On 2006-06-14 16:45, Jens Elmegård Rasmussen wrote:
> That will be the phonetic rationale: Since Meillet's Law is nothing
> but loss of accent, making the word an enklinomenon, failure of
> Meillet's law to work is retention of the accent, in this case
> obviously due to the additional weight caused by the clustering
> consonants. Clusters that did not annull Meillet's Law, such as sT-
> clusters, were apparently not heavy enough in real pronunciation.
That's a very common phenomenon. sT-clusters often behave
phonotactically like single consonants in IE languages, and the /s/ is
easily attracted, together with the stop, into the onset of the
following syllable. In that case of course the cluster fails to make the
preceding syllable closed. The "open syllable conspiracy" in Slavic did
not result in the simplification of sT- or even sTR-clusters, since they
could always be parsed as syllable-initial. Note the stress pattern of
nouns like <órchestra> or <prótestant> in the Latinate sublexicon of
English. If the penult syllable counted as heavy, it would attract
stress (cf. <agénda>, <deféndant>). English, by the way, is not entirely
consístent and hesitates between the <órchestra> and the <asbéstos> types.
Piotr