142 
Wisconsin Academy of Sciences, Arts and- Letters. 
language, and which therefore cannot be referred to the Ursprache for 
their origin. Such a formation is the Greek perfect in -xa. No other Indo- 
European language presents any trace of it, and we therefore assume that 
we have to deal with a formation which has developed itself on Greek 
ground by way of association or analogy with some other inflectional form 
of the Greek language. 
There have lately appeared several new theories in explanation of the -xa- 
perfect. The two most important of these have emanated from the school 
of the so-called “ Junggrammatiker,” and have been propounded by the 
heads of the school, one by Prof. Brugmann of Leipsic, the other by Prof. 
Osthoff of Heidelberg. A third is that of Felix Hartmann. 
Before taking up the consideration of these theories, let us state more 
precisely the problem to be solved. The oldest formation of the Greek 
perfect, was unquestionably by the suffix -a, appended to the strong form 
of the root, with the reduplication prefixed, e. g., 
root Xeitc-, perf. Xe-X ont-a. 
root 7t£i$-, perf. Tte-rtoi'S-a. 
root psvy-, perf. Tte-cpEvy-a , earlier ^ ite-cpovy-a. 
The Sanskrit formed its perfect in the same way, and there can be no 
doubt that the formation goes back to the Ursprache , for its origin. But 
by the side of this Greek perfect in -or we find in the oldest monuments 
of the language also a considerable number of perfects in -xa. As time 
goes on this number increases, and in the Attic of the 4th century B. C. 
the -xa perfect is the regular formation for all verbs except stems in 
labial and palatal mutes and a few others, which have retained the old 
-a- formation. 
The problem before us is to explain this -xa; why we have -xa in¬ 
stead of -a. 
BRUGMANN’S THEORY. 
Brugmann’s theory appeared first in Yol. XXV. of Kuhn's Zeitschrift 
fur Vergleichende Sprachforschung, p. 212 ff., 1879. 
As he omits some of the earlier theories we may mention them here. 
An examination of the perfects used by Homer discloses the fact that 
but an extremely small number of -xa- forms occurs. While perfects in 
-a, as TtertoiSa, Ei'Xr/xa, EiXr/pa, xexpaya, etc., are very numerous, the - xa - 
perfects found in the Iliad and Odyssey together number only twenty. As 
examples may be mentioned, ftefir/xa, ftefiXrjxa, SeSvxa , sdrr/xa, 7tepvxa, 
reSr/xa, rsrXr/xa. All the Homeric -%<n-perfects, moreover, are from 
vowel stems. There are no Homeric -%<n-perfects from consonant stems, 
such as the later Attic TtsrtEixa, sdraXxa spSapxa, rticpayxa. 
This circumstance led Thiersch and subsequently Ahrens to explain the 
x as inserted between two vowels for the purpose of avoiding hiatus, so 
that according to this theory fi£(dr/xa would be explained as arising from 
* fie (dr/a, by the insertion of x; so Edrr/xa from an imaginary *£drr/a. As 
