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 by JSTOR. 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.jstor.org/participate-jstor/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. THE BOOK DIVISIONS OF THUCYDIDES By Robert J. Bonner All the extant manuscripts of Thucydides are divided into eight books. But both a nine- and a thirteen-book edition are mentioned by ancient writers. Diodorus 1 refers twice to a nine-book edition, oiros niv ovv b woXepos dUnetvev iirl err] eUocn eirra, 6 8k QovKvdidrjs errj Svo irpds rots &uxn yeypas di rives haipovffLv, kvvia.. Marcellinus 2 reports that the eight-book division was the most common. He mentions a thirteen-book edition among others. The words &XX01 8i fiXXws may very well refer to the nine-book edition mentioned by Diodorus. Wilamowitz 3 doubts that there ever was a nine-book division. He imagines that there were two editions in which the books were numbered by letters, A to (8) and A to N (13). In transmission N became H. Then someone supposing that they were numbered by figures reported an eight- and a nine-book edition, a-77' (8), a'-d' (9). Wilamowitz is not so much interested in discrediting the nine-book tradition as in gaining a new witness for the thirteen-book edition. Conradt 4 rejects this theory and advances the following: One eight-book edition was marked by letters a-6, another by figures a'-rj'. Some- one supposing that both were numbered by figures reported an eight- and a nine-book edition. Both of these highly ingenious but utterly futile theories may be dismissed at once. They prove nothing. In view of Diodorus' two statements and the implication of Marcellinus' aXKot, 8e aWus the silence of the scholiast, who mentions only the thirteen-book edition, does not justify any doubt of the existence of a nine-book edition. Nothing is more natural than that some editor 1 xii. 37 ; cf . xiii. 42. There is evidence of a nine-book edition of Xenophon's Hellenica (Birt, Das antike Buchwesen, p. 448) . 2 Vita Thucydidis 58. Urriov di art. t^v rpayitarelav airrov ol flip Karkre/xov els Sitca rptis Urroplas, SXXot Si &XXc«. Sfiois Si i) ifhtiorn (or PeXrUmj) Kal 1} xoivij Kc/ipdrij/ce, to fiixpi tOd inri) &iypijriv eis rpurKaiSeKa, epravBa t6 t4Xos rfjs tpitjjs tcTTOpios &purav rat ipxh" tt}s rer&pTiis (ii. 78) . 4 tS)V (U t.y' reXos ttjs «', hp%i) rrjs C' (iii- 116). 5 tS>v tU TpuTKaldeica riXos tijs &tjjs, hpxh T V' e|386/«)S (iv. 78) . 6 Op. tit., pp. 3ff.; cf. Festa, "Sulla publicazione della storia di Tucidide," Bassegna Italiana, I (1918), pp. 1 ff. The Book Divisions of Thucydides 75 reconstructions that have been proposed are open to three objections: (1) They reject without adequate reasons part of the evidence of the scholiast. (2) The books are of disproportionate length. (3) They usually seek to apply some definite principle in the distribution of the subject-matter to the exclusion of other considerations. It is true that the scholiast is an indifferent authority in matters requiring critical scholarship; but in matters of fact, such as the point at which one of the thirteen books began, anyone who could read and write could observe and report the fact correctly. Such notes are quite different from serious speculations as to why Thucydi- des, whose name begins with 0, wrote ?/' books, while Herodotus, whose name began with t\, wrote 0' books, airopovo-'i Tivts ipvxpb-p awopiav, 5ia ri Oouki'SiStjj, ix av t6 bvofia airb rod 0, r{ crvveypaif/aro Icrroplas, 6 Se 'HpoSoros, ex^f airb rod rj, $knre . &v i/ Xutris \pv\po- ripa, 6n eireidr] <$>a, airb Se rod &os rod o 6/ctw drav0a rod bvbuaros ypafj,p,a aptfuriaav- rej icai rb 'iartpov (iv. 135). Similarly one might reasonably reject the statement that Thucydides did not divide his material into books but composed it as a unit. 6 QovKv5ld7}s ov SieTXep els Icroplas, aKka p,lav o-vveypayparo (iv. 135). But when in a note on ii. 78 it is said that the fourth book of the thirteen-book division began at this point one is scarcely justified in assuming that the note belongs to ii. 71, where a campaign ends. Precisely the same situation occurs at the end of the sixth book. The campaign of the year 415 ends at chapter 93, but the book division comes at the end of chapter 115. Had the situation been reversed and had a scholiast on a thirteen-book MS observed that the sixth book of a lost eight-book division ended at chapter 115 the modern scientific restorer of the eight-book edition would inevitably have disregarded the evidence and shifted the division point back to chapter 94 or forward to vii. 18. These con- siderations show that it is unsafe to reject the scholiast regarding a matter entirely within his knowledge and entirely beyond ours. While the book divisions of ancient authors are not of uniform length there is an approximation of uniformity. In Thucydides the average book length is 74 pages, the maximum 87, and the minimum 65. Kalinka, one of the more recent restorers of the thirteen-book scheme, has quite properly taken account of this factor. Starting 76 Robert J. Bonner with the principle that the end of a book should normally fall at the end of a year, he splits Book i, assigning five books to the Ten Years' War, one to the period of truce, and five to the remaining five years of the war. The result is a marked disparity of book lengths. The maximum is 66 pages and the minimum 19. Con- sequently he modifies 1 the scheme so as to increase the minimum to 36. Curiously enough, while refusing to credit the scholiast who reported the fourth book as ending at ii. 78, and shifting it back to chapter 71, where the year ends, he himself in four instances fails to make the book divisions correspond with campaigns. 2 In view of these inconsistencies Kalinka cannot expect his criticism of the scholiast to be convincing. It is clear that no mechanical scheme of book division will produce anything like the uniformity that is found in classical authors in general. The eight-book division of Thucydides is not consistently based on any single plan. Why should we expect to find one in the thirteen- book division? The introduction is contained in the first book. Books ii, iii, and iv each contain an account of three years; but the tenth year of the Archidamian War, which is a distinct unit, is combined in Book v with an account of the events of the period of the truce. The division between Books vi and vii does not occur at vi. 94, where the campaign of 415 ended. The editor preferred to break into the year 415 because, perhaps, as the scholiast remarks, ivTavda. 17 t&v 'Evpa.KovvUav &pxercu vint] nal t&v 'Adtjvaiuv ^rra. Similarly the account of the Sicilian War (vi and vii) neither begins nor ends with a year. 1 Zu Thukydides, Festschrift fur Gomperz., pp. 109 ff. PAGES PAGES l:\ i. Introduction 100 J 8. v. 25 — vi. 7, five years 50 9. vi. 8-93, one year 53 i 3. ii. 1-70, two years 44 10. vi. 94— vii. 18, one year 19 4. ii. 71 — iii. 25, two years 39 11. vii. 19 — viii. 6, one year 58* 5. iii. 26-116, two years 53 12. viii. 7-60, one year 34 6. iv. 1-116, two years 66| 13. viii. 61-109, one year 36 7. iv. 117 — v. 24, two years 28 Modifications: 6. iv. 1-77, 7. iv. 78— v. 24, 10. vi. 94— vii. 41, 11. vii. 41 — viii. 6, PAGES 44 i 49 36i 41 2 Books vi, vii, x, xi. Of ten books dealing with the war four depart from the scheme with which he begins. The Book Divisions of Thtjcydides 77 The following reconstruction of the thirteen-book edition gives full effect to the evidence of the scholiast and follows closely the divisions of the eight-book edition. PAGES 1 PAGES 1. i. 1-65 38 8. v. 26-116 48 2. i. 66-146 50 9. vi. 1-93 60 3. ii. 1-78 50 10. vi. 94— vii.41 38 4. ii. 79— III. 25 36 11. vii. 42-87 38 5. iii. 26-116 56 12. viii. 1-60 40 6. iv. 1-77 48 13. viii. 61-109 37 7. iv. 78— v. 25 53 The first book is divided at the point where military operations cease and the diplomatic history begins. 2 The unity of the Archida- mian War is preserved better than in the eight-book edition, though in two instances the division points do not occur at the end of a year. The eighth book, opening with the so-called second preface, covers the period of the truce. The unity of the Sicilian expedition as it appears in the eight-book edition is preserved in three books. 3 The ninth book ends with a year at a point where Wilamowitz found indications of a former book division in Vaticanus B. The eleventh book begins with the arrival of Demosthenes in Sicily. The twelfth and thirteenth correspond to the eighth. The division point occurs at the end of a year. The maximum number of pages per book is 65, the minimum 36, the average 45. Festa advances the novel theory that Thucydides himself began to publish his history in instalments after the conclusion of the war. 4 The basis of this view is the appearance of his name twelve times in the stereotyped expression with which he ends each one of the 1 Oxford text. 2 Kalinka does not specify a division point. Festa (op. cit., p. 7) would divide at i. 89, the beginning of the irtrrriKovTatTia. 3 A shift involving a few chapters at the beginning of Books vi and viii (vi. 8 instead of vi. 1 and viii. 7 instead of vii. 87) would bring the Sicilian expedition within the limits of two years. But why should we suppose that the editor of the thirteen- book edition disregarded the literary and historical unity of the story of the Sicilian war for the sake of a mechanical scheme which even the most zealous of its advocates admit cannot be consistently followed ? l Op. cit., p. 9: "Tucidide publicd di mano in mano le varie parti dell'opera. Certo egli attese la fine della guerra intera, prima di accingersi alia composizione vera e propria." 78 Robert J. Bonner twenty-one years covered by his history. 1 Festa argues that the words &v QovKvdidris t-vveypaxlsev in the formula like 6 Selva £woir)ii. 70; ii. 103; iii. 25; iii. 88; iii. 116; iv. 51; iv. 135; vi. 7; vi. 93; vii. 18; viii. 6; viii. 60; e.g., htiirtpov iros kreXeira t t4>8« iv 0ovKv5i5i]s £vviypatl/ei>. 2 " II valore preciso di questa formula va messo in chiaro, e non bisogna confondere con essa una dichiarazione come yiypafa St (ecu toOto 6 airrds ©oukuSiStjs. Quest' ultima riguarda un dato di fatto, strettamente connesso con l'espozione dei criteri dell'autore quanto alia materia e ai fini dell'opera. Nell'altra formula, invece, l'aorista dice chiaramente che abbiamo da fare con una dichiarazione di propriety letteraria. Se perd teniamo conto che il primo volume non aveva bisogno di quella dichiarazione, perche portava scritto da principio BovKv5i6r)s 'Afltjcolos ivviypa^e," etc. The dis- tinction is not convincing. 3 The books in Festa's scheme end as follows: (1) i. 88; (2) i. 146; (3) ii. 70; (4) iii. 25; (5) iii. 116; (6) iv. 51; (7) iv. 135; (8) vi. 7; (9) vi. 93; (10) vii. 18; (11) viii. 6; (12) viii. 60; (13) viii. 109. It will be observed that neither this scheme nor the Thucydidean puts the so-called second preface at the beginning of a book. * iii. 116. The Book Divisions of Thucydides 79 The appearance of an author's name twelve times in the course of a single work cannot be without significance. One is inevitably- reminded of the words of Theognis (19-20) : Kvpve, i.£onev<0 p.ev ep.oi ptjyls eiriKeicrdo) Totab' hretnv, X^priyis why was it not added to the formula at the end of every year t l For example, the name does not appear at all in the six-year endings included in the fifth book. These omissions can scarcely have been accidental. One very obvious weakness in Festa's theory may be indicated. The first occurrence of the tag is at ii. 70. And so the first Thucydidean book should include the introduction (Book i) and the story of the first two years of the war. But Festa, feeling that a book of 135 pages would be inordinately long compared with the subsequent books, seizes upon the introductory words QovKv8i8t]s ' KBrjvatos k.tX. as a substitute for the formula. This procedure is open to serious objection. It involves the awkward necessity of explaining why the sentence yiypae 81 /cat ravra 6 a&rds QovkvSiStis, v. 26, is not also a tag marking off a book. He cannot admit that it is, for he would then have two tags in one book or be obliged to select a division point somewhere between v. 26 and vi. 7. In this section there is no natural division point such as there is at the end of the first book, which constitutes an introduction easily separated from the rest of the work. In the first case there would be fourteen published sections and fifteen in the second. In either case the connection, psychological or sentimental, between the Thucydidean edition and the thirteen- book edition disappears. It is much simpler to end the first book at ii. 70. Festa is unduly influenced by the disproportionate length of the section. An author publishing serially would not attach so much importance to uni- formity of sections as an author dividing his entire work before publication as a whole. A further advantage of this modification is that the origin of the thirteen-book edition and its priority over the eight-book edition are established. The editor of Thucydides having before him thirteen sections, twelve published and one 1 The formula with slight variations appears twenty-one times. The name is omitted nine times. 80 ROBEBT J. BONNBK unpublished, might very well divide the whole work into thirteen books even if he felt obliged to rearrange them entirely. Further- more, an author publishing the first instalment of his history of a war would, if he desired to engage public interest, most naturally include the history of one or two campaigns. Thus modified Festa's theory becomes extremely plausible, if not convincing. The bearing of this theory on the much-vexed question of the composition (Entstehung) of Thucydides' history 1 would seem to be important, but Festa has made no attempt to exploit it in this direction. He offers no argument to support his statement that Thucydides did not publish until after the conclusion of the entire war. But this statement cannot be disproved. There are every- where throughout the whole work indications that the author knew the conclusion of the war. There are also passages that seem to have been written without such knowledge. This is to be expected in a history covering so many years. Beginning as he did to compose his history as soon as the war began it is natural that he should have written considerable portions before he knew the outcome of the war. The presence in these sections of references to the later phases of the war may be due to revision. Those who argue that Thucydides published the history of the Ten Years' War before the fall of Athens assume that an editor made the revision. The question regarding the composition of the history was first suggested by the so-called second preface. 2 Under the impression that the war was over in 421 he published an account of the Ten Years' War. Upon the resumption of hostilities after the truce he continued his work, prefacing it with a statement of his views as to the unity of the war. The difficulty of this theory is that the history of the ten years which we have could not have been published until the war was concluded. 3 Consequently it is assumed that it was revised and republished by Thucydides or by his literary executor. According to Festa's scheme this chapter came, not at the begin- ning of an instalment, but in the heart of it. It is not a preface but 1 Grundy, Thucydides and the History of His Age, pp. 387 ff. In an appendix Grundy has summarized the various views on this perplexing problem. 2 v. 26. • A number of passages show that the author knew how the war ended; cf. ii. 65. 12. The Book Divisions of Thtjcydides 81 an explanation of his views which were at variance with contemporary opinion. 1 This statement appears at the weakest point, the begin- ning of the story of the vtottos avonuxh- By combining in one instalment the story of this period and the last year of the Archi- damian War he emphasizes the unity of the war in a striking manner. In effect Thucydides says in this statement: "I am the man who started out to write the history of the war between the Peloponnesians and the Athenians. In my opinion the war did not end with the peace of Nicias. For during the six years of doubtful truce, although they did not invade each other's territory, they failed to carry out the terms and did each other all the harm they could. Therefore I am going right on to give an account of these years as part of the task I originally set myself." This is not the sort of statement a man would make on renewing a task which he imagined he had completed. If there was so sharp a distinction between the history of the Ten Years' War and the rest of the struggle one would expect to find some trace of it in the book divisions. There is none in the eight- book edition, where the second preface appears in the heart of the fifth book. This may not have much significance if the editor was mainly concerned with obtaining fairly uniform book divisions, for Book v without the chapters devoted to the tenth year of the war would be reduced to 48 Oxford pages as compared with an average of 82 if the history of the ten years was distributed over three books. Festa keeps the fifth book intact in his reconstruction of the thirteen- book division. Kalinka and the others begin a book at v.26. From a mechanical point of view this is a natural division point, for although the fifth book is the shortest in the eight-book edition it would be considerably longer than the others in a thirteen-book edition. By a slight variation in the reconstruction I have suggested above the eighth book could be increased so as to include our fifth book. If the eight-book edition grew out of the earlier thirteen-book edition the "second preface" made no impression on the editor responsible for it. In another direction Festa's theory is in conflict with views that are widely held. Many scholars have found evidences of incomplete- ness, particularly in the fifth and eighth books. In the latter book 1 Grundy, op. cit., pp. 391 S. 82 Robert J. Bonner are found " contradictions as to matters of fact as well as differences in tendency." Regarding the treaty quoted in Doric in the fifth book Bury 1 remarks: "It was not in accordance with the artistic method of Thucydides, or of ancient historians in general, to introduce into the narrative matter heterogeneous in style; and it is almost incredible that he would have admitted texts not written in Attic Greek." All the probabilities support this view but they do not amount to proof. His practice in reporting Doric speeches in Attic Greek is quite another matter, for he makes no attempt to report speeches verbatim. Nothing short of the appearance else- where of an Attic version of a Doric official document made by Thucydides could justify the unqualified assertion that the fifth book is in an unfinished state and could not have been published by Thucydides. University of Chicago 1 The Ancient Greek Historians; cf. Bonner, "The Mutual Intelligibility of Greek Dialects," Classical Journal, IV, 361.