Message-Id: <199602121519.RAA08261@xiron.pc.helsinki.fi>
Date:         Mon, 12 Feb 1996 16:11:43 +0100
Reply-To: Goran Topic <topic@STUDENT.MATH.HR>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: Goran Topic <topic@STUDENT.MATH.HR>
Subject:      Re: GEN: *please* read: PROPOSAL: intervals
To:         lojban@cuvmb.cc.columbia.edu
Content-Length: 4997
Lines: 104

> > Anything except 0da will satisfy me. I *need* such a mechanism, and
> > I believe this should be expressible within the tense system. It
> > ba'e belongs into the tense system.
>
>    This posting is NOT pro or con the proposal and
>    NOT A CHANGE PROPOSAL, just a technical feasibility study.
>
>    If we overload NOI and make the following changes to the YACC grammar,
>    we can express both an origin and an offset:

[li'o]

>    A more general form would replace 'NOI_585 sentence_40' with
>    'relative_clauses_121'.
>
>    Corresponding changes could be made to the grammar of VEhA/ZEhA.
>
>    All these modifications pass YACC without conflicts.
>
>    VA/ZI+NOI is more verbose than xe'i but also more expressive as it
>    combines subjective and objective distances.

I say, great!

>    I REPEAT:
>
>    THIS IS NOT A CHANGE PROPOSAL. I REFRAIN FROM MAKING ANY NEW PROPOSALS
>    -- AND I'M NOT PUSHING EVEN THE ONE I MADE ONE AND A HALF YEARS AGO.
>
>    In general, I'm more con than pro at the moment. We need the books
>    and we need the baselines, soon.

I know. What I am saying is: this gap was found a long time ago. It is
not a particularly hard one to fix. But, it was ignored. Only xorxes and
I paid any attention, AFAIK. Nothing was settled, not even discussed.
Now you say it is too late. Why is it too late? Are the books already
done? Maybe I am just dumb and don't know what I am talking about, but
the change you not-proposed is easy to implement in YACC form and needs
three more paragraphs written in tense.txt. That is all the work needed
to put it in the language as far as I can see. If you have tested this,
and it works, I don't see why it should be more than twenty minutes
work, providing people agree upon it. This solution is elegant, and I
really admire it: it requires no new cmavo, doesn't change the existing
corpus, it is versatile, just as you said people can describe objective
*and* subjective distances ({be'azavoi ki'omitre}, if I'm right), the
only two possible objections I can think of being "overloading" of NOI
and GOI and differing from other imaginary journey devices.

The former causes no ambiguity whatsoever, syntactic or semantic. This
use of relative clauses/phrases as defined in ma'oste does not differ
from this. This use of {noi} also "attaches subordinate bridi with
incidental information", {no'u} is "non-restrictive appositive phrase
marker" and marks ZI/VA interval/distance specifier as that "which
incidentally is the same thing as" a more precise specification of the
interval in question, etc. I find this very consistent. The only
difference is that it does not attach to a sumti in this case, but to a
imaginary journeys system cmavo. It all seems very logical to me, much
more so than BO and at least equally so as NAI and JOI. Just as those
selma'o, this does the same function wherever it occurs. We would just
add one more context where NOI is applicable *meaning*the*same*thing*. I
don't see this as overloading.

The latter possible objection isn't IMHO so important - this is
functional; it would make me able to say things I feel very clumsy
saying under current (incomplete) system. That's what I feel is
important.

Anyway, I have rambled long enough. I think that the only thing that
could pose an obstacle for inclusion of veion's change proposed by me
because he does not wish to do so himself zo'o is endless discussion.
There has been many endless discussions I have seen on this list and I
definitely do not wish to see another one. I see this as pretty
straightforward thing, and if I sense that things are heading for an
endless discussion, I will withdraw the proposal and continue to live in
sin, using my slang {te'i}, because I really feel imaginary journeys
system to be incomplete without a way to specify distances/intervals
more precisely than short/medium/long.

If I wrote anything stupid here, please, drop a mail, enlighten me. I
really think this is an ingenious solution, great thinking veion, and
almost trivial to implement, nothing like fuzzy logic and existential
import and opaque contexts and stuff. If you see a problem with this,
say so - y'all know I am reasonable enough. I do see the importance of
publishing the refgrammar, but I expect this proposal to be dealt with
in three or four days maximum.

I hope positively.

> > ERROR: the YACC modification I previously sent works at the YACC
> >       level but isn't feasible in practice as the modification
> >       ended up on the lexer side - I'm not yet sure whether it
> >       can be done on the parser side, probably not without a
> >       major modification.
>
> However,
>
>    adding the following two lines to the rules for 'modifier_82'
>
>          |  tense_modal_815  relative_clauses_121
>          |  tense_modal_815  relative_clauses_121 BO_508 sumti_90
>
>    (in the unmodified grammar) would give us a working solution.

co'o mi'e. goran. noi mutce pacna lenu do ze'i pensi gi'ebazibo datpai