Return-Path: Message-Id: <9110311659.AA05430@relay1.UU.NET> Date: Fri Nov 1 14:55:16 1991 Reply-To: cbmvax!uunet!pucc.princeton.edu!jimc Sender: Lojban list From: cbmvax!uunet!pucc.princeton.edu!jimc Subject: Re: Lojban duplications X-To: lojban@cuvmb.columbia.edu To: John Cowan , Ken Taylor In-Reply-To: Your message of "Wed, 30 Oct 91 20:01:00 EST." <9110310629.AA22854@julia.math.ucla.edu> Status: RO X-From-Space-Date: Fri Nov 1 14:55:16 1991 X-From-Space-Address: cbmvax!uunet!cuvma.bitnet!LOJBAN Art Protin writes: > I am not sure how I want to argue this. For one thing, I see that this > is a very logical extension of the distinction being made between > "klama" and "litru", and for the other, I strongly disapprove of the > distinction being made ... and want to abandon any gismu with the > sole distinction of having fewer places. I've reluctantly been dragged a small distance out of Art's camp by Lojbab's arguments and by a counterexample I came across in Nick's "Colossal Cave" text. My former position was that each gismu should have a definite and fixed list of essential places in the relation (and in consequence, klama vs. litru is a bogus distinction since both express the same relation and therefore must have the same places). Nick translated "The walls are frozen rivers of orange rock". rirxe = "x1 is a river of landmass x2". There is no for "made of material" (and I think there should be). 1. Nick used le bitmu cu dunja rirxe flecu lo narju rokci the walls are frozen river flows of orange rock where flecu = "x1 is a current of x2 flowing to x3 from x4 (no route)". Thus he injects a place for the material. I think that to use flecu is to cheat, and stubbornly insist on using rirxe as the main predicate as is done in the English. It's an outrageous restriction on our ability to express our thoughts if you can't assert that x1 is a (something) made of material x2, for any main selbri, not just the ones predefined with a material place. (No blame to Nick is intended for accomplishing the mission the only way he could.) 2. Per Lojbab, place structures are, at bottom, negotiated between the users of the language, and the dictionary entries merely reflect the negotiated real-world usage patterns. This is particularly true of BAI places, in particular, which BAI's (if any) are "essential" to the relation. In the rirxe example it's a numbered place which needs to be instantly renegotiated. So (jimc infers, definitely not agreeing), just as in English, you don't get anally retentive about landmasses in the dictionary and just say: le bitmu cu dunja rirxe lo narju rokci the walls are frozen rivers of orange rock 3. Per jimc, staying within the mainline grammar: le bitmu cu dunja rirxe fi'o sligu lo narju rokci the walls are frozen rivers of solid orange rock The problem with this is, while Lojbab assures me that "fi'o sligu X" is just like "cu'u X" or any other BAI phrase, the form of the phrase is a restrictive subordinate clause, as in "le du'u (le bitmu cu dunja rirxe) cu se cusku la nik." (...is said by Nick) But "le du'u (le bitmu cu dunja rirxe) cu sligu lo narju rokci" is ridiculously false. (Note, there is no gismu that I can see which means only "x1 is made of material x2"; all gismu with materials say "x1 is some kind of thing, made of material x2", e.g. solid.) 4. This construction sweeps the whole issue under the rug by making parallel assertions -- but again, that's cheating. le bitmu cu dunja rirxe gi'e narju rokci the walls are frozen rivers and are orange rock In all the gismu where materials arise, the material has a numbered place, as in tanbo = "x1 is a board of material x2", and thus is identified as an essential element of the relation (by Lojban Central, yes I hear Bob Chassell's argument, I think it was). I want Nick's off-the-wall material place to be equally essential. On the other hand, going back to the cusku example above, I claim that the speaker identification is never essential, so that the subordinate clause format or interpretation is not only correct but is preferred by me to the "real" BAI relation. Also, Mark Shoulson argues convincingly that if all possible BAI's and FI'O's are elliptically present as an essential element in every relation, the practical and philosophical job of analysing the relation becomes ridiculous. It sounds like BAI and FI'O signal one grammatical form that represents two meanings: an extra essential place in the relation, or a kind of restrictive clause. It will be necessary to distinguish clearly between these meanings. -- jimc