and so on, through all the range of compound tenses. The construction is the Bhâva or impersonal throughout, showing that the form originates from the participle, and is to be literally rendered "by me to be loosed it is, or was," which accounts for the neuter form being used.

§ 53. It is to this place that I would now refer the ba type of the future as used in B. and O., and in the Bhojpuri dialect of Hindi. It has been usual to compare these tenses with the Latin future in bo, as amabo, and the comparison is tempting, but, as I now think, delusive. It rests upon the supposition that the b of the termination represents the substantive verb bhú; but to this there are the seemingly fatal objections that bhû, in its modified form of bhava, had from very early times become ho, losing its labial element altogether, and that the present tense bharâmi, etc., though much changed and worn away in modern times, always retains its characteristic vowel o, sometimes shortened to u or resolved into ua. It is only when an i follows the o, that the combination oë is at times shortened to e, as in O. hebâ=hoïbâ.

Judging by the analogy of the cognate languages, it seems that we ought now to see in the B. and O. future the Skr. participle in tavya, in combination with the present tense of Vas. The forms are (dekh, "see")—

The 3 sing. and 3 pl. of Bhojpuri may be excluded from this

group, as they belong to the sa or ha type of the future (§ 35). The dialect of Riwâ has some forms of the ba type, as 2 sing. मारिबेस, 1 pl. मारिब, मारब, and मारिबे, 2 pl. मारिबा . The dialect of Oudh (Avadhi) has 1 sing. मारबूं, 2 मार बेस, मारवे, 1 pl. मारब, 2 मारबो, and in the old Purbi dialect मारब mârab is used for all three persons of both numbers.

There is thus apparent a general tendency to the use of the ba type of future throughout the eastern area of the Aryan territory in India, and it will be seen by comparing either B. or O. terminations of the future with those of the tenses formed by those languages from the present and past participles respectively, that they are almost, if not quite identical. Thus 0.

As the analogy of the other languages compels us to see in these terminations abraded fragments of the present tense of as, when combined with the present and past participles, the same process of reasoning leads us to see the same element in combination with the future participle, and the 2 sing. of the Riwâ and Avadhi dialects further confirms this view by having preserved, like Marathi, the characteristics of the Skr. 2 sing. asi. There is the same agreement of the final vowel in these three tenses of the B. verb, though it is not so accurately preserved as in the more archaic Oriya. Thus B. dekhib-a does not correspond with dekhit-âm, and dekhil-âm. So, also, B. dekhit-is differs from dekhil-i and dekhib-i. There is, however, sufficient general similarity, and the differences consist

mostly in this, that one tense has preserved a more archaic form than the other, thus dekhitâm preserves Skr. asmi, Pr. amhi, better than dekhiba, dekhitis preserves the s of asi, while dekhili and dekhibi have rejected it. Thus, while the abrasion of the substantive verb has been carried to so great a pitch in these two languages as almost to obliterate all traces of it, yet, from the general analogy of cognate forms, there is little doubt that we have in the ba future the Skr. participle in tavya. The final k in the 3 sing. of B. is a phenomenon for which I have in vain sought an explanation; the most probable one is, I think, that which considers it merely as a tag, or meaningless addition, but why a tag should have been added to this person merely, and not to others also, remains to be explained.

§ 54. The two remaining types of the future may appropriately be introduced here. They are certainly participial tenses, though not participial in the same sense or on the same plan as the other tenses discussed in this chapter; they are, in fact, exactly the reverse. Whereas, in the other participial tenses, it is the modern stem which is the participle, and the ancient verbal additions are a present tense, in the two types of future, which we shall now examine, the modern stem is a present tense, and the ancient verbal addition is a participle.

The first of the two is the ga type. This consists in adding H. sing. m., it ƒ., pl. a m., f., to the aorist, and the same in P. except the pl. f., which ist. In the pl. f. H. also ordinarily drops the anuswâra. This type is only found in H. and P. The tense runs thus (sun "hear"), "I will hear," etc.

If the reader will compare this example with that of the aorist in § 33, it will be at once apparent that this tense is formed by adding the syllables gå, gi, etc., to that tense. Among the Mahomedans of Delhi and other large cities, this form is used even as a present, and one frequently hears such a word as haigâ, for hai "it is." There can be little doubt that this gà is the Skr. p. p. गत, Pr. गदो, गयो, गओ In H., as mentioned in § 48, the p.p. of jânâ "to go," is gayâ m., and this in the medieval poets is often shortened to gâ. The f. is

gat, which easily becomes, so also pl. m. becomes 7. The meaning of the tense is thus, "I have gone (that) I may do," a construction which recalls our English idiom "I am going to do," and French "je vais faire." The participial nature of the affix is shown by its being inflected for gender and number in concord with the agent.

This type seems to be of late origin. It is not much, if at all, known or used by the early writers, who, except when they use the future of the ba type, generally express a future sense by the aorist only. As this method grew by degrees to be felt insufficient, the participle was added to give greater distinctness.

§ 55. The second type is that which has for its characteristic. Among the classical dialects Marathi only employs this form, and there has been much speculation about the Marathi future by those who only looked to the language itself. It had, however, long been known that a future with this type existed in the Marwâri dialect, belonging to the Hindi area, and spoken over a large extent of country in Western Rajputana. The able researches of Kellogg have recently placed us in possession of two more instances of a future of this type spoken in Nepal and by the mountaineers of Garhwâl, and Kumaon in the Himalayas, who are by origin Rajputs from the

VOL. III.

11

plains. Thus, the Marathi future now no longer stands alone, and we are in a position to compare the whole group of futures of the la type (par "fall," mâr "strike," ho "be," sut, sod).

In these dialects the aorist has had added to it a form sing. and pl., which does not appear to be inflected for gender, but has only sing. and pl. masc. So far as it goes, however, it directly corresponds to H. and P. gâ, ge, etc., and like them points to a participial origin. In Marathi the inflectional terminations have been dropped, and in some cases even the la itself. The 1 sing. of the neuter aorist is, and adding to this, we get, which, being difficult to pronounce, has glided into 2. In the 1 pl., however, the सुटेन. has simply been rejected, so that it is the same as the aorist. In the 2 sing. the aorist has, but, as has already been shown, this is a contraction from सुटसि, and सुटसि + ल =

ल