Accusative/Ergative/Semantic (was Re: Triggeriness)
|From:||Tim May <butsuri@...>|
|Date:||Wednesday, December 17, 2003, 3:26|
Andreas Johansson wrote at 2003-12-16 23:41:52 (+0100)
> Yea, verily, and English, Basque and Tagalog are all active
> languages in denial!
> (Seriously, how _do_ explain an accusative system under a Holy
> Agent-Patient Verbal System? An article on semantic roles I read
> some year ago stated flatly that it's impossible, if you mean what
> I think you mean by it.)
At a tangent here, something which struck me a while back... it's
quite obvious really, but I handn't really thought about it before, so
it seemed interesting at the time.
While the grammatical relations Nominative, Accusative, Ergative and
Absolutive are quite clearly not semantic matters of agent vs patient
- they're general-purpose categories for the argument structure of the
verb - the ability to tell the difference between an accusative system
and an ergative system _is_ semantic. If you can't tell which of the
arguments of a transitive verb is more "agent-like", it's impossible to
say which it is.
Well, I said it was obvious. But this means you could in principle
have a system (case, say) that's just like an ergative or accusative
system, but isn't either. In that of the two arguments of a
transitive verb one patterns with the argument of an intransitive and
the other doesn't, and this is part of the lexical definition of each
verb, but there's no correlation between these two formal roles and
the semantic agentiveness of the arguments.
This just struck me as a weird and interesting idea, at the time.