* Saturday, 2011-12-03 at 11:06 -0800 - John E Clifford <kali9putra@yahoo.com>:
> > From: Martin Bays <mbays@sdf.org>
> > Sent: Fri, December 2, 2011 9:24:56 PM
> > * Friday, 2011-12-02 at 09:20 -0800 - John E Clifford <kali9putra@yahoo.com>:
> > > From: Martin Bays <mbays@sdf.org>
> > > Sent: Mon, November 28, 2011 7:34:42 PM
> > > > * Sunday, 2011-11-27 at 22:59 -0600 - John E. Clifford
> > ><kali9putra@yahoo.com>:
> > > > > On Nov 27, 2011, at 10:36 AM, Martin Bays <mbays@sdf.org> wrote:
> > > > > > * Friday, 2011-11-25 at 12:38 -0600 - John E. Clifford
> > > > ><kali9putra@yahoo.com>:
> > > > > >> On Nov 24, 2011, at 9:44 PM, Martin Bays <mbays@sdf.org> wrote:
> > > > > >>> * Wednesday, 2011-11-23 at 13:34 -0800 - John E Clifford
> > > > ><kali9putra@yahoo.com>:
> > > > > > OK. So we need a meaning for {zo'e} which has each of these two as
> > > > > > special cases. This could be "particular quantifier over a domain
> > > > > > I (could) have in mind" - a special case being that it's quantification
> > > > > > over a singleton domain, so equivalent to it just being a constant
> > > > > > I (could) have in mind.
> > > > Well, what I really meant was the dreaded close-scope existentially
> > > > quantified plural variable - "close-scope" dealing with the interaction
> > > > with other quantifiers, and "plural" dealing with the bunch issue (i.e.
> > > > I did mean a domain of quantification consisting of a single bunch
> > > > when I said "singleton domain").
> > > Not sure why dreaded, they just don't fit sometimes.
> >
> > Could you give an example?
>
> Example: {xu do klama le zarci} (mi klama), which, under Gricean rules has to be
> either x2 = {le zarci} or is terribly rude.
Why doesn't that fit? It's just a constant, equivalently existential
quantification over a (plural) singleton.
> The whole distinction between singular and plural quantifiers is suspect. It
> may mean that the whole system has to be taken as second order, though I think
> that can be avoided.
I don't know what you mean by "suspect". Certainly plural quantification
is effectively equivalent to monadic second-order (Boolos has his name
attached to this). That's one reason for wanting to keep usual {su'o}
and {ro} as singular quantifiers - we don't want to be using
second-order quantifiers willy-nilly.
> What cases do you expect other than "something" and "what I have in mind"? I
> can't think of an example that doesn't fit those two, though I haven't thought
> very hard.
"something satisfying the predicate I have in mind". This would
handle cases where xorxes would appeal to kinds, and you to distributive
predication. I'm not sure how necessary it is.
> > {du jo du} is because we couldn't find an appropriate one-place
> > predicate.
> for {du jo du}, since all the work is done by {jo}, any one place predicate
> would do.
True, but {du} has the advantage of being one syllable long. For the
purposes of the present discussion, how about we declare {jai'a} to be
the always-true unary predicate (selma'o GOhA).
> > I have no modes. In your terminology, this probably does means that
> > I assume the mode is always collaborative.
> Using collaborative as the default mode rather weakens the force of
> collaboration, by allowing all sorts of extraneous (a value judgement) items
> in. In this sense, "We won" shouted by a drunken baseball fan here after the
> world series this year would be literally true, despite his never even having
> gome to a game or watched one. To be sure, we do sometimes want to bring in
> ancillary folk, but we usually set some limits. Of course, this is not so much
> a logical matter as a matter of what constitutes a 'real group'.
Well, what bunches satisfy a predicate is a matter for the dictionary
writers to decide (or decide not to decide).
> I'm not quite sure how kinds have much to do with disjunctive mode (nor what
> kinds are doing here at all).
I meant kinds in the xorxes-carlson-chiercha sense, so basically english
bare plurals. "lions are in my garden" means some lion is in my garden.
The point is that xorxes uses kinds in this way to explain cases of zo'e
which would otherwise look like existential quantification. (e.g. lo
tadni be zo'e)
> > > For the second, I don't see much hope, except, as you say, introducing
> > > a Skolem function -- and even that may not workm depending on the
> > > rules.
> >
> > I don't know what problem you're seeing here. The rules for having the
> > quantifiers be "innermost" are simple: given a lojban sentence, work
> > with any {zo'e}s as if they were constant terms; once we have translated
> > the sentence to a sentence in an appropriate logic, handle {zo'e}s
> > by replacing an atomic formula of e.g. the form
> > "broda(zo'e,a,b,zo'e,c,...)" with
> > "EX X:P(X). EX Y:Q(Y). broda(X,a,b,Y,c)".
>
> Problem with Skolem functions.
There are no Skolem functions here, in the end.
> Some definitions only require mention of all
> universal quantifiers, others require all terms (the resulting proofs go
> somewhat differently). If we go the first route, we cannot keep the little
> buried particular out of having scope over subsequent quantifiers. It is not
> easy even with the second solution, since, in normal form, this particular will
> then bbe in the scope of subsequent quantifiers, which isn't right either.
>
> > Anyway, I am unhappy to note that without allowing either a disjunctive
> > mode or kinds (and they come to approximately the same thing in this
> > case), I don't see a way to understand {lo tadni} in {lo tadni goi ty cu
> > sruri lo dinju}. Presumably it isn't assumed that they all study the
> > same thing, nor that they otherwise collectively study anything, and yet
> > we are claiming {ty tadni zo'e}.
> The problem with {lo tadni be zo'e} is back to the interplay of modes and
> quantifiers and all that. {zo'e} can perfectly well be a buried particular
> here, provided it is not allowed to be pulled across the conjunctive mode of the
> predication (here is a place where quantifiers can't do the work modes do; I
> think there are others). But that sort of restriction is hard to do in standard
> logic, where all quantifiers have to be prenex at least to the smallest wff in
> which they occur.
>
> Curiously, I see your maneuvering as exactly complicating a simple system. We
> have to have the various modes of predication just to handle plural reference
> and then some more to deal with generalizations of various sorts. So the only
> problem, given that, is how to fit modes into the satisfaction rules.
> Apparently, that requires either another set of rules (replacing the usual set)
> or a replcation of various predicates or some other complication. But that is
> unavoidable with plural reference, so it is just a cost to the system.
If this is true, it seems a very good argument against equipping lojban
with plural reference! If we're to have these modes in the semantics, we
surely have to be able to mark them. Lojban currently has no facility
for that, and adding one would involve complicating the language
significantly. Even if we imagine for argument's sake that we wanted
only to mark the two modes 'conjunctive' and 'disjunctive', consider the
problem we have to solve. In a sentence, which thanks to eks and giheks
etc may involve multiple selbri and multiple sumti for each place of
each selbri, we have to be able to specify for each selbri-sumti pair
the mode involved - and moreover, we have to be able to specify the
*order* of these modes, since they don't commute.
Compared to this, xorkinds seem innocent.
> Anything else, though, seem prodigal. And, if we don't mark somehow,
> the different modes then we have either ambiguous claims or we have
> a default value. Ambiguous claims sound terrible logically, but
> aren't usually too bad practically. We have gotten along with out
> explicit markers so far, so we can usually work things out from
> context, using common sense (and the Grice rules, if necessary).
> Picking a default is more difficult, unless we use the weakest for of
> collaborative predication (which I don't really like at all,
> preferring one in which the members of a collaborative group actually
> participate -- still vague, of course -- in whatever they are said to
> do/be collectively).
Attachment:
pgpJoIOYATr0q.pgp
Description: PGP signature