Message-Id: <199402150042.AA14241@eli.CS.YALE.EDU>
Date:         Mon, 14 Feb 1994 19:42:20 EST
Reply-To: jorge@PHYAST.PITT.EDU
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject:      Re: Foot-in-mouth disease (was: Place structures with {co})
To: Erik Rauch <rauch>
Status: RO

la djan di'e cusku

> Only after reading Veijo's post did I see that my last post, in response
> to Jorge, misreads him.  Yes, it is true that "broda brode" has a different
> place structure from "brode co broda", viz. the former has the places of
> "brode", whereas the latter has the places of "broda".

Only for the places that come after the tanru. The places before the tanru
remain always those of brode, the tertanru. This is what I find strange.

> So, in saying that Colin's original claim is wrong, I erred.  "Post in haste,
> repent at leisure."  My explanation, however, as distinct from my correction,
> is still true, and is the same as Veijo's.

I didn't take what you said as saying that Colin was wrong, because he
had said what you said, and what Veijo said. (Too many said's in that
sentence.)


I think I haven't managed to express myself clearly. Let me try again.

I used the form {broda co brode da}, intending for {da} to fill
the x2 of {broda}.

Colin read it as being the x2 of {brode}. When he explained it to me,
I remembered that this is how it is canonically interpreted. (I had
read "Pretty Little Girls' School",  :)

I thought about it, and now find the canonical use (which I do
understand, so please do not explain it once again, this was not my
problem) slightly illogical.

This is what I find illogical: the places of a {co} tanru are not those
of the tertanru, like for any old tanru, but a mixture of the places
of the tertanru and those of the rightmost seltanru: namely, the
preceding sumti fall in the tertanru places, and the trailing sumti
fall in the seltanru places (starting with x2, no matter how many sumti
were in front, I presume).

I also find that this restricts unnecessarily the use of the tertanru
places, without a big gain, since the places of the seltanru would be
equally available with simple {be} and {bei}s.

I don't mind if the interpretation is left as is, but then we should not
claim that this permits modificand-modifier order with a simple particle,
because it doesn't: the tanru with {co} is more limited than the ordinary
tanru, because the modificand loses all its trailing places.

> I would add that Ivan's research seems to show that asymmetrical tanru,
> which are the only ones where "co" is relevant, are in
> the great majority, both of tanru-types and of instances; and that lojbab
> has said that symmetrical tanru without a connective are deprecated.

Yes, let the symmetric ones be deprecated. But for the interesting ones,
the asymmetrical, I can't use Spanish word order. (I don't mind much, as
long as we are clear that this is so.)

Jorge