From cbmvax!uunet!cuvma.bitnet!LOJBAN Wed Jan 8 22:58:47 1992 Return-Path: Date: Wed Jan 8 22:58:47 1992 Message-Id: <9201082342.AA09341@relay1.UU.NET> Reply-To: Logical Language Group Sender: Lojban list From: Logical Language Group Subject: consonant clusters - response to S. Levy X-To: lojban@cuvmb.cc.columbia.edu To: John Cowan Status: RO From: LEVY%LENNY@VENUS.YCC.YALE.EDU >Why is it that Lojban has so many consonant clusters? The general form >of polysyllablic words (especially selbri) seems to be CVCCV. This is >especially surprising to me, since (1) There have been repeated >references to Lojban as a "vowel-rich" language, (2) There was some >mention of the language moving toward syllable timing, which would be >unusual in a language with so many syllable-final consonants >(syllable-timed languages tend to favor open syllables), and (3) If one >wants to use Lojban for spoken communication with computers (one of my >main interests), having syllable- final consonants only makes things >difficult. > >Given that Lojban has 6 vowels and 12 consonants, the number of possible >CV syllables is 6*12 = 72. Therefore, the number of possible >two-syllable words is 72*72, or 5184. > >Perhaps the goal is to distinguish selbri from sumti, since sumti seem >to be CV syllables for the most part. > >Any comments? You have the terms wrong, but that matters little. Yes, the goal is an unambiguous morphology such that there is a one-to-one correspondence between speech and text. Words in a stream must therefore break down only one way. You know it is a content word (brivla) by it having at least one consonant cluster and penultimate stress. cmavo, the function words, never have a consonant cluster. Names of course have the consonant final not permitted for the ther two. These distinctions are vital to the word resolution algorithm. Among brivla, the ones people normally see these days are the gismu roots which are all CVCCV or CCVCV. If CCVCV, you do have two vowel final syllables. If CVCCV, you may have one or two vowel-final syllables depending on whether the CC medial is a permissible initial or not. Thus in content words perhaps 3/4 of syllables are vowel final, and function words are all 1 syllable all vowel final. When syllables are consonant final, you will find that a large portion of them end in nasals and liquids. The letter 'n' is by far the most frequent consonant in Lojban words but yet is found rarely in word initial - whereas it is 25% of 3rd position consonant (and the other nasals/liquids cover another 25%). My understanding is that such syllables are nearly as open as vowel-finals. Compound words that will eventually make up the vast majority of the vocabulary are not that numerous or well-known yet. I haven't done an analysis but I think they will consist even more than 3/4 vowel-final syllables. They also have diphthongs and divowels that cannot be found in the gismu. Thus they are more vowel rich, and the cmavo are most vowel rich. I don't know anything about the phonology of stress-timing vs syllable-timing and any correlation with language structures. People who actually know something like you are welcome to correct me. I had observed that most languages listed as syllable-timed tended to have many short syllables in words rather than long, heavy syllables. Lojban avoids the latter. Syllable timing also looked easier as a method for Lojban poetry and song, something many people seem to be interested in these days. Trying to write Lojban according to any of the stress-timed poetry rhythms found in English runs afoul of the greater numbers of syllables, and Lojban's rigid stress pronunciation that leads to varying numbers of syllables between stresses. I also noted some changes in my speaking style when I use LogFlash, the only time when I say Lojban words at a rapid enough speed to approximate a fluent accent. Time will of course tell. lojbab