a 
| 
ry. 
1898-99. | THE ORIGIN OF GENDER. 69 
majority of adjectives in Latin or Greek have only two terminations, 
one for the masculine and feminine and another for the neuter, e.g., 
dhoyos-ov or hilaris-e. While these adjectives are not older than those 
belonging to the first class, they seem to follow an older type with 
regard to gender. We noticed already that some adjectives of three 
terminations had evidently but recently evolved the feminine form. We 
have in Homer forms like ?Avudzatos 6dy7 or doxdotos y7* which indicate that 
many Greek adjectives that usually have three endings belonged 
primarily to this class. Many adjectives, which have only two forms to 
begin with, develop a third, always a feminine. So zpégpwy-ov, develops 
a feminine form, zpégpacca, xtwy-ov a feminine zea. Comparatives like 
dyzivwv-ov become in Modern Greek adjectives of three terminations. So 
in Latin Az/aris-e changes to hlarus-a-um, tnermts-e to tmermus-a-um. 
In Greek the genitive dual of the article has soi and not tay for the 
feminine as well as the masculine ; and Biicheler seems a little hasty, 
when he concludes from the inscription hastis puris duobus that hasta 
was once masculine, as it seems far more likely that we have here a 
trace of the older state of affairs, when duodus was used for the feminine 
as well as the masculine. (3) The third class of adjectives are those of - 
one ending for all three genders, e.g., »azap or vetus. Some of these are 
used as substantives as well as adjectives, e.g. uber, fertile, is primarily 
uber, the udder, and vetus is the same word as ¢ros, the year. But many 
of them cannot be accounted for inthis way. Here, too, we have signs 
of a development of a second form; pdzap has at times a feminine 
pdzatoa. The most noteworthy instance of this is in the case of Latin 
comparatives. In classical Latin, mazor is masculine or feminine, sazus 
neuter. But the difference of ending between mazor and mazus has 
nothing to do with the distinction of gender to begin with. J/azos (the 
older form of mzazus) is related to mazor precisely as arbos is related to 
arbor ; that is to say, mazor is the rhotacized form of the older mazos, 
which follows the analogy of the remaining cases, mazoris, mazor?, etc., 
and in the historians of the second century B.C., we find such 
expressions as prior bellum, bellum posterior, showing that their forms 
were not then felt as representing different genders. The change from 
hallec, n., to hallex, m., seems to indicate that the form for the 
nominative of many of these adjectives, eg., fe%, is properly a 
masculine form, which is used for all three genders. 
An examination of the forms for gender of adjectives seems to afford 
reason for the opinion that behind the three-fold system of genders,which 
we find in all Indo-European languages, lies a twofold system that does 
*See Munro, Homeric Grammar, p, 112, for further examples. 
