On 2008-12-04 10:39, Arnaud Fournet wrote:
> I'm afraid your Judeo-Christian background is too sensitive to "Decay
> Rules".
I'm neither Jewish nor Christian, so please don't impute such motives to
me. Everybody has got _a_ background, and you are no exception, but I
don't speculate on how it might bias _your_ thinking.
> You once said you read Martinet "Economie des Changements phonétiques".
> It's much against Entropy.
Entropy inevitably increases in thermodynamically isolated systems.
Language, treated as a physical process in which information is
transmitted, is not such a system, but apart from the cases when we
"pump" information into it, it tends (for purely physical reasons)
towards states in which energy becomes maximally decoherent. Ease of
articulation is all about entropy. That doesn't mean that new structures
cannot emerge -- they do emerge now and then to restore sufficient
linguistic complexity, but "phonetic erosion" takes _all the time_.
Sooner or later, newly coined extended forms become telescoped into
shorter and simpler structures, so new ones have to be formed, and so ad
infinitum.
> I looked at it,
> It follows the general rule of about all those "short words", like apo, epi
> etc,
> Are you suggesting that epi should be reconstructed as ep(H)i ?
Certainly not. In all these cases the aspiration is secondary and the
plain stop is etymological. The odd think about accented <oukHí> is the
use of aspiration outside its original sandhi context. It has made some
linguists speculate that we have encliting *-gHí lurking there. But
Ionic Greek has <oukí>, which cannot be so explained (neither can
prevocalic <ouk>), while <oukHí> _can_ be analogical (say, as an
innovative strong form based on <oukH> and generalised in other contexts
in the Attic dialect), so it's more likely to be secondary.
> Ou(k)(h)(i) is not alone to exhibit that kind of (de)aspiration.
> Sometimes, unetymological consonants can intrude.
> For example, popular French mouche à-z ailes bleues "a fly with blue wings".
And non-rhotic varieties of English have intrusive /r/. But they
wouldn't have it without having etymologically justified linking /r/
first. By the same token, cases of intrusive /z/ or /t/ in French are
analogical extensions of "legitimate" (historically motivated) linking
consonants. If you claim that Greek /k/ in <ouk> is intrusive, what is
its historical source? Greek has some instances of intrusive /n/, but no
cases of intrusive /k/, at least to my knowledge.