To be sure, Bomhard admits -(V)C- suffixes
as root extensions, which greatly enlarges the encoding potential of CVC-. His
reconstructed _lexemes_ (as opposed to abstract "roots") are often of the form
CVC-(V)C-. You remarked that despite the number of posited protosegments there
is a lot of homophony in Bomhard's Nostratic. I would think homophony is a
near-constant element of lexical systems, no matter what the phonotactic
structure is. Note how much homophony exists in English despite its large
phoneme inventory and relatively complex phonotactics. One could easily compile
a list of _hundreds_ of homonymic doublets (and many triplets as well) involving
not loanwords (as Torsten would have it) but good Anglo-Saxon words that have
became homophones because of perfectly regular mergers (meat -- meet -- mete,
ewe -- yew -- you, wright -- right -- write, so -- soe -- sew, bare -- bear[1]
-- bear[2], roe -- row[1] -- row[2], weather -- wether -- whether,
etc.)
Piotr
----- Original Message -----
Sent: Monday, January 29, 2001 5:03 AM
Subject: Re: [tied] IS's "regular roots"
I was discussing the case of a hypothetical
Bomhard-like
reconstruction, but with the phonological inventory stripped,
and no
mention of tones, etc. Ehret's "Proto-Afrasian" is also CVC, but
at
least he posits 2 por 3 tones (in itself not unreasonable:
Chadic,
Cushitic and Omotic contain tonal languages). There are 42
consonants
and 10 vowels in his system.