From lojbab@lojban.org Mon Apr 30 15:41:37 2001 Return-Path: X-Sender: lojbab@lojban.org X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_1_2); 30 Apr 2001 22:41:33 -0000 Received: (qmail 80488 invoked from network); 30 Apr 2001 22:41:32 -0000 Received: from unknown (10.1.10.26) by m8.onelist.org with QMQP; 30 Apr 2001 22:41:32 -0000 Received: from unknown (HELO stmpy-5.cais.net) (205.252.14.75) by mta1 with SMTP; 30 Apr 2001 22:41:32 -0000 Received: from bob.lojban.org (209-8-89-145.dial-up.cais.com [209.8.89.145] (may be forged)) by stmpy-5.cais.net (8.11.1/8.11.1) with ESMTP id f3UMfTM94181 for ; Mon, 30 Apr 2001 18:41:29 -0400 (EDT) Message-Id: <4.3.2.7.2.20010430175548.00ace2b0@127.0.0.1> X-Sender: vir1036/pop.cais.com@127.0.0.1 X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Mon, 30 Apr 2001 18:44:27 -0400 To: lojban@yahoogroups.com Subject: Re: [lojban] Predicate logic and childhood. In-Reply-To: <20010430144802.B27753@digitalkingdom.org> References: <4.3.2.7.2.20010430171850.00b16cd0@127.0.0.1> <20010430155318.B504@twcny.rr.com> <20010430114521.C20818@digitalkingdom.org> <20010430155318.B504@twcny.rr.com> <20010430140028.A27753@digitalkingdom.org> <4.3.2.7.2.20010430171850.00b16cd0@127.0.0.1> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed From: "Bob LeChevalier (lojbab)" X-Yahoo-Message-Num: 7004 At 02:48 PM 04/30/2001 -0700, Robin Lee Powell wrote: >On Mon, Apr 30, 2001 at 05:36:31PM -0400, Bob LeChevalier (lojbab) wrote: > > At 02:00 PM 04/30/2001 -0700, Robin Lee Powell wrote: > > >On Mon, Apr 30, 2001 at 03:53:18PM -0400, Rob Speer wrote: > > > > Okay, I should have looked a little harder. Your C does have a 'ko' > in it. > > > > > >Correct. Note that either ko binding stops at the .ijo, or jbofi'e has > > >a bug in it. > > > > I believe that variable binding must continue across an .ijo boundary > > because such a sentence pair can in theory be re-expressed as a single > > sentence having identical meaning (subject to some special rules for > > negation and the like), using gi'o or using go...gi, and the > > presumption in either case would be that the ko has scope over the > > entire compound bridi. > >Why? Because that is what the refgrammar says? (pg 403-405 - the section title seems to imply it about negation, but the statements therein are not restricted to negations, but merely say that negation doesn't affect their applicability. See especially 10.2, which is not a negation and the statement following it, and then the example with a prenex in 10.6). If two logically-connected sentences are equivalent to a single sentence, then the logical scope has to always continue beyond the connection. It is possible, though not necessarily easy, to express shorter scope. >Once again, either your wrong or jbofi'e has a bug: > >go ko viska gi cusku >(0[go {ko viska} gi cusku])0 I don't know anything about jbofi'e or how it would indicate scope. The standard parser would not show the continuation of scope either. That is because the parser indicates grammatical structure/scope; logical scope is a semantics question. >And I see no indication of such scoping in the grammar, but that proves >very little. > >Note that, IMO, you're moving the definitions in the wrong direction. >IIRC, go...gi is _defined_ as a shortform for a two sentence connective. Forethought connectives are defined in the section following afterthought connectives, but there is no relationship between the two discussed at that point - they are merely alternate ways of expressing connection (the text seems to get distracted on why *English* chooses one or the other, but doesn't at that point say that they are equivalent). It is later, in the logic chapter after discussion of prenexes that the two are said to be equivalent and I see nothing there that suggests one is a short form of the other. GIhEKs on the other hand seem to be described as a short form when the left side of the selbri is shared in common. Now, there is one tidbit suggesting the opposite, but it is weak and indirect. On 468, in the discussion of topic/comment, it says that you need to use tu'e/tu'u to get a topic in the prenex to extend over more than one sentence. If one were therefore to interpret "ko" as a "topic" when exported to the prenex rather than as a "variable", then it would not extend across the sentence boundary without tu'e grouping. But this remains a semantic question, with two semantic principles competing, in this case a logical one (of scope) vs. a linguistic one of topic/comment. Like the story-time tense convention, one has to pay attention to the context and realize that there are times when usage will not fit the convention. > > However, missing the context that led to the current example, one > > should not being using a logical connective to express this room/park > > conditional, because there is time-order or causality involved. The > > room being cleaned may be necessary, sufficient, preceding, or > > causally determining whether the child goes to the park; the two > > clauses do not in fact have independent truth values as is necessary > > to use logical connectives. > >I disagree, because whether the child cleans eir room and whether the >parent gives permission to go to the park are independent by default, >but assuming you're right. > >Two questions, then: > >1. What's the point of allowing things like .ijobabo? Because it is hard to prohibit them? It is a short way of expressing two things at once - that the two sentences are independent and equivalent, and that the event of the second sentence takes place after the first. But I have never seen babo used other than with .ije so I can't necessarily come up with a reason why someone would use it with .ijo. Note that the point of the "bo" is to make sure that the "ba" does not attach to any sumti following the connective but instead remains part of the connective. There is another usage of bo, of course: to force tighter binding of connectives in contrast to the normal left-grouping, but that takes a series of three or more connected sentences before it comes into play, and we have ijobake for the same function with different grouping. >2. What is the use of conditionals other than .e? Well, "a" expressing alternation (and anai for conditional) has a pretty clear usage, and onai for exclusive-or does as well. But the use of logical connectives in general is based on the assumption, probably false, that language is attempting to express logical truth. By insisting that the language connectives follow the rules of logic strictly, people will either start thinking in a manner that reflects those rules or they will avoid the use of connectives (or they will break the rules). This is a key design feature for the original purpose of the language as a test of Sapir-Whorf, though I personally think that the formal logical aspects of Lojban don't come into play enough that it will matter much. I haven't done any examination of the corpus to see how people have used the other connectives, and whether they have done so correctly and consistently. There has been usage of jo and ju, but not much. We have concocted examples where they can be used generally in logical argument rather than in assertions about reality, but it isn't clear whether they will ever be used in the latter. lojbab -- lojbab lojbab@lojban.org Bob LeChevalier, President, The Logical Language Group, Inc. 2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273 Artificial language Loglan/Lojban: http://www.lojban.org