All illustrations from volumes other than Ka, Ga and A are courtesy of the British Library. Illustrations from Ka are courtesy of the Bodleian Library at Oxford. The Bodleian Library has agreed access to these miniatures from Ka following registration (register from the contents page or from a link under each controlled image). Images from the first folio of Ga and A are courtesy of the Victoria and Albert Museum. All other images from Ga and A are courtesy of the British Library. The Victoria and Albert Museum has agreed access to these miniatures from Ga and A following registration (register from the contents page or from a link under each controlled image). We appreciate the cooperation of these institutions.


"Not Wylie" Conventions

In due course, we hope to produce a Tibetan language version of our catalogue of the Rig 'dzin Tshe dbang nor bu edition of the rNying ma'i rgyud 'bum, making it more accessible to Tibetan (and other Asian) scholars who do not read English or who find it difficult to read Tibetan in transcription. In the meantime, our catalogue transliterates exact Tibetan spellings, as given in the original manuscripts (perhaps unusually, it also does this in the manner of an exhaustively detailed diplomatic edition) 1. The transliteration conforms to the system for the transliteration of Tibetan syllables which avoids diacritical marks and which has been widely used internationally in Tibetan studies since the late 1950s, especially in UK and North America (in continental Europe, systems using diacritical marks remain common). This particular system is widely referred to as, "Wylie transliteration", following the publication of Turrell Wylie's article in 1959 (A Standard System of Tibetan Transcription, Harvard Journal of Asiatic Studies: 261-7). However, Wylie did not develop the basic system, and was not the first to publish work using it. He acknowledges (p.267) that except in one significant respect (discussed below), his system is identical to that used earlier by René de Nebesky-Wojkowitz ( Oracles and Demons of Tibet, The Hague, 1956). One might add that other scholars had also adopted this system before Wylie's article: see, for instance, David L. Snellgrove's Buddhist Himālaya (Oxford, 1957: 299-300), and the same author's discussion of the issues in his Indo-Tibetan Buddhism (Boston and London, 1987: xxiv).

1 See "The Catalogue as a Partial Diplomatic Transcription" .

It might seem pedantic to question the naming of the system. However, the loose usage of the term, "Wylie", to refer to the general system of transliteration, can cause confusion, and it is necessary to clarify that we do not here follow the single significant suggestion which Wylie proposed in his article. This proposal was that where capitalization of a word is desirable in English prose, the first letter of the word should be capitalized, regardless of whether it is the root letter of the syllable (ming gzhi), or a prefix (sngon 'jug) or head letter (mgo). This convention has been widely adopted, especially in the works of North American scholars, but it has not been universally accepted. In our catalogue, capitalization is not generally needed, but where it is used, we follow Snellgrove, Nebesky-Wojkowitz, and other (especially European) scholars, who capitalize the root letter of the first syllable of the word. In the case of Tibetan syllables which are transcribed using more than one Roman consonant (such as nya, sha etc.), only the first Roman letter in the transcription of the root syllable is capitalized. Thus, we give rNying ma, which in strict Wylie transliteration would be given as Rnying ma.2 The capitalization of the root letter means that attention is drawn to it. In Tibetan writing, capitalization is not used, but attention is drawn conceptually to the root syllable or "basic letter" ( ming gzhi), and Tibetan dictionaries list words under these syllables.

2 An alternative convention, capitalizing the entire root letter, would give rNYing ma: this option is rarely adopted by contemporary scholars.

In one particular, the transliteration here departs from that used by Nebesky-Wojkowitz and Snellgrove:3 the subjoined syllable wa (wa zur btags pa) is given as "wa" and not "va", in conformity to its Tibetan language identification as a representation of an attached wa. In contemporary international usage, either alternative is common.

3 Wylie is silent on this issue, but since he claims his system to be identical to Nebesky-Wojkowitz apart from its use of capital letters, the implication would seem to be that he does not differ on this point.

Since our manuscripts include some constructions which are not used in everyday Tibetan written language (and not commented on by Nebesky-Wojkowitz, Wylie or Snellgrove), it is also necessary to include a word on how we transliterate more complex constructions. In some cases, where a word is written with all its syllables but in an unconventional manner, we transliterate in the usual way but add an explanatory note. For instance, where a final letter of a syllable is in fact subscribed to the root syllable, often to save space at the end of a line, we give the word in usual transliteration but add a note that the final letter is subscribed.4 Where an abbreviated form of a word is given, we tranliterate as exactly as possible,5 but note that the word is abbreviated: we have given a list of such abbreviations and the full words to which they correspond (see “List of Abbreviations used in the Edition” . Again, to save space, the Tibetan final letter, "ma", may be given as a small circle above the syllable (a practice borrowed from the nasalising Sanskrit anusvāra), and we distinguish this from the complete form of the letter by a full stop given after the "m" (eg. in gsum.). Occasionally, a sentence may be concluded with a syllable "sa" together with a subscribed, "ha". We have rendered this with a hyphen between the "s" and "ha" (eg. rdzogs s-ho), to indicate that we have a combination of two separate letters, and not the Tibetan syllable, "sha".

4 In one case, that of the word, "nyer", equivalent to "nyi shu" in compound numbers, this is almost invariably written in our manuscripts with the "ra" subscribed to the "nye" as an attached "ra" (ra btags). Thus, in this case alone, we do not note that the "ra" is subscribed but instead note when it is not subscribed.
5 Sometimes, exact precision produces strange results, since an abbreviation may combine syllables in a way which, following usual rules for transliteration, does not entirely correspond to their order in their full form. For example, we find the combination, bcyod, which represents an abbreviation of bco brgyad (eighteen), and thus is conceptually "bco" + "(brg)yad". In the case of abbreviations such as bcui (for bcu bzhi) and bcuis (for bcu gnyis), we have followed the conceptual order in giving the zhabs kyu ("u") before the gi gu ("i"), in the absence of any clear rule of precedence for transliterating more than one vowel on the same root letter.

Since it was desirable to avoid diacritics in our Web catalogue, we have not given diacritical marks for Tibetan renderings of Sanskrit letters, but have adopted simple pragmatic - if rather ugly - conventions. We use an apostrophe after the vowel to represent long vowels, such as Sanskrit "ā" or "ū" (eg. in ma ha'), which are given in Tibetan as a small letter "a chung" ("'a" in transliteration) subscribed to the root syllable. Thus, in this instance, since a Tibetan final "a chung" is also transcribed using an apostrophe, our usage approximates to the actual Tibetan rendering. 6 Similarly, where possible, we follow the letters given in Tibetan to represent Sanskrit syllables, including rendering as "ee" the double 'greng bu given for the Sanskrit diphthong "ai" (as in "Bee ro tsa na" for "Vairocana"), and giving "sha" for the Sanskrit palatal sibilant "śa". In the case of retroflex letters, such as the Tibetan "ta" written in reverse (as in 'mirror-writing') to represent the Sanskrit retroflex stop "ṭa", we have given a full stop after the letter (t.a). The Tibetan "sha" written in reverse to represent the Sanskrit retroflex sibilant "ṣa", is given as "s.a". Sanskrit anusvāra "ṃ", rendered in Tibetan as a small circle above the syllable - as in the shortened form of the Tibetan final letter "ma" (see above) - is given as "m." (eg. in hu'm.).

6 It should be obvious from the particular combination of letters and the context whether we have a final "a chung" as part of a Tibetan word, or a subscribed "a chung" representing a Sanskrit long vowel. Where a Tibetan final "a chung" is in fact subscribed, this is noted, in accordance with the practice explained above for subscribed final letters.

We have also aimed to indicate any punctuation found in the original. Thus, the shad is given as "/", and the occurrence of the usual yig mgo is noted, with an indication of whether this consists of a single, double, or triple yig mgo ("x 1", "x 2", or "x 3" respectively). More elaborate or ornamental shad or yig mgo are described, or where they are included in the unicode list for Tibetan, the appropriate unicode reference number has been noted.7 In particular, we find numerous instances of gter tsheg (unicode F14), which we have indicated by, "%", and of rnam bcad (unicode F7F), which we have rendered as, ":". Similarly, the presence of ornamental brackets and of other unusual features is noted.

7 Thus, when the catalogue is converted into a Tibetan language version, it should be possible to represent such features.


Return to Contents
Return to Tibetan Studies at Kent

All English text and transliterated Tibetan and other transliterated text is copyright 1999-2003 of Cathy Cantwell and Robert Mayer; the xml document structure and design is copyright 1999-2003 of Michael Fischer. The digital images of Volume Ka are the copyright 2001 of the Bodleian Library Oxford; all other digital images of the Tibetan manuscripts are the copyright 2001 of the British Library. Other artwork is copyright 2000-2003 by Cathy Cantwell. The web collection is copyright 2001-2003, Centre for Social Anthropology and Computing.