196 LANGUAGE OF THE NEW TESTAMENT.
determining the constr. in this point : compare the
two clauses with o^Xos in Mark iv. 1, John vi. 22
(best text), 24, xii. 9, 12. So it affects the gender
and number of a ptcp. standing as a secondary predi-
cate :
compare Acts xxi. 3G (true text) with Luke
xxiii. 1. An adj. or ptcp. forming a primary pre-
dicate would naturally be sing. (Luke i. 10, 21) :
yet we get the pi., and that before the pi. copula,
in John vii. 49.
With a compound subject, the verb can only be
sing, if it stand first (Matt. xvii. 2, John ii. 2, Acts
xx. 4), so that it is intelligible as constructed with
the first only of the nouns, and is supplied with the
rest. We notice a slight irregularity in the use of
a sing, verb with a pi. ptcp. as primary predicate in
Luke ii. 33, and as secondary in Matt. xvii. 2.
In general, the want of clear and straightforward
connexion between subject and predicate is the rarest
of faults in the style of the N. T. : the simplicity of
most of the sentences is a security for their correct-
ness and intelligibility. In the more periodic style of
the Acts, however, we get some entanglement : in
xvii. 2 it is hardly Greek to leave the subject to be
inferred from mention in an oblique case. Of viii. 7
we could at best say the same, if the T. K. were
right ; but as we must certainly read TroXXot, the
only choice is between saying that we have a mixture
of two constructions (TroXXoi ran/ c^. TTJ/. CLK. lOcpairevOrj-
o-av, and TroXXwv TTV. OLK. fto^vra <. /xey. c^p^ovro), and
taking cypx- i*1 a peculiar quasi-transitive sense,
" had spirits come out of them."
A converse case to this is the trajection of the
subject of clause to the beginning of the sentence, for