Traditionally, Semitic has 29 consonants, or all of the Arabic letters plus one,
the lateral fricative of Hebrew _s'in_.

However, I've read proposals of larger inventories. And after doing some
comparison of my own, I'm inclined to support more consonants. The "litmus
test" is how sibilants in Arabic and Hebrew are represented.

Arabic /s/ and Hebrew /s^/ are traditionally reconstructed /s^/, but Starostin
on his Babel website has /s/ ultimately.
Arabic /s/ and Hebrew /s/ ultimately point to /ts/, again according to
Starostin.
Arabic /s^/ and Hebrew /s'/ of course come from /s'/ which had the value of [L],
but an original consonant /tL/ is proposed.
Arabic /s^/ and Hebrew /s^/ lead to another /s'/, but this time a pure fricative
/L/, not an affricate.
Finally, there is the matter of the causative/reflexive prefix and definite
article, which is ?- in Arabic and h- in Hebrew. But Akkadian has /s^/. This
has been reckoned as /s^/ in Proto-Semitic.

My other question is about the possibility of an "emphatic" /p/ in
Proto-Semitic. As it is in IE, this is a quite rare consonant and a dubious one
at that. In Semitic, I've found some interesting cognate sets; I'll give a
couple here:

"branch": Hebrew se3i:p, Arabic s^u3ubat. p~b in other words. Two cognates
make a pretty weak argument however.

"to drip": Hebrew nTp, Arabic nTf, Ge'ez nTb. p~f~b in this case, instead of
the expected p~f~f.

"honey": Akkadian dis^pu, Arabic dibs, Syriac devs^a, Hebrew dbas^. Here,
p~b~b~b. (But the Akkadian data might result from devoicing assimilation from
the preceding s^, so this one's very doubtful.)

"iron": Arabic has both frzl and brzl; I'm suspecting the latter could be a loan
from Hebrew barzel though.

I seriously need more examples; I just scanned a homemade lexicon of Akkadian,
Arabic, Ge'ez, Hebrew and Syriac. This data could point in the direction of a
Proto-Semitic /p'/, or there may be a better explanation.

So what do you think?

~DaW~


_________________________________________________________
Do You Yahoo!?
Get your free @... address at http://mail.yahoo.com