Kenneth Whistler wrote: > The problem here is in part the result of too easily > using the term "letter" here. > > These things are really a Bengali orthographic solution > to the problem of representing vowel sounds (in borrowed > words) that are alien to the "slots" of the basic > phonology, and which don't have obvious representations > using the basic vowel letters of the system.
Yes this is the same as the innovation used with the Oriya letter Wa. (A semi vowel.) > As Michael > suggested, the solution makes use of an existing conjunct > form of ya, in combination with other vowel forms, and then > provides a reading for them. The Oriya letter Wa is the same. It uses an existing conjunct form of Ba (or Va if you prefer) and combines it with vowel 'O' > In principle, at some point in the future, either the > phonology or the orthography or both could evolve to > the point where the entire constructs start to get handled > as basic orthographic units (or "letters") for Bengali, > but it isn't really the place of the Unicode Standard to > try to push that evolution, if there is a well-defined > way to represent the sequence using the characters > already in the standard. And the well-defined way is? > In some respect, the problem is similar to arbitrary > orthographic adaptation [...]. An > example which comes to mind is using "kl'" to try to > represent a lateral affricate, for example. The fact that the > lateral affricate might be a phonological unit in the other > language, and that it might even have a unitary letter > representation (e.g., U+019A barred-l) in some other > orthography, doesn't mean that if people start to represent > it "kl'" in the "English alphabet" that we then have to turn > around and encode a "kl'" character in Unicode. > The above seems to be an argument as to why the Oriya Letter Wa should not have been encoded. (It is after all a modern innovation used to represent the foreign sound of 'W'. Andy
