Does a language exist whose older forms are known to have lacked the category of grammatical gender, and which proceeded to evolve one (perhaps from a non-gender-based system of noun classes)? Are "pre-gender" stages of language evolution, where they existed, universally a thing of such distant past as to be beyond reconstruction?

  • By "gender" do you mean sex-based, drawing from masculine, feminine and neuter?
    – user6726
    Feb 25, 2015 at 18:44
  • Or are you talking about a noun classifier system, like Navajo, Swahili, or Burmese? Feb 25, 2015 at 18:53
  • @user6726 Yes, I mean sex-based but extended to all nouns, as in most Indo-European languages, Semitic, etc. Feb 25, 2015 at 19:01
  • 2
    "Gender" always meant "kind" and was only used in connection with grammar. It acquired the secondary meaning of "sex" much more recently. Grammatical gender was not "based on" sex but on kind. The sexes may fit into the genders, the genders are not extensions of the sexes. Much confusion has resulted from the case of English semantic shift. Feb 26, 2015 at 8:55
  • 1
    @hippietrail "According to Aristotle, this concept was introduced by the Greek philosopher Protagoras. τὰ γένη τῶν ὀνομάτων ἄρρενα καὶ θήλεα καὶ σκεύη The classes (genē) of the nouns are males, females and things." en.wikipedia.org/wiki/Gender#Etymology_and_usage Feb 26, 2015 at 14:50

2 Answers 2


Surprisingly, Indo-European seems to be an example. Silvia Luraghi has an article "The origin of the Proto-Indo-European gender system: Typological considerations" (Folia Linguistica 45/2 (2011), 435–464) which discusses this, and it appears that there is agreement that the M/F/N system of later languages developed from a two-gender system where masculine and feminine were not distinguished, and the system was based on an animate / inanimate.

  • 2
    I knew someone would write something like this. You cannot "observe" a development from a hypothetical reconstructed language to a real language.
    – fdb
    Feb 27, 2015 at 22:42
  • I take it that based on the presence of the word "observe" title of the question you interpret his question as being about historically-attested changes. If that is your conclusion, then this would also be a problem with the Khasi example.
    – user6726
    Feb 27, 2015 at 22:52
  • In priniciple yes.
    – fdb
    Feb 27, 2015 at 23:11
  • 1
    @fdb Well, considering the answer to my exact question seems to probably be "no", anything coming close is interesting enough. Feb 28, 2015 at 4:49

Khasi, an Austroasiatic language spoken in Northeast India, differs from practically all other members of that family (at least those outside the Munda branch, as @user6726 rightly points out) in having grammatical gender. There is a paper about this: Lili Rabel-Heymann, ‘Gender in Khasi Nouns’, Mon-Khmer Studies Journal IV: 247-72 (1977), available online here.

Every Khasi noun is preceded by a gender indicator commonly known as an “article,” a term borrowed for reasons of convenience from the grammar of Indo-European. This gender indicator is repeated before the verb, and is then known as a pronominal verbal prefix. A morpheme identical with the nominal article and the verbal pronominal prefix functions in free form as a personal pronoun. It might therefore be said that the Khasi pronoun occurs as a free morpheme by itself and as a bound form before nouns and verbs. In any case, the prenominal and preverbal forms always agree with respect to number and gender. (Heymann p. 247)

While as far as I know there are no surviving records of earlier versions of the language that did not exhibit this feature, we can infer that what was originally a pronoun came to do double duty as a gender marker. And the reason for the innovation – contact with Indo-Aryan languages in the region – can also be inferred.

  • 1
    This might be an example, depending. S. Bhattacharya "Gender in the Munda Languages" indicates that sex-gender is widespread in Munda, and concludes that gender was marked (via suffixation) in Austroasiatic (lost in a number of languages due to contact with Chinese). It is possible that the original semantic basis was along the lines of rational / irrational or superior / inferior.
    – user6726
    Feb 27, 2015 at 18:37
  • @user6726 Good point! I don’t know how similar the gender systems of Khasi and the Munda languages are, and the other languages in its branch, Khasi-Khmuic, don’t have gender. It’s conceivable that Khasi had gender at some early stage of the language, lost it at the time that Khasi-Khmuic was one language, then innovated it anew in the way I’ve suggested.
    – neubau
    Feb 28, 2015 at 3:57
  • There doesn’t seem to be much of a consensus about the location of an Austroasiatic homeland, or about other issues of historical reconstruction for this family, but George Van Driem’s 2006 article in the M-K Studies Journal can provide some background.
    – neubau
    Feb 28, 2015 at 4:10
  • What does "gender" mean here? Do they distinguish three noun classes that distinguish male and female persons?
    – shuhalo
    Sep 9, 2015 at 21:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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