§ 57. The only participial tenses in the Gipsy verb are those formed from the past participle. This participle is sometimes regularly formed from the modern verbal root, and sometimes, as in the other languages, is an early Tadbhava, perpetuating the type of the Prakrit participle. There are, as in the other languages, three types of this participle ending in (1) to or do, (2) lo, (3) no. Examples of the first type are— The Aorist is formed by adding the terminations of the old substantive verb, thus from lino "taken ". Sing. 1. linom, 2. linân, 3. linâs, Pl. 1. linân, 2. linân, 3. linâs "I took," etc. So from kerdo "done," comes Sing. 1. kerdom, 2. kerdân, 3. kerdâs, Pl. 1. kerdâm, 2. kerdân, 3. kerdâs "I did," etc. And from muklo "abandoned" (Skr. mukta)— Sing. 1. muklom, 2. muklân, 3. muklâs, Pl. 1. muklâm, 2. muklân, 3. muklâs "I left," etc. This proceeding is strictly analogous in principle to the method employed in Sindhi, to which, of all the Indian languages, that of the Gipsies bears the closest relation. The future is formed by prefixing to the present tense the word kâma, Skr. kâm "desire," and thus means "I wish to do," etc. Thus kerâva "I do," kamakerâva "I will do,” i.e. "I wish to do." The prefixed word does not vary for number or person. This method of forming the future is, as Paspati (p. 101) points out, borrowed from modern Greek, in which θέλω contracted to θὲ and θὰ, is used in this way, as θὰ ὑπάγω “I will go." There is nothing strictly analogous to this method in our seven languages, though the futures of the ga and la types are formed on a not very dissimilar principle. CHAPTER IV. THE COMPOUND TENSES. CONTENTS.-§ 58. DEFINITION OF THE COMPOUND TENSES AND AUXILIARY ANCILLARIES. -§ 69. § 58. FURTHER removed from the old synthetical type than either of the preceding classes of tenses is that class which now comes under discussion. It is by means of this class that the seven modern languages, after having rejected the numerous and complicated formations of the Sanskrit verb, have secured for themselves the machinery necessary for the expression of very delicate shades of meaning. So numerous, indeed, are these shades of meaning, and so fine are the distinctions between them, that it is very difficult for a foreigner to catch them. The tenses in question are constructed by adding to the participles already mentioned various tenses of certain auxiliary verbs, and in a few instances by adding these auxiliaries to the simple present, or aorist. The auxiliaries themselves are modern formations capable of being used alone, and are traceable to well-known Sanskrit roots through processes partly Prakritic and partly post-Prakritic. Pali and the Prakrits carry the verbs in question through certain grades of change, and the modern languages either preserve the Prakrit form unchanged, or subject it to further changes of their own, such changes being often governed by laws unknown to the Prakrit stage of development. The roots so employed are √ अस्, √ भू, √स्था, √ या, and another, whose origin is somewhat obscure in Sanskrit, but which appears in Prakrit under the form. It will be necessary first to examine each of these roots and draw out the modern forms to be affiliated to each, after which the tenses formed by them may be arranged in order. § 59. 48. This root means "to be," and is the simple copula like Latin esse (see under sthâ in § 12). Only the present tense can be clearly traced in the modern languages, though there are some detached fragments here and there which may possibly represent other tenses. These will be noted further on. In Sanskrit the root belongs to the second or ad conjugation, in which the terminations are added direct to the root, thus giving rise to various euphonic changes in accordance with the laws of Sandhi. Pali and the scenic Prakrits, in contradiction to their usual practice of employing the bhû type for all roots, retain in this verb the ad type. Omitting the dual, the tense runsSkr. 2. asi, 3. asti. Sing. 1. asmi, 2. stha, 3. santi. 2. asi, 3. atthi. amhi. In Prakrit the initial vowel is often elided as 'mhi, 'mha. These forms, however, belong to the scenic Prakrit, which, as Pischel has shown, is really almost as artificial a language as Sanskrit, and on comparing the corresponding tense in the modern languages, it seems difficult, if not impossible, to derive it from the scenic forms. We are not justified in assuming that the modern tense was derived, according to different phonetic laws, from those which have guided and effected the transformations of other words in these languages. On the contrary, in the absence of a continuous chain of documents exhibiting the gradual changes that have taken place, we have nothing to guide us but the general principles of phonetic evolution, which we have been able to formulate for ourselves from undoubted instances. We have numerous well-established cases in which the Prakrit, followed by the moderns, has conjugated a verb according to the bhû type, though in classical Sanskrit it follows some other conjugation; indeed, it may, I think, be considered as proved that the forms of the bhû conjugation have swallowed up all other conjugational types, just as much as those of the as stem in nouns have driven out all other declensional forms. In this view there would be strong reasons for postulating the existence of a present tense of √ conjugated after the bhû type, thus Pl. 1. asâmaḥ, 2. asatha, 3. asanti. It is only from such a form as this, the existence of which, though I am not aware of any text in which it is found, may fairly be inferred from analogy, that the modern forms can, in accordance with the ordinary laws of development, be derived. Beginning with Sindhi as the most archaic, or nearly so, this tense runs thus Sing. 1. viftet, 2. आहें, आहीं, 3. आहे. 3. wifefa. Now, barring the troublesome superfluity of anunâsikas which the Sindhians have seen fit to bestow on this aorist, the forms are strikingly similar to those of the Sanskrit tense postulated above. The 3 sing. âhe is contracted from ahaï, which, again, is good Prakrit for asati, but it could hardly be deduced from asti, which, as we have seen, naturally results in |