^{page 204 note 1} Grammatik der Prākrit-sprachen, passim.

^{page 204 note 2} The only article by Jacobi that I actually have before me is that in ZDMG. xlvii, pp. 574 ffGoogle Scholar. It is from this and from antagonistic remarks in Pischel's Grammar, and from conversation with Professor Jacobi himself, that I have gathered what his views on the subject are. I must further apologize for the sad incompleteness of this article with regard to references. But in the case of comparative philology, unless I happen to possess a particular book myself, there is little hope that I shall find access to it in India. For further insufficiencies I must plead the exigencies of military service, which make a return to England and libraries a matter of doubt.

^{page 204 note 3} “On the Phonology of the Indo-Aryan Vernaculars”: ZDMG. xlix, pp. 393 ff.; L, pp. I ff.Google Scholar

^{page 204 note 4} La Formation de la langue marathe, §§ 32 ff.

^{page 204 note 5} Op. cit., § 36.

^{page 205 note 1} See Wackernagel, , Altindische Grammatik, i, §§ 243 ff.Google Scholar

^{page 206 note 1} Wackernagel, , op. cit., i, § 252d.Google Scholar

^{page 207 note 1} For all this see Wackernagel, loc. cit.; Macdonell, , Vedic Grammar, §§ 83 ff.Google Scholar

^{page 207 note 2} Pischel, , Grammatik der Prākrit-sprachen, § 46.Google Scholar

^{page 207 note 3} Op. cit., § 349.

^{page 207 note 4} Op. cit., § 80.

^{page 207 note 5} Op. cit., § 81.

^{page 207 note 6} Op. cit., § 90.

^{page 207 note 7} Op. cit., § 194.

^{page 207 note 8} Op. cit., § 108.

^{page 207 note 9} Op. cit., §§ 101 ff.

^{page 208 note 1} Jespersen, , Lehrbuch der Phonetik, 14. 9.Google Scholar

^{page 208 note 2} Op. cit., 12. 64.

^{page 209 note 1} La Formation de la langue marathe, § 94. It does not seem to me that Bloch's contention that Pischel is wrong in the particular case of deriving Prākrit mukka- from *mukna- is particularly happy in the face of Sanskrit forms like bhagna-, bhugna-, vṛkna-, akna-, etc.; in all there are about seventy such participles, and one at least, akna-, has a second form acita- añcita-. ee Whitney, § 957.

^{page 209 note 2} Loc. cit.

^{page 210 note 1} It must, however, be admitted that Latin similarly doubles consonants after short vowels: acqua < aqua, annima < anima (cf. jitta-< jita-?). But the majority of instances are of words with a long vowel, while words with the short vowel belong to the later period, when the stress was becoming strongly developed. See Sommer, , Handbuch der lateinischen Laut- und Formenlehre, § 160, 2 ff.Google Scholar

^{page 211 note 1} § 46.

^{page 211 note 2} ZDMG. xlvii, p. 247Google Scholar; cf. also his Ausgewählte Erzählungen in Mahārāṣṭrī, § 14, 3. Bloch quotes articles by him: Indogermanische Forschungen, xxxi, p. 219Google Scholar; Kuhn, 's Zeitschrift, xxxv, p. 568Google Scholar; but these I have not been able to see.

^{page 211 note 3} ZDMG. xlix, pp. 393 ff.; L,; pp. 1 ff.Google Scholar

^{page 212 note 1} La Formation de la langue marathe, § 34, a translation of Navalkar, The Student's Marāṭhī Grammar, § 39.

^{page 213 note 1} I have little doubt that Jacobi would agree that this stress developed in Sanskrit when the spoken language had already moved beyond the Sanskrit stage, and that it was introduced into the literary from the spoken language, just as all European countries have modelled their pronunciation of Latin on their own spelling; e.g., Lat. cīvis [ki-wis], pronounced in France [si-vis], in Germany [tsi-vis or tsi-fis], in England, except among a now fortunately increasing body of reformers [saivis]. Similarly, the stress accent has been introduced into the pronunciation of Greek in England and Germany; e.g., δός, modern Greek [ο^|ðος], in England and Germany [^|hɔdɔs].

^{page 213 note 2} This description of the stress in Gujrāti agrees in the main with that given by Taylor, Gujarātī Grammar, § 9, excepting that he makes no mention of an alternative even stress for words with final long syllable.

^{page 216 note 1} Cf. Gauthiot, , La fin de mot en indo-européen, p. 194.Google Scholar

^{page 216 note 2} Cf. Grierson, , Seven Grammars of the Bihārī Language, pt. i, § 36.Google Scholar

^{page 216 note 3} Cf. Ehrlich, , Untersuchungen über die Natur der griechischen Betonung, ch. i.Google Scholar

^{page 217 note 1} See below, §§ 27 ff.

^{page 217 note 2} Most of the evidence which I have available seems to show, as Bloch, op. cit., § 137, says, that Singhalese preserved -m- universally, even in inflected syllables: gamak: grāmaḥ; namanu: namati; ba ba: vyāmaḥ; kami: khādāmi; kamu: khādāmaḥ. There are, however, one or two words which show -v-: navanu beside namanu; and two with -m- for -v-, nama (subst.) beside nava (adj.) “9”; nimenu beside nivenu: nirvāti; ruvanmäli beside ruvanväli (see Geiger, , Literatur und Sprache der SinghalesenGoogle Scholar). In päminenu: prāpnōti, Pāli pāpuṇāti, I see a compromise between pam- (cf. G. pāmvū < prāpn-) and päviṇ- < pāpuṇ-. Cf. G. sāmṇū “dream”, a compromise between *sāmũ < svapnakam and *sivṇũ < siviṇa- ũ Idg. supəno-.

^{page 217 note 3} Bloch, op. cit., § 137, refers to Grierson, , Piśāca Languages of North-Western India, p. 118Google Scholar. Unfortunately I have not the book here.

^{page 217 note 4} e.g. German Gypsy, gāv;< grāma;ḥ; Rumanian Gypsy, gav; tu < Pkt. tumaṁ; kovlo < *kōmalakaḥ.

^{page 217 note 5} For full lists of words see my article in JRAS., 1915, p. 20 ff.Google Scholar

^{page 218 note 1} Fuller etymological lists of Gujrāti are necessary to solve this problem.

^{page 218 note 2} Cf. Jespersen, , Lehrbuch, 7. 32Google Scholar: “Akzent (Druck) ist Energie, intensive Muskeltātigkeit, die nicht an ein einzelnes Organ gebunden ist, sondern der gesamten Artikulation ihr Gepräge gibt. Soll eine starke Silbe ausgesprochen werden, wird in allen Organen die grösste Energie aufgewandt. In den oberen Organen zeigt sich die Energie in einer ausgeprägten Artikulation die alle Lautgegensätze scharf hervorstehen läset.”

^{page 218 note 3} In the “Rules of spelling for the Gujarātī reading series adopted by the Vernacular Text Books Revision Committee”, as quoted by Taylor, , Gujarātī Grammar, pp. viii ffGoogle Scholar., the following rules are laid down:— “vii (1) the i sound in any other syllable except the final or the penultimate is long or short according as the following syllable is short or long, and should be expressed in writing by i (long) or i (short). But before a conjunct consonant the i sound should be always short. viii (1) Except in monosyllabic words like sū, jū, lū, rū, and except in dissyllabic words like , etc. [These are, of course, really monosyllables: dūdh, cūk, mūk, s ḍh.—R. L. T.], the ū sound wherever it occurs is short in all words, and should be expressed by u (short).” In my opinion the committee was still under the influence of tradition when dealing with i and ī, and that the facts for i and ī are the same as for u and ū.

^{page 219 note 1} Phonology of the Indo-Aryan Vernaculars, § 13.

^{page 219 note 2} Above, § 13.

^{page 220 note 1} Cf. Sommer, , Handbuch der lateinischen Laut- und Formenlehre, § 160, 1.Google Scholar

^{page 220 note 2} Lehrbuch, 12. 23.

^{page 221 note 1} Cf. Blooh, , op. cit., § 43.Google Scholar

^{page 221 note 2} Cf. note to § 27.

^{page 221 note 3} Cf. Sommer, , op. cit., § 84, 2.Google Scholar

^{page 222 note 1} Lehrbuch, 14. 72.

^{page 222 note 2} Lehrbuch, 14. 73.

^{page 222 note 3} Below, § 34.

^{page 223 note 1} Above, § 23. Cf. also Prim. germ. *bíriði, bérome, γástimiz, náminiz > germ, bäret bärt, bären, gästen, namen. Kluge, Ungermanisch, § 89.

^{page 223 note 2} See below, § 40. A. iii.

^{page 225 note 1} + secondary unstressed ō > iyo: ^|vaṇiyo: *^|vāṇī (< ^|vānijaḥ); ^|sāthiyo: *^āthī (< ^|svastikaḥ). Similarly + secondary ō > uvo: jal j buvo: j bu ( < jambukaḥ).

^{page 225 note 2} See above, § 26.

^{page 225 note 3} For the position of the stress see below, § 40, A. i.

^{page 225 note 4} In a passage quoted by Maṇilāl Bakōrbhāī Vyās, Jūnī Gnjarātī Bhāṣā ane Jainasāhitya, p. 19.Google Scholar

^{page 226 note 1} Wackernagel, , op. cit., i, § 2406.Google Scholar

^{page 226 note 2} Jespersen, , Lehrbuch, 13. 61Google Scholar ff. It would, however, be unwise to dogmatize finally from this isolated piece of evidence as to the syllabic division in the Sanskrit stage. Apart from the statements of the grammarians, whom we know to have been keen observers, there are other indications (e.g. the assimilation and simplification of consonant groups, see Jespersen, 13. 64) that there was loose connexion generally in the language.

^{page 226 note 3} See above, § 26.

^{page 227 note 1} This may be a loan-form from another dialect, e.g. of Apabhraṁśa bandhimō -mu > b dh -ī (for loss of nasalization see my article JRAS., 1915, p. 29Google Scholar), and then to distinguish it from the absolutive badhī < bandhia, the ending -ē of 3rd plur. was added.

^{page 229 note 1} Cf. Jespersen, , A Modern English Grammar, 5. 58.Google Scholar

^{page 229 note 2} I leave the discussion of this question to a later article.

^{page 230 note 1} Literature und Sprache der Singhalesen, § 4 ff.

^{page 232 note 1} See Bloch, , op. cit., Introduction.Google Scholar

^{page 232 note 2} That the Indo-Germanic tone had become a stress in Primitive Germanic before the changes grouped under Verner's Law took place is plainly shown by Jespersen, , Lehrbuch, 7. 34Google Scholar, and Modern English Grammar, 6. 9Google Scholar. To my mind he completely answers Gauthiot's contention (Mémoires de la société de linguistique de Paris, xi, 193) that the accent was still a tone. Cf. also my article in the Classical Review, 08, 1912.Google Scholar

^{page 233 note 1} Cf. Vendryes, , L'Intensité initiale en latin, § 50Google Scholar; and Hirt, , Indogermanischer Akzent.Google Scholar

^{page 233 note 2} Cf. Kluge, , Urgermanisch, §§ 85–6.Google Scholar

^{page 234 note 1} Cf. Whitney, , §§ 611–732.Google Scholar

^{page 234 note 2} Cf. Pischel, , §§ 492–514.Google Scholar

^{page 234 note 3} Out of 181 Gujrāti verbs 94 are descended from simple thematic stems of the type dṓhati, námati, bándhati; 60 from stems with the formative suffix -aya-, such as pálayati, māráyati, lābháyati; 12 from stems with the suffix -ya-, such as yúdhyatē, truṭyati, mányatē; 15 from past participles, such as údgata-, vṛṣṭá-, labdhá-; 161 are simple verbs, 40 compound. There are no traces of any athematic verb; even ásti has been entirely replaced by *acchati, unless a last trace remains in the negative nahi.

^{page 235 note 1} Cf. Whitney, , §§ 745–9.Google Scholar

^{page 235 note 2} Cf. Whitney, , § 761.Google Scholar

^{page 235 note 3} Cf. my article in JRAS., 1913, p. 300Google Scholar, where a list is given of thirty-three causative verbs used in the Dvāviṇśatyavadānakathā with simple meaning.

^{page 236 note 1} The remarks above, § 27 ff., on ū and ī in Gujrāti apply equally to Marāṭhī.

^{page 238 note 1} I do not agree with Bloch, § 48, that this variety need be the result of dialectical mixing, or that a in this position may represent a long vowel with a different timbre from ā. To me the a of tapṇ , e.g., sounds as short [Λ].

^{page 240 note 1} Cf. the change in English from |autho|rize, etc. (still so spoken in Scotland and North Ireland) to |autho|rize, etc.

^{page 240 note 2} Cf. Whitney, § 1254.

^{page 240 note 3} See above, § 10, v.

^{page 242 note 1} See above, § 54.

^{page 243 note 1} See above, § 42. The subject needs further investigation.

^{page 243 note 2} Whitney, § 1148.

^{page 243 note 3} § 52, 4.

^{page 244 note 1} Whitney, § 1211a.

^{page 244 note 2} Whitney, § 1212.

^{page 244 note 3} Whitney, § 1267a.

^{page 244 note 4} Whitney, § 1270.

^{page 244 note 5} Whitney, § 1230.

^{page 245 note 1} Whitney, §§ 1222, 1227.

^{page 246 note 1} See above, § 64.

^{page 246 note 2} § 48 ff.

^{page 246 note 3} See above, § 23.

^{page 247 note 1} See above, § 55.

^{page 247 note 2} See above, § 54.

^{page 247 note 3} §49.

^{page 248 note 1} § 49, 1. Cf. also my article in JRAS., 1915, p. 23.Google Scholar

^{page 248 note 2} See above, § 54.

^{page 248 note 3} See above, § 65. i.

^{page 248 note 4} See above, § 67, A. i.

^{page 249 note 1} See above, § 64 B.

^{page 250 note 1} See above, § 13.