GIAN OR IBERIAN LANGUAGE.) It seems safe, therefore, to conclude, on such analogies as these, that the most primitive form of the Pre-Indo Germanic languages also had a natural rather than a grammatical gender. The question then arises as to the origin of grammatical gender. This problem, one of the most difficult of all those presented by linguistic science, has been answered in several ways; and all theories con cerning it must be regarded, in the present state of linguistic knowledge, as merely tentative. Of the two most plausible, the first isthe one de fended by Brugmann (q.v.). The masculine neuter must be set over against the feminine. In this all scholars of prominence are agreed. It is then to be noted that the nominative plural neuter and the so-called nominative singular feminine are identical in their termination, as Vedic Sanskrit yuga, yokes (Classical Sanskrit yugani), and sena, army (compare, with short ened final syllable, Greek 86 pa, gifts, with xd,pa, land; Latin oppidd, towns, with anima, soul, while Oscan and Old Church Slavic re tain representatives of original in the neuter plural, as Oscan praftfc, things proved, beside via, way, and Old Church Slavic roucha, gar ments, beside nogg, foot). Again, in Greek and Avesta a neuter plural subject takes a singular root, as Ta olKlyzarct the build ings fell, ya, varasaite, what things shall be done. In view of facts like these, the feminine singular is regarded as a collective, originally identical with the neuter plural. This has an interesting and suggestive analogue in Arabic, where the so-called broken plural, which is pre eminently a collective word, takes its verb in the feminine singular, as ja'a rajulun, there came a man, but ja'at rijalun, there came men. The termination of the broken plural is also often identical with that of the feminine singular, as ikhwatun, brothers, from akhun, brother (cf. malikatyn, queen, from malikun, king). This theory, however, is not altogether adequate, and it has been supplemented by such scholars as Wheeler and Jacobi. They have pointed out the influence of the pronoun, whose declension was perhaps once entirely unrelated to that of the noun, upon the noun. Here the origin of gram matical gender seems to lie. It is true that the feminine singular was originally a collective noun, merely differentiated in meaning from the neuter plural, and that it was occasionally con cretized to denote a female being, as in the case of the Greek woman, Bceotian fiavd, San skrit gna (originally 'bearings' in the discrete, then 'bearing' in the abstract, finally 'she who bears' in the concrete). From such instances
many words in were termed feminine by anal ogy (q.v.). On the other hand, the pronoun in all languages expresses natural; not grammati cal, gender, denoting male, female, and sexless. The feminine singular of the pronoun may have terminated originally in -a, like its neuter plural, but independently of it, differing herein from the noun, as already stated. On account of the true feminine termination of the pronoun, the collective noun, which chanced to coincide with it in form, was regarded as feminine, and by ana logical extension a numerous class of `feminines,' some female and others sexless, was evolved. In this way the so-called feminine gender probably arose. The so-called masculine gender was simi lar in development. The neuter originally dif fered from the masculine only in the nominative singular (as Latin servus, servum, slave, but templum, templum, temple), and its plural, ex cept for the collective form in -a (the so-called nominative and accusative, as templa, temples), was merely an extension analogical with the masculine. The neuter seems to have been a passive noun, while the masculine was active, and it was thus originally identical with the so called accusative or objective case (cf. Latin servus currit, the slave runs, but servum coedit, he kills the slave, with templum cadit, the temple falls, and templum eruit, he pulls down the temple). The principle of personification, on which Grimm laid such emphasis, was developed after, not before, grammatical gender. The original independence of natural gender is seen from the so-called epicene nouns, which have but one grammatical gender for both natural ones, as Latin lepus, hare (masculine), vulpes, fox (feminine), German Hese, hare (masculine), Hause, mouse (feminine), which leads to such apparent incongruities as vulpes mascula, male fox, weiblicher Hase, female hare. With the decay of the inflection grammatical gender is gradually disappearing, and the more primitive system of natural gender, so long superseded, is resuming its original position. so that the clas sification of nouns as masculine, feminine, or neuter is being based more and more on sex, and not gender.