198 LANGUAGE OF THE NEW TESTAMENT.
Matt. x. 14, Luke ix. 5. But sentences of this sort
shade off into such as Luke x. 8, where the rel. clause
has, if not a definite place in the sentence, a coherent
construction of its own, and from this into such as
ver. 10, where it even gets an antecedent clause at
last. In general, when we meet with anacolutha
more considerable than these, they are too closely
connected with the individual style of the writer to
rank as . characteristics of Hellenistic New Testament
Greek. We may however here notice the change of
constr. in Mark vi. 8, 9 from IW to the inf., and this
through a ptcp., which seems to presuppose an
earlier inf., and in its absence has no proper constr.
at all. This case is not unlike the common one (esp.
common in St. Luke), where a report of a speech
begun in or. obi. slides into or. recta : so indeed a v. I.
here. We have a solitary instance of the opposite
transition from or. recta to obi. in Acts xxiii. 23-4.
Somewhat similar to this, again, is the case of Horn,
ii. 7, 8, xi. 22, where the change from ace. to nom.
cannot be explained, likemost of St. Paul's anacolutha,
either by his losing his way in a long or involved
sentence, or by his wanting, before he had finished
saying one thing, to bring something else into relation
to it. See also p. 77, on Phil. iii. 18, 19. But we
throw no light on slight irregularities like these, by
correlating them with the mixtures of cases that we
get in the Apoc., e.g. vii. 9, xviii. 12, 13.
Of course no difficulty is presented by a sentence
where generally with a rhetorical purpose the
constr. is not changed, but left incomplete : e.g. Acts
xxiv. 19, where the "Jews of Asia who ought to
be here " never get a predicate the Apostle, instead