Six, -ts- > -ks-

It seemed to me that the missing link in Glen's argument to get the
word for "six" from Semitic to IE is that he has a hard time getting
to a -ks ending in the PIE for "six", since there is only a -t- in
the Semitic word. Here's something I found in Trask that might help:

Basque has <sei> "six". It can't have been loaned from Latin or the
surrounding Romance, since it hasn't gotten around to deleting
everything after the vowel. So it might be an earlier loan. Now I'll
make a bold assumption that some time in the grey past, after the
word was borrowed, all final clusters eg. with unvoiced stops were
deleted; this is safe since we're so far back in time there's nothing
to disprove it (Basque is an isolate), and, anyway, if we want
Basque "six" to be borrowed from Semitic as in so many other language
groups, we have to assume such a deletion.

BTW some linguists have proposed the loss of such clusters initially
in Basque.

So, by this assumption "six" ended in -ts in Basque at some remote
time. Here's the interesting part: it turns out that Basque -tz-
corresponds to Aquitanian <-x->, presumably -ks- (BTW for some odd
reason the French Basques still spell -x- for -tz-)eg.

haritz, Arixo "oak"
-belex, beltz "black"
Berhax-, beratz "soft, benign"
Bihox-, Bihos-, bihotz "heart"


Therefore, if Very Old Basque "six" ended in -ts-, in Aquitanian it
must have ended in -ks-.

Next we might want to assume that Vasconic (some hypothetical group
comprising Basque and Aquitanian stretched up further on the Western
European coast, cf. all the rivers named <Itter>; < Basque
<iturri> "spring"?) passed the word on to Western IE; the rest of IE
would have to be taken care of otherwise, but note the irregularity
of the "six" word in the east, suggesting another transmission route.


Torsten