A Manual of Linguistic Field Work and Indian Language Structures 3895864013, 9783895864018

Citation preview

For Satish, My friend, partner and my husband

Contents Acknowledgements ............................................................................................. ix Preface.................................................................................................................. x Abbreviations ...................................................................................................... xii 1. Linguistic Fieldwork and the Indian Scene ................................................ 1 1.1 All About Linguistic Field Research ................................................ 1 1.1.1

Field Linguistics As An Input System to Other Fields .... 2

1.1.2

What Does It Involve? ...................................................... 2

1.1.3

Participatory in Character ................................................. 2

1.1.4

Theory Independent .......................................................... 3

1.2 The Composition of Language Scene In India ................................... 4 1.2.1

The Constitution ............................................................... 4

1.2.2

The Minority Languages................................................... 8

1.3 The Hierarchical Structure of the Indian Society .......................... 9 1.4 Bilingualism and the Indian Society ................................................ 10 1.4.1

The Rural Scene ................................................................ 11

1.4.2

The Urban Scene............................................................... 13

1.4.3

The Cosmopolitan Cities of India (The Mahanagar) ....... 15

1.4.4

Bilingualism and Education .............................................. 16

1.4.5

Language Loyalty, Language Shift And Language Adoption ........................................................................... 17

1.5 The Contact Language/s of India .................................................... 19 1.5.1

All India ............................................................................. 19

1.5.2

Our Field Experience ......................................................... 21

2. Language Families, Language Contact, and Areal Universals.................. 24 2.1 Indian Language Features ............................................................... 24 2.1.1

Indo-Aryan ........................................................................ 25

2.1.2

Dravidian .......................................................................... 29

2.1.3

Austro-Asiatic .................................................................. 33

2.1.4

Tibeto-Burman .................................................................. 39

2.1.5

Andamanese ...................................................................... 42

2.2 India as a Linguistic Area ................................................................ 43 2.2.1

The Sound System ............................................................ 44

2.2.2

The Morphological System............................................... 44

ii A Manual of Linguistic Fieldwork and Structures of Indian Languages 2.2.3

The Syntactic System ....................................................... 45

2.2.4

Pragmatics and Sociolinguistics ....................................... 45

2.3 The Sub-linguistic Area .................................................................... 45 2.4 Contact and Convergence ................................................................. 45 2.4.1

Various Grammatical Levels ............................................ 46

2.4.2

The Restructuring of Grammars ....................................... 50

2.5 Conclusion ........................................................................................ 53 3. The Preparation ....................................................................................... 55 3.1 Budgeting and Reservation .............................................................. 55 3.1.l

Travel ................................................................................ 56

3.1.2

Board and Lodging ........................................................... 56

3.1.3

Remuneration for the Informants ...................................... 57

3.1.4

Communication Network Charges.................................... 58

3.1.5

Stationary .......................................................................... 58

3.1.6

Equipment and Accessories .............................................. 59

3.1.7

Data Processing ................................................................ 61

3.1.8

Word Processing/Typing .................................................. 61

3.1.9

Reprographic Services/Xeroxing ...................................... 61

3.1.10 Printing ............................................................................. 61 3.1.11 Books and Journals ........................................................... 62 3.1.12 Contingencies ................................................................... 62 3.1.13 Overheads ......................................................................... 62 3.2 Your Luggage ..................................................................................... 63 3.2.1 .. Materials to Take Along ................................................... 63 3.2.2 .. Dress Code ........................................................................ 64 3.3 Status of the Informants/Area to be Studied..................................... 64 3.3.1

Literature study ................................................................. 65

3.3.2

Ethnology .......................................................................... 65

3.3.3

Language study in big cities ............................................. 67

3.4 Fieldwork in the Class room ............................................................. 68 3.5 Preparation of the Questionnaire...................................................... 69 3.5.1

Goals Should be Clear ...................................................... 70

Contents iii 3.5.2

Each Questionnaire Should be Numbered. ...................... 70

3.5.3

Language of the Questionnaire ......................................... 70

3.5.4

Length of the Questionnaire ............................................. 71

3.5.5

Information Regarding the Informant ............................... 71

3.6 Contacts in the Field ........................................................................ 72 3.7 The Second Stage: In The Field....................................................... 72 3.7.1

What Language to Use for Eliciting Data? ....................... 73

3.8 Choosing Informants ........................................................................ 74 3.8.1.

Begin at School ................................................................ 74

3.8.2.

Avoid a Language Teacher .............................................. 75

3.8.3.

Choose Both Male and Female Informants ..................... 76

3.8.4.

Choose All Age Groups But Not Below Twelve............. 76

3.8.5.

Choose All Sections of the Stratified Society.................. 77

3.8.6.

One Willing Informant is Better Than Ten Unwilling Ones ......................................................... 78

3.9 The Role of the Interpreter ............................................................... 79 3.10 Your Own Behavior in the Field ..................................................... 80 3.10.1 The Investigator ............................................................... 80 3.11 Being a woman is a Blessing ........................................................... 82 4. Elicitation........................................................................................................ 84 4.1 Various Methods ............................................................................... 84 4.1.1

Observation Method ......................................................... 84

4.1.2

Interview Method.............................................................. 84

4.1.3

Sending Questionnaire Method ........................................ 85

4.1.4

Documentary Source Method ........................................... 85

4.2 Interviewing Informants ................................................................... 85 4.3 Interrogation Techniques ................................................................. 87 4.3.1

Translation ........................................................................ 88

4.3.2

Contact Language ............................................................ 90

4.3.3

Pictorial Representation .................................................... 91

4.3.4

Substitution Interrogation ................................................. 91

4.3.5

Associative Interrogation .................................................. 92

iv A Manual of Linguistic Fieldwork and Structures of Indian Languages 4.3.6

Paraphrase ......................................................................... 92

4.3.7

Cross Interrogation ........................................................... 93

4.3.8

Stimulus Interrogation ...................................................... 93

4.3.9

Examples and Illustrations ................................................ 93

4.4 Transcription ..................................................................................... 93 4.4.1

Narrow or Broad ............................................................... 94

4.4.2

IPA or American ............................................................... 94

4.5 Data Collection: Various Stages ..................................................... 97 4.5.1

Stage I: Basic Word List ................................................... 98

4.5.2

Stage II: 400 Word List .................................................... 99

4.5.3

Stage III: Small Phrases .................................................... 107

4.6. Morphological Topics ...................................................................... 109 4.7 Dichotomy between Noun and Verb ................................................ 113 5. Morphology And Word Formation .............................................................. 115 5.1 Inflection ........................................................................................... 115 5.1.1

Noun Morphology ............................................................ 115

5.1.2

Pronoun Morphology ........................................................ 124

5.1.3

Case Markings and Postpositions ..................................... 127

5.1.4

Morphology of Adjectives ................................................ 131

5.1.5

Stage IV Simple Sentences ............................................... 135

5.1.6

Morphology and Syntax of Adverbs................................. 137

5.1.7

Verb Morphology ............................................................. 139

5.1.8

Stage V: Complex Sentences ............................................ 152

5.2 Derivation .......................................................................................... 155 5.2.1

Particle –wala ................................................................... 157

5.2.2

Morphological Causatives ................................................ 159

5.3 Reduplication .................................................................................... 161 5.3.1

Morphological................................................................... 163

5.3.2

Lexical .............................................................................. 165

5.4 Compounds........................................................................................ 171 5.4.1

Endocentric ....................................................................... 171

5.4.2

Exocentric ......................................................................... 173

5.4.3

Appositional or Associative .............................................. 173

Contents v

6. Syntax and Semantics .................................................................................... 175 6.1 Inquiring into Syntax and Semantics ............................................. 175 6.2 Word Order Typology ....................................................................... 176 6.2.1

Characteristic Features of SOV ....................................... 177

6.3 Topic and Focus................................................................................ 178 6.4 Interrogation ..................................................................................... 178 6.5 Negation ............................................................................................ 180 6.5.1

Salient Features ................................................................. 180

6.5.2

Negative Copula ............................................................... 182

6.5.3

Deletion............................................................................. 184

6.5.4

Scope of Negation............................................................. 185

6.5.5

Other Related Features...................................................... 186

6.6 Complex Predicates........................................................................... 187 6.7 Explicator Compound Verbs ............................................................ 188 6.7.1

Aspectual .......................................................................... 189

6.7.2

Adverbial .......................................................................... 190

6.7.3

Attitudinal ......................................................................... 191

6.8 Dative Subjects .................................................................................. 192 6.8.1

Non Experiential ............................................................... 193

6.8.2

Experiential ...................................................................... 194

6.8.3

Subject Properties ............................................................. 197

6.9 Complementation .............................................................................. 199 6.9.1

Types of Complements ..................................................... 199

6.9.2

Varying Forms .................................................................. 200

6.9.3

Nominalized Complements .............................................. 201

6.9.4

‘whether’ Clauses ............................................................. 202

6.10 Converbs/Conjunctive Participles ................................................... 203 6.10.1

Semantic Functions ......................................................... 205

6.10.2

Some Idiosyncratic Constructions .................................. 207

6.11 Anaphora .......................................................................................... 208 6.11.1

Definition ........................................................................ 208

6.11.2

Reflexives ....................................................................... 208

6.11.3

Intensifiers ...................................................................... 211

6.11.4

Reciprocals ..................................................................... 211

vi A Manual of Linguistic Fieldwork and Structures of Indian Languages 6.12 Coordination ................................................................................... 213 6.12.1

Coordinands .................................................................... 215

6.12.2

Gapping........................................................................... 217

6.13 Adjectival Clauses ........................................................................... 218 6.13.1

Participial Constructions ................................................. 218

6.13.2

Relative-Correlative Constructions................................. 219

7. Social Aspects ................................................................................................. 221 7.1 Kinship Terms ................................................................................... 221 7.1.1

Non-Affinal...................................................................... 222

7.1.2

Affinal .............................................................................. 225

7.2 Forms of Address and Terms of Reference ..................................... 226 7.2.1

Forms of Address .............................................................. 226

7.2.2

Terms of Reference ........................................................... 230

7.3 Politeness Strategies ......................................................................... 232 7.3.1

Lexical .............................................................................. 233

7.3.2

Prosodic ............................................................................ 234

7.3.3

Syntactic .......................................................................... 234

7.4.4

Sociolinguistic .................................................................. 234

7.4 Language Shift, Retention, and Death ............................................ 235 7.4.1

Language Shift and Retention........................................... 236

7.4.2

Language Death ................................................................ 240

Appendices 1. IPA Charts for Consonants And Vowels [Including Blank Charts] .............. 242 2. The Basic Word List (Swadesh, Gudchinsky And Samarin) [A] ................... 244 3. The Basic Word List (300 Words) [B] ........................................................... 246 4. The Basic Word List (400 Words) [C] ........................................................... 247 5. The Basic Sentence List ................................................................................. 248 6. Cook’s Case Frame Matrix ............................................................................. 253 7. Word Order/ Topic And Focus/ Scrambling................................................... 254 8. Interrogation, Complementation ..................................................................... 255 9. Explicator Compound Verbs .......................................................................... 256 10. Dative Subjects ............................................................................................... 257 11. Language Shift And Retention/ Attitudes Of The Speakers ........................... 258 12. Language Death/ Obsolescence ...................................................................... 259 13. Complete Word Reduplication [Bilingual And Bi-Scriptal] .......................... 260 14. Map: Hindi Speaking States ........................................................................... 263 15. Map: Distribution of Tribal Languages………………………………………264 References ............................................................................................................ 265 Index..................................................................................................................... 283

Contents vii

Figures, Maps, and Tables Figures 1.1 Field linguistics as an input system to other fields of research .......... 1 1.2 Hierarchical status of Indian languages and dialects .......................... 9 2.1 Stages of Language Change in Contact Situation ............................... 51 4.1 Stages of Data Recording .................................................................. 98 4.2 Sample of recorded lexicon ................................................................ 101 5.1 Types of Reduplicated Structures ....................................................... 162 5.2 Sample information card on Echo formation ...................................... 170 6.1 Types of Constructions involving Dative Subjects ........................... 198 7.1 Address Forms ................................................................................... 231 7.2 Terms of Reference ............................................................................. 232 7.3 A five Value Scale used in Sociolinguistics Research ....................... 238

Maps 1.1 Official and Other languages by States and Union territories ............ 5 1.2 The Languages and Dialects of South Asia ........................................ 6 14. Appendix: Hindi Speaking States ....................................................... 263 15. Appendix: Distribution of Tribal Languages………………………...264

Tables 1.1 Scheduled languages 1991 ................................................................. 7 1.2 Non Scheduled languages spoken by more than a million ................. 7 1.3 Scripts Used in the Eighth Schedule languages .................................. 22 1.4 Language shift among Tribals in various States ................................. 18 1.5 Dialects of Hindi : 1991 ...................................................................... 20 2.1 Konkani : III Person .Pronoun. in Agentive Case............................... 27 2.2 Possessives in Maithili ........................................................................ 27 2.3 Telugu Personal Pronouns .................................................................. 31 2.4 Kharia Verb ‘to sit’ ............................................................................. 35 2.5 Kharia Transitive/ Intransitive Verb Inflection .................................. 36 2.6 Agreement Relations in Tibeto-Burman ............................................. 41 2.7 Gender Inflection in Kurux ................................................................. 47 3.1 Budget proposal for fieldwork in linguistics ...................................... 62

viiiA Manual of Linguistic Fieldwork and Structures of Indian Languages 3.2 Male and Female Speech in Kurux ..................................................... 76 4.1 Kurux Kinship Terms ......................................................................... 89 4.2 Santhali Locatives ............................................................................... 90 4.3 Magahi Verb forms ............................................................................. 92 4.4 Common Vowel Sounds in Indian languages ..................................... 96 4.5 Common Supra–segmental Sounds .................................................. 96 4.6 Common Indian Consonant Sounds ................................................... 97 4.7 Co-articulated consonants in Khasi .................................................... 105 4.8 Dipthongs and Tripthongs in Khasi .................................................... 105 4.9 Consonant Clusters ............................................................................. 106 5.1 Gender marking in Telugu ................................................................. 117 5.2 Nominal Inflection in Konkani ........................................................... 120 5.3 Sambalpuri Nouns .............................................................................. 120 5.4 Gender, Number and Case markings in Kurux male and female speech ............................................................................... 121 5.5 Gender, Number and Case markings in Kharia nouns........................ 121 5.6 Personal Pronouns in Santhali ............................................................ 125 5.7 Reflexives in Santhali ......................................................................... 125 5.8 Genitive/Possessive pronominals in Santhali ..................................... 126 5.9 Two Way Case System in human languages ...................................... 129 5.10 Semantic Prototype of Adjectives..................................................... 131 5.11 Possible Aspectual forms .................................................................. 149 5.12 Inflection and Derivation .................................................................. 155 5.13 Derivation from Verb to Noun in Kharia.......................................... 157 5.14 Areal Typology of South Asian Reduplication................................. 171 6.1 Marathi Negative Verbs ..................................................................... 183 6.2 Explicators indicating Aspect ............................................................. 190 6.3 Most Common Verbs Employed In ECV Constructions .................... 192 6.4 Reciprocals in Santhali ....................................................................... 212 7.1 Pronominal system of the Andamanese language .............................. 233

ix

Acknowledgements This book could probably not have been written without the help, support and love of many people. My foremost thanks and gratitude go to Bernard Comrie, who gave me the opportunity to visit the Max Planck Institute at Leipzig, Germany as a guest scientist where I wrote this manual. I am grateful to him for providing an environment in which ideas could be shared across traditional and national boundaries. The facilities and academic atmosphere of the institute were just right for such writing. I realized that I was among various field linguists with whom I could discuss, debate and share my experiences. I am especially indebted to Maria Polinsky and Martin Haspelmath who read through parts of this manual and gave insightful suggestions. Without the critical assessment of the first draft of the major portion of this manual by Maria, I could not have reached certain decisions about the organization and content of the book. Any remaining drawbacks are mine alone. The book has further benefited from helpful comments by Claus Peter Zoller and V. Prakasam. Thanks are due to Orin Gensler who readily made his personal library available to me and to Brigitte for going through the first draft of this book and suggesting several corrections. My special thanks are due to R. S. Gupta, colleague, friend and cosharer of the field trip blues. His help in the field in taking wise decisions at the right time and his sense of humor kept us, the fieldworkers, full of enthusiasm. I am grateful to Ayesha Kidwai who shared my joys and sorrows of teaching the Field Methods course and thus enriched my thoughts, first, as student and then, as colleague. I am also thankful to my students, past and present, who accompanied me on various field trips, discussed and shared their problems and experiences with me. Special mention should be made of Lakhan, Pradeep, Sunita, Priyanka and Shailendra. They constitute the invisible participants in the genesis of this book. What my children, Kani and Anuj must have gone through right from their childhood at not having their mother around for days and weeks each year, I am at a loss to say. Though they did accompany me on some field trips, more often than not, they were left behind. Yet, and I am grateful to them for this, they proudly believed that my work was of ‘national’ importance and eased the pain of leaving home. Finally and most important I am grateful to my husband Satish, to whom this book is dedicated. His constant encouragement, his complete understanding of the seriousness of my work, and his sacrifice in sparing me year after year for field trips, are qualities which are rare to find. I cannot find words eloquent enough to express how much I owe him. Anvita Abbi Leipzig, Germany

x

A Manual of Linguistic Fieldwork and Structures of Indian Languages

Preface Nothing is harder than opening up questions that we already think are solved. There is by now a huge literature on the problems of obtaining linguistic information from informants. Why another book on linguistic field methodology? Well, I always felt that none of the available literature caters to the needs of those scientists who want to work in the South Asian subcontinent. The South Asian peoples and their different and distinct cultures warrant a specific manual that helps the linguists avoid committing communicative blunders. When one decides to work on the languages of South Asia, it becomes increasingly evident that a knowledge of theoretical linguistics is essential, but not sufficient. What one needs to know in addition is the linguistic structures of the language families these languages belong to, the type of linguistic network in which the speakers interact, and the way each and every language is constantly facing impending change and restructuring of its grammar in intensive contact situations. This need has induced me to write this book. In it, I have tried to share all with my readers, my field experiences, my observations, my opinions and my judgements. The manual introduces the readers to linguistic field methodology which may be considered mundane by some, as collecting data is viewed as an intrinsically sound pursuit, but may be considered very significant by others, as the methodology is context-based and linguistic structures are varying and very challenging for theoretical linguists. After all, a meaningful contribution to the theory of grammar can only be made if we expand the data and the language samples, as looking at insufficient data from too few languages makes us ‘frogs in the well’. More than two decades of fieldwork on Indian languages have sensitized me to the ways Indian people participate in day-to-day communicative situations. Coupled with this, teaching a course on field methods at the Jawaharlal Nehru University for the last twenty-five years guided me to undertake this job of writing the manual. My fieldwork trips were rarely undertaken alone. We always went in a team that was comprised mainly of students enrolled for the course accompanied by an additional faculty member (if I was lucky). However, there were times when I alone had to tread the field. This was especially true of the cases when I worked in the northeastern area on Khasi and on the Tibeto-Burman languages. The manual is divided into seven chapters and detailed appendices. Chapter 1 introduces the reader to the language scene of India, the multiplicity of the languages in the region, the language policy of the Indian government, the question of major and minor languages, and the sociocultural makeup of the society. A knowledge of all these aspects is very essential before one decides to undertake fieldwork on these languages. Though I have not discussed the linguistic and social situations in South Asian countries other than India, yet information contained here should be relevant to other countries as well. Hierarchical social patterns, prevalent multilingualism, and recognition of only a handful of languages by their respective governments characterize all the South Asian countries. The second chapter is about the linguistic features. A discussion of the characteristic features of individual language families, the areal features of the languages spoken in the subcontinent, and the nature of the conflicting grammatical patterns that emerge in an intensive and long contact situation is relevant and pertinent to the entire subcontinent.

Preface The third chapter is all about how a field linguist should prepare herself for the fieldwork. The dos and don’ts of the field investigation are discussed here. My exposure to Indian societies and the blunders that many of our team members committed inadvertently in the past motivated me to write this chapter. The readers might find helpful hints here, which if remembered I expect will go a long way in smoothing the running of the project. The fourth chapter is devoted to the ways and means of eliciting data from native speaker informants. It is concerned with the elicitation of phonological structures of the language under consideration and the major morphological topics, the discussion of which constitute a significant part of the following chapter. Chapter five deals with word formation devices commonly used by Indian languages as well as the morphology therein. Elicitation and discussion on syntax, semantics and pragmatics constitute the sixth chapter. The topics covered here are not exhaustive but relevant from the point of view of the areal syntactic characteristics. The last chapter, though it is the shortest, deals with a very important aspect of language, i.e. its power of being the vehicle of communication and its ability to mark the identity and dignity of its speakers. It deals with the question of vitality of a language and fieldworkers are given guidelines on how to elicit information on various stages of language shift and ultimately on language death. Finally, a large number of questionnaires tested by my team members and me are given in the appendices to facilitate the fieldworker in designing her/his own. Unless otherwise specified, most of the reported data in the book was collected by the university students and the teachers including myself who conducted the course ‘Field Methods’. Most of the data is drawn from first-hand information and is in print for the first time. A penultimate note. I started writing this manual when I was visiting the University of Heidelberg in the spring of 1998 and was engaged in teaching a course on field methods. However, only a few pages were written then. Most of the writing (and rewriting) was done at the Max Planck Institute of Evolutionary Anthropology in Leipzig where I was a guest scientist between July 2000 to December 2000. This manual grew out of a major concern of preserving the linguistic diversity of India. Although Indian linguists have maintained that language maintenance has been a characteristic feature of the Indian languages, the last two hundred years have seen a fast-depleting stock of languages in urban India. If the pulse of a language lies in the youngest generation, then it is very sad to see that this pulse is beating very slowly and at times is non-existent.

Anvita Abbi December 25th, 2000 Leipzig

xi

xii A Manual of Linguistic Fieldwork and Structures of Indian Languages

Abbreviations 1 2 3 AA abl abs acc adj adv agr A/agt aux caus class clt cm comp cond corel cp/CP dat def det DR du ECV emph EP erg EW excl exmp exp ew f/fem foc fut GNP gen H hon hw IA imp impf incl

first person second person third person Austro-Asiatic ablative absolutive accusative adjectival adverbial agreement agent auxiliary causative classifier clitic comparative marker complement(iser) conditional correlative conjunctive participle dative definite determiner Dravidian dual explicator compound verbs emphatic emphatic particle ergative echo word exclusive example expressives echo word feminine focus future gender, number, person genitive honorific honorific head word Indo Aryan imperative imperfect inclusive

inf instr interro intr loc m/masc mh MIR Mn Neg nh nom nt O obl oblg P part pcl perf pl prog prp pr/prs prt/part ps pst refl rel S/subj sg skt SM TAM TB TN tr VR

infinitive instrumental interrogative intransitive locative masculine middle honorific mirativity meaning negative non honorific nominative neutral object oblique obligational participant partitive particle perfective plural progressive prepostion present participle part of speech past reflexive relative subject singular Sanskrit subject marker tense, aspect, mood Tibeto-Burman title name transitive verbal reflexive

Chapter I Linguistic Fieldwork and the Indian Scene

1.1

All About Linguistic-Field Research

In a land of multilingual, multicultural, and multi-religious environment, interlaced with diversity of food, dress, habits, flora, fauna and ailments, there runs a multicoloured thread which binds all these elements together, not because of the expected or inherent ‘unity’ in diversity but because of the fact that the Indian subcontinent is different from the rest of the world. This distinctiveness of the Indian character (which we believe will unfold as we proceed) forces any social scientist to investigate into the social configuration of the country/subcontinent independent of any other ‘standard’ model of investigation prevalent so far. This very need to examine ‘afresh’ any field of study that warrants empirical testing such as language science being the study of people’s language in their natural surroundings, certainly justifies the writing of this book 1.1.1

Field Linguisitcs As An Input System to Other Fields

Though language has interested linguists, anthropologists, and sociologists for more than a century, the study of language, its structure, and its contextual use occupied the minds of linguists as early as the fifth century B. C. In modern days, language research can be oriented either empirically or non-empirically, based on the interests of the scientist. Empirically oriented research in linguistics is commonly known as ‘field research’ or ‘fieldwork’. The branch of linguistics that deals with the knowledge of field research or fieldwork is known as ‘field linguistics’. Lately, there has been a revival of fieldwork as the importance of ‘first-hand information‘ is duly recognised. Regardless of the area of linguistics under study, field linguistics becomes indispensable. Descriptive linguists, topologists, universalists, sociolinguists, grammarians, and lexicographers, all need to verify their hypotheses, which can only be done with real-life language data from the field. The heterogeneity of linguistic data drawn from a large cross-section of a society is of paramount importance in studying language in society or the sociolinguistic nature of the community under consideration. Field linguistics serves as an indispensable input system for various types of linguistic research (see fig.1.1.). Theory of grammar

Typology

Language in society

Individual grammars

Field Linguistics

Universals

Sociolinguistics

Fig. 1.1 Field linguistics as an input system to various fields of research

2 A Manual of Linguistic Fieldwork and Structures of Indian Languages

In this manual, we would like to explore linguistic fieldwork. Is there a standardised methodology of conducting fieldwork? If yes, what does it involve and how should field research be undertaken in the Indian subcontinent? Considering the unique nature of the language scene in India, how should a standard field researcher orient herself/himself so as to undertake the task with a minimum of difficulties? What knowledge of the nature of Indian languages and their structures is helpful to conduct fieldwork in the South Asian countries? Is there a specific sociolinguistic environment that a field worker must be aware of before s/he begins her/his investigation? We would like to attempt to answer all these questions in the following pages. We shall discuss several issues relevant to conducting linguistic fieldwork in this chapter. First and foremost, we shall consider the nature of linguistic fieldwork. Secondly, we would like to introduce our readers to the constitution of India, especially the sections that pertain to the language policy of the country. We would like to familiarise our readers with the existing dichotomy of rural and urban settings and the linguistic behaviour of the communities therein. Closely linked to the two settings is the question of bilingualism and language shift or adoption which we will consider in some detail. The discussion will therefore also introduce the readers to the contact language/s of the country. The purpose of discussing these social and linguistic issues is that we believe that without the a priori knowledge of them a field investigator may commit communicative and linguistic blunders. In short, this is a manual that will tell you how to conduct fieldwork on Indian languages. 1.1.2. What does linguistic fieldwork involve? Field research involves observing and recording real spoken language/s as used by various members of a society in various contexts and times, and then analysing the data according to some specific linguistic methodology. The field researcher has to rely completely upon the community whose language s/he undertakes to study. For this reason, fieldwork becomes a total life. One has to develop self-criticism and selfawareness along with a strong appreciative perception of the speech community under study. This is a kind of work that demands total immersion in linguistic work comprising observing, interviewing, recording, eliciting and analysing linguistic data. The entire day in the field is spent in planning, observing, and eliciting data. Just as the concentration of the archer Arjuna is on the eye of the bird, one has to keep one and only ONE aim in mind, i.e. to collect and elicit linguistic data to the best of the workers’ ability, with utmost accuracy and with least distortion. Here, I would like to agree with the famous sociologist and fieldworker Evans-Pritchard (1973: pp 2-3), who advocated: Perhaps it would be better to say that one lives in two different worlds of thought at the same time, in categories and concepts and values which often cannot easily be reconciled. One becomes a sort of double marginal man, alienated from both worlds. 1.1.3 The Participatory Character One has to almost forget her/his identity to be a good fieldworker. Without such a commitment one can neither participate in the community‘s day-to-day life nor can one win the confidence of the members of the society under study. Field research, in

1. Linguistic Fieldwork and the Indian Scene 3

other words, is participatory in nature and hence demands understanding, friendship, and commitment on the part of the fieldworker. There is no other way to obtain insights into the language and “insides“ of the speech community under observation. Another important fact to be brought home is that a study, which is not based on real fieldwork, will run the risk of obscuring or distorting the basic characteristics of the language under consideration. Armchair linguists may float theories, but it is fieldwork, alone which can test such theories or generate new ones. Linguistic studies without the basis of first-hand data misses out specific characteristic features of language. We would like to emphasise that in the 21st century, when linguists are busy trying to identify language universals on the one hand and areal features on the other, field research on hitherto unknown languages and lesser known languages as well as on well-documented languages becomes an inevitable necessity. Field research demands linguists who are open-minded and not ‘hung’ upon a particular theory. Linguists who are participatory observers, who can break the social circles of apprehensions, who have a good knowledge of the discipline and possess a basic knowledge of the areal features of the languages under investigation, make good field linguists. Again, language is a living organism and thus has to be studied within its sustaining environment. If a linguist is not prepared to undertake fieldwork for the language s/he wants to analyse and write the grammar of, but wants to depend solely upon published material, it would be like seeing the picture of a garden without ever being there. One would have an idea of the language, but the language as a means of communication, as a means of sociosemantic nuances and, above all, as a means of marking the identity of the community under study would be missed. Languages change over the course of time and, like any other organism, may die, too. Hence it becomes of utmost importance to be exposed to the language at regular intervals which can be done only by undertaking fieldwork. India with its multitude of languages offers the best ground for theorists, topologists, and sociolinguists. However, most of the languages of India have yet to see the good fortune of linguists working on them. In the absence of much written material available, it becomes imperative for a linguist to take the path of field research. There is no substitute for acquiring first-hand data, which has not yet been distorted (unintentionally) by a grammarian or language-learner befitting some theory of linguistics. This brings us to a very relevant question. 1.1.4 Theory-Independent Should field research be theory-independent? My experience shows that going to field without any rigid framework of linguistic theory always helps in bringing out characteristic features of the language concerned, which perhaps could have gone unnoticed. Theory-dependency more often binds the fieldworker’s hands and restricts her/his vision. Of course, one has to be very clear about the grammatical axioms and semantic constructs (which are more or less universal) in mind. The experience shows that such an outlook helps a theory to ‘emerge’ right in the field. One should not associate theory-independency with vagueness of the linguistic work. A good training in linguistics is imperative. One has to give enough freedom to the

4 A Manual of Linguistic Fieldwork and Structures of Indian Languages

field linguist to evolve and write a new theory befitting the data rather than fitting the data to the theory. The same argument holds for dependency on a particular grammar. As we have witnessed in the past, all grammars are either written on the pattern of Sanskrit or English, consequently missing out on various salient features of individual languages. For instance, can one imagine that languages may not even distinguish overtly between nouns or verbs? Or can one think that by duplicating words languages may indicate aspectual functions, or that languages do not mark three basic types of tenses, or that languages do not make a distinction between a word and a sentence? How can all these types of languages be fitted into a grammar type? As we proceed this point will become clear. The possibility of various grammatical types has motivated us to discuss the structures of Indian languages in this manual. This will help field investigators shed some of her/his linguistic prejudices and adherence to a particular theory and type of grammar. We would like to concentrate on different aspects of fieldwork in an Indian context. We shall try to avoid repeating the already tested methodologies and techniques advocated by various linguists. Readers are advised to consult the wellknown field manuals available such as Pike (1943), Nida (1946,47), Kibrik (1977), Samarin (1967), Payne (1997), and Vaux and Cooper (2000), among others. Let us first familiarise ourselves with the language composition of the Indian society, both rural and urban, so as to prepare ourselves with the necessary background information before we begin our fieldwork. 1.2.

The Composition of Language Scene in India

India with its 114-odd languages (1991 Census) offers a rich ground for the linguist. The equally rich heritage of bi-/multi-lingualism coupled with steady language loss invites more linguists to undertake fieldwork than any other scientist. The unique nature of linguistic heterogeneity coexisting with the typological homogeneity among various languages and language families demands a different approach to fieldwork. Linguistic fieldwork in India cannot be undertaken unless the fieldworker has a basic knowledge of the Indian language scene, its composition, its structures, its areal features, and its vibrant nature. 1.2.1. The Constitution India represents five language families (Indo-Aryan, Dravidian, Austro-Asiatic, Tibeto-Burman, and Andamanese) and according to the 1991 Census, 114 mother tongues are spoken by 800 million people (96.29% of the population, estimated in 1991 at 838,583,988). The constitution of India recognises eighteen languages (generally termed ‘scheduled languages’ as they are listed under VIII Schedule, articles 343-51 of the constitution; see table 1.1) and Indian states are broadly organised on a linguistic basis (see map 1.1). There are many more languages and their varieties in every state and no state is thus monolingual in nature (see map 1.2: ‘Languages and Dialects of South Asia’). According to the 1991 census over 95% of India’s population speak and use one of the scheduled languages. This manual is an humble attempt to expose the readers to the methodology of fieldwork on the various Indian languages spoken by such a large population.

1. Linguistic Fieldwork and the Indian Scene 5

Map 1.1

6 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Map 1.2

1. Linguistic Fieldwork and the Indian Scene 7

Table 1.1 Scheduled languages in descending order of numbers of speakers 1991 Languages 1. Hindi (IA) 2. Bengali (IA) 3. Telugu (DR) 4. Marathi (IA) 5. Tamil (DR) 6. Urdu (IA) 7. Gujarati (IA) 8. Kannada (DR) 9. Malayalam (DR) 10. Oriya (IA) 11. Punjabi (IA) 12. Assamese (IA) 13. Sindhi (IA) 14. Nepali (IA) 15. Konkani (IA) 16. Manipuri (TB) 17. Kashmiri (IA) 18. Sanskrit (IA)

Mother Tongue Speakers 337,272,114 69,595,738 66,017,615 62,481,681 53,006,368 43,406,932 40,673,814 32,753,676 30,377,176 28,061,313 23,378,744 13,079,696 2,122,848 2,076,645 1,760,607 1,270,216 ++ 49,736

Percentage to total population 39.85 8.22 7.80 7.38 6.26 5.13 4.81 3.87 3.59 3.32 2.76 1.55 0.25 0.25 0.21 0.15 ++ 0.01

•

The percentage of speakers of each language has been worked out on the total population of India, including the projected population for Jammu & Kashmir where the 1991 census was not conducted due to disturbed conditions.

•

++ Full figures for Kashmiri language are not available as the 1991 census was not conducted in Jammu & Kashmir due to disturbed conditions.

The languages that are not listed in the VIII Schedule are no less populous. Some of the non-scheduled languages (those that are left out from the list of 18 mentioned above) are spoken by more than a million people (Table 1.2). Though spoken by such a large number of people, these languages in the Indian context are known as ‘minority languages‘, because they are non-scheduled. Table 1.2. Distribution of Non-Scheduled Languages spoken by more than a million in States and Union Territories. 1991 Languages Bhili (Indo Aryan) Bodo (TibetoBurman) Gondi (Dravidian) Kurukh/Oraon (Dravidian)

Total number of speakers in India 5,572308 1,221,881 2,124,852 1,426,618

States where spoken most Madhaya Pradesh Rajasthan Dadar Nagar haveli Arunachal Pradesh

Number of speakers in states 2,215,399 2,258,721 76,207 1,184,569

Madhya Pradesh Maharashtra Bihar Madhya Pradesh

1,481,265 441,203 681,921 393,825

8 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Munda/Mundari (Austro-Asiatic)

1,275,272

Santhali (AustroAsiatic)

5,216,325

Tulu (Dravidian)

1,552,259

West Bengal Orissa Assam Bihar Orissa Bihar Orissa West Bengal Karnataka Kerala Maharashtra

192,833 85,358 89,643 692,308 398,303 2,546,655 661,849 1,858,010 1,378,779 111,670 56,373

1.2.2 The Minority languages and Tribal languages It should be clear from the previous section that languages not included in the VIII Schedule are generally considered minority languages. Before we go into the assessment of these languages, it is necessary to understand the notion of tribal languages. There are four significant issues to be noted in this context. •

Tribes in India do not form a neat homogenous social and cultural category. These tribes are scheduled as per Article 342 of the constitution by the president and the parliament. The concept of tribe in India is an administrative, judicial, and political one.

•

The scheduled tribes constitute 623 varied communities (Singh and Manoharan 1997), but less than half of them speaks a tribal mother tongue.

•

Languages spoken by these scheduled communities are considered ‘tribal languages’. There is no linguistic definition of tribal language/s.

•

As we noted above, the government of India reports 114 mother tongues, of which only 18 have the status of a scheduled language. Out of 96 non-scheduled languages 92 are tribal. This implies that the tribal languages constitute almost 96% of the non-scheduled languages and approximately 81% of the total languages of India listed in the census.

Tribal languages as well as other minority languages have varied social status in Indian society. In addition to this there is a large number of languages that are classed as dialects or mother tongues of these 18 scheduled languages. These dialects as well as the tribal languages are seen as minority languages in Indian social and academic worlds. There is a layered hierarchy status that can be assigned to these languages as shown below in figure 1.2. The different layers are arranged in a descending order. The lowest in the figure (i.e. 149 languages) occupy the lowest social status while the next layer (i.e. 96 in all) enjoy a little higher status than the previous one and so on so forth. Fieldworkers should be aware of this cline of status before beginning their work on these languages and should be prepared to interact with the speakers of the community concerned.

1. Linguistic Fieldwork and the Indian Scene 9

SL 18 Dialects of SL 67 Non SL 96 Dialects of Non SL 149

Fig. 1.2 The Hierarchical status of Indian languages and dialects (330) The layered hierarchical status (see fig. 1.2) accorded to the languages of the Indian population stems from the current language policy, not from the number of speakers The government of India has labelled 312 (330–18) languages as ‘minority languages’ (although the government does not use the term minority. It uses the term non-scheduled and its varieties as mother-tongues and hence communities speaking these languages as mother-tongues are considered minority communities by the Indian administration). Indicative of the varying linguistic status in the society of their use, these four types of languages show a direct correlation to the social and economic status of the speech communities. For example, Indo-Aryan languages such as Maithili, Dogri, and Rajasthani are recognised by the Sahitya Akademi, although they occupy the second rung of the hierarchical ladder. This is a privilege not enjoyed by the rest of the ‘minority languages’ down the scale. Before beginning the fieldwork on any of these languages one must keep in mind this cline of status accorded to them. 1.3.

The Hierarchical Structure of The Indian Society

Indian society can be broadly divided into three types of regions/areas, each hierarchically divided into three or four types of groups. 1. Rural A) High Caste B) Low Caste C) Tribals 2. Urban A) Educated or High Caste or Both B) Uneducated C) Semi-Educated 3. Cosmopolitan Cities A) Educated B) Semi-Educated C) Uneducated D) Labour Class Although this division is an oversimplification, it will be functional for our purpose. There is some overlap between these sections. Thus, high caste groups would largely overlap with educated groups in cities. Conversely, uneducated or

10 A Manual of Linguistic Fieldwork and Structures of Indian Languages

semi-illiterates in cities might overlap with tribals and low caste in rural areas. In the wake of employment opportunities in urban areas and big cities, the low caste and tribal groups migrate to urban areas and thus merge with the semi-educated urban classes. Metropolitan cities have a large number of labourers employed in various factories and offices who might constitute a separate class because this category incorporates in its fold both educated and semi-educated people. There is a constant situation of ‘fluidity’ in the composition of Indian society. These ten basic divisions will eventually be refined by incorporating economic class and occupational structure also. What one has to remember is that before beginning linguistic fieldwork one must identify (1) which of the three areas and (2) which composition of the society, i.e. uneducated, semi-educated, or high or low caste you are investigating. Each division exposes you to different types of linguistic samples. These divisions become especially relevant in sociolinguistic research. We shall consider them in detail when we discuss sociolinguistic research in chapter 7. 1.4

Bilingualism and Indian society

The population of the Indian subcontinent has always been bilingual, as the languages of various linguistic families have been in contact with each other for more than 2500 years. India thus offers both high and low convergence areas, e.g. the states of Jharkhand and Maharashtra are high convergence areas while the states of the Northeast are low convergence areas. Unfortunately, the census no longer collects the figures for bilingualism, however, the study conducted by Singh and Manoharan (1993) reveals that Indians are highly bilingual in both cities and villages, with monolingualism being an exception. Indian bilingualism is rather fascinating and has recently been studied by various scholars1. We will not go into details of each and every aspect of this bilingualism. However, we would like to mention those features that concern a field linguist conducting linguistic research in India. •

Bilingualism among tribals is 50% higher than the national average. Two forces, retention and shift, coexist within the same language group. While the urban tribals of the Munda family and those of the Dravidian family are seen to shift easily, rural tribes of the same families prefer to maintain their indigenous languages. Details are given below.

•

A very low percentage of urban tribals in general are monolinguals in their ancestral languages. Ethnolinguistic minority status induces a negative attitude toward language loyalty.

1

•

Retention and shift varies from region to region.

•

Bilingualism among migrants of the scheduled languages shows a high percentage of mother-tongue retention.

Abbi (1986), Agnihotri (1992), Khubchandani (1983), Pandit (1972), Srivastava (1994), to a mention a few.

1. Linguistic Fieldwork and the Indian Scene 11

Let us consider some examples from a typical rural and urban area. Following sections describe the personal observation of the author and her students. 1.4.1 The Rural Scene As rural areas are constituted of caste and non-caste people such as tribals we would like to expose our readers to both of the groups. Societies that do not belong to either of the groups offer parallel examples, i.e. low class is equivalent to low caste, and high class is equivalent to high caste. 1.4.1.1

Bilingualism among tribals and non tribals

Most of the lower-caste people and tribals in rural areas are bilingual (or bidialectal) in their mother tongues and the regional language/s of the area. We notice a kind of societal or equitable bilingualism (where competence in the two languages is of equal degree) among tribals using two languages - one being their mother-tongue and the other the lingua franca of the region they reside in. In the rural area of Chotanagpur, Kuruxs use Sadari (the lingua franca) and Kurux (the name of the language is the same as that of the tribes, i.e. Kurux or Oraon) in various domains, such as at home, in playgrounds and market-places, as well as in dhumkuria (a kind of village community centre or chaupal which is used, among other purposes, for religious functions). 1.4.1.1.1

The Reality: Munda Experience

Another very interesting situation that has been observed and should be brought to the notice of any fieldworker is that often a tribe leaves its language for another tribal language which is equally marginalised. Thus, Mundari-speaking Oraons may be found in Ranchi City and its vicinity, the heart of Munda civilisation. An apparent change of identity is observed when one considers Oraons living in and around the villages of Mesra, Garihotwar and Angara. Oraons in these areas have not only changed their clan affiliation, but also speak a Munda dialect called Kera Munda. Culturally and linguistically it is difficult to distinguish these Oraons from local Mundas. This is a clear case of the culture and language of one minority community being superimposed on that of another minority. Our fieldwork in the Chotanagpur area reveals that Oraons are mostly multilingual, speaking Kurux at home, Sadri at the market, and Hindi for inter-group communication. People of high castes or those in power have passive bilingualsim in the speeches of tribal groups and those of the lower castes. In addition to this, the population of rural areas in the so-called ‘Hindi belt’ (see appendix 14) in general has some command over the spoken variety of Hindi of that region. Each ‘Hindispeaking’ state has Hindi or some other language variety of Hindi as a lingua franca which is used by all males and females. 1.4.1.2.

Gender basis for bilingualism

One must try to understand the structure of bilingual behaviour of males and females regarding their own language. The pattern of bilingualism in males vs. females differs greatly in some regions while being more or less the same in others. For instance, in villages in Ranchi district, Oraon (or Kurux) women outnumber men as

12 A Manual of Linguistic Fieldwork and Structures of Indian Languages

speakers of Kurux. In as many as 23 blocks it is the principal language of the female population. This could be ascribed to the absence of contact with caste people or to the lack of female education of Kurux women. By contrast, the marriage relations of the Oraon females in Basia, Khunti, and Tamar-I blocks extend to villages where other tribes and castes have Sadari as their mother tongue. The long contact with such tribes and castes has negatively influenced the language attachment of the Oraons. They have forgotten their own mother-tongue and are unable to teach it to their children. There is, thus, a constant increase in Oraons abandoning their language. Despite such contact pressures to lose the mother-tongue, in the core belt of the tribe men generally choose wives from among the women belonging to the above-mentioned 23 blocks so as to preserve the Kurux speech. As for the Oraon men, mobility is greater among them than among females. They are always moving from one place to another, either as government servants, teachers, compounders, doctors or domestic servants. As a consequence, they learn to speak other languages faster than their female counterparts. The trend is very conspicuous in urban areas where Oraons are mixed with other castes and tribes. A noticeable cultural characteristic of Oraons is the proficiency with which they take to other languages to the detriment of their own language. Interestingly, the situation of Kharia in the same region is just the reverse of the Kurux situation. It is the male population that has retained the mother–tongue, while the female population shows definite tendencies of moving towards adopting the regional language of the area (Malhotra 1982). This could be due to the fact that the Kharia female population is relatively mobile compared to the Kharia male population. In the entire Indian subcontinent, tribal-speaking areas are ethnically and linguistically heterogeneous. The coexistence of tribal and non-tribal populations, the latter of which are numerically, socially, and economically dominant, creates a multilingual milieu where the tribal languages cannot function as socially dominant languages. Therefore, the situation produces a large community of bilinguals. More often than not, rural women are not as highly bilingual as their male counterparts. In other words, women appear to be the conservationists of indigenous languages in general. A recent study by Singh and Manoharan (1993) has shown that the degree of bilingualism in Jharkhand alone is nearly 85% of the population. In some villages of the country, it is not surprising to find a large section of the population using three languages freely. For instance, in Southern Konkan area in Karnataka, one can find people speaking Konkan, Kannada, and Tulu (the first one is an Indo-Aryan language while the latter two are of Dravidian origin). 1.4.1.3

North East Tribal Community And Jharkhand Compared

The two largest areas with 70-100% of tribal groups speaking their indigenous languages are the Northeast and Jharkhand (Annamalai 1997; see appendix 15). However, the Northeast tribals and Jharkhand tribals operate on different sociolinguistic parameters. Their nature of bilingualism is also distinct. Jharkhandis are, to large extent compound bilinguals; nearly 85% speaking either two tribal languages of the region or their mother-tongue and a lingua franca called Sadari. It is

1. Linguistic Fieldwork and the Indian Scene 13

an area, which is inhabited by the tribals of the Munda language family and those speaking Dravidian languages along with speakers of the scheduled languages and their various dialects. Tribals of this area in contact with speakers of other language families, such as Dravidian and Indo-Aryan, to a greater extent as compared to those found in the Northeast. Their contact stretches back to more than 2000 years, leading to language change in the area of grammar2. Bilingualism in the Northeast, on the other hand is additive in nature. Most of the northeastern tribals are bilingual in Hindi, which they use for inter-tribal communication, or in another lingua franca such as Nagamese, Nefamese, Chakesang3, as well as in English if they are educated. The domains of various languages are well specified, and the mother-tongue does not clash with other regional language/s, as there is no other dominant language to clash with. A typical Naga in addition to her/his mother-tongue acquires the lingua franca Nagamese or Bazaar Hindi for inter-tribal communication and learns English at school. A typical Jharkhandi on the other hand, is bilingual in two tribal languages, say Kharia and Kurux (of which one could be her/his mother-tongue), the dominant regional language, e.g. a variant of Hindi or Bangla as the case may be, and in Standard Hindi if he/she gets an opportunity to go to school. The conflict between mother-tongue and other languages is greater, deeper and tenser in Jharkhand than in the Northeast. Linguistically, the forces of language loyalty and language revivalism as well as the need to communicate inter-tribally have given rise to two types of languages: one, the converged, hybrid varieties resulting from two or more language mixtures, such as Nagamese, Nefamese, Chakesang, Sadari/Sadani, and secondly, a variety of Hindi which is far from the Standard variety. 1.4.2 The Urban Scene The ongoing urbanisation and industrialisation create an inverse correlation between the maintenance of tribal languages or the languages of the uneducated classes and the level of industrialisation. It is not surprising to find that tribal languages are fast disappearing from most of the urbanised areas of cities and big towns. Yet, there is a large number of men and women working as household help who do maintain their mother-tongues along with the dominant regional language of the area they work in. For instance, a typical Kurux woman working in Delhi will have a good knowledge of spoken Hindi and of Oraon (in which church services are also conducted). There is another trend which can be noticed in the attitude of the migrant speech community of the tribal or low caste group, i.e. they seldom consider it their privilege to speak their mother tongues. On the contrary, ignorance of their native languages is regarded as an enhancement of their status and prestige. By speaking the dominant regional language they feel superior in comparison to other fellow-tribals/ low caste people who cannot speak it. While working on the languages of these communities, one has to be very cautious during interviewing, e.g. do you speak the language X? 2

See Abbi 1995. These languages are contact languages created out of the convergence of two or three languages. Thus, Nagamese arose from the convergence of Naga and Assamese, while Chakesang consists of structures drawn from three languages, such as Chokri, Kheza and Sangtam.

3

14 A Manual of Linguistic Fieldwork and Structures of Indian Languages

The answer will invariably be negative. A fieldworker might interpret this behaviour as that of non-cooperation but seldom it is so. By answering in affirmative the speaker of that particular language might run the risk of being isolated. Always remember that no one wants to become a ‘specimen’. The fieldworker has to learn to strike a balance between her/his goal of collecting data and maintaining the dignity and self-respect of the informant. 1.4.2.1 The Reality: Oraon Experience One of my students visited the Chotanagpur area to work on Kurux and had initial problems in locating informants, mainly because of their claims regarding their mother-tongues (see section 3.3.3). I would like to share the experience of my student with an educated Oraon research worker with my readers. “A young Oraon gentleman, who happens to be a research assistant, finally agrees to be interviewed after much reticence. I ask him the reason for this unwillingness- was it something we said or did which ruffled his feathers? ‘It‘s the attitude’ he replies slowly- “I hate being treated as a specimen,

it’s just years of being wondered at for being tribal finally takes its toll. On top, if you are highly-educated or well-placed in a professional service, you will never see the end of meeting people who will always wonder how you left your poison arrows and feather skirt and managed to come up so high. They will always have doubts about your competence - they would rather you stayed back in the forest or cave or wherever they think you are from, and dance wild war dances around bonfire than handle matters they believe you are not equipped to handle by dint of birth.” An agitated silence. One can see the man struggling to control his indignation. And one instinctively knows that one is seeing a man who is tired of having his dignity and self-respect undermined by existing prejudices and attitudes surrounding tribal identity. Yes, he carries a chip on his shoulder, but given his predicament, anyone would. He fights against an attitude which has only been strengthened down the ages by all kinds of forces: colonial forces legitimising their territorial expansion and mass destruction of indigenous people by branding them as savages; whole branches of academia devoted to studying exotic people; visual representation through art and films in which tribals are always depicted as a promiscuous, hedonistic lot peculiarly given to grotesque dancing at the slightest provocation - the list is endless. The appellations of “exotic” and “esoteric”, in fact, the very word ‘tribal’ has, if nothing else, tremendously pejorative connotations. One man fighting centuries of attitudes, I take my mental hat off to him. By now he has regained his composure and continues: “…as far as my language is concerned - it is very much alive in the Oraon-dominated districts of Gumla, Netarhat etc. Sometimes, it is such that even people whose mother-tongue is Sadri learn Oraon to communicate with their Oraon neighbours and friends”. He continues later with an opinion “…as far as urban areas are concerned, it is true that Hindi has made inroads in Oraon homes and replaced Kurux their mother tongue, but I do not see this phenomenon as very different from that in Delhi, where in most Punjabi homes the children switched to using Hindi. I don’t see how this ‘language

1. Linguistic Fieldwork and the Indian Scene 15

loss’ is peculiar to Oraons alone - it‘s symptomatic of any urban or cosmopolitan setting which is multilingual.“ One language’s meat is always another language’s poison” (Quite an apt sociolinguistic analysis!). He smiles sarcastically, “there is nothing tribal about it”. His manner changes as if he is feeling lighter after his outburst, but he pleads, “Please don’t go back and paint dark and dismal picture of the tribal situation in your report, you know, the usual ‘these tribals are happily abandoning their language now that they have got a taste of better things-syndrome’. It is a foregone conclusion that one must have more than a smattering of Hindi, and if possible English, if one has any ambition of moving up the social scale. Would you rather that we resisted Hindi with all our might and end up as ‘frogs-in-well’? Each society must pay a price for advancement. And if a language is a price we are paying for ours, I think it is a small one. Don’t you think so?” (Priyanka Bhattarcharya 1998) The statement speaks volumes and should guard any fieldworker against any preconceived stereotypical notions of about these people. We cannot afford to take anything for granted anywhere. 1.4.2.2

The Reality: Experience of Non-Tribal Migrants

Both bidialectalism and bilingualism mark Cities in India. The local population of the city speaks the local dominant language in addition to one of the varieties of that particular language for intra-community communication. English and Hindi can be added to the list if the speakers under consideration are educated. The migrant communities are seen to retain their mother tongues to a large extent for home use (this is more true of non-tribals than of tribals). Migrants are generally (nearly 100%) bilinguals. In addition to speaking their mother tongues and the local language of the city they migrated to, they may learn to use English and Hindi,4 which are taught in Indian schools. Migrant communities of non-tribals in different parts of India have been seen to maintain their mother-tongues despite the total lack of communication with the home town. A case in point are the members of the Naikan community of Kerala5 who are proficient in their mother-tongue, i.e. Telugu, as well as in the local dialect of Malayalam, despite the fact that Naikans migrated from Telangana districts some 1000 years ago and have lost contact with the original community since then. Such are the aspects of city bilingualism. 1.4.3

The Cosmopolitan Cities of India (The Mahanagars)

Delhi, Mumbai, Calcutta, Bangalore and Chennai are considered mahanagars of the country, cities with gaping economic and social disparities. These are the cities that 4

The Three Language Formula adopted by the government of India promotes the learning of Hindi and one other Indian language other than the mother tongue. Refer to C.V Bhuvaneshwari 1984. Bilingual Behaviour of the Naikan Community of Kerala. Ph.D. dissertation. Jawaharlal Nehru Univeristy. Unpublished

5

16 A Manual of Linguistic Fieldwork and Structures of Indian Languages

have created job opportunities for millions and still offer dreams for the population of rural areas and small towns. No wonder that these cities have developed a large migrant population speaking various mother-tongues. Here one can witness the great diversity of languages present in the Indian subcontinent in one place. Thus, Delhi has colonies of Sindhi-, Bengali-, Multani-, Tamil-, and Punjabi-speakers maintaining their respective mother-tongues, along with the natives of Delhi, who are speakers of Khari Boli. With such heterogeneity running throughout the city, it is bound to happen that a contact language serving different speech communities emerges from this mosaic of multiplicity. The inhabitants of Delhi and of similar big cities are not only multilingual in two or three languages, but they also use the contact language of the city which facilitates mobility across the region as well as among the various social groups. The following linguistic behaviour has been observed for all the mahanagars (metropolitan centres) of the country. • • • • •

•

Migrant groups retain their ancestral language for home use. Migrants acquire a working knowledge of the contact language Local residents of the area also acquire the contact language; however, they use it only with migrants. The language of the local residents has gone through various types of changes, depending upon the intensity and the duration of the contact with the languages of the migrant group/s. A variety of Hindi has emerged in Delhi, Mumbai, Calcutta and Bangalore which is far away from the standard variety, but nonetheless serves the purpose of contact language especially among semi-literate people. Educated people have a good knowledge of spoken and written English. English is the only contact language among the educated elite of the big cities.

A fieldworker must be aware of these facts before planning a trip to a mahanagar. S/he should also be acquainted with the locality to be visited, as migrant groups have a tendency to cluster in one particular area. Big cities serve as good areas for any survey related to bilingualism and language contact phenomena (see also 3.3.3). 1.4.4 Bilingualism and Education The speech of an average educated Indian is interlaced with code-mixing and codeswitching between their mother-tongues and the dominant regional languages of the region. The highly educated class is always bilingual in English and its mothertongue. In urban bilingualism, the proficiency level of bilingualism of the semieducated class is higher than that of the educated class. In addition, the former most often has command over at least three languages while the latter, at best over two, one being necessarily English. The semi-educated class (those who are educated up to standard VI) prefers to use the dominant regional language rather than its mothertongue even at home. This phenomenon is more noticeable in big cities than in small towns. Indian bilingualism, be it in a rural or urban environment, is very functional. Indians use a specific language for a specific purpose and thus may not have full ‘native’-like competence in L2 and L3. This role division is an important feature and

1. Linguistic Fieldwork and the Indian Scene 17

should be borne in mind before working with a specific linguistic community. As we shall see later, this aspect of bilingualism helps languages to homogenise typologically on the one hand and retain some specific language characteristic features on the other. Indian bilingualism, both rural and urban, is a way of life. The education system and the policies of the government of India reinforce multilingualism.6 The acquisition of the two contact languages English and Hindi is considered not only necessary but prestigious, too. There is another factor to this multilingualism, i.e. the multi-scriptal nature of the Indian languages. The literate and educated mass of India uses 14 different scripts to write 19 modern languages of India (including English). Table 1.3 [pp.22] demonstrates these scripts. Multi-scriptal India, however, should pose no threat to a fieldworker, as one is primarily concerned with the spoken language. 1.4.5 Language Loyalty, Language Shift and Language Adoption Among Minority Communities Dominance of one language is not the only factor that a migrant community or a minority community faces. There are various social and economic issues that are interlinked to motivate a community to shift to another language or change its loyalty to one particular language. We cannot go into the details of these issues here for lack of space, but would nonetheless like to draw the attention of our readers to some glaring facts that it is necessary to know before commencing any linguistic fieldwork. 1.4.5.1 Pattern of Language Shift and Language Retention Post-independent India has witnessed many language movements for political and financial gains. It has also seen language revivalism and revitalisation for the sake of establishing an independent identity despite the multilingual forces motivating speakers of minority languages to shift to major languages. The two forces, retention and shift, coexist within the same language group. Language shift, although a common phenomenon among the minority communities, does not have a uniform pattern all over India. It varies from language family to language family and from region to region. While tribals of the Munda family and those of the Dravidian family are seen to shift easily, a considerable resistance is seen among the tribals of the Tibeto-Burman family and those of the Mon Khmer, i.e. Khasi (see table 1.4) Similarly, the division between rural and urban areas also plays a role in language retention and shift. It has been revealed by a recent study made by M. Ishtiaq (1999) that more than 90% of Mundas of Ranchi city have declared Sadari as their mothertongue, while 100% of rural Mundas, especially in Kula Buru and Litingdish villages (Bundu Block) and Bande village (Murhu Block), have declared their traditional language Mundari as their mother-tongue in the absence of a regional dominant language. Our field study of Ranchi City also established that almost 90% of Kurux speakers declare Sadari as their mother-tongue. A very low percentage of urban tribal or non-tribal groups are monolingual in their ancestral languages. Ethnolinguistic minority status induces a negative attitude 6

See note 4.

18 A Manual of Linguistic Fieldwork and Structures of Indian Languages

towards language loyalty. A gradual adoption of a non-tribal language as a mothertongue presents a classic case of language shift. Theoretically, these symptoms are diagnostic of potential language death. It is to be noted that all these tribes are exposed to some variant form of Hindi. One can see the percentage of shift among the tribals of the five language families in various states in table 1.4 given below (adapted from M. Ishtiaq 1999). The People of India project POI) undertaken by Singh and Manoharan also indicates similar findings. South India has been known to show the least resistance to shift while Northeast has shown the maximum. Sikkim, Bengal and Bihar are known by the presence of peaceful coexistence of tribal and non- tribal languages. Urban migrant groups of major languages such as Gujarati, Marathi, Tamil etc. maintain their mother-tongues at home as they do not suffer from an inferiority complex. However, as stated earlier, they are not monolingual in their mothertongues and their competence in the dominant language of the region is at least functional, if not high. Table 1.4

Language shift among tribals in various states

Less than 1% 1-10% 10-20% 20-40% 40-80% More than 80% 1.4.5.2

Sikkim, Meghalaya, Arunachal Pradesh, Manipur, Mizoram, Nagaland Andaman and Nicobar Jammu and Kashmir Bihar Himachal Pradesh, Rajasthan, W.Bengal, Assam, Madhya Pradesh, Maharashtra, Orissa, Karnataka, Andhra Pradesh Gujarat, Kerala, Tamilnadu, Lakshadweep, Uttar Pradesh, Goa, Deu, Daman

Reasons for Language Shift and Retention

The following reasons are very significant in causing language shift among speakers of minority languages in Indian situation. (a)

Inferiority complex of the mother tongue speakers.

(b)

Economic benefits of learning the dominant language (DL), i.e. acquiring better job prospects.

(c)

Facility of learning DL through free schooling in the language concerned.

Conversely, the high percentage of language retention of minority languages can be ascribed to: (a)

Multilingualism of coexisting and not of competing nature.

(b)

Lack of educational and other opportunities.

(c)

Self-sufficient economic structures.

(d)

Absence of dominance of a particular regional language

Tibeto-Burman languages are retained in large number primarily because of lack of a dominant pressure of one particular language in the region and because of

1. Linguistic Fieldwork and the Indian Scene 19

the existence of nearly self-sufficient economic structure, the two factors missing in the Jharkhand area. Hence the difference between the two types of patterns of language maintenance. 1.4.5.3

The Paradox

The people of minority communities live in a constant state of bi-/multilingualism, with the result that these communities are always at a higher level on the ‘scale of bilingual proficiency’ than their dominant majority communities. The contactinduced changes in such minority languages lead to typological homogeneity on the one hand, and to a tendency of language attrition on the other. In the tug of war between language maintenance (retention of ancestral languages) and language proficiency in the dominant/contact language/s, many languages have begun to pass through a transition period of language change and language convergence, postponing or avoiding the expected language obsolescence situation. The paradoxical oscillation between language maintenance and language loss or shift is more visible in urban areas than in rural areas. Perhaps the motivation to shift languages is stronger in the urban areas because of the extensive contact and the prestige attached to the contact language/s (such as Sadari or Hindi in Jharkhand). In the context of such omnipresent multilingualism Indian languages have witnessed various stages of convergence from highly code-mixed stages to highly converged stages. Long and stable multilingualism over generations has been responsible for the genesis of ‘areal features’ in languages. It is utterly necessary to familiarise oneself with these areal features before undertaking any field trip, as advance knowledge of such features will prepare a field linguist about what to expect. We shall discuss these features in some detail in the next chapter. We shall also discuss individual characteristic features of each language family so as to facilitate the fieldworker to identify the direction and intensity of borrowing as well as of convergence and also perhaps to locate the source of the two. 1.5.

The Lingua Franca or Contact Language/s of India

When several languages and dialects are spoken in an area by diverse groups of people, it is an appropriate expectation to find one or more than one contact languages (one among non-educated people and another one among educated people in city environments) to emerge. India with its multicoloured mosaic of multilingualism offers several types of contact languages, each inviting an independent study. The case in point is the emergence of Sadari in the Jharkhand, Nagamese, Nefamese, and Chakesang in the Northeast and non-standard varieties of Contact Hindi across the country. Needless to say, such contact and converged languages will open new vistas for the theory of languages in contact. Readers are recommended to see maps exhibiting various languages used as lingua franca by Bertil Tikkanen (1996). These maps are good indicators of the degree of bilingualism in each region, too. This leads to the question whether there is a contact language used all over India for inter-community communication. We would attempt to answer this in the following section. 1.5.1 All India The study by Singh and Manoharan (1993) reports that a substantial portion of the Indian population uses Hindi (the second in the list is Marwari [Rajasthani]) and its

20 A Manual of Linguistic Fieldwork and Structures of Indian Languages

various forms for inter-group communication. This is true of both rural and urban communities. If we add the people using Urdu to the Hindi-speaking population (as, at the spoken level, the two languages do not differ at all), we reach the conclusion that more than 57% of the Indian population uses Hindi/Urdu for inter-group comunication (see review of Singh et al by Abbi 1995). It must be noted that Hindi is used as a cover term for conglomerates of mutually intelligible languages and language varieties (dialects). As many as 49 language varieties are subsumed under Hindi (see Table 1.5). We would like to add to this list Dakhini Hindi which is a lingua franca in Hyderabad and surrounding areas. Urdu is considered a separate language, but for all practical and linguistic purposes is closer to Hindi in its Khari Boli form than any other variety. In fact, it is the same language, as no one would be bi-dialectal or bilingual in Hindi and Urdu (other than educated elites who write in two different scripts and would assert the distinction by using different sets of lexical items.). Many Hindi speakers of rural areas understand and use two different styles of speech: one is the local dialect, typically used at home and in the local community, and the other is the more universal Khari Boli style to be used for inter-community communication. (See appendix 14). Refer also to Khubchandani (1991) under ‘Prominent Contact languages‘ to get an overview of the contact languages used in 1961 in various states and Union territories. The situation since then has changed. The latest situation is best described by Tikkanen (1996) in maps. Table 1.5 1991 Census. Varieties of Hindi Languages and mother-tongues grouped under Hindi HINDI 1 Awadhi 2 Bagheii/Baghelkhandi 3 Bagri [Rajasthani] 4 Banjari 5 Bharmauri/Gaddi 6 Bhojpuri 7 Braj Bhasha 8 Bundeli/Bundelkhandi 9 Chambeali 10 Chhattisgarhi 11 Churahi 12 Dhundhari 13 Garhwalii 14 Harauti 15 Haryanvi 16 Hindi 17 Jaunsari 18 Kangri 19 Khairari 20 Khortha/Khotta 21 Kulvi 22 Kumauni 23 Kurmali Thar 24 Labani

Number of Speakers 337,272,114 481,316 1,387,160 593,730 887,632 18,919 23,102,050 85,230 1,657,473 63,408 10,595,199 45,107 965,006 1,872,578 1,235,252 362,476 233,432,285 96,995 487,999 14,307 1,049,655 152,442 1,717,191 236,856 13,722

1. Linguistic Fieldwork and the Indian Scene 21

25 Lamani/Lambadi 26 Laria 27 Lodhi 28 Magadhi/Magahi 29 Maithili 30 Malvi 31 Mandeali 32 Marwari 33 Mewari 34 Mewati 35 Nagpuria 36 Nimadi 37 Pahari 38 Panchpargania 39 Pangwali 40 Pawari/Powari 41 Rajasthani 42 Sadan/Sadri 43 Sanori 44 Sirmauri 45 Sondwari 46 Sugali 47 Surgujia 48 Surjapuri Others 1.5.2

2,054,537 64,903 68,145 10,566,842 7,766,597 2,970,103 440,421 4,673,276 2,114,622 102,916 777,738 1,420,051 2,179,832 151,599 14,780 213,874 13,328,581 1,569,066 11,537 18,280 37,958 113,491 1,045,455 370,558 4,642,964

Our field experience

Our recent pilot survey (Abbi et al. 2000)7 reveals that among all the languages listed in the table it is Hindi-Urdu that takes the credit of serving as the lingua franca among the cities and towns of the country. Our fieldworks in remote areas of Kashmir, Northeast as in Manipur, or in Meghalaya, confirm that Hindi is understood and used by various community members for intra-community and especially for inter-community communication (Abbi et al. 2000). This Hindi may not be the ‘standard’ prescribed Hindi, but has the characteristic features of the local dominant language/s. In other words, if there is any contact language that India’s non-elite masses use for inter-group communication, it is Hindi in its various forms. The case of the Southern part of India is a little different. In big cities such as Hyderabad, Bangalore, and Mysore people understand and use Hindi; however, in the interior it is not used, though a large section of the population can understand it. The credit of this ‘spread’ of Hindi goes to Hindi films which are watched all over the country with great interest. I must caution the fieldworker that in Tamilnadu one faces a total rejection of the use of Hindi, although I have noticed that in the cosmopolitan city of Chennai educated shopkeepers, traders and businessmen not only understand Hindi, but can also use it quite proficiently. Our fieldwork across the length and the breadth 7

A Sociolinguistic Enquiry into the Acceptance Level of Hindi as a Pan Indian language: A Pilot Survey. Anvita Abbi, R.S. Gupta and R.Gargesh. 2000. Project funded by the ICSSR.

22 A Manual of Linguistic Fieldwork and Structures of Indian Languages

of this country has exposed one fact very blatantly: if one knows Hindi, one can visit any urban city of India. As we shall see later, this fact alone always prompts a field linguist to begin asking her/his questions in Hindi (if the regional dominant language is not known to her/him). In fact, we always make our questionnaires in Hindi. We shall talk about questionnaire framing in chapter four. In this context the Andamanese Hindi-Urdu must be mentioned, a widely used lingua franca of the region which consists of Indo-Aryan and Dravidian structures8. The enthusiasm shown by adults promoting the learning of Hindi among their children reminds me of an anecdote my student reported on her return from the field trip to Jharkhand area. On asking a Kurux woman why she discouraged her children from learning Kurux, she replied agitatedly “agar sirf Kurux sikhaate, aur baad men baccon ko naukri nahiin miltii to kyaa aap log dilli se aakar unhen khaanaa khilaatiin? Hindi aur Angrezi kii roti se vaastaa hai, isliye sikhaate hain. jahaan unkii zaruurat nahiin wahaan log ab bhii Kurux bolte hain. dehaat men jaakar dekh liijiye.” (If I had taught them only Kurux, and consequently they would have remained unemployed, would you then have come all the way from Delhi to feed them? Hindi and English are related to matters of food, that is why we teach these languages. Where these are not needed, people still speak Kurux - go and see in the villages.) (Reported by Priyanka Bhattacharya, 1998) Having examined the linguistic and social status of the languages of India, the nature of bilingualism and varying patterns of language loyalty it becomes imperative to consider the linguistic features of the major language families and significant areal features of Indian languages. We shall undertake this task in the next chapter. Table 1.3 DEV-Devanagri, GUJ-Gujarati, PUN-Punjabi, BEN-Bengali, ORI-Oriya, TEL-Telugu, KAN-Kannada, TAM-Tamil, MAL-Malayalam, SIN-Sinhalese, URD-Urdu, SIND-Sindhi.

8

Personal observation. This is one area that warrants an in-depth research.

1. Linguistic Fieldwork and the Indian Scene 23

24 A Manual of Linguistic Fieldwork and Structures of Indian Languages

24

A Manual of Linguistic Fieldwork and Structures of Indian Languages

Chapter II

Language Families, Language Contact, And Areal Universals Indian languages have always been a challenge to linguists for their diversity and multiplicity. In this chapter, we want to concentrate on the characteristic features of each language family of India and then briefly enumerate the areal features prevalent in these languages. We shall attempt to draw the attention of the readers to the omnipresent language contact situation due to heavy bilingualism that we discussed in the last chapter. The chapter ends with an exposition of how the contact situation has given rise to conflicting grammatical patterns and converged languages. You might ask why this in a field manual? Well, we believe that in order to do justice to the data that one collects in the field, and in order to understand why one sometimes finds aberrant patterns during fieldwork, one should be aware of the aspects of Indian languages that we discuss in the following pages. We are trying to prepare you for what you might expect in the field; although of course fieldwork always offers surprises, and perhaps that is the excitement one looks for in the study of linguistics. 2.1

Individual Language Features

India represents five language families, all very distinct from one another. These are: 1. 2. 3. 4. 5.

Indo - Aryan Dravidian Austro-Asiatic Tibeto-Burman Andamanese

These five distinct language families are distinguished on a linguistic basis and each family has been studied extensively except the Andamanese. While IndoAryan languages are highly inflecting, Dravidian languages are agglutinative, Munda (a major sub-branch of Austro-Asiatic) is of the incorporating and agglutinative type and Tibeto-Burman is of the isolating and agglutinative type. One cannot, however, compartmentalise them in so rigidly. Let us consider each of these language families in some detail. We shall first describe the geographical location and spread of languages belonging to each of these language families (see also map 1.2 given earlier in chapter 1). Having done that we shall consider the characteristic features of each language family under three major topics, i.e. (1) the sound system, (2) the lexical and morphological system, and (3) the syntactico-semantic system. We have tried to be brief and have restricted the discussion to the main defining features of each language family, rather than discussing in detail individual languages.

2. Language Families, Language Contact, and Areal Universals

2.1.1

The Indo Aryan Languages

2.1.1.1 The Geographical Spread: Indo-Aryan languages (IA) occupy the largest geographical area in the subcontinent. It stretches from Pakistan in the west, to Bangladesh in the east, from Nepal in the north and to Sri Lanka in the south. A chain of intelligibility can be established from the westernmost tip of India to the easternmost tips of Bangladesh, not, however, from the northern tip of Kashmir to the southernmost tip of Maharashtra. To the west of the IA belt lies the Iranian language family, its sister family from which it separated around 2000 BC. Aryan movement is generally believed to have taken place in successive waves between the period of 1700-2000 BC. Sinhalese, an IA language, dominates the region of Sri Lanka in the south of the Indian continent. 2.1.1.2 The Sound System: The Indo-Aryan vowel system is as varied as the number of languages in the family. Masica (1991: 109-113) discusses the range from the basic five vowel system of Roma (found in Europe) and a six vowel system of Oriya /i, e, a, u, o,!/ to an “authentic thirteen vowel system” found in Sinhalese, i.e., /i:, i, e:, e, E:, E, a:, a, «, o:, o, u:, u/. Konkani seems to have the largest number of vowel contrast based on quantity and quality, i.e., fourteen in all. The defining feature of these languages is that the vowel contrast is maintained more on the level of quality than on quantity. Thus, Hindi and various other languages in the Hindi-speaking area (see appendix 14) distinguish vowels on the scale of height and fronting, not according to the length of the vowel. Thus it is common to find /I, i, U, u, e, E, o, !, «, a/ – a ten-vowel set with length accompanying high vowels interconsonantally in a syllable. However, languages adjacent to Dravidian languages (see below) have incorporated length as the distinguishing phenomenon. Another characteristic feature of modern IA languages is the prevalence of nasalised vowels. Languages such as Hindi and its dialects offer a complete parallel set of nasalised vowels, though the phonemic status of each of them is in dispute (Kaye 1997). Nasalisation in IA languages is predictable in some positions and thus is allophonic, whereas in some positions it is not and thus may be regarded as phonemic. For instance, the plural distinction in Hindi is maintained by nasalisation. Thus: hE ‘is’ and hEâ ‘are’ or ayegi ‘she will come [fsg]’ and ayeâgi ‘they will come [fpl]’. Indo-Aryan consonant system is identifiable by the aspirated stops prevalent in all the languages. Most of the languages offer four-way contrast in stops, distinguishing voiceless and voiced unaspirated series from voiced and voiceless aspirated series, i.e. /p, pH, b, bH, t, tH, d, dH, ÿ, ÿh, ê, êH, c, cH, j, jH, k, kH, g, gH /. One should, if possible, consult Ladefoged (2001: 123-25) and listen to the four-way contrast recorded in the accompanying CD. There are languages such as Marwari, Sindhi, Gujarati, Marathi, and Nepali that offer aspiration contrast in nasals, too: /m, mH, n, nH /. Some dialects of Hindi and Rajasthani offer aspiration contrast in laterals, as well: / l, lH /. For fieldworkers the feature ‘aspiration’ can serve as a strong feature

25

26

A Manual of Linguistic Fieldwork and Structures of Indian Languages to identify a language as Indo-Aryan or as having been a language being influenced by it. Combinations of vowel sounds within a syllable, i.e. dipthongs, are mostly being replaced by a combination of a vowel and a glide y or w, e.g. cay ‘tea’ in most of the IA languages. 2.1.1.3

The Lexical and Morphological system

The Indo-Aryan languages have derived their lexicon from three main sources. The foremost source is Sanskrit, and these lexical items are generally termed as tatsama words. These words are used in their original Sanskrit forms with morphological endings according to each language. The second source is traditionally known as tadbhava, i.e. the words which were originally derived from Sanskrit but have gone through phonological changes over a period of historical development via the period of Prakrit and Apbhramsa. Thus Hindi hatH ‘hand’ is derived by various phonological rules (and is tadbhava) from Sanskrit h«st« ‘hand’. Then there are words called des&aj ‘local’, the etymology of which cannot be traced to Sanskrit origins. The fourth type are vides&i, ‘foreign’ words that are borrowed from Persian, Arabic, French, Portuguese, and English. Although this is a traditional way of classifying IA languages, it proves useful to have this knowledge while conducting field work on a new language, as the rules for phonology and morphology operate differently in each case. For instance, see Ohala 1983 for the schwa deletion rule and its applicational constraints in Hindi. As far as word formation is concerned, a large portion of the IA vocabulary is derived by expressive morphology and reduplication of words; especially those derived as echo formations. A detailed discussion on these can be found in chapter 5. Gender in Indo-Aryan languages is both a syntactic and a morphological category. Thus, in languages such as Hindi and Punjabi gender assignment of a noun controls the agreement feature in the rest of the sentence, resulting in each modifier inflecting for the gender. According to Masica (1991: 219): “In the New Indo-Aryan languages that have it, gender is an inherent and classificatory property of one class of words (nouns) and a variable or inflectional property of others (adjectives, certain verbal forms, sometimes pronouns and adverbs, and one extremely important postposition1).” On the other hand, in languages such as Bangla some nouns designating professions are classified as masculine and feminine without any syntactic implications. Gender in eastern Indo-Aryan languages has almost disappeared. Languages such as Konkani and Marathi have preserved the old Sanskrit distinction between masculine, feminine and neuter. The common pattern in the rest of the Indo-Aryan languages, however is the two-gender system where neuter merges with masculine and feminine. The other major defining characteristic feature of IA morphology is that the languages in this family are highly inflecting in nature, leading to a large number of portmanteau morphs. Thus in the Hindi form jaeâgi ‘they [f] will go’ ja- is the root, i.e. genitive such as in Hindi: ram ki ÿopi ‘ Ram’s cap [fsg]’ but ram ka gH«r ‘Ram’s house [msg]’.

1

2. Language Families, Language Contact, and Areal Universals

27

e- indicates 3rd person, nasalisation indicates plural, -g- is future tense marker and –iindicates feminine gender; in the word l«}ke ‘boys’ /l«}«k/ is the root; –e indicates both number and case, [in this case plurality and nominative or oblique]. We will discuss this feature in detail when we discuss word formation processes (Chapter 5). Agreement is an important feature of this language family, although not all IA languages show it in a uniform pattern. The languages that show agreement prominently are Western Indo-Aryan languages. The agreement is exhibited in a linear fashion, so that the gender and number of the modified noun is encoded in the modifiers and the finite verb, as for example in Punjabi: c«Ngiyaâ ku}iyaâ ‘good girls’, c«Nge mu÷êe ‘good boys’. In Hindi however, while verbs are encoded with the gender and number information of the argument, modifiers encode only gender but not number if the controller is feminine. If the controller is masculine, the modifiers encode number in addition to gender, e.g. «ccHe l«}ke ‘good boys’ but «ccHi l«}ki/l«}kiyaâ ‘good girl/girls’ Another feature which distinguishes languages of this family from others, is that nouns and pronouns change into the oblique form before taking any case marking. Thus, the Hindi word for boy l«}ka changes its form to l«}ke in the oblique form, as in l«}ke-ne ‘boy, agentive’. Pronouns of the first, second and third persons also change their form to the oblique case. Konkani offers the largest number of forms as the language has retained the old Sanskrit pattern of three genders. Consider a small sample of the agentive case (Table 2.1). Table 2.1 Konkani III Person Pronouns in Agentive Case Singular ta-÷e ti-÷i tEâ-÷E

Masculine Feminine Neuter

Plural tE-÷e ta-÷i tiâ-÷i

The second person non-honorific pronoun in Bangani tu changes to taâu in oblique case before taking any postposition. Indo-Aryan languages, as stated above, are highly inflecting. Interestingly, various discourse-related features, such as the status of the addresser and the addressee, are incorporated in the verbal morphology of some languages (see the example of Magahi given in Chapter 4 [table 4.3] and the example for Maithili given in chapter 6 [example 6.138]). A small set of 3rd person genitive pronouns is given below to draw attention to the fact that deictic elements are a rather a complex phenomenon to study in Indo-Aryan languages. Table 2.2 Maithili Possessive III Person Pronouns (Adapted from Ramawatar Yadav 1996:121) III person

HONORIFIC PROXIMATE hin-k-«r

Singular

REMOTE hun-k-«r

Plural

hun-k-«r s«bH-«k

hin-k-«r s«bH-«k

NON-HONORIFIC REMOTE PROXIMATE ok-«r ek-«r ok-«r s«bH-«k

ek-«r-s«bH-«k

28

A Manual of Linguistic Fieldwork and Structures of Indian Languages

A very significant Indo-Aryan feature is that languages derive transitive and causative forms by suffixing a long vowel -a: and – va: (or its cognate where vowel and consonant interchange, thus Bhojpuri –a:v) to an intransitive or transitive root, respectively. . Thus, Hindi-Urdu likH ‘write’ > likH + a: -‘ to make someone write’ > likH-va:- ‘to make someone make someone write’. In the scientific literature, the first causative with long vowel has been termed direct causative, contactive causative or causative 1 by various scholars (see Masica 1976 for a detailed discussion), and the latter form has been called indirect causative, distant causative or causative 2. Since the causative morpheme is an argument-increasing morpheme, an additional argument in the second causative distances the actor from the instigator, hence the terms ‘distant causative‘ or ‘indirect causative’. There are languages, however, that do not make a clear-cut distinction between direct and indirect causatives, and therefore an additional argument would not trigger a distinct morphological marking. In other words, the semantic distinction of direct and indirect causation is not overtly maintained by the morphology of the causative. Consider these sentences from Bangani: 2.1(a) auâ tEs!

!âs-a:u

I he-acc laugh-caus ‘I made him laugh’ (b) auâ guruji-kei tEs! !âs-a:u I teacher-instr he-acc laugh-caus ‘I made the teacher make him laugh’ Typically, Indo-Aryan languages use prefixes and suffixes, the former being used more in word forms of the noun and adjective class than in verbs. The affixes are freely borrowed from Sanskrit and Perso-Arabic (PA) stock. Thus Hindi-Urdu kar«÷ ‘cause’ but «-(Skt) kar«÷ ‘without cause’, p«}Ha ‘educated’ but b - (PA) p«}Ha ‘uneducated’. 2.1.1.4

The Syntactico-Semantic System

Languages of the IA language family exhibit SOV word order. Therefore, in a prototypical case modifiers precede the modified, and adjuncts precede verbs. Kashmiri (see Hook 1976) is an exception and Bangani also prefers verbs in the second position (see Abbi 2000), although verb-final sentences are also used. We shall consider the phenomenon of word order in chapter 6, where we mainly discuss syntactic issues. Although an SOV pattern is maintained, word order is comparatively free in IA languages, and movement of elements is used for strategies such as topic and focus, foregrounding and backgrounding information. Though the verb is an essential component, yet it is not surprising to find verb-less sentences in Eastern Indo-Aryan languages such as Oriya, Sampbalpuri, Bangla, and Assamese. Thus in Bangla: 2.2 kaal ke or kaaj tomorrow-acc he-dat work

f f

2. Language Families, Language Contact, and Areal Universals ‘he has work tomorrow’ A large number of languages are split ergative (a feature they share with Tibeto-Burman languages), i.e. show ergative features in some tensed-marked sentences, such as in the past tense in Hindi. Eastern Indo-Aryan languages are not ergative in nature. For a detailed discussion on ergativity please see section 5.1.7.2 and 5.3.1. Another interesting feature of IA languages is that experiential constructions taking typical psychological predicates mark their experiencer nouns with the dative or any other case. These constructions are popular in linguistic literature under the name of ‘dative subject’, as the dative-marked experiencer noun serves as the subject of the sentence. For details see chapter 6. The following sentence from Hindi will illustrate the issue for the time being. 2.3

radha –ko

bHukH l«gi hE

Radha-dat hunger strike aux ‘Radha is hungry’ A typical feature of Indo-Aryan is that negation is always preverbal. Languages in this family generally offer two types of negation, one prohibitive and another for simply negating the assertion. Thus Hindi m«t jao ‘don’t go’ while vo n«hiâ g«ya ‘he did not go’ type of constructions can be found in almost all IA languages. We shall discuss negation at length in all families in chapter 6. There are other constructions in the languages of this family that employ what is known as ‘conjunct verbs’. The conjunct verb is constituted of two elements, the first of which is generally an adjective or a noun, while the second element is the verb ‘to do’, ‘to be’, or ‘to happen’. This is a very productive way of forming verbs in these languages - so much so, that a noun or adjective can be borrowed from a foreign language and the indigenous verb ‘to do’ suffixed to it. For instance Hindi fon k«rna ‘to telephone’, or k«rrekt k«rna ‘to correct’. Typically, many languages take a dative subject with conjunct verbs. Finally, the IA languages typically have relative-correlative constructions, where the modifying clause marked by the relative marker is followed by the main clause preceded by a correlative. Consider the Hindi sentence given below. 2.4

jo b«cca baah«r kHel r«haa hE vo mere bHai ka hE Rel child outside play prog. Aux 3msg Corel my brother-gen Aux 3msg ‘ The child who is playing outside is my brother’s.

2.1.2

The Dravidian Languages

2.1.2.1 The Geographical Spread Dravidian languages occupy the southern part of India and a little pocket in West Pakistan inhabited by North Dravidian speakers called Brahui. In fact, this pocket offers a clue to the directionality of migration of the Dravidian speakers in the

29

30

A Manual of Linguistic Fieldwork and Structures of Indian Languages past, i.e. from the westernmost part of South Asia to the southernmost part of the continent. Two of the other North Dravidian languages, namely Kurux and Malto, are spoken in the Chotanagpur area of southern Bihar. These little pockets of Dravidian languages in the north and the central-east of India indicate that perhaps before the Aryans dominated the region, the entire north and central India was inhabited by the Dravidian speakers along with other tribes of the region. Dravidian languages, such as Tamil, are spoken in Sri Lanka, too. See map 1.2. 2.1.2.2

The Sound System

The feature of vowel length characterises the Dravidian languages. The basic fivevowel set exists in pairs of short and long duration in almost all of the languages of this family /a, a:, i, i:, u, u:, e, e:, o, o:/. As Steever (1998: 17) puts it: “The prosodic structure of Proto-Dravidian is quantitative rather than qualitative”. This situation still exists today in modern Dravidian languages. For instance, Tamil (Annamalai and Steever 1998: 102) makes a phonemic distinction between the long and short /«/ in addition to the phonemic quantitative distinction maintained in the standard fivevowel set. Among the consonant series, the Dravidian language family is marked by the large range of point of articulation for stops. Thus, Toda and Malayalam exhibit a six-way contrast in point of articulation, a rather rare feature in language typology. Malayalam has /p, t, t, ÿ, c, k/ as distinct phonemes. Nasal stops are also distinguished phonemically on the basis of six points of articulation, e.g. /m, n, n, ÷, ø, N/. Aspirates are absent in Malayalam, however, Telugu (Krishnamurti 1998: 206) and Kannada (Steever 1998: 130) have borrowed voiced and voiceless aspirates from Sanskrit and are used in educated speech. Retroflexes are the hallmark of these languages. Lateral /ñ/ and nasal /÷/, stop series / ÿ, ê, ÿH / as well as approximants / " /. are very common, especially in Tamil. North Dravidian languages exhibit velar fricative /x/ and glottal stop /// in their phonemic inventory. 2.1.2.3 The Lexical and Morphological System Although the Dravidian family is independent and has been present in the Indian subcontinent for a much longer period of time than Indo-Aryan, the word stock of languages such as Telugu and Malayalam is full of Sanskrit tatsama words. Fieldworkers should be aware of the fact if working on a new language closely related to these two languages. Dravidian morphology is agglutinative and suffixal in nature. The identifiable feature of Dravidian morphology is the neat fashion of words/morphemes strung together in a linear sequence. Morphemes appear in this order: lexical root + derivative suffix (optional) + inflectional suffix. Thus Kannada mara-d-alli ‘on the tree’ is constituted of the noun root ‘tree’ + derivative suffix –d + locative suffix – alli. Linguists generally maintain that prototypically, Dravidian languages exhibit only two word classes, nouns and verbs. The lack of the adverb category is immediately visible when one tries to elicit negative sentences. Negation is conveyed

2. Language Families, Language Contact, and Areal Universals

31

through verbal morphosyntax, i.e. verbs have two forms, positive and negative. Hence, the finite verbs designate two kinds of tenses: past and non-past and two kinds of proposition: affirmative and negative. We shall consider this in detail in chapter 6. Another interesting aspect of Dravidian languages is that they offer two types of interrogative pronouns (Asher and Kumari 1997), general and selective. “The selective type is used to ask a question about ‘which one’ of a specified group or entities, as contrasted with a general ‘who’ or ‘what’” (ibid: 267). This, too, will be discussed in chapter 6. Dravidian languages typically use postpositions to express semanticosyntactic relations between verbs and nouns rather than case markers. Consider the following sentence from Tamil (Ananmalai and Steever 1998: 107): 2.5

na:n avan- ukku anta.p pustakatt-ai.k koÿu-tt-e:n I-nom that man-dat that book-acc give-pst-1sg ‘ I gave the book to him’

Gender assignment is another characteristic feature that distinguishes these languages from Indo-Aryan. The basic division is between human and non-human and within the former between masculine and feminine. In some languages, such as Telugu, the gender assignment of a young girl-child coincides with non-human inanimate nouns. Dravidian languages generally exhibit addressee components marked in the first person plural form. This feature is one of the defining features of the Dravidian languages. Kurux and Malto, the two North Dravidian languages surrounded by Munda and Indo-Aryan languages, still preserve the feature. Let us consider the following example from Telugu (Krishnamurti 1998: 211). Table 2. 3. Telugu Personal pronouns

Pronouns

Singular

Polite Singular

Plural

First person

ne:nu ‘I’

memu (excluding the speaker ) manamu (including the speaker)

Second person

ni:wu/nuwwu

mi:ru

‘you Third masculine

wa:êu ‘he’

atanu, a:yana,

wa:ññu (human)

wa:ru Third non masculine

adi ‘s/he’, ‘it’

a:me, a:wiêa, wa:ru

awi (non-human)

32

A Manual of Linguistic Fieldwork and Structures of Indian Languages 2.1.2.4 The Syntactico-Semantic System Let’s start with the following statement by Zvelebil (1990: 39): There exists no doubt a syntactic pattern typical of Dravidian in a very general and deep sense, but it is very difficult to determine how much of it is inherited, and how much is influenced in the processes of diffusion and borrowing. The only Dravidian languages which vary significantly from the overall pattern (such as Kurux or Brahui), are clearly under the impact of Indo-Aryan or Iranian. We shall nonetheless try to discuss those features in Dravidian that can be easily traced to Proto-Dravidian and are still present in Modern Dravidian languages. Dravidian languages also follow the SOV word order and all the implicational universals emanating from it. However, as we said earlier, most of the scholars believe that the modifier category does not exist as a separate word class. Therefore, the right-most element of the sentence, i.e. the verb, encodes information of adjectival and adverbial nature as well. Consider a sample from Tamil: 2.6.

neNN« v«nta oru m«ntri yesterday come-prt one minister ‘ The minister who came yesterday’

2.7

cikki r«maka hurry to become ‘Hurry up’ or ‘fast’, or ‘hurry’

All Dravidian languages are identified as employing a series of non-finite verb forms followed by one finite form at the end of a sentence. The finite verb form is encoded with subject–object agreement, tense, aspect and mood as well as with a conditional suffix that identifies the protasis (see Tamil sentence 2.9 given below), or, in a clause, the conditional proposition. This is in addition to the negative polarity that we discussed above, which shows up in two alternate forms of verbs. The feature that distinguishes these languages from the rest of the Indian languages is their complex verb system. Non-finite verb forms typically do not mark person agreement, feature but may mark tense and polarity (Krishnamurti 1998:220). This may be seen in compound and conjunct verbs (converbs). Thus in Telugu the conjunctive form amm-i ‘having sold’ is perfective while amm-ite: ‘if one sells/sold‘ is a conditional form. Similarly, amm-ina ‘that which sold/is or was sold’ is in the past tense but amm-e: ‘that which sells’ is in non-past tense. The absence of a separate lexical category for conjunction and the possibility of several verbal units being coordinated in Dravidian, but not in Indo-Aryan, is another identifying feature of these languages. Consider the Tamil sentence (Annamalai et al 1998:120). 2.8.

mazai pey-tu veyil aÿi-ttu va:navil to:nr –i.y-atu rain-nom pour-CP sunshine-nom beat-CP rainbow appear-pst-3sg ‘It rained, the sun shone, and a rainbow appeared’.

2. Language Families, Language Contact, and Areal Universals 2.9

avan aNke po:n-a:l na:n avan-ai canti-pp-e:n that man-nom there go-cond I-nom that-man acc meet-fut-1sg ‘If he goes there, I will meet him’

Another interesting syntactic aspect of these languages is that they employ several embedded sentences in participial form [sentence 2.6] to modify a nominal category, rather than using relative–correlative structures typical of Indo-Aryan languages. We shall give one simple example from Tamil (Annamalai et al 1998: 121) to illustrate the point. Readers are advised to consult chapter 6 for details. 2.10

ne:rru

va-nt-a

anta.p paiyan

yesterday come-pst-prt that ‘The boy who came yesterday’

boy-nom

In addition to morphological causatives, the Dravidian languages make ample use of syntactic causatives. Consider a modern Tamil example given below. (ibid ) 2.11.

a:ciriyar ma:÷avan-ai veñiyil nir –ka vai-tt-a:r teacher-nom student-acc outside stand-inf place-pst 3hon ‘The teacher made/caused the student to stand outside’

A very important and widely researched characteristic of the languages of this family is the use of a quotative, the verb ‘to say’, as a complementiser. Consider the following sentence from Tamil (Annamalai et al 1998: 121). 2.12

na:n avan nallavan en-ru ninai-kkir –e:n I-nom that.man- nom good.man-nom say-cp think-prs 1sg ‘ I think that he is a good man.’

Actually, Dravidian languages use “quotative verbs for all questions, thoughts and similar embeddings” (Zvelebil 1990: 44). 2.1.3

The Austro-Asiatic Languages

2.1.3.1 The Geographical Spread: These languages are represented by three major branches, Mon-Khmer, Munda and Nicobarese. Surprisingly, all the languages of this family are grouped under ‘tribal’ languages. Khasi, spoken in the eastern part of hilly India, represents the MonKhmer. Principally tribal groups in various pockets of Central India and Eastern India speak Munda languages (the older term is KOL). The Munda language family is spoken only in India. Nicobarese is spoken on the Nicobar Islands. There is a small pocket of the Munda language ‘Nahali’ in the center of India. Historically, speakers of the Munda branch of the language family are considered the oldest inhabitants of India. It is known that these languages have been present in India since before the arrival of the Aryans and the Dravidians. They certainly once occupied a larger territory in India than they do now. Nicobarese, spoken on the Nicobar Islands of India, is another language that is cut off from the mainland. Some linguists (Grierson 1915, Vol. 1:34) established this language as a real link between Khasi (MonKhmer) and the Munda languages of this family.

33

34

A Manual of Linguistic Fieldwork and Structures of Indian Languages 2.1.3.2

The Sound System

Languages of this family have gone through massive typological changes because of intense and long contact with Dravidian and Indo-Aryan, and thus it is very difficult to isolate features pertaining to the Austro-Asiatic sound system at this stage. However, a couple of features that are easily identified can be given here. The foremost is that there is a strong vowel harmony in all the languages of this family. Thus, in Kharia jib ‘to touch’ > ji-ni-b ‘a touch’, or job ‘ to suck’ > jo-no-b ‘suck’ after the infix /n + vowel/ is inserted. The vowel of the infix is always the copied vowel of the stem. The other characteristic feature of these languages is that they have phonemic occurrence of the glottal stop ///. Thus in Santhali / u:p / ‘hair’ but / /u:p / ‘put’; in Khasi /ka-wa/ / ‘river’ or /ba-pu.« t' / ‘rotten’; and in Kharia /borol-da/ / ‘water’ and /la/-ta/ ‘feels’. Most of the languages have word-final unreleased stop consonants. Thus, the Santhali temporal adverb /j!kkHEt' / ‘during’, or the borrowed Hindi word /enneic'/ ‘dance’; or Khasi / yaid bak-bak'/ ‘walk hurriedly’, or / p«t'/ ‘ again, next time’; or in Kharia /terob'/ ‘gave’ or /muruk'/ ‘very’ are examples of unreleased stops. As a result of contact with both Dravidian and Indo-Aryan, these languages have both quantitative and qualitative distinction among vowels. Therefore, some of the languages, such as Santhali, have 22 distinct vowels, i.e. long and short versions of /I i, e, E, , «, a, !, o, U, u / In some languages, especially in Khasi, the frequency of occurrence of palatal and velar nasals is a striking feature. Khasi also offers various kinds of initial consonant cluster paradigms with the initial velar nasal [N]. 2.1.3.3 The Lexical and Morphological system Austro-Asiatic languages are highly agglutinative in nature (see examples given below). The verb encodes the information on subject (person and number) and in some cases (e.g. in Santhali) object (number and person), too. Thus Santhali /Ege:rked-iø-a / ‘ scold-pst-object marker [me]-definite’, i.e. ‘X scolded me’. Gorum, another Munda language spoken in Orissa and Andhra Pradesh, displays a full range of subject (nominative) and object (accusative) pronominal incorporation in the verb (For details see Arlene Zide 1997: 253-261). Consider the following sentences from Santhali. 2.13

g«yi- e kIri:ø-ket' – kin a cow- subj buy-pstobject (dual) -def ‘(He) bought two cows’

2.14

øel- ked- e a- iø see- pst -object 3sg def- sub 1sg ‘(I) saw him’ Khasi, a Mon-Khmer language of the Austro-Asiatic branch has similar structures:

2. Language Families, Language Contact, and Areal Universals

2.15

ka-kH«nna ka- la - kli«d' –ki-sintiyo fsg-girl

fsg-pst- pick – pl- flower

‘the girl picked the flowers’ According to Norman Zide (1997: 307-8): The incorporation of monosyllabic ‘combining forms’ of nouns (usually objects or instruments) into verb stems goes back to ProtoSouth Munda, and subject-marking enclitic constructions of several kinds (presumably innovated) are found in some of the Munda languages, both North and South. These are probably of comparatively recent origin. We reconstruct subject (verb) prefixes for Proto-South Munda, and probably for Proto-Munda as well. The Munda languages (Western Austro-Asiatic), like the other language families of non-peripheral India, Indo-Aryan and Dravidian, are SOV languages, but Munda shows evidence of the earlier SVO pattern common in Mon-Khmer (Eastern Austro-Asiatic) and perhaps ProtoAustro-Asiatic. Object suffixes on the verb probably go back to ProtoMunda. Unlike the two language families we have discussed so far, Austro-Asiatic languages offer a rich complement of all kinds of affixes, i.e. prefixes, infixes, and suffixes. Consult Table 5.13 given in chapter 5 for examples of infixes in Kharia. Gender is decided mostly on the parameter of animate/non-animate distinction. Khasi, however, has grammatical gender and thus entities are classified either as masculine or feminine in the singular. In the plural, no gender distinction is maintained. There are three numbers in Austro-Asiatic languages (barring Khasi), singular, dual and plural. Consider the Kharia verb paradigm given below (Malhorta 1982:60). Table 2. 4 Kharia Verb êoko- ‘sit’ I Person exclusive I Person inclusive II Person III Person

Singular êoko-ta-ø êoko-ta-m êoko-ta-f

Dual êoko-ta-jar êoko-ta-naN êoko-ta-bar êoko-ta-kiyar

Plural êoko-ta-le êoko-ta-niN êoko-ta-pe êoko-ta-ki/moy

A very prominent feature of Austro-Asiatic languages is the rich stock of personal pronouns distinguished at the level of number, distance between the addresser and the referent and the person-addresser component. Kharia has eleven different pronominal forms, and Santhali is equally richly endowed. Refer to tables 5.6, 5.7, and 5.8 given in chapter 5, and table 6.2 given in chapter 6 for some of the forms in Santhali. Typically, Austro-Asiatic languages, especially Munda languages, lack the dichotomy between morphology and syntax. There exists only a predication that

35

36

A Manual of Linguistic Fieldwork and Structures of Indian Languages incorporates all the information regarding the subject, indirect and direct object nominals. The verb complex thus assumes a very significant position in these languages. However, due to areal pressure these languages have also developed a syntactic structure with identifiable NP + VP mode, although the incorporation of subject or object information in the verb complex is still adhered to. Let us first see the verb complex of Santhali, the widely spoken language of Jharkhand. In Santhali the order of the morphemic elements within a verb complex is the following. S(ubject) M(arker) + Verb + Tense + O(bject) M(arker) + Definitiser. The order of the whole sentence is: S + IO + O + V(erb complex) When working on a new language, the feature of subject and object incorporation into the verb serves a good identifier of these languages. Consider the following sentence from Santhali, where neither the direct object (DO) nor the indirect object (IO) has any case markings. Instead, the last part of the sentence, i.e. the verb complex indicates both subject marker and IO marker (in this case the animate), as the language does not mark the object if it is inanimate. 2.16

i:ø d! ram mit-ten putt -i:ø- Em - ad - e -a 1sg pleonastic Ram one book-Subj- give- pst- IO3sg- def. ‘ I gave a book to Ram’

Please note that the archaic SVO word order of Austro-Asiatic languages is visible in the verb complex. The example from Kharia given below exhibits the redundancy of an independent subject category in the beginning of the sentence, as the person information is coded into the verb complex. 2.17

(aniN) musa hok«}-te gil-teniN (1pl incl) today he-acc hit-tr. pres- 1pl incl ‘ We hit him today’

It is worth mentioning here that Kharia verbal inflection indicates a basic dichotomy between transitive and intransitive roots that is manifest in two distinct sets of tense terminations used for transitive and intransitive verbs, respectively, as given below. (Malhotra 1982: 150). For details consult Biligiri (1965) and Pinnow (1966). Table 2.5

Kharia Intransitive/Transitive Verb Inflection Intransitive

Transitive

Present

-ta

-te

Past

-ki

-o/

Future

-na

-e

2. Language Families, Language Contact, and Areal Universals Another example from Kharia will suffice to demonstrate that a distinction between syntax and morphology is unnecessary, as the complex verb form contains all relevant information about the arguments. 2.18

musa- (aniN)- kH«}ialaN-buN kayom- na niN today- (1pl incl) Kharia –language –with speak-intr fut 1pl incl ‘ Today we will speak in Kharia’

Munda languages are characterised by highly ambivalent word classes. Theoretically, any word or concept can function as the verb base (Pinnow 1960, 1966). Thus, it is not surprising to find a nominal lexical base, or an adjectival lexical base or a quantifier lexical base or interrogatives operating as verbal bases in the language. Pinnow (1966:101) traces this functional ambivalence of word classes to Proto-Munda. Consider the following examples from modern Kharia (for details see Malhotra 1982: 136-139). 2.19

bHagwan lebu-ki ro êel-ki God man-pst and come-pst ‘ God became man and came’

2.20 pampapur-a/ peÿH-hin-te lebu-ki jugHay-ki-moy Pampapur-loc market-in-loc people much-pst-intr-3pl

‘ There was an increase in the number of people in the Pampapu market’. In other words, one of the most prominent features of the Munda languages that has occupied linguists lately is the fact that these languages prototypically do not distinguish between nouns and verbs as distinct word classes. Bhat (1994: 239) discusses the Mundari dictionary by Hoffman (1930) where he finds extensive “homophony” between the two. Consider: buru gapa haÿiN kum}u lutur

‘a mountain, to heap up’ ‘tomorrow, to procrastinate’ ‘a part, to divide’ ‘a thief, to steal’ ‘ear, to listen’

Bhat later elaborates on this issue (1997: 227-251). Hoffman (1930) points out that concrete nouns denoting living beings or inanimate objects, pronouns, numerals, postpositions, adverbs, and interjections can all be used as verbs for denoting states or actions in the language. This is an interesting and defining feature of Munda languages. A very productive and widely used morphological strategy is reduplication of the existing lexicon to derive new word classes. We will talk about this device at length in chapter 5 under section 5.3; therefore, a small set of examples should be sufficient for the present purpose. The following examples from Kharia indicate past participial types to express an accomplished state that can further be used attributively or predicatively.

37

38

A Manual of Linguistic Fieldwork and Structures of Indian Languages

2.20

ga/ ga/ jHula bes um-yo ta tear tear shirt good neg-looks ‘ The torn shirt does not look good’

2.21

hok}a- sob jHula ga/ ga/ aij 3psg-gen all shirt tear tear is ‘All his shirts are torn’

Reduplication of lexical items is used to derive agentive nouns, modifiers of nouns and verbs (i.e. adjectives and adverbs) as well as reflexive and anaphoric pronouns. The most interesting is the type that we saw in 2.20, where an adjectivelike (participial) construction is created by reduplication. We saw that for a similar purpose Indo-Aryan languages would use a relative-correlative structure, while Dravidian language would use a participial construction. Mon- Khmer languages such as Khasi employ the strategy of expressive morphology to derive various kinds of adverbial modifiers. During our fieldwork, we found 59 different adverbs of “crying” derived from expressive morphology. For details see Abbi and Victor (1997: 41-32). 2.1.3.4

The Syntactico-Semantic System

As the discussion above must have made clear, the boundary line between morphology and syntax in the languages of this family is blurred, so that we have already touched upon some of the identifying syntactic features of the Austro-Asiatic family. At this juncture, we would like to draw the attention of our readers to the fact that one language of this family, i.e. Khasi, has still maintained the SVO word order, and the implicational universals contained therein. Thus, the language has prepositions and modifiers following the modified in a typical SVO situation. 2.22

ukH«nna ba-j&r!n -u s!n -hap! msg-boyadj-tall –msg sit- loc ‘ The tall boy sat on the chair’

2.23

balai - pHI-m lEit -yen why- you-fut go – home ‘Why are you going home?’

2.24

Na la-ba:m ya-u-s!/ I pst-eat acc-msg-fruit ‘I have eaten the fruit’

As one can see, Khasi, being a typical SVO language,2 follows the pattern of prefixing the tense and aspect marker to the verb rather than suffixing it. Khasi, 2

However, Nagraja (1997:347-359) reports of variation in the dialects of Khasi that do not strictly adhere to SVO pattern. Instead, VSO, and VOS patterns are observed, and tense marking follows verb forms in some dialects, e.g. Lyngngam and Nongtung.

2. Language Families, Language Contact, and Areal Universals being geographically and culturally distanced from the rest of India also does not show the areal features of South Asian languages. 2.1.4

The Tibeto-Burman Languages

2.1.4.1

The Geographical Spread

The Tibeto-Burman family of languages is a sub-family of Sino-Tibetan for which one could agree with the statement “one of the greatest language families in the whole world—even Indo-European does not have more first language speakers…” (Matisoff 1991:469). As far as India is concerned, these languages mainly exist in the Himalayan ranges stretching from northwestern India to northeastern India and further extending to Nepal, Burma and China. The Tibeto-Burman family has the most numerous and highly differentiated languages in its fold. These languages probably appeared in India later than the Mon-Khmer language, i.e. Khasi of eastern India. They co-existed with Khasi for as long as we know, and once occupied a much larger territory than today. The significant characteristic feature of the languages of this family is that mutual intelligibility among them is almost nil. Perhaps this is also the reason for a high degree of maintenance of the languages. 2.1.4.2

The Sound System

In almost all the languages of this family tone is phonemic, e.g. most of the languages use falling, rising and level tones at lexical level. Tangkhul Naga distinguishes following words only on the tone contour. k«húN

k«huN ‘rotten’;

‘red’;

k«hùN ‘lime’.

Meithei tones have been studied acoustically by Chelliah (1997: 26-48) and the phonemic contrast is established on the basis of rising and falling tones. Consider: í kHóy

‘blood’

ì

‘thatch’ kHòy

‘navel’

‘bee’

lá

‘banana leaf’

là

‘shallow basket’

síN

‘firewood’

sìN

‘ginger’

These are ‘tone prone’ languages, pertaining to a word-tone system with “generally two way contrast of low functional load between higher and lower pitch” (Matisoff 1991: 491). The phonemic distinction between dental nasal [n] and velar nasal [N] is maintained in all positions. Thus in Meithei m«na ‘leaf’: m«Na ‘five’; na ‘ear’: Na ‘fish’; l«n ‘wealth’: l«N ‘thread’ and NraN ‘yesterday’. These languages provide vowel clusters, which can easily be distinguished when encountered for the first time. Thus in Meithei m«un ‘skin’, ain ‘rules, laws’,

39

40

A Manual of Linguistic Fieldwork and Structures of Indian Languages and m«ruoyob« ‘important’. Chelliah (1997:64) mentions 36 vowel sequences in Meithei. Consonant clusters in initial position are rare. Words can be as long as seven syllables such as in Meithei m«ripHut«ram«Na ‘ninety five’. 2.1.4.3 The Lexical and Morphological System Most of the languages of this family are isolating in nature. The following examples from Meithei will demonstrate this: [unbantHa] ‘winter’ / un-pan-tH-a/ ‘ice-rule-month’

[kHokto/e] ‘peel out’ /kHoktHok- l«- e/ ‘peel (out)-happen-perf-assertive

Tangkhul Naga spoken in Nagaland and Rongmei exhibit similar structures. Thus tiN-lum-g«n ‘ rain-hot-time’, i.e. ‘summer’ in Rongmei. These languages are highly affixal and derivations of various word classes such as nouns, adjectives, and adverbs can be achieved by prefixes or sequences of suffixes. Thus, a root such as cà ‘eat’ in Meithei can either be derived as a noun cà-b« ‘ eating’, «-cà-pot ‘food’ [‘with-eat-thing’] or as a verb cà-li ‘eating, eat + prog’. Consider another root k«n with a potential meaning of ‘strong’. «-k«n-b« k«n-b« k«n-n« k«n-li ma-k«lli

‘strong’ (adj) ‘strength’ (noun) ‘fast’ (adv) ‘be strong’ (verb prog) [n>l/-l] ‘he is strong’

These languages in general and Meithei in particular do not distinguish a separate word class of adjectives. Bhat (1994:194) reports that adjectives and verbs occurring in the adnominal position can retain all the tense, aspect, and mood distinctions that they show in their predicative use; and they can also retain all their complements and other adverbial modifiers. Consider one set of examples from Meithei from his work: 2.25

(a)

cay-« saN-b« «y-gi-ni stick-prs long-inf my –gen-be ‘The long stick is mine’

(b)

c«y saN-d«-b« ma-gi-ni stick long neg-inf his-gen-be ‘The stick which is not long is his’

Prototypically, a verb root may take several derivational affixes before it takes an inflectional suffix. Chelliah (1997) discusses basic three-levels-type derivational categories. First-level derivation (eight in all) is that which describes “the extent to which an agent desires or intends to affect some object and the direction and manner in which an action is performed”. Second-level derivation

2. Language Families, Language Contact, and Areal Universals consists of suffixes that have meanings, such as deictic markers, modality, such as negation, and valency increasing potentiality, such as causative. The third level of derivation consists of suffixes that signal aspect and modality (other than negation) (For details see Chelliah 1997: chapter 7). As in Austro-Asiatic languages, the verb occupies the central role and the predication is built upon it by sets of derivation and inflection. It is important to note that the verbal morphology of Meithei and its dialects does not indicate number, gender, person or pronominal agreement between the verb and its arguments. Temporal adverbs are also derived by suffixation to the root. Thus hay-r«-g« ‘say-perf- assocaitive’, i.e. ‘afterwards’ and hay-b«- si-d« ‘say-inf-det-dat’, i.e. ‘therefore’. A characteristic feature of Tibeto-Burman languages is that tense distinction is maintained between future and non-future which is symbolised by an adverb of time. Verbs indicate only aspect. Thus, the following may be ambiguous unless preceded by the appropriate temporal adverb in Meithei. 2.26

N«raN/N«si- «i -sinema -«du -yeN-y yesterday/today I cinema that see ‘I saw/ will see the cinema yesterday/today’

2.1.4.4 The Syntactico-Semantic System Languages of this family are of SOV nature, and at least for Methei we can say that it is a nominative-accusative language. However, if we consider the TibetoBurman family in general, it appears that “notions of subject and object are in fact alien to Sino-Tibetan grammar, as are such grammatical categories as active and passive voice” (Matisoff 1991: 494). Sino-Tibetan languages are considered “topicprominent“ languages. Agreement features of these languages are best summarised by Bickel (2000: 584), which I reproduce here for our readers. The details about each character feature may be found in the references given therein. Table 2.6 Agreement relations in Tibeto-Burman (after Bickel 2000) FN = FV FV as FN FN of FV FN re FV

Identificational Appositional (‘as NP’) Partitional (‘NP of’) Relational (‘NP with regard to’)

This is a summary table of the various agreement relations attested in TibetoBurman. Identificational corresponds to Hale’s (1983) terms ‘merged’ (or ‘argumental’), appositional corresponds to ‘unmerged’ (or ‘predicative’). ‘FN’ and ‘FV’ stands for referential features marked by nominals and verbal agreement markers, respectively.

41

42

A Manual of Linguistic Fieldwork and Structures of Indian Languages The use of a quotative, the verb ‘say’ for complement clauses (as we witnessed in the case of Dravidian) is a common Tibeto-Burman feature. A large sequence of verbs, all grammaticalised except the main predicating one, is another common feature of these languages. As stated above, languages of the Tibeto-Burman family are aspect-prominent, and not tense-prominent languages. Hence, notions like ‘causative’, ‘progression’, ‘duration’, ‘completion’, and ‘inchoation’ are shown by a lexical verb, which is semantically bleached and thus partially or completely grammaticalised. For details see Abbi 1999b. Bhat (2000 :100) calls these languages “mood prominent”. These are also the languages that employ a large number of directional verbs for such a purpose. In this feature these verbs come closer to the explicator-compound-verb function found in all Indian languages. Readers will find an extensive discussion of this in chapter 6. Tibeto-Burman languages are known to have an intricate system of evidential particles that characterise the nature of the speaker’s information (for details see 5.1.7). The pragmatic features such as ‘hearsay’, ‘preconceived knowledge’, ‘first-hand knowledge’, ‘visual vs. auditory information’ and other emotional attitudes are encoded in the verbal morphology. It is a daunting task to analyze verb sequences and their composition in the languages of the Tibeto-Burman family. Other than being a good grammarian, one needs to be sensitive enough to perceive finer shades of emotion and attitudes and above all the ‘world view’ of the speakers of these languages. 2.1.5 2.1.5.1

The Andamanese Languages The Geographical Spread

Spoken in the Andaman Islands in the Bay of Bengal, Andamanese is a generic term used for two distinct groups of the Negrito race that once flourished in the entire Southeast Asia region (Manoharan 1997: 457). Andamanese is also a name of one of the languages of this family. There is a Little Andamanese group speaking Onge, Jarawa and Sentinelese and the Great Andamanese group speaking ten different languages in an equal number of tribes. Very little is known about the languages of this family. Not more than a few hundred speakers remain today. The Andamanese language family is considered the oldest of the Indian language families and has been spoken by the Negritos in the area for last 3000 years. Migration of this group to any other region has not been heard of. 2.1.5.2

The Linguistic Structure

As our knowledge about these languages is very minimal, we shall discuss in general the linguistic structure of only Andamanese and Great Andamanese languages. As far as the sound system is concerned, languages of this family have phonemic vowel length. They also have a front mid-rounded vowel [ë]. The language is rich in velar and palatal nasals [N and ø], respectively. The frequency of occurrence of the voiceless bilabial fricative [¸] is very high. Another common fricative in the language is the velar [x]. The languages offer ample examples of voiced and voiceless retroflexes. Examples of voiceless aspirated sounds [tH] are not uncommon.

2. Language Families, Language Contact, and Areal Universals As far as the lexical morphology is concerned, from the scanty data (all examples are from Manoharan 1997) that is available it seems to be an isolating language which is affixal in nature. Thus, a root can take various kinds of formative affixes to derive nouns and verbs. Thus: /k!w/ k!w k!w-bu k!w-¸ o !t-k!w-bo Ne-rulu-to-k!w-bo

‘bow’ ‘umbrella’ ‘banana fruit’ ‘bark of a tree’ ‘your eye lid’ [you-eye-lid]

The genitive always precedes the names for the body part. Thus:

ÿe-NNe:ÿ ÿe-ba:la: ÿa:-tHa:t Neli-rulu akka:-¸ oN

‘my naval’ ‘my hand’ ‘my tongue’ ‘your (hon) eye’ ‘his/her/its mouth’

The pronominal prefixes have two functions. When they occur before a bodypart word, they signify a genitive construction, but when they occur before a verb, they assume the subject role. An interesting aspect of the language is that pronominal affixes can be classified into three basic classes on the basis of the phonetic shapes. Thus, though there are three variants of the first person singular pronoun “I”, as in /ÿ!-, ÿe-, and ÿa-/, each of them has a restricted set of verbs that it attaches to. Thus ÿ!-otterta:- ‘to tell a story’ but ÿe-øo- ‘to come’ and ÿa-øa:- ‘to eat’. That is, verbs are classified according to the different types of pronominal prefixes that they take. Pronominal prefixes and pronominals maintain the distinction between inclusive vs. exclusive first person singular and plural pronouns. Pronouns of both the second and third person are marked by the honorific. Consult chapter 6, table 6.5. According to Manoharan, the language is highly creolised and multiple affixes for the same function are used simultaneously. “In the process of creolisation the affixes of the different languages spoken earlier are retained resulting in multiplication of the number of affixes” (ibid: 471). 2.2.

India as a Linguistic Area

It is not surprising to find distinct languages of India sharing common linguistic features due to areal pressure. Long and stable language contact accompanied by bilingualism has generated many linguistic structures that are common at the structural as well as the semantic levels across languages of various families. It is a known fact that contact-induced structural convergence of languages from different families within a coherent and contiguous geographic area often results in a linguistic area or sprachbund. The current Indian languages share with each other several

43

44

A Manual of Linguistic Fieldwork and Structures of Indian Languages features at all levels of grammar, i.e. phonological, morphological, syntax, semantics, pragmatics and sociolinguistics. Since the first attempt of identifying India as a linguistic area by Bloch (1934) and subsequently by Emeneau (1956), many linguists have identified features at all levels, demonstrating that despite distinct language structures of various language families we have a fairly large number of shared areal features. Some of the well-researched areal features of Indian languages, diagnostic of geographic and linguistic contact among the major language families in the country, are the following: 1. 2. 3. 4. 5. 6. 7.

Retroflexion (Bloch 1965, Emeneau 1956, Ramanujan &Masica 1969) Morphologically derived causatives (Masica 1976) Reduplication of words (Abbi 1992) Expressive morphology in psychological verbs (Abbi 1992) Explicator compound verbs (Hook 1977; Masica 1976) SOV word order (Masica 1976) Dative/Non Agentive/Experiential subjects (Masica 1976; Verma and Mohanan 1990; Abbi 1994) 8. Conjunctive participles (Masica 1976) 9. Echo formations (Emeneau 1969, Abbi 1992, 1994) The specific references given for each topic should be of help for the field investigator. In addition to the references mentioned above, one should consult Andronov 1964, Kuiper 1967, Vermeer 1969, Southworth 1971, Pandit 1972, Southworth and Apte (edited) 1974, Klaiman 1986, Krishnamurti (edited) 1986, and Abbi (edited) 1991. As the process of convergence involves a simultaneous process of divergence from the archaic structures of the language family, a pattern of typological homogeneity is witnessed in the languages of a linguistic area. As we will discuss areal features at all levels in the following chapters, we will not go into detail here, but would like to briefly enumerate the salient areal features of Indian languages. We shall undertake this task by exhibiting features at phonological, morphological, syntactic and semantic levels. You will find the details in the chapters 4, 5, and 6. A very large number of languages share the following features: 2.2.1

Sound System a) b) c)

2.2.1

Retroflexion Aspiration of voiceless stops Syllable-timed rather than stress-timed

Morphological System a) b) c) d) e)

Suffixal causatives Postpositions Morphological reduplication Echo formations Exclusively suffixing agglutinative morphology

2. Language Families, Language Contact, and Areal Universals

2.2.1

Syntactic System a) b) c) d) e)

2.2.4

OV order of words Conjunctive participles/converbs Explicator compound verbs Syntactic reduplication (aspectual) Dative subject/non-agent experiential constructions

Pragmatic and Sociolinguistic Systems a) b) c) d)

Negation used as a tag for confirmation Three way distinction in 2nd person pronouns Marked +/- honorific feature (in pronoun/verb morphology) Extensive kinship terminology

In addition to the features listed above, there are some ‘areal lexicons’, i.e. words that cannot easily be translated into English, but can be translated very well across Indian languages. We are going to consider some of them in our list of 400 basic words in section 4.5. 2.3

Sub-linguistic Areas

Apart from areal features, there are also those features that identify a micro-area, and it will perhaps be of use to know about these sub-areal features if working in the particular area. We shall briefly enumerate these here for convenience. 1. 2. 3. 4. 5. 6. 7.

Nasalisation (Northern India ) Aspiration (Northern India) Gender agreement (Western India) Right hand is ‘eating hand’ (Central India) Quotative Verb ‘say’ as complementiser (Southern and Northeastern India) Relative-correlative pronoun (Northern India) Classifiers (Eastern and Central India)

We agree with Masica (1992: 38: Like all large linguistic areas with a complex history, South Asia shows many sub-areas of more intense or special convergence, or of partial convergence with other areas. It also has features that place it within larger areal configurations. 2.4

Contact and Convergence

When languages are in contact with each other for a very long period of time, they tend to assimilate several features of languages in contact at all levels of grammar and thus deviate considerably from the characteristic features of their own genetic stock. The Dravidian languages, the Munda languages and the Indo-Aryan languages have all been in such close proximity to each other as to induce changes in each of them to such an extent that one can perhaps identify a ‘core grammar of Indian languages’. We shall in this section probe into the various linguistic features that are the result of this language contact. We shall also look into the fact that

45

46

A Manual of Linguistic Fieldwork and Structures of Indian Languages language contact not only leads to convergence of languages, but gives rise to complex and redundant grammatical features. Let us first consider convergence of various linguistic features at all levels of grammar among Indian languages. Before we begin, we must point out the glaring fact that Dravidian and Munda languages have incorporated more linguistic features of Indo-Aryan languages than the latter have features from the former two families. 2.4.1

Various Grammatical Levels

We shall briefly consider the phenomenon of convergence in various languages of different families in their lexicon, sound system, morphological system and in their syntactic system. The discussion is in no way exhaustive, as it is a topic for a separate study. Our basic intention in raising the issue of convergence in this manual is to draw the attention of the field investigator to the fact that no Indian language has remained ‘typical’ of its language family. Languages have diverged from their original stock due to intense and long contact, and therefore one must be careful while working on a new language to identify converged features, as they speak volumes about the history of the community. The easiest part of a language to borrow and incorporate is the lexicon. The lexicon of various languages of the Munda and Dravidian families has been borrowed from the Indo-Aryan family, especially that of Hindi in Munda and of Sanskrit in Dravidian. In case of the Munda and the Dravidian languages spoken in Jharkhand, one can safely say that more than 50% of the basic words are replaced by Indo-Aryan words. This is especially true in the area of verbs. A large number of IA lexical items, both from basic vocabulary and non-basic vocabulary, are directly borrowed in the tribal languages. Our study (1991) on Kurux indicates that out of a total of 15,000 words collected, 2100 were found to be of IA origin. The words pertaining to numbers are borrowed from Hindi but the vigesimal counting system is borrowed from Munda. Words pertaining to natural objects, such as words for ‘grass’, ‘flower’, ‘dust’, ‘cloud’ etc., temporal organisation such as ‘day’, ‘month’, ‘year’ etc., a large number of nominal modifiers, and verbs of intransitive and transitive nature as well as of psychological predicates, kinship terms such as words for ‘brother’ ‘sister’, ‘father-in-law’ etc. are all directly borrowed from IA. Kurux, however, does maintain the indigenous address forms and kinship terms for ‘my younger brother’, ‘your elder sister’ (terms indicating the genetic relationship with genitives)3. Literary languages such as Malayalam and Kannada use Sanskrit vocabulary both in their written and spoken forms, more in the former than in the latter. The process of relexification4 is strong and is an ongoing process in all of these languages. For instance, directional words in Kannada (Sridhar 1978:201), in Kharia (Malhotra 1982) and in Kurux (Mishra 1992) are borrowed from Sanskrit and Hindi, respectively. Similarly, words pertaining to empathy are borrowed from Sanskrit into several Dravidian and Munda languages. As is to be expected, relexification takes 3

For details see Abbi 1993. Relexification refers to a process where the indigenous words of a language are replaced by borrowed lexical items along with grammatical morphemes. This paves the way for the productive use of borrowed morphemes with indigenous words, e.g. Hindi feminine marking –i in Kharia and Kurux.

4

2. Language Families, Language Contact, and Areal Universals place with complete adaptation of the host language’s phonological structures. Thus, when Dravidian languages borrow an IA word with a nasalised vowel, it loses the nasalisation, and when Khasi borrows an IA word with a palatal affricate it replaces it by a sibilant, and a retroflex sound is replaced by a dental or alveolar. 2.4.1.1

Sound System.

The wide-spread presence of aspirates in various Dravidian and Munda languages is due to convergence with Indo-Aryan. The occurrence of voiced and voiceless aspirates in Telugu is due to IA influence, and so are the occurrences of retroflex sounds in Munda, stemming perhaps from both Dravidian and Indo-Aryan influence. Kannada shows violation of some positional constraints due to Indo-Aryan contact. Thus, the presence of word-initial voiced consonants, as in gamak, or initial consonant cluster with non–homorganic plosive cluster such as in pra:r«bdH«, and word final obstruents such as in tak«t, among others, may be attributed to IndoAryan influence (Sridhar 1978). Kurux spoken among IA languages and its dialects and Brahui spoken near Balochi show ample traces of convergence towards IndoAryan. These languages have adopted vowel nasalisation, and for Kurux one can say with certainty that it has developed the qualitative distinction in vowels in addition to keeping the original Dravidian trait of maintaining vowel length. A recent study of Khasi, the Mon-Khmer language of the Austro-Asiatic branch, indicates that the language has developed retroflexes [ÿ] and [ê], sounds that were non-existent until 19705. Sinhala, the language spoken in close proximity with the South Dravidian language Tamil in Sri Lanka, has lost all aspirates (Gair and Paolillo 1997:02). 2.4.1.2

Lexical Morphology

There have been an equally large number of penetrations in the area of lexical morphology. The foremost is the IA infinitive suffix –na being borrowed by Munda languages such as Kharia and Santhali, as well as by Dravidian Kurux. Kurux as well as Kharia have developed the gender-marking suffix –i borrowed from Hindi. Consider the example where gender marking is represented on both nominals and modifiers. Table 2.7

Gender Inflection in Kurux from Indo-Aryan

Masculine

Feminine

a:l«s ‘boy’ beˇa: ‘son’ ma:m«s ‘mother’s br. ajjos ‘grand father’ sona:r ‘goldsmith’ loha:r ‘blacksmith’ hindu: ‘a Hindu man’

a:li beˇi: ma:mi: ajji: suna:rin loha:rin hinduwa:in

5

IA suffix

‘girl’ ‘daughter’ ‘mother’s br. wife’ ‘grand mother ‘wife of goldsmith’ ‘wife of blacksmith’ ‘wife of a Hindu’

-i -i -i -i -in -in -in

Personal field experience. In 1978 I witnessed that these sounds were in transition as they freely varied between alveolars and retroflex. However, a recent trip to Shillong made in 1999 revealed that retroflex sounds had gained phonemic status.

47

48

A Manual of Linguistic Fieldwork and Structures of Indian Languages kurux ‘a Kurux man’ kuruxni: caêra: xall ‘bare field’ caêri:mukka: bahila: a:l«s ‘impotent man’ bahili: a:li kanêa: a:l«s ‘blind man’ kanêi: a:li

‘wife of a Kurux’ ‘bald woman’ ‘barren woman’ ‘blind woman’

-ni: -i -i -i

Gender marking has infiltrated into CP constructions as well. es’a: kala: es’ai kala:

‘you (m) having broken (it) go’ ‘you (f) having broken (it) go’

Kharia c«ngna: ‘rooster’ baêa ‘father’s elder br.’

c«ngni: baêi:

‘hen’ ‘wife of baêa

-i: -i

The highly agglutinative structure of Kurux and the incorporating structure of Santhali and Kharia are being replaced by the inflecting and non-incorporating structure of Indo-Aryan. These are clear cases of structural borrowing leading to a restructuring of grammars. Similarly, the classifier system, a feature dominant in Eastern Indo-Aryan languages is found in Malto, another Dravidian language spoken in Jharkhand. Kharia and Kurux have also borrowed the classifiers from Indo-Aryan. Interestingly, language contact does not limit its influence to borrowing or incorporation of a grammatical element in a recipient language, but also provides the right environment for its further development. A case in point is the classifier system in Malto, the language which initially borrowed classifiers from IA, but then developed these to such an extent that they surpass the system of the donor (Mahapatra in Abbi 1997:56). Thus Malto jh«n < j«n IA, go/goˇa, ˇhong, and ˇhu used as numeral classifiers are all from IA (see section 5.1.1.2.) Consider some examples from Munda and Dravidian languages of Jharkhand. Classifiers: irb jh«n a:l«r tiin jh«n lebu-ki tiin go kun}u en} goˇa mann

‘two men’ ‘three men’ ‘three children’ ‘two trees’

(Kurux) (Kharia) (Kharia) (Kurux)

In Kharia the numeral classifier ˇhong occurs with Kharia numeral /ubar ‘two’. 2.27

hokra ?ubar-ˇhong kun}u aij-kiyar his two-clas. child be-dual ‘he has two children’

In Kurux an IA cognate ˇhu: appears for classifying non-human but animate nouns. 2.28

mu˜ê do:y hu: «lla: three ten class dogs ‘thirty dogs’

2. Language Families, Language Contact, and Areal Universals As said earlier, unlike Kurux and Kharia the other North Dravidian language Malto has a very elaborate and highly developed classifier system unparalleled even in the Magadhan dialects from which it is supposed to have borrowed (Emeneau 1956, 1980; Mahapatra : ibid.). Other than the classifier system one finds clear cases of formerly unknown derivational affixes, especially in literary languages of the Dravidian family. Thus Sanskrit suffixes such as -«nt«, -isÿH«, -kara are freely used in Telugu as well as in Kannada. Hindi-Urdu suffixes like –da:r (Ka. p«ttedara ‘detective’), -kHor (Te. l«ncako:ru ‘corrupt man’), -wala (Ka. mi:sevala ‘the moustached man’) etc. (Sridhar: 1978) and prefixes (not known to the language system as shown above) such as «ti-, su-, p«ri-, be-, etc. are borrowings from Sanskrit and Hindi. 2.4.1.3

The Syntactic Structure

The convergence at the level of syntax is no less significant. Munda languages such as Kharia and Santhali have deviated from their typical lack of distinction between morphology and syntax and have become analytical in nature, as mentioned earlier. This involves the constituent order of NP + VP. In contact situations, the word-phrase structure changes to the IA sentence-structure type, which occasionally sacrifices the expressive power manifested in the original syntactic structure. Consider the following sentence from Kharia. 2.29

/kol -ob’-no/ dom-dha:b-la?-kiyar/ ‘you were being fed by each other quickly’ /kol/

reciprocal marker

/ob’/

causative prefix

/no// verb root ‘eat’ /dom/ passive infix /dha:b/ verb modifier signifying quick action /la?/

past continuous

/kiyar/ dual number These types of phrasal constructions have now become extinct with the decline of the word-phrase pattern. Note that in a single word-sentence all the notions of reciprocity, causation, passivity, manner adverbial, the verb marker for the continuous aspect, past tense, and dual number are incorporated. Munda and Dravidian languages have also incorporated relative-correlative constructions under the influence of IA (see the following section). Conversely, Indo-Aryan languages have borrowed participial constructions and converbs from the Dravidian languages. The best examples of syntactic convergence can be seen in the structure of Dakhini, the variety of Hindi-Urdu spoken in Hyderabad and parts of Karnataka. We shall discuss this in a separate section. An interesting observation has been made by Gair and Polillo (1997) about Sinhala that the language has lost the

49

50

A Manual of Linguistic Fieldwork and Structures of Indian Languages typical relative-correlative pattern due to contact with Tamil spoken in Sri Lanka. Thus, while Indian Dravidian languages incorporate the Indo-Aryan pattern, the Indo-Aryan language in Sri Lanka incorporates the Dravidian pattern for the relativised construction. Our latest fieldwork on Kurux reveals that the language has already lost the subjunctive mood suffix nekka that Rev. Hahn (1901) mentioned in his grammar 100 years ago. Similarly, in conditional phrases the conditional marker hole has been replaced by Hindi to. Hence a sentence (2.30) from Hahn’s grammar is rendered as (2.31) by the current speakers. 2.30

idin eska:n hole nima:n la:u/on this break cond you- all-I beat

2.30

idin eska:n to nima:n la:u/on ‘If you break this I will beat you’

Such intense contact between languages of various families have given rise time and again to converged languages such as Sadari in Central India, Dakhini in South India, Konkani and Marathi in Western India, and Nefamese, Nagamese and Chakesang in the Northeast India. They form independent areas of study and should be investigated seriously, as they reveal intricate ways of the process of language change. Instead of going into details of each of these languages and their grammatical structures, we shall concentrate on two very important stages that these languages pass through on the journey of convergence resulting in a restructuring of their grammars. 2.4.2

The Restructuring of Grammars

Long and stable contact among different linguistic communities gives rise to various changes in grammatical structures not necessarily brought in by that of the dominant languages. Thus Badaga, previously thought of as a dialect of Kannada, has converged with surrounding languages because of the adstratum of other tribal languages (Raichoor: 1997); Gondi and other Dravidian languages of Central India have developed subject incorporation into finite verbs on the pattern of Munda languages (Andronov: 1997), and Brahui, the Dravidian language which is spoken in Baluchistan, has phonological and morphological traits of Balochi, an Iranian language, because of its proximity to the latter. Occurrence of the IA conjunctive -ki and the use of the IA verb k«r ‘to do’, for instance, are direct borrowings from Iranian. Brahui shows its proximity to Iranian not only by adopting Iranian features, but also by dropping typical Dravidian features such as the gender system, the exclusive-inclusive distinction in the first person plural and pronominal references (Emeneau: 1980, Andronov: 1997). Language contact thus brings about changes in grammars by providing a constant stimulus to add to, drop and internally innovate linguistic structures. It has been observed that languages pass through two important stages while they restructure their grammars, a conflicting stage that results in parallel structures and the compromise stage resulting in redundancies. The former, i.e. the conflicting stage gives the speakers the freedom to choose between indigenous and borrowed

2. Language Families, Language Contact, and Areal Universals structures, while the latter stage is seen as the stage when obligatory use of a particular structure is maintained. The two stages give different results. Thus, during the course of the language change some features might be at stage one and others might be at stage two. The net result is that at a particular point of time the grammars of languages in intense contact are complex and not simple. These stages can be diagrammatically represented as: Beginning®®®®®®®Stage 1®®®®®®®®®Stage 2®®®®®End Simple ®®®®®®Parallel structures ®®®®Redundancies®®®®Complex Non-converged®®®Conflicting stage®®®Compromise stage®®®Converged Figure 2.1. Stages Of Language Change in Contact Situations The first line indicates the time period of change, the second line indicates the nature of the linguistic structures resulting from language change, and the last line indicates the nature of the various stages a language passes through in a contact situation. The terms ‘simple’ and ‘complex’ are relative terms used in comparison to the structures of the language before and after the contact. Let us consider some examples of stage 1 and stage 2, i.e. parallel structures and redundancies that are created due to languages in contact. 2.4.2.1 The Parallel Structures The conflicting stage of language change gives parallel structures that can be observed in a language when the language concerned offers a choice between the two structures, one from the donor language and the other from the recipient language at any level of grammar. Thus Marathi, an Indo-Aryan language, offers two parallel structures of modifier-modified construction, one the indigenous relative-correlative strategy (as discussed above) and the other a participial construction borrowed from a Dravidian language. Speakers use either of these constructions, as both types are acceptable and grammatical. Marathi 2.32

(a)

jo ma˜us titha ubha ahe to ma:jha bhau ahe rel. man there stand is Correl my brother is

(b)

[titha ubha aslela] ma˜us majha bhau ahe (Jhungare 1997) [there stand part 3msg] man my brother is ‘The man who is standing there is my brother’

The sentence in (b) is constructed exactly on the model of Dravidian languages, as can be seen in the following Telugu example. Telugu 2.33

[akk«ê« nuncuna] «t«nu na: t«mm«êu [there standing]

he

my brother

‘The man who is standing there is my brother’.

51

52

A Manual of Linguistic Fieldwork and Structures of Indian Languages The adoption of a relative-correlative structure in the Dravidian language Kannada is reported by Sridhar (1978: 204). Kannada 2.34

ya:va hengasu hasiru si:re uÿÿiddaño avañu nanna he÷êati Rel woman green saree wears

Correl my (is) wife

‘The woman who is wearing the green sari is my wife’. A very fascinating example of optionality came to our notice when we worked on Bangani, an Indo-Aryan language spoken in the western Himalayan region of Garhwal with a history of contact with various languages over the past 2000 years (Zoller 1988). The language has the option of choosing not only between (a) and (b) sentences but also between the two redundant (see 2.35(a) and 2.36(a)) negative particles occurring in the same construction. The inter-sentential and intrasentential optionality may exist in any language in a stable and enduring contact situation. Consider: 2.35

(a).

ram (na) n thi aunde lagiu Ram neg. neg-aux come prog/dur

(b)

ram na r

and

lagi

Ram neg. dur. come prog/dur ‘Ram is not coming’ 2.36

(a)

au kha˜ (na) n thi th khandi lagiu I food neg. neg-aux aux (pst. msg.) eat prog/dur.

(b)

au

kha˜

I

food

na r

th

khandi lagi

neg. dur. aux (pst. msg.) eat

prog/dur

‘I was not eating food’ 2.4.2.2

The Redundant Structures

Redundancies can be identified by the existence of non-optionality between the two linguistic structures performing the same or similar function. Consider the example from Dakhini that incorporates both the participial form from Dravidian and an incomplete relative-correlative structure from IA in the same sentence, thus resulting in redundancy in the grammar. 2.37

[a:ye so6] vo wa:la: a:dmi came rel corel. modfier man ‘The man who came’

6

The perfect relative participle in Dakhini is formed by adding an element so to the verb in the past tense of the embedded sentence. (Arora : 349)

2. Language Families, Language Contact, and Areal Universals The sentence given above is redundant in one more aspect. It uses the relative–correlative strategy as well as the use of the definitiser –wala, both of IndoAryan stock. The best example of redundancy can be seen in the morphology of Kharia where it combines the typical causative prefix indigenous to the AustroAsiatic family with the Hindi suffix –wa, which is typical to the Indo-Aryan structure, in the same word. Interestingly, the language offers both parallel structures and redundant structures. Thus, the first two examples (a-b) are used as parallel structures while the third one (c) is a case of redundancy (marked in bold letters). Kharia 2.38

(a)

i

ciˇhi: ob-likh-o-i

I letter caus-write-pst-1sg ‘I got a letter written’ (Kharia paradigm) (b)

i

ciˇhi: likh-ay-o-i

I

letter write-caus-pst-1sg.

‘I got a letter written’ (IA paradigm) (c) (IA)-

i i -a? souêom-nai -bu ciˇhi: ob-likh--ay-o-i I I-say husband-mine-by letter caus(Kh.)-write –caus pst.-1sg. ‘I had my husband write a letter’.

Interestingly, similar to Dakhini this language also shows redundancy at several levels, e.g. morphological and syntactic. The fact that (c) uses the verb ‘to say’ [a device to cause someone to do something], perhaps a borrowing from the Dravidian language Kurux, there was no need to use morphological causative at all as the sentence such as ‘I told my husband to write a letter’ was sufficient. At present the example (c) simultaneously provides three causative strategies in the same sentence, one syntactic and two morphological in nature. Bangani examples that we saw earlier show redundancy of the durative markers r! and lagi (2.36b) and the auxiliary –thi as in n!thi and th! in the same constructions (2.36a). This exhibits the possibility of finding more than one redundancy in the same construction. Similarly, multiple affixes in Andamanese (2.1.5.2); the presence of subject pronoun as well as subject encoding in the verb in Kharia that we discussed in 2.17 are clear cases of redundancies and a restructuring of grammars. 2.5

Conclusion

Redundancies arise when the retention of indigenous structures and the adoption of contact language structures exist side by side. The oscillation between old and new culminates in redundant expressions that can be found at any level of grammar

53

54

A Manual of Linguistic Fieldwork and Structures of Indian Languages whether it is at the sound system, lexical and morphological system or syntactic system.7 Thus, the multilingual context of intense language-contact at times gives rise to complexities at the expense of the native simplistic structures. This, however, does not rule out the reverse process, i.e. gaining simplification and not complexity. Redundancies and parallel structures co-exist with independent innovations in both donor and recipient languages, as there is an expansion of language use semantically, morphologically and syntactically. An intensive study of these emerging and converging language structures serve as a view finder of the changing and converging social patterns of the communities using these speech forms.

7

For instance, modern speakers of Hindi use many compound formations which are indicators of such redundancies, e.g. dh«n d l«t (wealth< Skt. + wealth< Persian) ‘wealth’; ph«l fruiˇ ( fruit< Skt. + fruit< Eng.) ‘fruits’.

Deleted:

3. The Preparation 55

Chapter III The Preparation

Once a field investigator has equipped herself/himself with a basic knowledge of the language scene and linguistic structures of Indian languages as well as with the relevant ‘areal’ features therein, s/he has to concentrate on the concrete and not-soconcrete aspects of the fieldwork. It is always advisable to prepare oneself with these issues (as discussed below) before one sets out to tread the field. This is the first stage of preparation. We will discuss them under nine topics. 1.

Budgeting and reservation

2.

Dress code and other luggage to take along

3.

3.1.

Knowledge of the geographic, social and economic status of the area under consideration

4.

Field Work in the classroom

5.

Preparation of the questionnaires

6.

Contacts in the field

7.

Language of elicitation

8.

How to choose appropriate informants

9.

The code of conduct of the investigator Budgeting and Reservation

It is very important that budgeting is done at least a year in advance if you are expecting funding from the sources outside the institute you are working at. In case your fieldwork is related to your work in the institute/university you are employed in, give at least six months to get the budget approved, and money made available to you. Though this time period varies from one university to another in India, yet on the average, Indian universities take a minimum of six months to sanction any kind of budget for fieldwork. While writing out the budget, one has to take into account the current prices operating for the mode of transportation available in the area of fieldwork and rail fares (wait for the railway budget which is announced in April every year), as there is an increase in rail tariffs every year. It is important to have some idea of prices of commodities and lodging facilities before one writes the budget as these vary from city to city and from village to village. Similarly, if one is planning to work in the remote hilly areas, there is no point in asking for jeep or bus fares, as most probably one will have to either trek on foot or hire mules to reach the heights. Although in the last fifty years pukka and kuccha roads have reached the remote Indian villages, and buses do ply regularly (and so do the jeeps) on the hilly roads, in very remote areas one still has to rely on mules. At times we have realized that mules are surer modes of transportation than busses, as the bus service, if at all, is fixed for specific times and most often is not available after 4 p.m. Mules, on the contrary, are always available, even after dark. In towns and cities of course one can

56 A Manual of Linguistic Fieldwork and Structures of Indian Languages rely on any three-wheel- or four-wheel vehicles (see table 3.1. The budgeting has to incorporate the following: 3.1.1

Travel

(a) Before reaching the field as such, rail or bus fares as well as fares incurred in reaching the rail or bus station. (b) Within the field, costs for a jeep, bus fares, mules etc. 3.1.2

Daily Allowance (DA) (Board and Lodging)

(a) During the period of journey. It is not cheap to buy meals in trains. we have noticed that a per diem cost of food during a train journey is more than that in the actual field. The journey to the field may be undertaken by bus, jeep or taxi. In such cases, an approximate idea can be made of the cost of the meals during the journey. (b) In the field. This would include board and lodging charges. One must make reservations in advance (see below) for lodging facilities. In cities one must look for hotels, guest-houses, university guest houses (these are very reasonably priced), PWD (Public Works Department) guest houses etc. for lodging. In small towns and villages one can explore Daak bungalows, PWD guest-houses, municipality rest houses, and other guest houses that come under the jurisdiction of the State and Central governments. One should also look for the possibility of staying in schools (especially if they are closed for vacation) and dharamshalaas (in fact they are clean and very cheap). In Shiroda, a village in Goa, we found a very comfortable and clean Kamaakshi Dharamshalaa which offered us nononsense food, good classical temple music early in the mornings, and a large number of various types of devotees who turned out to be excellent informants later on. Most of the Dharamshalaas run on donations made by devotees and hence charge very nominal price for the services they render. One should not exploit the facilities and it is advisable to donate generously (if possible) before leaving the place. In Shillong (Meghalaya) we could not find any of the accommodation types mentioned above and thus rented a small house and cooked food ourselves. One has to keep all kinds of possibilities in mind while planning the budget. My experience is that unless someone is suffering from ill health, it is not advisable to run your kitchen in the field. It is too time-consuming and not without frustrations. Of course, if the fieldwork is of more than two months duration one could think of the possibility of running her/his own kitchen. My colleagues from other departments of Social Sciences prefer to take along cooks for the job. One has to use one’s own discretion to frame the budget according to the requirements, time, feasibility and goals. There have been cases where investigators have stayed with the contact person’s family, or with the family of one of the informants. If this arrangement works out, there can be nothing better. If one is working on a sociolinguistic topic, it would be of immense help to get an opportunity to stay with a family and

3. The Preparation 57 observe day-to-day life among the various members of the speech community. It is surprising to see how fast Indian communities treat an outsider as one of there own, after the initial problems of adjustment. In such a case, try to be one of them, eat what they eat, behave the way it is expected of you, sing and dance with them, and if there is a grief in the family try to share it with the members. You will be a member of the family sooner than you realize. 3.1.3 Money for the Informants1 It has been noticed increasingly that informants expect to be paid for the time they spend responding to the fieldworker’s queries in small towns and cities. We generally pay Rs.50 per session (on the average 90 minute long). However, this rate is flexible and more (but not less) can be given depending upon the nature of the questionnaire. The ideal rate for this kind of sitting is Rs. 100/- per session of 90 minutes duration. It is increasingly observed that city dwellers are hard pressed for time and the only incentive that works is the incentive of money. Contrary to the town people, villagers neither expect nor is it possible to pay each and every informant that one interviews in the field. As discussed later in the chapter, we generally interview a minimum of 50 informants from a location to achieve a good and representative sample, and each informant spends long hours with each of the field investigators. Also, we have noticed that villagers at times feel offended when given any monetary payment for an interview. It is humiliating for them to be paid for talking to a guest2. However, there are ways of showing your gratitude to the informants. One could ‘pay’ in form of tokens such as buying some eatables for the children of the family every time one visits the family, paying for snacks and tea at the tea stall (if the interview takes place at that location), paying for small items in the bazaar if your informant accompanies you, or just buying them a gift. One does find various ways to pay back if one wants to. It is difficult to specify each and every mode as each situation demands and dictates a different mode of payment in the field. What one has to remember is that at no point should it appear that the ‘outsiders’ (i.e. fieldworkers) exploited the villagers (i.e. the informants). In cities, if the informant travels far to reach you and in the process incurs some expenses, it is customary for the fieldworker to bear the cost of travel incurred by the informant. Enough money should be budgeted for this intended expenditure. No informant will visit you again if s/he spends her/his own money to reach you. On should roughly calculate the number of such visits required and the expected mode of transportation to finalize the budgeting of this aspect. 3.1.4 Communication Network Charges 1

An informant is a native speaker of the language in question who is employed (or provides vcoluntary service) by the field worker for the purpose of elicitation of data. The alternative terms used for the word 'informnant' have been 'assistant' or 'consultant'. However, the term 'informant' is still used widely in the literature. We will adhrere to this very use. 2 Indians at large still follow the Sanskrit motto atithi devo bhav ‘The Guest is God’. How can you expect money from God?

58 A Manual of Linguistic Fieldwork and Structures of Indian Languages (Telephones, Telegrams, Faxes, and other postage) One should keep a fairly large amount of money for sending telegrams, making long-distance phone calls and sending faxes to make arrangements in the field. One should also include the money needed for establishing telephone communication with the informants. Our experience shows that locating the right number and appropriate informants in itself involves a large number of telephone calls and letters. It is to be noted that unless one is working in very remote areas of the country, telephone facilities are available throughout India. STD (State Trunk Dialing) facilities are provided in all areas of India, be it in villages or in small towns. Even the hilly areas of UP and Himachal are very well linked by telephone. With STD facilities available in every corner of the country, one should expect to incur expenses on long- and short-distance calls. In cities and towns, telephone facilities are easily available, though not very cheap. You will be required to make phone calls to establish appointment dates while in the field as well as prior to reaching the field. It is always advisable to continue the contact with the parent institute, a colleague or any other person who is attached to your project, so that the whole group is informed at all times of your whereabouts. Of course, this is suggested only in cases where several members are involved in the project. The telephone directories may be consulted in advance to get the estimate of phone call rates. Keep enough money for postage. 3.1.5

Stationery

While budgeting under this head one should plan for the entire duration of the project, not only for the time spent in the field itself. Stationery items for fieldwork generally include: (i) index cards (3*5 and 5*7) (ii) pencils, sharpeners, erasers, pens, markers, crayons, highlighters, transparency writers, etc. (iii) floppies (for data recording) (iv) Reams of paper, carbons, color papers, chart paper, etc. (v) spiral notebooks, ruled as well as unruled (in all sizes) (vi) registers (preferably unruled) soft and hard bound (vii) slates, chalks, (especially used in village settings) (viii) pictorial cards (ix) envelopes of various sizes to keep cards and notes (x) Any other item specific to the needs of the fieldworker. One should keep a good amount of paper in stock, as it is advisable to make photocopies of the data and frequent analysis thereof from time to time lest it get

3. The Preparation 59 lost. It will be clear that part of the stationery items are to be used within the field, and part after or before conducting the field research. Except for (iii) all other items are needed both in the field as well as outside it. We find the use of slate and chalk (generally known as battii in UP) very useful in village settings. Firstly, the sight of it is very familiar to all, secondly, one can make all kinds of pictures and diagrams to explain oneself while interviewing the informants. Similarly, informants find it extremely helpful to draw and explain some concepts/items on a slate and then explain these in the contact language, or in their mother tongue. As far as the use of registers is concerned, one might find it very convenient to have all the data in one place rather than scattered in hundreds of cards. We personally find the spiral notebooks very useful as they fit in a purse (and one has minimum chances of losing them) and can be replaced by a new one when the old one finishes. They thus are flexible enough for further classification of data, either by topic or by the period of investigation. For instance, one spiral book is sufficient for the basic word list, basic sentences and general information drawn from at least 3 to 5 informants. This way the ‘Basic Information’ can be stored in one place. But of course it is up to every individual how to divide the data. Our students prefer registers to cards and spiral notebooks as they do not mind carrying them along all the time. The convenience of having all the data that has been collected ready at hand every time it is needed far outweighs the burden of carrying the register. The index cards are useful for collecting lexical items, especially with the project of writing a dictionary in mind. Each card should contain only one lexical item, its grammatical information, examples, derivational forms, etc. Later when the data is analyzed, large index cards should be used. See the section on Echo formations (4.5.7.1) to have an idea of preparing a lexical entry for dictionary purpose. 3.1.6 Equipment and Accessories My teacher, Prof. Charles Hockett, while teaching us ‘Field Methods’ course used to say that ‘the best equipment that one has to take to the field is your pair of ears’. There is no denying the correctness of this statement. Without a good training in phonetics and auditory perception, a fieldworker becomes a dangerous linguist, as s/he would bring forth data which does not really exist in the language. However, one has nowadays several devices which can store data electronically. These are not substitutes for good auditory training, however, but only complementary to it. One of the most important pieces of equipment for every fieldworker is a tape-recorder, a still camera, and a video camera (if one also wants to shoot a movie). A market survey would be needed, as tape recorders are available in all shapes and sizes these days. We personally prefer a small one that fits in a purse or a little bag. The very sight of a big tape recorder generally puts the informant in an uncomfortable situation. Keep money allocated for a good tape recorder that has at least the following features: • tape counter • automatic ‘shut off’ with little or no noise • built-in microphone

60 A Manual of Linguistic Fieldwork and Structures of Indian Languages • light in weight • clear sound production • operates on both electric power supply and battery cells • takes standard-sized cassettes These days the electronic market is full of tape-machines, and one should choose the one that is best suited for the job. Linguists who have internet facilities should go to the web site /www.sony.com/ to see the range of tape recorders available by Sony. If there are a group of investigators going to the field together, the ideal situation is to have one tape recorder each, as every investigator is independent about her/his movements in the field. While budgeting, the cost of tape recorders constitutes a major part of the expenditure. It is also not a bad idea to buy a two-inone (radio + tape recorder) as this would facilitate taping regional radio programs in the languages of the area you visit. We find this item quite useful as there are interviews, songs, plays, childrens’ programs and news bulletins on the medium wave which are easy to record without any background noise. Nowadays, even major cities broadcast programs in local dialects and language varieties on air. In addition to tape recorders, at least one still camera should also be considered a ‘must’ for fieldwork. This need not be an expensive one but should have a built-in flash system. Always remember that the fewer the gadgets you take to the field, the easier the informant will feel with you. However, our experience shows that people in general like to be pictured and taped. A big component of the budget should go for batteries, tapes (preferably 60 minutes ones) and films. Almost all cities of India sell cassette tapes and films and hence you need not take the burden of carrying them all around the country. Depending upon the duration of the project, enough money should be kept aside for these accessories. The best recording is done by your pencil and by your ears. Always keep in mind that a 60-minute filled tape of conversation/discourse may take more than 20 hours of transcription time. If the language is tonal, the period of 20 hours may stretch further. Among all the types of data that are recorded on tapes, the easiest to transcribe are the basic word list or any list of words as these are always spoken slowly and are punctuated at regular intervals by silences/pauses. 3.1.7

Data Processing

Depending upon the nature and size of the corpus one might have the need for data entry job taken up by an expert. The salary of such personnel varies from city to city and from institute to institute. While framing the budget the field investigator has to visualize in advance the nature and the size of the data. At times one might need a special programmer to suit your needs. There are some programmers who might charge more than the usual price, but may help you by designing morphological paradigms, or dictionary programs, or basic morph-to-morph-translated sentence types with complete search facility.

3. The Preparation 61

3.1.8 Word Processing/Typing The final job, as in any project, is the writing of results. One seldom realizes that writing up the final report turns out to be an expensive job. This is especially true of linguistic reports, as typing or word processing the report with various symbols and diacritic marks increases the cost by 50-100% of the regular word processing/typing. One should always consult the person who undertakes word processing/typing before writing the budget. Out of the seven types of expenditures described above, only the first five are incurred in the field, as well as before going to the field. The last two types are normally incurred outside the field (the accessories being an exception). 3.1.9 Reprographic Services/Xeroxing A special budget should be made for xeroxing facilities as all the elicited data should be copied from time to time, and sent back from the field to the host institute. This is recommended to minimize the risk of loss or theft, or simply ‘misplaced’ field data books. One should be prepared for all kinds of contingencies when one is in the field. Reprographic facilities will also be required for making copies of questionnaires, maps, and other relevant material. Finally, one would require this facility to make multiple copies of the report. This is one expenditure which is incurred throughout the fieldwork. Have a sufficient amount of money reserved for it. 3.1.10 Printing When the field is vast and the number of informants to be interviewed is large, it is recommended that the questionnaire be printed, as it is cheaper to print it than photocopying (xeroxing) a large number of copies of the questionnaire. To minimize the cost, printing may be undertaken on both sides of the paper. Most of the questionnaires in an Indian setting are bilingual in nature (English and the contact language). If one decides to have a bilingual questionnaire, one should take into account the cost of printing such questionnaires before framing the budget. Biscriptal (such as Urdu and Hindi) questionnaires are most expensive. 3.1.11 Books and Journals No matter how knowledgeable the investigator is, no-one can conduct fieldwork without incurring some expenses on relevant books, journals, and maps. Sometimes, one has to go some distance to acquire a manuscript from a personal collection, which might be rather an expensive venture. 3.1.12 Contingencies A certain amount of money for unforeseen circumstances should be kept aside. One must remember that this is generally 10% of the sub-total of the budget (see below). A contingency can be anything from sudden illness in the field to purchasing an extra mat to sleep on or the payment to local people for their services. Your imagination can run wild, yet it is difficult to identify beforehand what type of contingency one might meet. This also includes an unexpected rise in the price of rail and airfares. There will be many unanticipated costs in a field study.

62 A Manual of Linguistic Fieldwork and Structures of Indian Languages 3.1.13 Overheads Most of the Indian institutes and universities charge an overhead, which varies from 10-15% of the total budget of the fieldwork project to render their administrative and secretarial assistance. This amount, generally overlooked by the field investigator, might put her/him in very precarious situation later on. We would like to draw the attention of the readers to the fact that the budgeting proposal suggested here does not include the cost of and purchase of computers and printers. Nor does it include the cost of typewriters. It is believed that these facilities already exist with the institutes or that services of such types can be purchased (as included in word processing or typing). The final form of the budget proposal would look somewhat as suggested in table 3.1 given below. Table 3. 1 Budget Proposal for Fieldwork in Linguistics I. Travel (a) Before reaching the field…………………. (b) Within the field…………………………… II Daily Allowance (a) During train/bus journey…………………. (b) In the field……………………………….. III Payment to the Informants………………… IV Telephone, Faxes, and Other Postal Expenses…………….. V Stationery (a) Paper products, pens etc. … …………………. (b) Floppies……………………………………. VI Equipment and Accessories (a) Tape recorder/Two-in-one……………….. (b) Still camera/Video camera……………….. (c) Cassettes (audio/video)……………………. (d) Batteries…………………………………… (e) Film rolls………………………………….. VII Data Processing VII Word Processing/Typing IX Reprographic Services X Printing of questionnaire(s) XI Books and journals SUBTOTAL 1. :::::::::::::::::::::::::::::::::::::::: XII Contingencies (10% of the subtotal 1 )………………….. SUBTOTAL 2. ::::::::::::::::::::::::::::::::::::::::::::: XII Overheads of the University or Institute (10-15% of the subtotal 2)…… FINAL TOTAL

:::::::::::::::::::::::::::::::::::::::::::::::::::::

3.2 Your Luggage As in the case of accommodations, railway reservations should be done at least six weeks in advance, as the Indian railways are heavily booked in all seasons and on all

3. The Preparation 63 routes. Railways also provide the best and cheapest way to travel within the country. Fortunately, Indian railways connect all cities, towns and quite a few villages. See Indian Railway map. Under this section, we are going to discuss the material artifacts one should carry to the field. Some of the items may be acquired in the field itself, but our experience shows it is far better to carry your own. We shall also discuss what to wear and what not wear, as dress is a very significant aspect that helps you in being accepted in the society or conversely immediately dislodges you from the community. 3.2.1. Materials to Take Along In addition to the materials discussed above under ‘stationery’ and ‘equipment and accessories’, fieldworkers should take as little as possible luggage with her/him. Her/his luggage should basically consist of: • Personal effects (clothing, shoes, toilet items, your specific needs). • Bedding • Some basic utensils and apparatus (stove, electric heater to heat water) • First-aid kit • A copy of ‘Where There Is No Doctor’3 • Mosquito net and repellent • Maps (geographical, political, social, linguistic, road maps) • Pots and pans (if you intend to cook). • Questionnaire/s (see 3.5 discussed later in the chapter). • Tape recorder/s, tapes and cells • Camera, films ideal

• Small gift items for informants (wall posters of several types make and reasonably priced gifts) • Knowledge of local customs, social behavior, and relevant cultural facts (see also below)

3.2.2 Dress Code No matter how simple a problem this might appear, dress code is very significant in making a good impression early on. Women specially should desist from wearing jeans and pants even in cities. Though city girls of India have taken to jeans, it is the Indian dresses such as saaris or salwaar kamiz that are looked at with regard. Despite our warning against wearing jeans, some of our girl students landed in Hanumangarh (Rajasthan) in jeans and the consequence was that for first two days no one took them seriously. On the contrary, young boys of the town tried to flirt with the girls. Even at home the women of the house did not encourage these students to enter the 3

“Where There Is No Doctor” 1980. Indian edition. Delhi

64 A Manual of Linguistic Fieldwork and Structures of Indian Languages kitchen (see also ‘Being a woman is a blessing’, below). The surest of all dresses that brings respect and acceptability by all is the old traditional dress of India, i.e. the Saari. One could take a couple of synthetic (polyester) saaris that are easy to take care of (I have now a good stock of my ‘field methods’ saaris). If one works in Haryana, Punjab, Himachal Pradesh, Rajasthan, Jammu and Kashmir, salwar-suits are equally respectable and convenient dresses. Fortunately, male fieldworkers do not have to worry about their dress as trousers and shirts are acceptable and respected all over India in rural and urban areas alike. However, in some cities of Southern India trousers may not be accepted if the fieldwork has to be carried out in and around temples. In such cases, the fieldworker has to learn to wear a ‘lungii’ (a threeyard cloth tied around your waist). Such areas are rare, but do at times put a fieldworker in a precarious and embarrassing situation. It is always better to be prepared than to be in a position of a set back. 3.3. The Ethnic, Linguistic, Social, Political and Economic Status of the Informants/Area to be Visited Before we acquaint ourselves with the social norms, customs, economic and political situations of our informants, We must remind the readers that we, the ‘outsiders’, are always looked at with suspicion by several indigenous communities. In every tribal belt that we have visited, we have found a word which is used for outsiders, but which really means ‘exploiter’. For instance, Kurux use diku for outsiders or immigrants, but this has an extensive meaning now for ‘exploiters’ or for ‘the one who sucks blood’. This mind-set represented by the language forewarns the fieldworker to tread cautiously and take utmost care not to hurt or even seemingly exploit the local people under investigation. It is advisable to acquaint oneself in advance with the socioeconomic situation of the area that one is about to visit in order to be prepared. There are a number of published materials in the various libraries about the societies and their cultural, social, and economic make up. In addition to this, it is expected that fieldworkers are aware of or have had an opportunity of reading the following recent linguistic studies, depending upon the area s/he is about to investigate. 3.3.1. Literature Study For Dravidian 1. Sanford B. Steever. Dec.1997. The Dravidian Languages. Routledge, London. For Indo-Aryan 2. Colin P Masica, 1991. The Indo-Aryan Languages. Cambridge University Press, New York.

3. The Preparation 65

For Austro-Asiatic 3. Norman Zide. 1966. Studies in Comparative Austro-Asiatic Linguistics. Mouton, The Hague. For Tibeto-Burman 4. James Matisoff.1978. Variational Semantics of Tibeto-Burman. The Organic approach to Linguistic Comparison. Institute for the Study of Human Issues Publication. Philadelphia. 5. James Matisoff 1991. 'Sino-Tibetan Linguistics: Present State and Fut8re Prospectus'. Annual Review of Anthropology. Vol. 20, 469-504. For Tribal Languages (Dravidian, Austro-Asiatic, Indo-Aryan, TibetoBurman, and Andamanese) 6. Anvita Abbi. 1997. The Languages of Tribal and Indigenous Peoples of India. The Ethnic Space. Motilal Benarsidass, Delhi. Of course there are other very important works in each area that cannot be ignored, but the list given here is the most current so far and will give an insight to a fieldworker about the language group s/he is interested in. 3.3.2 Ethnology As far as the knowledge of economic status and other ethnic aspects are concerned, Our experience is that advance knowledge is always ‘old’ when one reaches the field, but nonetheless should be acquired. While working on Khasi, the language spoken in the northeast in Meghalaya, we were equipped with the knowledge that the society was matrilineal and matriarchal. However, it was only after reaching the field that we realized the gravity of the situation. A week had passed and most of my students (who were male) had had very bad experiences in locating informants, as both females and males were wary of the students lest they marry their girls and take away Khasi property (which by social custom belongs to the youngest daughter of the family). It was only after my reassurance (which, coming from a woman, was taken seriously), that Khasi people let the boys enter their homes. Another incident that one could mention of here when we visited Chotanagpur area to conduct fieldwork on Kurux (Oraon). Our reading had informed us that (a) missionaries have suppressed the language and culture in the area, (b) Oraons have all taken to Christianity. However, on our visit to the place, we found a rather peaceful co-existence of Christianity and the old tribal religion where people worshipped ‘Dharmesh’. There was no divide whatsoever between the two. In fact, we came across several families where one brother would practice Christianity while the other worshipped Dharmesh., Both brothers were seen to participate in ‘Sarhul’, which is the festival of the non-Christian tribals. This reminded us of Hindus and Sikhs in Punjab, where the usual practice of Hindu Punjabis is to convert their firstborn son into a ‘Sikh’ (devoted to the Guru), while the other sons remain non-Sikh

66 A Manual of Linguistic Fieldwork and Structures of Indian Languages (‘mona’). In Chotanagpur, missionaries, contrary to our belief, have encouraged the Kurux language among the Oraon, as they have brought literacy and education to the villages. We, in fact, attended a mass conducted in Kurux. One has to be somewhat cautious about confirming the previous knowledge of the social customs and practices, as one does not know in advance what kinds of questions might annoy the informants or make her/him uncomfortable. I remember an episode from our Bangan fieldwork. Bangans live in the lower Himalayan region of the Western Garhwal between the Tons and Pabar rivers. We had some information beforehand about polyandry still prevalent in some parts of Bangan. When we had been in Bangan for just two days, one of my male students, who was overly inquisitive, asked a Bangani in Arakot about the practice. Less than half an hour had passed when the premises (we were visiting a middle school) were in a commotion. The word had spread that we had come to Bangan, not to work on their language, but to get into their homes. The readers can imagine the rejection and insults that we had to face following this. We were told in no uncertain words that we should pack and leave the place. I remember I had to seek for ‘forgiveness’ from the principal of the school on behalf of my impertinent student and clarify my position. Finally, the principal relented and, ironically, in the latter part of my stay in Bangan invited me home to meet his family members over an afternoon snack. I learnt that people in general are nice to you, as long as you do not probe into their most honoured and defended secrets. The purpose of narrating this incident here is that we would like to draw the attention of the readers to the fact that some knowledge of the ethnic and social customs is necessary before undertaking the field trip. However, one should only observe these and not be verbally inquisitive about them. We shall talk about this in greater detail in the next section ‘In The Field’ (3.7). The recent increase in demands for separate states by various tribal groups all over India have in one way united people for a common cause, which at times changes the ‘identification label’. A fieldworker must be aware of these. For instance, on interviewing a Mundari speaker in Netarhat (North Bihar), I got the response that my informant was a ‘Jharkahandi’ and not Mundari. Jharkhandi is a heterogeneous identity, since it denotes an amalgam of people living in Jharkhand at different stages of socio-economic development and speaking mainly languages of three distinct families, such as Dravidian, Munda, and Indo-Aryan. A typical ‘Jharkhandi’ is multilingual in Kurux, Sadari, and Mundari or Kharia (depending upon the region of dwelling). A typical saying in Jharkhand, which is a reflection on the deteriorating socioeconomic status of the local people, is: “Jharkhand is developing fast, but not the Jharkhandis”. This paradox serves as an impetus for the militant uprisings of the sub-alterns and the demand for an independent state4. 3.3.3

Language Study in big cities

As far as fieldwork in cities and cosmopolitan cities are concerned, the preparation is decided by the object of the study. If one is trying to find out whether the migrant community has retained its language or not, the advance preparation would be by 4

As we write this, three new states have come into existence, i.e. Chattisgarh, Jharkhand and Uttarkhand. See map 1.1.

3. The Preparation 67 acquiring some knowledge of the reason for and circumstances of migration. If the goal of the study is the grammar of the language of the area, one has to study the social strata that the community is divided into. For instance, when we wanted to study the sociolinguistic and linguistic aspects of the Hindi that people residing in Hyderabad use, we had to acquaint ourselves with the social history and economic status of the community at large, which we divided into four classes as: (a) Hindi-speaking Kayasthas, who migrated 30 years ago and have established themselves economically and socially in the upper strata of the society. (b) Urdu-speaking Muslims mainly living in the Charminar area. (c) Non-Telugus who reside in Hyderabad and use Hindi as L2 or L3. (d) Local Telugu-speakers of Hyderabad and vicinity using Hindi as L2. The reaction towards ‘Hindi’ and the level of competence in it was found to be very different in (c) and (d). Among (b), (c), and (d) we had to further distinguish between (i) educated, (ii) uneducated, and (iii) semi-educated as education played a significant role in forming opinions about the uses of the language. This issue was discussed above (see 1.2.3 and 1.4.4). Most of the Indian cities are heterogeneously constituted, hence some prior knowledge of the social and historical aspects of the people of the area under consideration is always useful. An advance knowledge of the area to be investigated will be crucial to subsequent work in the field. The faceto-face contact with the informants might fill you in with varied and conflicting facts of the language and its uses, but this information itself will force you to look into the reasons for these conflicting uses, which in turn is an indispensable inquiry. Despite fast-receding mother-tongues of migrant community members in big cities, linguists have observed language maintenance at the home level (Bangla, Punjabi, Sindhi and Tamil in Delhi; Hindi, Marwari in Calcutta; Hindi and Punjabi in Mumbai etc.) As discussed earlier in chapter 1, as long as languages are maintained for home use, no threat of extinction can touch these languages. We were aware of the fact that minority languages and tribal languages were dying fast, but just how fast was brought home to us when we encountered families in villages, where between two siblings the language was getting extinct. One of our informants once narrated an incident of being beaten badly by his father when the latter discovered him speaking his native language at home with his mother. The informant reported that none of his younger brothers and sisters know the language. There is another very disturbing incident reported to us by students who went to Ranchi to conduct fieldwork on Kurux is worth sharing with the readers. Everyone these students met for an interview denied being a native speaker of Kurux. They were all proud to say that they had forgotten their native language. Interestingly, each of them tried to identify someone else who could speak the language. However, the reality was altogether different, as later the students found out that all of them spoke Kurux at home and for intra-tribal communication. With a few exceptions, most of the speakers of tribal languages and of those belonging to the group of non-scheduled languages prefer to conceal their-mother tongue. Most of the tribal language speakers consider it a privilege not to know their own language. The self-destruction of one’s own mother tongue is a typical factor in

68 A Manual of Linguistic Fieldwork and Structures of Indian Languages fast-changing industrialized societies of which we should be aware. It is unfortunate that such a stigma is attached to minority languages 5 However, we must remember to evaluate a minority language according to its status on the status-cline given earlier in chapter 1. This will indicate whether the community you are going to investigate belongs to community of thriving speakers or of terminal speakers. 3.4

Fieldwork in the Class Room and the Time Required to know the Basics

An important aspect of fieldwork is to undertake some study of the language under investigation prior to visiting the field. An informant may be hired or consulted and the preliminary work of eliciting data with regard to the basic word list, basic sentences, basic phonetic and phonological structures as well as basic morphological structures of the language can be completed (see also chapter 4 on elicitation). As the methodology of working on the basics of a language differ from linguist to linguist one can never be very sure of the total time period that is involved in such work. However, an approximate time schedule that we have tried out can be given as follows. • one

A basic word list of 300 words will take 20 hours of elicitation provided the language under investigation is not tonal. Tonal languages will take and a half times of this.

•

The basic phonetics and phonology, i.e. the sound system of the language under investigation will take approximately 32 to 36 hours of elicitation.

•

The basic morphology, i.e. the word formation processes, will take at least 48 hours of elicitation time. If the language is inflectional, it would require more than 48 hours.

•

Basic sentence types will require 40-48 hours of data elicitation.

This time schedule might be extended if there are a number of informants and each informant is asked the same question. It is advisable to take a basic word list from each of the informants but work on other aspects of grammar continuously. Also note that suggested time period may vary with many other factors such as overenthusiastic informants or unwilling informants. 3.5

Preparation of the Questionnaire

An investigation into the grammar of any language resides in the comprehensiveness, clarity and completeness of the questionnaire. The questionnaire is the most important tool one has to get the requisite information. Always remember that a good questionnaire ensures the basics of linguistic research, i.e. appropriate data. Questionnaires have to be prepared and printed in advance. While making the questionnaire, the goal of the research should be kept in mind. All theoretical arguments and hypotheses that one wants to confirm or investigate the validity of should be incorporated in the questionnaire/s. As most of the time you will be asking 5

For details see Malhotra 1982 and Abbi 1992 and 1995.

3. The Preparation 69 questions in the contact language, it is advisable to make the questionnaire in the contact language, with an English translation for your own use. To test the linguistic hypotheses, both grammatical and non-grammatical sentences should be included as well as those which differ semantically. It should be noted that no analysis can be based on only one set of questionnaires. Questionnaire-making is an exercise that is carried out several times during the course of the work. After the initial collection of data and preliminary analysis a second questionnaire is generally designed to test the reliability and acceptability of the sentences/phrases/words that have already been collected. This also ensures the authenticity of the data. This exercise also rules out the possibility of forced translation. The preliminary work on the questionnaire may be worked out first in the classroom, on an idiolect and then upgraded to the second set, which may be carried to the field, and still further to the third set which will eventually be prepared in the target language. Generally the second set of questionnaires is much smaller than the first one. There are other reasons for designing a second or third questionnaire. The first questionnaire is generally based on the data collected in the classroom, but it is only after going to the field that one realizes the gaping omissions in the data. To incorporate all that you have observed in the field, a second and much-modified version of the questionnaire is needed. As we said earlier, this exercise may sometimes be repeated several times, depending on the discovery of new facts and grammatical structures in the language. Occasionally one devises a questionnaire with a seemingly uniform pattern of linguistic structures for an areal study. However, while working in the field one may find conflicting or different structures. Once varying structures are found, one cannot but modify the original questionnaire accordingly. We will discuss the preparation of the questionnaire shortly, but at this point we must remind our readers not to go to the field without any questionnaire in hand. You might miss the ‘focus’ of the study (see also 3.7). 3.5.1 Goals should be Clear Before even attempting to make the questionnaire, one should be very clear about what one wants to study. The questionnaire for writing a grammatical sketch of a language will vary drastically from the one prepared for writing on a specific topic of linguistics. Similarly, if the aim is to study a particular language, then the questionnaire will be framed differently than that of studying areal linguistic features. The most difficult to write are those questionnaires which pertain to identifying areal features of the region. One can begin with the advance knowledge of areal features as discussed in chapter 2, and with the specific features of the language of the region under study; yet when it comes to actual data elicitation, the questionnaire is never sufficient to elicit complete information. Always remember that any questionnaire is only a guide to reach a language, not the path itself. Your questionnaire will inadvertently be full of linguistic features already known to you. The real challenge is to unravel the mystery of the ‘unknowns’.

70 A Manual of Linguistic Fieldwork and Structures of Indian Languages

3.5.2

Each questionnaire should be numbered.

A commonsense suggestion, nonetheless very important for sifting, compiling and computing the data later on at the analysis stage is numbering the questionnaire. Numbering can be in roman letters prefixed by the letter code of the region visited (optional). Marking the region in the top righthand corner of the page along with the number of the questionnaire proves very useful for linguistic surveys or whenever the area covered by the investigator is large. 3.5.3

Language of the Questionnaire

We will talk about this in detail in section 3.7.1. However, we would like to mention here that the language of the questionnaire depends upon the following factors: 1. The linguistic composition of the speech community 2. The language of the investigator 3. The nature of the inquiry A diglossic and bilingual community at times dictates the use of a bilingual questionnaire or one which is prepared in both languages individually. One also has to take into consideration the language of the linguist collecting the data. Occasionally, there are questions of which only s/he is aware of and which can perhaps be asked best when written in the language s/he feels at ease with. We have observed many American linguists who use English or German for the questionnaire, but take recourse to the recommended language of the area when eliciting language data (as discussed in 3.7.1 below). At times the questionnaire is prepared in a biscriptal form in order to make use of the help of literate informants (see the appropriate appendix). 3.5. 4. Length of the questionnaire As said earlier also, no questionnaire should be longer than one hour of data elicitation. we have observed that some linguists prefer 90-minute sessions with a break of 10 minutes in-between. Whatever schedule is followed, one should break the questionnaire into several parts. This helps breaking the monotony and later helps in sifting and analyzing the data. However, this restriction of timing may be relaxed in the case of elicitation of words and word list. Sentences should be short and easy to comprehend. Even at the stage V discussed later, complex sentences do not imply complexity of ideas. The field linguist should always aim at accuracy, and that is ensured only if s/he is understood clearly by the informant/s. If you are following the bilingual method, you will have to prepare the questionnaire in two languages. This might increase the length of the questionnaire but not the elicitation time, as one will try to elicit data only in one language. One may consider the questionnaire on ‘Reduplication’ given in the appendices to get an idea of a bilingual and biscriptal questionnaire. 3.5.5

Information Regarding the Informant

To get background information and a general idea about the status of the language concerned, the following information should be collected, though not necessarily

3. The Preparation 71 during the initial stages, as it might offend informants. The best course is to ask 1-4 initially and wait for the rest. 1.

Name of the informant:

2.

Primary language spoken at home (mother-tongue):

3.

Other languages spoken at home:

4.

Languages known to you: Read--------, Write--------, Understand----

5.

Age :

6.

Gender:

7.

Education:

8.

Medium of instruction at the primary level

9.

Profession, if any

10.

How long have you stayed in this place?

11.

Why and when did you migrate to this place? (if migration is involved).

This is the basic information about the informant that has to be collected. A purely linguistic questionnaire will differ from a sociolinguistic one. We have given various types of questionnaires in the appendices. Please consult them before you settle for one. As said earlier, there is no ideal questionnaire. The name of the investigator, place of the interview and the date/time of filling out the questionnaire should be recorded at all costs. Different locations can give different types of information. This is especially true of sociolinguistic inquiries, hence recording the place is important. We also make a practice of noting down in whose presence the data was collected. Informants are known to vary in their responses, depending upon who else is listening. Women members of various communities vary drastically in their linguistic responses in front of male members of their own or another community. 3.6

Contacts in the Field

Only a seasoned fieldworker knows how difficult and tiresome it is to look for the ‘right contacts’ in the field, and how much patience this requires. It is advisable not to approach a field without a recommendation of some contact person 6. Arriving at the field without the medium of a contact person or her/his recommendation letter is a risky job. People might not take you and your work seriously, or they might be hostile to you, as they may take you for an intruder or an exploiter. It is generally the latter that I am afraid of. The only way out seems to be to approach the society through the medium of the contact person. However, there are also some disadvantages in this approach, as people associate you beforehand with a certain set of ideas, beliefs and prejudices which are taken over from the particular contact person. 6

Contact person is the one who serves as a link between you and the informants in the field. This may or may not reside in the ‘field’ you are working in.

72 A Manual of Linguistic Fieldwork and Structures of Indian Languages I remember when we went to Manipur to work on Meithei, our contact person was not a very popular man and was considered politically aggressive. Obviously, our team was oblivious to the fact, but I had smelt a rat by the very first evening. Even though it was winter, which in Manipur can get very cold (that day the temperature must have been 5°C), we were served with ice-cold daal and equally cold rice with a handful of salt in it. It was too obvious a fact that the amount of salt was not just a slip of the finger but was added purposefully, perhaps to provoke us. On such occasions one has to keep cool and restrain oneself almost like a yogi. My revolutionary students of JNU were prepared to spit out the food and slam the plates on the floor, but that is exactly what I stopped them form doing. I remember no one ate anything that night and we had to sleep without food (that was after a long journey of two days and two nights). Of course, the situation improved the next day when I had a serious talk with the Rector of the University and the organizers of the mess and explained that we had nothing to do with the ideology of Mr. X., our contact person. Despite such incidents, which are not infrequent, we would advocate that one should approach the field with the recommendation of one, or better still, several contact persons to avoid the incidents mentioned above 3.7.

The Second Stage: In the Field

After reaching the field and moving into the place of residence (which was booked in advance) the first question that occupies one’s mind is food, water, and a certain standard of cleanliness. Always arrange to have your water boiled and go slow in trying out new food items that your digestive system is not familiar with. Your good health in the field is the key to the success of fieldwork. The first few days in the field are always full of charm, excitement and enthusiasm that may or may not last throughout the course of the fieldwork. It is advisable to watch the community, its members, its communicative networks, its conventions and surroundings for the first two days. The responsibility of making the fieldwork engrossing is the task of the fieldworker. This you may learn by and by as you get experience. 3.7.1

What Language/s to Use for Eliciting Data?

We discussed the language of the questionnaire earlier but now let us discuss the language to be used while conversing with the speakers which may or may not be the language of the questionnaire. However, most often than not the language of the questionnaire and the language used for elicitation are the same. Let us make a distinction between the target language, the contact language and the meta-language. The target language is the language being observed and analysed. The contact language is the language that is understood by the fieldworker and the middlemen (e.g. interpreters) at large, or by the people who speak the target language. Mutual intelligibility of the contact language by the fieldworker and by the people of the community obviously renders interpreters unnecessary (which is the ideal situation). In a multilingual Indian society, it is not impossible to have the situation where fieldwork is conducted without the assistance of a middleman. The meta-language is the language in which the analysis and final writing will take place. This is generally English or any other standard language. Target language ï Contact language Meta language

ð Elicitation Þ

Analysis

Þ

3. The Preparation 73 The contact language will help elicit the relevant information, which will serve as the input for writing the results in the meta language. The language of elicitation either by the questionnaire method or by direct questions should preferably be the contact language with English translation. The widely spoken regional language is the best choice (most often than not Hindi-Urdu; the reasons having been discussed in chapter 1). Thus, while working on Kui one can afford to prepare the questionnaire in Telugu, but if one is working on Kurux then it is advisable to prepare questionnaires in Hindi, Bangla or Oriya, as the case may be (depending upon the region of study, as Kurux it is spoken in Jharkhand, Bangal, Orissa and Madhya Pradesh). This suggestion holds for the first round of data collection. As the work progresses, we prefer to elicit data in the target language. This is not difficult, as soon after the first three stages discussed in the next chapter, investigators can converse in the language of the informants. At this stage, it is not impossible to prepare a second (or third if we count the one used in the class room situation) questionnaire or elicit directly in the language of the informant, if the questionnaire method is not followed. Whatever language the investigator chooses, s/he has to frame sentences in a particular context. Sentences are always difficult to elicit out of context. One may also consider modifying the sentence according to the location and context in which it is being asked. There are two kinds of contexts we have in mind, (1) pertaining to the social environment and (2) pertaining to linguistic construction. The religious and hierarchical patterns existing in the society should guide you to frame your sentences in such a way as to avoid offending someone’s sensibilities. For instance, it is absurd to ask a Hindu ‘How would you address your second wife?’ or to try to elicit from a vegetarian Hindu ‘How would you say “the meat did not cook well today”. Or even asking a woman the name of her husband (especially in a village setting), as traditionally women are not supposed to verbalize the names of their husbands in conservative Indian society. As to the linguistic context, two factors should be kept in mind. Firstly, the lexical items in the sentence used for elicitation should be drawn from the surroundings. Thus, a sentence like ‘my car is blue’ is absurd in a village where no one has ever seen many cars. Similarly, a sentence such as ‘this is ice-cold’ will fetch inquisitive glances from your informant if spoken in a desert area such as some areas of Rajasthan. Secondly, one should frame sentences that appear normal and occur frequently in the language. Thus, ‘Ram goes’, ‘Sita goes’ are not good sentences, as these are unnatural sentences never used in normal discourse. It is better to say ‘Ram goes to X (use a name familiar to the informant)’ or, ‘Sita goes to X every day’. Always try to use proper names drawn from the society. This will put the informant at ease immediately and give her/him the feeling you are talking about her/his people. Often, the contact language is of no help, and in such cases be prepared to use gestures. However, this should not be used very frequently. As we are not talking about the monolingual method of data elicitation in this manual, questions regarding gestures seldom arise. It is also observed that if fieldwork is being carried out among different caste groups, changing some of the language forms according to the specific situation is helpful. One must draw attention to the typical bilingual nature of Indian community, where elicitation at times takes place in code-switched language; one being the target language and the other being the contact language. The language of elicitation will depend upon the linguistic configuration of the speech community

74 A Manual of Linguistic Fieldwork and Structures of Indian Languages under investigation. Longitudinal studies at times require different languages for different informants. Fieldwork in India, we feel, is not only challenging, but also dynamic and thus far from boring. This brings us to a very relevant question. Who should we choose as our source of information, i.e. informant? 3.8.

Choosing Informants

Much has been written about how and who to choose as an informant (see Kibrik (1977), Nida (1946), Pike (1947) and Samarin (1967)) in the linguistics books so that one need not repeat those instructions. However, for the Indian context following principles should be kept in mind. 3.8.1. Begin at School Try to have as your first informant teacher or a clerk or peon from a primary or middle school. Our experience shows that school-teachers turn out to be the best informants, as they are generally bilingual in the target and contact language, hold an appreciative view of linguistic work and can explain the historical and social makeup of the society under consideration with great clarity and conviction. Occasionally, schoolteachers are a bit too talkative and also too willing to give more information than needed, but their helpful attitude and knowledge of the language offset these minor drawbacks. Another advantage of visiting a school is that one is introduced to a large number of students and teachers (both male and female) who can subsequently be very rich sources of information. This is made possible as a schoolteacher considers a fieldworker as a ‘fellow teacher’ and simple comradeship motivates her/him to invite fieldworkers home. My personal experience of more than 23 years is that it has never taken more than a week for me to be invited home by a teacher to interview her/his family members. The first contact in the field is of great importance because much depends upon the ‘contact person’ who serves as the mediator between the fieldworker and the informant, as we discussed earlier in this chapter. Teachers in general are respected in villages by the Sarpanch (the headman), Jamindars (the landlord) and by other people in power. This fact should be borne in mind so as to ensure a smooth operation in the field. Also, since teaching is one profession which is respected by housewives and powerless people in villages and cities alike, a teacher as your contact person in the society would not be turned down and in turn your own work will be regarded with due respect. Through local teachers one can be introduced to a large cross-section of the society. In rural settings, teachers have been of immense help to my team and me. On account of the commonality of the teaching experience, the sense of taking part in uplifting the society, the sense of pursuing the most noble profession recognized the world over, any rural teacher takes no time at all to establish a bond of friendship with me that often gives me the opportunity to participate in their social, political and above all personal lives. What more could a fieldworker want in the field?

3. The Preparation 75

3.8 2. Avoid a Language Teacher One should try to avoid a teacher of the language under study in the earlier part of the fieldwork to save oneself from being lectured on the grammar of the language concerned. Language teachers generally have opinionated views about the target language which are useful to collect, but not in the very beginning of data collection. Our personal experience shows that most of the language teachers give out the entire paradigm of verbs, for example, exactly on the basis of the English grammar or Sanskrit grammar that they are teaching in the school. This way languages that do not make many distinctions in the aspect or tense categories, as we saw in chapter 2, are rendered unnecessarily as if they have one in the English or Sanskrit pattern. The situation becomes worse when you resist following their instructions and try to apply your training in linguistics. If time permits, one should visit a language teacher after the preliminary analysis of the collected data is achieved. We would like to share our personal experience with the readers here. When we were working on explicator compound verbs as an areal phenomenon, a topic that demanded the understanding of minute semantic differences and varying pragmatic uses, we found the help of language teachers indispensable. The effortless explanation that a language teacher can offer for semantically complex units is worthy to be noted. Also, it is the language teacher who quotes profusely from the literature of the language for various explanations and descriptions. But, as we said earlier, no matter how tempting it is, one should avoid them in the beginning of the fieldwork. 3.8.3. Choose Both Male and Female Informants This is undertaken to get a full representation of the society. In some societies, women have a different style of speaking than men and still others offer two separate grammars, one for women and another one for men. See the Kurux examples given below, where female speech is marked by length as well as by compound formations. In addition to the fact that women should constitute a large body of the sample, one should not forget old women. Old women make good informants for two reasons: (a) being mostly house-bound they seem to preserve the specialized lexicon, and (b) they tend to use archaic structures in their discourse more than their male counterparts. Table 3.2: Male and Female Speech in Kurux Male Speech

Female Speech

Gloss

singular

plural

kukkos

kukkor

kukko: xaddae

‘boy’

kukkoe

kukkoer

kukkoe: xaddae

‘girl’

pacgis

pacgir

pacgi: a:lae

‘old men’

pacco

paccor

pacco: mukkae

‘old women’

nalus

nalur

nalu: a:lae

‘male dancer’

nalunalur be:cus

plural

nalu: mukkae be:cur

be:cu: a:lae

‘female dancer ‘male player’

76 A Manual of Linguistic Fieldwork and Structures of Indian Languages Language and gender has been a popular topic for research in the last two decades of the last century. Women’s and men’s speech differs not only in grammatical constructions, but also in the choice of words and sentences, style, tempo, and in various phonological aspects. Leaving them out of your investigation would result in a skewed analysis. It is also important that women-to-women’s, mento-men’s and women-to-men’s speech be recorded. Many societies maintain different registers for all these variations. 3.8.4. Choose All Age Groups, But Not Younger Than Twelve Though it is debatable still among linguists at what stage a child acquires the full grammar of her/his language, we make a safe policy of interviewing only those who are above twelve years of age, i.e. from their teens onward. The speech of teenagers as well as that of old people will make good samples in your corpora. Each has its own merit. The speech of teenagers seems to be highly converged, especially if the society is bilingual, which we are afraid is a norm more than an exception. The speech of very old people of 70 and above, on the other hand, seems to retain some of the indigenous grammatical structures and lexicon which have become extinct in the younger generation. While working in the Bangan area we came across some very archaic words in the speech of old people we interviewed7. These words could be traced back to the PIE stage (see Abbi 1997, Zoller 1993). We shall illustrate some striking examples that we acquired during our fieldwork. Bangani / kur! or kur / ‘brave, valour’ < OIA /s´u:ra/ ‘hero’ /g sti/ ‘guest’ < PIE /ghosti-/ ‘a guest’ /l!kt!/ ‘milk’ < PIE /glak-/ Latin /lacte/ ‘milk’ /p!rk!/ ‘ask, question’ < PIE /p(e)rek/ ‘to ask’ /dukti/ ‘daughter’ < PIE /dhugh2ter/ ‘daughter’ On checking these words with young informants, we were confronted with infinite giggles and were told that we were trying to speak like their grandparents, as no one in their peer group used such words anymore. Similarly, when we visited Shillong, the capital of Meghalaya, again after a gap of 21 years and wanted to check on the ‘lamino-dentals’8 (discussed in detail in the next chapter) and retroflex sounds in Khasi, a Mon-Khmer language of the Austro-Asiatic branch, old informants came to be very useful. They had all retained the former and not a single one of them used the latter. The young people, on the contrary, varied in their use of lamino-dentals but invariably had started using retroflex sounds, which were not part of the earlier Khasi phonology. It is imperative to take into account all age groups to study language change phenomenon.

7

Two of them, a woman and a man, both were monolingual. Unfortunately, our male informant died last year. 8 These sounds were first discovered by us while working on the language in JNU, and later verifying their existence when we visited Meghalaya in 1978.

3. The Preparation 77 Needless to say, if one is working on language acquisition or specifically on the difference of language variation between adults and children one has to take children into account as field informants. 3.8.5. Choose All Sections of a Stratified Society To get a fairly representative sample of the society, one has to draw her/his data from all social and economic classes in both rural and urban areas. This becomes especially important if one is conducting a sociolinguistic inquiry into the speech community. Avoid a multilingual informant in the beginning of your data elicitation, unless you are interested in studying the ‘domain analysis‘ of the target language visa-vis the contact language. If one is working on the grammatical structure of the language concerned, then it is better to find an informant who is bilingual in the target and the contact language. To achieve this try to locate an informant who has not traveled widely. Always keep in mind that rural and urban areas as well as big cities offer a distinct organization of the society (see chapter 1), and study of each kind and type is the only way to get representative data. Highly structured and stratified societies like ours pose a challenge to a fieldworker. Nonetheless, they provide interesting insights into the dynamics of language in society and the language of society. 3.8.6. One Willing Informant is Better Than Ten Unwilling Ones Fieldwork is like a joint project that is conducted only with the cooperation and support of the informants. An unwilling informant not only has the possibility of giving no information but also of distorting information, which may be dangerous and misleading. While working on Khasi, the only non-Munda language of the Austro-Asiatic branch spoken in India, one male college student was sent to us to be our ‘informant’. When we started our interview sessions of eliciting words from the basic word list, he refused to give us the kinship terms on the excuse that these were too ‘personal’ questions to be answered to strangers from a distant land. When we started inquiring about the other words, especially the words from the noun class, most often his response would be “I have no word for this in my language”. After a couple of sittings it was more than obvious that he was an unwilling informant. We had to subsequently drop him from our list of informants. You might find a willing informant but not always in a mood to cooperate. We all are humans and data elicitation is a boring job for an informant. S/he just has to sit there and keep translating what you require her/him to translate. In such cases one has to work as if you are working with children. It is better to crack a joke or tell the informant some amusing story to set her/his mood right before starting on elicitation. Informants generally get tired and bored and are condescending in interviews. Beware of their being over-helpful in translating model sentences word for word and thus depriving you of seeing the distinctive structure of the language under investigation. When we were working on Kashmiri, the Indo-Aryan language spoken in Srinagar and the surrounding area, most of the informants we interviewed were overly generous with their rendering of Kashmiri sentences which appeared to be exact translations of our model Hindi sentences. Our brief knowledge of Kashmiri

78 A Manual of Linguistic Fieldwork and Structures of Indian Languages structure made us aware of the fact that something was amiss. For instance, the elicited sentences also had verbs as final elements while we knew that Kashmiri is a SVO language. Later we had to collect the data anew from a different set of informants. Overly generous informants should be dropped from your list. It would not be surprising to find informants answering to every question that you pose “we also say exactly like this word to word” and then s/he translates verbatim each and every word you said in your contact language. Fieldwork is like running an experiment where the researchers have not only to conduct the experiment again and again to ensure the validity of the data, but also have to be alert about receiving too neat a pattern which may reflect errors of elicitation. Self-defense operates strongly and thus one has to make allowance for the human tendency to rationalize, project and sublimate their desires, wishes, dreams and fears. Kibrik (1977: 54-56) mentions following features of a good informant which can be added to the features discussed above. 1. Good knowledge of the target language. 2. Good knowledge of the contact language. 3. Translating talent. 3. Clear pronunciation 4. Associative mobility of thought. 5. Patience. 6. Honesty and a lack of feeling of ‘linguistic prestige’. 7. Strictness 8. Experience in working with the investigator (this quality cannot be expected for the first time informant). 9. Linguistically not sophisticated As can be seen, some of the features listed by Kibrik have already figured in our discussion above. What one must remember is that there is no ideal or optimal informant. One has to collect data from various types of informants to get the right view of the target language. Just as there is no ideal or optimal informant, there is no optimal number of informants. It depends on the nature of the study. A sociolinguistic study demands a very large number of informants, while writing a grammar or doing typological work requires a much smaller number. We generally interview a minimum number of 3550 informants in the field for a non-sociolinguistic study and at least three when one is not fortunate enough to go to the field. For sociolinguistics study, we take a minimum of hundred informants into account. Avoid working with a single informant, unless the language is on the verge of extinction and you have been introduced to a terminal speaker.

3. The Preparation 79 Also, the nature of the linguistic work generally determines the nature of the informant. Investigations into a purely synchronic grammar of the target language or into the historical aspect of the language or into the sociolinguistic nature of the language will each demand different types of informants. The field investigator has to use her/his discretion as to the selection of informants in the field according to the linguistic requirements. Though a large number of informants are always justified by all field linguists to avoid an ‘idiolectal’ study, one has to choose and rely on one key informant or set of key informants to build the corpus. The key informant/s can either be in the field or in the city you came from, but this should be a person/people who is/are easily available to you from time to time as well as having most of the features listed above for being a good informant. The key informant is the person one has to run to again and again at the ‘analysis stage’, when one is required to question and validate hypotheses. The key informant thus should be one who can converse in and understand your contact language and has enough intelligence to comprehend your questions. A fairly intelligent key informant is a real boon to the fieldworker. 3.9

The Role of the Interpreter

An interpreter is needed when the informants do not understand your contact language. The selection of an interpreter should be undertaken with care. Sometimes, one might be lucky in finding an experienced interpreter, otherwise one needs to train an interpreter before one starts eliciting data. The interpreter should be told in advance about your requirements, how much information is needed, how one can collect variant sentences with varying meanings, or different types of sentences with pragmatically determined contexts, or sociolinguistically determined constructions etc. The interpreter should be reliable in rendering your questions clearly and without adding any colour from her/his side. While working in remote areas an interpreter almost becomes your guide and assistant. All those qualities that we discussed above about the ‘contact person’ will apply to the interpreter also. Hence the prestige accorded to her/him by the local people will definitely affect your fieldwork. Our experience shows that there is a general tendency to rely upon the help of the interpreter for non-academic needs, too. The interpreter, thus, may become a good friend of yours whose friendship you might cherish even after the fieldwork is over. Fieldwork not only teaches humility but also gives you friends for life. 3.10

Your own behaviour in the field

3.10.1. The Investigator When the investigator reaches the field, the obvious first question that occupies the mind is how to address the local people, young and old, women and men. In a typical situation this information should have been collected from the pilot informant before coming to the field. If not, one can be assured of being treated appropriately if the investigator addresses old men by the kinship term for ‘uncle’ or ‘grandfather’, and younger men by ‘brother’, depending upon their age and the difference between their age and the age of the investigator. Similarly, for women addressees one could use the terms for ‘aunt’, ‘grandmother’, “sister’ etc. Such address forms immediately put the local people at ease, even when they encounter you for the first time in life. Similarly, prepare yourself to be called by these very address forms in return, depending upon your age and attitude (though this stage comes much later when people trust you and know you better). In return, initially local people might like to

80 A Manual of Linguistic Fieldwork and Structures of Indian Languages call you by your name plus the address term for ‘elder sister/brother’ as the case may be. Some informants might call you “professor saab” if you are a teacher, irrespective of the fact whether you are a woman or man. It is seen that female informants make use of some kind of kinship term faster than their male counterparts. I am always addressed as didi ‘elder sister’ or its equivalent in their language in the field. Needless to emphasize how easy the stage is set for elicitation once two interlocutors who had never seen each other before, and happen to be investigator and informant, establish this kind of rapport. Fieldwork involves long hours in an alien culture, with people with different attitudes and point of views. One must be patient and eager to learn rather than denouncing what is not acceptable. Be a student and not a teacher or a preacher. Be appreciative of what you see and experience. People read your rejection or discomfort much faster than you would imagine, which may consequently be harmful for eliciting data. Moreover, try to build a sixth sense of reading relaxation, acceptance, rejection, resentment, hostility, or neutrality in the behaviour of the members of the community. To achieve this one needs to be humble, observant, perceptive, and friendly. Try to share the joys, games, sports and cultural festivals of your host community, if permitted. These activities are not only enriching but also give one a mute sanction to be a participant in the local life. We would like to share a page from the diary of one of our students who attended the Christmas mass in the city of Chotanagpur. “It is Christmas eve and we have been invited to attend the Midnight Mass at one of the city’s big cathedrals. Well over 80% of the congregation is made up of Mundas, Santhalis, and Oraons (with the latter in maximum numbers). Mother Mary here has been completely tribalized. She wears a white and red saari, tribal style, has colourful feathers stuck into her bun and has baby Jesus tied around her waist. What catches my eyes is the expression she wears in her eyes: eloquent, stoical and even a trifle mournful. I have seen this very expression in the eyes of almost all the Oraon women I have met. I cannot get over the beautiful Oraon Mother Mary. The Mass begins. I let my eyes wander. The congregation is by and large poor, with some hovering around the poverty line. Two little children in tatters sit right in front of me. Then a little boy comes handing copies of carols to be sung; each and every hand, middling poor, poor and very poor reaches out for a copy, and each and every eye is able to read them. I cannot help but think of the missionaries whose tremendous efforts have made this possible. Christmas carols and hymns are all being sung either in Mundari, Sadari, or in Kurux, accompanied by the maddeningly exciting beats of the Maadal (a typical drum which looked like Dholak, extensively used in the entire region). My feet are dying to step out and dance, so lively are the carols. But for the Christian lyrics, the tunes and the rhythms are unabashedly “tribal”. Again my eyes wander to Mother Mary so completely appropriated by the indigenous culture, and I wonder whether the indigenous culture of the tribals has really been dealt a

3. The Preparation 81 ‘fatal blow’ by the coming of the missionaries, as some would have it. I can’t help but think otherwise. If anything, that culture is tantalizingly alive right here and now in this Cathedral, in the beats of the Maadal, in the tunes of the carols, in the eyes of the Virgin and in the hundreds of voices joining the choir and singing alone lustily”. .

Priyanka Bhattacharya’s field diary 1998. The readers should not get the false impression that in every fieldwork that they conduct, they will be allowed to participate in the local festival. The participation and its opportunity will depend upon several complex circumstances of the attitude of the members of the community and the readyness of them to accept the fieldworker. Our experience is that the people of northeast India, the speakers of the Tibeto-Burman languages and Khasi do not allow participation very easily. This should not be taken as criticism. The reasons for this are much deeper than can one analyze, as the long history of neglect and isolation from the government machinery has forced these people to live their own lives without much encouragement for intermingling with the rest of the Indians. In fact, some of the northeastern tribes think of ‘India’ and ‘Indians’ as not a part of their country and countrymen9. Sometimes, as said earlier, the attitude of the investigator her/himself determines the outlook of these people. I remember landing in the midst of the land of the rising sun, i.e. Kanchipur of Manipur in the winter of 1984 to work on Meithei, a language of the Tibeto-Burman family with a bunch of very enthusiastic students. Those were the days when insurgency in Manipur was in full swing, the campus of Manipur University was guarded very heavily by armed BSF (Border Security Forces), shops had to be closed by 4 p.m., and the local people always looked at us with great suspicion. Even with the best of my ‘know-how’ of conducting fieldwork in remote areas I had not been very successful in locating willing informants and willing households to let us in. I wondered why? I could not help asking a member of the household about the reason for such a ‘cold’ welcome. He pointed towards a jeep waiting outside the premises which had a driver all wrapped in a blanket. On inquiring I found out that he had been sent for our ‘protection from the violent tribals’ and was equipped with a rifle hidden in that blanket. That explained the indifferent behavior and cold attitude of the people. Never accept help from the outside unless you are in dire need of one. To be a participatory observer one has to be first accepted by the society. You should never be the one who is feared. In short, the field investigator should be friendly, understanding, humble, ready to be accepted by the society and above all should give the impression that s/he is there to learn and not teach. Fieldwork, in whatever part of the world, allows glimpses of different lives. The intrigue and complexity of social relationships, people’s humour and anger, as well as their hospitality and antipathy are worthy cherishing. You come home enriched.

9

Interestingly, I have time and again heard of this sentence in the northeast region: “We were never part of India, don’t call us secessionist.”

82 A Manual of Linguistic Fieldwork and Structures of Indian Languages

3.11

Being A Woman Is A Blessing

Whoever thought that fieldwork is not for women is wrong. On the contrary, being a woman has been a blessing for me, especially while working with strangers in the remote areas of this country. The initial response from the members of the society may be sceptical, as people do not expect a woman to tread the interiors of an unknown land, but this scepticism dies very fast and soon is replaced by a friendly and protective attitude by the male members of the society, especially. A married woman and a mother is respected and honoured all over India, especially in villages. This fact alone has given me social acceptance. People immediately regard my work seriously, as ‘no married woman with children will roam around villages unless it is very important’. This attitude has time and again given me the kind of entry into the society which my male colleagues find almost impossible to achieve. It has been observed that male members of the society are always eager to introduce their female relatives to me. This serves as a double boon. On the one hand, I get to know the social group which is considered most difficult to interview, as well as being introduced to the innermost part of the household where all the activities take place, i.e. the kitchen. I observe that it does not take more than 20 minutes of conversation and soon I am invited to take a seat right in the kitchen while the lady of the house cooks. I always make sure that they do not miss their household chores because of my presence, as life in village is tough and women do not always have spare time for conversation We were working on Bangani, a language of the Indo-Aryan family spoken in western Garhwal of the Himalayan region, in the month of September when we realized that we had not chosen a very appropriate season for fieldwork. The women (except for the very old ones) had to get up very early in the morning so as to complete their morning chores and then would take off to the fields to cut grass to be stored for the cattle for the winter season. On their return from the fields they would again get busy cooking for the entire family. This would hardly leave any time for elicitation of data. We had to practically sit in the kitchen every evening to do so. Women share recipes very willingly and I take advantage of this fact on every fieldtrip. The fringe benefit of being in the kitchen is that you are always served hot meals and tea, a privilege my male colleagues are always deprived of. Contrary to this pattern of women working all evening, male members of the society return from the work in the field and prefer to drink while the dinner is being cooked. We found this uniform pattern in almost all villages, be it in the Himalayan ranges or in Chotanagpur of Jharkhand (Ranchi district, Central India) or in the northeast. However, in Ranchi district, we noticed that women were very fond of drinking, almost to the extent that some of them would agree to be interviewed only if we promised them a bottle. Of course this is not a norm. When male members drink, they generally do not want to be disturbed by the female investigators. Yet, in Bangan I realized that this was not a taboo. One elderly and respected man was willing to be interviewed while he was drinking. Not only did he start talking and drinking, but he also invited me to join him. In such cases I generally pretend to drink lest the informant is offended. My experience is to avoid circumstances of drinking-time as neither can one be sure of what one is served, nor

3. The Preparation 83 can one discuss matters of importance with the informants. As a woman, I use this time to be with the women and children of the family, as they are always found to be relaxed and very cooperative, especially after dinner. In the matrilineal society of Kerala as well as among the Khasi (who are in addition, also matriarchal) it proved to be of immense help that I am a woman. On our arrival at Shillong (1978), I realized that being a woman is a blessing. Not only were the doors of every house opened for me, but I was respected for being a university teacher, educated and interested in their language. Khasi women exposed me to a world not known to me until then. Not only was I taken to their annual dance festival, but I was also introduced to their haat (bazaar) in Cherapunji. On the other hand, I remember my male students had several problems locating informants or convincing them to be cooperative. Ultimately, I had to intervene and help these students locate and convince informants, as an assurance given by me to the Khasis (that we were there to work on their language and write a grammar and nothing else) was taken seriously. We discovered another reason for the non-cooperative behaviour of the Khasis towards male investigators. Being a matriarchal and matrilineal society, according to the local custom youngest daughter of the family inherits the family property. Local people were afraid of the male members of our team, lest they get interested in their daughters and the property would transfer to an alien. While we were there, a common topic at various tea stalls used to be whose daughter has married a person from outside. Non-Khasis are looked at with suspicion. This feeling of suspicion against the outsiders actually runs throughout the northeast region. This is an important fact to be borne in mind as nowhere else in India does one encounter this unnerving feeling. In general, men in India are protective of women, especially those who have come to them for educational purposes. This attitude of theirs has helped me immensely in getting to know them, their women, their daily life patterns and ultimately the language in the real setting. I cannot forget the innumerable number of hospitable dinners, lunches and snacks that I was offered, from special rice cooked in blood in Cherapunji, or five different types of preparation of fish in Kanchipur, or special bread filled with poppy seeds in Bangan, or special wine in Chotanagpur, or fish cooked in coconut water and fenny in Goa, or sweet preparation in Bargarh in Sambhalpur, or simply a coarse ro i in Hanumangarh village of Kole. Indian people in general are hospitable and a female member of the society who has come from a city brings out this trait of them to the full.

Chapter IV Elicitation 4.1. Various Methods Having reached the field and identified informant/s, key informants or interpreters as the case may be, one is confronted with the real task, i.e. the elicitation of linguistic data. In this chapter, we shall discuss various methods of elicitation, followed by a discussion of various techniques of interrogation. We have tried our best to give suggestions as what to do and what not to do while in the field. In the following pages, we discuss briefly the entire task of data collection divided into different stages. However, we shall also discuss stage II in detail in this chapter. For this, we will show the reader how to transcribe the spoken language s/he is working on and how to reach a tentative judgement about the phonetic and phonological structure of the language. The chapter ends with a summary list of morphological topics that one has to cover in the ensuing course of investigation. Linguists generally adopt four types of elicitation techniques: 4.1.1 Observation Method. This requires fairly good advance knowledge of the language under consideration. While adopting this method, it is assumed that the basic grammar is already known to the fieldworker, yet to refute or validate hypotheses one can adopt this method to collect further evidence. Linguists working on sociolinguistic topics also adopt this method, since asking questions regarding various linguistic forms and styles which are socially determined may lead to grave problems (see 3.3.2). The observation method involves the participation of the field investigator in various social activities of the community. S/he may not play an active part in the day-to-day life, however, s/he will have to engage herself/himself in some social activities of the community life. 4.1.2 Interview Method. This is the most common and widely used method for field investigation, and is strongly recommended. The field investigator can interview informants with or without the help of questionnaire/s, (preferably both). We shall discuss the formulation of the questionnaire shortly. Interviews are generally taken directly. However, if no common contact language is known to the informant and the field investigator, interviews can also be conducted with the help of an interpreter. The style and format of the interviews will play a major role for the quality of the elicitation. Questions should be fairly simple and clear (even when you are working on the most complex structures of the grammar) and you must give plenty of examples. Illustrations, pictures, and diagrams can all be used in the interview method. Make plenty of use of a slate and chalk in village settings.

4. Elicitation 85 4.1.3 Sending Questionnaire Method. This method involves giving or sending questionnaires to the informants and asking them to fill them out. This is not a very reliable method, as it cannot ensure any control over the object of the study. Informants may or may not take the exercise of filling out questionnaires seriously. However, this method does ensure a large corpus within a short period of time, and we have noticed that it is quite popular with the NRI linguists. Questions like acceptability, grammaticality, variability of linguistic constructions, or some specific objective type of questions leading to Yes/No answers can be taken care of by this method. This method has a very limited range of uses and is applicable only in the advanced stages of your research. Since it depends upon the ability of the informant to fill out the questionnaire, one is also constrained by the necessity of looking for an intelligent and disciplined informant. This method is also used for confirming elicited data to test its reliability. 4.1.4 Documentary Source Method In the absence of real speakers of the language, either because they no longer exist, as the language is extinct, or because they are out of the bounds of the investigator, the documentary source method is adopted. One has to rely on published material and make a trip to several places where such material is available. Sociolinguists and pragmatists may find this method fairly useful as they can deduct relevant information from different genres of literature available. Historical linguists, for obvious reasons have always relied on this method. In field linguistics, we generally combine this method with the interview method. Research on the ‘sociology of language‘ can also take recourse to this method. Linguists working on demographico-linguistic research can adopt only this method, as their sole search for various parameters are located in published census reports, year books, and other related manuals. Out of these four methods, the most reliable and complete is the interview method. Charles L Briggs, ‘Learning How to Ask’ (1986) is an extremely useful guide in this area. However, no method is complete in itself. At the end of the study it dawns upon a field linguist that s/he has made use of all four methods, as they are complementary in nature. Adopting all methods can reinforce the judgement of the field linguist as well as help her/him in extending the study beyond the possibilities of the interview method. We would now like to focus on the most interview method, as it is the most significant, and discuss its various parameters to ensure that the reader is fully prepared for eliciting linguistic data from the informants. 4.2.

Interviewing Informants

Following points should be kept in mind while interviewing people in their local surroundings: 1. If you are a man, it is advisable to first approach the head of the household to seek permission to interview the family members. If you are a woman, it is advisable to approach women (of any age) first. It has been noticed that in the case of femaleto-female interactions encountering your own age group first helps in establishing

86

A Manual of Linguistic Fieldwork and Structures of Indian Languages contact. The same strategy might not work for male investigators, as bypassing the head of the family might not be taken as polite behaviour. Men do not mind if they are by-passed by a female investigator as these are expected to approach and talk to people of their own sex. 2. Always begin by describing your aims and objectives to the informant. You might want to say that you are studying their language and writing the grammar of it. Do not be perplexed by the reaction “but my language has no grammar”. You may also explain that you are interested in documenting the language, as it is part of the national heritage. 3. Start with direct questions regarding the background of the informants followed by indirect and other linguistic questions clothed in simple sentences in the contact language. 4. The location of the interview plays a major role in the elicitation of data. Try to avoid public places such as dhaabaas, tea stalls, hotels, pubs, school-grounds etc. in the early stages of the work. Public places are always crowded, and you would have a tough time keeping away the seemingly interested onlookers. You would find that during the interview there are many people who are willing to oblige you, even though you asked only one person. It is recommended that all initial interviews be taken in fairly isolated areas so that both the investigator and the informant can hear each other clearly without any type of channel disturbance (especially if you are working on phonetics, phonology and the lexicon of the language). On the other hand, at an advanced stage of elicitation and for sociolinguistic studies the same public places that we are advocating against are the typical and ideal places for interviewing informant/s. One is, at times, thronged with various opinions which might be very useful while analysing stylistic variation, grammaticality, acceptability phenomena etc. in syntax or simply the sociolinguistic patterns of the language. 5. Try to complete all your interviews before sundown in village settings and before 9 p.m. in city settings. Most people do not like to be disturbed in the evening. It is recommended not to elicit data later in the evening unless suggested by the informant. 6. In formal interviews, such as interviewing socially prominent people in cities, one must follow all the etiquette that the social norm of the city dictates. Introduce yourself properly and give a summary of the main objectives of the project you are conducting to the informant. 7. Do not criticise, oppose or contradict the informant while interviewing, even if you totally disagree with her/his views. Remember that you are like an archer, concentrating only on the target, and not get tempted to lose yourself in trivial issues 8. No interview should last for more than two hours per informant per day. The ideal duration for a linguistic interview is 60 to 90 minutes. Interviews should also not be conducted if the informant cannot allocate more than half an hour to you. It will be a rather futile exercise to elicit any data in such a short time.

4. Elicitation 87 Short interviews can take place only at an advanced stage of analysing the data for reconfirming your hypotheses. 9. One may tape the interview provided the informant does not object. While recording, one must not switch the tape off again and again during the pauses as this distracts the informants. 10. If you are using the questionnaire technique, you might want the informant to repeat the words or sentences several times. In such cases warn the informant in advance that s/he should not be offended when s/he is asked to repeat the spoken utterance. 11. You should make a note of all possible features (linguistic and non-linguistic) which come to your notice and which might be useful in writing the report later. This can be anything from the attitude of the informant, the surroundings of her/his dwelling, to her/his verbal interaction with other members of her or his society while s/he is interacting with you. After all, everything cannot be captured by the tape recorder. 12. One should always be prepared for the next interview. One should not be looking for a question at each subsequent visit. This implies one should do some basic analysis each night so that one is fully prepared the next morning for what to ask and what to expect. It is useful to keep a list of questions and topics ready that one should keep revising as the interviews progress. 13. If one is working on sounds alone, it is advisable to rent or arrange for a sound-proof room where the data elicitation can take place. Simple palatography can be conducted in any room which is away from the usual noises of the surroundings. For recording sounds in isolation or in words, switch off all electrical appliances, especially fans, before you begin recording. Remember to make a note of the various environmental noises that your ears can hear in your note-book so that when you have left the field and start transcribing the spoken words you will be duly prepared to sift out the unwanted background noises. The interview method proves to be very useful as it gives you an insight into the society of the informant. It has been observed that while engaged in conversation, informants generally volunteer information regarding their people, knowledge about who should be met and who should be avoided, or information regarding how the language has changed over the years etc. This is one method which generates friends (and also foes) in the field and durable relations. 4.3.

Interrogation Techniques

Interviewing people to get any kind of information has become an indispensable part of our life. Interviews are ubiquitous in modern society. It has been observed that interviewers rarely examine the compatibility of interviews as a means of acquiring information with the ways in which their subjects typically convey information to one another (Briggs 1984). Improper interviews, whether in the methodology or in the content, lead to “communicative blunders” (as discussed in the Bangani episode in the last chapter). As interview methods plays a crucial role in social science

88

A Manual of Linguistic Fieldwork and Structures of Indian Languages research in general and in linguistic research in particular, We would like to discuss some of the significant techniques of conducting interviews. A fieldworker may adopt one or several of the techniques suggested below. 4.3.1 Translation Translation of the words, phrases, and sentences given in the contact language with or without the help of the interpreter. This is the most widely and commonly used method. For this, one has to identify informants with good competence in both the target language and the contact language. As is already evident to linguists, translation is no mean job nor can every word be translated. Occasionally, one has to explain the situation of the utterance and the probable identity of the speakers who could have uttered the model sentence to the informant before s/he begins translating a text. Translation of words requires examples of similar kinds to make the informant understand what you want. Translation of sentences requires an explanation of the context in which the sentence could be uttered. Translation of phrases requires the explanation of the situation of the discourse where such phrases could be used. The fieldworker has to be very alert in this technique of elicitation, as the informant might give a wrong translation, not because s/he did not know the exact word but did not comprehend the model word in question. This problem is rather acute when working on deictic categories and pronominals. Languages seem to represent referential categories in different ways, and the fine distinctions that they make and perceive might be totally missing in the contact language. Santhali makes a large number of distinctions in locational pronouns, which do not exist in Hindi. Hence one Hindi word such as v«haâ ‘there’ might have several equivalents in Santhali as they perceive Hindi v«haâ as polysemous. For instance, Santhali distinguishes between distances which are within the visibility limit of the speaker and those which are not. Thus n ‘there (visible), and h n ‘there (remote, non-visible)’. The following statement from Charles Briggs (1984:50) is very appropriate in this context. Both interviewer and respondent may share a common interpretation of the referential meaning of what is said and yet may differ widely on their interpretations of indexical meanings. This can create severe problems for researchers who have mastered the semantics and syntax of another language without gaining sufficient competence in subtle indexical function. There is no exact translation of generic words like ‘father’ and ‘mother’ in Kurux, as all kinship terms in the language are accompanied by a genitive nominal or pronominal. Thus it could be ‘X’s father’, or ‘X’s mother’ etc. Whenever, an informant replies that there is no equivalent of the word of the contact language in the native language, and specially if the word belongs to the basic vocabulary, do not doubt the capability of the informant to understand you but think of the possibility that the same concept might be differently represented. At such times, try to give situations, or a part of a fable or a story, which might use the model word. For instance the following conversation took place between a Kurux informant and a field investigator, where both the investigator and the informant were not getting the point and the situation became more and more tense.

4. Elicitation 89 FI:

What is the word for ‘father’?

I:

We don’t have such a word.

FI:

What? You don’t have a word for ‘father’?

I:

(meekly) No.

FI:

How do you call your father?

I:

baba.

FI:

How do you call your friend’s father?

I:

caca

FI:

(confused). But there must be a general name, I mean a word for ‘father’.

I:

No, we don’t have. Conversation piece no.4.1

This conversation is of a typical example of turning in circles without realizing the problem. The informant is made to appear stupid which should be avoided at all costs. Secondly, the investigator is on the wrong track, as by asking the informant how does he ‘call’ his father she has entered a different arena of words, which are known as ‘address terms’. In such cases, when an informant reports the nonexistence of a particular word in her/his language there could be many unknown semantic and pragmatic parameters at work of which we have no idea. If the question had been in a simple sentence such as ‘my father works very hard’, or ‘your father works in the city’, etc. the field investigator would have received the answer s/he was seeking. As Kurux has all the forms for the kinship terms used with some genitive, the denial by the informant that the term did not exist is understandable. See a small example of such cases for ‘brother’ as exemplified below. It was the duty of the field investigator to comprehend immediately her/his own mistake. Apparently Kurux has no generic term for ‘father’, ‘son’ or ‘brother’ unlike most known languages of the world. Table 4.1 Kurux Kinship Terms i÷g}is ‘my younger brother’

ni÷g}is ‘your younger brother’

ta÷g}is ‘his younger brother’

ta÷g}i ‘ his younger sister’

e÷gdas ‘my son’

ni÷gdas ‘your son’

ta÷gdas ‘his son’

e÷gda: ‘my daughter’

90

A Manual of Linguistic Fieldwork and Structures of Indian Languages 4.3.2. Polysemous Contact Language Words Polysemous words in the contact language and related semantic fields in the two languages may pose a different kind of problem. Many times while asking for an equivalent of a word the informant might ask you what specific context you want. This problem arises because words often refer to a large semantic field. One word in the contact language may refer to several semantically related words in the target language. This is a case when the word in question appears to be of polysemous nature, as the informant might render several words in her or his language for one simple cue word in the target language. The reason for this discrepancy lies in two semantic fields that might not be very different in nature but certainly have different ranges. A case in point is locative postpositions or locative adverbs. Generally speaking, the locative postpositions in Indian languages are in a compound form. If an informant reacts strangely while giving the equivalent of ‘in/on’, you know s/he is finding it difficult to identify the exact nature of the ‘location’. Encourage her/him to think of various possibilities and give you as many forms as s/he can think of. Consider the locational adpositions discovered by our team in Santhali during fieldwork when we encouraged our informant to use her imagination and give as many instances as possible. Locatives in Santhali The form r can be used both for location ‘in’, ‘on’, as well as for ‘genitive’. It can occur alone or compounded with locational adverbs (as in Hindi ke + nice ‘of’ + ‘below’), but occurs post-adverbially unlike Hindi. Table 4.2 Santhali Locatives Locative PP

Gloss

c tr

‘on/above the referent’

c tan r

‘above’

latar r

‘under/below the referent’

sama:N r

‘in front of the referent’

ku

‘behind the referent’ (not visible)

mr

dena:N r t y mr

‘behind the referent’ (boundary walls, trees, etc.) ‘behind the referent’ (not necessarily

hidden) bhit ri r

‘inside the referent’

bahar r

‘outside the referent’

j lat r

‘near the referent’ (two entities actually touching each other)

4. Elicitation 91 n p

‘near the referent’ r

‘near the referent’

sur r

‘near the referent’

dei r

‘towards’

dei t

‘along, by the side of ‘

tala r

‘between two referents’

k

‘from a source of location’

n

te

‘up to a destination’

h bic

‘up to a destination or a point of time’

a: ir

‘on the edge’

i:p r

‘on the edge but closer than a: ir

s t

‘towards (in motion)’

somoi

‘during’

j kk

t

‘during’

4.3. 3. Pictorial Representation Pictorial representation of concepts and articles is another method which is used for elicitation, especially when the contact language is not known to the informant very well. Slate and chalks should be used to draw those items and concepts which can be drawn. This method, however, has many obvious limitations. 4.3.4 Substitution Interrogation Substitution Interrogation (both paradigmatic and syntagmatic). Though this is an old structuralist method, it is very useful if the language has heavy inflectional morphology and encodes various categories. Pure substitution of a particular word in a sentence by various other words will bring in changes elsewhere in the sentence. The substitution method is highly recommended, as only this method will help you reach the right judgement about phonemic contrasts (seen later in the chapter), morphological variations and agreement features (discussed in the following chapters). Sometimes the morphological shape is highly complex, especially in inflecting languages, as it incorporates several grammatical and at times sociogrammatical features. It is the substitution of various elements within a word that will provide you with the various form-meaning combinations. We will consider them in some detail in chapter 5, but for present reference we give a standard case from Magahi, a Bihari language of the Indo-Aryan family. For instance, the following Magahi verb forms (Verma 1991) all translate into one basic meaning, but have multiple agreement features including the addressee component, too. There are

92

A Manual of Linguistic Fieldwork and Structures of Indian Languages five different AGR morphemes or morpheme complexes for the same person (subject of the sentence). The various forms of the verb represent the nature of the object and the addressee component. Consider: Table 4.3 Magahi Verb Forms ham dekha-l-i

‘I saw’

neutral object

ham (okraa) dekha-l-ai

‘I saw’

3 Object - Honorific

ham (unkaa) dekha l-i-ain ‘I saw’ 3P object

+Honorific

ham (toraa) dekha-l-i-au

‘I saw’ 2P object

- Honorific

ham (tohraa) dekha-l-i-o

‘I saw’ 2P object

+Honorific

4.3.5 Associative Interrogation. As the name suggests, this refers to inquiring about concepts, terms, meanings and other words that demand further investigation. This is done when some knowledge of the language is already acquired. One might choose a word and then ask questions around it to know about a process or event. For instance, if you have a word for chapaati ‘Indian bread’ in a particular language, you might gather further information about the dough, its process of being kneaded and rolled and then baked or cooked. This method not only gives additional related words and phrases, but also those grammatical constructions that were perhaps not known to you earlier. Associative interrogation is also used for generating paradigmatic constructions, in addition to the substitution method discussed above. Once a rule of affixation is clear, one can form newer forms by logical paradigmatic additions. In this case, one would like to confirm with the informant again and again of the prospective forms. 4.3.6. The Paraphrase Method This method is used to get various transformations. The typology of word order can be accessed by this method. The field investigator does the transformation and the informant is asked for her/his judgement on the paraphrased construction. One should always ask the informant about any change in meaning that might have occurred due to paraphrasing. Paraphrased sentences, though grammatical, are encoded with various types of information regarding the sociolinguistic or pragmatic use of the sentence. Obviously, this method can be used only after the preliminary analysis of sentence structure has already been undertaken. Knowledge about various grammatical constructions can be sought by this method, as we discuss in the next chapter, as it gives you enough opportunity to test the native speaker’s intuitive judgement of the transformed sentences. During the late sixties some linguists relied heavily on this method to elicit complex sentences and in turn received many ‘transformed’ sentences. Since the Chomskyian revolution had become popular linguists were motivated to give a simple sentence in the target language and asked the informant to change it in various different ways (Hale 1965). We personally do not favour this method as it demands too much from the informant. One requires a highly educated and intelligent informant to cooperate in this technique of data collection. As said earlier,

4. Elicitation 93 this particular method could be followed for testing word order changes and sociolinguistic information. 4.3. 7. Cross interrogation This is used to confirm or reject a hypothesis or to ensure whether the informant was right in giving the forms s/he has given. One should avoid interrogation with a critical sound or a style even if you are convinced that the informant has given the wrong form. One may begin her/his sentence by: “That day you gave this form. I think I did not record it properly, or maybe I did not understand it fully. Can you please explain it again - what did you mean by X?” Cross interrogation should be done very aptly and politely. It should not appear as if you are testing the person. Instead, it should give the impression that you want to confirm some facts already in your possession. 4.3.8 Stimulus Interrogation This is used to test hypotheses. A fieldworker deliberately makes ungrammatical sentences to check grammaticality and acceptability. The informant is asked to give her/his opinion and judgement on the stimulus sentences or word forms. Under this interrogation method, one should warn the speakers that you will be giving wrong (ungrammatical) sentences on purpose and that s/he should feel free to correct the investigator. This warning is important as it is generally observed that informants normally have a condescending approach towards the person who has come from a ‘far-away place’ to work on the language. This method will certainly put your informant in good humour. 4.3. 9. Examples and Illustrations. Informants should be encouraged to use ample examples and illustrations to explain concepts and situations, but only after you have earned some confidence and reliability with them. In this way, a field investigator gets a wider and more specific perspective of the phenomena. This technique is useful throughout the period of investigation. I generally keep a slate and chalk ready for the informant and encourage her/him to use them amply. Women in general love to draw, and at times explain several things by drawing on the kitchen floor. 4.4.

Transcription

The correct recording of linguistic data ensures correct analysis. The task of recording is done on paper by pencil and is sometimes accompanied by simultaneous recordings on a tape-recorder. As said earlier, a trained phonetician does the best recording by perception. One should avoid the temptation of generalising and fitting everything into a paradigmatic pattern in the initial stages of recording. Try to record everything you hear as accurately as possible. This brings in the question of what transcription to use for recording the data. 4.4.1. Narrow or Broad It is often asked in what form the data should be elicited and whether the data should be recorded in phonetic or phonemic transcription. If phonetic, which particular

94

A Manual of Linguistic Fieldwork and Structures of Indian Languages format or fonts should be used? Should we use IPA symbols or American symbols or conventional symbols for transcribing Indian languages? There is no one standard style that can be used for the entire work of elicitation. One begins with a narrow phonetic transcription, but soon changes it to a broad phonemic one as soon as the elicitation of the basic word list is over and one has a rough idea of the phonological contrasts and allophonic variation in the language. Narrow transcription, in the early stage of recording is necessary in order to record every minute sound characteristic, though a field linguist will be prepared for what to expect after familiarising her/himself with the characteristic features of the sound system of the specific language family as discussed in chapter two. Thus, a word spoken in isolation and transcribed narrowly would have tone markings, stress markings, juncture if any, but perhaps no morphemic breaks before these haven been established properly. As said earlier, a narrow transcription is good enough until the preliminary stage of phonemic analysis as it is quite a daunting task to keep on recording data in narrow transcription. When working on Bagri, a language of the Indo-Aryan family we heard murmur vowels and transcribed sounds with a raised h as in aH even though our knowledge of Bagri dictated us not to transcribe aspirated kinds of structure on vowels. Later, we realised we were hitting upon hitherto unrecorded sounds for Rajasthani. We could find minimal pairs with and without murmur vowels helping us to establish these as distinctive phonemes, as has been done for Gujarati, a language geographically and typologically close to Bagri. The story of the discovery of laminodentals in Khasi (see below) is similar. The first encounter with these sounds made us use a diacritic on the final unreleased consonant, as such sounds had not been heard before, and after a conclusive phonological judgement was reached we used an underscore, as in t’. Suprasegmentals like length should be avoided from being recorded in the early stage of elicitation as constant repetition of isolated words by informants generates artificial length. However, if we find minimal-pair-like structures where length seems to be the only distinctive feature, we should start marking length on the vowels. The same is true of stress phenomena, which is normally present on the first syllable of every word uttered in isolation. It is seen that juncture also creeps in where it should not when words are repeated slowly several times. These are unavoidable suprasegmentals, which might play no phonemic or sociolinguistic role. One should wait until one reaches the stage of syntax and small phrases to decide about the phonological status of the suprasegmentals. 4.4.2 IPA or American There are two sets of symbols used for transcribing data in field linguistics. One is that suggested by the International Phonetic Association (IPA) and the other is recommended by the American School originally designed by Bloch and Trager. The symbols provided by the American School are much easier to remember than those of the IPA and take care of diacritic features needed for the narrow transcription. However, in linguistic literature the IPA symbols are preferred. Nowadays IPA fonts are electronically available and easy to download from the internet, too. I would leave the question of choice of symbols to individual linguist. The Appendix 1 gives the IPA system. One can also mix the two systems according to her/his convenience. For instance, we generally adopt the IPA system without the diacritics for retroflexion. As Indian languages are heavily retroflexing we would

4. Elicitation 95 suggest not to use IPA symbols for symbolising retroflexed sounds, as they are difficult to write and may produce confusion at times (if the hook below the consonant letter is not long enough it might be read as a dental). The most common type is to use a dot below the letter. This suggestion is of course for transcribing data in the field. Once the investigator is back to her place of work she can use the IPA fonts including retroflex symbols, as we have done in this monograph. Aspiration and other secondary articulation can be symbolised with superscripts. For vowels, the IPA system can be adopted without any modifications. Indo-Aryan languages are mostly nasalised and hence nasalisation tilda should be marked above the vowel letter. Length in Indian languages is not distinguished by degrees, only by a binary feature, i.e. length appears as plus long or minus long. Thus, we generally mark length with a colon symbol (:) after the vowel letter. Long consonants, which are also popular in Indian languages, are written with double consonant letters as geminates. Most of the Tibeto-Burman languages are tonal and distinguish between three types of tones that can be marked by numerals or by the conventional symbols. I personally find the numerals easier to write. Consult the charts of common Indian languages given below. It is to be remembered that the sounds represented in the charts are the most common ones and thus are not all-inclusive. Every language might have some specific consonants and vowels which might need special diacritic marks to symbolise them. Again, consult the IPA for such transcriptions. The only piece of advice is to not get lost into the jungle of diacritics. It is very tempting to do narrow transcriptions using various symbols. However, at the stage of analysis, you might regret doing so. It is better to stick to the standard diacritics and search for new ones only if you discover a sound which is lacking in the literature. I did face this problem once. In 1978, when our team was working on Khasi, we encountered parallels of t, d, and n which were not pronounced as usual dentals against the front teeth but the tongue was made to touch the lower teeth. Well, since these sounds always occurred word-finally and were preceded by a glide, we were tempted to do some palatography. Though we tried in earnest, we did not succeed fully due to the lack of appropriate facilities A phonological analysis, however, was carried out and in the absence of any sounds of this kind being mentioned anywhere in the literature, we adopted an easy symbol of underlining [t, d, n] these three sounds (to make it look distinct from dentals). We called them ‘lamino-dentals’ as the front of the tongue touches the lower teeth (See Abbi 1978). No other language of India uses these types of sound, therefore adopting the diacritic of underlining posed no problem. Even if we think of comparing all Indian languages on the same scale of diacritic standard, underlining a consonant letter does not pose any problem. For instance, Malayalam with its five-way distinction of obstruents marks the dentals as symbolised by an IPA diacritic underneath the consonant letter [t 9], alveolars are unmarked [t], and retroflexes are marked by a dot underneath or a hook as used by IPA [ÿ].

96

A Manual of Linguistic Fieldwork and Structures of Indian Languages

Table 4. 4 Common Vowel Sounds in Indian languages FRONT

CENTRAL

BACK

UNROUNDED

UNROUNDED

+/- ROUNDED

i

u

I

U «

e

o

a (unrounded)

Table 4.5 Common Supra-segmental Sounds in Indian Languages Nasalisation

Ú

Length

[:] (for vowels), [CC] (geminates)

Tones

Û

(rising)

$

(falling) Ü

(level)

It is advisable to carry blank charts for consonants and vowels (see appendix 1) so that they can be filled out as soon as the phonemic analysis is carried out. It can be seen that during the course of elicitation one changes and modifies the already obtained phonemic output several times. For this reason, it is quite normal to carry several blank charts along.

4. Elicitation 97

Table 4.6 Common Consonant Sounds in Indian languages Places

Bilabial

Dental

Alveolar

Retroflex

Manner

vl

Vd

vl vd

vl vd

vl

S –asp

p

b

t

t

T +asp

PH

bH

tH dH

O -asp m P +asp

d

M n

vd

vl vd

d H

H

Velar

Glottal

vl vd

c

j

cH

jH

k

g

/

kH gH

N n

M mH

Palatal

M nH

Affricates

ts

dz

Affricates + asp

tsH dzH

Lateral

Ò

l Lateral + asp lH

}

Trills r Trills + asp

}H

rH Fricatives

f

v

S s

Fricatives + asp

z

Z

S

§(s ) (z )

V vH º

Implosive º

ë

Implosive +asp H Approximant w

4.5

y

Data Collection: Various Stages

Data collection is done in various stages, beginning from isolated words to small phrases to long and complex sentences followed by discourse pieces (see below). Text collection of narratives may be done at any stage after two; however, linguists vary on this chronology. Many linguists prefer to start with texts after stage one and then translate each and every word to extract phonological, morphological and syntactic structures of the language concerned. We personally wait before text collection until we have worked out the basic syntactic structures of the language, i.e. until stage three, as discussed below.

h

98

A Manual of Linguistic Fieldwork and Structures of Indian Languages All the stages of data elicitation are in feeding order as explained below by the diagram. To acquire sociolinguistic information one can elicit data after stage 3 but this is independent of stages 4 and 5. The result of analyzing each phase is given in brackets. Fig. 4.1 Stage 1

Stages of Data Recording

basic word list

(basic sounds)

ß Stage 2

400 word list

(phonological structure)

ß Stage 3

short phrases

(morphological paradigms)

ß Stage 4

simple sentences

(syntactic structure I, i.e. word order, agreement)

ß Stage 5

complex sentences

(syntactic structure II)

We shall discuss stages one and two in this chapter and three in the next chapter. Stages four and five will be discussed in chapter six. 4.5.1. Stage I :Basic Word List One generally begins with what is popularly known in the literature as the “basic word list’, originally proposed by Sarah Gudchinsky and Morris Swadesh (1956; see Appendix 2[A]). We generally avoid collecting the a number of words from the original list, as the meaning of these are hard to explain to the informants, either because these are function words, or because the words are polysemous in English and are not easy to comprehend. Moreover, these words are not very explicit, unless one collects some sentences of both complex and simple types, which we cannot do in the initial stage. The words are at, because, blow, dull, grease, float (as it confuses with the next word in the list which is flow), hunt, (it pertains to a very specific activity not known to all communities) if (without reaching the phrase-stage this is hard to explain to the informant), in (the locatives are difficult to explain in the beginning stage), shoot (again a very specific activity not known to many people), smooth, snow and turn. We normally add 89 words which are pertinent to the Indian context (including some in Indian languages as there are no parallels in English) to the basic word list (see Appendix 2 [B]). The consolidated list of 300-odd words (the consolidated basic word list or CBWL), gives us enough opportunity to identify the phonemes in the language and their phonetic variants. It also gives us a basic idea of different grammatical categories existing in the language. The basic words can be rearranged according to the field linguist’s choice. Some of the suggestions are to divide them on the basis of semantic fields, such as body parts, kinship terms, numerals, seasons, natural forces, diseases and medicines, etc. or to divide them on

4. Elicitation 99 the basis of grammatical categorisations, e.g. nouns, verbs, modifiers, etc. We prefer to use different coloured pencils to underline words of various grammatical categories at this stage. This helps an investigator to build her/his future inquiry words and phrases. It is best to collect the basic word list on cards, with one word on each card, so that later there is enough space to write the derivational or inflectional affixes of each of the words the same card. You may use any of the interrogation techniques discussed above. Our observation is that some informants prefer to see the basic word list, take it home and then give the equivalents the next day. Personally, we don’t advocate this, but sometimes one has to concede to the wishes of the informant as this is the beginning stage, and we need her/his help more than s/he needs ours. 4.5.2. Stage II 400 Word List To get a satisfactory phonological sketch one must collect at least 400 words (see Appendix 2, [C]). However, the ideal figure is approximately 500 words, provided one has covered the various word formation structures and morphological processes the language makes use of (see below). Thus, there is no magic number that one can specify, but given the usual constraints of time and lack of knowledge of the language in the beginning stage, a word list of 500 words on various semantic fields prove to be rather useful for acquiring the knowledge of the sound system of the language as well as of the basic processes of word formation. This is not difficult, as it is observed that while collecting the CBWL one naturally arrives at collecting 400 words. This happens as many times the informants give more information than asked for, or at times one collect gender and number markings on nouns and verbs, or one collects grammatically and semantically related words to the words already listed in the CBWL. It is observed that while collecting words for cardinal numbers one tends to collect ordinals too, or while collecting words for kinship one ends up collecting a large number of new terms which are non-existent in English. Some culture-specific words are also inadvertently collected in the initial stages of data collection. What one has to keep in mind at this stage of data collection is that one should elicit words that exhibit the most important semantic entities. A short list of various semantic fields can be suggested to identify the areas in which these words can be collected. 4.5.2.1

Various Semantic fields

1. Kinship terms 2. Address terms 3. Pronouns 4. Professions and related words 5. Household items 6. Seasons, weather, time, celestial bodies 7. Pain, illness and medicines 8. Body parts 9. Numerals 10. Flora and Fauna

100

A Manual of Linguistic Fieldwork and Structures of Indian Languages 11. Adornment, costumes 12. Edible items and related words for cooking, spices 13. Expressives for various emotions 14. Games, toys, sports and related items While collecting data on body parts, remember that often languages have genitive nouns or pronouns attached to the term to show whose body part it is1. Thus one must collect data by naming a person known to the informant or by just saying ‘my such and such’ or ‘your such and such’. Concerning items of adornment, the names of jewellery pieces, and concerning food items, the names of sweets and spices are very useful for collecting compound nouns, as most of them are derived by compounding in Indian languages. Data elicitation should be done according to the location of fieldwork. If you are sitting in the kitchen, try to obtain words related to cooking, utensils, food items, taste etc. However, if you are sitting in the open, flora and fauna and other related items should constitute your field of inquiry. In this way you can help the informant relate to the existing world and in turn may get words for which perhaps you have no term in your contact language. In addition to the semantic fields listed above, we also take into account words related to some of the semantic fields suggested by Anna Wierzbicka (1991). These are: 1.

Substantives: I, YOU, SOMEONE, SOMETHING, PEOPLE

2.

Mental predicates: THINK, SAY, KNOW, FEEL, WANT

Most of the semantic primitives she suggests under the heading ‘Determiners and Quantifiers’, ‘Action and Events’, ‘Meta-Predicates’, “Taxonomy and Partonomy’ are extremely difficult to elicit until we reach stage III. Other primitives such as ‘Time and Place’ or ‘Evaluators and Descriptors’ are already covered in the basic word list. 4.5.2.2 Preparing the Lexicon Cards This is the stage when one can start recording data on index cards of the sise 15*10.5cm. Each card can be reserved for one head-word2 and subsequent inflectional and derivational units or any other relevant morphological information on the word can be added to the same card. The format we generally follow is the same as used by the electronic dictionary-making program called LEXWARE. We shall give a sample of this format as it can be adapted and modified to one’s focus of enquiry. In the initial stages it is helpful to take coloured cards, each representing one grammatical category or the other.

1

Andamanese, for example, has a very elaborate system of possessive affixes attached to names of specific body parts. Please consult Manoharan 1997:465 for details. 2 Head word is a term generally used in lexicography for each main word under which diverse information regarding inflection, derivation etc. can be added.

4. Elicitation 101 Fig. 4.2 Sample of recorded lexicon on 15*10.5 cm. Card: Language: Sambalpuri

Location: Bargarh Date:

h.w.

a:sbar

mn.

To come, arrival

p.s.

verb intransitive, noun abstract

exmp. 1

mui j´ldi a:smi (a:sijimi) I

soon

come fut.

‘I will come soon’ exmp. 2

a:ej ta:r a:sbar kichi ÿHI:k

nQâ n´

today his arrival some certain neg. tag ‘his arrival is not certain today …S.E.

a:sba b´r´s

mn.

coming year

ps

modifier

exmp. 1

a:sba b´r´s ke mui ´merika jimi coming year loc. I

America go fut

‘I will go to America in the coming year’. Simple derivations can be listed on the same card if and when they are found, always recording the morph-to-morph translation. It is always good to record the whole sentence even when your aim is to elicit only one particular word. 4.5.2.3

Phonological and Morphological Contrasts

Stage II is particularly useful for checking phonological and morphological contrasts, and variant forms existing in the language. Before we proceed we would like to suggest the following to a field investigator. Tips and suggestion to the fieldworker are given under the hand symbol throughout the manual.

! •

Sit close to the informant so that you can see her or his mouth so that various mouth orifications in the production of new types of vowels and consonants are not missed.

•

Ask each word to be repeated at least three times. While working on tone languages this might need to be even more often.

•

Write each phonetically transcribed word in square brackets [x] so as to distinguish them later from phonologically transcribed words written between slashes /X/.

102

A Manual of Linguistic Fieldwork and Structures of Indian Languages •

As soon as you find a minimal pair, list it separately on a different page of your notebook. Keep adding to this list as you proceed.

•

As soon as you identify a variation of a sound due to adjacent environment, circle the information. This will guide you to look for a similar phenomenon. For instance, if you have noticed that two varieties of the voiceless dental [t] occur, one type before the semivowels and the other type in the rest of the environments, circle this information next to the elicited data in your notebook and check against this information when you collect the voiced counterpart [d]. Most often than not, [d] might also have two varieties occurring in similar environments.

•

If you find a gap in the series, mark it. Languages generally follow pattern congruity. If gaps occur, they should be further studied for diachronic, areal and contact phenomena.

•

If the informant does not take offence, try to tape all that is elicited at this stage on a reasonably good tape recorder.

•

Carry a blank matrix-table of vowels and consonants with you at all times so that as soon as you identify a sound you can jot it down in the appropriate box of the matrix. You will require two each for vowels and consonants - one for the phonetic inventory and the other for the phonemic inventory.

•

Carry an IPA table of symbols for vowels and consonants (or any other according to your preference). This is given in the appendix for your reference.

How and with what should one begin? Firstly, the most frequent words should be included in the list, although it is difficult occasionally to know which words are frequent and which are not. Similarly, the words which are difficult to translate should be avoided at this stage. The best possible way is to rely upon your elicited 400-word list to begin with and soon you will be adding to this list as you proceed. We are not going to discuss various methods of isolating phonemes or methods of writing phonological rules as these may be adopted by the investigator according to her preference for a particular model or theory. At this juncture, we can suggest that it is always helpful to look for minimal pairs first, as not only do they help you reach a phonological decision, but they are also fantastic samples to prompt an informant to give you the appropriate data. For instance, observe the following conversation between an investigator and an informant. The language under observation is Hindi: FI:

Did you say that the word for half is adHa ?

I:

yes

FI:

Do you also have a word such as atHa?

I:

Not atHa but we have ata.

FI:

What does it mean?

I:

he comes.

FI:

Do you have a word like aÿa?

4. Elicitation 103 I:

(pleased) yes we do.

FI:

What does it mean?

I:

flour Conversation piece no. 4.2

By now you know that the voiced aspirated dental does not contrast intervocalically with the voiceless aspirated dental. But you have also got a clue that perhaps the voiceless series of dentals and retroflexes contrasts at least wordmedially. However, you should not give up hope so early, unless you have exhausted all the possibilities. Now try to find the voiced aspirated dental and its voiceless counterpart in initial and final positions. FI:

Can you think of a word which begins with dH? You know, words like dHata, dH«l, dH«n?

I:

yes we have dH«n meaning wealth.

FI:

Do you have tH«n ?

I:

yes, we do. It means udder.

FI:

Can you think of a word which goes like t«n ?

I:

Yes, there is t«n and it means ‘body’

FI:

Do you have similar words, you know can you rhyme it with any other word?

I:

j«n, m«n, pH«n, s«n…….(and the list is long) Conversation piece no. 4.3

This kind of conversation presumes some basic education on the part of the informant, as illiterate people do not understand your clues very easily. One can change various vowel-consonant combinations, and you will realise that it is not very difficult to find contrast and free variation. However, by asking for rhyming words you may not necessarily acquire minimal pairs. The last answer to your question sends the informant on a cooperative trail but what you get is the CVC structure of the same vowel intervening between two consonants out of which the last one is fixed (i.e.-n). They do not form a minimal pair, as the two sounds of a minimal pair should be phonetically similar. And there is no phonetic similarity between the initial sounds of j«n and pH«n or s«n or t«n, or at least these initial sounds do not form a natural class, except the fact that the first two examples are [+obstruents] while the third one is [+continuant], and the last two are [+anterior]. Each of them takes part in a bigger natural class, but fails to give you what is commonly known as minimal pair.3 One should try to locate sounds contrasting in identical environments as well as in analogous environments. It is always helpful to make a distinctive feature chart of all the sounds occurring in the language, as this can help you identify and write phonological rules. As languages of this subcontinent are heavy on aspirations and retroflexes, the typical feature terms for these suggested in SPE by Chomsky and

3

Consult Pike (1947), Nida (1946), and Hockett (1955).

104

A Manual of Linguistic Fieldwork and Structures of Indian Languages Halle (1968) are not very appropriate. I use the conventional articulatory terms [+/aspiration] and [+/- retroflex] respectively. Once contrasting sounds are identified and some phonemes (both vowels and consonants) are established, you can look for conditioned occurrences, i.e. sounds occurring in phonologically conditioned environments. For this the best method is to see what sound occurs in which position. That is, try to collect words with initial, medial, and final occurrences of each sound under observation. It is easy to establish the positional occurrence of sounds after acquiring the 400 word-list. This is the time when one can easily check the variant forms, i.e. sounds occurring in complementary distribution. This would require keen observation of the adjacent phonetic environment. As sounds change due to preceding and following sounds and syllabic structure, one must exhaust all the possibilities of establishing variant forms of an underlying structure. The allophonic distribution of sounds along with the establishment of phonemes will give us a fairly good idea of the sound system of the language. The entire picture will emerge only after you have studied all the variations and can formulate phonological rules4 of the language concerned. A word about recording suprasegmental features is warranted here. It is not easy to capture all the tone contours, stresses and junctures at the beginning, but with some ear training one may be able to distinguish the various patterns. A very good software program of studying tones is available from the Summer Institute of Linguistics [SIL] which helps you study the tones very accurately. It is a speech analyser. The informant speaks into a microphone connected with the computer, and one can have a visual presentation of speech contours. For the analysis of tone languages, the most important features are the pitch graphs, which display the spoken tonal contours exactly. One can also make measurements in magnitude, of spectra and spectrograms, and of formants. There is a device where one can, with the help of formant measurement, find out the exact places of vowels within the vowel "square"5. A similar program that can easily be downloaded from the web and is useful for analysing speech is ‘Gold Wave’. It helps you to hear the recorded speech again and again and dissect it anywhere you would like to analyse it. One can also superimpose phonetic transcription on the recorded speech for future reference6. 4.5.2.4.

Co-articulation and Consonant clusters

While we discuss consonants we must remind our readers that only few languages of India such as Khasi offers extreme cases of co-articulated sounds in all positions, but most frequently in initial positions. Similarly, this is also the language that offers the largest number of dipthongs. Some of the common co-articulated consonants and dipthongs are given below for your reference (the list is not exhaustive). Many languages do not release plosives word-finally (e.g. Konkani and Khasi). This, although it is a predictable feature, should be marked by a diacritic such as [ ] above the consonant letter.

4

For writing phonological rules consult any basic book on phonology. We recommend Schane, Sanford (1978); Kisseberth, Charles and M. Kenstovicz (1979). 5 This information has been made available to us by Claus Peter Zoller. 6 Demonstrated personally by Ian Maddison.

4. Elicitation 105 Table 4.7

Co-articulated consonant Sounds in Khasi

Distribution by the second consonant 1.

pr, br, kr, kHr, mr, rr

2.

bn, pn, pHn, dn, Sn, sn, tHn, hn, nn, kn, kHn

3.

st

3.

km, kHm, mm

4.

dk, rk, rkH, sk

5.

pl, pHl, bl, tl, tHl, kl

6.

rN, sN, tN, pHN, rN, lN

7.

st, kHt, tt, lt,

8.

ks, kHs

9.

kj, jj

10.

sp, kHp, pp

11.

bø, pHø, kø, kHø, rø, sø

12.

t/, l/, s/ Table 4.8

Dipthongs and Tripthongs in Khasi

1.

/I/ as a first member, I!, IE, Io,

2.

/I/ as a second member; eI, EI, aI, oI,

3.

/o/ as the second member «o, ao, Io,

4.

/u/ as the second member au

I,

Tripthongs 1.

oaI as in roaI ‘ to sing’

Tripthongs are common in Meithei and its related dialects. Sounds change in very many ways because of their adjacent environments. They also change in such ways that the syllable structure of the language under consideration also changes. All these kinds of changes in linguistics are studied under the term ‘phonological processes’. One must be vigilant to identify which phonological process is operating on what sounds or set of sounds or natural classes. Readers may consult S. Schane (1978) for phonological processes. Always try to write these phonological processes in the form of rules. If in future you ever decide to do a comparative study, written rules in a given format will be very useful. 4.5.2.5

Phonotactic constraints

106

A Manual of Linguistic Fieldwork and Structures of Indian Languages Phonotactic constraints on vowel and consonant clusters as well as their positional constraints can be checked here. If the language appears to be using several types of CC clusters as found in Khasi and in many Tibeto-Burman languages of the Northeast region, one may make a matrix of consonant sequences and elicit data on the basis of coined words. These coined words may sound nonsensical to the native speaker, but s/he will get the clue that you are looking for words which have many consonants in the beginning or at the end or in the middle of the words, as the case may be. What you have to concentrate on here is not alone on the number of consonants or vowels occurring medially, initially and finally but also which type of consonant (or vowel) combines with which type of consonant (or vowel). For instance, if the first consonant in initial position is a bilabial plosive you may ask yourself ‘can the second in the sequence be a dental fricative, and if yes, can these two combine with another dental nasal?’ You may then formulate your question in nonsensical words with the intended sounds used at the appropriate place. For instance, you may give a word beginning with “bsn…” for the informant to fill in the rest. But this is a tedious way of elicitation. Consonant clusters are best collected from texts or from sentences collected at the third and fourth stage. Research on phonotactic constraints on consonants and vowels is generally a bye-product of some other research. In our course on field methods, we make students working on phonotactic constraints sit and observe speech under investigation along with other students working on various other topics. Later, after substantial data has been collected, one can make a full matrix of consonant clusters at the analysis stage exhibiting cluster sequences in initial, medial and final positions of a word. Three consonant clusters in the middle of a word are not uncommon, either. Languages of the Northeast offer three consonant clusters, though languages of other families are not far behind. Khasi pHrtHat ‘thunder’, prkHat ‘ to think’, prsa:dmInsEm ‘to breathe’; Konkani (IA) tamb}E ‘red’, nustyak ‘a kind of fish’ mElkriat ‘fool’, l!Ng}! ‘lame’; Bagri jI÷ÿñI ‘disheveled hair’, cuâNg÷o ‘to suck’, ku}kli ‘earrings’; Kurux (Dravidian) kH«njpa ‘fruit’, g«rab?na ‘squeeze’, or four consonant cluster Ilcka ‘fear’ are some of the instances. Double consonant clusters or geminates in medial position are very common in all Indian languages. It is observed that geminates can be formed from the entire inventory of the consonants of the language in question. Consider some of the common geminates from Bagri (IA) Meithei (TB), and Khasi (AA). Table 4.9

Consonant Geminates

Bagri -bb-

pH«bbo

‘toes’

-tt-

cHatti

‘breast’

-êê-

h«êêI

‘bone’

-vv-

tovvo

‘girdle’

cakk«

‘with rice’

Meithei -kk-

4. Elicitation 107 -mm- pH«mmu

‘sit down’

-nn-

cann«b«

‘intimate’

-NN-

h«NNu

‘ask’

-ll-

pullu

‘blind’

-rr-

barr!H

‘all’

- ll-

kH«ll«n

‘child’

-pp-

h«pp«

‘inside’

Khasi

-mm- tImmEn

‘old’

Consult chapter 2 to locate the expected sounds and sound patterns as well as word structures in a specific language family. But obtaining an expected sound system or word formation is not the only goal of any field worker. One must be alert and enthusiastic in identifying ‘new’ elements in the language. Language study always brings in surprises. And it is these surprises which make the study of new languages more interesting. No language, as said earlier, fits into any straight-jacket compartment and the phenomenon of languages in contact changes the structures of languages in unpredictable ways. 4.5.3.

Stage III

Short Phrases

Stage three begins by collecting short phrases, preferably two, three or four words long formed from the limited vocabulary collected so far. Stage three can give information regarding morphological paradigms, various word formation processes used by the language concerned, small clauses, as well as word ordering of modifier and modified. We must remind our readers that while collecting data one should concentrate on all four types of word formation processes such as the following (it is taken for granted that the term ‘word’ is equal to the phonological word.): • Inflection deals with the addition of various morphological materials (affixes, such as pre-, in-, and suffixes) to the existing words but retains the grammatical category and referential or cognitive meaning of the original word. • Derivation deals with the various ways of creating distinct grammatical categories or changes in meaning by affixation and alternation. • Compounding deals with combinations of two different bases • Reduplication deals with the doubling of the whole or part of the base. These four form the ‘productive’ part of the morphology which give us rules for formulating new words (as suggested earlier, if we consider the definition of a word as including the phonological word, then inflectional processes also form new

108

A Manual of Linguistic Fieldwork and Structures of Indian Languages words) in the language. This implies that the elicitation of data for word formation will not only give us the entitative knowledge but will also inform us about the processual aspect. We agree with Uhlenbeck (1995) in pointing out that the status of ‘word’. “Word - the central unit of language and language use takes part in three different structurations. It is involved in morphonological structure as a unit with perceptible, identifiable form; it is involved in morphology as a unit of form and meaning; and it also participates in syntactic structure as meaningbearing unit combinable with other similar units into groups of varying complexity.” (Presidential address 1991 Meeting of the Societas Linguistica Europea). The technique of eliciting data for short phrases is that the field investigator should make three-word phrases out of the existing words already collected rather than giving the three-word phrase in the contact language for translation. There are two advantages to this method. First, as soon you make an error with the word order you will be corrected. Secondly you will learn the agreement pattern in the language concerned. Lastly, the informant will be elated to find out not only that you are making an attempt to speak her/his language but are at times rendering perfect short phrases. Let us narrate an incident here that pertains to what we said above. When we were working on the Roma language of Gypsies in Heidelberg in Germany in the summer of 1998 I was not at ease throughout the elicitation as our contact language was German, of which I had no knowledge. My students used to work as interpreters, first listening to my giving the cue word or phrase needing elicitation in English, then translating it into German for the informant (who was bilingual in Roma and German, but as my students informed me his German was not very fluent). In return, I transcribed whatever the informant said in Roma, but if he ever got into an explanatory mood in German I was lost and on such occasions felt very helpless. When we reached the stage of collecting small phrases I took charge and decided to make three-word sentences in Roma. Since many basic words in various Roma dialects are still Indic, especially Punjabi and Hindi, I kept on making phrases with words collected so far as well as with those which I could guess. Initially the informant corrected me for using the SOV pattern and I soon realised that the language has changed to the SVO pattern during their 700 years’ history of migration to an SVO belt. It was not a long session, and if I remember right after only 30 minutes of this technique of data collection he was in awe, literally flabbergasted by my rendering of Roma phrases. He could not believe his ears that a woman coming from India who knows no German and no Roma could speak “pure” (that is what he said) and good Roma. He was so overwhelmed that he got up from his seat, shook my hand and said with moist eyes in Roma “you are my sister’’, and I promptly replied “and you are my brother” in Roma, a sentence even I did not know till then. I narrate this incident also to suggest that if you are using an interpreter because of your ignorance of the contact language, perhaps this is the stage when you can say good-bye to your interpreter, at least until you collect complex phrases and narrative texts. Let us consider a list of topics that can be covered for eliciting information regarding the morphological structure of the language.

4. Elicitation 109 4.6

Morphological Topics to Be Covered For Data Elicitation I INFLECTION Nouns • Gender • Number • Person • Case (synthetic and analytic) • Definite and non-definite • Classifiers Pronouns • Reflexive pronouns (agentive and non-agentive) • Possessive pronouns • Demonstrative pronouns • Interrogative pronouns • Determinative pronouns • Indefinite pronouns • Reciprocal pronouns • Relative-Correlative pronouns • Anaphoric Pronouns can be marked for [- feminine], [- remote], [-proximate], [-dual], [oblique], [-ergative], [-case], [- honorific]. Verb • Voice (active, passive, and capabilitative) • Tense (present, past, future) • Aspect (perfect/perfective, imperfect, habitual, progressive) • Mood (indicative, obligative, imperative, optative) • Nonfinite (infinitives used in nominalisation, imperatives, participles, both present and past) • Conjunctive participles or Converbs Adjectives: (Types and agreement with the modified NP) • Shape

110

A Manual of Linguistic Fieldwork and Structures of Indian Languages • Size • Colour • Taste • Quality • Equative • Comparative • Superlative Case System • Case inflection (nominative, oblique, vocative) Quantifiers • Enumerative Numerals • Cardinals • Ordinals • Aggregative • Multiplicative • Non-enumerative • Partitive and non-partitive • Quantifier + numeral (Hi: h«r-ek ‘each’) Adverbs (types and agreement with the modified) • Primary • Derivative • Temporal • Spatial • Locational • Directional • Manner • Cause/Purpose • Negation

4. Elicitation 111 Particles/Emphatics • Contradictory • Non-contradictory II

DERIVATION

Noun • Noun > Noun (feminine, diminutive, agentive, professional, abstract, attributive) • Adverb > Noun • Adjective > Noun • Verb > Noun Verbs • Noun > Verb (conjunct verbs ) • Noun + infinitive > Verb (Hindi b«tiyana ‘to gossip’) • Verb > Verb (causative) • Adjective > Verb (Hindi l«ng}ana ‘to limp’) Adjectives • Noun > Adjectives (affixation) • Verb > Adjectives (past and present participial) • wala-constructions (from V, N, and Adj.) Adverbs • Noun > Adverbs • Verb > Adverbs (participles) • Adjective > Adverbs • Interjections Postpositions • Compound (Hi. ke liye ‘for’, ke satH ‘with’) • Complex • Case markers (agentive, ergative, accusative, ablative, dative, genitive, locative, allative, instrumental, comitative)

112

A Manual of Linguistic Fieldwork and Structures of Indian Languages III

COMPOUNDING

• Endocentric, Exocentric and Oppositional • Nominal • Verbal • Adjectival • Adverbial • Explicator compound verbs - semantic types • Aspectual • Adverbial • Attitudinal • Others IV

REDUPLICATION

• Morphological : Expressives 1. Kinship terms 2. Words of five senses of perception 3. Others • Lexical 1.

Complete: all grammatical categories

2.

Partial : Echo formations and syllable reduplication

3.

Discontinuous

• Semantic (these may be studied under compounds) 1.

Adjunctive (Hi. dH«n dçl«t ‘wealth’)

2.

Derivative (Hi sili-silai: ‘readymade’)

Information regarding these various topics will be elicited from various grammatical sources. Repetition is inevitable as language works in a cohesive fashion, and as you extract information about one structure you get information about many more. For instance, while collecting information regarding gender and number it is possible that case markings are also evident. For instance, Hindi l«}ke ‘boy/s’ could be either nominative plural or accusative singular. Languages have portmanteau morphs, especially those that are highly inflecting. One should be prepared to study these carefully.

4. Elicitation 113 The line between inflection and derivation is rather thin, yet for analytical purpose we keep these domains separate and will discuss what needs to be elicited to find out everything about inflectional units in the language under observation. Make use of appendix 5 for a basic sentence list. Elicitation of data using this list will give you much-needed information regarding morphology, syntax and semantics at the very beginning of your fieldwork. 4.7.

Dichotomy between Noun and Verb

We know that words can be classified in different grammatical categories, such as verbs, nouns, adjectives, adverbs, interjections etc., depending upon the specific functions they perform in a sentence. Our school grammars have always taught that. The reality, however, is different. Not all languages make a clear-cut distinction between nouns and verbs, or verbs and adjectives. There are languages in South Asia that use only one word that performs the function of reference (i.e. noun), modification of the participant (i.e. adjectives), and also simultaneously performs the function of indicating an action/state (i.e. verb). Languages may differ from one another in types of word classes they possess. This will depend upon the kinds of syntactic function each word class performs in the sentence. The variety of word classes that exist in Indian languages is best reflected in the following statement by D.N.S. Bhat (1994), which was made in the context of locating the category adjective in the languages of the subcontinent: The study of adjectives has been rather fascinating in this respect because the four major language families that are prevalent in India have turned out to be representing the four distinct language types that need to be postulated in order to describe the categorial characteristics of adjectives. That is, it has turned out that the category of adjectives is distinct from that of nouns and verbs in Dravidian, whereas in Indo-Aryan it is generally indistinguishable from that of nouns, in Tibeto-Burman it is indistinguishable from that of both (nouns as well as verbs). (Bhat 1994:xi) Universalists have maintained that notional distinctions between noun type and verb type or adjective type can be maintained even in those languages which do not provide distinct word classes. This theory has been in dispute as linguists working on American Indian languages (Thompson 1979, Kinkade 1983) and those who have worked on Munda (Hoffman 1903, Pinnow 1966, Bhat 1994, 1997) strongly believe that such a bifurcation is not justified. As we saw in chapter two, the lexical distinction between noun and verb in Munda languages is more or less absent. That is, “any word can be used for any concept” (Pinnow 1966:101), meaning that there exists only predication. Any information regarding reference and referent or modification, numeral, quantifiers, and clitics are incorporated in verbs alone. These are also the languages that do not follow the rigid distinction of word and sentence. One word is a sentence as it conveys the entire information about the arguments and predication along with various modifications. In fact, one can think of positing a cline of gradation of languages showing such distinction as suggested by Bhat (1997:231). The fluidity with which noun-ness and verbal-ness can be assumed by distinct word classes or by a single word is an interesting area of research in this part of the world. Having raised this issue of the dichotomy, one has to be forewarned and

114

A Manual of Linguistic Fieldwork and Structures of Indian Languages prepared about the area that one is going to research. Prototypically, word formation processes, as discussed earlier in section 4.5.3, will show up on different word classes (if the language allows them) by inflection, derivation, reduplication and compounding. Let us discuss each of these processes in detail in the following chapter. Identification of these processes will also expose you to the basic syntactic structure of the language under investigation.

Chapter V Morphology and Word Formation We shall now discuss the morphological structure of words and the four processes of word formation mentioned earlier in 4.5.3, and the elicitation techniques for each of them. Before we begin, let us recall that a phonological unit is taken as definition of a word. Traditionally, inflection is not considered one of the word formation processes. We, however, would like to incorporate it as one of the word formation strategies, because the line between derivation and inflection becomes very thin in some languages, e.g. in Munda. Furthermore, inflected words serve as bases for further word formation (and are therefore similar to derivation in function), and lastly, the definition of word that we have adopted has given us the freedom to do so. Independently of our observation, Booij (1996) advocates the allowance of inflection to interact with word formation, contrary to the predictions of the hypothesis of split morphology. In order to identify word formation processes you would be required to collect simple sentences. Please consult the list ‘basic sentences’ given in the appendices. This would help achieve some significant syntactic analysis, too. As we know, the distinction between morphology and syntax cannot be made in watertight compartments, therefore any analysis of word formation will throw light on the syntactic processes operating in the language as well. Though we mention four distinct word formation processes, it is difficult to find a language that keeps the distinction very clearly. The basic organisation of the chapter is as follows. We shall first define and describe each word formation process followed by varying examples from Indian languages. We will enumerate and at times elaborate on the unique and uncommon linguistic characteristics of Indian languages to familiarise the reader with the linguistic scene. This, we hope, will help the field investigator to take into account all kinds of linguistic variables to achieve her/his goal. Suggestions on elicitation methods are given after each section or sub-section under the symbol !. 5.1

Inflection

5.1.1 Noun Morphology In this part of the chapter we will be concerned with the overt markings that a noun category in any language can take. We will not exclusively restrict our discussion to inflection, but shall explore all the related ways that noun morphology is expressed in various languages. We shall see that a language may make use of inflection, reduplication and compounding, or of all three strategies at the same time. Sometimes two different languages from the same family form their noun class by two different types of word formation processes. We will draw examples from varied and lesser-known languages so as to familiarise our readers with the diversity of the Indian linguistic scene. A field investigator, we believe, should be aware of the nature and the complexity of the phenomenon that s/he is investigating.

116 A Manual of Linguistic Fieldwork and Structures of Indian Languages Noun morphology contains various kinds of information, which may be represented differently in different languages, as not all types of information show overt marking. This is also the easiest category to elicit. One should start with the already collected material, especially with the 400-word list. The first task is to find out whether there is any overt marking on nouns at all. To find this out, special attention should be given to the expression of gender (G) and number (N). Try to employ noun words already collected in plural number and feminine gender. While collecting words of the noun class, one must keep in mind the following possible classification of nouns. •

+ abstract

•

+ count

•

+ animate

•

+ human

•

+ feminine

•

+ plural

•

+ potent

Languages, at times, are seen to morphologically mark the various distinctions. Alternatively, languages do not overtly mark the distinction, but classify nouns according to the semantic distinctions, which may show elsewhere in the language, for instance on modifiers or on the verbal morphology or on the noun itself by various prefixes assigned for each type of noun class, as in Andamanese1. The semantic classification of nouns given above will help select the appropriate noun list from your database. 5.1.1.1

Gender and Number

This brings us to the basic question of determining the number of noun classes in a language. According to Corbett (1991, 1993), the number of genders in a language is determined by the types of agreement2. In this manual we would follow the concept of agreement as described by Steele (1978:585-623) “The term agreement commonly refers to some syntactic covariance between a semantic or formal property of one element and a formal property of another”. For instance, Hindi has two types of agreement, Bangla has one type and Konkani has three types. It is important to bring in the two terms used by Corbett here, i.e. the target gender and the controller gender. The target gender shows as the surface realisation in a number of ways on modifiers, on nouns itself (by some phonologically fixed shape) and on verbs (predication). Controller genders are those elements (generally in the noun phrase such as nouns, pronouns etc.) that control the agreement. In Telugu, nouns divide into three gender classes. There are two surface markers of genders in singular and two in plural, but three combinations of these (Krishnamurti and Gwynn 1985:56-80). Consider the following table and the examples given after the table. 1

Manoharan, S. 1997: 453-73 in ‘Languages of Tribal and Indigenous Peoples of India’. The concept of “agreement” has been defined by several linguists such as Matthews (1977:12), Steele (1978:610), Keenan (1978:167) Lehmann (1982:203), and Durie (1986), among others.

2

5. Morphology and Word Formations 117 Table 5.1

Gender marking in Telugu

Singular

Plural

-êu

males, moon, sun

-ei

females neuter

-ru

-yi

The following examples with the verb ‘to be’ will clarify the phenomenon. 5.1

wa:ê(u) unna:êu He

be+ msg

‘He is’ 5.2

ad(i)

unnadi

she/it be+f/nt sg ‘She/it is’ 5.3

wa:ru unna:ru they

be+m/f pl

‘They (those persons: m.f ) are’ 5.4

aw(i)

unna:yi

they

be+nt.pl

‘They (those things) are’

There are three issues involved here. (1) We must establish whether the language under investigation forms regular paradigms of gender (masculine and feminine or neuter), and number (singular, dual, plural) by the help of some inflectional categories. (2) We must ascertain whether the language concerned shows some kind of agreement between the noun and its modifier or the verb. There could be languages that do not show nouns overtly inflected for gender or number but may show agreement, e.g. Dravidian. However, Hindi shows both the features mentioned above, i.e. distinct noun shapes according to different gender as well as agreement on modifier and verb. (3) Language contact phenomena, as mentioned in chapter 2, may always bring about conflicting results (e.g. Kurux unlike its other Dravidian sister languages overtly mark gender distinction on nouns such as in Hindi) and the field investigator should always be prepared for idiosyncrasies. Please refer to the ‘contact and convergence’ (2.4) section of the manual.

118 A Manual of Linguistic Fieldwork and Structures of Indian Languages

!

Now the question is, how does one know whether the language makes any distinction in gender, masculine and feminine or in animate non-animate nouns. The first strategy is to collocate several nouns with selected modifiers from your list. Thus, phrases such as ‘red flower’, ‘blue river’, ‘black sheep’, ‘tall trees’, etc. will immediately reveal to you whether gender and number are marked overtly or not. The second best way to find out is to take two names, one for a girl and one for a boy, and frame a phrase like ‘X came’, ‘Y ate’. All inflections concerning these categories should be elicited in simple phrases and clauses. These phrases should be drawn from day-to-day situations to bring in the naturalness of the conversation. The other device is that you make simple clauses with different types of verbs collocating with different types of nouns. Highly inflecting languages, such as IndoAryan, would show in no time that nouns are distinguished on the basis of gender and number, sometimes by noun morphology and sometimes by the morphology of the accompanying modifiers and verbs. Amazingly, languages are not classified very simply on this basis. The determination of feminine and masculine gender itself is not fixed on some biological criterion, or on gender-specific entities. In Hindi, words for ‘beard’ and ‘moustache’ are feminine. Every language community perceives its world differently and thus decides what should be referred to as masculine and what as feminine or neuter. Languages may also have alternating patterns. Hindi, for instance, offers many nouns whose shape will give you the information about the gender and number. But then there are many nouns in the same language that do not reveal the gender and number by their phonological shapes. If the language does not overtly mark gender and number on nouns, one should observe the verb morphology that may indicate the gender and number of the argument nouns. Let us see what the possibilities of marking these categories in the various languages are. Languages such as Bangla, though Indo-Aryan, make no overt distinction on nouns or on verb morphology. Consider the examples from Hindi and Bangla given below. We will restrict these to singular nouns. Hindi: 5.5(a)

radHa «ccH-i bat k«rt-i hE Radha good-f talk-f do-3f sg aux

5.6(b)

ram «ccH-a

kam

k«rt-a

hE

Ram good-m. work-m do-3msg aux Bangla 5.7(a)

radHa bHalo ka!tHa kore Radha good talk

5.8(b)

ram bHalo kaj

does (unmarked for GNP) kore

Ram good work does (unmarked for GNP) (a) ‘Radha talks (of) nice things’ (b) ‘Ram works well’

5. Morphology and Word Formations 119 Thus, bat in Hindi is feminine and kam is masculine and govern the verb accordingly, while k!tHa and kaj respectively in Bangla are not marked, and as you can see the verb remains invariant in both forms. There never was any satisfactory way of identifying the class of nouns belonging to various genders. Sometimes the phonological shape of the noun under consideration gives a clue but not always. Thus words ending in -i in Hindi are generally considered feminine, and those ending in -a masculine, but the list of exceptions is long. Similarly, the -en suffix in Kashmiri indicates feminine gender, as in sa:luren ‘the female guest’ as opposed to sa:lur ‘male guest’ or ka:r ‘crow’ and ka:ren ‘female crow’ etc., but there are exceptions, too. Features such as animacy play an important role in deciding gender distinction, e.g. in Dravidian. However, languages that make a distinction between feminine, masculine and neuter (e.g. Sanskrit, Konkani, Kharia) do not decide neutrality only by the presence or absence of the factor inanimacy. In many languages, fruits and plants that are used for medicines are considered feminine. In Konkani, for instance, many fruit-bearing plants are [+feminine] while the rest are [+neutral]. Languages such as Telugu assign the inanimate gender to young girl children. Marathi offers three distinct prefixes, which denote the gender of the lexical items, thereby making the gender system morphologically traceable. Thus, te- is prefixed to masculine singular and neuter singular nouns and ti- is prefixed to feminine singular. Though the word for tortoise takes neuter ending and the one for ‘clay’ takes the feminine. In many Dravidian languages, the gender distinction gets represented only in personal pronouns and not in nouns. However, the agreement pattern visible on verbs does further reinforce this distinction. The question of what can be viewed as masculine, feminine or neuter depends on the world-view of the specific culture. One can only recommend to observe the distinction carefully by checking the verb endings and endings of the modifiers when one works on a new language. Austro-Asiatic and Tibeto-Burman languages form gender distinction by the compounding of nouns. Thus, in Mizo wuiH-pa ‘dog’ while wuiH-nu ‘bitch’. In these languages, the male/female distinction cuts across the animate/non-animate distinction. Interestingly, in Munda animate nouns are intellectual beings, animals, super-natural powers, and heavenly bodies. The criteria for deciding the gender of each noun vary from community to community, as the decision lies in the cognitive aspect underlying the specific speech community. These factors would have to be kept in mind while framing the questionnaire and eliciting data for noun bases. Nonetheless, gender assignment and gender representation in Indian languages are rather interesting phenomena and should be pursued seriously. Field investigators have to be extra careful when dealing with highly inflecting languages such as Bagri, Hindi, and Konkani, where nouns ending in a consonant decline differently than those ending in a vowel, which again are different from those ending in +/- back vowels. The final syllable vowel feature [+/- round] also makes a difference. Indo-Aryan languages are the most complex of all as far as the nominal and verbal paradigms are concerned. Overt markings of gender and number behave differently in various case forms, too. Let us consider some of the varying paradigms of noun declension in

120 A Manual of Linguistic Fieldwork and Structures of Indian Languages oblique and non-oblique cases in different languages. As we said earlier, not all languages make a three-way distinction in gender. Khasi always marks nouns by attaching a gender-marking clitic before the noun concerned. Thus: Nominative and oblique masc. Singular

u-briew ‘man’

Nominative and oblique fem. singular

ka- briew ‘woman’

Nominative and oblique masc./fem. plural

ki-briew ‘men/ women’

To distinguish ‘men’ from ‘women’, the word siNraN is suffixed to ki-briew as in ki-briew siNraN ‘men’. In contrast with this pattern, Konkani poses a different picture altogether. It should be reminded that Konkani like Sanskrit has many nountypes depending upon the quality of vowels and consonants the last syllable of the word contains. We shall give the following table to familiarise the readers with the complexity. Consider the morphophonemic changes that oblique forms have incorporated. Table 5.2 Nominal Inflection in Konkani (Choraon dialect, Goa) Nominative

Nominative

Oblique

Singular

Plural

Singular

Masculine ‘boy’

burgE-an

Feminine ‘sister’

boin-an

Neuter ‘animal’

monzat-an

burgiâ-an

Oblique Plural

burgE-

burgiâ-

boinn-iâ-an

boin-

boinni-â

monzatti-an

monzat-

monzatti-

Simplest among the Indo Aryan languages are Oriya, Samblapuri and Laria, all spoken in Orissa. Table 5.3 Nominative Singular

Sambalpuri Nouns Nominative

Oblique

Oblique

Plural

Singular

Plural

cHua-

cHua-mane

’boy’

cHua

cHua-mane

‘girl’

Tukel

tukel-mane

tukel-

tukel-mane

If you remember, we discussed male and female speech in section 3.8.3 with reference to Kurux. Let us see how the category noun is inflected for various number. Nouns in nominative and other cases are not distinguished phonologically.

5. Morphology and Word Formations 121 Table 5.4 Gender, Number and Case markings in Kurux Male and Female Speech Singular Plural Nominative/oblique

kukko-s

kukko-r

kukko-e

kukko-e-r

masculine (male speech) ‘boy’ Nominative/oblique feminine (male speech) ‘girl’ Nominative/oblique

kukko-s

kukko:-xaddae /kukko-r

kukko-e

kukkoe:-xaddae/

masculine (female speech) ‘boy’ Nominative/ oblique feminine (female speech) ‘girl’

kukko-e-r

Kharia, which is a Munda language, shows all the three numbers distinctively and in addition to this, because of being agglutinative, shows clear-cut morphemic boundaries. Also to be noted is that, unlike animate nouns, inanimate nouns in Kharia are not marked by the dual suffix -kiyar and plural -ki as shown below, but are rather prefixed by a numeral. The language does not make any overt distinction between the nominative and oblique forms. Consider: Table 5.5

Gender, Number and Case markings in Kharia Nouns

Nominative/Oblique Masculine

Nominative/Oblique Feminine

Nominative/ Oblique Inaniamte

Singular

Kun}u ‘child’

beÿi ‘child’

soreN ‘stone’

Dual

Kun}u-kiyar

beÿi-kiyar

do-go-soreN two-cl-stone

Plural

Kun}u-ki

beÿi-ki

paâc-go-soreN five-cl-stone

There is another category that one has to look for in noun inflection, and that is the distinction between definite and non-definite. The case in point is Sinhala (Gair and Paolillo 1997:17). Sinhala has a natural gender and the main distinction is between animate and inanimate. Sinhala nouns inflect for plural (exception is some mass nouns) and for singular. The singular nouns (both +/-animate) are further inflected for definiteness and indefiniteness in all cases. Thus pot«ÿ« ‘book’ (Dat.+definite sg.) but pot«k«ÿ« (Dat.+ indefinite sg.) but no distinction in plural which is potw«l«ÿ« (Dat. pl.).

122 A Manual of Linguistic Fieldwork and Structures of Indian Languages

!

After collecting different noun bases one must elicit data on nouns occurring in various positions, such as subject of, direct object of, indirect object of and complement of the main verb. This may give various inflectional paradigms. Various paradigms given here will inform the readers of the variety of the ways one could visualise noun morphology as it appears in Indian languages. Depending upon which language family and which geographical location you are investigating, you must modify the sample phrases for elicitation. One must keep in mind that gender and number show up in some united fashion that makes it difficult to segregate one from the other. Thus, when you make sample phrases include all the possible gender and number distinctions you have in mind. 5.1.1.2

Noun Classifiers

Nouns can be classified by a system of classifiers. Aikhenvald (2000) makes a very interesting statement in the preface of her book regarding classifiers. “Languages with extensive systems of noun classification devices, especially those which combine classifiers and genders, present a true challenge for the typologist.” Well, Indian languages are just these. Though not an areal feature, languages of the Munda family and languages surrounding it geographically, such as Malto, some Bihari languages, Assamese, and Bangla are known to use numeral classifiers. Some of the Tibeto-Burman languages such as Boro, Burmese, Kokborok, Garo, Apaatani etc. (Bhat 2000) also use a classifier system. Khasi has also been reported as using classifiers, though not very extensively. Malto uses a very complex and intricate system of numeral classifiers (see Mahapatra 1997). Classifiers in Malto are used to categorise noun in terms of animacy, form, shape and size. There are also a few unique classifiers, each used with a single noun. Nouns are divided into two genders: male human vs. the rest in singular; human vs. non-human in plural. If one is working on any of the languages listed here one should be prepared to find nouns classified in various groups depending upon the world view of the speakers and their cognitive ability to distinguish objects. The order of numeral and classifier generally but not necessarily is Number + Classifier + Noun, e.g. Bangla tin ÿa boi ‘three + cl + books’ ‘three books’. Mahapatra reports three basic types of numeral classifiers for Malto indicating (1) status, (2) state and (3) quantity, which are further divided into many sub-sections. For instance, classifiers of quantity can be seen as various kinds of measures measuring depth, length, volume, weight, girth, time, and an enumerative numeral used for counting. Similarly, shape classifiers can be further subdivided according to the size and shape of the object under consideration. Consider a small sample of Malto classifiers. The words that take part in forming the classifier need not be indigenous to that language. j«n (IA)

for human beings

maq/pa:n

for non-humans

da a (IA)

for long and large objects

kaˇi (IA)

for long, small objects

panda (IA)

for long, flexible objects

5. Morphology and Word Formations 123 pa}a (IA)

for long, pods/fruits

paˇa (IA)

for flat, broad objects

pata (IA)

for flat, thin objects

ka˜êa (IA)

for flat, broad cotton objects

goˇ (IA)

for round heavy objects

pula (IA)

for round, light objects

“It reveals that the classificatory system is not merely an arbitrary or ad hoc collection of phenomena, but the result of sustained intellectual preoccupation with the environment towards formulating a conception of reality and beyond it, a natural philosophy of ethno-metaphysical speculation.”

(Mahapatra. 1997: 221) Khasi adds the classifier si-kanhon to the +count noun class, as in si-kanhon ki hati ‘a herd of elephants’, si-kanhon ki- kHun- kHshan ‘ a litter of puppies’.

!

While collecting data on number phrases in languages of Jharkhand or of the Tibeto-Burman branch one should ask and test model sentence employing various kinds of nouns. Thus, objects of different shapes, sizes, surface structure, as well as those that can be counted etc. should constitute separate sentences for elicitation. There might be a possibility that initially you may not get all kinds of classifiers, but living in the field for an extended period and keen observation will ultimately get you what you want. Another very interesting language that has come to our notice for offering a variety of classifiers is Andamanese. There are formative affixes (Manoharan [1997] does not call them classifiers, but they behave almost like typical classifiers) that have specific semantic domains and co-occur with pronominal prefixes (as all nouns do) to classify different nouns. Consider the following examples: Andamanese i-be:c

‘honey comb’

ara:-bec

‘tail (of birds)

tot-be:c

‘feather’

ÿ!-ot-be:c

‘my hair’

ÿe-erta:p-be:c ‘my beard’ et-cow

‘fruit (of tree)

ettay-cow

‘kidney’

ekko-cow

‘knot’

(Manoharan 1997:469)

124 A Manual of Linguistic Fieldwork and Structures of Indian Languages

!

To elicit data on classifiers some of the basic questions that one should ask

are: 1. Is there any overt noun class marking? If yes, is it a bound or free morpheme? 2. Does the classifier occur only on the head noun or also on a modifier? If it occurs on both, what are the syntactic and semantic differences? 3. What is the correlation of classifiers with other categories such as number, gender, and adjectives? 4. What are the semantics of the classifiers? That is, on what basis are they assigned? 5. Are classifiers obligatory? If not, why not? 5.1.2

Pronoun Morphology

Pronoun morphology pertains to identifying the different forms of pronouns that are encoded with various kinds of information regarding gender, number, animacy, and several deictic categories. As in the case of nouns, pronouns in Indian languages also occur in a wide range of forms. Hence, data should be collected on personal pronouns for two/three genders, all the numbers, i.e. singular, dual and plural, and the person categories as the case may be. In all, data should be elicited to unearth the morphological structures of pronominals of the language. The overlap between inflectional and derivational categories as well as between morphology and syntax is inevitable. The field investigator must be aware of this. We will not discuss each and every topic listed here; however, the questionnaires given in the appendices will be of great help to identify the typology of inquiries that one must address herself/himself to. You may also refer to the ‘basic sentences’ given in the appendices.

!

The foremost question concerns how to recognise pronouns in a language. It is fairly simple to do so as pronouns, being the replacement of noun phrases, are never modified by any adjectives or adjectival clauses, and don’t carry any determiners or definitisers with them. For this purpose, the field investigator should collect information about various pronominal forms listed below. This is not an exhaustive list but nonetheless will help reach a fair amount of information on the pronominal morphology of the language concerned. Various Pronominal forms found in Indian languages. 1.

Personal pronouns

2.

Demonstrative pronouns

3.

Interrogative pronouns

4.

Indefinite pronouns

5.

Possessive pronouns

6.

Reflexive pronouns

7.

Reciprocal pronouns

8.

Deictic pronouns represented by the third person

5. Morphology and Word Formations 125 9.

Relative–correlative pronouns

We would like to draw some examples from Santhali, a language of the Munda family spoken by a large number of tribal communities of Jharkhand. Of all the languages that we have worked on up to now, Santhali offers the maximum of possibilities of distinct and varying pronominal forms. The varying forms listed here will enable the field linguist to prepare her/his questionnaire accordingly when collecting data on pronominals in any language of this family, or would at least prepare her/him for the various possibilities. Table 5.6 Personal Pronouns In Santhali I person

II person

III person

Inclusive Exclusive Singular

iN

Dual

ala:N

Plural

abu

a:m

uni/nui

ali:N

aben

unkin

al

ap

unku

It is to be noted that many Indian languages make a distinction between inclusive (i.e. addressee included) and exclusive (addressee excluded) first person dual and plural pronouns.

!

To elicit data on the inclusive/exclusive distinction one has to collect sentences and confirm with the speaker whether the word ‘you’ or ‘I/we’ included the addressee or not. This can be easily done by framing a sentence in a context first and then pointing towards the addressee whether s/he is also included. The same format can also be used for reciprocals and reflexives. Consider: Table 5.7

Reflexives in Santhali

I person

II person

III person

i: t

a:mt m

ai/tt

ala:Nt:N ali:Nt i:N

abent ben

akint kin

abut bu

ap t p

akut ku

Inclusive Exclusive Singular Dual Plural

al t l

The category of ‘inclusive/exclusive’ is relevant only in the first person dual and plural. A blank table drawn on the format given above for collecting reciprocal pronouns can also be made. Santhali distinguishes between various types of the possessed entity, which shows up in the distinct forms of the possessor pronouns. The following conversation revealed to us that the language makes very fine distinctions who possess what.

126 A Manual of Linguistic Fieldwork and Structures of Indian Languages FI

how would you say ‘ram killed my cat’

I

ram i:ørEn pusi goet’kedea

FI

How would you say ‘my cat’?

I

i:ørEn pusi

FI

can you say i:ørEn janamdin (a common word for ‘birthday’ in Hindi and Santhali)?

I

No, it should be i:øa/ janamdin

FI

How would you say ‘my house’?

I

i:øa/ oêa

FI

(knowing it might be wrong but still for the sake of testing) can we say i:øa/ ku}a (ku}a is ‘boy’ in Santhali)?

I

(laughing) No, we say i:ørEn ku}a. He is not a house. Conversation piece no 5.1

The informant gave two distinct genitive forms to us because ‘house’ and ‘birthday’ are inanimate nouns while ‘cat’ and ‘boy’ are animate nouns. One should be prepared to find such distinct forms in other languages of the region too. As pronominals in this language are marked for three numbers and first person makes a distinction between inclusive and exclusive one gets a large number of forms in the language each earmarked for distinct possessor-possessed relationship. Consider the Santhali pronominals of genitive/possessive given below in the form of a table. Notice that a distinction in forms is made due to the animacy of the possessed nouns, and the inclusion or exclusion of the addressee pronoun. Table 5.8 Genitive/Possessive Pronominals in Santhali Possessor 1 sg 1dual, excl 1dual, incl 1 pl excl 1pl incl 2 sg 2dual 2 plural

Possessed inanimate iøa/

Possessed animate i:ørEn

ala:Na/

ala:NrEn

abuva/

aburEn

ali:øa/ alEya/

ali:ørEn alErEn

ama/

a:marEn

apEya/

apErEn

abena/

abenrEn

3sg, masc fem

uniya/ aya/

unirEn a/it

3sg, neut

nuirEa/

nuirEn

3 dual

unkina/

unkinrEn

3 plural

unkuva/

unkurEn

5. Morphology and Word Formations 127 Santhali is an incorporating type of language and thus various pronominal affixes are added to the verb form without any subject or object pronouns overtly showing in the sentence. Please refer to chapter 2 for details.

!

When eliciting data on pronominals, especially on the languages of the Munda group, one must remember that pronominals have various ways of showing up in the language. Do not forget to collect possessive constructions with abstract nouns, e.g. ‘my love’, ‘his honesty’ etc., though we must warn the investigators that it is a daunting task to collect abstract nouns. There is no simple way of explaining the abstract nouns and most often than not your clue word for an abstract noun would be translated as an adjective in the language. Andamanese offers a very elaborate pronominal system with a distinction being made in [+/-honorific] forms for second and third person pronouns. This is in addition to the inclusive/exclusive distinction being made in the first person singular and plural pronouns (Manoharan 1997:464-65). Also to be mentioned here is that since interrogative pronouns have the potentiality of occurring in reduplicated forms carrying [+distributive] meaning as in Hindi k!n k!n aya ‘who all came’, one should try to reduplicate all the interrogative pronouns collected so far, keeping the rest of the elicited sentence intact. Test this with the informant, and if s/he approves of it ask for the semantic change that must have occurred due to reduplicated pronouns. You may realise that not all types of pronouns reduplicate easily (see also section 5.3). 5.1.3.

Case Markings and Postpositions

It is universal feature of human languages to have cases. Cases exhibit the relation between verbs and nouns (or pronouns) in a sentence. It may also indicate the relation between two nouns, as we saw in the locative in Santhali (section 4.3.2) and the genitive in Kurux (4.3.1). This relation can be exhibited by rich inflectional endings on nouns as in Sanskrit and Konkani. Sometimes it is demonstrated by adpositions known as pre- or postpositions occurring with or without any morphological change in the nouns they are attached to. Hence, one must distinguish between case and case markers. The former is a semantic relationship while the latter exhibits this very relationship by some phonological word. Most of the nonnominative case markings in Hindi are added to the nouns ending in –a only after they are transformed into an oblique form. Thus, l«}ka g«ya ‘the boy went’ but l«}keko «ccHa l«ga ‘the boy-Dat. felt good’. Pronouns can also take various postpositions exhibiting different cases. We shall discuss cases in some detail below. At present it is necessary to remember that one can elicit ten phonologically distinct case markers allowing for distinctiveness of every gender and number as in the case of Konkani genitive. We generally carry a blank matrix chart in the field and fill it up as and when we elicit the required form. Another fact to be kept in mind is that many languages offer several distinct case forms for the same case relation. For instance, Konkani marks the accusative by ar, or ka, or ak or ek; the dative by ak or ka-lagun; the instrumental by en, yEn, yan, ka÷ (pl) or by ap; the ablative by ace/aco or len, or by sun; the locative by cer, ka÷ (pl) eâk, yan (pl) or by ant; and the comitative by ak-aca, b«r!b«r, and s«Nata. Most

128 A Manual of Linguistic Fieldwork and Structures of Indian Languages often the different forms of the same case relation are dependent upon the semantic nature of the various nouns involved in the case relations. There could be very many possibilities and field worker should be open-minded to test these. S/he should try to test various combinatory possibilities of nouns and verbs and nouns and nouns (or pronouns). Let us consider how these case relations are marked by postpositions (Khasi being the exception here, since it has prepositions).

!

To elicit data on cases, begin with making six basic types of sentences, each involving a case frame for the nominative, accusative, instrumental, dative, locative, and genitive. After this you may add other cases such as comitative, allative, ergative, vocative etc. mentioned earlier in chapter four as well as discussed as below. 5.1.3.1

Postpositions

Generally speaking, Indian languages exhibit ten case relations with distinct postpositions (with the exception of the vocative; see below). Some languages first change the nouns into an oblique form as we saw in the case of Hindi above, before the various postpositions are attached to them. Some Indian languages, as discussed in chapter 2 also have an ergative case (Tibeto-Burman) or demonstrate split ergativity such as only in the perfective (see the Hindi examples given in 4.5.3.1). For the convenience of discussion we shall club nominative, agentive and ergative together as agentive. The ten commonly found case relations that may be marked by various postpositions are listed below. 1.

Agentive

2.

Accusative

3.

Dative/Experiencer

4.

Ablative

5.

Genitive

6.

Locative

7.

Instrumental

8.

Comitative/Associative

9.

Allative

10.

Vocative

Two important features of Indian data have to be kept in mind. Firstly, in languages that mark agreement heavily, the governing noun triggers various different forms of genitive markers that exhibit gender and number of the possessed nouns (we saw Santhali data in chapter 4, table 4.5, as well as in table 5.8). Secondly, it is generally observed that the locative case is overtly marked with various postpositional markers, each depicting the type of the location as we saw in the case of Santhali. All of them can be marked overtly by postpositions. Sometimes the same postposition may indicate two different case relations, i.e. they may be homophonous in nature, a situation reverse to what we saw earlier in the case of Konkani where one case relation was represented by several different postpositions.

5. Morphology and Word Formations 129 According to the case systems found in the languages of the world, one can classify languages into two broad categories. 1.

Nominative–accusative

2.

Ergative-absolutive.

The distinction is made according to the agreement rules governed by the subject and the object of the sentence. In nominative-accusative systems the subject of transitive and intransitive predicates are marked in a similar way while the object is marked differently. In ergative-absolutive systems the subject of the transitive predicate is marked differently from the subject of the intransitive and object of the transitive predicate that are marked similarly. For the sake of convenience we give these two types of systems in a tabulated form. Table 5.9

Two-Way Case System in Human Languages

Marked NomAcc

O

Erg-Abs

A

Unmarked A, S

Examples Mary-f saw Jane- na Jane- f ran

O, S

Mary- na saw Jane-f Jane-f ran [-na is the marked case ending]

Generally, in this system, the ergative replaces the nominative case marking for the transitive subject and the absolutive takes place of the accusative case marking of the transitive object as well as for the intransitive subject. Indian languages exhibit both types of case systems. Moreover, it is not necessary that each language follows only one type. We just saw that Hindi is a language with split ergativity. It shows the nominative-accusative system in the future and present tense and ergative-absolutive in the perfect tense. Two case relations, genitive and locative, need some clarification. These are the only two cases which specify the relation between two noun phrases/nouns rather than specifically between nouns and verbs. Hence the nature of the two nouns involved governs the form of the case marking. Most of the time, Indian languages make a distinction between inalienable and alienable possessions which are represented in distinct case markings. Thus, in Hindi ram ko do l«}ke hEâ ‘Ram has two sons’ but admi ke do hatH hEâ ‘men have two hands’. Interestingly, in modern Hindi this difference is disappearing. However, Meithei still makes this distinction. Moreover, the gender and number of the possessed nouns may also govern the form of the genitive (we saw this earlier in the case of possessive pronouns of Santhali). At times when both the possessor and possessed nouns are inanimate, a distinct genitive form is used. For example in Santhali, although the common genitive is a/, rEa is used when both the possessor and the possessed involved are inanimate, e.g.

130 A Manual of Linguistic Fieldwork and Structures of Indian Languages ga a rEa? da? ‘river’s water’, dari rEa? s«ka:m ‘leaf of the tree’, se:v rEa/ da:m ‘price of the apple’ but uni-a/ usu:l ‘his height’, sitta-a/ ca bç l ‘dog’s tail’ and khalid-a/ ti: ‘Khalid’s hand’. Konkani uses distinct genitive case markings for masculine singular and feminine singular, but for plural it uses the same –e case marker for masculine and feminine. Another interesting fact about the genitive is that in many inflecting languages the genitive morpheme inflects for gender and number of the head. Thus in Hindi ram ki ÿopi ‘Ram gen-fsg cap-fsg ’ ‘Ram’s cap’, but ram ka gH«r ‘Ram genmsg house-msg’ ‘Ram’s house’. In Punjabi the genitive inflects for masculine and feminine plural nouns, e.g. ram de mu÷êe ‘ram gen mpl boys’ ‘Ram’s sons’ but ram diyaâ t`iyaâ ‘ram gen-fpl daughters’ ‘Ram’s daughters’ The location of an object and the quality of the placed object are perceived differently by various Indian communities. It can be seen that locative case markers come in varying forms in Indian languages, depending upon the type and the nature of the location perceived by the speech community. We have seen examples of various locative case markers in Santhali in section 4.3.1. Bangani is another language that we have come across that makes minute distinctions in location types. Thus, the locative ‘in’ in Bangani has five equivalents manifested in -di, -kE,-ari poru, and -mai, the use of each decided by the container that holds the object. Locatives in Konkani are also worth noting. Consider: cer

‘on, upon’

eâk

‘at, upon’

ant…ka÷

‘from, in to, inside’

yan…an

‘in, on, to’

Another important aspect about case marking in Indian languages is that they may come in compound forms. For instance, the dative in Hindi is ke liye ‘for’ which is a compound form of genitive + dative, the comitative is ke sa:tH ‘with’ which is a combination of genitive and comitative, or various locative case markers come in compound forms of genitive + locative postpositions, such as ke «nd«r ‘inside’, ke picHe ‘behind’, ke pa:s ‘near’, ke ba:h«r ‘outside’, ke up«r ‘on’, etc. Konkani offers the following combinatory possibilities of case marking within a sentence: Accusative + locative ek + ka÷ Genitive +comitative a + b«rob«r Nominative +accusative + instrumental + locative Ablative + locative wailen + an Genitive + dative a + lagun Agentive + ablative an + le

an + ak + an + eâk

5. Morphology and Word Formations 131

!

While collecting data on case relations one must try to elicit wide-ranging possibilities of locational and possessive situations. The best way to elicit data on location and genitive cases is to start with a sentence referring to the pocket of your shirt, or contents of your purse, or ink in your pen, objects seen easily by the informant and the investigator alike. We generally start with: • Pocket of this shirt…. •

Money in this pocket….

•

Ink of this pen…..

•

Ink in this pen…..

•

Comb in my purse

•

People in this house/school

•

Paddy in the field

•

Fish in the water

From there you may continue. Think of as many possibilities as the situation demands and try to use as many types of nouns as possible showing various locations and various possessions 5.1.4 Morphology of Adjectives Typically, adjectives are modifiers of their head nouns and hence also dependent on them. Characteristic feature/s of modifiers in attributive position and predicative position should be kept in mind while working on a language of a specific family. Dixon (1991) establishes ten semantic types of adjectives, mostly based on English examples. We shall add one more on taste adjectives as they form a distinct class when reduplicated (Abbi 1992). If we add quantifiers of the numeral and enumerative kind to this list, as they modify the noun for specificity, the list will have twelve adjectives in all. We shall give them here for quick reference. Table: 5.10 Semantic Prototypes of Adjectives Semantic Types 1. Dimension

Examples big, great, short, thin

2.

Physical property.

hard, strong, clean, sick

3.

Speed

quick, fast, slow, sudden

4.

Age

new, young, old, modern

5.

Color

white, black, red, blue

6.

Value

good, bad, strange, odd

7.

Difficulty

easy, difficult, tough, simple

8.

Qualification

definite, possible, normal,

9.

Human propensity

angry, happy, eager, clever

Similarity

like, unlike, similar, different

10.

132 A Manual of Linguistic Fieldwork and Structures of Indian Languages

11.

Taste

sweet, sour, bitter, hot

12.

Quantifiers

two, twice, both, some

Languages of India give two possibilities. Either adjectives agree with the noun they modify in gender, number or person or they stay invariant. Highly inflecting languages such as Punjabi and Konkani inflect nominal modifiers for gender and number while Hindi inflects for gender and masculine number. Eastern Indo-Aryan languages such as Bangla, Sambalpuri or Oriya keep the adjectives invariant. Hindi 5.9

g«ndi l«}ki

‘dirty girl’

5.10

g«nda l«}ka

‘dirty boy’

5.11

g«ndi-f l«}kiyaâ

‘dirty girls’

5.12

g«nd-e l«}ke

‘dirty boys’

5.13

g«ndi ku}I

‘dirty girl’

5.14

g«nda mu÷êa

‘dirty boy’

5.15

g«nd-iyaâ ku}-iyaâ

‘dirty girls’

5.15

g«nd-e mu÷ê-e

‘dirty boys’

5.16

bHalo-f me:

‘good girl’

5.17

bHalo-f cHele

‘good boy’

5.18

bHalo-f me:y-ra

‘good girls’

5.19

bHalo-f chele-ra

‘good boys’

Punjabi

Bangla

Adjectives can be used in attributive position followed by a noun or in a predicative position followed by an optional auxiliary. Fieldworkers should try to elicit both types. Languages of the Austro-Asiatic branch would obviously not show much difference. Languages that show agreement features on attributive adjectives will show the same features if adjectives are used predicatively. Tamil, however has an interesting pattern. In attributive position adjectives are not marked for gender and number, but when they occur in post-nominal position they are nominalised and they agree with the preceding noun in gender and number (Annamalai 1997). Consider: Tamil 5.20 ki}inja saÿÿeye eêu ! saÿÿe

ki}injade

eêu

5. Morphology and Word Formations 133 tear-pst-adj shirt-acc take !shirt tear-pst-adj-it-acc take ‘take the torn shirt’ 5.21 nalla peenaavaa vangu ! peena good-adj pen-adv buy ! pen

nalladaa

vaangu

good-adj-it-adv

buy

(Annamalai 1197:21)

!

Elicitation for adjectives is best done when you try to collocate modifiers with appropriate nouns pertaining to the local situation. For instance, ‘sweet pineapple’ or ‘blue sofa’ has no relevance in a village of Central India, or ‘ice-cold’ in Rajasthan. Always look for names of the most commonly used items for eliciting appropriate modifiers. A Konkani speaker had a tough time once trying to give us the equivalent of ‘high waterfall’, since, as we realised soon, there is no term for a waterfall in the language, as perhaps none exists in the topography of Goa. 5.1.4.1

Qualifiers

Adjectives can always be preceded (in SOV languages) by a qualifier of increased or decreased intensity: Hindi: 5.22

ye

am

z«ra kH«rab hE

This mango little bad

aux 3msg

‘This is a little rotten mango’ Sinhala: 5.23 mee bohom« alut pot« This very

new book-def

‘This very new book’

(Gair & Paolillo 1997:30)

!

After collecting a fair number of adjectives one can elicit data on qualifiers by giving the sample sentence in the contact language or by explaining to the informant by facial expression. The latter strategy works faster and without fail. The very same qualifiers can be employed with verbal modifiers, too. So once you elicit these remember to use them again when eliciting data on adverbs. 5.1.4.2

Quantifiers

The reason for considering quantifiers under the head ‘adjectives’ is that quantifiers in Indian languages behave very much like any other nominal modifiers. Firstly, quantifiers in highly inflecting languages such as in Hindi and Punjabi inflect for the gender and occasionally also for the number of the noun they modify. Thus Hindi has: 5.24

sara

kunba

all-msg family-m-mass noun

‘entire family’

134 A Manual of Linguistic Fieldwork and Structures of Indian Languages 5.25

5.26

sari

daal

all-fsg

dal-f-mass noun

sare

b«cce

‘entire dal’

all-mpl children-m- pl 5.27

sari-f

‘all children’

roÿ-i-yaâ

all-fpl roti-f-pl

‘all rotis (bread)’

Punjabi shows additional inflection for feminine plural quantifiers just as it shows for other modifiers. Thus sariyaâ roÿiyaâ ‘all-f-pl bread-fpl’ would refer to ‘all the bread’. Another reason for considering quantifiers under adjectives is that in Dravidian languages such as Tamil, quantifiers can also occupy postposed nominal position just as any other adjective. In addition to this feature, Tamil quantifiers in predicative position can also agree in gender and person with the noun it quantifies. In other words, languages that are not heavily inflecting, such as Tamil, also treat quantifiers the same way they treat other nominal modifiers, i.e. adjectives. Tamil 5.28

oru maaêu

!

One cow

maaêu o÷÷u cow

one-it (non human)

‘one cow’ 5.29

oru payyan !

payyan oruÿÿan

one

boy

boy

one-he (human)

‘one boy’

!

The four types of numerals, i.e. cardinals, ordinals, aggregatives, and multiplicatives that are listed in section 4.5.4 may be collected at this time. It is easy to collect data on them as you don’t require long sentences to elicit numerals and you have an additional advantage of showing the quantity under consideration right on the spot. For this, any item should be used for elicitation, twigs, pebbles, small stones, or wheat grains. It is also interesting to study counting systems across languages. 5.1.4.2.1.

Partitive and Non Partitive

Numerals and quantifiers are generally divided into two kinds, partitive and nonpartitive. Data should be collected to elicit both kinds. In many languages, the distinction between the whole and its part dictates that the part follows the whole. Thus in Hindi (a plus symbol is used between the whole and the part): 5.30

klas meâ se+ do l«}ke aye

5. Morphology and Word Formations 135 whole

part

‘two of the boys from the class came’ 5.31

sari bHi} meâ + kucH logoâ ne……. whole

part

‘some people from the crowd…’ There is also the possibility of finding negative partitive or non-partitive quantifiers. 5.32

sari bHi} meâ + koi bHi n«hiâ….. whole

part

neg.

‘no one from the crowd….’ There is another very productive morphological marker working as a clitic in most of the Indo-Aryan languages which functions as a modifier of a referent in the phrase. For the sake of convenience, we will call it the wala-construction (see 5.2.1), as in Hindi phrase ane wali ga}I ‘the coming train’ or better ‘the train that is about to come’, or lal wali k«miz ‘the red shirt’. As it behaves like a restrictive relative adjective but has the structure of an emphatic particle and behaves morphologically like a derivational unit, we will consider it later on. While working on modifiers one must also elicit data on various grammatical strategies that a language might adopt to adjectivalise an NP. Thus sentences of relativisation both of restrictive and non-restrictive types as well as of participialisation of the NP type should definitely enter the questionnaire. As they involve long and complex sentences with embeddings we shall wait and discuss them in the next chapter. 5.1.5 Stage IV Simple Sentences Simple and short sentences are defined as those which do not involve embeddings. They are collected basically to study word order, government, coordination, placement of negatives and interrogatives, and adjectival phrases of non-relativised types. At this stage we begin building up on the phrases that we elicited in stage three. 5.1.5.1

Comparison

Generally three types of comparison are seen in languages, equative, comparative, and superlative. Equative words follow the same pattern of inflection as the other modifiers we have just seen. For instance, Hindi equative utna ‘that much’ and jitna ‘as much’ like any other modifier inflects for gender and number. Equative

136 A Manual of Linguistic Fieldwork and Structures of Indian Languages

5.33

meri

sa}i

utni

buri nahiâ hE jitni

radHa ki

my-fsg saari-fsg that much bad-fsg neg aux as much-fsg Radha’s-fsg ‘ My sari is not as bad as that of Radha’s’ Comparative Comparatives are formed with a comparative marker, which may be a postpositon in the instrumental or dative case and a degree-modifying lexeme for ‘much’ or ‘less’, as the case may be (which may be optional) in most of the Indian languages. Consider a sentence from Laria and Hindi each for reference. Laria 5.34

nu gora E

ram ke

b!hin !k!r

bHai

Ram-geni

sister hisj

brother cm fair aux

Hindi 5.35

ram ki

b«h«n uske bhHai se zyada gori hE

Ram-geni sister hisj brother cm much fair aux ‘Ram’s sister is fairer than his brother’ Superlative Unlike English, most of the languages have only one way of forming superlatives, i.e. by periphrastic constructions, and not by morphological inflection or derivation. Indo-Aryan languages such as Hindi and Punjabi use the comparative construction preceded by the lexeme for ‘all’ Consider: Punjabi 5.36

m«nnu s«b -toâ k«`ÿ miÿhiaii milli I-dat all-cm less sweets got ‘I got the least amount of sweets’

!

Degree modifiers for ‘much’, ‘less’, and intensifiers/qualifiers ‘very’ or ‘little’ should be collocated with various kinds of adjectives. These can also be used with adverbs or simply as adverbs also (e.g. in Hindi ve zyada bolte hEâ [they much speak aux] ‘they speak too much’). Elicitation of data on comparison should begin with comparing two items present on the scene of elicitation. This will let you decrease or increase an attribute according to what you desire to achieve. We generally start by drawing two lines of different length on the floor or on a slate with different coloured chalks, if possible, and start as follows: FI

How would you say the ‘white line is longer than the pink one’?

5. Morphology and Word Formations 137 Or

How would you say ‘the pink line is shorter than the white one’?

You can modify the drawing a bit and can further ask: FI How would you say that ‘the white line is more crooked than the pink one’? Or ‘The pink line is less crooked than the white one? Or ‘The white one is thicker than the pink one? And so forth. 5.1.6 Morphology and Syntax of Adverbs Verbal modifiers or adverbs of all the nine types listed in section 4.6 should be elicited at this stage. Languages vary in their strategies of marking adverbs. Most of them have one word operating as a modifier of the action or state depicted by the verb, but still others have an affix added to the verb under consideration. One should start with manner adverbs as these are commonly used by all. While eliciting data on manner do not forget to elicit material on expressives, as these form a very large inventory of manner adverbials in all Indian languages, specifically in the languages of the northeast. Languages such as Boro spoken in the state of Assam and Meghalaya form adverbs by suffixes added to verbs. Consider: Boro

za-glo"

‘to eat quickly’

za- bay"

‘to eat again and again

za-klay "

‘to eat from top to bottom’

za- kma"

‘to eat stealthily’

za-do"b

‘to sit properly and eat’(Bhat 2000)

It is equally important to elicit temporal, spatial, locational, and directional adverbs. One can make very relevant sentences pertaining to various situations to elicit data on adverbs. In addition to these strategies, i.e. words vs. affixes, one can also have small clauses as modifiers of the main verbs. Sentences (1) using conditionals, (2) purpose, (3) reason (4) showing cause and effect, can all be used as verbal modifiers. Most of them behave like complements which we are going to discuss at length in the next chapter. We postpone discussion on them now because they involve eliciting long and complex sentences. At present we are concerned with short sentences without embeddings. As we said earlier in the section of nominal modifiers, verbal modifiers can also take various types of intensifiers to further modify the modifiers. Consider:

138 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Hindi 5.37

z«ra ahista bolo little soft speak-imp ‘Speak a little softly’

Laria 5.38

ra:jdHani Rajdahani

kEte ni kEte jor jatHe very very fast go-3sg

‘Rajdahani goes very very fast’

! §

To elicit information regarding adverbs, one must first elicit short phrases without any adverbs attached to them and then the same phrases with adverbs attached to them. The best technique is to make sentences with already collected material and attach each adverb that was collected at stage one and two one by one. This way you don’t have to take recourse to the translation method. The most popular sentence with us is ‘please speak slowly slowly’.

§

Do not rule out the possibility of adverbials agreeing in gender in inflecting languages such as in Hindi. One can encounter the following sentence:

5.39

mEâne gend

ucâ i pHeâki

I-m/f-erg ball-f high-f throw-pst 1sgf ‘I threw the ball high’ §

While in the field you will be watching people involved in some activity or the other. Right in the field you can ask about or make half a sentence referring to the activity seen and the rest is supplied by the informant. Very often, one gets rare words that cannot be translated into the contact language by eliciting adverbs by this method. Our experience is that many manner adverbs, including those that are represented by expressive morphology, are best elicited by observing an activity and pointing to it to get an appropriate modifier.

Adverbs in Indian languages are not only depicted by a single word of adverbial component but also by the following grammatical constructions. An example each from Hindi is given for quick reference. •

Explicator compound verbs3:

5.40

vo ga

uÿHI

She sing RISE past 3fsg ‘She sang suddenly’ 3

The explicator or vector, as they are generally known, is given in capital letters in this manual.

5. Morphology and Word Formations 139 •

Reduplicated verbal adverbs

5.41

m«ntri ne

bHash«÷ diya

kH«}e kH«}e

minister-erg stand-past part stand past part speech GAVE ‘The minister delivered the speech standing(ly)’. •

Preverbal participialised constructions

5.42

Ragini h«âste hue

boli

Ragini laugh prs part 3sg spoke 3fsg ‘Ragini spoke laughingly’ •

Conjunctive participle

5.43

Arun ne musk«ra k«r

meri

t«r«f

dekHa

Arun-erg smile- cp

my-fsg direction-fsg saw

‘Arun looked at my direction with a smile (or smilingly)’ Expressive morphology 5.44

Ram h«}b«}a-k«r uÿHa Ram exp – cp

got up 3msg

‘Ram got up with a startling motion’

5.1.7 Verb Morphology The most crucial and significant of all parts of a sentence is the verb. The verb is for a sentence what the heart is for the human body. It governs all other grammatical elements in the sentence. It determines how many nouns there are in a sentence, which in turn determines what kinds of modifiers accompany these nouns. In other words, the valency of the verb determines the semantic and syntactic nature of the sentence. By now you have a good idea of various types of verbs in the language of your investigation, yet to come to grips with various kinds of encoding that verb can potentially take one has to elicit data very systematically. As we are at the stage of simple sentences, it is not necessary to collect all that was listed under 4.6. at this time. At this stage one can concentrate on finding out how the language under investigation marks the following on the verb: 1. Tense (present, past, future, any other) 2. Aspect (perfect/perfective, imperfect, habitual, progressive) 3. Mood

(indicative, obligative, imperative, optative)

Considering the nature and typology of the Indian languages, we must at this stage investigate whether the verb in a sentence is coded for:

140 A Manual of Linguistic Fieldwork and Structures of Indian Languages 1. Subject 2. Direct object 3. Indirect object 4. Speech act participant 5. Any other information

This implies we look for verb agreement, i.e. we identify the occurrence of covariance of the various elements listed above (and more) and the verb in a sentence. We have to see what element/s of verbal morphology agree with which elements occurring elsewhere in the sentence. The question is how one should go about collecting verbs to determine various forms of encoding or agreement. The following suggestions could be of some help:

! •

Do not collect verbs in infinitive forms any more as you have enough of those by now. Start collecting by employing various types of subject and object nouns within a sentence.

•

You may start by collecting along semantic paradigms, such as action verbs like go, come, run, and action process verbs such as give, eat, etc. or by syntactic functions such as intransitive verbs run, go, transitive verbs drink, eat, and ditransitive verbs give, write. We generally try to cover the case frame matrix suggested by W. A. Cook (see the appropriate appendix) as it exhausts all the possibilities of varying forms of verbs.

•

Try to get the entire paradigm with different types of subjects and objects employed in a sentence. This might be a tedious and a boring job, but remember it is more boring for the informant to supply these forms than for you as you are getting closer to your goal of discovering a grammar of a language. To achieve the paradigm you must come prepared with a list of model sentence/s, each time substituting either the subject or the object but keeping the verb intact. Start with the present tense. See the following example from Hindi:

The model sentence is X eats food. Verb forms are underlined. 5.45

maâ

kHana kHati

hE

mother food eat-impf fem sg aux-pr sg 5.46

pitaji

kHana kHate

hEâ

father food eat-impf msg-hon aux-msg-hon 5.47

l«}ka

kHana

kHata hE

boy eat-impf msg aux-pr sg

5. Morphology and Word Formations 141

5.48

l«}ki kHana kHati girl

5.49

hE

food eat-impf fsg aux- pr sg

lE}ke kHana

kHate hEâ

boys food eat-impf mpl aux-pr pl 5.50

lE}kiyaâ kHana kHati hEâ girls

food eat-impf fsg aux-pr pl

Some linguists do not advise you to collect verb forms using kinship terms as subject nouns. However, we prefer this, as verb forms in Indian languages are marked for +/- honorific status and there is no harm in getting information about this along with gender, number and person at this point. Austro-Asiatic languages, mainly of the Munda family, as well as Khasi incorporate subject information in the verb. Consider:

Khasi 5.51

u-mohan

u-

rukHey

msg clt-Mohan msg clt-laughs-msg ‘Mohan laughs’ 5.52

ka-

bHarti

ka-

ruai

fsg clt-Bharti fsg clt- sing-fsg ‘Bharti sings’

Santhali 5.53

dal-

ket`-ku ta:m-

strike- pst-pl gen-

a-

ø

2sg-1sgSM

I struck your oxen’ 5.54

siri:ø-ida-i:ø sing-prs-1sg SM ‘I sing (am singing)’.

•

To break the monotony we generally work on two types of verbs alternately. For instance, one can take a motion verb along with the verb ‘to eat’ and collect data on both at the same time. Whatever the method you follow, it is advisable to keep separate pages for each paradigm in your notebook. This will ensure that you

142 A Manual of Linguistic Fieldwork and Structures of Indian Languages have ready-made material for analysis later on. Whether you stick to one paradigm or work on two will depend upon the willingness and intelligence of the informant you have. Sometimes after three sentences the informant understands that you are looking for various paradigms and s/he volunteers the information on her/his own. This obviously makes the task easier. •

Try to add simple modifiers (adjectives and adverbs) gradually.

•

Try to avoid experiential verbs at this point. Most experiential verbs in Indian languages have a dative subject and the verb, instead of agreeing with the dative subject noun, may agree with the experiential noun, further complicating the matter at this stage of elicitation. For this reason you should wait for the next stage when you collect complex sentences.

By the end of stage four our knowledge about phonology, morphology, morphonology and word formation should be quite advanced. This is the most extensive and comprehensive section of the field investigation. We observe that most of the grammatical information comes from this stage of field investigation. As Indian languages are highly encoded on verbs, short phrases with verbs and their several variant forms give a fairly good account of the grammar as well as of the sociolinguistic features encoded in the language. This is also the stage when morpheme-to-morpheme translation becomes unavoidable. Let us consider this aspect in some detail.

5.1.7.1 Tagging and Annotation As one collects short phrases, one could easily mark the morphemic breaks even in highly inflected languages such as Hindi. Glossing and morpheme-to-morpheme translations are absolutely necessary to achieve any right grammatical judgement. The whole process takes some time, which one should not try to cut short. We begin with whole sentence translation and then achieve smaller and smaller units translated bit by bit. For instance, first we ask the informant to translate the whole phrase followed by a word-to-word translation (indicated here by small breaks-) and then we try to analyze the elicited material and compare it with other related forms and reach the judgment of morpheme-to-morpheme translation. Thus, the Hindi word khaeâgi ‘ (she) will eat’ can be shown as kha-e‚-g-i where each break shows a distinct morpheme (eat-3pl-fut-f) which contrasts with kha-e‚-g-e ‘eat-3pl-fut-m’ or with khae-g-a ‘eat- 3sg-fut-m’. However, portmanteau morphs refer to several items of grammatical information at the same time. In contrast to this, Santhali data will not give you a single translatable word, as the whole construction is a predication. In such a case you can collect similar-looking constructions by changing only one component at a time as discussed earlier, and soon you will be able to mark the morphemic boundaries. Eliciting and tagging data is a slow and multi-stage process. Your tagging becomes neater and neater as you proceed to analyze the already elicited words and phrases.

!

It is important to remember that tagging and analyzing should be done the same day you collect the sentences, as any discrepancy or gaps can be confirmed the next day. Do not leave this job for a later date “Oh I will do this when I am back at home”, as some of our students are tempted to do.

5. Morphology and Word Formations 143 Though tagging is shown as a three-line format above most of the time it is a five or six line process which ultimately is reduced to three lines as shown below. The third stage (line) given here is relevant only when there are complex words in a sentence. 1st line ------raw phonetic data 2nd line------word-to-word translation of the data 3rd line------break up of complex words into simple words (sandhi vigrah) 4th line------morph-to-morph break up 5th line------grammatical and other information 6th line-----running translation in contact language

One can illustrate this by an example from Hindi: 1st line--5.55 -- ram-moh«n ko kitabeâ deta hE (elicited sentence) 2nd line----

ram- mohan to books gives is (word-to-word translation)

3 +4 line--ram-mohan-ko kitab-eâ de-ta hE (word and morphemic break) rd

th

5th line----

ram-f mohan-acc book-pl give-3m imperf sg aux 3sg (grammatical information)

6th line----

‘Ram gives the books to Mohan’ or ‘Ram gives Mohan the books.’

It is a convention to use single quotation marks ‘’ for freely translated sentences. There is another format used in the linguistic literature that is equally good. In this, the second line gives the morphemic breaks, because morphophonemic operations change the shape of the words. we will reproduce this format from Chelliah’s work on Meithei (1997:163). kEr«mn«n«

irónb«ge

háyb«du

útlu

k«r«mn«-n«

irónba-ke

háy-p«du

út-l«-u

how –adv

ironba-optative say-det comp show-prospective-imp

how want

ironba

that

show

‘Please show me how to cook Ironba’

The attempt should be to record as much information as possible, which can only be ensured if a large amount of variable data is elicited and analyzed simultaneously (not immediately, but during the same course of data collection). All kinds of grammatical information should be elicited and no phonological form should be left unexplained. 5.1.7.2

Encoding on Verbs/Verbal Agreement

144 A Manual of Linguistic Fieldwork and Structures of Indian Languages Let us see now in how many ways a verb can encode various information about the nouns in the sentence (subject of, object of, complement of, speech act participants etc.). We shall also examine how the temporal, aspectual, and modal situations of the utterance in question, and pragmatic information regarding respect shown to the addressee or addresser of the utterance, are overtly marked on a verb. Verbs give all the relevant information regarding the (a) participants of an event, (b) the location of an event on a linear time scale, (c) the temporal structure of an event, (d) the actuality of an event (realis or irrealis) and (e) evidentiality of an event. To observe this we leave the stage of sentences in the present tense and come to the stage where we can collect simple sentences of several types showing various encodings on the verb. We still do not make complex sentences with several embeddings, but stick to simple constructions involving varying types of nouns, modifiers, tense, mood, aspect and any other speech act phenomena. This brings in the issue of simultaneous analysis, which needs to be undertaken in the linguistic investigation. One can never collect data just for the sake of collection. One has to analyze the data from time to time in order to get a clear picture of what to collect and what not to collect in the next meeting. One should always come prepared for the elicitation meeting, preferably with model sentences written on the notebook. There are several small stages within this fourth stage when data elicitation is geared in a particular direction dictated by the preliminary analyses. For instance, consider the steps given below, which direct us to the ‘split ergativity’ factor in Hindi collected through simple sentences. We shall now employ the model sentence in various tenses to see in how many ways the verb changes. Samples of Verbal Agreement Step 1.

5.56 ram- jata h [intransitive] Ram- go3m sg-impf aux prs-sg. Ram goes

5.57

sita- jati h Sita -go 3f sg impf - aux prs-sg. Sita goes

5.58

sita gH«r g«i Sita home go pst fsg ‘Sita went home’

Conclusion: Subject governs, i.e. verb agrees with the subject noun for gender assignment. Subject is not case marked in present or past tense. Step 2. 5.59

ram- ne ro i khai Ram -agt.- bread-f-f sg. eat pst f sg ‘Ram ate the bread’

[transitive]

5. Morphology and Word Formations 145 5.60

sita - ne- ro i khai Sita -agt- bread-f-f sg- eat pst f sg ‘Sita ate the bread’

Conclusion: Object governs. Object is not case marked, but subject is marked for agentive. Perhaps past tense has something to do with the case marking of the subject. Step 3. 5.61

sita –ne badam khaya Sita -agt- almond-f m sg- eat pst m sg ‘Sita ate the almond’

5.62

sita –ne a÷gur khaye Sita- agt grapes-f m pl. eat pst m pl ‘Sita ate the grapes’

Conclusion: Object governs. Object is still not case marked, but subject is. Past tense does trigger the subject (agent noun) to be case marked Step 4 5.63

ram ro i Ram bread-f f sg

khayega eat fut m sg

‘Ram will eat the bread’

5.64

sita ro i Sita - bread -f f sg-

khayegi eat fut f sg

‘Sita will eat the bread’

5.65

ve- l }ke- a÷gur- khayenge those-boys grapes-f m pl- eat fut m pl ‘Those boys will eat the grapes’

5.66

ve l }kIyaâ aNgur kHayeâgi those girls grapes-f m pl- eat fut. f pl ‘Those girls will eat the grapes’

Conclusion:

Subject governs. Neither subject nor object is case marked in future.

146 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Step 5 5.67

l«}ki ne girl-erg

billi ko cat-fem acc

dekha see 3msg

‘the girl saw the cat’

Conclusion: If subject and object both are case marked then the verb takes a default agreement of 3msg. Overall conclusions: (1) Unmarked object governs only in the past tense. A case of split ergativity. If the object is case-marked in the past tense then the default agreement applies, i.e. 3msg. (2) Absolutive-ergative distinction is maintained in the past tense only if the object noun is non-animate but in future and present the distinction is nominative vs. accusative.

Here we would like to draw the attention of our readers to the Magahi example considered in 4.3.4, where the verb inflection shows the addressee component as well. This significant information can only be deducted if a morpheme-to-morpheme translation is made and a large number of varying forms can be compared with each other. As we witnessed above in various steps, data elicitation is accompanied by marking grammatical information. Not only the grammatical categorisations such as verbs and nouns, but various inflectional units should be specified at all stages, too. Untagged or non-specified data is good for nothing.

! By the time the field investigator reaches this stage, s/he is equipped with the basic word ordering in the language, some morphology and limited lexicon by which s/he can generate sentences in the target language and can let the informant correct them. The elicitation techniques discussed in sections 4.3.3, 4.3.4 and 4.3.5 are especially useful. One should try to converse in simple sentences with the local people by the time one reaches this stage. There are two definite advantages to this suggestion. Firstly, the informant feels very proud of her or his labor on hearing you speak her/his language and draws immense satisfaction from the whole activity of elicitation. Secondly, this is the best way of getting the data corrected grammatically as well as semantically. Stage four is also useful in acquiring many inflectional paradigms, as now one can collect a large number of forms with the help of substituting subject and object nouns accompanied by verbs in varying tenses, aspects and moods (TAM). Let us consider this category in some detail. 5.1.7.2.1

Tense, Aspect and Mood (TAM)

Tense indicates the temporal location of an event or a state. In other words, the category tense indicates the time of the predication in relation to some particular moment. This moment is typically the moment of speaking or writing the utterance.

5. Morphology and Word Formations 147 Tense thus indicates whether the event happened prior to the moment of speaking (past tense), is contemporaneous with it (present tense) or subsequent to it (future tense). Languages may make a distinction between past: non-past, or future: nonfuture, i.e. real: unreal (realis:unrealis). While Indo-Aryan languages generally make a distinction between past, future and present, the Tibeto-Burman languages make a two-way distinction between past and non-past. In these languages tense marking is interlinked with evidentiality phenomena discussed a little later in the chapter.

! Make sentences with temporal adverbs such as ‘yesterday’, ‘today’, and ‘tomorrow’ and observe whether this brings about any change in the verb form. If not, you know the language does not overtly mark tense. If yes, observe what kind of change has occurred. If it is on the auxiliary and not on the main verb, maybe the language marks different aspect categories and not different tense categories. If the main verb form changes, then you can be sure that the language under consideration marks tenses distinctively on verbs. You will have to undertake many substitution exercises to reach a conclusive judgement. Do not forget to expect varying forms in male speech and female speech if the society offers such a distinction, as in the case of Kurux. Refer to the section on ‘male and female speech’ in chapter 3. The category of aspect goes hand in hand with tense. It highlights the internal temporal unfolding of the predication. Essentially aspect indicates whether an event, state, process or action that is denoted by a verb is completed or in progress. Consider the following Hindi examples to get an idea how the two categories combine in a verb complex. This will also give you an idea how Indian languages use series of verbs to indicate tense, aspect and also mood. For instance, sentence 5.68 and 5.69 have three verb combinations in the main clause in finite position, and sentence 5.70 has two verbs in the finite position. 5.68

j«b t«k mE v«ha p« hUca vo khana kha1 cuka2 tha3 when till I there reached-msg he food eat completive aux pst 3msg. ‘By the time I reached there he had eaten the meals’ (past perfective/completive)

5.69

jab mE v«ha p« hUca

vo khana kha1 r«ha2 tha3

when I there reached-msg he food eat prog. aux. pst.3msg. ‘When I reached there he was eating his meals’. (past progressive/imperfective) 5.70

y«h kitab pa}H1 jao2 this book read GO-imp ‘Read this book and (finish it)’. (future perfective)

Now is the time to get sentences with compound verbs, or those which employ experiential verbs. The former primarily show perfective aspect in Indian languages and the latter exhibit varying agreement patterns on the verb of the sentence. Agreement has been shown by drawing a line between the controller and the target. An example each from Hindi should demonstrate the issue.

148 A Manual of Linguistic Fieldwork and Structures of Indian Languages Hindi: 5.71

ram-ne moh«n ko

kitabeâ

de

diâ

ram-agt mohan-acc book-fpl give-abs GAVE-pst fpl ‘ Ram gave the books to Mohan (perfective)’ 5.72

sita ko

ê«r

Sita-dat fear-m

l«ga feel-pst 3msg

‘Sita felt scared’ Please refer to the specific questionnaire on explicator compound verbs (ECV) given in the appendices. A verb such as GO indicates perfective aspect in all Indian languages, specifically.

!

Do not count upon your informant to give you a morph-to-morph gloss, or the information why there is a dative marker attached to Sita in the sentence above. Informants can be trusted only to give glosses for open classes of words such as nouns, adjectives, verbs, adverbs etc. The words belonging to the closed class and those words that have been grammaticalised in the language such as the verb ‘give’ in the sentence above cannot be explained by the informant. One should also resist from asking simple grammatical information such as ‘what is the plural of X?’ It is advisable to find out what grammatical paradigms the words fall into by making an appropriate phrase or a sentence. Speakers of Munda languages anyway find it very hard to segregate even the open class of words from the utterance they have rendered. One has to be careful what language and from which language family one is eliciting data. As we saw in chapter two, Bhat (2000) maintains that Indo-Aryan languages are aspect prominent, while Dravidian languages are tense prominent and Tibeto Burman languages are mood prominent. Although over the years grammarians have been writing grammars of these languages as if all languages were alike in forming tense, aspect and mood, this is not so in reality. The English sentence I am going to Patna tomorrow with present continuous tense but indicating future meaning can be rendered in all Indo-Aryan languages employing continuous tense marking with the same meaning. Tibeto-Burman languages and Dravidian show the distinction between past and non-past. Hence a field investigator should be prepared to find the same form of a verb carrying semantic functions of present and future tense. How then can you distinguish the two? The simplest technique is to juxtapose an adverb of time such as ‘today’ and ‘tomorrow’ with the model sentence to get the desired result. The following excerpt from Bhat’s study of Kannada is worth mentioning here: Kannada has two-fold deictic and three-fold non-deictic distinction in its use of tense markers. Tense markers are also more pervasive than aspect and mood markers in Kannada. They occur obligatorily in finite forms of verbs, and are retained to the greatest degree in the various non-finite forms of verbs as well. (Bhat 2000:110)

5. Morphology and Word Formations 149 Tibeto-Burman languages such as Manipuri, Naga, Mizo etc. make a distinction between realis vs. irrealis in their verb system. If the speaker is sure about the occurrence of the event, it is represented in realis aspect despite the fact that the event under consideration is yet to take place. If the speaker is not sure of the occurrence of the event despite the fact that the event under consideration has taken place it will be rendered in the irrealis aspect. Generally a suffix attached to the verb concerned indicates the two aspectual notions. One must remember to use all the pronouns to elicit data in this section, as each type of pronoun will govern a distinct agreement pattern on the verb morphology. To begin with, the three tenses, past, present, and future as well as three aspects such as perfective, continuous, and habitual should be looked for. Later, as the language dictates, many more divisions can be made. Consult the appropriate appendix for a detailed questionnaire. A blank table for possible verbal morphological structures as given below might help the field investigator to frame an appropriate questionnaire. The table is only a guide to frame sentences. You may not get such fine distinctions in each language as discussed above. Table: 5.11 Pronoun/Noun

Possible Different Aspectual Forms Habitual/ Simple

Continuous

Perfective

11 singular masc. 1 singular fem. 1 plural (inclusive) 1 plural (exclusive) 2 singular masc. 2 singular fem. 2 plural m + f 3 singular masc. 3 singular fem. 3 singular neuter 3 plural m + f

Similar tables can be made for the various tense markings that a language allows. Another category that combines well with verbal morphology is mood. The function of mood is to describe an event in terms of whether it is necessary, possible, permissible, desirable etc. The following English sentences are distinguished on the basis of four distinct moods.

150 A Manual of Linguistic Fieldwork and Structures of Indian Languages 1.

You must go (necessity)

2.

You may go

(permission)

3.

You can go

(possibility)

4.

You ought to go (desirability)

Consider a sample sentence designating mood, negation and various subject pronouns in Hindi as well as verb agreement which shows on the modal and the following auxiliary, but not on the main verb ga ‘sing’. The first model sentence is glossed and the subject pronoun which governs the agreement is given on the extreme left-hand side of the sentence concerned. 5.73 sing ‘

1 msg m

gana (n hi ) ga s kta hu

‘I

can

(not)

I song (neg) sing can-msg aux-1sg 5.74

1 fsg

me gana (n hi)

ga s kti hu

5.75

1 mpl h m gana (n hi ) ga s kte h â

5.76

1 f pl h m gana (n hi ) ga s kti h â

5.77

2 msg/pl

tUm (s

5.78

2 fsg/pl

tUm (s b) gana (nahi ) ga s kti ho

5.79

3msg vo gana (n hi ) ga s kta h

5.80

3fsg

5.81

3mpl ve gana (n hi ) ga s kte hEâ

5.82

3fpl

b) gana (n hi ) ga s kte ho

vo gana (n hi ) ga s kti h

ve gana (n hi ) ga s kti

hEâ

Mood can also be ‘potential’, ‘epistemic’ or ‘deontic’, ‘evidential’, ‘debitive’, and ‘subjunctive’. Most of the mood categories in Indian languages are indicated by explicator compound verbs where V2 loses its lexical meaning and denotes modality. The Meitei example given below (Chelliah 1997) indicates that -t«w (derived from the verb t«w ‘do’) does double duty as an epistemic and deontic marker. The morpheme -t«w here indicates a strong possibility that an action will take place. 5.83

m« ha k

c« tt«wre

m«-ha k

c«t

-t«w

3P-here

go

oblg. -perf. -assertive

‘He must have gone.’

-l«

-e

5. Morphology and Word Formations 151 Chelliah (1997: 242) reports that the inflectional morphology of verbs in Meitei consists of eight illocutionary mood markers: the non-hypothetical –i; the assertive –e; the optative –ke; the imperative –u; the prohibitive –nu; the solicitive – o; the supplicative –si, and the permissive -s«nu. Closely related to tense, aspect and mood is the phenomenon of speech act participants and the belief of the speaker of the utterance. Such belief is transmitted via verb morphology. Let us consider it in some detail. 5.1.7.2.2

Speech Act Participants

Indian languages mark their verbs for various sociolinguistic and pragmatic functions such as speech act participants (SAP). Consider below the information about the SAP relating to ‘indexicality’, ‘evidentiality’ and ‘mirativity’ that are encoded in the verb of Tibeto-Burman languages and other languages of the Himalayan region. John Peterson (1998) reports that modern Nepali has following forms for denoting evidentiality. 1. The morpheme cha 3sg. of hunu ‘be’, unmarked with respect to evidentiality 2. The morpheme hola:, the future of hunu ‘be’ is used to denote inference through reasoning or general world-knowledge. 3. The mirative suffix –e-cha consists of the –e pariciple + cha. 4. The form rahecha, which is the inferential/mirative form of rahanu ‘be, remain’. It is used to denote inference through results (as opposed to reasoning) and as the mirative/inferential form of cha. 5. The form re-, which is reportive, seems to derive from rahecha, the mirative/inferential form of rahanu ‘remain, be’. This form denotes second or third–hand knowledge Delancey (1997) also reports mirativity markers in Dhimal. Consider: 5.84

ka- te jim-gha-rei I- P sleep-pst - 1- MIR ‘I have just slept’ (I discover)’

5.85

waraN - te si-hi-rei old man- P die-pst-3-MIR The old man died (just discovered)

Roland Bielmeier (1997) also brings in the distinction between the linking verb in ‘to be’, and ‘existential verb’ jot by quoting from A.F.C. Read’s ‘Balti Grammar’ (1934). The two examples given below from Balti suggest that the speaker

152 A Manual of Linguistic Fieldwork and Structures of Indian Languages would chose between the two constructions depending upon her/his personal experience. 5.86

di smanpo ljaxmo in ‘The medicine is good’

(implies that ‘the speaker suggests that the ingredients and the quality of the medicine are good, but that its value as a curative has not been proved’). 5.87

di smanpo ljaxmo jot ‘The medicine is good’

(Implies that the speaker has knowledge of the thing in question. This points to the healing property and suggests personal experience, i.e. ‘I have used this medicine and it is good’.). It must be obvious by now that eliciting data on this aspect is a daunting task. It is not the sentence complexity that is challenging you, as there is none to be seen here, it is how different individuals perceive the event that has to be grasped. My linguist friends who work on these types of languages tell me that living in the field and achieving near native-like competence in the language alone helps them. Some had the advantage of a linguist as a native speaker who could explain the nuances. However, not every one is this lucky. 5.1.8 Stage V: Complex Sentences This stage is marked by collecting a large number of long and complex sentences as now one can handle embeddings and raising phenomena, if any exist in the target language. There are two ways of collecting complex sentences. Either one gives a model sentence in the contact language for the informant to translate or one asks the informant to narrate a mythological story or some common incident known both to the speaker and to the fieldworker. The conditionals, the relative constructions, long coordinated sentences, embeddings dealing with double causatives, and anaphora, are the kinds of constructions one attempts to elicit. The task is not very difficult as one builds up on the elicited material from stage four. Once again it is advisable to begin the interrogation by framing the sentence in the target language and then waiting for the informant to correct it. Later, the informant may be requested to narrate a short story. If permissible, one is advised to tape-record the narration. 5.1.8.1 Emphatic Particles/Clitics Clitics show a low degree of selectivity in that the category of the host may be lexical or phrasal. Clitics may exhibit different phonological behavior from other affixes in the language. For instance, Chelliah (1997: 248) mentions that in Meithei most affixes are toneless, but clitics are lexically marked for tone. The language offers six categories of enclitics that exhibit (1) distant/proximate determiner, (2) case roles, (3) copula, (4) mood, (5) pragmatic (emphatic) choice, (6) evidential and attitudinal information. Santhali uses a focus clitic [d!]. At present we briefly discuss those complex sentences that involve emphatic particles and clitics widely used in Indian languages. We will draw examples from

5. Morphology and Word Formations 153 Hindi, but this will be equally true of other Indian languages as the phenomenon of using emphatic particles is an areal feature. There are three emphatic particles (EP) normally found in Hindi: (1) adversative such as Hindi to (2) inclusive such as Hindi bHi ‘also’ and (3) exclusive such as Hindi hi ‘only’. All three have a common property of scope marking. In Hindi, it is the preceding grammatical category that takes the EP under its scope. All emphatic particles, in addition to serving the basic function of emphasizing an element, presuppose a proposition which may or may not get represented on the surface. Adversative particles such as Hindi to are employed in negating the related event in the next clause. It is also used as a focus particle to focus the word or clause it immediately follows, and thus is very mobile in the language it is used. The use of the –to particle presupposes that the interlocutors involved in the discourse share the knowledge of the referent specified by –to. The specified element can either be a word or a phrase and generally precedes the occurrence of –to. Sentences with –to cannot be discourse openers4 because of the constraint of ‘shared knowledge’. Consider the Hindi sentences given below: 5.88

usko jana

to tHa

p«r vo n«hiâ g«ya

he-dat go-inf EP aux-pst but he neg go pst 3msg ‘ He was supposed to go but he did not’ 5.89

kHana to h«mne kha

liya

p«r peÿ

n«hiâ bH«ra

food EP we-erg eat TOOK but stomach neg fill-pst 3msg Yes, we have eaten, but our stomach is still craving for more’ For a detailed study of this aspect refer to Kidwai (2000: 41-8) and Nair (1991/1992: 19-28).

!

It is difficult to elicit these constructions from a framed sentence. You should wait till you come across such structures. You will soon note that these particles are very mobile and thus once you locate them you can freely modify your sentence with changed order and get different focussed elements in the sentence. For instance, moving -to from its original position to a post-subject position in sentence 2 above moves the focus from ‘food’ to the subject NP ‘we’ in the sentence, meaning thereby that ‘whether any one else ate or not but we have eaten’. The inclusive emphatic operates as a conjunctive coordination. The clause or NP that it is attached to it carries a presupposition of not being included initially. Consider the Hindi sentence: 5.90

4

v«haâ b«cce bHi aye there children EP come pst 3mpl

However, one should not confuse this –to with the interjection –to ‘then’ used for encouraging conversation and getting information in normal discourse.

154 A Manual of Linguistic Fieldwork and Structures of Indian Languages ‘Children even came there’ (presupposes that the children were not expected). Like –to, this EP is mobile and can be preposed to b«cce just after v«haâ, bringing v«haâ under its scope of focus, with a change in meaning ‘children came there too’. The exclusive emphatic particle operates as a partitive. The relationship between the whole and the partitive that we saw earlier in the context of quantifiers is operational in this context. Consider the Hindi sentence given below: 5.91

g«dHe hi t«mbaku n«hiâ kHate donkeys EP tobacco neg eat-prs 3mpl ‘Only donkeys don’t eat tobacco’ or ‘donkeys alone do not eat tobacco’

Emeneau (1974), in his study of areal features lists Sanskrit api and Dravidian um with common sets of semantic and grammatical properties. These are emphatic particles that are present in most of the Indian languages today. Also refer to other particles listed below under derivation.

! There is no better way of eliciting material on emphatic particles than to give a sentence in the contact language for the informant to translate. As your contact language would be Hindi or any other Indian language, such a phenomenon can be shown without much problem. Later on, when you collect data for all three types, you may frame a sentence that includes all three at the same time andtest its acceptability by the informant. A sentence such as: 5.92

ap to jante

hi hEâ

ki meâ bHI sh«rab pita huâ

you EP know-2msg EP aux-hon comp I EP wine drink-prs impf aux 1msg ‘Only you know (already) that I also drink wine’ or ‘ As you know that I also drink wine’ Interestingly, these emphatic particles have become prevalent enough in Indian languages that they have made inroads into Indian English. The use of ‘only’ in Indian English has been of particular interest to linguists lately. When Indians use ‘only’ in their English, it replaces all three of the emphatic particles that we discussed above. 5.2

Derivation

Now we come to the second process of word formation, i.e. derivation. The line between inflection and derivation is very thin, in fact it is a little blurred. Traditionally, it is maintained that derivation is a morphological process that helps languages change one grammatical category into another. Inflection, on the other hand, is a morphological process that does not change the grammatical category but defines the characteristics of a word class or indicates the grammatical relationship between words. For the convenience of the field investigator we shall

5. Morphology and Word Formations 155 reproduce the summary table (Aikhenvald 2000:30) that explains the salient features of identifying derivation and inflection. Table 5.12

Inflection and Derivation

Inflection

Derivation

1. Usually obligatory

Optional

2. Final process (if affix, on rim of word)

Pre-final process (if affix, between root and inflection)

3. Forms a complete word

Derives a stem which takes inflections

4. Defining characteristic of a word class (e.g. nouns inflect for case)

Usually specific to a word class

5. Does not change word class

May derive a stem of a different word class, or may add some semantic specification to a root without changing class

6.May indicate grammatical relationship between words, and/or participate in agreement

Never indicates grammatical relationship between words or participates in agreement

7. Tend to be smallish systems

May be large systems

8. Tend to have high frequency in language

Likely to have lower frequency

9. Tend to be monosyllabic. Likely to undergo phonological processes when combined with stem (such as assimilation or fusion)

May be longer and are less likely to undergo phonological processes

But, as we have seen in the context of Tibeto-Burman and Munda languages, such a clear-cut division between derivation and inflection does not exist. However, this table is reproduced here so as to familiarise the investigator with the muchrecognised difference between the two processes, and it might be useful to know when working on the other two language families of India, i.e. Indo-Aryan and Dravidian. Languages do not always keep the distinction in various word classes. It is observed that a language may use the same form either as a verb or as an adjective or as a noun. It is the word order of the grammatical categories and the question of which two grammatical categories follow each other in a sequence that determine the various readings. Thus in Kurux 5.93

roj pairi

indram luêlam ra/I

daily morning something write be ‘One should write something every morning’

156 A Manual of Linguistic Fieldwork and Structures of Indian Languages

5.94

aas hi xaxata luêlam dao ra/I he-gen hand writing good be ‘His hand writing is good’

Despite the confusion and uncertainty we will proceed with our investigation, as some languages may show the distinctive derivational unit in a clearer fashion. We must submit that it is not easy to elicit information under this category, as what constitutes a derived word in your language or the contact language might not be so derived in the target language. For instance, if the English word ‘honest’ can be derived into an abstract noun as ‘honesty’, the word “honesty” in the target language need not be derived in this manner. Our field experience shows that it is a daunting task to elicit abstract derived nominals. Languages that do not mark the boundaries of inflection and derivation such as Naga and other Tibeto-Burman languages employ various different affixes each ear-marked for a specific grammatical category. Therefore, to determine whether a word is a noun, an adjective or a verb, one has to judge it by the appropriate category affix. For instance, in Manipuri the basic morpheme unit (let us call it a root) k«n can take prefixes and suffixes to derive (or inflect?) the following forms: Basic Root

k«n

‘potential meaning of valour, strong’

k«n-n«

‘strong’ (adjective)

k«n-b«

‘fast’ (adverb)

k«n- li

‘be strong’ (verb)

«-k«n-b«

‘related to strength’ (abstract noun)

ma-k«n-li

‘ strong 3msg’ ‘he is strong’

Indo-Aryan languages as well as Munda languages, however, provide ample sources for derivation. In view of the highly ambivalent word classes in Kharia, a number of productive derivational and reduplicative (discussed later in the chapter) processes are employed to derive nouns from verb bases. Let us consider few cases from this language5. One popular process is by infixation of a nasal syllable (a combination of a nasal and a vowel). Consider the following examples and do not miss the vowel harmony as the base vowel is copied in forming the infix. Table 5.13

5

Derivation of Nouns from Verbs in Kharia

Verb

Gloss

Noun

jib

‘touch’

ji-ni-b

job

‘suck’

jo-no-b

juN

‘ask’

ju-nu-N

The Kharia data used in this chapter is taken from Malhotra 1982.

5. Morphology and Word Formations 157 kol

‘count’

ko-no-l

êeb

‘climb’

de-ne-b

êel

‘come’

de-ne-l

jo/

‘sweep’

jo-no-/

Agentive nouns in many languages are formed by suffixing a morpheme indicating the doer of an action. Kharia makes use of the suffix -k«} attached to a verb root to derive agentive nouns. Consider: ebo/

‘play’

ebo?-k«}

‘player’

aloN

‘sing’

aloN-k«}

‘singer’

But if the root is a monosyllable it first reduplicates the basic verb form. ao

‘live’

ao ao k«}

‘inhabitant’

bay

‘build’

bay bay k«}

‘builder’

ter

‘give’

ter ter k«}

‘benefactor’

The agentive morpheme in Hindi and in many of the Indo-Aryan languages is -wala and its cognates, which are considered later in the chapter.

!

Always remember that a very productive way of forming an agentive noun in a large number of languages is by suffixing the word for ‘man’ or ‘woman’ or ‘boy’ or ‘girl’ to the verb base. This process is closer to compounding than derivation. See the section on compounds given below. 5.2.1 The Particle -wala as a Derivational Unit Modern Indo-Aryan languages have developed a particle-like structure suffixed to any grammatical category of an open word class with a semantic function of restrictive relatives. This particle in Hindi is morphologised as -wala, and in other related languages one finds cognates, such as Bangla -uala, Punjabi -vala. The high frequency of occurrence of this particle makes it necessary for us to briefly mention here. Adjective + wala: Hindi lal wala ‘the red one’, cHoÿa wala ‘the small one’, gol wala ‘the round one’, miÿha wala ‘the sweet one’ etc. The characteristic feature is that the whole adjectival phrase with -wala can modify the noun in the attributive position, e.g. lal wala frock ‘ the red frock’. Noun + wala: When this particle is attached to nouns it functions as an agent marker. Thus k«p}e wala in Hindi would mean ‘ the cloth merchant’, pH«l wala ‘the fruit seller’, kam wala ‘ the worker’, and kitab wala ‘the book seller’.

158 A Manual of Linguistic Fieldwork and Structures of Indian Languages On the same paradigm, gH«r wala means ‘husband, the one who runs the house’, or gH«r wali would mean ‘wife’. Adverb + wala: Adverbs of time, space, location and manner can be followed by the -wala particle carrying the semantic function of ‘the one which is x’, i.e. up«r wala ‘the one which is up’, or aage wala ‘the one which is in front’, or j«ldi wala ‘the one who is in a hurry’ in Hindi. Verb + wala: When -wala is attached to any verb it can transform the verb into a modified noun and thus we get a nominalised structure. This can be used in an attributive position. A second option is that -wala can be followed by an auxiliary giving it the status of the finite verb. In either case, the verb it attaches to is always in the infinitival form. Some Hindi examples will clarify the statement. 5.95

aa---n-e wali ga}I come-inf-obl pcl-fsg train ‘the one which comes’ in

5.96

[p«ÿna se aa-n-e wali ga}i]NP leÿ hE [Patna from come-inf-obl pcl fsg train-fsg] late aux-3sg ‘The train that comes from Patna is late’

5.97

ga}I p«ÿna se [aa-n-e wali hE]VP train-fs Patna from [come-inf-obl pcl aux-3sg] ‘The train is about to come from Patna’

When a construction as in 5.97 constitutes part of the VP it functions both as a future tense marker and a probable action. Similarly the following Hindi sentence would refer to the immediate future 5.98

vo jane wala hE he go-inf-obl pcl-msg aux sg ‘he is about to go/leave’

!

The particle -wala has the potential to be inflected for GNP and case marked. Elicitation for this particle can be made only by the translation method. Take a common Hindi sentence with any word of open category and attach –wala to it. Ask your informant to translate it into her/his language. You will get the desired result. It is a very productive word formation strategy in most of the IA languages. 5.2.2 Morphological Causatives

5. Morphology and Word Formations 159 A large number of Indian languages derive transitive and causative forms of verbs by derivational affixes. “Indian languages typically have morphologically marked causative verb stems” (Masica 1976, author’s emphasis). Indo-Aryan and Dravidian languages form the causative with suffixes, while languages of the Tibeto-Burman and the Munda family take recourse to prefixes and infixes more often than suffixes. Language contact situations have often produced a conflicting stage where languages use both prefixes and suffixes at the same time within the same lexical unit (see the Kharia example below). Whatever the strategies followed, a field linguist has to remember that a causative affix increases the valency of the verb under consideration, and thus may be considered a valencyincreasing affix. Thus the verb ‘to sing’ has one obligatory argument, i.e. ‘the singer’, but the causative ‘made X to sing’ has two arguments ‘the singer’ and ‘the one who made the singer sing’. Prototypically then, an intransitive verb has the potentiality of becoming transitive and then causative, structures with two and three arguments, respectively. Both of these arguments can be added to a verb by a process of derivational affixation applied to that verb. Thus Hindi so-na ‘to sleep’ can be rendered transitive as su-la-na ‘to make someone sleep’ which itself can be further rendered as causative su-l-wa-na ‘to make someone sleep through someone’. In the South Asian literature the first step of derivation has been called by names like ‘transitiviser’ and the second step of derivation by ‘causative’. If the basic verb is transitive, such as Hindi khana ‘to eat’ with further derivatives kHil-aa-na and kHil-wa-na , the paradigm (for Hindi) is known as transitive, first causative and second causative (Kellogg 1875, Greaves 1933, Guru 1965, A. Sharma 1958, Fairbanks and Mishra 1966, Kachru 1966, among others). There are others who call the first step of derivation ‘simple causative’ or ‘direct causative’ and the second step causative as ‘derived causative’ or ‘indirect causative’. ‘Contactive’ and ‘distant’ causative (or mediative causative) are further names for the same phenomena. Rather than getting lost in the jungle of terminology and names, what one should remember is that both intransitive and transitive verbs can take valency-increasing affixes. The intransitive verbs can be derived maximally up to three argument structures and transitive verbs (which already structures with two arguments to begin with) can go maximally up to four argument structure types by adding derivational affixes, as shown in the example from Hindi given below. Each argument is given a subscript number after the case postposition, and the causative derivational suffix is in bold. 5.99

maâ-ne1 becc-e-ko2(e or a?) dudH 3 mother-erg child-obl-acc milk-f msg

pi-la-ya drink-caus-pst 3msg

‘The mother fed milk to the baby’ 5.100 maâ ne1 bacc-e-ko2 aya se3 dudH-f4 pi-l-wa-ya Mother-erg child-obl-acc maid-instr milk-f3msg drink-caus pst 3msg ‘Mother fed milk to the baby by the help of the maid’ ‘Mother made the maid feed milk to the baby’

OR

160 A Manual of Linguistic Fieldwork and Structures of Indian Languages But not all languages make morphologically overt distinctions between direct and indirect causatives. Consider the Konkani examples (foêl! in both cases) given below. 5.101 ars! fuÿ-l! mirror break-pst 3msg ‘the mirror broke’ 5.102 Raman ars! foê-l! Ram-erg mirror break-pst 3msg ‘Ram broke the mirror’ 5.103 syaman Shyam-erg

rama-k«êliyan ars! foê-l! Ram-instr. mirror break-pst 3msg

‘Shyam made Ram break the mirror’ Languages of the Munda family have the prefix ob- and its cognates as a valency-increasing affix, which adds only one argument to the existing argument structure. That means three-predicate transitive verbs and four-predicate transitive verbs are not distinguished by two distinct morphological affixes as we saw in the case of Hindi above (la- and wa-, respectively). Similarly, when one-predicate intransitive verbs (process or state verbs) change into two-predicate transitive verbs, they are derived by an infix –b- in Kharia. Thus bHore

‘full’

> bHo-b-re

‘to fill’

seNHor

‘straight’

> se-b-NHor

‘to straighten’

ului-na

‘boil’

> u-b-lui-na

‘to boil’

koyil-na

‘wither’

> ko-b-yil-na ‘to dry something to wither’

Contact induced changes have brought some redundancies in the grammar of Munda languages such as in the Kharia example (3) given below. 5.104 iø ciÿHI likH-o-iø I letter write-pst-tr-1sg ‘I wrote a letter’ 5.105 iø ciÿHI ob-likH-o-iø (original Kharia paradigm) I letter caus-write-pst-tr.-1sg ‘I got a letter written’ 5.106 iø iø-a/ souêom-naiø-buN ciÿHI ob-likH-ay-o-iø I I-say husband-my-by letter caus-write-caus-pst-1sg

5. Morphology and Word Formations 161 ‘ I got a letter written by my husband’ Sentence 5.106 shows redundancy as it incorporates both the Kharia causative prefix ob- as well as the IA (Hindi in this particular case) causative suffix – ay.

!

For elicitation one has to employ verbs of action, state, and process, (intransitive in nature) and then change them into respective action-process (transitive verbs with two or three place predicates) and action-process-mediative transitive verbs (with four place predicates). Try to collect natural sentences and remember that not all languages overtly mark direct and indirect causatives distinctively. Please consult the appropriate questionnaire given in the appendices. As said in the beginning of this section on derivation it is not easy to look for derivational categories. But once you hit upon one you may find others similar in type. Derivation of adjectives from nouns, and similarly derivation of nouns from adjectives are not only common in languages, but generally use more than one strategy to do so. The two derivational affixes discussed above can be combined to further derive a nominal, e.g. noun + caus + wala in Indo-Aryan languages. For instance in Hindi one can have:

5.3

kHI-l-ane wala l«}ka feeds’

‘the feeding boy’ or ‘ the boy who

su-l-ane wala gana

‘the song that puts you to sleep’

dHu-l-wane wale k«p}E

‘the clothes for washing’ or ‘the clothes that have to be washed by someone’

Reduplication

Although reduplication is one of the most important processes of forming words, most of the grammar books ignore the phenomenon, considering it rather trivial. We would like to emphasise here that any field investigator working on Indian languages should not consider it trivial at all. If s/he is working on languages of the Austro–Asiatic or Tibeto-Burman family, it is imperative to look for such structures, as the languages of these two families employ reduplication as one of the most productive processes of deriving new words. In addition to this, other languages such as those of the Indo-Aryan and Dravidian families use the phenomenon of reduplication syntactically and for exhibiting a variety of semantic functions. The concept of reduplication is generally undermined in the literature as one is not aware of the variety and multi-functionality of the structure (for details see Abbi 1992 and 1994). We shall in the following pages first describe what we mean by the concept ‘reduplication’ followed by some examples and then suggest ways of elicitation.

162 A Manual of Linguistic Fieldwork and Structures of Indian Languages Definition: Words formed either by duplicating syllables, or by duplicating a single word (phonological word), partially or completely are known as cases of reduplication. Reduplication may thus refer to all of the following: (a) The iteration of syllables which constitute a single word/lexeme, e.g. Hindi dH«k dH«k 'throbbing'; pH«} pH«} 'fluttering of wings or eyes' or (b) The repetition of the entire lexical item, e.g. Hindi gH«r ‘house’ > gH«r gH«r ‘house house’ ‘each and every house' (complete reduplication), or Kharia bor ‘to ask’ > bor bor ‘begging’. It is this kind of reduplication that is multi-functional and has a very large range of semantic functions in all the languages of our interest. (c) A part of a lexical item carrying a semantic modification, e.g. Hindi gH«r> gH«r v«r 'house etc'. (partial reduplication) , or da-dal ‘strike intensively’ from Santhali dal ‘to strike’. We shall consider partial reduplication in little detail in the following section. (d) Reduplicated words intervened by a syllable or a postposition in such a way that they appear as discontinuous, e.g. Hindi j«ldi ‘fast’ > j«ldi se j«ldi ‘the latest’. The reduplicated structure thus derived may or may not operate as a single structural category, meaning thereby that either both or only the second part of the reduplicated word may take any inflectional or derivational affix. However, it always operates as a single lexical category. This brings us to a two-way broad division in the typology of reduplication in South Asian languages: 1. Morphological 2. Lexical Reduplication

Lexical

Morphological

Partial

Echo formations

Complete

Syllable Reduplication

Class changing

Discontinuous

Expressives

Class maintaining

Fig. 5.1 Types of Reduplicated Structures

5.3.1 Morphological Reduplication This refers to the minimally meaningful and segmentally indivisible morphemes which are constituted of iterated syllables. Thus, the base and the iterated part

5. Morphology and Word Formations 163 together constitute a single morpheme, which is also a lexeme. I have used the term ‘expressives’ for such reduplication (Abbi 1992). Let us consider expressives in some detail as one must be able to recognise them when encountered. 5.3.1.1

Expressives

Reduplication of iterated syllables are generally onomatopoeic in nature but need not indicate sound symbolism only. Languages such as Austro-Asiatic are known to derive words by expressive morphology. Thus, Sora mel mel ‘to inspect’, di di ‘to count’, Khasi ra/ ra/ ‘flowers’ are words derived by morphological reduplication. An expressive derives its status of a word/lexeme only after it is duplicated, as the non -reduplicated syllable does not exist as a word. Following features can be associated with expressives. Features: • • •

Semantic

Kinship terminology Five senses of perception (panchendriya) States of mind and manner adverbs

Expressives in Indian languages are used to denote all the five senses of perception, the famous panchendriya, i.e. the senses of smell, sight, touch, hearing, and taste. Thus, gam gam 'aroma' in Maithili, cam cam 'glittering' in Hindi, las las ‘sticky' in Punjabi, khe khe in Meitei, ÿok ÿok ‘laughing sound' in Kurukh and kur kur-a 'crunchy' in Hindi are typical examples of expressives indicating distinct five senses of perception. While working on languages of the northeastern area of TibetoBurman languages and Khasi we realised that the area can be marked as a micro-area for expressives functioning as manner adverbs. We identified 59 expressives, all indicating the manner of walking in Khasi and an equal number in Tangkhul Naga ( Abbi & Victor 1997). Action verbs such as ‘crying’, ‘walking’, ‘running’, ‘laughing’ etc are coded with a high number of expressives. Expressive morphology is widely used to express various states of mind/feeling, perceptions as well as a large variety of 'manner' adverbs. To describe disorder, confusion, untidiness it is common for these languages to use expressive words. Structurally, a CVC form is duplicated once. What is to be noted is that expressives can work as any other word in the language and generally have following morphological and syntactic features. Morphological Features can be used in: • •

Verbal paradigms Noun derivation

They are employed in regular grammatical paradigms of the language and thus form an integral part of the lexicon. In other words, they do not necessarily have morphosyntactic characteristics distinct from the rest of the lexical items in these languages. For instance, many of the imitative expressives (acoustic noises) in Hindi operate as an ordinary verbal category taking the usual affixes of Hindi. Expressives used in various morphological paradigms

164 A Manual of Linguistic Fieldwork and Structures of Indian Languages Root Stem Infinitive

pH«} pH«} pH«} ‘flutter’ pH«} pH«}-ana

Past 3msg 3fsg 3mpl 3fpl

pH«} pH«}-ya pH«} pH«}-yi pH«} pH«}-ye pH«} pH«}-yiâ

Present

Future

pH«} pH«}-ta pH«} pH«}-ti pH«} pH«}-te pH«} pH«}-tiâ

pH«}pH«}-yega pH«} pH«}-yegi pH«} pH«}-yeâge pH«} pH«}-yeâgi

Derivative Noun

pH«} pH«}-a-h«ÿ

Derivative Modifiers

pH«} pH«}-ay-a/i-hu-a/i

‘fluttering’ ‘ fluttered’

Syntactic Characteristics Expressive verbs have following characteristics at the syntactic level in most of the South Asian languages: • As a finite verb • As a conjunct Verb, e.g. Hindi dH«k dH«k k«r (do)/ hona (be) • As a complex predicate in conjunctive participle form, such as in dH«k dH«k k«r ke (expr. + do.+ CP) 'having throbbed’ or ‘throbbingly', which serves as a manner adverb. They are extremely common in Bangla.

!

As expressives are difficult to translate, one has to be observant in the field to acquire them. The difficulty is duplicated by the fact that no two languages form their expressive morphology in a similar way. The best strategy is to start with kinship terms, as most of the languages have some kinship terms expressed in expressive morphology. As known, in Indian languages the five senses of perception such as touch, smell, taste, hear, and see, are designated by expressive morphology. Therefore, the second step is to make sentences employing words of the five senses (panchendriya) and you will be surprised how every language has something or the other to offer you in this field. You may start like this: FI

How would you say ‘clouds are thundering….’?

FI

How would you say ‘her bangles are…..’?

FI

How would you say ‘stars are…..’?

FI

How would you say ‘the dish (local preparation) is very hot/spicy…’?

FI

How would you say ‘a nice aroma from the kitchen is coming’?

FI

How would you say ‘garbage is stinking…’?

FI

How would you say ‘the satin cloth is …’?

While working in the northeast one should first collect one or two samples of expressives indicating ‘manner’ and then build on these by using different verb

5. Morphology and Word Formations 165 forms. Be prepared to be corrected by the informant, as there are strict co-occurrence restrictions between the expressives and the finite verb. At times you might be forced to demonstrate various actions yourself. Yes, one should be prepared to act also in the elicitation process. You will be surprised how soon the informant understands what you are trying to achieve and then s/he on her own will supply various relevant forms. You are advised to take help from a teacher in elicitation data on expressives. 5.3.2 Lexical Reduplication Lexical reduplication refers to a complete or partial repetition of a word/lexeme. Complete lexical reduplication is constituted of two identical (bimodal) words, e.g. Hindi g«r«m g«r«m ‘hot’ or baiÿHe baiÿHe' while sitting’ is derived from the modifier g«r«m ‘hot’ and verb baiÿh- ‘to sit’, respectively. Such derivation can either be of the class-changing or of the class-maintaining type. Munda languages use complete lexical reduplication for deriving words from one grammatical class to another. It is this family of languages that uses reduplication the most and, as seen earlier, the derivational unit is attached only after the word under consideration is reduplicated. Consider the following examples of verbs changing into nouns in Kharia.

su/ si/ kol øo êel

‘stitch’ ‘plough’ ‘count’ ‘eat’ ‘come’

> > > > >

su/ su/ si/ si/ kol kol øo øo êel êel

‘stitching’ ‘ploughing’ ‘counting’ ‘meal’ ‘arrival’

As has been demonstrated earlier in this chapter, agentive nominals are very commonly derived by duplicating verbal morphemes before they compound with a noun indicating ‘man’. Thus: dej ‘cut’ ului dej dej ÿhakur hair cut cut man (a caste name) ‘barber’

!

While working on Munda languages you could try to reduplicate whatever verb type morphemes you may have elicited and most often than not you might find that the reduplicated forms also exist in the language. 5.3.2.1

Complete word (lexical) reduplication

As said earlier is bimodal in nature, i.e. both form and meaning are repeated once. The combined meaning of the reduplicant and the reduplicator, at times is an extension of the meaning inherent in the single lexical entry (i.e. reduplicator) and at times is a contraction. The iconicity is manifested best in reduplicated adverbs and in any grammatical formative which is marked for emphasis. Repetition is a common stylistic device used for emphasis in various languages of the world, including those of the South Asia. Nouns, pronouns, modifiers, quantifiers, question words, relative, correlative, and finite verbs (only in Tibeto-Burman and Austro-Asiatic) can all be

166 A Manual of Linguistic Fieldwork and Structures of Indian Languages reduplicated in a sentence in order to emphasise the content (the central value) of the word in question. What makes South Asian reduplication interesting and unique is the kind of semantic functions it plays in the grammar of the languages of this subcontinent. Among various semantic and syntactic functions the following are the most common that reduplicated structures are embedded with. (1) (2) (3) (4)

Semantic Functions Emphasis Accentuation or attenuation Distributiveness Exclusiveness

Following examples from various languages will help the field investigator identify each semantic type. •

Nouns,

Kurukh: 5.107 mukk«r mukk«r mela era k«rr«r ladies ladies fair see went ‘Only ladies went to see the fair’ (Exclusiveness) •

Adverbs

(manner, temporal and spatial)

Kodagu: 5.108 i m«ddi g«÷ÿe g«÷ÿe buÿÿi ku}i this medicine hour hour after drink ‘Take this medicine every hour’ (distributive) •

Adjectives

Hindi: 5.109 mera kurta pila pila ho g«ya hE my shirt yellow yellow be GO 3msg past ‘My shirt has become yellowish’ [+attenuation] mujHe pile pile kele p«s«nd hEâ I-dat. yellow-mpl yellow-mpl like aux 3mpl. ‘I like yellow bananas’ [+distributive, emphasis] •

Verbs (finite) Rare. Only in TB and AA languages.

Taizang: 5.110 kei curacanpur-a ka pa pa-i I-nom Churachanpur-loc 1sg go go ‘I go to Churachanpur very often’ [+iterative]

5. Morphology and Word Formations 167 •

Quantifiers

Hindi: 5.111 ram ke tin tin gH«r hEâ Ram-dat three three house-m aux 3mpl. ‘Ram has three houses’ [emphasis] •

Question words Hindi: 5.112

k«l seminar meâ k!n k!n aya yesterday seminar in who who come 3msg past ‘Who all came to the seminar yesterday?

!

Following syntactic functions are performed by complete lexical reduplication. Field investigators will have to be careful and observant as they are generally missed for their syntactic functions, and the initial reaction is that they are emphasis indicators. Syntactic Functions (1)

They designate aspect such as simultaneity, continuity, iterativity, nonprecipitation, sequentiality, etc.(in all language families). It must be kept in mind that most of the languages reduplicate verbs and use the form in adverbial position to designate various aspects. An interesting fact about the reduplicated verbal adverbs is that they always co-refer to the subject noun and thus are also used as a disambiguating device. The Hindi example (5.113) given below designates manner as well as continuity, while (5.114) designates iteration, and (5.115) simultaneity: 5.113 vo kH«}e kH«}e bola he stand-pst part stand-pst part speak-pst 3msg ‘He spoke while standing (or in a standing position) 5.114 radha hindi pikc«r dekH dekH k«r ub g«I Radha Hindi pictures see see-cp bore Go-pst 3fsg ‘Radha got bored of watching Hindi movies (again and again)’. 5.115

(2)

baat k«r-te k«r-te uskii aâ:kHoâ-meâ aâ:su a g«ye talk do-prs impf do-pres-impf her eyes-loc tears come GO 3pl pst 'While talking tears came into her eyes'

Reduplicated pronouns indicate reflexives and reciprocals (Dravidian, TibetoBurman, Austro-Asiatic). For instance in Meithei : 5.116 a: isa-n« isa-bu fuj«I I I-agt I-acc hit ‘ I hit myself’

168 A Manual of Linguistic Fieldwork and Structures of Indian Languages

(3)

They form gerundives/absolutives (Munda and Kurux). For instance, em em ‘having taken a bath’ or ekk ekk ‘having walked’ in Kurux.

(4)

They form participial formations modifying nouns (Munda). Kharia ga/ ga/ jHula ‘torn torn shirt’, i.e. ‘the shirt which is torn’.

(5)

Khasi is observed to use reduplication for forming causatives. For instance, hap ‘to fall’ but haphap ‘cause someone to fall’.

!

For eliciting data on complete word reduplication, one has to duplicate already collected words starting from adjectives such as words for ‘hot’, ‘cold’, ‘long’, ‘sweet’ etc. followed by nouns like words for ‘boys’ ‘girls’ or words for edible items such as ‘mangoes’, ‘bananas’ etc. This can be further followed by adverbs of manner as in ‘slow’, ‘fast’ etc. Lastly you collect words in the verbal group and see whether the language concerned duplicates finite verbs or uses verbs as sentential adverbs denoting various aspect functions. For details see Abbi 1991. Refer to the detailed bilingual and biscriptal questionnaire given in the appendices. 5.3.2.2

Partial lexical reduplication

This is constituted of either phonological or semantic partial repetition of a word, e.g. (1) echo constructions such as Hindi khaanaa vaanaa 'food etc'. , or (2) semantically reduplicated compounds like khaanaa piinaa 'eat + drink =standard of living' (considered later under ‘compounds’) or (3) where only the initial syllable is repeated as in Santhali da dal ‘strike intensively'. Lexical reduplications whether complete or partial, unlike morphological reduplications, are not minimally meaningful and thus can be further divided morphologically, as they are formed of two identical words or two non-identical phonological words. From the grammatical point of view, however, they act as a single structural category. More often than not lexical reduplicated structures have distinct morphosyntactic properties that keep them distinct from the other lexical items in the lexicon of the language. Let us see how echo formation is constituted. 5.3.2.2.1

Echo Formations

An echo formation is a concatenation of a base word followed (in rare cases preceded also) by an echo word. An echo word has generally been defined as a partially repeated form of the base word, such that either the initial sound (which can either be a vowel or a consonant) or the initial syllable of the base word is replaced by another sound or another syllable (let us call it a replacer sound or sound sequence). These replacer sounds or replacer sound sequences are unique to each language. Their use does not change the canonical shape of the word in question. The echo word thus derived is placed after the base word (or before as the case may be), the combination then serves as a compound, e.g. gana vana 'song’+EW= ‘song and such thing'. An important feature to remember is that all grammatical categories are echoed.

5. Morphology and Word Formations 169

Three types of systems operate in South Asian languages: (1)

Initial sound replacement by a fixed sound, e.g. IA such as Hindi v-, Punjabi sh-, Bangla ÿ- etc.

(2)

Initial sound replacement by a fixed phonological syllable, e.g. Dravidian languages use gi-.

(3)

Vowel alternation by a qualitatively fixed vowel, e.g. in Tibeto-Burman languages and some of the IA languages such as Punjabi and Bangani (spoken in the Himalayan regions), for instance shakun shukun 'meat etc. in Bangani. In this type, the canonical shape of the base word is copied with a change of vowel in the first syllable, e.g. mar ‘beat’ >mar mur ‘beating etc.’ but tu`l ‘dust’ > tu`l tal ‘dust etc.’ in Punjabi. The vowel alternation strategy is gradually penetrating in languages that never used it before such as Hindi and its dialects.

Echo formations have a wide range of semantic, syntactic and pragmatic functions to perform in Indian languages and not much research has been done on these constructions. Readers may consult Abbi 1992, 1994 and Mohan 1999 for some of the aspects of these constructions. As they are used primarily in spoken form, a field investigator cannot miss them. In brief, they are coded with following semantic and pragmatic features. §

Semantic features: sense echoing, creating a super-ordinate structure, generality.

For instance pen ven ‘pen etc.’ in Hindi may mean ‘stationary items’ or ‘pen or any such writing instrument’ depending upon the context in which the word is uttered. §

Pragmatic features: casualness, non-specificity.

Echo formations are very interesting from the point of view of studying pragmatics. If you are writing a chapter on the pragmatics of the language under investigation, then echo formations, their multifunctional uses and pragmatic constraints should be studied intensively.

! As all Indian languages adopt echo formation processes to derive new words, and as said earlier all parts of speech are echo-formed, one can easily start coining new words by taking the cue from the existing words already collected. Substitute the initial sound or the syllable with the replacer phoneme or syllable as the case may be. It would be difficult to find out the very first time about the replacer phoneme or replacer syllable, but once you have it you can use it very productively again and again and you will be surprised to find out that a large number of lexical items can be echoed by some identifiable sound pattern with a specific meaning. For instance, once you have an echo-word such as Hindi kam vam ‘work etc.’ one can generate a large number of words on this paradigm, e.g. da:l va:l ‘dal etc’, sona vona ‘sleep etc., bura vura ‘bad etc.’. Wherever a specific sound or sound sequence is not permitted as the replacer, the informant will correct you. If you decide to work on echo formations alone, then it is suggested that all the words, i.e. of the CBWL can

170 A Manual of Linguistic Fieldwork and Structures of Indian Languages be written on cards (one word on each card) and then filled with information regarding semantics, pragmatics (which is yet to be worked out for Indian languages), and phonological constraints of each echo formation. The format of each card may look like the following. Fig. 5.2

Sample information card on Echo Formation

Head word

gana vana

Gr. Cat.

Verb or noun, derived from the verb gana ‘to sing’

Meaning

song etc., singing and related activities

Illustration

tumhe

Morph trans.

you dat. some song EW. come aux what

Free trans.

Do you know singing or any such activities (such as playing an instrument or dancing)?

Pragmatic use

informal, non-honorific

kuch gana vana ata hE kya?

Grammatical info. Can be used as a finite verb too, e.g. k l k ub gaya vaya isliye g la k

rab ho g ya

yesterday lot sing-pst EW therefore throat bad be GO-pst3sg ‘Yesterday I sang (for a long time and did similar things), that is why my throat is so hoarse’. It is obvious that such long illustrations cannot be elicited at first instance. One can prepare the cards, and as one progresses to the stage of long and complex sentences one can fill in the rest of the information. 5.3.2.3

Discontinuous Reduplication

Another type of reduplication used by Indian languages is when a postposition or any empty syllable interrupts two duplicated words. Discontinuous reduplication is used to refer to extremities. Hindi k«m ‘a little’ but k«m se k«m ‘minimum’; Punjabi v«d ‘more’ but v«d toâ v«d ‘maximum’. Quantifiers in several languages use this type of reduplication for indefiniteness, e.g. Hindi koi ‘someone’ but koi na koi ‘someone or other’, k«bHi ‘sometime’ and k«bHi na k«bHi ‘sometime or other’ etc. Khasi derives adjectives from verbs by this method. Thus to/ ‘steal’ but to/ sa to/ ‘crafty, malicious, stealthy’. Different infixes between discontinuous reduplicated words would refer to different concepts in Khasi. Thus: sano

‘where’

sano sano

‘anywhere’

sano pa sano ‘somewhere’ sano re sano ‘somewhere or other’

5. Morphology and Word Formations 171

!

To elicit data on discontinuous reduplication you may begin with extremities denoting words such as ‘maximum’ or ‘minimum’ by giving clues in Hindi. For other types as found in abundance in Khasi, one has to be observant, as there is no simple equivalent word in the contact language that one can use for elicitation. Keen observation and alert ears are the keys to successful data collection. The following table will help you identify all types of reduplication. Table 5.14: Areal typology of South Asian Reduplication Types

Morphological Subtypes

Morphological

Semantic Constructs Kinship Terms Perception Words

Expressives

Lexical (Word Reduplication)

Echo Formations

Generality, Non Specificness

Partial Syllable Reduplication Discontinuous Complete

Hedging, Grouping, Intensity Extremities, Non Specificness Aspectual Modification

Simultaneity Continuity Iteration etc.

Intensity Accentuation or Attenuation Distributiveness

[ Source: Abbi, 1994: 41] 5.4

Compounds

When two independent words in a language are joined to create a new word, this is known as compounding. Generally two significant factors are involved in creating compounds; firstly, the words that are juxtaposed to form a compound already occupy an independent status in the language, and secondly, after being coined as a compound the constituent elements of the compound lose their basic meaning and acquire a new reference. The latter statement does not mean that there is semantic bleaching of the constituents but that it is not always true that the construction is semantically transparent. It is the ‘new reference’ that is important here and not the meaning of the individual constituents. This is specifically true if the two constituents of a compound belong to the same grammatical category, i.e. noun + noun, adjective + adjective, verb + verb and so on. Hindi words like dal-cini ‘cinnamon’ < ‘pulses’ + ‘sugar’; gulab-jamun ‘a sweet preparation’ < ‘rose’ + ‘jamun fruit’; lal-pila (hona) ‘(be) angry’ < ‘red’ + ‘yellow’; rona-dHona ‘lamenting’ < ‘cry’ + ‘wash’ are not semantically transparent compounds. Compound-formation is a productive device in Indian languages to coin new words from the existing ones. Let us consider the process of forming compounds in some detail. 5.4.1 Endocentric Compounds Linguists generally try to describe the process of compound formation and identify the head of the compound; i.e. which of the two constituents that are joined together determines the grammatical class of the compound. Thus, steam boat is a kind of a

172 A Manual of Linguistic Fieldwork and Structures of Indian Languages boat and boat ride is a kind of ride. In English nominal compounds it is the rightmost constituent that governs the grammatical status of the compound. However, in verbal compounds such as give up, or put in it is the left-most constituent, i.e. a verb that governs the grammatical status. Compounds where only one constituent operates as a head are known in the literature as ‘endocentric compounds’. One of the constituents of these compounds acts as a modifier. Thus Hindi candni-ra:t ‘moonlit night’ is constituted of the modifier candni ‘moonlit’ that modifies ra:t ‘night’. Similarly Kharia dHu}i-lo/ ‘desert’ is compounded of words for ‘dust’ and ‘land’, i.e. ‘the land that is full of dust’. Compounds formed of a verb plus noun where the noun is the head are also available in Kharia, though they are rare in Hindi: goN-}i-o/ ‘kitchen’ < goN ‘to cook’ + o/ ‘house’. A very productive process under this category is that of deriving agentive nouns. If you recall we saw earlier in the chapter how languages derive agentive nouns by suffixation, such as -k«} in Kharia. The same languages may also employ the process of compounding to coin an agentive noun. Consider especially the semantic transparency that is held between the two lexemes that are juxtaposed. You can always derive these words on your own with the existing vocabulary and test them with the informant for her/his approval. In addition to the -k«} suffix that we saw above, Kharia uses the word for ‘man’ lebu and derives nouns from verbs as well as from nouns. For instance, soNgol ‘firewood’ can be converted into an agentive noun as in soNgol lebu ‘person who carries/sells firewood’. However, when an agentive noun is derived from a verb, it first goes first through the process of reduplication and then the lexeme lebu is attached. Thus: bor

‘beg’ >

bor bor lebu ‘beggar’

go/ ‘carry’

>

go/ go/ lebu/êu ‘man/woman who carries load’

say ‘cut’

>

say say êu

‘harvesting girl’

In such cases compounding is a two-fold process: the verb first goes through reduplication and then the noun word is suffixed. Verb > reduplication > noun > agentive noun Interestingly, Kharia derives a nominal construction from a verb form meaning ‘desire to x’ by suffixing the noun mo/ ‘desire’. Thus koN ‘to know’ and koN-mo/ ‘desire to know’, or yo/-mo/ ‘desire to see’. It might be useful to know that in many languages kinship terminology of feminine gender is coined by affixing a word for a female noun. Thus in Kharia kul«m ‘brother’

kul«m-êay

‘sister’

boker ‘brother-in-law’

boker-êay

‘sister-in-law’

5. Morphology and Word Formations 173 magra ‘man’s name’

magra-êay

‘Mâagra’s wife’

Many Tibeto-Burman languages follow the same pattern. 5.4.2 Excocentric Compounds Conversely, compounds which are formed in such a way that none of the constituents acts as a head are called ‘exocentric compound’. These are known as bahuvriha samasa in Indian grammatical tradition. Among the exocentric compounds we would like to include those compounds that are formed by joining two semantically identical or semantically related words and thus form an equational compound. Thus, Hindi-Urdu dH«n d!l«t ‘wealth’ is derived from Skt. dH«n ‘wealth’ and Persian d!l«t ‘wealth’; Tamil kai-k«ri ‘vegetables’ is derived from kai ‘raw vegetables’ and k«ri ‘cooked vegetables’; Hindi mol ‘value’ and tol ‘measure’ is compounded as mol-tol ‘bargain, evaluation’. Equational compounds can be formed by joining two semantically identical words but drawn from two different sources, e.g. of dH«n d!l«t ‘wealth’ cited above as well as Kharia p«isa-êHebua ‘wealth’ derived from p«isa ‘money (IA)’ + êHebua ‘money (Kharia)’. 5.4.3 Appositional or Associative Compounds There is a third type known as ‘appositional’ or ‘associative’ compound (or dwandwa in the Indian grammatical tradition) formed simply by conjunction of two elements without any dependency relation existing between them, e.g. Hindi rat-din ‘ around the clock’ < ‘night’ + ‘day’ or dal-roÿi ‘existence’ < ‘pulses’ + ‘bread’; utarc«}Hav ‘fluctuations’ compounded from the antonymical words utar ‘fall’ and c«}Hav ‘rise’. The two constituents in associative compounds could either be of polar quality of the same semantic range or incorporate the salient characteristics of the same semantic field. The referential range of such a compound form therefore includes the whole semantic field to which the two constituents belong. Indian languages of all types are rich in these associative compounds. A few more examples from Kharia will clarify the issue. Kharia

!

uslo/- tirib

‘earth-cloud (sky)’

‘universe’

iêib-meya

‘night-morning’

‘the whole day’

ti/-kaÿa

‘hand-leg’

‘body parts’

êa/-pe/

‘water-rice’

‘nutrition’

Now the question is how to recognise compounds. There is no way one can translate from the contact language to get an appropriate answer. The clue to compounding comes from the recognizable vocabulary, the words that you already have in your notebook. When a recognizable word appears to be longer than your phonological pattern allows and the word in question does not allow any intervening affix, you should have a clue that you might have hit upon a compound form. To

174 A Manual of Linguistic Fieldwork and Structures of Indian Languages confirm whether the word in question is a compound or not, ask what the word refers to from the informant. Compounds are not easy to elicit but are fun to work with. It is not necessary that the two constituents of a compound belong to the same grammatical class, however the ones belonging to the same grammatical class are more common than those that are compounded from different grammatical classes. You can think of any two grammatical categories combining to form a compound. In Nicobarese the word /elkE`:l for ‘armpit’ is constituted of a compound formed of El ‘in, on’ and kE`:l ‘arm’, and the word /«tvi`c ‘lobster’ is compounded of two independent words «t ‘not’ and vic ‘be quick’ (Braine 1970). Consult the relevant section on ‘compound verbs’ in the next chapter also. Recall also the compounded postpositions considered earlier in 5.1.5.1. In Indian languages it is not necessary that compounds be formed only of two constituents. Tibeto-Burman languages and Mon Khmer languages such as Khasi form compounds of three and four words also. Thus the word for ‘divorce’ in Khasi is pillaid-sun-shiyaN ‘release-five-shell’ and the word for ‘journalist’ is: u- n!N-

tHo/-

k!t-kubor

msg-person-write-person-news The non-transparency of the constituents of the compounds is at times so strong that one can never guess by hearing the compound what it could refer to. Consider the following compound constituted of V + N + Prp + N + N meaning ‘departed (soul)’ in Khasi. ba:m-kuwai-ha- y«N - ublai eat- betelnut-in-house-God ‘The one who is eating betelnut in the house of God’ ‘departed’

!

It is easy to recognise Khasi nominal compounds, as unlike words used in a sentence that take GNP markers before them, constituents used in compounds take GNP only once before the first noun. Also, since most of the words in Khasi are monosyllabic it is easy to identify a compound (Mayjee Philip 1997). As joining two independent words in the language forms compounds, it is advisable to study the morphophonemics of these words. There are specific syntactic and phonological constraints distinct from those operating on non-compounded forms that need to be investigated. Tone languages such as of the Tibeto-Burman family should be investigated carefully, as they exhibit various tone sandhis in the compounded forms of words. Also consult section 6.7 on Explicator Compound Verbs in the next chapter.

Chapter VI Syntax and Semantics The ultimate aim of any fieldwork is to write a comprehensive grammar. Fieldworkers should be well prepared to handle data for writing a grammar that not only covers syntactic and semantic aspects of the language concerned but also of pragmatics and sociolinguistic aspects as well. The typological studies undertaken in the last thirty years have given us a large number of varied kinds of grammars each pertaining to the characteristic features of the language under observation. It would be worthwhile to consult some of them. In this chapter, we shall focus only on some major aspects of language that help define the ‘Indian’ character of the Indian languages. A few select topics from the areas of syntax, semantics, and pragmatics are discussed in the following sections. These topics are a mere guide to writing a grammatical sketch of the language under investigation and as said may help to identify areal features commonly found in Indian languages. They should in no way be considered exhaustive. 6.1.

Inquiry into Syntax, Semantics, and Pragmatics

Though the collection of spoken narrative text cannot be equalled as an aid to unearthing the language structure, the fieldworkers may find the discussion that follows helpful for framing appropriate questionnaires. We have given samples of a few selected questionnaires in the appendix to help the fieldworker design her or his own. A very detailed format of linguistic questionnaire given by Comrie and Smith (1977), which the readers are advised to consult too, is generally followed around the globe. This has led to a series of grammars written in the last ten years.1 However, we would like to caution the readers that the said questionnaire is not very easy to use and is designed for global use. A fieldworker has to skip several entries for her/his purpose. Readers may benefit by referring to Payne (1997), as he discusses a large number of syntactic and pragmatic topics in detail. The queries he raises at the end of each major section, may serve helpful in identifying specific questions. In addition to the questionnaire method, we expect that the following discussion will also help to identify grammatical structures from a spoken text. We will now give a brief outline of various topics that should and could be covered while working on Indian languages. Detailed questionnaires pertaining to some of these topics are given in the appendices. These questionnaires should be able to help elicit data on morphology as well. The topics and the list that we considered at stage III may also be taken care of by these questionnaires. Ultimately, one must make her or his questionnaire depending upon the nature of the query and emphasis that one wants to give on a particular topic. The topics discussed here are such that one has to depend to some extent on narrative 1

‘Marathi” by Pandharipande (1997), ‘Punjabi’ by Bhatia (1993), “Kannada’ by Sridhar (1990), and ‘Malayalam’ (1997) by Asher & Kumari are a few among the Indian languages.

176 A Manual of Linguistic Fieldwork and Structures of Indian Languages pieces, too. A questionnaire alone would not be sufficient. As usual, helpful tips for fieldworkers are given under the sign !. Major Topics 1. Word order typology 2. Topic and focus 3. Interrogation 4. Negation 5. Complex predicates 6. Explicator compound verbs 7. Dative subjects/oblique marked subjects 8. Complementation 9. Converbs/Conjunctive participles [CP] 10. Anaphora 11. Coordination 12. Kinship terms 13. Terms of address and reference 14. Politeness strategies. 15. Language shift, retention and death These topics and the questions therein are not mutually exclusive of each other. In fact, there is immense overlapping between them which one should be aware of. For instance, while discussing complex predicates, much illustrative material could be brought in form of experiential verbs which are discussed under the ‘dative subjects’. We would therefore like to highlight only those salient features that are important for the particular topic under investigation. 6.2.

Word Order Typology

One has to investigate into the ordering of various constituents of a sentence. Word order typology involves the ordering of elements of noun phrases and verb phrases, including complement clauses. Since almost all the Indian languages (barring Kashmiri and Khasi) are of SOV type, we should frame our questionnaire to incorporate (a) – (d) as suggested below. It should be kept in mind that, while testing various universals proposed by Greenberg (1963), one also has to test the implicational universals that are suggested by the author and those suggested by Lehmann (1973, 1978), and Hawkins (1980, 1983). Also consult Hawkins 1988 and 1990. The basic question ‘what are the pairs of elements whose order correlates that of the verb?’ has been tackled by Dryer (1992). The study is based on the examination of word order properties of 625

6. Syntax and Semantics

languages. These studies will help the investigator to get a wider perspective of the phenomenon and to frame the appropriate questionnaire. To identify the basic word order one has to look for statistically frequent order as well as that used in simple declarative sentences. The test sentences should be stylistically neutral and least marked for regional or social variation. My experience is that elicitation of material for word order typology feeds into various other grammatical topics too, e.g. topic and focus, scope of negation etc. In all, one has to observe primarily: (a) The basic word order in simple sentences (b) The basic word order in complex clauses (c) The order of modifier and modified in noun and verb phrases (d) The morphotactics of affixes in bound morphology 6.2.1. Characteristic Features of SOV Some of the most robust features found in all verb final Indian languages are: 1.

Use of postpositions

2.

Modifiers such as adjectives, demonstratives and numerals precede nominals.

3.

The genitive precedes the governing noun

4.

The indirect object precedes the direct object

5.

The auxiliary verb follows the main verb

6.

Explicators follow the main verbs.

In addition, the common tendencies found in verb final languages of India2 are the following: 7.

Time adverbial precedes place adverbial

8.

Participle phrases with time and place adverbials precede the modified noun

9.

Occurrence of relative-correlative constructions

10.

Conjunctive and disjunctive markers follow respective conjunct or disjunct.

11.

Inclusive, exclusive and emphatic particles follow the element they specify.

12.

Conditional and counterfactual markers are post-verbal.

13.

Adjective follows the standard of comparison.

14.

Dubitative markers are post-verbal.

2

I am thankful to K. V. Subbarao for providing me a large part of this list.

177

178 A Manual of Linguistic Fieldwork and Structures of Indian Languages 15.

Sentential complements may follow the verb while nominal objects rigidly precede it.

16.

Postposed complementisers.

17.

Left Branching.

While testing these orderings one must also test the possibility of scrambling between different constituents so as to identify grammaticality range and semantic variance. At times, scrambling gives good information regarding focus and topic. Word order flexibility is used in most languages to code pragmatic distinctions such as topic/comment, focus/non-focus, definite/indefinite, emphatic/neutral etc. Indian languages in general allow a fair amount of scrambling. Please consult the appendices for the appropriate questionnaire to see the various possibilities. 6.3

Topic and Focus

This has been one of the most debated issues in the theoretical linguistic literature. Chafe (1976:50) defines topic as that constituent of a sentence “that sets a spatial, temporal, or individual framework within which the main predication holds… What the topics appear to do is to limit the applicability of the main predication to a certain restricted domain.” Focus, on the other hand, is a pragmatic aspect of the conversation. It involves the speaker’s intentions and hence is a discourse property. At times suprasegmental features such as stress and intonation exemplify focus. Languages may distinguish between contrastive [e.g. sentence 17 in appendix 7] and non-contrastive focus [e.g. sentence 13 in appendix 7], or restrictive focus [e.g. sentence 14 in appendix 7]. The information on topic and focus is significant for semantic and pragmatic judgement. One can achieve the information on them through word order constituents and interrogative sentences discussed below. The questionnaires given in the appendices on word order and interrogation include sentences that may help eliciting information about topic and focus.

!

Elicitation for word order will also give you information on topic and focus. Narration of all kinds proves to be a great help. Move elements around as well to notice any change in the pattern of stress or any other supra-segmental feature used by the speaker. 6.4

Interrogation

Interrogative is the sentence type used in the expression of questions. The function of questions is to elicit information. Usually questionnaires regarding interrogation will consist of contrasted statements, commands and exclamations in terms of their semantic function. Traditionally, the term interrogative is used to refer to construction types which have the semantic function of questions. Interrogatives can be divided into yes/no questions and wh- questions. This division does not take into account echo

6. Syntax and Semantics

questions. Yes/no questions or in Jespersen’s term ‘nexus-questions’ give an instruction of dysfunction. (Lyons 1977). The other category of questions is called by Jespersen ‘xquestions’, X is an unknown quantity as in an algebraic equation. The grammatical form of x-questions has interrogative pronouns or pronoun-type adverbs in them (Jespersen 1933). Yes/no questions and x-questions can be better distinguished on the basis of the type of variable contained. Yes/no questions contain a two-valued variable while x-questions contain many-valued variables. Most languages have a particle that indicates the yes/no function of the construction. Thus, in Hindi `kyaa’ and in Kurux ‘ender’ functions as yes/no operators. In English, yes/no questions are formed by auxiliary inversion, e.g. ‘Will he come?’ Such a phenomenon is non-existent in Indian languages. Intonation is widely used for the same purpose, even in languages which have syntactic means of expressing the yes/no questions. The ‘x-questions’ have a many-valued variable in it. In most languages, a particular class of words, interrogative pronouns, stand for these variables. In English, most words of this class begin with the letters ‘wh-’. Therefore, ‘x-questions’ are more commonly known as wh-questions. In Hindi, this class of questions words includes k!n (who), ‘kya (‘what’), kEse (‘how’), k«b (‘when’), k«haâ (where) and so on. Malayalam general question-words include aara (who), eyane (‘how’), eppol (‘when’), eviÿe (where) etc. There are two types of interrogative pronouns in Dravidian languages, viz. general and selective (Sridhar 1992: 212-13, Asher et al 1997: 267). Thus the selective set in Malayalam is: ev«n ev«ñ ev«r

‘which male person’ ‘which female person’ ‘which persons’

In Indo-Aryan languages similar strategies are available. Thus in Hindi the morpheme sa is inflected for gender, number and person and then is suffixed to the question word k!n ‘who’, e.g. k!n sa ‘which one [m]’, k!n si ‘which one [f]’. Generative syntax regards interrogative constructions as forming operatorvariable chains which delimit the propositional frame for the set of possible answers. The operators are the above-mentioned wh-words and are considered scope-forming elements. Whether they are assigned values or not is dependent on their scope. An operator with narrow scope is not assigned a value and if there is no wh-phrase with wide scope, the sentence is not interpreted as interrogative, e.g. in the sentence ‘I know who will come” ‘who, though a wh-phrase has narrow scope and the sentence is interpreted as a declarative. However, “who do you know will come” has a structure in which ‘who’ has a wide scope. ‘Who’ could be assigned values in this case and the sentence is an interrogative. Usually the domain which the wh-phrase commands is regarded as its scope domain. Only wh-phrases with matrix scope are necessarily assigned values.

!

To elicit data on interrogation, start with the model affirmative sentence and repeat it with rising intonation. A sentence such as Ram goes home can be rendered as

179

180 A Manual of Linguistic Fieldwork and Structures of Indian Languages Ram goes home? with a rising intonation. You will be surprised to know that this technique will help you identify the scope of interrogation as the informant might answer you in negative saying ‘no he does not go home but to the market’. Use question words from your basic word list and start framing sentences first by putting these words in the beginning of the sentence. That is, begin your sentence with question word. Only gradually move these words from the initial position to other positions within the sentence. Try to elicit all three types mentioned above. You may also use echo-questions such as ‘he ate what?’ to elicit data on wh-questions. Consult the appropriate questionnaire given in the appendices. 6.5

Negation

Negation opposes affirmation. By using negative expression a speaker may deny the hearer’s belief or the scope of a proposition. The dependence of negative expressions on affirmative ones is evident in sentence types in languages where a simple negative particle inserted in the affirmation sentence renders the expression negative. Consider the following Hindi sentences: 6.1

ram k«l gH«r aya Ram yesterday home came Ram came home yesterday ram k«l gH«r nahiâ aya

6.2.

Ram yesterday home neg came Ram did not come home yesterday. 6.5.1

Salient Features

Most of the Indian languages make a distinction between imperative negative [i.e. prohibitive] and non-imperative negative markers. For instance Kharia has a-bu for imperative and um for non-imperative negative. Some languages have different forms but use them interchangeably. For instance, although Hindi offers m«t for imperative negation, it uses the other negative marker n«hiâ interchangeably with m«t, e.g. m«t/n«hiâ jao ‘don’t go’. The questionnaire thus should have ample scope for testing prohibitive and affirmative sentences. There could be a range of negative markers in Indian languages, each reserved for a specific use. They could be broadly summarised as follows: •

Simple negative marker as in Hindi na or n«hiâ.

•

Imperative negative marker as in Hindi m«t.

•

Negating the scope of a proposition such as in Hindi bina [e.g. 6.27b]

•

Conditional negative marker such as in Hindi n«hiâ to.

6. Syntax and Semantics

•

Double negatives (showing discontinuously) appearing in both the matrix and the embedded clause but phonologically distinct, such as in Hindi n«hiâ………na. Such constructions generally have affirmative meaning. Consider:

Hindi 6.3

koii

v«j«h n«hiâ

some reason neg.

ki

kiraya gH«ÿa: na diya: ja:ye

comp rent

reduce neg. give- GO –subjunctive

‘There is no reason for not reducing the rent’

(Hook 1974)

Another interesting aspect of double negatives is the case of Bangani (Abbi 1999). This language has the French type of behavior of discontinuous negative clitics. seu kam na-kçrdç -na he work neg -do-pres-msg neg

6.4.

‘He does not work’ 6.5.

au kitab na- pç}di-na I book neg read-pres-fsg. neg ‘I don’t read books’

We witness three types of double negation in Bangani. 6.6

S---NEG+AUX---Neg.---V ram nçthi-na andE lagiu ‘Ram is not coming’.

6.7

S—NEG+AUX1---NEG-AUX2---V (second aux. marks the subject) ram nçthi nathç andE lagiu ‘Ram was not coming’

6.8

S---Neg.---NEG+AUX1---Aux2---V (second aux. marks the subject) ram na- nçthi- thç andE lagiu (discontinuous neg. and aux.) ‘Ram was not coming’ Sentences (6.7) and (6.8) coexist as parallel structures.

•

Neither ..nor situations where at times two distinct negative markers are used in a coordinated sentence.

181

182 A Manual of Linguistic Fieldwork and Structures of Indian Languages •

Reduced negative markers as a tag such as in Hindi na. Negative tags are an areal universal and are found in all Indian languages.

•

Those languages that make a distinction between realis (present and past) and irrealis (future) in their tense marking system may also make use of distinct negative markers for the two situations. For instance, in Manipuri (Bhat 2000) the negative particle de indicates realis and roy indicates irrealis situations. Consider: 6.9

m«hak hotel he

«du-d« ca-de

hotel that-loc eat – neg.

‘He did /does not eat in this hotel’ 6.10 m«hak hotel «du-d« ca roy he

hotel that –loc eat-neg

‘He will not eat in this hotel’

(Bhat 2000: 151)

6.5.2 Negative Copula There are Indian languages which show a negative copula that inflects for G N P. This phenomenon is available more often in Dravidian and Austro-Asiatic languages than in Indo-Aryan. However, converged languages such as Sadari and Marathi do employ a negative copula. Consider: Sadari 6.11

moeâ ÿic«r nal«goâ I teacher neg 1st sg. “I am not a teacher”

6.12. u mor bap nal«ge he mine father neg. 3rd sg ‘he is not my father’ 6.13. toeâ «fs«r nal«gis you officer neg. 2nd sg. ‘You are not an officer’ Marathi, although employing negative copula, drops some agreement features (Bhat 2000) from these. Consider the form for ‘to be’. Five distinct forms in the affirmative have been reduced to only three forms in the negative.

6. Syntax and Semantics

Table 6.1

Marathi Negative Copula Forms

Affirmative forms

Negative forms

PERSON

singular

plural

singular

plural

I

a:he

a:hot

na:hi:

na:hi:

II

a:hes

a:ha:t

na:hi:s

na:hi:

III

a:he

a:het

na:hi:

na:hi:t

Related to the phenomenon of negative copula is cliticisation of personal deictic markers on the negative particle in Munda. In the following sentences from Kharia, the personal deictic marker advances from the verb suffix position to the preverbal negative suffix position. Consider: Kharia 6.14. (iø)

øo?-t-iø

pe?

(I) food

eat-tr..prs.-1sg

‘I eat food’ 6.15. (iø) (I)

øo?-te

pe?

um-iø

food

neg-1sg eat-tr.pst.

‘I don’t eat food’ 6.16. (ele)

tuëa

co-na-le

(we excl.) tomorrow go-tr.fut.-pl.excl. ‘ We will go tomorrow’ 6.17. (ele)

tuëa

um-le

co-na

(we excl.) tomorrow neg. 1pl.excl. go-intr.fut ‘We will not go tomorrow’ Another interesting aspect about Kharia negatives is their imperative constructions where the prohibitive form a/-bu takes personal deictic markings. One should be prepared to find such constructions in other languages of the same family. Thus in Kharia:

183

184 A Manual of Linguistic Fieldwork and Structures of Indian Languages

6.18

a/ -bu cona

‘you (sg.) don’t go

a/- bar cona ‘you (dual) don’t go’ a/ -pe cona

‘you (pl.) don’t go’

6.5.3 Deletion of some of the grammatical categories This is not an unusual phenomenon. Negative markers occasionally trigger the dropping of an explicator or auxiliary from a sentence. This is especially true for Hindi compound verb constructions and many of the Indo-Aryan languages. Consider for instance: Hindi 6.19

r«him b ÿH g«ya: Rahim sit GO past Rahim sat down (perfective)

6.20 r«him n«hiâ b ÿHa: f

*n«hiâ b ÿH g«ya:

Rahim neg. sit pst 3rd sg.

neg. sit GO pst 3msg.

‘Rahim did not sit down’ Bangla 6.21

radha ba:}i

kore

che

Radha house built Aux pst ‘Radha has built a house’ 6.22. radha ba}i

kore ni f

Radha: house built neg. ‘Radha did not build the house’. Bagri 6.23

m«nn

lag

h

k

bo «b

b«ÿÿH koni jave (h )

me-obl seems Aux comp he now there neg. go-prs (aux) ‘It seems to me that he no longer goes there’.

6. Syntax and Semantics

6.5.4 Scope of Negation Scope of negation is another topic that has attracted the attention of many linguists. It is a popular phenomenon among Indian languages to move the scope of negation. Either the entire proposition may be negated, or a focussed constituent may be brought under the scope of negation, such as sund«r k«p}e in the following sentence: 6.24

usne

sund«r

kap}e nahiâ pahne

He erg. beautiful clothes neg. wear-pl pst ‘He did not wear beautiful clothes’ (but he wore clothes) Bhat (2000) reports of narrow and wide scope negative ambiguity with the use of quantifiers in Kannada. Consider: 6.25

bahaña j«na many

bar-al –illa

people come-purposive-neg.

(a)

‘There were many people who did not come’

(b)

The people who came were not many’

This kind of double interpretation may be possible in other languages too, as the scope of the quantifier may be varied. Thus in Hindi, which is an IA language, in the sentence bahut se log n«hiâ aye ‘many people did not come’, similar double interpretation is possible. Actually, the interpretation of type (b) would entail stress on the negative particle. The scope of negation should definitely be tested with conjunctive participle constructions (or converbs; see 6.10) designating coordination, manner, and cause. Indian languages rich in conjunctive participles shift the scope of negation depending on which constituent it is attached. Consider: 6.26

vo kHana kHa k«r gHumne n«hiâ g«ya he food eat + CP stroll-inf neg go-pst ‘After having a meal he did not go for a stroll’. verb)

(scope of neg. is the finite

6.27.a vo kHana na kHa k«r gHumne g«ya ‘Without having a meal he went for a stroll’ (scope of the neg. is the CP)

185

186 A Manual of Linguistic Fieldwork and Structures of Indian Languages However, the preferred construction is the following, where use of bina precedes the participial construction and thus negates the scope of the proposition inherent in the participial construction. 6.27.b

vo bina kHana kHaye gHumne g«ya he without food eat-part stroll-inf go pst3msg ‘Without having a meal he went for a stroll’.

Scope is an interesting aspect of negation and should certainly be tested. Intonation and stress also mark the scope and thus while eliciting data speech should be tape-recorded. One must take care of mobile negative particles that can be used for constituent negation. 6.5.5 Other related features •

Some interesting features of Indian negation can be seen in compounding, reduplication and other morphological strategies. In most of the Indian languages, negative particles collocate well with emphatic particles such as Hindi to bHi (particle of focus + also) and to hi (particle of focus + even) that perform various pragmatic functions. Discontinuous reduplication of the negative particle gives nonspecific but definite reference, e.g. koi na koi in Hindi would mean ‘someone or other’ and k«di na k«di in Punjabi would mean ‘some time or other’. As far as the derivational morphology is concerned, most of the languages have a process of prefixation to derive negative words. These can be numerous in Hindi because of heavy loans from Sanskrit and Persian. These words with negative markings (known also as antonyms) are difficult to elicit in the field as there never is a readymade translation for these to be inserted in the questionnaire for validation and confirmation. One has to be exposed to the speech community for a long time to elicit these affixes.

•

An interesting feature of Indian languages is that the matrix clause of the verb of ‘thinking’ type is never negated but the following complement clause is. An English sentence such as I don’t think he will come (where denial is on his coming but the negative particle is attached to ‘think’) would be rendered in Hindi, for example, as mera kH«yal h ki vo n«hiâ (Neg.) ayega. Also see sentence from Bagri (6.23) given above. However, under the influence of English, the negative particle has become mobile and most of the Indo-Aryan languages have been using the negative particle with ‘thinking’ type of verbs as an alternative choice. But this is not the preferred choice.

!

To elicit data on negation start with your affirmative sentences of all types and with all kinds of tense and aspect markings. If the negative morpheme in the target language inflects for tense and number it will show up. Also try to move the negative

6. Syntax and Semantics

particle to see the scope of negation. You will have to give several alternative sentences to the informant and ask her/him for different interpretations, if any exist. Imperatives and simple non-imperatives should be tested separately. This is also the time when you may collect some data on antonyms and negative morphology. 6.6

Complex Predicates

It is well known that Indian languages exhibit a complex predicate construction in which the first member is a noun, adjective or a nonfinite verb (in its root form), and the second member is a “light” verb that has lost part of its semantic content. The following examples from Hindi illustrate the constructions. 6.28

ram ne sari mI ai kha li Ram-erg. all sweets eat-take-pst 3fsg ‘Ram ate up all the sweets’

6.29. rahim sari mI ai kha cuka hoga Rahim-erg. all sweets eat- finish -pst, 3msg be-fut 3msg ‘Rahim must have already finished the sweets’ 6.30. ram ko dard hua Ram dat-pain happen- pst 3msg ‘Ram had a pain’ The first example is of an explicator compound verb, the second of a serial verb, and the third is of a conjunct verb used in a dative construction. In 6.28 and 6.29 the light verbs lena ‘to take’ and cukna ‘to finish’ combine with the non-finite verb form of khana ‘to eat’ while in 6.30 the light verb hona ‘to be’ combines with a noun. Frequently found in abundance, these light verbs in Indian languages are translatable as ‘do’, ‘be’, ‘become’, ‘have’, ‘go’, ‘remain’, ‘hit’, ‘come’, ‘put’, ‘give’, ‘be’ and ‘struck’. Some of the properties, which make these constructions both descriptively and theoretically interesting, are the following. a. Case marking as seen in 6.28-6.30. The light verbs influence the selection of case on the arguments of the complex predicate. b. Valency. The valency (the number of arguments) of the complex predicate is determined by both the light and its host verb. For details see the sample of sentences given in the appendix on ‘complex predicates’. For a theoretical discussion on this refer to Verma et. al (1990) and Abbi (1994). Constructions involving explicator compound verbs, conjunct verbs, experiential constructions, and non-agentive subject are considered complex predicate

187

188 A Manual of Linguistic Fieldwork and Structures of Indian Languages constructions. For details see the sample questionnaires under each topic given in the appendices. We will discuss each of the topics separately in the following sections. 6.7

Explicator Compound Verbs (ECV)

The explicator compound verb construction has been the focus of attention since its recognition as a major areal feature of South Asian languages. An ECV construction refers to a sequence of two verbs V1 and V2, in which the main verb of the sentence, generally V1 in SOV languages, is followed by another verb, i.e. V2, which is delexicalised in the construction (Abbi 1991, 1993, Gopalakrishanan 1992, Hook 1974, Masica 1976). For instance, in the Malayalam sentence given below the V2 is stripped of its basic meaning ‘to go’. (explicator verbs are translated with capital letters through the manual). Malayalam 6.31. kuppi poÿÿi pooyi Bottle break-CP GO pst ‘The bottle broke’ (perfective) The first member of the ECV has variously been called ‘main verb’, ‘polar verb’, and ‘principal verb’. The second delexicalised verb is known a ‘operator’, ‘vector’, ‘explicator’ ‘auxiliary’ (though this is incorrect), ‘intensive auxiliary’ and also ‘light verb (in more recent Government and Binding theory). Let us first define the construction so that it is easily identified once confronted. ECVs are concatenations or sequences of two verbal forms. Of these two, the first member is the main or predicating verb and in most languages is in stem or some nonfinite form (see the Kashmiri example given below). The second member, although homophonous with an independent verb in the language, does not appear in its primary lexical meaning. The latter is the morphologically finite verb that is marked for relevant grammatical categories such as person, number, gender, tense, aspect, and modality. The second important factor is that the ECV alternates with a corresponding simple verb with no apparent change in the basic (cognitive) meaning of the predicate. In other words, use of one or the other does not alter the truth-value of the predication made in either case. Thus the parallel Malayalam sentence 6.32 of the one given above would hold the same truth-value as above. Malayalam 6.32

kuppi poÿÿI bottle break-pst ‘The bottle broke’

6. Syntax and Semantics

The second part of the definitional constraint automatically excludes all tense, aspect and modal auxiliaries from the class of explicators. However, for details readers are advised to refer to Gopalakrishanan et al (1992). Kashmiri being an SVO language generally has the main verb at the end of the sentence and the explicator in the middle of the sentence. The compound verb is made up of the conjunctive participial form [V-itH] of the main predicate plus the finite form of an auxiliary or “vector” verb homophonous with a full lexical verb. The following illustration is worth considering. Kashmiri 6.33

von kyaa k«ri bi goos yi kE:m kEr-itH Now what do I WENT this work do‘Now what shall I do, I have done this work’

Among the many semantic functions that an following are very significant.

(Abbi et al 1991) ECV

is capable of imparting, the

(a) Semantic functions •

Aspectual

•

Adverbial

•

Attitudinal

Each of these semantic functions of ECV can be seen in terms of several discrete or sometimes partially overlapping subtypes. The aspectual subtype is the least differentiated one. A South Asian universal, ECVs in general exhibit overlapping meaning of ‘perfectivity’ or action being seen as a whole, or action drawn to the last point, or completion. Most commonly the action verbs and in particular the verb ‘Go’ is used for indicating perfectivity (refer to table 6.2 on aspectual). Adverbial subtypes indicate various manners of an action, or non-manner indicating the benefactive characteristic of an action, or introvert and overt action, an irreversible action, anticipatory action, emphatic and definite actions etc. The attitudinal subtypes mark the attitudes of the speaker or narrator towards the action or actor, such as those of humility, respect, contempt, surprise, censurability or undesirability and anger, disgust or exasperation. We shall give a few examples of each of the three types for ready reference.

189

190 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Table 6.2

Explicator Verbs Indicating Aspect In Various Languages

Languages INDO ARYAN Hindi Punjabi Kashmiri Bangla Marathi DRAVIDIAN Tamil Malayalam Telugu Kannada Kurux AUSTRO-ASIATIC Kharia Gta/ TIBETO-BURMAN Meithei Kabui

Perfective/Action Drawn To The Last Point/Action Seen As A Complete Whole ja ‘go’, a ‘come’, le ‘take’, de ‘give’ ja ‘go’, a ‘come’, lE ‘take’, de ‘give’ nyun ‘take’, yun ‘come’, tsHunun ‘put’/‘wear’ ja ‘go’, a:sh ‘come’, ne ‘take’, de ‘give’ za ‘go’, ye ‘come’, ka:ê ‘draw’ viÿu ‘leave’, po: ‘go’, iÿu ‘put’ iÿa ‘put’, po: ‘go’, ka:ñ« ‘throw’ po: ‘go’, peÿÿ ‘keep’, we:s ‘throw’ ho:gu ‘go’, biÿu ‘release’ ka:l ‘go’, ci/ ‘give’ goê ‘pluck’, ce: ‘go’ we: ‘go’, bi/ ‘give’ tHok ‘exit’, taN ‘do, finish’, lau ‘put’ [Adapted from Abbi and Gopalakrishnan 1991]

(1) Aspectual Hindi 6.34

usne kHana kHaya He-erg food eat-pst-3msg ‘He ate the food’.

6.35

usne kHana He-erg food

kHa liya eat TAKE-pst 3msg

‘He ate the food’ (perfective) (2) Adverbial Punjabi

6. Syntax and Semantics

6.36

toshi ne kam k«r suÿÿiya Toshi-erg work do THROW-pst.3msg ‘Toshi did the work (violently)’

(manner)

Kurux 6.37

nin eNage o÷ÿe sveÿ«r t«s/oi ci/oi You for me one sweater knit-interro GIVE-interro ‘Will you knit a sweater for me?’

(benefactive)

Hindi 6.38

sita ne maâ se kah ra«kHa hE ki…. Sita –erg mother-acc say KEEP-pst aux comp…. ‘Sita has already told her mother that ….’ (anticipatory action)

In the Kurux example both the main verb and the explicator take interrogative inflection. This is the language that takes TAM markers on both of the verbs, the fact that goes against our definition of ECV. It has been noted that this could be due to the fact that the constructions are new to the language and in the transitory phase of language convergence, or more plausibly that since Kurux is a Dravidian language it behaves like any other sister language in the family. Hindi 6.39

ye mEâ kya k«r bEÿHI This I what do SIT-pst 1fsg ‘Oh, what have I done!’ (regret, irreversible)

There is sometimes an overlap between the non-manner adverbial reading and the attitudinal meaning, however, as described above we will restrict ourselves to the definitions provided for each of these categories. Consider the following examples. (3) Attitudinal Tamil 6.40

l«kshmi paaÿÿ«-paaêI kilicca Lakshmi song sing-pst prt TEAR –pst 3fsg ‘Lakshmi sang a song’ (contempt)

191

192 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Sinhala 6.41

«ya b«êu kH«ê«l« he things break

dan«w« PUT –prog 3msg

‘He is breaking up everything’ (undesirable) (Gopalakrishnan et al 1992) Dravidian languages in general show tense markings on both of the verbs. Whether the languages show the tense markings on the explicator or not, the reference of both the verbs is to the same temporal situation and only the final verb operates as a finite verb. The following table 6.3 gives a brief summary of the various verbs that occur as explicators in ECV constructions. For a detailed list of individual types and individual languages under each family, one may consult Abbi et al (1991). Table 6.3

Most Common Verbs Employed in ECV Constructions in Various Language Families Aspect

Attitudinal

Indo-Aryan

GO

SIT

Dravidian

GO

CONTAIN, PUT

Austro-Asiatic Tibeto-Burman

GO EXIT, PUT

Not available GO, COME, GIVE

Adverbial (manner)

Adverbial (others)

RISE, FALL, PUT, THROW GO, LEAVE, THROW

TAKE, GIVE

STRIP STAY, BREAK

PLUCK GIVE, KEEP

GO, KEEP, CONTAIN, THROW

! To elicit data on ECV one must look for verb serials and try to distinguish explicator from non-explicator. The biggest clue is that explicator verbs can always be replaced by a simple verb without violating the basic semantic content of the sentence. You may begin with the verbs listed in table 6.3. Make sentences with and without them and check for variation in the meaning. It is easy and simple to observe the aspectual category, as it is very common in Indian languages to use a verb of motion, and especially the verb ‘go’, to indicate this. The strategy to confirm the availability of the aspectual category is to first render a sentence with a single verb, get the meaning and then add the verb for ‘go’ at the end of the sentence and observe the semantic difference that this change creates. More often than not you will get the expected results. Also refer to the questionnaires given in the appendices. 6.8

Dative Subjects

Languages of the south Asian subcontinent make a distinction between subjects who act, perform or do an action from those subjects who undergo, experience, have, become of something they have no control over. That is, languages in this subcontinent make a very sharp distinction between performative and non-performative actions, as

6. Syntax and Semantics

the effect of the two are reflected distinctly on the respective subjects of the sentences. We are concerned with the latter type here. Subjects3 of non-performative activities generally take an oblique marking manifested more commonly in dative markings, hence the term ‘dative subjects’. Languages, however, have been seen to take other markings such as genitive and instrumental, although the dative is more common. For this reason I had earlier termed these constructions as ‘obliquely marked subjects’ (Abbi 1994). Prototypically the verb involved here is experiential in nature, although verbs of possession, attribute (both of animate and non-animate nouns) and ‘out-of-control’ (see below) also trigger dative subjects in many languages of India. See fig. 6.1 to get an overview of types of verbs involved in such constructions. All the major semantic types of constructions involving dative subjects are given in bold italics in this figure. Semantically the constructions are non-agentive in nature and thus imply an ‘involuntary’, ‘experiencer’, ‘benefactor’, or ‘recipient’ subject. A kind of semantic passivity is indicated on the part of the non-agent (or dative) subject. A few examples from Indian languages will make it clearer. 6.8.1 Non-Experiential The following constructions take oblique markings on the subject nouns but prototypically do not take verbs that are experiential in nature. Meithei 6.42

ram-gi makut –ama l«y Ram –gen hand one be ‘Ram has one hand’ (possession, inalienable)

Hindi 6.43

tumh-eâ kya mila you-dat what get-pst 3msg ‘what did you get?’

6.44

(benefactive)

papa-se ga}i chuÿ g«I papa-instr. train-f leave GO-pst 3fsg ‘Papa missed the train’ (out of control)

3

For a discussion of the identity of the ‘subjecthood’ see Abbi 1988, 1994 and Verma et al 1990. Also refer to Ura 2000: 95-145.

193

194 A Manual of Linguistic Fieldwork and Structures of Indian Languages Marathi 6.45

tya:la himm«t a:he he-dat courage aux ‘He has courage’.

6.46

(attribute, animate )

zHa:êa-la pHula a:het plant-dat flowers are ‘The trees has flowers’

(attribute, non animate)

As said earlier, the defining characteristic feature is that these constructions employ a prototypical non-agentive noun marked by an oblique marking, thus describing what Masica (1976) calls ‘subjective experience’ from an ‘external point of view’. Let us consider in some detail the typology of experiential constructions and the marking of the subject associated with them. 6.8.2

Experiential

Although it is difficult to make fine-grained distinctions and list all the types of experiences that this type of constructions denote, yet the following list (Abbi 1994:71) is helpful in identifying a typical dative subject construction. 1.

Sensory and mental experiences

2.

Physical and biological experiences

3.

Emotional experiences

4.

Compulsive/Obligational experiences

5.

Temporal situational experiences

(1) Sensory and mental experience constructions employ verbs indicating internalised desire, want, and need, such as verbs for ‘hunger’, ‘thirst’, ‘love’, ‘like’ etc. (e.g. 6.47 from Munda, 6.48. from Kannada and 6.49 from Hindi). 6.47

«iyaâ-ke pita p«seind mena I- dat bread like aux

‘I like bread’ 6.48

«v«ni-ge

hasiva-gide

he -dat

hunger-strike

‘He is hungry’

6. Syntax and Semantics

6.49

mujH-e tum-se pyaar hE I- dat you-instr love aux ‘I love you’

(2) Physical and biological experience constructions employ verbs indicating physical ailments and biological experiences such as ‘cold’, ‘cough’, ‘fever’, ‘pain’, ‘heart attack’ etc. Consider 6.50 and 6.51 from Munda and Newari, respectively, indicating biological experiences and 6.52 from Hindi exhibiting physical ailment. Munda 6.50

«iyaâ -ke a:du l«gao t«ra I- dat urinate strike aux

‘I feel like urinating’ Newari 6.51

ji-ta casu (li) l« I –dat itch become ‘ I am itching’

Hindi 6.52

ram- ko bukhar or sar dard hE Ram-dat fever and head ache aux ‘Ram has fever and head ache’

(3) Emotional experience constructions employ verbs indicating ‘feeling’, the stimulus for which is outside the experiencer, such as verbs for ‘worry’, ‘expect’, ‘be angry’, ‘be happy’ etc. Consider the following examples. Bangla 6.53

or

l!jja hocce

3sg-dat shy be ‘’S/he is shy’

195

196 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Marwari mH«n-÷e niâd a:ve I-dat sleep come 3sg

6.54

‘ I am sleepy’ There are other constructions that engage verbs indicating obligation or compulsion on the part of the prospective doer or forced doer of the action concerned. These verbs, prototypically, demand dative subjects. Consider the following examples. Nepali 6.55

ma-lai gH«r janu

cHa

I –dat home go-inf aux ‘ I have to go home’ Sinhala 6.56

ma-ta yana una I –dat go forced ‘I was forced to go’

! While collecting data on the aspect of obligation or compulsion one can translate all types of ‘modal’ sentences in English and most often than not one will receive dative subject constructions. (4) Temporal situational experience constructions are those that employ verbs that designate the subject as a patient in a temporally situated experience. Such verbs are stative action-process-experiencer verbs that signify the mental state of an experiencer with respect to the temporal aspect of the action that s/he is involved in. Consider the Hindi examples. 6.57

bH«iya –ko sigreÿ -pine- ki

ad«t hE

brother-dat cigarette –drink-inf –gen habit aux ‘The brother is in the habit of smoking’ or ‘..is a habitual smoker’. 6.58

us-e yahaâ he-dat here

bEÿHe- hue sit- pst part

do gH«÷ÿe ho g«ye two hours be GO –pst 3pl

‘ It has been two hours since s/he has been sitting here’ or ‘ S/he has been sitting here for two hours’

6. Syntax and Semantics

Dative subject constructions, as visible from the examples given above, pose some very interesting problems for the theory of case and agreement. For details see Verma and Mohanan (1990) and Abbi (1994). 6.8.3 Subject Properties. It has been a topic of much debate to decide whether the obliquely (i.e. dative) marked non-agent nominals should be considered subject of the sentence or not. As the question touches several theoretical syntactic and semantic issues, We will not go into detail here, but nonetheless would like to submit the following. Obliquely (i.e. dative) marked non-agent nominals do have semantic and syntactic subject properties. The arguments in brief are listed as follows: 1. They serve as anaphoric controllers. 2. They govern conjunction reduction even when the two identical NPs are distinctively case-marked. 3. Prototypically they occupy the subject initial position 4. They are generally the topic of the sentence For a detailed discussion consult the literature cited above.

!

As there are specific semantic constructs that lead to dative constructions, it is rather easy to elicit data in this area. One should start with bodily ailments, which are plenty in anybody’s life, followed by emotional and other biological experiences. Look for infinitival verb forms in obligational and compulsive constructions. Nonexperiential types involving benefactive, attributive and ‘out-of-control’ can be built up from the elicited data. These are constructions for which you have to depend solely on your semantic knowledge. There are possibilities of various other types of constructions not listed here that may indicate dative subjects (consult Verma and Mohanan 1990). You may consult the appropriate questionnaire given in the appendices.

197

198 A Manual of Linguistic Fieldwork and Structures of Indian Languages Dative Subjects

Non Experiential

Experiential Attributive

Benefactive

State

Out Of Control State State

Process

Inchoative

Stative Action Process

Process State/Process

Possessive

Alienable

Loss

Gain

True Passives

Incapabilitative Passives

Inalienable

Compulsive And Obligational

Sensory And Mental

Physical Emotional And Biological Temporal Situational

Fig 6.1

Types of constructions involving dative subjects

It is to be noted that not all types of experiential constructions trigger dative subjects. You may not find such constructions in Khasi, or you may find alternating patterns with (e.g. 6.59a) and without dative subject (e.g.6.59b) for the same semantic content. Hindi 6.59

a) mujHe bHukH l«gi hE I –dat hunger-fsg strike-fsg aux-3sg b)

mEâ

bHukha

huâ

I-nom hungry-msg aux-1sg ‘I am hungry’ Similar examples can be found in most of the Indo-Aryan languages. Altogether, languages may offer the following possibilities: (a) Languages may offer direct forms. (b) Languages may offer only indirect (dative, genitive, instrumental) forms. (c) Languages may offer both direct and indirect forms which alternate freely.

6. Syntax and Semantics

(d) Languages may offer both direct and indirect forms but have specific restrictions on their occurrences. (e) Languages may offer both direct and indirect forms but alternate them in specific environments only. Most of the languages of South Asia fall under the category of (e). Most of the Tibeto-Burman languages fall under the category of (a). Various individual languages fall in between. The presence or absence of dative vs. nominative (or zero) marking is linked to the question of topic and focus which needs an independent study and should not be ruled out if the alternating patterns are visible in the language under investigation. 6.9

Complementation

Complementation is a grammatical process by which we can use a sentence or a clause as the argument of a higher sentence. In other words, any sentence can have an entire sentence or a clause functioning as a noun phrase (its argument). This argument is either attached to a complementiser and thus is retained as a sentence/clause or is changed in various ways (such as participial constructions, infinitival phrases etc.) in Indian languages. Several types of complements signify several types of meanings and they differ from language to language. Linguists such as Bhat (1994, 2000) have viewed this phenomenon as a case of decategorisation or recategorisation; “Shifting of entities from their own categorial function to the function of other entities generally involves both decategorisation as well as recategorisation…when they [sentences] are shifted from their own sentential function to that of nouns (argument position), they lose some of their sentential characteristics as a result of recategorisation” (ibid. 2000:162). Since it is viewed as a shifting category, the degree of shift from its host category varies from one Indian language to another. One should not be surprised if one finds a typical nominal-like character in the complement clause. Let us briefly consider some of the typical complements. 6.9.1 Types of Complements 6.9.1.1

Ordering

The complement may either be preceded or followed by a complementiser as the case may be. The following examples will demonstrate both phenomena 6.9.1.2

Sentential.

The entire sentence may be used as an argument of the matrix sentence without bringing about any change in the sentence as if it were reported speech.

199

200 A Manual of Linguistic Fieldwork and Structures of Indian Languages Hindi 6.60 tum-ne k«ha tHa

ki

you-erg. say aux (pst) that

tum

aoge

you

come will 2msg(fut)

[indirect speech] or 6.61

tum ne k«ha tHa ki maiâ auâga you-erg. say aux (pst) that I come 1msg will (fut) ‘you said that you would come’ [reported, direct speech]

Marathi 6.62

te mH«÷a:le ki they said

6.63

udya

that

udya tomorrow

amHii tikêe yeu

tomorrow we

amHii ikêe yeu

there come

we

here come

as«:

te mH«÷a:le

thus

they said

‘They said that they would come here tomorrow’ (Bernsten and Nimbkar 1982 reported in Bhat 2000)

Notice that Marathi gives an option between pre-sentential or post-sentential complementisers. Unlike English, both Hindi and Marathi exhibit no change in the tense marking of the complement clause. Hindi optionally changes the pronominal, as the sentence 6.60 can also be rendered as 6.61, i.e. as a case of reported speech. All SOV languages of the Indo-Aryan family as well as those under their influence (such as Kurux, which, being spoken in the midst of IA languages, behaves very much like Hindi, although is a Dravidian language), place the complement clause in post-verbal position. This in effect leads to SVO order in predominantly SOV languages. Consider sentences 6.60 and 6.62 again where the structure is subject + verb + complementiser + object (of complement). Analogously, sentence 6.63 actually exhibits the OSV order. This issue has been much debated lately, and linguists interested in word order would definitely find enough material in Indian languages. 6.9.2 Varying forms of Complementisers Languages may distinguish between various types of events and states, which are reflected by choosing different complementisers. Meithei is such an example. The language has two complementisers, an adverbial hayn« and a nominal hayb« both derived from the verb root hay ‘to say’. “The former is used when the complement indicates a statement whereas the latter is used when the complement denotes a fact (a state, process or action)” (Bhat 2000: 165). Consider.

6. Syntax and Semantics

6.64. m«hak he

«yn«

hay-n«

lak-u

come-imp. say-adv

t«k-i

I-nom perusade-non-fut

‘ I persuaded him to come’ 6.65

m«hak lak-k«ni hay-b« ya-y he come-fut say-inf. possible-non future ‘it is possible that he will come’

Once a complementised sentence works as an argument it can be case marked. Consider the Kannada sentences (Bhat 2000): 6.66. avanu p«÷êita emb-ud -annu na:vu tiñidiral-illa he scholar

that –it-acc

we

know-not

‘We did not know that he was a scholar’ 6.67. avanu p«÷êita emb-udar-alli sanshya illa he

scholar

that-it loc

doubt neg

‘There is no doubt that he is a scholar’ 6.9.3 Nominalised Complement A nominalised complement is another common feature of Indian languages where infinitival constructions are used as arguments and thus lose verbal characteristics such as tense and aspect distinctions. Consider the Hindi and Kurux sentences given here. Similar sentences may also be found in other Indian languages. Hindi 6.68

b«cc-e ka

zor zor se

ro-na

child-obl-gen loud loud-instr. cry-inf.

mujhe pasand nahiâ I-dat like neg

‘I don’t like the child crying loudly’ Kurux 6.69

ENgHe

is«n b«r-na

as ge mal dau l«ggya

my-gen here come-inf he to not hood feel ‘he did not like my coming here’ As these are nominal complements, they behave like any other nouns in the language and thus can be case marked by different cases as the situation demands. The

201

202 A Manual of Linguistic Fieldwork and Structures of Indian Languages genitive, however, seems to be more popular than any other case. Other than the Hindi and Kurux examples that we saw, consider the Kannada and Meithei examples given below (Bhat 2000): Kannada 6.70. avan-a sanjey-a ba:gil-ina tereyuvike nana-ge gottu he-gen evenig-gen door-gen opening I-dat know ‘I know his opening of the door (of the evening)’. Meithei 6.71

m«hak-ki ciN-gi he- gen hill-gen

ka-r«m-b« m«-ka climb-comp-inf prefix-climb

‘his (complete) climbing of the hill’. 6.9.4 Whether English constructions that take ‘whether’ are also considered as complement clauses, though Indian languages have equivalents of ‘or not’ suffixed at the end of the sentence to indicate the complement. Consider: Kurux: 6.72

en ba-a

pollom as bar-os

ka m«lla

1sg say-neg able- 3sg come msg or

neg

Hindi: 6.73

mEâ n«hiâ b«ta s«kta (huâ ) ‘I

ki

vo ayega

ya n«hiâ

neg tell modal aux comp he come fut or neg

‘ I do not know whether he will come’

!

To elicit data on complementation, you should combine two sentences from your notebook. That is, rather than making a new sentence, try to combine two sentences already collected so far. This will help you locate any change in the tense or aspect marking in the final clause. On the basis of the examples given above you can easily do this job. Think of making adjectival clauses and embed them with the main clause. Narration, folklore and idioms are sometimes helpful in eliciting data on complementation.

6. Syntax and Semantics

6.10

Converbs or Conjunctive Participles (CP for short)

In popular linguistic literature several names such a ‘gerundive’, ‘conjunctive participle’, ‘medial verbs’4 and more recently ‘converbs’ (see below) have been used for this phenomenon. Gerundive is the term more suitable for western European languages. For South Asian languages the term converb is more suitable as we do not have a prototypical gerundive. Conjunctive participle is the term mainly used, because the origin of the constructions under consideration is ascribed to conjunction of two clauses, and there is a participle suffix following the non-finite verb. We shall use CP for this suffix; however, a verb in non-finite position serving the various functions of a converb (as described below) will be referred to as a ‘converb’. A converb is defined as “a non-finite verb form whose main function is to mark adverbial subordination” (Haspelmath 1995: 03). Converbs or conjunctive participles are modifiers of verbs and thus have an adverbial function. The verb in the converb is inherently subordinate and linguists have started believing that converbs should form a separate class in grammar (Mescaninov 1945 reported in Haspelmath 1995). Most of the IA languages use the verb ‘to do’ in its basic or commonly termed ‘absolute’ form (non-finite) and attach it to the basic form of the verb under consideration, e.g. kha‘eat’+ k«r ‘to do’ = ‘having eaten’ in Hindi. Prototypically, conjunctive participle constructions or converbs should be non-finite and should perform adverbial functions. However, the demarcation between finiteness and nonfiniteness is not very clear. Many languages may fall in between the two poles of +/finiteness, where the converbs may not take tense, aspect or mood markings but still show subject encoding. Some languages, such as Kurux, do not use the verb in its root form; instead, they inflect it for tense marking. The participle suffix –ki is added to this inflected verb. Thus, tense markings are exhibited twice, once on the CP form and a second time on the finite verb form. Consider: 6.74

en nal«x n«nn-on ki cail kal-on 1sg work do- fut CP walk go-fut ‘I will leave after doing the work’

6.75

en d«vaiyaâ

ho- on ki

cail kal-on

1sg medicines take-fut CP walk go-fut ‘I will leave after taking medicines’ Now the question is whether we should or should not include these verbs (in bold) as converbs, as they do not appear in non-finite form and thus do not fit the definition of converbs. However, they function as any converb does. We would thus 4

As Haspelmath (1995) explains the term ‘medial verb’ comes from Papuan linguistics and was apparently first proposed by Stephen Wurm (cf. Thurman 1975)

203

204 A Manual of Linguistic Fieldwork and Structures of Indian Languages like to include them under converbs as they are adverbial in nature, and although they are inflected for tense they are semantically non–finite. Another important aspect to note is that the suffix –ki in Kurux is a recent borrowing from IA languages, and therefore perhaps the construction is undergoing a transition to a crystalised form. This also raises the important question as to whether all conjunctive participles or converbs are derived from conjunctive sources and have been gradually grammaticalised in the process of language evolution. If this is true, the Kurux construction would be much younger and might be stripped of tense markings in a short period of time, with areal pressures as well as with diachronic forces operating on it. However, we must not forget that tense marking on converbs is allowed in Dravidian. Whether both verbs are tensed or not, the reference of the non-finite verb coincides with the reference of the finite main verb in its tense and mood. Thus, in the following Hindi sentence ‘having eaten’ is indicative of the same tense as that of the finite verb, that is past. In most of the Indo-Aryan languages, the subject in CP constructions maintains its identity with the subject of the main finite verb. Hence, the scope of the CP construction is the same as that of the finite verb in the following Hindi sentence, i.e. the subject of the two actions ‘eat’ and ‘sleep’ is the same. 6.76

pitaji kHana kHa-k«r so g«ye father food eat-CP sleep GO 3msg hon. past ‘Having taken his meals the father went off to sleep’

The subject identity constraint is not necessarily maintained by Dravidian languages such as Malayalam or by those Indo-Aryan languages that are highly converged towards Dravidian, such as Konkani. An interesting rule operates here. If the two actions are undertaken by distinct agents, subject identity or co-referencing is necessary. If only one of the two nouns is agent and the second one is a dative or experiencer, co-referencing is not a necessary condition. The following examples are very natural in Dravidian languages. Kannada 6.77

mañe band-u kere tumbitu rain come-pst-CP tank fill-pst ‘The tank filled as a result of rain’

(Causal)

(Bhat 2000:197)

While talking about co-referencing it must be brought to the notice of the fieldworkers that the subject identity condition operates even if the two clauses are of a different nature, i.e. one is in active and the other is in passive voice. Thus, in the following Hindi example the converb and the finite verb refer to the same agent, although the first one is ellipsed (that itself is proof that it refers to the same agent) in this construction.

6. Syntax and Semantics

6.78

[f] gH«r p«huâc-k«r bina-se kHana n«hiâ kHa-ya g«ya home reach-CP Bina-instr food neg. eat-pst GO-pst 3msg ‘[f]Having reached home Bina could not eat the food’

OR

‘[f]Having reached home food could not be eaten by Bina’ 6.10.1 Semantic Functions In CP constructions verbs assume the role of verbal modifiers and thus are treated in recent literature as converbs. They generally exhibit the following semantic functions (for details see Abbi 1980,1984, Davison 1981, and Kachru 1981): (a)

Sequential, clause chaining (attained by medial verbs) (e.g. 6.74, 6.76, 6.79, 6.80)

(b)

Causal/temporal (e.g. 6.77)

(c)

Manner (e.g. 6. 81b, 6.82b, 6.83)

(d)

Antithetical/contrastive (e.g. 6.84)

(e)

Attributive (e.g. 6.85)

The most important and basic function of converbs is to coordinate two clauses in a sequential fashion. The examples seen so far indicate this basic function. Related to this function is the phenomenon of clause chaining that depicts multiple sequences. Let us consider clause chaining in some detail. Clause chaining is associated with several (medial) verbs each suffixed by a CP with one final finite verb following them. Dravidian languages allow series of converbs in a chain, unlike Indo-Aryan languages that do not normally allow more than two converbs. Thus, the following would be ungrammatical in Hindi but not in any Dravidian language. 6.79

*sur«j c«m«k-k«r,

dHup nik«lk«r h«wa c«l k«r

k«p}e sukHe

The sun shine-CP, sun-come-CP, wind flow-CP clothes dried ‘The clothes dried because the sun came out and the wind blew’ As far as clause chaining is concerned, unless the two actions are semantically related and belong to the same semantic field they are not conjoined by converbs in most of the Indo-Aryan languages (Marathi and Konkani are exceptions). Dravidian languages, on the other hand, allow two independent actions to be conjoined by a converb if they are in a cause and effect relationship, as seen in the Kannada and Malayalam sentences given above. Indo-Aryan allows multiple events in a chain provided the converb in the lower clause is adverbial in nature and thus corefers to the

205

206 A Manual of Linguistic Fieldwork and Structures of Indian Languages subject of the lower clause. This clause in turn is a complement of the higher clause. Thus, sentences such as the following (Hindi) are possible in all languages: 6.80

mEne [anuj ko bHag k«r aate hue] dekHa I -erg [Anuj-acc run-CP come-part] saw 3msg ‘ I saw Anuj coming running’

CP constructions of state verbs such as ‘to know’, ‘to think’, ‘to understand’, ‘to stay’ more often than not indicate both time (sequential, e.g. [a] sentences) and manner (simultaneous, e.g. [b] sentences) (Abbi 1984:256). Thus the following sentences from Hindi exhibit both of the readings.

6.81

firdaus ne dilli meâ r«h-k«r biznes kiya Firdaus –erg Delhi-loc stay-CP business do pst 3msg

(a) ‘Firdaus lived in Delhi and did her business’ (b) ‘Firdaus did her business while being in Delhi’ 6.82 radha-ne Radha–erg

samajH -k«r

j«wab diya

understand-CP answer GAVE

(a) ‘Radha first thought and then answered’ (b) ‘Radha gave the answer thoughtfully’ Let us now consider converbs used for other semantic functions. All examples are from Hindi. 6.83

vo musk«ra-k«r bola He smileCP speak –pst 3msg ‘He spoke with a smile’

6.84

vo p«tr« na p«}H-k«r «kHbar

(manner) p«}H-ne l«ga

he letter neg read-CP newspaper read-inf begin-pst3msg ‘Instead of reading the letter, he started reading the newspaper’ (antithetical. These are always accompanied by a negative particle) 6.85 vo parÿI meâ sa}I p«h«n-k«r ai tHI she party-in sari wear-CP come-pst 3fsg ‘She came to the party in a sari’ or ‘She wore a sari to the party’ (attributive indicating personal adornment)

6. Syntax and Semantics

!

Watch for the Inflection on CP. What has to be remembered is that Indian languages behave differently as far as the TAM categories shown on the converbs are concerned. As said earlier, some languages such as Hindi use the root form of the verb, while Dravidian languages retain some of the TAM markings on the converbs. Marking of GNP on converbs, however, is rarely seen. 6.10.2

Some Idiosyncratic Constructions

There are constructions in Indian languages that involve (a) time expressions. These always involve CPs and are adverbial in nature. 6.86

vo d«s

b«j-k«r

n! min«ÿ

p«r gH«r se

He ten o’clock-CP nine minutes-loc home-abl

nikla came out-pst 3msg

‘He left home at ten past ten’ (literally speaking, two actions, one of ‘the clock striking 10:10’ and, the other ‘his leaving the home’, are sequenced) In languages of the Indo-Aryan family, the conjunctive participle is attached to (b) partially reduplicated verbal compounds to indicate sequential action or manner as the case may be. This particular use of the conjunctive participle is very productive and should not be missed by the fieldworker. An example from Hindi will be sufficient here: 6.87

tum p«}H likH-k«r kya k«roge you read write-CP what do-fut2msg ‘What will you do after being educated?’

These compounds are frozen expressions and designate specific meanings such as ‘educated’ in this particular case.

!

If you have your questionnaire in Hindi or any other Indian language you will have ample examples of converbs and getting them translated should not be a problem. Considering the five semantic functions mentioned above you can make your own sentences from the collected data and get them approved by the informant. First try to identify what the CP suffix is. Once that is known, use it with past tense form of a verb which itself is followed by another past tense verb in the finite position. This holds for Dravidian languages. If you are working on Indo-Aryan languages, simply use the base form and add -k«r or its cognates in various dialects and languages of the same family and then complete the sentence with a finite verb. Find out what the meaning of the sentence is. Make sentences involving two actions, one the cause or manner of the other. Also make sentences with reduplicated verbs suffixed with a CP, such as Hindi dekH dekH k«r ‘see see CP’. Also ask the informant what time it is. These expressions invariably use conjunctive participles for

207

208 A Manual of Linguistic Fieldwork and Structures of Indian Languages indicating past minutes. Once you obtain the expression you can fruitfully use it in framing a new sentence with a finite verb following this time expression. Converbs are very common in Indian languages, thus any piece of narrative text should also give you these structures. 6.11

Anaphora

This topic has been discussed as much in structuralist as in generative grammar. Recently, the topic has been widely researched in the framework of “binding theory” of the universal grammar. We will not discuss the theory here, as this might restrict the manual being used by those who do not want to follow a specific theory. For a detailed study of South Asian anaphora the best reference is Lust et al (2000). Readers are recommended to consult this volume as it intertwines theoretical aspects with a South Asian typology of anaphoric structures. We, however, would like to enumerate and discuss what kinds of inquiry one should have in this area. 6.11.1. Definition Anaphora refers to a process by which a word derives its interpretation from previously expressed linguistic units. In modern grammars it refers to a type of noun phrase that has no independent reference but derives it from some other constituent, its antecedent. Prototypically, anaphors include among others, reflexive pronouns, reciprocal pronouns, and personal pronouns used in discourse that have antecedents existing in previous clauses. In short, by anaphor we mean referentially dependent NP types such as reflexives and reciprocals. We have already seen the variety and range of some of the anaphors while discussing pronouns in an earlier chapter. We shall now account for the types of anaphoric relationships and their signifiers available in Indian languages. 6.11.2. Reflexives There are various strategies employed by Indian languages for forming reflexives, some of which are given below. •

Reduplicating the pronouns in various case-marked positions. Munda and TibetoBurman languages follow this strategy. Thus, in Kharia the first and second person pronouns reduplicate for reflexive.

6.88

iø iø -te I I -acc

gil ti -ø hit-tr-prs-1sg

‘I hit myself’ 6.89. am am-te uwa -na -m you you-acc bathe-intr.-fut 2sg

6. Syntax and Semantics

‘You bathe yourself’ A discontinuous reduplicated structure is more common in Mizo. 6.90 . amah leh amahi zovai a-in kaap he self he self Zova 3sg-VR-shoot ‘Zovai shot himselfi 6.91 amaah leh amaahi –cunga zova-cui a- thinrim he self and he self

on

Zova-foc 3sg-angry

‘ Zovai is angry at himselfi’ (Murthy and Subbarao 2000) •

Making use of verbal reflexives (VR) [in] as seen in the Mizo example above is very common in Dravidian languages, too (with the exception of Malayalam). Consider the Telugu example.

6.92

radHa (tana-ni (tanu)) poguêu-kon-di Radha self-acc self praise VR agr ‘Radha praised herself’

There are exceptions where VRs cannot be used. For full details see Murthy & Subbarao in Lust et al 2000. •

There is normally a fixed form for reflexives which does not change with GNP, e.g. Hindi has «pne ap ‘self’, Telugu has tana and Kharia has êom, and Khasi has lada. This is true of most of the other Indian languages, too. Consider the Hindi (6.93) and Khasi (6. 94, 6. 95) examples:

6. 93. maine/tumne/usnef,m /unhoânef,m «pne-ap-ko kosa I-erg/ you-erg/s/he-erg/they-erg self cursed ‘I/you/s/he they cursed self’ 6.94. Na sN«ut«n«t ia-lade I like-1sg acc-self ‘I like myself’ 6.95 u j!n u-la piny«p ia-lade msc John msc-pst kill acc-self ‘John killed himself’ However, there are languages in the Munda family which encode subject/object GNP on reflexive. These are highly pro-drop languages and since the verb takes all the encoding as discussed earlier in chapter II, it is not very surprising for this trait to occur.

209

210 A Manual of Linguistic Fieldwork and Structures of Indian Languages Santhali 6.96. nua k«mi d!

6.97.

ali:ø-tEli:ø k«miya 1du-excl-self

nua k«mi d! abu-tEbun 1pl-excl-refl

6.98. nua k«mi d!

alE-tElE

k«miya

k«miya

1pl-incl-refl ‘Let us do this work ourselves’ 6.99

a:mi d! a:m tE-mi k«:mi mE you p.e. you-self work-imperative [p.e.= pleonastic element] ‘You do the work yourself’

This is the reason that the English ambiguous sentence (6.100) given below can never be read as ambiguous in Santhali (6.101) as each reflexive is assigned its antecedent marking. 6.100 Rami asked Mohanj to wear hisi/j coat 6.101 rami mohanj-e mitadiya je ama?i/iøa?j koÿgi h!rogmE Ram Mohan-SM say pst that hisi/hisj coat wear imp •

With a quantifier, the subject reflexive occurs in reduplicated form to indicate distributive function.

Hindi 6.102. s«b «pne «pne gH«r jao all self self home go-imperative ‘Each of you should go to his/her respective home’ Sadari 6.103. s!b iskuliyam«n «p«n «p«n kitab leke a:l«y all school children self self book take –CP come pst 3mpl ‘All the students brought their books’ ‘Each student brought his/her book’

6. Syntax and Semantics

•

An interesting typological feature mentioned by the editors (Lust et al 2000) is that unlike English and Hindi Indo-Aryan languages of the outer sub-branch, i.e. Marathi, Gujarati, Sinhala, and Oriya and Dravidian languages allow a nominative reflexive (anaphor) subject in finite embedded clauses. Consider:

Gujarati 6.104. sudhaai j«÷a:ve che ke p«dma:tHj potei,*j n«tHI ê«rti Sudha inform aux that Padma-inst self not feared ‘Sudhai informs that selfi, *j is not afraid of Padma’

(ibid :17)

6.11.3 Intensifiers or Emphatics. These and anaphor reflexives are homonyms in most of the Indian languages, and should therefore be carefully elicited, as one cannot be sure whether the form exhibits anaphoric relationship or not. In the following Hindi examples the bold forms are intensifiers. Hindi 6.105a sita- ne «pne-ap/kHud kHana b«naya Sita-erg self food cook-pst ‘Sita cooked the meal herself’ Siemund (2000) has discussed these intensifiers or emphatic reflexive “self forms” in detail in his latest work. Interestingly, the intensifier can float without bringing any difference to the focussed constituent. Thus, the following three options in Hindi do not bring any difference to the focussed subject noun Sita: 6.105b

(kHud) sita-ne (kHud) kHana (kHud) b«naya

6.11.4 Reciprocals A very productive strategy in various Dravidian, Austro-Asiatic and Tibeto-Burman languages is the reduplication of pronouns, or reduplication of the word for ‘one’. “Thus in Kannada obba ‘one (for +human)’ and ondu ‘one (for –human)’ will be rendered as obbar-anna-obbaru ‘one-acc one-nom’ ‘each other-acc’: ond-akke ondu ‘one-dat one-nom’ ‘one another dat’. In these phrases, the first obba or ondu is in the argument position and carries the case appropriate to its position in the sentence. The second occurrence of obba or ondu is in non-argument position, and its case is a copy of the case of the antecedent of the reciprocal phrase.” (Amritavalli in Lust et al. 2000). Tamil (6.106) and Meithei (6.107) have similar strategies. Consider:

211

212 A Manual of Linguistic Fieldwork and Structures of Indian Languages 6.106. ortt«r-kul-ortt«r peshi ko÷êa:rk«l one person-in-oneperson spoke did-pl ‘They spoke among themselves’ 6.107. m«khoi amm«g«- amm«g« nunski n«I they one –one love ‘They love each other’.

(Abbi 1992)

Khasi offers two strategies: (1) it reduplicates the person marker with an infix of a verbal reflexive or (2) it uses only the verbal reflexive. 6.108. ki masi ki la kiny«t iwei-ya-iwei 3pl cow 3pl pst kick one-VR-one 6.109. ki masi ki la ya 3pl cow 3pl pst VR

kiny«t kick

‘The cows were kicking each other’ Santhali also reduplicates its pronominals for reciprocal reading. Note that reduplicated pronouns encode the subject/object marker as well as the number. Thus: 6.110 unku d! aku-aku-ku 3pl

l«}H«ina

p.e. one another-sm fight-pst

‘They fought with each other’ The following table can be consulted for Santhali reciprocals. Table 6.4 Reciprocals in Santhali Dual

Plural

Exclusive

ali:N-ali:N

alE-alE

Inclusive

ala:N-alaN

abu-abu

2nd PERSON

Inclusive

abEn-abEn

apE-apE

3rd PERSON

Exclusive

akin-akin

aku-aku

1ST PERSON

Kharia, although belonging to the same family as Santhali, uses a VR prefixed to the verb for reciprocal meaning. Thus gil ‘to beat ‘ and kol-gil ‘beat one another’; lebui ‘love’ and kol-lebui ‘to love one another’.

6. Syntax and Semantics

!

Elicitation on anaphors is not very easy to begin with. You must begin by obtaining the reciprocal constructions, as they are easy to explain. To elicit information on specific anaphors you will have to take recourse to the translation method. Also remember that one has to employ two connected sentences, one in the affirmative and the next in the negative, to extract the right meaning of the anaphor. For instance, to ensure that you get the right anaphor, you might have to say: “how would you say that ‘mother wore her clothes and not her sister’s?’” 6.12

Coordination

The term coordination has been best defined as “syntactic constructions in which two or more units of the same type are combined into a larger unit and still have the same semantic relations with other surrounding elements” (Haspelmath: 2000) The identity of both semantic and syntactic functions is a necessary prerequisite for coordinate constructions (Schater 1977, Abbi 1980). The units in coordinate constructions may be words (e.g. nouns, 6.111a), phrases (e.g. verb phrases, 6.111b), subordinate clauses (e.g. 6.111c), or full sentences (e.g. 6.111d). Hindi 6.111a.

imtiaz !r Shgufta dilli aye Imtiaz (m) and Shagufta (f) Delhi come-pst 3pl Imtiaz and Shagufta came to Delhi

6.111b.

(conjunction)

imtiaz dHire -se bola p«r zor se h«âsa Imtiaz softly-instr. speak-pst 3msg but loudly-instr. Laugh-pst 3msg ‘Imtiaz spoke softly but laughed loudly’ (disjunction, adversative)

6.111c.

imtiaz ko malum tHa

ki Shagufta der se

ayegi

Imtiaz-dat know aux-pst3msg that Shagufta late-instr come-fut 3fsg aur der t«k ÿHhEregi and late-instr. till stay-fut 3fsg ‘Imtiaz knew that Shagufta will come late and stay for long hours’ 6.111d.

imtiaz ne p«}Hana kH«tm« n«hiâ kiya tHa p«r sare b«cce klas ke bah«r Imtiaz-erg. teach-inf. finish not did but all children class-gen outside a g«ye come pst 3pl ‘Imtiaz did not finish teaching but all the children came out of the class’

213

214 A Manual of Linguistic Fieldwork and Structures of Indian Languages All languages possess coordinate constructions and there may be several linkers or coordinators that serve as links between the two or more units. Thus, in Hindi coordinate conjunction forms are aur, aur--bHi, tis p«r—bHi, and similarly disjunction forms are ya, «tHwa, ki, c«he…cahe etc. In general, coordinators indicate the following common semantic functions in various languages: • Conjunction • Disjunction • Adversative • Negative coordination • Causal coordination Examples given below may familiarise the readers with the various semantic types of these constructions. For a detailed description of Hindi coordination see Abbi 1988 and of a general typology of coordination refer to Johannessen (1998) and Haspelmath (2000). Punjabi 6.112 tuâ ya

to

«ram nal

bE ja ya fer n«s ja

you either emph leisure-instr sit GO or then run GO ‘You sit down comfortably or get lost’. (Disjunction) Hindi 6.113. mEâ gHumna

cahta tHa

isliye

mEne cHuÿÿI li

I visit places want aux pst therefore I

leave took

‘I wanted to visit therefore I took leave’ (Causal Coordination) 6.114 mEâ janta tHa isliye na«hiâ bola I

knew therefore not speak-pst 3msg

‘I knew it therefore did not speak out’ 6.115 b«ccoâ -ne

bHikhari -ko pEse

n«hiâ diye

b«lki use mara

children-erg beggar-dat money not give-pst3pl instead he-acc hit ‘The children did not give any money to the beggar, but (instead) hit him’. (Adversative coordination)

6. Syntax and Semantics

The use of adversative conjunction b«lki can be used for additive meaning too. 6.116 usne ram ko n!kri hi

n«hiâ dilai

b«lki uski shadi bHI ki

he-erg Ram-dat job emph not get-caus-pst but his marriage also did-fsg-pst ‘He not only got Ram a job but also married him off’ (adversative coordination with positive meaning because of the emphatic particles [italicised] in coordinands) 6.117 vo na

kHana kHata hE

na kHelta hE

he neither meals eat-3msg prs. aux nor play-3msg prs. Aux ‘He neither eats nor plays’ (Negative coordination) The position of the first negative particle na is rather mobile in most of the languages. Thus the following is also possible: 6.118 na vo kHana kHata hE na kHelta hE 6.119 cahe

vo aye cahe

vo na aye..

whether he comes whether he not comes ‘Whether he comes or not.’ 6.12.1

Coordinands

Units taking part in coordination have been termed coordinands (Haspelmath: 2000), and coordinands in Indian languages more or less follow the universal pattern. The following features can be generalised: •

Conjunctions, disjunctions and causal coordinators may link two or more coordinands, e.g. sentences 6.120, 6.121, and 6.122.

•

Adversative coordination is always binary and thus cannot link more than two coordinands, e.g. ungrammaticality of sentence 6.123.

Hindi 6.120 ve gH«r

g«ye

!r

n«haye

!r

so g«ye

they home go-pst 3mpl and bath-pst 3mpl and sleep GO-pst 3mpl ‘They went home, bathed and slept’ (multiple events linked by conjunctions)

215

216 A Manual of Linguistic Fieldwork and Structures of Indian Languages

6.121 ve ya «Ngrez hEâ ya rusi ya phir «mriki they either British aux-pl or Russian or then American ‘They are either British, or Russian or American’ (multiple modifiers linked by disjuncts) Multiple phrases linked by causal coordinators can either be of the type given below or may use converbs (conjunctive participles). The latter is more common in Dravidian languages. See also the section on converbs. 6.122

mEâ gHumne n«hiâ jauâgi kyoâki m!s«m kH«rab hE, kyoâki meri t«biy«t ÿHik I stroll-inf not go-fut 1fsg because weather bad aux because my health good n«hiâ hE, aur kyoâki

gH«r meâ

koi n«hiâ hE

not aux, and because home loc no one

aux 3sg

‘I do not want to go out for a stroll as the weather is not good, as I am not feeling well, and as there is no one in the house’ 6.123* pitaji ne mujhe daâÿa p«r mEne bura n«hiâ mana p«r mEâ bHag g«ya father –erg I –dat scold-pst3msg but I-erg bad not feel but I ran away ‘Father scolded me but I did not mind but I ran away’ (multiple adversative coordinators) While working on coordination one has to observe the number of coordinators (co for short) and their positions in the sentences. For each type discussed above one may find two or more than two types distinguished on a semantic basis. Positional occurrence may also vary from language to language. The most common occurrence is obviously in-between the two coordinands. The following possibilities, however, may emerge: •

A B (without any overt coordinator, e.g. 6.124)

•

A co B (all examples given above except 6.121 and 6.122)

•

A co B co (e.g. Sanskrit, Kannada 6.125)

•

co A co B (e.g. 6.119, 6.121 given above)

•

co A B (e.g. 6.113 and 6.114)

Khasi 6.124 ka-sim

ka -hEr

l«n –ba ki-diyEn [ ] sa l!r-lum

F1sg-bird F1sg.-flies through F1sg-tree [ ] to top-hill ‘The bird flies through the trees [and] to the top of the hill’.

6. Syntax and Semantics

Kannada 6.125 prakasam-u: krishnan-u: pe:ÿe-ge ho:d-aru Praksam-and Krishnan-and market-dat. Go-pstp3pl ‘Prakasam and Krishnan went to the market’

!

Note the agreement of the finite verb with the GNP of the coordinands. Check whether the verb agrees with the first or second coordinand or is in general plural. Each language may offer different examples. 6.12.2 Gapping/Ellipsis It is observed that phrasal coordination necessarily involves ellipsis of identical repeated elements. Thus, the Hindi sentence 6.126 is a derivation of 6.126a and 6.126b, respectively. Such an ellipsis transforms an underlying biclausal structure into a monoclausal one. They are also known as conjunction reduction and can gap coreferential nouns, even if they are case-marked differently [e.g. 6.127] (Abbi 1994). 6.126 b«ccoâ- ne

aiskrim

kHai

!r

b«}oâ ne

miÿHai f

children-erg ice cream eat-pst fsg and adults-erg f sweets ‘The children ate the ice-cream and the adults the sweets’ (ellipsis of the identical verb ‘eat’) 6.126a b«ccoâ -ne aiskrim kHai

‘children ate the ice cream’

6.126b b«}oâ -ne miÿHai kHai

‘adults ate the sweets’

6.127 mujHe nind

aur f so g«ya

ai

I-dat. sleep-f come-pst f3sg and f (I-nom) sleep GO-pst 3msg I felt sleepy and went off to sleep’. As most of the Indian languages use forward ellipsis (termed as analipsis by Haspelmath), where the site for ellipsis is the second coordinand, rather than backward ellipsis (termed as catalipsis by Haspelmath), where the site for ellipsis is the first coordinand, one should concentrate on the ellipsis of the second coordinand.

!

Start with collecting natural conjunctions such as ‘man and woman’, ‘day and night’, ‘needle and thread’. These are the conjunctions (with no intonation break) that exhibit one conceptual unit (Mithun 1988:332), a kind of bahu:bri:h compound in Sanskrit grammar. After this exercise, substitute conjunctions by disjunctions and start

217

218 A Manual of Linguistic Fieldwork and Structures of Indian Languages building up long and longer sentences. The best way of eliciting data on conjunctions is to tape a spontaneous narration. To elicit data on gapping, combine two full phrases and sentences and ask the speaker what would be the natural way of saying both of the sentences together. Keep deleting the repeated units, whether verbs, nouns, or modifiers, and check for ungrammaticality. 6.13

Adjectival Clauses

Sentences can be embedded to modify a noun. Two significant strategies employed by South Asian languages are: (a) Relative-correlative constructions typical of IA languages, where the modifying clause, marked by a member of the “J” set of relative pronouns (which, what, where etc,), adverbs, and other words is represented by a correlative modifying the NP and preceding the main clause. (b) The adjectival participle or participial clause, typical of Dravidian languages, precedes the modified noun. Participles are either past perfective or present imperfective and concord with the modified noun in gender and number. Due to language contact, IA languages such as Konkani, Marathi or Sambalpuri use both strategies. Refer to the section on language contact in chapter 2. 6.13.1 Participial Constructions These are used very commonly as nominal modifiers. As mentioned above, Dravidian languages are very rich in these, where a long sentence with temporal and spatial adverbs can be participialised and then placed as a nominal modifier. Consider. Telugu 6.128.

niina

gla:su pagalagoÿÿina

yesterday glass

«bba:yi ippuêu vachchiaêu

break hit pst prt boy

now come-pst

The boy who broke the glass yesterday has come now’ Sadari 6.129 r«him ekÿHo bu}Ha-ke «p«n kHa÷ê meâ bHari bojHa leike jat

dekHl«k

Rahim one-cl old-acc self shoulder loc heavy load take go-prs see-pst 3msg ‘Rahim saw an old man carrying a heavy load on his shoulder’. Kurux 6.130 as b«gge ulla nu mal kH«kH«rka lekHa ona? l«ggiyas 3psg many days in neg eat partcip. like eat be prs. ‘he eats as if he has not eaten for many days’

6. Syntax and Semantics

Tamil adjectival clauses can either have a nominative, accusative, dative, allative, instrumental or locative case relationship with the following head noun, which in turn can be relativised as shown by Annamalai (1997:7-8). 6.131 naan saapÿa ele I eat-pst adj leaf “the leaf I ate (on)’ 6.132

avanukku basle vanda koob«m inn«m poog«le he-dat bus-loc come anger

yet

go-neg

(a) the anger that came to him in the bus hasn’t gone yet’ (b) the anger that he came in the bus hasn’t gone yet’ 6.13.2 Relative-Correlative Constructions. In most of the languages that permit these constructions, the modifying clause marked by a J clause may occur to the left of the main clause, or to the right of the main clause after the final verb, or after the head noun (center embedded). Consider some examples from Hindi: 6.133 (a)

maâ ne

[jo

puri b«nai] vo b «hut g«r«m tHi

mother-erg [J ]which poori made that very

hot aux pst fsg

The poori that mother made was very hot’ (b)

vo puri b«hut g«r«m thi [jo maâ ne b«nai] that poori very hot was which mother made

(c )

vo puri [jo maâ ne b«nai ] b«hut g«r«m tHi that poori [which mother made] very hot was

!

Other than by the translation method one can obtain these kinds of constructions very easily in narrative texts. Once you get one type, try to modify it to check whether the other type is also available in the same language. You can relativise any noun, whether in subject or direct or indirect object position. You can relativise the whole sentence and move it to the left or to the right or into the center of the main clause to see various possibilities. The movement of a relativised construction is generally associated with the notion of ‘presupposition’, ‘foregrounding‘ and ‘backgrounding‘ the relevant information.

219

220 A Manual of Linguistic Fieldwork and Structures of Indian Languages We have tried to introduce you to those aspects of syntax and semantics that are interesting and at times peculiar to South Asia languages. As said earlier, this section is not exhaustive in any manner. There are many more interesting aspects of grammar that you would encounter in the field and would perhaps like to describe. The whole area of pragmatics is very wide and can be studied in depth provided a field investigator is prepared to spend extended time in the field. Many aspects of pragmatics are difficult to elicit from the narrative text and observation of various types of discourses becomes almost imperative. Similarly, the sociolinguistic aspects of a language are acquired only through keen observation of the society. We have discussed these issues in the next chapter.

Chapter VII SOCIAL ASPECTS

No grammar is complete unless it tells us about how the language is used as a vehicle for social interaction. The grammar must be explicit about the use of the language in a particular society. After all, language is a view-finder of the social patterns and social behavior of a given society. The linguistic terminology and linguistic structures expose the cultural, social, cognitive and religious aspects of a community. This is the reason why linguists are so concerned about preserving linguistic diversity. When a language dies, along with it dies a whole system of human history of culture and civilisation. When a mother-tongue ceases to be anybody’s tongue it wipes away completely the possibility of reconstructing the past. A history of a civilisation is wiped out from the face of this earth. Linguists such as Nancy Dorian put it very succinctly: There is something inexpressibly sad about watching the disappearance of a unique local language that will never again be heard flowing in its full magnificence from the tongue of a verbally gifted speaker. (Dorian 1999: 39) When we talk of culture and civilisation we have to take into account the language in its totality. Without studying the aspects of language in use or language not in use our knowledge will be incomplete. The question of who speaks what to whom is important, and equally important are the issues why out of 6000 languages spoken on the earth half face impending extinction in the next two hundred years. This section is devoted to the study of both of these aspects, viz. sociolinguistics as well as the sociology of language. An attempt will be made to train you in finding out the vitality of a language and its acceptability in the society it is spoken in. We will first discuss the kinship terms, as they exhibit the composition of the society and the role relationship between various interlocutors. Terms of address and reference, true indicators of proximity or distance between the interlocutors are discussed next. This is followed by a small discussion of verbal politeness strategies used by various Indian languages. What follows this is the very significant aspect of language shift and retention. Lastly, we will touch upon the burning issue of language death. How one can recognise a dying language, and how one should elicit data on the languages that are soon going to be extinct are the concerns of the final section of the chapter. 7.1

Kinship Terms

This is one area where linguists should tread with some caution, as the field of kinship belongs to anthropology. Linguists may find a society a complete mystery to them. The

2 A Manual of Linguistic Fieldwork and Structures of Indian Languages

social organisation and the people living in the society may all be new to them. One of the ways that a linguist can look into the new society is by studying its kinship terms. Language, after all, is a window through which we can see into the society. Remember the discussion under 4.3 when we talked about elicitation. Refer back to table 4.1 on Kurux kinship terms. The intricate hierarchical society of India is reflected in its linguistic patterns and in the various terms that one uses to address and refer to different relations. As we have already seen, grammatical patterns indicate the hierarchical composition of the society (the verb endings in Hindi and Maithili, or the pronominals in Santhali, to name but a few). In addition to the grammar, one can witness the composition, directionality, and intimacy of social behavior of a speech community through kinship terms. Kinship terms are significant on their own as a field of study, but become a useful tool for us when we probe into the system of address and reference that a society uses (we discuss this in the following section). The importance of kinship is best explained by Edwards Jeanette (2000:27). Kinship embraces connections people trace to each other through notions of shared substance, be it blood, genes, flesh, or bone: at the same time it places a greater or lesser emphasis, at different historical moments and in different parts of the world, on the creation and maintenance of social relationships through intimacies of care and effort. This is a potent combination with which to imagine relations of all kinds, not just the family kind. The latest anthropological research around the globe has proved that the concept of kinship with which anthropologists have often worked cannot be universally applied (Schneider 1968, 1984; Handler 1995). Linguists’ interest in kinship, as said earlier, has been restricted to obtaining the linguistic data through which the social organisation can be familiarised. India, despite its presumably common sociocultural makeup, offers very distinct and varying kinship systems and kinship terms. Thus, the northeast presents a very different picture from the Mulukurumbas of the south; Santhals present another different picture from Kurux with whom they are neighbours and, as we have seen, share a lingua franca. Speech communities in the south have one kind of system which they do not share with the Nairs, who are Malayalam speakers and follow a matrilineal pattern. Khasi in the northeast, being a community of both matrilineal and matriarchal nature offers yet another pattern. In other words, the heterogeneity and diversity in this area are very noticeable. This heterogeneity translates itself into various types of forms that interlocutors address each other with or use for reference. Readers are advised to refer to the standard literature in anthropology and linguistic anthropology to have a good grounding on various kinds of kinship systems. For the sake of simplicity, we shall enumerate the various types of relationships in the following paragraphs. As said earlier, the variety and minute differences of each society is capable of giving you a larger list than what is provided here. 7.1.1

Non-Affinal

7. Social Aspects 3

Let us first consider non-affinal relations, those that an ego acquires by birth and not because of entering into a marriage alliance.

4 A Manual of Linguistic Fieldwork and Structures of Indian Languages

Great grand kinsmen Great grand kinsmen are kinsmen more than two generations distant from ego. The choices are: •

Great-grandfather (father’s side)

•

Great-grandmother (father’s side)

•

Great-grandfather ( mother’s side)

•

Great-grandmother (mother’s side)

•

Man’s great-grandchild (male)

•

Man’s great-grandchild (female)

•

Woman’s great-grandchild (male)

•

Woman’s great-grandchild (female)

Grand kinsmen Grand kinsmen are kinsmen more than one generation distant from ego. The possible options are: •

Grandfather (father’s side)

•

Grandmother (father’s side)

•

Grandfather (mother’s side)

•

Grandmother (mother’s side)

•

Man’s grandchild (male)

•

Man’s grandchild (female)

•

Woman’s grandchild (male)

•

Woman’s grandchild (female)

Parents and children •

Father

•

Mother

7. Social Aspects 5

•

Father’s elder brother

•

Father’s middle brother

•

Father’s younger brother

•

Father’s elder sister

•

Father’s middle sister

•

Father’s younger sister

•

Mother’s elder brother

•

Mother’s middle brother

•

Mother’s younger brother

•

Mother’s elder sister

•

Mother’s middle sister

•

Mother’s younger sister

•

Child (male)

•

Child (female)

•

Twins (male)

•

Twins (female)

Siblings and Cousins It is to be noted that societies make distinctions in terms referring to cousins that are children of various brothers and sisters of the parents of the ego. The various possibilities could be: •

Elder brother

•

Middle brother

•

Younger brother

•

The youngest brother

•

Elder sister

•

Middle sister

•

Younger sister

•

The youngest sister

•

Cousin (male, father’s side) (children of father’s brothers and sisters, all hierarchies)

6 A Manual of Linguistic Fieldwork and Structures of Indian Languages

•

Cousin (female, father’s side) (children of father’s brothers and sisters, all hierarchies)

•

Cousin (male, mother’s side) (children of mother’s brothers and sisters, all hierarchies)

•

Cousin (female, mother’s side) (children of mother’s brothers and sisters, all hierarchies)

7.1.2

Affinal Kinsmen

As a parallel to the non-affinal terms mentioned above, affinal terms which are acquired after establishing a marriage relationship pertain to two broad divisions, one from the ego’s husband side and the other from the ego’s wife’s side. Let us add these two terms to our list. One can add the same terms that we have seen above to each of these to get the terms in the affinal category, e.g. ego’s husband’s father, ego’s husband’s mother, ego’s husband’s elder brother, and so on so forth. This will give us 34 additional possibilities (counting from the ‘parents and siblings’). The other relationships are acquired through marriage of ego’s husband’s or ego’s wife’s siblings. The following can be considered. (a)

Own generation Husband Wife Husband’s brother [elder, middle, younger] Husband’s sister [elder, middle, younger] Wife’s brother [elder, middle, younger] Wife’s sister [elder, midle, younger] Son’s wife’s parents Daughter’s husband’s parents

(b)

First ascending generation Spouse’s father Spouse’s mother Father’s [elder, middle, younger] brother’s wife Father’s [elder, middle, younger] sister’s husband Mother’s [elder, middle, younger] brother’s wife

7. Social Aspects 7

Mother’s [elder, middle, younger] sister’s husband (c)

First descending generation Son’s wife Daughter’s husband

!

To elicit data on kinship, the best method is simple observation of the members of the society in their day-to-day lives. One could make a table of all terms referring to all kinds of relationship that exist in a particular society and ask the informants to translate directly. However, our personal experience is that the informant gets lost in such a case. The best recourse is to observe the members of the society interacting amongst themselves. Two types of relations are considered ‘joking’ relationships at least in north India, i.e. ‘sister’s husband’, and ‘wife’s brother or sister’. 7.2.

Forms of Address and Terms of Reference

Language builders are the speakers and the hearers of the society. No language form exists in a vacuum or in a grammar book. Language resides in the form of discourse and the rules of its grammar are thus governed by the social norms and patterns of the society concerned. Although every society has its norms for selecting appropriate forms of address and referral, the hierarchical structure of the Indian society offers a wide and complex range of options to choose from. We shall be concerned with the general norms and principles underlying such choices in this section. Before we proceed we must be clear about the following terms. •

Symmetrical relationship. The relationship in which the S(peaker) and the R(ecipient) of the address have equal status.

•

Asymmetrical relationship. The relationship in which either S or R has a lower status.

•

Social status. This is determined by occupational status, education, and age. The first two are achieved social status and the last is ascribed social status.

•

Intimacy. This is determined by social distance between the dyads as well as by the duration and frequency of contact between the two.

•

Social structure. This is constituted of social status and social relationship between the dyads.

•

Location of the discourse. Pertains to private or public domain. It is of utmost significance that the appropriate form of address is adhered to in discourse. No grammar can be considered complete without incorporating information regarding this area.

8 A Manual of Linguistic Fieldwork and Structures of Indian Languages

7.2.1

Address Forms

Address forms are generally of two kinds: (1) used in introducing a person, and (2) used in direct face-to face conversation. The former in Indian society can either be constituted of direct forms comprising various names and combinations thereof of the member to be introduced or indirect forms comprising phrases indicating the role of the person in the society to be introduced. 7.2.1.1

Direct Forms for Introductions.

Names in Indian society are constituted differently from region to region. In the north of India, the first (given) name precedes the middle name and the surname (the family name). In the south of India the order is reversed. The family name precedes the middle name and the given name may appear last. The name in each of the cases can be preceded by a title name (TN) and followed by an honorific (H). Both are optional, as the occurrence of each is dependent upon the symmetrical and asymmetrical relationship as well as the location of the introduction. In every Indian language there might be several possible forms of TN and H. Consider the Hindi structure with possible options. (TN) + [Full name/First name/Last name] + (H), where TN could be any of the following: Shriman Shrimati Kumar/Kumari Lala Babu The most common (H)onorific is ji used for introducing both female and male strangers. Other possibilities are: sah«b and s«rkar. 7.2.1.2

Indirect Address Forms

These forms are generally used for introducing woman and her relationship to the relevant male member of the society. Consider the following conversation from Hindi. 7.1

S:

in-se

miliye ,

ye hEâ

sh«rma ji ki p«tni

this-instr meet-hon these are Sharma+H gen wife Please meet her, she is the wife of Mr. Sharma’

7. Social Aspects 9

7.2

R:

n«m«ste shrim«ti sh«rma. sh«rma ji kEse hEâ Greetings Mrs. Sharma. Sharma + H how is-hon ‘Greetings Mrs. Sharma. How is (Mr.)Sharma?’

Male members can also be introduced by indirect forms. In this case, the phrase designating the relationship of the member to be introduced to some third party known to the dyad generally precedes the full name. Consider: 7.3

S

in-se miliye,

sharmaji ke

cHoÿe bHai

alok kumar

this-instr. meet, Sharma +H –gen younger brother Alok Kumar ‘Meet Alok Kumar, the younger brother of Mr. Sharma’ 7.2.1.3

Direct Address forms used in face-to-face conversation may be of the following three types:

(a)

Names and various combinations thereof.

(b)

Kin’s address forms

(c)

Second person pronouns

7.2.1.3.1

Names.

As far as the forms in (a) are concerned, various forms and their combinations can be used in asymmetrical relationships, depending upon the degree of formality, intimacy, and location. Consider the following pattern indicating an increasing degree of intimacy as we descend the scale. TN + Full name + (H)

[the longer construction is more formal than the shorter]

TN + Last name + (H) TN + First name + (H) Last name + (H) First name + (H) Professional names are also used instead of the honorific at the end. This is more common if one of the dyad belongs to a lower caste or is professionally considered from a lower group. In a rural setting, generally only the caste names with an optional honorific (depending upon a particular situation) are used. Address words denoting the

10 A Manual of Linguistic Fieldwork and Structures of Indian Languages

designation of the addressee are very common in military life, too. Thus a captain is addressed as ‘Captain’, and a soldier is addressed as ‘Soldier’ in respective languages. In a symmetrical relationship, on the contrary, the following pattern is observed as the degree of intimacy increases. Last name + H > Last name > First name > Pet name (nickname). 7.2.1.3.2

Kin’s Address Forms.

The various forms in (b) can be divided into two groups. The first type are those forms where address forms and reference forms are identical. These are those that belong to non-affinal relations, i.e. relations that one acquires by birth. The second set consists of those forms that distinguish address and reference forms. These pertain to affinal relations, i.e. those relations that one acquires by marriage. Thus, a ‘father-in-law’ (which is a reference term) is addressed by the address form ‘father’ in the society. Interesting forms and variations exist from society to society and thus a fieldworker should be careful studying these. 7.2.1.3.3

Second Person Pronouns.

Most of the Indian languages make a three-way distinction in the 2nd person pronoun as given below in Hindi example. Honorific

Ordinary

Intimate

ap

tum

tu

The honorific and ordinary forms are used on formal occasions in both symmetrical and asymmetrical relationships. Ordinary and intimate forms are used on informal occasions in both symmetrical and asymmetrical relationships. This overlap, which is true of many languages of India, can be demonstrated as such: Formal

Honorific

Informal

Ordinary

Ordinary

Intimate (Abbi 1971, 1985)

In a symmetrical relationship the address form is always reciprocal. Women in India are observed to use intimate forms much more easily than men do. One has to remember that it is the degree of intimacy that governs the choice of the address form.

7. Social Aspects 11

For details refer to Abbi 1985. Other factors that help govern the choice of 2nd person pronoun are education status and the occupational structure of the addressee. Combined with the intimacy factor the following representation holds true of most of the languages of the subcontinent. Each form is defined by a set of distinctive features. Honorific

Ordinary

Intimate

-intimacy

+intimacy

+intimacy

+educational status

+educational status

-educational status

+occupational structure

-occupational structure

-occupational structure

In case the dyad belongs to kin relationship, it is the age factor that governs the choice. 7.2.1.4.1 Indirect Address These word forms in conversation consist of verbal phrases of ‘listening’ type used by married women while addressing their husbands. In many societies in India, married women do not address their husbands by their first or last name. Instead, they use phrases of the following types. As most of these phrases involve the verb ‘to listen’ they are termed ‘listen phrases’. Consider. •

‘Please listen’

•

Are you listening?

•

See a little

•

Listen to me

•

I said

7.2.2

Terms of Reference

A situation that demands an S and an R as well as a common referent that may or may not be present at the scene of conversation is the focus of the ‘terms of reference’. If the referent is not present, s/he can be referred to by the referent’s first name or various combinations thereof, depending upon the social variables that we just considered above, or by the 3rd person pronouns. If the situation is such that the referent is present at the scene of the conversation, the social status of the referent will govern the choice of the personal pronoun. In most of the Indo-Aryan languages the 2nd person pronoun is the preferred choice to show respect to the person concerned. For instance, instead of inhoâ-ne

12 A Manual of Linguistic Fieldwork and Structures of Indian Languages

‘they-erg’ Hindi speakers use ap-ne ‘you-erg’ even when they are not addressing but referring to the person concerned. In short, a combination of social variables, such as age, education, social status and intimacy combined with the location of the discourse will govern the appropriate choice of the form of address and the term of reference in Indian society. The following figures 7.1 and 7.2 summarise the typology of factors governing the choice.

!

As in the case of the kinship terms, one has to be very observant to elicit data in this area. One can ask the informant to tell a personal life story involving various members of the family. This will automatically include many terms of reference for kin. One can also adopt the direct method of asking the informant ‘How do you call so and so?’ Or ‘how does X call so and so?’ It is helpful to make a list of address forms and terms of reference in the contact language and get it translated as the informant can relate to the relationship easily. Thus, rather than asking ‘what would you call your brother’s wife’, I find it easier to ask ‘what do you call your bHabHi?’ (Provided the informant’s contact language is Hindi).

Address Forms

Introduction Direct

Conversation

Indirect Direct

names and their combinations

role relationship

Figure 7.1

Indirect

names kin’s address

‘listen’ phrases II P.Pr.

Typology of Address Forms

role relationship

7. Social Aspects 13

Terms of Reference

Kin

affinal

Friends

Presence of the Non affinal

referent

[names, II/III pr.] Figure7.2 7.3

Absence of the referent

[names, III pr.]

Typology of Terms of Reference

Politeness Strategies

In a society like India, where hierarchical structures play a decisive role in governing social relationships, it is mandatory for speakers to learn the rules of verbal politeness. Paradoxically, studies on politeness strategies have been undertaken more in the western countries (a special mention should be made of Brown and Levinson [1978 and subsequent editions], Grice’s Cooperative Principles [1975], and Leech ‘s Politeness Principle [1980, 1983], Sperber and Wilson ‘Relevance Theory’ [1986] to mention a few from a long list) than in the South Asian countries. We grant that it is an highly relevant phenomenon and politeness has to be somehow communicated, as sentences/discourse lacking such communication may be taken as impolite. Though the concept of ‘politeness’ may differ from society to society, individual to individual and situation to situation, in this section we will attempt to restrict our discussion to the very basics of understanding and eliciting data on those language features that pertain to indicating politeness1. We shall avoid repeating what we already have said in the context of forms of address and terms of reference in earlier sections. Every language offers a wide range of options for showing politeness through its grammatical structure. Ranging from the choice of a specific word to the choice of a 1

Readers are advised to refer to Brown and Levinson (1999 edition: 68-74) for concepts like ‘positive politeness’ and ‘negative politeness’.

14 A Manual of Linguistic Fieldwork and Structures of Indian Languages

specific syntactic structure, all can be used to indicate different degrees of politeness according to the situation. The following are some of the linguistic devices that language speakers adopt for the purpose. 7.3.1

Lexical

1.1

Greeting words

1.2

Honorifics

1.3

Personal pronouns

1.4

Choice of specific lexical items

1.5

Kin terms

1.6

Title names

Out of all these the fieldworker must pay special attention to the use of personal pronouns. Although the use of plural pronouns in general is considered more polite than the singular ones such as use of ‘we’ instead of ‘I’ for a singular referent, interchange of pronouns also indicates politeness. For instance, a change of third person pronoun to second person pronoun, even in a situation of reference and not address (see 7.2.2) should be taken note of. Andamanese also offers a rich pronominal system indicating degrees of honorific. Consider the following table: Table 7.1 Pronominal System of Andamanese Singular

Plural

ÿIyo

miyo

ÿeNNiyo

meNNiyo

Second person

Niyo

Niliyo

Second person

Namale

Namale

o

niyo

«mma:ye [m], «mmimi [f]

niliyo

First person First person (inclusive)

Honorific Third person Third person Honorific

(Adapted from Manoharan 1997: 465)

7. Social Aspects 15

7.3.2

Prosodic Features

7.3.3

2.1

Tone/intonation

2.2

Pause

2.3

Length of the vowel

2.4

Stress2

Syntactic Features 3.1

Use of subjunctive and optative

3.2

Tags

3.3

Repetition of the last part of the phrase or of intensifiers

3.4

Word order or ordering of phrases

3.5

Future tense in imperative constructions

3.6

Agreement [e.g. the sentence from Maithili given below]

Politeness has many syntactic manifestations, and one of the most intriguing is the last one mentioned above. Verb agreement in Indian languages alone indicates whether the addressee and the referent are shown any respect or not. The simple word ‘do’ (imperative) may have three forms in Hindi k«r, k«ro, and kijiye, depending on who the request is made to and how much respect is being shown. Consider the Maithili sentence where the verb is not only coded for subject, direct object and indirect object, but also for honorific degrees (non-honorific, mid-honorific, high-honorific) assigned to each of the arguments under consideration (Bickel, Bisang and Yadav 1999:482). This is an extreme case of the multiple verb agreement system (the capital letter N stands for nominative case). 7.4

h«m tora:

k«niya:-ke dekH-au-l-i -au-nH

I-N 2nh/mh dat bride-dat

see-caus-pst-1N-2nh/mh-non N-3h non N

‘I showed you nh/mh the bride h’. 7.3.4.

2

Sociolinguistic Features 1.1

Code switching

1.2

Dialectal switching

Consult ‘Accommodation theory’ (Giles 1980, 1984; Trudgill 1981; Coupland 1980; Street and Giles 1982).

16 A Manual of Linguistic Fieldwork and Structures of Indian Languages

1.3

Hedging

1.4

Inclusion of the speaker [e.g. ‘let us do it’ rather than ‘you do it’]

1.5

Personal-center switch [used with tags]

!

While discussing politeness, one must remember to study the speech of men and women. In many societies women have different linguistic rules from men for observing politeness. For instance, male speakers of Hindi and its related dialects use the term yaar ‘buddy’ at the end of a sentence to show informality and politeness in a symmetrical relationship, while women speakers do not use such words. We must distinguish and ascertain behavior in same-sex dyads and cross-sex dyads to study differences. 7.4

Language Shift, Language Retention and Language Death

The issues of language shift and obsolescence have gripped the minds of linguists, anthropologists, sociologists, and geneticists for the last fifty years, as we all believe that diversity of languages is essential to human evolution. We find it absolutely necessary to investigate the status of a linguistic community regarding the vitality of the language concerned (see Kindell and Lewis 2000). The following excerpt from Dorian (1999: 3334) is very touching and we would like to share it with the readers. “...But an ethnic language once lost is far less easily recovered than other identity markers, and the cultural content that the ethnic language carried is never fully recoverable. Fieldworkers confronted, with this sort of situation can either bow to what seems to be the inevitable, taking the loss of the ethnic language to represent “the will of the people,” or they can resist and become advocates for the ethnic language, a tricky stance for an outsider to take”. We leave the choice of becoming advocates for the revival of the language under investigation to individual fieldworker concerned. However, our experience is that prolonged fieldwork in a specific area will automatically draw you to the greater task of preserving and helping the community revive the language. Let us be concerned for the present how one should study the vitality of a given language. Linguists such as Thomason and Kaufman (1988) have dealt extensively with the phenomena of shift and maintenance and readers are advised to refer to their work for a working model.

!

Collect data from as old a speaker as you have access to and then from a teenager to observe the shifting patterns, attitudes and fluency in the language concerned. Fieldworkers should be concerned with these questions. Are the speakers of a particular

7. Social Aspects 17

language shifting to another language? Are their some speakers left still speaking a particular language while the majority of them speak some other, preferably the dominant language of the region? Are speakers becoming bi- and multilingual and thus maintaining their heritage language? If the community is multilingual, what are the domains of use of various languages? What is the status of each language that is used in various domains? What is the attitude of the speakers towards their mother-tongue and towards other languages that they use? What are the impending signs of language obsolescence and death? To answer all these questions, a fieldworker has to proceed very systematically with great care and preparation. One has to be observant as well as be prepared with a set of questions (see the sample questionnaire given in the appendix) that one asks different members of the society concerned. 7.4.1 Language Shift and Retention. In a multilingual milieu such as India it can at times be trying to identify language shift. People speak different languages and dialects with varying degrees of competence. Language shift is more common among urban migrant communities or among minority communities than among rural populations or among majority language speakers. Other than the methodology followed by the sociologists (Ishtiaq 1999) which is dependent on head counts and is based on a census type of enumeration, linguists have evolved their own techniques to elicit data on language shift. The reason why direct questions as ‘what is your mother-tongue? or ‘what do you speak at home?’ are avoided in a minority community is because speakers in general want to associate with the majority group and therefore never give the true picture. Our experience with Kurux and Kharia (Abbi 1992) reveals that despite the fact that these languages were spoken at home there were few people claiming to do so. Several linguists have been studied the factors and variables associated with language shift (Giles, Bourhis and Taylor 1977, Harwood, Giles and Bourhis 1994, Paulston 1994, Fishman 1991, and Miller 2000, to mention but a few). However, although there is no set list of factors involving language shift, as much depends upon the society and the people, there is a sort of general agreement on some selected factors, the study of which may help identify language shift. Field linguists should test the following to determine language shift. This may be done either by observing or by directly asking questions, or better by combining both of these methods. 1.

The domain of use of the language in question could be the following: •

Home

•

Work

•

Market

•

Other public places

•

In interpersonal relations

18 A Manual of Linguistic Fieldwork and Structures of Indian Languages

•

In religious discourse

•

In education

•

In media

2.

Attitudes towards the language and its speakers: Positive and negative.

3.

Linguistic material: Basic word list, basic sentences, code switching, extent of lexical and syntactic borrowing.

7.4.1.1

Domains of Use.

A detailed questionnaire should be prepared to inquire about the various domains where a language is used. An intensive survey should be carried out about who speaks what to whom. Thus, while framing questions regarding language used at home and in interpersonal discourse, data has to be elicited about the use of language with all types of dyads of various hierarchical ranks. Some of the sample questions could be: •

In what language do you speak with your parents/grandparents/children? (for each dyadic relation a separate question)

•

In what language do your parents/ grandparents/children respond to you? (for each dyadic relation a separate question)

•

In what language do your mother and father speak to each other?

•

In what language do your grandmother and grandfather speak to each other?

•

In what language do you speak to your servants?

•

In what language do you speak to your friends?

•

In what language do you speak in family gatherings?

•

In what language do you speak with members of the society other than your own family?

As said earlier, more than asking such direct questions you also have to be observant about the linguistic behavior of the community, as many times you might find variation between the responses of the informant and the actual linguistic behavior.

7. Social Aspects 19

7.4.1.2

The Attitudes

The attitudes of the speaker towards the language and its speakers can be gauged by direct questions and disguised questions3. The latter almost invariably give correct reading. Positive attitudes signal language retention tendencies and negative attitudes signal language shift tendencies. Opinions about the language’s attributes such as ‘rich’, ‘sweet’, ‘harsh’, ‘powerful’, ‘prestigious’, ‘complex’, ‘useful’ etc should be collected. This will give you an indication about the degree of attachment that the speaker has towards her/his mother tongue. Opinions should also be sought about the speakers of the language concerned by testing attributes such as ‘polite’, ‘friendly’, ‘cultured’, ‘stupid’ ‘educated’, ‘aggressive’, ‘honest’, ‘progressive’, ‘miserly’, ‘generous’ etc. Disguised questions may look nonsensical, but they give us the true feelings of the speaker. A typical such question could be ‘If you had a choice what would you choose as a mother-tongue?’ or ‘if all could speak one language, which one do you think they would speak?’ A recently conducted pilot survey (Abbi et al 2000)4 of India gave us very accurate data about the attitude towards Hindi as a panIndian language by the speakers of non-Hindi and Hindi alike. Sociolinguistic research, unlike pure linguistic research, warrants a different type of questionnaire and a different type of methodology for gathering the data. Many of the questions such as those pertaining to attitudes or beliefs cannot be asked in a binary query. Generally, a five-value scale (fig. 7.3) such as the following is used by sociolinguists (Karan and Stadler 2000:195). Numbers are assigned to each of the values to give a comparative profile of the issue under investigation. Strongly

disagree

disagree

uncertain

1

2

3

Fig 7.3 7.4.1.3

Strongly

agree

agree

4

5

A five-value scale used in sociolinguistic research

Linguistic Material.

As far as the linguistic aspects are concerned, one should begin with the basic word list. Depleted use of basic words is a good sign of language shift and ultimately of language death. The process of borrowing words and phrases from the contact or 3

Readers are advised to refer to Karan and Stadler (2000) for the data-gathering techniques discussed in this section. 4 Abbi, A., R.S.Gupta and R. Gargesh 2000.

20 A Manual of Linguistic Fieldwork and Structures of Indian Languages

dominant language begins with the pattern of code-switching leading to complete take-over of the sentence structure of the target language. Divide the basic word list into the following groups and then test how many words from each of the sections have been retained. •

Numerals: the first three

•

Kinship terms: other than for ‘mother’ and ‘father’

•

Body parts

•

Objects of nature

•

Temporal organisation

•

Adjectives

•

Verbs

•

Adverbs

•

Specific household artifacts

•

Flora and fauna

•

Items of adornment used by women

A recent study of Laria by Naik (1999) shows only 6% retention of basic words in the language. Despite this decline, the language is not on the verge of extinction. On the contrary, it shows tendencies of survival because of its strong use in the home domain. In addition, the positive attitude of Laria speakers towards the language and its speech community will help sustain it. As said earlier, lexical borrowing is one of the ways of diagnosis of language death, but not all. In fact, there never is only one reason of language obsolescence, but a combination of various linguistic, social, cultural and psychological reasons that lead to language shift, and consequently to language obsolescence. Lexical borrowing results in a change in the phonological system of the language followed by a borrowing of the syntactic pattern. This second stage of borrowing is most often the result of communities turning into 100% compound bilinguals (e.g. Bangani, Laria, and urban tribal speakers in Jharkhand, to mention but a few). The last(?) component to be borrowed are the inflectional categories and this has been attested in the case of many urban speakers of tribal and minority languages. The linguistic factors, though important, do not serve as a complete guide for judging language shift. It is an important fact, as my experience in Jharkhand has shown, that languages such as Kharia and Kurux cannot be considered dying languages despite having heavily borrowed from Hindi in every realm of grammar including the inflectional endings (e.g. gender agreement). The attitudes of the speakers as well as the domain of use play a very major

7. Social Aspects 21

role in language retention. If urban Kurux shows tendencies of language death, it is because of these two factors i.e. reduction in domain of use and the negative attitude of the Kurux speakers towards their own mother-tongue, and not only because of the linguistic factor. As we mentioned in chapter 2, an urban Kurux feels it is prestigious to forget his mother-tongue and gets punished by his own people for using the language. Let us consider the question of language death in some detail. 7.4.2

Language Death

Language attrition may be due to linguistic and non-linguistic factors. Dorian (1989) has given some focus questions to make an assessment of language obsolescence, which we recommend our readers.5 Again, it is the combination of linguistic and non-linguistic factors that play a role in language death. A fieldworker working on language attrition/death phenomena has to be concerned with the following basic issues: 1.

Demographic figures of the speech community

2.

The degree of language shift

3.

Motivational factors for retaining/losing the language

4.

Strength of the support system

Let us consider each of them briefly. First, depleting population figures of a particular speech community are definitely of concern and are a sure sign of language attrition. The Andamanese and Nicobarese are communities who show a consistent decrease in population size, and thus these may be considered dying languages6. Demographic factors in general include the number of members of a speech community, their distribution, concentration, proportion, birthrate, and patterns of immigration and emigration. A large-scale migration of a community from its home state to another state may lead to language attrition, though Indian communities never came under this threat, despite large movements from one state to another. Language maintenance is the hallmark of Indian migrants. Secondly, the three major factors considered in 7.4.1 are important for identifying the degree of language shift. Thirdly, motivational factors play a significant role in deciding language attrition. Motivational factors include the attitude of the speakers towards their own language as 5

Naik (1999) adopted these questions to judge the status of Laria (spoken in Orissa) and reached very conclusive results. 6 The government of India does not publish the information on number of speakers, but a rough estimate is that there is no more than a couple of hundred Andamanese left in the Island.

22 A Manual of Linguistic Fieldwork and Structures of Indian Languages

well as the attitude of the speakers of the dominant language towards the speakers of the minority language. If the latter is negative, i.e. if dominant language speakers look down upon the speakers of the minority language, there is little chance of survival of the latter. At this point a speaker’s claim (even if it is untrue) that s/he does not speak the language is important. The inferiority complex towards your own language paves the path to language attrition sooner or later. Finally, various support factors involve the extent to which the group receives support for the language in both formal and informal institutions such as the home, at school, in government, business, and society, and so on. The language that enjoys the support of all three support institutions, i.e. personal, community and institutional, certainly has a better and longer life span. For the survival of any language it is necessary that it be used in the home environment and for inter-personal communication among the community members. This has been the sole reason why a large number of tribal and small community languages have survived in South Asia despite group bilingualism, large-scale migration, dominance of majority languages, marginalisation of minority communities and their languages and lack of formal institutional support.. The majority of the minority languages and most of the tribal languages are still spoken at home.

!

Fieldworkers are expected refer to the appropriate questionnaire given in the appendix. The fine-graded divisions under each of the rubrics considered above are given in the questionnaire and should be investigated. One has to make some contact with the local people before one starts investigating the questions of language death and language maintenance. Do not forget to include very old and very young people among your informants. Do not forget to interview women and be prepared to find a large range of variation between the responses of women and men. Do not expect that people will readily volunteer information on the questions asked here, as the issues are sensitive and people can be touchy. One needs to be extra-careful when conducting fieldwork on the topic of language death and shift, as what you observe may not tally with what you elicit, especially if you are working with marginalised communities. BEST OF LUCK

Appendix 1

Bilabial vl vd

Labiodental vl vd

Dental vl vd

Alveolar vl vd

Retroflex vl vd

Alveopalatal vl vd

Palatal

Velar

vl vd

vl vd

Stops -asp +asp Fricatives Affricates -asp +asp Nasals Flaps Laterals Approx.

Blank chart for Consonants

FRONT

CENTRAL

BACK

UNROUNDED

UNROUNDED

+/- ROUNDED

High Lower high Higher mid Mean mid Higher low Low

Blank chart for Vowels

Glottal

Appendix 2 The Basic Word List [A] This list includes words from Swadesh’s list (1955), and from Gudschinsky’s (1956) list. The words are alphabetically arranged by Samarin (1967). 1. all 2. and 3. animal 4. ashes 5. at 6. back 7. bad 8. bark 9. because 10. belly 11. big 12. bird 13. bite 14. black 15. blood 16. blow 17. bone 18. breast 19. breathe 20. burn 21. child 22. claw 23. cloud 24. cold 25. come 26. count 27. cut 28. day 29. die 30. dig 31. dirty 32. dog 33. drink 34. dry 35. dull 36. dust 37. ear 38. earth

39. eat 40. egg 41. eye 42. fall 43. far 44. fat/grease 45. father 46. fear 47. feather 48. few 49. fight 50. fire 51. fish 52. five 53. float 54. flow 55. flower 56. fly 57. fog 58. foot 59. four 60. freeze 61. fruit 62. full 63. give 64. good 65. grass 66. green 67. guts 68. hair 69. hand 70. he 71. head 72. hear 73. heart 74. heavy 75. here 76. hit

77. hold/take 78. horn 79. how 80. hunt 81. husband 82. I 83. ice 84. if 85. in 86. kill 87. knee 88. know 89. lake 90. laugh 91. leaf 92. leftside 93. leg 94. lie (i.e. be in lying position) 95. live 96. liver 97. long 98. louse 99. man/male 100. many 101. meat/flesh 102. moon 103. mother 104. mountain 105. mouth 106. name 107. narrow 108. near 109. neck 110. new 111. night 112. nose 113. not

114. old 115. one 116. other 117. person 118. play 119. pull 120. push 121. rain 122. red 123. right/correct 124. rightside 125. river 126. road 127. root 128. rope 129. rotten 130. round 131. rub 132. salt 133. sand 134. say 135. scratch 136. sea 137. see 138. seed 139. sew 140. sharp 141. short 142. sing 143. sit 144. skin 145. sky 146. sleep 147. small 148. smell 149. smoke 150. smooth

151. snake 152. snow 153. some 154. spit 155. split 156. squeeze 157. stab/pierce 158. stand 159. star 160. stick 161. stone 162. straight 163. suck 164. sun 165. swell 166. swim 167. tail 168. that 169. there 170. they 171. thick 172. thin 173. think 174. this 175. thou 176. three 177. throw 178. tie 179. tongue 180. tooth 181. tree 182. turn 183. two 184. vomit 185. walk 186. warm 187. wash

188. water 189. we 190. wet 191. what 192. when 193. where 194. white 195. who 196. wide 197. wife 198. wind 199. wing 200. wipe 201. with 202. woman 203. woods 204. worm 205. ye 206. year The following words are added by Gudschinsky (1956). 207. brother 208. clothing 209. cook 210. dance 211. eight 212. hundred 213. seven 214. shoot 215. sister 216. spear 217. twenty 218. work

Appendix 3 The 300 Basic Word List [B] (for the Indian context )

The following words are added to the basic word list [A] * Refers to the typical North Indian concepts 1. *bindi 2. *flour (kneaded) 3. *jura (bun) 4. *roÿi 5. banana 6. bangles 7. blouse 8. book 9. brother, elder 10. brother’s wife 11. cat 12. chilli 13. cold (ailment) 14. comb 15. copper 16. cough 17. cow 18. crow 19. cry 20. dog 21. door/entrance 22. down 23. drizzle 24. earring 25. elephant 26. eye-brow 27. fever 28. finger 29. fish 30. flour (dry) 31. food 32. forest 33. garlic 34. God

35. Goddess 36. gold 37. grand father (F) 38. grand father (M) 39. grand mother (F) 40. grand mother (M) 41. green vegetable 42. hot 43. house 44. house fly 45. itch 46. language 47. lion 48. lips 49. liquor 50. mango 51. medicine 52. milk 53. mirror 54. money 55. mosquito 56. mouse 57. nails 58. necklace 59. nose-pin 60. oil 61. onion 62. pain 63. peacock 64. plait 65. pond 66. potato 67. rice (cooked) 68. rice (raw)

69. road 70. round 71. run 72. sari 73. shawl 74. shirt 75. sister, elder 76. sister’s husband 77. small 78. snake 79. spectacles 80. spices 81. spoon 82. sugar 83. tasty 84. tea 85. teeth 86. thumb 87. turmeric 88. up 89. village

Appendix 4 The 400 Basic Word List [C] * Refers to the typical North Indian concepts. 90. bel«n (rolling pin) 91. *c«klaa (rolling board) 92. *charpoy 93. *dal 94. *k«}Haii 95. *neem 96. *t«wa 97. abuse 98. air 99. ant 100. aroma 101. bald 102. bathe [vcaus] 103. bathe [vi] 104. bathe [vt] 105. bazaar 106. begin 107. behind 108. blind 109. brass 110. brave 111. bull 112. butterfly 113. buy 114. cheap 115. coconut 116. corpse 117. costly 118. cry of x 119. curse 120. daily 121. danger 122. deaf 123. donkey

124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157.

double draught earn end enemy farmer feed flood front of goat gold smith halves hard heels her high his honey horn hunger ill iron smith king/chief land lord lazy letter lonely love mad maize memory mine monkey paint

158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189.

pig soft price priest read remember rice (crushed) rice (husk) rice (paddy) rice (puffed) sell sheep shop silver sometimes sparrow sugar cane sweet potato tailor teach teacher thirst tongs train upside down war weave weaver well wheat worship write yours

Appendix 5 The Basic Sentences for Collecting Field Data You might find some of these sentences funny or ungrammatical, however, when translated into the contact language and the target language they will appear all right. SIMPLE 1. Ram is eating a mango. 2. Sita is eating a ripe mango. 3. Ram cut the mango. 4. The children cut the mango with a knife. 5. Rizwan cut his fingers while cutting the mango. 6. Ruth ate the mango in the bus. 7. Salma ate the mango in the morning. 8. The child ate up all the sweets. NEGATIVES 1. I don't go to school. 2. We will not go to Jaipur tomorrow. 3. 1 did not go to school yesterday. 4. The child did not hit his/her sister. 5. Because they did not study they failed in the exams. 6. Neither my mother came nor my sister. 7. The old elephant did not die. Note: check the scope of negation. IMPERATIVES 1. Come in. 2. Don't come in. 3. Please sit down. 4. Sit/sit down. 5. Admission is prohibited. 6. Get lost! CONDITIONAL AND COORDINATION 1. If you don't give me the sweets, I will cry. 2. If you will hit me I will cry. 3. If he had come earlier he would have seen the letter. 4: If he asks me 1 will certainly tell the whole story. 5. If he can come then he should. 6. Ram and Sita both went to watch the movie.

7. Children ate and drank cold water. 8. Mary is beautiful but ill natured. 9. Sohail will help but not Reshma. INCLUSIVE/EXCLUSIVE 1. Yesterday we had gone to see the movie (-addressee) 2. Yesterday we had gone to see the movie (+addressee) 3. Let us go now (-addressee) 4. Let us go now (+addressee) 5. We got punishment, how bad! (-addressee) 6. We got punishment, how bad! (+addressee) Note: check for the dual number also INTERROGATIVES 1. What is your name? 2. Where do you stay? 3. How are you? 4. When are you going home? 5. Why are you going to Delhi? 6. Who is he/she? 7. How much did you get? 8. Have you eaten? 9. Shall I come tomorrow? 10. Did you see the papers today? 11. Will you be cooking today at home? 12. What do you think he was thinking? 13. Which one is your brother? RELATIVIZATION, PARTICIPIALIZATION, ADJECTIVES 1. The boy who had fever died yesterday. 2. Call the boy who broke the glass. 3. Call the boy who broke the glass day before yesterday. 4. Throw away the broken branch. 5. The tail cut monkey was a nuisance. 6. The cap which was hung on the nail, fell. 7. The fevered boy died. 8. Call the glass-breaker boy. 9. Call the tea-drinker boy. 10. The hung cup fell down. 11. The innocent child. 12. The child is innocent.

CAUSATIVES, PASSIVES/ INCAPABILITATIVES 1. Mother fed the baby. 2. Mother made the ayaa [nurse] feed the baby. 3. Father got all the papers thrown away by the servant. 4. The girl is getting the varandah cleaned by the maid. 5. Shila asked Ratna to make Sita rise. 6. The teacher made the child study. 7. The forest officer is making the wood-cutter cut the trees. 8. Trees were cut. 9. The sister is making the little brother sleep. 10. Rajiv Gandhi was killed by a bomb. 11. If he had made him study computers he would have got the job by now. 12. Alas, my parents had made me study science! 13. The bottle broke (by me). 14. Rice is eaten in the Southern India. 15. A girl was raped last night. 16. I was beaten furiously by the goondas/hooligans. 17.The food cannot be eaten fast by the child. 18.Because of cold, writing cannot be done by me. (Hindi: likH aa n«hiâ jata) CASES, AGREEMENT, COINDEXING 1. The boy is playing. 2. The boy is playing with a ball. 3. The boy ate a banana. 4. The boy played well. 5. The girl had eaten the meals. 6. Mother will cook now. 7. I am sad. 8. Ram is hungry/thirsty/in pain. 9. Pick up the book and keep it on the table. 10. Give the horse the feed. 11. Wash the clothes please. 12. Ram wrote a letter to his mother yesterday. 13. Curd is made from milk. 14. Ravan fought with Ram. 15. Hanuman burnt Lanka with his tail. 16. The leaves fall from trees in autumn. 17. Give me some money for my daughter. 18. Nobody is at home. 19. I bought everything in ten rupees. 20. Flowers are blooming in the garden. 21. The book is on the table. 22. Clothes are kept on the top of the box. 23. Books are kept at the bottom of the box. 24. There is a garden behind my house. 25. There is a mango tree in front of my house. 26. Sister will go shopping only with her friends. 27. Ram's sister's wedding is tomorrow.

28. Ram's cap is blue. 29. Buy twenty-five rupees’ sugar [i.e. for 25 Rupees.]. 30. Don't turn the book's pages. 31. Before 1 went to her house I changed my clothes. 32. After coming back from the movie 1 went to sleep. COMPARATIVE/CONTRASTIVE 1. River water is cleaner than lake water. 1 The Ganga water is the cleanest. 3. Bombay weather is wet but Delhi weather is dry. 4. India is corrupt but Japan is not. 5. My uniform is whiter than his uniform. ECHO-FORMATIONS 1. Please have some tea, etc. 2. He has gone to buy some stationary, etc. 3. He got some chairs, etc. on his marriage. 4. Sit, etc. (imperative) [try this with formal, informal and intimate forms.] 5. I cannot walk, etc, (Hindi: mujH se c«la n«hiâ jata) 6. Why do you get angry, etc? 7. Children go to study carrying bag, etc. 8. The throat does not go bad by singing, etc. CLASSIFIERS 1. Two books were stolen. 2. Longish bamboo was lying there. 3. Round moon. 4. One girl/boy. REDUPLICATION 1. He was walking slowly slowly. 2. He came again and again. 3. What all did he eat? 4. Who all came to the party? 5. When when (how often) will you go to Ranchi? 6. Where all Ram found Sita's ornaments? 7. He got tired (of) sitting sitting. 8. I got bored writing letters. 9. She spoke while eating. 10. The child fell down while walking on the footpath. 11. While/As he was watching, the thieves ran away with the cash. 12. Give me hot hot coffee. 13. Red red apples are juicy. 14. Bring sweet sweet mangoes. 15. How are you going to jump over these high high hedges?

16. Give me little little of every variety of sweets. 17. Don't come after me (Hindi: picHe picHe) 18. Sometime or the other (Hindi: k«bHi na k«bHi ) he will come back to me. COMPOUND VERBS 1. The old elephant died. 2. By the time papa came home his friend had left. 3. Give me the letter. 4. Get out! 5. Please knit a sweater [for me]. 6. Please knit a sweater [for yourself]. 7. See Lakshmi sing! 8. He sat down. CONJUNCT VERBS/ CONVERBS 1. Having eaten his meal papa went out for a stroll. 2, Read the letter loud (Hindi: p«}H k«r sunao). 3. See the house properly (Hindi: gHum k«r). 4. He did not do well having cheated his own brother. 5. Go and see (Hindi: ja k«r dekHo). 6. He hit me and ran away. INFINITIVES, COMPLEMENTS 1. I don't like your coming here. 2. Who does not like to eat well? 3. Because of his coming back his mother could survive. 4. He told me that he was leaving the town soon. 5. She said that she likes to cook. 6. My sister told me that it rained heavily last night. Total number of sentences 160

Appendix 6 The Case Frame Matrix

A Basic Verb Types State [+Ps] Process [+P] Action [+A] Action Process [+A, P]

B +Experiencer State Experiencer [+Ps, E] Process Experiencer [P, E] Action Experiencer [+A, E] Action Process Experiencer [+A,P,E]

C +Benefactive State Benefactive [+Ps, B] Process Benefactive [P, B] Action Benefactive [+A, B]

D +Locative State Locative [+Ps, L] Process Locative [P, L]

Action Process Benefactive [+A,P,B]

Action Process Locative [+A,P,L]

Action Locative [+A, L]

Adapted from the Case Frame Matrix suggested by W. A. Cook. All Os (for object) has been changed to P (for patient).

Appendix 7 Word order/ Topic and Focus/ Scrambling As word order is comparatively free in a majority of languages, try to move elements within the sentence to test topic and focus and the grammaticality of the scrambled sentences. Some sentences are given in Hindi using phonetic transcription so that the English translation does not elude the characteristic features. 1. She told me that the story was true. 2. I do not know whether he will come to my house. 3. Rani hated the boy who loved Mona. 4. It is easy for Ram to read fat books. [ ram moÿI moÿI kitabeâ asani se p«}H leta hE] 5. She is not a beautiful girl. 6. There are no trees in this jungle. 7. No this is not so as you think. 8. No, I don’t agree that no one is there. 9. Why did you not come? 10. You are not a fool! 11. Who did you show the film and to whom did you give the money? 12. I showed my friend (a) film. 13. I will go home but after finishing the job. [try this with emphatic particle] 14. No doubt it was a small house but was beautiful. [tHa to vo m«kan cHoÿa par sund«r tHa] 15. Every boy is not loved by his sister. [h«r l«}ke ko uski b«h«n nahiâ pyar k«rti] 16. I like the film that has no violence. [try scrambling with every constituent] 17. Book, I will take back. [kitab to mEâ le jauâgi] 18. Running water is never dirty. [try to form participle construction] 19. A man who was wearing a turban barged into our house last night. 20. I think he is mad. 21. He said that exams were not held today. 22. I wanted the exams to be over today. 23. Ravii saw Manojj coming while hei/j was crossing the road. 24. He wondered why were they fighting 25. I don’t think that he will help the organization.

Appendix 8 Interrogation 1. Do you know what did he do? 2. Who do you think will read the book? 3. Who will win the race? 4. How do you know that Ram is coming? 5. Do you know who will sing? 6. When can you tell if Ram is coming? 7. Which movie do you think is good? 8. Do you know whether he did it? 9. Why do you think whether Sita will eat? 10. Why don’t you think whether or not Sita will eat? 11. Ram ate what? [echo question] 12. Ram is going home? 13. Is that girl beautiful? 14. Where will you go in the middle of the night? 15. Who is that? 16. Will you buy anything? 17. What will you buy? 18. Why should I not go to Calcutta? 19. How will you go? 20. Who ate what? [Try it with reduplication: kis kis ne kya kya kHaya?] 21. Why do you think that Ram is angry? 22. What matter do you know is disturbing him? 23. Do you know whom you hit? 24. Could he sleep last night after all that noise? 25. Who knows who did what? 26. You want to go home, isn’t it? 27. He won’t call, would he?

Appendix 9 Explicator Compound Verbs It is difficult to render these sentences in English. Hindi equivalents are given within brackets and the English sentences are mere approximation of the sentences you would find in the Indian languages. Sample sentences 24-28 are common only in Dravidian languages. Please be seated! [bEÿH jaiye] He ate up all the sweets. He did not eat all the sweets. I will drink up all the milk. Close the door on your way out. Have you had your tea? Oh! What have I done? Beant Singh killed Indira Gandhi Ram had entrusted Sita with 500 rupees. [ram ne sita ko p«cas rupaye de r«kHe tHe] 10. Govind has seen the cat. 11. Rakesh read the entire book in one sitting. [ rakes& sari kitab ek bar meâ p«}H g«ya] 12. Rakesh read the entire book. [rakes& ne sari kitab p«}H li] 13. I went away. 14. Ram Lal will surely get your work done. 15. I am unable to stand.[mujHse kH«}a n«hiâ hua jata] 16. Walk carefully! [dekH k«r c«la k«ro] 17. Let me go home. 18. He left (abruptly). [vo c«l diya] 19. Will you knit a sweater for me? 20. Please read the letter (aloud). 21. Please read the letter. 22. Ram was about to reach home that the police caught him. 23. Wait here until he gives you the letter. 24. Bhaskaran wrote a letter to Raman 25. Bhaskaran wrote a letter for Raman 26. Just watch her sing! (contempt) 27. I welcome you (in all humility) 28. Well before the interview Mohan went and looked up the office. 1. 2. 3. 4. 5. 6. 7. 8. 9.

Appendix 10 Dative Subjects 1. 2. 3. 4. 5. 6. 7. 8.

I am sad/happy I have cold/cough/ fever/ pain [record separate answer for each]. Children like to play in the ground. Mr. Lal is in the habit of smoking (he smokes). Priya wants books from the store. As soon as I entered the garden I could smell the fragrance of rose. I am (feeling) tired. They have been sitting at the bus stop for two hours.[unheâ y«haâ bEtHe hue do gH«÷ÿe ho g«ye hEâ]. 9. I feel shy. 10. He got angry. 11. You will have to come tomorrow. 12. I have to go home now. 13. She caught cold in the trip. 14. Man has two hands and two feet. 15. My sister has three children. 16. My father has plenty of money. 17. I got a letter from an old friend. 18. She found a coin on the road. 19. Mary had a baby boy. 20. Ahmed has courage. 21. Then sense came to Farida and she said. 22. The house has two doors. 23. The plant sprouted branches. 24. After months of massaging the child became beautiful. 25. Abdul has a large family.

Appendix 11 Language Shift and Retention After the basic preliminary questions regarding name, age, education level, occupation, parents’ occupation, marital status one can proceed with the following inquiries. It is to be noted that you are interviewing speakers of the minority community of the region. 1. 2. 3. 4. 5. 6. 7.

How long have you lived here? [5-10, 11-24, over 25 years, always ] Place and reason for migration, if any. What do you speak at home? Or what is your native language? What is your parents’ mother tongue? What is your grandparents’ mother tongue? How many languages do you know? [understand, speak, read, write] How would you rate the competence in your own mother tongue? • Fluent • OK but not fluent • Have difficulty in speaking but not understanding 8. How and where did you learn the contact language/ dominant language? [in school, language institutions, with friends, in the market, play ground] 9. What language do you use at home? [record separate answer for each]

•

With parents/elder brother/ elder sister/younger brother/younger sister/grandparents/wife/husband/ servants/pets • With relatives more educated than you/less educated than you • For discussing professional matters • When you are loving/affectionate • When you are angry • For discussing political matters 10.In which language do you do the following? [record separate answer for each] • Think /Pray/ Curse/ Joke /Scold/ Sing /Count 11. What do you call your (both address and reference forms)? [record separate answer for each] • Father/ mother/ elder brother/ elder sister/parents of your wife/husband 12. What language do you use with your doctor? 13. What language do you use for asking time or directions from • Acquaintance (male/ female) • Stranger (male/female) 14. What language do you use with neighbors who are • Linguistically similar • Linguistically dissimilar 15. What language do you use in market place/ post office? 16. What is the medium of instruction of the school your children go to? 17. Given a choice, which language would you like to educate your children in and why? 18. Do you encourage the younger generation to speak your native language? 19. If you had a choice, what would you choose as your mother tongue? Why?

20. Your mother tongue is useful/suitable for………..[let the informant complete] 21. Do you feel prestigious/embarrassed/neutral to speak in your mother tongue in the presence of the speakers of the dominant language? 22. Have you ever had any problem as a result of being the native speaker of your mother tongue? Yes/No. If yes, what kind? 23. What is the language generally used in community gatherings? 24. What is the language generally used in community prayer? 25. Compared to the dominant language spoken here how would you rate your own mother tongue? Answer in yes or no. • Rich • Sweet • Harsh • Powerless • Non-prestigious • Useful

Appendix 12 Language Death In addition to the questions listed in appendix 10 one must collect demographic figures of the members of the speech community, the basic word list, numerals (first three), kinship terms, body parts and adornment items used by women. Degree of language shift, one of the most important factors to decide language death will be obtained by answers collected in the appendix 10.

Appendix 13 Complete Word Reduplication This is a sample questionnaire in bilingual and bi-scriptal format actually used in the field. Verbal Adverbs/ Verbs

1- YkM+dk pyrs pyrs Fkd x;k (The boy got tired of walking.) 2- YkM+dk pyrs pyrs fxj x;k (The boy fell down while walking.) 3- Okg cSBs cSBs Fkd x;k (He got tired of sitting.) 4- Okg Hkkxrs Hkkxrs Fkd x;k (He said while running or He said when he was about to start running.)

5- Okg xkuk lqu lqu dj Fkd x;k (He got tired of listening to songs.) 6- [kkuk [kkrs [kkrs er cksYkk djks (Do not speak while eating.) 7- ns’k NksM+rs NksM+rs mUgksus /ku dk caVokjk fd;k (They divided the money at the time they were leaving the country.)

8- esjs ns[krs ns[krs og Mwc x;k (He drowned in my presence.) 9- jke ds fdrkc Ik