History: Redo the words

Source of version: 1 (current)

According to the refgram, the gismu were designed in a highly illogical process. They mishmashed 6 of the largest spoken languages to generate most of the gismu. So we have the worst of both worlds--essentially very few words can be immediately recognized, and we get irregularities rampant through all the rafsi and cmavo. I propose they all be regenerated so as to be completely random yet subject to systematic criteria. Such criteria could establish a regular system for rafsi assignment. In the end, it should be possible to generate completely new words automatically--so essentially you'll be settling on a seed number and a random number function. Parsers wouldn't have to store the whole vocabulary. --((tinkit)) ((Gua\spi)) went through a random-word stage, when it was called "-gzn!gvr". Taste tests judged the random words to be unacceptably ugly: they have inhuman phonotactics. I don't think -gua!spi's phonotactics are any too wonderful (e.g. ''fkar'' 'car'). --((John Cowan)) They don't store the whole vocabulary as it is. The __lexer__ (not the parser) has to be able to map cmavo to their se cmavo, but all other word classes can be identified based merely on their morphological features. If you want to be able to look up semantic info, though, then once you've parsed it you'll have to have the completely vocabulary one way or the other. (Certainly you won't claim that you can use a PRNG to generate the meanings of the words, also.) --((Jay Kominek|jay)) Why don't you have a go at redoing some or all of the words? Obviously nobody's going to learn or use a new set of words, but some of us (e.g. me) would be interested to see what you come up with. --((And)) ---- ((tinkit)) may very well have left by now, but I think I know the gist of what he would have come up with. Behold, the __completely regular__ Lojban vocabulary: gismu rafsi cmavo meaning ----- ----- ----- ------- bilna bi'a bi'a "zero" xalnu xa'u xa'u "one" Of course, you can express anything with these. As long as you're talking to a computer. ''zo'o mi'e ((rab.spir))'' ---- Why waste disk space on this stuff? Can anyone really believe that this "proposal" is going to be remembered in two weeks? I've commented before on how there wasn't any sense in worrying about wasting disk space. But I feel as though someone is out to prove me wrong. I'd far rather be using the disk space to store bad polka MP3s than anymore of thinkit's one-person "movements". --((Jay Kominek|jay)) ''This isn't as small as some, like the hex thing. I would love to see rafsi and selma'o reformed, I doubt i'm the only one who dislikes the current rafsi besides tinkit, though I love the gismu as they are. - mi'e. ((kreig.daniyl.))'' This wiki community is very censorious! Why can't tinkit be accepted as a dissenting member of the lojban community? --((And)) Some of us who use the language, and who have invested into the corpus, are getting irritated by these silly tinkering attempts, none of which ever amount to anything. Each one makes me more conservative. Throw out all our gismu, indeed! --xod Agreed. What would the point be of writing if the next day you could discover that your innocuous poem had become a love letter? - mi'e. ((kreig.daniyl.)) Nobody is censoring anything, but when someone wants to come along and throw out a big chunk of the memorization required to learn the language because he thinks they're irregular (for some completely undefined value of "irregular"), when in fact they were created by an algorithmic process (and he wants to make them __COMPLETELY RANDOM__ (how does that jive with his desire to make them more regular?)), that person is putting on the loony hat and running around town naked. --((Jay Kominek|jay)) ---- I have a question. Why do we need ''te panci'' and ''sumne'' and ''vrusi''? Or ''cecla'' and ''renro'' and ''se danti''? The gismu list isn't advertised as being a minimal set of primitives, merely a complete (and convinent) set of primitives. --((Jay Kominek|jay))