free variation in pronunciation

[Lynne Murphy]

dInIs said:>
>
> Nothing can be said to be in free variation until extensive research
> on the probabilistic influences of phonological, morphological,
> syntactic, semantic, and pragmatic (including interactional,
> attitudinal, etc...) environments on the item in question has been
> carried out.

I take your point seriously.  I'm working from my own experience of having made
various hypotheses about my own use of the two pronunciations and having them
all disproved when someone's pointed out to me that I've used one where I've
said I use the other.  I'd love to see an in-depth study of this done, but I'd
want to see it done on the individual rather than language-community level
(i.e., consider several people, but consider their data separately), and the
fact that it's a pronunciation variation makes corpus studies hard to do for
it.

What I'm trying to react to is the claim that the two pronunciations of
neither/either (in free variation) flout the different forms-->different
meanings convention in language.  Obviously, for some people on the list
the two forms do have different social meanings, because one is more 'foreign'
and seemingly 'snootier'.  In my northeastern experience, there's more
opportunity for free variation, since one of the forms is not so 'otherly'.
I'm wondering whether such free variation is more tolerated in closed class
word categories, since the closed classes will not as easily allow for
semantic change (they allow for change, but not as readily as in the open
classes).  The other much discussed example of possible true synonymy in
English is another closed-class set, the pronouns ending in -one and -body.
As we've already discussed here, these two seem to be not so synonymous as
Jespersen inter alia claimed.  At the moment, the two pronunciations of
'either' and 'neither' are the closest things I have to examples of true
synonymy.

A lot of the things that people have offered to me as examples of free
variation do not meet the definition that I'm using--that they must be
freely variant in a single individual's idiolect.  If you say 'tomayto' and
I say 'tomahto', then the two forms are in complementary distribution across
idiolects (and probably dialects).  I, as a matter of fact, say 'tomayto'
and 'tomahto', but this still doesn't count because I say the first when
speaking to North Americans and the second when speaking to others--so it's
in socially/geographically complementary distribution.

So, while I agree with dInIs that it's too easy to claim that either/neither
are in free variation, even within my own idiolect, I am without the
wherewithall to disprove the claim that they are, and so (in order to
complete an argument regarding the nature of synonymy), am looking for
possible explanations and fellow examples in order to see whether the counter-
example that they present is a threat to the theory or not.

Does anyone have a student who wants a thesis topic?

Lynne