Some New Theories of the Greek Ka-Perfect. 
153 
Osthoff, however, maintains that the composition of the root with the 
particle antedates the historical period of the Indian language and that we 
have in da-ddu, etc., the legitimate successors of an Indo-European de-do-u. 
Whitney in his Sanskrit Grammar (§ 800 c of the German ed.) cites two 
forms from the Rig Veda of perfects from d-stems in simple a, not -du, 
viz.: papra, jaha ( pra-9 , hd- “follow”), instead of paprau, jahdu. This 
would seem to indicate that roots in - d originally formed the perfect in -a, 
and if that is so it would be natural to see in the forms in -du traces of an 
added -u as Osthoff does. If now, reasons Osthoff, the Sanskrit employs 
the particle -u in a perfect formation by adding it to the 1st and 3rd singu¬ 
lar of long vowel stems, what more natural than that the Greek should do 
the same with Tier? 
Osthoff’s theory then, briefly stated, is that the particle hev in its weak or 
unaccented form hoc, a form actually occurring in several Greek dialects, 
became added to the original perfect forms of long vowel stems, first to 
the 1st and 3rd singular, and that it afterwards extended its range over the 
entire perfect system; and the analogy of the Sanskrit perfects in -du is 
appealed to, to substantiate this view. 
Is this a probable theory? It 3&ems to me that it does have a good deal 
to commend it to our favorable consideration. 
1. As to the amalgation of different words into one word in the indi¬ 
vidual Indo-European languages we need have no hesitation. Instances 
are frequent of the union of an inflected form with a particle, or of two 
inflected forms with each other, as may be seen, e. g., in the Gothic liu- 
ganda-u, nimaida-u (Greek vep oiro), in the French j ’ aimer-ai , Latin veneo, 
(for venumeo), calefacio, fervefacio, etc., and, as I incline to believe, also in 
Sanskrit daddu, etc. As to the form xa, it is only what we should be 
compelled to assume as the weak form of hev, even did we not know of 
its actual existence in seve ral of the Greek dialects. That the form xa 
therefore should combine with the original perfect is not at all unlikely. 
2. As to the exact etymology of hev (xa) I conceive that that really 
concerns the main question very slightly. It is immaterial whether hev 
be the same as the Sanskrit kam or gam, or in fine neither the one nor the 
other. It is certainly an independent word of the Greek language and as 
* such can unite with an already existing inflectional form to create a new 
form. If it really be the same as the Sanskrit gam and has gone through the 
same semasiological development as the German wol so much the better, 
though neither of these facts is necessary to our theory. 
3. To my mind it can hardly be fortuitous that we find the origin of 
the -hoc- perfect in Greek in those very stems, which in Sanskrit we are 
led to look upon as compounded in their perfect formation, with the 
particle - u, viz., the long vowel stems. To be sure we find this -u in the 
Sanskrit perfect only in the 1st and 3rd singular; yet we must bear in 
mind that in Greek also the perfect in -xa was for a long time confined to 
the singular of the indicative, and that practically only in the post-Hom- 
•eric Greek did it extend its range beyond this. 
