^{page 531 note 1} The possibility that there may be languages on the border-line between Miao and Yao cannot be ruled out. More linguistic and ethnographic documentation is necessary before definitive characterization will be feasible. Southern Chinese use of the word ‘Miao’ to cover various hill-tribes adds to the difficulty—the so-called Miao of Hainan island appear to be Yao both linguistically and culturally.

Although the She tribes in the mountainous areas near the south-east coast of China are known to be culturally akin to the Yao, nothing has yet been published about their language.

^{page 531 note 2} For tribal distribution in China, see the Karta narodov Kitaya, MNR, i Korei, Moskva, 1959Google Scholar.

^{page 531 note 3} The material was gathered during 1959 and 1960, chiefly in the villages of P'u K'amteng and Chungliang, about four and seven hours from Muongluang Namtha on the Muongsai trail, in north-west Laos. My chief informant was Tsĕw Iw-táq , a young headman from the area, and the language described is essentially his. Highland Yao, however, appears to be remarkably uniform, and few phonological differences from his speech were noted in the area. Highland Yao seems to be almost identical with P'anku Yao of Kwangsi province, of which some words are noted in and , p. 26 ff. (CYYY Monograph, No. 2, 1929).

Except for a few short word-lists, the only published works on Yao languages hitherto are:

Savina, F., ‘Dictionnaire français-mán précédé d'une note sur les Man Kim-cli-mun et leur langue’, BEFEO, xxvi, 1926, 11–255Google Scholar.

Wong, S. L.‘Phonetics and phonology of the Yao language’, Ling-nan Science Journal, xviii, 4, 1939, 425 ffGoogle Scholar.

Chao, Y. R., Phonetics of the Yao folk-songs (CYYY Monograph, A, No. 1), Peiping, 1930Google Scholar.

Haudricourt, A. G., ‘Introduction à la phonologie historique des langues miao-yao’, BEFEO, XLIV, 2, 1947–1950, [pub. 1954], 555–76Google Scholar. This also contains a bibliography of manuscript material on the Yao dialects. Highland Yao is substantially equivalent to his ‘Mien’.

^{page 532 note 1} It occurs also in words such as kyan-tsâq ‘robber’. A special term is used for the Valley Yao—tshan-tsăy.

^{page 532 note 2} For descriptive convenience, the treatment of the finals precedes that of the nucleus. Throughout this article italic type is used for syllable components, bold-face type for IPA transcriptions of realizations when it is necessary to specify these. Where this is unnecessary, the syllable components are to be understood as having their corresponding IPA values.

^{page 533 note 1} From the synchronic, monolingual point of view, there is no reason for the order of tones given (except for the difference between continuant and stop finals); this order was first used in Chang K'un's article ‘On the tonal system of the Miao and Yao languages’ CYYY, xvi (Shanghai), 1947, 93Google Scholar, and is convenient for comparative purposes. There is, however, a phonological distinction between syllables with even-number tones and those with odd-number tones (see below, pp. 533–5). Note that although the syllables with final stops have the same tone marks as those with tone 5 and 6, this is merely for printing convenience. No identification is suggested.

^{page 533 note 2} See below, p. 538.

^{page 534 note 1} The following exceptions were noted; except for the first word, all are of Chinese origin. kεŋ-khyòŋ ‘mosquito’

phá:ŋ ‘raised sleeping-platform for guests’(Cantonese. pha:ŋ ‘platform, scaffolding’)

tsiəp-hm ‘fifteen’ (Cant. -sap,ŋ ‘fifteen’)

^{page 534 note 2} Words such as ùy ‘to surround’are also exceptions; these, however, do not seem to contrast with ^*-wi (as noted infra, p. 537, n. 2), and might be so analysed; in which case they follow the regular rule.

^{page 535 note 1} See p. 533.

^{page 536 note 1} There seems to be little point in associating such syllables with syllables having one of the vowel nuclei. Chao, Y. R. (Cantonese primer, 22Google Scholar) takes the parallel syllables in Cantonese, m ‘not’ and, ŋ ‘five’as mu and ŋu respectively, for which there is some historical justification, but no advantage in ease of description, since the vowel -u- does not otherwise occur after labial or velar initials. Moreover, such syllables are a common feature of the languages in this area; cf. Moiyan Hakka –m and ŋ (same words) which contrast with e.g. '‘mu ‘dream’ and ŋwu (in ̄ŋu-'fi ‘misunderstanding’), and for which, therefore, no such analysis is possible. It seems best to take the occurrence of such syllables as a recurrent feature of the area, which stands outside the normal system of vowels. (In HY these syllables contrast with mú ‘fifth term in the cycle often’).

^{page 537 note 1} Probably a borrowing from a Chinese dialect, cf. Fuehou ̄phiū (k) eiŋ) ‘ladle, spoon’.

^{page 537 note 2} -iw seems to be distinguished from -yu only after velars; e.g. gǏw ‘scissorst’ (giu) as against kyŭ ‘dog’(CU). After other initials, only syllables with the rhyme -iw occur. A parallel laok of distinction is found with the rhyme -uy, which never seems to contrast with a hypothetical *-wi.

^{page 538 note 1} Note that there is no trace of such lip-rounding with initial O-; ŭə ‘that, there’ is simply uə, with no W-like semivowel.

^{page 538 note 2} See p. 535.

^{page 539 note 1} e.g., in words of Chinese origin, Cant, –sok = HY tsûəq ‘ripe’, but Cant. ̄tsoŋ = HY tswáŋ ‘to plant’; Cant, – sap– lok = HY tsiəp- lûəq ‘sixteen’, but Cant, ' kwəŋ = HY kywâ:ŋy ‘together’; also, Cant, - pək = HY pîəq ‘thin’, but Cant. 'kwəŋ = HY gyā:ŋ ‘bright’; whereas the relationship is regular with other finals: Cant, –lap = HY lîəp ‘to set up’ just as Cant. lam = HY liəm ‘to drench’.

^{page 540 note 1} For example, for hnəy-hnāy ‘every day’, I recorded at various times. I had the strong impression that constantly recurring collocations occurred very often with extremely reduced forms, whereas more occasional collocations usually approxi-mated to the corresponding full forms except for tone; in other words, some disyllables were more ‘institutionalized’ than others. More research is needed on this point, however. In this connexion it is worth noting that in reduced syllables for which there exists no corresponding full form recognized by the speakers, a high proportion have the nuclei -a-, -o-, -u-, i.e., those resembling the lax vowels listed above. For example, fan-dày ‘potato’, yoŋ-īn ‘opium’, nom-tsīw ‘banana’; and the common prefix tom- ‘great’ as in tom-kīdn ‘high official’, tom-tsāŋ ‘high mountain’(the free word for ‘great, big’ is hlō). It may be conjectured that some of these are derived from the fast forms of reduced syllables with an originally wider range of nucleus.