Even if we accepted that the
p seen in
erepto:was original (I mean, if it was not derived from Gk.
ph+j, i.e.
*erephjo > erepto1), the word
roof would be just a cognate of Gk.
erepto:; in no case would it be a loan from Greek. Still I insist that
p is not original in
erepto:, as witnessed by its obvious cognates with
-ph- (which would be equivalent to Germanic
b).
On the other hand, the semantic aspect is of no importance at all, since many superficially similar words have identical meaning in several, even unrelated, languages, cf. Malay
mata, Greek
mati, both denoting an 'eye'. However, in the Malay word the root is
-ta (
ma- is an affix), while in the Greek word the root is
*oqw- > *op- (> *op-ma > omma > ommation > ommatin > mati).
Do you see the point?
And the last and most convincing argument is that OE shows the form
hrof. The initial Germanic
h should correspond to a Greek
k (cf. English
horn, Gk.
keras, L.
cornu or English
row <
hro-, Gk.
kruos, krustallos L.
cruor, crudus etc.)
1An original -j- following a labial consonant is regularly turned into -t- in Greek, cf. *kop-j-o: > kopto: (= I cut).P.S. I checked
http://ewonago.blogspot.com and I observed that the majority of "English" words of "no apparent Greek" origin mentioned there are mere cognates between Latin and Greek. Please, do not confuse a
cognate with a
loan/derivative and their ilk. This is an error made by those not acquainted with linguistic rules and patterns. Of course there are English words of Greek origin, but they entered English mostly through Latin. But there are many more Latin-derived or 'true' English (Germanic) words similar to Greek ones just because of the common origin of Greek, English, Latin etc.
