Approaching the historical phonology of three highly ...the Naish group and its position within...
Transcript of Approaching the historical phonology of three highly ...the Naish group and its position within...
Diachronica 28:4 (2011), 1–25. doi 10.1075/dia.28.4.02jac.additionalissn 017–4225 / e-issn 15–714 © John Benjamins Publishing Company
Appendix
Approaching the historical phonology of three highly eroded Sino-Tibetan languagesNaxi, Na and Laze
Guillaume Jacques and Alexis MichaudCentre de Recherches Linguistiques sur l’Asie Orientale / Langues et civilisations à tradition orale-Centre national de la recherche scientifique
Appendix 1. Background data about the Naish languages
This Appendix provides (i) information on the geographic coordinates of Naxi, Na and Laze, and a brief review of the literature; (ii) phylogenetic reflections on the Naish group and its position within Sino-Tibetan; and (iii) reasons why no comparison with languages closely related to the Naish languages was attempted in the present research.
Geographic coordinates and a brief review of the literature
Naxi is the best-documented of the three languages studied in the present article. This is due in part to the scholarly attention devoted to the Naxi scripts (picto-graphic and syllabic), which indirectly stimulated linguistic work (Fang Guoyu & He Zhiwu 1995, Li Lincan, Zhang Kun et al. 1953, Rock 1963–1972). Annotated editions of Naxi ritual texts also constitute important resources for linguists (see in particular Fu Maoji 1981–1984 and the 100-volume Annotated collection of Naxi Dongba manuscripts, 1999–2000). Specialised linguistic work includes reflections on the position of Naxi respective to the Yi (a.k.a. Ngwi, Lolo) subgroup of Tibe-to-Burman (Okrand 1974, Bradley 1975); preliminary field notes by Hashimoto Mantaro (Hashimoto 1988); and a book-length glossary (Pinson 1998) which pro-vides data on several dialects (see Pinson 1996). Finally, the rudimentary word lists collected at the turn of the 20th century provide a few useful hints: on this topic, see Michaud & Jacques 2010.
The specific language varieties studied here are indicated on the map (Fig-ure 1):
2 Guillaume Jacques and Alexis Michaud
i. Naxi (autonym: /nɑ˩hi˧/), as spoken in the hamlet of A-sher (/ɑ˧ʂɚ˩/); Chinese coordinates: Wenhua township, Lijiang Municipality, Yunnan, China.
ii. Yongning Na (autonym: /nɑ˩˧/), as spoken in Yongning township, Lijiang mu-nicipality, Yunnan, China.1 A neighbouring dialect is described by Lidz (2006, 2007, forthcoming).
iii. Laze (autonym: /lɑ˧ze˧/; referred to in China as Muli Shuitian 木里水田 or Lare 拉热), as spoken in Xiangjiao township, Muli prefecture, Sichuan, China. (See Huang Bufan 2009 for a general overview of a neighbou-ring dialect.)
The present research essentially relies on first-hand data collected by Alexis Mi-chaud from 2002 to 2009. With apologies for self-references, here is a list of pub-lished results: analyses of the phonemic system of Naxi (Michailovsky & Michaud 2006, Michaud 2006a) and of its tone system (Michaud 2006b, Michaud & He Xueguang 2007); a phonemic and tonal analysis of Yongning Na (Michaud 2008); and a tonal analysis of Laze (Michaud 2009).
The Naish group and its position within Sino-Tibetan
Although language classification is not the main focus of the present paper, it is essential to provide evidence of the close phylogenetic relatedness of Na, Laze and Naxi in order to legitimate the attempt made in the present article: to contribute to the reconstruction of their common ancestor, ‘Proto-Naish’, and to document the evolution from this common ancestor to Naxi, Na and Laze, which are referred to as ‘Naish languages’.
It is widely accepted in Chinese scholarship that Naxi and Na are closely re-lated. He Jiren & Jiang Zhuyi (1985: 107) consider them as dialects of the same language, which they call “Naxi”, even though speakers of Na do not call their own language ‘Naxi’. The boundaries of ‘Naxi’ as defined by He & Jiang are so broad that they actually coincide with what we call Naish languages. ‘Naxi’ in the sense used in the present article (i.e. restricting its extent to the area where speakers use the name ‘Naxi’ for their own language) coincides with what He & Jiang refer to as ‘Western Naxi’ (纳西语西部方言), whereas they consider Na as part of a looser set of dialects to which they refer as ‘Eastern Naxi’ (纳西语东部方言). Laze is not mentioned in He & Jiang (ibid.); the question of its inclusion within Naish (‘Naxi’ as defined by He & Jiang) has been the object of some controversy in Chinese scholarship. With fewer than 300 proficient speakers, Laze is less well documented than the other two varieties. In their History of the Naxi People, Guo Dalie and He
1. This language is also known as ‘Mosuo’; for a discussion of this exonym, see Yang Fuquan 2006.
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 3
4 Guillaume Jacques and Alexis Michaud
Zhiwu, adopting the same broad understanding of the term ‘Naxi’ as He & Jiang, classify the Laze as one out of eight subgroups within the Naxi ethnic group on the basis of cultural and linguistic similarities with another proposed Naxi subgroup, the Nari 纳日 (Guo Dalie & He Zhiwu 1994 [2nd ed. 1999]: 6–7). Huang Bufan (2009: 55) expresses reservations on this topic, concluding that “…the relationship [of Laze] with Naxi, and its position within Tibeto-Burman, call for more in-depth investigation”. Our own research results point to a degree of closeness between Naxi, Na and Laze which is clearly greater than with other languages of the area. In addition to a fair amount of basic vocabulary, they share some lexical innova-tions. A short list of such probable innovations is provided in Table 18, including two disyllables: “medicine” and “noble”. Not all the words in the list belong to the basic vocabulary, witness the word for the Bai ethnic group. On the other hand, their correspondences for initials and rhymes all coincide with one of the regular phonetic correspondences brought out in this article, suggesting that they may all be actual cognates.
Table 18. A short list of probable Naish lexical innovations.[2]
meaning Naxi Na Laze Proto-Naish
to stumble pe˧ kʰɯ.piM *(S)pa
cloud2 ki˩ tɕi˧ tɕi˩sɯ˥ *ki
village hi˧mbe˧ fv.biL ɖɯ˧bie˧ *mba
Bai (ethnic group) le˧bv˧ ɬi.bvM *Sla
noble sɯ.pʰiM sɯ˩pʰie˩ *si pʰa
medicine (2nd syllable) ʈʂʰɚ˧ɯ˧ ʈʂʰæ.ɯ#H tsʰɯ˧fi˧ *rtsʰi Swri
Moreover, Laze, Na and Naxi share structural properties of numeral-plus-classifier determiners which are not found in other languages of the area (Michaud forth-coming)
The boundaries of the Naish branch remain to be worked out in detail; the list of «subfamilies» (支系) of the “Naxi nationality” (纳西族) provided by Guo & He (1999: 5–9) can serve as a starting-point, keeping in mind that this list was es-sentially based on anthropological criteria, and that the inclusion of a language in the Naish branch requires a systematic comparative study such as the present one.
2. Lookalikes to this etymon are found in Lizu: /tɕe35/, Shangyou Shixing: /tɕi55ro21/, and Xiay-ou Shixing: /ti55ro21/, as pointed out by Katia Chirkova (p.c.). The Shixing form, however, is more profitably compared instead to Proto-Lolo-Burmese *C-dim¹ and Rgyalrong /zdɯm/. As for Lizu /tɕe35/, more research is needed to determine whether or not this could be an external cognate.
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 5
As for the position of the Naish languages within the Sino-Tibetan family, it remains controversial. Naxi was initially classified within the Loloish branch of Tibeto-Burman (Shafer 1955); however, Bradley (1975: 6) shows that it does not share the innovations that characterise this group and concludes that Naxi is “cer-tainly not a Loloish language, and probably not a Burmish language either”. Thur-good (2003: 19) lists Naxi among the unsubgrouped languages of the Sino-Tibetan family. This issue links up with more general uncertainties about subgroupings within a relatively large portion of the family, which encompasses Lolo-Burmese and Qiangic. The Naish languages appear closely related to the Shixing language, spoken in Muli county, Sichuan, and which was initially classified by Sun Hon-gkai 2001 within a ‘Southern Qiangic’ branch on purely typological grounds. A relatively close relationship with other languages likewise classified as ‘Southern Qiangic’, such as Namuyi (a.k.a. Namuzi, Namʑi) and Ersu, Tosu and Lizu, is also plausible; specific investigations are required to ascertain the degree of closeness between these languages. Bradley (2008) proposes the following set of hypotheses: Naxi and Na are closest to Namuyi, the second closest is Shixing, and the third closest is Ersu. In the family tree proposed in Figure 2, the name “Naic” is pro-posed for a node grouping Naish with Shixing and Namuyi.
Some of the groupings in Figure 2 are by now well-established, in particular the Rgyalrongic group (Sun 2000a). Higher-level groupings are more controversial. Under the present proposal, the Qiangic group only includes Rgyalrongic, Tangut, Pumi (a.k.a. Prinmi), Muya and Qiang, i.e. languages that can be shown to have an extensive amount of uniquely shared vocabulary (there remain doubts concerning Zhaba). Ersu, Tosu and Lizu are generally considered to be Qiangic languages, fol-lowing Sun Hongkai’s 1983 classification (see e.g., Yu 2009), but evidence for their inclusion in this subbranch is weak; our hypothesis is that these languages may in fact belong to the Burmo-Qiangic group but not to Qiangic proper; more research is needed before any conclusion can be reached on this issue.3
The family tree outlined in Figure 2 reflects the hypothesis that Naish is close-ly related to Lolo-Burmese and Qiangic, and that it belongs in an independent branch of a larger Burmo-Qiangic group. This Burmo-Qiangic group is close to ‘Eastern Tibeto-Burman’ as proposed by Bradley 1997. This hypothesis will be briefly defended here on the basis of lexical evidence, since Lolo-Burmese and Naic languages have not preserved much morphology.
3. Fieldwork on these languages is underway, so that the necessary basis for comparative stud-ies should become available in the near future: see in particular Chirkova 2008, 2009. Further research will also be necessary to clarify the relationship of Guiqiong and Tujia to the Burmo-Qiangic group as defined here.
6 Guillaume Jacques and Alexis Michaud
One such piece of evidence is the suppletion found for the noun “year”, with a labial-initial root (Proto-Tangut *C-pja) in “this year, next year, last year” and a different root (Proto-Tangut *kjuk) with numerals: see Table 19. Rgyalrong has generalized the labial form (“next year” is innovative) and the velar root was lost. In Lolo-Burmese languages, only the root related to Tangut *kjuk is found.
Figure 2. A tentative family tree showing the position of Naxi, Na and Laze within a Burmo-Qiangic branch of Sino-Tibetan.
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 7
Table 19. Suppletion for the noun “year” in several Burmo-Qiangic languages. About the Proto-Naish forms, see Appendix 2, items a7.20 and u3.14
meaning Tangut Rgyalrong Shuiluo Pumi Muya Proto-Naish
last year .jɨ².wji¹ japa ʑɛpə jø³³zɑ²⁴ …*C-ba
this year pjɨ¹.wji¹ ɣɯjpa pəpə pə³³βə⁵³ …*C-ba
next year sjij¹.wji¹ fsɤqʰe <*psaŋ-qʰo-j
ʑɛkʰiu sæ³³βə⁵³ …*C-ba
one year .a-kjiw¹ tɯ-xpa tɜ-ko tɐ⁵⁵-kui⁵³ …*kʰu
two years njɨɨ¹-kjiw¹ ʁnɯ-xpa ɲi-ko …*kʰu
Table 20 presents a preliminary list of common etyma between Qiangic, Naish and LB not found elsewhere in ST (to the best of our knowledge). It should be kept in mind that finding uniquely shared lexical innovations is a difficult task. This short list will require revision in future; if the hypothesis is correct, it is expected that an increasing number of cognates and uniquely shared lexical innovations will come to light.
Table 20. Correspondences for lexical items that may constitute Burmese-Qiangic in-novations. The Naish forms are Na, apart from those marked as NX, which are from Naxi. Achang belongs to Burmish, and Hani to Loloish.
meaning Rgyalrong(S=Situ)
Tangut Naish(NX=Naxi)
Proto-Naish
Burmese Achang Hani
copula ŋu ŋwu² ŋi˩˧ ? ŋɯ³¹
star ʑŋgri gjij¹ kɯ˥ *kri kray² kʰʐə⁵⁵ a³¹ gɯ⁵⁵
forget jmɯt mjɨ² mv.pʰæ L+MH# *mi me¹ ɲi³⁵ ɲi⁵⁵
be ill ngo < *ngaŋ
ŋo² gu˩ *go
flint ʁdɯrtsa tse.miH# *tsa
to hide nɤtsɯ tsɯ˥ NX *tsu
to swallow mqlaʁ ʁv˥ *NqU <*Nqak
dry spɯ pv˧ *Spu
thick jaʁ laa¹ lo˧˥ *laC2
jump mtsaʁ tsʰo˧ *tsʰaC2
winter qartsɯ tsur¹ tsʰi˥ *tsʰu cʰoŋ³ tɕʰɔŋ³¹ tsʰɔ³1 ga³³
knee tə-mŋɑ S ŋwer² ŋwɤ.ko H# *ŋwa
sun ʁmbɣi be² bi˧ NX *bi
8 Guillaume Jacques and Alexis Michaud
Note that the inclusion of Rgyalrongic within Qiangic contradicts LaPolla’s hy-pothesis of a Rung group, distinct from Qiangic, that would include Rgyalrongic as well as Kiranti and Dulong/Rawang. LaPolla’s proposed grouping is based on the hypothesis that the morphology found across these languages is a common innovation (LaPolla 2003: 30 and references therein). However, the comparison of Rgyalrong to Kiranti reveals very little common vocabulary: a careful examination of Boyd Michailovsky’s unpublished Kiranti etymological dictionary brought out less than 150 potential cognates, which are too widespread within the Sino-Tibet-an family to be convincing instances of shared innovation. If Rgyalrong and Ki-ranti were closely related in the Sino-Tibetan family tree, one would expect more cognate vocabulary, including some lexical innovations.
The view of the Sino-Tibetan family presented in Figure 2 has the important implication that any morphology that is found in both Rgyalrong and Kiranti, or Rgyalrong and Tibetan, must be of great antiquity (predating the split between Proto-Burmo-Qiangic and other branches), and that it was lost almost without traces in Lolo-Burmese and Naish. In this light, vestigial phenomena such as the traces of vowel alternation found in the Naic language Shixing (Chirkova 2009) deserve special attention: they may point to an earlier verb conjugation system.
Why no comparison with languages closely related to the Naish languages was attempted in the present research
The phylogenetic distance between Naish, Rgyalrong and Burmese is relatively great — although we believe that they belong together with the Naish languages in a Burmo-Qiangic branch of Sino-Tibetan, as explained above. The distance be-tween Naish and Tibetan is even greater. Some justifications must be provided for referring to these distant languages in the reconstruction of Proto-Naish, in-stead of relying on data from Shixing, Namuyi and Ersu/Tosu/Lizu, which, while they do not belong to the Naish branch by our criteria, appear to be its closest relatives and could belong in a Naic group (see Figure 2). There are in fact three pressing reasons not to attempt to incorporate data from these languages at the present stage. (i) Available phonemic analyses for these languages are not fully satisfactory. A thorough synchronic description, including a complete inventory of syllables, is required before these languages can be put to use in historical com-parison. In the case of the Naish languages, a preliminary to the present research consisted in elaborating a comprehensive synchronic phonological analysis. By ‘comprehensive’, we mean an analysis which, in addition to the inventory of vowel and consonant phonemes in the language, comprises a list of all attested syllables. As the Naish languages tend to present many phonological contrasts in restricted contexts, the inventory of syllables is necessary to study the full extent of gaps
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 9
in the combinations of onsets and rhymes. For Shixing, Namuyi and Ersu, such inventories are not yet available. (ii) In addition to this practical reason, there is a methodological reason for postponing comparison with these languages: they are almost as eroded as the Naish languages, and therefore extremely difficult to use for comparative purposes. Naish, Shixing, Namuyi and Ersu have undergone an enormous amount of phonological changes independently from one another, and do not share most of their phonological innovations. Comparing them di-rectly to one another only yields a lengthy list of opaque correspondences, offering precious few insights as to how these correspondences should be sorted out and reconstructed. Since these languages are mostly isolating and have almost no in-flections (except in their tonology), we cannot rely on the reconstruction of vowel alternations to solve these issues. (iii) Last but not least, areal diffusion has had a conspicuous influence on Shixing and Namuyi, whose speakers are currently mul-tilingual, raising with extreme acuteness the classical issue of inheritance versus borrowing (about which see Aikhenvald & Dixon 2001, among others).
Appendix 2. Examples of five rhymes of Proto-Naish (*a, *i, *o, *u and *aC1/*aC2) with comparanda in the conservative languages and proposed reconstructions.
The question mark after a reconstructed form indicates that this form has other possible origins, and that the form indicated is a rule-of-thumb hypothesis. The “Ref[erence]” consists of (i) the proto-vowel, (ii) the number assigned to the vowel correspondence among the three Naish languages under study, and (iii) the num-ber assigned to the cognate set. In the “HTB” column, we indicate the page num-ber corresponding to the etymon in Matisoff ’s handbook (2003). The words pro-vided in the “Rgyalrong” column are Japhug Rgyalrong forms, except those with the mention “(Situ)”, which are Situ Rgyalrong forms from Huang Liangrong & Sun Hongkai 2002. The notation of the tones for Na disyllables follows the conven-tions set out in Michaud (2008). Finally, it must be emphasised that the data in the “other languages” column are not part of the comparative study carried out here: these potential cognates are provided solely as stepping-stones for future com-parative work with these languages (Tangut, Pumi and Lisu). For Pumi, SL refers to the Shuiluo dialect (unpublished fieldwork data), and LP to the Lanping dialect (Lu Shaozun 2001). Personal communications from James Matisoff are labelled “(JAM)”.
10 Guillaume Jacques and Alexis Michaud
Tabl
e 21
. Rh
yme
*-a
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
othe
r la
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
to w
ina1
.01
βʁa
ŋgɑ˧
ʁɑ˥
ŋga/
aC1
stre
ngth
4a1
.02
170–
4Ta
ngut
ɣie
<*
C-k
akɑ˧tɯ
˥ʁɑ˥
ʁɑ˩z
i˩N
ka/a
C1
bitte
ra1
.03
164–
8kh
a³kʰ
akʰɑ˧
qʰɑ˥
kʰɑ˧
kʰa/
aC1
to st
ep a
c-cr
oss
a1.0
4mɢl
aʁɑ˧
(ɖɯ˧)
ʁɑ˧
ŋga/
aC1 (o
r N
ka/a
C1)
diffi
cult5
a1.0
5nq
adk
alo
.hɑM
lu˧hɑ˧
Cka
/aC
1
knee
6a2
.01
tə-m
ŋɑ
(Situ
)Ta
ngut
ŋw
er²
<*rŋ
wa
ŋwɤ.k
o H#
ŋwɑ˩
tu˥
ŋwa
bow
la2
.02
Pum
i kʰwǎ
kʰwɑ˥
qʰwɤ˩˧
kʰwɤ˩
kʰw
a
hoof
a2.0
317
0tɯ
-qa
Pum
i kw
akʰ
wɑ˧
qʰwɤ.ʂ
eL#kʰ
wɤ˥
bie˥
kʰw
a
a pa
ira3
.01
dze˩
dze˥
dza
to lo
cka3
.02
tse˩
tse˥
tsa
whe
at7
a3.0
316
2–5
ndza
ca³
zadz
e˧dz
e.lɯ
Mdz
e˥dz
a
naem
orhe
-du
s gor
ala3
.04
se˩
se˧
se˧
sa
stee
l (fo
r fli
nt)
a3.0
5ʁd
ɯrt
sats
e˧mɑ˩
tse.m
iH#
tse˧
mie˥
tsa
salt
a3.0
617
2ch
a³tsʰw
atsʰe˧
tsʰe˥
tsʰe˧
tsʰa
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 11
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
othe
r la
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
neph
ewa3
.07
171–
2tɯ
-ftsa
tsʰa
-bo
dze˧ɯ˧
ze.vL
ze˧
Cdz
a
to w
alk
a3.0
8se˥
se˧s
e˧sa
to b
orro
wa4
.01
162–
5hŋ
a³rɲ
aŋi˧
ŋi˥
ŋi˧
ŋi/a
fish
a4.0
216
2ŋa
³ɲa
ŋi˧
ŋi.zu
#Hɲi˩z
e˥ŋi
/a
span
a5.0
1tɯ
-tɣa
thw
a²m
tʰo<
*mtw
aʈi˥
twa
toot
ha5
.02
171–
2tɯ
-ɕɣa
swa³
so <
*sw
ahɯ
˧hi˥
fi˧tʰu
˧Sw
a
rain
a5.0
317
1–3
rwa²
hɯ˩
hi˩˧
fi˧Sw
a
lake
a5.0
4m
tsʰo
< *m
-sw
ahɯ
˥hi
.nɑ.
miLM
+#H
fi˩Sw
a
cow
a5.0
5Ta
ngut
ŋw
e² <
*ŋ
wa
ɯ˧
i˥vi˧
wa
to fa
ll (r
ain)
a6.0
1ŋg
ra “t
o fa
ll”ky
a¹gɯ
˧gi˥
gra
mea
ta6
.02
sa³
ɕaʂɯ
˧ʂe˥
ʂɯ˧
ɕa
eart
h8a6
.03
ʈʂɯ˧
ʈʂe˥
ʈʂɯ˧
tra
axe
a7.0
117
1–2
tɯ-r
palɑ˩m
be˧
bi.m
iL(S
)mba
to st
umbl
ea7
.02
pe˧
kʰɯ
.piM
(S)p
a
Bai (
ethn
ic gr
oup)
a7.0
3le˧b
v˧ɬi.
bvM
Sla
linen
a7.0
4pʰ
e˩pʰ
i˩˧(S
)pʰa
12 Guillaume Jacques and Alexis Michaud
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
othe
r la
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
knife
a7.0
516
2dh
a³zɯ
˧tʰe˧
sɯ.tʰ
iLtʰa
nobl
ea7
.06
sɯ.pʰiM
sɯ˩pʰie˩
pʰa
soul
9a7
.07
162
tɯ-r
lase
e foo
tnot
ebl
a/br
lao˩
he˧
æ.ɬi
L+#
Hʁɔ˩ɬi
e˩la
thin
a7.0
816
2m
bapa
³m
be˧
bi˥
tʰɑ˧ b
ie˥
mba
snow
a7.0
917
2tɤ
-jpa
mbe˧
bi˥
vie˧
Smba
villa
gea7
.10
hi˧m
be˧
fv.b
iLɖɯ
˧bie˧
mba
to d
oa7
.11
paby
edbe˧
i˥vi
e˧C
ba
tea
a7.1
248
le˥
li˩˧lie˩
l a
moo
na7
.13
162–
4tɯ
-sla
la¹
zla-
bale˩
ɬi.m
iMɬie˧m
ie˧
Sla
ear
a7.1
416
2–5
tɯ-r
nana
³rn
ahe˥tsɯ˩
ɬi.pi
L#
ɬie˧tu
˥la
trou
sers
a7.1
516
3–5
Tang
ut lj
ii¹
< *lj
aale˧
ɬi.qʰ
wɤL
ɬie˥kʰwɤ˥
Sla
fem
ale
a7.1
617
5m
a (s
uffix)
me˧
mi
mie
ma
ask
for
a7.1
7m
e˥m
i˩m
ie˩
ma
butte
rfly
a7.1
8qa
mba
lɯla
pʰe˧
le˩
pʰi.l
i L#
pʰie˧li
e˥pʰ
a la
rabb
it10a7
.19
qala
tʰo˧le˧
tʰu.li
Mtʰu
˧lie˥
la
this
year
a7.2
0tɯ
-xpa
ʈʂʰɯ
˧be˧
tsʰi.
i(M)
tsʰɯ
˧vie˧
Cba
who
a7.2
1ə˧
ne˩
ni˩˧
na
4. I
t is l
ikel
y th
at “t
o w
in”
*ŋga
/aC
1 and
“str
engt
h” *
Nka
/aC
1 orig
inal
ly b
elon
g to
the
sam
e ro
ot, b
ut th
ey n
eed
to b
e di
stin
guish
ed a
t the
Pro
to-N
aish
st
age.
A re
latio
nshi
p w
ith B
urm
ese
a³ a
nd it
s Lol
o-Bu
rmes
e co
gnat
es (M
atiso
ff 20
03: 1
70) i
s pos
sible
.
5. W
e al
so fi
nd fo
rms s
uch
as L
ahu
/ha¹
¹/ “d
ifficu
lt” (M
atiso
ff 19
88: 1
066)
, whi
ch co
uld
poin
t to
an a
ltern
ativ
e et
ymol
ogy.
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 13
6. Th
e Na w
ord
is pr
onou
nced
[ŋwɤ]
. Sin
ce th
e com
bina
tion
of an
initi
al v
elar
and
a rhy
me /
wɑ/
is n
ot at
test
ed in
Na,
one
may
cons
ider
that
the c
ontr
ast
betw
een
the
rhym
es /w
ɤ/ a
nd /w
ɑ/ is
neu
tral
ised
in th
is co
ntex
t, an
d he
nce
com
pare
Na
[ŋwɤ]
with
Laz
e [ŋ
wɑ]
.
7. Th
is w
as o
rigin
ally
a n
omin
alise
d fo
rm o
f the
ver
b ‘to
eat
’; a
sem
antic
cha
nge
from
‘foo
d’ to
‘whe
at’ o
ccur
red
in th
is et
ymon
. The
free
ver
b “t
o ea
t”
in N
aish
, /dzɯ˥/
in N
a an
d /n
dzɯ˧/
in N
axi,
poin
ts to
a re
cons
truc
tion
*ndz
i in
Prot
o-N
aish
, whi
ch is
not
com
patib
le w
ith th
e vo
wel
in th
e la
ngua
ges
of re
fere
nce.
The
*-a
/ *-i
alte
rnat
ion
foun
d in
this
pair
of w
ords
can
onl
y be
a tr
ace
of m
orph
olog
y. Th
e rh
yme
*-i o
f the
ver
b m
ight
be
the
resu
lt of
the
fusio
n of
the
root
with
a su
ffix.
Suc
h a
phen
omen
on is
foun
d in
Rgy
alro
ngic
lang
uage
s: in
Japh
ug R
gyal
rong
, tra
nsiti
ve v
erbs
with
ope
n-sy
llabl
e -a
fina
l (in
clud
ing
ndza
“to
eat”,
the
cogn
ate
of P
roto
-Nai
sh *n
dzi)
have
a n
on-p
ast f
orm
sing
ular
stem
in -e
(for
inst
ance
/ndz
e/ “h
e ea
ts”) t
hat r
esul
ts fr
om th
e fu
sion
of th
e roo
t vow
el w
ith a
suffi
x *-
jə at
test
ed as
a fr
ee fo
rm in
oth
er R
gyal
rong
ic la
ngua
ges (
Jacq
ues 2
004:
356)
. An
expl
anat
ion
for t
he fo
rm *n
dzi i
n N
aish
is th
at it
repr
esen
ts th
e ge
nera
lisat
ion
of th
e no
n-pa
st fo
rm o
f the
ver
b, th
ereb
y pr
eser
ving
a tr
ace
of a
hist
oric
al st
age
whe
n N
aish
lang
uage
s had
ve
rbal
mor
phol
ogy
of th
e ty
pe th
at R
gyal
rong
pre
serv
es to
this
day.
8. Th
is fo
rm is
per
haps
rela
tabl
e to
the
seco
nd sy
llabl
e of
Lah
u /m
i²¹ch
a⁵³/
“ear
th” (
JAM
).
9. T
wo
com
petin
g Bu
rmes
e et
ymol
ogie
s exi
st fo
r thi
s ety
mon
: lip
pra
¹ “so
ul” a
nd h
la¹ “
beau
tiful
” (M
atiso
ff 20
03: 6
2).
10.
Sim
ilar n
ames
are
foun
d in
oth
er la
ngua
ges,
for i
nsta
nce
Lahu
/tʰɔ
⁵³la
²¹/ “
year
of t
he ra
bbit”
(JA
M).
14 Guillaume Jacques and Alexis Michaud
Tabl
e 22
. Vo
wel
*-i
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erla
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
goat
i1.0
131
5tsʰɤ
tch
ittsʰɯ
˥tsʰɯ˧˥
tsʰɯ
˩tsʰi
oat11
i1.0
218
9mɯ˩zɯ˧
mv.
zɯL
mv˥
zɯ˥
zi
sout
hi1
.03
i˧ʈʂʰɯ˧m
ɯ˩
i.ʈʂʰɯ
.mi
MH
Li˧ʈʂʰɯ˥m
ie˧
tɕʰi
gras
si1
.04
zɯ˧
zɯ˥
zɯ˧
zi
skin
i1.0
518
9tɯ
-ndʐ
ire
²ɯ˧pʰi˩
ɯ˩˧
ʐɯ˧k
v˥ri
yello
wi1
.06
191
ʂɯ˩
ʂɯ˧ k
ɔ˥ læ
˥ɕi
to k
now
i1.0
720
6sɯ
ssi¹
ɕes
sɯ˧
sɯ˥
sɯ˩
si
to d
ie12
i1.0
818
9si
se²
ɕiʂɯ
˧ʂɯ
˧sɯ
˩rs
i
to th
read
(b
eads
)i1
.09
Lisu
sɯ³³
sɯ˧
sɯ˥
si
new
i1.1
034
4ɕɤɣ
sac
ʂɯ˥
ʂɯ˧˥
ʂɯ˩tsɑ˩
ɕi
to ti
ei1
.11
tsʰik
³³
(Situ
)tsɯ˧
tsɯ˥
tsɯ˧
tsi
mor
ning
i1.1
2m
v˧sɯ
˥tɯ˧
mv.
sɯ
L+M
H#
si
girl
i2.0
118
7tɯ
-me
min
³m
i˥m
v˩˧m
v˩m
i
fire
i2.0
120
6sm
im
i³m
em
i˧m
v˥m
v˧m
i
to h
ear
i2.0
2Pu
mi m
ɛ³kʰ
o˧m
i˧m
v˥m
v˧m
i
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 15
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erla
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
to fo
rget
i2.0
3jm
ɯt
me¹
le˧m
i˥m
v.pʰ
æ
L+M
H#
læ˥m
v˩m
i
low
er si
dei2
.04
mɯ˨tʰ
æ˧
mv.ʈʰæ
Mv˧ʈʰæ
˥m
i
nam
ei2
.05
296
tɤ-r
mi
maɲ
²m
ing
mi˩
mv.ʈʂ
æ M
H#
mi
star
13i3
.01
212
ʑŋgr
ikr
ay²
kɯ˩
kɯ˥
tsi˧
kri
gallb
ladd
eri3
.02
189
tɯ-ɕ
krɯ
tsaɲ³
khr
e²m
kʰris
kɯ˩
kɯ˥
tsi˩
kri
med
icin
ei3
.03
189
che³
rtsi
ʈʂʰɚ˧ɯ
˧ʈʂʰæ
.ɯ#H
tsʰɯ
˧fi˧
rtsʰ
i Sw
ri
tight
i3.0
430
5gr
im-p
okɯ
˥ts
i˥kr
i
liver
i4.0
129
7tɯ
-mtsʰi
saɲ³
mtɕʰin
sɚ˥
si˩˧si˧
siN
woo
di4
.02
347
sisa
cɕi
ŋsɚ˧
si˥si˧
siN
to sh
ave
i4.0
3si˧˥
si˧si˧˥
siN
hot
i4.0
4tsʰɚ˧
tsʰi˧
tsʰi˧
tsʰiN
to p
lane
do
wn
i5.0
1tʰi˩
tʰi˩
tʰi
wat
eri5
.02
tɯ-c
igi˩
dʑi˩
gi
to fl
owi5
.03
jiti˩
ʑi˧
ji
tong
uei5
.04
215
lhya
²ltɕ
ehi˥
hi.m
iLɕi˩m
ie˩
hi
two
i5.0
543
4ʁn
ɯs
hnac
gɲis
ŋi˧
ŋi˧˥
ŋi˥g
v˥ŋi
/a
swee
t14i5
.06
cʰi
kʰi˩
tɕʰi˩
tsʰi˥
kʰi
to se
lli5
.07
ntsɣ
etɕʰi˧
tɕʰi˧
tɕʰi
thor
ni5
.08
kʰi˧
tɕʰi˥
tɕʰi˧
tu˥
tɕʰi
16 Guillaume Jacques and Alexis Michaud
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erla
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
mun
tjac
i5.0
918
9kh
ye²
kʰi˩
tɕʰi˩
tsʰi˥
kʰi
saliv
a15i5
.10
tɯ m
ciki˥
tɕi˩
tɕi˩
tɕi
sadd
lei5
.11
Pum
i: stʃɛ
tʂho
(LP)
, ɕi
² (SL
)
ʐwɑ˧
ki˥
kʰo˩
tɕi˩
tɕi˩q
ʰɔ˥
tɕi
smal
li5
.12
xtɕi
ki˥
tɕiL
tɕi˩
tɕi
to sl
eep16
i5.1
350
0nɯ
ʑɯw
‘ipi˥
ʑi˧˥
zi˩
ji
to w
alk
i5.1
4ŋk
eŋg
i˧ŋg
i
to lo
sei5
.15
ŋi˥
ni
clou
di5
.16
ki˩
tɕi˧
tɕi˩sɯ˥
tɕi
urin
ei5
.17
tɯ-r
mbi
Pum
i bi²
(SL)
mbi˧
mbi
pus
i6.0
1pr
aɲ²
mbɚ˩
bæ˩˧
bæ˩
priN
shor
ti6
.02
ndɚ˧
ɖæ˧
ɖæ˧˥
rdiN
grai
ni6
.03
lɚ˥
ɻæ˩˧
læ˩
rliN
neck
17i6
.04
mgr
inPu
mi kɛ²
(S
L)kjɚ˧
pɚ˩
ʁæ.ʈv
H#
ʁæ˧sɯ˧pɑ˧
C-N
kriN
resin
i6.0
5th
o˧ŋg
jɚ˩
tʰo.ʁ
æL
thu˧ʁæ
˥C
-ŋgr
iN
rope
i6.0
6tɯ
-mbr
im
breŋ
bæ˧
briN
gues
ti6
.07
tɯ-p
ibɚ˧
hi.b
æ#H
briN
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 17
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erla
ngua
ges
Nax
iN
aLa
zePr
oto-
Nai
sh
long
i6.0
828
0–2
zri
riŋPu
mi ʂɛ¹
(S
L)ʂɚ˩
ʂæ˧
ɕiN
hunt
i6.0
9ʂɚ˩
kʰv˧ʂæ˧˥
ɕiN
artic
ulat
ion
i7.0
1tɯ
-rtsɤɣ
chac
tsʰig
sʈʂɚ˥
ʈʂæ˧˥
tsɯ˩
rtsi
was
hi7
.02
χtɕi
(Situ
rtɕi
)ʈʂʰɚ˧
ʈʂʰæ˧
bɑ˥lɑ
˥ tsʰɯ˩
rtsʰ
i
med
icin
ei7
.03
che²
rtsi
ʈʂʰɚ˧ɯ
˧ʈʂʰæ
.ɯ#H
tsʰɯ
˧ fi˧
rtsʰ
i
wai
sti7
.04
i.ʈʂæ
L+M
H#
i˩tsɯ
˩rt
si
11.
Prot
o-Lo
lo-B
urm
ese
*zəy
² “ba
rley”
(JA
M).
12.
The
reco
nstr
uctio
n of
the
clus
ter *
rs fo
r thi
s wor
d re
sults
from
the
appl
icat
ion
of th
e sa
me
prin
cipl
e as
for o
ther
cas
es w
here
a re
trofl
ex in
itial
in N
a an
d N
axi c
orre
spon
ds w
ith a
den
tal i
nitia
l in
Laze
. This
reco
nstr
uctio
n is
not s
uppo
rted
by
com
para
tive
evid
ence
from
the
cons
erva
tive
lang
uage
s. Th
e cl
uste
r in
*rsi
coul
d be
a tr
ace
of m
orph
olog
y th
at h
ad d
evel
oped
in P
roto
-Nai
sh.
13.
Mat
isoff
(198
0) h
as p
ropo
sed
a de
taile
d et
ymol
ogy
for t
his e
tym
on co
mm
on to
Nai
sh, L
olo-
Burm
ese
and
Qia
ngic
lang
uage
s.
14.
This
etym
on is
per
haps
rela
ted
to B
urm
ese
khyu
i² (c
f. M
atiso
ff 20
03: 1
82).
15.
This
root
coul
d be
rela
ted
to ‘w
ater
’ (cf
. Mat
isoff
2003
: 451
).
16.
The c
orre
spon
denc
e of i
nitia
ls fo
r thi
s ite
m is
pro
blem
atic
. The r
econ
stru
ctio
n pr
opos
ed h
ere r
ests
on
the h
ypot
hesis
that
*ji c
hang
ed to
/zi/
in L
aze.
Cru
cial
evi
denc
e w
ould
com
e fr
om o
ther
inst
ance
s of t
he co
rres
pond
ence
/i:ʑ
i:zi/.
17.
The
rhym
e in
the
Nax
i dia
lect
stud
ied
here
is /æ
/: /k
jæ˧pɚ˩
/; ho
wev
er, t
his i
s due
to a
n in
nova
tion
foun
d in
this
dial
ect:
the
mer
ger o
f /ɚ/
and
/æ/ (
to
/æ/)
afte
r S-,
TS- a
nd K
j-, w
here
S- s
tand
s for
cor
onal
fric
ativ
es, T
S- fo
r cor
onal
affr
icat
es, a
nd K
for v
elar
stop
s. Th
e co
nser
vativ
e fo
rm is
/kjɚ˧pɚ˩
/, as
fo
und
in th
e var
iety
of N
axi s
poke
n in
the c
ity o
f Liji
ang (
Fang
Guo
yu &
He Z
hiw
u 19
95: 4
32),
whe
re th
e con
tras
t bet
wee
n /ɚ
/ and
/æ/ i
s pre
serv
ed in
thes
e co
ntex
ts. N
ote t
hat *
Nkr
iN an
d *ŋ
griN
do
not f
ollo
w th
e sam
e pho
netic
evol
utio
n as
*kri,
oth
erw
ise o
ne w
ould
expe
ct th
e cor
resp
onde
nce ŋ
gɯ:kɯ
:ndz
i.
18 Guillaume Jacques and Alexis Michaud
Tabl
e 23
. Vo
wel
*-o
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
brig
ht18
o1.0
1m
bu˧
bu˧
mbo
dike
o1.0
2m
bu˧
bu˧
bu˩tu
˩m
bo
hom
eo1
.03
mi˧u
˩ɑ.ʁu
LMɑ˥
u˥o
youn
ger
siste
ro1
.04
gu˧m
e˧gu
.miM
gu˧m
ie˧
go
win
now
ing
fan
o1.0
5m
u˩pʰ
i.mu L
#m
u˥m
o
lunc
ho1
.06
ʐu˧
ʐu.d
zɯL
ʐu˧
Cro
cold
(w
eath
er)
o1.0
726
2ɣɤ
ndʐo
graŋ
-mo
ɖʐu˥
ɖʐu˧
ndro
this
mor
n-in
go1
.08
ʈʂʰɯ
.ʂu(M
)tsʰi˥ʂu˥
ɕo
to ru
n aw
ayo1
.09
294
pʰɣo
mbr
ospʰ
u˩pʰ
u˩pʰ
u˥pʰ
o
head
19o1
.10
tɯ-k
um
goku
˧ly˧
ʁu.qʰwɤ L
#u˧
tu˥
SNko
corp
seo1
.11
265
Tang
utm
jij²
ʂɯ˧m
u˩hi
.mu L
#hi˩m
u˩m
o
to sp
read
o1.1
2ɕkʰo
khaŋ
³kʰ
u˧kʰ
o˥kʰ
u˧kʰ
o
tom
orro
w20
o1.1
3fs
osa
ŋ ɲi
nso˩ŋ
i˧so
.ŋi H
#m
i˧su˥
so
pine
o1.1
426
4tɤ
tʰoth
aŋ³
tʰaŋ
tʰo˧n
dzæ˩
tʰo.d
zi L
#tʰu
˧si˧
tʰo
be il
lo1
.15
ngo
ŋgu˩
go˩
ŋgo
mus
hroo
mo1
.16
183–
4tɤ
jmɤɣ
hmui
²m
og ɕ
am
u˥m
u˧˥m
u˩ʈʂʰwɤ˩
mo
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 19
18.
Poss
ibly
rela
ted
to L
ahu
/ba³
³/ “b
right
”, th
ough
the
vow
el co
rres
pond
ence
s are
pro
blem
atic
.
19.
A co
mpa
rison
with
Tib
etan
dbu
“hea
d” a
nd B
urm
ese
u² “h
ead”
is te
mpt
ing,
but
the
vow
els d
o no
t mat
ch.
20.
In L
olo-
Burm
ese,
one
finds
cogn
ates
that
poi
nt to
a rh
yme
*-ak
rath
er th
an *-
aŋ a
s do
the
Nai
sh a
nd T
ibet
an fo
rms.
Tabl
e 24
. Vo
wel
*u
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
ring21
u1.0
1lɑ˩pɚ˥
lo.p
v L+M
H#
lɑ˩ʈʰɯ˩
pru
com
b22u1
.02
pɚ˥
pv.m
iLpr
u
yak
u1.0
3qa
mbrɯ
mbr
ibɚ˩
bv˩˧
bv˥
bru
whi
te23
u1.0
4wɣr
umm
pʰru
mpʰɚ˩
pʰv˩
pʰv˥
pʰru
to h
oe u
pu1
.05
pɚ˩
pv˧˥
pv˩
pru
vert
ical
u2.0
1tsɯ˩
gɤ.ts
iLMts
u
to si
tu2
.02
mdzɯ
ndzɯ
˩dz
i˩dz
y˥nd
zu
win
ter
u2.0
3qa
rtsɯ
choŋ
³mɯ˧tsʰɯ
˧tsʰi˥
mu˧
tsʰy˧b
ie˧
tsʰu
to h
ide
u2.0
4nɤ
tsɯ
tsɯ˥
tsu
bald
u3.0
1ʁu
.bv M
H#
bv˥
bu
bug
u3.0
1pu
i³m
bubi˧d
i˩bv˥
bu
to la
y eg
gsu3
.02
57Pu
mi p
ɜ³
(SL)
bv˩
bv˩
bu
pan
u3.0
3Ta
ngut
.wju
¹ < C
pobv˧
v˥v˧
mie˧
Cbu
mat
erna
l unc
leu3
.04
u³a-
kʰu
ə˩gv˧
ə.v M
H#
æ˧v˥
Cgu
20 Guillaume Jacques and Alexis Michaud
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
to se
wu3
.05
141
tʂɯ
βkh
yup
ndru
bʐv˧
ʐv˧˥
ʐv˩
C-r
u
crea
sed,
w
rinkl
edu3
.06
aʁrɯ
ʁru
lə.ʐv
.ʐvH
C-r
u
to p
ass (
time)
u3.0
7gv˧
gv˧
gv˧
gu
body
u3.0
819
8tɯ
-skʰ
rɯku
iy²
sku
gv˧
gv.m
iMɖɯ
˧gv˧ɖɯ
˧dzi˧
gu
plow
u3.0
9æ
.gv L
ɭɔ˥g
v˥gu
to fi
x24u3
.10
ku¹
gv˩
gv˩
gu
nice
u3.1
1gv˧
gv˧
gu
bent
u3.1
235
7ŋg
ɤɣgu
ggv˩
lɑ.g
vLMlɑ˥g
v˥gu
owl
u3.1
3pɣ
ɤ-kʰɯ
bu˩fv
˧bu
˧lu˧fv
˧hu
year
u3.1
4fsɤqʰe
Pum
i ko²
(S
L)kʰ
v˥kʰ
v˧˥kʰ
v˧˥kʰ
u
to st
eal
u3.1
518
2mɯ
rkɯ
khui
³rk
ukʰ
v˩kʰ
v˧kʰ
v˧kʰ
u
clas
sifier
(men
)u3
.16
kv
kv
kv
ku
Bai e
thni
c gr
oup
u3.1
7le˧b
v˧ɬi.
bvM
la b
u
to b
ark25
u3.1
8lv˩
kʰɯ˥ l
v˥lu
enou
ghu3
.19
357
lok
lv˩
lv˩
lv˩
lu
to w
ind
(thr
ead)
26u3
.20
ndv˧
lv˥
lu
larv
au3
.21
qajɯ
lv˥
mu˧
kʰv˥
lv˥
lu
to g
raze
u3.2
2lɤɣ
lv˥
lv˧˥
lv˩
lu
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 21
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
bear
du3
.23
mv˧
tsɯ˥
mv.
tsɯ
MH
#m
v˧tsɯ˥
mu
to d
igu3
.24
184
tu³
ndv˧
dv˥
ndu
poiso
nu3
.25
357
tɤ-n
dɤɣ
dug
ndv˩
dv˩
ɖv˩
ndu
win
gu3
.26
285
Pum
i dio
³ (S
L)nd
v˧pʰ
i˩dv˩
dv˩tsʰɯ
˩nd
u
sickl
eu3
.27
ʂv˥k
v˧ʂv.g
vLhu
˧gv˥
Nku
to st
eam
u3.2
8Pu
mi b
u³
(SL)
pv˥
bv˩
bv˧
Npu
sadd
lecl
oth
u3.2
9ki˥n
v˧tɕ
i.nv L
+MH
#nu
thun
der
u3.3
0mɯ˧ŋ
gv˧
mv.
gv#H
mu˧
gv˧
ŋgu
sinew
u3.3
1tɯ
-ŋgr
uŋg
v˧ŋg
u
nine
u3.3
218
2kɯ
ngɯ
tku
i³dg
uŋg
v˧gv(˧
)gv˥g
v˥ŋg
u
to cr
yu3
.33
182
ɣɤw
uŋu
i²ŋu
ŋv˩
ŋv˩
ŋv˧
ŋu
silve
ru3
.34
414–
41
5ŋw
e²dŋ
ulŋv˩
ŋv˩
ŋv˧
ŋu
pric
eu3
.35
183–
18
4ɯ
pʰɯ
phui
³kɑ˧pʰv˧
ʁɑ.pʰv
#Hʁɑ˧pʰv˧
pʰu
mal
eu3
.36
pʰo
pʰv˧
pʰu
uncl
e’s u
ncle
u3.3
7tɯ
-rpɯ
pʰv˧
ə.pʰvM
æ˩pʰv˩
pʰu
blad
der
u3.3
8Li
su
si³¹pʰu
³¹sɯ
˥pv˩
sɯ.p
v L#
sɯ˧p
v˥pu
ambe
ru3
.39
pv˧ʂɯ˩
pv.ʂɯ
L#
?pu
22 Guillaume Jacques and Alexis Michaud
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
rock
u3.4
0rd
o (?
)æ
.ʈv L
ML
rtu
inte
stin
eu3
.41
180
tɯ-p
uu²
pʰo-
babv˧
v˧nɑ
˥Sb
u
garli
cu3
.42
ɕku
sgog
kv˧
kv˥
fv˧
Sku
kidn
eys
u3.4
3tɯ
-m
bɤtɯ
mm
by˧ly
˧bv.lɯ
Lv˩
li˩Sm
bu
dry
u3.4
4spɯ
pv˩
pv˧
læ˩fv
˩Sp
u
to g
o ou
t27u3
.45
thut
tʰv˧
tʰv˧
ʈʰv˧
tʰu
to co
ntam
inat
eu3
.46
ʈʂv˩
ʈʂv˩
> ʈʂ
v˥ʈʂ
v˩tr
u
to p
lant
u3.4
7tv˩
tv˧
ʈv˧
Stu
stra
ight
u3.4
8as
tutv˩
tv.tv
Lʈv˧ʈv
˥læ˥
Stu
thou
sand
u3.4
929
4th
oŋst
oŋtv˧
ʈv˧
Stu
hole
u3.5
0do
ŋpɑ˥ t
v˥tu
sleev
eu4
.01
lɑ˩jɤ˥kʰo˧
i.qʰv
Ljɑ˩qʰɔ˩
qʰU
swal
low
u4.0
2m
qlaʁ
ko˧
ʁv˥
ʁɔ˧
NqU
cave
u4.0
328
5ŋg
y˩kʰ
o˧ʁw
ɤ.qʰv
Mlæ˩qʰɔ˩
qʰU
thro
atu4
.04
tɯ-r
qolk
ogqv.ʈʂ
æ M
H#
qɔ˧tsɯ˥
qU
horn
u4.0
518
2ta
-ʁrɯ
khyu
i²ru
kʰo˧
qʰv˧
qʰɔ˥
qʰU
flyu5
.01
255
mbɚ˧lɚ
˥bv.ɻ
#Hbɔ˧ɭɔ˥
bu r28
to h
old
u6.0
1ʈʂʰʷɚ˥
ʈʂʰɻ˧˥
tsʰv˩
*rtsʰU
lung
su6
.02
tɯ-r
tsʰɤ
s <*
rtsʰɔs
)ch
utʈʂʰʷɚ˥
ʈʂʰɻ˧˥
tsʰv˩
*rtsʰU
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 23
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
to co
ugh
u6.0
3ʈʂʷɚ˧
ʈʂɻ
tsv˧
*rts
U
Tabl
e 25
. Rh
ymes
*aC
1 and
*aC
2
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
ches
tC
1.01
Tang
utɣa
r <*C
-r-
kaC
kɑ˩
ʁɑ˧ (ʁɑ
.pvM
)N
ka/a
C1
to fe
ll (a
tr
ee)
C1.
02nd
ɑ˥dɑ˧˥
dɑ˧˥
ndaC
1
drum
C1.
03nd
ɑ˧kʰɤ˩
dɑ.kʰɤ
Ldɑ˥kʰɤ˥
ndaC
1
all
C1.
04tɑ˥
tɑ M
H# (ɖ
ɯ.
tɑ M
H# )
tɑ˥ (ɖɯ
˧ tɑ˥)
taC
1
slant
edC
1.05
lɑ.tɑ
LMlɑ˧tɑ
˥la
C1 ta
C1
how
muc
hC
1.06
qʰɑ.
kv M
H#
kʰɑ˧
i˥kʰ
a/aC
1
in fr
ont o
fC
1.07
ʁu.dɑM
u˥dɑ˩
daC
1
to co
ver
C1.
08fk
aβŋg
ebs b
kab
kɑ˥
qɑ˩
qɑ˥
ka/a
C1
to w
eave
C1.
0931
8taʁ
rak
btag
sdɑ˩
dɑ˩
daC
1
blac
kC
1.10
317
ɲaʁ
nak
nag
ponɑ
˩nɑ˧˥
naC
1
shar
pC
1.11
318–
319
thak
tʰɑ˥
tʰɑ˧˥
tʰɑ˧˥
tʰaC
1
to h
itC
1.12
lɤtlɑ˥
lɑ˧˥
lɑ˧˥
laC
1
wol
f29C
1.13
qapa
rm
pʰar
pʰɑ˩
kʰɯ˧
pʰɑ˥
pʰaC
1
24 Guillaume Jacques and Alexis Michaud
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
vat,
woo
den
basin
C2.
01lo˧
lo˧
lu˥
laC
2
to b
e as
ham
ed28
C2.
01ʂɤ˥n
do˩
ʂɤ.d
o MH
#hæ
˩du˥
ndaC
2
tast
yC
2.02
so˩
so˩
su˥
saC
2
to le
an
agai
nst
C2.
03tʰo
˩tʰo
˩tʰu
˥tʰa
C2
slope
C2.
04to˥
to˩
tu˧b
ie˧
taC
2
to h
ugC
2.05
to˩to
˧to
.toM
tu˧
taC
2
grue
lC
2.06
ho˥
hu˧
haC
2
to se
eC
2.07
do˩
do˩
daC
2
valle
yC
2.08
lo˧
lo˩
laC
2
to cl
imb
C2.
09nd
o˧do
˥gɤ˧d
u˥nd
aC2
to ju
mp
C2.
10m
tsaʁ
tsʰo˧
tsʰo˧
tsʰu˥
tsʰa
C2
need
leC
2.11
342
taqa
βap
kʰab
ko˩
ʁu˩˧
u˩N
kaC
2
to st
udy
C2.
12so˩
so˩
su˩
saC
2
to w
ork
C2.
13lu
plo˧b
e˧lo
.iMlu˧v
ie˧
laC
2
pig
318–
319
paʁ
wak
pʰag
bu˩
bu˩˧
wɤ˧
SbaC
hand
319
tɯ-ja
ʁla
kla
glɑ˩o˧
lo.qʰwɤLM
lɑ˩pʰie˩
laC
1/laC
2
brea
th31
7sa
ksr
ogsɑ˥
so˧˥
sɑ˩
saC
1/saC
2
thic
kjaʁ
lɑ˥
lo˧˥
ɑ˥pɤ˥ l
u˧la
C1/la
C2
Approaching the historical phonology of three highly eroded Sino-Tibetan languages 25
mea
ning
Ref
HTB
Rgya
lrong
Burm
ese
Tibe
tan
Oth
erN
axi
Na
Laze
Prot
o-N
aish
deep
317
rnaʁ
nak
ho˥
ɬo˧˥
hɑ˩
laC
1/ laC
2/Sn
aC1
to k
illqʰ
o˧˥kʰɑ˩
kʰaC
1/aC
2
21.
Mat
isoff
(200
3: 69
, ft.1
01) c
ites L
ahu
and
Pum
i wor
ds th
at co
uld
be co
gnat
e to
this
root
.
22.
This
form
is p
roba
bly
rela
ted
to B
urm
ese p
hri³
and
othe
r com
para
nda
cite
d in
Mat
isoff
(200
3: 25
–26)
, tho
ugh
the v
owel
corr
espo
nden
ce re
mai
ns to
be
exp
lain
ed.
23.
Ano
ther
pos
sible
etym
olog
y for
this
etym
on is
Bur
mes
e phr
u² an
d its
Lol
o-Bu
rmes
e cog
nate
s (JA
M).
How
ever
, the
Nai
sh d
ata d
o no
t allo
w to
choo
se
betw
een
thes
e tw
o hy
poth
eses
.
24.
The
corr
espo
ndin
g Lo
lo-B
urm
ese
root
mea
ns ‘r
ecov
er fr
om il
lnes
s’ (J
AM
).
25.
A co
mpa
rison
with
Pro
to-L
olo-
Burm
ese
*laŋ
(Mat
isoff
2003
: 495
) is t
empt
ing,
but
the
vow
els d
o no
t mat
ch, a
s Pro
to-N
aish
*lo
wou
ld b
e ex
pect
ed.
26.
Plau
sibly
rela
ted
to L
ahu
/lɔʔ⁵/
“spi
ndle”
(JA
M).
27.
The
Burm
ese
form
mea
ns ‘t
o ta
ke o
ut’.
28.
We
susp
ect t
hat t
he fo
rms f
or ‘fl
y’ in
Laz
e an
d N
axi r
esul
t fro
m ri
ght-
to-le
ft vo
wel
har
mon
y, a
spor
adic
phe
nom
enon
in d
isylla
bles
(the
mor
e fr
e-qu
ent a
wor
d, th
e m
ore
prop
ensit
y it
has t
owar
ds v
owel
har
mon
y), l
ikew
ise fo
r ‘ki
dney
s’ in
Nax
i.
29.
The
Tibe
tan
and
Rgya
lrong
cogn
ates
act
ually
mea
n ‘d
hole
(Cyo
n al
pinu
s)’.
30.
A re
latio
n w
ith th
e fo
rms c
ited
in M
atiso
ff (2
003:
317)
is p
ossib
le b
ut re
quire
s fur
ther
rese
arch
.