Pardon my word choice, since I'm obviously lacking the background in linguistics. I know that language-specific questions are off-topic, yet I still like to use Finnish as an example, since it spawned my interest in this question, however I think this is general enough to warrant this query. Anyhow:

The German sentences

Ich finde einen Baum. (I find a tree) Ich finde Wasser. (I find water)

take the accusative case of the object (einen Baum) and (Wasser, although here with the Nullartikel).

The Finnish language does know about the accusative, however - apart from personal pronouns - it does not have an extra declension. For that reason, the accusative is usually called "nominative-accusative" or "genitive-accusative". In Finnish the aforementioned sentences translate to

Löydän puun. Löydän vettä.

Puun (tree) is written the same as the genitive declension of puu (puun, tree's). The critical thing to note here is, that water takes the partitive case (I'd say somewhat similar to the Nullartikel above, since water is not really a measurable quantity) instead of the genitive version (veden). Does vettä here still function as a accusative object although it's written in the partitive case? Could you call it a "paritive-accusative", if that even makes sense?

  • 1
    Within the context of Finnish, no, you wouldn’t, and it wouldn’t make sense. Finnish objects can be either telic or atelic, and in traditional Finnistic terms, accusative refers specifically to a telic object (that is, a personal pronoun in the accusative or a noun phrase in the genitive singular or nominative plural), while partitive refers to atelic objects. So a ‘partitive accusative’ does not make sense, since that would be a telic atelic object. But this distinction/terminology is specific to Finnish (and perhaps related languages), not a universal thing. Nov 25, 2021 at 21:05
  • Also, while measurability does interact with both Nullartikeln in German and telicity in Finnish, they are not the same – the telic variant löydän veden ‘I find the water’ is also perfectly grammatical. Nov 25, 2021 at 21:07
  • @JanusBahsJacquet: As for related languages, I can say that Hungarian, at least, has a straightforward accusative case (and no partitive case).
    – TonyK
    Nov 25, 2021 at 22:15
  • 1
    IIRC, the genitive-accusative historically comes from a real accusative that also had a vowel + N suffix, and they merged. The partitive has existed before as a separate case (with ablative meaning, as you can still observe in the partitive/essive/translative postposition series). Nov 26, 2021 at 7:37
  • 1
    @infinitezero Yes, exactly – a telic object, which in this case would align with a definite form in German and English. In other cases, though, this alignment would not apply; e.g., luen kirjaa/kirjan could both be translated as either ‘I read a book’ and ‘I read the book’, depending on context. Nov 26, 2021 at 14:00

3 Answers 3


I think broadly speaking, yes. Forgive a misinterpretation here as I am unfamiliar with Finnish and only have a cursory knowledge of German, but it seems to me that there is a confusion here between the morphological construction of vettä (which may have some kind of semantic weight, I'm still not entirely sure as I haven't done much reading on the partitive), and the syntactic role of vettä in the sentence

Löydän vettä.

In this case, the word vettä is taking the inflections typical of a noun in the partitive. However, when analyzing the sentence, it is clear that vettä is the direct object (or patient). There is a mismatch between these two ideas, but we see mismatches like this in languages all the time.

In Chukchi, a Chukotko-Kamchatkan language spoken in northeast Siberia, there is no dedicated inflection for the ergative outside of pronouns. When in situations that would call for the ergative, most nouns take the instrumental case. For example, take the sentence

ətɬəɣe ekkin waɬə pənenin. (The father sharpened the son's knife)

I haven't bothered with a gloss because the verb is very complex, but the point is ətɬəɣe, father, has taken the instrumental case ending -e. Proper nouns such as names would take the locative case, so if we were to replace ətɬəɣ- with Rintə- (a name), it would be

Rintəne ekkin waɬə pənenin.

So in Chukchi, as in Finnish, different morphological cases (that is to say, the patterns nouns can take) are used for the same syntactic case (that is to say, the function of the noun in the sentence), and in both situations there is mismatch between morphological and syntactic case. The paper I've cited below (Spencer, 2006) goes into more detail about this and provides two more examples.

Ultimately, the reasoning behind this is the fact that most of the language we use to talk about grammatical categories is based off of old grammars of Latin and Greek, two languages where morphological case and syntactic case are effectively the same thing. It never occurred to these grammarians that these two concepts aren't necessarily always linked together, and it is useful in modern linguistics to understand how they differ from one another.

In regard to vettä, I would say it is in the partitive, rather than a supposed partitive-accusative or something of that kind.

Spencer, A., 2006. Syntactic vs. morphological case. Case, valency and transitivity, 77, p.3. Link

Vinyar, A.I., Kazakova, P.N. and Naletova, P.R., 2017. Chukchi Denominal Verb Construction: Overview and Relation to Noun Incorporation. Higher School of Economics Research Paper No. WP BRP, 58. Link

  • Hey nice answer, by the last part confuses me: "In regard to vettä, I would say it is in the partitive." Do you mean, you would say it's in the partitive taking an accusative role? Nov 25, 2021 at 19:30
  • @infinitezero yeah, I just mean I would call it partitive rather than partitive-accusative or some other label like that. (Edited question to show this) Nov 25, 2021 at 22:26

Case (including Finnish partitive) is a morphological form, and object is a grammatical function (if we set aside the theory of abstract case). I would say that partitive objects are partitive objects, and not accusative.

Usually, accusative is the term used for a case which is (i) distinct from the nominative, and (ii) typical for objects in simple transitive clauses (where nominative is the term used for a case which is used for subjects of transitive and intransitive clauses; an ergative pattern is different, because the ergative case on transitive subjects is different from the absolutive case on intransitive subjects). Once you have used this kind of rule of thumb to decide which case to call “accusative”, then morphology dictates which noun phrases are in that case, not function. Thus a language might allow an accusative measure phrase or an accusative subject (e.g., in Icelandic), and you can have a dative or genitive object (e.g., in Russian, or again in Icelandic).

Since the partitive is not the typical case for objects in Finnish, I don’t think there is any utility in calling partitive objects “accusative” -- it seems to be enough to say that they appear in the partitive case, and that they are objects.

Of course, Finnish does have the interesting complication you mentioned that it doesn't have a uniquely identifiable accusative for lexical nouns (only for pronouns), so that there is an analytic question of whether typical objects are genitive or nominative, or whether the accusative is syncretic with one or the other or both (as discussed by Kiparsky in 2001 in Lingua in 'Structural Case in Finnish'). But whatever solution you adopt for that complication, I wouldn’t think it would justify calling a partitive accusative.


Additionally, I can give you examples in Russian that pretty similar to the Finnish one:

Ya nashel derevo (accusative). Ya nashel vodu (accusative).


Ya nashel derevo (accusative). Ya nashel vody (partitive).

For much instances, they are differentiated stylistically.

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.