reservwah (was: paraplegic)

[Laurence Horn]

At 1:52 PM -0400 4/16/01, Douglas G. Wilson wrote:
>Webster's Third attributes this pronunciation not to misguided Francophilic
>hypercorrection but rather to "r-dissimilation" [another example of which
>(assuming a rhotic dialect): "governor" /gVv at rn@r/ > /gVv at n@r/].
>
>This can go two ways here, apparently: for example /rEz at rvwar/ > /rEz at vwar/
>OR > /rEz at rvwa/, says the book; i.e., either the second or the third "r"
>can be silenced. I remain agnostic.
>
>-- Doug Wilson

Now that I think of it, I drop that medial /r/ myself--never the
final one--and I'm usually rhotic, and I drop it in "governor" too,
but I wonder if this is really dissimilation, as opposed to a
least-effort reduction.  The non-pronunciation of the first -n- in
"government" isn't due to dissimilation, I'd wager, even though there
is another -n- in the word.  But in any case, the loss/omission of
the FINAL r in "reservoir" is NOT an instance of dissimilation, but
rather of "misguided Francophilic hypercorrection", or Mandel's rule.
If this WERE dissimilation, and paralleled by the same process in
"gove(r)nor", you'd expect to get "governuh" as an alternate option
in rhotic dialects (alongside "rese(r)vwah", and I'm willing to bet
you John Rowland (and I'll throw in George Pataki) that you never do.
(Of course you get "gov'nor", "governuh", and "guv'nuh", but only as
variable outputs in non-(fully-)rhotic dialects.)  I respect Doug's
agnosticism with respect to Webster 3, but on this point I'm an
atheist.

larry