Annotated Swadesh wordlists for the Hmong group (Hmong-Mien family).

Languages included: Western Xiangxi Hmong [hmo-xxi]; Eastern Xiangxi Hmong [hmo-xxe]; Qiandong Hmong [hmo-qia]; Chuanqiandian Hmong [hmo-cqd]; Diandongbei Hmong [hmo-hmd]; Hmong Daw [hmo-daw]; Hmong Njua [hmo-nju]; Bunu (proper) [hmo-bnu]; Baonao Bunu [hmo-bna]; Numao Bunu [hmo-nmu]; Longhua Jiongnai [hmo-lhj]; Liuxiang Jiongnai [hmo-lxj]; Xiaozhai Younuo [hmo-ynz]; Huangluo Younuo [hmo-ynl]; Northern Pa-Hng [hmo-phn]; Southern Pa-Hng [hmo-phs]; Hm-Nai [hmo-wun]; Lianhua She [hmo-lia]; Luofu She [hmo-luo]; Pa Na [hmo-pna].

DATA SOURCES

General.

MYYFY 1987 = 苗瑤語方言詞匯集 [Miao-Yao yu fangyan cihui ji] [A collection of vocabularies of Miao-Yao dialects]. Beijing: Zhongyang minzu xueyuan chubanshe. // Comparative vocabularies for seven varieties of Hmong-Mien spoken in the People's Republic of China, including Xiangxi, Qiandong, Chuanqiandian, and Diandongbei Miao. A significant amount of data, accompanied, however, only by brief listings of the corresponding dialects' phonemic inventories.

I. Xiangxi Hmong (Western, Eastern).

Xiang 1992 = 向日征. 漢苗詞典 (湘西方言) [Xiang Rizheng. Han-Miao cidian (Xiangxi fangyan)] [Xiang Rizheng (ed.). Chinese-Miao vocabulary (Xiangxi dialect)]. Chengdu: Sichuan minzu chubanshe. // A large dictionary of the Xiangxi dialect, accompanied by a brief sketch of Xiangxi phonology.

Yang 2004 = 楊再彪. 苗語東部方言土語比較 [Yang Zaibiao. Miaoyu dongbu fangyan tuyu bijiao] [Yang Zaibia. A comparison of the Eastern dialects of the Miao language]. Beijing: Minzu chubanshe chuban faxing. // A comparative description of several dialects of Western and Eastern Xiangxi Hmong, including a large comparative vocabulary for three Western and three Eastern Xiangxi locations.

II. Qiandong Hmong.

Zhang 1990 = 張永祥. 苗漢詞典 (黔東方言) / Hmub diel cif dieex (hveb qeef dongb). [Zhang Yongxiang. Miao-Han cidian (Qiandong fangyan)] [Zhang Yongxiang (ed.). Miao-Chinese vocabulary (Qiandong dialect)]. Guizhou minzu chubanshe. // A large dictionary of the Northern Qiandong dialect, accompanied by a brief sketch of Qiandong phonology.

III. Hmong Daw.

Heimbach 1979 = Heimbach, Ernest E. White Hmong - English Dictionary. Ithaca, New York: Cornell University. // Large dictionary of the Hmong Daw (White Hmong) language, based primarily on the Northern Thailand dialect.

Mottin 1978 = Mottin, Jean. Elements de grammaire Hmong Blanc. Don Bosco Press. // A detailed grammatical description of the Hmong Daw language.

IV. Hmong Njua.

Lyman 1974 = Thomas Amis Lyman. Dictionary of Mong Njua, a Miao (Meo) Language of Southeast Asia. The Hague: Mouton. // A large dictionary of the Hmong Njua language, illustrated by numerous textual examples and short notes on phonology and some grammar aspects.

V. Bunu (proper; Baonao; Numao).

Meng 2001 = 蒙朝吉. 瑤族布努語方言研究 [Meng Chaoji. Yao zu bunu yu fangyan yanjiu] [Meng Chaoji. Research on the dialects of the Bunu language of the Yao nationality]. Beijing: Minzu chubanshe. // Detailed analysis of the phonology and word structure of several closely related dialects of the Bunu cluster.

VI. Jiongnai (Longhua; Liuxiang).

Mao & Li 2002 = 毛宗武, 李云兵. 炯奈語研究 [Mao Zongwu, Li Yunbing. Jiongnai yu yanjiu] [Mao Zongwu, Li Yunbing. Research on the Jiongnai language]. Beijing: Zhongyang minzu daxue chubanshe. // Detailed analysis of the phonology, lexicon, and grammar of the Jiongnai dialect cluster.

VII. Younuo (Xiaozhai; Huangluo).

Mao & Li 2007 = 毛宗武, 李云兵. 蚥霾語研究 [Mao Zongwu, Li Yunbing. Younuo yu yanjiu] [Mao Zongwu, Li Yunbing. Research on the Younuo language]. Beijing: Minzu chubanshe. // Detailed analysis of the phonology, lexicon, and grammar of the Younuo dialect cluster.

VIII. Pa-Hng (Northern; Southern); Hm-Nai.

Mao & Li 1997 = 毛宗武, 李云兵. 巴哼語研究 [Mao Zongwu, Li Yunbing. Baheng yu yanjiu] [Mao Zongwu, Li Yunbing. Research on the Pa-Hng language]. Beijing: Minzu chubanshe. // Detailed analysis of the phonology, lexicon, and grammar of the Bahengic languages, including Northern Pa-Hng, Southern Pa-Hng, and Hm-Nai.

IX. She (Lianhua; Luofu).

Mao & Meng 1986 = 毛宗武, 蒙朝吉. 畲語簡志 [Mao Zongwu, Meng Chaoji. She yu jianzhi] [Mao Zongwu, Meng Chaoji. A brief description of the She language]. Beijing: Minzu chubanshe. // Detailed analysis of the phonology, lexicon, and grammar of the She language, concentrating on the Lianhua (Western) and Luofu (Eastern) dialects.

Hiroki 2003 = Nakanishi Hiroki. A She Vocabulary. Haifeng Dialect. Kyoto University: Institute for Research in Humanities. // A large classified Chinese-language vocabulary of the Haifeng dialect of She, collected by the author. This dialect is lexically closer to Lianhua She than to Luofu She.

X. Pa-Na.

Taguchi 2001 = 田口善久. 巴那語詞匯集. [Taguchi, Yoshihisa. A vocabulary of Pana]. In: Tasaku Tsunoda (ed.). Basic Materials in Minority Languages 2001 (ELPR Publication Series 2001). // A 200-item wordlist for Pana, personally collected by the author.

Chen 2001 = 陳其光. 巴那語概況. [Chen Qiguang. Banayu gaikuang] [Chen Qiguang. A general description of the Pana language]. Minzu yuwen 2001(2): 69-81. // A brief description of the phonology and grammar of Pa-Na, with a representative accompanying vocabulary.

NOTES

I. Xiangxi Hmong.

General.

Xiangxi Hmong may be considered a "macro-language", with close to a million speakers, most of which are assumed to be speakers of its "Western" variety. The precise number of individual lects that could be thought of as individual languages is not easy to ascertain: based on [Yang 2004], so far the most detailed account on the dialectal variety of this cluster, even lects spoken in geographically closed locations may have up to 10% lexicostatistical discrepancies between them (although much of this has to do with stronger/weaker influence of colloquial and literary Chinese on specific dialects).

Although in practice, due to the availability of [Yang 2004] it is technically possible to compile separate wordlists for as many as six different varieties (Jiwei, Yangmeng, Zhongxin dialects of Western Xiangxi; Danqing, Xiaozhang, Dengshang dialects of Eastern Xiangxi), we have not used this option, because it is not clear to what extent the actual vocabularies in [Yang 2004] are semantically accurate, and we would not want to confuse potential semantic inaccuracies with real historical distinctions. Therefore, for the sake of extra safety we have adopted the following strategy:

(a) For Western Xiangxi, our default source is [Xiang 1992], a special detailed dictionary of just one dialectal variety, with the data in [MYYFY 1987] and in [Yang 2004] used as control sources.

(b) For Eastern Xiangxi, where the only usable source is [Yang 2004], we have selected the Danqing dialect as the one that seems the least infiltrated with Chinese borrowings (although that does not mean that it is entirely free of them, either). Data from the other two dialects (Xiaozhang, Dengshang) are given in the notes section.

Transliteration.

Although the Xiangxi vocabularies in [Xiang 1992] and [MYYFY 1987] are almost identical from a lexical standpoint (with just a few minor discrepancies in the choice of desemanticized classifiers, etc.), the IPA-based transcription in [MYYFY] and Xiang Rizheng's alphabetic notation for Xiangxi sometimes reflect different phonetics, probably due to their describing different subdialects. In particular:

(1) Xiang's initial clusters of nasals + unaspirated / aspirated stops (nd, np, etc.) are consistently transcribed as voiced unaspirated / aspirated consonants (d, bʰ, etc.) in [MYYFY];

(2) Those finals in -n which, according to Xiang's conversion tables, represent nasalized vowels (e. g. Xiang's en = our ɛ̃), are consistently transcribed in [MYYFY] without nasalization (i. e. simply as ɛ, etc.).

Transcription in [Yang 2004] is IPA-based and largely the same as in {MYYFY].

The basic conventions for transliteration of [Xiang 1992] are as follows:

[Xiang 1992]	UTS transliteration	Notes
b	p
bl	pɹ
c	cʰ
d	t
g	k	The sequence ki̯- is transcribed as ɕ- [IPA c-] in MYYFY, reflecting palatalization. In fact, k- before front vowels in [Xiang 1992] seems to generally correspond to c- in [MYYFY].
gh	q
h	x
hl	l̥	The sequence l̥i̯- is transcribed as ʎ̥ʰ- in MYYFY.
hm	m̥
hn	n̥
j	ɕ
kh	qʰ
l	l	The sequence li̯- is transcribed as ʎ- in MYYFY.
m	m
ml	mɹ
n	n	The sequence ni̯- is transcribed as ɲ- in MYYFY.
nb	mp	Transcribed as b in MYYFY.
nd	nt	Transcribed as d in MYYFY.
ngh	nq	Transcribed as ɢ in MYYFY.
nh	ɲ
nj	ɲɕ	Transcribed as ʓ (dʑ) in MYYFY.
np	mpʰ	Transcribed as bʰ (bh) in MYYFY.
nq	ɲɕʰ	Transcribed as ʓʰ (dʑh) in MYYFY.
p	pʰ
q	ɕʰ
r	ʐ
s	s
sh	ʂ
w	w
x	ʆ
y	ʑ
z	c
zh	ʈ

a	a
an	ɛ̃	Transcribed as ɛ in MYYFY.
ang	aŋ
ao	ɔ
e	e
ea	ä	Transcribed as a in MYYFY.
ei	ei
en	en
eu	ɤ
i	i
ia	i̯a
ian	i̯ɛ̃	Transcribed as iɛ in MYYFY.
iao	i̯ɔ
iea	i̯ä
in	in
io	i̯o
o	o	Transcribed as ɔ in MYYFY.
ong	oŋ
ou	ɯ
u	u
ue	u̯e
uou	u̯ɯ

-b	1	Tone 1 is described as mid-rising (35).
-x	2	Tone 2 is described as mid-falling (42).
-d	3	Tone 3 is described as high level (44).
-l	4	Tone 4 is described as low level (22).
-t	5	Tone 5 is described as high-to-mid-falling (54).
-s	6	Tone 6 is described as mid-to-low-falling (31).

II. Qiandong Hmong.

Transliteration.

There are very few transcriptional differences between the orthographic system used in [Zhang 1989] and the IPA-based transcription in [MYYFY 1987]. Both sources represent the same dialect of Qiandong Hmong: Northern Qiandong, spoken in the Yanghao village and other locations.

[Zhang 1989]	UTS transliteration	Notes
b	p
c	cʰ
d	t
dl	ɬ
f	f
g	k
gh	q
h	h
hf	fʰ
hl	l̥ʰ
hm	m̥ʰ
hn	n̥ʰ
hs	sʰ
hv	xʰ
hx	ʃʰ
j	ɕ
k	kʰ
kh	qʰ
l	l
m	m
n	n
ng	ŋ
p	pʰ
q	ɕʰ
r	z
s	s
t	tʰ
v	ɣ
w	v
x	ʃ
y	ʑ
z	c

a	a
ai	ɛ
ang	aŋ
ao	ao
e	ǝu	Transcribed simply as ǝ in [MYYFY 1987].
ee	e
ei	ei
en	en
i	i
ia	ia
iang	iaŋ
iao	iao
ie	iǝ
iee	ie
in	in
io	io
iong	ioŋ
iu	iu
o	o
ong	oŋ
u	u
ua	ua
uai	uɛ
uang	uaŋ
ui	ui
un	un

-b	1	Tone 1 is described as mid-level (33).
-x	2	Tone 2 is described as ultrahigh-level (55).
-d	3	Tone 3 is described as mid-rising (35).
-l	4	Tone 4 is described as low-level (11).
-t	5	Tone 5 is described as high-level (44).
-s	6	Tone 6 is described as low-rising (13).
-k	7	Tone 7 is described as high-falling (53).
-f	8	Tone 8 is described as mid-falling (31).

III. Chuanqiandian Hmong, Diandongbei Hmong.

General.

There are two varieties of Hmong speech, covered in [MYYFY 1987], that both belong to the so-called "Chuanqiandian cluster": "Chuanqiandian proper" is spoken in Xùyǒng (敘永), in the southeast of Sichuan, and "Diandongbei Hmong", also known as "Large Flowery Miao" (dà-huā-miáo 大花苗), is spoken in Shíménkǎn (石門坎), in west Guizhou. Since these two dialects are quite close, and data for them comes from the same unified source, we discuss them together.

Transliteration.

The phonetic systems of Chuanqiandian and Diandongbei are relatively close (although the former is significantly more rich), and their phonetic representation in [MYYFY 1987], based on the IPA, has only minor differences from the UTS transliteration. Some general notes:

1. Chuanqiandian proper has phonologically relevant voiced (b, d, ʒ, etc.) and prenasalised voiced (mb, nd, nʒ, etc.) initials that are missing in Diandongbei.

2. Chuanqiandian proper has initial consonantal clusters of labial and alveolar stops with -l- (pl-, tl-), whereas Diandongbei generally transforms both of these series into lateral affricates (ƛ-), etc. In terms of transcription, therefore, [MYYFY 1987] tl = UTS tl, but [MYYFY 1987] tɬ = UTS ƛ, tɬh = UTS ƛʰ.

3. Both dialects, like most Hmong varieties, have 8 tones, but their phonetic realization varies. According to [MYYFY 1987], the system is as follows:

Tone number	Phonetic realisation in Chuanqiandian	Phonetic realisation in Diandongbei
1	High-falling (43)	High-falling (53)
2	Mid-falling (31)	Mid-rising (45)
3	High level (55)	High level (55)
4	Low-falling (21)	Low level (11)
5	Low level (44)	Mid level (22)
6	Low-rising (13)	Mid-falling (31)
7	Mid level (33)	Low level (11) [merged with Tone 4]
8	Mid-rising (24)	Low-falling (21)

IV. Hmong Daw.

Transliteration.

The transliteration of E. Heimbach's transcriptional system into UTS is based on his own phonetic commentary, which is sometimes approximate, and may not be 100% adequate. However, all the phonological oppositions (consonantal, vocalic, and tonal) have been carefully preserved.

[Heimbach 1979]	UTS transliteration	Notes
c	ɕ
ch	ɕʰ
d	ʼd	Preglottalized articulation.
dh	ʼtʰ
f	f
g	ŋ
h	h
hl	l̥
hm	m̥
hn	n̥
hnl	m̥l
hny	ɲ̥
k	k
kh	kʰ
l	l
m	m
n	n
nc	ɲʓ
nch	ɲɕʰ
nk	ŋg
nkh	ŋkʰ
nl	ml
np	mb
nph	mpʰ
npl	mbl
nplh	mbl̥
nq	nq
nqh	nqʰ
nr	ɳɖ
nrh	ɳʈʰ
nt	nd
nth	ntʰ
nts	nǯ
ntsh	nčʰ
ntx	nʒ
ntxh	ncʰ
ny	ɲ
p	p
ph	pʰ
pl	pl
q	q
qh	qʰ
r	ʈ
rh	ʈʰ
s	š
t	t
th	tʰ
ts	č
tsh	čʰ
tx	c
txh	cʰ
v	v
x	s
xy	ʆ
y	y
z	ž

a	a
ai	ay
au	ɔw
aw	ǝw
e	e
ee	ɤ̃
i	i
ia	i̯a
o	ɔ
on	ɔ̃
oo	õ
u	u
ua	u̯ǝ
w	ɨ
ww	ɨ̃
x	s
xy	ʆ

-b	1	Tone 1 is described as high-level (55).
-j	2	Tone 2 is described as high-falling (53).
-v	3	Tone 3 is described as mid-rising (34).
-0 (unmarked)	4	Tone 4 is described as mid-level (33).
-s	5	Tone 5 is described as low-level (11).
-g	6	Tone 6 is described as low-falling (21).
-m	7	Tone 7 is described as low-level checked (11).
-d	8	Tone 8 is described as low-rising (13).

Notes:

1. Word-initial vowels are automatically preceded by a glottal stop; it is not marked in Heimbach's transcriptions and is not reproduced in our transliteration either.

2. Tone 8 (Heimbach's -d) is described as a rare variant of tone 7 (Heimbach's -m) that "appears only on a certain class of words". We still mark it separately in the only word on the Swadesh list where it appears ('that').

3. It should be noted that, since the system of word-initial consonants and clusters in Hmong Daw is so complex, many of these elements do not appear in the database at all. Nevertheless, we offer a complete transliteration system for a better understanding of the phonological oppositions in this language.

V. Hmong Njua.

Transliteration.

Like most early researchers of Hmong languages, Th. Lyman uses a rather idiosyncratic transcriptional notation in his dictionary. For simplicity's sake, we list his entire system as transcoded to UTS.

[Lyman 1974]	UTS transliteration	Notes
ʔ	ʔ	The glottal stop is automatic before an initial vowel; it also sometimes replaces velar and uvular consonants in the specific "children's speech" variants of certain words.
c	č	Postalveolar affricate.
ch	čh	Postalveolar affricate.
č		Retroflex affricate.
čh	ʰ	Retroflex affricate.
ȼ	c	Alveolar affricate.
ȼh	cʰ	Alveolar affricate.
f	f
h	h
hl	ɬ
hs	sʰ
hy	ʆ
k	k
kh	kʰ
l	l
m	m
mb	mb	-b- in this nasal cluster is an allophone of -p-.
mph	mpʰ
kh	kʰ
n	n
ñ	ɲ
ñch	ɲčʰ
ñd̨	ɲʓ	-ʓ- in this nasal cluster is an allophone of -ɕ-.
ñj	ɲǯ	-ǯ- in this nasal cluster is an allophone of -č-.
ɲƫh	ɲɕʰ
ň	ɳ
ňčh	ɳʰ
ňǰ	ɳᶚ	-ᶚ- in this nasal cluster is an allophone of --.
ŋ	ŋ
ŋg	ŋg	-g- in this nasal cluster is an allophone of -k-.
ŋɣ	ŋɢ	-ɢ- in this nasal cluster is an allophone of -q-.
ŋkh	ŋkʰ
ŋqh	ŋqʰ
p	p
ph	pʰ
q	q
q	qʰ
s	s
š	š
t	t
th	tʰ
ƫ	ɕ	Palatal affricate.
ƫh	ɕʰ	Palatal affricate.
v	v
y	y
ž	ž
a	a
ai	ay
aŋ	aŋ
ao	ao
au	au
aw	aʉ
e	e
eŋ	eŋ
i	i
ɨ	ɨ
oŋ	oŋ
u	u
ua	u̯a
w	ʉ
wŋ	ʉŋ

V́	1	Tone 1 is described as high-level (55 or 54).
V̌	2	Tone 2 is described as mid-rising (35).
V	3	Tone 3 is described as mid-level (33, long).
V̀	4	Tone 4 is described as low-level (22).
V̂	5	Tone 5 is described as high-falling (41).
V̊	6	Tone 6 is described as low-level (11) + breathy.
V+	7	Tone 7 is described as low-rising (214), with additional length and glottal stop.

VI. Bunu.

Transliteration.

The transcription system used in [Meng 2001] to transcribe the three dialectal varieties of Bunu is generally based on the IPA, and consequently requires only cosmetic change when transliterated to UTS. Below we give an exhaustive list of such changes, without printing out the entire inventory of Bunu initials or finals.

[Meng 2001]	UTS transliteration	Notes
ph, th...	pʰ, tʰ...	Aspirated initials.
nts	nc
tɕ, tɕh, ɕ	ɕ, ɕʰ, ʆ	Palatal affricates and fricatives.
tl, tlh, ntl	ƛ, ƛʰ, nƛ	These initials are also transcribed in [Meng 2001] as tɬ, tɬh, ntɬ for the Bunu proper dialect. However, the difference between tɬ and tl seems to be purely graphical, without any underlying phonetic reality.
pj, kj...	pi̯, ki̯...	We interpret the "softened" (palatalized) initials of Bunu as combinations with the medial glide -i̯- and transcribe them accordingly.
ai, ei	ay, ey

Tones. All Bunu dialects have an 8-tonal system; additionally, some of the tones (1-4 in Bunu proper; 1-3, 6, 8 in Numao; none in Baonao) undergo further binary splits under certain conditions, usually through active or "fossilized" tonal samdhi (however, very few examples of this tonal split are encountered in the Swadesh list). The precise phonetic realisations, according to [Meng 2001], are as follows:

Tone	Bunu proper	Baonao	Numao
1	33	42	33
2	13	24	55
3	43	33	24
4	232	44	43
5	41	32	44
6	221	31	31
7	32	32	55
8	21	54	32
1a	55		422
2a	35		424
3a	54		53
4a	454
6a			423
8a			52

The "split" variants of the tones (1a, 2a, etc.) are marked as 1 ́, 2 ́, etc. in [Meng 2001].

VII. Jiongnai.

Transliteration.

The transcription used in [Mao & Li 2002] to render the forms in the Longhua and Liuxiang dialects of the Jiongnai language is generally consistent with the IPA, and requires only the usual IPA > UTS transliteration elements (the same as outlined above for Bunu).

It must be mentioned that Mao & Li mark the Jiongnai tones in accordance with their register/contour characteristics rather than with a general abstract scheme from 1 to 8. The standard eight tones of Jiongnai are marked as follows: (1) = 44; (2) = 33 (Longhua), 31 (Liuxiang); (3) = 53 (Longhua), 35 (Liuxiang); (4) = 31 (Longhua), 53 (Liuxiang); (5) = 35 (Longhua), 22 (Liuxiang); (6) = 22 (Longhua), 21 (Liuxiang); (7) = 43; (8) = 32.

VIII. Younuo.

Transliteration.

The transcription used in [Mao & Li 2007] to render the forms in the Xiaozhai and Huangluo dialects of the Younuo language is generally consistent with the IPA, and requires only the usual IPA > UTS transliteration elements (the same as outlined above for Bunu and Jiongnai).

As in the case of Jiongnai, tones in [Mao & Li 2007] are marked in accordance with their register/contour characteristics rather than with any general abstract scheme from 1 to 8. The eight-tonal scheme is reduced to six tones in the Xiaozhai dialect and to five tones in the Huangluo dialect. The correlations between the "ideal" eight tones and the actual phonetic values in both dialects are as follows: (1) = 33; (2) = 13; (3) = 22; (4) = 22; (5) = 35 (Xiaozhai), 53 (Huangluo); (6) = 31 (Xiaozhai), 22 (Huangluo); (7) = 53; (8) = 31.

IX. Pa-Hng and Wunai.

1. General.

Data on the closely related Pa-Hng (= Bāhēng) and Hm-Nai (= Wúnài) languages are drawn here from the most detailed and up-to-date source on both of them, [Mao & Li 1997]. According to this source, Pa-Hng shows a primary division into Northern and Southern dialects, represented in Mao & Li's description by the lects spoken respectively in Gundong (滚董) (Liping County, Guizhou) and Wenjie (文界) (Sanjiang County, Guangxi). Hm-Nai is represented in the monograph with the lect spoken in Huxingshan (虎形山), Longhui County, Hunan. Although Wunai is usually described as a separate language, the data in Mao & Li 1997 clearly show that it is very closely related to Pa-Hng.

2. Transliteration.

The transcription used in [Mao & Li 1997] is generally consistent with the IPA, and requires only the usual IPA > UTS transliteration elements (the same as outlined above for Bunu, Jiongnai, and Younuo).

It should be mentioned that Mao & Li treat the "voiced" simple and prenasalized initials of Pa-Hng as combinations of a voiceless initial with voiced aspiration (pɦ, tɦ, mpɦ, ntɦ, etc.). We preserve their transcription in this matter.

Tones in [Mao & Li 1997] are marked in accordance with their register/contour characteristics rather than with any general abstract scheme from 1 to 8. The eight-tonal scheme is reduced to seven tones in Northern Pa-Hng, six tones in Southern Pa-Hng, and seven in Wunai; however, Proto-Bahengic definitely had a complete eight-tone system, as is evident from the correspondences.

The correlations between the "ideal" eight tones and the actual phonetic values in both dialects are as follows: (1) = 35; (2) = 33; (3) = 22 (Northern Pa-Hng), 31 (Southern Pa-Hng), 13 (Hm-Nai); (4) = 22 (Northern Pa-Hng, Wunai), 31 (Southern Pa-Hng); (5) = 55; (6) = 44 (Pa-Hng), 31 (Wunai); (7) = 53; (8) = 31.

X. She.

1. General.

Our main source on the She language, once thought of as possibly forming its own (third) branch of the Hmong-Mien family, but now definitively established as a divergent branch of the Hmong group, is [Mao & Meng 1986], a stereotypical Chinese language description of the phonology and grammar of She, accompanied with an extensive vocabulary for two idiolects: Lianhua (蓮花) and Luofu (羅浮), respectively representing the Western and Eastern dialects of the language, although lexical discrepancies between the two varieties are absolutely minimal anyway.

As a control source, we also use [Hiroki 2003], a large classified dictionary of the Haifeng (海豐) variety of the She language that seems most close to Lianhua (and is therefore quoted in the notes section on the Lianhua dialect), despite a few important phonetic discrepancies (for instance, Lianhua n regularly corresponds to Haifeng l).

2. Transliteration.

The transcription used in [Mao & Meng 1986] is generally consistent with the IPA, and requires only the usual IPA > UTS transliteration elements (the same as outlined above for Bunu, Jiongnai, Younuo, and Pa-Hng).

Tones are marked in [Mao & Meng 1986] according to the schematic 1-8 notation. Actual phonetic values are as follows: (1) = 22; (2) = 53; (3) = 33; (4) = 42; (5) = 31; (6) = 35; (7) = 21; (8) = 54. In [Hiroki 2003], register/contour characteristics are chosen as the basis for the numeric index instead.

XI. Pa Na.

1. General.

Pa Na, a language spoken by 1,000 to 3,000 native speakers in Southwest Hunan, is closely related to She, but has not been as thoroughly described due to the significantly smaller number of speakers. Our main source for the original wordlist has been Taguchi 2001; unfortunately, although the author uses a diagnostic 200-item wordlist, it is not based on Swadesh and, therefore, contains a large number of gaps. These had to be filled in with data taken from the brief description of the language given in Chen 2001; fortunately, despite some transcriptional differences and occasional morphological discrepancies (e.g. Chen more often gives nominal roots in conjunction with desemanticized prefixes, while Taguchi often sticks to "bare" roots), in most situations the data agree with each other.

2. Transliteration.

[Taguchi 2001] uses a simplified transcription for the language, much of which has to be reconverted. [Chen 2001] is closer to IPA, but still uses certain idiosyncrasies of his own.

(a) aspirated consonants: Taguchi's, ph, th, etc. = Chen's ph, th, etc. = UTS pʰ, tʰ, etc.;

(b) coronal affricates: Taguchi's ts, tsh, dz = Chen's ts, tsh, dz = UTS c, cʰ, ʒ;

(c) lateral affricates: Taguchi's tl, tlh, dl = Chen's tɬ, tɬh, dl = UTS ƛ, ƛʰ, Ł;

(d) Taguchi's ng = Chen's ŋ = UTS ŋ;

(e) Taguchi's q = Chen's ʔ = UTS ʔ;

(f) Taguchi's nasalized vowels i~, u~, o~ = UTS ĩ, ũ, õ.

Additionally, Chen postulates a separate series of palatal initials for the language, which usually correspond to Taguchi's combinations of coronal initials with a medial glide: Chen's tɕ, tɕh, dʑ, ʑ, ɕ = UTS ɕ, ɕʰ, ʓ, ʑ, ʃ (= Taguchi's tsj, tshj, dzj, zj, sj).

Tones are marked in [Taguchi 2001] according to the schematic 1-8 notation. Actual phonetic values are as follows:

(1) = 24 (Taguchi), 13 (Chen); (2) = 12 (Taguchi), 313 (Chen); (3) = 44; (4) = 21 (Taguchi), 31 (Chen); (5) = 35; (6) = 22; (7) = 45 (Taguchi), 55 (Chen); (8) = 42 (Taguchi), 53 (Chen).

Database compiled and annotated by: G. Starostin (Last update: June 2017).