>What these grammars did not
>provide is an explanation of this phenomena.
It is interesting. It certainly makes it easier to identify its antecedent
(which is occasionally not immediately obvious, at least to me), even though
it muddies its case usage within its own clause. Could this not be a
plausible explanation, that the author (probably unconsciously) is making the
connection with the antecedent by keeping the same case? I notice also that
the reverse occassionally happens, at least classically. Goodwin (1892,
#1035) notes:
The antecedent occasionally is assimilated to the case of the relative, when
this immediately follows; as ELEGON hOTI PANTWN hWN DEONTAI PEPRAGOTES EIEN,
they said that they had done all things which (PANTWN hWN) they needed,
X[enophon] H[ellenica] 1.4. THN OUSIAN hHN KATELIPE OU PLEIONOS ACIA ESTIN
H TETTARWN KAI DEKA TALANTWN, the estate which he left is not worth more than
fourteen talents, L[ysias] 19,47.
No one's use of language is perfect and even the most skilled sometimes use
it wrongly. Of course we often see authors making deliberate errors for
effect but it's hard to see how such assimilation (or attraction or whatever
you want to call it) would be deliberate or what effect it would be used for.
Blessings,
Rev. Craig R. Harmon.