The Dravidian language

Dravidian language

Dravidian language introduction

The Dravidian language family is, in terms of speakers, the fourth or fifth largest in the world. The family comprises at least twenty-three languages spoken primarily in South Asia by as many as 220 million people. The majority of the Dravidian languages are concentrated in southern and central India, spreading south from the Vindhya Mountains across the Deccan Plateau all the way to Cape Cormorin. Elsewhere, they are spoken in Bangladesh, Nepal, Pakistan and Sri Lanka. Outside South Asia, the Dravidian languages, particularly Tamil, are also spoken in Fiji, Indonesia, Malaysia, Martinique, Mauritius, Myanmar, Singapore, South Africa and Trinidad.

The Dravidian language family has four subgroups: South Dravidian with Badaga, Irula, Kannada, Kodagu, Kota, Malayalam, Tamil, Toda and Tulu; South-Central Dravidian with Gondi, Konda, Küi, Küvi, Manda, Pengo and Telugu; Central Dravidian with Gadaba, Kolami, Naiki and Parji; and North Dravidian with Brahui, Kurux and Malto. Since the 1950s reports of other languages have appeared, but the lack of adequate descriptions prevents us from saying whether these are new, independent languages or merely dialects of ones already known. For South Dravidian, Bellari, Burgundi, Kaikudi, Koraga, Korava, Kuruba, Sholega, Yerava and Yerukula have been reported; for South-Central Dravidian, Âwë, Indu and Savara. Conversely, certain dialects of Gondi, Kolami, and Kurux could prove, under closer inspection, to be independent languages.

Apart from their intrinsic linguistic interest, the Dravidian languages constitute the single most important gateway to many aspects of Dravidian and Indie culture. They inform many facets of Dravidian culture, including literature and religion, the fine arts and philosophy. The ancient poems of love and war, the twin epics, the medieval devotional poems and magisterial Ramayana of Kampan, all landmarks of Tamil literature, are slowly coming to be known and enjoyed by a wider audience. The vacana literature of the Vlrasaiva saints, composed in Middle Kannada and Middle Telugu, are also beginning to cast light on vast literatures hardly known outside of South India. In more recent times, some of the novels of T. Pillai are being translated from Malayalam into languages more accessible to a wider audience. The oral poetry and tales of many Dravidian groups attest to the richness of literary activity of communities without a written tradition, as is revealed in reading Gondi legends, Kota folktales or Toda songs.

Further, speakers of Dravidian languages have contributed to world thought and literature through other languages, notably Sanskrit and English. Three of India’s best-known philosophers, Sankarächärya, Rämänuja, and Mädhvä- chärya, who developed the advaita, visista advaita, and dvaita philosophies, respectively, lived in South India, spoke Dravidian languages and wrote in Sanskrit. In modem times, we have such writers as R.K. Narayan, a speaker of Tamil, Kannada and English, writing a series of much-beloved novels in English. A fellow Mysorean, the late A.K. Ramanujan, speaking and writing in Tamil, Kannada and English, has composed hymns to Murugan in Tamil, poems in Kannada and English, and made Tamil and Kannada classics available to the larger world with his unsurpassed translations.

Religious institutions, such as the bhakti movement, the school of Saiva Sid- dhanta and Vlrasaivism, stand rooted in South Indian, hence Dravidian, culture. Kamatic music, as distinct from Hindustani music, originates in South India: its songs set to music verses composed in Telugu, Tamil, Kannada and, of course, Sanskrit. Classical dance forms such as Bharatanatyam and Kuchuppudi, largely shaped and preserved in the south, interpret these songs. The linguistic and cultural influence of Dravidian India on Southeast Asia is evident in the spread of writing and literacy beyond South Asia: the Pallavas, a Tamil dynasty from the early centuries of the Common Era (c. 275 to 500 ce), carried their writing system on their voyages to the east. The Pallava writing system forms the basis for several writing systems of Southeast Asia and beyond: Myanmar, Thailand, Laos, Cambodia, Malaysia, Indonesia, Sulawesi (the Celebes) and the Philippines. While South Asia has not been politically united - the time from the Mau- ryan Empire to the establishment of the Indian Republic has seen diverse kingdoms, dynasties and feudatory states - it has been united culturally. Speakers of the Dravidian languages have contributed to this cultural unity in these many ways.

Dravidian language speakers

The Eighth Schedule of the Indian Constitution (1951) mandates the creation of states within the Indian Union along linguistic lines. Four Dravidian languages, Kannada, Malayalam, Tamil and Telugu, serve as the basis for establishing the four states of Karnataka, Kerala, Tamil Nadu and Andhra Pradesh, respectively. These four languages are recognised as official languages of the Indian Union. Tamil also has the status of a national language in Malaysia, Singapore and Sri Lanka.

The number of both the Dravidian languages and their speakers have been fraught with uncertainty. Bishop Robert Caldwell (1814-91) listed nine languages in the 1856 edition of his comparative grammar of the Dravidian language family; that number had doubled by 1940 and has nearly tripled as we enter the twenty-first century. In the early years of the twentieth century, George Grierson (1851-1941) conducted a linguistic survey of India: he listed 179 languages and 544 dialects. The survey covered all of British India except the provinces of Burma and Madras and the princely states of Hyderabad, Mysore, Travancore and Cochin. With the exception of Burma, all the regions excluded from the survey lie in South India, the heartland of the Dravidian languages. Grierson nevertheless presented descriptions of the major Dravidian languages spoken in these excluded regions, such as Kannada, Malayalam, Tamil and Telugu, but omitted such others as Irula and Toda.

The advance of years has done little to remove this uncertainty over numbers. The Census of India is taken in the first year of each decade; accordingly, a Census is published for 1931, 1941 and so on. In the 1951 Census, the first taken in independent India, 782 language names were returned by individuals living in India as their mother tongue. In the 1961 Census that number rose to 1,652, and in 1971 it again nearly doubled to just over 3,000. The 1981 Census, by contrast, abruptly limited the number to 105 by applying the criterion that a language should be spoken by over 10,000 individuals before it might be included. Thus, some ‘small’ Dravidian languages, such as Toda and Kota whose speakers number less than 1,000, appear in some editions of the Census but not in others. The 1981 Census includes 17 Dravidian languages, giving the combined total of speakers as 157,836,723, or 23.9 per cent of the reporting population.

The Dravidian languages enumerated in the 1981 Census include the four languages scheduled in the Indian Constitution: Kannada (26,887,837), Malayalam (25,952,966), Tamil (44,730,389) and Telugu (54,226,227). The thirteen remaining, non-scheduled languages include Gondi (1,954,693), Jatapu (23,366), Kisan (155,283), Kodagu (93,116), Kolami (78,500), Konda (11,062), Kondh (204,501), Koya (242,534), Küi (507,639), Kurux (1,264,590), Malto (94,614), Parji (33,091) and Tulu (1,376,306). Even though this Census contains a more manageable set of languages, it has not eliminated some prior confusions in the identification of languages. Respondents may on occasion return the name of their community or ethnic group as a language name. As early as Grierson, Kisan was recognised as an alternate name for Kurux; the name is vague in any event since the word means ‘cultivator’. Further, many who identify themselves as Gonds speak no Gondi. The group that identifies itself as Gadaba speaks two distinct languages, one Dravidian (Konekor Gadaba) and absent from the Census, the other Munda (Gotub Gadaba) and included in it. Scholarly consensus now views Koya as a dialect of Gondi. Jatapu, spoken in the hills of Srikakulam District in Andhra Pradesh, as well as in the Koraput and Ganjam districts of Orissa, may well be related to Küi and Küvi. Kondh, spoken primarily in Orissa, may also be a variety of Küi or Küvi. To add to the confusion, the term Kondh has traditionally been applied to the larger ethnic community that includes both Küi and Küvi speakers. Conversely, the term Küi as reported in the 1981 Census seems not to distinguish among Küi, Küvi and, perhaps, Kübi, another name for Konda. Absent from this Census are Irula, Kota and Toda from the South Dravidian languages; Badaga is listed as another name (dialect) of Kannada. Also missing are Küvi, Pengo, Manda, Naiki and Gadaba (though the Munda language Gadaba is listed). Brahui is absent because it is the only Dravidian language that has no speakers living within the Indian Union. The Census of India naturally omits information on the number of speakers outside Indian territory: Kurux, Mialto, Malayalam, Tamil and Telugu.

The World Atlas 1991 estimates that 48 million people speak Tamil as a first language, and that 66 million speak it as a first or second language. It further estimates that 55 million people speak Telugu as a first language, and that an additional 14 million speak it as a second language. Similarly, Kannada is estimated to be the first language of 25 million, which number rises to 42 million when those who speak it as a second language are included. Malayalam is spoken, according to this source, by 34 million world wide. Thus, approximately 210 million people speak the four literary languages as a first or a second language. It should be recalled that on the Indian subcontinent, where multilingualism is the rule rather than the exception, those who speak one of the four major Dravidian languages as a second language generally speak another Dravidian language as a first language. Many speakers of Konda thus speak Telugu or Oriya as a second language, depending on whether they live in Andhra Pradesh or Orissa. Speakers of Tulu and Kodagu may also speak Kannada, the politically dominant language of Karnataka. Speakers of Toda may also speak Kota and Badaga, two groups with whom they have traditional economic relationships. And since Todas live in what is now Tamil Nadu, they may also speak Tamil. Speakers of Tamil in Mysore may also speak Kannada while speakers of Telugu in Tamil Nadu typically speak Tamil.

Unlike many of the better-known languages and language families of the world, e.g. Indo-European, the majority of the Dravidian languages are spoken and lack a written tradition. Only four have extensive written traditions. The first Tamil inscription dates to c. 254 bce, the first Kannada inscription to c. 450 c e, the first Telugu inscription to c. 620 ce and the first Malayalam inscription to c. 830 c e. In each of the four, literary composition begins - more properly, begins to be committed to writing - one or two centuries after the appearance of these first inscriptions. These four thus have longer literary traditions than most of the modem languages of Europe. The first inscription in Tulu dates to the fifteenth century while literary texts begin in the eighteenth century. Apart from these five, our knowledge of the Dravidian languages is quite recent, dating from the middle of the nineteenth century. A number have only been discovered in the latter half of the twentieth century. And it seems likely that new Dravidian languages remain to be discovered, particularly in Central India.

Each of the four literary languages has evolved its own distinctive writing system.all four writing systems are alphasyllabic, and can be traced back to southern varieties of the Ashokan Brahmi script. The four scripts have also been used for other Dravidian languages. Tulu has been written in both Malayalam and Kannada scripts, Kodagu is written in Kannada script and Konda is written in Telugu script. Kurux is written in the Devanagari script, which is more broadly associated with Sanskrit and Hindi. Brahui, because it is spoken in Pakistan and Afghanistan, is written in the Urdu script. A number of Dravidian languages discovered since Indian Independence, such as Parji and Pengo, have been committed to writing by linguists using phonemic transcriptions based on Roman script. As noted earlier, the writing system of the Pallava dynasty (275 to 500 ce), which ruled from Kan- chipuram in Tamil Nadu, forms the basis for many writing systems extending from Myanmar to the Philippines, from Sri Lanka to Java.

The Dravidian languages have shared the South Asian subcontinent with three other language families since prehistoric times: the Indo-Aryan branch of Indo-European, the Munda branch of Austro-Asiatic and Sino-Tibetan. The linguistic influence of Indo-Aryan and Dravidian on each other has been well studied and documented; as Burrow has shown, the influence was not just in one direction: Dravidian languages had already exerted some influence on Indo- Aryan even in prehistoric times. The long co-existence of these four language groups has resulted in linguistic cross-influence and borrowing, to the extent that South Asia is now a linguistically defined area (Emeneau 1956; Masica 1976).

In historic times, the Dravidian languages have been in contact with languages from the West. From both literature and the archaeological record, we know that Roman merchants journeyed to the Tamil coastal emporia in the early centuries of the common era to buy spices, fabrics and other luxury goods with Roman gold coins. The spread of Islam to the subcontinent in the medieval period carried with it such administrative languages as Persian and Arabic. These influences had reached as far south as Mysore in Karnataka and Tanjore in Tamil Nadu by the time the British appeared on the scene in the middle of the eighteenth century. English has had, and continues to have, an impact on the vocabulary and sound systems of Dravidian languages over the past two centuries.

Commerce and colonisation have carried a number of Dravidian languages, particularly Tamil, beyond their traditional spheres, e.g. to Myanmar, South Africa, Fiji and islands in the Caribbean. Trade links between South and Southeast Asia have been attested since the late classical period, and continue to the present day so that Tamil is an official language in both Malaysia and Singapore. Such words as catamaran (Tamil kattamaram ‘[boat made of] tied timbers’), mango (Tamil, Malayalam mänkäy), coolie (Tamil küli ‘wages’), curry (Tamil kari ‘meat preparation’), mulligatawny soup (Tamil milakaittanmr ‘pepper-water’) and coir (Tamil kayiru ‘rope’) have made their way into English from Dravidian sources.

South Dravidian

Tamil (tamiz), the best known of the Dravidian languages, belongs to the South Dravidian (SDr) subgroup. It is first recorded in a lithic inscription in a form of Ashokan Brahmi script which is dated to c. 254 b c e. It is therefore one of India’s two classical languages, alongside the more widely known Indo-Aryan language Sanskrit. However, Tamil is the only one of the two with a palpable continuity between its classical and modem forms. Records of Tamil reveal three distinct historical stages: Old Tamil (c. 300 bce to 700 ce), Middle Tamil (700 to 1600 ce) and Modern Tamil (1600 ce to the present). In all its forms, Modem Tamil consists of many geographic dialects, with the major distinction drawn between Sri Lankan Tamil and the continental dialects of India. Tamil also exhibits marked diglossia with a high and a low variety, called centamiz ‘pure Tamil’ and kotuntamiz ‘harsh Tamil’, respectively. The high variety is used in political oratory, belletristic writing and news broadcasts while the low variety is used in virtually all face-to-face communication. But in recent times, the high variety has begun to give way to the low. There are over 50 million speakers of the language today. Within India, it is spoken in the states of Tamil Nadu, Kerala, Karnataka and Andhra Pradesh, as well as the Union Territory of Pondicherry. Outside, the language is reported in Sri Lanka (3,346,000), Malaysia (274,218), South Africa (250,000), Singapore (191,200) Fiji (6,663), Thailand and Mauritius.

Between 800 and 1200 ce the western dialects of Tamil, geographically separated from the others by the Western Ghats, developed into Malayalam. The Vazappalli inscription, the first record of the language, dates to c. 830 c e. Malayalam has lost subject-verb agreement and has borrowed so many Sanskritic loans that the modem language now distinguishes aspirated from nonaspirated stops in its phonology. It is spoken in Kerala, Tamil Nadu and Karnataka, as well as Lakshadweep (the Laccadive Islands). As with the other major Dravidian languages, Malayalam has a number of geographic and caste- based dialects. With perhaps 34 million speakers, Malayalis boast the highest literacy rate in the Indian Union.

Not only do the Western Ghats separate Malayalam from Tamil, they are also home to several of the non-literary South Dravidian languages. Communities speaking Badaga, Kota, Irula and Toda live in and around the Nilgiri Mountains, a branch of the Western Ghats that rises sharply above the Coimbatore Gap. Relatively isolated from the mainstream of traditional Indian society, the Todas, Kotas and Badagas have over many generations developed close economic ties, even while speaking different languages. Toda and Kota are spoken by so few people that they appear and disappear in different editions of the Indian Census. No more than 800 people speak Toda while figures returned on Kota give a population of 700. Despite the small number of speakers, these languages have attracted great anthropological and linguistic interest, starting in the nineteenth century with British adventurer Sir Richard Burton’s observations on Toda. In this century, Emeneau (1944-6) recorded four volumes of Kota texts; his 1984 analysis of Toda grammar and texts is a model of scholarship, revealing the linguistic wealth of the ‘small’ Dravidian languages. While the grammar of Toda and Kota are relatively close, the phonology of Toda is perhaps the most divergent among the Dravidian languages. Badaga, often treated as a dialect of Kannada, may have as many as 50,000 speakers. This language has the distinction of having one of the most extensive dictionaries of any of the non-literary Dravidian languages (Hockings and Pilot-Raichoor 1992). Irula, a language with perhaps 12,000 speakers, is spoken in the hilly spurs of the Nilgiri Massif between Kerala and Tamil Nadu. It is closely related to Tamil and Malayalam. Zvelebil (1973) presents a sketch of this language. Toda and Kota, in particular, are threatened with extinction in the near term as cultural and social pressures compel their assimilation into the larger social matrix.

Tulu is spoken in between the Western Ghats and the coast of the Arabian Sea in western Karnataka, northern Kerala and southern Maharashtra. Although the language is not traditionally ranked as a literary language, many brahmin native speakers of Tulu have been literate in Sanskrit, Kannada and Malayalam. As a result, Tulu has been written down for nearly 250 years, first in Malayalam and then in Kannada script. Today there may be as many as three million people speaking Tulu in four or five dialects: one half speak it as a first language, the other half as a second language. Tulu appears to share several features with the South-Central Dravidian languages, so many in fact that some scholars place it in that subgroup. It may well be the first individual language to branch off of Proto-South Dravidian.

Inland in the Coorg District of Karnataka, Kodagu (kodava, Coorg) is spoken by as many as 95,000 people in three dialects. It is also spoken in neighbouring regions of Tamil Nadu and Kerala. It has recently come to be written in Kannada script, although this writing system is not sufficiently differentiated to handle the rich vowel system that has evolved in Kodagu.

Kannada 0kannada, Kannarese, Canarese) is spoken by perhaps 25 million people throughout Karnataka and in the neighbouring states of Maharashtra, Andhra Pradesh and Tamil Nadu. Written records of the language reveal three historically distinct stages: Old, Middle and Modem. It has at least four major geographical dialects, caste dialects and diglossia, although these differences do not seem as attenuated as in Tamil. For a full description of the language, see Sridhar (1990), to whose work the chapter on Kannada in this volume owes much.

Some of the isoglosses that define the southern subgroup are the general loss of *c- in such etyma as *cäru ‘six’ (DEDR 2485), *cïy ‘give’ (DEDR 2598) and *cuppu ‘salt’ (DEDR 2674), which in Tamil are represented by the reflexes äru ‘six’, ïy ‘give’ and uppu ‘salt’. These languages have also innovated the use of the auxiliary verb iru ‘be’ (DEDR 480) as the perfect tense auxiliary to compensate for the loss of man ‘be’ (DEDR 4778) in that group (see Steever 1993: 96- 102). Serial verb formations, though attested in Old Tamil and Old Kannada,have been lost in all the modem languages of this subgroup.

During the period of Early Old Tamil (100 bce to 400 ce), the Pântiya, Cëra and Cözä dynasties ruled over much of South India. These kings and other chieftains patronised many bardic poets. Two anthologies of love and heroic poems composed by these bards survive: they contain 2,381 poems ranging in length from 3 to 782 lines. Totalling 32,000 lines, this corpus is known as Cankam (‘academy’) literature. During this period, with the propagation of Jainism and Buddhism in South India, a number of Prakrit and Sanskrit borrowings entered Old Tamil and appear in the Cankam anthologies. The literature of Late Old Tamil (400 ce to 700) comprises the two epics Cilappatikäram and Ma nimé kal ai, several ethical texts and certain poems conventionally included in the Cankam anthologies. The language of Old Tamil is thus embodied in a fixed corpus of poetic texts; conversely, poetic usage characterises the grammar of the language.

From its beginnings in Tolkäppiyam, traditional Tamil language study has linked grammar (ilakkanam ‘that which characterises’) and literature (ilakkiyam ‘that which is characterised’) so closely that the texts and their commentaries have become symbiotic. As noted later, these texts are not readily accessible without the help of commentaries written during the ancient and medieval periods. Although the Cankam corpus is ancient, many texts and commentaries were only recently rediscovered in the last century through the efforts of such scholars as U. Ve. Cäminäthaiyar. So, paradoxically, the modem study of Old Tamil grammar is quite new.

Old tamil

Background and History

Of all the Dravidian languages Tamil has the longest literary tradition, covering more than two thousand years. The earliest records are cave inscriptions from the second century b c e; the earliest extant literary text is the grammar Tolkäppiyam (100 bce), which describes the grammar and poetics of Tamil during that period. During its two-thousand-year uninterrupted history, Tamil distinguishes three different stages: Old Tamil (300 bce to 700 ce), Middle Tamil (700 ce to 1600) and Modem Tamil (1600 ce to the present), each with distinct grammatical characteristics.

Causation, for example, is expressed lexically in Old Tamil, morphologically in Middle Tamil and syntactically in Modem Tamil. Old Tamil has verb bases whose causative stem is idiosyncratic and must be listed in the lexicon, e.g. iru ‘sit’ and iruvu ‘make sit’, akal ‘disappear’ andakarru ‘make disappear’. In Middle Tamil causative stems are productively formed by suffixing -vi, -pi or -ppi to a verb base where the suffix chosen depends on the phonology of the base, e.g. cey ‘do’ and cey-vi ‘make do’, aru ‘cut’ and aru-ppi ‘make cut’. And in Modem Tamil causation is expressed periphrastically by means of the auxiliary verbs vaikka ‘place’, ceyya ‘do’ and panna ‘make’ following the infinitive of the main verb, e.g. arukk-a vaikka ‘make cut’. The techniques used to form causatives in Middle and Modem Tamil occur sporadically in Old Tamil, e.g. cèr-vi ‘make join’ (pari 12.74) and väz-a.c cey ‘make live’ (pura 367.10). Thus, despite the grammatical individuality of each stage of the language, many features of Middle and Modem Tamil are anchored in Old Tamil, demonstrating a grammatical continuity from the old to the modem language.

Modern Tamil

Background and History

Tamil (tamiz) belongs to the South Dravidian subgroup of languages. Spoken in southern India and northeastern Sri Lanka from prehistoric times, Tamil literature and other documents reveal three distinct historical stages of the language: Old Tamil (300 bce to 700 ce), Middle Tamil (700 to 1600 ce) and Modem Tamil (1600 ce to the present). This chapter treats the modem language.

Lehmann traces both continuities and discontinuities from one stage to the next in the historical development of Tamil. Greater continuity may be observed between Old Tamil and modem Sri Lankan Tamil than between the old language and the modem mainland dialects spoken in India. Sri Lankan Tamil preserves the medial deitic series in μ -, as in u-van ‘man in between’, and the synthetic present perfect in -an-, as in con.n-än-än ‘I have said’. The mainland dialects have lost these forms. Modem Sri Lankan Tamil has also resisted the borrowing of contrastive voiced stops in the spoken register: mainland döcai ‘rice pancake’ corresponds to Sri Lankan töcai ‘id.’.

Although Old Tamil lacked a present tense form, Middle Tamil developed two: one with the present tense marker -(k)kir-/-(k)kinr-, the other with the marker -äninr-. Only the first survives in Modem Tamil, and has been phonologically reduced to -(Jc)kr- in several dialects (Steever 1993). The negative synthetic verb survived into Middle Tamil, e.g. vâr-êm ‘we did/do/will not come’, but is lost in Modem Tamil, surviving only in such stock phrases as avanai.k känöm ‘he is not around’ (lit. ‘we.don’t.see2 himj’). Old Tamil stands alongside Sanskrit as one of India’s two classical languages, but Tamil alone has a recognisable continuity with a modem Indie language.

The best known of the Dravidian languages, Tamil is spoken today as a first language by approximately 50 million people. During its history, but especially during the Chola Empire and the British Raj, it was transplanted to Malaysia, Mauritius, Myanmar, Guyana and Martinique. In the modem world it is an official language in India, Malaysia, Singapore and Sri Lanka. The Eighth Schedule of the Indian Constitution (1951) recognises Tamil as one of India’s eighteen constitutional languages. The Tamil Nadu Language Act (1956) accords Tamil the status of first official language within the state of Tamil Nadu, with English as the second.

Of the four literary languages, Tamil has the oldest history, dating from about 300 b c e. The earliest records, inscriptions in Ashokan Brahmi script, date to c. 254 bce. Its literature, preserved on palm-leaf manuscripts, in copper-plate inscriptions and by rote memory, covers 2,000 years. Three stages of Tamil appear in these records: old (300 bce to 700 ce), medieval (700 to 1600 ce) and modem (1600 ce to the present). Its closest relatives are Malayalam, spoken in neighbouring Kerala, and Irula, a tribal language spoken in the Nilgiris District of Tamil Nadu.

Kannada

Introduction

Kannada (kannada) belongs to the southern branch of the Dravidian languages. First attested in an inscription found near Halmidi village in Hassan District dating to c. 450 ce, Kannada has a continuous literary tradition from the ninth century ce to the present. It is one of the four literary Dravidian languages; only Tamil has a longer literary tradition.

The Eighth Schedule to the Indian Constitution, which mandates the creation of states along linguistic lines, lists Kannada among the official languages of the Indian Union. It is spoken today by over 25 million people throughout the state of Karnataka, where it is the official language. While most inhabitants of Karnataka speak Kannada as a first language, it is also the second language of many whose first language is Kodagu or Tulu.

Kannada varies along several dimensions: historical, geographic, social and register. The language shows three historically distinct stages: Old Kannada dates from 450 ce to 1200, Middle Kannada from 1200 to 1700 and Modem Kannada from 1700 to the present.

Kannada includes four major dialect areas, which may be subdivided according to finer gradations. The southern dialect area includes the varieties spoken in and around the major cities of Bangalore and Mysore (Bright 1958; Sridhar 1990). The northern dialect area is centred on the city of Dharwar (Hiremath 1980; McCormack 1966); the western dialect in Mangalore District; and the northeastern in and around Bijapur. For example, the northern and northeastern dialects show the influence of neighbouring Marathi, an Indo-Aryan language.

The existence of caste dialects in Kannada reflects the historical fact that much traditional social and economic interaction respected caste structures. The primary distinction appears to be among brahmin, non-brahmin and harijan varieties of the language. As Sridhar (1990) observes however, under the influence of modem trends, one of which is the propagation of universal education, social dialect differences are shifting from caste-based to class-based lines.

Tulu

Indroduction

Tulu (tulu) is spoken by more than three million people, half as their mother tongue and half as a second language. Its use is confined to a discrete area including most of the South Kanara district (except the north) in Karnataka and Kasaragod Taluk in the north of neighbouring Kerala. Tulu speakers have also migrated within India in pursuit of commercial opportunities; they are consequently found in most major cities even though the Census of India generally counts them as Kannada speakers.

Tulunad (tulunädm), where Tulu is traditionally spoken, is geographically and sociolinguistically compact. Its geographic compactness derives from the natural boundaries that enclose it: the rivers Suvama and Chandragiri form its northern and southern boundaries while the Western Ghats and the Arabian Sea set its eastern and western boundaries. Tulunad thus lies wholly on the coastal belt of Karnataka and Kerala. Its sociolinguistic compactness comes from the fact that throughout the region there is a uniform, two-layered system of communication consisting of two common languages, Tulu and Kannada. People who speak different mother tongues, including Tulu with its different dialects, Kannada (Havyaka, Gowda and Baira dialects), Konkani, Marathi, Malayalam and Koraga, all use a common variety of Tulu for informal communication, as in business, and standard Kannada for formal communication, as in education.

Tulunad was administratively cohesive when it formed part of the Madras Presidency; but the reorganisation of Indian states in 1956-57 disrupted this with the accession of Kasaragod Taluk in the southwest of Kerala. The remainder of the area joined Karnataka.

North Dravidian

The North Dravidian languages include Brahui, Malto and Kurux. Spoken by nearly two million people, Kurux {kurux, Oraon, Uraon, Kisan) comprises several dialects spoken in India, Bangladesh and, recently, in the Terai of Nepal (where it is known as Dhangar or Jhangar). Notable in its grammar is a distinction between men’s speech and women’s speech, one which is present also in Malto. Malto (mal saba) is spoken by approximately 100,000 people in India and Bangladesh. It has at least three dialects distinguished according to phonology, morphology and lexicon: Kumarbhag, Malpaharia and Sawriya. Building on noun-noun compound constructions inherited from Dravidian, Malto appears to have innovated a classifier system. The chapter on Malto in the present volume is endebted to Mahapatra’s 1979 authoritative grammar of the language.

Brahui, spoken in Baluchistan on the northwest frontier of South Asia, is so divergent from Kurux and Malto that it might well constitute its own subgroup, the first to branch off from the Dravidian proto-language. The geographic separation of Brahui from the remaining Dravidian languages has invited some to speculate that the Dravidian languages once blanketed the subcontinent and were broken up into linguistic islands with the advance of Indo-Aryan languages. This speculation has led, in turn, to the more remote conjecture that the language of the Indus Valley Civilisation was Dravidian. The position advanced in the chapter on Brahui in the present volume, less grand but perhaps more sober, is that the Brahuis migrated westward from a homeland somewhere in Orissa and Bihar to their current location in Pakistan, Afghanistan and Iran.

With a lack of adequate descriptions, the North Dravidian group is often treated as an ‘elsewhere’ subgroup. Emeneau (1962) discusses the most common phonological isoglosses that the North Dravidian languages share, namely the uniform treatment of initial velar stops in the proto-language: *k remains a voiceless velar stop before */- and *ί-, and has a variety of other reflexes in other environments. The languages of the other three subgroups tend to palatalise the initial velar stop before */- and *F-.

Malto

Background

Malto, the northernmost of the Dravidian languages in India, is spoken by nearly 100,000 people in the Rajmahal Hills of Bihar. Pockets of Malto speakers also live in the states of West Bengal, Tripura and Orissa. Speakers of Malto call their language mal sapa ‘man’s language’ or malto ‘of the man’, and they call themselves maler ‘men’. Three dialects have been identified: Kumarbhag, Malpaharia and Sawriya.

Brahui

Habitat and Speakers

Brahui, the conventional spelling for brähöi (Persian/Urdu brähüi), designates both a tribal group and a language. The two are not coterminous: many Brahui tribesmen do not speak Brahui even as a second language; further, some Balochi tribes use Brahui as a second language. The main habitat of Brahui tribesmen, as well as the main area where Brahui is spoken, extends continuously over a narrow north-south belt in Pakistan from north of Quetta southwards through Mästung and Kalat (including Nushki to the west) as far south as Las Bela, just inland from the Arabian sea coast. This belt is approximately 100 miles wide (from 65 degrees 30 minutes to 67 degrees east longitude, roughly as in Grierson (1921:327)).

Substantial numbers of nomadic or semi-nomadic Brahui speakers inhabit the Afghan Söräwäk desert northwest of Nushki, in an area extending west along the Hilmand River into Iranian Sistan. Brahuis are not considered to be ethnically distinct from the Baloch in Afghanistan or Iran. There seem no longer to be any Brahui speakers in Iran south of Sistan, although G.P. Tate (1909) mentioned colonies of Brahuis as far south as Khäs. A small number of Brahui speakers are settled in Turkmenistan, mainly in the Marw oasis: together with the Baloch, they are late nineteenth- and early twentieth-century ce arrivals from India and Afghanistan. Many Brahui speakers are more or less permanently settled in the large cities of the Pakistani Sind: Karachi, Hyderabad, Sukkur and Larkana, mainly as casual labourers.

The only census data for Brahui come from pre-Partition India, and all are marred by the confusion between ‘Brahui tribesman’ and ‘Brahui speaker’. For centuries, Brahuis have customarily described themselves as Baloch to outsiders, thus leading to consistent overestimates of the number of speakers and underestimates of the number of tribesmen. These issues are discussed in the literature (Elfenbein 1987: 221-2; Encyclopedia Iranica: 433-44; Bray 1913). The major conclusions, extrapolated to 1996, are as follows. There are approximately 700,000 Brahui tribesmen, mainly in Pakistani Baluchistan and in Afghanistan. Of these, approximately 100,000 are primary speakers of Brahui,mainly in Pakistan; perhaps 300,000 are secondary speakers of Brahui in Pakistan and Afghanistan. Among the secondary speakers, the primary language is the Iranian language, Balochi. Further, most primary speakers of Brahui speak some Balochi as well. But fully 300,000 Brahui tribesmen speak no Brahui at all.

At present 27 tribes constitute the ‘Brahui Nation’; eight are considered nuclear tribes and 19 peripheral. Some 90 per cent of Brahuis belong to the peripheral tribes. Only two of the nuclear tribes are primary speakers of Brahui; one consists of primary speakers of Balochi; four are bilaterally bilingual in Brahui and Balochi; and one is bilaterally bilingual in Brahui and Siraki. Of the peripheral tribes, only three seem to be primary speakers of Brahui while the remainder are either bilaterally bilingual in, or primarily speakers of, Balochi. The largest tribe, the Mengal, are primary speakers of Balochi.

History

What little we know of the history of the Brahuis is summarised in the Encyclopedia Iranica (436-8) and in Elfenbein (1987: 226ff). The Brahuis first emerge in an historical light in the seventeenth century CE when, in reaction to Mughal pressures, they participate in the well-organised polity of the Brahui Confederacy, or the Khanate of Kalat, in close alliance with the Baloch and Dehwars. Later called Kalat State, the khanate was absorbed into modem Pakistan in 1948 after some 250 years of semi-independent existence.

The name Brahui is probably non-Brahui in origin: it likely comes from Siraiki brähö, itself a Siraiki borrowing of Ibrâhïm, the name of a prophet. It appears to have become the native designation only after the Brahuis migrated into Sindh and became Muslims, perhaps 1,000 years ago.

Brahui prehistory is entirely obscure. I have argued against the traditional assumption that Brahuis are a relic of the original Dravidian migrations into India, c. 3000 b c e, who remained in Kalat as the first group to separate from other Dra- vidians. In my view a more prosaic history is far more likely. The Brahuis, never a very close-knit group, migrated northwest from the Central Deccan in India across Gujarat and into Sindh in many waves from about 800-1100 c e. Afterwards they entered the Kalat highlands. Their subsequent political importance developed out of early contact with the eastward-migrating Baloch (c. 1000 ce). This picture fits the grouping of Brahui in the North Dravidian group with Kurux and Malto, from whom the Brahuis probably separated more than a millennium ago.

Introduction

South-Central Dravidian

Telugu (telugu, tenugu, telungu), the language spoken by the greatest number of speakers (60 million), belongs to the South-Central Dravidian (SCDr) subgroup. The history of Telugu may be divided into three stages: Old Telugu (600 to 1100 c e), Middle Telugu (1100 to 1600 ce) and Modern Telugu (1600 to the present). Modem Telugu has four regional dialects in Andhra Pradesh (Krishnamurti and Gwynn 1985): northern, southern, eastern and central. It is also spoken in Karnataka, Orissa, Tamil Nadu and as far away as Kerala. The standard language underwent a number of reforms during the nineteenth century with the result that the written and spoken varieties of the modem language do not diverge as sharply as in Tamil. Old Telugu had a two-way tense distinction, past and non-past, while Middle and Modem Telugu have a three-way distinction, past, present and future. In Middle Telugu a new past tense form developed in all areas of Andhra Pradesh except the Rayalseema which has retained the past tense form of Old Telugu to the present day.

The six remaining languages of this subgroup are non-literary languages spoken in Andhra Pradesh, Madhya Pradesh, Maharastra and Orissa. Gondi in all its forms has approximately two million speakers. Its numerous dialects form two groups: those to the northwest and those to the southeast. Koya, sometimes listed as a separate Dravidian language, appears to belong to the southeastern group of Gondi dialects. The Gondi dialect detailed in this volume is that of Muria Gondi, described by Andres (1977) whose work forms the backbone of the present description. Muria Gondi, spoken in the Bastar District of Madhya Pradesh, belongs to the southeast group of dialects.

Konda, also known as Kübi, is spoken in northeast Andhra Pradesh and southern Orissa. It stands linguistically between Telugu and Gondi on the one hand, and the four remaining South-Central languages on the other. It appears to be the most phonologically conservative language of the group, preserving the Proto-Dravidian distinction between the tap *r and the stop *r.

Küi, Küvi, Pengo and Manda are collectively known as the ‘Kondh’ languages because they are spoken primarily in the region of the Kondhmal Hills in western Orissa. They are closely related: in finer gradations, Küi is paired with Küvi and Pengo is paired with Manda. These four are distinct from the other three South-Central languages and, indeed, from all other Dravidian languages in having innovated a system of object-verb agreement alongside the inherited system of subject-verb agreement (see Steever 1993).

One of the most distinctive isoglosses of the South-Central group is the neutralisation of the distinction between long and short vowels in all but radical syllables. All these languages have innovated, to a greater or lesser extent, a rule of Metathesis or Apical Displacement (Krishnamurti 1961, 1978), which now permits resonants and laterals to appear word-initially, as in Telugu räyu/wräyu ‘write’ from *varay ‘draw, make lines’ (DEDR 5263), Küi lävenju ‘youth’ and Gondi leyyoR ‘id.’ from *ilay-antu ‘young man’ (DEDR 513). Apart from Modem Telugu, which lacks the relevant verb base, all South-Central languages have a rule that permits the truncation or total deletion of the perfect tense auxiliary verb man ‘be’. The truncation of this auxiliary verb has led to the creation of a new present perfect tense series in Pengo (see Steever 1993).

Old telugu

introduction

Old Telugu belongs to the South-Central branch of the Dravidian languages. All the languages of this subgroup underwent the phonological change of metathesis or apical displacement whereby certain resonants and sonants which in Proto- Dravidian (PDr) occurred only in the offset of a syllable now appear in the onset. The PDr verb *varay ‘sketch, draw, write’ becomes varai ‘id.’ in Tamil, but becomes first vräyu ‘id.’ then räyu ‘id.’ in Telugu. While the six non-literary languages of this subgroup have contrastive vowel length only in initial syllables, Telugu words may have contrastive length in any syllable. Of the seven languages of this subgroup, it is the only one with a written literary tradition. Old Telugu and Gondi retain serial verb constructions in the negative, while the other languages, including Modem Telugu, have lost them.

Linguistic evidence divides Telugu into three historical stages. The Old Telugu period runs from the seventh century ce to about 1100; the Middle Telugu period runs from 1100 to 1600; and the Modem Telugu period runs from 1600 to the present. Our knowledge of the language before the twentieth century comes from two basic corpora: inscriptions and literary works. Beginning with the seventh century ce, Old Telugu first appears in inscriptions; and from 1100 onwards in a much larger literary corpus. The language stages in these two corpora do not coincide in time: the inscriptional corpus is the more innovating of the two, the literary corpus is more conservative and, indeed, resolutely archaising. Nannaya, writing in the early medieval period, uses forms attested in Old Telugu inscriptions and avoids the more innovating forms attested in contemporaneous inscriptions. With the exception of Vlrasaiva literature introducing colloquialisms into the language of medieval devotional texts, the literary language generally lagged behind the spoken language as reflected in inscriptions until the early twentieth century when the spoken form finally took precedence (see Chapter 8 on Modem Telugu). Literary forms of the early medieval period are often cited as instances of Old Telugu because they are more abundant, more cohesive and deliberately archaic.

The keystone of classical literature is the epic Ändhramahäbhärata (c. eleventh century ce to fourteenth century) written by three authors (kavitraya ‘trinity of poets’): Nannaya (eleventh century), Tikkana (thirteenth century) and Errana (fourteenth century). While establishing a division between Old Telugu and Middle Telugu remains somewhat problematic in view of the discontinuity between the inscriptional and literary varieties, the latest date ascribed to Old Telugu is the fourteenth century. Around this time, literary genres diversify to include the vacana literature (thirteenth century onwards), influenced by the Vlrasaiva movement that began in Karnataka. This introduces a more colloquial diction into the literature. The prabandha literature of the sixteenth century retains - perhaps self-consciously - many features of the epic language. The classical period, which includes both Old and Middle Telugu, begins to wind down in the late nineteenth century when reforms are initiated to write and print Telugu in a manner that more closely matches the language spoken in Andhra Pradesh.

Telugu

Background and History

Telugu, a literary Dravidian language, is spoken by 54.2 million people (1981 census) and is the official language of Andhra Pradesh. Its literary texts date from the eleventh century ce onwards. Telugu (telugu < telungu) is also known as Tenugu (tenügu < tenungu) and Ändhram. On comparative grounds, Telugu belongs to the South-Central subgroup of South Dravidian. Its other close sisters are Gondi, Konda, Kui, Kuvi, Pengo and Manda spoken to its north and northeast. Culturally it has had closer links for centuries with the other two literary languages to its south and west, i.e. Tamil and Kannada.

Dialects and Standard Language

There were two styles, literary and colloquial, even during the period of early Telugu. The differences between them became pronounced from the eleventh century onwards: the poets maintained the archaic literary style while prose inscriptions, some commentaries, chronicles and devotional songs reflected the spoken form.

Modem prose writings in the form of essay, fiction, play, criticism, etc., started mainly from the nineteenth century under the influence of English. At the beginning of the twentieth century there was a prolonged battle between the classicists who wanted the grammar of the literary language used for prose and modernists who wanted it closer to the educated speech. Finally, Modem Telugu became the vehicle of all prose from 1940 onwards, with the model set by creative writers, newspapers and radio.

Modem Telugu has four regional dialects: Northern (nine districts of Telan- gana), Southern (four districts of Rayalasima, Nellore and Prakasam), Eastern (Srikakulam, Vijayanagaram and Visakhapatanam) and Central (East and West Godavari, Krishna, Guntur) (see Figure 8.1). There are many social dialects, based on the level of education and sensitivity to standard written form. 

The educated speech has a larger inventory of phonemes, less variation in grammatical marking, and more borrowings from Sanskrit and English. Modem Standard Telugu (MST) has evolved from the writings of educated speakers of the central dialect. The geographical base of MST is far more widespread than that of standard Bengali (Calcutta based) or standard Marathi (Pune based) and thus its evolution is unlike that of English and many other modem European languages .

Konda

Background and History

Konda belongs to the South Central Dravidian languages; its sisters are Telugu, Gondi, Küi, Küvi, Pengo and Manda. Its speakers number about 34,000 (1971 Census) and live in the Agency tracts of Visakhapatnam, Vijayanagaram and Srikakulam districts. They are known by various names: Konda Dora (lit. Hill lord), Konda Käpu, Öja, Pändava Raytu. They call themselves Κοηςίβη or Küb^ (-η plural suffix) and their language Kübi. Konda means ‘hill’ in Telugu, the dominant regional language. Konda speakers, particularly men, are bilingual in Telugu or Oriya, depending on where they live. This chapter is based on Bh. Krishnamurti (1969), which consists of texts, grammar and lexicon. The main informants for the data come from Araku Valley, about 80 miles northwest of Visakhapatnam. However, lexical variants for a number of items have also been taken from other dialects.

Gondi

Introduction

Gondi belongs to the South-Central branch of Dravidian. It is spoken by perhaps two million people in the central Indian states of Maharashtra (1,300,000), Madhya Pradesh (450,000), Andhra Pradesh (270,000) and Orissa (84,000). Gondi’s extensive dialect variation may be attributed to several factors: it covers a wide geographic area, has no written tradition and lacks official status.

As a South-Central Dravidian language, Gondi belongs to the same subgroup as Telugu, Konda, Küi, Küvi, Pengo and Manda. Like Telugu and Konda, but unlike the other four, Gondi lacks a system of object-verb agreement. Like Old Telugu, but unlike Konda, Gondi still uses serial verb constructions with the auxiliary verb ä ‘become’ (DEDR 333) in the negative conjugation. And, unlike the remaining South-Central languages, certain dialects of Gondi have lost contrastive vowel length.

Traditional accounts hold that Gondi once extended from the Godavari River north across the Deccan Plateau to the Vindhya Mountains. Such a broad geographic dispersion would have naturally contributed to the proliferation of dialects. When Muslim writers chronicled the region in the late medieval period, they named it Gondwana after the people living there and the language they spoke. Gonds do not use the term Gondi themselves. The term they prefer, if they speak Gondi at all, is koytor, which probably derives from the Dravidian etymon for ‘mountain’ (DEDR 2176, *kö) or possibly ‘hill’ (DEDR 1864 *kuntu). In Muria Gondi, the dialect presented here, the term koytoR applies to males, koytor to females.

Gondi is a non-literary language: no written records of the language exist before the middle of the nineteenth century ce. Historical accounts are therefore fragmentary. Reconstruction of the linguistic milieu suggests an extended period of co-existence with other language communities. Languages from three families converge here: Indo-Aryan, Dravidian and the Munda branch of Austro- Asiatic. The linguistic complexity of the situation has been muddled by the indiscriminate application of the terms Gond and Gondi to all people who inhabit Gondwana, including Bhatras, Halbas, Kolams, Parjas and Gadabas. The first two groups speak Indo-Aryan languages, the next two speak Dravidian languages, and the last, the Gadabas, includes two subgroups, one speaking a Dravidian,the other a Munda language. Moreover, the Dravidian languages in this area belong to two subgroups: South-Central Dravidian, which includes Telugu and Gondi, and Central Dravidian, which includes Kolami, Parji and Konekor Gadaba.

Central India has witnessed political struggles for nearly two millennia. Different groups have successively ruled the territory in which Gondi is spoken, and have brought with them such languages as Hindi, Marathi, Telugu and English, which functioned more or less as administrative languages within the area. This linguistic contact has naturally left its mark on Gondi. With India’s independence and the creation of states along linguistic lines, Gondi, as a non-official language of the Indian Union, came by legislative fiat to be spoken in four different states. Today Gondi speakers live in eastern Maharashtra, southern Madhya Pradesh, southwest Orissa and northwest Andhra Pradesh whose respective state languages are Marathi, Hindi, Oriya and Telugu.

These factors contribute to the diversity of dialects, as well as the lack of a recognised standard variety of the language. Only half of those who identify themselves as Gonds speak Gondi. Conversely, certain groups of non-Gonds speak Gondi as a second language. Of these latter, the most striking group is the Pardhans, a band of singers and reciters who preserve traditional Gondi oral literature.

Though some word lists predate it, Hislop’s (1866) study is the first to describe Gondi as a separate language. Grierson’s (1906) linguistic survey records several varieties of the language, but an adequate dialect survey is still lacking. More recently, linguists have described or identified several dialects, but have yet to catalogue all varieties of the language. The isoglosses that differentiate these dialects are phonological, morphological, syntactic and lexical. Perhaps the best known are the Adilabad and Koya dialects, spoken in the Adilabad District in northwest Andhra Pradesh. P.S. Subrahmanyam (1968) describes and analyses these dialects, while Tyler (1969) describes the Koya dialect. Another dialect, based on a preliminary survey, is the Maria dialect spoken in the Chanda District of neighbouring Madhya Pradesh. Others spoken further afield in Maharashtra and Madhya Pradesh include the Betul and Dorli dialects. One further dialect, and the basis of this chapter, is Muria Gondi. Studied intensively by Andres (1977), Muria Gondi is spoken in the Bastar District of Madhya Pradesh. All dialects of Gondi appear to cluster into one of two sets on the basis of shared features: one set is spoken towards the northwest of the Gondi-speaking area, the other towards the southeast. Muria Gondi belongs to the latter set of dialects.

The Muria Gonds live on a plateau in the Bastar District. Their settlements usually stand atop ridges and their rice fields lower down the slopes. One social institution that distinguishes the Muria Gonds from other Gond communities (such as the Raj Gonds of Adilabad in Andhra Pradesh) but unites them with certain Munda-speaking groups, is the gottul - a dormitory in which unmarried young men and women live. Muria Gond clans are divided into two moieties, called par, probably from *pätu ‘share, lot’ (< DEDR 3852 patu ‘befall’, DEDR 4065). Membership of a moiety determines how Muria Gonds address each other and whom they may marry. For example, members of a single moiety address other members of that moiety with kin terms rather than names. Further, the different families in a moiety may share a religious centre, called a manda (DEDR 4777 *manru ‘assembly hall, court’, a nominal derivative of man ‘be located’, DEDR 4778).

Central Dravidian

The Central Dravidian (CDr) languages comprise four non-literary languages: Gadaba, Kolami, Naiki and Parji. These languages are spoken in a belt that runs from northwestern Andhra Pradesh up through the Bastar District of Madhya Pradesh into western Orissa. The languages are not spoken by large numbers of people; Kolami is spoken by approximately 90,000 people. Most of these languages are known from word-lists or a single grammar. Gadaba, also known as Ollari, Konekor Gadaba and Poya, is spoken in the Koraput district of Orissa and the Srikakulam District of neighbouring Andhra Pradesh. It is unusual in having been described in two grammars, Bhattacharya (1956)and Bhaskararao (1980). Lexical stems, both nouns and verbs, in this subgroup typically exhibit patterns of alternation between two stems, related by anaptyxis, so that the Gadaba verb idig ‘get down! ’ is opposed to the basic form idg ‘get down’. In Kolami, the noun stem tedp-ltedep- ‘cloth’ also exhibits this alternation: monosyllabic tedp- ul ‘cloths’ contrasts with disyllabic tedep-t ‘in the cloth’ and tedep ‘cloth’. As Bhaskararao (1980) observes, when inflected, the basic forms reveal consonant and vowel harmony operating to a greater extent than elsewhere in the family (Pengo, for example, has limited consonant harmony). Even though both Gadaba and Kolami have vowel anaptyxis, they differ in the rules that determine the identity of the anatyptic vowels. The Central Dravidian languages have not lost the perfect tense auxiliary verb man ‘be’, nor do they have rules that permit its truncation and deletion.

Finally, Kannada exhibits diglossia (Nayak 1967), although the differences between the two varieties appear not to be as prominent as in, say, Tamil. Examples of diglossic variation are given below in the finite verb paradigms. The variety of Kannada presented here, Standard Colloquial Kannada, reflects the speech of educated speakers in and around the two southern cities of Bangalore and Mysore. This standard is historically based on the speech of brahmin Kann- adigas. Though it differs from Standard Literary Kannada, it remains somewhat closer to the literary form than, say, the northern and northeastern dialects.

Kolami

Background

Kolami is spoken by 80,000 people who live near Wardha and Kinwat in Maharashtra and in the Adilabad district of Andhra Pradesh (1981 Census). The language comprises three dialects: Adilabad (A), Naikri (N) and Wardha (W). The speakers call themselves kölavar (W)/kolavar (A) ‘Kolam’ and their language kölav päna (W)/kolava gotti (A) ‘Kolami’. Most Kolams also speak Marathi, a fact that has naturally influenced the structure of Kolami. While Grierson (1906: 561-5) first recognised Kolami as a distinct Dravidian language, others had noted its distinctiveness earlier. As a Central Dravidian language, Kolami is most closely related to Gadaba, Naiki of Chanda and Parji.

Authoritative information on Kolami appears in Emeneau (1955), the standard work on the language. Emeneau gathered material in 1937-8 from Kolams of Mandwa village nineteen miles from Wardha; hence, the name Wardha dialect. Besides a description of the Wardha dialect, Emeneau (1955) includes a summary of Sethudmadhava Rao’s 1950 sketch of the Adilabad dialect, along with notes on the Kinwat and Pändharkavra dialects, based on materials collected by Burrow and Bhattacharya. It also addressses the comparative position of Kolami within Dravidian, and provides texts and a vocabulary list that includes cognates from other Dravidian languages and Indo-Aryan loans from Marathi and Hindi. Bright (1956) adds lexical items elicited from Kolams from Sun- gapuram village in Adilabad district.

Further material on the Adilabad dialect was collected by P.S. Subrahmanyam during field trips in 1965 and 1975, but this remains unpublished. K. Tho- masaiah (1986) reports material on the Naikri dialect which he and P.S. Subrahmanyam gathered in 1984-5 from inhabitants of Sinna Böddi village, four miles from Boath Road Station near Kinwat. The present chapter is based on Emeneau’s (1955) treatment of the Wardha (W) dialect. Minor differences of transcription are introduced: use of a macron for long vowels, and use of j and j for Emeneau’s z and j, respectively. Important features of the Adilabad and Naikri dialects are noted where pertinent.

Gadaba

Background and History

Speakers of Gadaba live in a continuous area that traverses the north of Andhra Pradesh and the southwest of Orissa. The word ‘Gadaba’, sometimes spelt as ‘Gadba’, is a term that outsiders use; the speakers themselves use two names, konekor and mundli. Of these two, mundli is probably the older and its etymology is unknown. In Gadaba konekor means ‘hill people’. To confuse matters somewhat, Gadaba is also applied to another group living in the same area: this group speaks a language called Gotub Gadaba. Gotub belongs to the Munda branch of the Austroasiatic languages, and has no connection with the Dravidian language. konekor gadaba has, in the past, also been called Poya in the literature, but this term is now avoided.

The earliest systematic study of this language is Bhattacharya (1957); in this work and some others, the language is called Ollari. Burrow and Bhattacharya (1962-3) note that Ollari and Konekor Gadaba are two local variants of the same language. While the Census of India does not distinguish between Konekor Gadaba and Gotub Gadaba, the size of the Konekor-speaking population is estimated at a few thousand.