From pycyn@aol.com Sun Jul 29 11:56:34 2001 Return-Path: X-Sender: Pycyn@aol.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_2_0); 29 Jul 2001 18:56:34 -0000 Received: (qmail 85469 invoked from network); 29 Jul 2001 18:56:32 -0000 Received: from unknown (10.1.10.27) by l10.egroups.com with QMQP; 29 Jul 2001 18:56:32 -0000 Received: from unknown (HELO imo-r04.mx.aol.com) (152.163.225.100) by mta2 with SMTP; 29 Jul 2001 18:56:31 -0000 Received: from Pycyn@aol.com by imo-r04.mx.aol.com (mail_out_v31.9.) id r.d6.9f0ecb9 (18709) for ; Sun, 29 Jul 2001 14:56:26 -0400 (EDT) Message-ID: Date: Sun, 29 Jul 2001 14:56:26 EDT Subject: Re: [lojban] Tidying notes on {goi} To: lojban@yahoogroups.com MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="part1_d6.9f0ecb9.2895b65a_boundary" X-Mailer: AOL 6.0 for Windows US sub 10531 From: pycyn@aol.com --part1_d6.9f0ecb9.2895b65a_boundary Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit In a message dated 7/28/2001 10:02:29 PM Central Daylight Time, jjllambias@hotmail.com writes: > But {da'o} clears the references of all pro-sumti, xy included. > My objection was to {da'o}, not to {da goi xy}. > I misunderstood your point, I think, and got off on a craziness of my own in addition. I could, I suppose, defend myself on a loophole (pc resorting to cauistry! How shocking!) by noting that {da'o} is only listed as clearing KOhA and GOhA and {xy} is BY. But we will want to clear the other as well, short of {ni'o} and this seems the right way to do it. So, how else can I get what I want here. The answer seem to lie in the matter of scope of quantifiers. xorxes has made one suggestion, which goes against 16:14 (pp 410-1). On the other hand, it fits in nicely with 7:13 (p. 162): "bound variable prosumti generally last until rebound," which then points to 16 for details. (That same section has a discussion of {da'o} supporting its more general cancelling role.) But 16:14 actually says that generally the scope of a bound variable carries over ijeks but not over {.i} alone -- although no one holds that that too strictly in informal style (talking here aways about bridi level -- i.e., officially prenex -- placement, not buried in abstracts or relatives [! is this right? another part of the problem to look at] ). Further, the shifty ones that started this whole problem have absolutely shortest scope for their shifted meaning (whence the {ge...gi...} in my example). Thus in {ci da darxi reda leda canpa}, the shovel(s) pertain(s) to the three, not the two, for the scope of the reda is just its occurrence, not anything around it: {ci da re de zo'u da darxi de leda canpa} (except that the reda probably can't be fronted as things are now set up). I screwed up on {da'o} even if I were right about what it cleared, since I in fact just wanted to clear the variables used in the general claims being brought into the proof from outside (we need to consider the effects of {da'i} on quantifier scope -- it clearly does not go from in to out and only sometimes from out to in) so that they would not affect the internal vartiables. But the {.i} rule would do that and generally if we want to go beyond that scope with variable, we are actually working with some kind of instance, wherther introduced by {goi} or not, so that we can do with those short scope rules. This isn't by the Book, of course, and parts of it are shakey altogether. {xy} is an autonomous term, however it came into the discourse, so it ought to be quantifiable in any way -- why would it be the only kind of term not? The recycled {da} with the same restriction would be handy sometimes, but almost as suspect as the subquantifier version (which may, come to think of it, having been aiming at something like this after all). And we have a variety of other -- an better -- ways to do the subquantifier trick and probably the "same restriction" cases as well (sweeping the mess into the obscurity of just what {le} and {lo} mean -- for which obscurity we are at least occasionally very thankful). The problem of really good ways to deal with pluralities, especially those where the members appear first in different structural places, remains and is, as I said, much more important that the quibbles in a corner about quantifiers. xorxes has a start; can it be genralized to a fulness (not -- ever in Lojban -- a completeness)? --part1_d6.9f0ecb9.2895b65a_boundary Content-Type: text/html; charset="US-ASCII" Content-Transfer-Encoding: 7bit In a message dated 7/28/2001 10:02:29 PM Central Daylight Time,
jjllambias@hotmail.com writes:


But {da'o} clears the references of all pro-sumti, xy included.
My objection was to {da'o}, not to {da goi xy}.


I misunderstood your point, I think, and got off on a craziness of my own in
addition.  I could, I suppose, defend myself on a loophole (pc resorting to
cauistry! How shocking!) by noting that {da'o} is only listed as clearing
KOhA and GOhA and {xy} is BY.  But we will want to clear the other as well,
short of {ni'o} and this seems the right way to do it.  So, how else can I
get what I want here.  The answer seem to lie in the matter of scope of
quantifiers.  xorxes has made one suggestion, which goes against 16:14 (pp
410-1).  On the other hand, it fits in nicely with 7:13 (p. 162): "bound
variable prosumti generally last until rebound," which then points to 16 for
details. (That same section has a discussion of {da'o} supporting its more
general cancelling role.)  But 16:14 actually says that generally the scope
of a bound variable carries over ijeks but not over {.i} alone -- although no
one holds that that too strictly in informal style (talking here aways about
bridi level -- i.e., officially prenex -- placement, not buried in abstracts
or relatives [! is this right? another part of the problem to look at] ).  
Further, the shifty ones that started this whole problem have absolutely
shortest scope for their shifted meaning (whence the {ge...gi...} in my
example).  Thus in {ci da darxi reda leda canpa}, the shovel(s) pertain(s) to
the three, not the two, for the scope of the reda is just its occurrence, not
anything around it:
{ci da re de zo'u da darxi de leda canpa} (except that the reda probably
can't be fronted as things are now set up).    
I screwed up on {da'o} even if I were right about what it cleared, since I in
fact just wanted to clear the variables used in the general claims being
brought into the proof from outside (we need to consider the effects of
{da'i} on quantifier scope -- it clearly does not go from in to out and only
sometimes from out to in) so that they would not affect the internal
vartiables.  But the {.i} rule would do that and generally if we want to go
beyond that scope with variable, we are actually working with some kind of
instance, wherther introduced by {goi} or not, so that we can do with those
short scope rules.

<xy works everywhere you would use the bare
da within the scope of the original quantifier. Sticking a second
quantifier on xy would be incorrect from my point of view. Sticking
a second quantifier on da would recycle that variable as a new one,
but as a bonus you keep the poi-restriction of the original so you
don't have to repeat it.>

This isn't by the Book, of course, and parts of it are shakey altogether.  
{xy} is an autonomous term, however it came into the discourse, so it ought
to be quantifiable in any way -- why would it be the only kind of term not?  
The recycled {da} with the same restriction would be handy sometimes, but
almost as suspect as the subquantifier version (which may, come to think of
it, having been aiming at something like this after all).  And we have a
variety of other -- an better -- ways to do the subquantifier trick and
probably the "same restriction" cases as well (sweeping the mess into the
obscurity of just what {le} and {lo} mean -- for which obscurity we are at
least occasionally very thankful).  

The problem of really good ways to deal with pluralities, especially those
where the members appear first in different structural places, remains and
is, as I said, much more important that the quibbles in a corner about
quantifiers.  xorxes has a start; can it be genralized to a fulness (not --
ever in Lojban -- a completeness)?
--part1_d6.9f0ecb9.2895b65a_boundary--