THE STATUS OF NIHALI [ON NIHALI] Norman Zide 3/95 University of Chicago The interest in Nihali, such as it is, in certain narrow academic quarters, lies in the fact - possible fact - that it is (in interesting ways) no language at all but a `so-called' or seeming language, and/or that it is a mystery, a lost - possibly `paleolithic' - language (something like the Tasaday of the Philippines, what Tasaday was purported to be but without the heavy public relations flak that surrounded it). It is, perhaps, the only remnant of an ancient - pre-Munda, pre-Dravidian, pre-Indo-Aryan language family, with no living relatives, but perhaps a sister language of the language the Bhils spoke before they lost their own language and it was supplanted by the various Indo-Aryan `Bhilis'. Nihali has been noticed by historical linguists for the very high percentage of borrowed vocabulary, and the variety of (proposed) sources for that borrowing, and the `suspiciously simplified' syntax of the language. What is a mystery academically and semipopularly can be an administrative headache. What early notices of the Nihals we have describe them as as nuisances, hill marauders and plunderers, `caterans' who were `incorrigible', and needed to be exterminated, and almost were on a couple of occasions. (I use Mundlay's spelling, Nihali, which represents the local pronunciation; Kuiper and others write `Nahali'. Berger's paper goes into the history of the name. The name the Nihals use for themselves is Kalto or Kaltu.) It is due to the work of Professor F.B.J. Kuiper that Nihali has been brought to the attention of Indologists, and what we say here addresses matters that Kuiper has been the first to foreground, and to treat in impressive detail. Nihali has been referred to several times as a `so-called (sogennante)' language or something similar by Koppers, Konow, Kuiper, and even Fuchs expressed doubts about the language. The new many-volumed epitome of gazetteers & tribes-and-castes compendia, People of India (being issued by the Anthropological Survey of India) in its ninth volume, Languages and Scripts, recognizes (and, on the strength of its own invesigations?, finds) the `Nahals' as speaking - in different regions - Nimari or Korku (but not Bhili), but there is no mention of a spoken Nahali language. What is defective or `so-called' about Nihali? Why is it not just a language, comme les autres? For Kuiper it is an argot (of what? or in what multilingual package?), and he talks of gaunersprachen (secret languages used by criminals); Koppers seems to doubt that it is a full (complete) native language, the first language of anyone, the assumption being that all Nihali-speaking Nihals (a small minority of those identified in official records as Nihals) are bilingual, their other - full - language presumably being the North Munda language Korku. Kuiper reports Koppers' mention of a collection of texts collected by Koppers and Fuchs, but Fuchs in a recent book where he presents a considerable amount of information about the Nihals and says something about their language makes no mention of any text collection. (Mundlay found and worked with Nihali monolinguals as well as bilinguals, and did collect texts. It is, as of the sixties and perhaps still, a first language, a home language, and most likely a `full language', however we choose to define such a phrase. We will come to that later.) Konow was responsible for the data on the Munda languages and Nihali (in volume 4 of the Linguistic Survey of India (LSI)) and Koppers quotes his 1908 article: (that there is)"... one tribe, the so-called Nahals of Nimar, who were stated to speak Kurku in 1870, but who now speak (c 1908) a mixture of Munda, Dravidian and Aryan dialects', - presumably our `Nihali', some variant of it. The evidence for their speaking Korku and nothing else in 1870, in any case, is shaky. The progress from monolingual Korku to some `Nihali' seems unlikely. `Nihali' has been in and out of the roster of Indian languages several times. Now you see it, now you don't . Discussions of Nihali presuppose conjectural histories of the `language', so that, for instance, it is not clear that calling it an `argot' refers to present day use of Nihali usage (if Mundlay's data on monolinguals are accurate, and I think they are, then at least for some group(s) of Nihals Kuiper is mistaken) or to some earlier stage (`argotization'??) in the formation of Nihali. Is an argot `stage' recognized in the formation of some component of other (`full') languages? If Kuiper is refering to phonological deformation and `mutilation' - and he does talk of mutilation - does `argot' have something to do with `pidgin'? Reasons adduced for doubting that Nihali - if there is one or a set of closely related dialects that are being consistently referred to in these publications - is a full-fledged language, and the first and /or only language of any speaker are the following: 1) the common association of Nihali-speaking Nihals (the estimate as of 1963 of the Nihal population was c 25,000, of which (Mundlay's estimate) perhaps ten percent spoke Nihali) with Korkus and Korku villages. Presumably elswhere in Nihal territory Nihals (i.e. deCandolle's Zones 2 and 3) they did not speak or know Nihali. (I doubt if anyone has made a careful investigation of Nihali language competence and use throughout the area. Mundlay has useful information of the Nihals she surveyed in the Melghat region.) Korku villagers I worked with when asked about the Nihali language (Mundlay assured me that there were Nihali-speaking Nihals living in the village) told me that the Nihals had no language of their own; they spoke Korku. The few extended descriptions of the Nihals are in books (Fuchs (1988), Hermanns, Koppers) primarily concerned with other groups: the Korkus or the Bhils, and this seems to be the characteristic angle of encountering and viewing the Nihals, when seeing them as as anything but a source of civil disturbance and disruption. The few exceptions, papers addressing primarily Nihali matters, include the much quoted brief piece - the one source known to adminstrators or scholars having to find out something to the Nihals - in the Tribes and Castes of the Central Provinces (1916), and de Candolle's paper). 2) The (more than) apparent secrecy about the language and the general ignorance of its existence suggest to some that it isn't a language, but an argot or jargon used for certain limited purposes, the real language of these people (in Melghat/Nimar) being Korku, or in other areas - is the argot there completely gone? - some form of Nimari - or Bhili or Hindi or Marathi. 3) More interesting is the judgment that the language is limited, defective, perhaps a broken down descendant of an earlier `full language', or a mixed language, and not adequate to the usual needs of linguistic communication. Thus the need for Korku, or some other `full' language. 4) An examination of the structure of the lexicon has led Kuiper to suggest that in fact Nihali is an argot (see below). But the information on which all these generalizations have been made is limited. Shafer and Kuiper independently exhumed Nihali from the brief description in the Linguistic Survey of India and observed that it was not - as Grierson/Konow implied (although not without reservations) - a North Munda language, probably closely related to Korku, but something else. The LSI has a few pages on Nihali. Bhattacharya on one short field trip collected a small amount of material on Nihali (see his article), but the further field trips to the Melghat area he intended (personal communication, S. Bhattacharya c 1966) were not allowed by his superior, the then Director of the Anthropological Survey, Nirmal Kumar Bose. Kuiper's thorough study of Nihali of 1962 was based on the (limited) Nihali materials of the LSI and Bhattacharya. That's all there was. All the surmises about the status of Nihali can be shot down by new and better observational data, and Mundlay provides some of that. NIHALS AND BHILS, NIHALI AND `BHILI'. For the connection of the Nihals with the Bhils, see Koppers (1948), and also Kuiper (1962) and Fuchs (1988). Koppers quotes Campbell (1880) who wrote that the Nihals `are the most savage of the Bhils', but this, Koppers, says is not to be taken literally, the Nihal problem is complicated. (Koppers has a few notes about the Nihals in this book and in his Geheimnisse.) There is an extensive ethnographic literature on the numerous Bhil groups, and something, but much less, on the language(s). The `Bhili language' is, apparently, a number of Indo-Aryan dialects of the regional languages in the extensive area of Bhil settlement (Gujarat, Rajasthan, Maharashtra, Madhya Pradesh - see Koppers' map)). In some regions Nihals have been long associated with Bhils, have lived with them and still do (see de Candolle). (There are no - or few? - Nihali villages; Nihals live in sections of Korku, Bhil, and other (which?) villages.) Fuchs also mentions Bhili-speaking Nihals as well as Korku-speaking and Nihali-speaking Nihals. The pre-Indo-Aryan language the Bhils probably spoke is lost, and we don't know its genetic affiliation. West central India is almost entirely Indo-Aryan speaking now. Presumably other linguistic families were more strongly represented in these areas in earlier times. There are other - fairly large - groups in central India, the Baiga for one, who now speak a variety of the local `Hindi dialect', but who probably had their own, non-Indo-Aryan, language earlier. That `Old Bhili' was related to `Old Nihali' - that there was an ancient Nihali-Bhili family - is a plausible surmise (this was suggested by Koppers and by Shafer and accepted by a number of others), but as yet there is no linguistic evidence for it, and I have seen no strong claims based on ethnographic materials to support the case. I examined one lexicon of Bhili - Thompson (1895) - and found no vocabulary cognate with the Old Nihali vocabulary identified by Kuiper, Shafer and myself in the data Kuiper used and in and Mundlay's data. DRAVIDIAN. TIBETO-BURMAN. Most of the Dravidian cognates adduced by Kuiper, Shafer and Bhattacharya seem plausible. Pinnow in his review of Kuiper summarized his (Kuiper's) material on Dravidian influence: there are four strata (schichten) - (Kuiper speaks only of sources of Dravidian words, not strata) with c 47-50 examples (9 per cent of the total) If borrowing from Kurukh - one of the four strata - is relevant to the history of Nihali, it may be possible to date (approximately) some of the borrowing, given a hypothesis about the location of the Nihals at the time of known movements north of the Kurukhs (Oraons).) Burrow wrote a short notice of the book, but had no comments on the Dravidian material. The review provides a concise description of Kuiper's intentions, materials, conclusions; the only comment he allowed himself was to remark Kuiper's `considerable reserve' - resistance - to accepting Nihali - the lexical remains after the borrowings have been extracted - as `a language which is in origin quite independent' (of Indo-Aryan, Dravidian, Munda and Austroasiatic, Tibeto-Burman), which to Burrow `seems most likely', and which,as I too see it, Kuiper's monograph makes a good case for. One would like a Dravidianist to go over the entire corpus (i.e. including Mundlay's material) and comment on the whole picture. I am no Tibeto-Burmanist but the Tibeto-Burman (possible) cognates - of Konow and Shafer and Kuiper - I find less convincing. Here too, much new material - data and analyses -is now available, and it would clarify several of the issues if someone familiar with it and the other Nihali-related material were to reevalulate the Tibeto-Burman connection. Certainly a few of the forms (Nihali sunum `oil', from Korku and North Munda sunum) seem to have related forms in Tibeto-Burman, and there are certainly old Proto-Munda (and Proto-Austroasiatic) loans - whichever way the borrowing went (e.g. PM *kuXla `tiger'), but, for instance, the geographical information Kuiper provides on the proximity of a Tibeto-Burman-speaking group, Limbu - `not greater than about 130 miles', to (present day) Santal groups may not be relevant to Nihali although it is to North Munda since there is no reason to think that the North Munda connection (borrowing or whatever) was not primary, and that Nihali borrowed the form - as it did so many - later, from Korku. Kuiper mentions Konow's views on `complex pronominalized Himalayan languages' and a Munda substratum as a contributor to their formation. Konow's views on these languages - and the Munda substratum - are not accepted by Tibeto-Burmanists today. Kuiper finds the Tibeto-Burman connection to be `the most puzzling problem' of Nahali contacts, but, with caveats, goes on to find grammatical morphemes in Himalayan languages as (possibly) connected with Nahali morphemes. AUSTROASIATIC (apart from Munda). This discussion of the possible connections of Nihali with Austroasiatic is based on material in Kuiper (1962) and on various papers of Pinnow's, which are also taken up in some detail by Kuiper (1972). The linguists who have done considerable work on Austroasiatic (primarily Mon-Khmer) in the last thirty-five years or so, i.e. Shorto, Diffloth, Ferlus have had nothing to say about Nihali, probably because they don't find it to be (interestingly or at all) Austroasiatic, if in fact they find it. The identification of West (Munda) and East (Mon-Khmer)) Austroasiatic cognates in general (of course some languages have undergone more obscuring sound changes) has not been difficult. That the establishing of plausible Nihali cognates - the paucity of data making things that much harder - has been difficult and uncertain could be a result of several states of affairs, one being remoteness of relationship. Pinnow (1963) proposes a provisional (`the present state of investigation' of the position of Nahali does not permit any definite judgment.')Western group of Austroasiatic languages which he calls Nahali-Munda, Nahali (now definitely judged not to be Munda) being Western N-M, and Munda being eastern N-M. He writes that the classification of Nahali is `particularly difficult', in large part due to lack of data. The more interesting and difficult to explain connection of Nahali with Munda has to do with its morphology (`Its morphological system ... is obviously connected with that of the Munda languages.' (See details in 1966a, and some remarks on Nahali verb morphology and its implications below under `What Kind of a (Contact-Shaped) Language is Nahali?')). Kuiper quotes Pinnow's 1963 conclusion `We may perhaps come closest to the truth if we assume that Nahali possesses an isolated non-Austroasiatic stratum that has been partially replaced by an Austroasiatic stratum which has also provided Nahali with its inflection.' I would ask why `Austroasiatic' here should not be replaced by `Munda'. In 1965 in the Austroasiatic pronoun paper Pinnow writes (again I am quoting Kuiper (1972) `the personal pronouns of the disputed language Nahali can be classified with those of the Austroasiatic family, even though they are rather markedly distinguished from the personal pronouns of the other groups.' The Nahali pronouns don't look like Munda pronouns, and Pinnow finds a few similarities of individual Nahali pronouns with forms of similar meaning in Austroasiatic languages, e.g. Khasi. I don't find these miscellaneous similarities indicative of genetic relationship, and Pinnow himself expresses doubts in the paper. But in his 1966b review of Kuiper Pinnow finds himself increasingly persuaded of Nahali's fundamental Austroasiatic character ("Der grundlegend austroasiatische Charakter des Nahali schalt sich so nach und nach immer mehr heraus "). Kuiper writes that `my provisional attempt at an analysis of the case-endings and the pronouns did not confirm this assumption of an Austroasiatic provenance'. I agree with Kuiper in finding little evidence of Austroasiatic provenance. Kuiper's `central problem' in 1962, `how we must conceive the relations between that oldest Austro-Asiatic stratum and the other unidentified component of the language' should perhaps now be decentered. MUNDA. Apart from the numerous (transparent) borrowings from Korku what has Nihali borrowed from Munda, or Munda from Nihali? First, of course there is no assurance that all the Korku borrowings have been identified. And in the absence of sufficient possibly cognate vocabulary, no setting up of sound correspondences (Nahali-Kherwarian, or Nahali-South Munda) is possible, so that one goes by one's own intuitions about relations of words - in one's own style of negative capability. Examining the sets of words on Kuiper's page 39, `A. More closely connected with North Munda (Kherwari)', and `B. More closely connected with Central and South Munda' I find several of the seven items in A. unacceptable or implausible, most importantly te- (Mundlay t'e-) `to eat', which does belong here, but in set B. Of the items in B, the word for `father', a-ba, is pan-Munda, reconstructible - and not a loan - for Proto-Munda. The most interesting - and to my eye the most solid - forms are be- `to give' (Mundlay b'e-), er, ier- `to go', piy- `to come', and t'e-. Although we have only these four words, the connection here is more persuasive to me than anything in Kuiper's Munda alignments and the claims of cognation that go with them. These four do have good parallel forms in one or another branch of South Munda, and apparently no related forms in North Munda. I don't accept Kuiper's Santali atin as likely to be connected with Nihali te-. (South Munda (SM) branches into Kharia-Juang (KJ) and Koraput Munda (KM). Koraput Munda branches into Gutob-Remo-Gta? (GRG) and Sora-Juray-Gorum (SJG). The actual forms and their antiquity - subfamily membership - will be discussed elsewhere. We give here rough reconstructions: GRG *bEd- `to give', GRG *p+ng- `to come', KM *tej- `to serve food', KM * +r/er- `to run, jump, move'.) What do we make of this? The critical question - crucial to a hypothesis of South Munda subfamilies having borrowed from Nahali words that have no congeners in Mon-Khmer - of Austroasiatic cognates for these SM forms is as yet unanswered. That Nahali could have come into contact with South Munda languages is not at all unlikely. Certainly, some of the Koraput Munda languages (e.g. Gutob Gadba, now spoken only in Koraput District, Orissa was spoken further east, in Bastar, c seventy-five years ago, if not more recently), and we have no realistic notions of where and how the Nahals earlier ranged or came from. What sort of contact situations between what sort of groups, Munda and Nahali, could have resulted in the borrowing of basic lexical items? If the power of the Nihals earlier (as it was some time later) was military, their success as marauders (like the (SM) Remo (Bonda) today but in a smaller way, contained by the local district administration, - or like the Comanche and other Plains Indians) what sort of linguistic impression would we expect the marauders' language to make and under what conditions? Was there (intermittent?) occupation of the raided territories? Intermarriage? The claim that some of the SM languages (Juang, for one) spoke something else before they adopted the ancestor of the Munda language they now speak is not new. That `something else' could have been Nihali, or a sister language of Nihali. There are many possible scenarios to account for the lexical similarities (borrowing, presumably), but I want to affirm the importance of the identification (by Shafer and Kuiper) of these forms; they are less questionable and (therefore) more important than the other putative linkages proposed. In Kuiper's discussion of Nahali and Austroasiatic he writes `... the circumstance that the non-Kurku elements of the Nahali vocabulary cannot be attributed to any one of the sub-groups would seem to point to the conclusion that the older Munda stratum in Nahali stands somewhat apart from the sub-groups into which Munda is divided. Berger arrived at the same conclusion.' If the Kherwarian similarities can be discounted, and I think they can, and the A and B sets are revised and realigned as proposed above, then perhaps (the corpus is still too small, but we can perhaps find more forms supporting this hypothesis) it is precisely one subgroup, South Munda, or perhaps some sub-family or subfamilies of South Munda that show(s) these lexical relationships, and it is South Munda (the SM family and/or one or another of its subfamilies) that has the connection with Nahali. I suggest that SM or KM has borrowed from Nahali, and Kuiper's and Berger's conclusion is wrong. ARGOT. In 1962 Kuiper writes `In the case of Nahali, it is true, there are no certain indications of an analogous origin (he has been talking in the previous paragraph of metonymy and mutilation in speech disguise in various secret languages of the subcontinent) of the names of parts of the body, etc., which categories are also in Nahali etymologically unexplained. Still, it may be useful not to forget that some of the obscure Nahali words may also belong to an argot, and need not necessarily date back to a linguistic pre-history of India.' This is an interesting and useful - warning. He mentions jiki `eye' as perhaps connected as connected with Santali jhiki miki, jiki miki `splendid, resplendent, shining, radiant' - and also notes Ainu shik(i). (Kuiper in his earlier work on Proto-Munda words in Sanskrit brings in echo pairs of this sort, none of which are, as yet, reconstructable for Proto-Munda). In his 1972 review he writes (in a discussion my inadequate treatment of Nihali in survey article on the Austroasiatic languages of India): (Zide's) `observations contain nothing new except the confirmation that Nihali is actually an argot, as had been suggested (by Kuiper) in 1962'. I did not think or say that Nihali was an argot, but that it was likely that `Nihali was used as a more or less secret language'; Navajo was used during World War II by the United States military as a `secret language' because it was unintelligible to the enemy, this did not make it an argot. It seems possible that some of the obscure Nahali words may belong to an argot, but there are - as Kuiper shows - other reasons for obscurity. Despite the interesting and not irrelevant discussion of gaunersprachen I see no good reason (the jhiki miki forms don't convince me) to claim that Nihali is an argot (now? at which previous stage? all of it? some section of the vocabulary? which?). It may be that the phonological distortations, lexical substitutions (rhyming slang, etc) found in (other) secret languages are responsible for some of the Nahali vocabulary - certainly `some of the obscure Nihali words' may be argot, i.e. the result of speech disguising transformation and substitution, but this is something suggested here, and in no way demonstrated, and if it was something like rhyming slang (as in Cockney) there would be no way of retrieving the baseforms, and thus of proving that there was, in fact, this sort of distortion. Kuiper's reasons for proposing his argot hypothesis seem to be, first, the social position and criminal activities of the Nihals (which don't guarantee their possession and use of an argot), and, second, certain speculations about a few words in the old Nihali lexicon. I find the case unproven. I learn from Hal Fleming about `jargons' is small, low status hunting and gathering groups in East Africa where a small stock (c 40 words) of `jargon' has been recorded and the casual conclusion drawn from this short vocabulary is that the language `is a jargon'. One needs to see how much and what segment of the lexicon is (speech disguise-derived) `jargon' and what else `the language consists of'. Kuiper's case for an `argot', more explicitly, is the following (1962, pp 11-16): he first takes up the low status of the Nihals as a `despised social group' and notes that other low status groups in India have secret languages. He then introduces various kinds of phonological `mutilation' found in such sectret languages. He notes that words for body parts are commonly replaced in secrtet languages by disguised forms, and goes on to gives the sources and deriations of some of the replacement forms, and, a bit later, suggests that Nihali jiki `eye' might perhaps (originally) be a descriptive term. All of this is suggestive, but hardly probative, and I don't find it persuasive. In 1972 he is more positive about the argot hypothesis, and adduces some material (e.g. on Vedda) that might suggest analogues for what happened to Nihali, but again with nothing closer to a proof. The quest for Nihali seems in some subsubtext to reveal a plot, one that Professor Kuiper most probably did not intend and would not accept: the voyage to Nahali as the grand occasion for wide-ranging and thoprough exploration of Indo-Aryan, Dravidian, Austroasiatic and Tibeto-Burman. Then, the winnowing of the Nihali lexicon, and the extraction of contact-derived matter. What is left is a small cache of semi-precious Old Nihali words, but this does not satisfy . A second voyage, on the Argo(t) - the golden fleece was plastic after all - leads to the discounting and discarding - throwing overboard - of some of that old Nihali vocabulary. Apart from these substantial, scholarly souvenirs de voyage - what is left of `Nahali'? More, I think, than the argot hypothesis seems to allow. Pinnow (1966b) agrees with Kuiper on the importance of the argot hypothesis (`Der Hinweis auf den m"oglichen Argot-Charakter des Nahali ist eins der wichtigsten Ergebnisse der Arbeit Kuipers, und seine Bedeutung kann nicht genug unterstrichen werden.') We need to distinguish a `functional argot' - i.e. the use of a (secret) language for concealment, from an `argot' (secret language) formed through processes of word-deformation, substitution, etc. (see Guiraud and Mehrotra. Mehrotra isn't aware of Kuiper's monograph. Kuiper is not aware of some of the earlier material mentioned in Mehrotra, e.g. Sleeman on the language of the Thugs. The two discussions and bibliographies taken together provide a good survey of secret languages in India through c 1966.) The parallels with Sri Lankan Vedda and Rodiya are interesting (Vedda - Kuiper quoting de Silva `is a creole based on an older Vedda language with Sinhalese as the second contributing factor', (Kuiper) `Rodiya is a secret language in which non-Sinhalese items are used in Sinhalese structures'). Can (our) Nihali be a creole based on an older Nihali language? Individual factors and contexts may be shared by Nihali (under various conjectures) and Rodiya, Vedda, etc. But, as Kuiper's data show, none of these cases is closely parallel to the Nihali situation. Nahali exhibits a wide range of linguistic contacts, many more than were available to Vedda or Rodiya. What the time scale is - in any of these cases - is still unknown. For Nahali, we assume that there was considerable mobility in a fairly extensive multilingual territory, so that such partly similar contact situations as that of Brahui or Vedda or Rodiya with massive borrowing (or deformation?) but less extensive linguistic contacts are only partly similar. Perhaps the language of the Thugs should be more closely examined. The Romani (Gypsy) sociolinguistic situtations - one or more of them - seem more like what the Nihali situation(s) may have been, but for Romani we know where the people came from and, roughly, when (it is relevant that earlier speculations about the Gypsies posited a much more ancient exodus than the actual one scholars later ascertained), and we know their original - pre-exodus and pre-wandering - Indo-Aryan language and a fair amount about the languages they came in contact with, whereas for Nihali the ancestor language is presumably unattested outside the (obscure) Nihali lexical corpus, and some of its possible contacts - as proposed by Kuiper - have yet to be more firmly demonstrated. And as Kuiper has shown the proportion of borrowed vocabulary in Nahali is very high, presumably much higher than in any of the Romani dialects (what the corresponding figures for Vedda are I can't say.) As with most everything else about the Nihals, we know little about their social or occupational history. They do not now or in the recent past own and cultivate fields (and there is no evidence that they practised slash and burn agriculture)or cattle. In the Melghat they seem to have been associated with the Korkus (themselves known earlier as freebooters, but now settled agriculturalists), but we don't know how far back the connection goes. They were probably hunters and gatherers, and did more and less raiding of neighboring sedentary communities. Fuchs mentions that they are skilled trainers of dogs, and this skill is appreciated by the Korkus. As `caterans' one wonders about their mobility. Did they have horses? They don't now, and neither do the Korkus, although (see Fuchs) there are representations of horses commonly on Korku wooden funerary tablets, and a taboo on eating horsemeat. Horses in that area would be expensive to keep, not particularly practical - bullocks are at least as efficient as ploughing and cart animals, and more docile, and healthier in that terrain. The word for `horse' (see Kuiper) is ma v, which Kuiper connects with Dravidian (e.g. Telugu ma vu) and possibly Tibeto-Burman and Tai forms. The Korku and other Munda forms (old borrowings) are not related. Hermanns has origin stories from Nihali informants according to which in earlier times the Nihals and Korkus were one people. (Not surprisingly, Korku informants deny this.) This means, I assume, that they (all) then spoke a form of Nihali. Later the Korku upgraded themselves (giving up beef, certain low occupations, etc) and, presumably, adopted a North Munda language from high status invaders/immigrants to the region. The Nihals were downgraded, and the Korkus have maintained the social distance (see Fuchs for Korku-Nihali interactions). There are references to Korku presence and activity in the fifteenth century, and to the Nihals `at the time of Akbar', i.e. the latter part of the sixteenth century, in both cases as hill robbers and freebooters. See Fuchs (1963) also on the antiquity of the Korkus (and, by implication, the Nihals who, these writers would claim, probably were in there earlier if in fact they were a different group) in the region, and that of the other North Munda (linguistic) groups in Bihar and adjacent regions. There are, of course, the usual putative identifications with peoples mentioned in the Ramayana (as ra ks asas - demons, see Fuchs (1988), but also Zide (1972) on Khara.)The social position and the marginal occupations of the Nihals suggest that they may well have had and used a secret language, early and late. Information on this may still be obtainable. That this is an argot (or that they use an argot - and what relation that argot has with `Nihali') has yet to be proved. WHAT KIND OF A (CONTACT-SHAPED) LANGUAGE IS NAHALI? A much more thorough treatment of this topic is called for, but I offer here one possible scenario showing schematically how Nihali may have come to its present state. (The data on the Nahali verb can be found in the Linguistic Survey of India, Pinnow (1966a), Bhattacharya and Kuiper (1962) as well as in Mundlay and Lynch, sources that were not available to Kuiper but that don't describe a system that is significantly different.) An examination of what Kuiper meant and might mean by `argot' might introduce various linguistic - sociolinguistic -historical scenarios and tentatively try to place various statements and implications of Kuiper's in such a scheme. A preliminary attempt at doing that (all of the assumptions and stages are arguable): 1) the (Old) Nahali language - not Austroasiatic - was spoken (where? - perhaps in west central India; when?) as a first language by a (perhaps nomadic) group, probably not agriculturalists, and probably not pastoralists either. These people may well have been bi- or multilingual. It was a representative of a family no longer found in India (apart from the words inherited from that older lexicon in modern Nahali); there may well have been related languages at earlier times. 2) In the course of wandering in or to the eastern parts of central India (I won't try to break this down into ordered stages) there was borrowing from South Munda, i.e. South Munda from Nahali (see the words discussed in the `MUNDA' section) and probably from South Munda into Nahali (not that can we identify which is which with much assurance). The Nahals may have been more powerful at this time, more dominant - powerful as raiders, and conceivably more technologically advanced in some ways (although this seems less likely) than the South Munda groups (but not in agricultural techniques?). The morphosyntax of Nahali, whatever it was - I am assuming not much of it is now left and/or identifiable - at this stage, whatever `natural' changes it may have undergone, is fairly intact. 3. Some borrowing from Dravidian, massive disruption of the Nihali community (or communities), perhaps the decimation of the community on the orders of local rulers and chiefs (see Fuchs (1988), or earlier traumatic reduction and breakup of the community (due to what? natural catastrophe, i.e. disease??). (Kuiper quotes Forsyth on `the aboriginal races' (having been)compelled to retire to the mountains before the Hindu invaders ... A few remained in the country occupied by the Hindus, chiefly in the position of agricultural serfs, or watchers of the village', a description which, somewhat modified, could apply to the Nihals in the (non-Hindu) Korku villages. Historical speculation about the position of the Nihals, early and late, is constrained - stymied - by the lack of information on where `they' were when.) The old morphosyntax breaks up. Probably the Nahals (some group of them, one that remains, later, `Nahali-speaking') are bi- or multilingual) and lean more heavily on one Other Language for many vital functions, but retain some older Nahali, e.g. at least some of the lexicon. 4. The older language is remade and socially reconfigured - with a `creole-style grammar', this perhaps for (some) use as a secret language, perhaps with argot-style speech deformation in some of the lexicon. The tense/aspect, etc markers - most of the morphemes used - as Pinnow points out (1966a = Kuiper's 1960a) are familiar to the linguist from a number of other languages in the area, but they are not used in the same ways and have different meanings; this doesn't look like a case of `natural genetic' inheritance. In the Gutob (SM) language (and probably, earlier, elsewhere in SM as well) the tense suffixes in the positive conjugation are identical with suffixes with unaccountably (so far) different meanings in the negative conjugation. How this developed is baffling. It makes the language a bit harder to learn, but this seems to be a `genetically natural development', whatever it developed from, whereas the Nahali salad of verbal morphemes seems to show a disruption, a bad break - and to be something else. The language is heavily relexified, Melghat Korku being the lexifier. (Kuiper's figures on borrowing from Korku are misleading in that much - most? - of the borrowing from Indo-Aryan has been borrowing from Korku also, borrowing of words Korku borrowed from Indo-Aryan.) The latter assumption has implications about the antiquity and nature of Korku-Nahali connections. If, anciently, the Korkus and Nahals were one people, speaking an old form of Nahali, and then the Korkus, upgrading themselves, separated themselves and adopted a North Munda language, were the Nahals and Korkus still in contact in the kind of relationship (Nahals as inferiors of and servants of the Korkus) we find today, and which the literature tacitly suggests has been the case for some time? If such was the case, the Nahals would have been - as they are now - bilingual in Korku (and, perhaps, as now, familiar with other local languages as well). This would suggest that the borrowing from Korku started very early, and not, as it looks, fairly late. It is simpler - until someone can show that this could not have been the case - to say that while the Korku-Nahali connection may be old, the heavy relexification dates to the comparatively recent period when the Nahal community was badly broken up, reduced, scattered, and that then some of the Nahals retreated to the Korku area, and recuperated a social organization and a Nahali language. As to possible influences of - direct contacts with - non-Munda Austroasiatic languages, Tibeto-Burman languages - as the above discussion indicates, I have my doubts about these - they can have been acquired in the wanderyears of the Nahals, and clearly the Nahals have moved around. This scenario rules out the `fundamentally Austoasiatic character' hypothesis of Pinnow's. If Nahali has a fundamentally Austroasiatic character, and acquired it in the usual historical linguistic ways, then my scenario is all wrong - unless the connection is remote indeed, in which case we want to hear more from Pinnow about that fundamental Austroasiatic character. On questions of syntax that have not been examined, and not mentioned in this introduction, they can be investigated more closely using Mundlay's material . In this `remaking', was what was not borrowed (at one stage or another) and is retrievable and assignable to `Old Nihali' just a small set of words? We have said nothing about Nihali syntax, and how it compares with Korku or Hindi or Marathi or (some) Dravidian. The SOV word order, the use of postpositions, etc is common to all the languages in the area (I do not speak of Tibeto-Burman). But it is not clear, for example, what relative clause(-equivalents) are like in Nihali, the Indo-Aryan pattern(s) being different from the Dravidian and Munda patterns. Perhaps a closer examination of Nihali will show traces of earlier structures. How `simplified' or reduced' (simplified or reduced from what)? or `creole-like' is Nihali grammar? Pinnow (quoted by Kuiper, who had just noted the absence of Dravidian influence in the (morphology of) the verb) suggests that the verb system is like that of Proto-Munda. I don't see this. Younger historians of the region (they need to know Marathi) might be encouraged to look at the Nahals, and their history and place in history. The currently familiar and congenial problems having to do with colonialism, peasant rebellions, kingship and the discoursing that goes with them probably won't take them very very far with Nihal history, but that would be one more reason to pursue it. If concentrated subalternity is of interest, the Nihals are the subaltern's (Hindi- and Marathi-speakers') subaltern's (Korkus') subalterns. How `other' in the (non-urban) Indian scene can one get? Perhaps some illumination of the linguistic problems will come out of a better understanding of where the Nahals were and what they were doing and saying to whom. The paucity of data on Nihali has not prevented wider comparativists - `long rangers' - from finding (tentative) homes for it. Thus, J. Bengtson (1994) finds a place for Nihali in his Macro-Australic superstock. The most substantial section of this paper, `What Kind of a (Contact-Shaped) Language is Nihali' is much reduced here. It grew too long, and needs to be longer. To take up some of the problems coming out of Kuiper's dense and thorough treatment of the Nihali material available to him and Mundlay's material as well will require more work, much more consideration of the morphology and syntax. It is to be hoped that Mundlay's sketch of Nihali grammar, and papers on the sociolinguistic situation of the Nihals will also be published in the near future. And more results of more research on the Nihals. Bibliography Bengtson, J., note in Mother Tongue, 23, 1994 on Macro-Australic, 73-75. Berger, H., Deutung einiger alter Stammesnamen der Bhil aus der vorarischen Mythologie des Epos und der Purana, WZKSOA, III, 1959, 34-82. Bhattacharya, S., Field-notes on Nahali, Ind. Ling., 17, 1957, 245-258. The Munda languages in ancient India, JIndHist, 55, 1977, 25-33. Burrow, T., Sanskrit and the Pre-Aryan Tribes and Languages, Bulletin of the Ramakrishna Institute of Culture, February, 1958. review of Kuiper's Nahali, BSOAS, 26, 1963, 436-7. deCandolle, B., Contribution a l'etude des Nahals, Anthropos, 56, 1961, pp 750-788 (with three pages of photographs). Forsyth, J., Report of the Revenue Settlement of British Nimar District of the Central Provinces, 1870. The Highlands of Central India, London, 1872. Fuchs, S., The Aboriginal Races in North-Eastern Central India, in Festschrift Paul Schebesta, Studi Instituti Anthropos, 18, St Gabriel, Verlag, Wien-Modling, 1963, 349-355. _____, The Korkus of the Vindhya Hills, New Delhi, 1988. Grierson, G. (general editor), Linguistic Survey of India, Volume IV, 1906 (the article on Nahali (prepared by Konow) is on pp 185-189.) Guiraud, P., L'Argot, Que Sais-je, No. 700, Presses Universitaires de France, Paris, 1956. Hermanns, M., Die religiose-magische Weltanschauung der Primitivstamme Indiens, Band 2, Die Bhilala, Korku, Gond, Baiga, Franz Steiner Verlag, Wiesbaden, 1966. Konow, S., Notes on the Munda family of speech in India, Anthropos, 3, 1908, 68-82. (Also see Grierson) Koppers,W., Geheimnisse des Dschungels, Verlag Josef Stocker, Luzern, 1947. _____, Die Bhil in Zentralindien, Verlag Ferdinand Berger, Horn-Wien, 1948. Kuiper, F.B.J., Nihali, a Comparative Study, Amsterdam, 1962. The Genesis of a Linguistic Area, IIJ, X, 2/3, 1967. review of Current Trends in Linguistcs, IIJ, XIV, 3/4, 1972 (see pp. 292-93). Lynch, Owen, Outline of Nihali Grammar, unpublished typescript, 1963 (?). Mehrotra, R.R., Sociology of Secret Languages, Indian Institute of Advanced Study, Simla, 1977. Mufwene, S., Jargons, pidgins, creoles, koines: what are they?, in Pidgins and creoles: structure and status (eds D. Spears and D. Winford), Benjamins, Amsterdam, 1995. Mundlay, Asha, Nihali (Kaltu-mandi): Its place among the Munda languages of India, unpublished typescript, 1964, revised 1972. _____, A Nihali lexicon, revised Appendix of the above, 1979, 54 pp. Pinnow, H-J., The position of the Munda Languages within the Austroasiatic Language Family, in Linguistic comparison in South East Asia and the Pacific, School of African and Oriental Studies, University of London, 1963, 140-152. _____, Personal pronouns in the Austroasiatic languages: a historical study, in Indo-Pacific Linguistic Studies, Part I, North-Holland, Amsterdam, 1965, 3-42. _____, A comparative study of the verb in Munda languages, in Studies, in Austroasiatic Linguistics, Mouton, 1966, 96-193. (1966a, Kuiper's 1960a). _____, Review of Kuiper's Nahali in OLZ, 1966, 61, 492-6. (1966b) Russell, R.J. and Hira Lal, Tribes and Castes of the Central Provinces, Volume 4, 1916 (Nahal, Nihal 259-261). Shafer, R., Nahali, A Linguistic Study in Paleoethnography, Harvard Journal of Asiatic Studies, 5, 1940, 346-371. _____, Ethnography of Ancient India, Wiesbaden, 1954, see pp 10-12. Singh, K.S. and S. Manoharan, People of India, National Series, Vol. 9, Languages and Scripts, Anthropological Survey of India, Oxford, Delhi, 1993, see pp 189, 325. Thomason, S. and Kaufman, T., Language Contact, Creolization and Genetic Linguistics, Berkeley, 1988. Thompson, Charles S., Rudiments of the Bhili language, Ahmedabad, 1895. Zide, N., Some comparative notes on Gta? pronouns, in Pratidanam, Mouton, 1968, 348-358. _____, Current Trends in Linguistics, Volume 5, Mouton, 1969 (article on Munda). _____, Some Munda Etymological Notes on Names in the Ramayana, Journal of the Ganganatha Jha Kendriya Sanskrit Vidyapeeth, XXVIII, 1972, see p 755. _____, Studies in the Munda Numerals, Central Institute of Indian Languages, Mysore, 1978 (see pp. 71-2).