I wrote:
<<To say that the tie is not "necessary" -- that it's not 100 out of 100 --
is not scientifically relevant in these instances, because the correlation is
something like 95 out of 100 and way beyond the levels of serious statistical
doubt. And this only means that the rare exceptions prove the rule.>>
"Brian M. Scott" writes:
<<Does this have a point?>>
Of course it has a point. You just choose to omit it from your reply. Here
it is again:
I also wrote:
<<On the other hand, there are other cases where correlating material culture
and language causes serious problems, even in historical contexts....
So the real issue is whether we can recognize when language correlates with
archaeology and when it does not. The farther back we go, obviously the more
difficult the discernment becomes.>>
"Brian M. Scott" writes:
<<Of course if you choose an example in which the correlation is known to be
good, you'll find that the correlation is good.>>
So, you'd prefer we only look at the cases where the correlation is bad. And
disregard the cases where your statement is demonstrably incorrect?
I also wrote:
<<It should also be noted that if there is no correlation between archaeology
and language, the whole Pontic-PIE theory loses all foundation.
Paleolinguistics is entirely linked to archaeology. You have no dates or
locations for horses, wheels, graves, beeches, beavers, etc., without
archaeology.>>
"Brian M. Scott" replied:
<<And this is of course irrelevant, as it deals with a completely different
type of relationship between linguistics and archaeology.>>
Yes. And it is much more dubious. Not only does paleolinguistics seek to
correlate archaeology with linguistics, it actually attempts to correlate
particular words in unrecorded languages with archaeology that is thousands
of years older than any direct evidence of such languages. This is
specifically where common sense would say any reasonable correlation between
language and archaeology will be at its very worse. And of course where it
will also be scientifically unverifiable.
Steve Long