From pycyn@aol.com Wed Oct 31 17:25:51 2001 Return-Path: X-Sender: Pycyn@aol.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-8_0_0_1); 1 Nov 2001 01:25:51 -0000 Received: (qmail 43905 invoked from network); 1 Nov 2001 01:25:51 -0000 Received: from unknown (10.1.10.27) by l9.egroups.com with QMQP; 1 Nov 2001 01:25:51 -0000 Received: from unknown (HELO imo-d01.mx.aol.com) (205.188.157.33) by mta2 with SMTP; 1 Nov 2001 01:25:50 -0000 Received: from Pycyn@aol.com by imo-d01.mx.aol.com (mail_out_v31_r1.8.) id r.13d.3c5b2a9 (4324) for ; Wed, 31 Oct 2001 20:25:33 -0500 (EST) Message-ID: <13d.3c5b2a9.2911fe8c@aol.com> Date: Wed, 31 Oct 2001 20:25:32 EST Subject: Re: [lojban] a construal of lo'e & le'e To: lojban@yahoogroups.com MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="part1_13d.3c5b2a9.2911fe8c_boundary" X-Mailer: AOL 6.0 for Windows US sub 10535 From: pycyn@aol.com X-Yahoo-Profile: kaliputra X-Yahoo-Message-Num: 11824 --part1_13d.3c5b2a9.2911fe8c_boundary Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit In a message dated 10/31/2001 10:47:28 AM Central Standard Time, arosta@uclan.ac.uk writes: > First of all, hang on: if what you're saying is addressed to me, then I had > explicitly > said to John that I don't think lo'e should be understood in terms of > typicality > or averageness -- at least not definitionally. > aHAH! Though I should have known it from past performance, I did entirely miss that what And was about was not explaining {lo'e} but rather proposing an entirely new use for {lo'e}, which he seems to think is not presently being adequately -- or appropriately -- employed. Thus, all of my comments about prototype theory not doing a good job of capturing typicality or averageness were irrelevant to his arguments(I wish And had noted that earlier; it might have shortened the time before I saw what he was about). So, "it doesn't explain 'typical'" is out; what is left? 1. Could Lojban advantageously use prototype theory? Yes. there are a variety of theoretical and even practical questions for which some version of other of prototype theory seems the most illumination approach: type-token, eme-allo, work and particular copy, to cite three that have come up over the years in Lojban. 2. Does Lojban already have the means to deal with these? Yes and no. It could easily create predicates that would cover these cases (a few are even around, I think, from previous efforts) and thus talk about the problem pretty clearly. This has not been done systematically in Lojban. And --and I take this to be And's point -- there is no consistent way to display the results of this discussion when it is relevant: no way to show -- when it is important -- whether we are talking about the highest prototype letter a or the particular smudge of ink at this point on this page -- or any of the prototypes (or versions, depending) in between. And, of course, no way to avoid making the distinction when it doesn't make a difference. 3. Is there a clear prototype theory? Wellll. As such things (philosophical or linguistic theories) go, several pretty clear ones, all similar enough to handle most of the same problems similarly, different enough to allow arguments about hard cases. 4. Does prototype theory require a new ontology? Not necessarily. Most say something vaguely like what And has (even more vaguely) presented, but some claim this is merely a convenience and then (try to?) unpack it in terms of various machines of linguistics or logic (finite state automata, transformations, Adjukeiwicz cancellation grammars, ....), whiles others really seem to believe the stuff they say (though what they say is rarely blatantly self-contradictory). 5. Does this mean that prototypes should replace typicals as the referents of {lo'e} expressions? Not obviously. On the one hand, it is not clear that prototypes are nearly as common as typicals as points we want to make or distinctions that cause problems. Further, it is not clear that gadri are the appropriate way to make prototype/version distinctions: these are relative distinctions after all, not absolute ones (like genus/species outside biological taxonomy), with a series of prototypes/versions between top and bottom (though usually transitively settled, so that this smudge on this page is a version of the letter a (as well as of my a's, script a's, italic a's ....). Of course, there is the other side; that, since typicals are not really individuals, they should not be represented by gadri at all. But much the same can be said about several versions of prototypes. In short, now that I understand what And is about, I am no closer to thinking he has made his case than I was before, I just know what the case is. --part1_13d.3c5b2a9.2911fe8c_boundary Content-Type: text/html; charset="US-ASCII" Content-Transfer-Encoding: 7bit In a message dated 10/31/2001 10:47:28 AM Central Standard Time, arosta@uclan.ac.uk writes:


First of all, hang on: if what you're saying is addressed to me, then I had explicitly
said to John that I don't think lo'e should be understood in terms of typicality
or averageness -- at least not definitionally.


aHAH!  Though I should have known it from past performance, I did entirely miss that what And was about was not explaining {lo'e} but rather proposing an entirely new use for {lo'e}, which he seems to think is not presently being adequately -- or appropriately -- employed.  Thus, all of my comments about prototype theory not doing a good job of capturing typicality or averageness were irrelevant to his arguments(I wish And had noted that earlier; it might have shortened the time before I saw what he was about).  So, "it doesn't explain 'typical'" is out; what is left?

1. Could Lojban advantageously use prototype theory?  Yes.  there are a variety of theoretical and even practical questions for which some version of other of prototype theory seems the most illumination approach: type-token, eme-allo, work and particular copy, to cite three that have come up over the years in Lojban.

2.  Does Lojban already have the means to deal with these? Yes and no.  It could easily create predicates that would cover these cases (a few are even around, I think, from previous efforts) and thus talk about the problem pretty clearly.  This has not been done systematically in Lojban.  And --and I take this to be And's point -- there is no consistent way to display the results of this discussion when it is relevant: no way to show -- when it is important -- whether we are talking about the highest prototype letter a or the particular smudge of ink at this point on this page -- or any of the prototypes (or versions, depending) in between.  And, of course, no way to avoid making the distinction when it doesn't make a difference.

3.  Is there a clear prototype theory?  Wellll.  As such things (philosophical or linguistic theories) go, several pretty clear ones, all similar enough to handle most of the same problems similarly, different enough to allow arguments about hard cases.

4.  Does prototype theory require a new ontology?  Not necessarily.  Most say something vaguely like what And has (even more vaguely) presented, but some claim this is merely a convenience and then (try to?) unpack it in terms of various machines of linguistics or logic (finite state automata, transformations, Adjukeiwicz cancellation grammars, ....), whiles others really seem to believe the stuff they say (though what they say is rarely blatantly self-contradictory).

5.  Does this mean that prototypes should replace typicals as the referents of {lo'e} expressions?  Not obviously.  On the one hand, it is not clear that prototypes are nearly as common as typicals as points we want to make or distinctions that cause problems.  Further, it is not clear that gadri are the appropriate way to make prototype/version distinctions: these are relative distinctions after all, not absolute ones (like genus/species outside biological taxonomy), with a series of prototypes/versions between top and bottom (though usually transitively settled, so that this smudge on this page is a version of the letter a (as well as of my a's, script a's, italic a's ....).  Of course, there is the other side; that, since typicals are not really individuals, they should not be represented by gadri at all.  But much the same can be said about several versions of prototypes.
In short, now that I understand what And is about, I am no closer to thinking he has made his case than I was before, I just know what the case is.
--part1_13d.3c5b2a9.2911fe8c_boundary--