[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
--- In jboske@yahoogroups.com, "Jorge Llambías" <jjllambias@...> wrote:
>
> On 9/21/06, John E. Clifford <clifford-j@...> wrote:
> > Note -- your deliberately hiding this point is getting annoying --
> > that it is not the sentence but the inference from {mi djica lo
> > pavyseljirna} to {da poi pavyseljirna zo'u mi djica da}. The first
> > sentence may well be true in the real world (I've known some people
> > who at least claimed to want a unicorn), but the second ssentence is
> > false, so the inference is invalid.
>
> No, in my interpretation of Lojban, there is no model for which
> the first sentence may be true and the second false. I don't know
> why you say that I'm hiding this point because I have already stated
> it many times before.
>
You wrote -- in the piece here omitted -- that a certain model (the
real world and maybe the numbers only ones) did not count, since in
them the first and second sentence were alike false (or perhaps
meaningless). The point was that there are models in which the first
sentence is true (the real world for one) and the second false. By
talking about the truth of the individual sentences, rather than the
validity of the the argument, you trivialize the point.
I find you interpretation of Lojan bizarre. Lojban does not seem to
me to be that different from English -- for which ther are many models
(the real world for one) that invalidate the argument. Now, I also
think that the fact that it is to be a logical language militates
against your interpretation (whatever it is -- we only know one upshot
of it so far)since the sorts of things that lie at the heart of the
problem with {mi djica lo gerko} are known and incorporated bieces of
logical apparatus. To do away with them by whatever means (and I take
it it involves changing the meaning of {lo gerku} -- against your own
claims about it -- in a way that changes the whole metalanguage -- the
very nature of the models used seems false to that "logical" goal. If
Lojban is properly a logical language, there is not problem and,
indeed, the inference is valid, because writing mi djica lo gerku}
rather than {mi djica X gerku} is to take the {lo gerku} in its
wide-scope, specific reading. If, contrary to expectations, that is
all you mean, then I aggree heartily -- but I have small hope it is.
>
> > > > In short, of course I can claim to want a unicorn in a
> > > > domain that contains no unicorns; to claim otherwise is to fly
in the
> > > > face of linguistic evidence (try picking up that "a unicorn"
with "it"
> > > > in a declarative sentence, for example, let alone by a
quantifier).
> > >
> > > That's not that difficult:
> > > A: I want a unicorn. It's something I have always wanted.
> > > B: But what is it? I've never heard of unicorns before.
> > > A: It's a mythical creature. It's similar to a horse with a horn on
> > > its forehead.
> >
> > There are two explanations for this anomolous case and I don't know
> > which is better. One is that the "it" doesn't pick up the reference
> > of "a unicorn" but rather its sense (what is the reference there).
> > The second is that the reference continues in the want world and does
> > not come out to the initial world at all. In fact, both explanations
> > are probably correct, since they fit together so smoothly. Now try a
> > case that clearly is not about the sense: describing the unicorn, say.
>
> "It's a mythical creature. It's similar to a horse with a horn on its
> forehead" describes a unicorn. You may want to call this an anomalous
> use of "it" with respect to some theory of interpretation, but it is an
> ordinary use of English. If the theory calls it anomalous, that
means that
> the theory has some trouble explaining perfectly ordinary usages.
That is not anomolous (nor is the the first as I think more about it),
they are merely cases of separte -- and well-known modi: that is, the
"it" is as much in a secondary role (governed partly by "imaginary"
and partly by the definition pattern). Now try something that is not
definitional.
> > > > > Two things can be the same F and different G's, that's the only
> > > > > significance of relative identity.
> > > >
> > > > But then it is odd to talk about a different kind of identity,
since
> > > > it is only about comparing things in different way or or different
> > > > things in the same way.
> > >
> > > It's not my choice of terminology.
> >
> > Well, it is you who in this discussion have chosen to introduce this
> > theory, so you are stuck with some responsibility.
>
> I am happy with the theory. I am not especially attached to the labels
> "absolute identity" and "relative identity". If those labels
irritate you, we
> could use "identity" and "sameness", or whatever you find more
> agreeable.
Well, in most cases it really does seem to be identity, so why not
call it that and say that the comparisons are between different
aspects, against different standards, etc., which seems to be true,
unlike the claim that there are two sorts of identities.
>
> > > People seem to be confused by absolute identity. For example, in:
> > >
> > > (1) la superman du la klark kent
> > > (2) ti pixra la superman
> > > ------------------------------------
> > > (3) ti pixra la klark kent
> > >
> > > People say that (1) and (2) seem to be true, that the reasoning
seems
> > > to be valid, and yet (3) appears to be false. Paradox!
> > >
> > > Some people try to get around the paradox with talk of intensional
> > > contexts and all sorts of contortions with (2) and (3).
> > >
> > > Recognition of relative identity allows us to see that (1) (absolute
> > > identity) is false. What is true is a relative identity claim:
> > >
> > > (1') la superman mintu la klark kent lo ka prenu
> > >
> > > but this fact is not enough to conclude that {la superman} and {la
> > > klark kent} are co-referential terms. (2) is still true and (3)
is still
> > > false, and there is no paradox because (1') does not warrant
> > > intersubstitution of terms like (1) with absolute identity does.
> >
> > Well, that seems to work -- except for introducing two kinds of
> > identity where there is only one. What we actually do in this case is
> > something like this "Well, they are the same being, but they look very
> > different, so a picture of one would not be a picture of the other."
>
> Where "they are the same being" is absolute identity, or relative?
I don't make that distinction, so I don't know how someone who does
would do so. They are the same being, they may be different in other
ways.
If
> absolute, you have not gotten away from the paradox, because:
>
> (1) ko'a du ko'e
> (2) ko'a simlu ko'i
> --------------------------------
> (3) ko'e simlu ko'i
>
> If ko'a=ko'e absolutely, then if ko'a looks like ko'i, it would
follow by
> intersubstitution of co-referential terms that ko'e has to look like
ko'i
> too. If by "they are the same being" you mean {ko'a mintu ko'e lo ka
> xadni} or something like that, then you are doing the same thing I'm
> doing.
I am not clear what you are here calling {ko'a}in the second sentence
and {ko'e} in the thrid. If they are the same characterization for
both as in the first sentence, then the argument is valid (though I am
a bit worried about whether {simlu2} is opaque, which might throw
things off a bit. If, as seems likely to get this going, they mean
different things (wholes and slices, say), the the argument is just
Four Terms and properly a fallacy. Notice, not need jhere for relative
identity and the like.
> > (Pictures are new here and they take a different kind of intensional
> > object -- probably -- see the article you recommended).So we identify
> > the body and differentiate the look and it is the look that counts for
> > pictures. That is< "Superman" and "Clark Kent" are ambiguous, not "is
> > identical with".
>
> I'm not saying {du}, or "is (absolutely) identical with", is
ambiguous. I'm
> saying it is false of la superman and la klark kent.
>
Well, that is certainly correct for some senses of those names. We do
generally take this tio be about the whole being, at least as a first
guess, thoug, and then te identity is (in the presumed world) true.
>
> > > > By the way, it seems to me that the right thing to use for
"qua" is a
> > > > modality (which I can't find, suggesting another screw up) for "by
> > > > standard." For the case of Clinton, something derived from {te
xamgu}
> > > > seems to be just right.
> > >
> > > {ma'i}?
> > Maybe {sema'i} but I don't know just what may be involved in
> > association with {manri}. The second place of {xamgu} also looks
> > useful in the particular case.
>
> {se va'u} then.
I guess I did not take the second place as being beneficiary, but
rather as area of value (which I suppose could be thought to require
"good at" though "good for governing the country" works OK in English).
> But in either case, you will need a restrictive: {la klinton pe se
> va'u lo nu...},
> i.e. of the referents of {la klinton}, we select those such that...
Well, I forget the exact rules: I take it I cannot simply attach a
"modal" the {la klinton}. But, come to think of it, it is mnot {la
klinton} that I want to mdoify but what ever the verb was I was using
("admire" "approve of"? -- as usual no Lojban word comes readily to
hand: {zanru} is the wrong "approve). That is, I come up with a
certain evaluation of Clinton against certain criteria -- which
criteria are relevant to only some of his actions, to be sure. But I
still have the sense that I am judging Clinton, not some portion of
one eight-year slice of his life. So the "by standards" really does
seem to be the relevant (though generally missing) notion (although,
if we had a word for "admire" or "approve" it might well have that as
a place already -- but we do this sort of thing with all manner of
expressions -- not just the opposites of approval but off any number
of othetr axes as well and I doubt that that place gets to all of these).
> I think what you may want is different, to restrict not among
referents of
> {la klinton} but among senses.
I don't see the need for senses here; I think these judgments are
really about the referent, viewed in different ways -- perhaps
spatio-temporarily or by roles but mainly against different standards.
> But the only way to do that is to go
> metalinguistic in some sense. You want to make a comment to the audience
> such as "[note that this expression is being used with such and such
sense,
> perhaps not the sense you are expecting]". I think that might be
done with
> {sei}:
>
> mi sinma la klinton sei merko jatna gi'enai speni be la xilaris
> I admire Clinton [US president, not husband of Hillary].
Or (I forgot "esteem" when I was running through synonyms), giving
{sinma} a third an forth place (what for and by what standards though
standards may not fit this case) {mi sinma la klinto lo merko jatna
enai lo speni be la ilaris}.
> It seems to me that any comment about the sense of an expression
> we are using will have to be metalinguistic (at least to the extent that
> anything said in the object language can be metalinguistic). If not, it
> always ends up being about some referent, even if you want to classify
> that referent as "intensional object" or whatever.
Referring expressions alway refer to their referent in the given
context; what that referent is is sometimes not what it usually is.
The goal that the {tu'a} moves were aiming for was to make it so that
the referent was always the standard referent of the expression (by
changing the exprssion, not the contextually specified referent).
{tu'a} is designed for one specific set of problems, opaque contexts.
There are any number of other situations, many not readily
classifiable but dealt with on a more or less ad hoc basis, where the
contextually intended referent is not the usual one. Because these
are not regular patterns, there is no general, prespecified trick for
dealing with them. For ordinary identitiesm we saw that weneeded to
distinguish chunks and guises sometimes -- and probably other things
as well. And certainly other things in other situations. In short,
we cannot always change an expression to account for the contextual,
strange, referent. But we can for one large, frequent, regular, and
logically significant class (and the change fits in with other things
logicians want to hae in their language).