66 TRANSACTIONS OF THE CANADIAN INSTITUTE, [Vou VI. 
Glycertum, and in Greek all diminutives that denote females are 
feminine, while many that denote males, e.g., ta:dfov, perpdztov, are neuter. 
So epithets of males are usually masculine and epithets of females 
usually feminine. A common ending for feminine epithets was a, as in 
Theodora and Agatha, and hence began a tendency to make all words 
ending in a feminine. These words further attracted others of similar 
meaning: sz/va, for example, attracted arbor, and all names of trees in 
Latin tended to become feminine, though many of them, like laurus, 
ficus, pinus, malus, pirus, have the masculine ending. So in Greek, 
dyutd, a street, attracts to this gender 00g apa5itos and ‘atpaxds. So that, 
while gender is primarily based on sex, it is soon determined rather by 
the groupings of certain words according to their meanings or endings. 
The gender of a word is at times determined by its meaning as in the 
case’ of arbor ; at times by its ending, as in that of sz/va, while at times 
it wavers, owing to the counter-balancing force of these tendencies, as in 
cypressus, which is sometimes masculine, sometimes feminine. 
This view is a distinct advance on that of Grimm, and seems to givea 
true account of the way in which names of lifeless objects become 
masculine or feminine. But why are personal epithets like MWezd in 
German, or za:dfov in Greek, not feminine or masculine to begin with, if 
gender is primarily based on sex? In Latin, servws is masculine, but 
the legal term mancupium is neuter, as in 40pdzodoy in Greek, and the 
latter seem to represent the older state of things. The answer will 
probably be, that mancupium is primarily “that which is held in the 
” and properly neuter. Weib is from O. H. G. /Vzp—inspiration, 
also a neuter. But according to the theory just stated, words used as 
epithets of males or femaies tend to become masculine or feminine. So 
in Sanskrit dpas, n., is the work, apds is the workman; d7déhma, n., is 
hand, 
worship, m., the priest. In Greek, according to Brugmann, »<4:as, the 
youth, is altered from the older form *veae, youth in the abstract. In 
Latin, it is held, agrzco/a, m., the farmer, is merely a personification of 
the older agricola, ploughing ; and opézo, f., choice, becomes m. when it 
comes to mean the centurion’s assistant. An older answer, still 
commonly given, is that mancupium or adpdzodov were not primarily 
persons but ves or «z7jva7a, and hence were neuter. Is this true of Wezd 
as well? We have every reason for thinking that the primary condition 
of woman is that assigned her by Roman law, ie., she was the property 
of her husband. But if this is the true reason for the gender here, 
eender cannot be based on the natural distinction of sex, but on the 
artificial one of legal status. 
*Historical Syntax I., pp. 89, ff. 
