Wylie Transliteration

A STANDARD SYSTEM OF TIBETAN TRANSCRIPTION

Turrell Wylie

University of Washington

[published in the Harvard Journal of Asiatic Studies, vol. 22 [Dec., 1959], pp. 261-267 ]

There is a Tibetan proverb which says:

“Every district it's own dialect;

Every Lama his own doctrine.”¹

to which might be added:

Every scholar his own transcription.

Perhaps no other academic endeavor evidences the spirit of scholarly independence more readily than the transcription of the Tibetan language into Roman script. A survey of publications by a dozen Tibetan scholars selected at random reveals a dozen varying systems of transcription — a profusion of apostrophes, diacritical marks, Greek gammas, capital and italic letters (see Chart I).

Some of the of the transcription practices are as intriguing as they are inexplicable. For example, in an article by Csoma de Körös,² the nasal velar ng is transcribed in two different ways. In initial position it appears as an n with a tilde, the diacritical mark for a palatalized nasal, not a nasal velar; but, in a final position, it is transcribed as ng. In the same article the velar fricative h and the a-chung are both transcribed as h. The a-chung as a prefixed consonant is distinguishable from the velar fricative in that it is italicized; but this is unintentional since all prefixes are italicized in that article. There is, however, no distinction made between the transcription of the a-chung in initial or final positions and that of the velar fricative.

These are only two examples of the innumerable incongruities encountered in the transcription of Tibetan language. More cases could be cited; however, the purpose of this paper is not to enumerate such inconsistencies, but rather to seek their elimination in future publications.

In view of the increasing interest in Tibetan studies, it is desirable now more than ever that serious consideration be given to the acceptance of a standard system of Tibetan transcription. It is time to trade transcriptional independence for uniformity in order to facilitate and standardize the advancement of Tibetan studies. Admittedly, no single system of transcription can accurately reflect both the orthography and phonology of Tibetan; but diversity of spelling and pronunciation is not uncommon in other languages for which standard systems of transcription have been developed. This paper is concerned with a standard orthographic transcription for scholarly publication.

What, then, should be the criterion for a standard system of Tibetan transcription, It should be of minimal complexity and capable of reproduction on a standard typewriter, i.e., one lacking special keys for diacritical marks. The addition of diacritics, either by hand or by machine, requires time-consuming attention from the scholar. Even more important, it makes the same demands on the typesetter, who may not always share the scholar’s enthusiasm for exactitude. As a result, typographical errors caused by the omission of such marks are frequently encountered in publications; the most common one being the omission of the dot over an n for the nasal velar ng.

Justification for a diacritical mark rests in its indication of phonetic value and such marks are indespensible in phonetic transcriptions; however, in an orthographic transcription, such marks are extrinsic. Any transcription system exceeding the limit of minimal complexity wastes the time of the scholar, the printer and the reader. Fortunately, such exotic elements as the Greek gamma for a prefix g and a miniature circle for a prefix a-chung as used by Jäschke,³ have not gained favor. On the other hand, the dual function of an apostrophe to distinguish aspirates from unaspirates and to indicate the a-chung is still in use by some scholars. The indication of aspiration by an h, as in the transcription of aspirates in Sanskrit, eliminates this dual function of the apostrophe and maintains minimal complexity.

The use of the apostrophe to indicate an a-chung in final position as in mda’ (arrow), is excessive, since mda, even without the apostrophe, can be reconstructed in only one correct way i.e., with an a-chung final; however, this use of the apostrophe should be retained for consistency on a one-for-one transcriptional ratio, even though it exceeds minimal complexity.

The last, and perhaps the most pedantic, practice to be discussed in this paper is internal capitalization. The desire to distinguish prefix letters from the real initial letter appears early in Tibetan studies. In the article of 1932 by Csoma de Körös mentioned earlier, all prefix letters are prnted in italics; a tedious practice that auspiciously faded from the scene. It is to be noted that the first letter of the first word is capitalized in that article, even if it is an italicized prefix letter. With the discontinuation of italicizing prefixes, capitaliztion shifted from the first letter of a word to the so-called “initial”; thus beginning the tradition of internal capitalization.

There are two basic arguments in favor of “internal” capitalization, neither of which retains abstract validity when examined i concrete practice. The lexicographic argument maintains that capitalization of the true initial indicates to the reader the letter under which a given word may be found in a Tibetan dictionary. If the reader is sufficiently familiar wit h the Tibetan language to use a dictionary, is it not safe to assume the reader could find the word without the benefit of the initial capitalization? Whether safe or not, this seems to be the assumption of those using such capitalization, for its function in relation to lexicographic use is not explained to the reader. Moreover a survey of material published in transcription reveals that only the first word is subject to such capitalization. For one example, in a recent publication in which initial capitalization is practiced,⁴ contains thirty-three pages of Tibetan text in transcription in which only one letter is capitalized; i.e., the first letter of the first word.

The practice of capitalizing the “initial” of only the first word in compounded names results in an unsatisfactory transcription of words having an initial y with a prefix g as in g.yu (turquoise). For example, if g.yu is the first word, then the y is capitalized and the prefix g is not, i.e., gYu; but, if it occurs as other than the first word, then both the g and y are in small type without a separating mark, making it seem as if g is the initial and the y is subjoined, i.e., -gyu, which is a different word. One solution to such problems lies in the capitalization of each and every word, a proposal too demanding of manual and visual effort to be taken seriously. On the other hand, a proper system of transcription eliminates such problems regardless of capitalization. In fact such capitalization is a concession to Western practices, for there is no such thing as s capital letter in Tibetan.

The lexicographic justification for “initial” capitalization depends on its identification of the orthographic initial consistently, a justification not substantiated in actual practice.

The phonetic argument maintains that the initial should be capitalized as an indication to pronunciation and in order to distinguish prefix letters, which are silent in the dialect of Central Tibet. The application to English of such “phonetic” capitalization would result in such spellings as: hOur, kNight, pNeumonia, pSychiatry, and phTisic.

The phonetic argument suffers from the same practices as the lexicographic argument in that only the first word of a textual passage or of compounded names is subjected to capitalization. This evidences the assumption that the reader is familiar with Tibetan pronunciation; and if this be the case, what need is there for phonetic capitalization? An attempt to eliminate this situation could be made by phonetically capitalizing every single word, but such an attempt would be futile. What should be capitalized to show that the word bod is pronounced pö, that dbang is wang, or bya is chya, and so on? For published examples of the inconsistency involved in this practice, consider Roerich’s capitalization of the sub-joined l as the the phonetic initial; e.g., gLo (pronounced Lo) except where when subjoined to the lexicographic initial z; e.g., Zla (pronounced Da).⁵ Unable to capitalize a non-existent d the initial z was capitalized, again evidencing the assumption that the reader knows that the combination zl is pronounced d. This is but one example; many more could be cited, but many would prove no more than one that the capitalization of the “phonetic initial” is even less justified than the capitalization of the “lexicographic initial.”

A reader with knowledge of the Tibetan language needs no capitalization; those with no knowledge may find such random and inconsistent capitalizations intriguing but of little value in pronouncing Tibetan properly or in using a Tibetan dictionary. Because of the diversity between Tibetan orthography and phonology, it is sometimes desirable to transcribe Tibetan according to its pronunciation. When writing for the non-specialist, it would be pedantic to insist on the spelling Bkra-shis-lhum-po for the name Tashilhumpo, or Bla-ma for Lama. On the other hand much data about Tibet has been prepared for the non-specialist in which only a phonetic transcription is given. Such transcription reduces the value of those works for the specialist since the reconstruction of the correct orthography is sometimes impossible. For example Chango, the name of a village appearing on some Western maps at approximately 100° 30′ East and 31° 30′ North, is actually Brag-mgo according to it's orthography. Wherever it is desirable to transcribe words phonetically, it is suggested that the proper orthography be added in parenthesis, e.g., “. . . in the village of Chango (Brag-mgo). . . . ”

Since internal capitalization at random, whether of the orthographic or the phonetic initial, is valueless and total capitalization too cumbersome, it is suggested that Csoma de Körös’ original practice of capitalizing the first letter, whether a prefix or an initial, be restored if only for the sake of visual conformity to Western capitalization practices.

In conclusion, the following system, devoid of diacritical marks and representing minimal complexity, is suggested for adoption as the standard for Tibetan orthographic transcription.

This is the system devised years ago by members of the Inner Asia Project at the University of Washington, with a single alteration to be proposed: the substitution of a dot for a dash in transcribing an initial y with a prefix g. This substitution is made for the sake of visual continuity. When components of proper nouns are joined by dashes, a dash between a prefix g and an initial y isolates the g. For example, the transcription Yar-’brog-g.yu-mtsho is visually preferable to Yar-’brog-g-yu-mtsho. The revised system is therefore identical to the one used by the late René de Nebesky-Wojkowitz, with the exception that he practiced “internal lexicographic capitalization,”⁶ a practice not advocated in this paper for the reasons given earlier.

Footnotes:

¹ Charles A Bell, Grammar of Colloquial Tibetan (Allipore 1939), p.v.

² Csoma de Körös, “Translation of a Tibetan Fragment,” Journal of the Asiatic Society of Bengal I (1832) .269 ff.

³ H. A. Jäshke, Tibetan-English Dictionary, (London, 1881).

⁴ Alfonsa Ferrari, mK ’yen brtse’s Guide to the Holy Places of Central Tibet, Serie Orientale Roma, XVI (Rome, 1958).

⁵ George N. Roerich, The Blue Annals, Vol. II (Calcutta, 1953), pp. 1154-5, 1234-5.

⁶ René de Nebesky-Wojkowitz, Oracles and Demons of Tibet (The Hague, 1956).

Home | Index