From LOJBAN@CUVMB.CC.COLUMBIA.EDU Sat Mar 6 22:58:56 2010 Return-Path: Delivered-To: veion@XIRON.PC.HELSINKI.FI Received: (qmail 15254 invoked from network); 13 Dec 1996 21:07:18 -0000 Received: from segate.sunet.se (192.36.125.6) by xiron.pc.helsinki.fi with SMTP; 13 Dec 1996 21:07:18 -0000 Received: from listmail.sunet.se by SEGATE.SUNET.SE (LSMTP for OpenVMS v1.1a) with SMTP id <10.5B59AF8C@SEGATE.SUNET.SE>; Fri, 13 Dec 1996 22:06:30 +0100 Date: Fri, 13 Dec 1996 12:50:09 -0500 Reply-To: Pycyn@AOL.COM Sender: Lojban list From: "John E. Clifford" Subject: Re: lojban imperfections? X-To: lojban@cuvmb.cc.columbia.edu To: Veijo Vilva Content-Length: 5575 Lines: 90 Message-ID: <_bLHyIusFCH.A.3_B.w60kLB@chain.digitalkingdom.org> coi kris To take your last question first, I don't think there is anyone who is native- fluent in Lojban. There are several people -- you will hear from some of them as soon as you put your first bit of Lojban on this list -- who are pretty good, though. There have been some moderately lengthy and not very halting, conversations at meetings and over the telephone as well as computer chats (which may or may not count). Transcripts of some of these are around somewhere. They do suggest that speakers rarely use all the places of predicates with more than four or five places and even then - - in fact after about three -- are liable to throw in a "preposition" just to be on the safe side. Pushing the limits of the 5+/- 2 and right branching rules is one possible part of a Sapir-Whorf test (or an IQ-enhancer for that matter). But, yes, Lojban is meant to be human-speakable and actually is on occasion (exactly what is required for a proof that it is is not clear). 6. Order is tricky. Two factors, at least, have been involved. One is a (necessarily informal) survey of the relevant orders in typical sentences in major languages (and minor when they were interesting and available). This showed that, for example, agents preceded patients and recipients pretty regularly and the reverse order was pretty rare, so SO got fixed. There are also pretty general cross-language order patterns for various kinds of additional terms, though they are not as clear cut. The second factor was an (possibly subjective -- though I think therehave been some tests) judgement about the centrality of various arguments in the predication, which ones need to be specified most often. This, of course, starts with the question of what places to specify and then goes on to their order -- less central are later (since they can more often be dropped without loss of intended information). English is a fairly typical language in both these respects, but not the only one consulted. And notice that all orders are legal in Lojban and require only minor tricks to produce. 5. Actually, all the roots (gismu -- we use this as a technical term, so as not to carry over any freight from other fields that may not fit) are unambiguous *in Lojban*, but get translated various ways in various contexts in English or whatever (the first native speaker will probably the first person who really senses this fact). Lujvo are, however, open to wider interpretation (as indeed are tanru), only their syntax, not their semantics (nor pragmatic neither), is fixed by their structures. So, again, it is a human language, since ambiguity is one of the essential of human language -- art, if nothing else(and law, of course). But there is no *syntactic* ambiguity, so that parsing by machine is possible, even if translation/interpretation is not. 4. Well, those most popular langauges represent a rather wide array of cultures, about as wide as you can get in the growing homogenized world. Also, no gism are taken from any of these langauges, they are merely meant to be accessible to them (contrast Esperanto, for example, or some of the even worse Interglossae). More importantly, however, the physical shape of the words is relatively unimportant from a cultural point of view, what is important is how each fits into the rest of the vocabulary and into the sentence structure. And then no Lojban gismu is like any word in any other language (try a fairly literal -- by cognates or by dictionary -- translation from English into Lojban; the Lojban never means much nearer than shouting distance of the English, as we all learn about week three). Here as (did I mention it back then?) in matters of word order and the like, Lojban tends to be neutral by including as many variations as possible, without significantly favoring any. 3. No, the S-W test has never been done, not even finally designed, so far as I know. The design must wait for the finished language to calculate where the most likely effects are to be found and then the test devised to hit those points especially. All plans I have ever seen also require a cadre of several dozen fluent Lojban speakers and a big grant. Neither of these is in the offing. 2. Logicians have been involved in the design of Lojban (and other Loglans) since the start. We (I am one, on the project since '76) do know the standard terms and also the theoretical freight they carry. So we use new terms, which we can load up the way we want to. There are not that many differences in some cases, but rather enormous ones in others (no logical or linguistic expression in the NP class catches what all is involved in a sumti, for example). Ditto linguists (me again, for one). 1. A mess of factors for SVO (but notice that VSO and SOV are usually equally as good and without any extra flags-- the three most common forms, over 90% of base forms worldwide, it is said). 1) It is the most common order worldwide (not theoretically significant, just handy for learning -- like the gismu forms). 2) In logic, while it is not the usual form, it is a common form for making non-standard points (negative empty universe claims, for example) and it is nice to have that ability in Lojban, since it saves a bit in the syntax (but is not much used anymore). 3) Again in syntax, SVO save a bit on various non-declarative forms like imperatives and observatives, though not so much in the current language as in earlier versions. Finally, word order is not thought to be culturally significant in the Sapir-Whorf sense. >|83