From cbmvax!uunet!cuvma.bitnet!LOJBAN Tue Oct 22 15:00:34 1991 Return-Path: Date: Tue Oct 22 15:00:34 1991 Message-Id: <9110221528.AA11390@relay1.UU.NET> Reply-To: Logical Language Group Sender: Lojban list From: Logical Language Group Subject: response to And Rosta X-To: lojban@cuvmb.cc.columbia.edu To: John Cowan , Ken Taylor Status: RO 1) I don;t feel that any criticism of Lojban is out of bounds. I would rather hear criticism from friends than later from people who are hostile. And any argument that And or someone else is likely to come up with now, will either be explained away in our materials because he pointed it out to us, or we will have to deal with it later. 2) On the other hand, while we will isten to criticism, in most cases it is impractical to change the language if a criticism is valid. We have committed to finishing the language in the very short term, and only an unthinkably major idiocy at this point might cause us to - say - redo the phonology and morphology. I see And's comments on the morphology NOT as a true change but another way of looking at the existing design - that there is potentially a buffer vowel between all clusters such that there are no consonant clusters. But even if so, I don;t speak the language this way noramlly, and don;t expectt to, and I don;t see it as all that useful a concept in teaching the language. 3) The exact buffer sound to be used should be dependent on the speaker, the listener, and the environment. That is why we don't specify the sound other than to say it is a non-Lojban vowel. If I am talking to And, and he is hearing my [I] as "i", then I must shade that sound more towards [y]. This is pragmatic phonology. We've had to change our explanation of Lojban 'e' as well. We used to say "Romance 'e'", but what most American's say when they try a Romance 'e' is indistinguishable to other Americans from "ei". So we instead now tell people to use the sound of 'e' in "bet". 4) Re kando/kandi, the final vowel is significant in the recognition score algorithm that generated the words. "kandi" for example got score from "the 'di' from "dim", because we mapped the Englsih [I] to Lojban i for recognition algorithm purposes. All of the cultural words were by dictum made to end with 'o',because they were not developed using the recognition score formula, but rather when possible, from the local name for the culture/ language. The purpose of these words is to assign them short rafsi, and we have a surplus in the letter 'o' which was unintentionally deemphasized by our word-making phonetic mappings. If we could give a major culture a gismu with a rafsi, such that the gismu did not conflict and the rafsi was unassigned, we did so, sometimes having to fiddle the word around as with "kadno" and "sralo" to get either the rafsi or the gismu to work. In some cases, it was impossible - we could do nothing with "Bharat" for India, and chose to fall back on "xindo". Likewise, for Japanese, from Nippon, there is no 'obvious word', and the best anagram "nipno" is something no Japanese is likely to pronounce. We decided that we needed an nC consonant cluster, and ended up taking the 'j' from Japan arbitrarily, to get ponjo. This also moved the word out from the 'n' initials, which were at that point very crowded in rafsi. In short, lots of tradeoffs, some good, most mediocre. The bottom line was that we needed a set of words that we weren't going to change again, and changes in the gismu list, now baselined, are very difficult to win favor for. There has been exactly one such change in 3 years, no deletions, and less than 2 dozen additions. The community will not tolerate even this rate of{ change to continue - and there have been no proposals even considered since June. At this point, we see the language design as done, with only some fine tuning, generally in areas that are ill-defined. This doesn't mean that suggestions are not welcome, but that at best they will usually cause us to sharpen our explanation of the relevant design feature so we can say WHY we did what we did, not to provoke a change. A final factor here is that the language is now spoken by a small community, and their usage is the most conservative force there is. Just as Esperanto would not easily change for Ido, those of us who can finally converse in a version of Loglan after 35 years are not willing to put up with changes that make the language 'better' on a purely theoretical basis when we very pragmatically can actually USE the language that we have. Enough soapbox. lojbab