A Grammar of the Great Andamanese Language

Brill’s Studies in South and

Southwest Asian Languages

Series Editors
John Peterson, University of Kiel
Anju Saxena, Uppsala University

Eitorial Board
Anvita Abbi, Jawaharlal Nehru University
Balthasar Bickel, University of Zurich
George Cardona, University of Pennsylvania
Carol Genetti, University of California, Santa Barbara
Geoffrey Haig, University of Bamberg
Gilbert Lazard, cnrs & École Pratique des Hautes Études
Harold F. Schiffman, University of Pennsylvania
Udaya Narayana Singh, Visva-Bharati, Shantiniketan, India

VOLUME 4

The titles published in this series are listed at brill.com/bssal

A Grammar of the Great Andamanese
Language
An Ethnolinguistic Study

By
Anvita Abbi

LEIDEN • BOSTON
2013
Cover illustration: Strait Island, Andaman Islands. Picture courtesy of the author.

Library of Congress Cataloging-in-Publication Data

Abbi, Anvita, 1949–

A grammar of the great Andamanese language : an ethnolinguistic study / By Anvita Abbi.
pages cm. — (Brill’s studies in South and Southwest Asian languages ; 4)
Includes bibliographical references and index.
ISBN 978-90-04-23527-4 (hardback : alk. paper) — ISBN 978-90-04-24612-6 (e-book)
1. Andamanese language—Grammar. 2. Andamanese language—Etymology. 3. Ethnology—
India—Andaman Islands (India) 4. Andaman Islands (India)—Languages. I. Title.

PL7501.A6A34 2013
495.9—dc23
 2013018345

This publication has been typeset in the multilingual “Brill” typeface. With over 5,100 characters
covering Latin, IPA, Greek, and Cyrillic, this typeface is especially suitable for use in the
humanities. For more information, please see www.brill.com/brill-typeface.

ISSN 1877-4083
ISBN 978-90-04-23527-4 (hardback)
ISBN 978-90-04-24612-6 (e-book)

Copyright 2013 by Koninklijke Brill NV, Leiden, The Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing,
IDC Publishers and Martinus Nijhoff Publishers.

All rights reserved. No part of this publication may be reproduced, translated, stored in
a retrieval system, or transmitted in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without prior written permission from the publisher.

Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV
provided that the appropriate fees are paid directly to The Copyright Clearance Center,
222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.
Fees are subject to change.

This book is printed on acid-free paper.

To Peter Austin
 CONTENTS

Acknowledgements ......................................................................................... xv
Preface . ............................................................................................................... xvii
List of Maps, Figures and Tables ................................................................ xxi
Abbreviations and Symbols .......................................................................... xxv

I. The Andaman and the Great Andamanese

Introduction ................................................................................................. 1
1.1 Geography and Topography . ........................................................ 1
1.2 Population . ......................................................................................... 3
1.3 Present-Day Great Andaman and the Great Andamanese .... 6
1.3.1 Great Andaman . ................................................................... 6
1.3.2 The Strait Island . .................................................................. 7
1.4 History of Language Studies . ........................................................ 9
1.4.1 Great Andamanese . ............................................................. 9
1.4.2 Study in Pre-Independent India ...................................... 12
1.4.3 Study in Post-Independent India .................................... 14
1.5 Genealogical Classification ............................................................ 14
1.6 Typological Differences . ................................................................. 16
1.7 The Last Decade . .............................................................................. 16
1.8 A Sociolinguistic Sketch of the Great Andamanese .............. 18
1.8.1 Background of the Great Andamanese Speakers ....... 20
1.8.2 Language Competence ....................................................... 22
1.8.3 Language Use . ....................................................................... 24
1.8.4 A Complex Situation ........................................................... 27
1.8.5 The Predicament .................................................................. 28
1.9 The Great Andamanese Culture: Some Observations . ......... 29
1.9.1 Naming a Person .................................................................. 29
1.9.2 Naming Seasons: The Honey Calendar . ........................ 30
1.9.3 Measuring Time in Great Andamanese . ....................... 31
1.9.3.1 Parameters of Temporal Categorisation . ..... 31
1.9.3.2 Natural Time Parameters .................................. 31
1.9.3.3 Evolutionary Period (Mythological Time) .... 32
1.9.3.4 Historical Parameters
(Pre-and Post-British Era) ............................ 32
1.9.3.5 Life Cycle (Age of a Person) ............................. 32
viii contents

1.10 The Beliefs of the Great Andamanese ..................................... 33

1.10.1 The Domain or Realm .................................................... 34
1.10.2 Major Life Forms and Ethno-Biological
Classifications . .................................................................. 34
1.11 The Biological Universe of the Great Andamanese . ........... 35
1.12 The Present Study .......................................................................... 36

II. Phonetics and Phonology

Introduction ................................................................................................ 37
2.1 Vowels ................................................................................................ 37
2.1.1 Phonemic Contrasts ........................................................ 38
2.1.1.1 Front Vowels ...................................................... 38
2.1.1.2 Back Vowels . ...................................................... 39
2.1.2 Length . ................................................................................ 39
2.1.3 Phonotactics of Vowels .................................................. 40
2.1.4 Vowel Sequences/Clusters . ........................................... 41
2.2 Semi-Vowels/Glides ....................................................................... 44
2.3 Consonants ....................................................................................... 46
2.3.1 Phonemic Contrasts: Minimal and Sub-Minimal
Pairs for Consonants ....................................................... 49
2.3.1.1 Nasals .................................................................... 49
2.3.1.2 Oral Stops ............................................................ 49
2.3.1.3 Fricatives . ............................................................ 52
2.3.1.4 Liquids .................................................................. 52
2.3.2 Phonotactics of Consonants ......................................... 53
2.3.3 Medial Consonant Clusters ........................................... 54
2.3.4 Initial Consonant Clusters . ........................................... 56
2.4 Syllables ............................................................................................. 56
2.4.1 Possible Syllable Structures . ......................................... 57
2.4.2 Constraints ......................................................................... 58
2.4.3 Length of a Word ............................................................. 58
2.5 Morphophonemics ......................................................................... 59
2.5.1 Gemination ........................................................................ 59
2.5.2 Degemination and Compensatory Lengthening .... 59
2.5.3 Homorganicity .................................................................. 60
2.5.4 Insertion of a Consonant ............................................... 60
2.5.5 Vowel Harmony ................................................................ 60
2.5.6 Vowel Lowering ................................................................ 60
2.5.7 Vowel Deletion ................................................................. 61
2.5.8 Syllable Attraction ........................................................... 61
2.5.9 Metathesis .......................................................................... 62
2.6 Acoustic Study of Problematic Sounds .................................... 62
 contents ix

III. Grammar Overview

3.1 The Structure .................................................................................. 65
3.1.1 General . ............................................................................... 65
3.1.2 Typological Background ................................................. 65
3.1.3 Core Arguments ................................................................ 67
3.2 Ambivalence of Verbs, Adjectives and Nouns . .................... 70
3.2.1 Noun Verb Ambivalence ................................................ 70
3.2.2 Adjectives as Verbs . ......................................................... 74
3.2.3 Adjectives as Adverbs . .................................................... 74
3.2.4 Other Unusual Features . ................................................ 75
3.3 Inalienability (ina), Body Division Classes and
Grammaticalisation ...................................................................... 76
3.3.1 Introduction ....................................................................... 76
3.3.2 Anthropocentrism ............................................................ 77
3.3.3 The Semantics of Inalienability . .................................. 78
3.3.4 Linguistic Manifestations of ‘Inalienability’ ............. 79
3.3.5 Body Class Markers and Other Nouns ....................... 81
3.4 The Semantic Role of Body Division Classes ........................ 81
3.5 Inalienability and its Representation in Verbs . ................... 82
3.6 Inalienability and its Representation in Modifiers ............. 84
3.7 Process of Grammaticalisation . ................................................ 85
3.8 The Status of Inalienability Markers in the Grammar:
Proclitics ........................................................................................... 87
3.8.1 Arguments for Treating Inalienability Markers as
Clitics .................................................................................... 87
3.8.2 Nature of Proclitics . ......................................................... 90
3.8.3 Clitic Sequencing .............................................................. 92
3.9 Summary and Conclusions . ....................................................... 94

IV. Word Formation Processes

Introduction .............................................................................................. 97
4.1 Affixation . ........................................................................................ 97
4.2 Attaching Clitics ............................................................................ 100
4.2.1 Proclitics .............................................................................. 100
4.2.2 Enclitics . .............................................................................. 102
4.2.3 Object Clitics ...................................................................... 102
4.3 Combination of Affixation and Clitics .................................... 103
4.4 Compounding ................................................................................. 104
4.4.1 Formation of Compounds ............................................. 104
4.4.1.1 Noun + Modifier ............................................... 105
4.4.1.2 Noun + Noun . ................................................... 105
x contents

4.4.1.3 Noun + Verb or Verb + Noun ...................... 106

4.4.1.4 Noun + Proclitic + Noun . ............................. 107
4.4.1.5 Noun + Proclitic + Verb/Modifier .............. 107
4.4.1.6 Combination of Proclitics, Affixes,
and Compounds .............................................. 108
4.4.1.7 Antonyms .......................................................... 109
4.4.1.8 Summary of Compounds .............................. 109

V. Nouns and Noun Phrases

Nouns ............................................................................................................ 111
5.1 Number and Gender ...................................................................... 112
5.1.1 Numerals .............................................................................. 114
5.1.2 Gender .................................................................................. 115
5.2 Case ..................................................................................................... 116
5.2.1 Arguments ........................................................................... 116
5.2.2 Case Markings .................................................................... 117
5.2.2.1 Intransitive Subject Marking ....................... 117
5.2.2.2 Transitive Subject Marking and Ergative/
Agentive ............................................................. 119
5.2.2.3 Object Marking ................................................ 120
5.2.2.4 Dative . ................................................................ 122
5.2.2.5 Instrumental ..................................................... 122
5.2.2.6 Ablative .............................................................. 124
5.2.2.7 Comitative ......................................................... 125
5.2.2.8 Purposive/Benefactive ................................... 126
5.2.2.9 Directional . ....................................................... 126
5.2.2.10 Locational .......................................................... 127
5.2.2.11 Genitive .............................................................. 130
5.2.2.12 Comparative ..................................................... 131
5.3 Noun Phrases ................................................................................... 131
5.3.1 The General Structure ..................................................... 131
5.3.1.1 Descriptive ........................................................ 132
5.3.1.2 Possessive .......................................................... 132
5.3.1.3 Appositional ..................................................... 133
5.3.1.4 Quantifiers . ....................................................... 133
5.3.1.5 Relativisation .................................................... 133
5.3.2 Coordinated Noun Phrases ............................................ 134
5.4 Conclusion ........................................................................................ 135
 contents xi

VI. Possession
Introduction . ............................................................................................. 137
6.1 Possessive Classification ............................................................ 137
6.2 Primary Possession ...................................................................... 138
6.2.1 Body Part Terms ............................................................ 139
6.2.1.1 Mouth Cavity (a=) . ....................................... 140
6.2.1.2 Major External Body Parts (ɛr= ~ er=) . .. 141
6.2.1.3 Extremities of the Body (uŋ= ~ oŋ=) ....... 142
6.2.1.4 External Body Products or Extension
(ɔt= ~ ut= ~ ot=) ............................................. 143
6.2.1.5 Internal Organs (e= ~ i-) . ............................ 144
6.2.1.6 Nodular or Curved Structure (ara=, ra=) . .. 145
6.2.1.7 Lower Parts of Body (o= ~ ɔ=) . ................. 146
6.2.2 Is There a Hierarchy? ................................................... 148
6.3 Kinship Terms ............................................................................... 149
6.3.1 Parallels between the Body Part Terminology and
Kinship Terms ................................................................ 150
6.4 The Twin Levels . .......................................................................... 150
6.5 Secondary Possession . ................................................................ 151
6.5.1 Juxtaposition/Compounds .......................................... 151
6.5.2 Complex Structures . ..................................................... 152
6.5.2.1 Lexical Compounding with
Class Markers . ................................................ 152
6.5.2.2 Double Marking and Clitic Sequencing . 153
6.5.2.3 Adverbial Function ....................................... 153
6.5.2.4 Double Class Markers .................................. 155
6.5.2.5 Summary .......................................................... 156
6.6 Animate vs. Inanimate Possessor and the Semantics of
Inalienability . ................................................................................ 156
6.6.1. Inanimate ......................................................................... 156
6.6.2 Alienated but Inherent ................................................ 157
6.6.3 Part-to-Whole or Part-to-Component ..................... 159
6.6.4 Inalienable Possessed Nouns: A Conspectus . ....... 160
6.7 Alienable Nouns ........................................................................... 163
6.7.1 Possessed Nouns ............................................................ 163
6.7.2 Twelve Different Varieties . ......................................... 163
6.8. Attributive Modification and Possession . ............................ 165
6.9. The “possessive” Relationship: The ‘Have’ Construction ..... 166
6.10 Conclusion . .................................................................................... 167
xii contents

VII. Pronouns, Pronominal and Object Clitics

7.1 Personal Pronouns . .................................................................... 169
7.1.1 Person and Number ...................................................... 169
7.1.2 Second Person Pronouns and Honorifics . ............. 171
7.2 Demonstrative Pronouns . ........................................................ 172
7.3 Pronominal Clitics ...................................................................... 173
7.3.1 Human Arguments (Subject and Object) .............. 175
7.3.2 Object Clitics . ................................................................. 176
7.3.2.1 Pronominal Objects ..................................... 176
7.3.2.2 Inanimate External Objects . ..................... 177
7.3.2.3 Internalised Objects ..................................... 178
7.3.3 Pronominal Clitics and Temporal Adverbs ........... 179
7.4 Interrogative and Indefinite Pronouns ................................ 180
7.5 Reflexive Forms ........................................................................... 182
7.6 Reciprocal Pronouns . ................................................................ 186
7.7 Locational Adverbial Demonstratives .................................. 187

VIII. Modification Including Temporal and Spatial Deixis

Introduction . .......................................................................................... 189
8.1 Adjectives ...................................................................................... 189
8.1.1 Characteristic Features . ............................................... 189
8.1.2 The Semantic Content . ................................................ 190
8.1.3 Attributive Adjectives ................................................... 190
8.1.4 Predicative Adjectives .................................................. 191
8.1.5 Colour Terms . ................................................................. 193
8.2 Dependency of Modifiers ......................................................... 193
8.2.1 Proclitics and Modification . ....................................... 193
8.2.2 Non-Dependent Modifiers . ......................................... 197
8.3 Manner Adverbs . ........................................................................ 197
8.4 Temporal Adverbs and Temporal Deixis . ........................... 199
8.4.1 Subjecthood and Temporal Adverbs . ...................... 199
8.4.2 Syntax of Words Indicating Time ............................. 200
8.4.3 Compounding and Lexicalisation ............................. 202
8.4.4 Multiplicity of Temporal Deixis ................................ 202
8.4.5 Celestial Bodies as Temporal Markers .................... 203
8.4.6 Temporal Categorisation and Hunting and
Gathering .......................................................................... 203
8.5 Spatial Deixis . .............................................................................. 204
8.5.1 Distance and Direction ................................................ 205
8.5.2 Vertical and Horizontal Space ................................... 206
8.5.3 Locational/Spatial Postpositions ............................... 207
 contents xiii

8.6 Body Division Classes and Adverbs ........................................ 209

8.7 Third Person Demonstrative Pronouns and Deixis ........... 209

IX. The Verb and Verb Complex

Introduction . ............................................................................................. 213
9.1 Verbal Proclitics . .......................................................................... 214
9.2 Intransitive Verbs ......................................................................... 215
9.3 Transitive Verbs ............................................................................ 217
9.4 Reflexive/Self-Directed ............................................................... 223
9.5 The Distancing of the Verbal Proclitic from its Host . ...... 224
9.6 The Causative/Applicative ........................................................ 226
9.6.1 The Morphological Causative . ................................. 227
9.6.2 Periphrastic Causatives .............................................. 230
9.7 Formative Affixes ......................................................................... 231
9.8 Tense, Aspect and Mood (TAM) ............................................. 233
9.8.1 Past Tense . ..................................................................... 233
9.8.2 Non-Past Tense ............................................................. 236
9.8.3 Aspect .............................................................................. 236
9.8.4 Mood ................................................................................ 238
9.9 The Copula Be or Jiyo . ................................................................ 240
9.10 The Verbs ‘Come’ and ‘Go’ ........................................................ 241

X. Syntactic Organisation
Introduction . ............................................................................................. 243
10.1 Word Order .................................................................................... 243
10.1.1 Adjectives and Nouns ................................................. 243
10.1.2 Genitive ........................................................................... 245
10.1.3 Case Markers ................................................................. 245
10.1.4 Numerals and Quantifiers ......................................... 246
10.1.5 Degree Words ................................................................ 246
10.1.6 Manner Adverbs ........................................................... 247
10.1.7 Temporal Adverbs . ...................................................... 247
10.1.8 Negative Verb ................................................................ 248
10.1.9 Relative Markers ........................................................... 248
10.1.10 Interrogatives . ............................................................... 249
10.1.11 Adpositional Phrases . ................................................. 249
10.1.12 Variability in Word Order . ........................................ 250
10.1.13 Deviation from the Standard SOV Pattern . ......... 250
10.1.14 Variability in Proclitics ............................................... 250
xiv contents

10.2 Coordination .................................................................................... 251

10.2.1 Nominal Coordination . ................................................ 252
10.2.2 Pronominal Coordination ........................................... 252
10.2.3 Adjectival Coordination ............................................... 252
10.2.4 Verbal and Sentential Coordination ........................ 253
10.2.5 Adversative Coordination . .......................................... 253
10.2.6 Comitative Conjunct ..................................................... 254
10.3 Negation ............................................................................................ 254
10.3.1 Emphatic Negative Coordination . ............................ 255
10.3.2 Prohibitive Negation ..................................................... 255
10.4 Interrogatives . ................................................................................. 256
10.5 Non-Finite Subordination ............................................................ 258
10.5.1 Verb Serialisation ........................................................... 258
10.5.2 Cause and Effect . ........................................................... 261
10.5.3 Conditional Clauses . ..................................................... 261
10.5.4 Action Nominals . ........................................................... 262
10.6 Clause Chaining .............................................................................. 263
10.7 Comparative Constructions ........................................................ 264
10.8 Relativisation ................................................................................... 267
10.9 Syntax of the Possessive Construction .................................... 269
10.9.1 Possessed NP as the Head ........................................... 269
10.9.2 Possessed NP as Object ................................................ 270
10.9.3 Identification Question ................................................ 270
10.9.4 Complement Phrase . .................................................... 270
10.9.5 Verbless Constructions ................................................. 270

Appendix A: Lico’s Genealogical Affiliation and Her Language

Profile ............................................................................................................. 273
Appendix B: An Acoustic Study of Problematic Laterals . .................. 275
Text: The Great Narrative of Phertajido . .................................................. 279
Bibliography ...................................................................................................... 291
Index .................................................................................................................... 297
Photographs of Contact Persons
 Acknowledgements

First of all, I have to express deep gratitude to the Great Andamanese

people, who made it possible for me to explore the moribund language
during my fieldwork from 2005 to 2009. The native speakers who helped
me collect first-hand data, especially those who spent long hours with me
on Strait Island and in the city of Port Blair are: Nao Jr., Boa Sr., Lico, Peje,
and Noe. I am very thankful to them and others for their patience and
willingness to show me the jungles, marshes, seashores, their dwellings
and burial places in the Strait, and places in Mayabunder. This enabled
me to collect a large number of relevant lexical items on location. Most
importantly, the members of the tribe accepted me as a family member
and this bond exists till date. But for the cooperation and understanding of
the Great Andamanese tribe, I could not have even begun to think of writ-
ing this grammar. Their contribution to this work is greatly appreciated.
I thank my sounding boards, Bernard Comrie, John Peterson, Balthasar
Bickel, Andrew Spencer, Tania Kuteva, Ayesha Kidwai and Pramod Pan-
dey. Their suggestions and criticism have helped me in the analysis of a
very unique language. I also wish to take this opportunity to thank Saman-
tha Goodchild, Karen Buseman and Alan Buseman, in helping me prepare
the final draft. Any errors that remain are mine alone.
I am very grateful to two anonymous referees for providing me detailed
comments and suggestions on a draft version of this book. Their laudable
comments on the first version inspired me to prepare the final copy for
publication.
I am most grateful to Andrej Malchukov for reading parts of this gram-
mar and for giving insightful comments and suggestions. The students
associated with the project VOGA deserve special thanks for assisting me
whenever required. Special mention should be made of: Pramod Kumar,
Alok Das, Bidisha Som, Abhishek Avatans, Mayank, Narayan Chaudhary
and Sandy Ritchie.
I also acknowledge the help at various levels provided by the Ministry
of Human Resources and Development (MHRD), the Ministry of Tribal
Welfare, the Andaman Adim Jan Jati Vikas Samiti (AAJVS), the Jawaharlal
Nehru University, New Delhi and the School of Oriental and African Stud-
ies, University of London, UK.
xvi acknowledgements

Books like this one cannot be completed without the assistance and
generosity of the funding agency, viz. ELAR, SOAS, University of London,
which granted me funds for the mega project Vanishing Voices of the Great
Andamanese (VOGA) www.andamanese.net and http://elar.soas.ac.uk/
deposit/abbi2006greatandamanese.
This grammar was written at three different places while I was on long
leave from my university. I would like to express my gratitude to Bernard
Comrie who provided me the right atmosphere at the Max Planck Insti-
tute of Evolutionary Anthropology, Germany where I commenced writing
this grammar. I am thankful to Alexandra Aikhenvald who invited me as
a Visiting Professor at James Cook University at Cairns, Australia so that
I could work on this grammar. I am grateful to her and Robert Dixon for
the discussion on the main tenets of the grammar of Great Andamanese.
I am also thankful to Barbara Lotz for providing me a peaceful environ-
ment at the University of Würzburg, Germany, to give the final touches
to the manuscript.
I am greatly indebted to Peter Austin of SOAS, University of London,
UK where I was associated as the Leverhulme Professor for nine months
for the provision of both physical and mental space to think and complete
the grammar. But for the help that he and his staff offered I could not have
accomplished the task at hand. I dedicate this work to him.
I will be indebted all my life to Satish Abbi, my husband, who not only
encouraged me to plunge into this venture but also accompanied me to
the field as much as possible. His help, understanding, and cooperation
kept the fire burning in me despite many official, physical and psychologi-
cal hurdles.
I feel very happy and satisfied that I could accomplish the task of writ-
ing this grammar before the journey of the language into oblivion.

Jawaharlal Nehru University

New Delhi
 Preface

This grammar is the result of a major language documentation project

Vanishing Voices of the Great Andamanese (VOGA), which was undertaken
from 2005 to 2009 in the Andaman Islands. Under the auspices of the Hans
Rausing Endangered Languages Project, I, along with a team of research
assistants spent approximately 40 months in the Andamans, document-
ing Great Andamanese and producing descriptive and theoretical work
on the language. The team members spent varying amounts of time on
the island during this four-year period, but I spent the longest as after the
initial phase of fieldwork, my research assistants could not sustain their
interest in the work primarily because of the threatening attitude of some
of the officials, the difficult living conditions in the islands and boredom.
Present-day Great Andamanese (PGA) is a highly endangered language;
when the VOGA team first visited Strait Island, the current home of the
tribe, there were nine speakers. By the time the team left the Andamans
for the last time in 2009 there were only six left. The last speaker of Bo, a
variety of Great Andamanese, also died in January 2010. At present there
are only five speakers with varying degrees of competence in the language.
PGA is a complex and diverse language; it is a koiné formed from four sur-
viving northern Great Andamanese languages: Khora, Jeru, Sare and Bo. It
draws its lexicon from all four of these dialects but its grammar is primar-
ily based on Jeru. It displays highly unusual grammatical features such as
body part proclitics modifying all parts of speech. Perhaps most impor-
tantly, PGA is a unique language; there is strong linguistic and genetic
evidence to suggest that the people of the Andaman Islands represent a
distinct genetic group who populated the islands from the mainland tens
of thousands of years ago. According to some geneticists, Andamanese
are the survivors of the first migration out of Africa 70.000 years BP. The
language is a fast-closing window on a very ancient form of cognition, or
as Nicholas Ostler puts it in his review of the A Dictionary of the Great
Andamanese Language. English-Great Andamanese-Hindi (Anvita Abbi
2012): “700 centuries of unique experience terminated in just two.”
In addition to the Great Andamanese, there are three other distinct
indigenous ethnolinguistic groups inhabiting the Andamans: the Jarawa,
the Onge and the Sentinelese. My first introduction to the tribes of the
Andaman Islands was made in 2000–2001 when I conducted a pilot survey
xviii preface

of the languages of the island titled Language Survey of Andaman sup-

ported by the Max Planck Institute of Evolutionary Anthropology, Leipzig,
Germany. The results of this survey not only opened the road to language
documentation but also paved the path to explore the grammar of the
Great Andamanese language.
All the languages spoken by these peoples are endangered as their
population bases dwindle and Hindi extends its reach ever further in the
islands. Great Andamanese is the closest to extinction, however; today
only two of the four northern varieties of the language survive. The other
two, Khora and Bo, became extinct with the passing of their last speakers
Boro Sr. and Boa Sr. respectively.
Little is known about the affairs of the inhabitants of the Andaman
Islands until the arrival of the British in the 1860s. The early colonists
reported around 5000 people living as hunter-gatherers. At that time
there were ten distinct varieties of Great Andamanese named after the
ten tribes which spoke them. An estimated 3000–3500 Great Andamanese
people lived right across the group of islands known as the Great Anda-
man. By the time of the 1901 Census of the islands, this number had fallen
to just 625. It seems that contact with mainlanders had a devastating
effect on the indigenous population, as Edward Horace Man noted: “The
interesting Negrito race inhabiting the Andaman Islands is doomed to
early extinction—save possibly the small section occupying Little Anda-
man.” (27th June 1918).
By the time of the VOGA team’s arrival in 2005, this ominous predic-
tion had sadly been borne out; the number of the Great Andamanese had
dwindled to just 51, with only ten members of the tribe still able to speak
some form of the language. When the team first encountered the people,
they found their lives further blighted by the 2004 tsunami, which drove
them from their homes in Strait Island and forced them into temporary
shelters in the Andaman capital of Port Blair. Many of the earlier record-
ings were made in this difficult period. The people were often withdrawn
and unwilling to collaborate with the VOGA team, preoccupied as they
were with dealing with the difficult situation in which they found them-
selves. In late 2005 the members of the tribe returned to their homes in
Strait Island, a small island to the northeast of Port Blair. The later record-
ings were all made there.
The Great Andamanese of today live on subsidies provided to them by
the Indian government, though some members still continue their tradi-
tions of fishing for crabs and other sea-life and hunting turtles. Gathering
tubers and potatoes, however, is now virtually non-existent. Most tribe
 preface xix

members understand Hindi, Bangla and a few words of English as well as

some words in Great Andamanese. Children under 16 now do not speak or
understand the heritage language. The main language of communication
amongst members of the tribe is the Andaman variety of Hindi.
The unusual structures of the language captured in the present gram-
mar motivate us to capture the ‘possible’ structures of human language.

SOAS, University of London, UK

 LIST OF MAPS, FIGURES AND TABLES

Maps

1. Southeast Asia .................................................................................. 2

2. Location of the Andaman Islands .............................................. 4
3. The 19th–20th century ................................................................... 5
4. The Strait Island .............................................................................. 8
5. Distribution of the Andaman tribes .......................................... 17

Figures

1.1. Two distinct language families in Andaman .......................... 10

1.2. Present Great Andamanese and its regional varieties . ....... 11
1.3. The number and age of the Great Andamanese people
(2005–2006) . ..................................................................................... 20
1.4. Number of speakers on the scale of competence ................. 23
1.5. Age-wise competence level of the speakers in 2007 ............ 23
1.6. The number and mean age of the speakers on the
competence scale ............................................................................ 24
2.1. Syllable breaks . ................................................................................ 57
3.1. Degree of grammaticalisation across grammatical
categories ........................................................................................... 86
3.2. The classificatory functions of body class markers in
PGA . .................................................................................................... 95
6.1. Body division classes and possession . ...................................... 162
6.2. Word formation processes involved in possessive
constructions .................................................................................... 164
A1. Lico’s genealogical affiliation and her language profile ...... 273
B1a–b. The spectrogram of lɛ:c ‘arrow’ as articulated by Peje . ....... 276
B2a–b. The spectrogram of lɛ:c ‘arrow’ as articulated by
Nao Jr. ................................................................................................. 276
B3. The formant structure of [l] (Peje) ............................................ 277
B4. The formant structure of [l] (Nao Jr.) ....................................... 277
xxii list of maps, figures and tables

Tables

1.1. The Andamanese languages in the 19th century ........................ 10

1.2. Comparative lexicon in Angan and Great Andamanese .......... 15
1.3. The lineage of the oldest and the best speakers . ....................... 21
1.4. Lineage of the semi-speakers ............................................................ 21
1.5. Stages of naming an individual ........................................................ 29
1.6. Honey calendar or names of the seasons . .................................... 30
1.7. Blooming of flowers and associated months of the year ......... 31
1.8. Measuring time in PGA ...................................................................... 33
2.1. Vowels of Great Andamanese . ......................................................... 37
2.2. Distribution of vowels ......................................................................... 41
2.3. Vowel clusters in PGA ......................................................................... 42
2.4. Phonotactics: two vowel sequences . .............................................. 42
2.5. Intervocalic semi-vowels .................................................................... 44
2.6. Words that end in semi-vowels . ...................................................... 45
2.7. Words that begin with semi-vowels ............................................... 46
2.8. Intra-community variation of unusual sounds . .......................... 46
2.9. Phonetic variation across community members ........................ 47
2.10. Consonants of Great Andamanese . ................................................ 48
2.11. Distribution of consonant sounds ................................................... 53
2.12. Consonant clusters within a lexeme . ............................................. 54
2.13. Geminates ............................................................................................... 56
2.14. Possible syllable structures ................................................................ 58
3.1. Seven basic zones in the partonomy of the body ...................... 80
3.2. Semantics of body division class markers and bound form
classes ....................................................................................................... 95
4.1. Spatial words and the combination of clitics and affixes ........ 103
4.2. Left-headed Compounds .................................................................... 105
4.3. Right-headed Compounds ................................................................. 106
4.4. Non-compositional compounds ....................................................... 107
4.5. Compositional compounds . .............................................................. 107
4.6. Combination of proclitics, affixes and nouns .............................. 108
5.1. Case markings . ...................................................................................... 117
5.2. Locative forms ....................................................................................... 127
6.1. Seven basic zones in the partonomy of the body ...................... 140
6.2. Body part terms with a= possessive class marker ...................... 141
6.3. Body part terms with e=-, ɛr= possessive class marker ............. 141
6.4. Body part terms with oŋ= possessive class marker .................... 143
6.5. Body part terms with ot=, ɔt=, ut= possessive class marker .... 144
 list of maps, figures and tables xxiii

6.6. Body part terms with e=, i= possessive class marker ................. 145
6.7. Body part terms with ara= possessive class marker .................. 146
6.8. Body part terms with o=, ɔ= possessive class marker . .............. 147
6.9. Frequency of occurrence of body division class markers with
body part terminology . ....................................................................... 148
6.10. Kinship terms . ....................................................................................... 149
6.11. Parallels between body parts and kinship terms . ...................... 150
6.12. Double marking and secondary possession ................................. 154
6.13. Terms for the parts of a tree or a plant ......................................... 160
6.14. Possessive class markers defining primary possession and
classes of nouns . ................................................................................... 161
6.15. Varieties of possession ........................................................................ 165
7.1. Pronominal forms in PGA . ................................................................ 169
7.2. Pronominal clitics . ............................................................................... 174
7.3. Indefinite/interrogative forms .......................................................... 181
8.1. Body division classes in adjectives .................................................. 194
8.2. Edge and side of an object . ............................................................... 197
8.3. Multiplicity of temporal deixis ......................................................... 202
8.4. Terms for direction and days of a month ..................................... 203
8.5. Names of flowers and hunting time ............................................... 204
8.6. Levels of the sea .................................................................................... 205
8.7. Landscape and coastal terms ............................................................ 205
8.8. Spatial deixis describing parts of a Great Andamanese
house . ....................................................................................................... 206
8.9. Spatial adverbs and their morphemic composition .................. 208
8.10. Body division classes designating spatial relations .................... 209
8.11. Body division classes in adverbs ...................................................... 210
8.12. Demonstrative pronouns . .................................................................. 211
9.1. The body division class markers with verbs ................................ 215
9.2. Verbs with Formative affixes, tense and mood suffixes ........... 239
9.3. Verbal affixes in PGA . ......................................................................... 241
 Abbreviations and symbols

1 first person
2 second person
3 third person
A agent
abl ablative
abs absolutive
acc accusative
adj adjective
adv adverb
agt agentive
appl applicative
c consonant
caus causative
cl 1, cl 2, etc. body division class 1, 2, etc.
class classifier
clt clitic
com comitative
comp complementiser
compr comparative
conj conjunct
cop copula
cond conditional
conv converb
d possessed
dat dative
dem demonstrative
dir directional
dis.vis distant visible
dis.invis distant invisible
dst.pst distant past
do Direct object
du dual
emph emphatic
erg ergative
excl exclusive
xxvi abbreviations and symbols

exist existential
F female
fa formative affix
gen genitive
h, hon honorific
hab habitual
imm.pst immediate past
imp imperative
io indirect object
ina inalienability
incl inclusive
instr instrument
int intimate
loc locational
M male
mod modifier
mt mother tongue
N noun
neg negative
nmlz nominaliser
nom nominal
np noun phrase
npst non-past tense
nrr.pst narrative past
O object
obj object clitic
PGA Present-day Great Andamanese
pcpl participial
pl plural
pp postpositional phrase
poss possessive
pro pronoun
prohb prohibitive
prox proximate
prox1 proximate, nearer to the speaker
prox2 proximate, nearer to the hearer
pst past
q question
R possessor
 abbreviations and symbols xxvii

rec reciprocal
rel relativiser
refl reflexive
result resultative
S subject
sg singular
Sp speaker
spec specific
stat stative
subj subjunctive
tam tense, mood and aspect
tr transitiviser
V verb
vt verb transitive
vi verb intransitive
X any entity
= clitic boundary
 Chapter one

The Andaman and the Great Andamanese

They don’t understand me. What can I do? If they don’t

speak to me now, what will they do once I’ve passed
away? Don’t forget our language, grab hold of it. Don’t
ever let it go.
Boa Sr., the last speaker of Bo.

Introduction

The Andaman Islands and their people have been the centre of interest for
a multitude of scientific and cultural reasons since their discovery in the
18th century. During the 18th and 19th centuries, researchers studied these
islands to classify and catalogue the various human races in the world,
with the underlying desire to scientifically confirm the racial superiority
of the white race over the others. The 20th century witnessed a change
in this attitude. Subsequent scientific, anthropological, linguistic and cul-
tural studies revealed that the islands and the people were unique from
an evolutionary perspective. These preliminary studies have revealed that
investigations concerning these small islands are important for under-
standing significant aspects of human evolution, migration and even the
peopling of the continents.

1.1 Geography and Topography

A cluster of approximately 324 islands and islets in the Bay of Bengal,

running from north to south and located southeast of the Indian sub-
continent, constitutes the Andaman and Nicobar Islands. The Andaman
Islands lie between 10°13’–13°30’ N latitude and 90°15’–93°10’ E longitude.
These are truly oceanic islands, never having been connected to the main-
land during Pleistocene glaciations (Ripley & Beehler 1989). They are sep-
arated from the Malay Peninsula by the Andaman Sea, an arm of the Bay
of Bengal and are part of the union territory of the Andaman and Nicobar
Islands which belong to India. The Andaman Islands are broadly divisible
into two sets of groups, Great Andaman and Little Andaman.
 2 chapter one

92○ 93○ 94○ 95○

14○

13○

Strait I.
Bangladesh 12○
China
Burma Andaman
(Myanmar) Islands
India 11○
N. Sentinel I.
Thailand Little Andaman
Bay of
Bengal A n daman 10○
S ea

Car Nicobar
Sri Lanka 9○
Malaysia Nicobar
Indonesia
Islands
8○

Great 7○
0 50 Miles Nicobar
0 50 Kilometres

Map 1. Southeast Asia

 the andaman and the great andamanese 3

The Great Andaman: the area covered by the Andaman Islands is made
up of island clusters. From north to south, the various islands are North
Andaman, Middle Andaman, South Andaman, Baratang, Ritchie Archipel-
ago and North and South Sentinel. Collectively, these are called the Great
Andaman. The close proximity of these islands to each other creates the
impression of one island, a fact that unfortunately motivated the govern-
ment to build the Andaman Grand Trunk road, which robbed the tribes
of their basic resources due to rampant deforestation. The capital city of
the Andaman Islands is Port Blair, located in the south of the islands at a
distance of 1255 km from Kolkata. Approximately 65 km south from the
city of Port Blair in the Great Andaman is the island called Little Anda-
man, which is a home to Onges.

1.2 Population

According to the 2001 Census, the total population of the Andaman

and Nicobar Islands was around 314,239, of which 203,968 constituted
the population of the Andaman Islands. However, the tsunami caused
by the earthquake in the Indian Ocean on 26th December 2004 decreased
the population of the island by approximately 7000.
The variety of people from mainland India and neighbouring countries,
such as Bangladesh and Burma, has helped in making the area plurilin-
gual, multilingual and multicultural. Most of the scheduled and some of
the non-scheduled1 languages of Jharkhand, such as Kharia, Kurux and
Mundari are spoken on the islands. It is not surprising that with all their
diverse and heterogeneous communities, the Andaman Islands are called
‘Mini India.’
The current state of knowledge of Andamanese languages has improved
significantly from what was known fifty, or even ten years ago. Of the
fourteen known Andamanese languages, only four are still spoken today:
Onge, Jarawa, Sentinelese and the Koiné form of present-day Great Anda-
manese (PGA henceforth). The PGA is breathing its last as it is a mori-
bund language. Most of the spoken Andamanese languages are in danger
of extinction due to a low population base and/or language replacement
by Hindi. The speakers of the four living languages that are mentioned

1 The Constitution of India recognizes only twenty-two languages generally termed

“scheduled languages” as they are listed under VIII Schedule, Articles 343–351 of the Con-
stitution and the rest as ‘unscheduled.’ The former are official languages of Indian states.
4 chapter one

Map 2. Location of the Andaman Islands

 the andaman and the great andamanese 5

Map 3. The 19th–20th century

above can be grouped into four major groups: the Great Andamanese, the
Jarawa, the Onge and the Sentinelese. Barring the Sentinelese, the other
tribes have been exposed to the mainlanders. Jangil, a language closely
associated with Sentinelese, and several other Great Andamanese lan-
guages have become extinct within the last eighty years.
The ten distinct languages of the Great Andaman are known by the
names of the respective tribes that speak them. The history of contact
with outsiders varies from tribe to tribe. Chronologically, the first group
to come into contact with the mainlanders was the Great Andamanese,
followed by the Onge and finally the Jarawa. All attempts to establish con-
tact with the Sentinelese have failed so far. The Jarawa came to be known
to the mainlanders only recently. It is generally believed that first con-
tact was established in 1998. The demographic scale of these islanders is
inversely related to the period of contact with the mainlanders; the longer
and deeper the contact, the smaller the population. It is reported that the
6 chapter one

population of the Great Andamanese fell from an estimated 3000–3500

in the early part of the 19th century to 625 by the 20th century (Census
1901).2 At present, the population figures of different tribes are as follows:
present-day Great Andamanese (56); Jarawa (300); Onge (105); Sentinelese
(not known). The Great Andamanese people were hunter-gatherers until
the end of the 19th century, just before their intensive contact with the
British, when the latter established a penal colony in Port Blair in 1858.
While the Jarawas are still hunter-gatherers, the other two tribes—the
Onge and the Great Andamanese—live a sedentary life, chiefly dependent
on government welfare.

1.3 Present-Day Great Andaman and the Great Andamanese

1.3.1 Great Andaman

The present-day Great Andaman is no different from any other Union Ter-
ritory of India in a judicial, political or academic sense. The difference
is the noticeably large population of ‘settlers’ from South India, Bangla-
desh and Sri Lanka. With eighteen to twenty languages represented in
the Islands it is justifiably known as ‘Mini India.’ Great Andaman has the
potential for new jobs as the development of the Islands is in full swing.
Lately some of the Great Andamanese have been inducted into the
police force and some have joined the shipping fleet, but the majority of
them stay at Strait Island, 53 nautical miles to the northeast of the city of
Port Blair. The tribes of Strait Island frequently visit Port Blair to receive
their monthly allowance, medical help and other necessary aids from the
government bodies. At present, the Great Andamanese have a dietary
habit similar to most Indians, with wheat, rice, vegetables, fish and meat
forming their staple diet. Nonetheless, the men still prefer to hunt turtles
in the sea and they also go fishing quite regularly; crab is also an integral
part of their diet. Birds are occasionally eaten. However, the gathering of
tubers and potatoes is virtually non-existent now.
The flora of the Andaman Islands is diverse and complex. The tropi-
cal forest is dense with very tall trees and thick undergrowth. As far as
marine fauna is concerned, Andaman has a rich diversity. However, the
forest fauna is restricted to varieties of birds, insects and reptiles. In spite
of the tropical forest, there is a great paucity of mammals in the Andaman

2 The website www.andaman.org quotes the figure to be much higher.

 the andaman and the great andamanese 7

Islands. Wild pig was the only large mammal until fifty years ago. There
are no monkeys, squirrels or other small mammals except bats. Deer and
dogs were only introduced recently. Among the fauna the Andamanese
recognise a large variety of birds, crabs, turtles, fish, snakes, ants and mos-
quitoes. Among the flora, the largest lexicon is that of trees and plants
which are used for either hunting, everyday use or for medicinal purposes.
It perhaps justifies the utilitarian view of the folk biology that people have
a more diverse vocabulary for those categories of biological entities which
are more useful to them. For information on these aspects, readers may
refer to A Dictionary of the Great Andamanese Language (Abbi 2012).
The Great Andamanese, especially those who visit Port Blair frequently,
have a functional knowledge of Hindi and Bangla. Some of them also
understand a few words of English. Children below the age of sixteen do
not understand or speak the heritage language. All of the members com-
municate in Andamanese Hindi among themselves. See §1.8 below.

1.3.2 The Strait Island

The area of the present research was Strait Island, a tribal reserve recog-
nised by the government of India. Strait Island is a small island of two
km² located in South Andaman, Andaman and Nicobar Islands. From Port
Blair there is a fortnightly ferry service which is the only mode of trans-
port available to visit the island. Strait Island is a small, comma-shaped,
forested island (refer to map 4 titled ‘The Strait Island’), known for its
caves of birds’ nests and deer. The settlement of the Great Andamanese
at Strait is administered by the Andaman administration (AAJVS). There
are only 8 households in total and people do not live in them permanently
as members keep travelling back and forth between Strait Island and Port
Blair. Some of the Great Andamanese members have been appointed gov-
ernment jobs in the city and have been provided housing in Port Blair.
Some members of the community who live on Strait Island prefer to
hunt in the sea for small catches. However, this is done more as a hobby
than for survival. Every Great Andamanese household is equipped with
modern gadgets such as CD players, televisions, refrigerators, washing
machines and electric irons. It is a lost society which loves to emulate
city-dwellers on the one hand and tries in vain to retain old practices on
the other. Most of the men smoke, chew tobacco and drink alcohol, which
is one of the causes of their short life span. Women do not find the men
acceptable because of these bad habits and thus, prefer to have intimate
relations outside of the tribe. This leads to further injury to the mental
8 chapter one

Map 4. The Strait Island

state of the men in the tribe. It is a vicious circle from which men find it
difficult to break out.
A large part of the island is occupied by the Public Work Department
(PWD) and the AAJVS (the Andaman Tribal Welfare Society) after the
last tsunami of 2004. There is a nursery school for children and a small
dispensary for primary health care. The island is lush green consisting of
coconut, banana, tamarind, mango, gurjan, palm, and several indigenous
trees. No vegetables are grown on the island: they are all transported from
the city of Port Blair. Solar powered equipment is seen to be used by the
Great Andamanese people. The local non-tribal population is larger than
the indigenous tribal population, a glaring fact obvious to anyone visiting
the Reserve. We made several trips to this island as the government does
not give permission to visit the island for longer than three weeks at a
time.
When we first visited Strait Island in 2001 there were nine speakers
(of PGA), among them three Jirake, the Chief (multilingual in several
languages/dialects of the Great Andamanese language family, but profi-
cient in Jeru), Boa Sr., (the last speaker of Bo) and Nao Jr. (the most flu-
ent speaker of Jeru) served us as the main repositories of information.
Unfortunately, they all died between 2005–2010. Another speaker Boro,
who came from a Khora background also died during our fieldwork.
At present, we have, in order of fluency in the language: Lico, Peje, Noe,
Surmei and Golat. They are all in the range of fifty to fifty-eight years of
age and speak the koiné variety of present-day Great Andamanese. Alas,
the language is breathing its last.
 the andaman and the great andamanese 9

1.4 History of Language Studies

Apart from Great Andamanese, the other three languages are vibrant
despite the small population base, as they are still passed on to the next
generation. Except for Sentinelese, we have fair amount of knowledge
about the Andamanese languages.
The following is a brief account of a linguistic study undertaken on the
Andamanese languages with special emphasis on the new perspectives
emerging from research undertaken in the last ten years. While discussing
the results of these investigations, we will be keeping the issues of genetic
classification and typology of the languages in mind. In this way the his-
torical analysis proceeds in tandem with the descriptive analysis.

1.4.1 Great Andamanese

Great Andamanese is a cover term that has been used for ten disparate
groups of the tribes living in the Great Andaman, each speaking a different
language but mutually intelligible to its neighbouring tribe. It constitutes
the sixth language family of India (Abbi 2006b, 2009, Blevins 2007). There
are ten languages, which can be grouped into three varieties: southern,
central and northern. There are two distinct language families in the
Andaman Islands, i.e. Great Andamanese and Angan, the latter being
associated to the Austronesian language family (Blevins 2007). Consult
Figure 1.1.
Except for Jeru and Sare3 all of the Great Andamanese languages are
now extinct. The Great Andamanese languages formed a ‘linguistic con-
tinuum’, so that each language was closely related to its neighbour on
each side but those at the extreme ends of the geographic continuum
were mutually unintelligible. Hence, Aka-Cari (Map 2), a North Great
Andamanese language, was mutually unintelligible with Âkà-Bêa, the
southern variety.
The present-day Great Andamanese language is a mixture of four
northern varieties4 with several linguistic inputs levelled to generate the
current speech: a koiné (Manoharan 1989). PGA draws its lexicon from

3 There is one speaker who claims to hail from a Sare background though she speaks
the modern version of the Great Andamanese language.
4 The last few generations of Great Andamanese speakers are descendants of inter-
marriages among North Andamanese tribes. The Government of India encouraged this
practice in order to preserve their dwindling numbers when the entire population was
settled on Strait Island.
10 chapter one

Andamanese

Western Eastern

Angan Great Andamanese [56]

Kede Kol Juwai Pucikwar Bea Bale Jeru Bo Sare Khora

extinct extinct extinct extinct extinct extinct extinct extinct

Central Western Southern Western

Onge [105] Sentinelese [?]

Jarawa [300]
Figure 1.1. Two distinct language families in Andaman (Adapted from Abbi 2003)

Table 1.1. The Andamanese languages in the 19th century

Great Andamanese Angan
North Central Southern Central-East South South-East
Aka-Kora Aka-Puchikwar Aka-Bale Jarawa Onge Sentinelese
(Khora) (Pujjukar) (Ang) (Ang)
Aka-Cari (Sare) Aka-Kede Âkà-Bêa
Aka-Jeru Aka-Kol Jangil
(Jeru/Jero)
Aka-Bo Aka-Jowoi
(Present-day pronunciations of the names by the tribes are given in brackets)

Jeru, Sare, Khora and Bo, but is primarily based on the grammar of
Jeru. As the present form is the amalgam of four languages, the author
has deliberately avoided naming this form of the language ‘Jeru’ as this
will unnecessarily place an arbitrary emphasis on one language over the
others. PGA is spoken on Strait Island. A recent study shows that it shares
a large percentage of its vocabulary with Aka-Kede, the central variety of
the Great Andamanese language (Mayank 2009).
 the andaman and the great andamanese 11

Aka-Khora

Present Great Aka-Jeru

Andamanese Northern Variety
Aka-Cari

Aka-Bo

Aka-Kede
Great
Andamanese
Aka-Jowoi
Central Variety
Aka-Pucikwar

Aka-Kol

Aka-Bale
Southern Variety
Aka-Bea

Figure 1.2. Present Great Andamanese and its regional varieties

The other languages of the islands are:

Jarawa two varieties, spoken in northern Middle Andaman and
southern Middle Andaman
Onge spoken in Dugong Creek, north of Little Andaman
Sentinelese spoken on the North Sentinel Island
Jangil became extinct by 1920
Historically, the research on Andamanese languages can be broadly
divided into two phases: pre- and post-independent India. Most of the
documentation of the extinct languages of the Great Andamanese, espe-
cially the South Andamanese languages, is available in the writings of the
British ICS officers who had training in anthropology (Man 1875–1879,
1923; Temple 1903; Portman 1887, 1898; Radcliffe-Brown 1922, 1929, 1948).
The living languages of the Great and Little Andaman Islands have
been documented by Indian linguists and anthropologists in the post-
independent era.
12 chapter one

1.4.2 Study in Pre-Independent India

Interest in the indigenous tribes of the Andaman Islands began in order
to reveal and solve the ‘mystery’ of the hitherto obscure hunter-gatherer
populations living isolated from the rest of the world for thousands of
years. Travellers and explorers, who met the fate of being shipwrecked
on the Andaman shores or who had a ‘view’ of the “savaged race having
heads, eyes, and teeth resembling those of the canine species” (Marco
Polo in Masefield 1908: 347), or who thought they were “wild because they
eat one another” (Frederike 1625: 1710),5 were prejudiced. Furthermore,
their impressions were not based on fact, nor could they comment on the
language of the tribes primarily because they were never allowed to step
foot on the island.
It was the British government officers, with high designations as Com-
missioners or Assistant Commissioners, who spent considerable periods
of time administering the region during the last three decades of the 19th
and early 20th century, who were able to unearth the mystery of the lan-
guages spoken by these tribes. The logistical proximity of South Andaman,
having Port Blair as the capital city, gave Man (1883, 1885, 1923), Temple
(1903) (the two worked in collaboration on linguistic and anthropologi-
cal analysis) and Portman (1898) the opportunity to describe the South
Andamanese languages such as Âkà-Bêa and Aka Bale in detail. The
dictionary of Âkà-Bêa by Man (1923) is an exhaustive and unparalleled
example of language documentation of one of the Southern Great Anda-
manese languages in this phase. It contains, among other items, a brief
grammatical sketch of Âkà-Bêa which brings forth the unique nature of
the language (Man 1875–1878).6 However, Portman’s (1887) A manual of
the Andamanese languages is the only document that included a com-
parative lexicon from North (Chari), Central (Kede) and Southern (Bea
and Puchikwar) languages including some sporadic words from Onge,
the language spoken in Little Andaman, and English-Andamanese sen-
tences in the four Great Andamanese languages. However, both Temple
and Radcliffe-Brown raise doubts about the authenticity of the data in
Portman’s Manual. In the absence of any other reliable records in the
North Andamanese languages one cannot be sure of the judgment passed

5 Radcliffe-Brown (1922: 8) very convincingly disputes these judgments.

6 The author had the opportunity to go through the handwritten manuscripts by Man
(1875–1878 and 1875–1879) which are, on numerous occasions, annotated and commented
on by Temple. These are archived in the Royal Anthropological Institute, London.
 the andaman and the great andamanese 13

by Temple and Radcliffe-Brown. One would like to give it the benefit of

the doubt as Portman was the only officer who spent ten years in the
Andaman and Nicobar Islands. Portman’s Notes on the Languages of South
Andaman Group of Tribes (1898) discusses some vocabulary of the Jarawa
and Bea languages and was the first to recount ground level knowledge
of Andamanese languages. A recent study of the select few lexical entries
(Mayank 2009) proved that the North Andamanese data is very close to
the present-day Great Andamanese language. Man and Portman were
drawn to the fact that the Great Andamanese languages offered very
unusual structures and thus attempted to describe them in their writings.
Portman (1899) gave a detailed description of the socio-anthropological
nature of the Great Andamanese and the Jarawa. His telling description of
the ‘Jarawa-hunt’ is a good indicator of the inhuman, insulting and intimi-
dating behaviour of British officers towards Jarawas.
Sir R. C. Temple’s works on Andamanese languages, in the Report of
the Census of India 1901, Vol. III gives a good, but brief, description of the
Andamanese languages comparing them to the language of the Semangs
of the Malay Peninsula and the Aetas of the Philippine archipelago. Tem-
ple (1903) proposes the theory of Andamanese being a language isolate
and an unusual ‘mixed media’ style of communication, although it is not
clear what the term ‘mixed media’ meant in those days. Radcliffe-Brown’s
(1922) The Andaman Islanders gives a very brief sketch (ten pages in all)
of the Andamanese languages, comparing forms from North Andamanese
languages such as Jeru, South Andamanese languages such as Bea and that
of Little Andaman, i.e. Onge and bringing out differences among them.
His analysis is not without critique of Man and Portman’s analysis, but
brings home some basic distinctions between various languages. Temple
(1903) maintained that “the languages [spoken by the Andaman tribes] all
belong to one family, divided into three groups, closely connected to the
eye, but mutually unintelligible to the ear” (1994 reprint: 14).7
In short, investigations carried out in pre-independent India focused
mainly on documenting the lexicon/vocabulary and grammar of the
Southern Great Andamanese languages, especially Âkà-Bêa. The North-
ern Great Andamanese languages were not dealt with in any detail. The
study of Onge and Jarawa grammars and lexicon/vocabulary were never

7 A detailed grammar was included in the Census Report of 1901: 98–121, which has
been left out from the reprint version undertaken by the Government of India, and hence
is not accessible to researchers.
14 chapter one

attempted. As far as the genealogical classification is concerned, all the

languages spoken in the Andaman Islands were considered as belong-
ing to one family, although the linguistic structures of the languages of
Great Andaman were considered different from the languages of Little
Andaman.

1.4.3 Study in Post-Independent India

By the time India gained independence, most of the central and southern
Great Andamanese languages were extinct. The North Andamanese
languages, such as Sare, Khora, Jeru and Bo were represented by a few
speakers and among them semi-fluent speakers outnumbered the fluent
ones. Intermarriage among the tribes generated a kind of ‘mixed’ tribe
speaking a ‘mixed’ language which draws its lexicon from these four
languages but which is primarily based on the Jeru grammar. PGA is thus
a koiné (Manoharan 1989) and as of today, only five semi-fluent speakers
are left in a community of fifty-six. Jarawa, Sentinelese and Onge are still
spoken and transmitted to the next generation. Our knowledge about all
of these languages—barring Sentinelese—is far more advanced than it
was sixty years ago.

1.5 Genealogical Classification

Linguists in the post-independent era classified all the languages spoken

in the islands as constituting one family (Basu 1952, 1955; Manoharan 1980,
1983). It was on the basis of a pilot survey of the Andamanese languages,
i.e., Onge, Jarawa and PGA (Abbi 2003, refer to Table 1.2 on a compara-
tive lexicon given below), that it was indicated that it may be possible to
establish the Great Andamanese language(s) as forming one family that is
distinct from the language family to which Jarawa and Onge belong.
Subsequent studies of a comparative, historical and typological nature
of the three languages based on extensive field work (Abbi 2005, 2006a,
2006b) and fresh data from Jarawa (Kumar 2005), as well as citation of
non-linguistic proofs (e.g., cultural, anthropological, archaeological and
genetic) substantiated the thesis (Abbi 2009) that the languages of the
Andaman belonged to two distinct families, i.e. Great Andamanese and
Angan.8 It was only after the publication of Abbi (2006) and the first-

8 The latter got its name because both Jarawa and Onge call themselves ‘Ang’, pro-
nounced as əŋ. Justin (2000) had also proposed that they should be called Ung because of
 the andaman and the great andamanese 15

Table 1.2. Comparative lexicon in Angan and Great Andamanese

English Angan Great Andamanese
Onge Jarawa
blood ɨnace:ŋ ɔcheŋe e-tei
boat ɖaŋɛ cɨ, (cagiya paɖa)-taŋ/daŋ rowa
bow iya a:w ko
child ɨcɨɭɨ ɨcɨɭə ʈʰire
crocodile ʈɔyəgɨ torogiyəi sare-ka-teo
crow wawa wa:raw phaʈka
dog wə:me, uame wɔm ca:w
goat ʈikʷabuli thikhwa-gopejayo —
laugh ɨɲya ənia: khole
water ɨɲe i:ŋ ino
1sg ‘I’ mi mi ʈʰu
hold caʔ kho: ye-ke
descend akobelaneka jagi-yə lɛbe
steal uigəle ʈop-om
swim ibolo waɽa/h-waɽa ŋɔʈo
snake toruba tobo ʃubi
(Based on Abbi 2003, 2006)

hand data on hitherto inaccessible Jarawa (Kumar 2005), that Blevins

(2007) came to the conclusion that the ‘Ang’ languages9 were affiliated
with the Austronesian language family. The categorisation of Onge-Jarawa
as ‘Austronesian’ (argued by Blevins 2007 as Ongan group) is far from
universally accepted. Although it is not conclusively established whether
the group Jarawa-Onge belongs to Austronesian, its typological and gene-
alogical distinction from Great Andamanese has been established by Abbi
(2006) as the ‘Ang’ group. Research by geneticists is consistent with lin-
guistic research. Thangaraj et al. (2005) indicated that the two ancient
maternal DNA lineages, M31 and M32 in the Great Andamanese and the
Onge respectively, have evolved in the Andaman Islands independently
from other South and Southeast Asian populations.
The Great Andamanese family is constituted of ten languages, which
can be grouped into three varieties: southern, central and northern. Refer
to Figure 1.1 and Map 5, the latter very clearly distinguishes the Great Anda-
manese languages from the Jarawa-Onge, the Ang group of languages.

the similar reason. The name Jarawa is given to the tribe by Âkà-Bêa to refer to them as
‘stranger’, ‘one you should be afraid of ’.
9 She uses the term ‘Ongan’ for these.
16 chapter one

A very valuable account of all the studies undertaken on the languages

of the Andamans up until 1988 is given in Zide & Pandya (1989). To date
this remains the only detailed and annotated bibliography available on
the topic.
A description of the Little Andaman language Onge is given by Das-
gupta & Sharma (1982) but is far from satisfactory. There is an urgent need
to study Onge in depth before it too becomes a moribund language. We
await the publication of a Jarawa grammar but an urgent need for a dic-
tionary of these two languages should be recognised.

1.6 Typological Differences

The Great Andamanese is a head-marking polysynthetic and agglutina-

tive language with an SOV pattern, and has a very elaborate system for
marking inalienability (Abbi 2006b, 2011), which is nested in seven posses-
sive markers designating different body parts (refer to Chapter 3). These
markers are further grammaticalised in the language and appear as pro-
clitics which classify dependent nouns including kin terms, attributive
and verbal modifiers, transitive verbs and intransitive verbs of experience,
ambience, and motion. For instance, extremities of body part terms are
preceded by the proclitic oŋ= as in oŋ=kenap ‘fingers.’ However, the noun
for ‘human blood’ tei, if oozing out of fingers, is oŋ=tei, but the word for
‘fever’ is i=tei as i= is used for internal body parts. As such, they demon-
strate the conceptual world of one of the most ancient cultures. The Great
Andamanese conceptualise their world through these interdependencies
and hence the grammar of the language encodes this important phenom-
enon in every grammatical category expressing referential, attributive and
predicative meaning. These are highly unusual structures not shared by
any other language studied so far.
The Angan languages on the other hand, are agglutinating languages
with an SOV clausal structure without the complexities of marking depen-
dencies on every grammatical category. Nor is the lexicon divided into
seven divisions as in the case of Great Andamanese.

1.7 The Last Decade

Improved accessibility to the region and the increased interest of young

scholars in the field have generated quite a few articles, dissertations
and conference papers as well as some very useful online material (www
.andaman.org and www.andamanese.net) on extinct and living languages
 Geographical Distribution of the Ang Family and Ten Di�ferent Tribes
Representing Great Andamanese in the 19th Century
Great Andamanese
Aka - Cari
Aka - Kora

Aka - Aka - Bale Aka - Bo

Aka - Kora
Aka - Jeru Kol Aka - Jeru
Aka - Kede
Aka - Juwoi
Aka - Bea
Aka - Kol

Jangil Aka - Pucikwar

Aka - Bale
Aka - Bea
Aka - Cari Onge

Aka - Bo
Aka - Kede Jarawa Onge - Jarawa

Aka - Juwoi
Jarawa

Aka - Pucikwar
the andaman and the great andamanese

Jangil
Sentineli Sentineli

NORTH
Onge

Map 5. Distribution of the Andaman tribes

17
18 chapter one

of the Andaman. Special mention should be made of Som (2006), Avtans

(2006, 2007), and Chaudhary (2007), Chaudhary et al. (2008). An inter-
active multilingual (English-Great Andamanese-Hindi) dictionary, which
contains more than 4000 words, 432 pictures, 900 sound files of individual
words and phrases/sentences, approximately 1100 sentences as illustra-
tions, and detailed notes on cultural and indigenous knowledge, is avail-
able (Abbi 2012).10 In addition, a CD of Great Andamanese songs and
the first “Book of letters” Varnamaala (Abbi et al. 2008), the first socio-
linguistic sketch of the Great Andamanese language (Abbi et al. 2007),
and an ethnolinguistic work on the names of the Great Andamanese birds
brought out by Pande & Abbi (2011) have all helped in archiving the dying
language.11 The ongoing work of Kumar (2005) and Abbi & Kumar (2010)
have given us indications for probable language contact that existed
between the extinct South Great Andamanese language, Âkà-Bêa, and the
Angan language spoken in the vicinity, i.e. Jarawa. Although the North
Andamanese languages are on their way to extinction, we can be assured
that the koinéised variety, which is known as present-day Great Andama-
nese, is adequately documented for posterity.

1.8 A Sociolinguistic Sketch of the Great Andamanese

The term ‘fluent speaker’, of any language, generally refers to a person

using the language spontaneously and continuously for any duration of
time in all domains and contexts. However, this is not the case with the
Great Andamanese. Their use of the ‘heritage language’ is sporadic and
not spontaneous. A great amount of coaxing was needed to make two of
the ‘best speakers’ converse in their own language. Within no time they
started taking recourse to Hindi. The only occasion where a member of
the tribe would spontaneously use the ‘heritage language’ with another
was to exclude a perceived ‘outsider’ from the conversation. It can be said
that the language at this stage is best used as a ‘secret code’ rather than
as a standard mode of communication. Even when using the language as
a secret code, speakers usually engage in short exchanges of words and

10 This dictionary is prepared in the specially designed Toolbox program of SIL, with
an accompanying Lexique Pro on a CD Rom, which gives in addition, etymology of words
wherever possible and sources of information.
11 The works cited here were the results of the project Vanishing Voices of the Great
Andamanese (VOGA) supported by the Hans Rausing Endangered Language Project, Uni-
versity of London (2005–2009).
 the andaman and the great andamanese 19

phrases and not in lengthy conversations. This too is the case with the
‘best speakers’ of the language. The continuum of language competence
among the Great Andamanese that I discuss below should be understood
against this background. I use the term ‘best speaker’ for speakers with the
competence level 5 (refer to §1.8.2) and the speakers at level 4 are termed
‘semi-speakers.’ The term ‘best speakers’ is used for those who were found
to be best among the few members who still use the language although
not very fluently and not in all domains. Apart from language loss, the
Great Andamanese people also suffer from a loss of cultural heritage, of
ancient practices and rituals, as well as a loss of the art of narration. The
few people who now speak the language did not remember any native
stories. It was noticed that storytelling as an activity does not exist
anymore.12 Neither the young mothers nor the old people of the commu-
nity ever narrate stories to their children. The loss of the mother tongue
has very serious implications as the very genre of narration has also been
lost in the contact language, i.e. Andamani Hindi. Thus, the present gen-
eration of Great Andamanese have never heard any story from their elders
neither in their heritage language nor in Andamani Hindi.
A thorough investigation of the linguistic behaviour of the Great Anda-
manese towards their own language and culture presents a bleak picture
and points towards a future when they will become a group of people
who would have totally lost their land, language, tales, songs, art of body
painting and many facets of indigenous cultural identity. These are the
perils of language loss.
The following is a sociolinguistic sketch based on initial fieldwork con-
ducted during 2005–2007, when the total population of the Great Anda-
manese was fifty. The number of children under the age of ten years was
twenty. There were thirteen members in the age range between twenty
and thirty. There were only ten members aged between thirty and eighty.
The number of infants, i.e. between the ages of zero to five years, was also
thirteen. See Figure 1.3.
The increase in population in the Great Andamanese tribe has been
tremendous especially in the last few years. There were only thirty-six

12 However, Nao Jr., one of the best speakers of the language, did manage to narrate ten
folk tales with great effort, although not all in the heritage language. These are submitted
for publication with NBT, India. The first folk tale narrated by Nao Jr., An Ancient Tale from
Andaman, was published in 2012. He narrated folk tales in Hindi after much prodding. It
took him, at times, many days to complete a story as he had never narrated one in his life
time and had heard them from his elders only when he was a child. Other old members
of the community reported loss of memory regarding folk tales. It was a very sad state of
affairs as a couple of speakers wanted to oblige us but could not remember anything.
20 chapter one

14
13 Number of
12 tribe members
12

10
8
8
7
Number

6
5
4
2 2
2
1

0
0 to 5 6 to 9 10 to 20 21 to 30 31 to 40 41 to 50 51 to 60 61 to 80
years years years years years years years years
Age

Figure 1.3. The number and age of the Great Andamanese people (2005–2006)

members in 2001 when the author conducted her first fieldwork on the
islands but at present there are fifty-six. This indicates that the commu-
nity is thriving.

1.8.1 Background of the Great Andamanese Speakers

The speakers of the Great Andamanese language hail from various back-
grounds, some of them have Jeru parents, and some others have a mixed
parentage. For example, the parentage of the seven ‘best speakers’ is very
varied as can be seen from the following tables. Apart from three of the
speakers, all of the others have mixed parentage, i.e. their parents spoke
different languages, even though they themselves claim to be Jeru speak-
ers. For example, refer to appendix ‘A’: the genealogical tree of Lico, one of
our main consultants. The reality is that although they claim Jeru as their
mother tongue, it may actually be a mixed language with inputs from vari-
ous languages that they have learnt in their childhood. Consider the par-
entage of the older members of the community. Their children are farther
away from the heritage language, although they claim Jeru as their mother
tongue. It can be seen from Table 1.4 that one of the younger speakers,
speaker no. 14, not only has mixed parentage but also has a wife who is
from outside the community and speaks Andamani Hindi.
 the andaman and the great andamanese 21

Table 1.3. The lineage of the oldest and the best speakers
Speakers Claimed MT Mother’s MT Father’s MT Spouse’s MT
Sp 1 (58) M Jeru/Bo Sare Bo Jeru
Sp 2 (80) F Jeru Bo Jeru Jeru
Sp 3 (74) F Khora Khora Khora Jeru
Sp 4 (60) M Jeru Khora Jeru Jeru
Sp 5 (45) F Jeru Jeru Jeru Jeru/Khora
Sp 6 (48) F Jeru Jeru Jeru Bo
Sp 7 (43) F Jero Khora Bo Jeru
(The number in brackets indicates the age of the speakers)

Table 1.4. Lineage of the semi-speakers

Speakers Claimed MT Mother’s MT Father’s MT Spouse’s MT
Sp 8 (43) M Jeru Jeru Jeru Jeru
Sp 9 (27) M Jeru Jeru Bo Jeru
Sp 10 (27) M Jeru Khora Jeru Jeru
Sp 11 (28) M Jeru Jeru Jeru Jeru
Sp 12 (45) M Jeru Khora Jeru Jeru
Sp 13 (30) F Jeru Khora Jeru Jeru
Sp 14 (38) M Jeru Khora Jeru Andamani Hindi
(The names of the speakers are concealed on purpose)

Interestingly, speaker no. 14 is not the only one who has married out-
side the community. One of the female members of the community, aged
twenty-six (who has a mixed parentage of Jeru & Bo) has recently mar-
ried a Bengali person. She is not a fluent speaker of the language; her
competence level is 3. Hence, an average Great Andamanese could be a
Khora speaker, married to a Sare, and speaks Jeru to her neighbours and
Andamani Hindi to her children. It is a complex web of multilingualism
that an average Great Andamanese builds around her/him.
Marriage with outsiders has resulted in interactions with other com-
munity speakers and in the learning of their languages. Although marriage
and cohabitation with members outside the community is a recent trend,
intermarriages among different sub-tribes of the Great Andamanese is
nothing new, as can be seen in Table 1.3. However, with almost every-
one claiming Jeru as their mother tongue, the other heritage languages
are clearly marginalised and ignored, indicating a ‘superiority’ com-
plex assigned to the heritage language Jeru. Members are proud to call
themselves Jeru speakers even when few of them actually speak it. This
22 chapter one

is an unexpected response from an endangered or moribund language

community.13 In mainland India tribals never like to be associated with a
dying language.

1.8.2 Language Competence

The analysis of the language competence of present-day Great Andama-
nese speakers reflects the language’s health. When we initially visited
the island there were fifty members in the community and among them
twelve were below the age of five. Our analysis was based on the popula-
tion that was above the age of five, for obvious reasons. At the time of
the fieldwork, the fluency level of the other thirty-eight members of the
speech community was judged, based on a calculation, on a scale of com-
petence ranging from 0-5. The key to the competency levels given below
shows the number assigned to speakers on the scale of fluency.
Key to the competency levels
0 = no knowledge of the language
1 = neither speaks nor understands, recognises only a few words
2 = low level of understanding, does not speak
3 = phrase level awareness, but does not speak
4 = semi-speaker, can understand but does not speak the language fully
5 = best speaker, can understand and speak the language
This scale of competence is used in the following graphs as well as in the
analysis of the comparative competence level of the Great Andamanese
speakers (Figures 1.4 and 1.5).
We can see that the numbers of ‘best speakers’ and ‘semi-speakers’ are
identical, and together constitute thirty-six percent of the community. The
number of speakers in the highest competence group, i.e. level 5, was only
seven. These were the best speakers of the language. Another seven mem-
bers fell into the next level of competence, i.e. level 4, the ‘semi-speakers.’
These fourteen members of the community, who in quantitative terms
constituted a rather small community, could speak the language in some
domains and registers. The analysis showed that the highest number of
people, i.e. ten in all, was in the group denoted by the number 1, signifying
that part of the society which has a very low or negligible understand-
ing of the language. These speakers possessed passive knowledge of only
some words of the language, without ever being able to use them.

13 A similar situation is observed in the mainland northern belt of India where speakers
of most of the marginalized languages claim ‘Hindi’ as their mother tongue.
 the andaman and the great andamanese 23

5 7 Number of
speakers
4 7
Competence

3 8

2 3

1 10

0 6

0 2 4 6 8 10 12
Number of speakers
Figure 1.4. Number of speakers on the scale of competence

70 Mean Age
60 58
50
Mean age

40 34
30 23
20 16
10 8.8
4
0
0 1 2 3 4 5
Competence level

Figure 1.5. Age-wise competence level of the speakers in 2007

The second highest category was occupied by number 3 in the graph,

representing the group with phrase level understanding whilst not being
able to speak the language. As no two speakers speak the same variety
of the language, though they all claim to be speaking Jero/Jeru, compre-
hensibility of each other’s speech is rather low. For instance, the oldest
member of the tribe was an eighty-year-old woman, our speaker no. 2
(Table 1.3), whose mother tongue was Bo. Her speech did not share mutual
intelligibility with the speech of any other member of the community.
Other members somehow ‘worked out’ what she said.
It is noteworthy that the best speakers were all above the age of forty.
The mean age of the semi-speakers was thirty-five. It is clear that the level
of competence goes down with age, the oldest having the best and the
younger generation having the least knowledge, which shows the gradual
 24 chapter one

5 58 7

4 34 7
competence

3 23 9

2 16 3

1 9 10 Mean Age
No. of speakers
0 4 6

0 10 20 30 40 50 60 70
Mean age

Figure 1.6. The number and mean age of the speakers on the competence scale

process of language erosion. This also clearly indicates that the language
is not transmitted from the older generation to the younger generation
effectively. The reluctance of the younger members of the community to
learn and use the heritage language is also responsible for this decline in
their competence level. The graph below (Figure 1.5) clearly indicates the
decline in competence level from the older to the younger generation.
Figure 1.6 given below sums up the situation of language competence
among the Great Andamanese with respect to the age of the speakers and
their number in the community. The members with the highest compe-
tence also have the highest mean age. However, the strength of level 5
speakers was very small in 2007 and it has since been further reduced
by the passing away of four elderly speakers. The members with a low
competence level have a low mean age and their strength is consider-
ably higher than the previous group, indicating the moribund status of
the language. Competency level is positively related to age: the higher the
age group, the higher the proficiency; the lower the age group, the lower
the proficiency in Great Andamanese.

1.8.3 Language Use

It is clear by now that the Great Andamanese are no longer monolin-
gual in their heritage language. The few members of the tribe (above the
age of forty) who can speak the indigenous language are bilingual, both
in Hindi and in the heritage language. The members who are below
forty are monolingual in Hindi, which they use as a lingua franca as well
 the andaman and the great andamanese 25

as the home language. Some of the members of the younger generation,

especially teenage children, also speak a few words and phrases of Eng-
lish and other Indian languages, such as Bangla and Tamil. Andamanese
people are employed in various government departments and hence are
exposed to the languages of their colleagues. Marriage outside of the com-
munity also further adds to this language pool.
This kind of linguistic mosaic is constantly in the background and
decides who uses what language of communication, when and with
whom. The intra-community use of Jeru is rare. As mentioned earlier, it
is used as a secret code among the oldest members of the tribe. Another
important observation is that very few registers are used in the heritage
language. For instance, none of the so-called best speakers can narrate an
incident or a story (see footnote 12).
With Hindi being the matrix language and the use of the heritage
language limited to a few words and phrases, Hindi is used along with
present-day Great Andamanese in most of the conversations. There was a
not a single instance where a member of the tribe was observed using any-
thing longer than a phrasal communication in the heritage language. Chil-
dren and young adults use Hindi in most of the domains. It was noticed
during the data collection that the older members of the tribe use Hindi
to communicate with young and old alike. Interestingly, when posed with
a fictitious situation of choosing one’s mother tongue, a large number of
speakers opted for Hindi.
This brings us to the question of code-switching and code-mixing.
Among the Great Andamanese it is more the norm, rather than the excep-
tion, to use more than one language in the course of a single conversa-
tion. Almost all of the speakers code-switch between Hindi and Jeru while
talking in addition to using them as a mixed-code (Avtans & Abbi 2006).
However, along with these two languages, their present-day speech indi-
cates the use of vocabulary from other heritage languages, such as Bo, Sare
and Khora. Code-switching between Jeru and other heritage languages is
very rare. The form of present-day Great Andamanese, at best, can be
described as a Koiné.
Our prolonged stay with the members of the community revealed that
the so-called ‘best speakers’ cannot communicate in their respective lan-
guage varieties in all areas. The non-homogeneity of the community, due
to the existing members being descendants of various sub-groups within
the Andamanese language family, and hence having different linguistic
heritages, could have led to an awkward situation of non-use of one par-
ticular indigenous language. It should be noted that the language situation
26 chapter one

is highly complicated and unique. Although most of the ‘speakers’ of the

language claim to speak and understand Jeru, our observation is that this
is untrue. No two speakers use the same register. Though some of them
might understand fellow members’ speech, it certainly does not mean
they fully comprehend one another. Each of the community members
speaks a mixed variety of languages drawing lexicon from four languages,
i.e., Sare, Jeru, Khora and Bo, although the grammatical structure does
appear to be based on Jeru, the language they all claim to be speaking. In
subsequent chapters it will be shown that even this, i.e. the grammatical
structure, is not based on one language, but has a mixed base of at least
two grammatical structures.
Present-day Great Andamanese is thus the result of leveling of a mini-
mum of two grammatical inputs and of the vocabulary/lexical items from
four or five different languages of the same family. The present form of
PGA is a Koiné variety which has been learnt by speakers above fifty with
varying degrees of lexical input from the language of their father and
mother.14 It is this language which is described in this grammar.
To summarise, the older generation mainly uses the heritage language,
with code-switching and code-mixing between Hindi and PGA. All mem-
bers use these two languages in free variation in intra-community com-
munication. It should be noted that storytelling as an activity hardly exists
among the Great Andamanese anymore. Only one member of the com-
munity could remember some of their traditional stories, as the traditional
art of narration is no longer practised.
However, it must be mentioned here that their traditional songs and
the art of singing are not lost. It was observed that the female members of
the community still remember a few old songs and they sang them for us.
The number of men who could sing these traditional songs was negligible
compared to the women. Hence, it is the female members of the com-
munity who have retained the treasure of songs: more so than the male
members. This observation proves that the life of songs in the heritage
language is longer than the life of the language itself. Songs in the origi-
nal language survive even when the language is no longer in use. While
the art of narration perishes in the process of language death, songs and
singing are not as adversely affected. Music is a memory-based activity
and narration is based on language use. This difference in their cognitive

14 It is interesting to note that tribes associate themselves by father’s tongue and not
by mother’s tongue.
 the andaman and the great andamanese 27

processes has a serious impact on their retention, as exemplified by the

Great Andamanese.
The Great Andamanese love Hindi for its prestige and would not mind
if they were given a choice to make Hindi their mother tongue. For further
details see Abbi et al. (2007). They do not encourage children to speak the
heritage language as they prefer to talk to them in Andamani Hindi. The
Great Andamanese prefer to be identified by tribe name rather than by
language or religion. One of the causes of the reluctance of the younger
generation to learn Great Andamanese is the fact that having knowledge
of the heritage language neither provides them with a livelihood nor does
it lend them any prestige in society.

1.8.4 A Complex Situation

The case of the Great Andamanese language is a complex one: a language
on the brink of extinction with merely fifty-six members who comprise
the entire speech community, and out of these fifty-six, only five mem-
bers have varying degrees of competency in PGA. The situation becomes
tougher because of the differences in register even among these few best
speakers. Although they all claim to speak Jeru, they do not fully under-
stand one another, thus proving that they actually speak various forms of
the former Andamanese languages. This was illustrated when two elderly
women, one a Bo speaker and the other a Khora speaker, proved to be
mutually unintelligible. Currently, it is not easy to completely establish
the antecedents of PGA, but similarities with Sare, Khora, and even Bo
can be attested through the use of certain words.15 In addition, the lin-
eage of some of the speakers shows that they had non-Jeru parents (refer
to Appendix ‘A’). Hence, the use of other languages, such as Khora, Sare,
Pujjikar and Bo, which are undoubtedly mixed with PGA in their speech.
It can be concluded from the discussion above that present-day Great
Andamanese is a moribund language. As reported earlier in §1.8.3, the
Great Andamanese people have lost various registers of their language;
only one or two registers remain in use today. As there are not many
members left in the community who share the same language, the lack of
use of PGA has contributed considerably to its loss. The only ‘best speak-
ers’ left are well over forty years of age, and coupled with the permeation
of Hindi even among the older and ‘best speakers’ of the community, the

15 The etymology of the words is given in A Dictionary of the Great Andamanese Lan-
guage English-Great Andamanese-Hindi’ (Abbi 2012).
28 chapter one

very need and domains in which to use their heritage language are vanish-
ing. Alongside education and the culture of the work place, daily contact
with mainstream people must surely have a direct effect on language ero-
sion. By their own admission, code-switching and code-mixing between
Hindi and PGA is far higher than the occasional mixture of a few words
that takes place between the other heritage languages.

1.8.5 The Predicament

The survival and revival of any language solely depends on its speak-
ers’ attitudes towards it. In the case of the Great Andamanese the above
mentioned facts clearly show how short the life is for PGA as a heritage
language. The only signs of ‘revival’ that we witnessed were in our subse-
quent visits during language documentation (Avtans & Abbi 2006). The
Great Andamanese were motivated to speak the heritage language among
themselves and with us too. However, as we are not permanent mem-
bers of the community, this interest cannot be sustained. Once a language
ceases to be the home language and stops being transferred from genera-
tion to generation, one cannot save a language from the imminent danger
of death. PGA is a classic example of a dying and moribund language.
To conclude this discussion, we can state that moribund languages are
characterised by a lack of will to learn and teach the heritage language.16
The language does not get transferred from the older to the younger gen-
eration. Another important feature of moribund languages is the loss of
registers and reduced domains of use. In the restricted domains that the
language is used, it shows traces of earlier varieties in lexical items and
grammatical structures. The loss of various registers also results in the
lack of total mutual intelligibility even among those speakers who have
retained the language. Present-day Great Andamanese is a conglomera-
tion of several varieties, having input from two or three other heritage
languages. It has taken on the shape of a bilingual mixture with Hindi as
a significant constituent. Hence, PGA is both an example of a Koiné as
well as of a bilingual mixture. We will describe this very language in the
following chapters.

16 First language attrition has been researched by many linguists in the past; Kees de
Bot (2004), Herbert W. Seliger, Robert Michael Vago (1991), and Monika S. Schmid (2011)
among others.
 the andaman and the great andamanese 29

1.9 The Great Andamanese Culture: Some Observations

1.9.1 Naming a Person

An average Great Andamanese person acquires several names during his/
her lifetime dependent upon various factors, including ecological. The fol-
lowing table describes these stages and the reason(s) for a change of the
given name. The first stage starts with a name being given when the child
is still in the womb; hence there is no gender distinction in the names of
a person. A female child is given a different name during the course of

Table 1.5. Stages of naming an individual

Stage 1 Child
(with a prenatal name)
Stage 2 Beginning of puberty
(Period of Aka-op)
Boys Girls
Tattooing on their back (boys are called Flower name, preferably a seasonal
ejido for a short while and then revert one
back to their prenatal name)
Between stages 2 and 3 some dietary restrictions are adhered to
Stage 3 Commencement of adulthood
(Ceremony of turtle eating by males and females)
The boy is called khimil for a short period. Some food restrictions are lifted for
both males and females. Females are still called by their flower name.
Stage 4 Attainment of adulthood
(Honey/dugong/pig eating ceremony)
All food restrictions are lifted. Men regain their prenatal names but women do
not.
Stage 5 Marriage
Condition for marriage: prospective spouse has to prove his/her hunting pro-
wess; names are not changed
Stage 6 Birth of the first child
Parents are addressed by the same term for a short period
Stage 7 Parenthood
subsequent children; women and men regain their prenatal names
Stage 8 Death / spirits
all prenatal names are prefixed by the honorific maya, e.g. maya jirake
‘the late Mr. Jirake’
30 chapter one

Table 1.6. Honey calendar or names of the seasons

Name of flowers Associated seasons
jili rɔ Onset of summer
tipok ʈɔlɔ Midsummer
phoco ʈɔlɔ Intense summer
rɛʈ cer End of summer and onset of rains
rea ʈɔlɔ Onset of mild rains
cɔkhɔro ʈɔlɔ Heavy rains

her life based on the names of flowers. Men undergo name changes more
often than women in their life cycle. Consider the following table describ-
ing different stages of naming a person.

1.9.2 Naming Seasons: The Honey Calendar

The Great Andamanese honey calendar is based on the names of flowers
in bloom during that particular season. This naming process is related
not only to the seasonal changes but also to the flowers’ inherent rela-
tion with the availability of honey. As observed by Radcliffe-Brown (1922:
39) honey occupies a special place in the Great Andamanese pattern of
subsistence and movement. The Great Andamanese sub-categorisation of
the primary seasons into minor seasons is perhaps more closely related to
the availability of honey and its taste or smell (Radcliffe-Brown 1922: 119).
Interestingly, the Great Andamanese elders of Strait Island claim to be
able to identify the origin of honey by tasting it, i.e. the particular kind of
flowers that bees used in making the honey. The blooming of each flower
is associated with a change of the season and hence the particular time of
the year. Consider Table 1.6 below.
The onset of the rains is symbolised by the compound noun ‘bamboo’
+ ‘rain’ the rɛʈ cer. Here no name of any flower is mentioned.
Man (1923: 182) mentions that the Âkà-Bêa people divided the primary
seasons into twenty minor seasons named after flowers in bloom during
that particular season. Radcliffe-Brown (1922: 118) provides a list of flower
names which were used by the Great Andamanese as reference points of
time. His list of flower names in Aka-Jeru and Âkà-Bêa, and the season
they are in bloom are given alongside the corresponding months from the
Gregorian calendar (see Table 1.7 below).
 the andaman and the great andamanese 31

Table 1.7. Blooming of flowers and associated months of the year

Âkà-Bêa flower Aka-Jeru flower Corresponding time
clilipa celebi Middle of November to middle of February
moda mukui
ora okor Middle of February to middle of May
jidga
vere jeru
pataka botek
balya puliu
rece re Middle of May to end of August
cagara cokoro
carapa carap September, October, and the first half of
cenra ʈorok November
yulu jili

1.9.3 Measuring Тime in Great Andamanese

The Great Andamanese measure the time of an event, the age of a per-
son, the history of their culture and civilization, the antiquity of human
evolution, as well as the temporal divisions of a day, according to vari-
ous parameters. The temporal organisation system is based on references
to particular events rather than being based on a particular digital or
numeral system. Day-to-day life is organised around temporal divisions
which are guided by the movements of the sun and the moon. Four basic
temporal categories are recognised.

1.9.3.1 Parameters of Temporal Categorisation

Various parameters are used to measure the time of an event. The scale
for measuring each event type is distinct. There are four basic temporal
zones:
1. Natural time (recurrent phenomena)
2. Life cycle (age of a person)
3. Historical (Pre-/Post-British era)
4. Evolutionary period (mythological source)

1.9.3.2 Natural Τime Parameters

The time of a day, week or month is measured along the parameters given
below:
1. The movement of the sun in the sky
2. The waning/waxing of the moon
32 chapter one

3. High and low tides

4. The seasons when flowers and fruits are in bloom
5. Hunting and gathering criteria

1.9.3.3 Evolutionary Period (Mythological Time)

Mythological beliefs play a very significant role in the hunter-gatherer
society. Seen from this point of view, the evolution of humans, the world
and the cosmos have seven reference points. Some of them intersect with
each other.
1. The origin of the earth
2. The origin of humans
3. The great flood
4. The great drought
5. The creation and saving of fire
6. The transformation of animals into humans/spirits
7. The transformation of humans into birds

1.9.3.4 Historical Parameters (Pre-and Post-British Era)

It is interesting that the Pre-British and Post-British periods are known by
the absence and presence of dogs, respectively. Dogs are considered very
important animals because of their use in hunting.

1.9.3.5 Life Cycle (Age of a Person)

The average age of a person is calculated on the basis of what rites (s)he
has performed or undergone. These can be listed as:
1. Birth
2. Puberty rites
2. Turtle eating ceremony (commencement of adult hood)
3. Attainment of adulthood (initiation ceremony: pig/honey/dugong eating)
4. Marriage
5. Birth of each child to the married person
6. Death
The measuring of time can be summarised in Table 1.8 below.
It is not very easy to decipher this table as the stages are not presented
in a chronological order. Some of the stages overlap. The Great Andama-
nese believe that when the Andamanese became birds, there were some
animals that transformed into humans or into spirits. However, the origin
of the earth precedes all latter stages.
 the andaman and the great andamanese 33

Table 1.8. Measuring time in PGA

Time Parameter of temporal categorisation
Natural time Movement of the sun in the sky
Waning/waxing of the moon
High and low tides
Blooming of flowers and fruits
Hunting and gathering criteria
Life-cycle Birth
(Age of a person) Puberty
Commencement of adulthood
Attainment of adulthood
Marriage
Birth of a child
Death
Historical Introduction of dogs to the island vs. period
(Pre-and Post-British era) without dogs
Evolutionary period Origin of the earth
(Mythological time) Origin of humans
The great flood
The great drought
The creation and saving of fire
Transformation of animals into humans/
spirits
Late in the period of evolution Transformation of humans into birds

1.10 The Beliefs of the Great Andamanese

The Great Andamanese were hunter-gatherers until a hundred years

ago, but they have now adopted an ‘urban’ way of life. The Great Anda-
manese do not worship any gods. They do, however, chant mantra-like
words before going out on long expeditions in the sea. These words are
meant to calm the sea and please the puluga (southwest wind). The cre-
ator, mautkochua, resides in the sky. The Great Andamanese also consider
bilikhu, the strong wind coming from the south in a godlike sense. Mano-
haran (1989) mentions a word adix for ‘god.’ But when we tried to confirm
this word, our consultants pronounced it as a-dik and it meant ‘devil’ or
‘demon.’ One speaker used the word bilikhu-boto for ‘god’, but literal trans-
lation or morphological analysis of this word is bilikhu + storm. They have
two more words—jurwacom (god or devil of the sea, i.e., sea god) and
motkocua (the one above all of us). Perhaps this may also mean ‘creator.’
The words puluga and bilikhu refer to the same entity. The two are varia-
tions in different dialects. The creator is generally not given any rank in
34 chapter one

the folk system of classification, or if given, it assumes the highest rank.

For further details on this, readers may refer to Pande & Abbi (2011).

1.10.1 The Domain or Realm

The Great Andamanese perceive the world to be made up of three unique
divisions or domains in which everything else is included. The highest
domains for the Great Andamanese are ʈɔ ‘sky’, where their ancestors
dwell. All other beings belong to the forest ʈhimikhu and the sea sare. They
describe trees in a similar way to how a human body is described, its legs
(trunk), hands (branches), head (top) etc. This is very common across the
world in languages whose communities are very close to the environment,
such as hunter-gatherers.

1.10.2 Major Life Forms and Ethno-Biological Classifications

The Great Andamanese consider ‘life forms’ or ‘living beings’ as superor-
dinate forms and the creatures of the sea, jungle and sky as subordinates.
Hence tajio ‘living beings’ are of various kinds: tajio-tut-bec ‘living beings
of hair/feathers’, i.e., ‘birds’, tajio-tot cor ‘living beings of scales’, i.e., ‘fish’,
and tajio-cɔla ‘living beings of land’, i.e., ‘animals of the land.’
The classification of ethno-biological entities in Great Andamanese,
like many other hunter-gatherer societies, is based on their perception of
the world around them. The Great Andamanese recognise at least three
kingdoms:
(a) ‘Animals’ tajio-cɔla
(b) ‘Plants’ or ‘trees’ ʈɔŋ
(c) ‘Humans’ i-shɔŋo, essentially meaning the ‘human body’
Within each kingdom there are further classifications. For example, under
(a) animals, the life form categories include:
(i) Creatures that fly—em-barate, such as ‘birds’ -tajio-tut-bec and ‘bats’
jibeʈ, are grouped into one category.
(ii) Creatures of the water—ino, such as ‘fish’ tajio-tot-cor, ‘frogs’ phorube,
‘turtles’ cokbi, ‘dugong’ kɔroiɲ, ‘shells’ bun, ‘corals’ buruku, ‘clams’ coay,
and ‘muscles’ diʈhu.
(iii) Creeping creatures—‘lizards’ teo or teyo, ‘snakes’ shubi, ‘geckos’ ʈoʈɛmo,
‘monitors’ and ‘crocodiles’ sare-ka-teo.
(iv) Large creatures of the land tajiochola, such as ‘hogs’ ra, ‘jungle cats’
kheŋe, ‘rat’ bɔrot-toɖe, and ‘dog’ cao.
(v) Small creatures of the land—taoe, such as ‘insects’ duŋ, ‘bees’ phulimu,
and ‘worms’ bullu, etc.
 the andaman and the great andamanese 35

Each life form category is usually divided into intermediates, which is the
rank between life form and folk-generic categories.
The Great Andamanese people recognise different folk specifics and
a subclass of folk generics, mainly on visible attributes. One can also
see that there is a high level of correspondence between folk generics in
Great Andamanese ethno-taxonomy and scientific taxa, (for details see
Pande & Abbi 2011). Our detailed study shows that the Great Andamanese
classification system is consistent with ethno-biological classifications of
other non-literate tribal societies (Berlin 1992; Forth 2004), indicating that
humans everywhere recognise nature’s structure in essentially the same
way (Bulmer 1974). It can be seen that the Great Andamanese ethno-
taxonomy is consistent with the general principles and ideology of sci-
entific classification, as observed analytically for several other non-tribal
societies (Berlin 1992).

1.11 The Βiological Universe of the Great Andamanese

The Great Andamanese used to live in the tropical forests of Andaman.

This gave them a natural habitat rich in a specific type of flora and fauna,
which were an integral part of their hunter gatherer life-style. What remains
today in the memories of the surviving tribe members are a few vocabulary
items without real world referents. As mentioned earlier in the chapter,
the present-day Great Andamanese are said to be the descendants of the
North Andaman group of tribes that comprised, predominantly, of coastal
dwellers. This is one of the reasons why the lexicon provided in the dic-
tionary (Abbi 2012) is so rich in the names of fish and other sea creatures.
The available literature on the Andamanese tribal groups, from the
1800s until the present date, has focused on the culture and traditions of
these groups. The early available studies by scholarly-minded missionar-
ies are, for the most part, anthropological in nature. There have been few
ethno-biological explanations. Although some studies mention certain
plants and animals, a detailed study of the tribes’ knowledge of this topic,
how this is represented in their language, and more significantly, how this
knowledge and their culture interact in their domain of conceptualisation
have thus far remained unexplored. This grammar is a mere reflection of
how the Great Andamanese conceptualise their world and organise their
language.
The conceptual structures of lexical items and how they form sentences
in an endangered language are extremely important as a reflection of the
36 chapter one

collective mental representation of ideas by the surviving members of a

speech community, who are negotiating meaning in a fiercely multilingual
world—a world which is far removed from their own conceptualisation.
As the Great Andamanese population are known to be the descendants of
the early migrations from Africa about 70,000 years ago, and as they were
living in isolation until the end of the 18th century (Kashyap et al. 2003),
their language exposes to us an ancient world view. It is unfortunate that
this language is on the brink of extinction.

1.12 The Present Study

The present study analyses PGA as spoken on Strait Island and in parts of
Port Blair. There are ten chapters to this grammar, followed by the appen-
dices including maps, graphs and a sample text. As certain structures of
the language were found to be unique, Chapter 3, which is a ‘Grammar
overview’, familiarises the reader with the basics in addition to the axi-
oms necessary to comprehend what follows in the subsequent chapters.
In this respect, it is the most important chapter of the present grammar.
It would be beneficial if the reader were to read this chapter before mov-
ing on to the others. Readers may find phonetic variation in forms within
and across chapters as our data comes from various speakers who have
retained some of the sounds/phones of their original heritage language.
These variations have been explained in the chapter on phonology. The
only text with interlinear translations that was available to us, after a long
period of working with the community members,17 is the Creation Myth,
the story of Phertajido, the first Great Andamanese or the first human
being. This is given at the end of the grammar.

17 For details on the elicitation of folk tales and the difficulties in acquiring them, one
can consult A Dictionary of the Great Andamanese Language. English-Great Andamanese-
Hindi (Abbi 2012).
 Chapter two

Phonetics and phonology

Introduction

In this chapter the sound system of present-day Great Andamanese (PGA)

will be presented. Firstly, the vowels and their distribution will be dis-
cussed, followed by vowel sequences and then phonotactics. Secondly, the
consonants, their distribution and phonotactics will be presented. PGA
has some very unusual consonants which warranted acoustic study. The
discussion and the results of this study are given in Appendix B. Next, the
nature of the syllable is discussed. A high variation among the speakers
in the inventory of vowels and consonants was noted, most likely because
of the ‘koiné’ and ‘mixed’ nature of the language. Another factor leading
to such variation could be that as the language is on the verge of extinc-
tion, community members do not remember many words and their exact
pronunciation, and therefore offer varied sounds for the same word. The
indifference of the speakers towards the language could also lead to such
variation.

2.1 Vowels

Great Andamanese has a seven-vowel system, as shown in Table 2.1. Many

combinations of vowel sequences and clusters are possible, as shown in
Tables 2.3 and 2.4. The phonemic inventories of the vowels and conso-
nants were established by eliciting minimal pairs for most of the sounds.
Where minimal pairs were not available, the judgments of the native
speakers about the phonological contrasts were taken into account.

Table 2.1. Vowels of Great Andamanese

Front Central Back
Close i u
Half close e o
Half open ɛ ɔ
Open ɑ
38 chapter two

During the first phase of fieldwork (2001–2002), I recorded the mid-central

vowel, schwa [ə], in the data. However, subsequent visits to the speech
community and digital recordings of the vowel have indicated that the
language, in fact, lacks this central vowel. There is only one unrounded
back vowel and that is a. The rendering of Hindi words such as bənao
‘make’, were realised as banao. It was observed that contrast was neu-
tralised across speakers. For example, we noticed that although e and
a stand in contrast, speakers use them interchangeably in some words.
There was a similar situation with the back vowels o and ɔ, as well as
with o and u, which were in free variation (at times within the speech of
the same speaker), despite the fact that the two sounds do offer contrast
in minimal pairs. Measured on the scale of The World Atlas of Language
Structures (Haspelmath et al., 2005), PGA’s vowel system falls under ‘large
vowel quality inventories’, a characteristic feature of about a third of the
world’s languages.

2.1.1 Phonemic Contrasts

Some of the minimal pairs and sub-minimal pairs that we could attest are
given below.

2.1.1.1 Front Vowels

/e/ : /i/
/issu-e/ ‘burn’ [Vt]
/esso-i/ ‘stitch’ [Vt]
/eji-bom/ ‘he flies’
/iji-kom/ ‘he eats’
/bule/ ‘big creek’
/buli/ ‘the one who takes back something or somebody’
/beno/ ‘to sleep’
/biŋo/ ‘to hear’, ‘think’

/i/ : /ɛ/
/iboi/ ‘cooked food’
/ɛ boi/ ‘spouse’

/ɛ/ : /e/
/bɛŋ/ ‘muck’
/beŋ/ ‘forehead’ ~ [be:ŋ]
/en/ ‘a kind of leaf ’
/ɛn/ ‘wound’
 phonetics and phonology 39

/ɛ/ : /ɑ/
/ɑrɑ-cɛ/ ‘sting’
/ɑrɑ-ca/ ‘nest’

2.1.1.2 Back Vowels

/ɑ/ : /o/
/ɲɑ/ ‘bark’
/ɲo/ ‘stay’
/leɑ/ ‘slow’
/leo/ ‘younger than’

/ɑ/ : /ɔ/
/phɑr/ ‘boil/pimple’
/phƆr/ ‘bamboo’
/kɑbo/ ‘a deity’
/kƆbo/ ‘skin’

/o/ : /ɔ/
/kobu/ ‘umbrella’
/kɔbu/ ‘plates made of silai leaf ’
/ʈole/ ‘pluck’ (imp)
/ʈɔle/ ‘flowers’
/ɛrphɔke/ ‘take out from hot fire’
/ɛrphoke/ ‘kill’ (imp)
/ɛrkɔtho/ ‘nose’
/ɛrkotho/ ‘mix’

/u/ : /o/
/ɑrɑpho/ ‘fell a tree’
/ɑrɑphu/ ‘latrine/stool’

2.1.2 Length
Length at the phonemic level could only be attested for a few vowels.
However, it was noticed that the speakers occasionally varied between
the absence and presence of vowel length. Vowel length has been speci-
fied where it is important for the pronunciation of the word, especially in
the words which are derived from the Bo variety. Words derived from the
Sare variety also show distinctive vowel length. The last speaker of the Bo
language, Boa Sr., had distinctive length in her speech. Vowel length has
been specified by a colon mark [:] placed after the long vowel. It can be
postulated that the earlier version of the language had phonemic length,
as was documented by Manoharan (1989), although this is apparently no
40 chapter two

longer the case. Some examples demonstrating vowel length are given
below:

/ɑ:/ : /ɔ:/ (sub-minimal pair)

/bɑ:ʈom/ ‘a kind of fish’
/bɔ:ɲom/ ‘ties up’
/kɑ:bɑ/ ‘edible fruit’
/kɔ:bo/ ‘skin’

/ɑ/ : /ɑ:/
/ɑrɑttɑ/ ‘convince’
/ɑrɑ:ttɑy/ ‘mensuration’
/cɑo/ ‘dog’
/cɑ:o/ ‘rain fish’

/e/ : /e:/
/jibeʈ/ ‘small bat’
/jibe:ʈ/ ‘swallow’ (N)

/i/ : /i:/
/jili/ ‘snail’
/ji:li/ ‘aunt’ (father’s brother’s wife)

/o/ : /o:/
/bol/ ‘fish’
/bo:l/ ‘cane, fresh water snake’

/u/ : /u:/
/burul/ ‘white fish’
/bu:rul/ ‘a kind of fish’
/bun/ ‘sea shell’, ‘a kind of snail’
/bu:n/ ‘spoon’

Although there is a minimal pair [bun] and [bu:n] available in the lan-
guage, distinction in length is neutralised for some speakers as we also
find [bun] ~ [bu:n] ‘sea shell.’

2.1.3 Phonotactics of Vowels

The phonotactics of the vowel sounds in initial, medial and final positions
can be presented as follows in Table 2.2.
 phonetics and phonology 41

Table 2.2. Distribution of vowels

Segments Initial Medial Final
/i/ ikko ‘burn’ biʈhe ‘ashes’ subi ‘snake’
/i:/ i:ople ‘light’ ʃi:d ‘hunt’ —
/e/ ekko ‘open’ ʈhumel ‘honeycomb’ i:eke ‘to hold’
/e:/ e:ʃɔrɔ ‘to grind’ le:c ‘arrow’ biʈhe: ‘ashes’
/ε/ εr ‘genitive’ rεphe ‘rice’ ʈɛlnε ‘type of
mosquito’
/ε:/ — jicε:r ‘rain’ εsɔrε: ‘sing’
/ɑ/ arakaʈa ‘dwarf’ kholam ‘to laugh’ ʈhεr boa ‘my lips’
/ɑ:/ a:ʈ ‘wood’ ʈhoŋ kena:p ‘my fingers’ -ɔkka: ‘benefactive’
/ɔ/ ɔn ‘a jungle tree’ ʈɔkhɔtec ‘leaf’ cɔrɔlɔ ‘bird’
/ɔ:/ ɔ:tɔ ‘day’ bɔ:r ‘air’ ʈɔ: ‘sky’
/o/ oʈʈa:lar ‘bald’ boco ‘peel’ εr-ino ‘tears’
/o:/ o:cɔ ‘net’ pho:ŋ ‘mouth’ kuro: ‘palm tree’
/u/ ulukhu ‘python’ ʈhuŋ kara ‘my nail’ ʈhiliu ‘cloud’
/u:/ — ʈhεr bu:o ‘my ear’ —

2.1.4 Vowel Sequences/Clusters

PGA is rich in vowel sequences or vowel clusters. Vowel clusters involv-
ing both short and long vowels were attested. In addition, some speakers
use an epenthetic semi-vowel between short vowels within a word. For
example, some speakers use the palatal semi-vowel y as in ia > iya, or the
rounded back semi-vowel w as in ua > uwa, perhaps because of the influ-
ence of Hindi. However, only a few speakers use such epenthetic inser-
tions and they do not use them consistently. One of our speakers, Peje,
never used a semi-vowel between two vowels. Some instances of three
vowel sequences were also attested. However, in the current form of the
language, variation exists, e.g.

[coɑi ~ cowɑi ~ cowɑ:i] ‘clam’

Sporadic incidences of four vowel sequences [ioeo] were also noticed in

compounds:

mio.eone ‘lemon juice’

Consider the following Tables 2.3 and 2.4.

42 chapter two

Table 2.3. Vowel clusters in PGA

Two vowel sequences Three vowel sequences
(mostly word-finally)
Initial front vowel Initial back vowel aie [ɛncɑie] ‘spoil’
aou [ɛr=lɑoyu] ‘tattoo on the
ia, iu, io , i:o, ie, i:e, iɛ, ua, uo, uɔ, ui, ue forehead’
ei, eo, eɔ, ea, eu oa, o:a, oɔ, o:ɔ, ou, oi, aia [mɑiɑ] ‘Sir’
oe, o:e, oε aeo [a-mɑe.ot-ɲo] ‘house of
εo, εo, ɛi, ɛ:i, ɛɔ ɔi, ɔ:e, ɔe, ɔɛ, ɔo, ɔa our forefathers’
eia [eiɑ] ‘OK’
eoɛ [pheoɛ] ‘movement of
turtle in water’
iue [ɛr=liue] ‘shave’ ‘clean’
ieu [finlieu] ‘pulsating pain’
ioi [e=mioi] ‘sour’
oio [e=boio] ‘ripen’
ao, a:o, a:u, au, aε, a:e,
ae, ai

A variety of combinatory possibilities of vowel clusters in all positions is

a rather striking feature of the language. Other languages of the Andaman
Islands, namely Onge and Jarawa do not show such a pattern (Abbi 2006).
Consider Table 2.4.

Table 2.4. Phonotactics: two vowel sequences

V+V Initially Medially Finally
ia [iɑretɑ:ke] ‘study’ [khiɑrɑsene] ‘tall person’ [buliɑ] ‘a kind
of song’
ie [ie] ‘venom’ [phinliem] ‘have throbbing [itkie] ‘turn’
pain’
iɛ [iɛke] ‘catch’ [uliɛm] ‘whistle’ —
iu [iulo] ‘loose’ [eliutɛse] ‘give someone a [biu] ‘incense’
name’
io [iome] ‘shower’ [lioʈ] ‘a kind of red fish’ [tɑjio]
‘creature’
i:o [i:ople] ‘light’ [cokbiot-homo] ‘turtle flesh’ —
ea — [reɑʈɔlo] ‘onset of ‘mild rain’ [ɑleɑ] ‘slowly’
ei [ei] ‘vomit’ [keip] ‘coloured mixture’ [birei] ‘bat’
eo [eone] ‘juice’ [ekceoxe] ‘row’ [meo] ‘stone’
eɔ [eɔle] ‘see’ [teɔleɔ] ‘small knife’ [jelojeɔ]
‘evening tide’
 phonetics and phonology 43

Table 2.4 (cont.)

V+V Initially Medially Finally
eu [eure] ‘song’ — [bɑleu] ‘go
hunting’
ɛi [ɛibiʈki] ‘carrier’ [mikhutɛic] ‘stomach pain’ [bɛi] ‘liquor’
ɛ:i — [lɑotɛ:itutbec] ‘a migratory [rɛ:i] ‘puberty’
bird’
ɛo [ɛople] ‘light’ [rɛobol] ‘rope attached to [ɛlɛo~ɔ] ‘infant’
spear’
ɛɔ [ɛɔrɔkhui] ‘young man’ — [rɛɔ] ‘spear’
oa — [ɑmɑetoɑthumimi] ‘aunt’ [ʈoɑ] ‘layers of
stone’
oi [oinʃiŋ] ‘pant’ [boitɛrcek] ‘kind’ [ɑrɑcoi]
‘outsiders’
oɛ/oe — [ juroetɑtoŋ] ‘dancing [iboɛ] ‘refined
ground’ oil’; [noe] ‘hit’
oɔ — — [ɑkɑtoɔ]
‘choke’
ou — [ɖekhoulekɔ] ‘ejaculation’ —
ɔe/ɔ:e — [kɔ:el] ‘later’ [mɔʈɔkɔrɔe]
[kɔemo] ‘lice’ ‘left leg’;
[cɔ:e] ‘a kind of
tree’
ɔɛ — — [ɛrbelɔɛ]
‘pimples’
ɔi — [kɔrɔiɲ] ‘dugong’ [mɔrɔi] ‘water
current’
ɔo [ɔo] ‘sky’ [ʈɔo.ukhe] ‘sky god’ [ʈɔo] ‘sky’
[ɖekʰɔokʰɔŋɑ] ‘aware’
ɔa — [rɔɑtɛrkhuro] ‘a large boat’ [okɔʈɔɑ]
‘dugong-chest’
ui [uilem] ‘piss’, [ϲɔkhuirɑbelo] ‘ring finger’ [bui] ‘North
[uiboʈ] ‘shoot’ Andaman’;
‘a kind person’
ue [uem] ‘draw water’ [eruluettol] ‘spectacles’ [ɑ:ʈtuʃue] ‘burn
logs’
uo — [ɖuoc’] ‘hear’ [ɛrbuo] ‘ears’
ua [uɑge] ‘a shell’ — [mɔtkocuɑ]
‘divine figure’;
[luruɑ] ‘Bluff
Island’
au — [ɑrɑuli] ‘tail’ —
ao [ɑone] ‘sit’ [tɑlɑog] ‘a kind of white [cɑo] ‘dog’
[ɑo] ‘a kind of leaf ’ clay’
44 chapter two

Table 2.4 (cont.)

V+V Initially Medially Finally
ai — [ɑkɑile] ‘return’ [bɑlɑʈbɑi]
[etɑiʃe] ‘keep’ ‘Andaman
Drongo’
ɑe ~ ɑ:e — [ɑrɑcɑecɑʈo] ‘slow learner’ [ɑlɑe] ‘palate’;
[lɑ:e] ‘sugar-
cane’

It is observed that no two identical vowels occur in any position. For fur-
ther information refer to section 2.6 on morphophonemics.

2.2 Semi-vowels/Glides1

In most Indian languages, except those of the Tibeto-Burman group, an

epenthetic semi-vowel can be inserted between two vowels. PGA has both
kinds of structures. It has words with two vowel clusters which form two
syllabic peaks, as well as structures with a semi-vowel inserted between
two vowels (refer to §2.1.4 above).
It has been noticed that PGA also offers ample examples of glide inser-
tion between two vowels. This appears to be a recent phenomenon due
to contact with Hindi which does not allow vowel clusters. Hindi is the
first language of the majority of the Great Andamanese members. Those
for whom it is not the first language use it for communicating with their
children. Documentation of some of the North and Central Andamanese
languages by Portman (1887) also does not show any glide insertion. One
can speculate that the former structure with VV clusters is older than
the latter.
Consider Table 2.5.

Table 2.5. Intervocalic semi-vowels

V-SV-V Glide insertion Semi-vowel
ɑo-u [erlɑoyu] ‘mark’, ‘wrinkles’ y
ɑ-ɑ [kɑcɑwɑ] ‘broken stones’ w, y
[mɑyɑ] ‘deceased’

1 I am using y for the palatal semi-vowel and j for the palatal voiced stop throughout
the text.
 phonetics and phonology 45

Table 2.5 (cont.)

V-SV-V Glide insertion Semi-vowel
ei-ɑ [eiyɑ] ‘interjection’ y
e-o [khotewo] ‘get away’ w, y
[obeyo] ‘sting’
ɛ-o [kɛwo] ‘a kind of red crab’ w
i-ɑ [diyɑ] ‘3pl.(proximate)’ y
i-e [itkiye] ‘turn’; [erɑɲiye] ‘gather’ y
i-u [ɑrɑliyu] ‘pull out’ y
i-ɛ [iyɛke] ‘roast’ y
i-o [iyome] ‘shower’; [firiyo] ‘boil’ y, w
[liwo] ‘say’
o-ɑ [rowɑ] ‘boat’, w, y
[boyɑ] ‘outrigger’
oi-o [olɑitoiyo] ‘stagnant’ y
o-e [soye] ‘the sui tree’ y
o-o [burɔkowo] ‘Andaman Crake’ w, y
[koyop] ‘impotent’
ɔ-e, ɔ:-e [cɔye] ‘catch by force’; [cɔ:ye] ‘father’s sister’s y
husband’
ɔ:-o [ettɑ:rɔ:wo] ‘fence’ w
ɔ-ɑ, ɔ:-ɑ [pɔwɑ] ‘leftovers’ w, y
[yerɔ:yɑto] ‘a kind of tree’
u-e [uyem] ‘draw water’ y

Word-final semi-vowels or glides have been attested in the language.

While the semi-vowel w can occur after both front and back vowels, its
palatal counterpart y occurs only after back vowels, word-finally. Consider
Table 2.6.

Table 2.6. Words that end in semi-vowels

Final -y/-w Examples
-ow [ʈow] ‘a kind of tree’
-oy [co:y] ‘a kind of tree’
-o:y [irxuɖo:y] ‘circle’
-ɔw [ebicɔ:w] ‘posterior of a dugong’
-ɔy [rɔy] ‘salty’; [ewoʈɔ:y] ‘shoulder bone’
-ɔ:y [cɔ:y] ‘grandmother’
-ɑw [formucɑ:w] ‘lion, tiger’
-ɑy [kolɑy] ‘a kind of shell’
-uw [ertiyuw] ‘cap’
-uy [ullu:y] ‘whistle’
-ew [kɔrole:w] ‘torch’
46 chapter two

The occurrence of word-initial semi-vowels is also observed in PGA. Note

that there are no words which begin with the labial semi-vowel w.

Table 2.7. Words that begin with semi-vowels

Initial y- Examples
ye- [yeke] ‘hold’
yɑ- [yɑkhi] ‘pain’, ‘hurt’
yo- [yoŋɑʈpo] ‘load’
yɔ- [yɔ] ‘god’
yu- [yulu jili] ‘a flower’

2.3 Consonants

As far as the consonants are concerned, some unusual sounds2 were

observed, such as bilabial fricatives, both voiced and voiceless [β], [Φ],
and the labialised lateral [lw] (at least in one speaker, Peje). (Refer to
Appendix ‘B’). PGA does not have the voiced velar sound [g] and the
voiceless glottal fricative [h]. The former sound was present in the Middle
Andaman languages such as Pujjukar/Puchikwar (Basu 1952) which is now
extinct and in the languages of South Andaman (Man 1875–1878, 1923).
This sound was observed in songs sung by old people, as well as in the
name of one of our consultants—‘Golat’, therefore the sound must have
existed in one of the Great Andamanese languages spoken in the past.
Conversely, we noticed the acquisition of [h] from Hindi in the speech of
some of the younger speakers. The following sounds are in free variation
at the intra-community level, i.e., within the same clan, sometimes by the
same speaker.

Table 2.8. Intra-community variation of unusual sounds

[Φ ~ ph ~ f ]
[β ~ l ~ w ~ lw]
[kh ~ x]
[s ~ ʃ ~ c ~ ch]
[r ~ ɽ]
[t ~ ʈ]

2 “Unusual” because no other Indian language contains these sounds.

 phonetics and phonology 47

Table 2.9 gives an overview of variation among the speakers. The claimed
father tongue(s) by each speaker is given in parentheses.3 The sounds in
free variation vary from speaker to speaker, thus confirming the hypoth-
esis that the existing speakers of the language are not descendants of the
speakers of one language, but instead are the descendants of speakers of
different varieties of the same language family. In other words, PGA rep-
resents inter‐group free variation. However, the substitution of sounds in
free variation does not disturb the comprehension of the word under con-
sideration. One can see that the speech of the speaker Noe is influenced
by her husband, Peje, who comes from a mixed background of Khora
and Jeru.

Table 2.9. Phonetic variation across community members

Surmai ( Jeru) F ʈ ʃ ph b l ɑ r
Boro Sr. (Khora) F t ch Φ b l ɑi ɽ
Boa Sr. (Bo) F ʈ ch ph b l ɑ r
Noe ( Jeru) F ʈ ʃ f β lw ~ l ɑ r
Peje (Khora and Jeru) M ʈ ʃ Φ β lw ɑi r
Lico (Sare and Khora) F ʈ s ph b l ɑ r
Nao Jr. ( Jeru) F ʈ ʃ ph b l ɑ r

The intra‐community variation results in a large number of sound inven-

tories, as shown in Table 2.10 on consonants. The non-phonemic sounds
that occur less frequently are given in brackets in the table. Our consul-
tant, Peje, for example, does not have the lateral [l] in his verbal repertoire
as he invariably uses [lw] in all positions where the other members use [l].
This could be an idiosyncratic feature of Peje’s speech or it could be a relic
of the extinct languages of the family. Refer to appendix ‘B’.
Considering the variations and the unusual sounds which are present
in PGA, it is categorised with certain languages that are marked as having
‘heavy’ or ‘larger’ inventories of sounds, and by implication, (Lindblom &
Maddieson 1988)4 PGA could be said to consist of complex consonants.
PGA seems to have changed since Manoharan (1989) reported his work.
He had transcribed many of the words with the labio-dental fricative [f ],

3 The name of the language(s) and the names of the original sub-tribe that a speaker
belongs to are the same.
4 As mentioned in The World Atlas of Language Structures (Haspelmath et al., 2005).
48 chapter two

while our speakers mostly used the voiceless bilabial aspirated plosive
[ph]. Some speakers also used the voiceless bilabial fricative [Φ]. Specifi-
cally, Peje and his wife Noe produced this sound.5 The voiced counterpart
[β] was also sporadically attested in the speech of one or two consultants.6
One can safely say that the bilabial fricatives are on their way to extinc-
tion as our database has very few words with these sounds. They seem to
be replaced by bilabial aspirates. Similarly, the voiceless velar fricative [x]
has changed over the last forty years to a voiceless velar aspirated plo-
sive [kh] due to Hindi influence. We have tried our best to capture both
of these sounds (refer to Abbi 2012). Only one speaker, a woman aged
76 called Boro, who was originally a Khora speaker, used the retroflex
trill (we use the symbols [ɽ] for this in the text): this also has been docu-
mented. Sounds such as [Φ, β, ɽ, x, lw, w] that occur least often are given
in parentheses. All other sounds occur at phonemic level.

Table 2.10. Consonants of Great Andamanese

Place: Bilabial Labio- Dental Alveolar Retroflex Palatal Velar
Manner: Dental
Plosives p b t d ʈɖ c j k
Aspirated ph th ʈh kh
plosives
Nasals m n ɲ ŋ
Trills r (ɽ)
Fricatives (Φ) (β) (f ) s ʃ (x)
Laterals (lw) l
Approxim- (w) y
ants/glides

The interchangeability of sounds, at times, occurred within the speech

of the same speaker, indicating the sound in transition. For instance, our
main consultant Nao Jr often varied between dental [t] and retroflex [ʈ] as
in the word ara-ɖileʈmo ~ ara-ɖiletmo ‘bladder’ etc. These variations exist
despite the fact that the two sounds stand in contrast. The only reason
that can be ascribed to such variation by the same speaker is the fact that

5 My recent visit to Andaman in 2009 showed that Noe had lost [Φ] in favour of [ph]
due to contact with her fellow community members who also use [ph] under the influ-
ence of Hindi.
6 Our consultant Golat used [β] although he was the son of Boro Sr., the Khora speaker
who had [b] in her verbal repertoire.
 phonetics and phonology 49

the language is dying fast and speakers no longer use it in their daily life.
This lack of use of the language and indifferent attitude towards the use
of the heritage language seem to be a significant reason for the genesis of
sound variation in the lexicon as well as in syntactic constructions.7

2.3.1 Phonemic Contrasts:

Minimal and Sub-Minimal Pairs for Consonants
Despite the intra-community variation, we could attest some minimal
pairs to show clear-cut phonemic contrasts.

2.3.1.1 Nasals
There are four distinct nasal consonants at the phonemic level. /m/: /n/:
/ŋ/ : /ɲ/

/m/ bilabial nasal

/n/ alveolar nasal
/ɲ/ palatal nasal
/ŋ/ velar nasal

These four nasal consonants contrast with each other in initial, medial
and final positions.

kʰuŋi ‘a fish’
kʰun ‘bring’
kʰum ‘side shoulder’
biŋo ‘think’
bino ‘sleep’
ben ‘back shoulder bone’
beŋ ‘forehead’
ŋa 2sg
ɲa ‘eat’
ɛkʈɛŋe ‘sputum’
ekʈɛne ‘pull it’

2.3.1.2 Oral Stops

There are thirteen oral stops, most occurring in contrasting pairs. The lan-
guage has ample examples of voiced and voiceless dental and retroflex

7 For further information on the speakers and instances of use of the above variation,
readers should consult A Dictionary of the Great Andamanese Language (Abbi 2012) which
contains all the citations for the aforementioned variations.
50 chapter two

stops. Only the voiceless sounds are aspirated. We will first consider ante-
rior and coronal sounds. /t/ : /th /: /ʈ / : /ʈh/: /d/

/t/ voiceless unaspirated dental stop

/th/ voiceless aspirated dental stop
/ʈ/ voiceless unaspirated retroflex stop
/ʈh/ voiceless aspirated retroflex stop
/d/ voiced dental stop
/ɖ/ voiced retroflex stop

Minimal and subminimal pairs can be found to determine the phonemic

status of these sounds.

tec ‘leaf ’
theca ‘intoxicated’
tɛkho ‘speech’
dekhɔ ‘now’
ʈhu 1sg
thu ‘reach’
bathe ‘hind of an animal’
bate ‘extinguish’
eka-ʈoke ‘necklace’
eka-ʈʰoke ‘close it’
kata ‘a piece’
kaʈa ‘girl’
dun ‘they’
ɖum ‘earthworm’
dakar ‘basket’
ɖakar ‘potato’

The velar sound exists only in its voiceless form. As we discussed in §2.3
the voiced velar [g] is conspicuous by its absence in the present form
of the language but the possibility of its existence in the earlier form can-
not be ruled out.

/kh/ : /k/
/kh/ voiceless aspirated velar stop
/k/ voiceless unaspirated velar stop
khu ‘drink’
ku ‘burn’
ɛr=ʈɔko ‘slowly’
ɛr=ʈɔkʰo ‘fishing with bamboo’
kharae ‘crab, found only in sand’
karai ‘black ant’
 phonetics and phonology 51

kɔrɔ ‘palm’
khɔrɔ ‘sand’

There are three contrastive bilabial consonants. Voiced aspirates could

not be attested in the language.

/p/ : /ph/ : /b/

/p/ Voiceless unaspirated bilabial stop
/ph/ Voiceless aspirated bilabial stop. This sound alternates with the following
two sounds in the intra-community speech. Refer to Table 2.9 to view the
pattern of variation.
[f ] Voiceless labio-dental fricative
[Φ] Voiceless bilabial fricative
Hence the word phile ‘teeth’ would have alternative forms such as: [phile] ~ [Φile]
~ [file].

The alternating pattern that we observed between [ph], [Φ] and [f ] in

Table 2.9 is also indicative of the age factor. The younger generation uses
the fricative [f ] word-initially while the older generation uses either [ph]
or [Φ] in all positions. Similarly, variation between the voiced bilabial
stop [b] and voiced bilabial fricative [β] was observed across different age
groups, the latter was used by elder and the former, i.e., [b] by younger
people.

/b/
[b] voiced bilabial stop
[β] voiced bilabial fricative

The following examples help us establish the phonemic status of /p/, /ph/
and /b/, the latter varying with its fricative counterpart.

puro ‘whetstone’
phuro ‘a kind of owl’
buro [buro] ~ [βuro] ‘hand movement’
bol ‘rope’
phol ‘love someone’

/c/ : /j/

/c/ voiceless palatal stop. It has a variant in [ch] which is a voiceless

aspirated palatal affricate. Consider:
52 chapter two

otcar [otcɑr] ~ [otchɑr] ‘middle of the chest’

onco [onco] ~ [oncho] ‘to stitch’

Aspirated and unaspirated versions vary freely between speakers. The aspi-
rated version was observed in the speech of Boa Sr., our only Bo speaker.
Sporadically the Khora speaker, Boro Sr. also used the sound [ch] where
other speakers used [c].8 This has been noted in Table 2.9 above.

/j/ voiced palatal stop

co ‘seed’
jo ‘song’
cul ‘near’
jol ‘crawl’

2.3.1.3 Fricatives
/s/ : /ʃ/
/s/ voiceless alveolar sibilant
/ʃ/ voiceless palatal sibilant
sorobul ‘enemy’
ʃoro-kom ‘kill with an arrow’

Speakers, such as Boa Sr., were observed using [ch] instead of [ ʃ ] word-
initially as in chime ‘soak’ while others varied between [sime] and [ʃime].
As observed in Table 2.9 above, the sounds [s], [ ʃ ], [c] and [ch] varied
freely between the members of the Great Andamanese community as
each of them had descended from a different linguistic background.

2.3.1.4 Liquids
/l/ : /r/
/l/ voiced alveolar lateral. This sound has an alternant in
[lw] which is a voiced labialised lateral.
luro [luro] ~ [lw uro] ~ [luɽo] ‘fire’
phal [phɑl] ~ [phɑ lw] ‘sea wave’

The laterals in PGA need further discussion. An acoustic analysis was

made of comparative laterals across the speakers and can be consulted
in appendix ‘B’. As stated earlier, except for one elderly speaker, Peje, no
one else used the sound [lw].

8 It was observed that the two speakers were intimate friends and spent most of their
time together. Influence of Boa’s speech on Boro’s cannot be ruled out.
 phonetics and phonology 53

/r/ voiced alveolar trill

[ɽ] voiced retroflex trill
ɖakar [ɖɑkɑɽ] ‘a kind of potato’
phar [phɑr ~ phɑɽ] ‘injury’

Boro Sr., who hailed from the Khora tribe, used the retroflex version of the
trill more often than not. The trill /r/ contrasted with lateral /l/.

juro ‘goddess’
julu ‘clothes’
le ‘crab’
re ‘flower’

2.3.2 Phonotactics of Consonants

PGA is a language that allows most of the consonants to appear in all
positions with varying degrees of frequency. However, aspirated sounds
do not occur word-finally. Final [d] and [ɖ] are rare as we have only two
words each for these sounds in our database. As can be observed, gemi-
nates do not occur word-initially nor word-finally, a phenomenon which
is shared across most of the world’s languages.

Table 2.11. Distribution of consonant sounds

Segments Initial Medial Final
p pɔrɔk ‘big’ εɽeptɔ ‘weak’ lep’ ‘smoke’
ph ~ Φ pha:r ~ Φar ‘wound’ ε-bɔpho~ɛ-bɔΦo ‘stupid’ ʈɔph ‘bathe’
b boʈho ‘fall’ cɔkbi ‘turtle’ kheb ‘left’
β ~ lw -βuc ~ lw uc ‘cold’ i:βe ‘mother-in-law’ phaβ ~phalw ~phal
βuɽo ~ lwuɽo ‘fire’ ‘sea wave’
ββ — kɔββɔ ‘bamboo’ —
t tole ‘potato’ oto ‘morning’ kɔt ‘cough’
tt — otto ‘see’ —
th thu ‘reach’ er=bathe ‘forelegs of —
turtle, fish fins, flippers’
d dakar ‘basket’ khidεr ‘coconut’ ethud ‘kill’
ɖ ɖakar ‘potato’ fuɖoʈɔŋ ‘mangrove tree’ ʃiɖ ‘hunt’
ʈↄkhɔtec ‘leaf ’ phaʈka ‘crow’ rεʈ ‘bamboo’
— oʈʈo ‘smell’ —
ʈh ʈhu ‘I’ boʈho ‘fall’ ɛrɖiʈh ‘pierce’
c cεrpho ‘sneeze’ εrlɔcɔŋ ‘horn’ βε:c ‘arrow’
cc — uccire ‘clean’ —
j jurε ‘dance’ ijurul ‘sweet’ teɲj ‘spit’
54 chapter two

Table 2.11 (cont.)

Segments Initial Medial Final

jj — ujjete ‘shy’ —
k kɔtɔrɔ ‘a kind of phorɔ:ke:ʈ ‘heaven’ ʈhεrcurɔk ‘my
fish’ knee’
kk — diyɔ:kkɑ ‘for this’ —
kh khole ‘to laugh’ εrkhuro ‘elder’ ~ xuro ‘big’ lebetɔkh ‘to fan’
m mu ‘we’ (excl) ʈhumεl ‘bee’ ɖirim ‘black’
mm — immixolo ‘inside the —
bark of a tree’
n nu ‘people’ ino ‘water’ kεren ‘whale’
nn — elinna:wbe ‘sleeping —
place’
ɲ ɲyɔ ‘house’ suɲu ‘upward’ kataɲ ‘star’
ŋ ɲu ‘you’ taŋol ‘rod’ ʈɔŋ ‘tree’
r rowa ‘boat’ phuro ‘owl’ ɖakar ‘a type of
potato’
ɽ ɽulu ‘eye’ εɽla ‘alone’ bɔɽtεɽ ‘storm’
l lico ‘basket’ jili ‘snail’ ɛmphil ‘die’
ll — jilli ‘menstruating girl’; —
‘a kind of blue flower’
lw lwec ‘arrow’ bi: lw u ‘ship’ bolw ‘rope’
s sare ‘salt’ isɔrɔke ‘kill’ ɖe:s ‘hazy’
ss — esso:ro ‘beat’ —
ʃ ʃi:d ‘hunt’ təraiʃitu ‘dusk’ —
y yejile ‘ate up’ thikyoŋ ‘to live’ rεoʈɔy ‘iron’
w wεtεŋe ‘smell’ rowa ‘boat’ ca:w ‘dog’

2.3.3 Medial Consonant Clusters

Great Andamanese has a wide variety of medial CC clusters within a lex-
eme. Refer to Tables 2.12 and 2.13.

Table 2.12. Consonant clusters within a lexeme

CC Lexeme Gloss
-pl- ɛople ‘light in weight’
-pt- eptaruoŋa ‘step-child’
-pb- bilupbu ‘remind’
-pl- kaplo ‘mangrove seed’
-tl- ecɛrʈɔk-totla ‘industrious’
-tp- kɔtpɛc ‘horn’
-tb- itbi ‘yellow’
 phonetics and phonology 55

Table 2.12 (cont.)

CC Lexeme Gloss
-tr- notroʃup ‘noisy’
-tc- rɛtcɛr ‘onset of rains’
-tk- itkie ‘turn itself ’
-dg- jidga ‘a flower’
-tc- enɔl-ʈɔtcɔ ‘very good’
-ʈm- kɛʈmo ‘bowstring’
-ʈp- yoŋaʈpo ‘load’
-ʈm- eʈɔlɔʈmo ‘white’
-ʈl- ciriʈli ‘turtle’
-ʈk- ebiʈki ‘balance’
-ɖd- biɖde ‘edge of the sea’
-ɖb- koɖbelo ‘Banded Sea Krait snake’
-cph- kharacphak ‘very far’
-cm- cɔkhɔcmo ‘a green seabird’
-kr- ekrɔ ‘got the prey’
-kl- cɔkle ‘an insect’
-kb- cokbi ‘turtle’
-mph- ikoɖumphu ‘big’ (animal)
-mb- ɛmboe ‘marry’
-mt- julumte ‘early tomorrow morning’
-nm- kɔnmo ‘white tuber’
-nl- phinli ‘throbbing’
-nt- dɔntɔ ‘therefore’
-nk- cɛnkolo ‘sea shell’
ekhonko ‘young girl’ (Bo)
-ŋʈ- ʈaŋʈɔ ‘lean’
-ŋk- tuŋkɛlo ‘all over’
khiriŋkoso ‘Strait Island’
-ɲʈ- boiɲʈaiŋa ‘name of a place’
-lph- ilphe ‘one who returns from hunting’;
‘turn over’
-lm- kɔlme ‘there’
-lt- olaltoiyo ‘stagnant’
eltite ‘egg yolk’
-lk- er-pholkɔe ‘bigamous’
-lŋ- bolŋo ‘a kind of big leech’
-ln- ilni ‘nerve’
-rʈ- nutlu:rʈɔ:y ‘most’
kɔrerʈo ‘a kind of game’
-rt- burto ‘a kind of fish’
-rʃ- mirʃi ‘uprooted’
-rb- korbo ‘a kind of sour fruit’
-rm- phormu ‘big’
56 chapter two

Table 2.12 (cont.)

CC Lexeme Gloss
surme ‘having a good voice’
-rn- kharne ‘a kind of Manta Ray’
-rŋ- kharŋe ‘sweet potato’
-ɽl- ɛɽla ‘alone’ [Khora]
-ɽc- khuɽcu ‘police’
-rl- cɛrle ‘a kind of fly’
-rk- burku ‘coral’
-sl- isluye ‘rotten’

Table 2.13. Geminates

CC Lexeme Gloss
-ββ- ɛββeo ‘a variety of bamboo’
-tt- ekka:la:ttoy ‘pungent taste’; ittakhe ‘right now’
-ʈʈ- eʈʈɔrom ‘salty’
-cc- era:cco:cce ‘on the whole’; uccire ‘clean’
-kk- ekka:ɖum ‘thin’; ikko ‘burnt wood’
-ll- iɖello ‘pebbles’
-jj- ujjete ‘shy’
-nn- elinna:wbe ‘sleeping place’
-mm- immixolo ‘inside the bark of a tree’
-ss- esso:ro ‘beat’

2.3.4 Initial Consonant Clusters

A few cases of two-consonant clusters in word-initial position were
observed. Only voiceless stops and trills constitute part of the onset of a
syllable.

trera ‘shining’
trɛpho ‘much’
trai ‘fever’
khriŋkoʃo ‘Strait Island’

2.4 Syllables

It is generally believed that complex syllable structures are those that

allow two or more consonants after the vowel (Maddieson 2005).
Following this definition, PGA has complex syllable structures. A large
 phonetics and phonology 57

number of consonants can be combined to form inter-vocalic consonant

clusters. See Table 2.12.
Syllables have the following structure:

(C) (C) V (V) (C) (C) (V) (V) (V) (C) or

C2 V1 C2 V3 C1

Thus, only one vowel (V) is obligatory to form a meaningful syllable in

PGA. Furthermore, as we saw earlier, syllables can involve consonants
(C) in the onset (i.e. the beginning) or in the coda (i.e. the end) of the
syllable. Consonant clusters are rare in the onset position except in the
combination [tr-] and [khr-]. There are ample examples of intervocalic
consonant clusters within a morpheme. Refer to Table 2.12 to see the vari-
ety of consonant clusters which occur morpheme-internally, but with a
syllable break in-between. Consider Figure 2.1 which shows the syllable
break.

/e biʈ ki/ /bi lup bu/ /ɛm bo e/

σ σ σ σ σ σ σ σ σ

/ɛr li u e/

σ σ σ σ
Figure 2.1. Syllable breaks

2.4.1 Possible Syllable Structures

Most simple syllable structures, i.e. one obligatory vowel, are seen in
monosyllabic words denoting possessive body class markers or proclitics.
The canonical form of a root is disyllabic. As a consequence, the following
syllable structures are possible in simple words:
58 chapter two

Table 2.14. Possible syllable structures

Forms Examples Meaning
V a- cl 1.poss
VV ie ‘catch’, ‘give’
VVV aia ‘Sir’
VC ot- cl 4.poss
CV ɲo ‘stay’, ‘live’
CVV boi ‘ask’
CVVC ɖuoc ‘hear’
CVC bɔr ‘breeze’, ‘air’
VCV uno ‘sit down’
VVCV eole ‘see’
CVCV cone ‘go’
CVCVC ŋemaʈ ‘run’
CCVCV trɛpho ‘climb (tree)’ [rare]
CVCCV jermo ‘worm’
CVVV taie ‘a kind of bird’

2.4.2 Constraints
All consonants can appear in the onset of a word. The presence of initial
[ɽ] was observed in a few words in Khora. All consonants, except aspi-
rated consonants and the sibilant [ ʃ ], can end a word. There are only
three words in our database that end in the voiceless sibilant [s]. The
large inventory of consonants that appear word-initially and word-finally
can be ascribed to the fact that the language is formed from four different
varieties and has included several kinds of words in its lexicon (see Abbi
2012). As mentioned in §2.3.4, initial consonant clusters appear sporadi-
cally in [tr-], as in trɛpho ‘much’ and in [khr-] as in khriŋkoʃo ‘Strait Island.’
Medial consonant clusters are in abundance.

2.4.3 Length of a Word

As affixation and compounding are both productive word formation
devices in PGA, a complex word can be as long as six or seven syllables:

a-ka-ker-tot-lɔcoŋ ‘uvula’
v-cv-cvc-cvc-cv-cvc
1 2 3 4 5 6

a ra- ɖo mo-ʈot-kɔ bo ‘skin of scrotum’

v-cv-cv-cv-cvc-cv-cv
1 2 3 4 5 6 7
 phonetics and phonology 59

Names of birds, fish, insects, reptiles and other jungle creatures provide
most of the complex words with long syllable structures.9 At times, two
consonant clusters may appear within the same word, as in trɔkʈɔimo
‘long-tail cricket’ or kulʈɛʈmo ‘wasp’ (Vespa affinis).

2.5 Morphophonemics

2.5.1 Gemination
We have seen that geminates occur intervocalically and that they also
conform to the syllable division given above. Geminates are created in
PGA by progressive assimilation. Thus:

t> k /--- +k = it + ko > ikko ‘burnt wood’

t> t /--- +t = it + tɑkhe > ittakhe ‘right now’
t> ʈ /--- +ʈ = it + ʈɔrɔm > iʈʈɔrɔm ‘salty’
t> c /--- +c = ut + cire > uccire ‘clean’
t> j /--- +j = ut + jete > ujjete ‘shy’
t> s /--- +s = et + so:ro > esso:ro ‘beat’

The gemination of sonorant sounds such as nasals and liquids occurs

within a lexeme. For instance, immixolo ‘inside the bark of a tree’; elinna:w-
be ‘sleeping place’; i-ɖella ‘pebbles’, etc.

2.5.2 Degemination and Compensatory Lengthening

Some speakers (Lico for example) were observed to degeminate conso-
nants with compensatory lengthening in the next syllable.

vciciv → vcv: = attaekhe > ata:ekhe ‘lie’

c v
[+ long] → [– long] / [+ long]

The compensatory length can be assigned either to the preceding or to

the following vowel which is rather unusual from a cross-linguistic per-
spective.

v + ccv → v: + cv
bi + ikko > be:ku-abs + burn ‘burnt’

9 See Pande & Abbi (2011).

60 chapter two

2.5.3 Homorganicity
Nasals acquire the features of the next consonant across the morpheme
boundary. That is, nasal consonants assimilate to the place of articulation
of the following obstruents. This can be stated in a rule:

+ consonant
α anterior + consonant
+ nasal
β coronal α anterior
γ back + β coronal
γ back

This can be exemplified by:

[ɑm + khui] > aŋkhui ‘embrace’

[ɑkɑ-εm-dukolƆt] > akandukolɔt ‘young girl’

2.5.4 Insertion of a Consonant

If two different vowels occur across lexemes or morphemes, most often
than not the consonant appearing in the second lexeme or morpheme is
inserted between the vowels.

[dilli + ek] > dillikek ‘towards Delhi’

2.5.5 Vowel Harmony

If the word is constructed of two morphemes with dissimilar vowels,
the vowel of the second morpheme agrees with the first one in fronting.
Thus:

[duoč + ɛm] > [duosɛm] > duosom ‘listen to me carefully’

2.5.6 Vowel Lowering

a) A high front vowel is lowered in front of another high front vowel across a
morpheme boundary.
i → e /--- + i
u → o /--- + u

For example,
bi + ikko > be + ikku abs + burn ‘burnt’
bi + ijiom > be + ijiom abs + eat ‘eats’
ʈhu + ut cone > ʈho + ut cone ‘I go’
 phonetics and phonology 61

2.5.7 Vowel Deletion

(a) Two identical vowels across morpheme boundaries become one. Thus:
Vi → ø / --- + Vi
aka + aom > akaom 3sg + comes ‘he comes’
ino + ot=khu > inotkhu water + cl 4=drink ‘drinking water’
(b) Non-identical vowels across morpheme boundaries tend to lead to a loss of
the second vowel before the consonant.
Vi → ø / Vx + --- C
be + iku > be:ku abs + burn ‘burnt’
be+ ijiom > bejiom abs + eat ‘eats’
ʈho + ut cone > thot cone ‘I go’
ara + ɛm-ben > aramben cl 6 + refl-rest ‘to rest’
khu + im > khum drink + neg ‘don’t drink’
kheŋe + ara + uli > khɛŋerauli cat + cl 6 + tail ‘cat’s tail’

When the absolutive -bi combines with the object proclitic, it follows
the same rules as in §2.5.7 (a) and (b). Thus:

bi + it/et > bit

bi + ik/ek > bik
bi + ɛn/en > bin

Vowel lowering, vowel deletion, degemination and compensatory length-

ening rules are ordered.

2.5.8 Syllable Attraction

When a pronominal proclitic, which is a reduced form of a pronoun
without a vowel, is followed by a lexical item with a proclitic or a prefix
beginning with a vowel, then this vowel is attracted by the pronominal
proclitic (refer to §3.8.3 for details on clitic sequencing). For instance, the
first singular pronomial proclitic is ʈh= ‘I’ and the word for ‘go’ is ot=cone
(proclitic ot= and the verb root cone).10 When the two words combine in
a phrase, the verbal proclitic attaches itself to the previous clitic, in this
case a pronominal clitic. The word is rendered as ʈhot as a phonological
word. This can be schematised as below:

ʈh= ot=cone > ʈhot cone

10 For details on proclitics refer to Chapter 3 and 6.

62 chapter two

The syllable attraction rule allows any two proclitics, or a combination

of a proclitic and a bound morpheme to generate a prosodic word. Thus
the resulting construction appears to be a one-word construction where
the second bound morpheme of the following word moves to its left and
forms part of the first word shown by the spacing symbol #. Thus:

ʈhu → ʈh = ot=cone > ʈhot # cone ‘I go’

ɖu → ɖ i=tei > ɖi # tei ‘his blood’
ʈhu → ʈh=ɛm- > ʈhɛm # ‘I myself ’
ŋu → ŋ= ara=karap > ŋara # karap ‘your waist’
khider-bi ut=phoke > khider # bit # phoke ‘break the coconut’

Refer to §10.1.14 also to see how this works for all parts of speech. The
constituents of this phonological word may or may not have an indepen-
dent meaning, e.g. ʈhɛm which is derived from the first singular pronomi-
nal proclitic ʈh= and the reflexive prefix ɛm- means ‘myself.’ However, ʈhot
which is derived from the first singular pronominal proclitic ʈh = ‘I’ and
possessive proclitic ot= carries the meaning of only the first morpheme ʈhu
‘I’ or may represent possessive ‘mine’ if followed by a possessed noun as
in ʈhot bo ‘my back.’ The attachment of clitics to various nouns and verbs
is discussed in the next chapter.

2.5.9 Metathesis
Sporadic examples of metathesis of the semi-vowel y and the locative
marker -l have been noticed in the speech of Jero and Khora speakers.
Thus:

cay + l > cyal ‘where’

2.6 Acoustic Study of Problematic Sounds

As noted above, one of the speakers, Peje,11 always used a labialised lateral
instead of the simple alveolar lateral [l]. An acoustic study was carried out
to analyse the sound.
The labialised lateral could have been an idiosyncratic phenomenon in
Peje’s speech. To rule out language contact effects, other languages of the
Andaman were also investigated. However, no other language spoken in

11 When we first met Peje in 2001 he was in his mid-fifties.

 phonetics and phonology 63

the Andaman was noticed to contain this particular sound. Despite this,
I undertook the acoustic study of laterals so as to give the benefit of the
doubt for establishing an archaic trait of the language. Peje had reported
to us that he had spent some time of his childhood with some Burmese
tribes. We, however, could not verify this.
Please refer to Appendix B for the spectrograms and further analysis
of laterals.
 Chapter three

Grammar overview

3.1 The Structure

3.1.1 General
As discussed in the previous chapter the language demonstrates large
phonetic variation, and contains varying vocabulary items used to express
the same meaning; this is likely because of its mixed nature or its status
as an endangered and moribund language. This chapter is an overview
of grammar highlighting the basic tenets of the grammatical structure of
PGA. This is done so as to prepare the readers to comprehend what fol-
lows in the subsequent chapters.

3.1.2 Typological Background

PGA is a double-marking polysynthetic and agglutinative language with
an SOV pattern. Examples in the text are given in four-line format and
three-line format. The former format represents the word boundaries in
the first line and does not give any morphemic breaks while the second
line gives the morphemic breaks.

(1) ɑjoe ɑtoŋnu tɑrɑcɔre eoleinciko

ɑ=joe ɑ=toŋ-nu tɑrɑcɔr-e eole-inci-k-o
cl 1=Joe cl 1=Tong-pl spring-abs see-go-fa- dst.pst
‘Joe and Tong went to see the spring.’

PGA has a dual semantic system for body part categorisation: one that
is expressed in concrete nouns and another more abstract one that is
expressed in grammaticalised morphemes indicating body divisions. These
are represented by body division possessive classes that classify body part
terms based on the area of the body they refer to. These classes also clas-
sify other inalienables, with some semantic connection to the body part
system. Furthermore, body part semantics pervade the lexical and gram-
matical system of the language as this dual system is extended to other
form classes, viz. verbs, adjectives and adverbs. The body division class
markers occur as proclitics attached to all classes of content words. Thus,
66 chapter three

all content words in PGA can easily be divided into bound and free, the
former necessarily imbued with the semantics of ‘inherency’ and ‘depen-
dency.’ This is a unique and important feature of the language: all parts
of speech occur as dependent and non-dependent pairs, the former being
obligatorily marked. Refer to §3.6 below.
Most of the nouns that refer to the typically inalienable possessed
items as well as those which refer to the objects or results of an action,
e.g., ik-jira ‘it-story.tell’, i.e. ‘a story’, are marked for dependency. All body
part terms, kinship terms, part-to-whole, part-to-component, as well as
nouns referring to time, direction, and depth are dependent nouns. The
dependent nouns are further divided into seven classes, each defined by
a distinct body division inalienability marker appearing as a class marker
proclitic.
There are three numbers encoded in pronominal forms. However,
nouns in general are not marked for duality and plurality. Number is
marked for plurality on a few common nouns designating animacy fea-
tures, e.g. ‘dogs’, ‘children’, etc.
PGA is agglutinative in terms of its treatment of morpheme boundar-
ies, but polysynthetic in its nature of word morphology. A large number
of morphemes, affixes, phonological words, clitics, and incorporation can
constitute a single word. The incorporation of reflexives and nouns is seen
in verb complexes. These verb complexes may constitute a verb phrase.
Thus, verbs are much more versatile and elaborate than nouns. PGA is a
prototypical ‘double marking’ language where the head or possessed noun
is obligatorily marked in inalienable possession, but it is the possessor, the
dependent noun, which is marked in alienable possession. In addition,
case markings are suffixed to the nouns. The verb complex includes a large
amount of information in multi-morphemic strings that include object clit-
ics, incorporated nominals in causative constructions, reflexive and recip-
rocal prefixes, as well as suffixes expressing tense, aspect, and mood. Overt
external NPs are present in addition to the verb complex. However, these
are optional and often dropped in discourse. PGA is a verb-final language.
Following are few sentences from PGA to exemplify the verb complex and
the pervasive body division class markers (symbolised as cl). One can see
that body class markers are also attached to verbs.

(2) oʈhobɔyamo ʈhuitɛrtakom

o=ʈh=o=bɔi-ɑmo ʈhu i(t)=tɛrtɑ-k-om
3sg=1sg=cl 7=ask-cond 1sg obj=tell-fa-npst
‘If he asks me, I will tell him (the whole story).’
 grammar overview 67

(3) ʈhamaikaʈhit bɔlo

ʈh=ɑ=mɑi (i)k=ɑ=ʈhi-(i)t=bɔlo
1sg=cl 1.poss=father obj=cl 1=search-obj=went off
‘(They) went off to search for my father.’

(4) u khidere tautphayom

u khider-e tɑ-ut=phɑy-om
3sg coconut-abs tra-cl 4=dry-npst
‘He is drying a coconut.’ ( jointly with other people)

Some of the other typological features are: (1) While the genitive phrase
precedes the head noun (as is typical of verb-final languages), other modi-
fiers follow the modified elements. (2) Nouns in general are poor in affix-
ation compared to verbs. (3) Cross-referencing proclitics occur on verbs
as well as on temporal adverbs.
The most striking feature of the language is the heavy occurrence of
proclitics. Most of the free personal pronouns exist in reduced forms as
simple proclitics and when they occur with other clitics in the language,
such as body class proclitics, they offer the possibility of clitic sequenc-
ing. In a large number of words, the clitics have fused with the noun or
verb and are no longer transparent. This can be seen as a result of the
process of grammaticalisation (discussed in detail later in the chapter in
§3.7) or from concrete lexicon to abstract grammatical concepts that nor-
mally takes place in the course of language change but more significantly,
during language evolution (Givon 2002; Heine & Kuteva 2007).
Proclitics are symbolised by the = symbol in the examples given in the
text. We will consider these in detail later in the chapter.

3.1.3 Core Arguments

The language has two different sets of markers, one for prototypical agent
nominals with -e suffixed to agent nouns and another one with -bi attached
to the subject of unergative intransitive verbs, the subject of unaccusative
verbs and the object nominals. That is to say, PGA maintains the erga-
tive vs. absolutive distinction with the absolutive being overtly marked.
It was observed during my fieldwork that in discourse and in fast speech
speakers tend to drop these case markings.
Pronouns, unlike proper nouns, are not marked for S or A function.
The appropriate argument markers are underlined in the following
examples.
68 chapter three

(5) kɔʈphɛcbi kanticole belekom

kɔʈ-phɛc-bi kɑntico-l e=bele-k-om
clay-vessel-abs fill-conv cl 5=overflow-fa-npst
‘Having filled, the vessel is overflowing with water.’

(6) billubi bithom

bilu-bi bith-om
ship-abs sink-npst
‘The ship is sinking.’

(7) thire-bi bas khuttral beno-k-o

child-abs bus inside sleep-dst.pst
‘The child slept in the bus.’

(8) aʃyame bas kuttaral konabit beliŋo

ɑ=ʃyam=e bɑs kotr-ɑl konɑ-bi it=beliŋ-o
cl 1=Shyam-erg bus inside-loc tendu1-abs 3obj= cut-dst.pst
‘Shyam cut the tendu fruit in the bus.’

(9) thirebi ŋolom

thire-bi ŋol-om
child-abs cry-npst
‘The child cries.’

Plural subject nouns are not marked for their argument functions. As
stated earlier, pronouns are unmarked for S and A functions (11).

(10) thirenu ŋolom

thire-nu-ø ŋol-om
child-pl cry-npst
‘The children cry.’

(11) o aʈ era phuvom

o-ø ɑʈ- erɑ=phu-om
3sg wood cl 2=cut-npst
‘He is cutting the trees.’

PGA is thus an ergative-absolutive language with a fluid S function.

The nominal agent of transitive verbs is optionally marked by -e (exam-
ples 12, 13, 14). At times, and quite surprisingly, the agent of transitive

1 There is no equivalent in English for this local fruit known as ‘Tendu’ in the Anda-
manese Hindi.
 grammar overview 69

verbs, i.e., the A function is marked by S marking—the marker -bi, which

otherwise marks S, as in (18) below. The agent marking -e is also missing
with pronominal agents and with plural nominal agents in the examples
given above.
All human subjects and objects, but more specifically, names of humans,
are preceded by the proclitic a= which represents the body division class
marker 1 to indicate ‘origin.’ Consider:

(12) athirenu naramliʃo boʃobo

ɑ=thire-nu n-ɑrɑ.m-liʃo boʃo-b-o
cl 1=child-pl 3pl.cl 6=.refl-sister beat-fa-dst.pst
‘The child did not beat his sister.’

(13) asalma kambikhir konabi jiyo

ɑ=sɑlmɑ kɑmbikhir konɑ-bi iji-o
cl 1=Salma-ø morning tendu-abs eat-dst.pst
‘Salma ate tendu fruit in the morning.’

(14) athirenu rɛfibi ijul nu ino julu khuo

ɑ=thire-nu refi-bi iji-ul nu -ino-julu khu-o
cl 1=child-pl rice-abs eat-conv 3pl water-cold drink-dst.pst
‘Children ate rice and drank cold water.’/‘Having eaten rice, children drank
cold water.’

(15) aʈɔŋ emɔto ie bi

ɑ=ʈɔŋ e=mɔto ie bi
cl 1=Tong-ø cl 5.poss=foot pain cop
‘Tong has pain in her foot.’/‘Tong’s foot hurts.’2

The absolutive -bi has the additional function of indicating ‘specificness.’

In the following sentences -bi in (17) is optional because it is not a spe-
cific tree. Once the noun is specified, the proclitic to the verb becomes
optional as in (16). However, many speakers use both, viz. -bi attached to
the noun and the verbal proclitic attached to the verb.

(16) ʈhu ʈɔŋbi rapho

ʈhu ʈɔŋ-bi (e=)rɑpho-ø
1sg tree-abs cut-pst
‘I cut the tree (the specific one).’

2 Possessive constructions are discussed in detail in Chapter 6.

70 chapter three

(17) o ʈɔŋe rapho

o ʈɔŋ (-bi) e=rapho-ø
3sg tree cl 5=cut-pst
‘He cut the tree.’

The following sentence gives both functions of homophonous -bi, viz. the O and
A functions. It is not clear why the subject of the transitive is marked by -bi case
marking. Is it an instance of case alternation?

(18) ɖowlu-bi ʈhɛ-bi ok-o

ant-abs 1sg-abs bite-dst.pst
‘The ant bit me.’

3.2 Ambivalence of Verbs, Adjectives and Nouns

The canonical semantic functions of word classes are observed in PGA.

These are:
Nouns: denote referential entities (‘things’).
Verbs: denote eventualities (actions, processes, and states).
Adjectives: denote properties and attributes.
These categories also operate in their pivotal places at the syntactic level.
That is to say, adjectives always follow the modified, verbs occupy the sen-
tence-final position and are marked for all possible TAM inflections, and
nouns take relevant nominal inflections such as case, gender and number.
Adjectives appear in bare forms as a modifier of the head. However, a
large number of words can function as both nominal categories, defining
objects and persons, and as verbal categories, defining states and events,
each taking appropriate affixes. Adjectives can also be derived as nouns
or verbs depending upon the requisite terminations. In short, a noun, an
adjective, a stative or an active verb, or a time word and a temporal adverb
can occupy the predicate slot. This is quite uncommon in languages.
This is not equal to categorical-indeterminacy as described by
Koptjevskaja-Tamm (1993) or the case of ‘categorical dual-citizenship’ as
described by Spencer (2011). Let us examine a few constructions to show
the ambivalence of the word class in PGA.

3.2.1 Noun Verb Ambivalence

ot/ut=ɲyo Noun: ‘house’
Verb: ‘to live’
 grammar overview 71

(19) eremla ʈhi (u)t=ɲyo-ø

alone earth cl 4=live-pst
‘He lived alone.’

(20) a=cao-no ʈhi-bi (u)t=ɲyo-am

cl 1=dog-pl place (live)-abs cl 4=home-npst
‘Dogs live with us.’
(Dogs are domestic animals and are treated as members of the family, hence
the a= proclitic is used.)

(21) ʈɔilccɔŋ tot=ɲyo ʈhimikhu bi

deer cl 4.poss=house forest cop
‘The forest is the deer’s house.’

er/ɛr=nol Noun: ‘pen, writing instrument’

Verb: ‘to write’

(22) thu er=ŋole-ta er=ŋole-me

1sg cl 2=write-instr cl 2= write-hab
‘I write with a pen.’

i/e=bɔr Noun: ‘wind, breeze’

Verb: ‘to flow’

(23) e=bɔr-ɔm
cl 5=wind-npst
‘Wind is blowing.’

(24) bɔr-be tambikhir e=bɔr-ɔm

wind-abs tomorrow cl 5=wind-npst
‘The wind will flow tomorrow = it will be windy tomorrow.’

e=cɛr Noun: ‘rain’

Verb: ‘to rain’

(25) e=cɛr trɔphe

cl 5=rain much
‘Too much rain.’

(26) e=cɛr-om
cl 5=rain-npst
‘It is raining/will/may rain.’

bino Noun: ‘sleep, dream’

Verb: ‘to sleep’
72 chapter three

(27) ʈhu bino-bi kalemu-o

1sg sleep-abs early-pst
‘I had some sleep very early.’

(28) thire-nu er=beno be

child-pl cl 2=sleep cop
‘Children are sleeping.’

e=boe Noun: ‘spouse’

Verb: ‘to marry’

(29) ʈh=em boe-b-om

1sg-refl marry-fa-npst
‘I am getting married.’

(30) ʈhe=boe nɔl pho

1sg=husband nice neg
‘My husband is not nice.’

ɛr=coic Noun: ‘embrace’

Verb: ‘to embrace each other’

(31) a=caci ɛk=tɛr= coic-b-om

cl 1= aunt obj=cl 2= embrace-fa-npst
‘The aunt embraces him/her.’

(32) ŋ=ɛr coic ʈha tabiŋom

ŋ=ɛr coic ʈha= ta-biŋ-om
2sg=cl 1 embrace 1sg.cl 1 tr-remember-npst
‘I remember your embrace (the way you used to hold me in your arms).’

ɖiʈh Noun: ‘hole’

Verb: ‘to make a hole’

(33) rowa-bi ram-ɖiʈh-o

boat-abs in the centre-hole-dst.pst
‘A hole developed in the bottom of the boat.’

(34) julu tɛr=ɖiʈh-(bi) khuro be

dress cl 2-hole-abs big cop
‘There is a big hole in my dress.’

e=jira Noun: ‘speech, saying’

Verb: ‘to speak, to tell’

(35) kholek ʈhe-jira-k-om

laughing 1sg.cl 5=say-fa-npst
‘He speaks to me laughingly.’
 grammar overview 73

(36) ʈhi ik-jira-(i)t= nɔl-e

1sg obj=story-obj= good-be
‘Tell me a good story.’

ʈɔl Noun: ‘flower’

Verb: ‘to blossom’

(37) ɖu ɛ=ʈɔl-ot= co
3sg-dem cl 5=flower-cl 4=seed
‘That is the pollen of the flower.’

(38) kona-bi ʈhimukhi-e ʈɔl-om

tendu fruit-abs jungle-loc flower-npst
‘Tendu fruits have blossomed in the forest.’

bolo Verb: ‘excrete’

Noun: ‘excreta’

(39a) bolo be
defecate imp.
‘go defecate’

(39b) ŋut bolo

2sg.cl 4.poss excreta
‘your stool.’

The dichotomy between noun and verb does not appear to be rigid
because the same lexeme can be used as a verb or as a noun depending
upon the process involved in forming a lexical unit. Verbality is assigned
to prototypical noun categories by attaching tense markings, and con-
versely, nominality can be assigned to verbs by attaching case suffixes and
by other means such as showing it as an object of possession. The same is
true for adjectives. That is, any content word in PGA is capable of moving
from its pivotal slot and can occupy the predicate position with appropri-
ate tense, mood and aspect inflections. Thus each of the following word
classes can be used as a predicate.

(a) phoŋ ‘hole’ n

(b) khuro ‘big’ adj
(c) ut=ɲo ‘live’ v
(d) lɔto ‘much’ adv

Let us consider adjectives now.

74 chapter three

3.2.2 Adjectives as Verbs

Adjectives can take the predicate position with or without any TAM mark-
ings, or occur with copula.

cɔfe Adjective: ‘much’, ‘a lot’

Verb: ‘exist a lot’

(40) mino cɔfe e=tɛʃ-e

potatoes much cl 5=give-imp
‘Give lots of potatoes.’

(41) nɛ=cɔfɛ=m fo non-to-foc

2sg.cl 5=much=refl neg only two
‘Yours are not many but only two.’ (In the context of children)

(42) ʈhut= thire cɔfe be

1sg.cl 4.poss= children many cop
‘I have many children.’

lɔto Adjective: ‘too much, a lot’

(43) cai ono lɔto-k-o

Q oil much-fa-dst.pst
‘Why have you put [in] so much oil (while cooking)?’

3.2.3 Adjectives as Αdverbs

Adjectives can be used in the adverbial category by placing them next to
verbs, quite common in many languages.

nɔl Adjective: ‘good’

Adverb: ‘well, nicely’

(44) buamo ra-ʈulu nɔl

musical instrument cl 6=play good
‘He plays the Buamo well.’

As seen above in (43) adjectives can take TAM inflections.

(45) ʈhi tuŋ-ul ʈhaono nɔl-o

earth refl-on sit like/good-dst.pst
‘I liked to sit on the ground’
 grammar overview 75

Consider the following constructions indicating that the adjective can be

used predicatively. Verbal class markers have the tendency of being long
distanced from their host (discussed later in §3.8.1). Hence verbal class 5
is distanced from the verb and appears before the object noun phoŋ. For
details refer to §3.8. In (47) a long distanced class marker is shown by an
arrow. It is attached to adjective which is used as a verb here. Obligatorily
marked adjectives are considered in detail in §8.2.

(46) ʈhu ephoŋom

ʈhu e=phoŋ-om
1sg cl 5=dig-npst
‘I am digging.’

(47) ʈhu e=phoŋe khurokɔm

ʈhu e=phoŋ-e khuro-k-ɔm

1sg cl 5=hole-abs big-fa-npst
‘I am digging a big hole.’

3.2.4 Other Unusual Features

There is another significant factor which cannot be overlooked: there are
grammatical categories which are pervasive in nature and are thus obliga-
torily assigned to nouns, adjectives and verbs. The body division class
marker clitics that we will discuss later in the chapter operate as proclit-
ics that attach to nouns, modifiers, and verbs. Thus, the proclitic ot= is
attached to the verb ‘go’ in [ot=cone]V, and is also found in ‘hair’ [ot=bec]N,
and ‘wide’ [ot=belo]adj. Although transparent in some respects, these have
been highly grammaticalised in the language. For further details see §3.3.
Furthermore, temporal deictic markers are also attached with pronom-
inal proclitics to their left which suggests an anthromorphic explanation
of deictic relations with the self. Thus, the word bikhir ‘morning’ is never
used in its bare form, but is necessarily preceded by a pronominal as in
ʈh.am-bikhir ‘1sg.refl-morning’, or ak.am-bikhir ‘3sg.refl-morning’, or
ŋ.am-bikhir ‘2sg.refl-morning’, as these phases of the day are referenced
to the speaker or hearer. Hence, the following sentence is spoken in the
context of the addressee’s reference.

(48) ʈhu ŋ.am-bikhir ŋ=ot=ɲo inci-k-om

1sg 2sg.refl-morning 2sg=cl 4.poss=house reach-fa-npst
‘I will reach your home in the morning.’
76 chapter three

Words for the phases of the moon and the sun can occupy the predicate
position by adding TAM categories. The word for ‘sunset’ is ɖiu-birate
which takes class 4 as well as verbal categories such as tense marking.

(49) ʈh=ut=ɖiu-birate-k-ɔm
1sg=cl 4=sun-set-fa-npst
‘It will take me the whole day (to finish the job).’

Hence, the absolute separation between the nominal and verbal domain
has weakened in the language.3 Practically any root word in the language
has the potential to be predicated.

In the beginning of the chapter in §3.1.2, we said that there was a dual
semantic system in the language. This system cannot be understood with-
out the discussion on the concept of inalienability and inherency as per-
ceived by the Great Andamanese people. Let us begin with ‘inalienability’
and ‘grammaticalisation.’

3.3 Inalienability (ina), Body Division Classes and

Grammaticalisation

3.3.1 Introduction
The concept of inalienability (henceforth ina) is very old in grammar
(Hyman et al. 1970; Allan 1975–1976; Voeltz 1976; Hawkins 1981; Hinne­
busch & Kirsner 1981; Diem 1986; Chappell & McGregor 1996, to name
a few) and has been much debated by linguists, anthropologists, phi-
losophers and ethnographers for many years. It is still a major field for
research and debate. One of the reasons for such a wide and sustained
interest in the topic has been that there are some yet unresolved issues
regarding the definition and identification of ina elements that become
encoded in language. Then, there is also the problem of finding the appro-
priate placements for ina elements in grammar.
There is a fairly general agreement among linguists that there are five
areas of inalienability:

3 Muysken (2008) observes similar phenomena for other languages.

 grammar overview 77

(i) Body parts

(ii) Kinship relations
(iii) Spatial relation terms
(iv) Objects closely associated to a person for his livelihood, e.g. boat,
home, etc.
(v) Part-whole relationships such as ‘branch of a tree’

3.3.2 Anthropocentrism
PGA exhibits a hierarchy within most of the ina domains such as ‘body
divisions.’ There are seven distinct divisions/areas that are recognised
within the human body and each one is assigned a hierarchically des-
ignated place on a scale. The specific area that a particular body part
belongs to is designated by a specific inalienability [ina] or inherency
feature represented by body division classes. These are grammaticalised
morphemes that attach to several nouns pertaining to body part or body
division terms. These serve the grammatical function of a prototypical
possessive association between two nouns that stand in a relationship of
possessor and the possessed, or R and D. Since these class markers are
obligatorily attached to the head noun, D, the nouns under category D are
dependent nouns. The class markers by their distinct nature behave like
proclitics (see below). As mentioned earlier, PGA maintains seven body
divisions. It then further extends the notion of these seven divisions to
other terms, i.e. kin terms, spatial relational terms, closely related object
terms, and furthermore, attributes (human propensity), objects, action
and state (denoted by verbs) with reference to the body. This implies that
the Great Andamanese people visualise their world from a point of view
centered in the ‘self’ or ‘ego.’ The perspective that human beings are the
centre governs the structure of the modules of grammar. However, each
division within the realm of inalienability is designated on a hierarchical
scale. We will see in Chapter 6 that inalienable areas other than those rep-
resented by nominals such as (i)–(v) mentioned above, are also preceded
by these body division class markers.
As mentioned above, verbs, adjectives and adverbs are capable of being
attached to the same seven division class markers. In other words, these
class markers play a significant role in word formation processes as
grammaticalised units and can attach to a large number of vocabulary
items signifying diverse meanings. As there is a distinct marker used for
all typical ‘alienable’ objects -ico or -iʃo, the language formally maintains
the distinction between the two kinds of possession, i.e., alienable and
78 chapter three

inalienable. We shall consider these in detail in Chapter 6. Let us briefly

discuss what leads to the appearance of these class markers and why they
are obligatorily attached to the following form classes. We shall begin with
the concept of inalienability and inherency, the two governing forces that
dictate the occurrence of body division class markers.

3.3.3 The Semantics of Inalienability

An obvious question is how one can recognise inalienability. Does physi-
cal separation alone license the notion of alienability? Does the concept
of inalienability belong to more areas than those covered by the five given
in §3.3.1? That is to say, all things being equal, can one justifiably include,
in the inalienability rubric, a list consisting of physical ailments, internal
and external feelings, emotions and their after effects? Can the notion
of inalienability be redefined so that it takes cognisance of the unique
system present in PGA so that it allows a proper description of the same?
Considering the various views held by linguists in analysing languages of
diverse natures, one is motivated to extend the definition of inalienability
or ina to include the following parameters and relations:

a) Inseparability of r and d
b) Inherent relationship between r and d
c) Permanency of the relation between r and d
d) Intimacy of the relations between r and d (language- and culture-
specific)
e) Conceptual dependency
f) Inextricably linked entities

As Langacker (1991, 1995) observed, one can say that inherently relational
nouns are characterised by a high degree of conceptual dependency. They
are conceptually dependent in the sense that they must be understood in
relation to something given (Velazquez-Castillo 1996: 34). It is difficult to
define and identify what is inherently related and what is not; just as it
is challenging to establish a correlation between inherent relations and
conceptual dependency. The biggest challenge that PGA poses is this:
how should one distinguish one kind of conceptual dependency from the
other? How is a body part term such as ‘head’ conceptually more depen-
dent or less dependent than terms like ‘stomach’ or ‘tongue’? Or, how
can one correlate this dependency with the kinship terminology so as to
establish a one-to-one relationship between two inalienable nouns, the
body part and the kinship term such as ‘head’ and ‘child’ respectively?
 grammar overview 79

As one proceeds it will become clear that PGA speakers conceptualise

these six kinds of relationships (listed in (a)–(f ) above) between any two
nominal categories in the relationship of R and D, or between the adjunct
and the verb, as well as between a particular attribute and the host of the
attribute. They then symbolise them in the form of distinct class markers
that are obligatorily attached to the D.
This implies that the bonding between R and D does not restrict to
possessions but stretches itself to other relationships symbolised by other
grammatical categories, viz. verbs and modifiers. The kind of relationship
that holds between the two, i.e. class marker and its dependent morpheme
is determined by culture-specific cognitive elements.
Since body division class markers occur with nouns, modifiers, action
and state verbs, and express the relationship between an action and its
object, or between an action and its result, or between an action and
its manner, or an object and its state, this relationship between the two
grammatical categories symbolised by class markers cannot be termed
as a possessor and the possessed in a prototypical fashion. It represents
the concept of ‘inherency’ that is perceived by the speakers of the lan-
guage. The dependency feature of the verbal root, modifier or noun on
the preceding body division class marker may be understood as the
‘inherency factor.’ In other words, the relationship between two nomi-
nal categories or between an action and its mode of operation or resul-
tant state or between the object and its attribute, is seen as inherent and
inextricable. This factor is more obvious in the case of nouns designating
separated body parts (example 51) and part-to-whole relationships (refer
to Chapter 6 for details). The notion of inherency further represents a
conceptual dependency between the object and its possessor. The Great
Andamanese conceptualise their world through these interdependencies
and hence the grammar of the language encodes this important phenom-
enon in every class form expressing referential, attributive and predicative
meaning.
It is to be noted that a varying degree of inalienability, primarily based
on frequency of occurrence, is present in the conceptual world of the
Great Andamanese people. That is to say, not all ina occur with the same
degree of frequency: some are more frequent than others. Refer to Table 6.9
given in Chapter 6.

3.3.4 Linguistic Manifestations of ‘Inalienability’

It was stated above that the Great Andamanese people perceive their
world through the divisions of their body. Each division is within the
80 chapter three

Table 3.1. Seven basic zones in the partonomy of the body

Classes Partonomy of the human body body class
markers
1 mouth and its semantic extensions a=
2 major external body parts ɛr=
3 extreme ends of the body, e.g. toes and fingernails oŋ=
4 bodily products and part-whole relationships ut=
5 organs inside the body e=
6 parts designating round shape/sexual organs ara=
7 parts for legs and related terms o= ~ ɔ=

realm of inalienability but is on a scale. The same division is carried

over in their perception of human relations and other objects which are
conceptually dependent. The binary distinction of alienable versus inalien-
able does not exist in a strict sense. The prototypical alienable nouns
which exhibit external possession are few, while all others fall within the
realm of inalienability. The Great Andamanese perceive the following
divisions of the body and symbolise each by a monosyllabic or disyllabic
morpheme which is attached to the appropriate term for the body part,
the head noun, D. The concept of inalienability is cardinally represented
in possessive constructions depicting various body parts. Table 3.1 sum-
marises these markers and the semantic domains they cover. Readers are
advised to refer to Chapter 6 for a detailed discussion on possession.
A word of caution must be given at this point: these divisions are not
very straightforward as there are many overlaps and unclear elements
that we will take in detail in Chapter 6 where the complete list of body
division terms is given. A list of kinship terms is also given in the same
chapter. An exemplary sentence will elucidate the discussion.

(50) ʈh=ot=bo
1sg=cl 4.poss=back
‘My back.’

The fundamental division of animacy plays an important role in deciding

the phonetic shape of the base form of the class marker. Thus a cut up
body part, say of an animal, belongs to the same class, but is preceded by
a dental t-. Consider:

(51) ra tot= bo
pig cl 4.poss=back
‘Pig’s back.’ (For consumption)
 grammar overview 81

These are discussed in detail in Chapter 6.

3.3.5 Body Class Μarkers and Οther Νouns

Various body division class markers can attach themselves to the same
nominal reference modifying it and further indicating various locations
of the object noun on the body. Thus:

(a) e=tei (class 5=blood) ‘blood inside the body’

(b) ot=tei (class 4=blood) ‘blood outside the body’ [when bleeding]
(c) oŋ=tei (class 3=blood) ‘blood on finger or from finger’
(d) ɛr=tei (class 2=blood) ‘blood from the head’

For details on noun marking see Chapter 5.

The class markers occur with terms for spatial distances and directions,
e.g., ʃiro ‘sea’ and ʃiro tɛr=likhui/likhu4 ‘sea class 2=lap’, or ‘deep sea’; ʃ iro
tara=cɛrɛl ‘sea class 6=green/blue’ or ‘open sea.’ Consider the range of
spatial characteristics of ‘sea’ given in Table 8.6 in Chapter 8.
The notions of ‘inherency’ or ‘conceptual dependency’ become clear
when one considers other constructions that involve spatial relations or
the size and shape of an object. As mentioned earlier, in PGA, the use of
class markers is not restricted to nominal categories, but also pervades
other levels of linguistic organisation, viz., adverbs, adjectives and verbs.
It is imperative to look into the morpho-syntactic manifestations of the
distinction and their correlations with lexico-semantic categories that are
considered to be prototypically inalienable.
We shall discuss nouns pertaining to spatial, depth and directional
words in Chapter 8 on deixis.

3.4 The Semantic Role of Body Division Classes

The body division class markers help us to classify nouns into eight divi-
sions: seven inherently related or representing relational nouns, and one
non-inherently related or representing independent nouns. The objects
that can be transferred, or which are expected to be transferred, such as
goods in the market, household goods not of an intimate nature, objects

4 The class marker is preceded by a dental sound t- in this and the following example
because of the inanimate nature of the modified nominal.
82 chapter three

belonging flora and fauna (see Pande & Abbi 2011) and other objects of the
natural environment do not obligatorily take class markers. These may be
considered as independent nouns. These are discussed in detail in Chap-
ter 6 on possession.
As seen in (51), the use of t- prefixed to the body division class
marker makes it clear that inalienability is not perceived as a physi-
cal non-separation of an object from its host but as an inherent part of
the host. The conceptual dependency between the two is absolute and
is thus recognised as being intimately bonded to each other. The lan-
guage has developed formal devices in the form of proclitics (see below)
that reflect the dependencies, both conceptual and inherent, between
the related objects and events. Furthermore, the variety of formal cat-
egories present in the inalienable category implies varying degrees of
inalienability present in the conceptual world of the Great Andamanese.
Between the varying degrees of inalienability lie various kinds of inher-
ent relations. The inherent relation is signified by the use of proclitics
not only with nouns but also with other content words from different
grammatical categories. This phenomenon can be captured better by
Figure 3.1 given below indicating varying degrees of grammaticalisation of
content words.

3.5 Inalienability and its Representation in Verbs

The analogy of semantics carried by class markers is further extended to

verbs. A large number of verbs are individuated by class marker proclit-
ics, in which the body part semantics shift into event type semantic cat-
egories of various kinds. For instance the word for ‘leave’ is ut=cone and
not simply cone. The class marker proclitics combine with verbal roots of
any valence. Similar phenomena have been observed in Papantla Totonac
(Levy 1992: 533), a language of northern Veracruz, Mexico and in Matses, a
Panoan language spoken in Amazonian Peru and Brazil (Fleck 2006), with
the difference that the prefixation in Panoan is of the body part term itself
and not of the grammaticalised morphemes pertaining to the division of
the body. Nor are the semantics in that language as intricate as in PGA.
Some languages of the world incorporate body part terms into verbs, as
in the languages of the Americas and Australia (Evans 1996: 66), however,
the structures of PGA are different from the phenomenon of incorpora-
tion mentioned by Evans. In PGA the class markers classify the verbs in
the same way as they are used for classifying nouns as described above. In
 grammar overview 83

addition, they can be seen as modifiers of the action because they encode
the manner of action.
Some transitive and intransitive verbs are attached to class markers
which account for a world of action and state that can be considered on
a par with the distinctions made on the scale of inalienability with regard
to the body part terms. Thus:

(a) ut=ʃile
class 4=aim class 5=aim
(b) e=ʃile
‘aim from above’ ‘aim to pierce’

Body part semantics also pervades intransitive verbs designating vari-

ous psychological predicates, experiences, and states. Hence, verbs with
the proclitic ot= ~ ut= would refer to an action of motion away from the
speaker, such as ‘go’ or ‘exit’; thus ʈh=ut=cone-bom (1sg=class 4=go-npst)
‘I am going’, or, where something (not necessarily tangible) is being gen-
erated as in experiential verbs ‘feeling sad/happy/hungry/thirsty/’, etc., as
in thire ut=ʈheʈe-bom (child class 4=hunger-npst) ‘the child is hungry.’
These objects of experience, namely ‘hunger’, ‘thirst’, etc., are an inherent
part of the experience (hence they are inalienable) and emerge involun-
tarily in a person. They are seen as products of the body or ‘self.’ Similarly,
verbs like ‘shake’ and ‘kiss’ have class marker 2 er= while verbs like ‘pound’
or ‘beat to a pulp’ uses e=, the class marker 5 which allows us to extend
the analogy given in the semantics of the partonomy of the body. These
are considered in detail in Chapter 9.
Semantically, the notion of dependency can be seen in two different
ways: (1) when class markers appear with verbs which are prototypically
transitive in nature, they signify the mode of operation and the effect of
an action, something which cannot be alienated from the action itself;
(2) when class markers appear with prototypical intransitive verbs of
motion, they signify ambience (Chafe 1970), the resultant experience, or
the patient noun which cannot be alienated from the verb. For example,
‘dance’ cannot be segregated from the ‘action of dancing’, the result of ‘roll
down’ cannot be separated from the ‘action of rolling down’, the expe-
rience of ‘hunger’ cannot be separated from the state of ‘being hungry.’
Although it is very difficult to distinguish one kind of meaning from the
other while analysing each of the seven divisions represented in the verbs,
partly because of the class markers having grammaticalised in varying
degrees over a period of time of language development, one can arrive at
a broad classification.
84 chapter three

3.6 Inalienability and its Representation in Modifiers

The case for adjectives is similar to the case for the verbs. A kind of
semantic transparency, drawn on the basis of the original classificatory
meaning assigned to the body part terms by the class marker, can be seen
between these and the host adjectives to a large extent. This, at times,
leads to the possibility of determining the choice of adjective with a par-
ticular class marker. For instance, it was observed that the class marker
i- ~ e- is attached to those terms for body parts which are inside the body,
e.g., ‘blood’, ‘intestines’, etc., and that the same class marker is attached to
adjectives defining internal human traits such as e=liu-ʃɔŋɔ ‘brave’, e=cay
‘bad’, e=ɖirim ‘black’ or ‘dark’, and ɛ=bopho ‘stupid.’ It may also signify
internal quality of an inanimate object such as in e=kokʰela ‘blunt’; i=boe
‘boiled’; e=mɔʈello ‘thick’ and i=phuŋ ‘fully ripe.’
One can explain that inherently relational elements are conceptually
dependent on each other and thus also define attributes. The inherent
attribute may include inherent personal attributes such as qualities,
nature, weight, height, size, shape, state of health, temperature, blood
pressure, energy, bodily functions (for example, movement, reaction, uri-
nation), consciousness, courage, fear, name, among others. Refer to Chap-
ter 8 on modification and deixis and Figure 3.2 at the end of this section.
Body parts and attributes of an object are both considered to be inalien-
able possession. Each class marker proclitic from the set of seven decides
the nature of the modifier. Please refer to Chapters 6 and 8 for the full
spectrum of the phenomenon.
The following sentences indicate attribution or inherent quality of the
argument, but are marked by different body division class markers signify-
ing different kinds of attributes.

(52) a=kɔbo ɛr=tɔlɔbɔŋ (be)

cl 1=Kobo cl 2=tall (cop)
‘Kobo is tall.’

(53) ʈele i=pheca khamo bi

elephant cl 5=old condition cop
‘The elephant is old.’

(54) a=loka er=biŋoi be ara=kata

cl 1=Loka cl 2=fat cop cl 6=dwarf/short
‘Loka is fat (and/but) short.’
 grammar overview 85

Modifiers of verbs, viz. adverbs, can be preceded by these class markers

designating various deictic meanings as well as designating the manner
of an action. In this function, these class markers are highly grammatica-
lised. For example, the ‘anteriority of an action’ is a=kaulu, ‘backwards’ is
ot=lo, ‘behind something’ is ara=balo, and the manner adverb ‘slowly’ is
a=lea. Refer to Chapter 8 for further details.

3.7 Process of Grammaticalisation

Since body division class markers are attached to all form classes, each
with a specific meaning, they appear to be grammaticalised to a large
extent. It is not easy to establish a one-to-one correspondence between
these class markers and the class markers used for body division/body
part terms. However, native speakers of the language have no problem
in assigning an appropriate class marker in the case of new adjectives,
new verbs and new nouns. Although it is very difficult to distinguish one
specific kind of meaning from another while analysing each of the seven
divisions represented in verbs, adjectives, adverbs and nouns, one can
represent the process of grammaticalisation for each category as follows.
The length of each arrow signifies the degree of grammaticalisation. The
longer the arrow, the higher the grammaticalisation. Refer to Figure 3.1.

V: origin, human names

Class 1
ADJ: mouth-related attributes
‘mouth’
a=, ta=
ADV: anterior action/event

V: activity related to the front part of the body

Class 2
‘external ADJ: external attributes
body part’
ɛr=, tɛr=
ADV: deixis of adjacency

V: hand-related activities
Class 3
‘hand’ ADJ: attributes related to limbs
body part’
oŋ=, toŋ=
ADV: ‘haste’, ‘hurriedly done action’
 86 chapter three

V: directional verbs of emission, experiential verbs

Class 4
‘body’ ADJ: attributes signifying truncation/diminished entity
products,
genesis
ut=, tut= ADV: deixis of direction away from the self

Class 5 V: ingestive verbs, action involving interior of an object

‘internal
organs of ADJ: inherent attributes
the body’
e=, te= ADV: deixis of internal space

Class 6 V: action involving side of the body

‘sexual
organs/ ADJ: belly-related attribute
middle part
of the body’ ADV: deixis of immediate vertical and horizontal
ar=, tara= space, ‘over’, ‘under’

V: resultative action
Class 7
‘legs’ ADJ: ‘shape’, ‘texture’
o=, to=
ADV: temporal deixis

Figure 3.1. Degree of grammaticalisation across grammatical categories

It should be clear from Figure 3.1 that the degree of grammaticalisation

varies from one form of a class marker to another and from one part of
speech to another. The inalienability marker symbolised by class marker 7
has attained an equal degree of grammaticalisation as for verbs, adjectives
and adverbs. In general, class markers used for adverbs are more gram-
maticalised than those used for nouns and adjectives. This is because the
meaning of each of them is not very distant from the original meaning
which specifies body division or a body part term.
The process of attaching the class markers to head nouns appears
to be an iconic way of representing the ‘inherency’ factor. It cannot be
denied that the perception of what is inherent, non-transferable, and thus
inalienable is governed by the specifics of the Andamanese culture and
society. A prototypical inalienable noun, verb or modifier in PGA is one in
which R and D are conceptually close, which are not distanced even after
 grammar overview 87

physical separation and which cannot be transferred from one person to

another. The relationship of the ‘self’ or ‘person’ to the parts of the body
is characterised by conceptualising them inherently but distinctly. This
very relationship is then analogically transferred to other nouns, events
and modifiers. A question could be raised here as to whether terms such
as ‘possessor’ and ‘possessed’ that are categorically used for nominal clas-
sification can be stretched to include the dependency relation exhibited
by verbs and modifiers.

3.8 The Status of Inalienability Markers in the Grammar:

Proclitics

We have been using the terms ‘class markers’ and ‘markers’ throughout
the above discussion, without giving a particular name to this form or
structure from the available range of grammatical categories, because we
first wanted to discuss the various ways in which these markers or class
marker forms appear in the language and the varied functions they per-
form. It was observed that the class markers not only define and clas-
sify nouns but also modify other word classes that they are associated
with. As observed above, they pervade adjectives and verbs defining and
distinguishing the semantic types of the verbs and adjectives concerned.
These are grammaticalised forms of the inalienability markers and oper-
ate more like clitics than prefixes. In the absence of any written record
one can speculate that these must be diachronically derived from full lexi-
cal items. Over time, each has developed its own semantics which gov-
erns the larger meaning of the dependent form class, D. This has become
evident as each part of speech form has the potential to take different
class markers designating distinct meanings. Conversely, the same class
marker can be attached to various forms belonging to the same part of
speech designating similar meanings. Morpho-syntactically, they seem to
be closer to the category of ‘clitics.’ We shall now discuss in detail why
these class marker forms operate as proclitics in the language and how
they are different from simple affixes.

3.8.1 Arguments for Treating Inalienability Markers as Clitics

There are strong arguments to justify the labelling of body class mark-
ers as clitics: (i) they are not as close to their host as affixes, since class
markers that attach to verbs can be distanced from the verbs by several
other functional categories (sentences 56a, 57, 59 and 61). (ii) They can
88 chapter three

be characterised as an underlying determiner of the semantics of the

grammatical category that they are attached to. Attributive adjectives of
human traits or a verb of transitive or intransitive nature (examples given
earlier in §3.6 and 3.5 respectively) only have their complete meaning
when class markers are attached to them. This concordance suggests that
grammatical categories such as nouns or verbs are typically marked by
different kinds of class markers, which designate in their role of modifiers
the perceptual analogy that a native speaker makes between a specific
class marker and the nature of the action or the nature of the modifier of
a noun. For instance, every class marker that is attached to a verb decides
the nature of the action and the object associated with it. Some examples
are cited below to illustrate this point. The various shades of meaning to
the verb ‘cut’ are assigned by the occurrences of the distinct class markers
attaching to the left of the verb root.

(B) Verbs of ‘cut’

ara=pho [class 6] ‘cut it down’, ‘fell’ (tree)
ɛr=pho [class 2] ‘hit with a stick’ (from the front)
ut=pho [class 4] ‘cut/hit from above’ (coconut)

These examples illustrate that the same verb root may take a variety of
class markers depending upon the nature of the action and its effects on
the object concerned. The inherent semantics of each of the class markers
determines the specific meaning of the host.
Seen from a structural point of view, the fact that there is no obligatory
rule that they immediately attach to a verb in all constructions and the
fact that they can be moved to a position before a verb phrase as shown
in (55a, 56, 58, 59, and 60) below, makes them good candidates for pro-
clitics. This argument has been widely accepted as the indicator of clitics
(Aikhenvald 2002; Heggie & Ordóñez 2005). They are moveable and are
attached to the subject as in the following case, distancing themselves
from the verbs they are classifying. This allows incorporation of the nomi-
nal object. Affixes certainly do not behave in this manner. In the following
examples the (a) form shows a distanced verbal clitic (the class marker)
and the (b) form represents the immediate attachment of the clitic to its
host.

(55a) ʈh=ut= tɔŋ-e phoke

1sg=cl 4= tree-abs cut
‘I cut the tree.’
 grammar overview 89

(55b) ʈhu tɔŋ-e ut=phoke

1sg tree-abs cl 4=cut
‘I cut the tree.’

(56a) ŋo i=cophe cɔŋ-o

2sg cl 5=enough/how much get-dst.pst
‘How much did you get?’ or ‘Did you get enough?’

(56b) ŋo cophe i=cɔŋ-o

2sg enough/how much class 5=get-dst.pst
‘How much did you get (or) did you get enough?’

The semantic difference between (a) and (b) is that of focus, viz. the focus
is on the verb when the proclitic is adjacent to it.
It was discussed in §3.5 that these class markers may also attach to
intransitive verbs. These can also be distanced from their heads, i.e. the
verbs, as the subject noun attracts the class marker or object clitic towards
itself as in sentence (57). Intransitive verbs that are experiential in nature
or those of motion such as, ‘exit’ or ‘leave’ also allow the markers to be
attached to the subject noun (58, 59).

(57) rea a=jo et= nɛ rence-o

Reya cl 1= Joe obj= 3pl fight- dst.pst
‘Reya and Joe fought with each other.’

(58) ʈh=ot= ʈheʈhe-b-ɔm

1sg=cl 4= hunger-fa-npst
‘I am hungry.’

(59) buli-ot= diglipur-ak cone-b-ɔ

Buli-cl 4= Diglipur-dir go away-fa-dst.pst
‘Buli went away to Diglipur.’

(iii) The third argument for naming these class markers as clitics and not
affixes comes from a prosodic phenomenon of the language. Pronomi-
nal clitics combined with the class markers are rendered as one or two
syllables—simple phonological words in fast speech. Hence, (a) and (c)
are common, but (b) and (d) are not (Note: a word boundary is shown by
the symbol #):
90 chapter three

(a) ʈhut # toa-tut=kaʈa # ‘My elder sister.’

(b) ʈhu # ut=toa-tut=kaʈa #
(c) ʈhot # conebom # ‘I go.’ or ‘I leave.’
(d) ʈhu # ot=conebom #

However, two prefixes combined together do not form a word. The assign-
ment of the status of the word is decided by the prosodic phenomenon of
pause and stress. For further details see §2.5.7 and §2.5.8.

3.8.2 Nature of Proclitics

It was observed that the dual semantic system in PGA offers us gram-
maticalised morphemes functioning as class markers which are necessar-
ily attached to words drawn from all form classes. We shall now discuss
the nature of these obligatorily attached grammaticalised morphemes,
i.e., ‘proclitics.’ PGA has two kinds of proclitics:

(a) Pronominal clitics (for details refer to §7.3) which are phonetically
reduced forms of pronouns. These are ‘simple clitics’ (Anderson
2005).
(b) The ‘special clitics’, the body division class marker proclitics which
specify classes of various form classes and are attached to the left of
the hosts.

Although all class markers have some distinctive features, they share most
of the properties of being clitics. Since they attach to various form classes
they may be termed as: (i) Nominal proclitics, (ii) Verbal proclitics, and
(iii) Modifier (adjectival/adverbial) proclitics. These are special clitics. All
of them are unstressed monosyllabic words. When added to their hosts,
they operate as compounds.
Out of the six well recognised criteria that Zwicky & Pullum (1983: 305)
suggest for the identification of clitics, four of them are very relevant for
PGA clitics.

(a) Clitics and not affixes are least selective and can attach to any word
of a major word class, such as noun, adjective, adverb and verb. By
this criterion, class markers in PGA are clitics. PGA class markers also
meet at least three other criteria. These are:
(b) There are no unexpected forms or irregularities in clitics. Hosts are
unaffected by clitics.
 grammar overview 91

(c) There are no semantic idiosyncrasies, as they are predictable to a

large extent. It has been observed that PGA speakers readily assign
an appropriate class marker to a new word because of its high pre-
dictability and semantic transparency.
(d) Clitics can attach to material already containing clitics, but affixes
cannot. The PGA class markers are attracted by the pronominal clitics
and can form a word as seen above.

The parameters suggested by Aikhenvald (2002: 42–78) for the identifica-

tion of clitics match the observations made in PGA. Consider the follow-
ing arguments.

1. Selectivity: they can attach to nouns, verbs, modifiers, and to other clit-
ics. The class markers, as discussed in the sections above, can attach
to all form classes. Moreover, pronominal clitics attach to other class
marker proclitics. The latter quality gives rise to the possibility of clitic
sequencing.
2. The direction of proclitics is prefixing whilst the language is suffixing
in nature. This is the reason they have been termed ‘proclitics’ in this
grammar.
3. Phonotactic restrictions are such that they are monosyllabic in nature,
except one, i.e., class 6 ara=.5 The class marker proclitics always start
with a vowel, unless specified for inanimate nouns. Despite the fact
that the language has ample examples of VV sequences and medial
consonant clusters, the proclitics have the shape of V(C). Pronominal
clitics are constituted of single sounds, either a vowel or a consonant.
4. Two clitics can combine and result in a lexeme, e.g., a pronominal clitic
can combine with body class marker proclitics as in ʈh + ɛr ‘1sg + class
2’ meaning ‘mine.’ PGA thus has ‘clitic only’ words. Also refer to §3.8.3
below. The possibility of lexicalisation of class marker proclitics gives
us constructions like tottaral ‘on’ comprising of tot class 4, tara class 6
and -l ‘locative’ marker. There are other deictic categories which are
derived by lexicalising the proclitics. These will be considered in sub-
sequent chapters.

5 The hypothesis is that this particular proclitic is derived from the combination of two
proclitics, i.e., cl 1 a= and cl 6 ra=.
92 chapter three

5. Clitics can float, unlike affixes. It was considered in the discussion

above that the class marker proclitics can be distanced from their host.
Pronominal clitics have the option of being attached to the distanced
class marker proclitics (56a above) or to a case marker (§7.3), but they
always occupy the clause-initial position.
6. Considering the relationship with affixes, the language has few prefixes
and infixes, but a large body of suffixes in the areas of number marking,
case marking and TAM categories. Barring one example of enclitics
PGA offers us evidence for proclitics.
7. Another factor which is usually mentioned in the context of clitics is
the degree of grammaticalisation. As noticed above, the degree of gram-
maticalisation varies from one class marker proclitic to another and
from one grammatical category to another. Adverbial proclitics seem
to be highly grammaticalised. Refer to §3.7. This factor in combina-
tion with the one considered above in (c), proves that not all proclitics
have the same degree of grammaticalisation. Some of them are highly
grammaticalised while others are still transparent. This fact confirms
that these features are part of the language evolutionary path and clitic
formation is a diachronic phenomenon.
8. Dixon (2010), while defining various features of clitics, suggests that
they have the tendency of occurring at the periphery of all the affixes.
That is to say, they are attached after all the affixes have been attached
to the root. In PGA the pronominal clitic attaches to the reflexive pre-
fix ɛm- ‘self’ to the left of the whole word as in ‘spouse’ as in ʈh=ɛm-boi
‘my husband.’ Two affixes, on the other hand, cannot combine to make
a lexeme.
9. Phonologically, clitics follow the same rules as other words in the lan-
guage. This is seen most clearly at the morpheme boundary. Thus, class
marker proclitic ut= class 4 drops the final consonant t= across the
morpheme boundary if the next syllable starts with a dental plosive.

In PGA proclitics have a very significant status and need to be studied in

detail to understand the operation of the grammar.

3.8.3 Clitic Sequencing

Pronominal proclitics and class marker proclitics may be sequenced
together. The two clitics, simple and special, are strictly ordered in this
combination, one of the essential conditions specified for clitic combina-
tions (Gerlach & Grizenhout 2000: 10). The prosodic phenomenon that
 grammar overview 93

dictates the occurrence of these two clitics as one phonological word also
qualifies the phenomenon to be that of clitic sequencing (Spencer 2000:
368). Reconsider the phrases given earlier in §3.8.1 using the criterion (iii)
given there for clitic sequencing. To recall, we repeat two such examples
to show clitic sequencing. We use the symbol = twice to show double
clitics.
The verb ‘cut’ is ut=phoke but the class marker can move and attaches
to the pronominal clitic.

(60) ʈh=ut= ʈoŋe phoke

1sg=cl 4= tree cut
‘I cut the tree.’

The verb ‘get’ is i=coŋ but as one can see below i= can move next to the
subject.

(61) ŋ=i= ʈɔkhotɛc coŋ-o

2sg=cl 5= money get-dst.pst
‘You got the money.’

Although the class markers are compatible with nouns, modifiers and
verbs as lexically governed categories, they move towards pronominal
clitics. This phenomenon is known in PGA as the syllable attraction
rule (refer to §2.5.8). In a construction with a proper noun or a common
noun as a subject, or when the pronominal is in its full form (see below)
the class marker proclitic is not attracted to the pronoun, but is instead
attached to its host as in ut=cuʃu ‘to kill someone with a gun’ or ot=cobi ‘to
kill a large crowd with a gun’, and i=rube ‘to kill by stoning’:

(62) ŋu lao-bi ut=cuʃu-e

2sg foreigner-abs cl 4=fire-imp
‘You shoot a foreigner (non-Great Andamanese).’

(63) a=khuɽuc lao-n ot=cobi-k-o

cl 1=police foreigner-pl cl 4=fire-fa-dst.pst
‘The police opened fire into a large crowd of foreigners.’

(64) ʈhu meo-ta tajio-bi i=rub-e

1sg stone-instr fish-abs cl 5=kill-imm.pst
‘I killed the fish with a stone.’ or ‘I killed the fish by stoning it.’

Readers may also refer to §10.1.14.

94 chapter three

3.9 Summary and Conclusions

We prefer to name the ina markers proclitics and not prefixes, not only
because they precede the host category, but also because of the reasons
mentioned above in §3.8.1–3.8.3, which exhibit characteristic features that
are never shared by affixes. It appears that these class markers function as
proclitics and are the nucleus to which all the other grammatical catego-
ries such as nouns, verbs and modifiers are attached as satellites.
We saw that the class markers are grammaticalised in the language and
represent an inherent part of the cognitive world of the speakers. While
eliciting these words, whether in the areas of nouns, adjectives or verbs, it
was observed that the native speakers always attached the content words
with these class markers. Thus, when asked for the word ‘slap’, speakers
gave the form ut=bate and not bate. Similarly, the word ‘good’ was not
rendered in isolation as nɔl but i=nɔl. Thus, the various class markers can
be analysed morphologically as:

(i) Possessive proclitics or attributive proclitics

(ii) Verbal proclitics
(iii) Modificational proclitics
(iv) Nominal proclitics
(v) Pronominal proclitics

Proclitics (i) to (iv) may be long distanced from their host as they have the
tendency to be attracted by the pronominal proclitics (v). Figure 3.2 and
Table 3.2 summarise our findings.
To summarise, the class markers define and classify nouns, attributes,
states and actions in the language. The relationship between two nominal
categories or between a nominal and a verbal category (including those
defining a state) is seen as inherent where physical separation, if it exists
at all, does not make them ‘alienable’, as seen in the case of terms for cut
off body parts of animals. The notion of inalienability in the language rep-
resents a conceptual dependency between the object and its possessor. It
pervades transitive verbs where it establishes an inextricable link between
the mode of operation and the effect of an action; or in the case of intran-
sitive verbs it signifies ambience, the resultant experience, or the inherent
patient/object noun which cannot be alienated from the verb. It appears
with modifiers designating an inherent attribute which cannot be sepa-
rated from its locus. Table 3.2 represents the semantics of various proclit-
ics and host nouns, modifiers, and verbs. Some markers, if not all, retain
the basic meaning of body division classes across form classes.
 grammar overview 95

Inalienability
INA

Classification
by
modification

SEMANTIC
FUNCTION

Noun classes Verb classes Adjective classes

1‒7 1‒7 1‒7

MORPHO-SYNTACTIC Proclitics
FUNCTION

Figure 3.2. The classificatory functions of body class markers in PGA

Table 3.2. Semantics of body division class markers and bound form classes
Class Body Body Verbs Adjectives Adverbs
division class
classes markers
1 mouth and a= mouth-related mouth-related, deictic meaning
its semantic activity, origin quality of a of front or back,
extension person anteriority of an
action
2 major ɛr= activity in attribute of deictic meaning of
external which the size, external adjacency or front,
body parts front part of beauty uncontrollable
the body is actions/emotions
involved
3 extreme oŋ= hand-related attributes manner:
ends of the activity, action related to ‘hurriedly’
body like to do with limbs
toe and extremities of
fingernails body
4 bodily ut= directional, quality of an emerging out
products and away from X after a part of something,
part-whole the self, is taken out deictic meaning
relationship experiential of it of ‘towards X’, ‘up’
 96 chapter three

Table 3.2 (cont.)

Class Body Body Verbs Adjectives Adverbs
division class
classes markers
5 organs inside e=, ɛ= internalised inherent deictic meaning
the body action, when attribute of X of ‘in the middle
the effect of of X’, manner:
an action can ‘slowly’
be seen on
the object, or
experienced
6 parts ara= action that attribute of deictic meaning
designating involves side size, ‘time’ of something in
round shape or middle and belly- contact with or
and sides portion of the related periphery
body
7 parts for leg o= ~ ɔ= action which external temporal
and related more often attribute of deixis relating
terms than not results an X to ‘sunrise’ or
in roundish vertical deixis
object or in a
definite result
 Chapter four

Word formation processes

Introduction

A word in present-day Great Andamanese consists of several bound and

free morphemes. It is safe to say that words are formed in four different,
but interrelated ways: (a) affixation (b) by attaching proclitics (c) com-
pounding, and (d) by a combination of affixation, proclitics and com-
pounding. There is no evidence of reduplication in the language, other
than that found in the words for ‘mother’ and ‘swing.’ There is also no
evidence of the expressive morphology that is found in abundance in all
South Asian languages (Abbi 1991). The absence of reduplication defies
one of the significant Indian areal universals as PGA being the only lan-
guage of India which does not have this feature.
We shall consider each of the four kinds of word formation processes
in the following.

4.1 Affixation

PGA has a number of prefixes and suffixes that attach to roots giving rise
to new words with different meanings from the original un-affixed root.
Sometimes one has to take recourse to a longer translation because equiv-
alents in English are not expressed in two morphemes. The affixes listed
below do not include those affixes which are used as case markers or for
forming plurals. Nor does the list include TAM suffixes as they do not,
prototypically, form new words. Refer to §5.2.2 for various case marking
suffixes and to Chapter 9 on verbs for further information on how the
verbal inflection suffixes operate. The following prefixes and suffixes are
used to form words in the language.

1. /ɛm- ~ em- ~ un- ~ um-/ is a reflexive, as in phe ‘drown’, but ɛm-phe ‘dive’,
bele ‘flow’ and ɛm-bele ‘overflow’, bolo ‘scratch’ and ɛm-bolo ‘scratch oneself ’,
phorol ‘turn’, but em-phorol ‘turn over while sleeping’, cɛkho ‘face’ and em-cɛkho
‘agreed’, etc. If the reflexive is prefixed to an inanimate noun it is preceded
by t-. The reflexive prefix can be infixed between a verbal proclitic and the
verb root, e.g., ɛr=khole ‘to laugh’ can be rendered as ɛr=ɛm-khole ‘to laugh at
98 chapter four

oneself.’ Refer to §7.5 also. For details on marking inanimate nouns refer to
§3.3.4. Consider the following sentence:

(1) roa-bi tum-tɛrkhɛʈ-o

boat-abs refl-turn over-dst.pst
‘The boat turned over by itself.’

The following two affixes are constituted of two elements: a body division
class marker proclitic and a reflexive affix. Each of them, i.e. the combina-
tion of a proclitic and an affix, operates as a single prefix.

2. /ɑrɑm-/ is derived from two bound morphemes, the body division class marker
proclitic ara= and the reflexive affix ɛm- respectively. This is a reflexive mor-
pheme which is attached to verbs designating that the action is directed to
‘self.’ For example laʈ ‘fear’, but aram-laʈ ‘be scared of ’, ben ‘sleep’, but aram-
ben ‘to take rest/lie down/sleep.’ More often than not, this is attached to action
verbs which involve the side of the body. For details refer to §6.2.1.6 and Table
6.14.

(2) ŋa-l aram-liʃu eola-ø

2sg-loc refl-end see-pst
‘You saw yourself (in the mirror).’

3. /ɛrem-/ is derived by a combination of the body division class marker proclitic

ɛr= and the reflexive affix ɛm- ~ em-. The combination of the two indicates
reciprocal meaning as in ceo ‘fight’, but ɛrem-ceo ‘fight with each other.’

4. /ɛn- ~ en-/ is a prefix indicating a resultative or an inchoative action as in

ɛn-oke ‘knit with cane’, ɛn-ʈɔl ‘bloomed flowers’, en-tasue ‘cooked food’ while
tasue is itself derived by a combination of the applicative ta- and sue ‘to cook.’
ɛn-ʈɔlo ‘bloomed ones’ as in caybi enʈɔlo-k-o ‘flowers have bloomed all over the
place.’

5. /-ukhi ~ -jukhi ~ -jukhe/ ‘agentive’ and ‘specifier’ suffix added to a verb, an

adjective or a noun indicates ‘the specific one.’ For instance, aono ‘sit’ and
aono-jukhi ‘the one who is/was sitting’, nol ‘good’ but nol jukhi ‘the good one’ or
‘the one which is good’, ʈoya jukhe athire ‘the standing child’ or ‘the child who
was standing’, etc. This is very much like the Hindi particle wala which is suf-
fixed to any content word, i.e., noun, adjective, adverb, or verb for specificity
and agentive functions (see Abbi 2001). Refer to §5.3.1.5, §10.1.9, and §10.8 on
relativisation.

6. /-bim ~ -bem ~ -im ~ -em ~ -um/ is used as a ‘prohibitive’ marker.

khu -im > khum ‘don’t drink’, ‘not to be consumed’
iji -bim > ijibim ~ ijim ‘don’t eat’, ‘not to be eaten’
 word formation processes 99

7. /-bi/ is a nominaliser suffix which, when attached to a verb root, changes it to

a verbal noun. For details see §10.5.4. Consider the following sentences:

(3) ʈhu beno-bi inci-k-o

1sg sleep-nmlz go-fa-dst.pst
‘I went to sleep.’

(4) ʈhu ʈh=aɲa-bi inci-k-o

1sg 1sg=eat-nmlz go-fa-dst.pst
‘I went to eat.’

(5) u ʈoph-i inci-k-om

3sg bath- nmlz go-fa-npst
‘He is going to take a bath.’

If the finite verb is transitive and is preceded by the object clitic it= the
nominalisation marker assimilates to it and the combination of the two,
i.e., clitic and affix contraction pronounced as [bit], constitutes a single
word. Please refer to §2.5.7 and §2.5.8 for details. Consider the following
sentences:

(6) thire ŋolo bit jukhu

child cry nmlz.obj about to
‘The child is about to cry.’

(7) a=thire er=eŋ-khole-bit nol-ɔ

cl 1=child cl 2=refl-play-nmlz.obj well-dst.pst
‘Child’s play was good.’

Thus, the word for ‘call’ tere becomes tere-bit ‘for calling’ and ‘burn’ has a
verbal nominal in irulu-bit ‘for burning.’

8. /-iʃo ~ -ico/ is a genitive suffix and indicates alienable possession, e.g. ʈh=ico
julu ‘my clothes’, surmei-iʃo bindi ‘Surmei’s bindi’, niʃo ‘theirs’, reya-ico bol
‘Reya’s rope’, etc. These also describe modification by possession. Since these
are on the borderline of morphology and syntax, they can be considered as
forming new phrases and not new words. These are discussed at length in
Chapter 6 on possession.
100 chapter four

4.2 Attaching Clitics

4.2.1 Proclitics
We have so far established that the language offers two types of word
classes, dependent and independent, which are observable in nouns, adjec-
tives and verbs. The dependent classes are marked by the body division
class markers representing special proclitics. Refer to §3.8.2. The com-
bination of various proclitics and content words gives us a large lexicon
spanning various and different semantic fields. For example, the lexeme
tei ‘blood’ when combined with different proclitics designates different
sources of blood. Thus oŋ=tei ‘blood from finger’, er=tei ‘blood from head’,
o=tei ‘blood from legs’, ut=tei ‘oozing out of blood’, e=tei ‘blood inside the
body’ etc. Similarly, the word khir means ‘hot’ but with proclitic 5 as in
e=khir it signifies the ‘state of being hot.’ However, the same lexeme with
proclitic 4 as in ut=khir changes it into a transitive verb meaning ‘heat it.’
These examples can be taken as being constituted of two phonological
words. Thus, ɛr=-co ‘head’ is constituted of two phonological words: ɛr=
‘class 2.poss’ and co ‘head’ and together they constitute a grammatical
word, a noun designating a body part term. Similarly, the word ot=cone
‘go’ is constituted of two phonological words: ot= class 4 and the stem
cone and together, they constitute the verbal category meaning ‘go’ or
‘exit.’ This process of attaching proclitics to dependent nominal, verbal
and modifier categories is a very productive process of deriving words in
the language. We can symbolise this structure as: proclitic + n/v/mod.
The readers may recall that these proclitics are the various body division
class markers that were considered earlier in §3.3. Hence, lea ‘small’, but
a=lea ‘slowly’; ara=karap ‘waist’, but a=karap ‘behind someone’; e=buŋoi
‘beautiful, inner beauty’, but ɛr=buŋoi ‘beautiful to look at’; kaʈa ‘young
girl’, but ara=kata ‘dwarf ’; luk ‘lap’, but ara=luk ‘weigh’; jire ‘speak/speech’,
and a=jire ‘abuse’; ɖelo ‘ball’ and ara=ɖelo ‘pregnant’; and tujuro ‘shake’,
but uŋ=tujuro ‘trembling of hands’, etc. Many adjectives can be derived
from nouns by attaching the special proclitics or object clitics. Consider:

sare ‘sea’ > e=sare ‘salty’ [cl 5=salt]

julu ‘cold weather’ > et=a=julu ‘cold, cool’ [obj=cl 1=cold] > eta=julu-e ‘cool it’(V)
kata ‘a piece’ > et=kata ‘round’ [obj=a piece]
co lobↄŋ ‘big head’ > e=co lobↄŋ ‘clever’ [cl 5=head long]
bo i=ter=cek ‘big heart’ > aka bo i=ter=cek ‘kind’ [3sg heart cl 5=poss 2=much]
tɛkho cae ‘talk bad’ > aka tɛkho cae ‘rude, impolite’
 word formation processes 101

tɛkho tara=kamo ‘talk cl 6=much’ > aka tɛkho tara=kamo ‘talkative’

jom ‘food, eat’ > aka-jom ‘greedy’ [3sg-food]
kaʈa ‘girl’ > ara=kaʈa ‘dwarf ’ [cl 6=girl]

In addition to body division class markers operating as proclitics, PGA offers

the proclitic /tɑ=/ as a causative or applicative; ta=aono ‘make someone sit.’ A
speculation can be made that the causative proclitic could have been derived
from the body division class marker 1 which is a= or ta=. Since the dichotomy
between noun and verb is not particularly strict, as we saw in the previous chap-
ter, the same proclitic can generate a nominal category. The proclitic /ta=/ can
be attached to the same root aono to refer to the object on which the action of
sitting is performed, e.g., ‘chair’ or ‘stool.’ It is the word order which disambigu-
ates the reference. The ta= proclitic with verbs of an intransitive nature is pro-
ductively used to generate transitive verbs such as thu ‘be born’, but ta=thu ‘give
birth to’, ŋolo ‘cry’ and ta=ŋolo ‘make someone cry’, etc. It is a very productive
proclitic as the generic words referring to ‘birds’, ‘fish’, and ‘animals’ are derived
by affixing this to the root morpheme meaning ‘living creature.’ Consider some
of the examples below:

ta=jio-cor [caus=living being-scales] ‘fish’

ta=jio-tut-bec [caus=living being-poss 4=feathers] ‘birds’
ta=jio [caus-living being] ‘animals’

For details on causative formation please refer to §9.6.

Readers should refer to Chapter 3 for more information on how the
attachment of proclitics generates new words with varying meanings and
semantic extensions.
The combination of seven body division class markers or proclitics and
a number of content words has given rise to a large vocabulary. As lan-
guage evolution is a diachronic process with increasing stages of grammat-
icalisation, and which eventually leads to the lexicalisation of morphemes
giving rise to independent words, PGA also has a large number of lexical
items where various proclitics and affixes have been lexicalised. Words
such as okara ‘sunset’ or taratal ‘right under’, otle ‘seaward’, amu ‘dumb’,
obaloŋ ‘round’ have been lexicalised to such an extent that it is difficult to
segregate affixes or proclitics from the bound stem. One can hypothesise
that the bound grammaticalised morphemes, i.e. proclitics, infused with
the host morphemes over a long period of development of the language,
and thus, created a large number of new words.
102 chapter four

4.2.2 Enclitics
There is one example of an enclitic ­=n in the language which indicates
plurality. It co- varies with the word nu ~ no for ‘people’ which is also used
as a second constituent of the compound indicating plurality. Thus, cao
nu ‘dogs’ can co-vary with the form cao=n ‘dogs’ without any change in the
meaning. Refer also to §5.1.

4.2.3 Object Clitics

In addition to special clitics symbolised in body division classes, there are
several monosyllabic grammaticalised1 bound forms that attach to verbs.
These are: it= ~ et=, ek= ~ ik=. Many transitive verbs are attached to the
right of these object clitics to symbolise the nature of the object and the
manner of the action. For instance, ʈↄl ‘flower’, but it=ʈɔl ‘break’ and ik=ʈↄl
‘pluck flowers.’

(8) u lɛc ik=tɛrta-l-om

3sg arrow obj=shoot-fa-npst
‘He shoots arrows.’

(9) ŋo ik=caʈ-o
2sg obj=work-dst.pst
‘What did you do?’

(10) a=thire mitɛ-khudi et=ŋolo-m

cl 1=child milk-for obj=cry-npst
‘The child is crying for milk.’

(11) ʈhu ek=boniŋo ot=nↄl-o

1sg obj=understand cl 4=well-dst.pst
‘I understood it well.’

(12) a=meo ŋu ek=tɛr-o

cl 1=Meo 2sg obj=push-dst.pst
‘Meo pushed you.’

To summarise, PGA has the following clitics, most of them are used to
form new words.

1 They also seem to be derived from some of the body division classes, but it is difficult
to trace their origin due to the limited data available from the last few speakers.
 word formation processes 103

• Body division class marker proclitics: seven in all: a=, ɛr=, oŋ=, ut=, e=,
ara=, and o=. Each of these has a number of allomorphs as stated ear-
lier in Chapter 3.
• The causative or applicative proclitic a= ~ ta= can derive transitive
verbs from intransitive or causative from transitive verbs. It can be used
to generate a nominal category as we saw above in §4.2.1 for the word
for ‘chair.’
• Object clitics: ik= and it= and their variant forms.
• Enclitics: only one, which is =n to indicate plurality. It does not form
new words.

4.3 Combination of Affixation and Clitics

PGA forms new words by combining affixes and proclitics, especially with
words in the deictic categories (refer to §2.5.8). For example, the locative
case marker, -l, which is suffixed to a locational noun, can be combined
with a variety of proclitics and affixes to form a word which specifies the
exact location of the object under consideration. Consider Table 4.1.
There are phonological words that are derived from two bound mor-
phemes, each of them bound by the preceding and the following bound
stems. However, the stress pattern and the pause between the words
render these bound morphemes as one phonological word. For instance,
we witnessed earlier in §4.1 that the absolutive -bi, which is suffixed to a
noun, and the object clitic, which is attached to the following verb, can
combine together to form one phonological word, viz. bit as in sentences
6 and 7 above. Similarly, other phonological words that can be derived by
combining various clitics and affixes are the following. Sometimes a word
is derived by combining two clitics and sometimes by combining a clitic
and an affix, and further, rarely by the combination of two affixes:

Table 4.1. Spatial words and the combination of clitics and affixes
English gloss Andamanese form Morphology
on tut=ara:=l cl 4=cl 6=loc
inside k=ut=ara:=l obj=cl 4=cl 6=loc
under tum-bo-l refl-heart-loc
near cu-l near-loc
away ta=cakhu-l cl 1=platform-loc
in front of tɛr=cokhe-l cl 2=face-loc
behind (lit. back) ut=bɔ-l cl 4=back-loc
104 chapter four

(a) ʈhut < ʈh=ut= (1sg= cl 4) ‘my’

(b) ʈhɛm < ʈh=ɛm (1sg= refl) ‘myself ’
(c) ʈhota < ʈho-tɑ (1sg-abl) ‘from me’
(d) bit < -bi + it= (abs + obj)
(e) bin < -bi + en- (abs + result)
(f ) bik < -bi + ek= (abs + obj)

In (a–b) the first constituent is a pronominal clitic and the second con-
stituent is either a proclitic (a) or a reflexive prefix (b). Example (c) dem-
onstrates that any pronominal clitic can be combined with case markers
to constitute one phonological word. In (d–f ) the first constituent is
bound to the word on the left and the second is bound to the word on
the right, e.g., [kona-bi] [it=phoke] [‘tendu fruit-abs’] [‘obj=cut’] ‘cut the
tendu fruit’, which, when rendered, has two bound morphemes appear-
ing as one phonological word as in kona bit phoke. Also refer to §2.5.7
and §2.5.8, the syllable attraction rule. These kinds of structures are also
discussed in §9.3.

4.4 Compounding

The productive process of deriving words by compounding is well cov-

ered in the linguistic literature. There are several definitions and explana-
tions given for the concept of ‘compounds.’ As PGA compounds do not
fall neatly into one specific pattern, I will describe the morpho-semantic
characteristics of the compounds rather than define or identify the defini-
tion of compounds. As mentioned earlier, compounding in conjunction
with affixation and cliticisation facilitates a large number of word deriva-
tions which will be discussed in the following.

4.4.1 Formation of Compounds

There is a broad two-way division in PGA compounds: (i) where the mean-
ing is compositional, i.e., where the composite meaning of a compound is
provided by the individual meanings of the constituents of the compound
as we saw in the case of antonyms given above; (ii) where meaning is not
compositional, i.e., where the meaning of a compound is not transparent
from its constituents. The compound in this case refers to a third indepen-
dent entity and is also known as ‘idiomatic’, referring to a specific kind of
X, or which constitutes a lexical class type (Waelchli 2005: 105). This
kind is also known as Bahuvrīhi in Panini’s grammar (5th century B.C.).
 word formation processes 105

Diachronically, the non-compositional compounds can be seen as the

last stage of a grammaticalisation process in language development where
the lexical meaning of each of the constituents is delexicalised to form a
new meaning. Most of the proper names or specific names of the objects
in PGA are derived from non-compositional compounds. Interestingly,
there is no fixed word order in the compound formation. Hence, we have
examples of right-headed and left-headed compounds.
The following is a list of potential candidates for compounds and the
order of their appearance in a word in PGA. The combination of gram-
matical categories is shown by a plus symbol.

• N + mod
• N + N (Lexical construct (Fabri 1996))
• N + V or V + N
• N + proclitic + N
• N + proclitic + V
• N + proclitic + mod

I will now discuss each of these types of compounds in the following

sections.

4.4.1.1 Noun + Modifier

These are left-headed compounds and conform to the word order of the
modified (noun) and the modifier. These are descriptive in nature.

Table 4.2. Left-headed compounds

Noun Adjective Gloss Meaning
1 khider phuŋ coconut + dry dry coconut
2 ino julu water + cold cold water
3 ʃiro liu sea + little low level sea
4 tatamo ɖirim lizard + black black lizard
5 buruin leo hill + small hillock
6 ʃiro cɛrel sea-green vast sea, expanse of sea

4.4.1.2 Noun + Noun

These are right-headed compounds as the leftmost constituent modifies
the rightmost constituent. These are Tatpuruṣạ compounds where the
second member has primacy. The first member of the compound is in a
case relationship with the second one. Thus in example 1 in Table 4.3 it
106 chapter four

Table 4.3. Right-headed compounds

Noun Noun Gloss Meaning
1 teo mulu crocodile + eggs eggs of crocodile
2 teo rulu crocodile + eyes eyes of crocodile
3 meo phoŋ stone + cave cave made of stone
4 murgi thire hen + children chicken
5 cokbi ʈhomo turtle + skin turtle’s skin
6 khider ʈↄŋ coconut + tree coconut tree
7 khider ino coconut + water coconut water
8 ʈↄkho tei tree + blood Gum from a tree
9 jukhu bec above lip area + hair moustache
10 tap bec jaw + hair beard
11 teo bukhu crocodile + female female crocodile
12 bucɔ ʈↄŋ mangrove + tree mangrove tree

is ‘eggs—genitive—crocodile’, while in the last two examples (11 and 12)

it is a variety of Tatpuru$ṣ known as karmadhāraya (descriptive) in
Sanskrit. In karmadhāraya the relation of the first member to the last is
appositional, attributive or adverbial.
This conforms to the word order in genitive and possessive phrases.

4.4.1.3 Noun + Verb or Verb + Noun

A compound in this category becomes non-compositional as the individual

constituent does not carry the meaning of each of the constituents used
in the compound formation. Instead, the compound refers to some spe-
cific meaning other than those specified by the constituents. Most of the
bird names are derived by this process. As can be seen from Table 4.4 the
names of birds seem to be derived by juxtaposing the attributes of various
kinds with particular reference to birds and their habits. It appears that as
time passed these compound words became highly grammaticalised and
became reference terms. The system of naming and classification of Great
Andamanese birds is dealt with in Pande & Abbi (2011).
Other than birds, the naming of various objects of flora and fauna fol-
lows the same principle of compounding. For example names such as rↄi
ban ‘a kind of greenish crab’ is derived from two words rↄi ‘crab’ and ban
‘a kind of leaf ’; the word comulu ‘a kind of tree that bears small fruits and
grows by the sea shore’ is derived from two independent words co ‘seed’
or ‘head’ and mulu ‘eggs.’
 word formation processes 107

Table 4.4. Non-compositional compounds

Noun Verb Verb Noun Gloss Names of Great
Andamanese birds in
English
1 cɛ lele — — pointed object + Oriental Plover
swing
2 rɛn kↄ pinch + black ticks Andaman Green
Pigeon
3 cↄ ʈoʈ tie up + bark of a Eurasian Woodcock
tree
4 kara aʈ-com ascend from water + White-bellied Sea
wood - betel nut Eagle
5 mili ɖu move + sun/glitter Emerald Dove
6 thirin cↄrↄlo babies + downward Indian Hanging Parrot
flowing in space (it hangs upside down
from trees)

Table 4.5. Compositional compounds

N-class=N Meaning of individual Meaning of the compound
constituents
lao-tara=ɲyo foreigner-cl 6=house residence of foreigners
(now known as the city of Port
Blair)
bↄ-tara=iu chest-cl 6=pain chest pain
khider-tot=kↄbo coconut-cl 4=skin coconut peel/husk of coconut
bilukhu-tut=ɲyo God-cl 4=house cobweb
lɛc-tara=thomo arrow-cl 6=flesh quiver
ɖiu-tara=le sun-cl 6=edge direction of sunset, sunset, West
ɖiu-ter=bec sun-cl 2=hair hiding of the sun in the clouds,
cloud cover
ɖiu-tara=bat sun-cl 6=dark dusk

4.4.1.4 Noun + Proclitic + Noun

When two nouns are attached to each other by an intervening proclitic
indicating body division class, they can form a compound. This process
gives us both right-headed compounds as well as compositional com-
pounds, where the meaning is transparent.

4.4.1.5 Noun + Proclitic + Verb/Modifier

Nouns and verbs, or nouns and modifiers with an intervening body
division class marker proclitic can be coordinated to form a Bahuvrīhi
108 chapter four

compound. Thus, ʃiro tɛr=cɛk (sea cl 2=far) ‘centre of the sea’; bo i=ter=cɛk
(heart cl 5=cl 2=much) ‘kind person, person with a large heart’; tɛkho
tara=kamo (talk cl 6=much) ‘talkative’, are derived by this method.

4.4.1.6 Combination of Proclitics, Affixes, and Compounds

Non-compositional compounds are also formed by (1) a combination of
two free nouns, or (2) two dependent nouns, i.e., each attached with a
proclitic or (3) a combination of affix and nouns or verbs. More often
than not, they are Bahuvrīhi compounds where the referent does not carry
the lexicalised meaning of the words in question, e.g. ut=julu ter=cↄkho
(cl 4=front of/adorn cl 2=face) ‘mirror.’
A proclitic can also be attached to two nouns in succession as in
tara=ca-jobu (cl 6=nest-jewel) ‘shining nest’, ‘Edible–nest Swiftlet’, or to
two different form classes in succession, e.g., e=co-loboŋ (cl 5=head-long)
‘big headed’ or ‘clever’; ot=tei-kha (cl 4=blood-hortative) ‘blood extract/
emit (voluntary)’, ‘murderer.’
Consider Table 4.6 for words derived from a combination of proclitics
and compounds or a combination of proclitics, affixes and nouns. These
are non-compositional compounds.

Table 4.6. Combination of proclitics, affixes and nouns

Number Great Morpheme-to- Free translation Gloss in English
Andamanese morpheme gloss in English
1 ɛr=i=kak cl 2=cl 5=dir towards some- Aim at
thing with inter-
nalised result
2 ot=tei-kha cl 4=blood- blood extract/ Murderer
hortative emit -voluntary
3 tara=ca-jobu cl 6=nest-jewel shining nest Edible-nest
Swiftlet
4 cↄ-ʈↄ-l seen suddenly- The one which Chinese Sparrow
sky-loc is seen very Hawk
rarely in the sky,
i.e. a migratory
bird
5 cɛ-le-ne pointed object Crab eater bird Crab plover
(beak)-crab-pl
6 ta-ie caus-pain The pain giver Little Tern
(named after its
shrill voice)
7 tɛr=en-cokh cl 2=refl-head/ Self face/head Looking glass,
-ole or tɛr= face-see OR seeing one OR mirror
en-co-phol-e cl 2=refl-head-
love-agt Self head lover
 word formation processes 109

Table 4.6 (cont.)

Number Great Morpheme-to- Free translation Gloss in English
Andamanese morpheme gloss in English
8 et-ta-phul rsult-abl-love The one who is The one who
separated from loses a sibling
the loved one
9 aka-bere 3sg.gen-toad/frog Someone who One who talks
tara=kɔmo cl 6=too much leaps like a frog about himself/
in talk herself only
10 aka-kɛr-tot=l 3sg-throat-cl Thing that hangs Uvula
:coŋ 4=hanging in the throat

4.4.1.7 Antonyms
In addition to these compounds, antonyms in PGA can be derived by add-
ing the negative verb pho ‘not’ to a noun or a modifier. Thus:

nↄl ‘good’ > nↄl-pho ‘bad’ [good + neg]

eleo ‘small’ > eleo-pho ‘big’ [small + neg]
caecaʈo ‘work’ > caecaʈo-pho ‘useless’ [work-neg]
ot=bo-nↄl ‘happy’ > ot=bo-nↄl-pho ‘unhappy’ [cl 4.poss=heart-good-neg]

The negative pho can assume a full lexical status and can be used as a
negative verb with appropriate tense marking. This will be discussed in
Chapter 10.

4.4.1.8 Summary of Compounds

There are many possible semantic relations between different parts of a
compound. PGA employs various kinds of compounds each explicating a
distinct semantic relation between any two or three phonological words
that enter into a compound relationship. These compounds are the most
productive process of word formation.

1. Endocentric compounds where the core meaning is carried by the

‘head’ of the compound.
2. Excocentric compounds, where the status of all the elements in a com-
pound formation is equal, without a head, e.g. the words for ‘quiver’ or
‘sunset’ or ‘cobweb.’
3. Non-compositional compounds, where the constituents of a compound
do not carry their individual lexical meaning. The compound in this
case refers to a new entity, as in the case of bird names.
110 chapter four

Since the language uses a dependent marking system for all content
words such as nouns, verbs and modifiers, it is not surprising to see a
combination of cliticised stems forming compounds. The examples that
are found without the body division classes are semantically either non-
compositional compounds, i.e., co-bol (seed–rope) ‘knot’ or endocentric
compounds, e.g., tajiocor-bathe ‘fish-fins’ or cokbi-ʈhomu ‘turtle’s flesh.’ The
endocentric compounds are the most frequent compounds found in the
language.
As stated at the beginning of the chapter, there are no cases of ‘redu-
plication’ in the language. The only examples that were attested were lele
‘swing’ or mimi ‘mother.’ Other Great Andamanese languages such as Bo
and Khora, that are now extinct, seem to have had reduplication of a syl-
lable to form a new word. For instance, a folk song in Bo attested words
like gila gila ‘sitting with arms around knees.’ There was one case of echo-
formation thire-thiro ‘children etc.’ perhaps from the influence of Hindi. All
in all, complete reduplication of a word or of a syllable, or partial redupli-
cation as in echo formation is not a strategy that PGA allows for forming
new words.
 Chapter five

Nouns and noun phrases

Nouns

There are two types of nouns in present-day Great Andamanese (PGA):

dependent and independent. Dependent nouns obligatorily take a body
division class marker denoting inalienable possession or an inherent rela-
tion between the possessor and the possessed as discussed in §3.3. The
nominal category which denotes the body, whether it designates a part of
the body, a product of the body or something that is related to the body
is considered to be dependent. All body part terms, kinship terms, parti-
tives, as well as nouns referring to time, direction, and depth are depen-
dent nouns and thus are obligatorily preceded by a body division class
marker as a proclitic. These are considered in detail in Chapter 3 as
well as in Chapters 6 and 8 on ‘Possession’ and ‘Modification and deixis’
respectively. It was mentioned in §3.3.3 that words for physical attributes,
physical and mental ailments are also preceded by a body division class
marker. These markers, in addition to representing ‘inherent’ aspects,
symbolise the location of the nominal object with respect to the body.
Some examples considered before are repeated below to recapture the
discussion:

(a) ot=cala (class 4=scar) ‘scar left by arrow-head’

(b) er=cala (class 2=scar) ‘scar on the head’
(c) oŋ=cala (class 3=scar) ‘scar on the limbs’
(d) e=tei (class 5=blood) ‘blood inside the body’
(e) ot=tei (class 4=blood) ‘blood outside the body’ [when bleeding]
(f) oŋ=tei (class 3=blood) ‘blood on finger or from finger’

Nouns designating different kinds of ailments take different class markers

depending on which part of the body is affected. For instance:

(a) ara=mikhu-tei (class 6=middle-blood/pain) ‘stomach ache’

(b) er=belɔɛ (class 2=pimples) ‘pimples’
(c) er=co bie (class 2=head-pain) ‘headache’
(d) er=e-ʈɛŋe (class 2=class 5-measles) ‘measles’
(e) ot=tei (class 4=pain) ‘splitting headache’
112 chapter five

All in all, PGA marks the following nouns as dependent categories and
thus body part semantics can individuate noun references. One of the
seven class markers discussed earlier precedes these noun categories. Also
refer to Chapter 6 on possession.

1. Kinship terms (exceptions are ‘son’ and ‘daughter’)

2. Body parts, body products, tattoos, body paint, body attributes
3. Home, village, courtyard
4. Language, words
5. Ailments, both physical and mental
6. Boat, head gear, cap, covering for private parts made of leaves, clothes
(in modern times) and boat
7. Spatial terms
8. Parts of objects or parts of a component
9. Terms relating to the sea and landscape terms
10. Incorporeal, viz. spirits, ghosts, supernatural beings and soul

Independent nouns are those that refer to the names of flora and fauna,
and generic names for celestial bodies and other words related to the
environment.
These nouns remain invariant in their shape when a case marker,
whether a suffix or an adposition, is suffixed to them.
A minimal nominal word can consist of just of a root, or a proclitic +
root, or root + root (as in compounds), or a combination of the two. Nomi-
nal compounds can be very long as they are formed of multiple roots.
These are quite frequent in the language. Refer to Chapter 4 on word
formation processes.

5.1 Number and Gender

There are three numbers, singular, dual and plural encoded in pronomi-
nal forms (see §7.1 on pronouns). However, nouns in general are not
marked for duality and plurality. The word nu ~ ne ~ ni ~ in ‘people’ is
suffixed to a few select animate nouns to mark plurality.1 As discussed in
the last chapter this form can also appear as an enclitic =n. For instance,

1 Interestingly, this feature of forming the plural by juxtaposing the word for ‘people’ to
animate nouns is very common in Hindi spoken in the Northeast of India. Thus kutta log
‘dog people’ refers to the additive plural ‘dogs.’
 nouns and noun phrases 113

bukhu-ne untele ‘call the ladies’, but mimi=n untele ‘call the mothers.’ Simi-
larly, ese ‘mother in law’, but ese=n ‘mothers-in-law.’ The forms nu, no and
ni can also be used for the word ‘people’ or in a generic sense of ‘and oth-
ers’ independently (4–5). We can safely say that the word nu ‘people’ is
grammaticalised as a plural marker enclitic =n.

(1) thire=ni kona-bi raliʃu-k-o

child=pl kona-abs finish-fa-dst.pst
‘Children finished the tendu fruit’

(2) ŋ=ut thire=n cophe bi

2sg=cl 4.poss child=pl much cop
‘You have many children.’

(3) khudi=no cae-bi caʈo-k-o

3prox.int=pl what-abs work-fa-dst.pst
‘What were they doing?’

The word nu for ‘people’, as can be seen in the following examples, assumes
the subject position and acts as an independent word.

(4) a= caʃiu nu porʈblɛr-ak neɲo-b-e < nu eɲobe

cl 1= who people Port Blair-dir people.leave-fa-npst
‘Who all will leave for Port Blair?’

(5) a= raja unni jira-l nu iku=beliŋ-ø

cl 1= king 3pl speak-nrr.pst people obj.clt=cut-pst
‘The king told them, the people cut (the tree).’ or
‘The king had the people cut the tree.’

The plural marker is also attached to proper nouns to indicate additional

non specified members. Consider the following sentences.

(6) Naobulini εk tεrta

Nao-Buli-ni ɛk= tɛrtɑ-ø
Nao-Buli-people obj.clt= narrate-pst
‘Nao narrated a story to Buli and others (other children).’

(7) ŋale amimin untɛleke

ŋɑle ɑ=mimi=n untɛle-kh-e
2pl.hon cl 1=mother=pl call loudly-fa-imp
‘Call your mothers.’ (Referring to several of them)
114 chapter five

Note further that many words denoting domestic animals, for example
‘dogs’, can have plural forms marked by nu, while others, such as the word
for ‘pigs’ or ‘cats’, cannot.

(8) cao-nu ra-be ɲa-om

dog-pl pig-abs bark-npst
‘Dogs are barking at the pigs.’

(9) cao–nu kheŋe-be ɲa-b-o

dog-pl cat-abs bark-fa-dst.pst
‘Dogs barked at the cats.’

As discussed in §3.2.4 temporal deictic markers are attached to pronomi-

nal proclitics on their left. In the following sentence plural pronominal
proclitic n= co-refers the plural subject of the sentence. Consider:

(10) reya rengi sʈreʈ-ak n=ambikhir ɔt=cɔni-b-ɔm

Reya Renge Strait-dir 3pl-morning cl 4=go-fa-npst
‘Reya and Renge will go to Strait tomorrow.’

Another strategy of forming the plural is to use the quantifier for ‘many’
or ‘much’ cɔfe to indicate inanimate as well as non-human animate plu-
rals. Thus, constructions such as pata cɔfe ‘many mushrooms’ or ra cɔfe
‘pigs’ are common. However, quantifier and plural markers may coexist
(11) with human nouns.

(11) thire=n cɔphe

child=pl many
‘Many children’

An alternative pattern is also emerging where the plural morpheme -ni is

totally dropped and is replaced by cɔfe even in human nouns. Hence thire
cɔfe ‘many children’ is not uncommon in the speech of some speakers.

(12) ŋu(t) thire cɔfe be

2sg.cl 4.poss child many cop
‘You have many children.’

5.1.1 Numerals
There are only two numbers available: ‘one’ and ‘two.’ No one in the com-
munity remembers whether they had more than two numbers. Any count
 nouns and noun phrases 115

beyond two is ‘many.’ However, one consultant gave us the word incinkɔ
for ‘three.’ Manoharan (1989) notes the word for ‘four’ as maɛxuccol which
was translated by our consultant as ‘the last one.’ An interesting feature
about numerals is that cardinals can act as predicates. Consider:

(13) ʈhu (t)=thire ʈɔplɔ

1sg.cl 4.poss=child one
‘I have one child.’

(14) ʈhu (t)=thir-in nɛrtaphul

1sg.cl 4.poss=child-pl two
‘I have two children.’

(15) ʈhu (t)=thire akaʈa nɛrtaphul

1sg.cl 4.poss=child girl two
‘I have two daughters.’

(16) ʈhu(t)=thire narakamu

1sg.cl 4.poss=child many
‘I have many children.’

5.1.2 Gender
There is no grammatical gender in PGA. It does not encode grammatical
gender in agreement, as can be seen from the following examples (17–18).
Nor is gender morphologically coded on nouns. However, the word bukhu
for ‘female’ or ‘woman’ is productively used in a compound formation to
signify female counterparts of male entities. For instance, cao ‘dog’, but
cao-bukhu ‘bitch.’

(17) sita konaphuŋbi jiyom

sitɑ konɑ-phuŋ-bi iji-om
Sita tendu-ripe-abs eat-npst
‘Sita is eating a ripe tendu.’

(18) ram konabi jiyom

rɑm konɑ-bi iji-om
Ram tendu-abs eat-npst
‘Ram is eating a tendu.’
116 chapter five

5.2 Case

As we saw in Chapter 3, ‘Grammar overview’, PGA is an ergative-absolu-

tive language with a fluid S function. The language offers two different sets
of markers, one for prototypical agent nominals, -e suffixed to the agent
nouns, and another one, -bi attached to the subject of unergative intransi-
tive verbs, the subject of unaccusative verbs and the object nominals. We
have symbolised this relationship as:

So = Sa = O -bi absolutive
A -e or unmarked ergative

This means that PGA maintains the ergative-absolutive distinction with

the absolutive being overtly marked. We also notice some instances of
case-alternation.
PGA does not have prototypical cases, but does have case markings
that appear as case suffixes. The case markings prototypically encode par-
ticipants’ roles and are more in the order of relators relating the noun
with the verb. Some of the relations are very detailed and specific while
others are generic. Hence, there are many locative markings as can be
seen in §8.4 on deixis, while the patient, agent or instrumental are not
classified in such detail. Moreover, some case markings are optional such
as ‘nominative’, ‘accusative’, ‘dative’, but others such as ‘instrumental’
and ‘ablative’ are obligatorily marked. Hence, many times an argument
appears without any case suffix and its meaning is designated by the con-
text. Another interesting feature concerning the noun-verb relationship
is that this relation is also designated by the proclitic attached to verb
roots. For instance, experiential verbs that co-occur with non-nominative
or dative subjects are marked by the class 4 proclitic.

5.2.1 Arguments
Human subjects and objects, particularly those with proper names, are
always preceded by class 1 proclitic a= in the Jeru and Bo varieties. This
means the marker is not present in the speech of Khora, Pujjukar and
Sare. Also refer to §3.1.3.

(19) a kaba a buro ikjiral a bie ta-elukho

ɑ=kɑbɑ ɑ=buro ik=jirɑ-l ɑ=bie tɑ-e=lukh-o
cl 1=Kaba cl 1=Buro 3sg.obj=tell-conv cl 1=Bie tr-cl 5=pick-dst.pst
‘Kaba asked Buro to pick up Bie.’
 nouns and noun phrases 117

(20) aboa ameo ikjiral ikubeliŋ

ɑ=boɑ ɑ=meo ik=jirɑ-l ik=u(t)=beliŋ-ø
cl 1=Boa cl 1=Meo 3sg.obj-tell-conv 3sg.obj=cl 4=cut-pst
‘Boa asked Meo to break the coconut.’

5.2.2 Case Markings

Case marks the relationship of a noun to a verb at the clause level or of a
noun to a postposition, or another noun at the phrase level. In PGA case
marking appears in two forms: (a) a bound form as a case suffix attached
to nouns and noun phrases and, (b) as an independent phonological word
as an adposition appearing after noun phrases. Adpositions indicate the
grammatical or semantic relationship of a noun phrase to the verb in the
containing clause.
The case suffixes are presented in the following table. Some of them
are homophonous.

Table 5.1. Case markings

Case/relators Suffixed post-nominally
Absolutive (So, Sa) -bi ~ ø
Agentive -e, ~ -ø
Dative -i, -bi
Instrumental -ta
Ablative -ta
Genitive, alienable -iʃo ~ -ico
Comitative -kɔc
Purposive/benefactive -ɔkka / -kocara
Directional -ak ~ -a
Locative -al ~ -il ~ -el ~ -l

Readers will notice that there exist, throughout the grammar of PGA,
many instances of optional overt representation of any argument mark-
ing, hence zero marking alternates with agentive -e and with absolutive
-bi. This could be due to the process involved in the genesis of koiné or
due to some pragmatic reasons.

5.2.2.1 Intransitive Subject Marking

The subject of an intransitive which has an O (So)-like function [non-
action verbs] as in ‘vessel filled up’, ‘sun showed up’, ‘moon appeared’,
etc., and the subject of an intransitive which has an A-like function (Sa)
118 chapter five

[action verbs, but involuntary and uncontrollable], ‘baby slept’, ‘hens

roamed around’, ‘they dreamt’, etc., are marked by the absolutive case
suffix -bi. The case markers are underlined in the following examples.

(21) ʃarebi jiokɔ

Sɑre-bi jio-k-ɔ
sea-abs come down-fa-dst.pst
‘The sea has come down.’

(22) thire-bi ŋol-om

child-abs cry-npst
‘The child cries.’

(23) beibi tunʈɔlo

bei-bi tun-ʈɔl-o
bottle-abs refl-break-dst.pst
‘The bottle broke (itself ).’

(24) mɔcɔbci ʈhikaːlaːm

mɔcɔ-bi ʈhi kɑːlɑː-m
hen-abs Earth roam-npst
‘The hen is roaming on the earth/ground.’

(25) bɔrbe bɔrom

bɔr-be bɔr-om
wind-abs blow-npst
‘Wind is blowing.’

Plural subject nouns are not marked for their argument functions. Absence
of markings are shown with a zero symbol ø. However, we could attest a
few cases of plural pronouns marked for argument structure in narration.
See ‘Text’ (sentences 81, 82 and 85) given at the end of the grammar. It
should be mentioned here that the majority of the speakers do not attach
any case marking to plural pronominal forms.

(26) thirenu ŋolom

thire-nu-ø ŋol-om
child-pl cry-npst
‘The children are crying.’

(27) o aʈ era phuvom

o-ø ɑʈ erɑ=phu-om
3sg wood cl 2=cut-npst
‘He is cutting the trees.’
 nouns and noun phrases 119

5.2.2.2 Transitive Subject Marking and Ergative/Agentive

The agentive marking, -e, is suffixed to the agentive noun. Pronominals in
their clitic forms can be suffixed by it as in ʈhe ‘I’ or ŋe ‘you.’ This feature
appears to be falling into disuse. The ergative marker, -e, is not obliga-
tory in every speaker’s speech. Thus, agent nouns are not marked by -e in
sentences (29–30).

(28) aʃyame bas kuttaral konabit beliŋo

ɑ=ʃyam-e bɑs kottrɑl konɑ-bi it=beliŋ-o
cl 1=Shyam-erg bus inside tendu-abs 3obj.clt=cut-dst.pst
‘Shyam cut the tendu fruit in the bus.’

(29) anilli ara-itɔile

ɑ=nilli ɑrɑ=i=tɔil-e
cl 1=Nili cl 6=cl 5=roam around-cop
‘Nili roams around.’

(30) athire ɖileʈmo-bi ereŋkholam

ɑ=thire ɖileʈmo-bi ereŋ-khol-ɑm
cl 1=child ball-abs refl-play-npst
‘The child is playing (with) a ball (all by himself ).’

One often hears a sentence without the class 1 proclitic a- and the erga-
tive suffix -e. The Bo variety does not mark the case suffix -e after an agen-
tive nominal (31, 32). Consider:

(31) alico rɛfi raʃuekom

ɑ=lico rɛfi rɑ=ʃue-k-om
cl 1=Lico rice cl 6=cook-fa-npst
‘Lico cooks rice.’

(32) tajiotɔtbɛc ʈautɛr rulɛk

tɑjiotɔtbɛc ʈɑu tɛr=rulɛk
Birds cloud/sky cl 2=glide
‘The birds are gliding in the sky.’

(33) ajoe ŋeleolebe

ɑ=jo-e ŋe-l e=eole be
cl 1=Jo-erg 2.abs-fa cl 5=see cop
‘Joe sees/will see you.’

(34) a-ram-e a-ravanek ɛrenceo

ɑ=rɑm-e ɑ=rɑvɑn ek=ɛrence-o
cl 1=ram-erg cl 1=ravan obj.clt=fight-dst.pst
‘Ram fought with Ravana.’
120 chapter five

5.2.2.3 Object Marking

Since the language maintains an ergative-absolutive distinction, all
direct object nouns marked by any of the case suffixes such as -bi, -be,
-e, -ø are marked as abs in the following sentences and elsewhere in the
grammar.

(35) ʈhu buruiŋbi eola

ʈhu-ø buruiŋ-bi eolɑ-ø
1sg mountain-abs see-pst
‘I saw the mountain.’

(36) ram meobi jiyom

rɑm meo-bi iji-om
Ram mango-abs eat-npst
‘Ram is eating a mango.’

In a ditransitive construction this case suffix occurs twice but can also be
dropped from one of the object nouns as in (38).

(37) ra-bi kona­-bi tɛʃ-e

pig-abs tendu-abs give-imp
‘Give the tendu fruit to the pig.’

(38) a-joe toŋ-e cota khidere ebe

ɑ=joe-e toŋ e=co-tɑ khider-ø ebe-ø
cl 1=Joe-erg Tong cl 5=near-abl coconut take-pst
‘Joe took coconut from Tong.’

The choice between object marking and non-marking is not free, but is
determined by specificity. This was especially observed in discourse. If the
object is overtly marked it specifies the object under consideration, if it is
unmarked the object under consideration is interpreted as generic.
Another important fact worth noting is that there exists at times, an
optionality between presence and absence of the absolutive marking -bi
on nouns especially if there is an object proclitic or a body division ver-
bal proclitic that attaches to the left of the verb in the same sentence.
In sentence 39 the absolutive -bi is optional and can be dropped (40).
While object marking in casal suffix form has become optional (as dis-
cussed above), in the speech of some speakers, obligatory attachment of
the body class proclitic or object clitic to verbs that indicate the transitiv-
ity of the verb is not optional and is still an inherent and necessary part of
the grammar. In Chapter 3 it was mentioned in §3.1.3 that some speakers
 nouns and noun phrases 121

used both markings. Without the sufficient statistical survey of such uses
one can not reach any conclusive judgment for this alternation. Refer to
Chapter 9 on verbs for details.

(39) ʈhu ʈɔŋbie rapho

ʈhu ʈɔŋ-bi e=rɑpho-ø
1sg tree-abs cl 5=cut-pst
‘I cut the tree (the specific one).’

(40) o ʈɔŋ erapho

o ʈɔŋ-ø e=rɑpho-ø
3sg tree cl 5=cut-pst
‘He cut a tree.’ (some tree)

(41) amime rɛfebe jikɔm

ɑ-mimi-e rɛfe-be iji-k-ɔm
cl 1=mother-erg food-abs eat-fa-npst
‘The mother is eating food.’

As noted earlier, in the case of subject marking with plurality in §5.2.2.1,

plural objects marked by the =n enclitic do not take object marking -bi.
However, the object is marked on the verbs with the proclitic er= which
indicates that it is a transitive verb with an object noun. This is an inter-
esting observation as it implies that case marking is indicated on verbs.
Thus:

(42) *thire mɔco-n-bi er=phuko-ø

child hen-pl-abs cl 2=hit-pst
‘The child hit the hens.’

But
(43) thire mɔco-n-ø er=phuko-ø
child hen-pl-(abs) cl 2=hit-pst
‘The child hit the hens.’

And
(44) thire mɔco-bi er=phuko-ø
child hen-abs cl 2=hit-pst
‘The child hit the hen.’

Interestingly, if the plural enclitic, =n, is replaced by the fully lexicalised

form, nu ‘people’, or if followed by the modifier, cɔphe ‘many’, then the
absolutive marker, -bi, can occur after the noun phrase. Consider:
122 chapter five

(45) thire mɔco cɔphe-bi er=phuko-ø

child hen-many-abs cl 2=hit-pst
‘The child hit the hens.’

Or
(46) thire mɔco-nu-bi er=phuko-ø
child hen-people-abs cl 2=hit-pst
‘The child hit the hens.’

This implies that (i) the absolutive -bi attaches to plain nouns which are
stripped of any inflectional endings. (ii) In the case of a noun phrase with
modification the absolutive case marker is suffixed to the whole noun
phrase.

5.2.2.4 Dative
The dative marking is -i suffixed to pronouns, as in ʈhi ‘to me’ or ŋi ‘to you’
as well as to proper nouns. Speakers have been observed dropping it in
conversation as the meaning is ascertained by the context.

(47) ʈhi miʈhaibi tɛʃeke

ʈh-i miʈhɑi-bi tɛʃe-k-e
1sg-dat sweet-abs give-fa-imp
‘Give me the sweets.’

(48) ʈhuʈhire ʈhi ʈɔkhotɛcbi tɛʃe

ʈh=ut=ʈhire ʈhi ʈɔkhotɛc-bi tɛʃ-e
1sg=cl 4.poss=child 1sg.dat money-abs give-imp
‘Give me money for my child.’

(49) naŋe ciʈhi-bi peje-i mok-e

2sg.h letter-abs peje-dat give-imp
‘You give the letter to Peje.’

5.2.2.5 Instrumental
The instrumental case marker is -ta ~ -te suffixed to nouns of instrument
and initiators in typical causative constructions. One of our speakers, Peje,
used the form te-.

(50) lɛc ʈɔkho-ta eban-o

lɛc ʈɔkho-tɑ e=bɑn-o
arrow wood-instr cl 5=make-dst.pst
‘Arrow was made of wood.’
 nouns and noun phrases 123

(51) ʈhu bol-ta cokbi-bi i=ʈhudo-ø

1sg bol-instr turtle-abs cl 5=kill-pst
‘I killed the turtle with the help of a Bol-rope (used in turtle hunting).’

(52) roa-ta sʈreʈ-ak ʈhut= cone-b-o

boat-instr Strait-dir 1sg.cl 4= go-fa-dst.pst
‘I went to Strait by boat.’

(53) athirenu ceuta konabit beliŋo

ɑ-thire-nu ceu-tɑ konɑ-bi it=beliŋ-o
cl 1=child-pl knife-instr tendu-abs obj=cut-dst.pst
‘The child cut the tendu (in pieces) with a knife.’

(54) ahanuman aramulibite lankabi ʃuiko

ɑ=hanuman ɑrɑ=em-ulibi-te lɑnkɑ-bi ʃui-k-o
cl 1=Hanuman cl 6=refl-tail-instr lanka-abs burn-fa-dst.pst
‘Hanuman burnt Lanka with his tail.’

(55) ʈhu bolte etcobɔ

ʈhu bol-te et=co-b-ɔ
1sg rope-instr obj=tie-fa-dst.pst
‘I tied (it) with the rope.’

(56) inota ʈhukhube

ino-tɑ ʈhu-khu-be
water-instr 1sg-drink-imp
‘I will drink with water (the alcohol).’

(57) uroʈɔyil ɲyotaratɛŋ ɲyotɛcta ɲyobi embanoma

uroʈɔy-il ɲyo-tɑrɑ=tɛŋ ɲyo-tɛc-tɑ ɲyo-bi
ancient-loc house-cl 6=pole house-leaf-instr house-abs
em-bɑno-mɑ-ø
refl-make-hab.pst
‘In ancient times, the houses were made of poles and leaves.’

(58) ʈhu lɛcta tajiocorbit okho

ʈhu lɛc-tɑ tɑjiocor-bi it=okho-ø
1sg arrow-instr fish-abs obj=kill-pst
‘I killed the fish with an arrow.’

There are no examples where both the instrumental -ta and the causative
ta= are used in the same sentence. It is not very clear whether the suffix
-ta or -te is instrumental or the causative proclitic to the verb ‘cut’ with the
direct object has an incorporated nominal which is not very uncommon
in the language. Refer to §9.6. Sentence (59) given below indicates the
instrumental function while (60) expresses causative function.
124 chapter five

(59) ʈhu nata ʈɔŋe rapho

ʈhu nɑ-tɑ ʈɔŋ e=rɑpho-ø
1sg people-instr Tree cl 5=cut-pst
‘I got the tree cut by the people.’

(60) ʈhu na ʈɔŋ tae rapho

ʈhu nɑ ʈɔŋ tɑ=e=rɑpho-ø
1sg people tree caus=cl 5=cut-pst
‘I got the tree cut by the people.’

They both have the same meaning.

5.2.2.6 Ablative
The ablative and instrumental are homophonous, i.e. -ta ~ -te suffixed to
the noun which indicates the source of separation.

(61) kɔʈphɛcta ekambelekom

kɔʈ-phɛc-tɑ ek=ɑmbele-k-om
soil-pot-abl obj=overflow-fa-npst
‘(Water) is overflowing from the clay-pot [pitcher].’

(62) ʈoŋ-ta ʈɔkho-tɛc-bi bothom/bɔrothe

ʈoŋ-tɑ ʈɔkho-tɛc-bi both-om/bɔro-th-e
tree-abl wood-leaf-abs fall-npst/blow-fa-cop
‘Tree-leaves are falling from the tree.’

(63) ʈhu ocɔta itcavom

ʈhu ocɔ-tɑ ut=cav-om
1sg net-inst/abl cl 4=spread-npst
‘I throw the net (in the water).’
(It is to be noted that no source is specified).

(64) ʈho porʈblɛrta ʈhɛɲobe
ʈho/u porʈ-blɛr-tɑ ʈh=ɛ=eɲo-b-e
1sg port blair-abl 1sg=cl 5=come from-fa-npst
‘I come from Port Blair.’

(65) aʈ lurota ʈole-bi elute

ɑʈ luro-tɑ ʈole-bi elut-e
wood fire-abl potato-abs take out-imp
‘Take the potato out of the fire.’
 nouns and noun phrases 125

5.2.2.7 Comitative
The comitative morpheme is an adposition, kɔc or bɔco ~ buce, and is
placed in between the two nouns that stand in a comitative relationship
with each other.

(66) ʈhu bɔco ŋu iji-b-om

1sg com 2sg eat-fa-npst
‘I will eat with you’ (‘we will have meals together’)

(67) joe kɔc meo unni-b-o

Joe com Meo return-fa-dst.pst
‘Joe came back (returned) with Meo.’

We noticed that most of the time speakers dropped the adposition as

in (68–69). At times, the adposition was used after the two coordinated
nouns (71). Thus:

(68) ʈho cao-ne ʃiʈ-b-om

1sg dog-pl hunt-fa-npst
‘I go hunting with dogs.’

Or
(69) a=cao-no ʈh e ʈhi-bi it=ɲyo
cl 1=dog-pl 1sg.abs place(=live)-abs cl 5=home/live
‘Dogs live with us.’ (bi + ot > bit)

(70) rɛfe-bi bɔcɔ cae-bi ji-o

rice-abs com what-abs eat-dst.pst
‘What else did you eat with rice?’

(71) thu alok-e buce kalom/kalebom/kalekom

1sg-alok-erg com stroll.npst
‘I am strolling with Alok.’ (I and Alok are strolling together.’)

Consider another sentence where the comitative meaning is expressed

without the use of kɔc. It appears that class marking a= with the verb ci
indicates comitativity. Again, the case is marked on verbs. Refer to Chap-
ter 9 for details.

(72) ʈho ŋ=a=ci-b-om ʈho kele pho be

1sg 2sg=cl 1=go-fa-npst 1sg stay neg cop
‘I will come with you but will not stay back.’
126 chapter five

5.2.2.8 Purposive/Benefactive
The postposition for the benefactive is khudi which also represents
‘purposive’ or ‘cause’ functions. These are underlined.

(73) mitɛ khudi et=ŋolo-m

milk purposive obj=cry-npst
‘(He) cries for milk.’

(74) dillikɛk caːy khudi ŋutconebom

dilli-kɛk cɑːy khudi ŋ =ut=cone-b-om
Delhi-dir what purposive 2sg=cl 4=go-fa-npst
‘Why are you going to Delhi?’

The following sentence can be interpreted as showing a ‘cause and effect’

relationship.

(75) tɛkhucaːy khudi ʈhɛren ʈhicaːy

tɛkhu-cɑːy khudi ʈh=ɛr=en ʈh=i=cɑːy
talk-bad because 1sg=cl 2=refl 1sg=cl 5=bad/mad
‘I became mad (rough) because this talk was bad (unacceptable).’

As in other case markings even this is no longer obligatory, and purposive

meaning is interpreted through the context. Consider:

(76) ʈhu khider-bi inci-k-om

1sg coconut-abs go-fa-npst
‘He goes for coconut.’

Also refer to §10.5.

5.2.2.9 Directional
The directional case suffixes combine with spatial deixis markers, which
will be referred to in Chapter 8. However, we give here the most common
directional suffix -ak that specifies the goal of an action. It is underlined
in the following examples.

(77) ʈhu ʈhimikhu-ak ʈh=ot=cɔne-b-ɔm

1sg forest-dir 1sg= cl 4=go-fa-npst
‘I am going towards the forest.’

(78) sʈreʈ-ak ɖun ot=cɔne-k-ɔm

Strait-dir 3pl cl 4=go-fa-npst
‘They will go to Strait.’
 nouns and noun phrases 127

(79) ʈh=ambikhir sʈreʈ-ak ʈh=ot=cɔne-b-ɔm

1sg= morning Strait-dir 1sg=cl 4=go-fa-npst
‘I will go to Strait tomorrow morning.’

(80) ʈhu ɲo-ak uni-b-om

1sg house-dir come-fa-npst
‘I will return home.’

(81) o roə-ta ʃare-a(k) ʈhut connom

o roə-tɑ ʃare-ɑk ʈh =ut=conn-om
3sg boat-instr sea-dir 1sg= cl 4=go-npst
‘He is going into the sea in a boat.’

(82) terkɔ-otɔ ʈɔrɔ-ak ʈh=ut=cɔne-b-ɔm

Day after tomorrow beach-dir 1sg=cl 4=go-fa-npst
‘I am going to the beach the day after tomorrow.’

(83) terkɔotɔ porʈbler-ak ʈh=u(t)=thu-ɔ

Day before yesterday Port Blair-dir 1sg=cl 4=reach-dst.pst
‘The day before yesterday I reached Port Blair.’

(84) khuro ɲo kotra-ak ci be

prox-loc house inside-dir come imp
‘Here, come inside the house’

(85) ŋu ɲo-ak ŋ=ut=cɔne- b-ɔm

2sg house-dir 2sg=cl 4=go- fa-npst
‘You are going towards home.’

5.2.2.10 Locational
The locative case is represented by -al ~ -il ~ -el ~ -l suffixed to specific
location words designating space. They are also suffixed to nouns which
specifically refer to the goal or source of the location. Some of these
words, as can be seen from the following table, are constituted by two
body class markers.

Table 5.2. Locative forms

English gloss Andamanese form Morphology
On tut=ara:=l cl 4=cl 6=loc
Inside k=ut=ara:=l obj=cl 4=cl 6=loc
Under tum-bo-l refl-heart-loc
Near cu-l near-loc
Away ta=cakhu-l cl 1=platform-loc
In front of tɛr=cokhe-l cl 2=face-loc
Behind (lit. back) ut=bɔ-l cl 4=back-loc
128 chapter five

This consists of a combination of orientation marking words and loca-

tive case markings. A combination of two body class markers along with
the locative suffix confirms the paradigm of a compound word as seen in
Chapter 3 on word formation processes. Some examples are given below.

(86) kitaːb ʈhica tuttaraːl jiyo

kitɑːb ʈhicɑ tuttɑrɑːl jiyo
book table on-loc be-exist
‘Book is on the table.’

(87) ʃup kuttaraː-l mino be

basket inside-loc potato cop
‘The potato is inside the basket.’

(88) o ʈɔŋ tumbol araːŋ karaːp iːt.talɛ

3sg tree under refl waist obj.take rest
‘He takes rest under the shadow of a tree.’

(89) ɖu boaː -l aka uno-o

3sg.dist.vis floor-loc 3sg sit-pst
‘He was sitting on the floor.’

The following sentences are without an overt copula, as in fast speech,

speakers tend to drop it.

(90) e=cul jiyo khider

cl 5=near be-exist coconut
‘There is a coconut nearby.’

(91) er=ŋolka biu-ta cɔkhe-l jiu (be)

cl 2=matchbox candle-abl face-loc be/exist (cop)
‘Matchbox is away from the (front of ) candle.’

In the following sentences, location markers such as ‘in front of ’, ‘behind’,

or ‘at the back of ’, are placed in front of the entire phrase as an adjunct
(92–94).

(92) tɛr=cɔkhe-l/tut-bɔ-l th=ut ŋyo ʈɔŋ jiyo

cl 1=face-loc/cl 4=behind-loc 1sg=cl 4=house tree exist
‘There is a tree in front of/behind my house.’

(93) boal ʈhaunobit jul ukuthuo

boɑ-l ʈh=ɑuno-bit jul uku thu-o
ground-loc 1sg=sit-nmlz soon 3sg reach-dst.pst
‘I was about to sit on the ground when he came.’
 nouns and noun phrases 129

(94) ʈhimikhu-l ʈɔilɔcɔŋ bi

ʈhimikhu-l ʈɔilɔcɔŋ bi
forest-loc deer cop
‘Deer are in the jungle.’

(95) tajiocɔr ʃare-l jiyu

tɑjiocɔr ʃɑre-l jiyo
fish sea-loc exist
‘Fish are in the sea’

(96) inobi tamrac kuttaral jiyo

ino-bi tɑmrac kuttɑrɑl jiyo
water-abs bucket inside-loc exist
‘Water is in the bucket.’

(97) mara-ʈhimikhul aʃiu-no ŋa-tɛkhok bɔ-phome

mɑrɑ=ʈhimikhu-l ɑʃiu-no ŋ=ɑ=tɛkho (i)k=bɔ pho-me
1pl.incl.cl 6.poss= who-pl 2sg=cl 1=language obj=understand
jungle-loc neg-hab
‘Nobody in our jungle (village) knows your language.’

(98) du tajeocƆr ɖiul taphaːyəm

du tɑjeocɔr ɖiul tɑ=phaːe-om
3sg.dist.vis fish sun-loc caus=dry-npst
‘She dries fish in the sun.’

(99) ʈh=e=cul ocƆ pho-be

1sg=cl 5=near net neg-cop
‘I do not have a net.’

(100) mia khriŋkoʃo-il ʈhi-bi ik=ɲo-k-Ɔm

1pl.excl Strait island-loc place-abs obj=live-fa-npst
‘We live in Strait Island.’

With interrogatives and indefinite pronouns which are identical in form,

i.e., ca:y, the locative suffix -l can follow the question word with metath-
esis of vowel and semivowel within the form (refer to §2.5.9). Thus ca:y-l
is rendered as cya:l ‘where.’

(101) cyaːl no=t=ɲo be

cyɑː-l ŋ=ot=ɲo be
what-loc 2sg-cl 4.poss=house cop
‘Where is your house?’

130 chapter five

(102) cyaːl ʈhoŋeolebo

cyɑ-l ʈho.ŋ=eule-b-o
what-loc 1sg.2sg=see-fa-dst.pst
‘I have seen you somewhere.’

Location will be further discussed in §8.5 on spatial deixis.

5.2.2.11 Genitive
PGA is a double marking language: the inalienable possession markings
are procliticised to the head noun, while alienable possession is repre-
sented by a dependent marking on the possessor noun. Possession has
been widely discussed in Chapter 3 and Chapter 6. We should recall that
there are in all, seven types of body class markers, serving the func-
tion of possession, which are attached to the head noun indicating mostly
inherent relationships between the R and the D.
As we saw in Chapter 3 PGA also has dependent marked possessor
nouns when dealing with the feature of alienable nouns. This implies that
the language has both systems: head-marked and dependent-marked; the
previous form is extensively used.
Forms that are not inalienable are marked by the genitive suffixes
-ico ~ -iʃo, which are suffixed to the possessor. These suffixes, in a real
sense, designate possessive relationships. Predictably, this is used only
with the animate possessors. Since the items listed as inalienable or alien-
able are culture-specific, the Great Andamanese have their own listing of
items as typical alienable nouns.2 These follow the possessor nouns with
the genitive -ico or -iʃo.

(103) ɖu-iʃo cƆkbi

3sg.dist.vis-gen turtle
‘His turtle.’

(104) ɖun-iʃo ko
3du.dist.vis-gen bow
‘Their bow.’

(105) ʈh-ico boa

1sg-gen land
‘My land.’

2 Hence, words designating ‘land’, ‘jungle’, ‘upper garments’, ‘lower garments’, ‘dog’,
‘friend’, ‘God’, as well as some kinship terms such as ‘son’ and ‘daughter’ are considered
alienable.
 nouns and noun phrases 131

(106) ʈh-ico ʈhimikhu

1sg-gen forest
‘My forest.’

(107) ʈh-ico a cao

1sg-gen arg dog
‘My dog.’

(108) ʈh-ico jo
1sg-gen song
‘My song.’

(109) ʈh-ico ʈƆʈo

1sg-gen dress
‘My dress.’

5.2.2.12 Comparative
Our database does not contain many examples of comparison. The marker
of comparison that sometimes appears is -ta, which is homophonous to
the ablative or instrumental and is suffixed to the noun.

(110) ʈhut=thu ta-i=pho ʈh=aːra=belo- ta-i=pho ʈh=erla-b-om

1sg.cl abl-cl 1sg=cl abl-cl 1sg=alone-fa-
4=born 5=neg 6=earlier 5=neg npst
‘There is no one either younger or older than me, I am alone.’
Lit. ‘No one was born later than me and no one is born earlier than me,
I am alone.’

For further discussion on comparatives please refer to §10.7 in the chapter

on syntax.

5.3 Noun Phrases

5.3.1 The General Structure

Noun phrases can be descriptive, possessive (alienable and inalienable),
appositional, quantifying, and relativised. The structure of the noun
phrase in PGA can be:

(S 5-1) pro/class=nom (case marking) (class=)adj (case marking)

132 chapter five

That is, a modified noun precedes the modifier. If the noun is obligatorily
marked by a body class marker then the following adjective is stripped of
any class marker.

5.3.1.1 Descriptive
Adjectives follow the head. In a descriptive noun phrase case markings
are suffixed to the modifier, i.e. to the whole NP (114) and not to the noun
alone.

(111) ɲo nƆl
house good
‘A good house.’

(112) phƆr lobuŋ-bi khulol jio

Bamboo long-abs there exist
‘There is a long bamboo.’

Adjectives may also be used predicatively, as in (113) and (114).

(113) surmai ɛr=buŋoi

Surmai cl 2=beautiful
‘Surmai is beautiful.’

(114) ɖulo tɛr=khudoi

moon cl 2=round
‘round moon’ / ‘The moon is round.’

5.3.1.2 Possessive
Alienable

(115) u-iʃo koʈhorɛmo

3sg-gen small bowl
‘His bowl.’

Inalienable

(116) ʈh=a= mimi

1s=cl 1.poss=mother
‘My mother.’

(117) ŋ=ot= bo
2s=cl 4.poss= heart
‘Your heart.’
 nouns and noun phrases 133

(118) cao tot= kƆbo e=cay

dog cl 4.poss= skin cl 5= bad
‘Dog’s bad/sick skin.’ or ‘Dog’s skin is not good.’

5.3.1.3 Appositional
An appositional phrase can be created when a non-finite verb followed by
a relative marker modifies the following noun.

(119) ʈoya ( j)ukhe a=thire ta-beno

stand rel/spec cl 1=child appl-sleep
‘The standing child (the one who was standing) had a sleep (went off to
sleep).’

Details on jukhe can be found out in §4.1 as well as in §5.3.1.5 given

below.

5.3.1.4 Quantifiers
As the language is in contact with Hindi we find parallel Hindi and PGA
structures where the modifier precedes, as well as follows, the modified.
The first sentence (120) is an exact Hindi translation, hence the use of the
Hindi numeral das ‘ten’, as well as the use of Hindi word order. These
kinds of constructions are becoming very popular in day-to-day speech.

(120) das ʈƆkho-tɛc-il

ten wood-leaf-loc
‘In ten rupees.’

(121) a=kaʈa ontoplƆ

cl 1=girl one
‘One girl.’

(122) [kitab tatɛrbui]-bi ʈob-o

[book two]-abs steal-dst.pst
‘Two books were stolen.’

5.3.1.5 Relativisation
Relativised constructions follow two different word orders. The most
common ones are those where the modified precedes the modifier (123),
a participial phrase which is formed by adding an additional suffix, -e,
indicating the state of the nominal in question (124). The other type of
construction, where the order of modifier and modified is reversed, i.e.,
the modified noun follows the modifier, has arisen due to contact with
Hindi (125).
134 chapter five

(123) ra tara=ulibi ikto(t)-kata-e

Pig cl 6=tail obj=cl 4=cut apiece-state
‘Pig with a cut tail’ or ‘the pig whose tail was cut.’

(124) ʈƆŋ-to ʈƆŋ-bi unɖu-e ek=tɛrʈƆ-e

tree-abl branch-abs break-state obj=throw-imp
‘Throw away the broken branch of a tree’ or
‘Throw away the branch broken from the tree.’

The relative marker jukhe is also used for specific reference of a nominal
category. This type of construction parallels Hindi specifier wala (Abbi
2001). The following sentence is an exact morpheme-to-morpheme trans-
lation of the Hindi sentence əccha wala ləɽka mera beʈa hɛ ‘the boy who
is good is my son.’

(125) nƆl jukhe ʈoʈa ʈh=ut=thire be

good rel boy 1sg=cl.poss 4=child cop
‘The boy who is good is my son.’

Relativisation is dealt with in detail in §10.8.

5.3.2 Coordinated Noun Phrases

Two nouns are coordinated by the optional use of a conjunction bo. In fast
speech this is dropped (127, 129, 130, 131).

(126) sare bo akajirabi tɛʃe

sare bo ɑkɑ-jirɑ-bi tɛʃ-e
salt and pepper-abs give-imp
‘Give salt and pepper.’

(127) torƆm kajira cƆkbithomo inol etɛʃe

torɔm kɑjirɑ cɔkbi thomo ino-l e=tɛʃ-e
salt chili turtle flesh water-loc cl 5=put in-imp
‘Put salt, chili and turtle flesh in water.’

(128) lico bo bea kona-bi it=ji-o

Licho and Bea tendu-abs obj=eat-dst.pst
‘Licho and Bea ate tendu fruit.’

When a coordinated subject is used with an intransitive verb, the con-

junction is generally omitted.
 nouns and noun phrases 135

(129) a=joe a=toŋ ɲyo-ak cone-b-om

cl 1=Jo cl 1=Tong house-dir go-fa-npst
‘Jo and Tong are going towards the house.’

The coordinated nouns can be marked for plurality by suffixing the plural
marker to the second noun in the linear order.

(130) a=joe a=toŋ-nu taracɔr-e eole-inci-k-o

cl 1=Jo cl 1=Tong-pl spring-abs see-go for-fa-dst.pst
‘Joe and Tong went to see the spring.’

The optional use of the conjunction can also be seen in phrases with a
comitative reading. Thus:

(131) ʈho abhishek ʈhi-bi ɲo-m

1sg Abhishek place-abs live-npst
‘I live with Abhishek at the same place.’

(132) ʈh-ico khider ʈƆŋ er=khuro boe=ʈɛlɛ pʰec be

1sg-gen coconut tree cl 2=big and cl 5=thick very cop
‘My coconut tree is big and very thick.’

5.4 Conclusion

To summarise, PGA is highly agglutinative in nature. Other than the plural

enclitic =n, the language displays a large variety of proclitics. There are
many case suffixes or case relators which are multifunctional in nature. As
we have noticed, the following case suffixes are multifunctional:

-be Absolutive, hence marks the S, and O

-ta Ablative, instrumental, comparative
-e Ergative, accusative

The following multifunctional verbal suffixes and proclitics are homopho-

nous to the nominal case suffixes. These are classic examples of ‘versatile
cases’ (Aikhenvald 2008: 565–603).3

-be ~ -e Completive, imperative, stative

ta= Applicative, causative

3 For the list of languages possessing this function see Aikhenvald (2008: 595).
136 chapter five

PGA offers evidence for the gradual loss of a large number of case mark-
ers ashere are alternating forms without any markings. The context of
conversation decides the pragmatics and semantics of the sentence. The
most interesting feature of PGA is that it denotes case relations on verbs
by attaching proclitics to verb roots. This feature makes overt marking on
nouns redundant. Please refer to Chapter 9.
 Chapter Six

Possession

Introduction

Possessive constructions are the most intriguing and interesting in the

present-day Great Andamanese language. The language has a wide variety
of possessives quite unlike any other Indian language. The distinctions
in various forms do not only depend upon the simple binary oppositions
of alienability/inalienability, but also on various diverse ethno-semantic
categories which define the relation between the possessor and the pos-
sessed nouns. Factors such as part-whole relationship, part-to-­component,
­intimate/non-intimate relations, human/non-human relations, and the
notion of possessum being part of the possessor, all play an important
role in deciding the appropriate inalienability marker (ina). The ina
marker that relates the possessor and the possessed is thus selected by the
semantic categorisation of the two nouns that it relates to. The concept of
inalienability and its grammaticalisation as possessive markers were dealt
with briefly in §3.3. Readers are advised to refer to Chapter 3. We will dis-
cuss the formation of possessive constructions in the following sections.
We start with the basic two divisions of alienable vs. inalienable for
lack of any other appropriate term describing such a variety of concepts
that play a role in describing the relationship between the possessor (R)
and the possessed (D) in the language, and then we will proceed further
on the basis of the concept of inalienability as already defined in §3.3.

6.1 Possessive Classification

As discussed in Chapter 3, one major deciding criterion that determines

the choice of a particular possessive marker is the partonomy of the
body. Under various possessive forms lies the semantic typology of cat-
egorisation of the human body parts and divisions.
This implies that no body part term can be rendered without a body divi-
sion class marker. This unique feature of the language is the division itself
within the body parts each being symbolised by a distinct class marker.
138 chapter six

Inalienability in the languages of the world studied so far maintains a

two-way distinction in dividing nouns into inalienable and ­alienable, the
­former comprising of one or all six areas mentioned in §3.3.3. PGA, as
stated earlier, maintains seven divisions within the terminology of the
body parts. It then further extends the notion of seven divisions (Table 6.1)
to the rest of the terms, i.e. kinship terms, spatial relational terms, closely
related object terms, and furthermore to attributes (human characteris-
tics), objects, actions and states (denoted by verbs) with reference to the
body. However, the biggest function of these markers is classificatory. In
this function they are lexically determined by the D and result in distinct
overt markings of possession.
The Great Andamanese perceive the following divisions of the body
and symbolise each by a monosyllabic or disyllabic morpheme, a proclitic
which is attached to the appropriate term for the body part, the head
noun, D on its left.
We shall first describe the various possessive markings used for posses-
sive constructions relating to body division/part terms and then the ones
used for the kinship terms. These possessive markings are symbolised
by body division class markers depicting inalienability. This will be fol-
lowed by the processes to derive possessive constructions. Finally, we will
discuss other nouns which take part in possessive constructions. While
inter-glossing, we will be using the specific number of the body division
possessive class markers by an abbreviated symbol cl followed by the
appropriate number so that readers can immediately associate the type or
class of the grammatical category that is being discussed. Thus cl 2.poss
will mean possessive class marker 2 which has been derived from a partic-
ular body division term. We will optionally specify the morpho-syntactic
function of this particular class marker.

6.2 Primary Possession

PGA distinguishes between primary and secondary possession. The pri-

mary possessions refer to those which are derived from the basic seven
divisions of the human body part terms.
The typical structure of a noun phrase with body part terminology is:

(S 6. 1) Possessor pronominal clitic/Noun_body division class marker

_dependent noun
 possession 139

The structure can be abbreviated as:

(S 6.2) R = class marker = D

An R is a possessor which, in this case, is a pronominal clitic or a proper

noun followed by an appropriate possessive class marker attached to
the dependent noun D, which can be a body division/part term or other
inalienable noun. They have an additional function in the grammar and
that is a classificatory one. The body division class markers classify
nouns into various classes, each symbolised by a class marker (class)
proclitic. This is true of all obligatorily possessed nouns, even those
which are not prototypical inalienable nouns. These are considered in
§3.3 and in §3.4. Let us first consider the simple possession of various
body parts.

(1) ʈh=ot=bo
1sg=cl 4.poss=back
‘My back.’

In (1) the first singular pronominal clitic ʈh= is followed by the appropriate
class marker, viz. class 4, ot= which is selected by the body part term bo
‘back.’ The clitic is attached to the left of the possessed noun, i.e., ot=bo
and not to the possessor noun. However, as seen in Chapter 3, the pro-
nominal clitic has the tendency to attract the possessive proclitic towards
it and hence may appear as if it is attached to the possessor and forms one
word (refer to §2.5.8 and §3.8 for details).

6.2.1 Body Part Terms

The body class markers, as stated earlier, perform the twin functions of
(a) possession and (b) classification. Here, Table 3.1 is presented again as
6.1 for easy reference.
We shall now examine each possessive class marker defining the
various body divisions which encompass several body parts. During the
description we will make occasional mention of the kinship terms which
are represented by the same class markers. One should recall that these
class markers operate as proclitics, as described in Chapter 3. Most of the
examples given below represent clitic sequencing.
140 chapter six

Table 6.1. Seven basic zones in the partonomy of the body

Classes Partonomy of the human body body division
class markers
1 mouth and its semantic extension a=
2 major external body parts εr=
3 extreme ends of the body like toes and oŋ=
­fingernails
4 bodily products and part-whole relationship ut=
5 organs inside the body e=
6 parts designating round shape/sexual organs ara=
7 parts for legs and related terms o= ~ ɔ=

6.2.1.1 Mouth Cavity (a=)

When the possessed entity or D is the mouth and its extensions (e.g.
‘tongue’ and ‘throat’) as well as one of the primary kin relationships like
‘mother’, ‘father’, ‘grandmother’, ‘grandfather’, this proclitic is attached to
the noun denoting a body part or kinship term.

(2) ʈh=a=tat
1sg=cl 1.poss=tongue
‘My tongue.’

(3) ʈh=a=foŋ
1sg=cl 1.poss=cavity
‘My mouth.’

(4) ʈh=a=kεr
1sg=cl 1.poss=throat
‘My throat.’

(5) ʈh=a=may
1sg=cl 1.poss=father
‘My father.’

(6) ʈh=a=mimi
1sg-cl 1.poss=mother
‘My mother.’

Table 6.2. presents some examples of the body part terms attached
to a=.
 possession 141

Table 6.2. Body part terms with a= possessive class marker

Present-day Great English gloss English translation
­Andamanese
ʈh=a=phoŋ 1sg=class 1=cavity My mouth
ʈh=a=ʈεiŋ 1sg=class 1=liquid My saliva
ʈh=a=phup 1sg=class 1=excretion My sputum
ʈh=a=lae 1sg=class 1=surface My palate
ʈh=a=tat 1sg=class 1=tongue My tongue
ʈh=a=ker 1sg=class 1=neck My throat/neck
ʈh=a=cɔkhɔ 1sg=class 1=face My area around face

6.2.1.2 Major External Body Parts (εr= ~ er=)

A large number of bound nouns are included in this class and these des-
ignate major body parts that pertain to the ‘head’, ‘brain’, ‘neck’, ‘face’,
‘arms’, ‘thigh’, ‘calf’, ‘knee’ and ‘bones.’ Refer to Table 6.3. This class is also
used with reference to certain kin relations such as ‘spouse’ (9).

(7) ʈh= εr=co

1sg=cl 2.poss=head
‘My head.’

(8) ʈh=εr=ʈɔe
1sg=cl 2.poss=bone (calf)
‘My bone.’

(9) ʈh=εr=boi
1sg=cl 2.poss=spouse
‘My spouse.’

Table 6.3. gives noun forms attached with εr= or er= class marker
­proclitic.

Table 6.3. Body part terms with er=, εr= possessive class marker
Present-day Great English gloss English translation
­Andamanese
ʈh=εr=co 1sg=class 2=seed/head My head
ʈh=er=kɔbɔ 1sg=class 2=skin My scalp, skin
ʈh=er=mine 1sg=class 2=brain My brain
ʈh=er=beŋ 1sg=class 2=forehead My forehead
ʈh=er=buo 1sg=class 2=ear My ear
142 chapter six

Table 6.3 (cont.)

Present-day Great English gloss English translation
­Andamanese
ʈh=er=jili 1sg=class 2=flower My area above eyebrow
ʈh=er=ulu 1sg=class 2=eye My eyes
ʈh=er=kɔʈho 1sg=class 2=nose/trunk My nose
ʈh=er=kɔʈho-ʈɔː 1sg=class 2=nose-bone My sinew
ʈh=er=tap 1sg=class 2=chin My lower jaw/chin
ʈh=er=tap-bεc 1sg=class 2=chin-hair My beard
ʈh=εr=nɔkho 1sg=class 2=cheeks My cheeks
ʈh=er=phile 1sg=class 2=teeth My teeth
ʈh=εr=jukhu 1sg=class 2=space above My area between upper
upper lip lip and nostrils
ʈh=er=boa 1sg=class 2=land My lips
ʈh=er=khum 1sg=class 2=side My shoulder’s edge
ʈh=εr=bala 1sg=class 2=arms My arms
ʈh=er=khit 1sg=class 2=biceps My biceps
ʈh=εr=ʈɔŋ 1sg=class 2=branch My forearm
ʈh=er=me-tei 1sg=class 2=­mother-liquid My breast
ʈh=er=lɔ 1sg=class 2=mole My mole
ʈh=er=layu 1sg=class 2=wrinkle My wrinkle
ʈh=er=belɔe 1sg=class 2=pimple My pimple

6.2.1.3 Extremities of the Body (uŋ= ~ oŋ=)

When the possessed entity is any part of the hand or arm, e.g. ‘finger’,
‘palm’, ‘wrist’, ‘nail’, or other extremity, the class marker uŋ= ~ oŋ= ~ ɔŋ=
is attached to the D. This marker does not attach to any of the kin terms,
as it is exclusively used for the parts of the body.

(10) ʈh=ɔŋ=korɔ
1sg=cl 3.poss=palm
‘My palm.’

(11) ʈh=uŋ=kaːra
1sg=cl 3.poss=nails
‘My nails.’

(12) ʈh=uŋ=kenap
1sg=cl 3.poss=finger
‘My finger.’

Table 6.4 gives other noun forms with uŋ= or oŋ= body division class
marker.
 possession 143

Table 6.4. Body part terms with oŋ= possessive class marker
Present-day Great English gloss English translation
­Andamanese
ʈh=oŋ=kenap 1sg=class 3=finger My fingers
ʈh=oŋ=kara 1sg=class 3=nails My nails
ʈh=oŋ=kɔrɔ 1sg=class 3=hand My palm, hand
ʈh=oŋ=ʈɔː 1sg=class 3=bone My wrist bone
ʈh=oŋ=kenap-cɔkhɔ 1sg=class 3=finger-face My thumb
ʈh=oŋ=kɔrɔ-tot=bɔ 1sg=class 3=hand=class 4=back My back of the hand
ʈh=oŋ=phoŋ 1sg=class 3=cavity My armpit

6.2.1.4 External Body Products or Extension (ɔt= ~ ut= ~ ot=)

This set of class markers attaches to the D for entities which include, (a)
those that can be considered to show part-to-whole or part-to-component
relationships, (b) the body parts forming the torso, such as the ‘chest’,
‘back’ and ‘heart’, (c) bodily products such as ‘hair’, ‘life’, ‘sweat’, and
‘breath’ and, (e) the kinship term for ‘child.’ It is interesting to note that
while ‘child’ is considered inalienable, ‘daughter’ and ‘son’ are not.

(13) ʈh=ut=bec
1sg=cl 4.poss=hair
‘My hair.’

(14) ŋ=ut=khirme
2sg=cl 4.poss=sweat
‘Your sweat.’

(15) lico ut=thire

Licho cl 4.poss=child
‘Licho’s child.’

(16) ʈh=ut=ʈhi
1sg=cl 4.poss=breath
‘My breath.’

Table 6.5 presents noun forms with class marker 4.

144 chapter six

Table 6.5. Body part terms with ot=, ɔt=, ut= possessive class marker
Present Great English gloss English translation
­Andamanese
ʈh=ot=bεc 1sg=class 4=hair My hair
ʈh=ot=ʈeŋ 1sg=class 4=branch My nape of neck
ʈh=ot=loŋɔ 1sg=class 4=lower part My lower part of neck
ʈh=ot=ʈɔː 1sg=class 4=bone My neck bone
ʈh=ut=khum 1sg=class 4=shoulder My shoulder
ʈh=ot=bɔ 1sg=class 4=back/heart My back (upper)
ʈh=ot=car 1sg=class 4=chest My chest
ʈh=ot=kɔrno 1sg=class 4=lungs My lungs
ʈh=ut=bo-it=dello 1sg=class 4=heart-obj=ball My heart
ʈh=ot=co-to=bat 1sg=class 4=seed=class 4=night My nipple
ʈh=ot=kɔbɔ 1sg=class 4=skin My skin
ʈh=ut=khirme 1sg=class 4=hot/heat My sweat

6.2.1.5 Internal organs (e= ~ i-)

The body division class marker 5 attaches to terms which pertain to enti-
ties inside the body. These include: ‘blood’, ‘ribs’, ‘liver’, ‘covering around
intestines’, ‘hip bone’, ‘belly/stomach’ and ‘bile.’ In other words, the con-
cerned entities are invisible body parts, mostly inside the stomach and
abdomen. However, words for ‘knee’ and ‘thigh’ are also expressed by this
class marker, whereas ‘heart’ and ‘lungs’ belong to class 4. Such anoma-
lies are there and difficult to explain as the cognitive world of Great Anda-
manese and its perception that determines these classes are very different
from ours. As the language is endangered and speakers are very few it was
impossible to comprehend the world view of the community fully and
comprehensively. Refer to Table 6.6.

(17) cao e=tei

Dog cl 5.poss=blood
‘Dog’s blood.’

(18) ʈh=e=sudu
1sg=cl 5.poss=intestine
‘My intestines.’

(19) ʈh=e=teɖu
1sg-=cl 5.poss=pancreas
‘My pancreas.’
 possession 145

Table 6.6. Body part terms with e=, i= possessive class marker
Present-day Great English gloss English translation
Andamanese
ʈh=e=tei 1sg=class 5=liquid My blood
ʈh=e=burɔŋo ʈɔː 1sg=class 5=fruit-bone My ribs
ʈh=e=philu 1sg=class 5=belly My stomach
ʈh=e=philu-phet 1sg=class 5=belly-big My belly
ʈh=i=ŋet 1sg=class 5=breathe My navel
ʈh=e=sudu 1sg=class 5=intestines My intestines
ʈh=e=baene 1sg=class 5=covering My covering around intestines
ʈh=e=bi-ʈɔlɔn 1sg=class 5=obj-flower My kidney
ʈh=e=meca 1sg=class 5=liver My liver
ʈh=e=teɖu 1sg=class 5=pancreas My pancreas
ʈh=e=bucɔ 1sg=class 5=lap/mangrove My lap
ʈh=e=cɔrɔkh 1sg=class 5=joint My knee
meŋe=i=ʃoŋo 1pl=class 5=body Our bodies

6.2.1.6 Nodular or Curved Structure (ara=, ra=)

These class markers of possession are placed before the possessed entity,
which is a circular and curved structure, such as ‘cheeks’, ‘bladder’, ‘scro-
tum’, ‘heel’, etc. Surprisingly, the words for ‘knee’ or ‘head’ are not in this
list. The terms for ‘sides’ or ‘curvature of the body’ may also be attached to
this class marker. These class markers signify a type of a relational posses-
sion, and are also used for indicating younger sibling relations.

(20) ʈh=ara=sulu-thu ʈɔʈa/kaʈa

1sg=cl 6.poss=after-born boy/girl
‘My younger brother/sister.’

(21) ŋ=ara=karap
2sg=cl 6.poss=rib cage
‘Your rib cage.’

(22) ŋ=ara=phu
2sg=cl 6.poss=stool
‘Your stool.’

Table 6.7 provides some of the noun forms with ara= noun class marker 6.
146 chapter six

Table 6.7. Body part terms with ara= possessive class marker
Present-day Great English gloss English translation
­Andamanese
ʈh=ara=ʈɔlɔ 1sg=class 6=flower My large intestine
ʈh=ara=karap 1sg=class 6=lower back My waist, lower back
ʈh=ara=thɔmo 1sg=class 6=fat/flesh My buttocks
ʈh=ara=karap-thomo 1sg=class 6=lower back-flesh My groin
ʈh=ara=karap-jiriŋe 1sg=class 6=lower back-?? My pelvis
ʈh=ara=ʈεt 1sg=class 6=anus My anus
ʈh=ara=ɖiletmo 1sg=class 6=ball small My bladder
ʈh=ara=ɖomo 1sg=class 6=testicles My testicles
ʈh=ara=ɖomo-tot=kɔbɔ 1sg=class 6=testicles=class My scrotum
5=skin
ʈh=ara=ili 1sg=class 6=urine My urine
aka-ara=phu 3sg-class 6=excretion His stool

6.2.1.7 Lower Parts of Body (o= ~ Ɔ=)

The class marker 7 classifies those body parts which are visible and refers
to lower parts of the body, such as ‘leg’, ‘toe’, ‘sole’, ‘heel’, etc. Refer to
Table 6.8.

(23) ʈh=o=mɔʈɔ
1sg=cl 7.poss=leg
‘My leg.’

(24) ʈh=o=roŋo
1sg=cl 7.poss=ankle
‘My ankle.’

(25) ʈh=o=mɔʈɔ-ʈo=mikhu
1sg=cl 7.poss=leg=cl 7.poss=centre
‘My sole.’

Class marker 7 classifies relations such as o=toni ‘son-in-law’ or ‘younger

sister’s husband.’
There are some anomalous class markings in the words for ‘urine’,
‘stool’ and ‘pubic hair’ which cannot be explained. Similarly, the word
for ‘armpit’ could have belonged to class 6 marked by ara= or the word
for ‘lungs’ could be marked by cl 5 e= to show internal organs. Or the
words for ‘knee’ and ‘head’ could have been preceded by body division
class marker reserved for ‘curvature.’ These are some of the unexplain-
able areas. These anomalies occur also because it is difficult to under-
 possession 147

Table 6.8. Body part terms with o=, ɔ= possessive class marker
Present-day Great English gloss English translation
­Andamanese
ʈh=ɔ=mɔʈɔ 1sg=class 7=leg My leg
ʈh=ɔ=mɔʈɔ-tu=jukhu 1sg=class 7=leg-class 4=extension My toe
ʈh=o=mɔʈɔ-to=mikhu 1sg=class 7=leg-class 4=centre My sole
ʈh=o=mɔʈɔ-tara=ɖole 1sg=class 7=leg-class 6=ball My heel
ʈh=ɔ=mɔʈɔ-ʈɔː 1sg=class 7=leg-bone My bone below
knee
ʈh=o=roŋo 1sg=class 7=ankle My ankle
ʈh=o=ʈɔnno 1sg=class 7=semen My semen

stand the clear-cut basis of the seven divisions. What is being presented
here is based on limited field data as language is highly endangered and
moribund. The nature of the language being a koiné has aggravated the
problem. No speaker could explain the justification for such divisions. All
we can say is that (1) there are, in general, seven primary divisions that
help the Great Andamanese to conceptually perceive the human body
and in turn, other associated relationships, and (2) it can be inferred
from the forms given in these tables that the body part terminology in
PGA represents the relationship of body parts to a ‘person’ or ‘self’ and
not as a part-whole relationship. This is reflected in the majority of sim-
plex morphemes.
In general, it can be argued that inalienable possession of anatomical
terms is the basic semantic relation of this domain. The possessor in this
context is the human being and not the body. Thus, ‘my leg’ or ‘his head’
are more readily elicitable expressions in PGA than say, ‘the leg is part
of my body.’ In fact, this is the only way body parts in this language are
expressed; the body parts are obligatorily ‘possessed.’
All of the primary body parts are possessed by the self and thereby
have the obligatory pronominal clitic or a noun as the possessor. It can be
argued in this light that there are certain body parts that are less salient
than others and hence are treated as second order body parts (see §6.5).
They derive their names either by means of descriptive terms or by jux-
taposing two primary part names, whereas the primary body parts are
primary lexemes used with a body division class marker with refer-
ence to the possessor. The choice of the class marker in the case of
the first order body parts is semantic in nature and varies according to the
perceptual division of the entire human body by the Great Andamanese.
In this sense the clustering of body parts into divisions/areas of the body
148 chapter six

and the consideration of each division/area as an inalienable entity is a

culture-specific phenomenon.
The significant finding about the cognitive world of the Great Anda-
manese is that the Great Andamanese people hold the anthromorphic
view of the world and thus relate the universe to the human body and
its various divisions. The grammar of the language contains the duality
of the system where there are concrete words for each body part in con-
junction with the abstract symbols for divisions of human body that con-
tain these parts.

6.2.2 Is There a Hierarchy?

All seven categories of body parts are directly possessed by a ‘being’ or
the ‘self.’ Now the question is: should we classify all of them on the same
level or consider them to be on varying levels of a hierarchy according to
the frequency parameter? It is tempting to consider the latter option. A
study made by Avtans (2006: 97) reports some interesting statistics about
the frequency of use of these body class markers as given in Table 6.9. This
implies that body class marker er= or εr= is most commonly used. This
has guided us to frame the hierarchical scale of the various body division
classes in the language.
The last two proclitics, o= and oŋ=, listed in Table 6.9, not only occur
less frequently as far as references to the body part terms are concerned,
but also, have a limited distribution across the lexicon of the language.
The potential of these becoming grammaticalised is low when compared
to the other five class markers. As we proceed, this fact will become
clearer. Hence, on a hierarchical scale of frequency, o= and oŋ= occupy
the lower ends of the scale.
We thus posit that there is a cline of grammaticalisation of the body
division class markers in the language.

Table 6.9. Frequency of occurrence of body division class markers with body
part terminology
Noun Class Body division class markers Frequency of use
class 2 er=, εr= 36%
class 5 e=, i= 17%
class 4 ot=, ɔt=, ut= 13%
class 6 ara= 12%
class 1 a= 9%
class 7 o= ɔ= 7%
class 3 oŋ=, uŋ= 6%
 possession 149

6.3 Kinship Terms

Only five out of the seven class markers are used to represent kinship
terms. It is challenging to discover why only these are reserved for kin-
ship terms while others are rejected. They are: ut=, εr=, a=, o=, and ara=.
An additional factor to be considered is that ‘higher (generation)’ kin are
designated by ‘mouth cavity and related body parts.’
With the exception of the terms for ‘son’ and ‘daughter’, which are
marked alienable genitive suffixes, all other terms are attached with
inalienable markers1 symbolised by body division class markers. While
speakers were able to provide kinship terms through direct elicitation,
they were not used in natural language. Elicitation of these terms was
not easy and this explains the significant variation in the terms elicited
from different speakers. Most, if not all of the speakers, use Hindi terms
now. Table 6.10 presents a variety of class markers attached to terms for
kinship relations.

Table 6.10. Kinship terms

Kinship terms English gloss
ʈh=ico a=kaʈa My daughter
ʈh=ico a=ʈoʈa My son
ʈh=o=toni My son-in-law / younger sister’s husband
ʈh=ε=toa thu-e My elder sibling
ʈh=ɔt=toa thu-e-akaoi My elder sister
ʈh=ε=toa thu ʈoʈʈa My elder brother
ʈh=a-=mai ra=tob Grandfather-like elderly person
ʈh=a=mai ka= ʈh=a=mai Grandfather
ʈh=a=mai ka= ʈh=a=mimi Grandmother
aka=maya (someone’s) late (deceased) old person, used as a
title
ara=lepha ka Widow or widower
ʈh=e=bɔe ~ boi My wife or husband (spouse)
korɔm olε bik Wife of the first man
ŋ=ara:=bε:loka Your wife’s younger brother
ŋ=a=mai exe Your wife’s father ‘father-in-law’
ŋ=a=mimi exe Your wife’s mother ‘mother-in-law’

1 However, the nouns for ‘son’ and ‘daughter’ were marked by ‘possessive prefixes’ in
the extinct South Andaman languages such as Âkà-Bêa. Man (1923: 158–159) cites examples
of kinship terms including the ones used for ‘son’ and ‘daughter’ that are preceded by
possessive prefixes.
150 chapter six

Table 6.10 (cont.)

Kinship terms English gloss
ʈh=ara=sulu thuo My younger sibling
ʈh=ara=sulu thu ʈoʈʈa My younger brother
ʈh=ara=sulu thu e=kata My younger sister
ʈh=a=mai My father
ʈh=a=mimi My mother

6.3.1 Parallels Between the Body Part Terminology

and Kinship Terms
Five body division class markers are used for kinship terminology that
exhibit cognitive parallelism between the inalienable body divisions and
a specific kinship relation terminology. Table 6.11 summarises this parallel
relationship.

6.4 The Twin Levels

As mentioned in §3.3, as well as in the introduction to Chapter 6 pos-

sessives function at twin levels in PGA: the primary and the secondary.
This merits some discussion. The primary ones are used with reference
to the self, which denotes the major body parts and the main kinship
terms. The secondary possessives are used for denoting those body parts
that are derived from the major parts and those kinship terms that are
descriptive, e.g. the ones used for siblings. In addition, they signify part-
to-whole relationships between non-human nouns. Thus, the forms dis-
cussed in §6.5 below fall into the class of secondary possession. In a
simplified way, one can state that the secondary divisions are added to
the basic ones to represent the extension of the body parts discussed
above. Thus, the ­language uses double markings to refer to words for
Table 6.11. Parallels between body parts and kinship terms
Class markers Body parts Kin terms
εr= major body parts spouse
a= mouth cavity parents
ut= extensions of body child
parts/body products
ara= + ut= nodular structure younger/older siblings
o= Lower body son in law/husband of younger sister
 possession 151

‘eyelashes’, ‘tears’, etc. The kinship terms, perceived as being of a sec-

ondary nature, are similarly marked. We shall now discuss the formation
of these constructions.

6.5 Secondary Possession

The possessives function on twin levels in PGA (Som 2006), namely the
primary and the secondary. Those with a primary level of function are
used with reference to the ‘self’ and denote the major body divisions/
areas and the main kinship terms that were considered above. Those with
a secondary level of possession are used for denoting those body parts
that are extension of the major parts, e.g. ‘eyelashes’, and those kinship
terms that are descriptive, e.g. the ones used for siblings. These are added
to the basic ones. The language uses double markings to refer to words
for ‘eyelashes’, ‘tears’, etc. and those kinship terms which are perceived
as of a secondary nature. We shall now briefly discuss the formation of
these constructions. It can be argued that there are certain body parts that
are less salient than others and hence are treated as second order body
parts. They derive their names either by means of descriptive terms or by
juxtaposing two primary part names, whereas the primary body parts are
primary lexemes used with an appropriate class marker with reference to
the possessor.
In addition to attaching the various class markers to the possessed
nouns, the language offers two more strategies to derive inalienable
constructions: (1) compounding or juxtaposition of two nouns and,
(2) syntactic derivation where a combination of more than two devices
is used.

6.5.1 Juxtaposition/Compounds
The dependent and the head nouns are juxtaposed (in that order) to form
a compound. However, it should be noted that the head and the depen-
dent noun do not stand in a ‘possessive’ relationship, but a descriptive
one where the dependent noun indicates the type of entity that is being
referred to by the head noun. In the examples given below the second
noun, i.e. the head, designates a generic entity while the first noun, which
is a dependent noun, indicates the type or class to which the designated
entity belongs.
152 chapter six

PGA Literal translation

(26) cokbi thomu turtle meat
(27) cokbi mulu turtle egg
(28) mɔcɔ mulu hen egg
(29) khidεr ʈɔŋ coconut tree
(30) khidεr ino coconut water
(31) ʈɔkho tei tree blood (gum)
(32) ra thire pig children (piglets)

It should be recalled that the ordering of the modifier and the modified
in compounds is the reverse of the order in adjective phrases where the
modifier follows the modified, i.e., the head. For example, cokbi nɔl ‘turtle
good’ and cokbi cae ‘turtle bad.’ Since the genitive precedes the possessed
noun in PGA (§3.1.2 and §6.7.1 below) and the order of the constituents in
possessive constructions is dependent followed by head (§6.2 above), one
cannot rule out the possibility that the typical structure of dependent and
head nouns with intervening class markers gave way to juxtaposition or
compounding in the order dependent-head while losing the class marker
in this journey of evolution. Compounding devices appear to be used as a
defining category rather than indicating the relationship of the possessor
and the possessed.

6.5.2 Complex Structures

Non-basic body part terms are derived by several morphological processes.
They are described below.

6.5.2.1 Lexical Compounding with Class Markers

This involves twin layers of possessive strategies, i.e. the first layer con-
sists of the compounding of two terms for body parts. The second layer
attaches the appropriate body division class marker along with the pos-
sessor pronominal clitic/noun to the compound noun.

(33) [ʈh=εr=[tap bec]]

1sg=cl 2.poss=[chin hair]
‘My beard.’

(34) [ʈh=εr=[jukhu bec]]

1sg=cl 2.poss=[above upper lip hair]
‘My moustache.’
 possession 153

(35) [ʈh=ara=[karap ʈↄ]]

1sg=cl 2.poss=[waist bone]
‘My waist bone.’

This implies that the head of the possessive construction is the leftmost
constituent of the compound (given in innermost brackets) because it
governs the type of the appropriate class marker. It is tap ‘chin’ and jukhu
‘the area above lip’ that dictate the presence of the class 2 εr= while the
word karap ‘waist’ governs the occurrence of the class 6 ara=. It is an
interesting finding as the semantic head in these constructions is the
rightmost nominal. Kinship terms such as ‘father-in-law’ or ‘mother-in-
law’ are also formed by this process.

6.5.2.2 Double Marking and Clitic Sequencing

There appear to be two hierarchical levels of syntactic organization of the
possessive constructions. First, the primary body part is possessed by the
self, where the self is the dependent noun and the specific body part is
the head noun. Then at the second level of derivation, the same head noun,
along with the previously dependent noun, functions as the dependent noun
to which the secondary body part is added by means of a secondary posses-
sive marker. This secondary marker is a grammaticalised form of the word
thu ‘born of.’ For instance, different parts and sub-parts of the eye are not
simply juxtaposed to the term for ‘eye’, rather, the nouns in this category
which are subordinate to the ‘part’ eye needs the obligatory use of the suffix
-thu which means literally ‘born of’ and has been grammaticalised to serve
the function of possession. This has been symbolised in small capitals.

(36) ʈh=εr=ulu thu ino

1sg=cl 2.poss=eye born water
‘My tears.’

(37) ʈh=εr=ulu thu bec

1sg=cl 2.poss=eye born hair
‘My eyelashes.’

(38) lico εr= ulu thu bɔːk

Lico cl 2.poss=eye born behind
‘Licho’s eyelids.’

6.5.2.3 Adverbial Function

Symmetric kinship relationships, such as those that represent sibling rela-
tions, are derived by using an adverbial phrase ut=toa thu ‘born before’
154 chapter six

or ara=sulu thu ‘born after’, which describes the temporal relationship

between the dependent noun, i.e. the possessor and the head noun, i.e.
the possessed. Thus, the adverbial phrase ‘boy born after me’ signifies
‘younger brother’, while ‘X born before me’ signifies ‘elder sibling.’

(39) ʈh=ut=toa thu kaʈa

1sg=cl 4=before born girl
‘My elder sister.’ (Literally: ‘Before me born girl.’)

(40) ʈh=ara=sulu thu ʈɔʈa

1sg=cl 6=after born boy
‘My younger brother.’ (Literally: ‘After me born boy.’)

It is interesting to note that two different class markers (i.e., ut= and
ara=) are chosen for designating elder and younger sibling relationships
although both of them describe temporality.
A combination of double marking complex structures and adverbial
use-constructions can be presented in a tabulated form. The lexeme thu is
grammaticalised and serves as the second possessive marker, symbolised
by small caps (born.poss), the first being represented by the specific body
division class marker. hence the construction here is:

(S 6.3) R class marker D born D

[poss I] [poss II]

Table 6.12. Double marking and secondary possession

Double marking Gloss
(41) ʈh=ut=toa thu ʈɔʈa/kaʈa my elder brother/sister
1sg=cl 4.poss i before born.poss ii (the one born before me)
boy/girl

(42) ŋ=εr=ulu thu b:ɔk your eyelids

2sg=cl 2.poss i =eye born.poss ii behind (the space born behind the eyes)

(43) ŋ=εr=ulu thu be:c your eyelashes

2sg=cl 2.poss i =eye born.poss ii hair (hair born of eyes)

(44) u=εr=ulu thu ino his tears

3sg=cl 2.poss i =eye born.poss ii water (water born of eyes)

(45) ʈh=ara=sulu thu ʈɔʈa/kaʈa my younger brother/sister

1sg=cl 6.poss i=after born.poss ii boy/girl (the one born after me)
 possession 155

Secondary possession indicates that the leftmost nominal of body part

terms determines the nature of the body class proclitic and is also the
possessor of the second body part term which is an extension of the for-
mer in physical sense, e.g., ‘eyelid’ is an extension of ‘eye’ and ‘beard’ is an
extension of ‘chin’ etc.

6.5.2.4 Double Class Markers

Double marking possessives can also be derived by employing two differ-
ent class markers in the same NP, without using the adverbial phrase
and the grammaticalised –thu ‘born.’ Consider the following where pos-
sessive markings are progressively decided by the head noun. The struc-
ture can be represented as:

(S 6.4) R-poss-D
{

R-poss-D
That is:
(S 6.5) [[R [R poss.1=DR] poss.2=D]]

(46) ŋ=er=phile tara=phoŋ

2sg=cl 2=teeth cl 6=cavity
‘Your dental cavity.’

(47) ʈh=er=ulu tut=ʈɔlɔtmo

1sg=cl 2=eye cl 4=white
‘The white of my eye (sclera).’

(48) ʈh=er=jili tot=bεc

1sg=cl 2=bone above eye cl 4=hair
‘My eyebrows.’

(49) ʈh=er=phile tara=tharale

1sg=cl 2=teeth cl 6=on (deixis of contact)
‘My gums.’

(50) ʈh=er=kɔʈho tara=phoŋ

1sg=cl 2=nose cl 6=cavities
‘My nostrils.’

(51) ʈh=o=mɔʈɔ tut=jukhu

1sg=cl 7=leg cl 4=end
‘My toes.’
156 chapter six

For an explanation as to why the poss tut= is used instead of ut=, or tara=
is used instead of ara=, refer to §6.6.

6.5.2.5 Summary
To summarise the discussion so far, one can claim that there are twin
levels of possession functioning in PGA: the primary and the secondary.
The primary level is used with reference to the self, which denotes the
divisions/areas of the body and the main kinship terms. The secondary
level of possession is used for denoting those body parts that are derived
from the major seven divisions/areas and those kinship terms that are
relational in nature, e.g. the ones used for siblings. Cliticisation, juxtaposi-
tion/compounding and syntactic derivation are three processes that are
employed in relating the possessor and the possessed nominals.
These different processes in combination with the seven body division
class markers discussed above result in as many as twelve different forms
of possessives. These are selected on the basis of ethno-semantic divisions
in the language. These are given in Table 6.15. Although some languages are
known to offer multiple grammatical devices to designate body part terms
(cf. Ndjebbana or Kunibidijig, a non-Pama-Nyungan ­language as described
in McKay 1996), PGA appears to be unique in its own devices.
Despite the fact that there was a substantial amount of variation of
forms noted among speakers, it was found that all speakers were consis-
tent in using the body division class markers with body part terms and
kinship terms. The available choice in class markers and their associated
D (for example, I have no idea why ‘head’ and ‘arm’ are expressed by the
same class marker) do not give a very coherent semantic and cognitive
explanation as of yet. This could be because the system is very ancient
and has acquired some irregularities over time. Or, as the language is
a koiné as well as being of a ‘mixed’ nature, different language systems
merged creating some inconsistencies. As mentioned in Chapter 1, speak-
ers of PGA are drawn from mutually intelligible different speech varieties.
Their common language, which is used as a lingua franca, bears traces of
their respective pre-existing dialects.

6.6 Animate vs. Inanimate Possessor and the

Semantics of Inalienability

6.6.1 Inanimate
As mentioned briefly in Chapter 3, the fundamental division of animacy
plays an important role in deciding the phonetic shape of the base form
 possession 157

of the class marker. If the possessor noun is inanimate, the class marker
is prefixed by a dental consonant t-, otherwise with all animate possess-
ors, both human and non-human, class markers begin with a vowel. Thus,
possessive class markers ara=, ot=, etc., which are indicators of animate
possessors, will be rendered as tara=, tot= respectively, if the possessors
are inanimate beings. Thus, live animals and their body parts will be
marked by a class marker without the initial t- sound as is the case with
human body parts. However, when the part is cut and segregated from
the body the associated marker will be prefixed with t‑. This entails that
an intact body part belongs to one particular class and a detached one
is treated differently, but still belongs to the same class. Consider the
following in bold:

(52) ra εr=co ‘pig’s head’ but ra tεr=co ‘pig’s head’ [cut]

(53) moco ara=moʈho ‘hen’s leg’ but moco tara=moʈho ‘chicken leg’ [cut]

(54) kƆrƆiŋ ot=car ‘dugong’s chest’ but kƆrƆiŋ tot=car ‘chest of dugong’
[cut, or shown from a distance]

(55) kheŋe ra=uli ‘cat’s tail’ but kheŋe tara=uli ‘cat’s tail’ [cut]

(56) εrεn e=meca ‘deer’s intestines’ but εrεn te=meca ‘deer’s intestines’
[extracted]

(57) kɔrɔiɲ e=tei ‘dugong’s blood’ but kɔrɔiɲ te=tei ‘dugong’s blood’
[extracted]

6.6.2 Alienated but Inherent

The concept of inalienability in PGA does not entail physical separation.
While (58) expresses a body part separated from its source, (59) describes
an intact body part. Both of them are attached with class marker 4 suggest-
ing that the physical separation of the object belonging to the ‘inalienable’
noun category does not require the category of ‘alienable.’ Consider:

(58) cokbi tot=ʈhomu

turtle cl 4=flesh
‘turtle’s meat’ [cut-up for consumption]

(59) cokbi ot=ʈhomu

turtle cl 4=flesh
‘turtle’s meat’ [still on its body]
158 chapter six

(60) cokbi ʈhomu

‘turtle’s meat’ (compounding: modification by noun defining nature), (i.e.,
used in a sentence that ‘he has gone for turtle meat’)

But not

*cokbi-ico ʈhomu
turtle-gen flesh (alienable suffix)

Sentences such as ‘my turtle’s meat’ will use the compound form in (60) pre-
ceded by the first person possessor as in ʈh=ico cokbi ʈhomu. Although gram-
matically correct, this form does not exist in the language. The author never
observed anyone claiming ownership of food items or other consumables.
Why is it that despite being separated from the body the cut up part is
still in the realm of inalienability and is not considered on a par with the
alienated possession for which there is a separate marking -ico ~ -iʃo? The
answer to this question lies in the fact that the speakers realise the inher-
ency of these parts being conceptually dependent on each other despite
being separated. There is less conceptual distance between R and D and
thus, more intimate bonding between the two is realised. This bonding
is retained even after physical separation. Conceptual dependency is
assigned by the prime factor of non-transferability. It is the inability to
transfer the D to different Rs that makes these objects and parts inalien-
able. It is not possible to transfer body parts, affinal or non-affinal kinship,
or the inherent attributes (as seen in §3.3) of an object or a person, even
if they are transient and ephemeral. This is the reason that even after
segregating the body part of animals as in the cases illustrated above it
still retains the body division class marker. As the discussion given in
Chapter 9 on adjectives reveals, the inherent qualifications of a person
or an object are signified by attaching a body division class marker to
each of the modifiers following the possessor noun as in kaba er=buŋoi
a=kata ‘Kaba is beautiful and short’ (class 2 and class 1). The objects
that can be transferred or are expected to be transferred, such as goods in
the market, household goods of a not very intimate nature, objects of the
jungle such as trees, plants, creepers, and other objects of natural envi-
ronment do not obligatorily take ina marker. These may be considered
independent nouns.
In this context one cannot understand why the words for ‘daughter’
and ‘son’ in the language are considered ‘alienable’ and thus demands
genitive marker -ico ~ -iʃo suffixed to the possessor while the word for
‘child’ takes the body division class marker ut= which is added to its
 possession 159

left. The South Andaman language Âkà-Bêa, which became extinct in

1930 is attested with an ut= type of proclitic (Man 1875–1879) for ‘son’
and ‘daughter.’ The absence of a proclitic in PGA with these forms could
have been the result of language change over time. The language is going
through a period of very fast change in the context of contact with other
Indian languages which are devoid of clitic type structures but exemplify
structures with genitives for all kinds of possessions.

6.6.3 Part-To-Whole or Part-To-Component

Part-to-whole or parts of a component are expressed by a possessive mor-
pheme symbolised by an appropriate class marker, but with a prefix t-
to indicate inanimacy. The same class markers as those used in classifying
body part terms as seen in §6.5 above are used. One can establish a one-
to-one correlation between the inanimate segregated parts of a body and
the non-segregated parts of the object in question, as they are considered
to be inherently and conceptually linked to each other.

(61) fεc ta=phoŋ

vessel cl 1=cavity
‘The mouth of the vessel.’

(62) bun tεr=phir

shell cl 2=sharp edge
‘The sharp edge of a shell.’

(63) jicεr tot=tεkho

rain cl 4=sound
‘The sound of rain.’

(64) khidεr tɛr=ʈɔŋ

coconut cl 2=branch
‘The branch of a coconut tree.’

(65) buruiɲ tɛr=pheʈ

mountain cl 2=back
‘Behind the mountain.’

Thus each body division class marker can be used for objects and their
various parts on the same paradigm as body and body parts. The analogy
is very clear in examples like (63) where sound emission is considered
equivalent to products of the body. The body products are symbolised by
the ut= ~ ot= proclitic. The word for ‘sound’ takes the same clitic since
160 chapter six

sounds ‘emerge’ or ‘are produced’ from some source, and hence is marked
by tot=, or in (61), the mouth of a vessel and mouth of a human take the
same class marker and in (65) ‘behind’ of a human body and ‘behind’ a
mountain are marked by the same class marker.
As far as the analogy is concerned, there is no surprise that the hunter-
gatherer society visualizes ‘tree’ as a body and its different parts as belong-
ing to the tree as a whole. Consider Table 6.13, which provides the names
of different parts of a tree each attached to an appropriate inalienable ina
marker designated in body division class markers. It is to be noted that
while describing the partonomy of a tree the word ʈɔkho ‘wood’ is used
more often than the word ʈɔŋ ‘tree.’
Some analogical comparisons are interesting. Consider that the word
ʈɔŋ which means ‘hand’ in PGA, but the same word is also used as a word
for ‘tree’ and ‘branch.’ In a similar manner the word tei means blood as
well as ‘gum’, and the word kɔbɔ means ‘skin’ but is also used to refer to
‘bark.’ Interestingly, these analogies are not unidirectional, since words
from the semantic domains of plants and trees are also used to denote
some animal body parts. For example ɔrɔ means ‘flowers with fruits’ but
can also be used to denote ‘tail’ of some animal as in cao tara=ɔrɔ ‘dog’s
tail’ and teo tara=ɔrɔ ‘crocodile’s tail.’

6.6.4 Inalienable Possessed Nouns: A Conspectus

PGA, in short, maintains that the following types of nouns are prototypi-
cally inalienable and are thus obligatorily prefixed by a body division
class marker. Practically any part-to-component is represented by a

Table 6.13. Terms for the parts of a tree or a plant

Great Andamanese English gloss English translation
ʈɔŋ Tree tree
ʈɔkho ter=tek Tree class 2=wood trunk of a tree
ʈɔkho ta=bεc Wood class 1=hair canopy of trees
ʈɔkho tara=cεʈho Wood class 6=root root of a tree
tɔkho tot=cε Wood class 4=thorn thorns of a tree or a plant
ʈɔkhɔ tεc Wood leaf leaf
ʈɔkho tei Wood blood gum
ʈɔŋ e=ka=thire Tree class 5=class 1=child saplings
ʈɔkho ot=ʈɔŋ Wood class 4=hands/tree branch
ʈɔŋ i=mikhu Tree class 5=middle inside of a tree
ʈɔkho et=kɔbɔ Wood obj=skin bark of a tree
 possession 161

class marker as these are considered inalienable. Inalienability is thus, not

measured on the semantics of ‘separation’ but on the basis of ‘inherency.’
Two nouns that are inherently related are considered to be conceptually
dependent on each other and take any of the seven proclitics considered
so far. Some examples are:

1. Kin terms, e.g. ʈh=a= mimi ‘my mother’ (exceptions are ‘son’ and
‘daughter’)
2. Body parts, body products, tattoo, body paint, e.g. ut=ʈole ‘tattoo’
3. Home, village, courtyard, e.g. ŋ=ot=ɲo ‘your house’
4. Associatives such as language, words, names, clan, community, e.g.
εr=liu ‘his name’, a=jirake ‘Jirake’
5. Ailments both physical and mental, e.g. er=eʈεŋe ‘measles’
6. Boat, head gear, cap, covering for private parts made of leaves, clothes
(in modern times) and boat, e.g. er=buke ‘my cap’
7. Spatial terms, e.g. e=julue ‘in front of’
8. Parts of objects or parts of a component, e.g. ɲo tot=tara ‘roof of the
house’
9. Seascape and landscape terms, e.g. buruin ter=tekh-il ‘in the middle of
mountain’

A consolidated list of body division class markers used as possessive

class markers to classify nouns is given in Table 6.14 and Figure 6.1.

Table 6.14. Possessive class markers defining primary possession and classes of nouns
body division class Body parts Kinship Other Types of Semantic
marker proclitics terms objects classes axioms
animate inanimate
a= ~ at= ta= ~ tat= tongue, mother language, Primary mouth and
mouth father, grand proper 1 extension
parents names of mouth
εr= ~ er= tεr= major body spouse side, near, Primary external
parts, head, name, above, 2 organs,
calf etc. cap, deictic
headgear
oŋ= toŋ= fingers, — — Primary extension
nails 3 of hand,
extremities
 162 chapter six

Table 6.14 (cont.)

body division class Body parts Kinship Other Types of Semantic
marker proclitics terms objects classes axioms
animate inanimate

u= ~ut= tu= ~ tut= chest, back, child, house, Primary extension

~ ot= ~ tot= hair, sweat brother ailments 4 of self,
products
of self
e= te= blood, spouse fever, Primary internal
pancreas, leaf-covering 5 organs
ailments
ara= ~ ra= tara= waist, side sister village, Primary deictic,
and hip boat 6 relational,
area and its nodulated,
extensions circular
like tail
o= ~ ɔ= to= ~ tɔ= leg, heels, — — Primary lower part
calf 7 of the body

Inalienable possession

Possessed
Possessor (inalienable)

-animate +animate Class 1 Class 2 Class 3 Class 4 Class 5 Class 6 Class 7

t+V/VC/VCV V/VC/VCV

Figure 6.1. Body division classes and possession

 possession 163

6.7 Alienable Nouns

6.7.1 Possessed Nouns

Alienable possession in the language is designated by a genitive mor-
pheme which is suffixed to the possessor noun. This genitive has two allo-
morphs; -ico ~ -iʃo. These variations occur across speakers from different
language backgrounds. For example, the Sare speaker always used -ico
while the Khora speaker used -iʃo. Most of the typical alienable nouns des-
ignating ‘land’, ‘jungle’, ‘upper garments’, ‘lower garments’, ‘dog’, ‘friend’,
‘God’, as well as some kinship terms, such as ‘son’, and ‘daughter’ govern
genitive markers -ico or -iʃo which are suffixed to the possessor noun. The
encoding of possession pertaining to alienable objects draws our atten-
tion to the fact that we find evidence of both head marking and depen-
dent marking in PGA; the former for inalienable possession and the latter
for alienable possession.

(66) nu iʃo julu

Nu gen dress
‘Nu’s dress.’

(67) n=iʃo ko
3pl=gen bow
‘Their bow.’

(68) ʈh=ico boa

1sg=gen land
‘My land.’

Figure 6.2 summarises the word formation processes involved in posses-

sive constructions in PGA. PGA is a rare language where the choice of
a possessive marking is decided by both the possessor (R) and the pos-
sessed (D). In addition, alienable nouns are represented by genitives while
the inalienable ones are represented by body division class marker pro-
clitics. Hence, the language offers both dependent and non-dependent
marking systems.

6.7.2 Twelve Different Varieties

To conclude, there are, in all, twelve different varieties of possessive
markers which, except for one, function as proclitics depending upon
164 chapter six

Syntactic Juxtaposition/
A���xation/cliticisation compounding

Adverbial
Alienable

Inalienable Partitive

Primary Secondary

Figure 6.2. Word formation processes involved in possessive constructions

the semantic nature of the possessor, the possessed, the relation between
the two, and the type of word formation processes. Out of these twelve,
only one marker is used for alienable possession: the genitive marker.
The inanimate possessor represented by initial t- is considered the twelfth
one which is demonstrated along with the forms used with animate pos-
sessor in Table 6.15. The rest are seen as variations of inalienable pos-
sessions. By taking into account a large inventory of nominal objects, it
becomes eminently clear that the terms alienable versus inalienable are
not very appropriate for the language under analysis. It should be under-
stood as the phenomenon of ‘inherency.’ The large variety of possessive
class markers as observed in PGA is unusual.
A summary of all the body division class markers serving as posses-
sive proclitics, and genitives is provided in Table 6.15. These categories
are divided into two hierarchical levels, the primary and the secondary.
The decision to categorise them as primary or secondary is based on their
morphological shape. Monomorphemic constructions are primary posses-
sions while those which use derivation or compounding are the second-
ary ones. Another reason to divide them into primary and secondary is
the semantics of body parts. The secondary body part terms define part
of a whole or a part of a component or a combination of two or three
body parts.
 possession 165

Table 6.15. Varieties of possession

Number Word possession Semantic Hierarchical
formation type and category & level
morphemic noun class
structure
1 Clitics a= ~ ta= Inalienable 1 Primary
2 Clitics εr= ~ tεr= Inalienable 2 Primary
3 Clitics oŋ= ~ toŋ= Inalienable 3 Primary
4 Clitics ut= ~ ot= ~ Inalienable 4 Primary
tut= ~ tot=
5 Clitics e= ~ te= Inalienable 5 Primary
6 Clitics ara= ~ tara= Inalienable 6 Primary
7 Clitics o= ~ ɔ= ~ Inalienable 7 Primary
to= ~ tɔ=
8 Affixation -ico ~ -iʃo Alienable Primary
9 Compounding R_D Inalienable Secondary
10 Compounding class marker=R_D Inalienable Secondary
with CLASS
11 Complex R _class Inalienable Secondary
double marking marker=adv_
born_D

6.8 Attributive Modification and Possession

The language considers all inherent and external attributes of a person or

of an object as inalienable attributes modifying the head noun designated
by an appropriate proclitic, i.e. the body division class marker. These
are the internal structure of an NP which includes possession. Thus, words
such as ‘boy’s courage’ or ‘courage of the boy’ express the possession of
the attribute ‘courage’ (D) by the possessor ‘boy’ (R). Other than numerals
and quantifiers, attributes of shape, size, colour and quality are attached
by an appropriate class marker representing the modified-modifier
order. All such constructions imply modification and can be used both
as an attributive modification and as a predicative modification because
all adjectives can take the predicate slot as seen in Chapter 3. It must be
noted that these are not prototypical possessor-possessed relationships
or ownerships, but explicate inherent modification. These are considered
separately in Chapter 8.
166 chapter six

6.9 The “Possessive” Relationship: The ‘Have’ Construction

In PGA the morpheme cul ‘have’, ‘near’ functions as a transitive verb and
thus the D is like an object noun following this morpheme. The construc-
tion is used only for the alienable possession and is in no way parallel
to the English ‘have’ constructions which, in addition to marking posses-
sion of alienable goods, marks body ailments too, e.g., ‘I have a headache.’
The morpheme cul signifies proximity of the object to the possessor. The
most interesting part is that this morpheme cul is itself preceded by a
body division class marker 2 εr= ~ er= which is used for face-related and
major body parts. There is another morpheme cuo with the same mean-
ing but without the obligatory class marker morpheme. Consider a few
examples.

(69) er=cul jiyo khider

cl 2=near be-exist coconut
‘(He) has a coconut.’ / ‘There is a coconut by his side.’

(70) a=boro cuo kider-bi jiyo

cl1=Boro have coconut-abs exist
‘Boro has coconuts.’

(71) ʈh=εr= cuo kapra cophe bi

1sg=cl 2= have clothes many cop
‘I have many clothes.’

(72) e(r)=cul o cɔ pho be

cl 2=have 3sg net neg cop
‘(He) does not have a net.’ / ‘He does not possess a net.’

As said earlier, inalienable possession does not take ‘have’ cul ~ cuo. Con-
sider the following sentence without cul ~ cuo.

(73) ʈh=u(t)=tεi tεr=biraŋ be

1sg-cl 4=blood cl 2=red cop
‘My blood is red.’ / ‘I have red blood.’

The word er/εr=cuo-l (class 2 =‘near’-loc) can employ either the copula
bi ~ be or the verb of ‘exist’ jiyo. In fast speech the locative suffix -l is
often dropped.
 possession 167

6.10 Conclusion

To conclude, PGA provides seven basic possession types which are derived
from the terms used for the seven divisions of the human body. These are
termed body division class markers which are grammaticalised further
to indicate possession of various inherently dependent objects, both phys-
ical and reified. The body division class markers are attached to the left
of the dependent nouns. Hence, they are proclitics. Thus, the language
has possessed entities as the head of the phrase which are marked by
the possessive morpheme. PGA has provision for further extending these
basic seven types by various word formation devices which indicate modi-
fication and part-to-whole or part-to-component dependency. NP internal
possession can serve as the head or the object of a verb or as a comple-
ment of a phrase. The language is unique in the sense that both possessor
noun as well as the possessed noun determine the choice of an appro-
priate possessive marker. PGA has only one true genitive form which is
suffixed to the possessor noun in alienable possession. This brings forth
another unique feature of the language: it has dependent-marked head
nouns in inalienable possession, as well as dependent-marked non-head
nouns in alienable possession. In all, the language has twelve different
varieties of possessions.
 Chapter Seven

Pronouns, pronominal and object clitics

7.1 Personal Pronouns

7.1.1 Person and Number

PGA recognises three persons and three numbers, which is evident in
the pronominal system of the language. The forms for first, second and
third person have three numbers. The bare pronominal forms in the lan-
guage are more elaborate than in the other two endangered languages of
the Andaman Islands, i.e. Jarawa and Onge (Abbi 2006a). A distinction
between inclusive and exclusive is maintained in the first person plural
pronouns.
Consider the forms in Table 7.1. It should be noted that there are minor
variations across speakers. Pronouns in PGA can occur in the bare form
or can be cliticised as the subject or object of a verb. The cliticisation
does not operate on dual and plural forms. This implies that dual and
plural pronouns occur in full forms only. We will first consider the list of
all these forms in Table 7.1 and will then give examples for each of the
forms (1–15).

Table 7.1. Pronominal forms in PGA

Personal Pronouns
1sg. (speaker) ʈhu
1sg. (emphasis, ‘I alone’) ʈhio
1du. (two speakers) ʈhεɲio
1pl.excl. (addressee not included) ma ~ mio ~ mia
1pl.incl. (addressee(s) included) meŋ ~ meŋamboro
1pl.incl. (emphasis) meŋio
2sg. (one addressee) ŋu
2sg. (emphasis, ‘you alone’) ŋio
2du. (two addressees) ŋole
2pl. (more than two addressees) ŋilie ~ ŋilio
2sg.h. (one honorific addressee) naŋe
2du.h. (two honorific addressees) naŋela ~ ŋoli ~ ŋamboro
2pl.h. (more than two honorific addressees) ŋale ~ ŋele
170 chapter seven

Table 7.1 (cont.)

Demonstrative pronouns
3sg.dis.vis. di ~ ɖi
3du.dis.vis. ɖuini
3pl.dis.vis. ɖunio
3sg.dis.invis. ɖu ~ ɖuio
3pl.dis.invis. ɖune
3sg.prox1 (nearer to the speaker) khidi
3sg.prox2 (nearer to the hearer) khudi
3pl.prox2 (nearer to the hearer) khudino ~ dia ~ diya

The pronominal system of the language offers focused pronominal forms

in addition to inclusive/exclusive distinction. Hence the distinction
between ʈhu and ʈhio ‘1sg’ is that the latter form carries additional empha-
sis or focus, something like English ‘I alone and no one else.’
Secondly, the first person plural has two forms: exclusive and inclusive,
demonstrating the presence or absence of the addressee. The 1pl.excl
form ma denotes the speaker and at least two other individuals, neither
of which is the hearer, whereas the 1pl.incl forms meŋ or meŋamboro are
associative plurals which refer to the speaker and at least two other indi-
viduals, one of which is the addressee (generally kin members).
Thirdly, other than the visibility factor to play a role in deciding the
demonstrative pronoun, it is the two levels of proximity which interplay
with the decision of the appropriate third person proximate forms. Thus,
proximate objects are distinguished between ‘very close’ which denotes
references near to the speaker, prox1 vs. ‘intermediate’ which is between
‘distant’ and ‘proximate’, which can be safely recognised as nearer to the
addressee prox2. Thus khidi refers to an object or a person which is very
close, almost touching the speaker, but khudi is used for an object or a
person which is proximate, visible, but closer to the hearer.
Sentences (1)–(7) exemplify first person pronouns.

(1) ʈhio jero be

1sg.emph Jero cop
‘I (alone) am a Jero.’

(2) meŋio meŋut=cɔne-b-e

1pl.incl.emph 1pl.incl.cl 4=go-imm.pst
‘We (inclusive) went.’
 pronouns, pronominal and object clitics 171

(3) mia khriŋkɔʃo-il ʈhi-bik ɲƆ-k-Ɔm

1pl.excl Strait island-loc place-abs.obj live-fa-npst
‘We (alone, only) live in the Strait Island.’

(4) ma jero be
1pl.excl jero cop
‘We (alone) are Jero.’

(5) mio ʈhotarphuc be

1pl.excl Andamanese cop
‘We (alone) are Andamanese.’

(6) meŋ-mam-bikhir cɔkbi-bi ji-k-ɔ

1pl.incl-refl-morning turtle-abs eat-fa-dst.pst
‘We (inclusive) ate turtle last morning.’

(7) meŋ-ot=bo et=ɖello

1pl.incl-cl 4=back obj=ball (heart)
‘Our hearts.’

7.1.2 Second Person Pronouns and Honorifics

Honorifics are represented in the second person singular, dual and plural
forms. Honorifics are indicated by the abbreviated small caps h. Sentences
(8) to (10) exemplify second person honorific pronouns while sentences
(11) to (15) exhibit varieties of other second person pronouns.

(8) ŋaŋe ŋam-bikhir cɔkbi-bi ji-k-ɔ

2sg.h refl-morning turtle-abs eat-fa-dst.pst
‘You ate turtle yesterday morning.’

(9) naːŋe ŋuʈhi lec-bi tεsaŋ

2sg.h 2sg1sg.dat arrow-abs give
‘Sir, will you give me an arrow?’

(10) ŋoli ut=bucɔ

2du.h cl 4=lap
‘Your (two of you) laps.’

(11) ŋu ɖunio ne-boʃo-k-e

2sg 3pl.dis.vis 3pl-thrash-fa-imp
‘You thrash them (nearer to the hearer).’

(12) ŋilio ŋalim-bikhir cɔkbi-bi ji-k-ɔ

2pl. 2pl.h.refl-morning turtle-abs eat-fa-dst.pst
‘You (pl) ate turtle last morning.’
172 chapter seven

(13) ŋole khringkɔʃo-il ʈhi-bi-k=ɲo-k-ɔm

2du. Strait Island-loc place-abs-obj=live-fa-npst
‘You (two) live in Strait Island.’

(14) ŋilio ŋili-ʃɔ-m rεfe-be raʃue-k-ɔm

2pl. 2pl-gen-refl food-abs cook-fa-npst
‘You folks cook your own food.’

(15) ʈhio ŋio ŋu ŋa=tεkho-bik

1sg.emph 2sg.emph 2sg 2.cl 1=language-abs.obj
bobi ŋon ci-k-o
learn/know come-fa-dst.pst
‘I alone had come only to you to learn your language.’

7.2 Demonstrative Pronouns

The defining properties of demonstrative pronouns in PGA are:

(a) Demonstrative pronouns are not different from any other personal
pronouns in forming the NP by themselves.
(b) The co-referential property of demonstrative pronouns can be
explained as such: they serve an anaphoric function whereby they
substitute a full NP, the reference of which is made earlier in the
discourse. The plurality of demonstrative pronouns such as ɖunio is
co-represented in the temporal adverb ‘morning’ as a plural prefix n-
(16) or the singularity of ɖu is co-represented in the temporal adverb
for ‘morning’ as aka- (17) explicating its full NP structure.
(c)  As is true for many languages, the demonstrative pronouns in PGA
can also occur with a noun in an NP (18, 19).
(d) They offer spatial contrast with proximity and distance defined along
parameters of the visibility and invisibility of the referent (16, 17) as well
as the proximity between the speaker and the referent (20, 21 and 22).

(16) ɖunio n=ambikhir cɔkbi-bi ji-k-ɔ

3pl.dis.vis pl=morning turtle-abs eat-fa-dst.pst
‘They ate turtle in the morning.’

(17) ɖu ak=ambikhir cɔkbi-bi ji-k-ɔ

3sg.dis.invis 3sg=morning turtle-abs eat-fa-dst.pst
‘He ate turtle in the morning.’
 pronouns, pronominal and object clitics 173

(18) ɖu (t)=ɲo buruin tot=bɔ-l ɲo jio

3sg.dis.invis(cl 4)=house mountain cl 4=behind-loc house exist
‘That house is behind the mountain.’

(19) ɖune ra moʈe lobe

3pl.dis.invis pig fat (<Hindi) exist
‘Those pigs are fat.’

(20) khudi-no cae bi caʈo-k-o

3pl.prox2 what abs do-fa-dst.pst
‘What were these (people) doing?’

(21) khidi ʈh=ot=ɲo be

3sg.prox1 1sg=cl 4=house cop
‘This is my house.’ (the speaker is in the house)

(22) khudi loka ot=ɲo be

3sg.prox2 Loka cl 4= house cop
‘This is Loka’s house.’ (pointing towards it)

(23) di kɔbɔ ot=ɲo be

3sg.dis.vis Kobo cl 4=house cop
‘That is Kobo’s house.’

Please refer to §8.5 for details on the spatial deictic functions of these
demonstrative pronouns. Also refer to Table 8.12 on demonstrative pro-
nouns in the next chapter.

7.3 Pronominal Clitics

First, second and third person singular pronouns can occur as reduced
single consonant or vowel clitic forms. Different case markings can be
attached to basic pronouns for an appropriate meaning only after they
are cliticised. Thus,

ʈhu + i > ʈhi ‘to me’ (dative)

ŋu + e > ŋe ‘you’ (agentive)
ɖu + i > ɖi ‘to him/her’ (dative)

Prototypically, the cliticised forms with the case markings designate

arguments in a sentence (24). There are restrictions in their occurrence
as clitics:
174 chapter seven

(a) Pronominal clitics are not attached to the absolutive -bi. However,
case suffixes as agentive -e, accusative/dative -i, genitive -ico/-iʃo, and
ablative -a can be attached to the pronominal clitics.
(b) It is the pronominal clitic which attracts the body class marker of
the following verb or a noun and helps in generating a phonological
word. For details see §2.5.8.
(c) The second person plural form does not have a clitic form.

Cliticised forms of pronouns followed by various types of body class mark-

ers can be attached to a verb, adverb or a noun phrase (25). In other words,
the pronominal clitics can combine with body class marking proclitics
(a case of clitic sequencing and climbing as we saw earlier in §3.8.3). Con-
forming to the SOV type, the nominal and pronominal arguments or their
clitic forms precede the verb as a subject and/or object of the verb. These
are in bold in the examples given below.

(24) ʈhi miʈhaibi tεʃeke

ʈh=i miʈhai-bi tεʃe-k-e
1sg-dat sweet-abs give-fa-imp
‘Give me the sweets.’

(25) roa-ta stret-ak ʈh=ut=cone-b-o

boat-instr Strait-dir 1sg=cl 4=go-fa-dst.pst
‘I went to Strait by boat.’ (allative)

The following are the pronominal clitics in subject and object positions.
Some positional verbs such as ‘sit’ necessarily take the subject pronomi-
nal clitic, e.g. ʈh=aono ‘I sit.’ The first and second pronominal clitic forms
given above can be attached to such verbs. However, the third person u=
or o= is rendered as aka= (29, 31). This form also occurs if the object is of

Table 7.2. Pronominal clitics

Pronominal clitic S and O functions
ʈh= 1sg
m= 1pl
ŋ= 2sg
u= 3sg.dis.vis
o= 3sg.dis.invis
n= 3pl
aka= 3sg.obj of possession/anaphoric
 pronouns, pronominal and object clitics 175

inalienable possession, or has an anaphoric reference. Most of the names

of the languages are preceded by the morpheme aka to indicate inalien-
ability or inherent possession of one’s language. Thus there are the names
such as ‘Aka-Bo’, ‘Aka-Kol’, ‘Aka Jeru’, etc. for the languages Bo, Kol, and
Jeru respectively. This practice of attaching aka to the language names is
no longer used. The form aka can also occur as a pronominal object (27).
Other examples of this form are (28, 29) and (31).
Instances of pronominal clitics can be presented as follows:

(26) maramliʃu mutconne ŋutconepho

m=ɑrɑm-liʃu m=ut=cone ŋ=ut=conne-pho
1pl.excl=refl-all 1pl.excl=cl 4=go 2sg=cl 4=go-neg
‘All of us will go but not you (you will not go).’

(27) o=ek= ak=aŋkhui-ø

3sg.dis.invis=3sg.obj= 3sg=embrace-pst
‘He embraced him.’

(28) ʈhεtujulul akaunobom

ʈh=ε=tujulul ɑkɑ ɑuno-b-om
1sg=cl 2=early/already 3sgsit-fa-npst
‘He sat earlier than me/he was already sitting there.’

(29) meo bas sʈopil akanʈɔyil o malaiyo

meo bɑs sʈop-il ɑkɑ-ʈɔy-il o mɑlɑi-o
Meo bus stop-loc 3sg-stand-conv 3sg tire-dst.pst
‘Meo was tired of standing (for a long time) at the bus stop.’

(30) ʈhu ʈh=aono-l ʈh=olam-o

1sg 1sg=sit-conv 1sg=tire-dst.pst
‘I got tired sitting.’

(31) o ak(a)=aono-l o=olam-o

3sg.dis.invis 3sg=sit-conv 3sg.dis.invis =tire-dst.pst
‘He got tired sitting.’

7.3.1 Human Arguments (Subject and Object)

Human arguments including proper nouns (32, 33) are preceded by the
body division class 1. As we mentioned earlier in §5.2.1 this is primarily
a feature of the Bo and Jeru languages. It was difficult to ascertain the
semantic status of this marker as not all speakers used it. Those who did
use it, considered it as a marker of specificity. Speakers like Boa Sr., the
last speaker of the Bo language, never completed her sentences without
176 chapter seven

the use of a before all human subject and object nouns while others in the
community thought it was optional to use it. It should be noted that pro-
clitics never precede pronominals. Thus, *a=ʈhu is not possible. As seen
in Chapter 3, the class 1 proclitic is grammaticalised for the semantics of
‘origin’, or ‘source.’ Interestingly, the class 1 proclitic is attached to nouns
referring to domestic animals such as ‘dog’, hence a=cao.

(32) athirenu churuŋbi raliʃu ejiu

ɑ=thire-nu churuŋ-bi erɑliʃu e=iji-o
cl 1=child-pl sweet-abs empty cl 5=eat-dst.pst
‘The children ate up all the sweets.’

(33) athirenu naramliʃo boʃobo

ɑ=thire-nu n=ɑrɑ.ɑm-liʃo boʃo-b-o
cl 1=child-pl 3pl=cl 6.refl-sister beat-fa-dst.pst
‘The children beat their sisters.’

(34) a=ram-e a-ravan-ek εrenceo

ɑ=ram-e ɑ=ravan-ø ek=εrence-o
cl 1=ram-erg cl 1=ravan-abs 3sg.obj =fight- dst.pst
‘Ram fought with Ravana.’

(35) a-boa a=nu ik=jira-ø tatoŋ-e it=ʃir-o

cl 1=Boa cl 1=Nu 3sg.obj-tell.conv courtyard-abs obj=wash-dst.pst
‘Boa told Nu to wash the courtyard.’

(36) mεŋo khider-bi mεŋo εt=bɔl-o

1pl.incl coconut-abs 1pl.incl 3obj=peel-dst.pst
‘We alone, we all [inclusive] peeled the coconut.’

7.3.2 Object Clitics

7.3.2.1 Pronominal Objects

As mentioned earlier, the first, second and third person pronominal clit-
ics more often than not appear as a single consonant or a vowel. These
can appear in the second position in the sentence with dative, accusative
or genitive case suffixes as object pronominal clitics. The second person
plural is not cliticised, but the first and third person plural can appear in
clitic forms as in m= and n= respectively. Consider

(37) ŋu me boʃo-b-e
2sg 1pl.obj thrash-fa-imp
‘You will thrash us.’
 pronouns, pronominal and object clitics 177

(38) ʈhu ŋele boʃo-b-e

1sg 2pl.obj thrash-fa-imp
‘I will thrash you people.’

(39) ŋu ɖunio ne boʃo-k-e

2sg 3pl 3pl.obj thrash-fa-imp
‘You will thrash them.’

(40) u ʈhi jira kɔil o au-b-om

3sg 1sg.dat tell later 3sg come-fa-npst
‘He told me that he will come tomorrow.’

7.3.2.2 Inanimate External Objects

Inanimate object clitics seem to be derived from the body division mark-
ers but substantial proof is lacking as the language is dying fast and many
structures have become obsolete. Some of these clitics can be recognised
as body division markers but not all are transparent. They show up as
clitics attached to various verb forms in several shapes according to the
shape or the quality of the object it represents and according to the nature
of the verb it is attached to. These are: uku=, kak=, εk=, ek=, ik=, it=, ut=,
et=, i=, e=, and ε=. Clitics such as it=, et=, i=, and e= can be speculated to
derive from body class marker 5 representing ‘inherency’ or ‘internal’ to an
object and similarly, ut= may have been derived from body class marker 4
to denote an object which is a result of emission. However, it is difficult
to decide the source of other clitics.
Since every transitive verb in PGA is preceded by one of the several
proclitics available in the language, the shape of each of these proclitics
is decided by the semantic nature of the host verb. For example, the verb
beliŋ ‘cut’ is preceded by the object clitic iku= ~ ikku= which signifies the
resultative object of that action. Similarly, the verb ‘to wash’ ʃir is accom-
panied by an object which bears the impact of the action of ‘washing’, i.e.,
with class marker 5 which represents internalised objects. Hence ‘wash’ is
a verb which shows an ‘internalised’ event while ‘cut’ is a resultative verb
giving rise to a distinct entity. Both are accompanied by different object
clitics. Another clitic ik= ~ ek= ~ εk= signifies the alienable direct object
which can precede verbs of ‘hunt’, ‘tell’, ‘fight with’, ‘catch fish’, etc. It can
be seen as ‘evidential’ too. Clitics thus can function as modals as well as
manner (see below (a) ‘various forms of ‘cut’). It can be speculated that
these are grammaticalised forms of lexical items which once existed in
the language. For instance, in the following examples the presence of the
clitic ik= signifies the successful catch and the absence of it signifies that
178 chapter seven

the hunters returned empty-handed. It is quite common to hear εk=rε?

with a rising tone to mean ‘did you get anything’ while hunting something
in the forest. Consider the difference in meaning between (41) and (42).

(41) nu cokbi-lara-bi ik= uni

people turtle-hunt-abs 3sg.obj= came back
‘People came back after hunting a turtle.’ (brought the hunt)

(42) nu cokbi-lara-bi ø-uni

people turtle-hunt-abs came back
‘People came back after hunting turtle.’ (did not get one)

(43) nu taŋol ik=ao-ø

people catch fish 3sg.obj=came back-pst
‘Having caught fish people came back.’

(44) ŋu cokbibi ŋuekphoʈε

Nu cokbi-bi Nu ek=phoʈε
2sg turtle-abs 2sg 3obj=cut
‘You cut the turtle.’

Earlier in Chapter 3, we considered while discussing the transitive verb

‘cut’ that it can be preceded by various forms of the object clitic depend-
ing upon the motion of the action ‘cut’ as well as the object of the result
of the action ‘cut.’ Hence, the same lexical unit pho ‘cut’ or kata ‘cut in
chunks’ will have different resultative objects as well as a different man-
ner of action. An equivalent English translation is sometimes long and
sometimes requires different lexemes.
Adding these clitics to transitive verbs is indicative of the inalienable
character in the way that the object of the verb and the manner in which
the action is performed cannot be alienated from each other. Consider:

(a) Various forms of ‘cut’

et=pho ‘cut or separate from the source’ (e.g., betel nut from a branch)
ara=pho ‘cut to make it fall on the ground’; ‘fell a tree’
ik=/ek=kata ‘cut a small piece’
okko=kata ‘cut a very small piece’
it=kata ‘cut a big piece’
εt=/ut=kata ‘cut a very big piece (something like 150 grams)’

7.3.2.3 Internalised Objects

We saw in Chapter 3 that if the object is internalised the body class marker
e- or i- is used for invisible body parts or for those attributes which are
 pronouns, pronominal and object clitics 179

inherent qualities such as ‘to be bad’ or ‘to be good.’ This implies that the
attributes of ‘good’ and ‘bad’ are internalised in a person. In addition, we
saw in the case of intransitive experiential verbs (§3.5) that all of them are
preceded by body class marker 5 symbolising the experience, emotions or
feelings which are internal to the body. The impact verbs such as ‘wash’
considered above also are accompanied by this object clitic. In the follow-
ing examples both the object noun ‘fever’ and the verb ‘to have fever’ take
the object clitic ε ~ i ~ e to signify an internalised object.

(45) ʈh=ε=ʈε i=bi

1sg=cl 5=fever 3obj=cop
‘I have a fever.’ (i.e., 3obj i- co-refers the noun tε ‘fever’)

The objects of ‘sing’, ‘embrace’ and ‘eat’ also share the same clitic.

7.3.3 Pronominal Clitics and Temporal Adverbs

An interesting feature of PGA is that first, second, and third person sin-
gular pronominal clitics can occur with words designating time, i.e., tem-
poral adverbs as a co-reference to the personal pronominal subject. This
has also been considered in Chapter 3 as well as in §7.2 above. Consider
some more examples.

(46) asalma kambikhir konabi jiyo

ɑ-sɑlmɑ ɑkɑɑm-bikhir kɔfo-bi iji-o
cl 1-Salma 3sg refl-morning banana-abs eat-dst.pst
‘Salma ate a banana in the morning.’

(47) ʈhambikhir jaipure ʈhutconne phobe

ʈh =ɑm-bikhir jɑipur-e ʈh=ut=conne pho-be
1sg=refl-tomorrow Jaipur-abs 1sg =cl 4=go neg-cop
‘I will not go to Jaipur tomorrow.’

(48) ŋambikhir alephai eolebo

ŋ==ɑm-bikhir ɑ=lephai e=ol-(e)-b-o
2sg=refl-morning cl 1=Lephai cl 5=see-fa-dst.pst
‘You saw Lephai yesterday morning.’

(49) a-ʈoʈa akambikhir tεyotrεpho akambikhir uemphilo/akaːle

ɑ=ʈoʈɑ ɑkɑɑm-bikhir tεyotrεpho ɑkɑɑm-bikhir u=akɑːl-e
cl 1=boy 3sg.refl- fever.dst.pst 3sg refl- 3sg=die-
morning morning imm.pst
‘The boy who had a fever yesterday died this morning.’
180 chapter seven

Also refer to Chapter §8.4.1. We can summarise the section on object clit-
ics by saying that the third person object clitics come in varying forms
depending on the size, shape, quality and location of the object nouns
that result from the action. These clitics attach to various verbs on their
left. In addition, they indicate the manner of an action. They attach to
temporal adverbs to designate concepts such as ‘morning’, ‘evening’, etc.
Clitics are important in order to fully comprehend the meaning of a verb
in PGA.

7.4 Interrogative and Indefinite Pronouns

In PGA the forms for the interrogative and the indefinite pronouns are the
same in non-human contexts. The pronoun cay is used for both of these
functions as seen in the following examples.

(50) das ʈɔkho-tεc-il ʈhu caːy cɔkho-bi bul-e

ten (<Hindi) tree-leaf-loc 1sg whatever all-abs buy-imm.pst
‘I bought what all (everything) in ten rupees.’

(51) ocay erŋolom

o cɑy erŋol-om
3sg what write-npst
‘He is writing something’ / ‘What is he writing?’

(52) ɖutɲo ca:y jukhe

ɖu (u)t=ɲo cɑ:y jukhe
3sg cl 4=house what rel/spec
‘Which one is his house?’

(53) ʈhu ca:y-bit kaɲom

ʈhu cɑ:y-bi it=kɑɲ-om
1sg what-abs obj=touch-npst
‘I am touching something.’

PGA has two basic interrogative pronouns: one is used for human referents,
such as aʃiu and the other for non-human referents, i.e., cay. The former
has two variants (used by different speakers) given in the following table.
The latter is productively used for deriving all kinds of WH constructions.
It should be noted that the interrogative pronouns can be combined with
a demonstrative pronoun (55). Interestingly, the human interrogative pro-
noun aʃiu can be combined with the interrogative pronoun cay to give the
 pronouns, pronominal and object clitics 181

Table 7.3. Indefinite/Interrogative forms

Human Non-human
Interrogative aʃiu ~ ʃɔ ~ aʃetani ‘who’ cay ~ ca:y ‘what’
Indefinite caʃiu ‘somebody’ cay bit ‘something’
Relativised ʃɔ jukhi ‘which one’ cay jukhi ‘which one’
Purposive ʃɔ khudi ‘which one’ cay khudi ‘for what’, ‘why’
Locative cayl ~ cya:l ‘where’
Manner ʃetani ‘how’ ʃetani ‘how’

indefinite form caʃiu ‘someone’ as in (57), or the interrogative inanimate

cay can be combined with the quantifier cophe ‘many’ to render the mean-
ing of an indefinite plural entity (59). Also refer to §10.4 on interrogatives.
Table 7.3 lists the various types of interrogative pronouns.
The human referent can be the ‘name of a person’ and not the person
as such. For example,

(54) ŋ=er=liu aʃyu bi

2.sg=cl 2=name what cop
‘What is your name?’

Other examples of interrogative and indefinite pronominal references

are:

(55) di aʃiu bi
3sg.dem who cop
‘Who is he?’

(56) ɲyol aʃiuphu bi

ɲyo-l ɑʃiu phu bi
home-loc who neg cop
‘No one is at home.’ or ‘Who is not at home?’

(57) acaʃiu itconebo

ɑ=cɑ-ʃiu ut=cone-b-o
cl 1=what-someone cl 4=go-fa-dst.pst
‘Someone went away.’

(58) dillikεk caːy khudi ŋutconebom

dilli-kεk cɑːy khudi ŋ=ut=cone-b-om
Delhi-dir what for 2sg=cl 4=go-fa-npst
‘Why are you going to Delhi?’
182 chapter seven

(59) thire=n caːy cɔphe

child=pl what many
‘How many children does he have?’

(60) ʈhu caːy-bit kaɲ-om

1sg what-abs.obj touch-npst
‘I will touch something.’

(61) acaʃiu narkha ʈhitʈuddilo kaːretta

ɑcɑ-ɑʃiu nɑrkhɑ ʈhi-t=ʈuddilo k-ɑːrett-ɑ
what-someone many land-cl 4=ball 3sg-tell-imm.pst
‘Many people informed me about the island.’

(62) u cay-bi ta=biŋo-ø

3sg.what-abs appl=think-pst
‘What did he think/remember?’ or ‘He thought of something.’

(63) ŋ=ut= thu a=ʃiu bi

2sg=cl 4=born cl 1=what cop
‘Who is your brother?’

(64) ŋu cay iji it=kalemo

2sg why/what eat 3obj=soon
‘Why did you eat so fast?’ or ‘What did you eat so fast?’

7.5 Reflexive Forms

PGA has both nominal and verbal reflexives. The language has four dif-
ferent kinds of reflexive forms in the area of nominal reflexives. These
are suffixed to the specific pronominal clitic. None of them designates
number or person:

(a) The form -tuŋ indicates a possessive relationship with human nouns
indicating kinship and body parts (65–70);
(b) Reflexive -em ~ -εm denotes specific identification with the subject or
object (71–73) and is attached to verbal forms;
(c) In addition to being suffixed to pronominal proclitics, alternatively, it
can be suffixed to the genitive morpheme -iʃo ~ -ico to show alien-
able reflexive possession (74–76);
(d) Another form is -encuo which co-refers to the pronominal subject
of the clause (82–86) to designate emphasis. The details are as
follows.
 pronouns, pronominal and object clitics 183

Let us consider each of them in little detail.

(a) -tuŋ ~ -tun

The reflexive is attached to the pronominal form on its right. The pres-
ence of different pronominal forms does not bring any change in its shape.
However, the pronoun is repeated once. Some examples are in order.

(65) ʈhu ʈhu-tuŋ ʈhire

1sg 1sg-refl children
‘My own children.’

(66) ŋu ŋo-tuŋ ʈhire

2sg 2sg-refl children
‘Your own children.’

(67) o u-tuŋ ʈhire

3sg 3sg-refl children
‘His own children.’

(68) nili ŋa= ŋo-tuŋ ʈhire-bi εk~εt=rɔšε-k-om

Nili 2sg.cl 1= 2sg- refl children-abs 3obj=love-cl=fa-npst
‘Nili, you love your children.’

(69) ʈhu tuŋ co-bi ɲao-co-ø

1sg self head-abs scratch-head-pst
‘I scratched my head.’

(70) aka-mimi u-tun-thire ta-ʈɔph-om

3sg-mother 3sg-refl-child caus-bathe-npst
‘His/her mother is bathing her child.’

(b) -εm-

This reflexive form is very commonly used in narration, especially if the

narration is in the first person. Instead of the simple first person pronoun,
ʈhu, speakers prefer to use ʈh=εm to indicate ‘I alone’ or for emphasis with
the English parallel ‘It was I.’ Thus, a phrase such as ʈh=εm-a=thudil ‘It
was I who killed it’ or ‘I alone killed it’ are the closest translations. This
form also generates reflexive verbs (74). It is safe to say that in case of
the reflexive ‑εm-, in addition to the meaning of ‘self-directed’, it can also
mark ‘emphasis.’
184 chapter seven

The reflexive morpheme, which co-refers to the subject pronoun is

infixed between the proclitic and its host which could be a verb root or
a noun root (71, 71a). Thus, the verb ‘laugh’ is εr=khole but with the reflex-
ive it is rendered as εr=εm-khole >εrεmkhole ‘laugh’, or the word for ‘tail’ is
ara=ulibi and with the reflexive the form is ara=εm-ulibi > arεmulibi ‘his
own tail.’ Refer to §4.1 as well as §7.6 below.

(71) ŋ=εr=εm khole-l

2sg=cl 2=refl laugh-nrr.pst
‘You were laughing’, ‘You were laughing to yourself.’

(71a) ahanuman aramulibite lankabi ʃuiko

ɑ=hɑnumɑn ɑrɑ=em-ulibi-te lɑnkɑ-bi ʃui-k-o

cl 1=Hanuman cl 6=refl-tail-instr lanka-abs burn-fa-dst.pst
‘Hanuman burnt Lanka with his own tail.’

(72) khudi ε=oli-ik εr=εm-ce-k-om

3sg.prox2 cl 5=see-obj cl 2=refl=angry-fa-npst
‘Look at her, she is angry (with herself).’

(73) ŋ=εm maʈh be

2sg=refl run imp
‘You go running.’

Reflexive morphemes such as -εm- generate reflexive verbs, where the

action is self-directed, e.g. εm-boe ‘to marry’, εm-eʈho ‘to recline’, εm-borace
‘get angry with oneself’, etc. Also consider:

(74) ʈh=a =(ε)m eʈh-om

1sg=cl 1=refl recline-npst
‘I am reclining.’

Reflexive forms also generate ambient verbs such as ‘become cloudy’ as in erem-
becom, ‘overflow’ embele.

(c) -iʃo-m

When the reflexive morpheme -εm- is suffixed to the genitive indicat-

ing alienable possession, the initial vowel is dropped. Thus -iʃo + -εm- >
-iʃom.
 pronouns, pronominal and object clitics 185

(75) ʈhu ʈh=iʃo-m julu

1sg 1sg-gen-refl shirt
‘My own shirt.’

(76) ɲe ɲεʃε/o-m ko
2sg 2sg-gen bow
‘Your own bow.’

(77) o iʃo-m biuεr

3sg 3sg-gen candle
‘His own candle.’

(d) -εncuo

This is derived from the combination of the reflexive morpheme -εm-

and the word cuo ‘near.’ The bilabial nasal m changes to a dental n when
followed by non-labial consonant as in cuo. We are not clear about the
semantic status of cuo which is obligatory when the action is self-directed.
The morpheme -εncuo can directly be suffixed to bare pronominals or pro-
nominal clitics.

(78) ʈh=εncuo εt=kaɲo-m

1sg=refl obj=work-npst
‘Let us do the work ourselves.’

(79) ɲ=εncuo εt=kaɲo-m

2sg=refl obj.work-npst
‘You do the work yourself.’

(80) ʈhu-εncuo raʃue

1sg-refl cook
‘I cook myself.’

(81) ɲe-εncuo raʃue

2sg-refl cook
‘You cook by yourself.’

(82) ne-εncuo ut=ʃerep

3pl-refl cl 4=cut
‘They cut (it) themselves.’
186 chapter seven

7.6 Reciprocal Pronouns

Related to the forms given in (b) above, reciprocal forms are derived from
the reflexive -εm- by suffixing it to the appropriate proclitic dictated by
the host verb. Hence, it is a lexicalised form of class marker + reflexive,
such as, ara= + -εm-. The initial vowel of the reflexive form -εm- is dropped
if the preceding class marker ends in a vowel. Consider (a)–(c) given
below. This structure is not different from what we considered above in
reflexive forms. It appears that the same form can designate both a reflex-
ive and a reciprocal meaning. The structure of the reciprocal can be speci-
fied as:

(S 7.1) Pronominal proclitic_reciprocal_V

(a) ʈhu ʈh=ara=m ‘I to myself’ as in ‘It is lonely here I talk to myself.’
(b) ŋu ŋ=ara=m ‘you with each other’
(c) o k=ara=m ‘he/she with each other’

The reciprocal form is prefixed to the verb. Consider:

(83) thire-bi (a)r=εm ceo

child-abs cl 2=rec fight
‘Children fight among themselves.’ / ‘Children fight with each other.’

The sentence given above has an alternate form with reciprocal aram if
the host verb dictates different proclitic as in ara=julo ‘slander’:

(84) thire-bi karam julo

child-abs 3sg.rec slander
‘Children are slandering with each other.’ or
‘Chidren are involved in a slandering match with each other.’

We have a word in our database which translates as ‘each other’ and

seems to be derived from the plural marker ne + object clitic et= and
the reflexive pronoun tuŋ. Thus ne + et + tuŋ > nettuŋ. Consider the
­following:

(85) nettuŋ roʃe t=em-boe

each other love obj=refl-spouse
‘Man and woman love each other.’
 pronouns, pronominal and object clitics 187

7.7 Locational Adverbial Demonstratives

These will be considered in detail in Chapter 8; however, some exam-

ples are cited here to show how these correlate with demonstrative
­pronouns.

(86) ɖi khudi-iʃo boa be

3sg.dis.vis 3sg.prox2-gen land cop
‘That is his land.’ (when the referent is near the hearer)

(87) khudi aka=ʈoŋ-el ʈɔkho ter=khuro be

3sg.prox2 3sg.gen=courtyard-loc tree cl 2=big cop
‘There is a big tree in his courtyard.’

(88) khidi cae be

3sg.prox1 what cop
‘What is this?’

Local demonstratives can occur with a noun in an NP.

(89) khudi a-ʈɔʈa ε(r)=co ie be

3sg.prox2 cl 1-boy cl 2=head pain cop
‘That boy has a headache.’

The first syllable in khudi or khidi can combine with the locative mor-
pheme -lel to give the meaning of spatial adverbs, e.g. khulol ‘there’ and
khilel ‘here.’

(90) khi-el ci-b-e

3sg.prox1-loc come-fa-imp
‘Come here to me.’

(91) khu-ol ŋ-ot= cɔne-b-e

3sg.prox2-loc 2sg=cl 4= go-fa-imp
‘(You) Go there.’
 Chapter eight

Modification including temporal and spatial deixis

Introduction

We shall consider all types of modifiers in this chapter, namely, those that
modify nominal categories and those that modify verbal categories. As far
as deixis is concerned, we will only discuss here those words that occur
independently of verbal morphology, but which modify an action, event
or a state temporally or spatially. Thus we exclude from our discussion
tense marking on verb forms, which are taken care of in Chapter 9 on
verbs. This implies that currently we are concerned with the following:

• Adjectives
• Adverbs
• Temporal deixis
• Spatial deixis

Adjectives and adverbs have similar morpho-semantic features in the

grammar, viz. they can be either dependently or independently marked,
i.e., they are marked by the presence or absence of the body class pro-
clitics as considered earlier in Table 3.2 in Chapter 3. Spatial deixis is
defined as expressions which locate speech act participants in space while
temporal deixis localises the speech event in time.

8.1 Adjectives

8.1.1 Characteristic Features

The following are the characteristic features of adjectives in present-day
Great Andamanese (PGA).

(a) They can function as attributive modifiers of the preceding noun

(1–6).
(b) They can be derived from stative verbs used in a participial construc-
tion (7).
(c) They can also function as predicates (9 and 10).
190 chapter eight

(d) They include not only descriptive adjectives but also quantifiers
and numbers such as ‘only’, ‘some’, ‘single’, ‘many’, ‘first’, ‘third’, etc.
(8 and 11–16).
(e) They can serve as the head of an intransitive predicate with an
optional copula (18, 19).
(f) They are never marked for number.
(g) As mentioned earlier, they are marked for dependency by an appro-
priate body class marker proclitic. If two adjectives are conjoined they
retain their respective proclitics (20).

8.1.2 The Semantic Content

The following are the kinds of adjectives used in the language. Some of
the proclitics are omitted here because they may vary depending on the
nature of the object the modifier is used for.

(1) dimension khuro ‘big’; εlea ‘small’

(2) age cakhamo ‘old’; kɔlɔt ‘new’
(3) value nɔl ‘good’; cae ‘bad’
(4) colour ɖirim ‘black’; cεrel ‘green’
(5) physical property ʈεl ‘hard’; khir ‘hot’; εr=ulu-pho ‘blind’
(6) human propensity aka=boi-tεrcek ‘kind’; er=bobiŋu ‘happy’
(7) speed ercek ‘fast’; ara=cae-caʈo ‘slow (in work)’
(8) similarity teka ‘like’
(9) qualification er=cop ‘correct’
(10) quantification cɔphe ‘many’
(11) position cul ‘near’; tεmic ‘right’; ɖulokara ‘west’
(12) cardinal numbers onʈoplo ‘one’, ‘single’; εr=tubui, nεrtaphul ‘two’

8.1.3 Attributive Adjectives

Let us first consider the attributive types. The quantifiers of various types
also have an attributive function. Adjectives follow the nouns they modify.

(1) sita konaphuŋbi jiyum

sitɑ konɑ-phuŋ-bi iji-om
Sita tendu-ripe-abs eat-npst
‘Sita is eating a ripe tendu fruit.’

(2) ʈhu tajio εtcalo cɔphe ji-k-o

1sg fish a lot enough eat-fa-dst.pst
‘I ate a lot of fish.’
 modification including temporal and spatial deixis 191

(3) golaʈ εr=ulu entoplo nɔl pho be

Golɑt cl 2=eyes single good not cop
‘Golat’s one eye is not good.’

(4) phertajiɖo ut=uŋ bo cay ik=aono

Phertajido cl 4=refl heart bad 3sg=sit.pst
‘Phertajido sat down with a heavy heart.’

(5) u biu ut=calo ut=cɔŋ-o

3sg incense cl 4=heap cl 4=find-dst.pst
‘He found a heap (a lot of) of incense.’

(6) u i=khuɖilo untɔplo i=ebi

3sg cl 5=tiny piece single cl 5=take.pst
‘He took a single small piece.’

As can be seen in example (1) the case marker (if any) is suffixed to the
entire NP. Adjectival meaning is also expressed by the verb used in a par-
ticipial construction indicating a stative situation, where it modifies the
preceding noun. Consider the following sentence.

(7) aʈoʈa lεcik unɖuke untεle

a=ʈoʈa lεc-ik unɖu-k-e untεl-e
cl 1=boy arrow-abs break-fa-stat call-imp
‘Call the boy who broke the arrow.’

These were considered earlier in §5.3.1.5.

Another common strategy is to use a relative marker jukhe ~ ukhe fol-
lowing the modified noun. For details on relativisation one should refer
to §10.8. The relative marker jukhe is also used for specifying the entire NP
as we saw in §5.3.1.5.

8.1.4 Predicative Adjectives

Adjectives including quantifiers can function as predicates. Consider:

(8) ʈɔŋ-el εr=tεc tarakamo

tree-loc cl 2=leaves many
‘There are many leaves in the tree.’

(9) loka ot=bo-tara=lam

Loka cl 4.poss=heart-cl 6.poss=valour/courage
‘Loka is courageous.’
192 chapter eight

(10) sulu-aram laʈ

Sulu-reflx fear
‘Sulu is a coward.’

(11) ʈh=e=ʈɔŋ cɔphe

1sg=cl 5=tree many
‘I have many trees.’

(12) ʈh=ut= thire-n nεrtaphul

1sg=cl 4= child-pl two
‘I have two children.’

(13) ʈh=ut= thire-n narakamu

1sg=cl 4 child-pl many
‘I have many children.’

The cardinal number ‘one’ can function as an indefinite entity or for

expressing the quantity ‘one.’ PGA offers only the two numbers 1 and
2 morphologically. More than one is ‘many.’ However, ordinal 3 exists (15).
Consider:

(14) ʈhire εntɔplo aka=uno-ø

child one 3sg=sit-pst
‘One child sat down.’

The ordinals behave the same way as the cardinals, i.e., they follow the
nouns they modify and can be used as predicates.
(15) ɲo mεxutta:wlu
house third
‘The third house.’

(16) khider tabεco

coconut first
‘The first coconut.’

The predication can also be indicated by an optional copula be as in the

following sentences.
(17) a=thire-nu ta=e=ole nol be
cl 1=child-pl tr=cl 5=see good cop
‘The children look good/well.’
 modification including temporal and spatial deixis 193

(18) a=loka ara=kata (be)

cl 1=Loka cl 6=small (cop)
‘Loka is a dwarf/short.’

(19) loka ot=bo-tara=lam (be)

Loka cl 4=heart-cl 6=courage (cop)
‘Loka is courageous.’

Two modifiers can be coordinated by the conjunct bo.

(20) a=loka er=biŋoi bo ara=kata

cl 1=loka cl 2=fat conj cl 6=small
‘Loka is fat and short.’

8.1.5 Colour Terms

Unlike many hunter and gatherer societies, PGA is marked by the presence
of several terms for colour. The community has added two colour terms
recently (6 and 7) to their list of colours. The following colour terms, like
any other modifier, are used post-nominally and can also serve as predi-
cates. At present there are seven colour terms in the language.

1. cεrel ‘green’, ‘blue’

2. itbi ‘yellow’
3. bira ‘red’
4. ɖirim ‘black’
5. ʈɔl ‘white’
6. liʈiyu:w ‘pink’
7. co:p ‘orange’ (derived from the name of the fruit)

As expected, they all are preceded by class 5 if used for internal attributes
or class 4 if used for external attributes to describe a noun, e.g., rae=/
i=ɖirim ‘black pig’ and ut=ʈɔle ‘blossoming white, very white.’

8.2 Dependency of Modifiers

8.2.1 Proclitics and Modification

The occurrence of modification by a combination of proclitics and modi-
fiers in the language reinforces the fact that attribution explicates depen-
dency between the modifier and the modified. As was discussed earlier
in Chapter 3, this relationship between the modified and the modifier is
194 chapter eight

of an inherent nature or it designates possession of an internalised kind.

Noun modifiers, e.g. adjectives, are preceded by the body class markers
distinguishing seven classes of adjectives within the realm of inalienabil-
ity. The following table gives the semantic range of the proclitics used
with adjectives:

Table 8.1. Body division classes in adjectives

Class Body class markers Semantics Examples
1 a= mouth-related a=mu ‘dumb’,
­attribute a=tutlup ‘greedy’
2 εr= ~ er= external attribute εr=buŋoi ‘beautiful’,
εr=achil ‘surprised’
3 oŋ= ~ on= attributes related oŋ=karacay ‘lame’,
to limbs ‘­handicapped’,
on=toplo ‘alone’
4 ut= ~ ot= negative attribute ot=lile ‘decay’, ot=lɔkho ‘nude’
5 e= ~ i= ~ ε= inherent attribute e=sare ‘salty’, ε=bεn ‘soft’
6 ara= belly-related ara=pheʈkhetɔ ‘big bellied’,
­attribute ara=kaʈa ‘stout/dwarf’
7 o= ~ ɔ= attribute of shape o=baloŋ ‘round’,
and texture o=phelaɲa ‘slippery’

The class markers originally refer to a specific body division. In this func-
tion of modification these are grammaticalised to denote various kinds of
attributes. However, as we saw in §3.7, the degree of grammaticalisation
varied from one class to another. Any noun modified by these adjectives
will have the following structure:

(S 8.1) N_CL=MOD

The examples given above indicate this structure.

Consider the following examples where adjectives are obligatorily pre-
ceded by various distinct class markers representing different meanings.

class 1 aka=, a=, ta=

aka=tutlup ‘dumb’
aka=mu ‘greedy’
aka=lekɔ ‘dead’
aka=kel ‘miser’
 modification including temporal and spatial deixis 195

a=thie ‘heavy’
aka=mu ‘dumb’ (Bo)
ta=iʃom ‘soft, slow’

class 2 εr=, tεr=

er=ulu-to-phui ‘alive’ (literally: flower of eyes)
er=ceotarakamo ‘enraged, angry, mad’
εr=buŋoi ‘beautiful’
er=khurɔ ‘older than someone, elder, huge’
er=biŋui ‘fat’ (Bo)
er=pholkɔe ‘bigamous’
ter=ʈɔ:lɔ ‘broken’ (Bo)
erɔsemɔ ‘delicious’
er=boi ‘obedient’
er=lobe ‘open’
εr=cara ‘wood cut into small pieces’
εr=buŋui ‘(be) happy’
εr=ʈɔbo ‘blind’ (Bo)
er=alcabu ‘tongue tied’
er=a=cil ‘surprised’
er=ɲum ‘sleepy’

class 3 oŋ=, toŋ=, un=

un=tɑbol ‘naughty’
on=toplo ‘all alone’
un=tepuc ‘alive’
oŋ=karacay ‘lame’, ‘handicapped’
tuŋ=kεlo ‘all around’

class 4 ut=, ot=, tut=, tot=, u=

kata ‘a piece’
ot=kata ‘a dwarf’, ‘a midget’
ut=lile ‘deaf’
ot=catɲa ‘adopted’
ut=širo-cerel ‘bluish green colour of the sea’
ut=cire ‘clean by washing’
uj=jete < ut + jete ‘shy’ (person)
ut=khuduŋ ‘heavy in uplifting’
ut=belo ‘wide’
ot=bata ‘short’
ot=cɔ ‘right’
ut=bo ‘reverse’
ot=kobɔlo ‘bald person’

class 5 e=, te=, i=, ti=

ε=nol-pho ‘bad’, ‘not good’
ε=kɛrɛk ‘bitter’
e=ɖirim ‘black’
196 chapter eight

e=kokhela ‘blunt’
i:=ople ‘full of light’
e=phuŋ ‘fully ripe’
e=khui ‘hard’
i=khimil ‘hot’
i=boe ‘boiled’ (Bo)
e=liu-sʃɔŋɔ ‘brave’ (Bo)
et=to:ʈte ‘clean grass’
e=ta-julu-e ‘cool it’, ‘cold’ [resultative]
e=cay ‘dirty’, ‘bad’
e=ʈɔlɔʈ-mo ‘fair’, ‘white’
i=belo ‘flat’
i=let-amo ‘flat as in levelled’ (Bo)
ɛ=tɔk ‘frail’ (Bo)
i:=ople ‘full of light’
e=phuŋ ‘fully ripe’
e=khui ‘hard’
e=kɔʈɔbo ‘hollow’
i=khimil ‘hot’
e=cɛrʈɔk-totla ‘light brown colour’
e=co-loboŋ ‘long headed’, ‘clever’, ‘intelligent’
e=sɑre ‘salty’
ε=lɔr ‘mixed taste’ [salty-sweet-sour]
e=phae-tec ‘dried leaves’

class 6 ara=, tara=

ara=tɔm ‘elderly’, ‘old’ (Bo)
ara=ka-ter-cek ‘lazy’ (Bo)
ara=kaʈa ‘short’ (person)
ara=betta ‘ancient’
ara=pheʈ-khetɔ ‘big-bellied person’
ara=jio ‘be scared of something’

class 7 o=, Ɔ=, to=, tƆ=

o=baloŋ ‘round’
o=laltoio ‘stagnant’
o=cae-ca-ʈopho ‘useless’
o=phelaɲa ‘slippery’

Modification by possession can also be achieved with nouns serving as

an R of the following adjective D with an intervening body class marker.
These were discussed in detail in Chapter 5.
We can safely claim that class markers have their own semantics
which are carried over the grammatical category they are attached with.
Consider the attribution of a dependent noun, such as an edge of some
 modification including temporal and spatial deixis 197

sharp object. These nouns are obligatorily preceded by a proclitic signify-

ing the type of modification.

Table 8.2. Edge and side of an object

e=burɔŋo ‘edge of the seashore’ [cl 5 = N]
ʈhi burɔŋo ‘edge of the land surface’ [N N]
ir=juxu ‘edge of an arrow’ [cl 2 = N]
e=traxo ‘edge of a bow’ [cl 5 = N]
ir=fide ‘edge of a rudder’ [cl 2 = N]
ir=file ‘edge of a weapon’ [cl 2 = N]
tot=taŋo ‘edge, side’ [cl 4 = N]

8.2.2 Non-Dependent Modifiers

So far we have considered dependent modifiers, which are obligatorily
attached to the proclitic. There are others which are not obligatorily pre-
ceded by a body division proclitic. One can speculate that during the
evolution of the language the markers became lexicalised and hence,
the presence of a class marker is no longer transparent in its morphemic
break. Another hypothesis is that these proclitics were dropped from the
language due to the language contact situation. However, the indepen-
dent modifiers occur infrequently. Consider the following examples.

intajionɔl ‘tasty’
lephai ‘thirsty’
phinli ‘throbbing [as in pain]’
bekha ‘useless’
eŋeʈhe ‘very dirty’
etino ‘wet’
ʈaŋʈɔ ‘lean, fatless’
mo ‘small’
tec ‘diminutive’
ilphe ‘indifferent’
totcyo ‘very’

8.3 Manner Adverbs

Adverbs or modifiers of verbs can either precede or follow the verb. This is
not an option with adjectives as the position of the latter is fixed in the lan-
guage, i.e. post-nominal. The most common adverbs are those that describe
an action, namely, manner adverbs. Others fall under the ­paradigm of
198 chapter eight

‘space’ and ‘time’ which are considered in §8.4 and 8.5. Adverbs of manner
follow the verb and may fill the predicate slot. Alternately, they may pre-
cede the finite verb. They also have the potential to occur without a verb
as the meaning of the word includes the meaning contained in the verbal
category (24–26). Adverbs are underlined in the following examples.
(21) ʈhiyo noɳʈokha nu-bi boʃuʈ-ø nol
1sg.emph Jarawa people-abs hit.pst well
‘I bashed Jarawa well.’

(22) buro erentεʃu iji-k-om

Buro slowly eat-fa-npst
‘Buro eats slowly.’

(23) ʈhu ŋɔʈ -bit cɔkho-m

1sg swim-abs.obj good/well-npst
‘I can swim well.’

(24) o ʈh=ɔŋ=ʈo εk=tεrnin-o

3sg 1sg=cl 3.poss=arm obj=hold tightly-dst.pst
‘He held my hand tightly.’

(25) o- ʈh=ɛ tɛr=ninucoic- ø

3sg 1sg=dat cl 2=embrace fiercely/hold tightly-pst
‘He embraced me fiercely.’

(26) enliɑ-k-om
(walk) slowly-fa-npst
‘He walks slowly.’

Some of the manner adverbs are listed below. Proclitics are not attached
to all of them. Where the words have proclitics they are shown by the
symbol =. Some proclitics in manner adverbs are grammaticalised and
hence it is difficult to segregate them from the root morpheme.

ikɑɲɔrɔ-k-e ‘completely’
ɛren-cɛthe ‘angrily’
ekɑ=jirɑ-cɛrel ‘greedily’
eren-tɑ=liu ‘inadvertently’
ɛr=ɔ=pili-phu ‘incessant’
nɔle ‘properly’
e=lia ~ ɑ=leɑ ‘slowly’
ɛr=ʈɔko ~ tɑ=iʃom ‘slowly’
kɑlemo ‘soon’; ‘fast’; ‘early’
itthɑxe ‘still’
er=ɛntɑ-lie ‘unknowingly’
 modification including temporal and spatial deixis 199

Except in a few instances, manner adverbs are not obligatorily attached to

proclitics as they are mostly independent words.
Manner of an action can also be expressed by the nature of the proclitic
attached to the verb root as it was observed in §3.3.3, §3.5, and §7.3.2.2.

8.4 Temporal Adverbs and Temporal Deixis

8.4.1 Subjecthood and Temporal Adverbs

We considered in the previous chapter in §7.3.3 that a major feature of tem-
poral adverbs designating ‘morning’ and ‘evening’ is that they are ­preceded
by the pronominal proclitic co-referring to the subject, a rare phenomenon
in the linguistic literature. Thus, ambikhir ‘morning’, ‘day’ has the follow-
ing forms depending upon whether the subject of the sentence is the first,
second or third person singular pronoun.

(i) ʈh=ambikihir ‘tomorrow’ (spoken by the speaker referring to himself as a

participant of the action)
(ii) ŋ=ambikhir ‘tomorrow’ (spoken by the speaker to the addressee)
(iii) ak=ambikhir ‘tomorrow’ (spoken by the speaker for a person other than
the addressee)

Another feature worth noting is that the word for morning refers to ‘yes-
terday’ as well as to ‘tomorrow.’ It is the tense marking on the verb that
clarifies the meaning. The word ambikhir appears to be the lexicalised
form of the ‘reflexive pronoun’ + ‘morning’, i.e. am + bikhir. Refer to §7.5.

(a) ʈh=ambikhir 1sg-‘morning, day’

(b) ŋ=ambikhir 2sg-‘morning, day’
(c) aka=ambikhir 3sg-‘morning, day’

but

(d) *nu-ambikhir 3pl-‘morning, day’

(27) ʈh=ɑmbikhir jɑipur-e ʈh=ut=cone-pho be

1sg-morning Jaipur-abs 1sg=cl 4=leave-neg cop
‘I will not leave for Jaipur tomorrow.’

(28) ɑ=sulu ɑka=ɑmbikhir sʈreʈ-ɑk ot=cɔne b-ɔ

cl 1=Sulu 3sg=morning strait-dir cl 4=leave-fa-dst.pst
‘Sulu left for Strait yesterday.’
200 chapter eight

The time before the ‘present’ or after the ‘past’ is expressed by affixing the
words for ‘before’ and ‘later’ to the word ambikhir. In forming the words
for ‘the day before yesterday’ or ‘the day after tomorrow’ the construction
is very transparent. Consider:

(29) tɑrɑ=sulu-ʈh=ambikhir
cl 6=earlier-1sg=morning
‘Morning before yesterday’s morning.’

(30) ʈhu julu-ta=ambikhir

1sg early morning-cl 1=morning
‘Morning after tomorrow’s morning.’

8.4.2 Syntax of Words Indicating Time

Words such as tεrkoɔto and ɖiɖek are used both for ‘yesterday’ and
‘tomorrow.’ The word for ‘day break’ is otɔ ~ oto and the generic word
for ‘morning’ is ɖiɖek. The word otɔ-tut=bo literally: ‘anterior to day
break’, or ‘day behind me’, i.e., ‘yesterday’ is derived by the body part
term ut=bo ‘behind.’
Temporal adverbs can fill in the predicate slot with the appropriate
tense category.

(31) nu bɑt-il jo-bi uru-ottɔ-k-o

Nu night-loc sing-abs sang-morning-fa-dst.pst
‘Noe sang the whole night till morning came.’
or
‘It was morning before Noe finished singing’

The position of temporal adverbs is not fixed and they can move to the
beginning of the clause or a sentence. They can behave like nouns taking
case markers1 (underlined words).
(32) ko-il ʈh=ɑ=meli-b-om koic
later-loc 1sg=cl 1=return-fa-npst again
‘I will come back here again later.’

1 This pattern is very much like Contact Hindi where case marking can be suffixed to
temporal adverbial words, e.g., dopahər men ‘in the afternoon’, saal men ‘in the year’ and
əndhere men ‘in the dark/night.’
 modification including temporal and spatial deixis 201

(33) julum-te ʈh=ut=cone-b-om

tomorrow morning-abs 1sg=cl 4=go-fa-npst
‘I will go early tomorrow morning.’

(34) tɑŋto-te ʈhu-ɑom

year-abs 1sg-come.npst
‘I come every year.’

(35) tɑrɑ=sulu-ʈɑŋtɔ-e ʈhitphile-k-o

cl 6=later-year-abs tsunami-fa-dst.pst
‘The tsunami came last year.’

(36) ŋɛʈɛŋ-il ʈh=ɛm lɑʈ

darkness-loc 1sg=refl fear
‘I fear the dark.’

(37) terkɔ-otɔ ʈɔrɔ-ɑk ʈh=ut=cɔne-b-ɔm

day after tomorrow shore-dir 1sg=cl 4=go-fa-npst
‘I am going to the beach the day after tomorrow.’ (otɔ ‘day break’)

(38) terkɔ-otɔ porʈblεr-ɑk ʈhu thu-ɔ

day before yesterday Port Blair-dir 1sg reach-dst.pst
‘I reached Port Blair the day before yesterday.’

(39) tɑrɑ=sulu-ɖiɖek-il ʈhu cɔkbi-bi i( ji)-k-ɔ

cl 6=earlier-day-loc 1sg turtle-abs eat-fa-dst.pst
‘I ate a turtle yesterday.’

(40) urɔʈɔe-tɑ ɲo-ʈhik ɲo

ancient time-abl house-land.obj live
‘He has been living in the house from ancient times.’

(41) ʈh=ɑrɑ=in(ci) jululu jicɛr-bi ɑo-Ø

1sg=cl 6=leave before rain-abs come-pst
‘As I was about to leave, the rain started.’

(42) tekhɑmo ŋ=ot= cɔne-be

immediately 2sg=cl 4=go-imp
‘You go right away.’ or ‘Leave immediately.’

The temporal adverb may be moved so that it occurs in the middle of the
clause.
(43) ʈhu ino-bi ɖikhɔmɑ i=khue
1sg water-abs already cl 5=drink.imm.pst
‘I have already had water.’
202 chapter eight

8.4.3 Compounding and Lexicalisation

The lexicalisation of inalienability markers produces compounds of tem-
poral adverbs. Not all constituents are transparent in their meaning.

(a) bɑt-ter=bel
night-cl 2=overflow
‘Late night.’

(b) ʈh=ut=ɖiu
1sg=cl 4=sun
‘Whole day.’

(c) it=ʈhɑxe
obj=1sg.now
‘Till now.’

(d) bɑt ɑtɑ=mikhu

night cl 2=middle
‘Middle of the night.’

(e) meŋ-er=co-ɖiu
1pl.incl-cl 2=head-sun
‘Midday’ Literally: ‘Sun on our head’

(f) ɛ=oro khui-l

cl 5=flower drink-loc
‘In youth’; ‘while young’ (in the stage of drinking nectar)

8.4.4 Multiplicity of Temporal Deixis

It is not unusual to find several distinct forms for the same temporal refer-
ence as the speakers came from different lineages as discussed earlier in
Chapter 1. We have noted all the variations in the following table.

Table 8.3. Multiplicity of temporal deixis

English gloss Andamanese form
ancient times urɔʈɔe
ancient times purɔʈɔe
daily, often ʈa:ɲorɔ
just now khelto
now it=tɑkhe
now dekhɔ
now khɑmu
prior to aka=sulu
today, now ɖiɖek
 modification including temporal and spatial deixis 203

8.4.5 Celestial Bodies as Temporal Markers

The Great Andamanese, like many other societies of hunter-gatherers, dis-
tinguish the day and the month on the basis of the appearance of the ‘sun’
ɖiu and the ‘moon’ ɖulo. Consult the following table.
The Great Andamanese do not use specific days of a year as deictic cat-
egories. Hence, there are no names for the days of a week such as ‘Sunday’,
‘Monday’, etc.
Other ways of marking time in PGA are based on the honey calendar
and the naming seasons which are based on the seasonal blossoms. This
topic has already been discussed in §1.9.2.

8.4.6 Temporal Categorisation and Hunting and Gathering

Apart from measuring and categorising time on the basis of natural events
like the blossoming of seasonal flowers, the movement of the sun, etc., the
Great Andamanese also categorise time on the basis of the availability of
food resources and the best time for hunting an animal or fish. We were
able to elicit some names for periods of time that were either coined after
a blooming flower or fruit or based on the availability of food.

Table 8.4. Terms for direction and days of a month

ɖiu-bala ‘afternoon’ Lit. ‘sun-arm’
ɖiu-beraceʈo ‘evening’ (around 3–4 pm)
ɖiu-bet-oncɔlɔ ‘solar eclipse’
ɖiu-bik-a-thu ‘around sunrise time’ Lit. ‘born of sun’
ɖiu-kara ‘sunrise, east’
ɖiu-karale ‘sunset, west’
ɖiu-mɛŋer-co ‘mid-day sun’ Lit. ‘sun above our head’
ɖiu-tara=bat ‘dusk’
ɖiu-tara=cɔl ‘sunlight’
ɖiu-tara=le ‘sunset, direction of sunset, west’
ɖiu-temec ‘focus of sun towards Port Blair’
ɖiu-ter=bec ‘overcast sky’ Lit. ‘hiding of the sun in the clouds’, ‘hair of the
sun’
ɖulo ~ ɖulɔ ‘moon, moonlight, month’
ɖulo-kan-tuse ‘fortnight’
ɖulo-kara ‘setting of moon’, ‘west’
ɖulo-phƆrƆk ‘full moon’
ɖulo-tarale ‘rising of moon’, ‘east’
ɖu:lo-tara:wto ‘third quarter of moon’
ɖu:lo-thirɛ ‘first quarter of moon’ Lit. ‘moon-child’
ɖulo-beton-cɔlo ‘lunar eclipse’
ɖulo-tara=cɔl ‘moon light’
ʈhi-ter=ɔtɔ ‘early morning, day break’
204 chapter eight

Table 8.5. Names of flowers and hunting time

Great Andamanese flower Relevance for hunting in the sea or in the forest
names
kɔrɔiɲ-ɔrɔ The period when the fat content in turtles and fish
is high
bɔp-ʈɔlɔ The period when bikhir, lioʈ and bere fish are found
in abundance
loro-ʈɔlɔ The period best for catching phiku and ɲuri fishes
cɔkhɔro-ʈɔlɔ The period when pigs have maximum fat. Therefore
the best time to hunt pigs.

Man (1883: 144) writes about the Great Andamanese belief that the best
time to hunt turtle or fish is during the ebb between dawn and the rising
of the waning moon. This period is therefore named aks-tig-pala in the
Âkà-Bêa language. Similarly Man (1923: 182) writes about the best time to
collect shell fish known as astoya in Âkà-Bêa.
The parameters of temporal categorisation have also been discussed in
§1.9.3.1 and §1.9.3.2 and should be referred to to assess the ways the Great
Andamanese measure time historically.

8.5 Spatial Deixis

Space occupies a pivotal role in Great Andamanese life which is encoded

in the language. Although the speakers are no longer hunter-gatherers, the
language they use reflects the life of the forest and the sea-shore dwell-
ing society. Spatial deixis is defined as expressions which locate both the
speech and narrated participants in space. Two significant aspects that
are encoded in the language is the idea of proximity of speech and nar-
rated participants and the visibility of the latter. These two aspects were
also considered to some extent in the previous chapter. We will concen-
trate on other spatial aspects which are encoded in the language. Most
of the speakers of PGA are descendants of coast dwellers, and thus there
are many elaborate terms for the seascape as shown in Table 8.6. Terms
that are related to land and forest are equally elaborate as can be seen in
Table 8.7.
 modification including temporal and spatial deixis 205

Table 8.6. Levels of the sea

ʃiro ‘sea’
ʃiro-tεr-likhu ‘deep sea’ (literally: ‘depth of sea’)
ʃiro tεr kεra ‘shallow sea’ (literally: ‘descend of sea’)
ʃiro tεr tεk ‘away from the coast’
ʃiro tεr cεk ‘right in the centre of the sea’ (literally: ‘much of the sea’)
ʃiro tara cεrεl ‘open sea’ (literally: ‘green of sea’), Blue ocean
ʃiro cεrεl ‘green/blue sea’
tot-pholo ‘seashore’ (literally: ‘emerging from the waves’)
burukhu-tara-boʈho ‘end of the rocky region’ (literally: ‘fall of rocky structure’)
sare ‘knee deep water or beginning of the sea’
kera ‘beyond sare with waist-high water’
ʈɔrɔ-burɔŋo ‘sandy beach’

8.5.1 Distance and Direction

A short list of seascape terms is given here to illustrate how conceptual
dependency and inalienability are intertwined to refer to distances and
directions of a space with reference to the self. The word order in the
following constructions is the same as in adjectival constructions, i.e., the
modified followed by modifier. The language uses single words as well as
larger constructions, those which are modified by possession.
Many locational modifications are specified either by compounding
or by juxtaposing two nominals with an intervening body division class
marker. Some of the constructions modify nouns by specifying the space
of the object and operate in the spatial deictic category. Consider the fol-
lowing table.

Table 8.7. Landscape and coastal terms

Inland PGA English gloss
domain
Land ʈhi ‘Land’, ‘earth’
ʈhi-tara-lε ‘Decline’ (literally: descend/ decrease of land)
ʈhi-tot-kara ‘Incline’ (literally: ascend/increase of land)
Forest ʈhi- mikhu ‘Forest’
aro ‘Light forest’
ʈhimikhu ŋero ‘Dense forest’
ʈhimikhu-ʈƆŋ-phoŋ ‘Heart of the jungle’
ʈottaŋo ‘Sides of the forest’
ʈhi-tɔt=be-ta=pho ‘Opening in forest with little undergrowth’
ino-ter=phoŋ ‘Fresh water hole’
206 chapter eight

Table 8.7 (cont.)

Inland PGA English gloss
domain
ʈhi tara=ta ‘Levelled ground’
ʈhi-tot=kɔrɔbo ‘Dried land’
ʈhi-tot=belo ‘Clear ground’
ʈɔrɔ-boa ‘Sandy ground’
Island ʈut ɖello ‘Small island’
ʈhi-tot=kata ‘Small islet’
Coastal buliu ‘Creek’
bule ‘Big creek’
tara cara ‘Small creek’
kara ŋik ‘Sea shore area’
bucɔʈɔŋ ‘Mangrove area’ (named metonymically after
mangrove tree)
Hill buruin ‘Hill’
buruin-to=kara ‘Steep of hill/climb hill’
buruin tara=le ‘Slope of hill/descend from hill’
buruin tom=phoŋ ‘Foot of the hill’
buruin leo ‘Small hillock’

8.5.2 Vertical and Horizontal Space

The Great Andamanese divide the space in a house or any dwelling ver-
tically and horizontally. Unlike in the case of the sea-space, some body
division terms are used in defining the parts and space of a house ɲo. In
other words, the conceptualisation of the house space is anthropocentric.
The Great Andamanese use human categories to describe non-human
ones. Consider the following table. Only those body part terms which
are transparent are translated. The individual meaning of the others is
not known.

Table 8.8. Spatial deixis describing parts of a Great Andamanese house

Specific part of a English gloss
house
ɲo-tara=taŋ ‘house-cl 6=‘top’ ‘Roof of the house’ (elevated)
ɲo-tara=teŋ ‘house-cl 6=‘arm’ ‘Roof of the house near the ground’
ɲo-phoŋ ‘house’-‘mouth’, ‘Entrance of the house’
ɲo-ta==cɔkho ‘house’-cl 1=‘face’: ‘Front of the house’
ɲo-ter=olot ‘house’-cl 2=‘octopus’, ‘Openings in the screens through
which smoke goes out’ (now used for windows)
 modification including temporal and spatial deixis 207

Table 8.8 (cont.)

Specific part of a English gloss
house
ɲo-ta=tecɔ ‘house’-cl 1=‘trunk’, ‘Middle post of the house’
ɲo-ta=ʈɔkho-phareɔ ‘house’-cl 1=‘wood-mat’, ‘Platform on which roofing mat
is spread’
ɲo-ta=khuro ‘house’-cl 1=‘shoulder’, ‘Slope of a roof’
ɲo-kotra ‘Interior of a house’

Although modern Great Andamanese houses are no longer made with slop-
ing roofs touching the ground, some terms listed in Table 8.8 are still used.

8.5.3 Locational/Spatial Postpositions

Locational or spatial postpositions are derived by (1) juxtaposition of one
or multiple class markers and the locative marker -l; (2) a single word
specifying the space; (3) a body part term juxtaposed to both a class
marker and a locative marker -l; and (4) by using the directional suffix
-ak. Most of the spatial postpositions are derived by the first method. As
in the case of temporal deixis, the spatial deixis is specified with reference
to the self. Among the body part terms ‘face’, ‘head’, ‘back’, ‘front’, and
‘waist’ are used most often to derive spatial deixis. Thus, εr=karap ‘waist’
but a=karap ‘behind someone.’ The extremities of the body are never used
for deriving spatial deixis.
In the following sentence the body part terms ‘face’ and ‘back’ are suf-
fixed by locational suffixes to express locational deixis.

(44) tεr=cɔkhe-l th=ut=ɲyo ʈɔŋ jiyo

cl 1=face-loc 1sg=cl 4=house tree exist
‘There is a tree in front of/behind my house.’

(45) ŋut=bɔ-l ʈhu ʈh=aono-m

2sg.cl 4=back=loc 1sg 1sg=sit.npst
‘I am sitting behind you.’

The table presented below lists different locative words in PGA. The suffix
-l indicates the locative marker. Some of the forms are lexicalised in the
current form of the language. Readers will notice multiple words for the
same concept as the words are drawn from different speakers who are
from different lineages as discussed in the first chapter. Also refer to case
in sections §5.2.2.9 and §5.2.2.10.
208 chapter eight

Table 8.9. Spatial adverbs and their morphemic composition

English gloss Andamanese form Morphology
above the head ut=kɔt-co cl 4=above-head: ‘heaven’
across εr=phεʈ cl 2=swell
at, in the direction of -ak deixis
away ta=cakhu-l cl 1=face-loc
before, preceding a= tujulu cl 1=early
behind ut=bɔ-l cl 4=back-loc
behind ut=khum cl 4=shoulder
behind someone a=karap cl 1=waist
behind something ara=balo cl 6=creeper
far away, last one tara=thuro cl 6=exit
front of a=tɔŋ cl 1=courtyard
having brought the boat kotεʃil deixis
to the side
having turned back bɔ-t-khɑcol back-obj-deixis
immediately under tara=te=ono cl 6=cl 5=forearm
in front of tεr=cokhe-l cl 2=face-loc
in the middle of something tekhil deixis
inside e=kotra cl 5=inside
near cul deixis
near e=ʈopho cl 5=deixis
near belo:pok deixis
near (touching) ε= cua cl 5=near
on tut=ara:=l cl 4=cl 6=loc
on something in contact tara=l cl 6=loc
other side (of jungle/sea) tεr=phεʈ cl 2=swell
side (of) X ε=phεʈ-εr= cl 5=swell-cl 2=
towards the land o=kara cl 7=deixis
towards the sea ot=le cl 4=deixis
under tumbo-l down-loc
under something tara=ta=l cl 6=cl 1=loc
upstairs suŋo deixis
very far kɑrɑ-it=pho-be sunset-obj=neg-cop
very far kɑra-tɑ=teŋo sunset-cl 1=?

Some examples:

(46) ɲo-khuttrɑl mo beno-b-ɔm

house-inside we sleep-fa-npst
‘We sleep inside the house.’

(47) a=buli ʈɔrɔ-ak ut =cɔne-k-ɔm

cl 1=buli beach-dir cl 4=go-fa-npst
‘Buli goes to the beach.’
 modification including temporal and spatial deixis 209

(48) kɔrɔiɲ sare tara =ta=l meo-phoŋ ʈhi bik ɲo-k-ɔm

Dugong sea cl 6=cl 1=loc stone-hole live abs.obj live-fa-npst
‘Dugong lives in the stone-cave of the sea.’

The various body division class markers explicate various spatial relations
as considered in Table 8.10.

8.6 Body Division Classes and Adverbs

As should be clear by now, adverbs of various types and some deictic

markers can be preceded by proclitics. Proclitics are highly grammatica-
lised in an adverbial function. Below are listed seven classes of adverbs
each preceded by an appropriate proclitic and its semantic character-
istics. On many occasions the same class marker signifies more than two
adverbial functions, such as ‘manner’ and ‘space’ or ‘time.’ This has been
noted in Table 8.11. These markers are obligatorily attached to the fol-
lowing words which assign the status of dependency to the modifiers of
action, time and space.
It can be seen that various deictic meanings, whether of space or time,
are signified by a specific body division proclitic. Although a= or ta= is
a body division class marker which refers to the body part term ‘mouth’,
the grammaticalisation of the term gives the manner adverb, a meaning
of anteriority or spatial deixis of front and back when it precedes a word
of action or state. Similarly, directional meaning is conveyed by using cl 4
ut= which originally refers to bodily products.

8.7 Third person Demonstrative Pronouns and Deixis

We would like to recall that third person pronouns offer the maximum
variety of forms. See §7.1. One important parameter that was discussed in

Table 8.10. Body division classes designating spatial relations

Class Body class Body division Spatial relations Reference points
markers
1 a= mouth cavity surface ‘front’
2 εr= face anterior, exterior ‘front’, ‘out’
4 ut= body products posterior, superior ‘up’
5 e= internal parts interior, centre ‘in’
6 ara= sides periphery ‘edge’
7 ɔ= lower parts inferior ‘down’
210 chapter eight

Table 8.11. Body division classes in adverbs

Class Body class Semantics Examples
markers
1 a= deixis of front or back; a=karap ‘behind’,
anteriority of an action a=-kaulu ‘prior to’
2 εr= ~ er= deixis of adjacency er=beʈʈo:fo ‘adjacent to/near X’
3 oŋ= ~ on= haste or hurriedly done oŋ=kocil ‘fast’, ‘hurriedly’
action
4 ut= ~ ot= directional deixis ot=le, ‘seaward’,
ot=bo ‘backwards’
5 e= ~ i= ~ ε= deixis of internal space te=khil ‘in the middle’,
e=kotra ‘inside’
6 ara= deixis of immediate vertical ara=balo ‘behind X’,
or horizontal space tara=tal ‘right under X’
7 o= ~ ɔ= temporal deixis o=ʈɔ: ‘day break’,
o=kara ‘sunset’

Chapter 7 was that the forms were distinguished on the basis of distance
and visibility. The distant but visible object is represented by three distinct
forms depending upon the number of the object in question (Table 8.12).
The distant but invisible objects are represented by two different forms
depending on the singular and plural number of the objects under study.
The subject or object third person is generally cliticised as u or o. Thus, u
~ o refers to any person who is not visible and is equivalent to the English
referent ‘he’, while ɖune ~ ɖuni is reserved for identifying the third person
plural referent ‘they.’ The accusative and dative form of this is unni.
The other category that plays a role in deciding the third person form,
as mentioned in the previous chapter, is the proximity to the addressor. If
the object is very close to the speaker, say, almost at arm’s length, and is
able to be touched, the form khidi is used. However, if the object is closer
to the hearer than to the speaker, the form would be khudi if singular, or
diya and its variant forms if plural.

(49) diya/khudi-no cae bi caʈo-k-o

3pl.prox2 what abs do-fa-dst.pst
‘What were these (people) doing?’
 modification including temporal and spatial deixis 211

(50) khidi ŋ=iʃo julu bireŋ be

3sg.prox1 2sg=gen dress colourful cop
‘This is your colourful dress.’

It should be noted that the plural forms for proximate and distant visible
objects were only attested for one speaker as other speakers no longer use
them. Demonstratives are often identical to locative pro-forms and usu-
ally make a two-way distinction based on either proximity to the speaker
or the relative distance based on the visibility of the object, or the prox-
imity to the hearer. The form marked for intermediate distance, i.e. the
one used for ‘closer to the hearer’, is the least marked and often serves
as the relative clause marker or occasionally a third person pronoun. In
all, the language provides eight different forms of demonstratives given in
Table 8.12. Some of them have more than one variant form.
Locational case suffixes and location words such as tuttara:l ‘on’, ‘on
top of’ and kuttara:l ‘inside’ have already been discussed in §5.2.2.10. Also
refer to Table 5.2 given in Chapter 5. In all, deictic categories are very
complex and are derived by multiple processes. They represent the inter-
play of anthropocentrism and the perception of distance and time by the
Great Andamanese.

Table 8.12. Demonstrative pronouns

Deictic distance Singular Dual Plural
Proximate1 khidi —
closer to speaker
Proximate2 khudi — khudino ~ dia ~ diya
closer to hearer
Distant ɖi ɖuini ɖunio
visible
Distant ɖu ~ ɖuio ~ u ɖuni ~ ɖune
invisible
 Chapter Nine

The verb and verb complex

Introduction

The verb in present-day Great Andamanese (PGA) occupies the sentence-

final position and consists of several morphemes, which are prefixed,
infixed and suffixed to the root morpheme. There is no agreement in a
prototypical sense, although direct object agreement exists in the form
of object clitics with all transitive verbs as discussed earlier in Chapter 3
and in the sections §7.3.2 as well as in §9.3. However, it was observed in
§3.3.3 and §3.5 that these verbal clitics are intertwined with the manner
of action. Refer to various forms of ‘cut’ and associated various object
clitics in §7.3.2.2 representing different ways of ‘cutting.’ Pronominal
subject agreement appears on the temporal adverbs as seen in §8.4, but
not on finite verbs. These are not prototypical cases of the agreement
phenomenon.
There are also reflexive verbs generated by adding the reflexive mor-
pheme Ɛm- to verb roots, e.g. Ɛm-boe, ‘to marry’, Ɛm-eʈho ‘to recline’,
Ɛm-borace ‘get angry with oneself ’, etc. as seen in §7.5 and reciprocal
verbs such as eren-ceo ‘to fight’ as discussed in §7.6.
In short, a verb root may or may not be bound. All bound forms are
obligatorily preceded by one of the seven body division class markers or
object clitics. In addition, verbs may optionally be preceded by a valency-
indicating morpheme, such as the causative or applicative, or by a reflex-
ive morpheme indicating self-directed action. If the verbal root morpheme
ends in an open syllable, a formative affixal consonant -b- or -k- or -l- is
infixed between the verb root and the following mood or tense marker.
This can be illustrated in a schema. Aspect markers are added directly to
the verb root without the formative affix.

(S 9.1)

(proclitic) (valency) (reflexive) verb root ([formative affix])

(mood/aspect) [tense]
214 chapter nine

Thus:

(1) a=narayan Ɛr=ŋol-om

cl 1=narayan cl 2=write-npst
‘Narayan writes.’

(2) ŋu ek=terol-e
2sg obj=sell-imp
‘You sell it.’

(3) thire rƐfe-bi iji-k-o

child rice-abs eat-fa-dst.pst
‘The child was eating rice.’

(4) oco-ta=cav-om
net-appl/caus=spread-npst
‘(He) casts the net.’

(5) ɑ= thire-n cɑe-b-o

cl 1= child-refl rot-fa-dst.pst
‘The child was aborted.’

A large number of verbs are individuated by the body division class marker
proclitics, in which body part semantics shift into event type semantic
categories of various kinds. These proclitics combine with verbal roots
of any valence. These are very significant features of the language. Let us
discuss them in detail.

9.1 Verbal Proclitics

As we observed in Chapter 3, PGA has a dual semantic system for body

part categorisation. Further, we have seen that body part semantics per-
vade the lexical and grammatical system of the language as this dual sys-
tem is extended to other form classes, viz. verbs, adjectives and adverbs.
Here we are concerned with the relationship of these class markers
with verbs.
Although most of the verbs in the language are obligatorily preceded
by these body division class proclitics, a few verbs in the language appear
as free forms. Thus, PGA offers both bound and free forms of verbs. The
seven body division class markers are grammaticalised to a large extent,
but the original semantics of each of the body division terms is retained
by some. The following table gives all the class marker proclitics for the
 the verb and verb complex 215

Table 9.1. The body division class markers with verbs

Class Body division Semantics of classes Examples
class markers and verb forms

Class 1 a= mouth-related a=jire ‘abuse’,

activity, origin, a=kɔpho ‘sprout’
name
Class 2 Ɛr= ~ er= action involving the era=luk ‘weigh’
front part of the
body
Class 3 oŋ= ~ on= hand-related oŋ=cho ‘stitch’,
activity un=tujuro ‘trembling of
hands’
Class 4 ut= ~ ot= directional, ot=cone ‘leave’,
experiential ʈheʈhe-bom ‘be hungry’
Class 5 e= ~ i- ~ Ɛ- action involving e=lƐco ‘suck’,
the interior of an Ɛ=rino ‘tear’
object
Class 6 ara= action involving ara=ɖelo ‘be pregnant’
middle and back
portions of the
body
Class 7 o= ɔ = resultative state o=cɔrno ‘make nest’, o=beo
‘sting’

verbs. While the semantics for class 1 are nearly transparent, class 7 is
highly grammaticalised. Not all seven are grammaticalised to the same
degree.
Let us consider verbs of all valencies, intransitive, transitive and caus-
ative, and their association with the body division class markers.

9.2 Intransitive Verbs

The body division class markers attach to intransitive verbs designating

various psychological predicates, experiences, and states. Some intransi-
tive verbs are attached to body division class markers explicating a world
of event and state which can be considered on a par with the distinctions
made on the scale of inalienability with regards to the body part terms.
216 chapter nine

Hence, verbs with a proclitic ot= ~ ut= originally signifying ‘genesis’ or

‘emerging out of the body’ would refer to an action of motion away from
the speaker, such as ‘go’, ‘exit’; thus ʈh=ut=cone-bom (1sg=cl 4=go-npst)
‘I am going’, or, where something is (not necessarily tangible) generated
as in experiential verbs ‘feeling sad/happy/hungry/thirsty’, etc., as in thire
ut=ʈheʈe-bom (child cl 4=hunger-npst) ‘the child is hungry.’ These objects
of experience, namely ‘hunger’, ‘thirst’, etc., are an inherent part of the
experience (hence they are inalienable) and emerge involuntarily in a per-
son. They are seen as products of the body or ‘self.’ Similarly, verbs such
as ‘shake’ and ‘kiss’ have class 2 er= while verbs like ‘pound’ or ‘beat to a
pulp’ use e=, the class 5 which allows us to extend the analogy given in
the semantics of the partonomy of body.

(a) a=jeth
cl 1=vomit like
‘to feel nauseated or uneasy’
(b) Ɛ=colol
cl 5=roll
‘to roll down’
(c) e=biŋe
cl 5=think/remember
‘to think’

(6) ʈh=e=ta=biŋe
1sg=cl 5=app=think
‘I am thinking (of something).’

Two different kinds of experiences may demand two different proclitics.

jome ‘scare’

(a) Ɛ=jome
cl 5=scare
‘be afraid’
(b) ut=jome
cl 4= scare
‘get startled’

The proclitics are very productive and are not very selective as they attach
to a large number of verbs. Their distribution makes sense in terms of the
basic seven body division classifications. Consider class marker 4 ut=,
ot= attached to different verbs retaining the primary meaning of ‘genesis’
 the verb and verb complex 217

and class marker 5 i=, e= indicating internalised experience. Although

many usages are lexicalised (h), in many cases the class markers can pro-
ductively be attached to several verb forms to create new lexemes.

(a) ut=thu ‘be born’ [cl 4=born]

(b) ot=cone ‘go’, ‘exit’ [cl 4=exit]
(c) ut= ʈheʈhe ‘feel hungry’ [cl 4=hunger]
(d) ut=phae ‘feel thirsty’ [cl 4=thirst]
(e) ot=bo:lo ‘peel off ’ [cl 4=peel]
(f) ot=cobi ‘shoot at a large crowd’ [cl 4=shoot out]
(g) i=ji:te ‘tremble’ [cl 5=tremble]
(h) ijube ‘fly’ N

9.3 Transitive Verbs

A large number of transitive verbs are preceded by body division class

markers. The nature of these decides the specific meaning of the verb and,
at times, signifies multiple locations and the manner of action as exempli-
fied below. In addition to the seven basic class markers considered above,
the language has additional object clitics that attach to transitive verbs.
These are: et= ~ it= ~ ik= ~ ek= and Ɛn= ~ en= indicating more often than
not a resultative action as in Εn=oke ‘knit with cane’, where the result is
symbolised by the object clitic. Also refer to §7.3.2 on object clitics. The
phonetic shape of each of these clitics is often decided by the nature of
the action designated by the verb and the associated object seen in the
context of the partonomy of the body. Consider the following examples.
The verb ‘aim’ can denote various ways of aiming at an object in a hunter-
gatherer society and in PGA each is marked differently. This implies that
the same verb root may take different proclitics each specifying the mode
of action as seen in the example of the verb ‘cut’ in the previous chapter
in §7.3.2.2. Attaching different proclitics has the potentiality of changing
the valency of the verb concerned. This is a significant characteristic of
PGA that any verb root can take a variety of proclitics governed by the
appropriate semantics. Proclitics decide the ultimate meaning of the verb.
See below.

ʃile ‘aim’

(a) ut=ʃile
cl 4=aim
‘Aim from above’
218 chapter nine

(b) ara=ʃile
cl 6=aim
‘Aim from a distance’
(c) ek=ʃile
obj=aim
‘Aim at’ (resultative)
(d) e=ʃile
cl 5=aim
‘Aim to pierce.’ (internalised action)

Consider the following sentence:

(7) bol-bi(u)t= tara=ʃil-e

bol-abs-cl 4= cl 6=aim-imp
‘Aim at the Bol fish from a distance.’

In many usages these proclitics seem to indicate the ‘manner’ of action.

The transitive verbs ‘cut’ as seen earlier in §7.3.2.2 and ‘hit/slap’ given
below can have many manifestations signified by the body division class
markers.

pho ‘cut’

(a) ara=pho
cl 6=cut
‘cut down’, ‘fell’
(b) Ɛr=pho
cl 2=cut
‘hit with a stick in the front’
(c) ut=pho
cl 4=cut
‘separate from the source’ (betel nut from its branch)

bate ‘hit’

(a) er=bate
cl 2=slap
‘slap on the face’
(b) ek=bate
obj=slap
‘slap suddenly, unexpected’
(c) ut=bate
cl 4=slap
‘slap (hard)’
 the verb and verb complex 219

(d) eren-bate
refl-slap
‘slap oneself ’

(8) meo-e nyaramo ut=bate-k-o

Meo-erg Nyaramo cl 4=slap-fa-dst.pst
‘Meo slapped Nyaramo hard.’

A related phenomenon with the word ‘pluck’ or ‘pick’ is designated by

different clitics. English equivalents emerge as phrasal verbs with spatial
terms. Each verb form occurs with a distinct class marker.

lub ‘pick’

(a) Ɛr/e=lubom
cl 2=pick up (from the ground)
(b) it=lubom
obj=pick out, (stones in ‘daal’), weed out, sort out

(9) u khider ut=lub-om

3sg coconut cl 4=pluck-npst
‘He plucks coconuts (from the tree).’

As said earlier, the body division class markers can change the valency of
the verb.

khir ‘hot’

(10) e=khir be
cl 5=hot cop
‘It is hot.’

(11) ut=khir be
cl 4=hot imp
‘Heat it up.’

bin ‘squeeze’

(12) julu e=bin

clothes cl 5=squeeze
‘Wrung out clothes.’

(13) julu it=bin-k-o

clothes obj=squeeze-fa-dst.pst
‘(She) wrung out the clothes.’
220 chapter nine

Both the object clitic and the body division class marker preceding the
verb can coexist in clitic sequencing as in the following examples (14–15).
Although case markings are dropped in fast speech as mentioned earlier
in §5.2 on case, they may also be dropped when the object noun is not
a specific one (14). The object noun ‘box’ in (15) is accompanied by an
absolutive case as it refers to a specific box.

(14) ik=tƐr= ɔlo-k-e

obj=cl 2=send-fa-imp
‘Send (goods/something).’

(15) pheʈi-bi ik=tƐr= ɔlo-k-e

box-abs obj=cl 2=send-fa-imp
‘Send the box.’

Each body division class marker expresses different body locations when
combined with the verb for ‘tattooing’, but with class 4 which is for ‘gen-
esis’ it has the meaning of ‘uproot.’

ʈɔl ‘tattoo’

(a) ot= ʈɔl-e

cl 4=tattoo-imp
‘Tattoo the back of the body.’
(b) Ɛr= ʈɔl-e
cl 2=tattoo-imp
‘Tattoo the forehead.’
(c) ut= ʈɔl-e
cl 4=pluck-imp
‘Pluck it.’

The verb ‘to pluck’, ‘break, or ‘disengage’ is ʈɔl which can be employed
with a large number of body division class markers and object clitics to
convey the nature of the action and the object concerned. In (c) above,
the basic meaning of ‘emission’ or ‘separation’ is retained by the clitic. The
combinations of various class markers and the verb root express diverse
meanings, sometimes rendering an idiomatic phrase as in (17).

(16) bei-bi it= ʈɔl-o

bottle-abs obj=break-dst.pst
‘The bottle broke.’ (into pieces)
 the verb and verb complex 221

(17) caybi Ɛn-ʈɔlo-ke <un- ʈɔloko ‘bloomed flowers’

Wherever result-bloom-cop
‘Flowers bloom everywhere.’

(18) Ɛn-ʈɔlo-bi ik/t=(Ɛ) ʈɔle <Ɛ-ʈɔl-e ‘pluck flowers’

result-bloom-abs obj=pluck
‘Pluck the bloomed ones.’

(19) ʈɔʈɔe ɔrobit teŋom

ʈɔʈɔ-e=ɔro-bi ut=teŋ-om
Pandanus-cl 5.poss=flower-abs cl 4=smell-npst
‘The Pandanus flower smells.’

The verb ‘to see’ which is marked by the internal body division class
marker 5, i.e., e= ~ Ɛ=, changes the meaning if the relational body part
class marker 2 Ɛr= is attached to it.

ole ‘see’

(a) Ɛr=ole
cl 2=see
‘Call someone by gesture.’
(b) Ɛ=ole
cl 5=see
‘see’

All these examples indicate that the body division class markers are
semantically coded with a specific meaning and thus can be attached to
a variety of verbs with the appropriate modification in each case.
As far as object clitics are concerned we observed that each one speci-
fies different readings and, at times, results in idiomatic phrases. For
instance, ek=tara=ʃile ‘to see straight ahead’, or ‘to aim straight ahead.’
However, it can be used as an idiomatic expression as in ‘to keep an eye
on someone.’ Consider:

(20) Ɛm-boe-bi (e)k=tara-ʃile

refl-spouse-abs obj=cl 6-aim
‘Keep an eye on your husband.’

The phonetics of these proclitics is interesting. In fast speech they attach

themselves to the preceding case marker or pronominal clitic (example
19 above and 21 below) and form a phonological word. Refer to §2.5.8,
§3.8.3 and §4.3 for details. Thus, the above sentence will be rendered as
222 chapter nine

Ɛm-boe-bik tara-ʃile. Similarly, the word e=ole ‘see’ will be pronounced as

ʈhe=ole ‘I see’ if the pronominal clitic is added, but in case of the full pro-
noun form it will be ʈhu cokbi e=ole ‘I see a turtle.’ The verbs ut=leb ‘sweep’,
et=mok ‘leave’ and it=nɔl ‘played X’ will be rendered as exemplified in the
following sentences, allowing the object clitic to be moved towards the
case maker -bi.

ut=lub ‘pick’, ‘sweep’

(21) ʈhu ʈhi-bit=leb-om

1sg earth-abs.cl 4=sweep-npst
‘I sweep the floor/ground.’

et=mok ‘leave’

(22) ak=arkha tƐkhamo u porʈblƐr-bit=mok-om

3sg-tell soon 3sg Port Blair-abs.obj=leave-npst
‘He says that he will leave Port Blair soon.’

it=nol ‘obj=well’

(23) a=thire eren-khole-bit=nol-ɔ

cl 1=child refl-play-abs.obj=well-dst.pst
‘The child played it well.’

Thus, what strikes us most from the examples given above is that:

(i) The basic division in the verbs is not between +/- transitive, but is
between +/- dependency, i.e., whether they are preceded by a class
marker proclitic or not. Verbs are either dependent or indepen-
dent. The dependent ones can take any of the seven body division
class markers and/or object clitics.
(ii) The aspect of identifying what kind of body division class marking
proclitic is attached to verb root is important, as the body division
marking changes the meaning and valency of the verb.
(iii) In a typical SOV construction, the object clitic or the body division
class marking proclitic of the verb has the tendency to attach itself
to the absolutive form -bi of the preceding object noun and thus the
object noun in the sentence has the terminal part sounding as [bit]
or [bik]. Refer to §2.5.8 for details.
 the verb and verb complex 223

Although it is very difficult to distinguish one kind of meaning from

another while analysing each of the seven divisions represented in the
verbs, partly because the body division class marking proclitics have been
grammaticalised to varying degrees over a period of time, we were able to
make some basic generalisations as presented in Table 9.1 above.
Verbs that do not obligatorily take proclitics are smaller in number
than those which do. Some of the examples of the former are: ci ‘come’,
khuro ‘come close’, ʈoph ‘bath’, ɖekɔc ‘learn’, cɔk ‘to do good’, caʈ ‘to do’, ʃit
‘to hunt’, khuro ‘come’, lara ‘hunt turtle’, kara ‘rising of the sun’, akale ‘die’,
etc. Some of these examples have the potential to take reflexive prefixes,
Ɛm- and em-, to specify that the action was directed to the subject, e.g.,
em- ʈoph ‘to take a bath.’ Let us now consider the reflexive prefixes.

9.4 Reflexive/Self-directed

We briefly discussed reflexives in Chapter 4. The verbs whose actions are

self-directed are marked by the reflexive prefix aran- ~ aram- or Εm-. The
prefix aran- changes to aram- if the following word begins with a labial.
Not all but most of these verbs are prototypically experiential verbs. Thus,
the following verbs indicate that the action is self-directed.

(a) aram boli-k-o-m ‘to take rest’

(b) aram ben-o-m ‘to lie down’
(c) aran laʈ ‘startle’, ‘get scared’
(d) aran tƐŋ ‘to howl’, ‘to scream’
(e) em-phorol ‘turn (over) by self ’
(f) Ɛm-rokho-b ‘get ready’
(g) Ɛm-boi ‘get married’
(h) em-bɔrʈhul ‘remember’
(i) ɛm-bɑrɑte ‘flutter’
(j) ɛm-bo ‘know’
(k) ɛm-bolo ‘bruise’
(l) ɛm-borɑcɛ ‘get angry’
(m) ɛm-pʰɛ ‘dive’
(n) ɛm-pʰil ‘die’
(o) em-khil ‘shake’

(24) ʈha (Ɛ)m eʈh-om

1sg refl recline-npst
‘I am reclining’
224 chapter nine

(25) ʈh=Ɛm boe-b-om

1sg=refl marry-fa-npst
‘I am getting married.’

It has been noticed that the reflexive morpheme can also be infixed
between the root and the tense, aspect and mood categories. Thus:

(26) ut=ʈheʈhe > ut=ʈheʈhe-em-be ‘(He) was hungry.’

(27) e=maʈ> e=maʈ-em-b-ɔm ‘He runs.’

If the verb has the overt causative/applicative marker ta= the reflexive
marker can be distanced from the verb root, as in:

(28) ʈɔya-bi ʈha/ŋa/a/ŋale/-Ɛm-ta=malai-o

stand-stat 1sg/2sg/3sg/2pl/1pl/-refl-appl=tire-dst.pst
‘I/you/he/you all/we all got tired standing.’

To summarise, PGA verbs can be preceded by:

(i) Proclitics to an indicate direct object and objects which are the result
of an action. These are:
{et= ~ it= ~ ik= ~ ek= ~ Ɛn= and ~ en=}.
(ii) Reflexive prefixes such as {Ɛm- ~ em-}.
(iii) The seven body division class marker proclitics, such as {a=, Ɛr=, oŋ=,
ut=, e=, ara=, and o=}, which are attached to the verb roots.
(iv) The causative / applicative proclitic ata= ~ ta=.

Let us now consider the adjacency phenomenon that occurs between

bound morphemes and the verb root.

9.5 The Distancing of the Verbal Proclitic from its Host

We discussed in Chapter 3 that there is no obligatory rule that proclitics

immediately attach to the verb in all constructions. The fact is that they
can be moved to a position before a verb phrase as shown in (29a, 30, 32,
and 34) below. Whenever a body division class marker is long-distanced it
shows discontinuity. For example, class 4 in sentence (29a) is a proclitic
to the verb ‘cut’ and not to the noun ‘tree’ which is an independent noun.
Thus the hanging = sign means that the proclitic is not adjacent to the
 the verb and verb complex 225

verb, its usual host, but is instead attached to the grammatical category
on the left of the verb form, which serves as the phonological host.1 This
category is often the subject of the sentence. We reiterate the discussion
for the purpose of clarity and to link it with the discussion that follows in
§9.6. Distancing is shown by the line drawing.

(29a) ʈh=ut= ʈɔŋ−e pho-k-e

1sg=cl 4= tree-abs cut-fa-imm.pst
‘I cut the tree.’

(29b) ʈhu ʈɔŋ-e ut=pho-k-e

1sg tree-abs cl 4=cut-fa-imm.pst
‘I cut the tree.’

(30) ŋo i=cophe cɔŋ-o

2sg cl 5=enough/how much get-dst.pst
‘How much did you get?’ or ‘Did you get enough?’

(31) ŋo cophe i=cɔŋ-o

2sg enough/how much cl 5=get-dst.pst
‘How much did you get (or) did you get enough?’

We saw in §9.1 that verbal proclitics also attach to intransitive verbs.

These can also be long-distanced from the head, i.e., the verb, as the sub-
ject noun attracts the verbal proclitic or object clitic towards itself as in
sentences (32, 33, 34 and 35).

(32) reya jo et= nƐ rence-o

Reya Joe obj= 3pl fight-dst.pst
‘Reya and Joe fought with each other.’

(33) ʈh= ot=ʈheʈhe-b-ɔm

1sg= cl 4=hunger-fa-npst
‘I am hungry.’

(34) Buli-ot= diglipur-ak cone-b-ɔ

Buli-cl 4= Diglipur-dir go away-fa-dst.pst
‘Buli went away to Diglipur.’

1 This, at times, gives the clitic sequencing that we discussed in Chapter 3.

226 chapter nine

These structures show that complements of the clause or adjunct can

operate as intervening elements between the clitic and the verb. The
distancing of the proclitic is optional as the sentence (34) can also be
rendered as:

(35) buli diglipur-ak ut=cone-b-o

Buli Diglipur-dir cl 4=go-fa-dst.pst
‘Buli went to Diglipur’

There is no difference in meaning between (29a) and (29b), (30) and (31),
and (34) and (35). The verbal proclitics are marked by their floating char-
acteristic feature and never create any ambiguity.

9.6 The Causative/Applicative

The distancing of the host from its clitic is not restricted to body division
class marker proclitics. It can also occur in the case of causative construc-
tions where the causative proclitic can be distanced from the verb root.
The causative/applicative morpheme may be considered as body division
class marker proclitic number 1, i.e., a= ~ ta= ~ ata= because of it homopho-
nous nature with class 1. However, there are some significant constraints.
The causative morpheme attaches to all kinds of verbs regardless of the
fact that they are related to the concept of ‘emission’ as expected of verbs
taking the number 1 proclitic. Secondly, it precedes independent verbs,
those that are free from any attachment of proclitics, as well as dependent
verbs. At the most, one can speculate that the causative morpheme ta=
and its variant forms were derived from the body division class marker
cl 1 a= ~ ta= and later have been grammaticalised as causation. How-
ever, we do not have any proof thereof. Let us now consider causative and
applicative morphemes in detail: the valency-increasing phenomenon.
PGA employs two strategies to derive causative or applicative forms.
One is by using the morphological causative and the other is by the peri-
phrastic causative. The valency-increasing morpheme, i.e., the causative,
applicative, or transitiviser, morpheme ta= or ata= is prefixed to the verb
root and is thus infixed between the verbal proclitic and the verb root, e.g.
e=bi:ɲe ‘to think’ but e=ta=bi:ɲe ‘make someone think’ or ‘remember.’ This
is applicable to the verbs that obligatory take proclitics and are dependent
verbs. The causative proclitic attaches directly to the verb form if it is an
independent verb without any class proclitic. Let us discuss them.
 the verb and verb complex 227

9.6.1 The Morphological Causative

The morphological causative is derived by attaching the causative pro-
clitic to the verbal forms. Some examples are given here:

(a) i=ata=ole cl 5=appl=‘see’ ‘show’

(b) e=ta=biʈhe cl 5=caus=‘drown’ ‘sink’ (transitive)
(c) e=ta=iʃe cl 5=caus=‘give’ ‘keep’
(d) e=ta=julue cl 5=caus=‘cold’ ‘cool (it)’
(e) e=ta=lek cl 5=caus=‘down’ ‘bring something down from up’
(f) e=ta=loʈa cl 5=caus=‘wear’ ‘make someone wear (clothes)’
(g) e=ta=meʈa cl 5=caus=‘stick’ ‘stick something’
(h) a=ta=uno cl 1=caus=‘sit’ ‘seat someone’
(i) e=ta=ʈoya cl 5=caus=‘stand’ ‘erect’

In the case of the verbs with a class marker 5 proclitic signifying inter-
nalised action such as e=khole ‘laugh’, or self-directed action with the
reflexive Ɛm-beno ‘sleep’, the verbal proclitic or verbal reflexive prefix is
dropped and the causative ta= or ata= is directly added to the verb root
as in the following.

(36) surme nu ta=khole-o

Surme Nu caus=laugh-dst.pst
‘Surme made Nu laugh.’

(37) araːmliʃu aranʃuluthu ta=beno-ø

refl.sister her younger brother appl/caus=sleep-pst
‘The sister put her younger brother to sleep.’

The position of the causative morpheme is not rigidly fixed and it can be
prefixed to the entire verb form, i.e., a combination of the verbal proclitic
and the verb root (39). If the indirect object is not specified overtly in a
clause and the verb has a causative marker ta= it implies its latent pres-
ence (40), e.g.:

(38) u khider-e ut=phay-om

3sg coconut-abs cl 4=dry-npst
‘He dries coconuts.’

(39) u khider-e ta=ut=phay-om

3sg coconut-abs caus=cl 4=dry-npst
‘He dries coconuts (collectively with the help of someone).’
228 chapter nine

The causative morpheme can distance itself from the verb with the direct
object or a whole noun phrase preceding the verb form. For example:

e=khu-(v)e ‘drink’ (imperative)

(40) ata=([ino]obj) khu-ve

caus=water drink-imp
‘Make (him) drink (water).’

(41) ata=[ino-julu]np khu-ve

caus=cold-water drink-imp
‘Make (him) drink cold water.’

There are instances of both indirect and direct object incorporation

between the causative morpheme and the verb form. Consider below
where (b) and (c) are variants of (a).

(42) (a) ʈh=a=mimi thire ino ta=khue-ø

1sg=cl 1.poss =mother child water caus=drink-pst
(b) ʈh=a=mimi thire ta=ino- khue-ø
1sg=cl 1.poss=mother child caus=water-drink-pst
(c) ʈh=a=mimi-ta= thire-ino-khue-ø
1sg-cl 1.poss=mother-caus= child-water-drink-pst
‘My mother fed the baby the water.’ or
‘The baby drank the water with the help of my mother.’

In sentence (42c), the second reading, i.e., ‘The baby drank the water with
the help of my mother’ is more appropriate than the first one ‘my mother
fed the baby the water’ as {ta=} is homophonous with the instrumental
suffix {-ta}.2 Also refer to §5.2 on case.
In the following example (43), the proclitic i= is distanced from its
host by the intervening object noun kitab ‘book’ as well as the caus-
ative marker ta=. In the next example (44), the causative marker ta=
is moved further from the verb incorporating the object noun, with the
verbal proclitic still being distanced from its host. In short, the causative
proclitic ta= or ata= is moveable and can be distanced from the verb form

2 This appears to be the ‘case on verbs.’ That is, “the meaning of a morpheme used
in different morphosyntactic contexts changes because of the context itself. . . . Such ‘ver-
satile’ cases can express (i) temporal, causal and other relationships between clauses”
Aikhenvald (2008: 565).
 the verb and verb complex 229

and the object noun can be incorporated between the two bound mor-
phemes, the verbal and the causative proclitics and the verb root.

i=ole ‘see’

(43) ʈhu buli i= kitab ta=ole-ø

1sg Buli cl 5= book caus=see-pst
‘I showed Buli the book.’

(44) ʈhu buli i= ta=kitab-ole-ø

1sg Buli cl 5= caus=book-see-pst
‘I showed Buli the book.’
‘I showed the book to Buli.’

As seen earlier, proclitics and object clitics have the tendency to be

attracted by the pronominal proclitic to generate a phonological word.
In the following example (45) the causative proclitic, like any other class
marker proclitic is attracted towards the pronoun. Hence, it attaches to
the preceding pronominal form ʈho ‘I.’ The causative ta= ~ ata= can also
be attached to the preceding word if it happens to be one ending in a
vowel. Fusion of a causative proclitic to a word on its left creates a word
boundary3 that ends at ta. This is observed in fast speech.

(45) lico ʈhota= ŋolɔ-k-o

Lico 1sg.caus cry-fa-dst.pst
‘Lico made me cry.’

(46) a=lico thire-bi ta=thu-e

cl 1=Lico child-abs.appl=born-imm.pst
‘Lico gave birth to a child.’

(47) mona-i inu juluta=khu-e

Mona-dat water cold.caus=drink-imp
‘Make Mona drink cold water.’

The applicative/causative ta= can also be distanced from the verb root
and can move to the beginning of the verb phrase as in the following
sentences:

3 The word boundary is measured by the presence of a pause. Thus, speakers give a
pause between ʈhota and the following verb ŋol ‘cry.’
230 chapter nine

(48) ʈhu tar-e ta=ʈhu-i=noku-ø

1sg wire-abs appl=1sg-cl 5=strengthen-pst
‘I strengthened the wire.’

(49) ʈhu a=jo ta=ʈhu-i=kon-be

1sg cl 1=Jo caus=1sg-cl 5=rise-npst
‘I make Jo rise/wake up.’

Sentences (48–49) are very interesting as the causative proclitic begins

the verb phrase which has the incorporated subject ʈhu ‘I’ as well as
the information about the object indicated by the verbal proclitic. This
phenomenon is unusual across the languages of the world. The semantic
difference between the caus proclitic preceding the verb phrase and the
one with the default position of the pre-verbal root as in (47) could not
be elicited because of the moribund nature of the language and the low
frequency of such constructions.

9.6.2 Periphrastic Causatives

PGA offers periphrastic causatives by employing the word for ‘ask’/ ‘tell’
in addition to the causative morpheme ta=. The subject of the verb ‘ask’ is
the causer of the event, the object of the verb ‘ask’ is the causee which
serves the A function and the object of the caused event serves the O
function.

(50) a=kaba a=buro ik=jira-l a=bie

cl 1=Kaba cl 1=Buro obj=tell-conv cl 1=Bie
ta=e=lukh-o
caus=cl 5=pickup-dst.pst
causer causee (A) caused (O)
‘Kaba asked Buro to pick up Bie.’

Many speakers were observed dropping the causative morpheme ta= and
instead used the verb for ‘ask’ in the first clause for causative reading.
Since the second clause lacks the overt subject in most of the construc-
tions the sentences appear to be sequential in nature.

(51) a=boa a=nu ik=jira tatoŋ-e it=ʃiro-ø

cl 1=Boa cl 1=Nu obj=told courtyard-abs obj=wash-pst
‘Boa asked Nu to wash the courtyard.’
 the verb and verb complex 231

(52) a=boa a=meo ik=jira-l iku=beliŋ-ø

cl 1=Boa cl 1=Meo obj=say-conv obj=cut-pst
‘Boa asked Meo to cut (the tree).’ Or
‘Having Boa asked, Meo cut the tree.’

(53) a=raja unni-jira-l nu iku=beliŋ-ø

cl 1=king people-say-conv people obj=cut-pst
‘The king made the people cut (the tree).’ or
‘The king got the tree cut by asking them.’

Another strategy used to form causatives was observed in the speech of

Boa Sr., our Bo speaker. Consider the following sentence where the verb
‘ask’ terminates the sentence but carries the meaning of causation of the
event. The caused event is in the non-finite nominalised form.

(54) lico ʈh=i aʈ-bi ʃere-bi liwo-ø

Lico 1sg=dat wood-abs cut-nmlz ask-pst
‘Lico made me cut the wood.’

(55) lico ʈhi ŋole liwo

lico ʈh=i ŋol-e liwo-ø
Lico 1sg=dat cry-nmlz ask-pst
‘Lico made me cry.’

As these constructions were not shared by the other speakers, one can
speculate that this was a feature of the Bo language, now extinct as we
lost Boa, Sr., the last speaker of the language, in 2010.
To summarise the discussion so far, a very significant fact about PGA
is: dependency in PGA does not necessarily license adjacency. Bound root
forms can be bound long-distance with clitics. The distancing of these
bound forms does not bring any change in meaning.

9.7 Formative Affixes

All verb forms that end in an open syllable are followed by a single con-
sonant -b, or -k or -l as a formative affix before the TAM morphemes are
attached to them.4 This consonant can never terminate the verbal form.

4 In my earlier works on the language (Abbi 2006a, 2006b, and 2009) I had considered
these formative affixes as verb class markers with some specific meanings. A closer look
into the language and subsequent visits to the field revealed that these are phonetic ele-
ments which serve as the linkers between the verb root and the TAM markings.
232 chapter nine

Hence, its status in the grammar is that of an infix. The variations between
these forms are not conditioned by any phonological or semantic environ-
ment. They appear to be remnants of various dialectal variations which
came together to form PGA. The author has observed that these are occa-
sionally dropped in current speech. We will use the non-past form or the
imperative mood marker to explain this phenomenon. Thus, the form for
‘goes’ is constituted of [a proclitic] [verb root] [the formative affix] and
the [non-past tense morpheme -om]. Consider:

(56) ut=cone-b-om
cl 4=go-fa-npst
‘(He) goes’, ‘leaves.’

(57) iji-b/k-om
eat-fa-npst
‘(He) eats.’

(58) ie-k-om
give-fa-npst
‘(He) gives’

(59) et=mo-k-e
obj=leave-fa-imp
‘leave (it)’ (imp)

The verb root ending in a closed syllable, i.e. ending in a consonant, does
not take the formative affix and the TAM suffixes are directly attached to
the verb root.

(60) buʈh-o
fall-dst.pst
‘(it) fell.’

(61) it=ʃir-e
obj=wash-imp
‘Wash it.’

(62) o cokbi it=coŋ-o

3sg turtle obj=find-dst.pst
‘He found the turtle.’

(63) u khider-bit= bol-om

3sg coconut-abs.obj= scrape-npst
‘He scrapes the coconut.’
 the verb and verb complex 233

9.8 Tense, Aspect and Mood (TAM)

9.8.1 Past Tense

The language makes a distinction between the past and the non-past,
the latter representing the present and the future, and is symbolised by
{-om}.
Following Dahl (1985: 23–26), the non-past tense can have a basic
meaning and an extended meaning which is decided by the context. As
we shall observe later in the chapter, this is also true for PGA.
PGA has four different past-tense categories, each specified differently.
-o distant past
-a ~ -e immediate past
-ø simple past
-il ~ -l It is mainly used as a past tense marker in a narrative style to show
ongoing events in the past.
The immediate past refers to an action or state that took place a few hours
prior to the time of utterance. Any other event beyond this time period
is distant past. The unmarked past, or when the verb occurs in the root
form without any marking (symbolised by ø in the grammar), indicates
simple past without the distant or immediate readings. It is also used for
simple statements that refer to a time anterior to the speech event. In
some ways sentences marked by ø signify an incomplete action or pre-
sume that something else happened after that past event.
The unmarked past and the one marked for narration -il/-l, are homoph-
onous to the verbal form in the subordinate clause indicating the adver-
bial or converbal function of ‘cause’, ‘manner’, ‘after’ and ‘while.’ Refer
to §10.5.1 for details.
The most common and widely used past tense marker is /-o/. It has
the greatest frequency of occurrence in our database. In rare cases, some
speakers used this even when the action was not undertaken in the dis-
tant past.
Tense markings are the terminal elements in the verbal form. As
explained earlier, the tense markings are suffixed to the verb root + for-
mative affix (if present). Consider:

(64) tara=sulu ʈaŋʈɔ-e ʈhit-phile-k-o

cl 6=later year-abs land.obj-teeth-fa-dst.pst
‘The tsunami came last year.’
234 chapter nine

(65) e=tei-kha ʈoʈa Ɛmphil-o

cl 5=fever-spec boy die-dst.pst
‘The boy who had a fever died (sometime ago).’

Let us consider a few sentences to elucidate constructions with a zero

past tense marker.

(66) nu aka-uno-ø
Nu 3sg-sit-pst
‘Nu sat down.’

(67) o ʈole iebi-ø

3sg potato bring-pst
‘He took/brought the potatoes.’

(68) u khider-e utun-toa e=mok-ø

3sg coconut-abs refl-brother cl 5=give-pst
‘He gave the coconut to his brother.’

(69) a=boa a=meo ik=jira-l khider-e iku=beliŋ-ø

cl 1=Boa cl 1=Meo obj=say-conv coconut-abs obj=cut-pst
‘Boa asked Meo to cut the coconut.’ or
‘After being told by Boa, Meo cut the coconut.’

(70) ʈho-ŋa= tƐkho-bi ik=jira-ø

1sg-2sg.cl 1= language-abs obj=speak-pst
‘I spoke in your language.’

Verbs can be compounded with the final verb taking a zero past tense
marker, e.g.:

(71) du bo5 o lɛc.ik= nɔne ince-ø

remaining 3sg arrow-obj= search go after-pst
‘He went to search for the remaining arrows.’

The following examples indicate the immediate past.

(72) u ʈh=i ino-bi tƐʃ-a

3sg 1sg=dat water-abs give-imm.pst
‘He gave me water.’

5 It is a combination of /du/ ‘deictic for invisible distance’ and /bo/ ‘and’, the
conjunction.
 the verb and verb complex 235

(73) atto ra-bi ʈhambikhir ʈhu eol-a

emph pig-abs 1sg.morning 1sg see-imm.pst
‘I saw the same pig this morning.’

Another feature that caught our attention was that some speakers, espe-
cially the Khora speaker, used the -e suffix indiscriminately for the imme-
diate past tense (sentence 74 and 75 below) as well as for the imperative
mood (see §9.8.4). Consider:

(74) u ʈhu it=mok-e

3sg 1sg obj=give-imm.pst
‘He gave it to me.’

(75) a=joe-e a=nu-co-ta khider-bi ie-b-e

cl 1=Jo-erg cl 1=Nu-near-abl coconut-abs take-fa-imm.pst
‘Jo took away the coconut from Nu.’

As stated above, the fourth type of the past marker -il ~ l is used in a nar-
rative style. Refer to the narrative text given at the end of the grammar.
Its usage as an adverbial subordination with the meaning of ‘after’ has
been termed converb in subordinate clauses (79, 80). For further discus-
sion refer to §10.5.1. Consider the following sentences extracted from a
folk tale:

(76) u tuŋ bo-kalui-il

3sg refl heart-break-nrr.pst
‘His was heart-broken.’ / ‘He became melancholic.’

(77) u ne boi-(e)m-il
3sg pl marry-rec-nrr.pst
‘They got married among themselves.’

(78) phertajiɖo em-boi a=kɔʈ-e ikjira-l

Phertajido refl-wife cl 1=Kot-abs told-nrr.pst
‘Phertajido asked his wife Kot.’

(79) bo lɛc-ik= kac-il o lɛc-it=cɔŋ-ø mino-tara cƐʈhul

and arrow-obj= make-conv 3sg arrow-obj=find-pst potato-under roots
‘Having made more arrows [he went to search for more of the shot arrows]
he found the arrow in the root of a potato plant.’

(80) o it=cɔŋ-il kɔʈ e=mƐle ʈh(i)it=cɔŋ-o

3sg obj=find-conv soil cl 5=fine land-obj=find-dst.pst
‘Having found it (the arrow) he found very fine soil.’
236 chapter nine

9.8.2 Non-Past Tense

As mentioned earlier, there is one suffix -om that is used both for the
future and present as well as for continuous actions or events.

(81) ʈhu ʈɔphɔ-l-ɔm

1sg bathe-fa-npst
‘I am going to take a bath.’ or ‘I am bathing.’

(82) ʈh=ot= ʈheʈhƐ-b-om

1sg-cl 4= hungry-fa-npst
‘I am hungry.’

(83) ŋ=aram marako-be ʈamʈam akaɲa-k-om

2sg=refl quiet-imp Tamtam eat-fa-npst
‘Don’t speak, Tamtam is eating.’

The non-past is not necessarily a primary deictic temporal reference. It

may refer to various types of imperfective situations with the moment of
speech as the reference point. Hence, it includes:

(i) ongoing activities

(ii) generic situations
(iii) habitual situations
(iv) future situations

Hence, the non-past tense -om signifies present imperfective situations as

well as an event that is yet to take place.

9.8.3 Aspect
The aspectual functions of ‘perfective’, ‘iterative’, and ‘prospective’ in PGA
are expressed by adverbial constructions and not by an aspect marker added
to the verb root. Let us first consider the use of adverbial constructions.

(84) ʈh=ara=ʃulu-thu karakha kɔil bath-il jicer-bi

1sg=cl 6.poss=sister 3sg.said yesterday night-loc rain-abs
kaɲɔro-k-e
frequently-fa-imm.pst
‘My younger sister said that it rained frequently last night.’

(85) a=thire thu kɔlot

cl 1=child born just now
‘The child is born just now.’
 the verb and verb complex 237

(86) tethamo ɲεr=cɔk-bi (e)ole-pho

Many days 2sg.cl 2.poss=face-abs see-neg
‘(I) have not seen your face for a long time.’

(87) ɖekho reya kona-bi iji-k-o

Already Reya tendu-abs eat-fa-dst.pst
‘Reya has already eaten tendu.’

Aspect can mark the stage of an action. The prospective aspect is indi-
cated by the combination of the verb ut=jukho ‘about to’ which specifies
anticipatory action, and the verb in a nominalised form (refer to §10.5.4
on action nominals).

(88) ʈhu boaː-l tuttaraː-l ʈh=auno-bi-(u)t= jukho-ø

1sg floor-loc on-loc 1sg=sit-nmlz-cl 4= about to-pst
‘I was going to sit on the floor.’

(89) thire ŋolo-bi-(u)t= jukho-ø

Child cry-nmlz-cl 4= about to-pst
‘The child was about to cry.’

The other aspect marker -me which indicates a habitual or generic mean-
ing is attested in the Bo language. Other speakers were not seen using this
suffix. In our database there are very few examples of -me. It has been
noticed that most of the current speakers use -om for indicating the itera-
tive and generic as well as habitual meaning.6 As stated earlier in §9.8.2
the non-past tense marker indicates several meanings and one of them
could be ‘generic.’ The following are examples from Bo.

(90) eƐran tƐro nukhe-me

Sea wave huge approach-hab
‘The big wave rises again and again.’

(91) ʈhu maram-bɔ pho-me

1sg self-heart neg-hab/npst
‘I don’t know.’

(92) bharat-il nu rƐfe i=iji-me

Bharat-loc people rice cl 5=eat-hab/npst
‘People eat rice in India.’

6 It is not uncommon to use present tense forms for generic or habitual meaning in
languages, e.g., Hindi mƐn məndir jati hun ‘I go to temple (everyday).’
238 chapter nine

(93) uroʈɔi-l ɲo tƐc-ta ɲo-bi Ɛm-bano-me

Old time-loc house leaf-instr house-abs refl-make-hab
‘In old times houses were made of leaves.’

(94) me ʈhi-bi it=phile-l m=ar(a)=Ɛm-bɔ pho-me

1pl place-abs obj=swell-conv 1pl=cl 6=refl-heart neg-hab
‘We had no knowledge (when the Tsunami came and) the place got
inundated.’

The post verbal {-me} is interpreted as a repeated action. The habitual,

thus, is an aspect which expresses action that takes place habitually or
repeatedly. It does not refer to a particular event.

9.8.4 Mood
Only one marker of mood was attested by all speakers, viz., imperative.
The imperative mood is symbolised by {-e ~ -be}.7 It is attached to the
formative affix or to the verb root as the case may be. Consider:

(95) u khider-e khu-a a=ʈoʈa untƐle-be

3sg coconut-abs drink-imm.pst cl 1=boy call-imp
‘Call the boy who drank coconut (water).’

(96) ŋa= uno-b-e

2sg.cl 1= sit-fa-imp
‘You sit down.’

(97) ʈh=e bet-b-e

1sg=abs press-fa-imp
‘Please press me a bit.’

(98) aʈ iku-b/k-e
Wood burn-fa-imp
‘Burn the firewood.’

(99) ata=metƐ khu-e

caus=milk drink-imp
‘Feed him milk.’

7 The form -be is homophonous with the copula be.

 the verb and verb complex 239

Table 9.2. Verbs with Formative affixes, tense and mood suffixes
Verb Non-past Various past Imperative English gloss
markers
1 cone cone-b-om cone-ø, cone-b-o cone-b-e go, leave
2 ɲa aɲa-b-om aɲa-ø, ɲa-b-o aɲa-b-e bark, eat
3 ŋol ŋol-om ŋol-o ŋol-e cry
4 caʈ caʈ-om caʈ-o caʈ-b-e do
5 ɖe ɖe-b-om ɖe-b-o ɖe-b-e shut up
6 boʃo boʃo-b-om boʃo-b-o boʃo-b-e beat, hit
7 ʈɔlo ʈɔlo-k-om ʈɔlo-k-o ʈɔlo-ke break
8 iye iye-k-om iye-k-o iye-k-e catch
9 iji iji-k-om iji-k-o iji-k-e eat
10 juvu juvu-k-om juvu-k-o juvu-ke fly
11 ʃolo ʃolo-k-om ʃolo-k-o ʃolo-k-e hang
12 tɔle tɔle-k-om tɔle-k-o tɔle-k-e mix
13 uno uno-b/k-om uno-k/b-o uno-k/b-e sit
14 beno beno-b/k-om beno-b/k-o beno-b/k-e sleep
15 eole eole-b-om eol-a, eole-b-o eole-b-e see
16 lele lele-b/k-om lele-b/k-o lele-k/b-e swing
17 Ɛmphil Ɛmphil-om Ɛmphil-o Ɛmphil-e die
18 araʈhul araʈhul-om araʈhul-o araʈhul-e kick
19 ʈhitbol ʈhitbol-om ʈhitbol-o ʈhitbol-e search the
ground/jungle
20 kac kac-om kac-il kac-e make

By prefixing various adverbial words to the verb with the imperative {-be
~ -e} suffix, one can get a reading of the hortative and optative mood.8
Thus:

(100) akhe ʈhu beno-be

For a while 1sg sleep-imp
‘Let me sleep for a while.’

The following table lists some of the forms discussed above.

8 We found only one example of the subjunctive mood koliŋe in our database as shown
in example (106) in the next chapter. Several attempts to obtain similar sentences failed as
speakers preferred to use paratactic constructions. Hence, we presume this mood marker
is no longer used.
240 chapter nine

9.9 The Copula Be or Jiyo

Languages offer nonverbal predicates with copula occurring with noun,

verb and adjective in the predicate slot (Dryer 2007: 224). PGA also offers
nonverbal predicates: nominal predicates, adjectival predicates and loca-
tive predicates which take the uninflected form of the copula.

(101) peje ɲyo be

Peje home cop
‘Peje is at home.’

(102) di ʈhi nɔl bi

this place good cop
‘This is a good place.’

The copula can also exist with negative verb pho as in:

(103) ʈh=Ɛ=cul tƐkhotƐc pho be

2sg=cl 2.poss=near money neg cop
‘I don’t have any money.’

The be or bi copula never takes tense, mood and aspect categories.

jiyo ‘exist’

PGA has a related word jiyo signifying ‘existence’ of an entity with refer-
ence to a ‘location’ at a particular point in time.

(104) ʃup macan tuttral jiyo

basket table under exist
‘The basket is under the table.’

(105) akacotoŋ.tot=ɲo tƐr=cokhe-l ʈh=ut=ɲo jiyo

doctor cl 4.poss=house cl 2=front-loc 1sg=cl 4.poss=house exist
‘My house is in front of the doctor’s house.’

The word jiyo ~ jiyu can optionally take the copula as in the following.

(106) er=ŋolka biu cul jiyu (be)

cl 2=matchbox candle near exist (cop)
‘The matchbox is near the candle.’
 the verb and verb complex 241

Table 9.3. Verbal affixes in PGA

Types of affixes Prefixes Infixes Suffixes
Formative -b-, -k-, -l-
Reflexive (moveable) Ɛm-, em- -Ɛm-, -em- -Ɛm-, -em-
Past tense -o, -a, -ø, -il
Non-past tense -om, -e
Imperative -be, -e
Habitual -me
Conjunctive participle -il, -l
Action nominal -bi

(107) tajiocor ʃare-l jiyo (be)

fish sea-loc exist (cop)
‘Fish live in the sea.’

As should be clear by now, the PGA verb does not agree with any of the
arguments in the sentence. Hence, the verb form remains the same in the
following sentence.

(108) Ram/sita/thire-n kona-bi iji-om

Ram/Sita/child-pl tendu-abs eat-npst
‘Ram/Sita/children are eating tendu fruit.’

To summarise the discussion, one can present the following table. All ver-
bal affixes and their variants are given in the table. Object clitics and body
division class marker clitics which precede the verbs are not included for
obvious reasons as each of them will vary according to the nature of the
verb. Causative proclitics are also not shown here as the table lists all the
affixes.
The moveable affixes have the potentiality to occur as prefixes as well
as infixes.

9.10 The Verbs ‘Come’ and ‘Go’

While collecting data for PGA, it was noticed that the two verbs ‘come’
and ‘go’ had several forms, each with a distinct meaning. They occur in
both forms, dependent and independent. The latter, the ones without the
body division class marker proclitics, are more common. Readers may find
this list interesting from the point of view of ethnolinguistic information.
Some of the structures have become obsolete as speakers who used them
are no more in this world.
242 chapter nine

(a) ci ‘come in or go to someone’

(b) ce (Bo) ‘come in or go to someone’, ‘go towards’
(c) aka=ci ‘to follow someone’
(d) unci ‘to come or go to someone’
(e) inci ‘to come or go with a purpose’
(f) uni ‘come back from somewhere’
(g) eɲo ‘come or go aimlessly’
(h) thu ‘exit’
(i) meli ‘return’
(j) mo ‘leave’
(k) ut=cone ‘leave’, ‘exit’, ‘go’
(l) ʈɔl ‘roam around’
(m) tƐbol ‘run away’
(n) e=maʈ ‘run (fast)’
(o) aka-ɲar ‘come frequently’
(p) khuro ‘come here (imperative)’
(q) bel ‘pass away’, ‘leave’
(r) aka-ao ‘come with someone’
(s) loto ‘come again and again’
(t) ʈhikhamol ‘come immediately’
(u) Ɛr=ʈɔk ‘come down slowly’
(v) phoro ‘come close’

The exact meaning of every form is not very clear. Sometimes the mixed
nature of the language gives us more forms than necessary. Consider some
of the examples.

(109) ŋu kotra-kak ci-b-im meŋ e=likh-om

2sg room-inside come in-fa-prohb 1pl cl 5=uncomfortable-npst
‘Don’t enter the room we feel uncomfortable/shy.’

(110) cay bi inci-k-om

Q abs go for-fa-npst
‘What are you going for?’

(111) ʈhu uni-b-om

1sg come back-fa-npst
‘I will come back.’

As can be observed, very few of these take proclitics and thus, are free
forms. Readers may refer to the dictionary of the language (Abbi 2012) for
the semantics of verb forms and their variations.
To summarise, a verb root in PGA is free to take any of the seven body
division class markers or any object clitic. It is the combination of the pro-
clitic and the verb root which gives the complete meaning to a sentence.
 Chapter Ten

Syntactic organisation

Introduction

Present-day Great Andamanese (PGA), as we have seen so far, has a

simple syntactic structure. Most of the complexities that exist are at the
morphological level. Speakers keep their sentences short and conjoined
by pauses. We shall discuss some of the structures that represent various
syntactic organisational characteristics in this chapter.

10.1 Word Order

PGA is a verb-final language and the order of the constituents is SOV (1).
However, there is some flexibility in the ordering of the word, not only
because of focus, but also because the speakers were observed varying the
order freely. As the language is not used anymore it was difficult to reach
any statistical analysis of the preference of aberration from the prototypi-
cal SOV pattern.

(1) ram konabi jiyom

ram konabi iji-om
Ram tendu-abs eat-npst
‘Ram is eating a tendu fruit.’

10.1.1 Adjectives and Nouns

Adjectives and intensifiers follow the noun they modify. However, in
inalienable possession the body division possessive marker precedes the
possessed element. For details see Chapter 6 on possession and §3.4. The
following phrase will clarify the picture.

(2) ʈɛlɛ taratom akaːlepho

ʈɛlɛ tɑrɑtom ɑkɑːle pho-ø
elephant old die neg-pst
‘The old elephant did not die.’
244 chapter ten

(3) sita kona phuŋ-bi iji-om

Sita tendu-ripe-abs eat-npst
‘Sita is eating a ripe tendu.’

(4) lico er=co

Lico cl 2.poss=head
‘Lico’s head.’

The direct object is next to the verb on its left. Hence the ordering is S IO
DO V.

(5) u utuntoa khidere umuk

u u-tun-toɑ khider-e u-muk-ø
3sg 3sg-refl-brother coconut-abs 3sg-give-pst
‘He gave a coconut to his brother.’

If the DO under consideration is modified by an adjective then the whole

noun phrase precedes the verb.

(6) ʈhui rathomo tɛʃulu ijo.

ʈhu i= rɑ thomo tɛʃulu iji-o.
1sg cl 5= pig-flesh stale eat-dst.pst
‘I ate stale pork.’

This is true of the constructions with intensifiers too.

(7) kɔʈ e=mele tɛrcɛk ʈhit=cɔŋ-o

Soil cl 5=fine much earth.obj=find-dst.pst
‘(He) found very fine soil on the ground.’

An NP with multiple adjectives precedes the verb form.

(8) kɔʈ e=mele e=nɔl ʈhit=cɔŋ-o

Soil cl 5=fine cl 5=good earth.obj=find-dst.pst
‘(He) found good and fine soil on the ground.’

In the case of multiple adjectives, each adjective is preceded by its appro-

priate proclitic as in the examples (8 and 9).

(9) aka=phup ut=ʈɛŋ e=cay

his=sputum cl 4=odour cl 5=bad
‘His bad smelly sputum.’
 syntactic organisation 245

The order among elements at the same level differs from one group of
languages to another (Shopen 2007: 111–113). Our database of PGA does
not contain demonstratives, adjectives, quantifiers and numerals in the
same phrase. Since adjectives and numerals follow nouns but demonstra-
tives precede nouns we have a mixed ordering system with elements at
the same level.

(10) (ɖuini) kitaːb biraŋ tatɛrbui-bi boʈho-ø

(3du.dis.vis) books red two-abs fall-pst
‘Those two red books fell down.’

Since the language has a dual demonstrative pronoun there is no need to

have the number ‘two.’ Alternatively, the dual demonstrative pronoun can
be dropped as it is redundant.

10.1.2 Genitive
The genitive -ico ~ -iʃo is attached to the possessor noun or pronoun. How-
ever as discussed in Chapter 6, the inalienable possessions are preceded
by possessive body division proclitics. Thus (11) has genitive suffix
but (12) has possessive proclitic.

(11) ɖu-iʃo cɔkbi

3sg.dis.invis-gen turtle
‘His turtle.’

(12) cokbi er=co

turtle cl 2=head
‘Turtle’s head.’

Refer to §6.7 in the chapter on possession for details.

10.1.3 Case Markers

As discussed in Chapter 5 case markers appear in the post-NP position.
For details refer to §5.2.

(13) aʃyame bəs kuttaral konabit beliŋo

ɑ=ʃyam-e bɑs kotr-ɑl konɑ-bi it=beliŋ-o
cl 1=Shyam-erg bus inside-loc tendu-abs obj=cut-dst.pst
‘Shyam cut the tendu fruit in the bus.’
246 chapter ten

However, the body class marker a= always precedes animate arguments

as in the case of a=ʃyam above. If the noun is modified by an adjective or
a quantifier or numeral, the case marker appears at the end of the NP as
in sentence (19) below. The demonstrative precedes the noun. Also refer
to §7.2 and §8.1.

(14) di kɔbɔ ot=ɲo be

3sg.dist.vis Kobo cl 4=house cop
‘That is Kobo’s house.’

(15) ɖu ɲo buruin tot =bɔl jio

3sg.dis.invis house mountain cl 4=behind exist
‘That house is behind the mountain.’

(16) ɖune ra mote lobe

3pl.dis.invis pig fat (<Hindi) exist
‘Those pigs are fat.’

10.1.4 Numerals and Quantifiers

As expected, numerals and other quantifiers follow the nouns they modify.

(17) ʈhu tajio ɛtcalo cɔphe ji- k-o

1sg fish a lot enough eat-fa-dst.pst
‘I ate a lot of fish.’

(18) golaʈ ɛr=ulu entoplo nɔl pho-be

Golat cl 2.poss=eyes single good not-cop
‘Golat’s one eye is not good.’

(19) kitaːb tatɛrbuibi ʈobo

kitɑːb tɑtɛrbui-bi ʈob-o
book two-abs steal-dst.pst
‘Two books were stolen.’

10.1.5 Degree Words

Degree words like intensifiers, such as ‘more’, ‘much’, ‘very’ follow the
modified. This is found in the superlative constructions too.

(20) ŋɛcul julu cɔphebe

ŋɛcul julu cɔphe-be
2sg-have cloth many-cop
‘You have many clothes.’
 syntactic organisation 247

(21) di ino nɔl totcɔ be

di ino nɔl totcɔ be
3sg.dis.vis water good very cop
‘This water is the best.’

10.1.6 Manner Adverbs

Manner adverbs, both derived and non-derived, precede the verb espe-
cially in imperatives. In other constructions they optionally occur in post
verbal position. Also refer to §8.3.

(22) ŋe erentɛʃu iju-k-e

2sg slowly eat-fa-imp
‘You eat slowly.’

(23) o aka-uno-l olam-o

3sg 3sg.cl1–sit-conv tire-dst.pst
‘He got tired of sitting.’

(24) o khole.k= ʈh=e jira-ø

3sg laugh.obj= 1sg=abs speak-pst
‘He spoke to me laughingly.’

However, Jeru speakers, such as Surmei, always used the adverb in post-
sentential position.

(25) ŋu cai jit (iji-it) kalemo

2sg why eat-obj soon
‘Why did you eat it so fast?’

(26) ʈhu ŋɔʈo cɔphu

ʈhu ŋɔʈo-ø cɔphu
1sg swim-pst well
‘I swam well.’

10.1.7 Temporal Adverbs

Temporal adverbs generally occupy the sentence-initial position (27 and
28), however, they can optionally occur after the subject (29). The tem-
poral adverb indicating ‘morning’ is obligatorily attached to a pronominal
proclitic on its left. The appropriate proclitic (here ʈh=1sg) represents the
subject of the sentence. Also refer to §8.4.
248 chapter ten

(27) ʈhikamo ɲ-ɛr=cɔk-bi (e)ole-pho

For many days 2sg-cl 2.poss=face-abs (cl 5)=see-neg
‘(I) haven’t seen your face for many days.’

(28) ʈh=ambikhir ʈh=ut= cone-b-om

1sg=refl.morning 1sg=cl 4= go-fa-npst
‘I will go in the morning.’

(29) peje taraottabike ʈhimikhu ot=cɔne-k-ɔm

Peje early in the morning jungle cl 4=go-fa-npst
‘Peje goes to the jungle early in the morning.’

10.1.8 Negative Verb

The negative verb pho always follows the main verb that it negates, but
precedes the copula.

(30) du dillikek utcɔnepho

du dilli-kek ut=cɔne pho
3sg.prox Delhi-dir cl 4=go neg
‘He does not go to Delhi.’

For a detailed account on negation refer to §10.3 below on negation.

10.1.9 Relative Markers

The order of relative markers and nouns is intriguing as most of the
speakers use two coordinated sentences such as the following without
any conjunction and with gapping of the subject noun. In addition, the
constructions given in (32, 33) are not very uncommon either.

(31) aka khuke utconneko

aka-khu-k-e ut=cone-k-o
3sg-drink-fa-imm.pst cl 4=go-fa-dst.pst
‘The boy who drank went away.’ Or
‘It had been a while since the guy who drank it went away.’

However, one Jero speaker, Surmei, did use the relative marker jukhe ~
ukhe preceding the noun it modifies as in the following.

(32) nɔl jukhe ʈoʈa ʈh=ut=thire be

good rel boy 1sg=cl 4=child cop
‘The good (one) boy is mine.’
 syntactic organisation 249

(33) akauno ukhe athire ʈhico khimil be

ɑkɑ-uno ukhe ɑ=thire ʈh=ico khimil be
3sg-sit rel cl 1=child 1sg=gen friend cop
‘The child who is sitting there is my friend.’

For details consult §10.8 below on relativisation.

10.1.10 Interrogatives
As far as the ordering of the interrogative particle/word and its proposi-
tion are concerned, PGA allows two strategies. Languages generally make
a distinction between question particles, words that signal polar questions
and those which are known as WH-words, which occur in content ques-
tions. In PGA the latter occurs in situ, i.e., in the same position in which a
corresponding non interrogative expression would occur (35). The former,
i.e., the interrogative clause of yes-no questions has no specific question
particle or word. Normally a declarative sentence is spoken with a rising
tone to ask the yes-no question. Thus:

(34) ŋ=(ut) thire-ne nɔl be

2sg=cl 4.poss child-pl good cop
‘Are your children fine?’

This finding is rather unusual as other languages of the Andaman, such as

Jarawa and Onge (Abbi 2006a) do have two different lexemes to denote
interrogative expressions and WH-words. The example for the latter is as
follows.

(35) ŋu cay iji-k-o

2sg what eat-fa-dst.pst
‘What did you eat?’

We shall consider interrogatives in more detail later in §10.4.

10.1.11 Adpositional Phrases

The order of the verb and adpositional phrase (postpositional in this case)
is the same as the order of verb and object, i.e., [pp] [V]. Consider:

(36) thire mitɛ khudi er=ŋol-om

child milk for cl 2=cry-npst
‘The child is crying for milk.’
250 chapter ten

(37) ʈhu tajio-bi inci-k-om

1sg fish-abs go-fa-npst
‘I am going for food.’

10.1.12 Variability in Word Order

We must address the question of variability in word order in PGA. As is the
case with most SOV languages, PGA also allows some flexibility, primarily
for focus, in the ordering of the elements in a sentence. For instance, to
indicate focus, the object noun can precede the sentence.

(39) a=meo ŋu ek=tɛro-ø

cl 1=Meo 2sg obj=push-pst
‘You pushed Meo.’ or ‘It was Meo that you pushed.’

Focus can also be indicated by moving the locative object before the
subject.

(40) terkɔ otɔ port blɛr-ak ʈhu thu-ø

Day before yesterday Port Blair-dir 1sg reach-pst
‘I reached Port Blair the day before yesterday.’
or ‘It was the day before yesterday that I reached Port Blair.’

10.1.13 Deviation from the Standard SOV Pattern

PGA deviates from the prototypical SOV languages in positing valency-
increasing markers such as the causative and applicative before the verb
and not after the verb. Refer to §9.6 on causatives in the previous chapter
for more details.

(41) peje thire-bi ata=ŋol-om

Peje child-abs caus=cry-npst
‘Peje is making the child cry.’

Another feature which does not adhere to the paradigm of verb-final

languages is that body division class marker clitics do not appear after
the host but before the host. Similarly, verbal object clitics attach to their
hosts in a preverbal position. Hence, they have been termed as proclit-
ics in the grammar.

10.1.14 Variability in Proclitics

Proclitics in PGA are like prepositional words and carry their own mean-
ing. Every content word, as we have seen so far, is attached to one of the
 syntactic organisation 251

seven proclitics on its left. There are two interesting aspects: one, verbal
proclitics are moveable and can be long distanced from their hosts as was
discussed in §9.5; second, if the preceding word to the proclitic ends in an
open syllable the proclitic combines with it like an enclitic in fast speech
§2.5.7 and §2.5.8 or may be rendered as clitic sequencing as discussed
in §3.8.3. Thus, the first line in the example below represents the pho-
netic realisation which is derived from the underlying form given in the
second line:

(42) nu kona-bit beliŋo

nu konɑ-bi it=beliŋ-o
Nu kona-abs obj=cut-dst.pst
‘Nu cut the Kona (fruit).’ <Object clitic>

(43) thirener benobe

thire-nu er=beno-be
child-pl cl 2=sleep-npst
‘The children are sleeping.’ <verbal clitic>

(44) ʈha tɛkho ɖuocil tɛmborace

ʈh= ɑ=tɛkho ɖuoc-il tɛm-borɑc-e
1sg= cl 1=say hear-conv refl-angry-imm.pst
‘He got angry hearing me.’ <nominal clitic>

(45) ʈhu ɲobir khurobit molo

ʈhu ɲo-bi ɛr=khuro-bi it=mol-o
1sg house-abs cl 2=big-abs obj=demand-dst.pst
‘I wanted a big house.’ <adjectival clitic>

10.2 Coordination

Following Haspelmath (in Shopen 2007: 1), coordinate constructions

have similar units that can combine in a larger unit and can have words
belonging to: (a) any part of speech (verbs, nouns, adjectives, adverbs),
(b) phrases (noun phrases and verb phrases), (c) subordinate clauses or
(d) full sentences. Let us consider all of these here.
The conjunction bo is inserted between two conjoined grammatical
categories. However, a more common phenomenon is to leave out the
use of the conjunction, i.e., there is no overt coordinator and conjoined
units are sequentially conjoined (50, 51).
252 chapter ten

10.2.1 Nominal Coordination

If two nouns conjoined in a coordinate construction are marked similarly
for case then the entire NP takes the case marking. However, the prefer-
ence is to leave the two nouns unmarked for case relations.

(46) sare bo akajira-bi tef-e

salt conj pepper-abs give-imp
‘Give salt and pepper.’

(47) liko bo bea kona-bit=jio-ø

Lico conj Bea kona-abs.obj=eat-pst
‘Lico and Bea ate kona.’

(48) torɔm kajira cɔkbi-thomo ino-l e=tɛʃ-e

salt pepper turtle-flesh water-loc cl 5=put-imp
‘Put salt, pepper and turtle flesh into the water.’

(49) a=joe a=toŋ ɲyo -ak cɔne-b-om

cl 1=Jo cl 1=Tong home-dir go-fa-npst
‘Jo and Tong will go home.’

10.2.2 Pronominal Coordination

Two pronouns can be combined in the same way as two nouns do in a
paratactic form.

(50) ʈhe ŋe ereŋ-khole

1sg.erg 2sg.erg self-play
‘I and you will play.’

10.2.3 Adjectival Coordination

The conjunction marker bo is used between two adjectives or they are
simply juxtaposed without any conjunction. The latter is the preferred
construction. Our observation is that conjoined adjectives are used more
in the predicate position than in the attributive and the preferred con-
struction is to drop the copula (54) in such cases.

(51) a=surmei er=buŋoi bo e=nɔl be

cl 1=Surmei cl 2=beautiful conj cl 5=good cop
‘Surmei is beautiful and nice.’

(52) a=loka er=biŋoi be ara=kata

cl 1=Loka cl 2=handsome cop cl 6–short
‘Loka is handsome and short.’
 syntactic organisation 253

10.2.4 Verbal and Sentential Coordination

Most of the verbal coordination appears to be sentential coordination as
the subject is repeated between the two actions and fulfils the condition
for sentential coordination.

(53) ʈhu rɛfi-bi ra=ʃue-k-o bo ʈhu ʈhi-bi bɛl-o

1sg rice-abs cl 6=cook-fa-pst conj 1sg place-abs clean-dst.pst
‘I cooked the rice and cleaned the place.’

(54) ʈhu ʈh-an-ɖiɖik ʈh=auno bo ʈhu cay-bi kaɲ-o

1sg 1sg-refl-whole day 1sg=sit conj 1sg whatever-abs do-dst.pst
‘I sat for the whole day and I did a whole lot (what did I not do?).’

These constructions are different from the serial verb constructions dis-
cussed below in §10.6.

10.2.5 Adversative Coordination

One affirmative and one negative sentence can be juxtaposed to each
other to represent adversative coordination. There is no specific word for
the disjunction ‘but’ in the language.

(55) m-ara=m-liʃu mut=cone ŋut=conne-pho-ø

1pl-cl 6=refl-all 1pl.cl 4=go 2sg.cl 4=go-neg-pst
‘We all went but you did not.’

Thus, adversative conjunctions or disjunctions are not represented

overtly:

(56) a= lephai-r=cɔk nol e=cai un-tabol-o

cl 1= Lephai-cl 2=face good cl 5=bad refl-naughty-dst.pst
‘Lephai is handsome but bad and naughty.’

(57) ʈho ŋa ci-b-om ʈho kɛle pho-be

1sg 2sg come-fa-npst 1sg stay neg-npst
‘I will come with you but won’t stay back.’

Adversative constructions can involve gapping of the verb.

(58) a= lephai er=eŋkhole-b-om a=bie pho

cl 1= Lephai cl 2=play-fa-npst cl 1=Bie neg
‘Lephai will play but not Bie.’
254 chapter ten

10.2.6 Comitative Conjunction

The comitative conjunction marker is kɔc which is inserted between the
two constituents that are conjoined.

(59) joe kɔc meo uni-b-o

Jo com meo return-fa-dst.pst
‘Jo returned with Meo.’

It has been noticed that speakers drop the comitative marker more often
than not and it is, instead, the context that delivers the appropriate
meaning.

(60) ʈho cao-ne ʃit-b-um

1sg dog-pl hunt-fa-npst
‘We go hunting with dogs.’

10.3 Negation

It behaves like a verb as it is capable of taking TAM markings (61–63). The

negative marker pho or phu is used post-verbally but before the copula be
(64). In our database some examples of negative verbs, as expected, have
the body division class marker 5. However, at present this marker is not
used by all of the speakers.

(61) ʈhu ŋe bɔ i=pho-ø

1sg 2sg.dat ask cl 5=neg-pst
‘I did not ask you.’

(62) eŋkaʈ pho-me

find someone neg-hab
‘(I) normally do not find anyone.’

(63) u ʃup-bi now-pho-m

3sg basket-abs weave-npst
‘He is not weaving the basket.’

Just like any other verb form the negative lexeme can define a state and
thus can be followed by a copula.

(64) e=cul oco pho be

cl 5=have net neg cop
‘(He) does not have a net.’
 syntactic organisation 255

One of the Jeru speakers used the negative word khum instead of pho for
negating the psychological verbs such as bɔ ‘understand’ or PGA bɔ ‘ask’
Consider:

(65) ʈhoŋa tɛkho-bi bɔ-khum-om

1sg.2sg language-abs understand-neg-npst
‘I do not understand / know your language.’

10.3.1 Emphatic Negative Coordination

Negative coordination of the type ‘neither . . . nor’ are represented as coor-
dinated phrases with emphatic negation taipho meaning ‘not at all’ or
‘absolutely not.’

(66) ʈh=ut=thu aʃio taipho ʈh=aːra=belo taipho ʈh=ɛrlaː-b-om

1sg=cl 4=born after neg 1sg=cl 6=earlier neg 1sg=alone-fa-npst
Literally: ‘No one is born after me, no one earlier, I am alone.’
‘There is no one either younger or older than me, I am all alone.’

10.3.2 Prohibitive Negation

Prohibitive imperative markers -bim ~ -em ~ -um are suffixed to the verb
root. As vowel clusters are allowed in the language (refer to §2.1.4) the
negative um in (68) forms a distinctive syllable.

(67) ŋ.ut= kona-bi beliŋ-em

2sg.cl 4= tendu-abs cut-prohb
‘Do not cut tendu.’

(68) ŋu bei-bi khu-um

2sg bottle-abs drink-prohb
‘You should not drink.’

(69) khuol ŋu ʈha=ce-bim

there 2sg 1sg.cl 1=come towards-prohb
‘You need not come to me there.’

(70) ŋu beno-bim
2sg sleep-prohb
‘Do not sleep.’

(71) iku (jukhi) aʈ-bi ɛr=ban-em

burn agent/rel wood-abs cl 2=hold-prohb
‘Do not hold the burnt wood.’
256 chapter ten

(72) ino-bi cay im

water-abs bad prohb
‘Don’t dirty the water.’

To show respect to the addressee the Jeru speakers use the negative verb
phu instead of the prohibitive negative marker.

(73) ŋu kaːʃar-bi khu-phu

2sg tea-abs drink-neg
‘(Please) don’t drink tea.’

10.4 Interrogatives

As stated earlier in §10.1.10, PGA has many WH-words, content question

words, but none like an interrogative particle, a word that signals polar
questions. Polar questions are framed by raising the tone of the declara-
tive sentence. Thus, the PGA version of ‘are you going to hunt’ will be the
equivalent of ‘you are going to hunt’ spoken with a rising tone.

(74) ŋu ʈh-i lɛc-bi tɛʃ-am

2sg 1sg-dat arrow-abs give-npst
‘Will you give me an arrow?’

(75) kitakhi ŋ=aono be

sometime 2sg=sit imp
‘Will you sit here for some time?’

The question words occur in the same position in which a corresponding

non interrogative expression would occur.

(76) ŋu cay-bi kaɲ-o

2sg what-abs do-dst.pst
‘What did you do?’

Another interesting feature is that PGA makes a distinction between

human and non-human referents (§7.4). The human interrogative pro-
noun is aʃiu which is marked for person. The non-human question word is
cay, which is productively used for deriving all kinds of WH constructions.
It is used for both ‘what’ and ‘where.’ When it connotes the meaning of
‘where’, the locative marker -l is suffixed to the question word, thereby
giving it the meaning ‘what in/at’ whereas the term ʃetani stands for ‘how.’
 syntactic organisation 257

Another interesting fact is that forms for indefinite pronouns and ques-
tion words are identical. See §7.4 for details. The question word cay can be
followed by any case marker as the situation demands; cay-bi ‘what-abs’,
cay-ta ‘what-abl/inst’, cay-khudi ‘what-dat’, etc.

(77) ŋu cay-bi ra=ʃue-ø

2sg what-abs cl 6=cook-pst
‘What did you cook?’

(78) meo-bi cay-ta et=pho-ø

mango-abs what-abl obj=cut-pst
‘With what did you cut the mango?’

(79) a=thire ca:y-khudi ŋolo-b-om

cl 1=child what-dat cry-fa-npst
‘Why is the child crying?’ or ‘What for the child crying?’

(80) di aʃiu be
3sg.dem who cop
‘Who is he?’

(81) a=caʃiu-nu porʈblɛr-ak nu eɲo be

cl 1=who-pl Port Blair-dir 3pl visit npst
‘Who all will visit Port Blair?’

The question word combined with the relative marker can be the final
word in the sentence. This means it can take the predicate position as
this is a verbless clause.

(82) ɖut= ɲo cay-jukhe

3sg.cl 4= house what-rel
‘Which one is your house?’

There are the modal interrogative words ʃitani and khitani that convey the
meaning of ‘how.’ The order of this word is not rigidly fixed as it can occupy
the initial position as well as the second position in the sentence.

(83) khitani ŋiyo be

how 2sg.exist cop
‘How are you?’

(84) ʃetani ŋe eɲo-be

how 2sg come-npst
‘How do you come?’
258 chapter ten

(85) ŋu ʃetani celɛcmo (u)t=lukhui-bit phɛn-e

2sg how bushes cl 4=high-abs.obj= jump over-imm.pst
‘How did you jump over the high bushes?’

(86) ŋut= thire ʃetani eɲo-be

2sg.poss4= child how come-npst
‘How does your child come?’

As stated earlier, the question word cay is suffixed with the locative
marker -l to represent a locative interrogative such as the equivalent of
English ‘where’. Thus, PGA has cay-l but since the metathesis rule applies
(see §2.5.9), the word is rendered as cyal ~ cya:l ‘where.’

(87) cya:l ut=ɲo be

where cl 4=live cop
‘Where does he live?’

(88) cyal ŋu a=jo e=ole-b-o

where 2sg cl 1=Jo cl 5=see-fa-dst.pst
‘Where did you see Jo?’

(89) cya:l ŋ=ut=coŋ-o

where 2sg=cl 4=find-dst.pst
‘Where did you find it?’

10.5 Non-finite Subordination

PGA uses non-finite strategies for building complex syntactic struc-

tures. In this section we shall consider all those constructions which are
derived by combining two verbal clauses in which the second clause in
the sequence is finite and the first one is either unmarked (symbolized
by ø) or is marked by the homophonous past marker of the narrative style
-il (see §9.8.1) to indicate the adverbial subordination to the main verb in
the second clause. The first verb in the two verb sequence is non-finite.
The explanation with examples for verb serialisation is given below. Non-
finite conjunctive verbs, those that conjoin two clauses, are symbolized
as conv.

10.5.1 Verb Serialisation

The combination of non-finite and finite verbs may denote any of the
following readings:
 syntactic organisation 259

(a) S equential or conjunctive. If the relationship between the two verbs is

of conjunctive denoting the adverbial function of ‘after’, the first verb
is suffixed by the homophonous past marker reserved for narration,
i.e., -il or -l. (90, 91). It can optionally appear in the (unmarked) simple
past tense (92, 93). These verb forms serve the functions of converb
as discussed previously in §9.8.1. The second and the final verb can
take any of the tense markers as the situation demands.

(90) o ʈh=ɛ boʃo-l tɛbol-ø

3sg 1sg=abs hit- conv run away-pst
‘After hitting me he ran away’

(91) ʈha=mai aka ɲa-l u khijire-kak ot=cone-ø

1sg.cl 1.poss=father 3sg eat-conv 3sg stroll-dir cl 4=go-pst
‘After eating (his food) my father went for a stroll.’

(92) khudi bo o lec-ik=tɛr=tola-ø eka ʈhι

3sg. prox2 and 3sg arrow-obj=cl 2=shot-conv 3obj land
(u)t=bol-o
cl 4=search-dst.pst
‘He searched for the shot arrows.’ (He shot more arrows, he searched for
them)
‘Having shot more arrows he searched for them on the land.’

(93) u kona ʈob-ø tɛbɔl-o

3sg tendu steal-conv run away-dst.pst
‘Having stolen tendu fruit he ran away.’

(b) C
 ause and effect. If the relationship between the two events is one
of cause and effect then the first verb is unmarked and the final verb
takes any of the tense markers as the situation demands (94).

(94) meo-bit bim thire nu ekhuku

meo-bi it=bim-ø thire-nu e=khu-k-u
orange-abs obj=squeeze-conv child-pl cl 5=drink-fa-dst.pst
‘Having squeezed the orange children drank it (the juice).’

(c) Purposive. If the final action is undertaken in order to fulfil a pur-

pose indicated in the action/event in the first verb then it may be
unmarked. The second and final verb can take any of the tense mark-
ers as the situation demands (95–97).
260 chapter ten

(95) em-boi kɔʈ-e ekjira-ø inci-l

refl-wife Kot-abs tell-conv go-nrr.pst
‘(He) went to tell his wife Kot.’ (Purposive)

(96) ʈh=ambikhir taracɔr-bi eole-ø inci-k-o

1sg=refl. morning spring-abs see-conv go-fa-dst.pst
‘Yesterday I went to see the spring.’ (Purposive)

(97) a thire phir ʃup bi konarlub incikom

ɑ=thire phir-ʃup-bi konɑ er=lub- ø inci-k-om
cl 1=child cane-basket-abs tendu cl 2=pick-conv go-fa-npst
‘The child goes with a cane basket to pick up tendu fruits.’ (Purposive)

(d) V
 erb serialisation can also be used as the first verb in the sequence
performs the manner in which the final verb is undertaken (98). In
such cases the first verb is suffixed by the homophonous past marker
reserved for narration, i.e., ‑il or -l and performs the function of the
converb.

(98) phertajiɖo e=maʈ-il unci-ø

Phertajido cl 5=run-conv go towards-pst
‘Phartajido went towards (her) running.’ (Manner)

(e) The subject identity is maintained in most of the examples unless the
sequence of events are undertaken by two distinct subjects (99–100),
in which case the construction represents chaining of two clauses.
Refer to §10.6.

(99) ŋu khiel ʈhu inci-l ʈh=ut=bɔ nɔl-pho-be

2sg here 1sg come-nrr.pst 1sg=cl 4.poss=heart good-neg-cop
‘I do not like your coming here’ (you came to me, I do not like).’

(100) meʈhibitphilel mərambɔphome

me-ʈhi-bi it=philel mɑ-ɑrɑm-bɔ-pho-me
1pl-place-abs obj=flood-nrr.pst 1pl.excl-refl-know-neg-hab
‘How our place flooded with water, we do not know.’

It can be generalised that if the two actions are undertaken by the same
subject the first event in a sequence of two events serves an adverbial
function followed by the second verb in a finite form with full TAM
markings.
It is observed that PGA does not have the prototypical specific word for
the complement conjunction ‘that.’ Hence, all direct speech involving the
 syntactic organisation 261

complementiser like the English ‘that’ is rendered as clause chaining (see

§10.6) without any intervening element in them.
However, there are two types of clauses that can be considered as com-
plement clauses from a semantic point of view. These employ the comple-
mentiser of ‘reasoning’ khude and the ‘conditional’ amo. The structure of
these can be described as:

(S 10.1) subordinate clause_comp_main clause

These are discussed below.

10.5.2 Cause and Effect

Clauses with a cause and effect situation or reasoning precede the main
clause. There is a comp khude ‘because’, ‘that’s why’, which is a link
between the two.

(101) milɛnephukhude ʈhumeteimulu phukhu

milɛne phu khude ʈhu metei-mulu phukhu
sweet neg comp 1sg milk-sugarless neg drink
‘Because there is no sugar, I will not drink milk.’

Sentence (101) can be rendered without the complementiser khude as an

example of clause chaining. See §10.6. Some more examples with comple-
mentiser links are:

(102) loka euretɛrcekpho khude o ektɛrʃɔropho

lokɑ eure-tɛrcek pho khude o ek=tɛr=ʃɔro pho
Loka body-strength neg comp 3sg 3obj=cl 2=push neg
‘Because Loka has no strength, he cannot push.’

In the following construction the subject is not overtly marked but it is

understood.

(103) e=buʈhu khude o=ʃolo pho

cl 5=weak comp cl 7=crawl neg
‘Because it was weak it could not crawl.’

10.5.3 Conditional Clauses

Conditional clauses are marked by the conditional marker -amo which is
suffixed to the root verb of the subordinate clause. As expected, the main
clause follows the subordinate clause.
262 chapter ten

(104) jicerbira kɔrophamo ʈhutconebe

jicer-bi rɑ=kɔroph-amo ʈh=ut=cone-b-e
rain-abs cl 6=stops-cond 1sg=cl 4=go-fa-imm.pst
‘If the rain had stopped for a while, we could have gone.’

(105) ŋu ʈh=i miʈhai-bi phu tɛʃ-amo ʈho ŋolo-b-om

2sg 2sg-dat sweets-abs neg give-cond 1sg cry-fa-npst
‘If you do not give me the sweets I will cry.’

(106)1 tujul-(u)l ŋu ʈhu bik=amo ŋ=a=ciʈhi-bi ole-koliŋe

early-loc 2sg reach abs.obj=cond 2sg=cl 1=letter-abs see-subj
‘If you had reached (here) earlier you would have seen the letter.’

(107) o ʈh=o=bɔi-amo ʈhu i=tɛrta-k-om

3sg 2sg=cl 4=ask-cond 1sg cl 5=narrate-fa-npst
‘If he asks me I will tell him (the whole story).’

(108) u ʈhu inci-k-amo u ʈhu-n-ce

3sg 1sg come-fa-cond 3sg 1sg-refl-come.imp
‘If he can come he should come to me.’

(109) sarephile-k-amo ŋu kɛo-bi it=cɔŋg-om

tide-fa-cond 2sg crabs-abs obj=find-npst
‘In case of tide you will find crabs.’

PGA is losing the complement markers -amo and -khude as these construc-
tions are being replaced by clause chaining. The context always serves as
the clue for the appropriate interpretation.
While discussing verb sequences one cannot ignore the action nominals
in the language as they are derived from non-finite verbs and indicate ‘pur-
pose’ of the action signified by the finite verb. Let us consider them now.

10.5.4 Action Nominals

Following Comrie (1976: 178) action nominals are nouns derived from
verbs with the general meaning of an action or process and are capable
of taking prepositions or postpositions in the same way as non-derived
nouns. The absolutive marker -bi in PGA is homophonous to the nomi-
naliser suffix -bi and is attached to a verb root to convert it into a nomi-
nalised form. This is a very productive strategy in the language to derive
nouns from verbs. It also indicates the state of the object. Consider:

1 This is the only example of the subjunctive mood in our database. Enquiries regarding
getting more constructions of this type did not bring any results.
 syntactic organisation 263

(110) ʈɔya-bi ŋale a=malai-o

stand-nmlz 2pl cl 1=tire-dst.pst
‘You all got tired while standing (for a long time).’

(111) ʈhu beno-bi inci-k-o

1sg sleep-nmlz go-fa-dst.pst
‘I went away to sleep.’

(112) ʈhu taŋa-bi inci-k-o

1sg eat-nmlz go-fa-dst.pst
‘I went to eat.’

(113) ŋu ŋoʈo-bi ut=cokh-om

2sg swim-nmlz cl 4=well-npst
‘You can swim well.’

10.6 Clause Chaining

PGA offers ample examples of clause chaining or clause linking, whereby a

number of finite clauses referring to distinct but related events are placed
together in a sequence. The main clause is placed last in the sentence.
Clause chaining is evident in reported speech where two sentences
appear to be joined together without a link. The statement sentence or
reported speech follows the main clause, i.e., the reporting speech with
a reporting verb ‘say’, ‘tell’, etc. Consider when the reported speech is a
declarative sentence:

(114) akarkha tɛkhamo u portblair etmokom.

ɑ=kɑrkhɑ-ø tɛkhɑmo u port blɛr et=ok-om
cl 1=tell-pst soon 3sg Port Blair obj=leave-npst
‘(He) said that he will leave Port Blair soon.’

(115) o akarkha rɛfiraʃue ɲyoak enoleme

o ɑ=kɑrkhɑ-ø rɛfe rɑʃue ɲyo-ɑk ɛ=nole-me
3sg cl 1=say-pst rice cook home-dir cl 5=good-hab
‘He said that cooking at home was good.’

(116) naŋe ŋu-i(k)=khider-e tɛʃa-l ŋu-i=

3pl.abs 2sg-obj=coconut-abs give-nrr.pst 2sg-cl5=
ŋ=ot=oborɛ-pho
2sg=cl 4=forget-neg
‘You did not forget that you gave them a coconut.’
264 chapter ten

The structure of juxtaposing the main clause, i.e., reporting speech and
the reported speech exists for all kinds of subordinate clauses involving
interrogatives, imperatives, and negatives in the reported speech.

(117) o jira-b-o phɛtɛr-il ŋa-uno be

3sg say-fa-dst.pst stone-loc 2sg-sit imp
‘He asked me to sit on the stone.’

(118) abhishek jira-b-o ʈamʈam berebe ɛr=bathe pho

Abhishek tell-fa-dst.pst Tamtam Birebe cl 2=slap neg
‘Abhishek told that Tamtam did not slap Birebe.’

(119) u ʈh=e= boi-om ŋ=ut=ʈhuthu a=ʃyu bi

3sg 1sg=cl 5= ask-npst 2sg=cl 4.poss=brother cl 1=who cop
‘He asks me who my brother is.’

(120) jicer-bi cer-om ʈh=ut=cone-pho-be

rain-abs rain-npst 1sg=cl 4=go-neg-npst
‘I will not leave since it is raining.’

Consider the following sentence where the subject pronoun (ʈhu) ‘1sg’ is
gapped.

(121) ʈha=mai jiret=al tocai-be-(ʈhu)

1sg.cl 1.poss-father tell.obj=nrr.pst whatever-abs
(i)t=likho-k-om
(1sg)-obj=make-fa-npst
‘Whatever my father told me I do it.’ or
‘I do exactly the same way as told by my father.’

Two imperative forms can be combined to represent a joint action. Here,

both verb forms take the imperative suffix {-be ~ -e}. However, since the
two verbs conjoined here belong to two different body division classes
the subject pronoun is obligatorily repeated. In English it would be con-
sidered an example of sentential coordination.

(122) ŋ=ut=cone be ŋ=e=eole e

2sg=cl 4=go imp 2sg=cl 5=see imp
‘You go and see.’

10.7 Comparative Constructions

There seems to be no overt morphological marking for comparative or

contrastive or superlative degrees. In our database we have very few sen-
 syntactic organisation 265

tences pertaining to this phenomenon. The mixed nature of the language

gives alternative structures and we do sometimes obtain a marker, the sta-
tus of which cannot be established as ‘comparative.’ The marker of com-
parison that appears sometimes is -ta ~ -te, which is homophonous to the
ablative/instrumental marking and is suffixed to the noun of the standard
of comparison. This results in three alternating structures.

(a) In the following sentence, as expected, the comparative marker is

post-nominal.

(123) ŋ=ot=ɲo-ta di ɲo nɔl totcyo be

2sg=cl 4=house-compr 3dem house good very much cop
‘This house is better than your house.’ or
‘This house is the best of all.’

(b) Alternatively, the standard of comparison can be moved to the final

position in the sentence without a finite verb, a rare phenomenon.

(124) inotaracɔr ɖuroʈmo bi ʃareta

ino-tɑrɑ=cɔr ɖuroʈmo bi ʃɑre-tɑ
water-cl 6=spring clean cop sea-compr
‘Spring water is cleaner than the sea water.’

Although the two sentences conjoined by the copula be below appear

to be independent sentences, they are grammatically and syntactically
related and conjoined as the proclitic appearing in the second sentence
co-refers to methil ‘dirty’ in the previous sentence.

(125) port blɛr methil be ɖiglipuril taipho

port blɛr methil be ɖiglipur-il tɑ-i=pho
Port Blair dirty cop Diglipur-loc compr-cl 5=neg
‘Port-Blair is dirtier than Diglipur.’ or
‘Port Blair is dirty, Diglipur is not.’

(c) Thirdly, there are constructions which match exactly the pattern of
Hindi syntax. Consider the following.

(126) ŋarajulu-te ʈharajulu ʈɔlɔʈmobe

ŋ=ɑrɑ=julu-te ʈh=ɑrɑ=julu ʈɔlɔʈmo be
2sg=cl 6.poss=clothes-compr 1sg=cl 6.poss=cloth white cop/exist
‘My clothes are whiter than yours.’
266 chapter ten

Constructions for indicating kinship relations younger and elder than the
speaker are either derived by primary possessive (127) or by secondary
possessive (128) devices as seen in Chapter 6 on possession.

(127) ʈh=ara=ʃulu
1sg=cl 6.poss=later
‘Younger than me.’

(128) ʈh=ut=toa-thu
1sg=cl 4–=before-born
‘My elder brother.’

(129) u-tun-toa-thu toka phu

3sg- refl-before-born like neg
‘(He) is not like his elder brother.’

Two consecutive sentences/phrases are juxtaposed to each other with

contrastive sets of adjectives. The sentences conjoined in this way may
have one finite verb at the end, or none of the phrases have finiteness
and hence they may be verbless constructions (131).2 These are considered
non-verbal clauses (Dryer 2007: 224).

(130) ajoeutkata anarayɔnertɔy lobuŋbe

ɑ=joe-ut=kata ɑ=nɑrɑyɑn er=ʈɔy lobuŋ be
cl 1=Joe-cl 4=dwarf cl 1=Narayan cl 2.poss=leg long be-exist
‘Joe is short. Narayan is tall.’

(131) diglipur tot=lukhui porʈblɛr tot=letamo

Diglipur cl4=high Port Blair cl4=flat
‘Diglipur is high. Port Blair is flat.’ Or
‘Diglipur is higher than Port Blair.’

Sentence (66) of emphatic negation coordination considered above repre-

sents another example of a comparative construction. It is repeated here
for convenience.

(132) ʈh=ut=thu aʃio taipho ʈh=aːra=belo taipho

1sg=cl 4=born after neg 1sg=cl 6=earlier neg
ʈh=ɛerlaː-b-om
1sg=alone-fa-npst

2 This means PGA has both non-verbal predicates as discussed in Chapter 3 and exem-
plified elsewhere in the grammar as well as non verbal clauses.
 syntactic organisation 267

Literally: ‘No one is born after me, no one earlier, I am alone.’

‘There is no one either younger or older than me, I am all alone.’

10.8 Relativisation

We briefly mentioned in §10.1.9 that the relative marker is not used much
except in Jeru. Other Great Andamanese languages use paratactic strat-
egies and participial constructions. In all, PGA uses the following three
structures for relativisation described below in (S 10.2–4):

(S 10.2) (ADJ) REL_N

There is a special lexeme jukhe ~ ukhe, which is a relativiser morpheme

that specifies the following noun and is optionally preceded by an adjec-
tive or a verb. It serves as a restrictive relative marker. This construction
was observed only in the speech of Jeru speakers as the other speakers
preferred to drop the relativiser from the sentence. It should be noted that
the order of the modifier and modified is reversed in these constructions.
The prototypical order, as we have seen so far is modified-modifier.

(133) akauno ukhe a=thire ʈh=ico khimil be

3sg.sit rel cl 1=child 1sg=gen friend cop
‘The child who is sitting is my friend.’

(134) nɔl jukhe ʈoʈa ʈh=ut=thire be

good rel boy 1sg=cl 4=child cop
‘The good (looking) one is my child.’

One of the readings of the following sentence is superlative.

(135) nɔl tocyo jukhi ɲo ʈh=ut=ɲo be

good very rel house 1sg=cl 4.poss=house cop
‘The house which is the best is mine.’

(136) ʈoya ukhe a=thire ta=ben-om

stand rel cl 1=child appl=sleep-npst
‘The child who had been standing went off to sleep.’ or
‘The standing child is dozing off.’

Constructions without the relativiser appear to be simple adjectival con-

structions but with reversed ordering, i.e., the modifier precedes the noun
rather than following it. In such cases it is always the verb that modifies
268 chapter ten

the following noun and the construction is considered a relativised nomi-

nal construction.3 Consider:

(137) e=tei-ka ʈoʈa emphil-o

cl 5=fever-gen boy die-dst.pst
‘The fevered boy died.’ / ‘The boy who had a fever died.’

The second structure for relativisation is of the following type:

(S 10.3) S_S

This is a case of sentential modification, where in a sequence of two finite

clauses, the previous one modifies the following one and the subject is
elided in the second clause provided it is shared.

(138) ko tɛr=ʃolo-k-a ø i=boʈho-ø

bow cl 2=hang-fa-imm.pst cl5=fall- pst
‘The bow which was hanging fell down.’

(139) a=to-ra be ʈh=ambikhir ø eol-a

cl 1=emph-pig cop 1sg=refl.morning see-imm.pst
‘This is the pig that I saw this morning.’

(140) ʈɔkhot-o(t)=ʈɔŋ-el ʈh=o(t)=ʈhomo e=boʈho-ø ø

piece-cl 4.poss=wood-loc 1sg=cl 4.poss=cap cl 5=fall-pst
ɛr=ʃolo-k-e
cl 1=hang-fa-imp
‘Hang my cap which fell down from the wooden nail.’ or
‘My cap, which was on the wooden nail, fell down, (please) hang it
(back).’

The third structure is very similar to the Dravidian and Indo-Aryan

languages.

(S 10.4) NP (Vstate)

It is a non-finite participial relativisation which modifies the preceding

noun phrase. The verb root is suffixed by -e and defines the state of the

3 These constructions are similar to participial constructions found in Dravidian languages

or in those Indo-Aryan languages which are influenced by Dravidian, such as Konkani.
 syntactic organisation 269

noun preceding it. Thus, it is post-nominal modifier as expected. The

modifier verb is in a non-finite form without any TAM markings. The par-
ticipial modification can modify the noun in A function as well as in O
function. Consider:

(141) a=ʈoʈa lɛc-ik= unɖu-k-e untɛl-e

cl 1=boy arrow-obj= break-fa-pcpl call-imp
‘Call the boy who broke the arrow.’

(142) lɛc-ik= unɖu-k-e e=kao-e

arrow-obj= break-fa-pcpl cl 5=bring-imp
‘Bring the broken arrow.’

(143) u khidire i=khu-e a=kaʈa untɛle-be

3sg coconut cl 5=drink-pcpl cl 1=girl call-imp
‘Call the girl who drank coconut (water).’

The following example does not have the verb ‘cut’ in the participial but in
the root form. It acts as a verbal modifier modifying the ‘pig’ ra. This looks
remarkably similar to the Dravidian relative participle strategy which has
no accessibility hierarchy restrictions.

(144) [ra [[tara=uli-bik=] tokata] [ɛrcek]]

[pig [[cl 6.poss=tail-abs.obj=] cut] [naughty]]
‘The tail-cut pig is a nuisance.’

10.9 Syntax of the Possessive Construction

Possessive phrases can take part in various kinds of constructions. These

are given here simply for the reason that readers are informed of the fact
that PGA follows the standard typology of languages of India.

10.9.1 Possessed NP as the Head

The possessed NP is the head of the phrase. All the examples of R - class
marker -D seen in Chapter 5 can be employed in a sentence with D work-
ing as the head of the phrase and the clause. For instance:

(145) ʈh=ara=kara-bik= ie-k-o

1sg=cl 6.poss-waist-abs.obj= pain-fa-dst.pst
‘My waist pained.’/ ‘There was pain in my waist.’
270 chapter ten

(146) ʈh=a=kɛr ca:e

1sg=cl 1.poss=throat bad
‘My throat is bad.’ / ‘I have a sore throat.’

10.9.2 Possessed NP as Object

The possessive construction can operate as an object NP of the verb
phrase as in:

(147) ʈh=ɛr=beŋ-bi it=oɖu-k-e

1sg=cl 2.poss=forehead-abs obj=paste-fa-imp
‘Put ointment on my forehead.’

(148) ʈho [ŋ=a=tɛkho-bi] ik=jira-k-om

1sg [2sg=cl 1=speech-abs] obj=speak-fa-npst
‘I speak your language.’

10.9.3 Identification Question

An identification question can be asked by a possessive construction.

(149) ŋ=er=liu a=ʃiu bi

2sg=cl 2.poss=name cl 1=who cop
‘What is your name?’

10.9.4 Complement Phrase

A possessive NP can operate as a complement phrase/clause:

(150) ʈh=u(t)=mikhu-l a=thire be

1sg=cl 4.poss=lap-loc cl 1=child cop
‘There is a child in my lap.’ ‘The child is in my lap.’

(151) ŋ=er=ka:ra ut=conne-b-om

2sg=cl 2.poss=birth place cl 4=go-fa-npst
‘You leave for the place where you were born.’

10.9.5 Verbless Constructions

A possessive phrase can exist without a copula and without a verb phrase
and in this way it is a complete sentence and contains information
structure.
 syntactic organisation 271

(152) thire er=liu bea

child cl 2.poss=name Bea
‘The child’s name is Bea.’

(153) kheŋe er=uli a=kaʈa

cat cl 2.poss=tail cl 1=short
‘The cat’s tail is short.’

To summarise the discussion for this section and the others considered
above, it seems that the impending death of the language is forcing the
language to adopt more paratactic structures and to do away with embed-
ding and complex structures. The simplification could also have been the
result of the ‘mixed’ nature of the language as variation is very evident
and to mitigate the variation speakers accommodate to simple structures.
One could also hypothesise that the simple structures are the remnant of
the old Great Andamanese language. The anthropocentrism that is repre-
sented in the body division classes pervading all content words appears
to be the clue for the initial stages of language evolution. The antiquity of
the language, genetic study of the community and the long isolation of the
community with the rest of the world add to this hypothesis.
 Appendix A
Appendix A
Lico’s genealogical a���liation and her language pro��le
Lico’s genealogical affiliation and her language profile

Biological maternal grandfather (Burmese) Biological maternal grandmother (Sare)

Loka (daughter) = Jirake (Bo)

(Mother died at her birth) Lico = Golat (Khora)

After 11 years Jirake married Surmei (Jeru)

Figure A1

Lico was brought up by foster parents whom she regarded as grandpar-

ents. They were:
Loka (Pujjikar) grandfather = Kobo (Sare) grandmother
She also had contact with her biological grandmother but was closer
to her foster grandmother. Her speech is largely of the Sare variety but
should have traces of Khora, Bo and Jeru. She knows Jeru better than the
other two as her father Jirake, who hailed from a Bo tribe mainly spoke
Jeru, at least claimed that he was a Jeru speaker.

Note: Names in parentheses indicate names of the tribe and the language
of the person. The symbol = means ‘married to.’
 Appendix B

An Acoustic Study of problematic lateral

The Laterals in Peje’s Speech

As noted in Chapter 2 one of the speakers, Peje, always used a labialised

lateral instead of a simple alveolar lateral [l]. An acoustic study was car-
ried out to analyse the sound. The following words which contain laterals
in initial, medial and final positions were elicited and recorded with all
the speakers. We also compared Peje’s speech with our main consultant,
Nao Jr. (male 55). A labialised lateral is clearly evident in Peje’s speech,
which is totally missing from the speech of the other speakers.

Methodology

The following words were tested to analyse initial, medial and final later-
als across speakers.

1. [lƐ:c] ‘arrow’ [ʈhƐphilu] ‘stomach’ [phal] ‘wave’

2. [luɽo] ‘fire’ [kɔ:lo] ‘bamboo’ [ʈhumƐl] ‘bee’
3. [lico] ‘basket’ [ilphe] ‘turn over’ [bol] ‘rope’
4. [laʈ] ‘fear’ [ɖulo] ‘moon’

Peje’s articulation of [l] was video-graphed. We also used a digital recorder

to record Peje and the other speakers. Instances of initial, medial and final
[l] in the words above were recorded to compare variation in the sound
[l]. The word lƐ:c ‘arrow’, as articulated by Peje and Nao Jr, is discussed
below.

Analysis

The laterals in Peje’s speech are the result of a combination of two dif-
ferent articulations, lateral [l] and an approximant [w]: the former is an
alveolar lateral and the latter is a bilabial approximant. In other words, the
tongue is positioned as if to produce a lateral, but the lips are configured
276 appendix b

Figure B1a–b. The spectrogram of lƐ:c ‘arrow’ as articulated by Peje

Figure B2a–b. The spectrogram of lƐ:c ‘arrow’ as articulated by Nao Jr.

as if to articulate a bilabial approximant. This is evident from the spectro-

gram where he articulated the word lƐ:c ‘arrow’. See B1a–b given below.
In the word lƐ:c, as articulated by Peje, the sound [l] has around 0.139ms,
whereas for other speakers’ it has around 0.097ms. This is because Peje’s
articulation of [l] is different from the other speakers. The narrow and
wide band spectrograms, of Peje in Figure B1a-b, and Nao Jr. in Figure
B2a–b, are presented above.
 an acoustic study of problematic lateral 277

lɛ:c arrow_mono
0.0987723133 0.23835005
5000
F4

4000

3000 F3

2000
F2

1000

F1
0
0.09877 0.2384
Time(s)

Figure B3. Formant Structure of [l] (Peje)

lɛ:c_arrow_Nao_mono
0.412049691 0.495232858
5000

F4
4000

3000 F3

2000
F2

1000
F1

0
0.412 0.4952
Time(s)
Figure B4. Formant Structure of [l] (Nao Jr.)
278 appendix b

Formant Structure

Acoustically, laterals have a formant-like resonance such as is found

in the vowels (Ladefoged & Maddieson 1996). Thus, it is important to
compare the formant structures of the laterals. The formant structure of
Peje’s [l] is different from Nao Jr.’s [l]. The formant structure of Nao Jr.’s [l]
is quite clear and straight, while Peje’s [l] has some deviations, see Figures
B3 (Peje) & B4 (Nao Jr.) given above. The labiality in the [l] pronounced
by Peje is clearly evident from the transitions plotted in formant tracings
in Figures B3 and B4.
 Text

The Great Narrative of Phertajido

This is a creation myth and was told to me by Nao Jr.1 on the night of 21st
January 2006.2 Nao was so fascinated by this particular story that he never
tired of telling it again and again as he thought this was one of the greatest
love stories that he ever heard in his life. He told me that creating one’s
own partner according to one’s own liking was the best part of the story.
In many respects it is a complete story as it mentions five basic elements
of life, i.e., fire, water, earth, sky/space and air.

1. phertajiɖo o tarphuc tunʃoŋ ot=untoplɔ ji-o

Phertajido 3sg Andamanese human cl 4=alone/first live-dst.pst
‘Phertajido was the first human from Andaman.’

2. phɔr kotra-ta thu-o

Bamboo from within-abl born-dst.pst
‘He was born out of bamboo.’

3. uɽoʈɔy-il eliu phertajiɖo

ancient times-loc his.name Phertajido
‘In ancient time, he was named Phertajido.’

4. eŋkalel eŋkaʈ-pho-me eremla ʈhit=ɲyo-ø

here and there find someone-neg-hab alone place.obj= live-pst
‘Finding no one around, he lived a solitary life.’

5. kacole-me
shave the arrow [make]-hab
‘He used to make bows and arrows.’

6. o-i= ta-eŋkocil-el o itaci-ø ek=tɛrʈok-ø

3sg-obj= appl-make arrow-conv 3sg test-pst obj=shoot arrow-pst
‘Having made arrows, he tested them by shooting them.’

1 Nao Jr. died in Port Blair in 2009 due to kidney failure.

2 Subsequently, the story was collected in full form in many sessions by me and Narayan
Chaudhary, the research assistant in the project.
280 text

7. o tuŋkɛlo lɛc-ik= tɛr=tole-me

3sg all over arrow-obj= cl 2=shot (arrow)-hab
‘He shot arrows all around.’

8. a=kambikhir o [lɛc-ik=tɛrtolaoika] ʈhit=bolo

cl 1=3sg.next morning 3sg [arrow-obj=shot behind] place.obj=find
inci-ø
go-pst
‘Next morning he went after the [to search for the] shot arrows.’
‘He went next day to search for the arrows that he had shot.’

9. o it=cɔŋ-el o ino-tara=cɔr et=cɔŋo-ø

3sg obj=get/found-conv 3sg water-cl 6.poss=source obj=find-pst
o- i=khu-l ino-t=khu cɔŋ-o
3sg cl 5=drink-conv water-cl 4.poss=drink found-dst.pst
‘Having found one (arrow), he found a water source. He drank the water and
thus discovered drinking water.’

10. bo lɛc-ik= kac-il o lɛc-it=cɔŋ-ø

and arrow-obj= make-conv 3sg arrow-obj=find-pst
mino-tara=cɛʈhul
potato-cl 4.poss=under roots
‘Having made more arrows [he shot more arrows] he found an arrow in the
roots of a potato plant.’

11. o i=phoŋ-il mino-t=cɔŋ-ø

3sg cl 5=hollow-loc potato-obj=find-pst
‘He found potato in the hollow (of the root).’

12. o ʈol-e iebi-ø

3sg potato-abs bring-pst
‘He took/brought the potato.’

13. du bo3 o lɛc i=nɔne-ø ince-ø

remaining 3sg arrow cl 5=search-pst go after-pst
‘He went to search for the remaining arrows.’

14. o it=cɔŋ -il bi-ut=calɔ ut=cɔŋ-ø

3sg obj=find-nrr.pst dhoop-poss4=heap cl 4=got/find-pst
‘On his search, he found lots of dhoop4.’
‘Having found it (arrow) he discovered incense (Hindi: dhuup).’

3 du ‘deictic for invisible distance’ and bo ‘and’, ‘then’ conjunction.

4 Incense.
 the great narrative of phertajido 281

15. o i=kuɖilo untɔplɔ iebi-ø

3sg cl 5=small piece single return-pst
‘He returned with a single small piece of incense.’

16. khudi bo o lɛc-ik=tɛrtola eka ʈhi-t=bɔlo-ø

3sg.prox2 and 3sg arrow obj=shot 3sg ground-obj=search-pst
‘He went to search for more of the shot arrows on the ground.’
‘He searched for more shot arrows on the ground.’

17. o it=coŋ-il kɔʈ-e=mɛle ʈhit=cɔŋ-ø

3sg obj=find-nrr.pst soil-cl 5=fine earth.obj=find-pst
‘He found it, the Kaut,5 a very fine soil.’

18. o it=cɔŋ-il u iebi-ø

3sg obj=find-conv 3 bring-pst
‘He found it, (Having found it) he brought it over.’

19. u ik=khuni-ø
3sg obj=return-pst
‘He (took some of it and) returned.’

20. kɔʈ ek=phɛc i=bano-ø

soil obj=pot cl 5=make-pst
‘(He) made a pot/vessel out of the soil.’

21. o i=ta=phai-ø
3sg obj=tr=dry-pst
‘He dried it.’

22. irem lam-il phɛc ta=ʈol er=aʈ-ɛʃe

refl hard-conv pot cl 1.poss=potato cl 2=put on fire.imm.pst
‘It (the pot) hardened, he cooked the potato of/in the pot on the fire.’
‘Having hardened the pot, he cooked the potato in the pot.’

23. phɛc ta =ʈole ra=iʃe iboi u iji-ø

pot cl 1.poss=potato cl 6=cook in the fire boil 3sg eat-pst
‘He put the pot with the potato on the fire, boiled/cooked it and ate it.’

24. u iji-te ut=bɔrʈhu-l

3sg eat-pcpl cl 4=think-nrr.pst
‘While eating, he thought.’

5 kɔʈ is a kind of clay used in making pots.

282 text

25. o-tum arpʰuc teka u ebukhu e=bano-ø

3sg-refl Andamanese like 3sg woman cl 5=make-pst
‘He made a dummy-Andamanese in the shape of woman.’

26. u ʈhical ara=caka6 o ara=aʈ-e kubi-ø

3sg above the earth cl 6=platform 3sg cl 6=fire-abs light-pst
‘He put it [the dummy] above the ground on the platform and lighted a
fire.’

27. ata=pʰai (i)t=nol-ø

tr=dry obj=good-pst
‘He dried it well.’

28. bo koka cɔl-il u tum bɔt khacol o eule-ø

and/while giving final touches-conv 3sgrefl back turn 3sg see-pst
‘While giving the final touches (to arrows) he turned back to see.’

29. ʈhica tuttara:l kɔʈ-tun ta=bino be

platform on Kaut-self appl=sleeping-cop
‘Kaut was lying (sleeping) on the platform.’

30. bo ara=aʈ eku=bi unci-ø

again cl 6=fire obj=light go-pst
‘He went again to kindle the fire.’

31. bo o koka-ø cɔl

and 3sg resum-pst peeling
‘And he resumed peeling the bow.’

32. bo o tum bɔt-khacol ʈhica eule-ø

and 3sg refl turn-back platform see-pst
‘He turned back again to see the platform.’

33. a=mimi a= kɔʈ ata=ʃuɲ-kɛloɲ-il eŋkhile-ø

cl 1=mother cl 1=Kaut cl 1.poss=body-turn side-conv shake-pst
‘The body of the mother Kaut turned sideways (and) shook.’

34. phertajiɖo u tum bɔrco

Phertajido 3sg refl satisfied/happy
‘Phertajido was happy with himself/was satisfied.’

35. ara=n-ʈoya-l u ara=aʈ-ikubi unci

cl 6=refl-stand-conv 3sg cl 6=fire-light went
‘Then he stood up and went to kindle the fire.’

6 Hindi machaan ‘wooden platform’.

 the great narrative of phertajido 283

36. kokacol ik=malai-il u ʃiɖ-ik ut=cɔne-ø

make arrow obj=tire-conv 3sg hunt-abs cl 4=go-pst
‘Having got tired by making the arrows, he went to hunt.’

37. o ra- (ɛ)k-rɛ-l u ikh-unni ɲyo-ak

3sg pig-obj=-hunt-conv 3sg came back home-dir
‘Having found the game, he came back home.’

38. akaŋ tut=ole-l ʈhica-kak

around cl 4.poss=see-nrr.pst platform-dir
‘He glanced at the platform/ He looked over to the platform.’

39. ʈhica tɛr=lɔkho-ø

platform cl 2=empty-pst
‘There was nothing on the platform/platform was empty.’

40. u tuŋ-bo kalui-il

3sg refl-heart break-nrr.pst
‘He was heart-broken/he became melancholic.’

41. o ra talɛ-l cyak khidi kɔʈ bel-ɔ

3sg pig put down-conv where dem.dis Kaut get lost-dst.pst
‘He put down the (hunted)pig, “Where did Kaut go/get lost?”

42. o tuŋ-bo ca:y ik=akauno-ø

3sg refl-heart bad obj=3sg.sit down-pst
‘With a heavy heart, he sat down.’

43. a=mimi a=kɔʈ ɲyo-kotrata o

cl 1=mother cl 1=Kaut house-inside 3sg
a=phertajiɖo.t=kholet=lameme
cl 1=Phertajido.cl 4=laugh.obj=tired
‘Mother Kaut tired herself while laughing at Phertajido from inside the house.’

44. phertajido erac-il a=tum-bo(u)t khacul eule-ø

Phertajido surprise-conv cl 1=refl-back turn see-pst
‘Surprised, Phertajido turned back to see [where the laughter was coming
from].’

45. a=kɔʈ-e eule-ø

cl 1=Kaut-abs see-pst
‘He saw Kaut.’

46. phertajiɖo e=maʈ-il unci-ø

Phertjido cl 5=running-conv go-pst
‘Phertajido went to her running.’
284 text

47. a=kɔʈ-ek=tercoic-il kɔʈ ek=ŋolo-me

cl 1=Kaut-obj=embrace-conv Kaut obj=cry-hab
‘Embracing Kaut, he burst out crying.’

48. dantɔ nencuo ʈhi-t=ɲyo-me

Thus own place-obj=live-hab
‘Then they lived in their own place.’

49. nutun-tʰire em-il nutun-tʰire cɔpʰe

3pl.poss-children refl-nrr.pst 3pl.poss-children many
‘Their children had their children and they had many children.’

50. u ne boi-(e)m-il
3 pl marry-rec-nrr.pst
‘They married each other.’

51. u thire-t-thire cɔphe-(e)m-il

3sg children-cl 4.poss=children many-refl-nrr.pst
‘Children gave birth to many more children. Children’s children were many.’

52. phertajiɖo em-boi a=kɔʈ-e ikjira-l

Phertajido refl-wife cl 1=Kaut-abs told-nrr.pst
‘Phertajido asked his wife Kaut.’

53. pharoko7-bi kɛʈ-e

rope-abs knit/make-imp
‘Make a rope.’

54. um- pharako inci-l pharako-kuni-ø

3sg.for her- pharako bring-conv pharako-3sg.come back-pst
‘He brought pharako twigs for her, came back with pharako.’ or
‘He came back with Pharako twigs for her.’

55. em-boi a=kɔʈ ek=jira-l pharako-bi bɔl-e

refl-wife cl 1=kaut obj=say-nrr.pst rope-abs make-imp
‘He said to his wife Kaut, “Make a rope out of it.”’

56. era=lobum o i=ta=uŋ-kɔc-il et=calo-ø

cl 2=long 3sg cl 5=tr=ref-make-conv obj=heap-pst
‘She made such a long (rope) that it coiled in a heap.’

57. pharako tu(t)=juphul phertajiɖo meo-bi kocob-il

rope cl 4.poss=end Phertajido stone-abs tie-nrr.pst
‘Phertajido tied a stone at the end of the rope.’

7 Pharako is a kind of rope which is made out of a creeper by the same name, i.e.,
‘pharako.’ The rope is especially known for its strength.
 the great narrative of phertajido 285

58. o er=kɛʈo-l o ek=terʈɔe-ø ʈau

3sg cl 2=swirl-conv 3sg obj=threw-pst sky/clouds
tu(t)=mikhu-(a)k
cl 4=middle-dir
‘He swirled it and then hurled it (the stone) at (in the middle of) the sky.’

59. cyal eren-ʃolok-il o ek=ʈɛno ataci

where refl-stuck/caught-nrr.pst 3sg obj=pulled check
‘He pulled it (the rope) down to check where it (the rope) was stuck.’

60. u ɛr=ʈeʈer-el
3sg cl 2=got caught/entangled-nrr.pst
‘It (the rope) was stuck.’

61. kɛnmo er=kɛʈ-il

thread cl 2=swirl/twist-nrr.pst
‘(He) twisted the threads of the rope.’

62. kɛnmo coʈɔ-me

thread curvy became
‘It (the thread) became taut and curled.’

63. (e)k.ɔ=cop-ø
3 obj.cl 7=tie-pst
‘He tied it (the rope).’

64. em-boi kɔʈ-e ekjira-(i)nci-l

refl-wife Kaut-abs tell-go-nrr.pst
‘He went to tell his wife Kaut.’

65. ʈhu ʃoŋak mɔtkɔco-kak ʈh=ut=cone-b-om

1sg human heaven/above-dir 1sg=cl 4=go-fa-npst
‘I will go up above, there in the heaven.’

66. ʈhu ʈhi-bi ole

1sg place-abs will see
‘I will see the place.’

67. ʃitane ʈhi bi

How place cop
‘How is that place?’

68. mɔtkɔco-kak ʈhambikhir ʈhut= cone-k-om

heaven/above-dir tomorrow 1sg.cl 4=go-fa-npst
‘Up above us I will go (to the heaven) tomorrow (next morning).’
286 text

69. akambikhir ut=cone-ø

3 sg.next day cl 4=go-pst
‘Next day he went away.’

70. o rɛphu-l mɔtkɔco-kak ʈhi ole mil-ø

3sg climb-conv heaven/above-dir place see return-pst
‘Having climbed up, he saw/discovered the place and returned.’

71. o tarphuc-ine cɔphe (e)t= cɔŋ-ø

3sg Andamanese like his own many obj=find-pst
‘He found many people like himself.’

72. aka mele ʈhi –(e)loe-l u uni- ø

3sg returned place see-conv 3sg come back-pst
‘Having seen the place he came back (returned).’

73. em-boi a=kɔʈ-e jiralol-kɔc mu-taraphuc khe-ɲyo

refl-wife cl 1=Kauto-abs said “up above” 1incl-like us there-live
‘He told his wife Kaut “people like us live up above us.”’

74. ʈhi nol-e me-be

Place good-abs allover-cop
‘That place is very good.’

75. em-boi kɔʈ-e (e)kjira-l khaʈɛ ʈhɛŋ.ot=cone

refl-wife Kaut-abs-say-nrr.pst let us 1sg incl.cl 4=go
‘He told his wife Kaut, “Let us go there.”’

76. a=kɔʈ-e aka-tɛkhu-k pʰ-il

cl 1=Kaut-abs his-proposal-obj neg-nrr.pst
‘Kaut did not like his proposal.’

77. ʃitane ʈhɛŋ-otun thire-ne ʈhi-bi.t= mɔ

How 1sg.incl-refl child-pl place-abs.obj=leave
“How shall we leave the place of our children?”

78. bo a=pʰertajiɖo ikjira-ø

then cl 1=Phertajido say-pst
‘Then Phertajido said.’

79. ʈʰɛŋ-otun-tʰire-ni arbitta kamo ʈhɛŋ ot=cɔne

1incl-refl-child-pl make understand cond 1incl cl 4=go
“If we convince our children we can go.”

80. nutun-thire.n araɲ-il u ni arɛtta-ø

refl-child.pl gather-conv 3sg 3pl.dat lecture-pst
‘Having gathered his children he addressed them.’
 the great narrative of phertajido 287

81. phɛrtajiɖo ikjira-m ʈhu(t) thire kho-ittakhe

Phertjido say-npst my children now-have patience
neli ʈhi-bi-ka lile-kɛ
2pl.h.dat place-abs-3sg peace-imp
‘Phertajido says “My children, have patience! You all keep silent. Let there be
peace on this land”’

82. ŋale-mimi ŋale-may ʈhiyo mu-ŋili arbitta-k m

2pl.h-mother 2pl.h-father exist 1du-2pl.dat speak-npst
‘We, are your mother (and) your father, both of us want to speak to you/
make you understand.’

83. khilele mu-ŋili ʈhi.bi.t=ɲyo pho

over here 1du-2pl.dat place.abs.obj=live neg
‘We will not stay over here with you all.’

84. mɔtkɔcua m.eɲo be

above 1pl.go up cop
‘We will go up above us (heaven).’

85. khilele ŋole ʈhi-bi.t=ɲyo-t=nole

Over here 2pl.agt place-abs.obj=live cl 4=happily
‘You all live here happily.’

86. ɖikhɔ mun-tara=liubo-ø.

Enough 1pl-cl6=finish-pst
‘That’s enough, we have completed our time here/our time is finished.’

87. ittɔ-tekhamo m=ut=cone-b-om

Right now 1pl=cl 4=go-fa-npst
‘Right now we are going / leaving.’

88. nu rɛphu-l pharako ikku=beliŋ-ø

3pl climb-conv rope obj=cut off-pst
‘Having climbed up (the rope) they cut off the rope.’

English translation of ‘The Great Narrative of Phertajido’

An Attempt Has Been Made to Keep the Translation as Close to the

Original as Possible
In ancient times there lived a man named Phertajido.8 Phertajido was the
first man of the Andaman Islands. He originated from the hollow of a

8 The literal meaning of pher-ta-jido in Great Andamanese is ‘born out of a bamboo.’

288 text

bamboo. He roamed here and there, searched for food and lived alone. He
spent his time making bows and arrows.
One day, he shot the arrows here and there in all directions. Next day,
he went to search for the shot arrows. When he searched for the arrows,
he found a spring and drank the water from it and thus discovered drink-
ing water.
He went to look for more of the arrows and found one hidden in the
roots of a potato plant. He thus found the potato and took some with
him.
He looked for more of the shot arrows. This time he found the third
arrow in heap of incense (dhoop). He took a bit of the dhoop with him.
He went to look for more of the shot arrows. This time he found a very
fine soil of Kaut. He took some of this also with him. He made pots out of
the soil. He dried them to harden. When the pots dried and became hard,
he put the potatoes in the pot and boiled them. He ate those potatoes and
thus enjoyed the meal of boiled potatoes.
While eating potatoes, an idea came to him to carve a sculpture from
the remaining Kaut.
He wasted no time and in a few days made a human look-alike dummy
out of the Kaut soil.
He put this dummy on a raised platform and burnt some fire under it
so as to dry it well. Thereafter, he resumed his bow-making.
He would look at the dummy on the platform again and again while he
was busy making his bows and arrows. He ensured that the Kaut was lying
on the platform and did not fall down. He would occasionally get up and
go to the platform to put some more wood into the fire and then would
come back to his bow-shaving job. After a while he looked back again.
He was surprised. The platform shook as the female figure of Kaut
turned over. Phertajido was overwhelmed. He was immensely satisfied
with his work. He stood up again to kindle the fire and complete the job
of drying the figure.
When he got tired of making bows, he went into the jungle to hunt,
leaving the Kaut on the platform for it to be dried completely. He found
game and proceeded home with it.
As he approached home, he glanced from a distance at the platform. He
found nothing on the platform. The platform was empty. He was shocked.
His felt dejected and lost.
He put down the game and sighed, “Where did Kaut disappear?” He
was feeling sad. He sat down on the ground with a heavy heart. He did
 the great narrative of phertajido 289

not know what to do next. He was oblivious of the fact that the lady Kaut
was inside the house.
Kaut saw Pertajido from inside the house and started laughing. She
laughed, and laughed until she got tired of it. Surprised, Phertajido looked
back. He saw Kaut sitting inside the house laughing merrily.
Phertajido ran to her. He embraced Kaut and burst in tears out of
the sheer joy of discovering Kaut. After that, both of them started living
together as husband and wife. They had many children. Their children
married among themselves and thus their clan increased by leaps and
bounds.
Phertajido once asked his wife Kaut to make ropes.
He went to bring a creeper (pharako), found in the jungle which is good
for making rope, and came back with it.
He asked his wife to peel the creeper and make rope of it.
His wife followed his advice and made a very long rope. It was so long
that it coiled in the shape of a heap.
Phertajido tied a stone at the head of the rope. He swirled and swirled
the rope several times and finally threw it up in the sky. He pulled back
the rope and found that the rope was entangled somewhere and would
not come down. He twisted the rope to make it harder. The rope tight-
ened and stiffened. He tugged at it, but the rope would not move. He
knew that the rope was stuck somewhere.
He went to call Kaut. He said to her, “I will go up above the clouds. I
will go and see the place above us. I will find out what the place looks like.
I will go there tomorrow.”
Next day he climbed up above the clouds.
He saw the place and was surprised to find many people like himself.
He came back to the earth and told his wife about this. He told her that
the place above them was nice and there were many people like the peo-
ple of Andaman. He suggested to her that both of them should go there.
Kaut did not like his suggestion. She said, “How can we leave our chil-
dren’s place?”
Phertajido said, “We will convince our children and then we will go.”
He gathered his children at one place. He tried to convince them. Pher-
tajido said, “My dear children, please keep silent for a while. Your father
and your mother are speaking to you. We will no longer stay here on this
earth. We will go up above the clouds. You should live your life well here.
Our time here is completed. Now we will go.”
Thus saying, they went up above the clouds through the rope. Once
they had reached the top, they cut the rope off from above.
 Bibliography

Abbi, Anvita. 1991. Reduplication in South Asian Languages. An Areal, Typological and His-
torical Study. New Delhi. Allied Publishers.
——. 2001. A Manual of Linguistic Field Work and Structures of Indian Languages. Munich:
Lincom Europa.
——. 2003. Vanishing Voices of the Languages of the Andaman Islands. Paper presented at
the Max Planck Institute of Evolutionary Anthropology, Leipzig, June 13.
——. 2005. “Is Great Andamanese Typologically Divergent from the Standard Aver-
age Andamanese?” Paper presented at ALT VI, Max Planck Institute, Padang,
Indonesia.
——. 2006a. Endangered Languages of the Andaman Islands. München: Lincom GmbH.
——. 2006b. “Vanishing Voices: A Typological Sketch of Great Andamanese.” In: Anju
Saxena & Lars Borin (eds.). Lesser-Known Languages of South Asia. Status and Policies,
Case Studies and Applications of Information Technology. Berlin / New York: Mouton de
Gruyter. 107–123.
——. 2008. Folk Songs of the Great Andamans CD-Rom. New Delhi: Jawaharlal Nehru Uni-
versity.
——. 2009. “Is Great Andamanese Genealogically and Typologically Distinct from Onge
and Jarawa?” Language Sciences 31: 791–812.
——. 2011. “Body Divisions in Great Andamanese: Possessive Classification, the Semantics
of Inherency and Grammaticalization.” Studies in Language 35/4: 735–788.
——. 2012. A Dictionary of the Great Andamanese Language. English-Great
Andamanese-Hindi (with CD Rom) Delhi: Ratna Sagar P. Ltd.
——. 2012. An Ancient Tale from Andaman. Narrated by Nao Jr. Delhi: National Book Trust,
India.
Abbi, Anvita, Bidisha Som & Alok Das. 2007. “Where Have All the Speakers Gone? A Socio-
linguistic Study of the Great Andamanese.” Indian Linguistics 68/3–4: 325–342.
Abbi, Anvita, Nao Jr. & Abhishek Avatans. 2008. Varnamala the Book of Letters of Great
Andamanese. New Delhi: Jawaharlal Nehru University.
Abbi, Anvita & Pramod Kumar. 2010. “In Search of Language Contact Between Jarawa and
Aka-Bea: The Languages of South Andaman.” Paper read at the First Conference on ASJP
and Language Prehistory (ALP-I), Max Planck Institute of Evolutionary Anthropology,
Leipzig, Germany on 18 September 2010.
Aikhenvald, Alexandra Y. 2000. Classifiers: A Typology of Noun Categorization Devices.
Oxford: Oxford University Press.
——. 2002. “Typological Parameters for the Study of Clitics, with Special Reference to
Tariana.” In: R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.). Word: A Cross-Linguistic
Typology. Cambridge: Cambridge University Press. 42–78.
——. 2007. “Typological Distinctions in Word-Formation.” In: Timothy Shopen (ed.), Vol.
3. 1–65.
——. 2008. “Versatile Cases.” Journal of Linguistics 44: 565–603.
Allan, Edward J. 1975–1976. “Inalienable Possession in Four Ethiopian Languages.” Afrika
und Übersee 59: 300–307.
Ameka, Felix. 1996. “Body Parts in Ewe Grammar.” In: Hilary Chappell & William McGregor
(eds.). 783–840.
Anderson, R. 2005. Aspects of the Theory of Clitics. Oxford: Oxford University Press (Oxford
Studies in Theoretical Linguistics).
Avtans, Abhishek. 2006. Diectic Categories in Great Andamanese. M.Phil. dissertation, Jawa-
harlal Nehru University, New Delhi.
292 bibliography

——. 2007. “Humane Face of Language Documentation: A Great Andamanese experience.”

In: Peter K. Austin, Oliver Bond & David Nathan (eds.). Proceedings of the Conference
on Language Documentation & Linguistic Theory. London: Department of Linguistics,
SOAS. 21–28.
Avtans, Abhishek & Anvita Abbi. 2006. “Language Documentation in Andamans: Highs
and Lows.” In: R. Elangaiyan, R. Mckenna Brown, Nicholas D. M. Ostler & Mahendra K.
Verma (eds.). The Proceedings of the Tenth Foundation of Endangered Languages Confer-
ence, Vital Voices: Endangered Languages and Multilingualism. Mysore: CIIL. 73–76.
Bally, Charles. 1996 (1926). “The Expression of Concepts of the Personal Domain and Indi-
visibility in Indo-European Languages.” In: Hilary Chappell & William McGregor (eds.).
31–61.
Barker, C. 1997. Possessive Descriptions. Stanford: CSLI Publications.
Baron, Irène, Michael Herslund & Finn Sørensen (eds.). 2001. Dimensions of Possession.
Amsterdam / Philadelphia: Benjamins (Typological studies in language, 47).
Basu, D. N. 1952. “A Linguistic Introduction to Andamanese.” Bulletin of the Department of
Anthropology, 1–2: 55–70.
——. 1955. “A General Note on the Andamanese Languages.” Indian Linguistics 16: 213–
225.
Berlin, Brent. 1976. “The Concept of Rank in Ethnological Classification. Some Evidence
from Aguaruna Folk Botany.” American Ethnology 3: 381–99.
——. 1992. Ethnobiological Classification. Principles of Categorization of Plants and Animals
in Traditional Societies. New Jersey: Princeton University Press.
Bladon, R. A. W. 1979. “The Production of Laterals: Some Articulatory Properties and their
Acoustic Implications.” In: H. Hollien & P. Hollien (eds.). Current Issues in the Phonetic
Sciences. Amsterdam: Benjamins. 501–508.
Blevins, Juliette. 2007. “A Long Lost Sister of Proto-Austronesian? Proto-Ongan, Mother of
Jarawa and Onge of the Andaman Islands.” Oceanic Linguistics 46/1: 155–198.
de Bot, Kees. 2004. “Language Attrition.” International Journal of Bilingualism 8:3. Kingston
Press.
Bulmer, R. N. H. 1974. “Memoirs of a Small Game Hunter: On the Track of Unknown Ani-
mal Categories in New Guinea.” Journal D’agriculture Tropicale et de Botanique Appli-
quée 21: 79–99.
Campbell, Lyle. 1997. “The Amerind Personal Pronouns: A Second Opinion.” Language
73/2: 339–351.
Chafe, Wallace. 1970. Meaning and the Structure of Language. Chicago: Chicago University
Press.
Chappell, Hilary & Sandra A. Thompson. 1992. “Semantics and Pragmatics of Associative de
in Mandarin Discourse.” Cahiers de Linguistique Asie Orientale 21/2: 199–229.
Chappell, Hilary & William McGregor. 1989. “Alienability, Inalienability and Nominal Clas-
sification.” Proceedings of the Fifteenth Annual Meeting the Berkeley Linguistic Society 15:
24–36.
——. 1996. “Prolegomena to a Theory of Inalienability.” In: Hilary Chappell & William
McGregor (eds.). 3–30.
—— (eds.). 1996. The Grammar of Inalienability. A Typological Perspective on Body Part
Terms and the Part-Whole Relation. Berlin / New York: Mouton de Gruyter.
Chaudhary, Narayan. 2007. Developing a Computational Framework for the Verb-Morphology
of Great Andamanese. M.Phil. Dissertation. Jawaharlal Nehru University, New Delhi.
Chaudhary, Narayan, Anvita Abbi & Girish Nath Jha. 2008. “Morphological Analyzer for
Great Andamanese Verbs: Implementing a Concatenate Template.” Vishwabharat April
2007–January 2008: 113–118.
Cipriani, Lidio. 1966. The Andaman Islanders. Ed. and translated by D. Tayler Cox. London:
Weidenfeld and Nicholson.
Comrie, Bernard. 1976. “The Syntax of Action Nominals: A Cross-Linguistic Study.” Lingua
40: 177–201.
 bibliography 293

Cooper, Zarine. 2002. Archeology and History. Early Settlements in the Andaman Islands.
New Delhi: Oxford University Press.
Crowley, Terry. 1995. “Inalienable Possession in Paamese Grammar.” In: Hilary Chappell &
William McGregor (eds.). 383–432.
Dahl, A. 1985. “The Non-Past Tense.” In: Timothy Shopen (ed.), Vol. 2. 23–26.
Das, Alok. 2006. “Directionality of the Pre-Neolithic Linguistic Spread: Gathering Evidence
Through the Andamanese Languages.” Journal of Philology 1(22): 128–134.
Dasgupta, D. & S. R. Sharma. 1982. A Handbook of Onge Language. Calcutta: Anthropologi-
cal Survey of India.
Diem, Werner. 1986. “Alienable und Inalienable Possession im Semitischen.” Zeitschrift der
deutschen Morgenländischen Gesellschaft 136: 227–291.
Dixon, Robert M. W. 2004. The Jarawara Language of Southern Amazonia. Oxford: Oxford
University Press.
——. 2010. Basic Linguistic Theory. Vol. 2. Oxford: Oxford University Press.
Dryer, Matthew S. 2007. “Clause Types.” In: Timothy Shopen (ed.), Vol. 1. 224–275.
Dutta, P. C. 2004. The Great Andamanese. Past and Present. Kolkata: Anthropological Sur-
vey of India.
Endicott, Phillip, M. Thomas P. Gilbert, Chris Stringer, Carles Lalueza-Fox, Eske Willerslev,
Anders J. Hansen & Alan Cooper. 2003. The Genetic Origins of the Andaman Islanders.
American Journal of Human Genetics 72, Report no. 178.
Evans, N. 1996. “The Syntax and Semantics of Body Part Incorporation in Mayali.” In: Hilary
Chappell & William McGregor (eds.). 65–109.
Fabri, Ray. 1996. “The Construct State and the Pseudo-Construct State in Maltese.” Rivista
di linguistica 8: 229–244.
Fant, Gunnar. 1960. Acoustic Theory of Speech Production. The Hague: Mouton.
Fleck, David. W. 2006. “Body Part Prefixes in Matses: Derivation or Noun Incorporation?”
International Journal of American Linguistics 72/1: 59–96.
Forth, Gregory. 2004. Nage Birds. London: Routledge.
Frederike, Caesar. 1625. Purchas: His Pilgrimes. Eighteen Years of Indian Observations.
Vol. II: 1710.
Ganguli, P. 1966. “Vocabulary of the Negritos of Little Andaman with Grammatical Notes
and Materials.” Bulletin of the Department of Anthropology XV: 1–30.
Gerlach, Birgit & Janet Grijzenhout. 2000. Clitics in Phonology, Morphology and Syntax.
Amsterdam / Philadelphia: Benjamins (Linguistik aktuell / Linguistics today, 36).
Givon, T. 2002. Bio-Linguistics: The Santa Barbara Lectures. Amsterdam / Philadelphia:
Benjamins.
Haspelmath, Martin. 2007. “Coordination.” In: Tomothy Shopen (ed.), Vol. 2. 1–51.
Haspelmath, Martin, Matthew S. Dryer, David Gil & Bernard Comrie (eds.). 2005. The
World Atlas of Language Structures. Oxford: Oxford University Press.
Hawkins, Roger. 1981. “Towards an Account of the Possessive Constructions: NP’s N and the
N of Np.” Journal of Linguistics 17: 247–269.
Heggie, Lorei & Francisco Ordóñez (Eds.). 2005. Clitics and Affix Combination: Theoreti-
cal Perspectives. Amsterdam / Philadelphia: Benjamins (Linguistik Aktuell / Linguistics
Today, 74).
Heine, Bernd. 1997 [2006 Reprint]. Possession. Cognitive Sources, Forces and Grammatical-
ization. Cambridge: Cambridge University Press (Cambridge Studies in Linguistics, 83).
Heine, Bernd & Tania Kuteva. 2007. The Genesis of Grammar: A Reconstruction. Oxford:
Oxford University Press.
Hinnebusch, Thomas J. & Robert S. Kirsner. 1981. “On the Reference of ‘Inalienable Posses-
sion’ in Swahili.” Journal of African Language and Linguistics 2: 1–16.
Hyman, Larry M., Danny Keith Alford & Elizabeth Akpati. 1970. “Inalienable Possession in
Igbo.” Journal of West African Languages 7: 85–101.
Justin, Anstice. 2000. “Who are the Jarawa?” Reproduced by www.andaman.org.
294 bibliography

Kashyap, V. K., T. Sitalaximi, B. N. Sarkar & T. Trivedi. 2003. “Molecular Relatedness of the
Aboriginal Groups of Andaman and Nicobar Islands with Similar Ethnic Populations.”
International Journal of Human Genetics 3/1: 5–11.
Koptjevskaja-Tamm, Maria. 1993. Nominalizations. London: Routledge.
Kumar, Pramod. 2005. Field Notes on Jarawa. Unpublished. Jawaharlal Nehru University,
New Delhi.
——. 2012. Grammar of Jarawa. Ph.D. Dissertation, Max Planck Institute of Evolutionary
Anthropology, Leipzig and Jawaharlal Nehru University, New Delhi.
Ladefoged, Peter & Ian Maddieson 1996. The Sounds of the World’s Languages. UK / USA /
Australia: Blackwell Publishing.
Langacker, Ronald W. 1991. Foundations of Cognitive Grammar. Descriptive Application. Vol
II. Stanford University Press.
——. 1995. “Possession and Possessive Constructions.” In: John R. Taylor & Robert E.
Maclaury (Eds.). Language and the Cognitive Construal of the World. Berlin / New York:
Mouton De Gruyter (Trends in Linguistics. Studies and Monographs, 82). 51–79.
Levy, Paulette. 1992. “Body Part Prefixes in Papantla Totonac.” Zeitschrift Für Phonetik,
Sprachwissenschaft Und Kommunikationsforschung 45/6: 530–542.
——. 1999. “From ‘Part’ to ‘Shape’: Incorporation in Totonac and the Issue of Classification
by Verbs.” International Journal of American Linguistics 65: 127–175.
Lévy-Bruhl, Lucien. 1914. “L’expression de la Possession dans les Langues Mélanésiennes.”
Memoires de la Société de Linguistique de Paris 19: 96–104.
Lindblom, B. & Ian Maddieson. 1988. “Phonetic Universals in Consonant Systems.” In:
H. Hyman & C. Li (Eds.). Language, Speech and Mind. London: Routledge. 62–79.
Maddieson, Ian. 2005. “Consonant Inventories.” In: Martin Haspelmath et al. (Eds.). 10–13.
Majid, Asifa. 2010. “Words for Parts of the Body.” In: Barbara C. Malt & Phillip Wolff (Eds.).
Words and the Mind. New York / Oxford: Oxford University Press.
Man, E. Horace. 1875–1879. Andaman Vocabulary. Part I (English-Andamanese); Part II
(Andamanese-English). Handwritten Manuscript. London: Royal Anthropological Institute.
——. 1883. “On the Aboriginal Inhabitants of the Andaman Islands.” Journal of the Anthro-
pological Institute of Great Britain and Ireland 12: 69–116 (Part I); 117–175 (Part II); 327–
434 (Part III).
——. 1885. On the Aboriginal Inhabitants of the Andaman Islands with Reports of Researches
Into the Language of the South Andaman Islands by A. J. Ellis. London: Anthropological
Institute of Great Britain and Ireland.
——. 1923. A Dictionary of the South Andaman (Âkà-Bêa) Language. Mazgaon, Bombay:
British India Press.
Man, E. Horace & R. C. Temple. 1875-1878. A Grammar of the Bojingyida or South Anda-
man Language. Handwritten Manuscript, Loose Sheets. London: Royal Anthropological
Institute.
Manoharan, S. 1980. “Language of the Present Great Andamanese.” Journal of the Indian
Anthropological Society 15: 43–55.
——. 1983. “Subgrouping Andamanese Group of Languages.” International Journal of Dra-
vidian Linguistics 12/1: 82–95.
——. 1989. A Descriptive and Comparative Study of the Andamanese Language. Calcutta:
Anthropological Survey of India.
Masefield, John (Ed.). 1908. The Travels of Marco Polo the Venetian. London: J. M. Dent
(Everyman’s Library).
Mayank. 2009. ComparatIve Lexicon of Great Andamanese Languages. M.Phil. Dissertation.
Jawaharlal Nehru University, New Delhi.
Mckay, Graham R. 1996. “Body Parts, Possession Marking and Nominal Classes in Ndjéb-
bana.” In: Hilary Chappell & William Mcgregor (Eds.). 293–326.
Muysken, Pieter. 2008. Functional Categories. Cambridge: Cambridge University Press.
Nichols, Johanna. 1988. “On Alienable and Inalienable Possession.” William Shipley (Ed.).
In Honor of Mary Haas: From the Haas Festival Conference in Native American Linguistic
S. Berlin: Mouton De Gruyter. 557–609.
 bibliography 295

——. 1992. Linguistic Diversity in Space and Time. Chicago: University of Chicago Press.
Nichols, Johanna & David A. Peterson. 2008. “N-M Pronouns.” In: Martin Haspelmath, Mat-
thew S. Dryer, David Gil & Bernard Comrie (Eds.). The World Atlas of Language Struc-
tures Online. Munich: Max Planck Digital Library. Chapter 136 and 137. Available Online
At Http://Wals.Info/Feature/137.
Pande, Satish & Anvita Abbi. 2011. Ethno-Ornithology. Birds of Great Andamanese. Names,
Classification and Culture. Pune: Ela Foundation, Bombay Natural History Society and
Oxford University Press.
Portman, M. V. 1887 [1992 Reprint]. A Manual of the Andamanese Languages. Delhi: Manas
Publications.
——. 1898. Notes On the Languages of the South Andaman Group of Tribes. Calcutta: office
of the Superintendent of Government Printing, India.
——. 1899 [1990 Reprint]. A History of Our Relations With the Andamanese. Vols. I–II. New
Delhi: Asian Educational Services.
Radcliffe-Brown, A. R. 1922 [3rd Reprint 1948]. The Andaman Islanders. Illinois: Free Press,
Glencoe.
Ross, Malcom. 2004. “Demonstratives, Local Nouns and Directionals in Oceanic Languages:
A Diachronic Perspective.” In: Gunter Senft (Ed). 15–36.
Sarkar, B. N. & R. Sahani. 2002. “A Demographic and Health Profile of the Jarawas of
Andaman Islands.” In: K. Mukhopadhyay, R. K. Bhattacharya & B. N. Sarkar (Eds.).
Jarawa Contact: Ours with them/theirs with us. Calcutta: Anthropological Survey of India.
132–147.
Schmid, Monika S. 2011. Language Attrition. Key Topics in Sociolinguistics. Cambridge: Cam-
bridge University Press.
Seiler, Hansjakob. 1983. Possession as an Operational Dimension of Language. Tübingen:
Gunter Narr Verlag (Language Universals Series, 2).
——. 2001. “The Operational Basis of Possession. A Dimensional Approach Revisited.” In:
Irène Baron, Michael Herslund & Finn Sørensen (Eds.). 27–40.
Seliger, Herbert W. and Robert Michael Vago Eds. 1991. First Language Attrition. Cam-
bridge: Cambridge University Press.
Senft, Gunter. 2004. “Aspects of Spatial Deixis in Kilivila.” In: Gunter Senft (Ed.). 59–80.
—— (Ed.). 2004. Deixis and Demonstratives in Oceanic Languages. Canberra: Pacific Lin-
guistics, Research School of Pacific and Asian Studies, the Australian National University.
Shibatani, Masayoshi & Sandra Thompson (Eds.). 1996. Grammatical Constructions. Their
Form and Meaning. Oxford: Oxford University Press.
Shopen, Timothy (Ed.). 2007. Language Typology and Syntactic Description. Vols. 1–3, 2nd
Ed. Cambridge: Cambridge University Press.
Siegel, J. 1985. “Koinés and Koinéization.” Language and Society 14: 357–378.
Som, Bidisha. 2006. A Lexico-Semantic Study of Great Andamanese: A Thematic Approach.
Ph.D. Dissertation, Jawaharlal Nehru University.
Spanoghe, Annie-Marie. 2001. “(In)alienability and (In)determination in Portuguese.” In:
Irène Baron, Michael Herslund & Finn Sørensen (Eds.). 227–242.
Spencer, Andrew. 2000. “Verbal Clitics in Bulgarian.” In: Birgit Gerlach & Janet Grijzenhout
(Eds.). 355–386.
——. 2011. Towards a Typology of ‘Mixed Categories’. Preprint Copy.
Temple, R. C. 1903 (1994 Reprint, New Delhi). “A Grammar of the Andamanese Language.”
Chapter IV of Part I, Volume 13 of The Census Report on the Andaman and Nicobar
Islands 1901. Port Blair: Superintendent’s Printing Press.
Thangaraj, Kumarasamy, Gyaneshwer Chaubey, Vijay Kumar Singh, Ayyasamy Vanniarajan,
Ismail Thanseem, Alla G. Reddy and Lalji Singh. 2006. “In Situ Origin of Deep Rooting
Lineages of Mitochondrial Macrohaplogroup ‘M’ In India.” Bmc Genomics 7: 151.
Thangaraj, Kumarasamy, Lalji Singh, Alla G. Reddy, V. Raghavendra Rao, Subhash C.
Sehgal, Peter A. Underhill, Melanie Pierson, Ian G. Frame & Erika Hagelberg. 2003.
“Genetic Affinities of the Andaman Islanders. A Vanishing Human Population.” Current
Biology 13: 86–93.
296 bibliography

——. Underhill, Melanie Pierson, Ian G. Frame & Erika Hagelberg. 2005. “Reconstructing
the Origin of Andaman Islanders.” Science 308: 996.
Tsunoda, Tasaku. 1996. “The Possession Cline in Japanese and Other Languages.” In: Hilary
Chappell & William Mcgregor (Eds.). 565–630.
Velázquez-Castillo, Maura. 1996. The Grammar of Possession. Inalienability, Incorporation
and Possessor Ascension in Guarani. Amsterdam / Philadelphia: Benjamins (Studies in
Language Companion Series, 33).
Voeltz, Erhard F. K. 1976. “Inalienable Possession in Sotho.” In: Larry M. Hyman, Leon C.
Jacobson & Russell G. Schuh (Eds.). Papers in African Linguistics in Honor of Wm. E.
Welmers. Los Angeles: University of California (Studies in African Linguistics. Supple-
ment 6). 255–266.
Waelchli, Bernard. 2005. Co-Compounds and Natural Coordination. Oxford: Oxford Univer-
sity Press.
Wierzbicka, Anna. 1976. “Mind and Body from a Semantic Point of View.” In: James Mcca-
wley (Ed.). Notes from the Linguistic Underground. New York: Academic Press (Syntax
and Semantics, 7). 129–158.
Wierzbicka, Anna. 1985. Lexicography and Conceptual Analysis. Ann Arbor: Karoma.
Wohlgemuth, Jan & Michael Cysouw (Eds.). 2010. Rethinking Universals: How Rarities Affect
Linguistic Theory. Berlin: Mouton De Gruyter.
Wunderlich, Dieter & Ray Fabri. 1995. “Minimalist Morphology: An Approach to Inflec-
tion.” Zeitschrift für Sprachwissenschaft 14: 236–294.
Zide, Norman & Vishvajit Pandya. 1989. “A Bibliographical Introduction to Andamanese
Linguistics.” Journal of the American Oriental Society 109/ 4: 639–651.
Zwicky, Arnold M. & Geoffrey K. Pullum. 1983. “Cliticization vs. Inflection: English N’t.”
Language 59/3: 502–513.

Websites with various kinds of information and database available on languages of the
Andaman Islands:

• www.andaman.org
• www.andamanese.net
 Index

Abbi, A. 14–15, 18, 27, 34, 106, 231 n. 4 Berlin, Brent 35

Absolutive 61, 67–69, 103, 116–118, Best speaker 18–27
120–122, 135, 174, 220, 222, 262 Bilingual mixture 28
Acoustic 52, 62–63, 275–278 Blevins, J. 15
Adjacency 224–231 Bo xvii–xviii, 8, 10, 14, 21, 23, 25–27, 39,
Adjunct 79, 128, 226 52, 110, 116, 119, 175, 231, 237, 273
Aetas 13 Boa Sr. the last speaker of Bo xv, xviii, 1,
Agent nominal 67, 116 8, 39, 47, 52, 175, 231
Agentive 98, 117, 119, 173–174 Body part categorization 65, 137, 214
Agglutinative 16, 65–66, 135 Brazil 82
Aikhenvald, A. 91, 135 n. 3 Bulmer, R. N. H. 35
Aka-Bale 12
Âkà-Bêa 12 Cardinal number 190, 192
Aka-Cari (Sare) (Chari) xvii, 9–12, 14, 21, Case alternation 70, 116
25–27, 39, 116, 163, 273 Case marker 92, 97, 103–104, 112, 118, 122,
Aka-Jowoi 10–11 136, 191, 200, 221, 245–246, 257
Aka-Kede 10, 12 Case relation 105, 136, 252
Aka-Kol 175 Categorical dual citizenship 70
Aka-Kora (Khora) xvii–xviii, 8, 10, 14, 21, Categorical-indeterminacy 70
25–27, 47–48, 52–53, 58, 62, 110, 116, 163, Causative 66, 101, 103, 122–123, 135, 213,
235, 273 224, 226–231, 250
Aka-Puchikwar (Pujjukar) 12, 46, 116 Chafe, W. 83
Alienable 66, 77, 80, 94, 99, 130–132, Chappell, H. 76
137–138, 149, 157–158, 163–164, 166–167, Chaudhary, N. 18, 279 n. 2
177, 182, 184 Clause chaining 261–264
Allan 76 Clitic sequencing 67, 91–93, 139, 153, 174,
Amazonian Peru 82 220, 225 n. 1, 251
Ambience 16, 83, 94 Code mixing 25–26, 28
Ambient verbs 184 Code switching 25–26, 28
Anaphoric reference 175 Comparative degree 131, 264–267
Andamani Hindi 19–21, 27 Competence level 19, 21–22, 24
Anderson, R. 90 Complement 167, 226, 260–262
Ang 14 n. 8, 15 Complement clause 261, 270
Angan 9, 14–16, 18 Complex consonants 47
Anthromorphic 75, 148 Complex syllable structure 56
Anthropocentric 206 Compositional compound 107
Anthropocentrism 211 Comrie, B. 262
Anticipatory action 237 Conceptual dependency 78–79, 81–82,
Antonym 104, 109 94, 158, 205
Applicative 98, 101, 103, 135, 213, 224, Conjunct 193, 254
226–230, 250 Conjunctive participle 241
Argument marker 67 Content question 249, 256
Assimilation 59 Contrastive degree 264
Austronesian 9, 15 Converb 233, 235, 259–260
Avatans, A. xv, 18, 25
Dasgupta 16
Bahuvrihi compound 104, 107–108 Degree of grammaticalisation 85–86,
Bangla 7, 25 92, 194
Basu 14, 46 Delexicalised 105
298 index

Dependency 66, 78–79, 81–83, 87, 94, 158, Homophonous 70, 117, 124, 131, 135, 226,
167, 190, 193–197, 205, 209, 222, 231 228, 233, 238 n. 7, 258–260, 262, 265
Descriptive adjectives 132, 190 Honey calendar 30–31, 203
Diem, W. 76 Human propensity 77, 190
Dimension 190 Hyman, L. M. 76
Disjunct 253
Dixon, R. M. W. 92 Imperative 238–239
Double marking 65–66, 130, 150–151, Inalienability 16, 66, 76–87, 94, 137–138,
153–155 156–158, 161, 175, 194, 202, 205, 215
Dryer, M. 240, 266 Inalienable 65–66, 77–78, 80–84, 86, 111,
Dual semantic system 65, 76, 90, 214 130–131, 137–139, 143, 147–151, 157–158,
160–167, 175, 178, 216, 243, 245
Enclitic 92, 102–103, 112–113, 121, 135, 251 Inchoative 98
Endangered language xvii, 35, 169 Inclusive 169–170
Endocentric compound 110 Incorporation 66, 82, 88, 123–124, 228
English 7, 12, 18, 25, 68 n. 1, 97, 166, 170, Independent form 113, 117, 241
178, 183, 210, 219, 258, 261, 264, 287 Inherency 66, 76–79, 81, 86, 158, 161, 164,
Epenthetic semi-vowel 41, 44 177
Ergative 67–68, 116, 119–120, 135 Intransitive predicate 190
Ethno-biological 34–36 Intransitive verbs 16, 67, 83, 89, 94, 116,
Ethno-taxonomy 35 134, 215–217, 225
Evans, N. 82
Evidential marker 177 Jangil 5, 11
Exclusive 169–170 Jarawa xvii, 3, 5–6, 11, 13–16, 18, 42, 169,
249
Fleck, D. W. 82 Jeru (Jero) xvii, 8–10, 13–14, 20–21, 23,
Focus 89, 170, 243, 250 25–27, 30, 47, 116, 175, 247, 255–256, 267,
Folk generics 35 273
Folk specifics 35 Justin, A. 14 n. 8
Formant structure 278
Formant transition 278 Kashyap, V. K. 36
Forth, Gregory 35 Key to competency levels 22
Frederike, Caesar 12 Kharia 3
Khora xvii–xviii, 8, 10, 14, 21, 25–27,
Generic 35, 101, 112–113, 116, 120, 151, 200, 47–48, 52–53, 58, 62, 110, 116, 163, 235,
236–237 273
Genitive 67, 99, 106, 130–131, 149, 152, Kirsner, R. S. 76
158–159, 163–164, 167, 174, 176, 182, 184, Koiné xvii, 3, 8–9, 14, 18, 25–26, 28, 37,
245 117, 147, 156
Gerlach, B. 92 Koptjevskaja-Tamm 70
Grammaticalised morpheme 65, 77, 82, Kumar, P. 18
90, 101 Kunibidijig 156
Grizenhout, J. 92 Kurux 3

Habitual 236–238 Labialised lateral 46, 52, 62, 275

Haspelmath, M. 251 Langacker, R. W. 78
Hawkins, R. 76 Language evolution 67, 92, 101, 271
Heggie, L. 88 Left-headed compound 105
Heritage language 7, 18–21, 24–28, 36, 49 Levy 82
Hindi xviii, 3, 7, 18–21, 22 n. 13, 24–28, Lexical class type 104
38, 41, 44, 46, 48, 68 n. 1, 98, 110, 112 n. 1, Lexicalisation 91, 101, 202
133–134, 149, 200 n. 1, 265, 282 n. 6 Lexicalised 101, 108, 121, 186, 197, 199, 207,
Hinnebusch, T. J. 76 217
 index 299

Lindblom, B. 47 Polar question 249, 256

Linguistic continuum 9 Polo, Marco 12
Location 81, 103, 111, 127–130, 180, 187, Polysynthetic 16, 65–66
205–209, 211, 217, 220, 240 Portman, M. V. 11, 12–13, 44
Locative case marker 103, 117 Possessive class 65, 137,
Possessive class marker 57, 137–148, 157,
Maddieson, I. 47, 56, 278 161, 164
Man, H. E. xviii, 12–13, 30, 149 n. 1, 159, Progressive assimilation 59
204 Prohibitive 98, 255–256
Manoharan, S. 33, 39, 47, 115 Prosodic 62, 89–90, 92
Masefield 1908 12 Prospective aspect 237
Matses 82 Proximity 166, 170, 172, 204, 210–211
Mayank xv, 10, 13 Psychological predicate 83, 215
McGregor, W. 76 Pullum, G. K. 90
Mixed base 26
Mixed language 14, 20 Radcliffe-Brown, A. R. 11, 12–13, 30
Monolingual 24 Referential 16, 70, 79, 172
Moribund 3, 16, 22, 24, 27–28, 65, 147, 230 Reflexive 62, 66, 92, 97–98, 104, 182–186,
Multifunctional 135 199, 213, 223–224, 227
Mundari 3 Reflexive prefix 62, 92, 97, 104, 227
Muysken, Pieter 76 n. 3 Reflexive suffix 182–185
Reflexive verb 183–184, 213
Ndjebbana 156 Register 22, 25–28
Nominaliser suffix 99, 262 Relators 116, 135
Non-compositional compound 105, Reported speech 263–264
107–108 Reporting speech 263–264
Non-Pama-Nyungan language 156 Resultative 98, 177–178, 217–218
Numbers 66, 112, 114, 169, 190, 192 Right-headed compound 105–107

Object nominal 67, 116 Scale of competence 22–23

Ongan 15 Scale of fluency 22
Onge xvii, 3, 5–6, 11–16, 42, 169, 249 Semangs 13
Ordinal 192 Semi-speaker 19, 22–23
Ordónez, F. 88 Sentinelese xvii, 3, 5–6, 9, 11, 14
Sharma, S. 16
Pande, Satish 18, 34–35, 82, 106 Shopen, T. 245, 251
Pandya, V. 16 Simple clitics 90
Panini grammar 104 Som, Bidisha 18
Panoan 82 Spatial word 103
Papantla Totonac 82 Special clitics 90, 102
Parameter of inalienability 78 Specificity 98, 120, 175
Paratactic strategies 267 Specifier 98, 134
Participial construction 133–134, 189, 191, Spencer, A. 70
267, 268 n. 3 Stative 70, 189, 191
Part-to-component 66, 137, 143, 159–160, Subordinate clause 233, 235, 251, 261,
167 264
Part-to-whole relationship 66, 79, 143, Superlative degree 264
159, 167 Syllable attraction 61–62, 93, 104
Perfectivity 236
Phonological host 225 TAM 66, 70, 73–74, 76, 92, 97, 224,
Phonological word 61–62, 66, 89, 93, 100, 231–240, 254, 260, 269
103–104, 109, 117, 174, 221, 229 Tamil 25
Plurality 66, 102–103, 112, 121, 135, 172 Tatpuruṣa compound 105–106
300 index

Temple, R. C. 11, 12–13 Veracruz 82

Temporal deictic marker 75, 114 Verb complex 213–242
Thangaraj, K. 15 Visibility 170, 172, 204, 210–211
Voeltz, Erhard 76
Unaccusative verbs 67, 116 Vowel sequence 37, 41–44
Unergative verbs 67, 116
Waelchli, B. 104
Valency 213, 217, 219, 222, 226, 250
Vanishing voices of the Great Andamanese Zero symbol 118
xvii–xviii, 18 n. 11 Zide 16
Velazquez-Castillo, M. 78 Zwicky, Arnold 90
Anvita Abbi collecting data from Nao Junior

Boa Senior
Boro, Anvita Abbi, Boa Senior and Peje at Strait

Great Andamanese community at Strait Island

Nao Junior and Lico