En réponse à Andreas Johansson <[log in to unmask]>: > > You were, but certainly its a drawback for a transliteration scheme not > to keep > the value of a grapheme when it occurs in both the source and target > character > set? Well, in this case I was influenced by KPA, which does the same, and nobody finds it a drawback :)) . (In this case, the benefits of turning "!" into a diacritic is > probably > greater than the drawback, but that's a question of weighing the pros > and cons > in a specific case, not something of principle.) > But C-IPA is all about principles ;))) . That's what's making it different from other transliteration schemes. If you simply want to weigh the pros and cons, X- SAMPA is just good enough ;)) . > > Yep, but surely it doesn't hurt if other signs to so too, where > practical? The > primary stress sign is a case in point. > True, and this one is taken as ' indeed, just like secondary stress is taken as , . But this has no character of necessity here, just personal dislike against X-SAMPA " and % ;))) . ? was another example, but like Joe suggested, h| works as well and it frees ?, which helps me with a little problem I had. Now C-IPA has distinct diacritics for implosives and ejectives (Kirshenbaum doesn't make a distinction, but all other ASCII-IPA schemes and IPA itself do, so I felt like I had to do so too, whether it was meaningful or not - is it? Are there languages with voiced ejectives or voiceless implosives? -). In some way I find it a bit embarassing not to have ? for the glottal stop, but until I find a better diacritic for implosives, it will have to do. For some reason, I mind less that ! doesn't have the meaning it has in IPA than ? ;))) . > > There's no reason it'd make everything much longer; I did suggest > removing the > C-IPA characteristics as a class, only the particular ones [! |]. > That would break the unity of the C-IPA diacritics and lose too much modularity :(( . > > Naturally not in the base form of C-IPA, just like [th] isn't acceptable > for > aspirated "t" in base IPA. I was merely pointing out how one could > simply the > system used for a specific language. > That's what modularity of C-IPA is for. To allow for language-specific (and person-specific ;)) ) schemes whithout "breaking" the rules. So you can do as you say indeed. You just have to always make it clear by beginning with a series of X := Y lines, with X being the character sequence you apply the meaning of the pure C-IPA Y sequence. In your case, you'd probably have two lines: redef ! := t! redef | := t+! The "redef" makes it clear that ! and |'s original meanings are overridden by your definition (a line defining a character which has not been defined before must normally begin with a "def", but you may omit it since there's no ambiguity). For a click-heavy language, this is a very good idea, as long as you *never* use ! and | for their original meaning. Still, I recommend keeping redefinitions to a minimum. Definitions are fine, but redefinitions play havoc with the C-IPA kernel (and now I'm really looking like a TeX geek ;)))) ) by making things difficult to read. Unless you have a very good reason to apply them (Describing !Xu~ is good enough ;))) ), you shouldn't. After all, with the correct redefinitions, you can transform C-IPA into X-SAMPA, Kirshenbaum or KPA or whatever ASCII-IPA you want! ;)))) That's not quite what C-IPA is for ;) . Christophe. http://rainbow.conlang.free.fr Take your life as a movie: do not let anybody else play the leading role.