On 1/9/08, komfo,amonan <[EMAIL PROTECTED]> wrote: > > {ndj} {mz} and {kv} cannot be reliably distinguished aurally from {nj} {mbz} > and {kf}/{gv}. I believe that's why those clusters shake out like that.
If this was based on some serious study reflecting human speech recognition capabilities, those restrictions would be easier to swallow, but since they seem to be based on the arbitrary impressions of someone or other, they are really a pain. I can accept the {kv} restriction as part of the general proscription of unvoiced-voiced pairs. At least it's a general enough principle. The eventual confusion of {mz} with {mbz} cannot really be the issue because {mbz} cannot appear in lujvo. Isn't {mt} just as likely to be confused with {mpt}, for example? Why single out {mz}? (The best theory I heard about this one is that it was introduced so that JCB, James Cook Brown, would not have an easy time lojbanizing his name.) Are ndj/nj really more easily confused than rdj/rj or ldj/lj? Are there any studies that show that? I wouldn't be surprized if even "s" and "f" are more easily confused than "ndj" and "nj". mu'o mi'e xorxes