[Corpora-List] Phonetic corpora as a tool for language classification.

[John F. Sowa]

JFS>> French and Italian, for example, are very close in vocabulary,
 >> but farther apart phonetically because the people in Gaul who learned
 >> Latin were native speakers of language(s) with very different
 >> phonological patterns.

MM> I vaguely recall an article (maybe a book review in Language)
 > a couple decades ago that started out by saying that the above is
 > what one might have expected, but....

I didn't want to suggest that was the only influence.  My main point was
to suggest more likely hypotheses than the following:

YT> Was Latin sound picture preserved in Moldavian better?

MM> BTW, if the above is true, then Italian (particularly the Italian
 > dialects spoken in central Italy) might be expected to be the
 > closest to Latin, because apart from the invaders at the end of
 > the Roman Empire, the people there were *native* Latin speakers.

Yes.  Italian diverged by simplifying many consonant clusters of
Latin.  In fact, most of the Romance languages have later technical
terms borrowed directly from Latin that have a greater phonological
similarity to Latin than the words that evolved from proto-Romance.

Any measure of phonological distance should distinguish words
that were present in the earliest versions of the language from
borrowed words from Latin and borrowed words from other sources.

John Sowa


_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora