How do your languages treat (in)definiteness?

[Snyexarosha]

I am deciding how my language will interpret nominals as definite or indefinite. The language is an article-less language with zero-marked singulars and a plural suffix. Right now I'm considering having overt plural marking necessarily result in a definite interpretation, with a zero-marking ambiguous between a general "type" interpretation (e.g., "the dog is a type of a mammal") and a singular that's ambiguous between definite and indefinite. To get a non-type indefinite plural, you might use a quantifier like "some" paired with a zero-marked nominal; quantifiers, in turn, obligatorily pair with zero-marked nominals.

These are just some initial ideas I've been considering for my article-less language, and I'm curious what you have all come up with. This is not something I generally see discussed among conlangers (nor is semantics at large, frankly). Are your languages article-less with nominals formally ambiguous between definite and indefinite? Do you mark the distinction with an article? Is there some other distinction that you mark (e.g., specificity)?

[Dormouse559]

Faux-Phrygian, the conlang I’ve been working on lately, has a definite article, with some restrictions. It marks referents that have already been mentioned in the discourse or which are physically present. Other possible definite categories, like shared cultural concepts, take no article on first reference. Indefinites are also unmarked. My thought is that the language is in the process of turning a demonstrative into a definite article, and at the moment, the article is still restricted mainly to places that would be appropriate for a demonstrative.

[Ahzoh]

Vrkhazhian has no definiteness marking. However, nouns modified by demonstratives are inherently definite. I've also considered that the subject/agents of verbs tend to be definite by default while objects/patients are indefinite by default.

[eldin raigmore]

I think Adpihi and Reptigan have obligatory demonstratives for specific/referential noun-phrases, including definite ones.
Pronouns, of course, are all definite already.

I’ll have to find out about Arpien; I don’t know yet.

[Salmoneus]

Faux-Phrygian, the conlang I’ve been working on lately, has a definite article, with some restrictions. It marks referents that have already been mentioned in the discourse or which are physically present. Other possible definite categories, like shared cultural concepts, take no article on first reference. Indefinites are also unmarked. My thought is that the language is in the process of turning a demonstrative into a definite article, and at the moment, the article is still restricted mainly to places that would be appropriate for a demonstrative.

Like Faux-Phrygian, Old Wenthish (a 1st-millennium Germanic language) is in the process of developing a demonstrative (the cognate of "that") into an article. It is used ostensively (i.e. for things physically present), in contrasts (that dog vs another dog, that dog vs my dog etc), or for emphatic reference back to something already mentioned.

However, being a Germanic language it already has a different system of definiteness already built into it, in the form of definite and indefinite declension paradigms for adjectives. The presence of a demonstrative (or a possessive) forces the noun to be definite and hence demands definite adjectives - but nouns can still be definite, with definite adjectives, even in the absence of the demonstrative. It's the adjectival marking that's the fundamental definiteness marking in the language at this stage - even though adjectives may be absent (and are ambiguous in some forms anyway).

Later stages of the language will evolve a full definite article. In fact, they'll evolve two - a true definite (from 'that') and a semi-definite or specific (from 'one', although it's probably reinforced from the Irish definite 'an').

--------------

Rawàng Ata has no definite article. However, definiteness can be inferred in several ways.

Firstly, initial-syllable reduplication, though for simplicity treated as a 'plural' marker, is actually a definite plural marker (indefinite plurals look like singulars), and can sometimes be used even to mark a definite singular when singularity is clear from context (though this would be regarded as poetic or vulgar, not the ordinary manner of speaking).

[so, "a cat eats mice", and "cats eat mice", and in this respect usually "the cat eats mice" all look the same, but "the cats eat mice" is marked, and can occasionally be used even when 'the cats' refers only to one actual cat. Although this marking is optional, and really means something more like "these (numerous) cats eat mice"]

More importantly, however, definiteness plays a role in one of the central concepts of the language: transitivity.

Briefly, there is a distinction between transitive and intransitive dyadic dynamic clauses. The clause is transitive only when certain conditions are met, one of which is that the agent and patient must both be specific (others are that the agent must be in control of the action, the action must be 'successful', the patient must be substantially affected by the action, and the patient must not be higher in the animacy hierarchy than the agent). Intransitive clauses have different agent marking: in transitives, the agent is in the nominative and the patient is (usually) in the accusative, whereas in intransitives the agent is in the ergative, and the 'patient', actually now considered an indirect object, is usually in an oblique case or governed by a preposition (although I think the accusative may still be possible in some situations).

As a result, "the cat eats the mouse" and "a cat eats the mouse" are actually very translated very differently: one as "cat eat mouse-ACC", and the other as "cat-ERG eat mouse-ERG" (I think - the second ergative here acting as a partitive).

However, you can't tell whether it's the cat and/or the mouse that's non-specific, and there are also other interpretations - it might mean "the cat nibbles a small bit of the mouse". You can't detect the non-specificity when there's an animacy violation (eg "the cat eats the man" and "a cat eats the man" are the same, because having a human as the patient and a non-human animate as the agent automatically requires intransitive marking). And with an inherently intransitive verb, like "chase", there is likewise no distinction. [I suspect that actually 'chase' is translated as 'catch' but in the intransitive, but I don't know, and there's probably another verb like 'run after' anyway]

More directly, however, definiteness itself (not specificity!) is required for topicalisation, so any overt topicalisation implies definiteness. Non-topicalisation, however, does not imply indefiniteness. [you can say "as for the mouse, a cat eats it", but not "as for a mouse, a cat eats it"]

This is all kind of an example of the general attitude of the language, which is to often avoid simplistic, obligatory one-to-one marking of things.

[WeepingElf]

Old Albic has a definite article, but no indefinite article.

[Reyzadren]

No definiteness. No articles.

My conlang is just like any language without them.

[Khemehekis]

No definiteness. No articles.

My conlang is just like any language without them.

Any language?

What about Japanese? Japanese is without definite and indefinite articles, but the order of nominals -- what's a topic before the "wa" versus what comes after the topic marker and takes "ga" and all that -- works in a way comparable to English "the" and "a".

[lsd]

no article in 3SDL, there is no definiteness mode,
it is the accumulation of qualifiers attached to a term that allows it to be considered definite...
in linear mode it's the anaphora system that enables this accumulation,
which could therefore be considered a definiteness system...

[Omzinesý]

Dlor has a definite article s-.

It is used of object that are mentioned or hinted to earlier in the discourse or that are language-externally known.
It is not used of generic entities.
I don't know how words like Sun behave.

I'm planing that associated motion markers also introduce new entities into the discourse sometimes ('N came and V-ed' though no concrete motion appears). So, it is one way to code indefinite subjects.

[thethief3]

Amarin marks its 1st deixis for specificness rather then definiteness. So specific indefinites will be marked differently from english.

[Creyeditor]

Omlueuet has an indedinite article but not a definite one. Kobardon has inflectional suffixes on nouns that express number and definiteness.

[Snyexarosha]

No definiteness. No articles.

My conlang is just like any language without them.

Agree with Khemehekis here, what do you mean "like any language"? My conlang also has no articles, but still has mechanisms for expressing (in)definiteness (as do many natlangs without articles).

On another note, I love to see other people's conlangs expressing other notions with articles, like specificity! Very cool. Out of curiosity, what do you mean exactly by "specificity"? (It can mean different things in linguistics literature.)

Also out of curiosity--how do your languages all express notions like "someone", "anyone", etc.? I personally have made a single word for "someone"/"anyone", "something"/"anything", etc. Things are easier that way. :)

[Reyzadren]

No definiteness. No articles.

My conlang is just like any language without them.

Any language?

What about Japanese? Japanese is without definite and indefinite articles, but the order of nominals -- what's a topic before the "wa" versus what comes after the topic marker and takes "ga" and all that -- works in a way comparable to English "the" and "a".

That other natlang that I natively speak has no definiteness and no articles. There are no topic markers there, and there is no such thing as doing it with order of nominals.

Moreover, I have encountered a foreign language that has no definiteness and/or articles. Also, I know lots of people who speak languages without them too. I'm quite sure that members here can name many other natlangs that also don't have definiteness/articles.

Agree with Khemehekis here, what do you mean "like any language"? My conlang also has no articles, but still has mechanisms for expressing (in)definiteness (as do many natlangs without articles).

I meant, "like any language that does not have definiteness/articles". If your lang has mechanisms like that, then that means your lang has definiteness within its grammar, but it just doesn't use articles.

[thethief3]

Take the following six examples divided into pairs. The english translations could be the same but in Amarin they would encode for specificity which is whether you well have a specific one in mind. These are demonstratives so are not as often used as articles but can still be used frequently. Also only the deixis 1 demonstratives encode for specificity and not the deixis 2 or 3 ones.

anarama ni ne kili to
give-IMP DAT 1.sg fish DEM.nonsp
“Give me a fish (any fish)”

anarama ni ne kili mee
give-IMP DAT 1.sg fish DEM.spec
“Give me a fish (a specific one)”

ne anaramais ni sua puki to
1.sg give-near.future DAT 2.sg gift DEM.nonsp
“I’ll give you a gift (i don’t have one picked out yet)”

ne anaramais ni sua puki mee
1.sg give-near.future DAT 2.sg gift DEM.spec
“I’ll give you a gift (i have it picked out)”

aki arana samba to
let talk.to-1.sg boss DEM.nonsp
“Lemme talk to the boss (whoever it is)”

aki arana samba mee
let talk.to-1.sg boss DEM.spec
“Lemme talk to the boss (who i know)”

Also see further examples the first one which uses both specific and nonspecific demonstratives and a rather compact translation of an english phrase.
pemiia śapi wii konna
COP.3.sg.speculative DEM.LOC.nonsp DEM.LOC.spec good
“She may be anywhere at the needed time”

śapi śapa
DEM.LOC.nonsp DEM.LOC.3
“Any place, any time”

[DV82LECM]

No "the," but Skasti has an indefinite article me- /me/ (one) and a negative sǫ- /θõ~θo/ (no(t any)). These two do not behave like the other number suffixes, where they have become more attributable to some degree of "definiteness." There apparently are 5 degrees of definiteness.

definite: y̌emóa /geˈmo̯a/ "the tree"
definite plural: ày̌emóa /ˌageˈmo̯a/ "the trees"
indefinite/singular: mèy̌emóa /ˌmegeˈmo̯a/ "a/one tree"
countable plural: èay̌emóa /ˌe̯ageˈmo̯a/ "three trees"
absolute negative: sǫ̀y̌emóa /ˌθõgeˈmo̯a/ "no tree(s)"

No definiteness. No articles.

My conlang is just like any language without them.

Any language?

What about Japanese? Japanese is without definite and indefinite articles, but the order of nominals -- what's a topic before the "wa" versus what comes after the topic marker and takes "ga" and all that -- works in a way comparable to English "the" and "a".

To me, it is all amount semantic focal points. One could assume Japanese's item-counting system is like a slightly-fusional (in)direct object marker plus a countable semantic aspect.

[lurker]

Commonthroat inflects nouns for both personal and spatial deixis, which includes definiteness. The suffixes I have defined so far are:

first person: -l /short low weak grunt/
second person: -qn /huff, short high weak grunt/
third person indefinate (the lemma for nouns): -g /short low weak growl/
third person proximal: -Mr /long low strong grunt, chuff/
third person distal: -p /short high strong grunt/
interrogative: -BD /long rising weak whine/

So the noun sFsFg /yip, long high strong whine, yip, long high strong whine, short low weak growl/, friend inflects as follows:

sFsFl: I, the friend
sFsFqn: you, the friend
sFsFg: a friend, some friend, friends in general
sFsFMr: this friend
sFsFp: that friend
sFsFBD: what friend? Which friend?

[eldin raigmore]

Take the following six examples divided into pairs. The english translations could be the same but in Amarin they would encode for specificity which is whether you well have a specific one in mind. These are demonstratives so are not as often used as articles but can still be used frequently. Also only the deixis 1 demonstratives encode for specificity and not the deixis 2 or 3 ones.

Spoiler:

anarama ni ne kili to
give-IMP DAT 1.sg fish DEM.nonsp
“Give me a fish (any fish)”

anarama ni ne kili mee
give-IMP DAT 1.sg fish DEM.spec
“Give me a fish (a specific one)”

ne anaramais ni sua puki to
1.sg give-near.future DAT 2.sg gift DEM.nonsp
“I’ll give you a gift (i don’t have one picked out yet)”

ne anaramais ni sua puki mee
1.sg give-near.future DAT 2.sg gift DEM.spec
“I’ll give you a gift (i have it picked out)”

aki arana samba to
let talk.to-1.sg boss DEM.nonsp
“Lemme talk to the boss (whoever it is)”

aki arana samba mee
let talk.to-1.sg boss DEM.spec
“Lemme talk to the boss (who i know)”

Also see further examples the first one which uses both specific and nonspecific demonstratives and a rather compact translation of an english phrase.
pemiia śapi wii konna
COP.3.sg.speculative DEM.LOC.nonsp DEM.LOC.spec good
“She may be anywhere at the needed time”

śapi śapa
DEM.LOC.nonsp DEM.LOC.3
“Any place, any time”

I like these!

[Glenn]

Responding to this thread a little late…

Chusole does not have either indefinite or definite articles; however, there are optional means of indicating (in)definiteness. A noun can be marked as indefinite by putting the word “one” or “some” in front of it, and definiteness can be confirmed by the use of demonstratives. (These can take one of two forms: either an independent word preceding the noun, with a three-way distinction (here/there/yonder), or as an enclitic following the noun or noun phrase, which has only a single form. I suspect that the latter is more common in colloquial speech.)

Snyexarosha’s mention of plural marking indicating definiteness also caused me to wonder whether this might be true to some extent in Chusole as well. My current thinking is that plural marking is mandatory in pronouns and optional in nouns, but that animate nouns (especially those referring to human beings) are more likely to be given plural marking than inanimate ones. I am now wondering whether definiteness (and/or specificity) might be a factor as well.

I have a few scattered notes and/or thoughts for at least three other conlangs, none of which have been developed at all. Of these, Ramīyan and Coastlander (Meritskilesi) have definite articles, but probably lack indefinite ones, while Northlander, which was my attempt at a “polysynthetic” language, so far lacks articles, but may be able to indicate (in)definiteness through word order, demonstratives, and the presence or absence of noun incorporation (although my thoughts on the matter are rather contradictory, and may change).

[Imralu]

In Hibuese, definiteness is indicated quite optionally using the singular 3rd person pronoun de or its plural form dwe.

Subject:
It tends not to be used in the subject because subjects tend to be assumed definite. The predicate is generally assumed indefinite, and this results in a cleft-like structure being used for indefinite subjects.

Ngoi zi lu.
ngo=i zi lu
person=COP PST LOC
The person was there.
The people were there

I ngo ye zi lu.
i ngo ye zi lu
COP person ATTR PST LOC
Someone was there.
A person was there.
People were there.
There is a person who was there.
There are people who were there.

When number is indicated, however, both nwa (indicating singular) and we indicating plural, as well as other quantifiers, generally trigger an indefinite reading, meaning that de may be used for explicitly singular, definite subjects (or before other quantifiers) and dwe is used before definite, plural subjects.

De ngoi ho.
The person is good. (Explicitly singular, definite).

Dwe ngoi ho.
The people are good. (Explicitly plural, definite.)

Nwa ngoi ho.
I nwa ngo ye ho.
One person is good. (Indefinite singular.)
Someone is good.
There is one person who is good.

We ngoi ho.
People are good. (Indefinite, generalisation.)

I we ngo ye ho.
Some people are good.
There are some people who are good.

Predicate:
Predicates are usually assumed negative. The 3rd person pronouns de (3S) and dwe (3P) may be used to indicate definiteness.

Nai zi vwe ya mbingu.
na=i zi vwe ya mbingu
1S=COP PST seer GEN dog
I (am someone who) saw a dog.

Nai de zi vwe ya mbingu.
na=i de zi vwe ya mbingu
1S=COP 3S PST seer GEN dog
I'm the one who saw a dog.

Genitive:
Genitive modifiers are generally assumed indefinite. Definiteness can be indicated using the genitive forms of de and dwe which are zyi and zywe respectively.

Nai zi vwe zyi mbingu.
na=i zi vwe zyi mbingu
1S=COP PST seer GEN.3S dog
I saw the dog.

Nai zi vwe zywe mbingu.
na=i zi vwe zywe mbingu
1S=COP PST seer GEN.3P dog
I saw the dogs.

The indefinite quantifiers nwa (SG) and we (PL) also have genitive forms which are mbya and vye respectively.

Nai zi vwe mbya mbingu.
na=i zi vwe mbya mbingu
1S=COP PST seer GEN.SG dog
I saw a dog.

Nai zi vwe vye mbingu.
na=i zi vwe vye mbingu
1S=COP PST seer GEN.PL dog
I saw (some) dogs.

In general, (in)definiteness is only explicitly indicated where it runs counter to the expectation that subjects are definite and predicates and modifiers are indefinite, and is also more commonly indicated when number is indicated at the same time.

Generalisations are often unmarked for number and definiteness (although we is often used in the subject, where it doesn't tend to have the "some" meaning), and genitive modifiers may also be replaced by appositional (i.e. unmarked) modifiers where allowed.

Mbingui mbi.
Dogs are animals.
The dog is an animal.

We mbingui mbi.
Dogs are animals.

Nai le ya mbingu.
I like a dog.
I like dogs.

Nai le mbingu.
I like dogs.
I am a dog lover.

Specificity:
Specific indefinites are optionally indicated with the forms nware (gen.: mbyare) and were (gen.: vyere), essentially just being the indefinite singular and plural markers suffixed with de. They are only used to disambiguate in context where non-specificity is possible, more or less as "certain" and "particular" are used in English.

Nai nguza (ya/mbya) ndyolya.
I'm looking for a book.

Nai nguza mbyare ndyolya.
I'm looking for a particular book.