From @YaleVM.YCC.YALE.EDU:LOJBAN@CUVMB.BITNET  Sat Feb 12 13:41:13 1994
Received: from ELI.CS.YALE.EDU by NEBULA.SYSTEMSZ.CS.YALE.EDU via SMTP; Sat, 12 Feb 1994 13:24:12 -0500
Received: from YALEVM.YCC.YALE.EDU by eli.CS.YALE.EDU via SMTP; Sat, 12 Feb 1994 13:24:07 -0500
Message-Id: <199402121824.AA01034@eli.CS.YALE.EDU>
Received: from CUVMB.CC.COLUMBIA.EDU by YaleVM.YCC.Yale.Edu (IBM VM SMTP V2R2)    with BSMTP id 6287; Sat, 12 Feb 94 13:22:18 EST
Received: from CUVMB.COLUMBIA.EDU by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with  BSMTP id 6896; Sat, 12 Feb 94 04:41:06 EDT
Date:         Sat, 12 Feb 1994 11:41:13 +0200
Reply-To: veion@XIRON.PC.HELSINKI.FI
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Subject:      [Re] Place structures with co
X-To:         lojban@cuvmb.cc.columbia.edu
To: Erik Rauch <rauch>
Status: RO
X-Status: 

Date:         Fri, 11 Feb 1994 16:42:27 EST
From:         Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject:      Place structures with {co}

> la kolin cusku lu
....
>> I cannot think of a way of inverting
>> broda brode da
>> as you are trying to do.
>>
>> brode co broda da
>> expressly means
>> broda be da co brode
>
> I guess you mean
> broda be da brode
>
> This is not the idea that I get from looking at the BNF,
> which has the trailing sumti attaching to the whole bridi,
>
>        brode co broda da
>
> parses as:
>
>        ({brode co broda} {da VAU})
>
> Which suggests to me something different from:
>
>        ({brode co <broda [be da BE'O]>} VAU)
>

It has been noted before that we mustn't take the parses too literally :-)

> If {brode co broda da} doesn't mean {broda brode da}, then
> the claim that {co} is there to permit the modifier to come
> after the modificand (is this the right word?), as in some
> languages (like Spanish :) is not quite true, because using
> it restricts what we can say with that tanru. (A lot of places
> become inaccessible, unless used before the selbri.)
>
> So {broda brode} has a different place structure
> than {brode co broda}. (If a tanru can be said to have a
> place structure.)
>
> I'm not saying it's wrong, just that it's a bit strange.
>
> Jorge

co is a mechanism to avoid infix sumti. Let a1, a2, ... be sumti of broda
and e1, e2, ... sumti of brode. Then for a tanru broda brode we have

    e1 broda brode e2 e3 ...  = e1 e2 e3 broda brode

Now if we want to specify a2 we must attach it infixed

    e1 broda be a2 [be'o] brode e2 e3

Sometimes we are not at all interested in e2, e3,... but would like to
specify (a1), a2, a3, ... (brode modified by broda [fa a1] a2 a3).
We do this by using co:

    e1 brode co broda [fa a1] a2 a3 ...

i.e. the purpose of co is to make the sumti places of the modifier more
easily accessible.

This is especially valuable in cases where e.g. a2 is so complex that
infixing it would make the tanru unrecognizable. In those cases co is
sometimes preferable even if we want to specify e2:

    e1 brode be e2 [be'o] co broda a2 a3

This, of course, is only necessary if the tanru is asymmetric, i.e.
broda brode != brode broda, and we want to have brode as the basic
relation.


co'o mi'e veion