Languages that mark case, but not on nouns?

[xBlackHeartx]

German has an unusual form of case marking. It has four cases (with a fair amount of syncretism), but only the genitive is ever really marked on nouns. Normally, the case marking falls on the article. In absence of that, it falls on the adjective. Pronouns also inflect for case (personal pronouns though are highly irregular). So, what happens if you have a noun that has no article or adjective? Well, then it just gets no marking. This is what happens with proper nouns all the time, though other things like uncountable nouns and nouns with numbers also go unmarked.

Thinking about it, if a language with this type of indirect case marking was to mark case on everything, it would end up rather complicated rather fast. Proper and uncountable nouns would have to receive articles or some kind of particle of their own. Worse yet, all the possible numbers would have to have their own forms! That alone could, obviously, get out of hand rather quickly. I don't think I've ever even heard of a natural language that marks case on number (with the possible exception of Russian, but only the numbers one, two, and three can take case).

Another weird thing is I've never heard of any natlang other than German that does this. Yeah, there's languages that use isolating particles, but I'm talking about inflectional marking here. Definite articles, indefinite articles, demonstratives, and sometimes adjectives are what take the case marking in a noun phrase. You simply can't have case without these (besides the genitive anyway, but the use of that has been diminishing, but truth be told, I don't think I've ever actually seen an isolated genitive noun in a sentence).

Are there any other languages with this strange indirect method for marking case? If so, are there languages where the marking is far more reliable? As I said, German nouns often just have to go without case marking since you can't actually mark any cases (besides the genitive, and the dative in a few dialects) directly on nouns; you can only do it on other elements in a noun phrase.

[linguoboy]

Proper and uncountable nouns would have to receive articles or some kind of particle of their own.

In fact, this is exactly what happens in colloquial German, where personal names routinely take definite articles, e.g. "Und sag dem Hans, die Huber Anni sucht ihn."

Worse yet, all the possible numbers would have to have their own forms! That alone could, obviously, get out of hand rather quickly. I don't think I've ever even heard of a natural language that marks case on number (with the possible exception of Russian, but only the numbers one, two, and three can take case).

Actually, this was a common Indo-European feature. It persisted into the 18th century in Standard German and is still preserved in several dialects. (Siehe hier: https://de.wikipedia.org/wiki/Zwei#Kardinalzahl.)

[awrui]

Worse yet, all the possible numbers would have to have their own forms! That alone could, obviously, get out of hand rather quickly. I don't think I've ever even heard of a natural language that marks case on number (with the possible exception of Russian, but only the numbers one, two, and three can take case).

Actually, this was a common Indo-European feature. It persisted into the 18th century in Standard German and is still preserved in several dialects. (Siehe hier: https://de.wikipedia.org/wiki/Zwei#Kardinalzahl.)

I consider myself dialect-free, but I also do that sometimes...
Some Uralic languages do that, too.

[Linguaphile]

I don't think I've ever even heard of a natural language that marks case on number (with the possible exception of Russian, but only the numbers one, two, and three can take case).

Actually, this was a common Indo-European feature. It persisted into the 18th century in Standard German and is still preserved in several dialects. (Siehe hier: https://de.wikipedia.org/wiki/Zwei#Kardinalzahl.)

I consider myself dialect-free, but I also do that sometimes...
Some Uralic languages do that, too.

Yes, Uralic languages do this. For Estonian, not only does it do this, but many of the numbers belong to one of the word classes that have two sets of plural forms (the so-called i-plural and the regular one), which doubles the number of possible plural case forms.

It persisted into the 18th century in Standard German and is still preserved in several dialects. (Siehe hier: https://de.wikipedia.org/wiki/Zwei#Kardinalzahl.)

In German does this happen only with the number two, or with other numbers as well? (Asking since the Wikipedia article is specifically about the number two.)

Worse yet, all the possible numbers would have to have their own forms! That alone could, obviously, get out of hand rather quickly.

Yes, they do; and yes, it does. (No more so than any other word, though.)
Let's take a random number, 18, in Estonian. These are the possible forms. (And these are just the cardinal numbers. Ordinal numbers also have their own set of forms....) I can't really think of many practical uses for a few of these forms, but they're real words - if the need to say "into the eighteens" is actually out there, there's a declension for that (actually, two of them: kaheksateistkümneisse and kaheksateistkümnetesse).

kaheksateist = singular nominative short form
kaheksateistkümmend = singular nominative long form
kaheksateistkümned = plural nominative
kaheksateistkümne = singular genitive
kaheksateistkümnete = plural genitive
kaheksatteist = singular partitive short form
kaheksatteistkümmend = first singular partitive long form
kaheksatteistkümmet = second singular partitive long form
kaheksateistkümneid = plural partitive
kaheksateistkümnesse = singular illative
kaheksateistkümneisse = plural illative short form
kaheksateistkümnetesse = plural illative long form
kaheksateistkümnes = singular inessive
kaheksateistkümneis = plural inessive short form
kaheksateistkümnetes = plural inessive long form
kaheksateistkümnest = singular elative
kaheksateistkümneist = plural elative short form
kaheksateistkümnetest = plural elative long form
kaheksateistkümnele = singular allative
kaheksateistkümneile = plural allative short form
kaheksateistkümnetele = plural allative long form
kaheksateistkümnel = singular adessive
kaheksateistkümneil = plural adessive short form
kaheksateistkümnetel = plural adessive long form
kaheksateistkümnelt = singular ablative
kaheksateistkümneilt = plural ablative short form
kaheksateistkümnetelt = plural ablative long form
kaheksateistkümneks = singular translative
kaheksateistkümneiks = plural translative short form
kaheksateistkümneteks = plural translative long form
kaheksateistkümneni = singular terminative
kaheksateistkümneini = plural terminative short form
kaheksateistkümneteni = plural terminative long form
kaheksateistkümnena = singular essive
kaheksateistkümneina = plural essive short form
kaheksateistkümnetena = plural essive long form
kaheksateistkümneta = singular abessive
kaheksateistkümneteta = plural abessive
kaheksateistkümnega = singular comitative
kaheksateistkümnetega = plural comitative

An article with another example: 75 Ways to Say 80; Estonian numeral examples

[linguoboy]

It persisted into the 18th century in Standard German and is still preserved in several dialects. (Siehe hier: https://de.wikipedia.org/wiki/Zwei#Kardinalzahl.)

In German does this happen only with the number two, or with other numbers as well? (Asking since the Wikipedia article is specifically about the number two.)

"Three" has the forms dreier (genitive plural) and dreien (dative plural), both of which can be found in contemporary formal German.

In Alemannic varieties, these inflect not for case but for gender, e.g. Züritüütsch (Zurich German):

zwee Mane "two men"
zwoo Fraue "two women"
zwäi Chind "two children"

drei Mane/Fraue "three men/women"
drüü Chind "three children"

[xBlackHeartx]

For the record, I've never heard of this in German, though I'm only really familiar with the 'standard' variety everyone teaches (and the one you normally see in popular media). I've also never seen anyone use a definite article with a proper noun in German, though I have heard Greek does do this.

A major problem I see is there isn't always a reason to use an article or some other 'modifier' (for lack of a better term). For instance, in the phrase 'Hans drinks milk', what article do you use with that? I think the biggest problem with sentences like that is its technically in the gnomic mood, and so you're not referring to any 'milk' in particular, or a group of 'milk' (so using 'some' which just seem strange).

Mainly, I was wanting to know if there's any other languages with this obviously strange feature. I've never heard of any language that does. The closest I've seen are languages like Basque which puts its case markings on whatever happens to be the last element in a noun phrase. Of course, case markings CAN fall on nouns in that language, so I don't really see Basque as counting.

[h34]

I've also never seen anyone use a definite article with a proper noun in German, though I have heard Greek does do this.

It is quite typical for colloquial German (as Linguoboy mentioned) but mainly in Southern Germany, Austria and Switzerland. In Northern Germany we generally don't use definite articles before proper names.

[kevin]

A major problem I see is there isn't always a reason to use an article or some other 'modifier' (for lack of a better term). For instance, in the phrase 'Hans drinks milk', what article do you use with that?

But where is the problem? Other languages like English get by without marking case at all. How could marking case "only" in almost all cases be more ambiguous or problematic than not marking case at all?

"Hans trinkt Milch" is just as easy to understand as "Hans drinks milk". In fact, context is enough that playing with the word order is possible evem though cases aren't marked. So "Milch trinkt Hans" would be understood just fine, or even "Es trinkt Hans Milch" or "Es trinkt Milch Hans" (though the last one sounds really strange, and in fact I really want to say "der Hans" in the last two, as non-standard as it may be).

[xBlackHeartx]

A major problem I see is there isn't always a reason to use an article or some other 'modifier' (for lack of a better term). For instance, in the phrase 'Hans drinks milk', what article do you use with that?

But where is the problem? Other languages like English get by without marking case at all. How could marking case "only" in almost all cases be more ambiguous or problematic than not marking case at all?

Technically no, but it could be a problem if you want free word order. I was mainly looking for a way to have cases in a conlang without having nouns take case endings. For some reason, I just never liked that. I am, however, used to German's system, weil ich Deutsch spreche.

[kevin]

Technically no, but it could be a problem if you want free word order.

Theoretically yes, but I don't think it matters much in practice. Your example where neither subject nor object have an article is already kind of exceptional, and context will usually clarify the intended meaning anyway. I think I mentioned all word order permutations allowed by German grammar, and while technically that makes the sentences ambiguous, milk just doesn't drink persons, but it's always the other way round.

I was mainly looking for a way to have cases in a conlang without having nouns take case endings.

If you really want to mark case without exception, I think I would just add articles to proper names then. For uncountable nouns like "milk" in your example, you could either reuse the indefinite article (that I assume you already have) or introduce a third type of articles.

[vijayjohn]

In the case of that particular example in German, maybe the animacy hierarchy would also help (milk is not animate, therefore it's surely not the milk that's drinking Hans).

[Synalepha]

Italian and I think other Romance languages too are developing quite an interesting strategy. I asked about this on reddit back when I used to use it and someone told be it was called object affixation or something like that (I really can't remember) and it can be found in Swahili and other Bantu languages. In Romance languages this feature seems to be in its embryonal phase and it's still debated by linguists (at least this is what I was told).

So basically, the way it works is: you use the object pronoun in combination with the noun in order to signal that the noun in question is the object of the sentence, by doing so, you can have a freer order even without cases. Hopefully, to make things a bit clearer, here's an example:

a) (io) prendo l'ombrello. - (I) take the umbrella
b) l'ombrello lo prendo (io). - The umbrella-it take (I)

[linguoboy]

Italian and I think other Romance languages too are developing quite an interesting strategy. I asked about this on reddit back when I used to use it and someone told be it was called object affixation or something like that (I really can't remember) and it can be found in Swahili and other Bantu languages. In Romance languages this feature seems to be in its embryonal phase and it's still debated by linguists (at least this is what I was told).

The term for this feature as exhibited by Bantu languages is "polypersonal agreement", as the verb complex shows agreement for all persons, and not just subject/agent.

The argument you mention is that the pronominal clitics of Romance are in the process of becoming bound morphemes comparable to the pronominal affixes of polypersonal languages. I've never heard this argument made for standard Romance languages other than French, as they seem to exhibit much less use of pleonastic clitics than colloquial French.

[Synalepha]

Italian and I think other Romance languages too are developing quite an interesting strategy. I asked about this on reddit back when I used to use it and someone told be it was called object affixation or something like that (I really can't remember) and it can be found in Swahili and other Bantu languages. In Romance languages this feature seems to be in its embryonal phase and it's still debated by linguists (at least this is what I was told).

The term for this feature as exhibited by Bantu languages is "polypersonal agreement", as the verb complex shows agreement for all persons, and not just subject/agent.

The argument you mention is that the pronominal clitics of Romance are in the process of becoming bound morphemes comparable to the pronominal affixes of polypersonal languages. I've never heard this argument made for standard Romance languages other than French, as they seem to exhibit much less use of pleonastic clitics than colloquial French.

In Italian it's relatively common to use OV(S) or V(S)O orders if you insert an object pronoun before the verb. It's not always possible though and I'm not sure I would be able to infere the exact rules without spending an insane amount of time on this task. This definitely has a lot to do with topicalization (I think that's the right word), but nowadays there are instances where (S)VO, OV(S) and V(S)O sound interchangeable to me (as long as you add the object pronoun to OV(S) and V(S)O). For example, to say "I take the umbrella", there are equal chances I might say "prendo l'ombrello", "l'ombrello lo prendo" or "lo prendo l'ombrello".

The same goes for partitive ne and locative ci.