[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
Nick:
> .... If propositionalism covers everything. Like I said, by the time you
> get to depiction, I doubt it does
Depiction/resemblance is the hard case, but we have no decent solution
for that yet, or at least none that have penetrated my skull.
> > Message: 7
> > Date: Thu, 26 Dec 2002 18:48:26 -0000
> > From: "And Rosta" <a.rosta@hidden.email>
> > Subject: propositionalism redux
>
> > 3. Depicting (e.g. "This depicts a snake"). As Nick suggests,
> > we can treat this as "There is something that in some world
> > (not necessarily the local world) is a snake and that in the
> > local world this depicts"
>
> Of course (weaseling), depiction need not be intended as depiction of
> any one individual anyway, real or imaginary. It can be just a
> construct based on the prototoype..
>
> > There was already a need for a way to do this in Lojban,
> > so that we could talk about imaginaries such as Sherlock Holmes
> > without having to abandon the distinction between the local world
> > and worlds that from the perspective of the local world are
> > fictional. The two ways that have been proposed for doing this
> > are {da'i} and {ka'e}/{nu'o}. Both are unsatisfactory for two
> > reasons. The first reason is that they mean, or should mean,
> > something else. {ka'e}/{nu'o} pertain to Possible Worlds
>
> How is this a problem? Because fictional worlds are not the same thing
> as possible worlds?
Yes. That's why I capitalized to Possible Worlds. PWs are relevantly
similar to the local world, differing only in specified and contextually
relevant ways. PWs give us various sorts of conditional meanings.
> > {da'i}
> > is in UI and therefore has something to do with illocutionary
> > meaning
>
> A constraint already violated with {kau},
But {kau} has no independent meaning and is in UI for syntactic reasons.
{da'i} has a mahoste gloss that is perfectly compatible with it having
illocutionary meaning.
> and indeed with several UI
> before that (since UI is a dumping ground for odds and sods.)
& even more disgustingly, with {po'o}.
> An illegitimate objection, especially for Microsoft Lojban
Yes, if you delete the "especially".
> > The second reason is that they don't allow us to
> > distinguish "For every x there is some world w such that in w
> > x is broda" and "There is some world w and for every x that in
> > w is broda". For example, "For every Danish mermaid, I will
> > write a poem about her", normally wouldn't mean I will write
> > an infinite number of poems, one for every imaginable Danish
> > mermaid
>
> (a) we can easily make the second the Gricean default;
We can't make anything be the Gricean default. If it's within our
control, it's not the Gricean default.
> (b) of course we
> will need to introduce machinery for possible-world reasoning into
> Lojban. Although..
>
> > What we need is a selbri, "x is world of which y is true"
> > It could be a lujvo, but I'll define a NU, {jei'u}, to do
> > the job: "x1 is a world of which the abstraction is true"
>
> ... it is only because you've been working in Academic Lojban for so
> long that you can do this so blithely. And, you know the fundies see
> you invent cmavo and think you're a crackpot. Try using the lujvo for a
> while, as a political move; you'll make it harder for the fundies to
> dismiss you. (And you know they're dismissing you.) In fact, stick
> close to English for a while, and use {munje} and {vasru}. (This goes
> for Jorge's loi'ei too.)
I am happy for the fundies to dismiss me. MS Lojban already makes me
feel palpably nauseous, and I don't really care about whether it
adopts one or two of my ideas or not. The people working to finish
SL are free to use my ideas or ignore them.
> > This then gives us:
> >
> > da ro de poi da je'u de is Danish mermaid zo'u I will write
> > a poem about de
>
> su'o da poi munje ku'o
> ro de poi gugdrdanska fipni'u
> zi'e poi da vasru de
> zo'u: mi te pemci de
You need to say that LEdu'u de gugdrdanska fipni'u is true of da.
> > This solution is not always satisfactory, though. Consider:
> >
> > This branch has the shape of three intertwined snakes
> >
> > It is not enough to say there are in some world three snakes
> > that have the shape of this branch. That statement would presumably
> > be true whatever the shape of the branch
>
> Eh?
>
> You gotta mold the snakes into the shape you want in some world or
> over.
>
> I simply cannot see the problem with:
>
> su'o da poi munje ku'o
> ci de poi since
> zi'e poi da vasru de
> zo'u:
> le ti tricyspi tarmi simsa
> lu'o ci de poi gunma torni [vi da]
The problem is that virtually any shape is the shape of some three
snakes in some imaginary world. I don't deny that Grice can come
to the rescue somehow, but He needs to be given a helping hand by
what we actually say.
My latest thought is something like "The branch is such that if
we were to judge by shape alone it would be true that it (the branch)
is three snakes" (= "The branch is shaped as though it were three
snakes").
> I agree that the hangup on the depiction actually being a depiction of
> individual, real or imaginary, is probably bad; but this is not your
> killer counterexample. You can make it the Unique; I like the referent
> of the Prototype, since we are dealing in mental constructs (and as a
> Kind, even if individuated, the Unique is not a mental construct)..
I have no idea what "the referent of the Prototype" should mean.
I am not satisfied with Unique as a solution. If branch X is shaped
like three snakes and branch Y is shaped like three snakes, I do
not want to have to claim that the two branches are shaped like
the same thing (which is what I would be claiming if I claimed
they were shaped like the unique snake trio).
Note that "is shaped as though it were three snakes" avoids this
problem.
> > 4. Psych-predicates
> > "John reveres the authors of the American constitution but John
> > doesn't know who authored the American constitution."
>
> Of course, it's a bit disingenous to make a special case of psych
> evaluation preds, because this is a classic extension/intension problem
> that obtains just with djuno:
>
> la djan djuno ledu'u ro finti be le merko flajicmu cu morsi
> kandyselskapi nakni
> ..i la djan na djuno ledu'u ma kau finti le merko flajicmu
It's not disingenuous, because with djuno we already have the
solution, thanks to the du'u complement. Your Lojban sentence
gives the opaque reading. To get the 'transparent' reading
you must change to:
ro da poi finti le merko flajimcu zo'u la djan djuno ledu'u
da morsi kandyselskapi nakni
or (which is not synonymous with the other 2 readings):
finti be le merko flajimcu be'o cei broda ro da poi broda zo'u
la djan djuno ledu'u da ge broda gi da morsi kandyselskapi nakni
The reason why psych-predicates are pertinent is that they don't
normally have a propositional argument.
> > 5. "mi djica LEnu broda" means "I want that LEnu broda be actual (be
> > fasnu)". But which nu broda? -- *Any* one nu broda. How do we
> > express this? There is no propositionalist solution. One solution
> > is to kill quantification by singularizing nu broda (e.g. by
> > {piroloinu}
>
> Or by your Unique, since Prototype is now out of bounds for this
Yes.
> > A better solution is to use a predicate that means
> > "I want that p be true", which has the benefit of allowing nu to
> > behave like all other predicates in having an extension that varies
> > from world to world. Thus: "mi -wants LEdu'u broda"
>
> You mean,
>
> mi djica ledu'u su'o da zo'u: da nu broda ?
>
> God I hope you're wrong..
Yes, except it doesn't have to be the gismu djica that does this.
> > Message: 8
> > Date: Thu, 26 Dec 2002 18:48:28 -0000
> > From: "And Rosta" <a.rosta@hidden.email>
> > Subject: RE: big rethink on Unique and other gadri
>
> > The pisu'o default on loi is unfortunate, then, not just because
> > piQ looks to be generally mabla but because we have the strange
> > situation whereby:
> > {loi} means "a fraction of loi"
> > {piroloi} means "loi"
> > So all {piro} does is cancel the implicit {pisu'o}
>
> Gricean salvator again. Don't read in any implicit quantifier where it
> makes no sense. And do no such reasoning, unless you insert all
> explicit quantifiers: {loi} by itself has no meaning, unless it's
> quantified
CLL
{loi} = {pisu'o loi} & means "pisu'o loi"
{piroloi} means "piro loi" = "loi"
you:
{loi} ambiguous between {pisu'oloi}, "pisu'oloi", and {piroloi}, "loi".
You are deviating from CLL. If CLL is to be changed here, better to
make bare {loi} mean "loi" (= "piroloi").
> .... Argh. Now I feel like crap. because Jorge liked my gadri summary.
> But obviously I am getting aggro because I'm seeing myself and Jorge in
> incompatible missions --- fundie vs, revisionist; so our respective
> keywords set the other off. This cannot happen in the BPFK; it needs to
> stop. We're working on Microsoft Lojban together, or not working
> together at all
We're doing both -- that is, some of the time we're working on MS
Lojban together and some of the time we're not all working together.
> > Message: 13
> > Date: Fri, 27 Dec 2002 01:02:00 -0000
> > From: "And Rosta" <a.rosta@hidden.email>
> > Subject: RE: propositionalism redux
>
> >>> It is not enough to say there are in some world three snakes
> >>> that have the shape of this branch. That statement would presumably
> >>> be true whatever the shape of the branch
> >> I'm not sure I see the difference between a picture
> >> depicting three snakes and a branch resembling three
> >> snakes
> >
> > A depictee can be the 'subject-matter' or the 'iconically
> > signified':
> >
> > "This picture is about X" (subject matter)
> > "This picture looks like X" (iconically signified)
> >
> > Texts usually don't iconically signify, but do have subject matter
> > Branches don't have subject matter, but do iconically signify
> > Pictures typically iconically signify their subject matter. So
> > 'picture' is ambiguous
>
> Yeah, so there's a difference in there being an intended referent for
> subject matter and an overt one for icons.
My kitchen tap sounds like an ululation of Xena Warrior Princess, but
I don't think of my kitchen tap having a referent. Rather than talking
of referents I'd rather say that my tap sounds as though it were
an ululation of Xena. (The fictionality of Xena is irrelevant to this
example.)
> I don't get how that affects
> whether imaginary-world quantification is sufficient or not to handle
> it. Sure, subject matter can include stuff you completely make up, with
> no referent in this world; but if you read enough fiction, so can
> icons. A tree branch can look like an elf too. In fact, it's the book
> rather than the branch that is likelier to have an individual referent:
> the book is about a specific elf, the branch looks like Any elf
Exactly. It's the branch that was the problem for propositionalism,
and the propositionalist solution I now want to push for the branch
is "shaped as though it were".
> So maybe depiction is intensional in the Any-x sense after all
Depiction in the sense of 'semblance' does (or can) have the any-x
sense, but I am not satisfied with a nonpropositionalist gadri
as an underlying solution for expressing any-x.
--And.