[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[lojban] Re: A (rather long) discussion of {all}
On 5/19/06, Jorge Llambías <jjllambias@gmail.com> wrote:
> I bring up certain examples where (it seems
> that) your methods don't have a capacity to be specific, unless it's
> through some verbose and still-context-dependant means.
Specificity is covered with {le}. You have something in mind and then
These domain-dependent definitions have gotten the better of me. I
don't mean "specific", I mean "precise".
you use whatever description you think is most useful to get your
listeners to identify what it is that you have in mind. But your proposal
concerns precision: a description so precise that no matter how picky
and lawyeresque your listeners get, the description can still refer to all
Whoa there. No. I explicitly stated that this would not cover all
issues of ambiguity in Lojban. In retrospect, "complete/absolute
restrictions" or "100% certain that the restriction..." were probably
not the best things to say, but really I don't feel like saying
"99.99999% certain in this situation that the restriction", or
explaining that I mean by it that the *restriction itself* would not
be at fault - the person's bounds may be at fault.
and only the things you intend it to refer. I don't believe such precision
is achievable.
In each person's mind at a certain time, I think that it is (though
obviously people don't spend all their days thinking about exactly
what is and isn't a bear). It's just that that certainty is, yeah,
unattainable, because the listener doesn't usually know the speaker's
exact model for "bears" or "cubs". (Unless they define it as having a
bear-pelt between under X inches across or whatever.)
It might be achievable by using super-absolutist dictionary
definitions like "a human-child is one who is under the chronological
earth-age of 18" or something, but that would be extremely
undesireable.
>Yes, context
> works most of the time. You can usually say with 90% certainty that a
> person means X. Occasionally, cases where there is 50~% certainty come
> up, but they seem to be bargained up to something "sure" like 90%
> through verbose ensuing explanations. But in either case, your method
> doesn't have 100% certainty regarding a referent for usual "LE" sumti.
That's right, I don't believe 100% certainty in an absolute sense is achievable,
no matter how much effort you put into definitions.
No, in some cases it isn't achievable. However, in some cases (i.e.
most actual cases) where your method alone could never be certain,
mine would be. Consider the game example.
Consider that at the end of the game, I've won and say "pick up all
the stones on the table, we'll start again" (pick up = make orderly),
and I mean *all* the stones on the table. If we've had it occasionally
where the loser must order the stones, there's a 50~% chance of me
being misunderstood. If we've played the last few games where the
loser always picks up all the stones, there's a 5~% chance of me being
misunderstood. If we've never played with that consequence, then
there's an 90~% of me being misunderstood.
Ok, so in that case with 90%, just say "pick up both all the stones
that I used and all the stones that you used on the table now". Which
is exactly my point. I shouldn't have to resort to something like
that, even in a simplistic example like this. And especially in more
complex and critical situations. My proposal reaches 100% certainty in
saying "make orderly all the stones cpana the table" or "make orderly
all your stones cpana the board" or "make orderly all stones related
to you", etc.
I say 100%, but that's not exactly true - there might be a chance that
my opponent does not recognize that the table is a table, because it
deviates too far from his model. The odds of there being a person who
doesn't recognize
http://www.greenfingers.com/images/superstore/RD0002D_l.jpg
the thing in that picture as a table are slim (one in a million), but
they do exist, after all, so yes, my restriction is still "open to
ambiguity".
> That sort of certainty is achievable, the very simplest one is {pami},
{pami} means "one of us, the speakers".
I thought I was told that that defaulted to the actual current
speaker. Oh well, that makes more sense anyway. By fluke it's right
regardless, because I'm the only one speaking, and I'm speaking for
nobody but myself (in which case just "mi" should have been good
enough...)
> > > > > {mu L_ ro cribe cu ba zasti}
> > > > > Five bears, out of all hypothetical future bears, will exist in the future.
> > > >
> > > > Yes, that's "out of all bears, exactly five will exist (at some
> > > > unspecified time in the future)".
> >
> > I was using it in my sense, i.e. "out of all things that count as bears".
>
> What counts as bears? Bears that will actually exist in the future?
It depends on the context. In many, perhaps most contexts only
bears that exist will count as bears, yes. But as you say below, when
the predicate of the claim is {zasti}, non-existing bears could be brought
into the picture by the very presence of that predicate.
"Unless you're dealing within a {zasti}-type relationship, zasti is
implicitly added as a restriction." This rule that I defined has
nothing to do with context, it's more like a rule of grammar than
anything.
> Surely not, consider:
>
> {mu L_ ro cribe cu ba zasti}
>
> "five of all-bears-such-that(-will)-exist will exist"? That wouldn't
> say anything at all. You need a hypothetical mega-set so that it
> becomes comprehensible: "five of all-hypothetical-future-bears will
> exist".
It could also be interpreted as "exactly five of all currently existing bears
will still exist in the future". You don't offer much context so there are
many open possibilities.
Well, your understanding is that the meaning of a word changes based
on context. My understanding opposes that - my definitions of words
change when I learn something new about how other people are using the
word. My point was that regardless of how you see it, the group that
you're choosing 5 bears (to say that they exist) from is not the group
defined by "such that will/do/did actually exist".
> > le tanxe cu vasru no da to no maisle .e no pulce .e zo'e si'a toi
> >
> > That is of course a much less likely thing to say than the usual
> > "nothing", so it is proper that it takes a bit more effort.
>
> Erm. "The box contains nothing" - how would I say that? I don't mean
> the nothing as in in-context, and I don't mean "not stuff like
> molecules and dust". Just "nothing" (no existing thing).
Plain "nothing" would be {no da}.
You've explicitly stated that your {no da} could mean "nothing" or
"not dust" or "not what we've talked about" etc., all based on
context. Here's where you stated this:
This is what I mean: In many/most contexts particles of dust won't
be available as possible referents without some extra work from the
speaker. So if you say {le tanxe cu vasru no da}, "the box contains
nothing", dust particles won't count as a disproof of the assertion. Of
course, any participant may bring dust particles into the discourse
and then they will have to be dealt with somehow, but until and unless
that happens, they don't count. You seem to want them to always count,
so that {le tanxe cu vasru no da} is practically always false.
I'm asking you how I can be specific that I mean *nothing*. Like, if I
by {no da} meant "no dust", I could be specific (via your method) that
I meant "no dust" by adding that {da} is dust or whatever. But I'm
asking how you make unambiguous the other case: that by {no da} that
you mean *nothing*.
My assumption is that you can't. This seems to be the clearest example
that I could find of the problem with your lone method - there are
many more practical examples where it exists, but I seem to be having
a hard time clarifying in them.
"No existing thing" would be {no zasti}.
You can't say anything without context in an absolute sense.
I don't know what you mean. But I'll make a guess and respond:
Yes, in your version you can't restrict without needing context to
finish the job, but in my version you can.
> (Point being that I
> want to be shown that this proposal does nothing significant, *not*
> that it doesn't cover every case where there's ambiguity in Lojban.)
How does one show that? It should be easier to show, first that it can
work, and second that it does something significant.
Show me how to say "nothing" (described above) via your method in a
way that isn't convoluted.
> > Right. Words are never defined with such precision that your proposed
> > {ro} could ever be applied with absolute certainty.
>
> I disagree. Lojban {punji} is explicit - you have to perform the
> action. That seems pretty certain to me.
It seems pretty certain to me too, but we were talking about infinite precision,
so that you can know with absolute certainty whether {punji} applies or not.
Applies to what? That we can't be sure that the definition of {punji}
that I have in my mind is what I'd call that thing which I want you to
do? (i.e. put X into Y.) No, I'm pretty certain of my definition,
especially in this into-bag-putting situation. What I'm not certain of
is that your model for {punji} will match my model for {punji}, though
with {punji} the chances of that occuring are a million times smaller
than our models for "bear" not matching. I am of course assuming that
neither of us are delusional or new to the concept of "put".
I'm not saying {punji} is not well defined, I'm saying it is not
infinitely precise.
> Regardless, you're trying to
> move out of the scope of my proposal. This can't cover every problem
> (namely the two objections that I just mentioned), but it does cover
> some rather important ones.
Which important problem does it cover? Neither the bears example nor
the stones example are the least bit problematic as far as I can see.
Neither requires an absolute inner {ro} to make yourself understood.
Indeed an absolute inner {ro} only brings confusion in both cases, as
you start to wonder about irrelevant stuff like imaginary bears or whether
the stones are or are not on the table in the most absolute and strict
sense possible.
I don't worry about imaginary bears because of language's hackish
"zasti rule". And I certainly don't worry that our models for "stone"
or "table" or "white" or "put away" differ in a way that would prevent
me from saying "put away all the white stones on the table now".
> I'm not adjusting because I screwed up on my restriction, I'm
> adjusting because I screwed up on my definition (I thought a word
> restricted something in some way implicitly, and it didn't.)
But how can you ever be certain about a restriction, unless you assume
that words have infinitely precise meanings? Do you think lojban gismu
Because I don't want to have to say I'm 99.999999% certain every time.
I think that the odds of you saying "there is no table in the picture
you linked" are sooo slim, that yeah, I'm pretty certain. Much, much,
more certain of that than I am that "pick up all the stones on the
table now" means "pick up all your stones" rather than "pick up all
our stones", or that you mean "the box contains *nothing*" rather than
"the box contains nothing that isn't really really small and probably
irrelevant" when you say "the box contains nothing".
have infinitely precise meanings? Can those infinitely precise meanings
Lojban gismu do not have absolutist meanings.
be obtained from the English definitions of the gi'uste?
I know of no absolutist dictionary. I wouldn't want to use one. Unless
it was really, really good, and basically had a whole volume dedicated
to every word. In which case I'd use it to resolve life-threatening
conflicts of meaning or something.
> > longer think of an objection, what does that prove? Only that I'm not
> > imaginative enough, not that the description you finally gave is precise to
> > an absolute degree.
>
> If I were you, I too would be anxious to show this pattern. But it
> hasn't emerged, so you're jumping the gun. And conversely, if it does
> emerge, you finding problems with my restrictions would only show that
> I'm not skilled enough in Lojban and in restrictions to provide a
> decent one. So hopefully I can show that you can't show anything wrong
> with these restrictions right off the bat.
Do you think that if I can't find an objection at some point, that shows that
the proposed description is complete?
No, but if neither you nor anyone can find one in the first example
that I give, I think this is adequate reason to think that my method
does a much better job of being certain when certainty is required
than your lone method.
> (1) A: Put (all the black stones that are cpana this table now) in
> this black bag
>
> Here's a rough idea of the situation, so that you don't think I'm
> changing anything: There's one table and one black bag in my
> perception. All the black stones supported-from-below-by the table are
> from the game-set (and therefore fit in their bag). There are no
> stones in the black bag, because they were dumped out, so saying just
> cpana is exactly what I want to say.
>
> So, what have I failed to restrict?
For all practical purposes, nothing. In an absolute sense, I don't know,
What do you mean you don't know? You think that the stones might
actually be dried berries/plastic imitations, or what? I assure you
that I'm talking about typical stones, and that they fit both our
models of what stones should be very well. Yes, a crazy person might
exist that calls them elephants, but between you and me, I think that
this restriction is 100% complete.
because I don't have full access to the situation that you are imagining.
I assume for example that you are not imagining any stones that can
just disappear before I grab them, or stones that may turn into bears, or ...
but how can I be absolutely certain?
I've said that my perception of any language requires that anything
that isn't part of a zasti-type relationship (..imaginary,
hypothetical...) have an implicit zasti, so no, they aren't imaginary.
And they aren't holograms, I would have called them stone-holograms if
they were.
> Yes, I see that it makes the spatial tense "sticky", among other
> things, which is how it would be translated into Lojban, so your
> objection that said something like "it's implied but never stated
> which bears, but we *assume* that it's bears in the..." is not
> accurate. We don't assume, it's stated right there.
It's part of the context, yes. It would not be part of the sumti in question,
No, it's not part of the context. Adding a sticky tense or redefining
a word has nothing to do with context. Context originates from things
that we say, stickyness and redefinitions originate from what we say.
Stickyness and redefinitions do not originate from context.
so inner {ro} could not be used to escape the context.
> If it's in his imagination, and he's not the one that spoke the
> command, when what business does he have telling the admiral (as I'll
> call him) what the admiral meant when he said it? Use the perceptive
> definitions of the *speaker*, as best as you can guess at them (yes, I
> did just admit that there is vagueness in Lojban that I'm not trying
> to cover).
If only the speaker can determine what they mean, then I don't understand
Unless we're on the "catch massively mutated bears, hyenas, and
tigers" squad, or one of us is all sorts of insane, I'm going to be
able to determine what you mean by "pass me that bear" just fine
(though odds are that I won't comply, since they're usually pretty
heavy and dangerous). Just as if we were sitting at a table in a
garden, and you said "this table", I'd know that you meant "the table
we are sitting at" just fine. It's not that only the speaker can
determine X, it's that we should be using the speakers determinations
of what X is in order to understand the speaker. Just as if we met a
crazy person who thought all knives were forks and all forks were
knives, we'd understand him best if we used his determinations of what
a fork is (i.e. a knife).
what you mean by a "complete" description. It seems that to everyone else
the description would be indistinguishable from an incomplete one, so why
bother with it?
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.