On Thu, 9 Aug 2012 01:44:42 -0700 (PDT), Butch Malahide
<
fred....@gmail.com> wrote in
<
news:9e913fd6-921d-4c51...@z11g2000yqa.googlegroups.com>
in rec.arts.sf.written:
[...]
>> (Except that English doesn't allow <vv>,
As you point out, that was slightly overstated. All four of
your examples are modern colloquial or slang. The same goes
for <savvy>. <Navvy> 'construction worker' (originally from
<navigator>) is the one exception that I've found: it's
recorded from 1829 and despite its rather informal origin,
it seems rapidly to have become the normal term.
> Except after <i> as in <chivvy>, <divvy>, <skivvy>,
> <flivver>? What a strange rule, I wonder how that came
> about.
English treats the letter <v> oddly in other ways as well.
For example, it does not appear finally: <live>, not <liv>
or <livv>; <love>, not <luv> or <luvv> (except in eye
dialect spellings); <relative>, not <relativ>; and so on.
This, I think, is largely a consequence of the fact that Old
English and early Middle English simply did not have final
/v/. In Old English the letter <f> represented both /f/ and
/v/, and the correct allophone was entirely predictable from
the phonetic context. In particular, final <f> was /f/, but
intervocalic <f> was <v>. The verb <live>, for example, is
from OE <lifian>, which was pronounced with /v/; it was
still disyllabic <live(n)> in Middle English. Although it
was not needed to signal a long /i/, the final <e> was
retained in the spelling even after it ceased to be
pronounced. I'm not sure why this occurred specifically
with <v>, but I suspect that it had to do with the fact that
<u> and <v> were not treated as distinct letters (as opposed
to variant forms of a single letter) until somewhat later.
In many medieval and even Elizabethan hands the distribution
of <u> and <v> has nothing to do with pronunciation and
everything to do with position in the word, just as long <s>
is found everywhere except at the end of a word, where short
<s> is used.
I suspect that the same factors somehow underlie the
avoidance of <vv>, probably along with the fact that at one
time <uu> = <vv> = <w>.
And of course once a convention is well-established, it
tends to be self-perpetuating.
Brian