In Kasshian, the copula could only connect nouns. To express a predicate adjective, the prefix sa- was added to an adjective, which converted it into a verb. E.g., the adjective kulas “sad” would become the verb sakulas “be sad”. Thus, as a result, gender agreement only occurred in attributive position.
By Old Ivetsian, the prefix sa- had ceased to be productive. However, adjectives were still unable to be used as copula complements. Instead, one would have to use a noun in the complement for the adjective to modify. E.g., the equivalent of “My sister is tall” would be literally “My sister is a tall woman”. To say “Ivets is large”, you would say “Ivets is a large city”, and so on. The preference was for fairly generic nouns like “woman”, “man”, “city”, “place”, etc., with a strong preference for nouns with the same gender, although in principle any noun could be used (thus, “person” could be used as a complement to “my sister”, but was disfavored since it would be epicene gender while “sister” would be human female)
In the Middle Ivetsian period, the number of possible head nouns in this construction became restricted, and those nouns became phonetically reduced, eventually becoming prefixes on the adjectives.
Finally, during the Transitional Ivetsian period, the distinction between predicate and attributive adjectives was lost. Ordinary adjectives became possible in predicate position. In addition, some of those prefixes that had been used in predicate adjectives were extended to attributive adjectives, effectively expanding the gender system, while other prefixes fell out of use.
Thus, in Modern Ivetsian, adjectives could exist, as adjectives, in both predicate and attributive position.
Most adjectives, anyways. That prefix sa-, while it was no longer productive, did survive in a number of adjectives, mostly those relating to emotions and other mental states and personality traits, which adjectives cannot be used as bare predicates. Thus, to say “My sister is happy”, you would use the verb “be happy” which preserves the old prefix sa-.
Kasshian is part of a larger family known as the Taremba Family. The name is derived from the reconstructed word for “language” in the protolanguage.
Proto-Taremba did not have a gender system, but it did possess a moderate number of classifiers, traditionally reconstructed as a set of 39 classifiers, although there is some uncertainty over the exact number, since some only survive in particular branches, and thus might possibly be later developments of those branches. In addition, of course, there is the possibility that there might’ve been classifiers not preserved by any descendant.
The classifiers agreed with the head noun in case and number. Most of them were clearly related to independent nouns. In some cases, the only difference is that they were unstressed. The classifier for celestial objects, for example, was *era, clearly related to the noun *erá meaning “star” (classical Kasshian wila), in other cases, with phonetic simplification, for example, the classifier for places was *bihu, related to the noun *mbihú (uninhabited place, probably originally generically “place”), or contracted, often with further simplifications, as in *daza, used for beginnings, origins, etc., from *ndázəta (root)
The classifiers are believed to have already become obligatory, used before all nouns. They could also be used as pronouns. The structure of the Proto-Taremba noun phrase is believed to have been (demonstrative) (number) classifier (adjectives) noun. There was no agreement yet, and thus they were not yet genders. Kasshian linguists, however, don’t treat nouns and classifiers as distinct entities, but rather, as different types of a single category.
In some branches, the classifiers remain as classifiers. In the Kasshian branch, however, they evolved into gender-markers.
By Early Old Kasshian, the number of categories had been dramatically reduced, to just 11 in the oldest recorded form (possibly 12 just before that, see below). The forms of the markers had been reduced to monosyllabic forms. In the earliest form of Old Kasshian, they were frequently (but still not obligatorily) used with predicate adjectives, the only form of agreement found in Early Old Kasshian. Later, agreement became obligatory with predicate adjectives and preposed adjectives, and optional with postposed adjectives (which were becoming the most common order, by Classical Kasshian, adjectives could only be postposed, and agreement was obligatory). They had already become prefixes by this point, no longer inflected for case, but number inflection remained. Singular was marked by no suffix, dual by -l(i), plural by -f (except for gender IX). The plural marker was originally a paucal in Proto-Taremba. The Proto-Taremba plural was, for the most part, lost. It survived in a few forms where it had acquired a collective meaning
The eleven genders of Old Kasshian are numbered by an adaptation of the system used for the classical-period dialects (note that the numbering was originally used mostly for comparison between genders; specific names were often used within a dialect). Gender III was a creation of Late Old Kasshian, and thus the number 3 is skipped in the numbering for Early Old Kasshian. There were a few allomorphic variations which, for simplicity’s sake, I’ll omit here
Gender I was marked by te-. It indicated female humans. It was derived from the PT *teba with the same meaning
Gender II was marked by na-. It indicated male humans. It was derived from the PT *nako with the same meaning
Gender IV was marked by ki- and was used for domestic animals, infants, and some culturally-significant animals; it was derived from the PT *kita, solely for domestic animals
Gender V was marked by rə- or pa-, varying by dialect, and was used for all animals not included in IV or XII; it was derived from the PT *rəba, which was originally specifically large animals; in the evolution from PT to OK, it absorbed other, more specific, classifiers
Gender VI was marked by (k)wa-, and was, in OK, a rather limited gender, used for fire, liquids, things that move, and, in some dialects, plants. It was derived from the PT *quha, used for fire and water (itself related to the verb “to burn”, and thus presumably originally *just* fire, with water added by semantic association)
Gender VII was marked by pi-, and was the default gender for inanimates. It was derived from PT *bite, used for irregularly-shaped objects and mass nouns
Gender VIII was marked by on-, and was used for instruments, tools, languages, cultural concepts, beliefs, etc.; it was derived from PT *hona, for instruments and tools
Gender IX was marked by ton-/tok- and was used for gods and other supernatural beings, weather phenomena, celestial objects; it was derived from the PT *tokan, used for deities, supernatural beings, and weather phenomena; its plural was derived from the PT -n(a) plural; dialects that used ton- had tonna- and dialects that used tok- had tonka- or tokən-
Gender X was marked by la- and was used for sacred objects, items used in religious ceremonies, etc.; it was derived from the PT *lase for religious objects, rituals, beliefs
Gender XI was marked by kla- and was used for groups of humans, settlements, peoples, political entities, etc.; it was derived from the PT *kula for groups of humans, settlements, buildings, trade, items used for trade, interpersonal relationships, etc.
Gender XII was marked by ka- and was used for birds; it was derived from the PT *katə with the same meaning. Birds were considered sacred animals to the Kasshi, possibly explaining why that was the only group to retain a specific gender (PT had other specific classifiers, such as *heka for fish, *dake for swarming animals, etc.)
Late Old Kasshian added a new gender, gender III, so- in southern dialects, do- in northern dialects, used as an epicene form for humans, as well as for androgynes. This was added after gender XI had been lost in all dialects and XII from most. The numbering is thus anachronistic in that there was never a stage that had all of I-XII.
Body parts had an unstable assignment. PT had a specific classifier for these, *kira, and its probable that that may have survived to a period just before the earliest written records (some linguists reconstruct a gender XIII, *kra-, for pre-Old Kasshian, others suggest that it may have merged phonetically with either the ki- of gender IV or the ka- of gender XII, explaining their instability as they were “ejected” from the gender it had become homophonous with). In Old Kasshian, they were sometimes placed into specific genders, generally VI or VIII, and sometimes made into “agreeing nouns”, nouns whose gender “agrees with” their owner, thus, gender I if they belonged to a woman, gender II for a man, etc. Body parts remained variable between dialects well past the classical period. Within individual dialects there was generally a degree of stability. Some treated them all the same, either placing them in a specific gender or making them all agreeing, others, such as Classical Kasshian, arbitrarily divided them into those placed into a specific gender (VI for CK) and those that agreed.
Genders XI and XII did not survive in any of the Kasshian dialects by the classical period, hence their being numbered after the others, as they were labelled only after grammarians began investigating earlier stages of the language. In addition, no dialect preserved both IX and X. Late Old Kasshian began to lose the distinction between the two, using them interchangeably. By the classical era, the two were thoroughly merged, and only one prefix was perserved, generally narrowed in meaning to just gods and supernatural beings. Early Kasshian grammarians thus referred to the two together as the “Divine Gender”. The first use of numbering divided the genders into four groups, Human I-III (genders I-III), Animal I-II (IV-V), Inanimate I-III (VI-VIII), and Divine. Later the first three groups were combined and renumbered. It was only later that the Divine gender(s) were added to the numbering scheme, hence their being placed after the other 8. Very few dialects preserved all eight of the non-divine genders, most losing one or two of them. Classical Kasshian, for example, lost gender VIII, and of the divine genders, preserved IX. Gender numbering in CK is thus I-VII and IX.
In Classical Kasshian, gender IV expanded somewhat, absorbing the old gender XII and acquiring a few other animals from gender V, mostly animals that live in or around human settlements and game animals. Gender VI expanded considerably at the expense of gender VII and the old VIII (divided between VI and VII), and absorbing the old genders IX-XII (except for gods/supernatural beings in IX). By Classical Kasshian, gender VI was used for all liquids, gasses, plants, groups of humans (peoples, cities, etc.), political entities (nations, provinces, etc.; but *not* the names of places with a purely geographical meaning, which are gender VII - sometimes a place name can be either gender VI or VII depending on its significance or ones political views; after the former Kingdom of Kalpan was divided between the Kasshi Empire and the Blafu Union, for example, the imperial government referred to it with gender VII, treating it as simply a geographical region divided between two nations, while Kalpanian nationalists insisted on using gender VI, emphasizing that they were a divided nation) insects, precious metals, movable objects (that is, objects in which motion outside of being moved by hand, was crucial to their use; for example, vehicles, arrows, throwing knives [but not knives used for cutting or stabbing], etc.), language, emotions, beliefs, etc. It can be summarized as things that move or grow, either literally or metaphorically, and is sometimes called “pseudo-animate”.
Gender assignments varied significantly between dialects and, to a lesser extent, within individual dialects. Gender VI is a great example. A rather large gender in CK, most other dialects used it much more narrowly, with a meaning closer to the Old Kasshian meaning, or had even lost it entirely, as in the Gitanilan dialect.
Actually, 7-11 are etymologically derived as well.
7 (blanta) and 9 (blamfli) are compounds of the word blan (the original word for "6" - which also shows up in blammandu "78") plus ta and fli, so six-one and six-three
8 (bichi - originally bīchi) and 10 (besshi) contain bi, abbreviated from kabi plus -chi and -sshi meaning "4" and "5", thus "two-four" and "two-five"
11 (duta) was originally tanduta from ta naduta "one from twelve"
