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. TWO TEXTUAL PROBLEMS IN THE DIALOGUS OF TACITUS By Alfred Gudeman I. THE EXTENT OP THE TWO LACUNAE Prior to the year 1901, when Sabbadini 1 discovered in an Ambro- sian MS some references which Pier Candido Decembrio (1399-1477) jotted down in his diary, our knowledge of the existence of an exten- sive lacuna after c. 35 (ad veros indices ventum) was entirely based upon the testimony of the following marginal notes in our MSS: A: hie desunt sex pagelle. B: deerant in exemplari sex pageUae vetustate consumptae. E: hie deest multum. in exemplari dicitur deesse sex paginas. A: hie est defectus uni' folii cum dimidio. V: hie est defectus unius folii cum dimidio. Ven. : hie deficiunt quattuor parvae pagellae. C (or C 2 ) : multum deficit in exemplaribus quae reperiuntur. In D only 12| lines are left vacant and in H a space equivalent to nearly ^ of the treatise was apparently reserved for the subsequent discovery of the lost portions. The printed editions finally, up to that of Beroaldus, give a continuous text, without the slightest intimation of a lacuna. The identical statements of ABE are, of course, not so many independent witnesses, but they were all derived ultimately from a common source, the codex Hersfeldensis, which, we now know, was brought to Italy by Henoch of Ascoli about 1455, 2 16 pages of the original being still extant in the cod. Aesinus, 3 dis- covered in Iesi in 1902. B's specific information, if it was not added conjecturally by Pontanus, from whose copy B is directly derived, may reproduce the marginal note of the archetype exactly, but whether it was this codex or its parent MS which preserved the evidence justifying the number six is no longer ascertainable. The determination of the extent of this large gap has, of course, considerably exercised the ingenuity of critics for a long time, but as ' Riv. di fit- class. XXIX, 262; Scoperti 141, 166. Cf. also Wissowa, Preface to the Sijthoff facsimile of the cod. Leidensis (B), pp. v-vii (1907). » Decembrio, as he tells us himself, saw the MS from which he drew his information in Rome in that year. » See Annibaldi's admirable edition of this MS and Wissowa, op. tit. [Classical Philology VII, October, 1912] 412 Textual Problems in "Dialogus" of Tacitus 413 the exact number of the pages of the archetype was unknown, all efforts in the above direction were necessarily as fanciful, as were the guesses at the content. All that is quite certain is, that the causes which made Greek oratory what it was were dealt with under the same point of view as in the discussion concerning Roman eloquence which follows the gap, this being made evident by the words "eadem ratio in nostra quoque civitate antiquorum eloquentiam provexit." Nevertheless, Brotier, whose slumbers were doubtless disturbed by the laurels of Freinsheim's Livy, attempted a restoration of the lost text, ludicrous both in style and substance, and Habbe, on the basis of wholly unwarranted premises, fixed the extent of the gap at \ of the entire treatise, others indulging in still other conjectures. (For the details, see my Prolegomena to the Dialogus lxxxiv ff .) In this perplexity the note of Decembrio on the Dialogus unexpectedly came to our aid. It reads as follows: "Cornelii Taciti dialogus de oratoribus. Incipit 'Sepe — retineat.' Opus foliorum XII II in columellis. Post hec deficiunt sex folia, nam finit 'quam ingentibus verbis prosequuntur. Cum ad veros iudices ventum.' Deinde sequitur: 'rem cogitare nihil abiectum, nihil humile.' post hec sequuntur folia duo cum dimidio et finit: 'Cum adrisissent disces- simus.' " As the MS from which Decembrio gleaned his information was written in double columns, a practice which had gone out of vogue in the fifteenth century, Sabbadini justly concluded that Decembrio'e MS was not an apographon of the Hersfeldensis but the original? and he further calculated the extent of the lacuna as amounting to f of the entire treatise 8 (14+2| = 16:6). So far as I am aware, this calculation has been universally accepted; at all events, it has nowhere been called into question, and yet the surprisingly huge extent of the lacuna ought alone to have aroused suspicion. For as Messalla had only reluctantly, on the urgent request of Maternus, consented to continue the discussion along the lines designated by the same Maternus, and as the history of Greek eloquence, serving merely as an analogous illustration, cannot possibly have been 1 It was doubtless merely for the sake of conformity that the additions to the original 16 pages in the cod. Aesinus were also written in double columns. 1 Op. tit.: "La lacuna comprende circa i I dell' intero (italics are mine) libro: aun di presso quattrocento righe di una ordinaria edizione di Lipsia." 414 Alfred Gudeman accorded a prominence disproportionate to the main design of the Dialogus, the problem as to what else could have been treated at considerable length, would be rendered still more perplexing than it had previously been. Happily for our peace of mind it turns out that Sabbadini's arithmetic was sadly out of joint, for the wording of Decembrio leaves no doubt whatever that the text of the Dialogus in the codex Hersfeldensis contained 14+6+2J=22| folia =45 paginae in all. The loss (6 folia) sustained constituted, therefore, only between f and \ and not f of the entire treatise, more exactly 12: 45 =.266 per cent. This gap is still sufficiently extensive to have included the close of Messalla's speech and the discussion of Greek oratory. At the same time it effectively refutes the opinion of those who, though they also assumed a second lacuna after c. 40, 6 admovebant, of which more anon, nevertheless assigned all the intervening portion to Messalla, to the exclusion of Secundus. Right here, however, a difficulty, which even Wissowa 1 overlooked, arises in connection with the testimony of the MSS cited above. Decembrio, it will be recalled, speaks of folia; the MSS ABE, however, mention only six paginae or pagellae! There is no evidence, so far as I have been able to ascertain, that these terms were ever employed as synonyms in the fifteenth century or earlier and, since Decembrio uses folia consistently and repeatedly in its proper sense, there is but one way out of the dilemma, and that is, to assume that the common source of ABE had inadver- tently substituted the possibly more familiar term pagina or pagella for folium. This explanation is all the more plausible, since even modern critics, like Habbe and Peterson, were guilty of the same error or, like Sabbadini, Annibaldi, Wissowa, and others, failed to notice the glaring discrepancy between the statements of the MSS and Decembrio now existing. While the unimpeachable testimony of Decembrio, based as it is on documentary evidence, enabled us to determine once for all the exact extent of a lacuna indicated in our MSS, thus disposing of a long-standing problem which would have remained insoluble without the accession of new material, the codex Aesinus, already referred to, renders still more valuable service in another direction. 1 Op. tit., vii. Textual Problems in "Dialogus" op Tacitus 415 Students of the Dialogus text are familiar with the equally time- honored controversy regarding the justification of assuming a second lacuna at c. 40, 6, after the word admovebant, albeit there is in this case no indication in our MSS of a gap at this point. Heumann (1719), who was the first to notice the "hiatus," did not pursue the thought any further, going quite astray at that, in attributing the preceding portion to Messalla. As a matter of fact, internal reasons of irrefutable validity make it impossible to assign cc. 36-42 (c. 42 finierat Maternus) to one and the same speaker. 1 That the failure of the MSS to indicate a lacuna does not militate against such a supposition, provided the character of the context imperatively demands it, is well known 2 and in the present instance certainly Bentley's famous dictum "ratio et res ipsa centum codicibus potiores sunt" can be accepted without compunctions of conscience. As, however, the absence of external evidence still remained the one apparently insuperable obstacle to a more general acquiescence in the assumption in question, it was a consummation devoutly to be wished that something might turn up which would explain how a lacuna could easily have arisen at c. 40, 6 without leaving any trace of its existence. This highly welcome evidence is now furnished by the codex Aesinus. By its aid it can be demonstrated with all but mathematical precision, that cc. 36,1 — 40, 6 took up exactly 4 pages or two folia of the archetype, which implies that faces admovebant formed the last words, and non de otiosa the beginning of a page. Now nothing is a priori more natural than that a MS gap would be overlooked if an entire leaf had become detached, and this would be all but inevitable if the words preceding the lacuna, as happens to be the case here, also closed a sentence. If it be added, that the majority of modern critics have either failed or refused to recognize a lacuna, we can certainly not be surprised, that a mediaeval scribe did not divine its existence. To suppose, on the other hand, that a leaf was accidentally lost at the precise juncture, where the strongest 1 See the full discussion of the subject in my Proleg. Ixxv-lxxxii. Dienel, in his frantic efforts to maintain the identity of the speaker, takes refuge in the paradoxical hypothesis that Maternus-Tacitus intentionally indulged in a series of antinomies (alias irreconcilable contradictions) . Not to mention that on this theory the otiose repetitions would still defy explanation, it is quite inconceivable what subtle purpose this early predecessor of Kant desired to subserve by so antithetical a process. 2 See also Proleg. lxxviii f . 416 Alfred Gtjdeman of internal reasons pointed to an interruption of the context, would be a coincidence too marvelous for credence. It remains for me to show how the surprising result stated above was reached. The codex Aesinus has preserved, as already remarked, 8 folia or 16 pages of the archetype. They cover c. 13, 2 munia — c. 40, 6 missum, of the Agricola, being written in double columns. The lines, of which there are thirty to the page, are of irregular length, very brief lines usually marking the end of a chapter, though they occa- sionally also occur elsewhere. The chapters themselves, moreover, are more numerous than those of our modern printed text (26 as against 36, with 33 short lines). The Dialogus and the Suetonius fragment and cc. 1-13, 1 ; 40, 7 to the end of the Agricola had become detached from the archetype shortly after the time of Decembrio, who still had the intact MS before him. The Germania and the missing portions of the Agricola were, indeed, subsequently copied from it by another scribe, but the Dialogus and Suetonius fragment were not also added, possibly because these had already been sold or otherwise disposed of and then lost. Such being the history of the composition of the Aesinus and its archetype, so far as it concerns us here, it is self-evident that the still existing portions of the archetype may be taken as accurately representing the character of the Dialogus text as well, including all the peculiarities mentioned above. Now the length of a column line in the original text of the Agricola varies between 4.4 and 4.85 or of a double line between 8.8 and 9.7 cm., the grand average being 9.4 cm. An average page, therefore, contained 282 cm. (9.4X30). On the other hand, a full line of a Teubner text measures 8.5 cm. The entire length of the printed Teubner text, from cc. 36 — 40,6 admovebant, is 1025.4cm., but as 1 cm. of this text is equivalent to 1.1 cm. of the archetype, the same amount of text covered 1127 .94 cm. in the MS. Dividing this total by 282, the number of cm. to a page, we find that cc. 36 — 40, 6 took up 8.999, or exactly four com- plete pages in the codex Hersfeldensis. The quite insignificant frac- tional difference — it actually amounts to less than a single letter! — is amply accounted for by the aforesaid inequality of the lines in the archetype, which in the present instance, of course, eludes an even approximate calculation. Textual Problems in "Dialogus" of Tacitus 417 ii. the validity of the x family v. y The apparently trivial notes in Decembrio's diary not only give us exact information regarding the space occupied by the Germania, the Agricola, Dialogus, and the Suetonius fragment in the archetype of all our MSS, but they are also of no little value in deciding, what has hitherto been altogether overlooked, the vexed question as regards the respective authority of the X class of MSS now repre- sented by A ( Vaticanus 1862) and B (Leidensis) and the so-called Y class which is generally made to include all the other MSS extant. Michaelis, as is well known, reached the conclusion that the former has preserved the better tradition, and his recension of the text was consistently constructed on that basis. Later editors did not, how- ever, follow in his footsteps, but have advocated with more or less fervor an eclectic method. The exhaustive discussion of the subject in my Proleg. cxxix-cxxxviii, concludes with these words: "The Y family, albeit it has suffered at the hands of learned and unlearned scribes, represents as a body the purer tradition A does, indeed, faithfully reproduce its archetypon. But that archetypon must yield the palm to Y, which has proved itself to be more accurate in preserving what Tacitus must have written." This verdict has not been overthrown; on the contrary, it can now be shown, what no one could have suspected before the discovery of Decembrio's notes, that the scribe of the lost MS X did not reproduce the codex Hersfeldensis as faithfully or as accurately as did Y, the parent of the other extant MSS. The 16 pages of the Hersfeldensis still extant would unquestionably have furnished superabundant proof of this; unfortunately, however, neither A nor B contains the Agricola. We are therefore confined to the few original passages which Decembrio copied from the archetype merely for the purpose, be it observed, of precisely marking the beginning and end of the treatises and in the Dialogus also the beginning and end of the lacuna. Happily these passages exhibit a number of most significant variants, which suffice to establish the statement made above. D. 35: rem cogitare nihil abiectum, nihil humile: Decembrio ex archetypo rem cogitant nihil humile vel abiectum: AB rem cogitare nihil h. nihil a.: o> 1 1 (a = reliqui vel omnes codices. 418 Alfred Gudeman That Decembrio found cogitare in his MS is made evident by the fact that he mistook nihil abiectum for its object, and in consequence omitted the words following, eloqui poterat. That the effective anaphora represents the genuine reading is rendered certain by a Ciceronian parallel, De fin. 5.57: "nihil abiectum, nihil humile cogitant," where, it will be noticed, not only the adjectives are in the same order as given by Decembrio, but we also have cogitare as their governing verb. This remarkable coincidence might at first glance suggest some causal relation between the two passages as the most plausible explanation of Decembrio's reading. But such an infer- ence is open to serious objections. For what we know of the work of this Italian humanist makes it extremely improbable that he was so profoundly conversant with Cicero's voluminous dialogue as to recall a phrase not intrinsically noteworthy, and that this untimely reminiscence then prompted him to make an arbitrary change of a passage which he cited, as he did eight others, for the purely technical reason pointed out above. It remains to suggest some plausible explanation for the unanimous reading of all our MSS. I can account for the transposition variant in no other way than by sup- posing, that nihil abiectum or nihil humile, having been accidentally omitted, owing to the anaphora, a frequent source of error, was subsequently placed above the line, XY inserting the words in one place, Decembrio in another, the archetype furnishing no clue to the original collocation. The habit of Tacitus in the Dialogus of inverting the order of synonyms in Cicero — my commentary fur- nishes very abundant illustrations of this practice — would to a certain extent favor the order given in the MSS. But be this as it may, it does not help the case of X, for the substitution of vel, though easily misread for nil, certainly introduces an inferior reading, while cogi- tant, found neither in Decembrio nor in Y, looks suspiciously like a deliberate attempt to round out the sentence syntactically. X fares no better in the following instances, where Decembrio, in agreement with Y, has alone preserved the genuine text. Sueton., p. 126 Rf . : Pisonem proconsukm: Dec. o>, personalem: AB Sueton., p. 127: excanduisset ut: Dec. w, excanduisse et ut: AB corr. B 2 (excanduisse ut et: C, excanduisse: G) The wrong reading of AB is here shared by two inferior members of the Y class. Textual Problems in "Dialogus" of Tacitus 419 While the genesis of the above variants is easily understood, they prove, in spite of their apparent insignificance, the untrustworthiness of X, the last available example pointing still more emphatically to the same conclusion, for here the paleographical origin of the cor- ruption of X is more difficult to explain, the resulting unintelligible reading being far removed from the genuine text: Sueton., p. 127: conspectu: Dec. ; the confusion between I and r is frequent) ac liber- tatis auctorem ac (Dec; et, a>) vindicem. Munich