346 G. A. Grierson — Suffixes in the Kagmiri Language. [No. 4, 
Apabhrarin^a Genitive PL amhahd , of which it is a much abraded form : 
H. C. iv, 897, allows a non-initial non-conjunct m to become a nasalized 
v. In mh the m is certainly compound, but such compounds with h 
are usually treated as single letters, not lengthening a preceding vowel 
by position, so that mh is treated as the aspirate of m, just as dh is the 
aspirate of d. This is of every day occurrence in the modern vernaculars. 
We may therefore assume a form # avhdhd, from which the transition to 
hvaha, liau, hu and U is simple. 
Second Person. — The Sanskrit Enclitics are, Sg., acc. tvd, dat. 
gen. te ; Pl., acc. dat. gen., vas. The corresponding Prakrit forms are, 
Sg. te; Pl. vo. The Kagmiri obi. form in t speaks for itself. It is 
the Sanskrit Prakrit te. The Kacmirl h is of doubtful origin. Gram¬ 
marians give the Ka^mM suffix as k, and say that, when not the final 
letter of a word, it becomes h. It is, no doubt, true that there is a 
tendency for a medial k to become h, but that would leave the 
existence of the k unaccounted for, nor do I know any Sanskrit or 
Prakrit form of the second personal pronoun from which it could be 
derived. On the other hand, if we take h as the original form, there 
is no difficulty in finding an original in Prakrit, and there is no 
inherent impossibility in a final h becoming k. A final h , unvocalised, 
does regularly become k in Sanskrit, but I must confess that I have 
not met any other example of the change in the modern Indo Aryan 
Vernaculars. Assuming, therefore, that h is the original form, it must be 
derived, not from the Sanskrit enclitics, but from the Prakrit bhe or uyhe 
(Cf. W. Panjabi ni-vhe ), the plural form, corresponding to the Sanskrit 
yuyam. 1 If & is the original form, I am, at present, unable to suggest any 
derivation. We next come to the forms, v§, o , I (direct), el, e (oblique), 
and (Western Panjabi) i and (Kagmiri) y (dative). Trumpp 2 derives 
these from the Skr.-Pr., enclitic te, assuming that the t has been elided, 
and the e has remained. I am, however, unable to admit this deri¬ 
vation as correct. In the first place, it would not account for the v in 
ve ; in the second place, it would be contrary to the analogy of the 
other Pronominal Enclitics. Accordingly to analogy, it is the e not 
the t which should be elided, and the t not the e which should remain. 
This has actually occurred in Kacmirl. It is much simpler to derive 
these forms from the Prakrit blie or uyhe already mentioned, or rather 
from their instrumentals bhehi, ubbhehi , umhehi or uyhehi . 3 The 
plural suffixes of this pronoun, va, ve, o, or u, present, no difficulty. 
They are all plainly derived from the Skr.-Pr. enclitic vas, vd. 
Third. Person. There is in Sanskrit a partially declined enclitic 
pronoun of the second person, having ena for its base. This becomes in 
1 H. C. iii, 91. 2 s. Gr., 225. 8 H. C. iii, 95. 
