From lojbab Tue Jun 11 02:44:04 1991 Return-Path: Message-Id: Date: Tue, 11 Jun 91 02:43 EDT From: lojbab (Bob LeChevalier) To: lojban-list Status: RO Subject: Loose Ends 5th of 6 related messages This message addresses a couple of loose ends that fall out of the previous discussions. Consider two apparently unrelated features of Lojban: - In discussing sumti-raising, we tried to note that "tu'a le sumti" is equivalent to "lenu le sumti cu XXX", where XXX is some unspecified selbri. You cannot use "broda" which is a nominally assigned word just as "ko'a" is. Nor is "bu'a" correct; it corresponds to the bound quantified variable "da". We needed a selbri equivalent to "zo'e", the 'unspecified because I'm not bothering to - but it has a specific value' variable. - You are allowed to put places on a bridi with no specified relationship in the place structure. This occurs naturally if you supply 3 places on a 2 place bridi. But how do you describe that place. There is no non-specific member of BAI. These two situations now come together as 5. 2 new cmavo for "unspecified bridi": co'e (selma'o GOhA): pro-bridi for an unspecified bridi do'e (selma'o BAI): case tag for an unspecified relationship The memory hook here is "zo'e", the elliptically unspecified sumti. "co'e" can be used as its own rafsi (affix) in compounds, though I'm hard pressed to come up with a useful example. Any takers? John suggested when proposing it: Note also that "co'e" (proposed for "unspecified bridi") can also serve as a rafsi: "co'epre" = "unspecified type of person". but this makes no distinction from simple "prenu", so why use it? [Nora opines that someone may wish to use parallel structures in lujvo, as in: ti xaupre This good-person. ti xlapre This bad-person. ti co'epre This unspecified-person. but to me this seems more reasonable in tanru than in lujvo - the 3rd brivla seems to be intended to mean 'unspecified as to goodness or badness - person' while a different set of examples would suggest a different meaning. This context-dependent meaning is appropriate for tanru but NOT for lujvo. Any other ideas?] 6. Much of this many-part proposal is tied up in the use of selma'o BAI, and we've been forced to re-examine what these are. The specific question is whether BAI is the exact equivalent of the gismu which we selected as a memory hook. Or in other words, is a member of BAI an abbreviation for a specific FIhO construct. This was not the original intent, which was to decide on certain useful or needed roles (yes, 'cases' jimc) that could or should be useful in expanding bridi, and then to pick words for them. As a basis for this we used Jim Brown's work for Institute Loglan, coupled with some research into case theory, and the everyday, if malglico, analysis of English prepositional and subordinate phrases. Institute Loglan has TWO sets of these - case tags that are usable only to label place structure places, and 'modal relative phrases' which are used to attach non-place structure terms. There is some overlap and some commonality between these and the two are not interchangeable - the case tags are more like our FA tags. We wanted to have the capability that Jim Brown intended for 'case tags' without the restrictions and duality. BAI was formed with the intention that every place structure place could be labelled more or less accurately with one of these. I know that Nick and Ivan D. has challenged us on this - and the problem in meeting this challenge is that we have drifted from this original, rather more metaphorical use of BAI; the current meanings are too closely tied to their associated gismu, and have lost touch with their original purpose. Thus we have John Cowan's proposal: 6. Bogus "ci'a" tag: This tag is really two different concepts: "fi'o cusku" (= "cu'u") and "fi'o finti" (which doesn't have a tag). One refers to the "expresser", the other to the "creator". "ci'a" is meant to mean "written by", but the question of whether the writing is recorded on a medium is secondary to the two other questions. The reason that "cu'u" is not the same as "fi'o finti" can be shown thus: mi nelci la .apasionatas. ne fi'o finti la betoven. I love the Appassionata, composed by Beethoven. mi nelci la .apasionatas. pe cu'u la artr. rubenstain. I love the Appassionata performed (expressed) by Artur Rubenstein. Note the "ne" vs. "pe"; there is only one Appassionata as composed, but there are many performances of which I am specifying Rubenstein's. There are no "f" cmavo available, so if "finti" is to get a tag (freeing "ci'a") something else must change. The candidates are: fa'i VUhU reciprocal of [fendi] fe'i VUhU divided by [fendi] fi'i COI hospitality, welcome to ... fo'i KOhA it8 fu'i UI freely (attitudinal scale) [frili] fi'a FA place structure ? fi'e GOhA parent bridi fi'o FIhO selbri to modal fi'u PA fraction slash [fendi] Alternatively, "ci'a" can be declared to be the tag for "finti", and bugger the morphology. I (lojbab) suspect a similar case could be made against "ja'o", and "du'o" is a metaphorical use of "djuno" at the very least. I support John's proposal, assigning "fi'e" and freeing "ci'a". As discussed in message 3, the current "fi'e" and "fai" get moved elsewhere. John, on the other hand looked elsewhere for a cmavo, resulting in the following: 7. Inverse fraction slash: This is "bi'u", and was intended for representing the reciprocal fractions like 1/2, 1/3, 1/4, etc. It also was meant to capture the reverse solidus, "\". However, it is not a part of standard mathematics. The reciprocal fractions can easily be handled with the ordinary slash, as /2, /3, /4, etc. As for "\", it has no standard mathematical meaning. Using "bi'u" for forward slash (instead of "fi'u") allows "fi'u" to be freed for the previous proposal. I support the argument for eliminating "bi'u", but we need not change "fi'u" given the other change proposed. Since John forgot that "fi'u" is derived from "frinu", this would have lost a valuable memory hook. This solves the problem of one word in BAI, but not the general problem. With regard to the BAI selma'o in general, we propose recognition that FIhO equivalence may be inevitable, although we would prefer to keep rather more metaphorical definitions. Recognizing that there are some horribly malglico metaphors, we intend to devise equivalences for BAI, not necessarily to a gismu but possibly to some compound (lujvo) from which we can derive the intended usage. With the exception of "ci'a", I suspect that most of these are best made as lujvo based on the current gismu hook, and this means no other BAI members need to be changed. However, some of the meanings of the BAI members, especially SE-converted ones from the 'main use' we had in mind when we selected the word, may change. I'm asking John to take a first cut at this, and it will be reviewed at LogFest prior to posting, given the shortness of time before that event. ---- lojbab = Bob LeChevalier, President, The Logical Language Group, Inc. 2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273 lojbab@snark.thyrsus.com