Message-Id: <9203172207.AA00260@relay1.UU.NET>
Date:         Tue, 17 Mar 1992 16:45:10 -0500
Reply-To: "Mark E. Shoulson" <cbmvax!uunet!ctr.columbia.edu!shoulson>
Sender: Lojban list <cbmvax!uunet!pucc.princeton.edu!LOJBAN>
From: "Mark E. Shoulson" <cbmvax!uunet!ctr.columbia.edu!shoulson>
Subject:      A pair of how-do-i-say-it's
To: John Cowan <cowan@snark.thyrsus.com>
In-Reply-To:  John Cowan's message of Tue, 17 Mar 1992 15:42:34 EST
Status: RO

John answers my questions:

>la mark. clsn. cusku di'e

>> The first is the use of {cei} and the {bu'a} series....
>> Here's an example of a sentence I was playing with:
>>
>> George Bush is to the United States what John Major is to Great Britain.
>> la djordj. buc. bu'a le merko gugde .i la djan. meidjr. bu'a le brito gugde

>Correct, except that ".i" needs to be ".ije", otherwise the "bu'a"s are
>separate.  This is a rule that applies to "da" also.

Say what?  That doesn't seem to make sense to me.  I can agree that {ni'o}
should cancel assignments, but {.i}?  Grammatically, there's no stronger
tie between {.i}-joined sentences and {.ije}-joined sentences.  Maybe I'll
go easy on you if you say I have to use {.ibo} or {.ijebo}; at least that's
technically a bit closer.

Then again, who wants to remember that he used {roda} last sentence and now
has to do {rode}, even though the two have nothing to do with each other?
Oh, well.

>Here's an example of "bu'a" within a prenex, from my Hakka story:

>   ro bu'a zo'u la .aniis. cu djica le nu bu'a .inaja bu'a
>   for-all <pred> [if] Anyi desires the event-of <pred> then <pred>

>The grammar demands that any bu'a-series variable appearing within a prenex
>must be quantified, typically with "ro" or "su'o".  A bare "bu'a" is a
>selbri and isn't allowed.

Hmmm.  I can see it, but it looks klugdy.  Actually, it may not be so hot.
Look: the way I see it, this sentence is: "For all
things-that-are-something1: [if] Anyi desires the-event: (something)
is-something1, then (something) is-something1", using "is-something" for
selbriness.  Leaving aside the fact that I agree with Nick that causality
is not something you should have to do truth-tables for, the prenex isn't
quantifying over all predicates, it's quantifying over all *things* that
satisfy this unspecified predicate, just as {ro prenu} is all things which
fulfil "prenu"itude, i.e. "all people". (ro broda ~= ro lo ro broda).  Does
this make sense to anyone else, or am I missing something?  I suppose I can
see how it can wind up meaning what you want, but not well.  Oh, how about
{ro nu bu'a zo'u}?  Hmmm.  doesn't look much better, same problem.  I can't
think of any way, grammatical or not, within Lojban's (or my own) framework
to get what I'm looking for; how do you quantify over relationships?
Requires some thought...

>> The other came up in a translation I was thinking about.  We have relative
>> clauses to specify sumti, but they only attach to sumti at a fairly low
>> syntactic level.  So let's say I mean to say "I meet the man and the woman
>> wbout whom you talked with me." (meaning you talked about *both* of them.
>> And for the sake of argument, I met them separately and unrelatedly, so
>> {.e} would be a reasonable conjunction).

>This is a known problem which I'm working on for the next release of the
>grammar.  Ideally, we should be able to say:

>   *mi penmi ke le nanmu .e le ninmu ke'e poi do tavla mi ke'a

Sorta what I thought; there should be a way.  I was kind of looking for a
way to parenthesize sumti, to make them one sumti of a lower level, just as
ke/ke'e braces parenthesize selbri to low-level selbri so the precedence
changes.  That's why I looked at termsets first, but they're still too
high-level.

>but so far I haven't been able to make that form work.

Maybe you can get some sort of parenthesizing scheme to happen.  Would be
most convenient if you didn't need more selma'o, though.

>> The solution I found was with LUhI:
>>
>> mi penmi lu'a le nanmu .e le ninmu lu'u poi do tavla mi ke'a
>> I meet the-individuals-of: the man and the woman (close-LUhI) which-are...

>Yes, it works, but I agree it's ugh.  (Hmm: we need a word to transform a
>UI into a selbri....)

You're joking, but I think you're right, we do.  Fortunately, from my
standpoint anyway, we do.  {mela'elu .a'unai li'u} works for me (assuming
that's the right UI).  Easier, of course, with non-compounded UI, so I can
use {mela'ezo .ai}.  Anyone's guess as to good place structures, though.
Maybe "x1 internally percieves emotion/discourse thingy/whatever (n)".
Steps a bit on the toes of {cinmo}, that one.  But taking it as it stands,
you could have your use of "ugh" in the above statement as {mela'elu
.a'unai li'u rinka}: repulsion-making.  Maybe a {ka} somewhere?  I'll leave
this to better stylists.  I like the idea, though.

~mark