Return-Path: Message-Id: From: cbmvax!uunet!count.tct.com!chip (Chip Salzenberg) Subject: Gismu uniqueness To: lojban-list@snark.thyrsus.com (Lojban mailing list) Date: Fri, 10 May 91 8:43:17 EDT X-Mailer: ELM [version 2.4dev PL19] Status: RO X-From-Space-Date: Fri May 10 14:56:45 1991 X-From-Space-Address: cbmvax!uunet!pdn!tct!count.tct.com!chip First, as an aside, I'd like to suggest that the word lists be sent to the mailing list. That's what a mailing list is for, right? Now, the point. I wonder how many unique gismu we can have, given that a gismu stripped of its final vowell is always a valid rafsi. For example, consider "brivla", which is equivalent to "bridyvalsi". If another gismu besides "bridi" matched the pattern "bridV", then "bridyvalsi" would be ambiguous, and ambiguity is the one thing la lojban. can't allow. So if no gismu like "bridV" other than "bridi" can be created, then why not have all gismu end with the same vowell, say "o", thus freeing up "CVVCa", "CVVCe", etc. for other uses? -- Chip Salzenberg at Teltronics/TCT ,