[YG Conlang Archives] > [westasianconlangs group] > messages [Date Index] [Thread Index] >
--- In westasianconlangs@yahoogroups.com, "Isaac Penzev" <isaacp@u...> wrote: > Ktebe David J. Peterson: > > > > Anthony wrote: > > << > > How many phonemes must a language which uses radicals possess in order > > to form sufficient triradicals? > > >> > > > > Well, if you admit for the possibility of any kind of root (including > > three of the same consonant in a row [Arabic has at least one: /j j j/, > > which forms the verb "to use the letter yaa"]), then it's simply a > > matter of multiplying it out. Let's say your low end is 20: > > > > 20 x 20 x 20 = 8,000 > > > > That's a fair number of roots, and quite possibly "enough". However, > > rules on forming triradicals may cut that number down, so you may > > want to increase the number of phonemes, and, with each phoneme > > you add, the number increases significantly. > > > > How many phonemes were you thinking having? > > I merely second David's opinion. > A typical Semitic lang (the only kind of natlangs that use this kind of > morphology AFAIK) uses 20-30 consonantal phonemes, e.g. Hebrew uses 23. > There are some restriction wrt phoneme combinations, but e.g. the whole > Tanakh (s.c. "Old Testament") uses ca. 1,700 roots, which is far less than > 12,167 possible combinations, but quite enough to express rather complex > ideas. The structure of the words of the language are CV1RV1V2C or CV1RV2V1C, where C= consonant, V= vowel, and R= root, which may be C, CC, or CCC. Although there is no assimilation, the clustering of consonants imposes practical phonological constraints. > > -- Yitzik