Gender category in Indo-European.
 
         § 1. Origin of gender in Proto-Indo-European.
 
It is proved that it would be wrong to mix the categories of gender and sex in Proto-Indo-European noun system. Modern Indo-European languages which have gender at all, usually make these two categories the same: what is female in sex, is female in gender. What has no sex, is neuter - for those languages which use neuter, like Slavic and German.

Proto-Indo-European shows a completely different system of gender. And though linguistic schools argue about the more likely structure of noun genders in the Proto-language, some facts can be stated for sure.

First, Proto-Indo-European had no gender. At that time the language consisted not of morphological items, like nouns, verbs and adjectives, but just of words which acted independently in the sentence. The main grammar means of the language was not the declension and conjugation, but the combination of words, and the word order. Nouns lacked any endings, case suffixes, formants of gender of number.

Later, when the language acquired the ergative structure, where all words should be clearly distinguished between active and inactive (or animate and inanimate) classes. Here, on this stage, nouns first get the declension. Most scientists believe there were two or three noun cases at that time in Proto-Indo-European: ergative case, which denoted the subject and indirect objects, absolutive case, which meant subject and direct objects, and maybe genitive, which could exist in the stage of forming. We must stress that only active, animate nouns could be declined, and nouns denoting things did not have cases at all.

There was still no gender, but it was already shaping. The ergative structure was slowly transforming into another type of the language. When a special case was invented for the direct object (accusative), the language could be already named nominative. Words were already divided into nominal and verbal parts of speech - nominal including modern nouns, adjectives, pronouns, numerals, and verbal uniting verbs and perfect verbs - two quite different sorts of words at that time. While the ergative structure was declining, the nominal part of speech was divided into two ones: nouns, which marked the subject and object, and the adjective marking the attribute of a noun.

This division was logical: people were going to separate attributive and subjective words, those which determine and those which state. Adjectives, nevertheless, could not completely avoid the influence of nouns, and since then they had to follow the noun in case and number.

This was the time for the gender to appear. And here the opinions of leading world linguists do not agree. Some of them try to prove, rather successfully, that there was two genders in Early Proto-Indo-European, and then one of them divided into two. Other argue there were three genders originally. Let us see the arguments of each of the sides, just in order to seem objective.

The first version means the following. Two genders which appeared in the Proto-Indo-European language were invented to separate active and inactive nouns from each other, to divide nouns meaning animate objects (people, sacred animals, deities) from simple inanimate things (trees, ground, weapons). The active gender, or active class of nouns, acquired the ending -s / -os in nominative singular and could be declined according to case and number. The plural number denoted several animate nouns, their real plurality. The inactive class, vice versa, could not be declined, its characteristic ending was -m / -om, and even though it had plural number, it did not mean plurality, but just collectiveness. Later processes of the language development generated a-stems of nouns, stems in long i and u. These three kinds of noun stems sooner or later started to denote feminine nouns - now they were equal to the sex. That is how the active class was divided into masculine and feminine.

There are several solid proofs of such a point of view. In most Indo-European languages many noun stems include both masculine and feminine nouns: like short i-stems (Latin hostis masc, ossis fem). Such nouns, though different in gender, have absolutely the same declension and only adjectives, attributes next to them, can distinct their gender. Even a-stems (Latin femina, Lithuanian motina, Greek gunh), made specially for feminine, include many masculine nouns (Latin nauta, poeta). Such a situation of mixture cannot exist with neuter nouns which do not mix in declension with masculine or feminine. While masculine and feminine nouns in most of stems (except a-) still use -s ending in all classical Indo-European tongues, neuter nouns have -m. This very -m is always repeated in accusative singular together with nominative singular: this never happens to other genders.

Greek and Latin, being typical Indo-European languages, show one more interesting feature, called "two ending adjectives". Ordinary adjectives use all three genders: Latin bonus, bona, bonum; but this happens only to adjectives of o- and a-stems which cooperate, and with all other stems adjectives use only two forms: the same for masculine and feminine, and another for neuter. Such adjectives as Latin fortis, fortis, forte, are called two ending adjectives and also show the main thesis of all who state 2 original genders of Proto-Indo-European: differences between masculine and feminine in Indo-European are not as big, as those between masculine-feminine and neuter.

The other argument in favour of this theory is the situation in Anatolian languages. It was proved that the Anatolian branch, later consisting of Hittite, Luwian and Palaic languages, was the first to move away from the Proto-Indo-European community, and did it even before the Proto-language acquired some of its significant morphological and phonetic traits. So Anatolian preserved some interesting archaic moments which we cannot found in other Indo-European languages. Anatolian languages show no feminine or masculine languages, and even no a-stems, and genders are two - exactly for animate and inanimate nouns. The animate gender used -s' ending (as's'us' - good), and the other one either had -n (derived from Indo-European -m), or no ending at all, subsequently did not decline (as's'u - good (neut.)). This makes us believe that such a system is the Early Indo-European one: Anatolian languages were the only group to preserve the 2-gender structure.

The third proof concerns the plural number. Masculine and feminine nouns and adjective in Indo-European use the ending -es in nominative plural, and -ns in accusative plural. All other cases also have their inflections. As for neuter words, their ending -a is used both in nominative and in accusative. Plural neuter nouns denoted collective unity in Proto-Indo-European, like modern Russian nebesa "skies" which does not mean "several skies" but "sky in a collective meaning". Such nouns in the Proto-language included many words which later will be included into the feminine a-stems - their endings in nominative were the same. For example, Latin aqua (water) is originally a collective neuter noun in plural: "waters". In Greek, ancient Indo-Iranian languages and in Hittite the subject in neuter plural form always use their predicate in singular. This is maybe the most important evidence of the special position of neuter in Proto-Indo-European: it was inactive.

Besides a-stems, Indo-European formed the feminine gender, different from masculine, with one more type of noun stems: long i-stems. Obviously, the suffix -i- meant the possession on archaic stages of the language (Sanskrit rathah "a chariot", rathih "something belonging to a chariot"). Later this meaning could go in two ways: it might begin to mean the possessive genitive case (remember -i in Latin, Venetic and Celtic) and to form the feminine gender, if Indo-Europeans looked at the woman as a possession of the man. That is how long i-stems turned out to be all feminine.

So from that moment, when Anatolian languages already migrated to Asia Minor, when a-stems and i-stems appeared, and the opposition emerged between masculine and feminine nouns, we can speak about three genders in Indo-European: masculine, feminine, neuter.

There are linguistic schools which do not agree to the fact that the Proto-language used two genders, and not three. They say that Anatolian just united masculine and neuter, for its a-stems phonetically coincided with masculine o-stems, so the genders coincided as well. Other explain this unification by the substratum influence. But anyway, the 2-gender version seems much better proven.
 

        § 2. Genders in ancient and modern Indo-European languages.
 
When the Proto-language disappeared, transforming into different groups, the gender system chose its own way in each of them. But the general trend in every Indo-European language was quite clear: the number of genders was going to reduce.

Hellenic languages in their ancient varieties (Ancient Greek dialects) preserves the three-gender structure, and its peculiarities often show the closer connections between masculine and feminine, than between them both and neuter (2-ending adjectives, etc., see above). Ancient Greek represents the classical neuter endings in -n < *-m. In Greek neuter plural subjects have usually a singular predicate: ploia plei (ships move). The New Greek language also shows three genders, and their opposition is strengthened by the extensive use of articles, also declined in three genders.

Italic can be called classical in this meaning as well. Latin shows no sign of reduction of genders, and its neuter plays an important role in the language. The reduction of some final consonants in Umbrian influenced the so-called Popular Latin, which was quickly moving towards the analytism in morphology, lost many endings and so many gender forms coincided with each other, which could not but lead new Romance languages in their majority to the 2-gender opposition with the loss of neuter. In many modern Romance tongues the function of differentiating genders passed from the inflection to the article usage: French has no other signs of genders, but un, une, le, la, some prepositional forms like du - de, and adjective endings gros - grosse. In some languages, however, the process of secondary morphologisation of genders is going on nowadays: Spanish generated new endings fro masculine and feminine: hermano - hermana, cabron - cabra. This interesting feature seems to make us think that the gender development history goes the same way as the case system development (see the Sinusoid Theory for Case development in Indo-European).

Germanic languages are called the most analytic among all modern Indo-European tongues. While all ancient varieties of Germanic (Gothic, Old Norse, Old English) used three cases, the modern languages reduced their number. English and Afrikaans removed genders at all from their morphology: English preserved them only as a "hidden category", which can be seen in personal pronouns he, she, it (for example, ships are always she). German uses both some infections and first of all the article, definite and indefinite (der - die - das, ein - eine). An interesting thing happened in Scandinavian languages: they lost the opposition between masculine and feminine and returned to the ancient system of two noun classes for animate - inanimate nouns. This proves once more that feminine and masculine in fact are similar in the language.

Iranian languages used to have all three genders in ancient tongues (Avestan, Old Persian), but under the influence of analytic trend the system was destroyed or reduced, and many modern Iranian languages do not have genders at all. It can be also connected with the adstratum and substratum languages, which influenced Indo-European structural features. Genders are completely lost also in Armenian.
 
Baltic linguistics shows that Old Prussian had neuter which was not about to disappear at all. But Lithuanian lost it somehow, not very long ago. Some relics take place, however, in adjectives and in pronouns, but in fact now Lithuanian knows only feminine and masculine (see Historical Grammar of Lithuanian).

The most complicated system exists in Slavic languages, where genders were not only preserved, but also developed. For example, in Russian the category of gender united with that of animateness, forming the common system of classes: at all there are 6 classes, three genders with animate - inanimate subgenders in each. See the forms of nominative and accusative plural in Russian: novye doma ("new houses", nom., masc., inanimate), novye doma (acc.); novye direktory ("new directors", nom., masc., animate), novyh direktorov (acc.). The same for feminine and neuter.

In Polish even further complication of gender structure happened: masculine has animate and inanimate forms, and animate has in its turn personal and impersonal forms in plural. So:

dom (nom. sg. inanimate), dom (acc. sg.) "a house"
pies (nom. sg. animate), psa (acc. sg.), psy (nom. pl.), psy (acc.pl.) "a dog"
chlop (nom. sg. animate personal), chlopa (acc. sg.), chlopy (nom. pl.), chlopow (acc. pl.) "a peasant"

According to these processes which independently move on in different groups of the Indo-European family we can make a conclusion that genders do not behave exactly the way noun cases do, i.e. the reduction process is sometimes substituted by the complex changes (Slavic, Spanish). But in general we can state that genders are nowadays less synthetic and less spread, than they used to be in the Proto-Indo-European language.

See in the next issue:

-    How many numbers were there in Proto-Indo-European?
-    What was the counting system and its changes in Indo-European languages?
-    What is the origin of numerals in Indo-European?
 

Next Issue