guages have passed by. The following are examples of the causal in these latter:

fa "write," H.P.B.O.S. f¶“cause to write," (H. Pres. likhátá, Pret. likhaya, Aor. likhae, or likháy, or likháve.)

In Hindi, as in the other languages, the causal of a neuter verb is, in effect, nothing more than an active, as

So also in the case of double verbs given in §§ 20, 21, the active form, with long vowel in the stem syllable, may be regarded as a causal. In fact, it might be said, looking at the matter with reference to meaning, that the modern languages have two ways of forming the causal, one in which the short vowel of the stem is lengthened, the other in which â or some other suffix is added. Looking at it in another way with reference to form, the division which I have adopted commends itself, the forms with a long vowel in the stem being regarded as actives, those with the added syllable as causals. In point of derivation, however, both forms are causals. There is a wonderful, though unconscious, economy in our languages; where Prakrit has more types than one for the same phase of a verb, the modern languages retain them all, but give to each a different meaning. For instance, Prakrit has three types for the passive, one in which the final consonant of the stem is doubled by absorption of the of Skr. as gamyate = gammaï, य्

a second in ta, as gamyate=gamîadi, and a third in ijja, as gamyate=gamijjadi. The first of these types, having lost whatever might remind the speaker of its passive character, has been adopted in the modern languages as the form of the simple neuter verb, the second survives in the Panjabi passive, as mârîdâ=mârîadi, the third in the Sindhi and Marwari passive given in § 25. So, also, it seems to me that the two types of the Prakrit causal have been separately utilized; that which corresponds to the Sanskrit type in aya with long or guņa vowel in the stem, has become in the moderns an active verb, as hârayati hâreï=hâr; troṭayati toṛeïtor; while that which takes the causal is preserved as the ordinary causal of the moderns, as kârayati (karâpayati) = karâvei = karâ.

=

Often, however, both forms exist together, and there is little or no apparent distinction between them; thus from are made both फाटना and फटाना, from हरना are made हारना and हराना, , and so in many other instances.

The causal, properly so called, namely, that with the suffix â, âv, etc., has always a short vowel in the stem syllable, except in a few instances where the stem vowel is vriddhi, in which case it is sometimes retained. Thus in the double verbs the causal suffix may be regarded as added to the neuter form, as in

In such cases, however, we more commonly find the double or passive causal.

Single verbs with a long or guņa stem-vowel have causals with the corresponding short or simple vowel, as in the examples बोलना and बुलाना, जागना and जगाना given above.

Verbs whose simple stems end in a vowel insert a semivowel before the termination of the causal, and change the vowel of the stem, if a, i, or e, into i, if û or o, into u. The semivowel

used is sometimes व or र, but more commonly ल. Thus, लेना

“ to take,” लिवाना “ to cause to take,” but—

In a few cases of stems ending in, or in aspirates, the is optionally inserted, as

A similar method exists in Sindhi, but with instead of,

as is customary with that language, as1

Here the is inserted after the causal suffix, and this was probably the method originally in force in Hindi, for we find in the mediaval poets such words as dikhârnâ "to show," and even in modern colloquial usage baithâlnâ is quite as common

1 Trumpp, Sindhi Grammar, p. 256.

as bithland. Gujarati forms its causal in an analogous way,

After words ending in a vowel, the suffix takes to prevent

This language, like Hindi, also reverses the position of the long vowel of the causal suffix, and uses such forms as dhavarâv, khavarâv, with change of to.

There is nothing remarkable about the Panjabi causal, which is identical with Old Hindi, merely retaining the junction vowel u, as khilâ-u-nâ, dikhâ-u-nâ. In both these languages the old form âu has, in a few instances, changed to o instead of â, as

भिगोना "to wet," from "to be wet."

Bengali and Oriya have only the causal form in â with junction vowel i, as B. karâ-i-te, O. karâ-i-bâ, and use this form in preference to that with the long stem vowel, even in those causals which are, in meaning, simple actives.

There are thus, independently of the stem with the long or guna vowel, which I prefer to treat as an active, two separate systems of forming the causal in the seven languages: one starting from the Prakrit causal in âre, and exhibiting the forms âva, ava, ira, îra, âv, âu, o, â; the other starting, I know

VOL. III.

6

not whence, but probably from a method in use in early Aryan speech, which has only been preserved by the classical language in a few instances, and exhibiting the forms år, åd, âl, râ, lâ. Whether these two forms are connected by an interchange between the two semivowels which must remain for future research. not impossible, and is even, in my opinion, highly probable.

and v, is a problem Such a connexion is

§ 27. The Passive Causal may be also called the double causal. The use of either term depends upon the point of view of the speaker, for whether I say, "I cause Râm to be struck by Shyâm," or, "I cause Shyâm to strike Râm," the idea is the same. As regards form, the term double causal is more appropriate in some languages. In H. and P. this phrase is constructed by adding to the stem H., P., in which we should, I think, recognize the syllable av of the single causal shortened, and another av added to it, thus from sun "hear," comes causal sunâ, "cause to hear," "tell," double causal sunvâ,1 "cause to cause to hear," cause to tell;" here, as sund is from the fuller form sunâv, so sunvâ is from sunâv+âv=sunav+â= sunvâ. This double or passive causal is in use mostly with neuter and active intransitive stems, whose single causal is naturally an active, as बनना “be made,” बनाना “make,” बनवाना cause to be made." Thus they say,

66

99 66

"The fort is being built ; " थवई गढ बनाता “ The architect is building the fort ;” and राजा थवई के द्वारा गढ़ बनवाता “ The king is causing the fort to be built by the architect." In this last sentence, and in all similar phrases, the nature of the construction is such that we can only translate it by the passive causal, we could not render "The king causes the architect to build," etc., by in any other way than by putting "architect" बनवाना

1 Generally, the semivowel in this form is pronounced softly, almost like the English w, so that sunwând would more nearly represent the sound than sunvânâ. The v, however, in all Indian languages is a softer sound than our v.