The defoliation of the Tani Stammbaum: A positive-minded exercise in contact linguistics Mark W. Post University of New England 1. About Tani • • • • • • speakers: around 800,000 (400,000 non-Mising) geographic location: Eastern Himalaya, central Arunachal Pradesh, few in SE Tibet, one populous tribe (Mising) in NE Assam. cultural profile: hillside grain agriculture mixed with foraging, hunting, limited husbandry, sub-tropical forest materials (bamboo, cane, wood, leaf), patrilineal, animist genetic (linguistic) profile: clearly T-B/T-H in some sense, seemingly partially creolized in early history of the subgroup, relatively recent spread and diversification, possible substrates (Koro, Milang, Digarish, other?) language contact: extensive early and ongoing Tani-Tani, some medium-recent Tani-Bodish, extensive recent Tani-IE (Assamese, Hindi, Indian English), marginal and sporadic recent Tani-other TB (poss. more in early times) language work: 19th/early 20th C Europeans, Indian and Chinese “language guides”. Turning point: Sun (1993), followed by Post and colleagues (I. Rwbaa, I. Rwbaa, B. Rwbaa, Modi, Kanno, Blench 2006-2013, Sun and Post 2014~2015) Figure 1 – Northern North East India, with Tani Language Area in rough outline 1 2. Tian-Shin Jackson Sun’s (1993) A Historical-Comparative Study of the Tani (Mirish) Branch of Tibeto-Burman 2.1. Background to Sun’s work Brown (1837) notes that “Abor” (= Padam) and “Miri” (= Mising) are closer than either is to “Dafla” (= Nyishi); Hamilton (1900) aligns Dafla with “Apa Tanang” (= Apatani). Marrison (1988) and Nishida (1984 - cited by Sun 1993, not seen by me) basically agree with the resulting incipient “Western Tani” vs. “Eastern Tani” split. But all of this was simply eyeballing the data - no systematic establishment either of the Tani (or Abor-Miri-Dafla) branch within Tibeto-Burman, nor of the internal structure of the branch, was attempted until Sun (1993). 2.2. Sun’s methods Surveyed five “key languages” – Apatani (Simon 1972), Na/Bangni (J. Sun fieldwork), Bokar (Ou-Yang 1985), Padam and Mising (Lorrain 1995 [1910]), supplemented by a dozen or so less comprehensive sources. Primary goals: (a) establish regular correspondences among shared vocabulary items (b) perform as much reconstruction as possible (c) develop subgrouping proposal on basis of (i) shared phonological innovations (ii) differential sharing of characteristic lexemes. Methodological caveats to the subgrouping proposal: Under ideal conditions, boundaries between distinct subgroups will be demarcated by bundles of isoglosses each of which is defined by a shared innovative linguistic feature. In practice, however, criss-crossing of isoglosses are the rule rather than the exception, and clear-cut dialect boundaries are rarely found, especially in compact groups like Tani. (Sun 1993: 224) We believe that, at least at the present stage of our comparative research, it may be more realistic to adopt a prototype approach to tackling Tani dialect affiliations. That is, selected linguistic isoglosses are used to define broad subgroups within Tani, each one with prototypical or central members where the characteristic features of the group are fully represented, as well as less typical or peripheral members where the defining features are only partially present. Put differently, we make provision for dialect subgroups with fuzzy edges and even borderline cases between major subgroups. (Sun 1993: 225) 2 2.3. Sun’s results Sun isolated a set of (a) four phonological innovations and (b) twenty-five lexical isoglosses, which were employed in the establishment of prototype-based subgroups. Phonological innovation (1) Velar palatalization (2) Labial palatalization (3) Deliquidation Schematic (4) Coda *s-drop *-s > Ø (WT) *VEL > PAL / _i, j, e (WT) *LAB > PAL / _i, j, e (WT) *rj- > j- (ET) Sample gloss ‘ill’ ‘know’ ‘eye’ ‘man’ ‘bow’ ‘pig’ ‘vomit’ PT Galo (WT) *ki *ken *mik *mi *rji *rjek *b(r)as čìčènɲ�ḱ ɲír�-́ rə́kba_- Minyong (ET) kikenmikmijijekbat- Table 1 – Phonological innovations used in Sun’s (1993) Tani subgrouping proposal (innovations in bold) Resulting Stammbaum: 2.4. Discussion Sun was able to establish a significant number of largely regular sound correspondences; the resulting reconstructions have held up well when tested against data from languages which were little-known or unknown to Sun at the time of his writing (Galo, Minyong and Tangam, though not the highly exceptional Milang). So, it’s clear that we’re dealing with a solid subgroup, with a plausibly reconstructible ancestor. What a relief! However, the matter of internal subgrouping has remained vexing (if only to me). Two issues stand out: (1) overlaps among 50% of subgrouping criteria (cross-branch sharing of innovations) and (2) strong correlation between subgroups and geographical areas – in particular, Western Tani with the Subansiri River area and Eastern Tani with the Siang River area… 3 3. Background to the present study Here, we deal with four problematic languages from the perspective of subgrouping: Galo, Bokar, Damu and Tangam. The case of Milang is more complex, and has a slightly different type of solution (Post and Modi 2011). 3.1. The problematic cases of Galo and Bokar Galo was one of four Tani languages considered by Sun to be problematic, in the sense that it failed to align perfectly with his subgrouping criteria. In particular, Weidert’s (1987) Galo data aligned with Eastern Tani with respect to (3) Deliquidation, as well as in sharing a larger-than-average number of ET lexical isoglosses. In Post (2013a), I showed that ET overlap was highest in the Minyong-bordering Pugo dialect of Galo, and lowest in the Tagin and Hills Miri-bordering Western and Northwestern dialects of Galo, and explained Galo’s ET convergence as due to language contact. Although I have no direct experience, by implication at least, the same should be true of the Siang-bordering, ET-abutting Bokar language. Ok, so change (3) is an areally-diffusing innovation associated with the Siang River valley and its tributaries. That still leaves us with three robust-looking phonological isoglosses… 3.2. The problematic cases of Damu and Tangam Damu is a somewhat mysterious language, apparently spoken in Tibet, with which I have no direct experience. Seemingly, there were only 80 speakers heavily influenced by their proximity to larger numbers of Tibetans in 1985; it is not known whether this community still exists, or not. Sun (1993) speculates that Damu may represent a northern variety of Tangam – a language for which no data existed at the time of Sun’s writing – but my (Post 2013b) Tangam data depart significantly from Ouyang’s (1985) Damu data in a number of respects. In any case, both Damu and Tangam problematically align with Western Tani with respect to phonological innovation (4) Coda *s-drop. Phonological innovation (1) Velar pal’zn (2) Labial pal’zn Schematic *VEL > PAL / _i, j, e (WT) *LAB > PAL / _i, j, e (WT) (3) Deliquidation *rj- > j- (ET) (4) Coda *s-drop *-s > Ø (WT) Sample gloss ‘ill’ ‘know’ ‘eye’ PT ‘man’ ‘bow’ ‘pig’ ‘vomit’ 4 *ki *ken *mik Galo (WT) čìčènɲ�ḱ - Minyong (ET) kikenmik- *mi *rji *rjek *b(r)as ɲír�-́ rə́kba mijijekbat Tangam kikinmik- ~ mitmijɨjɨkba Language contact again? This time, probably not. Tangam are exclusively on the eastern bank of the Siang River, and exhibit an overwhelming abundance of (a) Eastern Tani phonology and vocabulary and (b) unique innovations. There is little if any cultural, linguistic or oral-historical evidence that would support a putative contact relationship between Tangam and Western Tani languages, even in early times. So, we would need to re-draw the Stammbaum along these lines, with change (3) Deliquidation understood as an areally-diffusing innovation within the Siang valley: Proto-Tani Pre-Western Tani (4) (3) Milang Eastern Tani Western Tani (1) Apatani Subansiri (2) Tangam, Bori Damu? Bokar Bangni-Tagin NyishiHill Miri Siang Minyong Pasi Padam Galo Mising Lare Pugo (3) Figure 2 – Revised Tani Stammbaum, with Damu and Tangam re-classified as Western, and change (3) understood as an areal innovation which is non-criterial for subgrouping Ah! That’s better! Or is it…? 4. The evolution of Tani languages reconsidered 4.1. What happens when we map other known Tani sound changes to Sun’s Stammbaum? There’s a reason why Sun only considered four out of the available 40-odd identifiable Tani innovations as subgrouping criteria. Nothing else works! That is to say, we can certainly identify regular correspondences among languages, so we can do some reliable reconstruction and see where individual innovations happened in the histories of individual languages. In the same way, we can see how some languages retained archaisms. However, very few innovations or retentions (Sun’s (1), (2) and possibly (4) excepted) align with any others. This is suggestive of areal effects on a massive scale. 5 4.1.1. Case study 1: *-l > -r Only one core Tani language retains Proto-Tani *-l coda: Padam. 1 Every other Tani language merges *-l to -r. Gloss ‘disc; money’ ‘laugh’ ‘star’ PT *bal *ŋil *kar Galo (WT) barɲɨrkar- Minyong (ET) barŋirkar- Padam (ET) balŋilkar- Table 2 – Reflexes of PT *-l in WT, ET and Padam So, do we place the *-l > -r change in the earliest stage, in which the common ancestor of all Tani languages other than Padam split with Padam? Proto-Tani *-l Padam *-l All other Tani *-r Milang *-l If so, then we imagine that every other innovation which affected both Eastern Tani languages and Padam was an areal change, and/or that Padam remained mutuallyintelligible with other ET lects, and underwent the corresponding changes simultaneously. That’s fine, and some version of this story is probably correct. But it lands us with a pretty misleading-looking Stammbaum… 4.1.2. Case study 2: Fricative initials Sun (1993) reconstructs a much larger set of fricative initials to Proto-Tani than is found in any daughter language. Most of these were merged and/or lost, as follows: PT *f*v*ɦ*h*z*s- Bng fvh/Øhss- Apt hhhhhs- Dam x(Ø-) hxjs- Bkr ʱØ-/hhhjs/h- Gal ʱʱʱʱs/hs/h- Tng ØØddjs- Mny ØØØØjh- Pdm ØØØØjs- Mis ØØØØss- Unless I’m wrong, it isn’t possible to map any subset of these changes onto any one of the subgrouping proposals discussed above. 1 Milang also retains *-l; however, if Post and Modi’s (2011) reclassification of Milang as descending from a pre-PT position is tenable, then Milang’s retention of *-l has no bearing on Tani subgrouping. 6 Proto-Tani *f, *v LABIALS WT *f, *v Bangni f, v ET Ø Tangam, Minyong, Padam, Mising Damu x, Ø Apatani h Bokar ɦ, h Galo ʱ- Proto-Tani *ɦ, *h (?) GLOTTALS Tangam d WT *ɦ, *h Minyong, Padam, Mising Ø Galo ɦ Damu h, x Apatani, Bangni, Bokar h Proto-Tani *z CORONALS 1 WT *z ET j Damu, Bokar, Tangam, Minyong, Padam Apatani h Bangni, Mising s Galo s/h 7 Proto-Tani *s- CORONALS 2 Bangni, Apatani, Damu, Tangam, Padam, Mising s- Galo, Bokar s/h- Minyong h- Clearly, there’s something wrong with this approach. 4.2. Language contact and areal shifts In the majority of cases in Tani, the most reliable correlate of a phonological innovation is geographical proximity to another language which shares this innovation. If we draw a “tree” not on “genetic” lines, but instead on largely geographical lines (i.e., not a “real” tree, but rather an areal-distribution schematic): Proto-Tani Subansiri River Central Tributaries Bokar Tangam Damu Apatani (Tagin, Nyishi, etc.) Bangni Siang River Milang Galo Minyong Lare Pugo Padam Mising Then innovations readily cluster: Proto-Tani Sun’s 4 isoglosses Subansiri River Central Tributaries Siang River (4) *s-drop (1) Vel Pal Bokar Tangam Damu Apatani Bangni Galo (Tagin, Nyishi, etc.) Lare Pugo (2) Lab Pal (3) Deliq Minyong Padam Mising 8 Milang Labial and Glottal Fricatives Proto-Tani *f-, *v-, *ɦ-, *h- Subansiri River Central Tributaries Damu Apatani Bangni Galo (Tagin, Nyishi, etc.) Subansiri River Bokar Tangam Labials > Ø Milang Minyong Lare Pugo Labials, Glottals > h- or ɦ- Coronal Fricatives Siang River Labials, Glottals > Ø Mising Padam Proto-Tani *s-, *z- Central Tributaries Siang River Bokar Tangam Milang *z> jGalo *s- > h- Minyong Lare Pugo Padam Mising Damu Apatani Bangni (Tagin, Nyishi, etc.) *z- > s- 5. Conclusion: What have we learned? For nearly ten years, I’ve been trying to reconcile the fact that Sun’s (1993) Proto-Tani reconstructions hold up wonderfully well, enabling me to say a good deal about the evolution of individual Tani languages, with the fact that I seemed unable to further subgroup Tani languages on the basis of the many phonological innovations that have been discovered. Furthermore, I had to return time and again to language contact, which almost always explained “cross-branch” sharing of innovations. Tallying everything up, the “exceptional” “cross-branch” innovations outnumbered the Stammbaum-agreeing innovations by a factor of five or six. Might it not be simpler, and potentially more insightful, to assume that areal diffusion is the primary determinant of innovative feature-sharing in Tani, rather than sharing a common meso-ancestor? 9 This view would imagine that speakers of mutually-unintelligible Tani languages nevertheless find their languages changing in the same ways. At a distance, this might seem implausible. But in fact, this seems to be exactly the way things happen. Tani grammars are almost identical to one another. This suggests low time-depth, but it also suggests a continuing communicative compatibility. Tani language speakers who live next to other Tani language speakers almost invariably learn one anothers’ languages, intermarry, and culturally converge. Milang and Padam are certainly different enough as languages to be mutually unintelligible. But almost all Milang speak Padam fluently and use it on a daily basis; no doubt, speaking Padam influences the way Milang speak Milang. Similarly, Tangam and Minyong should be mutually unintelligible. But almost all Tangam are bilingual in Minyong, and sometimes forget which language they are speaking mid-sentence. In the same way, Galo and Minyong should be mutually-unintelligible, but almost all Galo in Minyong-bordering areas speak Minyong fluently, and this has clearly influenced the Galo dialects from these regions. I am nonetheless left with a few questions: what is a subgroup? Is it always the same type of thing? Can we model genetically-inherited and contact-acquired features in the same way (or in mutually-compatible ways), as part of a diachronically-oriented “classification” of languages? References Brown, N. (1837). 'Comparison of Indo-Chinese languages.' Journal of the Asiatic Society of Bengal 6: 1023-1039. Hamilton, R. C. (1900). An Outline Grammar of the Dafla Language: As Spoken by the Tribes Immediately South of the Apa Tanang Country. Shillong, Assam Secretariat Printing Office. Lorrain, J. Herbert (1995 [1910]). A Dictionary of the Abor-Miri Language, with Illustrative Sentences and Notes. New Delhi, Mittal. Marrison, Geoffery E. (1988). 'The Adi-Dafla group of languages of North-East India: a sketch.' In David Bradley, Eugenie J. A. Henderson and Martine Mazaudon, Eds. Prosodic Analysis and Asian Linguistics: To Honor R. K. Sprigg (Pacific Linguistics C-104). Canberra, Australian National University Press: 205-222. Ou-Yang, Jue-Ya (1985). 珞巴族语言简志（崩尼-博嘎尔语)[Luobazu Yuyan Jianzhi (Bengni Boga'eryu)/Brief Description of the Luoba Nationality Language (The Bengni-Bokar Language)]. Beijing, 民族出版社 [Minzu Chubanshe/Nationalities Press]. Post, Mark W. (2013a). 'The Siyom River Valley: An essay on intra-subgroup convergence in TibetoBurman.' In Gwendolyn Hyslop, Stephen Morey and Mark W. Post, Eds. North East Indian Linguistics Volume 5. New Delhi, Cambridge University Press India: 60-90. Post, Mark W. (2013b). 'The Tangam language of Kugɨŋ Təəraŋ.' Paper presented at the 46th International Conference of Sino-Tibetan Languages and Linguistics. Hanover, Dartmouth University, Jul 10. Post, Mark W. and Yankee Modi (2011). 'Language contact and the genetic position of Milang (Eastern Himalaya).' Anthropological Linguistics 53 (3): 215-258. Simon, Ivan Martin (1972). An Introduction to Apatani. Gangtok, Sikkim, Government of India Press. Sun, Tian-Shin Jackson (1993). A Historical-Comparative Study of the Tani (Mirish) Branch of TibetoBurman PhD Dissertation. Department of Linguistics. Berkeley, University of California. Weidert, Alfons (1987). Tibeto-Burman Tonology. Amsterdam, John Benjamins. 10