How should I assign gender to words?

[caters]

I am making a conlang in which the writing system is based off of Arabic but the grammar is not all that similar to Arabic or English(more justification for calling it a conlang if grammar has little similarity to the languages it is based off of).

I plan for it to be agglutinative but not necessarily to the degree that Hungarian has(I mean a sentence spanning word sounds pretty crazy to me but maybe that is just because I grew up speaking English).

Anyway, I thought about it long and hard and decided to assign gender to words of at least these 4 parts of speech:

- Nouns
- Verbs
- Adjectives
- Pronouns

Not all verbs will I assign gender to. For example linking verbs like is or looks will not be assigned gender because they are used to link the subject to the action verb or to adjectives and adverbs.

Pronouns practically already have gender in them. For example "he" is masculine. However "he" could also be neuter, at least in English because it can be used to refer to someone of unknown gender. "it" is clearly neuter.

It does make sense though for all nouns and adjectives to have gender. And again, there are clear examples of where 1 gender makes far more sense. An example of that is the adjective "pregnant" It makes far more sense to say "pregnant" is feminine because male pregnancy, while possible, is extremely rare.

But as for the assigning of gender I figured it would be like this:

- Feminine
- Masculine
- Neuter

Latin had this 3 gender system.

Now I am not so sure I would want inflections from grammatical case. I mean word order is a much simpler way to get across the same ideas as grammatical case. In other words for example instead of having the genitive case you could just use a possession marker on the possessor and either way, you get the same thing across, possession. But the possession marker is way simpler than inflecting every possessable noun for the genitive case.

But how should I decide which words are feminine, which are masculine, and which are neuter? I mean with the pronouns, it is obvious, with some nouns, adjectives, and verbs, it is obvious, but what about the majority of nouns, adjectives, and verbs which are not obvious? Sure, with some I could look at statistics(like for example how many females knit compared to males) but it is likely that for the majority of nouns, adjectives, and verbs, I either won't find statistics or the statistics will be inconclusive(thus me assigning neuter if I just went by statistics). A great example of that would be the word stitch. It really depends on why you are stitching(crafts or medical), thus it would most likely be inconclusive.

So for those words of which statistics won't help, should I look at related words to help me assign gender and if I can't find related words of which the gender is obvious, already assigned, or obvious from statistics, use randomness?

And should I strive for equal # of words in each gender?

My conlang is not naturalistic in the sense that adjectives have an inherent gender and thus do not have to agree with the noun's gender. For example, here is a translation specifying the noun and the adjective and their genders and also specifying pronoun gender:

My NEUT husband MASC is pregnant FEM. I NEUT know that is weird but he MASC is pregnant FEM.
So yeah, adjectives don't have to agree with the noun in gender.

 

[CupofJoe]

If you are going for a non-naturalist language I'd go for a random [or near random] allocation.

 

[DragonOfTheAerie]

It doesn't seem to follow any sensible pattern in real life languages. In Spanish, chair is feminine and coffee is masculine. Go figure.

 

[caters]

Yeah, being arbitrary for which words have which gender is exactly what natural languages do.

While seeing if any languages have their translation of pregnant being masculine or neuter, I noticed that the English word gravid has its roots in Danish where gravid means pregnant both in the sense of full of meaning(a pregnant idea is an idea with lots of meaning) and in the sense of carrying a child in the womb but in English it is often used to mean that a reptile has gotten bigger due to eggs it is going to lay.

But yeah, my approach that I posted(obvious, statistics, related words, randomness) I think is more appropriate for a non naturalistic language.

I mean if I am going as far as an Arabic writing system with non Arabic grammar or adjectives not having to agree with nouns in gender, why not tend more towards statistics and less towards randomness where possible?

 

[elemtilas]

I think one thing we need to understand is that grammatical gender is not the same thing as natural gender. So, "gendered" words like chair or coffee in Spanish don't really mean that chairs or girls or cups of coffee are boys.

Grammatical gender is simply a way that words (particularly nouns) have become organised in a language.

So, when you want to introduce gender into your words, grammatically speaking, what you'll want to do is look for patterns: do a large number of words share a characteristic inflexional pattern? Like in Latin, puellarum, stellarum, sellarum, musicarum, tibiarum vs puerorum, tubulorum, libellorum vs sedilium, canum vs rerum, sperum. Notice how the first group all have an -a- before the ending, the second group all have -o-, the third group has -i- or -/- and the last group has -e-. These are called declensions and are the nominal gender system of Latin. Or like in Spanish, flauta, pluma, computadora vs vaso, gato vs arbol, ordenador. Notice that the first group has -a, the second group has -o and the third group has -/-.

Once you get a good idea of what the patterns are, I would assign those as your gender classes. If you want to associate one or another of them with a social or natural gender as well, that's fine, but you don't really need to. I really wouldn't get into statistics or trying to artificially even things out. Just randomly assigning gender is better, perhaps more organic. I'd really expect one or perhaps two large genders and one or more smaller ones with perhaps several categories with single or few words in each.

Re grammatical case, I can see right away that you've fallen for the "grammatical case is hard" trap! You just said that having a "possession marker" is much simpler. But think about it: is it really not the same thing? Is a floating marker really not the same basic idea as a fixed case marker? We do both in English, and everyone seems to get along just fine!

Re assigning gender to verbs: that's an interesting concept. It will be interesting to see what you do along those lines!