GREEK PREHISTORY AND IE (EVIDENCE?)

[Hans Holm]

SG>not against Rosenfelder, whose attempt I've not found easily
SG>accessible, mainly because the link above is dead.

.. < The link via the homepage seems to be disconnected, indeed; but, the
URLdoes work now leaving out the 'l': "www.zompist.com/chance.htm";-))

SG>What most of the books and article I've seen *don't* address, however, is
SG>the question how "resemblances" are to be defined in the first place

.. < Rosenfelder tries to define them, I think. Another source is e.g.
Ringe in Diachronica 1992, passim, in his nearly famous
Greenberg-controversy; in abridged form to be found in Larry's textbook.
----------
SG>Your resemblance may not at all be mine,
.. < I never maintained any resemblance (as common heritage, you mean).
----------
SG>as if lexicon had *anything* to do with lg. classification,
.. < correct, regarding the percentage calculations à la Dyen.  Only by
understanding and applying the properties of the 'hypergeometric' and,
using a complete etymological dictionary, it is possible to compute
split-off bases between any two languages. Not more, and nothing less. And
the brain of homo sapiens has not been constructed to grasp the
hypergeometric ad hoc.
---------
SG>the calculation of probabilities
..<all probability calculations have and will be wrong because they
neglect the fact that two closely related languages may have lost
extremely different amounts of features, or even may have died out, in the
worst case. Again: Only the hypergeometric can solve this problem, not
percentages nor their probabilities.

Regards
Hans J. Holm
D-30629 Hannover