Re: Monster Raving Loony (was: Re: Ergativity)
| From: | Paul Bennett <paul-bennett@...> |
|---|
| Date: | Sunday, August 22, 2004, 19:40 |
|---|
On Sun, 22 Aug 2004 20:50:45 +0200, Carsten Becker
<naranoieati@...> wrote:
> Digging in my archive ...
>
> On Wed 25 February 2004 10:36, Andreas Johansson wrote:
>
> > Quoting "Mark J. Reed" <markjreed@MAIL.COM
> <mailto:markjreed@...>>:
> > [...]
> > > Construct Categories
> > > Name Groupings
> > > [...]
> > > Monster Raving Loony {A, P} {S}
> > > [...]
> >
> > It should also be possible to have a hierarchical system
> > which is morphologically and syntactically MRL. Given the
> > usual result of me expressing a wish that some feature or
> > pattern is absent from natlangs, I'm not going to say I
> > do hope no natlang is constructed thusly.
>
> Now I don't understand anything anymore. How can MRL
> languages work at all? The agent cannot be the patient at
> the same time! And why is the subject separate?
Nonononono. You misunderstand. The point of MRL languages is not that
something is the agent and the patient at the same time, but rather that:
In intransitive sentences, the subject is marked in case A.
In transitive sentences, both the patient and the agent are marked in case
B.
This is not always entirely looney, since word order, verbal agreement,
and other factors can differentiate, but it *is* true that there are
natlangs whereby context is the only possible disambiguator.
Paul
Reply