Return-Path: Message-Id: <9110041921.AA29950@relay1.UU.NET> Date: Fri Oct 4 18:33:28 1991 Reply-To: Logical Language Group Sender: Lojban list From: Logical Language Group Subject: VSO order X-To: lojban@cuvmb.cc.columbia.edu To: John Cowan , Ken Taylor , List Reader Status: RO X-From-Space-Date: Fri Oct 4 18:33:28 1991 X-From-Space-Address: cbmvax!uunet!cuvma.bitnet!LOJBAN Mark Shoulson says in response to Dean Gahlon: >Using "VSO" form, {citka le nanmu le cripu}, is quite grammatical, but >poses a different problem. By current usage, since VSO is not a common >word-order in many languages, the "selbri-first" word-order is reserved >for "observative" sentences--ones with the x1 place ellipsized. Thus, >the above sentence would probably be understood to mean "(something) >eats the man ??? the bridge"--since "citka" only has 2 places, it would >be unlear how the bridge related to it all. > >In private e-mail, jimc and I have discovered that we both would prefer >to allow VSO to enjoy the same treatment as anything else, making it >more consistent, as well as easier to use VSO (which we both seem to >like). What are everyone's opinions? I'm going to interpolate what you are proposing as being that citka le nanmu le cripu should be x1 x2 and not the current x2 x3 with x1 unspecified in other words, you want an >unmarked< VSO order. I note that you can express VSO with the minimally marked citka fa le nanmu le cripu x1 x2 So there is no constraint against using VSO. Why is the 'fa' necessary? The answer is the usage we call the 'observative', where the speaker omits x1 because the predicate ('verb') is most important in the sentence and the x1 value is obvious. I note that in old Loglan, the unpreceded predicate was the imperative, and it was the justification for the observative as an equally basic form that successfully out-argued the imperative command for this form. It cannot be argued that VSO is even more basic or necessary, so as to justify a less-marked form than the observative. I will aside here, since this isn't the first time this has come up, to acknowledge a previous proposal by Nick (I think, and I hope I got it straight) that would retain the observative for the unpreceded predicate if there is no trailing sumti, but if there were, require fe to skip to the x2 place. In other words, reversing the marking requirement so that the sumti-ed observative is more highly marked than the VSO form. The answer is still the same - the observative is more basic a language usage than the filled in VSO. This is hard to argue in the context of written text, because observatives tend to arise primarily in spoken language rather than written, and more often in children's talk than adult's. (But see below - Lojban HAS a major written use for observative/ ellipsized x1). We use a LOT of observatives in conversations around here. Loglan/Lojban is essentially a rigid word order language, designed by JCB to be rigidly SVO. The late development of the Hikson/Bonewitz tags (our fa/fe/fi series) was one way to change the order. We also early on noticed that the machine grammar did not care how many sumti are to the left or right of the selbri/predicate, making the SOV an equally valid unmarked order, and raising the question/possibility of VSO being unmarked. (JCB also added and still has in his version of Loglan, some funky sumti rotators that shift sumti around in still another way (with no particular use or basis in natural language) moving trailing sumti to the front in order, or leading sumti to the end in order. These went quickly in the Lojban redesign as they are redundant to the fa tags and virtually impossible to use without firm and memorized complete place structures.) So unlike Esperanto, for example, Lojban is not designed for freely expressing sumti in any order. Lojban allows it, but it is not basic to the language. Indeed, the rest of Lojban's grammatical structure remains well-linked with SVO order in ways that further argue for the current interpretation of "citka le nanmu" as "eater of the man". The first such argument is that of 'be' linkage used in specified descriptions. In Lojban, the following two structures mean the same thing klama le nanmu le cripu klama be le nanmu be'o le cripu goer to the man from the bridge But the second expression binds the sumti as x2 onto the selbri allowing le klama be le nanmu be'o le cripu the goer to the man from the bridge as a sumti. If VSO were the natural interpretation, these would have to be equated as: klama le nanmu le cripu * goes the man to the bridge klama be le nanmu be'o le cripu goer to the man from the bridge and the parallelism between these two basic forms is lost and you might even have to mark the latter even more highly as: klama be fe le nanmu be'o le cripu goer to the man from the bridge But this makes little sense because the essence of sumti description with 'le' is using a predicate to describe its x1 by stating its relationship with other things: the goer is defined by where she/he/it goes to/from/via/using, and by the basic fact that a relationship involving klama (going) applies. Another area, the one I promised above, is in abstractions. Even with all the gagalag about cleft place structures a few months ago and in JL15, there are still going to be many places where abstractions have 'obvious' or elliptically omitted x1 sumti (indeed, our analysis is finding that a lot of the apparent cleft structures still in the language truly aren't - but that's another thread) mi djica I desire the going to the market omits the desired goer, presumably the speaker 'mi' although some contexts might give different values. If VSO were a basic unmarked form of the language, you couldn't ellipsize the x1 without a marker 'fe' before 'le zarci'. Because x1 (the 'subject') has such a special role in the language grammar semantics and pragmatics, it is hard to imagine an aesthethic justification for unmarked VSO be approved by the speaking community. The minor easing of learning one less used form would be counter by making other forms more difficult to learn. Our language design so recognizes the importance of x1 at the expense of the other places, that there is little reason seen for any rearrangement of later places - hence my surprise a couple of days ago at Nick's conversion operator 'setese' in his 'Only' translations. That conversion basically switches x2 and x3 leaving all other sumti in the same order, but takes 3 syllables instead of the 1 syllable required to just say 'fi' before the x3 (his example had an elliptically omitted x2, but even if both x3 and x2 are expressed, it takes only 'fi' and 'fe',and you don't need to do the mental gymnastics of the triple conversion 'setese'. I personally see absolutely no need for such multiple conversion operators in usage unless some kind of dikyjvo system were adopted that forced certain unnatural and unuseful place structures that needed such conversion in order to be used in still more complex diklujvo. Oh - one last related point. One person not on the list - who was teaching English to Spanish speakers, noted that the ellipsized subject is even more used in Spanish and other inflected languages where the verb-ending gives some clue as to the intended referent, than in English. This is apparently true in Spanmish even when the verb-ending has multiple meanings - the language just tends towards ambiguous subjects. lojbab@grebyn.com