The Oxford Guide to the Languages of the Central Andes (Oxford Guides to the World's Languages) 0198849923, 9780198849926

Citation preview

the oxford guide to the

LANGUAGES OF THE CENTRAL ANDES

ox f or d g u ide s to t he w or l d ’s l a ng ua g e s

general editors Adam Ledgeway, University of Cambridge, and Martin Maiden, University of Oxford

advisory editors

Alexandra Y. Aikhenvald, Central Queensland University Edith Aldridge, University of Washington Stephen R. Anderson, Yale University Bernard Comrie, University of California, Santa Barbara Jan Terje Faarlund, University of Oslo Alice Harris, University of Massachusetts, Amherst Bernd Heine, University of Cologne Paul Hopper, Carnegie-Mellon University

Geoffrey Khan, University of Cambridge Lutz Marten, SOAS, London Marianne Mithun, University of California, Santa Barbara Irina Nikolaeva, SOAS, London Chris Reintges, CNRS, Paris Masayoshi Shibatani, Rice University David Willis, University of Oxford

published The Oxford Guide to the Atlantic Languages of West Africa Edited by Friederike Lüpke The Oxford Guide to Australian Languages Edited by Claire Bowern The Oxford Guide to the Languages of the Central Andes Edited by Matthias Urban The Oxford Guide to the Malayo-Polynesian Languages of Southeast Asia Edited by Alexander Adelaar and Antoinette Schapper The Oxford Guide to the Romance Languages Edited by Adam Ledgeway and Martin Maiden The Oxford Guide to the Transeurasian Languages Edited by Martine Robbeets and Alexander Savelyev The Oxford Guide to the Uralic Languages Edited by Marianne Bakró-Nagy, Johanna Laakso, and Elena Skribnik

in preparation The Oxford Guide to the Bantu Languages Edited by Ellen Hurst, Nancy Kula, Lutz Marten, and Jochen Zeller The Oxford Guide to the Papuan Languages Edited by Nicholas Evans and Sebastian Fedden The Oxford Guide to the Slavonic Languages Edited by Jan Fellerer and Neil Bermel The Oxford Guide to the Tibeto-Burman Languages Edited by Kristine Hildebrandt, Yankee Modi, David Peterson, and Hiroyuki Suzuki

the oxford guide to the

Languages of the Central Andes edited by

Matthias Urban

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © editorial matter and organization Matthias Urban 2024 © the chapters their several authors 2024 The moral rights of the authors have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2024943239 ISBN 9780198849926 DOI: 10.1093/oso/9780198849926.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

This volume is dedicated to the memory of three pioneers of Central Andean linguistics: Peter Landerman (1940–2023), Pieter C. Muysken (1950–2021), and Gérald Taylor (1933–2020).

Short contents Detailed contents Series preface List of abbreviations The contributors

PART I: Background and context 1. Introduction: Central Andean linguistic diversity and the diversity of Central Andean linguistics Matthias Urban

xi xlii xliii xlvi

1 3

2. Physical geography and cultural trajectory of the Central Andes Peter Kaulicke

38

3. Historical linguistics, philology, and the development of Andean linguistics Rodolfo Cerrón-Palomino

63

PART II: Language profiles

83

4. Huaylas (Ancash) Quechua Carlos Molina-Vital

85

5. Southern Yauyos Quechua Aviva Shimelman

139

6. Chachapoyas Quechua Aviva Shimelman and Jairo Valqui

167

7. Cuzco Quechua Raúl Bendezú-Araujo and Jorge Acurio-Palma

196

8. Aymara Matt Coler

219

9. Jaqaru Matt Coler

245

10. Uru and Chipaya Katja Hannß

265

11. Mochica Matthias Urban

311

12. Puquina Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar

335

vii

short contents 13. Híbito and Cholón Astrid Alexander-Bakkerus

392

14. Small and extinct languages of Northern Peru Matthias Urban

419

15. Kallawaya Pieter Muysken

438

16. The Andean Spanish of Southern Peru and Bolivia Luis Andrade Ciudad

469

PART III: Comparative studies

483

17. Central Andean Segmental Phonologies in Continental Perspective Lev Michael and Allegra Robertson

485

18. The morphology of the nominal domain in the languages of the Central Andes Olga Krasnoukhova

532

19. The grammar of the verb in the languages of the Central Andes Matthias Pache

551

20. Syntactic structures in the languages of the Central Andes Rik van Gijn

597

21. Discourse, information structure, and evidentiality in the Central Andes Karolina Grzech

635

22. Linguistic complexity in the Central Andes Johanna Nichols

667

PART IV: Language history

681

23. Expansions and language shift in prehistory Paul Heggarty

683

24. Language ecologies and dynamics in the ancient Central Andes Matthias Urban

725

25. Language diffusion and state agency: Quechuan in Inca and colonial times César Itier

739

PART V:

755

Language contact, sociolinguistics, and linguistic anthropology

26. The Quechuan–Aymaran relationship Nicholas Q. Emlen 27. Contact between Indigenous languages of the Central Andes and Spanish: Linguistic outcomes as cases of contra-hierarchical diffusion Anna María Escobar 28. Language ideologies and the Quechuan family Rosaleen Howard

viii

757

783 798

short contents 29. The Andean–Amazonian interface: Sociolinguistic relations and areal–typological patterns Nicholas Q. Emlen, Rik van Gijn, and Sietze Norder

818

30. Language and the Andean environment Joshua Shapero

839

References Index

860 923

ix

Detailed contents Series preface List of abbreviations The contributors

PART I: Background and context 1. Introduction: Central Andean linguistic diversity and the diversity of Central Andean linguistics Matthias Urban 1.1 Extent and characteristics of the Central Andes 1.2 Past and present linguistic landscapes of the Central Andes 1.3 Language contact, areal typology, and the fabric of Central Andean languages 1.4 The emergence and development of descriptive traditions in the languages of the Central Andes 1.5 The emergence and development of the historical linguistics of the Central Andes 1.6 Language, culture, and society in the past 1.7 Language, culture, and society in the present 1.8 Language shift and language endangerment 1.9 The broader relevance of Central Andean languages and linguistics 1.10 The present volume 1.11 Conventions 1.11.1 Glossonyms and spelling conventions for other proper names 1.11.2 Orthographies and other conventions for the representation of linguistic data

1.11.3 Terminology and morpheme glosses 2. Physical geography and cultural trajectory of the Central Andes Peter Kaulicke 2.1 Introduction 2.2 Northern perspectives 2.2.1 The Northern environment 2.2.2 Population and cultural trajectories 2.3 Central perspectives 2.3.1 Environment 2.3.2 Population and cultural trajectories 2.4 Southern perspectives 2.4.1 Introduction 2.4.2 Environment 2.4.3 Population and cultural trajectories 2.5 Summary and conclusions

xlii xliii xlvi

1 3 3 4 7 10 15 18 21 23 25 27 28 28 29 36 38 38 39 39 42 46 46 48 52 52 52 54 60

xi

detailed contents

xii

3. Historical linguistics, philology, and the development of Andean linguistics Rodolfo Cerrón-Palomino 3.1 Preliminaries 3.2 Philology and the Andean world 3.3 The written documentary record 3.4 Agreements 3.4.1 Diachronic questions 3.4.1.1 The Quechuan sibilants 3.4.1.2 Weakening of consonantal codas in Cuzco Quechua 3.4.1.3 The possessive phrase in Coastal Quechua 3.4.1.4 Stress and vowel elision 3.4.1.5 Evolution of accusative marking in Southern Aymara 3.4.1.6 Evolution of derivative suffixes 3.4.2 Questions of external relations 3.4.3 Lexico-semantic questions 3.5 Disagreements 3.5.1 Reconstructions of protolanguages 3.5.2 Philological and ecdotic problems 3.5.2.1 Recent relapses in the editions of chronicles 3.5.2.2 Misguided editions 3.6 Epilogue

63

PART II: Language profiles

83

4. Huaylas (Ancash) Quechua Carlos Molina-Vital 4.1 General information 4.1.1 Geography, environment, and economics 4.1.2 Speaker population and endangerment 4.2 Internal variation and classification 4.3 Phonology and phonotactics 4.3.1 Vowels 4.3.2 Consonants 4.3.3 Syllable structure, root structure, and phonotactic constraints 4.3.4 Important morphophonological processes 4.3.4.1 Syllable structure and the suffix -ni ‘epenthetic (ep)’ 4.3.4.2 Monophthongization 4.3.4.3 Vowel lowering in derivational morphology 4.3.5 Stress assignment 4.4 Parts of speech and transcategorical operations 4.4.1 Parts of speech 4.4.2 Transcategorical operations 4.4.2.1 Nominal derivation (nominalization) 4.4.2.1.1 -na ‘potential nominalizer (pot.nmlz)’ 4.4.2.1.2 -q ‘agentive nominalizer (ag)’ 4.4.2.1.3 -shqa ‘resultative nominalizer (res)’ 4.4.2.1.4 -nqa ‘actual nominalizer (act.nmlz)’ 4.4.2.1.5 -y ‘infinitive (inf)’ 4.4.2.2 Verbal derivation 4.4.2.2.1 -ya: ‘transformative (tf)’

85

63 65 66 67 68 68 68 69 69 70 71 72 74 76 76 78 78 80 81

85 85 85 86 87 87 88 89 89 89 90 90 91 92 92 92 92 92 92 93 93 93 93 93

detailed contents 4.4.2.2.2 -na: ‘desiderative (desr)’ 4.4.2.2.3 -tsa: ‘factive (fact)’ 4.4.2.2.4 -ta: ‘substance-covering (cov)’ 4.4.2.2.5 Voice as verbalizer: -tsi ‘causative (caus),’ -ku

‘middle voice, reflexive, medio passive (mid)’ and -ka: ‘spontaneous voice (spon)’

4.5 Nominals 4.5.1 Subclasses of nominals 4.5.1.1 Pronouns 4.5.1.1.1 Personal pronouns 4.5.1.1.2 Demonstrative pronouns 4.5.1.1.3 Nominally inflected pronouns (emphatic and quantifiers)

4.5.1.1.4 Interrogative and indefinite pronouns

4.5.1.2 Numerals 4.5.1.3 Gender 4.5.1.4 Order of constituents in the nominal phrase

4.5.2 Nominal derivation 4.5.2.1 -yuq ‘inalienable possession (inal.poss)’ 4.5.2.2 -sapa ‘multiple possession (mult.all)’ 4.5.2.3 -ntin ‘accompaniment, adjacency (incl)’ 4.5.2.4 -mayi ‘partnership (part)’ 4.5.2.5 -lla ‘restrictive, limitative (lim)’ 4.5.3 Nominal inflection 4.5.3.1 Person 4.5.3.2 Number 4.5.3.3 Case 4.5.3.3.1 -ta ‘accusative (acc)’ 4.5.3.3.2 -pa ‘genitive (gen)’ 4.5.3.3.3 -man ‘allative (all)’ 4.5.3.3.4 -chaw ‘locative (loc)’ 4.5.3.3.5 -paq ‘dative, allative, purposive (dat)’ 4.5.3.3.6 -piq ~ -pita ‘ablative (abl)’ 4.5.3.3.7 -naw ‘simulative (simul)’ 4.5.3.3.8 -yaq, -kama ‘limitative (lim)’ 4.5.3.3.9 -wan ‘comitative-instrumental (com)’ 4.5.3.3.10 -raykur ‘causal (csl)’ 4.5.3.3.11 -pura ‘exclusive (excl)’ 4.6 Verbs 4.6.1 Verb subclasses 4.6.1.1 Copulative verbs 4.6.1.2 Transitive verbs 4.6.1.3 Intransitive verbs 4.6.2 Verbal derivation 4.6.2.1 Aspect 4.6.2.1.1 Imperfectives 4.6.2.1.2 Perfectives 4.6.2.2 Voice 4.6.2.2.1 -tsi ‘causative (caus)’ 4.6.2.2.2 -ku ‘middle (mid)’ 4.6.2.2.3 -na and -ku ‘reciprocal (recp)’ 4.6.2.2.4 -ka: ‘spontaneous (spon)’

93 94 94 94 94 94 94 94 95 95 95 96 97 97 97 97 97 97 98 98 98 98 98 98 98 100 101 101 101 101 101 101 102 102 102 103 103 103 104 105 105 105 105 106 109 109 110 111 112

xiii

detailed contents 4.6.2.2.5 -pu ‘benefactive (ben)’ 4.6.2.2.6 -pa: ‘applicative (appl)’

4.6.2.3 Other derivational suffixes 4.6.2.3.1 -mu ‘cislocative (cisl)’ 4.6.2.3.2 -lla: ‘limitative (lim)’ 4.6.2.3.3 -rpu ‘downward (dw)’ 4.6.2.3.4 -rpa: ‘stationary (stn)’ 4.6.2.3.5 -paku ‘attentive action (atten)’ 4.6.2.3.6 -chaku ‘(care)’ 4.6.2.4 Order of derivational suffixes

4.6.3 Inflectional categories of the verb 4.6.3.1 Person and number 4.6.3.1.1 Subject marking 4.6.3.1.2 Object marking 4.6.3.2 Tense 4.6.3.2.1 Present 4.6.3.2.2 -rqa ‘experienced past (pst)’ 4.6.3.2.3 -rqu ‘recent past/perfective of immediateness (‘rec.pst)’ 4.6.3.2.4 -naq ‘non-experienced past or narrative past (narr.pst)’ 4.6.3.2.5 -shqa ‘past perfect (prf)’ 4.6.3.2.6 -q + kay ‘habitual past construction (hab)’ 4.6.3.2.7 ‘future (fut)’ 4.6.3.3 ‘conditionals (cond)’ 4.6.3.4 ‘imperatives (imp)’ 4.7 Independent suffixes (enclitics) 4.7.1 Connectors of sentence information 4.7.1.1 -raq ‘continuative (cont)’ 4.7.1.2 -na ‘discontinuative (disc)’ 4.7.1.3 -pis ‘additive (add)’ 4.7.1.4 -taq ‘contrastive (contr)’ 4.7.2 Negative and interrogative 4.7.2.1 -tsu ‘negative (neg)’ 4.7.2.2 -ku ‘interrogative (q)’ 4.7.3 Modality: topic and evidentials 4.7.3.1 -qa ‘topic (top)’ 4.7.3.2 -mi ‘evidential–direct experience (evd)’ 4.7.3.3 -shi ‘evidential–reportative/quotative (evr)’ 4.7.3.4 -chi ‘evidential–conjectural (evc)’ 4.7.3.5 -ri ‘corroborative (corr)’ 4.7.3.6 -tsuraq ‘dubitative (dub)’ 4.8 Basic syntax 4.8.1 Grammatical relations 4.8.2 Constituent order 4.8.3 Major clause types 4.8.3.1 Assertions 4.8.3.2 Negations and prohibitions 4.8.3.3 Questions 4.8.4 Complex clauses 4.8.4.1 Juxtaposition and coordination 4.8.4.2 Subordination 4.8.4.2.1 Relative clauses 4.8.4.2.2 Nominal clauses (embedding) 4.8.4.2.3 Adverbial clauses

xiv

113 114 116 116 118 118 118 118 119 119 120 121 121 122 123 123 123 124 124 124 124 125 125 125 125 125 125 126 126 126 127 127 127 127 127 128 128 128 129 129 129 129 129 130 130 130 130 131 131 132 132 133 134

detailed contents 4.9 Text sample 4.10 Literature 5. Southern Yauyos Quechua Aviva Shimelman 5.1 General information 5.1.1 Geography and environment 5.1.2 Speaker population and language endangerment 5.2 Internal variation and classification 5.3 Phonology and phonotactics 5.3.1 Vowel phonemes 5.3.2 Consonant phonemes 5.3.3 Syllable structure and root structure: phonotactic constraints 5.3.4 Major morphophonological processes 5.3.5 Stress assignment 5.3.6 Phonological developments from the protolanguage 5.4 Parts of speech and transcategorical operations 5.4.1 Parts of speech 5.4.2 Transcategorical operations 5.4.2.1 Nominal derivation 5.4.2.1.1 -na ‘nominalizer (nmlz)’ 5.4.2.1.2 -q ‘agentive (ag)’ 5.4.2.1.3 -sHa ‘resultative (res)’ 5.4.2.1.4 -y ‘infinitive (inf)’ 5.4.2.2 Verbal derivation 5.4.2.2.1 -cha ‘factive (fact)’ 5.4.2.2.2 -ku ‘reflexive (refl)’ 5.4.2.2.3 -tuku ‘simulative (simul)’ 5.4.2.2.4 -ya ‘inchoative (inch)’ 5.4.2.2.5 -na ‘verbalizer (vblz)’ 5.4.2.2.6 -naya ‘desiderative (desr)’ 5.4.2.3 Other derivational processes 5.5 Nominals 5.5.1 Subclasses of nominals 5.5.1.1 Pronouns and other pro-forms 5.5.1.1.1 Personal pronouns 5.5.1.1.2 Demonstrative pronouns 5.5.1.1.3 Dependent pronouns 5.5.1.1.4 Interrogative-indefinites 5.5.1.2 Numerals 5.5.1.3 Common nouns 5.5.1.4 Locative nouns 5.5.2 Nominal derivation 5.5.2.1 -kuna ‘non-exhaustivity (nonexhst)’ 5.5.2.2 -ntin ‘accompaniment, adjacency (incl)’ 5.5.2.3 -sapa ‘multiple possession (mult.all)’ 5.5.2.4 -yuq ‘possession (poss)’ 5.5.2.5 -masi ‘partnership (part)’ 5.5.2.6 -cha ‘diminutive (dim)’ 5.5.3 Nominal inflection 5.5.3.1 Person 5.5.3.2 Number

137 137 139 139 139 139 140 140 141 141 141 141 141 142 142 142 143 143 143 143 143 143 143 143 143 143 143 143 143 144 144 144 144 144 144 144 145 145 146 146 146 146 146 146 146 147 147 147 147 148

xv

detailed contents 5.5.3.3 Case 5.5.3.3.1 5.5.3.3.2 5.5.3.3.3 5.5.3.3.4 5.5.3.3.5 5.5.3.3.6 5.5.3.3.7 5.5.3.3.8 5.5.3.3.9 5.5.3.3.10

-ta ‘accusative (acc)’ -pa ‘genitive-locative (gen)’ -man ‘allative-dative (all)’ -pi ‘locative (loc)’ -paq ‘ablative, benefactive, purposive (abl, ben, purp)’ -hina ‘simulative (simul)’ -kama ‘limitative (lim)’ -wan ‘comitative-instrumental (com)’ -rayku ‘causal (csl)’ -puRa ‘exclusive (excl)’

5.6 Verbs 5.6.1 Verb subclasses 5.6.2 Verbal derivation 5.6.2.1 Affixes 5.6.2.1.1 -cha ‘diminutive (dim)’ 5.6.2.1.2 -chi ‘causative (caus)’ 5.6.2.1.3 -ka ‘accidental (passacc)’ 5.6.2.1.4 -kaćha ‘frequentive (freq)’ 5.6.2.1.5 -ku ‘reflexive, middle, medio-passive, passive (refl)’ 5.6.2.1.6 -lla ‘restrictive, limitative (lim)’ 5.6.2.1.7 -mu ‘cislocative, translocative (cisl)’ 5.6.2.1.8 -naku ‘reciprocal (recp)’ 5.6.2.1.9 -naya(:) ‘desiderative (desr)’ 5.6.2.1.10 -pa ‘repetitive (repet)’ 5.6.2.1.11 -pa(:)ku ‘joint action (jtact)’ 5.6.2.1.12 -pu ‘benefactive, translocative (ben)’ 5.6.2.1.13 -Ra ‘durative (dur)’ 5.6.2.1.14 -Raya ‘passivizer (pass)’ 5.6.2.1.15 -Ri ‘inceptive (incep)’ 5.6.2.1.16 -Ru ‘various (urgt)’ 5.6.2.1.17 -sHi ‘accompaniment (acmp)’ 5.6.2.1.18 -tamu ‘irreversible (irrev)’ 5.6.2.1.19 -yku ‘exceptional performance (excep)’ 5.6.2.2 Affix order 5.6.3 Inflectional categories of the verb 5.6.3.1 Person and number 5.6.3.1.1 Subject 5.6.3.1.2 Subject and object reference 5.6.3.2 Tense 5.6.3.2.1 Present 5.6.3.2.2 Future 5.6.3.2.3 Simple past 5.6.3.2.4 Narrative past 5.6.3.2.5 Perfect 5.6.3.2.6 Habitual past 5.6.3.3 Conditionality 5.6.3.4 Imperativity 5.6.3.5 Aspect 5.6.3.6 Subordination 5.6.3.7 Order of affixes

xvi

148 148 148 149 149 149 149 150 150 150 150 150 150 151 151 151 151 151 152 152 152 152 152 152 152 152 152 152 152 152 153 153 153 153 153 154 154 154 155 156 156 156 156 156 156 157 157 158 158 158 159

detailed contents

5.7 Clitics with phrasal or clausal scope 5.7.1 Clitics that position the utterance 5.7.1.1 -chu ‘interrogative (q)’ 5.7.1.2 -lla ‘restrictive-limitative (lim)’ 5.7.1.3 -ña ‘discontinuative (disc)’ 5.7.1.4 -pis ‘additive (add)’ 5.7.1.5 -qa ‘topic marker (top)’ 5.7.1.6 -Raq ‘continuative (cont)’ 5.7.1.7 -taq ‘sequential (seq)’ 5.7.2 Clitics that position the speaker 5.7.2.1 -Yá ‘emphatic (emph)’ 5.7.2.2 -puni ‘certainty, precision (cert)’ 5.7.2.3 -mi ‘evidential–direct experience (evd)’ 5.7.2.4 -shi ‘evidential–reportative/quotative (evr)’ 5.7.2.5 -ćhi ‘evidential – conjectural (evc)’ 5.7.2.6 -aRi ‘assertive force (ass)’ 5.7.2.7 -ik and -iki ‘evidential modifiers (ev.mod.2, ev.mod.3)’ 5.7.3 Combinations 5.8 Basic syntax 5.8.1 Grammatical relations 5.8.2 Constituent order 5.8.3 Major clause types 5.8.4 Coordination 5.8.5 Clausal complementation 5.8.5.1 -na ‘nominalizer (nmlz)’ 5.8.5.2 -q ‘agentive (ag)’ 5.8.5.3 -sHa ‘resultative (res)’ 5.8.5.4 -y ‘infinitive (inf)’ 5.8.6 Relativization 5.8.7 Negation 5.9 Lexical and semantic properties 5.9.1 Lexical differences among the SYQ dialects 5.9.2 Evidentials 5.10 Text sample 5.11 Literature 6. Chachapoyas Quechua Aviva Shimelman and Jairo Valqui 6.1 General information 6.2 Internal variation 6.3 Phonology and phonotactics 6.3.1 Vowel phonemes 6.3.2 Consonant phonemes 6.3.2.1 Consonant inventory 6.3.2.2 Controversial segments 6.3.2.2.1 Affricates 6.3.2.2.2 Laterals 6.3.3 Root structure and syllable structure 6.3.4 Major morphophonological processes 6.3.4.1 Occlusives 6.3.4.1.1 Voicing of occlusives

159 159 159 160 160 160 160 160 160 160 160 160 161 161 161 161 161 162 162 162 162 162 162 163 163 163 163 163 164 164 164 164 165 165 166 167 167 167 168 168 168 168 169 169 170 170 173 173 173

xvii

detailed contents 6.3.4.1.2 Fricatization of occlusives 6.3.4.2 Fricatives /s, sh/ 6.3.4.3 Nasals 6.3.4.4 Approximants /r, j, w/ 6.3.4.4.1 /r/ 6.3.4.4.2 /y/ 6.3.4.4.3 /w/

6.3.5 Stress assignment 6.3.6 Phonological developments from the protolanguage 6.3.7 Orthographic conventions 6.4 Parts of speech and transcategorical operations 6.4.1 Parts of speech 6.4.1.1 Subclasses of nominals 6.4.1.2 Subclasses of verbs 6.4.1.3 Subclasses of adjectives 6.4.2 Transcategorical operations 6.4.2.1 Nominal derivation 6.4.2.1.1 -na ‘nominalizer (nmlz)’ 6.4.2.1.2 -k ‘agentive (ag)’ 6.4.2.1.3 -shka ‘resultative (res)’ 6.4.2.1.4 -y ‘infinitive (inf)’ 6.4.2.2 Verbal derivation 6.4.2.2.1 -cha ‘factive (fact)’ 6.4.2.2.2 -ku ‘reflexive (refl)’ 6.4.2.2.3 -ya ‘inchoative (inch)’ 6.4.2.3 Other derivational processes 6.5 Nominals 6.5.1 Subclasses of nominals 6.5.1.1 Pronouns 6.5.1.1.1 Personal pronouns 6.5.1.1.2 Demonstrative pronouns 6.5.1.1.3 Dependent pronouns 6.5.1.1.4 Interrogative indefinites 6.5.1.2 Numerals 6.5.1.3 Common nouns 6.5.2 Nominal derivation 6.5.2.1 -ntin ‘accompaniment, stereotypical association (incl)’ 6.5.2.2 -yuk ‘possession (poss)’ 6.5.3 Nominal inflection 6.5.3.1 Person 6.5.3.2 Number 6.5.3.3 Case 6.5.3.3.1 -ta ‘accusative (acc)’ 6.5.3.3.2 -pa ‘genitive (gen)’ 6.5.3.3.3 -pi ‘locative (loc)’ 6.5.3.3.4 -man ‘allative (all)’ 6.5.3.3.5 -kaman ‘limitative (lim)’ 6.5.3.3.6 -manta ‘ablative (abl)’ 6.5.3.3.7 -pa ‘purposive (purp)’ 6.5.3.3.8 -wan ‘comitative-instrumental (com)’ 6.5.3.3.9 -shina and -yupay ‘simulative (simul)’ 6.5.3.3.10 -rayku ‘causal (csl)’

xviii

174 174 174 175 175 175 176 176 177 177 180 180 180 180 180 180 180 180 181 181 181 181 181 181 181 182 182 182 182 182 182 182 182 183 183 183 183 184 184 184 184 184 185 185 185 185 186 186 186 186 186 186

detailed contents

6.6 Verbs 6.6.1 Verb subclasses 6.6.2 Verbal derivation 6.6.2.1 -chi ‘causative/(di-)transitivizer (caus)’ 6.6.2.2 -ku ‘reflexive, middle, medio-passive (refl)’ 6.6.2.3 -mu ‘cislocative, translocative (cisl)’ 6.6.2.4 -naku ‘reciprocal (recp)’ 6.6.2.5 -naya ‘desiderative (desr)’ 6.6.2.6 -pu ‘benefactive, translocative (ben)’ 6.6.2.7 -ri ‘prolongation/persistence/continuation (cont)’ 6.6.3 Inflectional categories of the verb 6.6.3.1 Person and number 6.6.3.2 Tense 6.6.3.2.1 Present 6.6.3.2.2 Future 6.6.3.2.3 Past 6.6.3.3 Conditionality 6.6.3.4 Imperativity 6.6.3.5 Aspect 6.6.3.6 Subordination 6.6.3.6.1 -ti ‘subordinator-different subjects (subds)’ 6.6.3.6.2 -shpa and -shtin ‘subordinator-same subjects (subis)’ 6.6.3.6.3 -na-poss-kama ‘limitative construction’ 6.7 Clitics with phrasal or clausal scope 6.7.1 Clitics that position the utterance with regard to other utterances 6.7.1.1 -chu ‘interrogation, disjunction, negation (neg)’ 6.7.1.2 -ña ~ -na ‘discontinuative (disc)’ 6.7.1.3 -pish ‘additive (add)’ 6.7.1.4 -ra ‘continuative (cont)’ 6.7.1.5 -ta ‘sequential/interrogative (seq)’ 6.7.2 Clitics that position the speaker with regard to the utterance 6.7.2.1 -mi, -shi, and -chi (evidentials) 6.7.2.2 -a(riya) ‘emphatic (emph)’ 6.8 Text sample 6.9 Literature 7. Cuzco Quechua Raúl Bendezú-Araujo and Jorge Acurio-Palma 7.1 General information 7.2 Internal variation 7.3 Phonology and phonotactics 7.3.1 Vowel phonemes 7.3.2 Consonant phonemes 7.3.3 Syllable structure and root structure 7.3.4 Phonological processes 7.3.5 Stress assignment 7.4 Parts of speech and transcategorical operations 7.4.1 Parts of speech 7.4.2 Transcategorical operations 7.4.2.1 Nominal derivation 7.4.2.2 Verbal derivation

187 187 187 187 187 188 188 188 188 188 188 189 190 190 190 190 190 191 191 191 192 192 192 192 192 192 193 193 193 193 193 193 194 194 195 196 196 196 196 196 197 197 198 198 198 198 200 200 200

xix

detailed contents 7.4.2.2.1 7.4.2.2.2 7.4.2.2.3 7.4.2.2.4 7.4.2.2.5 7.4.2.2.6 7.4.2.2.7

-cha ‘factive (fact)’ -ya ‘transformative (tf)’ -lli ‘self-transformative (stf)’ -tuku ‘simulative (simul)’ -na ‘relocative (reloc)’ -naya ‘desiderative (desr)’ -raya ‘durative (dur)’

7.5 Nominals 7.5.1 Subclasses 7.5.1.1 Pronouns 7.5.1.1.1 Personal pronouns 7.5.1.1.2 Demonstrative pronouns 7.5.1.1.3 Interrogative pronouns 7.5.1.2 Numerals 7.5.2 Nominal derivation 7.5.3 Nominal inflection 7.5.3.1 Person and number 7.5.3.2 Case 7.5.3.2.1 -ta ‘accusative (acc)’ 7.5.3.2.2 -pa ~ -q ‘genitive (gen)’ 7.5.3.2.3 -man ‘allative-dative (dat)’ 7.5.3.2.4 -pi ‘locative (loc)’ 7.5.3.2.5 -paq ‘benefactive (ben)’ 7.5.3.2.6 -manta ‘ablative (abl)’ 7.5.3.2.7 -kama ‘limitative (lim)’ 7.5.3.2.8 -wan ‘comitative-instrumental (com)’ 7.5.3.2.9 -rayku ‘causal (csl)’ 7.5.3.2.10 -ø ‘nominative’ 7.6 Verbs 7.6.1 Subclasses 7.6.1.1 Copular verbs 7.6.1.2 Intransitive verbs 7.6.1.3 Transitive and ditransitive verbs 7.6.2 Verbal derivation 7.6.2.1 Grammatical suffixes 7.6.2.1.1 -chi ‘causative (caus)’ 7.6.2.1.2 -ysi ‘accompaniment (acmp)’ 7.6.2.1.3 -pu ‘benefactive (ben)’ 7.6.2.1.4 -ku ‘reflexive (refl)’ 7.6.2.1.5 -na ‘reciprocal (recp)’ 7.6.2.2 Aspectual suffixes 7.6.2.2.1 -ri ‘inchoative (inch)’ 7.6.2.2.2 -paya ‘frequentative (freq)’ and -pa ‘repetitive (rep)’ 7.6.2.2.3 -ypari ~ -rpari ‘sudden action (sudd)’ 7.6.2.3 Directional suffixes 7.6.2.3.1 -yku ‘inward (iw)’ 7.6.2.3.2 -rqu ‘sudden action (sudd)’ 7.6.2.3.3 -ykacha ‘multidirectional (mult)’ 7.6.2.3.4 -mu ‘cislocative, translocative (cisl)’ 7.6.3 Inflectional categories of the verb 7.6.3.1 Person, number, and tense 7.6.3.1.1 Person

xx

200 200 201 201 201 201 201 201 201 202 202 202 202 202 202 203 203 203 203 204 204 204 204 204 204 204 205 205 205 205 205 206 206 206 206 206 207 207 207 207 207 207 207 208 208 208 208 208 208 209 209 209

detailed contents 7.6.3.1.2 Tense 7.6.3.1.3 Aspect

7.6.3.2 Mood 7.7 Clitics with phrasal or clausal scope 7.7.1 Evidentials 7.7.2 -chusina ‘dubitative (dub)’ 7.7.3 -puni ‘certainty (cert)’ 7.7.4 -qa and -ri ‘topic (top)’ 7.7.5 -chu ‘interrogative (q)’ 7.7.6 -chu ‘negative (neg)’ 7.7.7 -pis ~ -pas ‘additive (add)’ 7.7.8 -taq ‘contrastive (contr)’ 7.7.9 -raq ‘continuative (cont)’ and -ña ‘discontinuative (disc)’ 7.7.10 -lla ‘limitative (lim)’ 7.7.11 -yá ‘emphatic (emph)’ 7.8 Basic syntax 7.8.1 Grammatical relations 7.8.2 Constituent order 7.8.3 Major clause types 7.8.4 Coordination 7.8.5 Clausal complementation 7.8.6 Relativization 7.8.7 Negation 7.9 Text sample 7.10 Literature 8. Aymara Matt Coler 8.1 Introduction 8.2 Internal variation 8.3 Phonology and phonotactics 8.3.1 Vowel phonemes 8.3.2 Consonant phonemes 8.3.3 Syllable structure and root structure 8.3.4 Major morphophonological processes 8.3.4.1 Syllabic reduction of /aya/ and /uyu/ 8.3.4.2 Vowel deletion 8.3.4.2.1 Phonotactic vowel deletion 8.3.4.2.2 Syntactic vowel deletion 8.3.4.2.3 Lexically specified vowel deletion 8.3.4.3 Laryngeal harmony 8.3.4.4 Uvular harmony 8.3.4.5 Neutralization of non-modal plosives 8.3.5 Stress assignment 8.4 Parts of speech and transcategorical operations 8.4.1 Parts of speech 8.4.1.1 Nouns 8.4.1.2 Verbs 8.4.1.3 Particles 8.4.2 Transcategorical operations 8.4.2.1 Verbalization of nouns 8.4.2.1.1 -: or -ø ‘copulative verbalizer (cop.vblz)’

209 210 210 211 211 211 212 212 212 213 213 213 213 213 214 214 214 214 215 215 216 217 217 217 218 219 219 219 220 220 220 220 220 220 221 221 221 221 222 222 222 222 223 223 223 223 224 224 224 224

xxi

detailed contents 8.4.2.1.2 8.4.2.1.3 8.4.2.1.4 8.4.2.1.5 8.4.2.1.6 8.4.2.1.7 8.4.2.1.8

-c ka ‘locative verbalizer (loc.vblz)’ -cha ‘factive (fact)’ -ncha ‘saturator (sat)’ -pta ‘transformative (tf)’ -kipta ‘accelerated transformative (atf)’ -si ‘reflexive (refl)’ -tata ‘propagative, dispersal or broadcast motion (incep)’ 8.4.2.2 Nominalization of verbs 8.4.2.2.1 -ña ‘action nominalizer (nmlz)’ 8.4.2.2.2 -ta ‘resultative (re)’ 8.4.2.2.3 -iri ‘agentive (ag)’ 8.4.2.2.4 -wi ‘locality nominalizer (loc.nmlz)’ 8.4.2.2.5 -sa, -sina, and -ipana ‘subordinators (sub)’

8.5 Nominals 8.5.1 Subclasses 8.5.1.1 Personal and demonstrative pronouns 8.5.1.1.1 Personal pronouns 8.5.1.1.2 Demonstrative pronouns 8.5.1.2 Interrogative pronouns 8.5.1.3 Numerals 8.5.1.4 Common and proper nouns 8.5.1.5 Semantic adjectives 8.5.2 Nominal derivational-like suffixes 8.5.2.1 -ita, -itu, -cha, -lla ‘diminutive (dim)’ 8.5.2.2 -c chapi ‘delimitative (dl)’ 8.5.2.3 -wisa ‘privative (priv)’ 8.5.2.4 -rara ‘augmentative (aug)’ 8.5.3 Nominal inflectional-like suffixes 8.5.3.1 Location suffixes 8.5.3.2 -ni ‘possessor (poss)’ 8.5.3.3 Possessive paradigm 8.5.3.4 -naka ‘plural (pl)’ 8.5.3.5 -c pacha ‘reciprocal (recp)’ 8.5.3.6 Case suffixes 8.6 Verbs 8.6.1 Subclasses 8.6.1.1 Interrogative pronouns 8.6.1.2 The verb sa- ‘to say’ 8.6.2 Verbal derivational-like suffixes 8.6.2.1 Direction of motion suffixes 8.6.2.1.1 -nta ‘inward (iw)’ 8.6.2.1.2 -c su ‘outward (ow)’ 8.6.2.1.3 -qa ‘downward and dividing (dw)’ 8.6.2.1.4 -c xa ‘atop (atop)’ 8.6.2.1.5 -c ta ‘upward (uw)’ 8.6.2.2 Spatial location suffixes 8.6.2.3 -c kata ‘motion across or in front (ma)’ 8.6.2.4 -kipa ‘perifactive or contour motion (prf)’ 8.6.2.5 -muku ‘distancer (dst)’ 8.6.2.6 -c nuqa ‘placer (pla)’ 8.6.2.7 -naqa ‘diffuse motion, oscillative (df)’

xxii

224 224 225 225 225 225 225 225 225 225 225 225 226 226 226 226 226 227 227 227 227 228 228 228 228 228 228 228 228 228 228 229 229 229 230 230 230 230 231 231 231 231 231 231 231 231 232 232 232 232 232

detailed contents

8.6.3 8.6.4

8.6.5

8.6.6

8.6.2.8 8.6.2.9 8.6.2.10 8.6.2.11

-tata ‘inceptive (incep)’ -c tapi ‘concentrative, congregative, gathering motion (cc)’ -c χata ‘above (above)’ -ni ‘cislocative, translocative (cisl)’ Multiplying and reversing suffixes 8.6.3.1 -ra ‘multiplier/reverser (mlt/rv)’ 8.6.3.2 -rpaya ‘multiplier/reverser (mlt/rv)’ Valency-increasing suffixes 8.6.4.1 -c xasi ‘perdurative (pdr)’ 8.6.4.2 -rapi ‘beneficiary (bn)’ 8.6.4.3 -raqa ‘detrimental (dtm)’ 8.6.4.4 -c t’a ‘applicative (mom)’ 8.6.4.5 -ya ‘causative (caus)’ 8.6.4.6 -si ‘reflexive (refl)’ Aspectual suffixes 8.6.5.1 -c ch’uki ‘fossilized intensifier (int)’ 8.6.5.2 -c chu ‘factive completive (fact.compl)’ 8.6.5.3 -wxa or -xa ‘completive (compl)’ 8.6.5.4 -c ka ‘non-completive (ncompl)’ 8.6.5.5 -ska ‘progressive (prog)’ 8.6.5.6 -c t’a ‘momentaneous (mom)’ 8.6.5.7 -tata ‘inceptive (incep)’ Other derivational-like suffixes 8.6.6.1 -xama ‘verbal simulative (vbl.simul)’ 8.6.6.2 -xwa ‘category buffer (bfr)’ 8.6.6.3 -paya ‘intensifier (int)’

8.7 Inflectional-like categories of the verb 8.7.1 Person/tense and evidentiality 8.7.2 Number 8.7.3 Mood and modality 8.7.3.1 Counterfactual mood 8.7.3.2 Evidentiality 8.7.3.3 Event modality 8.7.3.4 Imperative 8.7.4 Order of the inflectional-like suffixes 8.8 Suffixes not subcategorized for lexical category 8.8.1 Stem-external word-level suffixes 8.8.2 Phrase-final suffixes 8.9 Basic syntax 8.9.1 Grammatical relations 8.9.2 Constituent order 8.9.3 Major clause types 8.9.3.1 Main clauses 8.9.3.2 Complement clauses 8.9.3.3 Relative clauses 8.9.3.4 Adverbial clauses 8.9.3.5 Nominalizations that go directly with the verb 8.9.4 Coordination: conjunction and disjunction 8.9.5 Clausal complementation 8.9.6 Relativization 8.9.7 Negation

232 232 232 232 232 233 233 233 233 233 233 233 233 233 233 233 234 234 234 234 234 234 234 234 234 234 234 235 235 235 236 236 237 237 237 238 238 238 239 239 239 239 239 239 239 240 240 240 240 240 240

xxiii

detailed contents 8.10 Lexical and semantic properties 8.10.1 Ambivalent roots 8.10.2 Verbs of taking 8.10.3 Integration of loanwords 8.11 Text sample 8.12 Literature 9. Jaqaru Matt Coler 9.1 Introduction 9.2 Brief remarks on comparisons with Aymara 9.3 Phonology and phonotactics 9.4 Parts of speech and transcategorical operations 9.4.1 Verbalization of nouns 9.4.1.1 -wa ‘copulative verbalizer (cop.vblz)’ 9.4.1.2 -c k ‘locative verbalizer (loc.vblz)’ 9.4.1.3 -chi ‘factive (fac)’ 9.4.2 Nominalization of verbs 9.4.2.1 -ata ‘resultative (res)’ 9.4.2.2 -iri ‘agentive (ag)’ 9.4.2.3 -nushu ‘purposive (purp)’ 9.4.2.4 -c mashi ‘rartnership (part)’ 9.4.2.5 -c cha ‘delimitative (dl)’ 9.4.2.6 Subordination 9.5 Nominals 9.5.1 Subclasses 9.5.1.1 Personal and demonstrative pronouns 9.5.1.2 Numerals 9.5.1.3 Interrogative pronouns 9.5.1.4 Semantic adjectives 9.5.2 Nominal inflectional-like suffixes 9.5.2.1 Possessive paradigm 9.5.2.2 -c kuna ‘Plural (pl)’ 9.5.2.3 -ishi ‘reflexive (refl)’ 9.5.2.4 Location 9.5.2.4.1 -c nuri ‘inessive (iness)’ 9.5.2.4.2 -asu ‘adessive (adess)’ 9.5.2.4.3 -c sana ‘atop (atop)’ 9.5.2.5 Case 9.5.3 Nominal derivational-like suffixes 9.5.3.1 -uña ‘diminutive (dim)’ 9.5.3.2 -c cha ‘delimitative (dl)’ 9.5.3.3 -wishi ‘privative (priv)’ 9.6 Verbs 9.6.1 Subclasses 9.6.2 Verbal derivational-like suffixes 9.6.2.1 Direction of motion and spatial location suffixes 9.6.2.1.1 -c pta ‘upwards (uw)’ 9.6.2.1.2 -natsa ‘downward (dw)’ 9.6.2.1.3 -c nuqu ‘placer (pla)’

xxiv

241 241 241 241 242 243 245 245 245 246 247 248 248 248 248 248 249 249 249 249 249 249 250 250 250 250 251 251 251 252 252 252 252 252 253 253 253 254 254 254 254 255 255 255 255 255 255 255

detailed contents 9.6.2.1.4 9.6.2.1.5 9.6.2.1.6 9.6.2.1.7 9.6.2.1.8 9.6.2.1.9

-ru ‘i nward (iw)’ -c shu ‘outward (ow)’ -c kata ‘motion across (ma)’ -kipa ‘contour Motion (cm)’ -naqa ‘diffuse motion (df)’ -c muku ‘doing something rapidly on the way to do something else (dsr)’ 9.6.2.1.10 -c thapa ‘above (above)’ 9.6.2.1.11 -c kusu ‘as one goes (as.one.goes)’ 9.6.2.1.12 -c qhasa ‘intensifier (int)’ 9.6.2.1.13 -rpaya ‘next to (next.to)’ 9.6.2.2 Valency-increasing suffixes 9.6.2.2.1 -ya ‘causative (caus)’ 9.6.2.2.2 -ishi ‘reflexive (refl)’ 9.6.2.3 Aspectual suffixes 9.6.2.3.1 -w ‘completive (compl)’ 9.6.2.3.2 -c k ‘non-completive (ncompl)’ 9.6.2.3.3 -c tata ‘Inceptive (incep)’ 9.6.2.3.4 -qhullu ‘inceptive (incep)’ 9.6.2.2.5 -as ‘cessative (cess)’ 9.6.2.2.6 -t’a ‘momentaneous (mom)’ 9.6.2.2.7 -c kha ‘momentaneous (mom)’ 9.6.2.2.8 -c qha ‘repetitive (repet)’ 9.6.3 Inflectional-like categories of the verb 9.6.3.1 Person/tense and evidentiality 9.6.3.1.1 Simple tense 9.6.3.1.2 Proximal past tense 9.6.3.1.3 Distal past tense 9.6.3.1.4 Future tense 9.6.3.2 -rqaya ‘verbal plural (vpl)’ 9.6.3 Mood and modality 9.6.3.1 Counterfactual mood 9.6.3.2 Evidentiality 9.6.3.3 Event modality 9.6.3.4 Imperative

9.7 Suffixes not subcategorized for lexical category 9.8 Basic syntax 9.8.1 Grammatical relations 9.8.2 Constituent order 9.8.3 Major clause types 9.8.3.1 Main clauses 9.8.3.2 Relative clauses 9.8.3.3 Nominalizations that go directly with the verb 9.8.4 Coordination: conjunction and disjunction 9.8.5 Negation 9.9 Lexical and semantic properties 9.9.1 Ambivalent roots 9.9.2 Verbs of taking 9.10 Literature

255 255 256 256 256 256 256 256 256 256 257 257 257 257 257 257 257 257 257 258 258 258 258 258 258 259 259 260 260 260 260 260 261 261 261 262 262 262 263 263 263 263 263 263 264 264 264 264

xxv

detailed contents 10.Uru and Chipaya Katja Hannß 10.1 General information 10.2 Internal variation 10.3 Phonology and phonotactics 10.3.1 Vowel phonemes 10.3.2 Consonant phonemes 10.3.3 Syllable structure and root structure 10.3.4 Major morphophonological processes 10.3.4.1 Vowel elision 10.3.4.2 Assimilation 10.3.4.3 Vocalization and semi-consonantalization 10.3.4.4 Metathesis 10.3.5 Stress assignment 10.4 Parts of speech and transcategorical operations 10.4.1 Parts of speech 10.4.1.1 Nouns 10.4.1.2 Adjectives 10.4.1.3 Verbs 10.4.2 Transcategorical operations 10.5 Nominals 10.5.1 Subclasses 10.5.1.1 Pronouns 10.5.1.1.1 Personal pronouns 10.5.1.1.2 Possessive pronouns 10.5.1.1.3 Demonstrative pronouns 10.5.1.1.4 Interrogative pronouns 10.5.1.1.5 Indefinite pronouns 10.5.1.2 Numerals 10.5.1.3 Common nouns 10.5.2 Nominal derivation 10.5.2.1 Uru -chi, Chipaya -chis ́ ‘possessive (poss)’ 10.5.2.2 Chipaya -lla ‘diminutive (dim)’ 10.5.2.3 Chipaya -rara ‘augmentative (aug)’ 10.5.3 Nominal inflection 10.5.3.1 Plural 10.5.3.2 Case 10.5.3.2.1 Chipaya -t, -z´, and -a, Uru -t, -s, ‘genitive (gen)’ 10.5.3.2.2 Chipaya -kis ́ ‘proximal locative-allative

(loc/all.prox)’ and -kin(a) ‘distal locative-allative (loc/all.dist),’ Uru -kis ~ -ski ‘locative-allative (loc/all)’ and kina ‘allative (all)’ 10.5.3.2.3 Chipaya -kama, Uru -kama ‘limitative (lim)’ 10.5.3.2.4 Chipaya -kis ́tan(a), Uru -kistana ‘ablative (abl)’ 10.5.3.2.5 Chipaya -ke: ‘approximative (approx)’ 10.5.3.2.6 Chipaya -(ta)xapa ‘benefactive, purposive (ben),’ Uru -taki ‘benefactive (ben),’ -xapa ‘purposive (purp)’ 10.5.3.2.7 Chipaya -z´tan, Uru -stani ‘comitative-instrumental (com)’ 10.5.3.2.8 Chipaya -z´ta ‘simulative (simul)’ 10.5.3.2.9 Chipaya -layku ‘causal (csl)’ 10.5.3.2.10 Chipaya -pora ‘interactive (inter)’

xxvi

265 265 266 266 266 267 269 270 270 271 271 272 273 274 274 274 275 277 278 280 280 281 281 282 283 283 284 284 285 285 286 286 286 286 287 287 287

287 289 289 290 290 291 291 291 291

detailed contents

10.6 Verbs 10.6.1 Subclasses 10.6.2 Verbal derivation 10.6.2.1 Aspect: -ay ‘prioritizing (prior),’ -tan ‘obligative (oblg),’ -ta ‘inceptive (incep),’ -maya ‘imminent (imm)’

10.6.2.2 Valency-changing derivation: Chipaya -n(a) ‘transitivizer

(tr), -z´ ~ -s ́ ‘reflexive (refl),’ -as ‘reciprocal (recp),’ -qat ‘causative (caus),’ -z´hin ‘benefactive (ben),’ Uru -sna ‘causative (caus)’ 10.6.2.3 Movement: Chipaya -z´ki ‘cislocative, translocative (cisl), Uru and Chipaya -lay ‘aimless movement’ 10.6.3 Inflectional categories of the verb 10.6.3.1 Person and tense/aspect marking 10.6.3.2 Mood 10.6.3.3 Subordination 10.6.3.4 Order of the inflectional morphemes

10.7 Clitics with phrasal or clausal scope 10.8 Basic syntax 10.8.1 Grammatical relations 10.8.2 Constituent order 10.8.3 Major clause types 10.8.3.1 Declarative clauses 10.8.3.2 Non-verbal predicate clauses 10.8.3.3 Existential predicate clauses 10.8.3.4 Locational predicate clauses 10.8.3.5 Possessive predicate clauses 10.8.3.6 Attributive predicate clauses 10.8.3.7 Imperative clauses 10.8.3.8 Interrogative clauses 10.8.4 Coordination 10.8.5 Clausal complementation 10.8.6 Relativization 10.8.7 Negation 10.9 Lexical and semantic properties 10.10 Text sample 10.11 Literature

11. Mochica Matthias Urban 11.1 General information 11.1.1 Geography and environment 11.1.2 Speaker population and language endangerment 11.2 Internal variation and classification 11.3 Phonology and phonotactics 11.3.1 Vowel phonemes 11.3.2 Consonant phonemes 11.3.3 Syllable structure and root structure; phonotactic constraints 11.3.4 Major morphophonological processes 11.3.5 Stress assignment 11.4 Parts of speech and transcategorical operations 11.4.1 Parts of speech

292 292 292 292

292 293 294 294 297 297 298 299 301 301 302 303 303 303 303 303 303 304 304 304 305 305 306 307 307 310 310 311 311 311 312 312 313 315 315 318 318 318 319 319

xxvii

detailed contents 11.4.2 Transcategorical operations 11.4.2.1 Nominal derivation 11.4.2.1.1 ‘nominalizer (nmlz)’ 11.4.2.1.2 ~ ‘instrument/locative nominalizer (ins/loc.nmlz)’

11.4.2.1.3 ‘agentive (ag)’ 11.4.2.1.4 ‘resultative (res)’ 11.4.2.1.5 ‘accompaniment (acmp)’

11.4.2.2 Verbal derivation 11.5 Nominals 11.5.1 Subclasses of nominals 11.5.1.1 Pronouns and other pro-forms 11.5.1.1.1 Personal pronouns 11.5.1.1.2 Demonstrative pronouns 11.5.1.1.3 Interrogative and indefinite pronouns 11.5.1.2 Numerals and numeral classifiers 11.5.1.3 Common nouns 11.5.2 Nominal derivation 11.5.2.1 Possession 11.5.2.2 Other processes of noun–noun derivation 11.5.3 Nominal inflection 11.5.3.1 Number 11.5.3.2 Case 11.5.3.2.1 ‘nominative (nom)’ 11.5.3.2.2 ~ ~ ~ ~ ~ ‘oblique 11.5.3.2.3 11.5.3.2.4 11.5.3.2.5 11.5.3.2.6

(obl)’ ‘locative (loc)’ ≤n(i)c> ‘locative-allative (loc/all)’ ≤ich> ‘ablative (abl)’ ≤n> ‘agentive (agt)’

11.6 Verbs 11.6.1 Verb subclasses 11.6.2 Verbal derivation 11.6.2.1 ‘causative (caus)’ 11.6.2.2 ‘benefactive (ben)’ 11.6.2.3 ~ , , ‘passive (pass)’ 11.6.2.4 ‘continuative (cont)’ 11.6.2.5 ~ ‘impersonal (impr)’ 11.6.2.6 Affix order 11.6.3 Inflection categories of the verb 11.6.3.1 Number 11.6.3.2 Imperativity 11.7 Clitics with phrasal or clausal scope 11.7.1 Personal reference clitics 11.7.2 Copula clitics 11.7.3 Tense marking and discourse clitics 11.8 Basic syntax 11.8.1 The expression of tense, aspect, and mood 11.8.2 Grammatical relations 11.8.3 Constituent order 11.8.4 Major clause types 11.8.4.1 Declarative clauses

xxviii

319 319 319 319 319 319 320 320 320 320 320 320 321 321 322 322 322 322 322 323 323 323 324 324 324 324 324 324 325 325 325 325 325 326 326 326 326 326 326 326 327 327 327 328 328 328 329 329 330 330

detailed contents 11.8.4.2 Interrogative clauses 11.8.4.3 Exclamative clauses

11.8.5 Coordination 11.8.5.1 ‘sequential (seq)’ 11.8.5.2 ‘so, as’ 11.8.5.3 ‘also, and, too’ 11.8.5.4 ‘also, and, too’ 11.8.5.5 ‘or’ 11.8.6 Clausal complementation 11.8.6.1 Purposive clauses in and 11.8.6.2 Temporal, causal, and concessive clauses in and

11.8.6.3 Temporal clauses in 11.8.6.4 Infinitival clauses with 11.8.6.5 Conditional clauses in and

11.8.7 Relativization 11.8.8 Negation 11.9 Lexical and semantic properties 11.10 Text sample 11.11 Literature

12. Puquina Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar 12.1 General information 12.1.1 Geographical distribution and sociohistorical context 12.1.2 The Rituale seu Manuale Peruanum (1607) and other sources of information about Puquina

12.2 Internal variation 12.3 Phonology and phonotactics 12.3.1 Vowel phonemes 12.3.2 Consonant phonemes 12.3.3 Syllable structure and root structure 12.3.4 Major morphophonological processes 12.3.4.1 Phonotactic vowel deletion 12.3.4.2 Sandhi 12.3.4.3 Voiced and voiceless allomorphs and Puquina verb classes 12.4 Parts of speech and transcategorical operations 12.4.1 Parts of speech 12.4.2 Transcategorical operations 12.4.2.1 -no ‘infinitive (inf)’ 12.4.2.2 -eno ‘agentive (ag)’ 12.4.2.3 -so ~ -su ‘participial nominalizer (ptcp)’ 12.4.2.4 -nosu ‘future nominalizer (fut.nmlz)’ 12.5 Nominals 12.5.1 Subclasses 12.5.1.1 Pronouns 12.5.1.2 Numerals 12.5.1.3 Common nouns 12.5.2 Nominal derivation 12.5.3 Nominal inflection 12.5.3.1 Number

330 330 330 330 330 330 331 331 331 331 331 331 331 332 332 332 332 333 334 335 335 335 340 341 341 342 343 344 344 344 345 346 349 349 349 349 349 350 350 350 350 350 352 352 353 353 353

xxix

detailed contents 12.5.3.2 Case 12.5.3.2.1 12.5.3.2.2 12.5.3.2.3 12.5.3.2.4 12.5.3.2.5 12.5.3.2.6 12.5.3.2.7 12.5.3.2.8 12.5.3.2.9 12.5.3.2.10 12.5.3.2.11

-ø (no overt case marking) -ch(a) ‘hierarchy marker (hm)’ -wa ‘genitive (gen)’ -na and -(u)t ‘locative (loc)’ -guta ‘allative (all)’ -kama and -wen(o) ‘limitative (lim)’ -ch ‘ablative (abl)’ -m ‘comitative-instrumental, associative (com)’ -wa ‘benefactive, purposive (ben, purp)’ -wichna ‘causal (csl)’ -ye ‘feminine vocative (voc.f)’, -re ‘masculine vocative (voc.m)’ 12.5.3.2.12 Other case-like markers

12.6 Verbs 12.6.1 Subclasses 12.6.1.1 Copula constructions 12.6.2 Verbal derivation 12.6.2.1 ( h)e- ‘reverser (rev)’ 12.6.2.2 -toch(u) ~ -tos(u) ‘cislocative, translocative (cisl)’ 12.6.2.3 -schi ‘completive (compl)’ 12.6.2.4 -na ‘causative (caus)’ 12.6.2.5 -ska ‘reflexive (refl)’ 12.6.2.6 -si ‘medial, reflexive (med),’ -ti ‘reciprocal (recp)’ 12.6.2.7 -ke ‘applicative (appl)’ 12.6.2.8 -ke ‘future (fut)’ 12.6.2.9 -a ‘past (pst)’ 12.6.2.10 -lli ‘transformative, to act unnaturally (tf)’ 12.6.2.11 , , and 12.6.3 Verbal inflection 12.6.3.1 Person marking and the inverse construction 12.6.3.2 Tense and aspect 12.6.3.3 Mood 12.6.3.3.1 -nch ‘declarative (decl)’ 12.6.3.3.2 -ta ‘2nd person imperative (imp)’ and -suma

‘2nd person subject, 1st person object imperative (imp)’ 12.6.3.3.3 -iska ‘potential (pot) and -anta ‘optative (opt)’

12.7 Morphemes with clausal scope 12.7.1 Topic markers 12.7.2 -i ‘interrogative (q)’ 12.7.3 -hamp ‘additive (add)’ 12.7.4 -nch ‘declarative (decl)’ 12.7.5 -nk(a) ‘restrictive-limitative (rstr)’ 12.7.6 -ñao ‘personal belief (emph)’ 12.8 Basic syntax 12.8.1 Grammatical relations 12.8.2 Constituent order 12.8.2.1 Word order in the noun phrase 12.8.2.2 Word order in the clause 12.8.3 Major clause types 12.8.4 Coordination 12.8.4.1 Coordination of noun phrases 12.8.4.2 Coordination of clauses

xxx

353 353 354 354 355 355 355 356 356 356 356 356 356 357 357 357 358 358 359 359 359 359 359 360 360 360 360 361 361 361 363 363 363 363 363 364 364 365 366 366 367 367 368 368 368 368 368 369 369 369 369

detailed contents 12.8.5 Embedded clauses 12.8.5.1 Complement clauses 12.8.5.2 Conditional clauses 12.8.5.3 Purpose clauses 12.8.5.4 Same-subject clauses 12.8.5.5 Concessive clauses 12.8.6 Relative clauses 12.8.7 Negation 12.9 Lexical and semantic properties 12.10 Text sample: Pater Noster (Text R) 12.11 Literature 13. Híbito and Cholón Astrid Alexander-Bakkerus 13.1 General information 13.1.1 Sources 13.1.2 The Híbito–Cholón relationship 13.2 Phonology and phonotactics 13.2.1 Hypothetical vowel inventory 13.2.2 Possible consonant inventory 13.2.3 Syllable structure and root structure 13.2.4 Major morphophonological processes 13.2.4.1 Vowel harmony 13.2.4.2 Vowel suppression 13.2.4.3 Stem alternation 13.2.5 Stress assignment 13.3 Parts of speech and transcategorical operations 13.3.1 Parts of speech 13.3.2 Transcategorical operations 13.3.2.1 Nominal derivation 13.3.2.1.1 -(e)ch ‘factive (fact)’ 13.3.2.1.2 -lam ‘future nominalizer (fut.nmlz1)’ 13.3.2.1.3 -(ŋ)o ‘future nominalizer (fut.nmlz2)’ 13.3.2.1.4 -(w)uch ‘agentive (ag)’ 13.3.2.2 Verbal derivation 13.3.2.2.1 -(a)ŋ ‘imperfective (ipfv)’ 13.3.2.2.2 -w(e) ‘verbalizer (vblz)’ 13.3.2.2.3 -(k)e(h) ‘causative (caus)’ 13.3.2.3 Other derivational processes 13.4 Nominals 13.4.1 Nominal subclasses 13.4.1.1 Adjectives 13.4.1.2 Pronouns 13.4.1.3 Interrogatives 13.4.1.4 Numerals and classifiers 13.4.2 Nominal inflection 13.4.2.1 Person 13.4.2.2 Number 13.4.2.3 Case 13.4.2.4 Diminutive and indefinite marking 13.4.2.5 The morpheme -pulleŋ 13.5 Verbs

369 369 370 370 370 371 371 372 372 373 391 392 392 392 392 393 394 395 395 396 396 396 397 397 397 397 397 397 397 398 398 398 398 398 398 398 398 398 398 399 399 399 400 400 401 401 402 402 402 402

xxxi

detailed contents 13.5.1 Verb subclasses 13.5.2 Verbal derivation 13.5.2.1 -(k)e(h) ‘causative (caus)’ 13.5.2.2 -ka(h) ‘causative (caus)’ 13.5.2.3 -its ‘passive (pass)’ 13.5.2.4 -(k)ia(h) ‘repetitive (repet)’ 13.5.2.5 -n(o) ‘reflexive (refl)’ 13.5.3 Inflectional categories of the verb 13.5.3.1 Person 13.5.3.2 Tense 13.5.3.2.1 Past-tense marking and stem extension 13.5.3.2.2 Future 13.5.3.3 Aspect 13.5.3.4 Imperativity 13.5.3.5 Applicative 13.5.3.6 Order of affixes Suffixes and clitics with phrasal or clausal scope 13.6.1 Speech markers 13.6.2 Exclamatives 13.6.3 Question markers 13.6.4 Adverbial suffixes 13.6.5 Additive 13.6.6 Emphasis 13.6.7 Topic Basic syntax 13.7.1 Grammatical relations 13.7.2 Constituent order 13.7.3 Major clause types 13.7.4 Coordination 13.7.5 Clausal complementation and other subordination processes 13.7.5.1 Clausal complements 13.7.5.2 Case-marked nominalizations 13.7.5.3 Irrealis clauses 13.7.5.4 Purposive and resultative clauses 13.7.5.5 Concessive and conditional clauses 13.7.6 Relativization 13.7.7 Negation Lexical and semantic properties Text sample Literature Appendix 13.11.1 Híbito data from Mrs Gutiérrez 13.11.2 Híbito data from Martínez Compañón (1985[1782–1790]) and Günter

402 403 403 403 403 403 403 403 404 406 406 406 406 406 407 407 407 408 408 408 408 409 409 409 409 409 410 411 411 412 412 412 412 413 413 413 414 414 414 416 416 416

14. Small and extinct languages of Northern Peru Matthias Urban 14.1 Introduction 14.2 Overview and distribution of the languages 14.3 The documentary record and methods of analyzing linguistic data from the Northern Peruvian languages

419

13.6

13.7

13.8 13.9 13.10 13.11

Tessmann (1930)

xxxii

416

419 419 428

detailed contents 14.4 14.5 14.6 14.7 14.8

Common structural features and vocabulary Language and the pre-Hispanic cultures of Northern Peru Conclusion Further reading Appendix A. A new transcription of the Madrid version of Martínez Compañón’s “plan” 14.9 Appendix B. Tessmann’s (1930: 458–459, 547) Híbito and Cholón wordlists, collected by J. W. Harmston at Pachiza 15. Kallawaya Pieter Muysken 15.1 General information 15.1.1 Introduction 15.1.2 Contemporary sociolinguistic setting 15.1.3 Historical and linguistic background information on the Kallawaya people and language 15.1.3.1 Historical perspective on the region and the Kallawaya people 15.1.3.2 Puquina 15.1.3.3 Quechua 15.1.3.4 Aymara 15.1.4 Variation

15.2 Phonology and phonotactics 15.2.1 Vowel phonemes 15.2.2 Consonant phonemes 15.2.3 Word and syllable structure 15.2.4 Stress assignment 15.2.5 Orthography 15.3 Parts of speech 15.4 Nominals 15.4.1 Nominal derivation 15.4.2 Nominal inflection 15.4.2.1 Person 15.4.2.2 Number 15.4.2.3 Case 15.5 Verbs 15.5.1 Verbal derivation 15.5.2 Person reference 15.6 Basic syntax 15.6.1 Constituent order 15.6.2 Conjunctions and subordination 15.7 Lexical properties 15.8 Trying to interpret the genesis of Kallawaya 15.8.1 Original function of Kallawaya and possible associated properties 15.8.1.1 Was Kallawaya a community language? 15.8.1.2 Is the use of Kallawaya restricted by gender? 15.8.1.3 Does Kallawaya show signs of simplification, as in a pidgin? 15.8.1.4 Does Kallawaya show signs of conscious manipulation? 15.8.1.5 Does Kallawaya signal a specific ethnic identity? 15.8.1.6 How fixed or frozen was Kallawaya?

430 432 434 434 435 437 438 438 438 439 441 441 442 443 443 443 444 444 444 444 445 445 445 445 446 446 446 447 447 449 449 450 450 450 452 452 453 453 453 454 454 454 454 454

xxxiii

detailed contents 15.8.1.7 Does Kallawaya have the typical features of a secret language?

15.8.2 The emergence of Kallawaya 15.8.3 The unknown lexical component: how likely is it that it may contain 15.9 15.10 15.11 15.12

Puquina roots?

Conclusions and suggestions for further research Text samples Literature Appendices

457 458 458 460 462

16. The Andean Spanish of Southern Peru and Bolivia Luis Andrade Ciudad 16.1 Introduction 16.2 Phonetic–phonological features 16.3 Morphosyntactic features 16.4 Lexical and discourse-pragmatic trends 16.5 Concluding remarks

469

PART III: Comparative studies

483

17. Central Andean Segmental Phonologies in Continental Perspective Lev Michael and Allegra Robertson 17.1 Introduction 17.2 A qualitative overview of the phonological inventories of Central Andean languages 17.3 Quantitative explorations of phonological areality 17.3.1 Naive Bayes Classifiers and the “Core and Periphery” method 17.3.2 Constructing exclusive and shared profiles in three-way NBC analysis 17.3.3 Data and modeling 17.4 The phonologies of Central Andean languages in continental perspective 17.4.1 South-Central and North-Central Andean core analyses 17.4.2 The Southern Cone Phonological Area 17.4.3 The North-Central Andean Phonological Area 17.4.4 Pan-Central Andean phonological features and phonological Areality

485

in Western South America

17.5 Central Andean phonological areality: discussion and conclusion 17.6 Appendices 18. The morphology of the nominal domain in the languages of the Central Andes Olga Krasnoukhova 18.1 Introduction 18.2 (Pro)Nominal number 18.2.1 Pronominal number 18.2.2 Nominal number 18.3 Nominal possession marking 18.4 Case 18.5 Gender 18.6 Demonstratives 18.7 Adjectives 18.8 Summary of observations

xxxiv

454 455

469 470 472 476 480

485 486 488 494 495 497 497 497 502 507 510 516 518 532 532 532 532 533 537 542 544 545 547 549

detailed contents 19. The grammar of the verb in the languages of the Central Andes Matthias Pache 19.1 Introduction 19.2 Background 19.2.1 Characteristics of the verb in Central Andean languages 19.2.2 Verb classes and complexities 19.2.3 Features that are absent or marginally present 19.3 Participants 19.3.1 Prefixes, suffixes, unbound forms: indicating verbal person 19.3.1.1 Prefixes 19.3.1.2 Suffixes 19.3.1.2.1 Non-fusional suffixes 19.3.1.2.2 Suffixes fusing subject and object person 19.3.1.2.3 Suffixes fusing person and tense/mood 19.3.1.3 Unbound forms 19.3.1.4 Mixed patterns 19.3.1.4.1 Unbound forms and affixes 19.3.1.4.2 Unbound forms and clitics 19.3.1.4.3 Verbal person markers preceding and following

551 552 552 553 553 554 554 554 555 555 555 555 556 556 556 556

the root

556 557 557 558 558 559 559 561 561 561

beneficiary, and assistive

562 563 563 564 565 566 568 568 569 569 570

19.3.2 Participant-related distinctions reflected in verbal person marking 19.3.2.1 1st and 2nd person versus 3rd person 19.3.2.2 1st person vs non-1st person 19.3.2.3 Gender distinctions 19.3.3 Clusivity 19.3.4 Number 19.3.5 Valency and valency changes 19.3.5.1 Valency-increasing operations 19.3.5.1.1 Introducing an agent argument: causative 19.3.5.1.2 Introducing an object argument: applicative, 19.3.5.2 Valency-reducing operations 19.3.5.2.1 Reflexivity, reciprocity, and middle voice 19.3.5.2.2 Passives and passive-like constructions

19.3.6 Inverse constructions 19.3.7 Switch-reference 19.4 Time and space 19.4.1 Tense 19.4.1.1 Future tense 19.4.1.2 Non-future tenses (simple tense, past tenses) 19.4.2 Aspectual notions 19.4.2.1 Perfective and imperfective, completive and non-completive, progressive

19.4.2.2 Habitual

551

19.4.3 Spatial notions 19.4.3.1 Cislocative, translocative, and associated motion 19.4.3.2 Other directions 19.5 Inner states and attitudes 19.5.1 Cognition 19.5.1.1 Evidentiality and epistemic modality 19.5.1.2 Cognitive distance/mirativity

572 573 573 574 575 576 576 576 577

xxxv

detailed contents

19.6

19.7

19.8 19.9

19.5.2 Emotion 19.5.2.1 Emotional engagement 19.5.2.2 Emotional distance 19.5.3 Desire and reluctance 19.5.4 Orders and commands 19.5.4.1 Bound morphemes expressing orders and commands 19.5.4.2 Unbound morphemes expressing orders and commands 19.5.5 Politeness 19.5.6 Purpose 19.5.6.1 Purposive constructions and switch-reference 19.5.6.2 Purposive constructions and motion verbs 19.5.7 Necessity Actions, states, and events in the abstract 19.6.1 The action nominalizer/infinitive 19.6.2 Nominalized verbs as complements 19.6.3 Further nominalizing suffixes for lexical derivation 19.6.4 Differences and similarities in nominalization ‘to be’ and other semantically light or empty verbs: existentials, copulas, and auxiliaries 19.7.1 Auxiliaries and light verbs 19.7.2 Existential verbs and copulas Negation 19.8.1 Preverbal negative adverbs 19.8.2 Negative enclitics or suffixes Final remarks: a comparative perspective 19.9.1 Differences and domains of diversity 19.9.2 Similarities and correspondences

20. Syntactic structures in the languages of the Central Andes Rik van Gijn 20.1 Basic building blocks 20.1.1 Grammatical and phonological words 20.1.2 Major lexical classes 20.2 Phrasal structures 20.2.1 Noun phrases 20.2.1.1 Word order within the NP 20.2.1.2 Marking of morphosyntactic relations within the NP 20.2.1.3 NP-internal phrasal structures 20.2.2 The verb phrase 20.2.3 Other phrasal structures 20.3 Clausal syntax 20.3.1 Constituent order 20.3.2 Clitic placement 20.3.3 Main-clause morphosyntactic alignment 20.4 Complex sentences 20.4.1 Clausal coordination 20.4.2 Subordination 20.5 Conclusions: syntax in the languages of the Central Andes

xxxvi

577 578 578 578 580 580 580 581 583 583 584 584 585 586 586 587 587 588 588 589 591 591 592 593 593 594 597 597 597 598 601 601 601 604 606 607 609 610 611 616 620 625 625 626 633

detailed contents 21. Discourse, information structure, and evidentiality in the Central Andes Karolina Grzech 21.1 Introduction 21.2 Background: relevant notions 21.2.1 Evidentiality and related notions 21.2.2 Information structure 21.2.3 Switch-reference and tail–head linkage 21.3 Evidentiality, information structure, and reference tracking in Quechuan 21.3.1 Evidential markers: evidence, validation, shared and exclusive knowledge 21.3.1.1 Direct evidence/Best Possible Grounds: -mi 21.3.1.2 Inference/conjecture: -chá 21.3.1.3 Reportative/hearsay -si 21.3.1.4 Interpersonal semantics of evidentials 21.3.1.5 Evidentials in Central Andean Quechua: a summary 21.3.2 Information structure 21.3.3 Tense marking, evidentiality, and mirativity 21.3.4 Switch-reference and tail–head linkage

21.4 Evidentiality, information structure, and discourse in Aymaran 21.4.1 Evidential distinctions in TAM marking 21.4.1.1 Tense and personal/non-personal knowledge 21.4.1.2 Aspect and conjectural/inferential evidentiality 21.4.2 Reported evidentiality 21.4.3 Independent suffixes: evidentiality, information structure, and discourse

21.4.3.1 Sentence suffixes in Aymara 21.4.3.2 Sentence suffixes in Jaqaru

21.4.4 Reference tracking and discourse structure 21.5 Discourse, information structure, and evidentiality in Uru–Chipaya 21.6 Discourse, evidentiality, and information structure in Central Andes: a summary 21.7 Conclusions 22. Linguistic complexity in the Central Andes Johanna Nichols 22.1 Introduction 22.2 Method and definitions 22.2.1 Measuring complexity 22.2.1.1 Enumerative complexity 22.2.1.2 Canonical complexity 22.2.1.3 Sample 22.2.2 Geography 22.2.3 Sociolinguistics and sociohistory 22.2.4 Central Andean language history 22.2.5 Important nearby languages 22.3 Results 22.3.1 Complexity and geography 22.3.2 Sociohistorical factors 22.3.3 Typological overview 22.4 Discussion 22.4.1 Inferring past sociolinguistics from complexity

635 635 636 636 637 637 638 638 639 640 641 643 644 644 647 649 651 651 651 652 654 655 655 658 659 660 662 665 667 667 668 668 668 669 669 669 672 673 675 675 676 677 677 678 678

xxxvii

detailed contents 22.4.2 Which is the oldest highland language? 22.5 Conclusions

PART IV: Language history

681

23. Expansions and language shift in prehistory Paul Heggarty 23.1 Introduction 23.1.1 Aims and scope 23.1.2 Large-scale geographical patterns 23.1.3 Not just linguistics: a coherent, cross-disciplinary prehistory 23.2 General principles 23.2.1 Causation: why do language families exist at all? 23.2.2 How languages spread: demographic or cultural mechanisms,

683

23.3 23.4 23.5 23.6

683 683 683 684 686 686

language shift and diversity

687 687 688 690 690 690 691 692 692 692 695

Ecuador

703 704 706 708 708 713 713 713

Altiplano

713 714

hypothesis on the origins of Aymaran

715 719 721 724

23.2.3 How language expansions proceed: micro- and macro-scales Andean exceptionalism and the vertical dimension Methodology: how to uncover language prehistory 23.4.1 Time-depth 23.4.2 Language family trees, branches . . . and migrations? Deep time: before the major families Quechuan 23.6.1 Stage 1. Abandoning “Quechua = Incas” 23.6.2 Stage 2. The “traditional” linguistic vision 23.6.3 Stage 3. Challenges to the tree model for Quechuan 23.6.4 Stage 4. Archeological challenges to Quechua IIB and the spread to

23.6.5 Stage 5. Challenges on causation: Huari as Quechua? 23.6.6 Stage 6. Recovering the role of the Incas 23.6.7 Colonial-era expansions 23.6.8 Summarizing the new hypothesis 23.6.9 State of the art 23.7 Aymaran (and Puquina) 23.7.1 A superficial association: Aymaran = Tiahuanaco 23.7.2 Linguistic evidence: Aymaran origins in Central Peru, not the 23.7.3 Puquina as the (main) language of Tihuanaco 23.7.4 If not Tihuanaco, then where and when? The traditional linguistic 23.7.5 Weaknesses and alternatives 23.7.6 When did Aymara spread across the Altiplano? 23.8 Envoi 24. Language ecologies and dynamics in the ancient Central Andes Matthias Urban 24.1 Introduction 24.2 Discontinuous language distributions, discontinuous territoriality, and vertical complementarity 24.3 The social ecology of the Quechuan–Aymaran dualism

xxxviii

679 679

725 725 727 731

detailed contents 24.4 Envoi: language in social space today 24.5 Conclusion

736 738

25. Language diffusion and state agency: Quechuan in Inca and colonial times César Itier 25.1 Introduction 25.2 Vernacular and vehicular languages in the Inca empire 25.2.1 Functions of vehicular Quechua 25.2.2 The “language of the Inca” and the “language of the valleys” 25.3 The colonial lengua general: expansion, vernacularization, and regionalization 25.3.1 Cities as foci of Quechuan spread 25.3.2 The criollos as Quechua speakers 25.3.3 Vernacularization 25.3.4 Ayacucho Quechua 25.3.5 The central area 25.4 Conclusion

739

PART V. Language contact, sociolinguistics, and linguistic anthropology

755

26. The Quechuan–Aymaran relationship Nicholas Q. Emlen 26.1 Introduction 26.2 Quechuan and Aymaran: a shared history 26.2.1 Lexical and structural resemblances between the Quechuan and

757

739 740 742 745 747 747 748 749 750 751 753

757 758

Aymaran languages

760 764 765 765 769

Quechumaran hypothesis

769 771 771 773 774 775 775 779 780 781

26.3 Early speculation about the Quechuan–Aymaran relationship (1653–1950) 26.4 The early comparativist period and the turn to convergence (1950–1986) 26.4.1 Aspirated and ejective plosives 26.4.2 Positional analysis 26.4.3 Loans or cognates? Some concluding comments about the first 26.5 Periodization, directionality, and reformulation (1986–2010) 26.5.1 Relative chronology 26.5.2 Sorting and direction of borrowing 26.5.3 Searching for cognates in the non-shared lexicons 26.6 Reconstructing convergence (2010–present) 26.6.1 Innovations in the Quechuan inflectional complex 26.6.2 The polymorphemic genesis of some Proto-Quechua roots 26.7 The sociocultural context of the initial convergence 26.8 Conclusion and further directions 27. Contact between Indigenous languages of the Central Andes and Spanish: Linguistic outcomes as cases of contra–hierarchical diffusion Anna María Escobar 27.1 Sociolinguistic scenario 27.2 Social prerequisites for contact-induced linguistic phenomena 27.3 Types of contact-induced linguistic outcomes 27.3.1 Phonetic–phonological features

783 783 785 786 788

xxxix

detailed contents 27.3.2 Morphosyntactic features 27.3.3 Morphosemantic features 27.3.4 Syntactic–pragmatic features 27.4 The emergence of Andean Spanish as a case of contra-hierarchical diffusion 27.5 Final reflections 28. Language ideologies and the Quechuan family Rosaleen Howard 28.1 Introduction 28.2 Language ideologies in the colonial period 28.3 Quechuan in its contemporary sociolinguistic context 28.4 Language ideologies in the discursive expression of “multilingual subjectivities” 28.5 Ideologies in language legislation, policy, and planning 28.5.1 Governmental frameworks 28.5.2 The Academia Mayor de la Lengua Quechua 28.5.3 Language ideologies in relation to Quechuan literacy 28.5.4 Language ideologies in relation to corpus planning 28.5.4.1 Language ideologies and orthography 28.5.4.2 Language ideologies and lexical standardization 28.6 Concluding remarks 29. The Andean–Amazonian interface: Sociolinguistic relations and areal -typological patterns Nicholas Q. Emlen, Rik van Gijn, and Sietze Norder 29.1 Introduction 29.2 The areal typology of South American languages 29.2.1 Andean areal studies 29.2.2 Amazonian areal studies 29.2.3 Continental studies 29.2.4 Pan-western South American studies 29.2.4.1 Geographical distribution of structural features 29.2.4.2 Elevational distribution of structural features 29.3 Languages and people between the Andes and Amazonia 29.3.1 Some broad-scale patterns 29.4 Two Andean–Amazonian multilingual networks 29.4.1 Matsigenka, Quechua, and Spanish in the Urubamba valley, Peru

793 796 798 798 800 801 802 808 809 811 813 813 814 815 816 818 818 819 822 823 824 825 827 827 829 831 834

(20th and 21st centuries)

834

(15th–16th centuries)

836 838

29.4.2 A multilingual network between Lake Titicaca and the Amazon plain 29.5 Conclusion

30. Language and the Andean environment Joshua Shapero 30.1 Introduction 30.2 Spatial language in the Central Andes 30.2.1 Toponyms and other lexical items 30.2.2 Affixes marking location and direction 30.2.3 Demonstrative reference 30.2.4 Frame of Reference

xl

789 790 791

839 839 841 841 844 847 850

detailed contents 30.3 Environmental engagement in the Central Andes 30.4 Conclusion References Index

855 857 860 923

xli

Series preface We know that the close study of individual language families and linguistic areas is vital both to the synchronic and diachronic study of language and to cognitive science more widely. Comparative investigations of this type stimulate exciting synergies between different subdisciplines of linguistics, such as language change, contact linguistics, sociolinguistics, linguistic typology, textual philology, and microvariation in grammar, sound, and meaning within and across languages. Besides reflecting and encouraging the links between these subdomains, the fundamental goal of the series is the publish high-quality, substantial reference works which represent a set of theoretically informed and systematic guides to what is known about the world’s languages. Each Guide focuses on a particular language family, subfamily, or areal grouping, and is edited by leading authorities, who bring together contributions from the best international scholars in the field. The Guides aim to show the more general theoretical significance of the languages’ history, linguistic and sociolinguistic characteristics, and overall to provide an indispensable reference tool both to specialist scholars and students and to professional linguists. The approach adopted in all the Guides is systematic and comparative, informed by the latest research and theoretical and methodological perspectives, and, where appropriate, the authors draw on relevant work in such fields as anthropology, archaeology, and cognitive science. Adam Ledgeway and Martin Maiden University of Cambridge and University of Oxford

xlii

List of abbreviations – + (…) 1 2 3 A abl abs acc ACH acmp act ad add adess adj adjlz adv advlz af ag agt agr all AMV anim ant antip appl approx art as.one.goes assr atf atop att atten attn aug aux bc ben

feature partially absent feature partially present feature is marginally present 1st person 2nd person 3rd person agent-like argument of canonical transitive verb ablative absolutive accusative Azángaro, Chocos, and Huangáscar accompaniment actual adverbial direction additive adessive adjective adjectivalizer adverb(ial) adverbializer affected agentive nominalizer agentive case agreement allative Apurí, Madeán, and Viñac animate anterior antipassive applicative approximative article as one goes assertive accelerated transformative verbalizer atop attributive attention attenuative augmentative auxiliary binary choice benefactive

benv bfr bn BPG caus cc cert cess cf CH cisl clf cm cnj coll com comp comp compl cond conj cont cop cop.vblz corr cov csl culm custom cvb dat decl def dem deon desr det df dim dir disc distr dl

verbal benefactive buffer beneficiary Best Possible Grounds causative concentrative certainty, precision cessative counterfactual cfy confirmatory Cacra and Hongos cislocative, translocative classifier contour motion conjectural collective comitative comparative complementizer completive conditional conjunctive only continuative copula copulative verbalizer corroborative substance-covering causal culminative customary converb dative declarative definite demonstrative deontic desiderative determiner diffuse motion diminutive directional discontinuative distributive delimitative

xliii

list of abbreviations dsr dst dtm du dub dur dw emph ENSO ep erg evc evd evi evp evr ev.mod.1 ev.mod.2 ev.mod.3 ev.mod.4 exc excep excl F fact foc FoR freq fut gen ger HQ ik imm imp impr inal incep inch incl incr ind indf iness inf inform injunc

xliv

doing something rapidly on the way to something else distancer detrimental dual dubitative durative downward emphatic El Niño–Southern Oscillation epenthetic ergative evidential—conjectural evidential—direct evidential—inferential evidential—potential evidential—reportative evidential modifier—minimal evidential modifier—increased evidential modifier—maximal evidential modifier—shared knowledge exclamative exceptional performance exclusive feminine factive focus Frame of Reference frequentive future genitive gerund Huaylas Quechua, Huailas Quechua evidential modifier (strong) imminent imperative impersonal inalienable inceptive inchoative inclusive increased impact indicative indefinite inessive infinitive informal injunctive

ins int inter interr intj intr inv ipev irr irrev is iw hab hm hort hypo jtact lim loc LPC LT M ma masl mf>m mid mir mlt/rev mom myth mult.all n Nn.a. narr.pst NBC NCAPA ncompl n.d. neg/q next.to nom nonexhst nmlz npers nposs obj obl

instrumental intensifier interactive interruption interjection intransitive inverse imperfective irrealis irreversible indirect speech inward habitual hierarchy marker hortative hypothetical joint action limitative locative Linear Predictive Coding Lincha and Tana masculine motion across meters above sea level man or woman addressing man middle mirative multiplier/reverser momentaneous mythical multiple possessive neuter non- (e.g. NSG nonsingular, NPST nonpast) not applicable narrative past Naïve Bayes Classifier North-Central Andean Phonological Area non-completive no data negative/interrogative next to nominative non-exhaustive nominalizer/nominalization non-personal non-possessive object oblique

list of abbreviations oblg opt osv ovs ow p PA pass part passacc pdr perf perl pfv pl pla plv pol poss pot PP PQ pred prf prior priv prf prog proh prox prs pst ptcl ptcp punc purp Q quot rc rec recp red refl rel reloc repet res rev

obligative optative object–subject–verb object–verb–subject outward patient-like argument of canonical transitive verb Proto-Aymara passive partnership passive, accidental perdurative perfectivizer perlative perfective plural placer verbal plural politeness possessive potential Proto-Puquina Proto-Quechua predicate, predicative perifactive or contour motion priority privative perfect progressive prohibitive proximal/proximate present past particle participle punctual purposive question particle/marker quotative relative clause recent reciprocal reduplication reflexive relative relocative repetitive resultative reversive

rhet rpst rsn rstr sal s sbj SAP sat sbjv SCOPA se SG sim simul simult SPE SPL stat stf stn su sub subadv subds subis sudd super sv svo SYQ TAM tf top tr transl trib unint uninterr urgt uw vblz vo voc vs WSAPA

rhetorical recent past reason restrictive salience-marking enclitic single argument of canonical intransitive verb subject speech act participant saturator verbalizer subjunctive Southern Cone Phonological Area stem extender sep separative seq sequential singular simple tense simulative simultaneous subject pronoun expression San Pedro and Liscay stative self-transformative stationary surprise subordinator subordinator—adverbial subordinator different subjects subordinator identical subjects sudden action superlative subject–verb subject–verb–object, Southern Yauyos Quechuan tense–aspect–mood transformative topic transitive translocative tribulative unintentional persistence, uninterrupted action urgency, personal interest upward verbalizer verb–object vocative verb–subject Western South American Phonological Area

xlv

The contributors Jorge Acurio-Palma is a Quechua teacher with 15 years experience, and currently coordinates the Quechua major in the professional career of “Traducción e Interpretación Profesional” at the Universidad Peruana de Ciencias Aplicadas. He studied for a master’s degree at the Pontificia Universidad Católica del Perú, where he investigated the phonetic–phonological differences in the subdialects of Cuzco Quechua. His current research interests include the dialectology of Quechua (especially within Cuzco Quechua), the teaching of Quechua as a second language, and the correlation between linguistic space and cultural space (with an emphasis on that configured by folk dances). Willem F. H. Adelaar is Emeritus Professor of Amerindian Languages and Cultures at Leiden University in the Netherlands. He has conducted field research on different varieties of Quechua and on minor languages of the Andes. He has also worked on the genetic relations of South American languages of the Andes and the Amazonian region and has been involved in international activities addressing the issue of language endangerment. His further areas of expertise include linguistic reconstruction, contact and areal linguistics, oral literature and ethnohistory of South American and Mesoamerican peoples, as well as the interface of linguistic studies with archaeological and historical research. Astrid Alexander-Bakkerus is Master of Arts (French) and PhD in Comparative Linguistics, Amerindian Languages. She has published grammars of 18th-century Amerindian languages (Cholón, Ecuadorian Quichua, Jebero/Xebero), papers regarding the phonology, morphology, nominalization, subordination, and vocabulary of these languages, and papers in the field of colonial and missionary linguistics. Luis Andrade Ciudad is Full Professor in Linguistics at the Pontificia Universidad Católica del Perú (PUCP). His research interests include Andean linguistics, language contact, and social and ethnographic approaches to language history. He has published The Spanish of the Northern Peruvian Andes: A Sociohistorical and Dialectological Account (Peter Lang), and has recently co-edited (with Sandro Sessarego) Los castellanos del Perú: historia, variación y contacto lingüístico (Routledge). Raúl Bendezú-Araujo is scientific coordinator at the CRC1287 “Limits of Variability in Language,” hosted by the

xlvi

Universita¨t Potsdam and funded by the German Research Foundation (DFG). During his master’s studies at the Pontificia Universidad Católica del Perú, he carried out research on colonial Quechua and language and culture. Having completed his doctoral studies at the Freie Universita¨t Berlin, his current research interests include the semantics and pragmatics of Quechuan languages, language documentation, and language contact. Rodolfo Cerrón-Palomino received his PhD in linguistics from the Universidad Nacional Mayor de San Marcos in Lima and the University of Illinois Urbana-Champaign. An emeritus professor of the UNMSM, he currently holds the chair at the Pontificia Universidad Católica del Perú. He has been visiting professor and speaker at conferences at various universities of his home country and abroad, and has published around twenty books and numerous articles in his area of specialization in renowned national and international journals. His most recent books include Las lenguas de los Incas: el puquina, el aimara y el quechua (Peter Lang), Tras las huellas de Inca Garcilaso (Latinoamericana Editores), and El uro de la Bahía de Puno (Instituto Riva-Agüero). Matt Coler is an Associate Professor in the Research Group “Language, Technology and Culture” at the University of Groningen, where he is the Director of the Voice Technology Master’s program. In addition to Aymaran languages and Andean Spanish, his research interests include voice synthesis and speech recognition for under-resourced languages as well as auditory perception. Nicholas Q. Emlen is Assistant Professor of Language and Culture at the University of Groningen (Campus Fryslaˆn) and postdoctoral researcher at the Leiden University Centre for Linguistics. He has conducted extensive ethnographic field research about multilingualism on the Andean–Amazonian agricultural frontier of Southern Peru. He also works on the reconstruction of Quechua–Aymara language contact in the ancient Central Andes and on multilingualism among Quechua, Aymara, Puquina, and Spanish in the colonial Andes. His first book, Language, Coffee, and Migration on an Andean–Amazonian Frontier (University of Arizona Press), was published in 2020. Anna María Escobar is Full Professor at the University of Illinois at Urbana-Champaign (USA). Her research and teaching center on sociolinguistics and language change

contribu to rs in language contact scenarios, with a focus on grammatical categories. Her most recent book project is on the Indigenous origins of Andean Spanish, a sociohistorical study analyzing colonial and diachronic modern data. She is co-editor with Salikoko Mufwene of the Cambridge Handbook of Language Contact (Cambridge University Press, under contract). Karolina Grzech holds a PhD in Linguistics from SOAS, University of London. Since 2013, she has been documenting Quechuan languages in Ecuador, collaborating with speakers of Upper Napo Kichwa (since 2013) and Chibuleo Kichwa (since 2020). She held postdoctoral positions at the Universities of Stockholm (2018-20) and Valencia (2020– 2024, at the time of writing her contribution to this volume). Since 2024, she is a tenure-track professor at the Department of Translation and Language Sciences, Universitat Pompeu Fabra. Karolina’s research interests include evidentiality, epistemicity, pragmatics, and the methodology of linguistic fieldwork. In 2021 she received the DELAMAN award for the documentation of Upper Napo Kichwa deposited in the ELAR archive. Katja Hannß obtained her Master’s degree in Americanist Studies and Linguistics from Bonn University, before carrying out a PhD project at the Radboud University Nijmegen (Netherlands). Her PhD thesis is a grammatical description of the extinct Uru language (Bolivia), and was successfully defended in 2008. Further stations of the author’s career include Stirling (UK), Constance, Regensburg, and Cologne. Since April 2020 she has been interim professor at the Ludwig-Maximilians University at Munich. Central topics of her research are discourse studies, language contact, and language documentation. Paul Heggarty holds a cátedra de excelencia at the Pontificia Universidad Católica del Perú in Lima. His focus is on language (pre)history, aiming to ensure that the perspective from linguistics is better understood outside that field and to contribute towards a more coherent, cross-disciplinary vision of the human past. To that end he works closely with archaeologists, geneticists, and historians. Within interests that range worldwide, he specializes in the languages of the Andes, particularly the divergence history of the Quechuan and Aymaran families. Since 2008 he has convened a series of nine interdisciplinary conferences and symposia on the Andean past. Rosaleen Howard is Professor Emerita in the School of Modern Languages at Newcastle University, UK, where she was Chair of Hispanic Studies until her retirement in 2020. She works on the sociolinguistics and anthropological linguistics of the Andes and on Quechuan. Her books include Por los linderos de la lengua. Ideologías lingüísticas en los Andes (Instituto de Estudios Peruanos/Instituto Frances de Estudios Andinos/Pontificia Universidad Catolica del Peru) and

Kawsay Vida. A Multimedia Quechua Course for Beginners and Beyond (University of Texas Press). She is also the author of numerous articles in specialist journals. César Itier teaches Quechua at the Institut National des Langues et Civilisations Orientales (Paris). His research deals with Quechua oral and written literature, the social history of the language family, and its sociopolitical vocabulary. His books include El teatro quechua en el Cuzco (Centro de Estudios Regionales Andinos “Bartolomé de las Casas”), El hijo del oso: La literatura oral quechua de la región del Cuzco (Instituto Francés de Estudios Andinos/Instituto de Estudios Peruanos), Diccionario quechua sureño–castellano (Editorial Commentarios), Runasimita yachasun. Método de quechua (Instituto Francés de Estudios Andinos/Editorial Commentarios), the edition and translation of Juan de Espinosa Medrano’s El Robo de Proserpina y sueño de Endimión (Instituto Francés de Estudios Andinos/Instituto Riva-Agüéro), and Palabras clave de la sociedad y la cultura incas (Instituto Francés de Estudios Andinos/Editorial Comentarios). Peter Kaulicke (PhD 1980, Bonn University) has been Professor of Archaeology at the Pontificia Universidad Católica del Perú in Lima since 1982. His research foci include Archaic and Formative chronology, funerary contexts and analysis, art and religion, the origins of social complexity, the ethnohistory–archaeology relationship, and the history of archaeological research in Peru. He received several awards and was guest professor at many universities and research centers around the world. An Ordinary Member of the German Archaeological Institute and the Institute of Andean Studies, he is the author or editor of some 20 books and about 200 articles. Olga Krasnoukhova is a Marie Skłodowska-Curie Individual Fellow at Leiden University. She held previous appointments at the University of Antwerp and Radboud University Nijmegen. She is a typologist, with a focus on South American Indigenous languages. Her research interests include areal linguistics and language change, with an emphasis on the synchrony and diachrony of negation, as well as the nominal domain more broadly. Her current project aims to account for the dominant pattern of post-verbal negation in South American languages by exploring it from the perspective of language contact. Lev Michael is an anthropological linguist who has worked with speakers of several Amazonian languages, including Nanti and Matsigenka (both Arawakan), Iquito (Zaparoan), and Mã́ĩ́hĩk̵̀ ì (Tukanoan). Methodologically grounded in typologically informed language documentation and description, his work focuses on aspects of language history that shed light on the deep social history of South America, and on social and cultural aspects of language structure and use. He obtained his PhD from the Department of Anthropology at the University of Texas at Austin in 2008, and

xlvii

con tr ib utors is currently Professor in the Department of Linguistics at the University of California, Berkeley. Carlos Molina-Vital is the director of the Quechua Program in the Center for Latin American and Caribbean Studies at the University of Illinois, Urbana-Champaign. He is completing his doctoral dissertation on the middle voice in the grammatical voice system of Ancash Quechua (Huaylas) in the program of Andean Studies at the Pontificia Universidad Católica del Perú. His interests include linguistic typology, usage-based linguistics, grammatical description and linguistics applied to the teaching of Quechuan languages. Arjan Mossel is PhD candidate in linguistics at Leiden University, with a research project focused on place-names and the linguistic past of the Central Andes, and works as software developer in digital humanities at Utrecht University. His interests include Andean languages and linguistics, as well as the application of GIS, databases, and information technology more broadly for research in the humanities. † Pieter Muysken (1950–2021) was Professor of Linguistics at Radboud University, Nijmegen. His main research interests were Andean languages, creole languages, and language contact, and his work had a strong commitment to language contact and language history in South America. His books include The Languages of the Andes (Cambridge University Press, with Willem F. H. Adelaar), and he acted as co-editor for The Native Languages of South America: Origins, Development, Typology (Cambridge University Press, with Loretta O’Connor) and Language Dispersal, Diversification, and Contact: a Global Perspective (Oxford University Press, with Mily Crevels). Johanna Nichols is Professor Emeritus in the Department of Slavic Languages, University of California, Berkeley. She is currently Research Supervisor in the Higher School of Economics, Moscow, and is an AAAS and LSA Fellow. She works on typology, historical linguistics, linguistic geography, linguistic complexity, and their contemporary and prehistoric distributions. Sietze Norder is a biogeographer who is studying how the human past has been shaped by topography, climate, and ecology. He currently holds a postdoctoral position at the Leiden University Centre for Linguistics, where he adopts theories and methods from biogeography and macroecology to understand spatial patterns in the distribution of languages. He is also the developer of the “glottospace R” package for the geospatial analysis and visualization of linguistic data. Matthias Pache holds a PhD in linguistics from Leiden University (Netherlands) and has held research positions at the universities of Leiden, Bonn, Tübingen (Germany), Leuven Belgium, and Passau (Germany). His research

xlviii

specializations are South and Central American Indigenous languages, with a major focus on the Chibchan family and languages of the Andes—Aymaran, Mapudungun, and Quechuan in particular. He has published papers on topics in descriptive linguistics, contact linguistics, the historiography of American Indigenous linguistics, and typology. He particularly focuses on historical linguistics, including the investigation of distant genealogical relationships and the application of the comparative method in the context of South and Central American Indigenous languages. Allegra Robertson is a PhD student in linguistics at the University of California, Berkeley. Her primary research interests are in Indigenous language revitalization, phonetics, phonology, variation, and Amazonian languages. She has collaborated with speakers of Warao (isolate) on language documentation and revitalization projects, and is currently focused on morphophonological phenomena and dialectal variation in Yanesha’ (Arawakan). Joshua Shapero is assistant professor of anthropology at the University of New Mexico. His research examines the intersection of language, environment, and sociality among Ancash Quechua speakers in the central Peruvian Andes. He has published on topics such as ritual and the social ontology of Andean places, spatial language and cognition, and the use of gestures in Quechua. Currently, he is completing a monograph that examines the environmental grounds of communication, cognition, and social life among Ancash Quechua speakers in Peru’s Cordillera Blanca mountain range. Aviva Shimelman currently works as an ambulance paramedic and yoga therapist on Turtle Island. She has been affiliated with the research group “The Language Dynamics of the Ancient Central Andes,” hosted by the University of Tübingen, and the Department of Linguistic and Social Evolution at the Max Planck Institute for the Science of Human History. She held two “Documenting Endangered Languages” fellowships from the National Science Foundation (USA), and has taught linguistics at universities in the US and elsewhere. Her research focuses on language documentation and revival, principally in the Americas and Melanesia. She has compiled extensive corpuses with communities from these regions. Matthias Urban is researcher at the CNRS DDL lab, where he directs an ERC project on historical dynamics in language geography. He has held prior appointments at the Max Planck Institute for Evolutionary Anthropology in Leipzig and the universities of Leiden, Marburg, and Tübingen, where he was principal investigator of the Junior Research Group “The Language Dynamics of the Ancient Central Andes,” funded by the German Research

contribut ors Foundation (DFG)’s Emmy Noether Programme. His research interests include historical linguistics, in particular of the Andes, language contact, and linguistic typology. Jairo Valqui is principal investigator of the Research Group “Documentación lingüística de lenguas amenazadas en el Perú,” hosted by the Universidad Nacional Mayor de San Marcos. He has carried out extensive fieldwork on Quechuan and Arawakan languages. His research mainly focuses on phonetic and phonological aspects of the Andean and Amazonian languages. Among his current projects are “Vestiges of a Native Language in the Territory of the Chachapoyas Culture” and “Language Documentation of Chachapoyas Quechua.” Simon van de Kerke was, until his retirement, assistant professor at LUCL, the General Linguistics Department of the University of Leiden, The Netherlands. His research

focused on the morphology of a number of Indigenous languages of Bolivia: Quechua, Leko, Chipaya, and Puquina. He is currently working on a comprehensive linguistic description of the Leko language. Rik van Gijn is an assistant professor at Leiden University, and principal investigator of the ERC Consolidator project “South American Population History Revisited” (SAPPHIRE), focusing on uncovering the population history of the linguistically highly diverse Upper Amazon area in western South America. In his present and previous appointments at Radboud University Nijmegen, the Max Planck Institute for Psycholinguistics, and the University of Zurich, he has focused on a broad range of topics, including South American areal typology (in particular morphological and syntactic typology), interdisciplinary research into the human past, and the description and documentation of the Bolivian isolate language Yurakaré.

xlix

PART I

Background and context

chapter 1

Introduction Central Andean linguistic diversity and the diversity of Central Andean linguistics Matthias Urban

1.1 Extent and characteristics of the Central Andes The Andes are one of the most multifaceted, beautiful, and challenging regions of the world. Stretching from Venezuela in the north to Tierra del Fuego in the south along the entire latitudinal extent of the South American landmass, they are, like other high mountain areas, the result of worlds colliding: in this case, the Nazca plate, continuously moving westward, is gradually forced under the South American plate in what is called the South American subduction zone; as these tectonic giants clashed, the Andes, the second-highest and longest mountain chain in the world, were created. Given the huge variation in altitude and the fact that much of the Andes are located in tropical latitudes, a bewildering array of ecozones and environmental conditions was formed. Like microcosms, these represent much of the ecological diversity of the entire globe. But there is more to the Andes than just environmental diversity and beauty. They are a special place also for the human story. Not only was this challenging environment already encouraging the creation of cultural diversity and unique socioeconomic adaptations in prehistoric times, but the Andes are also one of the few “cradles of civilization”— places on earth where complex forms of social organization and state-level societies emerged in prehistory independently of outside influences (see e.g. Stanish 2001b). This, however, did not happen along the entire length of the Andes to the same extent; a nucleus of these developments was in the mid-latitudes of the Central Andes, which boasts a long-standing continuous cultural history involving complex societies beginning in the 3rd millennium bc. It came to an end when the Inca empire, the largest state ever witnessed in the pre-Hispanic Americas, extending from Southern Colombia to Northern Chile and Argentina, was crushed by the Spanish conquistadores in the 16th century. Many long-standing cultural traditions, however, continue to live

on until the present day, as do many (though far from all) Indigenous languages of the region. This culturally precocious region is known as the Central Andes. It is the linguistics of this region, in all its facets and diversity, to which the present volume aims to be a guide (leaving the Northern Andes of Ecuador and Colombia and the Southern Andes down to the Southern Cone and Tierra del Fuego for different volumes in this series). But this culture-based delimitation of the Central Andes is not unambiguous. The term is usually used in one of two ways, both related to the cultural history of the region: First, the extent of the Inca empire at its height in the early 16th century may be used as a reference point for linguistic explorations (e.g. Cerrón-Palomino 1985; Adelaar with Muysken 2004). However, the Inca presence in the empire's northern and southern peripheries (Ecuador, Colombia, Chile, and Argentina) lasted only few decades and was hence quite ephemeral. Prior to the Inca conquest, these regions had quite different cultural histories. Consequently, an alternative, more restrictive definition of the Central Andes may be adopted (e.g. Stanish 2001b): here, the Central Andes are principally defined as the coast and highlands of present-day Peru and Andean Bolivia, including the greater Titicaca basin and the Altiplano. The complex cultural history of the Central Andes thus defined, featuring several waves of homogenization due to expanding states as well as phases of development at a more regional level, has also had an impact on the languages spoken in the region. Therefore, while based on culture history and geography, the definition is not completely arbitrary from a linguistic point of view. It is this slightly narrower definition that is adopted in this volume. It includes an overview chapter on the physical geography and cultural history of the Central Andes thus defined (Chapter 2 by Peter Kaulicke), which explores the natural and cultural environment into which language use in the Central Andes is embedded in quite some detail.

Matthias Urban, Introduction. In: The Oxford Guide to the Languages of the Central Andes. Edited by: Matthias Urban, Oxford University Press. © Matthias Urban (2024). DOI: 10.1093/oso/9780198849926.003.0001

3

matthias urban

1.2 Past and present linguistic landscapes of the Central Andes Given their cultural and demographic weight, it is not uncommon, especially in publications from countries of the Central Andes themselves, for “Andes” to stand for “Central Andes.” Thus, lingüística andina “Andean linguistics” or castellano andino “Andean Spanish” may, with or without further explanation, meant to be understood as meaning “Central Andean linguistics” and “Central Andean Spanish” (e.g. Cerrón-Palomino 1985; 2003a) in either the broader or narrower definition of the term as sketched above. Such usage reflects perspectives on the Andes from the point of view of Peru or Bolivia and are, from a purely geographical perspective, inadequate as descriptive labels, as languages of the Andes to the north and the south of the Central Andes may not be comparable and are generally not implied. But this is not an issue only in Spanish-language publications from South America: it can occasionally lead to confusion and misunderstanding also in English language publications aimed at an international audience (as when “Andean” languages generally are said to be typologically like the major Central Andean languages in Dixon and Aikhenvald 1999; see Aikhenvald 2007 for a more cautious statement and Urban 2019a for discussion). So what languages are implied when talking about the Central Andes as defined above? In terms of Indigenous languages, at present and in the past, Quechuan and Aymaran are the dominant language families. As the map in Figure 1.1, which displays the present-day linguistic diversity in the Central Andes, shows, Quechuan in particular is very widespread within and beyond the Central Andes; and it is probably precisely because of this wide extension that both the Inca and the colonial Spanish regime adopted Quechuan varieties as lingua francas for the purposes of imperial administration and religious indoctrination (Chapter 25 by César Itier in this volume). To avoid false impressions right at the outset, Quechuan is a language family, not a single homogeneous language. Many native speakers who are only familiar with their local varieties are not aware of this, and common terminological choices made even in academic literature in which Quechua is called a language and individual varieties dialects continue to perpetuate this impression (Luykx et al. 2016). The map in Figure 1.1 does not represent internal variation within the Quechuan language family, partially because gradual rather than abrupt changes, as indeed in a dialect continuum, make it difficult to reify individual configurations as distinct types of speech; however, there are also sharper dividing lines within the family, and at any rate speakers of very different varieties have significant troubles

4

when trying to communicate with one another, showing clearly that internal variation is far from trivial (Luykx et al. 2016). Still today, Quechuan is the quintessential language family of the Central Andes, and Quechuan speech can be heard widely in everyday life: at tourist markets for craftgoods sold in Otovalo in highland Ecuador, cramped with 10 other people and a not much lower number of chicken in a minibus colectivo going from the Andean slopes to the city of Huaraz in Northern Peru, or on the chacras of the comunidades campesinas of Southern Peru and Bolivia as people go to work. In addition to Quechuan, Aymara also played a significant role in the culture history of the Central Andes. This language remains widespread today in Southern Peru, but especially on the Altiplano of Bolivia. Aymara’s sister language, Jaqaru, in contrast, is severely endangered and likely to become lost soon. Both together form the Aymaran language family. Alongside these two widespread families, there is the unrelated Chipaya language, vigorously spoken only in a single stronghold community, Santa Ana de Chipaya, on the Bolivian Altiplano. Chipaya, too, was originally part of a language family, but the more poorly documented sister language, Uru, went dormant recently (see Zurita Aguilar 2018 for revitalization efforts). The fate of Uru is shared with several other more or less local languages of the Central Andes. As the map in Figure 1.2, which shows an approximate reconstruction of the linguistic diversity of the Central Andes in earliest historical times, makes obvious, the linguistic situation was considerably more complex formerly. One particularly important language, of a large extension in a transect from the northern outliers of the Bolivian Andes through the Titicaca basin to the coast of Southern Peru and Chile, was Puquina, which was also unrelated to Quechuan and Aymaran. But loss of linguistic diversity is particularly acute in the northern part of the Central Andes, which boasted a notably different linguistic geography from southern Peru and the Titicaca basin. Unlike the Altiplano, which provides conditions that are conducive to the spread of languages or relatively homogeneous language families (Nichols 2015 and Chapter 22 by Johanna Nichols in this volume), here we find more in situ diversity, with as many as thirteen languages coexisting in a relatively limited geographical domain. What is known on these languages is unfortunately in many cases very little. Of the languages of the North Coast, only one, Mochica, is relatively well documented. In addition, on the eastern slopes, the Hí bito and Cholón languages were spoken; these probably formed a small local language family, a position that is supported by Adelaar with Muysken (2004: 461–2) as well as Astrid Alexander-Bakkerus in Chapter 13 of this volume. For all we know, most of these languages, like Puquina, were isolates.

Aymara

Peru

Chipaya

Chile

Bolivia

Lake Titicaca

Based on Lewis et al. (2016), with the Quechua and Aymara distribution in southern Peru and Bolivia refined according to Landerman (1994: 335–6, maps 1 and 2). Map is for illustration only. Note that Indigenous languages to the east of the Andes, of which there are many, are not shown on this map.

Figure 1.1 Approximate present-day distribution of Indigenous languages of the Central Andes.

Jaqaru

Quechuan

0 100 200 300 400 500 km

Jaqaru

Quechuan

Chile

Bolivia

0 100 200 300 400 500 km

Quechuan & Aymaran & Puquina

Chipaya

Aymara & Puquina

Quechuan & Aymara

Peru

Cholón

Hibito

Chacha

Chirino Patagón Copallín Bagua

Uru

Lake Titicaca

Puquina

Figure 1.2 Approximate reconstruction of the linguistic situation in the Central Andes in the early 16th century. Based on Torero (1986; 1989; 1993); Adelaar with Muysken (2004); Cerrón-Palomino (2010); Andrade Ciudad and Bell (2016); Lewis et al. (2016); Urban (2019b), and Wikipedia user Eredebhel’s map of Uru populations. Map is for illustration only. Note that Indigenous languages to the east of the Andes, of which there are many, are not shown on this map.

Quingnam

Culli

Mochica

Sechura

Tallán

Sácata

introduction It is all these languages, whether well described or only partially documented, and whether still spoken or already lost or dormant, that the present volume seeks to provide a guide to. One overarching pattern that is common to these original language geographies in north and south of the Central Andes, different as they are, is the complex entanglement of languages in geographical space. In spite of the fact that this has commonly been attempted in the literature, this entanglement is often so intricate that it is difficult to delimit sharp language boundaries and language areas, and if maps showing such boundaries seek to represent actual linguistic reality rather than being useful devices for getting a very rough impression on language distributions, they may even be inadequate. Part of the explanation for this is surely that the original picture must be reconstructed from incomplete and sometimes conflicting historical sources. But this picture likely also reflects a genuine pattern of discontinuous language distribution that dovetails with patterns of discontinuous territoriality, one of the specifically Andean patterns of socioeconomic organization (Mannheim 1991). I sketch related issues in Chapter 24 of this volume.

1.3 Language contact, areal typology, and the fabric of Central Andean languages Apart from the cultural trajectory, are there also more strictly speaking linguistic grounds for defining the Central Andes like in the present volume? Is there anything that holds, or held, the languages mentioned in Section 1.2 together as a group? One answer that could be given, and that in fact frequently is given for a linguistically-based definition of the Central Andes in either the narrow or the wider sense described in Section 1.1, makes reference to the dominance of Quechuan and Aymaran. Indeed, the distribution of Quechuan is strikingly similar to the maximal extent of the Inca empire, implying a convenient congruence between the linguistic and archeological records that might serve well for delimitative purposes. Howevers, this congruence is misleading and fallacious: while there is good evidence that Inca agency propelled Quechuan into its current peripheries in Colombia, Bolivia, and Argentina, historical linguists are quite confident that the family was already widespread when the Inca rose to power (Heggarty 2007; Urban submitted). Furthermore, as Section 1.2 has shown, the perceived dominance of the Quechuan and Aymaran language families is something that was accentuated through the loss of much of the former linguistic diversity. Once the much more fractal picture that could be observed in language geography and diversity is fully

appreciated, the danger of an anachronistic distortion of the linguistics of the Central Andes when relying on the extent of single linguistic lineages among many becomes obvious, however important they may have been or be at present. Yet, there is another manner in which Quechuan and Aymaran evidence can provide elements for a linguistically based delimitation of the Andes. To appreciate this, we have to once again travel to the past to the processes and interactions that shaped the languages of the Central Andes in decisive ways and that are of continuing relevance: while contact relationships between South American languages are “vast and almost intractable” (Muysken and O’Connor 2014: 10) generally, the Quechuan and Aymaran case furnishes “one of the most intriguing and intense cases of language contact to be found in the entire world” (Adelaar 2012b: 576). As Nicholas Q. Emlen shows in detail in Chapter 26 of this volume on the Quechuan–Aymaran relationship, there are striking similarities in the lexicon and the morphosyntactic structures between the languages of the two lineages that cry out for an explanation. Estimates of shared vocabulary reach up to 30% and include many items of basic vocabulary, and similarities in morphosyntactic organization reach down into minute details (Cerrón-Palomino 1994). Over the centuries in which scholars have occupied themselves with the origins of these similarities, the pendulum has swung back and forth between a preference for an explication in terms of genealogical relatedness and an alternative account in terms of contact-induced convergence. At the present stage of research, for reasons explained by Emlen, the consensus view is that a contact-based explanation is more adequate to account for the specifics of the similarities (however, see Campbell 1995 for a relatively recent opposite view, and note that Emlen discusses some new evidence that might in the future revive the genealogical hypothesis). Given that a significant subset of the mentioned similarities reconstructs to both protolanguages, one is forced to assume that already speakers of these protolanguages, when they were still spoken only in their presumed homeland in central Peru at a time deep in the prehistory of the Central Andes, began interacting intensively with one another, leading to the transfer of lexical and grammatical matter and patterns. In Chapter 24, I sketch a possible sociolinguistic scenario for this period of “initial convergence” (Adelaar 2012c) that is informed by and attuned to specific and apparently long-standing patterns of social differentiation and socioeconomic interaction in the Central Andes; in this scenario, linguistic difference is semiotically active in signaling socioeconomically distinct identities within society. In any case, the initial convergence led to a complex and multidimensional entwinement of the lineages (see Adelaar 2020a

7

matthias urban for an overview) that gave both families their characteristic imprint. As a result, one can speak of a common Quechuan and Aymaran language type (Adelaar 2017a). Then, when both protolanguages began to spread from their homelands to eventually become the most widespread language families of the Central Andes (on these language spreads, see Chapter 23 by Paul Heggarty in this volume), a process of diversification set in that eventually lead to the internal differentiation and dialectal configurations that are visible today. In this process, daughter languages and dialects inherited the common structures that resulted from the initial convergence: “there is no modern variety of Quechuan that does not bear the structural marks of intensive contact with Aymaran, and the opposite probably holds true for Aymaran with respect to Quechuan lexical influence” (Adelaar 2020a: 9). Thus, the spread of the families also spread the Quechuan-Aymaran language type resulting from the initial convergence through the Central Andes. Where daughter languages ended up spoken in adjacent regions, this led to secondary transfer of lexical items and grammatical structures (Emlen and Adelaar 2017), yielding the complex and multilayered relationship that deeply and virtually inextricably intertwines the two families. Importantly, while the onset of the special relationship between Quechuan and Aymaran is a matter of historical linguistics (more on such research in Section 1.5), it is of continuing relevance, as Quechuan and Aymaran contact continues in southern Peru and Bolivia (Narayanan 2018; see also Section 1.7), and further chapters in the history of Quechuan-Aymaran interaction are being written in the present. The Quechuan-Aymaran case presents a striking case of ultimately bilateral language contact, and understanding it in its full complexity is a continuous challenge in Central Andean linguistics. But there is more. Beyond bilateral cases of language contact, when transfer of matter and pattern from one language or language family to another can be studied in a focused manner, there are often larger scale patterns of typological and lexical affinity that involve more languages and the emergence of which is not traceable in the same way as contact effects in bilateral scenarios. At the greater time depths at which such more generalized patterns of affinity arose, areal typology is usually the method of study as far as grammatical structure is concerned (see Muysken 2008). Although there is considerable typological diversity among the languages of the Central Andes, we find typological similarities and evidence for linguistic convergence between subsets of these languages, concomitant areal substructure (Urban 2019a), and also some overarching commonalities and family resemblances in typological features (e.g. Adelaar 2008, 2012b, 2017a; Torero 2002; Urban 2019a). In Chapter 28 of this

8

volume, Nicholas Q. Emlen, Rik Van Gijn, and Sietze Norder review areal-typological studies for the Central Andes that corroborate such an observation, and also the typologicalcomparative chapters in this volume amply attest to it. For instance, as discussed by Olga Krasnoukhova in Chapter 18, languages of the Central Andes typically have rich systems of peripheral cases that allow to express highly specific ways in which dependent nouns stand in relationship to the verbal head of the clause. Another example, coming from verbal morphology that is discussed in Chapter 19 by Matthias Pache, is that many languages of the Central Andes have verbal person reference markers that are fused with tense, in particular future tense. Such markers are found in Quechuan (Chapter 4 by Carlos Molina-Vital, Chapter 5 by Aviva Shimelman, Chapter 6 by Aviva Shimelman and Jairo Valqui, and Chapter 7 by Raúl Bendezú-Araujo and Jorge Enrique Acurio-Palma in this volume) and Aymaran (Chapter 8 by Matt Coler), but beyond these, they are also characteristic of Puquina (Chapter 12 by Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar), Cholón (Chapter 13 by Astrid Alexander-Bakkerus), and UruChipaya (Chapter 10 by Katja Hannß). On the other hand— and this is something that may have been underestimated or understated until quite recently—Quechuan and Aymaran languages may share preferences that, within the broader Central Andean context, do not make them the most typical and characteristic representatives, but more the odd one out. For instance, languages of both families share strong disyllabic root canons, whereas other languages of the Central Andes, in particular in Northern Peru, show less clear consistent canons, but, if anything, have a predilection for monosyllabic roots that, in addition, allow for heavier syllables than can usually be observed in Quechuan and Aymaran; this gives Northern Peruvian languages a lexical profile that is notably different from Quechuan-Aymaran canons (Urban 2019a; 2019b; see also my Chapter 14 in this volume). The evidence for language contact beyond Quechuan and Aymaran is not restricted to patterns in grammatical organization, but sometimes extends to the phonological matter of relevant markers. In fact, in some cases there is not only striking similarity in the functions fulfilled by the morphological markers of languages of the Central Andes, but there are also resemblances in their form. Some of these resemblances are clear instances of grammatical borrowing, e.g. in the case of the case suffix -layku in Chipaya, which is clearly of Quechuan/Aymaran origin. But beyond these obvious cases, there are others where resemblances are vaguer and fuzzier. For instance, the Puquina agentive nominalizer -eno uncannily resembles the Cholón future nominalizer -(ŋ)o. Cholón in turn, features, alongside clearly borrowed material from Quechua, also similarities that are vaguer

introduction and that, if not due to chance, likely reflect older affinities, for instance the indefinite marker -pit that resembles the Quechua counterpart -pis (Alexander-Bakkerus 2005). Such resemblances are very difficult to make sense of. The same is true of wanderwo¨rter: lexical items that are common in the languages of the Central Andes across genealogical boundaries (see also Torero 2002: 29). A salient selection is in Table 1.1. While some of these similarities can be explained through bilateral language contact (for instance, the Uru form is most likely a borrowing from Aymaran), this is not possible at large. For instance, there is no systematic explanation available to account for the s : t correspondence in the initial consonant in the Quechuan and Chipaya forms for ‘maize,’ which are sara and tara respectively, and similar statements must be made for forms in the other languages. Such unsystematic correspondences suggest deep layers of lexical affinities, and since they are too scattered to suggest common descent, they may indicate linguistic interaction in prehistory in this culturally precocious region. The semantics of the lexical items in Table 1.1 is also revealing: maize is a crucial and culturally highly significant staple crop, bodies of water such as lakes are important in a

challenging environment where freshwater is a very scarce and precious commodity, and rainfall is absolutely vital for agricultural success. Accordingly, there are widespread and deep-rooted cosmological beliefs and cultural practices that attest to the importance of maize and water in the Central Andes (e.g. Weismantel 1991; Hastorf and Johannessen 1993 on maize-based chicha beer; Urton 1982b on water) that dovetail with the widespreadness of forms in Table 1.1. As is the case for the prehistoric cultural trajectory (Isbell and Silverman 2008), typological similarities and lexical affinities in culturally relevant vocabulary thus create a linguistic fabric that overarches considerable internal diversity and difference, and that leads to the impression of a more or less coherent “core” of Andean linguistic structure that is often aligned with Quechuan and Aymaran as the most important families (e.g. van Gijn and Muysken 2020; Heggarty 2020a), or even implicitly defined with reference to them (Aikhenvald 2007), but that in reality is more inclusive and diffuse, with considerable typological variation and substructure (Urban 2019a). The typological-comparative sketches in this volume are a testimony to this diversity. In particular, a north–south structure in language geography

Table 1.1 Three Wanderwo¨rter in the languages of the Central Andes Language (Source)

%kot(a) ‘(body of) water’

%sara ‘maize’

proto-Quechua (Emlen 2017a)

*qutʂa ‘lake’

*sara

%para ‘rain’

Ayacucho Quechua (Parker 1969)

para

Cuzco Quechua (Cusihuamán 1976a)

para ‘rain, shower’

proto-Aymara (Emlen 2017)

*qutʂa ‘lake, pool’

Chipaya (Cerrón-Palomino and Aguirre 2011) Uru (Vellard in Muysken 2000)

tara ‘lake’

Puquina (Oré 1607, Emlen et al., Ch. 12, this volume)

para ‘river’

Sechura (Martínez Compañón 1985[1782–1790], Urban, Ch. 14, this volume)

Culli (Martínez Compañón 1985[1782–1790], Urban, Ch. 14, this volume)

‘sea’

Cholón (Alexander-Bakkerus 2005)

kot ‘water’

Híbito (Martínez Compañón 1985[1782–1790], Urban, Ch. 14, this volume)

‘water’

Copallín (Anonymous 1965[~1570], Torero 1993)

‘water’

9

matthias urban and typology is mirrored in the archeological record, with nuclei of cultural development in north and south that have distinct cultural expressions and great consistency through time (see Chapter 2 by Peter Kaulicke in this volume) and, as is just becoming clear recently as genomic samples from northern Peru are becoming available, also in modern population structure (Urban and Barbieri 2020). However, the Central Andes cannot be described as a self-contained linguistic interaction sphere, as typological affinities extend beyond both northward and southward, for instance to the Barbacoan languages in Ecuador (Adelaar with Muysken 2004: 145), the languages of the Chaco (Michael et al. 2014), and the Amazonian fringe (Wise 2011; Valenzuela 2015; see also Chapter 28 by Nicholas Q. Emlen, Rik van Gijn, and Sietze Norder in this volume); generally, within the Andes, a picture of smooth rather than abrupt typological change as one moves through geographical space can be observed (Urban et al. 2019). Also lexical similarities such as those listed in Table 1.1 sometimes extend to the languages of the Southern Andes (cf. e.g. Kunza puri ‘water’ Vaïsse et al. 1896, Mapudungun pire ‘hail, snow’ Augusta 1916; see also Torero 2002: 29 and cf. Pache 2014: 350 fn. 2) and the adjacent Amazonian lowlands (cf. e.g. Munichi tsá’a ‘maize,’ Proyecto de Documentación del Idioma Muniche 2009). Adequately capturing how the languages of the Central Andes as a whole have a distinct areal-typological profile even beyond the Quechuan-Aymaran prototype against the background of their unique cultural trajectory, how that relates to interaction patterns, and how the region is nevertheless embedded organically into the broader Andean and South American context, is a continuing challenge. The comparative chapters in this volume, as well as Chapter 26 by Nicholas Q. Emlen on the Quechuan–Aymaran relationship, make significant contributions to this. Authors build on both their previous expertise in the areal-typology of South America, which often allows them to draw on established large-scale comparative databases. Together with the descriptive sketches contained in the present volume, most of which have been made available to authors in manuscript form when writing their chapters, such largescale comparative databases create a two-pronged basis for typological-comparative work that allow both the largescale comparative picture as well as micro-variation in the Central Andes themselves to become visible.1 Lev Michael and Allegra Robertson are concerned with phonological structures in Chapter 17, Olga Krasnoukhova with nominal 1 Chapter 4 on Ancash Quechua, by Carlos Molina-Vital, and Chapter 7 on Cuzco Quechua, by Raúl Bendezú-Araujo and Jorge Acurio-Palma, were commissioned at a late stage in the preparation of the volume to compensate for cancellation of other contributors on Quechuan varieties. For this reason, the information contained in these sketches could not be made available to the authors of the typological chapters as a source.

10

morphology in Chapter 18, Matthias Pache with verbal morphology in Chapter 19, Rik van Gijn with syntax in Chapter 20, and Karolina Grzech with discourse, information structure, and evidentiality in Chapter 21. In addition, in Chapter 22, Johanna Nichols explores differences in structural complexity in the languages of the Central Andes, which she interprets using a sophisticated theoretical model for language dynamics in relation to physical geography. Given the evidence for language contact as sketched above, complex and shifting patterns of bi- and multilingualism must have been the norm rather than the exception in the pre-Hispanic Central Andes, and they remain common today (Adelaar 2017b; Urban 2017). Language contact is thus a topic that looms large in Central Andean linguistics, and a privileged point of access for an approach to the language dynamics of the region that includes Quechuan and Aymaran as crucial players, but that is at the same time more inclusive and allows to incorporate evidence from all the now lost local languages that once were equally parts of the Central Andean linguistic ecology (Urban 2018a). This approach does not have to sacrifice the attunement to the historical development, to the contrary: As Peter Kaulicke emphasizes in Chapter 2 of this volume, throughout prehistory, the story of the Central Andes was one of interaction, whether local or large-distance, whether peaceful or conflictive, and whether intense or casual. It is this complex fabric of accumulated temporally layered patterns of interaction, some of which continuing right until the present day, that must form the joint background for the evidence for language contact that we observe in the linguistic record and that, in spite of the linguistic continuities that link them in all directions with neighboring regions, lends a certain coherence to the languages of the Central Andes.

1.4 The emergence and development of descriptive traditions in the languages of the Central Andes Of course, lexical and grammatical comparisons such as those sketched above presuppose lexical and grammatical descriptions. Compared to pioneering applications of modern linguistic frameworks to the description and documentation of unwritten languages (e.g. Boas and Deloria 1941) and to the early applications of the comparative method to unwritten languages like those of the Algonquian family (Bloomfield 1925), Central Andean linguistics as a scientific enterprise has a relatively short history that took off only in the 1960s. Paradoxically, in contrast, Central

introduction Andean linguistics as such has an unusually long tradition. The first Quechua grammar and dictionary were published in the year 1560 by the Dominican Friar Domingo de Santo Tomás, years before the first grammars of many major European vernaculars of the time, and the first Aymara grammar and dictionary by the Italian Jesuit Ludovico Bertonio, masterful and impressive works even today, followed suit in 1612. Also other (i.e. non-Quechuan, non-Aymaran) languages were documented by colonial grammarians, such as Mochica on the North Coast of Peru, for which de la Carrera’s (1644) grammar has remained the principal source until today. Unfortunately, not all achievements of colonial grammarians have been preserved until the present day, or they remain buried in ecclesiastic or secular archives. For some languages, we do know by later nominal references that grammars and other materials were composed, and we even know the names of the respective authors; if these materials were discovered, this would constitute a quantum leap in our understanding of languages like Quingnam, about which very little indeed is known. Colonial linguistics has developed elaborate language policies and reached impressive achievements in language documentation and description that are valuable for the purposes of historical linguistics, as Rodolfo CerrónPalomino shows at length in Chapter 3 of this volume. In other cases, where the language has been lost, such works are the only source of information that is available, underscoring the continuing value of this kind of work. To some extent the influence of pre-modern traditions of grammaticography and lexicography is still felt in Central Andean linguistics, e.g. when fused subject–object personal reference markers are called “transitions” still in present-day descriptive practice, following a terminology introduced by Santo Tomás. Nevertheless, pre-modern work cannot be taken at face value, as, in spite of intimate familiarity with the described languages, its authors did not have the descriptive and terminological apparatus of modern linguistics at their disposal, and were struggling in particular with distinctions or phenomena that the Latin-based grammaticographical tradition of the occident had not prepared them for, such as inverse systems. Also, there was no systematic phonetics or phonology, leading to impressionistic statements on the acoustic properties of sounds that seem clumsy to today’s linguists and force them to make educated guesses as to the articulatory and acoustic phenomena the authors of pre-modern works were describing. Such naïve impressionistic descriptions of phonetic and phonological variation continued to be commonly found in works dealing with linguistic variation in Central Andean linguistics well into the 20th century. For instance, the word “letters” often continued to be used with reference to phonetic and phonological

units, showing an insufficient awareness of the fact that the relationship between historically grown orthographies, phonetics, and phonology cannot be mapped one to one. One important class of pre-modern research on Central Andean languages after the colonial tradition represented by Santo Tomás, Bertonio, and many others, stems from the early 19th and 20th century, when German scholars like Walter Lehmann, Ernst Middendorf, Max Uhle, Hans Heinrich Brüning, but also Frenchmen like Jean Vellard, were carrying out primary fieldwork on the languages of the Central Andes. They produced documentation that is impressive in quantity and often also of a remarkably high quality by the standards of the time. For instance, Ernst Middendorf alone published a whole six-volume book series on Die Einheimischen Sprachen Perus ‘The Indigenous Language of Perus,’ including a grammar of Cuzco Quechua (Middendorf 1890a), a Quechua dictionary (Middendorf 1890b), a collection of Quechua verbal art (Middendorf 1890c; 1891a), an Aymara grammar (Middendorf 1891b), and studies on Mochica (Middendorf 1892). Ambiguities as to the correct interpretation of the data in such works remain in the case of languages which, unlike Quechuan and Aymaran, have no modern documentation. But they are an invaluable source of information for such languages, as shown the Chapter 10 by Katja Hannß on Uru and Chipaya and Chapter 11 by myself on Mochica. While there have been isolated works applying modern linguistic approaches to Andean languages before (including, for the Central Andes as defined here, Garro 1942; 1944; Rowe 1950; Harrington and Valcárel 1941; Wonderly 1952; Goins 1954; Solá 1958 on Quechuan; Tschopik 1951; Sebeok 1951 for Aymara), a breakthrough was achieved in the 1960s. Detailed descriptive data followed suit to the diachronic work of Parker (1963) and Torero (1964) that revolutionized Central Andean historical linguistics (more on this in Section 1.5), although data had been collected and analyzed at least partly hand-in-hand with the elaboration of the diachronic work. A collection of phonological sketches of highly divergent Quechuan varieties that allowed for an improved view on the internal diversity of the family were published in Escobar et al. (1967), and Parker (1965, 1969a) provides an in depth-description of Ayacucho Quechua; see Parker (1972a) for a more detailed review of research in this period. The increased visibility of the internal diversity of the Quechuan family was a further impetus for language documentation and language policy development. A particularly noteworthy achievement was the publication of grammars and dictionaries for six Peruvian types of Quechua that had been selected as regional quasistandards by the Peruvian Ministerio de Educación in 1976: Cuzco-Collao (Cusihuamán 1976a; 1976b), Ayacucho-Chanca

11

matthias urban (Soto Ruiz 1976a; 1976b), Junín-Huanca (Cerrón-Palomino 1976a; 1976b), Ancash-Huaylas (Parker 1976; Parker and Chávez 1976), Cajamarca-Cañaris (Quesada 1976a; 1976b), and the lowland variety of San Martín (Coombs et al. 1976; Park et al. 1976). Some of these seminal publications include data from heterogeneous lects from the regions covered, in spite of the fact that they were meant to have a normative character as well. As Adelaar with Muysken (2004: 193) note, “for highly interesting dialects such as Cajamarca, Cuzco and Huanca, they represent a half-way resting-place, rather than a terminus.” In fact, descriptive efforts have not ceased since 1976, and major contributions have been made since then to the description and documentation of Quechuan varieties (e.g. Adelaar 1977; 1982; Weber 1989; Taylor 1996a; 1999a; Hintz 2007; 2011; Shimelman 2017). At roughly the same time, studies of the Aymaran family took off thanks especially to the work of Hardman (1963; 1966), who was the first scholar to study the Jaqaru language of the central Peruvian highlands in depth. A description of Aymara followed by Hardman de Bautista (1974) and an important study on Aymara dialectology by Briggs (1976). This work then fed into diachronic comparative work (see Section 1.5). A major contribution in the new millennium is Coler’s (2014a) in-depth study of a southern Peruvian variety that also forms the basis for Coler’s Aymara sketch in Chapter 8 of this volume. While Uru has been the subject of attention of scholars in the early 20th century (e.g. Polo 1901; Métraux 1935; Vellard 1949; 1950; 1951), Chipaya had to wait longer for descriptions to become available, but now we have primary documentation that is superior to that for Uru, which has become dormant, in both quantity and quality (Cerrón-Palomino 2006b; Cerrón-Palomino and Ballón Aguirre 2011). Finally, of course, today Spanish is the dominant language of the Central Andes, and is in fact continuing to cause loss of Indigenous languages as it already has done in many cases in the past (see Section 1.8). At the same time, so-called Andean Spanish features a number of characteristics that it owns to the influence of Indigenous languages. In fact, research into these characteristics and their genesis is a topic on which a lot of research is carried out—more research, in fact, than on the description and documentation of Indigenous languages themselves. Accordingly, seminal studies such as Escobar (1978) for Peru have been followed up by further in-depth research on the characteristics of Andean Spanish by hispanists, often with a regional focus that allows to capture fine-grained variation and differences in Andean Spanish, e.g. Garatea (2010), Callisaya Apaza (2012), and Andrade Ciudad (2016). Willem Adelaar has repeatedly emphasized that the impression that Andean languages are generally well described, whereas the true terra incognita of South American

12

linguistics lies in Amazonia, is exaggerated. In particular, “[t]he relatively early date of most descriptive work on the major Andean languages necessarily implies that it was not yet heavily influenced by the advances in functional and typological linguistics that characterize most present-day descriptions of South American languages” (Adelaar 2012a: 17). As we have seen, major descriptive contributions that draw on the best of current descriptive practices have been made in recent years. Nevertheless, much work remains to be done (and this is even more true for Andean languages more broadly). For most other Indigenous languages, as for Uru, a particular challenge in gaining adequate descriptive knowledge is their early loss. Not all are blessed with colonial documentation of the same quantity and quality, and some were lost with only minimal or even with no (known) documentation. However, Central Andean linguistics has a long and fruitful tradition of making use not only of the rich colonial tradition of grammaticography and lexicography but also of ancillary data sources apart from primary documentation, such as colonial reports, lawsuits, and chronicles that contain linguistically relevant information (see Adelaar 2007b and Solís Fonseca 2009 for programmatic articles). For the Central Andes, this tradition is even more developed than for other parts of the world, including other parts than South America, for which similar historically oriented perspectives have a strong tradition as well. As Rodolfo CerrónPalomino shows in Chapter 3 of this volume, philology is therefore a crucial complementary discipline for Central Andean descriptive and historical linguistics. It is against that background that Cerrón-Palomino’s fervent plea for adequate and responsible treatment of such sources in attempts to understand the original linguistic situation better must be understood. Work on the analysis of such sources began more or less hand in hand with the elaboration of fieldwork-based sketches and the historical classification and reconstruction of Quechuan and Aymaran (see Section 1.5). In fact, the thesis of Alfredo Torero, one of the founding figures of modern Central Andean linguistics, was not on the classification of Quechuan for which he is best known, but a reconstitution of Puquina grammar and lexicon (Torero 1965), and throughout his career, he occupied himself with lost languages alongside those still spoken in the 20th century (Torero 1986; 1989; 1992; 1993; 1997). A similar broad orientation characterizes the work of two other major figures, Rodolfo Cerrón-Palomino and Willem F. H. Adelaar, who, alongside language documentation and historical-comparative work, have occupied themselves with the lost and poorly documented languages, the former especially in recent years focusing on Puquina (e.g. Cerrón-Palomino 2014; 2016f; 2020a; 2020b; 2020c), the latter especially on Culli (Adelaar 1988a;

introduction 2019). Even in the case of minimal documentation, such philological work with pre-modern sources does not have to be educated guesswork, but can rely on systematic and principled approaches. An example for such an approach comes from my own work; I call it “comparative colonial linguistics” in Urban (2020b). The basic idea is to attempt to contextualize puzzling aspects of pre-modern data concerning lost and poorly documented languages with the broader tradition of colonial linguistic traditions of the same region to arrive at sensible interpretations. The concrete example discussed in Urban (2020) is an alternation in cognates between the letter on the one hand with or on the other in 18th-century wordlists for two apparently closely related varieties of the Tallán language of northern Peru. Taken at face value, this alternation is difficult to make sense of indeed: through accumulated sound change over extended periods of linguistic development, regular sound correspondences that involve even highly heterogeneous types of sounds can arise, but it would be very surprising if [m] and [g] were implied in this particular case as otherwise the data suggest a very close relationship, possibly on the dialectal level. A key observation here is that colonial grammars often alternated between a letter for a nasal sound such as and and a (sometimes with diacritics like the tilde in the Tallán data) when trying to represent one particular sound for which the Latin alphabet does not have a dedicated letter: the velar nasal. Unlike for Tallán, for these ancillary sources one can be more confident about this identification because modern data that employ a more systematic transcription system is available in many cases. This underscores two points: on the one hand, the need for philological contextualization of data for poorly documented languages (see also e.g. Constenla Umaña 2000 for a similar statement in a different context), and, on the other hand, that philological work even for languages where modern documentation exists is worthwhile and can pay off in unexpected ways (Goddard 1973). The descriptive sketches in this volume contain the fruits of both types of work—the fieldwork-based descriptive and, where this is not possible or feasible, the philological reconstructive—and show the high standard of scholarship that has been achieved in Central Andean linguistics in both areas. The Quechuan language family is represented by sketches of four very different varieties that together cover the diversity of the language family as represented in the Central Andes. On the one hand, Carlos Molina-Vital, drawing on extensive primary fieldwork, provides a detailed sketch grammar of Huaylas Quechua, spoken in the northern Peruvian department of Ancash, in Chapter 4. His work on

this still vigorously spoken variety of Central Quechua provides many new and original insights into its phonological and morphosyntactic structure. In Chapter 5, Aviva Shimelman, drawing on her full-length fieldwork-based grammar (Shimelman 2017), provides a sketch of Southern Yauyos Quechua, part of a dialectally highly fragmented part of the Quechua-speaking area in Central Peru where the traditionally recognized major divisions of the family come together geographically and where their features locally overlap. Shimelman’s description reflects this fact and is therefore of considerable historical interest. In Chapter 6, Aviva Shimelman and Jairo Valqui portray Chachapoyas Quechua, an unusual Quechuan variety of the eastern cloud forest in Northern Peru that has undergone some very salient and unique changes that make it depart—in particular in its phonology—from more commonly encountered patterns in Quechuan. This chapter benefits from the first-hand experience of both scholars with the language, in particular Valqui’s phonetic and phonological research, but also draws on previous descriptive efforts by Gérald Taylor, who worked with the language in the 1970s when it was still more vigorously spoken than it is today. Finally, in Chapter 7, Raúl Bendezú-Araujo and Jorge Acurio-Palma describe Cuzco Quechua, a major variety of Quechuan in terms of speakers that is spoken not only in the city of Cuzco, but also in the extensive province of Cuzco of which it is the capital, and beyond. Cuzco Quechua is very similar though not identical to the Quechua spoken in Bolivia, and effortlessly mutually intelligible with it. Also this variety is of considerable historical and linguistic interest. For one thing, it is the only variety of Quechuan to fully distinguish between plain, aspirated, and ejective plosives, and it is still not clear whether this is an archaic feature retained from the protolanguage or a possibly contact-induced innovation (e.g. Landerman 1994; 1998; Campbell 1995; CerrónPalomino 2000b: 320–21). Being spoken in the former Inca heartland, it is probably the result of the chequered cultural history of this region in late pre-Hispanic and colonial times. A similarly fruitful combination of first-hand fieldworkbased expertise and sensible utilization of previous documentation is found in Matt Coler’s descriptions of the Aymaran family in Chapters 8 and 9 of this volume. Coler’s sketch grammar of Aymara draws on his full-length grammar (Coler 2014a) that is based on extensive fieldwork, whereas his sketch of Jaqaru in Chapter 9 is written on the basis of the seminal documentation by Hardman (1963; 1966) and subsequent work. Coler’s in-depth familiarity with Aymara puts him in a particularly advantageous position to make visible the similarities as well as the differences between these two languages.

13

matthias urban Katja Hannß’s treatment of Uru and Chipaya in Chapter 10 also relies on a mix of earlier documentation and own firsthand experience with one of the languages, Chipaya. In this particular case, such a mix is in fact necessary, since Uru is dormant and has to be made accessible through the utilization of relatively substantial documentation that has been realized mostly in the early 20th century. In addition to her experience with the documentation of Chipaya, Hannß can draw on significant earlier own efforts with regard to the reconstitution of Uru (Hannß 2008), putting her in an ideal position for a joint treatment of Uru and Chipaya. Philological research into colonial documentation and texts has also blossomed since the appearance of the last major overviews of the languages of the Central Andes (e.g. Alexander-Bakkerus 2005; Urban 2019b; Mossel et al. 2020). These major achievements are reflected in Astrid Alexander-Bakkerus’s Chapter 13 on Híbito and Cholón, and Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F.H. Adelaar’s description of Puquina in Chapter 12. The latter contribution in particular shows how very highquality descriptions can be achieved even on the basis of highly problematic and incomplete primary documentation without jumping to unwarranted conclusions. In addition, Chapter 11, written by myself, features a sketch of Mochica, one of the major languages of the Peruvian North coast, that is based on the efforts of a range of scholars who have explored the relatively rich array of pre-modern sources. Many of the descriptive sketches in this volume, however, are not (or not only) summaries of what was already known on the described languages, but also offer new analyses and significant new insights into their structure. One chapter where this is immediately visible is Chapter 12, in which Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar report significant new insights into the workings of Puquina grammar by recognizing for the first time the existence of two verb classes that highly regularly govern allomorph selection for a wide range of suffixes. On the basis of years of continuing philological research, they also offer a new and complete list of all lexical material available in the collection of texts in Oré (1607), the only dedicated documentation of the language. Likewise, Astrid Alexander-Bakkerus’s Chapter 13 provides a sketch of Cholón on the basis of her previous work (Alexander-Bakkerus 2005), but offers new insights especially into Híbito, underscoring that analysis of the minimal documentation available for this language is not exhausted yet and can still be pushed further. Although no new primary data have become available since the new millennium, Pieter Muysken’s Chapter 15 on Kallawaya offers new thinking on the possible origin of this remarkable mixed and secret language and its features, engaging

14

with significant recent research by Hannß (2017; 2019). Finally, while research on Andean Spanish is often aligned with the boundaries of modern nation-states, Luis Andrade Ciudad’s approach to Andean Spanish in Chapter 16 is unusual in several regards. His descriptive sketch specifically seeks to characterize the features of the Andean Spanish of Southern Peru and the Bolivian Altiplano across national boundaries, emphasizing the pre-Hispanic and colonial cultural-political continuities between these highland regions. Furthermore, while discussions of Andean Spanish often rightly emphasize the role of Quechuan and Aymaran substrate in the genesis of its dialectal configurations, Andrade Ciudad’s view is more balanced, pointing out that it is both interference-induced innovation as well as preservation of inherited peninsular structures that give it its characteristic imprint. While Andrade Ciudad’s contribution thus zooms in, in innovative ways, on a particular type of Andean Spanish and thereby provides fascinating perspectives on micro-variation, Anna María Escobar’s Chapter 27 provides more general perspectives on the features of Andean Spanish from the macro-level, focusing in particular on the Indigenous contribution to its emergence in the last centuries and highlighting the sociolinguistic aspects that led to a salient case of counter-hierarchical diffusion of linguistic features to a hegemonic language in a postcolonial setting. Another reason why some chapters add value beyond being mere summaries of extant research is that for the first time in a major English-language publication they make available and synthesize far-flung scholarship that has not hitherto been widely available. This is true of Carlos Molina-Vital’s description of Huaylas Quechua in Chapter 4, but also of Aviva Shimelman and Jairo Valqui’s sketch of Chachapoyas Quechua in Chapter 6, which draws on Valqui’s detailed acoustic phonetic analyses of this Quechuan variety and Shimelman’s unpublished work on morphosyntax. The result is a description of Chachapoyas Quechua that significantly goes beyond what had been achieved before, and is in fact the first such description of the language to become widely available in English. A third way in which the descriptive chapters in this volume are significant additions to the extant literature is that they will facilitate or even enable systematic comparison of the languages within and across language families in the first place. This is true of Matt Coler’s parallel description of Aymara and Jaqaru in Chapters 8 and 9, but is also and particularly strongly visible in Katja Hannß’s joint treatment of Uru and Chipaya in Chapter 10. Publications on Uru were scattered in far-flung and difficult to access publications from the late 19th and early 20th century, and her summary description of the language on the basis of these sources

introduction (Hannß 2008) used a somewhat different terminology and analytic approach from that of the major reference work on Chipaya (Cerrón-Palomino 2006b), making any direct comparison between the two languages a difficult enterprise. Hannß’s treatment makes such comparison much easier for the non-specialist, and at the same time provides a synthetic description of both languages that is an excellent starting point for further explorations of the languages’ structure in dedicated publications.

1.5 The emergence and development of the historical linguistics of the Central Andes Early scholarship, and in particular the clergy who required translations of Christian texts to Indigenous languages, were aware of linguistic variety and diversity in the Central Andes, within and across language groups. José de Acosta (1596), for instance, remarked that South America in general and the Andes in particular hosted “such a diverse forest of languages” (my translation from the Latin), to the dismay of the clergy for whose purposes homogeneous linguistic landscapes that would allow the indoctrination of the population more conveniently would have been preferable; indeed, as Adelaar (1999: 207) observes, “References to multilingualism are normally found embedded in complaints about linguistic heterogeneity as a major obstacle to christianization.” However, with Quechuan, Acosta and other churchmen were lucky in encountering relatively widespread and easily learnable languages that were understood by many Indigenous people, including those for whom it was not their first language. Accordingly, a form of Quechua was declared a lengua general ‘general language,’ as were Aymara and Puquina, granting them special status within the colonial administrative and ecclesiastic apparatus (see Itier 2011 and Adelaar 2017b for discussion; see also Chapter 25 by César Itier). But the clergy and its missionary linguists were—contrary to the portrayal of Quechuan as a single dialectally diverse language that is also implied by terminology that still prevails in the writings of some scholars—only too aware of internal differentiation within the family, because it stood in the way of the application of standardized materials and required some amount of adaptations to local speech in lexicon and grammar. Durston (2007b) provides extensive treatment of how the clergy dealt with this and other problems, and César Itier in Chapter 25 likewise provides an exemplary case of how language materials prepared on the basis of an artificial “Standard Colonial Quechua” (Durston 2014), created by the clergy on the basis of varieties of southern Peru and Bolivia (see Itier 2011),

were adopted to the Central Peruvian Quechuan varieties that are significantly different. That the adaptations were made in this direction—and materials were not, for instance, first prepared on the basis of Central Peruvian Quechuan varieties and then adapted to Southern Peruvian varieties—dovetails with a persistent language ideology in the Central Andes, according to which the Quechua of the Cuzco region in the Inca heartland, associated with the cultural splendor of these latest representatives of Andean statecraft, was equivalent to “proper” Quechua, in fact was Quechua, while the other varieties in the provinces, derogatorily called “dialects,” were considered degenerated forms associated with the uneducated, allegedly culturally inferior peasantry in the provinces (cf. Cerrón-Palomino 1985); similar Cuzco-centric language ideologies are sometimes entertained even in the present day, for instance by the Academia Mayor de la Lengua Quechua (Itiér 1992b; Niño-Murcia 1997; Marr 1999; Chapter 29 by Rosaleen Howard in this volume). This view is often associated with a crude diachronic theory that has been taken for granted for a long time. According to this view, the Quechuan family originated in the Inca heartland around Cuzco, and was spread throughout the Central Andes only in Inca times, and by Inca agency. Such a view (though without the ideological overtones) is still found as late as Rowe (1950), where at the same time the comparative method of historical linguistics was first applied to a limited selection of Quechuan varieties from the south. When samples of central and northern varieties were first published, this occurred in a form that was (even though perhaps with good intentions) to some extend standardized according to a Cuzco-centric model (Farfán 1952a according to Cerrón-Palomino 1985: 511–12), showing again and in a different way that issues with a Cuzco-centric view persisted until well into the 20th century even in academic contexts. This view (corresponding to “Phase 1” in terms of Paul Heggarty’s chronology in Chapter 23 of this volume) was shattered at least in academic circles in the 1960s, which brought an essential breakthrough in the historical linguistics of the Central Andes. Working independently of one another, Parker (1963) and Torero (1964) proposed classifications of the Quechuan family, worked out sound correspondences between the varieties, and, by a thorough application of the comparative method on behalf of both linguists, reconstructed the fundamentals of the protolanguage. The results of both authors were largely in agreement, forming what Paul Heggarty calls the “‘traditional’ linguistic vision” in Chapter 23 of this volume. Both ascertained hitherto unsuspected diversity in the dialects of Central Peru, whose importance had already been recognized by Ferrario (1956), and within this diverse region, a particularly dense array of variation in the highlands of Lima

15

matthias urban (prominently the Yauyos region, whose dialectal diversity is also sketched in Chapter 5 by Aviva Shimelman in this volume). Here, two major branches of the Quechuan language family, recognized in more or less the same way by both authors, were coming together in geographical space. These are called Quechua I by Torero and Quechua B by Parker on the one hand, and Quechua II and Quechua A by Parker on the other, with Quechua I/B covering the varieties of Central Peru, and Quechua II/A those to the north (even though their classification remains problematic to the present day) and those to the south. An alternative terminology that is also widespread emphasizes geographical distributions more than the purely arbitrary labels chosen by Torero and Parker, and does not necessarily imply that these groups can be defined by shared innovations in a tree-based classification.2 In this terminology, Central Quechua is used near-synonymously with Quechua I/B, whereas Southern Peruvian Quechua, or simply Southern Quechua, is used as a cover term for the varieties of Ayacucho, Cuzco, and sometimes Bolivia that immediately borders on Central Quechua/Quechua I/Quechua B in the south. Torero classifies these together as Quechua IIC. Together, the designations Central Quechua and Southern Quechua exhaust the varieties of the continuous zone. This newly emerging picture was a true revolution in prevalent thinking on Central Andean linguistic prehistory, for it showed clearly that the Cuzco-centric view of the Quechuan family was quite simply untenable, and at the same time that the Quechuan spread must have commenced long before the rise of the Inca empire. After the 1960s, historical work on the Quechuan family has not come to a stop, and has been slowly but steadily extended and refined (e.g. Adelaar 1984; Cerrón-Palomino 1987a, 1987b). The state of the art of Quechuan historical linguistics until the beginning of the 21st century, which attests to the high quality and depth of reconstructive and historical work that has penetrated many aspects of the internal history of the family, is exhaustively summarized in Cerrón-Palomino (2003[1987]). Since then, research has not stopped. Contributions such as Halm (2020) explain hitherto recalcitrant irregular correspondences between Quechuan varieties with reference to iconic values associated with palatal sounds in systems that are also documented synchronically by Adelaar (1977), De Reuse (1986), and Cerrón-Palomino (2016e), while Emlen and Dellert (2020) and Adelaar (2020a) push Quechuan reconstruction further by identifying submorphemic patterns in form–meaning correspondences in the mostly disyllabic lexicon of the protolanguage that strongly 2 Alternative Quechuan-based labels for the branches, i.e. Wáywash for Quechua I, Yúngay for Quechua II, Límay for Quechua IIA, and Chínchay for the node uniting Quechua IIB and Quechua IIC, suggested by Torero (2002: 55–8), have not caught on.

16

points to massive lexicalization processes in which earlier smaller monosyllabic lexical building blocks coagulated to yield the Proto-Quechua lexicon. It is these elements— rather than the Proto-Quechua lexicon, which to a large extent turns out to consist of secondary formations—that hold most promise as the input for a new search for external relationships of the Quechuan family and for opening a new chapter in the already mature field of Central Andean historical linguistics. On the other hand, the internal classification of Quechuan varieties, in spite of significant research efforts, has never been conclusively settled, and in fact, at the time of writing, is undergoing a phase in which the vision developed in the 1960s is re-examined from various angles and which might culminate in the first significant revision since the 1960s. Torero (1968) himself, in a consideration of the genealogical position of the Northern Peruvian dialects, was forced to admit, only some years after his seminal proposal for the classification of Quechuan, that there are signals in the data that are hard to reconcile with a dendritic (treebased) classification, and that are better described through criss-crossing dialect networks of affinities. This view has found supporters in Landerman (1991) and more recently in Pearce and Heggarty (2011), who break with the classification of Torero and Parker and suggest a generalized view of Quechuan internal diversity in line with a dialect continuum, only to attract a spirited defense of the reality of the basic genealogical split between Quechua I and II varieties by Adelaar (2013b). But as it looks at present, this was just the prelude to a more profound reconsideration of the very fundaments of views on the family structure of Quechuan that was more or less consensus from the onset of scientific inquiry in the 1960. Itier (to appear) argues that Ecuadorian Quichua descends from a low-level ancestor that featured ejective and aspirated consonants, which is in contrast to the until-then viable explanation of traces of these through an Inca-related Cuzco superstratum (Torero 1964). In a parallel vein, Floyd (2021) presents strong evidence for the presence of ejective and aspirated consonants in the Quechua documented by Domingo de Santo Tomás, suggesting, together with the Ecuadorian evidence, that this feature is one that might be entrenched in the family much more deeply than suggested by the alternative account in which these segments were adopted in Cuzco-Bolivian Quechua through Aymara contact—and indeed, Landerman (1994; 1998) and Campbell (1995) have long argued that an orthodox application of the comparative method requires positing of ejective and aspirated consonants for the protolanguage. At the same time, postnasal voicing and the merger of velar and uvular stops, the key defining features of the important Quechua IIB group, are seen with increasing skepticism as shared innovations that would reliably

introduction indicate common development and hence subgrouping— postnasal voicing is a cross-linguistically common phenomenon, and merger of velars and uvulars arguably a natural effect of contact with uvular-less languages in those areas covered by Quechua IIB, where we know of the former presence of non-Quechuan languages (CerrónPalomino 1990b and Paul Heggarty in Chapter 23 of this volume). Thanks to Hardman’s (1963; 1966) descriptive work on Jaqaru, not much later it became possible to establish firmly that there was a genealogical relationship between Jaqaru and Aymara (Hardman 1966b; 1975b; Torero 1972[1970]; Hardman-De-Bautista 1978). This language family is nowadays most commonly referred to as Aymaran, following Cerrón-Palomino (2000b); it has been, and occasionally still is, called Jaqi (a suggestion going back to by Hardman) and Aru (a suggestion going back to Torero), using the words for ‘person’ and ‘language,’ which are cognates shared by Jaqaru and Aymara, respectively, as designations for the linguistic grouping. The comparative material also allowed for the reconstruction of fundamental aspects of the protolanguage (Hardman 1975b; Hardman-De-Bautista 1978), parallel to the developments that took place slightly earlier in Quechuan linguistics. This paved the way for a modern Aymaran historical and descriptive linguistics; Cerrón-Palomino (2000b) summarizes the state of the art achieved in the 20th century, and also makes significant new contributions to the comparative study of the language family. Reconstruction of Uru-Chipaya is still in a more embryonic state, essentially restricted to a phonological reconstruction on behalf of Cerrón-Palomino (2007). The Quechuan–Aymaran relationship itself, which, as sketched in Section 1.3, is thought to be largely due to prehistoric and still ongoing language contact, is of course at the same time a problem of historical linguistics; in fact, it is the topic in Central Andean historical linguistics. Since initial formulations of the alternative genealogical hypothesis on the basis of the comparative method (Orr and Longacre 1968)—which still reflects to some extent a Cuzco-centric view of Quechuan is characterized by further problems with the use of comparative data— exploration of the relationship has become significantly more sophisticated (see Cerrón-Palomino 1982 and Proulx 1987 on the state of the art in the 1980s). As mentioned before, the genesis of ejectives and aspirated stops in Cuzco-Bolivian Quechua has been a matter of continuing debate (Mannheim 1991; Landerman 1994; 1998). It is still a topic of high relevance for Quechuan historical linguistics: Cerrón-Palomino (2000b: 320–21) believes he has found faint evidence for a reflex of aspirated stops as vowel length in Central Quechua, and Itier (to appear) argues for

a shallow, Inca-related chronology of Ecuadorian Quichua, which shows aspiration that relates in complex ways to that of Cuzco Quechua (Torero 1964). In another vein, a crucial step in understanding the complexities of the relationship was Adelaar’s (1986) innovative historical treatment of the Quechuan and Aymaran lexicon that led to the suggestion that lexical influence was mostly from the Quechuan to the Aymaran lineage; this work has been significantly deepened by Emlen (2017b), who also provides the most extensive Proto-Quechua and Proto-Aymara reconstructions to date. Research on language contact effects as described in Section 1.3 is highly relevant for understanding language history in the Central Andes too, as areal-typological patterns are the net accumulated result of events in prehistory, even though reconstructing these from the complex overall picture (Campbell 2006; 2017) may be difficult or impossible. Lexical borrowing also reflects actual interactions between speakers of different languages, and the semantic domains in which borrowings occur can afford important insights into the nature of the underlying interactions. For instance, from the mere fact that in Yanesha’, an Arawakan language of the eastern slopes of Central Peru, Quechuan loans from the Southern Quechua branch have to do with military force and societal rank, whereas loans from the neighboring Central Quechua have a different semantic profile, suggests two very different phases or types of interactions with speakers of different Quechuan varieties (Adelaar 2006a). And in a different vein, the fact that the Mochica word ⟨tūp⟩ ‘balsa, reed boat, totora reed’ (Brüning 2004) recurs in the lowland Quechuan variety of San Martín, on the other side of the Andes, as tupa ‘tree from which rafts are made’ (Park et al. 1976) is remarkable (Urban 2021e). Consistent with this, de la Carrera (1644) mentions a presence of Mochica speakers in a church outpost on the Marañón river that forms a major obstacle for east–west movement precisely in the area where San Martín Quechua is still spoken in colonial times; ethnohistoric sources (Zevallos Quiñones 1995) furthermore affirm a presence of mitmaqkuna—either forcefully or voluntarily resettled people—from the North Coast that were tasked with controlling a strategic crossing of the Marañón—possibly, as the linguistic evidence suggests, using rafts. In this way, linguistics can contribute to creating a very rich fabric of information on prehistoric interactions. A final broadly historical topic that has been explored for the Central Andes for many years, but that has gained in prominence especially in recent years, is the reconstruction of past linguistics landscapes. Scholars like Adelaar (1988a) for Culli, Torero (1986; 1989; 1993) for northern Peru, and Domínguez Faura (2014) for Puquina, among many others, have occupied themselves with reconstructing with as

17

matthias urban much precision as possible the former geographical extent of lost languages on the basis of historical records and toponymic distributions—an enterprise that is severely complicated as language distributions in the Central Andes often flowed into one another rather than being mutually exclusive, and more generally were frequently discontinuous (see Chapter 24 by myself in this volume). Such research has also led to the realization that, in certain parts of the Central Andes at least, there must once have been more languages than those mentioned in the colonial sources, betraying their former presence almost exclusively through a toponymic or anthroponymic fingerprint. Again, most prominent examples come from Northern Peru with its fragmented linguistic landscapes. Torero (1989) and Taylor (1990a) are pioneering studies here, but recent research has refined the picture and advanced new hypotheses (Andrade Ciudad 2010; 2011; Emlen and Mossel 2023; Urban 2019b; 2021a), even though uncertainties and insecurities concerning the proper interpretation of the often sparse available record remain. The relevant research is not only concerned with reconstructing linguistic landscapes themselves, but also with establishing the temporal chronologies of language replacement and the affiliations of the vanished languages. In this way, Andrade Ciudad (2010) sketches a scenario of language replacement in Northern Peru in which several languages succeeded one another in geographical space on the basis of the occurrence or non-occurrence of hybrid toponyms that are made up of elements of two different languages, and that hence suggest a period of temporal overlap of these in the area where such toponyms occur. Building on earlier suggestions by Torero (1989), Adelaar (2012b: 580), and Jolkesky (2016: 241: table 10), Urban (2021a) suggests that languages related to Híbito and Cholón were once much more widespread in Northern Peru, covering the linguistic void in the highlands of Northern Peru (see Figure 1.2). Such suggestions, too, are afflicted with considerable insecurity, even though, when several types of available evidence— toponymy and lexical traces, for instance—point in the same direction, and the evidence, according to several diagnostic criteria (Urban submitted), is consistent, they can at least be understood as the best available hypotheses for former language distributions in the Central Andes in prehistory. In this particular case, for instance, there are several possibilities of etymologizing Northern Peruvian highland toponymy through Cholón; one piece of evidence is that there are three place-names—Llacanora, Llacadén, and Llacamate—distributed widely over Northern Peru that all feature the initial sequence llaca-, and that all occur in or near areas of saliently reddish-ocher soil. It is therefore suggestive to isolate llaca as a toponymic element that reflects a morph in a now-dormant language, and, in this particular

18

case, it can be felicitously interpreted through Cholón ly aka ‘red’ (Alexander-Bakkerus 2005), yielding not only a very good formal match but also one that receives support from the real-world characteristics of the designated places. This volume reflects the strong historical orientation of Andean linguistics. Rodolfo Cerrón-Palomino provides many concrete examples of analyses of issues in Central Andean historical linguistics—not only from Quechuan, but also Aymaran—to illustrate the relevance of historical documentation and philological approaches in addition to the major analytic techniques of historical linguistics such as internal reconstruction and the comparative method. As already mentioned, Nicholas Q. Emlen (Chapter 26) provides an in-depth treatment of the history and current state of the art of research on the Quechuan–Aymaran relationship, distinguishing different phases and showing how, in each iteration, thinking on the issue changed, achieved improvements in the empirical data base, and grew in theoretical sophistication. How Quechuan and Aymaran languages spread is discussed in two chapters in the present volume: Paul Heggarty (Chapter 23) treats current views on the prehistoric spread of Quechuan and Aymaran before the Inca empire, whereas César Itier (Chapter 25) focuses on the further spread of Quechuan through and beyond the Central Andes in Inca and colonial times. These chapters both make rich reference to the non-linguistic social, political, and economical contexts. Theories on how language, culture, and society were interwoven in the past are discussed in more detail in the following section.

1.6 Language, culture, and society in the past Central Andean historical linguists have always sought to contextualize the historical linguistic evidence in an interdisciplinary fashion, and have thereby contributed to major and important reinterpretations of the prehistory of this culturally precocious region. The protagonists of the developments of the 1960s, Alfredo Torero in particular, were already well aware that their work had implications not just for Central Andean historical linguistics, but also for understanding Andean culture history at large. Torero’s and Parker’s work clearly demonstrated that the internal diversity in Quechuan was great enough to firmly reject a recent origin of the family with the Inca empire—a point already argued by Torero (1964) and developed further in Torero (1972[1970]; 1974) and Parker (1972b). The method of glottochronology, on which Torero

introduction (1972[1970]; 1974) relied for linguistic dating in his fundamental vision of the linguistic prehistory of the Central Andes in relation to its cultural trajectory, with its simplistic assumption of constant rates of change, is now widely rejected as inadequate (and more recent Bayesian approaches, while promising to eventually yield better estimates, are still fraught with methodological problems, Maurits et al. 2019). Nevertheless, it was becoming clear that the genesis of the Quechuan and Aymaran families goes back to deeper layers in the cultural chronology than the ephemeral Inca empire, and that their spread must instead be contextualized with much earlier developments in the cultural prehistory of the Central Andes. On the basis of glottochronological dates and other evidence, Torero thought that the expansion of Quechuan, Aymaran, and Puquina (apparently dialectally highly diverse, see also Chapter 12 by Nicholas Q. Emlen, Willem F. H. Adelaar, Simon van de Kerke, and Arjan Mossel in this volume) commenced approximately contemporaneously at the onset or during the so-called Middle Horizon in Andean culture history (see Chapter 2 by Peter Kaulicke in this volume): Quechuan would have started out from North-Central Peru, Aymaran from South-Central Peru, and Puquina from a homeland still further to the south. For those regions where representatives of more than one of the three lineages were present, their arrival, given their spatially staggered starting points, probably happened at different points of time. Indeed, Torero argued that the presence of Puquina on the Altiplano was more ancient than that of Aymara. In turn, Aymara, Torero argued, had a more ancient presence in Southern Peru—in particular the Inca heartland of Cuzco—than Quechuan. Thus, he established a relative chronology of language expansion and succession that in its essence is still accepted as valid, or at least considered plausible (see e.g. Heggarty 2008: 49). These purely linguistic considerations are discussed by Torero in concert with archeological evidence for the cultural prehistory of the Central Andes at the timedepths suggested by his glottochronological considerations. Given the supposed homeland in South-Central Peru, Torero pondered a relationship of the Aymaran lineage first with the Nasca archeological culture, which occupies roughly the right space at roughly the right time; at a slightly later point of time, as the lineages came to cover more extended geographical spaces, he considered an association of both Aymaran and Quechuan spreads with different archeologically visible aspects of the Huari phenomenon, in particular the centers and styles of Viñaque and Pachacamac. Torero considered Puquina, with its southern distribution, to have been associated with the Tiahuanaco society on the Altiplano (for more on all the archeological cultures, see Chapter 2 by Peter Kaulicke). This view of language succession and association with archeological cultures was, at this point in

time, also radically new, and clashed with an ideologically charged portrayal of Aymara as an autochthonous ancestral language of the Altiplano region, and in particular of the foundational Tiahuanaco civilization, that has been the dominant mode of interpretation earlier, especially in Bolivia (Uhle 1922; see also Cerrón-Palomino 1998a; 2000a). Also the association of Puquina with Tiahuanaco is the one that still prevails at least among linguists today (though the predecessor Pucará culture in roughly the same geographical space is now also frequently mentioned in the context of Puquina, e.g. Torero 1992: 183–4; Cerrón-Palomino 2010a: 258). Generally, the overall framework developed by Torero, including in particular the association of Aymara with Huari, remained prevalent among leading Central Andean historical linguists for decades to come. This does not mean that problems did not come to be felt: for one thing, as already noted, glottochronology has come to be viewed with increasing skepticism and has been virtually abandoned. Concrete aspects of the proposal—such as a certain lacuna regarding the force behind the original Quechuan spread, which Torero assigned to different agents for different branches and subbranches—remained somewhat unsatisfactory. Finally, central aspects of Torero’s vision are merely narrative sketches rather than empirically testable theories; Isbell (1983/1984: 256), in a scathing review, calls Torero’s (1972[1970]; 1974) and Bird et al.’s (1983/1984) interdisciplinary theorizing “examples of educated guess-work that employ unjustifiable assumptions and poorly documented interpretations.” Some refinements and alternative proposals have indeed been made in the following decades: Isbell (1974), for instance, argues for a view the genesis of the Quechua I/Quechua II divide that does not invoke archeological “cultures,” but advances in agriculture in different parts of the Central Andes that respond to different geophysical affordances as an explanans (an aspect that is also, in a different form, touched upon by César Itier in Chapter 25 of this volume); and, showcasing a link with agriculture in a different way, Bird et al. (1983/1984) suggest that maize cultivation and cultivars (land races) are a material correlate of Quechuan spread; see Isbell (1983/1984) for a methodologically oriented review of the time. Browman (1994) and Stanish (2003), both archeologists, meanwhile, continued to defend the older line of thinking that suggests a deep chronology for Aymaran on the Altiplano in general and an association with Tiahuanaco in particular. See Heggarty (2008), Cerrón-Palomino and Kaulicke (2010), and Heggarty and Beresford-Jones (2010b) for review of theorizing up to the new millennium. A major reconsideration of what Heggarty and BeresfordJones (2010b) call the “traditional model” only emerged

19

matthias urban in the late 2000s and early 2010s. Heggarty’s (2007; 2008) methodologically oriented papers on theorizing at the linguistics–archeology interface, with case studies from the Central Andes, can be thought of as a prelude to these developments. Inspired by processual archeology and the Language-Farming Dispersal Hypothesis that had been developed against a similar intellectual background (see also Renfrew 2010), Beresford-Jones and Heggarty (2010) insist that language spreads do not just happen, but that there are non-linguistic phenomena of commensurate scale that drive them—these can be of diverse kinds, and may involve ideological or religious movements, technological innovations, significant socioeconomic changes, and, most relevantly, political phenomena such as the expansion of polities and states over territories (Heggarty 2007: 48). The emphasis on processes and driving forces is the crucial aspect here, and shifts the focus from static views of linguistic and cultural distributions at a given point of time to a more dynamic view that has, on the linguistic side, language spread as the crucial explanandum. For the concrete case of the Central Andes, this immediately implies an emphasis on Quechuan and Aymaran, and, given the need for extralinguistic forces of commensurate scale that could account for their spread, a focus on the ‘horizons’ in Central Andean culture chronology; for a particularly clear statement of this, see Heggarty (2007: 37). Concretely, Beresford-Jones and Heggarty’s (2010) proposal is to turn the traditional model virtually on its head by assuming that it is the Quechuan lineage that commenced its spread in the context of the Huari cultural horizon, whereas Aymaran, in a bold move of theorizing, is brought into the context with the earlier Chavín phenomenon that radiated out from Northern Peru, where Aymaran languages are actually unattested, although faint toponymic traces mentioned informally by Cerrón-Palomino (2000b) suggest that Aymaran may once have been present (see now Emlen and Mossel 2023 for a much more detailed exploration of the toponymic record). Heggarty and Beresford-Jones succeeded in assembling a large array of scholars, including leading figures in the linguistics and the archeology of the Central Andes, to debate the proposal and more generally to rethink the linguistic and archeological evidence for prehistory in the Central Andes. Several conferences resulted in two edited volumes with contributions from participants (Kaulicke et al. 2010; Heggarty and Beresford-Jones 2012a). However, as discussed in the concluding chapter to the volumes by Adelaar (2010a), Heggarty and Beresford-Jones (2010b), and Heggarty and Beresford-Jones (2012b), the contributors were unable to agree whole-heartedly with Heggarty and Beresford-Jones’s model, and in fact a considerable diversity of opinions emerged. While some considered some form of an association of Quechuan with Huari to be plausible,

20

details differed considerably (Adelaar 2010b; 2012d; Isbell 2010; Makowski 2010), and others defended aspects of the traditional model or suggested revised versions thereof (Cerrón-Palomino 2010a). Other contributions consider the role of agriculture, in particular maize, agropastoralism, and cultural roles and identities associated with these practices (Heggarty and Beresford-Jones 2010a; Dillehay 2010; Lane 2010; Urton 2012b; Chapter 24 by myself in this volume). There have also been critical voices on the way the linguistics–archeology interface was conceived of in practice. One very contentious point pertains to the portrayal of Quechuan as a dialect continuum on behalf of BeresfordJones and Heggarty (2010) and Pearce and Heggarty (2011), an interpretation that forms part of the basis for their view of Quechuan prehistory; the basic split is fervently defended by Adelaar (2013b). This is a strictly linguistic problem. However, there are also more general methodological points. One is that the way linguistic developments were embedded in the broader context of cultural dynamics may have well involved more complex configurations than one reified archeological culture being associated with one single language—a frequent explicit or tacit assumption on the linguistics–archeology interface generally (see Urban 2021c for discussion). In fact, the case of the Incas, about which much more is known than on earlier political configurations because the last stages of the empire were directly observed by Europeans, evidence suggests that different languages in different sociolinguistic roles must be distinguished: a mysterious ceremonial and secret language that the ethnic Inca, according to the chronicler Garcilaso de la Vega (1829), used for special purposes, a language used for commercial and administrative purposes within the empire—thought to have been an Aymaran variety at first but later, as imperial ambitions became wider, Quechuan— and of course, the many Indigenous languages spoken in the Inca realm. This view of Inca linguistic history—which incidentally, mirrors the assumed language layering in southern Peru and Bolivia sketched above—is today associated mostly with Rodolfo Cerrón-Palomino (2013a; 2015), though it is developed already in Hardman (1985b). In a different vein, in Urban (2019b) and Urban (2021c), I develop a perspective which suggests that, rather than one language being associated with the expansion of one prehistoric polity, linguistic differences may have been active as semiotic indexes of social, economic, or political identity within societies, and that, especially for the North Coast, such linguistic differences may have enjoyed a high temporal consistency, possibly reaching back more than a millennium into prehistory. The principal interest of the dynamicized theorizing at the linguistics–archeology interface as proposed by Heggarty and Beresford-Jones lies in the languages that

introduction are spread by commensurate non-linguistic processes (Heggarty and Beresford-Jones 2010b); nevertheless, the Inca and Northern Peruvian cases show how this focus leaves out complexities that can contribute to a richer understanding of language distributions in the past. Kaulicke (2016b) has argued against the idea of an Aymaran spread with the Chavín phenomenon, which, given its nature, he considers to have been more compatible with a multilingual landscape (as was in fact observed in northern Peru in the 16th century). Moreover, a close look at the toponymic evidence does not strongly bear out the predictions made by the hypothesis (Emlen and Mossel 2023). Generally, one issue is that the strong focus on language spread as the explanandum in theorizing at the language–archeology interface and the focus on the horizons in Central Andean cultural chronology puts a very strong focus on the two major language families of the Central Andes. These are without doubt of high importance, but the theory necessarily leaves out less widespread and isolated languages—as the approach has no material to operate on in these cases (Heggarty 2007: 37)—and times in the Central Andean cultural chronology that were not characterized by imperial force and large-scale developments, but by different and more local modes of interregional interaction. This unnecessarily reduces the many ways in which the records of the two disciplines can be brought to bear on one another to examine language spreads and expansive processes. In the present volume, interdisciplinary research on the linguistic prehistory of the Central Andes is principally represented by three chapters: in my own contribution on language ecologies and language dynamics in the ancient Central Andes (Chapter 24), I am not so much concerned with the sociohistorical context of the spread of Quechuan and Aymaran, on which research has strongly focused, but rather in developing ethnographically, ethnohistorically, and archeologically informed models for the initial convergence between the lineages in their mutual homeland. This is a relatively neglected question (though see Adelaar 2012c; 2020a; Muysken 2012a). Here, in the vein of Urton (2012b), I develop an account of how a long-standing societal dualism between agriculturalists and pastoralists, who were united into a larger society linked by marriage ties, shared rituals, and exchange of produce, may have been aligned with linguistic differences. Two chapters focus on language spread. Paul Heggarty (Chapter 23), one of the protagonists of the interdisciplinary initiatives of the late 2000s and early 2010s, discusses language expansion and shift in prehistory up to the Inca empire. In Chapter 25, César Itier, a historian, philologist, and specialist in Quechuan linguistics, takes over. He discusses Inca policies and their impact on the linguistic landscapes of

the Central Andes, in particular the different ways in which these contributed to the further spread of Quechuan, in particular into the peripheries. Furthermore, “It stands to reason that a proper understanding of Andean colonial history is of major importance for the interpretation of the development and diversification of the main Andean language groups, because the effects of colonial interference have been considerable” (Adelaar 2020a: 6). This is a point that Itier emphasizes in Chapter 25 of this volume, too. Again, a complex network of policies and processes contributed to the further spread of Quechua in colonial times. These differed significantly in different parts of the Central Andes, depending on the nature of the religious, commercial, and economic policies and activities that prevailed within them. In my chapter on the small and lost languages of Northern Peru (Chapter 14), furthermore, I touch on the linguistics–archeology interface with reference to Northern Peru.

1.7 Language, culture, and society in the present Given that the volume aims to describe both past and present, it should contain an overview chapter on the present-day social context of Indigenous languages in the Central Andes, and the situation of their speakers, who are often the poorest and most underprivileged members of the societies of national-states that are themselves some of the poorest in the world. Here I provide a very brief outline; contributions by Rosaleen Howard (Chapter 29), on language ideologies, and by Joshua Shapero (Chapter 30), on the linguistically mediated engagement of Central Andean communities with the environment, contain material that also provides perspectives of this situation. Traditionally, communicative practices in the Central Andes resonate with the intense physical environment; these practices, as Joshua Shapero discusses in Chapter 30, “transform a semiotically heterogeneous physical world into an environment or landscape that yields resources, meaning, and social context to the humans that live within it.” Toponyms and systems of spatial language play a large role in the Central Andes, as they do elsewhere (see also Basso 1996; Urban 2018b); however, in at least parts of the Central Andes, language distributions traditionally themselves respond to altitudinal tiers of the Andes and associated socioeconomic specializations, although these patterns are now virtually disrupted as Indigenous communities become more and more integrated into national economies and political organizations (Chapter 24 by myself in this volume).

21

matthias urban Indeed, the societies of the Central Andes generally are undergoing rapid and potent social and economic change that affects all members of society in some way. In this context, increased mobility questions the traditional association between language, community, and places (Zavala 2020a), with cities like Lima harboring around half a million speakers of Quechuan alone (Chirinos Rivera 2001: 118). Meanwhile, the expansion of coffee farming on the eastern slopes of the Central Andes has created multilingual frontier societies as massive migrations have brought Quechuan and Aymaran speakers from the highlands to the region in search of economic opportunities; this and intermarriage with locals, also with Indigenous women who speak Arawakan languages, have led to complex ways of negotiating language use in the dynamic and complex social fabric of these frontier societies (Babel 2018; Emlen 2020a). Also more generally, in the context of the present-day life in the national states of the Central Andes, language, culture, and society stand in complex and occasionally tense relationships with one another. There are language ideologies in which Indigenous languages are a principal condensation point for social and personal identity, as when in Chapter 29 of this volume, Rosaleen Howard relates how Eliseo, a Quechua speaker from Tantamayo in Central Peru, describes himself as “born to Quechua,” and states that, for him, “Quechua means I am Eliseo. If I don’t speak Quechua, I am not Eliseo, I would be another person” (Howard’s translation from the Spanish). Such statements contrast with language ideologies that portray Indigenous languages as signs of uneducatedness, backwardness, and provinciality, as case studies in Howard’s chapter show. Such language ideologies are not exclusively associated with urban mestizos and criollos, but also have a reality for runa (mostly rural Indigenous people): being viewed as associated with an Indigenous language, even if only through an accent that suggests such an association, can have negative consequences for economic success and upward mobility. Linguistic features such as motoseo— the pronunciation of Spanish mid vowels that is influenced by the three-vowel systems of Quechuan and Aymaran— (see Chapter 16 by Luis Andrade Ciudad and Chapter 27 by Anna María Escobar) are extremely potent indexes of social identity in the Central Andes, and are strongly negatively evaluated in such language ideologies. The social enregisterment of linguistic features gains broader relevance in that these processes may influence their behavior in language contact (Babel 2011), and thereby, when accumulated over the long term, influence the typological trajectory of languages and thus contribute to the emergence of macro-patterns in areal typology as sketched in Section 1.3.

22

Linguistic relations are thus one of “dynamic tension, embedded in social interaction,” and are “indicative of the competitive forces at work within the culturally diverse and highly stratified society to which their speakers belong” (Howard-Malverde 1995: 162). The situation, however, often is not only predicated on a simple contrast between Spanish and Indigenous languages: In Bolivian provinces like Potosí (Howard-Malverde 1995), where Aymara was traditionally the main language of the rural populations and Quechua initially entered only in economically preferred areas (see also Chapter 25 by César Itier), Quechua enjoys more prestige and is associated with the modern urban lifestyle of Quechua–Spanish bilingual mestizo populations. In this part of the Central Andes, functional domains associated with Spanish and Quechua in the urban center are clearly recognizable, but boundaries are fluid and code-switching is frequent. Spanish is associated with upward mobility, yet there is no strong stigma on Quechua. Quechua, in turn, is dominant in the rural communities, being spoken by people whose grandparents most likely still spoke Aymara; Quechua–Aymara bilingualism is still frequent for the older generation, and Aymara—knowledge of which may be concealed from outsiders, presumably due to low prestige—is associated strongly with ritual contexts. Generally, Aymara is more associated with rural traditional life, a fact which is reflected in past and ongoing language shift from Aymara to Quechua. Relationships between Indigenous languages in Central Andean language ideologies may not be free of tension either: in the city of Puno, well known for being home to urban populations of both Cuzco-Bolivian Quechua and Aymara speakers, the languages are entwined by conflicting ideologies that, on the one hand, emphasize a common “indigeneity” but on the other project two distinct linguistically based ethnic identities that embrace the pre-Columbian past, in that the “Quechuas” are thought to descend from the Inca and “Aymara” from the Tiahuanaco empire and Qolla group (Naranayan 2018; see Section 1.5 on these ideologies from a historical linguistic point of view). “Tension” is also a word that Zavala (2020a) uses to describe the relationship between official policy toward Indigenous languages and linguistic rights, such as the 2011 Ley de Lenguas Indígenas in Peru, which recognizes language rights in a wide range of contexts (Howard 2011 and Chapter 29 by Rosaleen Howard discuss basics of official language policies in Peru and Bolivia). Policy-making in the 1970s and 1980s was spearheaded by Lima-based linguists and implemented in speaker communities that were not involved in decision-making (Zavala 2018). Efforts at intercultural bilingual education in rural areas with a high percentage of speakers of Indigenous first languages

introduction have acquired a “remedial and compensatory connotation” (Zavala 2019b: 60) in the context of the socially, economically, and politically marginal role of its speakers and the perceived association of Indigenous languages with poverty and lack of education. Sharp dichotomies such as urban vs rural, wealth vs poverty, associated linguistic differences between Spanish and Quechua, and the focus on L1 speakers arguably serve to cement imagined stereotypes (Zavala 2019a; note that, in reality, “there is not always a clear-cut class-based rural-urban dichotomy between Spanish and Quechua-speaking populations,” as Howard 2011: 190 writes). Language policies have become more “bottomup” in the past two decades, and programs have tended to become more community-based and led by Quechua speakers. The 2011 Peruvian policy of linguistic inclusion went hand in hand with a program for higher education for youth from disadvantaged backgrounds. However, instead of creating an inclusive language-based regional identity, social boundaries between the educated and the uneducated may only be recursively recreated in this context: Zavala (2018; 2019a) describes the language ideologies of educated Quechua speakers in Apurímac who emerged from study programs in intercultural bilingual education and who now work in language education for various governmental and non-governmental organizations. Often born in rural communities but now living in urban contexts where Spanish dominates, Indigenous language competence for these individuals is linked with some economic and social capital. Locally at least, such educators tend not only to create social boundaries between themselves, with local roots, and external linguists from Lima who would have implemented bilingual education programs earlier; they also assert a higher authority over Quechua than those speakers who do not have explicit awareness of the language’s grammatical structure and cannot write it, and those speakers who were not born like them in rural communities and who, while having linguistic competence, are said not to incarnate the authentic “Andean world-view” associated with Quechua that they purport to possess. This constellation of involved parties may (or may not) be a local one that is specific to the programs studied by Zavala; however, the case serves well to illustrate the complexities involved in the current sociolinguistic relations in which Indigenous languages and their speakers find themselves embedded. Meanwhile, young urban Quechua and Aymara speakers, with higher levels of education and access to social media, particularly in recent years, are developing new cultural practices and identities that transcend traditional dichotomies and stereotypes by eclectically drawing on elements from traditional Indigenous culture and language as well as modern pop cultures and media, thereby challenging

stereotypical semiotic roles of dress, language, and other indexes that typify particular social profiles (Hornberger and Swinehart 2012; Zavala 2018; 2019b).

1.8 Language shift and language endangerment From only the most superficial comparison of the maps in Figures 1.1 and 1.2, it becomes immediately obvious that language loss has impoverished the original linguistic landscape of the Central Andes, in particular but not only in Northern Peru. Adelaar (1999) discusses language loss in the Central Andes with special reference to this area; he (1999: 215–16) suggests a generally “pacific and gradual” transition to Spanish given that a lot of Indigenous aspects of culture are still visible until the present day. In some cases, the transition may even have been influenced by pre-Hispanic events: subjects of the Chimú empire on the North Coast, who had been subjugated by the Inca not long before the arrival of the Spaniards (see Chapter 2 by Peter Kaulicke in this volume), may have preferred Spanish to Quechua as their second language as “[i]nitially, it may have appeared less threatening to the survival of their culture” (Adelaar 1999: 216). While this is speculation, indeed the Quingnam language that once dominated in the Chimú heartland disappeared particularly early. While it was still documented in the 17th century, we can already trace the onsets of its demise through the travel diaries of Toribio de Mogrovejo (2006[1593–1605]), which attests to incipient loss already in the 16th century (Salas García 2010). In the south of the Central Andes, Puquina, in spite of its formerly wide spread and status as lengua general, was an early victim of language shift, too—a development that was in this case brought about by language shift to Aymara that probably continued seamlessly from late prehistoric into early colonial times and was further accelerated by a number of vicissitudes (Chapter 12 by Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar). With regard to Mochica, Salas García (2013) evaluates a rich array of colonial sources related to the gradual demise of the language's vitality in most of its former strongholds, becoming finally dormant when a railroad connected its last refuge at Eten on the North Coast of Peru with the outside world and brought it more and more in interregional contact and the confines of the modern economic and social system (Middendorf 1892: 45). A faintest last trace of survival of Mochica comes from recordings of Simón Quesquén, who passed in 1995 and who, while not being a speaker of the language, had, in a stereotyped form, acquired some aspects of the language from his

23

matthias urban grandmother as a child (Herrera Calderón et al. 2019). Urban (2019b) provides succinct discussions for all the North Coast languages; the overall picture is one of temporally layered processes of language loss from the 17th to the 20th century, as the different fates of Quingnam and Mochica already illustrate. The loss, in some cases, is very recent indeed: Culli was still documented in the early 20th century and probably survived until at least the mid-20th century. For Híbito and Cholón, thought for some time to have been lost already, Alexander-Bakkerus (2005; Chapter 13 in this volume) and Latorre (2014) were able to identify speakers who are now deceased. While Adelaar (1999: 217) expresses hopes that something similar could be achieved with regard to Culli, these have hitherto not been fulfilled. The fate of Uru, for which the situation is comparable to that of Culli in the north, was probably sealed only in the 20th century, when a severe drought brought an end to the communities that still used the language and when, without the support of daily use, the language began to become dormant (Vellard 1949). These late language losses, and the sometimes exceedingly poor state of documentation which could have been avoided, show “total lack of interest and support for the native languages on the side of the authorities” (Adelaar 1999: 216), suggesting “a combination of official neglect and a rapidly changing social context” (Adelaar 1999: 217) as the combined background for language loss. Compared to the fate of these languages, Quechuan and Aymaran seem to have fared better, and while it is true that they are still widely and sometimes vigorously spoken, this impression is deceiving, as the discussion in Section 1.7 already suggests. One linguist who continues to warn that the relatively large speaker base that Quechuan varieties still enjoy does not mean that Quechuan is “safe” is Willem Adelaar (2007a; 2014a). While on the national level, speaker numbers are actually increasing in absolute terms, this does not mean that the languages are gaining ground—this is just a concomitant of general population growth which in fact is much more pronounced than the growth in the number of speakers of Indigenous languages, which is, in comparison, rather modest (Adelaar 2007a: 15; 2014a: 5). Relative to the overall growth of the population, in fact, speaker numbers have consistently declined, and the same may be observed in the near future even in absolute terms (Adelaar 2007a: 15; see Howard 2011 for discussion of census data). Poverty increases infant mortality, reduces life expectancy, and makes Indigenous populations more vulnerable to violence, all contributing factors to the impediment of a stable and self-reproducing speaker population that keeps up with the general population growth in the countries of the Central Andes (Hornberger and Coronel-Molina 2004; Adelaar 2014a). The number of monolingual Quechua speakers in

24

Peru decreases steadily, ensuing bilingualism being a transient phase that leads to Spanish monolingualism. Overall Quechuan is yielding functional domains to Spanish (Von Gleich and Wo¨lck 1994). Locally, perspectives can be even more somber. Language shift is eating away in the continuous zone as speakers shift to Spanish in regions like Junín, Pasco, Tarma, and Yauli, which have seen a drastic decline in speaker numbers in the second half of the 20th century (Adelaar 2007a: 14; 2014a: 5). In fact, through the loss of the language in these areas, Quechuan’s continuous zone has been virtually bisected by now, and speakers in the two discontinuous remnant areas to the north and the south “will find it harder and harder to preserve their linguistic identities” (Adelaar 2014a: 6). Language ideologies of inferiority that often forestall individuals with an Indigenous linguistic background from economic opportunities, as described in Section 1.7 and illustrated in more detail by Rosaleen Howard in Chapter 29 of this volume, play a large role in the rupture in the transmission of the language to the next generation (Adelaar 2007a: 14; see also Hornberger and Coronel-Molina 2004: 14), leading to relatively rapid turnover of Quechuan-dominant parent generations and already Spanish-dominant child generations who no longer use the language even at home with their relatives (Adelaar 2007a: 14–15). “Centuries of oppression and social marginalization have convinced Quechua speakers that progress will only be possible if they shift to a dominant language first” (Adelaar 2014a: 9). In Chapter 29, Rosaleen Howard illustrates the point that even motoseo as a substrate feature is harmful, or even more harmful than being a speaker of a native language, to economic success. Migration to urban centers is another factor involved in thinning out the Quechua-speaking population in the countryside, and the urban context of Lima, where the language is particularly stigmatized, provides unfavorable conditions for its preservation (Hornberger and Coronel Molina 2004: 15; Marr 2011; Adelaar 2014a: 5). Hornberger and CoronelMolina (2004) provide a detailed discussion of patterns of use and decline of use in different communicative domains, and Howard (2004) contains a case study on the gradual loss of the language in Tantamayo in Central Peru. While Jaqaru is severely endangered, and in particular the Cauqui variety is close to becoming dormant (Hardman 2000: 1; Hardman et al. n.d.), the situation of Aymara is broadly comparable to that of Quechuan: while the overall speaker population is still large, Bolivia records declining speaker numbers and an increase of Spanish–Aymara bilingualism (Albó 1995), macro-parameters that are the telltale signs of impeding language endangerment. Also on the community level, patterns that are similar to those for Quechuan are observed, with largely negative language attitudes toward

introduction Aymara in light of the association of Spanish with economic opportunities and wealth; middle-aged people in one community in Southern Peru are passive bilinguals, and there is a rupture in the transmission of the language to the youngest generation (Coler 2014a: 24–5). Rural exodus likewise contributes to a thinning of the speaker base. In contrast with the situation of Quechuan in Lima, however, the situation in Bolivian cities (Howard 2011: 196)—and especially El Alto, a very fast-growing community next to La Paz, with which it is intimately linked—differs somewhat. With a population of nearly one million—the second-largest city of Bolivia—El Alto has an urban population that consists dominantly of recent immigrants from the rural Altiplano who are ethnic Aymaras and who often maintain residences also in the countryside (Albó 2016; Howard 2011: 196). While not everyone in El Alto speaks the language, Aymara is widely spoken and radio programs in Aymara support language use linked with an ideology of language purism (Swinehart 2009); a similar bilingual Quechua–Aymara radio program exists in the city of Puno in Peru, another town of the Central Andes known for an urban Indigenous population (Narayanan 2018). One point that Adelaar (2014a: 2) emphasizes is that precisely the dispersed settlement patterns of rural populations, where the same process of language shift takes place simultaneously in too many places to mount efficient countermeasures, is something that lowers the prognosis for the survival of Indigenous languages in the Central Andes in spite of still relatively high speaker numbers. This, he points out, is in contrast to smaller but more tightly knit speech communities in Amazonia, who are therefore more easily reached as a group. In addition, bilingual education is not usually seen positively given the experience that competence or even mere association with an Indigenous language through pronunciation features may impede social and economic success (Hornberger and Coronel-Molina 2004: 14; Adelaar 2014a: 9). There is one case in the Central Andes, however, where impending language loss has been successfully reversed: this is the case of Chipaya, which has experienced an impressive and unexpected increase in speaker numbers since the first half of the 20th century; in the community of Santa Ana de Chipaya, the language is now widely learned as a first language at home while Spanish is taught at school. According to Cerrón-Palomino (2018a), conducive factors include speakers’ awareness of ethnic identity, identification with the language as one of its emblems, and a concomitantly positive language attitude. This is in contrast with attitudes toward many of the declining Quechuan varieties, which, while varying (see Manley 2008 and von Gleich 2016 for nuanced regional perspectives), are often negative due to language ideologies that portray Quechua

as primitive and backward, and that influence economic opportunities and social upward mobility of speakers negatively. Whether efforts at language planning and preservation for Quechuan varieties as described in detail by Hornberger and Coronel-Molina (2004) will be sufficient to preserve Quechua in the long run, and if so to what extent of its original internal diversity, remains to be seen; a crucial factor to success seems to be not only the development of literacy, bilingual education, and language standardization to make Quechuan fit for those domains of life where Spanish dominates, but to provide support for the language in its traditional stronghold: the intimacy of households and community life (Luykx 2004). The Chipaya case underscores how crucial this may be for language preservation. At the same time, on the North Coast, where the last language, Mochica, went dormant around a century ago, a general revalorization of the rich pre-Hispanic heritage (Silverman 2005; Expósito Martín 2020) provides favorable conditions for language revitalization. In fact, several initiatives have been introduced to teach a new version of Mochica as a second language. Pedagogical materials (e.g. Chero Zurita et el. 2012; Chero Zurita and Peralta Vallejos 2017), designed specifically for teaching in classrooms, have become available to support this task. Vigorous revitalization and/or strengthening efforts are not only also underway for Uru (Zurita Aguilar 2018), but also for Aymara as spoken in Southern Peru (Arocutipa Saira 2020) and Chile (Abarzúa Silva 2016), where the language is under significant pressure from Spanish (Gundermann et al. 2009), as well as for some of those areas where Quechuan varieties are losing ground to Spanish, e.g. rural Cuzco (Yana Ccorimanya and Gutiérrez Mendoza 2011).

1.9 The broader relevance of Central Andean languages and linguistics Central Andean languages had, and continue to have, a considerable role in investigations into topics of general interest. Data from Central Andean languages have contributed to linguists’ understanding of phenomena as different as evidentiality, language contact, or numerical cognition, and they have also contributed to theoretical advances in linguistic theory. The considerable relevance of Central Andean data for contact linguistics is already evident enough from the discussion of the Quechuan–Aymaran relationship in Section 1.3 and in Nicholas Q. Emlen’s much more extensive review in Chapter 26 of this volume. Together

25

matthias urban with Dawkins’s (1916) study of Asia Minor Greek or Gumperz and Wilson’s (1971) work in Kupwar, it is a classic exemplification of the extent to which language contact can reshape a language. But, as reviewed in Section 1.3, language contact is a topic of broader relevance in the Central Andes. The Kallawaya language of Bolivia, described in Chapter 15 of this volume by Pieter Muysken, is of particular theoretical interest for being at the same time a mixed and a secret language. At the same time as offering many puzzles when it comes to origins (see Hannβ 2017; 2019, and Chapter 15 by Pieter Muysken), the mixed nature of the language, with lexical and grammatical material having different origins, is also of interest for questions of psycholinguistic processing (Muysken 2011b). In more recent years, it is especially Spanish–Quechuan language contact that has been analyzed in depth and from a wide range of different angles. Anna-María Escobar, in Chapter 27 of this volume, provides an overview of the most salient outcomes of the contact situation. The literature on the topic is vast, ranging from corpus-based (Van Hout and Muysken 1994) to experiment-based (Muntendam and Torreira 2016) approaches, and covers phenomena in phonetics, phonology (e.g. O’Rourke 2010; 2012), morphology (e.g. Rodríguez Garrido 1982), syntax (e.g. Muntendam 2013), and semantics and discourse pragmatics (e.g. Zavala 2001). Frequently, analyses of Spanish–Quechuan language contact phenomena are couched in generative frameworks, linking empirical data with linguistic theory (e.g. Sánchez 1999; 2003; Luján and Parodi 2001; Muntendam 2008b). Also more generally, generative analyses on languages of the Central Andes are available. Early formal work by Muysken (1979, 1981b) and reply by Weber (1983b) was relevant to the question of the (non-)autonomy of morphology and syntax (cf. D’Alessandro 2019: 11 for a contemporary review), and Quechuan data has also contributed to current and past issues in generative syntax, often—interestingly but not unsurprisingly given their agglutinative character— on questions at the interface of morphology and syntax. Topics where Quechua has contributed data include binding (Weber 1993, 1994), cyclicity (Cole 1982b), agreement (Myler 2016), and especially nominalization (Lefebvre and Muysken 1988; Hastings 2004; Cole and Hermon 2011). Generative research on Quechuan has not come to a halt, and in fact, has kept up with theoretical developments; in the process, attention has shifted more to pragmatics and information structure, which is analyzed in a minimalist framework for instance by Sánchez (2010). Furthermore, there is a relatively rich literature on formal semantic analysis of Quechuan morphology (e.g. Faller 2011; 2012; Martínez Vera 2020).

26

But also apart from formal and generative analyses, data from Central Andean languages have contributed to the clarification and sharpening of more generally relevant debates and issues. Thus, Quechuan data have been used to shed light on the adjective as a cross-linguistic category (Floyd 2011) and on the cross-linguistic range and semantic structure of grammatical morphology (Faller 2020); in particular, evidentiality is an area of research that has benefitted from data from Central Andean languages (e.g. Floyd 1999; Faller 2011; Howard 2012; 2018; Hintz and Hintz 2017; Andrade Ciudad 2020; see in particular also Chapter 21 by Karolina Grzech in this volume). In terms of our knowledge of linguistic diversity, especially also typologically unusual and exotic structures, Central Andean languages, including ones only documented in pre-modern sources, have shown their relevance. For instance, Hovdhaugen (2005a) discusses a very special case of case stacking (“Suffixaufnahme”) in Mochica, and Adelaar (1994a) shows a special grammatical treatment of a very specific type of phenomenon—manifestations of the supernatural—in the colonial Huarochirí manuscript, a most important primary source for pre-Hispanic Andean religion redacted in Quechua. This special grammatical treatment is obviously related to cultural factors. Indeed, the conceptual structures enshrined in the languages of the Central Andes have important insights into human culture and cognition to offer for anthropologists and cognitive scientists; these are of particular relevance, as cognitive scientists have become aware of the requirement for studying cognition in a wide array of cultural and linguistic contexts (Henrich et al. 2010; Majid and Levinson 2010). For instance, data from the Aymara language and gestural behavior of its speakers show, according to Núñez and Sweetser (2006), how the Aymara past is consistently conceptualized as lying in front of the individual and the future as lying behind, showing that cognitive models that prevail in languages like English where the reverse is the case (as seen in expressions like you have a bright future ahead of you or you must leave your past behind) are not universal. In a different vein, based on long-term ethnographic work in the Peruvian Andes, Urton with Nina Llanos (1997) provide a deeply insightful discussion of Quechuan numerals and numerical cognition that explains, among many other things, how Quechuan women can effortlessly master highly complex computations required in weaving, the Andean handicraft par excellence. The stylistic devices used in the verbal art of the languages of the Central Andes add to our understanding of humans’ creative use of language. Traditions such as semantic couplets—in which semantically similar but not identical words are paired to provide structure to a piece of verbal art in the same way as a rhyme in Western poetry—play

introduction against a more generalized background of Amerindian verbal art in which such phenomena are common. Semantic couplets provide rich means for Central Andean poets to express subtle nuances in meaning. At the same time, analysis of the ways in which near-synonyms are paired together in Central Andean verbal art provides unique opportunities for linguists and cultural anthropologists to investigate the semantics of Central Andean languages and the thought of their speakers (Mannheim 1986a; 1987a; 1998; 2015; Coler et al. 2020). Likewise, Andean storytellers have unique ways of creatively knotting together extant motives in Andean narrative in novel ways, and of reversing and altering relationships of actors that respond deeply to more general aspects of Central Andean culture and can best be understood against the background of these (Allen 2012; Marín-Dale 2016). Finally, as reviewed in Section 1.6, Central languages have also played a key role in exploring the interfaces of historical linguistics with archeology, in particular the relationship between the introduction of agricultural techniques and language expansion in prehistory (Heggarty and Beresford-Jones 2010a). Phenomena such as the Quechuan– Aymaran relationship demand careful and deep thinking to develop models for prehistoric interaction that are consistent with the observed intensity of contact effects (Adelaar 2012c, 2020a; see also Chapter 24 by myself in this volume); more generally, linguistic and cultural distributions in past and present suggest that evidence from the Central Andes has much to contribute to a richer picture of how linguistic and archeological data can correspond than simple one-to-one mappings of (proto-)languages onto archeological cultures (Urban and Barbieri 2020; Urban 2021c). In sum, the languages of the Central Andes have influenced and continue to influence linguistic thinking more broadly in multiple respects.

1.10 The present volume As the diversity of topics treated in this introduction shows, this volume aims to be a guide to the research on languages in the Central Andes in all its diversity and heterogeneity. Central Andean linguistics, in the minds of outsiders and perhaps also in the view of some insiders, is frequently equated with Quechuan and Aymaran linguistics. And while it is true that these families—with regard to descriptive efforts of individual varieties (Section 1.4), historical linguistics (Section 1.5), discussions of areal typology and language contact (Section 1.3), and sources of data for more general questions in anthropology and linguistics (Section 1.9)—are

a crucial focal point of attention, it is equally clear that they were not, and are not, the only languages of the Central Andes, and that a once much more diverse linguistic landscape has vanished over the past 500 years through language loss (Sections 1.2 and 1.8). Especially in recent years it has become obvious how lack of attention to, and often even basic awareness of, this fact has led to Quechuan and Aymaran-centric visions of the Central Andes that sometimes present incomplete pictures of the linguistic diversity of the Central Andes, including areal typology and interdisciplinary approaches to Central Andean culture history, and how more inclusive approaches can help paint a more nuanced picture of the linguistics and the human story in this part of South America (Urban 2019a; Urban and Barbieri 2020). One basic concern of this volume, therefore, is to make a case for the benefits of a more balanced and inclusive view that considers Central Andean languages under the umbrella of continuing cultural and linguistic interaction as sketched in Section 1.3, including the tense and complex current sociolinguistic fabric sketched in Section 1.7. Other than that, the volume is “non-partisan” and does not attempt to push a particular point of view on areal typology, historical linguistics, theorizing at the linguistics– archeology interface, or other matters on which there is no unanimity and consensus. There is quite substantial heterogeneity in opinions, with scholars emphasizing different aspects of the evidence and thus preferring different scenarios and narratives on a wide range of topics; in particular, the current state of the art on Quechuan classification is in such a flux that any a priori fixation on a particular view is rather likely to become outdated soon. I believe the field is at the moment better served by offering a balance of the state of the art that may serve as a springboard for further research. More generally, my conviction is that a good and substantial guide is more comprehensive and therefore more complex than an introductory volume, and that it should reflect the complexities of the field, including differing and possibly conflictive views on major topics, rather than presenting a simplified and homogeneous view that renders invisible loose threads of research and debates that are always present in a scholarly community to varying extents. Instead, I believe that a guide should introduce these, explain the reasons for diverging assumptions, and thus enable the reader to find their way through the primary literature on their own. This introductory chapter can hopefully serve to introduce readers to the general panorama of linguistic diversity in the Central Andes, past and present, and research topics in Central Andean linguistics; hopefully, it will enable them to contextualize the individual contributions to the volume and serve as a sort of guide to the guide that the volume as a whole aims to be. I am confident that the full set

27

matthias urban of contributions in this volume will do justice to the comprehensiveness and detailedness that the Oxford Guides to the Language of the World aim to offer, and that the richness of scholarship on Central Andean languages actually requires to be fully appreciated.

1.11 Conventions 1.11.1 Glossonyms and spelling conventions for other proper names Are Quechua and Aymara designations of (dialectally diverse) languages or language families? While any serious scholarship emphasizes that the internal differentiation within the families is far beyond the trivializing impression that the historically grown use of the term “dialect” for individual Quechuan varieties implies, as far as terminology is concerned, there are two schools, and both have arguments to bolster in defense of their position. In one solution, used in particular in Spanish-language publications and explicitly proposed by Cerrón-Palomino (1994), both “Quechua” and “Aymara” are used as the names for language families and individual languages and varieties—in the case of Quechua, preceded by a modifier that usually refers to the geographical region in which the variety is spoken, e.g. Yauyos Quechua, Ayacucho Quechua, etc. In the case of Aymaran, this terminological school usually calls Jaqaru “Central Aymara” and Aymara proper “Southern Aymara.” The other solution, also widely adopted now but with a shorter tradition, is to use “Quechua” for individual varieties, again with a modifier as to provide the differentia specifica, to use “Jaqaru” and Aymara” for the constituent members of the Aymaran family, and to use the designations “Quechuan” and “Aymaran” respectively for the families (Adelaar 2012b). This practice follows widespread similar usage around the world, especially in work that is written in English in using the suffixoid -an for names of language families. For the Central Andes, this solution has two distinct advantages: first, usage of “Aymaran” for the second major language family of the Central Andes makes for a direct analogy and parallel to “Quechuan” (see Cerrón-Palomino 1994); the terms Jaqi, as proposed by Hardman (1966), and Aru, proposed by Torero (1970[1972), are avoided here. On the one hand, for Quechuan, this usage renders immediately visible the fact that we are not dealing with a single language, and that description of Quechuan internal diversity is not only a matter of dialectology, a trivializing impression that prevailing terminology would suggest (see also Luykx et al. 2016: 162). Further reasons against talking

28

of internal variation in Quechuan (exclusively) in terms of dialects is that this is historically ideologically associated with the now outdated Cuzco-centric views of Quechuan history as reviewed in Section 1.5, and also with a more general and especially historically relevant contempt for Indigenous languages that would deny them that status of full-fledged languages. Another advantage, but also a disadvantage, is that distinguishing between “Quechua” and “Quechuan” forces precision in scope when making claims on the family and on individual varieties. As the above discussion makes clear, contributions to this volume use “variety” rather than “dialect” to refer to individual Quechuan lects. Aside from undesirable historical connotations, this terminology acknowledges that it is still unclear how many Quechua “dialects” should be distinguished. In that situation, the term “dialect” potentially unduly reifies what are really rather fluent continua; the term “variety” is more neutral in this regard. Glossonyms are, however, not only an issue for the two large language families, but also for the “minor” languages and language families. For some of these, a multitude of confusing designations coexist. Some of these are rooted in the history of research, as when Puquina and Uru-Chipaya were confused because, apparently, some speakers of Uru and Chipaya referred to their language as Puquina (Wachtel 1994: 123). Similarly, Kallawaya, the mixed and secret language, has sometimes been referred to as Pohena (Soria Lens 1951: 32), hinting at the relationship that exists with Puquina but falsely invoking the idea that the two can immediately be equated, as has sometimes been done prematurely (Brinton 1890; De la Grasserie 1894). In this volume, Uru-Chipaya (rather than Uruquilla or Uro, two of the most saliently available alternatives) is used to refer to the language family, following what seems to the most common usage at least in the English-language literature. Likewise based mostly on considerations of widespread use, but also because of the ambiguities associated with the most prominent alternative, Yunga (Itier 2015), the prominent language of the North Coast described by de la Carrera (1644) is referred to by the name Mochica in this volume, and its southern neighbor is referred to as Quingnam rather than the likewise more ambiguous alternative Pescadora (see Urban 2019a and further literature cited there for discussion). Another dimension of variation is that language names, even when they are in principle consistently applied, are represented inconsistently orthographically, with more traditional alternatives that are based on Spanish orthographic conventions, such as Puquina and Callawaya, being used alongside more phonologically oriented spellings, such as Pukina and Kallawaya. In this volume, generally the Spanishbased orthography is preferred, except for Kallawaya, where the phonemic alternative appears to have gained so much

introduction ground that it is now the most widely used orthographic representation of the name of the language. Spelling is not the only issue in the linguistics of the Central Andes, but also when it comes to placenames, names of archeological cultures, etc. Here, I generally follow Spanish orthographic conventions. Where the etymology of place-names or other proper names is at stake, they may be repeated a second time in a spelling that is appropriate to the Indigenous language from which it derives, e.g. “Cuzco (Qusqu, perhaps from Aymara qusqu ‘owl’ according to Cerrón-Palomino 2006a), was the capital of the Inca empire.” If for etymological purposes the spelling of a proper name as it appears in an early colonial source or other document characterized by nonstandardized orthography is relevant, may be placed around it.

1.11.2 Orthographies and other conventions for the representation of linguistic data However, an even greater issue is the orthographic representation of linguistic data itself. This is a particular vexed issue in the Central Andes, where “orthographic chaos” (Cerrón-Palomino 1992: 122, my translation from the Spanish) has reigned and often continues to reign. In spite of early orthographic standardization efforts by the clergy for both Quechuan and Aymaran in the 16th century, a wide range of different spelling systems came into being that— unsurprisingly given the generally unregulated orthography in use for the Spanish language itself—were inconsistent in use across different colonial grammarians and typically even within the works of one and the same colonial grammarian (see e.g. Calvo Pérez 2005 for some aspects of this). For instance, in colonial sources, doubling of consonants, e.g. , was often used to indicate glottalization, for instance in Bertonio (1612b), but not consistently; some modern orthographies have taken over this practice. In modern times, there are a number of reasons for the great number of different spelling systems that are, in spite of some general phonological similarities between the languages of the Central Andes (see Lev Michael and Allegra Robertson’s discussion in Chapter 17 of this volume), in use. One is that many practical orthographies are based on phonemic principles. However, speakers of mostly unwritten languages who (if at all) were accustomed to writing one particular language, Spanish, may struggle to adopt different logic in spelling; accordingly, Weber (1998a; 2006) presents a plea for the application of Spanish-based orthographies to Quechuan, and such a system is in fact

employed for Huallaga Quechua. Another issue is the proliferation of regional variation in pronunciation, in particular within the Quechuan language family. The gradually shifting properties of varieties make it difficult to isolate and reify them into separate linguistic entities for which orthographies could be designed (cf. Luykx et al. 2016, and note that speakers of individual varieties are often unaware of regional variation, instead assuming that Quechua was indeed the reified homogeneous language as which it has often been portrayed). This is exacerbated by the fact that regional language planning is carried out by both government and non-government actors, national and non-national linguists, and other interest groups that often differ in their backgrounds, goals, language ideologies, and, accordingly, orthographic preferences. Hornberger and King (1998) and Howard (2015) discuss the issues involved in the creation of a unified alphabet for Quechuan in Peru. First orthographic standardization initiatives date to the 1950s and, with official participation, various successor orthographies have been proposed through time, e.g. the 1975 Alfabeto Básico General del Quechua for use in the 1976 grammars and dictionaries mentioned in Section 1.4; as dissatisfaction with this orthography grew (cf. Cerrón-Palomino 1992: 125), it was revised again in 1985, leading to the Alfabeto Oficial or Panalfabeto, and again in 2000. However, NGOs such as the Academia Mayor de La Lengua Quechua or the Summer Institute of Linguistics keep advertising and promoting their own orthographic preferences, so that no consistently applied alphabet of universal use has emerged so far (Hornberger and Coronel-Molina 2012: 40; see Cerrón-Palomino 1992; Coronel Molina 1996: 9–12; Niño-Murcia 1997: 152–5; and Hornberger and Coronal Molina 2004: 39–40 for brief overviews of past initiatives). One particularly fiercely debated question is whether Quechuan varieties, featuring a three-vowel system with mid-vowel allophones in the native phonology (though see Weber 1998a), should be written with three- or five-vowel letters, with warring factions of “trivocalists” and “pentevocalists” (see Albó 1987; Hornberger 1995; and Chapter 29 by Rosaleen Howard for some discussion of this particular question, and see Zavala 2019b for the war metaphor). This is not just moot academic debate nor a matter of idiosyncratic preferences: as mentioned in Section 1.7, motoseo, i.e. interference from the Quechuan and Aymaran three-vowel system in the pronunciation of Spanish, is a common substrate feature in Central Andean Spanish (Chapters 16 and 27 by Luis Andrade Ciudad and Anna-María Escobar respectively in this volume) that is ideologically evaluated very negatively (see Chapter 29 by Rosaleen Howard) and that in fact can have significant practical consequences for social and economic opportunities in the context of the national states of the Central Andes (Adelaar 2014c). The decision is thus

29

matthias urban highly politically charged, and may have significant practical consequences. A further complicating factor, at least in the past, has been a lack of coordination within countries and especially between the principally relevant countries, Peru and Bolivia, even though for Aymara, there now exists the Alfabeto Unificado that is officially recognized in both countries. In spite of the heterogeneity, there are also some outlines of traditions in practical orthographies. Many Quechuan and Aymaran alphabets, in spite of the significant overlap in phonemic inventories, differ fairly systematically from one another. For instance, in Aymaran orthographies, including the Alfabeto Unificado (see Coler 2014a: 30– 32 for previous alphabets) and also in some Quechua orthographies in Bolivia (Albó 1987), is sometimes used to represent [h]. The advantage is that is, as a consequence, only used as a diacritic for secondary aspiration and, given this unambiguous role of the letter, is not required to be in superscript as is common elsewhere. Also the representation of long vowels differs: whereas for the Central Peruvian Quechuan varieties these are usually indicated by a doubling of letters (e.g. ), in the Alfabeto Unificado, they are indicated by a diaresis as diacritic (e.g. ). A wide range of variation is found not just in practical orthographies but also in academic contexts. While a version of the Americanist Phonetic Alphabet enjoys some popularity, Central Andean orthographic practices are so varied and inconsistent that there are, in addition, several further idiosyncrasies. One “specialty” of the Andean descriptive tradition—in particular the 1975 Alfabeto Básico General del Quechua that is used in the 1976 dictionaries and grammars and that has also been employed beyond these in academic writing—is to represent the retroflex affricate using the idiosyncratic digraph . In the context of this volume, the orthography problem is alleviated somewhat in that the goal is not to create an orthography for immediate use in a language community. Therefore, considerations of practical applicability need not be taken into account, even though some alignment with practical orthographies is desirable. However, the task is nevertheless a daunting one, as an orthography is needed that can be applied consistently not just to a single language or language family, which has proven to be difficult enough in the Central Andes, but actually across several different language families with different phonological structures and traditions of representation. In the end, I have settled on an orthography that takes over some easily interpretable conventions from Spanish

30

orthographies, which are already employed with high frequency in practical orthographies for Indigenous languages of the Central Andes and also in academic texts. This is combined with the practice of using an acute as a consistent diacritic for retroflex versions of consonants, as in practical orthographies, e.g. for Huanca Quechua and for Chipaya. In this solution, stands for a alveolar fricative, for its voiced counterpart, stands for a palatal fricative, for its voiced counterpart, for an alveolopalatal affricate, for its voiced counterpart, for a voiced palatal lateral approximant, and for a palatal nasal. In addition to an acute as a diacritic for retroflexion, an is used as a modifier for aspiration, an apostrophe for glottalization, and a superscript as a diacritic for labialization. For Jaqaru, the only language where such sounds figure, is used for an alveolo-palatal stop. Other conventions are fairly self-explanatory and oriented on IPA standards (e.g. the use of for the velar nasal). An overview of orthographic representations used in this volume across chapters and Central Andean languages is in Table 1.2. For ease of reference, IPA representations, description of articulatory properties, and representations of the same (or highly similar sounds) in other orthographic traditions—in particular the also widely used Americanist Phonetic Alphabet—and major descriptions of Central Andean languages covered in this volume are also provided. Data from Central Andean languages cited from primary descriptions in the chapters of the present volume are normalized to match the orthographic standard described above where authors felt this was unambiguously possible. Comparative data from surrounding languages to the north, south, and east in the comparative chapters in Part II are not normalized in this way. As Sections 1.2 and 1.4 make clear, pre-modern data play an unusually large role in Central Andean linguistics. Generally, such data are presented in orthographic , following widespread practice in Central Andean linguistics and beyond. These orthographic chevrons are a constant reminder that the data thus represented should not be interpreted phonetically or phonologically lightheartedly without accompanying analysis. Such an analysis, and a standardized orthography, is offered by authors of descriptive chapters of languages that are principally based on premodern sources—Mochica, Híbito-Cholón, and Puquina—in a separate line of representation, where morphological segmentation of the data also occurs. This standardized and analyzed representation is given in {curly brackets}, which are traditionally used for morphological units. Thus, an example sentence from Cholón, taken from de la Mata’s (2007[1748])

Voiceless aspirated bilabial stop

Voiceless ejective bilabial stop

Voicedless bilabial stop

Voiceless alveolar stop

Voiceless aspirated alveolar stop

Voiceless ejective alveolar stop

Voiceless labialized alveolar stop

Voiced alveolar stop

Voiceless alveolar affricate

Voiceless aspirated alveolar affricate

ph

p’

b

t

th

t’

tw

d

ts

tsh tsh

ts

d

tw

t’

th

t

b

p’

ph

p

Voiceless bilabial stop

p

ch

c

d

tw

t’

th

t

b

p’

ph

p

This Americanist volume phonetic notation as applied in the Central Andean literature (e.g. Adelaar with Muysken 2004)

IPA Description

-

ts

d

-

t’

th

t

b

p’

ph

p

1975 Alfabeto Básico General del Quechua

-

-

-

-

t’

th

t

-

p’

ph

p

1985 Alfabeto Unificado Oficial

-

-

d

-

-

-

t

b

-

-

p

Yauyos Quechua (Shimelman 2017)

-

-

d

-

-

-

t

b

-

-

p

tz”

tz

-

-

t’

P”

t

-

p’

P”

p

Chachapoyas Jaqaru Quechua (Hard(Taylor man) 1999)

-

ts

ts

-

d

-

-

-

t

-

-

-

p

-

-

-

-

t

-

-

-

p

-

ts

-

tw

t’

th

t

-

p’

ph

p

Cholón Mohica Uru (Alexander-(Cerrón- (Hannß Bakkerus Palomino 2008) 2005) 1995)

Continued

tsh

ts

t’

th

t

p’

ph

p

Chipaya (CerrónPalomino 2006)

Table 1.2 Orthographic conventions in this volume with IPA representations, descriptions of articulatory properties, and comparisons with other sources

Voiceless aspirated alveolo-palatal stop

Voiceless Glottalized alveolo-palatal stop

Voiceless postalveolar affricate

Voiceless chh aspirated palatal affricate

Voiceless ch’ Glottalized palatal affricate

Voiced palatal affricate

Voiceless retroflex affricate

Voiceless aspirated retroflex affricate

t̠ ʲ’

tʃ

tʃh

tʃ’

dʒ

ʈʂ

ʈʂh ćhh

ćh

zh

ch

ȶ’

ȶh

č̣h

-

tr

ll

dž č̣

ch’

chh

č ’

č h

ch

-

ty’

č

-

tyh

-

ty

t̠ ʲh

ȶ

Voiceless alveolo-palatal stop

t̠ ʲ

-

ts’

Voiceless ejective alveolar affricate

ts’

1975 Alfabeto Básico General del Quechua

c’

This Americanist volume phonetic notation as applied in the Central Andean literature (e.g. Adelaar with Muysken 2004)

IPA Description

Table 1.2 Continued

-

-

ll

ch’

chh

ch

-

-

-

-

1985 Alfabeto Unificado Oficial

-

tr

ll

-

-

ch

-

-

-

-

Yauyos Quechua (Shimelman 2017)

-

ĉ

gh

-

-

ch

-

-

-

-

-

-

ll

ch’

ch”

ch

tx’

tx”

tx

tz’

Chachapoyas Jaqaru Quechua (Hard(Taylor man) 1999)

-

-

ly

-

-

č

-

-

-

-

-

-

-

-

-

č

-

-

ty

-

-

-

-

-

-

tʃ

-

-

-

-

Cholón Mohica Uru (Alexander-(Cerrón- (Hannß Bakkerus Palomino 2008) 2005) 1995)

trh

tr

ch

ts’

Chipaya (CerrónPalomino 2006)

Voiceless ejective retroflex affricate

Voiceless retroflex stop

Voiceless retroflex aspirated stop

Voiceless retroflex ejective stop

Voiceless retroflex fricative

Voiced retroflex fricative

Voiceless velar stop

Voiceless aspirated velar stop

Voiceless ejective velar stop

Voiceless labialized velar stop

Voiced velar stop

Voiceless uvular stop

Voiceless aspirated stop

Voiceless ejective stop

ʈʂ’

ʈ

ʈh

ʈ’

ʂ

ʐ

k

kh

k’

kw

g

q

qh

q’

ṭh

ṭ’

th́

t’́

q’

qh

q

g

kw

k’

kh

k

ź

q’

qh

q

g

q’

qh

q

g

-

k’

k’

kw

kh

k

-

-

-

-

-

kh

k

ṣ̌

ṭ

t́

ś

č ’

ćh’

q’

qh

q

-

-

k’

kh

k

-

-

-

-

-

-

-

q

g

-

-

-

k

-

-

-

-

-

-

-

-

g

-

-

-

k

-

-

-

-

-

q’

q”

q

-

-

k’

k”

k

cx’

cx”

cx

-

-

-

-

-

-

-

-

k

-

-

-

-

-

-

-

-

-

-

-

-

k

-

-

-

-

-

q’

qh

q

-

kw

k’

kh

k

-

-

-

-

-

Continued

q’

qh

q

kw

k’

kh

k

sˆ, zh

tr’

Voiced alveolar fricative

Voiceless interdental fricative fricative

Voiceless ç palatal fricative

Voiceless palatal lateralized fricative

Voiceless alveolopalatal fricative

Voiced alveolopalatal fricative

Voiceless velar fricative

Voiceless labialized velar fricative

z

ʂ

ç

çl

ʃ

ʒ

x

xw -

x

zh

sh

ʂ

z

s

Voiceless alveolar fricative

s

-

Voiceless labialized uvular stop

xw

x

-

š

θ

z

s

qw

This Americanist volume phonetic notation as applied in the Central Andean literature (e.g. Adelaar with Muysken 2004)

qw

IPA Description

Table 1.2 Continued

-

-

sh

-

-

s

1975 Alfabeto Básico General del Quechua

j

-

-

-

-

s

1985 Alfabeto Unificado Oficial

x

-

sh

-

-

s

Yauyos Quechua (Shimelman 2017)

x

zh

sh

-

-

s

j

-

sh

-

-

s

Chachapoyas Jaqaru Quechua (Hard(Taylor man) 1999)

h

-

š

-

-

s

-

-

š

çl

-

-

s

̃

x

-

-

-

-

-

s

Cholón Mohica Uru (Alexander-(Cerrón- (Hannß Bakkerus Palomino 2008) 2005) 1995)

hw

h

ʂ

z

s

qw

Chipaya (CerrónPalomino 2006)

Voiceless uvular fricative

Voiceless labialized uvular fricative

Bilabial nasal

Alveolar nasal

Palatal nasal

Velar nasal

Alveolar lateral approximant

Alveolar lateral fricative

Palatal lateral approximant

Bilabial approximant

Alveolo-palatal approximant

Tap/Flap

Trill

Voiced bilabial fricative

Voiceless labiodental fricative

Glottal fricative

Glottal Stop

χ

χw

m

n

ɲ

ŋ

l

ɬ

ʎ

β̞

y

ɾ

r

β

f

h

ʔ

'

h

f

β

rr

r

y

w

ll

ɬ

l

ŋ

ñ

n

m

-

χ

?

h

-

rr

r

y

h

-

rr

r

y

w

ll

ly w

-

l

ɬ

l

ŋ

ŋ

ñ

n

n

y

m

-

n

m

x̣w

x̣

-

-

-

r

y

w

ll

-

l

-

ñ

n

m

x

-

-

rr

r

y

w

ll

-

l

-

ñ

n

m

-

-

v

r

y

w

lh

-

l

ŋ

ñ

n, n’

m

-

-

-

r

y

w

ll

-

l

nh

ñ

n

m

-

h

-

-

-

y

w

ly

l

ŋ

n

y

n

m

-

-

f

-

r̃

r

y

-

l̃

l

-

ñ

n

m

-

h

-

.

-

r

y

-

ʎ

-

l

-

-

n

m

-

?

r

y

w

ll

lj

l

n

ñ

n

m

jw

j

matthias urban grammar and analyzed by Astrid-Alexander Bakkerus in Chapter 13, looks as follows: (1) < … mellti, milaxi teputtam macjai cullha man miestege axman mige pallow cotan> {me-llt-iy mi-lash-iy 2sg.masc.sbj-be.weak-pst 2sg.masc.sbj-be.lame-pst te putam makhay high village joy

kullha-maŋ life-iness

mi-este-he ashmaŋ 2sg.masc.sbj-enter-sim first

pallow

ø-kot-aŋ} 3sg.sbj-be-ipfv ‘… it is better for you to enter into paradise weak and lame.’ (De la Mata 2007[1748]: 146) This practice, in fact, provides a convenient parallel to practices already in use for Aymara, which are also applied to both Aymara and Jaqaru data in Matt Coler’s contributions on these languages in Chapters 8 and 9 of this volume. Here, a special representation for morphophonologically underlying forms is required because of the rampant vowel deletion processes in the Aymaran family which obscure morpheme boundaries on the surface. Also here, the morphophonological representation is given between {curly brackets} in a separate line below the surface representation, which is given in italics as is usual for object-language data. Thus, an example from Coler’s Aymara data looks as follows (see also Coler 2014a: 438): (2) Tatalamax aliqampiw kasarasirkatamanx. {tala.la-ma-x(a) aliqa-mpi-w(a) dad-2.poss-top other-com-decl kasara-s(i)-irkataman(a)-x(a)} marry-refl-3.subj.1obj.past.cf.top ‘Your dad should have married you to someone else.’ Given that pre-modern (i.e. non-phonemic) orthographic practice prevailed for some languages well into the 20th century, not all examples even in relatively recent sources can be interpreted unambiguously in phonetic and phonemic terms. Where any doubts linger, such data are placed here between orthographic . This is for instance the case for the Kallawaya data in Chapter 15 by Pieter Muysken, where doubts as to the distinction between velar and uvular stops in the literature remain. Rather than standardizing in such cases, and thereby possibly distorting the data, original representations with their ambiguities are retained. On the other hand, Katja Hannß in Chapter 10 feels sufficiently

36

confident to standardize early 20th-century Uru data consistently, so that, for the benefit of clarity and unambiguous representation, she chooses to standardize Uru data. Rather than forcing decisions that may be inappropriate for the specific situation on each single language, in sum, a pragmatic approach was chosen that normalizes orthographies where this is feasible but refrains from doing so where there is a danger that this may lead to a misrepresentation of the language.

1.11.3 Terminology and morpheme glosses Descriptive sketches and also other chapters in this volume adhere to the glossing practices of the Leipzig Glossing Rules, which have become a widely used quasi-standard in the field. Authors were asked to adopt standard glosses from the Leipzig Glossing Rules where they make predefined suggestions. Authors were, however, free to define additional glosses as they saw fit for the purposes of their chapter. Often, different languages of the Central Andes feature morphemes with comparable functions—not just within language families, where this is to be expected, but because of the broader areal-typological affinities, also across languages. Therefore, where such conspicuous similarities in functions of grammatical morphology can be observed and no standard is predefined by the Leipzig Glossing Rules, some amount of consistency in terminology and glossing is desirable to facilitate comparison across languages. Consider, for instance, the case marker -rayku ~ -layku which is widely shared between languages of the Central Andes. In spite of still fulfilling basically the same function in the individual languages, this is called “causative” or “reason” in Shimelman’s (2017) description of Yauyos Quechua, “sake” by Weber (1989) for Huallaga Quechua, “motive” in Coler’s (2014a) Aymara reference grammar, and “causal” in Cerrón-Palomino’s (2006b) account of Chipaya. Given that we are dealing essentially with the same borrowed morpheme that has retained its function, it is desirable to refer to it using a consistent terminology to avoid confusion. Shimelman’s (2017) grammar of Yauyos Quechua, while also adopting the glossing system and the abbreviations of the Leipzig Glossing Rules, already defined a large set of terminological conventions and abbreviations beyond the Leipzig Glossing Rules for the specific purpose of describing Yauyos Quechua. This pioneering role, together with the fact that Shimelman’s sketch for Yauyos Quechua was among the first chapter to be submitted as this volume was being prepared, led to the wide adoption of conventions from Shimelman (2017) for other chapters where this was appropriate. This

introduction means that terminology in the descriptive chapters sometimes differs from that found in the published work of their authors. Care was taken, however, not to overdo standardization in terminology and abbreviations to avoid other languages being forced into the procrustean bed of (Yauyos) Quechua, and authors were free to choose abbreviations of

their own preferences where they deemed this more appropriate instead (and, of course, where neither the Leipzig Glossing Rules nor Shimelman 2017 provided usable precedents for the grammatical category to be glossed). A list of abbreviations used in this volume can be found in the front matter.

37

chapter 2

Physical geography and cultural trajectory of the Central Andes Peter Kaulicke

2.1 Introduction The Andes are a continuous mountain range that forms a kind of backbone of the western South American continent. With a total length of about 7,000km, they are the longest such range in the world. Their northern part adopts an almost vertical north–south orientation between 18° and 20° south. From there further south, the Andes widen to up to 700km on the east–west axis between the modern states of Peru, Bolivia, Chile, and Argentina. The term “Central Andes” refers to the coast and highlands of Peru, and often (for the purposes of this volume) also includes the highlands of Bolivia (Bennett 1948; Lumbreras 1981). Due to multiple factors such as altitude (from sea level to almost 7,000masl), distance to the equator, the cold Humboldt Current, and differing temperature and precipitation patterns, the Andes’ biodiversity is outstanding, featuring more than 80% of the total life zones of the globe. This diversity offers varying opportunities for and challenges to human populations. The modern capitals of Colombia, Ecuador, Bolivia, and Chile all lie in Andean highland valleys or, in the case of Lima, Peru, at the foot of the mountain range. It is obvious that environmental complexity, social complexity, and population density are interrelated factors, which grant the Central Andes a pivotal role in South American prehistory. Human presence in the Andes probably goes back to more than 15,000 years. In the following millennia, complexity came into existence in the sense that Late Pleistocene foraging groups transformed into complex societies culminating in the late Inca Empire. This remarkable path should in principle be comparable to other regions in the world, but in the case of the Andes, interpretable writing systems were absent, an obstacle that complicates the reconstruction of past societies and their languages. This negative precondition has led to much speculation about the degrees of social complexity, affected by modern nationalism or neo-evolutionism. Thus, terms like “state,”

“chiefdom,” “elite,” “migration,” and others are often uncritically applied. Degrees of social inequality usually are measured by the scale of monumental architecture and the presence of outstanding burials and complex art, as well as by site hierarchies. However, there are also more cautious hypotheses based on regional perspectives. In the context of these, the assessment of settlement patterns should be based on complete surveys and excavations in order to be interpreted in line with sound historical approaches. The enormous task to reach this goal, however, has not led to unequivocal results due to incomplete data. Assumed functions of archeological sites, therefore, have often led to hypothetical typological definitions of “cities” vs “ceremonial centers, temples, or palaces,” “lords or kings” vs. “commoners,” or the postulation of theocratic or militaristic empires as opposed to the assumption of lesser complexities in the case of their presumed absence. In this opening background chapter, regional perspectives on the basis of the extant evidence are preferred to such hypothetical theorizing. These regional perspectives necessarily consider environmental conditions taking into account small-scale ecological formations. These form geographically limited mosaics, and were linked by geographical corridors formed by rivers and other communication routes in other environments at different elevations. Mapping the interrelations of flora and fauna and other factors precisely is complicated by the severe deterioration of potential vegetation, particularly during the last halfcentury (Reynel et al. 2013: 96). Earlier intents of ecological subdivisions can be traced back to the 16th century. Pulgar Vidal’s (2014[1938]) distinction of eight main natural regions still is the most popular Peruvian elaboration. Brack Egg (1986) registered nine terrestrial and two marine ecoregions; these are also often referred to (Brack Egg and Mendiola Vargas 2010). The most up-to-date and precise map based on satellite images (see Reynel et al. 2013: 98–100: fig. 6) distinguishes 21 terrestrial ecoregions that can be assigned to ten major regions:

Peter Kaulicke, Physical geography and cultural trajectory of the Central Andes. In: The Oxford Guide to the Languages of the Central Andes. Edited by: Matthias Urban, Oxford University Press. © Peter Kaulicke (2024). DOI: 10.1093/oso/9780198849926.003.0002

38

physical geography and cultural trajectory (i) (ii) (iii) (iv) (v) (vi) (vii) (viii) (ix) (x)

Pacific desert with fog oasis and riparian formations Dry equatorial and mangrove forests Tropical Pacific forest Steppe with tropical seasonally dry forest, relics of subxerophilic forests of the western flank, and northwestern montane cloud forests High-altitude grasslands (puna) or tundra Páramo Humid forests, with montane cloud forests, premontane forests on the eastern flanks Amazonian forests with six subdivisions Savannah Oceanic ecoregions

The map in Figure 2.1 (see plate insert), in which most of these are shown, allows us to recognize (i) a rather complicated pattern in northwest Peru with extensions of different kinds of forests, (ii) a simpler central area of narrow vegetation belts, and (iii) a southern, more extensive highland part with puna formations and inter-Andean river basins. This general picture is adopted in the following discussion, enriched by data on fauna, flora, hydrology, precipitation, and climate, as well as information on their major changes from the Pleistocene to the 16th century. The conditions of ecological fragmentation generally favor small-scale polities, demand economic complementarity, necessitate the enlargement of cultivable lands through irrigation or terracing, and foster technologies related to improved intensified fishing, cultivation of plants, and pastoralism. They also incite negotiations between populations in communal feasting, facilitated by often impressive nondomestic architecture, as well as the intention to obtain commodities and captives through raids or other coercive means. The mentioned environments are cultural products and understood as landscapes, i.e. they are not passive conditioning factors. Climatic impacts of different scales, rather often affecting the entire Central Andes, can cause serious demographic losses including epidemics, severe negative effects on economic strategies, and the collapse of polities. They often are important factors in cultural and social change (Sandweiss and Quilter 2009; Brooke 2014). Fluctuations between local, regional, and interregional polities are contingent on specific settings, environmental and political changes in time, and their interrelations. With this in mind, pan-Andean phenomena known as “Horizons” probably are less cyclical and homogeneous than presumed, as will be shown later. Ceramics are the basic tool for stylistic chronologies because of their frequency in the archeological record. However, fine decorated wares are preferred in a culturehistory approach. This means that economic and environmental aspects—and hence the interrelations with more common subsistence and economic patterns, as well as

production, distribution, and consumption—are only approximately known (Kaulicke 2019). Contextualized calibrated 14 C dating allows to establish precise chronologies and time scales. Generally, for the Central Andes, the time-scale includes the following periods: the Archaic (or Lithic Periods) from the beginning of human colonization (c.13000 to about 1700 bc), the Formative, also known as Initial Period and Early Horizon (c.1700–200 bc), the Early Intermediate Period or Period of Regional Developments (c.200 bc–ad 600), the Middle Horizon (or Huari Empire) (c.600–1000 ad), the Late Intermediate Period (or Chiefdoms and Confederations, c.1000–1400 ad), and a short Late Horizon associated with the Inca Empire (c.1400–1532 ad), which is the latest Pre-Hispanic period before European invasion (Rowe 1960; Lanning 1967; Lumbreras 1981). These subdivisions are slightly different for Bolivian Altiplano archeology, which distinguishes a Late Archaic Period (c.5000–1500 bc), an Early Formative (c.1500–800 bc), a Middle Formative (c.800–100 bc), a Late Formative (c.100 bc–ad 500), Tiahuanaco 1 and 2 (c.ad 500–ad 1100), Pacajes (c.ad 1000–1400), and Inca (c.1400–1532 ad) (see Janusek 2008: 18–24: fig. 7.1). Urbanism and splendid burials are major topics in discussions about the presence or absence of states. The Inca Empire undoubtedly was an expansive state, a characterization strongly supported by colonial written sources and archeological evidence. Discussions of the precursor Huari polity of the Middle Horizon, for which written documents are not available, widely accept that it was an expansive state, too, but with differences and the coexistence with another expansive state, Tiahuanaco. Statehood is also claimed for lesser cultural (or political) formations, but often lacks precise definitions or convincing material evidence, as will be shown. In the following, the discussion of the Central Andes focuses on (i) modern and past environmental characteristics and (ii) the history of populations from the beginnings (Late Pleistocene) to the Inca Empire, which held sway over the Central Andes at the point of European contact. In order to discuss these topics in an interrelated way without the necessity for a large amount of cross-references within the chapter, a subdivision of the Central Andes in three major areas—North, Central and South—is made, and each of these portions is treated and discussed separately.

2.2 Northern perspectives 2.2.1 The Northern environment The coast between the Piura and the Casma river valleys covers a total length of about 400km (see Figure 2.2). Its width

39

peter kaulicke

Tumbes

4

ipe Chinch

Chira

Chulucanas Vicús 12

Piura Piura

Jaén Chamaya

Bagua Utc uba mb a

La Leche R.

Reque R.

on rañ

Ma

Chachapoyas

15 15 Lambayeque 13 3 9

Zaña R.

1 Chicama R.

Pucallpa

Santa. R Nepeña R.

Huaraz

Recuay

10

na

2 14

Mos

Casma R.

Ma rañ on

s ayla

a

250 km

Huaura R.

40

Lake junin

on rañ Ma

Fortaleza R. Pativilca R. Supe R.

Figure 2.2 The northern part of the Central Andes with sites mentioned in the text. Map by Karla Petroni.

Huánuco

Huarmey R.

ag all Hu

125

a ag all Hu

6

N 0

8 Trujillo 16 11 7

Moche R.

e Hu

Huaca Prieta Cerro Sechín Huaca Ventarrón Palanda - La Florida Pacopampa La Pampa Alto de las Guitarras Caballo Muerto Kuntur Wasi Chavín de Huántar Huacas de Moche Loma Negra Sipán El Purgatorio Batán Grande Chan Chan Huancabamba Depression Huascarán Mountain Modern towns

d ejón Call

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Cajamarca

Jequetepeque R.

physical geography and cultural trajectory varies from c.100km in Piura in the north, making it the most extensive coastal stretch in Peru, via c.20–30km in the Lambayeque hydrological basin, to Casma, where the cordillera almost touches the sea. Another important basin is the Chicama valley, which is connected to the smaller Moche basin. Further to the south, the Santa river flows in the longest valley of the North Central Andes, the elongated interandean Callejón de Huaylas, before turning to the sea. The Nepeña and Casma rivers originate in the Cordillera Negra, which flanks the Callejón de Huaylas. Continuous alluvial fans and dry rivers between Piura and Chicama complete this complex hydrology (Delavaud 1968). Rivers range in length between 100km and more than 300km. Apart from the mentioned Callejón de Huaylas, which has a length of 174km, another important topographical feature of the North Central Andes is the Huancabamba depression, an approximately 200kmlong corridor which crosses the Andes between Lambayeque and the Jaén-Bagua area (Reynel et al. 2013: 175: figs 8, 15, 16; Kaulicke 2020: figs 2.4.1–3). It is formed jointly by the Huancabamba and Camaya river basins; in the east, these rivers join the Utcubamba and the Maranón, where they produce extensive floodplains. Many other smaller fluvial plains are scattered over the entire zone; the largest one is the Cajamarca valley with a length of c.40km. In the corridor of the Huancabamba depression the Andean mountains reach their lowest elevation (at Porculla, 2,145masl), whereas Peru’s highest mountain peak, the Huascarán of the Cordillera Blanca to the east of the Callejón de Huaylas, rises to 6,757masl. Temperature and precipitation vary according to season, altitude, and proximity to the ocean or the Amazonian lowlands respectively. The climate is generally moderately warm, dry, and subtropical. Near the coast, at Chulucanas, temperatures range between about 15° C to about 40° C with extremely low precipitation, causing the rivers to dry seasonally. Highland Cajamarca (2,750masl) has a maximum temperature of about 20° C. Precipitation is lowest between June and August; its annual average is 795mm. Finally, Bagua (420masl), on the eastern slopes of the Andes, is very humid with permanent rainfalls and an annual average precipitation of almost 4,000mm. The maximum annual temperature oscillates between 20° C and more than 30° C. These general temperature and precipitation patterns are drastically altered by interruptions called ENSO (El Niño Southern Oscillations). ENSO has a warming phase called El Niño and a cooling phase called La Niña (Glantz 2002); both are global phenomena (Brooke 2014). More regional ones are known as Coastal Niños, and occur at intervals of three to five years; Mega-Niños or exceedingly strong ENSO events are less regular and more exceptional. The ENSO of 1982/83 registered rainfalls of more than 20 times the normal precipitation, inundating large parts of the Piura

coast (Novoa Goicochea 1998: 32: map 5). These impacts are strongest in the Tumbes, Piura, Lambayeque, and La Libertad departments, but reach the Central Coast and affect the South as droughts. The Piura region is a transitional mosaic of ecological landscapes (More et al. 2014), including tropical oceans, temperate oceans, islands, mangrove relicts, wetlands, deserts, dry scrubland, dry plains forests (with much of what is called the Sechura desert), and other six types of dry forests at different altitudes in the Huancabamba depression and in the Bagua floodplain. These dry equatorial forests, found between 0 and 2,800masl, are important formations from Tumbes to Lambayeque. Their fauna is rich in endemic species; in Piura and Tumbes, primates such as the mantled howler monkey (Alouatta palliata) and the white-fronted capuchin (Cebus albifrons), jaguar and other felines, peccaries (Pecari tayacu), and tayra (Eira barbara) (Brack Egg and Mendiola Vargas 2010; Reynel et al. 2013; More et al. 2014) lend the fauna a distinct Amazonian flair. Northwest cloud montane forests form a belt between 1,500 and 3,500masl from Piura to Cajamarca (Reynel et al. 2013; More et al. 2014). They grow on both flanks of the Andes. Palms, orchids, lichens, as well as endemic trees provide the habitat for the mountain tapir (Tapirus pinchaque), the spectacled bear (Tremarctos ornatus), the aforementioned primates, sloths (Choloeus hoffmanni), paca (Cuniculus paca), and many endemic birds (Brack Egg and Mendiola Vargas 2010; Reynel et al. 2013; More et al. 2014; Ministerio del Ambiente 2014). Páramo and puna are ecological formations above 3,200masl. The páramo is a northern type of ecoregion which is more typical for Colombia and Ecuador (Reynel et al. 2013: 109). It is more humid than the puna and features small rivers and lagoons. Scrubs, grasses, and orchids grow there, and there are relics of Podocarpus glomeratus forests (Sabogal 2014). The tapir (Tapirus pichanque) and the Páramo deer (Mazama rufina) are typical animals (Reynel et al. 2013: 109). The puna is more important for the ecoregions of the Central and Southern parts of the Central Andes (see Sections 2.3.1. and 2.4.2). The eastern Andes are characterized by eastern montane cloud forest between 3,500 and 2,000masl and the high jungle, the lower limit of which is at an altitude of 800masl. The flora of this region is extremely varied; the spectacled bear, puma (Puma concolor), and an endemic primate, the yellow-tailed wooly monkey (Oreonax flavicauda), live there (Reynel et al. 2013; Ministerio del Ambiente 2014). Gradually these ecoregions change as one moves south. At the latitude of Trujillo, the Humboldt Current and the increasing distance to the equator lead to drier coastal conditions which are shared with the Central area, while the described western Andean forests continue (see Figure 2.1). The dry forests of the Bagua floodplains follow the course

41

peter kaulicke of the Marañón river basin to the south up to the southern limit of the Callejón de Huaylas. Finally, the sea is of utmost importance in both its positive and negative aspects. It produces enormously rich marine resources, but it also brings disasters such as inundations, tsunamis, hurricanes, droughts, and earthquakes, partly as effects of ENSOs. The northern part of the ocean is affected by several currents, such as the Southern Ecuadorian Current, the Submarine Ecuadorian or Cromwell Current, and the Ecuadorian-Peruvian Countercurrent (Espino 2003). These produce variations in sea temperature, but generally warm waters which host about 70% of Peru’s littoral biodiversity; Northern Peruvian waters, however, are less productive than the cold sea to the south. There are more than 140 fish species, lobsters, oysters, and crustaceans. Spondylus and mother-of-pearl are important mollusks. Seabirds are numerous and humpback whales (Megaptera novaeanglia) arrive for reproduction (Schweigger 1964; Brack Egg and Mendiola Vargas 2010; More et al. 2014). An important topographical aspect of the North is that it affords relatively easy intercommunication. The courses of rivers, mountain passes, and the sea facilitate animal and human movements in different directions. The Chira river basin in the north connects with the Ecuadorian highlands, and the upper Piura basin with Lambayeque and the adjoining highlands via its numerous tributaries. The Lambayeque hydrological system gives access to the adjacent highlands and the Huancabamba depression to the eastern Andes, while the Jequetepeque river basin leads to the Cajamarca valley. The Callejón de Huaylas with the Santa valley, which is similar in shape to the smaller Piura, is connected with the coast by short river basins.

2.2.2 Population and cultural trajectories The earliest evidence for human occupation, dated to almost 15,000 radiocarbon years, comes from Huaca Prieta at the mouth of the Chicama Valley (Dillehay 2017b). Different early lithic industries might relate to different migration patterns and time-scales. The shorelines oscillated significantly from up to 35km away from the modern beach during the Late Pleistocene to inland ridges during the Early Holocene. This must have affected these early occupations. ENSO regimes existed between 11000 and 6800 bc, but are rarer thereafter (Sandweiss and Quilter 2009). Northern glaciers covered the high mountains in Pleistocene times, and the vegetation belts varied in extension according to differences in annual temperature and precipitation (Netherly 2011). The early inhabitants were not specialized megafauna hunters, as once believed, but

42

exploited a wide spectrum of resources from the sea as well as the adjacent highlands, and lived in clusters of often-reoccupied huts in what Dillehay (2011) calls “protohouseholds” for the period between 9200 and 7800 bc. In the Upper Zaña valley, between 7800 and 5800 bc, more formalized clusters of huts with small garden plots indicate an early “Neolithic” life style that involved early sedentism and cultigens. The cultivated plants suggest a connection with the eastern Andes, where their wild ancestors grow, probably via the Huancabamba depression. Mollusks from the sea and other resources from dry and humid forests are evidence for a wide range of interactions. Former domestic circular structures changed into rectangular buildings, and non-domestic mounds appeared about 5,700 bc in the Upper Zaña Valley and at Huaca Prieta in the Chicama Valley, where the main structure remained in continuous use up to Formative times as a ritual and mortuary mound. Apart from Huaca Prieta, early more formalized architecture emerged in the Casma valley at c.3,500 bc (Fuchs et al. 2010). At c.2200 bc a complex in the same valley, Cerro Sechín, with polychrome murals and characteristic internal features, constitutes an early representative example of what would become a long-standing tradition (Fuchs and Patzschke 2013). Cerro Sechín apparently forms part of an early core area in the middle Casma valley. Huaca Ventarrón in the northern Lambayeque region is a large complex of superimposed buildings with painted murals and reliefs that was constructed between c.2400 and c.1900 bc (Alva Meneses 2013; 2014). Its builders were probably in contact with the southern valleys and the Bagua region in the eastern Andes via the Huancabamba depression. Still undated architecture in the Bagua floodplains is similar to dated structures in the Upper Chinchipe region in modern Ecuador at Palanda (Olivera Nuñez 2014; Valdez 2014). These still incomplete data hint at long and steady processes leading to very early complexity as well as wide-ranging interactions in this time-range within the North of the Central Andes. Simple ceramics generally appeared at c.1500 bc, but they are slightly earlier in the Casma valley, where they are associated with huge architectural complexes. Dated to c.1700–1400 bc, they are adorned with impressive murals and sculptures. These murals, decorated lithic mortars, and bone spatulas retain earlier figurative elements that appeared between c.2200 and 1750 bc (Pozorski and Pozorski 1987; 2012; Lerner et al. 1995; Bischof 2009). These architectural complexes may reflect a large ceremonial city that was connected to other littoral settlements in the mid-Sechín river basin, but whose wider interaction sphere reached the Jequetepeque valley, the North-Central Coast, and probably the Huánuco region to the east, thus covering an area of more than 400km in length on the coast and more than

physical geography and cultural trajectory 200km in the highlands. The Early Formative in the North boasts other interrelated sets of sites, sometimes around large settlements, but the Casma area is without peers. Scattered non-domestic sites in the highlands, probably occupied between 1400 and 1200 bc, show decorated stone lintels and other architectural features in similar renderings between Pacopampa in Cajamarca and La Pampa in the Callejón de Huaylas (Bischof 2008; Morales Chocano 2008). Rock art at Alto de las Guitarras, above the Moche river basin, and several sites in the Jequetepeque valley basin (Pimentel 1986; Campana 2013) combine earlier figurative representations with other features later known as Chavín. During the Middle Formative (c.1200–800 bc), many of the coastal stretches between Casma and Lambayeque and adjacent headwaters from Piura to Lambayeque feature exuberantly decorated platform mounds with plazas of different scales. Earlier interaction networks become denser, though probably not in the sense of territorially unified polities, but more in the shape of perhaps segmentary states and/or other less complex political formations. The Chicama-Moche zone became the new power center, where Caballo Muerto seems to be the largest settlement of the Cupisnique culture (Pozorski 1975). It is around 1200 bc that Chavín de Huántar became a small sanctuary at the very margin of the North as defined here, to the east of the Callejón de Huaylas. It grew to an impressive center between 1000 and 800 bc, situated at a junction of connections between Casma, Huánuco, the North Coast (Jequetepeque to Chicama-Moche valleys), and the Central Coast (Lurín). These connections led to the assumption that Chavín was a center dominating other coeval sites during the Late Formative or Early Horizon (c.800– 600 bc), but it certainly was not a territorial state. Instead, Chavín was an interregional sanctuary. The overwhelming presence of material from the North and Central Coast, as well as minor other regions, hint at its attractiveness for foreign elites, some of who probably were responsible for the creation and maintenance of the center. Its impressive installations, including complex subterranean galleries and canals, and its impressively decorated architecture are evidence of sophisticated dramaturgies for rituals, sacrifices, processions, and large-scale consumption of food and beverages. Many of these activities probably were part of a water origin cult personified in a local deity that was depicted in an in situ cult image known as Lanzón and that was venerated until the definite demise of Chavín at c.500 bc (Burger 1984; 1992; 2008; Rodriguez Kembel 2008; Rick 2008; Lumbreras 2013; Rodriguez Kembel and Haas 2015). On the North Coast, in the time-frame between 800– 600/500 bc, valleys between the Jequetepeque and Piura and their headwaters boast rich burials at large centers or

ritually relevant places. Their size and complexity, as well as large funerary areas, hint at high population densities and ranked societies. Established and interrelated elites wore paraphernalia in gold and semiprecious stone (“crowns,” earspools, nose ornaments, pectorals) which are well known from later societies. The Bagua-Jaén center flourished with similar burial contexts, produced polychrome ceramics with emulated Cupisnique designs and forms widely distributed toward the Pacific coast and Southern Ecuador. In general, art expressions in metal, ceramics, lithic statuary, and monumental rock paintings share a common style of wide distribution, the Late Formative style (Alva 1986; 1992; Olivera Nuñez 1998; Onuki and Inokuchi 2011). At c.600 bc, the Late Formative gave way to more diversified developments in more or less well-defined areas. This is interpreted as the result of the breakdown of a “pax chavinensis,” caused by the collapse of a presumed pan-Andean Chavín cult that led to increased violence and political balkanization. The assumed crisis, however, could be understood as a mix of innovations, predominance of utilitarian pottery, and maintenance of earlier decorations and forms in architecture, site organization, and economic strategies. Vessel forms and decoration define interrelated complexes like Vicús, Salinar, or Gallinazo (Brennan 1980; Kaulicke 2006; 2009; Millaire and Morlion 2009; Ikehara and Chicoine 2011; Millaire and Eastaugh 2014). The Vicús style from the upper Piura incorporates some northern “Ecuadorian” elements, including architecture of earth mounds (tolas) with wattleand-daub structures (bahareque). Walled sites (“fortresses”) became more common in the southern part of the North Coast, where variegated architecture of an essentially nonmonumental nature and spatial organization began to develop in the Nepeña and Casma valleys at c.600 bc. In contrast, the centers in the headwaters of rivers in the north, at Pacopampa and Kuntur Wasi, retained much of their former traditions. In the Cajamarca highlands, the Layzón culture expanded to other parts of the highlands and adjacent coastal river basins. Huaraz is another tradition of this time-period in the Callejón de Huaylas (Mujica 1984; Terada and Onuki 1985; Lau 2002; 2011; Kaulicke 2006; Ikehara and Chicoine 2011). This complex network of different styles is related to varying but often relatively high population densities distributed over many small and a few large settlements, perhaps indicating early urbanization. Signs of what may have been ethnic identifications are increasingly in evidence, but it remains difficult to define corresponding polities. There might be a tendency toward more extended chiefdoms or simple state formations with clearer concepts of territoriality than is in evidence for earlier times. This tendency should be related to the expanding and diversifying economic space vis-à-vis the earlier Formative

43

peter kaulicke that involved competition and ensuing small-scale migration, colonization, and/or conquests. During the 1st or 2nd century ad, another cultural tradition emerged from the mentioned styles and subsisted until the 9th century: Moche or Mochica (Benson 1972; Bawden 1996; Pillsbury 2001; Uceda and Mujica 2003; Castillo Butters and Uceda Castillo 2008; Quilter and Castillo 2010). It covers the entire North Coast from Piura to Casma and extends even to Huarmey. Mochica centers boast impressive monumental stepped platform mounds, which result from superimposed architectural structures and are often adorned with spectacular murals. Some of these contain burial chambers with extraordinarily rich contexts for apparently outstanding individuals; less complex burials that are often associated attest to hierarchical societies (Donnan and Castillo Butters 1994; Alva 1998; Alva and Donnan 1993; Kaulicke 2000; Donnan 2007; Mujica Barreda 2007; Bourget 2014; Kaulicke and Kurella 2014). Much as in Late Formative times, each valley between the Piura and Jequetepeque boasts at least one outstanding monumental site, often associated with burials. In spite of intensive research in the last decades and the accumulation of abundant evidence, the nature of these societies remains a matter of debate. While some postulate the existence of one or two large territorial states, paramount chiefdoms, or kingdoms, others see city-states in the north and a territorial state in the south (Kaulicke 1992; 2000; 2006; Quilter and Castillo 2010). Elite material culture, the hallmark of the Mochica, varies in space and time. Art motives (particularly the spectacular pottery) appear to highlight military violence, a fact which is often interpreted as evidence for territorial expansion from the supposed capital, the Huaca del Sol/Huaca de la Luna complex near Trujillo; however, more small-scale conflict with ritual background is now the preferred interpretation. The idealized homogeneity and cultural superiority of Mochica society, however, collides with the nature of the evidence for the relationships with other societies or possible ethnic groups. In the Piura region, objects from Tolita, a culture from the coast between modern Ecuador and Colombia, appear in Mochica funerary contexts in the highlands, while Mochica elements are present in Ecuador. A major center in the Upper Piura region shows different types of monumental architecture—Vicús and Mochica—and extensive burial grounds with mainly Vicús structures containing Vicús and hybrid Vicús–Mochica ceramics, whereas an elite funerary area nearby at Loma Negra produced impressive metal objects in a definite Mochica style; but associated ceramics do not share the same elements as “classic” Mochica wares. This complex picture hints at the coexistence and interactions of various ethnic and political formations on the coast, the Piura highlands, and modern Ecuador (Kaulicke 1992; 2000; 2006).

44

The core area of northern Mochica, between the Piura and Jequetepeque valleys, is the same as occupied by Late Formative societies whose material cultures are similar, so that the origins of the Mochica culture ought to be sought there. The Mochica culture flourished between ad 300 and 600, the time-frame to which many outstanding burials and impressive superimposed platforms in city-like landscapes can be assigned. Most famous is the burial platform at Sipán in the Lambayeque Valley, which forms part of a large urbanized settlement and shows a dynasty-like sequence of various generations that have been buried there (Alva 1998; Alva and Donnan 1993; Chero 2013). The metal objects are clearly Mochica in style, but ceramics share many elements with the Upper Piura zone and the southern Zaña (Bourget 2014) and Jequetepeque valley basins (Donnan 2007). The 60ha Huaca del Sol/Huaca de la Luna complex near modern Trujillo (Uceda and Morales 2010), with dual pyramid mounds and an urban center in between, is often considered the capital of a unified Mochica state, but more probably it is another regional polity allied with the Chicama centers. Valleys to the south seem to have been areas of other similar but probably independent polities (Kaulicke 2000). Another regional center area was located in the Cajamarca valley and the higher parts of the adjacent river valleys. Based on the Layzón culture, it is characterized by an astonishingly long ceramic tradition of white (kaolin) wares from c.50 bc to Inca times (Terada and Onuki 1982; Terada and Matsumoto 1985; Matsumoto 1993; Watanabe 2009; 2019). Data, however, are insufficient to specify the nature of its populations and political structures. While many settlements are of reduced scale, some seem to be administrative or ceremonial centers. In spite of the apparent unity of ceramic art, the evidence probably reflects networks of interactions with properties that are difficult to discern. Recuay is another area with interrelated art styles and associated settlements, centered between the Cordillera Blanca and the Cordillera Negra in the Callejón de Huaylas and extending to the area between the Cordillera Blanca and the Marañón river basin and the upper reaches of the adjacent coastal basins from Moche to Casma (Lau 2011: 245: fig. 68). In several areas, denser agglomerations of sites with large and complex settlements seem to represent regional polities dominated by elites. These were characterized by complex burial structures, elaborate ceramics, metals, and textiles, as well as lithic statuary, tenon heads, and lintels, particularly between ad 300 and 600. In their early phases of development, the Gallinazo, Salinar, Vicús, and Mochica societies were interchanging cultural concepts and economical products with these highland polities. In spite of the emergence of large centers with often imposing ceremonial structures, they do not necessarily represent major political spheres.

physical geography and cultural trajectory The 6th century ad was full of events provoked by major natural impacts; ensuing social upheaval and transformation finally led to the collapse of former polities and the emergence of others. Especially strong Mega-Niños with major floods and long droughts occurred at c.ad 600 and affected the entire North as well as regions of the Central Andes beyond. Intensification of conflicts, increased frequency of sacrifice as a ritual response, and the opening of “frontiers” to movements of things and peoples, partially caused by the southern Huari phenomenon (to be discussed in Section 2.3.2), dominated the political scenario for about three centuries. Populations nucleated in large urban settlements in several valleys during the 6th century, but these were of relatively short duration. Elites in the Jequetepeque valley adopted a diversified material culture, as evidenced by stylistic hybridization, emulation of foreign styles, and actual foreign products in elaborate funerary structures. Earlier characteristic architecture survived, including Mochica material culture, but elements from various foreign cultures appeared at some centers (Castillo 2000; Pillsbury 2001; Chapdelaine 2010; Bracamonte 2015). At about 800 ad, the Huari presence became more widespread, but most polities of the time probably were relatively simple, with only sporadic instances of Huari control. Above-ground burial structures known as chullpas emerged and dominated the cultural landscapes in the Callejón de Huaylas and beyond, sometimes in impressive concentrations, scale, and complexity (Paredes et al. 2001; Tschauner 2003; Topic and Topic 2010). Huari influence in this region was sporadic, too, but strong enough to contribute to the demise of the Recuay culture (Lau 2014). At c.ad 900 new material cultures emerged which were characterized by different ceramics and textile styles whose decorations often were inspired by Mochica and Huari representational codes. A large center, El Purgatorio in the Casma valley, might have been the capital of a Casma state (Vogel 2016), presumably between Chao and Huarmey. Its ceramics, a fusion between earlier Mochica motifs and Huari, was widely distributed between the Jequetepeque valleys, the Central Coast, and the highlands. The Lambayeque or Sicán culture appeared in much of the North Coast from Tumbes to Chicama, but its center area was Lambayeque. The most imposing Sicán center is Batán Grande on the La Leche River in the Lambayeque valley. It is a huge settlement of approximately 2 × 1.5km with a temple core and associated burial grounds for elites. Some of these funerary structures are most impressive because of their enormous amount of valuables, often of exotic origins. Metalworking (mining, smelting, and processing) achieved almost industrial scales, though mostly destined for elite consumptions (Shimada 1995; 2014). Recent surveys defined a large surrounding area, Greater Sicán, with

a population of more than 26,000 inhabitants (Cervantes Quequezana 2020). This society bears the characteristics of a state with hierarchical organization. It extended from the Tumbes to the Moche valleys, and enjoyed wide-ranging direct contacts to regions in modern Ecuador, Colombia, or even Costa Rica, as well as several highland regions up to the Amazonas and the Central Coast (Shimada 1995). Its flourishing, however, was short-lived, from ad 900 to 1150, when the polity was strongly impacted by a Mega-Niño and long droughts. The Sicán culture contributed much to the Chimú in the Moche valley, where a large center, Chanchan, near Trujillo, with a core of c.6 km2 and an urban space of 14 km2 , grew to form the capital of the Chimor “kingdom” or “empire” (Rowe 1948; Ravines 1980; Moseley and Day 1982; Moseley and Cordy-Collins 1990) Chanchan’s impressive funerary platforms were already mentioned in written sources from the 16th century (Rowe 1948; Ravines 1980). These sources mention conquests from Tumbes to the North-Central Coast, essentially backed by archeological evidence. Centers from Paramonga in the Huarmey valley and others to the north in most of the mentioned valleys are particularly impressive by scale, complexity, and exuberance, in spite of Chimor’s brief existence from c.1375 to 1470 ad. The Inca finally conquered the entire region until they succumbed to the European invasion after an even shorter dominion (Cabello Balboa 1951[1586]; Ramirez 2016). The Lambayeque valley system hence experienced a long tradition of more than half a millennium of complex hierarchical societies with numerous often awe-inspiring, highly decorated mound complexes. It is mandatory, however, to consider that this evolution to major complexity began much earlier, at least at c.2400 bc. The imposing monumental platform mounds formed the center of settlements of varying scales, and are often interpreted as “temples” or ceremonial centers instead of towns or cities; however, the surrounding sociocultural landscapes are often less well known or ignored, even when non-monumental remains are visible, so that population estimates often are unrealistic. Monumental architecture is not only a form of social cohesion and power but also a memory place, and multigenerational use is indicated by superimposed buildings and continued patterns and themes of decoration in murals and associated burials. These centers also were nodes of networks, and were often ruled by elites that are recognizable by paraphernalia or power symbols from Formative times to Chimú. This is particularly true for the two main centers, Lambayeque and Chicama-Moche, which both feature outstanding complex irrigation systems (Ortloff et al. 1982; Netherly 1984; Huckleberry et al. 2012; Hayashida 2014). The highlands were less centralized, less hierarchical, and had restricted core areas. Camelids, intensive agriculture on terraces, and contacts with the coast guaranteed the

45

peter kaulicke This sector’s coast is relatively wide, with frequent promontories interrupted by relics of an ancient coastal cordillera. The river basins in the north—Fortaleza, Pativilca, Supe—form a complex that is connected with the northern Huarmey valley, whereas those in the south— Chillón, Rímac, Lurín—form another one that is connected with the southern Chancay river. As a result, the deserts between both are wider than those between their counterparts in the North. All northern rivers originate in a high plain to the south of the Callejón de Huaylas, while the southern complex is connected to large plains around Lake Junín. These river basins are similar in length and extension to those in the North, but differ in comparatively small amounts of water discharge. Lagoons are found in some parts of the coast, but much larger accumulations exist in the high Junín plain which, with a surface area of more than 500 km2 , is the second largest lake in Peru. It is drained by the important Mantaro River, which forms a large plain in the Jauja–Huancayo area. Nearby the Huallaga, a major tributary of the Amazon, rises and forms a plain around the modern town of Huánuco, in the eastern Andes, with almost 1,600 km2 in extension. Other rivers also flow to the east to join the Marañón, e.g. the Mosna River passing Chavín de Huántar. The maximum temperature on the coast at Lima is about 26° C from December to May and between 17° to 19° C

highland elites’ wealth. Intercommunication facilitated by road systems must have been essential, but these are only visible in the later impressive interregional Inca road system; the relative closeness of the coastal river basins, however, surely facilitated local communication. The sea was essential for coastal people, as it not only provided food, but was also was a substantial part of their cosmologies and allowed maritime contacts with other, distant parts of their worlds.

2.3 Central perspectives 2.3.1 Environment The Central sector of the Central Andes (see Figure 2.3) involves an essentially smaller space than its counterpart in the North, if we take into account that the distance between the Huarmey and Lurín rivers, which defines the north–south extension of the Central subdivision proposed here, is only some 250km in length compared to the c.400km which the North accounts for on the north–south axis. As will be shown, this subdivision is justified for its particular environmental and cultural aspects.

Pucallpa Santa R.

13

ylas Hua

5 12

Lake junin

2 1

Tarma

9

Jauja

Huancayo

ro ta an

7

M

N

125

Huánuco

on rañ Ma

11 Lima 8 Lurin R.

6

e ag all Hu

Uchkumachay Huarmey R. Telarmachay Caral Vichama Chawín Fortaleza R. Kotosh Pativilca R. Supe R. 3 Wariwillka 4 Pachacamac Huaura R. Hatun Xauxa Lauri Cajamarquilla Pumpu Chancay R. 10 Chavín de Huántar Huascarán Mountain Chillón R. Modern towns Rimac R.

0

a Mosn

Huaraz

Recuay 1 2 3 4 5 6 7 8 9 10 11 12 13

Ma ra ño n

de ejón Call

Nepeña R. Casma R.

250 Km

Figure 2.3 The central part of the Central Andes with sites mentioned in the text. Map by Karla Petroni.

46

physical geography and cultural trajectory during the rest of the year; the annual mean is 19° C. Thus, the summer months are warm and humid, while the winter months are cooler and misty. Humidity is extremely high during the entire year, but precipitation is almost absent. Differences to the North therefore are evident. These coastal conditions differ also from those in the highlands. At Cerro de Pasco, at 4,338masl, the maximum temperature is between 13 to 14° C, but can drop to –7.6° C; the annual average is 5.9° C. Precipitation is low and varies seasonally. 150 and 185mm per month between December and March drop to 30–50mm during June and August. Finally, Huánuco, at 1,859masl, has a maximum average temperature of 26.4° C and an average annual rainfall of about 379mm per month. These conditions shape the different ecological landscapes. Koepcke (1954: 18: fig. 2) describes 17 life zones from sea to puna. These are quite different from the aforementioned 17 life zones for Piura. Central ecoregions include the ocean, coastal lakes and lagoons, deserts, guano accumulations, lomas, tillandsia stands, riparian forests, columnar cacti formations, communities of small Carica and Jatropha trees, savannah, sporadic forests, grasslands with scrubs, puna grasslands, cushion plant communities, and glaciers with subnival zones. Upwelling of colder ocean water produces enormous quantities of plankton, particularly between Supe and Lima. Zooplankton and phytoplankton are nutrients for a huge biomass of fish. There are some 700 species of fish, 400 species of crustaceans, and sea mammals like whales, dolphins, seals, marine otters, as well as seabirds. Anchovies (Engraulis ringens), with a biomass of over 8 billion tons, are intensively exploited by modern fisheries, but many other species are also of outstanding economic importance. Coastal landscapes influenced by the sea include those that are formed by seasonal fogs (garúa), those that are part of deserts, and those that depend on fresh water. Particularly important are wetlands with lagoons. The abundant avifauna comprises migratory seagulls, ducks, and the osprey (Pandion haliaetus). Wetlands were much more important in the past, but have been much deteriorated or erased. They often coexist with saltpans and salty wetlands (gramadales) (Tovar Serpa 1977). Other formations of importance are the river deltas with lagoons, abundant fish, and waterfowl. More than 90% of the permanent bird species and the plants in the riparian ecosystem are independent of seasonal water, and plants also allow for the presence of deer, foxes, and skunk. Still another type of landscape is found between 300 and 800masl. It consists of fog oases flourishing in the winter months from April to November. These loma formations between Trujillo and Chile are highly differentiated, and their vegetation ranges

from permanent tree stands to xerophitic plant communities. Such loma formations expand significantly under ENSO conditions. Animals which also inhabit higher altitudes, like tinamous (Nothoprocta pentlandi), viscacha (Lagidium peruanum), deer (Hippocamelus antisensis and Odocoileus virginianus), guanaco (Lama guanicoe), puma (Puma concolor), and the mountain cat (Oreailurus jacobitus), seasonally live in these lomas. The steep western flanks between 1,000 and 3,800masl present a climatic inversion. Clouds cover the sky and produce frequent rainfall during the summer months, but remain dry when the coast is cloudy. The coastal desert conditions turn into semi-desert and steppe. Columnar cacti and fruit trees like Carica candicans (wild papaya) and Jatropha macrantha (huarnapu) are common. At higher altitudes, isolated pockets of rather dense forests form the habitat for many birds, rodents, puma, deer, skunk, viscacha, and opossum (Didelphis marsupialis). The western Andean flanks, hence, are dry and relatively poor in vegetation and fauna. The highest ecosystem is the tundra-like puna from 3,800 to 4,600masl, characterized by rolling grasslands, shrubs, wetlands, cushion plants, and “woolly” cacti (Tephrocactus sp.) with edible fruits. There are many lakes and lagoons, with or without reeds, at altitudes between 4,400–4,600masl. Lake Junín is the center of this landscape. Its shore forms belts of grasslands, muds, floating algae, dense totora (cattail) stands, and open water. It is particularly rich in waterfowl like ducks, coots, gallinules, grebes, cormorants, and gulls, which live there in the millions preying on fish. Frogs and toads (Bactrachophrynus spp., Telmatobius macrostomus) and wild guinea pigs (Cavia tschudi umbrata) are also common. Vicuñas (Vicugna vicugna), taruka (Hippocamelus antisiensis), puma, mountain cats, skunk (Conepatus chinga rex), and viscacha are living in the open grasslands. Thus, the Central lakes and grasslands were the home of especially dense biomasses, which in the early 19th century were much more abundant than nowadays and originally also included guanaco. Similar to the western flanks, the eastern side of the Andes and the inter-Andean valleys have riparian vegetation with trees and shrubs at the valley bottom and xerophitic zones and thorn forests further up to the puna rim. This is even true for the Huánuco basin, in spite of its closeness to the humid premontane forests (see Koepcke 1954; Rauh 1958; Koepcke 1961; Dourojeanni et al. 1968; Brack Egg and Mendiola Vargas 2010; Autoridad de Agua 2012). In summary, the Central zone is generally dryer than the North, and more dependent on seasonal changes with two dominant ecoregions: the sea and its shore and the puna. Both of them offer abundant resources but are subject to seasonal changes.

47

peter kaulicke

2.3.2 Population and cultural trajectories Compared to recent results regarding the beginnings of the human presence in the North, the state of research concerning the Central sector remains very much as it was in the 1980s, with relatively sparse and scattered data available. These data were extracted from sites on the coast (Patterson 1966; Lanning 1967; Kaulicke 1980) and in the puna (Cardich 1964; MacNeish 1971; Kaulicke 1980; Rick 1980). On the coast, the situation is similar to the one presented for the North in the sense that lithic industries with stemmed (so-called Paiján) points are known from quarries, workshops, and campments on the entire coastal stretch, but subsistence strategies are not well known (Kaulicke 1980). As most of these sites lie in modern and past lomas, the exploitation of resources depended on what was available seasonally there and nearby on river banks. These conditions led to the hypothesis that highland foragers followed migratory patterns of guanaco herds to the lomas in the winter months (Lanning 1963)—which, however, is no longer accepted. Some evidence for processing plants and the relevance of marine resources such as mollusks and fish hint at a combination of fishing and perhaps even early cultivation so that transhumance is unlikely as an economic option. On the other hand, better information is available for early puna occupations, thanks to the presence of numerous rock shelters or small caves with evidence spanning the entire Archaic from the Terminal Pleistocene to the Formative (Wheeler et al. 1976; Kaulicke 1980; Rick 1980; Lavallée et al. 1985; Rick and Moore 1999). Whether megafauna such as extinct horse and deer were associated with human occupation remains as at Uchkumachay in Junín at 4,050masl is still in need of clarification (Wheeler et al. 1976). The lithic puna industries are different from those on the coast, and were characterized by small projectile points, scrapers, and other implements that hint at a predominance of hunting. This specialization is confirmed by high amounts of camelid bones and, to a lesser extent, deer, viscacha, skunk, birds, felines, and frogs. Deer and camelid hunting in almost equal proportions characterize the early Holocene, but camelids became the preferred prey thereafter. Camelid domestication had been considered late, but evidence available now from Telarmachay, Junín, strongly suggests the presence of domesticated vicuña (alpaca) at c.4000 bc (Wheeler 1985; 1999). A certain disregard for early plant domestication has been proposed for coastal subsistence patterns. Moseley (1975) maintained that the focus on anchovy fishing triggered social complexity, while domestication and agriculture supposedly occurred much later. Early maize, however, appeared in several sites at the same time as animal domestication in the highlands. Probably it was not an essential part of the diet, but grown for

48

making fermented beverages known as chicha. Camelid domestication apparently did not lead to a decline of hunting on the wild species, but probably triggered interaction with other regions. Since the 1990s, these hypotheses have begun to require revisions, as the mentioned Fortaleza–Pativilca–Supe river complex in the north produced surprising evidence. From a still rather hypothetically dated onset at c.3000 bc, monumental architecture was detected in these valleys. By 2600 bc, numerous sites on the entire Central Coast, but particularly in both the Fortaleza–Pativilca–Supe and the Chillón–Rímac–Lurín river complexes, clearly show signs of sedentism with formalized architecture. A nucleation of sites in the middle Supe valley, with a major center called Caral, of 68ha in extension, connected to the upper valley and the delta; still other sites were located in the nearby Fortaleza and Huaura valleys. They flourished for about five centuries, when strong environmental impacts including earthquakes contributed to their demise. The sites share a number of architectural features, including stepped platform mounds with staircases and wattle-and-daub structures on their tops, annexed circular courtyards, and bags with stones as filling material, known as shicras. Minor structures in the same sites might have been domestic or residential units, which led researchers to consider them urban centers of a pristine state (Shady and Leyva 2003; Shady et al. 2015a). Caral supposedly was the capital of this pristine state at the origin of Andean civilization (a hypothesis not accepted unanimously). As has been shown, this core area was not the only one of its kind at this time: the Sechín core in the Casma Valley, discussed in the previous section, differed in architectural features and boasted highly adorned facades and a complex material culture that was almost entirely absent at Caral, but both areas were connected. Stronger connections can be inferred for the southern river complex because of similar but slightly later non-domestic structures, so that some population displacement from the North seems plausible. Sites located in the mid-valleys usually do not feature funerary areas, but on the littoral or in river deltas, such areas are common and relatively large. Complex funerary contexts comparable to those from the North are unknown. As far as the economy is concerned, fishing and foraging of marine resources was an important activity, supplemented by wetland resources where plants were also grown, as in the North at Huaca Prieta. As in the North, too, fruit trees like guava (Psidium guajaba), avocado (Persea americana), pacay (Inga feuillei), lúcuma (Pouteria lucuma), and a plum (Bunchosia armeniaca) probably were planted and used for different purposes apart from consumption; algarrobo (Prosopis juliflora) was another tree of great importance since earlier times. Beans (Phaseolus spp.), arrowroot (Canna edulis),

physical geography and cultural trajectory peanuts (Arachis hypogaea), and chili pepper (Capsicum spp.) were staples; potatoes (Solanum tuberosum), sweet potatoes (Ipomoea batata), and maize (Zea mays) probably were imported (Shady et al. 2015a). This situation is similar to the better-known Huaca Prieta site mentioned earlier. Dillehay (2017a: 581) correctly points out that “the Preceramic period of the littoral was a complex economy beyond a single specialized subsistence based on maritime resources.” Post-2000 bc settlements generally are not well known, but a site in the Huaura delta, Vichama, shows some significant changes between 1800 and 1500 bc which it shares with other inland sites. Unbaked polychrome figurines found in caches, decorated niches, and facades with reliefs began to appear, and bags with stone fillings used in constructions acquired enormous scales and weights (Shady et al. 2015b). These architectural and artistic innovations, partially an outcome of earlier practices, probably were stimulated by the coeval Sechín polity (1700–1400 bc), though its ceramics seem to be absent in Supe and probably further south. Between 1400 and 900 bc, the northern river complex seems to lose importance and probably a shift of gravity in cultural developments toward the south set in. Some 40 sites with distinctive monumental architecture were built and occupied between the Chancay and Lurín valleys, with minor extensions to the Huarmey in the north and Mala in the south (Burger and Salazar 2012). These monumental structures show some parallels with previous ones at Caral in the Supe valley, but the exuberant mural decorations and imported or emulated objects hint at a close connection with the coeval Cupisnique sphere on the North Coast between about 1000 and 800 bc. The spectacular ceramics, which also have been found in a gallery at Chavín de Huántar, differ slightly from styles of the North (Lumbreras 1993). They also were distributed on the North Coast, in the Huánuco region, and on the South Coast, but the social and political background does not hint at an eventual political unification: though some internal differences of monumental architecture in scale and distances are evident, often more than one monumental site occurs in each single valley. Most of the sites are unexcavated and associated non-monumental architecture is almost unknown, so that a supposedly exclusive function as ceremonial centers is questionable, as is their assumed dependence from Chavín de Huántar. This cultural area suffered changes after 800 bc when rural villages, some of them of major dimensions, proliferated (Palacios 1989; 1999; Goldhausen 2012). On the other hand, fortified sites, such as those known from the North Coast, extend to the Huaura Valley with Salinar-like pottery (Brown Vega 2009). Rather large funerary areas exist in the Lurín

valley, whose occupants perhaps were highlanders; contacts with the Mantaro basin were established (Makowski and Burger 2009). This basin counts with some minor non-domestic architecture from 1000 bc onward (Browman 1970). Its population subsisted on the basis of pastoralism and supplemental cultivated plants that became available thanks to agrotechnical improvements. Further to the east, the Huánuco basin produced a long archeological sequence that began at c.2500 bc and were thus coeval with the Supe area. The characteristics of its earliest non-domestic architecture are also present as a minor component at Caral, and some elements such as bone needles (tupus), unbaked figurines, and transversal flutes apparently were interchanged or emulated. At about 1800 bc, a distinctive ceramic style, associated with elaborate burial structures and little-known small settlements, emerged, and was produced until c.1200 bc. This Wairajirca style is one of the earliest pottery complexes in the Andes and is usually related with early Amazonian pottery found at Tutishcainyo near Pucallpa (Lathrap 1970), but its typical “Amazonian” forms and decorations only become apparent at c.1500 bc (Kanezaki et al. 2021). Between 1500 and 1200 bc this Wairajirca style pottery became widely distributed in highland settlements like Chawín (Brown 2017), rockshelters in the Junín puna (Lavallée 1977), and at some places on the coast. Jet mirrors, stone mortars, and elaborate stone architecture found at Wairajirca are features that were probably imported from the coeval Sechín polity (Izumi et al. 1972). The following Kotosh phase still has an Amazonian flair, but is much more widespread in the highlands and coast. It distinguishes itself by a particular funerary context associated with jet mirrors and decorated bone implements that resemble Chavín art. These bone implements suggest contact with Chavín de Huántar, where Kotosh pottery was found, too. Even the following, apparently brief, Chavín phase at the site of Kotosh, so called because of strong similarities in stamped pottery at both sites, does not necessarily imply foreign occupation or an implantation of a supposed Chavín cult, as decorated lithic art and other related traits typical for Chavín are absent. The following Sajarapatac phase at Kotosh retains some earlier forms, but the arrival of distinct ceramic forms, decorations, and architecture hint at the presence of different populations. Thus, these eastern Andean cultural developments are closely connected with the west, particularly the coast, but also with the Amazon, which is shown above all in minor decorated pottery types which probably are emulations of original wooden trays with carved designs (Izumi and Terada 1972).

49

peter kaulicke During the 1st and 2nd century bc and the first 2–3 centuries ad a remarkably fluid situation prevailed, characterized by different spheres of interrelated complexes that were connected through short-distance contacts or long-distance connections. These interrelations could include frequent small-scale population movements, but also exchange with regions farther away, such as the North Coast, the South Coast, and the Mantaro basin in the highlands. Economic strategies on the coast were based on a combination of agriculture and fishing, supplemented by camelid herding and occasional hunting of wild animals. Settlements were small and shifted frequently, with a tendency toward major structures at the end of the period. Violence and a growing relevance and frequency of different forms of warfare could be major preconditions for urbanism and state formation (Palacios 1989, 1999). A more long-term culture, known as Lima, appeared between the Chancay and Lurin river basins in the 3rd century ad and lasted until the 9th. Thus, this culture emerged slightly later than the Mochica and the Recuay societies in the North. Mochica probably contributed to the development of monumental architecture, while Recuay art motifs and techniques influenced murals, pottery, wood, and occasionally textiles. Impacts from the South also contributed to architecture and ceramic forms and decoration of the Lima culture. Its rise and demise are shared with the fate of the Mochica societies when ENSOs impacted them in the fourth and sixth centuries. Toward the end of the Lima culture, large urban settlements in the lower and middle Rímac valley emerged. While these data hint at the existence of a territorial state, the problems of definition are similar to those mentioned for Mochica. Outstanding elite funerals like those found in the North are not reported, and metalworking is not comparable in amount and sophistication to the contemporary North (Patterson 1966; Kaulicke 1997; 2001; Goldhausen 2012: 327–52). The roughly four centuries between ad 600 and 1000 are marked by the first expansive state of the Central Andes, Huari, whose core area lay in Ayacucho. Its impact was felt from Cajamarca in the North to Moquegua in the South, but its nature in the Central sector is complicated mostly because of deficient research and the dearth of detailed publications. The Lima societies did not succumb to Huari in its first extension (c.600–800 ad), but even entered a short florescence. A ceramic style called Nievería, an elaborate, probably prestige ware, pertained to the Lima material culture but adapted elements of other styles (Valdez 2010). It gained a wide distribution ranging from the Jequetepeque valley and the Callejón de Huaylas, to the North-Central Coast. An aspect usually less considered are connections within the

50

highlands, particularly between Jauja, Huancayo, and Ayacucho, the Huari homeland. These might have played a prominent role in the distribution of other Huari styles to the coast and the emergence of the later coastal Pachacamac style (Kaulicke 2000). Important changes occurred during the Middle Horizon in the Mantaro basin. Settlements at higher altitudes became abandoned in favor of larger constructions on the valley floor, including stone “temples,” storage buildings, and more elaborate funerary structures. This probably happened due to contacts with more complex societies, as documented by pottery from Cajamarca, Huamachuco, the Callejón de Huaylas, and Ayacucho, particularly between ad 800 and 1000 (Browman 1970). Rectangular annexes to the temples might represent early cores of larger Inca period settlements. One of the temple sites, Wariwillka, was an important oracle in Inca times. It shares this function with the much more famous Pachacamac the mouth of the Lurín Valley. This river valley functioned as a conduit for an Inca road about 230km long that connected to Hatun Xauxa near modern Jauja (Capriata et al. 2019). Pachacamac first was a funerary area at c.200 bc and remained in use as an oracle until the European invasion. A structure known as the Old Temple was established in Early Lima times and lasted to the 16th century as part of a series of superimposed buildings. Probably due to intense rainfalls, this complex collapsed, but was reconstructed. During the Middle Lima phase, from ad 300 to 500, monumental architecture and adjacent settlements began to form a florescent core area. At this time, ceramics and textiles appeared in offerings and burial contexts of several styles known from Ayacucho or the South Coast, but also with influences from the North (earspools with inlays of shell and semi-precious stones). These combinations led to the definition of a proper style, Pachacamac, that was probably used by members of an elite who were buried in ceremonial enclosures. Coeval architecture, however, is poorly known, and thus inconclusive. A wooden carved and painted object was venerated in the late 8th century. Carved in a style known as Casma, this cult image shows characteristic northern traits associated with Huari ceramics and painted textiles like Pachacamac (Franco Jordan and Paredes Botoni 2016; Pozzi-Escot 2017; Sepúlveda et al. 2020). Recently, similar associations have come forward from excavations of a spectacular funerary area with burial towers, known as chullpas, of different scales and complexity in the Huarmey Valley (Prümers 1990, 2001; Giersz 2017). Thus, the evaluation of the evidence and mechanisms for the spread of Huari elements in the Central sector is difficult. Pertinent evidence is available from about ad 600, when earlier powerful societies in urban sites still existed

physical geography and cultural trajectory or even flourished. At c.ad 800, after strong natural impacts, these elements certainly became widespread, but still sporadic. They are mostly known from funerary contexts that are intrusive to earlier ceremonial structures or from structures like burial towers, which are more common in the highlands. Funerary practices such as bundles of individuals with artificially modified crania are introductions from the South, but the associated objects are a mixture of imported, regional, and local items. Emergent elites might have sought new legitimation after the natural disaster before they themselves fell victim to another one. The environmental impacts should have caused some population movements, but military conquests from the core Huari area in Ayacucho or other foreign coercive measures are improbable for the Central sector. In the following centuries up to the European invasion, from c.ad 900/1000 to 1532, the climatic conditions first improved with the Medieval Optimum (ad 1000–1300), but then rapidly deteriorated with the onset of the Little Ice Age (ad 1300–1850). Population density increased to the point that settlements got much larger and more organized, with recognizable core areas often developing as extensions of earlier Lima culture centers: these include Lauri in the Chancay Valley with a surface area of 195ha, Cajamarquilla in the Rímac Valley, with 167ha, and Pachacamac, with c.300ha and a monumental core of c.150ha (Guzmán Dávila 2016). The shared construction techniques at these centers involve the use of rammed-earth blocks known as tapia that were rearranged in agglomerated room compounds and larger enclosures encasing “pyramids” with ramps. These latter are often interpreted as palaces. Such edifices are not only characteristic of the Central Coast, but appear in scatters from Chancay to the north; larger concentrations extend from Casma to Lambayeque. Regional differences in portable material culture are mainly in evidence from burial contexts in extensive but heavily looted funerary areas attached to settlements. A relatively simple ceramic style, called Ichsma, is common in the Rímac and Lurín valleys (Eeckhout 2004). Attractive pottery and fine textiles come from the Chancay and Huaura valleys, the core areas of their manufacture and distribution (Krzanowski 1991). More to the north, the Casma style with its different forms and decorations (Vogel 2016) continued to be produced. However, all these styles were widely distributed over the entire Central Coast. Other styles from the North (Lambayeque and later Chimú) are also commonly found on the Central Coast in this time period, particularly at Pachacamac. A different situation characterizes the highlands, mainly the Mantaro basin. The local Wanka style pottery is subdivided into Wanka I (ad 1000–1350) and Wanka II (ad

1250/1350–1470). The settlements, often located in higher altitudes, comprise circular stone structures of varying numbers that were encased by walls and ditches. During the Wanka II phase, settlements achieve dimensions of more than 30ha. They often are subdivided in two main sectors and probably were home to more than 10,000 inhabitants (D’Altroy 2015; Perales Munguía 2016). Other signs of duality are pairs of two settlements (cf. Chapter 24 of this volume, by Matthias Urban, for the structuring principle of asymmetric dualism and its relevance for social organization and, potentially, language dynamics). Most settlements, however, are much smaller and may be called villages or hamlets. Such settlements are found in Jauja, around Lake Junín, and Tarma (Parsons et al. 2000). The relatively simple ceramics associated with the Wanka style varies in decorations and forms; hence, these ceramics can probably be associated with different polities or ethnic groups. Wanka political organization still is a matter of debate, but the assumption of supreme elites with hegemonic rulers governing large states is unwarranted. The economy was based on a combination of agriculture and herding of camelids. Some kind of access to the eastern Andes is probable, though only sporadically. The expansion of the Inca Empire left its mark on the region during the last century of prehistory before the European invasion. During this short period, the political strategies in the Central part of the Central Andes varied. A road system connected the Central region, both coast and highlands, in horizontal and vertical directions. Sites at crossroads turned into large complexes. Thus, Pachacamac became an interregional oracle center, which is also relatively well known from early written sources (Pozzi-Escot 2017). Hatun Xauxa was built as a large administrative center of about 200ha in Inca times (Perales Munguía 2013). Pumpu, on the shore of Lake Junín, also on the main Inca road, is another important site (Matos Mendieta 1994). The Inca impact in the Mantaro valley was variegated, but brought significant overall changes, such as a nucleation of settlements and the emergence of centers, concomitant shift of settlements to lower elevations, and apparently forced deportations, known as mitmaq, which caused a drop in total population levels. On the coast, many settlements subsisted, some with only slight evidence of Inca influence, others seemingly without any at all. Inca ceramics are frequently hybrids with other foreign or local styles. Summing up, the Central sector of the Central Andes differs in major aspects from the North in geography, ecology, and cultural trajectories. It was culturally and socially less complex, largely without emblematic proper products,

51

peter kaulicke and often dependent on impacts from North and South in architecture and portable material culture. Periods of major developments, such as precocious Caral, were usually short and separated by periods in which smaller polities proliferated. On the coast, agriculture coexisted with extensive fishing, while herding and some agriculture characterized the highland basins until the 16th century ad. Economic and political interaction focused on the southern coastal sector from Chancay to Lurín. Although some connections with the coast existed much earlier, the Mantaro basin remained a kind of backwater until the 6th century, when it became integrated in interregional networks. The Central sector hence was a transitional region fully integrated in a Central Andean culture area only from the 9th century ad until the European arrival.

2.4 Southern perspectives 2.4.1 Introduction The Southern part of the Central Andes (see Figure 2.4) is often defined as limited to the approximately 300km long coastal stretch between the south of Lurín and the Yauca River to the south of Nazca. Especially from the 6th century ad to the European arrival, however, it is necessary to significantly expand the geographical scope and include the Peruvian/Bolivian Titicaca basin, where the expansive Tiahuanaco polity arose from different smaller polities. A largely coeval and interrelated trajectory that linked Huari and Tiahuanaco connects this area to the North. Taking into account the entire distance from Lurín to Iquique in modern Chile, to which Tiahuanaco extended in the south, the Southern part of the Central Andes extends across 700km. In the highlands, the distance between the Mantaro basin to Uyuni in Bolivia is even greater, c.800km as the crow flies.

2.4.2 Environment Generally, the coast widens to 20–30km between the Chincha and the Yauca river basins, but more to the south the coastal stretch is minor, as the cordillera often touches the sea. The Chincha river basin forms a complex with the Cañete and Pisco Rivers. Further to the south, another complex comprises the Ica, Río Grande de Nazca, Acarí, and Yauca rivers. Still further south, rivers cut deeply into the flat and dry Cordillera foothills. All these form hydrological basins with many tributaries flowing down from the highlands at above

52

3,000masl. The Chincha River has a length of 148km and, like others, runs almost dry during July and September (Autoridad de Agua 2012). Temperatures are above 30° C during most of the year. Precipitation is almost absent except for some humidity from eastern monsoon rains that reach the coast. The course of the South Highlands inclines toward the east and has a width that is about twice as large as that of the North and Central parts of the Central Andes; at the height of Arica in northern Chile, the Andes widen to about 700km. Thus, the highlands are much more dominant in defining the overall topographical profile in the South than in the North. An important topographical feature is the Abancay deflection, which is less well known than the Huancabamba depression in the North. It runs through the highlands from Paracas to the foothills of the eastern Andes and represents the northern limit of mostly inactive volcanoes (and therefore presence of obsidian) (Reynel et al. 2013: 178– 9). From this deflection onward to the south, an extensive puna (high-altitude grasslands) covers the main part of the highlands. Up to Ayacucho, it is interrupted by hydrological basins such as that of the Mantaro, where it joins the Pampas basin. Further to the south, other basins such as that of the Apurímac and Urubamba in the Cuzco region share the same southwestern direction. Still other rivers flow northwards to form tributaries of the huge Amazonas hydrological basin. Generally, temperatures are high, and aridity to hyperaridity dominate not only on the coast but also in the highlands. Thus, maximum temperatures in Ica, on the coast, generally are near to 30° C during most of the year, while minima are between 13 and 20.5° C. Precipitation is practically absent. In Ayacucho, at 2,761masl, maximum temperatures are between 23 and 27° C and minima between 7 and 12° C. Precipitation is almost absent during June and July, but reaches c.100mm per month from January to March. The maximum temperatures for Cuzco, at 3,399masl, are 18– 22° C, the minimum 2–8° C; precipitation is between 8mm and 1,982mm per month. Finally, Juliaca, at 3,825masl, has a maximal temperature of 18–22° and a minimum of 3–8° C. Precipitation is absent between May and July, and oscillates between 80 and 170mm per month during December and January. Drought and severe frost with snow, however, are relatively frequent. Humidity is mainly due to monsoon rains from the east, fogs from the sea, and evaporation of ground water. The ecoregions are adaptations to these extreme conditions. In the Ica valley, six landscapes types can be distinguished: dry huarango (Prosopis limensis) woodlands (huarango is a tree with up to 70m-long roots allowing access to ground water), riparian forests and scrubs, huarango stands on dunes, cacti and scrub forests, xerophitic scrublands, and

physical geography and cultural trajectory Lurin R.

15 11 Ayacucho

Cañete R. Chincha R.

Hu at an ay Cusco Ap urí 22 ma 16 c

23

a es op Or e cr Lu

Pampas

Pisco R. 10 Paracas

13 14

12

Palpa 3 Ica R. Nazca R.

9 78 Juliaca

Acarí R. Yauca R.

Puno 4 Siguas Majes Vitor

21 6 20 Catarí Taraco 18 19

Arequipa

5 17 Moquegua Cochabamba

Ilo Tacna

2 Africa

Azapa

24

At

ac

am

aD

es

er

t

Iquique

Atacama Desert

0

125

250 Km

San Pedro de Atacama

rt ese aD cam

Ata

N

Uyuni

Loa

1 Quebrada de los Burros 2 Asana 3 Pernil Alto 4 Toro Muerto 5 Huaracane 6 Chiripa 7 Qaluyu 8 Pucara 9 Cahuachi 10 La Muña 11 Wari Chonchopata 12 Huaro 13 Espíritu Pampa 14 Vilcabamba 15 Azángaro 16 Pikillacta 17 Cerro Baúl 18 Tiahuanaco 19 Khonko Wankane 20 Lukurmata 21 Pariti 22 Coricancha Sacsayhuaman 23 Machu Picchu 24 Camarones Abancay deflection Modern towns

Lake Titicaca e v Ila

Figure 2.4 The southern part of the Central Andes with sites mentioned in the text. Map by Karla Petroni.

wetlands with lagoons. Animals in these habitats are endemic birds, reptiles, and migratory species such as guanaco, puma, mountain cat, and some birds (Whaley et al. 2010; Beresford-Jones 2011).

The western flanks in the South of the Central Andes are even more barren than the Central ones. Above 3,000masl, large stands of tola (Parastrephia cuadrangulare) bushes grow in association with cacti. Typical puna vegetation occurs

53

peter kaulicke between 3,900 and 4,000masl and includes grasslands, cushion plants, wetlands, and remnants of queñua (Polyleps spp.) and quishuar (Buddleia incana) trees. Vicuña, deer, tinamous, the Andean ostrich (Pterocnemia pennata), and water fowl in the lagoons are typical animals (Rauh 1958; Brack Egg and Mendiola Vargas 2010; Reynel et al. 2013). The wide highland belt is interrupted by valleys. Some of these are enlarged to basins with valley floors at c.2,000masl, which support plant formations similar to the coast and the western flanks. In the Ayacucho highlands, these basins extend to the south, where they come close to the headwaters of the coastal Acarí valley. Nowadays the Ayacucho and the Urubamba basins have the highest population density. Most important is the Lake Titicaca basin, which is situated between the occidental and oriental cordilleras. It is over 1,200km in length from northwest to southeast and has a surface area of 57,340 km2 . Surrounded by mountains of up to more than 6,000masl, its valley floor ranges in elevation from 4,500 to 3,600masl. The lake itself covers 8,562 km2 , and thus is about 16 times the size of Lake Junín, the second largest lake in Peru. In the Titicaca basin, numerous rivers and springs form extensive wetlands with important reed stands. The large body of water produces microclimates with higher temperatures, which enable cultivation of plants that are normally grown in lower altitudes. Lake levels, however, varied considerably in the past. Low levels are in evidence between 900–800 bc, 400–200 bc, and ad 1–300, while wetter periods with higher levels occurred between ad 610–650 and ad 760–1040 (Binford and Kolata 1996; Stanish 2003). Summing up, hyperaridity characterizes the coast from Ica to Iquique (Atacama desert), and similar conditions extend to the Andean foothills and the south of the Titicaca basin (dry puna). Puna belts are much more important than further to the north and enable intensive pastoralism. InterAndean river basins of major extensions are found at greater distances from the coast (more than 400km) and relatively near to the eastern Andes, where agriculture is possible. The highlands thus are the major source of economic wealth and population concentrations in the South, in contrast to the coast and relatively nearby inter-Andean valleys which dominate the Northern and Central parts.

2.4.3 Population and cultural trajectories This huge area is of utmost importance for Central Andean prehistory as the homeland of the most expansive states in pre-Hispanic history: Huari in the north on the limits with the Central sector and a history linked to the North, Tiahuanaco, on the southern limit of the Central Andes, and finally,

54

the Inca state with its core area around Cuzco between its two predecessors. While the northern part of the South sector is still little known as far as its earliest inhabitants are concerned, shell heaps are in evidence from the South Coast since the Terminal Pleistocene. Early sites are also present in the highlands and in the Atacama desert (Latorre et al. 2013). More and better-known occupations on the coast, at Quebrada de los Burros and Tacna, and in the highlands, at Asana, emerged early in the Early Holocene; occupation at Asana spans the entire Archaic (Aldenderfer 1998; Lavallé and Julien 2012). At this early date, highlands and coast were already connected; obsidian was in much demand. Coastal populations were adapted to fishing and developed a sophisticated technology including watercraft for offshore captures. Hunting was complementary on the coast, as was the occasional use or consumption of wild and domesticated plants; in contrast, hunting was essential in the highlands. Between 5000 and 4000 bc, loma settlers near the Ica delta consumed wild plants and land snails (Beresford-Jones et al. 2015), but the inhabitants of an inland site, Pernil Alto, which features semi-subterranean circular structures similar to those on the Central Coast, depended on a wide range of cultivated plants like beans, potatoes, squash, arrowroot, guava, and probably jíquima (Pachyrhizus tuberosus). Products from the sea and hunting apparently were relatively unimportant (Gorbahn 2013, 2020). Quinua (Chenopodium quinoa) was grown and llama came to be domesticated in small highland and Altiplano hamlets or household clusters (Aldenderfer 2012). Thus, a panorama of growing complexity and increasing populations unfolds, which is not unlike what occurred in the areas more to the North, but at a notably minor scale and without any signs of non-domestic architecture. A remarkable mortuary treatment, the so-called Chinchorro Funerary Tradition, was practiced between Ilo in southern Peru and the Loa River in northern Chile, on a coastal stretch of about 600km in length. Between c.5000 and 1400 bc young individuals and fetuses whose bodies were modified in complicated ways were recovered from large burial grounds in the river deltas. They were transformed into effigies with masks and wigs, apart from other naturally mummified bodies of all ages. In some cases anthropomorphic figurines of clay or wood were associated, some with real body parts. Coeval settlements were simple and reduced. Economically, their inhabitants depended on abundant marine and wetland resources, but they had access to the highlands to up to 4,000masl (Sanz et al. 2014; Standen and Arriaza 2016; Kaulicke 2016a; Standen 2020). Subsequently, semi-subterranean circular houses turned into more permanent rectangular structures in villages. Storage facilities and associated funerary structures began to be constructed. Gold collars with sodalite beads were

physical geography and cultural trajectory recovered from such funerary structures near the Titicaca basin in the Ilave River, which probably indicate longdistance contacts with the eastern Andes. Agropastoralism became established in times when modern climatic conditions consolidated, corresponding to the Early Formative in cultural chronology (Craig 2012). During the Early Formative (1500–800 bc), simple and mostly undecorated ceramics and modest rectangular compounds prevailed. On the Peruvian South Coast at Pernil Alto, a wide range of cultivated plants and trees was grown, including maize. Hunting was possible due to the existence of large forests; settlements usually were small and without centers, and simple irrigation was practiced. Between 800 and 260 bc, the Paracas culture covered an extensive space between the Chincha and the Yauca rivers, but its core area was between the Paracas peninsula and Nazca. Ceramics with attractive form and decorations that were probably inspired by the North Coast Cupisnique style are diagnostic for this culture; it was recovered from elaborate burial chambers with decorated gourd vessels and textiles (Menzel et al. 1964; Kaulicke et al. 2010; Kaulicke 2015b; Mader 2019). After an initial input from Cupisnique at c.1000 bc, more independently emulated items prevailed. Settlements became larger, and non-domestic architecture as well as small burial grounds developed with emerging signs of social inequality. Monumental stone architecture in the Ayacucho highlands led to the postulation of direct influence from Chavín de Huántar, given that there are some similarities to architectural styles in the North. Other aspects of material culture, however, hint at a strong local background with ample evidence of contacts with the contemporaneous South Coast. Late Paracas (c.400–200 bc) represents an early climax of complexity, variety in material culture, settlement growth, and population density. Monumental architecture appeared in the Chincha valley in association with geoglyphs, large settlements emerged in the Ica valley, and defensive installations and relatively large sites with unusual ground plans in the headwaters of the Río Grande de Nazca came into existence. Pottery and textiles developed substyles of high quality. Some of these products were widely distributed on the Central Coast and the Central and South Highlands (Bachir Bacha and Llanos 2015; Canziani 2015; Tantaleán and Stanish 2017; Mader 2019). Some 350km to the south of Nasca, in the Siguas Valley, a large number of textiles from looted contexts show affinities to the Late Paracas style, but also to Altiplano iconography. Canes with engraved decorations are most similar to rock art sites, such as the famous Toro Muerto in the nearby Majes Valley, which is usually dated much later (Haeberli 2002; 2018). Another complex called Huaracane existed between Ilo and Camarones in northern

Chile between c.400 bc and ad 400. Its relatively simple ceramics share attributes with undecorated pottery of the Paracas culture. Settlements consist of terraces with circular domestic units and associated funerary areas, as well as areas of funerary mounds built up by alternating fill and vegetal layers. These mounds were probably used for reiterated use by extended families. Boot-shaped pits contain imported textiles and ceramics that are coeval with Nasca and Pucara (Goldstein and Owen 2002). In the Titicaca basin, centers emerge in its northwestern and southeastern extensions. Chiripa (type site and coeval settlements) on the Taraco peninsula in the south has a long history beginning at c.1500 bc. Its development is perhaps related to rising lake levels, which allowed access to rich lacustrine resources and to fertile agricultural land (Hastorf 1999; Bandy 2006). At c.800 bc the area became part of the Yaya-Mama Religious Tradition, whose characteristic traits are sunken courts, stone sculptures, ceramic pedestalbased burners known as incensarios, trumpets, and decorated serving bowls (Chávez 2018; Janusek and Ohnsted 2018). All of these are present in Late Chiripa or slightly earlier. The most elaborate and standardized architecture at Chiripa, rectangular structures on a platform called “temple storage complex,” date to 400 bc. Human burials beneath the floors of these structures included gold and copper objects (Bandy 2006). The northern Titicaca basin differed culturally, but was connected to the south. Qaluyu is the best-known site. With an extension of approximately 7ha, it is about the same size as Chiripa, but other similar less known and even larger centers coexisted together with many smaller settlements (Levine et al. 2012). At c.500 bc another center emerged at Pucara, which flourished between 200 bc and ad 200. Covering probably more than 1,5 km2 , this site is much larger than previous settlements, and features an impressive monumental core of terraces and structures (Klarich 2012; Klarich and Román Bustinza 2012). While formally a late manifestation of the Yaya-Mama Tradition, Pucara excels in the quality of stone sculptures with complex representations (Young-Sánchez 2004) and polychrome vessels show complex iconography that was not present in earlier sites. Motifs include anthropomorphic females and males as well as therianthropic beings, camelids, felines, and isolated human heads (Chávez 1992). Raised fields and artificial lagoons hint at intensive agriculture and a major role of domesticated camelids; non-monumental areas seem to have housed substantial populations who partially dedicated themselves to craft production. The distribution of specialized objects of Pucara origin in minor centers and hamlets led to Pucara being defined as a center of a state that covered the entire northern basin and part of the

55

peter kaulicke south. Pucara elements were also present in distant regions such as Cuzco (Rowe 1976; Chávez 1989; Zapata 1998; 2019) and the coast between Ica and Azapa in Chile (Conklin 1983; Agüero and Uribe 2001). Spectacular textiles from looted burial contexts in the Siguas, Majes, and Vitor basins, called Provincial Pucara, are coeval with the Pucara complex, but extend to about ad 500 after the definite demise of Pucara proper. Some of their iconography relate to later Tiahuanaco (Haeberli 2018). Pucara is an astonishingly advanced political formation with ample evidence for hierarchical structures and a wide interaction sphere, whose diverse mechanisms and extents are still in need of better definition. Before focusing on Tiahuanaco, the coeval trajectories on the South Coast and its relations to the highlands up to the emergence of Huari require attention. The Late Paracas phase is followed by what Carmichael (2019) calls the “Necropolis Era” from 200 bc to ad 100. Necropolis burial complexes on the Paracas peninsula are famous for their hundreds of funerary bundles that were wrapped in spectacular polychrome textiles with complex iconography and deposited in abandoned architecture. The associated fine ceramics is monochrome and belongs to the Topará Tradition, which is present in the Cañete, Chincha, and Pisco valleys and extended (by invasions?) to Ica and Nasca; polychrome Late Paracas and Initial Nasca ceramic traditions are partially contemporaneous. The center of the following Nasca Tradition are the Ica and Nazca Valleys, but, from the 1st century ad to ad 650, more local variants exist from the .Cañete valley in the north to the valleys in the Arequipa department in the south. Nasca is famous for its ceramics with polychrome painted decoration on simple but elegant forms, geoglyphs, and subterranean irrigation systems (Silverman and Proulx 2002; Schreiber and Lancho Rojas 2003; Lambers 2006; Pardo and Fux 2017; Reindel and Isla 2017). Previous centers in the Ica Valley were abandoned in Early Nasca times, and a new large center, Cahuachi, emerged in the Nazca valley (Silverman 1993; Orefici 2012). It grew to c.150ha in extension by ad 200, and while it lost its importance shortly thereafter, it survived into Middle Nasca times. Cahuachi’s political status is a matter of discussion; interpretations range from a vacant ceremonial center or temple city, paramount chiefdom, to capital of a unified state. Material culture is uniform, and the widespread distribution of ceramics and textiles in the Chincha, Acarí, and Arequipa valley basins is interpreted as evidence for colonies. The headwater sites disappeared, and the clearance of forests in the valley floors set in to facilitate colonization (Soßna 2015). New centers appeared in Middle Nasca times (ad 300–440). Many major sites were abandoned and new centers with rectangular

56

adobe structures and plazas, some with elite residences, emerged. At La Muña near Palpa, excavations revealed a large settlement and an extraordinary looted burial complex with rectangular enclosures built on platforms. Underneath them, shafts 5–7m deep lead to a looted quadrangular chamber with adobe walls and a roof of huarango trunks. The chambers still contained fine ceramics, Spondylus and semiprecious stone beads, obsidian points, and gold objects (Isla and Reindel 2006). During Middle Nasca times, the headwaters were resettled, but their circular architecture differs from building practices in the valley floors, so that their populations probably were not identical to the coastal ones. Finally, during the Late Nasca phase (ad 440–650), the population decreased, probably due to severe droughts. During this time, some areas were abandoned and there were generally fewer sites and centers with organization that differed from previous phases and included highstatus sectors. On the other hand, connections to other distant regions including the North Coast and the southern Arequipa valleys are evident in pottery and textiles (Soßna 2015). As mentioned, the Ayacucho region featured several monumental complexes, which are usually related culturally, politically, and chronologically to the Chavín cult. However, it seems that these did not disappear with the collapse of Chavín de Huántar at c.550 bc, but apparently continued to exist until at least 300 bc with continuous contact to Paracas on the South Coast. At c.ad 300, another culture complex, Huarpa, emerged to last until the end of the 5th century or even later. It is defined by ceramics with Late Nasca decorative traits. About 500 small sites as well as a relatively large center of c.7ha with residential compounds and a circular monumental structure were registered (Leoni 2006). Substantial architectural Huarpa remains at the site of Huari hint at a decisive role of this culture in Huari’s exceptional growth. During the following Early Huari phase (c.ad 600– 800), complex compounds with columns and niches were built at this site; the ceramics are in a Nasca–Ayacucho tradition and were often smashed intentionally (Ochatoma Paravicino et al. 2015; Cabrera Romero and Ochatoma Paravicino 2019). Another important site near Huari is Conchopata, where Huarpa burials lie beneath Early Middle Horizon architecture (Isbell 2001; 2018). Smashed fancy giant anthropomorphic urns in various offerings before ad 800 are early examples of a later more extended and elaborate tradition. Up to this date, Tiahuanaco influences in the shape of pottery or architecture were absent, so that such influences could not be responsible for Huari’s initial growth. Contacts between Nasca and Ayacucho in Late Nasca times could be due to crises in the coastal valleys and population displacement to other regions, including to

physical geography and cultural trajectory Ayacucho in Huarpa times. On the other hand, some architectural features on the coast are similar to Huarpa circular monuments, and pastoralists from the highlands entered the headwater regions (Soßna 2015). Huari is the center of a core area with a dozen major sites. This core area was about 35km in length, but its influence was much more widespread. The Cuzco basin and adjoining regions had shown connections with the Titicaca basin since the Formative, when many sites were just small hamlets, but some ceremonial centers were characterized by YayaMama traits (pottery, stone bowls, sunken plazas). These were succeeded by others with Pucara and Early Tiahuanaco elements (Zapata 1998). Thus, early Huari sites might have interrupted these Altiplano connections. They were quite numerous in clusters like Huaro, which had an extension of more than 200ha. Formal architecture and fancy ceramics that are almost identical to those found at Conchopata hint at close contacts with the Ayacucho homeland. Such clusters could reflect Huari colonies (Zapata 2019). They also reached the Eastern Andes at sites like Espíritu Pampa and Vilcabamba. Probably these sites served as outposts for acquiring feathers, monkeys, drugs, coca, or other lowland cultivated plants (Isbell 2016; Knobloch 2016). At about ad 800, a decisive change took place that converted Huari into a metropolis with a core area of 250–500ha and a total urban area of 1,000–1,500ha with continuous contact with Paracas on the South Coast. Large compounds enclose residential structures, perhaps palaces, sacred areas with D-shaped “temples,” and funerary areas with huge and complex constructions of various levels up to 15m beneath the surface. This monumental architecture used pozolana, a volcanic sediment and natural cement, which was cut into large building blocks, slabs, columns, and tables in ways similar to Tiahuanaco techniques and forms (see below). Ceramic styles underwent changes, and decorations adopted an iconography related to Tiahuanaco. Offering pits with large amounts of smashed pottery of outstanding quality and complex polychrome iconography are often found, particularly at nearby Conchopata, where they were elaborated in workshops. Beyond doubt, Huari was a huge city dominated by powerful elites that housed perhaps as many as 70,000 people. Complex canal systems for drainage and water supply and terraced slopes for agriculture around the city and its hinterland required major labor organization. New areas came under the sway of its polity, and extant distant centers grew into proper but dependent polities (Isbell and McEwan 1991; Ochatoma Paravicino et al. 2015; Isbell 2018; Cabrera Romero and Ochatoma Paravicino. 2019). Elite funerary structures of similar complexity existed in the Cuzco area near to larger and structured settlements with D-shaped buildings, e.g. at Espiritu Pampa

in the Eastern Andes (Isbell 2016; Zapata 2019). Pikillacta, a large site of 47ha near Cuzco, has an orthogonal cell architecture similar to Azángaro in the Huari core area and Viracochapampa in the northern highlands. These sites are interpreted as administrative centers, but at Pikillacta the most important features are dedicatory offering pits with hundreds of metal, Spondylus, and sodalite figurines together with sacrificed llamas and entire Spondylus shells (Arriola Tuni and Tesar 2011; Cuba Muñiz and Amachi Flores 2019). These figurines represent warriors, prisoners, felines, supernatural beings, and standing figures in different dresses. Another major Huari site is Cerro Baúl in the Moquegua valley. Its impressive setting and the many Tiahuanaco settlements and burial grounds on the other side of the river create the impression of a frontier situation. Cerro Baúl itself may have functioned as a Huari outpost (Williams et al. 2001). The Huari phenomenon has been subjected to diverse interpretations: style, city state, expansionist state, civilization, empire, or more generally a period of globalization. The empirical background of these discussions often lacks clear definitions, and models are biased. The Huari homeland lies near the southern border of the North, where, as mentioned, many complex polities with longstanding traditions existed, and thus was connected with the cultural development in the North. Thus, the introduction of advanced agricultural and extraction technologies, i.e. mining, from the North led to a considerable expansion of economic space for cultivation and pastoralism. These activities had to be organized, in some cases in the shape of client “kingdoms.” The elites in the Huari capital also were in need of large amounts of commodities such as coca and exotic items from the eastern Andes, as well as Spondylus from the north, among other things. Strategies, therefore, probably changed according to the possibilities of control, including coercion; military conquests of a major scale are, however, not sufficiently in evidence. The other major phenomenon in the South is Tiahuanaco. Since the 16th century (and earlier) Tiahuanaco has been considered an origin center, nowadays of New Age mysticism (Vranich 2013). It is the most famous pre-Hispanic site in South America, which during the centuries has produced myriads of myths, speculations, and nationalist exaggerations; more probable recent hypotheses are based on relatively limited research and archeological evidence. While Tiahuanaco indeed boasts a long history, its peak (Tiahuanaco VI and V or “Classic” Tiahuanaco) is shortlived and lasted probably not longer than two centuries from ad 800–1000, i.e. the time frame of Huari flourishment (Knobloch 2013). Late Formative occupations are small hamlets overshadowed by the large Pucara site and polity that

57

peter kaulicke had emerged earlier. By c.ad 550, Tiahuanaco was a village with a small monumental shrine and a funerary area, but the subsequent period until ad 800 is not well known (Marsh 2012; Isbell 2013). It shared decorated stone stelae with other centers like the better-known Khonko Wankane, which is coeval with Pucara, and whose statues are stylistically transitional to the impressive huge Classic Tiahuanaco statues (Ohnstad 2011; Janusek and Ohnstad 2018). These statues at Tiahuanaco (Posnansky 1945) wear clothing with the depiction of Staff God designs (Agüero Piwonka et al. 2003; Agüero 2017). Estimates related to the dimensions of the city and the number of its inhabitants vary from four to six km2 and 30,000 to 60,000 individuals, though about 10,000 to 15,000 would be a more realistic figure (Stanish 2013; Isbell 2013). Its center comprised a massive pyramid mound, a large semi-subterranean plaza, several minor structures which are interpreted as temples, and a palace oriented toward the surrounding mountains. This ritualized space together with statues and substantial offerings give Tiahuanaco the appearance of a ceremonial center or ceremonial city. Outstanding funerary remains are scarce, possible palaces are rather small (Couture and Sampeck 2003), and the interpretation of statues as representation of possible supreme rulers are weak evidence for the conception of the site as the capital of an imperial state (Berenguer Rodríguez 2000). Possible residential sectors and craft productions centers at the site are also in need of better definition, but ceramics and other cultural materials that probably were manufactured at the site are widespread (Janusek 2002). The Tiahuanaco valley was densely inhabited; a major center, Lukurmata, with a surface area of 150ha, lies in the nearby Catari valley (Bermann 1994). Many Tiahuanaco sites border the southwestern shores of Lake Titicaca. Several islands apparently were sanctuaries, some with spectacular offerings (Stanish 2003: 169–86). On the Pariti island near Tiahuanaco, two pits in a larger structure were excavated that contained more than 430 smashed high-quality ceramic vessels, gold objects, more than 40 sacrificed llamas and some waterfowl, all dated to c.ad 1000. Many anthropomorphic vessels seem to represent different ethnic affiliations, such as Colla, Puquina, and even early Inca (Korpisaari and Pa¨rssinen 2005; Korpisaari 2018; Pa¨rssinen 2018). At a distance of about 300km from Tiahuanaco, in the Moquegua Valley, many small sites and some towns, ranging in surface area from 5 to 10ha, are contemporaneous to the main site and can be dated to c.ad 800–1100. They were interconnected with the Altiplano by trails that were marked with giant geoglyphs, similar to those in northern Chile (Goldstein and Owen 2002; Goldstein 2005; 2018). Tiahuanaco was the homeland of most of the inhabitants of the Moquegua valley sites, who shared its material culture

58

even to the household level. As mentioned, they coexisted with much fewer Huari settlers, and interaction was limited. At Cochabamba, about 300km to the southeast of Tiahuanaco, Tiahuanaco influence is evident in the central valley of western Cochabamba (Anderson 2018: 245: fig. 8.2), but diminishes further to the east. Sites are relatively small; they have been occupied since Formative times, but occupations were densest between ad 800 and ad 1100. Rectangular clay architecture in compounds is similar to Tiahuanaco; non-ceramic artifacts, spoons, and flutes are also shared between the regions. Maize and camelid meat consumption was high. All this speaks in favor of strong ties with the homeland during the peak of the Tiahuanaco polity and the presence of highland settlers and traders. The less abundant evidence for Tiahuanaco presence in the northern Chilean valleys does not support massive colonization but more indirect interactions. San Pedro de Atacama, approximately 800km to the south of Tiahuanaco, is an important oasis and node of interaction between the Chilean coast, the Bolivian highlands, and northwest Argentina. Funerary goods include decorated wooden snuff tablets, textiles, and gold objects with Tiahuanaco-like iconography. These hardly indicate direct control, colonization, or other mechanisms related to imperial strategies, but instead are imported or emulated luxury goods for local elites (Uribe and Agüero 2001; Agüero and Uribe 2018). Thus, the Tiahuanaco sphere was centered in the circumTiticaca area, with colonies and outposts on the coast, in the eastern Andes, and further to the south, but without convincing evidence for the imposition or direct control of local elites from the center. To the north of the Titicaca basin, Tiahuanaco presence is less evident. As mentioned, the Cuzco basin boasts a massive Huari presence, but a style called Muyu Orqo is interpreted as strongly influenced by Tiahuanaco pottery with which it seems to be contemporaneous (Zapata 2019; Bélisle and Bauer 2020). What is more, Tiahuanaco ceramics are present in the area too, e.g. in the Sacsayhuaman complex near Cuzco, where they are probably related to structures beneath the famous Inca architecture (Bonnett Medina 2001; Paredes 2003: 104: fig. 27). This would mean that the Cuzco area was the northern part of the Huari–Tiahuanaco frontier documented in Moquegua. The most enigmatic case, however, is the city of Huari itself, where contemporaneous anthropomorphic stone statues (Lumbreras 1974) and emulated incensario vessels (Ochatoma Paravicino et al. 2015: 112) share a strong similarity with Tiahuanaco, and the iconography is one of the most outstanding expressions of the so-called Southern Andean Iconographic Series (Isbell 2018). No convincing explications for its emergence exist until now.

physical geography and cultural trajectory As is the case with polities in other Andean regions, the Huari and Tiahuanaco polities collapsed in the 10th century, probably due to serious environmental impacts combined with internal problems; the emblematic art forms were produced no longer. Documents from the 16th century, influenced by Inca interpretations of their past, create visions of political fragmentation and conflicts prior to the rise of their empire, but that is a simplified concept of imperial state formation, collapse, and its aftermath, which has been shown to be less useful. Variation in ceramic and textile styles hint at earlier internal variety in the Huari and Tiahuanaco realms or interaction spheres. This suggests the presence of diverse social groupings and including ethnic and linguistic diversity instead of ethnicities that only emerged during Inca dominion (Stanish 2003: 204–35; Janusek 2008: 301–5). Many of those ethnicities (or “nations,” to use the Colonial term) appear in colonial sources, but many more must have existed before (see Rowe 1946), as suggested by the aforementioned Pariti representations. These might have been minor polities, but some were kingdoms or empires like Chimor on the North Coast, or Huanca, Chankas, Lupaqas, Collas, and others in the South. Generally, a strong demographic increase set in over the whole of the Central Andes from c.ad 900 (Vega-Centeno 2019). The Inca phenomenon emerged against this complex background. Although written sources and archeological evidence are available, interpretations are diverse. Even more than Tiahuanaco, international fascination with the Incas is still present after 500 years since their first confrontation with Europeans in the 16th century. Scholarship produced multiple models, as well as eurocentric and neo-Indigenistic idealizations related to Inca history, culture, politics, and religion. Modern approaches are often based on a mixture of data from diverse written sources from the 16th and 17th centuries, material evidence (Inca or non-Inca) from the 15th century or even earlier, and unrelated anthropological theoretical models. While Inca society was and still is considered illiterate, a recording system called quipu was widely used for archival purposes and for recording historical information in Inca and early colonial times (Urton 2017). Quipus and additional devices, the tocapus, were already known to the Huari (Clados 2013; Urton 2014) and probably also to the Tiahuanaco society (quipus are not preserved in the environmental conditions of the highlands). But all these are not equivalent to written sources. Since the functioning of these devices is not well understood, the interpretation of Inca history is dependent on oral tradition and observations of the Europeans who first entered the area. One major problem is the emergence of the Inca phenomenon, or Inca ethnogenesis. Cuzco-oriented archeologists and ethnohistorians prefer a scenario of a local development out of multiple peer polities of limited

complexity (Covey 2006; Bauer and Smit 2015). Downplaying previous Altiplano contacts, their main argument is the temporal separation from the Huari presence in the region. Another approach is taking Inca traditions of a Tiahuanaco origin more seriously: the Tiahuanaco core did not disappear suddenly in the 11th and 12th centuries, but remained in the 13th century. Furthermore, Inca presence in the Altiplano is earlier by radiocarbon dates than suggested by the historical interpretation of some written sources. Additionally, many cultural elements, such as representations of elite headdresses and motifs or concepts related to religion or ritual as well as linguistic and genetic data clearly point at strong and rather direct connections with Altiplano populations (Cerrón-Palomino 2013a, 2015; Pa¨rssinen 2015a; Shinoda 2015). As mentioned, such contacts were already extant in pre-Inca times in the Cuzco area. The Inca homeland lies in the Cuzco or Huatanay basin and the adjacent Oropesa and Lucre basins. The capital, Cuzco, at 3,399masl, apparently has a pre-Inca origin, but the first Inca ruler, Pachacuti, who expanded the Inca state, rebuilt it as the center of his state in the early 15th century. The city grew to a core of approximately 40ha, but the surrounding urban space was calculated to have had an extension of about 500ha. Some 15,000 inhabitants lived in the central core, and about 125,000 in the entire city (Agurto Calvo 1987: 86–92). The center of the city and the center of the empire was the Coricancha, the main temple, from where the four world regions were thought to originate. They were structured and linked to the capital by a spectacular road system, as well as the ceques, imaginary lines connecting minor sanctuaries or memory places. Palaces, plazas, and residential compounds, structured by orthogonal street layouts, lie between two channeled rivers that delimit the city core. Separated by terraces, fields, and rivers, suburban villages, some with residential compounds, surrounded the core, as did planned workers’ villages, isolated buildings, storage facilities known as colcas, and agricultural landscapes (Bauer 2004; Farrington 2013; 2018; Crayla et al. 2014). Overlooking the city, a large and impressive ritual landscape, with clusters of sculpted rock outcrops and complex ritual centers like Sacsayhuaman and Muyuqmarca (Mar and Beltrán Caballero 2014), is still in need of detailed archeological documentation (Kaulicke 2015a; 2021). Cuzco was a ceremonial and political center, with people and material vestiges from the entire empire. The pivot of the Inca state was the Sapa Inca, the supreme ruler, who, as the son of the sun, was supported by the gods and his ancestors, whom he joined after his death. Burial customs differed, but the rulers’ bodies received special treatment in accord with complex body politics. Deceased rulers or their images were housed in special places in order to let them participate in major ceremonies (Kaulicke 2015a).

59

peter kaulicke Outside the core area of Cuzco, royal estates, including Machu Picchu, included palaces, installations for craft production known as aqllawasi, temples, burial areas, and memory places alongside extensive agricultural terraces. The road system, with a total length of some 40,000–60,000km, facilitated the movement of people and commodities for different purposes. Way stations known as tambos, storage facilities, and fortified sites in the frontier areas of what is today southern Colombia, northwestern Argentina, and Chile to the south of Santiago are signs of major political and economic organization and coercion (Hyslop 1984; 1990). “New Cuzcos” were replicas of the capital in the provinces. Coerced substantial population movements known as mitmaq served, on the one hand, the political purpose to reduce threats of rebellion. Other resettled people included specialist artisans such as potters, weavers, metalsmiths, and feather-workers. Such groups, known as mitmaqkuna or mitimaes, were brought from all over the empire, though they are often difficult to detect by archeological means (Turner and Hewitt 2018). These colonists were different from the aqllakuna, selected young women engaged in weaving and other activities in special buildings in estates and main centers, as well as the socially segregated yanakuna, who served the Inca aristocracy. The rest of the population were tax laborers and part-time soldiers. Agriculture and pastoralism were both essential for subsistence or transport, but also for different forms of feasting, ceremonial offerings, and religious life in general. Terraces, as well as sculpted rocks and architecture, were part of landscapes of great aesthetic appeal to modern observers (van de Guchte 1990; Kaulicke 2015a; 2021; Christie 2016; Bauer 2018). Textiles, metal objects, ceramics, and other crafted materials together with the characteristic cultural landscapes are symptomatic, easily recognized markers of Inca culture, but still difficult to grasp in their significance (for recent overviews see Pa¨rssinen 1992; D’Altroy 2014; Shimada 2015; Alconini and Covey 2018). Summing up, many problems persist in striking a balance when trying to make sense of Inca history between written and material evidence, between idealization and diversity, and between politically oriented priorities for Cuzco as opposed to the provinces; also, there is a paucity of serious interdisciplinary approaches related to these problems.

2.5 Summary and conclusions The interrelations of population and environmental histories in the pre-Hispanic Central Andes are so diverse that close investigation is required in order to grasp the main

60

aspects of complex phenomena. Commonly, extant narratives are stereotyped evaluations that selectively stress few aspects that are considered relevant while downplaying or ignoring others. In this sense, archeology is able to produce varied empirical data, but interdisciplinary approaches, including contributions from historical linguistics, are required in order to come to reasonably probable social histories (see Chapters 23 and 24, by Paul Heggarty and Matthias Urban, respectively, for some suggestions in this regard). Regional or even localized perspectives should be the base for reconstructing these histories. This has been the aim of the present chapter. The North as defined here is the most complex mosaic of ecological landscapes, dominated by wide littorals, narrow highlands, relatively dense networks of rivers, highly varied seasonally dry and humid forests, and transitional ocean temperatures. The Central sector differs in a spatially reduced and more arid coast, steep western and eastern Andean flanks, and a high puna plain with the large Lake Junín, which is drained by rivers which flow west and east and connect with inter-Andean basins to the south. Important wetlands and fog oases dominate the coast. Many animals are migratory between puna and coast; a major aspect is seasonality. Finally, the South is highland-oriented with important inter-Andean basins and the Altiplano. These basins are much nearer to the eastern Andean flanks than to the sea. Rivers connect inter-Andean basins as well as the Altiplano; coastal river basins are shorter and narrower, so that wide puna grasslands and arid plains at lower altitudes have to be crossed to reach these population centers. Initial colonization of the Central Andes dates to the Terminal Pleistocene in accord with the evidence from other South American countries. There is still relatively little data, though at least two routes can be suggested for initial settlements, one following the coastline and the other via the highlands, even passing altitudes of more than 4000masl. By c.6000 years bc a neolithic way of life set in, documented in two places in the North, Zaña and Huaca Prieta, where circular domestic clusters were superseded by rectangular units associated with non-domestic long-term architecture. This remarkable shift in domestic architectural ground plans set in much later in the Central and Southern parts of the Central Andes. Burial practices and ritualized objects reflect commonly shared ideologies; items from relatively distant places point to relatively large interaction spheres. Some plants were cultivated in garden plots during the Early Holocene, but foragers remained extant in most of the Central Andes. Cultigens appear sporadically on the coast, often as a minor supplement to a general foraging adaptation. Non-domestic architecture developed into more formalized structures

physical geography and cultural trajectory in the early 4th millennium bc on the North Coast, where in some cases walls were adorned with impressive murals. This accelerating process led to special burials, monumental art, and highly decorated portable items in an area between the Jequetepeque and Casma valleys. In the middle Casma valley, huge complexes with shared traits and related villages on the shore and others in the upper valley grew between c.1700 and 1400 bc. This is the outcome of a long history beginning about 2,000 years earlier. Over about a millennium and a half, this core area expanded from Casma in the south to Piura in the north. Thus, in various stages and with shifting power centers through time, it came to cover the entire North Coast. The generalized shared cultural background across this extensive space hints at relatively complex polities with growing inequalities, large centers and associated hinterlands, effective economies based on irrigation, fishing, and some camelid herding, and diverse craft production. There were shifting interaction spheres including highlands and coast. The North Coast as defined here is a space inhabited by complex societies with recognizable long-term cultural traits based on these Formative achievements for another millennium and a half until the 16th century. Despite shorter interruptions, cultural continuity is remarkable and without parallel in the pre-Hispanic Central Andean world. A similar early rise of complexity and nucleation took place in a river complex of the Central Coast with the socalled Caral civilization, between 3000 and 2000 bc; it lost importance after 1500 bc. Early monumental centers emerged in another river complex to the south with traits similar to those of Caral. Possibly their emergence is related to population shift. These traits were still visible in many ceremonial centers occupied between 1400 and 800 bc, when a general trend to rural settlements with minor centers set in. Only in the 3rd century ad did a more complex and expansive polity, the Lima culture, emerge in this southern complex, followed by varied post-Huari polities until the Inca conquest. The Central Coast trajectory is thus a significantly shorter, less steady process compared to the North that was characterized by frequent shifts in power centers. Strong influences from the North are visible in the Formative and later on, up to and including the expansion of the Chimor state to the Central Coast in the 14th century. In the highlands, smaller centers emerged at about 2500 bc in the headwaters of northern rivers, particularly in the Upper Santa and its tributaries as well as in the Huánuco basin at the foot of the eastern Andes. These apparently formed part of larger networks including Caral and Casma. The dense North Coast networks included centers in the headwaters and extended to the Bagua-Jaén area in the

eastern Andes. Chavín de Huántar emerged in a non-center area and became established as an interregional sanctuary in otherwise mainly rural highlands, some perhaps still marked by Archaic lifestyles at this time. At about ad 300, larger polities with ranked societies that were based on agropastoralism and characterized by relatively large centers and distinctive material cultures established themselves in large basins such as the Callejón de Huaylas and surrounding areas. Social complexity in the South of the Central Andes is in evidence even later. While ceramics seem to be present at about 1500 bc as in the North, these are not part of major changes in social and economic terms. Only between 1000 and 800 bc did monumental architecture in the Ayacucho area and in some other highland basins emerge, apparently stimulated by more complex North Coast structures and some imported items, but previous simpler regional traits prevailed. Probably they were part of an interaction network connected to the South Coast. On the coast, connections with the North were much stronger, and led to vigorous social complexity with proper elites and larger centers, from where fine ceramics and probably textiles were distributed across the coast and highlands at about 400 bc. The mechanisms through which these were distributed probably involved contacts between intermediaries who sought to purchase prestige items required by local elites. Generally, intensive maize cultivation, increased demand in obsidian, copper, and camelids for wool, transport, and rituals, as well as quinoa and potatoes in higher altitudes, were motors for increased, widened, and complex networks that led to an increase in social inequality and urbanism. These factors gave the highlands more economic and political weight than before. The Altiplano cultural development toward complexity set in with the so-called Yaya-Mama tradition at about 800 bc. Only around c.400 bc did a site in the southern basin grow to a ceremonial center of about 7ha in extension, with rather elaborate architecture and burials that bespeak social inequality. Other contemporary centers were built in the northern Titicaca basin, but one of these, Pucara, stood out with c.1.5 km2 surface area at its apogee between 200 bc and ad 200. Its complex iconography in stone and ceramics is without convincing regional antecedents, and thus probably was related to the South Coast. Pucara had an interaction sphere that was similar in extension to the much later Tiahuanaco polity. The latter shared the same cultural background, but grew between ad 600 and 800, when it reached its maximum degree of complexity and imperial extension, but disintegrated in the 10th century. The contemporary Huari polity has a different history based on connections with the South Coast, where the Nasca societies had built a center of c.150ha. While this center, Cahuachi, already was

61

peter kaulicke in use perhaps more than half a millennium earlier, its florescence lasted for only about a century, whereas Huari in the Ayacucho basin was only preceded by a regional polity of a couple of centuries before it turned into a city, and about ad 800 into a metropolis. While much of the nature of Huari’s and Tiahuanaco’s political, cultural, and ideological structures and the mechanisms and their mutual relationships are matters of discussion, it is beyond doubt that they were the most advanced polities in the Central Andes before the Incas. Their centers were of a hitherto unknown scale, their direct or indirect control extended to wide territories surpassing even the space known as the Central Andes for the first time. These successes were only possible thanks to previous achievements, fluctuating networks, long histories of changing powers, and materialized memories in spite of the lack of writing systems. The Inca Empire would be unthinkable without these antecedents. The narrative presented in this overview cannot hide some major deficiencies related to the lack or paucity of solid databases, above all concerning histories of populations and related social reconstructions. Inference of demographic increase or decline, densities in core areas, changes in demographic structures like migration, colonization, deportation, and their causes and scales usually rely on vague documentation or even generalizing guesswork. Wide areas in the highlands are practically without available archeological data, recent sporadic genetic and bio-anthropological research often analyzes statistically insignificant samples from heavily destroyed funerary areas in the entire Central Andes, many of them still without serious research. Thus, data on the extensions of these areas, the number of interred individuals, and the duration of their use, which should be one of the bases for population estimates, are mostly ignored factors. Studies of settlements often present estimates of extensions by surface surveys, but excavations that would be able to provide more sustained information are rare. In more

62

complex sites attention focused on the highly visible core areas, neglecting the surrounding non-monumental areas. Hence, inferences regarding site dimensions, functions, and population densities are debatable. Additionally, many examples presented here show that most complex sites have long histories before and after their relatively short peaks, while smaller sites probably were shifting after short durations, which could produce somewhat artificially high site density estimates. Finally, interaction is a major factor in the narrative of this introductory chapter. This encompasses many facets. Environmental factors such as resource distribution, scale of productivity, seasonality, and climatic irregularities led to strategies of intensive mobility that affected major groups in simpler economies or led to migrations during major climatic crises. In more complex societies, larger-scale interaction involved contacts with other regions and polities in order to satisfy needs for growing populations and elites. This took the form of peaceful exchange including intercommunal congregations, feasting, or pilgrimages, or less commonly coercive means like raids or conquests. When it comes to long-distance contacts, efficient means of transport are essential. Terrestrial movement depended on camelids and human bearers, and maritime transport on balsas or reed boats. The regularity, amount, and distance varied, but could have been considerable, as the wide distribution of Spondylus from the extreme north or semiprecious stones like sodalite from the southern Bolivian highlands indicate. These goods were procured on almost industrial scales by expansive states and thus are recognized as valuables in the entire Central Andes. In this sense, it was interaction in dynamic networks which created the Central Andean culture area from its beginnings. Therefore, it is crucial to be aware of this web of long interrelated histories in a wide and varied geographical background instead of ignoring them in favor of generalized Inca-centric approaches.

chapter 3

Historical linguistics, philology, and the development of Andean linguistics Rodolfo Cerrón-Palomino Translated from the original Spanish by Matthias Urban

L’herméneutique matérielle unifie l’herméneutique et la philologie dans une sémantique de l’interprétation. Si l’herméneutique matérielle n’est pas une philosophie, elle suppose cependant une épistémologie, une méthodologie et une déontologie. L’epistémologie est celle des sciences de la culture. La méthodologie unit a critique philologique et le comparatisme linguistique; elle suppose ou impose une conscience de la relativité historique. La déontologie est imposée par le caractère fondamentalement situé de l’activité interprétative; comme telle, elle n’échappe pas au problème de la responsabilité: ses deux principes immédiats sont le respect du texte, dans sa lettre comme dans son esprit, et la bienveillance dans la production du sens, pour créditer le texte et l’auteur des bonheurs de l’interprétation. (Rastier 2001: 101)

3.1 Preliminaries Historical-comparative linguistics in the Andean area, as in South America at large, was born in the 1960s. Its development is closely linked to the unfolding of descriptive and dialectological studies on the Indigenous languages, in particular the so-called “major” ones (in the Andean case, Quechuan and Aymaran). Thanks to these efforts, characterized by the analytic rigor which the discipline demands, researchers were better equipped to tackle the critical and principled examination of the colonial documentary record which, up to this point, had not been subject to a truly hermeneutical interpretation. In spite of that, diachronic studies initially privileged new and fresh materials from the field over the aged records of little-visited repositories and archives. In this context, there was a degree of contempt for the missionary grammatical production from colonial times, since it was considered, a priori and by prejudice, that such works were generally shaped in the mold of Latin grammar and were not reliable in the written representation. Work on the reconstruction and classification of the protolanguages started from this point, with written documents playing a rather vicarious role. Only later, once reconstructions had been proposed, and due to a certain dissatisfaction with aspects of some of them, did attention turn toward the documentary sources in search of evidence which would help to formulate alternative solutions, deal with

unresolved issues, and obtain clarification and refinements for which the mere consultation of synchronic and dialectal material was insufficient. In this manner, linguists specializing in the area, who had until that point been trained in the analysis of data obtained in their fieldwork, directed their attention to the examination and reappreciation of the colonial documentary record, either in manuscript or printed form. Thus, philological studies of the Andean area were initiated and employed in the rigorous intellectual examination of the colonial grammatical and lexicographical materials. The same, however, did not occur with respect to the no less important and valuable written production in Indigenous languages accessible at that stage. The interpretation and restitution of these texts remained outside the scope of the good intentions of historical linguists, and was left instead to scholars with backgrounds in disciplines outside of linguistics. In this manner, in the Andean area, a divide between the tasks of linguistically trained philologists and researchers from other disciplines emerged. Pro forma calls for interdisciplinary work, advocated with such insistence in the realms of the social sciences—simply combining common interests or, more dramatically, relegating the discipline of linguistics mostly to the status of a mere noninvolved onlooker—did not wholly succeed in resolving this divide. As the reader will note, in a certain sense I am here, cognizant of Baldinger’s (1987) celebrated article, concerned with the “splendor and misery” of the discipline of

Rodolfo Cerrón-Palomino, Historical linguistics, philology, and the development of Andean linguistics. In: The Oxford Guide to the Languages of the Central Andes. Edited by: Matthias Urban, Oxford University Press. © Rodolfo Cerrón-Palomino (2024). DOI: 10.1093/oso/9780198849926.003.0003

63

rodolfo cerrón-palomino philology as applied to the Andean world. In this contribution, I attempt to give an account of the development of linguistic-philological studies as they have come to be applied in the linguistics of the Central Andes. After offering a brief initial characterization of this matter, the contribution moves, as its title suggests, to reviewing in broad strokes the historical development of that scholarly activity, which I divide into two sequences: an initial one, ranging from early efforts that evolved in relation to the integration of historical linguistics and philology in scholarly activity, and a second stage, characterized by relaxation of the rigor demanded by the overall discipline, but which did not result in a total break from the previously initiated practices; this aspect is visible foremost in hermeneutical matters which are related to ecdotics and textual criticism. The critical and reflective evaluation which forms the main part of the contribution concludes by making a case for resuming the linguistic-philological work which has been begun under good and favorable auspices. Obviously, the continuation of the interdisciplinary approach, naturally observing in a rigorous manner the epistemological, theoreticalmethodological, and ethical principles which the relevant disciplines practice in their investigative activities, is urgent and indispensable.1 Before approaching the topic, it is necessary to agree on the concept of philology to be developed for its purposes, indicating its extent, objectives, and subject matter as far as it is put into the service of linguistics, especially in its historical-comparative dimensions. In doing so, I follow the classical distinction formulated by Yakov Malkiel (1953; 1968[1960]) with respect to the institutionalized practice of philology in the tradition of Romance studies. As far as the extent and objectives of philology (as I understand it here) is concerned, I need to state more precisely that I will limit myself to using the term in a narrow rather than broad dimension: in this sense, its task is the interpretation of the written documentation in service of the reasoned, strictly linguistic understanding of the history of a language and its different stages of development and, at the extralinguistic level, of the cultural and institutional history of the peoples who spoke it.2 The material studied in this context comprises every kind of record—whether in manuscript form or 1 The present contribution is a revised, augmented, and translated version previously published in Revista Andina 56, 101–28 in 2018 (CerrónPalomino 2018b). I thank Sergio Cangahuala Castro, my former student, for his invaluable help in rendering accessible a large part of the cited references which, since I was deprived of a large part of my library due to the quarantine imposed as a result of the pandemic at the time of writing, would otherwise have remained out of reach. 2 For an attempt to establish the kinds of problems which philological investigation faces, seen from what the author considers merits (“splendor”) and limitations (“misery”) of the discipline concerned, one can consult the excellent work of Kurt Baldinger (1987), with exemplifications from the tradition of Romance philology.

64

printed, literary or not, composed by a native speaker or dictated by way of an interpreter—that can give testimony of and information on the language and the period in which it was composed or deposited in court, whether directly or indirectly. Its scope is thus not necessarily restricted to works on grammatical or lexico-semantic subjects. In this sense, linguistics, especially historical linguistics, and philology come to be converging disciplines which mutually support each other as soon as both share an interest in the linguistic and cultural past of peoples. As my colleague Juan Antonio Frago points out: It is an urgent necessity that diachronic investigation treats the texts with the doubtlessly primordial importance which duly pertains to them in the elucidation of the diachronic issues; [in this sense] it would constitute scientifically suicidal demeanour to carry out historical reconstruction of a language while bypassing documentary sources where they exist. (Frago 1992: 33, my translation from the Spanish)

Now, this might appear to be a truism in contexts like those of the Old World, with languages that boast a millennialong written tradition and in which philology, in its wide sense, acted as the midwife of linguistics; not so, however, in the New World, where some languages at best enjoy a written record that begins only in the 16th century (with the exception of Maya, for which an Indigenous mixed syllabic-logographic writing system was available, cf. Campbell 1990). The situation was, accordingly, the opposite on the American continent, since here linguistics was born without its twin sister, philology. Concomitantly, there were worries about the quality of the available written records, including grammatical and lexicographical treatises, when compared with that of the materials collected in the field and later analyzed and set down in writing, with the latter generally taken to be more secure and trustworthy where the attested languages persisted. It is then in the context of the blossoming of descriptive linguistics in North America, which had its champions and most illustrious practitioners in Sapir and Bloomfield, that voices began to be raised calling the attention of linguists to the written records in Amerindian languages, which had not only been ignored but rejected on the basis of Eurocentric preconceptions. These voices denounced the negligence and the irresponsible inattentiveness on behalf of the practitioners of the moment, who, in bypassing such testimonies before their heuristic examination, instead practiced a version of the discipline alienated from its scientific congener. Such demands can be heard in the context of North American Amerindian linguistics in the 1970s, having Ives Goddard (1973) and Mary H. Haas (1978[1975]) as their most conspicuous representatives and practitioners, who demonstrated on the basis of examples the advantages

the development of andean linguistics of diachronic work on the basis of a fraternization of linguistics and philology. Thus, the former of these linguists would insist on the importance of the documentary records, especially of those documents that are written by Indigenous people themselves, since these “deserve more attention from linguists than they have been paid in the past” (Goddard 1973: 83). Mary Haas (1978[1975]: 178), in turn, was more optimistic: “The American Indian Linguist is no longer exclusively a descriptive and/or comparative linguist, he is often a philologist as well.”3 In relation to the philologicallinguistic approach as it was applied to Mesoamerican languages, the works of Canger (1990) on colonial Nahuatl and above all those of Campbell (1978; 1990), concerned with the likewise colonial textual records of Quiché Maya, are of singular value. Thanks to this documentation, these investigators could contribute information on different phonetic and phonological phenomena, among them the determination of segmental and prosodic units, the changes that operate on them, the recourse to these for the identification of extinct dialectal variants, and their ancillary potential as evidence for classification and subgrouping. All of these are topics on which information would be difficult to obtain beyond pure speculation by an application of the traditional methods of historical linguistics alone.

3.2 Philology and the Andean world In the prologue to their monumental bibliography on the Aymaran and Quechuan languages, the well-known Americanists Paul Rivet and George Créqui-Montfort (1951–1956) embraced, at the beginning of the second half of the 20th century, the hope that, with the copious documentary contribution they were facilitating, they would lay the foundations for undertaking historical-comparative linguistic work on the mentioned languages (cf. Rivet and Créqui-Montfort 1951–1956: xix). However, one would still have to wait one more decade until such studies were inaugurated, and not exactly thanks to the vast documentation assembled by these authors, as (the unequal value of the consigned documents aside) the largest part of it revolved, for reasons having to do with the research tradition, almost exclusively on the model varieties for both languages. These model varieties emerged as such already in colonial times, leaving varieties that diverged from such models to their fate if not 3 The relations between linguistics and philology was also a topic discussed in the Centro de Investigación de Lingüística Aplicada (CILA) of the Universidad Nacional Mayor de San Marcos in 1973; among the participants were none other than Gary J. Parker, one of the founding fathers of Andean linguistics, and also Enrique Ballón Aguirre and the present author (cf. Pozzi-Escot 1973). Given the time, any mention of the necessity of including philological perspectives in the diachronic studies of Andean languages was still absent from the debate.

to their extinction. As a consequence, it became imperative to direct attention precisely to those varieties that had been pushed aside until then, and for which, in the absence of major documentation, there was no other way to proceed than to carry out fieldwork in search for reliable and novel material that could be recorded on tape. This situation is certainly advantageous for the specialist whenever the use of the languages in question—Quechuan languages to a greater degree than Aymaran languages—still remain vigorous as an element of “control,” in contrast to what happened with extinct languages, without or with only sparse documentation, which is the case not only for Puquina, the third lengua general of ancient Peru (cf. Chapter 12 by Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar in this volume), but also for the regional and local speech of the Peruvian North Coast (see Chapters 11 and 14 by Matthias Urban in this volume). As far as the mentioned extant Andean languages are concerned, scholars who laid the foundation for their dialectal and comparative study proceeded by prioritizing oral sources in the procurement of fresh and diverse material, initially of Quechuan, in the first half of the 1960s, and then also of Aymaran, following in the 1970s. As far as Quechuan is concerned, I am referring to the works seeking to reconstruct the protolanguage (PQ), and the classification of the varieties that derived from it and which were proposed, independently of one another, by the linguists Gary Parker (2013[1963]), a North American and former student of Charles F. Hockett, and Alfredo Torero (1964), a Peruvian disciple of André Martinet. In this manner, strictly linguisticanalytical and reconstructive procedures that were applied in two distinct traditions converged in the field of Andean linguistics: the North American descriptive tradition and the French structuralist tradition, which yielded practically identical claims which the venerable historical-comparative method could vouch for. As far as diachronic studies on Aymaran are concerned, these did not enjoy the same luck, since, although they were announced in the same decade of the 1960s (cf. Hardman 1975a), they had to wait to be realized until the following decade (Hardman 1975b; Hardman-deBautista 1978). Until one could proceed to the reconstruction of the ancestral language (i.e. Proto-Aymara, or PA in the following) which has been proposed by the present author (cf. Cerrón-Palomino 2000b), one would have to wait until the last decade of the 20th century. After all, in this case, the linguistic and dialectal reality was completely different from that of Quechuan: not only do just two varieties, of unequal geographical scope, survive—Central Aymara or Jaqaru, in its small refuge in the highlands of Lima, and Southern Aymara in the vast Peruvian-Bolivian and Chilean Altiplano (see Chapters 8 and 9 by Matt Coler in this volume)—but also until that point there was no sufficiently

65

rodolfo cerrón-palomino abundant and reliable documentation of Jaqaru, material which only became available in the middle of the 1990s (cf. Belleza Castro 1995). What has been said until now is valid only for the two major languages of ancient Peru, not for the third language which enjoyed the same status: Puquina, which disappeared in the second half of the 19th century because of the Aymaraization or Quechuization of its speakers. Not only was the researcher confronted with a completely extinct language, but, more dramatically, given the failure to find traces of the possible grammatical and lexical documentation that has been attributed to the Jesuit priest Alonso de Barzana (or Bárcena), an excellent lenguaraz (cf. De la Viñaza 1977[1892]: 45), they had at their disposal only the religious texts that were compiled by the Franciscan Jerónimo de Oré (1607) which are partially indebted to de Barzana (see now Chapter 12 by Nicholas Q. Emlen, Arjan Mossel, Simon van de Kerke, and Willem F. H. Adelaar on Puquina).4 Having been confused with Uru (on which see Chapter 10 by Katja Hannß in this volume)—whose last shoots still survive in the Bolivian Altiplano around Oruro, where they are referred to by the treacherous name of Puquina—the two languages could definitely be distinguished in the 1960s, thanks to the work of Alfredo Torero (1965) on Puquina, which was presented as a thesis written under the direction of André Martinet. To be able to do so, the Peruvian linguist had to carry out a philological-linguistic interpretation of the pastoral material, extracting the fragments of the language’s grammar that underlie the texts as well as the lexicon that can be attributed to it after first purging it of the Quechuisms and Aymaraisms which are frequent in the texts (cf. Torero 1965). As far as Uru—which had never been the subject of documentation until the end of the 19th and the beginning of the 20th century—is concerned, all they had at their disposal until then were short wordlists, some of very dubious quality. Taken together, the materials at hand served to bring about the linguistic separation of Uru and Puquina (Torero 1992), thereby refuting the then very popular hypothesis of Créqui-Montfort and Rivet (1925; 1927), who postulated a genealogical relation between the two. With respect to Uru, whose documentation, as has been noted, only commenced toward the end of the 19th century, it was an urgent necessity to be able to rely on a grammatical 4 Considered as the third lengua general of ancient Peru, Puquina was not even taken into consideration as translation language by the Third Lima Council (1582–3), surely because, by then, the majority of its speakers already were Aymaraized or Quechuized to the extent that they could be evangelized in any of the languages concerned. This did not happen in the case of Mochica, as this language (mentioned in the conciliar documents as ) was taken into consideration as a target language by the translators (cf. Medina 1966[1904]: 17).

66

description of the only Uru-Chipaya variety that survives until the present: Chipaya, spoken on the steppes of Oruro in Bolivia. On the basis of fieldwork carried out in the first decade of the present century, we now possess a grammatical description of the language (Cerrón-Palomino 2006b) as well as a compilation of its lexicon (Cerrón-Palomino and Ballón Aguirre 2011). With Chipaya providing “checks and balances,” it was possible to tackle at a later point of time not only the “recomposition” or “restitution” of the first lexical records of Uru, which was accomplished in 1894 by the German scholar Max Uhle (Cerrón-Palomino 2008c), but also to offer a preliminary reconstruction of the phonological system of Proto-Uru-Chipaya (Cerrón-Palomino 2007). A similar approach of restitution of all Uru materials that have been recorded between 1894 and 1952, whether inedited or printed—among which the grammar of Iruhito Uru left behind by Max Uhle (1894), the vocabularies of Métraux, and the texts of Vellard stand out—was adopted by Katja Hannß. She attempted, not always successfully, to resolve the numerous problems of reading and interpretation which such material presents given its heterogeneity, corresponding to different documentary traditions with which its compilers were affiliated and of which some are more trustworthy than others (cf. Hannß 2008 and the exhaustive bibliographical references cited therein). Finally, and always utilizing Chipaya as the sourcebook in matters of reconstitution, not long ago I offered an interpreted version of the only Uru vocabulary of the variety of Ch’imu, from the Puno region in southern Peru, which was collected in 1929 by Walter Lehmann (cf. Cerrón-Palomino 2016b).

3.3 The written documentary record Before approaching the topic, I will classify the colonial documentary record, as I have suggested on an earlier occasion (Cerrón-Palomino 2009b), according to three broad categories: (i) circumstantial, (ii) textual, and (iii) properly linguistic. In the following, I will occupy myself briefly with each. Circumstantial sources are, as their name already indicates, those which supply all sorts of information, be it direct or indirect, with respect to the geographical distribution, spatial spread, and sociocultural status of the languages spoken in the Andean region when the Spaniards arrived in the 16th century. Such information, which belongs to various narrative genres, can be found dispersed in the chronicles, reports (relaciones), accounts, inspections (visitas), lists of parishes, registers of the Indigenous population, and tributary tax accounts. An unimprovable example of this kind of documentation is offered by the Relaciones Geográficas de

the development of andean linguistics Indias, compiled by Jiménez de la Espada (1965[1881–1897]), especially for Quechuan and Aymaran, but also the Memoria de las doctrinas (c.1600), made available by Josefina Ramos Cabredo (1950), for the languages of the North Coast, and, above all, the Copia de curatos y doctrinas (c.1600), for the distribution of Puquina, Aymara, and Quechua in the Altiplano region at the end of the 16h century, tracked down and studied by Thérèse Bouysse-Cassagne (1975). Sources of the textual kind comprise registers, whether pastoral or secular, written in an Indigenous language, be they reproduced from the oral traditions of the speakers, or collected and elaborated by Spaniards, criollos, mestizos, and Indigenous people in the colonial context. A paradigmatic example of oral literature—or rather a kind of Indigenous cosmovision, really unique of its kind—is the Huarochirí manuscript (c.1608), for which the native authorship is still debated (cf. Durston 2007a). Among the pastoral treatises, to mention only the oldest one, I shall single out the works (catechisms, confessionary, and collection of sermons) in a trilingual Spanish–Quechuan–Aymaran version, of the Third Lima Council (1584), the oldest printed documents from South America. In all these works, Quechuan and Aymaran are “intellectualized” to serve as the instruments of catechesis. Petitions and notarial (and even personal) protocols, written by Indigenous people, are also an important source of this type, and in recent times have increasingly been recalled from archives and judicial files (cf. Itier 1991; 1992c; 2005; Taylor 1985; Adelaar and Trigoso 1998; Durston 2013). Moreover, as far as Aymara is concerned, one has to mention the unique case, as far as is known up to now, of Martín de Santa Cruz, who was personally trained in the colonial town of Juli (Puno) in writing his language by the illustrious Aymaran scholar Ludovico Bertonio, who also translated the Vita Christi into Aymara (Bertonio 2014[1612]). According to Xavier Albó, his modern editor, the treatise is “the most accomplished of Bertonio’s Aymara texts” (Albó 2019: 117, my translation from the Spanish). Finally, sources that are properly linguistic in nature include the grammatical treatises and the lexical records of Indigenous languages, principally of Quechuan and of Aymaran, whose documentation goes back to 1560 in the case of Quechuan (cf. Santo Tomás 1994a[1560], 1994b[1560]), and to 1584 in the case of Aymaran (cf. Tercer Concilio Limense (1985[1584]: 78–84). In spite of being local or at most regional vehicles compared to Puquina, we include grammatical treatises on two further languages, one of them regional and the other local: I am referring to Mochica and Cholón, described by Fernando de la Carrera (1644) and Pedro de la Mata (1748), respectively (see Chapter 11 by Matthias Urban and Chapter 13 by Astrid Alexander-Bakkerus in this volume for sketches of the languages on the basis of these materials); of equal

importance for the study of these languages of the North Coast and the Northeast of Peru, especially for providing us a glance at the phonological evolution of Mochica, are the wordlists offered in the famous “Plan” of the languages spoken in the old bishopric of Trujillo, ordered to be compiled by the distinguished Bishop Baltazar Jaime Martínez Compañón (1985[1782–1790]) (see Chapter 14 by Matthias Urban in this volume for the relevance of this document for understanding former Northern Peruvian linguistic diversity). For the philologist and historical linguist, the two latter types of source obviously bear the highest importance for use as comparative and reconstructive source material; at the same time, the circumstantial documentation, which is no less valuable from a historical point of view, serves as a reference point to establish the localization and geographical spread of the languages involved as well as the boundaries of languages and language varieties. It remains to indicate that the three types of source that have been mentioned are not mutually exclusive, as, taken together, they contribute to the integral reconstruction of the linguistic and cultural history of the pre-Hispanic peoples. At any rate, for an efficient utilization of such materials, one must bear in mind their authenticity, their veracity, and their integrity before rigorous ecdotic and hermeneutic examination. Juan Antonio Frago reminds us of this when he insists that to tackle the documents and later interpret them, one has to “possess a good understanding of the history of the language in which the document under scrutiny is written, without having to discard from this diachronic knowledge either a refined paleographic technique or the comparative base which would be most suitable for the analysis undertaken” (Frago 1992: 63, my translation from the Spanish). As we will see later, an efficient handling of the analyzed document requires, in case of the Central languages that are still vigorous, a precise and well-informed knowledge of their history, evolution, and dialectology.

3.4 Agreements Once the ancestral languages of Quechuan and Aymaran had been reconstructed, as a result of this comparative work, consultation of the written documentary record—which, if not employed in an ancillary fashion, was postponed up until this point—became urgent to verify or clarify certain aspects for which external comparison or internal reconstruction alone turned out to be insufficient. In this manner, philological investigation was naturalized in this part of Andean linguistics with the objective, among others, to specify, and eventually explain, certain evolutionary processes of

67

rodolfo cerrón-palomino the languages and to resolve some problems that hitherto lacked a solution (e.g. the possibility of postulating the relative chronology of certain changes, and allowing in this manner an evidently more detailed comprehension of the involved issue). Thus, consultation of written sources, examined and interpreted critically, contributed decisively to a better and more detailed intellection of the evolutionary development of Quechuan and Aymaran, from both the formal and the semantic point of view, and in their immanent shape as well as in their sociohistorical institutional development.

3.4.1 Diachronic questions In this section I will offer, with examples, discussion of certain aspects of the evolution of Quechuan and Aymaran. Although some of these have been suggested and characterized in their general outlines by means of comparative work, formulated most of the time on the basis of inferences, thanks to the philological contribution it became possible not only to trace their operation empirically and in detail, but also to establish the relative chronology of their actuation. The phenomena addressed are, for Quechuan, the nature of the sibilants (Section 3.4.1.1), the weakening of the occlusives in coda position (Section 3.4.1.2), and the structure of the possessive phrase in coastal Quechua (Section 3.4.1.3), and, for Aymara, the role of stress in the process of vowel elision (Section 3.4.1.4), the evolution of accusative marking (Section 3.5.1.5), and the development of certain verbal derivational suffixes (Section 3.5.1.6). Incidentally, as will be seen in the discussed examples, it will be convenient to bear in mind that a “forgotten synchrony”, to employ a favorite expression of Rastier (2001), intractable to the documentary information, would let invaluable data (such as I will present shortly) escape attention.

3.4.1.1 The Quechuan sibilants There are Quechuan varieties which oppose two sibilants, /s/ and /sh/ (dorsal and palatal) on the one hand, and /s/ and /ś/ (dorsal and apical) on the other, as opposed to the rest of the varieties which only register the first member of the pairs, i.e. /s/. Therefore, it was unproblematic to postulate such a distinction for PQ; as a consequence both Parker and Torero reconstruct */s/ and */sh/ as the respective protophonemes, applying the “majority rule” in favor of the varieties which make a distinction between a dorsal and palatal sibilant (those in which a dorsal and an apical sibilant are opposed form a minority). This decision, however, was made without regard to the colonial documentation, leaving out of consideration the fact that in the first documentations

68

of Quechuan, which describe exactly the varieties which had lost the opposition, the distinction was still captured systematically, and in complementary distribution, by means of a coherent orthographic distinction between ~ on the one hand and ~ ~ (the latter before and ) on the other. The problem now was to identify the value of the first set of letters, but not so much that of the second, given that there was no doubt that one was dealing with a dorsal sibilant. Would it be a palatal sound, as Parker and Torero apparently thought? As it happens, both the anonymous author of a colonial Quechua grammar and vocabulary (2014[1586]: 361), whom we can now identify as no less a figure than the Chachapoyan mestizo Blas Valera and Gonçález Holguín (1952[1586]: “Al lector”), the first grammarian of Cuzco Quechua, categorically indicated that the varieties they described lacked , a letter which by then still represented the palatal sibilant /sh/ in Spanish orthography. In light of such documentary evidence, first Landerman (1982) and later Mannheim (1987b) proposed the revision of the postulation of */sh/ in favor of apical */ś/ for PQ, a proposal which today is beyond discussion.5 On the other hand, thanks to the textual evidence, one can likewise mention with complete confidence that the merger of the sibilants which distinguishes Southern Quechua from the remainder of its congeners, and whose inception seems to announce itself already between the end of the 16th and the beginning of the 17th century, reached its endpoint toward the end of the 17th century and the beginning of the 18th. This chronological dating would have been impossible to determine when not considering the documentary records. Incidentally, the same could be said about southern Aymara, which by the same period completed the merger of /sh/ and /s/, inherited from PA and still preserved in Jaqaru (cf. Cerrón-Palomino 1999a; see also Chapter 9 by Matt Coler in this volume).

3.4.1.2 Weakening of consonantal codas in Cuzco Quechua One of the properties which define Cuzco-Bolivian Quechua and set it apart from the rest of the southern Quechuan varieties is the weakening of consonants in syllable-final position. Thus, words like rapra ‘wing,’ utqa ‘quick,’ pichqa ‘five,’ chakra ‘seed plot,’ chuqllu ‘sweet corn,’ quimsa ‘three,’ and allqu ‘dog’ appear in as [raɸra], [usqh a], [pisqa], [t̠ʃaxra], [t̠ ʃoχλu], [kiŋsa] and [alqo], respectively. After studying the 5 For Julio Calvo Pérez (2013a: 37), the arguments marshaled in favor of the elucidation of the segments mentioned, and the following revision in the postulation of the sibilants that had previously been identified, would be mere Byzantine questions, and this he states in spite of embracing what is stated by Landerman. Unfortunately, throughout the “normalization” which he proposes, he interprets “fast and loose” the sibilants which the Dominican registers (cf. Calvo Pérez 2013b).

the development of andean linguistics behavior of such consonants in the dramatic works that represent the so-called Golden Age of Quechua, Mannheim (1990) succeeds in establishing the chronology of their erosion, which occurred toward the end of the 17th and the beginning of the 18th century. In this manner, an innovation which already cast a shadow from the second half of the 16th century onward, as Domingo de Santo Tomás appears to indicate (cf. Cerrón-Palomino 1990b: 349–53), reaches its culmination two centuries later—a phenomenon which can be traced thanks to the documentary record. Today we know, incidentally, that this change, with the exception of the nasal and the lateral, is the result of the agency of a southern Aymaran substrate, which is so powerful as to even define the Spanish of the region, as I have demonstrated elsewhere (Cerrón-Palomino 2003c; see also Chapter 16 by Luis Andrade Ciudad in this volume).

3.4.1.3 The possessive phrase in Coastal Quechua As is well known, the possessive phrase in Quechuan (in relating the possessor with the possessed or the whole with its part) is, with the particular exception of the Ecuadorian variety, characterized by double marking: the modifying element is marked by the genitive, which alternates in form between -p and -pa, and the modified, in turn, carries a reference marker which corresponds to the possessor or the whole on which the part depends. Thus, for example, the epithet of the Incas is inti-p churi-n ‘son of the Sun’ (literally ‘of the Sun its son’). Now, calculating the number of possessive phrases in the documentation of Domingo de Santo Tomás in the two parts of his Lexicón (Santo Tomás 1994[1560]b), and without counting obvious errors (of the type ‘planta del pie,’ i.e. ‘sole of foot,’ which is a linear calque from Spanish), we encounter 91 examples; 74 of these, or 81.3%, show cases of single marking (of the type {yutu-p hapi-q-ø}‘partridge hunter,’ {kuchi-p wawa-ø} ‘suckling,’ or {qucha-p pata-ø} ‘lake bank,’ and only 17 occurrences, or 18.68%, present complete forms with double marking of the type {wakcha-kuna-p wasi-n} ‘hospital for the poor,’ {piśqu-p qisha-n} ‘bird cage,’ or {paltay-pa yura-n} ‘avocado plant.’ Confronted with this very regular situation, the question is whether we are dealing with a grammatical trait of the described varieties which is not fully understood up to this point or whether it is simply the limited command of Quechua of Santo Tomás, as once suggested by Alfredo Torero (1990: 397), that is responsible for the alternations. That the Sevillan friar knew the complete double-marked forms is confirmed not only by the approximately 19% of correct forms which he cites, but also by the alternation between {runtu-p qillu-ø} and {runtu-p śungu-n} ‘egg yolk.’ One might therefore well be dealing with a dialectal trait of the coastal variety which he describes and which was distinct from the lengua general, with which he was also familiar (on which see Chapter 25 by César Itier in this volume). Perhaps this is an archaic trait that harks back to a stage that is attributable to the reshaping of Quechuan in the mold of Aymara (cf. Chapter 26 by Nicholas Q. Emlen in this volume).6 As far as we know, the only variety that shows dependent marking only in the possessive phrase was the Ecuadorian one, with which coastal Quechua has come to be connected.7

3.4.1.4 Stress and vowel elision The reconstruction of the accentual system of Quechuan and of Aymaran has still not been tackled by the specialists—in part because the grammars that provide some information on the matter do so in a not very explicit, if not confused, manner. Although all modern varieties of both language families, with the exception of some of the central ones and also Chachapoyas Quechua (see Chapter 6 by Aviva Shimelman and Jairo Valqui in this volume), stress assignment was without doubt complex in the past. This is suggested not only by the vestiges of different patterns in some of the central varieties, but especially also because of the somewhat imprecise information in the documentary record (with printing errors as far as the respective accent marks are concerned). Aymara is susceptible to vowel elisions, and, although the rules governing it are still not obvious synchronically, there is no doubt that they are conditioned by the stress placement in the structure of the word. To give just one example, let us take one of the various cases that are mentioned in the “Annotaciones” to the Aymaran variety that was used as the target for the translation of the Third Lima Council’s Doctrina Christiana (Tercer Consilio Limense 1985[1584]). In this particular instant, stress placement has to do with the grammatical information that is necessary for the correct assignment of prominence to the verbal forms. One is told that “the 1st and 2nd person singular of the present” has “its accent on the antepenult” (1985[1584]: 79v, my translation 6 Note, however, that the toponymy of the Altiplano, characterized by a Puquina and Aymara substrate, is replete with examples of possessive phrases like the mentioned ones, as one can see in the examples Incancala {inqa-n qala-ø} ‘Inca rock,’ Vilcanota {willka-n uta-ø} ‘house of the sun,’ umanamaru {uma-n amaru-ø} ‘water snake.’ 7 In this context it is surprising that Julio Calvo Pérez has not noted the recurrence of the discussed phenomenon in all its dimensions and consequences; further, he tells us that in Domingo de Santo Tomás’s data “there are times at which this interweaving [double marking—RCP] is lacking partially […] or totally,” concluding that the Dominican friar “sometimes appears to be guided by dialectal differences which, however, appear rather in the guise of lexicographic fluctuations” (Calvo Pérez 2013a: 44, my translation from the Spanish).

69

rodolfo cerrón-palomino from the Spanish). In this way, actual conjugated forms of the verb muna- ‘to love,’ like {mun-tha} ‘I love’ and {mun-ta} ‘you love,’ must have been *[ˈmuna-th a] and *[ˈmuna-ta], respectively, before the elision of the post-tonic vowel and the subsequent displacement of the stress to the penult, which occurred automatically and was still free from any kind of grammatical conditioning. As can be seen, the documentary record clearly illustrates the evolution of the phenomenon, which otherwise would remain at the level of conjecture. On the other hand, as far as syncope in nominal expressions is concerned, one of the rules with which the distinguished Bertonio provides us is one according to which the “nouns ending in a, in the singular or in the plural, undergo syncope, if one of the possessives ʃsa or Pa follows” (Bertonio 1603: 332–3, my translation from the Spanish). Subsequently, he gives the following examples: {ut-sa} ‘our house,’ {uta-nak-sa} ‘our houses,’ and {utpa} ‘his house,’ {uta-nak-pa} ‘his houses.’ The rule does not apply when the noun is followed by the possessive markers of the 1st and 2nd person, as in {uta-xa} ‘my house,’ {utanaka-xa} ‘my houses’ and {uta-ma} ‘your house,’ {utanakama} ‘your houses.’ This unequal behavior of the possessive suffixes as triggers of vowel syncope, a process which was already complete in Jaqaru, was paving its way progressively in the language starting from a situation in which the nominal possessive paradigm actually does not allow the phenomenon. If the Lupaqa variety of Aymara, which is the one described by Bertonio, had the rule at some point, it appears to have paved its way paradigmatically at some point of its evolution, something which only could be known thanks to the documentary record.

3.4.1.5 Evolution of accusative marking in Southern Aymara At present, accusative marking in southern Aymara is of the subtractive type, as can be seen in the case of {t'ant'a-c ø mun(a)-c th(a)-c wa} ‘I want bread,’ where t’ant’a ‘bread’ loses its vowel as a marker of the accusative. This occurs only if the object appears in unmarked position, that is, before the verb; in other, especially pragmatically marked contexts, the final vowel of the object is restituted, as is the case in muntha t’ant’a-wa ‘I WANT bread,’ where the vowel of the object t’ant’a is automatically recovered. Given that Jaqaru registers -ha to mark the respective case, it is natural to postulate *ha as the Proto-Aymara accusative marker, as I in fact did on the occasion (cf. Cerrón-Palomino 2000b: 206–8). However, since the oldest documentation of the language, in the materials of the Third Lima Council (Tercer Concilio Limense (1985[1584])), does not yet record the subtractive accusative

70

marking, it is logical to ask oneself regarding the temporal frame of the change (an issue I had obviously put aside when relying only on the modern dialectal sources). Fortunately, in this case, as I have demonstrated in another place (Cerrón-Palomino 1997a: 234–6), the same “Annotaciones” of the Doctrina Christiana from the Third Lima Council furnish us with a valuable piece of data to provide an answer to the question. In fact, even if it is sure that the translators of the mentioned pastoral document opt for the zero marker of the accusative case, they do so to avoid the polymorphism which the evolution of the morpheme in question had generated, as can be deduced from the relevant passage: Some [of the experts on the language—RCP] say that for the transitional accusative one has to put, conforming to the vowel in which it ends, ha, he, hi, ho, hu. But here [in the text— RCP] this is never used and neither is it common among the Indians. (Tercer Concilio 1985[1584]): fol. 78, my translation from the Spanish)

Thus, one clearly sees that those who advised marking the accusative in the indicated form still perceived the rests of the inherited morpheme as a mere lengthening of the vowel of the nominal root or stem when functioning as patient; its timbre could correspond to that of any of the vowels /a, i, u/, including that of the allophones [e, o]. That is, illustrating the situation with examples, we would have had the following situation: ‘village,’ ‘father,’ ‘field,’ ‘silver,’ and ‘provisions,’ i.e. phonetically [marka-ː], [awki-ː], [yapu-ː], [qoλqe-ː], and [qoqo-ː], respectively. In these examples the thus was only an orthographic device to indicate the vocalic lengthening, a means to which much recourse was made in the Aymara documentation of the 17th century. In light of such examples we can be sure that the change from *-ha > ø had an intermediate stage in the form of compensatory vowel lengthening, which was still vigorous in the Aymara of the 16th century, and which was later lost to give rise to the situation that is currently observed (with zero marking in unmarked syntactic position). Now we can, with a certain persuasiveness, say not only that the loss of the marker in question was not abrupt but that its gradual erosion involved two processes: (i) the evaporation of the glottal consonant and (ii) the subsequent contraction of the vowels that came in contact into one long one, whose dominant timbre was that of the vowel of the theme, as occurs as well in similar cases of intermorphemic contraction. These two processes are otherwise recurrent in the language and persist until the present day wherever the suitable context for their application arises. Allowing the “Annotacion” which I just interpreted to escape our attention in the wrong moment, we would not have been in the

the development of andean linguistics position to give an answer to the question that was initially formulated.

3.4.1.6 Evolution of derivative suffixes There are three Aymara suffixes whose evolution I would like to trace by means of the evidence which the colonial documentation makes available. I am referring, on the one hand, to the deverbal suffixes -c ta ‘upward’ and -c t’a ‘momentaneous,’ and, on the other, to the denominative suffix -pta ‘transformative.’ These suffixes appear listed, in a somewhat confused, repetitive, and even merged manner, in the table which the illustrious Bertonio offers in the first edition of his grammar (Bertonio 1603: 268). Actually, we set eyes here on , , , , and . Now, regarding the former morpheme, which he presents followed by the verbal person marker of the 1st person, i.e. , he informs us that “it is composed with the nouns and it means become or begin to do that which the nouns indicate” (Bertonio 1603: 268, my translation from the Spanish), suggesting clearly that one is dealing with the transformative. he assumes to be bimorphemic (we will see that the assumed suffix is a fiction), and gives examples using the verb ma- ‘to go’ (which we know goes back to PA *maya-), such as , of which we are told that “one applies this verb, composed in this way, to trees and plants to signify that they are growing” (Bertonio 1603: 289, my translation from the Spanish); in my opinion, reference is made to the upward motion suffix. As far as is concerned, Bertonio (1603: 291, my translation from the Spanish) would tell us that “the nouns composed with this particle signify to become what the composed form says.” With this statement he again presents us the transformative morpheme, only this time it is preceded by the perifactive or contour motion suffix , which the grammarian does not succeed in identifying when it co-appears with (that is to say, with apocope and consonantal gemination from the sequence {thuqri-ku-q} and of {tuku-y ri-ku-q},without noting that the second form is a forteriori the product of the Quechuization of the term thuqri ‘provincial governor,’ which is from Puquina rather than from Aymara as I had previously thought (cf. Cerrón-Palomino 2008b: 63–73) and was already incomprehensible to the Quechuanists of the 16th century. Pa¨rssinen (2003: 239–53), among others, follows Rowe in making the same mistake, but also, as a complete surprise, Szemński (Montesinos (2009[1644]: 122), who seems to be in complete disagreement with historical linguistics and authentic philology. For his part, the famed anthropologist Tom Zuidema, who seems to build interpretive models of pre-Hispanic Andean society on the basis of fairly erratic consultation of the colonial Quechua and Aymara vocabularies, postulates one more Inca religious festival: the (Zuidema 2010: 175–6), without noting that we are dealing with a typical case of a phonetic group, with erroneous formal ligature on behalf of the scribe of the expression , in place of , an Aymaraized version of the Puquina (for other similar cases of erroneous interpretations, sometimes performed in the tradition of historical studies, see Cerrón-Palomino 2020b). Finally, the third example which I offer also comes from the Finnish archeologist Martti Pa¨rssinen, who, in a recent article (Pa¨rssinen 2015b: 306), thinks of the as one of the three principal Aymara groups of the southern Altiplano, relying, as he claims, on Bertonio, but without citing him directly. However, the etymologies which he postulates for the names that would identify such groups (, , and ), with the exception of the third designation, turn out to be incorrect, since the Aymara element of the compounds, i.e. haqi ‘person, human being,’ has no bearing on e.g. the ethnonym , or, worse still, (for discussion of the point, see Cerrón-Palomino 2013c: 303–5). I note, incidentally, that the o spoke Puquina and not Aymara, a fact that does not exclude that they would later become Puquina/Aymara bilinguals. For similar distortions and arbitrary interpretations of the colonial text, see Cerrón-Palomino (2002a).

76

3.5 Disagreements While in the previous section I sketched the positive and encouraging panorama of the integrated development of historical linguistics and philology as applied to the Andean sphere in the last decades of the past century (see e.g. the reflection of Torero 1995a on the matter), in this section I should like to call attention to the state of pertinent scholarly activity in the current century up to this point. As in the previous case, the state-of-the-art survey which I should like to offer on this occasion focuses on work in historical linguistics proper as well as on the philological and ecdotic practice that is applied to the colonial documentary records. As will be seen, the offered panorama, especially in relation to the second topic, is not exactly as auspicious as one would have expected given its good start. It becomes necessary to attempt to explain the situation of stagnation, or even of regress—as paradoxical as it may seem—of the advances which had previously been achieved in the matter.

3.5.1 Reconstruction of protolanguages According to what has been mentioned, the comparative work on reconstruction and classification of the two major language families of ancient Peru culminated, in their substantial development, with the postulation of the basic structures of PQ and of PA. One of the great problems which remained unresolved—indeed a true challenge to the old comparative method as Campbell (1995) would indicate— lies in explaining the grammatical and partially also the lexical isomorphism between Quechuan and Aymaran. This had been the subject of an intense debate between those who, on the one hand, promoted a common origin which would account for the parallelism and those on the other who saw in it the result of ancient and repeated processes of convergence (see Chapter 26 by Nicholas Q. Emlen for more detailed review). At the present stage, the latter hypothesis is the one which, given the theoretical-methodological impossibility of proving that the two languages did not have a common origin, has become a kind of consensus view in the minds of the specialists of the area (see Cerrón-Palomino 2000b: 335–7; Adelaar with Muysken 2004: 34–6; Heggarty 2005). Moving forward, efforts should focus on trying to reconstruct the prehistoric scenarios, sociopolitically and culturally, in which the conditions that would have favored this convergence would have obtained. It is exactly in this direction that the recent works of Adelaar (2010b, 2012b) and Muysken (2012a), who seek to test interpretive models to help explain the grammatical parallelisms which both

the development of andean linguistics language families share, orient themselves (see now also Chapter 24 by Matthias Urban in this volume). They suggest the very plausible hypothesis that both the morphosyntactic re-accommodation and the lexical borrowings from one language into the other operated in two stages: the first is very old and affected the primordial pre-PQ, which came under the influence of the grammatical patterns of pre-PA, while the second, which operated in relatively more recent times, led primarily to lexical importations in the same direction, but this time pertaining only to the modern southern dialectal configurations of both families. In the lexical terrain, one of the recalcitrant problems which the comparative work has faced was the separation of the approximately 25% of the lexicon that both language families shared (see also Chapter 26 by Nicholas Q. Emlen in this volume). Once the existence of cognate elements in the non-cultural basic vocabulary was discarded, however, one important set of shared lexical items was left. The obvious principally cultural borrowings from one language to the other aside, the provenance of these items turned out to be difficult to establish and have, in the absence of pertinent etymological work, come to be designated, somewhat inappropriately, as “Quechumara lexicon.” Now, Adelaar (1986) suggested a methodology that was intended as an attempt to elucidate the provenance of this lexical conjunct and that relied on the idiosyncratic phonological and phonotactic properties of the inherited lexicon of both languages as a guide to separate that which is characteristic of each language family. Elaborating on Adelaar’s proposal, Nicholas Q. Emlen has recently delivered the results of his investigation which seeks to resolve the problem of the lexical separation and to establish distinct stages of linguistic contact in time and space, departing from the premise that the lexical heritage that is not shared would have to be attributed to the respective pre-protolanguages before their initial convergence. The diligence with which Emlen proceeds constitutes without doubt a commendable effort which overcomes certain methodological and analytic problems in the selection of the treated corpus and its interpretation. This work certainly contributes to a better phonological and lexical characterization of the respective protolanguages and induces a revision of some fundamental aspects of the postulates that have been formulated in the last third of the past century, such as the contentious issue of the origin of the laryngealized consonants in PA. This is not the place to evaluate the work in its entirety, which is something that would need to be done on another occasion; given the topic of this contribution, I only wish to call attention to one of the various problems that are related to the corpus selection and the formal and semantic etymologies that are postulated. I hasten to add, however, that the observations which I move on to formulate are the result of my own ongoing work which calls

into question the genuine character of certain lexical items that are attributed to Quechuan as well as to Aymaran. In fact, the investigations which are coming to be realized in matters of Andean linguistic prehistory show that a good part of the institutional and cultural lexicon of the Incas comes from Puquina, the third lengua general of ancient Peru (see e.g. Cerrón-Palomino 2015; 2016a; 2016f; 2017a; 2018c; 2020b), the same language which played a decisive role in the genesis and development of their empire. This demands a methodological requisite at the outset, to purge the Quechuan and Aymaran lexicon of all Puquinisms, which took root in the recipient language after adaptation, often a fortiori, to their sound system. This is a secure protocol to avoid falling prey to arbitrary postulations which attribute terms of intrusive provenance or which are, in the best of cases, calqued on the basis of a foreign mold, anachronistically in terms of the protolanguages.14 It is lamentable that Emlen’s work appears to illustrate this sowewhat erratic approach; thus, for example, *hukumari ‘bear,’ *mayruru ‘kidney,’ *qhapaqa ‘powerful, rich,’ *quyru ~ *quyllu ‘cloudy eye,’ are postulated for PA, and *awki ‘old man, grandfather, man of respect, spirit,’ *illa ‘amulet,’ *inti ‘sun,’ *quñu- ~ *quñi- ‘hot, heat,’ *wallpa ‘bird species,’ *wiraqucha ‘Andean deity,’ for PQ. All these items, as far as we know, are of Puquina origin and were plainly assimilated by both language families. However, on top of that, some of them are formally compounds, and nevertheless appear as if they were irreducible monads. Semantically, in turn, they bear glosses that are derived, if not extended and muddled with lately, but that are in any case in no way primary. Worse still, in at least two cases, *kurku ‘hunchback, hunch’ and *kintu ‘ritual offering of coca leaves, offering, small bundle,’ words are postulated for PA which are without doubt of Spanish origin and have been assimilated early by both languages, in the first case regressively from a source form *corcova and, in the second case, on the basis of the notion of the Spanish “King’s fifth.” Such 14 What some of my fellow linguists of the Andean area seem not to understand is the strong lexical and morphosyntactic influence (CerrónPalomino 2020a) which Puquina exercised initially on Aymaran and later on Quechuan, in distinct stages of direct and/or indirect contact with these languages. Such influences are notorious above all in the cultural, religious, and institutional lexicon of the Inca (see Itier 2023, who even prefers to invent Quechua suffixes, as I will demonstrate in some other place, so that he can persist in what I call the “quechuismo primitivo”, for this type of reluctance). Once this lexical patrimony was assimilated by both languages, it expanded throughout the Andean area with the speech of the conquering Incas and their legions of mitmaq and warriors, as toponyms which, having an Aymaran or Quechuan root, carry Puquina suffixes, also demonstrate, delimiting through their presence the spaces that were conquered by the Cuzqueños. Thus, when isolating the lexical stocks of Quechuan and of Aymaran in the service of the reconstruction of their respective protolanguages, one must not be fooled by the unanimous testimony of the lexical sources with respect to the evaluated terms, which in many cases might well be borrowings that were contributed by Puquina and generalized by the Inca civilization which mediated them.

77

rodolfo cerrón-palomino a manner of approach, which Parker (2013[1969]) had already lapsed into in postulating *illa, *inti, and *wallpa for PQ, in addition to the early hispanism *tayta (< *taíta) ‘father,’ could be perfectly explained by the incipient, if not to say virtually inexistent, state of onomastic studies on Puquina. However, there would be no place for lenience even in the present state of research in the mind of one of has come to insist (Cerrón-Palomino 2013b; 2013c) on the formal and semantic manipulation of the institutional lexicon of Puquina on behalf of the colonial Quechuanists and Aymaranists who, ignorant of the civilizing presence of Puquina in the Andean space, reduced the institutional lexicon wholesale to their respective plots of idiomatic competence. Here, as can be seen, philological, more precisely etymological work, has failed, but also, as far as language history is concerned, the recognition of the multilingual nature of Cuzco prehistory, and the language contact phenomena that arose from that multilingual history, has failed to be achieved.

3.5.2 Philological and ecdotic problems According to what has been previously indicated in Section 3.4.3, toward the end of the past century the anthropologist and historian Urbano celebrated the advent of a true revolution in the field of Andean studies, specifically concerning matters of comprehension and interpretation of the documentary sources. This revolution rests on the progress achieved in historical linguistics and philology— disciplines which he considered indispensable as hermeneutic tools in all findings regarding the pre-Hispanic past of the Andes, as Duviols (Itier 2016b) would also underscore some time later. In this sense, work undertaken by scholars with training in linguistics—like the classic works of Taylor on the Huarochirí manuscript, or those of Itier on the Quechua texts of the chronicle of Pachacuti, but also those of the historian Duviols himself, who researched Andean religion and the intents of its extirpation (Duviols 2003), which were familiar with and cognizant of the work of their colleagues in linguistics—appeared to definitely break with the prior practice of social scientists, especially historians, archeologists, and anthropologists. This prior practice involved the irresponsible and arbitrary handling of the sources, void of any philological learning; and yet the interpreters were swift to postulate interpretive systems of the sociopolitical and cultural reality of the preHispanic peoples on the basis of inappropriate and capricious readings of the sources which disrespected all ecdotic demands. Now, with little more than two decades passed since the Portuguese-born and prematurely deceased historian would

78

offer us this critical historical balance and programmatic interdisciplinary manifest, I believe that the practice he questioned seems to have deteriorated even further; a certain type of scholar without the necessary training in linguistics and philology is responsible for this state of affairs. The corpus delicti, if we may speak of it in metaphorical terms, was to be the recent editions of some of the written monuments of the Andean repertoire, basically chronicles and, in the particular case of which I will speak, two classical treatises on Quechua and Aymara.

3.5.2.1 Recent relapses in the editions of chronicles As far as editions of the corpus of Inca chronicles, traditionally in the hands of historians and lately literary critics, are concerned, one can say that there are numerous ones that could only to a limited extent deserve the qualification “critical,” since most of them are of poor or no reliability, lacking a rigorous ecdotic apparatus that would account, in a reasoned and explicit form, for the global and situational context of their discursive production. Generally, confronted with the abstention of true philologists who would be able to professionally recognize issues related to linguistic hermeneutics, the edition of chronicles will continue to suffer from the same errors which Arellano (1999: 55) indicated in his assessment of the texts in Ballesteros Gabrois (1985), published on the occasion of the quincentennial of 1492. According to Arellano, there are two fundamental aspects in which the editors of the chronicles of the Indies show glaring incompetence: (i) in linguistic terms, the “ignorance on vocabulary and phraseology of the times, which in turn frequently entails erroneous punctuation, or absurd notes” and (ii) in cultural terms, the absence of knowledge of the described reality, a fact that “provokes [a] series of fanciful notes which, instead of being helpful, confuse the reader” (my translation from the Spanish). As a result, it would turn out to be advisable to do without them, as paradoxical as that might seem. Now, if the just and apposite contestation that we just heard in relation to the editions of colonial documents which lack even basic philological rigor is directed principally to the exegesis of Spanish language texts, the situation is doubly critical when it comes to editions of documents that are redacted in Indigenous languages or which contain material of varying extent that appears in such languages. If specialists in Spanish philology in place are rare, the matter is much more dramatic in the domain of the Central Andean region. Recent examples of this procedural precariousness among the editions elaborated by persons who come from fields other than Andean linguistics and philology (and to

the development of andean linguistics mention just some) are those of the chronicles of Cristóbal de Molina (2010[1573]), prepared by Paloma Jiménez del Campo (which nevertheless has the advantage of offering a copy of the manuscript as facsímile), Polo Ondegardo, edited by Gonzalo Lamana Ferrario (2012), Cabello Valboa (2011[1586]), in the care of Isaías Lerner, and Ramos Gavilán (2015[1621]), of which Hans van den Berg and Andrés Eichman are in charge.15 In almost all of these the pertinent and erudite annotations on the Spanish text, which are to be welcomed with congratulations, contrast with the annotation on terms and expressions in Indigenous languages, which are generally supported by authors who are thought to be “philologists,” but who at best are no more than speakers of the native language in which they say they are experts.16 On my part I should like to indicate that the reason for that is not only the scarcity of linguists and philologists who specialize in the Andean world, but also a certain ideological and perhaps subconscious tendency among some hispanophile academics to disregard anything that forms part of the Indigenous legacy and, in the case of concern here, the language, as Father Acosta (1954[1588]: 519) already observed with reproach toward the end of the 16th century.17 A rare exception to such an ideological-linguistic attitude is one of the most brilliant students of the historian Porras Barrenechea, Carlos Araníbar, who recently vindicated the chronicler whose work appears to his master as that of a “savage” mind, in a monumental, diligent, and erudite edition, with a paleographic transcription of 15 Deserving mention is the edition of Santa Cruz Pachacuti Yamqui Salcamayhua which has been prepared by the linguist Rosario Navarro Gala (2007), since it is a work of self-contained paleographic transcription from the manuscript text, and not simply from a copy or microfilm of the manuscript (which naturally induces reading errors), with a meticulous critical apparatus to the text which calls to attention the different hands that intervened in its redaction, with their amendments, additions, and deletions, all of which are meticulously reproduced and compared with the original manuscript in hands. This is the kind of paleographic transcription which I would wish that linguists had at their disposal for all our handwritten documents, since only in this way will one be able to have confidence in the integrity of the edited text, with regard both to the passages written in Spanish and to those written in an Indigenous language. I understand, of course, that attention to philological details is currently not a preference of our social scientists. 16 An exception in this series of publications could be the next annotated edition of Sarmiento de Gamboa’s Historia Índica, prepared by Aleksín H. Ortega (2018), whose textual apparatus of the native language of the corpus comes supported by the available linguistic-philological information, although it too is not free of inevitable relapses to the traditional practice where the required pertinent information is unavailable. 17 And thus, to cite a paradigmatic example, we Peruvians have, among our “historiographical national heroes” as the semiologist Enrique Ballón Aguirre (2014: 34–5, my translation from the Spanish) calls them, someone who claimed that the chronicle of Huaman Poma de Ayala, and the Spanish in which it is ardously redacted, can only stem from “pure mental confusion” given the “disorder and barbarism in style and syntax” (Porras Barrenechea 1948: 7, my translation from the Spanish). This is describing someone who re-edited the first grammatical and lexicographical works on Quechuan, without understanding the language at all, as some of his third-generation followers continue to do to today.

the original and with a modernized version (cf. Huaman Poma de Ayala 2015[1613]).18 I can say the same of his major edition, assiduous like all the work he undertakes, of the complete works of Inca Garcilaso (2015). In both cases, one has to thank the author for the analytical indices and erudite and exhaustive onomastic glossaries which he offers, a truly indispensable source for the study of Andean language and culture, prepared during long years of patient dedication. For a summary appreciation of the work of the maestro Araníbar, see Cerrón-Palomino (2016d). The prevalent attitude among some Peruvian intellectuals toward native languages and culture is particularly evident in the way in which editors zealously retain archaic Spanish expressions and Latin phrases to avoid showing “uncultivatedness” while at the same time, when confronted with equivalent manifestations in Indigenous languages, they are condoned or bypassed in the absence of a counter-argument or a corrective canon, and even more so the absence of a philological tradition of the diminishing languages.19 As has been stated, the notable jump in the hermeneutic quality that was achieved in the interpretative works and editions of colonial Quechua texts in the 1990s were indebted not only to the informed scholarship of their authors in matters of historical linguistics, but also to the diligence applied to the cleansing of the lexicon of the language of the semantic contaminations that resulted from their manipulation for purposes of evangelization. However, this legitimate concern for semantic control over the religious and cultural texts is only one, albeit fundamental, aspect of what one faces in the heuristical work with the documentary material. Indeed, something unveiled by the etymological study of Andean onomastics in recent times is another type of manipulation of the language, apparently equally consciously but this time concerning its formal and semantic aspects, on behalf of the first lenguaraces and language experts on Quechua: Spaniards, criollos, and mestizos, but also Indigenous people who were officially recognized as authorities in the matter by the political 18 Something which is not worthy of the work, however, is the final appendix of the book, which presents a transcription, normalization, and translation of the Quechuan and Aymaran texts of the chronicler (cf. Huaman Poma de Ayala 2015[1613]: 511–48). This work was carried out by José Cárdenas Bunsen, an excellent exegete of the chronicle who, in this particular case, unfortunately does not appear to have dedicated the necessary diligence to the task. Not only is the appendix not exhaustive, as the title would seem to suggest, but it gives the impression that what is offered is a draft that is awaiting a complete revision, something that I hope will be done at some point. 19 It thus turns out to be natural to note that in the Peruvian–Bolivian Andean region there are, paradoxically, respectable translators, editors, and commentators of Graeco-Latin, and even classical Chinese works, but less so for Quechuan or Aymaran, languages which for such philologists would appear to continue to be barbarous and therefore lacking interest in the national memory.

79

rodolfo cerrón-palomino and religious administration of the Spanish crown. Without pointing out the profound linguistic complexity of the reality encountered at the time of the conquest, obscured and hidden under the treacherous veil of a Quechuabased lengua general which seemed to secure communication between the 200 ethnic groups which made up the Inca empire, such personalities—self-proclaimed authorities on the Indigenous language—initiated a massive drive toward coerced and arbitrary Quechuization of the Inca cultural and administrative vocabulary, which they hardly (or poorly) understood. However, as it did not conform formally to the language variety with which they were familiar, they proceeded to accommodate it as dabbling etymologists. This led to a procrustean formal and semantic treatment of the vocabulary under the assumption that the forms were incorrect or misrepresented in Spanish speech, with the objective of making them accessible and manageable in accordance with the idealized Quechua which these people created for themselves. That lexicon, Quechuized a fortiori, would come to be fixed and standardized as part of the institutional vocabulary of Inca civilization, and the Andean world as a whole. Historians like Inca Garcilaso are responsible for similar lexical institutionalization processes (cf. Cerrón-Palomino 2013c, 2019). We can say, however, that a good part of this repertoire is of Puquina, and to a lesser extent of Aymaran, origin, languages which the Incas traversed before their definite Quechuization (cf. Cerrón-Palomino 2013a; 2020a; 2020b).

3.5.2.2 Misguided editions The crass manipulations of the colonial texts in Indigenous languages, which have been noted by the few specialists in this area, keep being ignored or simply condoned by certain researchers. Lacking the good judgment required by linguistics, these figures consider themselves authorities in comparative and philological matters, and feel justified in etymologizing the cultural lexicon of the Inca, and even in editing some of the colonial classics, chronicles as well as linguistic documentation proper. Thus, the paucity of authentic specialists is exacerbated by the preponderance of work with a less robust basis. For example, Jan Szemiñski, in his recent edition of the second book of the Memoriales of the late and imaginative Montesinos, studies and analyzes the artificial and misrepresented list of 100 of the supposed rulers of ancient Peru (cf. Montesinos 2009[1644]: 293–349). Szemiñski is concerned more with the repetitions and erratic order of the names and their variants rather than with their etymology, without diminishing the value of some sparks of insight that occur in almost all his writings. Most concerningly, the editor

80

does not note the anachronisms nor the linguistic affiliations of the names that are contained in the list, in that he takes them all to be of Quechuan or Aymaran origin, whether they are misrepresented or not, whether they come from Puquina or not, and whether or not they contain hybridizations or not: instead he simply assumes that the etymologies of the colonial Quechuanists are set in stone. Two recent publications that are problematic in this regard are the editions of the chronicle of Cristóbal de Molina “el cuzqueño” (2008[1573]) and of the “Lexicón” of Father Domingo de Santo Tomás (2013[1560]) by Julio Calvo Pérez and Henrique Urbano. With respect to the edition of the Relación of Molina (2008[1573]), the coeditor responsible for the linguistic aspects of the work has extracted from the chronicle, as an appendix, a reasoned Quechua lexicon of the manuscript. Leaving aside the aspects that are related to the condition of the manuscript, with which I have briefly dealt with on an earlier occasion (Cerrón-Palomino 2016c), there are so many errors of interpretation in this edition that there is practically no entry that is free of ecdotic and heuristic mistakes. Indeed, Arellano's (1999: 55) comment on editions that “instead of being helpful, confuse the reader” applies here as well. I can say the same of the “critical edition” of the vocabulary of Domingo de Santo Tomás, the so-called “Indian Nebrija,” co-edited by the same specialists, in two large volumes: the first volume is preceded by a linguistic approximation to the work in which the original text is transcribed with significant errors and frequent misreadings. With respect to the second volume, following the declaration of the author, we are dealing with “a unique edition which is destined to fill a void for centuries” (my translation from the Spanish). The introductory study, in which the author proclaims his intention to analyze “in depth every entry of the work” (my translation from the Spanish), contains approximately 4,000 footnotes that refer almost entirely to his own work; it reveals gaps in understanding of basic aspects of Quechuan dialectology and historical linguistics.20 Here we have a clear example of how philological and linguistic work requires an in-depth understanding acquired through years of continuous study. Thus, the hermeneutic work also presupposes a responsibility “to confirm the

20 For the sake of illustration, here are some of his tenets: coastal Quechua would not have existed, and Quechuan and Aymaran would be “sister languages,” Santo Tomás would be the “creator of Quechua dialectology.” And, notwithstanding recognizing that he knows little of Quechuan historical linguistics, he embarks in disquisitions which explains his constant lapses. All of this is manifest in the arbitrary formal and semantic restitution of the entries, whereas really the ethical manner of proceeding would have been to seek the advice of any serious specialist of the discipline.

the development of andean linguistics vicissitudes of interpretation at the expense of the text and the author.”21

3.6 Epilogue As announced in the initial preview, my exposition has basically focused on offering a re-narration of the development of historical linguistics and philological studies in the Central Andean area, indicating within this development two periods that are clearly distinguishable by the quality of the works which have come to be realized within them: initially, from 1980 to 2000, it was characterized by a shared concern of historical linguists and philologists in a hermeneutical approach to the handling of the documentary sources, without neglecting the urgent necessity of archival research; and in a second period, which corresponds to the present century, in which a waiving of the practices that have been initiated in the previous period can be noted, in particular with regard to aspects related to the edition and interpretation of texts that are redacted in Indigenous languages. It is not that the focused disciplinary approach would have ceased to be practiced, but that the methodology, inaugurated with good auspices, was relaxed in the hands of practitioners who lacked good professional judgment both in linguistic-philological matters and in fundamental ecdotic questions. Thus, practices that had been considered largely overcome have returned to contaminate the disciplinary and academic area, sowing bewilderment and confusion not only among those who by their education are not in a condition to separate the wheat from the chaff but also among the specialists of the area themselves. Far from seeking to set in motion fluid communication between social scientists and literary critics, with shared concerns regarding the pre-Hispanic past and supporting themselves mutually where one requires the expertise of the other, it seems that blinkered work was accentuated, closing off disciplinary territories as if they were alien to each other. Thence the inevitable disagreements to which I have made reference and which result from not paying attention to what more than half a century ago the great Roman Jakobson (1960: 377) reminded us

when indicating that “a linguist deaf to the poetic functions of language and a literary scholar indifferent to linguistic problems and unconversant with linguistic methods are equally flagrant anachronisms.” What is true of men of letters and linguists can, naturally, be expanded to historians and archeologists who worked in a solipsist manner and ignored or condoned the contributions of their neighboring disciplines. In spite of everything, I harbor the firm conviction that, sooner or later, we will in the long run know to establish the collaborative work which is required by the adjoining disciplines, without discrediting the strict observation of its epistemiological, methodological, and deontological values, leaving behind the improvised work of those who prefer to continue to turn their backs on the achievements of the neighboring sciences. As stated in Section 3.3, this mission, and this time pertaining to the ambit of historians, linguistics, and philologists, must in some way leave behind the preoccupation with having at one’s disposal a “refined paleographic technique,” as the Romance scholar Frago reminds us. As we are dealing with colonial manuscript sources (mainly consulted by archeo-historians and ethnographers who were overtrustful of transcriptions carried out by wellmeaning paleographers who, however, lacked the skills of linguistic and philological investigation), errors in the reading and interpretation of texts, especially those recorded in Indigenous languages, will continue to haunt the editions of the colonial documents. Worse still, studies will continue to be carried out not only in Andean studies but also on topics of great validity and current relevance in Hispanic linguistics (such as those that concern Andean Spanish—especially its genesis and gradual propagation among bilinguals, Indians and mestizos alike—that can be traced through their written products which archival research is unveiling) on the basis of materials of such dubious paleographic treatment, and correspondingly bastardized content.

21 Since I am already speaking about classic vocabularies, I cannot avoid mentioning the recent publication of a “normalized” facsimile edition in two volumes of the Arte y Vocabvlario of the anonymous author (2019[1586]) which was edited by the Third Lima Council. This really is another senseless edition. To beat everything, the “normalized” and “modernized” version of the text was commissioned from a person completely foreign to Quechuan but who, being unprepared for the task, did not hesitate one moment to compound the offence by taking over, completely and unscrupulously, the normalized and modernized text which I myself had offered some years ago in my edition of the work (cf. Cerrón-Palomino 2014): a scandalous case of plagiarism, as Raúl Bendezú (2020) shows in the review of the text.

81

PART II

Language profiles

chapter 4

Huaylas (Ancash) Quechua Carlos Molina-Vital

4.1 General information Huaylas Quechua [HQ, ISO: 639-3 qwh, Glottocode: huay1240] is a dialect of the Ancash Quechua language [Glottocode: huay1239]. People refer to this language mostly as qichwa. Only a few people call their language nuna shimi ‘language of the people.’ The Huaylas variety of Ancash Quechua is one of the bestdescribed Central Quechua varieties, and has been studied particularly during the initial period of modern Andean linguistics (e.g. Parker 1967 and 1973a). Before that point, HQ terms were included, albeit not systematically, in the first Quechua grammar and lexicon by Domingo de Santo Tomás (1560a). Its author served as a priest in Yungay, and some terms from this variety of Ancash Quechua, known then as the Chinchaysuyo language, appeared in those foundational works. Another colonial source is the 1700 re-edition of Torres Rubio’s Arte de la lengua Quichua (1st edn 1619). There, Juan de Figueredo supplements the original, a description of Cuzco Quechua, with a list of lexical items and phrases from what is mainly, but not exclusively, an Ancash Quechua variety (Cerrón-Palomino 1987b: 89). More than 200 years later, the Vocabulario Poliglota Incaico (1905) included Ancash Quechua alongside the Quechuan languages from Cuzco, Ayacucho, and Junín. However, the Ancash variety it documented was the Southern Conchucos one, from the Huari province (Cerrón-Palomino 1987b: 386).

4.1.1 Geography, environment, and economics During early colonial times, administrative documents already attested the widespread presence of the two ethnic groups who spoke the main dialects of Ancash Quechua: the Huaylas and the Conchucos (Chirinos 2001: 54). The Huaylas variety is spoken along the Callejón de Huaylas (lit. ‘Huaylas alley’) region, a stretch of temperate interAndean valleys located between the eastern Cordillera Blanca (‘white-mountain range’) and the western Cordillera Negra (‘black-mountain range’). This fertile area is irrigated by the Santa river, which runs northwards from its origin in the

Conococha lake in the southern province of Bolognesi. This river steadily lowers in altitude across the whole Callejón de Huaylas, which makes the northern provinces’ valleys warmer in climate. From north to south, the following provinces make up this region: Huaylas, Yungay, Carhuaz, Huaraz, and Recuay.1 The capital of the region and center of the administrative and political power is the city of Huaraz (population 118,836 in 2017). Economically, the Callejón de Huaylas is very productive and diverse. Due to the beautiful ice-capped peaks in the Cordillera Blanca and its many glacial lakes, it is one of the most visited destinations in Peru. However, more traditionally, this region is an important center for production of tubers and corn (for internal consumption and to supply the city of Lima), as well as fruits and flowers (particularly blueberries). It also has a significant milk production in the lower-altitude provinces. Intensive mining of silver, copper, and zinc has changed the social dynamics between Quechua communities, the Peruvian government, and mining corporations.

4.1.2 Speaker population and endangerment Ancash Quechua, as a whole, is the largest Central Quechua dialectal area in terms of speakers. It is spoken by no more than 900,000 people (Pozzi-Escot 1998; Julca Guerrero 2010). However, among them, the majority speak the northern and southern Conchucos variety, which is one of the most vital Quechuan varieties in Peru when it comes to number of speakers (approx. 161,000 according to Ethnologue 2022) and transmission rates. The 2017 census shows that the lowest percentage of speakers for that variety is 79.92% in Huari, and the highest at 92.54% for Mariscal Luzuriaga (Instituto Nacional de Estadística e Informática 2018). In comparison, Huaylas Quechua (HQ) appears to be a threatened variety. Its 1 Julca Guerrero (2010: 280) indicates that the Huaylas variety includes also a Huaylas-Vertientes variety spoken in the areas right outside the Callejón de Huaylas, specifically the vertientes or areas located along the western slopes of the Cordillera Negra. This grammatical sketch does not include data from the Huaylas-Vertientes variety, only from the one spoken in the Callejón de Huaylas.

Carlos Molina-Vital, Huaylas (Ancash) Quechua. In: The Oxford Guide to the Languages of the Central Andes. Edited by: Matthias Urban, Oxford University Press. © Carlos Molina-Vital (2024). DOI: 10.1093/oso/9780198849926.003.0004

85

carlos molina-vital Table 4.1 Number of speakers of Huaylas Quechua in the Callejón de Huaylas provinces, Ancash region Province Population

% of Spanish speakers (age 5 and over)

% of Quechua speakers (age 5 and over)

Number of Quechua speakers (age 5 and over, rounded up)

Huaylas

46,649

42.50

57.05

26,613

Yungay

46,426

24.49

74.85

34,750

Carhuaz

15,855

27.60

71.78

11,381

Huaraz

149,713

60

39.11

60,050

Recuay

15,855

55.88

42.57

6,749

Source: Instituto Nacional de Estadística e Informática (2018).

total number of speakers has declined since 1940, when 80% of the population was Quechua-speaking (Julca Guerrero 2010: 63). However, HQ can be considered a vital language. Children still learn it and most provinces in this region display a clear majority of speakers learning it before Spanish. Table 4.1 summarizes the findings of the 2017 National Census regarding number of speakers in the provinces where this variety is spoken. The total number of speakers of a Quechuan language who are 5 years and older in the provinces where HQ is present is approximately 139,543. It must be noticed that this number does not represent the actual amount of HQ speakers, but an approximation. This is because Quechua speakers in the eastern districts in the provinces of Huaylas (parts of Yuracmarca, Santa Cruz, and Caraz), Yungay (Yanama), and Carhuaz (San Miguel de Aco) belong to the Conchucos variety of Ancash Quechua.2 Besides Julca Guerrero’s (2010) comprehensive dialectological survey of this region, little work has been done for those specific varieties.3 The presence of Quechua speakers is strong in many small rural towns located along ravines and ridges between 2,100 and 3,000masl. There it is still possible to find monolingual speakers of the Huaylas dialect. They tend to be elderly men and women, but if younger, it is almost exclusively women who remain monolingual. This is due to their more conservative economic role. Also, women still make up most of the illiterate population in the Central Andes, due to traditional prejudice against their education (cf. Julca Guerrero 2010: 2 Since the beginning of the 21st century, the Conchucos variety has been object of several specific studies by Daniel Hintz (2011; 2020), Diane Hintz (2003; 2006; 2007), and both together (Hintz and Hintz 2017). They have covered aspect, tense, modality, word order, and evidentiality. 3 An exception is the work of Francisco Carranza Romero (1979; 1998; 2003). He has produced several works based on his native Quechuan variety spoken in Quitaracsa, a village located in the Yuracmarca district, on the northeastern slopes of the Huaylas province. While the information in Carranza Romero (2003) shows that this variety is closer to the northern Conchucos one, it still has many lexical and syntactic commonalities with the Huaylas variety.

86

270–77). In any case, it is evident that HQ is threatened by the widespread use of Spanish (Menacho López and Villari 2019).

4.2 Internal variation and classification Ancash Quechua has been divided into four dialectal zones. Several studies (Torero 1974; Parker 1976; Landerman 1991; Julca Guerrero 2010) agree in dividing it into two main varieties, Huaylas and Conchucos, and two minor ones, the Sihuas-Corongo and the Bolognesi-Vertientes dialects. Within HQ, based on phonological, morphological, and lexical differences, Julca Guerrero (2010) distinguishes two areas: Southern and Northern Huaylas. The Southern variety is spoken in the high-altitude regions of the Callejón de Huaylas, in Recuay, Huaraz, and the southern extreme of Carhuaz, in the Marcará district, delimited by the river of the same name, and the Huasca Urán ravine. This dialectal zone also includes the communities located in the highlands of Carhuaz and Yungay, above 3,500masl. The Northern variety is spoken in the lower parts of the Callejón de Huaylas, in the provinces of Huaylas, Yungay, and Carhuaz (with the exception of its southernmost part). Its northern boundary is the area where the Cordillera Blanca and the Cordillera Negra meet: the Cañón del Pato (‘Duck Canyon’), where the Santa river splits the mountains moving westwards into the Corongo and Santa provinces. Based on phonological isoglosses,4 Southern and Northern HQ can be clearly distinguished because in the former 4 Relevant morphological isoglosses will be pointed out when specific suffixes are introduced in this grammatical sketch. By way of illustration of lexical differences, Northern Huaylas speakers say alla: pi ‘much,’ which is alla:pa in the Southern variety. Likewise, ‘dirty’ is qacha in Northern Huaylas and rakcha or taqra in the South, while ‘to dance’ is tushuy in the North and qatsway in the South.

huaylas (ancash) quechua the glottal fricative /h/ is used in word-initial position, while in the latter that segment has disappeared in that position. Thus, hirka ‘mountain’ and hatun ‘big’ in Southern Huaylas correspond to irka and atun, without initial [h], in Northern Huaylas. In addition to this, there are alternations for words that can be reconstructed with Proto-Quechua */s/ in wordinitial position. Thus, reconstructed forms such as *sacha ‘tree,’ *samay ‘to breathe,’ and *saqiy ‘to leave (something)’ became hacha, hamay, and haqiy in Southern Huaylas, and acha, amay, and aqiy in Northern Huaylas, where reduction of sibilants to Ø operates. However, it must be noticed that in both HQ varieties many terms maintain /s/ in wordinitial position. For instance, sutay ‘to pull,’ siqiy ‘to bring down,’ and suwa ‘thieve’ are rendered with an initial [s] in all of HQ. Finally, the other sibilant, the alveopalatal fricative */sh/, has experienced irregular changes in the Huaylas variety. For instance, shuti ‘name’ and shutu ‘drip’ are produced with their initial [ʃ] in the Northern Huaylas variety, but they are huti and hutu in the Southern one (Julca Guerrero 2010: 153–4). The retention of /sh/ for those words in the Northern dialect goes against its tendency toward the lenition and eventual suppression of sibilants. Conversely, other terms that show Proto-Quechua */sh/ became Ø in Northern Huaylas. *shiqta ‘mustard’ and *shuk ‘one’ became itqa and uk in that variety, while in the Southern variety the initial segment changed into /h/ (Julca Guerrero 2010: 155).

4.3 Phonology and phonotactics HQ has one of the most innovative phonologies within Central Quechua. Three distinctive sound changes in this variety are (i) diphthongs that result in monophthongied long vowels and, in some cases, middle long vowels (e.g. /aw/ → [oː], /ay/ → [eː]), (ii) the existence of a voiceless alveolar affricate /ʦ/ contrasting with a voiceless alveopalatal fricative /ch/ (e.g. tsaqa ‘dawn, dusk’ and chaqa ‘cup made of a gourd’), and (iii) widespread depalatalization of */ñ/ into alveodental /n/ (e.g. *ñawi ‘eyes’ > nawi).

4.3.1 Vowels The system of three phonological vowels (/i/, /a/, and /u/), common to every Quechuan language, is at the basis of HQ. However, as is the case for Central Quechua, the full set of vowels also includes their corresponding long forms (/iː/, /aː/, and /uː/). Due to the continuous contact with Spanish,

which led to widespread bilingualism between that language and HQ, it is common to find speakers who produce mid vowels like [e], [eː], and [o] with full regularity when using loans from Spanish. In any case, the native HQ vowel system has six vowels, short and long ones. Table 4.2 provides the inventory.5 Table 4.2 Vowel inventory of Huaylas Quechua Front High Mid Low a

i (e)

a

Central

Back

iː

u

u:

(e:)

(o)

(o:)

a

a:

Vowels from Spanish loans are in parentheses.

The high vowels /i/ and /u/ tend to have a lax midhigh realization close to [ɪ] and [ʊ], respectively. As in other Quechuan languages, the presence of the postvelar stop /q/ brings out a coarticulation process by which high vowels are produced like mid vowels [e] and [o] or mid-open vowels [ɛ] and [ɔ] respectively. Thus tsiqlla ‘waist’ becomes ['ʦɛqʎa] and shunqu ‘heart’ becomes ['ʃɔŋqɔ]. Notice that sonorant consonants (approximants, nasals, liquids, and rhotics) in between /q/ and a high vowel are no obstacle to the lowering process. This can be seen in tsiwqa ['ʦɛwqa] ‘teapot,’ sinqa ['ʃɛŋqa] ‘nose,’ ullqu ['ɔʎqɔ] ‘male,’ and murqu ['mɔrqɔ] ‘rusty.’ Compared to Southern Quechua, HQ sibilants appear to block vowel lowering (e.g. isqun ‘nine’ is mostly ['isqɔn]). Also, /i/ and /u/ are not always lowered in word-final position, something quite common in Southern Quechua. Moreover, some speakers do not lower them significantly even right after /q/ at the end of the word (thus, it is possible to hear ['ʃɔŋqʊ] for shunqu ‘heart’). Short vowels in HQ (like in other varieties of Central Quechua), can be clearly perceived as shorter than their unstressed Spanish counterparts. This seems to be even more pronounced with /a/, the least similar vowel to its Spanish equivalent, particularly syllable-finaly, wordfinaly, and before /m/ (Swisshelm 1971: 19; Parker 1976: 50–51). 5 The official orthography for Quechuan languages spoken in Peru has been established through the Decree no. 1218-85-ED, issued on November 18, 1985 by the Ministry of Education. The official orthography differs from the one used in this volume only in that long vowels are represented with a colon (, , ), instead of double vowels (, , ). Also, sounds that come from Spanish loans are written with their corresponding Spanish consonant if it is one not available in the native HQ inventory. Regarding Spanish vowels, loans are written using the native three-vowel system, including lengthening for stressed Spanish vowels, only if present in the corpus source. For instance zapato ‘shoe’ will be rendered sapatu, and vida ‘life’ will be bida. See Table 4.4 for the consonant inventory.

87

carlos molina-vital Table 4.3 Phonemic contrast between short and long vowels in Huaylas Quechua /a/

/a:/

/i/

/i:/

/u/

/u:/

Word initial

ashiy ‘to look for’

a:shi ‘they say yes’ (idiomatic expression)

ikiy, ikay ‘to chop (food) in little pieces’

i:gus ‘fig’ (from Spanish higo)

ushan ‘she finishes it’

u:shan ‘her sheep’

End of syllable

qara ‘skin’

qa:ra ‘plant similar to agave’ (Furcraea andina)

mitu ‘mud’

mi:tu ‘forest papaya’ (Vasconcellea candicans)

pukay ‘to turn red’

pu:kay ‘to blow’

Long vowels in HQ have an uneven distribution. In my corpus of approximately 100,000 words, /a:/ is found 16,149 times, while /i:/ and /u:/ are found only 2,375 and 423 times, respectively. Long vowels can open or close a syllable, but they never appear before a consonant in the same syllable. In other words, syllables ending in consonants only accept short vowels. Consequently, some roots or suffixes ending in a long vowel will frequently be rendered with a short one (e.g. puri-yka:-lla-n ‘he is just walking’ will shorten the continuous aspect marker -yka: in puri-yka-n ‘he is walking’). Also, long vowels are seldom found in word-initial position. When this happens, it is either in a Spanish loan or a word based on onomatopoeia (specifically /a:/ expressing agreement, assertion, or an affirmative answer). In any case, there are (near-)minimal pairs showing the phonological status of long vowels. The particularly short quality of /i/, /a/, and /u/ can be considered the main reason why Spanish loans that are fully incorporated into HQ use a long vowel corresponding to the one in the stressed syllable in the source. Table 4.3 includes i:gus, a clear loan from Spanish higo ‘fig.’ Other examples are caballo ‘horse,’ rendered as kawa:llu, and puro ‘pure,’ which is employed as pu:ru and functions as an augmentative modifier similar to ‘very.’

4.3.2 Consonants As can be seen in Table 4.4, HQ has 17 phonemic consonants in its native inventory. In addition to those, five more consonants originating from Spanish loans are also considered (included between parentheses): While most of the native consonant sounds are fairly similar to the Proto-Quechua inventory, some changes have occurred. Perhaps the most noticeable is the change from the Proto-Quechua voiceless alveopalatal fricative */ch/ (also

88

represented as /̌c/ in the Quechuanist literature) into the current voiceless alveolar affricate /ts/. This sound contrasted in Proto-Quechua with the voiceless retroflex affricate */ćh/ (usually represented as /ĉ/ in the literature). Furthermore, the latter sound lost its retroflex manner of articulation and became the alveopalatal affricate /ch/. Thus, Proto-Quechua pairs like *chaki ‘dry’ and *ćhaki ‘feet’ became respectively ʦaki and chaki in HQ. It is also worth noticing that HQ tends to keep /p/, /t/, /k/, and /q/ as stops in syllable-final position (i.e. they do not turn into a fricative). This can be heard in words like qapchi ['qapʧɪ] ‘thin,’ putka ['pʊtka] ‘turbid,’ tikpi ['tɪkpɪ] ‘pin,’ and suqpi ['soqpɪ] ‘loosely tied.’ Also, /r̝/, a raised post-alveolar non-sonorant trill (traditionally represented with /rˇ/ in the literature), is used to represent the Spanish voiced alveolar trill (e.g. carro ‘car’ is rendered as /'kar̝ ʊ/). This is fairly standard in Quechuan languages. The alveolar nasal /n/ has a rather restricted distribution as an alveolar. Only at the beginning of a syllable is it produced as [n], e.g. nina ['nina] ‘fire’ or nana ['nana] ‘pain.’ Also, a likely case of assimilation of point of articulation, /n/ remains [n] before the alveolar sounds /t/, /s/, and /ʦ/ as well as the alveolo-palatal affricate /ch/. Before /p/, /n/ assimilates to a bilabial nasal [m], e.g. mamanpis [ma'mampɪs] ‘also her mother.’ Likewise, before the velar and uvular plosives /k/ and /q/ it takes their point of articulation, e.g. yunka ['jʊŋka] ‘coast,’ qunqay ['qoɴqaj] ‘to forget.’ However, in several contexts where a conditioning environment for assimilation of the point of articulation is not apparent, /n/ is rendered as the velar nasal [ŋ]. This can be seen before other nasals, e.g. pupunmi [pʊ'pʊŋmɪ] ‘his navel (I know),’ laterals, e.g. pupunlla [pʊ'pʊŋʎa] ‘just his navel,’ rhotics, e.g. rinri ['ɾɪŋɾɪ] ‘ears,’ and semiconsonants, e.g. warminwan [waɾ'mɪŋwaŋ] ‘with his wife,’ tsunyaq ['ʦʊŋjaq] ‘uninhabited.’ Also, in word-final position, /n/ becomes [ŋ], e.g. tamyan ['tamyaŋ] ‘it rains.’ To sum up, alongside the

huaylas (ancash) quechua Table 4.4 Consonant inventory of Huaylas Quechua Bilabial

Labiodental

Alveolar

Post-alveolar ch

Voiceless plosive/affricate

p

t, ts

Voiced plosive

(b)a

(d)

Fricative Nasal

(f) m

s

Palatal

k

q

sh

Glottal

h ñ

Lateral

l

ll

Tap

ɾ

a

Uvular

(g)

n

Approximant

Velar

(r̝) y

w

Consonants from loans are in parentheses.

cross-linguistically well-known rule that assimilates a nasal’s point of articulation to that of the following obstruent (e.g. /np/ → [mp]), there seems to be another rule that connects the velar allophone [ŋ] with the continuous emission of airflow that characterizes sonorants and the syllable-final position. Finally, HQ has undergone a process of depalatalization of nasals (*/ñ/ > /n/) and, to a much lesser extent, lateral segments (*/ll/ > /l/). Thus, the palatal nasal /ñ/ in syllable-initial position has become /n/ for many lexemes throughout the Callejón de Huaylas. However, the change is not complete for those segments. In particular, it is not clear what conditions the change from */ñ/ into /n/. For instance, older forms like ñawi ‘eye’ or ñaña ‘sister (of a woman)’ have completely depalatalized in HQ and are nawi and nana. However, piña ‘angry’ and mañay ‘to request,’ among many others, remain palatalized. Still, there seems to be a distinction between Southern Huaylas, where depalatalization is more widespread, and Northern Huaylas, where forms with palatalized nasals alternate with their depalatalized equivalents (Julca Guerrero 2010: 121).

4.3.3 Syllable structure, root structure, and phonotactic constraints Syllable structure in HQ does not differ from the typical Quechuan one. A vowel is always the head, and it can be a syllable on its own (V), e.g. upa /'u.pa/ ‘deaf.’ The most frequent combination is that of a consonant in the onset (CV), e.g. pununa /pu.'nu.na/ ‘bed.’ More uncommon, but still widely present, are syllables with just a consonant coda (VC), e.g. uchpa ‘ashes,’ and ones with two consonants

surrounding the vowel (CVC), e.g. atuq ‘fox.’6 Notice that this syllable structure excludes CCV syllables. Thus, Spanish loans following that pattern are incorporated into HQ without the initial consonant (e.g. Spanish plátano ‘banana’ becomes latanus /la.'ta.nus/). The semiconsonants /y/ and /w/ behave like consonants for syllable structure. Thus upyay ‘to drink’ includes a CVC syllable at the end, and wawqi ‘brother (of a man)’ one at the beginning. As is customary for Quechuan languages, two vowels cannot appear contiguously, either in the same or different syllables. This is the reason why several words that can be reconstructed with an internal */s/ in Proto-Quechua ended up with a semiconsonant in HQ, e.g. *pusaq ‘eight’ > /'pu.haq/ > /'pu.waq/, *wasi ‘house’ > /'wa.hi/ > /wa.yi/).

4.3.4 Important morphophonological processes 4.3.4.1 Syllable structure and the suffix -ni ‘epenthetic (ep)’ If a suffix starting with a (semi-)consonant is added to a stem ending also in a (semi-)consonant, then a prohibited same-syllable CC sequence occurs. For instance, if yawar ‘blood’ were to take -yki ‘2.poss’ directly, it would yield impossible sequences such as *[ja.wa.'rj.kɪ], *[ja.war.'j.kɪ], or *[ja.war.'jkɪ], with CC, C, and CCV sequences respectively. To solve this, an empty-meaning suffix -ni is inserted as a “hinge” that helps restructure the syllable structure avoiding the forbidden sequences. Thus ‘your blood’ 6 Swisshelm (1971: 141–4) correctly noticed that /n/ and /s/ could be syllable heads in the final syllable. Thus, waqara: kun ‘she keeps crying’ can be rendered as [wa.qa.'ɾa:.kŋ], and nuqapis ‘me too’ as [no.'qa.ps]. This seems to be restricted to fast speech.

89

carlos molina-vital is expressed as yawarniyki with a syllabic structure /ja.war. 'nij.ki/. Notice that long vowels in HQ are considered consonants for syllabic structure. Consequently, following the monophthongization rules (see Section 4.3.4.2) llumtsuy ‘daughter in law’ is pronounced ['ʎumʦiː] with a long vowel at the end. If we were going to say ‘my daughter-in-law’ the 1st person possessive is marked by lengthening the last vowel in the root. This would lead to the impossible form *['ʎumʦiːː]. Instead, -ni is introduced to produce llumtsi:ni: [ʎum'ʦiːniː].

infinitive -y, imperative -y, emphatic perfective -yku, continuous -yka:, and second person nominal -yki). However, this is not always the case, particularly in the Northern Huaylas variety where, according to Julca Guerrero (2010) the diphthong is still present. Furthermore, expansion into lexemes is still very restricted. Julca Guerrero (2010: 132) only found two lexemes showing monophthongization: llumtsuy ‘daughter-in-law’ and luychu ‘deer,’ but not consistently in the region.7

4.3.4.3 Vowel lowering in derivational morphology

4.3.4.2 Monophthongization Perhaps the most innovative phonological process in HQ is the reduction of virtually all diphthongs into long vowels, i.e. monophthongization. This process was explicitly reported in early works for this variety (Parker 1967; Escribens and Proulx 1970; Swisshelm 1971), and it is currently dominant throughout the Callejón de Huaylas (Julca Guerrero 2010: 126); it is thus at a more advanced stage than similar monophthongization processes in Chachapoyas Quechua (see Chapter 6 by Aviva Shimelman and Jairo Valqui in this volume). Table 4.5 illustrates the instances of monophthongization and their contexts. The first instance of monophthongization is the most widespread one: /ay/ → [eː]. Since the 1960s it has been reported in lexemes and suffixes. The second instance, /aw/ → [oː], was reported as having restrictions such as a preference to occur in suffixes (e.g. the locative -chaw and the simulative -naw) and after palatal segments and /p/ (Parker 1976: 53−4). However, the process now is consummated through the region at the lexeme level (Julca Guerrero 2010: 128−9). Finally, the change /uy/ → [iː] is the most restricted one among the speakers of HQ. Specifically, monophthongization occurs when /y/ is the first segment in a suffix (e.g.

Verbal derivational suffixes ending on /i/ or /u/ undergo a lowering process that turn their high vowel into [a]. This happens when those suffixes are followed by a member of the set of “triggering” suffixes: these include the cislocative/translocative suffix -mu, the causative suffix -tsi, the 1st person object suffix -ma:, and the benefactive suffix -pu.8 Triggering suffixes, with the exception of -pu, which does 7 A related phenomenon is that the sequence /iy/ sometimes produces [i:] but most of the times is resolved as [i]. Thus, yaskiykuy /yas'kiykuy/ ‘answer, please’ is mostly produced as [jas'kikuj] and seldom as [ jas'kiːkuj]. 8 Parker (1973a: 40, 1976: 130) claimed that in HQ lowering effects were triggered also by the imperfective suffix -yka:, which lowers -ri and -pu; and by the middle voice suffix -ku, which lowers -pu. Those claims are incorrect. Regarding -yka:, Parker does not acknowledge the frequent combination of the durative suffix -ra: and the continuous -yka:, which, due to shortening of low vowels followed by a consonant coda, is rendered as -rayka:. It is clear that this instance of -ra does not come from the punctual aspect suffix -ri. The lowering of -pu to -pa is more problematic. Parker suggests that -yka: and -ku lower -pa. However, neither -yka: nor -ku has been reported as a trigger suffix in other varieties of Central Quechua, which casts doubts on that analysis. However, the main reason to reject the lowering of -pu before -ku or -yka: in HQ is that all the cases found in the corpus illustrate cases of -pa where this suffix has a different function from -pu: the latter is a benefactive, while the former could be either a general applicative that freely alternates with -pa:, or the first segment of the suffix -paku, which is cognate with the applicative (see Section 4.6.2.2.6 for the applicative suffix -pa: and Section 4.6.2.3.5 for -paku.)

Table 4.5 Monophthongization in Huaylas Quechua

Lexemes

/ay/ → [eː]

/aw/ → [oː]

/uy/ → [ɪː]

aytsa /'ajʦa/ > ['eːʦa] ‘meat’

awkish /'awkiʃ/ > ['awkɪʃ] ~ [oːkɪʃ] ‘old man’

luychu /'luychu/ > ['lujʧu] ~ ['lɪːʧu] ‘deer’ llumtsuy /’llumʦuy/ > ['ʎumʦɪː] ~ ['ʎumʦuj] ‘daughter-in-law’

tsaychaw /'ʦaychaw/ > ['ʦeːʧoː] ‘(in) there’ allqunaw /allqu'naw/ > [aʎqo'noː] ‘like a dog’

mikuy /mi'kuy/ > [‘mɪkɪː] ‘to eat,’ ‘eat!’ punuykun /punuy'kun/ > [punɪː'kuŋ] ‘she slept’

Morphemes takaykan /ta'kaykan/ > [ta'keːkaŋ] ‘she is hitting’ nanayki /na'nayki/ > [na'neːki] ‘your sister’

90

huaylas (ancash) quechua lower to -pa, do not undergo vowel lowering themselves. Some contrasting examples are in (1). (1) a. Rika:rin. {rika:-ri-n} see-punc-3.sbj ‘S/he sees briefly.’ b. Rika:ramun. {rika:-ri-mu-n} see-punc-cisl-3.sbj ‘S/he comes see briefly.’ (2) a. Haqipuy. {haqi-pu-y} leave-ben-imp ‘Leave it to him!’ b. Haqipamay. {haqi-pu-ma-y} leave-ben-1.obj-imp ‘Leave it to me!’ (3) a. Pukllaykuntsik. {puklla-yku-ntsik} play-excep-1pl.sbj ‘We played.’ b. Pukllaykatsintsik. {puklla-yku-tsi-ntsik} play-excep-caus-1pl.sbj ‘We made them play.’ (4) a. Rikarkunki. {rika-rku-nki} look.up-uw-2.sbj ‘You look it up.’ b. Rikarkapunki. {rika-rku-pu-nki} look.up-uw-ben-2.sbj ‘You look it up for her.’ The trigger suffix does not need to be placed immediately next to the affected one, as the contrast between the examples in (5) shows. (5) a. Rika:riyan. {rika:-ri-ya-n} see-punc-3.sbj ‘They see briefly.’

b. Rika:raya:mun. {rika:-ri-ya:-mu-n} see-punc-tf-cisl-3.sbj ‘They come see briefly.’ Also, the trigger suffix can affect a string of derivational suffixes. For instance, in Example (6), has -rpu, -rku, and -ri are affected by -mu. (6) hiqarparkaramuptin {hiqa-rpu-rku-ri-mu-pti-n} grow-dw-uw-punc-cisl-subds-3.sbj ‘when it has germinated fully’ (lit. ‘when it grows downwards and upwards,’ as a plant does when it germinates)

4.3.5 Stress assignment At the word level, the penultimate syllable is most generally stressed (as it is the norm for other Quechuan languages). Thus, adding suffixes will not displace stress from that position (e.g. pu-sha-n /'pu.shan/ ‘he guides’ and pu-sha-lla-n /pu.'sha.llan/ ‘he just guides’). However, there are several cases in which the primary accent does not fall on the penultimate syllable. First, toponyms ending in /n/ consistently display final stress (e.g. Tuspín /tus.'pin/, Parón /pa.'run/, Huancapetí /wan.ka.pi'ti/). Julca Guerrero (2009: 142−3) provides evidence for short vowels in a final stressed syllable originally coming from unstressed long vowels in final position. Other terms alternate between penultimate and final stress, e.g. turmanyay /tur.man. 'ye:/ or /tur. 'man.ye:/ ‘rainbow’ (Swisshelm 1971: 167). Finally, final stress, as is the case in other Quechuan languages, serves as an emotional device: vocatives and interjections are frequently associated with that stress pattern, e.g. tayta:! [teː'taː] ‘Sir!,’ alláw! [a.'llaw] ‘what a pity!’) The most complete analysis of stress patterns in relation with the length of a syllable is in Swisshelm (1971: 163–75). His proposal acknowledges several instances where penultimate stress does not occur. For instance, he observes that if there is only one long syllable in a word (i.e. CVC, VC, V:, or CV:) and it occurs anywhere before the last syllable, it will then carry the primary stress. Thus, warmi ['waɾ.mi] ‘woman’ or wara:nin [wa.'ɾaː.ni] ‘tomorrow’ follow this rule and have penultimate stress, while ampipa ['am.pi.pa] ‘at night’ or hirkakunapa ['hiɾ.ka.ku.na.pa] ‘through the mountains’ have stress in their initial syllable, the long one. Other rules Swisshelm (1971) proposes are that when several syllables before the final one are long, then the last one of

91

carlos molina-vital them will be stressed (e.g. siqaykarinqa [se.qeːka.'ren.qa] ‘he will fall down’ or aqaykaqtana [a.qeː.'kaq.ta.na] ‘the one already grinding it’); and if there are no long syllables before the final one, then the initial syllable becomes stressed (e.g. suwakuq ['su.wa.kuq] ‘the one who steals,’ waraka ['wa.ra.ka] ‘sling’).

4.4 Parts of speech and transcategorical operations 4.4.1 Parts of speech HQ is similar to other Quechuan languages in that it has four types of words: (i) nouns (e.g. hirka ‘mountain’), (ii) verbs (e.g. chura- ‘to put’), (iii) ambivalents (e.g. suwa ‘thief ’ and suwa- ‘to steal’), and particles (e.g. mana ‘not,’ hina ‘like, so’). Adjectives can be considered as a type of noun, since nothing in their morphosyntactic behavior distinguishes them from nouns. In Example (7), for instance, chukru ‘hard’ appears in object function without further modification. (7) Manam pay chukruta mikuyta puwidintsu. {mana-m pay chukru-ta miku-y-ta not-evd s/he hard-acc eat-inf-acc puwidi-n-tsu} be.able.to-3.sbj-neg ‘He can’t eat hard things.’ Some adjectives come from stative verbal roots (e.g. kushi ‘happy’), while others arise from nominalizations, some of which ended up lexicalized. For instance, shumaq ‘pretty’ and mushuq ‘new’ have clearly lexicalized the agentive nominalizer -q. This is no different from how lasa-q ‘heavy’ is created from lasa- ‘to weight.’ The main formal indicator of word-classhood is the type of suffix a root can take. Nouns take nominal suffixes (e.g. possessive inflection, plural, case) and verbs take verbal suffixes (e.g. tense, person, aspect, voice). Ambivalent roots take nominal or verbal morphology. There are also independent suffixes, known traditionally in Quechuan linguistics as enclitics (Cerrón-Palomino 1987b: 287). They convey notions related to the nature of information (e.g. evidentials), topic, focus, and other types of (inter)subjective stance. They are important to determine parts of speech because they can attach to nouns, verbs, and particles. What distinguishes particles is that they only take independent suffixes, never nominal or verbal ones (e.g. mana-m ‘no’). Interjections may be considered another type of word inasmuch as they do not seem to take any suffixes, and they have a consistent final-syllable stress pattern (e.g. atatáw ‘how ugly!,’ achalláw ‘how pretty!’).

92

4.4.2 Transcategorical operations In order to derive nouns from verbs, HQ uses the following nominalizers: potential -na, actual -nqa, agentive -q, resultative -shqa, and infinitive -y. Conversely, to turn nouns into verbs, the following suffixes are frequently used: transformative -ya:, desiderative -na:, factive -tsa:, causative -tsi, middle -ku, and spontaneous -ka: (which is also described as a reflexive, medio-passive, or passive in other varieties).

4.4.2.1 Nominal derivation (nominalization) 4.4.2.1.1 -na ‘potential nominalizer (pot.nmlz)’ -na nominalizes a verb from the perspective of its potential occurrence. Simply put, it expresses the possibility of an event. Frequently, it creates a referent that serves as an instrument, but also as an enabling location (e.g. tsapa-na ‘lid’ is the thing that allows the action denoted by the verb tsapa- ‘to cover’ to take place; but puklla-na denotes both the instrument ‘toy’ and the location ‘playground’). -na, expressing an irrealis event, is central to purpose constructions (see Section 4.8.4.2.3.2). It stands in direct contrast to -nqa, the actual or realis nominalizer, as Example (8) shows. (8) Kanan musyarirqunki imanaw rurana: kanqanta. {kanan musya-ri-rqu-nki now know-punc-rec.pst-2.sbj imanaw rura-na-: how make-pot.nmlz-1.poss

ka-nqa-n-ta} be-act.nmlz-acc

‘Now you know how my chores at home are.’ Here rurana: (lit. ‘my potential doing,’ i.e. ‘my chores’) serves as the modifier to kanqanta ‘what it actually is,’ the object of the cognition verb musya- ‘to know.’

4.4.2.1.2 -q ‘agentive nominalizer (ag)’ The nominalizer -q reifies an event based on the participant responsible for carrying out the event (the subject). For instance, in wamra salba-q ‘the savior of the child’ (from wamra ‘child’ and salba- ‘to save,’ a Spanish loan) the agentive nominalizer creates a referential noun. Notice that the modifier wamra corresponds to the object argument of salba-. Nominalizations with -q serve also as modifiers (e.g. qutsu-q yukis ‘singing thrush,’ from qutsu- ‘to sing’ and yukis ‘thrush’). In this way, -q is used for relative clauses (see Section 4.8.4.2.1.1). It is common to refer to -q as an “agentive nominalizer.” However, this is not entirely accurate. The argument which -q creates does not need to be volitional or affected by the

huaylas (ancash) quechua result of an event brought about by another participant. Consider qishya-q (from qishya- ‘to be sick’), a state verb whose subject is the affected participant. Nothing implies that something caused the participant to be sick, let alone that the subject is agentive/volitional.

4.4.2.1.3 -shqa ‘resultative nominalizer (res)’ The nominalizer -shqa (commonly known also as “past participle”) creates a referential expression understood in terms of a resulting state. For this reason it is widely used to create not only a referent, but also a deverbal adjective. For instance, wanu-shqa (from wanu- ‘to die’) can mean ‘dead one’ or simply denote the resulting quality, as in wanu-shqa allqu ‘dead dog.’ Moreover, -shqa commonly refers to whatever entity is affected, or results created during the event. Thus, muru-shqa-yki ‘what you sowed’ (from muru- ‘to sow,’ and with the 2nd person possessive suffix -yki) refers to a sowed entity. Constructions with -shqa can also be used as modifiers, which makes them analogous to relative clauses (see Section 4.8.4.2.1.1).

4.4.2.1.4 -nqa ‘actual nominalizer (act.nmlz)’ -nqa is used to express nominalized events corresponding to actual occurrences, in either the present or the past. In this sense, -nqa associates a “realis” modality to its referent. Parker (1976: 102) described it as an “imperfective nominalizer.” This description is rather odd, since Parker states that the nominalized event could also have ended (perfective aspect). In fact, -nqa is not used consistently as either a perfective or imperfective marker. Instead, what this suffix conveys is an event nominalization understood as an actual occurrence, which is frequently used with verbs of knowledge, communication, and perception. For this reason, the referred event can be in-progress or complete, as seen in (9). (9) Tushunqanta musya:. {tushu-nqa-n-ta musya-:} dance-act.nmlz-3.sbj-acc know-1.sbj ‘I know that s/he is dancing / that s/he danced.’ -nqa is used less frequently in HQ because -shqa has taken over the function of expressing an actually occurring event (by focusing on the perfective profile of this suffix, see Section 4.6.3.2.5).

4.4.2.1.5 -y ‘infinitive (inf)’ The infinitive suffix -y, common to every Quechuan variety, is used as the basic verbal noun. In other words, it typifies an event without implying any temporal or modal information about it. Thus, qutsu-y is not only ‘to dance,’ but also

‘dancing’ and the equivalent to the noun ‘dance.’ Likewise, achacha-y (from achacha:- ‘to be hot’) can mean ‘heat.’ The rather neutral properties of the infinitive make it very commonly used as the object of verbs like muna- ‘to want; to like,’ and yacha- ‘to know’ (see Section 4.8.4.2.2.1).

4.4.2.2 Verbal derivation Suffixes in this class, -ya:, -na:, and -tsa:, all end in a long vowel. However, they will be found with a short vowel whenever followed by a coda consonant in the same syllable (e.g. yaku-na:-nki ‘you feel thirsty’ becomes yakunanki [ya.kʊ.'nan.kɪ]).

4.4.2.2.1 -ya: ‘transformative (tf)’ As a productive suffix, transformative -ya: takes a noun in order to derive a spontaneous resulting state.9 This basic function can be clearly seen in forms like puka-ya:- ‘to turn red; to blush’ (from puka- ‘red’) or waktsa-ya:- ‘to get poor’ (from waktsa- ‘poor; orphan’). However, -ya: can be used in ways that are not transformative, but that imply a subject affected by the state resulting from an event having taken place. Thus, qanyas ‘the whole day’ and karu ‘far’ yield qanyas-ya:- ‘to spend the whole day,’ and karu-ya:- ‘to be far (from).’ Also, it is likely that this suffix is involved in turning onomatopoeic terms into verbs with imperfective lexical aspect (e.g. puqpuq-ya:- ‘to spurt out; to bubble with force’ and karkar-ya:- ‘to shiver’).

4.4.2.2.2 -na: ‘desiderative (desr)’ This suffix indicates that a subject is affected by the desire or propensity to experience an event. This last sense also includes that of an impending aspect (‘to be about to’). When it takes nouns as its basis, the resulting verb suggests that the subject wants the referent of the noun (e.g. yaku-na:‘to be thirsty,’ from yaku- ‘water,’ or warmi-na:- ‘to desire a woman,’ from warmi ‘woman’). However, -na: is more frequently found with two types of verbs. First, it combines with intransitive ones expressing an affected subject or an event whose direct object is produced or emitted by the subject (e.g. ishki-na:- ‘to feel like falling; to be about to fall,’ from ishki- ‘to fall,’ waqa-na:- ‘to feel like crying; to be about to cry,’ from waqa- ‘to cry,’ ishpa-na:- ‘to feel like urinating; to be about to urinate’). Second, it is used with atmospheric verbs. In those cases the event is impending or ‘about to happen’ (e.g. tamya-na:- ‘to be about to rain,’ from tamya ‘rain,’ ampi-na:- ‘to be about to get dark,’ from ampi- ‘to be dark’). 9 Sometimes -ya: is labeled “inchoative,” which should be reserved for an aspectual function. Since it only applies to nominal bases and not to verbs, as aspect suffixes do, “transformative” is a more accurate descriptor.

93

carlos molina-vital 4.4.2.2.3 -tsa: ‘factive (fact)’ This suffix turns noun bases into verbs; the meaning of the base noun then denotes a resulting product, performance, or an extended state. These three uses can be respectively seen in wayi-tsa:- ‘to make a house’ (from wayi ‘house’), haqa-tsa:‘to collect a debt’ (from haqa ‘debtor’), alli-tsa:- ‘to improve’ (from alli ‘good’), and puka-tsa:- ‘to make red’ (from puka ‘red’). Notice the contrast between -ya: and -tsa:: the first one is spontaneous, while the second implies a causer that brings the results referred by the base noun.10

4.4.2.2.4 -ta: ‘substance-covering (cov)’ This suffix has a very reduced distribution. It takes a nominal base referring to a substance (such as a color). The derived verb conveys the act of covering something with such substance (e.g. yawar-ta:- ‘to cover in blood,’ from yawar ‘blood,’ kachi-ta:- ‘to salt something, cover with salt,’ from kachi ‘salt,’ yuraq-ta:- ‘to paint white, to cover in white,’ from yuraq ‘white’).

4.4.2.2.5 Voice as verbalizer: -tsi ‘causative (caus),’ -ku ‘middle voice, reflexive, medio passive (mid)’ and -ka: ‘spontaneous voice (spon)’ It is possible to find many instances of nouns that take suffixes that otherwise are used with verbs to mark voice alternations (see Section 4.6.2.2). The causative -tsi combines sometimes with nouns producing verbs similar to those using factitive -tsa:. Examples are huti-tsi- ‘to christen,’ from huti ‘name,’ ladu-tsi- ‘to put in one side,’ from ladu- ‘side,’ misa-tsi- ‘to have mass celebrated,’ from misa ‘Catholic mass,’ and chichi-tsi- ‘to breastfeed,’ from chichi ‘breast’). The use of the middle voice suffix -ku as a verbalizer is a very productive mechanism in HQ. This construction indicates that the base noun is something procured, made, or worn by the subject. In this sense, the subject is always affected (as a beneficiary) by their own actions toward the noun referent. Examples are yuyu-ku- ‘to look for mustard leaves,’ from yuyu ‘mustard leaf,’ wayna-ku- ‘to procure a lover for oneself,’ from wayna ‘lover, boyfriend,’ aswa-ku- ‘to prepare chicha,’ from aswa ‘chicha,’ masa-ku- ‘to make dough (for bread),’ from Spanish masa ‘dough,’ and tsuku-ku- ‘to put on a hat,’ from tsuku ‘hat.’ In few cases, the spontaneous voice suffix -ka: is used with nouns to create a verb: qucha-ka:- ‘to become a lake’ (i.e. ‘to result a water accumulation’), kuta-ka:- ‘to end in a 10 The combination of the transformative -ya: with causative -tsi is also frequently attested with a similar function to that of -tsa:. For instance yana ‘black’ yields yana-ya:-tsi- ‘to make turn black,’ whose meaning is analogous to that of yana-tsa:- ‘to make black.’ However, -ya:-tsi- does not imply a direct and sustained action from the subject in order to bring about a progressive change (e.g. an otherwise spontaneous transformation is triggered by the subject). On the other hand, -tsa: implies that the subject is making the object change through all the steps of the process.

94

corner,’ tsillpi-ka:- ‘to become splinters; for an object to decay,’ and pirqa-ka:- ‘to end up formed a wall’ (from pirqa ‘wall,’ although it is used as an ambivalent stem). Those verbalizations imply the progressive passing of time until a result referring to the base noun takes place.

4.5 Nominals 4.5.1 Subclasses of nominals HQ includes the typical classes of nouns found in other Quechuan languages. These include common nouns (e.g. tika ‘mud brick’), personal pronouns (nuqa ‘I,’ qam ‘you,’ pay ‘s/he, it’), demonstrative pronouns (e.g. kay ‘this,’ tsay ‘that’), quantifiers (llapa ‘all’), emphatic pronouns (kiki ‘self ’), interrogative-indefinites (e.g. pi ‘who,’ ima ‘what,’ imay ‘when’), and numerals (e.g. huk ‘one,’ ishkay ‘two,’ kima ‘three’). Traditionally, a small class of pre-adjectives has been proposed. They include alla:pa ~ alla:pi ‘much,’ allibuwenu ‘very well’ (from Spanish bueno ‘good’), limpu ‘totally’ (perhaps from Spanish limpio ‘clean’), pasaypa ‘extremely’ (from Spanish pasar ‘to pass,’ ‘to surpass’), sayllama (['seʎama]) ‘very,’ shumaq ‘pretty; very,’ and mas ‘more’ (from Spanish). These terms profile the highest point of the scale implied by the term they modify. In other words, they stress a higher degree in, for instance, an adjective (e.g. alla:pa chukru ‘very hard’). Consequently, pre-adjectives are defined as adjective-modifying words. However, each and every one of the so-called pre-adjectives in HQ can be used with verbs too without changes (e.g. pasaypa ranti-n ‘he buys a lot,’ limpu ampikurin ‘it gets totally dark’).11 For this reason, it is best to view pre-adjectives as a type of adverb: specifically, a modifier sensitive to scalar properties that are central to both the temporal profile of verbs and the degree of the quality implied in some specific nouns.

4.5.1.1 Pronouns HQ, like other Quechuan languages, includes personal pronouns (which can take case marking), demonstrative pronouns (also determiners), pronouns that only appear with nominal inflection (emphatic and quantifiers), and interrogative pronouns (which double as indefinite ones).

4.5.1.1.1 Personal pronouns The main pronominal forms are nuqa ‘I,’ qam ‘you,’ and pay ‘s/he, it.’ As is common in most Quechuan languages, 11 The only exception is mas, which modifies verbs as mas-ta, in the accusative case.

huaylas (ancash) quechua the 1st person plural distinguishes between nuqa-kuna, the exclusive form (i.e. the addressee is not included in the pronoun referent), and nuqa-ntsik, the inclusive form. The plural for the second and third person are qam-kuna and pay-kuna, respectively. Like every noun, personal pronouns are marked for case (nuqa-ntsik-paq ‘for us, including you’). If the subject has been made explicit and is maintained as topic, it is frequently left unmentioned, unless the speaker wants to emphasize the identity of the subject. It must be noted that personal pronouns are seldom used as direct objects (only if there is need to contrastively emphasize the referent of a direct object). Instead, as it is common in Quechuan languages, there is an object-person marking system (see Section 4.6.3.1.2).

4.5.1.1.2 Demonstrative pronouns These pronouns have a deictic function based on the location of the referent in relation to the 1st, 2nd, and 3rd persons. While they can be used with the nominal plural suffix -kuna, this is infrequent. Thus, kay ‘this’ (e.g. kay wamra ‘this child’) points at a referent close to the speaker, but away from the speaker, and tsay ‘that’ (alternatively hay) indicates that the referent is close to the hearer, but away from the addressee (e.g. tsay/hay shipash ‘that young woman’). There are two evenly distributed 3rd person-based demonstratives. The first one, wak ‘that other,’ is based on the 3rd person considered as “the other” participant different from the speaker and the addressee (e.g. wak warmi ‘that other woman,’ wak bida-chaw ‘in the other life’ with bida ‘life,’ from Spanish, and locative -chaw.) The second one, taqay ‘that over there,’ does not imply the notion of “other,” and it seems more aligned with the deictic function kay ‘this’ and tsay ‘that’ have since it points to a referent located away from both the speaker and the addressee (e.g. taqay marka ‘that town over there’). Other demonstratives are wakin ‘the other’ and huk ‘a, another.’ Wakin ‘the other’ is related to wak ‘that other,’ and is much more frequently used. It refers to entities whose definiteness is relative to a previously defined part of the set. Thus wakin is best understood as ‘some’ or ‘the rest.’ The numeral huk ‘one’ in HQ can be used to indicate a definite referent that has not yet been made specific in speech. Consequently, it could mean both ‘a’ and ‘another.’ Thus, huk ra:tu huk sitiyuchaw, huk ra:tu huk sitiyuchaw means ‘one time (ra:tu) at one place (sitiyu), another time at another place.’

4.5.1.1.3 Nominally inflected pronouns (emphatic and quantifiers) Some pronouns must be inflected for nominal agreement in addition to case marking. The first one is the emphatic pronoun kiki ‘self.’ In this way, kiki-: ‘myself,’ kiki-yki ‘yourself ’

(pronounced most usually [kɪkɪ-kɪ] or [kɪkɪː-kɪ] in careful speech), and kiki-n ‘him/her/itself ’ indicate that their referent stands in sharp contrast to any other possible one (i.e. it is that person, and no other). In the 3rd person, this pronoun can be used as modifier (e.g. kiki-n tiyu-ntsik-paq ‘for our uncle himself ’). The second is the universal quantifier llapa ‘all’ (e.g. llapa-: ‘all of us,’ llapa-yki-kuna ‘all of you’). As a modifier it is found only in the 3rd person singular, llapa-n (e.g. llapa-n hara chakra ‘all of the corn plot’). Notice that plural marking is absent in the quantifier when the nominal head is marked (e.g. llapa-n pumakuna ‘all the cougars’). Thus, llapankuna only serves as a pronoun. Regarding the plural marking of llapa‘all,’ it seems that the presence of -kuna implies a precise reference to the individuals in a set. Thus, llapa-: alternates with llapa-:-kuna ‘all of us,’ but examples of the latter are found connected with an enumeration of individuals.

4.5.1.1.4 Interrogative and indefinite pronouns To form open questions, HQ uses the following basic set of interrogative pronouns: ima ‘what,’ pi ‘who,’ imay ‘when,’ may ‘where,’ ayka ‘how much/many,’ and mayqin ‘which.’ Ima ‘what’ is used with other suffixes or in composition with other words in order to create imanaw ‘how’ (using the simulative case -naw), and imanir ‘why’ (lit. ‘saying what,’ using the same-subject subordinate ni-r ‘saying’). Imanir must be considered a lexicalized form, since it yields the form imanirpis ‘for whatever reason’ (see below). While these pronouns can be used as free forms, that is considered too direct and even rude. Therefore, it is more common to find them in combination with the direct evidential -m ~ -mi for more frontal questions, where a specific answer is considered necessary (e.g. Ima-m parla-nki? ‘What are you saying?,’ with parla- ‘to talk’) or the contrastive -taq/-tan, which is used more frequently to connect a question to the ongoing dialogue, as e.g. in (10). (10) Aykatataq munanki? {ayka-ta-taq muna-nki?} how.many-acc-contr want-2.sbj ‘How many do you want?’ The continuative suffix -raq is present if the question implies an adverse situation or something that should not have happened, as in (11). (11) Imapaqraq takllata haqiramurqu:? {ima-paq-raq taklla-ta what-ben-cont foot.plow-acc haqi-ri-mu-rqu-:?} leave-punc-cisl-rec.pst-1.sbj ‘What for did I abandon my plow?’

95

carlos molina-vital Table 4.6 Interrogative and indefinite pronouns in Huaylas Quechua Interrogative pronoun

Gloss

Indefinite

Indefinite gloss

ima

‘what’

ima-pis

‘anything, whatever’

pi

‘who’

pi-pis

‘anyone, whatever’

imay

‘when’

imay-pis

‘any time, whenever’

may

‘where’

may-pis

‘anywhere, wherever’

ayka

‘how many’

ayka-pis

‘any amount’

mayqin

‘which’

mayqin-pis

‘whichever’

imanaw

‘how’

imayna-pis

‘anyhow, however’

imanir

‘why’

imanir-pis

‘whatever reason’

Interrogative pronouns serve as the basis for indefinite ones. To do this, they take the additive suffix -pis (e.g. ima-pis ‘anything,’ ‘whatever,’ may-chaw-pis ‘in any place,’ ‘wherever’). Negative indefinite pronouns are tied to the use of mana ‘not’ in the sentence (e.g. mana ima-pis ‘nothing’). However, the particle ni ‘nor’ (perhaps a Spanish loan) is frequently used to introduce a negative indefinite pronoun (e.g. ni ima-pis ‘nothing,’ ni may-pis ‘nowhere’). Sometimes, ni is used without -pis in the pronoun (e.g. ni pi-pa haqa-n ka-:-tsu ‘I am not anybody’s debtor,’ with pi ‘who’ in genitive case -pa, haqa ‘debtor’ marked with the 3rd person possessive suffix -n, and ka- ‘to be’ inflected for the 1st person by vowel lengthening with negative -tsu). Table 4.6 provides a summary of the interrogative and indefinite pronouns.

(but also ‘before’) and yapay ‘second’ (but also ‘increase’). For further ordinals, Spanish terms are widely used (sigundu ‘second,’ tirsiru ‘third,’ etc.). The Spanish number system is widely used for keeping tally of produce and to carry on business. Native Quechua numbers are productive, but restricted to age, calendar (though years are frequently in Spanish), and to household activities. Table 4.7 presents the numeral system.

Table 4.7 Numerals in Huaylas Quechua Number

Term

‘one’

huk

4.5.1.2 Numerals

‘two’

ishkay

HQ uses a decimal system. Once the number goes beyond a multiple of chunka ‘ten,’ units start being added. To do this, this variety simply indicates the number of units right after the number of tenths. Thus, ‘eleven’ is chunka huk. A unit before chunka ‘ten,’ pachak ‘100,’ or waranqa ‘1,000,’ is a multiplier. Thus, ‘25’ is iskay chunka pitsqa (lit. ‘two ten five’), ‘749’ is qanchis pachak chusku chunka isqun, and ‘2036’ is iskay waranqa kima chunka huqta. Hunu ‘1,000,000’ is mostly unknown in this variety. It is usual to find the agentive nominalizer -q combined with ishkay ‘two’ and kima ‘three.’ The forms ishkaq and kimaq mean ‘the two/three of them,’ and usually refer to people. Ordinal numbers are not used as a derived set from ordinal numbers. Instead, speakers use terms like punta ‘first’

‘three’

kima

‘four’

tawa

‘five’

pitsqa

‘six’

huqta

‘seven’

qanchis

‘eight’

puwaq

‘nine’

isqun

‘ten’

chunka

‘100’

pachak

‘1,000’

waranqa

96

huaylas (ancash) quechua 4.5.1.3 Gender As it is common for Quechuan languages, nominals are not marked for gender. However, some roots include gender as part of their core meaning. Thus, shipash is ‘young woman’ and cho:lu is ‘young man’ (from Peruvian Spanish cholo), while chakwas is ‘old woman’ and awkis is ‘old man.’ Also, it is frequent to introduce a gender distinction in terms of the biological sex of a person by using ullqu ‘male’ or warmi ‘woman.’ Thus, ullqu wawa and warmi wawa are ‘male baby’ and ‘female baby,’ respectively. For animals, ullqu is used, but for a female animal the term is china. Thus, ullqu allqu and china allqu are ‘male dog’ and ‘female dog.’ Many loans replicate the gender marking pattern of Spanish (e.g. lo:k-u ‘crazy-masc’ and lo:k-a ‘crazy-fem,’ duktur ‘doctor-masc’ and duktur-a ‘doctor-fem’).

4.5.1.4 Order of constituents in the nominal phrase Even if there is relative freedom of constituents in the clause, the nominal phrase order is quite rigid (Example 12 is adapted from Parker 1976: 31): (12) tsay dem i

llapan all ii

kima three iii

mana not iv

allapa very v

allim uchuk good little vi vii

hara chakra (-kuna) corn field (-pl) viii head ix ‘all three of those not very good little corn fields’ The position slots are thus: (i) demonstrative, (ii) quantifier, (iii) number, (iv) negator, (v) scalar adverb (i.e. preadjective), (vi) adjective (i.e. quality noun), (vii) adjective (i.e. size noun), (viii) attributive noun, and (ix) plural.

4.5.2 Nominal derivation This morphological process derives a noun from a base noun. There are five suffixes used for this process, (i) -yuq ‘inalienable possession (inal.poss)’ and -sapa ‘multiple possession (mult.all)’ use the base to indicate what is possessed. -ntin ‘accompaniment, adjacency (incl)’ presents a relation of inclusivity. The suffix -lla ‘restrictive, limitative (lim),’ which can also be found with verbs, restricts the reference of its base by making it unique. Finally, -mayi ‘partnership (part)’ suggests that the base noun is a shared characteristic between members of a group.

4.5.2.1 -yuq ‘inalienable possession (inal.poss)’ As it is common in most Quechuan languages, there is no verbal root meaning ‘to have.’ One of the ways in which

possession is expressed requires the use of the suffix -yuq with a noun base. Thus, wayi ‘house’ yields wayi-yuq ‘house owner, one who has a house.’ The possessive relation is expressed with the copula, as in (13). (13) Qam shumaq wasiyuq kanki. {qam shumaq wasi-yuq ka-nki} you pretty house-inal.poss be-2.sbj ‘You have a pretty house.’ (lit. ‘You are a pretty-house-owner.’) This suffix can be added to any type of nominal, including nominalizations, with the exception perhaps of personal pronouns (e.g. tsay-ni-yuq ‘that has that,’ ima-yuq ‘that has something,’ kawa-y-yuq ‘that has life,’ from kawa- ‘to live’ marked with the infinitive suffix-y). What defines -yuq is that it indicates that the nominal it marks is an inalienable possession of the nominal it modifies. In this way, the possessed entity is viewed as not freely transferrable, and helps characterize its owner. Consequently, body parts (e.g. piqa-yuq ka-: ‘I have a head,’ from piqa ‘head’), family members (e.g. warmi-yuq ka-nki ‘you have a wife,’ from warmi ‘woman’), and inherent qualities (e.g. chani-yuq ‘that has value, costly,’ from chani ‘value,’ and huti-yuq ‘that has a name, named,’ from huti ‘name’) use this suffix. Example (14) shows that the nominal head nuna ‘person’ is characterized by having the name ‘Abraham.’ (14) Abraham hutiyuq nuna shamun. {Abraham huti-yuq nuna Abraham name-inal.poss person

shamu-n} come-3.sbj

‘A person named Abraham has come.’ (lit. ‘An Abraham-name-having person has come.’)

4.5.2.2 -sapa ‘multiple possession (mult.all)’ This suffix also conveys a possessive notion. Thus, wayi-sapa indicates that someone possesses many houses. The difference between this suffix and -yuq is based on the quantity of the possessed entity: -sapa indicates a greater amount of it. This increased possession does not distinguish between individual entities (e.g. wa:ka-sapa nuna ‘person owner of many cows,’ from wa:ka ‘cow’) or substances (qillay-sapa nuna ‘rich person,’ from qillay ‘silver, money’).

4.5.2.3 -ntin ‘accompaniment, adjacency (incl)’ This suffix indicates that its base referent is included together with another entity (e.g. utsu-ntin ‘with hot pepper,’ from utsu ‘hot pepper,’ yawar-ni-ntin ‘with blood,’ from yawar ‘blood,’ which requires the epenthetic suffix -ni). Example (15) shows a context of use.

97

carlos molina-vital (15) Kashkitapis rurantsik aytsallantin. {kashki-ta-pis rura-ntsik aytsa-lla-ntin} soup-acc-add make-1pl.sbj meat-lim-incl ‘We also make soup with meat.’ -ntin also indicates that all the parts of the base referent are equally involved. This can be seen with the quantifier llapa ‘all’ (llapa-ntin ‘all together,’ ‘including everyone’), the particle hina ‘like, so’ (hina-ntin ‘everywhere’), and nouns like ampi ‘night’ (ampi-ntin ‘all night’).

4.5.2.4 -mayi ‘partnership (part)’ A noun marked with -mayi indicates that its referent shares with another individual a partnership defined by the base. Examples are marka-mayi ‘fellow town person,’ from marka ‘town,’ yawar-mayi ‘relative’ (lit. ‘blood mate’), llamkay-mayi ‘co-worker,’ from llamka-y ‘work-inf,’ wiñay-mayi ‘contemporary; age peer,’ from wiña-y ‘grow-inf.’

4.5.2.5 -lla ‘restrictive, limitative (lim)’ This suffix restricts the referent of the noun to either a single entity or a single class (e.g. tanta-lla ‘only bread,’ from tanta ‘bread,’ qam-lla ‘just you,’ from qam ‘you’). The delimitative sense takes place right before case marking. When this suffix takes the position immediately after the nominal stem, it takes a diminutive meaning (e.g. yanaqilla-n-kuna-lla-wan ‘just with your little friends,’ from yanaqi ‘friend’ with 3rd person -n, plural -kuna, and instrumental case -wan).12

4.5.3 Nominal inflection As is customary for Quechuan languages, nouns are inflected for person, number, and case. The nominal suffixes are distributed in the following order: derivational, person, number, and case. The complete distribution of derivational and inflectional suffixes is illustrated in Table 4.8.

4.5.3.1 Person Quechuan languages make extensive use of nominal person marking. Its most obvious function is that of indicating the possessor of an entity. For this reason, they are frequently labeled “possessive suffixes.” Table 4.9 presents the nominal agreement paradigm. The use of the plural marker -kuna produces ambiguity when combined with person. Thus, wayi-:-kuna (from wayi 12

The status of -lla as a derivational suffix is not entirely clear. An almost identical form -lla: can be considered a verbal derivational suffix (see Section 6.2.3.2.) Perhaps -lla should be analyzed as an enclitic.

98

‘house’) means not only ‘our (excl) house,’ but also ‘our (excl) houses’ with a plural possessed referent, as well as ‘my houses’ (where -kuna has scope only over the possessed referent, not the person/possessor). Person marking can be used with any nominal base, with the exception perhaps of personal pronouns (e.g. kay-ni-: ‘this thing of mine,’ tsay-ni-n ‘that thing of his/hers,’ ima-yki ‘what thing of yours,’ ishka-: ‘the two of us’). The quantifier llapa ‘all’ and the emphatic pronoun kiki ‘self ’ must always be marked for person in HQ (see Section 4.5.1.1.3). HQ, like most Quechuan languages, uses nominal person in several syntactic constructions. This is seen in the agreement between a genitive-marked person and the noun they are linked to (see Section 4.5.3.3.1), and in nominalized verbs using person marking to indicate grammatical relations (see Section 4.8.4.2.2.2)

4.5.3.2 Number The plural is marked through -kuna (e.g. shipash-kuna ‘young women,’ from shipash ‘young woman’). When another word indicating plurality is present, HQ speakers alternate between the use of -kuna and its absence. Thus, atska nuna-kuna and atska nuna are acceptable for ‘many people.’ Likewise, ‘four sheep’ can be chusku u:sha or chusku u:shakuna. Combining -kuna with person marking in a noun produces ambiguity with regards to what is pluralized, either the person, the base noun, or both (see Section 4.5.3.1).

4.5.3.3 Case As in any Quechuan language, this variety has a rich case system. Every noun is overtly marked to provide information about its role in the phrase or sentence (subject, object, goal, possessor, among others). In some cases, the case marker can be stacked (usually genitive -pa, instrumental -wan, similitude -naw). The nominative is zero-marked, as is the case for the subject in nominative–accusative languages (see Section 4.8.1). There are twelve case markers in HQ, and I will describe all of them separately, except the nominative. Nominative case is unmarked, as in every Quechuan language, and it expresses the subject argument of transitive and intransitive verbs (see Section 4.8.1). Table 4.10 introduces and exemplifies them.

4.5.3.3.1 -ta ‘accusative (acc)’ This is the case of the object argument of a transitive verb, as in (16). (16) Tantata rura:. {tanta-ta bread-acc ‘I make bread.’

rura-:} make-1.sbj

huaylas (ancash) quechua Table 4.8 Distribution of nominal suffixes in Huaylas Quechua Inflection Nominalizers

Derivation

Person

Number

Case

-na ‘pot.nmlz’ -q ‘ag’ -shqa ‘res’ -nqa ‘act.nmlz’

-yuq ‘inal.poss’ -sapa ‘mult.all’ -ntin ‘incl’ -mayi ‘mate’

-: ‘1.poss’ -yki ‘2.poss’ -n ‘3.poss’ -ntsik ‘1pl.poss’

-kuna ‘pl’

-chaw ‘loc’ -kama/-yaq ‘lim’ -man ‘all’ -naw ‘simul’ -pa ‘gen’ -paq ‘ben, all, purp’ -piq ~ -pita ‘abl’ -pura ‘excl’ -raykur ‘csl’ -ta ‘acc’ -wan ‘com’ -Ø ‘nom’

Note: Only some case suffixes are illustrated.

Table 4.9 The nominal person paradigm in Huaylas Quechua Person

Nominal agreement sg

1

-:

pl

verb is the theme, although sometimes it is located after the verb (even without -ta, making it a zero-marked object). (18) furnishes examples with qu- ‘to give.’ (18) a. Shanti tsurinta ruripata qurqan. {shanti tsuri-n-ta ruripa-ta Santiago child-3.poss-acc skirt-acc

incl

-ntsik

excl

-:-kuna

qu-rqa-n give-pst-3.sbj ‘Santiago gave his child a skirt.’

2

-yki

-yki-kuna

3

-n

-n-kuna

It is common to use -ta to indicate the ultimate goal of a motion event (17). In this sense, it contrasts with -man, which profiles the path that leads into the goal (see Section 4.5.3.3.3): (17) Wayiyikita purishun. {wayi-yiki-ta puri-shun} house-2.poss-acc walk-1pl.incl.sbj.fut ‘Let’s walk to your house.’ HQ does not distinguish between direct and indirect object. The latter is the argument affected in a way that is subordinate to another argument’s affectedness (the theme object). Consequently, ditransitive verbs like qu- ‘to give,’ qara- ‘to serve,’ or qatu- ‘to sell’ yield double accusative constructions. Usually, in these constructions the argument closer to the

b. Payta qun puka hilu. {pay-ta qu-n puka hilu-Ø} s/he-acc give-3.sbj red thread-acc ‘She gives him a red thread.’ While -ta is generally associated with the affected participant in an event, this suffix is found with nouns that are the source of the subject’s experience. With this function, -ta is tied to verbs with low transitivity like perception and emotion. As illustrated in (19), -ta crucially is also found with intransitive verbs without any valency-increasing voice suffix (Parker 1973a:16): (19) a. Na:nita rikaykaya:. {na:ni-ta rika:-yka:-ya-:} road-acc see-ipfv-pl-1.sbj ‘We are seeing the road.’

99

carlos molina-vital Table 4.10 Case marking in Huaylas Quechua Case

Suffix

Example

Gloss

Locative

-chaw

Wayi-chaw uryanki.

‘You work at home.’

Limitative

-kama/-yaq

Markara:-kama/-yuq aywayarqa:.

‘We went until (up to) Marcará.’

Allative

-man

Chuntayuq-man aywaykan.

‘He is going to Chontayoc.’

Simulative

-naw

nuqa-naw

‘like me’

Genitive

-pa

Shanti-pa wawqi-n.

‘Santiago’s brother.’

Benefactive, allative, purposive

-paq

Nuqantsik-paq uryaykayan.

‘They work for us.’

Ablative

-piq ~ -pita

Waras-pita/-piq shamu:.

‘I come from Huaraz.’

Interactive

-pura

yanasa-pura

‘among friends’

Causal

-raykur

yamta-raykur

‘because of the firewood’

Accusative

-ta

Kashki-ta miku:.

‘I eat soup.’

Instrumental-Comitative

-wan

Allqu-wan purinki.

‘You are walking with a dog.’

Nominative

-Ø

Peyu-Ø tushun.

‘Pedro dances.’

b. Warma mamanta waqan. {warma mama-n-ta waqa-n} child mother-3.poss-acc cry-3.sbj ‘The child is crying for her mom.’ (i.e. ‘due to her mom’) HQ speakers sometimes leave the direct object unmarked. As already seen in (18b) above, ditransitive verbs can leave their theme object unmarked. Also, a stereotypical object, or just an inanimate one, can be left unmarked (e.g. papa alla- ‘to dig out potatoes’ from alla- ‘to dig’). Finally, -ta is used to turn nouns of quality (i.e. “adjectives”) into verb modifiers (i.e. “adverbs”). This can be seen in alli-ta mikuy ‘eat well’ (from alli ‘good’ with mikuy ‘eat,’ a command) and atska-ta upyay ‘drink a lot,’ ‘drink much’ (from atska ‘many’ and upyay ‘to drink,’ a command).

4.5.3.3.2 -pa ‘genitive (gen)’ As in many Quechuan languages, the genitive -pa is used in agreement with the nominal person marking to indicate the possessor of the referred entity, as seen in (20). (20) Peyupa warmin {peyu-pa warmi-n} pedro-gen woman-3.poss ‘Pedro’s wife’ As shown in (21), this agreement process can be recursive.

100

(21) Peyupa warminpa turin {peyu-pa warmi-n-pa pedro-gen woman-3.poss-gen turi-n} brother.of.a.woman-3.poss ‘Pedro’s wife’s brother’ This suffix also indicates a path for motion events, specifically ‘through,’ or ‘along’ as in Example (22). (22) Hirkakunapa hananpa aywan. {hirka-kuna-pa hana-n-pa mountain-pl-gen top-3.poss-gen

aywa-n} go-3.sbj

‘It goes through the top of the mountains.’ This function of -pa might be active in expressions like ampipa (‘at night,’ from ampi ‘night’) or hunaq-pa (‘during daytime,’ from hunaq ‘day’). Perhaps it is also related to the use of -pa to create modifiers indicating the condition in which the subject is when doing or experiencing an event. It can also be used to modify how the event takes place. In this function, -pa is frequently found alongside the delimitative -lla. Examples are chaki-pa ‘by foot, on foot’ (from chaki ‘feet’), fiyu-pa ‘badly; strongly’ (from fiyu ‘ugly,’ a Spanish loan), waqa-lla-pa ‘crying’ (from waqa ‘cry’), paka-y-lla-pa ‘hiddenly, sneakily (from paka- ‘to hide’ in the infinitive). The next example illustrates the use of genitive -pa with a nominalized verb with infinitive -y.

huaylas (ancash) quechua (23) Huanshi mantsapaypa puriykan. {huanshi mantsa-pa:-y-pa Juan fear-appl-inf-gen

puri-yka:-n} walk-ipfv-3.sbj

‘Juan is walking afraid [of something].’

4.5.3.3.3 -man ‘allative (all)’ With this case, speakers mark the direction and endpoint into/onto which something moves (e.g. marka-man ‘to town,’ from marka ‘town,’ or uray-man ‘downwards,’ from uray ‘down’). A further example is in (24). (24) U:sha:kuna quchaman hiqaykurqun. {u:sha-ː-kuna qucha-man hiqa-yku-rqu-n sheep-1.poss-pl lake-all fall-iw-rec.pst-3.sbj ‘Our sheep have fallen into the lake.’ -man is never used to mark an indirect object in HQ (see Section 4.5.3.3.1).

4.5.3.3.4 -chaw ‘locative (loc)’ This case makes its base noun a location. It is commonly translated as ‘in,’ ‘on,’ or ‘at.’ For instance, chakra-chaw ‘in the crop field’ (from chakra ‘crop field, plot’), hunaq-chaw ‘in the day, during the day (from hunaq ‘day’).

4.5.3.3.5 -paq ‘dative, allative, purposive (dat)’ With animate referents, this case marks the beneficiary or maleficiary of an event, as in Example (23). (25) Kallmi paninpaq takiykan. {kallmi pani-n-paq carmen sister.of.a.woman-3.poss-ben taki-yka:-n} sing-ipfv-3.sbj ‘Carmen sings for her sister.’ In general, -paq conveys that a noun is the goal toward which an action aims. In this allative sense, there is no implicit path, only the -paq-marked endpoint of the action. This is illustrated in the examples in (26). (26) a. Limapaq purikuya:shaq. {lima-paq puri-ku-ya:-shaq lima-all walk-mid-plv-1.sbj.fut ‘We [will] go to Lima.’ b. Huk killana fiyistapaq faltan. {huk killa-na fistiya-paq falta-n} one month-disc party-all be.left-3.sbj ‘There is already one month left for the party.’

-paq is also used to indicate the purpose of non-motion events (see Section 4.8.4.2.3.2). While an animate endpoint tends to be interpreted as a beneficiary, affectedness of a dative participant is not guaranteed. To indicate full beneficiary status in an animate participant, the voice suffix -pu is used. For this reason, it is best to see -paq as a dative, and not as a benefactive (see Section 4.6.2.2.5).

4.5.3.3.6 -piq ~ -pita ‘ablative (abl)’ Either -piq or -pita mark a noun as the starting point of a motion, or the source of an event or state (e.g. Huaraspiq ‘from Huaraz,’ from the toponym Huaraz, waray-pita ‘since the morning,’ from waray ‘morning,’ or quri-pita ‘of gold,’ from quri ‘gold’) There is widespread alternation between -piq and -pita, which are indistinguishable in meaning. -pita is preferred in the Southern Huaylas variety, while -piq is preferred in the Northern one. Both forms originate from *-piqta, which is still used by few elderly speakers in the Northern Huaylas variety (Julca Guerrero 2010: 165–74). The ablative also indicates the reason for something (e.g. tsay-pita ‘due to that’). It is mostfrequent with nominalized clauses, as illustrated in the example 27. (27) aywakunqanpita {aywa-ku-nqa-n-pita} go-mid-act.nmlz-3.poss-abl ‘due to his leaving’ In this sense, -pita ~ -piq is replacing -raykur as a causal case (see Section 4.5.3.3.10).

4.5.3.3.7 -naw ‘simulative (simul)’ The simulative case indicates that its base noun serves as the reference point in a comparison (e.g. atuq-naw ‘like a fox,’ from atuq ‘fox,’ chukis-naw ‘like someone from Conchucos,’ from chukis ‘person from Conchucos,’ or kay-naw ‘like this,’ from kay ‘this’).

4.5.3.3.8 -yaq, -kama ‘limitative (lim)’ This case alternates in HQ between -yaq and -kama. They indicate that some path or time-period is covered until the limit determined by the base noun. It is commonly translated as until, e.g. hirka-yaq or hirka-kama ‘until the mountain,’ from hirka mountain,’ and waray-yaq or waray-kama ‘until dawn,’ from waray ‘dawn.’ While both forms are used in Northern and Southern varieties, -kama is more frequent in the Southern one. It is possible to use the limitative case meaning ‘while’ or ‘during.’ This seems to be possible only with the actual nominalizer -nqa, as illustrated in (28).

101

carlos molina-vital (28) Waska:ta apamunqa:yaq runkulla:ta umpakurillaqman. {waska-:-ta apa-mu-nqa-:-yaq rope-1.poss-acc bring-cisl-act.nmlz-1.poss-lim runkulla-:-ta bag-1.poss-acc umpa-ku-ri-lla-q-man} entrust-mid-punc-lim-1sbj.2.obj-cond ‘While I bring my rope, could I entrust you with my bag?’ (lit. ‘While I bring my rope, I would entrust you with my bag.’) As shown in (29), the combination -kama-yaq is possible in the same noun to emphasize a full duration of an event or a full covered path of a motion event. (29) Tsayshi kanankamayaq waqayan suwirtillankunata. {tsay-shi kanan-kama-yaq dem-evr now-lim-lim waqa-ya:-n suwirti-lla-n-kuna-ta} cry-plv-3sbj luck-lim-3.poss-pl-acc ‘Therefore, up until now they cry for their luck.’

4.5.3.3.9 -wan ‘comitative-instrumental (com)’ This case suffix has two main functions. First, it marks a noun as an instrument, that is, a co-causer subordinate to the subject participant, as in Example (30). (30) Llawiwan punkuta kicharqa:. {llawi-wan punku-ta kicha-rqa-:} key-com door-acc open-pst-1.sbj ‘I opened the door with the key.’ The second function, comitative, is to indicate that a noun referent is co-present alongside some other entity. This interpretation is preferred with animate beings. An example is in (31). (31) Malli taytanwan qatswan. {malli tayta-n-wan qatswa-n} maría father-3.poss-com dance-3sbj ‘María dances with her father.’ Also, comitative -wan is used for coordinating more than one nominal. Usually, it marks the first member of the set (e.g. ashnu-wan atuqkuna ‘the donkey and the foxes’). However, it is also possible to use it to mark all the members. In Example (32) i ‘and’ (from Spanish y) is used also as a coordinator.

102

(32) Yacha: Limachawmi papa:ni:wan, mama:ni:wan, wawqi:kunawan i pani:wan. {yacha-: lima-chaw-mi reside-1.sbj Lima-loc-evd papa:-ni-:-wan mama:-ni-:-wan dad-ep-1.sbj-com mom-ep-1.sbj-com wawqi-:-kuna-wan i brother.of.a.man-1.poss-pl-com and pani-:-wan} sister.of.a.man-1.poss-com ‘I live in Lima with my dad, my mom, my brothers, and sisters.’

4.5.3.3.10 -raykur ‘causal (csl)’ This case makes its base noun a cause, reason, or motivation (e.g. qillay-raykur ‘for money,’ from qillay ‘silver, money,’ yanasa-raykur ‘due to a friend,’ from yanasa ‘female friend,’ or tsay-raykur ‘because of that,’ a frequent consequence connective from tsay ‘that’). The next example shows how the noun tsuri-yki ‘your child’ serves as the cause for the verb urya- ‘to work.’ (33) Tsuriykiraykur uryaykanki. {tsuri-yki-raykur urya-yka:-nki} child-2.poss-csl work-ipfv-2.sbj ‘You are working for your child.’

4.5.3.3.11 -pura ‘exclusive (excl)’ The exclusive -pura establishes that only the elements within the set determined by the base noun participate in some event. This suffix emphasizes the members of the noun class excluding any entity outside that class.13 For instance, forms like yanaqi-pura ‘among friends’ (from yanaqi ‘male friend’), kiki-n-pura ‘among themselves’ (from kiki ‘self ’ with the 3rd person possessive suffix), and marka-pura ‘within towns’ (from marka ‘town’) specify who or what is exclusively involved in an event. Due to this involvement, -pura is labeled “interactive” by Julca Guerrero (2009: 191). With numerals (including fractions), -pura indicates the number of participants among which the amount gets distributed (e.g. pullan pura ‘half and a half,’ ‘between two,’ from pullan ‘half,’ pitsqa chunka pura ‘fifty per person,’ from pitsqa chunka ‘fifty’). Parker (1976: 89) calls this use “adverbial” because it specifies how the distribution is done (i.e. the noun with pura modifies the verb). Arguably, -pura refers instead to the 13 Some authors like Parker (1976: 89) consider -pura a particle. However, there is no clear indication that this form can be used independently of a nominal base immediately preceding it. Therefore, we classify it as a case suffix.

huaylas (ancash) quechua subject of the sentence to elaborate upon the number of participants exclusively involved. The distributive use of -pura and its relation to non-quantifying expressions still needs analysis.

‘He is my fellow townsman.’ However, whenever non-overt ka-n takes any other aspect, tense, or mood suffix, it gets overtly expressed, as in (36).

4.6 Verbs HQ’s verbs are bound stems which at least are inflected for person and tense. However, most frequently they are marked with aspect and voice derivational suffixes. As is usual in Quechuan languages, verbal stems are not a numerous class. My verbal database counts around a thousand of them. However, verbal derivational morphology allows for many meaning specifications that enhance a verb’s expressive potential. For example, apa- ‘to take’ yields apa-mu‘to bring’ (i.e. ‘to take toward the speaker,’ with -mu ‘cislocative’), and apa-tsi- ‘to send’ (i.e. ‘to make take,’ with -tsi ‘causative’), not to mention lexicalized derivational suffixes leading to forms like apari- ‘to carry on the back’ (likely from *-ri ‘upwards, above’).

4.6.1 Verb subclasses

(36) Pay llaqtamasi:mi karqan. {Pay llaqtamasi:-mi s/he fellow.townsman-1.poss-evd

ka-rqa-n} be-pst-3.sbj

‘He was my fellow townsman.’ It is common to find ka- marked in the continuous aspect -yka:. With this aspect, the copula is understood as an ongoing state that can be framed within an unbounded time period (see Section 4.6.2.1.1.1). Thus, ka-yka- conveys a predicate as a background. Example (37) illustrates this use. (37) Wak bidachawmi almayki mula kaykan. {wak bida-chaw-mi alma-yki that.other life-loc-evd soul-2.poss mula ka-yka:-n} mule be-ipfv-3.sbj ‘In the other life your soul is a mule.’

Based on how they relate to other nominal arguments, three verb subclasses have been proposed for HQ and perhaps every other Quechuan language: (i) copulatives relate nouns with predicates (e.g. ka- ‘to be,’ tuku- ‘to become’), (ii) intransitives, which denote events that require a single participant to take place or being experienced (e.g. aywa- ‘to go,’ wanu‘to die’), and (iii) transitive verbs that require the interaction of two participants (e.g. rura- ‘to make,’ chura- ‘to put’).

4.6.1.1 Copulative verbs Copulas are verbs that relate their subject to a nominal predicate. The most basic and frequent copula is ka- ‘to be, exist,’ which is illustrated in (34). (34) Nuqa chakra nunam ka:. {nuqa chakra nuna-m I crop.field person-evd

(35) Pay llaqtamasi:mi. {pay llaqtamasi-:-mi} s/he fellow.townsman-1.sbj-evd

ka-:} be-1.sbj

‘I am a peasant.’ (lit. ‘I am farm person’). As with most Quechuan languages, a 3rd person singular subject (pay ‘s/he, it’) does not take overt ka- in the present tense. Therefore, ‘he is my fellow townsman’ (with llaqtamasi ‘fellow townsman’) cannot use the 3rd person form ka-n. Instead, it must be rendered as in (35).

As shown in Example (38), when ka-n is overtly used, it only expresses the subject’s existence: (38) Allqu kan. {allqu ka-n} dog be-3.sbj ‘There is a dog.’ As shown in (39), in this function ka- remains uninflected for number. (39) Atska fiyistakuna kan. {atska fiyista-kuna many party-pl

ka-n} be-3.sg

‘There are many parties.’ The existential function of ka- ‘to be’ is crucial for the expression of possession in HQ. As characteristic in most Quechuan languages, there is no verb meaning ‘to have.’ Instead, to say “X has something” speakers literally say: “X’s something exists.” The genitive-marked possessor is optional in this construction, as in (40). (40) (Nuqapa) llanqi: kan. {(Nuqa-pa) llanqi-: I-gen sandals-1.poss

ka-n} be-3.sbj

‘I have sandals.’ (lit. ‘There are my sandals of mine.’)

103

carlos molina-vital It is possible to convey a stronger bond between the possessor and the possessed entity in HQ by using the benefactive suffix -pu with ka- ‘to be.’ Examples (41) and (42) illustrate this. (41) Luis Pardopa kapunaq huk kumpa:rim. {Luis Pardo-pa ka-pu-naq Luis Pardo-gen be-ben-narr.pst huk one

kumpari-m} compadre-evd

‘Luis Pardo had one compadre.’ (lit. ‘A compadre of Luis Pardoi belonged to himi .’) (42) Nuqapapis atskam kapaman. {nuqa-pa-pis atska-m ka-pu-ma:-n I-gen-add many-evd be-ben-1.obj-3.sbj ‘I also have a lot [of them]’ (lit.‘A lot [of them] belong to me too.’) Notice that the subject of ka-n is the possessed entity. However, no nominal agreement exists between the genitive possessor and the possessed entity (as in Luis Pardo-pa kumpari-n ‘Luis Pardo’s compadre’). It seems that in this construction the possessor is an appositional nominal construction sharing the same referent as the possessed entity. As such, it is optional (both huk kumpa:ri kapun and atska kapaman are acceptable). Finally, this construction is closer in meaning to English ‘to belong,’ since the possessor’s referent is given by an object introduced by the benefactive suffix -pu. In Example (41) it is a 1st person marked with -ma:, while in (42) it is a zero-marked 3rd person. Besides ka- ‘to be,’ the only other copulative verb is tuku‘to pretend to be,’ as in (43). (43) Qillaysapa tukunki. {qillaysapa tuku-nki} money-mult.all pretend.to.be-2.sbj ‘You pretend to be rich.’ In some sources (e.g. Pantoja et al. 1974), tuku- is used as a suffix attached to nouns and nominalized forms, as in (44). (44) wanuqtukur {wanu-q-tuku-r} die-ag-pretend.to.be-subis ‘pretending to be dead’

4.6.1.2 Transitive verbs Transitive verbs require two arguments for their elaboration. As a nominative–accusative language (see Section 4.8.1), HQ’s transitive verbs have a subject in the nominative case, and an object in the accusative case

104

(marked with -ta). Verbs expressing dynamic actions involving two participants, an agent and a patient, are the most representative of this class (e.g. taka- ‘to hit, knock,’ rura- ‘to make,’ or suwa- ‘to steal.’) Verbs that indicate a caused motion are prime cases of transitivity: the change of state they convey by placing the object in a different location or by causing a change in their posture is seen as mainly affecting the object participant. This is the reason why -ku—the middle-reflexive suffix that stresses the affectedness of the subject, and that is otherwise capable of combining with virtually every type of verb— seldom combines with caused motion events. Thus, caused motion verbs are at the very end of a semantic transitivity cline, fully distinguishable from intransitive ones. Understanding verbal transitivity as the presence of two participants distorts the conceptual flexibility of HQ verbs. For instance, verbs of consumption like miku- ‘to eat’ or upya‘to drink’ are listed as transitive because they take a direct object marked with -ta. Still, the main affected participant is the subject, which experiences the effects of consuming an object that becomes something of an inalienable possession. Thus, verbs of consumption are closer to intransitivity: quite frequently they appear without a direct object, not even one previously mentioned. Moreover, the use of the middle suffix -ku is customary for them, because the subject’s affectedness is more prominent than the object’s. Similarly, several cognition verbs (including perception and emotion) are viewed as transitive because they take a -ta-marked object. The examples in (45)−(47) illustrate this. (45) U:sha atuqta mantsan. {u:sha atuq-ta mantsa-n} sheep atuq-acc fear-3.sbj ‘The sheep fears the fox.’ (46) Michi aytsata ma:kurinaq. {michi aytsa-ta ma:ku-ri-naq} cat meat-acc notice-punc-narr.pst ‘The cat noticed the meat.’ (47) Peyu mankata rikaykan. {peyu manka-ta rika:-yka:-n} pedro pot-acc see-ipfv-3.sbj ‘Pedro is looking at the pot.’ The use of a direct object in HQ is not restricted to typically transitive dynamic actions. Instead, objects are complements that elaborate an event as the interaction of two participants. However, for some verbs like those of emotion, cognition, and perception, those objects are not affected ones, but stimuli for their experiencing subject. As stimuli, they must be considered as a cause for the event. Consequently, transitivity goes beyond a simple

huaylas (ancash) quechua agent–patient relation to include an experiencer–stimulus one. This is the reason behind Parker’s (1973a: 16) pioneering observation regarding the possibility of using -ta-marked direct objects with intransitive verbs like asi‘to laugh,’ waqa- ‘to cry,’ llaki- ‘to be sad,’ aha- ‘to be mad,’ or puklla- ‘to play’ (which becomes ‘to taunt’ or ‘to flirt with’ with a direct object). Furthermore, -ta can be deployed in order to create this type of construction with verbs that are otherwise basically intransitive (see Section 4.5.3.3.1 above, but also 4.6.2.2.6 for applicative voice). For ditransitive constructions involving double accusative marking, see Section 4.5.3.3.1.

4.6.1.3 Intransitive verbs An intransitive verb denotes an event that can be carried out sufficiently by a single participant. Usually, motion events are considered prototypical instances of intransitive verbs. Translational motion events like aywa- ‘to go,’ pa:ri- ‘to fly,’ hiqa- ‘to fall,’ and non-translational ones like ta:- ‘to sit,’ sha:ri- ‘to get up,’ and tuma- ‘to turn’ all have a single subject participant acting on its own. Other dynamic events commonly considered as intransitive are atmospheric ones (e.g. tamya- ‘to rain,’ wayra- ‘to be windy,’ and ampi- ‘to become night’). In spite of the more representative role given to dynamic intransitive events, non-dynamic or stative events are prominent among intransitives. They are events where the subject experiences an effect or state that comes from their own potential or tendency to experience it. Under this definition many subtypes of verbs can be included: emotions and their expression (e.g. kushi- ‘to be happy,’ waqa- ‘to cry,’ tiqna- ‘to be bored; to be fed up’), resulting states (illa- ‘to be absent; to disappear,’ chichu- ‘for an animal to be pregnant,’ macha- ‘to be drunk’), and inherent states (e.g. lasa- ‘to weigh,’ alala:- ‘to be cold,’ ana:- ‘to be hard,’ pishi- ‘to be short of; to be not enough,’ and qamla- ‘to be insipid’). Furthermore, numerous spontaneous events expressing a change are seen in HQ as caused by their experiencing subjects (e.g. wanu- ‘to die,’ puwa- ‘to boil,’ wayu- ‘to have abundant fruit, bear fruit,’ ismu- ‘to rot,’ and timpu- ‘to cook slowly’). Thus, many verbs listed as transitive in dictionaries because they can take a direct object are poor examples of transitivity. For instance, taki- ‘to sing,’ ni- ‘to say,’ and yaski‘to answer’ have their object (physical or cognitive) created through an action depending upon the subject itself. This is also the case for so-called transitive events with cognate direct objects (e.g. pirqa- ‘to build a wall,’ from pirqa ‘wall,’ and uchku- ‘to make a hole,’ from uchku ‘hole’). In sum, intransitive verbs are quite diverse and widespread in HQ. From a conceptual perspective, intransitivity is as prominent as transitivity. In my corpus I

have counted 469 intransitive roots (including spontaneous, atmospheric, stative, verbs of emission, and those with cognate objects, besides typical dynamic actions) against 522 transitive roots. Intransitivity is, then, not significantly less represented in the lexicon.

4.6.2 Verbal derivation There are 18 derivational suffixes that modify the basic meaning of the verb stem. They can be classified as belonging to the categories of aspect, voice, and general manner (including direction, coverage of the action, and attention and carefulness). Based on their position and interaction with other suffixes (e.g. vowel lowering, see Section 4.3.4.3), the verbal plural suffix -ya: and the 1st person object suffix -ma: should be considered derivational suffixes. However, they will be described under inflection. Derivation always precedes inflection. While not obligatory, derivational suffixes are frequently present following a mostly strict patterning, but not without exceptions and alternations. Moreover, their combination creates complex and subtle meaning distinctions.

4.6.2.1 Aspect Aspect can be easily classified in two main areas: imperfective (time-unbounded events) and perfectives (timebounded ones). HQ has a rich aspectual system with many further distinctions.

4.6.2.1.1 Imperfectives 4.6.2.1.1.1 -yka: ‘continuous (ipfv)’ The most basic imperfective marker is continuous aspect -yka: (shortened to -yka when followed by a consonant in syllabic coda). It can be described as a progressive when combined with dynamic events. An example illustrating its use is given in (48). (48) Pintiykarqayki. {pinti-yka:-rqa-yki} jump-ipfv-pst-2.sbj ‘You were jumping.’ However, -yka: frequently combines with stative events to indicate that the situation is effectively going on, as in (49). For this reason, this suffix is best viewed as encoding continuous aspect.14 14 This characterization is indebted to Hintz’s (2011) analysis of -yka: as continuous aspect in Southern Conchucos Quechua.

105

carlos molina-vital (49) Ama mantsaka:yaytsu, nuqaqa kawayka:mi. {ama mantsa-ka:-ya-y-tsu proh fear-spon-plv-imp-neg nuqa-qa kawa-yka:-:-mi} I-top live-ipfv-1.sbj-evd ‘Don’t be scared, I am alive.’ (lit. ‘Don’t be scared, I am living.’) Finally, it is frequent to use -yka: to introduce some implicit temporal boundaries to the event (i.e. an event is viewed as taking place as of a certain period), as in (50). (50) Kananqa yachaykan Yungaychaw. {kanan-qa yacha-yka:-n yungay-chaw} now-top reside-ipfv-3.sbj yungay-loc ‘She is currently living in Yungay.’

4.6.2.1.1.2 -ykacha ‘iterative (iter)’ This aspectual suffix alternates with -kacha and indicates that an event is done in a repetitive way. However, even if the event is composed of several episodes, it is different from the habitual aspect. -ykacha suggests that the event is divided into shorter episodes that are carried out in an unfocused manner, which makes their distribution somewhat unpredictable. Therefore, it suggests an action done carelessly (e.g. rika-ykacha- ‘to take a look around, here and there,’ from rika:- ‘to see,’ puklla-ykacha- ‘to play around,’ from puklla- ‘to play’).

4.6.2.1.1.3 -ra: ‘durative (dur)’ This suffix reduces to -ra when followed by a consonant in syllabic coda. It indicates that an event is extending through time without interruption. Thus, the durative aspect is commonly associated with states or actions done in ways that extend beyond normal, particularly resulting states (e.g. puñu-ra:- ‘to stay asleep,’ from puñu- ‘to sleep,’ apari-ra:‘to remain carrying something on the back,’ from apari- ‘to carry on the back’). This resultative sense is somewhat implicit in the way the event is presented: the subject remains doing or experiencing what the verbs indicates for an extended period. This leads to -ra: being used as a resultative passive voice, although restricted to transitive verbs whose object experiences a change in position or location. This is illustrated in (51). (51) Kiri: wichqaran. {kiri-: wichqa-ra:-n} wound-1.poss close-dur-3.sbj ‘My wound remains closed.’ (cf. kiri-:-ta kicha- ‘to open my wound’) Finally, -ra: is frequently found combined with -yka: as -rayka:. The meaning of the combination is perfectly predictable from its parts: it indicates that a continuous event

106

extends beyond what is assumed to be its basic implicit boundary. However, because -rayka: is always placed before the causative suffix -tsi, even when -yka: can occur before or after -tsi, it can be considered a fused form. (52) I wayichawqa arupakurayka:yan. I warmi cha:rin mana kasashpa. {I wayi-chaw-qa and house-loc-top aru-paku-ra:-yka:-ya-n cook-atten-dur-ipfv-plv-3.sbj i warmi cha:-ri-n and woman arrive-punc-3.sbj mana kasa-shpa} not marry-subis ‘And in the house, they kept cooking [for the party]. And the woman arrived without being married.’

4.6.2.1.2 Perfectives 4.6.2.1.2.1 -ri ‘punctual (punc)’ The most basic perfective marker in HQ is punctual -ri. It indicates that an event takes place in a brief and immediate manner. It is very frequently found with events that involve a quick change (e.g. qimtsi- ‘to close eyes’ > qimtsi-ri- ‘to blink,’ from qimtsi- ‘to close eyes,’ or tupa-ri- ‘to bump into,’ from tupa- ‘to run into’). Other examples showing how -ri conveys shorter duration and immediate occurrence are pa:tsa-ri‘to lie down’ (from patsa:- ‘to lay on the ground’), rika:-ri‘to notice; to glance’ (from rika:- ‘to see’), yarpa-ri- ‘to recall’ (from yarpa- ‘to remember’), cha:-ri- ‘to get there’ (from cha:- ‘to arrive’), ayllu-ri- ‘to get together’ (from ayllu- ‘to gather’). By conveying a punctual event, -ri marks the event as complete. However, this is not the same as completive aspect, where an event develops thoroughly to completion. Instead, -ri profiles the minimal instance of an event that can be considered complete. Consider hita- ‘to throw.’ Its broad meaning gets refined through the punctual aspect, so in (53), hita-ri- points at the final step in the process of throwing something (i.e. ‘to expel’): (53) Tsaytanash hitaraya:mun linu yakuta upurir. {tsay-ta-na-sh hita-ri-ya:-mu-n dem-acc-disc-evr throw-punc-plv-cisl-3sbj linu yaku-ta upu-ri-r flax water-acc drink-punc-subis ‘Then they expel it [the fetus] by drinking flax water.’ This example also shows the combination of a subordinating suffix -r with -ri. Subordinating suffixes are immediately preceded by this aspectual suffix to indicate that the event

huaylas (ancash) quechua in the subordinate clause is completed. This is evidence of the basic perfective status of -ri. With stative or copulative verbs, -ri profiles the point in which a state or experience comes to be, as in (54). (54) Ichikllachawmi imayka karin. {ichik-lla-chaw-mi imayka little-lim-loc-evd any.type

ka-ri-n} be-punc-3.sbj

‘In a little bit, anything comes up.’ (i.e. ‘problems appear’) Example (54) could be considered an example of the inceptive or inchoative function of -ri. However, my observations agree with Parker’s (1973a: 20) position that this suffix does not clearly convey the initial stage of an event. The inceptive function of -ri is the result of focusing on an abbreviated and perfective construal of the event. In Examples (53) and (54) -ri portrays those events as punctual changes. However, they are not equivalent to the beginning of those events. The inceptive function is best represented through a construction involving the auxiliary qalla- ‘to begin’ and the same subject subordinator -r (see Section 4.8.4.2.3.1, and Examples (107) and (175) in Sections 4.6.2.3.1 and 4.8.4.2.3.1 respectively)

4.6.2.1.2.2 -yku ‘exceptional performance (excep)’ One of the most frequently used suffixes in HQ is -yku. Its meaning is notoriously difficult to pinpoint, not only for this Quechuan variety (see Chapter 5 by Aviva Shimelman in this volume for Yauyos Quechua). Its origin is in a directional non-productive suffix -yku ‘inwards and downwards motion.’ With this function it is only found with lexicalized verbs (e.g. yayku- ‘to enter’) and a few motion events (e.g. muchki-yku‘to bow down and into something,’ from muchki- ‘to bow.’) Parker (1973a: 22–3) proposes that the directional sense of -yku is part of a Proto-Quechua directional paradigm based on the use of the proto-verb *ya- ‘to be at; to move into.’ Alongside yayku- ‘to enter,’ there are forms like yarku- ‘to go up,’ yarqu- ‘to exit,’ and yarpu- ‘to go down.’ Thus, -yku ‘into,’ -rku ‘up,’ -rqu ‘out,’ and -rpu ‘down’ can be reconstructed.15 -rpu ‘down’ is seldom used with its directional sense and it has not developed in HQ an aspectual function.16 On the other hand, -rku ‘up’ and -rqu ‘out’ are still found with their directional sense in limited instances of motion events; however, like -yku ‘into,’ they have evolved into new functions, an aspectual one -rku and a tense reading for -rqu (see Sections 4.6.2.1.2.3 and 4.6.3.2.3). 15 Other verbs that display the full directional paradigm of ProtoQuechua are qayku- ‘to put into a pen,’ qarku- ‘to hill a plant to earth up,’ qarqu- ‘to expel,’ and qarpu- ‘to push down.’ Parker and Chávez Reyes (1976: 138) propose *qa- with the meaning ‘to make go.’ 16 In Southern Conchucos Quechua, -rpu has developed a clear perfectivecompletive aspectual function (Hintz 2011: 27–9).

Based on the modal or subjective undertones it gives to an event, Parker (1973a: 34–5, 1976: 129) and Swisshelm (1974: 499–504) do not describe -yku as an aspectual marker. It can express carefulness, hastiness, unexpectedness, politeness, adverseness, or some type of particular importance or effort when carrying out the event. The following examples illustrate the frequent use in commands (Example 55), a special way of behaving (Example 56, based on the copula ka‘to be’), an adversative/accidental punctual event (Example 57), and an action carried out after no previous expectations (Example 58). (55) Munarqa tsayllachaw haqiriykuy yanasa. {muna-r-qa tsay-lla-chaw want-subis-top dem-lim-loc haqi-ri-yku-y yanasa} leave-punc-excep-imp friend.of.a.woman ‘If you want, please drop it here, friend.’ (56) Yapay tsaynawlla kaykunki. {yapay tsay-naw-lla ka-yku-nki} again dem-simul-lim be-excep-2.sbj ‘Again, you act in this way.’ (i.e. ‘Again, you are particularly in this way’) (57) Alfredo kasi ishkiykun. {Alfredo kasi ishki-yku-n} Alfredo almost fall-excep-3.sbj ‘Alfredo almost falls.’ (Context: Alfredo almost faints and falls down.) (58) Felismente payqa kichaykamarqan. {felismente pay-qa kicha-yku-ma:-rqa-n} fortunately s/he-top open-excep-1.obj-pst-3.sbj ‘Fortunately, he opened [the door] for me.’ (Context: The narrator had gone to a place he was supposed to lodge, but no one opened, so he goes unannounced to another friend’s place, who opens his door.) The diverse readings of -yku aside, it must be stressed that it is a perfective marker. This can be inferred from its complementary distribution with the continuous aspect suffixyka:. They are never found in the same verb (unless -yku is used with its non-productive directional sense). Even if perfective -rku also is in complementary distribution with -yka: (see Section 4.6.2.1.2.3), it also cannot combine with yku. From this evidence, it can be proposed that -yku must be a perfective whose meaning is also opposed to that of -rku.17 17 These arguments for the aspectual nature of -yku were first exposed for Southern Conchucos Quechua in Hintz (2011), who followed intuitions

107

carlos molina-vital The combination of modal/subjective interpretations and its perfective nature suggests that -yku is actually a durative telic aspect (i.e. an accomplishment): the event is presented as requiring some duration that leads to its endpoint (its perfective boundary). The implicit duration in -yku has experienced several semantic extensions. First, it is linked to the effort it takes to go through the process in order to reach the result. This can be read as carefulness, emphasis, a higher than normal degree in what is done, or any intensified subjectivity associated with the event. From this focus on the duration, it is possible to move into the notion of ‘contrary to what is expected.’ Additionally, an event marked for the effort it requires is easily recruited as a politeness device (the speaker uttering the command acknowledges how a subject will do something that costs them an unexpected effort, hence the ‘please’ translation). In sum, the ‘duration-aseffort’ meaning of -yku extends to a politeness function, and a ‘contrary-to-expectation’ meaning. From there, it moves on to include the ‘unexpected outcome’ (without referring to the basic duration of the event). This motivates the use of -yku with accidental situations, as in (57) and, arguably also (58).

4.6.2.1.2.3 -rku ‘perfective of achievement (uw)’ The suffix -rku originates in the locational suffix *-rku ‘upwards motion’. From this meaning, now restricted to motion events (e.g. hita-rku- ‘to throw upwards,’ from hita- ‘to throw’), and a few lexicalized verbs (yarku- ‘to go up,’ qarku ‘to hill plants; to earth up’). As an aspectual marker, the meaning of -rku is directly opposite to that of the perfective of emphasis function of the exceptional performance marker -yku discussed in Section 4.6.2.1.2.3. It conveys a punctual telic aspect (i.e. an achievement). Since there is no need to imply the duration of the event, the results are achieved without reference to any effort implied in the process. This semantic characteristic of -rku was first elaborated in Parker (1973a: 23–6). He concluded that -rku expressed an event that is completed “without resistance.” Examples (59)–(61) illustrate the way in which -rku is used. (59) Pukyuman cha:rirnash yakuta chaqarkamun. {pukyu-man cha:-ri-r-na-sh spring-all arrive-punc-subis-disc-evr yaku-ta water-acc

chaqa-rku-mu-n} gourd.ladle-uw-3.sbj

‘She arrived at the spring and took water with a gourd container.’ about the distribution of -yku in Huallaga Quechua from Weber (1989). Hintz (2011) considers -yku a “completive-perfective” that implies an obligation.

108

(60) Shiminta hirarkapurqunaq. {shimi-n-ta hira-rku-pu-rqu-naq} mouth-3.poss-acc sew-uw-ben-rec.pst-narr.pst ‘he sewed his mouth up.’ (61) Kumpari, ma: upurkushun. {kumpari ma: upu-rku-shun} compadre come.on drink-uw-1pl.incl.sbj.fut ‘Compadre, come on, let’s drink [up].’ These examples confirm that -rku conveys an action carried out in a straightforward manner. Completion is achieved with little effort or resistance that could delay or make the action difficult. With stative events -rku indicates that the event is fully experienced, as shown in (62). (62) Takllanqa lasaparkunaq. {taklla-n-qa lasa-pa:-rku-naq} foot.plow-3.sbj-top weight-appl-uw-narr.pst ‘Hisi foot plow was heavy for himi ’ (i.e. ‘was weighting him down’) Observe how in (63) and (64) apa-yku-r is connected to something done under pressure, while apa-rku-r points at an event done under the acceptance of everyone involved. (63) Paytana rogarqunaq atskata rurun apaykur. {pay-ta-na roga-rqu-naq atska-ta he-acc-disc beg-rec.pst-narr.pst many-acc ruru-n egg-3.poss

apa-yku-r} take-excep-subis

‘Carrying a lot of eggs, he begged him.’ (64) Tsaynin aparkur kushishqa hiqariq bruhoqa. {tsay-ni-n apa-rku-r kushi-shqa dem-ep-3.poss take-uw-subis happy-res hiqari-q bruho-qa} go-hab shaman-top ‘Thus the shaman left happy, taking that [his payment].’ In contrast with -ri and -yku, which combine productively, -ri and -rku do not.18 Being already a type of punctual aspect, it is incompatible with -ri. However, despite its clear punctual perfective nature, -rku can be found with durative aspect -ra: and with iterative -ykacha. This is impossible for -yku, which rejects any imperfective. For instance, among several examples with -ykacha, we have chunka-kacha-rku‘to roll something around’ (from chunka- ‘to roll by pushing’), or hicha-kacha-rku- ‘to throw away,’ ‘to wastefully discard something’ (from hita- ‘to throw’). With durative 18 Six cases of -ri-rku appear in our corpus. Five clearly show a use of -rku as indicating upwards motion. The only unaccounted instance is yarqu-rirku-r ‘going out’ marked also with same-subject subordinator -r.

huaylas (ancash) quechua -ra: only chukru-ra:-rku- ‘to stay hardened up’ (from chukru‘to be hard’) is registered. This type of data needs further analysis.

4.6.2.2 Voice Voice suffixes in HQ includes those in charge of altering the valency in the verb, as well as the grammatical relations of their arguments. Valency increasing suffixes are -tsi ‘causative,’ -pu ‘benefactive,’ and -pa: ‘applicative.’ On the other hand, -ku ‘middle, reflexive, medio-passive, passive,’ -naku ‘reciprocal,’ and -ka: ‘spontaenous’ can reduce valency. As for aspect, however, voice suffixes tend to convey subtle alternative ways in which the event is developed by their participants. Those interpretations are often at odds with their assumed basic valency-altering function.

4.6.2.2.1 -tsi ‘causative (caus)’ Causative voice is marked with -tsi, which introduces a participant in the subject position with the role of causer, the ultimate responsible for the occurrence of the event. As is common in many Quechuan languages, the causer does not need to forcefully enact causation upon the affected participant (direct causation): it is possible for it to simply let the event occur (permissive causation). Moreover, the causer can also act through a proxy or in any other way that separates it from the exact moment in which the caused effect occurs (indirect causation). These types of causation are exemplified in Examples (65)–(67) with wanu- ‘to die,’ with (65) showing direct causation, (66) permissive causation, and (67) indirect causation. (65) Tsay taksha wambrakuna mas wanutsiyan taksha pishqukunata. {tsay taksha wambra-kuna mas dem little child-pl more wanu-tsi-ya-n taksha pishqu-kuna-ta die-caus-plv-3.sbj little bird-pl-acc ‘Those little children kill smaller birds.’ (Context: little children use BB-guns or slings to kill smaller birds.) (66) Hampikuqkuna qishyaqkunata wanutsinaq. {hampi-ku-q-kuna qishya-q-kuna-ta cure-mid-ag-pl be.sick-ag-pl-acc wanu-tsi-naq. die-caus-narr.pst ‘The doctors killed the sick ones.’ (Context: The doctors didn’t have the resources to cure anybody, so they just let them die.)

(67) (Paykuna) mishita wawantin wanutsiyarqan. {(paykuna) mishi-ta wawa-ntin {they cat-acc baby-incl wanu-tsi-ya-rqa-n} kill-caus-plv-pst-3.sbj ‘They killed the cat along with her kittens.’ (Context: The enemies of the cat chopped down a tree upon which she lived with her kittens so that they all died with the fall.) The participant affected by the caused event is the causee. There are different ways in which the causee relates to the causer, and this is reflected in the morphosyntax of causatives. First, the most basic and frequent causative construction uses an intransitive verb. Examples (65)–(67) illustrated this with the spontaneous intransitive verb wanu- ‘to die.’ Example (68) uses a motion event. (68) Kallu Shantita pintitsiykan. {kallu shanti-ta pinti-tsi-yka:-n} carlos Santiago-acc jump-caus-ipfv-3.sbj ‘Carlos is making Santiago jump.’ This basic construction is frequently extended to transitive verbs in order to express that the causer gets the caused event done, as in (69) and (70). (69) Huwanshi runkukunata paka:tsirqan. {huwanshi runku-kuna-ta paka:-tsi-rqa-n} Juan bag-pl-acc hide-caus-pst-3.sbj ‘Juan made his bags be hidden.’ (i.e. ‘Juan had his bag hidden’) (70) Mama:ni: tantata rantitsirqun. {mama:-ni-: tanta-ta ranti-tsi-rqu-n. Mom-ep-1.poss bread-acc buy-caus-rec.pst-3.sbj ‘My mom had bread be bought.’ / ‘She ordered bread to be bought’.) The basic function of causativity in those examples is to express that the subject gets something done, and the means by which this happens are not relevant: there is no need to mention any other participant besides the causer and the direct object of the verb. Thus, a transitive verb marked with -tsi is close in meaning to a spontaneous or passive event because its agent is defocused. However, if the agent in the caused event must be overt, HQ, as is the case in other Quechuan languages, can code it as a secondary causer through the comitative-instrumental case -wan. This secondary causer does not oppose the control of the causer, making it an unaffected participant. Example (71) illustrates this.

109

carlos molina-vital (71) Ana paninwan waytata apatsirqan. {ana pani-n-wan ana sister.of.a.woman-3.poss-com wayta-ta apa-tsi-rqa-n} flower-acc take-caus-pst-3.sbj ‘Ana had her sister take a flower.’ As in (72), this construction can also be used with any intransitive verb. (72) Allquta puritsin yanaqinwan. {allqu-ta puri-tsi-n yanaqi-n-wan} dog-acc walk-acc-3.sbj friend-3.poss-com ‘She has her friend walk the dog.’ Another way to code what appears to be a secondary causer is through accusative -ta. In this case, this participant is not a secondary causer, but a causee: they are forced to carry out the event. Thus, they oppose the causer’s control, which makes them an affected participant. Example (73) illustrates. (73) Ana paninta waytata apatsirqan. {ana pani-n-ta ana sister.of.a.woman-3.poss-acc wayta-ta apa-tsi-rqa-n} flower-acc take-caus-pst-3.sbj ‘Ana made her sister take a flower.’ It is impossible to use this type of construction with an intransitive verb. Since intransitive verbs have only one participant, their argument structure cannot accommodate two affected ones. This suggests that this is a double accusative construction where the first object (panin-ta ‘her sisteracc’) is the most prominently affected participant, while the other object (wayta-ta ‘flower-acc’) is analogous to the theme in ditransitive constructions.

4.6.2.2.2 -ku ‘middle (mid)’ The suffix -ku is not only the most frequently used derivational suffix in HQ, but also the most polysemous one. This suffix is present in every Quechuan language, and in most of them it is labeled as “reflexive” (e.g. by Aviva Shimelman in Chapter 5 of this volume for Yauyos Quechua and by Aviva Shimelman and Jairo Valqui in Chapter 6 for Chachapoyas Quechua). This is also the case in many descriptions of HQ (e.g. Swisshelm 1974: 479; Julca Guerrero 2009: 238), according to whom -ku indicates that the subject and object of a verb are co-referential. This purported reflexive function is found in self-care actions such as those in Examples (74)–(76).

110

(74) Malli naqtsakuykan. {malli naqtsa-ku-yka:-n} María comb-mid-ipfv-3.sbj ‘María is combing [herself].’ (75) Mayuchaw armakurqantsik. {mayu-chaw arma-ku-rqa-ntsik river-loc bathe-mid-pst-1pl.sbj ‘We bathed in the river.’ (76) Bruhoqa hampikurqan. {bruho-qa hampi-ku-rqa-n} shaman-top cure-mid-pst-3.sbj ‘The shaman healed himself.’ Parker (1973a: 9) rejected the idea that -ku is a reflexive marker. Reflexivity requires an action that can only be carried out as if involving an external object participant, fully distinguishable from the subject. Under this basic definition, no actual reflexives can be linked to -ku in HQ. First, grooming verbs are fully compatible with a body-part direct object. Thus, in the previous examples naqtsa- ‘to comb’ could take aqtsa-n-ta ‘hair-3.poss-acc’ or piqa-n-ta ‘head-3.poss-acc,’ while hampi- ‘to cure’ could use rikra-n-ta ‘arm-3.poss-acc’ as its direct object. A verb like arma- ‘to bathe’ is best viewed as an event performed through the use of the subject’s entire body, like any other intransitive motion event. However, it could also take kuwirpu-n-ta ‘body-3.poss-acc’ as its direct object and still be marked with -ku. Note that these possible direct objects are inalienable possessions of the subject. Thus, they cannot be considered external and fully distinguishable from the subject. Other verbs that have been used to exemplify reflexivity like kuchu-ku- ‘to cut oneself ’ or rupa-ku- ‘to burn oneself ’ do not yield true reflexive constructions. Their basic transitive action never gets to be transmitted upon a different direct object when a body part of the subject is affected. Crucially, all of the previous examples illustrate how whenever an inalienable object becomes affected, the subject becomes automatically affected too. Although not a true reflexive, kuchu-ku- ‘to cut oneself ’ shows that -ku still reduces the transitive valency of an event by restricting the way this event is carried out to the subject’s own body. Example (77) illustrates this even when an inalienable object is not involved. (77) Tsay machay rurinllaman alli paka:kuyay. {tsay machay ruri-n-lla-man dem cave inside-3.poss-lim-all alli paka:-ku-ya-y} good hide-mid-plv-imp ‘Hide yourselves well inside of that cave!’

huaylas (ancash) quechua Pa:ka- ‘to hide’ is transitive in HQ. However, pa:ka-ku- ‘to hide (oneself)’ involves a single participant doing the hiding. Since this action is done by moving into any location that serves as a hideout, it is not truly reflexive, but an intransitive construal of this event. One central function of -ku is to indicate that the subject is also the most affected participant. However, the subject must be at the same time responsible for what it experiences. The overlap of affectedness and control is automatic, and not obtained through a subject acting upon itself. Therefore, -ku is a middle-voice marker that does not include reflexivity among its functions. Instead, it gives an event a meaning analogous to that of intransitive verbs. This middle construal of transitive verbs can be seen in paki-ku- ‘to experience a fracture’ (from paki‘to break something’) and alista-ku- ‘to get ready’ (from alista‘to prepare something,’ a loan from Spanish). Another common function of -ku is to indicate the subject’s self-benefit. This involves a direct object that is external to the subject. In those cases, a middle interpretation is present because the direct object becomes a possession of the subject. Also, -ku emphasizes how this object produces a prominent affectedness in the subject, which becomes a beneficiary. This is shown in the examples in (78) and (79). (78) Puma u:shantsikta apakurqun. {puma u:sha-ntsik-ta cougar sheep-1pl.incl.poss-acc apa-ku-rqu-n} take-mid-rec.pst-3.sbj ‘The cougar took our sheep away.’ (79) Qam mishki aytsata mikukunki. {Qam mishki aytsa-ta miku-ku-nki} you tasty meat-acc eat-mid-2.sbj ‘You eat tasty meat.’ (lit. ‘You eat for yourself tasty meat.’) There are virtually no restrictions in how -ku is used: it takes transitive or intransitive verbs, and it can be used with animate or inanimate beings (see Section 4.6.1.2 for a restriction in the use of -ku with caused motion transitive events). With intransitive verbs, and especially those in which the subject is an experiencer, -ku seems redundant. However, aywa-ku- (from aywa- ‘to go’), yacha-ku- (from yacha- ‘to know; to learn’), and rika:-ku- (from rika:- ‘to see’) suggest that their subject performs with deliberate intention in connection with an enhancement of what is experienced. Thus, aywa-ku- means ‘to leave, go away,’ yacha-ku- ‘to study,’ and rika:-ku- ‘to watch something fun,’ i.e. ‘to spectate.’ A deliberate-engagement reading of -ku is impossible with inanimate subjects. However, the basic middle function remains in that the subject has a clear tendency or propensity

to experience the effects of a verb. Moreover, those effects are considered more prominent (complete, total, or even lasting for an extended period). In (80), -ku indicates that the floor (an inanimate entity) moved noticeably due to the effect of an earthquake: (80) Ampipapis kada ratu kuyukarayka:murqan. {ampi-pa-pis kada ratu night-gen-add every time kuyu-ku-ra:-yka:-mu-rqa-n} move-mid-dur-ipfv-cisl-pst-3.sbj ‘[The floor] at night kept moving there every time.’ The use of -ku to indicate the propensity of a subject to experience the effects of the event has evolved into imperfective aspectual functions such as conveying that the event is customary for the subject. This motivates the frequent association of the agentive nominalizer -q with -ku (e.g. yacha-tsi-q ‘one who teaches’ becomes yacha-tsi-ku-q ‘teacher, one who customarily teaches,’ from yacha-tsi- ‘to make learn’).19 In sum, whatever its specific function, -ku indicates that the semantic roles of affectedness and control overlap on the subject participant. Moreover, -ku also emphasizes those roles by making them more prominent (i.e. adding deliberateness or propensity in inanimate subjects, as well as an effect that is subjectively stronger or more pronounced or long-lasting). My analysis is not much different from Parker’s (1973a): -ku emphasizes the role of the subject. However, this is a twofold role, based on (i) the subject’s control in the development of the event and (ii) its affectedness in an event that never fully projects upon a different entity.

4.6.2.2.3 -na and -ku ‘reciprocal (recp)’ Reciprocals are closely related to the middle voice. Alongside the middle suffix -ku, they require the presence of the distributive suffix -na, which only functions in this construction. It indicates that the members of a plural subject are acting reciprocally upon each other (e.g. maqa-naku‘to fight; to hit each other,’ from maqa- ‘to hit’; riqi-naku- ‘to meet; to become acquainted with each other,’ from riqi- ‘to know, be acquainted’). While the combination -naku is the way in which this suffix is mostly used, it is possible to insert the causative suffix in between. This will produce different interpretations regarding where the reciprocal interaction occurs. 19 The aspectual function of -ku was extended to its extreme in Ecuadorian Kichwa, where -ku is effectively an imperfective aspect marker (Cole 1982a).

111

carlos molina-vital (81) I yanaqinkunawanqa llapanmi pampatsa:natsikuyan. {i yanaqi-n-kuna-wan-qa and friend-3.poss-pl-com-top llapa-n-mi pampatsa:-na-tsi-ku-ya-n} all-3.poss-evd pair.up-recp-caus-mid-plv-3.sbj ‘And with their friends, all of themi match each otheri .’ In Example (81), pampatsa:-na-tsi-ku- ‘let each other pick a pair among themselves’ (from pampatsa:- ‘to even out; to match up’) includes the causative -tsi followed by -ku. This combination indicates that the subject allows something to happen to themselves (e.g. hayta-tsi-ku- ‘to let oneself be kicked,’ from hayta- ‘to kick.’) Thus, in -na-tsi-ku the distributive scope of -na is restricted to those causing the event. On the other hand, pampa:tsa-tsi-na-ku- ‘to make each other match up with someone else’ suggests that the subjects are letting or making each other pair up with a participant outside of those that make up the plural subject.

4.6.2.2.4 -ka: ‘spontaneous (spon)’ When followed by a consonant forming a syllabic coda, -ka: becomes -ka. This suffix has been documented in most varieties of Central Quechua (the exception are the Huanca varieties in central and southern Junín), in Northern Peruvian varieties, and in Yauyos Quechua (see Chapter 5 by Aviva Shimelman in this volume, who labels it, in contrast to the analysis for HQ offered here, ‘accidental’). For HQ, -ka: has been analyzed as a passive or semi-passive in all the previous grammatical descriptions. Accordingly, this suffix shows a clear preference to combine with transitive verbs. For instance, a transitive verb like paqa- ‘to wash’ yields paqa-ka:- ‘to be washed.’ Contrast the unmarked form, using kinwa ‘quinoa’ as the -ta-marked direct object in (82), and the form with -ka:, where kinwa is the nominative subject, in (83). (82) Kinwata paqarqantsik. {kinwa-ta paqa-rqa-ntsik} quinoa-acc wash-pst-1incl.sbj ‘We washed the quinoa.’ (83) Kinwa paqakarqantsik. {kinwa paqa-ka:-rqa-ntsik} quinoa wash-spon-pst-1incl.sbj ‘The quinoa was washed.’ Paqa-ka:- suggests that the quinoa was washed, but there is no way to indicate a specific agent involved. Thus, the use of -ka: is similar to the English get-passive: the quinoa “got washed” because something happened that produced such

112

change (e.g. rainfall). In fact, a definite animate individual serving as the agent is neither implied nor retrievable as an adjunct when -ka: is used. The sentence (83) could use an instrumental adjunct like tamya-wan ‘with the rain’ to indicate what caused the event, but this is unusual. An animate subject in a verb marked with -ka: is interpreted as not fully in control of the action they experience, as shown in (84). For this example, even a reflexive translation is possible. (84) Ashnu watakaykan. {ashnu wata-ka:-yka-n} donkey tie-spon-cont-3.sbj ‘The donkey is getting tied’ / ‘The donkey is getting itself tied.’ While many transitive verbs combine with -ka:, there are numerous instances of verbs where this suffix is never found or is considered odd. This is particularly noticeable with verbs that specify a way of doing things that requires a sentient subject in care of the event. Thus, verbs like lapta- ‘to touch with the hand,’ taki- ‘to sing,’ aru- ‘to work,’ muna- ‘to want,’ and qu- ‘to give’ are quite difficult to elicit with -ka: and never found in the corpus.20 To sum up, -ka: combines interpretations where (i) no external causer can be clearly identified, particularly as an animate individual, (ii) the subject is clearly affected as a result of the event taking place, (iii) it is possible to assign some level of responsibility to the subject, particularly if it is animate; however (iv) this responsibility (i.e. the subject’s control) seems to be hampered or diminished. Based on those characteristics, -ka: is best defined as spontaneous voice. This type of voice is conceptually related to other passive constructions (cf. Shibatani 1985; Toyota 2011). Canonical passive voices (e.g. John was seen) promote the patient argument of the active voice to the subject position, which leaves their original agent argument implicit. On the other hand, spontaneous voice constructions introduce a subject that experiences the event due to some external factor in the background. This factor is never an individual, but some force, situation, or context triggering the event experienced by the subject. Furthermore, the subject must find itself in a condition that makes it susceptible to experiencing the change brought by the external factor. In Example (85), a fox had his mouth stitched up by a trickster bird. After the fox walked around in pain for some time (the external factor), the stitches loosened up and the fox’s mouth tore open. This is represented by lliki-ka:- (from lliki- ‘to tear a thin surface’): 20 These restrictions can be overturned by the impersonal function of ka:, as shown further below in this section.

huaylas (ancash) quechua (85) I tsayna shimin llikika:kurinaq allibuwenu hatun. {i tsay-na shimi-n and dem-disc mouth-3.poss lliki-ka:-ku-ri-naq tear-spon-mid-punc-narr.pst allibwenu hatun} very big ‘And then his mouth tore open very big.’ -ka: also conveys true reflexivity in HQ. This means that it is used with transitive verbs representing events that strongly imply a transmission of force from one participant to another. Thus, in order to reduce the valency of a prototypical transitive verb, only -ka: is used, not -ku. Reflexive uses of -ka: are common for caused motion events. Examples are churaka:- ‘to put oneself ’ (from chura- ‘to put’), suta-ka:- ‘to stretch oneself, extend one’s own body’ (from suta- ‘to pull something’), and ayllu-ka:- ‘to get together, form an assembly’ (from ayllu- ‘to group things’). Example (86) provides further illustration. (86) Wamra kanastaman winaka:rqunaq. {wamra kanasta-man wina-ka:-rqu-naq} child basket-all fill-spon-rec.pst-narr.pst ‘The child got herself into the basket.’ (lit. ‘The child filled herself into the basket.’) The impersonal function of -ka: is the only way in which an animate agent can by implied in an otherwise spontaneous event. In impersonal situations, it is assumed that someone must have carried out the event, but its identity is irrelevant. With this function, any transitive verb can take -ka:. This seems to be a recent development in this variety, and it is found with some regularity in formal texts as in (87), perhaps under the influence of Spanish. (87) Ayan llapan hirkapa ashikaykayarqan. {aya-n llapan corpse-3.poss all-3.poss hirka-pa mountain-gen

ashi-ka:-yka:-ya:-rqa-n-mi} look.for-spon-ipfv-plv-pst-3.sbj-evd

‘His corpse was sought through all the mountains.’ Finally, although -ka: is generally associated with transitive verbs, it can be found with some intransitive verbs. Those cases have been regularly described as idiomatic expressions and deemed exceptional. Examples are punu-ka:- ‘to fall asleep’ (from punu- ‘to sleep’), wanu-ka:- ‘to wilt; to be exhausted’ (from wanu- ‘to die’), and yacha-ka:- ‘to become accustomed’ (from yacha- ‘to know’). Other apparent lexicalized, and quite frequent, forms are mantsa-ka:- ‘to get scared’ (from mantsa- ‘to fear something’) and usha-ka:- ‘to

get finished.’ With the exception of yacha-ka:- ‘to become accustomed to,’ all of these forms imply an adversative result. In other words, something overcame the subject and led to this result. In my corpus, there are at least 50 other cases of intransitive verbs and nouns that can take -ka: (to mention only a couple, tsaki-ka:- ‘to heal a wound by forming a scab,’ from tsaki- ‘to be dry,’ and pasa-ka:- ‘to pass a long time,’ from pasa- ‘to pass’). Virtually all of them combine the passing of time as the most common external factor with an adversative result.

4.6.2.2.5 -pu ‘benefactive (ben)’ This suffix alternates with -pa due to vowel lowering when followed by the cislocative suffix -mu and the 1st person object marker -ma: (see Section 4.3.4.3). The main function of -pu is to introduce an extra participant besides the basic arguments of a verb. This participant is interpreted as prominently affected by being the beneficiary or maleficiary of the event.21 In this sense, -pu introduces not only a goal of an event, but a highly affected one. The examples in (88)–(90) illustrate the use of -pu. (88) Hushi mamanpaq wayinta rurapun. {Hushi mama-n-paq José mother-3.poss-dat wayi-n-ta rura-pu-n} house-3.poss-acc make-ben-3.sbj ‘José makes her mother a house.’ (lit. ‘José made a house for his mother.’) (89) Wa:kaykikunata ta:papushunki. {wa:ka-yki-kuna-ta ta:pa-pu-shunki} cow-2.poss-pl-acc take.care-ben-3sbj.2.obj ‘She takes care of your cows for you.’ (90) Nunaqa pallarkapamurqun huk quri hachata. {nuna-qa palla-rku-pu-mu-rqu-n person-top pick-uw-ben-cisl-rec.pst-3.sbj huk quri one gold

hacha-ta} ax-acc

‘The person picked up a golden ax for him.’ The way in which -pu introduces the affected participant is based on the object person system (see Section 4.6.3.1.2). Thus, -pu uses -shu-nki to indicate that a 2nd person is the beneficiary in (89), while in (90) -pu refers to an unmarked 3rd person object. However, contrary to what we would expect, in (88) the overt beneficiary is not introduced by an object marked with the accusative suffix -ta; instead -paq is 21 Authors like Parker (1973a: 17) used the term “interpersonal” in order to avoid the positive connotation associated with “beneficiary.” However, due to its more widespread use in the literature, I will use the latter.

113

carlos molina-vital used (mama-n-paq ‘for his mother’). It appears as if only the basic direct object is marked with -ta, not the beneficiary. -pu is also frequently used with the copula ka- to express possession. The beneficiary marked with -pu is the possessor for whom something exists (see Section 4.6.1.1). There are two possible positions for -pu among derivational suffixes. In the examples above, it is in the second position. There, it can be preceded by aspect suffixes, the causative -tsi, and the inflectional verbal plural -ya:. In this position, its interpretation is based on the existence of an additional affected participant. However, when -pu takes the position next to the verbal stem, it implies that the event produces a negative or adversative affectedness toward a non-subject participant. An adversely affected direct object is introduced in (91), while in (92), the adverse reading of -pu applies to the core direct object or tari- ‘to find’:22 (91) Parlayarqan manana iglesyatapis kichapuya:ma:nanpaq. {parla-ya:-rqa-n mana-na iglesya-ta-pis talk-plv-pst-3.sbj not-disc church-acc-add kicha-pu-ya:-ma:-na-n-paq} open-ben-plv-1.obj-pot.nmlz-3.poss-dat ‘They talked so that they didn’t even open the church for us anymore.’ (92) Warmin amiganwan taripuykunaq. {warmi-n amiga-n-wan wife-3.poss mistress-3.poss-com tari-pu-yku-naq} find-ben-excep-narr.pst ‘Hisi wife caught himi with hisi mistress.’ Finally, the combination between -pu and -ku has three interpretations. First, it can convey that an action is carried out for the benefit of the subject in a way that is against the interests of another participant. This is an adversative reading. Thus miku- ‘to eat’ yields miku-pu-ku- ‘to eat someone else’s food; to drop uninvited to eat at someone’s place,’ and punu- ‘to sleep’ yields punu-pu-ku- ‘to crash at someone’s place.’ The second interpretation is based on transitive events that imply both self-benefit for the subject and benefit for a third party. Thus, naqtsa- ‘to comb (someone)’ yields naqtsa-pu-ku- ‘to cut hair for a living,’ and urya- ‘to work’ yields urya-pu-ku- ‘to work (as an employee).’ This could 22 Parker (1976: 118) presents a different instance of a suffix -pu and labels it “directive.” According to this author, this suffix only combines with verbs of bodily emission. Thus ishpa- ‘to urinate,’ isma- ‘to defecate,’ and tuqa- ‘to spit’ yield ishpa-pu-, isma-pu-, and tuqa-pu-, respectively. Clearly these cases are instances of adding -pu in the closest position to the stem to obtain an adversative interpretation: the subject has the intention to soil (or even to show contempt for) the target of those emissions.

114

be related to the customary action function of the middle -ku (see Section 4.6.2.2.2). Finally, -pu-ku- also indicates that an adverse or unexpected situation is brought upon an object and a subject, but due to the subject’s own responsibility. Thus, as expected in a middle voice, the subject is both the affected participant and the main responsible for what happens. However, the presence of -pu determines that what the subject experiences is whatever a beneficiary (or maleficiary) does. Examples (93) and (94) illustrate the contrast between a malefactive reading of -pu and the third function of -pu-ku, which conveys an adverse or unexpected situation that affects both the subject and an object: (93) Mamanta wanupurqan. {mama-n-ta wanu-pu-rqa-n} mother-3.poss-acc die-ben-pst-3.sbj ‘He died on his mother.’ (i.e. ‘His death affected his mother.’) (94) Wanupukuya:nanpaqqa sigarunkuna uchkurinshi. {wanu-pu-ku-ya:-na-n-paq-qa die-ben-mid-plv-pot. nmlz-3.poss-dat-top sigaru-n-kuna uchku-ri-n-shi} cigarettes-3.poss-pl be.a.hole-punc-3.sbj-evr ‘They say that, in order to have someone die on them, their cigarette gets a hole.’ (Context: Shamans cast a deadly spell by smoking their cigarettes in a way that it gets a hole. If this happens, their enemies will die, which benefits the shamans.)

4.6.2.2.6 -pa: ‘applicative (appl)’ This suffix loses its long vowel when followed by a consonant in a syllabic coda. However, in the Huaraz province it is frequent to find -pa alternating freely with -pa:. The only stable manifestation of -pa seems to be in the fused form -paku (see Section 4.6.2.3.5). -pa: has been described as conveying a repetitive, intermittent event, a careful or slow action, or an action directed to another participant (analogous to the benefactive suffix -pu). In addition, it can involve several affected objects. Clear aspectual properties for -pa: have been proposed (Julca Guerrero 2009: 240 and Parker 1976: 123–5). In (95)–(97) the iterative function is illustrated: (95) … waqtanta ashpipa:nanpaq {waqta-n-ta ashpi-pa:-na-n-paq} back-3.poss-acc scratch-appl-pot.nmlz-3.poss-dat ‘…so that he scratched her back.’

huaylas (ancash) quechua (96) Warminqa llullunta chichipa:rir punuratsinaq. {warmi-n-qa llullu-n-ta woman-3.poss-top new.born-3.poss-acc chichi-pa:-ri-r breastfeed-appl-punc-subis punu-ri-tsi-naq} sleep-punc-caus-narr.pst ‘His wife was making their new-born sleep after breastfeeding him.’ (97) Wawantana kutapan: riyakuy, nishpa. {wawa-n-ta-na kuta-pa:-n child-3.poss-acc-disc hit-appl-3.sbj riya-ku-y ni-shpa} wake.up-mid-imp say-subis ‘She shook her son saying: “wake up.”’ (Larsen 1976: 4) The directional function of -pa: is most apparent with verbs of emotion and communication. The former are considered typically intransitive, and the latter are arguably lower in their semantic transitivity. Examples of the directional function include e.g. asi-pa:- ‘to laugh at; to mock someone,’ from asi- ‘to laugh,’ aha-pa:- ‘to scold someone,’ from aha- ‘to feel mad,’ anku-pa:- ‘to feel compassion towards someone,’ from anku- ‘to feel nostalgic; to miss,’ tapu-pa:- ‘to question, inquire,’ from tapu- ‘to ask,’ or parla-pa:- ‘to talk to someone, to chat with,’ from parla- ‘to talk.’ Some uses of -pa: are listed in the dictionaries as lexicalizations too (e.g. taripa:- ‘to catch up with someone,’ from tari- ‘to find,’ or yachapa:- ‘to imitate, mimic,’ from yacha- ‘to know; to learn.’) Many cases of -pa:, however, do not fit the previous descriptions. In (98), kuti- ‘to go back,’ a motion verb, takes -pa: without involving any iteration, carefulness, several affected objects, or a directional function: (98) Tsaytanam kutipantsik huk killatanaw, mas allpan ayllupashpana i qurakunataqa hina ushaka:tsishpa. {tsay-ta-na-m kuti-pa:-ntsik dem-acc-disc-evd return-appl-1incl.sbj huk killa-ta-naw, mas one month-acc-simul more allpa-n ayllu-pa:-shpa-na dirt-3.poss gather-appl-subis-disc i qura-kuna-ta-qa hina and weed-pl-acc-top so usha-ka:-tsi-shpa} finish-spon-caus-subis ‘Later, we come back approximately in a month, and we lump more dirt [around the plant] and in the same way we exterminate the weed.’

Here kuti-pa:- suggests an action done to take care of the plants. The subject returns due to some motivation, but it does not suggest ‘coming back again and again.’ It only states that approximately after a month they go back to hill the plants. Contrast this with ayllu-pa:- (from ayllu ‘to gather’) in the same sentence, where an iterative meaning is evident. The notion of a motivation behind the event helps better understand the function of -pa:. Example (99) shows how an intransitive stative verb such as lasa- ‘to weight’ takes -pa: in order to add an object that motivates an experience associated with the weight of the subject taklla ‘foot plow’: (99) Takllaqa lasaparkunaq, Wallichuqa ninaq: “¿Imanirtaqshi taklla:qa lasapa:man?” {taklla-qa lasa-pa:-rku-naq foot.plow-top weight-appl-uw-narr.pst Wallichu-qa ni-naq valentín-top say-narr.pst imanir-taq-shi taklla-:-qa why-contr-evr foot.plow-1.poss-top lasa-pa:-ma-n} weight-appl-1.obj-3.sbj ‘The foot plow weighted down on him, and Valentín said: “Why could it be my foot plow weighs me down?”’ lasa-pa:- ‘to weight’ seems to have a directional function in (99) because it points at the experiencer of the weight of the foot plow. However, such a function falls under what -pu does (see Section 4.6.2.2.5, (94), wanu-pu- ‘to die on someone’). The reason why -pa: and not -pu is used in (99) is that the participant affected by the weight is not just receiving such weight; instead, the foot plow’s remarkable weight (the state it is in) comes from the perspective of the person feeling it (the 3rd person implicit direct object, and then -ma:, the 1st person direct object). Thus, it is only out of the object’s experience that the subject can be perceived in a certain state (in this case, as a particularly heavy foot plow). Every case of -pa: with an emotion or communication verb suggests that its object is the motivation behind what the subject does or experiences. Thus, asi-pa:- is not only ‘to laugh at something or someone that produces laughter’: that participant must also be affected by the subject’s laughter. Likewise, tapu-pa:- is to ask someone because they are a source of information; but this questioning is impertinent or inquisitive, which affects the object serving as source. This directive function of -pa: is more productive than acknowledged by Parker (1976). For instance, it appears in medi-pa:-, from medi- ‘to measure.’ Even if someone receives what the subject measures, -pu is not used, as shown in (100).

115

carlos molina-vital (100) I imallatapis siemprish warminta medipa:kuykaq. {i ima-lla-ta-pis and what-lim-acc-add siempre-sh warmi-n-ta always-evr woman-3.poss-acc medi-pa:-ku-yka:-q} measure-appl-mid-ipfv-hab ‘And he was always measuring anything he gave to his wife.’ The context for (100) is that a notoriously stingy man measured whatever he gave to his wife for cooking. By being given such limited ingredients she was very inconvenienced. Thus, medi-pa:- is not directional because the subject gives something that affects the object (beneficiary or maleficiary); instead, the subject is acting out of what he thinks the object deserves (i.e. is motivated by the object). The applicative voice turns oblique participants expressing circumstantial notions into core arguments of a verb. I argue that in HQ -pa: is an applicative that turns a participant serving as motivation or source into a prominently affected participant. Therefore, -pa: is the opposite of benefactive -pu, which introduces a participant as an affected goal. This contrast is not too different from that between por and para in Spanish, which both mean ‘for,’ but the former closer to ‘due to’ and the latter to ‘in order to.’ This subtle but real distinction can be confusing.23 By considering -pa: an applicative introducing an affected source or motivation for the subject’s actions or experiences, its iterative function can be accounted for: -pa: suggests that an event iterates as long as is required to achieve what the affected participant motivated. This is exactly what the ‘careful action’ function of -pa: suggests, but previous accounts failed to connect it to the iteration of the event. Thus, rika:-pa:-, from rika:- ‘to see,’ is ‘to watch carefully,’ which can be achieved by looking several times because of a reason that makes the subject pay such attention. Finally, the applicative voice function of -pa: prevented previous accounts from recognizing its conceptual proximity with the benefactive -pu. This made Parker (1973a; 1976) mistakenly identify some cases of -pa as lowering of -pu before the middle suffix -ku or the imperfective aspect marker -yka:. The only cases that correctly reflect -pu-lowering are cases where it has a clear beneficiary reading, and is followed by -ma:. All other examples Parker provides are either -pa: rendered as -pa because of vowel shortening (before -yka:) or instances of -paku. The latter form clearly uses -pa with an 23 It is likely that -pa: originates from *-pa-ya, a form that is commonly used in contrast with -pa in Southern Quechua. In any case, the fact that -pa has taken over the functions of -pu in Huallaga Quechua (cf. Weber 1989), or that in HQ adversative functions of -pu-ku are now virtually indistinguishable from -pa:-ku is evidence in favor of the strong connection between those two voice suffixes.

116

applicative sense, but it seems to have been competing with -pu-ku, at least in the Huaraz province (see Section 4.6.2.3.5)24 In any case, all of those cases are uses of -pa: or a cognate conveying an affected source, and not cases of underlying -pu.

4.6.2.3 Other derivational suffixes Here I include derivational suffixes like the cislocative -mu, the delimitative -lla:, and other forms that have a limited distribution like downward direction -rpu, stationary event -rpa:, attentive action -paku, and careful action -chaku.

4.6.2.3.1 -mu ‘cislocative (cisl)’ As discussed in Section 4.3.4.3,-mu is one of the suffixes that trigger vowel change into /a/ in derivational suffixes ending with /u/ placed in any preceding position. This suffix indicates that the event conveyed by the verbal stem is oriented toward the realm of a 1st person/speaker implicit in the sentence. This is not the same as “directed towards the subject” (which is associated with the middle -ku, see Section 4.6.2.2.2). With motion or caused-motion events its basic function is clear: apa-mu- ‘to bring’ (from apa- ‘to take; to carry’); ayllu-mu- ‘to gather and bring here’ (from ayllu‘to gather’); kuti-mu- ‘to come back’ (from kuti- ‘to return’). Motion events without -mu have their endpoint somewhere outside the realm of the 1st person. -mu is common with non-motion events, which makes them orient toward the speaker, as in (101) or (102). (101) Manam ni pi tsaynawqa ruramushqatsu. {mana-m ni pi not-evd nor who tsay-naw-qa rura-mu-shqa-tsu} dem-simul-top make-cisl-3.sbj.prf-neg ‘Nobody has come making it that way.’ (102) Paniykita ashiya:murqun. {pani-yki-ta sister.of.man-2.poss-acc ashi-ya:-mu-rqu-n} search-plv-cisl-rec.pst-3.sbj ‘They have come looking for your sister.’ The 1st person does not need to be a physical endpoint, but it could be a psychological one. Therefore, the subject’s interest can motivate using -mu. This is central to its use with verbs where the 1st person is not directly or even obviously aimed at in the motion, as in (103)–(106) 24 Swisshelm (1974: 491) was the first to voice doubts about -pa: lowering to -pu before -ku.

huaylas (ancash) quechua (103) Qullqapita hicharpamurqun chuskun sakuta kachita i utsuta. {qullqa-pita hicha-rpu-mu-rqu-n storage.room-abl throw-dw-cisl-rec.pst-3.sbj chusku-n saku-ta four-3.poss sack-acc kachi-ta i utsu-ta} salt-acc and hot.pepper-acc ‘From the storage room upstairs he was throwing down four sacks of salt and hot peppers.’ (104) Nikar nikarllanam tsuqllu yuriramun. {nika-r be.immediate-subis nika-r-lla-na-m be.immediate-subis-lim-disc-evd tsuqllu yuri-ri-mu-n} corn.cob appear-punc-cisl-3.sbj ‘In no time, a young corn ear appears.’ (105) Tsaynam rogayapti: bahamurqan kima waranqaman. {tsay-na-m roga-ya:-pti-: dem-disc-evd beg-plv-subds-1.sbj baha-mu-rqa-n kima lower-cisl-pst-3.sbj three waranqa-man} thousand-all ‘Then when we begged he brought [the price] down to three thousand.’ (106) Sharayka:mu:mi. {sha-ra:-yka:-mu-:-mi} come-dur-ipfv-1.sbj-evd ‘I am coming [for you]!’ Example (106) with shamu- ‘to come’ is revealing. -mu is virtually lexicalized to the stem *sha- (cognate with sha:- ‘to stand’). The position of -mu, however, can change in shamu-, particularly when the durative -ra: and imperfective -yka: suffixes appear, as seen in (106). The context of this sentence is as follows: A cougar has been challenged by a ram that wants to cross a bridge. The cougar is charging at the ram and says sharayka:mu:mi. Still, the actual motion moves away from where the speaker enunciates (the cougar is moving toward the ram). The cislocative suffix here implies the goal which the subject, who is also the speaker in (106), wants to reach. The reference point is construed in a subjective way, based on the speaker’s aim.

This observation is crucial for understanding other uses of -mu where it also indicates that, according to the speaker, the subject manages to do something. This is evident when the event’s occurrence is contrary to expectation, which affects in some way the speaker and those in their group. This is shown in (107) and (108). (107) Warmikunapis, tomar qallaykuyan… machashqana, warmikuna qallaykaya:mun takirnin. {warmi-kuna-pis toma-r woman-pl-add drink-subis qalla-yku-ya:-n… macha-shqa-na begin-excep-plv-3.sbj drunk-res-disc warmi-kuna qalla-yku-ya:-mu-n woman-pl begin-excep-plv-cisl-3.sbj taki-r-ni-n} sing-subis-ep-3.sbj ‘The women begin drinking… and once drunk, they begin to sing.’ (108) Waray mana mikuyta tarimuptiykiqa kikikitam mikushqayki. {waray mana miku-y-ta morning not eat-inf-acc tari-mu-pti-yki-qa find-cisl-subds-2.sbj-top kiki-yki-ta-m miku-shqa-yki} self-2.sbj-acc-evd eat-fut-1.sbj.2.obj ‘If tomorrow you don’t manage to find food, I will eat you and nobody but you.’ Furthermore, use of -mu as indicating that the subject manages what is denoted by the verb might be related to other uses of -mu where the action is regarded as urgent or pressing for the speaker. In this sense, -mu is still connected with the basic notion of an action moving toward the speaker, although in an entirely subjective way. In (109) the speaker is the one giving the command using -mu, tsariya:muy ‘you guys catch it!’ Nothing in the example suggests an actual motion toward the speaker. Instead, the speaker is interested in whatever the subjects in the imperative verb manage to do. Likewise, Example (110) indicates that the subject, who is also the speaker, is not simply going to spin something, but whatever they spin is motivated by a specific necessity pertaining to the speaker. (109) Kuchintsik tsariya:muy! {kuchi-ntsik tsari-ya:-mu-y} pig-1pl.incl.poss grab-plv-cisl-imp ‘Catch our pig!’ (adapted from Larsen 1976: 34)

117

carlos molina-vital (110) Imatataq puntata putskamushaq? {ima-ta-taq punta-ta what-acc-contr before-acc putska-mu-shaq} spin.yarn-cisl-1.sbj.fut ‘What [yarn] will I spin first?’ (adapted from Larsen 1976: 35)

4.6.2.3.2 -lla: ‘limitative (lim)’ This suffix seems to be the same as the limitative -lla for nominal derivation (see Section 4.5.2.5). In a few sources, -lla: becomes -lla before a consonant in coda position, but mostly, this suffix is always -lla. Its function with verbs includes that of restricting the event to a single instance or type (miku-lla-nki ‘you will only eat,’ from miku- ‘to eat’). However, a subjective function is more frequently found: the action is viewed as less face-threatening, implying reservations toward another participant, and as expressing general politeness (e.g. yayku-lla:-ya-nki ‘you will come in, please,’ from yayku- ‘to enter,’ inflected with the verbal plural -ya: and -nki ‘2nd person future’).

4.6.2.3.3 -rpu ‘downward (dw)’ This suffix is productive mainly, but not exclusively, with motion events. It indicates that the action is carried out downwards (e.g. yaka-rpu- ‘to store at the bottom,’ from yaka‘to put inside; to put away,’ cha-rpu- ‘to come down to,’ from cha:- ‘to arrive,’ qawa-rpu- ‘to watch downwards,’ from qawa‘to watch’).

4.6.2.3.4 -rpa: ‘stationary (stn)’ This unproductive suffix indicates that the subject remains affected for a long time or permanently. Its use is marginal in my corpus (hita-rpa:- ‘to remain laying down,’ from hita‘to throw,’ and kicha-rpa:- ‘to remain open,’ from kicha- ‘to open’). Parker (1976: 120) records it as -rpa in a couple of roots beside hita- ‘to throw:’ chura-rpa- ‘to place oneself permanently,’ from chura- ‘to put,’ and hama-rpa- ‘to rest (as a group), from hama- ‘to rest,’ which indicates a collective action. -rpa:, like the durative suffix -ra:, produces a resultative passive: It makes the direct object of the base verb appear as subject experiencer (see Section 4.6.2.1.1.3).

4.6.2.3.5 -paku ‘attentive action (atten)’ According to Julca Guerrero (2009: 245), -paku has a function similar to that of -chaku (see Section 4.6.2.3.6): it indicates that the action is carried out with special attention to a direct object that includes multiple details. However, this author suggests that -paku can refer also to an event performed

118

without volition and circumstantially (e.g. wiya-paku- ‘to pay attention to any message,’ from wiya- ‘to hear,’ and rika-paku‘to look around with attention,’ from rika:- ‘to see’). Larsen (1976: 7) is the only other author who considered -paku as an independent derivational suffix. In her view, this suffix suggests that the action has been carried to completion and done in a total manner, usually through multiple objects (something implied in Julca Guerrero’s previous definition). Examples (111)–(113) illustrate how a multiple object is fully affected, how a subject becomes a prominent experience through acting upon several objects, and how the subject is affected as a whole. (111) Tiyanqa uchkukunata latupakuq. {tiyan-qa uchku-kuna-ta latu-paku-q} aunt-top hole-pl-acc fill-atten-hab ‘Her aunt filled the holes.’ (i.e. all the holes that were out there) (adapted from Larsen 1976: 7) (112) Atalláw, upupakurishaqchi. {atalláw upu-paku-ri-shaq-chi} nice drink-atten-punc-1.sbj.fut-evc ‘How nice! It looks like I will drink as much as I can.’ (adapted from Larsen 1976: 7) (113) Virginqa hiqarpamun yana enteru limpu wankupakushqa. {virgin-qa hiqa-rpu-mu-n yana virgin-top fall-dw-cisl-3.sbj black enteru limpu whole totally

wanku-paku-shqa} wrap-atten-res

‘The Virgin fell from the skies completely wrapped in black.’ (adapted from Larsen 1976: 7) As Parker (1976: 123) noticed, -pa is nonexistent as an independent suffix in HQ, and it alternates with -pa:, which is considered the base form in that variety. It is possible, then, that -paku is only an alternation of -pa:ku. Compare (114) with (113) above: (114) Alli aqshupa:kurkur hiqarinaq uqi punchurishqa nunawan. {alli aqshu-pa:ku-rku-r good blanket-atten-uw-subis hiqari-naq uqi go-narr.pst grey punchu-ri-shqa nuna-wan} poncho-punc-res man-com ‘Wrapping up well [with her blanket], she left with the man in the grey poncho.’

huaylas (ancash) quechua Example (114) comes from a HQ author (and consultant) whose texts shows no alternation between -pa: and -pa (as found in corpus data from the Huaraz province). However, that very consultant still provides examples with -paku. The source participant in the event exists, but it is not prominently affected. As a source, it only provides some type of motivation (or support) for what the subject experiences. In this sense, -paku is exactly what would be expected from the combination of -pa (different from -pa:, which requires an affected source-object) and -ku (which places affectedness on the controlling subject). (115) —¡Ananáw, wawqi! Alla:pataq kayqa nanaykun, ¡ananáw! —Tsarapakuy wawqi, waray waratinqa puka qasqunam —puriykanki. {ananáw wawqi alla:pa-taq ouch brother a.lot-contr kay-qa nana-yku-n dem.prox-top hurt-excep-3.sbj ananáw tsara-paku-y wawqi ouch hold-atten-imp brother waray waratin-qa tomorrow day.after.tomorrow puka qasqu-na-m red chest-disc-evd puri-yka:-nki} walk-ipfv-2.sbj.fut ‘—Ouch, brother! This hurts a lot, ouch!’ ‘—Hold yourself, brother, tomorrow or the day after tomorrow you will be walking sporting a red chest.’ (Context: The fox complains because a red-chested bird is skinning him, so he also has a red chest.) (116) U:sha, kuchi, kakashqa mallaqa:yaqtsu qurakunata mikupakur. {u:sha kuchi kakash-qa sheep pig rooster-top mallaqa:-ya:-q-tsu qura-kuna-ta hunger-plv-hab-neg herb-pl-acc miku-paku-r} eat-atten-subis ‘The sheep, the pig, and the rooster were not hungry because they ate herbs.’ Perhaps the productive contrast that Parker (1976: 122, 124) found in Northern Conchucos Quechua between -pa (an action directed to multiple objects but not protracted, even immediate) and -pa: (virtually identical, but implying a protracted iteration) is still present in the combination -paku in

HQ. It is as if the use of -pa involves the full use of a source that motivates or supports the action, without this involving a prominent affectedness of the source, as implied by -pa:. Clarifying if -paku is actually different from -pa:-ku requires further analysis.

4.6.2.3.6 -chaku ‘care’ This suffix has an extremely reduced distribution, and it suggests that the action is done carefully, but also through repetition. In this regard, it is similar to the applicative -pa: (see Section 4.6.2.2.6). In my corpus it only appears as rikachaku- ‘to spot; to descry’ (from rika:- ‘to see’) and as yarpachaku- ‘to deliberate, reflect on’ (from yarpa- ‘to remember’). Other instances provided in previous works (Parker 1973a: 35 and Swisshelm 1974: 573) include wiya-chaku- (‘to be attentive to any sound,’ from wiya- ‘to hear’), apa-chaku- ‘to carry (a baby)’ (from apa- ‘to carry’), and muski-chaku- ‘to snoop’ (from muski- ‘to smell’).

4.6.2.4 Order of derivational suffixes Table 4.11 shows my proposal for the default distribution of HQ derivational suffixes. While Parker (1973a) and Swisshelm (1974) provided the foundations for my analysis, several modifications have been introduced. Table 4.11 includes the plural suffix -ya: and the 1st person object suffix -ma:. This is required because, even if these function as inflectional suffixes, their formal distribution is that of derivational ones. There are 12 slots, which are classified based on their overarching grammatical function (voice or aspect), their dynamic or stative meaning, or their specific function (plural, cislocative, and object marker). The following list introduces clarifications on how to read Table 4.11, as well as some specific restrictions that cannot be visually represented: (i) The distribution reads from left to right. For instance -pa:-ku, -ka:-ku, and -tsi-ku are valid combinations. *-ku-pa:, *-ku-ka:, and *-ku-tsi (i.e. *-ka-tsi, with vowel lowering) are not.25 (ii) A suffix’s meaning has scope over the items located to its left. (iii) Suffixes located one above the other, or in the same column do not combine. (iv) Suffixes in the same cell do not combine. This specifically applies to the “inverted-L” shaped cells covering columns one and two. 25 Some exceptions can be found for -ku + -tsi (e.g. pinqa-ka-tsi- ‘to make someone ashamed’), but they are few and likely to be lexicalized instances of -ku preceding the causative -tsi.

119

carlos molina-vital Table 4.11 Distribution of verbal derivational suffixes in Huaylas Quechua

1

2

Voice and stative meaning

3

4 Voice

5

6

7

Perfective aspect

8

9

10

11

12

lim

pl

cisl

ben

obj

Imperfective aspect Dynamic meaning -ykacha

(v) Suffixes in different-colored zones do not combine with each other due to their inherent semantic incompatibility. (vi) Suffixes located in the dark gray zone can’t combine among themselves, even if they appear in different columns. For instance, with its directional meaning, the upward suffix -rku can’t combine with the benefactive -pu when it has an adversative meaning. Likewise, the outward suffix -rqu can’t combine with careful action -chaku, nor -ykacha, nor with directional -yku. (vii) Columns have been grouped based on their shared function: (a) Columns 1 and 2 contain suffixes located next to the verb stem. The light gray area includes those with a stative meaning that imply a subject’s passive-like affectedness (-pa:, -ka:, -rpa:, -ra:, and -na:). The dark gray zone includes suffixes that indicate some dynamic and/or volitional meaning like direction of motion, as well as manner of activity. (b) Columns 3–5 include the causative, the reflexive, the reciprocal construction -naku, and the fused form -paku. They can be considered voice suffixes. The case of -paku as a fused suffix needs further study. (c) Columns 6 and 7 include suffixes that have a clear aspectual meaning: Perfective (-ri, -rku, and -yku) or basic imperfective (continuous -yka:).

120

(d) Columns 9 and 12 include those suffixes that are best understood as inflectional in terms of their function, but that align with derivational ones due to their formal properties (the plural suffix -ya: and the 1st person object suffix -ma:). (e) Column 10 includes the cislocative suffix -mu, which always occurs toward the end of the sequence of derivational suffixes, but never after the benefactive suffix -pu when it introduces an extra participant. (viii) The reciprocal suffix -na (column 4) must be combined with -ku (column 5) to form the reciprocal suffix-naku. While no other suffix occurs in between them, some exceptions with the aspect markers -tsi and -yka: are possible (see Section 4.6.2.2.3). (ix) In the same verb, -chaku never combines with -ku or -paku; nor do -rku and -yku or -mu and -ma:.

4.6.3 Inflectional categories of the verb HQ verbs inflect for person, number, and tense. Conditionals and imperatives instantiate grammatical mood, and can be considered inflectional in their distribution. Table 4.12 presents the inflectional paradigms for the basic forms. There are also periphrastic verbal constructions used to convey the present perfect (which can also use a basic paradigm), the non-experienced or narrative past,

huaylas (ancash) quechua Table 4.12 Basic inflectional paradigms in Huaylas Quechua Number

Singular

Plural

Person

Tense

Mood

Present -Ø

Experienced past -rqa

Recent past -rqu

Future

Conditional -man

Imperative -y

1

kuya-: ‘I love.’

kuya-rqa-: ‘I loved.’

kuya-rqu-: ‘I have loved.’

kuya-shaq ‘I will love.’

kuya-:-man ‘I would love.’

2

kuya-nki ‘You love.’

kuya-rqa-yki ‘You loved.’

kuya-rqu-nki ‘You have loved.’

kuya-nki ‘You will love.’

kuya-nki-man ‘You would love.’

kuya-y ‘(You) love.’

3

kuya-n ‘S/he loves.’

kuya-rqa-n ‘S/he loved.’

kuya-rqu-n ‘S/he has loved.’

kuya-nqa ‘S/he will love.’

kuya-n-man ‘S/he would love.’

kuya-tsun ‘Let him/her love.’

1 incl

kuya-ntsik ‘We (incl) love.’

kuya-rqa-ntsik ‘We (incl) loved.’

kuya-rqu-ntsik ‘We (incl) have loved.’

kuya-shun ‘We (incl) will love.’

kuya-ntsikman kuya-shwan ‘We (incl) would love.’

1 excl

kuya-ya-: ‘We (excl) love.’

kuya-ya-rqa-: ‘We (excl) loved.’

kuya-ya-rqu-: ‘We (excl) have loved.’

kuya-ya-shaq ‘We (excl) will love.’

kuya-ya-:-man ‘We (excl) would love.’

2

kuya-ya-nki ‘You love.’

kuya-ya-rqayki ‘You loved.’

kuya-ya-rqunki ‘You have loved.’

kuya-ya-nki ‘I will love.’

kuya-ya-nkiman ‘You would love.’

kuya-ya-y ‘(You) love.’

3

kuya-ya-n ‘They love.’

kuya-ya-rqa-n ‘He loved.’

kuya-ya-rqu-n ‘They have loved.’

kuya-ya-nqa ‘They will love.’

kuya-ya-n-man ‘They would love.’

kuya-ya-tsun ‘Let him/her love.’

the habitual past, and the perfect conditional. All of those forms use ka- ‘to be’ as their auxiliary. Table 4.13 presents the inflectional paradigms for the periphrastic forms.

alternations in the position of -ya: occur in the habitual past (see Section 4.6.3.2.6). As expected with other derivational suffixes with long vowel, it shortens if followed by a consonant coda.

4.6.3.1 Person and number

4.6.3.1.1 Subject marking

A three-person system, plus the distinction between the 1st person inclusive (-ntsik) and 1st person exclusive (-ya-:) is used (see also Section 4.5.1.1.1 on personal pronouns). HQ verbs use -ya: as their exclusive marker of plurality (the exception being 1st person inclusive -ntsik, itself a plural form). Its position aligns it with other derivational suffixes (specifically, it must be placed before the cislocative suffix -mu). Tables 4.12 and 4.13 illustrate the use of the verbal plural suffix -ya: with basic and periphrastic paradigms. Some

As is customary for Central Quechua languages, the 1st person subject is marked through vowel lengthening. The 2nd person subject uses -nki, except in the experienced past tense -rqa and one of the forms of the past perfect -shqa, where -yki is preferred. The 3rd person subject is marked with -n. However, periphrastic forms like non-experienced past, habitual past, and the past perfect do not use -n because their 3rd person depends on the unexpressed 3rd person form of ka- ‘to be,’ their auxiliary.

121

carlos molina-vital Table 4.13 Periphrastic inflectional paradigms in Huaylas Quechua Number

Singular

Plural

a

Person

Tense

Mood

Non-experienced past -naq + ka-

Habitual past -q + ka-

Past perfect -shqa or-shqa + ka-

Perfect Conditional -man karqan

1

kuya-naq ka-: ‘I had loved.’

kuya-q ka-: ‘I used to love.’

kuya-shqa-: or kuya-shqa ka-: ‘I have loved.’

kuya-:-man karqan ‘I would have loved.’

2

kuya-naq ka-nki ‘You had loved.’

kuya-q ka-nki ‘You used to love.’

kuya-shqa-yki or kuya-shqa ka-nki ‘You have loved.’

kuya-nki-man karqan ‘You would have loved.’

3

kuya-naq ‘S/he had loved.’

kuya-q ‘S/he used to love.’

kuya-shqa-n or kuya-shqa ‘S/he has loved.’

kuya-n-man karqan ‘S/he would have loved.’

1 incl

kuya-naq ka-ntsik ‘We (incl)a had loved.’

kuya-q ka-ntsik ‘We (incl) used to love.’

kuya-shqa-ntsik or kuya-shqa ka-ntsik ‘We (incl) have loved.’

kuya-nstik-man karqan or kuya-shwan karqan ‘We (incl) would have loved.’

1 excl

kuya-ya:-naq ka(-ya)-: ‘We (excl) had loved.’

kuya-ya-q ka(-ya)-: ‘We (excl) used to love.’

kuya-ya-shqa-: or kuya-shqa ka-ya-: ‘We (excl) have loved.’

kuya-ya-:-man karqan ‘We (excl) would have loved.’

2

kuya-ya:-naq ka-ya-nki ‘You had loved.’

kuya-ya-q ka-nki ‘You used to love.’

kuya-ya-shqa-yki or kuya-shqa ka-ya-nki ‘You have loved.’

kuya-ya-nki-man karqan ‘You would have loved.’

3

kuya-ya:-naq ‘They had loved.’

kuya-ya-q ‘They used to love.’

kuya-ya-shqa-n or kuya-ya-shqa ‘They have loved.’

kuya-ya-n-man karqan ‘They would have loved.’

Plural markers that are frequent but optional are given in parentheses.

4.6.3.1.2 Object marking As usual in many Quechuan languages, person inflections include a paradigm for object marking. The traditionally labeled “transitions” system conveys the interaction between subject and object. Thus a verb like kuya- ‘to love’ takes -q to indicate a 1st person subject and a 2nd person

122

object: kuya-q ‘I love you.’ Conversely, if a 2nd person targets a 1st person, -ma: is used to indicate the 1st person object, while the basic 2nd person subject marking -nki remains: kuya-ma-nki ‘You love me’ (-ma: undergoes vowel shortening). If a 3rd person targets a 2nd person, then -shunki is used: kuya-shu-nki ‘s/he loves you.’ Although -shu-nki

huaylas (ancash) quechua

SUBJECT

Table 4.14 Singular and plural subject–object agreement marking in Huaylas Quechua Subject-Object OBJECT agreement (singular)

OBJECT

SUBJECT

Subject-Object agreement (plural)

appears as a single suffix, it is actually divided in two suffixes, one before the tense marker and the other right after it (e.g. kuya-shu-rqa-yki, from kuya- ‘to love’ in past tense -rqa). Also, with non-inflected forms like different-subject subordinator -pti and other nominalizers, -shu-yki is used. Finally, a 3rd person object is never marked on the verb (zero marking). Thus, if its object referent is implicit, the verb is only marked for subject. This marking system is irregular. The 1st person object marker -ma: combines with the regular person subject markers. However, -q (1st person subject to 2nd person object) and -shu-nki (3rd person subject to 2nd person object) fulfill the subject and object function and are the only person marker allowed. Furthermore, in the future tense -yki is used instead of -q. Finally, in order to indicate interaction between plural subjects or objects, only -ya: is used. This produces several instances of ambiguity. For instance, kuya-ya:-ma-nki (2nd person subject and 1st person object) yields ‘you (pl) love me,’ ‘you love us,’ and ‘you (pl) love us.’ Table 4.14 shows the distribution of subject–object person markers including plurals.

4.6.3.2 Tense HQ, like other Quechuan languages, distinguishes between three basic tenses: Present (aywa-n ‘s/he goes’), past (aywarqa-n ‘s/he went’), and future (aywa-nqa ‘s/he will go’). This system shows a high degree of regularity with little fusion of tense and person. Tense suffixes are used first

among inflectional suffixes, with the exception of -shunki, whose first part immediately precedes tense (see Section 4.6.3.1.2).

4.6.3.2.1 Present Present tense is zero-marked, and it just needs person agreement added to the stem. Since the present is not strictly anchored in a precise time coinciding with the enunciation moment, it is frequently used to express an event located in the near past. This is common in narrations, where the present is deployed to create a sequence of events. Also, the present can express habitual situations such as customary actions or situations that extend through time, as in (117). (117) Tsay tsaka kawan. {tsay tsaka kawa-n} dem bridge live-3.sbj ‘That bridge exists.’ (lit. ‘That bridge lives.’)

4.6.3.2.2 -rqa ‘experienced past (pst)’ -rqa refers to an experienced event that is located clearly in a previous moment without impinging in the current moment of enunciation. -rqa conveys only tense, not aspect. Thus, its interpretation can be that of a perfective or imperfective past based on the context of use. In Example (118) uryarqan is best viewed as an imperfective that provides background to the events in the narration:

123

carlos molina-vital (118) Tsaypita uryayarqan pagadunam munaq kaqkuna. {tsay-pita urya-ya:-rqa-n pagadu-na-m dem-abl work-plv-pst-3.sbj paid-disc-evd muna-q want-ag

ka-q-kuna} be-ag-pl

‘After that, only those who wanted worked, but for pay.’ (Context: Changes in contracts were enforced, so from that point onwards only those wanting to work were doing it and getting paid.) It is characteristic for -rqa to refer to an event whose occurrence is considered certain by the speaker (see Section 4.7.3.2 for the notion of direct knowledge in evidentials). Example (118) above refers to the process of road constructions in Huaraz during the early 20th century.

4.6.3.2.3 -rqu ‘recent past/perfective of immediateness (‘rec.pst)’ The inflectional suffix -rqu originated from the derivational outwards movement -rqu. It is a perfective past tense that points at a past that is close to the moment of enunciation, or that impinges upon a current time. It takes the same position as -rqa, with which it never combines (e.g. rika-rquntsik ‘we just saw, have seen,’ rika-rqa-ntsik ‘we saw’). It is often found in combination with the non-experienced past -naq, giving it a perfective interpretation (see Section 4.6.3.2.4).

4.6.3.2.4 -naq ‘non-experienced past or narrative past (narr.pst)’ -naq indicates that an event belongs to a narration (i.e. is fictional), or it took place with its subject not being completely aware of it (e.g. dreams, hallucinations, drunkenness, or early childhood). Like the experienced past -rqa, it is not marked for aspect. To obtain a perfective interpretation, -rqu is used before -naq. This is a periphrastic form that requires the auxiliary ka- ‘to be’ (e.g. punu-naq ka-: ‘I slept,’ punu-naq ka-nki ‘you slept,’ and punu-naq ‘s/he slept’; see Table 4.13 for the full paradigm). Examples of -naq are in (41), (46), (60), and (62).

4.6.3.2.5 -shqa ‘past perfect (prf)’ The form -shqa is cognate with -sqa, the narrative or nonexperienced past present in Southern and Northern varieties (see Chapter 5 by Aviva Shimelman in this volume for Yauyos Quechua). The past perfect shows three different patterns in the Ancash region (Parker 1976: 109–10); my HQ corpus shows that only two of them are instantiated. The non-periphrastic one combines person agreement with the

124

perfect suffix -shqa (e.g. miku-shqa-: ‘I have eaten,’ miku-shqanki ‘you have eaten,’ and miku-shqa-n ‘s/he has eaten’). The periphrastic pattern is also attested, where -shqa co-occurs with the auxiliary ka- ‘to be’ conjugated for subject person (e.g. miku-shqa ka-: ‘I have eaten,’ muku-shqa ka-nki ‘you have eaten,’ and miku-shqa ‘s/he has eaten’). The past perfect has been considered as quite similar in meaning to the recent past -rqu (Parker 1976: 109; Julca Guerrero 2009: 207). My analysis is that this is a perfective past that does not make any assumptions about its proximity to a current situation. However, affectedness of the subject seems to be relevant up to what is the current time of enunciation, as (119) and (120) show. (119) Tayta:, maychawtaq tarishqa kanki? {tayta-: may-chaw-taq father-1.poss where-loc-contr tari-shqa find-prf

ka-nki} be-2.sbj

‘Sir, where have you found it?’ (Context: The speaker is impressed with the beautiful horse a stranger has arrived with.) (120) Llapan yachaynikiwanmi limpu trampaman ishkishqa kanki. {llapa-n yacha-y-ni-yki-wan-mi all-3.poss know-inf-ep-2.poss-com-evd limpu trampa-man ishki-shqa ka-nki} fully trap-all fall-prf be-2.sbj ‘With all your knowledge, you have completely fallen into the trap.’

4.6.3.2.6 -q + kay ‘habitual past construction (hab)’ A habitual event in the past can be expressed with a periphrastic construction involving the agentive nominalizer -q with the main verb, and the copulative verb ka‘to be’ in the present tense and marked for subject (e.g. kuya-q ka-: ‘I used to love,’ kuya-q ka-nki ‘you used to love,’ kuya-q ‘s/he used to love’; see Table 4.13 for the full paradigm). The use of the verbal plural marker -ya: in the habitual past construction shows dialectal variation, and my corpus reflects this. Some speakers use the plural marker with the main verb, before -q, others with the auxiliary ka-, while still others use it with both. Sometimes, even the same source shows those three distributions. For instance, ‘we (excl) used to play’ can be expressed as puklla-ya-q ka:, pukllaq ka-ya-:, or puklla-ya-q ka-ya-:. Regarding the use of subject–object markers, they go with the main verb (e.g. kuya-ma-q ka-nki ‘you used to love me,’ kuya-shu-q ka-nki ‘s/he used to love you,’ kuya-ma-q ka-ntsik

huaylas (ancash) quechua ‘s/he/they used to love us (incl)’). Due to the coincidence between the nominalizer -q and the 1st person subject and 2nd person object marker -q, only an accusative-marked overt 2nd person pronoun is used (e.g. qam-ta kuya-q ka-: ‘I used to love you’).

4.6.3.2.7 ‘future (fut)’ The future is the only irregular tense in HQ (see Table 4.12). The 2nd person is the same as the one in the present tense. Each other person has its own future marker (1st person -shaq, 1st person inclusive -shun, 3rd person -nqa), which makes the future a case of fusional morphology. See Table 4.12 for its full paradigm. The 1st person future suffix -shaq alternates with -shqa and with -sha when combined with the irregular 1st subject to 2nd object marker -yki (e.g. maqa-shqa-yki or maqa-sha-yki ‘I will hit you’).

4.6.3.3 ‘conditionals (cond)’ This irrealis mood category is translated through the auxiliaries would and could in English, as in Example (121). They express situations considered less likely than those marked with future tense. (121) Puklla:man paywan. {puklla-:-man pay-wan play-1.sbj-cond s/he-com ‘I would/could play with her.’ While its formation follows a regular pattern, this inflectional category comes from the regularization of a periphrastic pattern. Arguably, the illative case -man became the conditional marker. As such, it follows the normal person markers (1st person -:, 2nd person -nki, 1st person inclusive -ntsik, and 3rd person -n). See Table 4.12 for the full conditional paradigm. In my corpus the 1st person inclusive is expressed only through the fusional form -shwan (e.g. puwidi-shwan ‘we would be able to,’ from puwidi- ‘to be able to’). The perfective aspect is marked in the conditional by means of the invariably auxiliary karqan (i.e. ka- ‘to be’ in the 3rd person singular). Thus, forms like miku-:-man ka-rqa-n ‘I would have eaten,’ miku-nki-man ka-rqa-n ‘you would have eaten,’ and miku-n-man ka-rqa-n ‘s/he would have eaten’ are used (see Table 4.13 for the full paradigm). Regarding subject–object person markers, they take the place of the subject person markers (e.g. parla- ‘to talk’ produces parla-ma-n-man, ‘s/he would talk to me,’ parla-yaq-man ‘I would talk to you (pl),’ ‘we would talk to you,’ ‘we would talk to you (pl),’ and parla-shu-nki-man ‘s/he would talk to you’).

4.6.3.4 ‘imperatives (imp)’ Imperative mood is marked for the 2nd person with the suffix -y, the 1st person plural inclusive with the suffix -shun (which is the corresponding future form), and for the 3rd person (injunctive) with the suffix -tsun. Plural forms are created with -ya:, which is placed before the imperative suffix. The full paradigm is available in Table 4.12. Examples of imperatives are in (55), (77), and (97). HQ has a prohibitive construction combining the imperative suffixes with the particle ama ‘do not,’ which requires the use of the negative suffix -tsu. It is illustrated in (122). (122) Ama parlayaytsu! {ama parla-ya-y-tsu} proh speak-plv-imp-neg ‘Don’t you (pl) speak!’

4.7 Independent suffixes (enclitics) As in all Quechuan languages, HQ uses independent suffixes (or “enclitics” in traditional grammatical description). They can attach to nouns, verbs, or particles. Enclitics come after nominal or verbal inflectional suffixes. I follow Parker’s (1976: 145–9) three-way organization of independent suffixes based on their distribution. The first group includes those that connect the information in the sentence where they are placed to information in another sentence or information that is considered implicit. These include the continuative -raq (4.7.1.1), the discontinuative -na (4.7.1.2), the additive -pis (4.7.1.3), and the contrastive -taq (4.7.1.4). The second group features suffixes that provide a negative or interrogative focus to any part of the sentence. These include the negative -tsu and the interrogative -ku. The final group includes those suffixes conveying the stance of a speaker regarding the statement produced, and whether that information is old or new. This group includes the topic -qa and the evidentials -mi ~ -m, -shi ~-sh, -chi, -cha:, and -ri. Enclitics can be used only once per sentence, although some exceptions can be found.

4.7.1 Connectors of sentence information 4.7.1.1 -raq ‘continuative (cont)’ -raq conveys the notion of continuity. As such, it is commonly translated as ‘yet’ or ‘still,’ as in (123). (123) Punuykanraq. {punu-yka:-n-raq} sleep-ipfv-3.sbj-cont ‘She is still sleeping.’

125

carlos molina-vital However, this suffix is best understood in relation with some previous condition or assumption in spite of which something (an event or a nominal) is brought into the current focus of attention. Punuykanraq implies that sleeping time should be over, and manaraq that something should be the case already, but is not yet. This type of contrastive notion associated with -raq yields the idea of priority or even recentness, as illustrated in (124). (124) Na:nitaraq ruramuy. {na:ni-ta-raq rura-mu-y} road-acc-cont make-cisl-imp ‘Make the road first.’ In this example, the scope of the continuative is not on the action rura- ‘to make,’ but on the object nominal. What the speaker has in mind, here na: ‘road,’ takes precedence over anything else, thus the meaning intended is that of ‘first the road.’ Likewise, tsay-raq (from tsay ‘that’) is best understood as an adverbial modifier meaning ‘recently, just,’ or even ‘at last.’ In Section 4.5.1.1.4, we saw that -raq combines with interrogative pronouns. A form like ima-paq-raq ‘what for?’ (from ima ‘what’ and dative -paq) implies that some previous condition should not be expected to hold at the moment of the question, hence the uncertainty or rhetorical quality with which it has been described (Parker 1976: 146; Julca Guerrero 2009: 260).

4.7.1.2 -na ‘discontinuative (disc)’ This enclitic can be considered as having the opposite meaning to -raq. It conveys the idea of discontinuity: an assumption of continuity is contradicted by the term marked with -na. In other words, whatever is marked with -na stands in sharp contrast with information previously assumed or mentioned. Common translations are ‘already, anymore, now’ (e.g. paqas-na ‘already night,’ from paqas ‘night,’ mana-na ‘no more,’ from mana ‘no’). Another example is in (125). (125) Sha:rikuyna. {sha:ri-ku-y-na} get.up-mid-imp-disc ‘Get up now!’ -na is also used frequently with nominals. Arguably, in those cases the referent of the noun marked with -na is a new instance in a sequence and contrasts with whatever or whoever took place before. This seems to be the case in tsay-na ‘therefore’ (from tsay ‘that’). In (126), different animals are

126

making noises together in order to scare some people. The first instance, ashnu ‘donkey,’ is unmarked. Then, every animal making a noise is marked with -na. (126) Ashnu llapan kallpanwan ha:chinyarqan. Allquna anirqan… Mishina ñawyarqan. Galluna kantarqan. {ashnu llapa-n kallpa-n-wan donkey all-3.poss force-3.poss-com ha:chinya-rqa-n allqu-na ani-rqa-n bray-pst-3.sbj dog-disc bark-pst-3.sbj mishi-na ñawya-rqa-n gallu-na cat-disc meow-pst-3.sbj rooster-disc kanta-rqa-n} sing-pst-3.sbj ‘The donkey brayed with all his strength. And the dog barked… And the cat meowed. And the rooster sang.’

4.7.1.3 -pis ‘additive (add)’ The function of this suffix is to indicate that a term is also the case or also present alongside other information in the sentence (e.g. ruripa-n-pis ‘also her skirt,’ from ruripa ‘skirt,’ kanan-pis ‘also now,’ from kanan ‘now,’ qatsawa-n-pis ‘he also dances,’ from qatswa- ‘to dance’ in the 3rd person -n). However, this notion of addition is not restricted to objectively adding a new referent or notion. Instead, it can be used with the sense of something that gets added despite of a previous notion. Thus, the previous examples could also be translated as ‘although’ or ‘even’ (Example 91 clearly shows this function of -pis). For the use of -pis with interrogative pronouns to create indefinite pronouns, see Section 4.5.1.1.4.

4.7.1.4 -taq ‘contrastive (contr)’ This suffix conveys the idea that some information in the sentence contradicts previous information. In other words, -taq adds new information whose function is to be in contrast with some old information. -taq never appears with the topic marker -qa in the same word, since the latter links a term with current relevance to co-referential old information. The contrastive use of -taq is clearly illustrated in (115) above. In the narration, a fox says Alla:pa-taq kayqa nanaykun, ¡ananáw! ‘This hurts a lot, ouch!’ Alla:pa ‘a lot’ is marked with -taq because the fox was not expecting the process to be painful. The contrastive enclitic is frequently associated with interrogative pronouns as a way to link the question to previous information. -taq attaches to the head of the interrogative phrase, as in (127) and (128).

huaylas (ancash) quechua (127) Imataq tsay? {ima-taq tsay} what-contr dem ‘What is that?’ (128) Ima ninataq tsayqa? {ima ni-na-taq tsay-qa} what say-pot.nmlz-contr dem-top ‘What does that mean?’ The accusative suffix -ta can be omitted when -taq is used (Ima-ta-taq/Ima-taq rura-nki? ‘What do you do?’ with rura- ‘to do, make’ marked with 2nd person -nki). Finally, there is alternation between -taq and -tan. The latter is preferred in the Northern Huaylas variety (Julca Guerrero 2009: 260). Parker (1976: 147) argues that they are not exactly interchangeable because -taq has more contexts of use. I have not observed enough instances of -tan to determine what could differentiate it from -taq.

4.7.2 Negative and interrogative 4.7.2.1 -tsu ‘negative (neg)’ This suffix makes any word the focus of a negation (e.g. alli-tsu ‘not good,’ from alli ‘good,’ alla:pa-raq-tsu ‘not much yet,’ from alla:pa ‘much’ and continuative -raq). The negated term with -tsu can be introduced by a negator like mana ‘no’ or the prohibitive ama ‘do not,’ but only the second is mandatory. -tsu can combine in the same word with enclitics in the first group, with the exception of -taq. However, it seems that it never combines in the same word with enclitics in the third group (topic and evidentials), but more analysis is required. See Section 4.8.3.2 for examples and information on the use of -tsu in negative sentences.

4.7.2.2 -ku ‘interrogative (q)’ -ku is added to any word to make it the focus of a question whose answer can only be awmi ‘yes’ or mana ‘no.’ For instance, the question in (129) is asking whether that is your destination or not by marking the goal with -ku. (129) Tsaymanku aywanki? {tsay-man-ku aywa-nki} dem-all-q go-2.sbj ‘Do you go there?’ On the other hand, (130) focuses on whether you live or not at a place (ta:- ‘to live’ with middle -ku in the 2nd person -nki and tsay ‘that’ with locative -chaw).

(130) Tsaychaw ta:kunkiku? {tsay-chaw ta:-ku-nki-ku?} dem-loc live-mid-2.sbj-q ‘Do you live there?’ When the negator mana ‘no’ takes -ku, the resulting expression is a negative question, as exemplified in (131). (131) Manaku munanki? {mana-ku muna-nki} neg-q want-2.sbj ‘Don’t you want [it]?’26

4.7.3 Modality: topic and evidentials 4.7.3.1 -qa ‘topic (top)’ The main function of the topic marker -qa is to keep track of whatever has been mentioned before in the discourse, and that is brought back to the attention of the addressee. Thus, the function of marking old, known, or active information suits most uses of -qa. In Example (132) -qa marks taklla:qa ‘my foot plow,’ which refers to the previously mentioned term takllallata ‘a foot plow’ (old information). (132) ¿Manaku takllallata rikarqunki? ¡takllalla:qa manataq kantsu! {mana-ku taklla-lla-ta no-q foot.plow-lim-acc rika-rqu-nki taklla-lla-:-qa see-rec.pst-2.sbj foot.plow-1.poss-top mana-taq ka-n-tsu} no-contr be-3.sbj-neg ‘Haven’t you seen a foot plow? My foot plow is not there!’ Although the topic marker can appear with any word class, nominals are mostly marked. Very few instances are inflected verbs (unless they are nominalized). Subordinate constructions with -pti or -r (see Sections 4.8.4.2.3.1 and 4.8.4.2.3.3) and particles like itsan ‘instead’ or kanan ‘now’ are also frequently combined with -qa, and they take more specific meanings, arguably contrastive (itsan-qa ‘however’ and kanan-qa ‘right now’). In my corpus -qa is mostly used once per sentence (considered in terms of the presence of a main inflected verb). However, some exceptions exist. In (133), each instance marked with -qa stands in clear contrast to the way in which the person addressed by the speaker was cultivating potatoes: 26 The negator -ku has no relation with middle voice -ku. They are homonyms.

127

carlos molina-vital (133) Kananpitaqa kaynaw pachankunachawqa uraymanpam murunki… {kanan-pita-qa kay-naw now-abl-top dem.prox-simul pachan-kuna-chaw-qa slope-pl-loc-top uray-man-pa-m down-all-gen-evd

muru-nki} sow-2.sbj

‘From this moment on, in this way you will sow [potatoes] downwards on the slopes…’ Perhaps this contrastive function of -qa is related to bringing the attention of the hearer to something the subject considers important to elaborate upon. This is essentially the function of marking some information as the topic about which some comments are offered. In any case, it should be noted that -qa does not combine with -raq, -na, or -taq in the same word.

4.7.3.2 -mi ‘evidential–direct experience (evd)’ While -qa provides the topic or currently considered piece of information in a text, evidentials convey the information the speaker considers most relevant in order to comment upon the provided topic. In HQ, as in all Quechuan languages, the comment information is marked based on levels of certitude the speaker assigns to that information. The first evidential is -mi (-m after a short vowel). It indicates that the information offered comes from the speaker’s direct experience. By using it, the speaker presents themselves as a worthy source, a witness, or confident enough to vouch for the information. Frequently, it marks the information that answers a question, as in (134). (134) Pitaq tsay shipash —Payqa churi:mi. {pi-taq tsay shipash who-contr dem young.woman

pay-qa she-top

churi-:-mi} child-1.poss-evd ‘Who is that young woman’—‘She is my daughter.’ As with any enclitic, -mi is placed in any word the speaker marks as the comment. Thus, a sentence like Leonel tsayman away-n ‘Leonel goes there’ could place the direct evidential on each word. Leonel-mi would stress that it is LEONEL who goes there; tsay-man-mi that it is THERE that Leonel goes. If placed on the verb, the whole sentence is considered the scope of the evidential (Julca Guerrero 2009: 252). Other examples with -mi are given in (1) and (120).

128

4.7.3.3 -shi ‘evidential–reportative/quotative (evr)’ -shi (-sh after a short vowel) indicates that the information provided comes not from direct experience of the speaker, but from someone else’s testimony or report, hence the term “reportative” or “quotative evidential.” This source is a 3rd person who does not directly participate in the speech act. Like -mi, it is possible to place -shi on any word in a sentence, but only once. This evidential marker is closely associated with the meaning of -naq, the non-experienced past tense. In few cases, they are used together to emphasize the nonexperienced nature of what is narrated, as (135) illustrates. (135) Achikaypapis kapunaqshi huk warmi wamran. {achikay-pa-pis ka-pu-naq-shi huk achique-gen-add be-ben-narr.pst-evr one warmi wamra-n} female child-3.poss ‘They say that Achiqué had a daughter.’ Some degree of uncertainty is associated with -shi. Therefore, it can be used with interrogative pronouns to suggest perplexity from the person asking the question (e.g. in (99) imanir-taq-shi ‘why could it be?’) Examples of -shi in narrations (i.e. non-witnessed events) are found in (94) and (99). Uses of the indirect evidential -shi with tsay functioning as a connective of consequence (‘therefore, they say’) can be found in (29) and (53).

4.7.3.4 -chi ‘evidential–conjectural (evc)’ -chi conveys an attitude the speaker has toward the source of the information being presented. In this case, the information comes from a conjecture. This means that there is some degree of certainty associated with what is said, although not a complete one. Example (112) provided a use of -chi in upu-paku-ri-shaq-chi ‘it seems like I am going to drink a lot’ to express something that the speaker believed possible. Example (136) offers a case of likely conjecture: (136) Tsuri:taqa alla:pachi ankushaq. {tsuri-:-ta-qa alla:pa-chi child-1.poss-acc-top much-evc

anku-shaq} miss-1.sbj.fut

‘I will miss my daughter surely a lot.’ When used with an interrogative pronoun, -chi indicates that there is no complete certainty about what is being asked. The person asking is unsure about what they ask, or think that there is no answer to the question (ima-chi? ‘what could it be?’). Finally, the conjectural evidential -chi can be followed by -ya:, an emphatic suffix increasing the certainty of the information. This use is barely present in my corpus, but

huaylas (ancash) quechua Julca Guerrero (2009, 2010) registers its use in the Huaraz province; see (137).

4.8 Basic syntax

(137) Allquykichiya: wamralla:ta kanishqa. allqu-yki-chi-ya: wamra-lla:-ta dog-2.poss-evc-emph child-dim-acc

4.8.1 Grammatical relations

kani-shqa. bit-3.sbj.prf ‘It is quite likely that your dog has bitten my child.’ (adapted from Julca Guerrero 2009: 256)

4.7.3.5 -ri ‘corroborative (corr)’ This enclitic is not an evidential like -mi, -shi, or -chi, since it does not relate to the quality or validity of the information. Instead, it conveys the idea that the speaker is in agreement with some information the interlocutor has presented, as illustrated in (138). (138) —Aywankiku? —Aywashaqri! {aywa-nki-ku aywa-shaq-ri} go-2.sbj-q go-1sbj.fut-corr ‘Are you going?’—‘I will go, of course!’ As in (139), -ri can also imply that the addressee should agree with what the speaker thinks. (139) Amari pinqakuytsu, qichwa shimintsikta parlakushun. {ama-ri pinqa-ku-y-tsu qichwa proh-corr shame-mid-imp-neg quechua shimi-ntsik-ta language-1.incl.poss-acc parla-ku-shun} speak-mid-1.pl.incl.sbj.fut ‘Do not feel embarrassed, let’s speak our Quechua language.’ (Implicit: There is nothing to be embarrassed about in speaking Quechua.) In a way, -ri also indicates that something is the current focus of attention. As such, it doesn’t seem to combine with -qa, but it combines with the evidentials (e.g. tsay-mi-ri ‘therefore, of course’ from tsay ‘that’ or wawa-yki-sh-ri ‘they say that your child, of course,’ from wawa-yki ‘your child’).

4.7.3.6 -tsuraq ‘dubitative (dub)’ The combination of the interrogative -tsu and the continuative -raq can be considered a fused element -tsuraq. This suffix is used for yes/no questions. However, what the question implies is considered unlikely (e.g. ka-n-tsuraq? ‘could there be?,’ from ka-n ‘there is,’ or qam-tsuraq? ‘perhaps it is you?,’ from qam ‘you’).

Like every Quechuan language, HQ can be considered a typical instance of nominative–accusative alignment. The A (agent-like subject) argument of a transitive verb and the S (single argument) of an intransitive verb are marked equally in the zero-marked nominative case. The P (patient-like) argument receives the accusative case suffix -ta. Examples (140) illustrates the pattern with an intransitive sentence, while (141) presents a transitive one. (140) Malli puriykan. {malli-ø puri-yka:-n} María-nom walk-ipfv-3.sbj ‘María is walking.’ (141) Malli hakata aruykan. {malli haka-ta María guinea.pig-acc

aru-yka:-n} cook-ipfv-3.sbj

‘María is cooking a guinea pig.’

4.8.2 Constituent order HQ places the modifier before the head. This is clearly illustrated in Section 4.5.1.4 through the order of constituents in the nominal phrase, where the nominal head is preceded by different types of modifiers (e.g. demonstratives, numerals, property terms, or adjectives). This ordering takes place also with the genitive possessor preceding the possessed nominal, which is marked with a nominal possessive suffix (see Sections 4.5.3.1 and 4.5.3.3.2). Relative clauses (formed with nominalizers like -q, -shqa, -nqa, and -na) also precede the modified head (however, appositional modifiers are always found after the head: see Section 4.8.4.2.1.2). Likewise, adverbs (as well as so-called pre-adjectives) precede their head (see Section 4.5.1). The unmarked constituent order is SOV. Most examples including accusative-marked arguments illustrate this (see (1) and (12), to mention only the first ones with SOV order). However, it is not uncommon for the object to appear after the verb. This can be seen in (29), (65), and (90). No instances of OSV have been found in my corpus. As seems to be the norm for Quechuan languages, subordinate clauses (nominalized or using same-subject -r or different subject -pti) are much stricter regarding OV order. Examples (7) and (91) illustrate this.

129

carlos molina-vital

4.8.3 Major clause types

mana ‘no’ can be seen in (146) with a nominalized purpose subordinate (on which see Section 4.8.4.2.3.2):

4.8.3.1 Assertions Regarding assertions, also known as declarative clauses, Sections 4.7.1 and 4.7.3 offer several examples of how they are expressed through enclitics like the continuative -raq, the discontinuative -na, the contrastive -taq, the topic -qa, evidentials, and emphatic markers. While those enclitics are by no means mandatory to express an assertion, they are frequently associated with maintaining the flow of information. Also, HQ speakers, as is common in other Quechuan languages, tend to perceive the lack of enclitics as something that weakens the clarity of what is being said.

4.8.3.2 Negations and prohibitions Negative clauses are formed by using the enclitic -tsu (see Section 4.7.2.1). This suffix can mark any term in a sentence with a negative focus; however, if -tsu is on an inflected verb, then the scope of the negation will be the whole sentence. Examples (142)–(145) illustrate the different words or phrases to which -tsu can attach. (142) Tsayshi mana allitsu kanqa. {tsay-shi mana alli-tsu dem-evr no good-neg

ka-nqa} be-3.sbj.fut

‘Therefore, they say, it will be no good.’ (143) Kikinkuna pura mana akwirdutsu kayarqan. {kiki-n-kuna pura mana akwirdu-tsu self-3.poss-pl excl no agreement-neg ka-ya-rqa-n} be-plv-pst-3.sbj ‘They were in no agreement among themselves.’ (144) Mana allita rurarqantsu. {mana alli-ta rura-rqa-n-tsu} No good-acc make-pst-3.sbj-neg ‘She didn’t do it well.’ (145) Rukusqa kakashta mana willashqatsu. {rukus-qa kakash-ta mana sparrow.hawk-top rooster-acc no willa-shqa-tsu} tell-3.sbj.prf-neg ‘The sparrowhawk has not told it to the rooster.’ -tsu is used only within a sentence (i.e. a clause with an inflected verb). Mana ‘no’ becomes the mandatory negative operator outside of the sentence level (e.g. alli ‘good’ > mana alli ‘bad, no good’). The absence of -tsu and the need for

130

(146) Mana atuq rika:ma:nanpaq punchuykillawan tsapaykamay. {mana atuq rika:-ma:-na-n-paq no fox see-1.obj-pot.nmlz-3.poss-dat punchu-yki-lla-wan poncho-2.poss-lim-com

tsapa-yku-ma:-y} cover-excep-1.obj-imp

‘Cover me with your poncho please, so that the fox doesn’t see me.’ Despite the frequent use of mana in agreement with -tsu, it is common to find sentences in which only -tsu is used (e.g. musya-:-tsu ‘I don’t know,’ from musya- ‘to know’) Optionality in the use of mana can be considered as part of a diachronic cycle (Pineda-Bernuy 2014) in which languages with one negative marker (e.g. mana ‘no’) develop a negative agreement (e.g. mana + -tsu), and then reduce to a single marker (only -tsu). According to Villari’s (2017: 70) study of negation in the Huaraz province, a plausible synchronic motivation for the absence of mana ‘no’ before -tsu could be the presence of that negator or ni ‘nor’ (see Section 4.5.1.1.4) in the previous co-text. HQ uses the negative particle ama ‘do not’ as a prohibitive. This is observed with all the imperative forms, which must be negated with -tsu. See Section 4.6.3.4 and (49), (122), and (139) for the use of ama.

4.8.3.3 Questions Interrogative sentences have been presented in Section 4.5.1.1.4 on interrogative pronouns and in Section 4.7.2.2 on the interrogative enclitic -ku for yes/no questions. HQ, like several other Quechuan languages, has an interrogative verb, imana- ‘to happen; to do.’ It is used to ask about an indefinite action. As such, it can be used as an intransitive verb meaning ‘what’s going on?’, as in (147). (147) Imanaykan? {imana-yka-n} happen-cont-3.sbj ‘What’s going on?’ However, it also works with an overt or implied direct object, as in (148) and (149) (148) Nuqa imana:taq payta? {nuqa imana-:-taq pay-ta} I happen-1.sbj-contr s/he-acc ‘What do I do to her?’

huaylas (ancash) quechua (149) Imanankitaq? {imana-nki-taq} do-2.sbj.fut-contr ‘What will you do?’

mana allin-ta rika:-yka:-naq} no good-acc see-ipfv-narr.pst ‘He fell into an awfully deep ditch; he wasn’t seeing well.’

A common expression of resignation in the face of adversity, in (151), uses imana in the 1st person inclusive

Coordinated clauses in HQ that explicitly connect different information are traditionally called “conjunctive coordinate sentences.” They are obtained through enclitics like the continuative -raq, the discontinuative -na, the additive -pis, and the contrastive -taq. As seen in Section 4.7.1, they connect the information they mark to other implicit or overt information in a previous sentence. A true coordinating function for those enclitics was exemplified for -na in Example (126). A similar type of distribution can be obtained through the use of -raq. This is best used with the same subject performing different actions simultaneously: qotsu-:-raq, wichya-:-raq, tushu-:-raq ‘I sing, whistle, and dance at the same time,’ from qotsu- ‘to sing,’ wichya- ‘to whistle,’ and tushu- ‘to dance,’ all in the 1st person subject (Julca Guerrero 2009: 317). In (153) and (154), I offer two examples with -taq, the most frequently used coordinating enclitic.

(151) Imanashun? {imana-shun} do-1pl.incl.sbj.fut

(153) Añañáw! Achalláw! Kayqa wallqataq! Kayqa rusaryutaq! {añañáw achalláw kay-qa sweet nice dem.prox-top

This question does not ask about a concrete thing that the addressee will do. Instead, it refers to a whole action as a response to some issue. This is analogous to the role of the direct object as stimulus (see Section 4.6.1.2), which makes imana- a type of semantically intransitive verb (see Section 4.6.1.3). Subject–object agreement marking is also possible with imana-; see (150). (150) Imanashunkitaq? {imana-shu-nki-taq} happen-3.sbj-2.obj-contr ‘What’s the matter with you?,’ ‘Does it affect you?’

‘What could be done?’ (lit. ‘What will we do? What will happen to us?’)

wallqa-taq necklace-contr

kay-qa dem.prox-top

rusaryu-taq} rosary-contr

‘Sweet! Nice! This is a necklace! And this a rosary!’

4.8.4 Complex clauses Complex clauses include more than one verbal head. They can be formed through juxtaposition (the direct combination of two or more sentences), coordination (the assembly of clauses through a connecting suffix or particle), and subordination (the use of a verbal clause to fulfill the grammatical functions of modifiers or arguments within a sentence).

4.8.4.1 Juxtaposition and coordination Juxtaposition is the process of connecting two or more sentences without using any overt grammatical marker for this. Clauses are presented sequentially based on their semantic compatibility. In Example (152), two sentences with inflected verbs hiqakurqunaq ‘he fell’ and rikayka:naq ‘he saw’ are juxtaposed, with the second sentence being the cause of the first: (152) Hiqakurqunaq huk fiyu sanhaman, mana allinta rikayka:naq. {hiqa-ku-rqu-naq huk fiyu sanha-man fall-mid-rec.pst-narr.pst one ugly ditch-all

(154) Piru tsaychawtaq nuqaqa ka:. Apamurqa:taq mikuyniyki. {piru tsay-chaw-taq nuqa-qa but dem-loc-contr I-top ka-: apa-mu-rqa-:-taq be-1.sbj carry-cisl-pst-1.sbj-contr miku-y-ni-yki} eat-inf-ep-2.poss ‘But I was there, and I brought you your food.’ In HQ the use of coordinating conjunctions from Spanish is widespread. Thus, i (from y ‘and’), piru/pero (from pero ‘but’), u/o (from o ‘or’), and purki/porke (from porque ‘because’), among other are found throughout the corpus, often combined with the original HQ connecting enclitics. An example of disjunctive coordination using o is in (155). (155) Itsa kawan o wanushqa? {itsa kawa-n o Perhaps live-3.sbj or

wanu-shqa} die-3.sbj.prf

‘Maybe they live or maybe they have died?’ Disjunctive expressions in HQ are viewed as closely related to adversative ones. Itsa is a connective that alternates with the Spanish loan pero as in (156):

131

carlos molina-vital (156) Qamqa allilla kaykay; itsa nuqaqa hutsaykiraykurmi nakamushaqraq. {qam-qa alli-lla ka-yka:-y itsa nuqa-qa you-top good-lim be-ipfv-imp but I-top hutsa-yki-raykur-mi naka-mu-shaq-raq} sin-2.poss-cisl-evd suffer-cisl-1sbj.fut-cont ‘You stay calm; but I will suffer longer due to your sins.’

Subordination is obtained, as in most Quechuan languages, through the use of nominalized forms. They create relative, complement, or adverbial clauses. Still, the main function of nominalizers is to turn verbal stems into nominal referents, be they lexical (i.e. an entity) or grammatical (i.e. an argument or an event). The infinitive -y, the agentive -q, the resultative nominalizer -shqa, the potential nominalizer -na, and the less frequently used actual nominalizer (realis) -nqa are the HQ nominalizers (see Section 4.4.2.1). Although they share formal properties with nominalizers, the suffixes -pti, -r, and -shpa, discussed in Section 4.8.4.2.3.1, are not considered part of that category, but as “subordinating inflections” introducing adverbial clauses.

4.8.4.2.1 Relative clauses The first type of subordinate clauses are relative or adjective clauses. They can use any nominalizer to create a clause that specifies the referent of the noun they modify, or that can be used as an apposition to explain or elaborate upon that referent. There are no restrictions regarding the grammatical function of the relativized constituent. This means that the co-referent in the relative clause can be a subject, object, or an oblique.

4.8.4.2.1.1 specifying or restrictive relative clauses Example (157) illustrates how nominalizers can be used to restrict the referent of the nominal head they modify. First, I will show that the infinitive can be used to create a nominal that can modify another one. These are not usually presented as relative clauses, but they are not formally different from those created with other nominalizers: kanan} now

‘The day of your departure from the house is today.’

132

(158) … kakashyuq nunaqa alla:pa qishyaq nunata rika:rinaq … {kakash-yuq nuna-qa alla:pa qishya-q rooster-inal.poss man-top very sick-ag nuna-ta rika:-ri-naq} man-acc see-punc-narr.pst ‘… the man with a rooster saw a very sick man …’

4.8.4.2 Subordination

(157) Wayipita yarquyniyki hunaqmi kanan. {wayi-pita yarqu-y-ni-yki hunaq-mi house-abl exit-inf-ep-2.poss day-evd

Examples (158)–(161) illustrate relative clauses with the remaining nominalizers. It should be noticed that -na is only seldom used with this function27 :

(159) Murushqan kinwapaq manana yakuta apayamunkitsu. {muru-shqa-n kinwa-paq mana-na yaku-ta sow-res-3.poss quinoa-dat no-disc water-acc apa-ya:-mu-nki-tsu} carry-plv-cisl-2.sbj-neg ‘You won’t bring water anymore for the quinoa that he sowed.’ (160) Alla:pam mikuyanayki qurata muruyan. {alla:pa-m miku-ya:-na-yki many-evd eat-plv-pot.nmlz-2.poss qura-ta muru-ya:-n. plant-acc sow-plv-3.sbj ‘They sow a lot of the plants that you eat.’ (161) Yurikunqa: wayita tayta: rantikushqana. {yuri-ku-nqa-: wayi-ta be.born-mid-act.nmlz-1.poss house-acc tayta-: father-1.poss

ranti-ku-shqa-na} trade-mid-3.sbj.prf-disc

‘My father has already sold the house I was born [in].’ In (158), (159), and (160) nominal agreement indicates the subject argument of the relative clause. If the subject has to be overt, it will be marked with genitive -pa, as in (162). (162) qampa mikushqayki papa {qam-pa miku-shqa-yki you-gen eat-res-2.poss

papa} potato

‘the potato you ate’ However, with the agentive nominalizer -q, nominal agreement indicates the object argument, as in (163). (163) kuyaqniyki {kuya-q-ni-yki} love-ag-ep-2.poss ‘the one that loves you’ 27 Weber (1983c: 27–31) discusses the fact that Huallaga Quechua differs from Ancash Quechua in that the former cannot use nominalizations with -na as adjective clauses.

huaylas (ancash) quechua Other examples of nominalizers as relative clauses are found with -q in (118), and with -shqa in (64) and (114). Finally, internally headed relative clauses are frequent in HQ with all the nominalizers, except the infinitive -y. Example (164) shows a nominalized sentence with no overt modified nominal present. It is obvious, however, that this nominal must be wamra ‘child,’ the subject of the nominalized clause with -q. (164) Hurqurkaraya:mun nunakunata, warmikunata, wamra aywakuykaqta. {hurqu-rku-ri-ya:-mu-n nuna-kuna-ta take.out-uw-punc-plv-cisl-3.sbj man-pl-acc warmi-kuna-ta wamra aywa-ku-yka:-q-ta woman-pl-acc child go-mid-ipfv-ag-acc ‘They took out men, women, and the children that were going [to that school].’

4.8.4.2.1.2 appositional relative clauses The use of nominalized clauses after a nominal head to provide extra information about that head is common in HQ. Those appositional relative clauses should be considered as autonomous nominals sharing the same referent as the previously mentioned nominal they describe. Appositions must share the same case marker as their co-referential head. Example (165) shows the use of an apposition with -q introduced through a perception event: (165) Manam rikarqu:tsu ni pita kaypa pasaqta. {Mana-m rika-rqu-:-tsu ni pi-ta no-evd see-rec.pst-1.sbj-neg nor who-acc kay-pa dem.prox-gen

pasa-q-ta} pass-ag-acc

‘I didn’t see anyone who passes through here.’

Example (167) illustrates how this type of appositional construction (also known as “case-floating”) can occur even without a nominalization involved. (167) Pitsqa rapranta kukata tsarikun. {pitsqa rapra-n-ta kuka-ta tsari-ku-n} five leaf-3.poss-acc coca-acc grab-mid-3.sbj ‘He takes five leaves of coca.’ (lit. ‘He takes five leaves, the coca ones.’) Other probable instances of appositional modifiers were given in (41) and (42) with existential ka-n and the benefactive suffix -pu expressing possession (see Section 4.6.1.1). Also, (114) could be a case of apposition with -shqa further elaborating what the co-referential noun looks like.

4.8.4.2.2 Nominal clauses (embedding) In HQ, embedded nominal clauses are possible with any nominalizers described in Section 4.4.2.1. It is possible to find cases of embedding in subject position with verbs like gusta- ‘to like,’ as in (168). (168) Parlanqayki gustan. {parla-nqa-yki gusta-n} talk-act.nmlz-2.poss like-3.sbj ‘S/he likes that you talked.’ However, this is uncommon with other verbs. Thus, I will focus here on direct object embedding.

4.8.4.2.2.1 complement clauses with infinitive -y This common type of complement clause requires that the subject of the main (inflected) verb is also the A (agent-like) argument of the embedded infinitive verb. It is illustrated in (169).

Although perception events are commonly associated with this type of relative clause, other types of verbs can use them too. In Example (166), tishqa- ‘to dock something on an island’ (more commonly, tishqu-) takes two co-referential direct objects marked with -ta.

(169) Yarquyta munaya:. {yarqu-y-ta muna-ya-:} exit-inf-acc want-plv-1.sbj

(166) Tsaynash yakuqa wamrakunata aparaykashqanta, tishqaykurqunaq. {tsay-na-sh yaku-qa wamra-kuna-ta dem-disc-evr water-top child-pl-acc

Verb classes that commonly take an infinitive embedded object are cognition (e.g. gusta- ‘to like,’ yacha- ‘to know; to learn,’ musya:- ‘to think,’ yarpa- ‘to remember’), deontic (puwidi- ‘to be able to,’ a Spanish loan that has replaced ati- completely in HQ), and in very few cases, the aspectual auxiliary qalla- ‘to begin’ (see Section 4.8.4.2.3.1).

apa-ra:-yka:-shqa-n-ta take-dur-ipfv-res-3.poss-acc

‘We want to go out.’

tishqa-yku-rqu-naq} dock.on.an.island-excep-rec.pst-narr.pst

4.8.4.2.2.2 embedding with other nominalizers (-shqa, -na, and -nqa)

‘Then the water beached the children, which were being taken [by the water].’

When the main verb and the embedded clause have different subjects, the nominalizers -shqa, -na, and -nqa are

133

carlos molina-vital used. The agentive nominalizer -q could only appear in object position in a relative clause. Example (170) shows how -shqa profiles the affected object, Example (171) how -na profiles an event as a future occurrence, and Example (172) how -nqa profiles the fact that something took place: (170) Qutsushqanta yarpankiku? {qutsu-shqa-n-ta yarpa-nki-ku} sing-res-3.poss-acc remember-2.sbj-q ‘Do you remember what she sang?’ (171) Yarqunaykita munaya:. {yarqu-na-yki-ta muna-ya-:} exit-pot.nmlz-2.poss-acc want-plv-1.sbj ‘We want you to go out.’ (172) Mikuyanqa:ta yachan. {miku-ya:-nqa-:-ta yacha-n} eat-plv-act.nmlz-1.poss-acc know-3.sbj ‘She knows that we ate.’ Against Parker’s (1976: 164) observations, no limitations seem to exist regarding verbs that can take an embedded clause. While this type of complex construction is not frequent in oral styles, they are found more often in written text with a wider variety of verbs; as exemplified in (173).

(173) Cooperacion Popular kalaminata a:niya:mashqanta mana apuradu kumpliptin … {cooperacion popular kalamina-ta Cooperation public tin.roof-acc a:ni-ya:-ma-shqa-n-ta agree-plv-1.obj-res-3.poss-acc mana no

apuradu hurried

kumpli-pti-n} fulfill-subds-3.sbj

‘When Cooperación Popular (a government agency) didn’t hurry up to fulfill their agreement with us to a tin roof . . .’ (lit. ‘When Cooperación Popular didn’t hurry up to fulfil that they had said yes to us to a tin roof ’) Finally, I include among nominalized embedded clauses the basic obligation construction expressing ‘to have to.’ It uses the potential nominalizer -na to indicate the probable nature of an event. In addition to this, in the present tense it uses a zero 3rd person singular form of ka- ‘to be.’ For other tenses, a 3rd person singular form of ka- is overtly used. Examples of this construction were offered in (91), (94), (95), and (146). See Table 4.15 for the obligation construction paradigm in the three basic tenses.

4.8.4.2.3 Adverbial clauses In order to provide information surrounding an event’s occurrence such as cause, consequence, location, and duration,

Table 4.15 Obligation construction with potential nominalizer -na and auxiliary ka- in Huaylas Quechua Number

Person

Present Auxiliary: Ø

Past Auxiliary: ka-rqa-n

Future Auxiliary: ka-nqa

Singular

1

rima-na-: ‘I have to speak.’

rima-na-: karqan ‘I had to speak.’

rima-na-: kanqa ‘I will have to speak.’

2

rima-na-yki ‘You have to speak.’

rima-na-yki karqan ‘You had to speak.’

rima-na-yki kanqa ‘You will have to speak.’

3

rima-na-n ‘S/he has to speak.’

rima-na-n karqan ‘S/he had to speak.’

rima-na-n kanqa ‘S/he will have to speak.’

1.incl

rima-na-ntsik ‘We have to speak.’

rima-na-ntsik karqan ‘We had to speak.’

rima-na-ntsik kanqa ‘We will have to speak.’

1.excl

rima-ya:-na-na ‘We have to speak.’

rima-ya:-na-: karqan ‘We had to speak.’

rima-ya:-na-: kanqa ‘We will have to speak.’

2

rima-ya:-na-yki ‘You have to speak.’

rima-ya-na-yki karqan ‘You had to speak.’

rima-ya-na-yki kanqa ‘You will have to speak.’

3

rima-ya:-na-n ‘They have to speak.’

rima-ya:-na-n karqan ‘They had to speak.’

rima-ya:-na-n kanqa ‘They will have to speak.’

Plural

134

huaylas (ancash) quechua HQ uses specific subordinating suffixes or nominalized forms (with the exception of the agentive nominalizer -q) marked with case suffixes. In this section, I will present first the subordinating suffixes -r (same subject), -shpa (gerund) (4.8.4.2.3.1), and -pti (different-subject) (4.8.4.2.3.2), before moving on to purpose adverbial construction using nominalizers (4.8.4.2.3.2).

(176) Huklla takllayki katarkur uryaq ayway! {huklla taklla-yki kata-rku-r at.once foot.plow-2.poss carry.on.back-uw-subis

4.8.4.2.3.1 -r and -shpa ‘same-subject subordinator (subis’), -pti ‘different-subject subordinator’ (subds)

While nominal agreement is not required for same-subject reference, object agreement can be coded with -r by using the following combinations: -ma-r-ni-: (1st person object singular), -ma-r-ni-ntsik (1st person inclusive object), -r-ni-yki (2nd person object), -shu-r-ni-yki (3rd person subject and 2nd person object). Their plural forms take -ya: in its default position, before the object marker. Notice that if the subordinate sentence has a 3rd person object, then it can optionally take the form -r-ni-n, which is equivalent to just -r. Example (177) illustrates the use of -r with the 1st person object marker -ma:.

Adverbial subordinate clauses present a situation in a temporal or logical relation depending on the inflected (main) verb. They are translated with when, while, and because. Like most Quechuan languages, HQ uses a switch-reference system: -r and -shpa require that the subject in the main and subordinate clauses are the same; if they are different, -pti is used. The subordinating suffix -r is not marked for nominal agreement, since it has the same subject as the main clause. As illustrated in (174), it marks a subordinate event as taking place prior to the main one, sometimes as a cause to the main clause. (174) … tikakunata hurqur, allibwenu sasata paskaykurqa:. {tika-kuna-ta hurqu-r allibwenu sasa-ta bricks-acc extract-subis very difficult-acc paska-yku-rqa-:} release-excep-pst-1.sbj ‘… when I took out the bricks, I released her with a lot of difficulty.’ (i.e. ‘by taking out the bricks’) As mentioned in Section 4.6.2.1.2.1 in connection with (53), the presence of -ri immediately before -r (as well as -pti) indicates that the subordinate has taken place as a completed event. An example of this is given in (175). (175) Llapanta alistarirnam, qallaykuyan amararna. {llapanta alista-ri-r-na-m all-3.poss-acc prepare-punc-subis-disc-evd qalla-yku-ya:-n amara-r-na} begin-excep-pl-3.sbj3 tie-subis-disc ‘After having prepared everything, they began to tie it up.’ The previous example also illustrates how qalla- ‘to begin’ takes a subordinate clause with -r (amara-r- ‘tying’) as its complement instead of taking an infinitive complement. A very frequent type of coordination in HQ presents the main clause as immediately following the event marked with -r, as in Example (176).

urya-q away-y} work-ag go-imp ‘Hurry up, put your foot plow on your back and go to work!’

(177) Imanirtan mantsamarni: ayqinki? {imanir-tan mantsa-ma:-r-ni-: ayqi-nki} why-contr fear-1.obj-subis-ep-1.sbj flee-2.sbj ‘Why do you run away fearing me?’ The same-subject subordinator -shpa is less frequent than -r, and its use is restricted to the Huaraz province in the Callejón de Huaylas (cf. Parker 1976: 144). In contrast to -r or -pti, it never takes any type of subject or object agreement markers. While less frequent than -r, it is still common in my corpus. The relation between -shpa and the main clause is ambiguous, and it can be interpreted as simultaneous, consecutive, or previous to the main clause. In (178), due to context, only simultaneous and consecutive readings are possible: (178) Tsaynaw nirshi, mana mantsashpa pumapa wayinman hiqarinaq. {tsay-naw ni-r-shi mana mantsa-shpa dem-simul say-subis-evr not fear-subis puma-pa wayi-n-man hiqari-naq} cougar-gen house-3.poss-all leave-narr.pst ‘Having said this, he left for the cougar’s house without being afraid.’ The different-subject subordinator -pti is always marked for nominal person agreement to indicate the subject of the subordinate clause, as in (179). (179) Tsayshi patsa wara:riptin kuntuta rika:rinaq. {tsay-shi patsa wara:-ri-pti-n dem-evr world be.dawn-punc-subds-3.sbj

135

carlos molina-vital kuntu-ta rika:-ri-naq} jug-acc see-punc-narr.pst ‘Then, right after it was dawn, he looked at the jug.’ -pti can also be marked with subject–object agreement. Example (180) uses the person marker -q, which indicates a 1st person subject acting on a 2nd person object. (180) Yachatsiyka:yaptiq allinta yacharqanki. {yacha-tsi-yka:-ya-pti-q allin-ta know-caus-ipfv-plv-subds-1.sbj.2.obj good-acc yacha-rqa-nki} learn-pst-2.sbj ‘When we were teaching teach you, you learned well.’ A subordinate clause with -pti is viewed as occurring before the main verb. However, the use of aspectual suffixes further delimits the temporal relation between clauses. In (179), the punctual aspect suffix -ri indicates that the subordinate takes place right before the main event. On the other hand, continuous -yka in Example (180) suggests that the main event took place while the subordinate event was unfolding.28

4.8.4.2.3.2 purpose constructions with nominalizers (-na-paq and -q with motion event) Many types of verb modifiers can be created in HQ through the use of infinitives with case markers. Alongside use of the genitive case -pa and the accusative case -ta to do this (see Sections 4.5.3.3.1 and 4.5.3.3.2), the ablative case -pita or the locative case -chaw can be used to specify causes or specific circumstances connected to any event, e.g. in (181) or (182). (181) yarquyniykipita {yarqu-y-ni-yki-pita} depart-inf-ep-2.poss-abl ‘because of your departure’ (182) papa allaychaw {papa alla-y-chaw} potato harvest-inf-loc ‘at the potato harvest’ Case-based constructions include any nominalizer, except the agentive -q. The most frequent subordinate clause based on a case marker is the purpose construction with the potential nominalizer -na and the dative case -paq. The potential nominalizer indicates that the event has not taken place yet, but it is possible. At its most basic, -na-paq indicates a generic goal, e.g. in (183). 28 See (28) in Section 4.5.3.3.8. for other ways to convey simultaneous events using the limitative case -kama.

136

(183) papata murunapaq {papa-ta muru-na-paq} potato-acc sow-pot.nmlz-dat ‘in order to sow potato’ The nominal agreement markers specify the subject of the subordinate clause. However, if an overt subject is added, it remains in the nominative case (or as a modifier of the nominalizer), but it is never marked genitive, as usual with embedded clauses. This is illustrated in (184). (184) … mana gendarmikuna tsariya:nanpaq {mana gendarmi-kuna no police-pl tsari-ya:-na-n-paq} catch-plv-pot.nmlz-3.poss-dat ‘… so that the police didn’t catch him’ This frequent purpose adverbial construction has been illustrated before in (91), (94), (95), and (146). However, -na-paq is never used with verbs of motion. For those cases, a purpose construction with the agentive nominalizer -q conveys the goal. The subject is always identical in the motion verb and the nominalized form. An example of this construction can be found in (176),where the form urya-q ayway ‘go to work’ (lit. ‘go worker’) indicates the goal of the motion. The q purpose subordinate can be placed before or after the main verb, as in (185). (185) yarqunki pukllaq {yarqu-nki puklla-q} exit-2.sbj play-ag ‘You went out to eat.’

4.8.4.2.3.3 conditional clauses Conditional subordinate clauses introduce the logical condition for the main clause. There are two ways to introduce the conditional clause in HQ. First, the conjunction sitsun introduces this subordinate with an inflected verb. This conjunction comes from Spanish si ‘if ’ and the irrealis/subjunctive marker -tsun (likely to be cognate with the injunctive marker discussed in Section 4.6.3.4). Although (186) uses a perfect conditional in the subordinate, forms in the present, past, or future are allowed too: (186) Sitsun mana padrinu: kankiman karqan, ushaka:tsiqmanmi. {sitsun mana padrinu-: ka-nki-man if no godfather-1.poss be-2.sbj-cond ka-rqa-n usha-ka:-tsi-q-man-mi} be-pst-3.sbj finish-spon-caus-1.sbj.2.obj ‘If you hadn’t been my godfather, I would destroy you.’

huaylas (ancash) quechua Conditional clauses are also introduced with subordinators -r and -pti marked with the topic suffix -qa. Examples (55) and (108) illustrate muna-r-qa ‘if you want’ and tari-mu-ptiyki-qa ‘if you don’t manage to find (food),’ both conditional subordinate clauses.

4.9 Text sample The following text is taken from Voces Quechuas (Julca Guerrero 2017: 169), a collection of texts in Ancash Quechua, adapted to the orthography used in this volume and with minor adjustments for clarity. The translation is mine. (187) Kay ashmaqa ampipapis allim rikan. {kay ashma-qa ampi-pa-pis alli-m dem.prox animal-top night-gen-add good-evd rika-n} see-3.sbj ‘This animal also sees well at night.’ (188) Atuqpaqa chupanchawmi upaya:tsikuqnin kapun, … {atuq-pa-qa chupa-n-chaw-mi fox-gen-top tail-3.sbj-loc-evd upa-ya:-tsi-ku-q-ni-n ka-pu-n} stupid-tf-caus-mid-ag-ep-3.sbj be-ben-3.sbj ‘On its tail, the fox has something that stuns, …’ (189) … tsaywanmi runakunata punuka:tsin. {tsay-wan-mi runa-kuna-ta dem-com-evd person-pl-acc punu-ka:-tsi-n} sleep-spon-caus-3.sbj ‘… with which it makes people fall asleep.’ (190) Payqa allqukunatapis pantatsinmi. {pay-qa allqu-kuna-ta-pis panta-tsi-n-mi} s/he-top dog-pl-acc-add err-caus-3.sbj-evd ‘It also makes dogs confused.’ (191) Atuq hirkwan llallinakuyan. {atuq hirka-wan llalli-naku-ya-n} fox mountain-com defeat-recp-plv-3.sbj ‘The fox and the mountain compete against each other.’ (192) Manaraq tamya cha:muptin, atuq waqyaptinqa, … {mana-raq tamya cha:-mu-pti-n atuq not-cont rain arrive-cisl-subds-3.sbj fox waqya-pti-n-qa} cry-subds-3.sbj-top ‘If the fox howls before the first rain comes, …’

(193) … mana alli watapaqmi willamantsik; hirka llalliptinqa, … {mana alli wata-paq-mi willa-mantsik not good year-dat-evd tell-3.sbj.1.incl.obj hirka llalli-pti-n-qa} mountain win-subds-3.sbj-top ‘…it announces a bad year. If the mountain beats [the fox] …’ (194) … ñawpatam tamya cha:mun, tsayqa alli wata kananpaqmi. {ñawpa-ta-m tamya cha:-mu-n tsay-qa before-acc-evd rain arrive-cisl-3.sbj dem-top alli good

wata year

ka-na-n-paq-mi be-pot.nmlz-3.sbj-dat-evd

‘… and the rain comes before, then it is for a good year.’ (195) Tsaytam “puspa” patsa niyan. {tsay-ta-m puspa patsa ni-ya-n} dem-acc-evd puspa time say-plv-3.sbj ‘This they call “puspa” time.’ (196) Kayqa wayra killakunachawmi qallan. {kay-qa wayra killa-kuna-chaw-mi dem.prox-top wind month-pl-loc-evd qalla-n} begin-3.sbj ‘This begins in the windy months [August and September]. (197) Kay yachayqa, chakra nunakunapa rurayninkuna pushaqninmi. {kay yacha-y-qa, chakra nuna-kuna-pa.’ dem.prox know-inf-top farm person-pl-gen rura-y-ni-n-kuna pusha-q-ni-n-mi} make-inf-ep-3.poss-pl guide-ag-ep-3.sbj-evd ‘This knowledge is a guide for the actions of the rural man.’

4.10 Literature HQ has been the object of many studies, grammars, vocabularies, and texts since the late 1960s. I can only mention here the most influential works and those that have served as sources for those analysis and descriptions offered in this chapter that diverge from those found in the previous literature. Parker (1976) is the first substantial reference grammar for Ancash Quechua, and it is based on HQ. Julca Guerrero (2009) is a reference grammar that considers numerous contexts for dialectal variation. It also includes a chapter

137

carlos molina-vital devoted to the current phonological orthography of Ancash Quechua. While this work includes data from all the Ancash Quechua varieties, HQ is the main source. Early grammars and sketches of HQ are Escribens and Proulx (1970), Levengood de Estrella and Larsen (1972), and the extensive introduction in Swisshelm (1972). Regarding dictionaries, Swisshelm (1972) organized extensive fieldwork notes by José [Joseph] Ripkens in the Huaraz province. Parker and Chávez (1976) is a revision of Parker (1971). Amancio Chávez Reyes contributed terms from the Conchucos variety into a collection that was mostly a HQ lexicon. Carranza Romero’s (2003) dictionary, while based on a variety closer to the northern Conchucos ones, also includes terms from the Callejón de Huaylas. More recently, Julca Guerrero and C. Julca Guerrero (2016) have published a study of the lexical diversity and expressiveness in Ancash Quechua, based mostly on the Huaylas variety. Early works on phonology are Parker (1967; 1970) and Swisshelm (1971). The latter is a pioneering work containing many insights on stress, pitch, and phonotactics. Unfortunately, due to its limited availability, those observations have not been reviewed with modern methods yet. Vivas Bravo’s (2005) MA thesis studies vowel length using data from the Huaraz province. Morphosyntax has been studied mainly from a descriptive and typological perspective. Parker’s (1973a) groundbreaking work on verbal derivational morphology is based mainly on the Huaylas dialect, but he also uses data from his extensive fieldwork through Ancash during the 1960s. Swisshelm (1974) is also a study of verbal derivational morphology. It has the distinction of being the first corpus-based study ever done in Quechuan linguistics, and is one of the most comprehensive analyses of the distribution of those suffixes in

138

any Quechuan variety. Larsen (1976), also a study of derivational morphology, deserves mention due to the diversity of examples it includes. Villari’s dissertation (2017) studies the form and function of negation in the Huaraz province, while stressing the importance of shared knowledge and sociocultural settings. Villari and Menacho (2015) offer a general description of the middle-voice suffix -ku. This topic is expanded in Molina-Vital’s dissertation (in preparation), where this category is analyzed in terms of control and affectedness and in relation to causatives and the proposed category of spontaneous voice. The interaction between environment and spatial orientation in the southern Callejón de Huaylas is explored in Shapero’s (2017a) dissertation. While not based on HQ, but on the closely related Southern Conchucos variety, Daniel Hintz’s (2011) exhaustive monograph on verbal aspect, and Diane Hintz’s (2007) dissertation on the form and function of past tense markers under an interactional perspective are fundamental contributions to those areas of verbal morphosyntax. Many texts have been published in HQ. These include not only collections aimed for linguistic research, but textbooks for intercultural bilingual education (e.g. Camones Maguiña 2016), and numerous basic reading materials created by SIL and their associates (e.g. Granados Barreto 1995 and Colonia Macedo 2002). Among collections created to help linguistic research, Proulx and Méndez (1967) collects stories from the Yungay province. Pantoja Ramos et al.’s (1974) monumental collection of stories, memoirs, folklore, and instructive texts from the Huaraz province is still the most valuable source of data for HQ. More recently, Cháves Gonzales and Julca Guerrero (2009) and Julca Guerrero (2017) collected academic, narrative, poetry, and instructional texts from different varieties of Ancash Quechua, with a significant majority from the Callejón de Huaylas.

chapter 5

Southern Yauyos Quechua Aviva Shimelman

5.1 General information Yauyos [ISO 369-3: qux, Glottocode: yauy1235] is a critically endangered Quechuan language spoken in the Peruvian Andes, in the Province of Yauyos, Department of Lima.1 Yauyos comprises eight dialects. The northern three—formerly spoken in Alis and Tomas, Huancaya and Vitis, and Laraos—are extinct, and not enough is known about them to usefully include them in this analysis. This chapter, therefore, focuses exclusively on the five surviving dialects in the south. They are named for the districts in which they are spoken: Apurí, Madeán, and Viñac (amv), Azángaro, Chocos, and Huangáscar (ach), Lincha and Tana (lt), Cacra and Hongos (ch), and San Pedro and Liscay (spl). I refer to these collectively as “Southern Yauyos Quechua” (syq).2

district is more than a day’s walk from any other. The principal towns of all but three of the districts sit at altitudes around 3,300masl; the exceptions—Chocos, Huangáscar, and Tana—sit at just under 3,000masl. Also spoken in the region are two Aymaran languages, Jaqaru and Cauqui, in and around the villages of Tupe and Cachuy, respectively. For treatments of the relationships between the two language families, see Adelaar with Muysken (2004: 259–317) and Cerrón-Palomino (1994; 2000b). On Jaqaru, see Hardman (1966; 1983a; 2000) and Chapter 9 by Matt Coler in this volume. In my own analysis of almost 2,500 lexical items of SYQ, only two that were not cognate with terms widespread in Quechuan could be related to Jaqaru terms: kallwi- ‘to cultivate’ and liklachiku ‘underarm.’3

5.1.1 Geography and environment The first two sets of towns are located in the valley created by the Huangáscar river and its principal tributary, the Viñac river. The second two sets are located in the valley created by the Cacra river and its principal tributaries, the Lincha and Paluche rivers. The two valleys are separated by a chain of rather high and rocky hills, running from Cerro Shallalli in the west to Cerro Pishqullay in the east. The region can be contained within an area of 40km2 , so no 1 Notational conventions: In order to be able to refer to items with different realizations in different SYQ dialects, I use italicized capital letters. R indicates an alternation between the allophones of */r/, [ɾ] and [l], as in runa ~ luna ‘person.’ H indicates an alternation between the allophones of */h/, [h] and [ʃ], as in hamu- ~ shamu- ‘to come’; in the specific case of the perfect marker, H refers to an alternation between [h] and [ø], as in -sha ~ -sa ‘prf.’ Q, S, and K also indicate alternation with [ø] in dialectal variants of different morphemes, as in the simple past suffix -rqa ~ ra, the perfect suffix -sha ~ -sa, and the accusative suffix -kta ~ -ta (see Table 5.9). 2 As is to be expected, there is further microvariation within each of the dialects, both in the lexicon and morphology, with the most prominent variation being that between Cacra and Hongos. The former, for example, is more likely than the latter to employ -r (realized [l]) in place of -shpa for the identical-subjects subordinator, and -mish in place of -hina for the comparative. Space limitations preclude cataloging all such microvariation here.

5.1.2 Speaker population and language endangerment Yauyos now counts fewer than 400 speakers, most over 65 years old, and all but the most elderly are fully bilingual in Spanish. No dialect of the language is currently being transmitted. Most young people opt to leave their villages when they finish secondary school, generally heading to Lima or, to a lesser extent, Cañete. Those who remain in the Andes generally practice subsistence farming and maintain small herds of sheep, goats, and cows. They identify ethnically—to the extent that they do—as “Quechua.” They are, however, more likely to call themselves serrano (Sp. ‘mountain person), an identity with occupational/class, regional, cultural, and racial vectors.

3 In numbered examples, dialect is indicated in the first line, as is standard. In running-text examples, it is specified only in those cases in which the form is not attested in all dialects. I have purposely chosen examples that are indeed universally attested. Please be aware that this may exaggerate the degree of uniformity among the five dialects.

Aviva Shimelman, Southern Yauyos Quechua. In: The Oxford Guide to the Languages of the Central Andes. Edited by: Matthias Urban, Oxford University Press. © Aviva Shimelman (2024). DOI: 10.1093/oso/9780198849926.003.0005

139

aviva shimelman

5.2 Internal variation and classification There is very little socioeconomic variation among speakers, so many of the usual axes of sociolinguistic variation are not relevant in the context. Intergenerational variation is marked above all by degrees of fluency and, in particular, by the extent to which speakers integrate Spanish loanwords. Virtually all internal variation in the language is diatopic. Yauyos is located on the border of the two large, contiguous zones where the languages of the two different branches of the Quechuan language family are spoken. Central Quechua is spoken immediately to the north, in the Department of Junín and in the northern part of the Department of Lima; Southern Quechua is spoken immediately to the south, in the Departments of Huancavelica and Ayacucho.4 Yauyos manifests characteristics of both groups. Take 1st person marking: three dialects, Azángaro–Chocos– Huangáscar, Cacra–Hongos, and San Pedro–Liscay, use the same marking, vowel length, for the 1st person in both the nominal and the verbal paradigms and, further, indicate the 1st person object with -ma. These are the two characteristics that define a Quechuan language as belonging to the Central Quechua branch. The other two dialects, Apurí– Madeán–Viñac and Lincha–Tana, mark the 1st person differently in the nominal and verbal paradigms (with -y and -ni, respectively) and mark 1st person object with -wa. These two dialects, then, sort with the Northern and Southern Quechua languages respectively. That said, the Central Quechua-like SYQ dialects, Azángaro–Chocos–Huangáscar, Cacra–Hongos, and San Pedro–Liscay, manifest few of the other traits that set the Central Quechua languages apart from Northern and Southern Quechua and other Quechuan languages. They do use ñuqakuna in place of ñuqayku to form the 1st person plural exclusive, as well as -pa(:)ku to indicate the verbal plural. Crucially, however, so do both the SYQ dialects that resemble Northern and Southern Quechua. Further, none of the five manifest any other of the principal traits that generally set the Central Quechua languages apart from the rest: none of the five SYQ dialects uses -naw to form the comparative, -naq to form the narrative past, or 4

The standard circulated tree groups 9 of the 11 districts of southern Yauyos into 5 sets, assigning each of these sets the status of an independent language. Moreover, two of these sets are actually singletons, as Chocos is listed as independent of Azángaro-Huangáscar, to which it is essentially identical, and Apurí is listed as independent of Madeán-Viñac, to which it is essentially identical. Cacra-Hongos, the set that would deserve independent placement, if any did, appears nowhere at all. The fact that all of these “languages” are completely mutually intelligible does not justify this. It further seems unjustified to place the Quechua of single villages on the level of that of whole nations—Bolivia and Ecuador.

140

-piqta to form the ablative; nor do any, except for Cacra, use a cognate of -r in place of -shpa to form same-subject subordinate clauses. The two SYQ dialects that pattern with Northern and Southern Quechua do manifest several of the traits that set specifically the Southern Quechua languages apart from the rest: they use the diminutive -cha, the emphatic -aRi, the assertive -puni, and the alternative 1st person plural subject conditional -chuwan; the Apurí–Madeán–Viñac dialect additionally uses the alternative 2nd person subject conditional -waq. Crucially, however, the three Central Quechua-like SYQ dialects also use three of these: -cha, -aRi, and -chuwan. Further, all five share with Ayacucho Quechua the unique use of what I call the “evidential modifier” -ki. They share a second rare form, -tamu, a modal suffix indicating irreversible change, with others of the varieties of Southern Quechua spoken in the departments of Cuzco and Apurímac (cf. Itier 1997).5 Nevertheless, none of the five manifest any of the other defining traits of the Southern Quechua languages: none use -ku to indicate the 1st person plural exclusive or the 3rd person plural, none use -chka to form the progressive or -nka to form the distributive, and none suffered the fusion of */ch/ with */ćh/or */s/ with */sh/, as those varieties did. Rather, the dialects of SYQ are mutually intelligible, and they all share characteristics that set them apart from all the other Quechuan languages. With the single exception that Cacra–Hongos uses the accusative form -kta in place of -ta and the locative -ćhaw alongside locative -pi, all five dialects employ the same case system, which includes the unique ablative form -paq. All dialects use the progressive form -ya, all employ the plural -kuna with non-exhaustive meaning, and all employ the same unique system of evidential modification. Further, with the exception of the reflexes of /r, s, h/ in Cacra, all the five dialects are uniform phonologically, all employing a highly conservative system that retains all those phonemes hypothesized by Parker (1969d) and Cerrón-Palomino (1987b) to have been present in Proto-Quechua.

5.3 Phonology and phonotactics SYQ is, phonologically, extraordinarily conservative, with four of its five dialects essentially instantiating the systems proposed for Proto-Quechua in Landerman (1991). All SYQ dialects retain contrasts between 5 Thanks to an anonymous reader for suggesting the inclusion of this element.

southern yauyos quechua (i) (ii) (iii) (iv) (v)

/ch/ and /ćh/ /k/, /q/, and /h/ /n/ and /ñ/ /l/ and /ll/, and /s/ and /sh/.

None of the dialects include aspirated or ejective consonants in its phonemic inventory. Vowel length is contrastive in the grammars but not the lexicons of the dialects of Azángaro– Chocos–Huangáscar, Cacra–Hongos, and San Pedro–Liscay.6 In these dialects, as in all the Central Quechua languages with the exception of Pacaraos, vowel length marks the 1st person in both the nominal possessive and verbal paradigms (e.g. wasi-: ‘my house’ and puri-: ‘I walk’).

5.3.3 Syllable structure and root structure: phonotactic constraints

5.3.1 Vowel phonemes SYQ counts three native vowel phonemes: /a/, /i/, and /u/. Table 5.1 locates these along the standard axes. Table 5.1 Vowel inventory of SYQ Front Central Back High i Low

u a

In words native to SYQ, the closed vowels /i/ and /u/ have mid and lax allophones [e], [ɪ] and [o], [υ], respectively. As in other Quechuan languages, the alternations [i] ~ [e, ɪ] and [u] ~ [o, ʊ] are conditioned by the environment: the lax variants appear either (i) in syllables including [q] (e.g. /qiʎa/ ‘lazy’ → [qeʎa] and /atuq/ ‘fox’ → [atoq]) or (ii) across syllable boundaries preceding /q/ and either immediately adjacent to it or separated from it by a nasal or liquid (e.g. /hiqa-/ ‘to go up’ → [hɪqa] and /urqu/ ‘mountain’ → [ʊrqʊ]).

5.3.2 Consonant phonemes In all SYQ dialects, the consonant inventory counts 17 native and five borrowed phonemes. The native phonemes include the voiceless plosives /p/, /t/, /ch/, /ćh/, /k/, and /q/, the voiceless fricatives /s/, /sh/, and /h/, the nasals /m/, /n/, and /ñ/, the laterals /l/ and /ll/, the tap /r/, and the approximants /w/ and /y/. Borrowed from Spanish are the voiced plosives /b/, /d/, and /ɡ/, the voiceless fricative /f/, and the trill /rr/. Table 5.2 locates these along the standard axes. 6

In the Cacra–Hongos dialect, the proto-phoneme */r/ is generally but not uniformly realized as [l] (e.g. /runa/ ‘person’ → [luna]), and word-initial */s/ and */h/ are generally but not uniformly realized as [h] and [ʃ], respectively (e.g. /sapa/ ‘alone’ → [hapa], /hamu/ ‘to come’ → [ʃamu]). That said, some lexemes resist these mutations (e.g. riqsi- ‘to become acquainted,’ siki ‘behind,’ and hapi- ‘to grab’). In Lincha and Tana, the immediate neighbors of Cacra and Hongos to the northeast and southwest respectively, speakers may realize word-initial */r/ and */s/ as [l] and [h], respectively, in a few cases (e.g. /runku/ ‘bag’ → [lunku]). These substitutions are not systematic, however.

This is not to say that no lexemes or morphemes feature long vowels. A few examples, principally from Cacra and to a lesser extent Hongos, can be found (e.g. ta:- ‘to sit,’ -pa:ku ‘joint action’).

Syllable structure in SYQ, as in other Quechuan languages, is (C)V(C), except in borrowed words. That is, syllables of the form CCV and VCC are prohibited. One vowel does not follow another without an intervening consonant, i.e. sequences of the form VV are prohibited. Only the first syllable of a word may begin with a vowel (e.g. a.pa- ‘to bring,’ ach.ka ‘a lot’). The majority of verb roots are disyllabic and end in a vowel. All of these constraints are common to the Quechuan languages generally.

5.3.4 Major morphophonological processes In SYQ, as in other Quechuan languages, the 1st person object suffix -ma(:), the cislocative/translocative suffix -mu, the benefactive suffix -pu, and the causative suffix -chi trigger the lowering of a preceding vowel [u] to [a] in the suffixes -yku, -ku, -naku, -paku, -pa:ku, -Ru, and -pu. A handful of lexical items, including yayku- ‘to go in’ and Riku- ‘to see,’ may also undergo vowel lowering when followed by -chi, -mu, or -ma: (e.g. to yayka- and mika-, respectively). Other major morphophonological processes in SYQ include vowel reduction in the evidential clitics (see Section 5.9.2) and the alternation -yki ~ -ki in the 2nd person possessive (see Section 5.5.3.1). Additionally, when a consonant-initial affix abuts a consonant-final stem or sequence, SYQ inserts the semantically void particle -ni to avoid illicit consonant clusters (e.g. *huk-yuq ‘one-poss’ → huk-ni-yuq ‘person with one’).

5.3.5 Stress assignment Primary stress falls on the penultimate syllable of a word (e.g. contrast yanápa-n ‘he helps’ with yanapá-ya-n ‘he is helping’). The first syllable of a word with more than four

141

aviva shimelman Table 5.2 Consonant inventory of SYQ Bilabial Labio-dental Alveolar Post-alveolar Retroflex Palatal Velar Uvular Glottal Voiceless plosive/affricate p

t

Voiced plosive

(d)

(b)

Fricative Nasal

(f)

ch

sh

Lateral

l

ll

Trill

(r)

Tap

ɾ w

5.3.6 Phonological developments from the protolanguage As noted in Section 5.3, SYQ is, phonologically, extraordinarily conservative. First, in SYQ, Proto-Quechua */ćh/ did not suffer de-retroflexion (to /ch/), as it did in Colombia, Ecuador, Bolivia, and Argentina, as well as in the east and south of Peru. SYQ thus retains Proto-Quechua forms like *ćhupa ‘tail’ and *qućha ‘lagoon.’ Minimal pairs illustrating the contrastive nature of /ćh/ and /ch/ include ćhaki ‘foot’ vs chaki ‘dry.’ Second, in SYQ, Proto-Quechua */q/ was neither velarized (to /k/) nor glottalized (to /h/) (which is not to say that these processes are the norm among the Quechuan languages). The language retains, for example, the Proto-Quechua forms *qusa ‘husband’ and *waqa- ‘to cry.’ Minimal pairs illustrating the contrast include qilla ‘lazy’ vs killa ‘moon.’ */h/ appears in SYQ, as in Proto-Quechua, principally word-initially, as in hapi- ‘to

142

q h

ñ

syllables generally receives a weak secondary stress. There are three exceptions to this rule. First, in all dialects, exclamations (see Section 5.8.3) often receive stress on the ultimate syllable (e.g. ¡Achachák! ‘What a fright!,’ ¡Achachalláw! ‘How awful!’). Second, in those dialects where vowel length indicates the 1st person, stress falls on the ultimate syllable when person marking (i.e. the suffix -:) is not followed by any other suffix or clitic (e.g. contrast taytá-: ‘my father’ with taytá-:-ta ‘to my father,’ where stress behaves normally). Finally, the emphatic clitic -Yá is always stressed when it comes last in the phonological word (see Section 5.7.2.1).

k (g)

n

Approximant

m

s

ćh

y

grab’ and hampi- ‘to cure.’ Third, in SYQ, Proto-Quechua */ñ/ did not undergo depalatalization (to /n/) as it did in the Quechuan languages of Central Peru. /ñ/ in SYQ is featured in the personal pronoun ñuqa, as well as in lexemes such as ñaka-ri- ‘to suffer’ and ñañu ‘thin.’ Examples of /ñ/ vs /n/ minimal pairs include the lexemes ana ‘mole’ vs aña- ‘to scold’ and the dummy noun na vs the discontinuative suffix -ña. Fourth, in SYQ */sh/ did not suffer depalatalization (to /s/), as it did in Southern Quechua. SYQ retains ProtoQuechua forms such as *shunqu ‘heart,’ *washa ‘back,’ and *ishkay ‘two.’ Examples of /s/ vs /sh/ minimal pairs include suqu ‘gray hair’ vs shuqu- ‘to sip.’ See Section 5.3.2 for specific developments in the Cacra–Hongos and Lincha–Tana dialects.

5.4 Parts of speech and transcategorical operations 5.4.1 Parts of speech The parts of speech in SYQ, as in other Quechuan languages, are substantives (e.g. waRmi ‘woman’), verbs (e.g. allwi- ‘to warp [a loom]’), ambivalents (e.g. paRa ‘rain; to rain’), and particles (e.g. mana ‘no, not’). Adjectives form a distinct class. The principal distinction among these is syntactic: substantives and verbs are subject to different patterns of inflection, ambivalents are unspecified for part of speech and may inflect either as substantives or verbs, and particles and

southern yauyos quechua adjectives do not inflect. One subclass of particles, the interjections (e.g. ¡Alaláw! ‘How cold!’), can be set apart from the rest, morphologically: interjections may end in a consonant and bear principal stress on the final syllable. Adjectives are remarkable in that they are often reduplicated. The effect is augmentative (uchuk ‘small’ → uchuk-uchuk ‘very small’). When adjectives are reduplicated, the last consonant or the last syllable of the first instance is generally elided (e.g. alli-allin ‘very good,’ hat-hatun ‘very big’).

5.4.2 Transcategorical operations Four suffixes—the nominalizer -na, agentive -q, resultative -sHa, and infinitive -y—derive nouns from verbs in SYQ. Five suffixes derive verbs from substantives: factive -cha, reflexive -ku, simulative -tuku, and inchoative -ya. Additionally, two verbs can suffix to nouns to derive verbs: -na ‘to do, act’ and -naya ‘to give desire.’

5.4.2.1 Nominal derivation 5.4.2.1.1 -na ‘nominalizer (nmlz)’ -na derives nouns that refer to (i) the instrument with which the action named by the base is realized (e.g. alla-na ‘harvesting tool,’ from alla- ‘to harvest’), (ii) the place in which the event referred to occurs (e.g. michi-na ‘pasture,’ from michi‘to pasture’), and (iii) the object in which the action named by the base is realized (e.g. upya-na ‘drinking water,’ from upya- ‘to drink’).

5.4.2.1.2 -q ‘agentive (ag)’ -q derives nouns that refer to the agent of the verb to which it attaches (e.g. aRa-q ‘one that plows,’ from aRa- ‘to plow’). -q nominalizations may form adjectives and relative clauses (e.g. chinka-ku-q pashña ‘the lost girl,’ from chinka-ku- ‘to get lost’).

5.4.2.1.3 -sHa ‘resultative (res)’ -sHa derives stative participles that may modify other nominals (e.g. chaki-sHa ‘dried,’ from chaki- ‘to dry’). This morpheme is identical in form to the perfect tense marker (Section 5.6.3.2.5).

5.4.2.2 Verbal derivation 5.4.2.2.1 -cha ‘factive (fact)’ -cha suffixes to nouns to derive verbs: ‘to make N’ or ‘to make into N’ (e.g. siru-cha- ‘to form a hill,’ from siru ‘hill’), ‘to locate something in N’ (e.g. kustal-cha- ‘to put into sacks,’ from kustal ‘sack’),‘to locate N in/on something; to remove N’ (e.g. usa-cha- ‘to remove lice,’ from usa ‘lice’). Suffixing -cha to adjectives derives verbs: ‘to make A’ (e.g. qaćha-cha‘to make dirty,’ from qaćha ‘dirty’).

5.4.2.2.2 -ku ‘reflexive (refl)’ -ku suffixes to nouns referring to objects to derive verbs: ‘to make/prepare N’ (e.g. Madeán qisha-ku- ‘to make a nest’ from qisha ‘nest’). Suffixed to nouns referring specifically to clothing and other items that can be placed on a person’s body, -ku derives verbs: ‘to put on N’ (e.g. kata-ku- ‘to put on a shawl,’ from kata ‘shawl’). Suffixing to adjectives referring to human states—happy, guilty, envious—, A-ku has the meaning ‘to become A’ (e.g. piña-ku- ‘to become angry,’ from piña ‘angry’).

5.4.2.2.3 -tuku ‘simulative (simul)’ -tuku suffixes to nouns to derive verbs: ‘to pretend to be N’ or ‘to become N’ (e.g. maqta-tuku- ‘to pretend to be a young man,’ from maqta ‘young man’).

5.4.2.2.4 -ya ‘inchoative (inch)’ -ya suffixes to nouns to derive verbs: ‘to become N’ (e.g. Rumi-ya- ‘to petrify,’ from Rumi ‘stone’) and ‘to perform a characteristic action with N’ (e.g. kwahu-ya- ‘to add curdling agent,’ from kwahu ‘curdling agent’). Suffixed to adjectives, -ya derives verbs: ‘to become A’ (e.g. alli-ya- ‘to get well, to heal,’ from alli ‘good’).

5.4.2.2.5 -na ‘verbalizer (vblz)’ -na suffixes to demonstrative pronouns, yielding transitive verbs: ‘to be thus’ or ‘to do thus’ (e.g. ima-na- ‘to do something, to happen,’ from ima ‘what’).

5.4.2.2.6 -naya ‘desiderative (desr)’

5.4.2.1.4 -y ‘infinitive (inf)’

-naya suffixes to nouns to derive verbs: ‘to give the desire for N,’ as in (1).7 Note that, syntactically, the subject is impersonal.

-y indicates the infinitive (e.g. tushu-y ‘to dance, dancing,’ from tushu- ‘to dance’). -y nominalizations may also refer to the object or event in which the verb stem is realized (e.g. ishpa-y ‘urine,’ from ishpa- ‘to urinate’ or nana-y ‘pain,’ from nana- ‘to hurt’).

7 -naya is unusual in that it may immediately attach both to noun and verb roots (see Section 5.6.2.1.9).

143

aviva shimelman (1) Apurí–Madeán–Viñac Mishkinayaruwan. {mishki-naya-ru-wa-n} fruit-desr-urgt-1.obj-3.sbj ‘I have a craving for fruit.’ (lit. ‘It gives me the desire for fruit.’)

5.4.2.3 Other derivational processes Nouns referring to time may occur adverbially, without inflection. The genitive-locative and accusative may form adverbs from nouns and adjectives, respectively (e.g. tuta-pa ‘at night,’ from tuta ‘night,’ or allin-ta ‘well,’ from allin ‘good’). Additionally, adverbs can sometimes be derived from adjectives with the suffixation of the restrictive suffix -lla (e.g. sumaq-lla ‘nicely,’ from sumaq ‘pretty’). Example (2) gives a more extended illustration from a popular story about a fox. (2) San Pedro–Liscay “¿Imaynataq waqayanki qamqa? ¡Kuyayllata waqanki!” nin. {imayna-taq waqa-ya-nki qam-qa why-seq cry-prog-2.sbj you-top kuya-y-lla-ta waqa-nki ni-n} love-inf-rstr-acc cry-2.sbj say-3.sbj “‘Compadre, why are you crying? How beautifully you sing!” he said.’

5.5 Nominals 5.5.1 Subclasses of nominals In SYQ, as in Quechuan languages generally, the class of substantives is defined as including nouns (e.g. wasi ‘house’), pronouns (e.g. ñuqa-nchik ‘we’), interrogative-indefinites (e.g. may ‘where’), adjectives (e.g. putka ‘turbid’), preadjectives (e.g. dimas ‘too’), and numerals (e.g. kimsa ‘three’). A few substantives belong to multiple classes. There is a dummy noun, na. All substantives, with the exception of the dependent pronouns (e.g. Sapa ‘only, alone’), may occur as free forms.

5.5.1.1 Pronouns and other pro-forms In SYQ, as in other Quechuan languages, pronouns may be sorted into four classes: personal pronouns, demonstrative pronouns, dependent pronouns, and interrogativeindefinite pronouns.

5.5.1.1.1 Personal pronouns The personal pronouns in SYQ are ñuqa ‘I,’ qam ‘you,’ pay ‘s/he,’ ñuqa-nchik ‘we,’ qam-kuna ‘you(pl),’ and pay-kuna

144

‘they.’ SYQ makes no distinction between subject, object, and possessive pronouns. With all three, case marking attaches to the same stem (e.g. ñuqa-ta ‘me’ or ñuqa-pa ‘my’). Nominative may be considered zero-marked. SYQ makes available a three-way distinction in the 1st person plural among the dual form ñuqa-nchik, the inclusive ñuqa-nchikkuna, and the exclusive ñuqa-kuna. In practice, except in Cacra–Hongos, ñuqa-nchik is employed with dual, inclusive, and exclusive interpretations. Verbs and substantives appearing with the inclusive ñuqa-nchik-kuna inflect in the same manner as verbs and substantives appearing with the dual/default ñuqa-nchik; verbs and substantives appearing with the exclusive ñuqa-kuna inflect in the manner as those appearing with the singular ñuqa. Table 5.3 summarizes this information. Table 5.3 Personal pronouns of SYQ sg

du

pl

1 ñuqa ñuqa-nchik incl excl

ñuqa-nchik-kuna ñuqa-kuna

2 qam

qam-kuna

3 pay

pay-kuna

5.5.1.1.2 Demonstrative pronouns The demonstrative pronouns—identical in form to the demonstrative determiners—are kay ‘this,’ chay ‘that,’ and wak ‘that (other).’ Table 5.4 lists the demonstrative pronouns. chay may have proximal or distal referents. The demonstrative pronouns may substitute for any phrase or clause. They can (but need not) inflect for number. Table 5.4 Demonstrative pronouns of SYQ Proximal

kay

Unspecified

chay

Distal

wak

5.5.1.1.3 Dependent pronouns The dependent pronouns are kiki ‘oneself,’ sapa ‘only, alone,’ llapa ‘all,’ and kuska ‘together.’ These pronouns are dependent in the sense that they cannot occur uninflected: the suffixes of the nominal possessive person-marking paradigm attach to dependent pronouns indicating the person and, in the case of the 1st person, the number of the referent

southern yauyos quechua of the pronoun (e.g. with the possessive suffix of the 2nd person sapa-yki ‘you alone’ or with the possessive suffix of the 1st person plural llapa-nchik ‘all of us’). One additional pronoun may appear suffixed with substantive person inflection: wakin ‘some…,’ ‘the rest of…’ Dependent pronouns function in the same manner as personal pronouns do: they may refer to any of the participants in an event, whether subject or object, they inflect obligatorily for case and optionally for number, and they may be followed by enclitics, as in (3). All except kiki may occur as free forms as well; it is, however, only as adjectives that they may occur uninflected; as pronouns or adverbs, all still demand inflection.

5.5.1.2 Numerals SYQ employs two sets of cardinal numerals. The first is native to Quechua; the second is borrowed from Spanish. The latter is always used for time and almost always for money. Also borrowed from Spanish are the ordinal numerals, primiru ‘first,’ sigundu ‘second,’ and so on. There is no set of ordinal numerals native to SYQ. The set of cardinal numerals native to SYQ includes twelve members. Table 5.6 lists them. Table 5.6 Numerals of SYQ

(3) Azángaro–Chocos–Huangáscar Kikinkamaćh wañuchinakura. {kiki-n-kama-ćhi wañu-chi-naku-ra-ø} self-3.poss-lim-evc die-caus-recp-pst-3.sbj ‘They must have killed each other themselves.’

5.5.1.1.4 Interrogative-indefinites SYQ has seven interrogative-indefinite stems: pi ‘who,’ ima ‘what,’ imay ‘when,’ may ‘where,’ imayna ‘how,’ mayqin ‘which,’ imapaq ‘why,’ and ayka ‘how much or how many,’ as shown in Table 5.5. These form interrogative, indefinite, and negative indefinite pronouns. Interrogative pronouns are formed by suffixing the stem—generally but not obligatorily—with either the sequential suffix -taq, the continuative suffix -raq, or any of the evidential clitics, -mi, -shi, or -ćhi (e.g. pi-taq ‘who,’ ima-raq ‘what’). Indefinite pronouns are formed by attaching the additive suffix -pis to the stem (e.g. pi-pis ‘someone’ or ima-pis ‘something’); negative indefinite pronouns are formed by preceding the indefinite pronoun with mana ‘no’ (e.g. mana pi-pis ‘no one’ or mana ima-pis ‘nothing’).

Number

Term

‘one’

huk

‘two’

ishkay

‘three’

kimsa

‘four’

tawa

‘five’

pichqa

‘six’

suqta

‘seven’

qanchis

‘eight’

pusaq

‘nine’

isqun

‘ten’

ćhunka

‘100’

paćhak

‘20,’ ‘30,’ and so on are formed by placing a unit numeral in attributive construction with ćhunka ‘ten’ (e.g. ishkay ćhunka ‘twenty’). ‘21’ and ‘22’ etc., then, are formed by following

Table 5.5 Interrogative indefinites of SYQ Stem

Gloss

(Negative) indefinite (Negative) indefinite gloss

pi

‘who’

(mana) pi-pis

‘some/anyone (no one)’

ima

‘what’

(mana) ima-pis

‘some/anything (nothing)’

imay

‘when’

(mana) imay-pis

‘some/any time (never)’

may

‘where’

(mana) may-pis

‘some/anywhere (nowhere)’

imapaq

‘why’

(mana) imapaq-pis

‘some/any reason (no reason)’

imayna

‘how’

(mana) imayna-pis

‘some/anyhow (no how)’

mayqin

‘which’

(mana) mayqin-pis

‘whichever (none)’

ayka

‘how many’

(mana) ayka-pis

‘some/any amount (none)’

145

aviva shimelman this construction with another unit numeral using the possessive suffix -yuq or, following a consonant, its allomorph, -ni-yuq (e.g. kimsa ćhunka pusaq-ni-yuq ‘38’).

5.5.1.3 Common nouns The class of nouns itself may be divided into four subclasses: regular nouns (e.g. wayta ‘flower’), time nouns (e.g. kanan ‘now’), gender nouns (e.g. tiya ‘aunt’), and locative nouns (e.g. qipa ‘behind’). The class of regular nouns includes all nouns not included in the other three classes. Although in this sense it is defined negatively, it includes more members than any of the others by far. An example is provided in (4). (4) Azángaro–Chocos–Huangáscar Ukuchapa ćhupallanta palumaqa quykun. {ukucha-pa ćhupa-lla-n-ta paluma-qa mouse-gen tail-rstr-3.poss-acc dove-top qu-yku-n} give-excep-3.sbj ‘The dove gave them the tail of a mouse.’ Nouns indigenous to SYQ do not inflect for gender. SYQ indicates biological gender either with distinct noun roots (e.g. maqta ‘young man,’ pashña ‘young woman’) or by modification with qari ‘man’ or warmi ‘woman’ in the case of people (e.g. qari wawa ‘boy child’ or warmi wawa ‘girl child’) or urqu ‘male’ or ćhina ‘female’ in the case of animals. A few nouns, all borrowed from Spanish, are inflected for gender (masculine and feminine ).

5.5.1.4 Locative nouns Locative nouns indicate relative position (e.g. chimpa ‘front,’ hawa ‘top’). They are inflected with the suffixes of the nominal possessive paradigm of person marking which indicate the person—and, in the case of the 1st person, also the number—of the complement noun, as in Example (5). (5) Apurí–Madeán–Viñac Kalamina hawanta pasarachisa ukunman saqakuykusa. {kalamina hawa-n-ta metal.roof above-3.poss-acc pasa-ra-chi-sa-ø uku-n-man pass-urgt-caus-narr.pst-3.sbj inside-3.poss-all saqa-ku-yku-sa-ø} go.down-refl-excep-narr.pst-3.sbj ‘He made him go on top of the tin roof and he fell inside.’

146

5.5.2 Nominal derivation Four “characteristic” derivational suffixes exist in SYQ that derive substantives from substantives: -kuna, -ntin, -sapa, and -yuq. The first two of these, -kuna and -ntin, indicate accompaniment, adjacency, or completeness (e.g. llama-nkuna ‘with her llama,’ from llama-n ‘her llama,’ or amiga-ntin ‘with her friends,’ from amiga ‘friend’); -yuq and -sapa indicate possession (e.g. llama-yuq ‘person with llamas’ or llama-sapa ‘person with more llamas than usual,’ both from llama ‘llama’).

5.5.2.1 -kuna ‘non-exhaustivity (nonexhst)’ -kuna indicates that the referent of its base is accompanied by another entity, generally of the same class, as in (6). (6) Apurí–Madeán–Viñac Umaykikuna nananman. {uma-yki-kuna head-2.poss-nonexhst

nana-n-man} hurt-3.sbj-cond

‘Your head, among other things, could hurt.’

5.5.2.2 -ntin ‘accompaniment, adjacency (incl)’ -ntin indicates that the referent of the base accompanies or is adjacent to another entity (e.g. allqu-ntin ‘with her dog,’ from allqu ‘dog’).

5.5.2.3 -sapa ‘multiple possession (mult.all)’ -sapa derives nouns referring to the possessor of the referent of the base. It differs from -yuq (see Section 5.5.2.4) in that what is possessed is possessed in greater proportion than usual (e.g. uma-sapa ‘person with a head bigger than usual,’ from uma ‘head,’ or yuya-y-sapa ‘person with a memory better than usual,’ from yuya-y ‘memory’).

5.5.2.4 -yuq ‘possession (poss)’ -yuq derives nouns referring to the possessor of the referent of the base (e.g. ćhakra-yuq ‘landowner,’ from ćhakra ‘field’). Ownership applies to substantives, including interrogative indefinites, numerals, pronouns, and so on (e.g. kimsa-yuq ‘person with three,’ from kimsa ‘three’). In case the base ends in a consonant, the semantically vacuous particle -ni precedes -yuq. Allomorphs /yuq/ and /qu/ are in free variation following /i/, as can be seen in (7), taken from a story about two brothers. (7) Apurí–Madeán–Viñac Huknin kasa kaqniqu huknin mana kaqniqu. {huk-ni-n ka-sa-ø ka-q-ni-yuq one-ep-3.poss be-narr.pst-3.sbj be-ag-ep-poss

southern yauyos quechua huk-ni-n one-ep-3.poss

mana ka-q-ni-yuq} no be-ag-ep-poss

‘One of them was wealthy, one of them had nothing.’

Table 5.7 Nominal inflection of SYQ Stem Root (Derivation) Person

5.5.2.5 -masi ‘partnership (part)’ -masi derives nouns referring to the partner of the referent of the base. These can generally be translated as ‘N-mate,’ ‘fellow N,’ or ‘co-N’ (e.g. puñu-q-masi ‘sleep-mate,’ from puñuq ‘sleep’). -masi is not very widely employed. (8) provides an example. (8) Lincha–Tana Chaywan apakaćhakushpam rikakayachin runamasinchiktaqa. {chay-wan dem-com rika-ka-ya-chi-n

Case

-na ‘nmlz’

-y ~ -: ‘1.poss’ -kuna ‘pl’ -Hina ‘simul’a

-q ‘ag’

-yki ‘2.poss’

-kama ‘lim’

-sHa ‘res’

-n ‘3.poss’

-man ‘all’

-y ‘inf’

-nchik ‘1pl.poss’

-pa ‘gen’

-ntin ‘incl’

-pa, -pi ‘loc’b

-sapa ‘mult.all’

-paq ‘purp’

-yuq ‘poss’

-paq ‘abl, ben, purp’

-masi ‘part’

-puRa ‘excl’ -ta ‘acc’c

apa-kaćha-ku-shpa-mi

-wan ‘com’

bring-freq-refl-subis.simult-evd

-ø ‘nom’

runa-masi-nchik-ta-qa} ‘Carrying those [their arms], they made our fellow people look.’

Number

-Rayku ‘csl’

see-passacc-prog-caus-3.sbj

person-part-1pl.poss-acc-top

Inflection

a

In Cacra and occasionally in Hongos, -mish replaces -hina to mark the comparative. b In Cacra-Hongos only, -ćhaw alternates with -pa and -pi to mark the locative. c In Cacra-Hongos only -Kta replaces -ta to mark the accusative.

5.5.2.6 -cha ‘diminutive (dim)’ -cha derives nouns with the meaning ‘little N.’ The forms mama-cha and tayta-cha, derived from mama ‘mother’ and tayta ‘father’ respectively, are lexicalized, meaning ‘grandmother’ and ‘grandfather’ respectively.

5.5.3 Nominal inflection Substantives in SYQ, as in other Quechuan languages, inflect for person, number, and case. The complete paradigm is displayed in Table 5.7. All case marking attaches to the last word in the nominal phrase. When a stem bears suffixes of two or three classes, these appear in the order person–number–case.

5.5.3.1 Person The nominal possessive person marking suffixes of SYQ are the same in all dialects for all persons except the 1st

person singular. Two of the five dialects—Apurí–Madeán– Viñac and Lincha–Tana—follow the pattern of Northern and Southern Quechua, marking the 1st person singular with -y; three dialects—Azángaro–Chocos–Huangáscar, Cacra–Hongos, and San Pedro–Liscay—follow the Central Quechua pattern, marking it with -: (vowel length). The SYQ nominal suffixes, then, are: -y or -: ‘1.poss,’ -yki ‘2. poss,’ -n ‘3.poss,’ -nchik ‘1pl.poss’ (e.g. mishi-y ~ mishi-: ‘my cat,’ from mishi ‘cat,’ or asnu-yki ‘your donkey,’ from asnu ‘donkey’). Stems of the following substantive classes may be suffixed with person suffixes: nouns (e.g. wambra-yki ‘your child,’ from wambra ‘child’), general numerals (e.g. kimsa-nchik ‘the three of us,’ from kimsa ‘three’), dependent pronouns (e.g. kiki-n ‘she herself,’ from kiki ‘self ’), demonstrative pronouns (e.g. chay-ni-y ‘this of mine,’ from chay ‘this’), and interrogative-indefinites, as in Example (9).

147

aviva shimelman (9) Apurí–Madeán–Viñac Ishkaynin, kimsan kashpaqa mikunyá. {ishkay-ni-n kimsa-n ka-shpa-qa two-ep-3.poss three-3.poss be-subis.simult-top miku-n-yá} eat-3.sbj-emph ‘If there are two of them or three of them, they eat.’ SYQ predicative possessive constructions are formed by combining a noun marked with the possessive suffixes with the verb ka- ‘to be,’ as in (10). (10) Allqun karqa. {allqu-n ka-RQa-ø} dog-3.poss be-pst-3.sbj ‘She had a dog.’ (lit. ‘Her dog was.’) Additionally, suffixes of the nominal possessive personmarking paradigm attach to the subordinating suffix -pti as well as to the nominalizing suffixes -na and -sa to form subordinate, purposive, complement, and relative clauses (see Section 5.8.5). In SYQ, subordinate clauses are, syntactically, nominals and, as such, are case marked with the suffix attaching to the last word in the phrase, as in (11), which is taken from a very well-known story where a condor disguises himself as a man. (11) Apurí–Madeán–Viñac Chayshi yaćharun kundur kashanta. {chay-shi yaćha-ru-n kundur dem-evr know-urgt-3.sbj condor ka-sha-n-ta} be-perf-3.poss-acc ‘That’s how they found out he was a condor.’

5.5.3.2 Number The plural is marked in all dialects of SYQ with -kuna (e.g. malta-kuna ‘young animals,’ from malta ‘young animal’). While number marking is the norm, it is not obligatory. NPs introduced by numerals or quantifying adjectives generally are not inflected with -kuna (e.g. kimsa maćhay ‘three caves,’ from kimsa ‘three’ and maćhay ‘cave’).

5.5.3.3 Case All case-marking processes consist in adding a suffix to the last word in the nominal group. Inflection is obligatory, with nominative considered zero-marked. Most case suffixes are mutually exclusive. Combinations are sometimes possible with the genitive/locatives -pa or -paq, the comitative-instrumental -wan, and the simulative -hina (e.g.

148

karsil-pa-hina ‘like in prison,’ from karsil ‘prison’). Where a noun phrase marked with genitive-locative -pa or -paq functions as an anaphor, the phrase may be case-marked as its referent would be, as in (12), taken from a discussion of mutual aid. (12) Cacra–Hongos Paqarin yanapamay, paqarin ñuqapakta chaypaq talpushun qampaktañataq. {paqarin yanapa-ma-y paqarin tomorrow help-1.obj-imp tomorrow ñuqa-pa-kta chay-paq talpu-shun i-gen-acc dem-abl plant-1pl.sbj.fut qam-pa-kta-ña-taq} you-gen-acc-disc-seq ‘Help me tomorrow! Mine tomorrow, then we’ll plant yours just after.’ SYQ counts 10 case suffixes: -ta ‘accusative,’ -pa ‘genitivelocative,’ -man ‘allative-dative,’ -pi ‘locative,’ -paq ‘ablative, benefactive, purposive,’ -hina ‘simulative,’ -kama ‘limitative,’ -wan ‘comitative-instrumental,’ -rayku ‘causal,’ and -puRa ‘exclusive.’ -pa is the default form for the locative, but -pi is often and -paq is sometimes used. The Cacra–Hongos dialect uses a fourth form, -ćhaw, common to the varieties of Central Quechua. The Cacra–Hongos dialect is also unique among the five in its realization of accusative -ta as -kta after a short vowel. The exclusive -pura and the causal -rayku are employed only rarely.

5.5.3.3.1 -ta ‘accusative (acc)’ In the Cacra–Hongos dialect, the accusative is realized -kta after a short vowel and -ta after a long vowel or consonant; in all other dialects it is realized as -ta in all environments. -ta may mark the object or goal of a transitive verb or the goal of movement of a person. -n-ta may indicate path. And, as indicated in Section 5.4.2.3, -ta marks substantives—nouns, adjectives, numerals, derived nouns—when they function as adverbs, as in (13). (13) Apurí–Madeán–Viñac Ishkay ishkaytam plantaramuni. {ishkay ishkay-ta-mi planta-ra-mu-ni} two two-acc-evd plant-urgt-cisl-1.sbj ‘I planted them two by two.’

5.5.3.3.2 -pa ‘genitive-locative (gen)’ -pa generally indicates possession; it is often paired with person inflection, as in (14), taken from a well-known story about two brothers, one poor, one rich.

southern yauyos quechua (14) Lincha–Tana Manañam miranñachu ganawninqa paypaqa. {mana-ña-mi mira-n-ña-chu no-disc-evd reproduce-3.sbj-disc-neg ganaw-ni-n-qa pay-pa-qa} cattle-ep-3.poss-top 3sg-gen-top ‘His animals no longer reproduce.’

5.5.3.3.3 -man ‘allative-dative (all)’ -man generally indicates movement toward a point (e.g. qiñwal-man ‘to the quingual grove,’ from qiñwal ‘quingual grove’) or the endpoint of movement or action more generally, as in (15), taken from a popular story about a condemned soul. (15) San Pedro–Liscay Wak wasikunamanshi yaykurun kundinawqa. {wak wasi-kuna-man-shi yayku-ru-n dem house-pl-all-evr enter-urgt-3.sbj kundinaw-qa} zombie-top ‘The zombie entered those houses.’ With verbs of giving, -man marks the recipient; with verbs of communication, it marks the person receiving the communication, as in (16), from a conversation about the Shining Path. (16) Azángaro–Chocos–Huangáscar Chayllapaq willakurusa tirruristaman hinaptin chayta wañurachin. {chay-lla-paq willa-ku-ru-sa dem-rstr-abl tell-refl-urgt-narr.pst tirrurista-man hinaptin terrorist-all then chay-ta wañu-ra-chi-n} dem-acc die-urgt-caus-3.sbj ‘So, they told it to the terrorists and then they killed him.’

5.5.3.3.5 -paq ‘ablative, benefactive, purposive (abl, ben, purp)’ As an ablative, -paq indicates provenance in space (e.g. maypaq ‘from where,’ from may ‘where’) or time (e.g. kanan-paq ‘from now [on],’ from kanan ‘now’), origin or cause (e.g. alalay-paq ‘from [because of] the cold,’ from alala-y ‘cold’) or the material of which an item is made (e.g. millwa-paq ‘from [out of] yarn,’ from millwa ‘wool’). As a benefactive, -paq indicates the individual who benefits from—or suffers as a result of— an event, as in (17), taken from a discussion about medicinal herbs. (17) Apurí–Madeán–Viñac Allinmi chay qarikuna mana ishpayta atipaqpaq. {Allin-mi chay qari-kuna mana good-evd dem man-pl no ishpa-y-ta atipa-q-paq} urinate-inf-acc be.able-ag-ben ‘This is good for men who can’t urinate.’ As a purposive, -paq indicates the purpose of an event. (18) gives an example. -paq may also alternate with -pa and -pi to indicate the genitive (e.g. ñuqa-paq ‘mine,’ from ñuqa ‘I’) or locative (e.g. paqcha-paq ‘in the waterfall,’ from paqcha ‘waterfall’). (18) Apurí–Madeán–Viñac Qawanaypaq imawan wañurun nishpa kićhani. {qawa-na-y-paq ima-wan see-nmlz-1.poss-purp what-com wañu-ru-n ni-shpa die-urgt-3.sbj say-subis.simult kićha-ra-ni} open-prf-1.sbj ‘To see what he died from, I said, and I opened him up.’ Finally, -paq is employed in the formation of comparative expressions, where it attaches to the base of comparison, as in (19), where the speaker is referring to the volume of milk one of her cows produced that day.

-pa may also indicate temporal and spatial location (e.g. biranu-pa ‘in the summer,’ from biranu ‘summer,’ llaqta-pa ‘in town,’ from llaqta ‘town’). In the Cacra–Hongos dialect, -ćhaw is used in addition to -pa as a locative. All dialects also employ the locative -pi (see Section 5.5.3.3.4).

(19) Apurí–Madeán–Viñac Qayna punćhawpaq masmi. {qayna punćhaw-paq mas-mi} previous day-abl more-evd

5.5.3.3.4 -pi ‘locative (loc)’

5.5.3.3.6 -hina ‘simulative (simul)’

As a locative, -pi indicates temporal (e.g. uktubri-pi ‘in October,’ from uktubri ‘Octobre’) and spatial location (e.g. San Pedro–Liscay lasta-pi ‘in the snow,’ from lasta ‘snow’).

-hina generally indicates resemblance or comparison (e.g. yawar-hina ‘like blood,’ from yawar ‘blood’). It can generally be translated as ‘like.’ In Cacra and sometimes in Hongos,

‘It’s more than yesterday.’

149

aviva shimelman -mish is employed in place of -hina (e.g. ima-mish ‘like what,’ from ima ‘what’).

5.5.3.3.7 -kama ‘limitative (lim)’ -kama –sometimes realized as -kaman—generally indicates a limit in space (e.g. kay-kama ‘toward here,’ from kay ‘this’) or time (e.g. kanan-kama ‘until now,’ from kanan ‘now’). When time is delimited by an event, as in the English phrases so I can sing or so you can dance, SYQ employs the structure verb stem-nmlz-poss-kama, i.e. the verb stem is nominalized with -na, marked for person, and affixed with the limitative -kama, as in (20). (20) purinaykikama {puri-na-yki-kama} walk-nmlz-2.poss-lim ‘so you can walk’ -kama can form distributive expressions: in this case, -kama attaches to the quality or characteristic that is distributed, as in (21), taken from a discussion about the Shining Path. (21) Azángaro–Chocos–Huangáscar Ćhayaramun arman qipikusakama. Manchaku:. {ćhaya-ra-mu-n arma-n arrive-urgt-cisl-3.sbj weapon-3.poss qipi-ku-sa-kama carry-refl-prf-all

mancha-ku-:} scare-refl-1.sbj

‘They arrived, each carrying weapons. I got scared.’

5.5.3.3.8 -wan ‘comitative-instrumental (com)’ -wan indicates means or company, i.e. an instrument or item which is essential to the event (e.g. tikshi-wan ‘with bits of leftover wool,’ from tikshi ‘bits of leftover wool’) or an animate individual who takes part in an event together with the performer (e.g. ñaña-n-wan ‘with her sister,’ from ñaña ‘sister’). -wan may mark coordinate relations between nouns or nominal groups (e.g. llama-wan alpaka-wan ‘the llama and the alpaca,’ from llama ‘llama’ and alpaka ‘alpaca’).

5.5.3.3.9 -rayku ‘causal (csl)’ -rayku indicates motivation or reason. It generally but not obligatorily follows inflection for person with the nominal possessive paradigm, as in (22). (22) Cacra–Hongos shamunanrayku {shamu-na-n-rayku} come-nmlz-3.poss-csl ‘on account of their coming’

150

5.5.3.3.10 -puRa ‘exclusive (excl)’ -puRa indicates the inclusion of the marked individual among other individuals of the same kind, as in (23). It can be translated as ‘among’ or ‘between.’ -puRa is not commonly employed; more common is the particle intri ‘between’ (< Sp. entre ‘between’). (23) Cacra–Hongos Walmipula qutunakulanchik. {walmi-pula qutu-naku-la-nchik} woman-excl gather-recp-pst-1pl.sbj ‘We women gathered among ourselves.’

5.6 Verbs In SYQ, as in other Quechuan languages, verb stems always end in a vowel (e.g. yanapa- ‘to help’). Verb stems are bound forms: with the single exception of haku ‘let’s go,’ a nearly uninflectable verb root, they never appear in isolation.8 They are subject to both derivational and inflectional processes, both suffixing (e.g. wañu-chi-n ‘he kills,’ from wañu-chi- ‘to kill,’ itself from wañu- ‘to die’). The order of inflectional suffixes is fixed; the order of derivational suffixes is highly regular but admits exception. Inflection for person is obligatory (note, though, that 3rd person is zeromarked in the past tense); all other inflectional categories and derivational processes are optional. The different person, tense, and mood suffixes are mutually exclusive within each category; different derivational suffixes, in contrast, may attach in series, as in Example (24), taken from a popular story about a girl who gets a lift from a condor. (24) San Pedro–Liscay Qipirachikusa. {qipi-ra-chi-ku-sa} carry-urgt-caus-refl-narr.pst ‘She got herself carried.’

5.6.1 Verb subclasses Quechuan verb stems are usually classed as transitive (e.g. qu- ‘to give’), intransitive (e.g. puñu- ‘to sleep’), or copulative (ka- ‘to be’). A fourth class can be set apart: onomatopoetic verbs (e.g. chuqchuqya- ‘to nurse; to make the sound of a calf nursing’). Special cases include the deictic verb hina-, the 8 An anonymous reviewer points out that haku might be more satisfactorily analyzed as a regular interjection, despite the fact that it does sometimes appear inflected in the 1st person plural.

southern yauyos quechua dummy verb na-, and the combining verbs -naya- ‘to give desire’ and -na- ‘to do what, matter, happen.’ Transitive verbs are standardly defined for Quechuan languages as those that can take direct objects that are case marked with the accusative suffix -ta, as in (25). (25) Bikuñata rutushaq. {bikuña-ta Rutu-shaq} vicuña-acc shear-1.sbj.fut ‘I will shear the vicuña.’ Intransitive verbs, in contrast, are those that cannot (e.g. wiña- ‘to grow’). Also included among the intransitives are the impersonal weather verbs, like qasa- ‘to freeze,’ which do not take subjects. The weather verbs admit only their corresponding weather nouns for subjects (e.g. para paraya-n ‘the rain is raining’). Verbs of motion (e.g. Hamu- ‘to come,’ lluqsi- ‘to exit’) form a subclass of intransitive verbs. These often have adverbial complements marked with the accusative -ta, the allative-dative -man, the ablative -paq, and the limitative -kama, and they may occur in clauses that include a nominalization with the agentive suffix -q indicating the purpose of movement (e.g. qutu-q ‘in order to gather,’ from qutu- ‘to gather’). SYQ counts a single copulative verb, ka-. Like the English to be, ka- has both copulative (‘I am a llama’) and existential (‘there are llamas’) interpretations. ka- is irregular: the 3rd person singular present tense form, ka-n, never appears in equational statements, but only in existential statements. Onomatopoetic verbs can be distinguished from other verbs by the shape of their stem. The majority involve the repetition—two to four times—of a syllable or syllable group, most often with the suffixation of -ya. Four patterns dominate: Pattern (i) involves the repetition of a single syllable twice or three times, generally with -ya or, more rarely, -ku or -yaku, i.e. (σ)σσ(-ya)(-ku) (e.g. qurqurya- ‘to snore’ or luqluqluqya‘to boil’). Pattern (ii), like Pattern (i), involves the repetition of a single syllable generally with -ya or, more rarely, -ku or -ya-ku. Pattern (ii) differs from pattern (i), however, in that the repeated syllable is always repeated three times, it never includes a coda, and it is preceded by a non-cognate syllable, which generally, if not always, includes the same vowel as does the repeated syllable, i.e. σi σj σj σj (-ya)(-ku) (e.g. bunrururu- ‘thunder’). Pattern (iii) replaces the single syllable of pattern (i) with a two-syllable unit. A variation on Patterns (i) and (iii) is also common: the final consonant in the final iteration of the repeating unit is elided (e.g. waqwaqwaya- ‘to guffaw’).

5.6.2 Verbal derivation A set of 19 suffixes derives verbs from verbs. These are: -cha ‘diminutive,’ -chi ‘causative,’ -ka ‘passive, accidental,’ -kaćha ‘frequentive,’ -ku ‘reflexive, middle, medio-passive, passive,’ -lla ‘restrictive, limitative,’ -mu ‘cislocative, translocative,’ -naku ‘reciprocal,’ -naya(:) ‘desiderative,’ -pa ‘repetitive,’ -pa(:)ku ‘joint action,’ -pu ‘benefactive,’ -Ra ‘durative,’ -Raya ‘passivizer,’ -Ri ‘inceptive,’ -Ru ‘action with urgency or personal interest,’ -sHi ‘accompaniment,’ -tamu ‘irreversible,’ and -yku ‘exceptional performance.’

5.6.2.1 Affixes Of the 19, arguably only four—the causative -chi, the reflexive -ku, the reciprocal -naku, and the desiderative -naya(:)—actually change the root/stem’s theta structure and derive new lexical items. The rest specify mood or aspect or otherwise function adverbially. Each affix admits multiple interpretations; listed here are only the most common.

5.6.2.1.1 -cha ‘diminutive (dim)’ -cha indicates action performed by a child or in the manner of a child, or action of little importance (e.g. puklla-cha ‘to play like a child,’ from puklla- ‘to play’).

5.6.2.1.2 -chi ‘causative (caus)’ -chi increases the addicity of the root, deriving verbs with the meaning ‘to cause to V’ or ‘to permit to V, as in (26a). Compounded with the reflexive -ku, -chi derives verbs with the meaning ‘cause oneself to V’ or ‘cause oneself to be V-ed,’ as in (26b). (26) a. Ñakayachiwan. {ñaka-ya-chi-wa-n} hurt-prog-caus-1.obj-3.sbj ‘It’s making me suffer.’ b. yanapachiku{yanapa-chi-ku-} help-caus-refl‘to get oneself helped’

5.6.2.1.3 -ka ‘accidental (passacc)’ -ka indicates that the event referred to is not under the control either of a participant in that event or of the speaker (e.g. puñu-ka- ‘to fall asleep,’ from puñu- ‘to sleep’).

151

aviva shimelman 5.6.2.1.4 -kaćha ‘frequentive (freq)’ -kaćha indicates extended or repetitive action (e.g. kurrikaćha- ‘to run around and around,’ from kurri- ‘to run,’ qawa-kaćha- ‘to look around,’ from qawa- ‘to look’).

5.6.2.1.5 -ku ‘reflexive, middle, medio-passive, passive (refl)’

(28) Azángaro–Chocos–Huangáscar Huk punćhaw huk tuta nanapashunki. {huk punćhaw huk tuta nana-pa-shunki} one day one night hurt-repet-2.obj ‘One day and one night it’s hurting and hurting you.’

5.6.2.1.11 -pa(:)ku ‘joint action (jtact)’

-ku is rather polyvalent, deriving reflexives, middles, mediopassives, and passives (e.g. mancha-ku- ‘to scare oneself, to get scared,’ from mancha- ‘to scare,’ or suwa-ku- ‘to steal for oneself,’ from suwa- ‘to steal’).

-pa(:)ku indicates joint action by a plurality of individuals, e.g. trabaha-pa:ku- ‘to work (together with others),’ from trabaha- ‘to work.’

5.6.2.1.6 -lla ‘restrictive, limitative (lim)’

-pu indicates that an action is performed on behalf—or to the detriment—of someone other than the subject (e.g. priparapu- ‘to prepare (for someone else),’ from pripara- ‘to prepare’). The form -paku—possibly a compound of -pu with -ku—indicates that the action is performed as a means or preparation for something else which supersedes it, which includes all remunerated labor (e.g. awa-paku- ‘to weave (for others, to make money),’ from awa- ‘to weave’).

-lla indicates that the event referred to remains limited to itself and is not accompanied by other events (e.g. lluqsi-lla‘to just leave,’ from lluqsi- ‘to leave’).

5.6.2.1.7 -mu ‘cislocative, translocative (cisl)’ -mu indicates—in the case of verbs involving motion— motion toward the speaker or toward a place that is indicated by the speaker (e.g. apa-mu- ‘to bring here,’ from apa‘to bring’). In the case of verbs that do not involve motion, -mu commonly adds a vector of movement to the action; such movement is commonly away from ego, as in (27). (27) mañakaramuy {maña-ka-ra-mu-y} ask-refl-urgt-cisl-inf ‘to go ask for’

5.6.2.1.8 -naku ‘reciprocal (recp)’ -naku indicates that the participants act reciprocally on each other (e.g. willa-naku- ‘to tell each other,’ from willa- ‘to tell’). Compounded with the causative -chi, -naku derives verbs with the meaning ‘to cause each other to V’ (e.g. willa-chinaku- ‘to cause each other to tell,’ from willa-chi- ‘to cause to tell’).

5.6.2.1.9 -naya(:) ‘desiderative (desr)’ -naya(:) alters the theta structure of the verb, deriving a compound verb meaning ‘to give the desire to V’ (e.g. mikunaya- ‘to be hungry,’ from miku- ‘to eat,’ thus literally ‘to give the desire to eat’; see also Section 5.4.2.2.6).

5.6.2.1.10 -pa ‘repetitive (repet)’ -pa indicates repetitive action (e.g. yata-pa- ‘to fondle,’ from yata- ‘to touch’). In (28), the speaker is talking about labor pains.

152

5.6.2.1.12 -pu ‘benefactive, translocative (ben)’

5.6.2.1.13 -Ra ‘durative (dur)’ -Ra indicates sustained action, i.e. action that extends in time without interruption (e.g. tiya-Ra ‘to remain,’ from tiya‘to sit’).

5.6.2.1.14 -Raya ‘passivizer (pass)’ -Raya derives passive from transitive verbs (e.g. wata-raya‘to be tied,’ from wata- ‘to tie’).

5.6.2.1.15 -Ri ‘inceptive (incep)’ -Ri indicates that the event is in its initial stages, that it has not yet concluded (e.g. shinka-ri- ‘to begin to get drunk, be tipsy,’ from shinka- ‘to be drunk’). A further example is in (29). (29) Cacra Nina:qa manalaq lupaliyanchu. Manalaq shansha: kanchu. {nina-:-qa mana-laq fire-1.poss-top no-cont lupa-li-ya-n-chu mana-laq burn-incep-prog-3.sbj-neg no-cont shansha-: ember-1.poss

ka-n-chu} be-3.sbj-neg

‘My fire still isn’t starting to burn. I still don’t have any embers.’ -Ri is common in apologetic statements and supplicatory commands (e.g. tiya-ri-ku-y ‘kindly sit down,’ from tiya- ‘to sit’). It is unattested in Hongos.

southern yauyos quechua 5.6.2.1.16 -Ru ‘various (urgt)’

5.6.2.2 Affix order

-Ru indicates action with urgency or personal interest, as in (30), where the speaker is addressing a member of the Shining Path, which terrorized the region in the 1980s and early 1990s, killing many residents.

There are several reasons to be cautious in proposing any template for the ordering of derivational suffixes in SYQ. First, in SYQ, series of more than three modal suffixes are rare; second, many combinations are not attested; and third, some suffixes may commute, which may but need not have semantic effect. Causative -chi is among the most mobile; change in its placement results in a change in verb meaning. Contrast e.g. (32a) with (32b).

(30) Apurí–Madeán–Viñac ¡Willakurushaqmi gwardyanman {willa-ku-ru-shaq-mi gwardyan-man tell-refl-urgt-1.sbj.fut-evd police-all tirruku kasaykita! tirruku ka-sa-yki-ta} terrorist be-prf-2.sbj-acc ‘I’m going to tell the police that you were a terrorist!’ -Ru is very frequently used with a completive interpretation, i.e. it adds aspectual valence (e.g. kani-ru-n ‘bit,’ from kani- ‘to bite’; see also Section 5.6.3.5).

5.6.2.1.17 -sHi ‘accompaniment (acmp)’ -sHi indicates accompaniment, often with the purpose of helping, as in (31). -sHi is realized [si] in San Pedro–Liscay and [ʃi] elsewhere. (31) Cacra–Hongos Kwidashimanchu. Hapalla: kwidaku: hapalla:. {kwida-shi-ma-n-chu hapa-lla-: care.for-acmp-1.obj-3.sbj-neg alone-rstr-1.poss kwida-ku-: take.care-refl-1.sbj

hapa-lla-:} alone-rstr-1.poss

‘He doesn’t help me take care of them. Alone, I take care of them. Alone.’

5.6.2.1.18 -tamu ‘irreversible (irrev)’ -tamu indicates a change of state that is irreversible, e.g. puchuka-tamu- ‘to wipe out,’ from puchuka- ‘to finish off,’ or ñiti-tamu- ‘to crush beyond repair, to flatten,’ from ñiti‘to crush.’ Not surprisingly, it is most frequently used in spontaneous discourse with wañu- ‘to die.’

5.6.2.1.19 -yku ‘exceptional performance (excep)’ -yku is perhaps the derivative suffix for which it is hardest to identify any kind of central interpretation. With regard to cognates in other Quechuan languages, it is sometimes said that it indicates action performed in some special or unexpected way (Hintz 2007: 185: table 6.2). An example for the use of -yku is silla-ku-yku- ‘to gallop,’ from silla ‘seat’).

(32) a. Wañuchinayawan. {wañu-chi-naya-wa-n} die-caus-desr-1.obj-3.sbj ‘It makes me want to kill.’ b. Wañunayachiwan. {wañu-naya-chi-wa-n} die-desr-caus-1.obj-3.sbj ‘It makes me feel like I want to die.’9 -chi and the continuous -ya(:) regularly commute with each other with no detectable semantic reflex. Contrast e.g. (33a) with (33b). (33) a. Mikuykayachin. {miku-yka-ya-chi-n} eat-excep-prog-caus-3.sbj ‘He is making it eat.’ b. Suliyachiyanchik. {suli-ya-chi-ya-nchik} sun-excep-caus-prog-3.sbj ‘We are making it [dry in the] sun.’ The same applies to the exceptional-performance suffix-yku and both reflexive -ku and urgent -Ru. Compare e.g. (34a) with (34b). (34) a. apuntaykukushpa {apunta-yku-ku-shpa} aim.at-excep-refl-subis.simult ‘taking aim at’ b. mashtakuykushpa {mashta-ku-yku-shpa} spread.out-refl-excep-subis.simult ‘spreading it out’ That said, as a general rule, -naku, -pa(:)ku, -Raya, -tamu, and -mu are later occurring, while -ka, -pa, -Ra, and -kaćha occur earlier. A derivational morpheme does not appear more than once within a single verb complex. 9 The example, from Albó (1964), as cited in Cerrón-Palomino (1987b: 284), was verified in SYQ.

153

aviva shimelman Table 5.8 Verbal inflection of SYQ Present

Future

Simple past

Habitual past

Conditional

1.sbj

-ni/:

-shaq

-RQa-ni/:

-q ka-ni/:

-y/:-man

2.sbj

-nki

-nki

-RQa-nki

-q ka-nki

-nki-man, -waq (amv)

-y

3.sbj

-n

-nqa

-Rqa

-q

-n-man

-chun

1pl.sbj

-nchik

-shun

-RQa-nchik

-q ka-nchik

-nchik-man, -chuwan

-shun

2.sbj.1.obj

-wa/ma-nki

-wa/ma-nki

-wa/ma-RQa-nki

-wa/ma-q ka-nki

-wa/ma-nki-man

3.sbj.1.obj

-wa/ma-n

-wa/ma-nqa

-wa/ma-RQa

-wa/ma-q

-wa/ma-man

3.sbj.1pl.obj -wa/ma-nchik –wa/ma(:)-shun -wa/ma-RQa-nchik

(unattested)

-wa/ma-nchik-man, -wa/ma-chuwan

1.sbj.2.obj

-yki

-sHQa-yki

-RQa-yki

(unattested)

-yki-man

3.sbj.2.obj

-shu-nki

-shu-nki

-shu-RQa-nki

(unattested)

-shu-nki-man

5.6.3 Inflectional categories of the verb Verbs in SYQ, as in other Quechuan languages, inflect for person, number, tense, conditionality, imperativity, aspect, and subordination. Table 5.8 supplies the paradigm of verbal inflection. Dialects differ from each other in four sets of cases in this context. They diverge in terms of (i) the form they use to encode 1st person subjects and objects (see Section 5.2), (ii) their realization of the simple past-tense morpheme -RQa, (iii) their realization of the perfect morpheme -sHa, and (iv) whether they feature an alternative 2nd person subject marker in the conditional. Table 5.9 summarizes this information. Table 5.9 Differences among SYQ dialects in the realization of verbal inflectional morphemes 1st person Simple past Perfect Alternative 2nd person subject conditional amv -ni

-rqa

-sa

yes

ach

-:

-ra

-sa

no

ch

-:

-la

-sha

no

spl

-:

-ra

-sa

no

lt

-ni

-ra

-sha

no

154

Imperative

5.6.3.1 Person and number SYQ non-subordinate verbs inflect for subject and object reference, as detailed in Sections 5.6.3.1.1 and 5.6.3.1.2. Table 5.8 displays the paradigm.

5.6.3.1.1 Subject The 1st person is indicated in Azángaro–Chocos– Huangáscar, Cacra–Hongos, and San Pedro–Liscay by -:, in the manner of the Central Quechua languages; in Apurí–Madeán–Viñac and Lincha–Tana, 1st person is indicated by -ni, in the manner of the Northern and Southern Quechua languages. In all dialects the 2nd person is indicated by -nki and the 3rd person by -n. -nchik indicates 1st person plural. Although syq makes available a three-way distinction between dual inclusive, plural inclusive, and plural exclusive in the 1st person (with ñuqanchik, ñuqanchikkuna, and nuqakuna, respectively), in practice, in all but the Cacra–Hongos dialect, the form associated with dual inclusive may be employed with all three meanings; inclusive and exclusive interpretations may additionally be supplied by context, both linguistic and extralinguistic. Subject markers attach to verb stems, plus derivational or inflectional suffixes, if any are present.10

10 The inflectional affixes of the possessive paradigm, -: and -y, are employed instead (e.g. lluqsi-y-man ~ lluqsi-:-man ‘I could leave’).

southern yauyos quechua 5.6.3.1.2 Subject and object reference Transitive non-reflexive verbs may inflect for subject–object reference. -wa in Apurí–Madeán–Viñac and Lincha–Tana and -ma(:) in Azángaro–Chocos–Huangáscar, Cacra–Hongos, and San Pedro–Liscay indicate a 1st person object, as in (35). (35) Quwa/manki. {qu-wa/ma-nki} give-1.obj-2.sbj ‘You give me.’ -yki and shunki indicate a 1st and a 3rd person respectively acting on a 2nd person in all SYQ dialects (as in (36b)), with the exception that, in Cacra, a 1st person acting on a 2nd person is indicated by -k, as in (36a). 3rd person objects are not indexed on the verb in any SYQ dialect. (36) a. Quk. {qu-k} give-1.sbj.2.obj ‘I give you.’ b. Quyki. {qu-yki} give-1.sbj.2.obj ‘I give you.’11 -nchik pluralizes the object acted upon by a 3rd person subject in the present or past (37). (37) Quwa/manchik. {qu-wa/ma-nchik} give-1.obj-pl ‘He gives us.’ In contrast, -shun pluralizes the object acted upon by a 3rd person subject in the future, as in (38). (38) Quwa/mashun. {qu-wa/ma-shun} give-1.obj-1pl.fut ‘He will give us.’ The 2nd person suffixes are unspecified for number. All object suffixes succeed aspect suffixes; all except the 1st person object pluralizers, -nchik and -shun, precede tense and subordinating suffixes, as well as the nominalizing suffix -na and the conditional suffix -man, as in (39). 11 -shu-nki is unspecified for tense and has both present-tense and futuretense interpretations (e.g. qu-sHQa-yki ‘I will give you’).

(39) a. Quwa/mankiman. {qu-wa/ma-nki-man} give-1.obj-2.sbj-cond ‘You could give me.’ b. Viñac Ñitiruwanmanchik. {ñiti-ru-wa-n-man-chik} crush-urgt-1.obj-3.sbj-cond-1pl ‘It could crush us.’12 The 1st person object plural has an alternative form for the conditional: -wa/ma-chuwan, as seen in (40). (40) Viñac Chukaruwachuwan. {chuka-ru-wa-chuwan} strike-urgt-1.obj-1pl.sbj.cond ‘It can make us sick.’ 3rd person objects are indicated by standard case marking on the 3rd person pronoun pay (accusative -ta or allative/dative -man); see (40). (41) a. Payta qawanchik. {pay-ta qawa-nchik} 3sg-acc see-1pl.sbj ‘We see him/her.’ b. Paykunaman qunki. {pay-kuna-man qu-nki} 3sg-pl-all give-2.sbj ‘You give them.’ In the case of the 1st and 2nd persons, the actor-object suffixes may be reinforced—but not replaced—by accusativeand dative-marked personal pronouns, as in (42). (42) Viñac Ñuqata harquruwara. {ñuqa-ta harqu-ru-wa-ra-ø} 1sg-acc toss-urgt-1.obj-pst-3.sbj ‘He tossed me out.’ Actor–object suffixes are employed both with transitive and ditransitive verbs; see (43). 12 Note that when number is explicitly indexed, it succeeds indexing for object, tense, and subject. SYQ dialects, here, follow the pattern characteristic of Cuzco and Bolivia (obj–tns–sbj–num), not that of the Quechuan languages in the immediate region, Huanca in particular (num–obj–tns– sbj).

155

aviva shimelman (43) a. Mikurushunki. {miku-ru-shunki} eat-urgt-3.sbj.2.obj ‘He’s going to eat you.’ b. Qullqita qusqayki. {qullqi-ta qu-sqayki} gold-acc give-1.sbj.2.obj.fut ‘I’m going to give you money.’ There are no special verbal affixes for reflexive action.

present tense inflection, with the exception that the 3rd person is not indicated with -n, but -ø. In all five dialects, -RQa indicates tense but not aspect and is thus consistent with both perfective and imperfective aspect and can be assigned both simple past and present perfect interpretations (see (45)). (45) a. Viñac Taqtarqani. {taqta-rqa-ni} empty-pst-1.sbj ‘I emptied’ or ‘I have emptied.’

5.6.3.2 Tense SYQ broadly counts three tenses: present, past, and future (contrast maska-nchik ‘we look for,’ maska-RQa-nchik ‘we looked for,’ and maska-shun ‘we will look for’). With the exception of the 1st person plural, person suffixes in SYQ are unmarked for number. -nki and -n correspond to the 2nd and 3rd person singular and plural subject, respectively (e.g. yanapa-nki ‘you (sg/pl) help,’ taki-n ‘s/he/it/they sing(s)’).

b. Cacra Paćhyala. {paćhya-la-ø} explode-pst-3.sbj ‘S/he/it/they exploded’ or ‘S/he/it/they has/have exploded.’

The simple present is, in fact, unmarked for time. It will generally indicate temporally unanchored or habitual action but may also receive past- or future-tense interpretations in different contexts.

Perfective aspect is, rather, indicated by the derivational suffix -Ru. -RQa and -Ru are not in paradigmatic opposition but differ in their distribution. -RQa, but not -Ru, is used in the construction of the habitual past and the past conditional, while -Ru, but not -RQa, may be used in combination with -sHa as well as with -shpa and -pti, in which case it indicates the precedence of the subordinate clause event to the main clause event.

5.6.3.2.2 Future

5.6.3.2.4 Narrative past

5.6.3.2.1 Present

The future tense suffixes in SYQ are 1st person -shaq, 2nd person -nki, 3rd person -nqa, and 1st person plural -shun (e.g. saqma-shaq ‘I will slip,’ pawa-nki ‘you will jump,’ picha-nqa ‘s/he/it/they will sweep,’ tanqa-shun ‘we will push’). Thus, -shaq and -shun are portmanteau, simultaneously indicating person, number, and tense, -nqa indicates person and tense, and -nki person alone.

5.6.3.2.3 Simple past SYQ distinguishes between the simple past, the perfect, and the habitual past. The simple past is indicated by -RQa, as in (44). (44) Taqllarqanchik. {taqlla-RQa-nchik} clap-pst-1pl.sbj ‘We clapped.’ -RQa is realized as -rqa in Apurí–Madeán–Viñac, -ra in Azángaro–Chocos–Huangáscar, Lincha–Tana, and San Pedro–Liscay, and -la in Cacra–Hongos. In all five dialects, person–number inflection in the past tense resembles

156

In SYQ, as in other Quechuan languages, when speakers have only secondhand knowledge of the events they report, they may revert to another past-tense form, -sHQa, often referred to as the “narrative past” because it is used systematically in storytelling. In SYQ, -sHQa—occasionally realized as -shqa in all dialects but generally realized as -sa in Azángaro– Chocos–Huangáscar, Apurí–Madeán–Viñac, and San Pedro– Liscay and as -sha in Cacra–Hongos and Lincha–Tana—is used predominantly in storytelling, historical narrative, and in relating information received from others (46). (46) Azángaro Tanisa. {tani-sa-ø} cease.produce-narr.pst-3.sbj ‘They ceased to produce.’

5.6.3.2.5 Perfect Within the morphological paradigm, -sHa—realized as -sa in Azángaro–Chocos–Huangáscar, Apurí–Madeán–Viñac and San Pedro–Liscay and as -sha in Cacra–Hongos and Lincha– Tana—occupies a slot that seems to be reserved for the

southern yauyos quechua perfect and, indeed, is cognate with morphemes indicating the perfect in other Quechuan languages. That said, almost without exception, the non-nominalizing instances of -sHa in the corpus have more readily available interpretations as narrative pasts and are generally assigned simple past translations by consultants. -sHa is immediately followed by the same person–number suffixes as is the simple past -RQa (i.e. the 3rd person subject is realized as -ø).

5.6.3.2.6 Habitual past The habitual past is indicated by the combination of the agentive noun—formed by the addition of the agentive -q to the verb stem—and the relevant present tense form of ka- ‘to be’ (elided in the 3rd person), as in (47). (47) Takiq kanki. {taki-q ka-nki} sing-ag be-2.sbj ‘You used to sing.’ Object suffixes precede -q, as seen in (48) (48) San Pedro–Liscay Willamaqkanki. {willa-ma-q ka-nki} tell-1.obj-ag be-2.sbj ‘S/he/it/they used to tell me.’

5.6.3.3 Conditionality SYQ verbs inflect for conditionality, present and past. Two different forms indicate the conditional in SYQ. The regular conditional, -man, is attested in all persons, singular and plural, in all dialects. For the 1st person subject, it is the person–number suffixes of the nominal possessive person marking paradigm that are used in combination with -man, i.e. -y and not -ni is used for the 1st person singular in the dialects that are aligned with Northern and Southern Quechua, as in the Viñac example in (49). (49) Viñac Puñuyman. {puñu-y-man} sleep-1.sbj-cond ‘I could sleep.’ -man follows all other inflectional suffixes, as in (50). (50) Takpankiman. {takpa-nki-man} jinx-2.sbj-cond ‘You could jinx it.’

Alternative conditional forms, -chuwan and -waq, are attested in the 1st person plural in all dialects in the 2nd person (both singular and plural) in the Apurí–Madeán–Viñac dialect, respectively; see (51). (51) a. Lincha Pamparuwachuwan. {pampa-ru-wa-chuwan} bury-urgt-1.obj-1pl.sbj.cond ‘They could bury us.’ b. Madeán Alchawaq. {alcha-waq} fix-2.sbj.cond ‘You can fix.’ The past conditional is formed by the addition of ka-RQa—the 3rd person simple past tense form of ka‘to be’—to a present tense conditional form, as seen in (52). (52) Viñac Chawaruwaq karqa. {chawa-ru-waq ka-rqa-ø} milk-urgt-2.sbj.cond be-pst-3.sbj ‘You should have milked.’ The conditional (also called “potential” or “irrealis”) covers more territory than does the conditional in English. It corresponds to the existential and universal ability, circumstantial, deontic, epistemic, and teleological modals of English. SYQ modals are themselves unspecified for force: modal force is determined by context and is generally specified by the evidential modifiers, -ø, -ki, or -iki. Weak modal readings result when the modal is under the scope either of no evidential or of an evidential modified by the evidential modifier -ø; in contrast, progressively stronger universal readings result when the evidential is modified by the evidential modifiers -ik and -iki, see (53) and (54). (53) a. Siqaykunmanćhi. {siqa-yku-n-man-ćhi-ø} fall-excep-3.sbj-cond-evc-ø ‘It might fall.’ b. Siqaykunmanćhiki. {siqa-yku-n-man-ćhi-ki} fall-excep-3.sbj-cond-evc-ki ‘It will most likely fall.’ (54) a. Waytakunkimanmi. {waytaku-nki-man-mi-ø} wear.flowers-3.sbj-cond-evd-ø ‘You should wear flowers.’

157

aviva shimelman b. Istudiyankimanmiki. {istudiya-nki-man-mi-ki} {study-3.sbj-cond-evd-ki} ‘You must study.’ Under the scope of the conjectural evidential, -ćhi, conditionals are generally restricted to epistemic interpretations; under the scope of the direct evidential -mi, they receive all but conjectural interpretations. Ability modals also result from the combination of the infinitive and the verb atipa‘to be able.’ The verbs usHachi- and puydi- (< Spanish poder ‘to be able’), both translated as ‘to be able,’ as well as yaćha- ‘to know,’ may also be employed in this construction.

5.6.3.4 Imperativity The 2nd person and 1st person plural imperative are indicated by -y and -shun, respectively; 3rd person injunctive —the suggestion on the part of the speaker as to the advisability of action by a third party—is indicated by -chun (e.g. Wiksa-ku-y! ‘Get pregnant!’ Wishtu-shun! ‘Let’s limp!’ and Yarqa-chun! ‘Let him starve!’) Prohibitions and negative injunctives are formed by suffixing the inflected form with -chu and preceding it with ama, as in (55). (55) Ama millpuchunchu! {Ama millpu-chun-chu!} proh swallow-3.injunc-neg ‘Don’t let her/him/it/them swallow it!’ The 2nd person future tense, too, has an imperative interpretation. Haku! ‘Let’s go!’ is irregular: it cannot be negated or inflected, except, optionally, with the 1st person plural -nchik (with no change in interpretation).

(57) Pakiyachinki. {paki-ya-chi-nki} break-prog-caus-2.sbj ‘You are making him/her/it/them break [it].’ -ya marks both the progressive and durative components of the continuous. A second morpheme, -chka, which is very rarely employed in any but the San Pedro–Liscay dialect, also receives a durative interpretation. Perfective aspect is generally indicated by -Ru (see Section 5.6.2.1.16).

5.6.3.6 Subordination SYQ counts three subordinating suffixes, -pti, -shpa, and -shtin, and one subordinating structure, -na-poss-kama. In addition, the nominalizing suffixes -na, -q, -sa, and -y form subordinate relative and complement clauses. -pti is employed when the subjects of the main and subordinate clauses are different, as in (58). (58) Apurí–Madeán–Viñac Huk qawaptinqa, ñuqanchik qawanchikchu. Almanchik puriyanshi. {huk qawa-pti-n-qa ñuqa-nchik one see-subds.ant-3.sbj-top 1-1pl qawa-nchik-chu alma-nchik see-1pl.sbj-neg soul-1pl.poss puri-ya-n-shi} walk-prog-3.sbj-evr ‘Although others see them, we don’t see them. Our souls wander about, they say.’

5.6.3.5 Aspect

-shpa and -shtin, in contrast, are employed when the subjects of the two clauses are identical, as in (59) and (60). While it is indeed attested in spontaneous speech, -shtin is rare.

In SYQ, continuous aspect is indicated by -ya. This -ya, however, is probably better analyzed as a derivational morpheme: unlike inflectional morphemes, -ya can appear in subordinate clauses and nominalizations, as seen in (56).

(59) Apurí–Madeán–Viñac Vakata harkanchik puchkashpa millwata. Puchkapuchkashpa. {vaka-ta harka-nchik puchka-shpa cow-acc herd-1pl.sbj spin-subis.simult

(56) a. muytiyaptin {muyti-ya-pti-n} crouch-prog-subds-3.sbj ‘when he is crouching’ b. pakayaq {paka-ya-q} hide-prog-ag ‘one who is hiding’ It can—and, indeed, sometimes must—precede some derivational suffixes, including cislocative -mu and causative -chi, as seen in (57).

158

millwa-ta puchka-puchka-shpa} wool-acc spin-spin-subis.simult ‘We herd the cows while we spin wool. Spinning and spinning.’ (60) Apurí–Madeán–Viñac Awashtin mikuchini wambrayta. {awa-shtin miku-chi-ni wambra-y-ta} spin-subis.simult eat-caus-1.sbj child-1.poss-acc ‘(By) weaving, I feed my children.’ Cacra, but not Hongos, employs -R (realized [l]) in place of -shpa, as in (61).

southern yauyos quechua (61) Cacra Ćhaqnal pushalamun. {ćhaqna-l pusha-la-mu-n} bind.limbs-subis take.along-urgt-cisl-3.sbj ‘Binding his hands and feet, they took him along.’ -pti generally indicates that the event of the subordinated clause began prior to that of the main clause, as in (62), while -shpa and -shtin generally indicate that the event of the subordinated clause is simultaneous with that of the main clause, as in (63). (62) Apurí–Madeán–Viñac Urkista-qa ćhayamuptin tushurqanchik. {urkista-qa ćhaya-mu-pti-n band-top arrive-cisl-subds.ant-3.sbj tushu-rqa-nchik} dance-pst-1pl.sbj ‘When the band arrived, we danced.’ (63) Apurí–Madeán–Viñac Sapuqa kurrkurryashpa kurriyan. {Sapu-qa kurrkurrya-shpa kurri-ya-n} frog-top croak-subis.simult run-prog-3.sbj ‘The frog is running, going kurr-kurr!’ Depending on the context, both -pti and -shpa can be translated by ‘when,’ ‘if,’ ‘because,’ ‘although,’ or with a gerund. Negated, -shpa can be translated by ‘without,’ ‘although,’ or ‘despite.’ -shtin subordinates are adverbial and can generally be translated by ‘while’ or with a gerund. The three subordinating suffixes also differ in their inflectional patterns. -pti subordinates are inflected with possessive suffixes (e.g. tarpu-pti-nchik ‘when we plant’); in contrast, -shpa and -shtin subordinates do not inflect. Regardless, subordinate clause verbs are never suffixed with any other inflectional morphemes, with the exception of -ya. Rather, subordinate clause verbs inherit tense, aspect, and conditionality specification from the main clause verb, as Example (64) shows. (64) Apurí–Madeán–Viñac Rishpa qawayman karqa. {ri-shpa qawa-y-man ka-rqa-ø} go-subis.simult see-inf-cond be-pst-3.sbj ‘If I would have gone, I would have seen.’ In combination with the nominalizer -na and inflection for person with the nominal possessive paradigm, -paq forms subordinate clauses, indicating that the event referred to is either simultaneous with or limits the event referred to in the main clause, as in (65).

(65) Apurí–Madeán–Viñac Mikumuwananpaq kutimushpa ćhaqnaruwan. {miku-mu-wa-na-n-paq eat-cisl-1.obj-nmlz-3.sbj-purp kuti-mu-shpa return-cisl-subis.simult ćhaqna-ru-wa-n} bind.limbs-urgt-1.obj-3.sbj ‘In order to be able to eat me when he got back, he tied me up.’

5.6.3.7 Order of affixes Inflectional suffixes (is) follow derivational suffixes (ds), if any are present; derivational suffixes attach to the verb stem (vs). Thus, a SYQ verb is built: vs–(da)–ia The order of inflectional affixes admits little exception: the object suffixes precede tense and subordinating suffixes, as well as the nominalizing suffix -na; (66) furnishes an example. (66) Quwa/mananpaq. {qu-wa/ma-na-n-paq} give-1.obj-nmlz-3.sbj-purp ‘So s/he/they give(s) me.’

5.7 Clitics with phrasal or clausal scope SYQ, like other Quechuan languages, counts a number of enclitics that may attach to both nouns and verbs as well as to adverbs and negators. Enclitics always follow all inflectional suffixes, verbal and nominal. SYQ counts 15 enclitics. These can be divided into two classes: (i) those which position the utterance with regard to others salient in the discourse (-chu ‘interrogative/disjunctive/negative,’ -lla ‘restrictivelimitative,’ -ña ‘discontinuative,’ -pis ‘additive,’ -qa ‘topic,’ -Raq ‘continuative,’ and -taq ‘sequential’); and (ii) those that position the speaker with regard to the utterance (-aRi ‘assertive,’ -puni ‘certainty,’ -Yá ‘emphatic,’ and the evidentials -mi, -shi, and -ćhi along with their modifiers -ik and -iki).

5.7.1 Clitics that position the utterance 5.7.1.1 -chu ‘interrogative (q)’ -chu forms absolute and disjunctive questions; see (67). (67) a. Waqankichu? {waqa-nki-chu} cry-2.sbj-q ‘Do you cry?’

159

aviva shimelman b. QaRichu waRmichu? {qaRi-chu waRmi-chu} man-q woman-q ‘A man or a woman?’ On -chu and negation, see Section 5.8.6.

5.7.1.2 -lla ‘restrictive-limitative (lim)’ -lla generally indicates exclusivity or limitation in number. It is generally translated as ‘just’ or ‘only’ (e.g. uma-lla ‘just the head,’ from uma ‘head’). -lla may express an affectionate or familiar attitude.

5.7.1.3 -ña ‘discontinuative (disc)’ -ña indicates transition or change of state or quality. In affirmative statements, it is generally translated as ‘already’ (e.g. kundinadaw-ña ‘already condemned,’ from kundinadaw ‘condemned’); in negative statements, it is translated as ‘no more’ or ‘no longer,’ as in (68). (68) San Pedro–Liscay Unayćhik. Kananqa kanñachu imapis. {unay-ćhi-k kanan-qa ka-n-ña-chu before-evc-ik now-top be-3.sbj-disc-neg ima-pis} what-add ‘That would be a long time ago. Now, there isn’t anything anymore.’ In questions, it is translated as ‘yet,’ as in (69). (69) Pasarunñachu? {pasa-ru-n-ña-chu} pass-urgt-3.sbj-disc-q ‘Did they go by yet?’

5.7.1.4 -pis ‘additive (add)’ -pis indicates the inclusion of an item or event into a series of similar items or events. It is generally translated as ‘too’ or ‘also’ (e.g. pay-pis ‘s/he/it, too,’ from pay ‘s/he/it’) or, when negated, ‘neither, not … either,’ as in (70). Attaching to interrogative-indefinite stems, it forms indefinites and, with mana, negative indefinites (e.g. mana ima-pis ‘nothing,’ from ima ‘what’). (70) Apurí–Madeán–Viñac Paykunaqa manam qawarqapischu. {pay-kuna-qa mana-mi qawa-wa-rqa-ø-pis-chu} 3-pl-top no-evd see-1.obj-pst-3.sbj-add-neg ‘They didn’t see us either.’

160

5.7.1.5 -qa ‘topic marker (top)’ -qa indicates the topic of the clause. It is generally left untranslated, as in (70) immediately above.

5.7.1.6 -Raq ‘continuative (cont)’ -raq indicates continuity of action, state, or quality. It is generally translated as ‘still’ or, negated, ‘yet’ (e.g. taqsa-naraq ‘still having to be cleaned,’ from taqsa-na ‘cleaning’ or mana-raq ‘not yet,’ from mana ‘no’).

5.7.1.7 -taq ‘sequential (seq)’ -taq indicates the sequence of events. In this capacity, it is generally translated as ‘then’ or ‘so’ (e.g. yapa-taq ‘then once more,’ from yapa ‘again’). In questions introduced by an interrogative (e.g. pi- ‘who,’ ima- ‘what’), -taq may suffix to the interrogative phrase (e.g. ima-taq ‘what?’). It is unmarked in this capacity, contributing no valence to the interpretation.

5.7.2 Clitics that position the speaker 5.7.2.1 -Yá ‘emphatic (emph)’ -Yá indicates emphasis (e.g. mana-yá ‘no indeed, no never,’ from mana ‘no’). It takes on an allomorph -á after the evidential enclitics. As in (71), it generally has no reflex in translation, unlike the evidential modifiers (see Section 5.7.2.7). (71) Azángaro–Chocos–Huangáscar Chay wañuruptikiqa, ¿pimá qawashunki? ¿Yasqayaruptikiqa? {Chay wañu-ru-pti-ki-qa dem die-urgt-subds.ant-2.sbj-top pi-mi-Yá qawa-shunki who-evd-emph see-3.sbj.2.obj.fut yasqa-ya-ru-pti-ki-qa} old.woman-inch-urgt-subds.ant-2.sbj-top ‘When you die, who’s going to see to you? Or when you get old?’

5.7.2.2 -puni ‘certainty, precision (cert)’ -puni is generally translated as ‘necessarily,’ ‘definitely,’ ‘precisely,’ as in (72). -puni is attested only in the dialects of SYQ that are aligned with Northern and Southern Quechua, where it is infrequently employed.

southern yauyos quechua (72) Apurí–Madeán–Viñac Chay wiqawninchikmanpuni chiri yakuta ćhuranchik. {chay wiqaw-ni-nchik-man-puni chiri dem waist-ep-1pl.poss-all-cert cold yaku-ta ćhura-nchik} water-acc put-1pl.sbj ‘We put cold water right on our lower backs.’

5.7.2.3 -mi ‘evidential–direct experience (evd)’ -mi indicates that the speaker has personal evidence for the proposition under the scope of the evidential. In addition to indicating the speaker’s information type, evidentials also function to indicate focus or comment. Contrast the examples in (73). (73) a. Llamaykikunatash wañuchinqa. {llama-yki-kuna-ta-shi wañu-chi-nqa} llama-2.poss-pl-acc-evr die-caus-3.sbj.fut ‘S/he/it/they will butcher your llamas, they say.’

5.7.2.5 -ćhi ‘evidential – conjectural (evc)’ -ćhi indicates that the speaker is making a conjecture as to the proposition under the scope of the evidential. -ćhi is agnostic as to the evidence type of the propositions from which that conjecture is calculated, i.e. as to whether they are direct or not. -ćhi is generally translated in Spanish as seguro ‘for sure.’ Note that in Andean Spanish, this seguro does not reflect speaker certainty but, like the -ćhi it translates, simply indicates that the proposition in its scope is a conjecture. See (75). (75) Apurí–Madeán–Viñac Qiñwalman ćhayarachiptiki wañukunmanćhi. {qiñwal13 -man queñua.grove-all ćhaya-ra-chi-pti-ki arrive-urgt-caus-subds.ant-2.sbj wañu-ku-n-man-ćhi} die-refl-3.sbj-cond-evc ‘If you make her go all the way to the quingual grove, she might die.’

b. Llamaykikunata wañuchinqash. {llama-yki-kuna-ta-sh wañu-chi-nqa-shi} llama-2.poss-pl-acc die-caus-3.sbj.fut-evr ‘S/he/it/they will butcher your llamas, they say.’ In (73a), the focus is on the direct object, ‘llamas’; in (73b), it is the verb, ‘to butcher,’ that is marked as the focus. The evidentials are commonly employed as complete non-verbal copular predicates (e.g. kuchi-hina-m ‘It’s like a pig,’ from kuchi-hina ‘like a pig’). The evidentials are realized without their vowel nucleus (i.e. as -m, -sh, and -ćh), when they follow a vowel, as in the running-text examples in this section.

5.7.2.4 -shi ‘evidential–reportative/quotative (evr)’ -shi indicates that the speaker has non-personal-experience evidence for the proposition under the scope of the evidential. -shi appears systematically in stories. It is often translated as ‘they say,’ as in (74). (74) San Pedro–Liscay Radyukunapa rimayta rimayan. Lluqsiyamunshi. Tirrurista rikariyamunshi. {radyu-kuna-pa rima-y-ta rima-ya-n radio-pl-loc talk-inf-acc talk-prog-3.sbj lluqsi-ya-mu-n-shi go.out-prog-cisl-3.sbj-evr

tirrurista terrorist

rikari-ya-mu-n-shi appear-prog-cisl-3.sbj-evr} ‘On the radio they talk for the sake of talking. They are coming out, they say. Terrorists are appearing, they say.’

5.7.2.6 -aRi ‘assertive force (ass)’ -aRi indicates conviction on the part of the speaker. It is generally translated as ‘certainly’ or ‘of course,’ as in (76). (76) Qillakuyankićhari. {qilla-ku-ya-nki-ćhi-ari} lazy-refl-prog-2.sbj-evc-assr ‘You really must be getting lazy.’ Following -yá, the allomorph -Ri is employed (e.g. -ya-Ri).

5.7.2.7 -ik and -iki ‘evidential modifiers (ev.mod.2, ev.mod.3)’ SYQ is unusual in that each of its three evidentials counts three variants, formed by the suffixation of -ø, -ik, or -iki. The resulting nine forms are direct -mi-ø, -m-ik, and-m-iki, reportative/quotative -shi-ø, -sh-ik, and -sh-iki, and conjectural -ćhi-ø, -ćh-ik, and -ćh-iki. They indicate degrees of increasing evidence strength or conjectural certainty for -mi and -ćhi, respectively. In the case of universal-deontic-modal and future-tense verbs, -ik and -iki indicate increasingly strong obligation and increasingly imminent/certain futures, respectively. -ik and -iki are generally translated in Spanish as pues and seguro, respectively. (77) gives an example with -ćh-iki. 13

A small tree (Polylepis sp.).

161

aviva shimelman (77) Apurí–Madeán–Viñac Kabrapis kasusam. Riqsiyanćhiki runantaqa. {kabra-pis kasu-sa-ø-mi. goat-add obey-prf-3.sbj-evd riqsi-ya-n-ćhi-iki know-prog-3.sbj-evc-ev.mod.3 runa-n-ta-qa} person-3.poss-acc-top ‘The goats obey her. They must recognize [her as] their master, for sure.’

5.8.2 Constituent order Modifiers generally precede the elements they modify: adjectives precede the nouns they modify (79a), possessors precede the thing possessed (79b), relative clauses precede their heads (79c), and adverbs precede the elements they modify. (79) a. yuraq wayta {yuraq wayta} white flower ‘white flower’

5.7.3 Combinations The pair [-Raq, -ña] is mutually exclusive, as are the pairs [-ari, -iki] and [-Yá, -iki]; the evidentials are mutually exclusive with each other as well as with -qa. Combinations of individual enclitics generally occur in the order: -lla, -puni, -pis, [-ña, -Raq], -taq, -chu, [-qa, -mi, -shi, -ćhi], [-ik, -iki, -Yá], -aRi.

5.8 Basic syntax

b. paypa pupun {pay-pa pupu-n} s/he/it-gen navel-3.poss ‘her navel’ c. trabahasayki wasipa {trabaha-sa-yki wasi-pa} work-prf-2.poss house-loc ‘in the house where you worked’ In case an NP includes multiple modifiers, these appear in the order: demonstrative–quantifier–numeral–negation– pre-adjective–adjective–attribute–nucleus. Of course, not all combinations are possible.

5.8.1 Grammatical relations The unmarked constituent order in SYQ, as in other Quechuan languages, is SOV. That said, because constituents are obligatorily marked for case, a change in the order of constituents in an utterance will not necessarily change the sense of that utterance or change the interpretation of topic or focus, which are generally marked with enclitics. Nevertheless, the verb and the object cannot commute in subordinate clauses, where only the order OV is grammatical, as in (78). (78) Apurí–Madeán–Viñac Pampakuni frutachaykuna apasayta. {pampa-ku-ni fruta-cha-y-kuna bury-refl-1.sbj fruit-dim-1.poss-pl apa-sa-y-ta} bring-prf-1.sbj-acc ‘I bury the fruit I bring.’

162

5.8.3 Major clause types Major clause types in SYQ include declaratives, interrogatives, and exclamatives. On the formation of declaratives, see Section 5.7.2.3; on interrogatives, see Section 5.5.1.1.4. Exclamatives are described here. All spontaneously attested Indigenous exclamations share a common pattern: they begin with /a/ and end in /w/ (e.g. ¡Alaláw! ‘How cold!’ and ¡Ananáw! ‘Ouch!’), or, less commonly, in /k/ or /y/ (e.g. ¡Achachák! ‘What a fright!’). With the exception of the final /w/, they feature almost exclusively the alveolar and palatal consonants /ch, ll, l, n, ñ, t, y/ (which accounts for the entire catalog of native SYQ alveolars, post-alveolars, and palatals, with the exception of the voiceless fricatives /s, sh/ and the tap /r/). Exclamatives include no vowels except /a/, they bear stress on the final syllable, and syllable repetition is the norm (e.g. ¡Achachalláw! ‘How awful!’).

southern yauyos quechua

5.8.4 Coordination The enclitics -pis, -taq, and -Raq can all be used to coordinate NPs, AdvPs and VPs. The case suffix -wan can be used to coordinate the first two of these, as in (80), where the speaker is speculating about the cheese of a neighbor. (80) Apurí–Madeán–Viñac Kabranpaqwan vakanpaqwanćhi. {kabra-n-paq-wan vaka-n-paq-wan-ćhi} goat-3.poss-gen-com cow-3.poss-gen-com-evc ‘It’s from her goats’ [milk] and from her cows’ [milk], probably.’ -pis, -taq, and -Raq generally imply relations of inclusion, contrast, and contradiction, respectively. Thus, -pis ‘inclusion’ can generally be translated as and or also (e.g. tullu-pis tunshu-pis ‘bone and marrow’). -taq and -Raq ‘contrast’ and ‘contradiction’ can both be translated as but, while, whereas, and so on, as in (81). For the use of -wan as a coordinator, see Section 5.5.3.3.8. (81) Azángaro–Chocos–Huangáscar Wawanchikta idukanchik. Qillakunaqa manataqmi. {wawa-nchik-ta iduka-nchik qilla-kuna-qa baby-1pl.poss-acc educate-1pl.sbj lazy-pl-top mana-taq-mi} no-seq-evd ‘We educate our children, whereas the lazy ones don’t.’ Additional strategies employed for coordination in SYQ include (i) the employment of the Indigenous coordinating particle icha ‘or,’ which attaches to each of the coordinated elements, or any of the borrowed coordinators i ‘and,’ u ‘or,’ piru ‘but,’ or ni ‘nor’ (< Spanish y, o, pero, and ni respectively) and (ii) simple juxtaposition, i.e. the placement of coordinated elements in sequence.

5.8.5 Clausal complementation Syntactically, clausal complements are NPs, derived by the four deverbal nominalizing suffixes, concretizing -na, agentive -q, resultative -sHa, and infinitive -y (see Section 5.4.2.1). These four suffixes form subjunctive, purposive, indicative, and infinitive complement clauses, respectively. Nominalizing suffixes attach directly to the verb stem, with the exception that the 1st and 2nd person object suffixes -wa ~ -ma and -sHu intercede if present, as in Example (82).

(82) Cacra–Hongos inyiktamananchikpaq {inyikta-ma-na-nchik-paq} inject-1.obj-nom-1pl.sbj-purp ‘in order to inject us’ As these structures are formally nouns, they are inflected with substantive suffixes (e.g. tinku-sHa-yki ‘that you found’).

5.8.5.1 -na ‘nominalizer (nmlz)’ -na followed by a possessive person suffix indicates necessity (i.e. it forms universal deontic/teleological modals), as in (83). (83) Taqsanayki. {taqsa-na-yki} wash-nmlz-2.poss ‘You have to wash.’ Tense and aspect may be specified by the copula auxiliary inflected for 3rd person, as can be seen in (84). (84) Pallanayki karqa. {palla-na-yki ka-RQa-ø} pick-nmlz-2.poss be-pst-3.sbj ‘You had to pick.’

5.8.5.2 -q ‘agentive (ag)’ -q complement clauses, with verbs of movement, indicate the purpose of the displacement, as in (85). (85) Lincha Tana. Ullukuq hamunqa. {ullu-ku-q hamu-nqa} oil-refl-ag come-3.sbj.fut ‘He will come to baptize.’

5.8.5.3 -sHa ‘resultative (res)’ -sHa is perfective, forming indicative complement clauses. An example is given in (86). (86) Lincha–Tana Atipashaytaćhik ruwakushaq. {atipa-sha-y-ta-ćhi-k be.able-res-1.poss-acc-evc-ik

ruwa-ku-shaq} make-refl-1.sbj.fut

‘I’ll do what I can.’

5.8.5.4 -y ‘infinitive (inf)’ -y forms infinitive complement clauses, as in (87).

163

aviva shimelman (87) Lincha–Tana Tikayta qallakun. {tika-y-ta qalla-ku-n} coagulate-inf-acc begin-refl-3.sb ‘It started to coagulate.’ The latter are particularly common with the auxiliary verbs muna- ‘to want,’ atipa- ‘to be able,’ and yaćha- ‘to know.’ See (88). (88) Tinkayta yaćhanki. {tinka-y-ta yaćha-nki} spin-inf-acc know-2.sbj

-chu is also licensed by the enclitic additive suffix -pis and the borrowed particle -ni (from Spanish ni ‘nor’). In prohibitions, -chu co-occurs with the particle ama, which is translated as ‘don’t.’ See (94). (94) Ama wañuchiychu! {ama wañu-chi-y-chu} proh die-caus-imp-neg ‘Don’t kill [him]!’ -chu does not appear in subordinate clauses where negation is indicated with a negative particle alone.

‘You know how to spin [wool].’

5.9 Lexical and semantic properties

5.8.6 Relativization Relative clauses are formed with the four deverbal nominalizing suffixes: -na ‘nominalizer’ (89), -q ‘agentive’ (90), -sHa ‘resultative’ (91), and -y ‘infinitive’ (92). See also Section 5.4.2.1. Any substantive constituent—subject, object, or oblique—can be relativized. (89) wañukunayki punćhaw {wañu-ku-na-yki die-refl-nmlz-2.poss

punćhaw} day

‘the day that you die’ (90) chinkakuq pashña {chinka-ku-q pashña} lose-refl-ag girl ‘the girl who was lost’ (91) mayman yaykushayki {may-man yayku-sHa-yki} where-all enter-prf-2.poss ‘wherever you go’ (92) yanukuy tardi {yanu-ku-y tardi} cook-refl-inf afternoon ‘the afternoon that we cook’

5.8.7 Negation In SYQ, negation of declaratives is generally indicated by the enclitic -chu in combination with the particle mana, as in (93). (93) Mana suyankichu. {mana suya-nki-chu} no wait-2.sbj-neg ‘You don’t wait.’

164

5.9.1 Lexical differences among the SYQ dialects As explained in Section 5.2, Yauyos is located on the border of the two large, contiguous zones where languages belonging to the two major branches of the Quechuan language family are spoken: Central Quechua to the north, and Southern Quechua to the south. Grammatically, we have seen Apurí–Madeán–Viñac and Lincha–Tana sort with Northern and Southern Quechua, while Azángaro–Chocos– Huangáscar, Cacra–Hongos, and San Pedro–Liscay sort with Central Quechua. Lexically, however, the dialects cleave along different lines, lines defined not by grammar but by geography: the two northern SYQ dialects—the Central Quechua-like Cacra– Hongos and the Southern Quechua-like Lincha–Tana—sort together, while the three southern SYQ dialects—the Central Quechua-like Azángaro–Chocos–Huangáscar and San Pedro–Liscay and the Southern Quechua-like MadeánViñac—sort together. Below, I detail a dialectical analysis of the lexicon, which, at the time, counted 2,462 SYQ entries.14 After eliminating derived15 and non-native terms, there remained 1,940 lexical items in the set. In the vast majority of cases, the dialects feature cognate forms to denote a given referent. I could identify only 37 instances in which they employed non-cognate forms. In 32 of these instances, the dialects cleaved along north–south lines 14 The .xml database as well as 4 files formatted as EnglishYauyos/Yauyos-English and Spanish-Yauyos/Yauyos-Spanish dictionaries are available from the Archive of the Indigenous Languages of Latin America at the University of Texas, Austin, (https://www.ailla.utexas.org/ islandora/object/ailla%3A242751), as well as from the Language Archive of the Max Planck Institute for Psycholinguistics (https://archive.mpi. nl/tla/). 15 The lexicon includes both roots and derived terms. Thus, both the pairs mv, ah, sp sumaq and ch, lt tuki ‘pretty’ and amv, ach, sp sumaq-lla and ch, lt tuki-lla ‘nicely’ appear in the corpus. Only the root pair, however, was included in the analysis. “Multiples” such as these numbered 116. Other examples include: amv, ach, sp qawa- ∼ ch, lt rika- ‘to see’ → qawa-chi- ∼ rika-chi- ‘to show; to make an offering,’ mv, ah, sp, lt qishta ∼ ch, lt tunta ‘nest’ → qishta-cha- ∼ tunta-cha- ‘to build a nest.’

southern yauyos quechua (see examples in Table 5.10). In four of the five remaining instances, San Pedro–Liscay supplied the outstanding form. In 22 of the 28 cases for which cognate forms for the same referents could be identified in Junín-Huanca Quechua and Ayacucho Quechua (the former a variety of Central Quechua spoken immediately to the north of Yauyos; the latter, a variety of Southern Quechua spoken very nearby, to the south), the northern SYQ dialects employed a form cognate with the Junín-Huanca form, while the southern SYQ dialects employed a form cognate with that of Ayacucho (e.g. kasha, kichka ‘thorn,’ pukutay, puyu ‘cloud, fog’). Table 5.10 Examples of non-cognate lexical forms with a northern SYQ vs southern SYQ dialectal distribution Gloss

SYQ South Ayacucho SYQ North Junín-Huanca

‘cloud, fog’

puyu

puyu

pukutay

pukatay

‘nest’

qisha ~ qishta

qisha

tunta

tunta

‘thick’

rakta

rakta

tita

tita

‘sash’

chumpi

waćhakuq

‘to wean’

anu-

wasqi-

‘pretty’

sumaq

tuki

This does not mean that the SYQ dialects employed identical forms for all remaining terms. Of these, the phonological forms in all five dialects agreed perfectly for 1,081 items (e.g. sapi ‘root,’ ayqi- ‘to escape,’ and chaqchu- ‘sprinkle, scatter’). The remaining 859 items are accounted for as follows. In 154 cases the item had an identical form in all dialects where any native form for that referent was attested (e.g. the forms chaskay ‘morning star’ and uya ‘face’ are attested in Apurí–Madeán–Viñac, Azángaro–Chocos–Huangáscar, and San Pedro–Liscay, but Cacra–Hongos and Lincha–Tana use the loan words lusiru and kara ( 2 > 3, -ch(a) was suffixed to 1st or 2nd person subject pronouns in transitive constructions in which an agent acts upon a patient or a goal one step down in the ph. The marker ni-ch(a) refers to a transition between a 1st person subject and a 2nd person object, and pi-ch(a) refers to a transition between a 2nd person subject and a 3rd person object. The first line of the Rituale presents an example of this “downstep” phenomenon, in which the -ch marker on the 1st person pronoun indicates that the verb expresses a situation in which a 1st person subject acts on a 2nd person object, even though the 2nd person is not overtly expressed. This is shown in (30). (30) {ni-ch baptiza-ke-nch iki-m chu 1sg-hm baptize-1.sbj-decl father-com 3.poss sku-m spiritu sancto-m men-ut} son.of.man-com Spirit Holy-com name-com ‘I baptize you in the name of the Father, and his Son, and the Holy Spirit.’ (A.01) There are only 15 instances of the 1st person pronoun ni in the whole Rituale. Of these, seven present an instance of the relevant main tense transitive interaction involving a 1st person acting on a 2nd person.19 The fact that we also find cases of such transitions in which the personal pronoun does

{enahamp patero no hucha ore-ke-Ø-nch perhaps priest 1.poss sin tell-fut-3.sbj-decl

a-tawa} say-subis

‘…saying, “Perhaps the priest will tell my sins.”’ (K.13) This is a main-tense direct-quote clause in which, according to Torero’s analysis, the agent subject patero ‘the priest’ should have been marked with an ergative marker. However, it is not. 19 We find twice, twice, and and a single time each. In most of these cases we find an affirmative marker on the pronoun in the corresponding Quechua text; that would also be a plausible alternative interpretation for this suffix.

354

not occur at all shows that realization of ni-ch was not obligatory, in line with our claim in Section 12.6.3.1 that Puquina was a pro-drop language. Of the 39 instances of the pronoun pi ‘you’ in the Rituale, there is just one case in which we find pi-ch(a) in a main tense downstep environment in which a 2nd person acts on a 3rd person. Meanwhile, pi is realized in two instances of this main tense downstep environment. There are three other cases of pi-ch(a) as the subject of a nominalized clause. In 13 cases, pi is realized as the subject of an intransitive or reflexive verb, and in the other 20 cases it is realized as an object or an adjunct.

12.5.3.2.3 -wa ‘genitive (gen)’ The most common means for forming possessive noun phrases in Puquina was by juxtaposing the specifier and the head (e.g. Torero 2002: 143), as in Dios isu in (31). Such constructions cannot be distinguished from nominal compounds, so it is impossible to say whether this would be best translated as ‘God’s house’ or simply as ‘God house.’ (31) {suwa-pi-i iglesia-wanana dios isu-nana} steal-2.sbj-q church-abl God house-abl ‘… did you steal from the church, from God[’s] house?’ (L07.02) However, there is a small number of examples in which we find an element on the specifier that resembles a genitive case marker. The clearest such case is the morpheme -wa. It may be that there is a conceptual relationship between this suffix and the benefactive case marker -wa described in Section 12.5.3.2.9. An example is given in (32). (32) {Dios iki win atipa-eno-wa kupi God father all be.able-ag-gen right moha-na e-kuma-Ø-nch} hand-loc rev-stay-3.sbj-decl ‘He sits again at the right side of God, the almighty Father.’ (G.08) However, since we find two nearly identical versions of this sentence without a genitive marker (T.05, V.17), it appears that the realization of this suffix was optional. Two other markers are more uncertain with respect to their status as genitive morphemes. The first is , which occurs three times in the Rituale. Two of these instances are in the identical texts B (the baptism of boys) and C (the baptism of girls), so those really only represent a single attestation. This text is reproduced in (33).

puquina (33) {kiñ hata-Ø-i Dios-n Iglesia-wanana-k} what want-3.sbj-q God-gen church-abl-acc ‘What does s/he want from God’s church?’ [B.03/C.03] The other case, in (93), marks the subject of a nominalized clause, so it represents a rather different usage. This putative genitive suffix was first proposed by de la Grasserie (1894: 5). His analysis was adopted by Szemiński (1990: 384) in an examination of an Inca song (see discussion in Cerrón-Palomino 1998a; Torero 2002: 143). However, given that there are only three instances of this suffix in the Rituale, two of which are in identical texts—and since it is also interpretable as the Aymara genitive suffix -na—we find the evidence insufficient to consider it a Puquina genitive morpheme. The final element that appears in contexts where one expects a genitive marker is the suffix in (34) (note that -s had other functions in Puquina, including the inversive, to which we return in Section 12.6.3.1). (34) {Jesu Christo-s suma-eno oxa-ke-Ø-nch Jesus Christ-gen live-ag eat-fut-3.sbj-decl chu kam oqo-ke-Ø-nch} 3.poss blood drink-fut-3.sbj-decl ‘He will eat Christ’s living [body] and drink his blood.’20 (E.12) This sentence is a good example of the problems one encounters in analyzing the Rituale. In the Spanish, Quechua, and Aymara versions of this sentence, the word for ‘body’ is expressed and modified with the equivalent of ‘living.’ In Puquina, by contrast, the word for ‘body’ is absent, and we only find suma-eno ‘the living thing.’ So, on the basis of the occurrence of ‘body’ in the parallel texts it appears that this example is an isolated, but clear case of -s on the specifier of a noun. To summarize, possessive noun phrases in Puquina were usually formed by juxtaposing the specifier and the head. However, a morpheme -wa also seems to serve a genitive function. Evidence for genitive suffixes and is weaker.

12.5.3.2.4 -na and -(u)t ‘locative (loc)’ Puquina had two locative markers, -na and -(u)t. We identified a total of 71 cases of locative marking with these two 20 The only other possible, but clearly unintended, reading of this sentence would be ‘Jesus Christ will eat the living thing and will drink his blood.’

suffixes. Of these, 32 are instances of -(u)t, which are suffixed to 20 nouns. Examples of this suffix are found in (41), (66), (79), and (81). The suffix normally appears as after consonants and as after vowels. The remaining 39 instances of -na are suffixed to 19 nouns. This suffix is identical to its Aymara counterpart (see (32) for an illustration of the use of this suffix). Notably, among the 71 tokens of locative marking, there is no overlap among the nouns that take each marker, i.e. nouns predictably take one or the other locative suffix. The sole exception is the Spanish loan cruz ‘cross,’ which occurs as once and as three times. There does not appear to be a morphophonological explanation for this distinction, and the only pattern we could discern is that the nouns that select -(u)t are often religious in nature, while the nouns that select -na tend to refer to other types of concepts.

12.5.3.2.5 -guta ‘allative (all)’ The Puquina allative suffix -guta has a clear directional interpretation with verbs of motion. This function can be seen in (7) and (88), in which it refers to motion away from a deictic center. It also has a more metaphorical directional interpretation on complements of verbs like confesa- ‘to confess,’ kuha a- (or ti a-) ‘to say yes to, believe in,’ and holla- ‘to speak.’ With such usages, the variant -gutak may be used. It is likely that -gutak is in fact a combination of -guta and the accusative marker -k.

12.5.3.2.6 -kama and -wen(o) ‘limitative (lim)’ Puquina shared the limitative suffix -kama with Quechuan and Aymaran, but it also had its own limitative marker -wen(o), which is variously spelled and in the Rituale. These suffixes referred to a limit in both space and time (‘until’). (35) includes both -wen(o) and -kama. This is another case in which Oré provides equivalent lexemes (and in this case a suffix) from both Quechuan and Puquina. (35) {missa-k qalla-so titu-ch ho-no-weno mass-acc begin-ptcp moment-abl end-inf-lim puchuka-s-kama kata-lla-eno-pi-i} complete-inf-lim hear-?-ag-2.sbj-q ‘Did you listen from the moment he began mass, until the end, until the end?’ (L03.01) Notably, the Quechuan/Aymaran suffix -kama is occasionally voiced in the Rituale (i.e. ), especially after Puquina nominalizing suffixes -eno and -no. This suggests that the voicing phenomena described in Section 12.3.4.3 may also extend to -kama.

355

emlen, mossel, van de kerke, adelaar 12.5.3.2.7 -ch ‘ablative (abl)’ The ablative case marker -ch in Puquina expressed the concepts ‘from,’ ‘out’ in space, ‘after’ in time as in (35), ‘about’ a topic as in (51), the notion ‘reason’ (especially on nominalized clauses) as in (36) and (79), and others, all of which are similar to the uses of Southern Quechua -manta and Aymara -ta ~ -tha. A further example of the ‘reason’ sense of ablative -ch is given in (36). (36) {po hucha-wichna pi Dios-guta so 2.poss sin-csl you God-all two se'e ascha-so-ch huntu kichu-ke-pi-anch} heart be-ptcp-abl very be.sad-fut-2.sbj-decl ‘You must be very sad because of your sins and because of the fact that you have been with two hearts toward God.’ (K.05) Another ablative-like marker -(wa)nana(k) occurs less frequently in the Rituale, as in (31).

12.5.3.2.8 -m ‘comitative-instrumental, associative (com)’ The comitative-instrumental suffix -m functioned as a coordinator of nouns or nominal phrases (note that it was also commonly represented as ). Its instrumental function can be seen in (37), and its role in coordinating nouns or nominal phrases is illustrated in (30) above. (37) {Iesu Christo señ pip hiski-m apa Jesus Christ our flesh eye-com not koha-n-ki a-su-hamp} see-plv-1.sbj say-ptcp-add ‘Although we do not see Jesus Christ with our “bodily” eyes …’ (E.16)

(38) {señ anima-kuna-wa qollana kurinba 1pl.poss soul-pl-ben superior food iti-n-ki-nch receive-plv-1.sbj-decl} ‘We receive the mighty food for our souls.’ (E.12) On infinitival complements, -wa marks the purpose for which an action is performed, as we discuss in Section 12.8.5.3.

12.5.3.2.10 -wichna ‘causal (csl)’ There was also a Puquina case marker -wichna, which expressed the notion of ‘reason’ or ‘for the sake of,’ in the manner of Quechuan -rayku and Aymaran -layku or -supa. Examples are given in (36) and (83). Given that the consonant sequence is not found elsewhere in Puquina morphemes, it is possible that -wichna is a fossilized complex form containing the locative or genitive -na and an unknown form -wich.

12.5.3.2.11 -ye ‘feminine vocative (voc.f)’, -re ‘masculine vocative (voc.m)’ A special property of Puquina is that it had a vocative marker with distinct variants for male and female addressees. In the passage of the Rituale shown in (39), priests are instructed to address female parishioners with a form ending in -ye, and men with a form ending in -re. (39) {koha-ta haya-re (ſi ſuere muger) haya-ye see-imp child-voc.m (if it’s a woman) child-voc.f kata-lla-n-s-pi-ns} hear-?-plv-inv-2.sbj-decl ‘Look, child (m), (if it’s a woman) child (f), you have been heard [by God].’ (N.01) There was apparently another vocative marker -s, which was more limited in use. This is illustrated in (54) below.

12.5.3.2.9 -wa ‘benefactive, purposive (ben, purp)’

12.5.3.2.12 Other case-like markers

On nouns, the benefactive marker -wa marks the person for whom an action is performed, as in (38).21 Another case is given in (27).

There are some other case-like Puquina suffixes that bear mentioning in this section. For example, we do not consider -pura ‘among’ nor -gi’in ‘like’ as true case markers. However, since they appear in the same slot, it is relevant to discuss them here. -pura occurs with the same form and function in the Quechuan languages and in Aymara. We find -pura mainly in reflexive and reciprocal contexts, as in (40).

21 Nominal or pronominal benefactive arguments in Puquina were rare because it was a hierarchical inverse language, as we show in Section 12.6.3.1. 1st or 2nd person arguments were obligatorily realized as subjects of the phrase in which they occurred.

356

puquina Table 12.9 Verbal slot template Reversive

Verb root

Derivation

Plural

Inversive

(40) {chu’un meñ masi-pura 3.refl person mate-among cheqna-ska-n-Ø-nch} hate-refl-plv-3.sbj-decl ‘Among their own fellow people, they hate each other.’ (E.22) The suffix -gi’in ‘like’ occurs ten times in the text (one of which is shown in (26) above).

12.6 Verbs With respect to its verbal morphology, Puquina was a strictly suffixing language, with the exception of one prefix (e- ‘reversive’). Like many other Central Andean languages it had a strict suffix order, proceeding from derivational morphology, to tense, mood, and person markers, and finally to discourse marking morphemes (cf. Chapter 19 by Matthias Pache in this volume). Some of these discourse morphemes can be analyzed as sentential enclitics, since they can also be realized on nominal elements. For that reason, they will be treated in Section 12.7. The verbal slot template is shown in Table 12.9.

12.6.1 Subclasses In Section 12.3.4.3, we showed that particular Puquina verbs, with a very high degree of regularity, take either voiceless or voiced variants of the suffixes in Table 12.2. This fact allows us to speak of two verb classes in Puquina. It thus becomes necessary to ask what factors might explain membership in one or the other verb classes. A first hint regards loanwords. The loans in the Rituale (from Quechuan, Aymaran, and Spanish) are found almost entirely in the voiced class; thus, we can say that loanwords in Puquina are borrowed into the voiced class. However, this pattern does not explain the distribution of the (presumably) native Puquina verbs between the two classes. One hypothesis might be a phonological motivation; however, we find vowel-initial and consonant-initial roots in each of the classes, comprising one, two, and three syllables,

Person subject, imperative

Potential, optative

Clausal morphemes

and with the same classes of phonemes. For instance, the voiceless class includes kaña- ‘to bring’ and kata- ‘to hear,’ while the voiced class includes kara- ‘to cry’ and kaka- ‘to regret.’ It is not clear what kind of phonological distinction between the former and the latter might explain their class membership. A second possibility is that the syntactic characteristics of the verbs (e.g. transitive versus intransitive) play a role. However, we find verbs with such syntactic characteristics in both classes. A third possibility is that the semantic properties of the verbs are relevant to their class membership. Indeed, some of the verbs in the voiced class have to do with suffering, sickness, sadness, and regret, but this pattern is weak. Thus, it may be that the classes followed some pattern in the past, but that this pattern had simply broken down by the time the Rituale was composed.22 Until a pattern can be found that explains membership in the verb classes, we will treat them as arbitrary.

12.6.1.1 Copula constructions In Section 12.3.4.3, we saw that Puquina could attach main tense markers to bare nouns and nominalizations. It is a matter of debate whether we must stipulate zero copulas to account for this. However, relevant to this question is the fact that Puquina also had two overt copula verbs. The first is ka-, which Puquina shares with Quechuan and Aymaran. Puquina ka- also had a variant ha-. The second Puquina copula verb was ascha-. The first functioned as a true meaningless copula that served as the carrier of person and tense, but the latter also functioned as a locative verb. Compare (41) and (42): (41) {Hostia-t-i Iesu Christo-xe apa-nk host-loc-q Jesus Christ-top not-rstr ka-Ø-nch be-3.sbj-decl

tanta-n23 bread-decl

ka-Ø-nch} be-3.sbj-decl

‘Is Jesus Christ in the host? He is not, it is bread.’ (E.09-E.10) 22 This scenario would be somewhat like the fate of the Niagantsi Arawak A-class and I-class verb distinction. These used to regularly designate reflexive and irreflexive verbs, respectively (Payne 1981), but that distinction has become synchronically weak in Nanti and the Matsigenka varieties of the family (Michael 2008: 250 n. 26).

357

emlen, mossel, van de kerke, adelaar (42) {communion Sacramento-xe kiñ-i qapaq communion Sacrament-top what-q mighty qollana sacramento altar-na ascha-eno superior sacrament altar-loc be-ag hostia} host

Table 12.10 Puquina derivational verbal suffixes No. of unique stems that the suffix is attested with

Total occurrences

(h)e-

‘reversive (rev)’

5–6

20

-toch(u) ~ -tos(u)

‘cislocative, translocative (cisl)’

7

16

-schi

‘completive (compl)’

7

10

-na

‘causative (caus)’

28

50

-ska

‘reflexive (refl)’

45

120

-si

‘medial, reflexive (med)’

4

12

-ti

‘reciprocal (recp)’

2

2

-ke

‘applicative (appl)’

17

30

12.6.2 Verbal derivation

-ke

‘future (fut)’

50

101

-a

‘past (pst)’

7

12

We find a limited number of derivational morphemes in the corpus, shown in Table 12.10. There were likely more such morphemes in Puquina that are simply not attested in the Rituale. Given the unreliable nature of the Rituale as a source of linguistic data, we also give tallies of each derivational morpheme’s tokens, as well as the number of unique verbs that they appear with. Note that these tallies vary slightly depending on how one analyzes constructions that are not entirely clear. In addition to these affixes, there are others that appear only on Quechuan (and Aymaran) loans, particularly on verbs from the pan-Andean Catholic lexicon (for instance, hucha-ncha- ‘to sin’; see Durston 2007b for more about this sort of religious lexicon). It is not entirely clear whether suffixes like -(n)cha ‘factive,’ -chi ‘causative,’ and -ku ‘reflexive’ were fossilized elements of loan verbs, or whether they were

-lli

‘transformative 6 (tf)’

38

‘What is the communion Sacrament? The mighty superior Sacrament is the host located on the altar …’ (E.01-E.02) Example (42) shows that the 3rd person copula could be absent in predicative constructions; the ‘is’ in the English translation has no equivalent in the Puquina sentence. However, if the copula had to carry relevant information, as in (41), the copula could be realized. At the same time, in (41) we see that ascha- is not realized when the location of Christ in the host is questioned. The verb kaha- expresses the meaning ‘to be, exist.’ In our opinion, this verb has too much semantic content to be called a copula, as illustrated in (43): (43) {dios señ iki pi-m kaha-Ø-ñao} God 1pl.poss father you-com be-3.sbj-emph ‘God our Father is with you.’ (S.02)

23 The occasional spelling of ‘bread’ as in the Rituale suggests that the second consonant may have been glottalized, as it was in Southern Quechua at that time. However, since we do not find consistent evidence for ejectives in Puquina, we do not represent this term as ejective in this chapter.

358

productive in Puquina. The fact that these suffixes do not appear on native Puquina roots suggests the former scenario. Three other likely derivational affixes, , , and , appear once or twice in the text, making a reliable analysis of their semantics impossible.

12.6.2.1 (h)e- ‘reversive (rev)’ The only verbal prefix in Puquina is the reversive (h)e-. It is combined with several verbs, the most common of which (consistent with the religious nature of the text) is suma- ‘to live,’ giving the meaning ‘to come back to life.’ In almost all attested cases, it is found in combination with the cislocative/translocative morpheme -toch(u), as in (44).

puquina (44) {kuha-ñ a-pi-i win meñ like-decl say-2.sbj-q all person señ 1pl.poss

pip-m anima-m-hamp flesh-com soul-com-add

kachuntu-na end.of.world-loc

e-suma-toch-n-ki-nch} rev-live-cisl-plv-1.sbj-decl

‘Do you believe that we, all people, come back to life with our flesh and soul at the end of the world?’ (G.09)

12.6.2.2 -toch(u) ~ -tos(u) ‘cislocative, translocative (cisl)’ There are 16 occurrences of the derivational translocative morpheme -toch(u) ~ -tos(u) in the text, five of which are duplicates.24 These instances often correspond to Quechuan -mu ‘cislocative’ or -pu ‘translocative,’ or a combination thereof, which is the case of (44) above. We therefore believe that this suffix had both cislocative and translocative directional interpretations.

12.6.2.3 -schi ‘completive (compl)’ The morpheme -schi (which, as Torero 2002: 432 argues, may have had a voiced allomorph represented in the Rituale as ) was likely a verbal derivational suffix with a completive aspect. However, its exact function is unclear due to the small number of occurrences in the Rituale, and due to the fact that these occurrences fall within passages that are difficult to interpret. An example of -schi is given in (45). As in other cases in the Rituale, this passage gives equivalent Quechuan (picha- ‘to clean’) and Puquina (mata- ‘to wash’) lexical items. (45) {chu pip25 picha-schi-ke-Ø-nch 3.poss flesh clean-compl-fut-3.sbj-decl mata-schi-ke-Ø-nch} wash-compl-fut-3.sbj-decl ‘He has to clean up and wash up his body.’ (E.24c)

24 In addition to these 16 cases, there are four more likely instances of -toch in the enigmatic phrase ekitochkineno (see Mossel et al. 2020). 25 is apparently a misspelling of ‘his flesh,’ likely because of confusion with the following verb root picha- ‘to clean.’ This analysis is supported by the corresponding Quechua text in the Rituale, in which we find ukunta ‘his flesh.’

12.6.2.4 -na ‘causative (caus)’ The status of -na as the Puquina causative marker is clear. There are around 50 occurrences of this suffix, on some 28 unique verbs. An example is given in (46), in which the meaning ‘to kill’ is created by affixing the causative -na to the verb halla- ‘to die.’ (46) {meñ halla-na-eno-pi-i} person die-caus-ag-2.sbj-q ‘Did you kill someone?’ (L05.02)

12.6.2.5 -ska ‘reflexive (refl)’ Puquina had a reflexive suffix -ska. There are about 120 occurrences of this morpheme, suffixed onto some 45 unique verbs. The morpheme could also be used as a reciprocal marker, as in cheqna-ska- ‘to hate each other’ in (40) above. An example of the reflexive use of -ska is shown in (47). Note that -ska sometimes deletes the previous vowel, as in this example, but that this process is inconsistent. As mentioned in Section 12.3.4.3, it always triggers a voiceless variant in the following suffix (where relevant). (47) {ka po se'e sipi-ska-tawa a-ta now 2.poss heart beat-refl-subis say-imp apu-re Iesu Christo} Lord-voc.m Jesus Christ ‘Now, while beating your heart, you should say: “My Lord Jesus Christ …”’ (G.14a)

12.6.2.6 -si ‘medial, reflexive (med),’ -ti ‘reciprocal (recp)’ Puquina also had a medial/reflexive marker -si, which is shared with Aymara (see Chapter 8 by Matt Coler in this volume). Most occurrences are found in text L06, in which parishioners are questioned about their sexual habits. Here, we find -si in combination with the verbal root ro- ‘to have sex with,’ when addressing women (as well as two cases of ro-ska-, which includes the reflexive suffix -ska). By contrast, when men are questioned about their sexual contacts with women, the derived verb ro-ga- is used instead. The most obvious explanation is that for the priest, ro-ga- represented male participation in sexual intercourse, while ro-ska- ~ ro-si- represented female participation in sexual intercourse. In (48) we see that -si may also have an interpretation ‘for one’s own benefit’ that is commonly associated with the reflexive.

359

emlen, mossel, van de kerke, adelaar (48) {po meñ-wanana chu sqana … 2.poss person-abl 3.poss silver … chu 3.poss

he-hamp kaña-no nu- hamp llama-add exist-inf who-add

taku-ch too-si-eno-pi-i} force-abl bring-med-ag-2.sbj-q ‘… did you take for yourself, by force, from your people, their money …, their llamas, or anything else?’ (M.01) Torero (2002: 433) also proposes -ti as a possible reciprocal suffix. There are two such occurences in the text, in a single passage: in P.04 (where the suffix combined with -ska ‘reflexive,’and in P.05 (where -ri, a possible voiced allomorph of -ti, appears). This analysis remains tentative. The passage in question is shown in (49). (49) {meñ kuya-ska-ti-pi-nch oracion person love-refl-recp-2.sbj-decl prayer kallaka-no kor-na do-inf time-loc

ayuna-no fast-inf

kor-na-hamp ama tok-ti-pi-ns} time-loc-add not touch-recp-2-sbj-decl ‘You should love each other, and during prayer and fasting time, you should not touch each other.’ (P.04–05)

12.6.2.7 -ke ‘applicative (appl)’ The applicative suffix -ke is widespread in the Rituale, and in most cases it is a direct counterpart of the morpheme -pu in the corresponding Quechua text, which likewise adds an argument to the valency frame of the derived verb. Like that suffix in Southern Quechua, it may have had a benefactive/malefactive interpretation with transitive verbs, while with verbs of motion it seems to have had a regressive interpretation. It is identical in shape with the future marker -ke, which creates occasional problems of interpretation (as we noted above, the tallies in Table 12.10 vary slightly depending on how one analyzes particular instances). However, the co-occurence of the applicative and future suffixes clearly shows the existence of both, as in (50) (note that this exhibits the alternating voicing pattern described in Section 12.3.4.3)

360

(50) {so hor-su a-Ø-ñao ama qasi-nana two speak-ptcp say-3.sbj-emph not in.vain-abl dios men yakchi-ke-ke-pi-anch} God name misuse-appl-fut-2.sbj-decl ‘The second word says: you shall not misuse God’s name in vain.’ (W.04)

12.6.2.8 -ke ‘future (fut)’ The Puquina future morpheme -ke is also illustrated in (50) above (and in many other examples throughout this chapter). This suffix may not have belonged properly to the domain of tense. This is because it is realized before the inverse marker -s, along with the rest of the derivational morphology, which indicates that it may be part of that system instead. On the other hand, the 1st person singular future suffix -kina does in fact appear toward the end of the affix string.

12.6.2.9 -a ‘past (pst)’ We encountered several cases in the Rituale in which a suffix -a appears to mark past, as in miñ tuku-a-nch ‘he became a person.’ The corresponding Aymaran and Quechuan texts in the Rituale are also marked in the past tense. However, in (51), we find -a realized before the nominalizer -so, which makes a main tense interpretation for -a impossible. (51) {wakna stu hor-su-xe Iesu Christo other seven speak-ptcp-top Jesus Christ señ iki cheqa Dios meñ our father real God person tuku-a-so-ch holla-Ø-ñao} become-pst-ptcp-abl speak-3.sbj-emph ‘The other seven words speak about how Jesus Christ our father, the real God, became a person.’ (V.02) This leads us to the conclusion that -a is also part of the derivational morphology, like the future marker -ke. Notably, in 10 of 12 occurrences, this suffix combines with a verbal root shared with Quechuan and/or Aymaran.

12.6.2.10 -lli ‘transformative, to act unnaturally (tf)’ The derivational morpheme -lli is fairly common in the Rituale, but its function is difficult to discern. In one usage, it

puquina apparently referred to actions that are performed in a manner that is fake, unnatural, or uncommon, and it is found most often in graphic descriptions of prohibited sexual acts (especially in text L06). The same suffix serves similar functions in some Quechuan varieties; however, -lli appears on four native Puquina roots in the Rituale, so it also seems to have been productive in Puquina. An example is given in (52), where the root tak- ‘to work’ (otherwise realized as tak-ska-) has the derivation tak-lli- ‘to masturbate.’ (52) {po moha-m pi kiki tak-lli-ska-tawa 2.poss hand-com 2sg self work-tf-refl-subis raago-m-gi'in hucha-ska-eno-pi-i} man-com-simul sin-refl-ag-2.sbj-q ‘Did you sin by “working” yourself with your hand like with a man?’ (addressed to a woman) (L06.06) Most of the other Puquina cases of -lli in the Rituale appear in pan-Andean Catholic terms (which are also found in both the corresponding Quechua and Aymara texts), as in upalli‘to adore’ and huchalliku- ‘to sin.’ In these Quechua cases, -lli appears to serve a verbalizing function.

12.6.2.11 , , and As mentioned above, there are several phonetic strings in Puquina verbal constructions that are in the correct position to be analyzed as derivational suffixes, but their meaning cannot be interpreted because they occur only once or twice in the text. These include, for instance, , of which there is just one occurrence in the text (K.14). There is little context to tell us what its function might have been. However, two other suffixes are of interest because they are also mentioned by Pieter Muysken in Chapter 15 of this volume as existing in Kallawaya, though, as in Puquina, they are unanalyzable. These Kallawaya suffixes are -kha ~ -xa and -naxa. Evidence for as a derivational morpheme in Puquina (which may correspond to Kallawaya -kha ~ -xa) comes from the opposition between the terms for sexual activity, ro-gaand ro-si-, treated in Section 12.6.2.6. This element also appears in some verbs that may, in fact, be polymorphemic. These include, for instance, ‘to hide’ (cf. ‘to hide’). It is not clear what sound might have represented in these cases, especially when it appears in complicated graphic sequences like (see also the many spellings of ‘to hide’ in Table 12.1). The alternation of and in some verbs (e.g. ~ ‘to make’) suggests that this suffix may also have been subject to the voiced/voiceless allomorphy described in Section 12.3.4.3. However, its meaning remains unclear at this time.

The case of is also quite complex. There are only three cases in the Rituale, of which one is shown in (53). Here, the suffix appears to play a nominalizing role. Note that Aymara also has a suffix -naqa, which denotes “repetitive action performed in various places” (Coler 2014a: 323; see also Chapter 8 by Matt Coler in this volume), but this does not seem to fit the meaning of the Puquina suffix. (53) {misa qallara-so titu-ch ho-naxa-kama mass begin-ptcp moment-abl end-?-lim kata-lla-ke-pi-anch} here-?-fut-2.sbj-decl ‘You must listen from the moment the mass begins until it finishes.’

12.6.3 Verbal inflection Puquina verbal inflection consisted of person and number marking in interaction with an inverse system and some modality markers. We treated two tense-like morphemes as part of the derivational system in Sections 12.6.2.8 and 12.6.2.9.

12.6.3.1 Person marking and the inverse construction Puquina had a four-way distinction in personal pronouns, but person marking on the verb was three-way. There was also one fused form, -kina, which indicated a 1st person singular subject future. The pronouns and verbal markers are shown in Table 12.11. Puquina made no overt distinction between 1st person inclusive and exclusive verb agreement. The verbal marker Table 12.11 Puquina personal pronouns and verbal subject person marking Pronoun

Verb marking

Fused forms

1sg.sbj

ni

-ke ~ -ki

-kina ‘1sbj.fut’

2sg.sbj

pi

-pi

3sg.sbj

chu

-Ø

1pl.sbj

señ

-n-ke ~ -n-ki

2pl.sbj

(pi)

-n-pi

3pl.sbj

(chu)

-n-Ø

-n-kina ‘1pl.sbj.fut’

361

emlen, mossel, van de kerke, adelaar Table 12.12 Puquina person marking in transitive clauses 1sg.obj 1sg.sbj 2sg.sbj

V-s-ki

3sg.sbj

V-s-ki

1pl.sbj 2pl.sbj

V-n-s-ki

3pl.sbj

V-n-s-ki

2sg.obj

3sg.obj

V-ki

V-ki V-pi

V-n-s-ki

V-s-pi

V

V-n-s-ki

V-n-ki

V-n-ki

V-n-s-pi

V-n-pi

V-n-s-ki

V-n

V-n-s-ki

-n was used to express plurality of the subject. Another relevant observation is that personal pronouns rarely occurred in any syntactic position except as verbal subjects, where they were optional. Puquina was thus a pro-drop language, and pronouns were only realized for purposes of discourse coherence. Note that this system only encodes a single argument of intransitive verbs (whatever its thematic role) as a syntactic subject controlling agreement marking on the verb. Person marking on transitive verbs was more complicated, as shown in Table 12.12. Puquina did not have an extra slot in the verb matrix to express the second argument of transitive verbs.26 However, it did have an inverse rule in interaction with a 1 > 2 > 3 person hierarchy (ph). In this respect, Puquina can be characterized as a language with a direct-inverse system, as observed by Torero (2002: 410) (for more on the category of inverse in the Andean languages, see Adelaar 2009; Haude and Zúñiga 2016). This formed the basis of Puquina argument expression: the argument with the highest position in the ph, whatever its thematic role, had to be realized as the subject of the clause controlling agreement on the verb. Thus, interactions involving a 1st person singular or plural argument that acted on a lower-ranked patient or recipient (for instance) were simply expressed as V-ki or V-n-ki, since these are consistent with the ph. However, when an interaction violated the ph (i.e. when the agent argument was ranked lower than the patient or recipient argument), the inverse marker -s appeared in front of the person subject marker. This verbal agreement marker then referred to the patient or recipient argument in the transition. In this way, 2nd and 3rd person agents acting on 1st person patients/recipients are expressed as V-s-ki or V-n-s-ki. Similarly, interactions involving 2nd person arguments acting 26 In his analysis of a complex sentence in text E.11, Torero (2002: 429) proposes that -sin was a 1st person verbal object suffix. We are not convinced by this analysis, though the construction in question remains opaque.

362

1pl.obj

2pl.obj

3pl.obj

V-n-ki

V-n-ki V-n-pi

V-n-s-pi

V-n

V-n-ki

V-n-ki V-n-pi

V-n-s-pi

V-n

on a 3rd person patient or recipient were expressed as V-pi (sg) or V-n-pi (pl). Interactions involving 3rd person agents acting on 2nd person patients/recipients violate the ph and are expressed as V-s-pi (sg) or V-n-s-pi (pl). Note that all inverse constructions involve a speech act participant (1st or 2nd person); transitive interactions involving only 3rd person never involve inverse marking.27 In such cases, the agent argument is realized as a subject controlling agreement on the verb. In this respect, our analysis diverges from that of Torero (2002: 428–9). Torero claimed that Puquina did mark agents of transitive clauses (nouns and 3rd person pronouns) with ergative -s, in particular in interactions involving 3rd person in both agent and patient/recipient roles to avoid ambiguity. This last claim is hard to verify, since almost all of the examples of -s that Torero presents actually appear in inverse constructions involving speech act participants (1st and 2nd persons).28 The exact nature of -s marking is not yet fully understood, but we assume instead that the argument controlling agreement on the verb has the status of subject. In inverse constructions, this is the patient or recipient and, by consequence, the agent (if expressed) has to be realized as an adjunct. We analyze -s as the case marker for such agent adjuncts.29 There are two other reasons to doubt

27 In this respect, Puquina seems to be exceptional. Most languages with direct-inverse systems make a distinction within the 3rd person class, e.g. animate > inanimate or proximate > obviate. 28 Specifically, Torero presents 15 examples of a transitive verb with a 3rd person agent. 12 of these are inverse constructions involving a 1st or 2nd person patient/recipient. Only three examples express an interaction between 3rd persons. In two of these cases the agent is marked with -s. In the first it marks the agent adjunct in a causative reflexive clause (Torero 2002: 439). In the second it results from a mismatch in meaning suggesting that it would better be interpreted as a genitive. The only reliable case was presented in fn. 19 without an overt case marker. 29 In this respect, the inverse construction does not differ greatly from a passive construction. The crucial difference is that the passive is optional for the speaker, while the hierarchical inverse construction is obligatory.

puquina Torero’s ergative analysis: first, the marker was not obligatory, and indeed is rare in the Rituale. Second, it was used as a genitive to mark subjects in nominalized contexts, as we showed in Section 12.5.3.2.2, which is not easily reconciled with Torero’s ergative analysis. However, his proposal that -s marking played a role in avoiding ambiguity is relevant, since many instances of 3rd person -s-marked agents occur in sentences with non-standard OSV word order. Furthermore, some arguments are marked with -s for other reasons, as in (54), where it apparently functions as a vocative. These appearances of -s should not be confused with each other. (54) {raago-s po atago hata-ke-pi-ns man-voc 2.poss woman want-fut-2.sbj-decl ehe and

kuha atago-s-xe po raago like woman-voc-top 2.poss man

hata-ke-pi-ns} want-fut-2.sbj-decl ‘You, man, you shall love your wife; and in the same way, you [on the other hand], woman, you shall love your husband.’ (P.05)

12.6.3.2 Tense and aspect Tense and aspect marking is rare or perhaps even absent in Puquina (see discussion of tense-like future and past derivational markers in Sections 12.6.2.8 and 12.6.2.9, respectively). Present tense was not marked. The habitual construction with the agentive marker -eno is extensively used, but it does not have a meaning that is directly related to tense or aspect.

12.6.3.3 Mood Puquina distinguished the declarative mood (12.6.3.3.1), the imperative mood (12.6.3.3.2), and the potential/optative mood (12.6.3.3.3).

12.6.3.3.1 -nch ‘declarative (decl)’ The declarative suffix nch ~ -ns ~ -ñ is one of the most common morphemes in the Puquina sections of the Rituale, and it can be found in the majority of the example sentences in this chapter. It appears in nearly all affirmative and negative verbal constructions, but it does not cooccur with the imperative, interrogative, or personal belief markers. The declarative is explored in greater detail in Section 12.7.4.

12.6.3.3.2 -ta ‘2nd person imperative (imp)’ and -suma ‘2nd person subject, 1st person object imperative (imp)’ We find the suffixes -ta ‘2nd person imperative’ and -suma ‘2nd person subject, 1st person object imperative’ in the same slot in the affix string as the declarative person markers. These can also occur in inverse constructions: the 2nd person imperative had an inverse counterpart -s-ta, which refers to the transition between a 3rd person agent and a 2nd person patient. Both the inverse and the non-inverse constructions appear in the first line of the text for the wedding ceremony, in (55) and (56), respectively. (55) {suti-nk hor-ska-toch-n-s-ta} clear-rstr speak-refl-cisl-plv-inv-imp ‘… you (pl) must have been clearly told over there …!’ (O.01) (56) {suti hor-ska-n-ta} clear speak-refl-plv-imp ‘… speak out clearly (pl)!’ (O.01) An important difference between these examples is that in (56) expresses a normal 2nd person imperative, while in (55) includes an inverse -s that marks the subjects as the patients of the act of speaking (i.e. a 3rd person acting on a 2nd person transition). In addition, Puquina had an imperative involving a 2nd person subject interacting with a 1st person object, -suma. In three cases, -suma is represented as or , but it is not known what the difference in these forms might have been. It is possible that the final is also reflected in the personal belief suffix -ñao (Section 12.7.6), but data are too sparse to be sure. Examples of both -suma and -sumau are given in (57). (57) {ama ehe akro-suma hucha-guta not and let-2.sbj.1.obj.imp sin-all señ hoto-no-wa enahata we fall-inf-ben rather ento-nana qespi-na-sumau} bad-abl be.safe-caus-2.sbj.1.obj.imp ‘And don’t let us fall into sin, but rather, save us from evil.’ (R.06)

12.6.3.3.3 -iska ‘potential (pot) and -anta ‘optative (opt)’ We encounter a number of occurrences of the optative marker -anta in the Rituale. All of these cases have a 3rd

363

emlen, mossel, van de kerke, adelaar person subject, with one exception discussed below. Two instances of 3rd person constructions are shown in (58). (58) {no hucha pampacha-so ascha-Ø-anta no my sin forgive-ptcp be-3.sbj-opt my anima-xe qespi-Ø-anta} soul-top be.safe-3.sbj-opt ‘… may my sin be forgiven, and may my soul be saved.’ (E.26) Prior analyses have treated -anta specifically as a 3rd person marker (Torero 2002: 430; Adelaar and van de Kerke 2009: 135). However, (59) shows that it could also occur with the 2nd person marker (here, with an inverse suffix). This suggests that it did not specify person, as previous analyses have suggested; indeed, Puquina verbal constructions can only include a single person marker (Section 12.6.3.1). (59) {dios waqaycha-s-pi-anta} God protect-inv-2.sbj-opt ‘May God protect you.’ (L0.01) Therefore, it seems that the optative -anta and potential -iska together occupied a distinct slot in the suffix template. For this reason, both can follow a subject marker, as shown in (60), which gives two instances of the potential -iska. These also exhibit inverse marking. (60) {Dios iki-s kato pakas-[s]u30 God father-gen inner world-all paña-ke-s-pi-iska supaya-kuna-s throw-appl-inv-2.sbj-pot devil-pl-gen too-s-ke-s-pi-iska} bring-refl-appl-inv-2.sbj-pot ‘[Beware], you might be thrown into hell by God, you might be brought by the devils.’ (N.01)

12.7 Morphemes with clausal scope Puquina had several morphemes that functioned at or above the level of the sentence. In many cases these have regular Quechuan and Aymaran counterparts, but in other cases they do not. These include a set of topic markers (12.7.1), 30 Here we encounter an unexpected morpheme -u or -su as a directional case marker instead of the expected -guta.

364

an interrogative (12.7.2), an additive (12.7.3), a declarative marker that is mostly found on verbs, but occasionally on other elements (12.7.4), a restrictive (12.7.5), and an apparent epistemic or evidential morpheme that indicated personal belief or conviction (12.7.6).

12.7.1 Topic markers In the Rituale, topic marking is accomplished through a set of morphemes variously spelled , , , , -, , , , , , and . Due to the orthographic inconsistency of the text, it is impossible to be sure just how many separate morphemes these spellings represent. However, some regularities can be discerned, particularly when the Puquina sections are compared carefully with the Quechua and Aymara texts. First, the spellings , , , , , , and are all used to indicate contrastive topic reference, in the manner that English and on the other hand (or in questions, and what about) establishes a contrastive relationship between two referents or stretches of discourse. This can be seen in (54) above, in which the priest makes a statement to a man, and then turns to a woman, and says, ‘You, woman, on the other hand …’ It can also be seen in (61), which describes the first three commandments, and then turns, ‘on the other hand,’ to the last seven commandments. (61) {pesk kapa kamachi-so Dios señ first three order-ptcp God we upalli-no-wa ukru stu adore-inf-ben other seven kamachi-so-xe meñ-gata señ order-ptcp-top person-pl 1pl.poss kuya-no-wa} love-inf-ben ‘The first three commandments are for us to worship God; the other seven commandments, on the other hand, are for us to love our fellow men.’ (W.02) In the corresponding Quechua texts, all seven of these orthographic variants correspond to the Quechua contrastive topic marker -ri. It is possible that more than one morpheme is represented by these seven orthographic variants, but given their close overlap in function, and given their regular correspondence with Quechua -ri, it seems more likely that 31 is an apparent misspelling of pesk, written elsewhere in the Rituale.

puquina Oré simply struggled to represent this particular Puquina consonant with any consistency. The topic marker , which is rarer in the Rituale, served a similar contrastive function.32 However, it appears to refer more specifically to temporal sequences, in the manner of and then, at that point, or but then. It also frequently corresponds to Quechua -ri. A Puquina example is given in (62), with its corresponding Quechua text containing -ri in (63). (62) Puquina

{tanta-ga Iesu Christo pip-nk bread-top Jesus Christ flesh-rstr miñ tuku-a-Ø-nch person become-pst-3.sbj-decl

Rituale. A Puquina example is given in (64), and the corresponding Quechua text is given in (65). Note that the same correspondence between Puquina and Quechua (-qa) can also be seen on the term for bread, , in (62) and (63) respectively. (64) Puquina

{katir-ga Iesu Christo señ iki} but-top Jesus Christ our father ‘But Jesus Christ, our Father …’ (E.10b) (65) Quechua

{icha-qa Iesu Christo yaya-nchik} but-top Jesus Christ father-1incl.poss ‘But Jesus Christ, our Father … (E.10b)

vino-xatampe suma-eno kam wine-top live-ag blood tuku-a-Ø-nch} become-pst-3.sbj-decl

12.7.2 -i ‘interrogative (q)’

‘The bread became Jesus Christ’s human body, and then the wine became his living blood.’ (E.10c)

Puquina had an interrogative marker -i. Like Quechuan -chu and Aymaran -ti, it was used in polar questions. In (66), -i follows the locative suffix -(u)t.

(63) Quechua

{tanta-qa Iesu Christo-p uku-n-man-ña bread-top Jesus Christ-gen body-3.poss-all-compl tuku-n vino-ri kawsa-q become-3.sbj wine-top live-ag yawar-ni-n-man-ña-taq-mi blood-ep-3.poss-all-compl-top-evd

tuku-n} become-3.sbj

‘The bread became Jesus Christ’s body, and then the wine became his living blood.’ (E.10c) Similarly, in one case (E.24b), the Quechua discourse marker chay-manta-ri (that-abl-top) ‘and then’ aligns precisely with the Puquina construction ka-ch-xatampe (that-abl-top) ‘and then.’ Topic marking is also accomplished with another set of orthographic variants, , , and , which (confusingly) overlap with those of the contrastive marker described above. We use the normalized form -ga for the variants of this morpheme, which regularly correspond to the Quechua topic marker -qa in the corresponding texts in the 32 It is possible that -xatampe is a morphologically complex form including a topic marker -xa and the additive morpheme -hamp. However, since this leaves the intervening unexplained, we do not adopt that analysis here, and instead treat -xatampe as a single morpheme.

(66) {hostia-t-i Iesu Christo-ga} host-loc-q Jesus Christ-top ‘… is Jesus Christ in the Host?’ (E.09) However, unlike Quechua -chu and Aymara -ti, the interrogative suffix -i was also used in combination with the interrogative pronouns (see Table 12.6), as in (67). (67) {nu-i Iesu Christo-ga} who-q Jesus Christ-top ‘Who is Jesus Christ?’ (E.05) Another function of interrogative -i was to mark binary choice, in a manner similar to Quechuan -chu and Southern Aymara -cha. This is illustrated in (68). (68) {kiñ too-pi-i raago haya-i inke what bring-2.sbj-q man child-q or atago haya-i} woman child-q ‘What do you bring [to the church], a male child, or a female child?’ (C.01) 33

is apparently a misspelling of ‘Christ.’

365

emlen, mossel, van de kerke, adelaar

12.7.3 -hamp ‘additive (add)’

12.7.4 -nch ‘declarative (decl)’

The additive marker -hamp served several functions in the Rituale. First, -hamp meant ‘also’ or ‘and,’ as in ‘his body and his blood’ in (69).

A very common suffix -nch (spelled , , and ) was used in declarative sentences, and did not cooccur with the interrogative -i. This morpheme usually appeared on the verb, though in some cases it also appeared on other constituents in the sentence. This suggests that it was involved in information structure, but it is not clear how this might have worked. In declarative sentences, -nch frequently corresponds to the Quechuan direct evidential -mi ~ -m and the Aymara direct evidential -wa, though the functions of those morphemes do not overlap in other contexts. A typical example is given in (72).

(69) {chu pip chu kam-hamp hinanti-wa 3.poss flesh 3.poss blood-add totality-ben hiqi-ska-Ø-nch give-refl-3.sbj-decl

apa-nk not-rstr

killchi-tawa} divide-subis’

‘… he gives his body and his blood to everyone without dividing [himself].’ (E.14a) As in Quechuan and Aymaran languages, the additive marker was also used to express indefinites, as in ‘any of your sins’ or ‘even one of your sins’ in (70). In negated clauses, such constructions took meanings like ‘not even one.’ (70) {confesa-ska-tawa-xe huksto-hamp po confess-refl-subis-top one-add 2.poss hucha-k histikga-eno-pi-i} sin-acc hide-ag-2.sbj-q ‘And while you were confessing, did you hide even one of your sins?’ (L0.07) Similarly, the Puquina additive suffix -hamp was also used on interrogative pronouns, or in expressions including interrogative pronouns, to turn those into indefinites. For instance, nu ‘who’ becomes nu-hamp ‘someone’ (M.02), kiñ ‘what’ becomes kiñ-hamp ‘something’ (L06.07), and the noun patero ‘priest’ becomes nu patero-hamp ‘some priest’ (L08.01). Note that Quechua -pas ~ -pis and Aymara -sa serve precisely this function with interrogative pronouns as well (see Chapter 5 by Aviva Shimelman and Chapter 6 by Aviva Shimelman and Jairo Valqui in this volume). A Puquina example is given in (71). (71) {nu kiñ meñ masi-k-hamp who what person mate-acc-add sipsika-eno-pi-i} spread.rumors-ag-2.sbj-q ‘Have you spread rumors about any of your fellow men? (L08.06) Additive -hamp was used in concessive subordinate clauses, as we describe in Section 12.8.5.5.

366

(72) {wakna hucha Dios apa other sin God not pampacha-ke-s-pi-nch} forgive-appl-inv-2.sbj-decl ‘… God will not forgive you [for] your other sins. (I.06) In many cases, Puquina verbal constructions end in rather than . Though the data are inconsistent on this matter, generally appears in verbal constructions regarding a desired state of affairs (where it corresponds to the Quechua polite imperative -nki), and in prohibitive constructions with the negator ama (‘do not’). It also occurs in some conditional constructions. In all of these cases, deletes the preceding vowel in the 1st or 2nd person marker. Though there is too little information to be sure, it is possible that is composed of a verbal morpheme a- and the declarative suffix -nch (indeed, also only appears on verbs in declarative sentences, while can appear elsewhere). However, while corresponds regularly to these functions in the Rituale, it is also occasionally found outside of those contexts in a manner that is difficult to explain. For instance, Table 12.3 and (24) show cases of that do not correspond to ‘desired state of affairs,’ prohibitive, or conditional meanings. For this reason, we believe that the precise nature of must be resolved in further research, and we simply gloss it as a variant of declarative -nch in this chapter until more can be said about it. Examples of the ‘desired state of affairs’ usage of are given in (73) and (74), the prohibitive usage is shown in (74), and two conditional uses of are given in (75).

puquina (73) {po iki po imi-hamp 2.poss father 2.poss mother-add yupaycha-ke-pi-anch} honor-fut-2.sbj-decl

While the restrictive or limitative function of -nk(a) is clear throughout the Rituale, the same form also appears in contexts that do not suggest this function (most often as ), as in (55) and (62). These cases include apparent topic or emphasis marking. The available data are too sparse to de‘You must honor your mother and your father.’ (L04.00) termine whether this was a distinct morpheme, so we gloss these cases with the code rstr as well until more can be said about it. (74) {huksto-nka cheqa Dios upalli-ke-pi-anch 12.7.6 -ñao ‘personal belief (emph)’ one-rstr real God adore-fut-2.sbj-decl ama kaña-nu-hamp dios-gi'in not exist-inf-add God-simul upalli-ke-pi-anch} adore-fut-2.sbj-decl ‘You must worship the one true God, you may not worship anything else like God.’ (L01.01) (75) {kokasuske ehe ama skalli-ta and and not be.afraid-imp enahamp ko hucha-wichna patero-s perhaps this sin-csl priest-gen cha-ke-s-ki-anch punish-fut-inv-1.sbj-decl azote-ke-n-s-ki-anch whip-fut-plv-inv-1.sbj-decl

a-tawa} say-subis

‘And so do not be afraid, saying “Perhaps the priest might punish me, perhaps he might whip us, because of these sins.”’ (K.15)

12.7.5 - nk(a) ‘restrictive-limitative (rstr)’ The suffix -nk(a) appears frequently in the Rituale with a restrictive or limitative function, meaning ‘just’ or ‘only.’ This suffix could attach to a variety of elements. Examples include huksto-nka ‘just one’ in (74); this particular construction occurs frequently in the Rituale, and usually corresponds to the form sapa-lla ‘just one’ in the Quechua versions of the same texts. Another case is given in (76). (76) {sacerdotes-s holla-ska-s puchuka-so-nk} priest-gen speak-refl-inf complete-ptcp-rstr ‘… only once the priest has finished speaking.’ (E.10b)

A morpheme -ñao is only used in propositions for which personal belief, conviction, or knowledge are asserted by the principal (in the sense of Goffman 1981; note that this morpheme often appears in reported speech, so it does not mark the knowledge of the “speaker” per se). Most cases are found in the metatextual voice of the author, for instance when Oré introduces each commandment with a brief summary, as in (77). (77) {kapa hor-su a-Ø-ñao Domingo three speak-ptcp say-3.sbj-emph Sunday kamen fiesta kamen-na-hamp day feast day-loc-add ama tak-ska-ta} not work-refl-imp ‘The third commandment says: “On Sundays and on feast days, don’t work.”’ (L03.00) -ñao also appears in reported utterances—or more often, prescribed utterances—such as the formulae that parishioners are meant to use on particular occasions. For instance, when a man is asked whether he accepts his bride in marriage, he is instructed to say ‘Yes, I want [her],’ a fact to which he can uniquely attest. This is shown in (78). (78) {ti hata-ke-ñao} yes want-1.sbj-emph ‘Yes, I want [to marry her].’ (O.09) Similarly, -ñao is ubiquitous in the Credo text, in which Catholics attest to their personal faith. In this part of the Rituale, we find parishioners using -ñao to declare personal conviction and belief regarding their utterances, as seen in (79). (79) {Poncio Pilato chu ata-so-ch Pontius Pilate 3.poss judge-ptcp-abl

367

emlen, mossel, van de kerke, adelaar mutu-a-Ø-ñao cruz-t suffer-pst-3.sbj emph cross-loc halla-Ø-ñao taga-so kuma-Ø-ñao} die-3.sbj-emph bury-ptcp stay-3.sbj-emph ‘He suffered because of Pontius Pilate’s judgment, he died on the cross, and he remained buried …’ (T.03)

12.8 Basic syntax Our description of Puquina so far has focused on the language’s morphology and phonology. Less attention has been given to the language’s syntax. In the following sections we give a short overview of the basic syntax of Puquina, sometimes referring back to examples in the text.

12.8.1 Grammatical relations In Puquina verbs that have a single argument, the mapping of that argument is direct. The argument is realized as the subject of the verb, and it controls verb agreement. In transitive verbs, however, the mapping of arguments can become quite complex. Puquina is a hierarchical direct-inverse language, so the question of which of the arguments is mapped onto the subject position depends on their relative positions in the person hierarchy (ph). In the Puquina ph, Speech Act Participants (saps, 1st and 2nd person) are higher than nonsaps; among saps, 1st person is higher than 2nd person. If the ph is respected for a particular clause, a “direct” mapping takes place: agent to subject, and patient to object. If the ph is violated, the “inverse” mapping occurs: patient to subject, with the inverse -s suffixed to the verb. Puquina also allows recipients to be realized as subjects of inverse clauses. There is no subject case marking in Puquina, and object case marking is marginal and apparently influenced by Quechuan. All other arguments are treated as adjuncts, which receive a special case marker.

12.8.2 Constituent order Like its neigbors, Puquina was a head-final language. Section 12.8.2.1 treats word order in the noun phrase, and Section 12.8.2.2 describes word order in the clause.

12.8.2.1 Word order in the noun phrase Puquina had three different quantifiers meaning ‘all:’ win, puta, and koma. As modifiers, these appear in the left-most

368

position of the NP.34 In almost all cases, the quantifier comes first, followed by the possessive or demonstrative pronoun, followed in turn by the numeral and/or adjectival element preceding the noun. Puquina allowed rather complex noun phrases, as in (80). (80) {Dios chu iki win ati-eno kupi God 3.poss father all be.able-ag right yaha moha-na} side hand-loc ‘(He sits) at the right-hand side of his almighty father …’ (T.05) This can be represented in the following manner: [[Dios [chu [iki [win atieno]]]] [[kupi yaha] mohana]]].35

12.8.2.2 Word order in the clause The verb tends to be the last element in the clause, and direct arguments of the verb are normally realized directly to the left of it. Puquina shares this OV order with its neighboring languages. It is difficult to say much more about word order on the basis of the Rituale, for two reasons. First, Puquina was a pro-drop language, and person of subject was marked on the verb. For this reason, there are simply not many subject pronouns in the text. Second, religious manuals like the Rituale contain many imperatives and questions that priests are meant to direct to their parishioners. Imperative constructions do not include pronouns, so while there are many verbs marked for 2nd person, we only encountered 15 realizations of the 2nd person pronoun as a subject. The occurrence of 1st and 3rd person pronouns is even more rare. This means that clauses with fully expressed arguments are limited to interactions among 3rd persons, in which both arguments are expressed by an overt NP; only a few such cases are found in the text. Word orders other than S(O)V may be explained by the fact that words and clauses marked with interrogative morphemes tend to be realized sentence initially. This can be seen in (81) and (82). (81) {hostia consagra-so-t-i Iesu host consecrate-ptcp-loc-q Jesus

Christo} Christ

‘Is Jesus Christ in the consecrated Host?’ (E.16) 34 The only exception is the double occurrence of po koma hucha ‘all of your sins’ in Texts I/J, where koma ‘all’ is realized after the possessive pronoun po ‘your.’ 35 We assume that the position of win atieno as a right hand modifier of iki is a calque from yaya llapa atipaq ‘almighty Father’ in the corresponding Quechua text.

puquina (82) {upalli-eno-pi-i chata adore-ag-2.sbj-q mountain

chata} mountain

‘Have you worshipped mountains …?’ (L01.02) In some cases, topicalized arguments occur sentenceinitially in sentences with OSV or SVO order. It appears that a morpheme -s marked the subject of such sentences with non-standard word order.

12.8.3 Major clause types Given the Rituale’s aforementioned bias toward interrogative and imperative clauses, it is difficult to arrive at a definitive analysis of Puquina clause types. However, it is clear that Puquina had declarative, interrogative, imperative, potential, and optative clauses. Each of these had particular markers which we described in Sections 12.6.3.3, 12.7.2, and 12.7.4. In addition to these clause types, Puquina could use the infinitive nominalizer in deontic constructions to indicate obligation or necessity. This is illustrated in (83). (83) {ko-wichna win Christiano wata wata that-csl all christian year year confesa-no-ñao} confess-inf-emph ‘For that reason, all Christians must confess every year.’ (K.03) Similar deontic constructions can be found in neighboring Quechuan languages (with the future nominalizer -n(q)a) and in Aymara (with the nominalizer -ña).

12.8.4 Coordination We interpret coordination as the combination of two syntactically equivalent entities, i.e. [Noun Phrase] [Noun Phrase] or [Clause] [Clause]. These are discussed in Sections 12.8.4.1 and 12.8.4.2, respectively.

12.8.4.1 Coordination of noun phrases The comitative-instrumental case marker -m (see Section 12.5.3.2.8) is used to coordinate NPs. For instance, in (84), all of the coordinated NPs are marked with the comitative-instrumental. Another case can be seen in (30) above.

(84) {huntu huchancha-ska-ke-nch ki'illa-no-m very sin-refl-1.sbj-decl think-inf-com holla-su-m kallaka-no-m-hamp} speak-ptcp-com do-inf-com-add ‘I sinned terribly in thought, and in speech, and in deed.’ (F.02) There are a few places in the text where the concept ‘or’ is expressed by (e.g. in 68) or .

12.8.4.2 Coordination of clauses Clauses were linked by a number of different coordinators (some of which are listed in Table 12.8). The word ehe is extensively used on its own with the meaning ‘and,’ e.g. in (57) and (75), but it also appears in the combination ehe kuha ‘in the same way,’ as in (54). Different forms of kokasuske are also frequently used with a meaning ‘and,’ as in (75). Disjunctive coordination was expressed by means of apasu ‘if that is not the case,’ ‘or.’ A type of comparative clause was expressed by combining the question word kigo with the adverb kuha ‘just as,’ as illustrated in (85): (85) {kigo Sancta Catholica Apostolica Iglesia how holy catholic apostolic church Romana kamachi-n-s-ke-nch kuha} Roman order-plv-inv-1.sbj-decl like ‘… just as we are ordered by the Holy Catholic and Apostolic Roman Church?’ (O.08)

12.8.5 Embedded clauses There exists a basic distinction between complement clauses (with the status of an argument) and other types of embedded clauses (which have the status of adjuncts). Both are treated in this section.

12.8.5.1 Complement clauses A small number of verbs in the Rituale take a direct complement. For instance, infinitival complements could be formed with the verb hata- ‘to want,’ as in (86). (86) {confesa-no-k hata-pi-i} confess-inf-acc want-2.sbj-q ‘Do you want to confess?’ (L0.01)

369

emlen, mossel, van de kerke, adelaar Verbs that selected a participial complement with the participle -so ~ -su included holla- ‘to speak’ and si- ‘to know.’ Such a direct complement could refer to an event or action in its concrete, potentially person-related capacity, as in (87). (87) {ko-na ascha-so si-ska-n-ki-nch} this-loc be-ptcp know-refl-plv-1.sbj-decl ‘We know that s/he is here.’ (E.16c)

12.8.5.2 Conditional clauses The participle -so ~ -su could also be used to form conditional clauses, as in (88). (88) {po hucha ni-ch36 hor-ska-su 2.poss sin 1sg-hm speak-refl-ptcp Dios-ñao kato kachu-guta God-emph inner world-all paña-ke-s-ki-anch} throw.out-fut-inv-1.sbj-decl ‘… if I would tell your sin to someone, God would banish me to the lower world.’ (K.14)

12.8.5.3 Purpose clauses Puquina had two types of purpose clauses. The most common one consisted of a verb marked with the infinitive nominalizer -no, followed by the benefactive case marker -wa, as in (89). (89) {Iesu Christo señ iki señ hocha-no Jesus Christ 1pl.poss father 1pl sin-poss qespi-na-no-wa be.safe-caus-inf-ben

36

(90) {e-ata-s ata-eno e-wis-toch-ke-Ø-nch rev-judge-inf judge-ag rev-go-cisl-fut-3.sbj-decl was-ke-Ø-nch} go-fut-3.sbj-decl ‘… he will come back in order to judge, in order to judge [living and dead men].’ (V.14)

12.8.5.4 Same-subject clauses Puquina formed subordinate clauses in which the subject is the same as that of the main verb by means of the suffix -tawa, as shown in (47), (52), (69), (70), (75), and (94). Comparable structures are found in the Quechuan and Aymaran languages. However, Puquina differed from those languages in that it did not have a suffix that marked the subject of a subordinate clause that was different from the subject of the main verb (switch reference). For this function, Puquina had an element -aso, most likely a-so ‘say-ptcp,’ which gave a subordinate clause its own tense and person marking. The passage in (91) and (92), addressed to men, gives an impression of how the subordinators aso and -tawa ~ -rawa give structure to the text. (91) {pi-xe huchasapa-re hanigo pakas-guta you-top sinner-voc.m high world-all

cruz-t halla-Ø-ñao} cross-loc die-3.sbj-emph

was-kina a-pi a-so koma se'e-n go-1.sbj.fut say-2.sbj say-ptcp all heart-com

‘Jesus Christ Our Father died on the cross in order to save those of us with sin.’ (V.14)

kichu-tawa po hucha kaka-ska-tawa} be.sad-subis 2.poss sin regret-refl-subis

This is one of the rare cases in which we encounter -ch on a 1st or 2nd person personal pronoun that is the subject of a nominalized transitive clause. One might argue that -ch represents the hierarchy marker, although the 2nd person in (88) is not a syntactically relevant argument. Another explanation is that -ch in these cases was an ablative marker marking the agent argument. A third possibility is that in sentences with a transitive (nominalized) verb exhibiting exceptional OSV word order, the subject is marked with -ch.

370

The second type of purpose clause could only occur as the complement of a verb of motion. It took two different forms, as Torero (2002: 443) shows; both of these are found in (90), in which the verb of motion is was- ‘to go.’ One form, which is similar to a construction found in the Quechuan and Aymaran languages, combines the verb of motion with an agentive nominalized clause (in this case, ata-eno ‘in order to judge’). The other combines the verb of motion with a dependent verb marked with -s (in this case, e-ata-s ‘in order to judge again’).37

‘And you, [male] sinner, if you say “I will go to heaven,” being sad with all of your heart, and regretting all of your sins …’ (I.05) 37 This example is instructive since it shows not only two ways to express the purpose clause, but also two ways to express the concept ‘s/he came’ (i.e. with the verbs wis- and was-). It might be that two dialectal variants are represented here.

puquina (92) {po koma hucha hor-ska-ta 2.poss all sin speak-refl-imp ama huksto not one

hucha-hamp histikga-ke-pi-anch sin-add hide-fut-2.sbj-decl

(94) {kuha-ñ a-pi-i iki chu sku like-decl say-2.sbj-q father 3.poss son spiritu sancto Spirit Holy

kapa persona three person

huksto-hamp histikga-pi a-so wakna one-add hide-2.sbj say-ptcp other

ascha-tawa-hamp huksto sapa cheqa be-subis-add one single real

hucha dios sin God

dios God

apa not

pampacha-ke-s-pi-nch} forgive-fut-inv-2.sbj-decl ‘Tell me all of your sins. Do not hide even one of the sins, because if you hide even one of them, you will not be forgiven by God for your other ones.’ (I.06) Additionally, -tawa was used to form direct reported speech clauses when it was combined with the verb a- ‘to say’ (a-tawa), just as in Quechua ni-spa ‘saying.’ Direct reported speech constructions, unlike samesubject subordinate clauses, have independent person and tense marking. In (93), it is shown how aso introduces a conditional clause containing direct reported speech: (93) {sper hor-su a-Ø-ñao wata four speak-ptcp say-3.sbj-emph year watana year-loc

sancta Holy

imi ayuna-ta mother fast-imp

iglesia-n señ Church-gen our a-su ayuna-ke-pi-anch} say-ptcp fast-fut-2.sbj-decl

‘The fourth word says: “Every year you should fast, if our mother the Holy Church says “You should fast.””’ (X.05)

12.8.5.5 Concessive clauses The additive morpheme -hamp (Section 12.7.3) was used in concessive subordinate clauses that marked a contrast with the corresponding main clause, in the manner of English although or even though (a feature achieved in the neighboring Quechuan languages with the additive suffix -pis ~ -pas, and in the nearby Aymaran languages with the additive suffix -sa). This can be seen in (94), in combination with the same-subject marker -tawa.

ascha-so-k} be-ptcp-acc

‘Do you say “yes” [i.e. believe] that the father, his son, and the Holy Spirit, even though they are three people, are just one God?’ (G.04a) For a concessive subordinate clause with a past interpretation, a derivation with participle -so was used, as in (95). (95) {apa roga-su-hamp kuha kiñcha-eno-pi-i} not have.sex-ptcp-add like do.what-ag-2.sbj-q ‘Even though you did not have sex, you pretend that you did?’ (L06.08)

12.8.6 Relative clauses Puquina did not have relative pronouns, but it used the agentive nominalizer -eno and the participial nominalizer -so ~ -su to create relative clauses. These had an internal verbal structure. A participial relative clause is given in (96). (96) {Padre-s penitencia kalla-ska-ta priest-gen38 penitence do-refl-imp a-so-k poqe-pi-i say-ptcp-acc fulfill-2.sbj-q

puchuka-pi-i complete-2.sbj-q}

‘… have you fulfilled what was said by the priest [when he said] “Do penitence”? Have you fulfilled [what he said to do]?’ (L0.09) (96) is an instance of a headless relative clause, but Puquina also allowed relative clauses in which the antecedent was realized within the relative clause itself. An example is given in (81) above. 38 We analyze the -s on the subject of the nominalized phrase as a genitive, just as we do in comparable constructions in Quechua and Aymara.

371

emlen, mossel, van de kerke, adelaar

12.8.7 Negation Puquina, like the Quechuan languages, made a distinction between two negation markers: apa, seen e.g. in (37), (41), and (72) and ama, seen e.g. in (50), (57), and (74). Ama was used in imperative, unrealized or potential, and deontic contexts, while apa was used in declarative and interrogative contexts. This distinction might also be characterized as the opposition between possible and actual worlds. For the negation of adjectives and participles, apa was used. At the level of the clause, we find apa-nk, e.g. in (69), asserting ‘It is not the case that …,’ and its interrogative form apa-nka-i ‘Why is it not the case that…?’ Meanwhile, apa-su, as seen in (14), functioned as a clause coordinator ‘if that is not the case,’ ‘or.’

12.9 Lexical and semantic properties Table 12.13 is a list of all the non-Spanish lexemes for which we could determine a meaning in the Puquina sections of the Rituale (note that there are a number of forms in the text whose meanings remain unclear; these are not listed here). Many of the lexical items found in the Puquina portions of the Rituale are shared with other languages in the region. Most significantly, the Kallawaya language, used in ritual contexts by a group of Quechua-speaking healers north of Lake Titicaca, contains a great many terms that obviously originate in Puquina (see also Chapter 15 by Pieter Muysken in this volume). However, it is unclear whether these Kallawaya terms are from the variety of Puquina recorded by Oré; given the uncertain nature of this relationship, we do not call these terms “cognates,” since this implies a specific type of historical relatedness, but rather we use the more general term “correspondences.” The Puquina lexical presence in Kallawaya is substantial: among the 201 Puquina terms in our list that are not shared with Spanish, Aymaran, Quechuan, or Uru-Chipaya, 76 have corresponding Kallawaya terms (38%). According to Hannß (2017), around half of the Kallawaya lexicon remains unaccounted for etymologically; presumably, a great deal more of these Kallawaya terms could be attributed a Puquina origin if we had more than the meager list of Puquina lexemes yielded by the Rituale with which to compare them (see also Chapter 15 by Pieter Muysken in this volume). There are also many lexical items in the Puquina sections of the Rituale that are shared with Quechuan and Aymaran languages, and to a lesser extent with Uru-Chipaya and other languages in the region. Determining the direction of borrowing of these shared items is a task of considerable

372

difficulty. Some of these terms were adopted by European missionaries for the purposes of converting Indigenous Andean people to Christianity, and they were spread across Quechuan, Aymaran, and other languages during the early colonial period (see Durston 2007b regarding such terms). Other pan-Andean terms are more closely associated with the Inca period than the Spanish colonial period. On this matter, Cerrón-Palomino (2012, for example) notes ethnohistorical evidence that the Inca nobility spoke a lengua particular and that according to their own mythology, they came from the area around Lake Titicaca. On this basis, Cerrón-Palomino attributes many of these pan-Andean terms a Puquina origin, whether or not those terms are attested in the Rituale, the sole source of Puquina lexical data. In this scenario, those Puquina terms would have been dispersed across the entire Andean region after most of the Quechuan and Aymaran expansions had already taken place. This linguistic argument is tied up with ethnohistorical interpretations, and whether one accepts it depends on one’s evaluation of those interpretations. Still other terms are shared only by Puquina and the varieties of Aymara spoken around Lake Titicaca, as recorded in the works of Oré’s contemporary, Ludovico Bertonio (1612a; 1612b). Some of these Aymara terms do not have cognates in Jaqaru, which makes it possible that they were borrowed from Puquina into Aymara when Aymara arrived in the area around Lake Titicaca. Indeed, Cerrón-Palomino (2020a) identifies these terms as originating in Puquina. However, there is not much reason for certainty that they don’t originate in Proto-Aymara instead, and that the Jaqaru cognates were simply lost as the language underwent intensive subsequent contact with the Quechuan languages that now surround it. Given the complexities and uncertainties involved in attributing origins to particular terms in the Rituale, we do not attempt to do so here. Instead, we simply list all the non-Spanish terms and their correspondences with other languages, and refer readers to the works discussed in Section 12.11 for more information. For each of the 336 terms in the wordlist, four columns are given: (i) Normalized representations of non-Spanish terms (in italics), with glosses (between single quotes). This is our best attempt to approximate each term’s phonological form on the basis of Oré’s orthography, or from other sources, where relevant (also see discussion in Section 12.3). Note that in some cases, the great variation in Oré’s orthographic representations makes arriving at an adequate normalization a difficult task. For the sake of transparency, we list all of Oré’s orthographic representations in column (ii).

puquina (ii) Oré’s various orthographic representations (between chevrons), along with their token counts in the Rituale. Curly braces are used to indicate the orthographic context of a preceding or following segment where a character is relevant for both morphemes (e.g. in the sequence . Square brackets are used for morphemes that have been modified via regular morphophonemic processes (e.g ‘saying,’ ‘did you steal?’). (iii) 134 corresponding Kallawaya terms (in italics) and glosses (between single quotes), in cases where we deemed there to be a close enough formal and semantic resemblance to the terms in the Rituale. These are from the comparative dataset compiled by Hannß (2017), and are presented in her orthography; we left out some of her analyses, and proposed some new correspondences not mentioned in that work. (iv) Correspondences in Quechuan, Aymaran, and UruChipaya languages. Note that some of these terms are more widespread in the languages of the region, but we do not trace all of those connections in this chapter. This presentation draws from the etymological insights of Torero (2002), Hannß (2017), and Cerrón-Palomino (see citations in Section 12.11), as well as the comparative Quechuan–Aymaran dataset compiled by Emlen (2017b; see also Emlen and Adelaar 2017; Emlen and Dellert 2020). For Quechuan terms, we list the Southern Quechua (SQ) form whenever possible. However, many Quechuan terms are also attested broadly across that family. When Quechuan terms are attested in at least one source in each of the three accepted branches of the family— Quechua I (Central Quechua), IIB, and IIC (Southern Quechua)—we designate them in the wordlist as pan-Quechuan (pan-Q). Similarly, we list Aymara (A) terms as they are attested in those languages, but we also designate them as pan-Aymaran (pan-A) if they are attested in both branches of that family (i.e. in Aymara, from the Southern Aymaran branch, and Jaqaru, from the Central Aymaran branch). Note that for pan-Quechuan and pan-Aymaran forms, the forms cited in the wordlist may be different in the other languages from each family. Designation as pan-Quechuan and pan-Aymaran should not be taken to mean that these forms are universally attested across those families, nor that those terms can necessarily be reconstructed in the respective protolanguages. Other abbreviations used in this section are CB (Cuzco-Bolivian Quechua), CQ (Central Quechua), Ay (Ayacucho Quechua), SM (San Martín Quechua), Ec (Ecuadorian Quichua), Ch (Chipaya),

and U (Uru). Central Quechua data come from Ancash (Parker 1976) and Junín (Cerrón-Palomino 1976a), and Ecuadorian and Northern Quechua data come from Ecuador (Cordero 1992[1895]) and San Martín (Park et al. 1976). Southern Quechua data come from Academia Mayor de La Lengua Quechua (2005), Lira (1941), and Lara (1971). Aymara data are from Bertonio (1612b) and Huayhua Pari (2009). Jaqaru data are from Belleza Castro (1995). Uru–Chipaya correspondences are based in part on those identified by Torero (2002) and Hannß (2017), but we only list Chipaya terms if they are also attested in Cerrón-Palomino and Ballón Aguirre (2008). In addition, the following items are from other sources: Uru aps-ma ‘grandmother’ (La Barre 1941: 519) and upla ‘aunt’ (Uhle 1894), as well as Ayacucho Quechua wakna ‘that thing, that matter;another thing’ (Parker 1969a: 212).39

12.10 Text sample: Pater Noster (Text R) (97) {señ iki hanigo pakas-kuna-na ascha-eno our father high world-pl-loc be-ag po men upalli-so ka-Ø-anta} 2.poss name adore-ptcp be-3.sbj-opt ‘Our father, who is in the high places, may your name be worshipped.’ (98) {po qapaq ascha-no señ-guta 2.poss mighty be-inf 1pl-all wachu-Ø-anta} to.come-3.sbj-opt ‘May your greatness come to us.’ (99) {po hata-no kallaka-so ka-Ø-anta 2.poss want-inf do-ptcp be-3.sbj-opt kigo-ri hanigo pakasna How-? high world-loc ehe and

kuha ko like this

pakas-na-hamp} world-loc-add

‘May your will be done, in this world just like in the high world.’ 39 We are grateful to Rodolfo Cerrón-Palomino for his detailed and insightful comments on an earlier version of this word list.

373

emlen, mossel, van de kerke, adelaar

Table 12.13 List of Puquina lexical roots in the Rituale Puquina (normalized)

Orthographic variants

a- ‘to say’

(118), (23), (12), (4), (1)

aki- ‘to beget’

(1)

akro- ‘to leave, let’

(2)

allka-‘to be silent’

(1)

allka- ‘to be silent’ (CB)

ama ‘not’

(41)

ama ‘not’ (CB; pan-Q), ana ‘not’ (Ch)

ana ‘potato’

(1)

ani ‘cloth’

(2)

apa ‘not’

(55), (11), (1), (1)

apisa ‘brother of mother’

(1)

apu ‘lord’

(6), (1)

ascha- ‘to be’

(37), (12), (3), (2), (2), (1), (1), (1), (1), (1), (1)

asi ‘before’

(2)

ata- ‘to ask’

(1)

ata- ‘to judge’

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

atago ‘woman’

(44), (1), (1)

atasi ‘wife, woman, female’

ata-lo ‘mother’ (U)

ati-, atipa- ‘to be able’

(5), (5)

ati ‘possible’

ati(pa)- ‘to defeat, endure, be able’ (CB, A; pan-Q, pan-A), atip- ‘to beat in a competition’ (Ch)

awqa ‘enemy’

(3)

awka ‘enemy’

awqa ‘enemy’ (CB, A; pan-Q, pan-A, Ch.)

cha- ‘to call’

(1)

ch’a- ‘proclamation; to shout, lament, appeal’

374

Kallawaya

Other languages

ajna- ~ ajne- ~ ajni- ‘to abandon, pass, leave, distribute, forget, loose, free, separate’

ek- ‘to leave, abandon’ (Ch)

an ‘potato’

apasa ‘daughter-in-law’

asha ~ asuna ‘daughter-in-law’ (U) apu ‘lord’ (A; pan-Q)

acha- ‘to be (etc.)’

ata- ‘to inquire, consult, investigate, ask, interrogate’

asa- ‘to ask’ (U) atamasi- ‘to sue’ (A)

puquina

cha- ‘to punish, to beat (?)’

(3)

ch’aji- ~ ch’aje- ~ chajna- ‘to pound; to crush, destroy’

chakar ‘phenomenon (?)’

(1)

chakata- ‘to crucify’

(1)

chama ‘force’

(1)

chama-na- ‘to rule, dominate’

ch’ama ‘strength, energy’ (A; pan-A), ch’ama ‘hard-working’ (CB)

chama- ‘to enjoy’

(1)

chama ‘joy, happiness, applause, entertainment, enthusiasm, celebration, jubilation, congratulations, recreation; to make happy; to celebrate’

chama ‘elation, joy, happiness for some triumph’ (CB)

chani ‘value’

(3)

chani ‘portion, ration’

chani ‘price, worth’ (CB, A; pan-Q)

chapa ‘sexual organ’

(2)

chaske ‘relative (daughter of woman?)’

(1)

chata ‘mountain’

(2), (2)

chata- ‘to denounce’

(1)

cheka ‘nine’

(2)

cheqa ‘real’

(17), (1), (1)

cheqna- ‘to hate’

(2), (1), (1)

chexa- ‘shame; to be ashamed’

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

chichun ‘six’

(5)

china ‘star’

(1)

chia ~ chiya ‘star’

intichillay ‘shiny and vibrant glare of the sun’ (CB), qatachilla ‘a nebulous star in the Milky Way, or the stars above the nebula’ (A)

chinqi- ‘to have sex between women’

(1)

chinqi ‘female genitals’

chinqi ‘vagina’ (A; pan-A)

chok- ‘to offend’

(2), (1), (1), (1)

ch’ojna ~ ch’ojnana ~ chojna ~ chojña ~ chujna ~ ch’ohna ‘to flatten, slap, punch, punish, fight, combat, blow, beat, quarrel’

chhuxri ‘wound, bruise; wounded person’ (A)

chakata- ‘to crucify’ (CB; pan-Q)

ch’ata- ‘to denounce’ (CB, A; pan-Q) chiqa ‘real, true’ (CB; pan-A; Ch)

cheqa- ‘lesson, punishment’

Continued

375

emlen, mossel, van de kerke, adelaar

Table 12.13 Continued Puquina (normalized)

Orthographic variants

Kallawaya

chu ‘s/he, his/her’

(67), (1), (1)

chu ‘s/he, it’

chu’uñ ‘himself/herself ’

(4), (1), (1), (1)

ekitoch(kineno) ‘affine’

(2), (1), (1)

ehe ‘and’

(21), (2), (1)

eheku ‘like that’

(3), (1)

ekino ‘male progenitor’

(1)

enahamp ‘perhaps’

(2)

enahata ‘rather’

(4), (1)

enan ‘more’

(1)

enki ‘adolescent (?)’

(1), (1)

ento ‘bad’

(14), (1)

es- ‘to go’

(1), (1)

esku- ‘to put cloth’

(1)

etkiri- ‘to untie (?)’

(1)

garapa- ‘to help’

(1)

ghara ‘force (?)’

(1)

gio ‘sister (of woman ?)’

(1)

gora ‘world, time’

(2), (2), (1), (1)

guño ‘enemy (?)’

(2)

ha ‘yes’

(2)

haha- ‘to laugh (?)’

(1), (1)

haka- ‘to live, sustain’

(2), (1)

376

Other languages

inkila ‘adolescent’ isna- ‘to go, travel’ isi- ‘fabric, clothing, blanket type’ (A; pan-A), zkiti ‘clothing’ (Ch) rapa- ‘assistant; to help’

yanapa- (CB, A; pan-Q), yanap- ‘to help’ (Ch)

haka- ‘to cause, provoke; to start, life, live’

haka- ‘to live, exist’ (A; pan-A)

puquina

halla- ‘to die’

(24), (2), (1), (1), (1), (1)

jalla- ‘to die’

hampi- ‘cure; to cure’

(3), (1), (1)

hamp’i- ‘to cure’ (CB, A; pan-Q, pan-A)

hamu ‘thing, various’

(2), (1), (1), (1)

hamu ‘manner’ (CB, A)

hanigo ‘high’

(13), (3)

jana ‘high’

hanan, hanaq ‘high, above’ (CB; pan-Q)

hari ‘falsely’

(5)

kairo ~ qairo ‘sham, irreal, lie’

k’ari ‘lie’ (A)

harka- ‘to obstruct’

(1)

jirk’a- ‘to stop, stem’

hark’a- ‘to impede, block, detain, obstruct’ (CB, A; pan-Q, pan-A)

hata- ‘to want’

(22), (5), (1)

jata- ~ jat’a- ‘caress; to desire, like, want, love’

hatawallpa ‘chicken’

(1)

atawallpa ‘hen, chicken’

haya ‘child’

(17), (10), (1)

he ‘llama’

(2)

jee ‘llama’

hi-, hiqi-, hiye- ‘to give’

(10), (5), (3), (2), (1), (1), (1), (1)

heka- ~ jeka- ‘to reach, give, donate, equip, endow, send, burden, impose, offering, sacrifice, provide, obtain, supply, serve’

hila- ‘older; to enlarge’

(3)

julasma- ‘to grow, increase’

hilaqata ‘lord, hilaqata’

(2)

hilaqata ‘Indigenous political leader (second after cacique)’ (A)

hinanti ‘totality’

(7), (1)

hina-nti-n ‘all together (CB)’

hipa- ‘to be worthy (?)’

(1)

hisi ‘month’

(1), (1), (1)

hiski ‘eye’

(1), (1)

(ata)wallpa ‘chicken’ (CB, A; pan-Q, pan-A); kerwallpa ‘male chicken?’ (Ch.)

hila- ‘older; to grow, exceed’ (A; pan-A)

hi:s ‘moon, month’ (Ch) chej ‘eye(s)’

cheq- ‘to scrutinize, observe’ (Ch), ćhhuki ‘eye’ (Ch) Continued

377

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

histikga- ‘to hide’

(2), (2), (2), (2), (1), (1), (1), (1), (1), (1), (1), (1)

hit ‘between, half (?)’

(1)

ho- ‘to end, go’

(6)

eje ~ ejo ‘to end, finalize, terminate’

hor-holla-, ‘to speak’

(13), (12), (8), (6), (6), (4), (4), (2), (1), (1)

juya- ‘verse, speech, language, word, report; to talk, say, speak, modulate, narrate, tell, report’

hoto- ‘to fall’

(1)

hucha, huchacha-, huchalliku- ‘sin; to sin’

(98), (8), (1), (1), (1), (1)

jucha ~ juch’a ‘sin; to sin’

hucha ‘sin; to sin’ (CB, A; pan-Q, pan-A)

huksto ‘one, single’

(17), (3), (3), (3), (2), (1), (1)

ujsi ~ uksi ‘another, one’

huk ‘one’ (CB; pan-Q)

hulla ‘badly (?)’

(1), (1), (1)

huntu ‘very’

(9), (2)

ika- ‘to carry’

(1), (1), (1)

ika-na ~ jika-ku-na ‘to take, carry, transport’

ikha- ‘to drive cattle’ (J)

iki ‘father’

(35), (9), (3), (1)

ikili ~ ikile ‘father, friend (fig.), brother, dad’

iko ‘value (?)’

(1)

ille- ‘hot, to burn’

(1)

jilis ~ jillis ‘burning, fever, hot, heat’

imi ‘mother’

(10), (7), (4), (1)

mili ~ miilí ‘mother, mummy’

inch[a]- ‘to wound (?)’

(1)

inke ‘or’

(2), (2), (1)

inti ‘sun’

(2)

378

Kallawaya

Other languages

inti ‘sun’ (CB, A; pan-Q, pan-A)

puquina

iski ‘daughter (of man), sister (of man)’

(1), (1), (1)

isu ‘house’

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

ita- ‘to obey’

(1)

iti-, itlla- ‘to receive; to hold’

(17), (4), (2), (1), (1), (1), (1), (1)

iti- ~ ithi-ku-na ~ jiti-ku-na ‘to take, take root, stick, seize, hold up, possess, pin, hold in place, touch’

ka ‘now’

(4), (4), (1), (1)

khii ‘today, this day, recent’

kanan ‘now’ (CQ)

ka-, kaha- ‘to be’

(12), (8), (5), (1), (1), (1), (1)

ka- ‘to be’

kanka- ‘to be’ (A), ka- ‘to be’ (CB; pan-Q)

kachia ‘holy place’

(1)

kachikchi- ‘to hide (?)’

(1)

kachu ‘last’

(1)

kachu ‘world, time’

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

kachuntu ‘end of the world’

(2)

kagu ‘like’

(1)

kagu ‘self ’

(5), (1)

kaha- ‘debtor; to be in debt’

(3), (1), (1)

kaja ~ khaja ‘to owe, charge, pay for, debt, lend’

kax ~ kaxa ‘debt(s)’ (U), qaj(a) ‘debt’ (Ch)

kaka- ‘to regret’

(3)

kakohuch ‘and (?)’

(1)

kalla-, kallaka- ‘to do, make’

(9), (5), (3), (2), (1), (1), (1)

kam ‘blood’

(5), (1)

kama, kamachi‘order; to order’

(17), (5), (1), (1)

usi ‘house, building, public’

kambru ~ kampro ~ q’ambro ‘coloured, red; blood’ kama-chi- ‘to order’ (CB; pan-Q), kamachi- ‘to order’ (A) Continued

379

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

Kallawaya

kamen ‘day’

(11), (4), (1), (1)

kaman ~ qaman ‘day, time’

kaña- ‘to bring (?)’

(1)

kaña ‘to exist’

(13), (1)

kapa ‘three’

(11), (1), (1)

kapi ~ khapi ~ kjhapi ~ qapi ‘three’

kara- ‘to cry’

(2), (2), (1)

khowi ~ k’owi ~ qoni ‘tear’

kasi ‘sweet (?)’

(1)

k’osi ~ k’usi ~ q’osi ‘sweet’

kaski tikona ‘lord, authority’

(1), (1)

tikon ~ tikun ‘one hundred’

kasuske ‘and’

(3), (1), (1)

kata- ‘to hear’

(12)

kataña ‘breast’

(1)

katar ‘that’

(1)

katir ‘but’

(1)

kato ‘inner’

(4), (2), (1), (1), (1), (1)

katu ‘interior, deep, profound’

kaya ‘gold’

(1)

khaya ‘gold’

khaya ‘amulet’ (CB)

kenu ‘before’

(3)

ket, ketka- ‘happy; to be happy’

(2), (2), (1)

ki, kiñ ‘what’

(42), (12), (1), (1)

khi ~ khi’i ‘to be; then, so, for’

khiti ‘who’ (A)

ki’illa- ‘to think’

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

kee- ‘to imagine, logic, meditate, notion, think, sense’

kich, kichka- ‘to live’

(1), (1), (1), (1)

kichu- ‘sad; to be sad’

(8)

kigo ‘how’

(10), (3), (3), (2), (1), (1)

kiki ‘self ’

(19)

380

Other languages

chhep ‘three’ (Ch)

kata- ‘to hear’

kichu- ‘to be sad’ (A) khi-chay ~ khichaynataj ~ kichayna ‘how, like’ kiki ‘self ’ (CB; pan-Q, pan-A)

puquina

killchi- ‘to divide’

(2)

kina ‘eight’

(2)

kiña- ‘to go, leave’

(1), (1), (1), (1)

kinar ‘when’

(1)

khi ~ khi’i ‘to be; then, so, for’

kiñcha ‘why’

(2)

khi ~ khi’i ‘to be; then, so, for’

kiñcha- ‘to do what’

(1), (1)

khi ~ khi’i ‘to be; then, so, for’

kiogeo ‘any’

(1)

kiru ‘belly’

(2), (2), (1)

k’ere ~ keri ~ kheri ~ kjheri ~ qeri ‘belly, stomach’

kistis ‘why (?)’

(1)

khi ~ khi’i ‘to be; then, so, for’

kitkay ‘how much’

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

khi ~ khi’i ‘to be; then, so, for’

ko ‘brother, male parallel cousin’

(5), (2)

sejo ~ sejjo ‘friend, comrade, cast, lineage, colleague, compatriot, companion, fellowcountryman’

ko ‘penis’

(1)

ko ‘this, that’

(49), (2), (1), (1), (1)

ko- ‘to go’

(8)

koa ‘idol’

(4)

koha- ‘to see’

(6), (3), (2), (2), (2), (2), (1), (1), (1), (1)

kokasuske ‘and’

(5), (4), (1), (1)

kolla- ‘to hope, wait’

(2), (1)

koma, komen ‘all’

(18), (1), (1)

kumu ~ komu ‘global, all, everything, total’

konse ‘dog’

(1)

komsi ~ qomsi ‘dog, bitch’

kopuski ‘family (?)’

(1)

kho- ‘to go, march, walk, travel’ kho(a)ja- ~ kjhoha- ~ khura- ‘to look, see, watch’

qhawa- ‘to look’ (CB)

Continued

381

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

kora- ‘to contact’

(2)

kox ‘last’

(1)

kuha ‘like’

(23), (12), (3), (2), (2), (2), (1), (1), (1), (1), (1)

kuha- ‘to explain (?)’

(1)

kuma- ‘to stay’

(3), (1), (1)

kupi ‘right’

(3)

kurinba ‘food’

(1)

kusi ‘happy’

(1), (1)

kuy ‘female affine’

(1), (1), (1), (1)

kuya-, kuyapaya- ‘to love’

(5), (3), (2), (1)

khuya(paya)- ‘to love’ (CB, A; pan-Q)

llaki- ‘sad; to be sad’

(7)

llaki- ‘to be sad’ (CB, A; pan-Q, pan-A)

llulla- ‘lie; to lie’

(3), (2)

llulla- ‘to lie’ (CB, A; pan-Q, pan-A)

lluqi- ‘to hit with stick (?)’

(1)

ma- ‘to drink (?)’

(1)

macha- ‘to become drunk’

(2), (1)

maku ‘lord, authority’

(2)

mallku ‘leader’ (A), mallku ‘protector deity’ (Ch)

masi ‘mate’

(9)

masi ‘companion’ (CB, A; pan-Q, pan-A)

mata- ‘to wash’

(2)

mata- ‘to clean’

mati ‘pure, clean’

(1)

machi ‘clean, tidy, neat’

men ‘name’

(5), (5), (1)

mini- ~ mini- ‘name; to name’

meñ ‘person’

(15), (12), (11), (9), (5), (2), (2), (1), (1)

mee ~ mii ~ mil ‘people, individual, person, courtier’

382

Kallawaya

Other languages

kuma- ~ koma- ‘to live (in), place, seat, reside, sit, dwell’ kupi ‘right side’ (A; pan-A) k’usla ‘delight, pleasure’

macha- ‘to inebriate’

kusi- ‘to be happy’ (CB, A; pan-Q, pan-A)

macha- ‘to become drunk’ (CB, A; pan-Q, pan-A)

puquina

miha- ‘to sleep’

(3), (2), (2)

meja- ‘to lie down, sleep’

miqha- ‘to fornicate’ (A)

misto ‘poison (?)’

(1), (1)

moha- ‘hand; to hold, assemble’

(4), (3), (3), (1), (1)

mowe- ‘to stand (?)’

(1)

muchu ‘adolescent (?)’

(3), (1)

much’u ‘young woman’ (A)

mutu- ‘to suffer’

(7)

mutu- ‘to suffer’ (A; pan-A)

ñakari- ‘to suffer’

(1)

ñak’a(ri)- ‘to suffer’ (CB, A; pan-Q)

ni ‘I’

(14), (1)

no ‘my’

(19)

nu ‘who’

(25), (2), (1), (1)

oqo- ‘to drink’

(3), (3), (2)

ch’oqo- ‘to drink’

oxa- ‘to eat’

(2), (2), (1), (1), (1), (1), (1), (1)

oja- ‘meal, food; to eat, feed’

pacha ‘place’

(3)

pacha ‘place, time’ (CB, A; pan-Q, pan-A)

pachamama ‘mother earth’

(1)

pachamama ‘mother earth’ (CB, A)

paha- ‘to be absent’

(1), (1), (1)

pakas ‘world, time’

(15), (4), (3), (2), (1), (1), (1), (1)

pakas ‘field, estate, piece of land, region, earth’

pampa ‘land’

(1)

pimpi ‘plain, pampas, surface, plane, ground’

pampacha- ‘to forgive’

(11), (4), (1)

paña- ‘to throw out’

(5)

moxsa ‘mass rally, people’s assembly’

nichi ~ nisi ‘I’

pampa ‘flat place’ (CB, A; pan-Q, pan-A) pampacha- ‘to forgive’ (CB, A; pan-Q)

Continued

383

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

panta- ‘to err’

(2)

panta- ‘to err’ (CB, A; pan-Q, pan-A)

paqari- ‘to appear’

(3)

paqa(ri)- ‘to appear, be dawn, be born’ (CB, A)

paqta ‘beware!’

(4)

paqta ‘beware, perhaps’ (CB, A; pan-Q)

para ‘river’

(1), (1)

pata- ‘to go, leave’

(1), (1), (1)

patma- ‘to divide’

(1)

pe ‘before, last (?)’

(1)

pig ~ pik ‘forward, before, previous, first’

pego ‘first, former’

(2), (1)

pig ~ pik ‘forward, before, previous, first’

penqa- ‘shame, to be ashamed’

(2)

p’enka ‘shame’

p’inqa- ‘to be ashamed’ (CB, A; panQ, pan-A)

pesk ‘first’

(8), (3)

pi ‘you’

(39)

picha- ‘to clean’

(1)

picha- ‘to tidy, clean’

picha- ‘to clean’ (CB, A; pan-Q, pan-A)

pip ‘flesh’

(7), (1), (1), (1), (1)

pipi ‘flesh, meat, muscle’

po ‘your’

(137)

poqe- ‘to fulfill’

(1)

psami ‘female progenitor’

(1)

pta- ‘to lose’

(1)

puchu ‘rest’

(2)

puchuka- ‘to complete’

(3)

puchuka- ‘to fulfill’ (CB; pan-Q)

puka ‘red (?)’

(1)

puka ‘red’ (CB; pan-Q)

pukaw ‘full’

(1)

put- ‘to be sad’

(3)

384

Kallawaya

Other languages

waran ‘muddy river’

phatma- ‘to divide in parts’ (CB; pan-Q)

phoko- ‘height, full, ripe, ready, quorum, level, stuffing, limit, end; to fill’

puqu- ‘to mature, ripen’ (CB, A; pan-Q, pan-A), poq- ‘to ripen’ (Ch) aps-ma ‘grandmother’ (U)

puchu- ‘to remain’

phoko- ‘height, full, ripe, ready, quorum, level, stuffing, limit, end; to fill’

puchu- ‘rest; to be left over’ (CB, A; pan-Q)

puqu- ‘to mature, ripen’ (CB, A; pan-Q, pan-A), poq- ‘to ripen’ (Ch) phuti- ‘to be sad’ (CB, A)

puquina

puta ‘all’

(6), (1), (1)

qalla-, qallara- ‘to begin’

(2), (1)

qapaq ‘mighty’

(4), (1)

qasi ‘in vain’

(7), (1)

qatu ‘market place’

(1)

qhatu ‘market place’ (CB; pan-Q)

qespi-, qispichi- ‘to be safe, save’

(14), (1), (1)

q’ispi- ‘to be safe’ (CB, A; pan-Q)

qisinu ‘night’

(1)

qollana ‘superior’

(7)

quchu- ‘song; to sing’

(1), (1), (1)

q’ocho- ‘to sing verses’

raago ‘man’

(34), (2)

laaja ~ laja ‘man, male’

re ‘field’

(3), (2)

rei ~ rej ~ rey ‘field, land, terrain’

reega ‘witch’

(2), (1)

rek’a-ku-na ~ reka-na ‘wizard; to hypnotize’

rirpu ‘mirror’

(3)

ripsu ‘moment’

(1), (1)

ro-, roga- ‘to have sex’

(10), (2), (2), (1), (1), (1)

lluchka- ~ llochka- ~ lluka- ~ ijochka, loskakuj ‘coitus, sexual intercourse; to cohabit, fornicate, commit adultery; immoral’

rulli ‘female (cross?) cousin’

(1)

rullin ‘in-laws’

runa ‘day’

(3), (1), (1)

jona ‘day’

hunaq ‘day’ (CQ)

rutu ‘male (cross?) cousin’

(1)

saga- ‘to hide (?)’

(1)

saja ~ sajjai ~ chaka ‘to fence in, enclose, close, lock in, wallup, wall in’

saxa- ‘opening in the earth, crack, cave’ (A)

sallasi ‘family (?)’

(2), (1), (1)

kalla- ‘origin, introduction, preface; to inaugurate, initiate, start, begin’

qalla(ri)- ‘to begin’ (CB, A; pan-Q, pan-A), qallti ‘to begin, start’ (Ch) qhapaq (CB); qhapaqa (pan-A)

kase ~ khaje ~ k’aje ‘vainly, vain, gratis’

qhos ‘night’

qasi ‘in vain’ (CB, pan-A)

‘fish species’ (A) (cf. Cerrón-Palomino 2012: 279) qullana ‘superior’ (CB, A) q’uchu- ‘to sing’ (A)

layqa ‘sorcerer, wizard’ (CB, A, CQ, Ch) rirpu ‘mirror’ (CB, A)

saya ‘moiety’ (CB) Continued

385

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

Kallawaya

sama- ‘to give birth’

(6)

sama- ‘to give birth’

sapa ‘each, single’

(15), (3)

sapa ‘each, single’ (CB; pan-Q)

saqa ‘pimp’

(2)

saqha ‘pimp’ (A)

sasi- ‘to fast; abstinence’

(2)

sasi- ‘to fast’ (CB, SM)

se’e ‘heart’

(13), (3), (3), (3), (2)

sekka ‘valuable item’

(1)

señ ‘we, our, us’

(58), (5), (3), (2)

si- ‘to know’

(7), (2), (2), (2), (1)

sipi- ‘to beat’

(1), (1)

sipi- ‘to kill, butcher’ (CB, A; pan-Q)

sipsika- ‘to spread rumors’

(3)

sipsika- ‘to spread rumors about someone’ (CB), sipsik’a- ‘to spread rumors about someone’ (A)

skalli- ‘to be afraid’

(2), (1)

skama- ‘to be sick’

(2), (1), (1), (1), (1), (1)

kamin ‘illness, sickness; ill, sick’

skata ‘ten’

(3), (2)

katu ~ k’ata ~ k’ato ~ k’atu ‘high, ten, enormous, great’

skoman ‘pregnant (?)’

(2)

sku ‘son (of man)’

(9), (2)

kea ~ qea ‘baby, rearing, female, son, infant, offspring, baby’

so ‘two’

(5), (4)

so ~ so ~ su ~ su ~ suyu ~ suyo ‘two’

soknu ‘herb (?)’

(1)

soke- ~ socke-y ~ soq’e- ‘drug, medicine, remedy; to cure’

sper ‘four’

(5), (1), (1)

pil ~ pili ~ pilli ‘four’

sqana- ‘silver; to pay’

(6), (1), (1)

kena ~ khena ~ k’ena-; qena‘money, coin, nickel, brass (coin), silver, wealth; to pay, distribute, reward, gratify; pouch, bag’

sto- ‘to add’

(1)

stogun (?) ‘night’

(1), (1)

stu ‘seven’

(5), (4), (1)

386

Other languages

sau ~ saw ‘heart, breast’

seka- ~ seqa- ‘to learn, find out about something, know’

tutin ‘seven’

ziz- ‘to know’ (Ch)

puquina sua- ‘to fornicate; unmarried cohabitant’

(3)

suk- ‘to lie (?)’

(1)

suka ‘younger’ (suka iki ‘brother of father,’ suka imi ‘sister of mother’)

(8)

sullk’a ‘smaller, secondary, subordinate’

sullka ‘younger (brother)’ (A; pan-A), sullk’a ‘younger’ (CB; pan-Q)

sullu- ‘to abort pregnancy’

(2)

sullu ‘fetus’

sullu ‘fetus; to abort pregnancy’ (CB, A; pan-Q, pan-A)

suma- ‘to live’

(9), (6), (5), (2), (1), (1), (1)

supaya ‘devil’

(1)

supaya ‘devil’ (AS); supay (CB; pan-Q)

suti ‘clear’

(5)

sut’i ‘clear, logical’ (CB); suti ‘good’ (A)

suwa- ‘to steal’

(5), (3), (1), (1)

suwa- ‘thief; to steal’ (CB; pan-Q)

taga- ‘to bury’

(2)

tahe- ‘to dream’

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

teka ~ theka ‘to dream’

tak- ‘to work’

(10)

raka- ~ rajka- ~ tajra- ‘work, labor, mita (i.e. forced labor); to work, practice’

take- ‘to abuse’

(1)

takpa ‘five’

(7)

taku ‘force’

(2)

tanta ‘bread’

(3), (1)

tax- ‘to wash’

(1)

thi ‘agricultural crop (possibly maize)’

(1)

ti, tuy ‘yes’

(29), (1)

tikona ‘one hundred’

(2)

titu ‘moment’

(5), (1), (1), (1)

tok- ‘to touch (?)’

(1), (1)

t’anta ‘bread’ (CB; pan-Q); t’ant’a ‘bread’ (A; pan-A) t’ake- ~ t’aqe- ~ t’jana- ‘to wash’

t’aqsa- ‘to launder’ (CB; pan-Q), t’axsa- ‘to launder’ (A; pan-A)

tikon ~ tikun ‘one hundred’

Continued

387

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

too ‘adolescent’

(1)

too- ‘to bring’

(5), (1), (1)

tosta- ‘to hit (?)’

(1), (1), (1)

tot ‘big’

(3), (2), (2)

tuku- ‘to become’

(5), (4)

tuku- ‘to finish, become, pretend’ (CB, A; pan-A, pan-Q)

tumpa- ‘to bear false witness’

(2), (2)

tumpa- ‘to testify, accuse falsely, blame, slander’ (CB, A; pan-A, pan-Q)

tunt ‘already’

(6), (1), (1)

uba ‘good (?)’

(1)

uke ‘moon’

(1)

ukru ‘other’

(2), (1), (1)

ulli- ‘to succeed’

(2)

ulli- ‘to find, meet’

unancha- ‘sign; to sign’

(3)

unancha- ‘to guide’

upalli- ‘to adore’

(23), (2), (2), (1)

upa- ‘to kiss’

upra ‘relative (sister of father?)’

(1)

uqa- ‘to buy’

(2), (1)

urigas ‘hopefully’

(2), (1)

usu ‘girl’

(1)

uta ‘field (?)’

(1)

uta-n ‘agricultural field’

wachu- ‘to come’

(2), (2), (1)

waya- ~ wayaka- ‘appearance; to rise, go up, go down, run, race, descend, flow with, way out, float’

388

Kallawaya

Other languages thowa ‘young man’ (Ch), to: ‘young; child of less than three years’ (U)

tutu ‘great, thousand, maker, creator, monarch’

tutu ‘very large’ (A)

ok’o ~ oko ~ oqo ‘moon, month’

unancha- ‘sign; to signal’ (CB, A; pan-A)

upla ‘aunt’ (U) okha- ~ uka- ~ ucka- ‘realization; bankruptcy; to buy, transfer, expend, spend, realize, pay off, deal, sell’ ususi ‘daughter (of father), daughter’ (CB, Ec), uz ‘little girl’ (Ch)

puquina wago ‘good, valid’

(5), (1), (1), (1), (1)

wakagu- ‘to venerate huacas’

(1)

wakcha ‘poor’

(4), (1), (1)

wakina ‘part’

(1)

wakin ‘some, other’ (CB; pan-Q)

wakna ‘other’

(7)

wakna ‘that thing, that matter; another thing’ (Ay)

wanaka ‘new’

(1)

wañi- ‘to arrive’

(1), (1), (1)

wani- ‘to regret’

(4)

waqaycha- ‘to protect, respect’

(2), (1)

waqlli- ‘to distort’

(1)

waqlli- ‘to slip off, go astray, become corrupted’ (CB, A; pan-Q)

waruru ‘song’

(1)

waruru ‘song’ (A)

was-, wasu- ‘to go, go down’

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

wata ‘year’

(9), (2)

wata ‘year’ (CB, J; pan-Q, Ch)

watiqa- ‘to steal’

(1)

watiqa- ‘to stalk, spy on’ (CB, A; pan-Q, pan-A)

watu- ‘to fornicate’

(1)

wichiga- ‘to live (?)’

(1)

win ‘all’

(19), (2)

wiñaya ‘eternal’

(20), (2)

wiraqocha ‘gentleman, Spaniard’

(1), (1), (1)

wis- ‘to go’

(1), (1)

wisa- ‘to steal, take away (?)’

(1)

wak’a- ‘crack, cleft, special object or place, sacred, dangerous; to change state, turn wild’ wajcha ‘orphanhood’

wakcha ‘poor, destitute’ (CB, A; pan-Q, pan-A)

wane ~ wani ~ wana ‘new, brandnew’

wana-ku- ‘repentance, fear; to repent, reform, amend, fear’ (CB, A; pan-Q, pan-A) wakay ‘to conserve’

waqaycha- ‘to guard, preserve’ (CB, A)

waya- ~ wayaka- ‘appearance; to rise, go up, go down, run, race, descend, flow with, way out, float’

winña ‘eternity, everlasting, endlessly’

wiñay ‘eternal’ (CB), wiñaya (A) wiraqucha ‘gentleman, Spaniard’ (CB; pan-Q)

Continued

389

emlen, mossel, van de kerke, adelaar Table 12.13 Continued Puquina (normalized)

Orthographic variants

Kallawaya

xapne- ‘to open’

(1)

xumi ‘thief (?)’

(1)

xumna- ‘to become an enemy’

(2), (2)

yacha- ‘to know’

(1)

yaha ‘side’

(1), (1)

yakchi- ‘to misuse’

(2)

yalli- ‘to surpass’

(1)

yalli- ‘to exceed’ (CB, EC; A)

yana ‘servant’

(6)

yana ‘servant, companion’ (CB, A)

yanapa- ‘to help’

(2)

yanqaña ‘in vain, for nothing’

(1)

yanqa ‘in vain, for no reason’ (CB, A; pan-Q, pan-A)

yapo ‘agriculture (?)’

(1)

yapu- ‘agricultural plot; to plow’ (CB, A; pan-Q, pan-A)

yati- ‘to know’

(1)

yati- ‘to know’ (CB; pan-A)

yuke ‘face’

(1)

yuk(i) ‘face, forehead’ (Ch)

yuku- ‘to appear (?)’

(1)

yupaycha- ‘to honor’

(2)

yupaycha- ‘to honor’ (CB, CQ, A)

yuri- ‘to be born’

(1)

yuri- ‘to be born’ (CB, A; pan-Q)

suno ~ sunu ~ suu ‘rascal, (petty) thief; to rob’

yacha- ‘to know’ (CB; pan-Q)

yuna- ~ yana(pa)- ‘to help’

(100) {ka kamen-xe ehe hiqi-suma now day-top and give-2.sbj.1.obj.imp señ-guta kamen señ 1pl-all day 1pl.poss

Other languages

tanta} bread

yana(pa)- ‘to help’ (CB, A; pan-Q, pan-A), yanap- ‘to help’ (Ch)

pampacha-n-ki-nch kagu} forgive-plv-1.sbj-decl self ‘Forgive us our sins, just as we forgive those who have sinned toward us.’

‘And today, give us our daily bread.’ (101) {señ hucha-xe our sin-top pampacha-sumau kigo-ri señ to.forgive-2.sbj.1.obj.imp how-? we señ-guta hucha-cha-ska-eno-gata 1sg-all sin-fact-refl-ag-pl

390

(102) {ama ehe akro-suma hucha-guta not and let-2.sbj.1.obj.imp sin-all señ hoto-no-wa} 1pl fall-inf-ben ‘And don’t let us fall into sin.’

puquina (103) {enahata ento-nana qespi-na-sumau rather bad-abl be.safe-caus-2.sbj.1.obj.imp amen} amen ‘Rather, save us from evil. Amen.’

12.11 Literature The first linguistic analyses of the Puquina texts in the Rituale are found in de la Grasserie (1894) and Créqui-Montfort and Rivet (1925; 1926; 1927). The Ibero-Amerikanisches Institut in Berlin holds unpublished handwritten notes by Walter Lehmann (1930a; 1930b) analyzing Oré’s text (we thank Alejandra Regúnaga for bringing this to our attention). The most thorough and comprehensive sketch of the language to date is found in Torero (2002), which also contains a wordlist.40 Analysis of parts of the text and characterizations of the language are found in Adelaar with Muysken (2004) and Adelaar and Van de Kerke (2009). Faucet Pareja (2016) discusses Puquina possessive markers, and Mossel et al. (2020) offer an analysis of kinship terms found in the document. Aspects of the Puquina language derived primarily from other sources than the Rituale, in particular from onomastics, the Kallawaya lexicon, and Bertonio’s (1612b) Aymara dictionary,

are discussed by Cerrón-Palomino (2012; 2014; 2016a; 2018c; 2020a; 2020c). Given the scarcity of information on Puquina, the ethnohistorical question of where and by whom the language was spoken also forms an important aspect of the study of the language itself. Analyses of the Copia de los Curatos are presented by Bouysse-Cassagne (1975), Espinoza Soriano (1982; 2005), and Torero (1987; 2002); see Domínguez Faura (2010; 2014) for a comprehensive discussion. Other relevant ethnohistorical discussions are found in BouysseCassagne (1992; 2010), Galdós Rodríguez (2000), and Torero (1972[1970]; 1987).

Acknowledgments The authors thank Rodolfo Cerrón-Palomino, Pieter Muysken, and Matthias Urban for their thorough and insightful comments on this chapter. Thanks also to the Leiden University Centre for Linguistics for hosting the Leiden Puquina Working Group. The research leading to these results received funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation—project number UR 310/1-1) and from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 818854 —SAPPHIRE).

40 The density of the information presented in Torero’s grammatical sketch (2002) makes it difficult to digest, but it reflects a deep knowledge of the language, which he studied for more than 40 years. Our own chapter owes much to his descriptive work, and we hope that it will make his insights more accessible.

391

chapter 13

Híbito and Cholón Astrid Alexander-Bakkerus

13.1 General information Híbito ISO 639-3: hib, Glottocode: hibi1243, also spelled , , , , , , or , is an extinct language. The Híbito lived in the basin of the Rivers Bombanaje, Jelache, and Huayabamba, a tributary of the Middle Huallaga. They led a secluded life. For their living, the Híbito fished, hunted (monkeys and peccaries), and grew bananas, sweet manioc, peanuts, coca, cotton, and chonta palms. They traded coca for Spanish garments and iron (Steward 1948: 600–603). Steward also mentions that the Híbito were visited by the Jesuits around 1670, but converted by a Franciscan missionary, Joseph de Araujo, in 1676. Joseph de Araujo founded the mission of Jesús de Ochañache. He wrote a grammar, a vocabulary, and a catechism in the Híbito language. Unfortunately, these texts are lost. Subsequently, the Híbito were gathered in the missions Pajatén and Monte Sión, and, later, in the town of Pachiza. In addition to these missions, Pedro de la Mata mentions four other Híbito settlements in his Arte de la Lengua Cholona (de la Mata 1748: 249): Chillonya, Putonya, Ziumich, and Juanjuí. The names of the first three settlements are translated as ‘tree,’ ‘tree,’ and ‘palms,’ respectively. The names of the settlements Chillonya and Putonya, translated as ‘tree,’ may be compounds, consisting of the name of the tree and the lexeme ya, which connotes ‘tree, type.’ In Cholón yu meant ‘category, type.’ According to Steward (1948: 600–601), there were 205 Híbito people living in Sión and 372 in Valle in 1790, and 500 in Tocache, Lamasillo, Isonga, and Pisana in 1851. The last speaker of Híbito, Mrs Natividad Grández del Castillo, lived in Pachiza and was in her 90s in 1996. The language is extinct as of 2000. Cholón [ISO 639-3: cht, Glottocode: chol1284], autodenominated Seeptsá (Sofía Latorre, pers. comm.), was spoken in the valley of the Huallaga river and on the eastern slopes of the Andes. The habitat of the speakers, to the south of that of the Híbito, stretched out from Juanjuí in the north to Tingo María in the south (c.600km). The Cholón people were reported to be strong, cheerful, good-tempered, and cooperative. The Cholón men were good boatmen and pilots, mastering the tempestuous river with its rapids and gravel

banks. The people lived from hunting, fishing, farming, and trading. They traded their goods, especially salt, everywhere along the river (Herndon and Gibbon 1854: 136–9; Sobreviela et al. 1923[1787]; Steward 1948: 600–4). According to Poeppig (1836: 336), the Huallaga Valley is “one great bed of rock salt” (my translation from the German). It seems that the last speaker of Cholón, an old woman living in Juanjuí, died in 1993.

13.1.1 Sources Data on the Híbito language are scarce. We have a wordlist consisting of 43 items, composed by Martínez Compañón (1985[1782–1790]) (see Chapter 14 by Matthias Urban in this volume for more details on this important source), a wordlist of 33 items, published by Tessmann (1930), the names of three Híbito settlements (see Section 13.1), and six words and two expressions recorded in 1996 by myself during an interview with Mrs Aurelia Gutiérrez Cerquera, a then 76year old Cholón descendant who lived in Juanjuí. For Cholón, we likewise have a wordlist of 43 items and a wordlist of 33 items by Martínez Compañón (1985[1782–1790]) and Tessmann (1930) respectively. In 1996, two Cholón descendants, Mrs Aurelia Gutiérrez Cerquera and Mr José Santos Chapa Ponce, 76 years old, living in Valle, were interviewed. They remembered some Cholón words, mainly names of animals. Crucially, however, in 1748 Pedro de la Mata, a Franciscan friar, wrote his Arte de la Lengua Cholona, a grammar of the Cholón language. The following survey of the structure of Cholón is based on Alexander-Bakkerus’ (2005) description of the language, which is based on de la Mata (1748). All primary data on Híbito are given in the Appendix to this chapter.

13.1.2 The Híbito–Cholón relationship There has been some disagreement about the affiliation of Híbito and Cholón. Torero (1986: 533) ranks amongst the linguists who assumed that Híbito and Cholón are unrelated.

Astrid Alexander-Bakkerus, Híbito and Cholón. In: The Oxford Guide to the Languages of the Central Andes. Edited by: Matthias Urban, Oxford University Press. © Astrid Alexander-Bakkerus (2024). DOI: 10.1093/oso/9780198849926.003.0013

392

híbito and cholón Adelaar with Muysken (2004: 461–75) refute this. On the basis of lexical correspondences and the fact that Pedro de la Mata’s grammar of Cholón could be used in both Híbito and Cholón locations, Adelaar with Muysken argue that both languages must be related. As a matter of fact, at the end of de la Mata’s grammar a list of Híbito and Cholón-speaking places is appended, suggesting that the grammar was relevant for both languages. Adelaar with Muysken (2004: 461–2) establish conclusive proof that Híbito and Cholón were related to each other by comparing 19 items belonging to the core vocabulary of both languages: ‘black,’ ‘bone,’ ‘daughter,’ ‘to die,’ ‘earth,’ ‘father,’ ‘fruit,’ ‘head,’ ‘man,’ ‘mother,’ ‘one,’ ‘son,’ ‘stone,’ ‘three,’ ‘tree/stick,’ ‘two,’ ‘water,’ and ‘woman.’ They also propose the sound correspondences listed in Table 13.1. Table 13.1 Sound correspondences between Híbito and Cholón Híbito

Cholón

[ t͡ʃe], [ t͡ʃi]

[ta]

[c]

[c], [ t͡ʃ]

[a]

[o]

[o:], [u:]

[u]

Source: adapted from Adelaar with Muysken (2004: 462).

In support of Adelaar with Muysken’s (2004) evidence, the following arguments can be added:

(i) De la Mata (1748: fol. 248) declared that he used data of both Joseph de Araujo and Francisco Gutiérrez de Porres, the founder of Valle, a Cholón mission. This means that de la Mata’s grammar may include some Híbito material (Araujo’s data). However, since de la Mata called his grammar Arte de la Lengua Cholona, the major part of the work must be based on Gutiérrez’s material. (ii) Híbito and Cholón likely had an identical grammatical structure; see discussion of phrases provided by Aurelia Gutiérrez Cerquera in Appendix A. (iii) The lexical correspondences in Table 13.2 may be added to Adelaar with Muysken’s comparison of words belonging to the core vocabulary. The Cholón words are from de la Mata (1748).

Table 13.2 Lexical correspondences between Híbito and Cholón Híbito

Cholón

‘cassava, manioc’

̆ (Tessmann 1930)

‘classifier for fruits and round objects’

‘to cry’

(Martínez Compañón 1985[1782–1790])

‘to cry,’

‘to drink’

(Martínez Compañón 1985[1782– 1790])

~

‘to eat’

(Martínez Compañón 1985[1782– 1790])

‘to make eat’

‘fire’

(Martínez Compañón 1985[1782– 1790]), (Tessmann 1930)

‘heaven’

(Martínez Compañón 1985[1782– 1790])

‘sun’

‘sea’

(Martínez Compañón 1985[1782–1790])

‘tobacco’

(Tessmann 1930)

‘tongue’

(Tessmann 1930)

‘white’

, (Tessmann 1930)

‘wind’

(Martínez Compañón 1985[1782– 1790])

‘I cry’

13.2 Phonology and phonotactics As stated in Section 13.1.1, Híbito data are scarce: we only have 79 transcribed words and eight recorded items at our disposal. Due to the lack of sufficient Híbito material, we cannot really know how the language looked like, or how it may have sounded. We certainly do not have enough data for the establishment of minimal pairs whereby it would

393

alexander-bakkerus be possible to attribute a phonemic status to sounds. However, the material from Mrs Gutiérrez and the transcriptions of Martínez Compañón (1985[1782–1790]) and Tessmann (1930) invite us to take a closer look and try to make a tentative inventory of vowels and consonants. As far as Cholón is concerned, since there are no detailed recordings of spoken Cholón either, the transcriptions of Martínez Compañón (1985[1782–1790]), Tessmann (1930), de la Mata (1748), and especially the latter’s observations about the use of his symbols, and the few words spoken by both descendants, have been analyzed in order to establish a tentative sound system.

13.2.1 Hypothetical vowel inventory For Híbito, in the words spoken by Mrs Gutiérrez we can distinguish the following vowels: [i], [e], [a], [o], and [u]. Martínez Compañón (1985[1782–1790]) and Tessmann (1930) use the vowel symbols , , , , and in their Híbito transcription. They arguably represent the vowels [i], [e], [a], [o], and [u], respectively, as pronounced by Mrs Gutierrez. However, Martínez Compañón (1985[1782–1790])’s grapheme could also symbolize the vowel [u], as was usual in Latin inscriptions and in colonial documents. Besides the vowel symbols mentioned above, Tessmann used the symbols ̆ and , e.g. in ̆ ‘cassava’ and ‘tapir,’ respectively. The former possibly represents the open central vowel [ɞ] (see Alexander-Bakkerus 2005: 67–70), the latter may symbolize [ɑ]. Martínez Compañón additionally employs sequences of identical vowels in the items ‘daughter,’ ‘man,’ ‘mother,’ ‘son,’ and ‘rain.’ The sequences may indicate that the vowel is long, an interpretation that is supported by the fact that the word ‘man’ is transcribed by Tessmann with a superscript macron to indicate length, i.e. . Martínez Compañón likely doubled the vowel symbols , in closed syllables in the items , , and to avoid their interpretation as [ɛ], [ɑ], [ɔ], respectively. Martínez Compañón furthermore accents at the end of the words ‘moon,’ ‘brother/daughter,’ and ‘meat,’ likely to indicate stress on the last syllable. It is remarkable that Martínez Compañón employs both an acute accent and a grave accent. It is possible that he uses an acute accent to indicate that the vowel at issue was pronounced [a], as in an open syllable, and a grave accent to indicate a back articulation, i.e. [ɑ], in a closed syllable. Moreover, the double sequences may have been pronounced with or without an intermediate glottal stop, as [VʔV] or as [VV], respectively. The accentuation of the last segment of , possibly representing [ɑ], can also mean that the syllable

394

was closed by a glottal stop. Tessmann obviously employed the acute accent to indicate stress, as in ‘black.’ His apostrophe, closing a word, symbolizes in all likelihood a glottal stop, as in ‘dog,’ which may then be read as [ʃuʔ]. Sequences of similar vowels in Tessmann’s data at the end of the lexeme, as in ‘tapir,’ probably do not symbolize long vowels, since the vowels are not equally represented: versus . Such sequences may indicate that the sequence is disyllabic and that the vowel symbols are separated by a glottal closure: [VʔV], so that the word consists of four syllables: . This would be consistent with the fact that the item is a borrowed Quechua–Spanish hybrid of the form sachavaca, which also has four syllables and in which the last two vowels are also separated by a stop, in this case a velar stop. The correspondence between Tessmann’s derivative and the original lexeme support the assumption of a dissyllabic [VʔV] sequence. The alternation between and ‘man’ shows that length was not phonemic. Bearing in mind the qualities of the vowels as pronounced in Mrs Gutiérrez’s speech and the vowel symbols used by Martínez Compañón and Tessmann, we may assume that the Híbito vowel system was based on five vowels: [i], [e], [a] [o], [u], or six if we include Tessmann’s central mid vowel [ɞ]. This system is represented in Table 13.3. In closed syllables, the close-mid vowels and the open front vowels were presumably pronounced [ɛ], [ɔ], [ɑ], respectively. The front and the back vowels could also be long, but length was non-contrastive. Table 13.3 Tentative inventory of Híbito vowels Front

Central

Back

High

[i]

[u]

High-mid

[e]

[o]

Low-mid

[ɞ]

Low

[a]

As far as Cholón is concerned, a tentative inventory of distinctive vowels is presented in Table 13.4. For discussion of de la Mata’s vowel symbols , , , , and , their possible value, minimal pairs, and the deviant pronunciation of the vowels represented by and as [ɪ] and [ʊ], respectively, see Alexander-Bakkerus (2005: 58–72). In closed syllables the close-mid vowels and the open-front vowels were pronounced [ɛ], [ɔ], [ɑ], respectively. All vowels could

híbito and cholón be long, but length does not seem to have been contrastive (see Alexander-Bakkerus 2005: 58–72). Table 13.4 Tentative inventory of distinctive vowels in Cholón Front

Central

Back

High

i

u

Mid

ɪ

ʊ

Low

a

Source: Alexander-Bakkerus (2005: 72).

13.2.2 Possible consonant inventory With regard to Híbito consonants, in Mrs Gutiérrez’s speech the consonants [p], [b], [t], [c], [-j], [k], [s], [ʃ], [h], [m], [n], [ɲ], [ŋ], [l], [r], [w], and [j] can be distinguished. In Martínez Compañón’s (1985[1782–1790]) data, the letters , , , , , , , , ,