Reflexive The reflexive pronouns have developed
Pronoun. some unclassical uses, notably that in the
plural they are all fused in to the forms
originally appropriated to the third person. The presence
or absence of this confusion in the singular is a nice test of
the degree of culture in a writer of Common Greek. In the
papyri there are examples of it, mostly in very illiterate docu-
ments,2 while for the plural the use is general, beginning to
appear even in classical times.3 This answers to what we
find in the NT, where some seventy cases of the plural occur
without a single genuine example of the singular;4 late
scribes, reflecting the developments of their own time, have
introduced it into Jn 1834 and Rom 139 (Gal 514). As in the
papyri, e[autoua and some-
times is itself replaced by the personal pronoun. In
translations from Semitic originals we may find, instead of
e[auto5 thus Lk 925, compared
with its presumed original Mk 836. But this principle will
have to be most carefully restricted to definitely translated
passages; and even there it would be truer to say that e[auto
has been levelled up to th>n yuxh>n au]tou?, than that yuxh<
has been emptied of meaning.6
"Exhausted" In one class of phrases e[autou? is used
e[autou? and without emphasis, in a way that brings up the
i@dioj. discussion of its fellow i@dioj.b In sepulchral
inscriptions we find a son describing his
1 Transactions of Cambridge Philological Society, v. i., 1899.
2 See CR xv. 441, xviii. 154, Mayser 304. It is rather perplexing to find it
in literature: e.g. Lucian, Dial. Marin. iv. 3; Polybius 10; Marcus vii.
13; Aristeas 215.
3 Polybius always uses au]tw?n (Kalker, Quaestiones, p.
4 In 1 Co 1029 e[autou?="one's."
5 See J. A. Robinson, Study of the Gospels, p. 114.
6 On the shorter forms au]tou?, etc. see Mayser 305 ff. [a b See p, 240.
88 A GRAMMAR OF NEW TESTAMENT GREEK.
father as o[ pathdifference between the three is not very easily discernible.
In a number of these inscriptions contained in vol. iii. of the
IMA. I count 21 exx. with i@dioj, 10 with e[autou?, and 16
with neither. The papyrus formula used in all legal
documents where a woman is the principal, viz. meta> kuri
tou? e[auth?j a]ndro
rather faded use of the reflexive. It starts the more
serious question whether i@dioj is to be supposed similarly
weakened in Hellenistic. This is often affirmed, and is
vouched for by no less an authority than Deissmann (BS
123 f.). He calls special attention to such passages in the
LXX as Job 2412 (oi@kwn i]di15 (tou? i]di
912 (tou? e[autou? a]mpelw?noj. . . tou? i]di7
(i]di
ever answering to it in the original. He reminds us that
the "exhausted i@dioj" occurs in writers of the literary
Koinh<, and that in Josephus even oi]kei?oj comes to share this
weakening: a few Attic inscriptions from i/B.C. (Meisterhans3
235) show i@dioj with the like attenuated content. Our
inference must be that in Ac 2424 Luke is not ironically
suggesting the poverty of Felix's title, and that in Mt 225
there is no stress on the disloyal guest's busying himself with
his own farm instead of someone else's. (Cf p. 237 below.)
Perhaps, however, this doctrine of the exhausted i@dioj is
in some danger of being worked too hard. In CR xv.
440 f. are put down all the occurrences of i@dioj in BU vols.
i. and ii., which contain nearly 700 documents of various
antiquity. It is certainly remarkable that in all these
passages there is not one which goes to swell Deissmann's
list. Not even in the Byzantine papyri have we a single
case where i@dioj is not exactly represented by the English
own. In a papyrus as early as the Ptolemaic period we
find the possessive pronoun added—o@nta h[mw?n i@dion, which
is just like "our own." (Cf Pet 316, Tit 112, Ac 28.)
This use became normal in the Byzantine age, in which i@dioj
still had force enough to make such phrases as i]di
nomi
we cannot venture to deny in toto the weakening of i@dioj,
still less the practical equivalence of i@dioj and e[autou?, which
ADJECTIVES, PRONOUNS, PREPOSITIONS. 89
is evident from the sepulchral inscriptions above cited, as
well as from such passages as Prov 912 and 1 Co 72. But
the strong signs of life in the word throughout the papyri
have to be allowed for.
In correlating these perplexing phenomena, we may
bring in the following considerations:—(1) the fact that
Josephus similarly weakens oi]kei?oj seems to show that the
question turns on thought rather than on words. (2) It is
possible, as our own language shows, for a word to be
simultaneously in possession of a full and an attenuated
meaning.1 People who say "It's an awful nuisance," will
without any sense of incongruity say "How awfull" when
they read of some great catastrophe in the newspaper. No
doubt the habitual light use of such words does tend in
time to attenuate their content, but even this rule is not
universal. "To annoy" is in Hellenistic sku2 and in
modern French gener. There was a time when the Greek
in thus speaking compared his trouble to the pains of flaying
alive, when the Frenchman recalled the thought of Gehenna;
but the original full sense was unknown to the unlearned
speaker of a later day. Sometimes, however, the full sense
lives on, and even succeeds in ousting the lighter sense, as
in our word vast, the adverb of which is now; rarely heard
as a mere synonym of very. (3) The use of the English
own will help us somewhat. "Let each man be fully
assured in his own mind " (Rom 145) has the double
advantage of being the English of our daily speech and
of representing literally the original e]n t&? i]di<& noi~. What
function has the adjective there? It is not, abnormally, an
emphatic assertion of property: I am in no danger of being
assured in someone else's mind. It is simply method of
laying stress on the personal pronoun: e]n t&? noi~ and "in
his mind" alike transfer the stress to the noun.a This fact
at once shows the equivalence of i@dioj and e[autou? in certain
locutions. Now, when we look at the examples of "exhausted
i@dioj," we find that they very largely are attached to words
that imply some sort of belonging. Husband and wife
account for seven examples in the NT, and other relation-
1 Cf p. 237 below. 2 See Expos. VI. iii. 273 f. a See p. 246.
90 A GRAMMAR OF NEW TESTAMENT GREEK.
ships, including that of master and slave, for a good many
more. A large number come under the category of the
mind, thoughts and passions, and parts of the body. House,
estate, riding-animal, country or language, and similar very
intimate possessions receive the epithet. If occasionally
this sense of property is expressed where we should not
express it, this need not compromise the assertion that
i@dioj itself was always as strong as our English word own.
There are a host of places n the NT, as in the papyri,
where its emphasis is undeniable; e.g. Mt 91, Lk 641, Jn 141
(note its position) 518 etc., Ac 125, 1 Co 38, Gal 65, Heb 727,
and many others equally decisive. One feels therefore quite
justified in adopting the argument of Westcott, Milligan-
Moulton, etc., that the emphatic position of to>n i@dion in Jn 141
was meant as a hint that the unnamed companion of Andrew,
presumably John, fetched his brother. What to do in such
cases as Ac 2424 and Mt 225, is not easy to say. The Revisers
insert own in the latter place; and it is fair to argue that
the word suggests the strength of the counter-attraction,
which is more fully expressed in the companion parable,
Lk 1418. The case of Drusilla is less easy. It is hardly
enough to plead that i@dioj is customarily attached to the
relationship; for (with the Revisers) we instinctively feel
that own is appropriate in 1 Pet 31 and similar passages,
but inappropriate here. It is the only NT passage where
there is any real difficulty; and since B stands almost alone
in reading i]di<%, the temptation for once to prefer x is very
strong. The error may have arisen simply from the common-
ness of the combination h[ i]di
ferred to a context in which it was not at home.
[O i@dioj. Before leaving i@dioj something should
be said about the use of o[ i@dioj without a
noun expressed. This occur in Jn 111 131, Ac 423 2423
In the papyri we find the singular used thus as a term
of endearment to near relations: e.g. o[ dei?na t&? i]di<&
xaiExpos. vi. iii. 277 I ventured to cite this as a
possible encouragement to those (including B. Weiss) who
would translate Ac 2028 "the blood of one who was his
own." Mt 2724, according to the text of xL and the later
authorities, will supply a parallel for the grammatical
ADJECTIVES, PRONOUNS, PREPOSITIONS. 91
ambiguity: there as here we have to decide whether the
second genitive is an adjective qualifying the first or a noun
dependent on it. The MGr use of o[ i@dioj, as substitute for
the old o[ au]to
but in the papyrus of Eudoxus (ii/B.C.) we (find a passage
where th?i i]di
so that it seems inevitable to trace, with Blass, an anti-
cipation of MGr here. Perhaps the use was locally
restricted.
Au]to>j o[ and There is an apparent weakening of
o[ au]to>j. au]to>j o[ in Hellenistic, which tends to blunt
the distinction between this and e]kei?noj o[.
Dean Robinson (Gospels, p. 106) translates Lk 1021 "in that
hour" (Mt 1125 e]n e]kei12 (Mk 1311
e]kei7. It is difficult to be satisfied with "John
himself " in Mt 34; and in Luke particularly we feel that
the pronoun means little more than "that." Outside Luke,
and the one passage of Mt, au]to>j o[ has manifestly its full
classical force. From the papyri we may quote OP 745
(i/A.D.) au]to>n to>n ]Anta?n," the said A.": note also GH 26
(ii/B.C.) o[ aut]o>j $Wroj, "the same Horus," i.e. "the aforesaid,"
and so in BU 1052 (i/B.C.). We find the former use in
MGr, e.g. au]to> to> kri
have already seen (p. 86) that the emphatic au]to>j standing
alone can replace classical e]kei?noj (See now Wellh. 26 f.)
Relatives :— Turning to the Relatives we note the
Use of o!stij. limiting of o!stij, a conspicuous trait of the
vernacular, where the nominative (with the
neuter accusative) covers very nearly all the occurrences of
the pronoun. The phrase e!wj o!tou is the only exception in
NT Greek. The obsolescence of the distinction between o!j
and o!stij is asserted by Blass for Luke, but not for Paul.
A type like Lk 24 ei]j pod h!tij kalei?tai Bhqlee
may be exactly paralleled from Herodotus (see Blass 173)
and from papyri: so in an invitation formula au@rion h!tij
e]sti>n ie, "to-morrow, which is the 15th"—cf Mt 2762. Hort,
on 1 Pet 211 (Comm. p. 133), allows that "there are some
places in the NT in which o!stij cannot be distinguished from
o!j." "In most places, however, of the NT," he proceeds," o!stij
apparently retains its strict classical force, either generic,
92 A GRAMMAR OF NEW TESTAMENT GREEK.
'which, as other like things,' or essential, 'which by its very
nature.'" A large number of the exceptions, especially in
Lucan writings, seem to be by no means cases of equivalence
between o!j and o!stij, whether agreeing or disagreeing with
classical use. Some of them would have been expressed
with o!sper in Attic: thus in Ac 1128 we seem to expect
h!per e]ge
which can be brought out by various paraphrases, as in Lk 120,
"which for all that." Or o!stij represents what in English
would be expressed by a demonstrative and a conjunction, as
in Lk 1042, "and it shall not be taken away." In Mt we
find o!stij used four times a the beginning of a parable,
where, though the principal figure is formally described as
an individual, he is really a type, and o!stij is therefore
appropriate. We may refer to Blass 173, for examples
of o!j used for o!stij, with indefinite reference. The large
number of places in which o!stij is obviously right, according
to classical use, may fairly stand as proof that the distinction
is not yet dead. We must not stay to trace the distinction
further here, but may venture on the assertion that the
two relatives are never absolutely convertible, however
blurred may be the outlines of the classical distinction in
Luke, and possibly in sporadic passages outside his writings.
Milker (Quest. 245 f.) asserts that Polybius uses o!stij for o!j
before words beginning with a vowel, for no more serious
reason than the avoidance of hiatus; and it is curious that
among twenty-three more or less unclassical examples in the
Lucan books fourteen do happen to achieve this result. We
chronicle this fact as in duty bound, but without suggesting
any inclination to regard it as a key to our problem. If
Kalker is right for Polybius—and there certainly seems
weight in his remark that this substitution occurs just where
the forms of o!j end in a vowel--we may have to admit that
the distinction during the Koinh< period had worn rather
thin. It would be like the distinction between our relatives
who and that, which in a considerable proportion of sentences
are sufficiently convertible to be selected mostly according
to our sense of rhythm or euphony: this, however, does not
imply that the distinction is even blurred, much less lost.
The attraction of the Relative—which, of course, does
ADJECTIVES, PRONOUNS, PREPOSITIONS. 93
not involve o!stij—is a construction at least as popular in late
Attraction. as in classical Greek. It appears abundantly
in their papyri, even in the most illiterate
of them; and in legal documents we have the principle
stretched further in formula, such as a]rourw?n de
h} o!swn e]a>n w#sin ou]sw?n. There are to be noted some
exceptions to the general rule of attraction, on which see
Blass 173. In several cases of alleged breach of rule we may
more probably (with Blass) recognise the implied presence
of the "internal accusative": so in 2 Co 14, Eph 16 41, where
Dr Plummer (CGT, 2 Co i.e.) would make the dative the
original case for the relative.
Relatives and Confusion of relative and indirect inter-
Interrogatives roative is not uncommon. " !Osoj, oi#oj,
confused. o[poi?oj, h[li
interrogatives, and also—with the exception
of h[li
and in the papyri even o!j can be used in an indirect question.
Good examples are found in PP ii. 37 (ii/B.C.) kalw?j ou#n
poih
(iii/B.C.) fra e]n h$i kw
oi]kou?sin kai> p[oAntig.
542, OT 1068 (see Jebb's notes) ; and in Plato, Euth. 14E
a{ me>n ga>r dido dh?lon. It is superfluous to say
that this usage cannot possibly be extended to diect question,
so as to justify the AV in Mt 2650. The more illiterate
papyri and inscriptions show ti
seldom, as eu$ron georgo>n ti e[lkun xri
e@x^j--ti1 etc. Jebb on Soph. 0T 1141
remarks that while "ti
only where there is an indirect question, . . . Hellenistic Greek
did not always observe this rule: Mk 1436." There is no ade-
quate reason for punctuating Jas 313 so as to bring in this
misuse of ti19 and Lk 178 are essentially similar;2
nor does there seem to be any decisive reason against so reading
Ac 1325. Dieterich (Unters. 200) gives several inscriptional
exx., and observes that the use was specially strong in Asia
1 BU 822 (iii/A. D. ), BM 239 (iv/A.D.), JHS xix. 299. See p. 21 above. Gn 3825
is a clear ex. from LXX. 2 I must retract the denial I gave in CR xv. 441.
94 A GRAMMAR OF NEW TESTAMENT GREEK.
Minor. It is interesting therefore to note Thumb's statement
(ThLZ xxviii. 423), that the interrogative is similarly used in
Pontic now—a clear case of local survival. The NT use of
o!ti, for ti< in a direct question is a curious example of the
confusion between the two categories, a confusion much
further developed in our own language.
Developments MGr developments are instructive when
in MGr. we are examining the relatives and inter-
rogatives. The normal relative is pou?, fol-
lowed by the proper case of the demonstrative, as o[ giatro>j
pou? to>n e@steila, "the doctor whom I sent," etc. The
ingenious Abbe Viteau discovers a construction very much
like this, though he does not draw the parallel, in Jn 917 o!ti
h]ne<&cej o]fqalmou
opened": he cites Mk 617f. 824 as further exx. Since o! ti
and rw,xE are passable equivalents, we have here a "pure
Hebraism"—a gem of the first water! We might better
Viteaa's instruction by tracing to the same fertile source
the MGr idiom, supporting our case with a reference to
Jannaris HG § 1439, on MGr parallels to Mk 725 (h$j. . .
au]th?j) and the like.1 It will be wise however for us to sober
ourselves with a glance at Thumb's remarks, Hellen. 130,
after which we may proceed to look for parallels nearer home
than Hebrew. In older English this was the regular con-
struction. Thus, "thurh God, the ic thurh his willan hider
asend waes" (Gen 458); "namely oon That with a spere
was thirled his brest-boon " (Chaucer, Knightes Tale 1851 f.).
Cf the German "der du bist" = who art.2 The idiom is
still among us; and Mrs Gamp, remarking "which her
name is Mrs Harris," will hardly be suspected of Hebraism!
The presence of a usage in MGr affords an almost decisive
disproof of Semitism in the Koinh<, only one small corner of
whose domain came within range of Semitic influences; and we
have merely to recognise afresh the ease with which identical
idioms may arise in totally independent languages. It does
not however follow that Blass is wrong when he claims
1 See below, p. 237; also Wellh. 2, who adds exx. from D.
2 See Skeat's Chaucer, Prologue and Knightes Tale, p. xxxvi. I owe the
gestion to my friend Mr E. E. Kellett.
ADJECTIVES, PRONOUNS, PREPOSITIONS. 95
Mk 725 17 1319, Lk 316, and passages in Rev, as "specialy
suggested by Semitic usage." The phenomenon is frequent
in the LXX (see WM: 185), and the NT exx. are nearly
all from places where Aramaic sources are presumed. A
vernacular use may be stretched (cf pp. 10 f.) beyond its
natural limits, when convenient for literal translation. But
Blass's own quotation, ou$ h[ pnoh> au]tou? e]n h[mi?n e]sti1 comes
from a piece of free Greek. That this use did exist in the
old vernacular, away from any Semitic influence, is proved
by the papyri (p. 85). The quotations in Kuhner-Gerth
§ 561 n.2, and in Blass and Winer ll.cc., show 'that it had
its roots in the classical language. As was natural in a
usage which started from anacoluthon, the relative and
the pleonastic demonstrative were generally, in the earlier
examples, separated by a good many intervening words.
The modern Interrogative is mostly poio
practically worn down to the indeclinable ti<, just as our
what (historically identical with the Latin quod) has become
indifferent in gender. The NT decidedly shows the early
stages of this extension of poi?oj. It will not do for us to
refine too much on the distinction between the two pronouns.
The weakening of the special sense of poi?oj called into being a
new pronoun to express the sense qualis, namely, potapo
was the old podapo
etymology to suggest po
tion in meaning with a]llod-apo2
Numerals :— We take next the Numerals. The use
ei$j as ordinal; of ei$j as an ordinal is "undoubtedly a
Hebrew idiom," according to Blass, p. 144.
Our doubts, nevertheless, will not be repressed; and they
are encouraged by the query in Thumb's review. To
begin with, why did the Hebraism affect only the first
numeral, and not its successors? If the use was vernacular
Greek, the reason of the restriction is obvious: prw?toj is
the only ordinal which altogether differs in foam from the
1 Clement ad Cor. 21 fin. (Lightfoot, p. 78). Nestle (ZNTW i. 178 ff.)
thinks the writer was of Semitic birth. Gal 210 will serve instead.
2 The suffix is that of Latin prop-inquos, long-inquos, Skt. anv-anc, etc.: pod-
and a]llod- are quod, what, aliud, while h[med-, u[med-, answer to ablative forms
in Skt.
96 A GRAMMAR OF NEW TESTAMENT GREEK.
cardinal.1 When we add that both German and English say
"page forty" (WM 311), we are prepared for the belief that
the Greek vernacular also had his natural use. Now, although
ei$j kai> ei]kostounus et vicesimus, one and twentieth, are (as
Blass says) essentially different, since the ordinal element is
present at the end of the phrase, this is not so with t^? mi%? kai>
ei]ka2 BU 623 A.D.). But the matter is really settled
by the fact that in MGr the cardinals beyond 4 have ousted
the ordinals entirely (Thumb, Handbuch 56); and Dieterich
(Unters. 187 f.) shows from inscriptions that the use is as old
as Byzantine Greek. It would seem then that the encroach-
ment of the cardinal began in the one case where the ordinal
was entirely distinct in form, spread thence over other
numerals, and was finally repelled from the first four, in which
constant use preserved alike the declension and the distinct
ordinal form. Had Semitic influence been at work, there is
no conceivable reason why we should not have had t^? pe
at the same time. Simultaneously with this process we note
Simplification the firm establishment of simplified ordinals
of the “teens”; from 13th to 19th, which now (from iii/B.C.
onwards) are exclusively of the form triskai-
de
Similarly we find de
papyri, and de3a These phenomena
all started in the classical period: cf Meisterhans3 160.
ei$j as Indefinite There is a further use of ei$j which calls
Article. for remark, its development into an indefinite
article, like ein in German, un in French, or
our own an: in MGr the process is complete. The fact that
1 Deu
connect them. Curiously enough, Hebrew shares the peculiarity noted above,
which somewhat weakens our argument Aramaic, like Latin and English, uses
a word distinct from the cardinal for second as well as first. Hebrew has lost
all ordinals beyond 10, and Aramaic shows them only in the Jerus. Targ. See
Dalman, Gramm. 99 f. For clays of the month, the encroachment of cardinals
has gone further still in both dialects. The fact that the ordinals up to 10 are
all treated alike in Hebrew, reinforces our view.
2 Ei]kaNo. 20 or a set
of 20, though used only for the 20th of the month. Cf in Philo tria(LS), and tetra
3 Wellhausen notes that D has only deADJECTIVES, PRONOUNS, PREPOSITIONS. 97
ei$j, progressively ousted tij in popular speech, and that even
in classical Greek there was a use which only needed a little
diluting to make it essentially the same,1 is surely enough to
prove that the development lay entirely within the Greek
language, and only by accident agrees with Semitic. (See
Wellh. 27.) We must not therefore follow Meyer (on Mt
819), in denying that ei$j is ever used in the NT in the sense
of tij: it is dangerous to import exegetical subtleties into the
o[ ei$j NT, against the known history of the Common
Greek. The use of o[ ei$j in Mk 1410 is, as
noted in Expos. VI. vii. 111, paralleled in early papyri.2
In Blass's second edition (p. 330) we find a virtual sur-
Distributives. render of the Hebraism in dusumpo39f.), desma>j desma30
in Epiphanius --a very probable reading, as accounting for the
variants): he remarks on mi
that "Atticists had evidently complained of it as vulgar, and
it was not only Jewish-Greek." Winer compared Aeschylus
Persae 981, muriThLZ,
1898, p. 631) cites dh
and (as W. F. Moulton noted WM 312 n.) the usage is
found in MGr.3 Thumb is undeniably right in calling the
coincidence with Hebrew a mere accident. In the papyri
(e.g. Tb P 635 --ii/B.C.) the repetition of an adjective produces
an elative = mega
that in Lk 101 we have a mixed distributive a]na> du
(B al): so in Ev. Petr. 35, as Blass notes, and Acta Philippi
92 (Tisch.).4 See Brugmann, Distributiva (cites above, p. 21).
"Noah the Two single passages clai a word before
eighth person. we pass on from the numerals. @Ogdoon
Nw?e e]fu5 presents us with
1 It is difficult to see any difference between ei$j and tij in Aristophancs,
Av. 1292 :—
pen ei$j ka
xwlon tou@noma, k.t.l.
From the papyri we may cite as exx. AP 30 (ii/B.C.) Konduj tw?n a[liei
(Sc. prosklhqe
2 We may add good exx. from Par P 15 (ii/B.C.) to>n e!na au]tw?n $Wron—tou? e[no>j
tw?n e]gkaloumej au]tw?n patro
3 Thumb, Hellen. 128, Handbuch, 57.
4 See W. Schulze, Graeca Latina 13. Add now Wellh. 31.
98 A GRAMMAR OF NEW TESTAMENT GREEK.
a classical idiom which can be shown to survive at any rate in
literary Common Greek: see exx. in WM 312, and Schaefer l.c.
I have only noticed one instance in the papyri (p. 107), and
in 2 Pet we rather expect bookish phrases. The AV of
this passage is an instructive illustration for our inquiries
as to Hebraisms. "Noah the eighth person" is not English,
for all its appearing in a work which we are taught to regard
as the impeccable standard of classic purity. It is a piece of
"translation English," and tolerably unintelligible too, one
may well suppose, to its less educated readers. Now, if this
specimen of translators' "nodding" had made its way into
the language—like the misprint "strain at a gnat"—we
should have had a fair parallel for "Hebraism" as hitherto
understood. As it stands, a phrase which no one has ever
thought of imitating, it serves to illustrate the over-literal
translations which appear very frequently in the LXX and in
the NT, where a Semitic original underlies the Greek text.
(Compare what is said of Gallicisms in English on p. 13.)
" Seventy times Last in this division comes a note on
seven." Mt 1822. Blass ignores entirely the ren-
dering "seventy-seven times" (RV margin),
despite the fact that this meaning is unmistakable in Gen 424
(LXX). It will surely be felt that W. F. Moulton (WM
314) was right in regarding that passage as decisive. A
definite allusion to the Genesis story is highly probable:
Jesus pointedly sets against the natural man's craving for
seventy-sevenfold revenge the spiritual man's ambition to
exercise the privilege of seventy-sevenfold forgiveness. For
a partial grammatical parallle see Iliad xxii. 349, deka
kai> Fei
Prepositions :— It will be worth while to give statistics
Relative for the relative frequency of Prepositions in
Frequency. the NT, answering to those cited from Helbing
(above, pp. 2 f.) for the classical and post-
classical historians. If we represent e]n by unity, the order of
precedence works out thus:-- ei]j 64, e]k 34, e]pi< 32, pro
25, dia 24, a[po< 24, kata< 17, meta< 17, peri< 12, u[po<
08, para< 07, u[pe
a]na< 0045. We shall have to return later to prepositions
compounded with verbs, following our present principle of
ADJECTIVES, PRONOUNS, PREPOSTTIONS. 99
dealing with them in connexion with the parts of speech
with which they are used. A few miscellaneous matters
come in best at this point. First let us notice the pro-
Prepositions minence in Hellenistic of combinations of
joined with prepositions with adverbs. In papyri we
Adverbs. find such as e]k to
peBS 221), and even a]f ]
o!te e]lousa
NT we have a]po> to pe
a!pac, e]pi> tri
the classical e]j a]ei<, and the like. Some of these combinations
became fixed, as u[poka
be set beside the abundance of "Improper" prepositions. All
of these, except e]ggu1 Thumb
comments2 on the survival of such as e!wj, e]pa
u[poka
have been responsible for the coining of e]nw
mann proved it vernacular.3 The compound preposition a]na>
me
in the papyri,—not however in any use which would help
1 Co 65, where it is almost impossible to believe the text
sound. (An exact parallel occurs in the Athenaeum for Jan.
14, 1905, where a writer is properly censured for saying,
"I have attempted to discriminate between those which are
well authenticated," i.e. (presumably) "[and those which are
not]." It is hard to believe Paul would have been so slovenly
in writing, or even dictating.) We have a further set of
"Hebraisms" in the compound prepositions which are freely
made with pro
above, p. 81. Even here the Semitism is still on the
familiar lines: a phrase which is possible in native Greek
is extended widely beyond its idiomatic limits because it
translates exactly a common Hebrew locution; and the
conscious use of Biblical turns of speech explains the appli-
cation of such phrases on the lips of men whose minds are
saturated with the sacred writers' language. As early as iii/B.C.
1 Paraplh27. xACD has dat. 2 TLZ xxviii. 422. 3 BS 213.
Cf Expos. vii. 113: add OP 658 (iii/A.D.), and Tb P 14 (114 B.C.) parhggel-
ko
earliest ex. Cf Par P 63 (ii/B.C.) e]nopi
100 A GRAMMAR OF NEW TESTAMENT GREEK.
in a Libyan's will, we meet with kata> pro1 and
in mercantile language we constantly find the formula dia>
xeiro
hand to hand," as contrasted with "through an intermediary."
We may refer to Heitmuller's proof2 that the kindred phrase
ei]j to> o@noma< tinoj is good vernacular. The strong tendency
to use compound prepositional phrases, which we have been
illustrating already, would make it all the easier to develop
these adaptations of familiar language.
Prepositions The eighteen classical prepositions are,
with one case. as we have just seen, all represented in NT
Greek, except a]mfi<, which has disappeared
as a separate word, like ambi in Latin, and like its correlative
in English, the former existence of which in our own branch
is shown by the survival of um in modern German. It
was not sufficiently differentiated from peri<, to assert itself
in the competition; and the decay of the idea of duality
weakened further a preposition which still proclaimed its
original meaning, "on both sides," by its resemblance to
a]mfo
use, which accounts for seven instances, the phrase a]na> me
for four, and a]na> me
but a]nq ] w$n reduces the number of free occurrences to 17.
Rare though it is, it retains its individuality. "In front of,"
with a normal adnominal genitive, passes naturally into "in
place of," with the idea of equivalence or return or substitu-
tion, our for. For the preposition in Jn 116, an excellent
parallel from Philo is given in WM (p. 456 n.).3 Pro< occurs
48 times, including 9 exx. of pro> tou? c. inf., which invades
the province of pri21 we have pro> e{c h[merw?n
tou? paante diem tertiwm,
Kalendas. The plausible Latinism forces itself on our
attention all the more when we compare IMA iii. 325 (ii/A.D.)
1 Deissmann BS 140.
2 Im Namen Jesu 100 ff. So p. 63, for e]n o]no41.
3 Blass compares gh?n pro> gh?j e]laue]lpiDe
Poster. Caini § 145 (p. 254 M.): dio> ta>j prw xan koresqee]cubrij laxon kai> tamieusa
kai> tri tw?n deute ai]ei> ne palaioteADJECTIVES PRONOUNS, PREPOSITIONS. 101
pro ie Kalandw?n Au]gou
documents to be seen in Viereck's Sermo Graecus (see pp. 12,
13, 21, etc.). And yet it is soon found that the same
construction occurs in phrases which have nothing in
common with the peculiar formula of Latin days of the
month. In the Mysteries inscription from Andania (Michel
694, i/B.C.) we recognise it in Doric—pro> a[mera?n de
musthri
prw> du o]rniqa
fowls two days before the feast"), when combined with Jn l. c.,
makes the hypothesis of Latinism utterly improbable. The
second genitive in these three passages is best taken as an
ablative—"starting from the mysteries," etc. It is found as
early as Herodotus, who has (vi. 46) deute
the second year from these events": cf also OP 492 (ii/A.D.) met ]
e]niauto>n e!na th?j teleuth?j mou, "a year after (starting from)
my death." See also the note on o]ye<, supr. p. 72. There
remains the idiomatic use of pro<, seen in 2 Co 122 pro> e]tw?n
dekatessar
cites pro> a[mera?n de
1001), written in the Doric of Thera, "end of iii/B.C. or
beginning of ii/B.C., therefore pre-Roman"—to cite Blass's own
testimony.1 It becomes clear that historically the resem-
blance between the ante diem idiom and the Greek which
translates it is sheer coincidence, and the supposed Latinism
goes into the same class as the Hebraisms we have so often
disposed of already.2 This enquiry, with the general con-
siderations as to Latinisms which were advanced above (pp.
20 f.), will serve to encourage scepticism when we note the
1 Add FP 122 (i/ii A.D. ), BU 180 (ii/iii A.D.), 592 (ii/A.D.), NP 47 (iii/A.D.),
Ch P 15 (iv/A.D.), BU 836 (vi/A.D).
2 W. Schulze, Graec. Lat. 14-19, has a long and striking list of passages
illustrating the usage in question, which shows how common it became. His
earliest citation is pro> triw?n h[merw?n th?j teleuth?j from Hippocrates (v/B.C.),
which will go with that from Herodotus given above. We have accordingly
both Ionic and Doric warrant for this Koinh< construction, dating from a period
which makes Latin necessarily the borrower, were we bound to deny independent
development. Schulze adds a parallel from Lithuanian! Our explanation of
the dependent gen. as an ablative is supported by pro> mia?j h[meacc. et inf.,
in OGIS 435 (ii/B.C.) and Jos. Ant. xiv. 317: h@ replaces the ablative genitive
exactly as it does after comparatives.
102 A GRAMMAR OF NEW TESTAMENT GREEK.
resemblance of w[j a]po> stadi13) to a milli-
bus passuum duobus (Blass 95). Blass cites Jn 218, Rev 1420,
and the usage of Koinh< writers like Diodorus and Plutarch.
Mutatis mutandis, this idiom is identical in principle with that
just quoted for pro<. After noting the translation-Hebraism
fobei?sqai a]po< in Mt 1028 ( = Lk 124),1 we proceed to observe
the enlargement of the sphere of a]po<, which encroaches upon
e]k, u[po<, and para<.a The title of the modern vernacular
Gospels, "metafrasme to>n ]Alec. Pa
that a]po< has advanced further in the interval. Already in
the NT it sometimes expressed the agent after passive verbs
(e.g. Lk 843), where it is quite unnecessary to resort to
refinements unless the usage of a particular writer demands
them. The alleged Hebraism in kaqaro>j a]po< is dispelled by
Deissmann's quotations, BS 196. The use of prepositions,
where earlier Greek would have been content with a simple
case, enables e]k in NT to outnumber a]po< still, though
obsolete to-day,b except in the Epirot a]x or o]x.2 Thus a]po<
is used to express the partitive sense, and to replace the
genitive of material (as Mt 2721 34); e]k can even make a
partitive phrase capable of becoming subject of a sentence, as
in Jn 1617. For present purposes we need not pursue further
the NT uses of a]po< and e]k, which may be sought in the
lexicon; but we may quote two illustrative inscriptional
passages with e]k. Letronne 190 and 198 have swqei>j e]k,
"safe home from" (a place), which has affinity with Heb 57;
and u[paj e]k qeou? kai> qea?j, from the Rosetta stone
(OGIS 90—ii/B.C.), will elucidate Phil 35, if the reader of
the Greek should, conceivably, fall into the misconceptions
which so many English readers entertain. It gives us an
unpleasant start to find the language of the Nicene Creed
used centuries earlier of Ptolemy Epiphanes!3
We have already (pp. 62 f.) sketched the developments of
1 Were the active fobei?n still extant (below, p. 162), this might be taken as
"do not be panic-stricken by." It is like prose1. See p. 107.
2 Thus o]x to> bouno<, " from the hill," occurs in a modern song, Abbott 128 f.
3 Epiphanes=Avatar: the common translation " illustrious " is no longer
tenable. See Dittenberger's note, OGIS p. 144. So this title also antici-
pates the NT (e]pifa
terms, above, p. 84. (On a]po< see also below, p. 237.) [a b See p. 246,
ADJECTIVES, PRONOUNS, PREPOSITIONS. 103
ei]j, and need say no more of the single-case prepositions,
with one very large exception.a The late Greek uses of
Further uses e]n would take too much space if discussed in
of e]n. full here. It has become so much a maid-of-
all-work that we cannot wonder at its ulti-
mate disappearance, as too indeterminate. Students of Pauline
theology will not need to be reminded of Deissmann's masterly
monograph on "The NT Formula e]n Xrist&? ]Ihsou?," with its
careful investigation of LXX uses of and proof of the
originality of Paul's use. But SH (on Rom 611) seem rightly
to urge that the idea of the mystic indwelling originated with
the Master's own teaching: the actual phrase in Jn 154 may
be determined by Pauline language, but in the original Aramaic
teaching the thought may have been essentially present.
While there are a good many NT uses of e]n which may be
paralleled in vernacular documents, there are others beside
this one which cannot: in their case, however, analogy makes
it highly improbable that the NT writers were innovating.
If papyri have probebhko
we need not assume Hebraism in Lk 17 merely because the
evangelist inserts e]n: his faithful preservation of his source's
h[me14
(LXX) we have e]n = "amounting to," from which that in
Mk 48 bis does not greatly differ. This is precisely paralleled
by BU 970 (ii/A.D.) prooi?ka e]n draxmai?j e]nnakosi
(ii/A.D.) e@sxej th>n prw
BU 105 0 (i/A.D.) i[ma
the value of"). The use in Eph 215 e]n do
in," is akin to this. For e]n toi?j = "in the house of," as in
Lk 249, we have RL 382 (iii/B.C.) e]n toi?j ]Apollwni
(ii/B.C.) e]n toi?j ]Amenne
e]n toi?j Klaudi Prwta
katalu
official documents e]n meaning "in the department of": so
Tb P 27 (ii./B.C.) to> e]n au]tw?i o]feilo
topogrammatei?, al. I do not recall an exact NT parallel, but
1 Co 62, ei] e]n u[mi?n kri
have another use of e]n with a personal dative in 1 Co 1411
"in my judgement": possibly Judel e]n qe&? is akin to this.
Such uses would answer to para< c. dat. in classical Greek
a See v. 246.
104 A GRAMMAR OF NEW TESTAMENT GREEK.
The last might seem to be expressed more naturally by the
"dative of person judging" (like Ac 720 a]stei?oj t&? Qe&?, or
1 Co l.c. e@somai t&? lalou?nti ba
uses of dative and locative have some common ground, which
is indeed the leading cause of their syncretism. Thus we find
loc. in Sanskrit used quite often for the dat. of indirect object
after verbs of speaking. How readily e]n was added to the
dative, which in older Greek would have needed no preposi-
tion, we see well in such a passage as OP 48 8 (ii/iii. A.D.),
where " more . . . by one aroura" is expressed by e]n. This
particular dative is an instrumental—the same case as our
"the more the merrier"—, and is therefore parallel to that
of e]n maxai
mentioned (pp. 12, 61). We may fairly claim that "Hebraistic"
e]n is by this time reduced within tolerably narrow limits. One
further e]n, may be noted for its difficulty, and for its bearing
on Synoptic questions,--the i[mmologei?n e@n tini which is common
to Mt 1032 and Lk 128: this is among the clearest evidences
of essentially identical translations used in Mt and Lk. W. F.
Moulton (WM 283 n.) cites, apparently with approval, Godet's
explanation—"the repose of faith in Him whom it confesses":
so Westcott, quoting Heracleon, who originated this view
(Canon5 305 n.). Deissmann (In Christo 60) quotes Delitzsch's
Hebrew rendering ybi hd,Oy , and puts it with Mt 317 934 116
2321, as an example of a literal translation "mit angstlicher,
die hermeneutische Pedanterie nahelegender Pietat." Dr
Bendel Harris recalls the Graecised translation in Rev 35, and
gives me Syriac parallels. On the whole, it seems best not
to look for justification of this usage in Greek. The agreement
of Mt and Lk, in a point where accidental coincidence is out
of the question, remains the most important element in the
whole matter, proving as it does that Luke did not use any
knowledge of Aramaic so as to deal independently with the
translated Logia that came to him.1
Prepositions Of the prepositions with two cases, di
with two and meta< show no signs of weakening their
Cases; hold on both; but kata< c. gen. and peri<
u[pe
1 Cf the similar agreement as to fobei?sqai a]po<, above, p. 102.
ADJECTIVES, PRONOUNS, PREPOSITIONS. 105
We may give the statistics in proof. Dia< gen. 382, acc,
279; meta< gen. 361, acc. 100; kata< gen. 73, acc. 391;
peri<, gen. 291, acc. 38; u[pe
165, acc. 50. Comparing this list with that in a classical
Greek grammar, we see that meta<, peri< and u[po<1 have been
detached from connexion with the dative a fact in line
with those noted above, pp. 62 ff. Turning to details, we
find that kata<, (like a]na<, Rev 2121) is used as an adverb
distributively, as in to> kaq ] ei$j or ei$j kata> ei$j Mk 1419, [Jn] 89,
Rom 125. The MGr kaqei
which probably started from the stereotyping of to> kaq ] e!na,
e{n kaq ] e!n, etc., declined by analogy: cf e@ndhmoj from e]n
dhproconsul from pro console. The enfeebling of
the distinction between peri< and u[pe
some importance in the NT, where these prepositions are
used in well-known passages to describe the relation of the
Redeemer to man or man's sins. It is an evident fact that
u[pe23: it is used,
for example, scores of times in accounts, with the sense of
our commercial "to." This seems to show that its original
fullness of content must not be presumed upon in theological
definitions, although it may not have been wholly forgotten.
The distinction between a]nti< and the more colourless u[pe
applying the metaphor of purchase, is well seen in Mk 1045
( Mt 2028) lu pollw?n, and the quotation of this
logion in 1 Tim 26 a]ntir pa2 Dia< c. acc.
mostly retains its meaning "for the sake of," "because
of," distinct from "through," "by the instrumentality of,”
which belongs to the genitive. As early as MP 16 and
20 (iii/B.C.), we have i!na dia> se> basileu? tou? dikai
but if the humble petitioner had meant "through you,"
he would have addressed the king as a mere medium of
favour: referring to a sovereign power, the ordinary meaning
"because of you" is more appropriate. This applies exactly
to Jn 657. So Rom 820, where Winer's explanation is correct
(p. 498). In much later Greek, as Hatzidakis shows (p. 213)
1 For u[po< c. dat. can be quoted OGIS 54 (iii/B.C.) u[f ] e[autw?i poihsaand OP 708 (as late as ii/A.D.) e]k tou? u[po> soi> nomou?. LXX has peri< c. dat.
2 Note that dou>j e[auton
yuxh>n au]tou?: on this see above, p. 87. See further on u[pe
106 A GRAMMAR OF NEW TESTAMENT GREEK.
dia< c. acc. monopolised the field, which it still holds in
MGr.1 With the genitive, dia< is often contrasted with
e]k, u[po<, etc., as denoting mediate and not original authorship:
as 1 Co 86, Mt 122. In Heb 210 it is used of God, who is "the
final Cause and the efficient Cause of all things" (Westcott).
There seems no adequate reason for accepting Blass's con-
jectural emendation, di ] a]sqenei13: "because of an
illness" is an entirely satisfactory statement (see Lightfoot
in loc.), and the Vulgate per is not strong enough to justify
Blass's confidence.2 Meta< c. gen. has in Lk 158 a use
influenced by literal translation from Semitic.a Its relations
with su
very much the commoner way of saying with. Thumb
points out (Hellen. 125) that MGr use disproves Hebraism
in polemei?n meta< tinoj, Rev 127 al.b Thus, for example, Abbott
44: pole trei?j xilia
3000 Turks."
and with The category of prepositions used with
three. three cases is hurrying towards extinction,
as we should expect. Meta<, peri< and u[po<
have crossed the line into the two-case class and in the NT
pro
c. gen. 1 (Ac 2734, literary), dat. 6 ( = "close to" or "at,"
in Mk, Lk, Jn ter and Rev), acc. 679. With the dative,
however, it occurs 104 times in LXX, and 23 times c. gen.:
the decay seems to have been rapid. Cf however PFi 5
pro>j t&? pulw?ni, as late as 245 A.D. For para< the numbers
are, c. gen. 78, dat. 50, acc. 60. Blass notes that c. dat. it
is only used of persons, as generally in classical Greek, except
in Jn. 1925. One phrase with para< calls for a note on its
use in the papyri. Oi[ par ] au]tou? is exceedingly common
there to denote "his agents" or “representatives.” It has
hitherto been less easy to find parallels for Mk 321, where
it must mean "his family": see Swete and Field in loc.
We can now cite GH 36 (ii/B.C.) oi[ par ] h[mw?n pa1 Contrast Ac 242 with OP 41 (iii/iv A.D.) pollw?n a]gaq?n a]polau
dia> sai<.
2 Ou] dunaand a like phrase from OP 261 (i/A.D.), but of course they prove little of
nothing. [a See pp. 246 f.; b see p. 247.
ADJECTIVES, PRONOUNS, PREPOSITIONS. 107
BU 998 (ii/B.C.), and Par P 36 (ii/B.C.).1 Finally we come
to e]pi<, the only preposition which is still thoroughly at home
with all the cases (gen. 216, dat. 176, acc. 464). The
weakening of case-distinctions is shown however by the very
disproportion of these figures, and by the confusion of meaning
which is frequently arising. In Heb 810 1016 we construe
kardi th>n dia
it in the latter passage: on the other hand, the original in
Jer 31(38)33 is singular, which favours taking it as genitive.2
Our local upon can in fact be rendered by e]pi< with gen.,
dat., or acc., with comparatively little difference of force.
Particular phrases are appropriated to the several cases, but
the reason is not always obvious, though it may often be
traced back to classical language, where distinctions were
rather clearer. Among the current phrases we may note
e]pi> to> au]to< "together," "in all," perpetually used in arith-
metical statements: see Ac 115 247. Cf Blass2 330. The
common e]f ] &$ c. fut. indic. "on condition that," does not appear
in the NT. But with a pres. in 2 Co 54, and an aor. in Rom 512,
the meaning is essentially the same ("in view of the fact that"),
allowing for the sense resulting from a jussive future.
1 Expos. vi. vii. 118, viii. 436. See Witkowski's note, p. 72.
2 For Mk 639 e]pi> t&? xo
in D. In Ac 711 D has gen. for acc., and in 816 acc. for dat. In Eph 110 it
seems difficult to draw any valid distinction between the cases of e]pi> toi?j
ou]ranoi?j and e]pi> th?j gh?j. Nor can we distinguish between e]p ] e]sxa1
and the dative in Tb P 69 (ii/B.C.), w$n h[ dioi
ADDITIONAL NOTES.—P. 79. Mr Thackeray says prw?toj is used for pro
regularly in LXX. The latter occurs not infrequently in Ptolemaic papyri, but
seems to have weakened greatly in the Roman period.—P. 98. The Ptolemaic
PP iii. 28 has e]dragmatokle
JG 562 on p. mo
Jn 615x. On Mt 1822, W. C. Allen takes 70 x7 in Gen and Mt ll. cc. alike.
A further parallel for cardinal in place of adverb is BU 1074 (late D.)
trispuqioneiSyll. 3859 Hadrian says
he could not find e]k po h@rcasqe. This is a fairly close parallel to
the e!wj po
If it "may be quotable from early Greek," I cannot quite see why it is for
Dr Nestle "a Hebraism, even if it is still used by Palls in his MGr translation."
I seem to hear the shade of Hadrian demanding "Am I a Jew?"—P. 102.
BU 1079 (41 A. D. ) ble
n a]po> tw?n ]Ioudaithe Jews (i.e. moneylenders)," contains an idiom which the Hebraists will
hardly care to claim now!—P. 103. Fresh exx. of e]n accumulate in a great
variety of meanings. Amongst them I have only room for the Delphian inscr.,
Syll. 8508 (iii/B.C.) kriqejudges," a good illustration of e]n in Ac 1731.
CHAPTER VI.
THE VERB: TENSES AND MODES OF ACTION.
OUR first subject under the Verb will be one which has
not yet achieved an entrance into the grammars. For
the last few years the comparative philologists—mostly in
“Aktionsart.” Germany—have been busily investigating
the problems of Aktionsart, or the "kind of
action" denoted by different verbal formations. The subject,
complex in itself, has unfortunately been entangled not a
little by inconsistent terminology; but it must be studied by
all who wish to understand the rationale of the use of the
Tenses, and the extremely important part which Compound
Verbs play in the Greek and other Indo-Germanic languages.
The English student may be referred to pp. 477 ff. of Dr P.
Giles's admirable Manual of Comparative Philology, ed. 2.
A fuller summary may be found in pp. 471 of Karl Brug-
mann's Griech. Gramm., ed. 3, where the great philologist sets
forth the results of Delbruck and other pioneers in compara-
tive syntax, with an authority and lucidity all his own.
Conjugation The student of Hebrew will not need
and Tense telling that a Tense-system, dividing verbal
Stems. action into the familiar categories of Past,
Present and Future, is by no means so
necessary to language as we once conceived it to be. It
may be more of a surprise to be told that in our own
family of languages Tense is proved by scientific inquiry to
be relatively a late invention, so much so that the elementary
distinction between Past and Present had only been developed
to a rudimentary extent when the various branches of the
family separated so that they ceased to be mutually intel-
ligible. As the language then possessed no Passive whatever,
and no distinct Future, it will be realised that its resources
108
THE VERB: TENSES AND MODES OF ACTION. 109
needed not a little supplementing. But if they were scanty
in one direction, they were superabundant in another. Brug-
mann distinguishes no less than twenty-three conjugations,
or present-stem classes, of which traces remain in Greek;
and there are others preserved in other languages. We
must add the aorists and perfect as formations essentially
parallel. In most of these we are able to detect an
Aktionsart originally appropriate to the conjugation, though
naturally blurred by later developments. It is seen that the
Point Action; Aorist has a "punctiliar" action,1 that is, it
regards action as a point: it represents the
point of entrance (Ingressive, as balei?n "let fly," basileu?sai
"come to the throne"), or that of completion (Effective, as
balei?n "hit"), or it looks at a whole action simply as having
occurred, without distinguishing any steps in its progress
(Constative,2 as basileu?sai "reign," or as when a sculptor
says of his statue, e]poi
|