From @gate.demon.co.uk,@uga.cc.uga.edu:lojban@cuvmb.bitnet Thu Jun 22 23:30:22 1995 Received: from punt2.demon.co.uk by stryx.demon.co.uk with SMTP id AA3564 ; Thu, 22 Jun 95 23:30:20 BST Received: from punt2.demon.co.uk via puntmail for ia@stryx.demon.co.uk; Thu, 22 Jun 95 16:46:08 GMT Received: from gate.demon.co.uk by punt2.demon.co.uk id aa06918; 22 Jun 95 17:45 +0100 Received: from uga.cc.uga.edu by gate.demon.co.uk id ab07810; 22 Jun 95 17:45 GMT-60:00 Received: from UGA.CC.UGA.EDU by uga.cc.uga.edu (IBM VM SMTP V2R2) with BSMTP id 7498; Thu, 22 Jun 95 12:43:30 EDT Received: from UGA.CC.UGA.EDU (NJE origin LISTSERV@UGA) by UGA.CC.UGA.EDU (LMail V1.2a/1.8a) with BSMTP id 3537; Thu, 22 Jun 1995 12:42:20 -0400 Date: Thu, 22 Jun 1995 12:42:49 -0400 Reply-To: John Cowan Sender: Lojban list From: John Cowan Subject: Re: Another question about imperatives X-To: Lojban List To: Iain Alexander In-Reply-To: <199506220747.DAA16699@locke.ccil.org> from "jorge@PHYAST.PITT.EDU" at Jun 21, 95 09:49:55 pm Message-ID: <9506221745.ab07810@gate.demon.co.uk> Status: R > la kris cusku di'e > > > I wasn't sure about that -- I thought .e connected two sumti within the > > context of a particular place, rather than conjoining the places themselves. > > I.e. you could say "ne'i le botpi .e le tanxe" for "in the bottle and also > > in the box" but not "ne'i le botpi .e ne'i le tanxe". Or can you do both? la xorxes. cusku di'e > The first is definitely right. I think you can't do the second (although > the parser accepts it, but comes up with some strange stuff) so my translation > was wrong. {ne'i le botpi e ne'ibo le tanxe} is right too, but I have > no idea what it means. It means "within the bottle and the cup (where the bottle is inside the cup)" and so refers to a nested state of affairs. :-) The parser accepts "ne'i le botpi .e ne'i le tanxe" because of a long-standing and apparently ineradicable misfeature. Because we are using yacc, which is designed for programming languages, we get into trouble. Inserting elided terminators is done using yacc's error recovery, which sometimes has a habit of skipping three tokens in an attempt to avoid too many closely nested errors. This is fine for error recovery, but bad for a language feature. To take an unambiguous example: bai le botpi .e bau le tanxe cu co'e parses as: ({ } cu {co'e VAU}) because the error at ".e bau" is referred back three tokens earlier to the recovery step of inserting "KU" after "bai", and then the three tokens get tossed. I really must put in a check to see to it that the parser outputs every word it inputs. -- John Cowan cowan@ccil.org e'osai ko sarji la lojban.