Consider a system whereby any VP with exclusively third person arguments form your average nominative-accusative (or even erg-abs or whatever) structure. However, whenever a first or second person is involved, the system is inverse instead. Thus, the inverse and the direct markers also are, in some sense, a person marker: they signal the presence of at least a first- or second-person participant in the VP.
However, let's imagine further that even the presence of an indirect object in the first or second person triggers this. Let's, however, still have the inverse alignment be the only role-assigning marker even in that case.
Cut off?
ReplyDeleteHuh? No, I actually like to leave thoughts hanging in exactly that way - i.e. the reader must figure out the implications to some extent. I was hoping the reader would figure out that e.g. 'She gave me the documents' would have three arguments - give(she, I, documents, inverse): inverse, because "I" am not the subject (she is), and inverse, rather than 3sg because "I" am involved. However, the listener has to assign she and documents the correct roles by guesswork. With 'she gave me the documents' this is pretty easy - but imagine 'he provided me a contact'; here, we only know that 'me' is not the subject (inverse) - but he and contact - either the contact provided me him, or he provided me a contact. Who knows, at least without contextual knowledge.
DeleteSo, I extrapolated correctly; I didn't understand why you found it interesting. This is equivalent to a system which marks 1st and 2nd persons either nominative or oblique, and 3rd persons not at all. (Well, you omitted what happens with e.g. a 2p subject and a 1p in the VP. I should expect an animacy hierarchy in that case.) To my understanding, this is just a standard direct-inverse system conditioned on nominative–oblique rather than nominative–accusative.
Delete