46 LANGUAGE OF THE NEW TESTAMENT.
Matt. i. 2-16. (Tov ~Bapa(3/3av in Luke xxiii. 18,
John xviii. 40, is used with dramatic fitness ; it is the
first time the readers have heard of Barabbas, but
the speakers and their hearer know all about him.)
There are however instances where the position,
or the use or omission, of the art. cannot be justified,
according to strict grammatical rule; or only by
supposing a far-fetched and improbable refinement
of sense to be intended. E.g. in John vi. 32 the
sense the only one giving a reason for the actual
order of clauses appears to be TOV IK TOV ovpavov
aprov. Still more decidedly in xii. 9, 12, it seems
impossible that 6 o^A-os TTO\VS can mean anything
different from 6 iroXvs o^Xos : to say (with Butt-
mann) that o^Xos TroXv's ranks as one word seems
arbitrary, and would prove too much. Ibid. viii. 44
we should, in better Greek, have had for the first
clause IK Trarpos TOV 8., if the sense be, as commonly
understood, " Ye are [born] of the devil as father."
And in the last clause, if 6 waTrjp OLVTOVbe a predicate,
co-ordinate with I^CVCTT^S, it should like it have been
anarthrous. But there is no doubt that, from a
purely grammatical point of view,* the easiest trans-
lation of the verse would be "Ye are of the father
of the devil ... he " (the devil at least as probably
as his father) "was a murderer . . . because his
* The verse was probably thus understood by the author of
the Acta Tlumwe (c. 32 in Tisch. Acta Apost. Apocrypha).
The arguments are incommensurable with each other, in
favour of this interpretation, that it was adopted by a Greek-
speaking Christian of the second century, and against it,
that any one who adopted it did so as harmonising with the
strange superstitions and heresies of that work. Origen in loc.
considers the constr. ambiguous : he does not speak of the
startling interpretation as though it were exclusively Gnostic.