Gender in conlangs

[Ahzoh]

Like I said, "girl" for whatever reason was in a less animate gender. That doesn't seem to have bothered the language speakers. Of course, your scenario is just as valid. (In English, try referring to a person as "it".)

But the way the language was portrayed, the key of the gender change was "girl", which caused the other nouns to be changed by analogy.

Well, maybe, since Vrkhazhian neuter is primarily for unknown/mixed sex rather than objects (thus objects are simply classified neuter because you can't find out the sex, though some inanimate nouns have vestigial gender marking, such as "moon" and "sun"), using it on known persons would indicate affection if one was pretending that they have no relationship with the person... perhaps in some ironic fashion.

[Sḿtuval]

Ythnandosian has a common/neuter system (from PYT's masculine/feminine/neuter system) but gender isn't all that important in that language anymore. Pronouns don't mark gender at all (but this feature is shared with all Ydtobogȧntiaky languages).

Gender is marked on adjectives, articles, and demonstratives, but I tend to avoid articles and demonstratives when translating stuff. Some adjectives don't even mark gender at all, and those that do normally don't when used substantively (i.e. "it is red", "i don't like the small ones")

Ythnandosian is the only one of six Kauzic languages to actually continue marking gender though. Languages in that subgroup have much simpler declension paradigms than other Ydtobogȧntiaky languages.

[k1234567890y]

I am not a fan of sex-based grammatical gender, maybe most conlangers are not either, as it seems that conlangers are more likely to have a leftish or liberal(in the US sense) ideology which can make them doubt traditional or conventional views about gender and sex, however, I do have at least one a priori conlang that has sex-based gender.

[Isoraqathedh]

My grammatical genders can get pretty wild.

Hesmai Iok uses aspects on top of genders, which is a plist with 55 set elements. The 55 in question are really just taken more or less wholesale from Thaumcraft's aspect system, but in Hesmai Iok they perform completely different tasks, from lexicographical sorting to divination and numerology.

A word's aspect is determined by what it names, so expect gears and conveyor belts to have a lot of motus, textiles and fabrics to have fabrico, and sand to contain a little terra. The numbers are a bit aarbitrary but they're not completely so.

HI's genders are split to nouns and verbs, and both have a role to play in the language itself. They are creatively named P, Q, R and S, and the rule is "no sentence may have any head nouns share the same gender as the head verb." There's more detail, including how you can fudge that by using alternate methods to derive a gender from the word and various lexical hacks that one can use to get around the limitation, but that's only seen in some random posts I make, so I'll dig it up if anyone asks.

---

On the other hand, Uvbraot's gender system is a much milder system only having ten genders, split into half-genders of two and five: pick one of {masculine (written Aoʃ), feminine (written Yyʃ)}, and one of {Stone, Sea, Sky, SPAAACE, Society} (the alliteration is preserved from the original). Both halves strongly influence word order but very little else. Unlike Hesmai Iok, which prefers gender dissonance, Uvbraot prefers gender consonance, putting together words that share the same half-gender.

[Adarain]

Semũr has 14 noun classes. Word order and morphosyntactic alignment depend on them, verbs take a different ending based on them, and most have different pronouns.

[sangi39]

My three main conlangs, Proto-Sirdic, Proto-Skawlas and Lesi Kirra all have some kind of noun-class/agreement system, of varying sorts.

In Lesi Kirra, nouns are divided into 23 semantic groups marked out by distinct "generic" words, but there's no agreement between nouns and associated adjectives, verbs or the like. Nouns have to appear alongside a generic word, but the generic words can appear as stand-alone words. For example:

Kixo xu k'äqä yuga yeexa. Pä yuga kela deero.
AUX.STAT.PST.EGO 1SG to.see [DOMESTICATE] dog. AUX.STAT [DOMESTICATE] to.sit red.
I saw the dog. It was black.

In the second sentence, the generic word yuga appears as a stand-alone word, referring to yeexa, "dog", which itself cannot appear without the word yuga.

The only form of agreement you'll see in Lesi Kirra between verbs and their arguments is between the auxiliaries and a first person singular pronoun, e.g. kixo xu k'äqä yuga yeexa (I saw the dog) vs. päpä ka k'äqä yuga yeexa (you saw the dog).



In Proto-Skawlas, nouns aren't really marked any differently from each other in any form, but there is an animacy distinction which affects verb conjugation. In a basic sentence, the word order is SOV, but the verb agrees with "most animate" noun, regardless of whether that noun is the subject or the object. If the subject is less animate than the object, then an inverse suffix must be used:

haw dwal hịwanị
1SG.NOM 2SG-ACC move.away.from.something-1SG
I move away from you

lak kirḷ hịwammëj
2SG.NOM 1SG-ACC move.away.from.something-INV-1SG
You move away from me



Proto-Sirdic has a slightly more "familiar" noun class system, with six noun classes which affect nominal declension, adjectival agreement and verbal agreement in relation. The six noun classes are:

1) masculine, animate
2) feminine, animate
3) neuter, animate
4) masculine, inanimate
5) feminine, inanimate
6) neuter, inanimate

Adjectives agree with the gender of their associated nouns, regardless of animacy while number agreement (on verbs and adjectives) is determined by animacy and the presence/lack of a "plural quanitifier", all of which is described here.

There are semantic reasons behind the assignment of nouns into each noun class. Nouns referring to adult human males are usually masculine and animate while nouns referring to adult human females are feminine and animate. Body parts and tools associated with men and women are also associated with masculine and feminine noun classes on a similar basis, although it's less clear-cut than with humans and animals.



One conlang I'm planning on working on at some point, Proto-Mesit, lacks noun classes on any semantic basis, but animacy is a feature of a number of conlangs within the world I'm working on, as are gender and other noun class systems.

[Sights]

I have a conlang in which animates and inanimates mark (singulative) number differently, and each "gender" has a specific set of determiners. Does this mean my conlang has gender? Or is this too marginal to be considered a gender system?

[k1234567890y]

I have a conlang in which animates and inanimates mark (singulative) number differently, and each "gender" has a specific set of determiners. Does this mean my conlang has gender? Or is this too marginal to be considered a gender system?

I think your conlang has gender or a classification system.

Proto-Sirdic has a slightly more "familiar" noun class system, with six noun classes which affect nominal declension, adjectival agreement and verbal agreement in relation. The six noun classes are:

1) masculine, animate
2) feminine, animate
3) neuter, animate
4) masculine, inanimate
5) feminine, inanimate
6) neuter, inanimate

Adjectives agree with the gender of their associated nouns, regardless of animacy while number agreement (on verbs and adjectives) is determined by animacy and the presence/lack of a "plural quanitifier", all of which is described here.

There are semantic reasons behind the assignment of nouns into each noun class. Nouns referring to adult human males are usually masculine and animate while nouns referring to adult human females are feminine and animate. Body parts and tools associated with men and women are also associated with masculine and feminine noun classes on a similar basis, although it's less clear-cut than with humans and animals.

that reminds me of the system seen in some slavic languages

[Aleks]

Gender is useful the same way obviation is, and the more genders you have, the more useful it is. In a proximate/obviate system without gender, "the hero saw a girl rest on a tree before a rock hit it" would be ambiguous between the rock hitting the girl or the tree because both of those would be obviate in the narrative about the hero, but even in English with its limited gender system it's not.

It's quite apparent the rock hit the tree. Why would you be confused? If it was the girl it would have been "the hero saw a girl rest on a tree before a rock hit her".

Anyways on topic I felt genders didn't fit well in mine because what do you do with transexual people? Not knowing what to refer to them as, why not just use animate and inanimate genders? I feel this makes it better for conversing with people and makes it logical. The only thing I can think of that messes with this is AI and androids. With mine I have animate, inanimate, animate plural, and inanimate plural.

If a synthetic entity is self aware or shows signs of consciousness then they are treated as animate same with some kind of flora. I do have alternate 3rd person singular pronouns for synthetics that don't have a defined gender or if someone's gender is ambiguous. These are meant to be a polite version and don't really mean "it" which can be seen as offensive to use to refer to someone.

[sangi39]

I have a conlang in which animates and inanimates mark (singulative) number differently, and each "gender" has a specific set of determiners. Does this mean my conlang has gender? Or is this too marginal to be considered a gender system?

I think your conlang has gender or a classification system.

Proto-Sirdic has a slightly more "familiar" noun class system, with six noun classes which affect nominal declension, adjectival agreement and verbal agreement in relation. The six noun classes are:

1) masculine, animate
2) feminine, animate
3) neuter, animate
4) masculine, inanimate
5) feminine, inanimate
6) neuter, inanimate

Adjectives agree with the gender of their associated nouns, regardless of animacy while number agreement (on verbs and adjectives) is determined by animacy and the presence/lack of a "plural quanitifier", all of which is described here.

There are semantic reasons behind the assignment of nouns into each noun class. Nouns referring to adult human males are usually masculine and animate while nouns referring to adult human females are feminine and animate. Body parts and tools associated with men and women are also associated with masculine and feminine noun classes on a similar basis, although it's less clear-cut than with humans and animals.

that reminds me of the system seen in some slavic languages

That, I think, was part of the inspiration behind the system, which led to I read about a language from Africa, an Afro-Asiatic language if I recall correctly, although it could have belonged to another family altogether, which has a masculine-feminine gender system with differential number marking/agreement on the basis of animacy and gender, although there was no inflectional difference on the basis of animacy, which does occur in Proto-Sirdic*. Annoyingly, I accidentally removed the link to the GoogleBooks thingy, so I'll have to dig that one up. It was quite an interesting read.

*Different inflectional paradigms are a feature of Proto-Sirdic inspired by Proto-Indo-European. The paradigms of each gender are fairly similar, but animate nouns take a thematic vowel and distinguish between a nominative and accusative case while inanimate nouns do not.

[Dormouse559]

It's quite apparent the rock hit the tree. Why would you be confused? If it was the girl it would have been "the hero saw a girl rest on a tree before a rock hit her".

They were talking about a language with no gender, but with obviation, so imagine we used "xe/xem" for all third-person obviates. "The hero saw a girl rest on a tree before a rock hit xem." You can't know whether the girl or tree got hit because they're both obviate, and no there's no gender contrast.

Anyways on topic I felt genders didn't fit well in mine because what do you do with transexual people? Not knowing what to refer to them as, why not just use animate and inanimate genders? I feel this makes it better for conversing with people and makes it logical. The only thing I can think of that messes with this is AI and androids. With mine I have animate, inanimate, animate plural, and inanimate plural.

I don't know the full backstory of your conlang, but natural languages evolved long before transsexuality as we conceive of it today existed. And no natural language ever has developed with androids in mind. And it isn't necessarily naturalistic for a conlang to reflect your modern values or anyone's values really. (But all this is assuming naturalism is your goal.)

[sangi39]

Actually, just thinking about this given the recent posts, how do natural languages with gender (either as just a noun class system, or something that triggers various kinds of agreement) handle members of a "third sex", like the hijras or fa'afafine, which (from what I've read) are a "recognised" (although not always necessarily well-treated) part of that society?

I haven't had any thoughts regarding how people who identify as another gender might be treated within any of the conlangs I've been working on, or even how they might be perceived within the respective cultures identified with those languages*, but it would be interesting to get a look at some more natlang examples.

*Well, other than in Proto-Skawlas which lacks much more than a minimal animacy system, so in terms of pronouns, they probably wouldn't be referred to any differently than anyone else. Proto-Sirdic and Lesi Kirra, though, have distinct classes for men and women, but only Lesi Kirra has an "unspecified" class for "person" which covers children, groups of people of mixed sex, personifications and the like, but the pronoun system lacks gender, so I guess they'd fit in fairly easily. Seems Proto-Sirdic would be the more "troublesome" of the three

[k1234567890y]

*Different inflectional paradigms are a feature of Proto-Sirdic inspired by Proto-Indo-European. The paradigms of each gender are fairly similar, but animate nouns take a thematic vowel and distinguish between a nominative and accusative case while inanimate nouns do not.

Distinguishing between nominative and accusative forms in animate words, while not doing so in inanimate words, is what Slavic natural languages do. :) however, comparing your conlang with slavic natural languages while your conlang is not slavic, is probably not a good idea. :)

[Omzinesý]

Mhilva (the newest version of it) has four genders/or noun classes.

I most humans and higher animals, inanimate nouns as well
II no clear semantic criteria
III no clear semantic criteria
IV abstract concepts (often derived ones like mothernity, leadership etc.), other nouns as well

The system thus isn't too semantic based. The animate-inanimate distiction is quite clear though.

The classes can be marked with affixes on nouns. Class IV is the only one that always demands a marker. If the markers are explicit, they are usually derivational morphemes.

Adjectives agree with their head in gender. Verbs agree with their intransitive subject in person and gender (you and me are always of the gender I). They argee with their transitive object if it is definite. Transitive sunject can only be animate. I'm not sure if that's about grammatical or natural gender. There isn't a gender agreement for it, anway. When intransitive nouns are transitive subjects, the verb is passivized and the subject appears in the ergative/instrumental case.

[Micamo]

All in all we seem to be debunking Hoskh's claim that non-romlangs with grammatical gender are virtually nonexistant.

[k1234567890y]

Mhilva (the newest version of it) has four genders/or noun classes.

I most humans and higher animals, inanimate nouns as well
II no clear semantic criteria
III no clear semantic criteria
IV abstract concepts (often derived ones like mothernity, leadership etc.), other nouns as well

The system thus isn't too semantic based. The animate-inanimate distiction is quite clear though.

The classes can be marked with affixes on nouns. Class IV is the only one that always demands a marker. If the markers are explicit, they are usually derivational morphemes.

Adjectives agree with their head in gender. Verbs agree with their intransitive subject in person and gender (you and me are always of the gender I). They argee with their transitive object if it is definite. Transitive sunject can only be animate. I'm not sure if that's about grammatical or natural gender. There isn't a gender agreement for it, anway. When intransitive nouns are transitive subjects, the verb is passivized and the subject appears in the ergative/instrumental case.

your system is somewhat similar to that of Bantu languages and maybe Northeastern Caucasian languages, maybe also has some similarities to Athabaskan languages.

[Sḿtuval]

All in all we seem to be debunking Hoskh's claim that non-romlangs with grammatical gender are virtually nonexistant.

---

I forgot to add this to the last post:

Ythnandosian is extremely gender neutral, partly because of the masculine/feminine merger I mentioned earlier. However, the common/neuter system inherited from Kauzasian has been slowly shifting to an animacy distinction. Although this is generally limited to loanwords, some native words have shifted gender (such as gagy sibling, which used to be neuter but in Ythnandosian became common).

ka he, she, it
eŋŋ man, woman, person
mind boy, girl, child
efünd son, daughter, offspring
efuy father, mother, parent
gagy brother, sister, sibling
efagagy uncle, aunt
kaze boyfriend, girlfriend
maďiril husband, wife, spouse

Pretty much any word describing a person or animal is applicable to those of any gender or of non-specified gender (a great language for those who don't identify with a gender).

If gender must be specified, one may use the adjectives niŋŋ male and maŋŋ female. Alternatively, one may use them as nouns (as is allowed with most adjectives) and use diminutives or augmentatives.

niŋŋ male person
niŋŋyso boy
niŋŋyni man
maŋŋ female person
maŋŋyso girl
maŋŋyni woman

[elemtilas]

All in all we seem to be debunking Hoskh's claim that non-romlangs with grammatical gender are virtually nonexistant.

I guess I can put in my nonromlangs with grammatical gender as well. Talarian of course has an animate - inanimate distinction; Avantimannish has vestiges of grammatical gender (m/f/n) though this is passing away; Loucarian has vestiges of the old m/f/n as well, but is evolving into an epic/neut or perhaps an ungendered system; Iconian (a non-romlang Italic conlang) has the usual m/f/n; Queranarran has a system of natural gender. Hotai languages seem to distinguish masculine, animate and inanimate genders. And then there's Kalchian which has masculine, feminine and epicene, but the fem. is wanting in the 2nd person pronouns and nominal forms of direct address, so you have to circumlocute. On account of you can't speak directly to a lady. But these are also the people that have an eight sided web of animacy domains, an eight place scheme of evidentiality and can't conjugate first person verbs of perception in the active voice -- they use a third person middle construction with verbal suppletion to do that.

[Aleks]

I don't know the full backstory of your conlang, but natural languages evolved long before transsexuality as we conceive of it today existed. And no natural language ever has developed with androids in mind. And it isn't necessarily naturalistic for a conlang to reflect your modern values or anyone's values really. (But all this is assuming naturalism is your goal.)

To be fair the conlang kinda comes from an advanced human like species. Maybe they thought to change their language like that.

[Xing]

Anyways on topic I felt genders didn't fit well in mine because what do you do with transexual people? Not knowing what to refer to them as, why not just use animate and inanimate genders? I feel this makes it better for conversing with people and makes it logical. The only thing I can think of that messes with this is AI and androids. With mine I have animate, inanimate, animate plural, and inanimate plural.

I don't know the full backstory of your conlang, but natural languages evolved long before transsexuality as we conceive of it today existed. And no natural language ever has developed with androids in mind. And it isn't necessarily naturalistic for a conlang to reflect your modern values or anyone's values really. (But all this is assuming naturalism is your goal.)

To be fair, there have probably been "borderline cases" - people who don't identify easily as either males of females even in pre-modern times. However, the vast majority of humans can be identified as male or female, and it's a relatively constant trait, so I can understand that it's handy to have natural gender/sex as a basis for a grammatical gender system.

The problem of "borderline cases" is not unique to masculine/feminine gender system. There will always be some arbitrariness in where exactly you draw the line between, say, "animate" and "inanimate", or between "rational" and "non-rational" beings.

It's not even a problem unique to grammatical gender systems. All (or at least, nearly all?? Are there any known exceptions??) human languages have some "naturally gendered" words, such as "man", "woman", "boy", "girl", "wife", etc – whether or not they have grammatical gender. As soon as you can make a lexical distinction between "A" and "B", there's always the risk that you run across some entity that cannot easily be classified as either an "A" or a "B".


As for my languages, neither Waku nor Nizhmel has gender. Waku had a distinction between animate and inanimate nouns when it came to the article. But since the distinction did not show up anywhere else than in the choice of article, I though I could as well ditch it. There wasn't much interesting stuff I could do with it. I think it'd be interesting to work on a language with gender, however. Some languages among the tonnes of sketches I've made during the years have had gender of some kind, but I haven't got very far with any of them.