Virgil, Aeneid 2.

567-88
Author(s): R. G. Austin
Source: The Classical Quarterly , Nov., 1961, Vol. 11, No. 2 (Nov., 1961), pp. 185-198
Published by: Cambridge University Press on behalf of The Classical Association

Stable URL: https://www.jstor.org/stable/637710

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

The Classical Association and Cambridge University Press are collaborating with JSTOR to
digitize, preserve and extend access to The Classical Quarterly

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88

FEW critics can ever have shown more light-hearted thoughtlessness towards an
anxious posterity than Servius in his casual preservation of the 'Helen-episode',
lacking in our ancient manuscripts of Virgil and primarily extant only in this
precarious form. A pity that Servius spoke at all, if he could not tell us more;
and to make matters worse, he ignored the lines in his commentary. Aeliu
Donatus says nothing of them. Tiberius Claudius Donatus passes peacefully
in his interpretatio from 2. 566 to 2. 589; his prosy conscientiousness nowhere
else allows him to skip so much. The passage so rashly preserved forms an
exasperating Tummelplatz' for students of Virgil: 'qua de re viri docti iam pridem
inter se certarunt semperque, ni fallor, certabunt.'2 The purpose of this paper
is to suggest that Servius told the truth about the lines, and was not planting
a forgery on a credulous world.
The passage is quoted by Servius auctus on 2. 566, with the comment 'post
hunc versum hi versus fuerunt, qui a Tucca et Vario obliti sunt'.3 The Servian
Preface states, in its fuller version,4 'et in secundo libro aliquos versus posuerat
quos constat esse detractos, quos inveni<e)mus cum pervenerimus ad locum
de quo detracti sunt': this is in illustration of Augustus' instructions to th
executors hac lege emendare, ut superflua demerent, nihil adderent tamen; the half-lines
are an example of their obedience to the second of these injunctions, the ille ego
passage and 2. 567-88 were removed in obedience to the first. On 2. 592, after
explaining dextraque prehensum, the commentator adds: ut enim dictum est, versus
illos qui superius notati sunt, hinc constat esse sublatos, nec immerito.5 nam et turpe est
viro forti contra feminam irasci, et contrarium est Helenam in domo Priami fuisse il
rei, quae in sexto dicitur, quia in domo est inventa Deiphobi, postquam ex summa arc
vocaverat Graecos. hinc autem versus esse sublatos Veneris verba declarant dicentis 'no
tibi Tyndaridisfacies invisa Lacaenae'. Thus we are told that Virgil had originally
placed (posuerat) the passage where it is quoted, that it was common knowledg
(constat) that Varius and Tucca removed it, and that they were justified on
two grounds; further, that the excision is detectable from 1. 60oi. We are not
told that Virgil himself deleted the lines, nor that he had marked them i
anyway, as has sometimes been supposed:6 contrast 3. 204, where Servius
auctus quotes three uncanonical lines with the comment 'hi versus circumducti
('ringed') inventi dicuntur et extra paginam in mundo'.
I Cf. Norden, Aeneis VI, p. 262; Bichner,4 Ed. Harv., p. 2; Thilo-Hagen's presenta-
R.-E. viii A, col. I353. tion of the text is confusing (see Fraenkel,
2 Wiechmann, De Aeneidos libri II com- J.R.S. xxxviii [1948], 131, where 567-88 are
positione, progr. Potsdam, 1876, p. I7. referred to as 'the interpolation').
3 qui a Tucca et Vario Bergk; q Tucca et s So Servius auctus: the shorter version
Varius C (but Bergk emended obliti, un- runs 'ut enim in primo diximus, aliquos
necessarily, to sublati). It is amusing to hinc versus constat esse sublatos, nec
find Gruppe (Minos, p. i75) defending immerito' (the rest is common to both ver-
Virgil's executors from a charge of amnesia, sions).
and Hartmann (MVnemosyne N.s. xxxiii [I 905], 6 Cf. Kvivala, Neue Beitridge zur Erkldrung
444) happily accepting the possibility as der Aeneis (Prag, i881), pp. 33 f.; Pascal,
clearing up the whole problem. Birt (Kritik Graecia Capta (Firenze, I905), p. 121;
und Hermeneutik, p. 16o) apparently swallowed Sabbadini states roundly (in his text of I937)
the MS. reading ('Varius und Tucca haitten 'hos versus Vergilius ipse delevit'.
sie vergessen').

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 186 R. G. AUSTIN

Servius' grounds for the d

though he held the passag
carpendi poetae causa a sc
paraphrase of 583-4, and d
been argued, for example,
of killing Helen. As for th
could not see the contradict
that it mattered, for the m
pended on argument about
to display his subtle skill,
possible critics with such a
removed the lines, they ca
The poetic worth of the
been as hotly upheld as its
The passions aroused have b
The lines seemed to Coning
of Virgil', though 'disfigur
that no protest would have
passage' if it had been tran
Aeneid. Baehrenss found th
critica recte utare) per om
man ... ffir gutes Eigenthu
'lo squarcio poetico e, per
gilio, e Vergiliano ne e lo s
found in the passage 'etw
and thought the opening 's
geschrieben haben kann, n
all defence as showing igno
nique. Heinze'I commented:
Verst6sse, sondern 1isst di
Tritt vermissen.' Above all
574 und so trocken und feh
und sicher nicht in die Er
'die elende Sprache mit ih
Stils'. It is quite a relief to
nonnulli pulchri, ac Virgili

I Prolegomena, p. Forschungen
" Plautinische 92 ('propter
(2nd ed., 1912), t
causas'). p. 42, and see his even more acrimonious re-
2 e.g. by Gossrau (1846); Mancuso, marks in Der Monolog im Drama (1908), p.5, n. i.
Classici e Neolatini, 1911, pp. 3I ff.; Noack,Leo classes the passage with the false ending
Rheinisches Museum N.F. xlviii (1893), 420 ff.of the Andria, or with 'Lucili quam sis men-
3 Hartmann, 1.c., p. 441. dosus', etc. (Hor. Sat. i. io, lines which still
4 Classical Philology i (90o6), 222. have defenders: see Burck's supplement to
s Fleckeisens Jahrb. fiir classische Philologie, Kiessling-Heinze, 6th ed., p. 411).
xxxi (1885), 399; unfortunately Baehrens's 12 Peerlkamp had a simple solution: he
own ideas of how to rewrite Virgil scarcelyejected this passage, and 589-623, and vi.
recommend him as a critic. 493-547. He made an illuminating comment:
6 1.c., p. 431. 7 1.c., p. 27. 'ipsi recentiores Europaei, Vida, Fracas-
8 Op. cit., p. I80. torius, Lotichius, multique Neerlandi, ea
9 Aeneis VI, p. 262. saepe protulerunt quae Virgilius libens pro
0o Virgils epische Technik (3rd ed. 1928), suis agnosceret.' Libens, indeed!
p. 45-

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88 187

varied judgements have been passed on the relation

text: some critics hold that without it the mention
ligible, others that this line needs no previous p
claimed that the episode forms a landmark in Aene
great spiritual crisis, victory in which gave him st
trial in Carthage.'
The passage must stand or fall on its style, langua
else is relevant, if this cannot be found satisfactor
Norden was very rash in his application of metr
in 573, 580, and 587 there is elision of a spondaic w
claims to occur on an average once in 89 lines in no
it has been shown by Shipley3 and Johnson4 that h
average is no evidence against the passage: the m
occur in groups,5 and there are often long stretch
confined himself to spondaic words in this position
-m is included in the examination, the groupings c
(sulcum ac), 69 (vero et), G. 3- 433 (siccum et), 439 (s
214 (ipsae et), 400 (circum haec), 402 (herbae et), 426
tains also a grouping of a different pattern: 572 (D
588 (iactabam et) : this can be paralleled from this bo
484 Priami et, 495 Danai et), and elsewhere.7 Clearl
is on no firm ground.8 He questioned also the encli
584 (feminea in poena est), claiming that there is onl
Virgil (G. I. 83) for such an enclisis in this position
captious: Norden might just as well suspect his par
conclusive weight towards damnation.9
Let me now turn to the diction. It is unfortunate that the German critics
nowhere specify what they object to in the words that they stigmatize. Norden
and Heinze brand praemetuens (573). This verb occurs in Lucr. 3. 0Io9 (the
participle, also at the beginning of a line), Caes. B.G. 7. 49, Phaedr. I. 16. 4,
and nowhere else in classical Latin.'0 But praefodere (I 1.473), praenatare (6. 705),
praesciscere (G. 4. 70), all of which occur once only in Virgil, are equally rare:"
and here praemetuens is not only effective but psychologically right, marking
Helen's terrified anticipation of what she deserves.'2 I cannot see that its
I L. R. Palmer, Mnemosyne, 3rd series, vi lines, and immediately preceding the caesura
(1938), 368 ff. But he does not explain why in six-might seem suspicious. But in 6.
so cardinal a passage should have had so 212-31 (a highly finished passage, which
casual a history. even Peerlkamp did not eject) et occurs eight
2 Aeneis VI, Anhang xi. I. 3 (p. 454). times (and there is another in 211), in four
3 T.A.P.A. Ivi (1925), 172 ff., an interest- cases immediately before the caesura.
ing and important paper. 9 Cf. Shipley, 1.c., p. 182.
4 C.R. xli (1927), 123. 'o It reappears in Auson. 31. 306, Prudent.
s e.g. 9. 677, 678, 12. 757, 769, 771 (with Symm. 2. 68I.
fixam et, 773). Neither Shipley nor Johnson " Shipley notes how many 'common'
takes account of Virgilian 'paragraph'- compound verbs appear once only in Virgil:
structure (e.g. in the group 3. 188, 2oo, 222), 'one might as well question 4. 297 because
but this scarcely weakens their case. of praesensit'.
6 Cf. O. Braum, De monosyllabis ante 12 It seems to me highly probable that
caesuras hexametri latini collocatis, diss. Mar- there is an actual reminiscence here of Lucr.
burg,; 19o6, p. 22. 3. zoz8 ff. 'at mens sibi conscia factis / prae-
7 e.g. 5. 737, 743, 744, 747, 7. 66, 69, 70; metuens adhibet stimulos torretque flagellis,/
G. 1. 149, 154, 159; G. 2. 347, 348, 355- nec videt interea qui terminus esse malorum /
8 The frequency of et-nine times in these possit nec quae sit poenarum denique finis.'

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 188 R. G. AUSTIN

occurrence makes the passa

again here only in Virgil. B
that with explere, as it is
Cicero's translation from H
Virgil it occurs often in Ov
used with a genitive in O
cineres satiasse as un-Virgi
certainly, but personal fee
Virgil has cineri ingrat
nequiquam adloquerer cin
ipsa cinis', Cicero says of o
cineri atque ossibus filii s
think that cineres satiasse i
an accompanying ablative,
'satiatos iam suppliciis noc
4- 735 'patrias satiabis cae
(see later) makes this an un
I pass now to a number o
with the conclusion 'das kl
Wortwahl, Bedeutung ode
(a) servantem (568). This is
ing', like the serpent in G
More often in such a con
custodia qualis / vestibulo
382 f. 'nutricis ... /... lime
per me, qui limina servo /
Am. I. 9. 8. 'ille fores domin
But the basic idea in the ve
'guarding' is a secondary n
in Hor. Epp. I. 5- 31 'atria s
the entrance of the shrine
a reasonable picture in itse
(b) aspicio (569). Did Heinze
This use of aspicio, 'I catch
clear in Ecl. 7. 8 (though th
frequent in Comedy (e.g. P
'laetae exclamant "venit",
its use here in a dramatic context.
(c) erranti (570). An obvious difficulty: Aeneas is still on the palace roof (he
does not leave it till 632). It cannot refer to the arx in general, as some have
tried to argue, nor is 4. 691 (oculis errantibus) a relevant parallel, as has been

I For a remarkable passage involving 2 All these passages show how appropriate
cinis see 5. 785 ff. 'non media de gente satiare is in the context of 587.
Phrygum exedisse nefandis / urbem odiis satis 3 Op. cit., p. 45, n. I.
est nec poenam traxe per omnem / reliquias 4 Cf. Madvig, Adversaria Critica, ii. 107,
Troiae: cineres atque ossa peremptae / in-against the reading focum servat in Ovid,
F. 6. 317: 'non agitur de assidua ad focum
sequitur': how the wolves might have fallen
on these lines if there had been any hint commoratione.' I suspect that Aeneas' in-
of mystery about their genuineness--ex- struction in 2. 711 'longe servet vestigia
edisse, poenam traxe per omnem, the ossa of anconiunx' really means that Creusa is to
urbs! keep fairly close behind.

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88 189

suggested. Its meaning is so closely defined bypassim

be taken simply to imply 'bewilderment' (cf. Ov
patres ambagibus errant'). It can only mean that A
roof, looking everywhere: not unintelligible, perha
the strange setting of the whole scene from 460 o
(d) sibi infestos (57I). If sibi is taken closely with i
natural compression of language (for qui sibi infesti
with praemetuens. For ob with infestus cf. Sall. lug
amicitiam infesti' (but it is a comparatively ra
parallel in verse).
(e) invisa (574). It is, of course, tempting to take
auctus on 592, ed. Harv.), which would give a c
abdiderat sese, with the adjective in a natural relatio
quite improbable: apart from Cato, agr. 141. 2 (wit
(with inauditus), and Apul. Met. 5. 3. 5, this sense oc
in other late texts. Further, invisa clearly applies no
abdiderat sese as well, and the meaning 'hateful' pr
preceding lines: there she sat in hiding, the hated
loathed by Greeks, loathed especially at this mome
text, 'unseen' would be intolerably lame and tam
occurred to Virgil, or to any Latin writer, that in
(f) exarsere ignes animo (575). Virgil does not use
nor is exardescere found elsewhere in classical Latin
literally or metaphorically (but cf. Sen. Epp. 91.
he has exarsit dolor (5. 172, cf. 8. 219 f.), exarsit violen
445), and the phrase in question is far less rema
muros et castra tuenti / ignescunt irae, duris dolor o
irae is a striking experiment in language.
(g) subit ira . . . ulcisci (575-6). The infinitive
although there is no exact parallel with ira: cf. 9. 7
rumpere claustra manu',6 Ovid, ex Pont. 4. 7. 30 'v
viros'.
(h) parto triumpho (578.) For the use ofparto cf. pa
term seems incongruous in a Trojan context, wh
triumphis / dives'?
(i) Iliadum turba (58o). Wagner derides this ('nonn
salutatione propinquorum turbam Iliacam com
appropriate unless 579 is ejected (see later): but wh
of Helen free and triumphant, returning home wi

be ablative.
' The curious suggestion was made by
Heidtmann that erranti . . . 3 Note the refers
ferenti juxtaposition
to in 7. 570 f.
Helen, not to Aeneas. This scholar, whose 'Erinys, / invisum numen': Helen is no
text of this book was published at Wesel in numen, but she is an Erinys.
1882, ejected 230 lines as spurious, boasting 4 Unless possibly in 7. 577, where Servius
that he had beaten Peerlkamp's record. interprets igni as 'ipso fervore deditionis'.
2 It has been suggested that aris is dative, Cf. Val. Flacc. I. 748 'saevos irarum concipit
with invisa (cf. Gerloff, Vindiciae Vergilianae, ignes'.
diss. Iena, 1911, pp. 35, 38, and Palmer, s Note Ovid, Met. 6. 708 'arserunt agitati
1.c., p. 375, n. 2). But this is no more justifi- fortius ignes' (of love).
able than it would be to take ratibus superbis 6 Perhaps also Prop. I. I I. 5 (but see
as dative with invisam in 4- 540. Aris must Shackleton Bailey, Propertiana, p. 32).

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 19o R. G. AUSTIN
Trojan women (cf. 785 f. 'non ego...
Iliades cf. 3. 65, I I. 34 f. 'Troianaq
usage in this sense, and occurs elsew
not be Ovid's work).
(j) exstinxisse nefas (585). It has b
person (8. 688 is no parallel); but
appears to use patricium nefas of
(Th. 7. 514) 'nupsi equidempeperiqu
certainly remarkable; but Aeneas is
was in 157 ('fas mihi Graiorum sa
nefas is equivalent to upholding wha
with reference to a person, it might
the embodiment.
(k) ferebar (588). The meaning is doubtful. Actual movement may be meant
(cf. 337, 511, 725), though one would expect a goal to be added: if so, Aeneas
begins to plunge forward to attack Helen, forgetting that he is on the roof. But
ferebar is better taken of Aeneas' frenzied state (cf. 4. 376 'heu furiis incensa
feror'). The line may well be a stopgap (see later).
So much for Heinze's specific points; I cannot see that his presumable
objections amount to anything like cumulative evidence against the Virgilian
character of the passage. Much more serious are the problems raised by
sceleratas poenas (576), merentis (585), and by the text of 579, 584, and 587. I
shall now consider these.
(a) sceleratas sumere poenas. 'A variety, perhaps a harsh one, for sceleris poenas
sumere orpoenas ex scelerata sumere' (Conington) ; 'an artificial but quite Virgilian
way of saying "punishment of a criminal" ' (Mackail) ; 'sceleratas sumere poenas
kann Vergil so wenig sagen wie sumpsisse merentes [sic] poenas, und gewiss nicht
beides hintereinander' (Leo);3 'unanst6ssig' (Birt).4 The sentiment is clearly
that of 12. 949 'poenas scelerato ex sanguine sunmit' (which Leo infers was the
'interpolator's' model). But the phrase is remarkable, and suggested parallels
are irrelevant (6. 563 sceleratum insistere limen;s G. 2. 256 sceleratum frigus;6
7. 766 patriasque explerit sanguine poenas).7 A faintly possible analogy is Prop. I.
13. 10 'multarum miseras exiget una vices', where Shackleton Bailey suggests
that miseras may be equivalent to a misero, and quotes this line in illustration.8
I think that it bears all the marks of a Virgilian invention: it does not seem in
the least like a coinage by some poetaster. Conington's interpretation is
reasonable enough: I do not believe the view9 that sceleratas means that Aeneas
is judging his own intention as a scelus which must nevertheless be committed.
Such self-condemnation is ruled out by laudabor (586), and there is no moral
content in nullum memorabile nomen (583).

' Birt, op. cit., p. I61, n. i. 7 Belling, Studien fiber die Compositionskunst
2 This parallel was noted by Forbiger; butVergils in der Aeneide (Leipzig, 1899), p. 178.
exstinctum following nefas there is a con- 8 Propertiana, p. 39; he regards vices as
jecture only, though a probable one (by equivalent to ultionem, noting Servius on
Gronovius), based on this passage, for MS.2. 433-
est tunc. 9 Cf. Palmer, 1.c., p. 377; Pascal, op. cit.,
3 Plautinische Forschungen, p. 43, n. p. I2o, n. 2; Hatch, Classical Philology liv
4 Op. cit., p. 16x, n. i. (1959), 255 f. Gossrau comments 'sceleratae
poenae sunt eae, quibus novum scelus com-
s Fairclough, 1.c., p. 223; see Norden,
ad loc. mittitur . . .; ita Orestes in Clytaemnestra
6 Noack, 1.c., p. 425, n. sceleratas poenas sumpsit'.

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88 191

(b) sumpsisse merentis / laudabor poenas. The cons

infinitive uses with dicor, existimor, as in Persius I
laudatur' (Prop. I. 7. II is analogous, 'me laud
puellae'); for other verbs cf. Prop. 4. I I. 36 'un
Am. 2. 6. 6I 'colligor ex ipso dominae placuisse
implicit), Donatus ad Ter. Eun. 479 'reprehensu
looks conservative enough beside Ovid, Am. i. 14.
umenti sustinuisse manu'.
Merentis is far more difficult. If it is genitive singular, it is an unparalleled
construction; Gossrau explains it as equivalent to de merente (similarly Heyne,
Forbiger) : Belling saw in the genitive construction 'einen Hauch echt vergili-
schen Geistes',' Peerlkamp commented 'poenas alicuius sumere vix Latinum est'.
If it is accusative plural, as most now take it, and as the run of the lines suggests,
it gives little less trouble: it has been thought to bear a passive sense (so
Wagner), which may perhaps be supported from Statius, S. 4. 3. I45 'seriem
merentis aevi';2z more probably it is an experiment like sceleratas, with the
epithet transferred from the culprit to the punishment. I prefer to regard it
as accusative. If this is right, it finally excludes the interpretation of sceleratas
as a piece of self-condemnation: the same act cannot be at once a scelus and
merens in the eyes of the same person. Again, the expression looks to me more
like the work of Virgil than like a clever idea on the part of a 'sciolus'.
(c) The text of 579. This line has been rejected even by some who do not
reject the whole passage: Wagner tartly enjoins 'abeat ineptissimus versiculus,
unde irrepsit'. Patris is a late correction of patres, but the run of the line
supports it, as well as the natural sense. The pattern in coniugium, domum patris,
natos-husband (for this use of coniugium cf. I I. 270), ancestral home, children-
is a familiar triple arrangement:3 so Sinon sighs gustily for his patria, his nati,
and his parens (137 f.); cf. Stat. Th. 3. 578 f. 'liquere domos dilectaque laeti /
conubia et primo plorantes limine natos', ib. 7. 719 'et domus et coniunx et
amantes litora nati', Minucius Felix 2. I 'relicta domo, coniuge, liberis'. If
patres is read, the pattern becomes coniugium domumque, patres natosque, making
a jerky rhythm and isolating videbit from the first pair of objects, as well as
spoiling the natural triad.
The use ofpatres for parentes (cf. soceros of both parents-in-law, 457) is frequent
in inscriptions,4 but there is no certain example of it in literary Latin,s and
there seems no reason why either Virgil or an interpolator pretending to be
Virgil should have wanted to use it here. Mackail's idea that patres natosque
may mean 'the older and younger generations of her family' is astonishing.
But it is an irrelevant objection to patres that Helen's mother is to be presumed
dead, or that as her father was really Jupiter she could hardly visit him

x Op. cit., p. I81. ibit and videbit; cf. 6. 683 'fataque fortunasque
2 See Helm, Philologische Wochenschrift,virum moresque manusque', where the
1934, col. 1419, where this line of Virgil isgrouping is similar (fata, fortunas virum,
adduced; Frere concurs; Vollmer, more moresque manusque).
probably, takes seriem as object of merentis. 4 L6fstedt, Syntactica, i. 69; it seems
Immerens is passive in Fulgentius, aet. mund. especially common near Trier (Fahnestock
p. 153. 4. Wackernagel, who holds these and Peaks, T.A.P.A. xliv [1913], 80).
lines to be interpolated, seems to take s Ovid, Met. 4. 61 has been adduced, but
merentis in 585 as a passive (Vorlesungen iiber patres need not mean parentes either there or
Syntax, i. 286). in the passages noted by Langen on Val.
3 -que with coniugium is the link between Flacc. I. I50.

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
i92 R. G. AUSTIN
(a bright thought of Wagner's, appare
of patres, without the aid of contorti
The main objection brought against
nexion between 578 and 580 (so Kv
(Wagner). But there is no interrup
return in triumph (578) to the boso
captive slaves (580)? Nor is the line
is a natural and moving reflection on
deserta Creusa, / et direpta domus et pa
the line.
(d) The text of 584. Servius gives hab
habet for habet haec, accepted by man
trusion,2 arising clearly from the neg
thought-connexion if habet haec is ke
proper punctuation of 585. Haec is nec
victoria, and the form that it takes is
belongs primarily to habet, and its
and Janell, who both retain habet h
punctuation, which makes it apply to

namque etsi nullum memo

feminea in poena est, habet ha
exstinxisse nefas tamen, et sump
laudabor poenas.
The long postponement of tamen is u
but the whole makes sense: for a near
cf. 9. 44 f. 'ergo etsi conferre manum
tamen et praecepta facessunt'.
Victoria is curious at first sight. It m
so in Io. 409, when a shepherd has fir
sedens victor flammas despectat ovan
he has safely got himself and the bab
(e) The text of 587. This is probably
vary between famam, famae; flammae
text retains. Some emendations offer
(Belling) ;6 ultrici etflamma cineres (Pee
ultricemque famem (Leo);8 altricis fam
iuvabit).9 These are eloquent of despai

' Natos too has been criticized,

h. v. laudem, since
extinxisse nefas, tamen; et .... in
Homer there is an sonly In his textchild, Hermione;
of I886, inaccessible to me.
6 Op. cit., p. a
but other legends mention I82;son,
B. ingeniously observes
Nicostratus
'die Wiederholung
(Apollodorus, Bibl. 3. I I. I, des Wortes
where . .. ist in
see
Frazer's note for this and other versions of
dem ganzen Abschnitt so wenig vermieden,
dass es nichts ausmacht, ob es noch einmal
Helen's family). But this is pedantry: nati
is conventional .and a natural convention. mehr steht'.
7 P. explains that Aeneas intended
2 See Fraenkel, J.R.S. xxxviii (1948), I37.
3 Mackail: habet h. v. laudem, / exstinxisse
Helenam in flammas inicere.
8 Plautinische
nefas; tamen et ...; Janell: habet h. v. laudem: / Forschungen, p. 42, n. 3; L.
extinxissenefas tamen et... (similarly Ladewig-
adds 'das ist aber wohl zu gut flir diesen
Poeten'.
Deuticke, but with a full stop after laudem).
4 Birt, op. cit., p. 161, n. I, reads: habet
9 Ancient Lives of Vergil, p. 24.

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88 193

that ultricis flammae can be taken as genitive afte

the line as an assortment of glosses, concluding
nimirum captu breviter lector quidam adiecerat, m
versus speciem redegit'.2 These scioli were a hard-w
Flammae is an obvious stopgap. Ultrixflamma can
as has been imagined, for flamma never means
this sense cannot be justified here from ignes in 5
muros ultricis semita flammae', the reference is t
and so offers no more support than does the occur
in Christian writers3-indeedflammae here might w
his mind fixed on hell fire.
Birt4 defended the expression ultrix fama as an invention for fama ultionis,
comparing Prop. 2. 32. 2I 'famae iactura pudicae', where he tookfama pudica
as meaningfama pudicitiae:s but this is no parallel. It is explained in Thes. L.L.6
as 'ea laus quae ultorem sequetur', surely impossible. Ultrix means 'vengeance-
bringing'. It is used (a) of the Furies, passim; (b) of war (Stat. S. 5. 2. 150, Sil.
2. 423); (c) of weapons (Aen. II. 590, Stat. Th. 5. 117, IO. 911, Sil. I. 539,
12. 414); (d) of instruments, etc., of punishment (catenae, Stat. Th. 2. 271;
flamma, ib. I I. 4; manus, ib. I I. 693; dextera, Quintil. decl. 3. 8; umbra, 'a ghost',
ib. i o. 19 ; undae, Ovid, Met. 3. 190; rates, id. ex Pont. 5. 90; daps, Sen. Thy. 894) ;
(e) of places (Tartara, Stat. Th. Io. 25; Europa, id. Ach. I. 397); (f) of time
(dies, Lucan 7. 782; hora, Sil. 5. 655); (g) of emotions (curae, Aen. 6. 274; rabies,
Stat. Th. 4- 360; irae, ib. i1. 563; paenitentia, Quintil. decl. 12. 11). There
is no place here for ultrixfama. But even if the phrase is conceded as a possible
invention, its only meaning could be 'fame that brings vengeance'. How could
Aeneas'fama bring vengeance on Helen ? How, for that matter, canfama 'bring
vengeance' at all?
There is the further problem of the construction, if ultricisfamae is swallowed
as a possible reading. Explere occurs nowhere else with a genitive, and it could
only be explained here by analogy with the genitive use after implere (which
is found with abstracts first in Livy). But even if this in turn were thought
possible, is it a probable supplementation of animum explesse? For animum
explere in itself means 'to satisfy the feelings', and the context normally shows
what those feelings are: thus, Ter. Andr. 188 'dum tempus ad eam rem tulit,
sivi, animum ('pleasures') ut expleret suom', Hec. 755 'exple animum ('mis-
givings') eis', Cic Clu. 202 'nemo huic tam iniquus praeter parentem fuit, cuius
non animum ('enmity') iam expletum esse putemus'. An ablative (voluptate,
gaudio, etc.) is sometimes added, marking the means of satisfaction. Here, then,
we should have to assume that ultricisfamae takes the place of such an ablative,
which seems to me in the highest degree unlikely. I cannot believe in ultricis
famae as a sound reading. The sense is complete at iuvabit: 'animum explesse'
I Pascal, op. cit., p. 12o, n. 2, following a Cemenelensis, hornm. 8. 5 'ultrices criminum
suggestion by Heyne. flammae'. But it must be granted that
2 1.c., pp. 399 f. Firmicus Maternus has two clear Virgilian
echoes in the de errore (3. I 'pro iniuria
3 Firmicus Maternus, de err. prof. relig.
16. 2 'quid enim meretur aliud parricida, spretae fecerat formae', 6. 7 'Liber... cum
nisi ut perpetua <con)tinuatione flammarum semiviro comitatu fugiens').
ante sententiam dei cotidie flammis ultricibus
4Op. cit., p. 161, n. I.
concremetur?'; Rufinus, hist. 3- 7. 8 'si s Shackleton Bailey, Propertiana, p. 126,
possent ultrices poenarum flammas lacri-explains differently.
marum ubertate restinguere'; Valerianus 6 s.v. expleo, col. 1717- 42.
4599.3 O

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 194 R. G. AUSTIN

means 'to have satisfied my hat

suggesting a lacuna before 587,
should be retained, though uninte
Such are the main problems o
minimize them. In addition to his
the point that 577-87 forms an
narrative to Dido, producing an eff
This is mechanical criticism. Wha
envisaged is strikingly dramatic:
of the drama, heightening the qu
appears. Why should such an effe
precedented ?z
There are, however, other proble
expression: aspicio (569), aspiciet (
sceleratas sumere poenas (576), femin
(585-6) ; deserti coniugis iras (572), s
animo (575), Troia arserit igni (58 I,
(573, of Greece), patriam (576, of
laudabor (586), famam (?) (587),
Aeneas sees Helen limina Vestae ser
in 572 she is afraid of Menelaus, i
him. The last line (588) looks like a
if the passage had been fully inte
expected talia iactabam et ... but t
(cf i. . 02, 9. 621; Ovid, Met.
I I. I55; Stat. Th. 7. 668). I suspe
the way in which the passage sh
But there are clear Virgilian tou
manner;4 so is the pattern of 568,
and end of the line) ;5 so is the pat
cf. 61, 413, 514, etc.). There is a c
two short interrogative clauses ra
line. Alliteration6 and assonance
controlled: the lines do not read
position which it occupies in 584:
Lucan, and Silius occasionally plac
Ennian (Sc. 181 V.) and Lucretia
giving Aeneas his glimpse of Helen
Virgil does so well (cf. 2. 694, 7
Let us now assume that nothing
We shall find that (a) cum (589) is
(593) and her words in 594 have
z Op. cit., p. 47.
3 Again in 407, and nowhere else in Virgil;
2 Birt (op. cit., it recurs
p. 161, in Sil. n.
2. 210,
i)Val. Flacc. 8. 445- to
objects
absence of inquam or the like; but its 4 Cf. Leo, Ausgewiihlte kleine Schriften, ii.
presence would have been frostig in the52.
extreme. Yet Birt himself (p. I3') adduces s See Norden, Aeneis VI, Anhang III. A. 2.
this passage to illustrate Hor. Epod. 4. 11 ff., 6 Cf. Kvivala, op. cit., pp. 33 ff.; with
where liberrima indignatio breaks into speech exstinxisse . . . sumpsisse . . . explesse compare
(with no inquit) in a context not unlike in 4. 603 ff. (fuisset ... tulissem ... implessem . .
tone.
exstinxem . . . dedissem).

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88 195

follows awkwardly close upon 560 ff., where Aenea

for his family; (d) the mention of Helen (6oi) is
though it can be explained (especially with the a
reflection on the ultimate human cause of the
stood, or was planned to stand, where the 'Helen
been assumed (e.g. by Heinze) that Virgil had or
as intending to die, either in battle or by his own
he removed the lines, leaving a lacuna which was lat
But such a representation would have been very
Aeneas' mind is filled with fear for his family; he l
circum copia (564), clearly with the intention of goin
himself quite alone-and did he then decide that
Impossible. The theory goes back to Ti. Claudius Do
in 565-6 evidence that Aeneas is patting himself on
constantia, proceeds at once to comment on the
which he treats as a piece of whitewashing to conc
for Aeneas' misfortunes: 'omnes ... duplicem vol
nancti . . . aut praecipites dederant sese aut pet
alterum facerem aut, quod supererat tertium, ho
mater mea ... me ab ipsis inefficacissimis actibus rev
indeed: and the dreary 'interpretatio' of Venus'
plainly what heavy weather the commentator was
supposed continuity from 566 direct to 589. The d
gruous: the passage preserved by Servius contain
congruous but heightens the tension with true poe
we are told stood there, in favour of what surely n
An interesting theory was produced by K6rte,3 wh
eye passes directly from 566 to 632, the narrative r
ness from Aeneas' realization of his position to his d
Korte deduced that 589-631 was a later addition
the book to Augustus, to strengthen his defence o
that Virgil never completed the transition needed
with what he had previously composed; and that an
plan 'das Loch im Gewebe der Dichtung mit de
Gespinst zu stopfen'.4 Korte had two main grounds
596-600 to 560-3: he feels that the two passages can
original plan, and concludes that Virgil must h
earlier lines. But the difficulty is more apparen
realizes the danger of his family: in 596-600 he lea
are yet safe, and his search to find quae sit me circ
has lost all human aid, but divinity is at his side. I
two passages they would jar: but it is not necessary
something did stand between them. K6rte's second
' Cf. Heinze, op. cit., p. 50.
Tyndaridis facies (601) occurs in a papyrus
2Cf. O. Walter, Die Entstehung
dated by E. der Halb-
G. Turner to the first century
A.D.: see Studi
verse in der Aeneis, diss. Giessen, in onore
1933, P- di Aristide Calderini e
34.
3Hermes li (1916), 145 ff.
Roberto Paribeni (I957), ii. I57 ff., and J. G.
4 Peerlkamp regarded 589-623 as Milne, J.H.S. xxviii (1908), 125. Even
spurious. If the Venus-scene had been inter-
Heidtmann retained the scene, in a mangled
form.
polated, which of course is nonsense, it must
have happened very early, for non tibi

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 196 R. G. AUSTIN

reading ducente deo in 632. Here

a good example of an early Virg
justifiable without the Venus-scen
This is a stronger argument: for
is indeed odd when Aeneas has jus
Troy, quite apart from its strange
said tutum patrio te limine sistam
here for the later composition of
to be an interpolation; if anythin
fact a Virgilian draft of a transit
but takes it for granted that they
language and metrical technique
The supposed interpolator wou
ungebildeter Mann'. The passage s
It shows a Virgilian familiarity w
clear affinity with Euripides (Or.
the striking phrase of Helen as Tr
I 0o. 59, where Cleopatra is dedecus
like a downright imitation of thi
It shows one plain reminiscence o
well be an adaptation of the leg
Helen, but put it up at the sight o
had to be more than merely learn
imagination, and deep dramatic in
I read these lines, the less they
Gruppe thought that they were c
pretty nineteenth-century phanta
ever seems properly to have faced
arrestingly treated. It is not re
matically; we should consider why
he had written, or why his execu
removed any passage that they
would not have acted so becaus
writing nonsense, or impieties, o
a shrewd executor could spot imp
dissatisfaction with something

' deo V, above dea; in M dea has been on 7. 498); this may have been a neoteric
corrected to deo; in P dea is a supralineal mannerism, since the corresponding Greek
correction of de; Schol. Veron. on v. 467 usage of e'ds is frequent in Callimachus
quotes the line with deo. Servius (here and on (see Pfeiffer's index).
I. 382, 4. 228, 7. 498), Macrobius (3. 8. I), 3 See Heinze, op. cit., p. 48.
Aelius Donatus (on Ter. Ad. 894, Eun. 875) 4 Heinze denies this, adducing instead
support deo. Ti. Claudius Donatus read Eur. Or. 1388. There is no reason why both
abducente dea (wicked Venus persuading her passages should not lie behind the phrase, as
son to desert his post). The balance is well as Ennius, Sc. 71 V. 'Lacedaemonia
strong in favour of the masculine. mulier, Furiarum una', which Fraenkel (on
Aesch. l.c.) thinks may go back to Euripides.
2 See Thes. L.L. s.v. deus, col. 890. I6 ff.,
for the masculine used of a male or female s Eur. Andr. 627 ff., Ar. Lys. I55 f.; the
numen (but Catull. 61. 64 is wrongly classi-story goes back to the Ilias Parva. Cf. Gerloff,
fied). Both Servius and Macrobius quote
op. cit., pp. 25 if.; Mancuso, 1.c., p. 25;
Calvus 'pollentemque deum Venerem' Noack, 1.c., p. 428.
(these scholia are closely related, cf. Servius

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 VIRGIL, AENEID 2. 567-88 197

handled. Obviously there were such passages. The S

that the two passages named as deleted are illustra
on which Virgil's editors worked ('ut in principio'
into the vortex of the ille ego controversy (and it
that both passages were removed for the same rea
claim that the 'Helen-episode', in the state in whic
is precisely the kind of thing that Virgil wished m
authorized, finished work, and that it was recogn
The lines could have reached Servius in various wa
workshop must have been littered with rough dra
would have been discarded by Virgil in his lifetim
had access to actual manuscripts of the poet,3 m
or some friend, recognizing the power behind
obviously unrevised piece, may have preserved it
uncanonical.4
The authority of the chief denigrators of the passage is such that it needs
courage to disagree with them. But they have been strangely dogmatic, and it
would have been better if they had done something more explicit and con-
structive than merely listing words and expressions that in their view merit
abuse. Others have already shown that Norden's metrical argument is a
fallacy; I hope that my own detailed examination of the linguistic problems has
shown that, though some of them are formidable, they do not put the passage
beyond the Virgilian pale. Paradoxical as it may seem, I feel that the very
imperfections of these lines speak for Servius' good faith. I think that it is a
mistake to regard the passage as a unit, a single draft: it looks much more like
a collection of drafts, stages, or strata in a scheme, with a clear disintegration
towards the end. The opening lines (567-70), far from being stiimperhaft, are
firmly wrought, and, taken alone, present no insoluble problem: they might
even be the beginning of a final draft. The next four lines (571-4) may also
belong to this stratum, but aris invisa sedebat suggests the beginning of an
alternative plan to limina Vestae servantem. From 575 onwards the idea of
'punishment' is worked out in various forms, starting with the experimental
sceleratas sumere poenas (576). The opening of the soliloquy (577-82) looks like
something fairly definitive. But from 583 to 587 we have what is obviously
tentative and incomplete :s there is the repeated poena, poenas, with sumpsisse
merentis poenas plainly an experiment on the lines of sceleratas sumere poenas and
surely an alternative idea to it-the two could not have stood in a final draft, as,
to do him justice, Leo pointed out; habet haec victoria laudem exstinxisse nefas
starts a train of thought to which laudabor is alternative. This tentative set of
notes may have ended with animumque explesse iuvabit, and the corrupt 587 may
represent the start of a new plan (famam is an obvious variant on the theme
of laus), with a reworking of the idea of vengeance (ultricis here, ulcisci above,
576). Cineres satiasse meorum might be an alternative to animum explesse; it might

ambigebat, quo magis iudicium hominum

SCf. Bill, Classical Philology xxvii (1932),
I68 f.
2 T.A.P.A. Ivi (1925), 184.
3 See Aulus Gellius, 13. 21. 4 (Probus), I.
21.2 (Hyginus); cf. Bill, l.c.
4 Cf. Donatus, vita 33 'recitavit et pluri-
bus, sed neque frequenter et ea fere de quibus
experiretur'.
s Jahn, retaining habet haec in 584, adds
'ich habe . .. die folgenden drei Verse als
Dittogramm ausgesondert, bin aber meiner
Sache keineswegs sicher'.

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms
 198 R. G. AUSTIN

also be a jotting of a new idea, b

meorum'. The last line of the pa
link with what follows. The wh
mind at work. The two most di
poenas (really one difficulty,
Aeneid equally hard), suggest hi
has handed down a piece of raw
it can be judged in a proper per
been capable of the imaginativ
thought and in expression; no
flawless forgery, would he have
tions and contradictions.' Virg
doing when they did not include
what Virgil himself would have

The University of Liverpool R. G. AUSTIN

Note. The time needed for the preparation of this pape

my tenure of a Leverhulme Research Fellowship, for
grateful.

Cf. Funaioli, Studi di letteratura antica II. i. 245.

This content downloaded from

132.248.9.41 on Fri, 13 Aug 2021 18:32:16 UTC
All use subject to https://about.jstor.org/terms