* Monday, 2011-10-17 at 19:46 -0700 - John E Clifford <kali9putra@yahoo.com>:
> ----- Original Message ----
> From: Martin Bays <mbays@sdf.org>
> To: lojban@googlegroups.com
> Sent: Mon, October 17, 2011 5:37:30 PM
> Subject: Re: [lojban] {zo'e} as close-scope existentially quantified plural
> variable
>
> > * Monday, 2011-10-17 at 15:25 -0400 - John E. Clifford <kali9putra@yahoo.com>:
> > > {mi nelci lo ka cinfo} doesn't get at any meaning of "I like lions" as
> > > far as I can see; it's about liking something else, lionness. Now,
> > > I suppose that liking lioness might be one factor in liking lions, but
> > > surely not the whole of it. I might like their feet or the way they
> > > taste roasted or ..., none of which are lionness.
> >
> > Maybe it isn't a very good translation of the english. But I don't see
> > why {mi nelci lo cinfo} with the kind (in xorxes' sense) 'lions' as the
> > referent of {lo cinfo} shoud be any different.
>
> Nor do I; I hqave never -- except for a few periods when he was working around
> to something that seemed to make sense -- been a fan of xorxes' kinds (or any of
> the other names the same sort of thing has been called over the years). As
> usual, I am unsure what exactly he means this time, so I can't say it is wrong,
> but I certainly won't say it is right.
>
> > > Actually, I have no problem with interpreting some occurrences of {lo}
> > > as kinds (as understand that notion). But, the nice thing about
> > > slipping into intensional contexts is that words now stand for their
> > > intensions, not their extensions. But, even if they did refer to
> > > their extensions, they would still be extensions in some idealized
> > > range of worlds.
> >
> > Maybe I finally understand what you mean with your "kinds = maximal
> > bunches" idea. Let's see.
> >
> > I've been implicitly assuming that in {lo broda}, the tense inside the
> > description is by default copied from outside it. So {mi ca ca'a nelci
> > lo pavyseljirna} == {mi ca ca'a nelci lo ca ca'a pavyseljirna}, which is
> > false if there are no unicorns.
>
> I suppose the tense (if there is one) is as contextual as everything else about
> descriptions. The same as the bridi surely is a good guess in general, but may
> be obviously wrong in other circumstances. For example, in generalities, the
> tense (if that is the right notion) is probably past, present, future and
> possible.
Right, so I think I do understand you.
Does this work?
Let's try "dodos are extinct", which I'll render lazily as {lo cipnrdodo
cu tcinrekstinkta}. So we might reasonably glork this as {lo pu
cipnrdodo cu tcinrekstinkta}. If {tcinrekstinkta} just means what {ba'o
morsi} does, then we win. If it's true of a plurality iff no instance of
any of the species there represented have living representitives, we
also win. If it's only true of pluralities formed from all ever-living
representatives of a species, which maybe is closer to the english, then
we win as long as {lo [pu] cipnrdodo} is glorked sensibly.
Similar story with 'widespread'.
{lo cinfo cu citka lo mirli} could work too.
(mirli == deer/elk/moose, in case you don't have your lists to hand.)
So we can get the effect of kind/property predication by just using
plural semantics. Neat.
The 'resolution to generic predication' aspect would have to just be
part of the semantics of the selbri... so we retain xorxes/chierchia's
idea that pure-kind predication blocks generic predication.
Similarly with existential resolution as disjunctivity. As, I now see,
you were saying all along.
We can't get {lo cinfo cu na zvati le mi purdi} meaning "no lions are in
my garden" unless {zvati} is disjunctive in x1, which it blatantly
isn't. But that's no great loss - we have {su'o} (or {pi za'u}) for
that.
We don't replicate xorxes' "paired predications" like "dodos ate my
ancestor's face; thankfully, they're now extinct" - but I think we can
live with that.
And arguably we do get it in some circumstances anyway: in that example,
if {lo cipnrdodo} gets all dodos which ever actually existed, as above
above, then perhaps it's true that they ate my ancestor's face, because
at least one of them did. I'm uneasy about allowing "ivities" to vary,
though, and we might not always want the x1 of {citka} to be
disjunctive.
Disambiguation becomes just a matter of specifying the tenses in the
description, which is theoretically easy but may be painful in practice.
Maybe other gadri could indicate some default assumptions about the
inner tenses.
I think I like this solution.
> > Now I'm guessing that you want to exploit the fact that no tense is
> > actually specified in {lo pavyseljirna}, and hence it could very well be
> > {lo pavyseljirna be ca da bei ca'a de} - "things which are unicorns at
> > some time in some possible world". Liking that plurality, even if it be
> > in a situation in which none of its constituents zasti, could indeed
> > mean something. It could plausibly even mean whatever it is that it
> > means to "like unicorns".
>
> Well, the use of a description in a primary context means that something
> satisfying that description is in the domain, so, if I am going to say {mi nelci
> lo pavyseljirna}, I am committed to there being unicorns (not necessarily
> existing ones). And, of course, things mentioned in secondary places are going
> to be in the domains associated with those secondary structures, without having
> any necessary effect on the primary domain.
Why would you want a separate domain for the {lo} structure?
> I note, by the way, that {nelci} has no place for what one likes a thing for,
> which explains some of my problem with questions like whether I like lions: for
> some things yes, for others not so much. Lacking this information makes it hard
> to know what to put in the object place. The [tu'a} approach has the virtue of
> covering a lot of cases, but not the one that is apparently central to {nelci},
> eating, since that is very concrete.
Right, and now {nelci tu'a lo pavyseljirna} does make sense.
> > Is this what you've been getting at? If so, sorry that I've been too
> > dense to see it before.
> >
> > I'm not sure what to make of it as a solution to kinds issues... let me
> > ponder.
> >
> > > Of course, we can do without the {tu'a}, but it is
> > > less plausible, even with lion, let alone unicorns. How would an
> > > abstraction involving lionness be closer to liking lions than one
> > > involving lions?
> >
> > Because the abstraction in question could be {lo nu citka lo ckaji be lo
> > ka cinfo ku ku joi lo xamgu ke xunre vanju}...
>
> Even without taking the time to translate this, it is clear that this is going
> to have very little to do with liking lions, except in the sense of providing a
> recipe for lion ragout, which I might like. Ahah, maybe this is getting back to
> the issue of what I like lions for and spelling it out in as part of what I
> like. But, of course, this would work just as well as {lo nu citka lo cinfo ku
> joi lo xamgu ke xunre vanju}
I now agree.
> (actually, isn't it white wine with cats?)
That's bats.
Martin
Attachment:
pgp12WZ4y86_f.pgp
Description: PGP signature