The verb, which is properly a kind of noun or participle, has no element of person, and denotes the conditions of tense and mood by an external and internal inflexion, or the addition of auxiliary verbs and suffixes when the stem is not susceptible of inflexion, so that instead of saying “I go,” a Tibetan says “my going.” The conditions which approximate most closely to our present, perfect, future and imperative are marked either by aspiration of the initial or by one of the five prefix consonants according to the rules of euphony, and the whole looks like a former system thrown into confusion and disorder by phonetic decay. As to the internal vowel, a or e in the present tends to become o in the imperative, the e changing to a in the past and future; i and u are less liable to change. A final s is also occasionally added. Only a limited number of verbs are capable of four changes; some cannot assume more than three, some two, and many only one. This deficiency is made up by the addition of auxiliaries or suffixes. There are no numeral auxiliaries or segregatives used in counting, as in many languages of eastern Asia, though words expressive of a collective or integral are often used after the tens, sometimes after a smaller number. A good deal of new research on the grammar is to be found in Grierson's Linguistic Survey of India, part III., 1908. In scientific and astrological works, the numerals, as in Sanskrit, are expressed by symbolical words. In the order of the sentence the substantive precedes the adjective and the verb stands last; the object and the adverb precede the verb, and the genitive precedes the noun on which it depends—this contrasts with the order in the isolating Chinese, where the order is subject, verb, object. An active or causal verb requires before it the instrumental instead of the nominative case, which goes only before a neuter or intransitive verb. The chief differences between the classical language of the Tibetan translators of the 9th century and the vernacular, as well as the language of native words, existed in vocabulary, phraseology and grammatical structure, and arose from the influence of
the translated texts.
The Tibetan language, presenting such marked differences between its written and spoken forms, has a great interest for philologists, Philology. on account of its bearing on the history of the monosyllabic languages of eastern Asia, with their so-called “isolation” or absence of form-words and consequently of grammatical forms. Is the Tibetan a monosyllabic language passing to agglutination, or the reverse? The question has turned mainly upon the elucidation of the phenomenon of the silent letters, generally prefixed, which differentiate the spelling, of many words from their pronunciation, in the central dialect or current speech of Lhasa. Rémusat rather dubiously suggested, while Schmidt and Schiefner maintained, that the silent letters were a device of grammarians to distinguish in writing words which were not distinguished in speech. But this convenient opinion was not sufficient for a general explanation, being supported by only a few cases. Among these are—(a) the addition of silent letters to foreign words in analogy with older terms of the language (e.g. the Persian tadjik was transcribed staggzig or “tiger-leopard,” because the foreign term left untouched would have been meaningless for Tibetan readers); (b) the addition for the sake of uniformity of prefixed letters to words etymologically deprived of them; (c) the probable addition of letters by the Buddhist teachers from India to Tibetan words in order to make them more similar to Sanskrit expressions (for instance rje- for “king,” written in imitation of raja, though the original word was je or she, as is shown by cognate languages). On the other hand, while phonetically the above explanation was not inconsistent with such cases as rka dkah, bkah, bska, and nga, rnga, ngad, sngags, lnga, ngad and brtse, brdzun, dbyar, &c., were the italicized letters are renounced in full and the others are left aside, it failed to explain other cases, such as dgra, mgron, spyod, spyan, sbrang, sbrul, bkra, k'ri, krad, k'rims, k'rus, &c., pronounced ḍa, ḍon, čod, or šwod, čen, dang, ḍeu, ṭa, t'i, tad or teh, t'im, tu, &c., and many others, where the spoken forms are obviously the alteration by wear and tear of sounds originally similar to the written forms. Csoma de Körös, who was acquainted with the somewhat archaic sounds of Ladak, was able to point to only a few letters as silent. Foucaux, in his Grammaire (1858), quoted a fragment from a native work on grammar several centuries old, in which the pronunciation of the supposed silent letters is carefully described. Since then the problem has been disentangled; and now minor points only remain to be cleared up. Jaeschke devoted special attention to the dialectical sounds, and showed in several papers and by the comparative table prefixed to his dictionary that in the western and eastern dialects these sounds correspond more or less closely to the written forms.
Jaeschke first noted the existence of tones in Tibetan, and these have been found by Professor Conrady to have developed on the Tones. same lines as in Chinese. Thus intransitive bases seem to have begun only with soft consonants, and it is doubtful whether the parent tongue possessed hard consonants at all; while transitive bases were formed by hardening of the initial consonants and at the same time pronouncing the words in a higher tone, and these two latter changes are supposed to have been indicated by a prefix to the base-word. Many of these old soft initial consonants which are now hardened in the modern dialects are preserved in classical Tibetan, i.e. in Tibetan of the 7th to the 9th century A.D. The old language seems to have pronounced prefixes extensively which in modern pronunciation in central Tibet are largely lost, whilst the soft initials have become aspirated or hardened and tones have developed, and in the west and east, where prefixes and soft initials have been preserved, there are no tones. Thus the valuable testimony of these dialects may be added to the evidence furnished by foreign transcriptions of Tibetan words, loan words in conterminous languages, and words of common descent in kindred tongues. And the whole shows plainly that the written forms of words which are not of later remodelling are really the representatives of the pronunciation of the language as it was spoken at the time of the transcription.
The concurrence of the evidence indicated above enables us to form the following outline of the evolution of Tibetan. In the 9th century, as shown by the bilingual Tibeto-Chinese edict at Lhasa, there was relatively little difference between the spoken and the written language. Soon afterwards, when the language was extended to the western valleys, many of the prefixed and most of the important consonants vanished from the spoken words. The ya-tag and ra-tag, or y and r subscript, and the s after vowels and consonants, were still in force. The next change took place in the central provinces; the ra-tags were altered into cerebral dentals, and the ya-tags became č. Later on the superscribed letters and finals d and s disappeared, except in the east and west. It was at this stage that the language spread in Lahul and Spiti, where the superscribed letters were silent, the d and g finals were hardly heard, and as, os, us were ai, oi, ui. The words introduced from Tibet into the border languages at that time differ greatly from those introduced at an earlier period. The other changes are more recent and restricted to the provinces of Ü and Tsang. The vowel sounds ai, oi, ui have become ē, ō, ū; and a, o, u before the finals d and n are now ä, ö, ü. The mediae have become aspirate tenues with a low intonation, which also marks the words having a simple initial consonant; while the former aspirates and the complex initials simplified in speech are uttered with a high tone, or, as the Tibetans say, “with a woman's voice,” shrill and rapidly. An inhabitant of Lhasa, for example, finds the distinction between s' and z', or between s and z, not in the consonant, but in the tone, pronouncing s' and s with a high note and z' and z with a low one. The introduction of the important compensation of tones to balance phonetic losses had begun several centuries before, as appears from a Tibetan MS. (No. 462b St Petersburg) partly published by Jaeschke (Monatsber. Akad. Berl., 1867). A few instances will serve to illustrate what has been said. In the bilingual inscriptions, Tibetan and Chinese, set up at Lhasa in 822, and published by Bushell in 1880, we remark that the silent letters were pronounced: Tib. spudgyal, now pugyal, is rendered suh-pot-ye in Chinese symbols; khri, now t'i, is kieh-li; hbrong is puh-lung; snyan is sheh-njoh and su-njoh; srong is su-lun, su-lung and si-lung. These transcriptions show by their variety that they were made from the spoken and not from the written forms, and, considering the limited capacities of Chinese orthoepy, were the nearest attempt at rendering the Tibetan sounds. Spra or spreu (a monkey), now altered into deu at Lhasa, ṭeu in Lahul, Spiti and Tsang, is still more recognizable in the Gyarung shepri and in the following degenerated forms—shreu in Ladak, streu-go in Khams and in cognate languages, soba in Limbu, saheu in Lepcha, simai in Tablung Naga, sibeh in Abor Miri, shibe in Sibsagar Miri, sarrha in Kol, sara in Kuri, &c. Grog-ma (ant), now altered into the spoken t'oma, is still kyoma in Bhutan, and, without the suffix, korok in Gyarung, k'oro- in Sokpa, k'orok, k'alek in Kiranti, &c. Grang-po- (cold), spoken t'ammo, is still grang-mo in Takpa, k'yam in Burmese, &c. A respectful word for “head” is ü, written dbu, which finds its cognates in Murmi thobo, Sibsagar Miri tub, &c. Bya (bird), spoken cha, is still pye in Gyarung. Brjod (to speak), pronounced jod, is cognate to the Burmese pyauhtso, the Garo brot, &c. The word for “cowries” is 'gron- in written, rum- in spoken Tibetan, and grwa in written Burmese; slop (to learn), spoken lop, is slop in Melam. “Moon” is zlava in written and dawa in spoken language, in which -va is a suffix; the word itself is zla-, cognate to the Mongol ssara, Sokpa sara, Gyarung t-sile, Vayu cholo, &c. The common spoken word for “head” is go, written mgo, to which the Manipuri moko and the Mishmi mkura are related. Sometimes the written forms correspond to double words which have disappeared. For instance, gye (eight), which is written brgyad and still spoken vrgyad in Balti in the west and Khams in the east, is gyäd in Ladak, Lahul, Tsang and Ü. The same word does not appear elsewhere; but we find its two parts separately, such as Gurung pre, Murmi pre, Taksya phre and Takpa gyet, Serpa gye, Garo chet, &c. Rta (horse) is reduced to ta in speech, but we find ri, rhyi, roh in Sokpa, Horpa, Tochu, Minyak, and tā, tah, teh, t'ay in Lhopa, Serpa, Murmi, Kami, Takpa, &c., both with the same meaning. Such are the various pieces of evidence obtained from an endless number of instances. The cases referred to above do not, owing to the difference of the causes, yield to any explanation of this kind. And it must be admitted that there are also many cases, some of them caused by irregularities of writing, modification of spelling by decay, and by a probable use of prefixes still unascertained, which also resist explanation, though the
