[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >

[Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: [engelang] Xorban: Semantics of "l-" (and "s-" and "r-")

Mike S., On 08/09/2012 01:11:



On Fri, Sep 7, 2012 at 1:17 PM, And Rosta <and.rosta@hidden.email
<mailto:and.rosta@hidden.email>> wrote:

Jorge Llamb�as, On 07/09/2012 03:21:



On Thu, Sep 6, 2012 at 10:48 PM, Mike S.<maikxlx@gmail.com
<mailto:maikxlx%40gmail.com>> wrote:

On Thu, Sep 6, 2012 at 7:40 PM, And Rosta<and.rosta@hidden.email
<mailto:and.rosta%40gmail.com>> wrote:


Hence it behaves like a constant. There are (rightly, IMO) no
constants in Xorban, but their crucial logical property is
that when there is only one X, some X is Y iff every X is Y;
everything true of some X is true of every X.


I tried something like that yesterday and it doesn't work
formally:

Assume def: li Ri Pi<=> jo si Ri Pi ri Ri Pi<=> jo ri Ri Pi si
Ri Pi


The "everything true of some X is true of every X" is a conventional
implicature, which in logical terms means it's outside the scope of
any illocutionary.

E.g. "ca li bcdi fghi" = "je jo si bcdi fghi ri bcdi fghica sibcdi
fghi"


Unfortunately, I don't believe that "jo si bcdi fghi ri bcdi fghi"
stops encoding the useless meaning that I demonstrated that it
encodes merely by shuffling it outside of some scope, that of an
illocutionary or anything else.


But your demonstration requires precisely that the "jo" element not be presupposed.

You said, correctly, "It should follow that na li Ri Pi <=> li Ri na Pi", but then you made the improper move of replacing "li Ri Pi" by "jo si Ri Pi ri Ri Pi", which is valid only if the jo phrase is not presuppositional.

Have a misunderstood you?


I don't think "la bcda fgha" implies "sa bcda fgha" much less
"ra bcda fgha", though either of the latter does imply "la bcda
fgha".


Not in the same context. You can only move from "sa bcda fgha" or
"ra bcda fgha" to "la bcda fgha" by changing the universe of
discourse to one where bcda (and to some extent fgha as well) is
no longer dividuated, which is typically not the case in universes
where it makes sense to use s- or r-. So it's not a logical
implication that can take us from s-/r- to l- unless we are
already in l-'s territory, in which case you wouldn't be using
s-/r- in the first place.


But Mike was saying you can't go from lV to rV/sV, and I think that
lV does entail both sV and rV. "la ma xrxe rgntna, ra ma xrxe
rgntna, sa ma xrxe rgntna" (is Argentinian) are each true of the
same construal of the same world.

It's pretty clear that "la mlta xkra" doesn't entail "ra mlta xkra",
 unless the universe of discourse is ham-handedly purged of non-black
 cats, and even in that case that entailment seems purely
circumstantial.


"la mlta xkra" *does* entail "exactly one thing is mlta and ca ra mlta xkra". "That thing such that its sole defining characteristic is felinity is black" entails (no?) "everything feline is black". Just as "that thing such that its sole defining characteristic is Jorgehood is Argentinian" entails "everything that is Jorge is Argentinian".


I really dislike this pretense that the universe can vacillate so
drastically with every new sentence added to the discourse,


But I think this bears more on the usage of lV than the meaning of lV. If a speaker says "la mlta xkra", it's clear what the *sentence* means, and it's up to the speaker to decide what sort of match the speakers wants between the proposition expressed by the sentence and the proposition the speaker is trying to communicate or the state of affairs the speaker is trying to describe.


not only because that does violence to the very concept of a
universe, but also because the underlying formalized meaning of "cat"
itself depends on positing countless possible worlds full of cats,
not just the handful of cats that might happen to comprise the
current discourse topic. So constraining the model in this way not
only feels rather arbitrary, but it wreaks enormous havoc on the
underlying semantics as well.


I don't understand this (because the issues are ones I'm not accustomed to thinking about, not because you're opaque), but I also think the notion of an underlying formalized meaning of "cat" is extrinsic to core Xorban, tho it might play a part in some large loglang project based on Xorban lexicogrammar.


I should note that ' Semantics of "l-" ' was less than ideal a
choice for a thread title;  I am mainly interested in how "l-" might
be understood to work within FOL, at least for right now.  Thanks to
your ideas and Jorge's, I am fairly confident that "lV RV" can be
defined as "sV [[R]]V" and equivalently "rV [[R]]V" and get that job
done with relative ease.  I think we have a good impression of
what's inside the black box [[R]]; if not, a fully satisfying
explanation can wait until another day.


I'm glad you think there's consensus. I haven't yet got my head around "[[R]]" -- you'd said it's not an intension, right? --  but very possibly your [[mlt]] is my "that whose sole defining property is felinity".

--And.