[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
--- In jboske@yahoogroups.com, "Jorge Llambías" <jjllambias@...> wrote:
>
> On 9/27/06, John E. Clifford <clifford-j@...> wrote:
> > --- In jboske@yahoogroups.com, "Jorge Llambías" <jjllambias@> wrote:
> > >
> > > I'm not sure how the purport of a sentence could be that one of the
> > > referring expressions used (as opposed to mentioned) in the sentence
> > > have no referent.
> >
> > (no da me lo broda} or even {no da du lo broda)
>
> OK.
>
> Unless {no da} is interpreted as being implicitly restricted (i.e as
> {no da [poi ...]) then I would take those sentences to be
uninterpretable
> or necessarily false.
>
> People don't usually utter uninterpretable or necessarily false
sentences,
> so in ordinary circumstances if someone utters such a sentence they
> will be understood as having some unexpressed restriction in mind.
Well, I can see them being uttereed just after someone said something
involving {lo broda}. Of course, they are both equivalent to {no da
broda} and so are true in every model that contains no brodas. They
are neither unintelligible nor false, when there are no brodas. We use
such locutions all the time in English and they don't seem to create
any problems; simply let us know something about the model we are
working in. I knoow that Meinong got into a twist about some of these
sentences, but negative references -- and wishworlds -- aren;'t ones
that raise any problem -- and sure don't require that there be
something that {lo broda} refers to. Work it out theough the standard
definitions.
>
> > > > > Model_0: Does not contain p in its theory.
> > > > > Model_1: Includes p and turns out to be inconsistent, so the
model
> > > > is ditched.
> > > > > Model_2: Includes the theory of Model_0 plus ~p. The
examination of
> > > > Model_1
> > > > > was helpful to determine that Model_0 could be expanded to
Model_2.
> > > >
> > > > "Had to be expanded" is perhaps more to the point. A situation
> > > > ultimately must contain all the logically entailled + claims. But
> > > > that is not entirely relevant here.
> > >
> > > p could be a new proposed axiom rather than just a theorem of the
> > theory.
> > > If it was logically entailed by the truths of Model_0, it was as you
> > say,
> > > already in Model_0.
> >
> > The fact that the reductio works ed showed tht it was.
>
> What if p makes reference to some new object not included in the
> domain of Model_0?
>
> > > When it comes to ordinary discourse, seen as a process of model
> > > building, most new sentences will be incorporated as new axioms
since
> > > what already follows from already extant axioms are usually
trivialities
> > > that noone would bother uttering.
> >
> > The point is, as usual, that we can sometimes add these new axioms
> > without having to expand the domain, even though the addition appears
> > to contain a referring expression.
>
> And from my point of view, Lojban sentences that appear to contain a
> referring expression always do in fact contain a referring expression
> (or else are uninterpretable).
"referring expression" is a technical term in some linguistics' it
means an expression designed to refer, not one that actually does. As
witness the fact that we can ask the question: What is the status of a
referring expression when it doesn't in fact refer? As for what a
sentence means that has a non-referring referring expression in it
means, you can handle it either way for simple sentences (Fa): they
are either false or meaningless (they are false on the standard
definitions -- except that the definitions are not quite standard) or
they are meaningless. In the latter case you then have all sorts of
options about very complex znd compound sentences. In the former
there are no problems (which is why I tend to stick with it).
>
> > I gather that you think that the problem only arises
> > when there are explicit quantifiers involved, which is a useful way to
> > deal with the deep structure.
>
> Basically yes.
>
> > But ordinary languages seem to create
> > these contrasts without making the the quantifiers explicit and I
> > think Lojban should be able to do this as well-- while still keeping
> > -- and making explcit -- the distinction betweeen the two readings.
> > Note that the official line does just that; the problem being only
> > that what that official pattern points to does not seem to be the
> > right sort of thing.
>
> Yes, the official (CLL) line has {lo broda} = {su'o lo broda}. That's
> why I don't like implicit quantifiers.
I was actually thinking about the difference between {mi djica lo
broda} (wide scope) and {mi djica tu'a lo broda} (narrow). I agree
that {lo broda} and {su'o lo broda} have different pragmatic functions
-- and both are different from {su'o da poi broda} (they are, however
materially equivalent). I don't like implicit quantifiers either,
mainly because they seem always to be the wrong ones on {lo}.
> > Still, we can mark the difference and worry about
> > what is actually going on later.
> > The interesting fact is that the two quantified forms are not
> > interchangable even when there is only one unicorn in the domain. I
> > can want a unicorn (narrow reading) other than the unicorn there
> > happens to be -- to deny this flies in the face of what we all do
> > constantly.
>
> From my perspective, "a unicorn other than the unicorn there
> happens to be" will have a referent too, if expressed with a referring
> term. (Of course you could always say something like
> {ko'a goi lo pa pavyseljirna zo'u mi djica lo du'u su'o da pavyseljirna
> gi'e na du ko'a}, but then there is no reference to a unicorn other than
> the only unicorn.)
>
Of course, generally virtually all words are referring expressions,
you mean only those that refer to individuals, like sumti, not to
classes (or properties), like brivla. But notice that your claim that
every referring expression has a referent means that every predicate
has a non-null extension{da broda} is eqivalent to {da me lo broda)
and {da du lo broda}. So, in the end you mean all content words refer
if they get used at all. Now, admittedly, no one would say some of
these things in the normal course of a conversation, but that doesn't
keep them from being true.
I think you have a pragmatic notion of the domain: it doesn't contain
things, it contains ezxpressions and yhou keep track of all these as
the legitimate expressions in the langauge. That they refer is then
just a wayof saying they are on the list. But that doesn't justify
and the semantic tricks you would play with them, unless quantifiers
and the like are pragmatic too -- just about expressions. There are a
number of logics that have something like that as one possible
interpretation (free logics, empty univers, even intuitionist) but I
don't see any reason to bring them into Lojban.
> > > One can claim that someone wants it or that someone doesn't
> > > want it, but it makes no sense to distinguish "a particular one
of them
> > > is such that the person wants it" from "the person wants just
any one
> > > of them".
> > Sure, if you are talking about wanting *it,* but wanting a unicorn is
> > not the same as wanting it, even if it is the only unicorn there is.
>
> You may not like this usage of English "it", but:
>
> A: I want something.
> B: What is it that you want?
> A: A unicorn.
> (later)
> B: Did you get it?
> A: What?
> B: What you wanted, a unicorn.
> A: No, I didn't get it yet.
>
> sounds perfectly normal English to me, even with the "any unicorn"
sense.
I agree but the question is -- as always with thee cases -- just what
"a unicorn" means. "Any unicorn" generates a peculiar set of
transformations -- actually, two sets and I am unsure which one works
here. At the very least, this about what I eant throughout and that
has to be some intensional object -- whether a unicorn type or an
event sense or whatever. Clearly, so long as we keep referrring back
(and foreward) to objects of desire, this discussion presents no
problems -- and doesn't require that there be any unicorns at all. If
A dioes get it, then there is a unicorn.
>
> > > Our
> > > disagreement is about whether or not we can have a model where
> > > a single thing is the referent of an expression with that single
thing
> > > having a generic or type interpretation.
> >
> > In fact, we haven't talked about this at all for some time, and not
> > very precisely even earlier. If this is what you think the arguemnt
> > is about, then you really need to explain what a type or a generic
> > interpretation of a thing is.
>
> I thought that's what I have been doing.
I don't remember anything about types for several rounds nor about
generics at all, I certainly haven't seen any need for them and you
have neither shown a need for them nor shown how they would help (nor,
of course, what the Hell they are).
> > {lo broda} refers to brodas, as you
> > kept insisting a few threads ago, to the several of them (how many,
> > which ones, even whether existent or not, not specified),
>
> Not "the several of them". It may be the one and only too. Number is
> not specified in any way.
Right -- I am just trying to avoid vcalling them a bunch, which used
to get you upset -- though without any cause.
> > not to some
> > one thing of which these were parts or some such relation. Has all
> > this changed? Why? And why would we want this unknown object in
Lojban?
>
> I want it because it makes the language expressive and easy to use,
without
> losing its logicality.
Well, from what I have seen of this notion in the past, it destroys
the inherent logicality of Lojban by introducing an entirely new thing
into the semantics, breaking all the associations of various forms.
But the real problem with it is that it is pointless. There is no
problem that this is going to solve. The only problem right now is to
find a good clear way to distinguish two inferences. I can't figure
out whether you have a solution to that or not -- you seem generally
to be saying that there aren't the two inferences or that only one of
them occurs in Lojban. And then go haring off after some other
inference that doesn't occur in Lojban because it involves things not
in Lojban.