In his phonemic analysis of Lhasa Tibetan in Love songs of the sixth Dalai Lama Jaw Yuanrenn (Y. R. Chao) notices ‘variations’ of some of the phonemes, and ascribes some of these ‘variations’ to differences in tempo or to chance: of the phoneme that he writes as ‘e’ he says ‘ i, 10 3 khrel gzhung tş'ilcuῃ tş'ilcuŋ’ [it may sometimes change to l, e.g. in song 10, line 3, khrel gzhung tş'cuῃ is pronounced tş'ilcuῃ]; of the phoneme ‘a’, ‘ә 1 2 lam buhi lamp'ø lәmpø’ [it may change to ә in rapid speech, e.g. in song 4, line 2, lam buhi lamp'ø is pronounced lәmpø]; and of the phoneme ‘o’, ‘u, 36 3 rlung po luŋpo luŋbu’ [it may occasionally be pronounced u, e.g. in song 36, line 3, rlung po luŋpo is pronounced lunbu]; but a phonological analysis of the speech of Rinzin Wangpo (rig-'dzin dbang-po) (R.), a Lhasa-dialect-speaking Tibetan, overwhelmingly suggests that vowel alternations such as these should be attributed to vowel harmony.
1 This article is based on a paper, ‘Vowel harmony in Lhasa Tibetan’, read at the twenty-fifth International Congress of Orientalists, Moscow, in August 1960. Vowel harmony was chosen as the subject of this paper as a compliment to the late Professor Georges de Roerich (Ю). H. Pepиx), of the Institute of Peoples of Asia of the Academy of Sciences of the U.S.S.R., with whom its author had discussed this aspect of Tibetan phonology some ten years ago in India.
In the phonological analysis presented in this article, following F. R. Palmer, ‘“Openness”: in Tigre: a problem in prosodic statement’, BSOAS, XVIII, 3, 1956, 561, vowel and consonant, and their derivatives (vocalic, consonantal), are used as purely phonetic terms; for phonological units of Syllable structure V and C are used. A particular term in a V system, e.g. A, may have one or more vowels stated as its phonetic exponents. With vowel used in this phonetic sense the term vowel harmony refers to a relationship not of structural units but of vocalic, and therefore phonetic, features; but Lhasa Tibetan is also a vowel-harmony language in the traditional use of this term: ‘the peculiar restriction… of tolerating only certain combinations of vowels in successive syllables’ (L. Bloomfield, Language, London, 1950, 181). There is thus a twofold sense in which the title of this article is appropriate to the material presented in it.