[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
Nick:
> --- In jboske@yahoogroups.com, "And Rosta" <a.rosta@l...> wrote:
>
> > So where you and I differ is with the counting types, which IMO are these:
> >
> > A. you can start to count them and you can stop counting at su'eci'ino
> > B. you can start to count them and can't stop until ci'ipa
> > C. you can't even start counting (but {da poi broda} still holds)
> >
> > Individual is A. Bit of Substance is B. Substance is C. Most noematic
> > things are Substance/C
>
> Individual is A. Bit of Substance (which I call Substance) is B. Kind
> is C, and
> includes Substance. (I'm not convinced da poi broda does hold, but
> whatever.)
I'll come back to this in another message.
> > Collective (as I understand it) is not a special counting type
> > Collectives can be either A or C. Member-of-Collective is A or
> > B.
>
> Collectives, I think, are still B, because both Collectives and bits
> of substance
> are Masses (they have memberships); but I'll have to think some more.
But collectives of a fixed internal cardinality can be counted and
are type A.
> > Without wanting to be unduly negative, (a) "the whole" can mean
> > "every bit" in English,
>
> But is this pragmatically conditioned? As in, it only does so when the
> predicate is distributive?
Good question. Certainly "the whole of the cake contains a raisin"
!= "every bit of the cake contains a raisin".
I suppose what I had in mind is that "the whole" is used only to
mean "not only part of", in contexts where "only part of" would
otherwise be plausible. Perhaps I was overhasty in glossing this
as "every bit".
> > (b) taking it to mean something other
> > than "every bit" simply does not square with it being a fraction
> > and a quantifier,
>
> I now think something deeply, deeply different is going on with fractional
> quantifiers. They describe portions; they don't actually quantify jack. The
> quantification isn't there, because their referents are really Kinds.
> Maybe of
> different kinds to the Kind of Substance. As in, we have Kind of
> Atom, Kind of
> Stuff, and then we have the fractionally quantified Kind of Substance
> of Stuff,
> and the quantified Kind of Individual(s) of Atoms and Kind of Collection of
> Atoms (I seek three mermaids, and so on.)
From your description of them in another message, they sound like
what we could call 'Extensional Kinds' -- the manifestations I've
been talking about.
> > (c) if it doesn't mean "every bit" then there
> > is no difference between "the whole of X" and just "X"
>
> As long as X is extensionally defined, that is indeed the case
>
> > > Especially not when I've used it to finally reconcile with
> > > collectives and emergence
> >
> > Ideally the reconciliation would be one that is based on an
> > internally consistent system rather than one that applies
> > idiomatic meaning to constructions that could have had a
> > sensical compositional meaning
>
> My story is now, there is a sensical compositional meaning, and it's rather
> more intricate than we'd thought
That's good news, but I haven't grasped it yet.
> > > With respect, Jorge, that's bullshit. p. 130 of CLL says "the whole".
> > > CLL makes explicit that collectives are always fractionally
> > > quantifiable. And collectives are utterly analogous to sets (we
> > > considered conflating them, and sometimes the gi'uste does); p. 131
> > > explicitly says that pimu lo'i remna is a set. So lojbanmass is closed
> > > under fractional quantification, just like individuals are closed under
> > > natural number quantification
>
> > Without going into a detailed response at this stage, you must
> > consider what strategy to take when what the CLL says makes sense
> > only as a stipulation of idiomatic meaning that overrides sensical
> > compositional meaning.
>
> Two strategies. One, I lump it, and say that's life (and lose you
> guys). Two, I
> say CLL is shite, and lose SL-compatibility (which I will do only on extreme
> brokenness, because otherwise it loses everyone else.) Three, I perform a
> miracle
>
> I think I've got such a miracle cooking now; we'll see..
You never know, especially if it is possible to exploit areas where
CLL left things undefined. (I mean: yes, the miracle can happen.)
> > When xorxes talks about his understanding of piro, he is talking
> > about his understanding of what it would mean if CLL stipulations
> > didn't substitute an idiomatic meaning for a compositional meaning
> >
> > Here we are running up against exactly the problems I foresaw. I
> > thought you could create a sensible gadri system ab initio, but
> > not one that would be conformant with CLL.
>
> You keep saying English is more logical than people give it credit
> for. Half the
> glass of water means something. Lojban pimu loi djacu means the same
> thing. It doesn't mean what you would do in straightforward extensional
> quantification using 1/2. It describes the portion. If Lojban is
> broken here, so is English
Taking the Lojban alone, if {pimu loi djacu} means half the glass of
water (though it doesn't, does it -- it means half of all water), then
it would be what you get if you take pa loi djacu and chop it in
half. And re loi djacu would be what you get if you take pa loi
djacu and duplicate it. In that case, {piro} would mean the same
as {pa}, which is not unreasonable, since n/n = 1. So Q lo(i)
would mean "Q * a single". That is perfectly coherent to me (today).
But you say it works like English. I don't actually know how this
bit of English works; I've not studied it except cursorily. Although
I do in fact think English is more logical than people give it
credit for, by Ideal Loglan/Lojban standards of logicality it is
not so logical, and the claim I usually mean to make is in fact that
when people speak English they are using English to express logical
meanings -- that is, logic infuses all our thoughts and some of
that logic filters into English itself, though with an as-yet
undiscovered degree of ambiguity and kludginess.
> > Sure, but me and xorxes have been enticed into these exchanges by
> > the appearance of you trying to create a rational system. If you
> > say "Sorry -- CLL makes a balls-up here, but it's got to stay",
> > then I'd probably be better off just backing off and leaving it
> > to you to do on your own
>
> That is an an acceptable outcome of the Yuletide accord: we can part ways if
> you think I'm doing crap. But I think I've got something now..
Yes.
> > Sorry -- I can't remember which messages this was in. It wasn't
> > a single thread. But we've been talking one another in and out
> > of two coherent positions:
> >
> > I. piro != every bit, & is equivalent to the absence of quantification,
> > so should always be the default
> > II. piro = every bit, is not equivalent to the absence of quantification,
> > and should not be a default
> >
> > Currently we favour II, because it is far more consistent with
> > compositionality
>
> I don't think what I'm now saying fits this, and I think it's a false
> extensionalist dichotomy
I think you are (or at least have recently been) saying I. "piro loi"
= "loi when you don't chop it in twain". My objection to a redundant
piro being a default quantifier for loi is somewhat voided by the
reinterpretation of defaults as pragmatic defaults. My only remaining
objection is that it has yet to be proved to me that piPA makes
no sense as a quantifier, but in fact if re loi is two jbomasses,
then I think piPA is a quantifier and does exactly what you want it
to.
> > > I define a substance as an infinitesimal collective, then. I can deal
> > > with that. This may end up meaning that while there are three
> > > ontological types --- atom, stuff, group --- there are only two
> > > counting types: individual... and mass
> >
> > That's an improvement. But what is the counting type of the set
> > {Nick, And, Xorxes}? Neither individual nor mass, IMO
>
> .... or both. What count do you want? One, Whole, or Three?
I see three counting types, as I've said elsewhere.
su'eci'ino
ci'ipa
Mr PA
the counting type of the set is Mr PA. The essential idea of Mr PA
is that it conflates all real numbers greater than 0 into a single
number.
--And.