Moulton (Gr. 98) gives the statistics for the relative frequency of prepositions in the N.T. For every 100 times that evn occurs he finds the relative frequency of the prepositions with which we are here concerned as follows: eivj( 64; evk( 34; evpi,( 32; pro,j( 25; dia,( 24; avpo,( 24; kata,( 17; meta,( 17; u`po,( 8) Calculating J in the same way (though the numbers are to be taken as only approximately correct): evn( 100; eivj( 83; evk( 73; pro,j( 45; dia,( 26; meta,( 25; avpo,( 18; evpi,( 16; kata,( 4) Here we observe that evk is nearly as frequent as eivj, that evpi, is half as frequent as it is normally throughout the N.T. In fat the numbers vary in every case. A comparision of the numbers (which are only approximately trustworthy) in our author is instructive: evn( 100; evpi,( 89; evk( 87; ei,j( 49; meta,( 33; avpo,( 23; dia,( 11; kata,( 5; pro,j( 5.⁽¹⁾ Here the most notable differences are in the case of evpi, (J^ap 89 - J 16), dia,{cxxviii} (J^ap 11 - J 26), pro,j (Jap 5 - J 45). Also the order of priority in frequency is very different. In the three classical historians (Herodotus, Thucydides, and Xenophon), according to Helbing (quoted by Moulton, 62 n) eivj slightly exceeds evn in frequency, whereas in twelve writers of literary koinh, it occurs nearly twice as often as eivj, while the koinh, uses eivj nearly twice as often as koinh,) On the other hand, our author approximates closely to the koinh, in his frequent use of evpi,, and therein diverges strongly from the rest of the N.T. See also Robertson, Gr. 556 sq. But these differences between J^ap and J are not half so striking as those that emerge in the individual treatment of the prepositions.

(i.) avna, = "apiece," in 4⁸ avna. pte,rugaj e[x) Cf. J 2⁶. Found also in Matthew and Luke. The phrase avna. ei-j e[kastoj in 21²¹. These latter uses not in J.

(ii.) avpo,. 36 times. (a) with makro,qen, 18^10.15.17 (source). Not in J.

(b) = "at a distance from," 14²⁰avpo. stadi,wn, cf. J 11¹⁸ 21⁸. Not elsewhere in the N.T. It is not necessary to explain it as a Latinism; cf. Moulton, Gr. 101 sq.; Robertson, Gr. 575; Abbott, Gr. 227. It is found in Strabo, Diodorus, and Plutarch. For an analogous construction with meta,, cf. Test. Reub. 1² meta. e;th du,o th/j teleuth/j: T. Zeb. 1¹ meta. ou=n du,o e;th tou/ qana,tou -- a construction also found in Plutarch. And with pro,, cf. J 12¹, Amos (o ,) 1¹ 4⁷.

(c) avpo. prosw,pow) This phrase occurs three times, 6¹⁶ 12¹⁴ 20¹¹. In the last instance, however, it has a strange form, avpo. tou/ prosw,pou, to which we shall return presently. In all three cases the phrase is the equivalent of ynpm. In 6¹⁶ 20¹¹ it = "from the presence of." It could b e taken in this sense also in 12¹⁴ if it is  connected with pe,thtai, but the fact that sixteen words intervene is against this explanation in our author. Hence the phrase, owing to the Hebrew it presupposes = "because of." The woman's stay of three and a half years in the wilderness is "owing to" or "because of the serpent." This is an ordinary meaning of ynpm in Hebrew. avpo, alone is used in this sense in Matt 18⁷. In 20¹¹ the art. in avpo. tou/ prosw,pou is quite exceptional. It appears only a few (three or more) times in the o , so far as I am aware, and in two of these some MSS omit it. In our text also 046 and many cursives omit. But since Aa 025. 2040 attest it, it goes back to the archetype as edited by the Seer's disciple. For two other departures from the Seer's usage in 20^4-11, see vol. ii. 182. This phrase is absent from J. 

{cxxix} (d) Abnormal use of avpo, before o` w;n. This is deliberate on our author's part.

(e) After passive verbs: avpekta,nqhsan, 9¹⁸; h`toimasme,non, 12⁶. This came to be the rule in later writers.

(f) After avpe,rcesqai and avpollu,nai, 18¹⁴: avfairei/n, 22¹⁹: kru,ptein, 6¹⁶ (avpo. prosw,pou, where J 12³⁶ has simply avpo,): feu,gein, 9⁶ 20¹¹ (J 10⁵).
 

(iii.) a;cri. 2^10.26 12¹¹ 14²⁹ 18⁵ (source).

(iv.) dia,. (a) with gen. 1¹ 21²⁴. In J 15 times. (b) With acc. 16 times and 45 in J. 

(v.) eivj. eivj follows ba,llein when the noun after eivj is not a person, cf. 2^10.22 8^5[7.8] 12^4.9.13 14^{19 (bis)} 18²¹ 20^3.10.14.15, save in 14¹⁶ (interpolated) where we have ba,llein ) ) ) evpi. t) gh/n) Contrast 14¹⁹. But evpi, when the noun is a person, cf. 2²⁴ ba,llw evfV u`ma/j (cf. 1¹⁷). Similarly after katabai,nein we have eivj th.n gh/n, 13¹³, but evpi. tou.j avnqrw,pouj, 16²¹. Our author uses either eivj th.n gh/n, 5⁶ 6¹³ 8⁷ 9^1.3 12^4.9.13 14¹⁹ 16^1.2 etc., even after pi,ptein, 6¹³ 9¹, though this verb in other phrases is followed by evpi,, 6¹⁶ 7¹¹ [8¹⁰] 11¹⁶, or evpi. th/j gh/j (see on evpi, below). eivj occurs about 78 times.

(vi.) evk) This preposition is of very frequent occurrence -- about 135 times.

(a) Partitive Genitive. As subject, 1¹⁹ble,pousin evk tw/n law/n: cf. J 7⁴⁰ 16¹⁷. As object, 2¹⁰ evx u`mw/n( 3<sup>9</sup> 5<sup>9</sup> (in 2<sup>17</sup> we have genitive alone -- tou/ ma,nna: cf. 2 J<sup>4</sup>evk tw/n te,knwn). evk occurs often after ei-j in a partitive sense: cf. 5<sup>5</sup> 6<sup>1</sup> 7<sup>13</sup> etc., but in 17<sup>11</sup> (source) evk tw/n e`pta, = "one of the seven." For ei-j evk, cf. J 1⁴¹ 6^8.70.71 7⁵⁰ etc. This appears to be the best explanation of 2⁹ th.n blasfhmi,an evk tw/n lego,ntwn,⁽²⁾ "the blasphemy of certain people who say"; or the evk may be simply a sign of the genitive. Hence "the blasphemy of," etc.: cf J 3¹ a;nqrwpoj evk t) Farisai,wn: or better, Aesch. Eum. 344, u[mnoj evx VErinu,wn, "hymn of the Erinyes"; Soph. Ant. 95, h` evx evmou/ dusbouli,a.

(b) evk ) ) ) avpo,, 3¹² 21^2.10, where the prepositions may signify respectively heavenly origin and divine mission. But in J 1⁴⁴ 7^41.42 11¹ (Abbott, Gr. 227 sqq.) these mean respectively "native of" and "resident in." 

(c) evk follows a variety of verbs, gemi,zein( evkporeu,esqai( evkdikei/n (involving a Hebraism), exalei,fein( evxe,rcesqai( e;rcesqai( kri,nein (18²⁰ (a source) involving a Hebraism), lamba,nein( lu,ein( metanoei/n,⁽³⁾{cxxx}pi,nein( poti,zein( fagei/n( corta,zesqai) It follows avgora,zein( 5⁹; but this verb is followed by para,, 3¹⁸, and avpo,, 14^3.4. In 18^3.19 ploutei/n is followed by evk and in 18¹⁵ by avpo,.

(d) evk is used after a passive: cf. 13¹⁸ pepurwme,non evk puro,j, 18¹ evfwti,sqh evk t) do,xhj auvtou/)

(e) evk = "by reason of," 8¹³ evk t) fwnw/n, 16¹¹ evblasfh,mhsan ) ) ) evk t) po,nwn auvtw/n)

(f) evk is used with the material of which anything is formed: cf. 18¹²pa/n skeu/oj evk xu,lou. This usage is common to Greek and Hebrew: cf. Xen. Symp. 8, stra,teuma evx evrastw/n: Aesch. Suppl. 9⁵³, evk kriqw/n me,qu) See (a) above ad fin.

(viii.) evn. This preposition occurs nearly 157 times. (a) The most noteworthy use of evn in our author is its instrumental use. Thus it occurs 33 times, whereas it does not occur at all in J (save in a quasi-instrumental sense in the phrase evn tou,tw|: see Abbott, Gr. 256), nor yet in the Pauline or Catholic Epp. save once in 2 Pet. It is found 34 times in Acts, and 3 in Hebrews. Moulton (Gr., pp. 12, 61, 104) thinks that the publication of the Tebtunis Papyri (1902) has "rescued the instrumental evn from the class of Hebraisms" in the case of evn macairh|, Lk 22⁴⁹, and evn r`a,bdw|, 1 Cor. 4²¹. To this claim Abbott (Gr. 256 n.) rejoins effectively. But even though the instrumental evn does occur in the papyri sporadically (where the influence of Jewish traders may have been at work), this fact cannot account in any case for the preponderating use of evn in our author. No adequate explanation can be found save in its origination in a mind steeped in Semitic. Even Moulton (p. 61 n.) concedes that this evn "came to be used rather excessively . . . by men whose mother tongue was Aramaic." But this concession in the case of our author is quite inadequate. evn is used instrumentally after avgora,zein( 5⁹: avdikei/n( 9¹⁹: avpoktei,nein( 2²³ 6⁸ 9²⁰ 13^{10 (bis)} 19²¹: basani,zein( 14¹⁰: kai,ein( 19²⁰; but without evn, [8⁸] 21⁸ (due to editor?): katakai,ein( 17¹⁶ 18⁸: kaumati,zein, 16⁸: khru,ssein( 5²: kiqari,zein( 14²: leukai,nein, 7¹⁴: lu,ein( 1⁵: mignu,nai( 8⁷: pata,ssein( 11⁶ 19¹⁵: plana/n, 19²⁰ 18²³: periba,llesqai( 3⁵ 4⁴ (> evn, A): poimai,nein, 2²⁷ 12⁵ 19¹⁵: polemei/n( 2¹⁶ (19¹¹): crusou/n, 18¹⁶. evn is used locally after kaqi,zein in 3^{21 (bis)} (but evpi, c. acc. 20⁴): {cxxxi} after katoikei/n, 13¹² (but this is not our author's use. He uses evpi, c. gen.).

(b) evn is used temporarily in 1¹⁰ 2¹³ 9⁶ 10⁷ 11¹³ etc.: see temporal phrases without evn in 18^10.16.19 mia/| w[ra| (source).

(c) evn is used generally after gra,fein, 1³ 13⁸ 20^12.15 21²⁷ 22^18.19 (but eivj is found in 1¹¹, and evpi, 17⁸: see under evpi,).

(d) evn is found in the phrases evn th/| dexia/| ceiri,( 1¹⁶: evn th/| dexia/|( 2¹: evn t) ceiri,( 6⁵ 7⁹ 10² etc.; but evpi. th.n dexia,n, 5¹. Also in evn fwnh/| mega,lh|, after le,gein, 14^7.9 (but without evn in 512 813). evn is never used in this phrase after kra,zein, 6¹⁰ 7² 10³ (see vol. i. 260 ad fin., ii. 22 ad init.) except in passages from another hand or source, 14¹⁵ 18². It is also omitted in this phrase after fwnei/n, 14¹⁸. evn me,sw| is always followed by gen. 1¹³ 2¹ 4⁶ etc.; hence 2⁷evn me,sw| tw/| paradei,sw| in a^cc 025 is either a conflation of two texts or a correction of the later.

(x.) e;xwqen, 14²⁰.

(xi.) evpa,nw) Only twice. Really an adverb but  used as a preposition, 6⁸ 20³.

(xii.) evpi,. About 143 times⁽⁴⁾ in all (74 with acc., 13 with dat., 56 with gen.). This preposition is used very idiomatically by our author, and several of the uses are of his own devising. It is therefore of primary importance to be acquainted with these.

(a) evpi, in various phrases:

(a) evpi. th/j gh/j, 5^3.10.13 7¹ 10^2.5.8 etc. -- never evpi. th.n gh/n (for 14¹⁶ is an interpolation). If our author wishes to use gh/n he writes eivj th.n gh/n, 5⁶ 6¹³ 8⁵ 9¹ etc. See vol. i. 191.
(b) evpi. th/j qala,sshj -- so always. 5^13(*) 7¹ 10^2.5.8 except in 15², where the evpi. th.n qa,lassan seems due to its  being preceded by i`sta,nai, which always in the case of other nouns is followed by evpi, with the acc. See vol. i. 262 ad med., ii. 34 ad init. Our author's use comes out forcibly in 7<sup>1</sup> i[na mh. pne,h| a;nemoj evpi. th/j gh/j mh,te evpi. th/j qala,sshj mh,te evpi. pa/n (a 025: cf. 7<sup>16</sup> ouvde. mh. ) ) ) pa/n kau/ma: 9<sup>4</sup> 21<sup>27</sup>) de,ndron) Observe the evpi, with the acc. at the close.
(g) evpi. th.n $ta.j% kefalh,n $-a,j). Only in 12<sup>1</sup> do we find evpi. th/j kefalh/j. See vol. i. 300 sq., 303.
(d% evpi. to. me,twpon, or {cxxxii} if he uses the pl. evpi. tw/n metw,pwn. See vol. i. 206 ad med. In 14<sup>9</sup> we find // evpi. tou/ metw,pou/ //; but this verse is corrupt. See vol. ii. 15 ad fin.
(e) The above forms are rigid. But in phrases composed of evpi, or h` dexi,a our author uses the gen. or acc.: cf. evpi. th/j ceiro.j auvtw/n th/j dexia/j 13¹⁶, evpi. th/j dexia/j 1²⁰, and evpi. th.n cei/ra, 14⁹ 20^1.4: evpi. th.n dexia,n, 5¹. See vol. i. 335 ad med.

(b) evpi, with some case of qro,noj (or nefe,lh) determined by the case of the preceding participle kaqh,menoj. This is one of the most remarkable idiosyncrasies of our author. When the part. is in the nom. or acc. it is followed by evpi. to.n qro,nou: when in the dat. by evpi. tw/| qro,nw|.⁽⁵⁾

 

$a% o` kaqh,menoj  to.n kaqh,menon

evpi. to.n qro,non (or evpi. th.n nefe,lhn) (or evpi. to.n i[ppon).

So in 4^2.4 6^2.5 11¹⁶ 14¹⁴ 19¹¹. This usage of our author is generally not observed in the interpolations or edited portions. Thus 9¹⁷ t) kaqhme,nouj evpV // auvtw/n // seems due to a reviser of the preceding words: 14¹⁶ o` kaqh,menoj evpi. t) nefe,lhj (Aa: t) nefe,lhn, C 025) occurs in the interpolation 14<sup>15-17</sup>: 20<sup>11</sup> to.n kaqh,menon evpV auvtou/ (A: evpa,nw auvtou/, a), and 7<sup>15</sup>o` kaqh,menoj evpi. // t) qro,nou // (Aa: tw/| qro,nw|, 025. 046), are due to the editor of 20⁴-22. 215 o` kaqh,menoj evpi. t) qro,nw|, is a primitive corruption. On 14⁹ see vol. ii. 12.

(b) tw/| kaqhme,nw| evpi. tw/| qro,nw|. So 4⁹ 5¹³ 7¹⁰ 19⁴. In 6⁴ tw/| kaq) evpV // auvto,n // is a primitive corruption, while tw/| kaq) evpi. t) nefe,lhj occurs in the interpolation, 14^15-17.

(g% tou/ kaqhme,nou evpi. tou/ qro,nou. So 4¹⁰ 5^1.7 6¹⁶: cf. 17¹ (th/j kaqhme,nhj evpi. u`da,twn) 19^19.21 (tou/ kaqhme,nou evpi. tou/ i[ppou both times). Hence 19¹⁸tw/n kaqhme,nwn evpV // auvtoi/j // (A: auvtou,ja) seems to be a primitive corruption. 025. 046 and cursives read rightly evpV auvtw/n) These MSS may have preserved the original reading here, and A may be corrupt.

(c) evpi, is used after certain verbs.

(a% ba,llein evpi,with acc. 2²⁴ 18¹⁹ (source)

(b% gra,fein evpi, with acc. 2¹⁷ 3¹² 17^5.8 (source) 19¹⁶. In 14¹ the gen. evpi. tw/n metw,pwn after gra,fein is due to our author's predilection for the gen. pl. in this phrase: see under (a) above.

$g% evkce,ein evpi, with acc. 16^8.10.12.17.

{cxxxiii}$d% i`sta,nai evpi, with acc. 3<sup>20</sup> e[sthka evpi. th.n qu,ran (contrast J 18<sup>16</sup> ei`sth,kei pro.j th/| qu,ra|), 7¹ 8³ 11¹¹ 12¹⁸ 14¹ 15².

(e% kaqi,zein evpi, with acc. 20⁴.
 

(z% katoikei/n evpi, with gen. See vol. i. 289, 336, ii. 12 ad fin. This construction is characteristic alike as to meaning and form. Two other constructions are found in 13¹² 17² where they appear due to sources.

(h) ko,ptesqai evpi, with acc. 1⁷ = "to wail because of" (but in Zech. 12¹⁰ (o ,), 2 Sam. 11²⁶ (A) "to wail for"). So far as I am aware this usage is not Greek. wyl[ dps could be rendered "wail over him," as in Zech. 12¹⁰, or "wail because of him," as the text requires here. Has our author assigned to evpi, a meaning that belongs only to l[? We could also render the Greek "to wail in regard to him." In 18⁹ this phrase = "to wail over."

(q% pi,ptein evpi, with acc. 6¹⁶ 7^11.16 8¹⁰ 11^11.16, but with eivj th.n gh/n, 6¹³ 9¹, since our author does not say evpi. th.n gh/n (see (a) above).

(i% skhnou/n evpi, with acc. 7¹⁵.

(k% tiqe,nai evpi, with acc. 1¹⁷, but in 10² with evpi. th/j qala,sshj in conformity with his usage (see (a) above).

(l) marturei/n and profhteu,ein are followed by evpi, (= "concerning") with dat. in 22¹⁶ (a 046) 10¹¹. evpi, has this meaning in J 12¹⁶ evpV auvtw/| gegramme,na. But in 22¹⁶ A vg bo read evn. See evpi, with dat. after dede,sqai, 9¹⁴; ovrgi,zesqai, 12¹⁷; euvfrai,nesqai, 18²⁰.

(d) After evxousi,aevpi, there follows sometimes the gen. 2²⁶ 11^6b (source) 14¹⁸ 20⁶: sometimes the acc. 6⁸ 13⁷ 16⁹ 22¹⁴. J has neither of these constructions, but the gen. without evpi,, 17², or the inf. 1¹² 5²⁷ 10^{18 (bis)} etc. A similar usage occurs in 17¹⁸ basilei,an evpi. (= "over") tw/n basile,wn: cf. Rom 9⁵.

(xiv.) kuklo,qen as a prep. in 4^3.4: as an adv. in 4⁸.

(xv.) ku,klw| as a prep. 4⁶ 5¹¹ 7¹¹.

(xvi.) meta,. 52 times (41 with gen. and 11 with acc.). (a) meta, with gen. after avkolouqei/n [6⁸] 14¹³ (= "to accompany"): deipnei/n, 3²⁰: e;rcesqai $meta. tw/n nefelw/n), 1⁷: kaqi,zein, 3^{21 (bis)}: lalei/n, 1¹² 10⁸ 17¹ 21^9.15: moiceu,ein, 2²²: [molu,nesqai, 14⁴]: poih/sai po,lemon, 11⁷ 12¹⁷ 13⁷ 19¹⁹: polemei/n, 2¹⁶ 12⁷ 13⁴ 17¹⁴ -- a decided Hebraism, only in our author in the N.T. An occasional instance of it has been found in the papyri: porneu,ein, 17² 18^3.9 (source). This construction is not classical Greek, which requires the acc. So also moiceu,ein.⁽⁶⁾

(b) meta, with acc. is only found in the phrase meta. tau/ta, except in 11¹¹ meta. ta.j trei/j {cxxxiv}h`me,raj. meta. tau/ta has two meanings in our author -- its ordinary one, "after these things," 1¹⁹ 4² 9¹² 20³, and a technical one, which, when combined with ei=don, always introduces a new and important vision, 4¹ 7^1.9 15⁵ 18¹ 19¹. On the value of this phrase as a canon of criticism, see vol. i. 106, footnote. This usage is found in J: (cf. 2¹² 3²² 4⁴³ 5¹ 6¹ 7¹ 19³⁸) as introducing a new section.

(xviii.) pro,j. 8 times (1 with dat. and 7 with acc.). In J, on the other hand, pro,j with acc. occurs about 100 times, and with the dat. 4. pro,j c. dat. is found in our author only once, 1¹³; elsewhere in N.T., Mark 5¹¹, J 18¹⁶ 20^{11.12 (bis)}. He uses pro,j with acc. after verbs of motion, 3²⁰ 10⁹ etc. (6 times). pro,j = "against," in 13⁶ h;noixen to. sto,ma auvtou/ eivj blasfhmi,aj pro.j t) qeo,n. Here eivj would be more natural: cf. Mark 3²⁹, Luke 12¹⁰, Acts 6¹¹. This preposition is much more varied in meaning in J.

(xix.) u`po,. Only twice, and one of these in an interpolation, 6⁸.

(xx.) u`poka,tw) 4 times. Really an adverb but used as a preposition.
 

NOTES
1. These numbers refer to the entire text, including sources and interpolations.
2. This phrase is explained also as "blasphemy arising from" (cf. J 3²⁵); but in our author we should expect in this case blasfhmi,an th.n evk) In 6⁴ the evk is rightly omitted by A after th.n eivrh,nhn [evk] th/j gh/j) If the evk is retained it is to be taken with labei/n, as in 5⁷ 10¹⁰ 18⁴ (source).
3. Cf, 2^21.[22] 9^20.21 16¹¹. metanoei/n avpo, is found in Acts 8²² and Jer 8⁶ (LXX). But menanoei/n evk does not occur in the LXX. It probably represents !~ bwv in our author's mind.
4. These numbers are only approximately true. Different texts yield different results.
*. The context would suggest here the rendering "in the sea." Such was the view of many of the ancients. Thus a reads evn th/| qala,ssh|, and is supported by Pr gig vg s^1.2 arm bo eth.
5. It is noteworthy that this participle in the nom. and acc. is followed by evpi, with the acc. in five passages of the six where it occurs in the rest of the N.T., Matt 9⁹, Mark 2¹⁴, Luke 5²⁷ 21³⁵, J 12¹⁵: exceptions, Acts 8; and that when it is in the gen. it is followed by evpi, with the gen. in Matt 24³ 27¹⁹: exception, Mark 13³. But whereas these may be coincidences, in our author the use is a law. In Mark 13³ we have kaqhme,nou followed by eivj, whereas Matt 24³ has evpi. t) o;rouj t) evlaiw/n.
6. Perhaps we might trace it to such an expression as that in Is. 23¹ #rah twklmm lk ta htnz. porneu,ein meta, is found in Ezek. 16³⁴, but the Hebrew does not explain the meta,. Similarly @an (= moiceu,ein) is followed by ta (= meta,) in Jer. 29²³: but not o ,, which gives evmoicw/nto ta.j gunai/kaj)
 
 

Scanned and edited by Brad Johnson