In the comments of another question about animate as noun gender in some Slavic languages an interesting point was raised.

Many if not most Indo European languages exhibit grammatical gender for nouns. Also it's known that for many other languages there is a broader realm of "noun classes", certainly in African and Australian languages and the related concept of "counter words" common in East and Southeast Asian languages.

So are there languages in the Indo European family which exhibit noun classes in the broader sense rather than just grammatical genders? (I'd like some references rather than just speculation please, and please only submit answers about nouns and not pronouns)

There is one complicating factor as mentioned in Wikipedia:

Some authors use the term "grammatical gender" as a synonym of "noun class", but others use different definitions for each.

From what I have read a noun class, like a gender would be an attribute of each word at the lexical level and not dependent on semantics, syntax, usage, etc. This is what makes noun classes distinct from classifiers if I am not wrong.

  • 3
    Note that animate v. inanimate is semi-native to many IE languages, since they already distinguish between neuter on the one hand and masculine/feminine on the other: neuter words are very often inanimate (though the converse is hardly true).
    – Cerberus
    Commented Sep 22, 2011 at 0:52
  • Ah well that is what makes it grammatical gender as opposed to natural gender. The former is largely arbitrary and abstract unlike the latter. Commented Sep 22, 2011 at 0:55
  • 4
    Just to second @Cerberus' remark, AFAIK the inanimate/animate distinction pre-dates the M/F gender distinction in the evolution of IE languages. For instance Hittite, believed to have branched out early from theoretical PIE, features only the inanimate/animate classes. So do a lot of non IE languages categorised as archaic (e.g. Elamite, Sumerian in Ancient Middle East). However, the hypothesis that the M/F gender distinction results from a split of the animate class is contradicted by various Slavic languages (Czech, Polish or Slovak) having both inanimate and animate Masculine classes. Commented Sep 22, 2011 at 15:20
  • 1
    @AlainPannetier. I know it says this all over the internet, but it is wrong. Hittite had two genders: common and neuter. Inanimate nouns are distributed between the genders, as they are in other IE languages.
    – fdb
    Commented Feb 2, 2015 at 15:32
  • 2
    @fdb. That looks like nitpicking to me. See for instance "A GRAMMAR OF THE HITTITE LANGUAGE" p64 (ch 3) by Harry A. Hoffner Jr. and H. Craig Melchert. "Gender. Hittite recognizes two grammatical gender classes, traditionally called common and neuter, alternatively animate and inanimate". Although I agree Olivier Lauffenburger's grammar does not cite the animate/inanimate "genders". Body parts for instance being neuter it's probably more accurate indeed, to talk about common and neuter then. Let's correct the "Internet"! Commented Feb 2, 2015 at 17:20

5 Answers 5


Yes, the same answer that was provided in the previous question applies here. Several Slavic languages have an animate/inanimate distinction in addition to masculine/feminine/neuter. In some languages, only certain combinations are permissible - for example, Russian only distinguishes between animate and inanimate in the masculine.

It's also been hypothesized that Proto-Indo-European possessed only an animacy distinction, with masculine/feminine developing out of the animate class of nouns and the inanimates becoming neuter. This idea appears to originate from Meillet (1926), and was further developed by Gamkrelidze and Ivanov (1973, 1984).

Gamkrelidze, T.V. & Ivanov, V.V. (1973). "Sprachtypologie und die Rekonstruktion der gemeinindogermanischen Verschlüsse". Phonetica 27. 150-156.

Gamkrelidze, T.V. & Ivanov, V.V. (1984). "Indoevropejskij jazyk i indoevropejcy". Rekonstrukcija i istoriko-tipologicheskij analiz prajazyka i protokul'tury. Tbilisi: IzdatePstvo Tbilisskogo Universiteta

Meillet, A. (1926). Linguistique Historique et Linguistique Generale. Honore Champion, Paris.

  • 2
    Interestingly, Danish and Swedish have gone to a animacy distinction but a grammatical one. The masculine and feminine have merged into a "common" gender but neuter remains. It's grammatical in that not only inanimate things are neuter and not only animate things are common. Would that mean that PIE had noun classes but Scandinavian has gender? Commented Sep 22, 2011 at 1:11
  • 1
    Can you reword your first sentence? I can't parse it. Gender is a feature that divides nouns into two or three classes - this is often arbitrary (compare German der Tisch (m.) and Spanish la mesa (f.) "table"), but it is so named because human males and females will each select a pronoun or agreement paradigm from exactly one of the classes. So both PIE and Scandinavian have noun classes - the former divides by animacy and the latter by gender.
    – Alek Storm
    Commented Sep 22, 2011 at 1:30
  • 1
    Well it gets fuzzy because the very term "gender" has changed meaning. It originally only meant the grammatical category but has been extended to cover sex so now when people use it it's hard to tell when they're making a connection which was not originally there. Commented Sep 22, 2011 at 1:36
  • To clarify when you say "it is so named because ..." do you mean gender was so named, or that the specific genders, masculine, feminine, etc were so named. If the former then this is actually wrong, if the latter then yes certainly for the case of the named 2/3 gender languages we're familiar with but I'm not sure about languages with a greater number of noun classes. Commented Sep 22, 2011 at 1:38
  • 1
    I guess I'm still not being clear. The term gender in linguistics does not refer to sex difference, it comes from the same root as "genera" and "generic". It seems counterintuitive but it's the other way around. The sex sense borrowed it from the linguistics sense. Commented Sep 22, 2011 at 1:59

Modern Spanish has a construction (called a personal in Spanish) that (in some dialects) has an animacy requirement. This involves the insertion of a (homophonous with the preposition translated "to"; whether it is a preposition in this context is debatable) before a direct object:

(1)  He     visto a Maria
    *He     visto   Maria
     I.have seen  A M.

(2) *He     visto a la  mesa
     He     visto   la  mesa
     I.have seen  A the table

(The judgments given correspond to most but not all dialects of Spanish. Whether a non-human living being gets a personal is, as I understand, a particular locus of dialect and speaker variation.)

This is of course orthogonal to gender, which is marked in the usual Romance way on nouns, adjectives, determiners and pronouns.

  • 5
    Yes indeed in fact Romanian has the same thing with the preposition/particle "pe". They certainly reflect animacy or human-ness, but is that the same as noun class? For instance each noun in the dictionary has a gender but would it make sense for these languages to assign each noun also a +/- (animate/human) attribute also? Commented Sep 22, 2011 at 2:36
  • 2
    @hippietrail, yes. The information is not predictable (dialects differ as to what is animate; there are also idiomatic exceptions -- I think "fire" is animate is some Slavic languages), so the information has to be encoded lexically. Many linguistic theories regard words as bundles of "features," of which "gender" and "animate" are two. (Pro)ouns also have "person" and "number"; verbs have case features (and transfer these to nouns via syntactic agreement); these then influence the ending of the noun. So "gender" is just one of many possible ways to chop nouns up into classes.
    – Aaron
    Commented Sep 22, 2011 at 2:47
  • 1
    I've been thinking about it and personal a doesn't seem to be lexical but syntactic with maybe a semantic component. It's used when referring to people whether by name or words like quien. For instance there won't be cases of needing a with some synonyms and not others like two synonyms for the same thing can have different genders: la estrella, el astro. Commented Sep 22, 2011 at 9:31
  • 1
    @CesarGon, the contrast is between "He visto a Maria" and "He visto Madrid," either sentence becoming ungrammatical if you add/remove "a". There is dialect variation, but this is the general pattern – does it hold in your dialect? (You seem to be a native speaker.) It is also worth mentioning that the article with personal names is a vernacular feature in Spanish, so one could also say "he visto a la Maria" which would be prescriptively marked but not ungrammatical.
    – Aaron
    Commented Sep 24, 2011 at 16:59
  • 1
    @Aaron: Indeed, "He visto a María" and "He visto Madrid" both become ungrammatical if you remove/add "a". And, as far as I know, this wouldn't change depending on dialectal forms. Regarding the article with personal names, yes, indeed; see my answer to linguistics.stackexchange.com/questions/5/…
    – CesarGon
    Commented Sep 24, 2011 at 17:03

In Asturian, there is such a thing as a different noun class than masculine / feminine. You can mark a noun with a different suffix to indicate that it is an abstract version of that noun. This is the so-called neutru de materia, used for non-countable mass nouns. This presents a five-fold paradigm morphologically:

  1. -u for masculine singular
  2. -os for masculine plural
  3. -a for feminine singular
  4. -es for feminine plural
  5. -o for the “material neuter”

Nouns are still either masculine or feminine, but you can make them abstract / material nouns. So for example, if you had a red hair, it would be un pelu roxu, but if you were talking about red hair in general, it becomes pelo roxo.

It’s a bit more complicated than this (for example, feminine nouns don’t change their endings), but that’s the basic idea. You can find vestiges of this in the Castilian ése / ésa / eso distinction, and in lo necesario.

If you read Castilian, you should be able to piece your way through texts in Asturian that talk about this very interesting subject. One such is El llamáu neutru materia.

  • Very interesting! Commented Aug 4, 2012 at 21:25

Welsh shows some signs of a semantic classifying system in its noun plurals. There are a great many plural affixes in Welsh, partly reflecting the old stem-endings of nouns (which have often disappeared in the singular forms due to the loss of final vowels) and partly reflecting new analogical processes based on semantics.

For example, the suffix -od very often occurs with animal names: llewod "lions", cathod "cats", llwynogod "foxes", tyrchod "hogs, boars", crancod "crabs", llyffantod "toads", drywod "wrens", etc. etc.

The most widely-known animal names tend to resist this pattern, perhaps through the force of frequency: ceffylau "horses", cwn "dogs" (pronounced [kun]), defaid "sheep", ieir "hens", etc.

Many of the nouns with plurals in -od also have an alternate (and possibly older) variant of the plural: e.g. llyffaint "toads", crainc "crabs".

Another pattern can be seen with the plural suffix -ydd, which seems to be frequently used with nouns referring to pathways (roads/rivers/etc.), fields and certain other geographical locations: meysydd "fields", heolydd "roads", afonydd "rivers", corsydd "swamps", dolydd "meadows", neintydd "brooks, streams", camlesydd ”canals” and so on. (I haven't seen this pattern with -ydd pointed out in a reference grammar yet, but based on my survey of the lexicon, it seems pretty likely that there is or has been a semantic aspect to its use.)

As with the -od suffix, there are words for which -ydd coexists with a possibly-older plural form: cyrs "swamps", naint "streams, brooks", camlesi "canals".

Alongside these features, Welsh retains the IE masculine and feminine gender classes (the neuter having dissolved into masc. and fem.). In the modern language, there is no longer much systematic difference in the way that masculine and feminine plurals are formed, although some patterns can be observed. For example, the vowel -e- is generally more likely to appear in feminine nouns (both in the root and in suffixes): cf. dinasoedd "cities", modrabedd "aunts". (This is because the old feminine ending *-a, which disappeared a long time ago in Welsh, had the effect of lowering the vowel *i in a preceding syllable to *e.)


Pontic Greek, historically spoken in Northern Turkey, has acquired an animacy distinction on top of its inherited gender system; see e.g. http://westminsterresearch.wmin.ac.uk/16114/ :

  • Human nouns have full grammatical agreement, and can be of any gender. i mikresa i nifæ eton ki alo poniresa, "the.FEM small.FEM bride.FEM was even more cunning.FEM"
  • Inanimate nouns retain gender in their morphology; but any adjective or pronoun that aligns with them is in the neuter. t asimenion o mastrapas pali kremete "the.NEUT silver.NEUT the.MASC pitcher.MATCH still hangs"
  • Animate non-human nouns retain gender in their morphology; any adjective or pronoun that aligns with them is in the neuter. Masculine such nouns take masculine articles; but feminine such nouns take neuter articles. ta katas eksenitepsan ki i pendiki xorevne, "the.NEUT cats.FEM went.away and the.MASC mice.MASC dance"

The assumption I've seen in older accounts was that this was triggered by language contact with Caucasian languages. The paper I've linked to seems to prefer internal causes, and places Pontic on a continuum with Cappadocian Greek (historical spoken in Central Turkey), which has lost gender agreement altogether.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.