 203 Original quote: “Es posible que también en Curazao se haya usado primitivamente un relativo ke o ki, como en las islas de Cabo Verde. Algunos textos antiguos (…) usan ki por ku”

3.5. Miscellaneous

3.4.3.

129

Final remarks on conjunctions

Table 30 shows the overlap between PA and Upper Guinea PC in the category of coordinate and subordinate conjunctions: Table 30. Corresponding coordinate and subordinate conjunctions204 Etymon

PA

Upper Guinea PC

Gloss/function

Sp./Port. con/com Sp./Port. mas, o/ou, si/se Sp./Port. ora + que –

ku ma, o, si

ku ma, o, si

‘with’ = ‘and’ ‘but, or, if’

ora (ku) pa/na/te/te na + ora (ku) promé (ku)

‘hour’ = ‘when’ Discussed in §3.4.2.1 ‘first’ = ‘before’

manera (ku) modi (Early PA)

ora ki/ku pa/na/te/te na + ora ki/ku promeru (Early SCV) ∼prumeru∼purmeru (ku) GBC manera (ki/ku) SCV modi

pa bia ku

GBC pabia (ku/ki)

ki (Early PA) ku (Modern PA) pa

SCV ki GBC ki∼ke∼ku pa

‘because’ (§3.2.8.3.) relative pronoun

Sp./Port. primero/primeiro Sp./Port. manera/maneira, modo de (?) Sp./Port. por via Sp./Port. que Sp./Port. para

‘manner’= ‘how’

complementizer (§3.2.5)

3.5. Miscellaneous 3.5.1.

Reciprocity and reflexivity

In creole languages, the reciproque and reflexive pronouns seldom represent replications from the lexifier (Muysken & Smith 1994: 45). Correspondences between PA and Upper Guinea PC in these categories are therefore unlikely to be due to a shared lexifier language and might instead be revealing of deeper historical ties.

204 Not discussed in this section is the complementizer ‘that’ following verbs of saying and thinking, which is ku in PA, kuma in GBC and kuma∼ma in CV.

130

Selected parts of speech

3.5.1.1. Reciprocal ‘other’ Reciprocity in PA and GBC is expressed with a lexeme meaning ‘other’(Dijkhoff 2000: 87; Scantamburlo 1999: 167, 168), as in (42).205 (42)

a. PA

Nos no ta skirbi otro mas we NEG IMP write other more ‘We do not write each other anymore’ (Goilo [1951]1975: 110) b. GBC no na oja un utru we FUT see a other ‘we’ll see each other’ (Kihm 1994: 168)

Reciprocal ‘other’ also characterizes Gulf of Guinea PC (Maurer 2009a: 153), but here the word for ‘other’ appears obligatorily both in subject and object position (43): (43)

PRI

ôtô sa faa ôtô bê other PROG tell other hello ‘they were greeting each other’

(Maurer 2009a: 153)

3.5.1.2. Reflexive ‘body’ One of several ways to reflexivize in PA and Upper Guinea PC is by means of the equivalent of ‘body’ (PA kurpa, GBC kurpu, SCV korpu). While this feature is definitely absent from the lexifiers Spanish and Portuguese (Muysken 1993: 298; Kramer 2004: 186), it is in fact quite common among creoles206 and noncreoles (Holm 1988: 205; Parkvall 2000: 57–60 & 2001: 199–201; Muysken & Smith 1994: 47; Heine 2005: 216; Heine & Song 2010: 127–128). It is nonetheless important to discuss the feature, as its presence in PA is often thought to correlate with a Kwa/Bantu substrate, while in fact the overlap between PA and particularly GBC provides quite straightforward clues for a historical link with Upper Guinea. This overlap concerns, first and foremost, the lexical shape of the reflexive pronoun (PA kurpa, GBC kurpu). The raising of the stressed etymological /o/ to /u/ in both GBC kurpu and PA kurpa (< Port. corpo) phonologically sets these forms apart from, for instance, ST klôpo (cf. §2.1.1.2). It must furthermore be borne in mind that the lowering of the unstressed etymological /o/ to /a/ in PA 205 Modern SCV differs from both PA and GBC with reciprocal kumpanheru ‘comrade, partner’ (< Port. companheiro) (see examples in Quint 2000a: 179; Baptista 2002: 57). 206 Cf. Papia Kristang eu lavá corpo 1sg-wash-body ‘I wash myself’ (Holm 1988: 205).

3.5. Miscellaneous

131

kurpa also affected other Portuguese-derived words such as PA biña ‘wine’ and bisiña ‘neighbor’ (< Port. vinho and vizinho), where GBC has biñu and bisiñu. In other words, it is likely that PA kurpa was once identical to GBC kurpu. Note, furthermore, that both PA kurpa (45a) and GBC kurpu (45b) combine optionally with a possessive pronoun (for PA, see Muysken 1993: 300; van Putte & van Putte-de Windt 2005b: 673; for GBC, see Kihm 1996: 222f.n.; Scantamburlo 1999: 167; Peck 1988: 279; do Couto 2009b: 53). (45)

a. PA

laba

(su)

kurpa

b. GBC

laba (si) kurpa wash his/her body ‘to wash oneself’

Compare also the idiomatic expressions PA sinti su kurpa bon/malu/kansá feelhis-body-good/bad/tired ‘to feel good/bad’ and GBC sinti si kurpu kansadu feel-his-body-tired ‘to feel tired’ (Scantamburlo 2002: 321). Other examples of the reflexive use of PA kurpa are given in (46). Although GBC kurpu in its role as a reflexive pronoun is almost exclusively illustrated in the literature with the verb GBC laba ‘to wash’ (45b), other (quasi) reflexive constructions involving kurpu can easily be identified, as in (47). (46)

PA PA PA

defendé su kurpa ‘to defend oneself’ (Maurer 1988: 44); kaña kurpa ‘to get drunk’ (Muysken 1993: 287) sofoká kurpa ‘to exert oneself’ (Muysken 1993: 287) (cf. Muysken 1993: 289 and van Putte & van Putte-de Windt 2002: 123 for a wealth of additional examples)

(47)

GBC

distindi si kurpu (cf. Port. estender-se) ‘to stretch oneself’ (Bull 1989: 189) tursi kurpu (cf. Port. torcer-se) ‘to twist’ (Scantamburlo 2002: 611) larga kurpu (cf. Port. abandonar-se) ‘to let oneself go’ (Scantamburlo 2002: 331) buli kurpu (cf. Port. mover-se) ‘to move oneself’ (Scantamburlo 2002: 94) alimenta si kurpu (cf. Port. alimentar-se) ‘to feed oneself’ (Scantamburlo 2002: 431) spindi kurpu (cf. Port. levantar-se ligeiramente) ‘to get up slowly’ (Scantamburlo 2002: 567)

GBC GBC GBC GBC GBC

132

Selected parts of speech

As the examples provided thus far suggest, both in PA and GBC, the ‘body’ reflexive is largely restricted to physical action verbs.207 For PA, Kouwenberg & Ramos-Michel (2007: 327) specify that it is “restricted to verbs denoting physical action, where it is often ambiguous between a reflexive interpretation and a literal interpretation, and a few idioms (e.g. yuda su kurpa help-his-body ‘help oneself’)”. The same ambiguity described here by Kouwenberg & RamosMichel for PA is described for GBC by Kihm (1996: 222f.n.): “Mary na laba si kurpu Mariya-PROG-wash-her-body may be interpreted as either ‘Mary is washing herself’ or ‘Mary is washing her body’”. On the whole, though, the restrictions on GBC kurpu appear to be more severe than on PA kurpa. For GBC, kabesa ‘head’ is normally given as the principal reflexive pronoun (Kihm 1994: 165; Baptista, Mello & Suzuki 2007: 77, 78).208 An unambiguously reflexive concept such as ‘to commit suicide’, for instance, translates into GBC as mata (si) kabesa (Scantamburlo 2002: 243, 244), while PA speakers say mata (su) kurpa. Compare also, for instance, PA kansa kurpa with GBC kansa kabesa, both ‘to get tired, to tire oneself’ (van Putte & van Putte-de Windt 2002: 123; Scantamburlo 2002: 258). Also in SCV we find korpu ‘body’ in structures that are ambiguous between a reflexive and a literal interpretation. As far as I am aware, the feature has not been described in the literature on SCV, but the following examples suggest at least a moderate grammaticalization of SCV korpu as a reflexive (48): (48)

SCV laba korpu ‘to wash oneself’ (cf. Port. lavar-se) (Nicolas Quint p.c.) SCV larga korpu ‘to go to sleep’ (cf. Port. deitar-se) (Brito 1887: 394) SCV stafa (si) korpu ‘to strain/exert oneself’ (cf. Port. estafar-se) (Lang 2002: 744) SCV suste korpu ‘to remain standing’ (cf. Port. manter-se de pe) (Lang 2002: 760) SCV torse korpu ‘to stretch oneself’ (cf. Port. espreguicar-se) (Lang 2002: 795)

207 In a typological study of reflexive pronouns in creole languages, Heine (2005: 209) observes that body part reflexives “appear to be confined to uses in specific contexts in all creoles concerned (…); we have no explanation why these reflexives have not been further grammaticalized”. 208 According to Scantamburlo (1999: 167), GBC kabesa is optionally preceded by a possessive pronoun, just as GBC kurpu/PA kurpa. Kihm (1996: 222f.n.) contradicts this claiming that a possessive is obligatory with kabesa. In support of Scantamburlo’s view, I found examples of reflexive kabesa with and without a possessive pronoun. Also in SCV, reflexive kabesa is optionally accompanied by a possessive pronoun (Quint 2010: 140; Veiga 2000: 175; Baptista 2002: 55, 56).

3.5. Miscellaneous

133

Just like GBC, however, speakers of SCV principally employ constructions with reflexive (si) kabesa ‘head’.209 According to van Putte & van Putte-de Windt (2002: 124), PA kabes ‘head’ can also occur in quasi reflexive clauses. The example they provide is the idiomatic expression PA preokupá bo kabes preoccupy-your-head ‘to worry’, which is also found with the English-derived verb wòri: PA wòri bo kabes ‘to worry’. As far as I can tell, however, the reflexive use of PA kabes is restricted to this idiom. In sum, the available linguistic data are indicative of a once incipient process of grammaticalization in (Early) Upper Guinea PC of the word for ‘body’ into a reflexive pronoun. This process was abandoned in Upper Guinea PC in favor of a ‘head’ reflexive, while in PA, grammaticalization of kurpa was taken several steps further and the ‘body’ reflexive became operative with a wider range of verbs. It is conceivable that convergence with Kwa and Bantu patterns contributed to the broader context of use of the ‘body’-reflexive in PA relative to Upper Guinea PC. In various Kwa and Bantu languages as well as in Gulf of Guinea PC (with an acknowledged Kwa/Bantu substrate), the primary reflexive pronoun is an equivalent of ‘body’ (49) (Parkvall 2000: 59; Heine 2005: 247f.n.)210 : (49)

n sa ke211 kanx’ igbê me I PROG IMP+go rest body my ‘I’m going to rest’ (Maurer 2009a: 152) b. ANG ê mata ôngê rê he kill body his ‘He killed himself’ (Maurer 1995: 145) a. PRI

It is important to note, however, that this construction contrasts with PA / Upper Guinea PC not only phonetically, but also in terms of word order: the word order in Gulf of Guinea PC, ‘body + possessive’, is reverse to that in PA and Upper Guinea PC ‘possessive + body/head’. Note, finally, that PLQ makes no use of pronouns or other overt markers to convey reflexivity (Dieck 2008: 45). 209 For more details on reflexivity in CV, see also Fiéis & Pratas (2007) and Veiga (1995: 361, 362). 210 Maurer (2009a: 152) also recorded examples of reflexive kabese ‘head’ in PRI, apparently co-occurring with reflexive igbê ‘body’. This may (but need not) be the result of contact with CV, speakers of which migrated to Príncipe in the second half of the 19th and early 20th century to work on cacao plantations (Batalha 2004: 38; Nascimento 2008: 55). 211 PRI ke is a contraction of ka we, with ka = imperfective aspect marker, and we = ‘to go’.

134

Selected parts of speech

3.5.1.3. PA mes / SCV me ‘self’ As noted, the use of PA kurpa is largely restricted to physical action verbs and some idiomatic expressions. In PA, then, the most frequent way of making a verb (whether physical or non-physical) reflexive is by means of the identifier mes ‘self’ < Port. mesmo (= Sp. mismo) preceded by a personal (possessive or object) pronoun: mi mes ‘myself’; bo mes ‘yourself’, etc. (Kouwenberg & Murray 1994: 40; Kouwenberg & Ramos-Michel 2007: 327; Kramer 2004: 185, 186).212 Also in SCV, a derivative of Port. mesmo is preceded by a personal pronoun to create compound identifiers: SCV mi-me, bo-me, etc. (Baptista 2002: 57; Baptista, Mello & Suzuki 2007: 77).213 Reflexives based on an equivalent of ‘self’ are of course rather unmarked across creoles214 and non-creoles.The chief observation is rather that the etymon selected to perform this task in PA is once more Portuguese-derived and once again has a close cognate in Upper Guinea PC. That PA mes does indeed derive from Port. mesmo rather than Sp. mismo finds support not only in the preservation of the etymological stressed /e/ but also in its adverbial use meaning ‘really, indeed’. This use is typical of Portuguese mesmo, not of Spanish mismo. Compare, for instance, Port. (muito) velho mesmo ‘really (very) old / (very) old indeed’ with PA (masha) bieu mes very-old-self ‘idem’ (van Putte & van Putte-de Windt 2005a: 284, 290).215 It is not surprising to find the equivalent SCV me with the same adverbial value: SCV bedju mé old-self ‘really old’ (Lang 2002: 434). 3.5.2. The deictic marker Early PA / Upper Guinea PC es Different from Upper Guinea PC, modern-day PA is characterized by the use of a definite article, e, as in PA e kas ‘the house’. PA demonstratives are construed

212 Kouwenberg & Ramos-Michel (2007: 327) observe: “Where the ‘pronoun + kurpa’ strategy can be used, the ‘pronoun + mes’ strategy is equally acceptable, but the reverse situation does not hold, showing ‘pronoun + mes’ to be the main reflexive form.” 213 Unlike PA mes, SCV me is only preceded by an oblique pronoun, not by a possessive. Hence, we find both e mes and su mes in PA, but only el-me (*si-me) in SCV (Baptista 2002: 56; Quint 2000a: 176). In GBC, the identifier pronoun is propi ‘self’(mi-propi, bo-propi) (Noël Bernard Biagui, p.c.), which also exists in SCV (Quint 2000a: 176). 214 For cross-creole data on reflexive strategies, see Holm & Patrick (2007: feature 17.7), Muysken & Smith (1994) and Heine (2005). 215 In Spanish, this meaning is typically achieved by means of the adverbs realmente/verdaderamente and/or adverbial expressions such as de verdad / de veras.

3.5. Miscellaneous

135

with the definite article e and one of three locative adverbs, aki, ei216 , or aya (< Sp. aquí, ahí, allá) enclosing the noun, as in PA e kas aki ‘this house’. Note that in the autonomous determiners PA esaki ‘this one’, esei ‘that one’ and esaya ‘that one (over there)’, as well as anaphoric esun and its plural variant esnan (= es + plural marker nan), the Early PA demonstrative es (< Sp. ese/este or Port. esse/este) still shines through, out of which the modern definite article e must have developed. I will return to this further below. In Upper Guinea PC, a specialized definite article like PA e does not exist. Rather, the Upper Guinea PC morpheme es is ambiguous between a definite article and a demonstrative (Kihm 1994: 139 for GBC; Quint 2000a: 186 for SCV). Thus, Upper Guinea PC es lugar can mean either ‘the place’ or ‘this place’, as the context demands. A circumlocution very similar to the PA demonstrative (PA e kas aki) is commonly used in Upper Guinea PC to give more deictic emphasis: es kaza li ‘this house’. Upper Guinea PC furthermore has es-li as an autonomous determiner, similar to PA esaki. On the basis of these correspondences, Birmingham (1970: 146), comparing the deictic systems of present-day GBC and PA, asserted that “the method is essentially the same”. Present-day PA’s deictic system is thus in several ways comparable to that of Upper Guinea PC, the main difference being the lack of a specialized definite article in the latter. However, an analysis of 19th-/early 20th-century PA texts shows that the Early PA deictic system overlapped with Upper Guinea PC more considerably still (cf. Martinus 1996: 34, 35). Crucially, the modern PA definite article e was originally realized as es and, like Upper Guinea PC es, was very much ambiguous between a definite article and a demonstrative: (51)

a. EPA es soberbia ‘the/this pride’ (Jansen 1911: 19) b. GBC Es abitu ‘the/this habit’ (Scantamburlo 2002: 3) c. SCV Es káru ‘the/this car’ (Quint 2000a: 184)

(52)

a. EPA

Na ees momeentoe in-DEF-moment ‘At that moment’ (Putman 1852: 27). b. SCV Na es momentu in-DEF-moment ‘At that moment’ (Lang 2002: 729) (cf. ex. [4] in §3.2)

To be sure, these examples are ungrammatical in modern-day PA, even though PA es appears to have existed in allomorphy with e well into the 20th century (cf. its use in e.g. Geerdink-Jesurun Pinto 1952[= Baart 1983]). Also, this use of 216 The older variant PA is steadily used in Early PA texts and ai is still attested in Bonairean PA.

136

Selected parts of speech

PA es can still be found in ‘modern’ PA evangelical texts (which are renowned for retaining conservative features): e.g. PA na es ora di purbamentu ‘in this testing hour’217 (cf. SCV n’es ora, Lang 2002: 101). Scholars such as Maurer (1998: 155) and Munteanu (2004) have correctly commented that no diachronic data is required to assume the development of PA’s definite article e out of the Early PA demonstrative es, as this morpheme is still visible in modern PA not only in the afore-mentioned autonomous demonstratives PA esaki, esei, esaya, esun and esnan, but also in lexicalized compositions such as the PA adverbs turesten ‘in the meantime’, esta218 ‘in other words’ and pesei ‘therefore’ (< *pa es ahi, discussed below), or the PA conjunction pasobra (< *pa es obra) ‘because’. Still in 1953, moreover, the PA specialist Maduro (1953: 74) used the adverb esor’ei ‘then, at that moment’ (< es + ora + ei ‘that hour there’, cf. SCV es ora li idem; Lang 2002: 95). Nowadays, PA e or’ei is the norm. The sum of these data puts beyond doubt that the modern PA definite article e “does not derive from the Spanish definite article, but from the Portuguese or Spanish demonstrative determiner este or esse/ese”219 (Maurer 1998: 155f.n.) and thus renders untenable Kouwenberg & Ramos-Michel’s (2007: 323) claim that PA e “is assumed to derive from the masculine form of the superstrate singular definite article el”.220 Early PA es is thus reminiscent of the Upper Guinea PC determiner es, both either functioning as a definite article or taking on a demonstrative value, as the context demands. Moreover, like Upper Guinea PC es, Early PA es only combined with locative adverbs in order to create a stronger deictic contrast. Compare, for instance, (53a) with (53b):

217 Online source: http://www.bendishon.com/bendishon/manageDocument.do?dispatch=view&id=1110, accessed on 01-04-2011. 218 PA esta is composed of the early determiner es and the copula ta (lit. ‘that is’) and is typically used as an adverb in the sense of ‘namely, in other words’ (cf. van Putte & van Putte-de Windt 2005a: 128). 219 Original quote: “no proviene del artículo definido español, sino del determinante demostrativo portugués o español este o esse/ese” 220 Cross-linguistically, the development of a demonstrative into a definite article is in fact quite common (cf. e.g. Heine & Kuteva 2002: 4, 106). Bruyn (2009) provides an interesting discussion of this diachronic development in Sranan. Besides (Early) PA and Upper Guinea PC, many other creoles have a deictic marker that unites the functions of a definite article and a demonstrative (Holm & Patrick 2007: x). Interestingly, in most creoles this deictic marker is not derived from the superstrate’s definite article but from the superstrate’s demonstrative (e.g. Perl 1989a: 374, 375).

3.5. Miscellaneous

(53)

137

a. EPA

Ees falso tistimonio DEF false testimony ‘The/this false testimony’ (from the 1803 Aruban letter, in Martinus 1996: 34) b. EPA solamente na ees caso allie only in DEF case there ‘only in that case’ (from the 1803 Aruban letter, in Martinus 1996: 34)

This pattern is in fact typical of most if not all Early PA texts analyzed for the present study and, as hinted at previously, this is still the norm in Upper Guinea PC, where es karu ‘this car’ and es karu li ‘this car (over here)’ can both occur, with the adverb li merely adding deictic emphasis (Wilson 1962: 16 for GBC; Quint 2003: 222 for SCV). Interestingly, as late as 1928, this pattern is still described as typical of PA by Lenz (1928: 114), who provided PA es kas and es kas aki and translated both as ‘this house’, thus showing his awareness of the lack of a ‘true’ definite article. To be sure, in modern-day PA we now find e kas ‘the house’ in complementary distribution with e kas aki ‘this house’. It is revealing to note that the deictic system of (Early) PA and Upper Guinea PC resembles that of French rather than Spanish or Portuguese. In describing PA’s deictic system, Lenz (1928: 114) therefore provided French translations: “es kas = cette maison; es kas akí = cette maison-ci”, etc. (cf. Munteanu 2004: 528, 529). Although such deictic systems are not rare world wide (Heine & Kuteva 2002: 172), it shows that the closeness of the lexifiers Spanish and Portuguese cannot be held responsible for the close parallels between (Early) PA and Upper Guinea PC in this domain. Martinus (1996: 36) asserted in this respect: “It is hard to believe any kind of (…) ‘Spanish derived creole’ would go to such a length to invent the same analytic deictic system as CV [Cape Verdean] (…), while it could simply stick to an equally transparent system as the Spanish one.” Martinus’ assertion is supported by the fact that, unlike PA and Upper Guinea PC, the Colombian creole PLQ did adopt the Spanish demonstrative system with ete (< Sp. este ‘this’), ese (< Sp. ese ‘that’) and aké (< Sp. aquel ‘that’) (Schwegler & Green 2007: 293). More contrast is provided by Gulf of Guinea PC. Just as for Early PA / Upper Guinea PC es, it has also been noted for the Gulf of Guinea PC definite marker se (< Port. esse) that it can function as both a definite article and a demonstrative pronoun (Janson 1984: 309; Valkhoff 1966: 93; Günther 1973: 58; Ferraz 1978: 239, 240). However, significant differences between PA and Upper Guinea PC on the one hand and Gulf of Guinea PC on the other are once more found in the

138

Selected parts of speech

domain of word order: PRI omi sê man-this ‘this man’ = Upper Guinea PC es omi / Early PA es homber.221 A notable difference between PA and Upper Guinea PC is that PA has no cognate of Iberian aquel, which has become integrated in Upper Guinea PC as the distal demonstrative kel. Instead, PA incorporated three locative adverbs (aki [proximal], ei [intermediate] and aya [distal]), where Upper Guinea PC has only two (li [proximal] and la [distal]).222 Assuming PA’s deictic system developed out of Upper Guinea PC’s dual demonstrative system, the loss of *kel in PA can be explained in terms of ‘narrowing’. Heine & Kuteva (2005) explain this process and provide an example which nicely fits the present context: Grammaticalization frequently involves narrowing, in that out of a range of lexical or grammatical items forming a paradigm, one is further grammaticalized. For instance, the evolution of definite articles is likely to be characterized by a process where there is a paradigm of demonstrative attributes and where one of these attributes is selected to serve as a definite marker. (Heine & Kuteva 2005: 269)

We can accordingly hypothesize that in Early PA both *es and *kel were available as demonstrative attributes and that, subsequently, *es was selected to evolve into a general definite marker at the expense of *kel. The subsequent erosion of es to e in PA (in the course of the 20th century223 ) has an interesting parallel in the Casamance variety of GBC, where the determiner es obligatorily becomes e when used in circumlocutions such as GBC e lugar li ‘this place (here)’ (Noël Bernard Biagui, p.c.; cf. Almeida 1991: 10). In autonomous position, es-li is used. Thus: GBC e lugar li GBC es-li

= =

PA e logá aki PA esaki

To close this section, I focus briefly on Quint’s (2008b: 136) discussion of Brito’s (1887) grammar of SCV and the use therein of the adverb pâ ês ‘therefore’(54a), 221 Janson (1984: 311) provides a comparison of deictic markers in creoles and also comments on this typological difference between PA and Gulf of Guinea PC. 222 In SCV, es combines exclusively with the adverb li, while distal kel combines with both li and la (Quint 2000a: 184). In the Brava dialect of CV, however, es does also combine with la (Meintel 1975: 229). 223 PA es must have dropped the /s/ in the course of the 20th century: while Putman’s (1849: 23) grammatical sketch provides only PA es, van de Veen Zeppenfeldt’s (1928: 22) and Lenz’s (1928: 97) grammars mention the variation between PA es and e; Goilo’s method ([1951]1975), finally, makes no more mention of es as an independent deictic marker. In that same period of time, e must have acquired the function of a ‘pure’ definite article.

3.5. Miscellaneous

139

which according to Quint no longer exists in modern-day SCV. Strikingly, we find the same adverb in Early PA in the speech of Lenz’s (1928) informant Natividad Sillie (54b): (54)

a. SCV Pâ ês nu ta tratâ di kada u separadamenti therefore we IMP treat of each one separately ‘Therefore, we will treat each one separately’ (Brito 1887: 362) b. EPA pa es tata no tin náda di bísa tokante di therefore father NEG have nothing to say regarding es kriatura this creature ‘Therefore, the father did not have anything to say about this creature’ (Lenz 1928: 288; cf. an example of p’es in Jansen 1911: 22)

Note that, different from SCV, the adverb pa es has not fully disappeared from PA, although it is nowadays only discernible in the PA adverb pesei ‘therefore’, which should diachronically be analyzed as the contraction of Early PA pa + es + ai as it can be found in several 19th-century Early PA texts. Compare: (55)

a. EPA

b. PA

Pa ees ahi mi ta bisa therefore I IMP say ‘Therefore, I tell you (…)’ pesei nos ta onr’ therefore we IMP honor ‘therefore, we honor him’

bosonan (…) you+PL (Conradi 1844: 25) é him (Baart 1983: 81)

In GBC (54c), the adverb pa es appears to exist still in its original shape, although a recent innovation modeled on Port. por isso cannot be discarded. (54)

c. GBC Pa es i prisizu kolaborason Therefore be necessary collaboration ‘Therefore, collaboration is needed’ (Scantamburlo 2002: 284)

3.5.3.

Negation

Arguably the most striking difference between the morphosyntax of PA and Upper Guinea PC is the negator PA no versus Upper Guinea PC ka. Upper Guinea PC ka is generally believed to derive from Portuguese nunca ‘never’ (e.g. Lopes da Silva 1957: 150; Bartens 2000: 57; Kihm 2007: 256), which was also the etymon for a verbal negator in several Asian Portuguese creoles (Holm

140

Selected parts of speech

1988: 172). PA no, on the other hand, quite probably derives from the Spanish homophone negator. However, apart from the two obviously differing morphemes, a look at the general negation patterns in PA and Upper Guinea PC reveals that these patterns largely coincide (cf. Quint 2000b: 141, 142), while differing in several respects from Gulf of Guinea PC and PLQ as well as from Spanish and Portuguese. For instance, PA and Upper Guinea PC show preverbal negative concord (56): (56)

a. PA

nunka

mi

b. SCV

nunka m- ka never I NEG ‘I never went’

no

a

bai

(Maurer 1988: 41)

bai go

(Quint 2000b: 142)

PFV

Clearly, this pattern differs from the mainstream varieties of Spanish and Portuguese: (56) may translate into any variety of Spanish as nunca fui and/or no fui nunca, but never as *nunca no fui. Schwegler, a connoisseur of (nonstandard) Latin American Spanish, indeed confirms the ungrammaticality of Spanish *jamás no nos pusimos bravos ‘we never became angry’, which must instead be rendered as jamás nos pusimos bravos (Schwegler & Green 2007: 287). In non-standard varieties of Portuguese, double negation patterns do exist, but these are all of a discontinous nature (i.e. with the second negating element following the verb224 ), which is not the case for PA and Upper Guinea PC. But even if preverbal double negation patterns such as that exemplified in (56) were to be found in one or the other non-standard Iberian variety, it will presumably never be obligatory, whereas it is considered obligatory in PA and Upper Guinea PC (Maduro 1971: 47; Baptista 2002: 119). We may furthermore point out the contrast between the preverbal negation patterns in PA / Upper Guinea PC and the postverbal negation and/or discontinuous double negation patterns found in PLQ and Gulf of Guinea PC. To start with PLQ, unlike PA and Upper Guinea PC, the principal negator nu normally follows the predicate, although a second pleonastic nu may optionally appear in preverbal position225 (57). When negative concord is applied, the principal negator remains in postverbal position (58): 224 Cf. Holm et al. (1998: 104–106) for double negation patterns in rural varieties of Brazilian Portuguese and Liliana Inverno (forthcoming) for double negation patterns in Angolan Portuguese. 225 Although negations of the type nu + V + nu and V + nu are considered to be the typical negation pattern in PLQ, Schwegler & Green (2007: 287) argue that strictly preverbal negation nu + V is possible as a third variant.

3.5. Miscellaneous

141

(57)

PLQ

Bo (nu) kelé nda mi na nu You NEG want give me nothing NEG ‘You don’t want to give me anything’ (Schwegler & Green 2007: 287)

(58)

PLQ

Nunka suto poné brabo nu never we become angry NEG ‘We never became angry’ (Schwegler & Green 2007: 287)

Though the negating morphemes differ, the syntax of negation in Gulf of Guinea PC is very much like in PLQ, while characterized by the use of a postverbal negator and/or discontinuous double negators. In ANG, the standard negation pattern is [na + predicate + wa∼va] (59a), in PRI we find [predicate + fa] (59b), and in ST the pattern is [na + predicate + fa] (Lorenzino 2007: 13; Maurer 2009a: 133). Like PLQ, and unlike PA and Upper Guinea PC (56), double negators in Gulf of Guinea PC are always discontinuous, with the principal negator occurring postverbally as in (60). (59)

(60)

a. ANG e na ka me ombo wa he NEG IMP eat goat NEG ‘He doesn’t eat goat’ b. PRI E xiga ontxi fa he arrive yesterday NEG ‘He didn’t arrive yesterday’

(Lorenzino 2007: 13)

(Maurer 2009a: 133)

a. ANG nunka a ka be wa never impersonal IMP see NEG ‘I never see anything’ (Lorenzino 2007: 13) b PRI M maxi ve ninge nhon na te I not.yet see person no in country se ki ka gbo xi xin fa. this REL HAB shit without press NEG ‘I have never seen anybody in this country who could defecate without pressing down’ (Maurer 2009a: 137)

In sum, notwithstanding the different negating morphemes (PA no versus Upper Guinea PC ka), the syntax of negation coincides fully in PA and Upper Guinea PC while differing from the lexifiers as well as from PLQ and Gulf of Guinea PC.226 226 Cf. Baptista, Mello and Suzuki (2007: 66), who note that “Neither CV nor GB have discontinuous double negation of the type found in Angolar and the other Gulf of Guinea Creoles”.

Chapter 4 Morphology

Introduction While the fate of morphology in creolization is hotly debated, the matter of the fact is that both PA and Upper Guinea PC are characterized by a productive set of derivational and inflectional morphemes. These might trace back to a common stage of development and will therefore be analyzed and compared below.227 It goes as common sense in historical, genetic and contact linguistics that morphology is among the most stable domains of a language undergoing change, be it contact-induced or not (Thomason & Kaufman 1988: 14; Comrie 2008: 15). Owing to this low borrowability rate, morphology is among the most reliable indicators of genetic relatedness between a given pair of languages: It is quite likely that we might have two pairs of languages, X and Y sharing complex morphology, P and Q sharing only lexicon, both in fact separated by equivalent amounts of differentiation from their respective common ancestor, and that we would be able to establish the genetic relationship of X and Y but not of P and Q. (Comrie 2000: 40)

A well-known case in point is the Indo-European family tree, which was established to an important extent on the basis of morphological correspondences, the nature of which, Comrie (2000: 40) affirms, “cannot be plausibly recon-

227 A widely discussed topic within creole studies is the complex relationship between pidgins and creoles and (the lack of) morphology (DeGraff 2001; Plag, ed. 2003; Plag 2005; Booij 2005: 259; Parkvall 2006; Bhatt & Plag 2006; McWhorter 2001, 2005; Holm 2008). Though the issue is certainly fascinating, it will not play any role in this chapter. Suffice it to stress here that the hypothesized presence of productive morphology in the Upper Guinea PC ancestor from which I assume PA descends does not imply that proto-Upper Guinea PC started life with productive morphology. In fact, I assume it did not and that, instead, it acquired this morphology after the actual process of creolization. After all, if the formation of proto-Upper Guinea PC was completed already by the early 16th century (see the introductory chapter of this study), there was plenty of time for it to borrow morphology from Portuguese (with which it was in ongoing contact) subsequent to its nativization and prior to its transfer to Curaçao in the latter half of the 17th century.

144

Morphology

ciled with any assumption other than genetic relatedness of the languages in question”.228 Within the domain of morphology, inflectional morphology (barring plural morphology) is considered to be more complex, and hence more stable, than derivational morphology: “no inflectional morphology is borrowed unless derivational morphology is also borrowed” (Moravcsik 1978 in Matras 2007: 33). In increasing order of importance, then, I will first discuss derivational in (§4.1), followed by inflectional morphology (§4.2). To close this chapter, §4.3 provides an elaborate diachronic discussion of passivization in PA and will argue that the creole started out with auxiliary-less passives (or so-called passive verbs), a feature that is still very typical of Upper Guinea PC.229

4.1. Derivational morphology In this section, I will first describe PA’s two most productive deverbal nominalizing suffixes, -mentu (§4.1.1) and -dó (§4.1.2), then draw a comparison with Upper Guinea PC (§4.1.3), and close with salient data from Early PA texts (§4.1.4). 4.1.1.

PA -mentu

A good indication of the productivity of the characteristic PA suffix -mentu is that it combines just as easily with native as with nonnative, borrowed verbal roots, allowing for the creation of coinages, viz. root-affix combinations not attested in the lexifier(s). In other words, the class of deverbal action nouns derived with -mentu230 is potentially open. Some of the most clear-cut coinages of the PA lexicon created with -mentu are those where it attaches to verbal roots of Dutch origin:

228 Comrie (2000: 38) provides a second interesting example of the utility of morphology in establishing kinship: “The overall effect of word taboo, at least in a small community like that of the Haruai, is to accelerate rates of lexical replacement. Indeed, the way in which I was able to establish that Haruai is genetically more closely related to Hagahai was by comparing the verb morphologies”. 229 Although passivization thematically fits in with §4.2, the analysis of passives in early 19th-century PA texts and the comparison with Upper Guinea PC proved to be revealing to the extent that a separate subsection is warranted, which will be §4.3. 230 In the Bonairean variety of PA, -mentu is typically realized as -/men/ (Lucille BerryHaseth, p.c.).

4.1. Derivational morphology

145

PA snuimentu ‘the act of cutting’ (PA snui ‘to cut’ < Du. snoeien idem) PA wakmentu ‘the act of watching’ (PA wak ‘to watch’ < Du. waken idem) PA skrufmentu ‘the act of screwing’ (PA skruf ‘to screw’ < Du. schroeven idem) etc. (cf. Dijkhoff 1993: 148–151; Birmingham 1970: 127; Kowallik & Kramer 1994: 157) 4.1.2.

PA -dó

In one of the most recent works on PA morphology (Sanchez 2005), the agentive nominalizer suffix PA -dó (< Sp./Port. -dor)231 is inaccurately said to be “[a]n example of a morpheme with limited integration” since it “is found only with Spanish or Iberian roots (Dijkhoff 1993)” (Sanchez 2005: 67). Sanchez’ reference to Dijkhoff (1993) is misleading, since Dijkhoff (1993: 148–151) is not only one of the richest sources (if not the richest) on word formation with the deverbal agentive nominalizer -dó in PA, she explicitly states that “-dó may be attached to almost any Papiamentu verb” (Dijkhoff 1993: 149)232 , regardless of the verb’s etymology. The following are just three examples out of a potentially open class of coinages based on the nominalizer -dó, demonstrating its productivity and degree of integration (cf. Kowallik & Kramer 1999: 157): PA yagdó ‘hunter’ (PA yag ‘to hunt’ < Du. jagen idem) PA hürdó ‘tenant, renter’ (PA hür ‘to rent’ < Du. huren idem) PA fèrfdó ‘painter’ (PA fèrf ‘to paint’ < Du. verven idem) 4.1.3.

Upper Guinea PC -mentu / -dor

To summarize the above, the most productive PA suffixes are doubtlessly -mentu and -dó and both stand out for potentially combining with all verbs (Dijkhoff 1993: 141; cf. Maurer 1998: 181, 182; van Putte 1999: 94; Bartens 1996: 255, 256). To come straight to the point, Kihm (1994: 131) points out that in GBC, “[t]wo suffixes should be set apart as they are used to derive nouns (…) from verbs. One is -dur (…). The other is -menti”.233 He affirms that both suffixes may 231 The loss of the final -r of PA -dó is a relatively recent phenomenon. In the 19th-/ early 20th-century PA texts analyzed for the present study, the suffix -dó appears consistently in its original form . 232 Moreover, Dijkhoff (1993: 150) shows that the suffix is productive both on Aruba and Curaçao, giving interesting examples such as double agentive Aruban PA wèlderdó meaning ‘welder’, whereas on Curaçao the term lasdó (idem) is more likely to be heard (with las ‘to weld’ < Du. lassen ‘idem’). 233 Note that in the more conservative Casamance variety of GBC, -menti and -dur are still realized as -mentu and -dor (Rougé 1988).

146

Morphology

attach to all verbs (‘free derivation’), meaning that “the stock of derived nouns is potentially open” (1994: 131). As outlined above, the same observations apply to PA -mentu and -dó. Examples of coinages (i.e. derivations with no cognate in the lexifier) shared by PA and GBC include the following: – PA siña ‘to teach, learn’ GBC nsina idem – PA pidi ‘to ask’ GBC pidi idem – PA pursiguí ‘to persecute’ GBC pursigui idem – PA ski(r)bi ‘to write’ GBC skirbi idem

> PA siñamentu ‘education’ > GBC nsinamentu idem = Sp. enseñanza, Port. ensino > PA pididó di limosna ‘beggar’ > GBC pididor di smola idem = Sp./Port. mendigo, Port. pedinte > PA pursiguimentu ‘persecution’ > GBC pursiguimentu idem = Sp. persecución, Port. prosseguimento > PA ski(r)bidó ‘author’ > GBC skirbidor idem = Sp./Port. escritor

Obviously, this list is not exhaustive in any way, since, as mentioned, the category of nouns derived with these suffixes is open in both creoles. For SCV, the literature suggests a more limited productivity of the suffixes -mentu and -dor compared to PA and GBC. This seems to correlate with the fact that modern SCV has incorporated an extensive series of additional derivational suffixes from Portuguese (see for instance Lang 2002: XLIff; Quint 2000a: 144, 145; Veiga 2000: 132, 133). Furthermore, in several cases where PA and GBC use regularized derivations, SCV has borrowed the irregular, suppletive forms from Portuguese. Compare, for instance, PA ski(r)bidó and GBC skirbidor (both ‘author’), on the one hand, with SCV skritor (< Port. escritor) on the other. Nonetheless, some petrified coinages testify to the fact that in a previous stage -mentu and -dor were productive also in SCV. Take, for instance, the noun SCV ndjutumentu ‘impoliteness, discourtesy’, where -mentu is attached to the verbal root ndjutu ‘to treat without respect / disdain’, in turn derived from the Mandinka verb jutu idem (Lang 2002: 476; Quint 2006: 79). 4.1.3.1. On the exclusivity of PA / Upper Guinea PC -mentu Although creole languages certainly borrow derivational morphology (see Parkvall 2006), the similarities between PA and Upper Guinea PC are in fact quite meaningful. In relation to the Saramaccan pair paká-pakamentu and Sranan paypayman (both ‘to pay’vs. ‘payment’), for instance, Cardoso & Smith (2004: 121) observe that “[s]uch derivational pairs are very rare, and can be counted on the

4.1. Derivational morphology

147

fingers of two hands” and thereby speak of “the non-productive derivational relationship between nouns (…) and verbs” in the Suriname Creoles. In Chabacano, a creole rich in productive, Spanish-derived derivational affixes, the suffix -dor is in use, but the language lacks a suffix corresponding to Sp. -miento (Steinkrüger 2003: 257–259). The Spanish-lexifier creole PLQ seems to lack productive denominalizing morphology all together. In Gulf of Guinea PC, the picture is less clear. For ST, Ferraz (1979: 57–59) makes no reference to a possibly productive deverbal nominalizer and the nouns found in Fontes’ (2007) ST dictionary derived with -mentu and -dô are limited to root-suffix combinations found in the lexifier. In ANG, the suffix -rô∼-dô is productive, as suggested by coinages such as kokolarô ‘someone who scratches’ (with kokola ‘to scratch’ of African origin, Maurer 1995: 90). ANG -me(n)tu is also said to be productive (Maurer 1995: 91), but coinages that might prove this are not given: ANG kadhamentu ‘marriage’ and kuzimentu ‘decoction’ (Maurer 1995: 91) could be direct loans from Portuguese, which has casamentu and cozimento with the same meanings, though of course this does not prove ANG -mentu to be unproductive either. For PRI, the same picture emerges: “agentive -dô is productive”, whereas derivations with -mentu are “relatively rare” (thus Maurer 2009a: 93). In sum, within the branch of Iberian-based creoles, a productive -DOR suffix is not unique; at least Chabacano and Gulf of Guinea PC have it. But the characteristic and productive use of the derivational suffix -MENTO appears to be exclusive to PA and Upper Guinea PC. 4.1.3.2. On the etymology of PA -mentu It is unclear why Sanchez (2005: 24) gives Spanish -miento as the etymon of PA -mentu; the monophthong clearly points towards a link with Portuguese. And although it has on rare occasions been issued that PA -mentu derives from an undiphthongised Old Spanish suffix -mento (see Grant 2008a: 54), it is a wellestablished fact that the diphthongization of -mento to -miento in Old Spanish had already been completed by the end of the 14th century (Cano Aguilar 2004: 313, 435; Penny 2002: 51–53). Of course, in several lexical items, modern Spanish has preserved the undiphthongized suffix -mento, but this form represents what Penny (2002: 290) refers to as a ‘learned suffix’, i.e. a suffix that has “passed into Spanish as a result of the borrowing of Latin words” (Penny 2002: 290). In other words, as of the 15th century and perhaps earlier, only diphthongized -miento is truly productive in Spanish. This and the above-mentioned fact that none of the Spanish-based creoles has a productive -m(i)ento suffix leave little ground to argue for a Spanish etymology of PA -mentu.

148

Morphology

In the case of the suffix PA -shon, PA did in fact preserve a Spanish diphthong (Sp. -ción = Port. -ção; cf. Upper Guinea PC -son). However, although PA -shon created a few coinages (e.g. moveshon ‘movement’ = Sp. movimiento), it does not suffix to non-native (e.g. Dutch-derived) verbal roots, suggesting its productivity is limited and its time-depth reduced (Kouwenberg & Murray 1994: 23, 27; Dijkhoff 1993: 151; Lenz 1928: 148; Maurer 1998: 182; van Putte 1999: 94). In other words, in contrast to -mentu and -dó, -shon more likely constitutes an example of non-native morphology borrowed into PA from its number one contact language, Spanish.234 4.1.3.3. Digression: On the glossonym Papiamentu The etymology of the glossonym Papiamentu has had a central position in the debate on the origins of PA for over a century (Jesurun 1897: 95; van Wijk 1958: 175; Maduro 1965: 26; Rona 1971: 11; Munteanu 1991: 50) and is still an outstanding issue (Kramer 2004: 97; Megenney 2007: 55; Martinus 2007: 10, 11). When Megenney (2007: 55) asserts that PA papia derives from Port. papear ‘to chatter’, he does so without explaining why Spanish papear (which, according to the Diccionario de la Real Academia Española, means ‘balbucir’, ‘tartamudear’, ‘hablar sin sentido’) could not be the etymon. Kramer (2004: 97) affirms that the verb can be attributed to either Sp. papear or Port. papear 235 , with equal likelihood. A fact not considered by both authors but that clearly strengthens the claim of a Portuguese rather than a Spanish etymon, is that Port. papear has a highly frequent cognate not only in PA but also in CV and GBC (Rougé 2004a: 223; Bartens 1996: 265) and in at least one Asian Portuguese creole, indeed, Papia Kristang, whereas reflexes of Sp. papear are to my knowledge not attested in the Spanish-lexifier creoles PLQ and Chabacano. It is interesting to note, on the other hand, that a cognate of Port. papear is absent in Gulf of Guinea PC (Rougé 2004a: 223; Maurer 2009a: 231), as this contradicts the widespread tendency to consider the verb papia as “gemeinsames element Portugiesischer Kreols” [‘shared element of the Portuguese creoles’] (Bachmann 2005: 47; cf. Bartens 1996: 19; Viaro 2005: 84). 234 In addition to PA -shon, several other affixes listed by Dijkhoff (1993: 81–85) appear to be in the process of becoming productive. Not listed by Dijkhoff are the PA diminutive suffixes -chi and shi (< Du. -tje and -sje). Their productivity is visible in coinages such as PA yuchi ‘little child’ < PA yu (< Sp. hijo) ‘child’ + -chi, PA avochi ‘forefather’ < *avo (< Port. avo ‘grandfather) + -chi. PA garoshi ‘small car’ < Sp./Port. carro ‘car’ + -shi. 235 Schuchardt (1882: 895) had already observed that papear is found in both Old Spanish and Portuguese.

4.1. Derivational morphology

149

Whereas the verb Sp./Port. papear (also ∼papiar) is still an entry in the larger dictionaries of both Spanish236 and Portuguese237 (suggesting it is still used in one or the other Iberian variety), the following derived nominalizations shared by Upper Guinea PC and PA are not: PA / Upper Guinea PC papia ‘to talk’

> PA papiadó ‘chatterbox’ > SCV papiador idem (Lang 2002: 540) > GBC papiador idem (Rougé 1988: 115) = Sp. hablador, Port. falador

PA / Upper Guinea PC papia ‘to talk’

> PA papiamentu ‘the act of talking, chatter’ > SCV papiamentu idem (Veiga 2000: 133) > BaCV papiamente idem (Veiga 2000: 133) > GBC papiamentu idem238 = Sp./Port. charla, Sp. hablamiento/Port. fofoca

Note that the noun papiamintu is also provided in the Papia Kristang dictionary by Baxter (2004: 69), endowed with the label “archaic”. It suggests that both the verb papear and the derived noun *papeamentu may once have been quite common in colonial maritime Portuguese. 4.1.4. The suffix -dadi in Early PA texts Several of the 19th-century PA texts examined for the present study show the nominalizing suffix -dadi (< Port. -dade) appearing instead of the modern suffix -dat (< Sp. -dad). Thus, where modern PA has skuridat ‘darkness’, kapasidat ‘capacity’, falsedat ‘falseness’, fermidat ‘illness’ and bondat ‘benevolence’, we find , , , and in a number of Early PA sources (van Dragt 1847: 2; van Dissel 1865: 35, 39, 48; Conradi 1844: 12–14, 21, 25, 46, 53, 55–57, 59, 61, 62). The parallel with the Upper Guinea PC suffix -dadi is evident, as in Upper Guinea PC kapasidadi, bondadi, falsidadi, etc. (Compare, by contrast, ST -dadji.) We may hypothesize that PA originally had -dadi, only to gradually have it replaced by -dat in the 18th/19th century. This replacement must have been 236 See for instance the Diccionario de la Real Academia Española (http://buscon.rae.es/draeI/SrvltConsulta?TIPO_BUS=3&LEMA=cultura) 237 See for instance the Dicionário Priberam da Língua Portuguesa (http://www.priberam.pt/DLPO/) 238 Though the entry does not appear in GBC dictionaries, the word papiamentu would be perfectly acceptable and comprehensible according to native speakers in Coimbra (Portugal).

150

Morphology

completed by the end of the 19th century, as I did not find any instances of PA -dadi in texts post-dating van Dissel (1865). It is relevant to note that modern PA -dat is not productive, so we may assume that Early PA -dadi was not either. In keeping with the assumption that functional (and thus productive) morphemes are less likely to be replaced than lexical items, the non-productive (and thus strictly lexical) status of -dadi could explain why it came to be substituted by Spanish-derived -dat, whereas the highly productive suffix PA -mentu, for instance, was not replaced by Spanish -miento. One furthermore encounters the noun ‘youth’ (< Port. mocidade; cf. Upper Guinea PC mosindádi) in van Dissel (1865). This noun did not survive into present-day PA, which has hubentut (< Sp. juventud) instead, although PA did retain the nouns moso and mosa ‘young man/lady’ (< Sp. mozo/moza or Port. moço/moça). It is also interesting to note the Early PA noun ‘care’ (< Sp./Port. cuidado) found in Conradi (1844). The change from the etymological /o/ to /i/ in is also visible in GBC kudadi (< Port. cuidado, Scantamburlo 2002: 309) and may result from analogy with nouns ending in -dadi (Nicolas Quint, p.c.). Present-day PA only has kuido (< Sp. cuido). It is important to stress, finally, that the existing literature dealing with the personal and/or professional background of the Dutch Catholic and Protestant priests, fathers and pastors (e.g. Hartog 1968; Eckkrammer 1996; Lampe 2001; Fouse 2002; Coomans & Coomans-Eustatia 2005; Bachmann 2005; Kramer 2008) provides no indications whatsoever to assume that any of these authors had additional knowledge of Portuguese. Rather, the title of the Dutch Bishop van Ewijk’s word list (1875) – “Nederlandsch-Papiamentsch-Spaansch woordenboekje” – sums up which three languages were dominant on the ABC-Islands in the 19th century as well as among the clergy and authors of these works. It is implied that the use of the Portuguese-derived suffix -dadi (whether productive or not) in 19th-century PA texts cannot be dismissed as an interference phenomenon.

4.2. Inflectional morphology As noted in the introduction to this chapter, the exact relationship between creolization and (inflectional) morphology and the degree to which one excludes the other are not yet fully understood and continue to be hotly debated.239 While 239 A recent trend in studies on creole morphology is to distinguish between inherent and contextual inflection, in keeping with the theoretical framework outlined by

4.2. Inflectional morphology

151

there is much to say for Grant’s (2001: 84) claim that in most creoles “the inflectional morphology of the lexifier language has not been transmitted”, scholars also recognize that “[t]he Portuguese-based creoles of West Africa have what appears to be an inflectional marking of the past participle” (Holm 1988: 315). Significantly, this bit of Iberian-derived verbal inflection singled out by Holm as typical of Afro-Portuguese creoles has also been retained in PA. Below, I will focus on the diachrony of PA’s past participle morpheme and demonstrate the morphosyntactic overlap with past participles in Upper Guinea PC. In a separate section (§4.3), I will address the typical application of past participles in the formation of passives in PA and Upper Guinea PC and, on the basis of 19th-century written PA records, postulate the claim that PA started out with auxiliary-less passives, a feature that is still quite characteristic of Upper Guinea PC. 4.2.1. The diachrony of PA’s past participle morpheme -/Ø/ In PA, past participles of disyllabic verbs are productively derived by shifting the stress from the first to the last syllable, i.e. by adding the suffix  -/Ø/ to the verbal root: PA kanta ‘to sing’ > kantá ‘sung’. In Upper Guinea PC, the Iberian past participle suffix -DO was retained, so that we find kanta ‘to sing’ > kantádu ‘sung’. As noted in §2.4, the infinitives of longer verbs in PA are already oxytonic and are thus homophonous with their past participles. Hence, PA batisá will mean either ‘to submit’ or ‘submitted’, according to the context. In Upper Guinea PC, this homophony does not exist: SCV botisa ‘to baptize’ > botisádu ‘baptized’. Diachronically, PA’s past participle morpheme is not a null morpheme, but rather a phonetically eroded reflex of the Iberian past participle suffix -DO. An important clue in this respect is that the Early PA texts and grammars analyzed for the present study testify to the use of an Early PA past participle morpheme -/r/, either instead of, or in variation with -/Ø/. Examples of verbparticiple pairs such as ‘to call’ vs. ‘called’ (= modern PA yamá) or ‘to make’ vs. ‘made’ (= modern PA hasí) abound in these Early PA texts and can still be found as late as in Lenz (1928). Martinus (1996: 9) therefore speaks of “the typical passing of final /d/ > /r/ of older Papiamentu” visible in “participles like coraar (…), lamantaar (…), sintir (…), konosir (…)”. For PA’s past participle morpheme -/Ø/, then, we can hypothesize an erosion path from etymological -DO via Early PA -/r/ to modern PA -/Ø/. Interestingly, this development had already been described for PA by Van Name (1869: 151): “Final d, with d which has become final by the loss of a following Booij (1993 and beyond). For examples of the relevance of this distinction to creole studies, see e.g. Kihm (2003), Luís (2010), or McWhorter (2011).

152

Morphology

vowel, either passes into r or is dropped” (cf. also Schuchardt 1882: 141). This has nonetheless largely been ignored in 20th-century scholarship on PA (with the above-mentioned exception of Martinus 1996): several scholars confronted with the frequent occurrence of word-final -r in the Early PA texts have erroneously claimed it to have been a purely orthographic tool240 or a result of decreolization241 . Tables 31, 32 and 33 are meant to further support a phonetic reduction path from Iberian -DO to Early PA -/r/ and Modern PA -/Ø/. The forms in the second column are reconstructed, but one should recall that, as mentioned above, most of these participle forms can actually be found written as such (i.e. with wordfinal ) in 19th-century Early PA records. Table 31 shows the sound change at issue: Table 31. Examples of etymological -DO > Early PA -/r/ > modern PA -/Ø/ Etymon

Early PA

Modern PA

mhline Port. corado ‘coloured, red’ (= Sp. rojo, colorado) Port. apertado ‘tight’ (= Sp. apretado) Sp./Port. rolado ‘turned’ Port. lambido ‘licked’(= Sp. lamido) *242

*korar

korá

Upper Guinea PC koradu

*pertar

pertá

pertadu

*lorar *lember *mester

lorá lembé mesté∼mester

loradu lembedu mestedu

Table 32 shows that the change from etymological -/dV/ via Early PA -/r/ to modern PA -/Ø/ also affected words other than past participles. Again, several of the second column’s forms are actually found as such in Early PA texts.

240 Some have claimed that the word-final -r must have been an orthographic tool used by Dutch authors of PA texts to indicate stress on the final syllable and thus to make up for the fact that the Dutch orthography does not include written accents of the type or . That this analysis is incorrect follows from the fact that the word-final -r also appears in texts that do have written accents (e.g. Kuiperi 1962; Sintiago 1898; Jansen 1911, etc.). Moreover, such an account does little to explain why the final -r only appears on Early PA words that etymologically end in -r, -rV, -d or -d(V), but never on any other oxytonic words ending with a vowel. 241 Wood (1972a: 22), for instance, draws attention to the occurrence of PA promer in Lenz (1928), referring to it as “the Hispanizing use of -r”. 242 This item is discussed in §4.3.5.2 & §5.7.1.3.

4.2. Inflectional morphology

153

Table 32. Examples of etymological -/dV/ > Early PA -/r/ > modern PA -/Ø/ or -/r/ Etymon

Early PA

Modern PA

Upper Guinea PC

Sp./Port. madrugada ‘dawn’ Port. tudo ‘all, every’ (= Sp. todo) Port. metade ‘half’243 Sp./Port. medida ‘measure’ Port. pode ‘[3s] be able’ (= Sp. puede) Sp./Port. criado ‘domestic servant’

*mardugar *tur *mitar *midir *por

mardugá tur mitá∼mitar midí por

mardugada tudu mitadi midida podi

*kriar

kriá

kriadu

Table 33, finally, shows that the step in PA from -/rV/ to -/r/ and ultimately -/Ø/ is similarly common, if not regular. Again, the lexemes in the second column are found in Early PA texts. Table 33. Examples of etymological -/rV/ > Early PA -/r/ > modern PA -/Ø/ Etymon

Early PA

Modern PA

Upper Guinea PC

Port. fora ‘outside’ (= Sp. fuera) Sp./Port. primero/primeiro ‘first’ Sp./Port. oscuro/escuro ‘dark’ Sp./Port. becerro/bezerro ‘calf’ Sp./Port. cachorro ‘puppy’

*for (di) *promer *sukur *bisher *kachor

fo∼for (di) promé sukú bishé kachó

fora (di) promeru sukuru bixeru katxor

The sum of the evidence suggests that PA’s past participle morpheme -/Ø/ is an eroded reflex of an early form *-du from Iberian -DO, which has been retained in full in Upper Guinea PC -du. Note that the diachronic erosion of this suffix in PA is in no way surprising, as it is and doubtlessly has always been one of the most frequently used morphemes of the PA grammar. Its productivity will be discussed in more detail in the next section.

243 Spanish mitad cannot be discarded as an etymon for PA mitar. However, the -r of PA mitar makes Portuguese mitade (typically pronounced with a plosive intervocalic /d/) a more plausible etymon than Spanish mitad (in which the word-final -d is pronounced more typically as a weak approximant).

154

Morphology

4.2.2. The regularization of past participle morphology in PA and Upper Guinea PC Regarding the formation of past participles in GBC, Kihm (1994: 243) points out that “[r]egularization has been thorough (…), so that there is no trace in Kriyol of the so-called ‘irregular’ forms such as escrito ‘written’ (Kriyol skirbidu ‘be written’), feito ‘done’ (Kriyol fasidu ‘be done’), and so forth”. The same is true for SCV and PA: past participle morphology has been regularized making the suppletive superstrate participles redundant. Consider the following coinages shared between PA and Upper Guinea PC244 : PA skirbi ‘to write’ GBC skirbi idem PA hasi ‘to make, do’ Upper Guinea PC fasi idem PA habri ‘to open’ Upper Guinea PC abri idem

> PA skirbí ‘written’ > GBC skirbidu idem = Sp./Port. escrito > PA hasí ‘made, done’ > Upper Guinea PC fasidu idem = Sp. hecho / Port. feito > PA habrí ‘opened’ > Upper Guinea PC abridu idem = Sp. abierto / Port. aberto

The regularization (and hence productivity) also manifests itself in participles whose corresponding Iberian infinitives end in -er: PA and Upper Guinea PC again coincide, giving these participles a regular stress bearing [e] instead of the irregular /i/ of the Iberian lexifiers (Quint 2000b: 146).245 Compare, for instance: PA kose ‘to sow’ SCV kose idem PA lembe SCV lembe

> PA kosé > SCV kosedu = Sp./Port. cosido > PA lembé > SCV lembedu = Sp. lamido / Port. lambido

Clearly, this high degree of productivity and regularization suggests that past participle morphology is an original part of these creoles’ grammars, an idea 244 The unusually high productivity of past participles in Upper Guinea PC has been pointed out by scholars such as Holm (1988: 274; 2008), Bal (1983a: 19), Wilson (1962: 26, 27) and Grant (2001: 81). 245 Some PA participles have two forms, one regular and one following the Iberian model, such as PA kome ‘to eat’ > komé∼komí ‘eaten’. An exhaustive listing of verbs that allow for this variation is provided by Maurer (1988: 70).

4.3. Passivization in (Early) PA and Upper Guinea PC

155

that is further supported by the fact that past participle usage is typical of all (including the most basilectal) varieties of both PA and Upper Guinea PC (unlike, for instance, gerund morphology, found only in the meso- and acrolect). To summarize, the mechanism of past participle formation is identical and equally productive in PA and Upper Guinea PC and the respective past participle suffixes, though synchronically distinct due to erosion in PA, are historically cognate. By way of contrast, neither PLQ (Schwegler & Green 2007: 287) nor Chabacano (Lipski & Santoro 2007: 385) have productive past participle morphology. In Gulf of Guinea PC, on the other hand, past participle formation with -du is productive at least in PRI (Holm 1988: 95, 96; Maurer 2009a: 93) and ANG (Maurer 1995: 91, 92), so that we may conclude that the productivity of past participles based on etymological -DO is a typical Afro-Portuguese feature. However, the use of past participles in the formation of passives appears not to be characteristic of Gulf of Guinea PC (cf. Maurer 2009a: 153 for PRI; Lorenzino 2007: 13, 14 for ANG), while in fact it is highly characteristic of both PA (Kouwenberg & Ramos-Michel 2007: 319) and Upper Guinea PC (Baptista, Mello & Suzuki 2007: 67, 68). I will therefore discuss passivization in some detail below.

4.3. Passivization246 in (Early) PA and Upper Guinea PC 4.3.1.

Passivization in present-day PA

A modern PA passive predicate consists of a TMA marker, an auxiliary verb and a past participle. A speaker of PA can choose its auxiliary from either ser (< Sp. ser)247 , worde∼wordo∼wòrdu248 (henceforth wòrdu) (< Du. worden), or keda (< Sp. quedar). Passives with keda are clearly a recent innovation

246 This section is concerned exclusively with morphologically marked passivization. I will thus not consider pseudo-passive techniques such as active sentences with a generic 3pl pronoun or with an unexpressed subject with arbitrary reference, both of which are common passivizing techniques in creoles, including in PA (Kouwenberg & Ramos-Michel 2007: 319). 247 Disregarding fossilized adverbial expressions such as PA podiser ‘maybe’ or pa ser exakto ‘to be correct’, PA ser (< Sp./Port. ser ‘to be’) has no function in PA other than that of a passive auxiliary (Munteanu 1996: 344). 248 Just as PA ser, wòrdu has no function other than that of a passive auxiliary either (Munteanu 1996: 344).

156

Morphology

(Maurer 1988: 329, cf. §5.7.2.3)249 and will not be further considered below. The auxiliaries PA ser and wòrdu are in free variation (Kouwenberg & Ramos-Michel 2007: 319) and are marked like nonstative verbs. PA’s passive can schematically be represented as follows: [TMA + Vaux + past participle]. Examples of PA’s passive are given below (61–64) involving the four basic TMA markers, ta (imperfective), tabata (imperfective past), a (perfective past) and lo (future). (61)

PA

Papiamentu no ta ser duná na skol Papiamentu not IMP Vaux given in school ‘Papiamentu is not taught in school’ (Kouwenberg & Ramos-Michel 2007: 319)

(62)

PA

e muchanan di e famianan protestant tabata the children of the families Protestant PST+IMP ser batiza Vaux baptized ‘the children of the Protestant families were (being) baptized’ (Mansur 1994: 402)

(63)

PA

E edifisionan a wòrdu restourá resientemente the buildings PFV Vaux restored recently ‘The buildings have recently been restored’ (van Putte & van Putte-de Windt 2002: 77)

(64)

PA

tur kos lo ser husgá every thing FUT Vaux judged ‘everything will be judged’

4.3.2.

(Lenz 1928: 133)

Passivization in Upper Guinea PC

PA is not seldom described as unique among Atlantic creoles for having a productive morphologically marked passive (cf. e.g. Markey & Fodale 1983: 73; Dijkhoff 1993: 19; Sanchez 2005: 76; Kriegel 2006: 131; Winford 2008: 25; Crowley 2008: 82) and passives of the type [TMA + Vaux + past participle] are indeed rare across creoles (cf. Holm & Patrick 2007: feature 11). However, in the Atlantic realm, fully productive passive morphology is typical not only of 249 Indeed, keda was for instance not yet mentioned as a passive auxiliary in Lenz (1928) or in Goilo’s ([1951]1975: 114) didactic grammar. Modern passives with keda are reminiscent of Spanish resultative constructions with quedar ‘to remain’ and accordingly occur mostly in perfective clauses modified by a (Dijkhoff 2000: 22), whereas its use in imperfective passive clauses is rare.

4.3. Passivization in (Early) PA and Upper Guinea PC

157

PA, but also of Upper Guinea PC.250 As in PA, past participles are at the heart of the Upper Guinea PC passive, the only syntactic difference being that, unlike PA, the Upper Guinea PC passive lacks an auxiliary verb. Schematically, Upper Guinea PC’s passive can thus be represented as [TMA + past participle]: (65)

a. SCV Na kel kaza ta bendedu fazenda in that house IMP be sold fabric ‘In that house, fabric is sold’ (Baptista, Mello & Suzuki 2007: 68) b. GBC kriol ta faladu li Kriol IMP be spoken here ‘Kriol is spoken here’ (Peck 1988: 147, 148)

Since the past participles in these passive constructions are marked by preverbal TMA markers just like regular verbs, they are often also referred to as ‘passive verbs’ (e.g. Abraham 2006; Holm 2008). In the remainder of this chapter I will mainly speak of ‘auxiliary-less passives’instead of ‘passive verbs’, though these are essentially one and the same thing. 4.3.3. Auxiliary-less passives in Early PA texts As hinted at above, the only syntactic difference between the modern-day PA passive and that of Upper Guinea PC is the presence versus absence of a passive auxiliary verb. As I will illustrate in detail below, however, passivization patterns found in Early PA follow the same auxiliary-less passive template [TMA + past participle] as just described for Upper Guinea PC. First, a methodological remark is in order. Since PA ta functions not only as an imperfective aspect marker (cf. §5.1) but also as an adjectival copula (cf. §5.7.2.1), the Early PA examples of the type [ta + past participle] are potentially ambiguous between a resultative (copula + adjectival participle) and a passive (imperfective aspect marker + past participle) interpretation. Conveniently though, many of the evangelical 19th-century PA texts can be put side by side with the original Dutch Bible text, so as to assure that we are indeed dealing with passives, not with resultatives. Therefore, the respective passage from the Authorized Version of the Dutch Bible251 (AVDB) is provided in a footnote whenever possible. Note, furthermore, that in the examples that fol250 On passivization in Upper Guinea PC, see e.g. Doneux & Rougé (1988: 21), Lang (1993: 158–160), M’Bodj (1993: 334), do Couto (1994: 88), Kihm (1994: 243), Scantamburlo (1999: 165, 166), Veiga (2000: 336), Baptista (2002: 112, 113), Quint (2003: 232, 233, 250), Baptista, Mello & Suzuki (2007: 67) and Holm (2008). 251 Digital source: www.statenvertaling.net

158

Morphology

low, an underscore _ marks the position where, according to modern-day PA passivization rules, we would expect the auxiliary verb wòrdu/ser. (66)

mi sanger, (. . .) koe ta _ dramaar pa moetsjoe my blood that IMP _ be shed for many ‘my blood, that is (being) shed for many’252 (Kuiperi 1862: 30, Gospel of Matthew 26: 28) b. GBC kana ta darmadu liquor IMP be spilled ‘liquor is (being) spilled’ (Scantamburlo 2002: 104)

(67)

a. EPA

ora eel ta _ plantaar deen tera when it IMP _ be planted into ground ‘when it is (being) sown into the ground’253 (van Dissel 1865: 14, Gospel of Mark 4: 31) b. SCV ta plantadu otu Polon IMP (be) planted other Polon tree ‘another Polon tree is (being) planted’ (Moser 1992: 51)

(68)

a. EPA

a. EPA

toer ees (…) deespuees ta _ tiraa na all this later IMP _ be thrown in moondi wilderness ‘all this will later be thrown into the wilderness’ (Conradi 1844: 36, Matthew 15: 17) b. SCV N ta tiradu di nha trabadju? I IMP be thrown from my work ‘Will I be fired?’254

Note that these auxiliary-less passives are not found exclusively in evangelical PA texts. The following example (69a) is from Sintiago (1898), which is a brief grammatical sketch of PA by a native speaker and assistant editor of the newspaper Civilisadó: (69)

a. EPA Ta _ pronunciá consonantenan mescos IMP _ be pronounced consonants same thing ‘Consonants are (being) pronounced the same way’ (Sintiago 1898: 8)

252 AVDB: “mijn bloed, hetwelk voor velen vergoten wordt”. 253 AVDB: “Wanneer het in de aarde gezaaid wordt.” 254 Source: http://www.mass.gov/Elwd/docs/dia/publications/ls_cape_verdean.pdf

4.3. Passivization in (Early) PA and Upper Guinea PC

b. SCV ta bendedu kuazi tudu kusa IMP be sold almost all thing ‘almost everything will be sold’

159

(Lang 2002: 159)

To be sure, the equivalents of (66a), (67a), (68a) and (69a) in modern PA must be construed with the auxiliary wòrdu/ser placed between ta and the passive participle. Unambiguous passives with the preverbal past imperfective marker tabata + past participle are similarly rife in the Early PA texts (70, 71). Again, modern PA requires an auxiliary verb in the empty slot: (70)

a. EPA

doos brigan tabata _ kroesoefikaa two thief PST+IMP _ be crusified ‘Two thieves were being crusified’ (Conradi 1844: 71, Gospel of Matthew 27: 38)255 b. SCV kantiga staba ta kantádu song ANT IMP be sung ‘The song was being sung’ (Victor Barros, p.c.)

(71)

a. EPA

eel tabata _ trokaar di figoera 3s PST+IMP _ be changed of figure ‘He was transfigured’ (van Dissel 1865: 33, Gospel of Mark 9: 2)256 b. SCV kel mininu staba ta batizadu that child ANT IMP be baptized ‘That child was being baptized’ (Victor Barros, p.c.)

Commonly attested in Upper Guinea PC are modal verbs followed by a past participle turning the subject into the logical object. I found one such example in Early PA: (72)

a. EPA

ningun hende no por _ salvá sin NEG people NEG can _ be saved without Bautismo baptism ‘no one can be saved without baptism’ (Niewindt 1837: 15) b. SCV istréla d’ oxi podê pagadu star of today can be extinguished ‘the star of today can be extinguished’ (Brito 1887: 393)

255 AVDB: “Twee dieven werden gekruisigd”. 256 AVDB: “Hij werd van gedaante veranderd.”

160

Morphology

c. GBC Kiriol pudi uzadu na tarbadju Kriol can be used at work ‘Kriol can be used at work’ (Scantamburlo 2002: 136) One noticeable difference between Early PA and Upper Guinea PC is that passives construed with a perfective past marker (PA a, Upper Guinea PC Ø) abound in Upper Guinea PC (e.g. 73b), while being rare in Early PA. This can at least in part be accounted for by the incorporation in PA of the overt marker a (< Sp. ha ‘[3s] have’) in place of the original zero (cf. §5.2 on this issue).257 Nonetheless, I found examples (73a) and (74a) which seem to qualify as a passive of the type [a + past participle] in Early PA: (73)

a. EPA

tera a teembla, i piedra nan a _ heendee. earth PFV shake, and stone+PL PFV _ be cleft ‘The earth shook and the rocks were split apart.’ (Conradi 1844: 72, Matthew 27: 51)258 b. SCV porta Ø abridu door PFV be opened ‘the door was (/has been) opened’ (Quint 2000a: 266)259

257 Assuming that the marker PA a derives from Spanish ha ‘[3s] have’, it would be unlikely for a + past participle predicates to take a passive reading, given that the Spanish equivalent ha + past participle typically renders an active reading (Markey & Fodale 1983: 70). 258 One might be tempted to analyze piedra nan a heendee as an active rather than passive clause; however, two facts justify the assumption that this clause is (semantically as well as syntactically) passive. First, the doubly written vowel in suggests word-final stress. This, in turn, suggests that is a past/passive participle (viz. passive verb) rather than an active verb. Compare, by contrast, active in that same sentence, where the singly written final vowel reflects that the final syllable is unstressed. Secondly, PA hende is a transitive verb, just as its etymon Sp. hender ‘to cleft’. The lack of a direct object in this clause is thus suggestive of a passive rather than active reading. 259 Note that (73b) cannot be read as a resultative clause, since resultative clauses in SCV are built with sta. Moreover, the adjectival participle of SCV abri ‘to open’ is suppletive abertu, not abridu (Quint 2000a: 266).

4.3. Passivization in (Early) PA and Upper Guinea PC

(74)

161

a. EPA

Na dia 30 di September di aña pasá a _ In day 30 of September of year past PFV _ poeblika e ley, pa kwal be published the law for which ‘On the 30th of September of last year, the law was published, by which’ (Proclamasjon 1863; translation from Kouwenberg & RamosMichel 2007: 319; cf. Abraham 2006: 24)260 b. SCV Ø kumpradu avion ku barku PFV be bought plain with boat ‘planes and ships were bought’ (Martinus 1996: 85)

Example (74) furthermore shows that also in terms of word order, there is overlap between passivization in Upper Guinea PC and Early PA. As noted by Baptista (2002: 113) for SCV: “In a number of cases, personal passives may involve subject-verb inversion (…) under the condition that the subject be a full NP”. Along the same lines, Abraham (2006: 24) analyzed (74a) and observed that this sentence “contains the passive verbs poeblika (publiká in modern spelling) (…) without an auxiliary or a BY-phrase. Strikingly, the passive subject e ley appears in postverbal position”. (Note that this subject-verb inversion is also visible in the examples (69a, b).) Finally, consider (75a).This Early PA passive clause lacks not only the passive auxiliary but also the overt perfective past marker a, which, together with a passive auxiliary, would be obligatory in the modern PA translation of (75a): e ku a wòrdu/ser mandá. The overlap between the Early PA variant and Upper Guinea PC is compelling: (75)

a. EPA

Eel (…), koe Ø _ mandar di soe maama he who PFV _ be sent by her mother ‘She (…), who had been sent by her mother’ (Conradi 1844: 34, Matthew 14: 8)261 b. GBC delegadus ki Ø mandadu pa Nasons Unidas delegates who PFV be sent by Nations United ‘Delegates who have been sent by the United Nations’ (Scantamburlo 2002: 105)

260 The Proclamasjon is the public announcement of the abolition of slavery on the ABC-Islands in 1863 and is accessible online: http://www.extrabon.com/edishon/edishon2006/extra2006-10-28.pdf 261 AVDB: “En zij, te voren onderricht zijnde van haar moeder” (‘And she, previously having been educated by her mother’).

162 4.3.4.

Morphology

Digression: On the reliability of Early PA evangelical texts

As noted, the evangelical PA texts from which most of the examples presented above were drawn offer an obvious advantage in that they can be compared with the source text, in this case the Dutch Bible from which they were translated, facilitating interpretation while minimizing ambiguity (cf. remarks made in this respect by Bachmann [2005: 71], van den Berg [2007: 1], and Eckkrammer 1996). On the other hand, scholars recurrently stress that one must be rather cautious when exploiting data from early creole texts, since, as e.g. Bickerton (1999: 98) notes, “[s]uch texts are produced in the majority of cases by nonnative speakers, less concerned with accurately replicating native structures than the desire to amuse or to impart information” (see also Hazaël-Massieux 2006 on the question of authenticity of early creole texts). Consequently, the evangelical sources cannot just like that be assumed to accurately reflect the usage of passive morphology in spoken Early PA; the authors Niewindt, Conradi, van Dissel and Kuiperi were certainly no native speakers of PA. However, the evangelical Early PA texts need not be dismissed as unreliable either. For instance, clues are available suggesting that the authors received assistance in writing from native PA speakers, in the person of slaves, fellow clergymen and/or church wardens. The 1825 PA catechism written by Niewindt provides one such clue: “Since this was one year after Niewindt’s arrival, it is somewhat open to question how much of this work was his own” (Fouse 2002: 128). Coomans & Coomans-Eustatia (2002) express similar doubts about the authorship: “It is surprising that at that time Niewindt was already publishing in Papiamentu. He had hardly been on Curaçao for one year and about his visit to Bonaire in November 1824 Niewindt says that he couldn’t do much because he didn’t understand the language”. They therefore assume the obvious, namely that Niewindt received assistance in writing from either the church wardens or from (one of) the nine Spanish priests already present on Curaçao at that time and probably more fluent in PA.262 In support of this view, it may be noted that, 262 At the same time, this anecdote shows Niewindt’s awareness of the need to learn PA in order to successfully spread the word of faith among the locals. As a consequence, most clergymen must have made serious efforts to learn PA once they arrived on the Antilles. Fouse (2002: 127) confirms that as early as 1769, “the Jesuit priest Rodier sent a letter to the Dutch Parliament in which he mentioned the need for priests to know Papiamentu”. Reverend Kuiperi, for instance, already preached in PA four months after his arrival on Aruba in 1858 (Coomans 2001, in the postscript to Kuiperi [1862], no page numbers). As to Niewindt, for the present study, one example, (72a), was taken from his 1837 catechism. This was 13 years after his arrival on Curaçao, and it seems likely that by that time he had mastered the language enough to write

4.3. Passivization in (Early) PA and Upper Guinea PC

163

as Hazaël-Massieux (2006: 37) points out, many early creole texts signed by ‘foreign’ white officials or clergymen were in fact written by “Creole whites (i.e. whites born in the colony) or educated mulattoes”. Note also that the authors, being clergymen, were concerned with the spread of faith amongst the (largely illiterate) working class and slave population: “[T]heir use of Papiamentu was nothing but instrumental to their social and religious power” (Broek 1988: 178). While this obviously does not guarantee the flawlessness of their writings, we may at least conclude that it would have been of little use to them to integrate new constructions – such as auxiliary-less passives – into their sermons, catechisms and Bible translations if nobody were to understand these constructions. But regardless of the exact authorship, Bickerton believes that “if a construction turns up at all in early texts, it surely must have existed at that time” (1999: 98, emphasis in original). If that is the case, we need not doubt the existence of auxiliary-less passive morphology in Early PA. And it is useful to recall the occurrence of auxiliary-less passives in non-evangelical texts such as Sintiago (1898) (e.g. 69a), showing that auxiliary-less passivization was not a phenomenon confined to the sociolect of the Dutch clergy. 4.3.5. Auxiliary-less passives (/passive verbs) in present-day Papiamentu The idea that auxiliary-less passives were widespread in Early PA is further supported by a small and apparently lexicalized set of auxiliary-less passives that appears to have been retained in present-day PA. It concerns the past participles of two speech verbs, PA papiá and bisá, as well as the past participle of the verb PA meste (PA mesté). 4.3.5.1. PA papiá / bisá Examples (76) and (77) show how the past participles PA papiá (from papia ‘to speak’) and bisá (from bisa ‘to say’) form auxiliary-less passives in modern PA yielding a reading very similar to the Iberian reflexive passives Sp. se dice que / Port. diz-se que. Again, an underscore _ indicates the position where, according to modern PA passivization rules, one would expect the auxiliary ser/wòrdu to appear. (76)

a. PA

ta _ papiá ku guera a ser deklará IMP _ be said that war PFV Vaux declared ‘it is said that the war has been declared’ (Martinus 1996: 85) (cf. Spanish se dice que / Port. diz-se que)

a catechism without assistance. Such affirmations remain tentative, however. To be complete, it is noted that Conradi arrived on Curaçao in 1837; van Dissel in 1856.

164

Morphology

b. SCV ka ta papiadu na mesa NEG IMP be spoken at table ‘no talking at the table’ (Bartens 1996: 43; cf. Veiga 1982: 175) (77)

a. PA

pasobra ta _ bisá ku e muhenan aki ta because IMP _ be said that the women here be mas salú more healthy ‘because it is said that the women here are healthier’ (Martinus 2007: 17) b. GBC I ta faladu kuma Bolama tene sesenta i sinku he IMP be said that Bolama has sixty five pursentu di analfabetus percent of illiterates ‘It is said that Bolama is sixty five percent illiterate’ (Scantamburlo 2002: 632)

To my knowledge, in modern-day PA, this auxiliary-less passive usage is limited to the past participles of papia and bisa, and does not apply to other perception or speech verbs, but more research is needed in this respect.263 4.3.5.2. PA mesté Further synchronic evidence of the use of auxiliary-less passives in PA is provided by the modal verb mesté (< Old Port. mester = Sp. menester), at least if we analyze it as a cognate of SCV mestedu, the passive verb pertaining to the active verb meste (cf. remarks on PA meste∼mesté and SCV meste-mestedu in §5.7.1.3). Etymologically, one may insist on the possibility that PA mesté descends directly from the Old Portuguese oxytonic noun mester, rather than being related to SCV mestedu. Interestingly, however, PA mesté has a paroxytonic allomorph, meste, homophonous to SCV meste. Moreover, PA meste∼mesté264 has both an active and passive meaning, ‘to need’ and ‘to be needed’, depending on the context. Both these facts (i.e. the existence of two differently stressed allomorphs and the coincidence of active and passive meaning in one verb) can historically be accounted for by assuming that PA meste∼mesté is a merger of an original 263 In place of (76a) and (77a), one may equally hear PA ta wòrdu/ser papiá and ta wòrdu/ser bisá with no change in meaning. Indeed, I assume the auxiliary-less variants to be the ‘deeper’, more original variants. 264 A third variant, ∼mester, will not be considered here.

4.3. Passivization in (Early) PA and Upper Guinea PC

165

verb/participle pair PA *meste ‘to need’ – *mesté ‘(to be) needed’, cognate with the SCV verb-participle pair meste ‘to need’ – mestedu ‘(to be) needed’. Morphologically, the derivation of the past participle PA *mesté from the verb *meste is of course completely regular, following the patterns of past participle formation described previously (e.g. PA papia ‘to speak’ – papiá ‘spoken’ = SCV papia – papiadu, etc.). Before providing examples, a word of caution is in order. As hinted at above, in SCV, the active/passive distinction has been maintained in meste ‘to need’ and mestedu ‘(to be) needed/necessary’, whereas in present-day PA, meste and mesté are mere allomorphs of one modal auxiliary, with either an active (78a) or a passive (79a) meaning, according to the context (van Putte & van Putte-de Windt 2005a: 290, 291)265 . Hence, if PA meste∼mesté is indeed related to SCV meste-mestedu, as I believe it is, it is implied that the original split between the active verb PA *meste and the passive verb *mesté was cancelled, with both forms merging semantically into one verb, meste∼mesté. (Note that the lack of a TMA marker in the following examples is due to the fact that PA meste∼mesté / SCV meste-mestedu are stative verbs.) (78)

a. PA

bo meeste tien miedoe you need have fear ‘you must be afraid’ (Conradi 1844: 10, Gospel of Matthew 5: 25) b. SCV bu ka meste ten medu you NEG need have fear ‘you must not be afraid’ (Pratas 2007: 317)

(79)

a. PA

b. PA

na teempoe koe meestee in time that be needed ‘in the time that is needed/necessary’ / ‘in due season’ (Conradi 1844: 60, Gospel of Matthew 24: 45) kiko mester hasi pa logra esaki what be necessary do for achieve this (lit. ‘what is needed to do to achieve this’) ‘what needs to be done to achieve this’ (Èxtra 21-03-2009)

265 I had fruitful exchanges with the native speakers and scholars Richard Hooi and Igma van Putte-de Windt about the use of PA meste-mesté (for which the existing literature provides no detailed description). I am indebted to both. Any errors or misinterpretations of course remain my responsibility only.

166

Morphology

c. SCV pa ofisializa kiriolu (…), so mestedu disiplina for make official creole only be needed discipline língua i kíria kondiso mental language and create condition mental ‘to make [Cape Verdean] Creole official, all that is necessary is to discipline the language and create mental conditions’ (Silva 2005: 145) If the diachronic analysis of PA mesté as a cognate and derivative of SCV mestedu is correct, this constitutes salient evidence not only of the link between the two creoles but also of the use of auxiliary-less passives in present-day PA. 4.3.6.

On the incorporation of wòrdu and ser

The data provided thus far suggest that PA started out with a passive of the type [TMA + past participle]. How this passive template subsequently developed into the ‘full’modern PA auxiliary-based passive [TMA +Vaux + past participle] can conveniently be exemplified with examples from Sintiago (1898), who alternated between auxiliary-less passives (80) and ser-based passives (81):266 (80)

EPA

Imperfecto ta _ formá coe (…) imperfective IMP _ be formed with ‘The imperfective is formed with (…)’ (Sintiago 1898: 22)

(81)

PA

Presente ta ser formá coe (…) present IMP be formed with ‘The present tense is formed with’

(Sintiago 1898: 20)

266 The texts analyzed for the present study suggest that PA started incorporating and productively using variants of Dutch worden in the second half of the 19th century. In the early documents dating from 1775, 1776 (e.g. in Maurer 1998: 203–206) and 1803 (e.g. in Martinus 1996: 33, 34) as well as in the longer texts by Conradi (1844), Niewindt (1833, 1837) and van Dissel (1865), the use of passive auxiliaries is still entirely absent. I found the first mention of a passive with wòrdu in van Dissel (1857: 127), who, in a brief two-paged grammatical sketch of PA, gives the-boat-IMP-AUX-watched ‘the boat is being watched’. (The variant is not attested in any other PA source.) Kuiperi (1862, 1864; in Coomans & Coomans-Eustatia 2005) provided the first texts with a strong tendency towards the use of passives with the auxiliary wòrdu (cf. Sanchez 2005: 161). The auxiliary ser integrated somewhat later. According to Sanchez (2008: 240) this happened in the mid-twentieth century. However, examples with ser are already found in Sintiago (1898), van de Veen Zeppenfeldt (1928) and Lenz (1928: 131–134). Lenz (1928: 132) did stress, though, that the use of ser was still rare in his days.

4.3. Passivization in (Early) PA and Upper Guinea PC

167

In other Early PA texts, similar hesitation between auxiliary-less and auxiliarybased passives is attested. Kuiperi’s (1862) catechism, for instance, is rife with wòrdu-based passives, but also provided the auxiliary-less passive in (66a). This and the examples from Sintiago suggest that for some time the two passive templates coexisted in written (and possibly spoken) discourse in the absence of a clear standard. In fact, as late as in 1928, alternation between the two passives is still attested: van de Veen Zeppenfeldt (1928: 58), in discussing the PA passive, for instance, provided the auxiliary-less clauses mi ta _ stimá and mi tabata _ stimá with the Dutch translation ‘Ik word/werd bemind’ (‘I am/was loved’). A few pages further down, the same author gives the auxiliary-based variants mi ta ser stimá and bo tabata ser stimá, with the same translations (1928: 66). The incorporation of the passive auxiliaries wòrdu and ser appears to have been triggered by a need for disambiguation on two levels. First, as noted previously, without a proper context, Early PA predicates of the type [ta + past participle] could have both a passive and a resultative interpretation, due to the ambiguous status of PA ta as both an imperfective aspect marker and a resultative copula.267 Secondly, the erosion of the past participle morpheme (discussed in §4.2.2) must have complicated the distinction between active (e.g. *mi ta bati ‘I am hitting’) and passive/resultative clauses (e.g. *mi ta batí ‘I am being hit’ / ‘I am hit’). The introduction of the passive auxiliaries wòrdu/ser cancelled this double ambiguity: active mi ta bati ‘I am hitting’; resultative > mi ta batí ‘I am hit’; passive > mi ta wòrdu/ser batí ‘I am being hit’. In addition to disambiguation, the incorporation of the auxiliaries can be explained in terms of ‘equivalence’, which is defined by Heine & Kuteva (2005: 133) as a process by which “an existing category is restructured to be equivalent to a corresponding category of the model language”.268 In the case of the PA passive, the existing passive template was in all likelihood restructured (viz. complemented with auxiliaries) in order to increase equivalence with the passive template of the model languages Dutch and Spanish.269 267 Similar ambiguity between passive and adjectival/resultative predicates is of course patent also in English (see e.g. Markey & Fodale 1983: 71) and, incidentally, is described also by Quint (2000a: 266; 2008b: 20) for modern SCV, where sta can be both a progressive aspect marker and a copula, causing confusion between passive and resultative clauses. 268 Note that there is some natural overlap between the notion of ‘equivalence’ and that of ‘gap filling’ (as applied e.g. in §3.1.5.1). 269 The degree of equivalence between the PA passive and that of Dutch and Spanish has further been increased by the generalization of BY-phrases in PA. Although it has been claimed that BY-phrases are avoided in PA (e.g. Dijkhoff 1993: 19;

168

Morphology

4.3.7.

Digression: The presumed non-nativeness of passives in PA

In the literature on PA, the view is held that its passive of the type [TMA +AUX + past participle] is a non-native feature: “in harmony with the majority of creole languages, [Papiamentu] originally did not have any passive” (Eckkrammer 1993: 140). Accordingly, morphologically marked passivization is considered a non-native feature that did not exist in PA prior to the introduction of the auxiliaries wòrdu and ser in the 19th century: “[T]he Papiamentu passive (…) is in fact (…) a syntactic structure of chiefly Spanish derivation. (…) [T]he accretion of a passive, and of a past participle inflectionally distinguished from the hitherto uninflected, invariable Papiamentu verb, is one of the principle examples of the decreolizing Hispanization of literary Papiamentu” (Wood 1972c: 859; cf. Wood 1972b: 645). Similar affirmations are easily found, for instance, in Andersen (1974: 263), Clemesha (1981: 12), Busche (1993: 80), Eckkrammer (1993: 140), Kramer (2004: 110) and Sanchez (2005: 68), so that it seems legitimate to speak of a consensus. This consensus is tangible also in the Antillean academic realm: Eckkrammer (1993: 141) speaks of “efforts of numerous Antillean linguists to eradicate passives” (cf. Maurer 1988: 331) and students are indeed generally discouraged to use passives (Birmingham 1970: 95, Richard Hooi, p.c.). The influential language purist Antoine Maduro also pleaded for avoiding passives as much as possible, as according to him “Papiamentu is a supporter of active forms” (Maduro 1971: 43–45; cf. 1991: 162). While the primary purpose of the data presented in §4.3.3 and 4.3.5 was to adduce evidence of a historical link between passive morphology in PA and Upper Guinea PC, the data additionally show that the consensus described above needs to be revised. The auxiliaries wòrdu and ser understandably give rise to suspicions of artificiality, but while these passive auxiliaries doubtlessly constitute a 19th-century borrowing, passive morphology appears to be a native, original feature of the PA grammar. 4.3.8.

Final remarks on passivization in PA and Upper Guinea PC

In addition to the data presented above, there is circumstancial evidence that passive morphology was indeed part of PA’s original morphosyntactic make-up. Eckkrammer 2004: 214), at least two types of BY-phrases are quite frequently used, one introduced by pa, the other by dor di (< Du. door ‘by’ + PA di) (Kouwenberg & Muysken 1995: 211). Just as the auxiliary PA wòrdu, the preposition dor di is used only in passive propositions, suggesting that it is a relatively recent innovation. According to Quint (2000a: 234), BY-phrases are avoided in SCV.

4.4. Final remarks on morphology

169

One such indication is, for instance, the fact that passives in PA are by no means marginal to the language but, as Wood (1972b: 645) notes, are in fact “very widespread and (…) not confined to urban or high-prestige use”. In addition, passives are frequent in written as well as in spoken discourse (Maurer 1988: 331; Busche 1993: 79). Importantly, the same observations apply to Upper Guinea PC: a random screening of any text written in either SCV or GBC suffices to yield a wealth of examples confirming Thiele’s (1991a: 73) observation that passivization in Upper Guinea PC is “impressively well developed”. In sum, unlike in most creoles, the use of passives abounds both in PA and in Upper Guinea PC and cuts across the variational continua, facts that can be taken to support the originality of passive morphology in PA and the historical-linguistic link with Upper Guinea. It is furthermore useful to point out that in both PA and Upper Guinea PC the distribution of passives is often at odds with (mainstream) Spanish and Portuguese: quite commonly, where the latter would take recourse to active reflexive constructions to convey a passive meaning, both PA and Upper Guinea PC use a passive construction. In reference to SCV, for instance, Bartens (1996: 43) observed that an utterance such SCV ka ta papiadu na mesa NEG-IMP-spokenat-table ‘no talking at the table’appears odd, “because, unlike Portuguese, there is no reflexivization but passivization” (cf. Quint 2000a: 234 for similar remarks). Note that the same observation in fact applies to most of the examples provided in this section, most of which would commonly be translated into Spanish and Portuguese by means of pseudo-passive clauses with the reflexive pronoun se. In short, unlike most creoles, both PA and Upper Guinea PC are characterized by the usage of an identical type of passive morphology in all lects and discourse, while the distribution of passives differs from the Iberian lexifiers’ passive. Not only are these strong arguments in favor of the native character of passive morphology in PA, they also underpin the hypothesized historical link between passivization in PA and Upper Guinea PC and confirm Quint’s (2000b: 206) observation that “a possible relation with the passive in SCV and GBC still shines through in PA’s participle morphology”270 .

4.4. Final remarks on morphology In §4.1, the two derivational morphemes -MENTO and -DOR were discussed and shown to constitute the two most productive morphemes in PA and GBC. 270 Original quote: “la morphologie du participe papiamento laisse encore entrevoir une possible parenté avec le passif du badiais et du CGC”

170

Morphology

In addition, the use of the suffix -dadi in 19th-century PA texts was highlighted, showing once more the value these old texts have in the research on the origins and historical development of PA. Section 4.2 outlined the correspondences between PA and Upper Guinea PC in the domain of past participle morphology, showing its remarkable productivity in both creoles and providing evidence that PA’s past participle morpheme -Ø goes back to an early suffix *-du, as still found in Upper Guinea PC. In §4.3, this past participle morphology was shown to play a crucial role in the productive formation of passives, a highly characteristic feature that binds PA and Upper Guinea PC and sets them apart from most (if not all) other Atlantic creoles. I drew mainly on Early PA texts to substantiate that Early PA was equipped with an auxiliary-less passive template very similar to Upper Guinea PC, only to complement this template with auxiliaries borrowed from Spanish and Dutch in the course of the 19th century. The compelling conclusion is that each bit of bound morphology that is characteristic of, and productive in PA (barring plural morphology, as accounted for in §3.1.5) can directly be traced back to Upper Guinea.

Chapter 5 Verbal system

Introduction Since they are creoles, it is obvious that the comparison of the verbal systems of PA and Upper Guinea PC will be largely focused on TMA marking. The usefulness of TMA markers in reconstructing earlier stages and establishing the genetic affiliation of creole languages is evident: it is widely recognized that they constitute the basic building blocks of any given creole’s grammar. Therefore, in order to test the hypothesis that PA and Upper Guinea PC are descendants from a shared ancestor creole, an analysis of their TMA markers is indispensable. This chapter aims to identify the most remarkable similarities in the domain of TMA marking in PA and Upper Guinea PC and simultaneously address and account for the most patent divergences between them. On several occasions, I will draw on Bickerton’s creole prototype in order to make predictions about the earlier stages of the TMA systems of the creoles in question. I should stress that corroborating this prototype is not an aim in itself. Like many before me, I will merely use it as a typological reference point for creole TMA marking.271 An added advantage of drawing on Bickerton (1981) is that he himself was compelled to account for some of the more conspicuous ways in which PA diverges from the prototypical creole TMA system. The first four subsections of this chapter discuss the linguistic relationship between each of the four markers that together constitute the spine of PA’s modern-dayTMA system, ta (general imperfective aspect), a (perfective past), lo (future/irrealis) and tabata (imperfective past), and their respective counterparts in Upper Guinea PC. The final three sections deal with topics that concern the verbal system as a whole. To be more precise, §5.1 focuses specifically on certain properties of PA ta that have been said to be uncharacteristic of Upper Guinea PC ta, and vice versa, in order to show that they essentially cover the same aspectual ground. §5.2 and §5.3 deal with the markers PA a and lo respectively, both of which have no direct equivalent in Upper Guinea PC. On the basis of typological data, both markers are argued to have developed after the separation 271 The relevance of the Bioprogram to the diachronic study of creole TMA systems is pointed out by Singler (1990a: x): “in almost all studies of individual pidgin and creole TMA systems since 1974, comparison with Bickerton’s prototypical TMA system is the diagnostic, the starting point from which further analysis proceeds”.

172

Verbal system

of PA from Upper Guinea PC due to more or less regular processes of language change. Following this, §5.4 focuses on the PA imperfective past markers taba and tabata and their Upper Guinea PC counterparts and presents the claim of a former [+anterior] rather than [+past] status of PA taba. The discussion of anterior marking provides a bridge to the analysis of relative vs. absolute tense marking in PA and Upper Guinea PC, in §5.5. The final two sections highlight remarkable lexical correspondences in the categories of stative (§5.6) and auxiliary verbs (§5.7) respectively.

5.1. PA / Upper Guinea PC preverbal ta For both PA and Upper Guinea PC, the morpheme ta is a central element of the verbal system performing tasks of a copula (§5.7.2.1), a future tense marker (§5.3) and a preverbal aspect marker. In this subsection, I am only concerned with the last function. To express habitual aspect272 , both PA and Upper Guinea PC take recourse to the marker ta (82) : (82)

ilha ta papia a. SCV Tudu algen di kesotu All person of that.other island HAB speak sanpadjudu sanpadjudu273 ‘All people from the other islands speak sanpadjudu’ (Lang 2002: 684) b. GBC N ta bibi binyu I HAB drink wine ‘I drink wine’ (Kihm 1994: 93) c. PA Wan ta kanta tur dia Wan HAB sing every day ‘Wan sings every day’ (Kouwenberg & Ramos-Michel 2007: 310)

The habitual marker ta exemplified here distinguishes PA and Upper Guinea PC on the one hand from Gulf of Guinea PC (which has ka; Maurer 2009b: 44), and 272 In the study of language contact (e.g. Holm & Patrick 2007: vii; Bybee et al. 1994: 151), the scope of imperfective aspect is traditionally divided into habitual and progressive aspect. Although more complex distinctions are made in the expert literature on aspect, I will here stick to the basic habitual-progressive division. 273 Sanpadjudu is the name used by natives of Santiago to designate both the people and the speech varieties of all other Cape Verde Islands (Lang 2002: 684).

5.1. PA / Upper Guinea PC preverbal ta

173

PLQ (which has asé; Schwegler & Green 2007: 279) on the other. Since this contrast largely speaks for itself, it will not be elucidated any further below. This section will rather be centered on two ways in which PA ta and Upper Guinea PC (particularly CV) ta have been claimed to be different, while in fact, as I hope to show, behaving quite similarly in these respects. In §5.1.1, I argue that, unlike what is commonly asserted in the literature, CV ta is fully operative as a preverbal progressive aspect marker, just like its PA cognate. Subsequently, the aim of §5.1.2 is to show that, just like Upper Guinea PC ta, PA ta is essentially [+imperfective] rather than [+present] as is often stated in the literature. 5.1.1. Analyzing CV ta as a progressive aspect marker The need to reassess the aspectual properties of CV ta arises in light of the fact that in the literature on CV, ta is typically categorized as [+habitual] and [−progressive] (e.g. Silva 1985: 158 & 1990: 157; Lang 1993: 150; Pereira 2000: 32; Bartens 2000: 51; Rougé 2004: 25; Baptista, Mello & Suzuki 2007: 56–58; Veiga 2000: 199). For PA, on the other hand, it is uncontroversial that ta can yield both habitual and progressive readings, depending on the context (e.g. Kouwenberg & Ramos-Michel 2007: 310). The superficial impression is thus created that PA ta and CV ta qualitatively differ in that the first is [+progressive] while the latter is [−progressive]. For example, in a recent (though brief) discussion of PA’s origins based on the comparison of TMA markers of PA, GBC, CV, Gulf of Guinea PC, and the Asian Portuguese creoles, Maurer (2009b: 44) asserts that “[t]he Cape Verdean marker ta (…) does not function as a progressive marker, in contrast to Papiamentu”. The data presented below, however, show that in certain contexts CV ta is fully operative as a preverbal progressive aspect marker and allow for the hypothesis that CV ta once operated as a general imperfective marker quite similar to its PA cognate. For PA, as noted previously, scholars agree that ta encodes not only habitual (82c), but can, depending on the context, just as easily encode progressivity (83). For the Sotavento varieties of CV (exemplified in (84) by SCV), sata, sta and sta ta are commonly put forward as progressive markers. There is no agreement among scholars on how these three variants correlate semantically and etymologically. Some see sta as a reduction of sata (e.g. Lang 2009: 167), others rather take sata to be a variant of sta ta (e.g. Silva 1985: 141); yet others analyze sta ta as a phonetic variant of sta (e.g. Suzuki 1994, in Baptista 2002: 93)274 . For the 274 See e.g. Quint (2000a: 258–260), Baptista (2002: 91–93) and Pratas (2007: 63–97) for discussions of these markers, their dialectal distribution, and their possible etymologies.

174

Verbal system

varieties of Barlavento (exemplified in (85) by the São Vicente dialect (SVCV)), ta ta and ti ta are generally provided. (83)

PA

Wan ta kanta awor-akí Wan PROG sing now.here ‘Wan is singing right now’ (Kouwenberg & Ramos-Michel 2007: 310)

(84)

SCV

Gósi-li m- sata papia ku Djom right now I PROG speak with Djom ‘I am speaking with Djom right now’ (Quint 2005: 44)

(85)

SVCV

Bo tio e k tita falá your uncle be who PROG talk ‘It’s your uncle who’s talking’

(Octavio Fontes, p.c.)

However, speakers of all varieties of CV not seldom take recourse to ta in what seem to be progressive contexts (86): (86)

a. SVCV Es ka ta falá ma bo! they NEG PRES.PROG speak with you ‘They are not talking to you’ (example, gloss and translation from Swolkien 2009: 2) b. SCV i mi li si275 tâ morrê di fome and I here thus IMP die of hunger ‘and I’m here dying of hunger / starving’ (translation of Port. eu aqui estou morrendo à fome!, in Costa & Duarte [1886]1967: 308, 309) c. SNCV amin ali tâ morrê de fôme I here IMP die of hunger ‘I’m here starving’ (translation of Port. eu aqui estou morrendo à fome!, in Costa & Duarte [1886]1967: 308, 314)

On the basis of similar examples, Bartens (1996: 40, 47) is among the few who explicitly characterized CV ta as [+progressive] (cf. Thiele 1991a: 53, 54; Fanha 1987: 304). Still, however, in the absence of a proper context, examples such as (86a–c) are open to different interpretations and by themselves do not suffice to bolster the claim that CV ta is operative as a preverbal progressive aspect marker. But more robust evidence is available. 275 SCV si < Port. assim ‘thus, therefore, still, anyway’ (Nicolas Quint, p.c.).

5.1. PA / Upper Guinea PC preverbal ta

175

5.1.1.1. Progressive ta + V complement clauses in CV and PA This evidence comes in the shape of complement clauses of the type ta + V in which ta is unambiguously [+progressive]. These complements, highly productive in CV and PA, have been referred to as ‘gerundive complements’ or ‘gerundive clauses’ by Kihm (1994: 210, for GBC, see further below) and Kouwenberg & Muysken (1995: 213, for PA) respectively. Accordingly, Muysken (2001: 404, on PA) speaks of complements in which “preverbal ta resembles English -ing (…) or Spanish -ndo”. The feature is discussed in detail below, the aim of which is twofold: first, to show that CV ta can mark progressive aspect; secondly, to illustrate the formal, functional and semantic similarities between ta + V complements in PA and CV. Note, first of all, that ta + V complements may cover the whole spectrum of imperfectivity so that their meaning may be habitual, iterative or progressive, always in function of the context (cf. Pereira 1999: 114). Since I am concerned with CV ta in its role as a progressive marker, the examples provided below aim at showing progressive ta + V complements only. I thereby opted for examples as little ambiguous as possible, basing my selection on the authors’ translations whenever one was provided and on my own judgement in all other cases. Needless to say, any semantic misinterpretations remain my responsibility. In many respects, ta + V complements behave similar to English and Iberian gerundives. For instance, in ways quite similar to English and Iberian gerundives, ta + V complements typically, but not exclusively, occur – as complements of complex predicates of the type Vaux + ta + Vinf 276 : (87)

kúme a. BCV el kumesa ta he begin IMP eat ‘he began eating (and went on eating)’ (example and translation from Meintel 1975: 220) b. PA Rosamalia a kumisa ta yorá Rosamalia PFV begin IMP cry ‘Rosamalia started crying’ (Lenz 1928: 128)

276 Pending further investigation, it seems that, in principle, any infinitive-governing verb can take the AUX position, though it is typical for both in PA and CV to find such predicates headed by verbs meaning ‘to start’ or ‘to stop’ as well as movement verbs and verbs meaning ‘to continue, keep on’.

176

Verbal system

(88)

a. BCV El fika ta kuzinha, ta lava ropa He stay IMP cook IMP wash clothes ‘He kept cooking, washing clothes’ (example and translation from Baptista 2007a: 75) b. PA kachó a keda ta hou dog PFV stay IMP bark ‘the dog kept on barking’ (Allen 2007: 239)

– in accusativus cum infinitivo constructions, as complements of perception verbs: (89)

a. SCV N xinti algen ta pintxa pórta! I feel person IMP push door ‘I felt somebody pushing the door!’ (Lang 2002: 671) b. PA El a sinti e logá ta sagudí he PFV feel the place IMP shake ‘He felt the place shaking’ (Maurer 1988: 268)

(90)

a. SCV el obi Manel ta sibia he hear Manel IMP whistle ‘he heard Manel whistling’ (Lang 2002: 511) b. PA el a topa Koma Baka ta kana buska he PFV meet Mother Cow IMP walk search kuminda food ‘he met Mother Cow looking for food’ (Maurer 1988: 264)

– as adjectival complements: (91)

a. SCV Anton, (…) ku gentis tudu ta kume (…) Thus with people all IMP eat ‘Thus, with everybody eating (…)’ (Lang 2002: 576) b. PA e plantashi ta yen di hende ta kòrta (…) the plantation be full of people IMP cut tabaku tobacco ‘the plantation was full of people cutting (…) tobacco’ (Maurer 1988: 264)

5.1. PA / Upper Guinea PC preverbal ta

177

– as adverbial complements: (92)

a. SCV El fika la, ta skuta-m. he stay there IMP listen+me ‘He stayed there, listening to me.’277 (example and translation from Baptista 2002: 78) b. PA Nanzi a keda atras ta pensa Nanzi PFV stay behind IMP think ‘Nanzi stayed behind thinking’ (Maurer 1988: 264)

(93)

a. SCV N sta na kasa ta I be in house IMP b. PA mi ta na kas ta I be in house IMP ‘I am at home studying’

(94)

studa study studia study

(Bernardino Tavares, p.c.)

(Rigmar Haynes, p.c.)

a. SCV e ten tres noti di fiu ta toka unbes he have three night in a row IMP play music nonstop lit. ‘he has three nights in a row playing music nonstop’ ‘three nights in a row, he has been playing music nonstop’ (Lang 2002: 776) b. PA Marcel tin basta ratu ta kome djente Marcel have quite a while IMP eat tooth lit. ‘Marcel has quite a while eating teeth’ ‘Marcel has been grinding his teeth for quite a while’ (Maurer 1988: 265)

It is interesting to notice that Lang (2002, for SCV) and Maurer (1988, for PA) have independently described the stative possessive predicate structure exemplified above in (94): – SCV ten diterminadu ténpu (ta/na) fase algun kusa lit. ‘to have an X amount of time doing s.th.’ > ‘to be doing something for a certain amount of time’ (Lang 2002: 776); – PA tin X tempu ta hasi algun kos lit. ‘to have an X amount of time doing s.th.’ > ‘to be doing s.th. for an X amount of time’ (Maurer 1988: 265; this structure is also discussed by Muller 1989: 235, 238). 277 Note that in (88a,b), fika/keda serves as an auxiliary meaning ‘to keep on’, while in (92a,b), it is a full verb meaning ‘to stay, remain’. An ambiguous reading can be avoided by the insertion of locative complements, as in (92a,b), or temporal complements, which would render a continuous (‘to keep on’) reading of fika/keda.

178

Verbal system

Note, furthermore, that both in PA and CV, the ta + V complements can occur in series expressing simultaneity of events. In such cases, there is normally no conjunction linking the complements (95): (95)

a. FCV el ta ficâ na ratrato ta abri odjo ta fitchâ, he IMP stay in picture IMP open eye IMP close ta abri boca ta fitchâ IMP open mouth IMP close ‘he appears on the pictures opening and closing his eyes, opening and closing his mouth’ (Macedo 1979: 205) b. PA Tur dia mi ta tendé-bo, ta pasa ku piská, All day I IMP hear.you IMP pass with fish ta grita riba kaya. IMP yell in street ‘Everyday I hear you passing by with fish yelling in the street.’ (Lenz 1928: 265; cf. examples in Maurer 1988: 262, 264)

As noted, it is not the case that any given ta + V complement necessarily yields a progressive reading; they may, if the context requires it, just as well express habitual or iterative aspect. The main point to grasp is rather that such complements (and thus CV ta) have the capacity to convey progressive aspect. Further evidence of the capacity of CV ta to mark progressive aspect is that the presence versus absence of ta in an auxiliary’s verbal complement allows to distinguish between an ongoing [+progressive] process (96) and a [+punctual] (97) event respectively. Interestingly, Pereira (1999: 112) and Maurer (1988: 267) have pointed this out independently for SCV (96a, 97a) and PA (96b, 97b). To facilitate comparison, I will provide the authors’ French translations: (96)

a. SCV E ba ta fase si kasa he go IMP make his/her house ‘Il a construit sa maison petit à petit (progressivement)’ (Pereira 1999: 112) b. PA M’ a mir’ é ta kap e palu I PFV see he IMP cut the tree ‘Je l’ai regardé en train d’abattre l’arbre’ (Maurer 1988: 267)

(97)

a. SCV E ba _ fase si katchupa he go make his/her catchupa ‘Il est allé faire sa catchupa’

(Pereira 1999: 112)

5.1. PA / Upper Guinea PC preverbal ta

b. PA

M’ a mir’ é _ kap e palu I PFV see he cut the tree ‘Je l’ai regardé abattre l’arbre’

179

(Maurer 1988: 267)

Note that in SCV (and possibly in other varieties of CV), the locative preposition na can occasionally replace ta in progressive ta + V complements (J. Lang 2000: 27; 2002: e.g. 557, 580, 675, 717, 776). Significantly, however, the markers canonically presented as CV’s primary progressive markers (e.g. sata, sta and sta ta) cannot (cf. Quint 2010: 141). The particularities of progressive ta + V complement clauses have been discussed in some detail for PA by Kouwenberg & Muysken (1995: 213–214), Andersen (1990: 73–75), Muysken (2001), and particularly Maurer (1988: 262– 269). For ta + V complements in CV, Pereira (1999) provides an interesting and quite detailed analysis, while several other authors have furthermore referred to the phenomenon in passing. Some examples: – Meintel (1975: 220, 220f.n.) discusses the “the uses of ta to express progression”, after having drawn attention to the occurrence of progressive ta + V complements in the Brava variety of CV (87a); – With respect to the predicate bira ta perdi turn-IMP-disappear ‘to start disappearing’, Quint (2010: 141) mentions that, where one would expect the progressive marker sta or sata, “only the imperfective marker ta can be used to express the fact that the action expressed by the verb (perdi) is taking place”278 (cf. Quint 2003: 139 & 2005: 43); – Using the phrase CV el sei ta ri he-leave-IMP-laugh ‘he left laughing’ as an example, Almada (1961: 123) observes that the Portuguese gerund morphology was not integrated in CV; instead, “it is the construction of the infinitive preceded by ta which is used, this rule being more within the spirit of the creole”279 ; – Pratas (2007: 41) provides the accusativus cum infinitivo construction N odjau ta fuma I-see-you-IMP-smoke ‘I saw you smoking’ accompanied by the following comment: “In embedded contexts which look like non-finite environments, ta is (…) needed, yielding some value of Progressive” (cf. Pratas 2010).

278 Original quote: “só a marca de imperfectivo ta pode ser utilizada para expressar o facto de que o processo expresso pelo verbo (perdi) está a acontecer” 279 Original quote: “é a construção do infinitivo precedido de tâ a usada, regra essa que está mais dentro do espírito do crioulo”

180

Verbal system

In fact, the type of progressive complements discussed here seems to be common across creoles, as has been pointed out by Bickerton (1981: 99–104) with particular respect to complements of perception verbs. After discussing this structure in Guyanese Creole, he stated: “Similar structures are found in other English creoles, such as Belize Creole (…); in French creoles, such as Haitian Creole and Guyanais; and in Portuguese creoles like Sãotomense” (Bickerton 1981: 103). Recently, Maurer (2009a: 85, 86), has described the feature under discussion for PRI, the big difference being that the progressive aspect marker is ta in PA and CV, but sa in PRI: Ê vika sa gô he-come-IMP-cry ‘He came crying’ (Maurer 2009a: 86). In addition to the use of CV ta in the progressive complements discussed above, there are other synchronic indications that it can mark progressive aspect, such as its use in a series of compound progressive markers to be discussed next. 5.1.1.2. The use of ta in compound progressive markers in CV With the progressive ta + V complements in mind, it is interesting to look at (the origins of) the stock of progressive markers found throughout the CV dialect cluster, such as SCV sata, sta ta, SVCV ti ta, ta ta, or the past forms SCV staba ta, FCV stá ta, BaCV tá ta, taba ta ∼ tava ta and tive ta ∼ tibe ta.280 Though the variation is considerable, all these markers have in common that the second (and truly preverbal) element is, indeed, ta. The variation is caused by the first element, which in all cases – except perhaps sa in sata, a particular case to which I will return later – corresponds to a locative copula (or past/anterior form of a copula) in the respective variety of CV. This is unsurprising, considering that locative copulas are regularly selected as auxiliaries for periphrastic progressives (e.g. Bybee et al. 1994: 129–131; Heine & Kuteva 2002: 276–282). In other words, the true progressive aspect marking morpheme in these complex markers is ta, while the first morpheme is a (locative) copula: Vcop + taprog . In the literature on CV, these complex markers are still commonly characterized or glossed as single units or as the combination of two TMA markers. However, such a characterization does not account for the fact that all the complex progressive markers (again, with the exception of sata) are discontinuous: e.g. (a) SCV El sta senpri ta kume he-be-always-IMP-eat ‘He’s always eating’ (Baptista 2002: 93) or (b) SVCV Um tava so ta falá I-be+PST-only-IMP-speak 280 The following sources were consulted: Lopes da Silva (1957: 139, 140); Almada (1961: 111, 112); Costa & Duarte (1967: 273); Morais-Barbosa (1975: 138); Cardoso (1990: 65, 66); Pereira (2000: 33f.n.); Bartens (2000: 51); Veiga (2000: 198); Quint (2000a: 258–260; 2000b: 75, 76); Baptista (2002: 91–93); Pratas (2007: 63– 97).

5.1. PA / Upper Guinea PC preverbal ta

181

‘I was merely chatting’ (Octavio Fontes, p.c.).281 Thus, although it is not uncommon to find sta in (a) and tava in (b) glossed as TMA markers, this is at odds with the fact that creole TMA markers occur before verbs rather than before adverbs; clearly, sta and tava here function as (locative) copula, whereas preverbal ta marks progressive aspect. The foregoing analysis implies that progressive predicates such as sta ta + V ‘to be V-ing’ should be analyzed on a par with the periphrastic Vaux + taprog + V predicates described in §5.1.1.1. The complex progressive aspect markers found throughout the CV dialect cluster thus provide additional synchronic evidence for a [+progressive] status of CV ta (cf. Meintel [1975: 220] and particularly Pratas [2007: 63–65, 87] for similar analyses). Finally, with the discontinuity of the CV complex progressive markers in mind, the comparison of the commonly heard string PA ta bezeg ta + V (bebusy-IMP + V) ‘be busy V-ing’ (with PA bezeg < Du. bezig ‘busy’; cf. Joubert 1999: 87; van Putte & van Putte-de Windt 2005a: 62), exemplified in (98a), with an example of discontinuous ta ta in São Nicolão CV (98b) is compelling: in both cases, ta1 serves as copula/auxiliary, while ta2 marks the complement verb for progressive aspect: (98)

a. PA

e ta bezig ta traha he be busy IMP work ‘he’s busy working’ b. SNCV Kacor ta mitid ta ladra dog be occupied IMP bark ‘The dog is busy barking’

(G. Lang 2000: 93)

(Cardoso 1990: 65)

5.1.1.3. On the origin of SCV sata One of the particles most typically found marking progressive aspect in main clauses in SCV is sata (Quint 2000a: 240; Rougé 2004a: 151; Pratas 2007: 63; Lang 2002: 670). As hinted at above, unlike the compound progressive markers discussed there, the marker sata is not discontinuous and cannot (or, as I will argue, no longer) be analyzed as Vcop + taprog . Synchronically, the first element, sa282 , only occurs in this very combination, i.e. prefixed to ta 281 The discontinuous character of the complex progressive markers is pointed out, for instance, by Cardoso (1990: 65), Baptista (2002: 87, 93), Solovova (2004: 164, 166) and Baptista, Mello & Suzuki (2007: 56). 282 Scholars do not agree on the etymology of sa. While most seem to believe it is a remnant of Port. estar (e.g. Lopes da Silva 1957: 139; Bal 1983a: 21, 23; Baptista 2002: 94; Lang 2002: 670; Hagemeijer 2007: 58f.n.), others analyze it as a cognate

182

Verbal system

(e.g. Lang 2002: 670): e sata kume senpri ‘he is always eating’, *e sa senpri ta kume. Diachronically, however, sata needs not be excluded from the analysis provided above: sa still functions as a locative copula in the conservative Casamance variety of GBC – si kása sa na metádi tera his-house-be-in-middle-of-village ‘his house is in the middle of the village’ (Bal 1983b: 35) – and appears as such in clearly progressive clauses – e sá na kõpo kása he-be-in/at-build-house ‘he is building a house’ (1983a: 23).283 It is thus possible that, in an early stage, sa also served as a locative copula in SCV. We can then hypothesize a process in which a frequent use of the periphrasis sa + taprog + V ‘to be V-ing’ led to the cliticization of sa to taprog and the crystallization of sata as a synthetic progres-

of Port. ser (Rougé 2004a: 261) or its 3pl present form são (Holm 1988: 181); yet others affirm that sa represents a fusion of ser and estar (Lipski 1993: 218). The use of sa as a locative copula in GBC (Bal 1983b: 23, 35; Noël Bernard Biagui, p.c.) seems to favor Port. estar as the etymon rather than Port. ser, the latter not being a locative copula. The fact that, cross-linguistically, locative copulas are particularly suitable for building progressives (Bybee et al. 1994: 128–130) also favors Port. estar as the etymon of sa, as it would account for its auxiliary function in progressive SCV sata + V and GBC sa na + V clauses. Notice in this respect that, in GBC and SCV, the copulas known to derive from Port. ser (GBC sedu/i and SCV ser/e) are not used as an auxiliary in progressive periphrases, while GBC/SCV sta (< Port. estar) is. Note, furthermore, that sa is used as a portmanteau copula in all varieties of Gulf of Guinea PC (Lipski 2002: 66). In addition, many scholars have drawn attention to the occurrence of sa(r) as a portmanteau copula verb (i.e. with properties of both ser and estar) in several 15th-16th century Língua de Pretu texts (Lipski 1993: 218; Rougé 2004a: 261). Lipski (2002) exhaustively documents the attestations of sa (and its variants) in these texts. One may of course speculate about a possible transfer of sa from the Língua de Pretu (viz. reconnaissance language) to the Portuguese-based creoles. See Lipski (2002) for an interesting discussion of this issue. 283 To my knowledge, the productive use of sa as a copula in Casamance has only been documented by Bal (1983a, b). Although Rougé (2004a: 261) mentions the existence of sa as a copula in Casamance, he suggests that its use is restricted to fixed expressions such as I sa ‘that is, in other words’. Noël Bernard Biagui (p.c.), a native speaker of the Casamance variety, confirmed the synchronic productive use of sa as locative copula, as documented by Bal. This seems to contradict Lipski (2002: 72), who speaks of “la ausencia de sa/sã en los criollos de Cabo Verde y Guinea-Bissau”.

5.1. PA / Upper Guinea PC preverbal ta

183

sive aspect marker in SCV, while sa did not survive as an autonomous copula verb, at least not in Santiago (cf. Rougé 2004a: 261 for a similar point).284 To close, we may take the opportunity to discuss two progressive markers of unascertained etymology found in the speech of elderly speakers of SCV, áta and áita (Quint 2000a: 259 & forthcoming; Lang 2002: 10). In his forthcoming Atlas de la langue capverdienne, Quint reports on the use of these two progressive markers in almost two thirds of the island of Santiago, though mainly in the speech of elderly people. As to the origins of these markers, it is arguable that they are construed of taprog with the prefixation of a locative adverbial element *a∼ai, which is otherwise unattested in SCV, but may well derive from the Portuguese locative adverb aí ‘there’. This hypothesis connects well with the cross-linguistic tendency of progressives to convey “some sense of location” (Bybee et al. 1994: 130; cf. Holm 2000: 180, 181). Incidentally, ata exists also in PA, as a locative (ad)verb: ata bo kòfi ‘here is your coffee’; ata mi akí ‘here I am’ (van Putte & van Putte-de Windt 2005a: 46). 5.1.1.4. Final remarks on CV ta as a progressive aspect marker The synchronic distribution of CV ta as a habitual and progressive aspect marker in all CV dialects clearly points towards an early status of CV ta as a general imperfective marker, quite like the portmanteau nonpunctual marker once identified by Bickerton (e.g. [1974]1980: 6) as a prototypical creole feature. Interestingly, as early as the late 19th century, Schuchardt (1882, in Hagemeijer & Holm 2008: 148) had expressed similar thoughts: “Originally, CV N ta da means, in effect, Port. eu estou dando ‘I am giving’. However, this meaning has blurred to Port. eu dou ‘I give’”.285 A scenario can thus be unfolded in which CV ta started out as a general imperfective marker [+habitual, +progressive] (cf. also Pereira 2000: 33). Gradually, then, alternative periphrastic progressives would come to absorb parts of progressive aspect in CV. If this scenario is correct, and if we note that CV ta also marks future tense (e.g. Baptista 2002: 79), an interesting parallel can be drawn with the Dravidian language Kui. In Kui, Bybee et al. (1994: 156) note, one single form was once used “for habitual, progressive, and future. With the 284 Note that this diachronic analysis of sata supports the orthographic tendency to separate sa from ta by a space, as advocated e.g. by Lang (1993: 150) and others. 285 Recognizing the status of CV ta as a general imperfective – and hence as [+progressive] – is in harmony with this marker’s ability to mark future tense: imperfectives and progressives developing into future tenses is an unmarked phenomenon in creoles (e.g. Parkvall 2000: 84; Holm & Patrick 2007: vii) and non-creoles (Dahl & Velupillai 2005c; Bybee et al. 1994: 276, 277).

184

Verbal system

development of a periphrastic progressive, this older form came to signal just habitual and future”. The parallel is striking, though with this difference: unlike the Kui marker, CV ta has maintained a good portion of its original [+progressive] qualities. The central aim of this section was to demonstrate that CV ta is fully operative as a progressive aspect marker in complement clauses and as such functions in ways very similar to its PA cognate. The development in CV of alternative, more emphatic progressive aspect markers such as sata alongside ta is by no means exceptional: it is quite common for a language to develop more than one way of encoding progressive aspect (Bybee et al. 1994: 129). In fact, PA provides a case in point: although in PA, ta is the only preverbal progressive aspect marker, the Iberian gerund morphology (suffix -ndo, e.g. mi ta papiando ‘I am talking’) is becoming increasingly popular as a way of marking progressive aspect more emphatically (cf. Sanchez 2000; Kouwenberg & Ramos-Michel 2007: 310). 5.1.1.5. Digression: Progressive aspect in GBC Also in the sister creole of CV, GBC, ta principally marks habitual aspect, progressive aspect being more typically encoded with na (99). GBC naprog results from the reduction of the periphrastic marker sta na286 ‘be in/at’ (Rougé 1994: 147), which is still attested in full in the early description of GBC by Bertrand-Bocandé (1849: 75) as well as in the speech of elderly people (Rougé 1993: 323).287 (99)

GBC

N na bibi binyu I PROG drink wine ‘I am drinking wine’

(Kihm 1994: 93)

Like in CV, however, clear traces can be found of a previous stage in which GBC ta served as a general imperfective. To start, Peck (1988: 273) mentions that “we have found ta to have a progressive meaning in some cases”, providing example (100):

286 In addition to encoding progressivity, GBC na functions as a future marker. Analogously to the development of GBC na out of sta na, the Haitian French Creole progressive aspect and immediate future marker ap results from the erosion of the preposition après ‘after’ which in turn was used in the verbal periphrasis être après a + V ‘to be (at) V-ing’ (DeGraff 2007: 104). 287 Also in SCV, periphrastic sta na can be used to mark progressivity and is occasionally reduced to na, though without reaching the levels of grammaticalization of GBC naprog (Quint 2000a: 264–265).

5.1. PA / Upper Guinea PC preverbal ta

(100) GBC

nho parbai ta chora suma minina Mr. Parbai IMP cry like child ‘Mr. Parbai was crying like a child’

185

(Peck 1988: 273)

More significantly, just like in CV and PA, GBC ta serves as a preverbal progressive aspect marker in complement clauses (101). Note that not only example (101) but also its translation is Wilson’s: (101) GBC

i kumsa ta coora he begin IMP cry ‘he began to cry (and continued crying)’ (example and translation from Wilson 1962: 22)

On the basis of this example, Wilson (1962) drew the conclusion that “ta has progressive meaning when used after an auxiliary verb” (Wilson 1962: 22; cf. Boretzky 1983: 130), which corresponds to the situation described above for CV ta. The use of progressive ta + V complements is confirmed by examples in Kihm (1994: 122) and Scantamburlo (1981: 74; 1999: 159; 2002: e.g. 445), although all these examples involve inchoative predicates headed by the auxiliary GBC kumsa∼kunsa (cf. §5.7.3.1). With other auxiliaries, na + V complements appear to be preferred (cf. e.g. Kihm 1994: 210, 211). Interestingly, for Casamance, Noël Bernard Biagui (p.c.) reports that periphrases of the type Vaux + taprog + V (e.g. bida/fika/sinta ta + V ‘start/stay/sit V-ing’) can still be heard in the speech of elderly people, but that variants with na are far more frequent nowadays. The comparison of (101) with GBC kumsa na cora provided by Rougé (2004a: 113) suggests that, indeed, na is on its way to fully displace ta as progressive marker, also in progressive complement clauses. Finally, it is interesting to note that in 19th-century GBC, the periphrastic marker sa ta was still in use (Barros 1899: 274, 275; Rougé 1999): “[I]n the description of GBC by Barros (1887–1908), na appears as the marker used in the areas outside the forts, whereas sa ta (…) is associated with the ‘inside’ variety” (Solovova 2004: 161, drawing on Rougé 1999: 53). The diachrony of progressive aspect marking in GBC can thus be concisely summarized as follows: in line with Kihm (1994: 92), it is plausible to assume that, “in some previous stage of the language [GBC], ta covered the whole field of the imperfective”. Periphrastic sa ta and (sta) na emerged subsequently to mark a more emphatic progressive aspect. Of these two, na has come out most dominantly, assuming considerable parts of the domain of progressive aspect marking in present-day GBC (cf. Bartens 1996: 69 for similar remarks).

186

Verbal system

5.1.2. Analyzing PA ta as [+imperfective], rather than [+present] To further support the view that PA ta and Upper Guinea PC ta stem from a shared proto-form, this subsection reassesses the status of PA ta as a TMA marker. In the literature on PA, ta is traditionally either explicitly labeled as a present tense marker or simply glossed as such. Upper Guinea PC ta, on the other hand, is without exception classified and glossed as an aspect marker (e.g. Lang 2002: 762; Rougé 2004a: 25; Kihm 1994: 92; Quint 2000a; Baptista 2002; Baptista, Mello & Suzuki 2007, etc.). From this one might be inclined to infer that PA ta qualitatively differs from Upper Guinea PC ta in that the former primarily marks tense ([+present]) while the latter primarily marks aspect ([+imperfective] or [+habitual]). This impression is reflected, for instance, in the back-to-back descriptions of the two creoles in Holm & Patrick (eds. 2007), where ta is glossed as a (present) tense marker for PA (label ‘non-past’, Kouwenberg & RamosMichel 2007) but as a (habitual) aspect marker for Upper Guinea PC (Baptista, Mello & Suzuki 2007). In keeping with e.g. Andersen (1990, 2000) and Faraclas, Rivera-Castillo & Walicek (2007), the remainder of this section argues that PA ta is not genuinely [+present], but genuinely [+imperfective], and as such should be categorized on a par with Upper Guinea PC ta as an aspect rather than as a tense marker288 . This is of course not to say that PA ta does not correlate with present tense; it undoubtedly often does. This comes forth mainly from the fact that whenever a present situation is described, the imperfective marker ta is likely to appear, since a present situation can by definition not be perfective (Bybee et al. 1994: 126). However, as I aim to show below, the reverse is not true: whenever ta appears, 288 In this section, following Comrie (1985: 6), the category tense is taken to convey the temporal location (e.g. past/anterior, present or future) of a certain predicate, whereas aspect expresses the internal temporal constituency (imperfective, perfective, etc.) of that predicate. For the sake of convenience, I will use the terms tense and aspect in opposition to each other. Of course, in reality the categories of tense and aspect overlap conceptually in as far as both concern the expression of temporal relations. For instance, present situations are inherently imperfective and perfective actions are inherently past (Bybee et al. 1994: 126; Dahl & Velupillai 2005a). This semantic overlap is one of the reasons why, as Dryer (2005 WALS feature 69) observes, “it is frequently difficult to determine from descriptive grammars whether a category ought to be considered tense or aspect. Different descriptions of the same language often differ in whether they characterize a category as one of tense or as one of aspect”. Suffice it to note that the same caveats are in order in the current discussion of the preverbal marker PA / Upper Guinea PC ta.

5.1. PA / Upper Guinea PC preverbal ta

187

the situation described is not necessarily present, but can just as well be past (or even future, cf. §5.3). What all predicates marked by ta have in common, then, is not that they occur in the present, but that the situation or action described is imperfective. 5.1.2.1. PA ta in complement and subordinate clauses with a past reference One example of PA ta marking imperfective aspect in sentences with a past reference has of course been reviewed in the previous subsection: the ta + V complements discussed there confirm ta in its role as what Andersen (1990: 73) calls a “tense-neutral imperfective aspect morpheme”. In these complements, ta merely marks the complement clause for [+imperfective], whereas the main clause can be either present tense (e.g. PA e ta bezeg ta traha he-be-busy-IMPwork ‘he is busy singing’), past perfective (e.g. PA el a kumisá ta kanta he-PFVstart-IMP-sing ‘he started singing’), past imperfective (marked by tabata: e.g. e tabata bezeg ta traha he-be+PST-busy-IMP-work ‘he was busy working’) or future (lo mi ta bezig ta kanta FUT-I-be-busy-IMP-sing ‘I’ll be busy singing’), with respect to the moment of speech. Clearly, this state of affairs underlines the status of ta as a tense-neutral imperfective aspect marker. Note, furthermore, that in past-tensed clauses with an aspectual perfectiveimperfective contrast, such as PA el a kumisá ta kanta he-PFV-start-IMP-sing ‘he started singing’, the replacement of ta by the past imperfective marker tabata is unacceptable (Maurer 1988: 269). A similar perfective-imperfective contrast is found also in indirect speech of the type He said that (…). Here too, the subordinate clause, if expressing an imperfective action, is obligatorily marked by PA ta, not by tabata, even though the reference time is past. As Maurer (2003: 244) puts it, “in perfective pasts, there is no tense agreement when the governing verb is a verb referring to a speech act”289 . An example would be: PA El a bisa ku e ta (*tabata) bai he-PFV-say-that-he-IMP-go ‘He said he was going’ (Maurer 2003: 244). These – certainly frequent – uses of ta in complement and subordinate clauses with a past reference are clearly at odds with the common practice of labeling and glossing PA ta as a present tense marker, which can be illustrated with paradoxical affirmations such as: “the present tense can also refer to a past situation”290 (Maurer 1998: 163). Note, furthermore, that Kouwenberg & Ramos-Michel’s (2007:e.g. 311) ‘non-past’ label does not obtain either. 289 Original quote:“Dans les complétives, il n’y a pas de concordance des temps lorsque le verbe régissant est un verbe renvoyant à un acte de parole” 290 Original quote: “el presente (…) puede también referirse a una situación pasada”

188

Verbal system

5.1.2.2. The absence of ta before stative verbs A second clear indication that ta is primarily [+imperfective] rather than [+present] lies in the fact that in PA (as well as in Upper Guinea PC and most other creoles) a series of stative verbs do not take ta, also in clearly present situations, e.g. PA mi (*ta) tin dos yu = SCV N (*ta) ten dos fidju I-have-two-child ‘I have two children’. The reason is that stative verbs are imperfective by nature, so that an imperfective aspect marker is superfluous. Clearly, if ta were a true present tense marker, the omission of ta before stative verbs would be unaccounted for. Rather, this reaffirms that ta represents what Bickerton ([1974]1980: 6) referred to as “a marker of nonpunctual aspect (…); this marker cannot normally cooccur with state verbs” (cf. Andersen 1990: 71, 72 for a similar argument). (The stative-nonstative distinction will receive more attention in the remainder of this chapter.) 5.1.2.3. The use of ta in narrative contexts Also indicative of ta’s [+imperfective] rather than [+present] properties is its use in stories and descriptions of past situations. In such narrative discourse, a past time reference is established either by means of temporal adverbs or by using a past/anterior marker at the onset. In other cases, the topic of conversation is already clearly linked to the past, so that no overt past marking is needed. Andersen (2000: 360, 361) sums up: “[O]nce the speaker is confident that the time frame is shared by all participants in a conversation, the speaker can suspend explicit marking of (…) tense (…) in subsequent clauses, regardless of whether the clauses are independent or dependent (…).” Examples (102–104) show ta as a tense-neutral aspect marker in story telling in PA and SCV: bisa, ku e ta bebe hopi nan tabata they PST+IMP say that he IMP drink much ‘They said he drank a lot’ (Maurer 1998: 163) b. SCV nu tenha xefri, ki ta leba libru we had leader who IMP bring book ‘We had a leader who would bring books’ (Baptista 2002: 78)

(102) a. PA

5.1. PA / Upper Guinea PC preverbal ta

189

(103) a. PA

Tabatin hende ku ta koba salu ku piki had people who IMP dig salt with pickaxe Nos ta yena makutu ku salu. (…) Despues we IMP fill basket with salt then nos ta pone salu na monton. we IMP put salt in pile ‘There were people who dug salt with a pickaxe. We would fill the baskets with salt. Then the salt was placed in piles.’ (Allen 2007: 72) b. SCV No mãi ta daba nos katchupa. E ta Our mom IMP give+ANT us cachupa. She IMP kaba, e ta chinta, e ta tchora stop she IMP sit she IMP cry ‘Our mother gave us cachupa [typical Cape Verdean dish]. She stopped, sat down, and cried’ (Fanha 1987: 301)

(104) a. PA

Den kunukunan (…) nan ta pag’é in plantation+PL they IMP pay+him trinta sens. Nan ta dun’é un kan’i hariña (…). thirty cents they IMP give.him a tin+of flour Esei ta e tempunan That be the times ‘On the kunukus [plantations] of the big landowners they’d pay him thirty cents. They’d give him a tin of flour (…). Those were the days’ (Andersen 1990: 81) b. SCV Nu ta konbersa tudu dia, Brankinha ta konta-m, we IMP talk all day Brankinha IMP tell-me N ta konta-l I IMP tell-her ‘We talked every day, Brankinha would tell me things, I would tell her things.’ (Baptista 2002: 78)

Maurer (2003: 241) confirms that in narratives, “ta refers to both progressive and habitual situations, present as well as past”291 , a fact he had already drawn attention to in earlier work (Maurer 1988: 312–322). Furthermore, in keeping with Bickerton (1981), who spoke of “the hard-dying creole myth about ‘narrative tenses’ and ‘historical presents’” (1981: 84), Maurer (2003) makes the valid observation that the narrative use of ta with past reference is not a par291 Original quote: “ta renvoie à une situation progressive ou habituelle tant présente que passée”

190

Verbal system

allel to the historical present of Indo-European languages. For instance, while aspectual relations are neutralized by the European historical present, this is not the case in PA narratives, which instead maintain a clear aspectual perfectiveimperfective opposition: PA un dia, e hòmber ta bini atardi ku su bestianan, el a tende un kehamentu den un buraku one-day-the-man-IMP-come-eveningwith-his-animals-he-PFV-hear-IND-moaning-in-IND-boat ‘one day, (while) the man came in the evening with his animals, he heard moaning in a hole’ (Maurer 2003: 240). The comparison with the Indo-European historical present is further flawed by the fact that the afore-mentioned distinction between statives (zero-marked) and nonstatives (markerd by ta) also applies to PA narratives (cf. Bickerton 1981: 84, 85 for a similar argument). Of course, in past-tensed narrative discourse such as exemplified above (102– 104), it is very well possible to find the past imperfective marker PA tabata instead of tense-neutral ta. In this regard, it is crucial to distinguish between basilectal and meso-/acrolectal varieties of PA (cf. Andersen 1990, 2000)292 . The frequency of tabata in narrative discourse rises proportionally as we move towards the acrolect, whereas basilectal speakers continue to show a strong preference for ta where, according to the rules of absolute tense marking, one would expect tabata. To the illustrations provided above (102–104) and in Andersen (1990, 2000), I would like to add one more. Here, Allen293 interviews a speaker of basilectal PA, hoping to learn more about the social situation on Curaçao in the mid-20th century. She asks: Kantu sèn señora ta haña? how.much-money-madam-IMPearn ‘How much did you earn?’ The informant answers: Kasi pòrnada bo ta traha almost-for free-2sg-IMP-work ‘you would work almost for nothing’(Allen 2007: 204). Space limitations do not permit further discussion of Allen’s (2007) field work notes here; suffice it to note that similar examples of the tense-neutral use of ta are rife therein. 292 Andersen (1990) is considered by some to be “[a] significant milestone in the treatment of TMA in Papiamentu” (Faraclas, Rivera-Castillo & Walicek 2007: 273). Andersen argued that, due to ongoing contact with Spanish and Dutch, PA has doubtlessly acquired the characteristics of a tense-prominent language. However, this prominence of tense is perceivable mainly in more educated formal registers, the users of which tend to be bi- or multilingual in PA and Spanish, Dutch or English: “[I]t appears that, rather than Papiamentu intrinsically having an absolute tense system (…), in this register Papiamentu is made to conform to the absolute tense system they use for formal Spanish, Dutch or English” (Andersen 1990: 68). The basilectal varieties, in turn, “show greater aspectual focus” (Andersen 1990: 61). 293 The sociolinguist Rose-Mary Allen has done extensive field work in remote, rural areas of Curaçao, where more basilectal varieties of PA are spoken.

5.1. PA / Upper Guinea PC preverbal ta

191

5.1.2.4. Final remarks on PA ta as [+imperfective], rather than [+present] This section has argued for the status of PA ta as a tense-neutral imperfective aspect marker. With that purpose, I have highlighted the use of ta rather than tabata in past-tensed narrative discourse, a feature that is more frequent in basilectal than in meso-/acrolectal PA. Also, I have stressed the obligatory use of PA ta in past-reference ta + V complements and subordinate clauses of indirect speech as well as its absence before stative verbs. These native structures show that the tense-neutral use of PA ta is in fact part of the daily speech of all PA speakers irrespective of their social background. In light of the data presented above, the tradition of labeling and glossing ta as PRESENT or ‘non-past’ should be reconsidered (at least in scholarly, nondidactic works), as these labels fail to account for the synchronic distribution of ta in sentences with a past reference. Treating ta as a [+imperfective] aspect marker, on the other hand, accounts both for its distribution in sentences with a present tense reference (given the obvious semantic correlation between imperfectivity and present tense) and for its distribution in clauses with a past reference. Whether PA ta is on its way to becoming a general present tense marker is open to discussion (cf. particularly Maurer 2003). Such a development, as shown above, has certainly not been completed yet, but due to intensive contact with prestigious languages with an absolute tense system (Spanish, Dutch and English), it is merely predictable, and indeed observable, that the use of PA ta in main clauses with past reference (such as in narratives) is decreasing in favor of a more generalized use of the past marker tabata. It is also true that certain stative verbs, counter to what one would expect from an aspect-prominent creole, require preverbal ta instead of zero for present reference (more to this in §5.6) and that PA, unlike Upper Guinea PC, has developed a clearly demarcated future tense marker (lo, discussed in §5.3) in opposition with ta. All these are signs of an ongoing development from aspect- to tense-prominence (cf. §5.5) and from ta [+imperfective] to ta [+present], although the before-mentioned distinction between basilectal and meso-/acrolectal varieties must at all times be taken into consideration. In any case, the frequent and clearly tense-neutral use of ta in past-reference ta + V complements (discussed at length in §5.1.1) is unlikely to disappear from the language. In this respect, the discussion by Maduro of an odd example of Vaux + tabata + V (instead of Vaux + ta + V) provided by Lenz’s (1928: 176) informant Sillie is revealing. With respect to the proposition E pilote a kuminsá tabata lamantá kalumia the-pilot-PFV-start-tabata-lift-calumny ‘The pilot started slandering [the girl]’, Maduro (1967: 24, in Maurer 1988: 269) crit-

192

Verbal system

ically observes: “By erroneously writing tabata lamantá instead of ta lamantá, Sillie has given Dr. Lenz more work than necessary”294 . 5.1.3.

Final remarks on PA / Upper Guinea PC preverbal ta

Section 5.1 challenged the alleged status of Upper Guinea PC ta as [−progressive] as well as of PA ta as [+present]. Linguistic data have been put forward to show that, in both regards, the properties of the particle ta in fact are nearly identical in PA and Upper Guinea PC. If we consider that PA / Upper Guinea PC ta also marks futurity (cf. §5.3), the marker meets the criteria of a true imperfective marker as defined by Bybee et al. (1994: 287) in that it is used as a progressive, habitual and future, and as an imperfective with both present and past reference. It is a small step from this analysis to the hypothesis that PA ta and Upper Guinea PC ta have one and the same origin. To close, Table 34 gives a summary of the occurrence of the imperfective marker ta within the Atlantic branch of Iberian-based creoles. Table 34. The imperfective aspect markers employed in the Atlantic Iberian-based creoles295

habitual progressive

PA

CV

GBC

ST

PRI

ANG

PLQ

ta ta & ta + V-ndo

ta various & ta

ta na & ta

ka (sa)ka

ka ∼ sa sa

ka θ aka

asé ta

294 Original quote: “Skirbiendo eróneamente tabata lamantá enbes di ta lamantá, Sillie a duna Dr. Lenz mas trabou pa analisá ku tabata nesesario” 295 If we take into consideration non-Atlantic creoles, it becomes clear why the particle ta has been a central piece of evidence in monogenesis frameworks developed since the 1950s. Though absent in Gulf of Guinea PC, ta is found with different functions in Chabacano, PLQ and several of the Asian Portuguese creoles (e.g. Hancock 1975: 221, 222; Quint 2000b: 224, 225; Lipski & Santoro 2007: 378, 379). In PLQ, the function of ta is restricted to progressive aspect; asé encodes habitual (Schwegler & Green 2007: 280). The aspect marker ta (or tá) appears also in Saramaccan encoding both progressive and habitual (Narrog 2008: 361). However, as noted in Chapter 1, scholars generally agree that Saramaccan is English-based; correspondingly, the etymon of ta is thought to be English stand rather than Port. estar (e.g. Bakker 2008b: 52). Lipski (1993) compares the functions of ta in PA, PLQ and Chabacano.

5.2. The diachrony of the PA perfective past marker a

193

5.2. The diachrony of the PA perfective past marker a The PA perfective past marker a296 is one of the features that most obviously set PA apart from Upper Guinea PC, which, like most Atlantic creoles, has a zero perfective marker instead. This characteristic PA feature has indeed, at least implicitly, been adduced as evidence against genetic ties with Upper Guinea PC, making the need for a diachronic discussion obvious. Below, typological evidence will be discussed suggesting that a was added to a pre-existing TMA system with a zero perfective marker. One of the main pillars of Bickerton’s creole prototype consists of three strongly correlating features pertaining to, or resulting from, the stative-nonstative distinction (after Bickerton 1981: 88; [1974]1980: 5, 6): 1. zero-marked (/unmarked) statives yield a present reading297 ; 2. present-reference nonstatives are obligatorily marked by an imperfective marker; 3. zero-marked (/unmarked) nonstatives yield a perfective past reading. PA and Upper Guinea PC both comply with the first two features. This is conveniently exemplified by the verb PA yama / SCV txoma, which can mean either stative ‘to be called’ or nonstative ‘to call’. As a stative verb, the unmarked form will yield a present reading (105); as a nonstative verb, the imperfective marker is required to yield a present reading (106): (105) a. PA

E Ø jama Maria 3s be called Mary ‘She is called (named) Mary’ (Birmingham 1970: 123; cf. Maurer 1988: 62 & 2003: 237f.n.; Holm et al. 2000: 137)

296 In the literature, PA a is alternately described as a perfective aspect marker (e.g. Kouwenberg & Ramos-Michel 2007: 312), as a perfect/perfecto (Maurer 1998: 161), or as a simple past (Bickerton 1981: 87). Judging from Maurer’s (1988: 107–111) description and analyses, PA a seems to cover both the Spanish pretérito (in)definido (e.g. comí ‘I ate’) and the presente perfecto (he comido ‘I’ve eaten’) and may thus be translated with either an English simple past or a present perfect, depending on the context. To reflect this, I will refer to a as a perfective past, in keeping with e.g. Faraclas, Rivera-Castillo & Walicek (2007: 274). 297 Note that ‘present reading’ here means a temporal reading simultaneous with a contextually determined reference point, which can, but needs not be, the moment of speech (cf. e.g. Singler 1990a: xi, xii).

194

Verbal system

b. SCV N Ø tchoma Leonor 1sg be called Leonor ‘I am called Leonor’

(Mendes et al. 2002: 121)

(106) a. PA

E ta jama Maria 3s IMP call Mary ‘He calls Mary’ (Birmingham 1970: 123; Maurer 1988: 62 & 2003: 237f.n.; Holm et al. 2000: 137) b. SCV tem algem ki ta tchomâ-l di fitiseru have people who IMP call-him [of] sorcerer ‘there are people who call him a sorcerer’ (Mendes et al. 2002: 121)

The comparison becomes problematical, however, when one considers that creoles (including Upper Guinea PC) that display features number one and two typically also display feature number three (unmarked nonstatives yield a perfective past reference), as in (107) (cf. e.g. Bickerton [1974]1980: 5). PA here deviates from the creole prototype, as a perfective past on nonstatives can only be conveyed with the overt marker a, as in (108).298 (107) SCV

Maria Ø tchomâ-l má e ka bem Maria PFV call-him but he NEG come ‘Maria called him but he didn’t come’ (Mendes et al. 2002: 121)

(108) PA

el a yama su kasá (*el Ø yama su kasá) he PFV call his wife ‘he called his wife’ (Maurer 1988: 297)

On the surface, PA a is doubtlessly one of the features that most clearly separates the two creoles synchronically. However, as is usual in establishing genetic ties between languages, the diachronic state of affairs, if reconstructable, is of overriding importance. Now, since, as pointed out above, synchronic TMA marking in PA displays the first two features pertaining to the typical (Atlantic) creole pattern of stative vs. nonstative marking, it must be assumed that, at an earlier stage of development, PA also displayed feature number three. This typological reasoning is summarized in Table 35: 298 Predicates in irrealis/subjunctive constitute an exception: in irrealis, all PA verbs can in principle have a present reference when marked by zero (for PA, cf. especially Maurer 1993). In this section, I will be concerned only with what Bickerton (1975: 27) calls the ‘realis’ subsystem of the creole.

5.2. The diachrony of the PA perfective past marker a

195

Table 35. Perfective aspect marking in Upper Guinea PC, PA, and Early PA

present-reference statives present-reference nonstatives perfective past nonstatives

Typical Atlantic creole

PA

Early PA

zero IMP zero

zero IMP a

zero IMP zero

A similar point was made by Bickerton (1981: 88): “The distribution of zeros and nonpunctuals in Papiamentu is identical to that of synchronic Sranan or basilectal Guyanese Creole, except for one thing: in the latter, past punctuals, as well as present statives, are zero-marked (…). In Papiamentu, a has moved in to fill the ‘vacuum’”. To be sure, this typology can be extended to non-creoles which have a stativenonstative distinction similar to creoles: if they comply with the first two features, they will most likely also have a zero marker to encode perfective on nonstatives: “A zero form, in languages in which the imperfective is overtly marked but does not occur on stative predicates [i.e. feature one, BJ], will indicate perfective for dynamic verbs [feature three, BJ], but present for stative ones [feature two, BJ]” (Bybee et al. 1994: 77; cf. Bybee 1994: 251)299 . Again, it is only the overt perfective marker a that separates PA from a typological fit with these languages, at least, in the domain of TMA. One may object to the above argumentation by pointing out that among the languages based on an aspectual perfective-imperfective opposition, several do in fact take recourse to an overt (rather than zero) morpheme to mark perfective on nonstatives, just as modern PA does. Crucially, however, in these languages – but not in PA – the same overt marker is also used to mark present-reference on stative verbs (Bybee et al. 1994: 90). A case in point is Bambara (Mande), where the overt marker y(é) encodes perfectivity on nonstative verbs as well as present-reference on stative verbs (Quint 2008a: 84). This, in turn, means that, if 299 One example of a non-creole conforming to this pattern is Mandjaco (WestAtlantic): zero marking yields a perfective past with nonstative verbs (bu Ø tsëp they-go ‘they went’) and a present with statives (a Ø me he-know ‘he knows’) (Doneux & Rougé 1993: 52). Both Bybee (1994: 251) and Bybee et al. (1994: 90) list several other non-creoles of this type. For data on Mande and West-Atlantic languages following this pattern, cf. Quint (2000a: 248, 2008: 82–86), Doneux & Rougé (1993: 52), and Rougé (2006: 71). For creole languages, besides Bickerton (1974[1980], 1975, 1981 and 1984), discussions and data on the typical creole TMA marking are found, for instance, in Singler (ed. 1990), Holm et al. (2000), Faraclas, Rivera-Castillo & Walicek (2007) and Holm & Patrick (2007).

196

Verbal system

a were an original part of PA’s grammar, one would expect it to encode not only perfectivity on nonstatives but also present-reference on stative verbs, which however it does not. As Faraclas, Rivera-Castillo & Walicek (2007: 269) put it: “If a has always been in opposition to ta, it would presumably have a wide distribution before [+stative] verbs (…), but this is not the case”. Along the same lines, Bickerton (1981: 87, 88) commented that, if PA a were an original creole feature, “the synchronic distribution of zero forms becomes merely a mysterious anomaly”. Citing Faraclas, Rivera-Castillo & Walicek (2007: 274), we may conclude that the presence in PA of a instead of zero “is a relatively superficial feature that masks fundamental congruity between TMA in PA, West African languages, and other Caribbean Creoles”. More precisely, the backbone of Upper Guinea PC’s verbal system, the imperfective:perfective opposition realized by ta:zero, must once have been central to PA as well. The postformative introduction of PA a replacing the original zero is likely to result from contact with Spanish. The etymon of a is generally (and correctly, I believe) thought to be the 3rd person singular of the Spanish perfect auxiliary haber ‘to have’, ha (e.g. Labov [1971]1990: 42, 43; Maurer 1988: 348; but see Rona 1971: 20300 and Bickerton 1981: 86301 ). Due to the acquisition of the overt perfective past a and the development of tabata into a true imperfective past tense marker, PA’s system has acquired an opposition that largely resembles the Spanish pretérito (in)definido – pretérito imperfecto opposition. This is unlikely to be a coincidence. As such, the introduction of a seems to be a classical case of what is often referred to as ‘gap filling’ in the literature on contactinduced change and grammaticalization: “gaps in the morphological inventory

300 Rona (1971: 20) argued in favor of an African etymology noting that Akan and other Volta languages have a prefix a- that marks perfective aspect. Indeed, Akan has a perfective aspect prefix a- (Osam 2008: 79) and speakers of Akan are indeed likely to have been brought to Curaçao, given that the Dutch drew significant numbers of slaves from Ghana. Thus, it seems not at all impossible that the Akan prefix a-, perhaps in convergence with Spanish ha, contributed to the development of PA a. 301 Bickerton (1981: 86) finds that PA a may stem from the adverb Port. já ‘already’. Though this cannot be disproven, I assume that if PA a were derived from the adverb Port. já, it should at least optionally precede the subject, whereas in reality it is rigidly positioned after the subject and before the verb. By comparison: the SCV cognate of Port. já, SCV dja (used to indicate the present relevance of a past action) consisently precedes the subject: SCV dja-m perdi nha kamisa ‘I’ve lost my shirt’ (Quint 2000a: 238); and so also the future marker PA lo (derived from the adverb Port. logo) optionally appears before the subject (PA lo mi bai FUT-I-go ‘I’ll go’).

5.3. Future tense marking in PA and Upper Guinea PC

197

of a language facilitate the importation of new categories from another language” (Heine & Kuteva 2005: 124, drawing on Winford 2003: 96, 97).302 I close with Bickerton’s (1981: 88) description of this particular case of contact-induced gap filling: “In Papiamentu, a has moved in to fill the ‘vacuum’ created by zero-marked past-reference nonstatives, thereby bringing the P[apiamentu] TMA system closer to European models, but (…) leaving clear traces of the more creole system that must have existed at an earlier stage.”

5.3. Future tense marking in PA and Upper Guinea PC In PA, the marker ta and the auxiliary bai ‘go’ may function as markers of immediate and certain future events (Maurer 1988: 88–100; Kouwenberg & Ramos-Michel 2007: 312). However, the marker generally associated with future/irrealis marking in PA is lo. Compared to PA ta and bai, PA lo encodes more distant and uncertain future events (Kouwenberg & Ramos-Michel 2007: 312). Also, unlike PA ta and bai, PA lo is typically used to mark conditional clauses. (Note that, in the following discussion, I am primarily concerned with future tense marking and less so with irrealis and conditionals.) A discussion of the diachrony of PA lo is of importance to the present study, given that a functional equivalent is not attested in Upper Guinea PC. Below, it will be argued that PA lo (quite like PA a) is a nonnative feature which developed after the separation of PA from Upper Guinea PC as a result of regular processes of language change and, hence, does not challenge the hypothesized genetic ties between the two creoles.

302 The idea that the original zero constituted a ‘gap’ of course does not imply that the ancestor variety was somehow defective. We may only speak of a ‘gap’ in Early PA with respect to its contact languages, in this case Spanish, which did (and does) have overt perfective marking. Still, the idea of a ‘gap’ is important as it suggests that there was scope for an overt perfective marker to be integrated without it having to replace an already existing overt morpheme. I thereby assume that the integration of a new overt morpheme will be more complicated if it requires the substitution of an already existing overt morpheme. It is furthermore debatable whether the hypothesized change from zero to a should be analyzed as the introduction of a new category or merely of a new morpheme. The answer, I suppose, depends on the extent to which the marker a differs semantically from the original zero. I will not address this issue here, but see Bickerton (1981: 86–88) and Faraclas, Rivera-Castillo & Walicek (2007) for ideas about the functions and semantics of PA a in comparison to the more widespread zero in creoles.

198

Verbal system

5.3.1. The PA future tense marker lo vs. its absence in Upper Guinea PC Due to its adverbial properties, its clearly Portuguese etymon (< Port. logo ‘soon, then, immediately, later’) and its resemblance to the future marker lo(gu) attested in several Asian-Portuguese creoles, PA lo has been a theme of discussion in the works of scholars such as Schuchardt (1882: 911, 912), Labov ([1971]1990: 42, 52),Alleyne (1980: 129), Bickerton (1981: 77, 81, 82) and Goodman (1987: 374) (cf. Bartens 1996: 259 and references therein). Interestingly, several scholars have reported the existence of a future marker lo also in CV.303 However, it can with some degree of certainty be affirmed that this is false (Nicolas Quint p.c.). Although an adverb derived from Port. logo exists in CV and GBC (both logu), it did not grammaticalize in these two creoles, or in any other West African Portuguese-based creole for that matter. Rather, CV and GBC have the imperfective marker ta covering the domain that is covered by lo in PA (e.g. Baptista, Mello & Suzuki 2007: 59; Bartens 2000: 52 for BaCV).304 Upper Guinea PC ta, then, is what Bybee et al. (1994: 275) refer to as an aspectual future.305 In typological terms, Upper Guinea PC appears to resemble what Detges (2001: 150f.n.) refers to as ‘futureless’ or ‘future-weak’ languages, i.e. languages that have not developed specialized future markers, but instead mark future contextually or aspectually, taking recourse to structures or items (in this case the imperfective aspect marker) whose primary function is not the encoding of future tense. The presence of lo in PA vs. its absence in West African Portuguese-based creoles led Maurer (1988: 348) to question: “Now, if PA was derived from an Afro-Portuguese proto-pidgin, how could we then explain the presence of lo in PA and several Asian-Portuguese creoles, whereas this morpheme is not

303 For some time, the erroneous assumption circulated that CV had a future marker lo (e.g. Thompson 1961: 110; Hancock 1975: 222; Ferrol 1982: 42; Lipski 1988: 28). The ultimate source of this piece of data is unknown, as none of these authors reveal their source. The spread of this piece of misinformation is discussed also by Holm (2009: 14 in reference to Thompson 1961), Maurer (1985: 56, in reference to Hancock 1975) and Bartens (1996: 40, in reference to Lipski 1988). 304 Though less typically, the progressive markers GBC na and SCV sta∼sata as well as several combinations of TMA markers appear also to allow a future reading (see Baptista, Mello & Suzuki 2007: 59, 60). 305 This definition reflects the fact that its primary role is that of an aspect marker, whose capacity to mark future tense “arises as a contextually determined use, and not, as is the case with primary futures, as an evolutionary endpoint in the unfolding development of originally lexical material” (Bybee et al. 1994: 275).

5.3. Future tense marking in PA and Upper Guinea PC

199

attested in the Afro-Portuguese creoles?”306 Though the question is legitimate, it can be answered by taking into account the fact that all languages, including creoles, undergo regular linguistic change. Thus, it is perfectly possible that after the separation of PA from its Upper Guinea PC ancestor, the future marker lo developed in PA through regular processes of either internal change or syntactic borrowing. (Both these options are discussed in §5.3.2.) First, we must recall that a specialized future marker (i.e. a marker exclusively encoding futurity) is not at hand in the future-weak TMA systems of CV and GBC, speakers of which take recourse to aspectual or contextual future marking instead. This implies that the TMA system of proto-Upper Guinea PC probably did not possess a specialized future marker. Clearly, if PA inherited this system, there was scope for the specialized future marker lo to develop (viz. to fill the ‘gap’). A strong indication that PA lo did not develop during creolization is its adverbial origin. Whereas temporal adverbs are a common lexical source for TMA (including future) markers cross-linguistically (Bybee et al. 1994: 244; Heine & Reh 1984: 132), creole languages clearly deviate in this respect: “Adverbs (…) are not a good source for creole TMA markers” (Bickerton 1981: 86). Of course, besides PA lo, there are other reported cases of adverbs turning into creole TMA markers, but if we may believe Bickerton (1981: 79), these constitute “the creole exception rather than the rule”. Thus, PA lo is unlikely to have emerged during creolization, but more likely some time afterwards. Also, as has often been remarked upon (e.g. Alleyne 1980: 129), reflecting its adverbial origin, PA lo can precede the subject, as in PA lo mi bai kas mas lat FUT-I-go-house-more-late ‘I’ll go home later’ (Kouwenberg & Ramos-Michel 2007: 313), and in negative clauses obligatorily precedes the negator: e.g. PA lo mi no bai [FUT-I-NEG-go] ‘I will not go’ (cf. PA *mi no lo bai) (Maurer 1988: 52f.n.; 1998: 161).307 As is well known, these positions outside the verb phrase are unusual for creole TMA markers, which, with very few known exceptions, are typically positioned rigidly after the subject (and negator, if at hand), imme306 Original quote: “Or, si le papiamento était dérivé d’un proto-pidgin afro-portugais (…), comment expliquer la présence de lo en papiamentu et dans certaines langues créoles luso-asiatiques, alors que ce morphème ne se retrouve pas dans les langues créoles afroportugaises?” 307 In several other aspects, PA lo behaves like the other members of PA’s TMA system. For instance, just like PA ta, lo can occur in tonic position in the answer to a yes-no question: PA lo bo bai? – sí, mi lo FUT 2sg-go-yes-1sg-FUT ‘will you go? – yes, I will’ (Quint 2000b: 151). See Maurer (1988: 52, 53) and Kouwenberg & RamosMichel (2007: 313) for further considerations regarding the rules that govern the position of PA lo.

200

Verbal system

diately before the verb (e.g. Bickerton 1981: 77). Again, this suggests that PA lo was integrated subsequent to, rather than during, creolization.308 As with PA a, the hypothesized postformative development of lo fits in well with cross-linguistically unmarked patterns of language change. Here too, we seem to be dealing with a case of gap filling, in as far as lo did not replace an existing morpheme, but rather took on a position that was vacant in the original TMA system. (Again, the notion of ‘gap’ must not be taken to imply that the original system was somehow defective; it merely indicates that a given language does not contain a category present in one of its contact languages.) Also, as was the case with PA a, there is a good likelihood that contact motivated and accelerated the introduction of lo: “we see contact-induced similarities in the organization of the future tense in several languages” (Matras 2007: 44). Correspondingly, it is conceivable that the integration of PA lo, be it through borrowing or through internal grammaticalization (discussed below), was stimulated by pressure from Spanish and Dutch, indeed, languages endowed with a clearly demarcated, absolute future tense category. 5.3.2.

On the origin of PA lo

As hinted at above, at least two scenarios can be hypothesized to account for the addition of lo to PA’s stock of TMA markers: one in which PA lo results from the internal grammaticalization of Early PA *logu (< Port. logo ‘soon, immediately, 308 The Tok Pisin future marker bai (and variants), derived from the (archaic) English adverb by and by ‘later’, is an often-cited example of an adverb turning into a creole TMA marker. Tok Pisin is different from PA in that the former went through an extended pidginization period during which pidgin features such as semigrammaticalized temporal adverbs could stabilize (Faraclas 2007: 355). To account for the development of PA lo (which the Bioprogram does not foresee), Bickerton (1981: 82) claimed that PA had gone through a very similar extended pidginization period (or what he refers to as “delayed creolization”). This account obviously differs from my own, since I rather believe PA lo either developed within, or was added to, a pre-existing creole TMA system. Nevertheless, the parallels between the syntax of PA lo and that of Tok Pisin bai are noteworthy, as e.g. Garcia (1985: 209) points out: “in Tok Pisin bai generally precedes personal pronouns, but follows lexical expression of the Subject (…); similarly, lo in Papiamentu generally precedes personal pronouns but follows lexical subjects”. Moreover, “in Tok Pisin, the order ‘future marker-personal pronoun-verb’ is unmarked, whereas the order ‘personal pronoun-future marker-verb’ is ’emphatic’ (…). Exactly the same is true in Papiamentu” (Garcia (1985: 210). These parallels caused a supporter of monogenesis such as Whinnom (cited in Labov 1971[1990]: 52) to speculate that the Tok Pisin marker “may be a re-lexification of Portuguese logo”.

5.3. Future tense marking in PA and Upper Guinea PC

201

later, then’) and another in which PA lo constitutes a wholesale grammatical borrowing. Note that both scenarios, to be discussed consecutively below, are perfectly compatible with the claim posited above that PA lo developed after PA’s separation from Upper Guinea PC. Maurer (1998: 197) asserts that “in Africa and America there exists no other trace of lo∼logo as a future marker, so that it is difficult to claim an AfroPortuguese origin for the PA future marker lo”309 . Although this is basically correct, it must also be noted that the lexical source of PA lo, Port. logo, does have a cognate in Upper Guinea PC (as well as in Gulf of Guinea PC, for that matter). Thus, there is no reason why the source material for PA lo could not have been the Upper Guinea PC adverb logu, which, though without having grammaticalized, is still a solid part of Upper Guinea PC’s adverbial inventory.310 If Early PA’s feature pool contained a cognate of the Port. adverb logo, its subsequent language-internal development into a future marker is not very spectacular from a cross-linguistic perspective: “the development of temporal adverbs to future tense markers constitutes a cross-linguistically well-established path of grammaticalization” (Heine & Kuteva 2005: 243). Semantically, the path from an Early PA form *logu ‘soon, afterwards, later’ to lo [+FUT] (and later also [+IRR]) is relatively straightforward. It is not a coincidence that all adverbs reported by Bybee et al. (1994: 253) as having grammaticalized into a future marker have meanings close to Port. logo (ranging from ‘then’, ‘thereafter’ and ‘afterward’, to ‘just now’ and ‘soon’).311 How closely related Upper Guinea PC logu and PA lo still are semantically can be illustrated by these parallel examples: 309 Original quote:“En Africa y América no existe ninguna otra huella de lo∼logo como marcador del futuro, así que es difícil atribuir un origen afroportugués al marcador del futuro lo del papiamentu” 310 The Negerhollands preverbal habitual and progressive marker lo can also encode futurity (Muysken & Meijer 1979: xvi; Holm 1988: 164), but likely has its etymon in Dutch lopen ‘to walk’; the correspondence with PA lo is likely to be a mere coincidence (cf. Holm 1988: 165). 311 A comparable case of internal grammaticalization is also described by Heine & Reh (1984: 132): “In Bari, an Eastern Nilotic language, the adverb dé ‘then, afterwards’ was desemanticized to a future marker. Since tense markers are placed between subject and verb, while adverbs occur sentence-initially, or sentence-finally, desemanticization was followed by permutation”. With ‘permutation’ the authors refer to the process of shifting and subsequent rigidification of the syntactic position of the item undergoing grammaticalization (cf. Bybee et al. 1994: 7). As noted, this process has not (yet?) been completed in the case of PA lo, which can still precede the subject.

202

Verbal system

SCV Logu ten un bádju di funaná ‘(Later) there will be a funaná festival’ (Lang 2002: 400) ≈ PA lo tin yoraméntu ‘there will be weeping’ (Eybers 1916: 60) GBC vingansa logu ta tciga ‘vengeance will come’ (Scantamburlo 2002: 342) ≈ PA tur bo curpa lo ta luminá ‘your entire body will be illuminated’ (Eybers 1916: 13) Also phonetically, the development from Early PA *logu to PA lo is not unexpected, as the loss of (final) unstressed syllables is quite typical of grammaticalization (see e.g. Bakker 2008b: 47). In other words, the development of the future marker PA lo out of the adverb *logu is an unmarked development which does not in any way contradict the hypothesis that PA descends from Upper Guinea PC. If anything, it strengthens the link. An alternative hypothesis holds that PA lo was borrowed wholesale from another language, most likely an Asian Portuguese creole. This hypothesis is nourished by the afore-mentioned presence of a preverbal future marker derived from Port. logo in the branch of Asian Portuguese creoles, including Korlai (Clements & Koontz-Garboden 2002: 220), Papia Kristang (Baxter 2004: 57), Macanese (Thompson 1961: 109), Sri Lanka Creole (Maurer 2009b: 44), as well as the extinct creole of Batavia (Maurer 2003b: 2).312 Accordingly, several scholars have found PA lo to constitute a “link with the Pacific” (Alleyne 1980: 129). The most concrete formulation of this hypothesis is offered by Maurer (1988, 1998) and holds that the European colonizers involved in the settlement of Curaçao may have brought along the future marker lo after having spent time in the Dutch Asian Portuguese colonies. One interesting historical fact supports this hypothesis: the Asian Portuguese creoles that have a future marker derived from Port. logo are indeed all spoken in territories formerly occupied by the Dutch (Maurer 1988: 348, 1998: 161, 162, 197, 198).313 This borrowing hypothesis doubtlessly has merit. Note, however, that the grammatical borrowing of a TMA marker requires that there was relatively intensive contact (say, scale four on Thomason & Kaufman’s [1988] borrowing 312 Two Asian Portuguese creoles which, as far as I know, do not have a future marker derived from Port. logo are the creoles of Daman (Clements & Koontz-Garboden 2002: 220) and Diu (Cardoso 2009: 152, 288, 289). Hancock (1977: 221) and Holm (1988: 267) provide overviews of future markers in Portuguese-based creoles. 313 Several scholars have furthermore suggested that the attestation of lo(gu) as a future marker in PA and the Asian Portuguese creoles may ultimately be related to the semigrammaticalized use of Port. logo in Vicente’s 16th-century Lingua de Preto texts (e.g. Naro 1978: 324, 329; Maurer 1986: 140; Kramer 2004: 127–129).

5.3. Future tense marking in PA and Upper Guinea PC

203

scale) between the replica language, PA in this case, and the donor language (presumably an Asian Portuguese creole variety). Other than PA lo, there are no very clear (linguistic) indications that this precondition was ever met, although it cannot be disproven either, of course. For now, then, there seems to be no a priori reason to discard either one of the two hypotheses outlined above for the origin of PA lo. The point to bear in mind here is that neither of the two scenarios (internal grammaticalization or borrowing) have any bearing on the hypothesis that PA descends from Upper Guinea PC. In both scenarios, the marker lo was integrated in PA only after its separation from Upper Guinea PC, possibly as a result of (or to support) the general shift of PA’s verbal system towards the absolute tense systems of the European languages with which it was in contact. 5.3.3. The diachrony of future tense marking in PA and Upper Guinea PC According to Bybee et al. (1994: 21), languages quite often have more than one way of marking future tense: “[E]specially in domains such as the expression of future and modality, it is not unusual to find an array of grammaticized and grammaticizing constructions of different ages and sources sharing or competing for overlapping territories” (cf. Bakker [2008a: 49] for a similar point on creoles). This is to say that if, as argued above, PA lo indeed was not part of the original stock of TMA markers, this does not mean that its predecessor had no means of encoding future tense. In fact, if we look at alternative ways of encoding future tense in PA, several parallels with Upper Guinea PC surface that allow for a tentative reconstruction of future marking in Early Upper Guinea PC. In CV, the imperfective marker ta marks immediate, certain events as well as less immediate, more uncertain events. In PA, as noted, ta is available to mark the immediate, specific future314 . Thus, there is a partial overlap between PA ta and CV ta in the domain of future tense marking (109–111): (109) a. SCV Manha, N ta kume pexe na djanta tomorrow I IMP eat fish in dinner ‘Tomorrow, I will eat fish at dinner’ (Pratas 2010: 221)

314 PA ta can mark more immediate and/or more certain future events: PA nos ta gana sigur-sigur we-IMP-win-sure-sure ‘we’ll surely win’; mi ta bai otro aña I-IMP-goother-year ‘I’ll go next year’ (Maurer 1988: 91, 93).

204

Verbal system

b. PA

Mañan mi ta kome dos biaha so tomorrow I IMP eat two time only ‘Tomorrow, I will eat no more than two times’ (Maurer 1988: 88)315

(110) a. SCV argen ki entra li nunka más ta sai somebody who enter here never more IMP leave ‘anybody who enters here will never leave again’ (Lang 2002: 745) b. PA Mi sí n’ ta bini Punda mas I yes NEG IMP come Punda more ‘I will never come to town [Punda] again’ (Allen 2007: 202) (111) a. SCV Si N ben más prumeru (…), el ta fika If I come more first he IMP stay ‘If I come earlier, he’ll stay’ (Lang 2002: 192) b. PA si e bini, mi ta keda if he come I IMP stay ‘If he comes, I’ll stay’ (Maurer 1988: 233) Admittedly, the future use of PA ta exemplified above is not highly remarkable, given that the use of a present or imperfective marker for future events is quite common cross-linguistically (Bybee et al. 1994: 275). Note, however, that the position taken by ta in (111) (i.e. in main clauses with a conditional if -clause) is the position taken typically by creole future/irrealis markers, as this example from Guyanese Creole (GU) shows (111c): (111) c. GU

mi go tel am if mi sii am I FUT tell him if I see him ‘I’ll tell him if I see him’ (Bickerton 1981: 93)

315 I take advantage of this example to point at the fact that PA biaha ‘journey’ is also the typical word for ‘time’ in expressions such as PA dos biaha ‘two times’. The same is true for GBC bias (Scantamburlo 2002: 32) and SCV biaxi∼viaxi (Mendes et al. 2002: 391), both from Port. viagem ‘journey, time’. This semantic extension is not found in Gulf of Guinea PC, where vyaji∼vyazi only means ‘journey’ (Rougé 2004a: 284). Rougé (2004: 284) notes that this semantic extension is typical of several African languages including Wolof and furthermore adds that in CV the use of biaxi ‘journey’with the meaning of ‘time’“est assez marquée comme rurale, voire archaïque”. Hesseling (1933a: 45) pointed out that archaic Dutch also displayed the use of reis ‘journey / voyage’ in the sense of ‘time’, as in dit werd enige reizen herhaald ‘this was repeated a few times’.

5.3. Future tense marking in PA and Upper Guinea PC

205

At least one other structure involving PA ta is indicative of its once more widespread use as a future marker and reminds one of the more idiosyncratic aspectual future use of SCV ta. It concerns the stative verb PA sa∼sabi ‘know’, which, when preceded by ta, typically receives a future reference, as in (112a, 112b) (Maurer 1988: 61; Muller 1989: 324). The same feature has been described for Upper Guinea PC, where ta (also na in GBC) typically combines with sabi316 (sebe∼sibi in GBC) to yield future-tensed clauses such as SCV e ta sabi ‘he’ll know’ (Quint 2000a: 241) / GBC I na sibi idem (Doneux & Rougé 1988: 21). Compare: (112) a. PA

mi ta sa unda nan a skonde e botin I FUT know where they PFV hide the key ‘I’ll find out where they hid the key’ (Muller 1989: 324) b. PA bo ta sa ku ta úniko moda di hasi you FUT know that be only way of do ‘you’ll find out that it’s the only way to do (/be)’ (Maurer 1988: 61) c. SCV Manhan nu ta sabe si e Pedru Pires ki tomorrow we FUT know if be Piedro Pires who ganha win ‘Tomorrow we’ll find out if it’s Pedro who won’ (Pratas 2007: 55) d. GBC i ta sebe kuma él ki meste kume she IMP know that she who want eat ‘she will find out that it is her whom he wants to eat’ (Chataigner 1963: 71)

In GBC, the progressive marker na has moved into the domain of future tense to mark the immediate, certain future, whereas ta marks the less certain, more distant future (Solovova 2004: 162). However, GBC ta is still considered the primary future marker in GBC (Morais-Barbosa 1975: 139, Peck 1988: 264, 265; Baptista, Mello & Suzuki 2007: 59) and, as expected of a future/irrealis marker, is typically selected in main clauses with a conditional if -clause (Doneux & Rougé 1988: 38):

316 Interestingly, as Quint (2009: 263) points out, the truncated form sa is also attested in SCV. Thus, SCV sa∼sabi = PA sa∼sabi. However, in PA the truncated form appears to be more frequent than the full form, whereas in SCV this ratio is reversed.

206

Verbal system

(111) d. GBC si bu ka laba rostu, bu ka ta bonitu if 2sg NEG wash face 2sg NEG FUT beautiful ‘if you don’t wash your face, you will not be pretty’ (Scantamburlo 2002: 167) In addition to ta, under certain circumstances and with subtle semantic differences, both PA and Upper Guinea PC may appeal to the auxiliaries ‘go’ (bai in PA, bai∼ba in SCV and GBC)317 and ‘come’ (PA/GBC bin; SCV ben), preceded by an imperfective marker, to mark future events (cf. Bartens 2000: 52 for BaCV; Quint 2000a: 263, 264f.n. for SCV and GBC; Kihm 1994: 108–110 for GBC; Kouwenberg & Ramos-Michel 2007: 313 and Maurer 1988: 96, 97 for PA). This is exemplified below with bin∼ben (113–114): (113) a. PA

Mi a bin komprondé (…) ku ta bin tin I PFV come understand that IMP come have mashá opstákulo many obstacle ‘I have come to understand that there will be many obstacles’ (Maurer 1988: 241) b. SCV N ta ben ten un fidju I IMP come have a child ‘I am going to have a child’ (Fanha 1987: 303)

(114) a. PA

Mi ta bin mira Mama kada un día pasa I IMP come see Mama every one day pass ‘I will (come and) visit you [Mama] each second day’ (Maurer 1988: 137) b. GBC kin kunsi si nomi, el ku na bin kasa who know his/her name he who IMP come marry -l him/her ‘Whoever knows her name will (come to) marry her’ (Kihm 1994: 108)

317 Upper Guinea PC uses bai as a lexical verb mainly, whereas a shorter variant, ba, is more typically used in auxiliary position (Wilson 1962: 24; Kihm 1994: 114; Quint 2000a: 286). Compare, for instance: GBC I bay matu he-go-bush ‘He went to the bush’ vs. i ba panya pirkitu bonitu he-go-catch-parrot-nice ‘he caught a nice parrot’ (Kihm 1994: 114). In SCV, an even further eroded form of bai, /a/, is now gaining ground in spoken discourse (Lang 2002: 40; Quint 2000a: 286). In PA, rapid speech occasions the contraction of ta + bai to /tei/, suggesting incipient grammaticalization (Kouwenberg & Ramos-Michel 2007: 313; cf. Maurer 1988: 262).

5.3. Future tense marking in PA and Upper Guinea PC

207

Of course, just like the future use of the imperfective marker, having go- and/or come-futures is not spectacular from a semantic or syntactic point of view and is in fact quite common cross-linguistically (Bybee et al. 1994: 266–270; for creoles, cf. Holm 1988: 164–65; Plag 1994: 13). Perhaps the main point to grasp is that, as with ta, the PA auxiliaries bai / bin can again be traced back to Upper Guinea PC, while differing, for instance, from Gulf of Guinea PC, which have ba∼be∼we for ‘go’ and bi∼bu∼vi for ‘come’ (Rougé 2004a: 180, 285). In light of the data presented in the foregoing, it is perfectly conceivable that both PA and Upper Guinea PC stem from an original TMA system that used the imperfective marker *ta as a primary future/irrealis marker and additionally took recourse to lexical (or semi-grammaticalized) items such as the auxiliaries *bai and *ben∼bin or the temporal adverb *logu.318 5.3.4.

Digression: SCV al and PA lo

Not yet discussed above is the preverbal marker SCV al (< Port. ha de ‘he/she/it has to’). In his 1887 SCV grammar, Brito (1887: e.g. 367, 371) described al as a true future marker and used it in positions where modern-day SCV would use ta. Although SCV al is used rather infrequently nowadays, Brito’s description suggests that it once competed more seriously with ta in the domain of future tense: “the material contained in Brito (1887) is sufficient to show that its use was clearly much more common at that time than it is now”319 (Quint 2008c: 135). While there is clearly no genetic or etymological link whatsoever between SCV al and PA lo, we may speculate that SCV al was designed to fill roughly the same ‘gap’ that lo has come to fill in PA. For instance, SCV nu al papiâ weFUT-speak ‘we will speak’ (Brito 1887: 355) translates into PA as nos lo papia (Goilo [1951]1975: 12).320 And while the exact synchronic status of SCV al is 318 If this reconstruction is correct, it means that proto-Upper Guinea PC deviated from Bickerton’s creole prototype in that, unlike this prototype, it did not have a marker designed specifically for marking future/irrealis. It is possible that the verb for ‘to go’ (reconstructed *bai) initially had this role, like in so many other creoles (Labov [1971]1990: 23; Bickerton 1981: 79). However, this would leave unaccounted for the fact that, synchronically, bai ‘to go’ does not function as such either in PA or in any variety of Upper Guinea PC. 319 Original quote: “Quelle qu’ait été la valeur exacte de al en capverdien du 19e s., le matériel contenu dans les Apontamentos est de tout façon suffisant pour montrer que son usage était à l’évidence beaucoup plus fréquent à l’époque que maintenant” 320 Particles clearly recognizeable as reflexes of Port. ha de have been reported with future marking functions in severalAsian Portuguese creoles (Lipski 1988: 29, drawing on Schuchardt 1883 and Dalgado 1906). Lipski (1988: 29) furthermore speculates

208

Verbal system

not agreed upon – some denote al as a type of mood marker with no inherent time reference (Quint 2000a: 258; cf. Lang 2002: 12), but others follow Brito in labeling al as a true future marker (e.g. Thiele 1991b: 50; Baptista, Mello & Suzuki 2007: 59) –, most of the examples provided by Lang (2002) in fact have a clear future reading, making al look like a rather typical future/irrealis marker with a semantic scope quite similar to PA lo. Furthermore, the examples provided by Quint (2000a) to illustrate the difference between SCV ta and al seem to correspond to semantic distinctions that speakers of PA can make by choosing between ta and lo. For instance, the modal difference between SCV m ta bem manhan I-IMP-come-tomorrow ‘I will come tomorrow’(Quint 2000a: 236) and SCV m al bem manhan I-POT-come-tomorrow ‘perhaps I come tomorrow’ (Quint 2000a: 240) can be achieved in PA by saying either mi ta bini mañan or mi lo bini mañan (Maurer 1988: 89). As mentioned, the use of al in modern-day SCV has diminished considerably. But speakers of SCV continue to look for alternative and formally distinctive ways of encoding future tense. For instance, Quint (2000a) recorded contracted forms – suggestive of grammaticalization – of periphrases with the verb ‘to come’, SCV t’em and sat’em (< (sa)ta + bem ‘come’): “This usage is not yet fixed. If the use of these contractions becomes widespread, the aspectual system of SCV could be quite significantly changed by the appearance of a formal future”321 (Quint 2000a: 263, 264).

5.4. PA / BaCV taba – tabata World-wide, a formal, semantic and morphosyntactic equivalent of the PA imperfective past marker tabata (∼tawata in Aruba) is found only in CV, which has staba ta in Sotavento322 and taba ta in Barlavento323 . Ever since Lopes that the Gulf of Guinea PC particles te∼di [+FUT] and possibly Chabacano di and ay∼ey [+FUT] might also be cognates of Port. ha de. 321 Original quote: “Cet usage n’est pas encore fixé. Si l’emploi de ces contractions se généralise, le système aspectuel du badiais pourrait être assez profondement modifié par l’apparition d’un futur formel” 322 Besides staba ta + V, the anterior progressive structure sa ta + V-ba is quite common in SCV (cf. e.g. Baptista 2002: 44) and in several places in the interior used more frequently than staba ta (Nicolas Quint, p.c.). 323 The CV varieties of São Vicente and Santo Antão have the variants tava and tava ta∼tava te, while São Nicolau has taba and tabata (Almada 1961: 116; Cardoso 1990: 65; Bartens 2000: 51). In São Vicente, moreover, the short forms tá (for tava) and tá ta∼tá te (for tava ta∼tava te) are common (Holm & Swolkien 2006: 209).

5.4. PA / BaCV taba – tabata

209

da Silva (1957: 140) provided examples such as el taba ta kanta he-be+PSTIMP-sing ‘he was singing’ in his well-read CV grammar, several scholars have drawn attention to the remarkable parallel with PA (e.g. Birmingham 1975: 20; Maurer 1988: 349; Green 1995: 65). In fact, as early as the late 19th century, Schuchardt (1882: 911), drawing on Coelho (1880: 5), had already observed this correspondence. Providing the example SCV en staba ta fazel falta I-ANTPROG-make.him-lack ‘I was missing him’, he commented: “Here we appear to have an imperfective which has its exact counterpart in PA: mi tabata hasi”324 (boldface mine). McWhorter (2006: 117) explains particularly well why the correspondence is so meaningful: “the parallel between [Cape Verdean] e taba ta kanta (…) and Papiamentu e tabata kanta is striking (…), since this usage is impossible to derive from any Iberian construction and is only one of many possible reconceptualizations of the lexifier material”. It should be noted that BaCV distinguishes between taba [+past/anterior, +habitual]325 and taba ta [+past/anterior, +progressive] (e.g. Bartens 2000: 51; Pereira 2000: 32), whereas present-day PA only has tabata [+past, +imperfective] covering the whole domain of imperfective past. However, scholars agree that also in PA, the base form taba must once have existed (Birmingham 1970: 87; Maurer 2003a: 247; Kramer 2004: 197; Kouwenberg & Ramos-Michel 2007: 309). Faraclas, Rivera-Castillo & Walicek (2007: 261f.n.), for instance, find it “tempting to see tabata as a compound” of an earlier base form taba and the imperfective aspect marker ta. They add that “we do not have synchronic evidence of this”. However, synchronic evidence as well as historical attestations of PA taba are readily available. For instance, there is the short form PA ta’a /taa/, which is frequent in present-day spoken discourse and which may very well result from the contraction of the earlier base form PA taba326 (see Maurer 1988: 50; Kramer 2004: 197). (Note, however, that PA ta’a and tabata differ from one another only stylistically, not semantically.) Furthermore, as for instance Maurer (1998: 164; 2003a: 237) has pointed out, the base form taba is still found petrified in the lexicalized past form of the stative verb tin ‘to have, be (existential)’, giving PA tabatin. Similarly, we find 324 Original quote: “Wir würden hier ein Imperfectum haben, das im Curazoleñischen sein genaues Gegenstück fande: mi tabata hasi” 325 Note that the available literature on BaCV is not unanimous as to whether taba and taba ta should be labeled as past tense or as anterior markers. 326 The hypothesized reduction of PA */taba/ to modern /taa/ has a parallel, for instance, in Saramaccan, where, as Bakker (2008b: 48) explains, the “loss of the second syllable combined with a lengthening of the vowel of the first syllable” led to the reduction of /kaba/ to /kaa/.

210

Verbal system

tava ten ANT-have (rather than *tava ta ten) in São Vicente CV (Veiga 2000: 268), although suppletive forms such as tinha (< Port. tinha) have become the norm in that variety nowadays (Dominika Swolkien p.c.). Furthermore, taba is visible in the past tense form of PA ta in its role as a copula ‘to be’ (cf. §5.7.2.1), giving tabata ‘to be [past]’. Tellingly, the original, chrystallized pasts PA tabatin and tabata are now in variation with the regularized forms tabata tin and tabata ta (Birmingham 1970: 87; Maurer 1986: 59f.n.; 2003: 238). In fact, the short form taba and the compound form tabata still appear side by side in the near-extinct Sephardic PA sociolect as described by Henriquez (1988: 83). About taba, Henriquez (1988: 83) specifies in telegram style: “antes hopi uzá, awor ménos i ménos” (‘much used in the past, now less and less’). This connects well with the fact that taba is found in the 1775 PA love letter written by a Sephardic Jew. But the form must also have existed in ‘standard’ (viz. non-Sephardic) PA at an earlier stage, which is suggested by its attestation in other historical, non-Sephardic sources. For instance, Van Name (1869: 156) reported the use of taba in the speech of his informant Camps327 . (Schuchardt [1882: 912] also mentioned the existence in PA of taba, but Van Name may well have been his source.) And last but not least, the PA phrase si mi taba sabi ku vo taba asina if-I-taba-know-that-you-taba-like.that ‘if I had known that you were like that’ is found in a short song fragment recorded in Coro (Venezuela). The song traces back to as early as the 18th century, when Curaçaoan runaway slaves settled there (see Lispki 2005: 192328 ; the song fragment is provided in the appendix to Lipski 2005, who draws on Domínguez 1989).329 Bickerton (1981: 86, 87) provided convincing creole typological evidence suggesting not only that tabata is indeed a compound of an early base form taba plus ta, but also that this base form must have started out as a true anterior marker330 , rather than as an absolute past, which tabata doubtlessly is nowa327 Van Name’s informant Camps was not a native PA speaker but had lived on Curaçao for several years (Bachmann 2005: 66). 328 Lipski (2005: 192) elaborates: “A number of the loangos had escaped from nearby Curaçao beginning in the eighteenth century, and are the principal source for the various Papiamento song fragments that survived in this region well into the second half of the twentieth century.” 329 Unfortunately, both the 1775 letter, Van Name’s speech sample, and the Coro PA song fragment are all too short to establish whether, at that time, taba was (still) in complementary distribution with tabata. 330 A classical creole anterior marker is defined by Singler (1990a: xii) as one that “locates an action as occurring prior to some reference point, this reference point in turn having occurred prior to the moment of speaking”.

5.4. PA / BaCV taba – tabata

211

days: creole “[a]nterior markers are most often recruited from a past copula form (…). Tava/taba is therefore what one would predict as an anterior marker in the original stage of any Portuguese creole” (1981: 87). Indeed, tava/taba [+ANT] is not only found in BaCV, but also in Gulf of Guinea PC (tava). Moreover, the conjunction of PA taba + ta, i.e. of an anterior marker + imperfective marker, is similarly predictable from a creole typological point of view: according to Bickerton, it is attested “for every creole without exception” (1981: 86) and typically marks anterior progressive or iterative situations. Indeed, not only in CV (taba ta), but also in Gulf of Guinea PC (tava ka / tava sa < tava [+ANT] + ka / sa [+IMP], Maurer 2009b: 44) or e.g. Guyanese English Creole (bina < bin [+ANT] + a [+IMP], Bickerton 1975: 37), the anterior tense and imperfective aspect markers combine to express an anterior progressive (cf. Holm 1988: 267 for a discussion of this feature in creoles). Bickerton (1981: 87) therefore proposes the following scenario to which I subscribe: “Originally, [Papiamentu] had taba anterior and ta nonpunctual, permitting the formation of tabata (it is hard to think of any other way in which this form could have been derived)” (cf. Kouwenberg & Ramos-Michel 2007: 310 for a similar point). (Incidentally, in his renowned article on the development of tense in creoles, Labov (1971[1990]: 43, drawing on Lenz 1928) drew a parallel between the elaboration of taba to tabata in PA and wen > wen go in Hawaiian English Creole and suspected stylistic motives behind this development.) If the historical analysis of PA tabata as *taba + ta is correct, it is implied that PA tabata was once a discontinuous marker, which BaCV taba ta of course still is (as for instance in SVCV Um tava so ta falá I-be+PST-only-IMP-speak ‘I was only chatting’, cf. §5.1.1.2). Its extension as a general imperfective past at the expense of taba followed by an increased frequency of use must have led to the reanalysis of tabata as a single grammatical unit. The fact that progressive tabata (rather than habitual taba) was selected to develop into PA’s general imperfective past sits well with the cross-linguistic tendency of progressives to evolve into imperfectives (Bybee et al. 1994: 20; cf. Dahl & Velupillai 2005b). 5.4.1.

Digression: On the diachrony of preverbal taba and postverbal -ba

The remarkable fact remains that PA lines up with the Barlavento varieties of CV in that both mark anterior/past tense preverbally, while the Sotavento varieties of CV and GBC take recourse to postverbal -ba (thus, PA/BaCV taba(ta) kanta = SoCV/GBC kanta-ba). This raises the intriguing question of how anteriority was marked in proto-Upper Guinea PC, i.e. pre- or postverbally. It should thereby be borne in mind that a direct genetic lineage between features found in BaCV and PA is unlikely in light of the chronology of the emergence of PA (late 17th

212

Verbal system

century) and the Barlavento varieties (through the 18th century) respectively. Based on the assumption (which I share) that the Sotavento varieties represent the oldest and most conservative varieties of Upper Guinea PC, Quint (2000b: 149) assumes that postverbal -ba was the original anterior marker and that, in the transition from Upper Guinea PC to PA, “the marker -ba shifted from postto preverbal position”331 . However, this does not yet account for the presence of preverbal taba∼tabata in both PA and BaCV: accepting Quint’s account means assuming that this shift happened twice independently. Though that is not impossible, we may alternatively speculate that the shared ancestor creole had both preverbal taba and postverbal -ba at its disposal at least until the Barlavento dialects separated (in the 18th century), with -ba coming out on top in SoCV and GBC at the expense of taba, which was however consolidated in PA and BaCV. To be sure, postverbal -ba is not completely absent from BaCV. According to Dominika Swolkien (p.c.), postverbal -ba “categorically does not occur in São Vicente or Santo Antão”, but she has heard it being used in São Nicolau, where it was also registered by Cardoso (1990:e.g. 62, 67; cf. Quint 2000b: 75).332 In fact, also in PA, vague remnants of postverbal -ba may have survived in the lexicalized Sephardic PA forms mesteriba (from mester ‘need, must, be necessary’) and poriba (from por ‘can, be able’) as documented by Henriquez (1988) and discussed in Martinus (1996: 187). Though clear-cut examples are scarce, these forms appear to have, or have had, a type of conditional meaning in phrases such as the following: PA e mesteriba bai kas 3s-must.iba-go-house ‘he should actually go home’ (Henriquez 1988: 48); bo poriba a laga mi mira e buki ayer 2sg-can.iba-PFV-let-me-watch-the-book-yesterday ‘you could have let me look at the book yesterday’ (Henriquez 1988: 62).333 A related, controversial issue is the etymology of postverbal -ba (cf. Baptista 2007b: 42–44 for a review of hypotheses). Unsurprisingly, -ba is generally thought to derive from the Portuguese postverbal imperfective past marker -va (as in cantava ‘(I/he/she) sang’) (e.g. Almada 1961: 116). However, a postverbal past marker is not found in any creole other than SoCV and GBC (Kihm 1994: 103), a circumstance that requires an explanation. 331 Original quote: “la marque -ba (…) du passé (…) était remontée du verbe à la particule” 332 I am indebted to Dominika Swolkien for a discussion on the topic dealt with in this section as well as for the reference to Cardoso (1990). 333 These forms are, however, very difficult to analyze since, clearly, the suffix is not -ba but -iba. A proper analysis of these forms seems to depend very much on the position of the stress: if the stress falls on the penultimate syllable (mester’iba), I assume the suffix iba could also be related to Spanish iba ‘it/he/she went’, rather than being a remnant of postverbal -ba.

5.4. PA / BaCV taba – tabata

213

Bickerton’s (1981: 80, 81) proposal to analyze postverbal -ba as an eroded form of the creole verb kaba ‘finish’ (an option considered also by Baptista 2007b: 43, Kihm 1989: 365/1994: 103 and McWhorter 2006: 118) provides such an explanation. His account relies mainly on the fact that kaba has been retained in PA, both as a lexical verb ‘finish’ and, more relevantly, as a postverbal completive aspect marker with a meaning similar to ‘already’, e.g. PA mi a kome bonchi kaba I-PFV-eat-bean-already ‘I’ve already eaten beans’ (Kouwenberg & Ramos-Michel 2007: 312). Following Bickerton (1981: 81), we may then conjecture that the postverbal completive marker *kaba was part of the original verbal stock of Upper Guinea PC, subsequently developing into the postverbal anterior marker -ba in SoCV and GBC. Perhaps it is important to note here that there has been two-way migration between Santiago and the Guinea-Bissau region well into the 19th century. This fact allows us to hypothesize that postverbal -ba developed on Santiago only after the separation of PA (17th century) and BaCV (17th/18th century). The feature could then still have spread easily to the Guinea-Bissau region, even though I assume GBC’s ‘true’ cut-off point to have been around the turn of the 17th century when Cacheu was founded (see details in Jacobs 2010). Kihm (1994: 103) notes in support of this etymology that “there is no semantic or phonetic impediment, indeed, such that, e.g., konta ba ‘had said’ could not have evolved from something like *konta kaba”. Phonetically, the wordfinal stress of the etymon, Portuguese acabar ‘to finish’, would indeed have allowed for the derivation of -ba along the following path: proto-Upper Guinea PC *kabá ‘to finish’ (full verb) > postverbal *kabá [+completive] > postverbal -ba [+anterior]. Semantically, moreover, the development from a verb ‘to finish’ into a completive and/or anterior marker is indeed comprehensible, and in fact quite common cross-linguistically (Bybee et al. 1994: 64, 105).334 But whichever etymology one prefers, it remains to be investigated (a) what factors caused GBC and SoCV to select postverbal -ba while PA and BaCV selected preverbal taba and (b) which of these two markers was Upper Guinea PC’s original anterior marker. 334 Kabá (< Port. acabar) as a postverbal completive marker is found also in Saramaccan (Kihm 1989: 365). Kihm (1989) furthemore affirms that several Mande languages such as Manjaku have a verb -ba ‘to finish, be finished’ “meaning completion when used with another verb” (Kihm 1989: 366). He uses this piece of data in support of the idea that the postverbal anterior marker GBC -ba resulted from the conflation of the Portuguese past ending -va, the proto-Upper Guinea PC completive marker *kaba, and the Mande completive marker -ba: “The syntactic peculiarities of Kriyol /ba/ would be readily explained given such an origin” (Kihm 1994: 103).

214

Verbal system

5.5. The issue of relative versus absolute tense marking in PA The data discussed so far have direct consequences for the more far-reaching issue of relative (viz. aspect-prominent or anterior-based) versus absolute (viz. tense-prominent or past-based) tense marking in PA, which has been a point of debate in several publications postdating, and including, Bickerton (1981).335 Due to the lack of a synchronic anterior marker as well as the fact that the PA preverbal markers a (§5.2) and lo (§5.3) deviate from the more typical creole TMA markers (cf. Bakker 2008b: 44), PA has often been characterized as an outsider among the Atlantic creoles and put forward as counterevidence to Bickerton’s prototype creole. Bickerton, however, accounted for PA’s apparently deviant (while absolute) TMA system by discussing evidence in favor of “an original anterior-nonanterior distinction [i.e. relative tense system] in Papiamentu” (1981: 87). Due to decreolization towards Spanish, he claimed, this relative tense system has made room for absolute tense marking. However, in a reply to Bickerton (1981), Maurer (1985) disputed the former’s conclusions, instead positing that PA likely never had anything other than an absolute tense system, which it could normally have inherited from its lexifier languages Spanish and Portuguese (1985: 60, 64). Although Maurer (e.g. 1988, 2003) has since presented particularly nuanced characterizations of PA’s TMA system, PA is still quite commonly described as a tense-prominent, past-based language, and thus seen as atypical among Atlantic creoles. Upper Guinea PC, on the other hand, at least GBC and the Sotavento varieties of CV336 , is generally considered as aspect-prominent and 335 In the comparison of creole TMA systems with those of their lexifiers, a crucial distinction is that between relative and absolute tense systems, alternatively referred to as aspect- vs. tense-prominent systems, or anterior- versus past-based systems. (Some scholars speak of ‘deictic’ [absolute] vs. ‘non-deictic’ [relative] tense [e.g. Bhat 1999: 14], but this terminology is not common in creole studies.) Comrie (1985: 56) specifies that in relative tense systems (such as displayed by most creoles), “the reference point for location of a situation is some point in time given by the context”, whereas in absolute tense systems, “the reference point for the location of a situation in time is the present moment”. Of course, there is no strict dichotomy between aspect(-prominence) and tense(-prominence). As Holm (2000: 180) points out: “There has been some mystification of aspect in creoles as if this made their verbal systems totally unlike those of their European source languages, but of course both tense and aspect are dimensions in the semantics of the verbal systems of both groups of languages” (cf. similar comments made by Parkvall [2000: 87]). 336 In the assessment of Upper Guinea PC’s verbal system, a clear distinction between the Barlavento and the Sotavento varieties is essential: the former have remained

5.5. The issue of relative versus absolute tense marking in PA

215

anterior-based (e.g. Quint 2000a: 229–240; Baptista 2002: 84; Lang 2009: 156– 163). For instance, while Kouwenberg & Ramos-Michel (2007: 309) describe PA as having “an absolute rather than a relative tense system”, Upper Guinea PC is described by Baptista, Mello & Suzuki (2007: 55) as having an anteriorbased, relative tense system. Since both these descriptions appear in the same comparative volume (Holm & Patrick 2007), the impression is created that the nature of PA’s TMA system is essentially different from that of Upper Guinea PC. In order to show that this impression is incorrect, I will briefly review the (creole) typological data in favor of the view that PA started out with an anteriorbased TMA system. Just like the typical anterior-based creole TMA system: – PA has a tense-neutral general imperfective marker (§5.1); – PA distinguishes morphologically between statives (zero for present) and nonstatives (imperfective aspect marker for present) (§5.2); – PA started out with a zero marker for perfective aspect (§5.2); – PA started out with an anterior marker derived from a past form of the lexifier’s copula (§5.4); – PA’s anterior marker combined with the imperfective aspect marker to signal that an event is [+anterior/past] and [+progressive]. On the basis of these features, Faraclas, Rivera-Castillo & Walicek (2007: 257) affirm that “Papiamentu TMA operates essentially on the basis of the same system found in most Atlantic Creoles (…)”. Indeed, this typological make-up can only be plausibly explained by assuming that PA, like all other Atlantic creoles, in closer contact with Portuguese and consequently undergone more intensive processes of decreolization, which is reflected particularly in the verbal system. The latter, on the other hand, have retained an intrinsically relative tense system, as shown in great detail by Quint (2000a: 225–273) and Baptista (2002: 75–100). This distinction has not always been made. Ferraz & Valkhoff (1975: 20), for instance, erroneously take Barlavento CV as representative of the whole of CV when they conclude that, of all Afro-Portuguese creoles, “Guineense would be the most aspectual of the Portuguese Creole languages and the nearest in structure to the West African languages, whilst Cape Verdean would be closest to Portuguese in this respect and has kept – or re-established – most of the Portuguese verbal system”. The need to distinguish between the Barlavento and Sotavento varieties of CV is nowadays better recognized: “le badiais et les créoles du Sud ont un système verbal original, dont la morphologie est dominée par les marques d’aspect (…). Les créoles du Nord ont un système verbal en grande partie aligné sur le modèle portugais” [‘SCV and the Sotavento varieties have an original verbal system, whose morphology is dominated by aspect markers. The Barlavento varieties have a verbal system largely aligned with the Portuguese model’] (Quint 2000b: 74).

216

Verbal system

started life with an anterior-nonanterior opposition (Bickerton 1981: 88). The data furthermore allow us to discard the view that PA’s verbal system stems directly from its Spanish/Portuguese lexifier system: neither Spanish nor Portuguese contains a tense-neutral general imperfective aspect marker (§5.1), a morphological stative-nonstative distinction such as displayed by PA and other creoles (§5.2), let alone a compound imperfective past/anterior marker of the likes of tabata (§5.4). On the other hand, it must be recognized that PA’s modern-day TMA system displays several features typical of absolute tense systems. For instance, the marker tabata is nowadays clearly an imperfective past rather than an anterior. Also, a series of stative verbs obligatorily take ta for present (discussed below in §5.6), which can be and has been taken as a sign that ta is developing in the direction of a present tense marker (§5.1.2) and that the stative-nonstative distinction in PA is merely historical (Maurer 1998: 163). However, all these changes from relative to absolute ways of marking tense can plausibly be explained as the result of the severe linguistic pressure from Spanish (and, to a lesser but not insignificant extent, Dutch) that PA has been undergoing ever since the late 17th century. Similar paths of decreolization involving a shift from relative to more absolute tense marking have been described in plenty for other creoles that have remained in contact with their lexifier (cf. not only Bickerton [1975, 1981] but also e.g. Alleyne [1980: 189, 190], Singler [1990b: 217] or Patrick [1999: 15]). In fact, the change from an originally relative to a more absolute tense system, as well as the gap filling (in the case of PA, the introduction of a [§5.2] and lo [§5.3]) that accompanies that change, match cross-linguistically common patterns of contact-induced change: Another domain where gap filling appears to be not uncommon is tense marking. There are some languages whose verbal system is characterized by an aspectual distinction imperfective vs. perfective and where tense is not a significant notion, being marked perhaps in the form of minor use patterns but not by means of distinct categories. Now, when such languages are in contact with languages having a clear-cut system of tense categorization it may happen that speakers of the former (aspect) languages replicate tense distinctions of the latter. (Heine & Kuteva 2005: 126)

We may conclude, then, that a correct characterization of PA’s verbal system depends crucially on the distinction between Early PA and modern-day PA. For Early PA, a relative, aspect-prominent and anterior-based TMA system can plausibly be hypothesized, closely conforming to forms (zero / ta / taba / tabata) and patterns (e.g. zero is present for statives, perfective for nonstatives) found

5.6. A comparison of stative verbs in PA and SCV

217

in Upper Guinea PC. In modern-day PA, the markers a and lo as well as the lack of a true anterior marker partially mask that fundamental parallelism.

5.6. A comparison of stative verbs in PA and SCV 5.6.1. The stative – nonstative distinction in creoles In §5.2, it was pointed out that creoles, including PA, distinguish morphologically between stative and nonstative verbs in that the latter obligatorily take an imperfective aspect marker for present reference (e.g. PA mi ta skirbi ‘I write’), whereas the former do not (e.g. PA mi tin ‘I have’). This is semantically motivated by the fact that stative verbs are inherently imperfective and hence need not be morphologically marked for imperfective aspect (cf. Singler 1990a: xiii). It is often said that this morphological stative-nonstative distinction in creoles is related to patterns found in the Indo-European lexifiers (e.g. Doneux & Rougé 1993: 52; Kramer 2004: 198). This claim is questionable at best, however. For instance, none of the European lexifiers distinguish morphologically between stative ‘I want’ and nonstative ‘I write’ in the way creoles do: PA mi Ø ker ‘I want’ vs. mi ta skirbi ‘I write’ (Bickerton 1981: 87). Bickerton (1981: 160) therefore reminds us that “there is a significant difference between creole and Indo-European systems (…). In the latter, the same morphological marking applies to both statives and non-statives; this seems so obvious that it is never even remarked on. In creoles, however, present-reference statives and presentreference nonstatives cannot be marked in the same way.” Moreover, whereas the distinction between present progressive (‘I am writing’) and present habitual (‘I write’) is marked on the verb in the Indo-European lexifiers, this is not so in creole languages, which (barring the usual exceptions) deploy one invariant verb form (Bickerton 1981: 87). These observations bear relevance to the discussion that follows in as far as they suggest that the correspondences between PA and Upper Guinea PC in the domain of stativity, to be discussed below, cannot plausibly be attributed to the closeness of their lexifiers.337

337 For typological and other details on stativity in creoles, see e.g. Holm et al. (2000) and Singler (ed. 1990).

218 5.6.2.

Verbal system

Strong vs. weak stative verbs

The class of stative verbs has been described extensively for both PA (e.g. Birmingham 1970: 81, 82; Maurer 1988: 56–63 & 2003a; Andersen 1990; Kramer 2004: 196) and SCV (e.g. Silva 1985 & 1990; Suzuki 1994; Baptista 2002: 95–101; Pratas 2007 & 2010), with two of these descriptions (Andersen 1990 and Silva 1990) appearing in the same volume (Singler, ed. 1990). However, it was not before Quint (2000b: 151–153) that the remarkable correspondences between the two creoles in this category were noticed and described. This section elaborates on Quint’s (2000b) findings, adding to it, among other things, contrastive data from other Iberian-based creoles. In several creoles (typically those that have lost direct contact with their substrate languages and have undergone at least some degree of decreolization), including PA and SCV, but also, for instance, Gulf of Guinea PC and PLQ, the typical morphological distinction between stative (zero → present338 ) and nonstative verbs (zero = present; imperfective marker → present) has weakened. Some clearly stative verbs are not (or no longer) morphologically marked as such, in as far as their zero- or unmarked form does not yield a present time reference. Rather, just like nonstative verbs, these stative verbs are obligatorily marked by an imperfective marker for present, even in unambiguously stative propositions. In other words, it appears as if the typical morphological marking of stative verbs (zero → present) is no longer productively applied in PA and SCV (and several other creoles), but applied only to a selection of stative verbs, and not to others. The verb ‘have’, for instance, is duly marked as a stative verb in PA (mi Ø tin ‘I have’) and SCV (N Ø ten idem); the verb ‘believe’, on the other hand, though stative beyond doubt, is not morphologically marked as such (115): (115) a. PA

mi ta (/*Ø) kere den I IMP believe in ‘I believe in God’ b. SCV bu ta (/*Ø) kre na you IMP believe in ‘do you believe in God?

Dios God (Maurer 1998: 163) Dés ? God (Mendes et al. 2002: 157)

Following Quint (2000a: 248–252, 2000b: 151–153), I will use the term ‘strong stative verbs’ to refer to those stative verbs that are still morphologically marked as such, i.e. whose unmarked form yields a present reading (e.g. PA tin, SCV ten). 338 Recall that ‘present’ here is used in a relative way, meaning that the time reference is simultaneous with a contextually established reference point.

5.6. A comparison of stative verbs in PA and SCV

219

The term ‘weak stative verbs’ is used in reference to all other stative verbs, i.e. those that are (or can be) semantically stative, but whose unmarked form does not yield a present reading (e.g. PA kere, SCV kre). What follows is a comparison of these two classes of stative verbs in PA and SCV.339 5.6.3. The class of strong stative verbs In terms of the features defined in Holm & Patrick (2007) comparative creole syntax, strong statives correspond to “stative verbs with a non-past reference when unmarked”. A word of caution is in place, however. While the unmarked form of strong statives will indeed yield a present reading, this does not imply that they are obligatorily unmarked for present. Some invariably take either a zero or an imperfective marker in the present. For instance, the strong stative verb PA gusta ‘like’ / SCV gosta di ‘idem’ can be either marked by zero or by ta, without any difference in meaning: PA mi Ø gusta bo = mi ta gusta bo ‘I like you’ (Maurer 2003a: 236) and SCV Pedro Ø gosta di Maria = Pedro ta gosta di Maria ‘Pedro likes Maria’ (Silva 1990: 149; cf. Pratas 2007: 66).340 Table 36 below is adapted from Quint (2000b: 152, 153) and shows that nearly all verbs that are strong stative in SCV are also strong stative in PA. Note that in some cases, the presence or absence of ta allows to distinguish between a stative or nonstative reading respectively, as was illustrated in §5.2 by way of the verb PA yama / SCV txoma ‘to be called’ (105) vs. ‘to call’ (106). To this category we must also count the PA and SCV equivalents of ‘to know’ (PA/SCV sabi and PA konosé/SCV konxe), which, in line with their Iberian etyma (Sp./Port. saber and Sp. conocer/Port. conhecer), may have both a stative 339 Quint’s (2000a, b) distinction between “verbes forts et faibles” in SCV is based on Guillaume (1973). It is important to not confuse this terminology with the categories of strong and weak verbs in Germanic languages. 340 For PA, Maurer (2003a: 236–238; cf. Maurer 1988: 56–63) identified eight strong stative verbs that can be modified by ta as well as by zero without difference in meaning. In addition to gusta ‘to like, love’, these are bal ‘to be worth’, debe ‘to must, need’, dependé ‘to depend’, falta ‘to lack’, gusta ‘to like, love’, meresé ‘to deserve’, parse ‘to appear, seem’ and stima ‘to love’ (cf. Andersen 1990: 71, 72). For SCV, Pratas (2007: 66) notes that “[a] few verbs, like gosta ‘to like’, have an ambiguous behavior”. In addition to SCV gosta, one such verb is SCV mora ‘to reside’, which occurs both with Ø and with ta without any apparent difference in meaning: N Ø mora na Santiago = N ta mora na Santiago ‘I live/reside in Santiago’ (Nicolas Quint, p.c.; cf. examples provided by Lang 2002: 455). Jürgen Lang (p.c.) specifies that, in principle, all strong stative verbs can be preceded by ta in SCV, with different semantic effects, including the expression of futurity, insistence, or to contradict and/or correct the interlocutor (for details, see Lang 1993: 148).

220

Verbal system

Table 36. Strong stative verbs in SCV and PA. (NB: […] = weak stative verb) SCV

PA

konxe ‘to know’ debe, ten ki ‘to have to, must’ sta, ser ‘to be’ gosta (di) ‘to like’ txoma ‘to call, be called’ merese ‘to deserve’ pode ‘to can, be able’ kre ‘to want, love’ sabe ‘to know’ ten ‘to have (possess)’ bale ‘to be worth’ kusta ‘to cost’ meste ‘to need’ – mestedu ‘be needed’ básta ‘to be enough’ dipende ‘to depend’ fálta ‘to lack, miss’ parsi ‘to appear’ [stima] ‘to like’ tene ‘to have (dispose of)’ fika ‘to stay, remain’ mora ‘to reside’

konosé idem341 debe, tin ku idem ta idem gusta idem yama idem meresé idem por, mag (< Du mag ‘may’) idem ke ∼ ker ∼ kié ∼ kier ‘to want’ sa, sabi idem tin idem bal idem kosta idem meste ∼ mesté ‘to need, be needed’ basta idem dependé idem falta idem parse idem stima ‘like, love’ [tene] idem [keda] idem [biba] idem

(‘to know’) and nonstative (‘to get to know / to learn / to find out’) meaning and are morphologically marked accordingly (zero → present for statives; imperfective marker → present for nonstatives). (The ambiguous stative-nonstative properties of SCV konxe and sabi have been discussed and analyzed by Pratas [2010]. For PA konosé and sa(bi), Maurer [1988: 59–61] provides very similar details.) The main discrepancies between PA and SCV are to be found in the four bottom rows of Table 36. First, there is the fact that etymological equivalents of SCV mora (< Port. morar) and fika (< Port. ficar) are lacking in PA. Note, however, that the lack of an equivalent of SCV mora in PA is not very surprising, given the well-known fact that content words (such as the full verb mora) are generally more susceptible to change than function words. Rather, it is striking 341 As noted in §2.1.1.2, Martinus (1996: 175) reported the Aruban/Bonairean form kunse, which does not appear in standard dictionaries and may well be related to GBC kunsi.

5.6. A comparison of stative verbs in PA and SCV

221

to note that nearly all PA verbs listed in Table 36 which are not only full verbs but also auxiliary verbs, can etymologically be traced back to SCV (which will be illustrated systematically in §5.7). One exception is SCV fika, where modern PA has keda (< Sp. quedar). However, an earlier form fika is still found in certain registers as well as in several 19th-century written PA records (cf. §5.7.2.3). Secondly, there is the weak marking of the stative verb PA tene versus the strong marking of the SCV equivalent tene; vice versa, the strong marking of PA stima contrasts with the weak use of its Cape Verdean cognate. But these discrepancies notwithstanding, Table 36 reveals that PA has preserved the zeromarking for present on largely the same set of stative verbs as SCV. Below, I will try to further emphasize the idiosyncrasy of this parallel. For that purpose, it is equally relevant to analyze the class of weak stative verbs. 5.6.4. The class of weak stative verbs Weak stative verbs are verbs generally considered to be stative (e.g. ‘to believe’), or which can, depending on the context, either be stative or nonstative (e.g. most perception verbs), but which, nonetheless, are obligatorily marked by an imperfective marker for present, even in clearly stative propositions. Examples are (115) above and (116) below. (116) a. PA

Ken sa unda e ta (/*Ø) biba? who know where he IMP live ‘Who knows where he lives?’ (Muller 1989: 322) b. SCV e ta (/*Ø) bibe na Mérka he IMP live in America ‘he lives in America’ (Lang 2002: 473)

Silva (1990: 149) comprehensively defines weak statives as “stative verbs (…) that behave like nonstatives”. While this is true, it introduces a comparative problem in as far as nonstatives behave differently in PA and SCV. In PA, due to the introduction of a where there used to be a zero, zero-marked nonstatives are ungrammatical, whereas in SCV, they yield a perfective past reading. Like nonstatives, zero-marked weak statives are ungrammatical in PA but yield a perfective past reading in SCV. This can be exemplified with the stative verb ‘to understand’, which is a weak stative in PA and SCV: thus, PA *mi Ø tende tur kos 1sg-understand-all.thing is ungrammatical, whereas SCV N Ø ntende tudu kuza 1sg-understand-all-thing (Pratas 2010: 217) is acceptable yielding the perfective past reading ‘I understood everything’. This, in turn, means that Maurer’s (1998: 163) definition of weak statives in PA as “stative verbs that can only be modified by ta and not by zero” is correct for PA but does not apply to SCV, and that,

222

Verbal system

reversely, Pratas’ (2010: 220) definition of weak statives as “stative bases with a past reading for non-overtly marked forms” does not apply to PA. This distinct behavior, however, should not conceal the fact that both creoles have weak stative verbs, i.e. statives that require an imperfective marker to yield a present reading. Table 37 below provides a (by no means exhaustive) list of such weak statives. Table 37. Weak stative verbs in SCV and PA (cf. Pratas 2010: 220 for SCV) SCV

PA

GLOSS

komprende kre ∼ kredita ∼ kerdita pertense sirbi bibe (kun)fia odia spéra pensa (∼ kuda) lembra ntende/ubi odja ∼ djobi

komprondé kere pertenesé sirbi biba konfia odia spera pensa korda tende/oi weta ∼ mira

‘to understand’ ‘to believe’ ‘to belong’ ‘to serve’ ‘to live’ ‘to trust’ ‘to hate’ ‘to hope, wait’ ‘to think’ ‘to remember’ ‘to hear’ ‘to see’

Arguably the most interesting member of this class of statives is PA kere / SCV kre ‘to believe’, exemplified previously in (115). Due to its unambiguous semantic stativity, it is often cited as a paradigm example of how certain stative verbs are no longer marked as such in PA (cf. e.g. Birmingham 1970: 83; Maurer 2003a: 235; Kramer 2004: 196). Clearly, the weak marking of PA kere / SCV kre does not correlate to any patterns found in Spanish (*estoy creyendo) or Portuguese (*estou a crêr) and the explanation has to be sought elsewhere. One rather compelling explanation is provided by Birmingham (1970: 83), who notes that in colloquial speech, PA kere ‘to believe’ is often homophonous with ker ‘to want’, both being reduced to /ke/.According to Birmingham, then, the absence or presence of ta serves to distinguish between them: “Since kjer ‘to want’ and kere ‘to believe’ tend to fall together in ke, the distinction is usually made as follows: mi ke ‘I want’, mi ta ke ‘I believe’” (Birmingham 1970: 83). Interestingly, this account can be applied to SCV as well, where kre means both ‘to want’ and ‘to believe’, while the distinction between the two meanings is made by the absence or presence of ta: SCV N kre ‘I want’ vs. N ta kre ‘I believe’. In fact, on the basis of the different forms for ‘to want’ and ‘to believe’ in PA (ker and kere), SCV (kre), Fogo CV (kere), and GBC (kere), the ambiguous proto-Upper Guinea PC

5.6. A comparison of stative verbs in PA and SCV

223

form *kere ‘to want / believe’ can be hypothesized (cf. also §5.7.1.2). It may thus well be that the practice of distinguishing between ‘to want’ and ‘to believe’ by means of the absence or presence of ta, as displayed by PA and SCV, dates back to the proto-creole. 5.6.5.

Contrastive analysis

Below, several examples are given of how the strong/weak marking of stative verbs coincides in PA and SCV, while differing from other Iberian-based creoles: – While the stative verb ‘to believe’ is weak in PA and SCV (115), it is strong in PLQ342 (117a) and in Gulf of Guinea PC, exemplified here by ST (117b) and Annobonese (117c): (117) a. PLQ I Ø kreé lo nu I believe it NEG ‘I don’t believe it’ (Armin Schwegler, p.c.) b. ST N Ø klê n desu I believe in God ‘I believe in God’ (Rougé 2004b: 123; cf. Major 2006: 86; Maurer 2009a: 72 for PRI) c. AB Xo sai m’ Ø kele thing DEF I believe ‘This I do believe’ (Post 2000: 137) – For PRI and ANG, Maurer (1994b: 418 & 2009a: 72 on PRI; 1995: 69 on ANG) identifies weak stative verbs, meaning that they require an imperfective marker for present reference. In this class we find the equivalents of ‘to appear, seem’, ‘to be worth’, ‘to be’ and ‘to be called’, all of which are in fact strong in PA and SCV (cf. Quint 2000b: 181). Compare, for instance (118) with (119): (118) a. PA

E Ø jama Maria she be called Mary ‘She is called (named) Mary’ (Birmingham 1970: 123; cf. Maurer 1988: 62 & 2003: 237f.n.; Holm et al. 2000: 137)

342 More about stativity in PLQ can be found in de Friedemann & Patiño Rosselli (1983: 126, 127), Maurer (1987a: 42, 43), and Schwegler & Green (2007).

224

Verbal system

b. SCV N Ø tchoma Leonor I be called Leonor ‘I am called Leonor’ (119) PRI

(Mendes et al. 2002: 121)

N ka sama Mene I IMP be called Mene ‘I’m called Mene / My name is Mene’ (Maurer 2009b: 73; cf. Lorenzino 2007: 8 for ANG;)

– In PLQ, the verb siribí ‘to serve’ is strong: in stative propositions, its unmarked form yields a present reading (120). By contrast, PA / SCV sirbi ‘serve’ is weak (cf. Table 37) and thus at all times requires an imperfective marker to yield a present reference, even in clearly stative propositions (121): (120) PLQ

ané Ø siribí nu they serve NEG ‘they don’t serve (i.e. are useless)’

(Dieck 2002: 156, 157)

(121) a. PA

e binaguer aki no ta (/*Ø) sirbi (pa nada) DEF vinegar here NEG IMP serve for nothing ‘this vinegar is useless’ (Goilo 1974: 56) b. SCV e ka ta (/*Ø) sirbi pa nada he NEG IMP serve for nothing ‘that doesn’t serve much’ (Mendes et al. 2002: 362)

– Baxter (1988: 60, 61), in reference to the Asian Portuguese creole Papia Kristang (PK), notes that the perception verbs PK lembrá ‘to remember’ and sintí ‘to feel’receive a present reading when unmarked in stative propositions (122). By contrast, as shown in Table 37, the PA/SCV equivalents of ‘to remember’ and ‘to feel’ are weak statives as they do not allow for a present reading when unmarked. (122) a. PK

b. PK

eli Ø sintí bos keré ngganá ku she feel you want deceive with ‘she feels that you want to deceive her’ n´ gka Ø lembrá NEG remember ‘I don’t remember’

eli her (Baxter 1988: 129)

(Baxter 1988: 135)

5.6. A comparison of stative verbs in PA and SCV

225

Although the contrastive analysis just provided is no where near exhaustive343 , the sum of the negative evidence and the correspondences in Tables 37 and 38 amply support Quint’s observation that the lists of strong stative verbs of PA and SCV are extremely similar to one another as well as his compelling conclusion: “This extraordinary resemblance cannot be due to mere chance. How, without a common origin, could SCV and PA have recreated the same list of strong stative verbs with so many etymologically close forms at a distance of 6.000 kilometers?”344 (Quint 2000b: 153). It is interesting to note, finally, that with only six strong stative verbs, the Barlavento varieties of CV have a considerably more reduced list of strong statives than SCV and PA (cf. Table 37). For instance, SCV e Ø konxi ‘he knows’ requires an imperfective marker in São Nicolau CV: el ta konxê ‘(él) conoce’ (Quint 2000b: 77; cf. Almada 1961: 117, 118). This can be taken to suggest that the paradigm of strong statives in PA has remained remarkably stable after its hypothesized split from Upper Guinea PC. 5.6.6.

Digression: The case of GBC

It is revealing to note that the strong-weak opposition described above for PA and SCV is not found in GBC. In PA and SCV, the strong marking (Ø → present) of stative verbs, as Maurer (1998: 163) puts it, “es solamente histórica”, in as far as it is restricted to a closed, more or less fossilized set of stative verbs. This is also why Quint speaks of “la classe fermée (…) des verbes forts” in SCV (2000a: 241f.n.). In GBC, on the other hand, the morphological [±stative] marking of a given predicate is entirely dictated by the semantic contents of the proposition in which it is embedded: predicates take a zero for present in stative propositions and require an imperfective marker for present in nonstative propositions. Perception verbs, for instance, known to have both a stative and nonstative use, are weak in PA and SCV (see Table 37), meaning that they require an imperfective marker for present both in stative and nonstative propositions. The corresponding verbs in GBC (e.g. ntende ‘understand’, pensa ‘feel’, ubi ‘hear, understand’, 343 It might be fruitful for future research to include contrastive data from creoles from different lexifier languages. One example from Gullah suffices to illustrate this point: Frank (2007: 162) asserts that in Gullah the verb meaning ‘to think’ receives a present reading when unmarked. In PA and SCV, on the other hand, pensa is obligatorily marked for present. 344 Original quote: “Cette ressemblance extraordinaire ne peut pas être due au seul hasard. Comment, sans une origine commune, le badiais et le papiamento auraientils pu recréer, à 6.000 km de distance, la même liste de verbes forts avec tant de formes étymologiquement proches?”

226

Verbal system

kuda ‘to think’, lenbra ‘to remember’, oja ‘to see’, and spera ‘to wait, expect, hope’), on the other hand, will be marked as the semantic contents of the proposition ([±stative]) demands (cf. Peck 1988: 229; Quint 2000a: 251, 2000b: 107; Solovova 2004: 163; Baptista, Mello & Suzuki 2007: 54). While it is widely recognized that GBC and SCV are sister creoles, the divergence in the area of stativity can be readily explained by the fact that GBC, unlike PA and SCV, has remained in contact with several West-Atlantic and Mande languages, where the strong-weak distinction does not apply either. Rather, in these languages the semantic contents of a predicate dictate what preverbal marking will be used (cf. Yaguello’s comments on Wolof, cited in Quint 2000a: 250). Bickerton (1975: 30) described the same state of affairs for basilectal Guyanese English Creole: “the stative – non-stative distinction in Guyanese Creole is a semantic one entirely: that is to say, it is not the case that specific lexical items are marked unambiguously [+stative] or [−stative], rather that these categories apply to propositions irrespective of their lexical content”. We may summarize that in GBC all predicates are marked in function of their semantic content. This also explains why adjectives, which are analyzed as verbs in GBC, are marked by zero (viz. do not require a copula) when they express a state (cf. §5.7.2.1), a feature found in other basilectal creoles as well (cf. Holm 1988: 85, 86 for details, cf. also Sebba 1997: 186, 187). It is plausible to assume that the state of affairs described here for presentday GBC reflects the state of affairs in proto-Upper Guinea PC. At some point in the development of Upper Guinea PC (at least prior to the separation of PA), a closed class of strong statives was formed, while others would lose their morphosyntactic stativity due to the extension of the imperfective marker to stative predicates.345 GBC, which after its separation came in renewed contact with African languages, then fully restored the original, semantically driven, marking of stative vs. nonstative propositions, whereas the closed class of strong statives was preserved, and has remained remarkably stable, in PA and SCV.

5.7. Auxiliary verbs As noted previously, if PA and Upper Guinea PC are genetically related, we can expect to find evidence for this in the functional domain rather than in the purely lexical domain of their vocabulary. Translated to the category of verbs, the lexical/functional distinction conforms to that between main verbs (viz. 345 Maurer (2003a: 235) comprehensively describes this diachronic process as the absorption of Ø by ta.

5.7. Auxiliary verbs

227

verbs with lexical meaning only) and auxiliary verbs (verbs with relatively little lexical meaning and capable of taking a verbal complement). Consequently, the hypothesis examined in the present study predicts that the majority of main verbs in the PA lexicon will have Spanish (and to a lesser extent Dutch) etyma, while the auxiliary verbs are closer to Portuguese, or rather to Upper Guinea PC. This is indeed the distribution we find in PA’s verbal inventory. Of course, it is often futile to debate the Spanish vs. Portuguese etymology of a given PA verb, especially there where the possible etyma are (nearly) homophonous in Spanish and Portuguese (e.g. PA sinti ‘to feel’< Sp./Port. sentir, etc.). Nevertheless, in the category of PA main verbs, it is relatively easy to establish a clear predominance of items with a probable Spanish etymology. For instance, high frequency verbs such as PA yama ‘to call, be called’, hasi ‘to make’, donar ‘to give’, mira ‘to watch, look’ are all quite clearly derived from Spanish (llamar, hacer, donar, mirar) rather than from Portuguese (chamar, fazer, doar, olhar). In fact, drawing on a list of 375 verbs composed by Hoyer (1918), Lenz (1928: 250, 251) noted that “[l]a enorme mayoría son voces que guardan casi por completo la forma española” (cf. his etymological statistics, Lenz 1928: 260). A focus on auxiliary verbs provides the reverse pattern: as I will show below, the dominant lexifier language in this category is Portuguese. More precisely, the auxiliaries discussed below show a significant overlap in form, meaning and function with their Upper Guinea PC cognates. I will subdivide the auxiliary verbs into modal auxiliaries (§5.7.1), copular verbs (§5.7.2), and other auxiliaries (§4.7.3). 5.7.1.

Modal auxiliaries

The importance of modal verbs in establishing genetic ties lies in their typically low borrowability rate and, consequently, high stability in cases of language change (Muysken 2008: 9; but see Matras 2009: 185–187), properties which may of course correlate with their being among the verbs with the highest frequency in many languages (cf. e.g Krug 2001: 311 for English; Lipski 1988: 31 for Spanish, Portuguese and Iberian-based creoles). The three modals to be discussed below have been discussed jointly for PA in an interesting paper by Maurer (1986c; cf. 1988: 274–312). Note that all three take zero for present (cf. Table 36 in §5.6.3). 5.7.1.1. PA por, Upper Guinea PC podi For phonetic reasons, the PA form por ‘to can, be able to’ is likely to derive from an earlier paroxytonic proto-form *pode. We can discard the oxytonic Sp./Port. infinitive poder as the etymon: if that had been the source, then the stressed /e/ of the final syllable would in all likelihood have been preserved in

228

Verbal system

the derived PA form, as suggested by forms such as PA pone ‘to put’ and kore ‘to run’, from the infinitives Sp. poner and Sp./Port. correr respectively. Cf. also the PLQ verb polé ‘to can, be able to’ from Sp. poder (Schwegler & Green 2007: 283). The monophthong /o/ in PA por, moreover, allows us to discard a derivation from Spanish puede ‘[3sg] can’, which, for instance, was the source for Zamboangueño pwéde (Lipski & Santoro 2007: 377). If these considerations are correct, the most plausible remaining etymology of PA por is provided by the paroxytonic form Port. pode ‘[3s] can’, which was integrated into Upper Guinea PC as podi. Conveniently, the hypothesized Early PA form *podi is still reflected in the lexicalized PA adverb podisé ‘may be’ < Port. pode ser. Moreover, the hypothesized development from Portuguese pode (or Upper Guinea PC podi) to PA por involves the apical change /d/ > /r/, which is mirrored in almost all PA words etymologically ending in unstressed -/dV/ as well as the Early PA past participle ending -r, discussed at length in §4.2. 5.7.1.2. PA ke∼ker, Upper Guinea PC kere The form kere is the norm in the Casamance variety of GBC (∼kiri in GuineaBissau) meaning both ‘to want’ and ‘to love’ (Chataigner 1963: 64; Rougé 1994: 146; 2004a: 244).346 In the phonetically faithful transcriptions of Bal (1983b), the Casamancese form kere is alternately realized as ke, so that we can juxtapose GBC ke∼kere to PA ke∼ker. Compare: (123) a. GBC N ké papiá ku bo. I want talk with you ‘I want to talk to you’ (Bal 1983b: 26) b. PA e ke papia kuné he want talk with+him ‘he wants to talk to him’ (Kouwenberg & Murray 1994: 37) Though present-day SCV has kre ‘to want’ (< Port. querer), the original form *kere was (and may still be) found in the Fogo variety of CV, at least judging from Parsons (1923) who spelled the verb as . Also, as argued in §5.6.4, the idea that PA ke∼ker and SCV kre trace back to the same proto-form *kere 346 There is another verb available in GBC to denote ‘to want’, namely meste in Casamance and misti in Guinea-Bissau (cf. also §5.7.1.3). I am not sure how GBC ke∼kere semantically relates to GBC meste∼misti, nor do I have concrete data on their geographic or dialectal distribution. As far as I can tell, however, ke∼kere is more typical of the Casamance variety and less frequent in Guinea-Bissau (Bal 1983b: 26; Rougé 2004b: 157).

5.7. Auxiliary verbs

229

is suggested by the way speakers distinguish between ‘to believe’ and ‘to want’, namely by means of the respective presence or absence of the marker ta. By contrast, a cognate of Port. querer is not attested in Gulf of Guinea PC, which instead uses mêsê (< Port. mester ‘need’) in the sense of ‘to want’(Maurer 2009a: 222, 223; Rougé 2004a: 244). On the other hand, Port. queria (the 3sg past imperfective of querer ‘to want’) was probably the source of an auxiliary kia, found in ST, denoting the unfulfilled intention to perform the action expressed by the complement verb (Rougé 2004a: 244). 5.7.1.3. PA meste – mesté, SCV meste – mestedu From a historical-linguistic point of view, PA meste∼mesté constitutes one of the most interesting items of the PA grammar. The etymon of PA meste∼mesté is the Old Portuguese noun mester ‘to need’ (= Sp. menester) which fell out of use in the course of the 16th century (Kihm 1994: 4; cf. Rougé 2004a: 207). Since the settlement of Curaçao did not start prior to the mid-17th century, the presence of 15th-16th century Portuguese items in PA requires an explanation. A plausible one is that these entered PA by way of Upper Guinea, where a creole based on 15th-16th century Portuguese is and was spoken. It is unlikely to be a coincidence, therefore, that we find the only plausible cognates of PA meste∼mesté in the SCV verb-participle pair meste-mestedu (see also §4.3.5.2). (Note that PA meste∼mesté is but one of several PA features that can be attributed to 15th-16th century Portuguese and which all have an equivalent in Upper Guinea PC; cf. §6.4.) Perhaps the Old Portuguese noun mester 347 already functioned as an autonomous verb in some colonial varieties of Old Portuguese (Quint 2000b: 230). This is suggested by the fact that mester has verbalized variants not only in PA and Upper Guinea PC but also in several other Portuguese-based creoles (e.g. Gulf of Guinea PC mêsê ‘to want’ [Rougé 2004a: 207]; see Quint [2000b: 230] for a good overview). Note, however, that the formal, semantic and syntactic match between PA meste∼mesté and SCV meste-mestedu is as perfect as no other.348 347 This noun (with the variant mister) was used in Old Portuguese in combinations with the auxiliaries haver and ser in the verb phrases haver mester ‘to need’ (later ter mester) and ser mester ‘to be needed/necessary’ (Lang 2002: 441; Rougé 2004a: 207). Combinations with mester later fell out of use in favor of precisar ‘to need’ and ser preciso ‘to be necessary’. 348 PA mesté also occurs as a noun meaning ‘need, desire’ (van Putte & van Puttede Windt 2005a: 290). As such, mesté occurs particularly in combination with the verb tin ‘to have’: PA tin mesté di algun kos/hende ‘to (be in) need (of) some-

230

Verbal system

5.7.2.

Copular verbs

5.7.2.1. PA / Upper Guinea PC ta PA ta functions not only as a general imperfective (§5.1) and a future tense marker (§5.3.3.1), but also as a portmanteau (equative, locative, and adjectival) copula. The discussion of ta as a copula is of some significance to the hypothesis examined in the present study, since this property of ta is often erroneously assumed to be alien to Upper Guinea PC. I will show below that the copula ta in fact links PA and Upper Guinea PC, while setting them apart from the Portuguese-based creoles of the Gulf of Guinea and Asia. The distribution of copulas in CV and GBC presented in Maurer (2009b: 43) is as follows: Table 38. Copulas in Upper Guinea PC and PA according to Maurer (2009b: 43)

CV GBC PA

nouns

adjectives

locatives

e/ser – sta zero – i/sedu ta

e/ser – sta zero ta

sta sta ta

This table is incomplete for several reasons. First, it does not mention the fact that ta functions as a locative (124) and adjectival copula (temporary states only) (125) in BaCV (Lipski 2002: 67; Holm & Swolkien 2006: 213; Swolkien 2009: 5): (124) a. SVCV El ta lá na kaldéra it COP there in pot ‘It [the food] is there in the pot’ b. PA Nos ta na Kòrsou we COP in Curaçao ‘We are in Curaçao’

(Swolkien 2009: 5)

(Maurer 1988: 328)

thing/somebody’. The noun mesté may well have its equivalent in the Casamance variety of GBC, where the noun mesteda ‘need, necessity, desire’ (derived from the verb meste ‘want’) is in use (Peck 1988: 146; Rougé 2004a: 207). The hypothesized path from GBC mesteda to PA mesté is completely regular and follows the typical path of erosion from etymological -/dV/ to PA -/Ø/, which has been discussed in §4.2.2. Moreover, just like PA mesté, GBC mesteda can follow the verb ‘to have’: GBC e tem mesteda pa fasi … 3s-have-need-to-make ‘he needs to make…’ (Schuchardt 1889a: 302). Note that GBC meste and mesteda are realized as misti and mistida in Guinea-Bissau (Rougé 2004a: 207).

5.7. Auxiliary verbs

(125) a. SVCV N ta doente / kansód I COP ill tired b. PA

231

(Dominika Swolkien, p.c.)

Mi ta malu / kansá I COP ill tired ‘I am ill / tired’

Secondly, adjectives in GBC behave like verbs (e.g. Baptista, Mello & Suzuki 2007: 68), which implies that adjectives may be marked not only by zero, as Table 38 correctly indicates, but also by ta. The distribution of zero and ta before adjectives in GBC is generally as follows: adjectives expressing non-inherent, temporary qualities are preceded by a zero copula, as in (126); adjectives expressing permanent qualities or states, on the other hand, take ta (Kihm 1994: 91 and 2007: 251), as in (127a) and, further down, (127b). (126) GBC

N Ø dwenti I ill ‘I am ill’

(127) a. GBC N ta dwenti I IMP/COP ill ‘I am (incurably) ill’

(Kihm 1994: 92)

(Kihm 1994: 92)

It is important to add that, although GBC ta ultimately derives from Port. estar (rather than from Port. ser), to translate (127a) above and (127b) below speakers of Portuguese take recourse to ser (rather than to estar). Clearly, this suggests that the copular use of GBC ta illustrated here is an original creole feature rather than a feature modelled on the lexifier (Kihm 1994: 92). (127) b. GBC Lubu ta branku hyena IMP/COP white ‘Hyenas are white’ (Kihm 2007: 250; cf. examples in Kihm 1994: 34; 91, 92) c. PA Lechi ta sano milk COP healthy ‘Milk is healthy’ (Kester & Schmitt 2007: 120) Given the [+verbal] properties of adjectives in GBC, one could argue about whether to define GBC ta in (127a, 127b) as a copula or as an imperfective marker.349 But the exact lable should not concern us here; the main point is 349 To reflect this functional ambiguity, Kihm (1994: 92) defines GBC ta as an imperfective auxiliary.

232

Verbal system

that one can easily imagine how these predicative properties of ta could have facilitated its reanalysis as a true adjectival copula, a task it performs in presentday PA (127c). A third shortcoming of Table 38 is the lack of a diachronic perspective. If scholars dealing with adjectival and nominal predication in Upper Guinea PC agree on anything, it is that the copulas listed for Upper Guinea PC in Table 39 (SCV e/ser 350 –sta and GBC i/sedu351 ) constitute postformative innovations (Quint 2000a; Kihm 2007; Baptista 2007c; Lipski 2002: 67). Below, I will discuss the central arguments in support of that assumption. A first indication of the relatively reduced time-depth of the copulas SCV e and GBC i is that they precede the negator rather than the adjective in predicates, as illustrated here for SCV: kumida e ka sabi food-COP-NEG-tasty ‘the food isn’t tasty’ (Quint 2000a: 267; cf. Kihm 2007: 256 for GBC). In SCV, moreover, the copula can optionally be omitted, e.g. bo e ka dodu = bo Ø ka dodu ‘you’re not stupid’ (Baptista 2007c: 195), without a change in meaning. Moreover, as Kihm (2007: 256) points out, suppletion is rare in Upper Guinea PC (like in most creoles), and where it is attested, one can be quite certain that it concerns postformative innovations due to pressure from Portuguese. Indeed, the suppletive forms of the ser-derived copula in SCV (e, era, ser) and GBC (i, (y)era, sedu) are suspiciously close to Portuguese (é, era, ser) (Kihm 2007: 249, 250). With respect to SCV, it must still be added that the semantic distribution of sta vs. e/ser in SCV, though not fully, still quite closely reflects that of estar vs. ser

350 The short form e (alternately written as ê or é) has a past/anterior counterpart era and alternates with the full form ser (past serba). The full form ser is confined to the following contexts: following other verbs (e.g. bu ten ki ser rápidu 2sg-have-RELbe-quick ‘you have to be quick’, Lang 2002: 636) and TMA markers (e.g. Shell s’ta ser renubádu Shell-PROG-be-renovated ‘Shell will be renovated’, Lang 2002: 196), or in subordinate clauses introduced by a preposition (e.g. pa ser presidenti for-bepresident ‘in order to be president’, Baptista 2002: 156, based on Veiga 1995: 367). In some contexts, the locative copula SCV sta can take on the role of an equative copula (Quint 2000a: 256). For further discussion, see Lang (2002: 164, 165), Quint (2000a: 252–257), Baptista (1999, 2007c) and Kihm (2007: 249–252). 351 Roughly the same distributional rules as those described in the previous footnote for SCV e/ser apply to GBC i/sedu(∼sedi). Just as SCV, GBC also has two past forms, seduba and (y)era. For discussions and analyses of the synchronic distribution and the (controversial) diachronic development of these markers, consult, for instance, Peck (1988: 133–142), M’Bodj (1993: 329, 330), Kihm (1994: 37–40) and Baptista, Mello & Suzuki (2007: 70, 71). Cf. the specialized papers on copulas in Upper Guinea PC by Baptista (1999, 2007c) and Kihm (2007).

5.7. Auxiliary verbs

233

in Portuguese (Baptista 2002: 102) and that no other Iberian-based creole has preserved this copular distinction (Bartens 1996: 18).352 All these observed patterns of course suggest that the copulas in question did not form part of the original Upper Guinea PC grammar. This is why Quint (2000a: 267; cf. 2000a: 257; 2000b: 154) believes that the use of the copula e/ser was “très reduit, voire nul, en créole ancien” (‘very restricted, or none, in the ancient creole’) and why Baptista (2002: 109), similarly, assumes that “Cape Verdean Creole (…) started out with copulaless constructions”. Based on what has been said thus far, it is not proven, but nonetheless very likely that the overt copulas SCV e/ser, GBC i/sedu and SCV/GBC sta are postformative acquisitions. From a creole typological perspective, this assumption is not particularly far-fetched, as the lack of lexifier-derived copulas is typical of many creoles (McWhorter 2011: 36). Instead, it seems highly likely that the GBC predicate system exemplified in (126) and (127) reflects the original state of affairs for Upper Guinea PC, which then must have had a predicate system based solely on zero and ta (cf. Quint 2000b: 207), the latter being ambiguous between an imperfective aspect marker and a copula. While GBC still productively draws on this predicate system (126, 127), SCV does not. However, we do find crystallized traces of zero predicates as well as of ta predicates also in SCV. As for zero, we find a lexicalized set of adjectives that still take a zero copula in basilectal varieties of SCV where, according to modern-day SCV predication rules, one would expect either e(∼ser) or sta to appear (Quint 2000a: 256). As for ta, Quint (2000a: 254; 2000b: 154) reports that in the interior of Santiago a fossilized set of adjectives (typically doxi ‘sweet’, limpu ‘clean’, and moli ‘soft’) still take ta, rather than e(∼ser) or sta, as an adjectival copula. Now, if PA inherited a predicate system based on zero and ta, the subsequent extension of ta to those contexts in which the Upper Guinea PC ancestor used zero would be a relatively minor – and, arguably, linguistically economic – change respective to the original system. 352 What Table 38 does not mention is that GBC sta is not only a locative copula but may occasionally also serve as an adjectival copula (Kihm 1994: 91; cf. Holm 2000: 198; Mbodj 1991: 70; Peck 1988: 133). An example is GBC N sta dwenti ‘I’m ill’ (Kihm 1994: 91). Crucially, this predicate is not a variant of N ta dwenti ‘I am (incurably) ill’ (127a), but of N Ø dwenti ‘I am ill (and hope to recover soon)’ (126) (cf. Port. estou doente). In other words, GBC sta appears only in adjectival predicates where Portuguese would use estar, and where GBC would more commonly use a zero copula. Clearly, this suggests that GBC sta was borrowed with the purpose of filling a ‘gap’ (the zero copula) in the original grammar (Kihm 1994: 92), although the zero copula illustrated in (126) is still more common in adjectival predicates than sta is.

234

Verbal system

Presumably, then, whereas PA chose to generalize the use of ta as a copula, SCV and GBC developed the copulas SCV e(∼ser)/sta and GBC i/sedu/sta with the purpose of bringing their predicate system more in line with that of Portuguese. Note that it is not unusual for languages to adjust their copula patterns to that of surrounding languages. For instance, Heine & Kuteva (2005: 129) discuss how the Arawakan language Tariana “appears to be drifting from the copula-less North Arawak profile towards a copula language akin to East Tucanoan languages”. For PA, it is not so clear whether contact somehow facilitated the extension of ta to all predicate types. Contact with Spanish is unlikely to have been an important factor here, since, just like Portuguese, Spanish distinguishes between permanent, inherent states (ser) and temporary, non-inherent states (estar), while PA does not. On the other hand, the Dutch copula zijn ‘to be’ is used for all predicate types and may have served as a model. We may also point out that Gulf of Guinea PC, just like PA, has one copula for all predicate types (Maurer 2009b: 43). As with preverbal taba (§5.4.1), this parallel need not be a coincidence, although any conclusions in this regard are rather tenuous (cf. §6.3.1). Table 39. The copula in (proto-)Upper Guinea PC, PA, Gulf of Guinea PC, the Asian Portuguese Creoles, and PLQ353

SoCV BaCV GBC proto-Upper Guinea PC PA Gulf of Guinea PC Asian Port. Creoles PLQ354

nominal copula

adjectival copula

locative copula

e(∼ser) e(∼ser) zero / i / sedu *zero

e(∼ser) / sta e(∼ser) / ta zero(∼sta) / ta *zero / *ta

sta ta sta *ta

ta sa

ta sa

ta sa

zero / (cognate of Port.) tem ta∼é/era∼sendá∼fue

zero / (cognate of Port.) tem ta∼é/era∼sendá∼fue

(cognate of Port.) tem ta∼é/era∼sendá∼fue

353 The sources drawn upon are M’bodj (1991: 70), Holm (2000: 198), Quint (2000b: 207), Maurer (2009b: 43), Lipski (2002) and Holm & Swolkien (2006: 213) and Schwegler & Green (2007: 288, 289) as well as the other sources mentioned throughout this subsection. 354 One interesting issue that for reasons of space has not been dealt with here is the variety of copulas found in PLQ, one of which is ta (see e.g. Lipski 2002 and Schwegler & Green 2007: 288, 289 for discussions).

5.7. Auxiliary verbs

235

Even if the diachronic analyses presented above were proven to be inaccurate, PA still shows considerably more overlap with Upper Guinea PC than with the Portuguese-based creoles of the Gulf of Guinea and Asia. Whereas the use of ta as a copula in PA, GBC and BaCV is a synchronic reality, Gulf of Guinea PC uses the copula sa355 (Lipski 2002: 67; Maurer 2009b: 43) while the Asian Portuguese creoles all have a copula derived from the Portuguese verb ter ‘have’ (Lipski 2002: 69; Maurer 2009b: 43). Table 39 presents a reassessment of copula marking in PA and Upper Guinea PC (including its hypothesized protoforms) and includes the contrastive data from Gulf of Guinea PC and the Asian Portuguese creoles. 5.7.2.2. PA / Upper Guinea PC bira The high-frequency copular verb PA bira ‘to become, grow, turn into’ has its equivalent in SCV bira and GBC bida. As a main verb, PA / SCV bira has the physical meaning ‘to turn’, just like its etymon, which I assume to be Portuguese virar ‘to turn’ rather than Spanish virar ‘to tack’ (see further below). GBC integrated the main verb vira ‘to turn’, assigning the exclusive task of copular verb to the older form bida (Scantamburlo 2002: 77, 632). As a copular verb, PA / Upper Guinea PC bira∼bida ‘to turn into, grow, become’ combines with adjectives (128, 129) and nouns (130) (Scantamburlo 2002: 77 for GBC; Lang 2002: 76, 77 for SCV; Birmingham 1970: 96 for PA). Note that, to translate (128–130), mainstream Spanish and Portuguese would take recourse to reflexive verbs such as Sp. volverse/hacerse/Port. fazerse/tornar-se or the copular verbs Sp. quedar(se)/Port. ficar(-se). (128) a. PA

e mucha a bira grandi DEF child PFV turn big ‘The child has grown big’ (van Putte & van Putte-de Windt 2005a: 64). b. GBC I bida dja garandi he turn already big ‘He has grown big (already)’ (Incanha Intumbo, p.c.)

355 In PRI, ta exists but according to Maurer (2009: 102) appears “exclusively in comparative sentences with mo∼modi ‘like’”. One may speculate that this isolated use of ta results from contact with speakers of CV. In the second half of the 19th and the first two decades of the 20th century, a considerable number of Cape Verdeans migrated to Principe to work on the local cacao plantations (Batalha 2004: 38; Nascimento 2008: 55).

236 (129)

Verbal system

a. PA

bira

šegu

(Birmingham 1970: 96)

b. GBC bida segu (Scantamburlo 2002: 77) turn blind ‘to go blind’ (130) a. PA

e hòmbër a bira un tronkon’i palu DEF man PFV turn a stump-of tree ‘the man has become/turned into a great big fellow’ (Maurer 1988: 319) b. SCV Enton, kabalinhu bira un ponba Then little horse turn IND dove ‘Then, the little horse turned into a dove’ (Lang 2002: 96)

The following examples show how the analytic creole structure [bira∼bida + adjective] substitutes for the synthetic Iberian inchoative verbs, such as Sp. oscurecer(se)/Port. escurecer(-se) (131) and envejecer(se)/envelhecer(-se) (132): (131)

a. PA

bira

skur

(van Putte & van Putte-de Windt 2005b: 162)

b. GBC bida sukuru (Scantamburlo 2002: 77) turn dark ‘grow dark’ (132)

a. PA

bira bieu

(van Putte & van Putte-de Windt 2005b: 634)

b. SCV bira bedju (Rougé 2004a: 285) turn old ‘grow old’ Markey & Fodale’s (1983: 74) PA-related claim that “bira can only be followed by an adjective or a noun, but never by a verb” is inaccurate: PA / Upper Guinea PC bira∼bida can indicate the starting point of a nonpunctual action expressed by ta + V complement clauses (133, 134) (cf. Martinus 1996: 172 for PA; Pereira 1999: 114, 115, Lang 2002: 77 and Quint 2010: 141 for SCV): (133) a. PA

el a bira ta bai dòktër he PFV turn IMP go doctor ‘he started seeing the doctor’ b. SCV Rapás bira ta grita boy turn IMP scream ‘The boy started screaming’

(134) a. PA

El a bira ta hari he PFV turn IMP laugh ‘he started laughing’

(Maurer 1988: 266)

(Lang 2002: 149)

(Martinus 1996: 172)

5.7. Auxiliary verbs

b. FCV ele bira ta papia sima papagaia he turn IMP talk like parrot ‘he started talking like a parrot’

237

(Parsons 1923: 27)

Although the verb virar exists also in Spanish, Spanish is unlikely to have been the source for PA bira. Afterall, virar is much more typical of mainstream Portuguese (where it has the general meaning ‘to turn’ as well as several related meanings) than of mainstream Spanish (where it is primarily used as a nautical term meaning ‘to tack’).356 Moreover, although PA bira has several lexical meanings, the principal meaning of Spanish virar ‘to tack’ is not one of them. Also, as is widely agreed upon, the less specific the lexical content of a given (motion) verb, the better this verb lends itself for grammaticalization (e.g. Brinton & Traugott 2005: 109). Thus, while it is easy to imagine a grammaticalization path from Port. virar ‘to turn’ > PA bira ‘to become (aux)’, the path from Sp. virar ‘to tack’ > PA bira ‘to become (aux)’ is rather unlikely. The claim of a Portuguese origin for PA bira is further strengthened by the fact that virar acquired copular functions not only in PA and Upper Guinea PC, but also in Brazilian Portuguese (e.g. virar um homem ‘to become an adult’, Holm et al. 1998: 95), Gulf of Guinea PC (bila∼vya) (Ferraz 1979: 85–87; Rougé 2004a: 286) and Papia Kristang (Baxter 2004: 16). To the best of my knowledge, the copular use of virar is not documented for any – be it creolized or otherwise non-standard – variety of Spanish. Note, finally, that the use of a motion verb ‘to turn’ as a copular verb ‘to become’ is not uncommon cross-linguistically (as English ‘to turn into’ shows) and is attested for creoles with different lexifiers (Holm et al. 1998: 95), including English-based, where ton ‘to turn’ is used in this manner, as in Jamaican Creole Fani ton waiz ‘Fanny has become wise’ (Ferraz 1979: 86). The use of ‘to turn’ in the sense of ‘to become’ is thus far from idiosyncratic, though that should not disguise the perfect match between PA bira and Upper Guinea PC bira∼bida. 5.7.2.3. Early PA / Upper Guinea PC fika The equivalent of the Upper Guinea PC auxiliary fika ‘to stay, remain’ (< Port. ficar) in present-day PA is Spanish-derived keda (< quedar). However, the Portuguese-derived equivalent fika is still attested in PA, although typically with the restricted lexical meaning ‘to get bogged down, be delayed’, as in PA por 356 According to Penny (2002: 281), virar was in fact borrowed into Spanish from Portuguese along with several other nautical terms. However, this borrowing presumably occurred as early as the late Middle Ages (Penny 2002: 280), so that this fact alone does not suffice to exclude a Spanish etymology for PA virar.

238

Verbal system

ta ku el a fika na kas di su amigu can-be-that-3s-fika-in-house-of-his-friend ‘he probably got delayed at his friend’s house (Ratzlaff 1992: 78). PA fika is included also in Henriquez’s (1988: 21) Sephardic PA dictionary with the annotation ‘archaic’ and a meaning description close to that given by Ratzlaff, as well as in Maduro’s (1953: 76) orthographic dictionary, which provides “fica (poco usá) [‘infrequent’]” (see also Maurer 1998: 199f.n.). According to van Putte & van Putte-de Windt (2005a: 137), fika in the sense of ‘to get bogged down’ is typical of Sephardic PA, while fika in the original sense of ‘to stay’ is labeled ‘archaic’, but not restricted to Sephardic PA. That fika ‘to stay’ (rather than keda) indeed used to be the norm in Early PA is shown by the fact that (as already hinted at in §5.6.3) the verb appears frequently in several 19th- to early 20th-century written records (e.g. Conradi 1844; Putman 1952 & 1953; Frederiks & Putman 1859; van Dissel 1865; Civilisadó 7-10-1871; La Union 19-03-1889; Jansen 1911) in contexts where modern-day PA speakers would use keda. Interestingly, Van Name (1869: 151) attested the verb fika in the speech of his informant Camps, but mistakenly analyzed it as a derivation from Sp. fijar ‘to fix’. Schuchardt (1882: 895) later gently corrected Van Name: “Van Name must have overlooked that Papiamentu contains numerous Portuguese elements, otherwise he surely would have attributed fika to Port. ficar”.357 The following correspondences with SCV are self-explanatory: (135)

(136)

(137)

a. EPA

fika boon

(Conradi 1844: 20)

b. SCV

fika bon ‘stay/remain/be well’

(Lang 2002: 339)

a. EPA

fica contentoe

(Jansen 1911: 33)

b. SCV

fika kontenti ‘stay/remain/be happy’

(Lang 2002: 189)

a. EPA

fika ketoe

(van Dissel 1865: 65)

b. SCV

fika ketu ‘stay quiet’

(Baptista 2007a: 78)

357 Original quote: “Dass das Curazoleñische zahlreiche portugiesische Elemente enthält (…), hatAddison van Name in seinen sehr werthvollen ‘Contributions to Creole Grammar’ (…) übersehen; sonst würde er wohl ficá (…) zu port. ficar (…) gestellt haben”

5.7. Auxiliary verbs

(138)

(139)

239

a. EPA

fika admirá

(Jansen 1911: 55)

b. SCV

fika dimirádu (Mendes et al. 2002: 49) ‘stay/remain/be astonished/marvelled’

a. EPA

fika para

(Jansen 1911: 16)

b. SCV

fika parádu ‘stand/hold still’

(Lang 2002: 657)

To be sure, (135–139) would be translated into present-day PA by way of keda ‘to stay, remain’. The substitution of Early PA fika by modern keda can be clearly illustrated by juxtaposing Conradi’s (1844) and van Dissel’s (1865) Gospels of Matthew and Mark with the corresponding Gospels from the 1916 PA Bible (Eybers 1916): (140) a. EPA b. PA

(141) a. EPA

b. PA

i toer doos ta fika boon and all two IMP stay well

(Conradi 1844: 20)

y tur dos ta queda bon and all two IMP stay well ‘and so both are preserved’

(Eybers 1916: 19)

fika aja te hora boso ta sali stay there until hour you IMP leave (van Dissel 1865: 20) queda ei te ora boso sali stay there until hour you leave ‘stay there until you leave’ (Eybers 1916: 83)

Around the turn of the 20th century, fika and keda still appear to have been in free variation in PA. This is particularly striking in Jansen (1911: 51, 54, 55, 66, 72, 108, 120), who produced predicates such as fica fiel and keda fiel ‘stay loyal’, fika admirá and keda admirá ‘remain marvelled’, or fika para and keda para ‘hold/stand still’, all in one text and without any (apparent) difference in meaning. Afterwards, the use of fika seems to have declined and Lenz (1928), for instance, makes no more mention of the verb. We may therefore assume that the first third of the 20th century was decisive regarding the replacement of fika by keda in mainstream PA.

240 5.7.3.

Verbal system

Other auxiliaries

5.7.3.1. PA kumisá, Upper Guinea PC kumesa In PA, ‘to begin, start’ is expressed by both a Spanish- and a Portuguese-derived variant, kuminsá (< Sp. comenzar) and kumisá (< Port. começar), which appear to be in free variation (Maurer 1988: 422). For Upper Guinea PC, Rougé (2004a: 113) provides the allomorphs komesa∼kunsa∼kumsa (SCV) and kumsa∼kunsa (GBC). However, the most commonly heard variant on Santiago seems to be kumesa, as found in e.g. Pereira (1999), Bartens (2000: 46), Veiga (1982: 119), Pratas (2007: 316), Lang (2002: 323) and Mendes et al. (2002: 132). This variant is also used in Brava CV (Meintel 1975: 220) and BaCV (Lopes da Silva 1957: 247). Moreover, Chataigner (1963: 57) attested the form kumesa in the conservative Casamance variety of GBC. Thus, we may juxtapose PA kumisá with Upper Guinea PC kumesa (cf. Martinus 1996: 175). In its role as an inchoative auxiliary, PA / Upper Guinea PC ‘to begin’ takes imperfective ta + V complements358 (cf. also [87] in §5.1.1): (142) a. PA

mediatamente el a kumisá ta tráha immediately he PFV start IMP work ‘he started working immediately’ (Lenz 1928: 298) b. SCV cˇ uba kumesa ta baza na cˇ on rain begin IMP fall on floor ‘(the) rain began to fall on the floor’ (Veiga 1982: 119) c. GBC i para, i kunsa ta sibi he stop he start IMP climb ‘he stopped [and] started to climb’ (Scantamburlo 2002: 547)

In present-day PA, the Spanish-derived form kuminsá appears to be significantly more frequent than the Portuguese-based variant kumisá. The latter is not found, for instance, in the dictionaries by van Putte & van Putte-de Windt (2005a & b), Ratzlaff (1992) or Joubert (1999), all of whom provide kuminsá. On the other hand, some Early PA sources such as Conradi (1844) and Lenz (1928) exclusively draw on the Portuguese-based variant kumisá ( in Conradi 1844). This and the fact that for instance Martinus consistently uses kumisá in his PA 358 In SCV, the short form kunsa has grammaticalized further into an auxiliary expressing what Lang (2002: 368) calls “a posterioridade do processo em relação a outro processo mencionado anteriormente” [‘the posteriority of the action in relation to another action mentioned earlier’]. A discussion of this particular use of SCV kunsa is provided by Pereira (1999). As far as I know, this meaning – “presque intraduisible en français” (Quint 2000a: 269) – is not typical of the PA / GBC cognates.

5.7. Auxiliary verbs

241

writings (e.g. Martinus 2007), suggests that the Portuguese-based variant was and continues to be part of the PA lexicon. The Gulf of Guinea PC form komesa (∼kometha in ANG) shows no raising of the etymological unstressed /o/ as in PA (kumisá) and Upper Guinea PC (kumesa∼kumsa). On the other hand, Gulf of Guinea PC komesa can take imperfective ka + V complements, similar to the PA / Upper Guinea PC ta + V complements: ST N komesa ka xina mindjan I-start-IMP-learn-medicine ‘I started learning medicine’ (Rougé 2004a: 113). 5.7.3.2. PA bai – bin, GBC bai – bin, SCV bai – ben The sound equivalence between the verbs ‘to go’ and ‘to come’ in PA and Upper Guinea PC is noteworthy: Table 40. Verbs for go and come in PA and Upper Guinea PC ETYMON Port. vai = Sp. va Sp./Port. ven/vem

359

PA

SCV

GBC

MEANING

bai bin

bai ben

bai bin

‘to go’ ‘to come’

One should recall that, as lexical verbs, PA / Upper Guinea PC bai and bin∼ben do not take prepositions when followed by a place indication (§3.2.6). Another interesting lexical parallel is found in the serial construction PA / SCV kore bai (143), with a meaning close to ‘to run off’. For the sake of comparison, I present the French translations provided by the respective authors: (143) a. PA

el a kore bai he PFV run go ‘Il est parti en courant’ (‘he ran off’) (Maurer 1988: 258) b. SCV E kore bai he run go ‘Il s’en est allé en courant’ (‘he ran off’) (Pereira 1999: 110)

359 As in most creoles, verbs in PA and Upper Guinea PC are generally derived from the lexifier’s infinitives, with the exception of several high-frequency verbs such as ‘to go’, ‘to come’ and some modal verbs (Lipski 1988: 31). It is not agreed upon, however, whether the verbs for ‘to go’ and ‘to come’ in Upper Guinea PC originally derive from the lexifier’s 3sg present or rather from the imperative forms. See e.g. Scantamburlo (1981: 39, 40) and Quint (2000a: 92) for opposite thoughts in this respect.

242

Verbal system

As auxiliaries, which is what concerns us most here, we find both bai and bin∼ben in positions where literal translations with English ‘to go’ and ‘to come’ are often – though certainly not always – possible: (144) a. PA

El a disidí di bai buska Shon Arei pa e he PFV decide to go search Shon Arei for him mes bin mira self come see ‘He decided to go look for Shon Arei, so that he could come and see for himself’ (Baart 1983: 50) b. GBC tudu dia e ta bai piska na mar all day he IMP go fish in sea ‘every day he goes fishing in the sea’ (Scantamburlo 2002: 467) c. SCV dipos, e ben kontra ku pastor afterwards he come meet with priest ‘afterwards, he came to meet the priest’ (Lang 2002: 552)

(145) a. PA

Yuchi Prince ku ta bin buska hende pa bai Yuchi Prince who IMP come search people for go kòrta kaña na Cuba cut sugar cane in Cuba ‘the little boy Prince who comes looking for people to cut sugar cane in Cuba’ (Èxtra 31-03-2006)

b. GBC Kal dia ku n pudi ba ta bim buska which day REL I can go IMP come search ropa? clothes ‘What day can I come to pick up the clothes?’ (Almeida 1991: 52) Also, the auxiliaries may appear in serializations of more than two verbs: (146) a. PA

e pais kolonisadó ta bolbe bin dikta riba the country colonising IMP return come rule over nos us ‘the colonising country will come to rule over us once again’ (Èxtra 02-09-2008)

5.7. Auxiliary verbs

243

b. SCV E ta torna ben toma konta he IMP return come take charge ‘he will end up taking charge once again’ (Baptista 2002: 114) c. GBC E linya li, n bin bay bindi dinoti this line here I come go sell tonight ‘This thread, I’ll go and sell it tonight’ (Kihm 1994: 218) In the remainder of this paragraph, I will focus solely on the parallels between PA bin and Upper Guinea PC bin∼ben as this verb discloses more idiosyncrasies than bai does. Examples (147, 148) illustrate the use of PA / Upper Guinea PC bin∼ben as an auxiliary that is best translated by means of adverbs such as ‘finally’ (Kihm 1994: 109), ‘at last’ (Martinus 1996: 189) or ‘eventually’ (van Putte & van Puttede Windt 2005a: 64), or, alternatively, by means of predicates such as ‘to come to V’, ‘to end up V-ing’ or ‘to start V-ing’: (147) a. PA

nos a bin sali nuebor y algu we PFV come leave nine o’clock and something ‘[we had agreed to leave at eight, but] we ended up leaving at nine something’ (Heuvel, van Wel & van der Wal 1989: 113) b. SCV N ben sai di vila I come leave of city ‘I ended up leaving the city’ (Baptista 2002: 114) c. GBC fulas bin kunsigi ciga Kansala Fulas come manage reach Kansala ‘the Fulas finally managed to reach Kansala’ (Kihm 1994: 109)

(148) a. PA

aña pasá, pa kasualidat mi a bin kontra un year past, for accident I PFV come meet a Indjan bieu Indian old ‘last year, by accident, I met an old Indian’ (Rosenstand 1970: 69)

244

Verbal system

b. GBC i na anda na kamiñu te i bin kontra he PROG walk on road until he come meet ku lyon with lion ‘he’s walking (on) the road until (/and then) he (suddenly) meets the lion’ (Kihm 1980: 149) Bin∼ben can also take ta + V complements, as in (149). The difference with (147, 148) is that the presence of ta in (149) causes the action expressed by the complement verb to be ongoing rather than completed. This suggests that bin∼ben is not really a “gradual completive” as Martinus (1996: 189) defines it, but rather a type of transition marker which marks the arrival at, or the starting point of, an action expressed by the complement verb (cf. Kihm 1994: 110). This action may then either be completed (147, 148, without ta) or ongoing (149, with ta). (149) a. SCV E ben ta fase kusas ki ningen ka sunha he come IMP do things REL nobody NEG dream ‘he ended up doing things that nobody thought possible’ (Pereira 1999: 113) b. PA un mucha a bin ta kanta keseyó ke tal IND child PFV come IMP sing God-knows-what ‘a child started singing God knows what’ (Rosalia 1983: 28) Admittedly, the examples provided thus far reflect universal tendencies involving the verb ‘to come’and, indeed, also the lexifiers Spanish and Portuguese have semantically comparable periphrases expressing the arrival at, or starting point of, a new action or state, as in Port. vim a saber que… ‘I came to know that…’ (Kihm 1994: 113). However, quite a few propositions involving the creole auxiliary bin∼ben in fact cannot be literally translated and would require additional adverbs or conjunctions for the translation to make sense. For instance, it is not uncommon in both PA and Upper Guinea PC for the lexical verb ‘to come’ to combine with its auxiliary counterpart as in (150). Clearly, a literal translation into Portuguese, Spanish or English would be ungrammatical here: (150) a. GBC n na bin bin toma di mi I IMP come come(AUX) take of me ‘I’ll come and take what is mine’ (Kihm 1994: 109) b. PA Bo no por a bini bin bisami? 2sg not can PFV come come(AUX) tell me? ‘Why didn’t you come and tell me?’ (Allen 2007: 183)

5.7. Auxiliary verbs

245

In (150), the auxiliary bin∼ben appears to function merely as a type of transition marker meaning ‘and then…’, viz. as an event sequencer marking the successive occurrence of two events (cf. Martinus 1996: 189). Kihm (1989: 367; cf. 1994: 110) refers to bin in this role as a posteriority marker. Martinus (1996: 189) also compared PA bin and GBC bin as auxiliaries, thereby drawing on Wilson (1962: 25) for GBC. According to Martinus (1996: 189), the phrase GBC i bay Mansoa, i bin bay Bisaw, i bin bin li ‘he went to Mansoa, then he went to Bissau, and then he came here’(Wilson 1962: 25) would translate into PA as el a bai Mansoa, despues el a bin bai Bisau, despues el a bin bin aki. The only difference with GBC, he asserts, is that “in Papiamentu an explicit time sequencing would perhaps be mandatory” (Martinus 1996: 189), which in his PA translation is achieved with PA despues ‘then’. However, as the following example shows (151), as in GBC, such a sequencing adverb is not always required in PA either: (151) a. PA

el ta bini purá bin contra bo coe he IMP come rashly come meet you with 400 homber 400 man ‘he is coming rashly (and) will meet you with 400 men’ (Jansen 1911: 31) b. GBC Bariga pega kirsi i bin fika barigona Belly take grow he come stay huge belly ‘the belly started to grow (and/until) it became a huge belly’ (Scantamburlo 2002: 66)

In such examples, according to Kihm (1994: 110), “bin denotes more successivity like its English adverbial equivalent then, or like Portuguese pois, and its privileged use is thus in narratives where sequences of events are reported, and where each successive event becomes a new time of reference for a following bin.” Indeed, this use of bin∼ben is widespread in story telling in GBC (cf. e.g. Montenegro & de Morais 1995) and it is similarly common in SCV (cf. e.g. Silva 1987) and PA (cf. e.g. Baart [1983], the stories collected by Maurer [1988: 352–366], or the interviews with basilectal PA speakers documented by Allen [2007]). The following comment by Kihm (1994: 112) quite appropriately summarizes the distribution of the auxiliary bin in GBC, which applies also to PA and SCV: bin, he notes, often “appears as a quite Kriyol way of expressing something like a prospective-successive infinitive. To put it more accurately (…), it shows the need on the part of a Kriyol speaker to encode these categories, even though they receive no expression in the Portuguese sequence he is translating”.

246

Verbal system

To account for the many subtle differences in behavior between GBC bin and its etymon, Port. vem, Kihm (1989: 367, 368) argues that conflation with the verb bi∼ben ‘come’found in several of the surrounding West-Atlantic languages was a crucial factor in the development of GBC bin as an auxiliary. The same West-Atlantic substrate might be responsible for (some of) the properties of PA bin, showing again the need to consider West-Atlantic and Mande as possible contributors to PA’s substrate. One may insist that the parallels between PA bin and Upper Guinea PC bin~ben result from the prototypical semantic properties of the verb ‘to come’ allowing for it to engage in such structures. The parallels gain relevance, however, when we notice that PA has two verbs for ‘come’: the short form bin (< Sp./Port. ven/vem) and a longer form bini (< Sp. venir = Port. vir) (cf. ex. [150b]). Interestingly, as a lexical verb, either Spanish-derived bini or the short form bin may occur: PA el a logra bin/bini he-PFV-manage-come ‘he managed to come’ (Martinus 1996: 189). In auxiliary position, however, only the short form bin is allowed. Thus, we find PA el a bin logra ‘he finally managed’, rather than *el a bini logra (Martinus 1996: 189). Though Spanish-derived bini may appear in serial verb constructions, it does so only as a main verb, not as an auxiliary, which is illustrated by the difference in meaning between PA el a bini bai ‘he came and went’ and el a bin bai ‘he finally went/ended up going’ (Martinus 1996: 189; cf. Maduro 1991: 16; Muller 1989: 344 for similar remarks). This lexical/functional distribution is exactly what the hypothesis examined in the present study predicts: the form bini is strictly lexical and clearly Spanishderived, whereas the short counterpart bin is mainly functional and shows a close formal and semantic-syntactic resemblance to Upper Guinea PC bin.

5.7.3.3. PA tin, Upper Guinea PC ten The etymon of PA tin and Upper Guinea PC ten (‘to have, exist’) is Port. tem (= Sp. tiene). Port. tem, in turn, is the 3sg present of the infinitive ter ‘to have’. Unsurprisingly, like Sp. tener and Port. ter, PA tin and Upper Guinea PC ten can combine with a relative pronoun to express ‘to must, have to’ (PA tin ku / SCV ten ki / GBC ten ki∼ku). In addition, tin and ten take past participles: PA tin papiá ku have-said-that ‘it is said that…/ the rumor has it that…’ (van Putte & van Putte-de Windt 2005a: 330); SCV N ten papiadu ‘I have spoken (Veiga 2000: 216; cf. examples of ten + past participle in Fogo CV in Meintel 1975: 220). The main functions of PA tin / Upper Guinea PC ten, though, are that of expressing possession (152) and existentiality (153):

5.7. Auxiliary verbs

247

(152) a. PA

mi tin dos yu I have two child ‘I have two children’ b. SCV N ten dos fidju I have three child ‘I have three children’

(153) a. PA

tin un hômbër ku yama Airu di Dios have a man who be called Air of God ‘there is a man who is called Air of God’ (Maurer 1988: 167) b. SCV ten un ómi ki txoma Djon Grándi have a man who be called John Big ‘there is a man who is called Big John’ (Lang 2002: 775)

It is is well known that many creoles have one lexical item for both possessive ‘to have’ and existential ‘to be’ (cf. Holm & Patrick 2007: feature 13.6) and, although European Port. tem does not express existentiality, Brazilian Port. tem also combines these two properties (Holm 1992: 60). In addition, the use of Spanish tener as an existential copula is attested in Dominican Spanish (Holm 2004: 76) as well as in the Spanish-based creole Zamboangueño (a variety of Chabacano), where Sp. tiene became tyéne, which indeed also means either ‘to have’ or existential ‘there is/are’ (Lipski & Santoro 2007: 387). The point to observe is thus not so much the polyfunctionality of PA tin / Upper Guinea PC ten, but rather the fact that, if PA had been an originally Spanish-based creole, we would have expected it, in line with Zamboangueño, to select Spanish rather than Portuguese source material for a feature as important as this one. 5.7.3.3.1. Two verbs ‘to have’: PA tin – tene, Upper Guinea PC ten – tene As just noted, many other creoles have one lexical item to express possession and existentiality. However, unlike most creoles, both PA and Upper Guinea PC have two formally related verbs in complementary semantic distribution together covering the ground of possession: PA tin/tene, SCV ten/tene and GBC ten/tene. The resemblance between PA tin/tene and GBC ten/tene was pointed out by Birmingham (1975: 21). More recently, Quint (2000b: 154, 155) provided a comparison of PA tin/tene and SCV ten/tene on which I elaborate here. For SCV, the semantic opposition between the short and the longer form is commonly defined as one between inalienable (viz. essential) and transitory (viz. occasional/accidental) possession (Quint 2000a: 244; Rougé 2004a: 270; Lang 2006: 53). Compare, for instance, (152b) provided above with SCV bu teni troku? 2sg-have-money ‘do you have the money?’(Quint 2000b: 154). For GBC,

248

Verbal system

the distribution of ten vs. tene is not homogeneously described and appears to differ from source to source. I will return to this briefly at the end. In PA, as we will see, the distribution of tin vs. tene is not exactly as in SCV either. Nonetheless, the following interesting semantic parallel surfaces: just like SCV tene, PA tene typically expresses transitory possession and may never express essential, inalienable possession. This can be seen, for instance, in the primary meaning of PA tene, which is ‘to hold (in one’s hand)’, quite similar to French tenir (Maurer 1998: 219; van Putte & van Putte-de Windt 2005a: 452; Quint 2000b: 154) and a meaning which SCV tene can also assume: (154) a. PA

un hende ta tene su man bon duro IND person IMP hold his hand good tight ‘[he suddenly feels that] somebody is holding his hand very tightly’ (Baart 1983: 44) b. SCV es ta tene -l sénpri fitxadu, marádu na they IMP have/hold him always locked tied up in korenti chain ‘they (will) always keep him [the dog] locked and tied up with a chain’ (Lang 2002: 346)

Illustrative of the distribution of tin vs. tene in PA is the expression tin/tene relashon, which can mean either ‘to be in a steady relationship’ or ‘to have a transitory (sexual) relationship (/ to have a fling)’, depending on whether tin or tene is selected: (155) PA

Sharlinda tin un relashon ku Luli Sharlinda have IND relationship with Luli ‘Sharlinda is in a relationship with Luli’ (Èxtra 13-02-2008)

(156) PA

Tim di Aresto a tene relashon seksual den Tim di Aresto PFV have relationship sexual in outo di trabou car of work ‘Tim di Aresto has had a sexual affair in (his) company car’ (Èxtra 10-03-2011)

In the three creoles, the short form (PA tin / Upper Guinea PC ten) typically occurs with nouns expressing physical and mental states:

5.7. Auxiliary verbs

(157)

a. PA tin forsa ‘(to) be/stay strong’ b. SCV ten forsa idem c. GBC ten forsa idem

(158)

a. PA

tin kurashi ‘(to) be brave’

b. SCV ten koráji idem c. GBC ten korajen idem

249

(Goilo 1974: 108) (Lang 2002: 462) (Kihm 1994: 213) (van Putte & van Puttede Windt 2005a: 253) (Quint 2000b: 155) (Scantamburlo 2002: 590)

However, tin in PA appears to cover more semantic ground than ten does in SCV. For instance, to express the transitory notions of being thirsty, cold and hungry, SCV uses tene sedi/friu/fómi ‘to be thirsty/cold/hungry’, whereas PA has tin sedu/friu/hamber (Quint 2000b: 154). Note, on the other hand, that in SCV it is not always clear whether to use ten or tene. Coelho ([1880]1967: 8), for example, provides SCV tudo alguên tenê médo di Duco all-person-have-fear-of-Duco ‘everybody is afraid of Duco’, whereas Costa & Duarte ([1886]1967: 274) give in tên mêd’êl I-have-fear-him ‘I am afraid [of] him’, with no apparent semantic difference between the two propositions. One remarkable phenomenon for which I lack an explanation and which appears to be restricted to PA is that, when a physical or mental state is expressed (as in [157–158]), tin is selected in the indicative mood, while tene is chosen in the imperative mood (a fact also observed by Birmingham 1975: 21 and Maurer 1988: 56, 72). Compare, for instance, the following enunciations in PA opposing indicative and imperative mood: (159)

a. PA b. PA

(160)

a. PA

b. PA

(161)

a. PA b. PA

lo mi tin miedu FUT-I-have-fear ‘I will be afraid’ (Lenz 1928: 291) No tene miedu NEG-have-fear ‘Don’t be afraid’ (Lenz 1928: 173) Mester tin kurashi be.necessary-have-courage ‘One must be courageous’ (Èxtra, 10-03-2011) Tene coerashi, roemannan have-courage-brother.PL ‘be brave, brothers’ (Jansen 1911: 101) Bo tin pasenshi you-have-patience ‘You are patient’ (Maurer 1988: 56) Tene pasenshi! have-patience ‘Be patient!’ (Maurer 1988: 56)

250

Verbal system

Nicolas Quint (p.c.) asserts that there exists no such correlation between imperative mood and the selection of tene in SCV, which rather takes recourse to auxiliary-less imperatives of the type ka-u médu NEG-2sg-afraid ‘don’t be afraid!’, koráxu mós = courage-friend ‘be brave, my friend!’. It is interesting to reflect on the etymology of the two verbs ‘to have’, since a good case for substrate influence can be made. While Upper Guinea PC ten in all likelihood derives from Port. tem, no Portuguese etymon can be identified for the long form tene. This fact, together with the semantic distinction encoded by the two variants, has been a cause for debate in the literature on Upper Guinea PC: “One wonders why two verbs, when neither Portuguese nor, so it seems, the African languages with which it is in contact use this semantic distinction”360 (Rougé 2004a: 270). However, both Quint (2000b) and Lang (2005, 2006) have pointed out the semantic equivalence between the SCV pair ten/tene and the corresponding Wolof pair am/ame. Like the SCV pair, the Wolof pair expresses an opposition between essential (am) and occasional (ame) possession (Lang 2006: 53). What is more, on a phonetic level, the Wolof derivation am > ame is reminiscent of, and plausibly accounts for, that of ten > tene in SCV (Quint 2000b: 55).361 For PA tin/tene, some synchronic facts seem to complicate the comparison with Upper Guinea PC/Wolof. First, unlike Upper Guinea PC tene, PA tene is obligatorily marked by ta for present (cf. §5.6.3). Secondly, PA has a third (though significantly less frequent) variant, tini ‘to have’ (van Putte & van Puttede Windt 2005a: 455), whose distribution remains to be investigated. Thirdly and finally, whereas Spanish tener ‘to have’ could, for obvious reasons, not have been the source for Upper Guinea PC tene, it cannot be discarded as the etymon of PA tene. On the other hand, it is important to recall that, unlike Sp. tener, PA tene can never express permanent possession. In fact, PA tin is exclusively assigned for that. In other words, PA tene is semantically closer to Upper Guinea PC tene than to Spanish tener, whereas the ground covered by Spanish tener is covered in PA mainly by Portuguese-derived tin. Consequently, it is not all too 360 Original quote: “On peut se demander pourquoi deux verbes alors que ni le portugais ni, semble-t-il, les langues africaines en contact n’opèrent cette distinction sémantique” 361 The modern SCV form tene is often realized as /teni/. However, Quint (2000b: 55) considers the variant /tene/ to be more conservative. Moreover, as Lang (2005: 45, 46) points out, Brito (1887) still used the written from and the original word-final /e/ is heard also when the verb receives the anterior marker or a clitic pronoun causing a stress shift from the first to the second syllable, e.g. te’neba or te’ne-l (Lang 2005: 45, 46).

5.7. Auxiliary verbs

251

far-fetched to assume that in an initial phase, PA tin and tene were distributed similarly to SCV ten-tene and Wolof am-ame. If there is indeed a historical-linguistic connection between PA tin/tene and Wolof am-ame, it is interesting to recall once more that, with very few exceptions (e.g. Intumbo 2006), Wolof and other Upper GuineanAfrican languages from the Mande and West-Atlantic branches are generally not considered as contributors to PA’s substrate. To close, I focus briefly on GBC. The available literature does not give a definite answer to how GBC ten and tene are distributed over the domain of possession. According to Rougé (2004a: 270), GBC ten and tene express the same permanent:nonpermanent distinction as in SCV. However, examples of GBC tene expressing permanent possession are rife: (162)

a. GBC kantu fidju ki bu tene ? how:many-child-REL-you-have ‘how many children do you have?’ (Doneux & Rougé 1988: 14) b. GBC tene kabesa have-head ‘(to) be paranormal’ (Scantamburlo 2002: 591) c. GBC Tene tiw bon have-uncle-good ‘It’s good to have an uncle’ (Kihm 1994: 51)

To be sure, to translate (162a-162c), both PA and SCV would take recourse to the short form tin∼ten. In fact, Noël Bernard Biagui (p.c.), a linguist and native speaker of the Casamance variety of GBC, comments that the distribution of ten/tene in his variety is quite the opposite of that found in SCV: “In the Casamance variety of GBC, ten is used to express a transitory, ephemeral situation”, whereas “tene expresses an ongoing, permanent state”. Still, however, it is noteworthy that the phonetic realization of GBC tene is /tene/ not only in Casamance but also in Bissau, since a strong tendency in Bissau is to raise the unstressed etymological /e/ to /i/ followed by vowel harmony, as in misti ‘want’ (= meste in Casamance), miti ‘to put’ (= mete in Casamance), bindi ‘to sell’ (= bende in Casamance), from Port. mester, meter, vender respectively (Rougé 1988). Accordingly, we would have expected to find */tini/ in Bissau. The fact that we don’t, added to the before-mentioned fact that no Portuguese etymon is available for tene, favors a historical connection between GBC ten/tene and the Wolof pair am/ame. If that connection exists, it is implied that after the separation of GBC from Early Upper Guinea PC, the semantic ground originally covered by each of its two verbs ‘to have’ was reshuffled.

252

Verbal system

Note, finally, that in the Barlavento varieties only the short form tem∼ten has survived (Quint 2000a: 244f.n., 2000b: 79). Also in Gulf of Guinea PC, only one verb for ‘to have’, tê, is attested (Rougé 2004a: 271). 5.7.3.3.2. Digression: The case of Palenquero As noted in Chapter 2, the famous remarks of Father Sandoval regarding the presence of ST speakers in El Palenque (Colombia) in 1627 has led scholars to speculate about possible genetic ties between PLQ and ST (cf. Schwegler 1998: 230 for a discussion). The presence in PLQ of several Portuguese-derived items has further fuelled such speculation (cf. Schwegler 1993; McWhorter 2000: 17–19). To my knowledge, however, a connection between PLQ and Upper Guinea has never been hypothesized, even though, as I will argue here, such a link makes good sense from a linguistic and historical point of view. For instance, it is interesting to observe that PLQ makes use of two forms of the verb ‘to have’ which are phonetically quite close to Upper Guinea PC: PLQ ten and tené (Schwegler & Green 2007: 290). This sound correspondence could of course be a coincidence. However, it is just one of several salient verbal features that PLQ shares with Upper Guinea PC – and not with Gulf of Guinea PC. Others include the progressive aspect marker ta (= sa∼saka in Gulf of Guinea PC), the post verbal anterior marker -ba (= preverbal taba in Gulf of Guinea PC), the auxiliary verb bai ‘go’ (= ba∼be∼we in Gulf of Guinea PC, Rougé 2004a: 180), and the copulas ta and é (= sa in Gulf of Guinea PC) (McWhorter 2000: 18; Lipski 2005: 297). In a rebuttal of PLQ’s Afrogenesis hypothesis, Bickerton (2002: 39) critically questioned: “Why, if we find Afro-Portuguese phonological and morphological features [in PLQ], do we not also find at least some traces of an amply distributed syntactic construction, nor traces of any system or subsystem grammatically similar to those of the Portuguese-based creoles?”362 (cf. §1.6.1). But the features shared by PLQ and Upper Guinea PC could of course be argued to constitute traces of a shared grammatical subsystem, in this case the verbal system. Historically, the possibility of linguistic transfer from Upper Guinea to Colombia is well founded, as the 16th-century arrival of considerable numbers of Wolof slaves to Colombia is well documented: Schwegler (1998: 224) notes that Wolof slaves predominated in Cartagena in the period from 1533 to 1580, which 362 Original quote: “ninguna de estas semejanzas liga el PLQ con los criollos de base portuguesa acertada (…). ¿Porqué, si existen rasgos (afro)portugueses fonológicos y morfológicos, no existen también por lo menos algunos rasgos de una construcción sintáctica tan ampliamente distribuida, ni tampoco rasgos de algún sistema o subsistema gramatical parecido a los de los criollos de base portuguesa acertada?”

5.7. Auxiliary verbs

253

is confirmed by Parkvall (2000: 137): “imports to Colombia were dominated by Yolofs until about 1580, whereas late 16th and early 17th century arrivals were predominantly Bantu”. The slave trade data and the linguistic features shared between PLQ and Upper Guinea PC allow for the tentative hypothesis that an unknown number of these imported Wolof slaves contributed to the formation of PLQ in the 17th century. It is important to note in this respect that, in the context of slave trade, Wolof should be taken as a geographic-ethnic rather than as a linguistic indication (Lang 2011: 68). Since Capeverdeans have traditionally been identified, and identified themselves, with Wolof people (Barros 1899: 283, 284; Lang 2006; 2009), an unknown number of the slaves listed as Wolof in Cartagena may well have stemmed from the Cape Verde Islands and, consequently, have had Upper Guinea PC as their mother tongue. It should be clear that I do not wish to claim that PLQ is a relexified offshoot of Upper Guinea PC; they differ too greatly in several domains of the grammar. I am merely assuming that, if speakers of Upper Guinea PC were present at the right time and the right place, they may well have contributed features, such as the ones mentioned above, to the emerging PLQ grammar.363 To return to the original topic, Schwegler & Green (2007: 290) assert that the PLQ forms ten∼tené are allomorphs of one verb. Given the parallel with Upper Guinea PC, however, it is not impossible that in an earlier stage PLQ speakers distinguished semantically between the two verbs. 5.7.3.4. PA / Upper Guinea PC parse The Sp./Port. verb (a)parecer ‘to appear, seem’ has provided PA parse, SCV parse and GBC parsi. Not so much their semantic scope (which largely coincides with the etymon), but the sound equivalence between them is worth focusing on. The syncope of the middle unstressed vowel of words with three or more syllables has given rise to several other lexical correspondences (cf. Quint 2000b: 128): Sp./Port. parecer ‘to appear, seem’ Sp./Port. América ‘America’

> PA parse

Upper Guinea PC parse∼parsi

> PA Merka

SCV Mérka,

363 Based on the analyses of late 17th-century court orders found in the Archivo General de Indias of Seville and in the Archivo General de la Nación of Bogotá, Gutiérrez Maté (2010; to appear b) casts interesting new light on the possible relationships between PLQ and 17th-century Colombian/Caribbean Afro-Hispanic speech.

254

Verbal system

Sp./Port. dia/día sábado ‘Saturday’ Sp./Port. levantar ‘to raise’

> PA (dia)sabra

SCV sabru

> PA lanta

GBC lanta

By contrast, in Gulf of Guinea PC we find ST palisê and PRI parêsê (< Port. parecer) (Rougé 2004a: 224; Maurer 2009a: 245). We may furthermore recall that PA / SCV parse is a strong stative verb whereas the Gulf of Guinea PC cognate is weak (§5.6.5). 5.7.4.

Final remarks on auxiliary verbs

Table 41 summarizes the PA auxiliaries discussed above and their Upper Guinea PC counterparts. Table 41. Etymological survey of PA auxiliaries and Upper Guinea PC cognates ETYMON

PA

Upper Guinea PC

MEANING

Port. pode (= Sp. puede) Port. quer (= Sp. quiere) Port. mester (= Sp. menester) Port. ficar (= Sp. quedar) Port. começar (= Sp. comenzar) Port. vai (= Sp. va) Port. tem (= Sp. tiene) Port. virar (= Sp. virar ‘to tack’) Sp./Port. estar Sp./Port. vem / ven Sp./Port. parecer

por ke∼ker meste∼mesté

podi kere meste-mestedu

‘to can, be able’ ‘to want’ ‘to need, be needed’

Early PA fika kumisá

fika kumesa

‘to stay, remain’ ‘to begin’

bai tin bira

bai ten bira

‘to go’ ‘to have, exist’ ‘to turn, become’

ta bin parse

ta ben∼bin parse

‘to be’ ‘to come’ ‘to appear, seem’

Provided that the etymologies are accurate, it is plain to see that Portuguesederived auxiliaries predominate in PA. Out of eleven high-frequency auxiliaries, seven are phonologically closer to Portuguese forms than to Spanish, one (PA bira) is semantically closer to Portuguese than to Spanish, whereas the remaining three are equally close to Portuguese and Spanish. Clearly, this distribution would not be expected if PA were an originally Spanish-based creole. More significantly, all eleven auxiliaries show a strong phonetic overlap with Upper Guinea PC, perhaps the only noticeable difference being the raising of the etymological /e/ to /i/ in PA tin (≈ Upper Guinea PC ten) and kumisá (≈ Upper Guinea PC kumesa).

5.8. Final remarks on the verbal system

255

Note that the choice not to list the Spanish-derived auxiliaries PA keda, kuminsá and kier in Table 41 is diachronically motivated. As noted in §1.5.1, as of the early 18th century, Portuguese was in all likelihood no longer used as a community language on Curaçao – if it was ever used as such in the first place. Consequently, in those cases where PA’s vocabulary offers both a Portugueseand Spanish-derived variant, such as PA fika vs. keda, kumisá vs. kuminsá and ke∼ker vs. kier, there is a good likelihood that the Portuguese-based variant is the oldest and most original. Table 41 offers other interesting clues. As noted, in most (Iberian-based) creoles, including PA and Upper Guinea PC, verbs generally derive from the respective lexifier’s infinitives, with the exception of a small subset of high frequency auxiliaries (Lipski 1988: 31). In Table 41, we find no less than five of these: PA por, tin, bai, meste∼mesté and bin. Four of these have a probable Portuguese etymology, while only one (bin) can be either Portuguese- or Spanish-derived. Moreover, all five have syntactically, semantically and phonetically equivalent forms in Upper Guinea PC. We may also isolate the three modal verbs, PA por, ke∼ker and meste∼mesté, and recall their phonetic, semantic and syntactic overlap with Upper Guinea PC, while bearing in mind the cross-linguistic stability of modals in cases of contact. Interestingly, these are exactly the three verbs singled out by Lipski (1988: 28) as verbs belonging to the core of the grammar of the Spanish-based creole Zamboangueño. Their realization in that creole – tyéne (< Sp. tiene), kyére (< Sp. quiere) and pwéde (< Sp. puede) (Lipski & Santoro 2007: 377, 384, 387) – provides ideal contrastive material showing what PA’s modal verbs may have looked like if it had been an originally Spanish-based creole.

5.8. Final remarks on the verbal system Assuming that “previous changes in any language inevitably leave their footprints behind them” (Bickerton 1981: 105, drawing on Givón 1971), the comparisons and analyses presented in §5.1–§5.5 were an effort to bring to the fore the TMA system that is likely to have characterized PA in its initial stages. This system must have been anterior-based (taba) and characterized by a zero marker indicating present before statives and perfective before nonstatives, as well as by a general, tense-neutral imperfective marker ta. The overlap with Upper Guinea PC is structural and leaves only to account for the fact that, in the domain of anterior marking, SoCV and GBC have postverbal -ba where PA and BaCV have preverbal taba(ta). Furthermore, though a marker such as ta is not at all unique among Iberian-based contact varieties as well as nonstandard varieties

256

Verbal system

of Spanish and Portuguese, its absence in Gulf of Guinea PC provides strong contrast and further supports the assumption that PA is not genetically related to that branch of Portuguese-based creoles. Additional value of this chapter lies in the argument that the PA markers lo and a are unlikely products of creolization and rather appear to be the results of unmarked processes of language change, be it borrowing or internal grammaticalization, carried out after the separation of PA from its Upper Guinea PC ancestor. Thus, I believe it is no coincidence that the two PA markers targeted by Bickerton (1981) as unexpected and untypical of creolization (lo and a) are indeed those that do not have a parallel in Upper Guinea PC, whereas all other markers (PA ta, taba and tabata) do. Though my account differs in several ways from that of Bickerton (1981), we agree that both lo and a did not result from creolization.The absence of these two markers in Upper Guinea PC therefore cannot be adduced as counterevidence to the hypothesized kinship of PA and Upper Guinea PC. It is important to stress here that, although I believe lo and a were acquired after the separation of PA from Upper Guinea PC, this of course does not eliminate the possibility that these two markers had become solidly integrated in the PA grammar already in the 18th century. In fact, that is likely to have been the case, as both markers appear in Early PA texts with a frequency more or less equal to their modern-day usage. In §5.6 and §5.7, a comparison of strong stative verbs and auxiliary verbs was offered. A structural, near-paradigmatic overlap between PA and Upper Guinea PC in these two verbal categories could be shown, with the individual constituents coinciding on the level of form, meaning and syntactic behavior. In light of the fact that these two categories form part of the fundamental part of the grammar of both creoles, the likelihood that this overlap is due to chance is rather small.

Chapter 6 Summary and interim analysis of the linguistic results

Introduction In chapters 2–5, a linguistic comparison of PA and Upper Guinea PC was offered the aim of which was to show that the two creole languages overlap in the fundamental parts of their grammar. Chapter 2 focused on a series of shared vowel, consonant and syllable changes the sum of which is indicative of a strong correlation between the two phonological systems. The chapter was organized with the secondary purpose of identifying a significant part of PA’s Portuguesederived vocabulary. It was shown that words of Portuguese etymology were integrated into PA and Upper Guinea PC in quite similar ways, which brings us closer to understanding the sources of the Portuguese elements in PA. In addition, the chapter aimed to reduce the likelihood that the phonological parallels between PA and Upper Guinea PC are due to chance. This was achieved by presenting negative evidence from the phonological systems of Gulf of Guinea PC and PLQ. Most attention, however, was paid to the comparison of PA’s and Upper Guinea PC’s morphosyntactic skeleton (chapters 3–5). Below, I will provide a summary of the most salient findings, accompanied by some additional remarks and conclusions.

6.1. Predominance of Portuguese-derived function words in PA A first observation of relevance is that the functional categories discussed above show a strong predominance of Portuguese-derived items (Table 42), with their number clearly exceeding those of probable Spanish origin. To properly assess the time-depth of these function words, we may recall that, if Portuguese was ever commonly used on Curaçao, which is doubtful in the first place, it was probably abandoned by the early 18th century (cf. §1.5.1). The items listed here thus belong to the most conservative layers of the PA grammar.

258

Summary and interim analysis of the linguistic results

Table 42. PA functional items of Portuguese etymology and their Upper Guinea PC cognates ETYMON

PA

Upper Guinea PC

MEANING

Port. nos (= Sp. nosotros) Port. corpo (= Sp. cuerpo) Port. mesmo (= Sp. mismo) Port. na < em + a (= Sp. en + la) Port. até (= Sp. hasta) Port. fora de (= Sp. fuera de) Port. quem (= Sp. quien) Port. onde (= Sp. donde) Port. -mento (= Sp. -miento) Port. logo (= Sp. luego)

nos

nos

‘we’

kurpa

kurpu (GBC)

‘body’

mes

mes

‘self’

na

na

‘in’

te

te

‘until’

for di

fora di

‘out (of), since’

ken

ken

‘who’

unda

unde

‘where’

-mentu

-mentu

-ment

lo

logu

Port. pode (= Sp. puede) Port. quer (= Sp. quiere) Port. mester (= Sp. menester) Port. ficar (= Sp. quedar) Port. começar (= Sp. comenzar) Port. virar (= Sp. virar ‘to tack’) Port. vai (= Sp. va) Port. tem (= Sp. tiene)

por

podi

future marker in PA; adverb ‘soon, later’ in Upper Guinea PC ‘to can, be able’

ke∼ker

kere

‘to want’

meste∼mesté

meste-mestedu

‘to need, be needed’

Early PA fika

fika

‘to stay, remain’

kumisá

kumesa

‘to begin’

bira

bira

‘to turn, become’

bai

bai

‘to go’

tin

ten

‘to have, there is/are’

6.2. Structural overlap between PA and Upper Guinea PC

259

6.2. Structural overlap between PA and Upper Guinea PC As noted in previous parts of this study, in the literature on PA it is recognized that, in order to gain insight into its origins, a main concern should be to identify the source of the Portuguese elements in the fundamental part of its grammar. Table 42 provides an important contribution to that quest: PA’s functional items of Portuguese origin can all be traced back to one source, Upper Guinea PC, whose cognates show a strong overlap in terms of phonetic shape as well as the associated syntactic and semantic properties. In a number of cases, these cognates have moved away from the Portuguese etymon in identical ways and also differ crucially from the cognates found in other Portuguese-based creoles. But the foregoing chapters managed to show that the overlap between PA and Upper Guinea PC is not limited to function words of probable Portuguese origin; also those of an indecisive Spanish/Portuguese etymology could in the majority of cases be shown to have an equivalent in Upper Guinea PC. Thus, while one can undoubtedly point out the occurrence of certain shared PA / Upper Guinea PC forms or features in one or the other nonstandard or overseas Iberian speech variety or Iberian-based creole, it is the sum of the correspondences between the two creoles that constitutes the ‘smoking gun’ evidence. In other words, although there is clearly truth in Rona’s (1971: 10) statement that “the Portuguese character of a few words does not prove anything”364 , it requires quite a stretch of imagination to assume that the structural correspondence of paradigmatically organized categories such as morphology, pronouns and TMA is due to chance. A case in point is provided by Bickerton, in reference to the alleged AfroPortuguese roots of PLQ: “if an Afro-Portuguese contact language had really existed in Spanish America, this language would have had its own structures and substructures. In that case, we would expect to find, in modern PLQ, at least some remnants of these structures and substructures”365 (Bickerton 2002: 37). While Bickerton’s argument may very well apply to PLQ (but see §5.7.3.3.2) it is difficult to maintain that Upper Guinea PC structures and substructures are not found in modern PA.

364 Original quote: “La portuguesidad de unas pocas palabras no prueba nada” 365 Original quote:“si verdaderamente existiera una lengua de contacto afroportuguesa en Hispanoamérica, dicha lengua hubiera tenido su propia estructura, más o menos bien hecha, con sus propias subestructuras (…). En tal caso, se esperaría hallar, en el Palenquero actual, por lo menos algunos vestigios de estas estructuras y subestructuras”

260

Summary and interim analysis of the linguistic results

In addition to structural correspondences, the comparisons also brought to the fore a number of more or less isolated but significant synchronic morphosyntactic differences between PA and Upper Guinea PC, the most noticeable of which are summarized below: – – – –

PA nan (3pl = PL) has no equivalent in Upper Guinea PC; PA a (perfective past marker) = Upper Guinea PC Ø; PA lo (future/irrealis marker) = SCV ta, al(?) / GBC ta, na; PA no = Upper Guinea PC ka

The first three features (nan, a and lo) were argued to result from processes of ‘gap filling’, which implies that there was scope in the original grammar for these items to either develop internally or to be borrowed from any of PA’s contact languages, not by replacing existing morphemes, but simply by ‘filling gaps’366 This is an important conclusion, as it implies that the presence of these three morphemes in PA versus their absence in Upper Guinea PC does not in any way constitute evidence against the hypothesized kinship between the creoles. The fourth feature listed here, PA no versus Upper Guinea PC ka, cannot be accounted for in terms of gap filling. Here, we must acknowledge that an original, highly functional and frequent morpheme was replaced by another. This replacement contradicts the working hypothesis of the present study, according to which functional morphemes are generally more stable than content morphemes and thus less likely to be replaced. For reasons of space, I will not venture into speculation about the possible motivation behind this replacement (but see Matras 2009: 208, 209 for pertinent remarks on the cross-linguistic borrowing of negators). Suffice it to recall here that the syntactic negation patterns of PA still coincide neatly with those of Upper Guinea PC, while differing crucially from PLQ and Gulf of Guinea PC as well as from the lexifiers (§3.5.3).

6.3. Negative evidence from PLQ and Gulf of Guinea PC Throughout chapters 2–5, contrastive data have been adduced mainly from PLQ and Gulf of Guinea PC. The aim of these data was to emphasize the idiosyncrasy of certain features shared between PA and Upper Guinea PC and thereby reduce the likelihood that these are due to chance or otherwise attributable to universal linguistic principals. As far as PLQ is concerned, the contrastive data presented 366 Recall that the notion of ‘gap’must not be taken to imply that a language is somehow defective. Saying that a language has a ‘gap’ merely means it lacks a category or feature present in one of its contact languages.

6.3. Negative evidence from PLQ and Gulf of Guinea PC

261

in the foregoing chapters complement and confirm the conclusions drawn by Maurer (1987), who had convincingly shown by means of a comparison of the verbal systems that PA and PLQ are unlikely to share a common origin. (On the other hand, in §5.7.3.3.2, I tentatively considered the possibility that certain verbal properties of PLQ were inherited from Upper Guinea PC, but I do not wish to make any claims in that respect.) As far as Gulf of Guinea PC is concerned, several drastic disparities between its grammar and that of PA and Upper Guinea PC have surfaced in the previous chapters. Among other things, it was pointed out that Gulf of Guinea PC differs considerably from PA / Upper Guinea PC in the domain of word order as well as in terms of the phonetic realization of Portuguese etyma. One of the most striking differences is constituted by the absence of the preverbal marker ta in Gulf of Guinea PC (which has ka and sa instead) versus the central role of this marker in the grammar of PA and Upper Guinea PC. 6.3.1.

Digression: What sets PA and Upper Guinea PC apart from Gulf of Guinea PC

I will digress here, in order to further focus specifically on the linguistic distance between PA and Gulf of Guinea PC and highlight some differences that could not be discussed within the structure of chapters 2–5. These merit elaborating on, since, as pointed out in §1.6.3, scholars concerned with the origins of PA have on more than one occasion hypothesized about possible genetic ties between PA and Gulf of Guinea PC. It is indisputable that PA and Gulf of Guinea PC share several features in the domain of the lexicon, phonology and occasionally even morphosyntax. However, this is quite unremarkable, given that, already in the second half of the 17th century, Curaçao started receiving considerable numbers of slaves from Kwaand Bantu-speaking areas who must have had a considerable linguistic impact. Kwa and Bantu are the two branches that crucially defined Gulf of Guinea PC’s substrate and it is therefore merely predictable to find several features shared by PA and Gulf of Guinea PC at all levels of the grammar. Examples of PA features shared in common with Gulf of Guinea PC and possibly resulting from shared Kwa/Bantu substrate influence include the vowel harmony discussed in §2.1.1 and the paragogic vowels on words such as PA solo ‘sun’ and santu ‘sand’ (< Sp./Port. sol, Du. zand). However, neither the vowel harmony nor the paragogue of vowels was ever applied systematically in PA. Interference from either Gulf of Guinea PC or a Kwa/Bantu substrate may furthermore have played a role in the extension of PA for di as a temporal preposition (§3.2.8.2), the regularization of preverbal taba(ta) at the cost of postverbal *-ba (§5.4.1), the extension

262

Summary and interim analysis of the linguistic results

of ta as a general copula (§5.7.2.1), but also (and this has not been discussed in the foregoing) the use of verbal serializations367 , although regular processes of independent, language-internal change can just as plausibly account for these parallels. It must be acknowledged that an (albeit relatively limited) number of parallels between PA and Gulf of Guinea PC appear to be idiosyncratic to such an extent that they suggest direct contact between speakers of Gulf of Guinea PC and of PA. This is a possibility that indeed needs to be considered, although it should be clear that this is no indication of genetic relatedness, but merely indicative that speakers of Gulf of Guinea PC may once have arrived on Curaçao and subsequently shifted to PA. One such parallel (PA nan ≈ Gulf of Guinea PC nan and variants) was discussed in §3.1.6. Two other parallels are also striking, though purely lexical: they concern the PA verb landa ‘to swim’ < Sp./Port. nadar, which is homophonous with Gulf of Guinea PC landa idem, and the noun PA arei ‘king’ (< Sp./Port. rey/rei), which has a prosthetic a- that is reminiscent of the corresponding Gulf of Guinea PC forms alê∼arê. Although these PA lexemes may very well be related to the Gulf of Guinea PC cognates, we also know that lexical items are relatively susceptible to replacement and that minimal forms of contact may in some cases be sufficient for lexical items to find their way into a recipient language (cf. Parkvall 1999: 205). It is implied that the correspondence of a few lexemes can hardly be put forward as proof of a genetic relation between a given pair of languages. Within the literature that has dealt with the correspondences between PA and Gulf of Guinea PC, Maurer (2002, 2005) and Quint (2000b: 167–183) stand out for having provided the most complete assessment of shared features. Their respective approach to the origins of PA, however, is quite different. Where Maurer speculated about possible genetic ties between PA and Gulf of Guinea PC, Quint (2000b) reached the following conclusion: The similarities between PA and Gulf of Guinea PC are essentially phonetic and lexical and above all much less frequent than the similarities between PA and SCV. There are commonalities between PA and Gulf of Guinea PC, but these are 367 Serial verbs are typical of Gulf of Guinea PC and are common in PA as well, and doubtlessly more frequent and longer in PA than in Upper Guinea PC. Quint (2000b: 178–180), however, shows that the use of serial verbs (of up to four verbs) in Upper Guinea PC is not as marginal to the language as is often suggested in the literature. Moreover, he reveals a possibly meaningful parallel between SCV and PA to which he refers as the use of ‘serial imperatives’, that is the ability to conjoin two verbs in imperative mood with a personal pronoun preceding the second verb: SCV xinta bu kumi! = PA sinta bo kome! sit-you-eat ‘sit down and eat! (Quint 2000b: 155).

6.3. Negative evidence from PLQ and Gulf of Guinea PC

263

very limited in number and the Creole of the ABC Islands should not in any way be considered a direct extension of this family of Creoles.368 (Quint 2000b: 180)

The data revealed in chapters 2–5 of the present study corroborate Quint’s assertion. An alternative way of emphasizing the closeness of PA and Upper Guinea PC, while at the same time assessing the distance between them and Gulf of Guinea PC, is by way of Ferraz (1987). In that publication, Ferraz has persuasively shown that “there was no single common source for the Portuguese Creoles of Africa” (1987: 356) by comparing Upper Guinea PC and Gulf of Guinea PC on a number of particularly criterial features. To my knowledge, his conclusion that “the P[ortuguese] Creoles of West Africa [i.e. Gulf of Guinea PC and Upper Guinea PC] fall into two distinct groups” has not been contested since and is still considered authoritative (cf. Holm’s [2009: 14, 15] discussion of Ferraz [1987]). Now, if we take the features that Ferraz put forward in support of his division and bring PA into the equation, we see that in the vast majority of cases PA lines up with Upper Guinea PC rather than with Gulf of Guinea PC. Among others, Ferraz discussed the following features: – Depalatalization of /S/ to /s/ before /a, o, u/. This feature is typical of Gulf of Guinea PC, while absent from Upper Guinea PC (Ferraz 1987: 342). As noted in §2.2.1, PA here coincides with Upper Guinea PC, although GBC must be exempted from this comparison for reasons explained in §2.2.1.2. – Word/syllable structure constraints. According to Ferraz (1987: 342), “Words in the Gulf of Guinea Creoles typically end in a vowel”, whereas in Upper Guinea PC there is “no constraint to words ending in a consonant”. As in Upper Guinea PC, there is no constraint on consonants in word-final position in PA either (cf. §2.3.4). – Unmarked personal pronoun a. This pronoun is typical of Gulf of Guinea PC, but not attested in Upper Guinea PC (Ferraz 1987: 343). Such a pronoun is not used in PA either (§3.1). – Word order. According to Ferraz (1987: 344), “In the Upper Guinea Creoles, demonstratives and possessives precede the noun (…). The Gulf of Guinea 368 Original quote: “Les coïncidences entre papiamento et les créoles du Golfe de Guinée sont essentiellement phonétiques et lexicales et surtout beaucoup moins fréquentes (…) que les coïncidences observées entre papiamento et badiais. (…) Il y a des points communs entre le papiamento et les créoles du Golfe de Guinée, mais ces points communs sont très limités en nombre (…) et le créole des îles ABC ne saurait en aucun cas être considéré comme un prolongement direct de cette famille de créoles.”

264

Summary and interim analysis of the linguistic results

Creoles have the reverse word order, demonstratives and possessives following the noun”. Clearly, PA again agrees with Upper Guinea PC rather than with Gulf of Guinea PC (cf. §3.1.5, §3.2.2, §3.2.7, §3.4.2.1, §3.5.1.2 and §3.5.2). – Diminutives. Ferraz (1987: 344) asserts that “diminutives do not occur in the Gulf of Guinea Creoles. Diminutives are, however, productively preserved in the Upper Guinea Creoles”. Just as in Upper Guinea PC, the formation of diminutives is a productive and quite typical feature in PA (e.g. Kramer 2004: 142, 143; cf. footnote 234). – Particle of Obligation: In all members of Gulf of Guinea PC, the particle sela (and variants) is used to express obligation. Such a particle is not attested in Upper Guinea PC (Ferraz 1987: 344, 345), nor in PA. Instead, to express obligation, PA and SCV share the particular use of the modal verb meste ‘to need, must’ (cf. §4.3.5.2 and §5.7.1.3). These distinctive features add to the body of negative evidence presented in the previous chapters, the sum of which shows that there is a rather limited linguistic basis for sustaining a genetic relationship between PA and Gulf of Guinea PC. To summarize, although some more or less idiosyncratic correspondences between PA and Gulf of Guinea PC can be observed, these should not be analyzed as remnants of a shared ancestor creole but rather point towards the fact that speakers of Kwa and Bantu languages contributed to the shaping of both these creoles. In the case of PA, this Kwa/Bantu contribution must be analyzed as postformative. It must furthermore be acknowledged that a small number of isolated features shared between PA and Gulf of Guinea PC may well have resulted from direct contact between speakers of the two creoles, though again, the linguistic evidence put forward in the preceding chapters suggests that this contact took place some time after the actual formation and stabilization of PA.

6.4. Old Portuguese features in PA and Upper Guinea PC Throughout chapters 2–5, a number of features shared by PA and Upper Guinea PC have been identified as being traceable back to Old (15th-/16th-century) Portuguese. These features include: – PA djadumingu/SCV diâ dumingu = Old Port. dia domingo (§2.1.1.2) – Preservation in PA and Upper Guinea PC of the Old Portuguese phonemic contrast between /tS/ and /S/ (§2.2.2); – The modal verb PA / Upper Guinea PC meste < Old Port. mester (§4.3.5.2 and §5.7.1.3);

6.5. The value of historical PA and Upper Guinea PC texts

265

– The preposition PA / Upper Guinea PC banda di ‘next to, besides, around’ < Old Port. banda de (§3.2.8.1). Clearly, the presence of these Old Portuguese features in the core of PA’s grammar is difficult to reconcile with any hypothesis that situates the birth of PA on Curaçao in the second half of the 17th century. Since all the Old Portuguese features that we have been able to identify in PA’s grammar are also found in Upper Guinea PC, it is plausible to assume that PA inherited the features from Upper Guinea PC. The role of Old Portuguese in the formation of Upper Guinea PC is of course quite well recognized: scholars generally agree that creolization on Santiago started in the late 15th century and, consequently, that the lexifier of proto-Upper Guinea PC was Old (15th- to 16th-century) Portuguese (cf. Quint 2000a: 55 and Lang 2001 for SCV; Kihm 1994: 4, 5 for GBC).

6.5. The value of historical PA and Upper Guinea PC texts In the linguistic comparisons carried out in chapters 2–5, I have repeatedly drawn on Early PA writings comprising grammatical sketches, letters and evangelical texts and covering a period from 1775 (the PA love letter) to 1928 (Lenz’s grammar). The purpose of this was to scrutinize the idea that, if PA and Upper Guinea PC share ancestry, the creoles must have resembled each other more closely in that early period than they do nowadays. Indeed, the data retrieved from the available historical texts has thoroughly substantiated that premise. Below is a summary of features found in Early PA as well as in Upper Guinea PC but no longer attested in present-day PA, arranged according to the chapter in which the feature was discussed: – Chapter 2: • Vowel harmony in Early PA trubunal, krusufiká = SCV trubunal, krusufiká; • Monophthongization in Early PA prinsipi = SCV prinsipi; • Paroxytonic stress on longer verbs in Early PA and SCV; – Chapter 3: • Early PA desdi∼des di∼desdi di = Upper Guinea PC desdi∼des di∼desdi di (preposition ‘since’) • Early PA pabia = Upper Guinea PC pabia (preposition and conjunction ‘because (of)’); • Early PA modi = Upper Guinea PC modi (conjunction ‘how, like, as’); • Early PA ki = Upper Guinea PC ki (relative pronoun); • Early PA es = Upper Guinea PC es (deictic marker); • Early PA pa es = Upper Guinea PC pa es (adverb ‘therefore’)

266

Summary and interim analysis of the linguistic results

– Chapter 4: • Early PA -dadi = Upper Guinea PC -dadi (deverbal nominalizer); • Use of auxiliary-less passives in Early PA and Upper Guinea PC; – Chapter 5: • Early PA taba = BaCV taba (preverbal past/anterior tense marker) • Early PA fika = Upper Guinea PC fika (copula/auxiliary verb ‘to stay, remain’) Vice-versa, we showed how the study of Early Upper Guinea PC texts may reveal features found in modern PA but no longer in modern Upper Guinea PC, such as the Early SCV form promeru (conjunction ‘before’ = PA promé), where modern-day SCV has prumeru∼purmeru. Clearly, future research into the linguistic ties between PA and Upper Guinea PC will benefit greatly from the recovery and analysis of additional historical texts and grammars. In this respect, we may notice that the oldest known Early PA text, the 1775 Sephardic love letter, consisted of several pages; only one fragment (half a page) of the original letter has hitherto been uncovered and analyzed, while the other pages remain hidden in the archives of the Mikve Israel synagogue in Willemstad (cf. Maduro 1971; Salomon 1982; Kramer 2004: 218f.n.).

6.6. West-Atlantic and Mande features in PA and Upper Guinea PC In chapters 2–5, at least three features shared by PA and Upper Guinea PC have been discussed and said to qualify as possible Upper Guinean (viz. WestAtlantic/Mande) substrate features: – The lack/rejection of voiced fricatives in the conservative part of PA’s vocabulary (§2.2.3); – Several properties of the auxiliary verb PA bin / Upper Guinea PC bin∼ben (§5.7.3.2); – Two verbs ‘to have’ (§4.7.3.3.1). I should stress that it is not the aim of this study to prove that the three features listed above are indeed related to a West-Atlantic/Mande substrate; doing so would require a significant amount of additional negative evidence from other (West) African languages as well as from non-standard varieties of the lexifiers. Despite this, the features suffice to point out the necessity of considering West-Atlantic and Mande (recognized as the two main contributors to Upper Guinea PC’s substrate) as possible contributors to the substrate of PA in future

6.6. West-Atlantic and Mande features in PA and Upper Guinea PC

267

studies. As noted previously, and barring occasional exceptions such as Intumbo (2006) and Bartens (1996: 246), West-Atlantic and Mande have hitherto not been considered in the literature on PA. This negligence appears to be directly related to the historical assumption that the slaves brought to Curaçao in the 17th and 18th centuries were drawn almost exclusively from Kwa- and Bantu-speaking areas. In the next chapter, I will address this consensus on the Curaçaoan slave trade in detail and show that it needs revision.

Chapter 7 The historical ties between Upper Guinea and Curaçao

Introduction In a critical discussion of alleged substrate influence on Saramaccan, Bickerton (1994: 65) once commented that “one must show that the right speakers were in the right place at the right time”. In line with e.g. Smith (1999: 252–254) and Parkvall (2006: 329, 330), however, I believe that in cases such as the present one the linguistic data must provide the conclusive proof of whether the right speakers were in the right place at the right time. Nevertheless, whenever possible, it is obviously desirable to support the linguistic evidence by showing that the right speakers were at least going to the right place at the right time. Here, this means demonstrating that slaves were transported from Upper Guinea to Curaçao in the latter half of the 17th century, which is the aim of this chapter. To my knowledge, most of the historical data to be discussed in this chapter have never been associated with the 17th-century settlement of Curaçao before. Their potential value to a better understanding of the origins of PA is hinted at by Lipski & Schwegler (1993: 420): “The knotty issue of Papiamento genesis cannot be resolved satisfactorily without additional study of language-external factors.” While a solid historical framework is in principle beneficial to any given claim of genetic relatedness between languages, it is particularly vital to the present study, in which the hypothesis is defended that PA is genetically related to Upper Guinea PC. The historical ties between Upper Guinea and Curaçao have not received any serious attention in Dutch colonial historiography (or, by extension, in the literature on PA), which has nourished the tacit assumption that such historical ties were simply non-existent, or otherwise insignificant. Owing to this historical assumption, the linguistic ties between the two regions have largely been ignored as well. After all, if there was no historical contact, then how could there have been any linguistic transfer? The structure of this chapter is as follows: first, I will illustrate the consensus according to which the importance of Upper Guinea (particularly Senegambia369 , see Map 4 below) to the history of the Dutch West India Company 369 Following Kihm (1994: 4f.n.), the term Senegambia is used to refer to the region in Upper Guinea that encompasses “the so-called Petite Côte south of Dakar, the

270

The historical ties between Upper Guinea and Curaçao

(and by extension to the history of Curaçao and the emergence of PA) was virtually negligible. Subsequently, in §7.2, it is shown that the Dutch were in fact a dominant force in Senegambia until well into the latter half of the 17th century, largely owing to their possession of Gorée, a strategic harbor situated near the Cape Verde Islands. In §7.3, little-known data on the slave trade between Gorée and Curaçao is presented and analyzed. I then focus on the little-known but pronounced presence of Sephardic Jewish communities in 17th-century Senegambia, their contacts with the local Dutch, and their trade and social networks with other Jewish communities in Curaçao and Amsterdam in §7.4. To close this historical chapter, §7.5 digresses by discussing the likelihood that the Upper Guinea area controlled by the Dutch in the 17th century was simultaneously populated by former inhabitants of Santiago and, thus, speakers of Early CV.

7.1. On the presumed insignificance of Upper Guinea to the history of Curaçao Although we factually know rather little about the origins of the first generation of slaves who arrived on Curaçao, the Upper Guinea region has quite systematically been overlooked as a possible source in works concerned with either the history of Curaçao or the emergence of PA, or both. This negligence, I assume, correlates strongly with the historiographic consensus that Upper Guinea played no role of significance in the colonial scheme of the Dutch West India Company (WIC). In this section, I will reflect on how this consensus could have come about and on how it manifests itself in the study of (the origins of) PA. Following the 1st WIC’s bankruptcy in 1674, most of its archives and records (including most pre-1674 records pertaining to the Chamber of Amsterdam, which was in charge of both Curaçao and the Petite Côte; see Map 3) were either lost or destroyed (see APPENDIX II). As a consequence, little is known about the pre-1674 Dutch period in West Africa. Typically, the majority of publications on, and assessments of the Dutch transatlantic slave trade deal with the post-1674 period corresponding to the 2nd WIC. In the scholarship that does address the activities of the 1st WIC, relatively little attention has been given to the Dutch presence in the Petite Côte region (Maps 3 & 4). Discussions of the strongholds in Elmina (Ghana) and Luanda Gambia, Casamance, and Guinea-Bissau”. The Petite Côte, in turn, refers to the part of the Senegalese coast that stretches from Dakar to the Gambia river. This area is occasionally also referred to as the Cape Verde region. Map 4 illustrates the regions in question.

7.1. On the presumed insignificance of Upper Guinea to the history of Curaçao

271

(Angola) as well as the Dutch colonies on the other side of the Atlantic (Curaçao) including the adventures in Brazil (of which valuable archives do in fact survive) traditionally predominate. Gorée is mentioned (if at all) only in passing and references to the Petite Côte ports of Rufisque, Portudal and Joal are usually absent. Illustrating this tendency, van Welie (2008: 58) notes that “victories at Arguin (1633), Elmina (1637) and São Paulo de Luanda and São Tomé (1641) guaranteed an unprecedented Dutch dominance on the African coast”. And Silva & Eltis (2008: 98) insist that “[t]he Dutch obtained a foothold in West Africa only with the taking of Elmina in 1637” (cf. e.g. Grant 2008a: 80f.n.), apparently trivializing the fact that Gorée was occupied by the Dutch already in 1621, facilitating Dutch participation in the Upper Guinean and transatlantic trade (to which more in the remainder of this chapter). Postma (1990: 57), moreover, in his authoritative assessment of the slave trade of the 1st and 2nd WIC, claimed that “the Senegambia region held little significance for the Dutch slave trade”. Postma’s assessment received wide acceptance in subsequent literature. For instance, drawing on Postma (1990), Rawley & Behrendt (2005: 72) explain: “The Dutch in the second half of the seventeenth century looked to the Slave Coast (…) for Africans to export to the New World. WIC slavers in the last quarter of the century loaded their human cargoes mainly at Offra and Whydah, and transported them in large proportions to the island of Curaçao and Surinam (…)”. And Parkvall’s (2000: 137) assessment of the number of Senegambian slaves imported into Curaçao as “virtually negligible” was also based on the figures provided by Postma (1990). However, Parkvall (2000: 137) adds an important caveat, namely that data for pre-1674 arrivals are hardly available (as a result of the afore-mentioned loss of 1st WIC archives and records following its bankruptcy in 1674; see APPENDIX II) As a consequence, Parkvall (2000: 137) notes, “we will have to make do with Postma’s (1990: 112) estimated figures for the early period [1660–1674]” (emphasis mine). These estimated figures, in turn, are based on the assumption that there were no systematic deviances in the Dutch slave trade before and after 1674 (Parkvall 2000: 137). Thus, due to the fact that after 1674 the Dutch slaving activity in Senegambia was indeed virtually negligible, Postma estimated a similarly negligible percentage of Senegambian slave exports for the pre-1674 period. The historical consensus described above (i.e. the presumed insignificance of Upper Guinea to the history of the Dutch WIC and thus to the history of Curaçao), I assume, is crucial in understanding why the linguistic ties between PA and Upper Guinea PC have hardly been investigated. After all, if the Dutch were never active in Upper Guinea, a significant contribution from that area to the linguistic landscape of Dutch-controlled Curaçao would not be expected.

272

The historical ties between Upper Guinea and Curaçao

How strong the correlation between this historical consensus and the study of PA actually is, can be illustrated by the fact that linguists concerned with the origins of PA traditionally do not consider the Mande and West-Atlantic languages spoken in Upper Guinea (the main substrate languages of Upper Guinea PC) as possible contributors to PA’s substrate. Maurer (1991: 126) provides an illustration: “Primarily Bantu dialects from the Congo/Angola region and Kwa dialects from the region stretching from Ghana to Nigeria should be considered as Papiamentu’s base languages” (cf. comparable assertions in e.g. Barme 2003: 237, Munteanu 1991: 22, Portilla2008a: 165). Note that Maurer’s affirmation was based largely on Postma’s (1974) assessment of the Dutch slave trade. In a similar vein, Lefebvre & Therrien (2007b: 230) assert that, “[b]esides Fongbe, there were other Gbe languages, other Kwa languages and possibly some Bantu languages involved in the formation of Papiamentu”. The fact that Lefebvre & Therrien mention no source or reference for their assessment shows that we are indeed dealing with a consensus. In sum, the historiographic tendency to ignore the 1st WIC’s presence and activities in 17th-century Senegambia or to portray these as insignificant has had a direct impact on the field of PA studies, which at least in part accounts for the fact that the historical-linguistic ties between Curaçao and Upper Guinea have been ignored in recent scholarship. Below, I will argue that this consensus needs reassessment by showing the importance of Upper Guinea to the history of the 1st WIC and to the rapid development of Curaçao in the second half of the 17th century. At the same time, however, it is important to acknowledge the correctness of Postma’s figures for the period after 1674, i.e. the data on the origins of the slaves imported into Curaçao by the 2nd WIC: it is very likely that, for reasons discussed below, the 2nd WIC came to purchase slaves almost exclusively in Lower Guinea and Bantu-speaking areas (see figures in Postma 1990: 112; van Welie 2008: 55). Acknowledging this adds to a better understanding of the chronology of the language transfer from Upper Guinea to Curaçao: given that Curaçao’s development as a slave depot started in the early 1650s, this language transfer must have occurred in the course of the period stretching roughly from 1650 to 1675. As the next sections will demonstrate, this was not only the peak period of Curaçao’s slave imports (§7.3), but also the period in which the Dutch were in control of significant parts of the Senegambia region (§7.2).

7.2. The Dutch presence in Senegambia in the 17th century

273

7.2. The Dutch presence in Senegambia in the 17th century Interestingly, the tendency to marginalize the relevance of Senegambia to the history of the WIC is noticeable mostly in scholarship concerned with assessing the Dutch WIC (slave) trade as a whole (but see Ribeiro da Silva 2009 & forthcoming370 ). Scholars whose primary concern is not with the WIC but rather with (aspects of) the history of Upper Guinea, on the other hand, in fact speak frankly about the pronounced Dutch presence and activities in 17th-century Senegambia, particularly the Petite Côte371 In this section, using 17th-century accounts as well as recent scholarship, I focus mainly on the Dutch possession of Gorée (from 1621 to 1677) and the surrounding factories, which guaranteed their participation in and control over the Senegambian trade in a period of some five decades, justifying Lemos Coelho’s (1669[Peres 1953]: 6) metaphorical description of Amsterdam as “the city with the Upper Guinea stock-market”372 . While the Dutch dominance in Senegambia very much depended on the seizure of Gorée in 1621, their interest and involvement in the regional trade clearly precedes this event. The late 16th-century presence of Dutch vessels along the Senegambian coast is noticed in Thilmans’ (1968: 17) discussion of a 1598 report by the Dutch trader E. van Reyd: “The navigations in Guinea had become so common that the coasts of this country were not without Dutch ships from summer to winter”373 . In that report, van Reyd orders the Dutch skipper W. Lodewycx to anchor at the harbors of Joal and Portudal, where he would be welcomed by Dutch factors (Thilmans 1968: 17), clearly suggesting that in 1598 the Dutch already possessed a number of factories on the Petite Côte (Thilmans 1968: 18). In this light, it is not surprising to learn that in 1605 another Dutch trader, Peter van den Broeck,

370 Research conducted by Ribeiro da Silva (2009; forthcoming) into the Dutch notarial archives of the Gemeente Archief in Amsterdam is now providing new detailed information about Dutch commercial activities on the Petite Côte for the period from ca. 1600 to 1674. 371 See, for instance, the works of Curtin (1975), Boulègue (1989), Brooks (1993, 2003), Havik (2004), Mark & Horta (2004, 2008, 2011), Green (2005, 2007, 2008, 2011) and particularly Moraes (1969, 1972, 1973, 1993, 1995, 1998a, b, c, d). 372 Original quote: “a cidade (…) donde tem a bolsa da costa de Guiné” 373 Original quote: “les navigations en Guinée étaient devenues si communes que les côtes de ce pays n’étaient dépourvues ni d’été ni d’hiver de vaisseaux néerlandais. Vingt à vingt-cinq à la fois jetant l’ancre près de la côte”

274

The historical ties between Upper Guinea and Curaçao

Map 3. Settlements and strongholds discussed in this chapter. In gray: strongholds with a pronounced Dutch commercial activity in the second half of the 17th century.

Map 4. Senegambia in the 17th century. In squares: strongholds controlled by the Dutch WIC

7.2. The Dutch presence in Senegambia in the 17th century

275

received orders to call at Portudal, the mainland village where most trade was conducted. There he rented a house, if it deserves that name, he said, a type of straw hut. He also rented a Portuguese slave to cook and serve as interpreter. According to him, most of the Portuguese living in the Cape Verde region are true thieves. He finds most of them in Portudal and Joal, where they trade with the English and the Dutch. (Silla 1969: 79) 374

Before proceeding, it is useful to note that the 1st WIC was subdivided into five Chambers, each of which were granted the right to trade in a given area375 . The Chamber of Amsterdam is of particular interest to this study. This Chamber has under its direction Cape Verde and the neighboring places, where the principal stronghold was situated on a little island off the Cape Verde Peninsula named Gorée, and the less important mainland strongholds of Portudal, Rufisque and Joal. Furthermore, the Chamber of Amsterdam was in command of New Netherlands and Cayenne, but she now only has Curaçao, Aruba and Bonaire. (Anonymous account, drawn up in 1670 or 1671, printed in Moraes 1998a: 245–247)376

In other words, the Chamber of Amsterdam was in charge of the ABC-islands on the one end and of the Dutch strongholds along the Petite Côte, Upper Guinea, on the other, a fact of obvious importance to understanding how the language transfer from Upper Guinea to Curaçao could come about. Furthermore, knowing that Curaçao was under control of the Chamber of Amsterdam allows us to narrow down the historical analysis of the Dutch in Upper Guinea, excluding from the discussion the Dutch stronghold in Arguin and the lesser known factory in Portendieck, both north of Senegambia on the Mauritanian coast. Since these strongholds fell under the jurisdiction of the Chamber of Zealand, I will assume for now that their role in the language transfer from Upper Guinea to Curaçao was reduced. 374 Original quote: “reçut l’ordre de se rendre à Portudal, ville du continent où se fait le principal commerce. Il y loua une maison, si l’on peut donner ce nom, dit-il, à des cabanes de paille. Il y loua aussi une esclave portugaise pour l’office de la cuisine et pour lui servir d’interprète (…). Selon lui la plupart des Portugais qui résident aux environs du cap Vert sont de véritables brigands. Il s’en trouve plusieurs à Portudale et à Joale, où ils exercent le commerce avec les Anglais et les Hollandais”. 375 For more information on the structure of the Dutch WIC and its division into Chambers, see, for instance, den Heijer (1994) and Ribeiro da Silva (2009 & forthcoming). 376 Original quote: “a sous sa direction Cabo Verde et les places circonvoisines plus éloignées dont le comptoir principal est situé sur une petite île située au Cabo Verde, nommé Goereede, et les comptoirs moins importants de la terre ferme du Cap, Pourto Dalié, Refisco, Sjouale (…). Pour la reste, la Chambre d’Amsterdam a eu (…) sous sa direction Neuw Nederlandt et Cajana, mais elle n’a plus maintenant que les îles Curaçao, Aruba et Buonaire”

276

The historical ties between Upper Guinea and Curaçao

The Dutch possessions along the Gambia River were not supervised by the Chamber of Amsterdam either but rather by the Chamber of Stadt ende Lande/Groningen (Moraes 1998a: 252). The Dutch presence in the Gambia River area warrants a brief discussion, however, since this river with Casão as its commercial terminus offered invaluable access into the interior where all sorts of merchandise (including slaves) were traded. An interesting reference to the Dutch presence along the Gambia River was made by a certain Francisco Pirez de Carvalho in a 1635 account: “Along the Gambia River, where the most important trade is conducted, there must be some twenty Portuguese among Flemish377 and creoles who are mulattos”378 (cited in Moraes 1995: 318). Still in 1660, the Dutch must have engaged in commerce along the Gambia River. In that year, the French reportedly raided Saint Andrew’s Island, situated in the mouth of the river, thereby defeating a Dutch battalion stationed there (Moraes 1995: 402). Not without a sense of irony, the French offered to sell Saint Andrew’s Island back to the Dutch. These, however, rejected the offer, after which the English bought the island in 1661 renaming it James Island (Moraes 1995: 402). A more detailed discussion of the Dutch presence and activity along the Gambia River cannot be offered in this section, however, due to a lack of data. The remainder of this section focuses on the Dutch presence in the Petite Côte region, starting with the island of Gorée (§7.2.1), then moving to the neighboring factories in Rufisque, Portudal and Joal (§7.2.2). Section 7.2.3 describes how the Dutch were ousted from Gorée and the other strongholds in 1677 and how this event heralded the permanent Dutch withdrawal from the Senegambia region. The section closes with an assessment of the Dutch ties with Cacheu (in modernday Guinea-Bissau) and the Cape Verde Islands (§7.2.4), two strongholds of obvious importance to the present study, given that Upper Guinea PC was, and still is, spoken there.

377 The terms ‘Flemish’ and ‘Flanders’ are frequently used in early accounts to refer to the Dutch and the Seven Provinces. While the shared history of Flanders and the Seven Provinces is complex, the two were apparently still perceived of as a political and commercial unity in the times of these early writings. 378 Original quote: “Do rio de Gambia adonde ha o maior Comercio havera couza de vinte Portuguezes os mais delles framengos [e] Crioilos que sao mulatos”

7.2. The Dutch presence in Senegambia in the 17th century

277

7.2.1. The Dutch in Gorée References to the island of Gorée (< Du. Goede Reede ‘Good Port’)379 are absent in several standard works on the Dutch (Atlantic) slave trade as well as in writings concerned with the historical relations between Curaçao and West Africa. Below, I will show that its importance to the history of the 1st WIC in Senegambia and, ultimately, to the early history of Curaçao and the emergence of PA cannot be gainsaid. Gorée lies just below the tip of the Cape Verde Peninsula at a “cannon shot distance” from the mainland (thus Donelha 1625[Mota & Hair 1977]: 127) and at the same latitude as the Cape Verde Islands (see Maps 3 & 4). The island was officially claimed by the Dutch in 1621380 . It offered an ideal port where vessels sailing under different flags would stop to refresh, refill food supplies, and/or buy slaves (see §7.3). The possession of Gorée brought strategic advantages to the 1st WIC boosting the Company’s ambitions in Senegambia. According to Brooks (2003: 139), “Dutch traders conveniently dispatched vessels to the Petite Côte and Gambia River from their base on Gorée Island”. Furthermore, “from Gorée, one could regulate all the traffic coming from Europe”381 (Moraes 1998d: 54) and to make a good situation even better, the island “was at the beginning of the shortest route to the West Indies” (M’bow 1985: 11). Thus, connecting the Petite Côte with Europe and the West Indies as well as with the Senegambian interior, 17thcentury Gorée can legitimately be referred to as “the meeting point of major shopping routes” (M’bow 1985: 11). By the mid-17th century, due to a variety of adversities (cf. §7.2.4.2), Santiago had lost its central position in the West African and transatlantic slave trade 379 Brooks (2003: 130) explains that the toponym Gorée is “a contraction of Goede Reede, the Dutch word for good roadstead, like the island of Goeree at the mouth of the Schelde River”. In documents dealing with the pre-Dutch period the island is variously referred to as Palma or Palm Island (cf. Mota & Hair 1977: 278f.n.). Among the Wolof, the island was initially known as Beer (M’bow 1985: 11). By the end of the 16th and for some time into the 17th century, the island went by the name Berzeguiche (orthography varies), after the African chief of the Cape Verde region at that time (Moraes 1969: 989; Thilmans 2006: 65f.n.). Confusion may arise, since in early accounts the toponym Berzeguiche is occasionally also used to denote the harbor of modern-day Dakar (Silla 1969: 78, 79; Moraes 1995: 314f.n.). 380 The years 1617, 1627 and 1629 are occasionally mentioned as the year of the Dutch occupation of Gorée, but the year 1621 is mentioned most frequently, which I therefore adopt in the present study. 381 Original quote: “À partir de Gorée, on pouvait (…) contrôler tout le trafic qui venait de l’Europe”

278

The historical ties between Upper Guinea and Curaçao

(e.g. Green 2007: 286, 287). The island’s decline went hand in hand with the demise of Portuguese control over the Senegambia region, which, in turn, was closely related – and inversely proportional – to the increasing Dutch dominance in that area. Mark (2002: 13) foregrounds the importance of Gorée in this shift of power: By the early 1600s (…), the joint Spanish-Portuguese monarchy’s financial difficulties, combined with the rise of Dutch commerce, had begun to undermine Portuguese supremacy. This process was abetted by the Dutch conquest of northeastern Brazil beginning in 1630 and by the establishment of a Dutch trading post on Gorée Island off the Senegalese coast in 1621. (Mark 2002: 13)

Or, as Green (2007: 241) puts it, “[g]rowing Dutch competition with Portugal meant that, with the definitive loss of Gorée to the Dutch in 1629, [Portuguese] influence over the Senegambia region from Cabo Verde effectively ceased”. There is a valuable account from 1684 by the Cape Verdean merchant Lemos Coelho (1684[Peres 1953]: 96, 97). Contemplating his time as an Upper Guinea trader, he introduces the island of ‘Bersiginche’, “which, in my time, was called Gur˜e by its Dutch masters”382 , and provides insight into the profit the Dutch took from the island: Either ships from Cacheu called here, or he [the Dutch Captain of Gorée] would send there [to Cacheu] his merchants to trade, and from there [Gorée] they drew (…) wax and ivory each year; and each year two to three big ships would come here from Holland to load the mentioned merchandise which they brought to Amsterdam (…) and they took much profit from this island, so much that when the English took it in my time, in 1663, they did not hesitate to strike back furiously, so that later in the following year General De Ruyter was sent with a squadron of 14 warships to retake the island; so here we see the benefits obtained from the trade they conducted here.383

382 Original quote:“a qual os olandezes, que erão em meu tempo senhores della, chamãvao a ilha de Gur˜e” 383 Original quote: “aqui vinhão os navios de Cacheo, ou elle mandava la os seus a comerciar, e de lá tirava todos os annos novecentos e mil quintais de cera e marfim; e aqui vinhão todos os annos de Olanda, duas e tres naos grandes a carregar dos ditos generos que lavavão para a cidade de Amsterdam donde tinhão o asento da Companhia de Africa que asim lhe chamavão; e tiravão tanto interesse desta ilheta que tomando-lha o ingles em meu tempo, no anno de 1663, não repararão em andarem as guerras muy acesas entre essas duas naço˜es, para que logo no anno seguinte não mandassem o seu general Rut com uma esquadra de quatorze náos de guerra a restauralla; assim que por aqui se verá os lucros que tiravão dos negocios que aqui farião.”

7.2. The Dutch presence in Senegambia in the 17th century

279

Gorée soon became the target of nations competing with the Dutch in Upper Guinea and realizing the island’s potential. Having resisted a series of Portuguese, English and French attacks, the Dutch finally lost Gorée to the French in 1677 (see §7.2.3). 7.2.2. The Dutch on the Petite Côte (Rufisque, Portudal and Joal) In spite of the pronounced Dutch presence in Rufisque, Portudal and Joal in the 17th century, references to these three Petite Côte ports are typically absent in scholarship on the Dutch WIC. The Dutch control over the main three Petite Côte ports of Rufisque, Portudal and Joal has been documented in detail by the above-mentioned trader Lemos Coelho (1669[Peres 1953]: 7–10). According to his (still authoritative) historical account, the Dutch (to which he alternately refers as Flemish) had factories in all three ports and regulated the trade with the local ‘Portuguese’ (many of whom were in fact Cape Verdeans from Santiago; see §7.5) and Luso-Africans (mulattos, lançados and tangomaos), as well as with the local Africans. Moreover, Sephardic Jews and New Christians (converted Jews) with ties to Amsterdam were active in these ports (see also §7.4).384 In fact, Gorée’s value to the WIC trade would have been minimal if the Dutch had not also secured their presence in and control over these Petite Côte ports: “Gorée does not serve the Dutch except as a general warehouse, and all the trade is done by their factors stationed in Rufisque, Portudal and Joal, as well as by their vessels sailing along the Coast continuously to register the merchandise traded by these factors”385 (in Moraes 1998a: 278, from an anonymous French report dated 1672). In the 1630s, a letter reached the Portuguese Crown with the warning message that in the Cape Verde region “there must be four to five sites where the Dutch have a fortress with a military camp and much commerce with much merchandise to optimalize their commerce”386 (in Moraes 1995: 313) and a 1635 report similarly spoke of “the Dutch, hoarders of riches, who are growing

384 Of these three ports, particularly Portudal seems to have flourished commercially (Lemos Coelho 1669[Peres 1953]: 10). 385 Original quote: “Gorée ne sert aux Hollandois que de magasin général, et tout le trafique se fait par leurs Commis qu’ils entretiennent à Rufisque, à Portudale et à Jouale, et par les barques qu’ils envoyent incessamment le long de la Côte, pour en rapporter les marchandises, que leurs Commis y ont traité” 386 Original quote: “avera quatro ou sinco legoas adonde o olandez tem hua fortaleza com prezidio de soldados com muita fazenda com que fazem seu comercio”

280

The historical ties between Upper Guinea and Curaçao

in number”387 (in Moraes 1995: 311). The Portuguese concerns grew not only with respect to the military and commercial threat posed by the Dutch, but also with respect to the possible spread of Protestantism in the Petit Côte region. The Portuguese therefore considered establishing a mission in that area “in order to prevent the heretic Dutch, who have built factories there, from corrupting the Christians with their heresies”388 (Moraes 1995: 290, citing from an anonymous 1636 account). At its peak, the Dutch trade in the Petite Côte area was a point of reference against which the trading intensity and product quality of other commercial entities was measured. In Portudal, in 1670, for instance, “a Cape Verdean mestizo installed in this port did more commerce than the Dutch here”389 (from a French report cited in Boulègue 1989: 64), and Lemos Coelho (1669[Peres 1953]: 8) found that the Mandingas “make a butter of a quality that beats the butter from Flanders”390 . Moreover, if Guèye’s (1998: 27) observation that “from 1621 to 1677, the Dutch masters of Gorée allowed the other nations to trade alongside them in Rufisque, Portudal and Joal”391 (emphasis mine) is correct, then this suggests that during some five decades the Dutch were basically in charge of the Petite Côte. Throughout the 1660s, the French had become increasingly interested in the Petite Côte and carefully registered Dutch trading patterns. The following passage was drawn up in 1670 by the Vice-Admiral of the French holdings in the Petite Côte, Jean d’Estrées (who some years later would be responsible for the seizure of Gorée; cf. next paragraph). It illustrates the balance of power in the Petite Côte region at that time and reveals how the chain of Petite Côte strongholds allowed the Dutch to put their mark on the local trade: Although all foreigners are welcome to trade along the Coast up until the entrance of the Gambia River, the Dutch do not let them pass openly and do not allow any fraud or deceit, which is not too difficult for them thanks to their establishment on the island of Gorée off Cape Verde and to the factors they maintain in Rufisque, Portudal and Joal. They transport their merchandise in vessels which they load

387 Original quote: “les Néerlandais, accapareurs des richesses qui viennent d’être énumérées” 388 Original quote:“afin que les Hollandais hérétiques qui y ont fixé des établissements” 389 Original quote: “un marchand métis capverdien installé dans ce port y faisait (…) plus de commerce que les Hollandais” 390 Original quote: “fazem mui boa manteiga, que a de Flandres lhe não ganha” 391 Original quote: “De 1621 a 1677, les Hollandais, maitres de Gorée, laissaient les autres nations trafiquer a Côte d’eux a Rufisque, Portudal, Joal”

7.2. The Dutch presence in Senegambia in the 17th century

281

by means of their customs and their friends, on whom they leave the imprint of the great riches of the States of Holland.392 (In Moraes 1998a: 238–239)

If we recall that the commercial development of Santiago stagnated in the course of the 17th century (e.g. Boxer 1963: 14), the Dutch in the Petite Côte seemed to be prepared to take over and reanimate significant parts of the Upper Guinean colonial trade. 7.2.3. The loss of Gorée and the Dutch retreat from Senegambia It is indicative of the importance of Gorée in facilitating the WIC’s grip on the Petite Côte trade that throughout the 17th century the Dutch suffered and resisted a series of Portuguese, English and French attacks on the island. The first of these attacks was mounted by the Portuguese in the late 1620s. A Portuguese official first informed his superiors of the possible benefits of taking up arms against the Dutch in the Petite Côte area: This is the land from which Your Majesty should expel the Dutch. And when Your Majesty takes this land and its commerce from them, you will take all the profits that the States of Flanders have from it, because the truth is that everything from the Castle of Arguin to the Castle of São Jorge de Mina belongs to the Dutch. Your Majesty could deny this commerce to the Dutch at very little cost and with much profit.393 (In Moraes 1995: 318)

The advice was taken at heart and in 1629 preparations were made to dislodge the Dutch from Gorée (Moraes 1969: 997). But besides inflicting material damage, the resulting Portuguese expedition against Gorée had little effect, witness the report of a Cape Verdean priest (cited in Moraes 1969: 1004) still in 1629: 392 Original quote: “Quoyque tous les estrangers soient bien receus pour traiter le long de la Coste jusque à l’entrée de la rivière de Gambie, et que les Hollandois ne les traversent pas ouvertement, ils ne laissent pas de le faire soubs mains, ce qui ne leur est pas difficile par l’establissement qu’ils ont à l’isle de Gorée au Cap Vert qui soutient les facteurs qu’ils ont à Rufisque, à Portudalle et à Joualle. Ils (…) traverser celles [leurs marchandises] des vaisseaux qui y abordent par le moyen de leurs habitudes et de leurs amis, outre qu’ils impriment une grande opinion de la puissance des Estats d’Hollande. (…) Les Hollandois ont un facteur à Rufisque (…), un autre à Portugalle et un à Joualle comme aux autres lieux.” 393 Original quote: “Estas são as terras que V. Magestade avera de vedar e proibir ao olandez E quando V. Magestade lhe tire estas terras e commercio lhe tira todo o Rendimento que tem os estados de frandes porque e Verdade he que do Castello de argi Como asima digo ate o Castello de Sao Jorge de mina tudo he dos Olandezes. V. Magestade pode fazer E Vedar este comercio aos Olandezes com muito pouco custo e sera muito grande proveito”

282

The historical ties between Upper Guinea and Curaçao

“[T]he Dutch have resumed the building of a fortress at Berzeguiche [Gorée] and it is reasonable to fear that they will become masters of the whole of Guinea and of this island (S. Tiago)”. The fear of a Dutch raid on Portuguese holdings proved to be ungrounded, however, and from 1630 onwards, according to Moraes (1998d: 59), “[a] certain modus vivendi between the Portuguese and the Dutch seems to have been found, since from that year (1630) onwards the Portuguese started trading again along the Petite Côte”. This apparent truce lasted for over a decade: Moraes (1969: 1008) makes mention of a second Portuguese raid on Gorée in 1645. The raid was this time supported by French mariners under the command of Etienne Duquesne, but must have been as ineffective as the first, judging from the fact that no reference to it was made in WIC reports post-dating this event (Moraes 1969: 1009). The same Duquesne made a renewed effort in 1659 (Benoist & Camara 2003: 18), this time in service of the Swedes394 . In the 1660s, the English were compelled to take up arms against the Dutch in order to get access to the Upper Guinea slave market and satisfy their increasing need for slaves in the West Indies (Rawley & Behrendt 2005: 77). The subsequent territorial struggles in the Petite Côte included the English taking of Gorée in 1663/1664 under the command of Robert Holmes and its recapture by Michiel de Ruyter’s armada only shortly after.395 Although they failed to dislodge the Dutch from the Petite Côte, the English by then were in control of the Gambia and continued to launch attacks against the Dutch from their fort on James Island (Rawley & Behrendt 2005: 78). The sum of these attacks had financially and militarily weakened the Dutch WIC. Rumors about the Company’s precarious situation spread rapidly so that in 1670 a French Petite Côte official believed that “the Dutch are so weakened that I do not believe that by the end of this month anybody will speak of them again”396 (cited in Moraes 1998a: 276). D’Estrées, the vice-admiral of the French holdings in the Petite Côte region, was now making plans to more permanently expel the Dutch from the Petite Côte area. But a 1670 memoir shows that the French were initially hesitant to use military force against the Dutch: “Her Majesty is not convinced that 394 The alliance between Duquesne and the Swedes against the Dutch must be seen in the light of the Swedish-Danish-Dutch rivalry that had developed on the African coast mostly as a result of the disputed possession of Cape Corso on the coast of Ghana (Moraes 1995: 398). 395 On the other side of the Atlantic, also in 1664 and as part of the same Anglo-Dutch conflict, New Netherlands was permanently ceded to the English in exchange for Surinam (e.g. Rawley & Behrendt 2005: 78). 396 Original quote: “Les Hollandois sont si bas que je ne crois pas qu’avant la fin de ce mois on parle plus d’eux dans l’Europe”

7.2. The Dutch presence in Senegambia in the 17th century

283

the proposal (…) to let them [the ships of a French fleet] depart and to send them to the Cape Verde Islands to undertake action against the forts owned by the Dutch in this land, can be executed”397 (Moraes 1998a: 275). Instead of attacking the Dutch strongholds, therefore, d’Estrées initially considered a more indirect approach by teaming up with the captain of the Cape Verde Islands, who proposed building a fortress between Portudal and Rufisque for the purpose of undermining Dutch trading patterns. D’Estrées reported: The governor of the Cape Verde Islands obviously does not like the Dutch, whose commerce fills him with aversion and jealousy, so that, in order to be protected from them, to take their riches and make them retreat, he proposed to me the building of a fort between Portudal and Rufisque. He then explained to me that such an establishment would produce great advantages and that it would render useless the Dutch warehouse on Gorée.398 (D’Estrées, 1670, cited in Moraes 1998a: 235, 236)

This plan appears never to have been executed, however, and in 1671, the French declared war on the Dutch on the Petite Côte (Moraes 1998a: 268). Since the Dutch WIC hadn’t been able to shake off the English either, they soon faced hostilities from two sides: “Acting in concert with the French, England waged another Dutch war (1672–1674), which diluted the Dutch menace to English trade and precipitated the bankruptcy of the WIC” (Rawley & Behrendt 2005: 79). France exploited the circumstances by ousting the Dutch from Gorée in 1677 and the surrounding Petite Côte strongholds before the end of the decade. This time the Dutch lacked the resources to try and reclaim their former Senegambian strongholds and were compelled to permanently withdraw from the region: “Until 1679 the WIC regularly sent ships to the isles of Arguin399 and Cabo Verde. After the conquest of its forts and factories in the Senegambia 397 Original quote: “Sa Majesté n’est pas persuadée que la proposition (…) de les [the ships of a French fleet] faire partir et de les envoyer aux Isles du Cap Verd pour faire entreprise sur les forts que les Hollandois ont en ce pays puisse estre exécutée” 398 Original quote: “Le gouverneur-général des isles du Cap-Vert (…) n’ayme pas naturellement les Hollandois, outre l’aversion que luy donne la jalousie de leur commerce, de sorte que, dans la veue de se fortifier contre eux et d’assurer ses richesses et une retraitte, il me fit la proposition de bastir un fort entre Portudal et Rufisque, sur les confins des deux estats (…). Il faisoit considérer ensuitte que cet establissement produiroit de grands avantages, que par là on rendoit inutile le magazin qui tiennent les Hollandois au fort de Gorée.” 399 The Dutch managed to retain Arguin on the Mauritanian coast up until the early 18th century (Barry 1988: 86). Note, however, that this port served the gum trade rather than the slave trade.

284

The historical ties between Upper Guinea and Curaçao

region by the French (…), trade and shipping to this territory was limited” (den Heijer 2003: 149). Similarly, according to Curtin (1975: 103), the taking of Gorée by the French “ended strong Dutch participation in Senegambian trade”. If the conquest of Gorée by the Dutch in 1921 had marked the beginning of a successful WIC campaign in Senegambia, it was now its loss to France in 1677 that marked the end of that campaign. The run up to Gorée’s changing hands and the far-reaching consequences for the region are described from a French perspective by Silla: The French looked at the Dutch outposts on the coasts of Senegal with great distrust. This had to create hostilities between them. A. Villard wrote in this respect: ‘On the 1st of November 1677, the Vice-Admiral d’Estrées hunted down the Dutch, whose squadrons were crushed near Gorée. The island was occupied. The French commerce obtained the right to trade in the three Petite Côte factories, Rufisque, Joal and Portudal.’ At that very moment the history of the French on the Cape Verde peninsula begins.400 (Silla 1969: 80; see also Lang 2006: 58)

It was indicative of the impact of the Dutch episode in Senegambia that the regional trade collapsed after their withdrawal in the 1670s. Whereas the WIC’s commercial policy in Senegambia had been based on tolerance towards other commercial entities401 , the French Company of Senegal opted for a strict monopoly, which, according to Boulègue (1989: 97), “provoked a considerable weak-

400 Original quote:“Les Français voyaient d’un mauvais oeil l’implantation hollandaise sur les côtes de Sénégal. Ce qui devait créer des hostilités entre eux. A. Villard écrit à ce sujet: ‘Le 1er novembre 1677 le vice-amiral d’Éstrées (…) pourchassa les Hollandais et leurs escadres furent dispersées près de Gorée. L’île fut occupée. Le commerce français avait le droit de trafiquer dans les trois factoreries de la Petite Côte, Rufisque, Joal et Portudal.’ C’est à ce moment que commence pour la France l’histoire de la presqu’île du Cap-Vert. Des rapports étaient inévitables entre l’île et le continent voisin.” 401 Moraes (1972: 112), for instance, describes how, in the period of Dutch control over Gorée and the Petite Côte, ‘foreign’ ships would trade freely along the Petite Côte. Furthermore, references to the healthy (commercial) relationships between the Dutch and local Africans are easy to find in contemporary reports. In one such report, it is observed that ‘[l]es Nègres de cette région portaient d’ailleurs grande amitié aux Néerlandais” [‘the negroes of this region maintained close friendship with the Dutch’] (Moraes & Thilmans 1973: 102). Rawley & Behrendt (2005: 72, 73) summarize: “The Dutch were skilful diplomats and traders in dealing with African coastal authorities”.

7.2. The Dutch presence in Senegambia in the 17th century

285

ening of the commerce on the Petite Côte. Rufisque, Portudal and Joal declined; the other trading points became inactive”402 . In the period after 1677, Gorée would become one of France’s main slave trading stations along the African West Coast (Parkvall 1999: 194).403 7.2.4. The Dutch ties with Cacheu and the Cape Verde Islands As noted in the introductory chapter of this study, it is likely that creolization in Santiago (viz. the birth of proto-Upper Guinea PC) had been completed by the mid-16th century and possibly earlier.404 Several historical documents discussed in the works of Carreira (e.g. 1983a, b) and others furthermore suggest that by the end of the 16th century, most of the original inhabitants of Santiago spoke the creole as their native language (cf. Mark 2002: 14). Having been founded and settled mainly by former Cape Verdean merchants in the late 16th century (Brooks 1993a: 237–244; Boulègue 2006: 49; Green 2007: 41, 257, 277f.n.), Cacheu is thought to have been the first settlement on the mainland where Upper Guinea PC (viz. Early GBC) was widely used as a vehicle of communication (do Couto 1994: 31, 32; Ladhams 2006: 90). If in the 17th century, Upper Guinea PC was indeed the native tongue of segments of the population of Santiago and Cacheu, the relevance to the present study of discussing the commercial ties between the Dutch WIC and these two ports is evident. Note that, unlike the Petite Côte, the Dutch never controlled Cacheu or the CapeVerde Islands. Nonetheless, as the discussion provided below is meant to illustrate, the WIC was actively involved in the Cacheu (slave) trade and also had access to the Santiago slave market. 7.2.4.1. Cacheu With Portuguese commerce shifting from the Cape Verde Islands to the continent in the late 16th and 17th centuries, the continental port of Cacheu (see Maps 3 & 4) grew rapidly to become the slave trading capital of Upper Guinea (Green 2007: 183; see also Castillo Mathieu 1982: 26). While Portugal offi402 Original quote:“provoqua un affaiblissement considérable du commerce de la Petite Côte. Rufisque, Portudal et Joal déclinèrent; les autres points de traite perdirent toute activité”. 403 Interestingly, post-1677 Gorée plays an important role in Parkvall’s (1999) afrogenesis hypothesis for the French Antillean creoles. For details on the French period in Senegambia following their conquest of Gorée, one can consult, for instance, Faure (1914). 404 Cf. the compelling historical and linguistic evidence for this claim provided by Lang (2006).

286

The historical ties between Upper Guinea and Curaçao

cially controlled Caheu, the port in reality hosted practically all nations and/or commercial entities involved in the slave trade at the time. The literature on Cacheu suggests that the Dutch were never as dominant a force there as in other parts of Senegambia. The WIC does seem to have made plans to gain more control over Cacheu, or at least that is what the Portuguese were preparing for: in a 1644 letter to the captain of Cacheu Gonçalo de Gamboa de Aiala, the Portuguese King himself, João VI, asks about the status of “the fortifications (…) I ordered to be made in order to resist a Dutch armada that was said to be coming from Holland towards these parts” (in Esteves 1988: 76). A Dutch strike on Cacheu, however, has never been registered to my knowledge. While the influence of the Dutch in Cacheu was thus in no way comparable to their authority in the Petite Côte, they did own a trading post there (Moraes 1995: 259, 264, 1998c: 55, den Heijer 2003: 141, Boulègue 1989: 27). In fact, reports of Dutch trading activity in Cacheu go back to as early as 1611 or 1612 when a reference is made to the arrival there of “a ship from Flanders loaded with merchandise”405 (Moraes 1972: 39) and the commercial ties between the Dutch settlements in the Petite Côte and Cacheu appear to have thrived throughout the 17th century, as testified to on more than one occasion by Lemos Coelho. In addition to his quote provided in §7.2.1, Lemos Coelho (1684 [Peres 1953]: 98) noted that, “in my times”, the island of Gorée offered “a beautiful harbor for the ships coming with blacks from Cacheu to this island, because here they refreshed their weaponry, took a fresh bath and bought on the island as many goods as they needed and as many presents as they wanted, being well received by the Flemish”. It is not far-fetched to assume that the WIC used its contacts with and trading post in Cacheu primarily for the purchase of slaves, for the simple reason that “there was nothing else to be bought in Cacheu” (from a contemporary account, cited in Green 2007: 245). That the Dutch had access to the Cacheu slave market is confirmed in the late 1660s, when, because of a scarcity of slaves in the Petite Côte, Dutch privateers calling at a WIC factory on the Cape Verde Peninsula in search of slaves were ordered to sail to Cacheu and buy them there (Moraes 1998b: 323). Moreover, for the period between 1650 and 1675, the Voyages database of transatlantic slave trade (www.slavevoyages.org) lists three ships sailing under the Dutch flag as having purchased their slaves in Cacheu as well as one Dutch ship (the Vredenburg) having taken on slaves in Cacheu, Gorée and the Gambia. Further research into the Dutch WIC’s influence in and ties with Cacheu must take the so-called Company of Cacheu into consideration (cf. Martinus 1996: 405 Original quote: “un navire des Flandres (…) chargé de marchandises de traite”

7.2. The Dutch presence in Senegambia in the 17th century

287

143–145). According to van Dantzig (1968, cited in Martinus 1996: 145), “The Portuguese licensed Company of Cacheu, founded in 1692 mainly for slave trade, had intimate links with the Portuguese Jews of Amsterdam and had its own office in Curaçao”. Note that under different names the same company had already been active in Cacheu from 1664 onwards (Martinus 1996: 144) and was designed to ratify already existing (slave) trading networks (Pereira 1993: 202). For obvious reasons (such as the widespread use of Upper Guinea PC in Cacheu), this direct connection may shed important light on the historical and linguistic ties between Curaçao and Upper Guinea. The Company of Cacheu’s example also shows how difficult it is to discuss the (slave) trade between Upper Guinea and Curaçao without addressing the involvement of the Sephardic Jewish communities on both sides of the Atlantic. Klooster, for instance, introduces the Sephardic Jew Phelipe Henriquez, born in Amsterdam, who came to Curaçao in the late 17th century to become “one of Curaçao’s most successful Jews” (Klooster 1998: 66). Interestingly, “Henriquez was involved in transatlantic slave trade for the Portuguese Company of Cacheu” (Klooster 1998: 66), became that Company’s director in 1692 and was furthermore authorized by the WIC to purchase slaves in West Africa and transport them to Curaçao (Böhm 1992: 190). I will return to the Sephardic Jewish networks between Upper Guinea, Curaçao and Amsterdam in §7.4. 7.2.4.2. The Cape Verde Islands Unlike in the Petite Côte and Cacheu, there is no historical evidence that the Dutch were ever physically established on the Cape Verde Islands. Rather, the earlier cited 1670 account from d’Estrées (“The governor of the Cape Verde Islands obviously does not like the Dutch”) suggests that the Dutch were unwelcome on Santiago and were seen as a commercial and military threat (cf. Mota & Hair 1977: 367). Any commercial exchanges between the Dutch WIC and the Cape Verde Islands are therefore likely to have been contraband. The circumstances were certainly favorable for contraband trade between the Dutch and Cape Verde Islanders to occur. For instance, throughout the 17th century, Portuguese ships coming from Cacheu and destined for the Cape Verde Islands, or vice-versa, would regularly replenish and restock at Gorée (Moraes 1997: 185). What we furthermore know is that in the first half of the 17th century and possibly beyond, the Dutch frequently visited the Islands of Sal, Maio and São Vicente in order to obtain salt, even though this practice did not require the Dutch to go ashore or to deal with Cape Verdean Islanders directly (Mota & Hair

288

The historical ties between Upper Guinea and Curaçao

1977: 367)406 . In fact, Dutch ships (by that time exceeding the Portuguese ones in size and fire power) could freely call at the unfortified (and/or uninhabited) parts of the Cape Verde Islands and did so not only to collect salt, but also to stock up before crossing the Atlantic. For instance, “In January 1624 (…), a fleet of twenty-six Dutch ships watered at S. Vicente, before sailing across the Atlantic to capture Bahia in Brazil in May of that year” (Mota & Hair 1977: 367). While the Dutch habit of calling at the Cape Verde Islands was clearly not appreciated by the Cape Verdean/Portuguese authorities, the undeniable fact that they frequently did suggests that there may very well have been other illicit commercial contacts between the Dutch and native Cape Verdeans involved in the contraband trade. Arguably more interesting still is Moraes’ (1995: 318f.n.) discussion of an account dated 1641 from a certain Jorge de Castilho “in which he reports that a Dutch ship had previously called at Santiago, ‘to load Negroes for Pernambuco’”407 . Though this is the only reference I encountered to the Dutch purchasing slaves on Santiago, it shows that it occasionally occurred. It gains further relevance in light of the fact that this occurrence is not confirmed in any official record of the slave trade to Pernambuco in the 17th century: “there is no evidence of any slaves from Cape Verde or, later, Cacheu and the southern rivers of Senegambia entering Pernambuco” (Silva & Eltis 2008: 118, 119). Thus, the Dutch apparently purchased slaves on Santiago without this being officially documented. Neither one (the Dutch buying slaves on Santiago) nor the other (this practice being contraband) is really surprising, however, considering the powerful position of the WIC in Senegambia at the time and the resulting animosity of the Portuguese authorities towards the Dutch, forcing them to trade clandestinely with Santiago. 406 Mota & Hair (1977: 375) cite from an anonymous Portuguese/Cape Verdean account from shortly after 1626, in which detailed light is shed on the Dutch practice of obtaining salt from the Cape Verde Islands: “As for the salt on Maio Island, they [the Dutch] load seventy or eighty ships each year with it – and who can stop them? Only a royal fleet or a fortress there, or the salt-flats being worked out, would stop them from carrying away the salt. But although so many ships and people come to collect the salt, not a single person dares to sleep on shore because they know the island is inhabited. Only on São Vincente, because it is unoccupied and has no people, do they sleep on shore and go from their ships. They do this nowhere else. It can be seen that since they do not do these things elsewhere, for the reason given, it would be wrong to evacuate the population of these islands, for this would make the enemy master of them, and then it would be necessary to compensate the owners.” 407 Original quote: “dans lequel il signale qu’un vaisseau hollandais s’était précédemment rendu à S. Tiago ‘qui chargea des nègres pour Pernambuco’”

7.3. Dutch slave trade from Upper Guinea to Curaçao

7.2.5.

289

Final remarks on the Dutch presence in Senegambia in the 17th century

This section provided a reassessment of Dutch presence in 17th-century Upper Guinea, particularly in the Senegambian Petite Côte. The presence of the Dutch in that area and that period must be seen as a historical precondition for the transfer of speakers of Upper Guinea PC to the Dutch-controlled Island of Curaçao to have occurred in the second half of the 17th century. I presented a scenario in which the taking of Gorée in 1621 began a fivedecade period of Dutch control over the Senegambian trade, including commercial ties with Cacheu, and gave the WIC full access to the region where Upper Guinea PC was and is spoken natively (cf. §7.5). The intensity of the Dutch episode in Senegambia obliges us to reconsider Silva & Eltis’ (2008: 98) claim that “[t]he Dutch obtained a foothold in West Africa only with the taking of Elmina in 1637”. Below, I will show that this episode also included slave trade from Gorée to Curaçao, a historical fact that has never before been related to the early history of Curaçao let alone to the origins of PA.

7.3. Dutch slave trade from Upper Guinea to Curaçao According to Martinus (1996: 5), “[t]he populating of Curaçao seems really to have started around 1660”. If we take the loss of Gorée in 1677 as marking the Dutch WIC’s withdrawal from Senegambia, this leaves a window of some seventeen years in which the language transfer from Upper Guinea to Curaçao could have come about. (This is assuming, of course, that slaves purchased by the Dutch in Upper Guinea were the principal transmitters.) How crucial this 17-year-window (1660–1677) may have been to the linguistic history of Curaçao is hinted at by Buddingh (1994: 35): “The years of most importance to the slave trade were the years from 1667 to 1674 when an average of 3,000 slaves per year arrived at Curaçao”408 . Allen (2007: 65) confirms these data: “From 1667 until 1675 (…) around 24,000 African enslaved people were shipped to Curaçao”. She tellingly calls this period the ‘peak period’ of Curaçao’s 17th-century slave trade. An immediate cause for this peak was the fact that in 1662 the WIC was allotted the asiento409 for supplying the Spanish colonies with slaves (Postma 1990: 111; Green 2007: 285). 408 Original quote: “De belangrijkste jaren voor de slavenhandel waren de jaren 1667 tot en met 1674 toen er gemiddeld 3.000 slaven per jaar op Curaçao aankwamen” 409 The so-called asientos were contracts or trading agreements that would give their owner the exclusive right to trade with Spain in a specific region.

290

The historical ties between Upper Guinea and Curaçao

As noted, mainly due to the bankruptcy of the 1st WIC in 1674, it is presumed that we lack data about the initial period of Curaçaoan slave trade. Indeed, much if not most of the proper 1st WIC documentation on the pre-1674 period is no longer available (see APPENDIX II). This makes Moraes’ research (1993, 1995, 1998a, 1998b, 1998c) all the more valuable: she has recovered manuscripts and documents from various Dutch municipal archives dealing with the activities of the 1st Dutch WIC along the Upper Guinea Coast all through the 17th century. The data she presents on the slave trade principally concern the 1660s and 70s, i.e. two decades which happen to coincide with Curaçao’s ‘peak period’ defined above. Among her most important findings are these: – In 1659, the Dutch ship Gideon brought 28 slaves from CapeVerde to Curaçao (Moraes 1998b: 42). The same ship was again reported as transporting slaves from the Cape Verde region410 to Curaçao in 1671 (Moraes 1998a: 326). – Moraes (1998b: 42) introduces Matthias Beck, the governor of Curaçao between 1657 and 1668, who, in a letter to the directors of the WIC in 1659, expresses his hopes “to bring from Cape Verde ‘a good lot of Negroes”411 . – Moraes (1998b: 51) describes the activities of the WIC vessel Casteel van Curaçao (what’s in a name?), which in 1674 transported around 200 slaves from Gorée to Curaçao and made a second, third and fourth return trip GoréeCuraçao, before returning to Amsterdam in 1676. – Among other Dutch WIC vessels setting sail from Gorée and/or the Cape Verde region to Curaçao transporting ‘mixed cargoes’ (i.e., slaves + other merchandise) are the Gerechtigheyt in 1672–1673, the Morgenstar in 1673– 1674 and the Elisabeth in 1674 (Moraes 1998a: 328–330). The few archival documents analyzed by Moraes are probably minor compared to the amount of documents lost or destroyed following the 1st WIC’s bankruptcy (see APPENDIX II). How incomplete the available data are can be illustrated by comparing Moraes’data with those of the online Voyages slave trade database412 . For instance, the Elisabeth, discussed by Moraes as transporting slaves from Gorée to Curaçao in 1674, is not registered as such by the Voyages. Reversely, the Voyages lists an unnamed slave vessel owned by a certain Jacob Foreest as 410 According to the Voyages database (www.slavevoyages.org), the principal place of slave purchase of the Gideon in 1659 was Portudal. 411 Original quote: “d’amener du Cap-Vert ‘un beau lot de Negres’” 412 The data collected by the Voyages database (www.slavevoyages.org) on the Dutch slave trade between Upper Guinea and Curaçao appear by using the following variables: Flag: Netherlands; Principal place of slave purchase: Senegambia & offshore Atlantic; Principle place of slave landing: Caribbean.

7.3. Dutch slave trade from Upper Guinea to Curaçao

291

having delivered slaves from Gorée to Curaçao in 1670, which is not discussed by Moraes. Also, where the records analyzed by Moraes mention no less than four return trips for the slaver Casteel van Curaçao, only one return trip is listed in theVoyages. On the other hand, the number of slaves theVoyages lists as having disembarked in Curaçao (394) exceeds Moraes’ number of 200. Similarly, the number of slaves brought to Curaçao by the Gideon in 1659 amount to 28 according to Moraes, whereas the Voyages database estimates that no less than 300 slaves disembarked on Curaçao. In contrast to its precursor, the 2nd WIC did properly preserve its records, which start in 1674 and end in 1740. The records are available in the Nederlands Historisch Data Archief in Leiden. The patterns that emerge from a basic analysis of these records confirm that, as argued in §7.2, the loss of Gorée to the French was indeed a terminal point in the history of the Dutch in Upper Guinea, with the number of 2nd WIC ships departing from the Senegambia/Cape Verde region at its peak prior to, and rapidly declining after the loss of Gorée in 1677. Between 1674 and 1677, a total of twelve WIC ships are marked as ‘CV’, meaning that they departed from the Cape Verde region. Four of these carried slaves, the remaining eight left with unspecified cargo. Of the four registered slavers, three had Curaçao as their destination. Then, from 1678 to 1681 (i.e. following the loss of Gorée), the number of WIC ships leaving the Cape Verde region is almost halved (seven). One, charged with slaves, headed for Curaçao. Eventually, from 1682 onwards, not more than one WIC ship (with unspecified cargo and destination) is registered as departing from the Cape Verde region. 7.3.1.

Other factors relevant to the Dutch slave trade from Upper Guinea to Curaçao

7.3.1.1. Portuguese rebellion against Spain and the demise of Cartagena The development of Curaçao as a principal slave trading port from the 1650s onwards (to which Castillo Mathieu [1982: 98] refers as “the Dutch eruption”) was boosted by the commercial paralysis of Cartagena in that same period (Castillo Mathieu 1982: 93). Until then, Cartagena had been the principal slave entrepôt in the area and, relevant in the present context, had been a major buyer and distributor of Senegambian slaves, particularly from Cacheu (Castillo Mathieu 1982: 44; Green 2007: 183).413 The demise of Spanish-controlled Cartagena was directly related to the Portuguese rebellion against the Spanish Crown in 1640. Until then, Portugal had been the major supplier of slaves to Cartagena, but after that event, Spanish trade 413 Another important reference on the rise and fall of Cartagena is Böttcher (1995).

292

The historical ties between Upper Guinea and Curaçao

with Portugal came to a halt (Castillo Mathieu 1982: 92). Green (2007: 244f.n.) cites from an anonymous account from the early 1640s illustrating the sudden lack of incoming slaves in Cartagena: “[F]rom the kingdoms of Guiné and other parts large numbers of blacks used to come to the city of Cartagena, but with the uprising of the kingdom of Portugal [in 1640] this trade has ceased with the result that the said city and its province are in great want of the said blacks.” By that time, Curaçao was organized to fill the commercial gap created by the downfall of Cartagena. The paralysis of trade between Portugal and Spain after 1640 was as much of a blow to the Portuguese as it was to the Spanish. The Portuguese traders had now lost their most significant sales market: the Spanish Americas in general and Cartagena in particular. As a consequence, the Upper Guinean slave export market, which had previously relied on Portuguese buyers, came to rely more heavily on the Dutch WIC (cf. Green 2007: 287, 274), whose colonial expansion and relatively healthy contacts with Spain in the New World correlated with an increasing need for slaves. The new reliance of the Upper Guinea slave market on Dutch buyers is nicely illustrated by Green (2007: 287, 274), when he remarks that the WIC’s loss of its settlements in Angola and São Tome in 1648 to the Portuguese was a blessing to Cacheu’s slave export economy, since this meant that the Dutch would become more active in Cacheu. 7.3.1.2. The 1662 asiento The WIC’s competitiveness vis-à-vis the Portuguese in Upper Guinea as well as that of Curaçao vis-à-vis Cartagena was further boosted in 1662, when the WIC received the much-desired asiento for providing the Spanish Americas with slaves. It is important to note that this asiento was obtained by the Chamber of Amsterdam (Rawley & Behrendt 2005: 75), which (as noted in §7.2) was in control of the Dutch Antilles on the one side of the Atlantic, and the Dutch strongholds in Upper Guinea on the other. 7.3.1.3. Contraband slave trade In the period at stake, undocumented contraband slave trade was the rule rather than the exception. The data presented above on the slave trade between Upper Guinea and Curaçao may thus only represent the proverbial tip of the iceberg. Not only the number of slave vessels that shuttled between the two regions, but also the number of slaves that embarked on each vessel is likely to have been higher than traceable. Roitman (2006: 8), for instance, cites a former governor of Cape Verde: “It is customary that the vessels in the port of Cacheu go to the Indies after having registered a cargo of 100, 120, or 150 pieces [slaves] (…)

7.3. Dutch slave trade from Upper Guinea to Curaçao

293

when, in fact, they are taking 800 to 1,000 pieces in each ship”. While that might be an exaggeration, it is nonetheless likely that the number of slaves purchased in Upper Guinea with destination Curaçao was much higher than officially reported. Perhaps a more realistic example is the slave ship De Diligent, which, departing from Gorée, only declared a 100 slaves to customs out of a total of 250 (Guèye 1998: 37). In the Caribbean, according to Castillo Mathieu (1982: 101), Curaçao was the principal center of contraband trade in the second half of the 17th century. In 1676, an official Spanish document shows officials in Cartagena complaining about Dutch contraband trade, which, they say, is a daily practice (1982: 102). It is revealing to note that slaves from the Cape Verde region were among the merchandise illicitly traded from Curaçao to Cartagena: in the latter part of the 17th century “the unfriendly slaves from Elmina and the strong slaves from Cape Verde were denied entrance to Cartagena, which proves they had been arriving in significant quantities”414 (Castillo Mathieu 1982: 110). While Curaçao thus seems to have been a breeding ground for clandestine (slave) trade, on the Upper Guinea Coast this role was reserved for Cacheu (cf. Green 2007: 183). Green highlights the proportions of the contraband slave trade between Cacheu on the one hand and Cartagena on the other in the first half of the 17th century: The main interest of the trading networks was the contraband trade in slaves (…). Alvarez Prieto, for instance, was constantly bringing slaves and trying to ensure that he declared as few officially as possible. (…) The official visits by the judicial and inquisitorial authorities to the slaving ships in these years make it apparent that it was routine behaviour to smuggle in as many as 510 slaves without declaring them, while 300 and 400 slaves arriving as contraband was quite normal. (Green 2007: 245)

Note that the contraband practice described here by Green covers the period up until the 1640s, i.e. shortly before Curaçao came to replace Cartagena as the principal slave trade center in the Caribbean. Had Cartagena been the principal outlet for contraband slaves from Cacheu up until the 1640s, Curaçao may well have come to fulfil this role afterwards. So when in the latter part of the 17th century a certain Ambrosio Gomes (a Sephardic Jew and former captain of Cacheu) engaged in voluminous contraband trade with ‘non-Portuguese’ traders (Moraes 1998a: 185), it is not far-fetched to assume that the Dutch were among his principal ‘non-Portuguese’ customers. 414 Original quote: “se prohibió (…) la entrada a las Indias de los ariscos minas y de los fuertes esclavos de Cabo Verde, lo que prueba que habían estado llegando antes en cantidades significativas”

294

The historical ties between Upper Guinea and Curaçao

Although ultimately we can only ponder about the true dimensions of the contraband slave trade between Upper Guinea and Curaçao, the linguistic evidence discussed in the linguistic chapters of this study suggests that it may have been significant. 7.3.1.4. Sailing from Upper Guinea to Curaçao: the normal route Also from a geographical point of view it should not be surprising that, in order to supply Curaçao, the WIC initially relied on their outposts in Senegambia. After all, to reach the Caribbean from West Africa, sea currents strongly favored a sailing route from Cape Verde rather than from Lower Guinea or Angola: “From the Cape Verde Islands, mariners could either sail south (…) along the African coast, or they could ride the North Equatorial Current westward toward the Caribbean” (Walker 2005: 60; cf. Roitman 2006: 4, 5). In other words, sailing from Upper Guinea to Curaçao was the normal route. Dutch settlements in Brazil, on the other hand, were mostly supplied with slaves from Lower Guinea and Angola for the same reasons: a favorable sea current

7.4. Sephardic Jewish networks linking Upper Guinea to Curaçao With the important exception of Martinus (1996: 120, 121, 142–145), no attempts have as yet been made to shed light on the historical ties between the Sephardic Jewish community of Curaçao and that of Upper Guinea/Senegambia. A preliminary discussion of these ties is now possible owing to the increasing body of scholarship concerned with the presence of the Sephardim in 16th/17th century Upper Guinea as well as with their familiarly and commercial connections with the Jews in Amsterdam. The importance of the role of the Sephardic Jews415 in the early settlement of Curaçao cannot easily be exaggerated and has been well-documented in the literature416 . In 1650 to 1652, when the island was still largely uninhabited, twelve Jewish families were given land to cultivate in the interior of the island (e.g. Joubert & Perl 2007: 45; see also Schors 2004: 200). Towards the end of the 17th century Curaçao harbored the largest Sephardic community in 415 In this study, the term Sephardic Jews or Sephardim is reserved for those Jews whose roots go back to the Iberian Peninsula, from where they were expelled towards the end of the 15th century. The term Sephardim derives from the name these Jews gave to the Iberian Peninsula: Sepharad (Bossong 2008: 13). 416 See, for example, Klooster (1998, 2001, 2003, 2005, 2006, 2009), Schors (2004), Granda (1974), Castillo Mathieu (1982) and Karner (1969).

7.4. Sephardic Jewish networks linking Upper Guinea to Curaçao

295

the Caribbean realm (Drescher 2001: 450). Shortly after settling on Curaçao in the 1650s, the Jews engaged in trans-Atlantic and intra-Caribbean (slave) trade (Schorsch 2004: 295). An example of the importance of the Sephardim to the commercial development of Curaçao in the second half of the 17th century is given by Israel (2009: 14): “[F]rom the 1660s onward a new system of transatlantic commerce emerged in which the Sephardim at Amsterdam and Curaçao acted as crucial intermediaries. At the height of its prosperity, at the end of the seventeenth century, Curaçao boasted a fleet of some eighty barques and sloops, a substantial proportion owned by Sephardic Jews.” The majority of the Sephardic Jews to settle on Curaçao in the 1650s, including many of those who had previously been in Brazil, are thought to have had familiarly ties with Amsterdam (cf. Goldish 2009: 5; Klooster 2009: 34). But while the Jewish Amsterdam-Curaçao connection is widely recognized, it is the purpose of this section to show that their networks extended to large parts of the Senegambia region. According to Green (2005: 166, 167), “[t]here is almost a complete absence of reference to the Jewish presence in West Africa among historians of the Sephardim”.417 Indeed, in elaborate works on the Sephardic Jews and their trans-Atlantic social and trading networks, references to the Cape Verde Islands, the Petite Côte or Cacheu are mostly absent (but see Lobban 1996). However, the importance of the Sephardim in Upper Guinea in the early 17th century, according to Green (2005: 168), is “impossible to gainsay”. Of particular interest to the present study is Green’s research into the “networks of Sephardic slavers, which were centered (…) on the Cape Verde islands and the Guinea coast” (2005: 175f.n.) as well as “previously unknown connections between Cabo Verde and the Jews of Amsterdam” (2007: 266). I will very briefly summarize these Upper Guinea-Amsterdam connections in the next paragraph (§7.4.1) and then provide examples of how the Sephardic community of Upper Guinea was related (both directly and via Amsterdam) to the Sephardic community of Curaçao, all within the 17th century (§7.4.2).

417 Green (2005: 166) thinks this absence of references to the Sephardim in West Africa “is not entirely unrelated to fears as to what might be uncovered, since it is notorious that one of the major activities of Europeans in Africa at this time was slaving. The implication of a significant number of Sephardim being involved in this activity would not sit comfortably with the traditional interpretation of many historians of the Sephardim that their subjects were, essentially, victims of persecution, and that, where they were slave owners, they treated their charges much better than did Christians”.

296

The historical ties between Upper Guinea and Curaçao

7.4.1. Ties between the Sephardim in Upper Guinea and Amsterdam According to Brooks, in the early 17th century, the Portuguese Crown decided to increase its efforts to persecute New Christians (forcibly converted Jews) residing in Upper Guinea, as a response to “allegations of commercial ties between New Christians and Jews residing in Holland” (Brooks 2003: 61; cf. Gueye 1998: 26, 27). Mark (2007: 27) shows that these allegations were not based on false arguments: “These Jewish merchants traded for ivory, which they obtained from (New) Christian traders – presumably lançados [free-lance traders often with a Sephardic Jewish background, BJ] – living in Senegambia. The lançados had established contact with the Dutch Jews as soon as the latter settled in Senegal, a fact that strongly suggests pre-existing contact.” The link between the Upper Guinea Sephardim andAmsterdam is extensively highlighted in Mark & Horta (2004): At the moment that an important Portuguese Jewish community was being established in Amsterdam, two communities of Portuguese Jews, closely affiliated with their counterparts in Holland, were growing in Senegal. (…) Dutch Jews chose to settle there [in Upper Guinea], for they were among friends, and, in many instances, co-religionists (Mark & Horta 2004: 232, 234)

Mark (2007: 206, 207) elaborates on New Christian and Sephardic Jewish merchants settled in Joal and Portudal in the early 17th century: “These traders were the descendants of forcibly-converted Portuguese Jews, but they had returned to their ancestral Jewish faith in Amsterdam. In Senegal they lived publicly as Jews and maintained close religious and commercial ties to Amsterdam. They had their own rabbi, sent in 1612 by the Sephardic community in Amsterdam.” Not only in Joal and Portudal but also in Rufisque and particularly Cacheu, Sephardic Jewish merchants were commercially active (Mark 2007: 206f.n.; Roitman 2006: 11; Green 2007). Interestingly, as pointed out in §7.2 (cf. Maps 3 & 4), the Dutch WIC had trading posts in Joal, Portudal, Rufisque and Cacheu. It is not far-fetched to assume, as does for instance Boulègue (1989: 39), that the Jewish networks between Amsterdam and Upper Guinea were beneficial to the Dutch WIC’s commercial interests in the region. And owing to the Dutch religious and commercial tolerance418 , the benefit must have been mutual. For example, Moraes (1995: 321; 1998a: 267) describes how Jewish communities on the Petite Côte would commonly receive goods from Amsterdam transported in Dutch WIC ships. Nor should it come as a surprise that lançados and other

418 Indeed, it is for this same reason – Dutch tolerance towards the Jews and their religion – that so many Jews had settled in Amsterdam in the first place.

7.4. Sephardic Jewish networks linking Upper Guinea to Curaçao

297

Luso-Africans with a Jewish background were reported as actively trading in Gorée (Brooks 2003: 63, Lobban 1996). Several individual examples, such as the Peregrino-Querido-Belmonte connection discussed in Green (2005: 170) and Mark (2007: 206, 207), or the Gramaxo family discussed in Roitman (2006), Moraes (1995: 305) and Green (2007: 3) will not be further highlighted here for reasons of space but demonstrate how closely related the Sephardim in Amsterdam and Upper Guinea actually were both on a social and commercial level, throughout the 17th century. These connections place the well-known ties between the Sephardim in Amsterdam and those in Curaçao in a different light: we may assume the existence of triangular social and (slave) trading networks connecting Senegambia, Amsterdam and Curaçao.419 Below, I will provide some concrete examples to illustrate that such networks indeed existed. 7.4.2.

Sephardim networks directly linking Upper Guinea to Curaçao

The interesting case of the Jewish Henriquez family, active in the slave trade, and with ties to Amsterdam, Cacheu and Curaçao, was already touched upon briefly in §7.2.4.1. An analysis of the relevant literature brings to light several other relevant historical links between the 17th-century Sephardim community in Upper Guinea and that residing in Curaçao. Some examples of results of what is a preliminary survey: – According to Moraes (1998a: 323, 324), the Sephardic Jewish slave trader Abraham Drago was accustomed to buying slaves in and from Cape Verde. Crucially, Abraham Drago was one of the first Jews to settle on Curaçao in 1651 (Klooster 2006: 139). The Voyages database (www.slavevoyages.org) lists two ships owned by Drago. Both ships sailed under the Dutch flag and had indeed purchased their slaves in Senegambia. Especially interesting is the slave ship Groene Viskorf, which was owned not only by Abraham Drago but also by David Nassy – another recognized founding father of the Curaçaoan Jewish community (Böhm 1992: 171). In 1664, the Groene Viskorf purchased a total of 337 slaves (destined for French Guiana) in the Dutch entrepot of Gorée. Note, furthermore, that in 1671, the same Abraham Drago requested the Dutch WIC’s permission to purchase slaves in West Africa and

419 For reasons of space, I will not consider the role of the Sephardic Jewish community in Brazil in any detail. See Goodman (1987) and Smith (1987, 1999) for discussions of this community and their ties to the Dutch WIC.

298

The historical ties between Upper Guinea and Curaçao

was subsequently told that “if he wants to trade a few Negroes in Cabo Verde, he will be treated favorably”420 (Moraes 1998b: 324). – Moraes (1998a: 183–185) analyzed a 1670 document that describes the voyage of a Portuguese slaving vessel owned by a certain Manuel Correa which, after purchasing slaves in Cacheu, anchors at Gorée and then crosses the Atlantic towards the Antilles. This piece of data gains relevance in light of the fact that the same Manuel Correa was a Sephardic Jew with family ties to Amsterdam and had become active in the slave trade after settling on Curaçao in the 1670s (Böhm 1992: 189; Goldish 2009: 166). In addition, on board Manuel Correa’s slaver was a certain Laurens de Mattos, son of Ambrosio Gomes, a former Captain of Cacheu. The same Ambrosio Gomes was also reported as clandestinely trading with ‘non-Portuguese’ powers on the Petite Côte in that period (Moraes 1998a: 185; cf. Brooks 2003: 133). – Böhm (1992: 172) introduces the Sephardic Jew Isaac de Costa, who in 1659 signed contracts with the Dutch WIC to bring to Curaçao a new group of Sephardic Jewish families. Isaac de Costa, in turn, was the nephew of a certain Bento Osorio, who was not only one of Amsterdam’s wealthiest Jews (Moraes 1995: 321), but was also reported as regularly sending merchandise to his family members residing on the Petite Côte (Moraes 1995: 321). In the context of language transfer from Upper Guinea to Curaçao, it is relevant to stress that, as the first two examples above illustrate, the networks between the Sephardic communities on both sides of the Atlantic involved slave trade. Although the literature on the history of the Sephardim is divided about the actual importance of the Jewish involvement in the slave trade (cf. footnote 417), the available historical records do suggest that slave trade was an important source of income to the Jews in that period (cf. e.g. Castillo Mathieu 1982: 55; Moraes 1995: 300; Green 2007: 96, 97; Mark & Horta 2011: 165–167). It is not surprising that the Sephardic Jews (including many lançados and New Christians) were involved in large-scale contraband (slave) trade. This was a direct result of the Portuguese inquisition: the Portuguese did not allow the Jews to engage in trade in those regions that fell under the Portuguese sphere of influence. In large parts of Upper Guinea (and particularly also in Cacheu), therefore, “the Jewish activity (…) was of necessity clandestine (…). This clandestine Jewish community coincided with the area which had become the centre for slaving operations in Guiné” (Green 2007: 183, 184). Recent archival research done by Mark & Horta (2011) allowed them to arrive at a new 420 Original quote: “s’il veut commercer quelques Nègres à Cabo Verde, il en sera traité favorablement”

7.4. Sephardic Jewish networks linking Upper Guinea to Curaçao

299

assessment of the scale of the slave trade illicitly conducted by the Jews in Upper Guinea: “These new slave trade statistics (…) help to settle the debate about the magnitude of early seventeenth-century slave exports from Upper Guinea. (…) Contraband voyages are missing, and there were a lot of them” (Mark & Horta 2011: 169).421 7.4.3.

Partnership between the Dutch WIC and the Sephardim

It is well-known that the Sephardic Jews were stockholders in the Dutch WIC. Even though their exact share in it is open to debate (Williams 2001: 391f.n.), one of the compelling facts that emerge from the historical data discussed above is indeed that both in Curaçao and in 17th-century Upper Guinea, the Dutch WIC and the Sephardic Jews seem to have been commercial allies. As we saw in §7.4.1, possibly without exception, in those parts of Upper Guinea where the Dutch WIC were commercially active, Sephardic communities are reported to have existed and, in many if not most cases, to have had ties with Amsterdam. Curaçaoan Jewish (slave) traders such as Drago and Nassy or later Henriquez sailed under the Dutch flag and Upper Guinea Jewish communities were supplied with ships from the WIC. We furthermore know that the settlement and economic development of Curaçao in that early period was facilitated by the joint efforts of the WIC and the Sephardic Jews. In light of the above, it is not far-fetched to assume that a considerable part of the slave trade conducted from Upper Guinea to Curaçao may have been jointly coordinated with success depending on the close collaboration between both parties. Interestingly, although the first generation of Jews had arrived on Curaçao already in 1650, they were effectively granted the right to import slaves only in 1659 (Emmanual 1970: 46; Böhm 1992: 172). It may not be a coincidence that this year coincides with the year in which the first ship (the Gideon) was registered as supplying Curaçao with slaves from the Petite Côte region (cf. §7.3). The 300 slaves delivered to Curaçao by the Gideon had been purchased in Portudal. Portudal, as noted in §7.2.2, was not only controlled by the Dutch WIC, but was also home to a significant Sephardic Jewish community with close ties to Amsterdam (Green 2007: 178). It is no less revealing that still in the pivotal year of 1659, Isaac de Costa headed a second generation of Jews settling on Curaçao. As it happens, Isaac de Costa was on close terms with the Governor of Curaçao at the time, Mathias Beck (Böhm 1992: 172). Recall that Mathias Beck – indeed, in 1659 – was 421 Unfortunately, I have not been able to more exhaustively analyze and integrate the research results presented in Mark & Horta (2011), as that work appeared only briefly before the completion of this study.

300

The historical ties between Upper Guinea and Curaçao

reported as filing a request to bring to Curaçao “a good lot of slaves from Cape Verde” (cf. §7.3). Incidentally, previously in Brazil, the Dutch had similarly relied on their partnership with the local Jewish planters. Newitt (2005: 181) provides interesting details about this partnership: “In 1590 it is thought that a third of all sugar mills in the Reconcavo hinterland of Bahia were owned by New Christians and in the 1620s the Dutch believed that, once they decided to invade Brazil, the New Christians would act as a fifth column in their support.” Moreover, prior to the foundation of the 1st Dutch WIC in 1621, we learn in Newitt (2005: 224) that “Dutch traders had already opened up a profitable trade with Brazil through the ports of northern Portugal where they co-operated with the local New Christian community”.

7.5. Diffusion of Upper Guinea PC to the mainland, 16th and 17th centuries As noted in §7.2.4.2, I found only one reference to the Dutch buying slaves directly from Santiago. However, even if the Dutch were no frequent customers on Santiago, they may still have had easy access to the Santiago slave market and, hence, to speakers of Upper Guinea PC, given that, as I hope to show below, (descendants from) native Cape Verde Islanders were omni-present in those parts of the Petite Côte controlled by the Dutch. The data discussed below furthermore suggest that Upper Guinea PC rather than Portuguese was the community language of these former inhabitants of Santiago. For some Cape Verde Islanders, the choice to migrate to the mainland may have been born purely out of the need to survive. It is well-known that the population of Santiago was struck by famines, droughts and epidemics in the late 16th and 17th centuries and scholars such as Green (2007: 257), Carreira (1983b: 73), Patterson (1988: 303, 306), Semedo (1993: 9) and Barbe (2002: 47) all affirm how this adversity led to a significantly increased migration rate from Santiago to the mainland. The most clear-cut illustration of this migration is the harbor of Cacheu, which, as mentioned, was settled by Cape Verdeans in the late 16th century and henceforth continued to massively attract natives (traders and free blacks) from Santiago (Green 2007: 257, Mendy 1993: 143). Of course, the fact that GBC, a sister variety of CV, is still spoken vividly today in Cacheu and surroundings provides straightforward linguistic evidence of the claim that Cape Verde Islanders once settled there (cf. Jacobs 2010). Unlike modern-day Cacheu, the Petite Côte area does not harbor any creolespeaking communities at present. However, historical documents clearly suggest

7.5. Diffusion of Upper Guinea PC to the mainland, 16th and 17th centuries

301

that at one time it did, at least until the French Company of Senegal expelled the Dutch and took control of the region in the late 17th century. Boulègue (1989: 31), for instance, mentions that in 1634 in the village that is now Dakar, “Father Alexis found only one single Christian family, who came from the Cape Verde Islands”422 . Moreover, the mayor of the village of Gaspar, also located on the CapeVerde Peninsula, was a native from the CapeVerde Islands (Boulègue 1989: 16) and in the neighboring harbor of Rufisque, around 1600, again, “there are blacks, mulattos and creoles from the Santiago Island of Cape Verde”423 (anonymous traveler cited in Boulègue 1989: 39). Furthermore, a certain Rossinde is described in an account dated 1670 as trading and competing with the Dutch in Portudal: “This Rossinde is a mulatto from the Santiago Island of Cape Verde, a righteous man and of good spirit, rich and loved by all the peoples of this coast and considered the King of Portudal”424 (Moraes 1997: 239, citing the French general d’Estrées). In order to correctly assess the presence of Cape Verdeans on the mainland in the 17th century, an identity-related remark is in order. In the 16th and 17th centuries, for reasons explained in detail in Mark (2002: 14, 15), “Cape Verdeans (…) resolutely maintained that they were ‘Portuguese’” and were denoted accordingly in travel accounts and other records from that period. The following fragment from the 1623 account of the Guinea Coast by the Dutch sailor Dirck Ruiters (cited in Brooks 1993b: 40) demonstrates this: “The trade we called ‘coastal’ is mostly undertaken (…) by Portuguese who live on Santiago Island”. This identity-related fact puts the numerous references in historical (travel) accounts to the presence of ‘Portuguese’ traders and communities along the Petite Côte in a different light and suggests that in many of these cases, reference was actually made to natives of Santiago. But not only is it historically documented that the Petite Côte was home to immigrants from Santiago, the contemporary accounts also give evidence that these immigrants had brought to the coast their native tongue, Early CV. At least, that is what the various references to a restructured variety of Portuguese spoken along the coast in the 17th century can be taken to suggest. Ladhams (2006) provides perhaps the most exhaustive collection of such historical references. Some ex422 Original quote: “le Padre Alexis ne trouva en 1634 qu’une seule famille chrétienne, qui était originaire des îles du Cap-Vert” 423 Original quote: “il y a des noirs, des mulâtres et des créoles naturels des îles de Santiago et Cabo Verde” 424 Original quote: “Ce Rossinde est mulâtre de l’isle de St. Jaque du Cap Vert, homme adroit et de bon esprit, riche et aymé de tous les peuples de cette Coste et considéré du Roy de Portudalle”

302

The historical ties between Upper Guinea and Curaçao

amples: in 1669, in Rufisque, “everyone, both men and women, speak a corrupt Portuguese” (Villaut de Bellefond 1669: 59, in Ladhams 2006: 101) and also in 17th-century Portudal and Joal the use of an ‘unintelligible form of Portuguese’ among the local population has been documented (Ladhams 2006: 90). It is important to stress that my interpretation of these accounts differs from Ladhams’ (2006). Ladhams and several others before him have interpreted these accounts as references to a widespread West African Pidgin Portuguese (cf. §1.6.1). I rather assume these accounts to be references to an early but full-fledged, native variety of CV (/ Upper Guinea PC), which makes sense if, as argued, the communities at stake were indeed composed largely of natives of Santiago (or direct descendants thereof), most of whom must have had CV as their mother tongue. Of particular interest to the present study is the historical possibility that CV was spoken among the population of Gorée. Though I found no contemporary eye-witness reports affirming this, several scholars (e.g. Curtin 1975: 114; M’bow 1985; Haardt 1992, and particularly Carreira 1983a: 63, 64) state it as likely that the society of Gorée, though limited in size it was, spoke a restructured variety of Portuguese (viz. CV) during the Dutch occupation. Parkvall (1999: 197, drawing on Curtin 1975: 114), for instance, notes that the local population of Gorée encountered by the French after having conquered the island in 1677 consisted of “Wolof-speaking Lebus or Portuguese Creole-speaking mulattoes”. An important correlation may furthermore have existed between the spread of natives from Santiago to the continent and the Christianization of the Guinea region by Cape Verdean missionaries starting in the mid-16th century. The bishop of Santiago would regularly dispatch to the continent so-called visitadores, priests baptizing the newly born and instructing the local Christians how to organize their own masses. One of the settlements receiving primary attention from the Diocese of Santiago was Cacheu (Benoist 2008: 28), the others were Gorée, Rufisque, Portudal and Joal: “These towns were inhabited by local people, but especially by mestizo traders from the Cape Verde Islands who had remained deeply attached to Catholicism”425 (Camara & Benoist 2003: 118). In this context, it is interesting to point out the documentary evidence provided by Carreira (1983b: 72) that on Santiago, by the early 17th century, most religious services were already carried out by crioulos. Soares (2006: 194) refers to these creole clergymen as ‘fathers of this land’, born and raised ‘in the creole world’, and emphasizes the likelihood that quite early on CV, rather than Portuguese was used in clerical circles and to chatechize Africans (cf. Baleno 2006: 154–156). 425 Original quote: “Ces agglomérations étaient habitées par des populations locales, mais surtout par des commerçants métis venus principalement des îles du Cap-Vert et restés très attachés au catholicisme.”

7.5. Diffusion of Upper Guinea PC to the mainland, 16th and 17th centuries

303

Interestingly, the chronology of the testimonies gathered by Ladhams (2006) suggest a correlation between the presence of the Dutch on the Petite Côte and the presence of ‘Portuguese’ communities in that same region: with the withdrawal of the Dutch from the Petite Côte in the late 17th century, references to ‘Portuguese’ communities and ‘corrupted Portuguese’ also disappear. For Rufisque, Ladhams (2006: 89) notes that “the last such reference dates from 1695”; in Portudal, the ‘Portuguese’ are reported to have withdrawn around 1700426 . It seems, then, as if the withdrawal of the Dutch (who had allowed and possibly even welcomed other commercial and religious entities to live and trade in the region) after 1677 and the subsequent French monopoly (already hinted at in §7.2.3) was a significant deterrent to the local Luso-African and other ‘Portuguese’ trading communities, whose subsequent departure from the region would only have been a matter of time.427 This historical scenario accounts for the absence of CV- / Upper Guinea PC-speaking communities in the present-day Petite Côte region, where they once cohabited and traded with the Dutch. To summarize, while the sections 7.2–7.4 showed that the Dutch and Sephardic Jews were trading (slaves) between Senegambia and Curaçao from the mid to the late 17th century, this section showed that in that same period, that same area was home to pockets of Cape Verde Islanders.428 This suggests, first of all, that the Dutch may have had access to the Santiago slave market even without being physically present on Santiago. On a linguistic level, I argued that a full-fledged native variety of CV (/ Upper Guinea PC) must have been quite widespread in the 17th-century Petite Côte region. This leaves room for thought-provoking speculation in light of the thriving connections between the Dutch, the local Sephardim, and other Luso-African locals in that area and period, for instance, about the possibility that Dutch and Sephardic Jewish traders themselves had gained L2 knowledge of Early CV. In any case, so it seems, both the Dutch WIC 426 Only in Joal, an ‘unintelligible form of Portuguese’ was apparently still used until the late 18th century. In the mid-19th century, however, French missionaries stated that also in Joal “the use of Portuguese was a mere memory” (Ladhams 2006: 90, drawing on Boulègue 1989). 427 Like their contrasting commercial policies, the Dutch and French sociolinguistic attitudes towards creolophone communities may also have diverged significantly. I assume, for instance, that the Dutch must have attempted to learn the local trade language(s), whereas, as history has shown, the French implemented their own language as the main vehicle of communication on the Petite Côte. 428 In some cases, references to Jewish families in one source and to families stemming from the Cape Verde Islands in another source may of course represent references to the same families, given that many of those who exchanged the Cape Verde Islands for the mainland in fact had a Jewish background.

304

The historical ties between Upper Guinea and Curaçao

and the Sephardim, i.e. the two entities mainly responsible for the 17th-century development of Curaçao, had direct access to, and were in close contact with, speakers of Early Upper Guinea PC.

7.6. Summary, conclusions, and final remarks The primary purpose of this historical chapter was to show that all conditions were met for a linguistic transfer from Upper Guinea to Curaçao to take place in the second half of the 17th century, which is the period in which PA is generally thought to have emerged on Curaçao. First (§7.1), I noted the consensus that Senegambia was insignificant to the history of the Dutch WIC and, by extension, of Curaçao and PA. One of the aims of the present chapter was to show that this consensus needs to be reconsidered. With that purpose, §7.2 discussed a precondition for the language transfer from Upper Guinea to Curaçao to take place: the presence of the Dutch in 17thcentury Senegambia. It was shown that over a period of some five decades, the Dutch controlled large parts of the Petite Côte and, in addition, enjoyed thriving commercial ties with Cacheu, the principal slave market of the region, as well as with local trading communities composed of (descendants of) natives from the Cape Verde Islands. In §7.3, little-known data on the slave trade between Senegambia and Curaçao was offered. It was shown that, starting in the 1650s and ending with the Dutch withdrawal from Senegambia in the late 1670s, Curaçao received several shipments of slaves purchased in Senegambia. We furthermore pointed out that this period of intense contact between Upper Guinea and Curaçao coincided with the island’s peak period of 17th-century slave trade. One important conclusion to be drawn from these data is that the language transfer from Upper Guinea to Curaçao must in all likelihood be situated in the window from, roughly, 1650 to 1680. Section 7.4 drew attention to 17th-century Sephardic social and (slave) trading networks linking Amsterdam to Upper Guinea and Upper Guinea to Curaçao. Though much research is still needed to fully understand the role of the Sephardim, the data presented in that section suggest they were important agents in the historical framework that encompasses the transfer of speakers of Upper Guinea PC to Curaçao. Finally, § 7.5 showed that those parts of the Upper Guinea coast characterized by Dutch-Sephardic commercial activity were also the parts where numerous pockets of former inhabitants of Santiago, viz. native speakers of Early CV, had chosen to migrate to in search of better fortune. One of many possible

7.6. Summary, conclusions, and final remarks

305

implications is that knowledge of Early CV may not have been uncommon among Dutch and Sephardic traders. This idea remains rather tenuous, however, and the thrust of the historical framework presented in this chapter does not in any way depend on it. For future study, Quint’s (2000b: 196) proposal for genetic research in order to pinpoint more exactly the origins of the early Curaçaoan population should be pursued. Gonçalves et al. (2003), for instance, have established a Sephardic Jewish contribution to the Cape Verdean gene pool, giving an indication of the results such research might provide. Similar research is currently being conducted in relation to the Spanish-based creole PLQ (Yves Moñino p.c.) and the Portuguese-based creole of SãoTomé (Coelho et al. 2008) with the purpose of gaining a better understanding of the ethnic origins of the speaker communities and ultimately of these creoles’ genesis and structure.

Chapter 8 Discussion: The development from Upper Guinea PC to Papiamentu

Introduction Thus far, we have been able to demonstrate a considerable overlap between PA and Upper Guinea PC in the fundamental parts of their grammar, particularly in the morphosyntax (chapters 3–5), which is the domain where we could expect to find the primary evidence for kinship, but also in the phonology, discussed in chapter 2. Chapter 7 was organized with the purpose of showing that the historical conditions were met for the linguistic transfer from Upper Guinea to Curaçao to occur in the second half of the 17th century. This tells us that we need not doubt the kinship suggested by the linguistic data and that we do not have to appeal to chance to explain the structural correspondences. Thus, if the aim of the present study was to scrutinize the hypothesis that PA and Upper Guinea PC are genetically related, the linguistic and historical data analyzed in the previous chapters have corroborated the plausibility of this hypothesis. On the basis of the linguistic and historical evidence amassed thus far, I will make the following assumptions: 1. PA is genetically and historically related to Upper Guinea PC; 2. The transfer of speakers of Upper Guinea PC to Curaçao took place roughly in the period between 1650 (the approximate start of slave trade on Curaçao) and 1680 (which is when the Dutch withdrew from Upper Guinea and thus presumably ceased to ship slaves from that region to Curaçao). 3. Taking 1680 as the approximate cut-off point of the two creoles, a time-depth of slightly more than three centuries separates the two creoles. In this closing chapter, I will take these assumptions as a point of departure and address several remaining issues that still require an explanation if we are to understand how Upper Guinea PC could establish itself on Curaçao and develop into what we now know as PA. One of these issues is the sociolinguistic factors that caused speakers of Upper Guinea PC to be wanted on Curaçao and to have a decisive impact on the linguistic scene (§8.1). Then, in §8.2, I will provide an elaborate discussion of the process of relexification towards Spanish that I believe underlies the development of PA out of Upper Guinea PC. Amongst other things, I will examine the benefits of analyzing PA as a mixed

308

Discussion: The development from Upper Guinea PC to Papiamentu

or intertwined language with an Upper Guinea PC morphosyntactic matrix and a Spanish embedded content lexicon (§8.2.3).

8.1. Sociolinguistic considerations While it was shown in chapter 7 that slave trade from Upper Guinea to Curaçao in the second half of the 17th century was probably much greater than traditionally assumed, it is indisputable that slaves from that region were not the only ones being shipped to Curaçao in that same period. It must be acknowledged that a considerable number, and perhaps the majority, of slaves were drawn from WIC entrepots along the coasts of Lower Guinea, Congo and Angola, which means that Upper Guinea PC-speaking slaves may well have constituted only a minority on the total of slaves that arrived on Curaçao in that early period. Furthermore, as Mikael Parkvall (p.c.) points out, one must distinguish between net and gross slave imports into Curaçao, “since many or most slaves did not stay on the island, but were re-exported”. These circumstances bring up two pivotal and interrelated questions: 1. Of the many thousands of slaves brought to Curaçao in the early period, why were Upper Guinea PC speakers selected to stay on the island rather than to be resold? 2. How could an assumed minority of Upper Guinea PC speakers, amidst newly arriving slaves from other linguistic areas consolidate its language and impose it onto the remaining parts of the rapidly growing and linguistically heterogeneous Curaçaoan slave society? These questions will be addressed in §8.1.1 and 8.1.2 respectively. 8.1.1.

On the choice of slaves in the early period of Curaçao’s settlement

It is important to recall that Curaçao started out as a transitory slave trading post. Most of the slaves that were brought in were held in slave camps and subsequently resold to third parties in the Spanish Americas and the Caribbean.429 In the first two to three decades of slave trade, between the 1650s and 1680s, the vast majority of slaves were resold, while only few would remain on the island. Consequently, the historical questions arise as to (a) which slaves were chosen to be kept on the island and which were preferred for resale and (b) which sociohistorical and/or socio-linguistic motives underlie this preference. Given the 429 Van Welie (2008: 62) describes this situation as follows: “Except for their short ‘layover’ at Curaçao and their impact on the island economy, most of these Africans quickly disappeared from the Dutch colonial realm. And while this transit trade enhanced the historic reputation of the Dutch as slave traders, the slaves themselves ended up in Spanish, not Dutch colonies”.

8.1. Sociolinguistic considerations

309

fact that, as Postma (1990: 106) notes, “such preferences might differ from one market area to another, and also from one time to another”, these two correlating issues are not easily resolved and one is inclined to agree with Hoetink (1958: 69): “The question of positive or negative selection of the Curaçaoan negro slaves cannot be answered due to a lack of data”430 . In that same vein, Allen (2007: 68) claims: “The question as to why certain enslaved people were kept on the island while others were transported, is a difficult one to answer due to the lack of available historical sources”. Fortunately, knowing which slaves remained on the island allows us to more effectively assess the reasons why these slaves, and not others, were chosen to stay. Here, in virtue of the lack of explicit historical sources, the value linguistics may have to historiography becomes clear: if the linguistic data presented in the present study and conclusions drawn from them are correct, we must necessarily assume that Upper Guinea PC-speaking slaves were particularly sought after by Curaçao’s first generation of colonizers and thus often kept on the island rather than being (re)sold to third parties, which, as noted, was the fate of the vast majority of slaves shipped to Curaçao. Below, I will discuss four circumstances that help understand why speakers of Upper Guinea PC were particularly in demand by the Dutch and Jewish settlers and thus selected to remain on Curaçao. 8.1.1.1. Preference for ‘experienced’ slaves One such circumstance is hinted at by Postma (1990) when he explains why slaves from Calabary (on the coast of what is now Nigeria) were disliked by the Dutch: “One can speculate that Calabary slaves originated from the predominantly stateless societies of that region, people who were accustomed to a great amount of individual freedom and were thus, like most native Americans, less able to adjust to the degradation and regimentation of enslavement” (Postma 1990: 107, 108). Conversely, we may postulate that, owing to their own history of and experience with slavery starting in the 1460s, slaves from the Cape Verde Islands (as well as Cacheu for that matter) were, if not already accustomed, particularly able to adjust to the degradation and several other aspects of the slave society that Curaçao was rapidly turning into. At least, Curaçao’s Dutch/Jewish slave owners may have considered this to be the case and therefore, as a type of low risk investment, have preferred slaves from the Cape Verde Islands and/or Cacheu to work their plantations and perform housekeeping tasks.

430 Original quote: “De vraag naar positieve of negatieve selectie der Curaçaose negerslaven kan bij gebrek aan gegevens niet beantwoord worden”

310

Discussion: The development from Upper Guinea PC to Papiamentu

8.1.1.2. Preference for light-skinned slaves Another circumstance that may underlie the early Curaçaoan colonizers’ preference for Upper Guinea PC-speaking slaves is related to skin color. The literature on the early Curaçaoan slave society warrants the assumption that Dutch and Jews in search of slaves generally preferred light-skinned over dark-skinned slaves. This preference must have favored the selection of the racially-mixed and thus generally lighter-skinned slaves from the Cape Verde region. I elaborate on this below. That a person’s degree of whiteness was directly proportional to his/her social prestige and status within the Curaçaoan society is often implied in descriptions of this society by scholars such as Hoetink (1958), Marks (1976) and Allen (2007). Hoetink (1958: 81), for instance, described the establishment within the early Afro-Curaçaon society of a hierarchy based on skin color with a clear positive correlation between the high social ranks and a light skin (1958: 74). That a skin-color-based social hierarchy had been established on Curaçao by the 18th century is noted also by Schors (2005: 253): “On Curaçao, nonwhites were simply and consistently denied participation in the community”.431 From these observations, it follows that light skin was an essential part of the European colonizers’ ideal of beauty (cf. Hoetink 1958: 74, 75; Marks 1976: 86; Allen 2007: 73, 74). We may assume that this ideal was an important selection criterion applied by the Dutch/Jewish settlers in their choice of slaves, especially as regards the house slaves (cf. Hoetink 1958: 74; Bartens 1996: 246), and, consequently, that light-skinned slaves were more likely to be kept on the island than dark-skinned slaves. Against this background, it becomes increasingly relevant to recall that the racial mixing of blacks and whites had started on the Cape Verde Islands as early as the late 15th century, so that by the late 17th-century, significant portions of the Cape Verdean society are likely to have been of mixed descent and, consequently, lighter skinned than slaves from, say, the Gold Coast, Congo or Angola, where the mixing between blacks and whites was far less advanced. In brief, being lighter-skinned than slaves from other areas, a relatively high proportion of Upper Guinea PC-speaking slaves is likely to have been kept on the island rather than to be resold.

431 If we may believe Cuales (1998: 89, 90), a color-based hierarchy is still prevalent on the Netherlands Antilles, “with the more light skinned mixtures (…) being placed in a higher social category of the continuum”. Cf. also the comments made by Holm (2004: 62, drawing on Cohen 1980) regarding the generally positive attitude of Iberian colonizers towards light-skinned slaves.

8.1. Sociolinguistic considerations

311

8.1.1.3. Preference for slaves with knowledge of an Iberian-based speech variety A third reason why Upper Guinea PC-speaking slaves may have been more in demand than slaves from other areas, I assume, is the very fact that they spoke Upper Guinea PC. One can easily imagine that, if Dutch or Jewish settlers were to pick one slave out of ten, they would pick the one with whom mutual intelligibility was the least problematic. Note in this respect that not only the Sephardic Jews but also many of the Dutch officials had knowledge of either Spanish or Portuguese (whether restructured or not), or both (cf. Kramer 2004: 122–137), and in some cases may even have had L2 knowledge of Upper Guinea PC, as tentatively argued in §7.5. The point is, even if a slave master’s knowledge of Portuguese was only basic, Upper Guinea PC-speaking slaves must have offered him great communicative advantages over slaves with only an African linguistic repertoire. Put differently, at all times, Upper Guinea PC would have been more intelligible to the Dutch and Jewish settlers than any given African language. Note that this argument cannot be challenged by the claim that slaves from other West African colonies spoke Portuguese. Very few actually did. In Angola, for instance, Portuguese only became the language of the local blacks in the 20th century (Liliana Inverno p.c.). 8.1.1.4. Preference for christianized slaves Although many of the first Dutch settlers of Curaçao were presumably Protestant, there must have been tolerance towards Catholicism within the ranks of the Dutch WIC, judging from the fact that it had Portuguese and Spanish Catholics in its service (Kramer 2004: 124). But whatever their religious conviction, we may assume that Dutch and Jewish colonizers in search of domestic slaves and nannies would have felt more comfortable with the systematically baptized and catechized creole slaves from the Cape Verde region than with non-christian slaves from other regions. 8.1.1.5. Summary Taken together, the above arguments can of course be rephrased into the more general claim that the linguistic, cultural and religious distance between Europeans and creoles from Upper Guinea was simply more easy to overcome than that between Europeans and other slaves, a fact favoring the selection of Upper Guinea PC-speaking slaves over other slaves, particularly in as far as domestic slaves were concerned.

312 8.1.2.

Discussion: The development from Upper Guinea PC to Papiamentu

Sociolinguistic issues relevant to the consolidation of Upper Guinea PC on Curaçao and its diffusion among the (slave) population

As noted previously, we cannot but acknowledge that in the first decades of slave trade on Curaçao, besides slaves from Upper Guinea, considerable numbers of slaves originating from Lower Guinea, Congo and Angola were brought to the island to satisfy the transit slave trade. The previous section discussed which properties most plausibly account for the fact that Upper Guinea PC-speaking slaves, in spite of constituting only a small percentage of the gross slave imports, were more likely to be kept on the island than others. These properties include: 1. their experience with, and capability to adapt to, the degrading conditions of slavery; 2. their relative whiteness compared to other slaves; 3. their command of a Portuguese-based speech variety; 4. their having been baptized and catechized. At the same time, however, we must assume that no or hardly any more Upper Guinea PC-speaking slaves were brought to Curaçao after the loss of Gorée in 1677 and the subsequent Dutch withdrawal from Upper Guinea. One can imagine that, as a consequence, the Upper Guinea PC-speaking slaves present on Curaçao would soon, perhaps already in the 1680s, be significantly outnumbered by slaves from Kwa- and Bantu-speaking areas. This leads to the question as to how they could nonetheless retain, consolidate and diffuse their speech among subsequent arrivals. In order to understand this, it is crucial to acknowledge in line with Parkvall (1999: 205) that “a relatively small group of people, given favorable conditions, can (…) have an influence upon (…) the emerging language that exceeds their numerical importance”. Below, I will hypothesize about what these ‘favorable conditions’ may have been. In order of presumed importance, I will first argue that the variety of Upper Guinea PC brought to Curaçao was a full-fledged creole and mother tongue of the people who spoke it (§8.1.2.1); secondly, that these people constituted the founders of Curaçao’s slave population (§8.1.2.2) and, finally, that being a speaker of Upper Guinea PC (§8.1.2.3) and being lighterskinned (§8.1.2.4) were important markers of prestige. 8.1.2.1. The imported variety: a full-fledged creole, not a jargon or pidgin In the 1970s, some consensus had in fact been reached about the Portuguese pidgin-origin of PA: “The Portuguese element in Jamaican Creole could have entered as a part of Portuguese Pidgin spoken in the Caribbean: it formed the basis of Papiamento” (Cassidy 1971: 207). In a similar vein, Hancock (1971: 513) classified PA as a “Spanish creole derived from an earlier Portuguese pidgin”. The examples reflect the fact that the claim that PA has Afro-Portuguese origins has traditionally entailed the idea that the proto-variety transmitted from

8.1. Sociolinguistic considerations

313

West Africa to Curaçao was a pidgin or a jargon, rather than an elaborated creole.432 In fact, this consensus has survived well into the 1990s. For example, Maurer summarizes the scholarship that claims a Portuguese origin for PA as follows: “According to some authors, an Afro-Portuguese pidgin was relexified by Spanish before stabilizing; according to others, an Afro-Portuguese pidgin stabilized prior to being relexified”433 (Maurer 1998: 196, emphasis added). As this summary indicates, in the body of literature concerned with the alleged Portuguese origins of PA, the possibility that PA actually derives from a Portuguese-based creole has remained significantly undervalued (see also van Wijk 1958: 181; Thompson 1961: 112; Feingold 1996: 80; Eckkrammer 1996: 59; Kramer 2004: 122–138). Sometimes, a clear distinction between pidgins and creoles is simply not made when PA’s origins are discussed. For instance, the use of terms such as ‘Afro-Portuguese proto-pidgin/creole’ is not uncommon, while in yet other cases, the terms ‘proto-pidgin’ and ‘proto-creole’ are used indiscriminately within the same publication. In his assessment of the state of the research on PA’s origins, Busche (1993: 109) provided similar criticism: “The boundaries between an Afro-Portuguese dialect, trade jargon or pidgin on the one hand and an Afro-Portuguese proto-creole on the other have as yet not been drawn clearly. It is also noticeable that all terms are often mentioned in one breath”434 (cf. the remarks made in §1.6.1). Below, I will claim that the variety of Upper Guinea PC brought to Curaçao was a full-fledged creole rather than a pidgin or trade jargon and argue that recognizing this not only puts an end to the terminological inconsistency described

432 This section takes as a starting point the traditional view that creole languages derive from pidgins (e.g. Hall 1966, Bickerton 1981, Thomason & Kaufman 1988). An opposing line of thought holds that creoles are direct descendants of their lexifiers (Chaudenson 1995, Mufwene 2000, DeGraff 2003). For compelling arguments in favor of the pidgin-creole cycle, I refer the reader to McWhorter (2001), McWhorter & Parkvall (2002), Parkvall (2008b) and Bakker, Daval-Markussen, Parkvall & Plag (2011). 433 Original quote: “Según algunos autores, un pidgin afroportugués fue relexificado por el español antes de llegar a estabilizarse; según otros, un pidgin afroportugués se estabilizó antes de ser relexificado” 434 Original quote:“Es konnte bisher immer noch nicht genau herausgearbeitet werden, wie die Grenze zwischen einem afroportugiesischen Dialekt / Handelsjargon / Pidgin (…) einerseits und einem afroportugiesischen Protokreol andererseits zu ziehen sind (…). Es fällt auch auf, daß oft alle Termini in einem Atemzug genannt werden.”

314

Discussion: The development from Upper Guinea PC to Papiamentu

above, but, more importantly, allows us to gain insight into how Upper Guinea PC managed to survive and consolidate itself on Curaçao. That the Upper Guinea PC variety brought to Curaçao had a particularly elaborate grammar is suggested by the linguistic correspondences between PA and Upper Guinea PC described in chapters 3–5. Several of these correspondences concern features that are generally thought to be absent or only marginally present in pidgins. The nature of these correspondences clearly suggests that nativization and subsequent grammatical expansion had already occurred prior to the separation of PA from Upper Guinea PC and thus warrants the assumption that the variety of Upper Guinea PC brought to Curaçao was a full-fledged creole rather than a rudimentary pidgin or trade jargon. Some examples: – According to Siegel (2004: 146), there may be exceptions, but in general pidgins appear to be characterized by “the virtual absence of productive bound morphology – inflectional or derivational” (cf. e.g. Bickerton 1995: 1452; Veenstra 1996: 259; Sebba 1997: 116; McWhorter 2005: 10; Parkvall 2006: 324) As was shown in §4.1 and §4.2, PA and Upper Guinea PC show a significant resemblance exactly in the domain of productive derivational and inflectional morphology; – Still with respect to inflectional morphology, Parkvall (2000: 20) identifies “the lack of passives as one of the most typical features of Pidgins”. PA’s passive morphology, as analyzed in §4.3, seems to have been inherited from Upper Guinea PC; – According to Bickerton (1995: 1452), “superstrate articles, auxiliary verbs and complementizers, as well as some question words, prepositions and conjunctions” are not found in the grammar of a typical pidgin (cf. Parkvall 2000: 67 and Bruyn 2008: 390, 391 on the lack of conjunctions in pidgins). As demonstrated in chapters 3 and 5, PA’s and Upper Guinea PC’s stocks of prepositions (§3.2), question words (§3.3), conjunctions (§3.4) and auxiliary verbs (§5.7) show a near-paradigmatic correspondence and the creoles furthermore share the complementizer pa (§3.2.5) as well as an article/deictic marker derived from the superstrate (§3.5.2); – With particular respect to interrogatives, it appears that those found in pidgins are typically bi-morphemic (Bickerton 1995: 1458). As shown in §3.3, however, PA and Upper Guinea PC share not only a set of bi-morphemic, but crucially also their monomorphemic interrogatives, which I take as a sign that the shared ancestor language was well beyond the pidgin stage by the time it arrived on Curaçao; – In keeping with numerous other scholars, Plag (1994: 9) asserts that “Pidgins lack a grammaticalized TMA system, i.e., TMA categories are not expressed

8.1. Sociolinguistic considerations

315

by obligatory grammatical morphemes”. Rather, the tense, mood and/or aspect relations of an utterance “are either not expressed at all and must be inferred from the context, or are optionally expressed by adverbials” (Plag 1994: 9; cf. e.g. Labov [1971]1990: 16; Bakker 2008b: 49, 50; Muysken 2008: 191; Bruyn 2008: 390, 391). The commonalities between the fully grammaticalized TMA systems of PA and Upper Guinea PC were extensively highlighted in chapter 4; – According to Bickerton ([1974]1980: 4), Hawaiian Pidgin English was characterized by being “deficient in the distinctions it is capable of drawing between different types of state, event or action”. Peter Bakker (p.c.) confirms that pidgins generally do not make a stative-nonstative distinction in the way this is typically done in creoles and, indeed, also in PA and Upper Guinea PC (cf. §5.6). Of course, Heine & Kuteva (2002: 10f.n) correctly remind us that “the boundary between pidgins and creoles is not seldom fuzzy” and it is nowadays finding recognition that the grammar of stable and expanded pidgins such as Tok Pisin and West African Pidgin English may contain morphology and other more complex functional categories (Bakker 2008b: 49; Lefebvre 1998: 4).435 But none of that should in any way affect the overall impression that the grammar of the ancestor variety from which I assume PA descends was characterized by a particularly high degree of grammatical elaboration, which, in my view, warrants the assumption that this variety was a full-fledged creole, transferred to Curaçao by fully competent native speakers.436 The idea that the variety of Upper Guinea PC transferred to Curaçao was a natively spoken creole with an elaborate grammar perfectly matches the widespread assumption, referred to in the introductory chapter of this study that creolization had started on Santiago already in the latter stages of the 15th century and was completed by the mid-16th century, if not earlier. Thus, by the time speakers arrived on Curaçao (1650–1680), it had already had plenty of 435 Note, however, that the classification of such stable expanded varieties as pidgins becomes problematic when they have acquired a native speaker community in addition to their original function as an L2 and out-group speech variety (Bakker 2008b: 49). The modern literature on pidgins therefore distinguishes more carefully between pidgins and pidgincreoles (e.g. Bakker 2008a). 436 Incidentally, in the context of his afrogenesis hypothesis for the Atlantic Englishbased creoles, McWhorter (2005: 223) draws on very similar linguistic arguments (i.e. the relatively complex nature of the correspondences between them) to support the claim that the shared ancestor variety “had passed far beyond the stage of a jargon”.

316

Discussion: The development from Upper Guinea PC to Papiamentu

time to expand syntactically and morphologically owing to the ongoing contact with a morphologically complex language such as Portuguese, but possibly also to internal, contact-independent processes of complexification. The importance of recognizing that PA’s proto-variety was a natively spoken creole (rather than a pidgin or a jargon) – and the reason why it is argued for in this particular part of the present study – is that it significantly increases our understanding of how Upper Guinea PC could survive and consolidate itself on Curaçao and subsequently spread to other layers of society. Compared to an unstable pidgin, by definition spoken only as an L2, an elaborated, stable, and by definition natively spoken creole language is much more likely to have a linguistic impact on the region its speakers move to. The following comment by Alleyne (1980: 138) is clarifying in this respect: “It is axiomatic of all (…) change arising out of language contact that there will be transmissions or continuities from the native language of the people undergoing linguistic change. There has never been any community known that has moved from one native language to another without there having been such transmissions and (dis)continuities.” To summarize, whereas it is clear that native speakers of a creole may crucially influence the linguistic landscape of the place where they settle, this is far less clear for L2 pidgin speakers. Consequently, recognizing the nativized and full-fledged status of the Upper Guinea PC variety brought to Curaçao contributes to understanding how this variety could survive on the island and subsequently be diffused to other segments of the expanding society. 8.1.2.2. The Founder Principle In addition to Upper Guinea PC being a full-fledged creole, for its speakers (whom I assume were slaves) to have had such a decisive impact on Curaçao’s linguistic landscape, the assumption that these speakers were the very first to populate the island might prove crucial. The demand for slaves on the island started to increase rapidly in the 1650s with the launching of active settlement policies on the part of the WIC in cooperation with the Sephardim. As shown in §7.3, this initial period of settlement coincided with some two decades of slave trade between Gorée and Curaçao. Although in that early booming period, the Dutch doubtlessly also brought slaves to Curaçao from other locations in West Africa, the previous section (§8.1.1) showed that Upper Guinea PC-speaking slaves were always the most likely ones to remain on the island, and may therefore well have been the first to work on the farms and plantations and as domestic slaves. If speakers of Upper Guinea PC indeed constituted the founding population of the Curaçaoan (slave) society, various aspects of the so-called Founder Principle, as outlined by Mufwene (1996), may have applied to the language

8.1. Sociolinguistic considerations

317

contact situation that subsequently emerged on the island. One particularly relevant observation made by Mufwene (1996: 92) is that “it was generally more cost-effective for subsequent generations of immigrants (free, enslaved, and indentured) to learn the emerging local vernaculars than to develop new ones from scratch”. In other words, it is quite possible that, being the first to have populated the interior, a relatively small number of Upper Guinea PC-speaking slaves was able to have a disproportionately great influence on the resulting linguistic scene by passing on their language to newcomers. As a case in point, we may take Lipski’s (1987b) description of how the Zamboanga dialect of Chabacano – the only form of Chabacano to survive the Spanish colonial period “including the last 100 years, when the Spanish language was (…) taught intensively to the native population” as well as “the American occupation, the ravages of World War II, and the nationalistic campaigns in favor of Tagalog” (1987b: 91) – is passed on to newcomers: “Chabacano is the first language of nearly all natives of the city [of Zamboanga] (…) and newcomers who have moved to Zamboanga City find they have to use at least some Chabacano in order to adequately carry out their activities. Among native Zamboangueños, Chabacano is the preferred language in nearly all informal situations” (Lipski 1987b: 92). Assuming that native speakers of Upper Guinea PC were indeed the first to populate the big houses, plantations and farms in the interior of Curaçao, a similar scenario as sketched here by Lipski may have applied to the developing slave society on Curaçao, with new arrivals having no choice but to acquire Upper Guinea PC (viz. Early PA) in order to become properly integrated. 8.1.2.3. The advantages of knowing Upper Guinea PC Another factor of importance to understanding the retention and expansion of Upper Guinea PC on Curaçao in the early period was hinted at in §8.1.1.3 and can be found in the nature of the creole itself. The very reason why creole languages have often come out on top in linguistically heterogeneous settings is of course that they are more accessible (sociolinguistically as well as typologically) to African slaves than any of the languages spoken by the European colonizers and, vice versa, more intelligible to the colonizers than any of the African languages could ever be (cf. Parkvall 1999: 205). But understanding and speaking Upper Guinea PC not only offered obvious practical advantages; we may also assume that it was a marker of status: “Knowing a speech form that allowed communication with both the masters and fellow slaves of other ethnolinguistic backgrounds may in itself have entailed a certain status” (Parkvall 1999: 205). Translated to Curaçao’s slave society in the latter stages of the 17th century, we may assume that being a speaker of Up-

318

Discussion: The development from Upper Guinea PC to Papiamentu

per Guinea PC was beneficial from both a practical and a social point of view, making its development into the principal vehicle of communication among the slaves quite predictable. 8.1.2.4. The advantages of being lighter-skinned The idea of speaking Upper Guinea PC as a marker of status can be added to the previously-mentioned fact that being a lighter-skinned slave entailed social prestige on Curaçao (§8.1.1.2). As noted, since they were creoles, speakers of Upper Guinea PC must in general have been lighter-skinned than slaves from other regions. Significantly, Hoetink (1958: 73, 74) affirms that, although the percentage of light-skinned slaves on the plantations and farms was often reduced, they would typically occupy the higher ranked positions, which, in turn, must have allowed them to have a considerable linguistic impact. The benefits of being lighter-skinned in the early Curaçaoan slave society are also noted by Allen (2007: 73, 74), though with particular respect to domestic slaves: The term ‘coloured people’, hende di koló, meant that one had a lighter and therefore more ‘appropriate’ skin colour, thus with slightly better chances for upward mobility.Afro-Curaçaoan women especially could reach certain positions on account of their lighter complexion. They were more likely to be chosen to work within the household of the enslaved people’s owner; they would work, for example, as yayas (nannies) and as domestic servants.

An important related question that remains for future research is whether Upper Guinea PC (viz. Early PA) spread from the plantations and farms into the houses or the other way around. Scholars are divided in this respect. Some have claimed an important role for the Afro-Curaçaoan yayas (nannies) in facilitating the emergence and spread of PA: “A key role within the linguistic output of the African slaves can be attributed to the jaja, the black nurse who brought up the master’s children” (Eckkrammer 1999: 61; cf. van Wijk 1958: 181). Martinus (1998: 114), however, believes that “the language moved from the fields [i.e. the plantations and farms] into the Big Houses – not the other way around”, a view shared by the historian Krafft (1953, cited in Maduro 1965: 18). For demographic reasons, however, I am inclined to line up with Eckkrammer and van Wijk: in 1683, no less than 75% of all the slaves residing on Curaçao worked as domestic slaves (Hoetink 1958: 58; Bartens 1996: 243). Moreover, while the number of light-skinned slaves (read: potential Upper Guinea PC speakers) on plantations was relatively low, light-skinned slaves constituted the majority among the group of domestic slaves (Hoetink 1958: 75; cf. Allen’s observations above). It is a small step from assuming that the majority of the first Curaçaoan

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

319

yayas were Upper Guinea PC-speaking women to understanding how Upper Guinea PC could be consolidated and diffused on Curaçao.

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish While some of the arguments put forward above are more tentative than others, their sum allows us to better understand how, even if the percentage of Upper Guinea PC speakers out of the total of imported slaves was relatively minor, Upper Guinea PC was able to establish itself on Curaçao, only to subsequently develop into what we now know as PA as a result of linguistic change. The particularities of this change will be discussed in this section. The comparisons of the grammatical categories generally considered to constitute the morphosyntactic skeleton of a (creole) language (TMA markers, auxiliaries, conjunctions, pronouns, demonstratives, derivational and inflectional morphology, but also the prepositions) revealed that PA and Upper Guinea PC correspond structurally in these domains, with grammatical morphemes coinciding both in form, semantic scope and syntactic behavior, while often neatly differing in those regards from their Iberian etyma as well as from the parallel morphemes of other Iberian-based creoles (cf. the summary provided in chapter 6). However, in spite of the structural morphosyntactic overlap, mutual intelligibility between uneducated speakers of both creoles is likely to be complicated if not impossible (cf. APPENDIX I)437 , due to far-reaching differences not only in the periphery but also in the fundamental parts of the content vocabulary. In line with e.g. Martinus (1996) and Quint (2000b), I believe that these differences are attributable first and foremost to a process of relexification of Upper Guinea PC by and towards Spanish. Below, I will provide a preliminary discussion of the conditions under which this relexification process was carried out. More precisely, I will argue that the development of PA out of Upper Guinea PC is the result of a rapid and farreaching process of relexification which must have taken place shortly after the arrival and settlement of speakers of Upper Guinea PC on Curaçao. In (§8.2.2), I will make clear why the process of change from Upper Guinea PC to PA should indeed be labeled ‘relexification’ as opposed to ‘heavy bor-

437 Both scholars and non-scholars have claimed that mutual intelligibility between PA and CV is possible. In APPENDIX I, I discuss this claim in some detail and explain why I am skeptical about it.

320

Discussion: The development from Upper Guinea PC to Papiamentu

rowing’ and argue that this relexification process was carried out in a relatively short time span of no more than three to four decades. In §8.2.3, I will situate the hypothesized case of relexification in the wider field of contact linguistics and consider the theoretical advantages of analyzing PA as a so-called mixed language. Finally, in §8.2.4, I will tentatively argue that the actual process of relexification was carried out in a relatively small community consisting of Spanish-speaking masters (either Dutch or Jewish) and Upper Guinea PC-speaking slaves. First, however, it is necessary to account for the fact that the hypothesis that PA results from the relexification of an originally Portuguese-based variety – a quite popular school of thought up until the 1970s (cf. §1.2 & §1.6.1) – has received only little attention in recent PA-related scholarship. 8.2.1.

PA, monogenesis, and the notion of relexification in creole studies

In the review of the literature (chapter 1), it was noted that a number of scholars from Lenz (1928) to e.g. Kramer (2004) have drawn attention to the fact that Portuguese-derived function words predominate in PA’s grammar and that the Spanish-derived lexical material, although quantitatively superior, is largely confined to the purely lexical part of the vocabulary and peripheral function words. In light of the relative stability of function words vis-à-vis content words, this particular distribution of Spanish- versus Portuguese-derived material in PA’s grammar suggests quite clearly that the Spanish vocabulary was embedded later into an already existing Portuguese-based morphosyntactic framework, which, in turn, lends support to the hypothesis that a ready-made Portuguese-based creole was imported into Curaçao and subsequently shifted its lexicon towards Spanish. Furthermore, this hypothesis also provides an answer to what Kramer (2004: 136) refers to as Curaçao’s ‘Paradoxon’, i.e. the fact that a Dutch creole never emerged on Curaçao in spite of the fact that the Dutch were at all times the dominant slave keeping power on the island and although Dutch creoles did in fact emerge in other colonies settled by the Dutch.438 In sum, the hypothesis that PA resulted from the relexification towards Spanish of an imported Afro-Portuguese variety provides a plausible and straightforward account for the particular distribution of Portuguese- and Spanish-derived 438 In Dutch Guyana, the creoles Skepi Dutch and Berbice Dutch emerged. The two are mutually unintelligible, suggesting independent processes of creolization. In the Virgin Islands, Negerhollands emerged (Holm 2000: xviii). The number of Dutchbased creoles that have emerged in the Caribbean thereby actually exceeds the number of Spanish-based creoles (arguably only PLQ) in that region. Unfortunately, all three Dutch-based creoles are by now extinct.

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

321

material as well as for the fact that an Iberian- rather than Dutch-based creole is spoken on Curaçao. These are of course two reasons why, as outlined in the review of the literature (§1.2), the supporters of the once popular model of creole monogenesis quite consistently put forward PA as a prime example of the diffusion and subsequent relexification of the alleged West-African protopidgin/creole in the New World. One may now legitimately wonder why, with the exception of Martinus (1996) and Quint (2000b), PA’sAfro-Portuguese relexification hypothesis – plausible from a (socio-)linguistic point of view and fervently adhered to in the 1950s, 60s and 70s – has hardly been taken seriously in recent scholarship. An important part of the answer lies, I believe, exactly in this firm association between PA’s Afro-Portuguese relexification scheme and conventional monogenesis. After all, as was pointed out in §1.2, the popularity of monogenesis drastically plummeted from the 1980s onwards due to the fact that other, more satisfactory and less far-fetched439 , theoretical models emerged to explain the similarities among creoles world-wide. It may not be a coincidence that, since the 1980s, the idea of seeing PA as a relexified offshoot of an Afro-Portuguese (proto-)variety is no longer subscribed to either. In other words, it seems as if, when monogenesis lost popularity and credibility, it dragged PA’s relexification hypothesis along in its downfall; though scholars traditionally agreed that PA is a plausible candidate for a relexification scenario, its firm inclusion in conventional monogenesis prevented such a scenario from being more widely studied in recent scholarship. In the 1990s, the term ‘relexification’ was shortly revived by the extreme Relexification Hypothesis (dubbed the ‘Strict Relexification Hypothesis’ by DeGraff 2002: 323), as proposed by Lumsden (e.g. 1991) and defended particularly by Lefebvre (1986 and beyond). This polygenetic model for creole genesis views creole languages as the complete (or at least extreme) relexification of one of their African substrate languages by the dominant European language. According to this hypothesis, the extreme relexification of Fongbe by French resulted in Haitian Creole. Todd (2004: 23) strikingly describes the extreme Relexification Hypothesis as the “rewriting [of] the relexification theory, making an African substratum rather than a Portuguese pidgin the syntactic base of today’s pidgins and creoles”. As far as I am aware, both monogenesis and the extreme Relexification Hypothesis enjoy little support if any among creolists nowadays. Consider now the following comment by Todd (2004: 23, emphasis in original) in describing the demise of monogenetic relexification schemes for 439 Traditional monogenesis went as far as to propose that relexification of a Portuguesebased proto-variety had given birth to Afrikaans (cf. Thomason & Kaufman 1988: 251).

322

Discussion: The development from Upper Guinea PC to Papiamentu

Caribbean creoles: “Paradoxically, the relexification theory would probably be more acceptable if it had been less successful; if it had, in other words, left larger traces of the original pidgin, particularly in terms of core vocabulary.” While Todd is spot on for most Caribbean creoles in as far as these show no or hardly any tangible traces of a Portuguese-based pidgin, chapters 2–5 of the present study showed that PA in fact continues to offer quite straightforward linguistic arguments in support of a scenario of relexification of an Afro-Portuguese variety, though not of an ill-defined pidgin, but, indeed, of an early, full-fledged variety of Upper Guinea PC. The account of PA’s origins offered in the present study thus coincides with monogenesis in that both accounts rely on the notion of relexification. In view of the above, however, it seems necessary to stress the various ways in which the present account differs from monogenesis. Firstly, the scheme of diffusion and genetic relatedness proposed here exclusively entails the path from Upper Guinea/Upper Guinea PC to Curaçao/PA (rather than from the wider West African coast to all corners of the world). Secondly, the hypothesis postulated here does not, like monogenesis, depend on the alleged existence and transoceanic diffusion of a proto-pidgin or jargon (cf. §1.6.1.1), but instead postulates that a full-fledged natively spoken variety of Upper Guinea PC was transferred to Curaçao. Thus, rather than relying on a largely undocumented proto-pidgin, we were able to compare PA with what seems to be its sister creole in all areas of grammar. The results of this comparison, in turn, provide us with the tools to build a well-balanced relexification hypothesis, one that is supported by linguistic and historical data and that fits in with acknowledged cases of language change and far-reaching lexical shift. Below, I will explain why the development of PA out of Upper Guinea PC warrants the lable ‘relexification’ rather than ‘heavy lexical borrowing’. 8.2.2.

From Upper Guinea PC to PA: ‘relexification’ rather than ‘heavy borrowing’

As noted by Bakker (1997: 23), the term ‘relexification’ occasionally appears in the literature “as an alternative for lexical borrowing, especially when this is a massive process”. In the current field of contact linguistics, however, there is a need to distinguish clearly between the two processes in order to better understand and analyze the development of languages such as Media Lengua (Muysken 1981), Michif (Bakker 1997), Saramaccan (Good 2009) and others that clearly defy the laws of ordinary lexical borrowing (cf. Tadmor 2009: 61). I will distinguish here as well, as I believe such a distinction is necessary also to understand the development of PA out of Upper Guinea PC.

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

323

Below, I discuss two arguments why this development does indeed qualify as a case of relexification by (or towards), rather than heavy borrowing from, Spanish. The first argument is that the degree of lexical shift was much more far-reaching than the degree witnessed in normal cases of heavy borrowing (§8.2.2.1); the second argument is that this shift occurred abruptly, rather than gradual (§8.2.2.2). 8.2.2.1. Replacement of basic content vocabulary Lexical borrowing, even when heavy or even massive, typically does not affect the most basic parts of the vocabulary: “there are definite bounds to borrowing, since it tends to cluster in nonfundamental vocabulary and makes only rare and sporadic inroads into basic vocabulary” (Greenberg [1971]2005: 16, 17). By contrast, there are no such constraints on relexification, which by definition involves the substitution of native lexical items in all areas of the lexicon including, and prominently, in the most fundamental parts of the vocabulary (e.g. Tadmor 2009:59). Accordingly, French influence on the English lexicon, for instance, does not qualify as relexification: though English borrowed heavily from French, Thomason & Kaufman (1988: 308) note that this borrowing process was “no more extreme than the kinds found in many other normal cases in history”. Bennett & Smithers (1968) (cited in Thomason & Kaufman 1988: 314) had already drawn attention to the fact that “one can compose a piece of English conversation without using a single French word; and one can often hear or take part in conversations containing very few”. Furthermore, the number of French loans in English is inversely proportional to the basicness of the vocabulary, i.e. the more basic the vocabulary, the smaller the number of French loans (cf. Bakker 2003: 121; Grant 2009: 375). For instance, the number of French loans on the English Swadesh-100 list (still commonly used as a yard stick for identifying basic vocabulary) is estimated at below 10% (Thomason & Kaufman 1988: 329; Parkvall 2006: 325). Similar observations apply to the cases of Italian borrowings in Maltese, Spanish in Chamorro (see Stolz 2003 for both), German in Danish (Bakker 2003: 120) or Arabic in Swahili (Schadeberg 2009: 88): all these recipient languages show a much greater percentage of borrowings in the periphery of the lexicon than in the basic vocabulary, so that the term ‘relexification’ is not called for (Bakker 2003: 121). The Spanish contribution to the PA lexicon is quite a different matter all together and differs significantly from the above mentioned cases. By this, I am not referring to the fact that Spanish-derived items predominate in the periphery of PA’s lexicon or that a number of adverbs and peripheral prepositions and

324

Discussion: The development from Upper Guinea PC to Papiamentu

conjunctions are derived from Spanish; though that is a fact, that could in principle all be the result of ‘heavy borrowing’. The lable ‘relexification’ is rather justified by the fact that the most essential part of the original basic content vocabulary of Upper Guinea PC has been replaced by Spanish-derived lexemes. Unfortunately, unlike in the case of French influence on English, a percentage count of the Spanish contribution to PA’s Swadesh 100 list (see APPENDIX III) is bound to be inaccurate, due to the fact that roughly half of the items on it can be attributed to either Spanish or Upper Guinea PC with equal likelihood (including a few loanblends). Still, however, a look at the remaining 50% of PA’s Swadesh 100 list is quite revealing: while only 10% is native (i.e. inherited from Upper Guinea PC), no less than 40% is clearly borrowed (i.e. Spanishderived). Among that 40% borrowed share, we find the PA words for the most basic concepts, such as man, woman, child, person, bird, tree, skin, blood, bone, hair, tooth, foot, hand, breast, to see, moon, star, water, stone, night, warm and name, all of which are of probable Spanish origin and etymologically unrelated to the corresponding lexemes of Upper Guinea PC.440 As a consequence, conversing in PA without taking recourse to borrowed Spanish lexical items is pretty much impossible and it is why so many linguists have classified and continue to classify PA as a Spanish-lexified creole without hesitation. Also, as noted, in heavy borrowing languages such as English, Chamorro or Maltese, the number of loanwords is high in the peripheral vocabulary, but diminishes as one approaches the basic vocabulary. This is clearly not the case here: Spanish loans are predominant both in the periphery and in the fundamental part of the PA lexicon and, to my knowledge, even languages such as Romanian or Swahili do not display the borrowing of such an extensive subset of the most basic vocabulary items.441 In sum, the amplitude of the borrowed Spanish component in PA is such that it warrants the label ‘relexification’ rather than ‘heavy borrowing’.442 440 To be complete, two items on PA’s Swadesh 100 list are derived from Dutch, and two others can be either English or Dutch loans. 441 This can be checked quite easily via the splendid online World Loanword Database (WOLD), http://wold.livingsources.org, referenced under Haspelmath & Tadmor, eds. (2009). 442 Note that the hypothesized lexical shift of Upper Guinea PC towards Spanish, besides the actual relexification (i.e. the replacement of basic and native content vocabulary), inevitably also involved regular lexical borrowing with the mere purpose of vocabulary expansion, but since we do not know the original size of the vocabulary of the Upper Guinea PC variety that arrived on Curaçao, it is impossible to estimate the relative proportions of relexified versus borrowed vocabulary in that early period.

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

325

One could still argue that, if the pressure from Spanish was sufficient, the three centuries separating PA and Upper Guinea PC would have sufficed for PA’s basic content lexicon to be replaced through a ‘normal’ process of heavy lexical borrowing, say, scale 5 on Thomason & Kaufman’s (1988) borrowing scale. However, if this were the case, we would have expected the early variety of Upper Guinea PC that arrived on Curaçao to not only replace its basic content vocabulary, but to also substitute “major structural features that cause significant typological disruption” (Thomason & Kaufman 1988:75). As chapters 2–6 of this monograph have shown, however, this has not happened in the transition from Upper Guinea PC to PA. Moreover, for scale-5-borrowing to occur, strong cultural-linguistic pressure over a prolonged period of time is required (Thomason & Kaufman 1988: 94). The available historical and linguistic clues, however, suggest that the lexical shift towards Spanish had already been completed by the turn of the 18th century, on which more in the next subsection. 8.2.2.2. Abrupt lexical shift Another property that crucially distinguishes relexification in the strict sense from more normal processes of lexical borrowing is the rate at which the lexical shift occurs. While relexification is considered to be an abrupt process that may take no more than a few decades (more on this in §8.2.3), processes of (heavy) lexical borrowing (e.g. the massive integration of Chinese loans into Classical Japanese [cf. e.g. Schmidt 2009: 549] or of French into English [cf. e.g. Thomason & Kaufman 1988: 326]) will usually take much longer (cf. also Golovko [2003: 191] and Haspelmath [2009:41] on the gradualness of regular lexical borrowing). As shown below, also in this respect, the development of PA out of Upper Guinea PC is classifiable as a case of relexification by, rather than as heavy borrowing from, Spanish. In order to understand the time span in which the original Upper Guinea PC lexicon was relexified towards Spanish to form PA, one should recall, firstly, that the first Upper Guinean slaves probably arrived on Curaçao in 1659 (and stopped arriving after ca. 1680). We may thus take 1659 as a terminus post quem. Secondly, as discussed in §1.5.1, the 1775 PA letter and earliest available written PA document shows roughly the same quantity of Spanish-derived lexical material as modern-day PA does. Clearly, this suggests that the most important lexical shift towards Spanish had already been completed by that time, which leaves us with a time span of some 115 years or less for the relexification towards Spanish to have occurred. A second historical-linguistic clue, however, allows us to significantly reduce the terminus ante quem: as early as 1704, the polyglot Father Alexius Schabel reported in his travel account that what the slaves of Curaçao

326

Discussion: The development from Upper Guinea PC to Papiamentu

spoke sounded much like ‘broken Spanish’ (cf. §1.1). Provided that the Father’s assessment is reliable (i.e. that he was capable of distinguishing Spanish from Portuguese, more on this in §8.2.3), we may assume that a far-reaching lexical shift towards Spanish had already been accomplished by the turn of the 18th century. If these calculations are correct, we are left with a remarkably short space of time of some four and a half decades (1659–1704) or less for this lexical shift to have been carried out. This provides another argument in favor of labeling the referred lexical shift as ‘relexification’ rather than as ‘(heavy) lexical borrowing’, which as noted, occurs at a much slower rate. (In the next section, §8.2.3, I will show in more detail that this short time interval is indeed perfectly reconcilable with recognized cases of drastic language change including relexification.) Based on the above, we can now make a rough time-based division between the earliest, most decisive lexical shift towards Spanish – the actual process of relexification, completed by the turn of the 18th century – and all changes that occurred afterwards up to present. Note that a similar assessment of PA’s development was posited by Holm (1988: 315): “Papiamentu’s historical movement towards Spanish has included its early relexification and lexical expansion as well as later structural borrowing”. This division implies, for instance, that the substitution in the course of the 19th century of lexemes such as Early PA fika by modern PA keda (§5.7.2.3) or of the suffix -dadi by -dat (§4.1.4) were not the result of the early process of relexification by/towards Spanish, but rather fall into the category of regular borrowing from Spanish. The same is true for the significant contributions made to PA by Kwa and Bantu languages and Dutch as well as, to a lesser extent, English. These contributions, too, fall under the heading of ‘borrowing’ rather than ‘relexification’. The latter term, then, is reserved for the early, rapid and far-reaching process of lexical shift of the original Portuguese-based lexicon towards Spanish, a process which, I assume, was activated and completed in the period between 1660 and 1700, and primarily involved two languages: Upper Guinea PC and Spanish. A word should still be said with regard to basic functional items such as PA a (PFV), nan (3pl = PL) and no (NEG), i.e. those features which most saliently separate modern PA from Upper Guinea PC and which were in all likelihood not present in the ancestor variety that arrived on Curaçao. Though the very concept of ‘relexification’ suggests a process of change that affects only the purely lexical part of the vocabulary, we may assume that in the period in which the relexification was carried out, the language must have been particularly unstable and hence receptive to influences on all levels of the grammar. The actual relexification may thus have been accompanied by the incorporation of structural features from Spanish such as PA a, no < Sp. ha, no replacing original Upper Guinea PC Ø and ka as well as from other languages present at the time

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

327

(e.g. PA nan < Gulf of Guinea PC nan). In any case, the 1775 PA love letter and other Early PA texts suggest that these features were already thoroughly integrated in the language in the latter stages of the 18th century.

8.2.3. Analyzing Papiamentu as a mixed language To review PA within the broader field of contact linguistics offers several descriptive advantages. Most importantly, it allows us to rehabilitate the notion of relexification by emancipating it from the field of creole studies: whereas relexification is a divisive and rather unpopular concept in creole studies – owing to its association with monogenesis as well as Lefebvre’s extreme Relexification Hypothesis (cf. §8.2.1) –, it is in fact a well-defined and by no means rare concept within contact linguistics in general and in the mixed language debate in particular. Indeed, if our relexification scheme proposed for the development of PA – i.e. rapid relexification of Upper Guinea PC towards Spanish affecting the basic content vocabulary but leaving intact the original morphosyntax – is correct, this brings up several interesting parallels with so-called mixed (or intertwined) languages. In spite of some controversy as to which criteria define a ‘true’ mixed language (cf. e.g. Bakker & Mous, eds. 1994; Thomason, ed. 1997; Matras 2000; Bakker & Matras, eds. 2003), the classical definition is that of a language which consists of “the grammatical system of language A (…) with the lexical stock of language B” (Bakker 1997: 213). According to this definition, PA may well constitute a paradigm example of a mixed language, with a morphosyntactic matrix inherited from its Upper Guinea PC ancestor while Spanish provided the bulk of lexicon. Consider also Comrie’s (2008: 21) description of a mixed language as “one which reflects, historically, (…) the combination of different major components from different historical sources”.Again, we can translate this to the case of PA and grosso modo distinguish between two major components – a functional and a purely lexical component – and two corresponding historical sources: Upper Guinea PC and Spanish respectively. Another way of saying the same is that, just like mixed languages, PA is characterized by a functional:lexical split in its grammar. Matras (2003: 156) refers to it as “the split between INFL and the content-lexicon”, whereby the INFL-language typically supplies the verbal syntax and morphology, the rules for word order, as well as “the typology and structures of clause combining” (2003: 155). As was shown in chapters 3–6, PA indeed structurally as well as typologically overlaps with Upper Guinea PC in these domains (cf. §3.4 for clause combining, chapter 5 for the verbal system, and §6.3.1 for word order).

328

Discussion: The development from Upper Guinea PC to Papiamentu

The hypothesis that PA results from the relexification of Upper Guinea PC towards Spanish may thus be compared with acknowledged cases of relexification such as Media Lengua, a mixed language spoken in Central Ecuador with a lexicon/grammar split comparable (though not identical) to that of PA. While Media Lengua’s morphosyntactic framework is Quechua, Spanish provided 90% of the lexicon (adapted to Quechua phonology), except affixes, which are still Quechua (cf. Muysken 1981, 1997; Dikker 2008)443 . The analysis of PA as a mixed language proves to be particularly fruitful in as far as it allows us to account for the remarkably short time span of less than four decades in which Upper Guinea PC must have been relexified. As noted previously, in light of Father Schabel’s 1704 description of PA as ‘broken Spanish’, we must assume that the relexification of Upper Guinea PC towards Spanish had already been largely completed by 1704. That this is perfectly possible is shown by other cases of drastic language mixing. For instance, the formation of Media Lengua (i.e. the relexification of 90% of the original Quechua vocabulary by Spanish) must have occurred sometime between 1920 and 1940 (Muysken 1997: 374) and most likely in a time span of less than a decade (Pieter Muysken p.c.). But not only Media Lengua, also several other recognized mixed languages are thought to have emerged in remarkably short periods of time (cf. Bakker 2003: 136). Copper Island Aleut (Aleut + Russian) and Michif (Cree + French) were presumably created in the space of a single generation according to Comrie (2008: 25) and in “probably two or three decades at most” according to Thomason (2007: 55). Moreover, Gurindji Kriol, a mixed language composed of Gurindji (an Australian aboriginal language) and Kriol (an Australian Englishbased creole), developed in the 1960s-70s out of the regularization of codeswitching patterns and subsequent nativization. This process took no more than two decades (McConvell & Meakins 2005: 15). Another interesting example of rapid and intense mixing is provided by the Surinamese creole Saramaccan, which, besides being a creole, can be argued to be a mixed language consisting of two major components: early Sranan and Portuguese (cf. e.g. Bakker, Smith & Veenstra 1995: 168; Muysken 2008: 198–201; Good 2009: 926, 927). The available historical and linguistic documentation suggest that the process that gave rise to Saramaccan, i.e. the partial relexification of Sranan towards Portuguese, took slightly longer than one decade (between the late 1660s and 1680) to be carried out (Cardoso & Smith 2004: 117).

443 Media Lengua was formed within a Quechua-speaking group of labor migrants temporarily settled in the capital of Ecuador (Muysken 1997: 373–376).

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

329

It is now interesting to briefly return to the debate on PA’s Spanish versus Portuguese origins. Father Schabel’s previously mentioned description of PA as ‘broken Spanish’, in the year 1704, had been put forward by several proponents of PA’s Spanish-origin-hypotheses as a decisive argument against PA’s alleged Afro-Portuguese roots. Maduro (1965: 5), for instance, commented that, “if Father Schabel recognized a broken Spanish in the speech of our people, then that is good reason to believe that this was what our people spoke”444 . Van Wijk (1958: 169), and other supporters of PA’s Afro-Portuguese origins, could only object by arguing that the Portuguese language was probably not part of Schabel’s otherwise formidable linguistic repertoire. The evidence from other mixed languages, however, shows that thorough relexification can, under the right circumstances, be a very rapid process, which means that we need not insist (as did van Wijk) on Schabel’s failure to recognize an earlier more pronounced Portuguese character of PA, but may instead admit to the soundness of the Father’s observation without this jeopardizing PA’s hypothesized Afro-Portuguese origins. Related to the debate about Schabel’s testimony is the discussion of the 1775 PA love letter and its implications for the origins of PA. As noted previously, this letter shows no quantitative difference with respect to modern PA in terms of Spanish- versus Portuguese-derived vocabulary. According to Wood (1972a), just like Schabel’s testimony, this too was different from what one would expect if PA had really developed out of a Portuguese-based variety. The lack of additional Portuguese material in the 1775 letter, he argued, is a “decisive factor in any consideration of the theory (…) that Papiamentu is a direct descendant of a Portuguese creole which has, in comparatively recent times, been decreolized through the influence of Spanish at all linguistic levels” (1972a: 24). Again, however, the evidence from Media Lengua and other mixed languages sheds new light on the issue and suggests that the 1775 PA love letter does not in the least challenge the hypothesis that PA started out as a Portuguese-based creole. In other words, instead of analyzing Father Schabel’s testimony and the 1775 PA letter as evidence against PA’s Afro-Portuguese hypotheses, they in fact provide valuable clues as to the rate at which Upper Guinea PC shifted its original Portuguese-based core vocabulary towards Spanish: this process must have been carried out in the time span of three to four decades at most. If we elaborate further on the parallel between PA and mixed languages, some interesting sociolinguistic questions emerge for future research to deal with. For instance, it is widely assumed that the process of language mixing/intertwining 444 Original quote: “Si Pader Schabel a distinguí un spañó quibrá den habla di nos pueblo, anto tin tur motibu pa kere qu t’eséi nos pueblo tábata papia”

330

Discussion: The development from Upper Guinea PC to Papiamentu

is strongly related to the formation of a new group identity (cf. e.g. Bakker & Matras 2003: 13; Muysken 1997: 376;Thomason 2003: 32; but see McConvell & Meakins 2005: 16 for an opposing view).445 It is implied that the creation of (most) mixed languages is to an important extent a conscious process and that the need or desire to mark a new ethnic identity can be the driving force behind cases of far-reaching language restructuring. We may now speculate about the extent to which this applies to the case of PA. We know, for instance, that PA had become an important identity marker of large parts of the Curaçaoan society already in the early 18th century and that the vitality of the creole is remarkable also at present, especially compared to the endangered status of other creoles (cf. Parkvall 2003). While one may point at other socio-linguistic factors which may have contributed to this vitality (e.g. the position of PA vis-à-vis the impopular language of the colonizers, Dutch), it does correlate nicely with the assumption that the language had indeed been created more or less deliberately with the aim of forming a new ethnic group identity. Assuming a conscious process of relexification makes sense even more if we take into consideration that there must already have been some degree of mutual intelligibility between Spanish and Upper Guinea PC due to the closeness of Spanish and Portuguese. As Pieter Muysken (p.c.) observes, this intelligibility may have overridden the strictly communicative necessity for relexification. If correct, it is implied that the condition triggering the relexification was not the lack of mutual intelligibility; rather, the principal motivation behind the relexification is likely to have been the desire to mark a new group identity and/or to flag acculturation (cf. Matras 2003: 155 for further discussion). Furthermore, though to my knowledge this is not discussed in the existing literature on mixed or intertwined languages, I assume that the conscious, intentional character of language mixing is one reason why it can be carried out in such a brief space of time. If language mixing were an unconscious and unintended by-product of (intensive) language contact, one would expect this to occur, if at all, at a much slower rate. The correlation between intentionality/deliberateness and rate of lexical replacement can be illustrated by means of taboo languages. These are languages in which the use of certain words can, for different culturalreligious reasons, suddenly become forbidden obliging their speakers to replace these words by others, be it through borrowing or through coining (cf. Comrie 2000; Mous 2003: 219). Evidently, the word replacement in taboo languages is 445 E.g. Mous (2003) speaks of ‘lexical manipulation’ in order to reflect the intentional character of process of language mixing. Thomason (1999) provides a detailed discussion of how language change in general correlates with speakers’ conscious and/or deliberate choices.

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

331

a conscious and deliberate speech act (viz. an act of lexical manipulation) and, indeed, can occur at a remarkably high rate (Comrie 2000: 38). While the motivation behind language mixing/intertwining (to which more in §8.2.4) is of course different from taboo-motivated change, the two phenomena are comparable in that they concern drastic linguistic change which is both intentional/deliberate and abrupt.446 In the case of PA, the fact that the relexification was indeed abrupt can thus reasonably be taken to suggest that this process was indeed also carried out deliberately. If the relexification of Upper Guinea PC towards Spanish, from which Papiamentu resulted, was indeed a conscious process, the notion of ‘lexical reorientation’ as defined by Matras (2000) might turn out to be applicable: Lexical re-orientation is the conscious shifting of the linguistic field that is responsible for encoding meaning or conceptual representations away from the language in which linguistic interaction is normally managed, organised and processed: speakers adopt in a sense one linguistic system to express lexical meaning (…) and another to organise the relations among lexical symbols, as well as within sentences, utterances, and interaction. The result is a split, by source language, between lexicon and grammar. (Matras 2000: 82)

As these examples show, the perspective of PA as a part of the wider field of contact linguistics and particularly the parallel with mixed languages offer a great deal of new starting points for future linguistic and socio-historical research into the history of the language and the circumstances under which it developed into the mother tongue of the people of Aruba, Bonaire and Curaçao. 8.2.4. The source(s) of the Spanish elements in PA’s basic content vocabulary As noted in the review of the literature (chapter 1), the search for the origins of PA has long been equated with the search for the source(s) of the Portuguese elements in its core vocabulary. In chapters 2–6, evidence was provided in favor of the claim that this source was Upper Guinea PC. If that is correct, ironically, that brings up the equally intriguing question as to who provided PA’s Spanish-derived lexicon. In other words, who provided the linguistic input and/or target necessary for the rapid and far-reaching relexification of Upper Guinea PC towards Spanish to occur? Below, I will focus on arguably the two most plausible candidates – the Dutch and the Sephardim – and discuss some other factors of possible relevance. 446 Of course, this is not to say that all intentional change will be abrupt, but rather that, in the absence of other factors, intentional change is more likely to be abrupt than unintentional change.

332

Discussion: The development from Upper Guinea PC to Papiamentu

Since, for quite obvious reasons, mixed languages are thought to typically emerge in bilingual communities (Bakker & Matras 2003: 1), we should search for a community of speakers bilingual in Spanish and Upper Guinea PC on Curaçao in the relevant period. We may exclude from this search the (presumably Spanish-speaking) Amerindians, regardless of their actual demographic relevance, since contact between the African slaves and Amerindians is thought to have been minimal in the early period. And even if there was contact, there is no reason to believe that Upper Guinea PC-speaking slaves would have drastically adapted their speech habits under the influence of these Amerindians, who would simply have lacked the sufficient degree of social prestige. With the exception of the Amerindians, pockets of native Spanish speakers on Curaçao are not officially documented during the second half of the 17th century. However, it is uncontroversial that the Spanish language enjoyed a great amount of both economic and social prestige on 17th-century Curaçao, whose growing society depended significantly on trade and cultural (including religious) contacts with the Venezuelan mainland. We may furthermore refer the reader to arguments provided in §1.5.1 suggesting that many of the Sephardic Jews who settled on Curaçao must have been fluent in Spanish. If Spanish was not already their home language, it is not hard to imagine them shifting to Spanish upon settlement on Curaçao. Thus, we may tentatively surmise that PA developed among the (offspring of) Upper Guinea PC-speaking slaves and Jewish masters with an L1 competence in Spanish. In this light, it is revealing to discuss a receipt for the payment of 200 pesos of ransom for the male slave Donqui. This document, from the year 1736, was brought to my attention by Han Jordaan, who retrieved it from the Dutch National Archives in The Hague. The receipt was meant to settle a financial dispute between the free colored woman Anna Maria Koningh and her former slave Donqui. The document is written in Spanish and was presumably drawn up by Nicolaas Henricus, who may very well have been part of the Sephardic Henriquez family. Han Jordaan noted that, given the underhand nature of the case, it is safe to assume that the document was drawn up in a language spoken both by the slave and by the mistress. It is furthermore known that Anna Maria Koningh had been the lover of a Sephardic Jew.447 Although this event postdates the period in which the relexification was presumably carried out, it allows us to hypothesize (a) that intimate relationships between Sephardic Jews and their slaves were all but seldom and (b) that both parties knew and used Spanish, probably in addition to Upper Guinea PC / PA. 447 I am indebted to Han Jordaan for sharing this piece of historical documentation and for providing the valuable background information.

8.2. From Upper Guinea PC to PA: a case of rapid relexification towards Spanish

333

Another possible source for the Spanish is the Dutch WIC officials, who were also the principal slave owners. Though we know little with certainty about their actual Spanish competence, it is historically documented that some of them, especially those occupying the higher ranks within the WIC, knew Spanish fluently (Kramer 2004: 135). One can imagine that these Dutch slave owners would have seen the benefits of speaking Spanish with and to their slaves. Note, furthermore, that some of the higher ranked WIC officials were in fact Spanish Catholics (Kramer 2004: 124). If we compare the Dutch with the Jewish scenario, arguably the latter seems more logical linguistically, given that the Sephardim must, in general, have been significantly more competent in Spanish than the Dutch. From a demographic point of view, on the other hand, one could argue that the scenario involving the Dutch slave owners is more plausible, in light of the fact that they owned a far greater number of slaves than the Sephardim did (Maurer 1998: 199). This potential demographic argument against relexification among the Sephardim and their slaves is however not as sound as it seems. In fact, a relatively small mixed bilingual community with speakers of Spanish and Upper Guinea PC may in principle have sufficed, or even have been favorable, for rapid relexification to take place: as Comrie (2000: 37) affirms, “in the absence of other factors, one would expect more rapid change the smaller the community involved”. The reason for this, he explains, is that, “[i]n a small community, innovations introduced by one individual can rapidly become known to all members of the community; the chance of some other member of the community rejecting the innovation is relatively low because the number of other members of the community is low” (Comrie 2000: 36; cf. Thomason 2003: 32; Bowern [2010]448 provides a detailed discussion of the correlation between population size and rates of linguistic change). Accordingly, it is unsurprising to learn that mixed languages indeed typically arise in relatively small communities of perhaps only a few hundred speakers (cf. e.g. Golovko 2003: 198). In this context, then, we may hypothesize that the relexification of Upper Guinea PC by Spanish was carried out and brought to completion in a small but influential bilingual speaker community (e.g. one of the agricultural farms in the interior449 ), small enough for the relexification to be rapid and influential enough to subsequently spread their speech to newly incoming slaves and other segments of the society.

448 I thank Peter Bakker for bringing this publication to my attention and am indebted to Claire Bowern for generously sharing a copy of it. 449 Note that, while sugar plantations did not exist on Curaçao, the larger agricultural farms in the interior sometimes harbored up to ca. 400 slaves.

334

Discussion: The development from Upper Guinea PC to Papiamentu

Finally, it is relevant to point out that full bilingualism is not always required for a community language to undergo far-reaching change. As is well known, prestige offers crucial (and sometimes sufficient) motivation for speakers to carry out and/or accept far-reaching linguistic changes towards a target language, even if the direct access to that target is limited. Haspelmath (2009: 48), for instance, in discussing loanword integration, observes that “‘prestige’ is a factor that is very difficult to measure independently (…). However, it seems to me undeniable that prestige is a factor with paramount importance for language change, going far beyond our current topic of loanwords”. Parkvall & Goyette (forthcoming) provide a case in point when pointing out that the cultural prestige of Latin in the post-Renaissance era was so high as to trigger numerous borrowings of words and productive morphemes from Latin into several IndoEuropean languages, in spite of the fact that only very small percentages of the recipient speaker populations had knowledge of Latin (cf. e.g. the particular case of French influence on Romanian discussed in Schulte [2009: 237, 238]). Another clarifying and in the present context particularly relevant case is discussed by Thomason (2001) when describing “the unnamed mixed language spoken in Ternate in North Maluku (Indonesia)” (2001: 213), which is a mixture of Ternatean (lexicon) and Malay (morphosyntax). This mixed language arose abruptly and along the following lines: [Malay] is the children’s first and only language until they reach an age when they find themselves in social situations where Ternatean is the appropriate language. They (…) do not bother learning Ternatean grammar, but instead simply attach Ternatean lexicon onto Malay grammar so that they can imagine that they are speaking Ternatean. (Thomason 2001: 214)450

Translated to the present case, one can easily imagine a scenario in which the members of a community of Upper Guinea PC speakers realize that Spanish is the appropriate language in certain social settings on Curaçao, and, instead of fully switching to Spanish, commence attaching Spanish lexicon onto their native grammar, so as to pretend to speak Spanish, viz. to flag acculturation (cf. Matras 2003: 155). Like Ternatean in the example given above, Spanish was doubtlessly the most appropriate language in certain social situations on Curaçao. Particularly in the trade with the Venezuelan mainland and other Spanish American colonies, on 450 Tadmor’s (2007) account of the core vocabulary borrowings from Malay into the Austroasiatic language Ceq Qong is quite comparable. Haspelmath (2009: 46, f.n.) summarizes this account as follows: “Speakers tried to assimilate to the strongly dominant Malay people, but had very little access to the Malay language, so they borrowed what they could, the basic vocabulary that they knew.”

8.3. Summary of the discussion

335

which the rapidly growing society of 17th-century Curaçao depended increasingly, a solid command of Spanish was advantageous, if not indispensable. But the Curaçaoan slaves also came in close contact with Spanish in the domain of religion. Although, unfortunately, quite little is known about the religious habits of the slaves, we do know that already in the late 17th century Spanish-speaking Catholic priests would occasionally come over from the Venezuelan mainland to perform a variety of clerical services and to christen the Curaçaoan slaves as Catholics (Hartog 1968: 145–154; Munteanu 1996: 85, 86). These Catholic missions targeted not only the transit slaves (i.e. those destined for resale), but notably also those residing on Curaçao (Hartog 1968: 152). The Catholic priests were given freedom to operate on Curaçao, firstly, because this was included in the terms of the asiento contract451 obtained by the Dutch WIC and, secondly, because Curaçao’s Protestant Dutch and Sephardic Jewish elites did not want their slaves to be members of their church (Hartog 1968: 148). Early on, as a consequence of the Catholic missions, Curaçao’s slave society had indeed become predominantly Catholic, rather than Protestant or Jewish (Hartog 1968: 178). Thus, the fact that Spanish was heavily associated with the slaves’religious instruction may have granted it the necessary prestige and accessibility necessary to initiate far-reaching relexification of the basic vocabulary. On the other hand, as noted, the interaction between the Catholic, Spanish-speaking priests and the Curaçaoan slaves in the early period has remained scarcely documented and warrants future research.

8.3. Summary of the discussion This chapter has attempted to relate the linguistic and historical evidences from chapters 2–6 and embed these in a framework that accounts for the actual development of PA out of its Upper Guinea PC ancestor. In §8.1, some of the sociolinguistic factors that may have been at play in facilitating the arrival and survival of Upper Guinea PC on Curaçao as well as its subsequent consolidation and spread among the emerging slave society have been discussed. First (§8.1.1), I discussed why, out of an impressive total number of slaves brought to Curaçao from different regions mostly for resale to third parties in the Caribbean, those that spoke Upper Guinea PC were preferred for keeping on the island. I argued that this preference for Upper Guinea PC-speaking slaves was 451 The so-called asientos were contracts or trading agreements that would give their owner the exclusive right to trade with Spain in a specific region.

336

Discussion: The development from Upper Guinea PC to Papiamentu

due to (a) their ability to adapt to the social climate of a slave society, (b) their being lighter-skinned than other slaves, (c) the communicative advantages offered by their mother tongue, and (d) their being properly christianized. Secondly (§8.1.2), I hypothesized about the favorable conditions that enabled a small-sized Upper Guinea PC speaker community to retain their speech in a multilingual setting and pass it on to newcomers and other segments of the Curaçaoan society. The most important of these conditions, I believe, were the following: – The Upper Guinea PC variety brought to Curaçao was a full-fledged creole (rather than an ill-defined pidgin or trade jargon) transmitted by native speakers; – Speakers of Upper Guinea PC constituted the Founding Population of the Curaçaoan slave society, putting them in the position to pass on their mother tongue to newcomers; – On the whole, slaves from the Cape Verde region were lighter-skinned than slaves from other West African regions. Lighter-skinned slaves, in turn, enjoyed greater social prestige on Curaçao; – Speaking a creole with a Portuguese-based lexicon allowed for at least basic communication with incoming slaves on the one hand and the Dutch and Jewish masters on the other; Needless to say, these sociolinguistic factors are likely to have interacted and reinforced one another at all times. In §8.2, I presented a scenario in which contact between Upper Guinea PC and Spanish led to the relexification of the former towards the latter in the space of time of perhaps no more than three to four decades. This process was not only rapid but also thorough, affecting a large portion of the most basic content vocabulary, while largely leaving intact the inherited functional categories. I argued that, if this scenario is correct, PA can be analyzed as a mixed language with a lexical/functional split in the grammar, which allows for fruitful parallels to be drawn with acknowledged cases of language intertwining. While the discussion on the origins of PA has thus far been dominated by the quest for the source(s) of its Portuguese elements, §8.2 closed with a search for the historical source(s) of the Spanish elements in its core vocabulary, speculating that either the Dutch or, arguably more likely, the Sephardim offered the target for relexification or that, as a third possibility, Spanish was accessible to the slaves by way of Spanish-speaking Catholic priests. Clearly, future research is required to tackle this issue more thoroughly, although the relative scarcity of historical documentation will likely complicate such future endeavors.

Chapter 9 Conclusions

The aim of this study was to examine and provide evidence for the hypothesis that PA, the language of Aruba, Bonaire and Curaçao, and Upper Guinea PC, the language of the Cape Verde Islands, Guinea-Bissau and the Senegalese province of Casamance, are genetically related. While the ties between these languages have been studied in some detail by Martinus (1996) and Quint (2000b), the conclusions reached by these scholars (and therefore the claim of kinship between PA and Upper Guinea PC) have received little to no support in the recent body of scholarship concerned with PA. The present study drew on these two pioneering studies, elaborated on them, and provided new linguistic and historical evidence in favor of a genetic and historical relationship between PA and Upper Guinea PC. In the linguistic chapters (2–6) of this study, it was shown that the phonology, function words, morphology and verbal system of PA significantly overlap with Upper Guinea PC. The correspondences were shown to be structural and/or idiosyncratic to such an extent that chance could not satisfactorily account for them. The historical chapter 7 showed that, indeed, we need not appeal to chance, as the historical connections between Upper Guinea and Curaçao were such that they provided the conditions for the linguistic transfer from one region to the other to occur in the latter half of the 17th century. Chapter 8 discussed and analyzed, firstly, the sociolinguistic and socio-historical circumstances that I believe contributed to the survival and consolidation of Upper Guinea PC in the multilingual setting found on Curaçao in the latter half of the 17th century and, secondly, the factors at play in the relexification process that facilitated the original language to develop into PA. The controversial debate on the origins of PA has been dominated by the quest for the source(s) of the Portuguese materials in the core of its grammar. As early as 1928, Rodolfo Lenz in his excellent study of PA had already claimed that this grammar is principally Afro-Portuguese, surrounded by a predominantly Spanish-based lexicon. The present study has corroborated his findings and has furthermore identified the source of the Afro-Portuguese materials as being Upper Guinea PC. I accounted for the high percentage of Spanish-derived lexemes in the core of PA’s vocabulary by postulating a rapid and thorough relexification process which was completed in the period of no more than three to four decades.

338

Conclusions

If this reconstruction of events is accurate, more than three centuries separate modern-day PA from Upper Guinea PC, a time span in which very different contact situations have led these creoles to grow apart considerably. Nonetheless, clear traces of their shared ancestry remain in the morphosyntax and phonology. In fact, the correspondences between PA and Upper Guinea PC at present are still of such a nature that they warrant the synchronic subgrouping of PA and Upper Guinea PC as a separate branch of creoles within the wider family of Iberian-based restructured varieties. One of the implications of recognizing their shared ancestry is that it will facilitate the reconstruction of Upper Guinea PC as spoken in the mid-17th century, i.e. prior to PA’s separation. This may prove useful not only for the historical and linguistic study of PA and Upper Guinea PC, but eventually also, I assume, for testing hypotheses of creole genesis. A second implication is that Curaçao was not the scene of creolization, but of language transfer and a subsequent process of rapid and far-reaching lexical replacement. Thus, representing what John Holm (p.c.) aptly refers to as a case of ‘language contact reduplicated’, PA can offer new insights into two unique types of extreme language contact. Cases of creole transfer are of course not unique. Parkvall (1995: 41), for instance, lists several “acknowledged cases of Creole languages having been introduced from the territory in which they originated into another. Examples include Seychellois (introduced from Mauritius), Principense (from São Tomé), and Aku (of the Gambia; from Krio)”. Also recognized is the historical spread of varieties of Asian Portuguese creole to Ceylon, Batavia and Malacca (Holm 1988: 288–294). Nevertheless, the case of PA is unique in that it shows that, under the right circumstances, the transfer of a West African creole across the Atlantic into the Caribbean area was possible. This possibility has long been, and continues to be, debated in the context of monogenesis as well as afrogenesis and with regards not only to Portuguese-, but also to French- and English-based creoles (cf. e.g. the contributions to Huber & Parkvall, eds. 1999; McWhorter 2000). It is important to stress, finally, that the hard evidence for a link between PA and Upper Guinea PC is primarily provided by the linguistic data. The primary aim of the historical chapter (7) and the discussion in chapter 8 was to offer a sound (socio-)historical framework in which these linguistic data could be placed and accounted for.

Appendix I

The myth of mutual intelligibility Although the kinship between PA and Upper Guinea PC is yet to receive (wider) recognition among linguists, their resemblance has not gone unnoticed among native speakers. Below, I provide a survey of testimonies in which mutual intelligibility between speakers of the two creoles is claimed and explain why I remain skeptical of such claims. According to Efraim Frank Martinus, linguist, native PA speaker, and strong proponent of kinship between PA and Upper Guinea PC, “the only barrier for a PA speaker to communicate directly with a Cape Verdean is the fact that CV uses ka instead of no”452 (Martinus 2007: 7). In an interview (de Wit 1997), Martinus explains how in 1953 he traveled to Holland from Curaçao, stopped over on the Cape Verde Islands and, having disembarked, was stunned to find himself understanding the language of the local Cape Verdean people. In The Associated Press (02-04-1985), furthermore, he claims that “even today the Caribbean’s Papiamento speakers can talk to Cape Verde people”. The claim of mutual intelligibility is also echoed firmly in this statement by the linguist Macedo (1979: 26): “As a native Capeverdean speaker, I have had no difficulty understanding both the written and the oral forms of Papiamento” (cf. also Peck 1988: 91). In a similar vein, Eva Martha Eckkrammer, a trained speaker of Papiamentu, asserts: Personal experiences with speakers of Portuguese Creoles living in Portugal, for instance from the Cape Verde Islands or Guinea-Bissau, brought to the fore striking parallels between Papiamentu and these West African creoles. It can without a doubt be claimed that conversation is possible, when one speaker uses PA, the other CV or GBC453 (Eckkrammer 1996: 59)

452 Original quote: “e úniko barera ku un papiadó di papiamentu tin pa komuniká immediatamente ku un kaboberdiano ta e echo ku kabobèrdiano ta uza ka pa no” 453 Original quote: “Persönliche Erfahrungen mit in Portugal lebenden Sprechern portugiesischer Kreolsprachen, z.B. der Kapverdischen Inseln und Guiné Bissaus, brachten eklatante Parallelen zwischen dem Papiamentu und diesen westafrikanischen Kreolsprachen zutage. Es kann ohne weiteres behauptet werden, daß Konversation möglich ist, wenn ein Sprecher Papiamentu, der andere Crioulo do Cabo Verde oder das Kreolische Guiné Bissaus verwendet”

340

Appendices

Quite recently, the CV specialist Pereira (2006: 78) gave an indication of the scope of these rumors of mutual intelligibility: Drawing on Maduro’s (1987a) juxtaposition of a PA and CV text, she notes: “A quick perusal of the following texts in CV and PA respectively gives us an immediate impression of identity which has aroused intense curiosity not only among scholars but also among the speakers themselves”454 . Ironically, however, as noted in §1.3, the intention of Maduro’s (1987a) juxtaposition was to illustrate that “an inhabitant of Cape Verde and one of our Leeward Islands will never be able to understand a single bit of what they say to each other”455 (1987a: 22). The Dutch city of Rotterdam would be interesting ground to test some of the assumptions regarding (the lack of) mutual intelligibility. This harbor is home to both a CV- and a PA-speaking community and there is said to be a good deal of contact between these groups. Here too, rumor has it that the PA speakers “have discovered a measure of mutual intelligibility with Cape Verdean Creole speakers residing there” (Fouse 2002: 95) and Linda Rupert (p.c.), a historian of the Dutch Antilles, was also confronted with rumors about Cape Verdeans in Rotterdam, “who find that they can communicate easily with the Antilleans there”. The firm claims of mutual intelligibility such as Martinus’ and Macedo’s are of course fascinating in the context of the kinship between PA and Upper Guinea PC and clearly warrant future research. Until that happens, I remain skeptical of these claims, though. In fact, I believe mutual intelligibility between uneducated viz. monolingual speakers of Papiamentu and of Upper Guinea PC to be rather unlikely, which is due not only to the different negator, but also to the overall differences in the (basic and non-basic) content vocabulary. I assume that only if a speaker of Upper Guinea PC has additional knowledge of Spanish and Dutch, or, vice-versa, a speaker of Papiamentu additional knowledge of Portuguese, could they perhaps converse more freely with one another. For the claims made in the present study, however, the (lack of) mutual intelligibility between PA and Upper Guinea PC is not crucial: their being mutually intelligible or not has no bearing on the hypothesis that they are genetically related, since, as I hope to have shown in the present study, this relatedness is revealed in the functional categories rather than in the ‘ordinary’ lexicon. 454 Original quote:“Uma leitura rápida dos textos seguintes, respectivamente em Caboverdiano (…) e em Papiamento (…) dá-nos uma imagem imediata de identidade que tem despertado enorme curiosidade não só entre os estudiosos mas também entre os próprios falantes” 455 Original quote:“un abitante di CaboVerde i unu di nos IslananAbou por komprondé un pátaka di loke un ta bisá otro”

Appendix II

On the loss of pre-1674 records of the Chamber of Amsterdam After the 1st WIC’s bankruptcy in 1674, most of its documents on the slave trade were lost or destroyed. The 1st WIC’s archives that have survived belong mostly to the Chamber of Zealand and include, for instance, the archives that pertain to the WIC’s period in Brazil. However, both the Netherlands Antilles and the Petite Côte, the areas that concern us here, fell under the jurisdiction of the Chamber of Amsterdam. Unfortunately, “[a]s far as the Chamber of Amsterdam is concerned, for the period prior to 1668, not more than one single register exists”456 (Moraes 1995: 319, my translation from the original). According to Römer (in the preface of Gehring & Schiltkamp 1987), “[n]eglect and, regretfully, destruction of documents are the main reasons for our lack of knowledge today of the First West India Company’s history in New Netherlands, the Caribbean area and, in particular, of the settlements on Curaçao”. Gehring & Schiltkamp explain what exactly is understood by ‘neglect and destruction’: On Curaçao, the lack of suitable quarters for its archives caused many of the early records to become food for white ants and cockroaches. The deterioration of such important archival material prompted the home government to remove all records dating before January 1, 1864, from Curaçao for storage in the Algemeen Rijks Archief in The Hague. However, the damage had already been done to the early records, so that very little has survived in the Netherlands from the early years of Dutch occupation. The correspondence and administrative papers sent by the vice-directors of Curaçao to the directors at Amsterdam did not experience a much better fate. Rather than suffering the ravages of insects, these papers suffered from the eternal human compulsion to rid itself of burdensome materials, or more directly put, to clean house. (Gehring & Schiltkamp 1987: xix, xx)

An effective method applied to get rid of these ‘burdensome materials’ was to simply sell them as old paper457 . In other cases, letters, reports and records “were put through the paper shredder” (Moraes 1998b: 317). Adding to the misfortune, “a fire destroyed part of the remaining pieces” (Moraes 1998b: 317). Gehring & Schiltkamp provide more detail: 456 This register is analyzed in detail in Moraes (1998a: 321–329). 457 This is mentioned in the introduction to the archives of the first Dutch WIC (National Archives, The Hague, no. 1.05.01.01, no author mentioned).

342

Appendices

It was first reported by John R. Brodhead on his trip to the Netherlands in the 1840s, while he was searching for early records relating to New Netherland, that the records of the first West India Company were ordered destroyed by ministerial decree in 1821. However, inventories of records surviving from the end of the 18th and early 19th centuries clearly indicate that most of the first West India Company archives had already been destroyed before this time. The papers destroyed in 1821 were all dated later than 1674 and were for the most part concerned with financial affairs. Most likely, the majority of the records of the first West India Company had already been destroyed upon the occasion of the West India Company’s reorganization in 1674. (Gehring & Schiltkamp 1987: xx)

Appendix III

Papiamentu Swadesh-100 list 1 I

mi, ami

2 you (singular) 3 we 4 this 5 that 6 who 7 what

bo, abo

18 man (human being) 19 bird

nos, anos esaki esei, esaya ken, kende ki, kiko

20 dog 21 louse 22 tree 23 seed 24 leaf

8 not

no

9 all 10 many 11 one 12 two

tur hopi / masha un dos

13 big

grandi

30 bone

14 long 15 small

largu chikí, chikitu muhé hòmber

31 fat (noun) 32 fire

16 woman 17 man (adult male) 52 breast 53 heart (organ) 54 liver 55 to drink 56 to eat 57 to bite 58 to see

persona

35 tail

rabu

paha

36 feather

pluma

kachó pieu palu simia foya/o / blachi 25 root reis / rais / wòrtel 26 bark (of a tree) kaska 27 skin kuero / piel

37 fish 38 hair 39 head 40 ear 41 eye

piská kabei kabes orea wowo

42 fingernail

uña

43 nose 44 mouth

nanishi boka

28 meat 29 blood

45 tooth 46 tongue (organ) 47 foot

djente lenga

48 knee 49 hand

rudia man

33 egg 34 horn

karni sanger / sangro wesu / bòt / stòmpi vèt / sebu kandela, yama webu kachu

50 belly 51 neck

barika nèk

pechu kurason

70 to stand 71 to give

ta (pará) duna

88 green 89 yellow

bèrdè hel / gel

igra / higra bebe kome morde mira, weta

72 to say 73 sun 74 moon 75 star 76 water

bisa solo luna strea awa

90 white 91 black 92 night 93 warm 94 cold

blanku pretu nochi kayente friu

pia

344

Appendices

59 to hear 60 to know 61 to sleep 62 to die 63 to kill 64 to swim 65 to fly 66 to walk 67 to come 68 to lie (as in a bed) 69 to sit

tende, 77 rain skucha, oi sa, sabi 78 stone drumi muri mata landa bula kana bin, bini drumi

79 sand 80 earth 81 cloud 82 smoke 83 ashes 84 to burn 85 road 86 mountain

awaseru / yobida pieda, piedra santu tera nubia huma shinishi kima kaminda seru

sinta

87 red

kòrá

95 full

yen

96 new

nobo

97 98 99 100

bon ront, rondó seku nòmber

good round dry name

References

Abraham, Werner 2006. Introduction: Passivization and typology. Form vs. function – a confined survey into the research status quo. In Werner Abraham and Larisa Leisiö (eds.), Passivization and typology, 1–27. Amsterdam: Benjamins. Allen, Rose Mary 2007. Di ki manera? A Social History of AfroCuraçaoans, 1863–1917. Ph.D. dissertation, University of Utrecht. Alleyne, Mervyn C. 1980. Comparative Afro-American. An Historical-Comparative Study of English-Based Afro-American Dialects of the New World. Ann Arbor: Karoma. Almada, Dulce 1961. Cabo Verde: Contribuição para o estudo do dialecto falado no seu arquipélago. Junta de Investigações do Ultramar, Centro de Estudos Políticos e Sociais. Almeida, Sabou Espirito Santo Vaz de 1991. Crioulo Grammar made simple. GuineaBissau: Peace Corps. Álvarez Nazario, Manuel 1974. El elemento afronegroide en el español de Puerto Rico. San Juan: Instituto de Cultura Puertorriqueña. Ambrosius, Jacobus 1905. Boeki di oracion pa catholicanan di vicariato apostolico di Curaçao. Nijmegen: L.C.G. Malmberg. Andersen, Roger W 1974. Nativization and Hispanization in the Papiamentu of Curaçao, Netherlands Antilles. Ph.D. dissertation, University of Texas. Andersen, Roger W 1990. The Papiamentu tense – aspect system, with special attention to discourse. In John Victor Singler (ed.), Pidgin and Creole Tense / Aspect / Mood Systems, 59–96. Amsterdam: Benjamins. Andersen, Roger W. 2000. Temporal frames in spoken Papiamentu discourse. In John Rickford and Suzanne Romaine (eds.), Creole genesis, Sociohistory and Aesthetics: Studies celebrating Charlene Sato, 353–372. Amsterdam: Benjamins. Andrade, Ernesto d’ and Alain Kihm 2000. Stress in three creoles. In Ernesto d’Andrade, Dulce Pereira and Maria Antónia Mota (eds.), Crioulos de base portuguesa, 97–109. Lisbon: Associação Portuguesa de Linguística. Arends, Jacques 1999. The origin of the Portuguese element in the Surinam Creoles. In Magnus Huber and Mikael Parkvall (eds.), Spreading the word. The issue of diffusion among the Atlantic Creoles, 195–208. London: University of Westminster Press. Arlotto, Anthony 1972. Introduction to historical linguistics. Lanham, MD: University Press of America. Aschenberg, Heidi 2002. Kommunikation im nationalsozialistischen Konzentrationslager. Zeitschrift für romanische Philologie 118: 529–572. Baart, Willem J. H. 1983. Cuentanan di Nanzi: een onderzoek naar de oorsprong, betekenis en functie van de papiamentse spinverhalen. Oegstgeest: Baart. Contains a reissue of Geerdink-Jesurun Pinto, Nilda M. 1952. Cuentanan di Nanzi. Curaçao: Drukkerij Scherpenheuvel.

346

References

Bachmann, Iris 2005. Die Sprachwerdung des Kreolischen. Eine diskursanalytische Untersuchung am Beispiel des Papiamentu. Tübingen: Gunter Narr. Baird, Keith 1975. Anticipations of Papiamento in the Afro-Portuguese of Gil Vicente. CAAS Linguistics 6: 1–6. Baker, Philip 2009. Productive bimorphemic structures and the concept of gradual creolization. In Rachel Selbach, Hugo Cardoso and Margot van den Berg (eds.), Gradual Creolization. Studies celebrating Jacques Arends, 27–54. Amsterdam: Benjamins. Bakker, Peter 1997. A language of our own. New York/Oxford: Oxford University Press. Bakker, Peter 2000. Rapid language change: Creolization, intertwining, convergence. In Colin Renfrew, April McMahon and Larry Trask (eds.), Time Depth in Historical Linguistics, 585–620. Oxford: Oxbow Books. Bakker, Peter 2003. Mixed languages as autonomous systems. In Peter Bakker and Yaron Matras (eds.), The Mixed Language Debate. Theoretical and Empirical Advances, 107–150. Berlin: Mouton De Gruyter. Bakker, Peter 2008a. Pidgins versus Creoles and Pidgincreoles. In Silvia Kouwenberg and John V. Singler (eds.), Handbook of Pidgin and Creole Studies, 130–157. Oxford: Wiley-Blackwell. Bakker, Peter 2008b. Tense, mood, and aspect, in creole languages and the typology of affix order. In Folke Josephson and Ingmar Söhrman (eds.), Interdependence of diachronic and synchronic analyses, 43–60. Amsterdam: Benjamins. Bakker, Peter 2009. The Saramaccan lexicon: Verbs. In Rachel Selbach, Hugo Cardoso and Margot van den Berg (eds.), Gradual Creolization. Studies Celebrating Jacques Arends, 155–173. Amsterdam: Benjamins. Bakker, Peter, Norval Smith and Tonjes Veenstra 1995. Saramaccan. In Jacques Arends, Pieter Muysken and Norval Smith (eds.), Pidgins and Creoles. An Introduction, 165– 179. Amsterdam: Benjamins. Bakker, Peter and Yaron Matras 2003. The study of mixed languages. In Peter Bakker and Yaron Matras (eds.), The Mixed Language Debate. Theoretical and Empirical Advances, 1–20. Berlin: Mouton De Gruyter. Bakker, Peter and Yaron Matras (eds.) 2003. The Mixed Language Debate. Theoretical and Empirical Advances. Berlin: De Gruyter. Bakker, Peter, Aymeric Daval-Markussen, Mikael Parkvall and Ingo Plag 2011. Creoles are typ ologically distinct from non-creoles. Journal of Pidgin and Creole Languages 26(1): 5–42. Bal, Willy 1980. Devinettes en créole portugais du Sénégal recueillies par Abel Chataigner. In Jean-Marie d’Heur and Nicoletta Cherubini (eds.), Études de Philologie Romane et d’Histoire Littéraire offertes à Jules Horrent à l’occasion de son soixantième anniversare, 543–550. Liège. Bal, Willy 1983a. À propos du système verbal du kriôl ou créole à base portugaise du Sénégal. In Eugeenn Roegeist and Liliane Tasmowski (eds.), Verbe et phrase dans les langues romanes. Romanica Gandensia XX, 15–27. Gent: Rijksuniversiteit te Gent. Bal, Willy 1983b. Textes en Kri(y)ôl, créole portugais du Sénégal. In Homenaje a Álvaro Galmés de Fuentes, vol. 2, 19–38. Oviedo/Madrid: Editorial Gredos.

References

347

Bal, Willy 1983c. Une enquête d’Abel Chataigner sur le créole portugais de la Casamance. In Langue, dialecte, littérature. Études romanes à la mémoire de Hugo Plomteux, 175–186. Louvain. Bal, Willy 1985. Le verbe en créole à base portugaise du Sénégal. In Linguistique descriptive. Phonétique, morphologie et lexique. Actes du XVIIème Congrès International de Linguistique et Philologie Romanes, vol. 3, 113–119. Aix-en-Provence: Publications de l’Université de Provence. Balen, W.J. van 1940. Papiamentoe en Portugeesch. West-Indische Gids 22: 371–376. Baleno, Ilídio 2006. A afirmação da sociedade cabo-verdiana a partir da crise dos século XVII e XVIII. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria João Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 45–52. Tübingen: Narr. Baptista, Marlyse 1999. On the nature of the morpheme e in Capeverdean Creole: to be or not to be. In Klaus Zimmermann (ed.), Lenguas criollas de base lexical española y portuguesa. Frankfurt am Main: Vervuert, 25–47. Baptista, Marlyse 2000. Le Créole de Guinée-Bissau: Genèse et Situation Sociolinguistique. Africultures 26, 69–72. Baptista, Marlyse 2002. The Syntax of Cape Verdean Creole. The Sotavento Varieties. Amsterdam: Benjamins. Baptista, Marlyse 2006. When substrates meet superstrate: the case of Cape Verdean Creole. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria João Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 91–116. Tübingen: Narr. Baptista, Marlyse 2007a. On the syntax and semantics of DP in Cape Verdean Creole. In Marlyse Baptista and Jacqueline Guéron (eds.), Noun phrases in Creole languages: a multi-faceted approach, 61–105. Amsterdam: Benjamins. Baptista, Marlyse 2007b. Selection and Competition in Creole Formation: A Case Study. University of Pennsylvania Working Papers in Linguistics 13(2): 38–50. Baptista, Marlyse 2007c. Étude comparative des phrases copulatives: facteurs de variation et ramifications typologiques. In Karl Gadelii and Anne Zribi-Hertz (eds.), Grammaires créoles et grammaire comparative, 183–208. Saint-Denis: Presse Universitaire de Vincenne. Baptista, Marlyse, Heliano Mello and Miki Suzuki 2007. CapeVerdean, or Kabuverdianu, and Guinea-Bissau, or Kriyol (Creole Portuguese). In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 53–82. London: Battlebridge. Barbe, André 2002. Les îles du Cap-Vert. De la découverte à nos jours, une introduction. Paris: L’Harmattan. Barme, Stefan 2003. Zum (ibero-)romanischen Ursprung einiger (morpho-)syntaktischer Strukturen des Papiamentu und des español caribeño. Zeitschrift für romanische Philologie 119(2): 232–255. Barme, Stefan 2008. The indirect address construction in Papiamentu and its origin. In Nicolas Faraclas, Ronnie Severing and Christa Weijer (eds.), Linguistic Studies on Papiamentu, 7–26. Curaçao: Fundashon pa Planifikashon di Idioma.

348

References

Barros, Padre Marcelino Marques de 1899. O Guineense. Revista Lusitana 5: 271–300. Barry, Boubacar 1988. La Sénégambie du XVe au XIXe siècle. Paris: L’Harmattan. Bartens, Angela 1996. Die iberoromanisch-basierten Kreolsprachen. Frankfurt am Main: Peter Lang. Bartens, Angela 2000. Notes on Componential Diffusion in the Genesis of the Kabuverdianu Cluster. John McWhorter (ed.), Language Change and Language Contact in Pidgins and Creoles, 35–61. Amsterdam: Benjamins. Batalha, Luís 2004. The Cape Verdean Diaspora in Portugal: Colonial Subjects in a Postcolonial World. Lanham: Lexington. Batalha, Luís and Jørgen Carling (eds.) 2008. Transnational Archipelago. Perspectives on Cape Verdean Migration and Diaspora. Amsterdam: University of Amsterdam Press. Baxter, Alan 1988. A Grammar of Kristang (Malacca Creole Portuguese). Pacific Linguistics, Series B - No. 95. Canberra: Australian National University Press. Baxter, Alan 2004. A dictionary of Kristang (Malacca Creole Portuguese) – English. Canberra: Pacific Linguistics. Benoist, Joseph Roger de 2008. Histoire de l’Eglise catholique au Sénégal du milieu duXVe siècle à l’aube du troisième millénaire. Paris/Dakar: Karthala/Clairafrique. Benoist, Joseph Roger de andAbdoulaye Camara 2003. Histoire de Gorée. Paris: Maisonneuve/Larose. Berg, Margot van den 2007. A grammar of Early Sranan. Ph.D. dissertation, Universiteit van Amsterdam. Biagui, Noël Bernard Forthcoming. Description générale du créole afro-portugais parlé à Ziguinchor (Casamance), Dakar/Paris: Université CheikhAnta Diop/Institut National des Langues et Civilisations Orientales. Biagui, Noël Bernard and Nicolas Quint Forthcoming. Casamancese Creole. In Susanne Michaelis, Philippe Maurer, Martin Haspelmath and Magnus Huber (eds.), Atlas of Pidgin and Creole Structures. Bhat, Shankara 1999.The prominence of tense, aspect and mood. Amsterdam: Benjamins. Bhatt, Parth and Ingo Plag (eds.) 2006. The structure of creole words: Segmental, syllabic and morphological aspects. Tübingen: Niemeyer. Bickerton, Derek [1974]1980. Creolization, Linguistic Universals, Natural Semantax and the Brain. In Richard R. Day (ed.), Issues in English Creoles: Papers from the 1975 Hawaii conference, 1–18. Heidelberg: Groos. Bickerton, Derek 1975. Dynamics of a creole system. London: Cambridge University Press. Bickerton, Derek 1981. Roots of Language. Ann Arbor: Karoma Publishers. Bickerton, Derek 1984. The Language Bioprogram Hypothesis. Behavioral and Brain Sciences 7(2): 173–222. Bickerton, Derek 1995. The syntax of creole languages. In Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld and Theo Vennemann (eds.), An International Handbook of Contemporary Research, 1451–1462. Berlin/New York: De Gruyter.

References

349

Bickerton, Derek 1998. A sociohistoric examination of Afrogenesis. Journal of Pidgin and Creole Languages 13(1). 63–92. Bickerton, Derek 1999. Perspectives on Creole language history. New West Indian Guide 73: 97–102. Bickerton, Derek 2001. Okay for content words, but what about functional items? Behavioral and Brain Sciences 24: 1104–1105. Bickerton, Derek 2002. Sobre los pretendidos portuguesismos de la lengua palenquera. In Yves Moñino and Armin Schwegler (eds.), Palenque, Cartagena y AfroCaribe, 35–42. Tübingen: Max Niemeyer. Bickerton, Derek and Escalante, Aquiles 1970. Palenquero: A Spanish-Based Creole of Northern Colombia. Lingua 24: 254–267. Birmingham, John C. 1970. The Papiamentu language of Curaçao. Michigan: University Microfilms. Birmingham, John C. 1971. Still more on Papiamentu. Hispania 54(2): 300–301. Birmingham, John C. 1975. Papiamentu’s West African cousins. In Frances M. Aid (ed.), Colloquium on Hispanic Linguistics, 19–25. Washington: Georgetown University Press. Birmingham, John C. 1976. Papiamentu: the Long-Lost Lingua Franca? The American Hispanist 13: 8–11. Birmingham, John C. 1978. Lexical Decreolization in Papiamentu. Kristòf 4(2): 49–59. Boer, Harm den 1996. La literatura sefardí de Amsterdam. Alcalá de Henares: Instituto Internacional de Estudos Sefardíes y Andalusíes. Böhm, Günter 1992. Los sefardíes en los domínios holandeses de Amrérica del Sur y del Caribe: 1630–1750. Frankfurt am Main: Vervuert. Booij, Geert 1993. Against Split Morphology. In Geert Booij and Jaap van Marle (eds.), Yearbook of Morphology, 27–49. Dordrecht: Kluwer. Booij, Geert 2005. The Grammar of Words. An Introduction to Morphology. New York: Oxford University Press. Boretzky, Norbert 1983. Kreolsprachen: Substrate und Sprachwandel. Wiesbaden: Harrassowitz. Boretzky, Norbert and Birgit Igla 1994. Romani Mixed Dialects. In Peter Bakker and Maarten Mous (eds.), Mixed languages. 15 Case Studies in Language Intertwining, 35–68. Amsterdam: IFOTT. Bossong, Georg 2005. Die Sepharden. Geschichte und Kultur der spanischen Juden. München: C.H. Beck. Boulègue, Jean 1989. Les Luso-Africains de Sénégambie. Lisbon: Instituto de Investigação Científica. Boulègue, Jean 2006. Présences portugaises et sociétés africaines sur la côte de la ‘Guinée du Cap-Vert’. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria João Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 45–52. Tübingen: Narr. Bowern, Claire 2010. Correlates of Language Change in Hunter-Gatherer and Other ‘Small’ Languages. Language and Linguistics Compass 4(8): 665–679.

350

References

Boxer, Charles R. 1963. Race Relations in the Portuguese Colonial Empire, 1415–1825. Oxford: Clarendon Press. Böttcher, Nikolaus 1995. Aufstieg und Fall eines atlantischen Handelsimperiums. Frankfurt am Main: Vervuert. Brenneker, Paul 1959. Benta. Dos cien cantica dje dushi tempu bieuw. Curaçao: Boekhandel St. Augustinus. Brenneker, Paul 1986. Zjozjoli. Gegevens over de volkskunde van Curaçao, Aruba en Bonaire. Curaçao: author’s edition. Brinton, Laurel J. and Elizabeth Closs Traugott 2005. Lexicalization and language change. Cambridge: Cambridge University Press. Brito, de Paula A. [1887]1967. Apontamentos para a gramática do crioulo que se fala na ilha de Santiago de Cabo Verde. In Jorge Morais-Barbosa (ed.), Estudos Linguísticos crioulos, 329–404. Lisbon: Academia Internacional da Cultura Portuguesa. Broek, Aart 1988. Review of Peasants and capital: Dominica in the world economy, Michel-Rolph Trouillot, 1988, New West Indian Guide, 62(3/4): 165–209. Brooks, George E. 1993a. Landlords and Strangers: Ecology, Society and Trade in Western Africa, 1000–1630. Boulder: Westview Press. Brooks, George E. 1993b. Historical perspectives on the Guinea-Bissau region, fifteenth to nineteenth centuries. In Lopes (ed.), 25–54. Brooks, George E. 2003. Eurafricans in Western Africa: Commerce, Social Status, Gender, and Religious Observance from the Sixteenth to the Eighteenth Century. Oxford: James Currey. Bruyn, Adrienne 2008. Grammaticalization in Pidgins and Creoles. In Silvia Kouwenberg and John V. Singler (eds.), The Handbook of Pidgin and Creole Studies, 385–410. Oxford: Wiley-Blackwell. Bruyn, Adrienne 2009. Grammaticalization in creoles: ordinary and not-so-ordinary cases. Studies in Language 33(2): 312–337. Buddingh, Bernard R. 1994. Van Punt en Snoa. Ontstaan en groei van Willemstad, Curaçao vanaf 1634. Den Bosch: Aldus Uitgevers. Bull, Benjamin Pinto. 1989. O crioulo da Guiné-Bissau. Filosofia e Sabedoria. Lisboa/Bissau: Instituto de Cultura e Lingua Portuguesa/Instituto Nacional de Estudos e Pesquisas. Busche, Christian 1993. Die portugiesische Entstehung des Papiamentu in der Kritik. In Matthias Perl, Axel Schönberger and Petra Thiele (eds.), Portugiesisch basierte Kreolsprachen, 103–113. Frankfurt am Main: TFM. Busse, Winfried 1999. Die Sprache(n) der Sepharden: Ladino, Ladino. In Norbert Rehrmann and Andreas Koechert (eds.), Spanien und die Sepharden. Geschichte, Kultur, Literatur, 133-143. Tübingen: Niemeyer. Buurt, Gerard van 2009. Caquetío Indians on Curaçao during colonial times and Caquetío words in the Papiamentu language. In Nicholas Faraclas, Ronald Severing, Christa Weijer and Liesbeth Echteld (eds.), Leeward voices: Fresh perspectives on Papiamentu and the literatures and cultures of the ABC Islands, vol 1, 57–82. Curaçao: Fundashon pa Planifikashon di Idioma/University of the Netherlands Antilles.

References

351

Buurt, Gerard van and Sidney Joubert 1997. Stemmen uit het verleden: Indiaanse woorden in het Papiamentu. Alphen aan den Rijn: Van Buurt. Bybee, Joan L., Revere Perkins and William Pagliuca 1994. The evolution of grammar. Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Bybee, Joan L. 1994. The Grammaticalization of Zero. In William Pagliuca (ed.), Perspectives on Grammaticalization, 235–254. Amsterdam: Benjamins. Cardoso, Pedro [1933]1983. Folclore CaboVerdeano. Paris: Solidariedade Caboverdiana. Cardoso, Eduardo 1990. O crioulo da ilha de S. Nicolau de Cabo Verde. Lisboa: Instituto de cultura e Língua Portuguesa. Cardoso, Hugo and Norval Smith 2004. A new look at the Portuguese element in Saramaccan. Journal of Portuguese Linguistics 3: 115–147. Carreira, António 1983a. O crioulo de Cabo Verde: Surto e expansão. Lisbon: Europam. Carreira, António 1983b. Documentos para a História das Ilhas de Cabo Verde e ‘Rios de Guiné’ (Séculos XVII e XVIII). Lisbon: author’s edition. Cassidy, Frederic G. 1971. Tracing the Pidgin Element in Jamaican Creole. In Dell Hymes (ed.), Pidginization and Creolization of Languages, 203–221. Cambridge: Cambridge University Press. Castillo Mathieu, Nicolas del 1982. Esclavos negros en Cartagena y sus aportes léxicos. Bogotá: Instituto Cao y Cuervo. Castro, Jorge G. de 1976. Un bosqueho riba origen di Papiamento. Aruba: Estilo di Combersashon. Chataigner, Abel 1963. Le créole portugais du Sénégal: Observations et textes. Journal of African languages 2(1): 44–71. Chaudenson, Robert 1995. Les créoles. Paris: Presses Universitaires de France. Clements, Joseph Clancy 2007. Korlai (Creole Portuguese), or No Lin. In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 153–174. London: Battlebridge. Clements, Joseph Clancy and Andrew Koontz-Garboden 2002. Two Indo-Portuguese Creoles in Contrast. Journal of Pidgin and Creole Languages 17(2): 191–236. Clemesha, Josephine 1981. Hispanización y desacriollamiento en Papiamento. Utrecht: Instituto de Estudios Hispánicos, Portugueses y Iberoamericanos. Coelho, F. Adolpho [1880]1967. Os Dialectos Románicos ou Neo-Latinos na África, Asia e América. Boletim da Sociedade de Geografia de Lisboa. In Jorge Morais-Barbosa (ed.), Estudos linguísticos crioulos, 1–108. Lisbon: AICP. Coelho, Margarida, Cintia Alves, Valentina Coia, Donata Luiselli, Antonella Useli, Tjerk Hagemeijer, António Amorim, Giovanni Destro-Bisol and Jorge Rocha 2008. Human microevolution and Atlantic slave trade: a case study of S. Tomé (Gulf of Guinea). Current Anthropology 49(1): 134–143. Coffen, Béatrice 2002. Histoire culturelle des pronoms d’adresse. Vers un typologie des systèmes allocutoires dans les langues romanes. Paris: Honoré Champion. Comrie, Bernard 1985. Tense. Cambridge: Cambridge University Press.

352

References

Comrie, Bernard 1988. Genetic classification, contact, and variation. In Thomas J. Walsh (ed.), Georgetown University Round Table on Languages and Linguistics, 81–93. Washington D.C.: Georgetown University Press. Comrie, Bernard 2000. Is there a single time depth cut-off point? In Colin Renfrew, April McMahon and Larry Trask (eds.), Time Depth in Historical Linguistics, vol. 1, 33–43. Cambridge: McDonald Institute for Archaeological Research. Comrie, Bernard 2008. Inflectional morphology and language contact, with special reference to mixed languages. In Peter Siemund and Noemi Kintana (eds.), Language Contact and Contact Languages, 15–32. Amsterdam: Benjamins. Conradi, Cornelis 1844. Ewanhelie di San Matheo. Curaçao: Fundashon pa Planifikashon di Idioma. Coomans, H. and Coomans-Eustatia, M. (eds.) 2002. Mgr Niewindt en de Drukpers op Curaçao. Postscript to Prefecto Apostolico di Curaçao na Cristian di su mision. E teksto imprimí di mas bieu na Papiamentu di 1833, 26–32. Bloemendaal: Stichting Libri Antilliani. Coomans, H. and Coomans-Eustatia, M. (eds.) 2005. Aruba en Papiamento: twee vroeg Papiamentse Protestantse uitgaven van Aruba / door Abraham van Dragt (1847) en Nicolaas A. Kuiperi (1864). Curaçao: Fundashon pa Planifikashon di Idioma. Costa, J.B.V. da and C.J. Duarte [1886]1967. O crioulo de Cabo Verde: Breves estudos sobre o crioulo das ilhas de Cabo Verde. In; Jorge Morais-Barbosa (ed.), Estudos Linguísticos Crioulos, 235–328. Lisboa: Academia International da Cultura Portuguesa. Couto, Hildo Honório do 1992. Lançados, grumetes e a origem do crioulo português no noroeste africano. In Ernesto d’Andrade and Alain Kihm (eds.), Actas do do Colóquio sobre Crioulos de base lexical portuguesa, 109–122. Lisbon: Colibri. Couto, Hildo Honório do 1994. O crioulo português da Guiné-Bissau. Hamburg: Helmut Buske. Couto, Hildo Honório do 2007. The Ecology of Spatial Relations: The Case of Kriol Prepositions. In Martina Schrader-Kniffki and Laura Morgenthaler García (eds.), La romania en interacción: entre historia, contacto y política, 479–514. Frankfurt/Madrid: Iberoamericana/Vervuert. Couto, Hildo Honório do 2009a. On the So-Called Complex Prepositions in Kriol. Revue Roumaine de Linguistique 3/4: 279–294. Couto, Hildo Honório do 2009b. As narrativas orais crioulo-guineenses. Papia 19: 51–68. Crowley, Terry 2008. Pidgin and Creole Morphology. In Silvia Kouwenberg and John V. Singler (eds.), The Handbook of Pidgin and Creole Studies, 74–97. Oxford: WileyBlackwell. Cuales, Sonia Magdalena 1998. Gender in the Netherlands Antilles. In Patricia Mohammed (ed.), Rethinking Caribbean difference, 86–100. New York: Routledge. Curtin, Philip D. 1975. Economic Change in Precolonial Africa: Senegambia in the Era of the Slave Trade. Madison: University of Wisconsin Press. Dahl, Östen and Viveka Velupillai 2005a. Tense and Aspect. In Martin Haspelmath, Matthew Dryer, David Gil and Bernard Comrie (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, introduction to chapters 65–68. http://wals.info/supplement/intro65-68. Accessed on 13-10-2010.

References

353

Dahl, Östen and Viveka Velupillai 2005b. Perfective/Imperfective. In Martin Haspelmath, Matthew Dryer, David Gil and Bernard Comrie (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter 65. http://wals.info/feature/65. Accessed on 13-10-2010. Dahl, Östen and Viveka Velupillai 2005c. The Future Tense. In Martin Haspelmath, Matthew Dryer, David Gil and Bernard Comrie (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter 67. http://wals.info/ feature/67. Accessed on 13-10-2010. Dalgado, Sebastião 1906. Dialecto indo-portugues do norte. Revista Lusitana 9: 142– 166, 193–228 DeBose, Charles 1975. Papiamentu: a Spanish-based creole. Michigan: University of Michigan dissertation. Debrot, Cola 1973. Kelki na boka: pasatempu ku ta konsisti di shete diálogo. Curaçao: author’s edition. Debrot, Cola 1986. Verzameld werk: Dagboekbladen uit Genève. Vol. 6. Amsterdam: Meulenhoff. DeGraff, Michel 2001. Morphology in creole genesis : Linguistics and ideology. In M. Kenstowicz (ed.), Ken Hale: A life in language. Cambridge MA: The MIT Press, 53–121. DeGraff, Michel 2002. Relexification:A reevaluation. Anthropological Linguistics 44(4): 321–414. DeGraff, Michel 2003. Against Creole exceptionalism. Language 79(2): 391–410. DeGraff, Michel 2007. Kreyòl Ayisyen, or Haitian Creole (Creole French). In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 101–126. London: Battlebridge. Delgado, Mariano 2003. Der Traum von der Universalmonarchie Zur Danielrezeption in den iberischen Kulturen nach 1492. In Mariano Delgado, Klaus Koch and Edgar Marsch (eds.), Europa. Tausendjähriges Reich und Neue Welt. Zwei Jahrtausende Geschichte und Utopie in der Rezeption des Danielbuches, 252–305. Freiburg/Stuttgart: Kohlhammer. Detges, Ulrich 2001. Grammatikalisierung. Eine kognitiv-pragmatische Theorie, dargestellt am Beispiel romanischer und anderer Sprachen. Habilitation, University of Tübingen. Dieck, Marianne 2002. Distribución y alcance de la negación en palenquero. In Yves Moñino and Armin Schwegler (eds.), Palenque, Cartagena y AfroCaribe, 149–167. Tübingen: Max Niemeyer. Dieck, Marianne 2008. La expresión de la reflexividad en Palenquero. Papia 18: 45–69. Dijkhoff, Martha B. 1993. Papiamentu Word Formation. Ph.D. dissertation, Universiteit van Amsterdam. Dijkhoff, Marta, Silvia Kouwenberg and Paul Tjon Sie Fat 2004. The Dutch-speaking Caribbean. In Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics. An international handbook of the science of language and society, vol 3, 2105–2114. Berlin/New York: Mouton De Gruyter.

354

References

Dijkhoff, Mario 2000. Gramatica di Papiamento. Oranjestad, Aruba: author’s edition. Dikker, Suzanne 2008. Spanish prepositions in Media Lengua: Redefining relexification. In Thomas Stolz, Dik Bakker and Rosa Salas Palomo (eds.), Hispanisation: The Impact of Spanish on the Lexicon and Grammar of the Indigenous Languages of Austronesia and the Americas, 121–148. Berlin/New York: Mouton De Gruyter. Dissel, Sijbrand van 1857. Curaçao. Herinneringen en Schetsen. Leiden: A.W. Sythoff. Dissel, Sijbrand van 1865. Evangeli segoen Marko. New York: American Bible Society. Doneux, Jean Léonce 1979. Systèmes phonologiques des langues de Casamance et français parlé par les élèves casamançais du CM1 (cours moyen première année). Dakar: Centre de Linguistique Appliquée de Dakar. Doneux, Jean Léonce and Jean-Louis Rougé 1988. En apprenant le creole a Bissau ou Ziguinchor. Paris: L’Harmattan. Doneux, Jean Léonce and Jean-Louis Rougé 1993. Gramática das línguas do país, gramática do crioulo. Papia 2(2): 50–58. Dragt, Abraham van [1847]2005.Aruba en Papiamento: twee vroeg-Papiamentse Protestantse uitgaven van Aruba / door Abraham van Dragt (1847) en Nicolaas A. Kuiperi (1864). Bloemendaal/Curaçao: Stichting LibriAntilliani/Fundashon pa Planifikashon di Idioma. Drescher, Seymour 2001. Jews and New Christians in the Atlantic Slave Trade. In Paolo Bernardini and Norman Fiering (eds.), The Jews and the Expansion of Europe to the West, 1450–1800, 439–470. New York/Oxford: Berghahn Books. Dryer, Matthew S. 2005. Position of Tense-Aspect Affixes. In Martin Haspelmath, Matthew Dryer, David Gil and Bernard Comrie (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter 69. http://wals.info/ feature/69. Accessed on 13-10-2010. Eckkrammer, E. Martha 1993. Interferencias entre el papiamento y el español: ¿testigos de una coexistencia pacífica o de una lucha encarnizada? In Jens Lüdtke and Matthias Perl, Lengua y Cultura en el Caribe Hispánico, 131–146. Tübingen: Niemeyer. Eckkrammer, Eva Martha 1996. Übersetzung als Werkzeug des Sprachausbaus: Am Beispiel Papiamentu. Bonn: Romanistischer Verlag. Eckkrammer, Eva Martha 2004. Passive voice in Papiamento: A corpus-based study on dialectal variability. In Geneviève Escure and Armin Schwegler (eds.), Creoles, Contact, and Language Change: Linguistic and social implications, 209–219. Amsterdam: Benjamins. Eckkrammer, Eva Martha 1999. The Standardisation of Papiamentu: New Trends, Problems and Perspectives. Bulletin suisse de linguistique appliquée 69(1): 59–74. Einarsdóttir, Jónína 2004. Tired of weeping: mother love, child death, and poverty in Guinea-Bissau. Wisconsin: The University of Wisconsin Press. Emmanuel, Isaac S. 1957. Precious stones of the Jews of Curaçao. Curaçaon Jewry 1656–1957. New York: Bloch Publishing Company. Emmanuel, Isaac S. and Suzanne A. Emmanuel 1970. History of the Jews of the Netherlands Antilles. Cincinnati: American Jewish Archives.

References

355

Endruschat, Annette 2004. Elementos preposicionais nas línguas crioulas de base portuguesa. Papia 14: 47–70. Endruschat, Annette 2005. Präpositionen in portugiesisch-basierten Kreolsprachen. In Annette Endruschat and Axel Schönberger, Portugiesische Kreolsprachen: Entstehung, Entwicklung, Ausbau und Verwendung, 121–160. Frankfurt am Main: Domus Editoria Europaea. Esteves, Maria Luísa 1988. Gonçalo de Gamboa de Aiala, Capitão-Mor de Cacheu, e o comércio negreiro espanhol (1640-1650). Lisbon: Centro de Estudos de História e Cartografia Antiga. Evertsz, N.J. 1898. Compendio de la gramaìtica del papiamento, oì sea meìtodo para aprender aì hablarlo y aì escribirlo en corto tiempo. Curaçao: Bethencourt. Ewijk, Henricus van 1875. Nederlandsch-Papiamentsch-Spaansch woordenboekje. Curaçao: author’s edition. Eybers, G.J. 1915. Iets over het Papiamentu en zijn schrijfwijze. Amigoe di Curaçao (newspaper), 02-10-1915. Eybers, G.J. 1916. E testament Nobo di nos Señor y Salbador Jesus Christo: Fielmente traduci segun e texto original. Amsterdam: Nederlands Bijbelgenootschap. Fanha, Dulce 1987. Aspectos do contacto entre o português e o crioulo de Cabo Verde. In Actos do congresso sobre a situação actual da língua portuguesa no mundo, vol. 2, 292–309. Lisbon: ICALP. Faraclas, Nicholas 2007. Tok Pisin (Pidgin/Creole English). In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 355–372. London: Battlebridge. Faraclas, Nicholas, Yolanda Rivera-Castillo and Don E. Walicek 2007. No exception to the rule.The tense-modality-aspect system of Papiamentu reconsidered. In Magnus Huber and Viveka Velupillai (eds.), Synchronic and Diachronic Perspectives on Contact Languages, 257–278. Amsterdam: Benjamins. Faure, Claude 1914. Histoire de la Presqu’île du Cap Vert et des Origines de Dakar. Paris: Émile Larose. Feingold, Eduardo, D. 1996. Child language, creolization and historical change. Spanish in contact with Portuguese. Tübingen: Narr. Ferraz, Luís Ivens 1978. The Creole of São Tomé. African Studies 37: 3–68 / 235–288. Ferraz, Luis Ivens 1979. The Creole of São Tomé. Johannesburg: Witwatersrand University Press. Ferraz, Luis Ivens 1983. The Origin and Development of four creoles in the Gulf of Guinea. In Ellen Woolford and William Washabaugh (eds.), The Social Context of creolization, 120–125. Ann Arbor: Karoma. Ferraz, Luiz Ivens 1987. Portuguese Creoles of West Africa and Asia. In Glenn Gilbert (ed.), Pidgin and Creole Languages: Essays in Memory of John E. Reinecke, 337– 360. Honolulu: University of Hawaii Press. Ferraz, Luiz Ivens and Marius F. Valkhoff 1975. A Comparative Study of São-Tomense and Cabo-Verdiano Creole. In Marius F. Valkhoff (ed.), Miscelânia luso-africana: colectânea de estudos, 15–39. Lisbon: Junta de Investigações Científicas do Ultramar.

356

References

Ferrol, Orlando 1982. La cuestión del origen y de la formación del papiamento. The Hague: Smit. Fiéis, Maria Alexandra and Fernanda Pratas 2007. Reflexivity in Capeverdean. Predicate properties and sentence structure. In Magnus Huber and Viveka Velupillai (eds.), Synchronic and diachronic perspectives on contact languages, 117–158.Amsterdam: Benjamins. Filho, Wilson Trajano 2000. Outros rumores de identidade na Guiné-Bissau. Brasília: Universidade de Brasília Press. Fokker, Anthony Abraham 1914. Het Papiamentoe of basterd-Spaans der West-Indiese Eilanden. Tijdschrift voor Nederlandsche Taal- en Letterkunde 33: 54–79. Fontes, Carlos 2007. Estudo do léxico do São-Tomense com diccionario. Master’s thesis, University of Coimbra. Fontes, Carlos and John A. Holm 2008. Pan-Creolisms in Santomé. Paper presented at the annual meeting of the Associação de Crioulos de Base Lexical Portuguesa e Espanhola, Lisbon 2–4 July 2008. Fouse, Gary C. 2002. The Story of Papiamentu. A Study in Slavery and Language. Lanham: University Press of America. Fouse, Gary C. 2007. Slavery and the development of Papiamentu. Journal of Caribbean Literatures 5(1): 61–72. Frank, David B. 2007. Three irregular verbs in Gullah. In Magnus Huber and Viveka Magnus (eds.), Synchronic and Diachronic Perspectives on Contact Languages, 159– 173. Amsterdam: Benjamins. Frederiks, Bernardus and Jacobus Putman [1859]2004. Woordenlijst der in de landstaal van Curaçao meest gerbuikelijke woorden met Zamenspraken. Curaçao/Bloemendaal: Fundashon pa Planifikashon di Idioma/Stichting Libri Antilliani. Friedemann, Nina de and Patiño Rosselli, Carlos 1983. Lengua y Sociedad en el Palenque de San Basilio. Bogotá: Instituto Caro y Cuervo. Fuchs, August 1849. Die romanischen Sprachen in ihrem Verhältnisse zum Lateinischen. Halle: Schmidt. Garcia, Erica C. 1985. Shifting variation. Lingua 67(2–3): 189–224. Gatschet, Alb. S. 1884. The Aruba language and the Papiamento jargon. Proceedings of the American Philosophical Society 22: 299–305. Geerdink-Jesurun Pinto, N.M. 1954. Corsouw ta conta. Hengelo: Drukkerij Smit. Gehring, Charles T. and J.A. Schiltkamp. 1987. Curaçao Papers 1640–1665. Series: New Netherland Documents, vol. XVII. Interlaken, New York: Heart of the Lakes Publishing. Ginneken, Jacobus van 1928. Handboek der Nederlandsche taal, I. Den Bosch: Malmberg. Giusti, Emilio 1981. Contes créoles de Guinée Bissau. Paris: Edicef. Givón, Talmy 1971. Historical Syntax and Synchronic Morphology: An Archeologist’s Field Trip. In Papers from the 7th Regional Meeting of the Chicago Linguistic Society, 394–415. Chicago: Chicago Linguistic Society. Goilo, Enrique R. [1951]1975. Papiaments Leerboek. Aruba: VAD Drukkers.

References

357

Goilo, Enrique R. 1974. Hablemos papiamento. Aruba: De Wit Stores. Goldish, Josette C. 2009. Once Jews: Stories of Caribbean Sephardim. Princeton, New Yersey: Markus Wiener Publishers. Golovko, Evgeniy V. 2003. Language contact and group identity: The role of “folk” linguistic engineering. In Peter Bakker andYaron Matras (eds.), The Mixed Language Debate. Theoretical and Empirical Advances, 177–208. Berlin: Mouton De Gruyter. Gomes Casseres, Charles 1990. Istoria kòrtiku di Hudiunan di Korsou. Curaçao: Scherpenheuvel. (http://www.sefarad.org/publication/lm/042/2.html) Gonçalves, Rita, Ana Freitas, Ana Fernandes, Toomas Kivisild, Richard Villems and Antonio Brehm 2003. Y-chromosome lineages in Cabo Verde Islands witness the diverse geographic origin of its first male settlers. Human Genetics 113: 467–472. Good, Jeff 2009. Loanwords in Saramaccan, an English-based creole of Suriname. In Martin Haspelmath and Uri Tadmor. Loanwords in the World’s Languages: A Comparative Handbook, 918–943. Berlin: Mouton De Gruyter. Goodman, Morris 1987. The Portuguese Element in the American Creoles. Pidgin and Creole Languages. In Glenn G. Gilbert (ed.), Essays in Memory of John E. Reinecke, 361–405. Honolulu: University of Hawaii Press. Goslinga, C. CH. 1979. A short history of the Netherlands Antilles and Surinam. The Hague: Martinus Nijhoff. Granda, Germán de 1968. La tipología ‘criolla’ de dos hablas del área lingüística hispánica. Thesaurus 23: 193–205. Granda, Germán de 1973. Papiamento en Hispanoamérica (siglos XVII-XIX). Thesaurus 28: 1–13. Granda, Germán de 1974. El repertorio lingüístico de los sefarditas de Curaçao durante los siglos XVII y XVIII y el problema del origen del papiamento. Romance Philology 28(1): 1–16. Granda, Germán de 1978. Estudios lingüísticos hispánicos, afrohispánicos y criollos. Madrid: Gredos. Granda, Germán de 1986. Retenciones africanas en la fonética del criollo portugues de annobón. Revista de filología románica 4: 111–124. Grant, Anthony P. 2001. Language intertwining: Its depiction in recent literature and its implications for theories of creolisation. In Norval Smith and Tonjes Veenstra (eds.), Creolization and contact, 81–112. Amsterdam: Benjamins. Grant, Anthony P. 2008a. The Portuguese elements in Papiamentu. In Nicholas Faraclas, Ronnie Severing and Christa Weijer (eds.), Linguistic studies on Papiamentu, 47–72. Curaçao: Fundashon pa Planifikashon di Idioma. Grant, Anthony P. 2008b. A constructivist approach to the early history of Papiamentu. In Nicholas Faraclas, Ronnie Severing and Christa Weijer (eds.), Linguistic studies on Papiamentu, 73–112. Curaçao: Fundashon pa Planifikashon di Idioma. Grant, Anthony 2009. Loanwords in British English. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook, 360–383. Berlin: Mouton De Gruyter.

358

References

Green, John N. 1988. Romance Creoles. In Martin Harris and Nigel Vincent (eds.), The Romance Languages, 420–473. London: Oxford University Press. Green, Kate 1995. Sobre la evolución y status de la cópula en criollos de base española y portuguesa. Papia 4(2): 56–70. Green, Toby 2005. Further Considerations on the Sephardim of the Petite Côte. History in Africa 32: 165–183. Green, Toby 2007. Masters of Difference: Creolization and the Jewish Presence in Cabo Verde, 1497–1672. Ph.D. dissertation, University of Birmingham. Green, Toby 2008. Amsterdam and the African Atlantic: The Role of Sephardim from Amsterdam in Senegal in the 17th Century. In Hilary Pomeroy, Christopher J. Pountain and Elena Romero (eds.), Proceedings of the Fourteenth British Conference on Judaeo-Spanish Studies, 85–94. London: Queen Mary and Westfield College. Green, Toby 2011. The Rise of the Trans-Atlantic Slave Trade in Western Africa, 1300– 1589. Cambridge: Cambridge University Press. Greenberg, Joseph H. [1971]2005. Language, culture, and communication. Stanford: Stanford University Press. Greenberg, Joseph H. 1987. Language in the Americas. Stanford: Stanford University Press. Guèye, Moustapha 1998. Gorée et la traite négrière. In Djibril Samb (ed.), Gorée et l’esclavage, 21–40. Dakar: Université Cheikh Anta Diop – IFAN. Guillaume, Gustave 1973. Principes de linguistique théorique. Paris/Québec: Klincksieck/Les Presses de l’Université de Laval. Günther, Winfried 1973. Das portugiesische Kreolisch der Ilha do Príncipe. Marburg an der Lahn: Selbstverlag. Gutiérrez Maté, Miguel 2010. Génesis de los pronombres sujetos obligatorios del español del Caribe: la hipótesis del contacto afro-hispánico sometida a revisión. In Teresa Encinas Manterola et al. (eds.), Ars longa. Diez años de AJIHLE, vol. 2, 387–414. Buenos Aires: Voces del Sur. Gutiérrez Maté, Miguel To appear a. El pronombre usted en el español de Cartagena de Indias del siglo XVII y su divergencia de vuestra merced. In E. Montero et al. (eds.), Actas del VIII Congreso Internacional de la Asociación de Historia de la Lengua Española. Gutiérrez Maté, Miguel To appear b. Palenquero y español colombiano atlántico en el siglo xvii. Conciencia lingüística y testimonio directo en documentos de archivo. Revista Internacional de la lingüística iberoamericana. Haardt, Caroline 1992. Goree, island of slaves. UNESCO Courier. http://findarticles.com/p/articles/mi_m1310/is_1992_Oct/ai_13979342/?tag= mantle_kin;content. Accessed on 14-09-2010 Hagemeijer, Tjerk 2007. Clause Structure in Santome. Ph.D. dissertation, Universidade de Lisboa. Hagemeijer, Tjerk 2009. Creole languages and pro. Berlin: Humboldt Universität invited talk. http://www2.hu-berlin.de/asaf/Afrika/Dokumente/Kolloquien/Hagemeijer% 202009%2010%2027%20Creoles%20Pro-drop.pdf. Accessed on 13-11-2010.

References

359

Hagemeijer, Tjerk and John A. Holm 2008. Schuchardt, Hugo. On the Creole Portuguese of São Tomé (West Africa). Annotated translation of “Ueber das Negerportugiesische von S. Thomé (Westafrika).” Sitzungsberichte der kaiserlichen Akademie der Wissenschaften zu Wien 101(2): 889–917. [1882]. In John A. Holm and Suzanne Michaelis (eds.), Contact Languages: Critical Concepts in Linguistics, vol. 1, 131– 156. London/New York: Routledge. Hall, Robert 1966. Pidgin and Creole Languages. Ithaca: Cornell University Press. Hancock, Ian F. 1971. A survey of the Pidgins and Creoles of the World. In Del Hymes (ed.), Pidginization and Creolization of Languages, 509–523. Cambridge: Cambridge University Press. Hancock, Ian F. 1975. Malacca Creole Portuguese: Asian, African or European? Anthropological Linguistics 17(55): 211–236. Hancock, Ian F. 1984. Review of: Orlando Ferrol, La cuestión del origen y de la formación del papiemento, 1982. New West Indian Guide 58(2): 139–142. Hartog, Johan 1968. Curaçao, from colonial dependence to autonomy. Aruba: De Wit. Haseth, Carel de 1990. Een bijdrage in de discussie over het ontstaan van het Papiaments. De Gids 53: 548–557. Haseth, Carel de 1995. Feiten, cijfers en data. Kruispunt 36: 29–34. Haspelmath, Martin 2009. Lexical borrowing: Concepts and issues. In Martin Haspelmath and Uri Tadmor (eds.), Loanwords in the World’s Languages: A Comparative Handbook, 35–54. Berlin: Mouton De Gruyter. Haspelmath, Martin and Uri Tadmor (eds.) 2009. Loanwords in the World’s Languages: A Comparative Handbook. Berlin: Mouton De Gruyter. Online World Loanword Database (WOLD): http://wold.livingsources.org Havik, Philip 2004. Silences and Soundbites. The Gendered Dynamics of Trade and Brokerage in the Pre-colonial Guinea Bissau Region. Münster: Lit Verlag. Hazaël-Massieux, Marie-Christine 2006. Using and interpreting historical texts to analyse the formation and development of creole languages. In Clancy Clements, Thomas A. Klingler, Deborah Piston-Hatlen and Kevin J. Rottet (eds.), History, Society and Variation in Pidgin and Creole Languages, 29–45. Amsterdam: Benjamins. Heijer, Henk den 1994. De geschiedenis van de WIC. Zupthen: Walburg Pers. Heijer, Henk den 2003. The West African Trade of the Dutch West India Company, 1674–1740. In Johannes Postma and Victor Enthoven (eds.), Riches From Atlantic Commerce, 138–169. Leiden: Brill. Heine, Bernd 2005. On reflexive forms in creoles. Lingua 115: 201–257. Heine, Bernd and Mechtild Reh 1984. Grammaticalization and reanalysis in African languages. Hamburg: Buske. Heine, Bernd and Kyung-An Song 2010. On the genesis of personal pronouns: Some conceptual sources. Language and Cognition 2(1): 117–147. Heine, Bernd and Tania Kuteva 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press.

360

References

Henriquez, May 1988. Ta asina? O ta asana? Abla, uzu i kustumber sefardí. Curaçao: author’s edition. Henriquez, May 1991. Lo ke a keda pa simia. Curaçao: author’s edition. Henríquez Ureña, Pedro 1921. Observaciones sobre el español de América. Revista de Filología Española 8: 357–390. Hesseling, Dirk Christiaan 1933a. Een Spaans boek over het Papiaments. Tijdschrift voor Nederlandsche Taal- en Letterkunde 52: 40–69. Hesseling, Dirk Christiaan 1933b. Papiaments en Negerhollands. Tijdschrift voor Nederlandsche Taal- en Letterkunde 52: 265-288. Heuvel, P., F. van Wel and A. van der Wal 1989. Met eigen stem: herkenningspunten in de letterkunde van de Nederlandse Antillen en Aruba. Den Haag: Kabinet voor Nederlands-Antilliaanse Zaken. Hock, Hans Heinrich 1991. Principles of Historical Linguistics. Berlin: Mouton De Gruyter. Hoetink, Harmannus 1958. Het patroon van de oude Curac¸aose samenleving : een sociologische studie. Assen: Van Gorcum. Holm, John A. 1988. Pidgins and Creoles. 2 vols. Cambridge: Cambridge University Press. Holm, John A. 1992. Popular Brazilian Portuguese. A semi-creole. In Ernesto d’Andrade and Alain Kihm (eds.), Actas do Colóquio sobre Crioulos de Base Lexical Portuguesa, 37–66. Lisbon: Colibri. Holm, John A. 2000.An introduction to pidgins and creoles. Cambridge University Press. Holm, John A. 2001. Chabacano versus related creoles: (Socio-)linguistic affinities and differences. Estudios de Sociolingüística 2(2): 69–93. Holm, John A. 2002. Research on African-American language varieties, 1956–1996. In Yves Moñino and Armin Schwegler (eds.), Palenque, Cartagena y AfroCaribe, 51–63. Tübingen: Max Niemeyer. Holm, John A. 2004. Languages in contact. The partial restructuring of vernaculars. Cambridge: Cambridge University Press. Holm, John A. 2008. Creolization and the fate of inflections. In Thomas Stolz, Dik Bakker and Rosa Salas Palomo (eds.), Aspects of Language Contact. Berlin: Mouton De Gruyter. 299–324. Holm, John A. 2009. Atlantic features in Asian varieties of Creole Portuguese. Journal of Portuguese Linguistics 8(2): 11–22. Holm, John A., Alan Baxter, William Megenney and Heliana de Mello 1998. Panorama do português vernáculo do Brasil. In Matthias Perl and Armin Schwegler (eds.), América negra: panorámica actual de los estudios lingüísticos sobre variedades criollas y afrohispanas, 71–137. Frankfurt: Vervuert Holm, John A. et al. 2000. The Creole Verb. A Comparative Study of Stativity and Time Reference. In John McWhorter (ed.), Language Change and Language Contact in Pidgins and Creoles, 133–161. Amsterdam: Benjamins. Holm, John A. and Incanha Intumbo 2009. Quantifying superstrate and substrate influence. Journal of Pidgin and Creole Languages 24(2): 259–262.

References

361

Holm, John A. and Peter L. Patrick (eds.) 2007. Comparative Creole Syntax. London: Battlebridge. Holm, John A. and Dominika Swolkien 2004. The vernaculars of São Vicente (Cape Verde) and Brazil: demographics and degrees of restructuring. Revista Internacional de Ligüística Iberoamericana IV(1): 72–86. Holm, John A. and Dominika Swolkien 2006. A expansão do crioulo cabo-verdiano para São Vicente: Factores sócio-históricos na difusão. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria Jão Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 199–220. Tübingen: Gunter Narr. Holm, John A. and Dominika Swolkien 2010. Inflections in the 19th c. Upper Guinea Creole texts. Reveue Roumaine de Linguistique LIV(3-4): 231–252. Huttar, George L. 1989. The Portuguese contribution to the Ndjuka lexicon. In Estudios sobre español de América y lingüística Afroamericana, 263-79. Bogotá: Instituto Caro y Cuervo. Intumbo, Incanha 2006. Papiamentu, Guiné-Bissau Creole Portuguese, and its substrate, Balanta: a comparison of the noun phrase. Revista internacional de lingüística iberoamericana 4(1). 107–115. Intumbo, Incanha 2007. Estudo comparativo da morfossintaxe do crioulo guineense, do balanta e do português. Munich: Lincom Europa. Intumbo, Incanha 2008. Balanta, Guiné-Bissau Creole Portuguese and Portuguese: A comparison of the noun phrase. In Suzanne Michaelis (ed.), Roots of creole structures 263–278. Amsterdam: Benjamins. Inverno, Liliana Forthcoming. A transição de Angola para o português vernáculo: estudo morfossintáctico do sintagma nominal. InAna Carvalho (ed.), Português em contacto. Madrid/Frankfurt: Iberoamericana/Vervuert. Israel, Jonathan 2009. Jews and Crypto-Jews in the Atlantic World Systems, 1500–1800. In Richard L. Kagan and Philip D. Morgan (eds.),Atlantic Diasporas Jews, Conversos, and Crypto-Jews in the Age of Mercantilism, 1500–1800, 3–17. Baltimore: The Johns Hopkins University Press. Jacobs, Bart 2010. Upper Guinea Creole: Evidence in favor of a Santiago birth. Journal of Pidgin and Creole Languages 25(2): 289–343. Jansen, Alphons L.M.J. 1911. Historia di Testament Bieuw, 1. Lectura Cristiana 3. Curaçao: Vicariato Apostólico. Janson, Tore 1984. Articles and plural formation in creoles: Change and universals. Lingua 64(4): 291–323. Jesurun, Abraham 1897. Eenige beschouwingen over de volkstaal op Curaçao. In Eerste jaarlijksch verslag van het Geschied-, Taal-, Land- en Volkenkundig Genootschap gevestigd te Willemstad, Curaçao, 95–110. Amsterdam: J.H. De Bussy. Jeuda, David M. 1990. Papiamentu Genesis. Unpublished manuscript. Joubert, Sidney M 1999. Handwoordenboek Nederlands-Papiaments. Zutphen: Koninklijke Wöhrmann. Joubert, Sidney and Matthias Perl 2007. The Portuguese Language on Curaçao and Its Role in the Formation of Papiamentu. Journal of Caribbean Literatures 5(1): 43–60.

362

References

Karner, Frances P. 1969. The Sephardics of Curaçao. A study of socio-cultural patterns in flux. Assen: Van Gorcum & Comp. Karttunen, Frances E. and James Lockhart 1976. Nahuatl in the middle years: language contact phenomena in texts of the colonial period. Berkeley: University of California Press. Kerkhof, Maxim P.A.M. 2003. Over het Portugees en het Spaans van de Sefardische Joden in Amsterdam in de 17e, 18e en 19e eeuw. Maastricht: Shaker. Kester, Ellen-Petra and Cristina Schmitt 2007. Papiamentu and Brazilian Portuguese: A comparative study of bare nominals. In Marlyse Baptista and Jacqueline Guéron (eds.), Noun Phrases in Creole Languages: A multi-faceted approach, 107–143. Amsterdam: Benjamins. Kihm, Alain 1980. Aspects d’une syntaxe historique: études sur le créole portugais de Guinée-Bissau. Thèse de troisième cycle, Université Paris III. Kihm, Alain 1989. Lexical Conflation as a Basis for Relexification. Canadian Journal of Linguistics 34(3): 351–376. Kihm, Alain 1994. Kriyol Syntax. The Portuguese-Based Creole Language of GuineaBissau. Amsterdam: Benjamins. Kihm, Alain 1996. Reflexivity in Kriyol: A case of half-hearted grammaticalization. In Philip Baker and Anand Syea (eds.), Changing meanings, changing functions: papers relating to grammaticalization in contact languages, 219–223. London: University of Westminster Press. Kihm, Alain 2003. Inflectional categories in creole languages. In Ingo Plag (ed.), Phonology and Morphology in Creole Languages, 333–363. Tübingen : Max Niemeyer Verlag. Kihm, Alain 2007. “être” créole: pourquoi la copule est-elle visible dans les langues créoles? In Karl Gadelii and Anne Zribi-Hertz (eds.), Grammaires créoles et grammaire comparative, 237–276. Saint-Denis: Presse Universitaire de Vincenne. Kleinmoedig-Eustatia, Adriana 1982. Mi koto di kuenta. Willemstad: Salas. Klooster, Willem 1998.. Illicit Riches. DutchTrade in the Caribbean, 1648–1795. Leiden: KITLV Press. Klooster, Willem 2001. The Jews in Suriname and Curaçao. In Paolo Bernardini and Norman Fiering (eds.), The Jews and the Expansion of Europe to the West, 1450– 1800, 350–369. New York/Oxford: Berghahn Books. Klooster, Willem 2003. Curaçao and the Caribbean TransitTrade. In Johannes Postma and Victor Enthoven (eds.), Riches From Atlantic Commerce, 365–384. Leiden/Boston: Brill. Klooster, Willem 2005. The Rise and Transformation of the Atlantic World. In Willem Klooster and Alfred Padula (eds.), The Atlantic World, 1–42. Upper Saddle River: Pearson. Klooster, Willem 2006. Communities of port Jews and their contacts in the DutchAtlantic World. Jewish History 20: 129–145. Klooster, Willem 2009. Networks of Colonial Entrepreneurs: The Founders of the Jewish Settlements in Dutch America, 1650s and 1660s. In Richard L. Kagan and Philip

References

363

D. Morgan (eds.), Atlantic Diasporas: Jews, Conversos, and Crypto-Jews in the Age of Mercantilism, 1500–1800, 31–49, 226–236. Baltimore: The Johns Hopkins University Press. Kouwenberg, Silvia 1990. Complementizer pa, the finiteness of its complements, and some remarks on empty categories in Papiamento. Journal of Pidgin and Creole Languages 5(1): 39–51. Kouwenberg, Silvia and Eric Murray 1994. Papiamentu. Munich: Lincom. Kouwenberg, Silvia and Pieter Muysken 1995. Papiamento. In Jacques Arends, Pieter Muysken and Norval Smith (eds.), Pidgins and Creoles: An Introduction, 205–218. Amsterdam: Benjamins. Kouwenberg, Silvia and Claire Lefebvre 2007. A new analysis of the Papiamentu clause structure. Probus 19: 37–73. Kouwenberg, Silvia and Abigail Ramos-Michel 2007. Papiamentu (Creole Spanish/Portuguese). In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 307–332. London: Battlebridge. Kowallik, Sabine and Kramer, Johannes 1994. Influencias del neerlandés en el papiamento. In Jens Lüdtke and Matthias Perl (eds.), Sprache und Kultur in der Karibik, 147–165. Tübingen: Niemeyer. Kramer, Johannes 2004. Die iberoromanische Kreolsprache Papiamento. Hamburg: Helmut Buske. Kramer, Johannes 2005. Indianische Wörter im Papiamentu. In Volker Noll and Haralambos Symeonidis (eds.), Sprache in Iberoamerika, 153–161. Hamburg: Buske. Kramer, Johannes 2008. Die früheste Erwähnung und das früheste Wörterbuch des Papiamento. Romanistik in Geschichte und Gegenwart 14(1): 99–114. Kriegel, Sibylle 2005. Contact de langues en papiamento: deux exemples de la syntaxe. Travaux interdisciplinaires du Laboratoire parole et langage d’Aix-en-Provence 24: 85–99. Kriegel, Sibylle 2006. Grammaticalisation: Dans quelle langue? Quelques exemples des langues créoles à base romane. In G. Lazard and C. Moyse Faurie (eds.), Linguistique typologique, 127–140. Villeneuve d’Ascq: Septentrion. Kuiperi, N.A. 1862. Katekismoe of sienjansa nan die religioon di kriestiaan nan pa oesoe di protestant nan na Aruba. Curaçao: Muller/Neuman. Krug, Manfred G. 2001. Frequency, iconicity, categorization: Evidence from emerging modals. In Joan Bybee and Paul Hopper (eds.), Frequency and the emergence of linguistic structure, 309–336. Amsterdam: Benjamins. Labov, William [1971]1990. On the adequacy of natural languages I: the development of tense. In John V. Singler (ed.), Pidgin and Creole Tense-Mood-Aspect Systems, 1–58. Amsterdam: Benjamins. Laclé, Digna, Ernesto Rosenstand and Ramon Todd Dandaré (eds.) 1983. Contami un storia: Un coleccion di e storianan ganador di e Concurso “Storia pa nos muchanan” teni na aña 1981. Aruba: Fundacion pa comunicacion grafico/Beatrixfonds. Ladhams, John 1999a. The Pernambuco connection? An examination of the nature and origin of the Portuguese elements in the Surinam Creoles. In Magnus Huber and

364

References

Mikael Parkvall (eds.), Spreading the word. The issue of diffusion among the Atlantic Creoles, 209–240. London: University of Westminster Press. Ladhams, John 1999b. Response to Norval Smith. In Magnus Huber and Mikael Parkvall (eds.), Spreading the word. The issue of diffusion among the Atlantic Creoles, 299– 304. London: University of Westminster Press. Ladhams, John 2006. In search of West African Pidgin Portuguese. Revista Internacional de Lingüística Iberoamericana 1(7): 87–105. Lampe, Armando 2001. Mission or Submission. Moravian and Catholic Missionaries in the Dutch Caribbean during the 19th Century. Göttingen: Vandenhoeck-Ruprecht. Lang, George 2000. Entwisted Tongues: Comparative Creole Literatures. Amsterdam: Rodopi. Lang, Jürgen 1993. Das Verbalsystem des kapverdischen Kreols (Variante von Santiago). In Matthias Perl, Axel Schönberger and Petra Thiele (eds.), Portugiesisch-basierte Kreolsprachen, 137–166. Frankfurt am Main: TFM/Domus Editoria Europaea. Lang, Jürgen 2000. Kurze Skizze der Grammatik des Kreols von Santiago (Kapverde). Neue Romania 23: 15–43. Lang, Jürgen 2001. La Problématique de l’origine du créole de Santiago (Cap-Vert). In Françoise Massa (ed.), Les îles atlantiques: réalités et imaginaire, 181–187. Université de Haute Bretagne Rennes 2: Équipe ERILAR. Lang, Jürgen 2002. Dicionário do crioulo da ilha de Santiago (Cabo Verde). Tübingen: Gunter Narr. Lang, Jürgen 2004. Affinités grammaticales entre le créole portugais de l’Île de Santiago (Cap-Vert) et le wolof (Sénégal). Aspect et temps. In Mauro Fernández, Manuel Fernández-Ferreiro, Manuel Veiga and Nancy Vázquez (eds.), Los criollos de base ibérica, 137–147. Madrid/Frankfurt am Main: Iberoamericana/Vervuert. Lang, Jürgen 2005. Zwei Fälle, in denen das portugiesische Kreol von Santiago (Kapverde) mit dem Wolof übereinstimmt. In Annette Endruschat and Axel Schönberger (eds.), Portugiesische Kreolsprachen: Entstehung, Entwicklung, Ausbau und Verwendung, 45–56. Frankfurt am Main: Domus Editoria Europaea. Lang, Jürgen 2006. L’influence des Wolof et du wolof sur la formation du créole santiagais. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria João Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 53–62. Tübingen: Gunter Narr. Lang, Jürgen 2009. Les langues des autres dans la créolisation. Théorie et exemplification par le créole d’empreinte wolof à l’île Santiago du Cap Vert. Tübingen: Gunter Narr. Lang, Jürgen 2009. A Wolof trace in the verbal system of the Portuguese Creole of Santiago island (Cape Verde). In Claire Lefebvre (ed.), Creoles, their substrates, and language typology, 61–80. Amsterdam: Benjamins. Lapesa, Rafael 1970. Personas gramaticales y tratamientos en español. In Homenaje a Menéndez Pidal, IV. Revista de la Universidad de Madrid 19: 141–167. Latour, Mamertus Dominicus 1936. Vreemde invloeden in het Papiamento. De WestIndische Gids 17: 387–396.

References

365

Latour, Mamertus Dominicus 1937. Portugese taalresten in het Papiamento. De WestIndische Gids 19: 212–214. Latour, Mamertus Dominicus 1940. Het Papiamento. De West-Indische Gids 22: 220– 225. Lauffer, Pierre 1971. Di nos: antologia di nos literatura. Curaçao: Libreria Salas. Lefebvre, Claire 1986. Relexification in creole genesis revisited: the case of Haitian Creole. In Pieter Muysken and Norval Smith (eds.), Substrata Versus Universals in Creole Genesis, 279–301. Amsterdam: Benjamins. Lefebvre, Claire 1998. Creole genesis and the acquisition of grammar: the case of Haitian Creole. Cambridge: Cambridge University Press. Lefebvre, Claire and Isabelle Therrien 2007a. The multiple facets of Papiamentu ku. Creolica. http://www.creolica.net/article.php3?id_article=58 Lefebvre, Claire and Isabelle Therrien 2007b. On the properties of Papiamentu pa: Synchronic and diachronic perspectives. In Magnus Huber and Viveka Velupillai (eds.), Synchronic and diachronic perspectives on contact languages, 215–255.Amsterdam: Benjamins. Lefebvre, Claire and Isabelle Therrien 2007c. On Papiamentu ku. In Terry Crowley, Jeff Siegel, John Dominic Lynch and Diana Eades (eds.), Language description, history and development: linguistic indulgence in memory of Terry Crowley, 169– 182. Amsterdam: Benjamins. Lenz, Rodolfo 1928. El papiamento: la lengua criolla de Curazao. Santiago de Chile: Balcells & Cia. Lessa, Orígenes 1975. Presença do português no papiamento. Rio de Janeiro: Cátedra. Liba, Moshé 2008. Contribución sefardí al idioma papiamentu. Maguén-Escudo 147: 30–35. Lima, Humberto 2000. Un bes tinha nho Lobu ku Tubinhu. Praia: Instituto Nacional de Investigação Cultural. Lipski, John M. 1986. A test case of the Afro-Hispanic connection. Syllable-final /s/ in Equatorial Guinea. Lingua 68: 209–222. Lipski, John M. 1987a. On the origin and use of lan/nan in Caribbean bozal Spanish. Beiträge zur romanischen Philologie 24: 281–290. Lipski, John M. 1987b.. Modern Spanish once-removed in Philippine Creole Spanish: The case of Zamboangueno. Language in Society 16(1): 91–108. Lipski, John 1988. Philippine creole Spanish: assessing the Portuguese element. Zeitschrift für Romanische Philologie 104: 25–45. Lipski, John M. 1989. The speech of the negros congos of Panama. Amsterdam: Benjamins. Lipski, John M. 1991. On the emergence of (a)mí as subject in Afro-Iberian pidgins and creoles. In Ray Harris-Northall and Thomas Cravens (eds.), Linguistic Studies in Medieval Spanish, 39–61. Madison: Hispanic Seminary of Medieval Studies. Lipski, John M. 1993. Origin and development of ta in Afro-Hispanic creoles. In Francis Byrne and John A. Holm (eds.), Atlantic meets Pacific: a global view of pidgnization and creolization, 217–231. Amsterdam: Benjamins.

366

References

Lipski, John M. 2000. Bozal Spanish: Restructuring or creolization? In Edgar Schneider and Ingrid Neumann-Holzschuh (eds.), Degrees of restructuring in Creole Languages, 437–468. Amsterdam: Benjamins. Lipski, John M. 2002. Génesis y evolución de la cópula en los criollos afroibéricos. In Yves Moñino and Armin Schwegler (eds.), Palenque, Cartagena y AfroCaribe, 65–101. Tübingen: Max Niemeyer. Lipski, John M. 2005. A History of AfroHispanic Language. Five Centuries, Five Continents. New York: Cambridge University Press. Lipski, John M. 2008. Spanish-Based Creoles in the Caribbean. In Silvia Kouwenberg and John Victor Singler (eds.), The Handbook of Pidgin and Creole Studies, 543–564. Oxford: Wiley-Blackwell. Lipski, John M. and Armin Schwegler 1993. Creole Spanish and Afro-Hispanic. In John Green and Rebecca Posner (eds.), Bilingualism and linguistic conflict in Romance, 407–432. Berlin: Mouton De Gruyter. Lipski, John M. and Maurizio Santoro 2007. Zamboangueño Creole Spanish. In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 373–398. London: Battlebridge. Lobban, Richard 1996. Jews in Cape Verde and on the Guinea Coast. Paper presented at the University of Massachusetts-Dartmouth, 11 February 1996. http://www.saudades.org/jewscapev.html. Accessed on 14-08-2010. Lopes da Silva, Baltasar 1957. O dialecto crioulo de Cabo Verde. Lisbon: National Press. Lorenzino, Gerardo A. 1998. The Angolar Creole Portuguese of São Tomé (West Africa): Its grammar and sociolinguistic history. New York: Ann Arbor. Lorenzino, Gerardo A. 2007. Angolar (Creole Portuguese), or Lunga Ngola. In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 1–24. London: Battlebridge. Luís,Ana R. 2008. The loss and survival of inflectional morphology: contextual vs. Inherent inflection in creoles. In Sonia Colina, Antxon Olarrea and Ana Carvalho (eds.) Selected Proceedings of the 39th Linguistic Symposium on Romance Languages, 323–336. Amsterdam: Benjamins. Lumsden, John S. 1991. On the aquisition of nominal structures in the genesis of Haitian Creole. Travaux de recherche sur le creole haitien 4: 37–56. Macedo, Donaldo Pereira 1979. A linguistic approach to the Capeverdean language. Ann Arbor, Michigan: University Microfilms. Maduro, Antoine J. 1953. Ensayo pa yega na un ortografia uniforme pa nos Papiamentu. Curaçao: Scherpenheuvel. Maduro, Antoine J. 1965. Papiamentu, origen i formación. Curaçao: author’s edition. Maduro, Antoine J. 1966a. Procedencia di Palabranan Papiamentu i otro anotacionnan. I (Letter A te M). Curaçao: author’s edition. Maduro, Antoine J. 1966b. Procedencia di Palabranan Papiamentu i otro anotacionnan. II (Letter N te ZJ). Curaçao: author’s edition. Maduro, Antoine J. 1967. Observacion- i apuntenan tocante ‘El Papiamento. La lengua criolla de Curazao […]’ di Dr. R. Lenz. Curaçao: author’s edition.

References

367

Maduro, Antoine J. 1969. Algun observashon tocante e articulo ‘origenes y evolucion del Papiamentu’ publicá den e numeronan 2, 3 i 4 di Watapana. Watapana 2(5): 13–15. Maduro, Antoine J. 1971. Bon papiamentu (i un appendix interesante). Curaçao: author’s edition. Maduro, Antoine J. 1987a. Kaboberdiano i Papiamentu. Curaçao: Maduro & Curiel’s Bank. Maduro, Antoine J. 1987b. Palenkero i Papiamentu. Curaçao: Maduro & Curiel’s Bank. Maduro, Antoine J. 1991. Papiamentu: Indagando i Ilustrando. Curaçao: Scherpenheuvel. Major, Ayres Veríssimo 2006. Dicionário Lexical. Santomé – Portugués. São Tomé: Ministerio da Educaçao e Cultura. Mansur, Jossy M. 1994. Historia di Aruba: 1499–1824. Miami: Hallmark Press Inc. Mark, Peter 2002. ‘Portuguese’Style and Luso-African Identity. Precolonial Senegambia, Sixteenth-Nineteenth Centuries. Indiana: University Press. Mark, Peter 2007. Towards a reassessment of the dating of the trading and the geographical origins of the Luso-African Ivories, 15th–17th centuries. History in Africa 34: 189–211. Mark, Peter and José da Silva Horta 2004. Two Early Sephardic Communities on Senegal’s Petite Côte. History in Africa 31: 231–256. Mark, Peter and José da Silva Horta 2008. Catholics, Jews, and Muslims in Early Seventeenth Century Guiné. In Richard L. Kagan and Philip D. Morgan (eds.), Atlantic Diasporas. Jews, Conversos, and Crypto-Jews in the Age of Mercantilism, 1500– 1800, 170–194. Baltimore: The Johns Hopkins University Press. Mark, Peter and José da Silva Horta 2011. The Forgotten Diaspora: Jewish Communities in West Africa And The Making Of The Atlantic World. Cambridge: Cambridge University Press. Markey, Thomas and Fodale, Peter 1983. Lexical Diathesis, Focal Shifts and Passivization: The Creole Voice. English World-Wide 4(1): 69–84. Marks, A.F. 1976. Male and female and the Afro-Curaçaoan household. The Hague: Nijhoff. Martinus, Efraim Frank. 1980. The Antillean Guinea creole: its practical and theoretical consequences for better understanding Papiamentu and other Portuguese-based creoles. Paper presented before the Society for Caribbean Linguistics, Third Biennial Conference. Martinus, Efraim Frank 1996. The Kiss of a Slave: Papiamentu’s West-African Connections. Amsterdam: Universiteit van Amsterdam dissertation. Martinus, Frank Martinus 1998. Victory of the concubines and the nannies. In Kathleen M. Balutansky and Marie-Agnès Sourieau (eds.), Caribbean Creolization: Reflections on the Cultural Dynamics of Language, Literature, and Identity, 110–117. Florida: University of Florida Press. Martinus, Efraim Frank 2003. The origin of the adjectival participle in Papiamentu. In Klaus Zimmermann (ed.), Lenguas criollas de base lexical española y portuguesa, 231–250. Madrid/Frankfurt am Main: Iberoamericana/Vervuert.

368

References

Martinus, Efraim Frank 2007. Djuku Kaiman ‘Sklabitut a kaba’: orígen i evolushon di papiamentu ku guene. Curaçao: Kas di Kultura. Martinus, Efraim Frank 2008. And Guene is… well, what exactly? In Nicolas Faraclas, Ronnie Severing and Christa Weijer (eds.), Linguistic Studies on Papiamentu, 157– 170. Curaçao: Fundashon pa Planifikashon di Idioma. Matras, Yaron (ed.) 1995. Romani in contact: the history, structure and sociology of a language. Amsterdam: Benjamins. Matras,Yaron 2000. Mixed languages:A functional-communicative approach. Bilingualism: Language and Cognition 3: 79–99. Matras, Yaron 2003. Mixed languages: Re-examining the structural prototype. In Peter Bakker and Yaron Matras (eds.), The Mixed Language Debate. Theoretical and Empirical Advances, 151–176. Berlin/New York: Mouton De Gruyter. Matras, Yaron 2007. The borrowability of structural categories. In Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross-Linguistic Perspective, 31– 74. Berlin/New York: Mouton De Gruyter. Matras, Yaron 2009. Language Contact. Cambridge: Cambridge University Press. Maurer, Philippe. 1985.. Le système temporel du papiamento et le système temporel proto-créole de Bickerton. Amsterdam Creole Studies 8: 41–66. Maurer, Philippe 1986a. El origen del papiamento desde el punto de vista de sus tiempos gramaticales. Neue Romania 4: 129–149. Maurer, Philippe 1986b. Le papiamento de Curaçao: un cas de creolisation atypique? Études creoles 9(1): 97–113. Maurer, Philippe 1986c. Los verbos modales por, mester y ke del papiamento: un caso de transparencia semántica? In Norbert Boretzky, Werner Enninger and Thomas Stolz (eds.) Akten des 2. Essener Kolloquiums über “Kreolsprachen und Sprachkontakte”, 135–156. Bochum: Brockmeyer. Maurer, Philippe 1987a. La comparación des morphèmes temporels du papiamento et du palenquero: arguments contre la théorie monogénétique de la genèse des langues créoles. In Philippe Maurer and Thomas Stolz (eds.), Varia Creolica, 27–70. Bochum: Brockmeyer. Maurer, Philippe 1987b. Substrate Influence on the Semantics of the Papiamentu Particle di: A Reply to Bickerton 1986. Journal of Pidgin and Creole Languages 2(2): 239– 249. Maurer, Philippe 1988. Les modifications temporelles et modales du verbe dans le papiamento de Curaçao (Antilles Néerlandaises). Hamburg: Helmut Buske. Maurer, Philippe1989. Les reiterations et reduplications lexicalisees du papiamento: influence du substrat africain? In Norbert Boretzky, Werner Enninger and Thomas Stolz (eds.), Beiträge zum 5. Essener Kolloquium über Kreolsprachen und Sprachkontakte, 95–118. Bochum: Brockmeyer. Maurer, Philippe 1991. Der Einfluss afrikanischer Sprachen auf die Wortsemantik des Papiamentu. In Norbert Boretzky, Werner Enninger and Thomas Stolz (eds.), Beiträge zum 6. Essener Kolloquium über Kreolsprachen und Sprachkontakte, 123–137. Bochum: Brockmeyer.

References

369

Maurer, Philippe 1994a. L’influence des langues africaines sur la signification des lexèmes du papiamento de Curaçao. In Ramón Lorenzo (ed.), Actas do XIX Congreso Internacional de Lingüística e Filoloxía Románicas, 771–779. Universidad de Santiago de Compostela. Maurer, Philippe 1994b. Tense-Aspect-Mood in Principense. Journal of Pidgin and Creole Languages 19: 415–435. Maurer, Philippe 1995. L’angolar. Un créole afroportugais parlé à São Tomé. Hamburg: Helmut Buske. Maurer, Philippe 1998. El papiamentu de Curazao. In Matthias Perl and Armin Schwegler (eds.), América Negra: panorámica actual de los estudios lingüísticos sobre variedades hispanas, portuguesas y criollas, 139–217. Frankfurt am Main: Vervuert. Maurer, Philippe 2002. La pluralización nominal en los criollos atlánticos con base léxica ibérica. In Yves Moñino and Armin Schwegler (eds.), Palenque, Cartagena y AfroCaribe, 129–136. Tübingen: Max Niemeyer. Maurer, Philippe 2003a. Procès de grammaticalisation dans le systèsme TAM du papiamento de Curaçao. In Sibylle Kriegel (ed.), Grammaticalisation et réanalyse. Approches de la variation créole et française, 233–247. Paris: CNRS Éditions. Maurer, Philippe 2003b. Les marques TAM du créole portugais de Batavia (Indonésie). Creolica. http://www.creolica.net/Les-marques-TAM-du-creole Maurer, Philippe 2005. La préposition for di du papiamento et le verbe fô des créoles du golfe de Guinée. La Linguistique 41(1): 57–66. Maurer, Philippe 2009a. Principense. Grammar, Texts, and Vocabulary of the AfroPortuguese Creole of the Island of Príncipe, Gulf of Guinea. London: Battlebridge. Maurer, Philippe 2009b. A short note about Papiamentu’s origin: Upper Guinea, Gulf of Guinea, or elsewhere? In Nicholas Faraclas, Ronald Severing, Christa Weijer and Liesbeth Echteld (eds.), Leeward voices: Fresh perspectives on Papiamentu and the literatures and cultures of the ABC Islands, vol. 1, 39–47. Curaçao: Fundashon pa Planifikashon di Idioma/University of the Netherlands Antilles. Mbodj, Chérif 1991. Esboço do verbo crioulo. Papia 1(2): 68–79. MBodj, Chérif 1993. Esquisse du verbe kriol de Guinée-Bissau. In Carlos Lopes (ed.), Mansas, escravos, grumetes e gentio, 325–342. Bissau, Guinea-Bissau: INEP. McConvell, Patrick and Felicity Meakins 2005. Gurindji Kriol. A Mixed Language emerges from Code switching. Australian Journal of Linguistics 25(1): 9–30. McWhorter, John 1995. Sisters under the skin: a case for genetic relationship between the Atlantic English-based creoles. Journal of Pidgin and Creole Languages 10(2): 289–333. McWhorter, John 1997. Towards a new model of creole genesis. New York: Peter Lang. McWorther, John 2000. The Missing Spanish Creoles: Recovering The Birth Of Plantation Contact Languages. Berkeley: University of California Press. McWhorter, John 2001. The world’s simplest grammars are creole grammars. Linguistic typology 5(3): 125–156. McWhorter, John 2005. Defining creole. Oxford: Oxford University Press.

370

References

McWhorter, John 2006. Creole transplantations: a source of solutions to resistant anomalies. In J. Clancy Clements (ed.), History, society and variation. In honor of Albert Valdman, 103–134. Amsterdam: Benjamins. McWhorter, John 2011. Linguistic Simplicity and Complexity: Why Do Languages Undress? Berlin: Mouton De Gruyter. McWhorter John and Mikael Parkvall 2002. Pas tout à fait du français: Une étude créole. Études Créoles 25(1): 179–231. Meeteren, Nicolaas van 1947. Volkskunde van Curaçao. Curaçao: Scherpenheuvel. Megenney, William W. 1983. La influencia del portugués en el palenquero colombiano. Thesaurus 28: 548–563. Megenney, William W. 1984. Traces of Portuguese in Three Caribbean Creoles: Evidence in Support of the Monogenetic Theory. Hispanic Linguistics 1(2): 177–189. Megenney, William W. 1985. La influencia criollo-portuguesa en el español caribeño. Anuario de Lingüística Hispánica 1: 157–180. Megenney, William W. 1993. Elementos criollo-portugueses en el español dominicano. Montalbán 25: 149–171. Meintel, Deirdre 1975. The Creole Dialect of the Island Brava. In Marius Valkhoff (ed.), Miscelânia luso-africana: colectânea de estudos, 205–256. Lisbon: Junta de Investigações Científicas do Ultramar. Mendes, Mafalda, Nicolas Quint, Fátima Ragageles and Aires Semedo 2002. Dicionário Práctico Português-Caboverdiano. Lisbon: Verbalis. Mendes, António de Almeida 2005. Elements pour une histoire socio-economique de la “Guinée du Cap-Vert” aux XVe-XVIIe siècles. In Le Portugal et l’Espagne dans leurs rapports avec les afriques continentale et insulaire, 229–237. Université de Haute Bretagne Rennes 2: Équipe ERILAR. Mendy, Karibe 1993. The tradition of resistance in Guinea-Bissau: the PortugueseAfrican encounter in Cacheu, Bissau and “suas dependências”, 1588–1878. In Carlos Lopes (ed.), Mansas, escravos, grumetes e gentio, 135–169. Guinea-Bissau: INEP. Menkman, Willem R. 1936. Curaçao, zijn naam en zijn taal. De West-Indische gids 18: 38–50. Menkman, Willem R. 1940. Stem uit de redactie. De West-Indische Gids 22: 374–376. Montenegro, Teresa and Carlos de Morais 1995. Uori: stórias de lama e philosophia. Guinea-Bissau: Ku Si Mon Editora. Moñino, Yves and Armin Schwegler (eds.) 2002. Palenque, Cartagena y AfroCaribe. Tübingen: Max Niemeyer. Moraes, Nize Isabel de 1969. Sur les prises de Gorée par les Portugais au XVIIe siècle. Bulletin de l’IFAN. T.XXXI(4): 989–1013. Moraes, Nize Isabel de 1972. Le commerce des peaux à la Petite Côte au XVIIe siècle (Sénégal). Notes Africaines 134: 37–45 & 136: 111–116. Moraes, Nize Isabel de 1973. Le commerce des tissus à la Petite au XVIIe siècle (Sénégambie). Notes Africaines 39: 71–75. Moraes, Nize Izabel de 1993. À la Découverte de la Petite Côte au XVIIe Siècle (Sénégal et Gambie). Vol 1. Dakar: Université Cheikh Anta Diop/IFAN.

References

371

Moraes, Nize Izabel de 1995. À la Découverte de la Petite Côte au XVIIe Siècle (Sénégal et Gambie). Vol 2. Dakar: Université Cheikh Anta Diop/IFAN. Moraes, Nize Izabel de 1998a. À la Découverte de la Petite Côte au XVIIe Siècle (Sénégal et Gambie). Vol 3. Dakar: Université Cheikh Anta Diop/IFAN. Moraes, Nize Izabel de 1998b. À la Découverte de la Petite Côte au XVIIe Siècle (Sénégal et Gambie). Vol 4. Dakar: Université Cheikh Anta Diop/IFAN. Moraes, Nize Izabel de 1998c. La traite des esclaves effectuée par les Hollandais à Gorée (1637–1677). In Djibril Samb (ed.), Gorée et l’esclavage, 41–52. Dakar: Université Cheikh Anta Diop/IFAN. Moraes, Nize Izabel de 1998d. Les puissances européennes, Gorée et la Petite Côte aux XVIe et XVIIe siècle. In Djibril Samb (ed.), Gorée et l’esclavage, 53–62. Dakar: Université Cheikh Anta Diop/IFAN. Moraes, Nize Izabel de and Guy Thilmans 1973. Jan Kompany, notable de la baie de Hann. (1664–1697). Notes Africaines 140: 101–105. Morais-Barbosa, Jorge (ed.) 1967. Estudos linguísticos crioulos. Lisbon: AICP. Morais-Barbosa, Jorge 1975. Cape Verde, Guinea-Bissau and Sao Tomé e Príncipe: the linguistic situation. In Marius F. Valkhoff (ed.), Miscelânia luso-africana: colectânea de estudos, 133–151. Lisbon: Junta de Investigações Científicas do Ultramar. Mota, Avelino Teixeira da and P.E.H. Hair 1977. Descrição da Serra Leoa e dos Rios de Guiné do Cabo Verde (1625), por André Donelha. Lisbon: Junta de Investigações Científicas do Ultramar. Mous, Maarten 2003. The linguistic properties of lexical manipulation and its relevance for Ma’á. In Peter Bakker and Yaron Matras (eds.), The Mixed Language Debate. Theoretical and Empirical Advances, 209–236. Berlin: Mouton De Gruyter. Mufwene, Salikoko 1996. The Founder Principle in Creole Genesis. Diachronica 13(1): 83–134. Mufwene, Salikoko 2000. Creolization is a social, not a structural, proces. In Edgar Schneider and Ingrid Neumann-Holzschuh (eds.), Degrees of restructuring in Creole Languages, 65–84. Amsterdam: Benjamins. Mühlhäusler, Peter 1997. Pidgin and Creole Linguistics. London: University of Westminster Press. Muller, Enrique 1989. Inleiding tot de syntaxis van het Papiamentu. Ph.D. dissertation, University of Amsterdam. Munteanu, Dan 1974. Observaciones acerca del origen del papiamento.Anuario de Letras 12: 83–115. Munteanu, Dan 1991. El Papiamento, Origen, Evolucion y Estructura. Bochum: Brockmeyer. Munteanu, Dan 1992. Apuntes sobre el origen del papiamento. Anuario de Lingüística Hispánica 8: 189–199. Munteanu, Dan 1996a. El papiamento, lengua criolla hispánica. Madrid: Gredos. Munteanu, Dan 1996b. Papiamento. In Manuel Alvar (ed.), Manual de dialectología hispánica. El Español de América, 68–78. Barcelona: Ariel.

372

References

Munteanu, Dan 2004. El artículo definido y el pronombre demonstrativo papiamento y románico. In David Trotter (ed.), Actes du XXIVe Congrès International de Linguistique et Philologie Romanes, 525–532. Tübingen: Niemeyer. Munteanu, Dan 2007. Lenguas criollas de base románica. In José Enrique Gargallo Gil and Maria Reina Bastardas (eds.), Manual de lingüística románica, 427–451. Barcelona: Ariel. Muysken, Pieter 1981. Halfway between Quechua and Spanish: the case for relexification. In Arnold Highfield and Albert Valdman (eds.), Historicity and variation in Creole studies, 52–78. Ann Arbor: Karoma. Muysken, Pieter 1993. Reflexes of Ibero-Romance reflexive clitic + verb combinations in Papiamentu: thematic grids and grammatical relations. In Francis Byrne and Donald Winford (eds.), Focus and grammatical relations in creole languages, 285–302. Amsterdam: Benjamins. Muysken, Pieter 1997. Media Lengua. In Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective, 365–427. Amsterdam: Benjamins. Muysken, Pieter 2001. The creole languages of the Caribbean. In Albert James Arnold (ed.), A history of literature in the Caribbean, 399–414. Benjamins: Amsterdam. Muysken, Pieter 2008. Functional Categories. Cambridge: Cambridge University Press. Muysken, Pieter and Guus Meijer 1979. Introduction. In Thomas L. Markey and Paul T. Roberge (eds.), Hesseling, Dirk Christiaan: OnThe Origin and Formation of Creoles: A Miscellany of Articles. Ann Arbor: Karoma. Muysken, Pieter and Norval Smith 1990. Question words in pidgin and creole languages. Linguistics 28: 883–903. Muysken, Pieter and Norval Smith 1994. Reflexives in the creole languages: an interim report. In Dany Adone and Ingo Plag (eds.), Creolization and Language Change, 45–64. Tübingen: Max Niemeyer. Myers-Scotton, Carol 1993. Duelling Languages. Grammatical Structure in Codeswitching. Oxford: Clarendon Press. Nagelkerken, Wilhelmus Petrus 1980. Coral reef fishes of Aruba, Bonaire, and Curaçao (Netherlands Antilles). Curaçao: Island Territory of Curaçao. Naro, Anthony 1978. A study on the origins of pidginization. Language 54: 314–347. Narrog, Heiko 2008.. A note on modality and aspect in Saramaccan. In Werner Abraham and Elisabeth Leiss (eds.), Modality-aspect interfaces: implications and typological solutions, 359–370. Amsterdam: Benjamins. Nascimento, Augusto 2008. Cape Verdeans in São Tomé and Príncipe. In Luís Batalha and Jørgen Carling (eds.), Transnational Archipelago. Perspectives on Cape Verdean Migration and Diaspora, 55–60. Amsterdam: University of Amsterdam Press. Navarro Tomás, Tomás 1953. Observaciones sobre el papiamento. Nueva Revista de Filologia Hispanica 71: 183–189. Ndiaye, Moussa D. 2004. Éléments de morphologie du wolof. Munich: Lincom. Newitt, Malyn 2005. A history of Portuguese overseas expansion, 1400–1668. New York: Routledge.

References

373

Niewindt, Martinus Joannes [1833]2002. Prefecto Apostulico di Curaçao na Cristian di su Mision. Bloemendaal: Libri Antilliani. Niewindt, Martinus Joannes [1837]2001. Catecismo Corticu pa uso di catolicanan di Curaçao. Curaçao: Fundashon pa Planifikashon di Idioma. Osam, Kweku E. 2008. Akan as an aspectual language. In Felix Ameka and Mary Esther Kropp Dakubu (eds.), Aspect and modality in Kwa languages, 69–90. Amsterdam: Benjamins. Otheguy, Ricardo 1973. The Spanish Caribbean: A Creole Perspective. In Charles-James N. Bailey and Roger W. Shuy (eds.), New Ways of Analyzing Variation in English, 323–339. Washington: Georgetown University Press. Parkvall, Mikael 1999a. On the Possibility of Afrogenesis in the Case of French Creoles. In John R. Rickford and Suzanne Romaine (eds.), Creole Genesis, Attitudes and Discourse, 187–214. Amsterdam: Benjamins. Parkvall, Mikael 1999b. Feature selection and genetic relationships among Atlantic Creoles. In Magnus Huber and Mikael Parkvall (eds.), Spreading the word. The issue of diffusion among the Atlantic Creoles, 29–68. London: University of Westminster Press. Parkvall, Mikael 2000. Out of Africa. London: Battlebridge. Parkvall, Mikael 2001. Reassessing the role of demographics in language restructuring. In Ingrid Neumann-Holzschuh and Edgar W. Schneider (eds.), Degrees of Restructuring in Creole Languages, 185–213. Amsterdam: Benjamins. Parkvall, Mikael 2003. The sociolinguistics of young languages. In Joseph M. Brincat, Winfried Boeder and Thomas Stolz (eds.), Purism in Minor Languages, Endangered Languages, Regional Languages, Mixed Languages. Papers from the Conference on Purism in the Age of Globalisation, Bremen, September 2001, 141–154. Bochum: Diversitas Linguarum. Parkvall, Mikael 2005. Review of Bettina Migge. 2003. Creole Formation as Language Contact. The case of the Suriname Creoles. Studies in Language 29(3): 700–706. Parkvall, Mikael 2006. Was Haitian ever more like French? InAna Deumert and Stephanie Durrleman-Tame (eds.), Structure and Variation in Language Contact, 315–335. Amsterdam: Benjamins. Parkvall, Mikael 2008a. Which parts of language are the most stable? STUF 61(3): 234–250. Parkvall, Mikael 2008b. The simplicity of creoles in a cross-linguistic perspective. In Matti Miestamo, Kaius Sinnemäki and Fred Karlsson (eds.), Language Complexity: Typology, contact, change, 265–285. Amsterdam: Benjamins. Parkvall, Mikael and Stéphane Goyette Forthcoming. Why mixedness isn’t what it’s all about. Parsons, Elsie Clews 1923. Folk-Lore from the Cape Verde Islands. Cambridge, Massachusetts: The Cosmos Press. Patiño Rosselli, Carlos 2002. Sobre origen y composition del criollo palenquero. InYves Moñino and Armin Schwegler (eds.), Palenque, Cartagena y AfroCaribe, 21–34. Tübingen: Max Niemeyer.

374

References

Patrick, Peter L. 1999. Urban Jamaican Creole: Variation in the mesolect. Amsterdam: Benjamins. Patrick, Peter L. 2007. Jamaican Patwa (Creole English). In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 127–152. London: Battlebridge. Patterson, K. David 1988. Epidemics, Famines, and Population in the Cape Verde Islands, 1580-1900. The International Journal of African Historical Studies 21(2): 291–313. Peck, Stephen Madry, Jr. 1988. Tense, Aspect and Mood in Guinea-Casamance Portuguese Creole. Los Angeles: University of California dissertation. Penny, Ralph 2000. Variation and Change in Spanish. Cambridge: Cambridge University Press. Penny, Ralph 2002. A History of the Spanish Language. Cambridge: Cambridge University Press. Pereira, Daniel 1993. A fundação da companhia de Cacheu 1671–1676. In Carlos Lopes (ed.), Mansas, escravos, grunetes e gentiu: Canheu na encruzilhada de civilizações, 199–247. Guiné-Bissau: Instituto Nacional de Estudos e Pesquisa. Pereira, Dulce 1999. Les verbes auxiliaires en créole de Santiago du Cap-Vert: Études Créoles, XXII(1): 108–118. Pereira, Dulce 2000. Um crioulo de outro planeta. In Ernesto d’Andrade, Dulce Pereira and Maria Antónia Mota (eds.), Crioulos de Base Portuguesa, 27–46. Lisbon: Associacao Portuguesa de Linguistica. Pereira, Dulce 2008. As múltiplas semelhanças de ku em Papiamento e em Caboverdiano. Paper presented at the ACBLPE Annual Meeting, July 3–5, Lisbon. Peres, Damião 1953. Duas Descrições Seiscentistas da Guiné, de Francisco De Lemos Coelho (1669 & 1684). Lisbon: Academia Portuguesa da História. Perl, Matthias 1982. Die Bedeutung des Kreolenportugiesischen für die Herausbildung der Kreolensprachen in der Karibik (unter besonderer Berücksichtigung der kubanischen ‘habla bozal’). Leipzig: Universität Leipzig dissertation. Perl, Matthias 1989a. Algunos resultados de la comparación de fenómenos morfosintácticos del ‘habla bozal’, de la ‘linguagem dos musseques’, del ‘palenquero’ y de lenguas criollas de base portuguesa. In Estudios sobre español de América y lingüística afroamericana, 369–380. Bogotá: Instituto Caro y Cuervo. Perl, Matthias 1989b. Zur Präsenz des kreolisierten Portugiesisch in der Karibik – Ein Beitrag zur Dialektologie des karibischen Spanisch. Beiträge zur romanischen Philologie 28:131–148. Perl, Matthias 1990. A reevaluation of the importance of early pidgin/creole Portuguese. Journal of Pidgin and Creole Languages 5: 125–130. Plag, Ingo 1994. Creolization and language change. A comparison. In Dany Adone and Ingo Plag (eds.), Creolization and language change, 3–21. Tübingen: Niemeyer. Plag, Ingo (ed.) 2003. The morphology of creole languages. Special section in Yearbook of Morphology 2002. Dordrecht: Foris. Plag, Ingo 2005. Morphology in pidgins and creoles. In Keith Brown (ed.), Encyclopedia of Language and Linguistics, vol. 8, 304–308. Oxford: Elsevier.

References

375

Pijpers, Antoine 1898. Theoretische spraakkunst der Papiamentsche taal. Curaçao: Neuman. Pla Cárceles, José 1923. La evolución del tratamiento “Vuestra-Merced”. Revista de Filología Española 10: 245–280. Porras, Jorge 2006. A Time-based Semantic Analysis of Selected Afro-Iberian Creole Narratives. The Journal of Pan African Studies 1(3): 86–101. Portilla, Mario 2007. Reconstrucción fonológica del pidgin afroportugués americano: las vocales. Filología y Lingüística 33(2): 179–203. Portilla, Mario 2008a. Reconstrucción fonológica del pidgin afroportugués americano: las consonantes. Filología y Lingüística 34(1): 157–192. Portilla, Mario 2008b. Protopidgin afroportugués americano. Revista Internacional de Lingüística Iberoamericana 6(1): 163–185. Post, Marieke 2000. The use of generic and impersonal pronouns in Fa d’Ambô. In Ernesto d’Andrade, Dulce Pereira and Maria Antónia Mota (eds.), Crioulos de base portuguesa, 127–142. Lisbon: Associação Portuguesa de Linguística. Postma, Johannes 1974. The Origin ofAfrican Slaves: The Dutch Activities on the Guinea Coast, 1675–1795. In Stanley L. Engerman and Eugene D. Genovese (eds.), Race and Slavery in the Western Hemisphere: Quantitative Studies, 33–49. Princeton: Princeton University Press. Postma, Johannes M. 1990. The Dutch in the Atlantic Slave Trade 1660–1815. Cambridge: Cambridge University Press. Pratas, Fernanda 2007. Tense Features and Argument Structure in Capeverdean Predicates. Ph.D. dissertation, Universidade Nova de Lisboa. Pratas, Fernanda 2010. States and temporal interpretation in Capeverdean. In BokBennema, Reineke, Brigitte Kampers Manhe and Bart Hollebrandse (eds.), Romance Languages and Linguistic Theory, 215–232. Amsterdam: Benjamins. Putman, Jacobus Josephus 1849. Proeve eener Hollandsche spraakkunst, ten gebruike der algemeene armenschool. Curaçao: author’s edition. Putman, Jacobus Josephus [1850]2001. Kamiena di Kroes. Koe historia, meditasjon I orasjon kortiekoe. Bloemendaal: Stichting Libri Antilliani. Putman, Jacobus Josephus 1852. Hesoe Kriestoe, noos dibienoe, adorabel Salbadoor i Libradoor. Curaçao, St. Rosa: Imprenta di Vicariato. Putman, Jacobus Josephus 1953. Meditasjon Arieba Soefrimeentoe di noos senjoor Hesoe Kristoe. Curaçao, St. Rosa: Imprenta di Vicariato. Putte, Florimon van 1999. Dede pikiña ku su bisiña: Papiamentu-Nederlands en de onverwerkt verleden tijd. Zutphen: de Walburg Pers. Putte, van Florimon and Igma van Putte-de Windt 2002. Dòsplak ‘i boka. Basiscursus Papiaments. Zutphen: Walburg Pers. Putte, Florimon van and Igma van Putte-de Windt 2005a. Dikshonario PapiamentuHulandes. Zutphen: Walburg Pers. Putte, Florimon van and Igma van Putte-de Windt 2005b. Dikshonario HulandesPapiamentu. Zutphen: Walburg Pers.

376

References

Quesada Pacheco, Miguel Ángel 2000. El Español de América. Cartago: Editorial Tecnológica. Quint, Nicolas 2000a. Grammaire de la Langue Cap-Verdienne. Paris: L’Harmattan. Quint, Nicolas 2000b. Le Cap Verdien: Origines et Devenir d’une Langue Métisse. Paris: L’Harmattan. Quint, Nicolas 2001a. Langues Créoles, Diachronie et Procédés de Reconstruction. Bulletin de la Société de linguistique de Paris 1: 265–284. Quint, Nicolas 2001b. Vowels as a morphological tool in Santiago Creole Portuguese (Cape Verde). Journal of African Languages and Linguistics (22)1: 69–80. Quint, Nicolas 2003. Parlons Capverdien. Langue et culture. Paris: L’Harmattan. Quint, Nicolas 2005. Le créole capverdien de poche. Chennevières-sur-Marne: Éditions Assimil. Quint, Nicolas 2006. Un bref aperçu des racines africaines de la langue cap-verdienne. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria João Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 75–90. Tübingen: Narr. Quint, Nicolas 2008a. Coordination et parataxe en capverdien moderne (dialecte santiagais ou badiais). In Bernard Caron (ed.), Subordination, dépendance et parataxe dans les langues africaines, 29–48. Louvain/Paris: Peeters. Quint, Nicolas 2008b. L’élément africain dans la langue capverdienne / Africanismos na língua caboverdiana. Paris: L’Harmattan. Quint, Nicolas 2008c. Les apontamentos de António de Paula Brito (1887) ou la naissance d’une tradition grammaticale capverdienne autochtone. Histoire Épistémologie Language 30(1): 127–153. Quint, Nicolas 2009. Les formes tronquées en capverdien santiagais. Revue Roumaine de Linguistique LIV(3–4): 253–277. Quint, Nicolas 2010a. Review of Magnus Huber and Viveka Villupillai (eds.), Synchronic and Diachronic Perspectives on Contact Languages, 2007. Journal of Language Contact Varia 3: 136–154. Quint, Nicolas 2010b. Vamos Falar Caboverdiano, Língua e cultura. Paris: L’Harmattan. Quint, Nicolas Forthcoming. Atlas linguistique de la langue capverdienne. Ratzlaff, Betty 1992. Dikshonario Papiamentu – Ingles Dikshonario. Bonaire: TWA Dictionary Foundation. Rawley, James A. and Stephen D. Behrendt 2005. The Transatlantic Slave Trade: A History. Lincoln: University of Nebraska Press. Reinecke, John E. 1975. A Bibliography of Pidgin and Creole Languages. Hawaii: Hawaii University Press. Ribeiro da Silva, Filipa 2009. Dutch Labor Migration to West Africa (c.1590–1674). In Wim Klooster (ed.), Migration, Trade, and Slavery in an Expanding World. Leiden: Brill. Ribeiro da Silva, Filipa Forthcoming. Dutch and Portuguese in Western Africa: Empire Building and Atlantic Systems (1580–1674). Leiden. Brill. Richardson, Linda 1977. The phrasal conjunctor and the comitative marker in Papiamentu. Amsterdam Creole Studies 1: 55–68.

References

377

Roitman, Jessica Vance 2006. The Portuguese, the Dutch and the Sephardim: Networks and cross-cultural mediation in the Atlantic slave trade, 1580–1654. Paper presented at the African slave trades and African slavery in a global comparative perspective conference, Leiden, African Studies Centre, 18 May 2006. Rona, José Pedro 1971. Elementos españoles, portugueses y africanos en el papiamento. Watapana 3(3): 7–23. Rona, José Pedro 1976. Réhispanisation des langues créoles aux Antilles. Étude sur la divergence et la convergence. In Actes du XIIIe Congrès Internationale de Linguistique et Philologie Romanes tenu à l’Univesrsité Laval (Québec, Canada), 1015–1025. Rosalia, Rene Vicente 1983. Shon Kolá: Labariano di rasa: Mei siglo na kustia di barí. Curaçao: Sekshon di Kultura di Teritorio Insular di Kòrsou. Rosenstand, Ernesto E. 1970. Kuentanan Rubiano. Oranjestad: Rosenstand. Ross, Malcolm 2005. Pronouns as a preliminary diagnostic for grouping Papuan languages. In Andrew Pawley, Robert Attenborough, Jack Golson and Robin Hide (eds.), Papuan pasts: cultural, linguistic and biological histories of Papuan-speaking peoples, 15–65. Canberra: Australian National University Press. Rougé, Jean-Louis 1985. Formation et évolution du créole de Guinée Bissao et de Casamance. Ph.D. dissertation, Université de Lyon II. Rougé, Jean-Louis 1988. Petit dictionnaire etymologique du Kriol de Guinée-Bissau et Casamance. Bissau: INEP. Rougé, Jean-Louis 1993. Les 400 ans du Kriol de Cacheu. In Carlos Lopes (ed.), Mansas, escravos, grumetes e gentio: Cacheu na encruzilhada de civilizações, 317–324. Guinea-Bissau: INEP. Rougé, Jean-Louis 1994. A propos de la formation des créoles du Cap Vert et de Guinée. Papia 3(2): 137–149. Rougé, Jean-Louis 1999. Apontamentos sobre o léxico de origem africana dos crioulos da Guiné e de Cabo Verde (Santiago). In Klaus Zimmermann (ed.), Lenguas criollas de base lexical espanola y portuguesa, 49–67. Frankfurt am Main: Vervuert. Rougé, Jean-Louis 2004a. Dictionnaire étymologique des créoles portugais d’Afrique. Paris: Karthala. Rougé, Jean-Louis 2004b. Le Catéchisme Créole de Ziguinchor. In Mauro Fernández, Manuel Fernández-Ferreiro, Manuel Veiga and Nancy Vázquez (eds.), Los criollos de base ibérica, 149–158. Madrid/Frankfurt: Iberoamericana/Vervuert Rougé, Jean-Louis 2006. L’influence manding sur la formation des créoles du Cap-Vert et de Guinée-Bissau et Casamance. In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 63–74. London: Battlebridge. Rougé, Jean-Louis 2008. A inexistencia do crioulo no Brasil. In José Luiz Fiorin and Margarida Petter (eds.), Africa no Brasil, 63–73. São Paulo: Contexto. Salas, Marius 1976. Le judéo-espagnol. Den Haag: Mouton. Salmons, Joseph 1994. Naturalness and Morphological Change in Texas German. In Nina Berend and Klaus J. Mattheier (eds.), Sprachinselforschung. Eine Gedenkschrift für Hugo Jedig, 59–72. Frankfurt am Main: Peter Lang.

378

References

Salomon, H.P. 1982. The Earliest Known Document in Papiamentu Contextually Reconsidered. Neophilologus 66: 367–376. Sanchez, Tara 2000. The Papiamentu progressive: Syntactic conditioning and dialect variation. Master’s thesis, University of Pennsylvania. Sanchez, Tara 2005. Constraints on structural borrowing in a multilingual contact situation. Ph.D. dissertation, University of Pennsylvania. Sanchez, Tara 2006. Papiamentu. In Keith Brown, Anne H. Anderson and Laurie Bauer (eds.), Encyclopedia of language and linguistics, 157–158. Amsterdam: Elsevier. Sanchez, Tara 2008. Accountability in morphological borrowing: Analyzing a linguistic subsystem as a sociolinguistic variable. Language Variation and Change 20: 225– 253. Sarna, Jonathan D. 2001. The Jews in British America. In Paolo Bernardini and Norman Fiering (eds.), The Jews and the Expansion of Europe to the West, 1450–1800, 519– 532. New York/Oxford: Berghahn Books. Scantamburlo, Luigi 1981. Gramática e Dicionário da Língua Criol da Guinea-Bissau (Gcr). Bologna: Editrice Missionaria Italiana, 1981. Scantamburlo, Luigi 1999. Dicionário do Guineense. Introdução e notas gramaticais. Bubaque, Guinea-Bissau: Edições Colibri/FASPEBI. Scantamburlo, Luigi 2002. Dicionário do Guineense. Dicionário Guineense-Português. Bubaque, Guinea-Bissau: FASPEBI. Schadeberg, Thilo, C. 2009. Loanwords in Swahili. In Martin Haspelmath and Uri Tadmor. Loanwords in the World’s Languages: A Comparative Handbook, 76–102. Berlin: Mouton De Gruyter. Schmidt, Benjamin 2001. The Hope of the Netherlands: Menasseh Ben Israel and the Dutch Idea of America. In Paolo Bernardini and Norman Fiering (eds.), The Jews and the Expansion of Europe to the West, 1450–1800, 86–106. New York/Oxford: Berghahn Books. Schmidt, Christopher K. 2009. Loanwords in Japanese. In Martin Haspelmath and Uri Tadmor. Loanwords in the World’s Languages: A Comparative Handbook, 545–574. Berlin: Mouton De Gruyter. Schmidt-Riese, Roland 2009. Comparando os sistemas pronominais do português e de três línguas crioulas de base portuguesa. Paper presented at the VIII. Congreso dos Lusitanistas Alemães, Munich. Schorsch, Jonathan 2004. Jews and Blacks in the Early Modern World. Cambridge: Cambridge University Press. Schorsch, Jonathan 2005. Early Modern Sephardim and Blacks. In Zion Zohar (ed.), Sephardic and Mizrahi Jewry, 239-258. New York: New York University Press. Schuchardt, Hugo 1882. Kreolische Studien I. Ueber das Negerportugiesische von S. Thomé (Westafrika). Sitzungsberichte der philosophisch-historischen Classe der Kaiserlichen Akademie der Wissenschaften in Wien 105: 889–917. Schuchardt, Hugo 1883. Kreolische Studien VI, über das Indoportugiesische von Mangalore. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien 105: 881–904.

References

379

Schuchardt, Hugo 1888. Beiträge zur Kenntnis des kreolischen Romanisch III. Zum Negerportugiesischen der Kapverden. Zeitschrift für romanische Philologie 12: 313– 322. Schuchardt, Hugo 1889. Beitrage zur Kenntnis des Kreolischen Romanisch. Zeitschrift fur Romanische Philologie 13: 476–516. Schulte, Kim 2009. Loanwords in Romanian. In Martin Haspelmath and Uri Tadmor. Loanwords in the World’s Languages: A Comparative Handbook, 230–259. Berlin: Mouton De Gruyter. Schwegler, Armin 1990. Analyticity and syntheticity. A diachronic perspective with special reference to Romance languages. Berlin: Mouton De Gruyter. Schwegler, Armin 1993. Rasgos (afro)portugueses en el criollo del Palenque de San Basilio (Colombia). In Díaz Alayón (ed.), Homenaje a José Pérez Vidal, 667–696. Tenerife: Litografía A. Romero S.A. Schwegler, Armin 1996. Chi ma nkongo: lengua y rito ancestrales en El Palenque de San Basilio (Colombia). Frankfurt am Main: Vervuert. Schwegler, Armin 1998. Palenquero. In Matthias Perl and Armin Schwegler (eds.), América Negra: panorámica actual de los estudios lingüísticos sobre variedades hispanas, portuguesas y criollas, 220–291. Frankfurt am Main: Vervuert. Schwegler, Armin 1999a. On the (African) origin of Palenquero subject pronouns. Ms. Schwegler, Armin 1999b. Monogenesis revisited: The Spanish perspective. In John R. Rickford and Suzanne Romaine (eds.), Creole genesis, attitudes and discourse, 235– 262. Amsterdam: Benjamins. Schwegler, Armin 2002. On the (African) origins of Palenquero subject pronouns. Diachronica 19(2): 273–332. Schwegler, Armin and Kate Green 2007. Palenquero (Creole Spanish). In John A. Holm and Peter L. Patrick (eds.), Comparative Creole Syntax, 273–306. London: Battlebridge. Schwegler, Armin 2007. Bare nouns in Palenquero: A fresh consensus in the making. In Marlyse Baptista and Jacqueline Guéron (eds.), Noun phrases in Creole languages: a multi-faceted approach, 205–222. Amsterdam: Benjamins. Sebba, Mark 1997. Contact languages: pidgins and creoles. New York: Palgrave. Seibert, Gerhard To appear. Creolization and Creole communities in the Portuguese Atlantic. São Tomé, Cape Verde and the Rivers of Guinea in comparison. In Tobias Green and Jose Lingna Nafafe (eds.): Proceedings of the Brokers of Change Conference: Atlantic Commerce and Cultures in Pre-Colonial “Guinea of Cape Verde”. Cambridge: Cambridge University Press. Semedo, José Maria 1993. Un archipel volcanique dispersé. In Littérature du Cap-Vert, de Guinée-Bissau, de São Tomé et de Principe. Special edition of Notre Librairie, 6–11. Siegel, Jeff 2004. Morphological simplicity in pidgins and creoles. Journal of Pidgin and Creole Languages 19(1): 139–162. Silla, Ousmane 1969. Essai Historique sur Portudal. Notes Africaines 123: 77–89.

380

References

Silva, Izione Santos 1985. Variation and Change in the Verbal System of Capeverdean Crioulo. Ann Arbor: University Microfilms. Silva, Izione Santos 1990. Tense and Aspect in Capeverdean Crioulo. In John Victor Singler (ed.), Pidgin and Creole Tense /Aspect / Mood Systems, 143–169.Amsterdam: Benjamins. Silva, Luiz Antônio da 2005. A língua que falamos: Português, história, variação e discurso. São Paulo: Globo. Silva, Tomé Varela da (ed.) 1987. Na bóka noti. Un libru di stórias tradisional organizadu y prizentadu pa T.V. da Silva, vol. 1. Praia: Instituto Kauberdianu di Libru. Silva, Tomé Varela 2005. (Kon)tributu: pa libertason y dizanvolviméntu. Praia: author’s edition. Silva, Daniel Barros Domingues da and David Eltis 2008. The slave trade to Pernambuco, 1561–1851. In David Eltis and David Richardson (eds.), Extending the Frontiers: Essays on the New Transatlantic Slave Trade Database, 95–129. Yale:Yale University Press. Silva Rêgo, Antonio da 1998. Dialecto Português de Malaca. Lisbon: CNCDP. Singler, John Victor (ed.) 1990. Pidgin and Creole Tense / Aspect / Mood Systems. Amsterdam: Benjamins. Singler, John Victor 1990a. Introduction: Pidgins and Creoles and Tense-Mood-Aspect. In John Victor Singler (ed.), Pidgin and Creole Tense /Aspect / Mood Systems, vii–xiv. Amsterdam: Benjamins. Singler, John Victor 1990b. The impact of decreolization upon T-M-A: Tenselessness, Mood and Aspect in Kru Pidgin English. In John Victor Singler (ed.), Pidgin and Creole Tense / Aspect / Mood Systems, 203–230. Amsterdam: Benjamins. Sintiago,Alfredo 1898. Gramatica Corticoe di Idioma Papiamentoe. Curaçao: University of the Netherlands Antilles. Smith, Norval 1987.The Genesis of the Creole Languages of Surinam. Ph.D. dissertation, Universiteit van Amsterdam. Smith, Norval 1999. Pernambuco to Surinam 1654–1665? The Jewish slave controversy. In Magnus Huber and Mikael Parkvall (eds.), Spreading the word. The issue of diffusion among the Atlantic Creoles, 251–298. London: University of Westminster Press. Smith, Norval 2009. English-speaking in early Surinam? In Rachel Selbach, Hugo Cardoso and Margot van den Berg (eds.), Gradual Creolization. Studies celebrating Jacques Arends, 305–326. Amsterdam: Benjamins. Soares, Maria João 2006. Igreja, crioulo e línguas em Cabo Verde. In John A. Holm, Jürgen Lang, Jean-Louis Rougé and Maria João Soares (eds.), Cabo Verde: origens da sua sociedade e do seu crioulo, 179–198. Tübingen: Narr. Solovova, Olga 2004. Progressive Markers in the Creoles of Cape Verde and GuinéBissau: The Outcome of Different Sociolinguistic Histories. In Mauro Fernández, Manuel Fernández-Ferreiro, Manuel Veiga and Nancy Vázquez (eds.), Los criollos de base ibérica, 159–169 Madrid/Frankfurt: Iberoamericana/Vervuert

References

381

Stassen, Leon 2008. Noun Phrase Conjunction. In Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie (eds.), The World Atlas of Language Structures Online. Munich: Max Planck Digital Library, chapter 63. http://wals.info/feature/63. Accessed on 18-08-2010. Stewart, William A. 1962. Creole languages in the Caribbean. In Frank Rice (ed.), Study of the role of second languages in Asia, Africa and Latin America, 34–53. Washington: Center for Applied Linguistics. Stolz, Thomas 2003. Not quite the right mixture: Chamorro and Malti as candidates for the status of mixed language. In Peter Bakker and Yaron Matras (eds.), The Mixed Language Debate. Theoretical and Empirical Advances, 271–315. Berlin: Mouton De Gruyter. Suzuki, Miki 1994. The markers in Cape Verdean CP. Manuscrito inédito, CUNY. Swolkien, Dominika 2004. Factores sociolinguísticos no povoamento da ilha de São Vicente de Cabo Verde. In Mauro Fernández, Manuel Fernández-Ferreiro, ManuelVeiga and Nancy Vázquez (eds.), Los criollos de base ibérica, 171–184. Madrid/Frankfurt: Iberoamericana/Vervuert Swolkien, Dominika 2009. Cape Verdean Creole of São Vicente: Verb Phrase – problems and findings. Paper presented at the II International Creole Seminar for Postgraduates, Ludwig-Maximilians-Universität München. Tadmor, Uri 2007. Is borrowability borrowable? Paper presented at the symposium Language Contact and the Dynamics of Language, Max Planck Institute for Evolutionary Anthropology, Leipzig, May 2007. Tadmor, Uri 2009. Loanwords in the world’s languages: Findings and results. In Martin Haspelmath and Uri Tadmor. Loanwords in the World’s Languages: A Comparative Handbook. Berlin: Mouton de Gruyter, 55–76. Taylor, Douglas 1971. Grammatical and lexical affinities of Creoles. In Dell Hymes (ed.), Pidginization and Creolization of Languages, 293–296. Cambridge: Cambridge University Press. Taylor, Douglas 1977. Languages of the West Indies. London: Johns Hopkins. Teenstra, Marten Douwes 1836. De Nederlandsche West-Indische eilanden, in derzelver tegenwoordigen toestand. Amsterdam: S. Emmering. Teyssier, Paul 1959. La langue de Gil Vicente. Paris: Klincksieck. Teza, Emilio 1863. Il dialetto curassesse. Il Politecnico 21: 342–352. Thiele, Petra 1991a. Kabuverdianu: Elementaria seiner TMA-Morphosyntax im lusokreolischen Vergleich. Bochum: Brockmeyer. Thiele, Petra 1991b. El caboverdiano. ¿Una lengua criolla en controversia con el sistema TMA clásico? Papia 1(2): 49–55. Thilmans, Guy 1968. Sur l’existence, fin XVIe siècle, de comptoirs néerlandais à Joal et Portudal (Sénégal). Notes Africaines 117: 17–18. Thilmans, Guy 2006. Des Négro-Africains devant l’Inquisition au XVIe siècle. In Senegalia: Études sur le patrimoine ouest-africain. Hommage à Guy Thilmans, 52–70. Paris: Éditions Sépia.

382

References

Thomason, Sarah G. (ed.) 1997. Contact Languages. A Wider Perspective. Amsterdam: Benjamins. Thomason, Sarah G. 1999. Speakers’choices in language change. Studies in the Linguistic Sciences 29(2): 19–43. Thomason, Sarah G. 2001. Language Contact: An Introduction. Washington D.C.: Georgetown University Press. Thomason, Sarah G. 2003. Social factors and linguistic processes in the emergence of stable mixed languages. In Peter Bakker andYaron Matras (eds.),The Mixed Language Debate. Theoretical and Empirical Advances, 21–40. Berlin: Mouton De Gruyter. Thomason, Sarah G. 2007. Language Contact and Deliberate Change. Journal of Language Contact 1: 41–62. Thomason, Sarah G. 2008. Pidgins/Creoles and Historical Linguistics. In Silvia Kouwenberg and John V. Singler (eds.), The Handbook of Pidgin and Creole Studies, 243–262. Oxford: Wiley-Blackwell. Thomason, Sarah G. and Terrence Kaufman 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Thompson, Robert W. 1961. A note on some possible affinities between the creole dialects of the Old World and those of the New. Creole Language Studies 2: 107–113. Todd, Loreto 2004. Pidgins and Creoles. New York: Routledge. Todd Dandaré, Ramón 1979. Analyzing variations in creole languages: the Papiamento case. Paper presented at the Conference on Theoretical Orientations in Creole Studies, St. Thomas. Valerianus-Fermina, Imelda 1992. Bolo pretu. Curaçao: SEAK. Valkhoff, Marius 1960. Contributions to the study of creole.African Studies 19(2): 77–87. Valkhoff, Marius 1966. Studies in Portuguese and Creoles. Johannesburg: Witwatersrand University Press. Van Name, Addison 1869. Contributions to Creole Grammar. Transactions of the American Philological Association 1: 123–167. Veen Zeppenfeldt, Antonio van de 1928. Praktische Handleiding der Papiamentsche Spraakkunst. Curaçao: author’s edition. Veenstra, Tonjes 1996. The pronominal system of Saramaccan. In Daniel Véronique (ed.), Matériaux pour l’étude des classes grammaticales dans les langues créoles, 259–272. Aix-en-Provence: University Press. Veiga, Manuel 1982. Diskrison strutural di lingua kabuverdianu. Praia: Instituto Caboverdiano do Livro e do Disco. Veiga, Manuel 1995. Introdução à Gramática do Crioulo. Praia: Instituto Caboverdiano do Livro e do Disco. Veiga, Manuel 2000. Le créole du Cap-Vert. Étude grammaticale descriptive et contrastive. Paris: Karthala. Velleman, Barry L. 2008. La imagen y los ecos del lingüista profesional: la correspondencia de Rodolfo Lenz. Revista de Lingüística Teórica y Aplicada 46(1): 11–28. Viaro, Mário Eduardo 2005. Algumas considerações acerca do português falado quatrocentista e quinhentista. Papia 15: 80–101.

References

383

Voorhoeve, Jan 1973. Historical and linguistic evidence in favour of the relexification theory in the Formation of Creoles. Language in Society 1: 133–145. Wagner, Max Leopold 1949. Lingua e Dialetti dell’America Spagnola. Florence: Le Lingue Estere. Walker, Timothy 2005. Lisbon as a Strategic Haven in the Atlantic World. In Willem Klooster and Alfred Padula (eds.), The Atlantic World, 60–75. Upper Saddle River: Pearson. Weinreich, Uriel 1953. Languages in contact: Findings and problems. The Hague: Mouton. Welie, Rik van 2008. Slave Trading and Slavery in the Dutch Colonial Empire: A Global Comparison. New West Indian Guide 82: 47–96. Whinnom, Keith 1965. The origin of the European-based creoles and pidgins. Orbis 14: 509–527. Wijk, Henri L.A. van 1958. Orígenes y evolución del Papiamentu. Neophilologus 42: 169–182. Wijk, Henri L.A. van 1968. Orígenes y evolución del Papiamentu. Watapana 1(2): 13–16. Wijk, Henri L.A. van 1969. Orígenes y evolución del Papiamentu. Watapana 1(3): 13–18. Williams, James Homer 2001. An Atlantic Perspective on the Jewish Struggle for Rights and Opportunities in Brazil, New Netherland, and New York. In Paolo Bernardini and Norman Fiering (eds.), The Jews and the Expansion of Europe to the West, 1450–1800, 369–393. New York/Oxford: Berghahn Books. Wilson, William A.A. 1962. The Crioulo of Guiné. Johannesburg: Witwatersrand University Press. Winford, Donald 2003. An introduction to contact linguistics. Oxford: Blackwell. Winford, Donald 2008. Atlantic Creole Syntax. In Silvia Kouwenberg and John Victor Singler (eds.), The Handbook of Pidgin and Creole Studies, 19–47. Oxford: WileyBlackwell. Wit, Bernadette de 1997. Archeoloog van de taal. De Groene Amsterdammer (8)1. Wood, Richard E. 1970. Papiamentu: Dutch Contributions. Ph.D. dissertation, Indiana University. Wood, Richard E. 1971. The English loanwords in Papiamentu. Nieuwe West-Indische Gids 48: 173–189. Wood, Richard E. 1972a. New light on the origins of Papiamentu: an eighteenth-century letter. Neophilologus 56: 18–30. Wood, Richard E. 1972b. Dutch Syntactic Loans in Papiamentu. Revue des langues vivantes 38(6): 635–647. Wood, Richard E. 1972c. The Hispanization of a Creole Language: Papiamentu. Hispania, (55)4: 857–863.

384

References

Newspapers The Associated Press 02-04-1985 Boletin di Notisia, 29-1-2010 Civilisadó 7-10-1871 Èxtra 25-10-2003 Èxtra 31-03-2006 Èxtra 02-09-2008 Èxtra 13-02-2008 Èxtra 21-03-2009 Èxtra 10-03-2011 La Union 19-03-1889 La Union 19-04-1889 Solo 15-07-2009

Internet sources Ethnologue: www.ethnologue.com http://www.curacao-encyclopedia.com, accessed on 05-08-2010 http://en.wikipedia.org/wiki/Sephardi_Jews, accessed on 05-08-2010 http://portugesnacrioulo.blogspot.com/2007/11/daki-nada.html, accessed on 19-08-2010 http://pensaodbertabissau.wordpress.com/amigos-da-avo/, accessed on 19-08-2010 http://articles.orlandosentinel.com/1999-08-28/lifestyle/9908270501_1_wide-webweb-today-world-tour, accessed on 24-03-2011 http://www.bendishon.com/bendishon/manageDocument.do?dispatch=view&id=1110, accessed on 01-04-2011 Diccionario de la Real Academia Española (DRAE): http://buscon.rae.es/draeI/SrvltConsulta?TIPO_BUS=3&LEMA=cultura Dicionário Priberam da Língua Portuguesa: http://www.priberam.pt/DLPO/ Authorized Version of the Dutch Bible: www.statenvertaling.net http://www.mass.gov/Elwd/docs/dia/publications/ls_cape_verdean.pdf Proclamasjon: http://www.extrabon.com/edishon/edishon2006/extra2006-10-28.pdf The Transatlantic slave trade database (Voyages): www.slavevoyages.org

Index

a (see under ‘perfective past’) Amsterdam (Chamber of) 34, 35, 270, 273, 275, 278, 279, 287, 290, 292, 294–297, 299 anterior tense / imperfective past (tabata) 193–196, 208–217 Aruban Papiamentu, 6, 6 (fn 6), 53 (fn 95), 54, 63, 84 (fn 137), 220 (fn 341) Asian Portuguese Creoles 86, 139, 148, 198, 202, 203, 207 (fn 320), 224, 234, 235, 338 auxiliary verbs 155–159, 161, 163, 164, 166, 167, 168, 175, 177, 178, 180, 181, 182, 185, 206, 207, 221, 226–255 – modals 164, 165, 227–229, 255 – copulas 230–239 Bantu (see under ‘substrate’) Barlavento varieties of CV 6, 7, 11, 76, 81, 126, 128 (fn 202), 149, 174, 180, 208, 209–214 (fn 336), 225, 230, 231, 234, 240, 252 Bioprogram 23, 27, 171, 180, 183, 188, 189, 193–197, 199, 200, 207, 210, 211, 214, 216, 217, 226, 256 Bonairean Papiamentu 6 (fn 6), 54, 135 (fn 216), 144 (fn 230), 220 (fn 341) Cacheu 213, 274, 276, 278, 285–289, 291–293, 295–298, 300, 302, 304 Cartagena 252, 253, 291–293 Casamance variety of GBC 8, 9, 11, 53 (fn 97), 80, 115 (fn 183), 127, 138, 145 (fn 233), 182, 185, 228, 230 (fn 348), 240, 251

Chabacano 55, 64, 85 (fn 138), 86, 104, 107, 117, 119, 122, 147, 148, 155, 208 (fn 320), 247, 317 conjunctions – subordinate 99, 111, 112, 120– 129 – coordinate 118, 119, 120 – comparative 115, 126 creole typology 94 (fn 154), 103, 119, 134 (fn 214), 136 (fn 220), 156, 183 (fn 285), 195, 219, 247, (see also under ‘Bioprogram’) definite article 90 (fn), 91 (fn), 134– 138 demonstrative (deictic/definite marker) 91, 134–138, 263, 264 Dutch (the) 4, 5, 37, 270–294, 296– 305, 309–312, 320, 321, 333, 335, 336, 341, 342 Dutch (language) 37, 37 (fn 76), 70, 74, 79, 82, 89, 144, 145, 148 (fn 234), 152 (fn 240), 157, 166 (fn 266), 167, 234 Early Papiamentu 11, 54, 56, 60, 61, 66, 79, 98 (fn 160), 110, 115 (fn 184), 116, 117, 124, 126–129, 134–139, 149–153, 157, 159–163, 166, 167, 195, 216, 228, 237, 265, 266, 326, 327 future tense 183, 197–208 Gorée 14, 271, 273–286, 289–291, 293, 297, 298, 302 grammaticalization 88, 96, 99, 105, 117, 119 (fn 191), 121, 122 (fn 194),

386

Index

132, 133, 138, 196, 200–202, 206 (fn 317), 208, 237 – equivalence 93, 167 – gap filling 92, 167 (fn 268), 196, 197, 200, 216 – reanalysis 88, 104, 105, 232 Gulf of Guinea PC (ST, ANG, PRI, Annobonese), 12, 40, 41, 45, 47, 52– 55, 61, 67–69, 71, 72, 76–78, 85, 88, 90, 91, 94, 96, 97, 99–101, 104, 106, 108–110, 112, 117, 120, 122, 123, 125, 126, 128, 130, 133, 137, 140, 141, 147, 148, 155, 172, 182 (fn 282), 192, 207, 211, 223, 229, 234, 235, 237, 241, 252, 254, 260– 264, 338 imperfective aspect 167, 172–193, 195– 198, 203–207, 209, 211, 217–226, 231, 233 – progressive 173–175, 178–185, 192, 211 – habitual 172, 173, 186, 192 imperfective past (see under ‘anterior tense’) interrogatives 114–117, 126 intertwiner (see under ‘mixed language’)

– past participles 79, 151–157, 159, 160, 163–168, 246 – plural (see under ‘nan’) nan (3pl = PL) 86, 90–95 negation 139–141, 260 Old Portuguese 53, 62, 64, 68, 69, 107, 148 (fn 235), 164, 229, 264, 265 Old Spanish 26, 29, 63, 64, 147, 148 (fn 235) Palenquero (PLQ) 13, 22–24, 27, 41, 42, 54, 55, 58, 61, 67, 72, 73, 77, 80, 85, 86, 88, 92, 94, 96, 101, 104, 107, 112, 116, 117, 119, 122, 125, 126, 133, 137, 140, 141, 147, 148, 155, 173, 192, 223, 224, 228, 234, 252, 253, 259–261, 305 perfective past 161, 221 personal pronouns 83–95, 263 Petite Côte 270, 271, 273–275, 277, 279–287, 289, 295, 296, 298–301, 303, 304, 341 pidgin 39–43, 200 (fn 308), 312–316 prepositions 95–111, 113, 114, 118, 122, 123

Kwa (see under ‘substrate’) lingua de pretu (see under ‘reconnaissaince language’) lo (see under ‘future tense’) Mande (see under ‘substrate’) mixed language 327–334, 336 monogenesis 20–24, 39, 41, 46–48, 98, 192 (fn 295), 320–322, 327 morphology 143–161, 163–170 – derivational 143–150 – inflectional 150–155 – passive 155–161, 163–169

reciprocity 129, 130 reconnaissaince language 22, 23, 42, 43, 87 (fn 141), 182 (fn 282), 202 (fn 313) reflexivity 129–134 relative vs. absolute tense 214–217 relexification 9, 20, 47, 48, 83. 319– 333, 335 Santiago 6, 7, 277, 279, 281, 285, 287, 288, 300–304 Senegambia 269–279, 281, 283–286, 288, 289–291, 294, 295, 296

Index Sephardim 30–38, 44, 67, 89, 279, 287, 293–299, 303–305, 311, 332, 333, 335, 336 Sephardic Papiamentu 89, 108, 210, 212, 238 slaves, slavery, slave trade 4, 5, 7, 37, 44, 161 (fn 260), 210, 253, 269–277, 282, 285–295, 297–300, 303, 304, 308–312, 316–320, 325, 332, 333, 335, 336, 341 Sotavento varieties of CV (other than SCV) – Fogo CV 107, 109, 180, 178, 222, 228, 237, 246 – Brava CV 81 (fn 135), 138 (fn 222), 175, 176, 179, 240

387

stative verbs 188, 191, 193–196, 216– 226 substrate – West-Atlantic (Wolof), Mande (Mandinka) 7, 8, 60, 64, 71, 72, 94 (fn 154), 106, 146, 226, 246, 250–253, 266, 272 – Kwa/Bantu 12, 13, 45, 71, 72, 94 (fn 154), 130, 133, 261 ta (see under ‘imperfective aspect’) West-Atlantic (see under ‘substrate’) Wolof (see under ‘substrate’) word order 94, 99, 106, 123, 133, 138, 263, 264