TheNaga Language Groups within the Tibeto-Burman

George van Driem

The Nagas speak languages of the Tibeto-Burman fami­ Ethnically, many Tibeto-Burman tribes of the northeast ly. Yet, according to our present state of knowledge, the have been called Naga in the past or have been labelled as >Naga languages< do not constitute a single genetic sub­ >Naga< in scholarly literature who are no longer usually group within Tibeto-Burman. What defines the Nagas best covered by the modern more restricted sense of the term is perhaps just the label Naga, which was once applied in­ today. Linguistically, even today's >Naga languages< do discriminately by Indo-Aryan colonists to all scantily clad not represent a single coherent branch of the family, but tribes speaking Tibeto-Burman languages in the northeast constitute several distinct branches of Tibeto-Burman. of the Subcontinent. At any rate, the name Naga, ultimately This essay aims (1) to give an idea of the linguistic position derived from nagna >naked<, originated as a titu­ of these languages within the family to which they belong, lar label, because the term denoted a sect of Shaivite sadhus (2) to provide a relatively comprehensive list of names and whose most salient trait to the eyes of the lay observer was localities as a directory for consultation by scholars and in­ that they went through life unclad. The Tibeto-Burman terested laymen who wish to make their way through the tribes labelled N aga in the northeast, though scantily clad, jungle of names and alternative appellations that confront were of course not Hindu at all. The diverse N aga language any interested reader of ethnographical and linguistic liter­ communities observed their own indigenous religions and ature on the Nagas and their closest Tibeto-Burman neigh­ represented a lineage of cultural traditions entirely dis­ bours in northeastern , and (3) to address the issue of tinct from the Hindu, Jain and Buddhist traditions ofln­ the provenance of the various Tibeto-Burman language dia proper. Other etymologies have been proposed relating communities designated collectively as Naga. the ethnonym N aga to words or expressions in one or other language of the many diverse Naga languages. One such The Tibeto-Burman family vies with lndo-European for etymology proposes that the term derives from an expres­ the title of the language family with the most speakers in sion denoting piercing, a practice prevalent amongst some the world. Numerically, however, most speakers ofTibeto­ of theNaga tribes. Whether or not any such alternative ex­ Burman languages are represented by a single branch. The planations in fact represent the genuine etymology of the Mandarin, Hakka, and the name, or just reflect folk etymologies, they evidently take other Chinese >dialects< account for most speakers as a their inspiration from popular perceptions by outsiders of historical consequence of the southward expansion ofH<'m salient features ofNaga material culture. language and culture from northeastern from the 3rd century BC onwards. Yet Sinitic is just one of many branch­ es. Most of the branches of the Tibeto-Burman language

311 George van Driem TheN aga language groups within the Tibeto-Burman language family

Kazakhstan Tibeto-Burman Mongolia

Uzbekistan Kyrgyztan WestHimalayish G~ ~ Sinitic ... and all languages Chinese Inner Mongolia ~8 Turkestan which can be ~ Tajikistan demonstrated to be 0 genetically related to Tibetan Chinese Burmese Afghanistan these three 8 ~ ~8 ~s~ China G ~a~~ b. Fig. 1. One of the language families identified byJulius Heinrich ~Digarish ~~ Pakistan von Klaproth in his polyphyletic view ofAsian linguistic stocks (1823a1 ~ Chepangic Midzhuish 8Nungtsh 8 . 1823b). He explicitly excluded languages today known to be Kra-Dai or 8 ~ ~G Kareruc e I e e 1 •e Daic (e.g. Thai, Lao, Shan) and known to be Austroasiatic (e.g. M on, ••••• •e ee••• e • Vietnamese, Nicobarese, Khmer). ~e~~ • · 8 e e •• Bangla- e•• • e ~ r::0 o ~s 8 ~ ~ Brahrnaputr•n G~ ~gS desh e Burma family are represented solely in India. In fact, the linguistic India e Karbi Kukish Laos and prehistorical centre of diversity of the language phy­ 8Mru lum lies decidedly within the Indian subcontinent. b. Fig. 2. Major branches of the Tibeto-Burman language family Thailand The Tibeto-Burman family oflanguages was first identi­ (van Driem 2001). fied in Paris by the German scholar Julius von Klaproth in his Asia Polyglotta in 1823. The first explicitly polyphyletic attention back to the centre ofTibeto-Burman linguistic view of Asian linguistic stocks had been presented over a diversity, which lies in the eastern and the In­ century earlier, in 1692, by Nicolaes Wits en, former mayor do-Burmese borderlands. The patch of fallen leaves on the of Amsterdam. Yet by the beginning of the nineteenth cen­ forest floor provides a more informative framework than a tury, enough language documentation had accumulated in false tree, such as the misleading and now defunct Indo­ Europe that the well-travelled and knowledgeable Klaproth Chinese or >Sino-Tibetan< model (van Driem 2003). Each branch of the family shown in Fig. 1 contains between b. Fig. 3· Geographical distribution of the major branches ofTibeto­ was able to identify and distinguish twenty-three Asian one and over forty completely different languages. The Burman. Each dot represents a major subgroup, not a single language. (vanDriem 2002). language families based on his systematic comparison of The metaphor of fallen leaves implies the existence of a tree. representation is schematic and neutral with respect to lexical roots. Some of his families have been augmented, Yet we cannot lift our gaze from the forest floor to see the different higher-order subgrouping hypotheses about the diminished or redefined, but today his Tibeto-Burman tree because we cannot look directly into the past. Instead, branching of the family, e.g. Sino-Bodic, Newaric, Brah­ >N agas< in the past, and other groups of northeastern India model remains the most well-informed model and also still historical comparative work will hopefully enable us to see maputran. The subgroups Sinitic, Bai, Tujia, Qiangic and that sometimes tend to be confused with the Nagas. The represents the most agnostic and thus inherently least con­ the shadows which the branches cast between the leaves on the Ersu cluster lie within the present-day political bor­ most common alternative names for the various language troversial theory of the genetic relationship between the the forest floor. Whether a language family appears to be ders of China. The Bodish, Tshangla and Lolo-Burmese communities are provided in order to aid the unravelling diverse Naga languages and languages such as Burmese, more rake-like or more tree-like is often a function of the branches are represented both inside China and on the In­ of the complex ethnic and linguistic nomenclature used lo­ Chinese, Tibetan, Sherpa, Limbu, Newari and other lan­ state of the art in historical comparative linguistics rather dian subcontinent. All other groups lie within India, Nepal, cally and in the specialist literature. The geographical loca­ guages of the same linguistic stock. than a statement about linguistic phylogeny. It is relevant Bhutan, Burma and . tions of the language communities are also indicated. to note, however, that with the inexorable progress oflndo­ We know more now than Klaproth did. Many previously European linguistic studies, even the twelve branches of The key to understanding the prehistory of this language The groups numbered here from 1 to 8 live to the west or unknown Tibeto-Burman languages and subgroups have this most well-studied language family, once depicted in family therefore lies in India, and the centre of diversity to the north of the N agas. Group 9 comprises the com­ been identified since 1823. In 2001 in Cambridge, I intro­ the pleasing shape of a branching oak, have gradually as­ lies specifically in the northeastern corner of the Subcon­ munities usually called >Northern Nagas<1 even though duced the metaphor of fallen leaves illustrated in Fig. 2 sumed a more rake-like appearance and so come closer to tinent, where we also find the Naga languages. Below are Northern N a gas are linguistically more closely affiliated to (van Driem 2001). The model attempts to identify all the the fallen leaves model. listed the groups in India which are called Naga, some Brahmaputran, i.e. Group 4 1 and perhaps also to Kachin­ constituent branches of the family and draw the focus of neighbouring groups that have been loosely categorised as ic, i.e. Group 17. Groups 10 to 12 are the linguistic groups

312 313 George van Driem TheN aga language groups within the Tibeto-Burman language family

narrowly defined as Nagas, comprising the languages of Tibeto-Burman subgroups and language communities 4b-4 (a.k.a. Lalung)-in the Karbi sb Eastern Tani (a.k.a. >Abor<, Adi, Padam- the Ao branch, the Angami-Pochuri branch and the Zeme Lepcha- and Darjeeling district Anglong or Mikir Hills around Umswai Minyong group) - east of the Subansiri branch. Groups 13 to 16 represent the immediate southern 1a. Lepcha (a.k.a. Rong) 4b-s Dimasa (a.k.a. >Bodo<)- sb-1 Damu-perhaps near the Subansiri, and eastern neighbours of the Nagas. Groups 18 to 38 are 2 Hrusish- in Kameng district in northern Cachar Hills precise whereabouts unknown branches of Tibeto-Burman languages spoken outside easternArunachal Pradesh 4b-6 Hojai (a.k.a. >Bodo<)- sb-2 Bori- along the Siyom and Sike in an area of northeastern India. More information on all of these 2a Dhfmmai (a.k.a. Miji)-between northern Cachar Hills enclosed by the Luyor hills on the east, the groups can be found in my 2001 handbook. Dirang, Seppa and Riang 4b-7 Kachari (a.k.a. >Bodo<)- Goalpara Piri hills on the west and on the north by the 2b Levai (a.k.a. Bangru)- northeast of 4b-8 - closing together ofthese two ranges, esp. Riang, northwest ofYomtam 4c Chutiya (a.k.a. Deori Chutiya) -in Payum, Dupu, Yiyo, Pame, Gasheng, Paying, 2c Hruso (a.k.a.Aka)- southwest Lakhimpur and Sibsagar districts Gatte, Gameng, Bogu and Mega villages ofSeppa, east ofShergaon s Tani (a.k. a. >Ab or-Miri-D afla <) -in Arunachal Pradesh sb-3 Mishing (a.k.a. Plains Miri)-East 3 Kho-Bwa-in Kameng district in sa Western Tani (a.k.a. >Miri-Dafla<, Nishi Siang at Oyan, Mer and N amsing easternArunachal Pradesh group) -west ofthe Subansiri river sb-4 Padam (a.k.a. Bor Ab or, i.e. >Great 3a Khowa (a.k.a. Bugun)- around Bomdila, sa-1 Apatani-between the Panior ,Ab or<)- the Dibang, Siang and Yamne esp. in Wanghoo and Singchungvillages and Kamla rivers valleys in East Siang, from the 3b Sulung (a.k.a. Puroit, Puroik)- north ofRiang sa-2 Bengni (a.k.a. Bangni, >WesternDafla<) border in the south to the Sidip river in along the upper reaches of the Par river sa-3 Nah (a.k.a. Chendar, Haphi, Hari and the north, and in western Lohit between 3c Lishpa (pass themselves off as Tisi) -Upper Subansiri district in seven the Siang and Sisiri or Sikang river, esp. >Monpa<)- aroundDirang villages ofTaksing administrative circle Anpum, Bijari, Bolung, Bomjir, Dambuk, 3d Sherdukpen- mainly in and around sa-4 Nishing (a.k.a. Nishi, >EasternDafla<)- Kangkong, Kappang, Tap at and Poblung. Shergaon, Rupa and Jigaon Kamengvalley in sb-s Minyong (a.k.a.AdiMinyong) -west bank 4 Brahmaputran (a.k.a. Kachari-Koch or Bodo-Koch) sa-s Nyisu-possibly a sub-branch ofNishi of the lower Siang river, downstream of 4a Koch sa-6 Yano-possibly a sub-branch ofNishi the Bori and Karko language communities 4a-1 Atong (a.k.a. At'ong, A tong >Garo<) sa-7 Tagin- northeastern Subansiri and to the east of the Gallongs, esp. Ledum, -Somasvara and Baghmara district, inc. Denekoli and Taliha Sido, Mirem, Yagrung, Rengin, Rotung, in the lower Garo Hills sa-8 HillMiri (a.k.a. Sarak) -enclaves Kebang, Yemsing, Pangin, Yeksing, Boleng, 4a-2 Ruga (a.k.a. Ruga >Garo<, Rugha) between Nishi and Gallong territory Dosing, Pankang, Parong, Yebuk, Sitang, -downhill from the Atong sa-9 Gallong (a.k.a. Galo) southern half of Riga, Riu, Komsing, Rumgong, Pes sing, 4a-3 Rabha- skirting the north of the Garo West Siang district, west ofthe Siyom river Molom, Moput, Lorging and J omlo Hills, on both sides of the Brahmaputra sa-10 Pailibo (a.k.a. Libo)- north ofGallong sb-6 Shimong (a.k.a. Simong) -left just before it bends south territory along banks of the Siyom bank of the Siang in and around 4a-4 Pani Koch- western fringe of the or >Yomgo<, esp. Tato andPayum Yingkiong, esp. N gaming,Jido, Anging, below the Garo Hills circles ofWest Siang district Singiang, Palin, Likor, Puging, Gete, 4a-s Rajbangsi (no longer speak anyTibeto­ sa-u Ramo- Mechukha subdivision ofWest Shimong and Gobuk villages Burman language but a dialect ofBengali, Siang district to the northwest of the Pailibo sb-7 Pasi-Pasighat area in the foothills, historically Rajbangsi are Koch assimilated area, esp. between Machukha and Tato esp. Gine, Kelek, Roing, Monku, linguistically in the Kamarupan period) sa-12 Bokar-Monigong Circle ofMachukha Balek, Rasam and Tigra - Siliguri,Jalpaiguri, Alipur duar (trade subdivision in West Siang district, esp. sb-8 Panggi -lower Yamne valley above the route) and other former Bhutanese duar Gesing, Pangri, Yangrang, Taihiyong, Rote, confluence ofthe Yamne and the Siang, esp. 4b Bodo-Garo Pidi, Ruying, Kate, Luto, Ramni, Hemi Geku, Sumsing, Sibum,Jeru and Pongging 4b-1 Mech-Jalpaiguri district, Bhutanese duar or Mote, Pote, Karle, Monigong, Ingo, sb-9 Tangam- northernmost Siang district in 4b-2 Garo - Garo Hills Pulom, Simegong, Papigro, Tadadege, the villages ofKuging, Ngering and Mayum 4b-3 Bodo-Darrang Lapugora, Yorkongdo and N amasiba

314 315 George van Driem TheN aga language groups within the Tibeto-Burman language family

sb-10 Karko (a.k.a. Karka)- across the river from (Lungchang), Longphi, Hawi (Have, Hf Kezhama (a.k.a. Kezha) ISP Zo- northeastern and the the Shimong to the west, mainly in Karko Havi),Jogli (Jugli, Yogli), Tikhak, Longri, Hg Senkadong neighbouring portion ofBurma village and nearby Ramsing and Go sang Rangpang, Moshang, Rangpang, etc. Hh Mao 1sq Vaiphei- northeasternMizoram sb-H Ashing (a.k.a. Asing)- headwaters 9a-2 N octe- in Hi Ntenyi and in Burma around Tiddim of the Siang near the Tibetan border, between Pong and Tirap Hj Maluri 1sr Tiddim Chin- northeastern Mizoram from Ramsing in the south to Tuting 9b Konyak-Wancho cluster 12 Zeme branch-the southwestern tip of and in Burma around Tiddim in the north, esp. Pango, Bomdo, 9b-1 Wancho-west ofTirap district in N agaland and the northw~stern corner of ISS Falam Chin-in southeastern Mizoram Ninging, Minging and Mosing. Arunachal Pradesh around Pong M anipur, starting from Tamenglong and and in Burma around Falam se Milang 9b -2 Konyak- the northern tip ofN agaland moving to the north and north-northeast 1St Haka Chin-in southeastern Mizoram sc-1 Milang- three villages ofMilang, Dalbing 9b-3 Phom- northern tip ofNagaland, 12a Mzieme and in Burma around Haka and Pekimodi in the upper Yamne valley in north ofTuensang 12b Khoirao 16 Mru (not to be confused withMaru) Mariyang subdivision ofEast Siang district 9b-4 Khiamngan- northern tip ofNagaland 12c Maram central hills of the Chittagong 6 Digarish (a.k.a. >Northern Mishmi<)-Lohit 9b-s Chang- northern N agaland, in 12d Puiron 17 Kachinic district ofArunachal Pradesh and around Tuensang 12e Zeme (a.k.a. Empeo Naga, Kacha 17a Luish 6a Idu (a.k.a. Chulikata >cropped hair< 9c UnclassifiedNorthernNaga Naga,KochuNaga) 17a-1 Andro [inManipur; theAndro Mishmi, >BebejiaMishmis<)-in the 9c-1 Kuwa- Burma east ofTirap district 12f Nruanghmei (a.k.a. Rongmai, Kabui) now speakMeithei] Dibang, Mithun and Dri valleys 9c-2 Haimi- Burma east ofTirap district 12g Liangmai (a.k.a. Kwoireng) 17a-2 Sengmai [inManipur; the 6b Taraon (a.k.a. Digaro, DigaruMishmi)-in 9c-3 Ponyo-Burma aroundLahe, east 13 Tangkhul branch- the northeastern corner ofManipur Sengmai now speak Meithei] the Lohitvalley, between the Delei and ofTuensang inNa gal and around Ukhrul and in neighbouring portions ofBurma 17a-3 Kadu [in Upper Burma; the Lati rivers in the east, the Kharem in the 9c-4 Welam- Burma around Lahe, east 13a Tangkhul Kadu now speak Burmese] south and the Digaru in the west ofTuensang inNagaland 13b Maring 17a-4 Chakma [in of 7 Midzhuish (a.k.a. >SouthernMishmi<)­ 9c-s Nokaw-Burma around Lahe, east 14 Meithei (a.k.a. Manipuri)- in Bangladesh, now speakBengali] ofArunachal Pradesh ofTuensang inNa gal and IS Kukish (a.k.a. Mizo-Ku~i-Chin) 17a-s Sak [southern Chittagonghill tracts 7a Kaman (a.k.a. Miju, MijhuMishmi)­ 9c-6 Htangan-Burma aroundLahe, IS a Mizo (a.k.a. Lushai)-Mizoram ofBangladesh, still speak Sak] upper reaches of the Lohit on both banks east ofTuensang inNagaland ISb Hmar- northern Mizoram and Cachar 17b Jinghpaw (Kachin) ofthe river around Parsuram Kund 10 Ao branch- centralNagaland, north and east ofWokha ISC Anal (a.k.a. LamgangKuki) -in 17b-1 Singpho-in theArunachalhills to the east 7b Zaiwa (a.k.a. Zakhring andMeyor dialects) -less 10a Yacham southeastern Manipur ofSadiya, north ofthe Tangsa area (see sa-1) than two hundred members ofthe Zakhring wb Tengsa 1sd Paite- in southeastern Manipur 17b-2 otherJinghpaw dialects in northern and Meyor clans in the vicinity ofWalong we Ao Chungli ISe Gangte-in southeastern M anipur Burma and southwestern China 8 Karbi (Mikir) wd Ao Mongsen ISf Thado (a.k.a. Thadou) -insoutheasternManipur 18 Raji-Raute- Pithauragadh district in Uttar­ Sa Karbi proper- the territory all we Lotha (a.k.a. Lhota) ISg Kom-in southern Manipur khanda and the western Nepalese Terai around Hojai in Assam wf Sangtam (a.k.a. Thukumi) ISh Purum- in southern Manipur 19 West Himalayish- in and Garhwal Sb Amri Karbi -lower Meghalaya wg Yimchungrii (a.k.a. Yachumi) ISi Chiru - in southern Manipur and an 20 Bodish-in Tibet and neighbouring portions hills northeast ofNongpoh wh Nruanghmei enclave south ofSilchar in Assam ofPakistan, Nepal, India and Bhutan 9 Konyak a.k.a. Northern N aga 11 Angami-Pochuri branch-southern N agaland ISj Biete-in an enclave northeast ofSilchar in Assam 21 Newaric- in Nepal. 9a Tangsa-Nocte cluster around , neighbouring portions of 1sk Chorei- in an enclave southwest 22 Kiranti- inN epal. 9a-1 Tangs a-in Changlang and Miao Burma and the northern fringe ofManipur ofSilcharinAssam 23 Tamangic- inN epal. subdivisions ofTirap district in Ha Angami ISl Bawm- on the Tripura-Assam border 24 Dura-inNepal. Arunachal Pradesh around Tirap Hb Chakri (a.k.a. Chokri) ISm Hrangkhol-in an enclave in northern Tripura 26 Magaric-inNepal. and also in neighbouring parts of Hc.Sema 1sn Lakher (a.k.a. Mara) -in southern Mizoram 26 Chepangic-inNepal. Burma: including Moklum, Longcang Hd Rengma (a.k.a.Anzang) ISO Simte- northeastern Mizoram and the 27 Dhimalish- in the southeastern Nepalese He Pochuri (a.k.a. >Southern neighbouring portion ofBurma Terai and on the Bhutanese duar Sangtam<, >EasternRengma<) 28 Lhokpu- in Bhutan

316 317 The Naga language groups within the Tibeto-Burmanlanguage family George van Driem

29 Gongduk- in Bhutan A perennial issue is the likely provenance of the N a gas, a 30 Tshangla- in Bhutan question that we can ask about any people or language 31 Nungish- in Yunnan province in community in the world. Inevitably, an ethnolinguis­ China and in northern Burma tically informed answer to such a question is necessar­ 32 Karenic- in eastern Burma and western Thailand ily many-facetted because the prehistory of any language 33 Pyu [extinct]- central Burma community is likely to have been complex. More funda­ 34 Sinitic-in China mentally, the linguistic ancestors of a community were 35 Bai-in Yunnan province in China not necessarily the same people in time and space as the 36 Tujia- in China biological ancestors of that community. The material cul­ 37 Q!angic- in Sichuan province in China ture and agricultural package of a community represents 38 Ersu cluster-in Sichuan and Yunnan provinces in China a third quantity altogether. Nonetheless, a detailed under­ standing of the linguistic affinities and population genet­ ics oflanguage communities and the archaeological record of the region which they inhabit sheds much light on the possible prehistory and provenance of peoples and on the interactions which these peoples may have had in the past with other groups.

The N eolithic cultural assemblage which is associated with northeastern India is the Eastern Indian Neolithic. This cultural assemblage is entirely distinct from other N eo­ lithic traditions of the Indian Subcontinent. The anteced­ ents of the Eastern Indian N eolithic lie to the northeast, i.e. possibly in what today is Sichuan province. This is shown in !:. Fig. 4· Brahmaputra basin and surrounding hill tracts colonised inextricably connected with the question of the homeland Fig. 4· In several publications (vanDriem 1998,2001, 2002), by ancient Tibeto-Burmans from Sichuan (van Driem 2001: 415). and provenance of the Austroasiatics, whose current geo­ The dating of the Eastern Indian Neolithic is currently a matter graphical centre of gravity lies in this very region. In addi­ I have argued for the hypothesis that the intrusion of this of archaeological conjecture and still not yet based on any sound cultural assemblage into the northeast of the Subcontinent stratigraphy or reliable radiocarbon or accelerator mass spectroscopy tion to the geographical distribution of modern Tibeto­ could represent Proto-Tibeto-Burmans originating from dates. Burman and Austroasiatic language communities, we can Sichuan, spreading southwest onto the lower Brahmapu­ examine the findings of population genetics and attempt to tran plain and introducing themselves and their Eastern trace the spread of agriculture and crops. Indian N eolithic culture to resident Austroasiatic popula­ before they appeared in . Therefore, most tions. Indeed, the two major linguistic phyla Tibeto-Bur­ archaeologists now tend to ascribe younger dates to the In­ Linguistic palaeontology, a term introduced by Adolphe man and Austroasiatic meet in this area. More recently, I dian Eastern Neolithic. Yet such archaeological dating esti­ Pictet in 1859, is an attempt to understand ancient mate­ have compared and discussed alternative and competing mates lie in the realm of conjecture, since a solid stratigra­ rial culture on the basis of the lexical items which can be interpretations for the spread ofTibeto-Burman as reflect­ phy and calibrated radiocarbon dates are still unavailable reliably reconstructed for the oldest level of the common ed in the archaeological record (van Driem 2007a). for this major South Asian cultural assemblage. ancestral language. Linguistic palaeontology and native lore both suggest that the ancient Tibeto-Burmans were The main uncertainty about this scenario is the dating of What arguments can be advanced to decide whether the cultivators ofbroomcorn millet panicum miliaceum and the relevant cultural assemblage. Archaeologists have esti­ advent of ancient Tibeto-Burman language communities foxtail millet setaria italica, whereas linguistic palaeontol­ mated the Indian Eastern N eo lithic to date from between to the Indo-Burmese borderlands took place in the Neo­ ogy qualifies the ancient Austroasiatics as the most likely w,ooo and 5,ooo BC (Sharma 1989b; Thapar 1985). If these lithic or much later, say, in the Bronze Age? The provenance candidates for the first cultivators of rice. estimates are taken at face value, it would mean that north­ ofTibeto-Burman groups in northeastern India and the In­ eastern India had shouldered adzes at least three millennia do-Burmese borderlands and the timing of their arrival is

319 318 George van Driem The Naga language groups within the Tibeto-Burmanlanguage family

The oldest millet cultures have been found in the ancient More recently, scholars have increasingly begun to take The seedlings are not transplanted but ripen in just four genetic findings and new archaeological research in the Yellow River basin and on the North China Plain. Cur­ note of findings that would move the original homeland of months in fields which must be constantly weeded. In bao area compel us to embrace another interpretation of pre­ rently, the earliest attested domestic millet dates from be­ rice cultivation back to the Indian subcontinent. At the La­ kheti, the rice seedlings are sown from mid March to mid history? New population genetic findings, especially on Ti­ fore 6ooo BC at Jl[.l5~dft Xinglonggou near~~ Chifeng, huradewa site (26°46' N, 82°57' E), the early farming phase, April in ploughed wet fields and likewise do not need to be beto-Burman populations in the Himalayas, e.g. Parkin et where there was aN eolithic culture without sickles (Fuller corresponding to period 1A in the site's clear-cut stratig­ transplanted. In sali kheti, the rice is sown from mid May to al. (2oo6a, 2oo6b), Kraaijenbrink et al. (2oo6, 2007), have et al. 2007). The rice story, however, is more complex, and raphy, has radiocarbon dates ranging from c. 5300 to 4300 mid June, and the seedlings are transplanted. Sali kheti rice shed light on the biological ancestry of populations in the the plot of the story has changed more than once in recent BC. Carbonised material from period lA was collected by varieties are held to have been derived from the wild offi­ northeastern portion of the Subcontinent. I have had oc­ decades. Whereas the oldest rice cultivation was once held the flotation method, yielding Setaria glauca and Oryza ru­ cianalis rice still widely found in swampy village areas. The casion to discuss the possible interpretations of emergent >incontestably< to lie in the Indian subcontinent (Haudri­ fipogon as well as a morphologically distinct, fully domesti­ wild rufipogon rice cannot be used for human consump­ population genetic findings with respect to the ethnolin­ court & Hedin 1987), subsequent scholarship moved the cated form of rice »comparable to cultivated Oryza sativa« tion because the plants shed their seeds before they ripen, guistic prehistory of Tibeto-Burman and Austroasiatic homeland of rice agriculture from the Ganges to the Yang­ (Tewari et al. 2002). More recently, accelerator mass spec­ so that rufipogon rice is used in Assam and other parts of populations at some length (van Driem 2oo6, 2007b, 2008), tze. For years conventional wisdom in archaeological troscopy dates were obtained on the rice grains themselves, northeastern India as cattle feed (Hazarika 2oo6b). and new results of genetic assays will soon appear in print. circles dictated that rice was domesticated in the Middle corroborating the antiquity of rice agriculture at the site. Ultimately, the answers to our questions about the prov­ Yangtze, perhaps as early as the sixth millennium BC. Most recently, new radiocarbon dates for rice agriculture So, might ancient Tibeto-Burmans originating in what enance of theNa gas and other peoples of the northeast will have been coming from the Ganges basin, with the Tokuva is today China indeed have spread southwest onto the have to come from a careful reconstruction of the disparate Currently available evidence indicates that immature site near Allahabad now yielding similar dates (Vasant Brahmaputran plain and introduced themselves and the societal and migrational processes which have shaped the morphologically wild rice may have been used by for­ Shinde [Vasant Sivaram Sinde], personal communication Eastern Indian Neolithic cultural assemblage to resident ethnolinguistic prehistory of all of northeastern India and agers before actual domestication of the crop, e.g. at 27 November 2007). Austroasiatic populations, or will emergent population adjacent regions. the A +tl Bashidimg site (7ooo-6ooo BC) belonging to the ~E] 0J Pengt6ushan culture in the Middle Yangtze and Turning to northeastern India and the Indo-Burmese bor­ at sites in the Yangtze delta area such as ~{lj!jtiJ Kuahuqiao, derlands, where we find the Naga peoples and related Tibe­ Ji~*f~ Majiabang {RJfr£.1:{£ (5000-3000 BC) and Hemudu to-Burman language communities, we must recognise that, (5oo0-4500 BC). However, only c. 5000 BC was the actual notwithstanding the excellent archaeological work done cultivation of rice probably first undertaken by people in in the Ganges and Yangtze river basins, the archaeology the Lower Yangtze, who at the time relied far more heavily of ancient rice agriculture is simply not known because no on the collecting of acorns and water chestnuts. substantive archaeological work has been done on theNeo­ lithic in the most relevant areas, e.g. Bangladesh, northeast­ Today, our understanding of the palaeoethnobotanical ern India and Burma. The sheer dearth of archaeological picture is more complex. The two main domesticated va­ research in these areas leaves entirely open the possibil­ rieties of rice, oryza indica and oryza japonica, are phyla­ ity that rice cultivation may have originated in this region. genetically distinct and would appear to have been do­ Perhaps the remains of the very first rice cultivating cul­ mesticated separately. Oryza indica derives from the wild tural assemblages lie buried forever in the silty sediments progenitor oryza nivara and was first cultivated in South of the lower Brahmaputran basin or were washed out long Asia or western Southeast Asia, perhaps in two separate ago into the depths of the Bay of . domestication events. Oryza japonica derives from the wild progenitor Oryza rufipogon and was first cultivated to At least four species of wild rice are native to northeastern yield the early Oryza japonica along the Middle Yangtze. India, viz. Oryza rufipogon, Oryza officianalis, Oryza peren­ Harvey et al. (2oo6) have critically reassessed the morpho­ nis, Oryza meyeriana, and over a thousand varieties of do­ metries of rice finds associated with various N eo lithic sites mesticated rice are currently in use in the region (Hazarika throughout the Yangtze basin in the light of recent genetic 2oo6a). The different varieties of rice in northeastern India findings. The wild progenitor Oryza rufipogon was not fully are cultivated in three periods by distinct cultivation proc­ domesticated in the Lower Yangtze to yield the early Oryza esses. In the process of ahu kheti, the rice is sown in the japonica until c. 4000 BC. months ofjagun and sot, i.e. mid February to early April.

320 321