In which languages do non-proper NPs have a default unmarked interpretation as definite, rather than indefinite?

In English, the default interpreration of plural and uncountable NPs is normally indefinite (thus, stars = "some stars", and water = "some water"). Countable singular NPs generally have no stable interpretation without a determiner, unless they are proper nouns (thus, you generally do not see a plain form like frog, as opposed to a frog, the frog, etc.).

Are there any languages that show the opposite tendency to English in this case — not necessarily for all forms of NPs, but for some of them?


  • In Finnish, there is a case to be made (although the consensus doesn't seem settled) that plural and uncountable NPs without any determiners are seen as definite. However, whether or not this description of Finnish is accurate, it only applies to subject/object NPs: the definite/indefinite distinction is fully neutralized in the oblique case forms, and must be assessed through context, just like singular countable NPs. – user8017 Nov 10 '15 at 2:22
  • An admin changed the title of this thread to "Are there any languages in which definiteness is unmarked". I changed it back to the original title, because any language lacking definite articles could satisfy the criterion of definiteness being unmarked -- I am asking about something slightly different. – user8017 Nov 10 '15 at 2:53
  • Doesn't it depend highly on the context and the type of noun? A previously-mentioned noun, a proper noun, or a noun like "Sun," Moon" or "sky" that refers to a single object is surely more likely to be interpreted as definite. – brass tacks Nov 10 '15 at 4:19
  • Actually, how do you even define definite or indefinite? – brass tacks Nov 10 '15 at 4:24
  • 1
    @jlawler I am only a very amateur and informal speaker of Turkish but I believe that default definiteness is indeed the case for subjects in Turkish (and other Turkic languages). Rather it is the objects (accusative case only) that can be marked for definiteness. – Adam Bittlingmayer Nov 13 '15 at 21:38

A complete list or fraction will be subjective, because each case has nuances - for one thing even languages with definite articles use them differently - and because it is always subjective what should count as a separate language.

Roughly speaking, among the languages of Eurasia and the Western world, essentially Slavic, Ural-Altaic (regardless of the dispute around it), Indo-Aryan, Dravidian and Berber languages can have definite NPs by default or no articles at all, including for countable nouns. The oldest forms of Greek and Latin had no definite article, but they later developed them. Please pardon my ignorance on the rest of the world.

Definite and indefinite articles, World Atlas of Language Structures

It seems to be a partly areal feature, for example it includes the Eastern Indo-European languages that had long and intimate contact with Altaic and Dravidian languages.

That said, no matter the counting methodology used, default definite NPs ie implied definite articles are rare, much rarer than languages with no articles at all. Furthermore, the overall distribution seems to be consistent between Eurasia, Austronesia, Africa and the New World.

To understand the nuance, let us examine some of well-known languages with NPs that are definite by default. In formal Persian it is so:

In the literary language, no definite article is used; rather, it is implied by the absence of the indefinite article.

In Turkic languages it is similar, with some nuances. For example:

Araba sokakta. (The car is in the street.)

Breaking it down a bit, we have Car street-IN. Araba simply means car, and is in its pure nominative and canonical form. (You can regard -ta and its variations either as a locative case ending or as a postposition meaning in, for our discussion it does not matter.)

Furthermore, notice that sokak (street) is also a definite NP.

So this is a good example of default definite interpretation of NPs.

In both language groups, however, there are many cases of countable, and of course uncountable, NPs without any article that would be translated into, say, English with the indefinite article or with no article, for example:
آیا اتومبیل دارید؟ | ("Do you have a car?")
Kar var mı? | ("Is there snow?")

Only in some context would one interpret this as Do you have the car?, and it is hard to dream up a context where "Is there the snow?" would make sense. Thus in such instances it functions more like a language without any articles.

Compare this to the following, with the indefinite article is strange or emphatic, a bit like English one (as opposed to a/an):
آیا یک ماشین دارید؟ | ("Do you have one car?")
Bir kar varmı? | ("Is there one snow?")

Again, it only makes sense in context or almost never.

In Slavic languages, there are generally neither definite nor indefinite articles, which is actually very common among the world's languages. In these cases, the default interpretation is mostly a question of the context, although naturally native speakers do not bother to map every NP to the concepts of definite and indefinite in an English speakers' mind.

As you are interested in the overview, some exceptions to the groupings: among Iranian languages, Ossetian and informal Persian and Kurdish can mark definiteness explicitly, and apparently Tajik marks only objects, like Turkic languages. Hungarian has very simple and explicit definite and indefinite articles, like Western European languages. Among Slavic languages, Macedonian and Bulgarian and related dialects do have a definite article, like other Balkan languages. Again, this is definitely an areal feature. Also, I cannot find a colonial-era creole that does not have a definite article.

In any case, there are always ways to be explicitly definite, eg with demonstratives, ie by saying that frog, and indeed definite articles in many languages were derived from such ways. To me anyway it is not clear where to draw the line.

  • Please note that my Persian and Turkish are not very good, if you have better examples then please tell me or edit the question. – Adam Bittlingmayer Nov 13 '15 at 21:48

I realized that French may also qualify as an answer to this question. Although indefinite singular countable nouns are marked with un (un canard, une voiture, etc.), I think that indefinite mass-nouns and indefinite plurals are marked with de "of" plus the definite article: e.g., du lait "(some) milk" is (at least historically) made up of de + le + lait "of the milk", and des voitures "cars" is composed of de + les + voitures "of the cars".

Although nouns without articles are not interpreted as definite in French, this does seem to be a case where indefiniteness is formed on a definite foundation of some kind.

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.