From: cbmvax!uunet!math.ucla.edu!jimc
Message-Id: <9104181535.AA22615@luna.math.ucla.edu>
To: lojban-list@snark.thyrsus.com
Subject: Re: oops! correction 
In-Reply-To: Your message of "Thu, 18 Apr 91 12:46:43 +1000."
             <9104180246.29217@mullian.ee.mu.OZ.AU> 
Date: Thu, 18 Apr 91 08:35:51 -0700
Status: RO

> Date:  Thu, 18 Apr 91 12:46:43 +1000
> To:  lojban-list@snark.thyrsus.com
> From:  nsn@mullian.ee.mu.OZ.AU
> Subject:  oops! correction
   
> Stylistics, I say, and I get a grammatical correction. *sigh*. Thanks
> anyway, John. Well, if that didn't generate any traffic, how about the
> way in which I run my cmavos together all the time? Does this irritate
> people? DO PEOPLE EVEN READ THIS?! Oops - getting carried away again.
> But I think writing things like {sefinti}, instead of going for the lujvo
> {selfinti}, is an underestimated resource.

My own preference is to stick some cmavo together, as in "lemi", but to
regard each one as a separate word rather than requiring the parser and
dictionary to honor the unit just because it lacks a blank -- this
being the position in Institute Loglan.  Being a native English writer
I tend to separate the article from the predicate, e.g. "la lojban"
rather than "lalojban", logical though the latter might be.  However, I
tend to interpret the article as being much tighter in the "core sumti"
than a case tag, etc., so "fila lojban" doesn't look right to me; here
I would put the blank: "fi la lojban".  This is due to both English
influence and a certain historical perspective on the Old Loglan
grammar.  

> >More importantly, "ma" and the other question words signal direct, not
> >indirect questions.  English indirect questions like "They asked me how
> >to write news" are hard to render in Lojban, even once you resolve the
> >ambiguity of "how" (in this case "ta'i ma" = "With what form?").
> >As yet there is no generally accepted way of writing indirect questions.

I'm getting into this a bit late, but I think an acceptable choice is
clear: "he requests (I say/talk about/teach/etc. (the form (which is
correct) of (a typical journalist writes a typical news[-item])))"  In
other words, "request an action" rather than "ask a question".  Sorry
for not producing proper Lojban but, well, I could do it in -gua!spi if
anyone's interested...

> Overall a medium bad effort, but I'll keep it up anyway. But does anyone
> out there deviate from SVO1O2O3O4 in order to get elisions? That's all I want
> to know. 'Cause if you don't, maybe you should. And maybe you shoudln't.

Normally I use SVO order out of English habit, but often for emphasis I
will do VSO (which is legal with no markings in -gua!spi and I don't
see why not in Lojban too).  And I will freely scramble the order to
get an emphasized sumti in the first position, or a particularly
complicated one at the end.  To elide an unwanted 1st case I typically
convert the predicate (this is more effective in Lojban which has
exchange conversion).  But when the sentence has two or more actual
sumti plus an elided one before/among them, I often prefer to drop an
explicit "something" in the elided case rather than muck up the rest of
the sentence with conversions and caselinks.  It's simpler and shorter,
particularly if it's the only quantified item.

Speaking of which, a special "something" analogous to the anonymous
variable of Prolog could be very useful (and is, in -gua!spi).  Each
instance is chosen independently; it cannot be copied anaphorically
(because at least in -gua!spi the anaphor is replaced by a copy of the
antecedent words, not the referent thereof; I don't know what the exact
policy on this is in Lojban); and in quantification the anonymous
variable(s) come last regardless of order in the sentence.  

James F. Carter        (213) 825-2897
UCLA-Mathnet;  6221 MSA; 405 Hilgard Ave.; Los Angeles, CA, USA  90024-1555
Internet: jimc@math.ucla.edu            BITNET: jimc%math.ucla.edu@INTERBIT
UUCP:...!{ucsd,ames,ncar,gatech,purdue,rutgers,decvax,uunet}!math.ucla.edu!jimc