While the relation between script and pronunciation in Tibetan is complicated it is generally possible to deduce the pronunciation of a given letter combination by fixed rules. The relation between orthography and pronunciation is thus more reminiscent of the situation in Francien than in English.
Transliteration and transcription practice
In general each foreign language uses its own system for transcribing the pronunciation of spoken Tibetan, but the one preferred by most Tibetans is the one devised in the 1860's by the Scandinavian missionary Thorvald Jacobsen Rømmestad (Tibetan ཐུར་དྦལ་ཡ་ཀོ་པ་ Thur-dbal-Ja-ko-pa [Thúrwǽl Jakóba]) based on Norwegian and Danish letter values and originally printed in Fraktur.
In the following all examples will be shown both in Xrírámpur romanization, which reflects Tibetan spelling and in Rømmestad transcription which reflects pronunciation.
The relation between complex spellings and pronunciation
The main source of complication is the fact that in the dialects of central Tibet, among them that of Hlása, ancient consonant clusters and syllable-final consonants have become reduced. As a result of this most single consonants of the language can be spelled in several ways, often with clusters of conjoined consonant letters where only one consonant is pronounced. Some of these conjunct spellings are equivalent to single-consonant spellings corresponding to one of the letters of the conjunct, but in other cases the lost consonants have affected the pronunciation of the preserved consonant, so that many conjunct spellings correspond to distinct sounds of the language which cannot be written with a single consonant letter, but only with various conjuncts; this is particularly the case with voiced stops and affricates which are all written with conjuncts.
Voiced stops and affricates
The single letters of the third grade, corresponding to the Sanskrit voiced stops, are pronounced as unaspirated or aspirated voiceless stops or affricates, but when written with superscript or prefixed letters they are pronounced as the corresponding voiced stops or affricates. Thus ga za da ba z̒a are pronounced [ka tja ta pa za] but rga rza rda rba rz̒a are pronounced [ga dja da ba dza].
The letter z
It should be noted that the letter z has very different meanings in the Zrírámpur romanization and Rømmestad transcription. In the Zrírámpur system the unadorned z corresponds to alveopalatal affricates and the z̒ with a superscript inverted comma corresponds to dental affricates. In Rømmestad's system the unadorned [z] on its own indicates an unaspirated voiceless dental affricate, while the digraphs [tz] and [dz] stand for an aspirated voiceless dental affricate and a voiced dental affricate respectively. Rømmestad avoided the according to him 'uncouth' digraph [zh] in favor of a distinction between the to him familiar [z] and [tz], common in German and in Scandinavian family names where they both corresponded to the same unaspirated or weakly aspirated alveolar affricate. The rare z̒h̦ corresponding to Sanskrit zh he was forced to render by [dzh], but it hardly ever occurred in the Christian texts which he wrote and translated. Despite its illogic this way of transcribing the dental affricates has remained in use.
It should also be noted that the sounds stranscribed as digraphs with a following -j, -r or -z are all perceived as single sounds by Tibetan speakers, as are the aspirated sounds written as digraphs with -h or h-. The digraphs are partly derived from Tibetan spelling, since the modern simple sounds like [c cʰ ɟ ɲ ʈ ʈʰ ɖ] go back to ancient clusters or where written as such kj khj gj ñ mj kr pr khr phr gr br in the Sanskrit-derived script. The alveopalatal sounds c ch z ñ x [tj thj dj nj sj] in fact go back to clusters or palatalized dental stops in pre-Tibetan, but are all written with single letters which were used for alveopalatal sounds in Sanskrit. Rømmestad's transcription with [tj thj dj nj sj] is due to his native Norwegian pronunciation of such spellings in Danish and Riksmål, which in turn is due to a historical phonology parallel to that of Tibetan! The spellings are unambiguous in Tibetan and have therefore remained in use among Christian Tibetans and when Buddhist Tibetans transcribe their language into Roman letters.
In a few cases the pronunciation of a conjunct is further removed from the ancient cluster represented by the spelling, as with zl which is pronounced [da] or dba, dbja and dbra which are pronounced [wá já rá]
Among the stops and affricates grade one and two of the alphabet, corresponding to the Sanskrit unaspirated voiceless and aspirated voiceless stops occur in high-tone syllables while grade three and four, corresponding to Sanskrit voiced unaspirated stops and to nasals occur in low-tone syllables.
Lost ancient consonants have also in many cases affected the tone of the syllable. In particular former clusters involving nasals and semivowels have resulted in high-tone syllables, while syllables written with a single nasal or semivowel letter are pronounced with a low tone.
In the Rømmestad transcription high tone is indicated with an acute accent on the stressed vowel, while low tone is unmarked. Thus ka kha sña bla become [ká khá njá lá] while sga ña la become [ga nja la].
Low-tone syllables with aspirated stops
In the pronunciation of some Tibetans the unconjoined grade three letters ga za da ba z̒a are pronounced as low-tone syllables with aspirated voiceless stops or affricates, and the same letters with the superscripts or prefixes r-, l-, s-, g-, d- and b- are pronounced as low-tone syllables with unaspirated voiceless stops or affricates, and only grade three letters with prefixed h or m are pronounced as low-tone syllables with voiced stops or affricates. This pronunciation is fully valid and is even regarded by some with a degree of prestige, and it is entirely possible to represent this pronunciation in Rømmestad transcription, although usually the more common pronunciation with unaspirated grade three sounds and all conjuncts corresponding to voiced sounds is followed in transcriptions.
|Spelling||Common pronunciation||Variant pronunciation|
|ka ca ta pa c̒a||ká tjá tá pá zá|
|kha cha tha pha c̒ha||khá thjá thá phá tzá|
|ga za da ba z̒a||ka tja ta pa za||kha thja tha pha tza|
|rga rza rda rba rz̒a||ga dja da ba dza||ka tja ta pa za|
| mga mza mda mz̒a |
hgah hzah hdah hbah hz̒ah
|ga dja da ba dza|
TO BE CONTINUED