----- Original Message -----
From: "Andreas Johansson" <andjo@...>
To: <CONLANG@...>
Sent: Thursday, August 07, 2003 1:15 PM
Subject: Re: Help in Determining Asha'ille Typology
> Quoting Joe <joe@...>:
>
> > ----- Original Message -----
> > From: "Andreas Johansson" <andjo@...>
> > To: <CONLANG@...>
> > Sent: Thursday, August 07, 2003 10:47 AM
> > Subject: Re: Help in Determining Asha'ille Typology
> >
> >
> > > Quoting "Thomas R. Wier" <trwier@...>:
> > >
> > > > Quoting Andreas Johansson <andjo@...>:
> > > >
> > > > > > 1) I eat food.
> > > > > > 2) I run.
> > > > > > 3) I fall.
> > > >
> > > > [...]
> > > >
> > > > > What would we call a language that marks "I" from (1) the same as
"I"
> > in
> > > > > (3), and "I" in (2) the same as "food" in (1)? Beyond weird, that
is.
> > > >
> > > > This would still be a split-S language. Split-S languages are
> > > > defined, in contrast to fluid-S languages, by the fact that verbs
> > > > simply subcategorize for whether the single argument patterns as
> > > > the NP-1 of transitives or NP-2 of transitives. It is also
> > > > characteristic of such languages that many verbs take the unexpected
> > > > marking, such as patientive for run or agentive for fall.
> > >
> > > I didn't state my question clearly enough to exclude the possibility
of a
> > > language with semantic marking. You could have language that uses
> > >
> > > I:PAT fall
> > >
> > > for a voluntary falling and
> > >
> > > I:AGT fall
> > >
> > > for involuntary. (Case names assigned to make sense with transitives.)
> >
> > Wouldn't it be the other way round? I mean, as far as I can gather, the
> > patient is the thing that something happens to, and the agent is what
causes
> > it to happed.
> >
> > So, for instance
> >
> > I.PAT fall
> >
> > could be said to mean 'I am being caused to fall', whereas
> >
> > I.AGT fall
> >
> > would be said to mean 'I am causing myself to fall'.
> >
> > If the other way around, using the terms 'Agent' and 'Patient', I think,
> > would be unsatisfactory.
>
> But if we switched the labels PAT and AGT, the hypothetical language would
> express "I eat food" as
>
> I:PAT eat food:AGT
>
> which to my mind is even less satisfactory. One could, of course, simply
call
> the cases "1" and "2", if one felt like.
>
Hmm? I'm sorry, I didn't say switch cases round in every case, merely in
the example you gave.
> Andreas
>