[Lingtyp] differential argument marking conditioned by modification

[matic at uni-muenster.de]

Dear typologists,

Has any of you ever seen a case of differential argument marking conditioned by whether the argument is modified (i.e. has an attributive modifier or some other adnominal dependent) or not?

I have this in Tundra Yukaghir. The story goes roughly as follows: If I want to say "I see a reindeer", "reindeer" will take the case suffix A; if I say "I see a good reindeer", "reindeer" takes the case suffix B. In both instances, the argument behaves syntactically the same, the only difference between the two types of case forms being the presence vs. absence of an attributive modifier.

I have been wondering (a) if any of you know of similar cases in any other language, and (b) how to name and gloss the case form that is used only with modified arguments.

(I am aware that there is a similar situation in the only relative of Tundra Yukaghir, Kolyma Yukaghir, and that there have been some analyses based on these data, but that's not what I am asking about.)

Thank you for your answers!

Best,
Dejan

-- 
Prof. Dr. Dejan Matic
Institut für Sprachwissenschaft
Westfälische Wilhelms-Universität Münster
Aegidiistr. 5
48143 Münster
Germany
tel. +49-251-8324494