--- In
cybalist@yahoogroups.com, "P&G" <petegray@...> wrote:
> Hebrew shows minimal lexical pairs, e.g.:
> yo:m = day; ya:m = sea
> (ayin)e:z = goat; (ayin)o:z = strength
> ?el = to ?e:l = god
> etc
> So I think the starting point for this discussion is misguided.
> Vowels are used lexically (and morphologically) in at least one
> Semitic language.
But that was not the starting point of the discussion. The question
was whether there is at least one Semitic language where vowels are
*not* used lexically. This is (almost) the case in Classical
Arabic, and reportedly in Southern Peripheral Semitic languages.
Certainly in the case of the verbs, the root of an Arabic verb is
given by its consonantal "skeleton". Vowels are inserted and
affixed purely as a function of the morphology.
The claim is somewhat harder to make in the case of (pro)nouns and
adjectives, where there are a number of vocalic patterns (for the
singular and the "broken" plural) to which the forms may belong, and
which are not always predictable based on the class (pronoun, noun
or adjective, etc.) or function of the word.
This is in principle completely parallel to the situation as
reconstructed for PIE. The vocalism of verbs is largely predictable
based on the morphology (e-grade/zero-grade in the present, o-grade
in the perfect sg., zero-grade in the middle, etc.), while the
vocalism of (pro)nouns and adjectives is somewhat less predictable
in general.