Re: THEORY: Verb-Medial Languages, Case-Marking and Agreement

>
> But, if that were so, wouldn't SVO languages case-mark the subject and
> have
> the verb agree with the object? But mostly they don't; rather, they tend
> to
> have the verb agree with the subject (if anything), and to case-mark only
> the
> object (if either grammatical relation).

I think this is sort of crazy-logic. If we're to accept the idea that
languages tend to establish roles quickly (which I am dubious of), then
that's fine, but I don't suppose that we should posit a unitary idealised
implementation. It seems to me that one SVO language I know (such as
English) are pretty efficient at assigning roles without the need for
morphology to do so; word order serves this purpose instead. Regardez:

Anna sees a frog.

ANNA is assigned to subject role by position. SEES calls for something to be
seen, so when we encounter FROG we havea role to slot it into.

That's awfully convenient, but it works for verbs with less strict
assignments too:

Anna supposes a frog into the pond.

This is clearly wrong! There is no role assigned previously that "into the
pond" latches onto appropriately. I imagine then that the answer to your
question is, "Historical accident." Given all the marking and agreement
tools available to languages, it doesn't make a lot of sense to ask, "Why
didn't extremely specific solution A arise in response to problem B?"

--
Shreyas