Gender in conlangs

[Xing]

What genders/noun classes do you tend to have in your conlangs? Do you prefer to have few or many? Semantically or formally based? How are your genders/noun-classes manifested? (What kind of agreement do they trigger?)

If you don't have "pure" genders, do you tend to have something similar (like classifiers of various kind)?

An earlier stage of Wakeu made a distinction between animate and inanimate nouns, but I decided to scrap that distinction. Maybe I will reintroduce some kind of distinction between animate and inanimate nouns - though not a fully developed gender system.

I hope to be able to make a language with a masculine/feminine distinction one day.

[eldin raigmore]

What genders/noun classes do you tend to have in your conlangs?

If I had more than one conlang for which I'd gotten that far, I guess I'd prefer four sex-based genders: Masculine, Feminine, Neuter, Epicene.
But I actually have only one conlang whose gender-system I've decided on; and it's not sex-based.
Adpihi's gender-system is actually a cross-product of several. Most have two values, but "animacy" has three.

1. 1. inanimate
   2. sessile or bound animate
   3. translocomotive or free animate
2. 1. nonliving
   2. living
3. 1. abstract
   2. concrete
4. 1. nonrational
   2. rational

TTBOMK that's all of Adpihi's gender system.
In modern Adpihi all 24 combinations are actually attested, but in Old Adpihi many of them weren't (for instance anything rational was concrete and living and not inanimate.) The nouns in the newer combinations are newer nouns.

Do you prefer to have few or many?

Again, if I had several conlangs I think I'd say "I prefer few"; but the only conlang I've actually worked out a gender-system for has "many".

Semantically or formally based?

Semantically based.
(BTW isn't it a universal that in languages with genders (concordial noun-classes) all genders or all with at most one exception have a "semantic core" of nouns classed that way because of their meaning, whether or not the bulk of nouns get their genders for phonological or morphological reasons? (In some languages the way a noun declines determines its gender instead of the other way around!))

How are your genders/noun-classes manifested? (What kind of agreement do they trigger?)

Mostly the verbs, in Adpihi at least. I haven't gotten that far yet, though.

If you don't have "pure" genders, do you tend to have something similar (like classifiers of various kind)?

Like numeral classifiers and possessive classifiers?
And there's another, too, isn't there, that WALS.info doesn't have a feature or chapter for yet?

An earlier stage of Wakeu made a distinction between animate and inanimate nouns, but I decided to scrap that distinction. Maybe I will reintroduce some kind of distinction between animate and inanimate nouns - though not a fully developed gender system.
I hope to be able to make a language with a masculine/feminine distinction one day.

On CONLANG-L I once sketched out a 25-gender sex-based gender system that would cover species in which individuals changed sex as well as those in which individuals could be bi-functional hermaphrodites. But it's a gender system without a conlang right now.

I collaborated once on a (still incomplete AFAIK) conlang which wound up having a 9-gender system combining the four sex-based classes {Masculine, Feminine, Neuter, Epicene} and the four-valued scale {Human > Animate non-human > Living inanimate > Non-living}.
If I remember correctly:
Singular humans would always be Masculine or Feminine.
Singular animate non-humans could be any of the four sex-based classes.
Singular living inanimate beings (e.g. plants, also some fungi and some sedentary/sessile animals?) could be any of them but Neuter. Non-living were always neuter.
So that's 2+4+3+1 which is ten; obviously I mis-remembered something.

Anyway, any bi-capable hermaphrodite individual was Epicene; any individual of unknown or unspecified sex was Epicene; and any group of mixed genders was Epicene.
I think that means a man and his stallion would be Epicene Associative Dual rather than Masculine Human Associative Dual; and a woman and her mare would be Epicene Associative Dual rather than Feminine Human Associative Dual.
(Clearly I don't remember that conlang as well as I thought I did when I began this post.)

[k1234567890y]

most systems of my conlangs, if they have, are related to animacy, as most of the conlangs are for my con-sentinents, and the characteristcs of most of my con-sentinents make their languages impossible to have a noun class system based on sex.

however, I do try to create some conlangs for conpeople that are human beings, and some of them have noun classes based on sex.

[Serena]

What genders/noun classes do you tend to have in your conlangs?

I do not tend in each way. I like experiments...

Do you prefer to have few or many?

I usually have 1 or 2. Never had three...

Semantically or formally based?

Formally based.

How are your genders/noun-classes manifested? (What kind of agreement do they trigger?)

Hanilian has different endings for the two genders (High and low gender). They agree with their adjectives.

[Squall]

Thing: any thing
Animate: person, animals, monsters
"Person": any thing that "speaks" (humans, elves, dwarves, aliens, ghosts, vampires, gods...)
Male: any male thing
Female: any female thing
Inanimate: plants, objects, places, tools, stone...
Abstract: concept, action, event, information...

They are manifested only in the pronouns and determiners as a suffix. If the suffix is omitted, the gender is "thing".
You cannot use Male and Female if you do not know if it is male or female.

The genders use a subclass system and you can use the gender of the superclass. The genders of a female dog are Female, Animate and Thing. The genders of a woman are Female, Person, Animate and Thing.

[Thakowsaizmu]

The current language I am working on, Xanh, has four noun classes. Each class is strictly semantically bound, so if a noun is class I it is sapient.

The Classes are:
Class I Sapient
Class II Animate
Class III Inanimate
Class IV Intangible

In world, and I will relate this in further detail as I create the language more, the language actually came out of a constructed version of one of the most predominant language families on the planet, and was initially very strictly governed. As time went by and the language became the L1 for a substantial population, it became somewhat more organic. But this is the reason that there is almost no deviation from the noun classes.

[Sangfroidish]

Vorynthel has a rather standard masculine - feminine - neuter arrangement. A noun's gender is determined by which phoneme it ends in, basically just for ease of making a set of declension endings for each that I know will fit onto the root; masculine nouns end in sonorants, feminine nouns end in open syllables, neuter nouns end in obstruents. At some point I may introduce irregular nouns whose roots end in phonemes contradictory to their gender and mess up the entire declension system, but I'm sticking to refining the non-exceptions for the majority of the near future.

Verbs agree with their subject's gender in the third person, though in many cases the agreement is the same for both masculine and feminine nouns, and I'm considering devising a system whereby most adjectives agree with their referents too.

[Creyeditor]

If I use gender I tend to use animate vs inanimate, so i.e. I prefer few, semantically based.
Omlüüt has three genders, labeled masculine, feminine, which are a mix of sex-based and formally based (action nominals are e.g. always feminine) and manifest in adjective agreement and in the inflection classes of nouns.
Kobardon does not have genders, but some of the daughter languages will develop nominal classifiers.
Mamambam has a distinction between masculine human, feminine human, animate non-human and inanimate, but this is rather used as a very productive derivation tool although adjectives agree with their nouns.
Pöplish does not have gender.
Ponuhi has a human vs. non-human animate vs. inanimate distinction which is manifested in the number paradigms, case paradigms and the verbal alignment.

Well, the conclusion on the top does not seem to match the examples, but since I am scrapper, there are a lot of minor projects which would confirm my hypothesis

[Egerius]

Romanz is SAE, even if it's not from Earth, using feminine, masculine and (somewhat crumbling) neuter (in this case, I point at Romanian).
Adjecrives normally adhere to their nouns in number and in suffix and case (lu lupu buonu - the good wolf, de les amigas fermosas - of the beautiful friends), but third-declension adjectives only in number and case (lu caballu ifferme - the ill horse).

[chris_notts]

What genders/noun classes do you tend to have in your conlangs? Do you prefer to have few or many? Semantically or formally based? How are your genders/noun-classes manifested? (What kind of agreement do they trigger?)

A number of my conlangs have had systems of clitic/suffix classifiers. These tends to be more-or-less semantically based and used for agreement on modifiers within the NP as well as anaphoric reference and derivation of new nouns. The number of classifiers is difficult as the class is semi-open. For natural languages with a similar system see many members of the Arawak and Tucano families.

The following are the dominant distinctions for the most common classifiers:

Code: Select all

ANIMATE - 8 classifiers in total, based on a combination of the following distinctions:

- human
     - male / female
     - in-group member / out-group member
     - respect (but then other distinctions are neutralised)
     - child (same classifier as small 3d objects, other distinctions neutralised)
- animal
- plant

SHAPE-BASED - 9 classifiers in total, based on a combination of:

- dimensionality - long/flat/3d
- size - big/small
- rigidity (other distinctions neutralised in flexible classifier)
- straightness (1d only)
- roundedness (2d only)
- hollowness (3d only)

There are also more abstract, function-based and/or specific classifiers, e.g. -do, a classifier for locations and spaces.

[prettydragoon]

What genders/noun classes do you tend to have in your conlangs? Do you prefer to have few or many? Semantically or formally based? How are your genders/noun-classes manifested? (What kind of agreement do they trigger?)

Nouns in Rireinutire may be divided in four genders, rational, feminine, masculine, and inanimate (R, F, M, I). This is a system of natural gender, in that the gender of a noun depends on the qualities of its referent. Thus, words for sentient beings and female and male animals belong in rational, feminine, and masculine gender respectively. However, the names of natural astronomical objects are also in rational gender, because they were of old seen as deities. Also, ships are often personified as female animals, and so given feminine gender. Gender is only manifested in third-person pronoun forms.

[Yačay256]

Feom, being analytic, lacks gender, even in its pronouns; however, it does have a large array of classifiers (for both nouns and verbs). Feom classifiers are based mostly on substance or material as opposed to shape: For example, éx is for locations, vu is for water and liquids, sam is for elastic things and putties and jì is for vessels and/or containers. I was inspired to use this system (as opposed to the more familiar shape-based system of East Asian languages) from a lecture in my linguistics class two semesters ago, where my professor explained the Jakaltek Maya substance-based classifier system.

In Qiŝn, on the other hand, each nominal stem has exactly one of four genders: The rational, the zooic, the phytic and the irrational. These are both semantic and arbitrary in nature so that, while all nominals referring to people - like "person, human", "warrior/soldier" or "worker", are without exception rational, a nominal describing a human quality - like ratskğ "stupid" or kotŝ "mean", could easily be, for example, irrational (as in the case of "stupid") or zooic (as in the case of "mean").

[Illuminatus]

In the conlang I'm creating right now (it's called Sevon/sɛvɔn) there are 3 genders. Originally they stem from an ancient religion which divided things, animals and people into "sacred", "neutral" and "unsacred" ones, but now that religion is gone and the genders are just relicts from that time. You can't guess the gender of the word by its ending or stem, similar to the German gender system haha. The declension of nouns itself isn't affected by these 3 genders, but there has to be a concord between the adjectives and substantives, in addition there still exist different forms for the personal pronouns in the 3rd singular.

ma-ɛv-a riθ-a, ma-ɛv-i skɔn-a, ma-ɛv-ï bast-Ø
[maˈʔɛːva ˈriːθa maˈʔɛːvi skʰɔːna maˈʔɛːvɨ bastʰ]
NOM-small-G1.SG woman(G1)-NOM.SG, NOM-small-G2.SG stone(G2)-NOM.SG, NOM-small-G3.SG house(G3)-NOM.SG
'a small woman, a small stone, (and) a small house.'

[eldin raigmore]

I forgot to mention:
In some conlang I collaborated on (which probably still isn't finished), there were four semantically-based genders:
1. Animate -- conceived of as anything with a soul or a spirit.
2. Vegetable -- conceived of as anything which would be useful for supporting and sustaining life if ingested. (So, for instance, that subcomponent of the UV component of sunlight which promotes the synthesis of Vitamin D, would be "vegetable light".)
3. Mineral -- conceived of as anything which originated by being dug out of the ground.
4. Other -- anything else, e.g. any abstract noun.

In case a noun could belong to two genders it belonged to the highest one.
Meat-animals and dairy-animals would be animate while their meat and milk would be vegetable.
Salt would be vegetable, as would edible roots, rhizomes, and tubers, in spite of digging for them.
And so on.

Presumably if a newly-coined or newly-borrowed noun "sounded like" it should belong in one of the genders, it might be made to fit into that gender, regardless of its semantics; or, might not.

[Ahzoh]

In Arkheo, pronouns are distinguished by whether the person being referred to was alive or dead, for example, "ér" vs. "éya". Both refer to 3rd person masculine, but "éya" refers only to someone who once lived in the past but not anymore.
Does that count as "gender"?

Do you think it plausible to have sex-based gender marking on plural forms of verbs and not singular?

[Micamo]

Do you think it plausible to have sex-based gender marking on plural forms of verbs and not singular?

There's some amazonian language that does this (all 3rd-person singular agreement forms are null, but the plural forms distinguish gender), but I can't quite remember which one it was. I'll browse through my grammars and see if I can't find it. I know of none where third singular forms are overt, however.

[clawgrip]

In Naduta there are three genders in two classes:

1. animate
2. common:
   1. known
   2. unknown

The animate gender is not limited to living things but also includes things that tend to move or be in flux, so this includes water, air, fire, emotions, etc. A few bugs and things are not in the animate gender, but most moving things are.

The known and unknown genders are somewhat mixed. The known one tends to include body parts, tools and buildings, and a number of uncountable/mass nouns. Unknown tends toward nature as well as countable nouns. The two genders are pretty mixed though. Their case marking is similar and it looks like the unknown gender declension is based off of the known one, though I haven't considered the relationship clearly yet.

Naduta was initially nothing more than a throwaway language meant to support the script I was developing, so I didn't think particularly hard about the gender system, but now I am developing the language more seriously and I like how the gender system turned out.

[moose]

I played around with a wet/dry noun class system in a recent project. Not sure how realistic is was, but I wanted something simple, yet non traditional. Wet class nouns covered liquids, aquatic animals and plants, fruits, and some body parts. Dry covered tools, land animals and plants, meats, strong solid materials, and kinship terms. Maybe it seems a little arbitrary, but I kind of liked it.

[Znex]

Proto-Oeg has three noun classes that are mostly only indicated by verbs:

1. Sentient
2. Animate
3. Inanimate

The sentient class is pretty self-explanatory: this covers most humans and entities comprised of humans. The animate class includes anything that moves by itself and has life but not sentiency, such as animals or plants. Among these nouns, the Oeg also include most of the natural forces, emotions, internal organs, the dead, and "sub-humans" (such as murderers). The inanimate class covers anything that does not move of its own accord or isn't covered in the other noun classes, including body parts, concepts, "sub-human" behaviour and emotion (such as idiocy and raw hatred), edible things, rocks, etc.

[Thrice Xandvii]

I am not a huge fan of noun class systems in my constructed languages... but when/if I do use them, they are definitely not going to be of the completely arbitrary masc/fem system (á là Spanish). It always bugged me that things like "the bra" are somehow masculine words (el sostén) for really no apparent reason. It seems a purely semantic argument, but I almost would have preferred that they simply be called Class I and Class II nouns, at least that way, I wouldn't be annoyed with the illogical terminology!

Where was I now? Oh yes. Noun class.

I initially avoided noun class since the languages I tend to make don't have many inflection forms and thereby wouldn't have much in the way of subject/verb agreement or adjective/noun agreement. So, to me, it seemed somewhat pointless to have a system in which few to no uses for it would exist. However, I did come up with a system that I kinda liked for Ĉètkãdo in which there were three noun classes:

• Solid: Things that are solid, hard or unchanging. It covered the literal as well as the metaphoric sense. Rocks, turtles, cheese, man, Earth, etc.
• Fluid: This was for liquids, soft things and things that flowed or changed. Water, fire, air, lava, women, sponges, etc.
• Æthid: This was for ethereal things and things that had to do with gods or heavenly bodies. Stars, seasons, weather, gods, goddess, etc.

While this too seems somewhat "illogical" how some things like the seasons or emotions might get classified, it also helped to define my concept for the conpeople who spoke the language. This is an idea I will likely recycle into another project at some point in time.

Agreement in this language included suffixes attached to the verb that agreed with the class of both the subject and object of the verb (when applicable).

For my current project, I may well introduce classifiers of some kind. Since it is very much inspired by Chinese, I may as well go all the way with it, eh? Though if I do introduce a system of classifiers for numbers, it will be a far less involved one than in Chinese.