Does anyone know how typologically common it is to have number agreement on nouns even when that noun's plurality is implied by numerals?

For instance, in Spanish "two houses" is "dos casas", i.e, the noun "casa" is marked for plural even though we already know there's two of them. Why not "dos casa"?

Are there more languages in which plural marking is always obligatory or more in which plural marking becomes optional with numerals? Is either strategy related to other features?

  • 1
    Of languages that I know about with number agreement, Turkish does not pluralize nouns with numbers. Of course in a language like Malay or Mandarin, there isn't any number agreement to start with.
    – jlawler
    Mar 17, 2017 at 0:32
  • 4
    Related, but no clear answer: linguistics.stackexchange.com/questions/20688/… Mar 17, 2017 at 1:23
  • 2
    Turkish is interesting. With numerical determiners, the noun remains in the base form. For example, "araba" (car) ~ arabalar (cars), but "Benim üç araba var" (or "Üç arabam var".) = "I have three cars".
    – BillJ
    Mar 18, 2017 at 9:38
  • Brazilian Portuguese is like Spanish but appears to be in the process of shedding the agreement (duas casas > duas casa is easy to spot in colloquial speech). Jun 16, 2017 at 9:36

2 Answers 2


Pluralising after a number is an arial feature, not necessarily a question of language family. For example, Armenian, Georgian, Persian and Turkish do not mark after a number (or after the word "many/much"), although they are from 3 completely separate language families.

But there are many borderline cases. Some languages never pluralise with morphology anyway. Some languages have a singular, a dual or paucal and a plural. Some languages do not pluralise after a number in some cases. For example, in German, it is correct to say "2 Bier".

Even if we somehow handle the edge cases, the definition of "common" is arbitrary, because there is no objective way to count languages. Does Luxembourgish count? Does it have the same weight as Bengali with hundreds of millions of speakers?

So it probably makes more sense to look at map than at a number or even a list. But we should beware of false precision in answers to questions like this.

  • "edge cases" > "borderline cases". Otherwise a good answer.
    – fdb
    Jun 18, 2017 at 20:37
  • @fdb Good point, edited. Jun 19, 2017 at 7:19
  • My intuition is that there are correlations between this pluralisation feature and others. For example, "I like apples." sounds quite strange in some languages, they would say "I like apple." But again this is based on a laughably small sample size. Jun 19, 2017 at 7:26

The languages put under the Altaic group, generally have optional or non plural with nouns. For example Japanese doesn't have it all, or Turkic languages usage of plural nouns are discouraged and limited to the couple of cases. Sama goes for Ugro-Finnic languages.

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.