STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world byJSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.istor.org/participate-istor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. DEMOSTHENES' AVOIDANCE OF BREVES By Chables D. Adams In the second edition of the Attische Beredsamkeit (1893) Friedrich Blass gave his final statement of the "law" of the avoidance of a succession of three or more short syllables in the speeches of Demos- thenes. He pointed out certain classes of apparent exceptions to the law, which he would remove by punctuation, change of spelling, elision, and similar means; he formulated other classes of exceptions which he held to be real, but permissible, and in some cases unavoid- able. After eliminating these classes of exceptions, Blass showed how few cases of breves remained even in so long a speech as the De corona. But here, as in the formation of his Teubner text and that of his edition of the De corona, he assumed the right to use the rhythmical law as a determining factor in many cases of textual criticism. If the reading of S gave breves^ in any given case, while the vulgate reading or a quotation in an ancient author offered a reading without the breves, or if the bracketing of an unnecessary word or phrase would remove the series of short syllables, Blass did not hesitate to accept the inferior manuscript authority, or to emend the text. This freedom in textual emendation in favor of his "law," and the vagueness of the formulation of some of his permissible exceptions, have tended to react against confidence in the law itself.* It has seemed to the writer that it would be well to attempt to establish the range of this law more definitely. There is no question that the strictness of the application of the law, or principle, or rule, or tendency varies in different speeches of Demosthenes, and in different parts of the same speech. A detailed examination should begin with a speech of the highest rhetorical perfection, carefully revised for pubKcation by Demosthenes, and of sufficient length to give variety of style. All this is found in the De corona. If we can • In the following, the word breves and its abbreviation, brr., are used for a series of three short syllables in the restricted sense defined by Blass. 'In a review of Fuhr's Teubner text (Classical Philology, X, 473), I have dis- cussed at length the application of this law to textual criticism. [Clabsicai, Philologtt XII, Jtdy, 1917] 271 272 Charles D. Adams determine the range of the avoidance of breves in this speech, we may fairly use the result as a norm to be applied to other speeches, and perhaps to some extent as a test of genuineness in the case of speeches of disputed authorship. As a text for such a study of the De corona we must not take one which, like Blass's Teubner text (1884), has been revised under the dominant influence of the very theory that is under investigation.* Butcher's Oxford text (1903) furnishes pre- cisely what we need: a text which follows the authority of S steadily unless the evidence, aside from all rhythmical considerations, is con- clusive against it. While Butcher usually introduced such regular elision and punctuation as would eHminate breves which were only apparent, he never pressed even this practice into the service of a theory, and he never recognized rhythmical considerations as justi- fying the acceptance of inferior manuscript readings. His text is therefore a safe one for our investigation. It is to be remembered, however, that even the best manuscript testimony as to elision and the use of p movable is of no value, and is properly ignored by Butcher. S itself is utterly inconsistent in such matters. Blass's law of the avoidance of a succession of short syllables as formulated in the Attische Beredsamkeit (3. Abt., 1. Absch., 2. Aufl., 1893, "Demosthenes," pp. 105-12) is, " Dass die Anhauf ung von mehr als zwei kurzen Silben moglichst vermieden wird, wobei naturUch solche Silben, die durch EKsion in Wegfall kommen, nicht mehr zahlen." Blass defines the following classes of "apparent excep- tions": 1. Breves caused by writing in one form a word which should be written in another form. Here belong words which should lose a short syllable by eUsion, or lengthen a final short syllable by the use of V movable (before a word beginning with a consonant). Other words should be differently spelled, as elpeKa for 'iveKa, avrov for iavTov, ■fi^ovXdn'qv for e^ov\6)xr}p. Sometimes a slightly different form of the word is to be assumed, as Trorep' for Trdrepop, ttSs for airas. 2. Breves where the third short syllable comes before a pause (where we may always assume lengthening in delivery) or where a pause breaks the series. ' It is also true that Blass's Teubner text suffers from his attempts to secure rhythmical correspondences of kola according to a theory which he himself soon greatly modified. Demosthenes' Avoidance of Breves 273 3. Breves that are removed by crasis or aphaeresis. 4. Breves that are contained in a quotation, a citation, the address to the reading clerk, i.e., any matter that falls outside the speech proper. The apparent exceptions under 1, 2, and 3 disappear in proper delivery. Much more important are Blass's "real exceptions": 1. A series of short syllables lying wholly within a word and not including its final syllable must be permitted if the word is to be used at all; many such words must be used. (The series should not include the final syllable, for a short final syllable may be lengthened by proper arrangement of words.) Blass further says (p. 108): "Ob dann zu dem hiermit gegebenen Tribrachys noch von den benachbarten Worten Kurzen hinzukommen, ist gleichgtiltig, falls nicht etwa damit noch eine zweite Hebung aufgelost erscheint." This statement is not justified; the breves within the word cannot be avoided; the addition of other short syllables preceding it can be, for the most part. 2. Breves are permissible in a preposition and its noun, the two being almost like one word; the same is true of article and noun, and of article followed by preposition and noun. 3. Breves of which Sri is a part are permissible. 4. Breves are permissible in certain compact phrases, for the same reason as within a single word; so oStos 6 vd/Ms, tp'ltov iros. 5. Breves are permissible when two words are amalgamated by elision, or brought together in hiatus with shortening of the final syllable of the first word. Blass gives no grounds for this exception, and none can be given. Such grouping of short syllables is not necessary; when it does occur it is a real violation of the law. In the following pages I give the instances of groups of three or more short syllables in the De corona, as a means of determining the extent of the permissible exceptions to the law of breves, and the extent and character of the violations of the law, after these excep- tions have been eliminated. I have not recorded the breves occurring within a word unless they include its final syllable nor breves occurring at a pause that is marked by punctuation in Butcher's text, as neither group is significant for our purpose. We shall find that even a text prepared by an editor who recognized Demosthenes' tendency to 274 Charles D. Adams avoid breves, and who often used elision and punctuation with this in view, still contains a considerable number of passages with apparent breves that are to be removed by elision or punctuation. The following abbreviations are used : p. A pause to be assumed, whether marked by punctuation or not. sp. Use different spelling or a variant form. V. Use V movable. e. Elide. aph. Use aphaeresis. or. Use crasis. ext. The breves are in matter that is external to the speech proper, as in quoted words, directions to the clerk, etc. Gd., Goodwin's edition; BL, Blass, Teubner text; H., Humphreys' text; F., Fuhr's Teubner text (1914). "Bl." at the left of a reference in the column of real exceptions means that Blass in the Attische Beredsamkeif (1893) recognizes this as a real violation of his "law." In all other cases "Bl." refers to Blass's Teubner text (1884). In his Attische Beredsamkeit Blass did not always follow the readings of his Teubner text, hence in the column of real exceptions some of the breves recorded as recognized by Blass as real violations of the law are cases that in his earlier (Teubner) text he had removed by one means or another. De corona, xviii [The asterisk marks an instance that is not to be counted in reckoning the minimum number of breves. See p. 283.] 1 . 1—13 . 10. Fkst 100 Oxford Lines Beal Exceptions Apparent Exceptions *i. iT€pl iixavTOV 2. cr. Tg anoXoyia 5. iravras av 6/*oXoyi}a-ai (vulg.; 3. p. /«.«v | on oi (, BL, H.) iravras after v/xas S, L, Gd., 9. p. StWov a/m | fipdxe' (, Bl.) H.) 10. e. Koiva w«p£u(KotVBl.,H.,F.) •7. n-epi diravTcov (S, L, al. ; irtpl 10. cr. y^eipovaKoXiiJik{KaiJi.(B\.,Ji.) jrdvTuv vulg.) 9. €vdvi av airtXoyov/Jiriv ■^10. Trepl i/Mtv *13. ypd6T(pa Tavr' Bl. 23. ovTos vyws irepi 25. Trdvra u BL, H.) 25.7—41.2 26. ^XirtiraTe rijv (om. T^v B, al., BL, H.) *32. /li) Sxi/jnv ' iavTiS (aw Apparent Exceptions 63. sp. TTOTtpOV avT^v (iroTtp' ?) 64. sp. 5ro'Xtv()8o«A«T' (v/3ovX- Bl.) 64. p. ytyvo/«va | cjri Tg (, Bl.) 69. p. avayKatov afux | iroo'tv (,B1.) ^ _ 69. sp. dAAa Ti txi"^'' (XP^'' Bl') H.) 70. e. yeyowi/oiSa (yryov'Bl., H.) 71.8—89.5 *75. tiTa #iXoKpaTJ7S 79. [a Ti jrepi e/iov ■ycypa^ev] Bl. 81. avTos 6 ^iXwriros *82. irepi (/«)v *87. xdpaKa /3aXo/XEVos {xapoKm/ta Syp. A: \apaKa>fux,Ta U', F.) 76. sp. amarai ?/*' {jrtp Tov (/i' Bl.) 82. p. oiScTTor' I ikv (, BL, H., F.) 83. sp. l/i' «5ri Tovrois (/i' Bl.) 89. p. TToXenios 1 a.v€v (, BL, H., F.) 89.6—103.5 93. 'EXX^oTTovTov wo ^iXimrtg 93. cr. 98. KivSweixrerc 8ia\oyurdfuvoi 97. sp. *99. ijTO 0«/iicro)Vos 99. ToT« TpiYjpdpxiov (S, L, H.; 103. cr. Tpvr)pdp\ti>v Tore vulg., Bl.) 100. e. BL 102. Tovs Sk fjLfTpi' ^ 101. p. Bl. 107. Bl. 107. Bl. 108. 103.6—119.3 OVTOV direXeii^di? 104. cr. Bwa/Jieini avdyevdoj. 105. ext. TaSwara a~uvipaxvfv ([0W£/3-] 105. ext. Bl.) 109. cr. Bl. 111. /3iov VTrvodvufK elvau. bpjoKoyu> oTi tm Tw 111. V. Sta y« toSto (S, L, B, F, al., Gd., H., F. (toCt'); Si' ailTo' 111. sp. ye TovTo L*, A, vulg.; hia TovTo y' Bl.) 111. p. TTcpi t/xov 118. ext. a jucv cn-eocoKa Ta SiSofitv' o/toXoyfiv (SiSo- jucva o/JMk- Bl.) 113. 113. *113 BL 119. 119 irpoaipco-ts ij €/x^ KaTa^evyovTas c<^' iavTovi (ouToiis BL) Kai w7rfvyov, (Kair- Bl.) €iruTT(v$-qT€ woXoyiuTixa KaO' o Kara rov ifwv iroXiT€vfia(ri Kal iv tois (iroXiTtujoiao-iv F.; [iroX-] Bl.) cXcye ircpi t5v (eXtytv BL, H., F.) avToi tSwd/uiyv (^8w- BL, F.) 6/xoXoy5 I alv (, BL, H., F.) fiapTvpci €' ois Demosthenes' Avoidance of Bbeves 277 119.4—132.4 Seal Exceptions 120. oiTOV av avapprjOrj *125. wXtOV ^ fXaTTOV 128. XeyovTOS ipvOptamic {i]pv6- pCatrev Bl.) 130. SiyXovoTi TaVTJJS Apparent Exceptions 120. p. koyCa-aveai \ on (, BL, H., F.) 121. ext. Ttvas o SijfWi 123. p. as ravra | Sltto (, Bl., H.; toot' F.) 124. aph. fx9pov ^ i/wv (^ 'fwv) 127. ext. Kot Apery 127. ext. Ta KoAo. Kttl Tot atcrxpa 8io- ■yiyvioaiceToi 130. e. aiv €TV)(ev ^v (irvx Bl.) 130. e. iiroiTj vvv in vulg., Bl.) *139. £>(? vSan A^X^ 1 dn-a rrji (, Bl., H., Koi on KCU, OTl jrdXe/itov tis TToAt/iOv 'Aix(J>i,ktvovik6v 144.6—161.4 151. KarriKOVTUTav airavros (S corr., pr. L, A; awuvT. kot. F., vulg.: om. on-ovTos S', Bl.) 160. paXuTTa &' OTl (8c, OTl Bl.) 161. Trap' iKaripoK 147. V.(p.)'A/i<^lKTVO(n I Kol (-0-IV BL, F.) 160. p. di/rofio-fj. I «'s (, Bl. H., F.) 161.5—177.9 161. yevoiTO irapaTrjpmv (S, L', 162. aph. fiaXXov ^ i/xov (jxov Bl.) Gd., H.; ytviJo-tToi irapoT. 169. cr. A, vulg., Bl.) 170. ext. Bl. 162. jrpoTcpov ^ (irpoTtpiov Bl.) 174, e. 163. oTt Tov iv ' A.pxf>UTa-rj apa TTj rip^pa Ti's ayopeutw iTVY)(a.vtv ?XOV, (eTvy^ov' Bl.) 278 Charles D. Adams Beal Exceptions Bl. 168. Oopvfiov ttrre tare ijkv airavres {66pvj3ov, t6Ttpa Tavra *172. lirpaTTiv 6 ^tXtJTTros {firpaTTt Bl.) Bl. 173. oCtos iv fKilVrj Ty '^p-ip^ iy$ (kotti- Bl.) Bl.) 196. cr. Kal oKpowftevovs ((cdxpo- TOV ayadov iroXlryiv Bl.) 196.4r-208.10 BL197. oo-ov iyio 197. p. OKvrjirai tStov | ov8'(, BL H.) ♦B1.202. . irporepov icrxyptov {irporepwv 197. cr.,e . Kol dpa 'Api(TTpaTpovripa 211. owoOev «VTaC^* (S, L; odfv ets (,B1.,H.,F.) _ TavTa cett. ; iirdvap,' otv odev 210. p., e. , i€v ovTo' Bl., H.) avTt St TOV tv Trpoa'Ti$rifU oTt r^s (wpotTTi- ^iu.'Bl.,H.) 234.8- 238. 239. *244. 244. 245. *247. *248. 248. 252. Bl. 258. 259. SiwXdxna Ttav air€p tveS^cro -248. 236. wapa roui (tveSex- perhaps gloss from 241. ivvv Weil.) 243. wapa <^iXiinrov 246. ovSi rrapa t&v (ovS« S, L, 246. Gd.: oi cett., BL, H.) re imkaKiav TO kot' «ju.e. irept €/«)C Kai iopaKUis (cupafcus codd., Bl.) 248.6— 260. i avOpunrivioTtpov iyla 249. Tivt Ki)^rj Bl., H., F.) 273. p. SrjXovoTi 1 Kol 276.2—288.11 284. toBto lUTaTiOffieyoi to. 276. aph 288. olvwoTov (S,L,Gd.,H.; [] Bl.) 278. cr. Bl. 288. alpctfo/TCS em rots ([iwt Tas 279. P- . . . . ]B1.) 282. P- 288. TO iripiSatrvov 284. cr. 288. p. 288.12- -302. 1 288. jUaAxCTTa Blfip€V, 288. ext. *288. VTTtp awdvTiav (iravrwv Bl., 291. P- H.) 296. p. 291. uero piv e/xoS ([cjuoti] Bl.) 298. P- 293. iirel t/wiy' 298. P- 297. p.ev wo ^iXiirwov ([ ] BL, 298. P- Butch, after Dobree) 298. iirl TO X^/xjua 276. aph. imWov ^ i/wl ('/no' BL, H.)_ Kal ayaJdov dv7}Xa)KO>ai | tSuxs (, Bl.) prfTOpOi I ^ a (, Bl.) TO ovofiara aStXoi I ol vjTO (, F.; [ol . . . .] Bl.) TO tiriypaij,ita tavTOV I oTt rots (, Bl., F.) iraTpuTW I S>vmp (, BL, F.) /teyc^os I ovr' (, Bl.) ovTt Kpo^oi I ovr' (, Bl.) irporjydytTO \ Sw (, BL, H., F.) ^ irarpiSi | ovStv (, BL, H., F.) ^ Kal aSiad6pov (xaSt- BL, H.) 298. 298. cr. 302.2—316.6 Bl. 303. ov8i irpoidevd' {irpoeBivTa S, A, Y; vpohodhna vulg.; TTpaOevTa Dobree; irpaOevd' Bl.) Bl. 307. Si Tis iSuj. Bl. 312. a)(ei ipavov BL 316. t5v vpoTtpov eiipyiorlas (S, L, al.; irpoT«pv 305. ext.,cr. (caTct to. i/xa \f/ri<^UritaTa {T&pk BL, H.) 310. e. iSioKtv 6 irapeXBiav («8«)x' BL, H.) Demosthenes' Avoidance of Breves 281 316.7—324.6. 58 Lines Beal Exceptions Apparent Exceptions 317. ■yap oTi KOI 321. p. Xeyovri | aveTnf>6vr}fj,a § 209), although his failure to do so not only leaves the misleading groups of breves, but such harsh hiatus as KexpT}fi.i:vov avkyKif §34; diMKoyu &v § 111; 'Xoyiffaa-dai on § 120; AiiX dxo § 134; afanevu eis § 160. We must not, of course, assume that every pause that would be sufficient in delivery to pre- vent a succession of breves is to be marked by punctuation in our texts; we are dealing here with rhetorical considerations; the printed punctuation takes account both of rhetorical and of gram- matical demands. We should hardly punctuate in such expressions as the following, although there is a clear rhetorical pause: 6pyi.^6- fjtevoL I irolnus (brr. and hiatus) § 20; ael | ofwKoyu {brr. and hiatus) § 31; a5ekoi \ ol iirb {brr. and hiatus) § 288. The assumption of the use of crasis removes 23 more of the appar- ent exceptions. We have no means of proving how far an orator used crasis in delivery, but we must assume that it was used fre- quently. 282 Chakles D. Adams We are certainly to assume the free use of elision also in delivery, as Butcher does in his printing as a rule; yet even in his text 15 cases of breves appear where elision should be reognized in the printed text. Humphreys prints 9 of these with elision, Fuhr, 7 of them. Good- win's text seems to have been prepared with no reference to the effect of eUsion on breves; in the first 200 (Oxford) lines of the speech, Goodwin's text shows 11 cases of breves that would be removed by eUsion,' and are so treated by Humphreys, and by Butcher with one exception (koivol iiiripev § 10). Of course all these are cases of appar- ent hiatus also, therefore doubly offensive. The fact that Goodwin does use elision freely in his printed text makes his failure to use it in cases like these the more misleading. II. THE REAL EXCEPTIONS The 137 real exceptions, an average of 6 to the hundred Oxford lines,^ form the evidence on which we must base the attempt to deter- mine the range of the "law" or tendency of Demosthenes' avoidance of breves in the De corona. It may be, however, that we ought to reduce this list of 137 real exceptions, for included among them are some cases of breves that are not entirely certain, and others that are practically unavoidable, and therefore perhaps to be treated like breves within a word. 1. In 5 cases there is some question of the pronunciation: in § 4 and § 321 irepi ifiavrov may have been pronounced irepl 'navTov; in § 32 nil Sacinev may have been pronounced as though written iiTfTrinev; in § 125 ri eKarrov may have been treated in a similar way; in § 172, while we should hardly print eirpar^' 6, that may have been the pronunciation (Blass, Rede vom Kranze, p. xiv). 2. In some cases we may be justified in assuming a form of the word that is slightly different from that of the manuscripts. So fiov for the unemphatic efiov in xepi ifwv §§ 10, 57, 82, 113, 248. In §§ 7 and 59 Trepi airavruv may have been wepl iravTOiv; so also mep airav- Twv §288. For irpbrepov §202, Blass would write irporepcav; for ayacrakTo § 204, he accepts Cobet's emendation, ayaxrOeit]. ^ T€ir pay /lira iiw.VTQ §4; re ISlov §8; KOivk inripev §10; ivSiSuxBe iirl §10; fieyiXa fxovcai § 14; Tpdy/tara iX^x<"" 5 l^! *"■" i^Tr^p § 17; iyuye iToXirev&iirjv § 18; (5(rre dveXeiv § 18; xP^M<*^<* d.va\i(rKojv § 19; SiJttotc Ivc/ca § 21. 2 In the following discussion I shall occasionally use the term "per cent" to indicate the number of groups of brenes per hundred lines. Demosthenes' Avoidance of Breves 283 3. The breves in the following fixed phrases are unavoidable: yp6iovTa irapavofjia § 13; x^^paica fia^fievos § 87; t<^ ifiQ ibari, § 139; hrl TO firjp.' § 209; rd Kar' ifie § 247. In some cases, though by no means in all, the breves that lie in part in proper names are practically unavoidable. Such cases are : fieri ^iXoKpoLTovs § 17; inrd ^iXLtttcp § 48; elra ^iXoKpdrijs § 75; inrd Qefucruvos §99; fierd, ^iKiinrov § 229; cure 'P6dos §234 (possibly also a pause after ToSos) ; irapi. ^iXlinrov § 244. If we deduct the 27 breves of these three groups (marked by an asterisk in column 1), 110 breves remain in the 2,258 lines, an average of 4 . 9 per hundred. • This may be taken as the minimum average for the De corona. But in comparing this speech with others it is gen- erally better to use the maximum figure, 6; for something of the subjective element enters in, the moment one begins to reduce the number on the ground of possible change of pronunciation or of the supposed necessity of the use of set phrases, etc. In comparing large blocks the relative frequency of breves will be about the same, whether we reckon the maximum or the minimum number in each block that is involved in the comparison. In dealing with a short block, like one of the Philippics, the repeated use of some one set phrase or some one proper name may considerably affect the result; here it is safer to use both minimum and maximum figures. Wher- ever in the following discussion minimum figures are given, they are obtained by excluding groups that correspond to those analyzed above for the De corona. Our results are affected, of course, by some questionable manu- script readings. Such variants in the De corona are recorded in our analysis. The only safe course here for our investigation is to use the reading that is best on grounds other than those of rhythm. How far, after the facts of rhythm have been established, these facts themselves may be used in deciding between variant readings in any given case should become clear as our investigation proceeds. The fact that there is some uncertainty of manuscript readings of course makes our conclusions to that extent doubtful, but the number of such cases is so small in comparison with the mass of material included in our study that the results are not seriously affected. 284 Charles D. Adams Our next step is to determine to what extent Demosthenes was indifferent to the occurrence of breves in groups caused by the lighter words, like the prepositions, the articles, the particles, etc. The article enters into only 19 such groups (including 5 where it follows a preposition of two short syllables, and 2 where it precedes one).^ In 3 of these the article stands with a proper name, and in 4 it is in a set phrase («rt r6 ^rjfia §§ 60, 171, 209; t6 Kar' ijil § 247). Considering the frequency of the article, the fact that it occurs so seldom in breves warrants the conclusion that its occurrence was objectionable to the ear of Demosthenes. Blass implies that the orator allowed himself a much greater freedom than we find in this examination. He says (p. 108): "Ferner gilt bei ihm wie bei den Tragikem die Verbindung einer Praposition mit einem Nomen einem Worte gleich, und bei dem Redner auch die Verbindung eines Artikels mit dem Nomen, oder mit Praposition und Nomen, wahrend die Dichter bei ihrer Freiheit in der Auslassung des Artikels in diesen Zwang nicht kamen." This statement is to be modified by Blass's qualifying words, "Jedoch was an und fur sich erlaubt ist, wiirde fehlerhaft werden, wenn es zu haiifig kame." Blass's suggestion that when Demosthenes does allow the article in breves it is from the feel- ing that article and noun are practically one word is misleading. Breves that occur within a word must be allowed, or the word cannot be used at all; but breves caused by an article followed by its noun or an attributive modifier of the noun can usually be avoided, as we see by the fact that only 12 such groups of breves occur in the De corona, three of which are with proper names, and three in the set phrase ^Tri rd Prjfia. The same criticism applies to Blass's statement as to breves formed by a preposition and its object. Demosthenes' use of such groups is very restricted. There are 10 instances of Trepl in breves in the De corona, but they are all in the phrases xepl e/xoO (5 times, per- haps to be read wepl pov, see p. 282), Trept airiLvrosv (twice, perhaps to be read repl iraproiv), irepl ifiavrov (twice, perhaps to be pronounced irepi 'navTOv), and irepi iKaripov (once).^ The 10 instances of irepi in J Cases where the article is subject to crasis are included in the "apparent excep- tions." « I do not count irepl i/Mu in % 79, bracketed by Blass and Butcher. Demosthenes' Avoidance of Breves 285 breves form not quite 12 per cent of all the instances of irepi in the speech; but in Aeschines Vs. Ctesiphon 34 per cent of the occurrences of irepi are in breves, and in Lysias Vs. Eratosthenes, 42 per cent. From this it is clear that Demosthenes distinctly restricts himself. 'Tx6 is in breves 9 times; two of these cases show the preposition followed by a proper name; three others have mo ^iKlirirov or ^iKlinr(fi preceded by one or more short syllables; in the 4 remaining cases the breves are made by viro with preceding syllables, not with its own object, and do not therefore fall under Blass's principle of treating preposition and object as essentially one word. The 7 cases of vir6 in breves (excluding the 2 with proper names only) form ISJ per cent of the occurrences of inro in the speech, as against 32 per cent in Aeschines Vs. Ctesiphon, and 43 per cent in Lysias Vs. Eratosthenes. Demosthenes therefore shows a clear tendency to avoid breves with ii5r6. There are 7 cases of hri in breves, but 3 of these are in the fixed phrase M rb fiijua. In 3 of the 4 remaining cases the breves are formed by eiri with the preceding word, not with its object. In the 4 cases of irapd in breves one is with a proper name, and two are with the preceding word. Two of the 3 cases of inrip in breves have the breves in the preposition and its object. Of the 2 cases of Kara in breves (both in set phrases) one has the breves wholly in the preposition and its object. The 2 cases of ixeri, in breves are with proper names. The only instances of 5id in breves (out of 44 occurrences of 5iA in the speech) are in the phrases 5iA ye tovto (§ 113) and Sid rd . . . . ^X*''' (§ 213). We find iv in the speech 162 times, but only 5 times in breves; in no one of these groups are the breves wholly in the prepo- sition and its object. In the single instance of &w6 in breves (out of 22 occurrences of the word) the group is made by the preposition with the preceding word (§ 230). We must conclude that Demosthenes generally avoided groups of short syllables with prepositions. In case of prepositions with proper names or fixed phrases the use was unavoidable. In other cases the breves are as likely to be due to the word before the prepo- sition as to its object. All instances of breves with prepositions, except with proper names or in fixed phrases, must be included in reckoning the minimum number of breves. 286 Charles D. Adams Blass treats Srt as an exception to the law of breves, but it does not occur in breves in the De corona often enough to justify his posi- tion. The 9 instances are only 15 per cent of all the occurrences of 6ti. in the speech, as against 47 per cent in Aeschines Vs. Ctesiphon, and 45 per cent in Lysias Vs. Eratosthenes. M^v occurs in breves 5 times, 5^ 3 times, and &p 8 times. While our analysis of these instances shows that Demosthenes did not admit the article, the prepositions, and the lighter particles freely in breves, yet the fact that the groups containing these words make up about 60 per cent of all the breves of the speech, shows that he did admit them much more freely than he did the more substantial words, viz., verbs, substantives, adjectives, demonstratives, etc. The bearing of these facts on textual criticism is evident. Breves contained in the lighter words of the sentence, while restricted, are nevertheless not so rare as to warrant textual alteration in any given case, or to have much weight in the decision between variant read- ings. But breves in groups of the stronger words, or at the end of one of them, are rare enough to warrant giving considerable weight where good manuscripts dififer; they are not rare enough to warrant textual alteration on the ground of the breves alone. In any attempt to determine the frequency of ejJceptions to the "law" of the avoidance of breves, the groups with the article, prepo- sitions, and the other lighter words must be included with the groups involving the stronger words; neither sort of breves is freely admitted, neither one is entirely excluded; the question is one of relative fre- quency, and for that no precise statement can be made. If we include all real breves in our discussion, we have a safe criterion for the com- parison of different parts of the same speech with one another, and for the comparison of one speech as a whole with another speech. An examination of the several divisions of the speech shows a marked tendency to avoid breves as the thought becomes more impres- sive and the style more elevated; yet this is not without exceptions. In the proem the breves are considerably less than the average of the speech, 3 cases in 58 lines (5 per cent) ; two of these are of the least objectionable type, irepl ifxavrov § 4, irepl airLvrav § 7 (see p. 282). The fine discussion of the situation at the time of the making of the peace, §§ 18-25, 66 lines, has 2 cases (3 per cent): a/juttdTtpa ravr' § 20, and o5to$ vyUs § 23. Demosthenes' Avoidance of Beeves 287 The narrative and discussion of the acts following the conclusion of the peace (§§ 25-50) show only 3 cases of breves in the first 105 lines (§§25-41): -irapra crwayaivL^ofiepos §25 (a very doubtful reading), ^Xxttrare rrip § 26 (another very doubtful reading), and fj.ri iiniiev § 32 (see p. 282). But as the argument and attack proceed we find 4 cases in 24 lines (§§ 41-44) ; then after 29 forcible lines that are entirely clear we are surprised to find 5 cases in only 15 lines (§§ 48 f .), and that in a passage of deep feeling and fine rhetorical finish. The section as a whole, 173 lines, with 12 breves, has a rate of 7 to the hundred lines, a rate slightly greater than that of the speech as a whole. The eloquent discussion of Demosthenes' services up to the cam- paign at Byzantium (§§ 60-87, 160 lines) shows only 6 cases, and most of these of a minor sort. The discussion of the rescue of Byzantium (§§ 87-102) has 5 cases in only 106 lines, but only one of these (nvdwevaere diaSoyur&fJtevoi, § 98) is in the more elevated part. The discussion of the naval reforms, a part perhaps the least impressive in the speech (§§ 102-10), has 4 breves in 53 lines (a high rate), including 2 bad cases of hiatus. The answer to the technical charges under the laws, another of the less elevated sections, has a like rate, 9 breves in the 109 lines (§§ 110-26). The solemn and elevated passage in which Demosthenes accuses Aeschines of having been the guilty cause of the Amphissian war (§§ 140-60, 129 lines) has only one group of breves and that of doubt- ful manuscript authority. The transition passage that follows (§§ 160-69) has 7 cases in only 37 lines, a very high rate. The story of the news from Elateia, a famous masterpiece of narrative, has no breves in the strictly descriptive part (§§ 168-71), but as narrative gives place to argument the breves become comparatively numerous. The 87 lines (§§ 171-81) show 14 cases, a rate of 16 to the hundred, a surprisingly high rate. The elevated passage in which Demosthenes discusses his services in the crisis (§§ 188-213, 201 lines) is in its earlier part almost wholly free from breves, having but 2 cases in the first 123 Knes, ocrop eyc!> § 197 and irpdrepov laxvpuv § 202; but after this they become more frequent — 6 cases in 78 lines. The "Marathonian oath," however (§§ 206-8 incL), is entirely free from them. 288 Charles D. Adams The effective passage (§§ 227-52) in which Demosthenes reviews his career in the form of the Koyt^f^s suggested by Aeschines shows more breves than the average, 16 cases in 189 Knes, nearly 82 per cent. The long discussion of Tvxn (§§ 252-76), where we might expect more freedom of style, has remarkably few breves, only 5 in 166 lines, 3 to the hundred — only half the average of the speech. In the indignant passage (§§ 276-85) in which Demosthenes takes up Aeschines' caution to the jury to beware of Demosthenes as a crafty rhetorician, there is only one group of breves (ravTa fterariSinevos §284) in the 69 lines. The outburst against the charge of Philippizing (§§ 294-301) has only 2 breves in the 61 lines, fnh mb ^CKlt^tov § 297, bracketed by Butcher after Dobree, and ^xi t6 X^p^ua § 298. The closing paragraphs, §§ 301 ff., have 7 breves in 165 lines, a rate somewhat less than that of the speech as a whole. If, with Schafer and Blass, we consider the epilogue as beginning with § 252, the epilogue as a whole has 22 breves in 525 lines, about 4 . 2 to the hundred, a rate considerably less than that of the speech as a whole. Having determined with a reasonable degree of precision the extent of Demosthenes' avoidance of breves in a speech that was pre- sumably prepared with the greatest care, and that represents his matured art, I proceed to an examination of the whole Demosthenic corpus. We may hope to learn whether there was any advance in Demosthenes' style in this particular between his earlier and his later works; whether in the private speeches, written for other men, he used as much care in this matter as in the case of speeches for public suits or for the bema; and whether the rate of breves in any speech attributed to Demosthenes may safely be used as a criterion of genuineness. I have recorded only the real exceptions; in determining these, I have followed steadily the same principles as in the preceding investigation. It is not likely that any two investigators would always agree as to which class of exceptions are real, and which only apparent ; but if each follows his own principles consistently through- out the whole corpus, the results will not differ materially. A tend- ency, e.g., on the part of one to assume less freedom in the use of crasis would give a higher total of breves for any one speech, but would not materially alter the ratio of breves between one speech and Demosthenes' Avoidance of Breves 289 another, or between the speeches of Demosthenes and those of another writer — and this ratio is the essential point of our study. For speeches i to xxvii, inclusive, I have used Butcher's Oxford text; for the rest, the text of Baiter and Sauppe, Zurich, 1838. The Zurich text follows S closely, and is not influenced by rhythmical theories of the editors; its lines have been reduced to the Oxford standard of length. I have added a study of breves in selected parts of other prose writers; such comparison is necessary for any just estimate of the significance of the results of the study of Demosthenes. It is here to be remarked that the tendency of any writer to avoid hiatus within the kolon will of itself reduce the number of accidental breves, for in writers who admit hiatus freely a group of short syllables is often due in part to the shortening of a long vowel in hiatus. In all the authors studied the lines have been reduced to the Oxford standard of length. In Aeschines the numerous phrases that were bracketed by Blass have been treated as belonging to the text. It will be found that in the passages from other authors the breves are distributed fairly evenly throughout the passage; the number in any one block of 100 lines does not vary greatly from that in any other; in the part of Isocrates examined the range is from 35 to 43; in Aeschines it is from 32 to 49; whereas in Demosthenes they are distributed unevenly, the breves in the De corona reckoned by suc- cessive hundred lines being 6, 7, 2, 10, 5, 4, 6, 10, 4, 8, 3, 14, 5, 6, 7, 10, 8, 3, 2, 4, 6, 4. An even distribution indicates accidental occur- rence, uneven distribution suggests conscious avoidance, now more careful, now less. BREVES IN THE DEMOSTHENIC CORPUS Real Exceptions Rate per Hundred Lines [Titles in parentheses are of speeches regarded by Blass as not genuine, or of doubtful genuineness. Figures in parentheses, after the number of breves in certain speeches, give the minimum number of breves, reckoned as explained on p. 283.] OP THE FIRST PERIOD Pleas in Private Cases (Vs. Callippus. lii.) 36&-368 B.C. brr. 55.3 (Vs. Nrcostratus. liii.) 366-365 b.c. brr. 60.6 [The above-named are too early to be genuine.] 290 Charles D. Adams First vs. Aphobus. xxvii. 363 B.C. ftrr. 44.8 Second vs. Aphobus. xxviii. 363 B.C. 6rr. 43.2 Vs. Aphobus for Phanus. xxix. 363-362 B.C. 6n-. 20.9 First vs. Onetor. xxx. 362-361 B.C. brr. 36 Second vs. Onetor. xxxi. 362-361 b.c. brr. 23 (Vs. Timotheus. xlix.) 362 b.c. brr. 60.5 (Vs. Polycles. 1.) ca. 358 B.C. brr. 57 OF THE SECOND PERIOD Political Speeches and Pleas In Public Oases On the Trierarchical Crown, li. 359-358 b.c. Vs. Androtion. xxii. 355 B.C. Vs. Leptines. xx. 354 B.C. On the Symmories. xiv. 354 b.c. Ilepi SwTo^eo)?. xiii. 354-351(?) b.c. Vs. Timocrates. xxiv. 353-352 b.c. For the Megalopolitans. xvi. 353 b.c. Vs. Aristocrates. xxiii. 352 b.c. First Philippic, iv. 351 B.C. For the Rhodians. xv. 351-350 b.c. First Olynthiac. i. 349 b.c. Second Olynthiac. ii. 349 b.c. Third Olynthiac. iii. 349 B.C. Vs. Midias. xxi. 348-347 b.c. Pleas in Private Cases Vs. Conon. liv. ca. 356-355 B.C. brr. 16.5 (Vs. Euergus and Mnesibulus. xlvii.) 353- 352 B.C. brr. 49.3 For Phormion. xxxvi. 350-349 B.C. brr. 15.6 First vs. Stephanus. xlv. 349-348 B.C. brr. 9 . 1 (Second vs. Stephanus. xlvi.) 349-348 B.C. brr. 39 First vs. Boeotus. xxxix. 348 b.c. brr. 17.2 (Second vs. Boeotus. xl.) 347 B.C. 6n-. 49.8 OF THE THIBD PERIOD Political Speeches and Pleas in Public Cases On the Peace, v. 346 b.c Second Philippic, vi. 344 B.C. On the Embassy, xix. 343 B.C. On the Chersonese, viii. 341 B.C. Third Philippic, ix. 341 b.c. Fourth Philippic, x. 341-340 b.c (Vs. Philip's Letter, xi.) 340-339 b.c brr'. 33 brr. 2.4 brr. 5.2 brr. 7.3 brr. 9.6 (8.3) brr. 11 brr. 6.8; 37.5 brr. 8.4 (7.5) brr. 7.1 brr. 6.1 (3.4) brr. 5 (4.2) brr. 2.8(2.8) brr. 2.5(2.1) brr. 7 (4.8) brr. 6.8 brr. 4.5 (3.4) brr. 6 (3.8) brr. 7.5(5.5) brr. 4.3 (3.5) brr. 9.4 (7) brr. 6.9 Demosthenes' Avoidance of Beeves 291 On the Crown, xviii. 330 b.c. hrr. 6(4.9) First vs. Aristogiton. X3cv. ca. 325 B.C. hrr. 9 (Second vs. Aristogiton. xxvi.) ca. 325 B.C. hrr. 36 Pleas in Private Cases Vs. Pantaenetus. xxxvii. ca. 346-345 B.C. hrr. 10 Vs. Nausimachus. xxxviii. ca. 346-345 b.c hrr. 8.3 Vs. Eubulides. Ivii. 345 B.C. hrr. 32 (Vs. Theocrines. Iviii.) 344-339 b.c brr. 37.6 (Vs. Neaera. lix.) 343-340 b.c. hrr. 44.9 (Vs. Macartatus. xliii.) ca. 342 b.c 6rr. 56.8 (Vs. Olympiodorus. xlviii.) ca. 341 B.C. hrr. 60.4 (Vs. Lacritus. xxxv.) 342-341 (?) b.c hrr. 54.5 (Vs. Phormion. xxxiv.) 327-326 b.c 6rr. 49.1 (Vs. Apaturius. xxxiii.) After 341 B.C., perhaps of the time of Alexander hrr. 49.8 OF THE PEHIOD OF AUEXANCEB Fleas In Private Cases (Vs. Phaenippus. xlii.) brr. 42. Z (Vs. Zenothemis. xxxii.) hrr. 16.3 (Vs. Dionysodorus. Ivi.) 323-322 b.c trr. 55.4 OF UNDETERMINED DATE Pleas in Private Cases and Epideictic Speeches Vs. Spudias. xli. hrr. 49.5 Vs.Callicles.lv. 6rr. 44.5 (Vs. Leochares. xliv.) hrr. 45.6 (The Eroticus. Ixi.) hrr. 33 (The Epitaphius. Ix.) hrr. 11.9 THE PBOEMS [The table gives the number of the proem, the number of (Oxford) lines, the number of breves, and the rate per hundred.] xi 13—0— xii 16—3—19 xiii 12—0— xiv 19—2—10.5 XV 15—1— 6.6 xvi 13—1— 7.7 xvii 13—1— 7.7 xviii 14 — 1 — 7 xix 13—1— 8 XX 17—0— i 25—1— 4 ii 26—2— 7.7 ill 9—0—0 iv 11—1— 9 V 20—1— 5 vi 18—1— 5.5 vii 18—0— viii 19—1— 5.3 ix 18—0— X 17—1— 5.8 292 Charles D. Adams xxi 35—2— 5.7 xxxix 19—1— 5 xxii 17—2—11.8 xl 19—2—10 xxiii 18—1— 5.6 xli 26—3—12 xxiv 33—1— 3 xlii 16—0— XXV 24—1— 4 xliii 16-0— xxvi 22—0— xliv 17—2—12 xxvii 13—2—15 xlv 26-0— xxviii 18—0— xlvi 32—0— xxix 20—3—15 xlvii 18—0— xxx 22—0— xlviii 21—2— 9.5 xxxi 17—1— 5.9 xlix 24—0— xxxii 30—0— 1 25—2— 8 xxxiii 22—1— 4.5 li 9—1—11 xxxiv 26—3—12 m 13—0— XXXV 30—1— 3.3 mi 35—3— 8.6 xxxvi 16—2—13 Hv 9—1—11 xxxvii 14—0— Iv 24—1— 4 xxxviii 22—2— 9 Ivi THE EPISTLES 25—3—12 i 112— 7—6.2 iv 80—3— 3.75 ii 173—14—8 V 34—9—26 iii 295—11—3.7 vi 16—5—31 THE RATE OF BREVES IN OTHER WRITERS, FOR COMPARISON WITH THAT OF DEMOSTHENES Thucydides. The Funeral Oration of Pericles. 220 (Oxford) lines; brr. 110. Rate 50 per cent. Lysias. Vs. Eratosthenes. First 300 lines; brr. 48, 47, 47. Isocrates. Panegyric. First 600 lines; brr. 41, 43, 35, 43, 38, 37. Average 39.5 per cent. Isaeus. xi. 403 lines; brr. 40, 32, 41, 47 (+3 lines with 2 brr.). 40 per cent. Isaeus. i. 321 lines; brr. 27, 39, 44 (+21 lines with 7 brr.). 36 per cent. [Speech I shows the most definite attempt to avoid hiatus of any of Isaeus' speeches, and it is nearly the latest in time.] Aeschines. Vs. Ctesiphon. First 500 lines, drr. 33, 36, 49, 45, 43. §§152.1 —162.9, 100 lines, brr. 49. Last 200 lines, brr. 37, 32. Average 40.5 per cent. Hyperides. Epitaphius. 259 lines; brr. 115. 44 per cent. Hegesippus (?). On Halonnesus. 287 lines; 6rr. 126. 44 per cent. Anonsrmous. litpl rSiv wpoi 'AXt$avSpov 2,vv6ii)kS>v. 217 lines; brr. 114. 52 per cent. Demosthenes' Avoidance of Beeves 293 The space limits of this article forbid the detailed discussion of the results here tabulated. It is plain that in the first two published speeches, xxvii and xxviii, there was no attempt to avoid breves; but with the third speech, xxix, the avoidance begins, and in a few years it has become fully established as a rule of composition. The public speeches naturally show a lower rate than those in private suits. The speeches Vs. Spvdias, xli, and Vs. Callicles, Iv, generally recognized as genuine works of Demosthenes, offer no evidence of date save their style. They are almost universally regarded as youthful works of the orator, and as such their high rate of breves is not sur- prising; they belong with the speeches Vs. Aphobus and Vs. Ortetor. It is seen from the table that two speeches which cannot safely be ascribed to Demosthenes follow the rule of avoidance of breves. The Epitaphius, Ix, probably by a writer after the time of Demos- thenes, shows that at least one rhetorician knew this secret of the Demosthenic composition. The speech Vs. Zenothemis, xxxii (brr. 16.3), was delivered, and probably written, by a kinsman of Demos- thenes; it is not surprising to find an imitation of this refinement of his style. All other speeches of the corpus which on other grounds must be declared non-Demosthenic show a high rate of breves. On the other hand, almost all that are certainly Demosthenic have a low rate. Two offer difiiculties which may be removed on the supposition of composite authorship; the detailed discussion of the grounds for such a hypothesis would be too long for this article. These are Vs. Evbulides, Ivii {brr. 32), and Vs. Timocrates, xxiv {brr. 6.8 in one part, 37.5 in the remainder); both certainly show the hand of Demosthenes; but difficulties both of composition and of argument are best solved on the supposition that in each case the speaker received from Demosthenes only a part of his material. The bearing of our study on the question of the genuineness of the Fourth Philippic is evident. The results for the Midiana, a speech that manifestly lacks final revision, show that at an early period the habit of avoidance of breves had become so established with Demosthenes as to prevail even in the first draft of a speech, a fact that has not been sufiiciently recognized. 294 Charles D. Adams The rate of breves in the Proems has about the same range as that of the speeches. Of course the small number of lines in each proem makes the percentage of breves less significant — ^there is more possi- bility of a chance absence or a chance abundance; yet it is certainly significant that so many are entirely free from breves. No one of the fifty-six proems has a rate so high as in itself to raise the question of genuineness. Epistles V and vi, on the other hand, pass beyond the rate of Demosthenes' matured style. Whether the low rate of the other epistles is, like that of the Epitaphius, due to the skill of an imitator, well instructed in the niceties of the Demosthenic style, is a question that must be settled on other evidence. Our study as a whole has shown that the avoidance of breves is a real characteristic of Demosthenes' style. A wide interval sepa- rates the least-polished product of his pen from the best work of his great contemporaries. Apparently there were some men of his own time — perhaps a narrow circle of personal friends — and some of a later time, who understood and imitated this peculiar device of prose composition, but none of the great writers were among them, and the principle did not find its way into the rhetorical treatises. Dabtmouth College