Some New Theories of the Greek Ka-Perfeet. 14& 
hence have become united in one verb; just as frequently the same thing 
occurs where the different roots are clearly distinct, e. g., 
opaGo, oipojuai, eiSov 
q)spcj, ol'6gd, rjveyxov. 
deSooxa therefore is de-Saw-a, i. e., a regularly formed perfect from the- 
root 8gox- with the ending -a, the regular ending of the perfect both in 
Greek and Sanskrit, as seen in 7t£itoi%a, XeXonta, Tticpevya and numer¬ 
ous other similar forms; and it is from this one form that Brugmann con¬ 
ceives the entire category of -xa- perfects to have arisen, and in the fol¬ 
lowing way. After eSooxa and deSooxa had become part and parcel of 
the verbal system of 8l-8gq-jui they were mentally referred to the root 8go- 
also, and the ending -xa was felt as a tense-suffix. 
As soon as it was felt as such it began to be employed in the formation 
of other perfects with long vowel roots, e. g., s6n/xa {6rp-), f5s(3pxa 
(fii)-), Ttecpvxa ( cpv -). 
Brugmann thinks he discovers a particular reason why the -xa- forma¬ 
tion was so readily adopted by the long vowel stems. The original inflec¬ 
tion of the singular of the perfect indicative of the root 8 go-, for example* 
must have been: 
* 8s- 8 go- a 
* 8s- 8 gj- ? (-Sd) 
* 8£- 8go- £ 
By contraction this must have early given: 
* 8£8GO 
* 8£8gd$ 
* 8 £8 go 
These forms were felt to lack the distinctive perfect character and were 
likewise identical in the 1st and 3rd singular. Hence it was that the lan¬ 
guage readily availed itself of such formations as £dr?-/xa, fisfirjxa- 
Ttscpvxa. It is at about this stage in the history of the perfect that th 
Homeric poems belong. The -^-perfects in Homer are almost exclusively 
from long vowel roots; but short vowel roots are also just beginning to 
adopt the -xa- formation, e. g., fi£0ir/xa, 8£8£i7tvr/xa, etc. Still later and 
subsequent to Homer are the formations from consonant stems, such as 
£6raXxa, Gcp^apxa, 7t£7t£ixa, pyy£Xxa, rG^avyaxa. 
The theory that one form like 8£8 coua should have been the type that 
has called hundreds of others into existence seems at first, perhaps, too 
remarkable to be lightly credited; yet similar phenomena are well attested 
for other languages. Thus in Old Bulgarian the first singular of the verb 
in the present indicative ended almost always in a— original Indoger- 
_ r 
manic o. Only four verbs had the ending -mi, viz:jesmi, vemi, dami, jami. 
Yet these four verbs have been the type from which all verbs in the modern 
Slovenian and Servian languages form their 1st person singular present 
indicative in -m. So we have hundreds of forms from four alone. So also 
10—A. & L. 
