Re: Lenition or Elision or What?

Mark Reed wrote:
> On 12/9/05, R A Brown <ray@...> wrote:
> > For similar reasons, Mark's "approximantification" is not accurate
> > either, since there is no approximant. There is merely the labialization
> > or palatalization of the [m] (and presumably of any other consonant in
> > that position).
>
> True.  I read it the same way you did - as [mi] -> [mj], rather than
> [mi] -> [m_j].
> (Underscore?  What underscore?)  And I don't even have the
> bunged-up-sinuses excuse.
Me too:-((  But actually, is there any great difference between [mj] and
[m_j]? I can see that palatalization, in a language like Russian, is indeed
a feature of the consonant; also I can see that there's a difference between
(non-labial) [Cj] vs. [C_j]. But in Charlie's example it's a case of (1) the
C+i+V..-  sequence reducing to C+j+V.. (bearing in mind that there are
morpheme boundaries), which then (2) just happens to be realized on the
surface as [C_jV..]-- I feel that information is being lost by shifting the
present-tense morpheme //i// to a [+pal.] feature of the consonant (Russian
palatalization is creates no morphemic change AFAIK). _Phonemically_ the
example word must be /m+i+a.../, and I'd bet that speakers could vary
between [CjV..] ~[C_jV...]. Presumably in case of /m+i+C.../ the phonetic
outcome must be [miC...], not [m_j(V?)C...].

Altogether it doesn't strike me as any more phenomenal than fast-speech
Spanish "mi amigo" > {mja'miGo] or "su amigo" > [swa'miGo].

And altogether, I think And's term fits best here.