Re: [lojban] Re: A (rather long) discussion of {all}

[John E Clifford <clifford-j@sbcglobal.net>]

--- Maxim Katcharov <maxim.katcharov@gmail.com>
wrote:

> On 5/15/06, Alex Martini <alexjm@umich.edu>
> wrote:
> > This is getting to be a really good example
> of a too-long thread.
> > Which I normally wouldn't point out, except
> that I really don't see
> > that we're clearing anything up in the
> process. I think a short study
> > break might be in order.

Sorry about the length, but, since {lo} is
probably the most common word in the language and
the quantifiers are also very central, it seems
worth a bit of effort to get them right.

> > First, regardless of how we like it, {all}
> doesn't have a single
> > definition. Rather, it has many
> interpretations depending on how it's
> > used. This is actually not that unusual; in
> my first Intro to
> > Linguistics course, we found quite a large
> number of these only
> > looking at English. These English cmavo-type
> words (for lack of a
> > better term) are exceptionally difficult for
> even the best linguists
> > to nail down entirely because they are very
> fluid around the other
> > words in the discourse. If you want another
> example, write down every
> > use of {so} in a discourse. By my count it
> has at least  4 different
> > functions, most of which have nothing to do
> with the core meaning of
> > {x so therefore y}.
> 
Well, strictly, "so therefore" is not the core
meaning of "so,"  which is closer to "so, thus"
or "so, is the case."  But we could probably get
into an argument about that, too.
> 
> >
> > Also, here are a few useful websites that
> describe how lo/le are
> > currently defined in Lojban and some proposed
> (and possibly accepted)
> > modification to the rules. I would ask that
> we take a short
> > intermission and (re)read these.
> >
> >         From CLL
> http://www.lojban.org/tiki/tiki-
> > download_wiki_attachment.php?attId=192
> >
> >        
>
http://www.lojban.org/tiki/tiki-index.php?page=How+to+use+xorlo&bl
> >        
>
http://www.lojban.org/tiki/tiki-index.php?page=BPFK+Section%3A+gadri&bl
> >        
>
http://www.lojban.org/tiki/tiki-index.php?page=Robin's%20gadri%
> > 20Proposal
> >
> >         Maybe this too:
>
http://ptolemy.tlg.uci.edu/~opoudjis/lojbanbrochure/
> > lessons/less4articles.html
> >
> > Finally, let's look at exactly what we want
> to define. So far, I see
> > two big questions. What's the difference
> between {le} and {lo}, and
> > their counterparts {lei} and {loi}? 

These seem to be pretty much in order (excet for
a few minor special cases).  The fact that people
are still confused about them is a PR problem,
not one for Lojban per se.

>Also, how
> do the inner and outer
> > quantifier, especially {ro} interact with
> each of these.

Again, aside from some minor special cases, there
is little real controversy here -- or was until
now.

> >
> > I really don't feel that I know Lojban well
> enough yet to make a good
> > explanation of the difference between {le}
> and {lo} without lapsing
> > into malglico Lojban, which isn't very
> helpful.

{le} for specific items, {lo} everywhere else.

> > [ li'osai ]
> >
> > mu'omi'e .aleks.

> 
> I think that we're slowly making some progress.
> It's a hefty subject, after all.
> 
> Answers.com gives >10 definitions for "so"
> ({so} = "nine"), and that
> probably doesn't even cover them all. One of
> the things that Lojban
> tries to solve are these hacked-together usages
> of words. All words
> are not inherantly vague, we can make words
> have good definitions.

Which, alas, involve more words that are
inherently vague (not to mention ambiguous).  but
that is not the current problem

> That's what all sides of this debate would like
> to do.
> 
> We won't be able to define both an inner {ro}
> and {le}/{lo} at the
> same time. Since the last explanation I offered
> for {le}/{lo} used an
> inner {ro}, I think that we should clear {ro}
> up first.

Good, since {le} and {lo} are taken care of.  So
is {ro} for that matter, but you are wanting to
change that and so we have to see what your
suggestion amounts to and whether it fills and
need and can be made to fit into Lojban as it
exists otherwise.
 
> I think that the problem may be that we don't
> have any definite way to
> know which usage is better. Perhaps we should
> set up some criterea
> under which to judge:

Given that there is established usage, making
acceptable changes in that has to play some
significant role.  It is a factor that has been
ignored in the present discussion.
 
> 1: Can a speaker actually restrict down to what
> their referant is?
> (i.e. "make a complete restriction"?)

For a given description of a situation, yes.  For
one that will meet the various ways the actual
situation may develop, not in general.  Once the
situation has developed in a certain way, we can
retrospectively create restrictions which would
have gotten the right thing, but that didn't help
the actual case.

> 2a: Is it important to be able to make complete
> restrictions?

Rarely (and rarely possible in the strong sense).
 We need ones that are good enough to do the job,
anything beyond that is excess verbiage.  Of
course, given that a situation may have several
descriptions, what is enough in a given situation
is hard to determine before it is put to the
test.  If the conversants end up at the same
referent, the description was adequate, if they
don't, not.  But we cant generally know
beforehand which it will be.

> 2b: Is there something that the current model
> doesn't handle well that
> is handled by the proposed usage?

None has been demonstrated; specific claims to
the contrary have been false on their face (if a
case can be described, the description can be
mirrored in Lojban). Nor does the proposed
handling really make much sense in Lojban.

> 3: Is there room within the current model for
> the proposed usage?

Maybe.  But the present proposal is to put it
into a space that is already occupied by a
reasonably clear system.
> 
> If the answer to 1 is "no", I will give up my
> position. If the answer
> to 2 is a definite "no" (as opposed to a "well,
> we can sort of hack
> around it, most of the time"), I will give up
> my position. 

Well, it seems a bit of a case of simple
name-calling to say that standrad procedures --
in Lojban as in many natural languages -- is a
hack.   It is the proposal, after all, that is
arbirrariky changing established things, pushing
usage to achieve an already achieved end.

>If the
> answer to 3 is "no", then I will try to think
> of some other way to
> propose this feature.
>
That might be you best move, provided that some
need for your proposal could be established. 

As I have said before, your proposal is
unnecessary, in that we already have good enough
descriptions, and futile, in that "improvements"
on what have will not significantly improve our
chances of shifting referents successfully.