Discrepancy by Design in Virg., Aen. VI 562-600 [Discrepancia ex profeso en Virg., Aen. VI 562-600]

As the context of Virg., Aen. VI 565 requires, the ancient interpretation of this line is correct, while the current consensus is mistaken. The Sibyl has never seen Tartarus. Subsequently, when she repeatedly asserts that she has witnessed the punishment of criminals confined in Tartarus, this blatant contradiction imitates the famously contradictory location of Odysseus outside Hades initially but later firmly within it.

In his learned new book on inconsistency in Roman epic, James O'Hara asserts that «uncertainties» in Odyssey XI also «mark Virgil's picture of the underworld, perhaps in part in imitation of Homer» 1 . In line with O'Hara's suggestion of a Homeric model for some of the «uncertainties» in Virgil's Underworld, this note will first correct a longstanding misinterpretation of Aen. VI 565. Then, it will argue that the internal discrepancy created by this revised interpretation finds a Homeric model, heretofore unnoticed, in a corresponding discrepancy in Odyssey XI. Emerita LXXX 1, 2012, pp. 171-178 ISSN 0013-6662 doi: 10.3989/emerita.2012.10.1104 As Aeneas and the Sibyl approach the gate of Tartarus, the Sibyl explains as follows why she and her companion are barred from entering that realm, and how she has come to know about it (Aen. VI 563-565):

N O T A S E I N F O R M A C I Ó N
Nulli fas casto sceleratum insistere limen, Sed me cum lucis Hecate praefecit Auernis, Ipsa deum poenas docuit perque omnia duxit.
The sequence of thought in these lines is altogether clear. The Sibyl, being casta beyond any doubt 2 , has never actually seen Tartarus. Rather, Hecate has personally told her about the punishments that the gods inflict there (ipsa deum poenas docuit), and she has omitted no detail (perque omnia duxit). In VI 565 in particular, the second clause re-states the first, and the line as a whole, consisting of two clauses that are syntactically parallel and similar semantically, exhibits the «theme and variation» structure of which Virgil is fond 3 .
The inclination to discount the testimony of Tiberius Claudius Donatus is widespread but unwarranted a priori 4 . Donatus' mother tongue was also Virgil's; Donatus, as he states in his epilogue, followed an exegetical tradition dating back to the immediate aftermath of the publication of the Aeneid; and Donatus was separated from Virgil by roughly the same span of years as separates us from Leonardo da Vinci and Christopher Columbus. These facts entitle Donatus' testimony to be taken at least as seriously as 2 If proof be required, the Sibyl has already been called casta in Aen. V 735. As Phoebi longaeua sacerdos (VI 628), she also belongs to the class of sacerdotes casti mentioned in VI 661.
3 See Henry 1873-1892, I, pp. 745-751 ad Aen. I 546-551. 4 Signs of a nascent reversal of this trend include Squillante Saccone 1985, wherein (pp. 9-10, n. 9) Aldo Setaioli is cited to the effect that Donatus solves some of the thorniest exegetical problems in Virgil, and that he possesses a sensitivity that is rare among ancient exegetes. Emerita LXXX 1, 2012, pp. 171-178 ISSN 0013-6662 doi: 10.3989/emerita.2012.10.1104 the exegesis of modern scholars whose mother tongue is not Latin, and whose separation in time from Virgil's day exceeds by a millennium and a half the number of years standing between Virgil and Donatus. If Servius was aware of one or more alternatives to the figurative sense of perque omnia duxit adopted in Donatus' paraphrase, he could be expected to have commented on these words. In fact, he did not. Roughly one millennium later, Donatus' reading remains unchallenged in the 17 th -century commentary of De la Cerda, who takes pains to elucidate how the Sibyl can know about Tartarus despite, as De la Cerda states explicitly, never having seen the place 5 .
If it is objected that the use of ducere per in a pedagogical sense is unprecedented and not securely attested until Quintilian 6 , a similar objection will also apply to arare 'wrinkle' in Aen. VII 417, to arua 'shore' in Aen. II 209, and to many other cases of a usage unprecedented before Virgil and sparsely attested thereafter. In all of these, context is decisive 7 . Just so does the context established by nulli fas casto in Aen. VI 563 require a figurative, pedagogical sense for ducere per two lines later.
In sum, in the context of Aen. VI 565, the figurative sense of perque omnia duxit is clear, and it was analyzed accordingly in late antiquity. Yet a consensus has now evolved around a quite different interpretation, the origin of which is difficult to estab lish. As early as François Catrou's translation of Virgil's works (1716), perque omnia duxit is suddenly rendered «m'a conduit dans tout le Tartare», as if Hecate once gave the Sibyl a guided tour of Tartarus, physically taking her there and showing her the sights 8 . In Heyne's edition, first published in 1767-1775, the ancient, figurative reading still survives, but now as an alternative to the new, literal sense 9 . Finally, by the end of the century, the literal interpretation seems to have taken over completely. In Nöhden and Heinrich's Erklärende Anmerkungen on the Aeneid, for example, published in 1794, perque omnia duxit is glossed tout court as follows: «Proserpina führte die Sibylla im Tartarus herum, und zeigte ihr alles» 10 . In the two centuries since, dissent from this view is nearly impossible to find 11 . It is difficult to understand how the current, literal interpretation of perque omnia duxit is even conceivable in the wake of an absolute and unconditional prohibition barring the blameless from ever setting foot in Tartarus. Has an exception somehow been made for Hecate and the Sibyl? There is nothing in the text to support this notion. Quite to the contrary, the Underworld is known for its strict enforcement of iron-clad laws: desine fata deum flecti sperare precando, says the Sibyl herself to Palinurus in Aen. VI 376. As a virgin goddess, Hecate ipso facto belongs among the casti no less than the Sibyl does 12 . Like the Sibyl, she too is vulnerable to the pollution emanating from Tartarus. In Aen. VI 563, the pronoun nulli is emphatic and subject to no exception either mortal or divine.
Eduard Norden in particular, acknowledging no other possibility, assumes a priori that the Sibyl has entered Tartarus and seen its denizens with her own eyes. Indeed, this premise leads Norden to a further assertion concerning Virgilian innovation: Is Norden's prestige a factor in the persistence of a literal sense being attached to perque omnia duxit? If it is, it is unlikely to be the only factor. In lines 563-565, by her own account, the Sibyl, being casta, is forbidden to enter Tartarus and has only learned of it secondhand from Hecate. Some seventeen lines later, however, with uidi in line 582, she abruptly-and, more to the point, inconsistently-claims to have seen firsthand the two Aloidae confined in Tartarus 15 . This claim of autopsy is repeated in line 585, where the same verb uidi is emphatically placed first and applied to the punishment of Salmoneus. Finally, eleven lines later, in Aen. VI 596, the Sibyl re-asserts for a third time her claim to have witnessed the punishment of sinners, in this case, of Tityus.
A likely reason now becomes clear for a critical consensus having evolved around taking perque omnia duxit literally. The Sibyl claims to have actually witnessed the torments in Tartarus. In order to reconcile this bald fact with perque omnia duxit, these words are taken to indicate, line 563 notwithstanding, that Hecate once guided the Sibyl through all of Tartarus 16 .
Unwilling to tolerate inconsistency in so fastidious a poet as Virgil, scholars have imposed on perque omnia duxit a meaning that its context does not allow it to have. This retrospective approach also runs counter to the actual experience of readers and of listeners alike. We retain what we have just read or heard, but we do not know what lies ahead, much less seventeen lines ahead. A reader or a listener will understand line 565 in a manner consistent with line 563, and this understanding will not necessarily be congruent with lines 582, 585, or 596 farther on 17 . 14 O'Hara 2007, p. 92. Even in a paper that insists on the Sibyl's ambiguousness and minutely documents various aspects thereof, the usual, quite unambiguous sense is attached to perque omnia duxit. «The Sibyl condenses the comprehensive tour she once had from her mentor Hecate» (Gowers 2005, p. 177). I have found no dissent from the communis opinio earlier than an obiter dictum in Fratantuono 2007, p. 182. In «the Sibyl gives him a summary of what she learned long before from Hecate herself», «learned» would seem to imply duxit taken figuratively. 15 For formulaic uidi, see La Penna 1987. 16 A connection between perque omnia duxit and lines 582, 585, and 596 must have been made as early as C. G. Heyne's first edition of 1767-1775. It is mentioned in the second edition, which is the earliest available to me. There, the implication of the aforementioned three lines for the sense of perque omnia duxit receives equal treatment with duxit taken figuratively and glossed as «narrando et percensendo singula». See Heyne 1787-1789 For a discussion of the same principle as it applies to other passages, see Weber 1998Weber -1999. Emerita LXXX 1, 2012, pp. 171-178 ISSN 0013-6662 doi: 10.3989/emerita.2012.10.1104 Far from seeking to eliminate the conflict between lines 563-565 on the one hand and lines 582, 585, and 596 on the other, one ought rather to look for a reason for this discrepancy. In fact, it is explained by O'Hara's dictum, cited above, that some of the «uncertainties» in the Virgilian Underworld are likely to be due to Homeric influence. In Od. XI 568-627, Odysseus, having heretofore stood at the threshold of Hades, suddenly finds himself within Hades itself. It is significant for the present discussion that this shift of Odysseus' location occurs precisely at that point in the narrative where the canonical sinners of the Underworld are introduced. Virgil was surely aware of this change in Odysseus' location, for it was a basis for Aristarchus' well-known rejection of Od. 11.568-627 as spurious. The Sibyl's location vis-à-vis Tartarus parallels that of Odysseus in Od. 11.568-627 18 . Like Odysseus, the Sibyl is not an on-site observer initially, but she abruptly becomes such with vidi in line 582, just as ἴδον in Homer likewise signals Odysseus' comparable change of location in Od. XI 568. Imitation of a familiar Homeric discrepancy would be especially appropriate here, where there is a further correspondence between the Sibyl's first-person narrative and that of Odysseus in Od. XI 568-627 19 .

N O T A S E I N F O R M A C I Ó N
This brief discussion has led to three main conclusions. First, for reasons that include, above all, the unconditional prohibition in Aen. VI 563, perque omnia duxit in VI 565 cannot be taken to mean that Hecate once guided the Sibyl through all of Tartarus. From this, a second conclusion follows: the Sibyl's claim of autopsy, thrice repeated in lines 582, 585, and 596, conflicts with what she has earlier said about being banned from Tartarus. Finally, not only is it mistaken to try to eliminate this conflict by retrospectively re-interpreting perque omnia duxit, but the discrepancy itself finds a raison d'être in Virgil's imitation of Homer-warts and all, as Fiachra Mac Góráin has described it per litteras. The Sibyl's change of location parallels the comparable shift in the location of Odysseus in Od. XI 568-627. The abode of canonical sinners described in the first person is the context common to both Virgil and Homer 20 . This parallelism 18 See Kirk 1962, pp. 236-237, andPage 1955, pp. 25-26. Aristarchus' misgivings are reflected in the question posed in two scholia on Od. XI 568: «how does he know that these people or the others are within the gates of Hades and the rivers?» For Virgil's familiarity with ancient scholarship on Homer, see Schmit-Neuerburg 1999, Lamberton 2001, Schlunk 1974, and Heinze 1915 It cannot be known, of course, to what extent Virgil's Odyssey differed from ours. Therefore, nothing can be made of the coincidence that Virgil's Tartarus episode and its Homeric analogue both end with line 627. The view advanced in Knauer 1964, p. 123, that both episodes are of comparable length (62 and 68 lines respectively), requires that the Virgilian scene not begin until Aen. VI 566. 20 For the role of the Sibyl's first-person narrative in recalling the Homeric realm of criminals, see Knauer 1964, p. 119, andNorden 1927, p. 359. The anaphora of uidi in Virgil finds a Homeric parallel in the fourfold repetition of ἴδον/εἶδον in Od. XI 568, 576, 582, 593. Emerita LXXX 1, 2012, pp. 171-178 ISSN 0013-6662 doi: 10.3989/emerita.2012.10.1104 of situation extends also to language, uidi = ἴδον signaling in each case the shift from outsider to appalled eyewitness. Equivalent verbs create equivalent discrepancies in the location of the first-person narrator. Ancient critics of the Aeneid were keenly aware of internal inconsistencies, real or imagined. Roughly sixty of these are noted in Servius' commentary, usually signaled by a telltale contrarium or atqui 21 . From the remarks of Tiberius Claudius Donatus transcribed above, it is clear that, as early as the fifth century, the Sibyl was understood to be barred from Tartarus, and thus that perque omnia duxit was taken fig uratively. The discrepancy thus created with uidi, and with cernere erat, some lines later could be expected to be noticed by critics who were habitually on the lookout for such things. It is remarkable, then, that in spite of his insistence that the Sibyl was never allowed inside Tartarus, her words indicating otherwise elicit no comment from Tiberius Claudius Donatus. This oversight is not unparalleled. In his commentary on Aen. IV 450, Servius Danielis elucidates an apparent contradiction with IV 696 but has nothing to say about this in his note on IV 696 itself. There, he is rather concerned with a different discrepancy, that between IV 696 and X 467.
With one exception, ancient analyses of Virgilian discrepancies also overlook Homeric analogues. When Cretheus is killed twice in Aen. IX 775 and XII 538-539, scholiasts fail to invoke such Homeric precedents as Pylaemenes, killed in battle in Il. V 596 but revived to mourn his son in XIII 658. Concerning the salient temporal contradiction between Aen. V 626 and I 797, nothing is said about chronological discrepancies in Homer. Yet Servius' note on Aen. VI 650 stands out as a unique exception to the general rule. There, noting an inconsistency between Dardanus in the Underworld and Dardanus in heaven, as he is said to be in Aen. VII 211, Servius asserts Homerum sequitur («he follows Homer»), and he cites as Virgil's model the dual incarnations of Homer's Heracles as both an Olympian and a chthonic simulacrum in Od. XI 601-604. In proposing a Homeric precedent for a discrepancy in Virgil, the present discussion itself finds a precedent in Servius.
In summary, this note proposes very little that is new. Indeed, the view that Aen. VI 563 excludes the Sibyl from Tartarus is by now, more or less, sixteen centuries old. When the Sibyl herself later contradicts this notion by claiming on-the-spot autopsy of criminals in Tartarus, it is argued above that this is modeled on a corresponding discrepancy in Odysseus' location in Odyssey XI. Only in revealing this «warts-andall» imitation of Homer does this note break any new ground 22 .