From phma@webjockey.net Thu Feb 26 14:46:17 2004 Return-Path: X-Sender: phma@ixazon.dynip.com X-Apparently-To: lojban@yahoogroups.com Received: (qmail 43065 invoked from network); 26 Feb 2004 22:46:16 -0000 Received: from unknown (66.218.66.216) by m2.grp.scd.yahoo.com with QMQP; 26 Feb 2004 22:46:16 -0000 Received: from unknown (HELO neofelis.ixazon.lan) (208.150.110.21) by mta1.grp.scd.yahoo.com with SMTP; 26 Feb 2004 22:46:15 -0000 Received: by neofelis.ixazon.lan (Postfix, from userid 500) id 34BF33C497; Thu, 26 Feb 2004 17:46:13 -0500 (EST) Content-Type: text/plain; charset="iso-8859-1" To: lojban@yahoogroups.com Date: Thu, 26 Feb 2004 17:46:08 -0500 X-Mailer: KMail [version 1.2] References: In-Reply-To: X-Spamtrap: fesmri@ixazon.dynip.com MIME-Version: 1.0 Message-Id: <04022617460813.02957@neofelis> Content-Transfer-Encoding: 8bit Sender: phma@ixazon.dynip.com X-eGroups-Remote-IP: 208.150.110.21 From: Pierre Abbat Subject: Re: [lojban] Lojban question X-Yahoo-Group-Post: member; u=92712300 X-Yahoo-Message-Num: 21622 On Thursday 26 February 2004 17:16, stewartberntson@yahoo.com wrote: > Is anything being done on a speech-recognition grammer for Lojban? > Since the language designed for understandability in high-noise > environments, it could be perfect for use as a remote speech > interface (i.e. computer has a problem, calls the owner's cell > (using SAPI likely), the owner could tell the computer what to do in > lojban, which would reduce the errors in recognition with similar > words (for example, reboot and refresh or something like that) I don't think Lojban actually is more noise-tolerant than other languages. The gismu list was designed so that two gismu would not differ in the final vowel (except for brod*) or by one similar-sounding consonant in the same position, but that didn't keep me from mishearing {kansa} as {tamca} on a jbofongri call where the bandwidth was low. {ractu} and {ratcu} both designate mammals and can easily be confused. The Queen and the Duchess in Alice respectively noltruni'u and noltroni'u. {pixra} "picture" can be confused with {spixra} "mosaic". {ko'a} through {ko'u} ("he", "she", "it", but assigned arbitrarily, not by gender) differ only in the second vowel, which is usually unstressed (if it's stressed and followed by a brivla of more than two syllables, the word split can be misheard unless a pause is inserted). Substitution of /'/ for /x/ or one consonant for another which changes the initial-cluster status of a consonant pair can move word boundaries. And so on. phma