52 LANGUAGE OF THE NEW TESTAMENT.
But 6 Saravas, taking it as actually a pr. n. ? cannot
be right : perhaps however St. John remembers that
in the Hebrew Scriptures the word is still on the same
footing as AtaySoXos that it is only with the art. that
it is equivalent to a pr. n.
The primitive* pronominal use of the art. has
come a step nearer to extinction in the !N. T. than
in Attic. Not only is it confined to the phrases with
the particles //,eV and Se not, it is true, to the par-
ticular case where these are opposed to each other ;
for we have both ol //,eV answered by aAXot [8e] in
John vii. 12 (if not in Matt. xvi. 14), and ol 6V,
6 Se, 77 6V, more rarely, 01 /x,eV, in plain narrative in
Matt. ii. 5, 9, Luke i. 29, Acts i. 6, etc. : but it is
apparently, with pne exception, only used in the
nom. masc. and fern. sing, and pi., in the forms, in
fact, that begin with the aspirate, not with T. In
Eph. iv. 11 we have TOVS p\v . . . rovs Se . . . TOVS 8e
. . . but in Mark xii. 5 we should read ovs, and so
everywhere where neuters or oblique cases occur in
phrases of this type. And even in the nom. it is
only in the masc. sing, that we can tell (since the
accents of our modern text do not represent a primi-
tive tradition) whether the pron. is of the form
identical with the art. or with the rel. The latter,
after becoming almost extinct in the earlier Attic
(KOL 05 was rare, ^ 8' os apparently only colloquial),
becomes frequent from Demosthenes onwards with
[jLtv and Se in antithetical clauses.
tN
Os Se a7rKpiOrj
avTols, which is strongly attested in John v. 1 1 , would
* Of course the poetical archaism of Aratus, quoted in Acts
xvii. 28, is no instance of N, T. nor even of real Alexandrian
idiom,