From @YaleVM.YCC.YALE.EDU:LOJBAN@CUVMB.BITNET Thu May 27 08:16:26 1993 Received: from YALEVM.YCC.YALE.EDU by MINERVA.CIS.YALE.EDU via SMTP; Thu, 27 May 1993 08:16:26 -0400 Received: from CUVMB.CC.COLUMBIA.EDU by YaleVM.YCC.Yale.Edu (IBM VM SMTP V2R2) with BSMTP id 3521; Thu, 27 May 93 08:15:37 EDT Received: from CUVMB.COLUMBIA.EDU by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with BSMTP id 4760; Thu, 27 May 93 08:16:52 EDT Date: Thu, 27 May 1993 13:07:36 BST Reply-To: Ivan A Derzhanski Sender: Lojban list From: Ivan A Derzhanski Subject: self-segregating morphemes To: Erik Rauch In-Reply-To: Logical Language Group's message of Sun, 23 May 1993 14:59:27 EDT <9267.9305232127@cogsci.ed.ac.uk> Status: RO X-Status: Message-ID: > Date: Sun, 23 May 1993 14:59:27 EDT > From: Logical Language Group > > Part of the reason for 'all that complexity' is the requirement that ALL > gosmu have combining forms, not just some of them, It is still only some of them that have 3-letter rafsi. > and that the gismu/rafsi list have some (if not a lot) > expansibility, so that new gismu can be added when new concpets arise. But it can be reasonably expected that the most fundamental gismu have been on the list from the very beginning, and surely it is the most fundamental gismu that warrant 3-letter rafsi. I agree that the current system is astonishingly complex and constitutes an unnecessarily great memory load, but I wouldn't go for even a low degree of ambiguity. I would still have preferred no more than one 3-letter rafsi per gismu, determined in a fixed way; say {qhawi} always yields {qha} and {qahwi} yields {qai} or {qa'i} (and there might be a way to determine which {hw} becomes zero and which a voiceless glide). The vocabulary could have been so constructed that the most fundamental gismu would each get a different 3-letter rafsi, built by the above scheme. For the rest one would use 5-letter rafsi. (I'm not saying `let us start it all from scratch', just thinking aloud.) > On this subject, though. I was thinking last night, and I realized that > there is another system that has two-level segregation, and which people > seem to learn without too much trouble: the chemical elements and their > symbols. You effectively have some 100 'gismu', the full element names, > and their atomic symbols, as rafsi. People studying chemistry have to learn > both to effectively learn material. While the numbers are smaller, I do not > recall any particular difficulty in learning atomic symbols as I needed them, > and thus being able to identify the components of a chemical compound (a > 'lujvo'), even though several of the symbols are not even mnemonic. All of them are mnemonic, actually, though based on the Latin names, not the English ones. Now consider what life would be like if the people who designed that system had came up with anywhere between zero and three variant atomic symbols for each element. Would that improve the readability of chemical formulae? Far from it. This is where chemical notation beats Lojban. Mind you, there are better examples. English uses a number of Latin and Greek morphemes which in many cases are counterparts to full English words. (You can think of _kilo_ as a rafsi for _thousand_ or _scop_ and _vid_/_vis_ as rafsi for _see_.) Japanese and Korean have rafsi of Chinese origin, and many American Indian languages have rafsi denoting directions, body parts and lots of other stuff that are only distantly (if at all) related to the independent words. None of this makes those languages any easier, though, let alone more logical. Ivan