^{page 702 note 2} See p. 717 below.

^{page 702 note 3} Here and later, all Bodish verb forms are given in the following order unless otherwise noted: present, perfect, future, imperative.

^{page 703 note 1} As by Conrady, August, Eine indochinesische causativ-denominativ-Bildung und ihr Zusammenhang mil den Tonaccenten, Leipzig, 1896, p. 26.Google Scholar

^{page 703 note 2} As Conrady's explanation of -s of āk'eṅs k'eṅs, op. cit., 19. Native grammarians had no clear conception of the difference between a functional, suffixed consonant and the final consonant of a stem, between a stem and a non-syllabic morpheme.

^{page 703 note 3} Körösi, Sándor Csoma (Alexander Csoma de Koros), A Grammar of the Tibetan Language in English, Calcutta, 1834Google Scholar, and Cordier, Palmyr, Cours de tibétain classique, Hanoi, 1908, 49 ff.Google Scholar, came the closest to a scientific classification. Ph(Ilippe) Éd(Ouard) Foucaux, Qrammaire de la langue tibétaine, Paris, 1858Google Scholar, repeated Csoma. Conrady, op. cit., and Li, Fang-Kuei, “Certain Phonetic Influences of the Tibetan Prefixes upon the Root Initials”, GYYY 4 (1933)Google Scholar, arranged their materials to prove their contentions rather than as analyses. In Jäschke's, Tibetan Grammar, Berlin and Leipzig, 1929, pp. 99 ff.CrossRefGoogle Scholar, verbs are arranged according to the number of stems.Bacot, Jacques, Grammaire du tibétain littéraire, I, Paris, 1946Google Scholar, noted consonantal and vocalic alternation in some intransitive-transitive forms (p. 58) and made some general observations regarding the functions of prefixes, alternation, etc. (pp. 49 ff.), but did not attempt a classification of verb forms. Native grammarians seem generally to have confined themselves to generalizations regarding the use and function of prefixes and suffixes (affixes), as in Schubert, Jöhannes, “Tibetische Nationalgrammatik”, MS0S 31 (1928), 1–59Google Scholar, and 32 (1929), 1–54; but since no attempt was made to classify the forms found in a large representative sample of all types of verbs, such generalizations are generally true only of certain types of verbs, not of others.

^{page 703 note 4} Such verbs may not have been unalterable in proto-Bodish; for corresponding to O.B. mt'oṅ “see”, is West Himalayish taṅ.

^{page 703 note 6} See note at the beginning of this article.

^{page 704 note 1} Desgodins: pf. tś'ag, im. tś'og(s).

^{page 705 note 1} <*-sdre; <*-stre?

^{page 706 note 1} Desgodins: pf. byug.

^{page 706 note 2} The term “root” is necessarily used rather loosely, in this exploratory stage, of certain predominant characteristics of the forms of the present, perfect, future, and imperative. The root may be said to correspond to the perfect except for b- and -s affixes.

^{page 706 note 3} See below.

^{page 708 note 1} See p. 709–11. In reconstructions t^d represents a dental stop.

^{page 709 note 1} Desgodins: pf. brgyoṅs, fut. brgyoṅ.

^{page 709 note 2} I. to rkyoṅ above.

^{page 709 note 3} “prob. the original form, but of rare occurrence, for rdab-pa” (Jäschke).

^{page 709 note 4} It is uncertain whether verbs of this sort belong to Type 1 or Type 4.

^{page 710 note 1} Desgodins: pf., fut. bskyor.

^{page 710 note 2} Desgodins: pf., im. spogs.

^{page 710 note 3} Desgodins 604 b.

^{page 711 note 1} For other examples see Fang-kuei Li, op. cit. (p. 703, n. 3), p. 138 ff.

^{page 711 note 2} The ld- of the intransitive has nothing to do with the appearance of z- in this verb; cf. sloṅ, (b)slaṅns, (b)slaṅ, sloṅ(s) “cause to rise”, i. Idaṅ, laṅs or ldaṅs.

^{page 711 note 3} “Morphophonic” designates morphophonemic alternation of vowels; “morphosymphonic”, the morphophonemic alternation of consonants. These are two of the principal problems in Bodish, and the latter in Sino-Tibetan comparative grammar. A single term is needed to designate each of these phenomena.

^{page 711 note 4} Desgodins throws doubt on this whole series: śoṅ, bśoṅs, bśoṅ śod, im. of ātś'ad; to śom, bśams or bśoms, bśam, bśom, he added “vulg. ad praet. et fut.”; and considered śu vulgar.

^{page 712 note 1} p. 299.

^{page 712 note 2} Besides the above irregular -a- stem verbs with g- prefix in the present, the only probable four-stem verbs found were gso, (b)sos or gsos; i. āts'o, sos, im. sos “feed” glśoṅ, bśoṅs or ††bśaṅs (although attributed to Jäschke) “excavate” However, only the presents of the transitive verb are given in the examples. One may note which are irregular -a- stems, as in Type 7. We have unfortunately nothing to indicate whether d- was a root prefix or the result of levelling of original dpog, *p'ag, dpag, *p'ogs.

^{page 713 note 1} Except verbs with labial initials and roots with sibilant initials.

^{page 713 note 2} dg represents d or g.

^{page 713 note 3} Examples only of k'ebs.

^{page 713 note 4} Reconstruction based on Sbalti kuk'-pa, Sikkim kug-gu, Dbus kug-pa “bend”.

^{page 714 note 1} Jäschke bgod; but in examples nearly always bkod. Desg. bkod only.

^{page 714 note 2} Das, p. 68; substantiated by West Bodish kye-.

^{page 716 note 1} See BSOAS., xiii, pt. 4.

^{page 716 note 2} According to Conrady, op. oit. (p. 703, n. 1 above).

^{page 717 note 1} From T. Grahame Bailey, “Kanauri Vocabulary in Two Parts: English-Kanauri and Kanauri-English”, JRAS (1910), 659–705; (1911), 315–364; Bailey, Linguistic Studies from, the Himalayas, Asiatic Society Monographs, Vol. 18, London, 1920; and Joshi Tika Bam. “A Grammar and Dictionary of Kanauri”, JASB, n.s. 5, extra number, 1909.

^{page 717 note 2} Probably actually tság-, not tśdg-, as it is taken from Tika Ram's vocabulary, where both tś and ts are recorded as ch. Bailey, who distinguished the two phonemes, did not give this word.

^{page 717 note 3} Bailey, 1920 (see n. 1 above).

^{page 718 note 1} tś and ts probably not distinguished. These words are from A. Henry, “The Lolos and other Tribes of Western China”, JAI 33 (1903), 101. Henry did not specify the dialect from which these words were cited. He studied the Lolo of Szemao and Mengtse and knew the vocabulary of the Lolos of Taliang. I can only state that the words are not from Woni or Pula.

^{page 718 note 2} Op. cit. (p. 703, n. 1), pp. 108–111.

^{page 718 note 3} p. 53.

^{page 718 note 4} ā, m- are classified here as nasal prefixes; all others as non-nasal, except 0. (zero).

^{page 718 note 5} pp. 73 ff.

^{page 719 note 1} The demonstration of this is too long to be presented here, but will appear in another work.

^{page 720 note 1} This development found in Sbalti and Burig probably extended over all Tibet. In East Bodish the best preserved dialect is Khams, recorded only by Jäschke, who noted vocalic changes and extreme consonantal shifts such as O.B. gr to Khams dr, O.B. d- prefix to Khams γ-, O.B. b- prefix to Khams v-, etc.; but like some modern phonemicists, Jäschke was inclined to write in a conventional manner if possible. Consequently it is only by the slips he made that we can infer that surd prefixes became sonant before sonant initials in Khams, and that sonant prefixes became surd before surd initials, as O.B. sgra, Khams zdra; O.B. sbal-ba, Kh. zual-wa; O.B. sbyar-ba, Kh. zuar-wa; O.B. dpyid, Kh. xśid.

Unduly influenced by Conrady, A. H. Francke, in collaboration with W. Simon, made the statement in the addenda to Jäschke's Tibetan Grammar (Berlin and Leipzig, 1929), 139, that “a prefixed generally raises a media to a tenuis”, which is belied by the much greater number of contrary examples following.

Some minor points in Conrady's work need correction: s- prefix probably never caused aspiration of even a few forms in Old Bodish, as Conrady believed; those he cited are probably modern dialect forms which have occasionally found their way into the dictionaries. His statement that some prefixes became s- in Ladwags should be modified to read only before surd and nasal initials.

^{page 720 note 2} Op. cit. (p. 703, n. 3), p. 152.

^{page 721 note 1} An alternative possibility will be suggested in the concluding portion of this article (BSOAS. xiii/4).

^{page 721 note 2} Op. cit, p. 28.

^{page 721 note 3} p. 138. Li did not seem aware that Conrady (p. 28) had suggested this. This is not surprising when one considers the confusion of Conrady's ideas and his disorganized presentation.

^{page 722 note 1} Absolute initial unaspirated surd stops tend to become sonant in the Ladwaga dialect; see Sten Konow, LSI, Vol. 3, Pt. 1, p. 53. We may surmise that this is an extension to surd initials in a modern dialect of a phonetic shift which transformed surd stop prefixes (at that time absolute initials) of proto-Tibeto-Burmic into sonant stops, in this case *p- > O.B. b-, before the introduction of writing into Tibet. All stop consonant prefixes are sonant in Old Bodish, which can hardly have represented a primitive condition.

^{page 722 note 2} As the equivalences for these phonemes, even when initial, have not been determined, an attempt to discover their equivalences when prefixed has small chance of success at present.

^{page 722 note 3} One may note here that some languages recorded in Tibetan writing in Chinese Turkestan in medieval times, and closely resembling Old Bodish phonetically, employ ã- prefix much more frequently than Old Bodish. In Nam, an Ancient Language of the Sino-Tibetan Borderland, London, 1948, Prof. Thomas, F. W.Google Scholar has revealed a language which is almost certainly Bodish, i.e. very closely related genetically to Old Bodish but which has an extraordinary development of ã- prefix. This may indicate that Old Bodish lost some of the proto-Bodish ã- prefixes; but knowledge of these languages has not advanced far enough to discuss their morphology.

^{page 723 note 1} Except that kl-, not k'l-, is found.

^{page 724 note 1} Neither the ā- nor the g- prefix so frequently found in “Nam” (see p. 722, n. 3) control the aspiration or non-aspiration of a following surd initial, as in Old Bodish. “Nam” may be of prime importance for the determination of primitive Bodish aspiration when translation of the new language has progressed further.