177

CHAPTER VI

ORTHOGRAPHY AND PHONETICS

The term orthography is used to include all that pertains to the spelling of Greek words. Phonetics deals with the sounds of the letters. The orthography was constantly changing, but not so rapidly as did the sounds. Each had an independent development as is seen very strikingly in the modern Greek vernacular (Thumb, Handbook of the Mod. Gk. Vernac., p. 6). There has never been a fixed orthography for the Greek tongue at any stage of its history. There has always been an effort to have new phonetic spelling to correspond to the sound-change. Cf. Blass, Gr. of N. T. Gk., p. 6. The confusion in spelling grew with the centuries as in English. Many delicate questions confront us at once. It has not seemed possible to give the explanation of all the varied phonetic (true or merely analogical) and orthographic changes in the use of the vowels and consonants. An orderly collection of the facts with historical side-lights is all that is attempted.

I. The Uncertainty of the Evidence. It is difficult to tell what is the vernacular usage in N. T. times on many points, though somewhat less so since the discovery of the papyri.

(a) THE ANCIENT LITERARY SPELLING. The difficulty is much increased by the comparison of the phonetic spelling of the modern vernacular with the historical orthography of the ancient literary Greek.1 This method applied to any language may lead one into error. Modern conversational English differs widely in orthography from Spenser's Faerie Queene. For most of the history of the Greek language no lexicons or grammars were in use. There were the schools and the books on the one hand and popular usage on the other. The movement of the Atticists was just the opposite of the modern phonetic spelling movement in English. The Atticists sought to check change rather than hasten it. It is to be remembered also that the Atticists were the cloister

178 A GRAMMAR OF THE GREEK NEW TESTAMENT

copyists of the ancient Greek writings and of the N. T. Later copyists reflect local types, some more conservative, some less so. The law of life is best here, as always, without artificial impulse or restraint. In seeking to restore the orthography of the vernacular of the first century A.D. one must not be handicapped by the literary Attic nor the modern Greek vernacular, though each will be of service. In simple truth one has to be less dogmatic these days concerning what could or could not have been in the past. Breasted2 calmly assures us that before 3000 B.C. "the alphabetic signs, each of which stood for one consonant," were in use in Egypt. He adds: "Had the Egyptian been less a creature of habit, he might have discarded his syllabic signs 3500 years before Christ, and have written with an alphabet of 24 letters." The Greek language was a growth and did not at first have 24 letters. E, even in early Attic,3 not to mention Cretan, had the force of e, h and sometimes ei. Indeed Jannaris4 asserts that "the symbols h and w, in numerous cases also i, originated at school as mere compensatory marks, to represent positional or 'thetic' e or o." It is not surprising with this origin of vowels (and consonants do not differ) that variations always exist in the sound and use of the Greek letters. Blass5 is clearly right when he points out that in changes in the sounds of words "it is usual for the spelling not to imitate the new sound off-hand," and in the case of the N. T. writers there was "no one fixed orthography in existence, but writers fluctuated between the old historical spelling and a new phonetic manner of writing." Moulton6 adds that the N. T. writers had to choose "between the literary and illiterate Greek of their time," and "an artificial orthography left the door open for not a few uncertainties." Here is a "letter of a prodigal son" (B.G.U. 846 ii/A.D. See Milligan, Gk. Papyri, p. 93 f.) in which we have "phonetic" spelling in abundance: Kai. dia. pa,ntw[ n] eu;comai, sai u`geiai,nein. To. prosku,nhma, sou [ poi] w/ kat v ai`ka,sthn h`mai,ran para. tw|/ kuri,w| [ Ser] a,peidei) Geinw,skein sai qe,lw ktl) There is here interchange of e and ai, of i and ei.

(b) THE DIALECT-COLOURED VERNACULAR. The dialects explain some variations in orthography. One copyist would be a better representative of the pure vernacular koinh, while another might

ORTHOGRAPHY AND PHONETICS 179

live where Attic, Ionic, Doric or Northwest Greek had still positive influence. Often what looks like a breaking-down of the language is but the survival or revival of old dialectical forms or pronunciation. But these variations are mainly due to the personal equation. It was not till the time of Marcus Aurelius that the learned grammarians succeeded in formulating the artificial rules which afterwards prevailed for writing the old classical Greek. The first century A.D. was still an age of freedom in orthography. Even in the fourth century A.D. the scribe of a prefers i rather than ei, while in the case of B ei often occurs where i, is the rule elsewhere. This is not mere itacism, but is also individual preference.7 "The oldest scribes whose work we possess (centuries 4 to 6) always kept themselves much freer from the schools than the later."8 But, even if Luke and Paul did not know the old historical spelling in the case of i mute (subscript) and ei, it is merely cutting the Gordian knot to "follow the Byzantine school, and consistently employ the historical spelling in the N. T." and that "without any regard to the MS. evidence." It is not the spelling of the Byzantine school nor of the Attic dialect that we are after, but the vernacular Greek of the first century A.D., and this is not quite "the most unprofitable of tasks," as Blass would have us believe.9

(c) THE UNCIALS. They do complicate the situation. On some points, as noted above, the great uncials a and B differ, but usually that is not true. There is a general agreement between the older uncials in orthography as against the later uncials and the cursives which fell under the spell of the Byzantine reformers, who sought to restore the classical literary spelling. The Syrian class of documents therefore fails to represent the orthography of

180 A GRAMMAR OF THE GREEK NEW TESTAMENT

Addenda 2nd ed.

Addenda 2nd ed.

the vernacular koinh, of the first century A.D. The Syrian class, for instance, reads Karernaou,m, not Kafarnaou,m. But do the MSS. which give us the pre-Syrian types of text preserve the autographic orthography? The fourth century is a long time from the first and the presumption might seem to be to some extent against the Neutral, Alexandrian and Western classes also. The temptation is constant to spell as people of one's time do. This difficulty is felt by every editor of classical Greek texts and often purely arbitrary rules are used, rules made by modern critics. Hort10 is willing to admit that in some instances the spellings found in the great uncials which are at variance with the Textus Receptus are due to the "literary spellings of the time" when the MSS. were written, "but for the most part they belong to the 'vulgar' or popular form of the language." Hort could see that before we had the new knowledge from the papyri and inscriptions. He adds11: "A large proportion of the peculiar spellings of the N. T. are simply spellings of common life. In most cases either identical or analogous spellings occur frequently in inscriptions written in different countries, by no means always of the more illiterate sort." This fact showed that the unclassical spellings in the uncials were current in the Apostolic age and were the most trustworthy even if sometimes doubtful. "Absolute uniformity belongs only to artificial times," Hort12 argues, and hence it is not strange to find this confusion in the MSS. The confusion existed in fact in the first century A.D. and probably the autographs did not follow uniform rules in spelling. Certain it is that the N. T. writings as preserved in the MSS. vary. But itacism applies to all the MSS. to a certain extent and makes it difficult to know what vowel or diphthong was really before the scribe. In general the N. T., like the LXX, is grounded in matters of orthography on the rules of the grammarians of the time of the Caesars (Apollonius and Herodian) rather than upon those of the time of Hadrian, when they had an archaistic or Atticistic tendency (Helbing, Grammatik d. LKX, p. 1). Moulton (Prol., p. 42) thinks that "there are some suggestive signs that the great uncials, in this respect as in others, are not far away from the autographs." But Thackeray (op. cit., p. 56) denies that this

ORTHOGRAPHY AND PHONETICS 181

conclusion can be drawn ipso facto of the LXX, since it was translated (the Pentateuch certainly) some three centuries earlier than the N. T. was written.

(d) THE PAPYRI. They strengthen the case for the uncials. Deissmann13 and Moulton14 show that the great uncials correspond in orthography not only with the contemporaneous inscriptions as Hort had seen, but also with the papyri of the better-educated writers. Among the strictly illiterate papyri writers one can find almost anything. The case of eva,na;n in relative clauses is worked out well by Moulton to prove this point. In the papyri dated B.C. the proportion of eva,n to a;n in such cases is 13 to 29, while in the first century A.D. it is 76 to 9. But in the fourth century A.D. it is 4 to 8 and the usage disappears in the sixth century A.D. Thackeray (Grammar, vol. I, pp. 65 ff.) shows (after Deissmann15) how the LXX confirms this conclusion for eva,na;n. The usage appears in B.C. 133; copyists are divided in different parts of the same book as in Exodus or Leviticus; it is predominant in the first and second centuries A.D., and then disappears. Thackeray (p. 58) traces ouvqei,j $mhqei,j) "from its cradle to its grave" (from 378 B.C. to end of ii/A.D.) and shows how in ii/A.D. ouvdei,j is supreme again. This point very strikingly confirms the faithfulness of the uncials in orthography in a matter out of harmony with the time when the MSS. were written. We may conclude then that Hort is right and the uncials, inscriptions and papyri give us the vernacular orthography of the koinh, with reasonable correctness.

II. Vowel-Changes ( stoicei/a fwnh,enta). In the old times the vowels underwent many changes, for orthography was not fixed. Indeed is it ever fixed? If the Atticists had let the koinh, have a normal development, Dr. Rutherford would not have complained that Greek was ruined by their persistence "in an obsolete orthography instead of spelling as they speak."16 But as early as 403 B.C. the orator Archinos17 had a law passed in Attica prescribing the use of the Ionic alphabet in the schools. The early Greek used only a e i o u and no distinction was made in writing be-

182 A GRAMMAR OF THE GREEK NEW TESTAMENT

tween long and short vowels, as indeed was never done in "the case of i and u. The Ionic invented18 W for long o. Before the introduction of the Ionic alphabet, I.E. a and e were represented by e. H was at first the aspirate like Hebrew h and then now aspirate and now long e or a as the inscriptions amply show. It is very common in the early inscriptions to see e thus used as long and o likewise, as in e=nai and toj. Cf. e, o for spurious diphthongs ei ou. The kinship of these vowels with the Phoenician alphabet is plain, as a is from a, e from h, i from y, o from [, u from the doubling of w (and so a Greek invention). It is interesting to note that the Sanskrit has three pure vowels, a, i, u, while e and o are diphthongs in origin. In Sanskrit a far surpasses all other vowel-sounds, more than twice as many as all other vowelsounds put together.19 Schleicher20 speaks of the weakening of a into i and u, and thus he, goes back to an original a sound for all the vowels. In Latin also a breaks into e, i and u.21 Even in Attica in the first century B.C., in spite of Archinos' law, the inscriptions use sometimes ai and ae, ei and i, h and i, u and i u and ui i and ei interchangeably.22 Uniformity did not exist in one dialect, not to mention the persistent differences between the various Greek dialects. These changes were going on constantly all over the Greek world in the first century A.D. For the alphabetical changes in the dialects see Buck's Greek Dialects, pp. 15 ff. These interchanges between vowels are interesting.

(a) THE CHANGES (INTERCHANGES) WITH a. The first sound made by a baby is a. These changes became dialectical peculiarities in many words like the Lesbian kre,toj ( kra,toj, "ablaut" variations), the Boeotian a[teroj ( e[teroj), Doric i`aro,j ( i`ero,j).23 So in the vernacular Attic we find , evreth, ( avreth,) where a breaks to e before e (vowel assimilation), as in the Ionic-Attic a sometimes changes to e after i and u.24 See Kuhner-Blass25 for many examples.

ORTHOGRAPHY AND PHONETICS 183

a and e. vAggareu,w appears as evggar. in a (Mt. 5:41) and aB (Mk. 15:21).26 The New Ionic ei[neken (more commonly e[neken% has nearly displaced the Attic e[neka which Blass27 admits only in Ac. 26:21. Ei=ten for ei=ta appears in Mk. 4:28 as a rare Ionic form. Herodotus28 had both ei=ta and e;peita) Kaqari,zw in the aorist (active and passive) and perfect middle has e for the second a in many of the best MSS. both in LXX and N. T. (cf. Mk. 1:42; Mt. 8: 3 W. H.). Gregory, Prolegomena, p. 82, gives the facts. Blass29 points out that Pa,tera $Pa,tara% occurs in AC in Ac. 21:1. Tessera,konta is the form given always by W. H. This is an Ionic form (vowel assimilation) which is not so common in the papyri as in the N. T. MSS.30 In modern Greek both sara,nta and sera,nta survive. Likewise W. H. always give the preference to te,ssera, though the papyri do not use it till the fourth century A.D.31 But in the inscriptions te,ssera is found several times,32 one case in the first century A.D.33 Te,sseraj, however, does not occur in the N. T. MSS., though the papyri have it in the Byzantine age.34 The Ionic and the modern Greek have te,ssarej and te,ssera. The N. T. thus differs from the koinh, papyri, but is in harmony with the Ionic literature and inscriptions. In some MSS. in both LXX and N. T. in Doric and Boeotian, while ge is found in the Ionic, Attic and Cypriote (Meister, Griech. Dial., Bd. II, p. 29).

184 A GRAMMAR OF THE GREEK NEW TESTAMENT

te,ssarej is accusative as well as nominative, like the Achaean dialect, but this is another story. a in Rev. 3:16 has cliero,j. The common (Ionic and Northwest Greek) use of - e,w instead of - a,w with verbs as in evrwte,w will be discussed in the chapter on Verbs.

Conversely e is sometimes changed to a. vAmfia,zei is accepted by W. H. in Lu. 12:28 rather than either the late avmfie,zei or the early avmfie,nnusi. The form evrauna,w instead of evreuna,w W. H. have everywhere received into the text, and so with evxeranuna,w and avnexe rau,nhtoj. aB always read it so, sometimes AC. It is supported by the papyri. Cf. Mayser, Gr., p. 113; Helbing, Gr. d. LXX, p. 7, for similar phenomena in the LXX.

Initial e often becomes a in modern Greek vernacular, as avla fro,j $evlafro,j% a;ntera $e;ntera), etc. Cf. Thumb, Handbook, p. 14. So the Doric pia,zw is used in the N. T. everywhere save in Lu. 6:38, where, however, pepiesme,noj has the original idea ('pressed down,' not 'seized'). Both occur in the LXX. The Attic forms fia,lh u[aloj are retained in the N. T. (as in LXX) rather than the Ionic and vernacular koinh, forms in e, a mark of the influence of the literary35 koinh,.

Some verbs in - e,w also use - aw forms, like evlea,w evlloga,w xura,w) See the chapter on Verbs.

Changes in a take place in a few Hebrew proper names. Kaper naou,m, is the Syrian reading for Kafarnaou,m (W. H.). So W. H. read Maleleh,l in Lu. 3:37, not Mel. (Tisch.), and Naqanah,l. Selaqih,l (instead of Sal.) appears in B. Thumb36 remarks that these changes between a and e occur to-day in the Kappadocian dialect.

a and h) The Doric forms o`dago,j o`dagw/ are found in the koinh,, though Schweizer37 calls it hardly a Dorism. So in N. T. MSS. we have prosace,w in B (Ac. 27:27) and r`a,ssw in D (Mk. 9:18). The Ptolemaic papyri regularly have avnhli,skein till ii/A.D. (Mayser, Gr., p. 345). For a and a| see h and h| under (c).

a and o. The changes38 between these two vowels are seen in the Lesbian uvpa, ( u`po,), Arcadian triaka,sioi, Doric ei;kati $ei;kosi%, etc. W. H. give battaloge,w in Mt. 6:7 (cf. battari,zw) instead of bat tologe,w. ABK and twice a and many cursives have pro.j Kolassaei/j <LE>

ORTHOGRAPHY AND PHONETICS 185

as the title, while in Col. 1:2 nearly all MSS. read evn Kolossai/j. Blass finds the title in o also in accordance with the coins and the profane writers; Xen., Anab. I, 2. 6, has a variant reading in Kolas sai,. In Mk. 13:35 B has mesanu,ktion and D in Lu. 11:5 instead of mesonu,ktion.39 In 1 Tim. 1:9 W. H. give mhtrolw|,aij and patro lw|,aij (instead of - aloi,aij) on the authority of aDFGL. Blass40 compares patrokto,noj.

a and w. vAna,gaion is read by the most and the best MSS. in Mk. 14:15; Lu. 22:12. vAnw,geon avnw,gaion avnw,gewn avna,geon have only "trifling authority."41 Gai/oj is Doric and Ionic.

a and ai. The papyri42 sometimes have the Epic and Ionic aivei,, though the N. T. only reads avei,. The i early dropped out between the vowels. Cf. Mayser, Gr., p. 103. B has aivei, in 1 Esd. 1:30. The N. T., like the LXX, has kai,w and klai,w, though the Ptolemaic papyri rarely have ka,w and kla,w.

a and au. In Lu. 2:1 aC D have vAgou,stou instead of Auvgou,stou. This spelling of a for au is found in Pergamum by Schweizer43 in the reflexive pronoun e`ato,n, while Meisterhans44 gives examples of it as early as 74 B.C. in the Attic inscriptions. Moulton45 is probably correct in saying that we need not assume the existence of this spelling in the N. T. autographs, though it is not impossible. He indorses Mayor's suggestion (Exp., VI, x, 289) "that avkatapa,stouj in 2 Pet. 2:14 AB may be thus explained: he compares avcmhrw|/ 1:19 A." This dropping of u between vowels extended to the dropping of u before consonants. In the modern Greek we have auvto,j (aftos) and avto,j (in Pontus), whence comes to, (not the article).46 The examples of vAgou/stoj and avto,j $avtogen nhto,n, once) in the papyri are very common.47 Thackeray (Gr., p. 79) finds no instances in the LXX.

186 A GRAMMAR OF THE GREEK NEW TESTAMENT

ai and e. ai was written ae in early Boeotian and Attic inscriptions (cf. Latin transliteration) and so gradually was pronounced as e (Jannaris, Hist. Gk. Gr., p. 28). By 100 A.D. in the koinh, ai was the mere equivalent of e. The Egyptian papyri show abundant illustrations of it. Especially do the LXX MSS. exhibit it (Thackeray, Gr., p. 78). The modern Greek pronounces both these vowel-sounds alike, as indeed did the Boeotian dialect long before the koinh,. Numerous examples of this interchange of spelling exist in the Pompeian wall-inscriptions and in the vernacular koinh, from 100 A.D. on.48 Indeed in the N. T. MSS. it is very common to find - sqai and - sqe used indiscriminately, probably representing the common later pronunciation which was already developing in the first century A.D. Hort49 compares this "shortening of an identical sound" to the late stu,loj for stu/loj and kri,ma for kri/ma. So common did this blending become that Blass50 places little confidence in the N. T. MSS. on this point. Such readings occur as evtei/sqe for aivtei/sqe and gune/kaij for gunai/kej. Sometimes only the context51 can decide between e and ai where different forms result, as in avna,pese or - ai (Lu. 14:10), e;geire or - ai (Mt. 9:5), evpa,nagkej (Ac. 15:28),52 e;rcesqe or - sqai in aADL (Lu. 14:17), e`te,roij or e`tai,roij (Mt. 11:16 Syrian reading), pare,negke or - ai (Mk. 14:36), etc. In Gal. 4:18 both a and B read zhlou/sqe for zhlou/sqai. B reads Aivlami/tai in Ac. 2:9, from ~l'y[;, the rest vEl. The authority according to Hort53 is "usually preponderant" for evxe,fnhj and evfni,dioj instead of aivf. So kere,a, for kerai,a is accepted54 in Mt. 5:18; Lu. 16:17, and krepa,lh for kraipa,lh in Lu. 21:34. Likewise W. H. receive Lase,a for lai/lay in Ac. 27:8. aAC in 2 Pet. 2:17 read le,lapoj, but lai/lay is the undoubted reading in Matthew, Luke. The uncials all have r`e,dh, not r`ai,dh, in Rev. 18:13. So all the early uncials but A have Sukomore,a (not - ai,a) in Lu. 19:4. Hort55 accepts also felo,nhj for failo,nhj (2 Tim. 4:13), though Moulton56 doubts, because of the Latin paenula.

ORTHOGRAPHY AND PHONETICS 187

(b) THE CHANGES WITH e. The interchanges of e and a have already been discussed under (a), but others took place with h i o.

e and ei In the Boeotian these were freely interchanged57 and the same interchange occurs in the Doric, New Ionic and Attic as ple,wn, or plei,wn. The Attic inscriptions58 show this common phenomenon. The i before a vowel easily and early loses its force and drops out. Before the adoption of the scholastic orthography at Athens (B.C. 403) e stood for e h ei. Sooner or later ei became everywhere a monophthong (Buck, Greek Dialects, p. 28). But the koinh, usually wrote ei before vowels rather than e (Thackeray, Gr., p. 81). The LXX MSS. reveal the same traits as the N. T. vAreopagi,thj is in Acts 17:34, but ;Areioj occurs (Ac. 17:19, 22). vAcrei/oj is uniform in the N. T., but in Ro. 3:12 we have hvcrew, qhsan ( aABDG). In Lu. 3:13; Jo. 21:15; Ac. 15:28, W. H. print ple,on (Attic has even ple,onoj),59 but elsewhere the N. T. has forms in ei. The derivatives all have e like pleonekte,w. But the N. T. has only te,leioj teleio,w, though Herodotus always and the Attic usually used teleo,w. Dc has telew/sai in Heb. 10:1.60 Of words with e and ei before consonants one may note that avpo stei,lw in Ac. 7:34 is aorist subjunctive. (Cf. Ex. 3:10.) Both e[neken and ei[neken occur in the N. T. (both Ionic and Attic). The N. T. never has evj but always eivj. However, e;sw is the uniform reading in the N. T. Homer used either ei;saw or e;sw.

e and h) Numerous examples of long e occur in the inscriptions like mete $mh,te%.61 These changes are probably all analogical and not phonetic. But in the N. T. we have only the shortening of h, back to short e in some words like avna,qema, though this particular word ('curse') came to be distinct from avna,qhma ('votive offering'). vAna,qhma occurs only once in the N. T. (Lu. 21:5), and even here aADX, etc., have avna,qema. Tisch. quotes Moeris as saying avna, qhma avttikw/j avna,qema e`llhnikw/j. But the use of avna,qema as 'curse'

188 A GRAMMAR OF THE GREEK NEW TESTAMENT

"is not an innovation of biblical Greek" (Moulton, Prolegomena, p. 46). In Ac. 11:11 aABDGr read h=men, not h;mhn. Perhaps this exchange between e and h bears on the use of sth,kete with i[na in Mk. 11:25; 1 Th. 3:8, and of MS. evidence for qauma,zete in Jo. 5:20 and evxomologh,setai in Ph. 2:11. Cf. also o;yhsqe and o;yesqe in Lu. 13:28. So in 13:25. Mayser (Gr., p. 64) thinks that sometimes e represents an original open h as in paresteko,tej. The koinh, shows quite a preference for words in - ema rather than -- hma (Mayser, Gr., p. 65 f.), and the LXX has new words in - ema, though some words have both forms (Thackeray, Gr., p. 80).

In the papyri this shortening (as in the LXX) appears in words like evpi,qema pro,sqema, etc.62 The interchanges between h and ei hi, and ei will be discussed under h (c). Mayser (Gr., p. 63 f.) thus ( h for e) explains plh,rhj as an indeclinable neuter form.

e and i. Dieterich63 mentions as one of the marks of the Attic and Egyptian koinh, the fact that i and e interchange when used with l and n. Cf. the modern Greek, and the Lesbian Greek used te,rtoj for tri,toj, and the Thessalian qio,j for qeo,j. It is a Doric characteristic. This variation appears in the inscriptions64 and in the papyri,65 especially in the case of legiw,n, which is also legew,n and even legeiw,n, not to mention a genitive legio/nwj ( o and w having the same sound). Legiw,n, is the reading of the best N. T. MSS. ( aBDL; cf. Latin legio), as in the papyri. Especially in the case of the Latin short i does the koinh, have e. `Aleei/j, not e`liei/j, is the reading in the N. T. according to the best MSS. (Mk. 1:16, etc.).66 This is a natural assimilation after a liquid. The frequency of e for i in the Egyptian papyri may be due in part to the Coptic, which has no short i (Steindorff, Kopt. Gr., p. 13). Note a soldier's use of ce/ran for cei/ra( n), B.G.U. 423 (ii/A.D.). Le,tion (Jo. 13:4, Latin linteum) is a change in the other direction, Latin i to Greek e. Blass67 says that le,nteon would have looked

ORTHOGRAPHY AND PHONETICS 189

unnatural to a Greek. Nhfa,lioj also is alone well-attested,68 not nhfa,leoj (1 Tim. 3:2, etc.). Poti,oloi in Ac. 28:13 represents the Latin Puteoli, using i for e (cf. Dittenberger, p. 145). Simiki,nqion (not - eon) is the N. T. reading (Ac. 19:12) for Latin semicinctium. So Tibe,rioj (not Tebe,rioj) is the N. T. rendition of Tiberius in Lu. 3:1, though the later Greek writers used Tebe,rioj, Dome,trioj, etc.69 It is really surprising that more examples of this exchange of e and i do not appear. The interchanges between ei and i are discussed under (d), those between eu and u under (f).

e and o. The Lesbian AEolic had stro,fw for the Doric stra,fw. The Ionic-Attic made it stre,fw. Meisterhans70 gives numerous examples of this change in e and o: ovbolo,j for ovbelo,j as early as the middle of the fourth century B.C. Dieterich71 mentions the assimilation of e and o as one of the marks of the Egyptian koinh,. In Ac. 18:24 a 15. 180. Cop. arm. and in 19:1 a 180. read vApellh/j for vApolw,j, though D has vApollw,nioj in 18:24. The Doric and the Attic inscriptions72 had vApe,llw vApellw,nioj vApe,llioj, etc. In 1 Cor. and Titus we have only vApollw,j. Indeed Blass73 suggests that vApellh/j is the reading of the a text in Acts and that vApollw,j is an interpolation from 1 Cor. It is more likely to think that the two old forms of the name were still in use, though vApol lw,j is the correct text in Acts also. The MSS. of the N. T., even good uncials, have ovloqreu,w evxoloqreu,w ovloqreuth,j as well as the usual ovleqreu,w, etc. (cf. ovbolo,j for ovbelo,j by assimilation), and Hort74 accepts the e form only in Ac. 3:23. The Syrian class has the o form. Blass,75 who usually cares little for such points, properly insists on the documentary evidence. In Heb. 11:28 only ADE have the e form, while in 1 Cor. 10:10 DFG read e.

190 A GRAMMAR OF THE GREEK NEW TESTAMENT

Addenda 3rd ed.

The LXX according to aAB reads e, though the modern Greek has xoloqreu,w. But o;leqroj is the uniform spelling in the N. T. and is the rule in the LXX (Thackeray, Gr., p. 88).

In Mk. 8:14 B has evpela,qento as is common in the LXX (Thackeray, Gr., p. 89). Cf. also avpe,deto (Heb. 12:16, LXX), evxe,deto (Mk. 12:1), diedi,deto (Ac. 4:35), paredi,deto (1 Cor. 11:23), and evxekre,meto (Lu. 19:48 aB). Hort (Appendix, p. 167 f.) explains these changes as "euphonic," but it is a change of the rootvowel of do, a confusion of thematic and athematic conjugations.

eva,n and a;n) See also I (d) under Papyri. This is as good a place as any to say a word further on the interchange of these two forms, not strictly vowel-changes, however. We have also eiva,n, (really eiv+ a;n) as in P Eleph. 1 (B.c. 311). See also aiva,n for eva,n B.G.U. 530 (i/A.D.). The use of eva,n= modal a;n in relative sentences, so common in the LXX, N. T. and papyri of i/ii A.D., is not an exchange of vowels, but possibly a slurring over of the e before a. ;Aneva,n survives from the ancient Greek in a few instances, as Jo. 5:19 ( aB); 12:32 (B and accepted by W. H.); 13:20 DEFG, etc., have eva,n, but aBC a;n and accepted by W. H.); 16:23 (BACD, accepted by W. H.); 20:23 (twice and accepted by W. H., though AD have first eva,n and aAD second). In Ac. 9:2 only aE have a;n and W. H. read eva,n) Blass76 thinks that as eva,n made encroachment into the province of a;n "a kind of interchange of meaning between the two words" grew up. The modern Greek vernacular uses a;n for 'if.' Hort77 considers the whole subject of the interchange between eva,n and a;n after relatives "peculiarly irregular and perplexing. Predominantly a;n is found after consonants, and eva,n after vowels, but there are many exceptions." Cf. eva,n in Mt. 20:4 and a;n in Mt. 20:26 f. Moulton78 has shown that eva,n= a;n is scarce in the papyri save from 100 B.C. to 200 A.D. In the Magnesian inscriptions79 only eva,n appears, not a;n nor h;n, h;neva,n is not in the N. T. But in the Herculaneum papyri these particles interchange freely.80 The Attic inscriptions uniformly have a;n with relatives.81

ORTHOGRAPHY AND PHONETICS 191

Indeed Attic often contracts this particle eva,n h;n.82 But modal a;n is found in Xen. Mem., w|- eva.n a`rmo,tth|, in Lysias, oua}j eva.n boulhqw/sin, etc. (see Jannaris, Hist. Gk. Gr., p. 421). This use of eva.n occurs sixty-one times in the N. T. Examples occur in late Greek of eiv- eva,n, as well as eiv - a;n, instead of eva,n. Cf. Reinhold, De Graecitate Patrum Apost. etc., p. 35; Moulton, Classical Review, 1901, p. 32. Thackeray (Gr., pp. 65 ff.) finds that in the ii/B.C. the papyri nearly always have oa}j a;n, while in the i/A.D. they nearly always have oa}j eva,n. In the books of Exodus and Leviticus he notes that in the first half of each book both forms occur while in the second part oa}j eva,n almost vanishes. Each book may have been written on two rolls.

(c) THE CHANGES WITH h. The changes between h and a, h and e have already been discussed.

h and i. As already stated, originally H was merely the rough breathing, but the Ionic psilosis left a symbol useless, and heta was called eta.83 Thus the new letter took the old long e value in Ionic and Attic and also largely supplanted the long a where a became e. The Sanskrit used long a, the Greek h and the Latin either e or i This new (in spelling) h (v/B.c.) gradually turned more to the i sound in harmony with the growing itacism of the language, though there was some etacism on the other hand.84 As early as 150 B.C. the Egyptian papyri show evidence of the use of i for h.85 By the middle of the second century A.D. the confusion between h and i, h and ei, hi and ei is very general. By the Byzantine times it is complete and the itacism is triumphant in the modern Greek.86 Reinhold87 thinks that the exchange between h and i was natural in view of the relation between h and e and the interchange between e and i) As early as the fifth century B.C. the change between h and i is seen on vases and inscriptions. But the Ptolemaic papyri show little of it and it is rare in the LXX MSS. aAB (Thackeray, Gr., p. 85). In the N. T. times the interchanges between h and i, h and ei, hi and ei are not many. In 1 Cor. 4:11 W. H. read gumniteu,w, though L and most of the cursives have h.

192 A GRAMMAR OF THE GREEK NEW TESTAMENT

The N. T. always has dhna,rion, though dina,rion appears very early.88 For ka,mhloj in Mt. 19:24 and Lu. 18:25 a few late cursive MSS. substitute ka,miloj ('rope'), a word found only in Suidas and a scholium on Arist. But "it is certainly wrong,"89 a mere effort to explain away the difficulty in the text, an effort as old as Cyril of Alexandria on Luke. For Kurh,nioj B90 it. vg. sah. have Kuri//noj, while B* has Kurei/noj and A has Khru,nioj, a striking example of itacism, h, i, ei u having the same sound in these MSS. The N. T. MSS. give simiki,nqion in Ads 19:12, but Liddell and Thayer both suggest shm. as an alternative spelling like the Latin semicinctium. So also the best MSS. in Rev. 18:12 read siriko,j, though some cursives have shriko,j (like Jos. and others), and still others suriko,j.91 Indeed in 1 Pet. 2:3 for crhsto,j L and many cursives have Cristo,j. The heathen misunderstood the word Cristo,j and confounded it with the familiar crhsto,j, pronounced much alike. Suetonius (Claudius 25) probably confused Christus with Chrestus. In Ac. 11:26 a 61 have Crhstianou,j, while B has Creist. So in Ac. 26:28 a has Crhstiano,n for Crist) while B has again ei. The same thing occurs in 1 Pet. 4:16.

h and ei. The Boeotian and the Thessalian dialects early changed92 h for ei, ti,qeimiti,qhmi. Schweizer93 gives para,dhsoj for para,deisoj (Byzantine inscription). In Lu. 14:13 (21) we have avna,peiroj (ABDEL), avna,phroj (GHK, etc.), and -- pir- ( aR). This itacism is condemned by Phrynichus the Atticist as vulgar.94 In the LXX a has avna,peiroj in Tob. 14:2 and AV show it in 2 Macc. 8:24 (Thackeray, Gr., p. 83). In Heb. 6:14 W. H. follow aABD in reading ei= mh,n rather than h= mh,n. This form occurs in the LXX and in the papyri. Moulton95 has shown that several times in the papyri it is obviously for h= mh,n by mere itacism, and so is not due to a confusion between the Hebraistic use of eiv mh, = alo mai, thus correcting Hort. The uncials and the

ORTHOGRAPHY AND PHONETICS 193

papyri here agree. Deissmann96 calls attention to the use of eiv ma,n in a Doric inscription of the first century B.C. Blass (Gr. of N. T. Gk., p. 306) observes that a papyrus reads khri,a for keiri,a (cf. Jo. 11:44, keir-, khr-, kir- i,aij).

hi and ei. In the old Attic there was no hi in writing, only ei, since h was not used as a vowel. As early as 400 B.C. the Attic used hi and ei interchangeably, klh|,w becoming klei,w klh|,jklei,j lh|tour go,jleitourgo,j, etc.97 This usage was not very common in Pergamum98 nor in Magnesia.99 Cronert finds this interchange in the Herculaneum papyri only in the papyri copies of Epicurus and Polystratus.100 In the N. T leitourgo,j - i,a, - ei/n iko,j are taken over from the Attic, but they occur also in Pergamum101 and Magnesia.102 The Attic indeed carried the fondness for ei so far that it was used always in writing in the second singular indicative middle everywhere, the other dialects using h| save the Ionic. The koinh, has h| save in bou,lei oi;ei o;yei. In the N. T. h| is universal according to W. H. save in Lu. 22:42 where bou,lei is genuine, though some MSS. have ei in other passages. Blass103 observes that this is a literary touch in Luke for the colloquial qe,leij. Hatzidakis104 notes how difficult this process made it to tell the difference between poih,sh|j and poih,seij, for instance, because of this Attic intermixture of the diphthongs. Blass105 will not hear of this as a possible explanation in any cases, but one must remark how well this vowel-blending harmonized with the kinship in meaning between the aorist subjunctive and the future indicative (cf. dw,sh| in some MSS. for dw,sei in Jo. 17:2) and made it easy for the later so-called future subjunctive (cf. Latin) to develop. WinerSchmiedel indeed accept as possible this vowel confusion in several instances.106 In Mk. 8:35 (Lu. 17:33) oa}j a'n avpole,sei, Lu. 12:8 oa}j a'n o`mologh,sei, 2 Cor. 12:21 mh. tapeinw,sei, Ro. 3:4 (Ps. 51:6)

194 A GRAMMAR OF THE GREEK NEW TESTAMENT

nikh,seij (cf. diakiwqh|/j), Ac. 5:15 i[na evpiskia,sei, 8:31 eva.n o`dhgh,sei. Winer-Schmiedel would find the aorist subjunctive and not the future indicative. This is possible but by no means certain, since the future indicative was undoubtedly used both with eva,n and i[na ( o[pwj). W. H. read vIwa,nei instead of h| in Mt. 11:4 = Lu. 7:18. Tw|/ diokhtei/ occurs in papyri Brit. Mus. I, Nr. 2. 135. In 2 Coy. 2:9 AB 109 have h|- where eiv is probably correct.

h and h|) Irrational Iota. The iota subscript was iota adscript till the twelfth century A.D., but as early as the third century B.C. it was not pronounced.107 When a was practically equal to h in sound, it was natural that h| ( hi) should be. The i was then dropped in sound long before it was subscript.108 Gradually it was felt to be a matter of indifference in some words whether this iota was written or not. Examples of h instead of h| occur in the inscriptions of Pergamum109 as evn h= as well as in the Attic.110 Moulton finds irrational i adscript ( e;cwi, for instance) abundant in the Ptolemaic Tebt. Papyri (Classical Review, 1904, p. 106). Cf. Mayser (Gr., pp. 122-126) who gives many examples. In the N. T. i has dropped from qnh,skw. Indeed since the second century B.C. i adscript in the diphthongs a|, h|, w| had become mute. Hort,111 however, argues for the retention of i in zh|/n6 and infinitives in - a|/n instead of the Doric-Attic form, as well as in avqw|/oj eivkh|/ zw|/on `Hrw|,dhj krufh|/ la,qra| pantach|/ pa,nth| prw|,ra sw|,zw u`perw|/on zw|/on, though he hesitated to put sw|,zw in the text. It is just as well to finish the discussion of the iota subscript here, though some of these examples go beyond the range of h|. The best editors print also shmosi,a| ivdi,a| mhtrolw|,aij patrolw|,aij patrw|/oj pezh|/ Samoqra|,kh Trw|,aj, though mimnh,skw and pra/oj. W. H. have forms in - oi/n also, as kataskhnoi/n (Mt. 13:32). Moulton112 gives a curious example of the loss of the irrational i in the case of the subjunctive h|= which sometimes in the papyri appears as h=n, having lost the i, and taken on irrational n. As a matter of fact iota adscript (iota

ORTHOGRAPHY AND PHONETICS 195

subscript not yet, of course) does not appear in the great uncials save h;idisan in D (Mk. 1:34) and xu,lwi in K (Lu. 23:31).113 Forms with and without the mute iota appear in the Herculaneum papyri,114 as eivkh/i or eivkh/) Blass115 would also restore i to avntipe,ra$a|%. He doubts if i was written in such new optative forms as dw,hn ( doi,hn Attic) though it should be put in the text.

h and u. Since these two vowels came to be pronounced alike as in modern Greek,116 it was to be expected that some interchange would come, though any early examples are wanting. However, by the second century A.D. the inscriptions give many instances such as qh,ra $qu,ra% mhsth,rion $must)% sku/ptron $skh/ptron%, etc.117 It is already in the Egyptian koinh, according to Thumb.118 Hence we are not surprised to see the N. T. MSS. get mixed over h`mei/j and u`mei/j. Especially in 1 Peter does this itacism lead to a mixing of the historical119 standpoint as in 1:12, where u`mi/n is read by aABCL, etc., h`mi/n by K and most cursives Syrsch Cop. In 1 Pet. 5:10 the MSS. similarly support u`ma/j and h`ma/j. In 2 Cor. the personal relations of Paul and his converts are involved in this piece of orthography as in 8:7 evx u`mw/n evn h`mi/n ( aCDE, etc.) or evx h`mw/n evn u`mi/n (B 30, 31, 37, etc.). See especially kaq v h`ma/j in Ac. 17:28 (B 33 Cop., etc.) which reading would make Paul identify himself with the Greeks on this occasion.

(d) THE CHANGES WITH i. For i and e see under (b); for i and h see under (c); for iota subscript (adscript), mute or irrational i, see under (c). For irrational iota see also Infinitive under Verb. The papyri show it in queer forms like avlhqh/i, le,gwi, P. Oxy. 37 (A.D. 49).

i and ei. The interchange between these vowel-symbols began very early (certainly by the sixth century B.C.120) and has been very persistent to the present day. The inscriptions give numerous examples121 in the fifth century B.C., such as avpoktinh vEpafro,deitoj. This was apparently the beginning122 of itacism which was extended to u h and then to h| oi ui. Jannaris123 thinks that the introduc-

196 A GRAMMAR OF THE GREEK NEW TESTAMENT

tion and rapid spread of h contributed to this confusion as by that time ei was pronounced like i, and h was taken by many, not as long e, but equal to i. The confusion apparently began in the Boeotian dialect124 and in postclassical times, but swept the field in all the dialects till every ei (closed and open) was pronounced as i. By 100 B.C. the Attic inscriptions show a general interchange between ei and i, and in the second century A.D.125 the confusion exists between ei and i. Dieterich126 thinks that this itacism had its widest development in Egypt. The Ptolemaic papyri of ii/B.C. show itacism very frequently. It is only the more illiterate scribes that use ei for i, though B has o;reion (Thackeray, Gr., p. 86 f.). Thumb127 considers the interchange between i and ei in the koinh, on a par with that between o and w. In Pergamum128 the change from i to ei is much more common than that from ei to i, though forms in - i,a for - ei,a occur, as avmeli,a. The same thing is true in Magnesia, where h`mei/n ( h`mi/n) is common.129 The Herculaneum papyri tell the same story,130 while it is so common in the Egyptian papyri that Moulton131 is unable to set much store by the minutiae gathered by W. H. from the great uncials, "for even W. H. admit that their paramount witness, B, 'has little authority on behalf of ei as against i.'" Clearly the partiality of a for i and of B for ei throw them both out of court as decisive witnesses on this point.132 So it is not merely itacism that we have to deal with in the numerous N. T. examples of exchange between i and ei, but "genuine peculiarities of original orthography" also.133 Whatever Dr. Hort meant, all that is true is that different scribes merely preferred one or the other method of representing i. The whole matter therefore remains in doubt and one is prepared for all sorts of variations in the N. T. MSS., because the koinh, no

ORTHOGRAPHY AND PHONETICS 197

longer insisted in the vernacular on the distinction between long or short i and ei. The examples here presented will give a fair idea of the situation. For the textual evidence see careful discussion by Gregory.134 Where ei is written for i it is to be pronounced like i. Ei is shortened to i in some abstract substantives, -- i,a instead of - ei,a, as135 vAttali,a a`gni,a (possibly), perhaps avkribi,a avlazoni,a avnadi,a avreski,a perhaps avpeiqi,a evqeloqrhski,a (but qrhskei,a), eivdwlolatri,a (but latrei,a), eivlikrini,a perhaps evkteni,a evpieiki,a evriqi,a e`rmhni,a i`erati,a Kaisari,a kakohqi,a kakopaqi,a kolaki,a kubi,a Laodiki,a magi,a meqodi,a ovfqalmodouli,a $douli,a doubtful), possibly paidi,a (cf. Ps. 53:5), politi,a pori,a ptwci,a pragmati,a prau?paqi,a probably Samari,a Seleuki,a perhaps strati,a farmaki,a Filadelfi,a wvfeli,a. Deissmann136 shows that it is logei,a, not logi,a in the papyri and so in 1 Cor. 16:1 f. Some MSS. have evpa,rceia (for - ia), euvtrtape,leia (for - ia), late MSS. kolwnei,a.

The endings - eion, and - eioj appear sometimes as - ion, -- ioj. So ai;gioj, [Arioj $Pa,goj% a;stioj da,nion (cf. dani,zw danisth,j% eivdw,lion vEpikou,rioj evpith,dioj mega,lia (cf. megalio,thj), pandoki,on stoici,on. Strong testimony exists for all these. So also - ino,j for - eino,j appears in ovrino,j skotino,j fwtino,j.

Further examples of i for ei are found as in the MSS. in avdia, liptoj avne,kliptoj avli,fw avpiqe,w avpiqh,j avpiqi,a avpodedigme,noj ;Areopa gi,thj di,gma evxali,fw katalelimme,noj (Ac. 25:14), even kri,sswn li,mma litourgo,j margari,thj (cf. poli,thj tecni,thj), mesi,thj oivkti,rw para digmati,zw piqo,j u`po,limma filo,nikoj filoniki,a creofile,thj. This is not to mention the verb-forms i;don i;dan i;den which W. H. count alternate forms in Revelation, but which are pure examples of itacism. In the case of vIko,nion (Ac. 13:51; 14:1) the inscriptions give both vIk. and Eivk.137

The use of eiv for i is seen in several ways also in N. T. MSS. In Mt. 28:3 W. H. give eivde,a, not ivde,a. Gei,nomai and geinw,skw are very common in the best MSS. `Hmei/n and u`mei/n are rarely seen, however. vAxei,nh Galeilai,a vElamei,thj Leuei,thj Leueitiko,j lei,an Nineuei,thj Peila/toj Samarei,thj all are found, as well as trapezei,thj, Fareisai/oi. Ta,ceion appears in John and Hebrews. In the Pastoral Epistles, Hort138 finds - leip- for - lip- forms. Keiri,aij is correct in Jo. 11:44. Hort139 also prefers panoikei,, but pamplhqei, is undisputed. Such verb-forms occur as mei,gnumi teima,w tei,sw.

198 A GRAMMAR OF THE GREEK NEW TESTAMENT

Semitic proper names in y have ei as vAddei, vArnei vEslei, vHlei, Melcei, Nhrei,) Cf. also140 vAdmei/n vAcei,m beniamei,n Dauei,d vEliakei,m vIwrei,m Kei,j Leuei,j Nefqalei,m Salei,m Semeei,n ceroubei,n Corazei.n) So also vEleisabe,t vHlei,aj qua,teira vIa,eiroj vIereicw, vIwsei,j vOzei,aj Sa,pfeira Tabeiqa,. Cf. also hvlei, r`abbei, r`abbounei, sabacqanei,. But y appears as i in vAminada,b Melcisede,k Sina, Siw,n) Likewise the MSS. usually read vAnani,aj Baraci,aj vEzeki,aj Zaca ri,aj vIeremi,aj vIerconi,aj Maqqi,aj Mattaqi,aj Ouvri,aj.

In many of these examples of changes in i and ei the testimony is greatly divided and one must not stickle too much for either spelling. The papyri and the inscriptions have nearly all of them. See 1 (c) for remarks on the difficulty of relying on the uncials in the matter of orthography. It is impossible to be dogmatic on the subject.

i. and o. It is a peculiar change, as Blass141 observes, that we have in ovmeiro,menoi for i`meiro,menoi (1 Th. 2:8). It appears in the LXX (some MSS. for Job 3:21 and Symm. at Ps. 62:2). The only example so far brought to light is u`peromei,resqai in Iren. 60. Winer-Schmiedel142 sees no comparison in katantroku, for katantikru,. Meisterhans143 gives avpantroku, for avpantikru,.

i and o. Jannaris144 defends the exchange of i and oi possibly as early as the fifth century B.C. Certainly in the first century B.C. Auvgoustoi/noj occurs in the inscriptions.145 Oi was exchanged with ei and h| as well as with i. In the N. T. the only example is in Mk. 11:8 where ACSVTX G Or. have stoiba,j for the usual stiba,j (from stei,bw). N and a few other MSS. read stuba,j. Zonar. illustrates this also by using stoiba,j. Cf. also stoibh,, stoiba,zw, etc. This word thus illustrates well the common itacistic tendency, showing forms in - i, - oi, - u and - ei (in the verb). The LXX has only sti,coj and stici,zw, not stoic. (Thackeray, Gr., p. 92).

i and u. These two vowels sometimes have the force of the consonants146 j (y) and u (cf. Latin). Cf. au- (af) and eu-- (ef) in modern Greek, and e in po,lewj. In modern Greek "every i- or e-sound which collides in the middle of a word with a succeeding

ORTHOGRAPHY AND PHONETICS 199

vowel, loses its syllabic value and becomes consonanted" (Thumb, Handb., p. 10). So a[gioj = ayos. The i is the last of the five original vowel-sounds in this order: a o u e i. This relative value has persisted in modern Greek (Thumb's Handbook, p. 12 f.). Jannaris147 gives avpwqou,menoi as an illustration of this gradation in sound. But as a matter of fact the interchange between i and u is not frequent. Meisterhans148 finds only five examples in the Attic inscriptions, two of which, bublion and Mitulhnai/oj, are found in N. T. MSS. (assimilation). Examples occur in the koinh, of Asia Minor, though Thumb149 agrees with Kretschmer in calling it a "barbarism." Still the old distinction in sound between i and u slowly broke down till in modern Greek the two vowels have the same sound. bh,rulloj in Rev. 21:20 is spelled also in MSS. bh, rilloj bu,rilloj biru,llioj, a fine illustration of itacism. D reads bu,bloj for bi,bloj in Mk. 12:26 and Lu. 20:42: In Ac. 20:14 Mitulh,nh is the correct text for the old M,ut., but AE have Mitu li,nh and L Mutuli,nh. For the Trwgi,lion of Strabo and the Byzantine writers the Textus Receptus addition to Ac. 20:15 has Trwguli,a, other MSS. Trwgu,llion Trwgu,lion.150 The LXX shows also h[musu in q Dan. 7:25 (13). The Ptolemaic papyri vary in this word (Thackeray, Gr., p. 95). In Lu. 19:8 D has h[musoi.

(e) THE CHANGES WITH o. For changes with a see under (a), for o and e under (b), for o and i under (d).

o and ou. The old Attic used Dio,skoroj, which Phrynichus151 prefers, though Thucyd. and Plato have the form in - ouroj also (Epic or Ionic). In Ac. 28:11 only some of the cursives have the form in - oroj. Both forms appear in the inscriptions.152 This exchange is rather common in the Ptolemaic papyri (Mayser, Gr., pp. 10 f., 116 f.). In the LXX a shows sometimes ovk for ouvk (Thackeray, Gr., p. 91). The modern Greek dialects have much diversity of usage on this point. Cf. Thumb, Handb., p. 8.

200 A GRAMMAR OF THE GREEK NEW TESTAMENT

o and u. The MSS. vary between153 pra/oj (Syrian) and prau<j in Mt. 11:29; 1 Pet. 3:4, as well as between prao,thj and prau<thj Pauline Epistles. W. H. adopt the form in - u. Von Soden varies between these forms, giving no reasons. It is the old distinction surviving in the koinh,. The LXX has the u form. The papyri have other illustrations (Mayser, Gr., p. 97). Cf. Poti,oloi in Ac. 28:13 for the Latin Puteoli.

o and w. Originally o represented both the short and long noun so that it was easy with careless pronunciation for more or less confusion to exist after w came into use. The Boeotian Pindar, for instance, has Diw,nusoj instead of Dio,nusoj.154 The New Ionic zo,h (parox.) appears in lieu of zwh,. However, the introduction of the Ionic alphabet in 403 B.C. kept the two vowels pretty distinct in Attic till the Roman time, though the change began in the third century B.C.155 After the second century B.C. the exchange of these two vowels was indiscriminate in the more illiterate vernacular.156 The confusion was earliest in Egypt, but the Attic inscriptions kept the distinction well till 100 A.D. The early uncials for the LXX and the N.T. show little evidence of the interchange (Thackeray, Gr., p. 89). Jannaris finds it common. The modern Greek makes no difference in sound between o and w except medial o as in not. "In the early papyri the instances of confusion between o and w are innumerable."157 The inscriptions tell the same story about the koinh, in Magnesia158 and Pergamum.159 In some instances,160 like do,ma for dw/ma and pro,doma an w is shortened to o after the analogy of e from h in qe,ma. In the N. T. MSS. "probably the commonest permutation is that of o and w, chiefly exemplified in the endings - omen, and - wmen."161 It is useless to follow the MSS. through their variations on this point. In Ro. 5:1 e;cwmen is supported by all the best documents and gives a difficult sense at first, though a better one on reflection than e;comen. In 1 Cor. 15:49 the evidence is so nearly balanced that

ORTHOGRAPHY AND PHONETICS 201

W. H. cannot decide between fore,swmen and fore,somen (the latter in the margin). Von Soden gives - sw-. This difficulty of distinguishing between o and w in the indicative and subjunctive increased in later koinh, times.162 Several further N. T. examples of interest are avgora,swmen (Lu. 9:13), is i[na avnapah,sontai (Rev. 14:13), i[na avnapau,sontai, (Rev. 6:11), eva.n avpoqnh,skomen as read by Lachmann (Ro. 14:8), i[na ginw,skomen (1 Jo. 5:20), i[na diw,kontai according to Tisch. (Gal. 6:12), i[na die,rcomai according to Treg. (Jo. 4:15), dw,swmen according to Treg. and Tisch., and preceded by avgora, swmen (Mk. 6:37), iva,somai (Mt. 13:15; cf. Is. 6:10), i[na kauqh,swmai or kauch,swmai (1 Cor. 13:3), i[na xurh,sontai (Ac. 21:24). In all these instances syntactical questions enter also besides the mere question of vowel interchange.163

The o appears instead of w in po,ma (1 Cor. 10:4; Heb. 9:10), pro,i?moj (Jas. 5:7), Stoi?ko,j (Ac. 17:18),164 sukomore,a, not - mwre,a (Lu. 19:4), creofile,thj according to W. H. and not creofeile,thj (Soden) nor crewfeile,thj according to LU, etc. (Lu. 7:41; 16:5). But w is correct apparently in avgaqwsu,nh a`giwsu,nh evndw,mhsij (Rev. 21:18, Soden - do,m--), i`erwsu,nh megalwsu,nh prwi?no,j. So also the LXX, but pro,i?moj (Thack., Gr., p. 90). Codex B shows others in the LXX (ib.). In Lu. 18:5 and 1 Cor. 9:27 the MSS. vary between u`pwpia,zw (from u`pw,pion) and u`popia,zw (-- pei,zw old form), though the best MSS. read u`pwp)165 In Ro. 13:3 tw|/ avgaqw|/ e;rgw| may possibly be tw|/ avgaqoergw|/. So in 2 Pet. 3:6. di v w-n may be166 for di v o[n. In Rev. 4:7 f. e;cwn, not e;con (Soden), is read by the best MSS., though the substantive is zw|/on. Now second century B.C. papyri have u`po,mnhma e;cwn where w and o are exchanged.167

(f) THE CHANGES WITH u. For the changes with u and i see under (d), u and o under (e).

u and eu. Only one example of this exchange appears in the N. T., that of presbu,thj in Phil. 1:9. Here the sense seems to demand presbeuth,j. Bentley suggested it long ago and Lightfoot (comm. in loco) collected a number of instances of the omission

202 A GRAMMAR OF THE GREEK NEW TESTAMENT

Addenda 2nd ed.

of e from eu, in single MSS. Hort168 thinks it due to a scribe and not to Paul, since the earlier Greek shows no examples of this interchange. However, Wood169 has found presbeu,teroj for presbu, teroj in an Ephesian inscription (analogy: in modern Greek eu=ef). Thackeray (Gr., p. 97) finds this "natural error" in the LXX MSS.

u and ou. This has always been a rare exchange in the Greek, the Boeotian dialect having retained the original u sound of u after the Attic gave it up.170 The Zaconian preserves it in the modern Greek.171 The koinh, has sometimes crouso,j for cruso,j.172 But ou was rather frequent in the koinh, to represent the Latin u as Drou/soj.173 In Rev. 3:18 the MSS. have kollou,rion kollu,rion koul lou,rion, etc. (Latin collyrium). W. H. prefer kollou,rion though aBC read - u,rion (so Soden).-Blass174 observes that we have long u - u,rion. B in the LXX shows the same variations (Thack., Gr., p. 92). The Ptolemaic papyri have few instances. Cf. change of u and ou (Mayser, Gr., p. 118). Thumb (Hellen., p. 193 f.) thinks that u in the koinh, was pronounced like German u ?, i and also u. In Rev. 1:5 the distinction between lu,santi ( aAC) and lou,santi (BP) is more than mere orthography, though the confusion was rendered easy. UI is always so written in the N. T. uncial MSS.,175 though the iota was sometimes dropped in, the inscriptions.

(g) THE CHANGES WITH w. For changes with w and a see under (a), for w and o under (e).

w and ou. The Thessalian dialect176 changed w to ou as in tou/ koinou/ for tw/ koinw/. This change reappears in Rhodes and the AEolic-Doric.177 Buresch178 finds the change between w and ou common in the Egyptian vernacular, as in the Sahidic dialect oo is often used for w.179 It is, of course, possible, according to the view of Winer-Schmiedel,180 that some indicatives in ou may really

ORTHOGRAPHY AND PHONETICS 203

be subjunctive as a result of this vowel-interchange. The contract form for the present participle tw|/ nikou/nti is read by AC in Rev. 2:17 and A in 2:7, a change more likely due to confusion of - a,w and - e,w verbs. So with i[na zhlou/te (Gal. 4:17) and i[na fusiou/sqe (1 Cor. 4:6), but the present indicative can be used with i[na, and one is slow to credit this form to a mere vowelexchange. The same remark applies to i[na tre,fousin (W. H. marg. Rev. 12:6) as well as i[na ginw,skousin (Tisch. and Treg., Jo. 17:3) and i[na swfroni,zousin (Tisch. and Treg., Tit. 2:4). The future indicative with i[na as katadoulw,sousin (Gal. 2:4), proskunh,sousin (Rev. 9:20), staurw,sousin (Tisch., Treg., Lach., Mk. 15:20), sfa,xousin (Rev. 6:4) has rival readings with w, aorist subjunctive. It is hardly mere vocal similarity. Similar instances are mh,pote katapath,sousin (Mt. 7:6), eva.n metanoh,sousin (Rev. 2:22), w|- eva,n dou leu,sousin (Ac. 7:7). In these and similar examples where the MSS. vary between w and ou it is probable that, as with h and e o and w, the difference in mode may have been blurred by the tendency to exchange these vowels. But the syntactical question is not essentially altered by this incidental orthographical problem.

w and wu?. Lachmann, Tregelles, W. H. all write wu in Mwush/j, but Thayer urges that the word is a trisyllable Mwu?sh/j (Fritzsche, Gesenius, Tisch., Soden). The Ionic e`wutou/ is a trisyllable. Cf. Mayser, Gr., p. 138. Blass181 indeed says that the diphthong wu is non-existent in the N. T. as in the Attic. The Text. Rec. reads Mwsh/j, following Strabo and Josephus in the Antiquities, though in the LXX and Josephus elsewhere we have Mwu?sh/j.

(h) CONTRACTION AND SYNCOPE. In general the koinh, uses contraction of vowels from the standpoint of the Attic,182 though a strong Ionic infusion183 is present also as in forms like ceile,wn ovre,wn, etc. The N. T. examples of unusual contraction find illustration184 in the koinh,. In the N. T. contraction is rarely neglected, as Winer saw, though evde,eto ( aC for Lu. 8:38, though BL 33 read ev`dei/to), noi< (1 Cor. 1 : 10, etc.), ovste,a (Lu. 24:39), ovste,wn (Mt. 23:27, etc.), ovre,wn (Rev. 6:15, Attic as well as Ionic), ceile,wn (Heb. 13:15), cruse,wn (Rev. 2:1, Lach., Treg.) show that the N. T. in this respect was like the koinh, and not the literary Attic. Blass185 observes that the N. T. Greek did not go quite as far in

204 A GRAMMAR OF THE GREEK NEW TESTAMENT

contracting vowels as the Attic did. In illustration can be mentioned avgaqoergei/n (1 Tim. 6:18), though avgaqourgw/n is the correct text in Ac. 14:17. But we have avmpelourgo,j i`erourgei/n, kakou/rgoj oivkourgo,j panou/rgoj, not to mention the conjectural reading avgaqoergo,j for Ro. 13:3 on the other hand. In Col. 2:16 neomhni,a for the Attic noumhni,a is read by W. H., though supported only by BFG 121 f g vg. So the LXX (Thack., Gr., p. 98). In the case of evleino,j W. H. have the regular form in Rev. 3:17, but evleeino,j in 1 Cor. 15:19. Blass186 reminds us, however, that even evleino,j may represent evlei?no,j. The N. T. likewise has nosso,j in Lu. 2:24 (like the LXX) and nossi,a (or nossia,) in Lu. 13:34; Mt. 23: 37. Phrynichus187 condemned this dropping of e in neosso,j. Kammu,w (Mt. 13:15; Ac. 28:27, both from Is. 6:10) comes from the Epic and the old vernacular. Kat was an old form parallel with Kata,.

There are several noteworthy points about i. The i is retained in avllotriepi,skopoj (1 Pet. 4:15). The same thing is true with h`mi,wron (Rev. 8:1), like h`miw,bolon in the Attic inscriptions.188 The form e;stwn in Mk. 1:6 (already in Homer) is a twin rather than a syncopated form of evsqi,wn (Mt. 11:19).189 In the N. T. the i is not dropped in such forms as biw,sesqe evnu,pnion siwpa/n ui`o,j. Blass190 calls the contraction of ieiii= i "an entirely new kind," though it appears in the koinh,, as in evpeikw/j tamei/on u`gei/a, etc.191 When ei came to be equal to i, the two sounds naturally blended into one. Cf. the Ionic dative po,li for po,lii. So in the N. T. we find pei/n (BCD), even pi/n ( aAL) for piei/n in Jo. 4:9, and elsewhere in the N. T. In Mt. 6:6, etc., tamei/on is read for tamei/on.192 On the other hand in Rev. 21:20 A reads sardio,nux for sardo,nux. W. H. read tetraarce,w tetraa,rchj rather than tetrarce,w, etc. The use of glwsso,komon instead of the earlier glwssoko,meion (- ion) should be noticed also. For the use of eva,n = modal a;n see under (b), p. 190.

(i) DIPHTHONGS AND DIAERESIS. The Boeotians monophthongized the diphthongs ai ei oi ou in the fourth and fifth

ORTHOGRAPHY AND PHONETICS 205

Addenda 3rd ed.

centuries B.C.193 The Boeotians pronounced cai,rei=cheri as the vernacular koinh, did. Thumb (Hellenismus, p. 228) objects to "this emphasizing of Boeotian" by Kretschmer (Die griech. Vaseninschriften; Einleit. in d. Gesch.). Moulton (Prolegomena, p. 33 f.) allows this Boeotian influence on the koinh, with a "perhaps." The itacising process still further developed this use of the diphthongs as monophthongs. Indeed Jannaris194 insists that the term di,fqog goj as applied to sullabh, concerned the eye rather than the ear and meant more biliteral than bivocal. The spurious diphthongs show the process in a state of completion. The papyri, unlike the inscriptions, do not dissect a diphthong at the close of a line.195 Where two vowels do not blend into one syllable, it is necessary to indicate it. Hence from very early times marks of diaeresis were used to show that each vowel has its own sound. The mark is put over the i or u which might otherwise be considered to unite with the preceding vowel. These marks are found in the oldest N. T. MSS. with such words as a`llhlou,i?a, (Rev. 19:1; but in the case of proper names transliterated from the Hebrew or Aramaic W. H. follow the Hebrew or Aramaic spelling. Cf. Hort, Intr., p. 313. So in other examples below), vAcai<a, vAcai?ko,j (1 Cor. 16:17), Bhqsai?da, Ga,i?oj (also Gai/oj in Ac. 20:4, etc., but cf. Allen, Harvard Studies in Class. Philol., ii, 1891, pp. 71 ff.), diu?li,zein (Mt. 23:24), vEbrai?ti, evlwi> (Mk. 15:34), vEf rai,m, however, or vEfre,m ( aL in Jo. 11:54), vHsai<aj, though B usually without,196 vIoudai?kw/j ivscu<i (2 Pet. 2:11), Kai?a,faj Ka,i?n (W. H. Kai,n), so W. H. Kaina,n (not Kai?na,n nor - a,m), Leuei,thj and not Leui<thj in W. H., Lwi<j (W. H. - i,j), Mwush/j in W. H., not Mwu?sh/j Nineuei,thj and not Nineui<thj pro,i?moj according to W. H., but prwi, prwino,j. W. H. have Ptolemai<da in Ac. 21:7 and `Rwmai?sti, in Jo. 19:20. D reads Corazai<n. The Semitic etymology complicates the matter with some of these words.197 Many of the MSS. use diaeresis at the beginning of words as in i<na.198 aA regularly write hu?, while wu? is correct also.199 See Giles200 on the subject of diphthongs. For iota subscript see under (c).

(j) APHAERESIS AND PROTHETIC VOWELS. qe,lw, not evqe,lw, is the only form in the N. T., as it is the common form in the koinh, and is that used in modern Greek. It is as old as Homer, and since

206 A GRAMMAR OF THE GREEK NEW TESTAMENT

250 B.C. is the only form in the Attic201 and Ionic202 inscriptions. The augment, however, is always h. Cronert203 finds evqe,lw after consonants. The koinh, does not follow the Ionic in the use of kei/noj for evkei/noj. Aphaeresis is frequent204 in the modern Greek vernacular, kei/ and evkei/ de,n for ouvde,n, etc. But the N. T. has only evcqe,j (so LXX) in the best MSS. (cf. Jo. 4:52 aABCD; Ac. 7:28 aBCD; Heb. 13:8 aACD), the usual Attic form, though the papyri sometimes have cqe,j instead of the common evcqe,j. The N. T. does not have du,romai ke,llw mei,romai, where o is dropped. Cf. Kuhner-Blass, Tl. I, Bd. 1, p. 186. The form mei,romai (cf. ovmeiro,menoi, in 1 Th. 2:8) occurs in Nicander for i`mei,romai. It is possible that in ov$o`%mei,romai we have prothetic o instead of apharesis. Cf. Hort, Notes on Orth., p. 152; WinerSchmiedel, p. 141. See Additional Notes for full list.

(k) ELISION. Besides the use of the movable final n and j the Greeks had two other methods of obviating hiatus (elision, crasis). The hiatus was distasteful to the finished writers, though more freedom was exercised in poetry. The avoidance of hiatus was always a more or less artificial matter and hiatus was unavoidable in the most careful Attic writers, as in the case of o[ti, peri, pro, ti, ti, the article, relative, the small "form-words" ( kai,, eiv mh,), etc. But the harsher hiatus like evdi,doto auvtw|/ would be avoided by the literary koinh, writers as well as by the Atticists. The inscriptions and the papyri show far less concern about hiatus than do the literary writers of the koinh,. As might be expected the N. T. books agree in this matter with the vernacular koinh, and the MSS. vary greatly among themselves. Blass205 considers this situation in harmony with the tendency to greater isolation of the words in the later language. Indeed he thinks that only one206 book in the N. T. (Hebrews) shows the care of an artistic writer in the avoidance of hiatus. By omitting the O. T. quotations and chapter 13 he finds that hiatus where there is a pause is a matter of indifference, as also with kai,. He finds fifty-two other instances of hiatus, whereas Romans goes beyond that num-

ORTHOGRAPHY AND PHONETICS 207

ber as far as ch. 4:18. But even then Blass has to admit cases of harsher hiatus in Hebrews, like avdelfoi. a[gioi e;nocoi h=san, etc.

The Attic inscriptions show that the vernacular tongue did not care much about hiatus.207 The lighter elisions like d v were used or not at will, while the heavier ones like di,kai v o[pwj were rare. The same indifference to elision appears in the koinh, inscriptions208 and in the papyri.209 In general in the N. T. elision takes place regularly before pronouns and particles and before nouns in combinations of frequent occurrence210 like katv oi=kon. Blass211 has carefully worked out the following facts in the N. T. MSS. Te ou=te mh,te a[ma a;ra ge evme, e;ti i[na w[ste, etc., do not undergo elision nor do noun- or verb-forms . The verse of Menander quoted in 1 Cor. 15:33 is properly printed crhsta. o`mili,ai by W. H.212 Even the compound words tesserakontaeth,j (Ac. 7:23) and e`katontaeth,j (Ro. 4:19) do not suffer elision, while tetraa,rchj has no elision in aC D (Alexandrian, Hort). Tou/t v e;sti or toute,sti is the only example in the pronouns that we have in the N. T.213 It is in the particles then that most N. T. elisions occur, though there are comparatively few. vAlla,, according to Gregory,214 has elision in 215 cases and fails to have it in 130, though the MSS. vary much. Hort215 observes that in avlla, elision is usual before articles, pronouns and particles, but rare before nouns and verbs. Ro. 6: 14-8:32 has many non-elisions of avlla,, and the elision varies before the different vowels except that it is constant before rarely suffers elision outside of o[j d v a;n, but here frequently, while W. H. read de. auvto, in Ph. 2:18 after aBP. In 2 Cor. 3:16 W. H. put h`ni,ka d v a;n in the margin, text h`n) de. eva,n (so Tisch, Nestle). In ouvde, elision takes place several times, as in ouvd v a;n (Heb. 8:4), ouvd v eiv (Ac. 19:2, aAB), ouvd v i[na (Heb. 9:25), ouvd v o[ti (Ro. 9:7), ouvd v ouv (Mt. 24:21; Heb. 13:5), ouvd v ou[twj (1 Cor. 14:21). Blass216 further notes that prepositions seldom use elision

208 A GRAMMAR OF THE GREEK NEW TESTAMENT

Addenda 3rd ed.

with proper names, since it was thought better, as on the inscriptions, to keep the name distinct and readily discernible, though W. H. read di v vAbraa,m in Heb. 7:9. Elision is most common with dia, as di v evso,ptrou (1 Cor. 13:12), "because there were already two vowels adjacent to each other" Blass217 thinks. vAnti, has elision only in avnq v w-n (Lu. 1:20, etc.). Elsewhere the prepositions show elision with pronouns and in current phrases, as in avp v avrch/j avp va;rti avp v auvtou/ avp v evmou/ evp v auvtw|/ kat v evme, kat v ivdi,an $kaq v i`di,an% kat v oi=kon met v evmou/ par v w-n u`f v h`mw/n v $u`mw/n% u`p v ouvdeno,j (1 Cor. 2:15).218 So the LXX (Thackeray, Gr., p. 137).

(1) CRASIS. The Attic official inscriptions make little use of crasis, though it is fairly common in the vase-inscriptions of the fifth century B.C.219 In Magnesia Nachmanson finds only a few examples of kai, and the article.220 The same thing is true of Pergamum.221 In the N. T. it is confined also to kai, and the article. And in the case of kai, crasis only occurs if the following word is a pronoun or a particle. Kai, thus often, though not always, coalesces with evgw, and the oblique cases, as kavgw, kavmoi, kavme,. If there is a "distinct co-ordination of evgw. with another pronoun or a substantive," crasis does not take place.222 Even the MSS. vary greatly.223 Kavkei/noj also is found as well as kavkei/ and kavkei/qen) Kai, likewise blends only occasionally with eva,n in the sense of 'and if,' as in Mk. 16:18; Lu. 13:9; Jas. 5:15. In the sense of 'even if' the crasis is more common, as in Mt. 26:35; Jo. 8:14. In the sense of 'if it be but' or 'if only' the crasis is uniform as in Mk. 5:28; 6:56; 2 Cor. 11:16.224 Cf. ka;nkai. eva,n (Jo. 8:14, 16). The article suffers crasis very often in the older Greek, but in the N. T. it is seldom so. Hort 225 declines to accent tauvta, for tau/ta in 1 Cor. 9:8 or tauvta, for ta. auvta, in Lu. 6:23, 26; 17:30, though supported in Luke by some good MSS. He does, however, accept tou;noma in Mt. 27:57 and touvnanti,on in 2 Cor. 2:7; Gal. 2:7; 1 Pet. 3:9 ("stereotyped as a single word," Blass226). Crasis is quite rare in the LXX (Thackeray, Gr., p. 137).

ORTHOGRAPHY AND PHONETICS 209

III. Consonant-Changes ( stoicei/a su,mfwna). The Greek, like other Indo-Germanic tongues, wrote out both vowels and consonants save in the case of iota adscript, which was not always used. But, as with the Phoenician and Hebrew, which wrote only consonants, the consonants form the backbone of the language. Both consonants and vowels are originally pictographic. "Beth" ( bh/ta) is 'house,' "gimul" ( ga,mma) is 'camel,' "daleth" ( de,lta) is 'door,' etc.227 The Greek indeed developed the vowels a e i o out of the Phoenician consonants aleph, he, yod, ayin.228

(a) ORIGIN AND CHARACTER OF THE CONSONANTS. Though the Greek consonants undoubtedly come chiefly from the Phoenician symbols, they were not all used at once nor in the same places. At first the digraphs K?, T? , x? were used for the later X, q, F, and even after these letters won a foothold KS CS, PS FS were used in Attic for x y. It is only since 403 B.C. that the Greek alphabet ( a;lfa bh/ta) has had regularly twenty-four letters. Jannaris229 gives an interesting study of the way the Greek letters looked in eighth, sixth, fifth and fourth centuries B.C. as shown by the inscriptions. In the inscriptions, however, ko,ppa continued to be used (like Latin Q) and bau/ or di,gamma. This last, though called double ga,mma, perhaps represents the Phoenician vau. On the use of digamma in Homer see Kuhner-Blass.230 It is a half-vowel in fact, as i and u are partly consonant in force, like Latin u (u) and i (j).231 The dropping of digamma affected many words, some of which have the rough breathing, though Thumb232 and Moulton233 think that this is an accident simply, and the rough breathing is due to analogy and not to the digamma in cases like kaq v e[toj, etc. But changes in the use of the consonants did not cease when the Euclidean spelling reform was instituted 403 B.C. As the vowels underwent steady development, so it was and is with the consonants. B early began occasionally to have the force of u, and g sometimes the j value of i as in modern Greek, and it was even inserted (irrational g).234 In general in the koinh, the

210 A GRAMMAR OF THE GREEK NEW TESTAMENT

consonant-changes are much fewer thhn those of the vowel. Such peculiarities as ss, gi,nomai lh,myomai are common (Thackeray, Gr., p. 100).

(b) THE INSERTION OF CONSONANTS. In the older Greek d is inserted in avndro,j, and so with b in meshmbri,a.235 The Attic used either form in evmpi,$m%plhmi evmpi,$m%prhmi. So in Ac. 14: 17 DEP read evmpimplw/n, (D evn), and in Ac. 28:6 acBHLP most cursives have pi,mprasqai. The LXX MSS. show the same variation. D in Lu. 2:32, etc., has vIstrah,l. The retention of m in all the forms (derivatives also) of lamba,nw (root lab) is in accord with the usage of the papyri ("almost invariably")236 and the inscriptions of the koinh,, and is due to the Ionic la,myomai.237 Hence lh,myomai evlh,mfqhn, etc. In the Ptolemaic age (iii/i B.C.) the papyri give both forms. From i/iv A.D. the papyri and uncials (LXX and N. T.) give almost wholly forms. In the Byzantine period (vi/viii A.D.) the classic lh,yomai reappears. Cf. Thackeray, Gr., p. 108 f.; Mayser, Gr., p. 194 f.; Cronert, Mem., p. 66. In the LXX the uncials give the spelling of their own date, not that of the translation. In Mk. 7:32 the extra g in mog$g%ila,lon is inserted by the Syrian class only and is not to be accepted. In Heb. 11:32 p is added to Samisw,n ( Samyw,n). So also in Ac. 3:7 ( aABC) d is added to sfu$d%ro,n which is as yet "unexplained."238 In the case of `Adramunthnw|/ (Ac. 27:2), read by W. H. on authority of AB 16 Copt. instead of `Adramutthnw|/, a slightly different situation exists. Two ways of pronouncing and spelling the name of the city existed.

(c) THE OMISSION OF CONSONANTS. There are not many cases where a consonant drops out of a N. T. word. In Rev. 13:2 the correct reading (all the uncials) is undoubtedly a;rkou, not a;rktou. This form is found also in the LXX and in inscriptions of the first or second century A.D.239 W. H., following B and a, also (save in Mk. 3:22) read beezebou,l instead of beelzebou,l. Gi,nomai and ginw,skw are the exclusive forms in the N. T., though some MSS., as in the papyri and inscriptions, have gein-. Nach-

ORTHOGRAPHY AND PHONETICS 211

manson240 states clearly the facts. The Ionic as early as the fifth century B.C. used the gin forms, and the Doric shows the same situation in the fourth century. Even in Athens the gin, forms appear, and in the koinh, the gign forms vanish. Golgoqa, follows the Hebrew tl,Gl.gu rather than the Chaldaic ax'l.G'l.Gu in having only one l. According to Winer-Schmiedel241 the two forms kau/da and klau/da (Ac. 27:16) represent two different islands near each other, which were confused in the MSS. It is hardly worth while to remark that sa,rdion (correct text in Rev. 4:3) is a substantive, while sa,rdinoj (Text. Rec.) is an adjective.

(d) SINGLE OR DOUBLE CONSONANTS. Blass242 and WinerSchmiedel243 comment on the obscurity concerning the use of single or double consonants in the koinh,. The phenomena in the N. T. in general correspond to the situation in the koinh,.244 In the modern Greek vernacular (cf. Thumb, Handbook, p. 27) the double consonants, except in Southeastern Greek dialects, have the value of only one. In the oldest Attic inscriptions in most cases where the doubling of consonants was possible the single consonant was used.245 The rule with initial r` was that when it passed to the middle of a word as a result of reduplication or the prefixing of a preposition, etc., it was doubled. But rverantisme,noj is read by aACDP in Heb. 10:22 as in Ionic and late Greek, r`erimme,noi in D (Mt. 9:36), and perireramme,noj in a (Rev. 19:13). Blass246 observes

212 A GRAMMAR OF THE GREEK NEW TESTAMENT

that the Syriac versions use amwhr for `Rw,mh, though some Attic inscriptions use initial pp. In Mt. 9:20 ai`morrou/sa is correct ( aL one r). In Ac. 10:29 BD 61 read avnatirh,twj, and in Ac. 19:36 BL have avnantirh,twn. In Ac. 27:43 W. H. follow aC in avpori,yantaj, and in Lu. 19:35 all but the Syrian class read evpi ri,yantej and aAB have the same form in 1 Pet. 5:7. In Mt. 9:36 the Neutral (and Alexandrian) class has evrimme,noi, the Syrian evrr), while D has r`erimm-. In Mt. 15:30 aDL read e;riyan, while B and the rest have e;rriyan, but see Ac. 27:19. But in Lu. 17:2 e;rriptai is supported by all MSS. save II and pscr. In Jo. 19:23 a;rafoj is read by W. H., though B has app. In 2 Cor. 12:4 a;rrhta is right as a;rrwstoj in Mk. 6:5, 13, etc. In 2 Cor. 1:22 W. H. follow BCD vs. aAL in reading avrrabw/na, a Semitic word which in its Semitic form has the doubling of the consonant and the metrical prosody - ? - according to Blass,247 who compares also the Latin arrha. W. H. have diar,xaj in Mk. 14:63 after BN, while in Lu. 8:29 diarh,sswn is supported by ABCRU D. In Mt. 26:65 W. H. give die,rhxen on the authority of only qf according to Tisch., though BL read dierh,sseto in Lu. 5:6. But prose,rhxen in Lu. 6:48 is supported by aBDL and in 6:49 by BDL. In Ac. 16:22 perirh,xantej is the reading of all uncials save P, but most cursives follow P. But in Ac. 14:14 all MSS. have diarrh, xantej and in Lu. 9:42 the same thing is true of e;rrhxen. In Mk. 2:21 evpira,ptei is read by all the best MSS. and the Syrian class is divided, and the same is true of Mt. 26:67 evra,pisan. In 2 Cor. 11:25 evrabdi,sqhn, is correct, while likewise evra,ntisen (Heb. 9:19, 21) has all save late Syrian support. So - rr- in evrre,qh (BD evrrh,qh, not W. H., Mt. 5:21, etc.) is the constant reading in the N. T. In Eph. 3:17 (18) and Col. 2:7, all MSS. have evrrizwme,noi. W. H. follow B alone in 2 Cor. 1:10; 2 Pet. 2:7 with evru,sato, while in Col. 1:13 B is joined by FGP. In 2 Tim. 3:11 AD read evru,sato, and aAC 37 give evru,sqhn in 2 Tim. 4:17. All MSS. have e;rrwsqe (Ac. 15:29). Mu,rra (B) is changed to Mu,ra in the Syrian text (Ac. 27:5; cf. Hort, Notes on Orth., p. 160), but Winer-Schmiedel (p. 58) found only Mu,ra in the inscriptions. Pararuw/men (Heb. 2:1) is read by all the pre-Syrian classes. Parrhsi,a parrhsia,zomai (from pan rhsi,a), not parh-, is the usual reading in the N. T. (see Additional Notes), as occasionally in the inscriptions.248 W. H. read purro,j in

ORTHOGRAPHY AND PHONETICS 213

Rev. 6:4 and 12:3, though the evidence is pretty evenly balanced.249 The Alexandrian class has pura,zei in Mt. 16 : 2, but W. H. reject the passage. The MSS. all have Ceima,rrou in Jo. 18:1.

The other instances outside of r are not so numerous. The MSS. (all but late Syrian) support ( balla,ntion, not bala,ntion, as do the papyri.250 Blass251 argues for it also on metrical grounds. Ge,nhma, because given by no grammarian, was "attributed by Fritzsche (on Mark, pp. 619 ff.) to the carelessness of transcribers" (Thayer), but as sometimes in the LXX (Ezek. 36:30) so in the N. T. the best MSS. distinguish between ge,nnhma (from genna,w), 'living creatures,' as gennh,mata evcidnw/n (Mt. 3:7) and ge,nhma (from gin,omai, 'the fruits of the earth,' as evk tou/ genh,matoj th/j avmpe,lou (Mk. 14:25). Phrynichus252 condemns the use of ge,nnhmakarpo,j (Diodorus, Polybius, etc.). Root of both verbs is gen. This distinction between ge,nhma and ge,nnhma appears in the papyri also, though genh qe,nta occurs in the Fayum Papyri (B.U. 110. 14) "undoubtedly from genna,w."253 So N. T. MSS. vary254 about ge,nnhma. The grammarians (Lobeck, ad Phrynichum, p. 726) reject evkcu,nw for evkce,w, but the best MSS. give evkcu,nnw everywhere in the N. T. W. H. accept this AEolic form in Mt. 23:35; 26:28; Mk. 14:24; Lu. 11:50 marg.); Lu. 22:20 (bracket the passage); and Ac. 22:20. So also suncu,nnw (W. H.) in Ac. 9:22; 21:31. Cf. u`perekcunno,me non in Lu. 6:38. Likewise MSS. support avnabai,nnw ovpta,nnomai while the AEolic avpokte,nnw is received by W. H. in Rev. 6:11 and avpoktennu,w in Mk. 12:5, though rejected elsewhere in N. T. on divided testimony. ;Enatoj has been restored throughout the N. T. by W. H. instead of e;nnatoj of the Text. Rec. The inscriptions support the N. T. MSS. in this change (Thayer). So W. H. give evnenh,konta (Mt. 18:12 ff.; Lu. 15:4, 7) but evnne,a always. vEneo,j, not evnneo,j, W. H. give (Ac. 9:7) as the LXX (Is. 56:10), a word possibly identical with a;newj ( a;naoj). W. H. present255 kra,bat toj instead of the krabbatoj of the Text. Rec., though kra,batoj would more nearly represent the Latin grabatus as it appears in Etym. M. (154. 34; 376. 36). Kraba,trioj is found also for the

214 A GRAMMAR OF THE GREEK NEW TESTAMENT

Latin grabatarius (CIGII 2114 d i). a, however, has 10/11 times the strange form kra,baktoj (- tt- only in Ac. 5:15). Lase,a (Ac. 27:8) is Lassai,a in some MSS. Mamwna/j, from Aramaic an'Amam', is correct. Masa,omai is the right reading in Rev. 16:10 ( aACP). Only the Western class has plhmu,rhj for plhmmu,rhj in Lu. 6:48. W. H. properly have r`a,koj, not r`a,kkoj, from r`h,gnumi (Mt. 9:16; Mk. 2:21). In the Western interpolation in Ac. 20:15, W. H. read Trwgu,lion, not - u,llion nor - i,lion. Some Latin MSS. read hysopus for u[sswpoj in Jo. 19:29 and Heb. 9:19. Fu,geloj, not - elloj, is read in 2 Tim. 1:15 by all save A and most cursives. Cf. Fuge,lioj in CIGII 3027.

The Hebrew and Aramaic proper names call for special remark. [Annaj = !n'x' (Josephus ;Ananoj) may be due to the dropping of a or to the analogy of [Anna hn'x;. W. H. (Ac. 1:23; 15:22) prefer Barsabba/j (from aB'v;rB;, 'son of the Sabbath') to Barsaba/j (from ab'v. rB;, 'son of Saba').256 The Text. Rec. has Fenh sare,t (W. H. Gennhsare,t) in Mk. 6:53, elsewhere - nn-.257 Go,morra is read in LXX and N. T. (Mt. 10:15, etc.), hr'mo[]. W. H. accept vElisai/oj, not vEliss. (Syrian) in Lu. 4:27= [v'ylia)gt. ) vIessai, (Lu. 3:32, etc.) comes from yv;yi. The N. T. and 1 Macc. have vIo,pph, but the ancient grammarians and lexicographers prefer vIo,ph.258 In Lu. 3:27 vIwana,n (indeclinable) is the right text. W. H. prefer vIwa,na ( wh;Ay) to vIwa,nhj in Lu. 8:3; 24:10. But more doubt exists concerning vIwa,nhj, which W. H. read everywhere save in Ac. 4:6; 13:5; Rev. 22:8, following B and sometimes D. The single n prevails in D in Luke and Acts, while vIwa,nnhj is more common in D in Matthew, Mark, John.259 a has the single n in the part written by the scribe of B.260 The inscriptions have it both ways. Blass261 finds the explanation in the Hebrew termination -an, which was treated as a variable inflection in the Greek, the LXX MSS. having now vIwana,n and now vIwa,non. This fact opposes the derivation of the name vIwa,nnhj from vIwana,nhj, leaving the - hj unexplained.262 Maria,m $ ~y'r.mi % = Maria,mmh in Josephus.263 Messi,aj is from the Aramaic axy'vim. = Hebrew h;yvim'h; but the Syr-

ORTHOGRAPHY AND PHONETICS 215

ian class reads Mesi,aj in Jo. 1:41 (42); 4:25. Sa,rra, Heb. hr'f': (feminine of rf;), is read by MSS. generally in N. T., though L has Sa,raj in Ro. 4:19 (vulg. Sarae). All the MSS. have nn in Sousa,nna (Lu. 8:3) after the Heb. hn'v;Av ('a lily'). Carra,n is supported by most MSS., though D and a few cursives have Cara,n in Ac. 7:2 after the Hebrew !r'x'. The LXX has Carra,n and the Greek writers (Strabo, etc.) have Ka,rrai, Latin Carrhae.

Doubling of the Aspirate. As a rule the aspirated mutes ( q, c, f) are not doubled in more correct writing either in early or late Greek, but N. T. MSS. give examples of qq, cc, ff. In Philemon 2 D has vAffi,a, while 3 has vAppi,a (so vulg.) and FG, etc., even vAmfi,a. In Mk. 7:34 all MSS. have ovffaqa, (or evffeqa,) save D and two Coptic MSS. which have evpfaqa,. W. H. give Maqqai/oj= Hebrew hy'tim;; in the N. T. (Mt. 9:9 ff., etc.), and Maqqa,n in Mt. 1:15. W. H. read Matqa,t in Lu. 3:24, but Maqqa,t in Lu. 3:29. In Ac. 1:23, 26 W. H. have Maqqi,aj, but in Lu. 3:25 f. they prefer Mattaqi,aj to Maqqaqi,aj. In Ac. 5:1, W. H. consider Saffeira Western and read Sapfeira (either Aramaic ar'yPis;, ' beautiful,' or Hebrew ryPis; 'precious stone').264 The LXX MSS. show the same variations. Cf. Thackeray, Gr., p. 121.

(e) ASSIMILATION OF CONSONANTS. In the early period of the Greek language the inscriptions often show assimilation of consonants between separate words. The words all ran together in the writing (scriptura continua) and to some extent in pronunciation like the modern French vernacular. Usage varied very early, but the tendency was constantly towards the distinctness of the separate words (dissimilation). However, evx came finally to be written evk before consonants, though evg evkk evc evgk and even ev (cf. Latin) are found in Attic inscriptions,265 as evg nh,swn, etc. Only sporadic examples outside of evx and evk appear in the N. T. as avne,gliptoj in D (Lu. 12:33), avpegdu,sei in B (Col. 2:11), e;ggona in D (1 Tim. 5:4), eggona, not engona.266 The Attic inscriptions even have j assimilated in tou.l li,qouj. The most

216 A GRAMMAR OF THE GREEK NEW TESTAMENT

common assimilation between separate words is in words ending in -- n especially with the article and evn) Examples like th.m po,lin, to.l lo,gon to.r `Ro,dion evl Le,sbw| evs Sidw/ni, etc., are very common.267 Similar phenomena occur in the koinh, inscriptions, though the failure to assimilate is far more noticeable. See list of examples in Nachmanson.268 As a rule the papyri do not assimilate such cases.269 In the N. T., as in the later koinh, generally, only a few remnants survive of this assimilation of n between words. Blass,270 who has used the MSS. to good purpose, finds several, as, for instance, evg gastri, in A (Lu. 21:23), evg Kana/ in AF (Jo. 2:11), evm me,sw| in AC (Rev. 1:13; 2:1, etc.), in AP (Heb. 2:12), in L D (Mt. 18:2; Lu. 8:7), evm prau,thti in a (Jas. 1:21), su.m Maria,m in AE, etc. (Lu. 2:5), su.m pa/sin in EG, etc. (Lu. 24:21). The earlier papyri (up to 150 B.C.) show a good deal of this assimilation between words (Thackeray, Gr., p. 131). This assimilation between separate words is common in modern Greek (cf. Thumb, Handb., pp. 16 ff.). So to.n pate,ra=tombatera. But a much more difficult matter is presented in the case of evn and su,n in composition, though in general "assimilation is the rule in compounds of evn, retention of n in those of su,n."271 But in 1 and 2 Peter assimilation is the rule (only two clear exceptions) for both su,n and evn due possibly272 to the absence of uncials. The later papyri as a rule do not assimilate su,n, though often evn.273 In the N. T. no examples occur of evn or su,n before x or r.274 Hort275 gives a list of what he considers "the certain and constant forms" of evn and su,n in composition. "All other compounds of su,n and evn are included in the list of alternative readings." Hort thus reads evm- before the labials ( p b f) and the liquid m except evnperipath,sw (2 Cor. 6:16), possibly evnpne,wn (Ac. 9:1), and e;nprosqen once (Rev. 4:6) and Western class elsewhere. So assimilation takes place before the liquid l, as evlloga,w). But before the palatals k g the usage varies, though before c we have evgcri/sai (Rev. 3:18) with reading evn)

ORTHOGRAPHY AND PHONETICS 217

We read evngegramme,nh in 2 Cor. 3:2 f. ( aABCDFG) and evnkai, nia evnkaini,zw evnkatoike,w evnkaucw/mai evnkentri,zw evnkri,nw, though evg kale,w e;gklhma, etc., and evgkatalei,pw except in Acts.276 As to su,n here is Hort's decision. Sunp- he accepts save in sumpo,sia. On the other hand Hort has only sunbasileu,w sunbiba,zw elsewhere sumbas in sumbai,nw; only su,nfhmi sunfu,w, but sumf- as in sumfe,rw. With the palatals Hort reads sunk- always, as in sunka,qahmai, only sugge nh,j sugkalu,ptw but suncrw/mai and su,gcusij. He has both sunlale,w sunlupou/mai and sullamba,nw sulle,gw* sunmaqhth,j, etc., but summor fizw su,mmorfoj. Hort has sunzw/, etc., but su,zuge* su,nyucoj, but has both sunstauro,w, etc., and sustre,fw, etc. For the detailed W. evidence see Gregory.277 Hort also prefers palingenesi,a, but is doubtful about kencreai, panplhqei,.

(f) INTERCHANGE AND CHANGING VALUE OF CONSONANTS. One cannot here go into the discussion of the labial, palatal, dental, velar stops, the spirants, liquids, nasals. One can give only the special variations in the N. T. The b sound was rare in the older Indo-Germanic languages and easily glided into u or v.278 The Greek bai,nw is like venio in Latin, bi,oj is like vivus though different in history. In modern Greek b has sound of v. In the N. T. as in the LXX all the uncials have u in Dauei,d (W. H.) where the minuscules read Dabi,d.279 In the case of beli,ar (2 Cor. 6:15) it is from r[;y; lBe ('lord of the forest'), while the Text. Rec. beli,al is from l;[;YliB. ('worthlessness').280 The variation between rs and rr, Moulton281 observes, runs down to modern Greek. The Attic rr did not displace the Ionic and early Attic rs entirely in the Attic inscriptions.282 In the N. T., like the rest of the koinh,, usage is divided.283 Hort (p. 149) prefers a;rshn except a;rrhn perhaps 4/4 times in Paul. In the Gospels and Acts qa,rsoj and the two imperatives qa,rsei qarsei/te are uniform, but in 2 Cor. 2 Cor.(5:6, 8; 7:16; 10:1, 2) and Heb. Heb.(13:6)

218 A GRAMMAR OF THE GREEK NEW TESTAMENT

qarrei/n is the correct text. z displaces s in a few words. Voiced s in union with voiced consonants had the sound of z, and z was pronounced sd.284 ;Azwtoj (Ac. 8:40) dADv.a;, Ashdod. Lagarde's LXX has vAseddw,d in Josh. 11:22 (A has vAshdw,d, B vAseldw,). ar'z.[, is rendered also ;Ezraj or ;Esdraj. But in the N. T. period z is changing from the ds sound to z. `Armo,zw, not the Attic a`rmo,ttw, is the N. T. form.285 Lachmann has mazo,j for masto,j in Rev. 1:13. In 1 Th. 5:19 BDFG (Western class) read zbe,nnute,286 simply phonetic spelling. Hort287 considers Zmu,rna as Western only in Rev. 1:11; 2:8, but the papyri and inscriptions both give it.288 The most noticeable feature of all is, however, that the Attic and Boeotian tt did not hold against the Ionic ss (though even Thucydides and the Tragic poets used ss). Papyri, inscriptions and N. T. MSS. all unite in using ss as the rule, though all occasionally have tt. It does not seem possible to reduce the usage to an intelligent rule.289 vEkplhtto,menoj is accepted by W. H. in Ac. 13:12, elsewhere ss) Both evla,sswn (Jo. 2:10; Ro. 9:12) and evla,ttwn (1 Tim. 5:9; Heb. 7:7) are found, but only the "literary" (so Blass) words evlatto,w (Jo. 3: 30; Heb. 2:7, 9) and evlattone,w (2 Cor. 8:15). Similar diversity exists between h-sson (1 Cor. 11:17; 2 Cor. 12:15) and h`ssw,qhte (2 Cor. 12:13) on the one hand and h[tthma (1 Cor. 6 : 7; Ro. 11:12) and h`tta/sqai (2 Pet. 2:19 f.) on the other. In Heb. 6:9; 10:34 W. H. read krei,ssona, elsewhere krei,ttona (Heb. 1 : 4; 7: 7, 19, 22; 8: 6; 9:23; 11:16, 35, 40; 12:24), and Hebrews has some literary influence, an argument for Blass' idea above. Paul has krei/tton only in 1 Cor. 7:9, while krei/sson is found in 1 Cor. 7:38; 11:17; Ph. 1:23. Hort accepts krei/tton in 1 Pet. 3:17

ORTHOGRAPHY AND PHONETICS 219

and 2 Pet. 2:21 (doubtful). Cf. sh,meron for the Attic th,meron. ;Ornix (Lu. 13:34) is called Western by Hort, though Moulton,290 observes that it has some papyrus support and is like the modern Greek (Cappadocian) ovrni,c.

(g) ASPIRATION OF CONSONANTS. There is besides some fluctuation in the aspiration of consonants. See under (d) for the double aspirates like vAffi,a, etc. This uncertainty of aspiration is very old and very common in the inscriptions and papyri,291 though the N. T. has only a few specimens. W. H. read `Akeldama,c in Ac. 1:19, am'D. lq;x]. So r`aka, (Mt. 5:22), aq'yre but sabacqanei, (B has - kt-) in Mt. 27:46. Gennhsare,t is correct; the Syrian class has - e,q in Mt. 14:34. W. H. have uniformly Kafarnaou,m, and read Nazare,t save in four passages, Nazare,q in Mt. 21:11; Ac. 10:38, and Nazara, in Mt. 4:13; Lu. 4:16. In Lu. 11:27; 23:29 DFG have masqoi for mastoi,, likewise a in Rev. 1:13. vEqu,qh is read by cursives, Clem., Or., etc., in 1 Cor. 5:7. In ouvqei,j and mhqei,j after elision of e the d has blended with the ei-j as if it were t and become q. It is first found in an inscr. 378 B.C. and is the usual form in the pap. in iii/B.C. and first half of ii/B.C. By i/A.D. the d forms are supreme again (Thack., Gr., pp. 58 ff). Blass292 finds ouvqeno,j in Lu. 22:35 (ABQT); 2 Cor. 11:8 ( aBMP); ouvqe,n, in Lu. 23:14 ( aBT); Ac. 15:9 (BHLP); 19:27 ( aABHP); 26:26 ( aB); 1 Cor. 13:2 ( aABCL); mhqe,n in Ac. 27:33 ( aAB). But evxouqene,w in the LXX and the N. T. prevails, though W. H. (after BD) read evxoudenhqh|/ in Mk. 9:12. a and aD read the Attic pandokei/on, - eu,j in Lu. 10:34 f., but W. H. accept pandocei/on eu,j (from de,comai). Sa,repta in Lu. 4:26 is the LXX rendering of tp;r.c. Tropofore,w and trofofore,w are two distinct words, though the MSS. differ widely in Ac. 13:18, the Neutral and Western supporting trop Hort considers sfuri,j for spuri,j right (Mt. 15:37, etc.). It is well attested by the papyri.293 W. H. read fo,bhqron, not fobhtron, in Lu. 21:11.

(h) VARIABLE FINAL CONSONANTS. The use of n evfelkustiko,n (paragogic n) cannot be reduced to any clear rule. The desire to avoid hiatus extended this usage, though it probably originally had a meaning and was extended by analogy to cases where it had none. Cf. English articles a, an (Giles, Man. of Comp. Philol., p. 208).

220 A GRAMMAR OP THE GREEK NEW TESTAMENT

The same thing is true of movable final j. In the old Attic before 403 B.C. this movable n was seldom used. It is more frequent in the new Attic up to 336 B.C., and most common in the koinh,, vanishing again in the modern Greek, as n easily disappears in pronunciation. Meisterhans294 has an interesting table on the subject, showing the relative frequency in different centuries. This table proves that in the koinh, it came to be the rule to use the movable n both before consonants and vowels. This is shown also by the inscriptions295 and the Ptolemaic papyri. Per contra note the disappearance of final n in modern Greek vernacular, when not pronounced (Thumb, Handb., pp. 24 ff.). However, as a rule, this movable final n occurs only with the same classes of words as in the Attic as after - si esti, and e in verbs (3d sing. past tenses). The irrational n mentioned as common later by Hatzidakis296 is rare. The older N. T. MSS. ( aABC) are in harmony with the koinh, and have the movable n and j both before consonants and vowels with a few exceptions. The later N. T. MSS. seem to feel the tendency to drop these variable consonants. Moulton297 mentions mei,zwn (Jo. 5:36) as a good example of the irrational n in N. T. MSS. (ABEGMA). Cf. also the irrational n with the subjunctive in the papyri. So eva.n h=n a;rsenon P. Oxy. 744 (i/B.C.) for h|=) See Moulton, Prol., pp. 168, 187, for further examples. The failure to use this n was originally most common in pause, sometimes even before vowels.298 Blass299 observes that it was only the Byzantine grammarians who made the rule that this n should be used before vowels and not before consonants, a rule of which their predecessors did not have the benefit, a thing true of many other grammatical rules. We moderns can teach the ancients much Greek! Since the N. T. MSS.300 show no knowledge of this later grammatical "rule," W. H. follow a mechanical one indeed,

ORTHOGRAPHY AND PHONETICS 221

but the only practical guide under the circumstances. They go by the testimony of the oldest uncials. Hort gives a considerable list of examples where the n is wanting in one or more of the older uncials, but where W. H. have n, as in avrou/sin (Mt. 4:6), pa/sin (Mt. 5:15), etc. But in Lu. 1:3 e;doxe is read by aBCD. In Ac. 24:27 kate,lipe is supported by aB. There are about a dozen more instances in Hort's long list of alternative readings where W. H. prefer the form without n, rather more frequently after at, than after e.301 W. H., however, have ei;kosi everywhere, as was usually the case in the Attic inscriptions and always in the Ptolemaic papyri and the LXX MSS. both before vowels and consonants.302 So e;mrposqen e;xwqen o;pisqen in the N. T. Likewise pe,rusi is correct in 2 Cor. 8:10; 9:2.303

The variables calls for a few words more. All good MSS. give a;ntikruj Ci,ou in Ac. 20:15.304 But as in Attic, the N. T. MSS. usually have a;cri and me,cri even before vowels. ;Acri (always before consonants) thus precedes vowels some fifteen times, and once only do we certainly305 have a;crij (Gal. 3:19), though it is uncertain whether it is followed by a;n or ou-. Me,cri is always used in the N. T. before a consonant and once before a vowel, me,cri vIwa,nou (Lu. 16:16). The early N. T. editors used to print ou[tw before consonants and ou[twj before vowels, but W. H. print ou[twj 196 times before consonants and vowels and only ten times ou[tw (all before consonants). These ten instances are Mk. 2:7; Mt. 3:15; 7:17; Ac. 13:47; 23:11; Ro. 1:15; 6:19; Ph. 3:17; Heb. 12:21; Rev. 16:18.306

(i) METATHESIS. Failo,nhj (2 Tim. 4:13), Latin paenula. See Additional Notes.

IV. Breathings.

(a) ORIGIN OF THE ASPIRATE. As is well known, in the modern Greek no distinction is made in pronunciation between spiritus asper and spiritus lenis, or pneu/ma dasu, and pneu/ma yilo,n. That

222 A GRAMMAR OF THE GREEK NEW TESTAMENT

is to say, the "rough" breathing is only a conventional sign used in writing. This sign is indeed a comparatively modern device, 'and', in use in the MSS. generally since the eleventh century A.D.307 This form was an evolution from H (Phoenician ?, he), then ? and ?, then ? and .308 This breathing (rough or smooth) did not find a place in the Greek alphabet, and so is not found in the early uncial MSS. It becomes therefore a difficult question to tell whether the modern ignoring of the rough breathing was the rule in the first century A.D. The MSS., as Hort309 points out, are practically worthless on this point. The original use of H as equal to h or the rough breathing was general in the old Attic and the Doric, not the AEolic and Ionic. And even in the Attic inscriptions the usage is very irregular and uncertain. Numerous examples like HEKATON occur, but some like HEN also, so that even H was not always rough.310 The modern English cockneys have no monopoly of trouble with h's. In French h is silent as l'homme. The Greeks always found the matter a knotty problem. The use of H= h in the Ionic and Attic (after 403 B.C.) left the Greeks without a literary sign for h. The inscriptions show that in the vernacular H continued to be so used for some time.

(b) INCREASING DE-ASPIRATION (Psilosis). But there was a steady decrease in the use of the h sound. The Ionic, like the AEolic, was distinguished by psilosis, and the koinh, largely311 followed the Ionic in this respect. More certain is the use of the aspirated consonants c q f, which succeeded the older KH, TH, PH.312 But certainly the rough breathing was in early use as the

ORTHOGRAPHY AND PHONETICS 223

inscriptions show, though not with much consistency.313 Sometimes the rough breathing may be due to the disappearance of a digamma, though sometimes a smooth breathing displaces it, as e;rgon from F e,rgon314 (cf. English 'work'). Then again the disappearance of s has the same result, as ivsaro,ji`ero,j.315 It is not strange therefore that usage in the koinh, is not uniform. Examples like u`po v auvtou/ u`f v auvtou/ ouvk e`wrw/men, etc., appear in the Pergamum inscriptions, not to mention kaq v e[toj kaq v i`di,an, etc.316 The same story of uncertainty is told elsewhere in the koinh, as in Magnesia,317 Herculaneum.318 Some of this variation is probably due to analogy,319 so that though "de-aspiration was the prevailing tendency,"320 yet the N. T. shows several examples in the opposite direction.

(c) VARIATIONS IN THE MSS. (Aspiration and Psilosis). The aspiration of the consonants k p t in case of elision is therefore a matter of documentary evidence321 and occurs in the case of avnti, evpi, kata, meta, ouvk u`po,. The N. T. MSS. vary considerably among themselves as in the LXX, though some like D in the Gospels and Acts are wholly untrustworthy about aspiration.322 In general Attic literary usage cannot be assumed to be the koinh, vernacular. Hort323 prefers `Adramunthno,j (Ac. 27:2) like Hadrumetum. vAloa,w (1 Cor. 9:9 f.; 1 Tim. 5:18) is connected with a[lwj or avlwh, and may be compared with avphliw,thj ( h[lioj).324 Hort (p. 144) prefers avats (Mk. 5:3), but eivlikrinh,j and eivlikrini,a, though ei`l. has ancient authority. vAfelpi,zontej is read by DP in Lu. 6:35 and the LXX has several similar instances,325 not to mention one consonants see Riem. and Goelzer, Phonet., pp. 194 ff., and for the dialects and interaspiration see K.-B1., Bd. I, pp. 107-114.

224 A GRAMMAR OF THE GREEK NEW TESTAMENT

Addenda 3rd ed.

in Hermas and in the Attic.326 In Ro. 8:20 W. H. accept evf v evlpi,di, while various MSS. support it in Ac. 2:26; 1 Cor. 9:10; Ro. 4:18; 5:2; Tit. 1:2, and FG have kaq v evlpi,da in Tit. 3:7. Hort327 thinks this is due to digamma dropped as well as in the case of avfi,dw (Ph. 2:23), but analogy to acpopav may be the explanation.328 ;Efode is read by a few MSS. in Ac. 4:29 as evfi/den in Lu. 1:25. Gregory329 gives many examples of avf evf kaq with evlpi,zw and ei=don in the LXX. W. H. offer ouvc i`dou, as an alternative reading in Ac. 2:7, while B reads ouvc i`do,ntej in 1 Pet. 1:8 and ouvc ei=don in Gal. 1:19. A has ouvc o[yesqe in Lu. 17:22. W. H.330 put ouvc `Ioudai?kw/j in the margin in Gal. 2:14. Kaq v i`di,an appears in a once, in B eight times, in D three times, in D once (Mt. 14:23; 17: 1, 19; 20:17; 24:3; Mk. 4:34; 6:31; 9:28; 13:3). But W. H. nowhere accept it, not even when B combines with a or D. aB have it in Mt. 24:3. The form kaq v i`di,an is common in the koinh, inscriptions and the papyri. Kaqei,dwlon is read by M in Ac. 17:16. On the other hand kaq v e[toj, so common in the koinh, (cf. Latin vetus), is not found in the N. T., all MSS. in Lu. 2:41 reading kat v e;toj. Hort331 considers ouvk e;sthken (Jo. 8:44) to be merely the imperfect indicative of sth,kw. So also as to e;sthken in Rev. 12:4. a has evfiorkh,seij in Mt. 5:33, a form common in the Doric inscriptions.332 DP have evfi,orkoj in 1 Tim. 1:10. In Rev. 12:11 A reads ouvc h`ga,phsen, while ouvc o`li,goj is read in the LXX and papyri as well as a number of times in Ac. Ac.(12:18 by aA, 14:28 by a, 17:4 by B, 19:23 by aAD, 19:24 by a, 27:20 by A). In Ac. 5:28 D has evfagagei/n. W. H. print on the other hand avpokatista,nei in Mk. 9:12 rather than avpokatasta,nei though with hesitation.333 So likewise W. H. give evpi,statai instead of evfi,statai

ORTHOGRAPHY AND PHONETICS 225

in 1 Th. 5:3 (like B in Sap. 6:8), a wholly unusual' absence of aspiration in compounds of i[sthmi. For the LXX phenomena see Thackeray, Gr., p. 127 f. It is wholly doubtful whether ovmei, romai or o`mei,romai is right (1 Th. 2:8). Ouvk eu=ron in some MSS. in Lu. 24:3, and ouvk e;neken in 2 Cor. 7:12, Blass334 considers as clerical errors, though they are common in the LXX and in the inscriptions.335 N. T. MSS. (late cursives) even have ai`te,w o`stew,n, o[cloj, etc. For mhqei,j ouvqei,j see this chapter p. 219, the Interchange of Consonants and chapter on Pronouns, pp. 750 f.

(d) TRANSLITERATED SEMITIC WORDS. The aspirate in the case of transliterated Semitic words (chiefly proper names) causes some difficulty. Blass336 calls it "insoluble," though he accepts Hort's practice as rational,337 expressing a and [ by the smooth breathing and h and x by the rough breathing. The MSS. disagree and are not consistent, but Blass calls the result of this procedure "strange." Hence Hort argues for [Abel ( h) , vAbraa,m ( a), [Agaboj ( [), [Agar ( h), `Akeldama,c ( x), a`llhlou,i?a ( h), `Alfai/oj ( x), `Anani,aj ( h), [Anna ( x), `Are,taj ( x), `Arimaqai,a ( h), [Ar Magedw,n ( h), vEbe,r ( [), vEbrai/oj ( [), vEbrai<j ( [), vEbrai?sti, ( [),338 vElisai/oj ( [), vElmada,m ( a), evlwi ( a), `Emmw,r ( x), `Enw,c ( x, but vEnw,j, a), `Errw,m ( h, but vEslei,, a), Eu[a ( x), hvlei, ( a), but `Hlei, ( h), vHlei,aj ( a), ;Hr ( [), u;sswpoj ( a),339 w`sanna, ( h), `Wshe, ( h). Hort340 gives, moreover, the smooth breathing to all names beginning with as y as vHsai,aj. Besides he considers it a "false association"341 to connect vIeremi,aj, vIerieicw, vIeroso,luma $mei,thj% vIerousalh,m with i`ero,j, though Blass retains `Ieroso,luma rather inconsistently.342

(e) THE USE OF BREATHINGS WITH r AND rr. W. H. follow Tischendorf and Lachmann in dropping the breathings in rr as in a;rrhta (2 Cor. 12:4), though retaining the rough breathing with initial r as in r`h,mata (Ib.). Winer343 argued that the Romans heard an aspiration with rr, since they used Pyrrhus, Tyrrhenus, etc. W. H. seem justified in using the smooth breathing with the first r in the word rverantisme,noi. (Heb. 10:22) by old Greek cus-

226 A GRAMMAR OF THE GREEK NEW TESTAMENT

tom.344 The MSS., of course, give no help in the matter. The breathing with r is not written in the modern Greek vernacular text as in Pallis or Thumb.

(f) THE QUESTION OF Au`tou/. This is somewhat knotty. It seems clear that as a rule auvtou/ and not au`tou/ is to be printed in the N. T. A number of reasons converge345 on this point. The older Greek often used au`tou/ rather than e`autou/ as shown by the aspiration of the prepositions like avf v au`tou/, etc. In the N. T. there is not a single case of such aspiration after elision save in a few single MSS. Add to this the fact that the N. T. uses the reflexive pronoun much less than the earlier Greek, "with unusual parsimony" (Hort). Besides the personal pronouns of the first and second persons are frequently employed (Buttmann) where the reflexive might have been used. Buttmann urges also the point that in the N. T. we always have seautou/, not sautou/. The earliest uncial MSS. of the N. T. and the LXX that use the diacritical marks belong to the eighth century, but they all have auvtou/, not au`tou/. Even in the early times it was largely a matter of individual taste as to whether the personal or the reflexive pronoun was used. Blass (p. 35) indeed decides absolutely against au`tou/) But the matter is not quite so easy, for the Kotin' inscriptions give examples of u`f au`tou/ in first century B.C. and A.D.346 Mayser347 also gives a number of papyri examples like kaq v au`tou/ meq v au`tou/ u`f v au`tw/n, where the matter is beyond dispute. Hort agrees with Winer in thinking that sometimes au`tou/ must be read unless one insists on undue harshness in the Greek idiom. He instances Jo. 2:24, auvto.j de. vIhsou/j ouvk evpi,steusen a`uto.n auvtoi/j and Lu. 23:12, prou?ph/rcon ga.r evn e;cqra| o;ntej pro.j au`tou,j. There are other examples where a different meaning will result from the smooth and the rough breathing as in 1 Jo. 5:10 ( au`tw|/), 18 ( auv to,n auvtou/, Eph. 1:5 ( auvto,n), 10 ( auvtw|/), Col. 1:20 ( auvto,n), 2:15 ( auvtw|/). W. H. print au`tou/ about twenty times. Winer leaves the matter "to the cautious judgment of the editors."

V. Accent.

(a) THE AGE OF GREEK ACCENT. The MSS. are worth as little for accent as for breathings. The systematic application of accent in the MSS., like the regular use of the spiritus lenis, dates

ORTHOGRAPHY AND PHONETICS 227

from the seventh century A.D.348 Hort349 caustically remarks that most modern grammarians have merely worked out "a consistent system of accentuation on paper " and have not recovered the Greek intonations of voice, though he has little to offer on the subject. Chandler350 indeed laments that modern scholars scatter their Greek accents about rather recklessly, but he adds: "In England, at all events, every man will accent his Greek properly who wishes to stand well with the world." It is a comfort to find one's accents irreproachable, and Chandler rightly urges that the only way to use the accents properly is to pronounce according to the accent. The ancients were interested in Greek accent. Herodian in his Kaqolikh. prosw|di,a investigated the accent of 60,000 words, but the bulk of his twenty books is lost. Chandler351 found most help from Gottling, though others have written at length on the subject.352 There are no accent-marks in the early inscriptions and papyri; in fact tradition ascribes the invention of these signs as a system to Aristophanes of Byzantium in the third century B.C., though the beginnings appear in the preceding century.353 He and his disciple, Aristarchus, made the rules at any rate.354 The Alexandrian grammarians developed these rules, which have shown a marvellous tenacity even to the present day in the modern Greek, though, of course, some words would naturally vary in accent with the centuries.355 There is the Harris papyrus of Homer in the first century A.D. which has accents, and clearly the word had the accent in pronunciation like English long before it was written out. After the fourth century A.D. the use of accentual rhythm in Greek in place of quantitative rhythm had a tendency

228 A GRAMMAR OF THE GREED NEW TESTAMENT

to make the accent rather more stable.356 "Of all the phonetic peculiarities of a language accent is the most important."357 The earlier use of accents and breathings was probably "for the text of poetry written in dialect"358 (cf. our reading-books for children). They were not written out "in ordinary prose till the times of minuscule writing," though Euthalius (A.D. 396) made use of them in his edition of the N. T.359 The Christian hymns early show signs of changing from tone (pitch) to stress as is the rule in modern Greek. Cf. Thumb, Handb., p. 6.

(b) SIGNIFICANCE OF ACCENT IN THE Koinh,. In Greek it is pitch, not stress, that is expressed by the accent, though in modern Greek the accents indicate stress. "In the ancient Sanskrit and the ancient Greek the rise and fall in musical tone was very marked."360 In English we are familiar with stress-accent. "Hadley has ably argued that the compass of tone used by the Greeks was a musical fifth, i.e. from C= do to G= sol, involving also the intermediate third or E= me."361 It was not a stronger current of breath,362 but a higher musical note that we have. It was in a word "das musikalische Moment."363 Hadley ("Nature and Theory of Gk. Accent," Essays Philol. and Crit., p. 111 f.) points out that prosw|di,a comes from a root meaning 'to sing' (like the Latin accentus) and so ovxu,j and baru,j answer to our high and low pitch. Giles364 thinks that in the original Indo-Germanic language pitch and stress-accent were more evenly balanced. The accent singles out one syllable sharply and raises it higher than the rest, though as a matter of fact each syllable in a word has an accent or pitch lower down in the scale: Cf. the secondary accent in the English "incompatibility." The Harris papyrus of Homer even accents every syllable in each word.365 Then again "the accent of a sentence is as much under the influence of a law of some kind as the accent of the word."366 Language without accent or musical va-

ORTHOGRAPHY AND PHONETICS 229

riety in tone would be hopelessly monotonous and ineffective. An instance of the importance of accent and breathings is seen in ou- ouv Ac. 19:40.

(c) SIGNS OF ACCENT. In practical usage (in our school grammars) there is only one distinction, the accented syllable and the unaccented syllables. The Greeks themselves distinguished the pronunciation of the acute and the circumflex. The difference is well illustrated by ei=mi and eivmi,. The three signs (acute or ovxei/a, grave or barei/a, circumflex or perispwme,nh) come to symbolize the higher pitch of the accented syllable.367 Originally the accented syllable was marked by the acute and all the unaccented syllables by the grave (merely the absence of the acute), but by and by this use of the grave accent was felt to be useless and was dropped.368 Then the grave accentual mark of falling inflection was used for the acute when an oxytone word comes before another word (not enclitic), though this "grave" accent has the pitch of the unaccented syllable. Similarly in contraction of two syllables with acute and grave ( , .) arose the circumflex, the grave and the acute making acute still. The actual use in pronunciation of both acute and grave in the contracted syllable disappeared, so that the circumflex in pitch differed little, if any, from the acute. The difference, for instance, between the acute in dhlw,sai, and the circumflex in dhlw/sai at was not perceptible in sound.369 The Greek and the Latin agree in having the accent only on one of the three last syllables and thus differ from English and French for instance. It is not necessary here to go into the rules (not wholly arbitrary) which the Greeks developed for the accent of words. In the use of unaccented words (proclitics or enclitics) Greek does not differ radically from English. If the Greek has evn oi;kw|, the English has "at-home." If the Greek has eivpe, moi, the English has "tell-me."370

(d) LATER DEVELOPMENTS IN ACCENT. There was not indeed uniformity among the dialects in the use of accent. They agreed only in the one point of not accenting further back than the third syllable from the end. "In other respects the Greek dialects show the widest divergencies in their accentuation. The two antipodes are AEolic and Doric, which are so closely allied phonetically: AEolic throws the accent as far back as possible in

230 A GRAMMAR OF THE GREEK NEW TESTAMENT

all words, e.g. basi,leuj= basileu,j, . . .; Doric, on the contrary, faithfully preserves the original oxytone accent. Between these two dialects lie Ionic and Attic, which, however, are much nearer to Doric than to AEolic. But all the dialects, including Doric, observe the rule that, in those forms of the verb which are capable of being conjugated, the accent goes back as far as possible."371 AEolic, for instance, has h[ sh where the Attic has h` sh,) But all the dialects372 have evgw, e;gwge. On this point in general see Kuhner-Blass, I, pp. 323 ff. The Dorians even had avnqrw,poi, evlu,san, etc. Perfect uniformity was no more possible in Greek than in English. The modern Greek preserves the three-syllable accent rule. Examples like e;piase evbra,duase are not exceptions, since the i and u count as consonants. Cf. Thumb, Handb., p. 28. French follows tone like the ancient Greek. Pecheur is 'fisher,' while pecheur is 'sinner,' for example, a difference only in quality, not in accent.

(e) N. T. PECULIARITIES. Where so much is in doubt, excessive refinement is certainly not desirable. But the following points call for remark, and Gregory373 can be consulted for the actual evidence (very slight) from the N. T. MSS. on the subject of accent. D alone among the older uncials has the accent (and that the occasional circumflex) save by the hand of a corrector.

1. Shortening Stem-Vowels. There is quite a tendency in the koinh, towards shortening some of the stem-vowels, especially in words in -- ma. Hence W. H. do not follow the Attic accent here, but that of the koinh,, and give us kli,ma kri,ma mi,gma (cf. e[ligma), po,ma cri,sma, though as to cri,sma Blass374 suggests that cri/sma is correct because of cristo,j and because B (1 Jo. 2:20, 27) has crei/sma. Analogy plays havoc with rules. Herodian375 says that i and u were usually shortened before So W. H. give us kh/rux, khru,xai sthri,xai (Ro. 16:25), probably foi/nix coi/nix. According to Winer-Schmiedel376 this rule applies to y also, but W. H. and Blass377 do not agree. So W. H. have qli,yij r`i,yan (Lu.

ORTHOGRAPHY AND PHONETICS 231

4:35). By parity of reasoning W. H. reject the circumflex accent in e`lku,sai li,non mu,ron spi,loj stu,loj suntetri,fqai (Mk. 5:4), though suntri/bon (Lu. 9:39) and sku/la (Lu. 11:22). Cf. mu/qoj margari/tai ni/koj si/toj su/kon etc. W. H. read yu,coj also. The length of u in ku,ptw is uncertain; avnaku,yai and paraku,yai usually appear in the N. T. W. H. have, however, kra/zon in Gal. 4:6 and lai/lay in Mk. 4:37. But e`sta,nai (Ac. 12:14) is right, though a=rai (Mt. 24:17), qumia/sai (Lu. 1:9) because of long a. Cf. also evpa/rai (Lu. 18:13), evpifa/nai (Lu. 1:79), pra/xai. (Ac. 26:9), but pia,sai (Jo. 7:30). So katalu/sai (Mt. 5:17), kateuqu/nai (Lu. 1:79) and kwlu/sai (Ac. 10:47).

2. Separate Words. These are not so easily classified. W. H. read avgorai/oi not avgo,raioi; a;ntikru not avntikru,* avnti,pera, not avnti pe,ra$n%* avpo,dektoj, not avpodekto,j but evklekto,j euvloghto,j misqwto,j* avreski,a (from avreskeu,w) with which compare evriqi,a (from evriqeu,w); avreski,a (Attic a;creioj), as also e;rhmoj (Attic evrh/moj), e[toimoj (Attic e`toi/moj% mwro,j (Attic mw/roj), o[moioj (Attic o`moi/oj), clwro,j (Attic clw/ roj%* braduth,j (3d decl.), but a`dro,thj (3d decl.); gazofula,koin, not ei/on and eivdw,lion, with which compare telw,nion glwsso,komon being for the earlier glwssoko,mion* de,smh not desmh,* dieth,j (Mt. 2:16), not die,thj (Attic), and so with other compounds of - ethj, like e`katontaeth,j, etc., but e`katontarcw/n (Ac. 23:17) is from - a,rchj, not arcoj* eivpo,n is the imperative (Mt. 18:17), for ei=pon is only Attic, and Charax calls eivpo,n Syracusan,378 with which one may compare i;de ( ivde, only Attic according to the Alexandrian grammarians, though Bornemann urged ivde, when verb and i;de when exclamation) and la,be ( labe, only Attic); qrhsko,j (Jas. 1:26), not qrh/skoj* i`drw,j (Lu. 22:44), not i`drw/j* i`ma,nta (Mk. 1:7), not the Attic i`ma/nta* i;soj, not the i=soj379; ivcqu,j (Mt. 7:10), not ivcqu/j; ovsfu,j (Mt. 3:4), not ovsfu/j; ivscu,j, not ivscu/j; klei,j in nominative singular (Rev. 9:1), though klei/jgrk grk(1:18) and klei/daj (Mt. 16:19) in accusative plural, etc., with which compare pou,j (Mk. 9:45), not pou/j and sh,j (Mt. 6:19), not sh/j* kti,sthj (1 Pet. 4:19), not ktisth,j, as gnw,sthj, etc.; kru,pth, not krupth, (Lu. 11:33); mogi la,loj (Mk. 7:32), not - la/loj* mulw,n (Mt. 24:41) is read only by DHM and most of the cursives, mu,loj being correct; muria,dwn ( a,j) as in Lu. 12:1; Rev. 5:11, not the Attic muriadw/n, and so as to cilia,dwn* ovrguia, (Ac. 27:28), not o;rguia; ouva, (Mk. 15:29), not ouva/* poi,mnion (Lu. 12:32), not poimni,on, and tru,blion in Mk. 14:20

232 A GRAMMAR OF THE GREEK NEW TESTAMENT

Addenda 3rd ed.

(called no diminutive by some),380 but tekni,on always; plh,mmura (Lu. 6:48) is preferred by Winer-Schmiedel381 as nominative to plhm mu,rhj rather than - mu,ra* ponhro,j always, not po,nhroj in the physical sense (Rev. 16:2) and ponhro,j in the moral (Gal. 1:4)382; prw|/ra (Ac. 27:41), not prw,ra* spei/ra (Mk. 15:16), not spei,ra* flu,aroj (1 Tim. 5:13), not fluaro,j. The compound adverbs evpe,keina u`per e,keina have thrown back the accent.

3. Difference in Sense. With some words the accent makes a difference in the sense and is quite important. We have, for instance, [Agia, not a`gia., in Heb. 9:2. W. H. read avlla,, not a;lla, in Jo. 6:23. In Jas. 1:15 W. H. have avpokuei/ (from - e,w), not avpoku,ei (from - ku,w). So W. H. print a=ra (interrog.) in Gal. 2:17, not a;ra (illative). Auvth, and au[th are easily confused, but W. H. prefer au[th to auvth, in Mt. 22:39 ( auvth|/ in margin); Ro. 7:10; 1 Cor. 7:12; and auvth, to au[th in Lu. 2:37; 7:12; 8:42; Ro. 16:2. In Rev. 2:24 the adjective baqe,a is correct, not the substantive ba,qea (uncontracted from ba,qoj). Dexiola,boj or dexio,la boj is possible in Ac. 23:23 (cf. Winer-Schmiedel, p. 69). So W. H. give us evgcri/sai (infinitive) in Rev. 3:18, not e;gcrisai (imperative). Cf. also evpitimh,sai (Jude 1:9), optative, not infinitive h/sai. Note the difference between fobhqh/te (subjunctive) and fobh,qhte (imperative) in Lu. 12:5. In Jo. 7:34, 36, W. H. prefer eivmi, rather than ei=mi (not elsewhere used in the N. T. save in composition with prepositions avpo, eivj evx evpi, su,n). In Mk. 13:28 and Mt. 24:32 W. H. have evkfu,h| (present active subjunctive), not evkfuh|/ (second aorist passive subjunctive). In Lu. 19:29; 21:37 W. H. prefer vElaiw/n, not vElaiw,n (the correct text in Ac. 1:12, and possibly in Luke also according to the papyri, though vElaiw/na would be the form expected).383 In Mk. 4:8, 20, W. H. put evn in the text and e[n in the margin. ;Eni, not evni,, occurs with ouvk several times, once (1 Cor. 6:5) ouvk e;ni evn) In Lu. 9:38, W. H. read evpible,yai (infinitive), not evpi,bleyai (imperative). In 1 Cor. 5:11 W. H. read h|= (subjunctive), not h; (conjunction as Rec.). In Ro. 1:30 W. H. follow most editors in giving qeostugei/j (passive), not qeostu,geij (active sense of the adjective). In Mk. 5:29 all editors have the perfect i;atai, not the present iva/tai. In Lu. 22:30 W. H. read kaqh/sqe (subjunctive), not ka,qhsqe (indicative) nor kaqh,sesqe (future, margin). In 1 Cor. 9:21 W. H. prefer kerdanw/ (future indicative) to kerda,nw (aorist subjunctive), and in

ORTHOGRAPHY AND PHONETICS 233

Addenda 3rd ed.

1 Cor. 6:2 krinou/sin (future) to kri,nousin (present indicative in marg.). In Mk. 12:40 we have makra,, not makra|/. In 1 Cor. 3:14 W. H. prefer menei/ (future) to me,nei (present), and in Jo. 14:17 they have me,nei. In 1 Cor. 4:15 (14:19) and Mt. 18:24 no distinction can be made in the accent of muri,oi ('innumerable') and mu,rioi ('ten thousand') because of the cases. Dr. E. J. Goodspeed, of Chicago University (Expository Times, July, 1909, p. 471 f.), suggests wvfelh,qhj in Mk. 7:11 instead of wvfelhqh|/j. It is entirely possible. In 1 Cor. 14:7 o[mwj is correct, not w`mw/jo`moi,wj. In Jo. 18:37 W. H. give ouvkou/n, not ou;koun, in Pilate's question. In Ac. 28:6 W. H. print pi,mprasqai ( mi verb), not pempra/sqai ( w verb). In Rev. 17:5 pornw/n (feminine) is probably right, not po,rnwn (masculine). Prwto,tokoj (Col. 1 : 15), not prwtoto,koj, is manifestly right. The difference between the interrogative ti,j and the indefinite ti.j calls for frequent attention. In Heb. 5:12 W. H. have tina,, not ti,na, but in Heb. 3:16 ti,nej, not tine,j, and in 3:17 ti,sin, not tisi,n. While in Mt. 24:41, 1 Th. 4:6, 1 Cor. 15:8 and 16:16 the article tw|/ is to be read, not the indefinite tw|, which form does not occur in the N. T. In 1 Cor. 10:19 ti, evstin (twice) is not interrogative, but the enclitic indefinite with the accent of evstin. In Jas. 3:6 troco,j ('wheel') is properly read, not tro,coj ('course'). In Mk. 4:12 W. H. read suni,wsin, not suniw/sin, as suni,ousin in Mt. 13:13. Winer384 considers the suggestion of fwtw/n for fw,twn in Jas. 1:17 "altogether absurd."

4. Enclitics (and Proclitics). Proclitics are regular in the N. T. The accent of enclitics calls for comment. As a rule W. H. do not accent them. So we have auvto,n tinaj (Mk. 12:13), ei=nai, tina (Ac. 5:36), ivdou, tinej (Mt. 28:11), o`do,n eivsin (Lu. 8 : 12), avsu,netoi, evste (Mk. 7:18), ga,r evste (Mk. 13:11), kai, fhsi (Ac. 10:31; 25:24). However, plenty of cases call for accent on the enclitic, as, for example, in eu`rei/n tina,j (Ac. 19:1) for emphasis, ga,r fhsi,n (Heb. 8:5 and cf. Mt. 14:8; Ac. 25:5, 22; 26:25; 1 Cor. 6:16; 2 Cor. 10:10) for clearness in punctuation, kai. eivsi,n (Mt. 19:12 and cf. Ac. 5:25) for emphasis, qeou/ evsme,n (1 Jo. 3:2), u`po. tinw/n (Lu. 9:8) likewise, ouvk eivmi, (Jo. 1:21). In o[pou eivmi, (Jo. 7:34, 36) the accent is regular, though some critics wrongly prefer ei=mi.

The use of evsti,n and e;stin demands special comment. When unemphatic, not at the beginning of a sentence, not preceded by avll v eiv kai, ouvk o[ti tou/t v or a paroxytone syllable, as, for example, in vIoudai,wn evsti,n (Jo. 4:22), we have unaccented evstin as in avgro,j evstin (Mt. 13:38, 39), kaqw,j evstin (1 Jo. 3:2), etc. In some ex-

234 A GRAMMAR OF THE GREEK NEW TESTAMENT

amples of mild emphasis W. H. have evsti,n, as in nu/n evsti,n (Jo. 4:23; 5:25), pou/ evsti,n (Mt. 2:2; Mk. 14:14). But the cases are numerous where e;stin is correct, as when it is emphatic, and expresses existence or possibility, as in ei=dej e;stin (Rev. 17:18), auvtou/ e;stin (Ac. 2:29), a[gion e;stin (Ac. 19:2), o` ei-j e;stin (Rev. 17:10), ouvdei.j e;stin (Lu. 1:61; 7:28; 18:29). ;Estin is also the accent at the beginning of sentences, as in Jo. 21:25; 1 Cor. 15:41; 1 Jo. 5:16 f.; Heb. 11:1. Cf. evsti,n in Col. 1:15 and e;stin in 1:17. Then again we have, according to the usual rule, e;stin after avll v (Jo. 13:10), eiv (1 Cor. 15:44), kai. (Mk. 12:11; 2 Cor. 4:3), o[ti (2 Th. 2:4; Mk. 6:55; Heb. 11:6), but o[ti evsti,n (Ac. 23:5) when the idea of existence is not stressed, ouvk (1 Cor. 11:20; Ro. 8:9, etc.), tou/t v (Mk. 7:2; Ro. 7:18). W. H. give only evsti,n after pou/ (Jo. 9:12; 11:57; Mk. 14:14).

Sometimes two enclitics come together. Here the critics differ and W. H.385 do not make clear the reasons for their practice. In Ac. 13:15 W. H. have ei; tij e;stin and in Gal. 6:15 peritomh, ti e;stin, because they take go e;stin to be emphatic in both instances. In Jo. 6:51 W. H. have sa,rx mou evsti,n. But in many examples the first enclitic is accented and the second unaccented as in Lu. 8:46 h[yato, mou, tij, 10:29 ti,j evsti,n mou, Jo. 5:14 cei/ro,n soi, ti, 8:31 maqhtai, mou, evste, 12:47 eva,n ti,j mou, 14:28 mei,zwn mou, evstin Ac. 2:25 dexiw/n mou, evstin, 25:5 ei; ti, evstin 25:14 avnh,r ti,j evstin, 1 Cor. 10:19 eivdwlo,quto,n ti, evstin and ei;dwlo,n ti, evstin, 11:24 tou/to, mou, evstin 2 Cor. 11:16 mh, ti,j me Ro. 3:8 kaqw,j fasi,n tinej Heb. 1:10 ceirw/n sou, eivsin, 2:6 de, pou, tij, Tit. 1:6 ei; ti,j evstin. Modern Greek only has a second accent when the accent is in the third syllable as in t v a[rmata, maj (Thumb, Handbook, p. 29).

The personal pronouns now have the accent in W. H. and now are without it, as ovfqalmw|/ sou/ and ovfqalmou/ sou (both in Mt. 7:4). Cf. also evgw, se (Jo. 17:4), su, me grk(17:8), but ti, evmoi. kai. soi, (Lu. 8:28). With prepositions generally the enclitics are accented, as evn soi, (Jo. 17:21), though e;mprosqe,n mou and ovpi,sw mou (Jo. 1:30 both, and so continually with these two prepositions). vEnw,pion evmou/ (Lu. 4:7) and evnw,pio,n mou (Ac. 2:25) both appear. With the prepositions usually evmou/, not mou, occurs as e[neka evmou/ (Mt. 5:11). It is only with pro,j that we have much trouble. The N. T. editors have generally printed pro,j se, but W. H. have that only in Mt. 25:39, elsewhere pro,j se, as in Mt. 26:18. Usually we have, according to W. H., pro,j me as in Mt. 25:36; Jo. 6:65; 7:37, etc., and where the "me" is emphatic in sense,

ORTHOGRAPHY AND. PHONETICS 235

as Mt. 3:14; 11:28, in the first of which Tisch. and Griesbach have pro.j me,, a usage not followed by W. H., though kept in the LXX text of B, as in Is. 48:16; etc.386 W. H. a few times prefer pro.j evme, (not enclitic) as in Lu. 1:43; Jo. 6:35, 37 (both ways here), 44 (marg.), 45; Ac. 22:8, 13; 23:22; 24:19. Occasionally the enclitic tine.j is found at the beginning of a sentence, as in Mt. 27:47; Lu. 6:2; Jo. 13:29; Ph. 1:15; 1 Tim. 5:24.

5. Proper Names cannot always be brought under rules, for in Greek, as in English, men claim the right to accent their own names as they will. On the accent of the abbreviated proper names see chapter V, v. It is difficult to make a clear line of distinction as to why vAnti,paj (Rev. 2:13) is proper, but vArtema/j (Tit. 3:12), save that in vArtemi,dwroj the accent was already after m. But cf. Kleo,paj (Lu.. 24:18) and Klwpa/j (Jo. 19:25).387 In general one may say that proper names (geographical and personal) throw the accent back, if the original adjectives or substantives were oxytone. This is for the sake of distinction. vAlexan drino,j (Ac. 27:6; 28:11) is the adjective. ;Assoj (Ac. 20:13 f.) is doubtless correct, though Pape gives vAsso,j also.388 In vAcai?ko,j (1 Cor. 16:17) the accent is not thrown back nor is it in vApollw,j (1 Cor. 16:12). vAsu,nkritoj (Ro. 16:14) retains the accent of the adjective, like Tro,fimoj (Ac 20:4) and `Ume,naioj (1 Tim. 1: 20). But we have Bla,stoj (Ac. 12:20), Diotre,fhj (3 Jo. 1:9), vEpai, netoj (Ro. 16:5), ;Erastoj grk(16:23), `Ermoge,nhj. (2 Tim. 1:15), Eu;tucoj (Ac. 20:9), Ka,rpoj (2 Tim. 4:13), probably vOnhsi,foroj (2 Tim. 1:16; 4:19), Pa,tara (Ac. 21:1), Pu,rroj (Ac. 20:4), Suntu,ch (Ph. 4:2), Swsqe,nhj (1 Cor. 1:2), Ti,mwn (Ac. 6:5), Tu, cikoj (Ac. 20:4) Fi,lhtoj (2 Tim. 2:17). But Cristo,j always retains the oxytone accent whether proper name (1 Tim. 1:1) or verbal adjective (Mt. 16:16). In 2 Tim. 4:21 Li,noj, not Li/noj, is read. So Ti,toj (2 Cor. 2:13, etc.). In Ac. 27:17 Su,rtij is read by W. H. But fh/lix in Ac. 24:22, etc.

6. Foreign Words. These always give occasion for diversity of usage in transliterating them into another tongue. Blass389 lets the quantity of the vowel in Latin determine the accent in the Greek equivalent for Latin words. So Marcus, Ma/rkoj, etc., but W. H. do not accept this easy principle and give us Ma,rkoj in Ac. 12:25, etc., Kri,spoj (1 Cor. 1:14), etc. W. H. likewise

236 A GRAMMAR OF THE GREEK NEW TESTAMENT

throw the accent back on Latin names like Kou,artoj (Ro. 16:23), Pri,skilla (Ac. 18:2), Se,koundoj (Ac. 20:4), Te,rtullojgrk grk(24:2), but we have on the other hand Gai/oj (Ro. 16:23), not Ga,i?oj Ouvr bano,j (Ro. 16:9), Silouano,j (2 Cor. 1:19), Skeua/j (Ac. 19:14).390

But not even Blass attempts to bring the Semitic words under regular rules. Still, it is true, as Winer391 shows, that indeclinable Semitic words (especially proper names) have the accent, as a rule, on the last syllable, though the usage of Josephus is the contrary, because he generally inflects the words that in the LXX and the N. T. are indeclinable. So vAarw,n vAbaddw,n vAbia, vAbiou,d vAbraa,m, to take only the first two pages of Thayer's Lexicon, though even here we find on the other side [Abel and vAbia,qar. If you turn over you meet [Agar vAda,m vAddei, vAdmei,n vAzw,r, etc. It is not necessary here to give a full list of these proper names, but reference can be made to Lu. 3:23-38 for a good sample. In this list some indeclinable words have the accent on the penult, as vElie,zer (29), Zoroba,bel (27), La,mec (36), Fa,lek (35).392 The inflected Semitic words often throw the accent back, as ;Azwtoj, vIa,kwboj La,zaroj. Many of the Aramaic words accent the ultima, as vAbba, Golgoqa, Korba,n vElwi, sabacqanei,, etc. For further remarks on the subject see Helbing, Gr. d. Sept:, pp. 26-31. The difficulties of the LXX translators are well illustrated here by Helbing.

VI. Pronunciation in the Koinh,. This is indeed a knotty problem and has been the occasion of fierce controversy. When the Byzantine scholars revived the study of Greek in Italy, they introduced, of course, their own pronunciation as well as their own spelling. But English-speaking people know that spelling is not a safe guide in pronunciation, for the pronunciation may change very much when the spelling remains the same. Writing is originally an effort to represent the sound and is more or less successful, but the comparison of Homer with modern Greek is a fruitful subject.393 Roger Bacon, as Reuchlin two centuries later, adopted the Byzantine pronunciation.394 Reuchlin, who introduced Greek to the further West, studied in Italy and passed on the Byzantine pronunciation. Erasmus is indirectly responsible for the current pronunciation of ancient Greek, for the Byzan-

ORTHOGRAPHY AND PHONETICS 237

tine scholars pronounced ancient and modern alike. Jannaris395 quotes the story of Voss, a Dutch scholar (1577-1649), as to how Erasmus heard some learned Greeks pronounce Greek in a very different way from the Byzantine custom. Erasmus published a discussion between a lion and a bear entitled De Recta Latini Graecique sermonis pronuntiatione, which made such an impression that those who accepted the ideas advanced in this book were called Erasmians and the rest Reuchlinians. As a matter of fact, however, Engel has shown that Erasmus merely wrote a literary squib to "take off" the new non-Byzantine pronunciation, though he was taken seriously by many. Dr. Caspar Rene Gregory writes me (May 6, 1912) : "The philologians were of course down on Engel and sided gladly with Blass. It was much easier to go on with the totally impossible pronunciation that they used than to change it." Cf. Engel, Die Aussprachen des Griechischen, 1887. In 1542 Stephen Gardiner, Chancellor of the University of Cambridge, "issued an edict for his university, in which, e.g. it was categorically forbidden to distinguish at from e, ei and oi from i in pronunciation, under penalty of expulsion from the Senate, exclusion from the attainment of a degree, rustication for students, and domestic chastisement for boys!"396 Hence though the continental pronunciation of Greek and Latin was "Erasmian," at Cambridge and Oxford the Reuchlinian influence prevailed, though with local modifications. Geldart,397 however, complains that at Eton, Rugby and Harrow so little attention is paid to pronouncing according to accent that most Greek scholars handle the accents loosely. The Classical Review (April, 1906, p. 146 f.) has the scheme approved by the Philological Societies of Cambridge and Oxford for "The Restored Pronunciation of Latin," which is the virtual adoption of the Continental principle. The modern Greeks themselves rather vehemently insist that ancient Greek should be pronounced as modern Greek is. Muller,398 for instance, calls the "Erasmian" pronunciation "false" because it treats Greek "as dead." Geldart (Modern Gk. Language in Its Relation to Ancient Gr., p. vii) says: "Modern Greek is nothing but ancient Greek made easy." It is not

238 A GRAMMAR OF THE GREEK NEW TESTAMENT

quite as simple as that. Foy399 properly distinguishes between the old Greek vocal sounds and the modern Greek and refers to the development of Latin into the several Romance languages. There is this difference in the Greek, however, that it has only one modern representative (with dialectical variations) of the ancient tongue. One must not make the mistake of comparing the pronunciation of the modern Greek vernacular with the probable pronunciation of the literary Attic of the fifth century B.C. Then, as now, there was the literary and the vernacular pronunciation. The changes in pronunciation that have come in the modern Greek have come through the Byzantine Greek from the koinh, and thus represent a common stream with many rills. The various dialects have made contributions to the pronunciation of the koinh, and so of the modern Greek. In cultivated Athens at its best there was a closer approximation between the people and the educated classes. "Demosthenes, in his oration peri. stefa,nou called AEschines a misqwto,n, but had accented the word erroneously, namely, mi,sqwton, whereupon the audience corrected him by crying misqwto,n."400 Like the modern Italian, the ancient Greek had a musical cadence that set it above all other European tongues.401 We can indeed appeal to the old Greek inscriptions for the popular pronunciation on many points.402 According to this evidence in the first century B.C. in Attica ai=ae, eii hi ui uiu oii, bu (English v).403 Clearly then in the koinh, the process of itacism was already at work before the N. T. was written. What was true of the koinh, vernacular then does not of course argue conclusively for the pronunciation of cultivated Athenians in the time of Socrates. In versatile Athens "a stranger, if introduced on the stage, is always represented as talking the language or dialect of the people to which he belongs."404 Blass indeed thinks that in Tarsus the school-teacher taught Paul Atticistic Greek! " ;Ismen,

ORTHOGRAPHY AND PHONETICS 239

i;ste i;sasin, he must have said, are the true forms which you must employ if you care to be considered a cultivated speaker or writer." Yet in Paul's Epistles he constantly has oi;damen ate asin. The Atticistic pronunciation was no more successful than the Atticistic spelling, forms and syntax. We maybe sure of one thing, the pronunciation of the vernacular koinh, was not exactly like the ancient literary Attic nor precisely like the modern Greek vernacular, but veering more towards the latter. In Greek as in English the pronunciation has perhaps varied more than the spelling. Giles405 observes that English pronunciation "is really a stumbling-block in tracing the history of the English language." Hadley406 has a very able and sane discussion of this matter of changes in Greek pronunciation. He insists on change all through the centuries (p. 139), which is the only rational position. If we turn to the earliest N. T. MSS. we shall find undoubtedly traces of this process of change from the old Attic toward the Byzantine or modern Greek pronunciation. Indeed in the fourth and fifth centuries A.D.,407 the date of the earliest uncials, the process is pretty well complete. The N. T. scribes make no hesitation in writing ai or e* i ei h h|* oi or u according to convenience or individual taste.408 Blass,409 contrary to his former view about Tarsus, says that it is impossible to suppose that there was anybody in the schools at Tarsus who would have taught Paul the correct historical spelling or pronunciation. To the student of the koinh,, as to us, in a sense "the Greek gra,mmata were dead symbols, from which must be recovered the living sounds."410 Of one thing we may be sure, and it is that other dialects besides the Attic contributed to the koinh, pronunciation. The koinh, would be dialect-coloured here and there in its pronunciation. Alexander's conquest, like the railroad and the steamship of the present day, levelled the dialectical variations in many points, whereas before every valley in Greece had its own pronunciation of certain words.411 One taught the koinh, in a Doric environment

240 A GRAMMAR OF THE GREEK NEW TESTAMENT

would show it somewhat. As a matter of fact the Boeotian dialect contributed largely to the koinh, vernacular pronunciation (and so the modern Greek) in points where the Boeotian differed radically from the old Attic.412 Boeotian Greek "modified its vowel-system more than any other Greek dialect."413 Thus already in Boeotian we find both ai and ae in the earliest inscriptions and finally h. So in Boeotian h became414 ei in sound, as evpidei,evpeidh,. The early Greek generally, as already shown, made no distinction in sign between o and w, and h was a slow development from e) The Ionic dialect never took kindly to the rough breathing and greatly influenced the koinh, and so the modern Greek. By the Christian era b is beginning to be pronounced as n, as the transliteration of Latin words like Bergi,lioj shows. Z is no longer ds, but z, though d seems still usually d, not th. Who is right, therefore, the "Erasmians" or the Reuchlinians? Jannaris415 sums up in favour of the Reuchlinians, while according to Riemann and Goelzer416 the "Erasmians" are wholly right. As a matter of fact neither side is wholly right. In speaking of ancient Greek one must recognise other dialects than the literary Attic of the fifth century B.C. If you ask for the pronunciation of the vernacular koinh, of the first century A.D., that will be found as a whole neither in the literary Attic alone nor in the N. T. MSS. of the fifth century A.D. The papyri and the inscriptions of the time throw light on a good many points, though not on all. But even here the illiterate papyri do not furnish a safe standard for the vernacular of a man like Paul or Luke. It is small wonder therefore that N. T. MSS. show much confusion between - sei (future indicative) and sh| (aorist subjunctive), omen (indicative) and - wmen (subjunctive), -- sqai (infinitive) and -- sqe (indicative middle), etc. It is possibly as well to go on pronouncing the N. T. Greek according to the literary Attic, since we cannot reproduce a clear picture of the actual vernacular koinh, pronunciation, only we must understand frankly that this

ORTHOGRAPHY AND PHONETICS 241

is not the way it was done. On the other hand the modern Greek method misses it by excess, as the literary Attic does by default. There was, of course, no Jewish pronunciation of the koinh,. The Coptic shows the current pronunciation in many ways and probably influenced the pronunciation of the koinh, in Egypt. Cf. a German's pronunciation of English.

VII. Punctuation. In the spoken language the division of words is made by the voice, pauses, emphasis, tone, gesture, but it is difficult to reproduce all this on the page for the eye. Many questions arise for the editor of the Greek N. T. that are not easy of solution. Caspar Rene Gregory insists that whenever N. T. MSS. have punctuation of any kind, it must be duly weighed, since it represents the reading given to the passage.

(a) THE PARAGRAPH. As early as Aristotle's time the paragraph ( para,grafoj) was known. A dividing horizontal stroke was written between the lines marking the end of a paragraph. Some other marks like > ( diplh/) or ? ( korwni,j) were used, or a slight break in the line made by a blank space. Then again the first letter of the line was written larger than the others or even made to project out farther than the rest.417 The paragraph was to the ancients the most important item in punctuation, and we owe a debt to the N. T. revisers for restoring it to the English N. T. Cf. Lightfoot, Trench, Ellicott, The Revision of the N. T., 1873, p. xlvi. Euthalius (A.D. 458) prepared an edition of the Greek N. T. with chapters ( kefa,laia), but long before him Clement of Alexandria spoke of perikopai, and Tertullian of capitula. These "chapters" were later called also ti,tloi.418 The sti,coj of Euthalius was a line of set length with no regard to the sense, like our printer's ems. W. H. have made careful use of the paragraph in their Greek N. T. The larger sections are marked off by spaces and the larger paragraphs are broken into smaller sub-paragraphs (after the French method) by smaller spaces.419 Another division is made by W. H. in the use of the capital letter at the beginning of an important sentence, while the other sentences, though after a period, begin with a small letter. This is a wholly arbitrary method, but it helps one better to understand W. H.'s interpretation of the text.

242 A GRAMMAR OF THE GREEK NEW TESTAMENT

W. H.420 have also printed in metrical form passages metrical in rhythm like the Magnificat of Mary (Lu. 1:46-55), the fragment of a hymn in 1 Tim. 3:16, etc., while Lu. 2:14 and the nonmetrical hymns in Revelation are merely printed in narrower columns. The Hebrew parallelism of O. T. quotations is indicated also.

(b) SENTENCES. The oldest inscriptions and papyri show few signs of punctuation between sentences or clauses in a sentence,421 though punctuation by points does appear on some of the ancient inscriptions. In the Artemisia, papyrus the double point (:) occasionally ends the sentence.422 It was Aristophanes of Byzantium (260 B.C.) who is credited with inventing a more regular system of sentence punctuation which was further developed by the Alexandrian grammarians.423 As a rule all the sentences, like the words, ran into one another in an unbroken line (scriptura continua), but finally three stops were provided for the sentence by the use of the full point. The point at the top of the line ( \% ( stigmh. telei,a, 'high point') was a full stop; that on the line (.) ( u`postigmh,) was equal to our semicolon, while a middle point ( stigmh. me,sh) was equivalent to our comma.424 But gradually changes came over these stops till the top point was equal to our colon, the bottom point became the full stop, the middle point vanished, and about the ninth century A.D. the comma (,) took its place. About this time also the question-mark (;) or evrwthmatiko,n appeared. These marks differed from the sti,coi in that they concerned the sense of the sentence. Some of the oldest N. T. MSS. show these marks to some extent. B has the higher point as a period, the lower point for a shorter pause.425 But still we cannot tell how much, if any, use the N. T. writers themselves made of punctuation points. We may be sure that they did not use the exclamation point, the dash, quotation-marks, the parenthesis, etc.426 Parenthetical clauses were certainly used, which will be discussed elsewhere, though no signs were used for this structure by the ancient Greeks. W. H. represent the parenthesis either by the comma (Ro. 1:13) or the dash with comma (1 Tim. 2:7). Instead of

ORTHOGRAPHY AND PHONETICS 243

quotation-marks W. H. begin the quotation with a capital letter with no punctuation before it, as in Jo. 12:19, 21. One way of expressing a quotation was by to,, as in Ro. 13:9. In the case of O. T. quotations the Scripture is put in uncial type (Jo. 12:13). The period ( peri,odoj) gives very little trouble to the modern editor, for it is obviously necessary for modern needs. Here the editor has to make his interpretation sometimes when it is doubtful, as W. H. give e[n) oa} ge,gonen evn, not ea}n oa} ge,gonen) evn (Jo. 1:4). So W. H. read qauma,zete) dia. tou/to Mwush/j in Jo. 7:22, not qauma,zete dia. tou/to) Mwush/j, etc. The colon ( kw/lon),427 'limb of the sentence' formed a complete clause. See Jo. 3:31 for example of use of colon made by W. H. The comma ( ko,mma) is the most common division of the sentence and is often necessary, as with the vocative. So Dida,skale ti, poih,swmen* (Lu. 3:12) and many common examples. In general W. H. use the comma only where it is necessary to make clear an otherwise ambiguous clause, whether it be a participial (Col. 2:2) or conjunctional phrase (Col. 1:23), or appositive (Col. 1:18), or relative (Col. 2:3). The first chapter of Colossians has a rather unusual number of colons (2, 6, 14, 16, 18, 20, 27, 28) as Paul struggles with several long sentences, not to mention the clashes (21, 22, 26). The Germans use the comma too freely with the Greek for our English ideas, leaving out the Greek! Even Winer defended the comma after karpo,n in Jo. 15:2 and o` nikw/n in Rev. 3:12, not to mention Griesbach's "excessive" use of the comma, Winer himself being judge.428 My friend, Rev. S. M. Provence, D.D. (Victoria, Tex.), suggests a full stop before maqw,n in Ac. 23:27 f. That would help the character of Claudius Lysias on the point of veracity.

(c) WORDS. The continuous writing of words without any space between them was not quite universal, though nearly so.429 The oldest Attic inscription (Dipylon vase, probably eighth century B.C.) is written from right to left. With the common method it was not always easy for the practised eye to distinguish between words. Hence there arose the diastolh, or u`podiastolh, a comma used to distinguish between ambiguous words, as o[ ti not o[ti. But W. H. make no use of this mark, not even in o[ ti to distinguish it from the conjunction o[ti. They print uniformly o[ti (Lu. 10:35; Jo. 2:5; 14:13; 1 Cor. 16:2, etc.), not to men-

244 A GRAMMAR OF THE GREEK NEW TESTAMENT

tion doubtful cases like those in Mk. 9:11, 28; Jo. 8:25; 2 Cor. 3:14.430 As to the marks of diaeresis ( ? ) reference may be had to the discussion of diphthongs and diaeresis in this chapter under II (i). W. H., like other modern editors, use the apostrophe ( v ) (or smooth breathing) to represent elision, as avp v avrch/j (Mt. 24:21).431 The coronis is the smooth breathing used also to show when crasis has taken place, as in kavmoi, (Lu. 1:3).432 The hyphen, a long straight line, was used in the Harris-Homer MS. to connect compound words, but it is not in the N. T.433 The editors vary much in the way such words as avlla, ge, i[na ti, tou/t v e;sti etc., are printed. The MSS. give no help at all, for tou/to de, evstin in Ro. 1:12 is not conclusive against tou/t v e;stin elsewhere.434 W. H. prefer a`lla, ge (Lu. 24:21; 1 Cor. 9:2), a=ra, ge (Ac. 8:30), dia, ge (Lu. 11:8; 18:5), ei; ge (2 Cor. 5:3, etc.), kai, ge (Ac. 2:18; 17:27), o[j ge (Ro. 8:32), dia. panto,j (Mk. 5:5, etc.), dia. ti, (Mt. 9 : 11, etc.), i[na ti, (Mt. 9:4, etc.), ei; pwj (Ac. 27:12), mh, pote (everywhere save in Mt. 25:9 where mh,pote% mh, pou (Ac. 27:29), mh, pwj (1 Cor. 9:27, etc.), mh, tij (1 Cor. 16:11, etc.). So also dh/lon o[ti in 1 Cor. 15:27, o[stij ou=n (Mt. 18:4). But on the other hand W. H. print dio,ti as well as ei;te ou;te mh,te w[ste kai,per mh,pote (once), mhde,pote mhde,pw ouvde,pote mhke,ti ouvke,ti mh,pw ou;pw mh,tige, even mh,ge (Mt. 6:1), kaqa, kaqo, kaqw,j kaqa,per kaqo,ti kaqo,lou w[sper w`sei, w`sperei, (1 Cor. 15:8), etc. But W. H. give us kaq v ei-j in Ro. 12:5, avna. me,son in Mt. 13:25, etc.; Kata. mo,naj in Mk. 4:10, kaq v o[son in Heb. 3:3. Adverbs like evpe,keina (Ac. 7:43), u`pere,keina (2 Cor. 10:16), parekto,j (2 Cor. 11:28) are, of course, printed as one word. W. H. properly have u[per evgw, (2 Cor. 11:23), not u`peregw,. In Ac. 27:33 tessareskaide,katoj is one word, but W. H. have `Iera. Po,lij in Col. 4:13 and Ne,a po,lij in Ac. 16:11. It must be confessed that no very clear principles in this matter can be set forth, and the effort of Winer-Schmiedel435 at minute analysis does not throw much light on the subject.

(d) THE EDITOR'S PREROGATIVE. Where there is so much confusion, what is the editor's prerogative? Blass436 boldly advances

ORTHOGRAPHY AND PHONETICS 245

the German idea: "The most correct principle appears to be to punctuate wherever a pause is necessary for reading correctly." But Winer437 shrinks from this profusion of punctuation-marks by the editors, which "often intruded on the text their own interpretation of it." The editor indeed has to interpret the text with his punctuation, but certainly good taste demands that the minimum, not the maximum, of punctuation-marks be the rule. They must of necessity decide "a multitude of subtle and difficult points of interpretation."438 Hort indeed aimed at "the greatest simplicity compatible with clearness," and this obviously should be the goal in the Greek N. T. But the editor's punctuation may be a hindrance to the student instead of a help. It is the privilege of each N. T. student to make his own punctuation.

1 Jannaris, Hist. Gk. Gr., p. 19 f.

2 A Hist. of Egypt, 1906, p. 45.

3 Meisterh., Gr. etc., p. 3; Jann., Hist. Gk. Gr., p. 26 f.; Solmsen, Inscr. Graecae etc., pp. 52 ff.

4 Op. cit., p. 27.

5 Gr. of N. T. Gk., p. 6.

6 Prol., p. 42.

7 Hort, The N. T. in Orig. Gk., App., Notes on Sel. Read., p. 152. But in the Intr. (p. 304) Hort is not willing to admit "peculiarities of a local or strictly dialectic nature" in the N. T. Still Hort (Notes on Orth., p. 151) allows the Doric o`dage,w ( o`dhge,w) in "single MS." like B and D, prosacei/n in B, r`a,ssw in D, etc. Hirt (Handb. d. Griech., p. 53) attributes much of the vocal change to dialect-mixing and analogy. On a and B see Hort, op. cit., p. 306 f.

8 Blass, Gr. of N. T. Gk., p. 6 f.

9 Ib., p. 7. Hort (p. 302 f. of the Intr. to the N. T. in Orig. Gk.) makes a strong defence of his effort to give as nearly as possible "the spelling of the autographs by means of documentary evidence." There must not be "slovenly neglect of philological truth." But Moulton (Prol., p. 47) does not "set much store by some of the minutiae which W. H. so conscientiously gather from the great uncials." Certainly "finality is impossible, notwithstanding the assistance now afforded by the papyri" (Thack., Gr., p. 71).

10 Op. cit.,. p. 303 f. Jann. (Hist. Gk. Gr., p. 35) calls attention to the fact that the professional copyists not only had to copy accurately, but "in the received uniform spelling." Cf. also Helbing, Gr. d. LXX, p. 2. For further remarks on the phenomena in the LXX MSS. see Swete, 0. T. in Gk. p. 300 f.

11 Op. cit., p. 304.

12 Op. cit., p. 308.

13 B. S., pp. 202 ff.

14 Prol., pp. 42 ff.

15 B. S., pp. 202 ff. On the whole subject of the difficulty of N. T. orthog. see W.-Sch., pp. 31 ff. Deiss. (B. S., p. 180) is clearly right in denying a "N. T. orthography" save as individual writers, as now, have their peculiarities. For general remarks about vowel changes in LXX MSS. see Swete, 0. T. in Gk., p. 301 f.; Thack., Gr., vol. I, pp. 71-100; Helbing, Gr., Laut- u. Wortl., pp. 3-14.

16 Nicklin, Cl. Rev., 1906, p. 115, in review of Rutherford's A Chap. in the Hist. of Annotation, 1905.

17 Cf. Bekker, Anec. Gr., vol. II, p. 783.

18 Riem. and Goelzer, Gr. Comp. du Grec et du Lat., Phont., p. 38. Cf. also Donaldson, The New Crat., pp. 207 ff.; K.-B1., Griech. Gr., Tl. I, Bd. I, pp. 39 ff.; Earle, Names of the Orig. Letters of the Gk. Alph. (ClassPapers, 1912, pp. 257 ff.); Flin.-Pet., Form. of the Gk. Alph. (1912). But Sir Arthur Evans gets the Gk. Alph. from Crete.

19 Whitney, Sans. Gr., p. 10.

20 Vergl. Gr., p. 55. His opinion is now considered antiquated.

21 Giles, Comp. Philol., p. 149 f.

22 Telfy, Chron. and Topog. d. griech. Ausspr. etc., 1893, p. 39. See also Larsfeld, Griech. Epig., 1892, pp. 494 ff.; King and Cookson, Sounds and Inflex. in Gk. and Lat., 1888.

23 K.-B1., Tl. I, Bd. I, p. 115 f.

24 Hirt, Handb. der griech. Laut- u. Formenl., pp. 115, 119. Ga,, is the form

25 Deiss., B. S., p. 182, gives euvgari,aj in a pap. (iv/A.D.).

26 Gr. of N. T. Gk., p. 20. Cf. Note in W.-Sch., p. 50; Thack., pp. 82, 135; Mays., p. 14.

27 According to Phrynichus (Rutherford, New Phryn., p. 204) both of these words are evsca,twj ba.rbara.

28 Gr. of N. T. Gk., p. 20.

29 Moulton, Prol., p. 46.

30 Ib. For assimilation between a and E in modern Gk. dialects see Dieterich, Unters. etc., pp. 272, 274. In mod. Gk. vernacular a frequently displaces initial e or o. Cf. Thumb, Handb., p. 14.

31 Dieterich, Unters. zur Gesch. der griech. Spr., p. 4; also Schweizer, Gr. d. perg. Inschr., p. 163.

32 Nachm., Laute and Formen d. magn. Inschr., p. 146.

33 Moulton, Prol., p. 46. For further evidence see Cronert, Mem. Graeca Hercul., 1903, p. 199. In the Apostolic Fathers and the N. T. Apoc. te,ssera and tessera,konta are common as well as evkaqeri,sqh (Reinhold, De Graecitate Patr. Apostol. etc., p. 38 f. On the whole subject of a and e in the papyri see careful discussion of Mayser, Gr., pp. 54-60, where he mentions evkou,w evggareu,w evpeleu,sasqai (for similar confusion of aorist and fut. inf. see evkfeu,xasqai, 2 Macc. 9:22 V). Te,ssera and tessera,konta are very common also in the LXX MSS. Cf. Helbing, Gr. d. LXX, p. 5; Thack., Gr., p. 62f. This spelling occurs as early as iv/B.C. in Pergamum (Schweizer, Gr. d. perg. Inschr., p. 163 f.). In Egypt it hardly appears before i/A.D. and is not common till ii/A.D. (Thack., Gr., p. 62). The uncials give the later spelling. See "Additional Notes."

34 Dieterich Unters. etc., p. 70. Cf. Thack., Gr., vol. I, p. 75 f. So Dalmati,a in 2 Tim. 4:10, though C has Delm. as Lat. has both. Blass, Gr. of N. T. Gk., p. 21. Both forms are in the pap., Deiss., B. S., p. 182.

35 Hellen. (Griech. Spr.), p. 76. See also Rademacher, N. T. Gr., pp. 34 ff.

36 Gr. d. perg. Inschr., p. 49. Cf. Mayser, Gr., p. 62, cra/sqai for crh/sqai. So A in 2 Macc. 6:21.

37 K.-B1., Tl. I, Bd. I, p. 117 f. Cf. Meisterh., Gr. etc., p. 117, where Attic inscr. are shown to have Neopoli,thj.

38 Hort (Notes on Orth., p. 152) compares me,sabon, and Blass (Gr., p. 21) mesastu,lion. Metoxu, $metaxu,% is in 1 Clem. and Barn. (Reinhold, De Grace., p. 40 . Cf. Mayser, Gr., p. 60 f., o[lloi for a[lloi. Illiterate scribes confused a and o, a and e in the LXX (as metoxi,) and in the pap. (Thack., Gr., p. 77).

39 Gr. of N. T. Gk., p. 21.

40 Hort, Notes on Orth., p. 151. W.-Sch., p. 51, compare katafaga/j and katwfaga/j as as parallel. Cf. Meisterh., Gr., p. 17.

41 Moulton, Cl. Rev., 1901, p. 31, 1904, p. 107.

42 Gr. etc., p. 91 f.

43 Gr. etc., p. 61. Cf. also Dieterich, linters. etc., p. 78.

44 Prol., p. 47.

45 Moulton, Exp., 1904, p. 363. So also in the Rom. period occasionally evmatou/ e`atou/. Cf. Blass, Gr. of N. T. Gk., p. 35; Wack., Kuhn's Zeitschr., xxxiii, pp. 2

46 Moulton, Cl. Rev., 1901, p. 33; 1904, p. 107. He quotes Laurent (B.C.H., 1903, p. 356) as saying that this phenomenon was very common in the latter half of i/B.C.

47 W.-Sch., p. 47.

48 Notes on Orth., p. 150. Cf. on at and E, Mayser, Gr., p. 107.

49 Gr. of N. T. Gk., p. 9.

50 W.-Sch., p. 47.

51 vEp v avna,gkaij "Alexandrian only" according to Hort, Notes on Orth., p. 151.

52 Ib.

53 Ib. Cf. the Western kainofwni,aj for kenofwni,aj in 1 Tim. 6:20. In 1 Th. 3:3 instead of sai,nesqai FG read sie,nesqai. Nestle (Neut.-Zeit., 1906, p. 361) finds parallels in the forms sianome,nwn and sianqei,j.

54 Notes on Orth., p. 151.

55 Cl. Rev., 1904, p. 107. The pap. give faino,lion.

56 Brug., Griech. Gr., p. 28, as qeio,j = qeo,j; Thumb, Handb., p. 220.

57 Meisterh., Gr., p. 20 f. Cf. Schweizer, Gr. etc., p. 44 f. The change in e and ei was very common in vi/iii B.C. Jann., Hist. Gk. Gr., p. 37.

58 But even the Arcadian dial. has ple,ona pleo,nwn (Solmsen, Inscr. Grace., p. 4). Ple,on is common in the N. T. Apoc. (Reinhold, De Grace. Patr. Apost. etc., p. 40). Cf. Meisterh., Gr. d. att. Inschr., p. 40 f. On the whole subject of e and ei in the pap. see Mayser, Gr., pp. 67-73. They arc very numerous indeed, these changes in the pap., both ways.

59 Blass, Gr. of N. T. Gk., p. 22.

60 Solmsen, Inscr. Graecae etc., p. 1. Arcadian dial. Cf. also Meisterh., Gr., p. 3. In the Pontic dial. to-day there is a wide-spread use of e instead of h as in se,pomai (Thumb, Hellen. [Griech. Spr., referred to hereafter usually as Hellen.] p. 149).

61 Moulton, Cl. Rev., 1904, p. 108. Cf. also Moulton, Prol., p. 46, and Schweizer, Gr. d. perg. Inschr., pp. 47 ff., has good discussion of this shortening of h to e and also w to o. "E and h interchange times without number from v/B.C. down to ix/A.D." (Jann., Hist. Gk. Gr., p. 36). Reinhold (De Graec. Patr. etc., p. 101 f.) shows howl the confusion between h and e led to forms like eva.n avga,gete. Cf. the mod. Gk. ste,kw ( sth,kw) and qe,tw ( qh,tw).

62 Unters. etc., p. 136.

63 Schweizer, Perg. Inschr., P. 43 f.

64 Moulton, Cl. Rev., 1901, pp. 33, 434; 1904, p. 107. Cf. Mayser, Gr., p. 80 f.

65 `Aliei/j occurs in pap. also. Blass, Gr. of N. T. Gk., p. 307; Thackeray, p. 84.

66 Gr. of N. T. Gk., p. 22.

67 Notes on Orth., p. 151.

68 Blass, Gr. of N. T. Gk., p. 21. But always Ti,toj. Cf. Nachm., Magn. Inschr., p. 22, in discussion of e for Lat. i. Both legiw,n and le,ntion are read in Magn. inscr. (Thieme, Die Inschr. von Magn. etc., p. 8). Cf. also Schweizer, Gr. d. perg. Inschr., p. 46. For assimilation between e and i in mod. Gk. see Dieterich, Unters. etc., p. 272 f.

69 Gr. d. att. Inschr., p. 22. Cf. also K.-B1., Tl. I, Bd. I, p. 118.

70 Unters. etc., p. 135 f. Cf. Hirt, Handb. d. Griech. etc., p. 115.

71 K.-B1., Tl. I, Bd. I, p. 118, and Hirt, op. cit., p. 115.

72 Gr. of N. T. Gk., p. 21. Cf. Mayser (Gr., pp. 94-97) for a discussion of the pap. situation.

73 Notes on Orth., p. 152.

74 Gr. of N. T. Gk., p. 21. He quotes Buresch, Rhein. Mus., p. 216 f., as in favour of e in the N. T. as well as the LXX. vOleq. appears in the Apost. Fathers (Goodspeed, Index) and ovloq. in N. T. Apoc. (Reinhold, p. 40). For assimilation between e and o in mod. Gk. see Dieterich, Unters. etc., p. 274.

75 Gr. of N. T. Gk., p. 60. Omitted by Debrunner in ed. 4.

76 Notes on Orth., p. 173. Hort has a curious error here, for the references under a;n and eva,n should be exactly reversed. ;An = eva,n ('if') is rarely found in the pap. also. Moulton (Cl. Rev., 1901, p. 434) gives a'n mh. avpodw/i. (AP 43, ii/B.C.). Cf. also Cl. Rev., 1901, p. 32; Mayser, Gr., p. 152 f. Mayser gives exx. of eva.na;n and of a;neva,n.

77 Prol., p. 43; Cl. Rev., 1901, p. 32, etc.

78 Nachm., Magn. Inschr., p. 68. See Gregory, Prol. (Nov. Test. Gr.), p. 96, for the facts about the N. T. MSS. and eva,n.

79 Cronert, Mem. Graeca Here., p. 130.

80 Dieterich, Unters. etc., p. 326.

81 Thumb, Hellen., p. 92.

82 Hirt, Handb. d. Griech. etc., p. 63.

83 Thumb, Hellen., p. 98 f.

84 Brug., Griech. Gr., p. 29. Cf. also Thumb, Hellen., p. 138. In Boeotia also h and i interchange in ii/B.C. Cf. W.-Sch., p. 46. Mayser (Gr., p. 82) cites from a Hom. pap. of i/B.C. e;qike for e;qhke, and per contra (p. 84) avfh,keto

85 Schweizer, Gr. d. perg. Inschr., p. 47. He gives evph, for evpi, from a Byz. inscr.

86 De Graec. Patr. etc., p. 41. Cf. also Meisterh., Gr. d. att. Inschr., p. 34 f.

87 Blass, Ausspr. d. Griech., pp. 37, 94.

88 Hort, Notes on Orth., p. 151.

89 Ib., refers to sirikopoio,j in Neap. inscr. (C. I. G. 5834). In the mod. Gk. h= i in pronunciation. Cf. Thumb, Handb. d. neugr. Volkerspr., p. 2. W.-Sch. (p. 46) mention qh,bhn qi,bhn qei,bhn, in Ex. 2:3-6.

90 Cf. Blass, K.-B1., Tl. I, Bd. I, p. 135.

91 Perg. Inschr., p. 47. Cf. also p. 56. See numerous exx. of this change in Meisterh., Gr. d. att. Inschr., p. 47 f.

92 Cf. Bekker, Anec., I, pp. 9, 22. It is found also in 2 Macc. 8:24. Hort (Notes on Orth., p. 15) shows that a;peiroj (not a;phroj) is read in Herod. i. 32.

93 Prol., p. 46; Cl. Rev., 1901, p. 33. See also Thackeray, p. 83.

94 B. S., pp. 205-8. Cf. Dittenb., Syll., No. 388, p. 570. See also Mayser, Gr., pp. 74-79, for careful discussion.

95 Meisterh., Gr. d. att. Inschr., pp. 36 ff. Cf. Jann., Hist. Gk. Gr., pp. 39 and 49. See also Mayser, Gr., pp. 79 f., 126-131.

96 Schweizer, Gr. d. perg. Inschr., p. 60 f.

97 Nachm., Magn. Inschr., p. 50 f.

98 Mem. Graeca Hercul., p. 37.

99 Schweizer, op. cit., p. 60.

100 Nachm., op. cit., p. 51.

101 Gr. of N. T. Gk., p. 8. bou,lei oi;ei o;yei in Ap. Fathers (Goodspeed, Index).

102 Einl. in d. neugr. Gr., p. 306. He gives exx. from the N. T. Apoc.

103 Gr. of N. T. Gk., p. 8.

104 W.-Sch., p. 47. Moulton (Prol., p. 168) would take indifferently u`pa,gei or u`pa,gh| in Rev. 14:4. For many similar exx. in the inscr. see Dittenb., o[pwj a'n u`pa,rcei (117. 17), ei`re,qhsan (352. 66), etc.

105 Blass, Pronun., etc., p. 50.

106 Hirt, Handb. d. Griech., p. 114.

107 Schweizer, Gr. d. perg. Inschr., p. 65.

108 Meisterh., Gr. d. att. Inschr., p. 64. In the iv/n.e. the Attic often wrote ei for hi, but not for h|. In the Thess., AEol and Ionic inscriptions the i with a h w is freely omitted or wrongly inserted (irrational i), as in th/ po,lei ta. o;rh| as early as vi/B.C. Cf. K.-B1., Tl. I, Bd. I, p. 183 f. Strabo (14. 41) says that many regularly dropped the i in spurious diphthongs. pol loi. ga.r cwri.j tou/ i gra,fousi ta.j dotika,j kai. evkba,llousi de. to. e;qoj fusikh.n aivti,an ouv`k e;con. Cf. Jann., Hist. Gk. Gr., p. 29 f. Schweizer (Perg. Inschr., p. 47) cites th.in eu;noian.

109 Introd. to N. T. Gk., p. 314.

110 Mayser, Gr., p. 121, finds no i with a/n in the pap.

111 Prol., pp. 49, 168, 187.

112 Gregory, Prol. (New Test. Gr.), p. 109.

113 Cronert, Mem. Graec. Hercul., pp. 41 ff.

114 Gr. of N. T. Gk., p. 7. The LXX phenomena are similar. Cf. Helbing, Griech. d. LXX, pp. 3

115 Hatz., Einl. in neugr. Gr., p. 304.

116 Jann., Hist. Gk. Gr., p. 48.

117 Hellen., p. 171.

118 Hort, Intr. to Gk. N. T., p. 310. On the subject of h and u see Mayser, Gr., p. 85 f. He denies (p. 86) that the itacising pronunciation of h prevailed in the Ptolemaic period.

119 Jann., Hist. Gk. Or., p. 47.

120 Ib.

121 Ib.

122 Ib., p. 41.

123> K.-B1., p. 131. Mayser (Gr., pp. 87-94) has a full discussion of the problem in the pap. of the first three centuries B.C. and finds that in Egypt the pronunciation of ei closely approached that of i.

124 Meisterh., Gr. d. att. Inschr., p. 49. In the succeeding pages he gives numerous exx. in chron. order of the various interchanges between i and ei, many of them identical with the N. T. exx.

125 Unters. etc., p. 45.

126 Hellen., p. 172. The next most common interchange of vowels in the N. T. MSS. are ai and e, h and i or ei oi and u (Warfield, Text. Crit. of the N. T., p. 103).

127 Schweizer, Perg. Inschr., p. 53 f.

128 Nachm., Magn. Inschr., p. 35 f. Cf. Egyp. pap. also.

129 Cronert, Mem. Grace. Hercul., pp. 27 ff.

130 Prol., p. 47. For the LXX see Helbing, Gr. d. LXX, pp. 7 Thack. (Gr., p. 86 f.) thinks that the orthography in this point is older than that of a and A.

131 Warfield, Text. Crit. of the N. T., p. 103.

132 Hort, Notes on Orth., p. 152.

133 Prol., pp. 83-90.

134 According to Hort, Notes on Orth., p. 153.

135 B. S., pp. 142 f., 219 f.

136 Blass, Gr. of N. T. Gk., p. 8.

137 Notes on Orth., p. 155.

138 Ib., p. 154.

139 Hort, Notes on Orth., p. 155.

140 Gr. of N. T. Gk., p. 22. But it is quite possible (see j) that this is a case of prothetic o.

141 W.-Sch., p. 52.

142 Gr. d. att. Inschr., p. 81.

143 Hist. Gk. Gr., p. 53. Cf. on the other side K.-BI., I, 3, p. 53.

144 Jann., ib., p. 52. Cf. Mayser, Gr., p. 112.

145 Jann., Hist. Gk. Gr., pp. 27, 55, etc.

146 Ib., p. 84.

147 Gr. d. att. Inschr., p. 28 f.

148 Hellen., pp. 139, 193 ff. Cf. Kretschmer, Einl. in d. Gesch. d. griech. Spr., p. 225 f. Cronert (Mem. Grace. Hercul., p. 21 f.) gives exx. in Hercul. pap. Cf. Mayser, Gr., pp. 100-103, for exx. like bu,bloj bubli,on, etc., in the pap.

149 Cf. Blass, Gr. of N. T. Gk., p. 22. In Athens before 403 B.C. o stood for o w ou (Jann., Hist. Gk. Gr., p. 24).

150 Lobeck, p. 235; The New. Phryn., p. 310. Cf. K.-BI., I, p. 140 f., for this change in Old Attic and New Ionic. The N. T. Apoc. (Reinhold, De Graec. etc., p. 41) has exx. like evbolo.mhn as the mod. Gk. vernac. (Thumb, Neugr. Volksspr., p. 6). Cf. Buresch, Phil. li, 89. Most common bet. vi/iii B.C. acc. to Jann., Hist. Gk. Gr., p. 37.

151 Schweizer, Perg. Inschr., p. 66 f.

152 Gregory, Prol., p. 82.

153 K.-B1., I, p. 141.

154 Meisterh., Gr. d. att. Inschr., p. 24 f., gives numerous exx. of the exchange in inscr. of various dates.

155 Jann., Hist. Gk. Gr., p. 37. Jann. quotes a Louvre pap. (165 B.C.) which has to/ auvto/ tro,pwi. Mayser (Gr., pp. 97 ff.) finds only two exx. of this confusion of o and w in the Ptol. pap. of iii/B.C., but seventy in the next two.

156 Ib. Cf. Cronert, Mem. Graec. Hercul., p. 19 f.

157 Nachm., Magn. Inschr., p. 64.

158 Schweizer, Perg. Inschr., p. 95. Cf. Thumb, Hellen., pp. 143, 172.

159 Reinhold, De Graec. Patr., p. 41, and Moulton, Cl. Rev., 1904, p. 108.

160> Hort, Intr. to Gk. N. T., p. 309.

161 Cf. Reinhold, De Graec. Patr., p. 102; Hatz., Einl. etc., p. 306.

162 W.-Sch., p. 48.

163 Hort thinks so "perhaps." The Doric had stoia,. Blass (Gr. N. T. Gk., p. 22) prefers the correct Stwi?ko,j, Von Soden Stoi?ko,j

164 Acc. to W.-Sch. (p. 48 f.) this is not orthographical at all, but etymological. Why not both?

165 Ib., p. 48.

166 Jann., Hist. Gk. Gr., p. 37. Doubtless other vowel-exchanges in Rev. may have a similar explanation and so do not violate concord of gender.

167 Notes on Sel. Read., p. 136.

168 Disc. at Ephesus, App., p. 24.

169 Thumb, Hellen., p. 31. Cf. Brug., Griech. Gr., 4th ed., p. 32 f.

170 Hatz., Einl. etc., p. 103.

171 Thumb, Hellen., p. 85.

172 Cf. Nachm., Magn. Inschr., p. 62. Schweizer, Perg. Inschr., p. 71 f.

173 Gr. of N. T. Gk., p. 22. Cf. Mayser, Gr., p. 118.

174 Cf. Nachm., Magn. Inschr., p. 46 f.; Blass, Gr. of N. T. Gk., p. 9 f., observes that B occasionally divides thus u`io,j at end of a line and so practically A and D.

175 K.-B1., p. 135. Common in mod. Gk. (Thumb, Handb., p. 8).

176 Schweizer, Perg. Inschr., p. 70 f.

177 Jahrb. f. klass. Philol., 1891, p. 434.

178 Tattam's Egyp. Gr., p. 5.

179 P. 52. Reinhold (De Graec. Patr. Apost., p. 41) gives similar exx. Sunku rw/nta a appears in Egyp. pap. (B. M., vol. II, cliv). Cf. Mayser, Gr., p. 99 f.

180 Gr. of N. T. Gk., p. 10.

181 Schweizer, Perg. Inschr., p. 100.

182 Thumb, Hellen., p. 237. Cf. also ib., p. 63. For the mod. Gk. contraction see p. 249. Cf. K.-B1., Bd. I, pp. 201-218.

183 Schweizer, Perg. Inschr., pp. 100 ff.; Nachm., Magn. Inschr., pp. 68 ff.

184 W.-Th., p. 46; W.-M., p. 51.

185 Gr. of N. T. Gk., p. 22 f.

186 Gr. of N. T. Gk., p. 23.

187 Rutherford, New Phryn., p. 287. For other syncopated forms in the LXX see Thack., Gr., p. 99.

188 Meisterh., Gr. etc., p. 23.

189 Hort., Notes on Orth., p. 145.

190 Gr. of N. T. Gk., p. 23. Omitted by Debrunner.

191 Schweizer, Perg. Inschr., p. 101. Cf. Dittenb., Or. Graec. Inscr. Sel., evpeikw/j (565. 19), tamei/on, (515. 26 ff.), u`gei,aj (618. 2). For the same phenomena in the LXX see Helbing, Gr. d. LXX, p. 10 f.

192 See Deiss., B. S., p. 183, for pap. illustrations of pei/n pi/n tamei/on. Moulton, Prol., p. 45, calls this coalescence of two successive i sounds "a universal law of Hellenistic phonology." Cf, for the LXX Thack., Gr., pp. 22, 63 f., 98.

193 Hatz., Einl. etc., p. 304. Cf. K.-BI., Bd. I, pp. 243 ff.

194 Hist. Gk. Gr., p. 29.

195 Ib., p. 43. Cf. Mayser, Gr., p. 153 f.

196> Blass, Gr. of N. T. Gk., p. 17. So vIessai,.

197 Ib. Cf. W.-Sch., p. 34.

198> Gregory, Prol. etc., p. 108.

199 Blass, Gr. of N. T. Gk., p. 10.

200 Comp. Philol., pp. 158 ff.

201 Meisterh., Gr., p. ,178.

202 Smyth, Ionic Dial., p. 482. Cf. Nachm., Magn. Inschr., p. 155.

203 Mem. Graec. Hercul., p. 133 f.

204 Cf. Thumb, Handb., p. 13.

205 Gr. of N. T. Gk., p. 18. Cf. on hiatus K.-B1., I, pp. 190 ff.

206 Ib., p. 296 f. On indifference of later Gk. to hiatus see Bischoff, Neut. Wiss., 1906, p. 268; Thieme, ib., p. 265. Moulton (Prol., p. 92) quotes Kaelker (Qumst., p. 245 f.) as saying that Polyb. uses o[stij for o[j merely to avoid hiatus. Cf. Mayser, Gr., p. 160.

207 Meisterh., Att. Inschr., p. 69 f.

208 Schweizer, Perg. Inschr., p. 134; Nachm., Magn. Inschr., p. 71 f.

209 Cronert, Mem. Graec. Hercul., p. 138 f. Cf. also Thumb, Hellen. etc., p. 82.

210 Hort, Notes on Orth., p. 146.

211 Gr. of N. T. Gk., p. 18. Cf. also Gregory, p. 93 f.

212 Moulton (Cl. Rev., Feb. 31, 1901) finds that the pap. like the Lat. have a vowel not used in the metre. The inscr. concur in this practice. Moulton, Prol., p. 45. Cf. also Mayser, Gr., pp. 155-158, 160-162. He shows that in the pap. it is largely a matter of indifference. On the scarcity of elision in the LXX see Helbing, Gr. d. LXX, p. 12 f.; Thackeray, pp. 22, 136 f.

213 Blass (Gr. of N. T. Gk., p. 306) refers to the Oxyrhynchus pap., which have tou/t eivpw,n in Jo. 20:22

214 Prol., p. 93 f.

215 Notes, p. 146.

216 Blass, Gr. of N. T. Gk., p. 18.

217 Gr. of N. T. Gk., p. 18. See Additional Notes.

218 For more minute details about the prep. see Gregory, Prol., pp. 94 ff.

219 Meisterh., Att. Inschr., pp. 70 ff.

220 Magn. Inschr., p. 74.

221 Schweizer, Perg. Inschr., p. 133. Cf. Mayser, Gr., pp. 158 ff., for the common pap. exx. like kavgw, tavlhqe,j, etc.

222 Hort, Notes on Orth., p. 145.

223 See Gregory, Prol., p. 96; Von Soden, I, p. 1380.

224 See Blass, Gr. of N. T. Gk., p. 18, and W.-Sch., p. 38; Von Soden, I, p. 1380. Blass gives kavpequ,mei from D (Lu. 15:16).

225 Notes on Orth., p. 145.

226 Gr. of N. T. Gk., p. 19. For scarcity in LXX see Helbing, Gr. d. LXX, p. 13 f.

227 Cf. Jann., Hist. Gk. Gr., p. 21.

228 Ib. Cf. Meisterh., Gr. etc., p. 3.

229 Ib., p. 24 f. On the whole subj. of changes in the pap. see Mayser, Gr., pp. 163-248. For general remarks about consonant-changes in LXX MSS. see Swete, 0. T. in Gk., p. 301.

230 Bd. I, pp. 85-101.

231 Ib., pp. 77-85, 101-103. The mod. Gk. pronounces auvto,j =aftos. The inscr. give the form avF utou/. Cf. Riem. and Goelzer, Phonet., p. 34.

232 Hellen., pp. 245 ff.

233 Prol., p. 44. But Sommer, Gr. Lautstudien, shows that the rough breathing is sometimes due to digamma.

234 Thumb, Hellen., p. 187 f.; cf. p. 134 f. for intervocal g.

235 Blass compares the insertion of consonants in Semitic names like ;Ead raj Mambrh/.

236 Moulton, Cl. Rev., 1901, p. 34.

237 Schweizer, Perg. Inschr., p. 179 f. Cf. W.-Sch., p. 64, for full references concerning the use of m with lamba,nw. Cf. Gregory (Prol., p. 72) for list and references of the various compounds of lamba,nw and lh/myij in the N. T., avna avnepi avnti avpo kata meta para pro pros) The LXX MSS. have lh,myomai (Q lh,yontai) and evlh,mfqhn. Cf. Helbing, Gr. d. Sept., p. 22.

238 Blass, Gr. of N. T. Gk., p. 24; W.-Sch., p. 64.

239 Ib., p. 65.

240 Magn. Inschr., p. 108. Cf. also Hoffmann, Griech. Dial., Bd. III, p. 173; Meisterh., p. 128; Mayser, Gr. d. griech. Pap., p. 165; Schmid, Atticismus, Bd. IV., p. 579 (for the Atticistic gign); Cronert, Mem. Graec. Hercul., p. 91 f.; Reinhold, De Graec. Patr. etc., pp. 46-48. In the LXX gi,nomai and ginw,skw are uniform. Cf. Helbing, Gr. d. Sept., p. 21. Thack. (Gr., p. 111 f.) finds illustrations of the omission of intervocalic 7 in the LXX uncials as in the pap. (Mayser, Gr., p. 167 f.).

241 P. 65, where a full discussion of the geographical points is given.

242 Gr. of N. T. Gk., p. 10.

243 P. 55; cf. also Riem. and Goelzer, Phonet., pp. 225 ff.

244 See Thumb, Hellen., pp. 20 ff.; Schweizer, Perg. Inschr., pp. 122 ff.; Nachm., Magn. Inschr., pp. 88 ff.; Cronert, Mem. Graec. Hercul., pp. 74 ff. Cf. Mayser, Gr., pp. 211-219. For the LXX see Helbing, Gr. d. Sept., pp. 14-16. The MSS. of the LXX are largely the same as those of the N. T. and show similar phenomena in orthography. So in Ex. 7:10 B has e;riyen vArr. Both avrrabw,n, and avrabw,n occur, and it is in the pap. that we can often find the true Ptolemaic spelling. A curiously has usually ge,nhma and B ge,nnhma.

245 Meisterh., Gr. d. att. Inschr., p. 93.

246 Gr. of N. T. Gk., pp. 10, 328. Similar variations in usage as to r or rr appear in the inscr. of the koinh, (Schweizer, Perg. Inschr., p. 124, avnantirh,twj etc.; Nachm., Magn. etc., p. 91) and even in the Attic inscr. (Meisterh., p. 95, avnarhqe,ntej, etc.). Cf. Reinhold, De Graec. etc., p.42, for exx. of evru,sato, etc.

247 Gr. of N. T. Gk., p. 10. vArabw,n "only Western," Hort, Notes on Orth., p. 148. But the pap. (Moulton, Cl. Rev., 1901, p. 33; Deiss., B. S., p. 183 f.) frequently have avrabw,n, and, as Deissmann remarks, people are not always particular to preserve mere etymology.

248 CIGII, 2722. 5. Cf. W.-Sch., p. 56,

249> The inscr. show puro,j also (Dittenb., 177. 15; 748. 20).

250 Cronert, Mem. Graec. Hercul., p. 76.

251 Gr. of N. T. Gk., p. 11.

252 Rutherford, New Phryn., p. 348.

253 Deiss., B. S., pp. 109 f., 184. Cf. Thackeray, p. 118.

254 Gregory, Prol., p. 79.

255 In Mk. B (5) has kra,batoj, but is not followed by W. H. in Jo. and Ac. (6). Thumb, Hellen., p. 22, argues for bb as the correct form from mod. Gk. usage. Blass (Gr. of N. T. Gk., p. 328) cites both kra,battoj and kraba,tion from Arrian's Diss. Epict. and kra,battoj from the pap. Cf. Moulton's note in Einl.

256 Cf. W.-Sch., p. 57.

257 Cf. Pliny (Nat. Hist., V, 15. 71 for Genh.) also. In W.-Sch., p. 57, the point is made that the unpointed Targums do not distinguish between rs;yneG. and rs;yneGi.

258 W.-Sch., p. 56, = Apy' or y'b;y'. Cf. on this subject Helbing, Gr. d. Sept., p. 26 f.

259 Blass, Gr. of N. T. Gk., p. 328, quoting E. Lippett.

260 Hort, Notes on Orth., p. 159.

261 Gr. of N. T. Gk., p. 11.

262 W.-Sch., p. 57; E. Bibl., p. 2504 f.

263 Blass, Gr. of N. T. Gk., p. 11.

264 On the whole subject see Hort, Notes on Orth., p. 159, and Blass, Gr. of N. T. Gk., p. 11. Cf. also Schweizer, Perg. etc., pp. 110 f., 114 f. Cf. for the pap., Mayser, Gr., pp. 190-224; Soden, I, pp. 1372 ff.

265 Cf. Meisterh., pp. 105-109. In North Engl. one hears "ith wood" for "in the wood." The MSS. of the LXX show the same phenomena as one sees in the N. T. MSS. and the pap., like evg gastri, evm me,sw| suggra,fein etc. Cf. Helbing, Gr. d. Sept., p. 16 f.; Thack., Gr., pp. 130 ff.

266 Cf. Blass, Gr. of N. T. Gk., p. 12; Ausspr, etc., p. 123. Alexandrian writers followed the Attic in this assimilation. Blass compares the guttural use of a in avh`li, (Mt. 27:46) in L and in the LXX vAermw,n vAendw,r.

267 Meisterh., p. 110 f. Cf. Jann., Hist. Gk. Gr., p. 97.

268 Magn. Inschr., p. 100 f. Cf. also Schweizer, Perg. etc., p. 127; Jann., Hist. Gk. Gr., p. 92.

269 Cronert, Mem. Graec. Hercul., p. 57; Blass, Gr. of N. T. Gk., p. 12.

270 Ib., pp. 11 f., 306.

271 Hort, Notes on Orth., p. 149.

272 Ib. In general see Wecklein, Curae Epigr. ad Gr. Graeca.e etc., 1869, p. 47 f.

273 Blass, Gr. of N. T. Gk., p. 12. Cf. Cronert, Mem. Graec. Hercul., p. 61.

274 Hort, Notes on Orth., p. 149. See for LXX Thackeray, pp. 132

275 Ib. For the inscr. see Nachm., Magn., p. 104 f. The Coptic shows similar variation. For the loss of final n in mod. Gk. vernac. see Thumb, Handb., p. 24 f.

276 About evn, in composition see Gregory, Prol. etc., p. 76 f.; Soden, I, p. 1383. vEn in MSS. appears in composition as evn--, evg and even evk as evkko,phn. On e;nprosqen in the pap. see Mayser, Gr., p. 45.

277 Prol. etc., p. 73 f. Cf. Jann., Hist. Gk. Gr., pp. 91-97, for the history of this subject during various stages of the language.

278 Cf. Giles, Man. of Comp. Philol., pp. 98, 124

279 Cf. W.-Sch., p. 66 note.

280 Cf. ib., p. 58 note, for further discussion.

281 Prol., p. 45. Cf. also Thumb, Theol. Literaturzeit., XXVIII, p. 422.

282 Meisterh., Att. Inschr., pp. 99 f.

283 Schweizer, Perg. Inschr., p. 125; Nachm., Magn. Inschr., p. 94. In the pap. a;rrhn "greatly preponderates over. a;rshn" (Moulton, Cl. Rev., 1901, p. 33). Cf. also Reinhold, De Graec. etc., p. 44 f. Thumb, Hellen., p. 77 f.

284 Giles, Man. of Comp. Philol., pp. 113, 115. On the whole subject of the exchange of consonants in the pap. see Mayser, Gr., pp. 169-188, 219-224. For the LXX exx. ( ouvde,n ouvqe,n* glw/ssa glw/tta* fula,ssw fula,ttw* evla,sswn evla,ttwn* a;rrhn qarrw/), etc.) see Helbing, Gr. d. Sept., pp. 17-20; Thack., Gr., pp. 100-124.

285 Cf. Rutherford, New Phyrn., p. 14.

286 Cf. a;zbestoj in N (Mk. 9:43), evgnwzme,noj, etc., in pap. (W.-Sch., p. 59).

287 Notes on Orth., p. 148.

288 Deiss., B. S., p. 185. Cf. Moulton, Prol., p. 45; Dittenb., 458. 41, iv Zmu,rnh|.

289> Cf. Thumb, Hellen., pp. 53, 78 ff.; Schweizer, Perg. Inschr., p. 125; Nachm., Magn. etc., p. 95 f.; Moulton, Cl. Rev., 1901, p. 32; Prol., p. 45; Blass, Gr. of N. T. Gk., p. 23; Hort, Notes on Orth., p. 148; Reinhold, De Grace. etc., p. 43 f. Giles (Man. of Comp. Philol., p. 115) thinks that the ss in Athens was a literary mannerism and pronounced just like tt.

290> Prol., p. 45. Cf. Thumb, Hellen., p. 90.

291 Cf. W.-Sch., p. 59..

292 Gr. of N. T. Gk., p. 24; W.-Sch., p. 61. Cf. Meisterh., p. 48, for this interaspiration in the old Attic inscr. Cf. Mayser, pp. 180

293 Moulton, Prol., p. 45. The Ptol. pap. have both spellings, Deiss., B. S., p. 185. Cf. Mayser, Gr., p. 173.

294 Att. Inschr., p. 114.

295 Schweizer, Perg. Inschr., p. 137, whose. table confirms that of Meisterh. Cf. also Thieme, Inschr. von Magn., p. 8; Nachm., Magn. Inschr., p. 110, with similar table. The pap. agree, Cronert, Mem. Graec. Hercul., p. 137, and Mayser, Gr. d. griech. Pap., pp. 236 ff. In the LXX n evfelk. occurs before consonants also. Cf. Helbing, Gr. d. Sept., pp. 22 ff.; Thack., Gr., pp. 134 ff. So as to movable j. Cf. me,cri u`mw/n and me,crij ou- in LXX.

296 Einl. etc., p. 111, like i`storh,qhn o` nao,j. Cf. Schweiz., Perg. Inschr., p. 137.

297 Prol., p. 49. Cf. also Reinhold, De Graec., p. 37.

298 W.--Sch., p. 62.

299 Gr. of N. T. Gk., p. 19.

300 Hort, Notes on Orth., p. 147 f.; Gregory, Prol., p. 97 f. In simple truth n movable was not so uniform in the earlier Gk. (esp. Thuc.) as the grammars imply. Cf. Maasson, De littera v Graec. parag., 1881, pp. 47, 61.

301 See Blass, Gr. of N. T. Gk., p. 19; Gregory, Prol., p. 97.

302 Blass, Gr. of N. T. Gk., p. 328, and references there given. Cf. Thack., Gr., p. 135.

303 Blass (Gr. of N. T. Gk., p. 19) quotes Attic usage for pe,rusin before vowels.

304 For the Hom. a;ntikru and further items see W.-Sch., p. 63 and note. ;Antikruj ( katantikru,) in Attic is 'downright,' not 'over against' (Blass, Gr. of N. T. Gk., p. 20). Cf. for the pap. Mayser, Gr., pp. 242 ff.

305 Hort, Notes on Orth., p. 148. But W. H. read a;crij ou- in Heb. 3:13, elsewhere a;cri ou-) For further discussions of a;cri and me,cri see W.-Sch., p. 63 note.

306 For illustrations from the koinh, inscr. see Nachm., Magn. Inschr., p. 112. Cf. Reinhold, p. 37 f.

307 Jann., Hist. Gk. Gr., p. 63. The marking of the rough breathing was general in the earlier forms in vii/A.D., ib., p. 65.

308 Cf. Bekker, Anec., II. 692, and Jann., Hist. Gk. Gr., p. 63.

309 Intr. to Gk. N. T., p. 310. Cf. also Sitterley, Praxis in MSS. of the Gk. Test., 1898, p. 32. See Helbing, Gr. d. Sept., p. 25 f., for remarks on breathings in the LXX MSS., where Aolic and Ionic psilosis occur in evp v o`dou/ kat v e[na as well as exx. of aspirated consonants like kaq v ovfqalmou,j kaq v evniauto,n evf v ei=den, not to mention ouvk e`wra,kasin and ouvc ivdou,. For further remarks on breathings in the LXX see Swete, 0. T. in Gk., p. 302.

310 Giles, Man. of Comp. Philol., pp. 81, 91. The stop for the opening of the glottis (lenis) easily becomes breathed (rough). Cf. also Thumb, Unters. uber d. Spir. Asper. im Griech., 1888, p. 63.

311 Cf. Thumb., p. 73 f. The Laconic Gk. used H in interaspiration as well as at the beginning (ib., p. 8). Dawes (Pronun. of the Gk. Aspirates, 1894, p. 103) is not able to reach a final decision as to whether the Gk. aspirates are genuine aspirates like the Sans. according to Brugmann, Curtius, etc.

312 Cf. Jann., Hist. Gk. Gr., p. 91. On the whole subject of the aspirated

313 Cecil Bendall, Jour. of Philol., 1904, pp. 199 ff.

314 R. Weiss, De Dig. etc., 1889, p. 47. Cf. also Paues, De Dig. Hesiodes Quest., 1887, p. 48.

315 Cf. Sommer, Griech. Lautstudien, 1905, p. 2. On metathesis in aspiration, as e[cw ( e;cw), see Meisterh., p. 102, exx. of e[cw in Attic inscr. v/B.C. See also article by Pernot in Rev. des Et. Grq., 1906, pp. 10-23, on La Metathese dans les Dial. de Chio.

316 Schweizer, Perg. Inschr. etc., pp. 116 ff. The Attic had only i;dioj, but (Meisterh., p. 87).

317 Nachm., Magn. Inschr., p. 83.

318 Cronert, Mem. Graec. Hercul., p. 152 f.

319 Thumb, Hellen. etc., p. 64.

320 Moulton, Prol., p. 44. Cf. also for the inscr., Dittenb., evf v e[toj (458. 71), kaq v i`di,an (233. 49), and for the pap., Moulton, Cl. Rev., 1901 (pp. 33, 434) and 1904 (p. 106). Cf. also Hort, Intr. to Gk. N. T., p. 312.

321 Ib., p. 311.

322 Blass, Gr. of N. T. Gk., p. 15.

323 Intr. to Gk. N. T., p. 313; App., p. 160.

324 W.-Sch., p. 40.

325 Gregory, Prol., p. 91; Thack., p. 125.

326 Blass, Gr. of N. T. Gk., p. 16. Cf. Thumb, Unters. d. Spir. Asper, p. 65.

327 Notes on Orth., p. 143.

328 Moulton, Prol., p. 44; Thumb, Spir. Asper, p. 71. Moulton (Cl. Rev., Mar., 1910, p. 53) now says: "I am quite willing to be convinced that the long-lost digamma was an accessory here if no better explanation turns up." Thumb (Spir. Asper, pp. 11, 71) admits the possibility of the digamma explanation in some cases.

329 Prol., p. 91.

330 Cf. Intr. to Gk. N. T., p. 313 f., where Hort really favours ouvc `Ioud) and the rough breathing for all the forms of vIou,daj vIoudai/oj, etc. For the variations in the LXX MSS. see Thack., p. 125.

331 Intr. to Gk. N. T., p. 312.

332 Rutherford, New Phryn., p. 363. For this transfer of aspiration cf. Curtius, Gk. Verb, II, 109. Nestle (Am. Jour. of Theol., July, 1909, p. 44S) urges that, since the Gk. of the Bible is an "east-west language," attention must be paid to oriental tongues. He notes that the Coptic has aspiration in helpis, hisos, for evlpi,j i;soj.

333 Notes on Orth., p. 168.

334 Hort, Notes on Orth., p. 144.

335 Gr. of N. T. Gk., p. 16.

336 W.-Sch., p. 39.

337 Gr. of N. T. Gk., p. 16. Hort, Intr. to N. T. Gk., p. 313. Cf. also Gregory, Prol., p. 106 f., for list of these words.

338 Strange as it may seem, "Hebrew" rather than "Ebrew" is modern (Hort, Intr. to Gk. N. T., p. 313).

339 Hort (Notes, etc., p. 144), however, merely follows custom and prints u[ss)

340 Intr. to N. T. Gk., p. 313.

341 Ib.

342 Gr. of N. T. Gk., p. 16. Cf. Helbing, Gr. d. Sept., p. 30 f.

343 W.-M., p. 53.

344> Cf. W.-Sch., p. 40 f.

345 On the whole matter see Hort, Notes on Orth., p. 144 f.; W.-M., p. 188 f.; Buttmann, Gr. of N. T. Gk., p. 111; Blass, Gr. of N. T. Gk., p. 35.

346 Nachm., Magn. Inschr., pp. 84, 144; Schweizer, Perg. Inschr., p. 161.

347 Gr. d. griech. Pap., p. 306.

348 Jann., Hist. Gk. Gr., p. 66. Cf. also pp. 507 ff. on the Origin and History of Accent.

349 Intr. to Gk. N. T., p. 314.

350 Gk. Accentuation (1881), p. xxiii.

351 Ib., p. xvii.

352 Cf. Meister, Bemerk. zur dorischen Accentuation (1883); Hadley, On the Nat. and Theory of the Gk. Accent. (Ess. Phil. and Crit., pp. 110 ff.); Wheeler, Die griech. Nominalaccente (1885); Bloomfield, Study of Gk. Accent (Am. Jour. of Philol., 1883); Wack., Beitr. zur Lehre vom griech. Akzent; Brugmann, Griech. Gr. (1900), pp. 150 ff.; K.-B1., I, pp. 317 ff.; for further lit. see Brugmann above. On accent changes in mod. Gk. see Hatz., Einl., pp. 418-440; Thumb, Handb., p. 28 f. For the accent in the LXX see Helbing, Gr. d. Sept., p. 24. Here the same MSS. present the same problems that we have in the N. T.

353 Jann., Hist. Gk. Gr., p. 62.

354 Riem. and Goelzer, Phonet., p. 77.

355 Krumb., Beitr. zu einer Gesch. der griech. Spr., Kuhn's Zeitschr. fur Sprachl., 1885, p. 521. Cf. also Hats., Einl. etc., p. 418; Chandler, Gk. Accentuation, p. v; Brugmann, Griech, Gr., p. 150.

356 Sophocles, Lex. of Rom. and Byz. Period, p. 48.

357 Giles, Man. of Comp. Philol., p. 91.

358 Blass, Gr. of N. T. Gk., p. 14.

359 Ib. Cf. Gregory, Prol., p. 114, for specimen from Euthalius.

360 Giles, Man. of Comp. Philol., p. 92.

361 Harris, MS. Notes on Gk. Gr. Cf. Riem. and Goelzer, Phonet., p. 77 f., for a discussion of the musical aspect of the matter.

362> Arnold and Conway, The Restored Pronun. of Gk. and Lat., 1895, p. 18.

363> Schweizer, Perg. Inschr., p. 129.

364 Man. of Comp. Philol., p. 94.

365 Jann., Hist. Gk. Gr., p. 65.

366 Bloomfield, Study of Gk. Accent, Am. Jour. of Philol., 1883, p. 22. Cf. Plato, Crat., 399 A-B. Hirt (Der Indoger. Akzent, 1895, p. 17) contends for the two-tone principle.

367 Jann., Hist. Gk. Gr., p. 66.

368 Ib., pp. 65, 68.

369> Hadley, Uber Wesen and Theorie der griech. Beton., 1872, pp. 409, 415.

370 Giles, Man. of Comp. Philol., p. 96. Giles thinks that words like evfero,meqa originally had the accent further back. Cf. Riem. and Goelzer, Phonet., p. 80, for Plato's word of 17 syllables and Aristophanes' word of 78.

371 Henry, Comp. Gr. of Gk. and Lat., Elliott's transl., 1890, p. 93 f. Cf. Meister, Bemerk. zur dorischen Accentuation, p. 1.

372 Cf. Wheeler, Griech. Nom. etc., p. 11, and Wack., Beitr., p. 19.

373 Prol., p. 99 f.

374 Gr. of N. T. Gk., p. 15. Cf. W.-Sch., p. 67, for further parallels. Also W.-M., p. 57.

375 Blass, Gr. of N. T. Gk., p. 15.

376 P. 68.

377 Gr. of N. T. Gk., p. 15. Blass urges that B has qlei/yij, but W. H. refuse to follow B in matters of orthography. But the Herculaneum rolls here reinforce B with ei before y) On the whole subject see Lipsius, Gr. Unters., pp. 31 ff.; Lobeck, Parall., pp. 400 ff.; Cobet, N. T. Vatic., pp. xlix ff.

378 Cf. W.-M., p. 58.

379 As shown in W.-M. (p. 60), the N. T. MSS. have e;sw, not ei;sw, though eivj not evj)

380 Cf. W.-S., p. 73.

381 Ib., p. 72.

382 Ib., p 69.

383 Cf. Moulton, Prol., p. 69. On accent of the vernac. see Apostolides, Glkwssikai. Mele,tai (1906).

384 W.-M., p. 62.

385 Cf. W.-Sch., p. 77.

386 Cf. Lipsius, Gr. Unters., p. 61. Cf. also W.-Sch., p. 78.

387 In W.-Sch., p. 74 f., see remarks on the subject.

388 Cf. W.-Sch., p. 73. This word is, of course, not to be confounded with a=sson (Ac. 27:13) as Text. Rec. did.

389 Gr. of N. T. Gk., p. 15.

390 Cf. W.-Sch., p. 75.

391 W.-M., p. 59.

392 Cf. also Gregory, Prol., p. 102 f.; W.-Sch., p. 75; Westcott, Notes on Orth., pp. 155, 159; Thackeray, pp. 150 ff.

393 Blass, Ausspr. des Griech., 1888, p. 7.

394 Nolan, The Gk. Gr. of Roger Bacon, p. xx.

395 Hist. Gk. Gr., p. 31 f. Cf. Mayser, Gr., pp. 138-151.

396 Blass, Pronun. of Anc. Gk., Purton's transl., p. 3.

397 Guide to Mod. Gk., p. X.

398 Hist. Gr. der hell. Spr. (pp. 26, 36). In pp. 35-40 he states the case against the squib of Erasmus. Cf. Engel (Die Ausspr. des Griech., 1887) who defends the mod. Gk. method, as already stated.

399 Lautsystem der griech. Vulgarspr., 1879, p. 83 f.

400 Achilles Rose, Chris. Greece and Living Gk., 1898, p. 61.

401 Cf. Mure, A Crit. Hist. of the Lang. and Lit. of Anc. Greece, I, p. 99; Bolland, Die althell. Wortbet. im Lichte der Gesch., 1897, p. 6. Cf. Pronun. of Gk. as deduced from Graeco-Latin Biling. Coins. By Cecil Bendall in Jour. of Philol., vol. XXIX, No. 58, 1904. Here the rough breathing is represented by h, q=th, f = ph.

402 Thumb, Unters. etc., 1888, p. 1. Cf. Sophocles, Hist. of Gk. Alph. and Pronun., 1854.

403 Telly, Chron. and Topog. der griech. Ausspr. nach d. Zeugnisse der Inschr., 1893, p. 39.

404 Rutherford, The New Phryn., p. 32.

405 Philol. of the Gosp., p. 9,

406 Man. of Comp. Philol., p. 103. Cf. also Ellis, Early Eng. Pronun.

407 "Gk. Pronun." in Ess. Philol. and Crit., pp. 128-140. Hatzidakis, Einl. etc.

408 Cf. Moulton, Prol., p. 34 f.

409 Blass, Gr. of N. T. Gk., p. 6 f.

410 Nicklin, Cl. Rev., Mar., 1906, p. 116. This is precisely the objection that Jannaris (Hist. Gk. Gr., p. 33) brings against the ancient grammarians as "post-Christian scribes" and unable to "speak with authority of the pronunciation of classical Greek."

411 Giles, Man. of Comp. Philol., p. 75. Cf. Oppenheim and Lucas, Byz. Zeitschr., 1905, p. 13, for exx. of phonetic spelling.

412 Cf. Riem. and Goelzer, Phonet., p. 41.

413 Giles, Man. of Comp. Philol., p. 540.

414 Riem. and Goelzer, Phonet., pp. 41, 46. Thumb (Hellen., p. 228) warns us against overemphasis of the Boeotian influence.

415 Hist. Gk. Gr., p. 31. "The pronunciation of ancient Gk. in the manner of the present Greeks had been traditionally accepted at all times, before and through the Middle Ages, as a matter of unquestioned fact."

416 Phonet., p. 56. "En resume, la prononciation grecque ancienne etait, sur presque tous les points, differente de la prononciation moderne."

417 On the paragraph see Thompson, Handb. of Gk. and Lat. Palaeog., pp. 67 ff. Occasionally the double point (:) was used to close a paragraph.

418 Cf. Warfield, Text. Crit. of N. T., pp. 40 ff.

419 Hort. Intr. to Gk. N. T., p. 319. For the sti,coj see further Gregory, Prol., p. 112 f.

420 Intr. to Gk. N. T., p. 319 f.

421 Jannaris, Hist. Gk. Gr., p. 62.

422 Thompson, Handb., etc., p. 69.

423 Ib., p. 70; Jannaris, Hist. Gk. Gr., p. 67.

424 I follow Thompson (Handb., etc., p. 70) on this point instead of Jannaris (pp. 63 and 67), who makes the u`postigmh, = our comma.

425 Cf. Gregory, Prol., pp. 345, 348; Blass, Gr. of N. T. Gk., p. 17. D has the sti,coi, in the way of sense-lines (Blass, ib.).

426 Jannaris, Hist. Gk. Gr., p. 67.

427 Thompson, Handb., etc., p. 81. So Suidas. The colon is the main semidivision of the sentence, but mod. Eng. makes less use of all marks save the period and comma.

428 W.-M., pp. 63, 67.

429 Thompson, Handb., etc., p. 67.

430 W.-Sch., p. 35.

431 See this ch. ii (k) for discussion of elision. For origin and early use of the apostrophe see Thompson, Handb., etc., p. 73.

432 See this ch. II (1) for discussion of crasis. Cf. Jannaris, Hist. Gk. Gr., p. 88.

433 Thompson, Handb., etc., p. 72.

434 Blass, Gr. of N. T. Gk., p. 14. For the usage of Tisch. in the union and the separation of particles see Gregory, Prol., pp. 109-111. In most cases Tisch. ran the particles together as one word.

435 P. 35.

436 Gr. of N. T. Gk., p. 17. Left out by Debrunner.

437 W.-M., p. 63.

438 Hort, Intr. to Gk. N. T., p. 318.