From pycyn@aol.com Mon Oct 01 09:11:52 2001 Return-Path: X-Sender: Pycyn@aol.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_4_1); 1 Oct 2001 16:11:52 -0000 Received: (qmail 60717 invoked from network); 1 Oct 2001 16:11:52 -0000 Received: from unknown (10.1.10.142) by m8.onelist.org with QMQP; 1 Oct 2001 16:11:52 -0000 Received: from unknown (HELO imo-r05.mx.aol.com) (152.163.225.101) by mta3 with SMTP; 1 Oct 2001 16:11:51 -0000 Received: from Pycyn@aol.com by imo-r05.mx.aol.com (mail_out_v31_r1.7.) id r.33.1bc17509 (3927) for ; Mon, 1 Oct 2001 12:11:40 -0400 (EDT) Message-ID: <33.1bc17509.28e9efbc@aol.com> Date: Mon, 1 Oct 2001 12:11:40 EDT Subject: Re: [lojban] Re: noxemol ce'u To: lojban@yahoogroups.com MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="part1_33.1bc17509.28e9efbc_boundary" X-Mailer: AOL 6.0 for Windows US sub 10535 From: pycyn@aol.com --part1_33.1bc17509.28e9efbc_boundary Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable In a message dated 9/29/2001 9:33:33 PM Central Daylight Time,=20 a.rosta@dtn.ntl.com writes: > You > could then press on with your indagations into ce'u, while Jorge and > I strive to make sure you don't demolish more than you add. Nice word; thanks. That is exactly what I was doing and still am, despite=20 many helpful suggestions about what I might be doing, but am not. Since I= =20 have added very little and tossed nothing out, I feel fairly secure from an= y=20 intervention from &&X.=20=20 What do we know for sure about {ce'u}: the accepted wisdom. First of all, = it=20 occurs in {ka} (and perhaps other NU) to mark the open places in a predicat= e=20 expression, a property. This is in the Refgram right after the explanation= =20 of {ke'a} as the specialized anaphora in relative clauses (anaphorizing the= =20 head). {ke'a} is anaphoric and is explicitly said to occur only in relativ= e=20 clauses of the sort described. {ce'u} is obviously not anaphoric and is n= ot=20 explicitly restricted in any way, although only examples with {ka} are give= n.=20 At a later point, in discussing {ka}, we get a thoroughly confused=20 presentation in which first properties are presented without {ce'u} -- and= =20 indeed without any place for a mark of what they are properties of, then wi= th=20 one (obligatory?) {ce'u} and then with the possibility of several {ce'u}. = =20 {ce'u} is not mentioned anywhere else in Refgram. Thus we have one paradig= m=20 case of {ce'u} but no indication of the scope of the paradigm: {ka} only,=20 other NU (this is at least mentioned as a possiblity and uses with {du'u}, = at=20 least, have occurred in And's canon), all NU, only NU,..... The other thing we "know" about {ce'u} is that it is a "lambda variable". = =20 This is nowhere mentioned in Refgram, even in the short reasonably accurate= =20 discussion of the lambda calculus in MEX -- which is about how to speak it = in=20 Lojban, not about using it at all. Apparently, Cowan used the notion of a= =20 lambda variable in the proposal introducing {ce'u} and has made use of it=20 since on occasion, though I cannot find any messages in which it is employe= d=20 in any significant way. Clearly the lambda notion does not ahve the same=20 authoritative force as citations from the Refgram.=20=20 So, the argument for {le mamta be ce'u} and for it not being a main bridi u= se=20 (aside from direct observation) must look elsewhere. The Refgram evidence = is=20 inconclusive: the fact that it is next to {ke'a} and that only one kind of= =20 example is given points to restricted applications, but the fact that, unli= ke=20 {ke'a} and a few other clear cases, there is no explicit restriction on the= =20 {ce'u}'s use tends to neutralize that claim.=20=20 My actual argument for {le mamta be ce'u} come from the way I came to it,=20 namely expanding a generalization that had emerged from dealing with indire= ct=20 questions: a rule that worked there looked incomplete, so I tested varous=20 ways of extending it, several of which -- including {le mamta be ce'u} --=20 worke dout jsut as they should. so, I suggested that they be taken in, On= ly=20 {le mamta be ce'u} caused problems immediately. I think some of the others= =20 are more suspect, but I'll not mention which, since I like them all and don= 't=20 feel like fighting more than one battle at once. But the problem with=20 arguing from this general theory to this aprticular case is that the genera= l=20 theory of indirect questions (all questions, actually) is itself in=20 contention, between a clear and sucessful set-of-answers view (from logic,= =20 loosely) and a muddled and failing extension-claims view (from God knows=20 where). So, to make the case for {le mamta be ce'u}, I first need to turn = to=20 giving extension-claims indirect questions the appropriate burial (as if=20 anyhting ever died in a Lojban discussion -- there are more=20 ghouls/vampires/ghosts in these messages than in all of literature). "function" in this context, esepcially since all of the examples I=20 > have given have not counted as cases, apparently (even though they=20 > are generally your examples with minor modifications),=20 I mean whatever you mean -- whatever it is that you think we have no agreed way of rendering in Lojban.> Clearly you don't mean that, since I have given you a dozen cases by now,=20 each of which you have rejected the Lojbanic version of an propsoed soemthi= ng=20 else which does not fit the case. I have tried to figure out what you take= a=20 function to be from the proposed answers you give, but they are not mutuall= y=20 consistent, so I am at a dead loss. <[1]> "Bobby, recite your times tables"?=20 [2]> "Sum is symmetric but power is not"?=A0=20 [3]>"Hyperpower is hard to define, since the 0 case is undetermined."?=A0=20 [4]>"They differ in their mothers"?=A0=20 [5]> "They are interchangeable in their friends"?=20 So if we could find ways to lojban these, we'd satisfy you?> Not necessarily, since these are only examples. But, since you can't do=20 them, the point is moot. Aside from not knowing what {ce'u - timestable} is -- "is a times table" --= I=20 suppose, so the extensuion of the property (I assume there is only one, so= =20 {tu'o} is not misleading, as if often is in these cases) would be what? 3x= ,=20 4x, 5x? Clearly not what is wanted. What is wanted is {li ce'u pi'i ceu}= =20 for a suitable range of both=20 {ce'u}. In short what is wanted in 1-3 is something that gives a number, n= ot=20 a sentence as all {du'u} do eventually. As for 4 and 5, of course functions are not necessarily involved, any more= =20 than properties and indirect questions are in other versions of this same=20 sort of thing. They all come down to case of identity/equivalence or their= =20 denial and so involve only fully specified sentences . But if we are to=20 allow one kind of abbreviation (actually it often ends up longer than the=20 original), why not all the others that come under the same rules? The need is not very pressing, any more than the need to fiddle with {goi}= =20 is. And, since nothing is thrown out in adding these, there is no place to= =20 look for the baby except safe in the crib it was in all the time. --part1_33.1bc17509.28e9efbc_boundary Content-Type: text/html; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable In a message dated 9/29/2001 9:33:33 PM Central Daylight Time, a.rosta@dt= n.ntl.com writes:


You
could then press on with your indagations into ce'u, while Jorge and
I strive to make sure you don't demolish more than you add.


Nice word; thanks. That is exactly what I was doing and st= ill am, despite many helpful suggestions about what I might be= doing, but am not.  Since I have added very little and tossed nothing= out, I feel fairly secure from any intervention from &&X.  

What do we know for sure about {ce'u}: the accepted wisdom.  First= of all, it occurs in {ka} (and perhaps other NU) to mark the open places i= n a predicate expression, a property.  This is in the Refgram right af= ter the explanation of {ke'a} as the specialized anaphora in relative claus= es (anaphorizing the head).  {ke'a} is anaphoric and is explicitly sai= d to occur only in relative clauses of the sort described.  {ce'u} is =  obviously not anaphoric and is not explicitly restricted in any way, = although only examples with {ka} are given.  At a later point, in disc= ussing {ka}, we get a thoroughly confused presentation in which first prope= rties are presented without {ce'u} -- and indeed without any place for a ma= rk of what they are properties of, then with one (obligatory?) {ce'u} and t= hen with the possibility of several {ce'u}.  {ce'u} is not mentioned a= nywhere else in Refgram.  Thus we have one paradigm case of {ce'u} but= no indication of the scope of the paradigm: {ka} only, other NU (this is a= t least mentioned as a possiblity and uses with {du'u}, at least,  hav= e occurred in And's canon), all NU, only NU,.....
The other thing we "know" about {ce'u} is that it is a "lambda variable= ".  This is nowhere mentioned in Refgram, even in the short reasonably= accurate discussion of the lambda calculus in MEX -- which is about how to= speak it in Lojban, not about using it at all.  Apparently, Cowan use= d the notion of a lambda variable in the proposal introducing {ce'u} and ha= s made use of it since on occasion, though I cannot find any messages in wh= ich it is employed in any significant way.  Clearly the lambda notion = does not ahve the same authoritative force as citations from the Refgram. &= nbsp;
So, the argument for {le mamta be ce'u} and for it not being a main bri= di use (aside from direct observation) must look elsewhere.  The Refgr= am evidence is inconclusive: the fact that it is next to {ke'a} and that on= ly one kind of example is given points to restricted applications, but the = fact that, unlike {ke'a} and a few other clear cases, there is no explicit = restriction on the {ce'u}'s use tends to neutralize that claim.  
My actual argument for {le mamta be ce'u} come from the way I came to i= t, namely expanding a generalization that had emerged from dealing with ind= irect questions: a rule that worked there looked incomplete, so I tested va= rous ways of extending it, several of which -- including {le mamta be ce'u}= -- worke dout jsut as they should.  so, I suggested that they be take= n in,  Only {le mamta be ce'u} caused problems immediately.  I th= ink some of the others are more suspect, but I'll not mention which, since = I like them all and don't feel like fighting more than one battle at once. =  But the problem with arguing from this general theory to this aprticu= lar case is that the general theory of indirect questions (all questions, a= ctually) is itself in contention, between a clear and sucessful set-of-answ= ers view (from logic, loosely) and a muddled and failing extension-claims v= iew (from God knows where).  So, to make the case for {le mamta be ce'= u}, I first need to turn to giving extension-claims indirect questions the = appropriate burial (as if anyhting ever died in a Lojban discussion -- ther= e are more ghouls/vampires/ghosts in these messages than in all of literatu= re).

<Since I am not clear what you mean by=20
> "function" in this context, esepcially since all of the examples I= =20
> have given have not counted as cases, apparently (even though they= =20
> are generally your examples with minor modifications),=20

I mean whatever you mean -- whatever it is that you think we have no
agreed way of rendering in Lojban.>

Clearly you don't mean that, since I have given you a dozen cases by no= w, each of which you have rejected the Lojbanic version of an propsoed soem= thing else which does not fit the case.  I have tried to figure out wh= at you take a function to be from the proposed answers you give, but they a= re not mutually consistent, so I am at a dead loss.

<[1]> "Bobby, recite your times tables"?=20
[2]> "Sum is symmetric but power is not"?=A0=20
[3]>"Hyperpower is hard to define, since the 0 case is undetermined.= "?=A0=20
[4]>"They differ in their mothers"?=A0=20
[5]> "They are interchangeable in their friends"?=20

So if we could find ways to lojban these, we'd satisfy you?>

Not necessarily, since these are only examples.  But, since you ca= n't do them, the point is moot.

<For [1], I'd say the recitee is the extension of tu'odu'u ce'u -tim= estable.

For [2] and [3] I'd render Sum, Power and Hyperpower by tu'odu'u ce'u -= sum
ce'u, etc.

For [4] and [5] I don't yet understand how functions are necessarily
involved; certainly you and I have agreed that certain functionless
logical renditions are in themselves adequate.>

Aside from not knowing what {ce'u - timestable} is -- "is a times table= " -- I suppose, so the extensuion of the property (I assume there is only o= ne, so {tu'o} is not misleading, as if often is in these cases) would be wh= at?  3x, 4x, 5x?  Clearly not what is wanted.  What is wante= d is {li ce'u pi'i ceu} for a suitable range of both=20
{ce'u}.  In short what is wanted in 1-3 is something that gives a = number, not a sentence as all {du'u} do eventually.
As for 4 and 5, of course functions are not necessarily involved, any m= ore than properties and indirect questions are in other versions of this sa= me sort of thing.  They all come down to case of identity/equivalence = or their denial and so involve only fully specified sentences .  But i= f we are to allow one kind of abbreviation (actually it often ends up longe= r than the original), why not all the others that come under the same rules= ?

<Well give us an indication of how pressing the need is, so we can j= udge
how much baby we can allow to be thrown out with the bathwater in the
search for a way to talk about them.>

The need is not very pressing, any more than the need to fiddle with {g= oi} is.  And, since nothing is thrown out in adding these, there is no= place to look for the baby except safe in the crib it was in all the time.





--part1_33.1bc17509.28e9efbc_boundary--