From LOJBAN%CUVMB.BITNET@UBVM.CC.BUFFALO.EDU Mon Jan 15 14:51:51 1996 Received: from VMS.DC.LSOFT.COM (vms.dc.lsoft.com [205.186.43.2]) by locke.ccil.org (8.6.9/8.6.10) with ESMTP id OAA10485 for ; Mon, 15 Jan 1996 14:51:45 -0500 Message-Id: <199601151951.OAA10485@locke.ccil.org> Received: from PEACH.EASE.LSOFT.COM (205.186.43.4) by VMS.DC.LSOFT.COM (LSMTP for OpenVMS v1.0a) with SMTP id 8EE8AA39 ; Mon, 15 Jan 1996 14:23:52 -0500 Date: Mon, 15 Jan 1996 14:22:55 -0500 Reply-To: Logical Language Group Sender: Lojban list From: Logical Language Group Subject: Re: SNU: ki'e doi skot. X-To: pbarreto@UNISYS.COM.BR X-cc: lojban@cuvmb.cc.columbia.edu To: John Cowan Status: OR X-Mozilla-Status: 0011 Content-Length: 6061 >That's a very poor standard. I know many (non-English-native) people with a >rich English vocabulary but a terrible knowledge of English grammar; the net >result of this combination is at most funny. > >Master vocabulary indicates good memory, not language proficiency. >Furthermore, the gi'uste is only a part the vocabulary a person must >memorize. What do you think is harder to remember, a single word "xorvo" or >the list "{gugdrxrvatska, >kulnrxrvatska, bangrxrvatska, ...}"? And words for those concepts must be >known (or looked up) anyway when we want to express the associated concepts. > >I doubt mastering the gi'uste weighs more than 20% in measuring Lojban >knowledge. >This means that adding, say, 50 new gismu would make the language less than >1% harder, if that much (no, I'm not proposing 50 new gismu, that's a >hypothetical number). > >IMHO freezing the number of gismu means resigning to accompany the evolution >of the concepts needed by human expression, therefore killing the language. >If youdeny >this, try to express "software" in Latin (Hmm, perhaps Nick wants to try :-) 1) Almost every analysis of semantic primitives in the various world languages, as well as roots, etc. tends to come up with between 1000-2000. Many so-called primitive languages have even fewer. 2) why do I need to memorize Xxrvatska unless I am talking to someone about Croatia, which in turn presumes that I know what/where Croatia is? I think it is likely that if I want to talk about a specific country, i can look up the name of the country a couple of times. Note that you DON'T need to memorize more than one word - it is "xorvo" or "-xrvatska" that you have to memorize. I think it likely that -xrvatska is easier to memorize since you might actually see it spelled like that somewhere else. 3) we are treading into an especially controversial ground with an example from the cultural gismu. There are some (Ivan for one) who feel we have far too many cultural gismu in the first place - indeed that we shouldn't have any. Adding cultural gismu is thus something that gets higher resistance than any other kind of gismu. 4) I think it is true that people will have to memorize a lot more than the gismu in order to be fluent users of the language, but in order to be NON-fluent users, most of the gismu and a subset of the cmavo are fine (that is as far as Jorge or I have gotten, for example). The lujvo, if not memorized, can be reconstructed for meaning on the fly from rafsi analysis, coupled with context. Likewise with fu'ivla. I may not have the vaguest idea where Croatia is, but i do know that gugdrxrvatska is a country - I do not necessarily know that about xorvo. In the meantime, there is something psychological about having a long list of words to memorize, and in my experience, ALL the difficulty in learning Lojban is in the word list. You can "fake it" on the grammar, on place strutures, and rafsi (using 4 letter forms) but you still have to know the gismu and some subset of the cmavo in rder to read any Lojban text. My experience is that learning the word list takes FAR longer than any other part of learning the language, even with the able help of LogFlash (LogFlash took me some 2 months of an hour a day, and it would likely take 3 months now given the increased number of gismu since Nora and I did that initial learning.) I should note that other than Nora and I, only 2 other people have ever REPORTED completing a first pass through the words using LogFlash, and no one besides me have reported getting to the 97% mastery level on the gismu list. The payoff is that I seem to be one of the few who doesn't ever have to have a wordlist in order to converse or read Lojban even though I haven't actively used the language more than incidentallyin over a year and am getting heavily crossfeed from my last 4 years working on learning Russian. 5) The problem with adding gismu is "where do you stop?" If you are to add a word for Croatian, you might need one for every country of the world, and then because of US dominance in the community, then maybe every state in the US. And that is just the cultural words. How about jargon terminology like software? How many fields do you need to cover to thoroughly cover jargon? And then why did we BOTHER to put in borrowing and compounding methodologies into the language. There are literally BILLIONS of lujvo with less than 4 terms, just waiting to be discovered; there is no language on earth with a significant fraction of that total. They might not help with Croatia (or maybe they would - you could make some appropriate lujvo from X-Y-South-Slavic, where Goran and friends chose the X and the Y to be meaningful and non-offensive to their culture). But outside of culture- specifric words, the language SHOULD be quite robust enough to handle concepts internally. If it is not, we have more wrong than the set of gismu. But seriously, we DID have the problem of "where to stop" a few years ago. As long as we had not said "no more", then every couple of weeks a new gismu was proposed. Each was duly debated at great length, and most debates were totally inconclusive. (Sound familiar from the current grammar discussions? Then you know why we want to freeze on prescribing grammar!) The last gismu votes were true parodies, with people giving blanket votes and proxies to vote yes or no on all proposals as a matter of principle. Since then, we decided to not have such votes take place in the formal meetings of the LLG, but still when half the community is voting totally on the basis of principles unrealted to the specific proposal, we have gone far beyond the point where meaningful debate is useful. 6) Then there is the problem of standardization. I'm not worried if we miss a lujvo of fu'ivla or two in compiling the dictionary, but we cannot miss any gismu. JCB had a real problem with gismu making during his dominance. From 1974-1984, perhaps a couple hundred new gismu were proposed by the community. To put it simply, most of the proposals were pure malkalci. But they were proposed, and they were often used. Unfortunately, almost no one hadd the resources to do a proper 8 language analysis (corresponding to Lojban's 6 languages), so most words were calques based on the 2 or 3 languages that the proposer knew. Many of the proposals had no business in gismu space - i think theremay have been several dozen computer words proposed as gismu. A few had no counterpart in any natural language - one person with a politico-moral agenda proposed a gismu for "space" - the personal quasi-territory that Americans refer to with the idiom "keep outta my space". Someone else, at least thinking of International language issues, proposed a gismu corresponding to krokodilo (or whatever the Esp-o word is to not use Esperanto when it is appropriate). The TLI community is still heavily biased towards adding gismu right and left. Every issue they have a "new words" column with many malglico lujvo and a few poorly thought out gismu. One other danger in such haphazard addition of gismu is that the recurrent arguments over gismu place structure would be noisier and more justifiable. I cna say that we have analyzed the WHOLE set of gismu for place struture consistency twice since the last gismu was added. If we were constantly adding gismu, i could not say this, of course, and such analysis would be getting harder and harder as the list grew longer. Yes there is saome fear that the language could stagnate if it is frozen while incomplete. But I don;t think that any incompleteness lies in the gismu list, or that this would be where we would face stagnation. I also think that the ability to add words with lujvo and fu'ivla argues that the language vocabulary will never truly stagnate. As for "software" - why should Latin have any more problem than English, which clearly used a two-part lujvo. Latin surely has a root for "soft" and another for "ware". So does Lojban, not that we would want to do a loan translation - such are not too Lojbanic, though Latin and most natlangs have no trouble with them. (For Lojban, something like commanded-system - I wouldn't even include the "computer" - would handle the concept nicely.) lojbab