[lojban] Re: Usage of lo and le

[John E Clifford <lojban-out@lojban.org>]

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

> On 5/11/06, John E Clifford
> <clifford-j@sbcglobal.net> wrote:
> > Well, I am not sure how much more clearly
> since I
> > am not clear what the two approaches are that
> you
> > see as being used and as bsiing inconsistent
> with
> > one another.  It seems to me that if the
> Lojban
> 
> Quoting xorxes:
> 
> <<Position 1: There is a determinate number of
> things that satisfy
> the predicate {cribe}, independent of any
> context whatsoever. Therefore
> in any context {lo ro cribe} refers to all and
> exactly those things.

But who exactly holds this?  To be sure, we can
construct such a domain for special purposes, but
it is only the referent of {lo ro cribe} when
those purposes are activated.  It happens that
logicians and philosophers have those purposes
often before them and so sometimes talk as tough
this were THE referent of {lo ro cribe}.  It is
handy to talk this way to do their kinds of
things, but that does not affect what happens in
ordinary conversation -- even among philosophers
and logicians.  Looking back, I see that (being
both) I have slipped occasionally into that
logician-philosopher talk to deal with extreme
cases (these being as extreme as possible,
presumably) but I even then can't find a place
where I said that this is what {lo ro cribe}
always means.  If I did, it was accidental and I
apologize for the confusion it seems to have
caused (if I didn't, then I apologize for not
having caught on more quickly that this was the
confusion going on).

> Position 2: The things that satisfy any
> predicate may vary with
> context. In a given context {lo ro cribe}
> refers to all and exactly the
> things that in that context satisfy the
> predicate {cribe} (not just the
> things present where the speaker is, mind you,
> all the things that
> relevantly satisfy the predicate).>>

I think there may be some confusion in
terminology going on here, between what is in
context and what is relevant.  I suspect I have
been using them interchangeably and xorxes seem
to be doing so, too, at least until recently. 
The context involved  and the relevance are both
features of the flow of conversation, not just
the physical location or the like; the domain of
discourse (which is again the things relevant in
the context) is ideational (or linguistic), not
physical or sensory, expands and contracts with
what is being said (and charting the exact flow
is not something that anyone has tried to do
beyond noting what is in or out at a given point
in the conversation -- and whether the steps have
been taken to assure that all conversants have
essentially the same view).
  
> > use of {ro} is inconsistent, then so is the
> > English use of "all," the main difference I
> see
> 
> Yes, it is. Just like the English use of most
> English words is inconsistent.

This seems a very odd thing to say.  It appears
to be either false -- since English shows no
signs of collapsing in the way that inconsistency
would be expected to create -- or to involve a
sense of "inconsistent" which is innocuous --
since generating none of the expected problems. 
Does this mean anything more than that most
English words have several meanings and that we
need context to tell which meaning is in use?  If
so, then I take back my remark about {ro} and
"all," since {ro} is not ambiguous in some ways
"all" is (mass v count, for example).  And I find
this a very strange sense of "inconsistent."


> > being that we have tried to lay out the
> Lojban
> > use with some care, whereas -- linguists
> aside --
> > English speakers are pretty unaware what is
> going
> > on.  {ro} (like "all") means everything (or
> the
> > named sort) in the domain of discoure.  What
> is
> 
> I understand this usage of {ro}, and my {L_
> cribe} (a lack of ro)
> implements this usage. 

But, since it does not involve {ro}, it does not
implement that usage.  Do you mean (I suppose you
do) that your expression {lo cribe} ({le} doesn't
really have a role in all this) says the same
thing as standard {lo ro cribe} (or, as you would
say, "{lo ro cribe} in this sense")?  This flies
in the face of established Lojban usage, where
{lo cribe} is (loosely speaking, since this is an
area of controversy -- but your suggestion is not
in the range) an unspecified bunch of bears,
maybe one, maybe all, maybe somewhere in between,
and maybe some suprabear entity that is
represented by those things or functions
autonomously.  In no case is it delimited as all
the relevent bears in the context. 

However, your {__ ro
> cribe} implements this
> usage, /and/ another usage - either "jump out
> of context", or "all
> ever" -- the former I consider basically the
> same as the first usage
> (and thus useless), the latter I implement with
> {L_ ro cribe}.

The standard {lo ro cribe} refers always and
everywhere, so far as I can tell, to all the
relevant bears in the context, but, notice, may
itself be part of setting that context, if the
domain is to shift with repect to bears.  Now,
the process of shifting domains is a tricky one,
as are all oves involving the Gricean
conversational conventions, and we have to pick
our way through them with care.  You have shown
in other places that you are aware of the
minefield and have done some picking -- not, I
think totally accurately, but usefully for making
your point. Starting with the bears currently in
the zoo, who are in the instant case clearly in
the domain, we want to jump to, say, all
currently existing bears.  So we say {lo ro cribe
poi zasti}.  This might, of course, be to
restrict the previous range of bears to the
existing ones (it previously included some
defunct and some imaginary ones, and so on), but
the speaker (and presumably the rest of the
conversants) hold that domain was already only
existing bears (the question "Did you see all of
them?" pretty much sets that in place), so this
is an expansion of the domain.  If this does not
work, then the description can be modified as
needed.  the point ids that this new description
sets a new domain that is not necessarily
relativized to the previous domain -- or, at
least, aims to do that (it may fail, but that is
a practical matter, not a theoretical one  --
though the theory has eventually to account for
the failure).

> > in that domain -- particularly of the sort in
> > question -- varies with the context (which is
> a
> > fairly broad concept, involving what is said,
> > what has been said, what is in the attention
> of
> > the conversants, and probably countless other
> > things encompassing the whole range of the
> > conversant's experiences and knowledge).  The
> 
> Yep. These factors are used to help the
> listener pick out which things
> are being spoken of. Another (or at least
> another use of) context is
> using it to place things relative to it: if we
> didn't have this
> context, we wouldn't know what "now", or
> "before", or "tomorrow", or
> "here" meant, because they're all relative to
> the current context. The
> latter is necessary just about always, but the
> former is only
> necessary when you aren't being precise enough
> (when you aren't
> restricting enough).

Actually, in the theory, moist of these examples
are relative not to the whole context but only to
the occasion of utterance, a relatively
controllable component of the context.
 
> When I say "this pen here-now on my table is
> blue", I use relative
> context (the latter), and disambiguating
> context is unnecessary
> altogether.

I don't see the point here (though I am glad to
see that specificity is no longer a part of the
issue).  I can often be precise enough to get
down to a single possible referent within the
context (the occasion of utterance is a part of
the context, after all, however separately
treatable). Are you saying only that we can
always (a rather risky claim, so "usually")
specify the referent of an expression uniquely,
so that there is no possibility of anyone being
in any doubt.   Given the things one can doubt if
one sets one's mind to it, I'll withdraw that
last bit, but surely we can specify so thoroughly
that all reasonable uncertainties are assuaged. 
Sure. We don't usually go to such extremes,
however, unless forced to it by the failure of
more normal means.
But has anyone suggested otherwise.  The most I
have seen is just that such precision is rarely
called for -- which seems correct.
     
> > speaker needs to be sure that what he says
> fits
> > into the existing context or change it in
> > recognizable ways.  He does not always do so
> --
> > or does not do so successfully -- so muddling
> can
> > -- and does -- occur.  And then the
> adjustments
> > have to be made.  But in all this, where is
> the
> > inconsistency?  Indeed, where the two usages?
> 
> Quoting xorxes and myself:
> 
> <<me: How would you say "let's talk about all
> bears that have ever existed"?
> >xorxes:
> > Something like:
> >
> >   e'u mi'o casnu lo ro cribe poi pu ja ca
> zasti
> >
> me:
> Not by your rules. Here you are inviting me to
> talk, out of the bears
> that are in context, of the ones that have
> existed and exist. 

No, he is setting a new context.  To be sure, it
*could* be a subdivision of the old domain, but
-- as noted earlier -- the old domain already had
only existing bears in it so this must be taken
as an extension.

This is
> clearly inconsistent. When does {__ ro} refer
> to all bears? When
> someone includes the word zasti after a poi?

If that works, then "Yes."

> When all bears in context
> clearly already exist? "Aha, clearly he's not
> talking about all bears
> already in context, because I thought that they
> all exist... wait, was
> he talking about more than existing bears
> then"? 

In this case, no, since they were all things I
could have seen.

And what if all bears
> in context don't exist now-before, and I want
> to suggest talking of
> the ones that do? Do the rules of Lojban change
> based on the context
> (all bears in context meet restrictions = new
> context, if they don't =
> modification of current context)?>>
> 
> There.
> 
> It's impossible for you to specify the context
> (disambiguating-context), because you can't
> jump out of the current
> context in order to do so. 

Of course we can and do all the time.  You seem
to have locked us into a position that no one
actually held and, should we have said something
once that might have been read as that, surely
subsequent discussion -- unless filtered through
your idee fixe -- should have long ago corrected.
I'm sorry if I was the one who said the
misleading thing, but surely I have said enough
since to indicte the way things really go and
even more surely xorxes has.

You basically have
> to make a statement
> that's nonsensical when applied to the current
> context, and then the
> listener says "uh, ok, I guess we're not
> talking about bears under
> that context anymore (or maybe my notion of the
> previous context was
> wrong?), so I'll make a guess as to what the
> context is now"

Yeah, in a sense that is the way that
conventional shifts work, although I would hope I
give better guidance than this suggests (well, I
can always hope...) and that it does not always
involve paradox (although paradox resolution is a
fabulous model for parqadim shifts, which this is
rather like in extreme cases). 
  
- and
> there's still no guarantee that that the
> listener will choose the
> correct context: 

Nope, no gurantees, but there never are in any of
this.  We do the best we can and patch up as we
go along.  And that is just the way it almost
always works.  We can take the time to be so
precise that no reasonable uncertainty remains,
but that usually takes too long and is too
complex to do on the fly (and if you had toime to
work that all up you could probably find a better
transition within normal means).

If I try to jump out with your
> {ro} (or your
> {rosai}), there's still no guarantee that it'll
> be all bears - maybe
> you're talking about all bears in some other
> "contextual sense".
> 
> > > Precision in picking out a referant has
> nothing
> > > to do with describing
> > > the referant down to the last molecule.
> It's
> > > enough to give a
> > > description that only
> the-thing(s)-you-refer-to
> > > can meet.
> >
> > Enough for what?  If you get down to a
> 
> Enough to let the listener know that you're
> talking about this pen and
> no other, the two cubs that are yours and no
> other, all bears that
> ever existed and not a subset, etc.

Yes.  And so?
 
> > description which only one thing meets, it is
> > pretty pointless to add to it if your purpose
> is
> 
> Yeah, it is. You misunderstand my position if
> you think that I'm
> saying we need to be needlessly specific.

Well, it was unclear just why you wanted to be so
precise (unclear to me anyhow, but I still had
specificity -- which your last remark here
reintroduced -- in part of my mind.) (Note that
you have not yet been specific, merely precise --
in the peculiar technical way we are using those
expressions: you may narrow things doen to (at
most) one things and still not be specific since
you may have no idea what that one thing is (or
even if there is one).)

> > to pick out exactly one thing (or, more
> likely,
> > at most one thing), if your purpose is to
> give a
> > complete description of a thing then
> obviously
> 
> No, I'm not at all talking about giving an
> object a complete description.



To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.