Message-Id: <m0tJqnW-0000ZUC@xiron.pc.helsinki.fi>
Date:         Sun, 26 Nov 1995 18:46:20 -0500
Reply-To:     Jorge Llambias <jorge@PHYAST.PITT.EDU>
Sender:       Lojban list <LOJBAN@CUVMB.BITNET>
From:         Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject:      Re: self-descriptions?
To:         lojban@cuvmb.cc.columbia.edu, jorge@minerva.phyast.pitt.edu
Content-Length: 2916
Lines: 79

And:
> (a) The only relatively unlojbanic order is V-initial (where V=selbri),
> which wd actually be my favoured order (& Mark's [Culsn]); it's
> disfavoured by the necessity of using {fa} for post-selbri x1.

I agree. I would also prefer that V-initial not be singled out like
that.

> (b) Colin Fine has used V-final a lot. And he's much more responsible than
> I am.

I am not saying that V-final by itself is bad. But if you have several
arguments, some of which with complex substructure, and on top of that
you scramble them away from their "natural" order, then you end up with
something hard to understand. Things like {la and mi di'e spuda} on the
other hand, are perfectly nice.

> (c) V-final is quite common among languages. In Japanese, for instance,
> you get the equivalent of fa-fe-fi sumti in any order, plus final selbri.
> I'm amazed that Japanese people can manage to speak it, but speak it
> they do.

If there is only three arguments it is much more manageable, and also if
the arguments have "cases". Only in the most superficial sense can the
FA-tags be called cases, since by themselves they have no semantic content.

> > (2) the x1 of the main selbri gets kicked out to the fai-place,
> So what?

Supposedly the x1 is the most important, or at least the most frequent.
By kicking it out to limbo the place structure becomes abnormal with
respect to the underlying gismu.

> (And I don't see why it can't be labelled {fa} rather than
> {fai}.)

{fa} labels the new x1. {jai} is like a SE, so the transformed selbri
has its arguments relabeled.

> > and worse of all (3) all the arguments have to be be-bei-linked.
> Well by that reasoning we ought to have a NU for every sumti.

Maybe we ought to, but that's no reason not to take advantage of
what's there already.

> If
>    le nu broda koa koe kei = le jaifau broda be koa bei koe,
> then let
>    le fa'a'a broda koa koe kei = le broda be koa bei koe
>    le fe'e'e koa broda zoe koe kei = le se broda be koa bei koe
> etc. That's actually a good idea.

I agree. How about using the convention:

su'uxipa = fa'a'a
su'uxire = fe'e'e
su'uxici = fi'i'i
su'uxivo = fo'o'o
su'uximu = fu'u'u

> But I don't like the event
> argument being singled out for special treatment.

One reason could be that it is the one that most often requires
the presence of the other arguments, but you will accuse me of
being a functionalist.

> > > What possible advantage offsets these disadvantages?
> A better match between syntax and semantics. If one didn't care
> about that sort of thing then I imagine one would be less fond
> of lojban.

Well, I do care about it, and I do understand your point, but
there is more to it than matching syntax and semantics. The
sentences with {jai fau} are simply too cumbersome for my taste,
because you are forced to use the be-beis. Maybe it is just that
they are unusual, but I think it is more than that.

Jorge