From pycyn@aol.com Sat Sep 22 14:42:27 2001 Return-Path: X-Sender: Pycyn@aol.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_3_2_2); 22 Sep 2001 21:42:27 -0000 Received: (qmail 26476 invoked from network); 22 Sep 2001 21:42:27 -0000 Received: from unknown (10.1.10.142) by 10.1.1.222 with QMQP; 22 Sep 2001 21:42:27 -0000 Received: from unknown (HELO imo-d04.mx.aol.com) (205.188.157.36) by mta3 with SMTP; 22 Sep 2001 21:42:22 -0000 Received: from Pycyn@aol.com by imo-d04.mx.aol.com (mail_out_v31_r1.7.) id r.116.4fe5a59 (2612) for ; Sat, 22 Sep 2001 17:42:18 -0400 (EDT) Message-ID: <116.4fe5a59.28de5fb9@aol.com> Date: Sat, 22 Sep 2001 17:42:17 EDT Subject: Re: [lojban] Set of answers encore To: lojban@yahoogroups.com MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="part1_116.4fe5a59.28de5fb9_boundary" X-Mailer: AOL 6.0 for Windows US sub 10535 From: pycyn@aol.com --part1_116.4fe5a59.28de5fb9_boundary Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable In a message dated 9/22/2001 1:25:58 PM Central Daylight Time,=20 a.rosta@dtn.ntl.com writes: > da zo'u la djan jinvi/krici tu'odu'u da -extension tu'odu'u > ce'u pa moi merko -president >=20 > -- i.e. as not substantively different from the treatment of djuno. >=20 Not so: the corresponding analysis would be=20 {da de poi du'u da extension of ledu'u ce'u pamoi merko president zo'u la=20 djan jinvi de} However, in working the details of this out, I see that it does work as wel= l=20 as well in this case as in the {djuno} one. I keep forgetting that when yo= u=20 say "extension" you mean the propositions that such-and-such is the=20 extension, not the set itself (I think this ahs been a problem for a while = --=20 back to whether the things that go in for {makau} are answers). I'll try t= o=20 remember this is the extension-claim theory, not the extension theory, whic= h=20 is very different. I am glad to hear that you now reject the rivalry which you fostered and th= at=20 you see the two as different formulations of the same general answer. I=20 agree that (after finally getting the details of the formal, but very messy= ,=20 presentation down) this looks plausible and hope that together we can get a= =20 reasonable set of rules laid out soon. I now begin to see how to convert extension-claim analysis to set-of-answer= =20 analysis fairly simply. Well, I am not sure what you mean by "standard sorts" -- both reduction and= =20 higher order lambdas are non-standard in a pretty clear sense (most logic=20 books don't have them) but everyday in the appropriate fields within logic= =20 (and used a variety of different ways, then). Reduction is pretty clearly= =20 the Lojban way, but higher levels might be clearer. If we need them at all= ;=20=20 I kinda like the treatment I have given to the {frica} and {dunli} cases,=20 which get rid of even the original selbri. I see that you do too, though=20 along a different route: Note, though, that, despite what you said at the beginning, you end up with= =20 "they" being {ce'u} and "who" and "what" {makau}. Now I know you want=20 {makau} to be a sort of {ceu}, but from your ussage it appears to be simply= =20 one with longer scope, able to step out of its immediate matrix. And very= =20 useful it is too. =A0=A0 answers -- illocutionary answers are a red herring.=20 >=20 > A nice distinction.=A0 I think that we will need to allow illocutionary=20 > answers, so as to encompass cases like "believes" and maybe "same"=20 > and "different."=A0 We'll see what happens as more cases come under the=20 > microscope.=20 I'll need to see concrete examples of where illocutionary answers might be needed, because I'm finding of any myself. Let me try. You want to be able to sat "John has an opinion about who went", such that this covers a case where John's opinion is that the set of goers includes da poi ninmu. I would handle this case=20 not by using indirect questions that allow for i-answers, but rather by, say, "la djan jinvi tu'odu'u lo'i klama mo kau", or =A0 da zo'u la djan jinvi tu'odu'u de ge cmima tu'o -extension tu'odu'u=20 =A0 ce'u klama gi da ckaji de> I think what I had in mind (that "extension" rather than "extension-claim"= =20 again) was the case where John opined that someone went who in fact did not= .=20=20 That case aside now, what is the one you have set up and why might I think = it=20 a problem for either of us? I don't see aproblems either way for this=20 translation of "John has an opion of what goes are like" or some such=20 phrasing.=20=20 Well, if we can intertranslatem perhaps we will come to agree that it does= =20 (cf. Turing and abacus). As for that being the way to say "the set of=20 answers,"=20 at least it's got set and proposition right, the rest is open to some=20 criticism. But then, we don't have "is the extension of du'u makau broda"= =20 down yet either, so we are about even.=20 I think that rules for generating presuppositions and consequences of qkau= =20 would be quite enough, even if we never found a way to say the same thing= =20 without using qkau. It does look like total elimination -- except for=20 rhetorical effect and convenience -- is possible, however. --part1_116.4fe5a59.28de5fb9_boundary Content-Type: text/html; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable In a message dated 9/22/2001 1:25:58 PM Central Daylight Time, a.rosta@dt= n.ntl.com writes:


 da zo'u la djan ji= nvi/krici tu'odu'u da -extension tu'odu'u
 ce'u pa moi merko -president

-- i.e. as not substantively different from the treatment of djuno.

Not so:  the corresponding analysis would be=20
{da de poi du'u da extension of ledu'u ce'u pamoi merko president zo'u = la djan jinvi de}
However, in working the details of this out, I see that it does work as= well as well in this case as in the {djuno} one.  I keep forgetting t= hat when you say "extension" you mean  the propositions that such-and-= such is the extension, not the set itself (I think this ahs been a problem = for a while -- back to whether the things that go in for {makau} are answer= s).  I'll try to remember this is the extension-claim theory, not the = extension theory, which is very different.

<I yet again reject the idea that Set-of-Answers and Extension can b= e
treated as competitors.

I think I can translate any Set-of-Answers analysis into an Extension
one, so whatever Set-of-Answers works for, so does Extension.>

I am glad to hear that you now reject the rivalry which you fostered an= d that you see the two as different formulations of the same general answer= .  I agree that (after finally getting the details of the formal, but = very messy, presentation down) this looks plausible and hope that together = we can get a reasonable set of rules laid out soon.
I now begin to see how to convert extension-claim analysis to set-of-an= swer analysis fairly simply.

<The solution does require [reducing all qkau to {makau} and/or high= er level lambdas] if that by definition is a criterion of
what counts as a solution -- for me to understand qkau I need to feduce
it to a logical formula that contains logical elements only of standard
sorts.>

Well, I am not sure what you mean by "standard sorts" -- both reduction= and higher order lambdas are non-standard in a pretty clear sense (most lo= gic books don't have them) but everyday in the appropriate fields within lo= gic (and used a variety of different ways, then).  Reduction is pretty= clearly the Lojban way, but higher levels might be clearer.  If we ne= ed them at all;  I kinda like the treatment I have given to the {frica= } and {dunli} cases, which get rid of even the original selbri.  I see= that you do too, though along a different route:

<Ah.=A0 What is it then that it gives you a two-ce'u problem?

"differ in who they love". Standard lojban renders "they" as ce'u. I
think "who" should be rendered as ce'u. Hence apparent two-ce'u
problem. Solution: "they" is not ce'u. The solution was spelt out in an
earlier message, which I repeat here, for your convenience:

#I think I am now able to offer a halfway decent analysis:
#
#no da ro de poi ke'a cmima la dybiyb ce la tcelsik [-- or cmima of wha= tever
#class of differers --] zo'u
#da -extension-of tu'odu'u ce'u mamta de
#
#=3D D frica C tu'odu'u ma kau mamta ce'u
#=3D Dubya and Chelsea differ in who their mothers are
#
#Now that can be done more simply as:
#
#no da ro de poi ke'a cmima la dybiyb ce la tcelsik zo'u da mamta de
#
#or indeed
#
#no da mamta ge la dybiyb gi la tcelsik
#
#But the longerwinded method comes into its own in cases like:
#
#=A0 X and Y differ in who gave them what
#=3D ... frica tu'odu'u ma kau dunda ma kau ce'u
#=3D ... da -extension of tu'odu'u ce'u dunda ce'u de
#
#Admittedly, this "halfway decent analysis" does not use {frica}, but t= here
#was no guarantee that {frica} is logically sound, and hence no guarant= ee
#that frica could be used in a logically explicit formulation.>

Note, though, that, despite what you said at the beginning, you end up = with "they" being {ce'u} and "who" and "what" {makau}.  Now I know you= want {makau} to be a sort of {ceu}, but from your ussage it appears to be = simply one with longer scope, able to step out of its immediate matrix. &nb= sp;And very useful it is too.

<my feeling is that interrogatives and qkau involve only logical=20
>=A0=A0 answers -- illocutionary answers are a red herring.=20
>=20
> A nice distinction.=A0 I think that we will need to allow illocuti= onary=20
> answers, so as to encompass cases like "believes" and maybe "same"= =20
> and "different."=A0 We'll see what happens as more cases come unde= r the=20
> microscope.=20

I'll need to see concrete examples of where illocutionary answers might
be needed, because I'm finding of any myself.

Let me try. You want to be able to sat "John has an opinion about
who went", such that this covers a case where John's opinion is that
the set of goers includes da poi ninmu. I would handle this case=20
not by using indirect questions that allow for i-answers, but rather
by, say, "la djan jinvi tu'odu'u lo'i klama mo kau", or

=A0 da zo'u la djan jinvi tu'odu'u de ge cmima tu'o -extension tu'odu'u= =20
=A0 ce'u klama gi da ckaji de>

I think what I had in mind (that "extension" rather than "extension-cla= im" again) was the case where John opined that someone went who in fact did= not.  That case aside now, what is the one you have set up and why mi= ght I think it a problem for either of us? I don't see aproblems either way= for this translation of "John has an opion of what goes are like" or some = such phrasing.  

<I don't accept that the set-of-answers theory gets us to a satisfac= tory
analysis of qkau. But what I meant was the latter -- that I don't accep= t
that "lo'i du'u ma kau broda" is the way to say "the set of answers to= =20
the question 'ma broda'". I don't reject it outright, though; my positi= on
is that that meaning does not automatically follow from the constituent
parts, but that since I can think of no other sensible meaning it is
an unobjectionable interpretive convention.>

Well, if we can intertranslatem perhaps we will come to agree that it d= oes (cf.  Turing and abacus).  As for that being the way to say "= the set of answers,"=20
at least it's got set and proposition right, the rest is open to some c= riticism.  But then, we don't have "is the extension of du'u makau bro= da" down yet either, so we are
about even.=20

<I think I've said in other messages that I don't think qkau can be
explained unless we can eliminate it (from underlying logical forms).&g= t;
I think that rules for generating presuppositions and consequences of q= kau would be quite enough, even if we never found  a way to say the sa= me thing without using qkau.  It does look like total elimination -- e= xcept for rhetorical effect and convenience -- is possible, however.
= --part1_116.4fe5a59.28de5fb9_boundary--