From LOJBAN@CUVMB.BITNET Sat Mar 6 22:46:11 2010 Reply-To: Logical Language Group Sender: Lojban list Date: Wed Dec 6 07:03:57 1995 From: Logical Language Group Subject: Re: TECH: new cmavo "ju'e" X-To: cowan@LOCKE.CCIL.ORG X-cc: lojban@cuvmb.cc.columbia.edu To: John Cowan Status: OR X-From-Space-Date: Wed Dec 6 07:03:57 1995 X-From-Space-Address: LOJBAN%CUVMB.BITNET@UBVM.CC.BUFFALO.EDU Message-ID: John missed one of my arguments regarding "stag bo" in after thought. Currently causal connectives are expressed using "A stag bo" or "JA stag bo" or "GIhA stag bo" etc. in order to clearly specify the scope of the afterthought. It is a fundamental design principle in Lo??an that such scope be absolutely unambigous, with the umpteen possible variable-scope endings of "John went to the window and ..." each reflected in a clearly different way. AS disucssed in another thread, JOI has represented a slight muddying of this design principle, since it has the grammar collectively of two differently scoped logical connectives, E and JE, in order to save cmavo in an ill-defined set of connectives that could be open-ended. In explicitly non-logical circumstances, a little smudging of the rules seemed OK. BUt people have of late been finding more and more uses for non-logical connectives, some of which are not necessarily quite without logical import. But giving JOI this broad grammar had a price, partly in terms of complexities in elidable terminators, and partly in terms of significantly increasing the overall complexity of the grammar. Put simply - I don't LIKE the grammar of JOI, and consider it non-Lojbanic, but accepted it as a price that had to be paid to keep from having a separare set of JOIKs for each of the different logical scopes. Even WITH the current design, there are no JOIK equivalents for GUhEks and GIhEks (fortethought tanru and bridi-tail connection). The complex grammar of JOI makes errors in use and resulting scope likely unless the proper terminators are used. At least, with the limited set of JOI members, though, the listener can be warned that the speaker is treading on dangerous ground, allowing for some quick analysis and a requesto for error correction. The proposed usage of stag or stag-BO as a connective, without a preposed connective scope marker, would add an infinite set of cmavo-compounds to JOI, and they would be difficult to resolve on the fly since they make use of the BAI words that already have two other usages with different scopes (sumti tcita and selbri inflection). Just as Jorge's proposal to allow JE the full range of usage of JOI is unacceptable then, so is a major increase of JOI-scoped connectives, especially since there already IS a means to connect using causals or other BAIs in after thought: we insert the BAI/stag in between a connective of the right scope and BO/KE. I would normally have done this using ".e ri'a bo", but Cowan convinced me last night that using a specifica logical connective like ".e" is logically risky - a causal does not necessarily want to claim ".e" truth conditions. One could decide on a specific and appropriate logical connective that HAS the desired truth table (perhaps a conditional), but the analysis might be different for different contexts. Ideally, what we need is a non-specific LOGICAL connective for each of the different scopes. Using JOI, with its overstretched scope rules is dangeorous and confusing, ESPECIALLY when the problem being posed is specifically one of scope. But adding a non-specific logical connective to each connective selma'o is impractical - we don't have a lot of cmavo free, and the logical connective paradigms are in a pattern that has no room for expansion. I am thus RELUCTANTLY conceding to John's proposal for ju'e BECAUSE it makes no grammar change and uses only one cmavo, and more so, because it preserves the existing principle of clear scope marking in afterthought connection. Thta it has the side benefit of allowing Veijo's sentence compression makes it a little more palatable. Myself - I'll try to figure out and put in the appropriate logical connective for each usage. I will note that the proposal is NOT a complete solution in any case, either for Veijo's compression or the causal problem in general. There is no JOI construct as counterpart to GIhEKs and GUhEks, and you cannot compress some constructs where the logical connectives are analyzed in the lexer/compounder. (e.g. you cannot compress constructs that would require other than "simple --- sorry, I mean you cannot compress constructs that ONLY can be expressed using a non-simple JOIKJEK as defined in the YACC grammar). Since GIhEks and GUhEks are both candidates for causal connection, and for compression, a JOI-based solution is only a partial one. But it will have to do. The only other solutions I can thing of would be to extend into the now experimental CVVV cmavo space (perhaps using gi'e'i gu'e'i ji'i'i and je'i'i for non specific connectives in GIhA GUhA A and JA respectively), and that also is something i am reluctant to do before the 5 year baseline - it may be the proper long term solution though. But at least ju'e, by being explicitly a memberof JOI preserves the pricniple of explicitly marking connective scope with one of a small set of cmavo. lojbab