In head-marking languages, the head usually takes the marking based on the dependent's intrinsic properties. For example, every English noun/pronoun has an intrinsic property of person: "bear" is always 3rd person, and this characteristic is wired into the word itself. When a verb governs the word "bear" as a subject, it kinda reads its person property, and then becomes head-marked with /-s/ that essentially means "my dependent is 3rd person subject". The same applies to head-marking in possessive constructions: in most languages with head-marked possessives, the head noun takes a prefix based on the person property of the dependent (e.g., "my dependent possessor is second-person")

Meanwhile, dependent marking generally works in a completely different way. Instead of the dependent agreeing with some property of the head, the head lisences a slot for a dependent marked in specific way (a case marking). Basically, the head "says": "I can only govern an object / a possessor if it's marked with an accusative/genitive case, and I won't accept it as my dependent unless it is". All verbs demand accusative of objects, and all nouns demand genitive of possessors: the dependent-marking doesn't tell you anything about the head.

I wonder: are there any languages where the dependent is marked based on an intrinsic property of the head rather than based on the case marking lisenced by the head? Say, in some languages, verbs are divided into two subclasses: stative and active. When a noun becomes the subject of a stative verb, it must take an /-s/ ending that correspond to the stative property of the head, essentially saying "my head isn't any verb, but a stative verb" (and some other ending when governed by active verbs). Something like that. Or perhaps there are languages with different verb classes that force the nouns to take different endings.

So far, the only example I know that qualifies is the adjective-noun agreement in many Indo-European languages: where an adjective attributing a feminine gender noun must also take a feminine ending. I wonder if there are other examples: in particular, I'm interested in the verb-noun relationship.

  • UPDATE: I see that suffixaufnahme is an example of this for possessive constructions, which works similar to adjectival agreement.
    – Slavus
    Apr 29, 2023 at 23:01
  • 2
    In Swahili, the analog of English ‘of’ is the genitive particle “-a” which agrees in the noun class with the possessor noun by attaching a prefix which corresponds to the possessor noun class, but not necessarily coinciding with it: watu wa Tanzania “people of Tanzania”, miti ya msitu “trees of the forest”, vitu vya Bwana Smith “things of Mr. Smith”.
    – Yellow Sky
    Apr 30, 2023 at 12:01
  • 1
    Your stative/active verb example sounds a lot like how ergativity works (subject of intransitive verb = absolutive case; subject of transitive verb = ergative case). But how is that ultimately different from dependent marking, with the stative verb saying, “I can only govern a subject if it takes the /-s/ ending”? In Icelandic, objects are dative with verbs meaning ‘throw’, but accusative with verbs meaning ‘catch’, so is the object reflecting a property of the head, or does each verb have specific “I will only govern an object if” rules? Apr 30, 2023 at 12:59
  • @JanusBahsJacquet, the difference is somewhat cosmetic, but still. 1) The ABS ending doesn't tell you anything about the verb per se. If you see a word "wolf" in ABS, you can't tell anything about the verb from it alone, even transitivity: it could be either the sole actant of an intr verb or the obj of a trans verb. 2) The case of the agent can actually dictate the transitivity of an ambitransitive verb. So in the end, ABS is about sentence structure, not the verb category. The verb doesn't determine the cases - not alone at least.
    – Slavus
    Apr 30, 2023 at 22:45
  • @JanusBahsJacquet, as for your 2nd example, the issue is specificity. I don't know Icelandic, but if it works roughly as English, DAT can introduce the obj of the Icelandic verb "throw". It can also introduce actual dative of "give". Also benefactives (I assume) for pretty much any verb (which aren't even proper actants). And even args of nouns ("a gift for you"). As I said, the difference is cosmetic, but to me the non-specificity of DAT shows that it's not about agreement, but role. Some heads lisence roles for DAT, but in the end DAT tells you nothing about the head like agreement does.
    – Slavus
    Apr 30, 2023 at 22:57


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.