SOME NEW THEORIES OF THE GREEK KA-PERFECT. 
By CHAS. E. BENNETT. 
In spite of all that comparative grammar has done to throw light upon 
the morphology of the different Indo-European languages, there yet re¬ 
main in the individual languages a goodly number of formations without 
correspondence in any other language. As examples of such isolated for¬ 
mations may be cited the Greek passive aorists in -rjv and -27 jv, the so- 
called * 72 - or weak declension of the Germanic languages, and the gerundive 
of the Latin. 
With regard to these and similar formations two theories might at first 
suggest themselves: 
First, that we have to do with primitive Indo-European formations, 
the traces of which have disappeared in the other languages and are pre¬ 
served in one alone, or 
Secondly, that the formation does not go back to the Indo-European 
Ursprache, but is a new creation of the particular language in which it is 
found. 
As a matter of fact we at once discard the theory that any formation of 
the nature in question existed as such in the Indo-European Ursprache . 
Unless we have the combined testimony of at least two groups of lan¬ 
guages as to the primitive character of a formation we shall not be justi¬ 
fied in attributing it to the Ursprache. We must rather view such a for¬ 
mation as due to the specific agency of the particular language in which it 
is found, as produced by way of analogy or association with other forma¬ 
tions of the same language, by the use of inflectional material already at 
hand. 
It is in this way that we explain the origin of the weak or -n- declension 
in the Germanic group. The -n- declension arose simply from a wider ex¬ 
tension and more frequent application of endings which were originally 
peculiar to -72-stems, such as Latin nomen, Greek Saip.Gov, etc. So in Greek 
the so-called 1st and 2nd aorists passive in -r/v and -Srjv were originally 
nothing more than intransitive 2nd aorists active of the -pi form, i. e., 
constructed after the analogy of forms like efiXrjv e6rrjv, etc. The for¬ 
mation in -r]v was the earlier, as in k-rtXdnpv, kdraXrfv , kcpdvrjr, etc., 
while the so-called first aorists in -Sr/v are formed after the analogy of 
those in -rfv, under the influence of such presents as dxs 2 go, 7t\r/%Go, <p2z'2(», 
fiifiddSco, rejus^GO, dx^ojuai, vi j$GQ, pirv^oo, 7tpr/%GD, kv?'/^go, etc. 
The above are cited simply as examples to illustrate the methods by 
which we may expect to explain similar formations peculiar to a single 
