* Wednesday, 2011-10-19 at 21:27 +0100 - And Rosta <and.rosta@gmail.com>:
> Martin Bays, On 19/10/2011 06:11:
> > * Wednesday, 2011-10-19 at 04:59 +0100 - And Rosta<and.rosta@gmail.com>:
> >
> >> Martin Bays, On 18/10/2011 04:26:
> >>>>>> For example, {na ku lo cinfo cu zvati lo mi purdi}
> >>>>>> has at least the two following meanings in terms of actual lions:
> >>>>>> 1. {lo cinfo} is interpreted as a plurality of mundane lions, giving
> >>>>>> roughly:
> >>>>>> For L some (contextually relevant) lions: \not in(L, my garden)
> >>>>>> (which probably means that there exists a lion among L which is not in
> >>>>>> my garden)
> >>>>>> 2. {lo cinfo} is interpreted as the kind Lions, giving
> >>>>>> \not in(Lions, my garden)
> >>>>>> which is then resolved existentially, giving
> >>>>>> \not \exists l:lion(l). in(l, my garden) .
> >>>> Sorry, I was unclear. I meant that English seems to allow only
> >>>> reading (2), and that the same might go for Lojban.
> >>> Ah! Have {lo} *only* able to get kinds, you mean?
> >> Yes. With anything that looks like a 'mundane' reconceived as a kind.
> >
> > So how would you rule out interpretation 1 in the above?
>
> By whatever rules out "it is not the case that Obama is in my garden"
> or "it is not the case that chlorine is in my garden" from being true
> in a circumstance in which some (but not all) Obama/chlorine is in my
> garden. I suppose the principle is that referents are treated as atoms
> rather than as complexes some bits of which do broda and other bits of
> which don't necessarily broda; but I'm really only thinking aloud in
> saying this.
I don't understand how this fits with your
whole-sort-of-general-mish-mash metaphysics and {lo} handling.
> > "It was the exuberant Obama who spoke today rather than the dour Obama
> > we're used to"? That kind of thing? The hacky solution still seems
> > reasonable.
> That kind of thing, yes. A hacky solution may or may not be
> reasonable, but legitimate justifications for seeking a hackysolution
> do not include the alleged absence of this phenomenon from natural
> language.
All I was doubting was that english *forces* us to use stages to
understand it, i.e. when developing a model-theoretic formal semantics
for it.
> >> You may be likely to be misunderstood, but that's because of
> >> philosophical differences between you, not linguistic differences.
> >> You don't have to agree on whether{mi zukte da poi do zukte} couod be
> >> true.
> > If that counts as philosophy, then it seems we do have to make
> > philosophical pronouncements if we want to well-specify lojban.
> It would be interesting and instructive if that turned out to be the
> case, though it's not yet apparent to me that it is. I think rather
> than talking about "well-specifiedness", we should distinguish (A) the
> rules mapping between phonological form and logical form from (B) the
> rules mapping between logical form and the universe. For everybody who
> wants a logical language, it is important that (A) be well-specified.
What use is a logical form if we don't know what it means? How is that
an improvement on unprocessed lojban?
> But I'm not sure there's anything remotely approaching a consensus on
> whether (B) must be well-specified. I myself incline to the view that
> it needs to be specified with a certain looseness, partly for
> practical reasons -- because while (A) can be specified to perfection,
> (B) can never be finished
I'm not so sure I agree that it can't, actually. Of course it depends
exactly what you mean by (B). I take it to mean a notion of satisfaction
for the logic in question, i.e. a way of telling whether a sentence is
true in a given model, where we allow some things to be handled by
pragmatics (a separate module wherein all ambiguity lies, and which is
less likely to be susceptible to clear and definite description). So
e.g. (B) is complete for the formal logics logicians talk about (with no
pragmatic component).
Completing (B) for lojban is clearly a non-trivial task, and we may well
find that we're better off settling for a partial solution (e.g. which
assumes boolean truth values, and so doesn't really handle {jei}) than
a complete one... but I don't see any obvious reason for considering it
impossible.
> -- and partly because speakers with different views on the nature of
> the universe ought still to be able to speak the same language.
That seems reasonable only if "nature of the universe" is understood in
its everyday sense, rather than the technical logical sense we're
talking about here.
Or at least, which I'm talking about here. I'm not entirely sure that
we're not talking at cross-purposes. In the case of Chierchia's kinds
(which I'm mostly identifying with xorxes', modulo his domain-switching
addition), it's fairly clear: he has us adding to the universe for each
(sufficiently regular) expressible unary predicate a new entity, in
a new sort over which some but not all quantifications range. Whether we
have such is clearly a question of the setup of the logic. I don't know
whether the same applies to your types.
> >>> My question is whether you perceive a "jump" between individual lions
> >>> and the kind 'lions' of a different kind from that between the kinds
> >>> 'fierce lions' and 'lions'. I don't think it's actually a precise
> >>> question about the structure of the partial order... it's rather that
> >>> I'd split "subtype" into two relations - "instance of" and "subclass
> >>> of".
> >>
> >> I understand your questions. The answer is a very definite No. There
> >> are only types, related by the Subtype relation; and there are no
> >> instances.
> >
> > Then I don't think I know at all what your "types" are. They seem to be
> > different from xorxes' kinds, which seem (or at least so my
> > uncontradicted impression was) to correspond to properties of
> > individuals at the level below.
>
> Hmm. I don't consciously find myself disagreeing with xorxes. Are
> there further diagnostic questions you could pose in order to
> discriminate between my view and the one you attribute to xorxes?
Well, I understood him as agreeing with Chierchia that for certain
predicates, like x1 of "is in", kind predication always ends up being
equivalent (switching domains as necessary) existential quantification
over instances of the kind - as in the "lions are in my garden" example.
If your kinds don't have instances, presumably you don't agree with
this!
> >> I think it would be good to have other gadri based on a model in which
> >> there are no types, only instances.
> >
> > And not worry about interactions?
>
> Between what? Different types of gadri? Probably yes -- don't worry.
> Or at least, it's interesting to discuss, but doesn't have to be
> addressed as part of the basic specification of Lojban.
But then it seems more that you're talking about forking the language
into multiple sublanguages, with the gadri you use as the only
indication as to which language you're using.
> >>>> The objections to that are that it is metaphysically biased,
> >>>
> >>> Why is that a problem?
> >>
> >> Avoidance of metaphysical bias was one of Lojban's aims. A fairly
> >> obvious and sensible one, since the language should not tell the
> >> speaker how the universe is, but rather should allow the speaker to
> >> describe how the speaker thinks the universe is.
> >
> > This seems to be in direct competition with an aim of lojban with which
> > I'm more familiar, namely that it be well-specified. Having a thorough
> > model-theoretic formal semantics seems to me an important part of
> > satisfying that aim - and it would involve specifying a metaphysics (by
> > your definition of metaphysics).
>
> See my comments above about the two types of specification. I think
> human languages are thoroughly specified for type (A) (even tho the
> rules allow ambiguity)
They are? Potentially or actually? With what "logical form"?
> but not for type (B). So I understand the goal of a logical language
> as to be like a human language, but for the type (A) rules to exclude
> ambiguity.
>
> Nevertheless, I can understand how you might want not only a fully
> specified language, but also a fully specified model of the universe,
> because it promises perfect communication not only at the level of
> logical form but also at the level of semantics.
>
> But the only Lojbanists I've ever seen ask for fully specified
> semantics are John Clifford and you, so I'd say that your
> understanding of well-specifiedness is not the normal one.
Well, plenty of people want the language to be better-specified than it
is. I'm not sure how to view this push to specification if not as a push
towards (B).
Martin
Attachment:
pgpZwZDQnBSRG.pgp
Description: PGP signature