^{page 954 note 1} Cf. p. 386. See also here below, pp. 964 and 973.

^{page 954 note 2} Jäschke, , Tibetan Grammar, 3rd ed. (reprinted Berlin, 1929), p. 20. See also the “Addenda” to this Reprint (by A.H. Franeke), pp. 112–13.CrossRefGoogle Scholar

^{page 954 note 3} Franeke, , Addenda, p. 112.Google Scholar

^{page 955 note 1} See here below, pp. 968, etc.

^{page 955 note 2} Cf. Grünwedel, A., Legenden des Naropa (Leipzig, 1933), p. 175Google Scholar. I presume that the group cog ge , occurring several times after a negative imperative in a song which is transmitted in the bTsun mo bkaẖi yig (see Laufer, B., Roman einer Tibelischen Kōnigin (Leipzig, 1911), p. 122, n. 2Google Scholar) belongs here as well, developed by reduplication from *cog ge cog. See here later, p. 961, n.2. For the treatment of the finale, e, cf. yig-ge and in compounds yig “letter”, lhag-ge and lhag-pa “superior”, and “clear”, śig-ge-ba (also ljig-ge-ba, cf. Grünwedel, loc. cit., p. 180), and śig-śig “close together”, gsal-le-ba and gsal-ba “to be clear”, rtog-ge-ba and rtog-pa “reasoning”, log-ge-ba and log-pa “reversed, inverted”.

^{page 955 note 3} The word-family is also among those adduced by St. N. Wolfenden in his paper 'Concerning the Variation of Final Consonants in the Word Families of Tibetan, Kachin, and Chinese” (JRAS., 1937, pp. 629–630), though I should disagree as to the inclusion of some members (especially Icag “rod”).

^{page 956 note 1} With ẖdogs-pa (Perf. btags, Fut. gdag(s), Imper. thogs) is connected probably thag-pa “rope”, thags “texture”, tha-ga-pa “weaver”.

^{page 956 note 2} I am indebted to Dr. A. Waley for recalling Japanese kokorobosoi (“heartnarrow”) “anxious”. I may also refer to Ch. Féré, “La Physiologie dans les métaphors” (Revue de Philosophie, xl (1895), pp. 352, etc., esp. p. 356).

^{page 956 note 3} Cf. Meyer-Lübke, W. in his Romanisches etymologisches Wōrterbuch (3rd ed., Heidelberg, 1935, Mo. 468)Google Scholar, who renders its meaning by “Angst”, which is itself related to angustia.

^{page 957 note 1} No. 6323 in Sakaki's edition (Kyôto, 1916–1925). The Sanskrit original (yasyedânîm kâlam manyase) has clearly not influenced the Tibetan idiomatic rendering.

^{page 957 note 2} Cf. here below, p. 974.

^{page 957 note 3} Grammaire de la langue tibélaine (Paris, 1858), p. 20.Google Scholar

^{page 957 note 4} Tibetan Grammar, 3rd. ed., and reprint, pp. 19–20.

^{page 957 note 5} His translation there of mi tsam by “some five people” is brought about by tsam.

^{page 958 note 1} Cf. Tôkyô-Tripiṭaka, vol. iv, p. 401a, 1. 8.

^{page 958 note 2} The Chinese version has in fact ma pien san shih êrh mei About the classifiers in colloquial Chinese, cf. for example Mateer, C. W., A Course Mandarin Lessons, rev. ed. (Shanghai, 1900), Lesson I, p. 2.Google Scholar In literary Chinese the classifier follows the numeral.

^{page 958 note 3} Cf. Das, S. Ch., Tibetan-English Dictionary (Calcutta, 1902), p. 370.Google Scholar

^{page 958 note 4} Mostly sum-mdo. Cf. Mahâvyutpatti (Sakaki), No. 5621.

^{page 958 note 5} Cf. Mahâvyutpatti (Sakaki), No. 5620.

^{page 958 note 6} For further examples of this type see Francke, A. H. in “Addenda” to Jäschke's Grammar (Reprint, Berlin, 1929), p. 126.Google Scholar

^{page 958 note 7} Cf. Jäschke, , Grammar, p. 32.Google Scholar

^{page 959 note 1} Cf. German “(ein) Stücker zehn, hundert, tausend”. As is well known, also the words for two and three are found without a prefix before these numbers. Cf. Foucaux, Grammaire, pp. 40–1Google Scholar.

^{page 959 note 2} Also may be found.

^{page 959 note 3} Dictionnaire Thibétain-Latin-Français par les Missionnaires Catholiques du Thibet (Hong Kong, 1899), p. 200: aliqui, aliquot; quelques, plusieurs. Also Baeot, J. in the “Index des particules et locutions”, included in his edition of Les Ślokas Grammaticaux de Thonmi Sambhoṭa (Paris, 1928), p. 95,Google Scholar renders ẖgaẖ by “beaucoup, plusieurs”.

^{page 959 note 4} S. Ch. Das' Tibetan-English Dictionary incorporates neither ẖgaẖ nor ẖgaẖ .

^{page 959 note 5} In the case of ẖgaẖ, the use as a singular can already be inferred from the phrase ẖgaẖ followed by a negation, correctly rendered by Jäschke as “no, no one, not any, none”. Bacot, loc. cit., p. 95, translates ẖgaẖ, ẖgaẖ tsam by “peu”. I note from the Tibetan version corresponding to the Divyâvadâna (ed. by Cowell, E. B. and Neil, R. A., Cambridge, 1886, p. 34, 23/4), KanjurGoogle Scholar, Narthang print, ẖDul, Kha 46 A 7, 46 B 1: … khyed cog gis baẖam/thos pa ẖgaẖ yod dam (“Gentlemen, have you seen or heard of anyone”: bhavanto'sti yushmâbhir dṛishṭaḥ śruto vâ From the Tibetan version of the Karmaśataka (Kanjur, mDo, Sa 84 A 4):bdag gis gdul bar ẖgyur ba ẖgaẖ Ita ci yod (“Is there then anyone whom I have converted ?”). Examples for ẖgaẖ used as a singular: Saddharmapuṇḍar7îika-Sûtra (ed. by U. Wogihara and C. Tsuchida, Tôkyô, 1934, p. 362, 20/1 = Kanjur, Narthang print, mDo, Ja 256 A 4/5): de na sems can gcig ẖgaẖ gis spyan ras gzigs (5) phyng la bos na (“if there is anybody invoking Avalokiteśvara”: tasmiṃś ca kaś-cid evaikaḥ sattvaḥ syâd yo ' … âkrandam kuryât. (Schmidt's edition, p. 128, 11 = Kanjur, Narthang print, mDo, Sa 305 B 7, 306 A 1): gal te ma ẖgaẖ rnyed na (“If you find anything” (lit. “any kind”). Vinayavastu (Kanjur, Narthang print, ẖDul, Kha 511 A 3: Bris pas dgos pa ẖgaẖ gi phyir gi pa gnyis la btugs pa las (“Likhita touching feet while asking him for something” (lit.“on account of some need”). For the Chinese version see Tôkyô-Tripiṭaka, vol. xxiv, p. 77 C 19; for the reconstruction of the names cf. also Nishio, R. in Kokuyaku-issaikyô, Ritsu-bu, vol. xxiii (1933), p. 279Google Scholar, notes 25–6. Also the etymology of ḥgaḥ would hardly justify an exclusively plural usage. ḥgaḥ is evidently a prefixed derivative of ga which is itself the etymon of and synonymous with the interrogative-indefinite pronoun .

^{page 960 note 1} Kanjur, Narthang print, ẖDul, Kha 34 A 2: ẖbab chu ẖtsheẖu dag gi ẖgram žig tu (“when we went to the benches of the rivers, lakes, and ponds”). Ibid., 411 B 1: de nas mi gis …thos/ (2)… / …. des smras pa / / (“Thereafter some people (first mi hearing and seeing … asked somebody (second mi )… he said: ‘Gentlemen, do you not know …. ? ’ ” ).

^{page 960 note 2} The Tibetan-Mongolian edition of the , the Bod kyi skad las gsar gi brdaẖ khyad par ston pa legs par (see Schmidt- Boehtlingk's Verzeichnis, Tibetan-Mongolian prints, No. 43, Bull. Hist.-phil. Acad. St. Petersb., T. iv (1848), cols. 124–5) has on page 3 A 1 the remark: gyur cig gu lia buẖi gu ni smon paẖi tshig.

^{page 960 note 3} Karmaśataka, Kanjur, Narthang print, mDo, Sa 7 B 4: sems can ẖdi Ita bu gyi tshe rabs dran la / tshigs smra / bcom gtam ẖ gtam zer ba su gu snyam mo (“They thought, ‘Who is this being that remembers his previous existence, speaks and converses with the Bhagavat ?’ ”). Ibid., 24 A 3: … de nas deẖi mas bsams pa / ci ẖdi (4) bya ba byas pa yin nam / ẖon te so soẖi skye bo yin gu / gal te so soẖi skye bo yin nani … (Then his wife thought: “Can he do (extraordinary) things or is he a common man, If he is a common man …’ ”). Ibid., 26 A 1 de nas nag pa chen gnas kyis bsams pa / ẖod ẖdi suẖi yin gu / bcom Idan ẖdas kho naẖi gor ma chag na / bcom Idan ẖdas ga la te / bdag la gu snyam pa (“Then (cf. Feer, Journ. As., 1901 (Janv.-Juin), p. 436) thought: “From whom does this light emanate ? (lit. whose is this light ?). As it surely comes from the Bhagavat himself, may the Bhagavat remember me where(ever) he be!’”). 83 A 5 khyim bdag des bsams pa / thabs ji (6) Itar byas na / khyeẖu ẖdiẖi tshe bar ẖgyur gu snyam nas(“The householder thought: ‘What can I do to ensure a long life to this boy (lit. If I do in what manner, this boy's life (is) turning into a long one)’ ’).Ibid. 143 B 3 ẖjig rten na mi ẖjig paẖi chos ẖjig par mi ẖgyur ba ẖgaẖ lta yod dam gu snyam mo (“he thought: 'Is there in the world then (lta) any (?ga?, see here above, p. 959, n. 5) imperishable doctrine which will not perish ? ’ ”).

^{page 961 note 1} Cf. Gardiner, A. H., The Theory of Speech and Language (Oxford, 1932), pp.311–12Google Scholar “Almost everywhere there may be observed a tendency to replace the imperative by other forms, or at least to mitigate its peremptoriness by the addition of some courteous word or phrase.” I may add that Franeke, A. H. (Addenda to the repr. of Jäschke's Grammar, 1929, p. 154Google Scholar) compared the German einmal after the imperative. His suggestion is of course, linked up with the idea that cig means primarily “one”.

^{page 961 note 2} It would seem as if originally the limited extent requested for an action could be contrasted by a sweeping prohibition. This, at least, would account for the appearance of cog ge < *cog ge cog in the examples referred to in note 2 on page 955. Cf. English at all with the negative imperative. (For the word-order, see below.) The usual appearance of cig with the negative imperative can, of course, also well be reconciled with the meaning “piece”: ma zer cig “do not say a bit”. (I do not discuss here the difference in stem of the negative imperative.)

^{page 962 note 1} The enclisis is again borne out by the phonetic changes; cf. here above, p. 954.

^{page 962 note 2} Cf., e.g., French tue-le! side by side with je le tue.

^{page 962 note 3} See also here below, p. 974.

^{page 962 note 4} Cf., e.g., Buck, C. D., Comparative Grammar of Greek and Latin (2nd Impr., Chicago, 1937), p. 134.Google Scholar

^{page 962 note 5} Cf. A.Walde, Lateinisches etymologisches Wörterbuch, 3rd ed., by Hofmann, J. B. (Heidelberg, 1938), p. 501.Google Scholar The Tibetan series includes also rdzu-ba “to give a deceptive appearance” and other derivatives which will be discussed on another occasion rdzu-ba obviously recalls a similar meaning of Latin fingo.

^{page 963 note 1} For the corresponding Chinese passage see Tôkyô-Tripiṭaka, vol. xxiv, p. 47 C 19.Google Scholar

^{page 963 note 2} Observe also the addition of par after cig. For a similar function of pa after numerals, cf. Jäschke, , Grammar, pp. 32–3Google Scholar. A further passage with nyin re (= Sanskrit dine dine) is to be found in the Tibetan version corresponding to Divyāvadana ed. Cowell and Neil), p. 540,1. 15: Kanjur, ẖDul, Kha 430 B 2.

^{page 963 note 3} Also cig occurs as indefinite pronoun. Laufer, , Roman e. Tib. Königin, Leipzig, 1911, p. 57,11Google Scholar. 6/7 mi khyod re dgaḦ ẖdug pa / cig rnyed pa yin nam ci la dgaẖ (“You, man, are very cheerful. Have you found anything ? or why are you (so) cheerful ?”).

^{page 963 note 4} For example, in the frequent sentence about the Buddha's silently accepting invitation, included in the Mahâvyutpatti (ed. Sakaki, No. 6451): (tûṣṇîm bhâvenâdhivâsayati).

^{page 963 note 5} Laufer, , Tib. Kōnigin, p. 58, 17Google Scholar: ci gu thams cad brtags nas mi zer bar / bandes brkus so rgyug cig zer zer nas (“You would not have said anything, if you had examined everything (lit. all little bits), instead of that you kept on saying: ‘The priest stole it, pursue him!’”) The parallel version of the story in the Gzer-myig (ed. by Francke, A. H., Asia Major, vol. iv (1927), p. 500 = MS., Fol. 79 B 1Google Scholar) has, however, ji dgu, which would be “the nine pieces”, dgu being used in a generalizing way (see Jäschke, , Grammar, Repr., Addenda, pp. 113 and 126Google Scholar) and ji being a (correlative) variant of ci (Jäschke, Diet., p. 172).

^{page 964 note 1} See my paper “Has the Chinese Language Parts of Speech ? ” (Transact. Philol. Soc, 1937, p. 106), where this development has been touched upon.

^{page 964 note 2} Linguistic Survey of India, vol. iii, P. i (1909), p. 26.Google Scholar

^{page 964 note 3} Ibid., vol. iii, P. iii (1904), p. 6.

^{page 964 note 4} Epigraphia Birmanica, vol. i (1919), pp. 35–6Google Scholar.

^{page 965 note 1} Cf., e.g., Jäschke, ,Grammer 3rd ed. (and repr.), pp. 46–7 and 60Google Scholar.

^{page 965 note 2} Cf. also Francke, A. H. in Addenda to Jäschke's Grammar (Repr., 1929), p. 156.Google Scholar

^{page 965 note 3} Cf. Jäschke, , Grammar, p. 21Google Scholar. The rules can, of course, be found already in the works of the Tibetan grammarians. I refer to the recent editions and translations by Baeot, J. (Les Ślokas grammaticaux de Thonmi Sambhoṭa (Paris, 1928), pp. 24, etc.Google Scholar), and Schubert, J. (Mitt. d. Sem. f. Orient Sprachen, 1928, I. Abt., p. 43, and Artitms Asiae, 1st Supplement, 1937, p. 46).Google Scholar

^{page 966 note 1} See, e.g., Jäschke, , Dictionary, p. 505Google Scholar. See also his Grammar, p. 75, n. 1, and Baeot, Ślokas grammaticaux, pp. 27–8.

^{page 966 note 2} See Franoke, A. H., Tibetische Handschriftenfunde aus Turfan (Sitzungsberichte d. Preuss. Akad. d. Wiss., Phil.-hist. Kl., 1924, iii, p. 16.Google Scholar

^{page 967 note 1} See my Tibetisch-chinesische Wortgleichungen (Berlin, 1930), pp. 6–7, and p. 6, n. 6.Google Scholar Cf. also Baeot, , Les Ślokas grammaticaux, p. 25, n. 1, p. 45, n. 1, and p. 187.Google Scholar

^{page 967 note 2} I find, however, one example of gyis after final g in the Old-Tibetan texts published by A. H. Francke. See p. 9 of the paper quoted here, p. 966, n. 2. Note that so far there is no example for as an alternative form of '.

^{page 967 note 3} See the remark in the preamble to this paper.

^{page 967 note 4} Cf., for instance, Welton, J., Manual of Logic, vol. i, 2nd ed., 6th Impr. (London, 1922), pp. 198–200Google Scholar, and Keynes, J. N., Studies and Exercises in Formal Logic (London, 1928), pp. 181, etc.Google Scholar

^{page 968 note 1} Grammar, 3rd ed., p. 52.

^{page 968 note 2} Cf., e.g., the first example adduced here in n. 1 on p. 960, or the sentence ẖdiẖi gtan tshigs (“what is the proof of this?”), included in Thonmi Sambhoṭa's Ślokas (ed. Bacot, J., Paris, 1928, p. 41Google Scholar).

^{page 968 note 3} When writing that paper, I failed to realize that ẖo itself is recognized as a demonstrative pronoun, and that only its identity with the final o has to be proved. I wish, however, to point out that Csoma in his Dictionary clearly differentiates between ẖu “this”, and ẖo “that”, though the two pronouns have been mixed up by Jäschke, who quotes Csoma as his authority (Dict., pp. 499 and 500).

^{page 968 note 4} See above, p. 965.

^{page 968 note 5} About the assimilation of h to the final consonant of the preceding word, see here below, p. 969, n. 4.

^{page 969 note 1} Grammar, 3rd. ed., p. 53.

^{page 969 note 2} Gesenius, , Hebrew Grammar, 2nd English edition by Cowley, A. E. (Oxford, 1910), p. 453, § 141, g—a.Google Scholar

^{page 969 note 3} “Addenda” to Jäschke's Grammar, repr., 1929, p. 147.

^{page 969 note 4} The treatment of the initial ẖ here clearly differs from that suggested for ẖi < 'yi (see above, p. 966). But there we had to do with a transient sound, whereas herewe are confronted by the assimilation of a word starting with ẖ See also here later.

^{page 970 note 1} “I think that he will come” < “I think that. He will come.” See also here below, p. 971, n. 1.

^{page 970 note 2} Cf., e.g., the description by Jäsehke, s.v. cog (Dict., p. 142): “When affixed to a word, it must be preceded by the vowel o, the final consonant of the root being at the same time repeated. Affixed to verbs, it seems to convert them into participles: ẖo s so cog la ‘to those arrived’.” On p. 499 Jäschke wrongly inserts ẖo cog under ẖu cog, giving it the meaning “we”.

^{page 970 note 3} See above, p. 955, n. 2, and p. 961, n. 2.

^{page 970 note 4} One must go back to the oldest European dictionary of Tibetan to find a separate entry ẖo cog. It is included in Schroeter's, F. C. G.Dictionary of the Bhotanta or Boutan Language (Serampore, 1826, p. 474)Google Scholar and translated there as “all, the whole”. This is in agreement with a gloss in the brDaẖ gsar gi skor, where it is rendered by ma lus, which stands evidently for ma lus par “entirely” (see p. 2 B of the Tibetan-Mongolian edition of the work, which constitutes the last part of the Dag yig mkhas paẖi byu gnas)

^{page 970 note 5} In addition to the example quoted by Jäschke from the Dza -lun, or that adduced by S. Ch. Das under cog-cig-car (Diet., p. 384), I wish to refer to a number of examples in the , where, however, also de dag thams cad occurs in the same syntactical situation. A passage like kye ma rgyal po chen po rnam thos kyi bu gnod sbyin che bar grags so cog da lhan (104 A) cig tu na(“Oh, if the great Kuvera together with all the famous Yakshas came to see me”, mDo, Sa, p. 103 B 7, Narthang print of the Kanjur) can be compared with bdag cag gi bzaẖ ba / bcaẖ ba sta gon byas pa de dag thams cad … phulo (“I gave away all the food and drink which I had prepared”, Ibid., p. 86 A 2). For byas pa de dag thams cad also byas so cog could be found. I give the following further examples: bu bu mo btsaso cog ẖchi bar ẖgyur ro (“all the sons and daughters that he had died”; Ibid., 83 A 3); de nas deẖi khyim bdag paẖi gos rgyan bskon no cog mi sna bar gyur te (“then all the clothes and the adornment of a householder with which he had dressed (him) vanished”, Ibid., p. 83 B 5). I have only one example for ẖo cog after an adjective: nyin cig du ri dags cog bsad de (“daily killing good game”, Ibid. 140 A 6). As can be gathered from the examples, after a verb ẖo cog plays often the part of what would be the antecedent of a relatival clause. This may throw light on an interesting passage of the gZer Myig (Asia Major, iv (1927), p. 498, 1. 7): slob dpon Ihaẖis na ci ma nyan no chog (or cog). A. H. Francke (Ibid., vol. v (1928), p. 13, translates: “If the divine Master does not speak, I wish to hear no other speech.” I think the passage rather means: “When the divine Master spoke I did not hear all that he said.” That the passage is difficult can also be inferred from the fact that it is changed in the version reproduced in Laufer's, Roman einer Tibetischen Kōnigin (Leipzig, 1911, p. 57, 11. 2–3)Google Scholar [A number of examples for ẖo cog (which I discovered only after this article had gone to press) are included in Schiefher's, A. paper “über Pluralbezeichnungen im Tibetischen” (Mèm. Acad. St. Petersb. xxv (1878), No. I, pp. 15–16).Google Scholar Schiefner also explains the ẖo of ẖo cog as demonstrative pronoun.]

^{page 971 note 1} That kyi is used “when two sentences are in contrast to each other”, has already been pointed out by Franeke, A. H. in the “Addenda” to Jäschke's Grammar (Repr., 1929, p. 156).Google Scholar But his two examples fail to show the correlative usage of kyi and ẖo. Jäschke combines kyi with kyis both in his Grammar (3rd Ed., pp. 59–60) and in his Dictionary (pp. 6–7) without mentioning ẖo in either place. As for kyi and kyis occurring at the end of sentences and allegedly expressing the future tense (Jäschke, Dict., p. 6; Francke, Addenda, p. 153), I should like to point out that the examples are apparently always followed by a verb of saying, so that kyi and kyis might properly belong to this verb, in the case of kyi in a similar way as this was suggested as a possible explanation for ẖo after a verb (see here above, p. 970, and n. 1), in the ease of kyis in a manner which is reminiscent of ces (see here above, p. 964). It need hardly be said that there is, of course, no connection between gyis and the verb bgyid-pa “to make”, as suggested by A. H. Francke. To his example from the gZer-Myig (As. Major, iv, p. 512, 12; cf. also p. 492, 13, and Laufer, Tib.-Konigin, p. 67, 16) add ẖgroẖis in As. Major, iv, p. 175, 15, which shows that his suggestion cannot be maintained.

^{page 972 note 1} Cf. how Professor Bacot, (Les Ślolcas grammaticaux de Thonmi Sambhoṭa, p. 31, n. 2)Google Scholar expresses a similar point of view concerning the Chinese and Tibetan questions, without, however, discussing the etymology of ẖam. Concerning the Chinese double question I may refer to my paper in Sinica, viii (1933), pp. 216–220.

^{page 973 note 1} Note that we have yu-bu-cag side by side with ẖu-bu-cag (“we”).

^{page 973 note 2} yo “oblique” belongs in a series with labial final: yom “to swing, to totter, to be unsteady”, “to move, incline, tilt”, yob “stirrup”, perhaps also “deceit”. But the yo of yo-byad “to prepare” and “to prepare” may belong in the local series (= “to lay the foundations”).

^{page 973 note 3} According to my paper in H.J.A.S. also ẖo “to come” and ẖo in the meaning “time, turn” would belong here.

^{page 974 note 1} Cf. also German überdies, or obendrein.

^{page 974 note 2} Cf. above (p. 962) the etymology of , originally meaning “to be on top”.

^{page 974 note 3} ye in ye śes, etc., may be another variant, going back to *'ye.

^{page 974 note 4} Note, however, that there is no definite connection between the quality of the vowels and the place, as opposed, for instance, to Manehu, where we have the light e for wesimbi “to rise”, and the dark a for wasimbi “to descend”, etc.

^{page 974 note 5} Cf. with de the last but one note about ye.

^{page 975 note 1} Observe, however, the shortening of ma in -ma kho > mkho. The latter spelling can be observed twice in one line of the Narthang print of the (mDo, Sa, p 201 A 7).