Mike S., On 08/09/2012 03:29:
> On Fri, Sep 7, 2012 at 9:15 PM, And Rosta <and.rosta@hidden.email
> <mailto:and.rosta@hidden.email>> wrote: But we should remember the
> distinction between (I) rules that map a sentence's phonological
> form to its logical form, and (II) rules that map logical forms to
> (stuff in) (possible) worlds. We do know from past discussions on
> Lojban list that some folk, including John Clifford, want of a
> logical language that it do not only (I) but also (II), whereas
> others, including me and Jorge, want only that it do (I), considering
> (II) to be essentially extralinguistic and unachievable and
> potentially overly restrictive for users.
>
> Therefore, folk who are after (II) are really after a different
> project than Xorban is. However, to make a Xorban-B they could still
> use the grammar of Xorban and simply add on type-(II) rules.
>
> I find this (I) versus (II) distinction very puzzling. As far as I
> can see, possible worlds have an essential part to play in the
> logical form underlying human language semantics, and they must be
> incorporated in Xorban if it is to be a human language. Every human
> being without the slightest formal linguistic training easily grasps
> lines like "I coulda been a contender", and that line clearly
> invokes a possible world. What on earth is "essentially
> extralinguistic and unachievable" about that?
I expressed myself badly. (II) involves formally modelling the extralinguistic world, including possible worlds, and defining correspondence rules between logical forms and, let's say, models of states of affairs. To take a concrete example, proponents of (II) want a rigid definition of cathood to be part of the language (hence questions about what exactly is la mlta). In contrast, for proponents of (I)-only, mlt in logical form is a pointer to the mlt entry in the encyclopedia of everything, but the contents of that entry are external to (I)-only language.
I don't particularly care for a rigid definition of cathood. but I would like a consistent set of rules about la mlta xkra. Talking
about pointers to encyclopedia entries presupposes a whole lot more about what is outside the language than I have ever pushed for. I have not asked for the conditions on a thing being a mlt (what I take to be what you mean by "the contents of that entry") nor what things are actually in the extension of mlt (another possible meaning), except, if it occurs, la mlta. Now, my question is really, is what makes la mlta xkra true in fact a mlt in whatever is the normal sense of that (about the details of which I do not care)?
Mike S., On 08/09/2012 08:06:
> On Fri, Sep 7, 2012 at 11:29 PM, John E. Clifford 1. No you can't,
> as witness the various interplays among the quantifiers.
>
> What exactly are you asking for? If you want FOL, it's there;
> Xorban "s-" and "r-" are in full force.
I agree. If s- and r- are inadequate, we're open to having it explained to us why they're inadequate and we're open to considering solutions to whatever problems there are. With, in my case, the caveat that I want to deliver a (I)-only language -- call it a "worldblind" language in contrast to an "encyclopedical" language in which the rules of the language include the contents of encyclopedia entries.
I don't have any problems with r and s until they get mixed with l. Then, since what they say together seems to make no sense, I have to wonder if it is only l that is odd or whether r and s are, too. This is all a question about this world (without any regard to what this world may
be. But it would be as much a problem in any other world.
John E. Clifford, On 08/09/2012 05:00:
> So la mlta xrka ascribes blackness to the object whose sole
> *defining* characteristic is catness. And what's this object? Is it a
> member of the extension of mlt in the relevant universe? Then it may
> be incidentally black, but that would hardly justify ra mlta xrka. Is
> it a something else in extension mlt by courtesy and also in all
> supersets of mlt? Then, of course, if it is in extension xrk , then
> the universal is justified and the particular, too, since extensions
> are L-sets. But, of course, this is rarely the case: cats aren't
> usually uniform for color. What's left? Catness? There is (or, I
> assume, will be) an _expression_ for that which says so clearly, which
> this does not. The mystery continues.
I want to create a loglang that expressly doesn't answer questions of this sort, at least not in any formal way free of contradictions, squintings and so forth.
So, then, how do we tell if la mlta xkra is true? We squint, apparently, but at what? I don't even know where to look. I would have thought that cats had something to do with it, but apparently not.
So let's collaborate where we can, and where our goals are too incompatible, we can make different loglangs (which may still have a lot in common).
I suppose my fear is that you will peddle this whateverthefuckitis as a logical language, where as it is not at all logical (in the logicians' sense) and, at a guess, not really a language, since it cannot convey basic messages with any assurance to the speaker that he said what he meant or to the listeners as to either what the speaker said or meant.
John E Clifford, On 08/09/2012 14:49:
> *From:* Mike S. <
maikxlx@gmail.com> What exactly are you asking for?
> If you want FOL, it's there; Xorban "s-" and "r-" are in full force.
> If you want to speak in human instead, then why not freely use "l-"?
> If you don't like my proposed formalization, then feel free to give
> us yours. I want to hear your ideas. If you've been thinking about
> this for thirty years, then surely you have some constructive
> observations.
>
> I've given you one plausible interpretation, that l is an ever
> leftmost quantifier that picks out a bunch of items and sticks with
> thm throughout its scope. It does the trick and is a familiar feature
> of natural languages. There are problems with it, xorxes insists, and
> so it may need some work, but it is not obviously crippled from the
> getgo.
I don't see problems with it. Crucially, if Y is true of something that is that bunch, then Y is true of everything that is that bunch -- given that it's a single bunch.
This seems to me a rather different point than what has been said formally. The formalism has been la Ra Pa <=> re RePe and la Ra Pa <=> se Re Pe. But what you are saying seems to be la Ra Pa <=> re e=a Pe and similarly for s. That -- given context leaping -- is clearly unobjectionable, but a very different claim.
There's a slightly different quantifier that I'd been meaning to mention, but have kept on forgetting. It's an existential quantifer with scope over the whole discourse:
"Once upon a time, there was a poor woodcutter. He lived in a hut with his dutiful daughter."
In Livagian I treat it as a separate quantifier in its own right.
Yeah, this is just the l that I take this the present l to be (as it is in Lojban, rightly understood), the Montagovian constant.
> Well, being a logical language poses some restrictions (just what
> depends upon what we mean by "logical"). One that I think is pretty
> generally accepted is that the operators are clearly defined and
> consistent.
We shouldn't waste time *arguing* over what we mean by logical. It is enough that we each, x, simply define what characteristics x requires of a loglang if it is to satisfy x. Where our requirements are the same or at least compatible, we can collaborate. Where they're not, we can create alternative versions.
Well, I'll let you start, since I tend to think that I mean what everyone means once they focus away from the popular uses of "logical" to mean reasonable or some such thing. And you *seem* to have something more restrictive in mind.
John E. Clifford, On 08/09/2012 04:39:
> [And Rosta wrote:
>> But we should remember the distinction between (I) rules that map
>> a sentence's phonological form to its logical form,]
>
> Well, so far you seem to be 1) screwing up on part I and 2) the ones
> appealing to metalanguage and even pragmatic items to save the crap
> (well, make it look like at least useful manure). Yes, a type II
> program needs a successful type I, which (were I working on a type
> II program or even thinking of it as a separate program) is why I
> object to messing stuff up at the start.
I don't want to close down fruitful debate, but on this specific issue that bothers you so, I've been seeing eye to eye with Jorge for twenty years(!), and for most of that time you've been objecting. I think this has been demonstrated to be a case where we aren't going to reach consensus, so we need to look for alternative solutions. Here are some suggested solutions:
A. If you can live with lV and are happy with just sV and rV, just let the matter rest.
B. If you can live with lV, but want another new quantifier, have a go persuading the rest of us that it is needed and not already covered by lV.
C. If you can't live with lV, define a variant dialect of Xorban in which lV is abolished or given a different definition.
Sorry, I will continue to insist that you actually provide the purported lV, as you have refused to do for twenty-five years or so. In the meantime, I will work with the lV you have actually provided (which is not the one you are talking about) and clean up some remaining messy edges. The fact is that, with a little care, there are practically no problems left using good old extensional {lo} -- and a lot of possible worlds, to be sure. So, I do not quite see why you insist on pushing for this chimera, rather than just fixi9ng the few remaining problem with the real thing you have.
--And.