When a Parent Becomes a Daughter

[Miguel Carrasquer Vidal]

JoatSimeon at aol.com wrote:

>>X99Lynx at aol.com writes:

>>E.g., -if Slavic loaned the word for wheel from Greek - or vice versa -
>>could we mistake it as being a gift to both from the reified parent - PIE?

>-- no.  The sound-changes would reveal the date of the borrowing.

This is an oversimplification.  It depends on whether there *had*
been any relevant sound changes in the intervening period, and in
general we can only establish *relative* dates for sound changes,
if we're lucky (many times not even that).  Sound changes can of
course not be dated absolutely if there are no written documents.

In the case of closely related languages or where bilingualism is
the norm, things are further complicated by people's awareness of
the sound laws themselves.  Basques still borrow Spanish words
ending in -(i)o'n as ending in -(i)oi, despite the fact that
Castilian final -e after /n/ had already dropped in the early
Middle Ages, and likewise the Basque rule to drop -n- between
vowels has stopped working many centuries ago (OSp. -one -> Bq.
-oe > -oi :: Sp. -o'n -> Bq. -oi).  If we're lucky we can catch a
mistake (e.g. when two sounds have merged in one language but not
in the other, and "folk etymological borrowing" assigns it to the
wrong prototype), but not in general.

In this case, of course, Slavic (*koles-) has not borrowed the
word from Greek (*kw(e)kwl-o-).  I wonder, though, whether
another word for wheel, *rot(H)o-, might not be a (pre-)Celtic
borrowing in the other IE lgs. that have it (Latin, Germanic,
Baltic, Indo-Iranian).  The root *ret(H)- "run", besides the word
for "wheel", does not have any semantic development (or e-Stufe
forms) outside of a bit in Baltic and Germanic, but especially in
Celtic.  On account of the *o, the word can't be Germanic or
Baltic (with the above caveats, but this is a merger *o > *a).
If the word is a borrowing from Celtic, we can also dispense with
the laryngeal.  Celtic, like Armenian and Germanic, probably had
started aspirating the IE tenues at an early stage (which would
account for the loss of *p in Celtic [and Armenian]).  A Celtic
*rotos ([rothos]) would have been borrowed as *rathas in
Indo-Iranian, and as there was no root *ret- (*rat-) in I-I,
there would have been no pressure to make the word conform to its
non-existent native cognates.

=======================
Miguel Carrasquer Vidal
mcv at wxs.nl