Notes on the Text of Lygdamus

The text of the Appendix Tibulliana is not less corrupt than one would expect from a tradition whose earliest complete extant witness dates to 1374/5. The object of the following notes is to reconstruct the autograph of [Tib.] 3.1–6 in a handful of places where a plausible reading has been overlooked or is yet to be conjectured. Five of those reconstructions are quite new; four either have never been printed in modern editions, or have never been printed more than once. The text of passages cited is that of Luck’s second Teubner edition, to which are appended his apparatus and sigla.^[1] To the MS readings which he reports I have added the findings of Navarro Antolín, as well as some forgotten conjectures from older books and periodicals.^[2]

3.1.15–20: Lygdamus asks that the Muses deliver his poetry to Neaera directly.

per uos, auctores huius mihi carminis, oro 15

Castaliamque undam Pieriosque lacus,

ite domum cultumque illi donate libellum,

sicut erit: nullus defluat inde color.

illa mihi referet, sit nostri mutua cura

an minor, an toto pectore deciderim. 20

19 referet Z+ : referat Guelferbyt. 1 et 4 | sit nostri mutua cura Lee : si nostri mutua cura est Z+ : nostri sit mutua cura Voss || 20 minor AVX+ : maneam G | anne amer Heinsius

Editors who adopt Lee’s emendation^[3] of v. 19 have misapprehended the poet’s intention.^[4]si nostri mutua cura est is the protasis of a condition completed by illa mihi referet ... an etc.; it is not, what many conceive it to be, the first clause of an indirect question depending on referet and continued by an minor an toto pectore deciderim.^[5] The construction is this: “if she is concerned for me as I for her (si nostri mutua cura est), then she will say (illa mihi referet) whether she loves me (an †minor†) or whether she loves me not (an toto pectore deciderim)”.^[6] In other words, even if Neaera does not love Lygdamus, provided she still has concern for his well-being,^[7] she ought at least to tell him whether she loves him or not, and thus to spare him the anxiety of not knowing where he stands in her estimation. This analysis conveniently does away with all three objections commonly raised against the tense of referet and the mood of est: (1) that Lygdamus cannot be certain if Neaera will reply (not an issue if referet is part of a condition); (2) that in disjunctive questions si ... an ... an is un-classical (si is part of no question); (3) that the mixture of moods is odd (est and deciderim are not of the same construction).^[8] Emending si nostri mutua cura est to sit nostri mutua cura with Lee or to nostri sit mutua cura with Voss is unattractive for another reason, namely that the second hemistich of v. 19 is copied verbatim from an epigram of Martial (10.13(20).9–10):

si tibi mens eadem, si nostri mutua cura est,

in quocumque loco Roma duobus erit. ^[9]

There, as here, the present protasis si nostri mutua cura est is followed by a future apodosis: in quocumque loco Roma duobus erit, “then any place will be Rome for us two”, corresponds to illa mihi referet, “then she will tell me”. The only problem with this conditional interpretation of Lygdamus’ si nostri mutua cura est lies in v. 20, for there minor (AVX+) continues to give awkward sense, as it requires cura to be supplied from the protasis of v. 19, and thus means “if she has any care for me in return, then she will say whether it [sc. her care for me] is less [sc. than my care for her], or whether I have fallen entirely out of her heart”. As a disjunction between equivalents, this is nonsense;^[10] and if Neaera’s concern for Lygdamus is potentially minor, can it really be mutua? A promising alternative is suggested by G’s maneam, which if adopted would give the first an-clause a finite verb to oppose deciderim in the second: “if she cares for me in return, then she will say whether I abide in her heart, or whether I have fallen entirely out of it”.^[11] Printed, I believe, only by Baehrens and Fabricius, maneam was ridiculed by Magnus in a review of the latter on the ground that fulfilment of the initial condition (si nostri mutua cura est) presupposes a positive answer to the first alternative and a negative one to the second, thus rendering the whole sentence illogical.^[12] But this is to draw a false equivalence between cura and amor: mutual concern, such as that which might subsist between peaceably separated partners, may imply but does not entail mutual love. But if even on this reading mutua continues to seem a not entirely apposite epithet of cura, it is easily accounted as a by-product of Lygdamus’ desire to incorporate matter procured from Martial (10.13.9) into a context where it does not fit perfectly. This is suggested also by the omission of tibi, which is necessary in the original and whose place must here be supplied by <illi>, and by the unusual interposition of the protasis between both halves of the apodosis.^[13]

If maneam be right, pectore must be mentally supplied from the second an-clause.^[14] My one gripe with this reading is its distance from A’s minor, the older and more authoritative lection, into which it is not easy to see how maneam might have been corrupted. If, as seems likely, maneam is only a conjecture of Pontano’s, a critic may be justified to look for a verb that bore a similar sense to maneam but might have been more easily corrupted into minor. Heinsius, for example, proposed to read anne amer; but this form of amare is unknown in classical literature, and the coordination of anne with an is hard to parallel.^[15] A likelier reading I think is morer, to be taken either in the sense of “retain the attention of, hold” (OLD s.v. 3) and “fix the attention of, delight, amuse, entertain” (L&S s. v. II 2) with eam supplied in our heads,^[16] or, what is perhaps slightly easier, in the sense of “linger” with pectore supplied ἀπὸ κοινοῦ: i. e. “she will tell me whether I linger in her breast, or whether I have fallen entirely out of it” (cf. Hazlewood, The Mother’s Dying Child, Act 3 Scene 3: “Can it be possible that any spark of love for him still lingers in your heart?”). Something of an analogy to this use of moror can be found in Ov. Ars am. 3.435–436:

quae uobis dicunt, dixerunt mille puellis:

errat et in nulla sede moratur amor.

If amor there be taken to mean the focus of one’s love, which never “lingers” in any place because it is always fixating on some new object, then it may not be a complete abuse of Latin idiom for Lygdamus to enquire whether he still “lingers” in Neaera’s heart, inasmuch he conceives himself to be a potential object of her affections, the Latin for which is also amor (OLD s. v. 3d). The confusion of moror and minor is frequent: cf. e. g. Claud. Carm. mai. 15.370 minas] moras and 22.226 moras] minas.^[17]

3.2.15–18: Lygdamus’ mother and sister are envisaged as performing the rites of his funeral.

praefatae ante meos manes animamque recentem 15

perfusaeque pias ante liquore manus,

pars quae sola mei superabit coporis, ossa

incinctae nigra candida ueste legant.

5 praefatae ante Z+ : praefataeque Huschke | recentem Bach : rogatae AG²VX+ : togatae G : precatae Y

For rogatae of AG²VX, admitted by all editors to be a nonsensical word, Bach proposes to read recentem, which finds a strong parallel in Ov. Met. 8.488 fraterni manes animaeque recentes.^[18] Lee, Tränkle, Luck and Maltby print it; Navarro Antolín returns to the old correction, precatae, but gives no reason for doing so.^[19] If his apprehension is owed to the dissimilarity of rogatae to recentem, that is fair enough; but it is an apprehension easily dispelled: recentem has come to rogatae by a scribe’s perseveration of praefatae in the same verse. Connected with this corruption, I believe, is the superfluity of ante before meos manes. praefatae, deponent, normally takes an accusative directly.^[20]ante then must mean antea or else be locatival,^[21] but why the pleonasm after prae-? and why at the cost of an elided dipthong?^[22] Says Maltby: “[t]he pleonastic expression praefatae ante is characteristic of our author in the Lygdamus poems”.^[23] He cites for parallels 4.47–48 aeuique futuri / euentura, 4.93 longe ante alias omnes mitissima, and 5.27 uano nequiquam;^[24] but the pleonasms in these examples are different in kind to that in 2.15, for in none of them does a word attempt to usurp the grammatical function of another. praefatae expects a direct object; ante wants to deprive it of one. Balance with the participle in v. 16, perfusaeque, urges us to adopt Huschke’s correction, praefataeque.^[25] He explains ante as having arisen by an accidental dittography of the final letters of p̄fate, thus creating p̄fateate (for p̄fateq;), which a later scribe, under the influence of ante in v. 16, then interpreted as p̄fateāte, viz. praefatae ante.

3.4.5–8: Lots and entrails tell true, but dreams tell false.

diui uera monent, uenturae nuntia sortis 5

uera monent Tuscis exta probata uiris:

somnia fallaci ludunt temeraria nocte

et pauidas mentes falsa timere iubent.

5 diui uera monent uenturae nuntia sortis] diuae u. m., uenturi nuntia, sortes Navarro Antolín

Problems with the sense and punctuation of v. 5 are acutely identified and explained by Navarro Antolín.^[26] The first difficulty concerns punctuation: does uenturae nuntia sortis go with uera₁ in v. 5 (“the gods send true communications of coming fortune”), with uera₂ in v. 6 (“entrails approved by Tuscan seers send true communications of coming fortune”) or with exta (“entrails approved by Tuscan seers, communicative of coming fortune, tell true”)? To take uenturae nuntia sortis with uera₂ or with exta is unattractive for two reasons: (1) the two resulting sentences, diui uera monent and uenturae nuntia sortis / uera monent Tuscis exta probata uiris, would be grossly disproportioned; and (2) the bald statement diui uera monent would contradict Lygdamus’ later claim (7–8) that gods who speak through dreams, such as that in which Apollo will soon address him, give false indications of the future, whereas the drawing of lots and inspection of entrails can be relied upon for true prophecies.^[27] To have uenturae nuntia sortis modify uera₁ runs into the same objection: Lygdamus’ point is that the gods tell true not in every medium, including dreams, but only through proper types of divination, such as augury.^[28] Navarro Antolín seems right in wanting sortes to be the subject of uera monent in v. 5, thus balancing uera monent ... exta in v. 6 (“lots tell the truth ... entrails tell the truth”). But his solution, which is to read

diuae uera monent, uenturi nuntia, sortes ;

uera monent Tuscis exta probata uiris,

I am unable to accept as the poet’s words without further alteration. That sortes (for sortis) should be the subject of monent, and that uenturi (for uenturae) should depend upon nuntia, are certain propositions; but the mere change of diui into diuae is unsatisfactory for reasons which are two. Firstly, the resulting word-order is not in keeping with Latin poetic style. To indicate that one noun phrase stands in apposition to another, such as uenturi nuntia to diuae ... sortes, the poets typically enclose it between the constituent elements of the phrase to which it is apposed; what critics call ‘inserted apposition’.^[29] I quote by way of example Verg. Ecl. 2.3 tantum inter densas, umbrosa cacumina, fagos; Prop. 4.9.3 uenit ad inuictos, pecorosa Palatia, montis; Epic. Drus. 283 adice Ledaeos, concordia sidera, fratres; and Ov. Met. 11.412 consulat ut sacras, hominum oblectamina, sortes.^[30] This is a universal tendency, in view of which the order diuae uera monent, uenturi nuntia, sortes seems highly anomalous, especially given how easily diuae might have been transposed to sit before uenturi, and given that, had Lygdamus written the words in this order, uera monent would stand at the beginning of two consecutive verses. That Lygdamus and the other authors of the Appendix Tibulliana are interested in such anaphoric repetition is shown by the following parallels:

saeuus amor docuit ualidos temptare leones,

saeuus amor docuit uerbera posse pati. (4.65–66)

nec bene mendaci risus componitur ore,

nec bene sollicitis ebria uerba sonant. (6.35–36)

urit, seu tyria uoluit procedere palla:

urit, seu niuea candida ueste uenit. (8.11–12)

huc ades et tenerae morbos expelle puellae:

huc ades, intonsa Phoebe superbe lyra. (10.1–2)

Lygdamus did not write mendaci bene nec in 6.35, nor would he have written diuae uera monent in v. 5 with uera monent Tuscis in v. 6. Two arguments, then, against the final shape of Navarro Antolín’s emendation, and implicitly against the text as printed by recent editors: (1) the failure to attain the word order expected of an inserted apposition, and (2) the desirability of anaphora in consecutive verses. I for my part find it almost inconceivable that Lygdamus should here forego both effects, of which he happily avails himself elsewhere, when by a very slight rearrangement he could have procured them thus:

uera monent diuae, uenturi nuntia, sortes;

uera monent Tuscis exta probata uiris. ^[31]

But diuae (= diuinae) is an inapposite epithet of sortes, with which it is not elsewhere collocated.^[32] Far more probable, in view of Ovid’s sacras, hominum oblectamina, sortes (Met. 11.412), is the claim that Lygdamus wrote

uera monent sacrae, uenturi nuntia, sortes;

uera monent Tuscis exta probata uiris.

True are the warnings given by sacred lots, tidings of the future; true are the warnings given by entrails examined by Tuscan seers.

The plural of sors is scarcely ever accorded an epithet other than sacrae: cf. Tib. 1.3.11 sacras ... sortes; 2.5.69 sacras ... sortes; Ov. Met. 1.368 sacras ... sortes; 11.412 sacras ... sortes; Pont. 3.1.131 sacras ... sortis. For (prae)nuntia in this connection, cf. Verg. Aen. 11.139 et iam Fama uolans, tanti praenuntia luctus; Ov. Met. 5.549 foedaque fit uolucris, uenturi nuntia luctus; Fast. 6.207 hinc solet hasta manu, belli praenuntia, mitti. If the corruption of this into diui uera monent uenturae nuntia sortis wants explaining, I imagine one scribe’s eye skipped from monent directly to uenturi, leaving out sacrae by accident, and that another scribe interpolated diui to fill out the metre, changing uenturi and sortes to suit his sense of the meaning.^[33] This sort of error occurs elsewhere in the Appendix Tibulliana: through the testimony of F and the Florilegium Gallicum, we know that at 3.7.96 the words ueniat grauis impetus hastae (F) were corrupted in an ancestor of Z+ to ueniat grandis impetus hastae and then purposely rearranged into grandis uenit impetus hastae (Z+).

3.4.43–48: Apollo explains his powers of prophecy to a dreaming Lygdamus.

“salue, cura deum: casto nam rite poetae

Phoebusque et Bacchus Pieridesque fauent:

sed proles Semeles Bacchus doctaeque sorores 45

dicere non norunt, quid ferat hora sequens:

at mihi fatorum leges aeuique futuri

euentura pater posse uidere dedit”.

45 Semeles Η : -is G : -(a)e AVX+ | bachus Z+ : mollis Damsté || 46 non norunt Z+ : non ualent C | quid ferat hora sequens Z+ : quid sequens hora ferat C || 47 (a)euique AX+ : cuique V : cuiusque GV²

There is an art to vv. 44–45. In the pentameter Lygdamus states each person and group by name: “Phoebus, Bacchus, the Muses”.^[34] In the following line he turns to periphrasis: “the offspring of Semele, the learned sisters”. But this elegant variation is blasted to bits by the reappearance of Bacchus in v. 45.^[35] “Poets are favoured of Phoebus, Bacchus, and the daughters of Pierus. But the offspring of Semele (that’s Bacchus by the way) and the learned sisters know not how to say etc.”. If Lygdamus through Apollo is glossing his own allusion, why does only Bacchus get the advantage of it? Why is the reader left in the dark as to the identity of these doctae sorores? Does not the foregoing list of Phoebus, Bacchus, and Pierides make obvious all their identities? Navarro Antolín’s C gives Liber as a variant for Bacchus (cf. Ov. Met. 3.520 proles Semeleia, Liber), which, going by the quality and character of its other readings, is almost certainly an intrusive gloss.^[36] Yet to gloss proles Semeles by this or another name is hardly an improvement, when the “learned sisters” receive no like elucidation. Observe that bachus in v. 44 lies directly above bachus in v. 45, and the cause of corruption discovers itself: a scribe has erred by concentrating overmuch on the preceding line and bouncing its contents into the next.

Damsté wondered if bachus in v. 45 had forced out an epithet of proles, perhaps mollis.^[37] If this were so, we would expect the adjective to stand before the noun (sed mollis proles), not after it. I prefer to think that bacchus has ousted a verb in kind to dicere non norunt, with quid ferat hora sequens depending on it ἀπὸ κοινοῦ. That verb I conceive to be nescit. The sense is “but the offspring of Semele does not know, and the learned sisters know not how to say, what the next hour will bring”. Construed in this way, Bacchus and the Muses’ respective shortcomings are more particularly distinguished. Bacchus does not know the future, and the Muses cannot articulate it­; an ironic situation for the latter, and only the latter, so often invoked by dicite, Pierides (Phoebus: “can’t dicere this, can you?”).^[38] For nescit + non nouit having separate subjects and the same object, cf. Ov. Fast. 2.83–84:

quod mare non nouit, quae nescit Ariona tellus?

carmine currentes ille tenebat aquas.

In Fast. 2.83 two verbs in asyndetic coordination together govern a single noun, but clauses of course can be objects too.^[39] For the exact construction posited here (“A does not know, and B cannot say, X”), I find a curious parallel in a statement of Housman’s, which, though it cannot serve for an illustration of Latin usage, may yet be worth adducing in conclusion: “[a]s regards IV 7 20 ‘iustoque uindicauit exemplo impiam’, apart from the unexpectedness of this mythological detail, I do not know, and the lexicons do not say, what authority there is for the construction uindico impiam in the sense of uindico in impiam or uindico impietatem”.^[40] There is a similar but not identical opposition between nescio and dico in the first sentence of Livy’s Ab Vrbe Condita (praef. 1): facturusne operae pretium sim, si a primordio urbis res populi Romani perscripserim, nec satis scio, nec, si sciam, dicere ausim. For nescire applied to quid ferat hora sequens, cf. Varro’s Sat. Men. 333 nescis quid uesper serus uehat (= Gell. 1.22.4; 13.11.1; Macr. Sat. 2.8.2–3). Granted, bachus and nescit do not look terribly alike; but dissimilitude of appearance is no sure protection against error, especially here, with one bachus hanging so precariously above the other.^[41]

3.5.5–8: Lygdamus prays to avert his death.

at mihi Persephone nigram denuntiat horam: 5

inmerito iuueni parce nocere, dea.

non ego temptaui nulli temeranda uirorum

audax laudandae sacra docere deae.

7 uirorum EW, Scal. ex Cuiac. rec. in quo ḍẹọṛ uirorum : deorum Z+ : piorum exc. Colotii : reorum Sandbach || 8 laudandae Z+ : celandae exc. Petrei, Broukhus. ex uno Vaticano : uelandae Hall | sacra docere Z+ : sacra mouere Voss. 2 Heynii : sacra uidere Erath : cernere sacra Voss : sacra subire uel prodere sacra Heinsius : noscere uel discere sacra Heyne

The identity of this laudanda dea has long been sought for. Some believe her to be Persephone, others the Bona Dea. On the ‘pro’ side of seeing in this goddess the person of Ceres’ daughter are, firstly, the epithet laudandae, which is thought to allude to the epic formula ἐπαινὴ Περσεφόνεια (cf. Hom. Il. 9.457; Od. 10.491; Hes. Th. 768), and secondly, the verb docere, which may suggest the illicit communication of the Eleusinian mysteries to uninitiated persons (cf. Hor. Carm. 3.2.267 qui Cereris sacrum / uulgarit arcanae; Ov. Ars 2.601–602 quis Cereris ritus ausit uulgare profanis).^[42] On the ‘con’ side are the following considerations: (1) Roman literary authors invariably identify the Eleusinian mysteries with Persephone’s mother Ceres, not Persephone herself; (2) that laudanda = Homer’s ἐπαινή, in the opinion of Tränkle, seems an allusion almost too learned for Lygdamus, who does not elsewhere exhibit familiarity with archaic Greek epic formulae; (3) the phrase nulli temeranda uirorum is an inaccurate description of the Eleusinian mysteries if uirorum means only “men/males”, and a superfluous description if uirorum means “persons”.^[43] Sandbach answers the last of these objections by emending deorum (due probably to dea ending v. 6 and deae ending v. 8)^[44] to reorum, “guilty persons”, on the ground that persons accused of crimes were not permitted to join in the rites of Demeter (cf. Suet. Ner. 34).^[45] But reorum presents its own difficulties. The point of Lygdamus’ long list of exculpations is to show that he is not a guilty man. Would it not then be a most curious thing for him to say that he “did not attempt to spread rites that ought not to be violated by guilty persons”, as if it were by reason of being a guilty man that he abstained from betraying these rites? A fourth reason to be suspicious of laudanda dea = Persephone is this: it would be a strange thing if Lygdamus were to begin his speech by addressing Persephone in the second person (6 parce ... nocere), only then to describe her rites as if she were a third disinterested party (7–8 nulli temeranda ... sacra ... deae). There are better reasons for identifying the goddess of v. 8 as the Bona Dea. In the first place, if deae = Bonae Deae, then as the main characteristic of her mystery cult was its limitation to members of the female sex, uirorum in v. 7 would be allowed to bear its proper sense of “men, males”, and nulli temeranda uirorum ... sacra would mean no more than “rites not to be violated by men”.^[46] In the next place, the necessity of not violating the mysteries of the Bona Dea is several times mentioned in the sort of elegiac verse upon which Lygdamus models his own. Compare Tib. 1.6.21–24:

exibit quam saepe, time, seu uisere dicet

sacra Bonae maribus non adeunda Deae.

at mihi si credas, illam sequar unus ad aras:

tunc mihi non oculis sit timuisse meis.

Prop. 4.9.25–26:

femineae loca clausa deae, fontesque piandos

impune et nullis sacra retecta uiris.

Ov. Ars Am. 3.633–638:

quid faciat custos ...

...

cum fuget a templis oculos Bona Diua uirorum,

praeterquam siquos illa uenire iubet?

Ov. Fast. 5.147–154:

quo feror? Augustus mensis mihi carminis huius

ius habet: interea Diua canenda Bona est.

...

templa Patres illic oculos exosa uiriles

leniter acclini constituere iugo.

These are affirming parallels. Tränkle, Navarro Antolín and Maltby therefore settle on Bona Dea as the referent of laudanda dea.^[47] But two problems with the text subsist. Firstly, as all these commentators note, “divulging” the mysteries of the Bona Dea is one step further than the main offence, viz. entering her temple and beholding her mysteries. Tränkle therefore obelizes docere, while singling out two conjectures, Erath’s sacra uidere and Voss’s cernere sacra,^[48] as solutions attractive in sense but too distant from the ductus litterarum to be printable. Maltby and Navarro Antolín, while they acknowledge the inconcinnity, yet allow it to stand. That docere looks suspiciously like the nocere of the previous pentameter gives me reason to claim it for another Perseverationsfehler, and if that assumption holds true, then whatever word it ousted need not bear a shadow of resemblance to the transmitted text. As Tibullus (1.6.24), Propertius (4.9.53) and Ovid (Ars am. 3.637; Fast. 5.153) allude specifically to the punishment of blindness, the best solution is surely Erath’s uidere. Tränkle adduces the parallel of Cic. Har. resp. 37:

quod quidem sacrificium [sc. Bonae Deae] nemo ante P. Clodium omni memoria uiolauit, nemo umquam adiit, nemo neglexit, nemo uir aspicere non horruit.^[49]

The phrase sacra uidere appears in this connection a paragraph later in Cic. Har. resp. 38 quis enim ante te sacra illa uir sciens uiderat? and is paralleled also by Ov. Met. 3.518 ne Bacchica sacra uideres.

The second problem of this pentameter concerns laudandae. This is not an epithet elsewhere accorded to the Bona Dea, nor is it one that would seem adequately to distinguish her from other gods and goddesses, many of whom could surely be considered as “praiseworthy” as she.^[50] In search of a more relevant epithet some early humanists proposed celandae, “that must be hidden”, Hall uelandae, “that must be veiled”.^[51] Since after temeranda another genitive might appear clumsily repetitious (“the rites that must not be violated of the goddess that must be concealed/hidden/praised”), I was for a time disposed to suggest arcanae, a word often used of gods and goddesses in their mysterious aspects. For this see Klotz, ThlL II 436.41–44: “de deis mysteriorum: Hor. Carm. 3.2.26 qui Cereris sacrum uulgarit arcanae; Ov. Am. 1.7.10 ausus in arcanas poscere tela deas; Sil. 3.394 redimitaque sacrā / nebride et arcano (sic C recte: ortano S) Maenas nocturna Lyaeo”; OLD s.v. arcanus 3: “(in general) mysterious, magical, mystic ... CIL 14.2852 fortvnae simvlacra colens et apollinis aras / arcanvmqve iovem”. If laudandae were another Perseverationsfehler composite of audax and temeranda, I would impute the corruption arcanae → laudandae to the same scribe who in 1.10 wrote pumicet (Z+) for pumex (G²) under the influence of indicet in v. 12. But after further consideration I think laudanda may be able to stand, if with Heyne and Brouwer we connect it with an idea of moral excellence: the goddess is “to be praised” (laudanda) because she is “good” (bona).^[52] An interesting but little-known parallel for the attribution of laudanda to a goddess is offered by a coin issued in the reign of Pertinax (RIC IV.1 7), on the reverse of which are inscribed the words menti lavdandae, which is presumably another way of denominating Mens Bona.^[53] But though the stem of this word is surely right, the dependence of one gerundive-noun phrase (laudandae ... deae) upon another gerundive-noun phrase (temeranda ... sacra) continues to seem a little clumsy, especially given that the author could have written, what comes to much the same thing, laudatae, “distinguished, lauded”. The past participle of this word is often used attributively: see Kamptz, ThlL VII.2 1046.73–76, and add Stat. Silv. 3.4.6; 5.1.11 laudati iuuenis rarissima coniunx; 5.3.113; Mart. 1.93.5; 4.60.4; 6.43.7; 9.17.3; and Sil. Pun. 2.512.^[54] For the confusion of laudanda and laudata, here possibly occasioned by the like termination of temeranda, cf. Cic. Fam. 1.9.21 laudata] laudanda, Hor. Sat. 2.5.75 laudato] laudando, Luc. 8.485 laudata] laudanda.^[55]

3.6.9–12: Lygdamus enjoins his friends to remain at a dinner party.

uos modo proposito dulces faueatis amici

neue neget quisquam me duce se comitem, 10

aut, si quis uini certamen mite recuset,

fallat eum tecto cara puella dolo.

11 mite Z+ : inire Burman, Schrader, Mitscherlich suis Martibus | recuset Scal. ex edd. : recusat Z+

mite is a strange epithet to apply to certamen. Drinking games in antiquity were not usually regarded as “soft” affairs: prizes were awarded for imbibing as much as one could eat (Plin. Nat. 14.140), and vomiting seems to have been of regular occurrence (cf. the young komast retching on the Brygos Kylix, Würzburg L 479 = ARV² 372, 32).^[56] The Latin poets tend to describe such contests as “battles” (Mart. 4.82.6 sua ... proelia Bacchus amat), where Bacchus is said to “rage” (Mart. 10.20.19 furit Lyaeus) or to “go a-revelling” (Mart. 9.61.15 commissatore Lyaeo). The common discourse of drinking games thus puts us in expectation of a ‘rowdy’ epithet, but this expectation is upset by mite, which also unhappily anticipates the appearance of mites two verses later in 13 ille facit mites (Lipsius : dites Z+) animos deus.^[57] Fulkerson regards the epithet as deliberately oxymoronic: “compared to the certamina of love, this one is gentle”.^[58] That the certamen bibendi is a “gentler” business than the certamen amoris might be Lygdamus’ opinion, but is it the opinion of the man who excuses himself from the party? The idea here seems not to be that the man exempts himself from the gentler contest of wine in order to enter the more brutal contest of love, but rather (in Lygdamus’ view) that to desert one’s boon companions for a girl is an ignoble behaviour that ought to be met with contempt and disappointment (12 fallat eum tecto cara puella dolo); cf. Ambrosius, De Helia et Ieiunio 13 nota grauis, si quis se excuset, si quis temperandum forte uinum putet.^[59] Navarro Antolín connects mite with the wine itself (cf. Verg. Georg. 1.344), but this hardly helps to clarify its literal meaning as an attribute of certamen.^[60] Maltby takes mite to mean iucundum: “the epithet here simply signifies that wine drinking is a pleasurable activity, especially for those whose pain Bacchus is relieving”.^[61] If Lygdamus had intended this, it is strange that he did not use a more direct word to convey it, rather than one whose common associations (“soft, gentle, meek, nonviolent”) are antithetical to the word it qualifies. As mite upsets the contrast established between the riot of uini certamen and the softness of cara puella, it seems best to posit corruption.

Three scholars, Burman, Schrader, and Mitscherlich, for mite each independently conjectured inire, “to enter upon”.^[62] This is an attractive change, for inire and mite are practically identical in minuscule script (ın = m, t = r), and parallels for the collocation of inire with certamen are found in Ov. Am. 2.2.61 certamen inis and Sil. Pun. 10.69 certamen inire. As for the metre, Hermann’s bridge was not quite so hallowed in Latin verse as it was in Homer’s, and one might consider the rhythm as peculiarly matched to sense, the bouncy amphibrach evincing the fun of the party. A contrary effect is achieved by Virgil (Ecl. 1.53–55) through similar artifice:

hinc tibi, quae semper, uicino ab limite saepes

Hyblaeis apibus florem depasta salicti

saepe leui somnum suadebit inire susurro. 55

The neoclassical poet Vincent Obsopoeus in his poem De Arte Bibendi (1536) five times describes the start of a drinking game by the words certamen inire:

aut si nemo tuo vasto cum gutture iustum

certamen vini potor inire potest? (2.661–662)

sic libet atque iuvat vini certamen inire

claraque suffuso ferre tropaea mero. (3.23–24)

nunc inquis “vini certamen inire laboro,

actutum veniat potus abire volens”. (3.63–64)

nec plures tecum certamen inire Falerni

optes audaci sustineasque manu. (3.187–188)

solus vobiscum certamen inire recuso;

solus cum multis morio bella gerat. (3.657–658)^[63]

A sixteenth-century humanist writing a didactic poem in the style of Ovid (cf. 1.1 si quis in hac artem non nouerit urbe bibendi etc.) does not possess the same evidential value in matters of textual criticism as, say, a poem by Ovid himself; but the resemblance between these lines and Burman, Schrader and Mitscherlich’s version of 6.11 seems too remarkable to be set entirely at naught.

3.6.53–56: Lygdamus impugns the honesty of his girlfriend.

quam uellem tecum longas requiescere noctes,

et tecum longos peruigilare dies,

perfida, nec merito nobis nec amica merenti, 55

perfida, sed quamuis perfida, cara tamen.

55 nec amica W, exc. Pocchi, ed. Ven. a. 1475 : inimica Z+ : et amica Postgate dub. | merenti Z+ : petenti P : querenti Ox₂ (merenti in mg. add. Ox₂²) : ferenti C

Verse 55, which the best and oldest MSS give in the form perfida, nec merito nobis inimica merenti, contains a long-standing crux. Before proceeding to discuss this problem, it may be worth noting that Postgate and Tränkle enclose the last words of this verse, inimica merenti, within obeli; that Navarro Antolín and Luck print the emendation of W and the editio Venetis of 1475, nec amica merenti; that Lenz/Galinsky and Maltby retain the paradosis, unintelligible though it is.^[64] “A difficult line”, says the latter, “meaning literally, ‘faithless girl, hostile to me (nobis) unjustly, though deserving well’, where nec merito = ‘undeserving’, and merenti = ‘though deserving well (of you)’”.^[65] This translation is more generous than literal. Neither “well” nor “of you”, which are essential to make this version work, are present in the Latin.^[66] Literally this verse means “faithless girl, and unfairly hostile to me who am deserving”. Deserving of what? Most readers (this one certainly) are likely to supply as the object of merenti the enmity suggested by inimica, but the resulting sense is very strange: “undeservedly hostile to one who deserves [it]”. If the unintelligibility of this description were not already apparent, the unexampled attribution of a singular participle (merenti) to a plural substantive (nobis) should make it quite plain. That nobis merenti may equate to nobis merentibus in, say, the Latin of early comedy, is perhaps conceivable; but the fact that not one parallel for the concord of a singular pronoun and plural participle subsists in poets whom Lygdamus either appears to imitate or is himself to be counted among, lays it firmly under suspicion. So much is admitted by Tränkle, Navarro Antolín, and Maltby too.^[67] Postgate and Tränkle, as I have said, therefore obelize the final two words of the verse; Navarro Antolín and Luck on the other hand print the reading of W, nec amica for inimica, which is apparently meant to yield the sense “unjustly faithless to me and friendly to one who does not deserve [it]” (i. e. nec amica merenti = et amica inmerenti).^[68] Such an oddly contorted phrase, in which the first nec is unaccountably postponed after perfida and the second goes with a word which it does not stand before, is hardly in the plain and simple style of Lygdamus.

The suspicious likeness of merenti to merito suggests that a scribe has accidentally transcribed the beginning of one of these words over the beginning of the other. If that is true, then the seat of corruption lies probably in merenti, not, as commonly supposed, in inimica. But simply to change this into another participle, such as petenti or ferenti with the later MSS, will not alone suffice, for the discord of number between this and nobis remains impossible. A glance at two parallel constructions may guide an answer:

perfide, nec cuiquam melior sperande puellae. (Prop. 4.7.13)

perfide, nec sedes umquam meriture quietas. (Stat. Theb. 11.569)

Analogy suggests that nobis and merenti conceal a pre-caesural epithet and a verse-terminal noun, such as cuiquam and puellae are in Prop. 4.7.13. The possibility that Lygdamus has Propertius’ syntax in mind is supported by the curious fact that his pentameter, perfida, sed quamuis perfida, cara tamen (56), seems vaguely to be based on Prop. 1.17.16 quamuis dura, tamen cara (ς : rara Ω) puella fuit. Taking this evidence into account, and having thought upon this problem much longer than anybody really ought to, I am fairly certain, so far as certainty can be reached in such matters, that the best conceivable emendation of v. 55 is this:

perfida, nec merito nostris inimica querelis,

perfida, sed quamuis perfida, cara tamen.

“Faithless girl, and unfairly ill-disposed to my complaints; faithless, but though faithless, yet dear”. Lygdamus is recalling the instructions of Apollo in 4.75–76 and regretting that they have not availed him:

ergo ne dubita blandas adhibere querelas:

uincuntur molli pectora dura prece.

So do not scruple to press your gentle complaints: hard hearts are conquered by soft pleading.

In 6.55, as in 4.75, querelae will denote the lover’s complaints, a thing so common in love elegy as hardly to require illustration here.^[69] That such complaints are liable to cause displeasure (55 inimica) is the penetrating observation of Propertius (2.18.1–2):

assiduae multis odium peperere querelae:

frangitur in tacito femina saepe uiro.

Complaining all the time has instilled hatred in many hearts: often a woman is broken by a man’s silence.

For nostrae + querelae arranged as I propose to arrange them, compare the following:

ultima tu nostris accedis causa querelis. (Ov. Her. 15.71)

et tacitus nostras intra te fige querelas. (Juv. 9.94)

nimirum timuit nostras Fortuna querelas. (Mart. 1.12.9)

For the vocative of inimicus + a dative noun qualified by noster, cf. Ciris 287 o iterum nostrae Minos inimice senectae. For inimicus + the dative of querelae/preces/uota, cf. Ov. Pont. 2.8.38 numina sint precibus non inimica meis. For the meeting of -īca quer-, cf. Mart. 1.53.10 inproba Cecropias offendit pica querelas. The corruption itself, of nostris to nobis and querelis to merenti, is much more straightforward than it may at first appear. By perseveration of merito the poet’s querelis devolved first into merentis and then into merenti (or, if not that, first into *merelis and then by emendation into merenti). The reading of Ox₂, querenti, unknown to me when querelis first entered the brain, shows how easily quer- and mer- may be exchanged; the same MS commits a similar error at 3.5.33 (promittite Z+ : perquirite Ox₂).^[70] The corruption would be of the same kind as that which editors implicitly suppose to have afflicted 3.2.15 (recentem → rogatae due to praefatae preceding), as discussed above. The corruption of nostris to nobis on the other hand is trivial. I cite the following parallels: nobis in error for nostris at Ov. Met. 8.126, 13.372 and Tr. 1.6.3 (all contrary to syntax); nobis in error for nostri at Tr. 5.3.34; nobis and nostro in error for nostra at Her. 16.323.^[71] The reverse error, nostris for nobis, is committed at Ov. Her. 17.81; nostro is given in error for nobis at Prop. 3.2.9.^[72]

3.6.57–8: Lygdamus asks for his wine to be diluted with water.

Naida Bacchus amat: cessas, o lente minister?

temperet annosum Marcia lympha merum.

“Are you lagging, my slow attendant?” The question is pointless. A positive answer is implied by the epithet of the addressee. We should read not cessas, o, “are you idle?”, but rather quid cessas, “why are you idle?”, a reading discovered in the margin of Bern, Bongars VI 155 and in the margin and the text of the Editio Gryphiana (1573), as reported by H. Dixon.^[73] This correction, though it was known also to Huschke,^[74] has not found its way even into the apparatus of modern editions; and yet it gives vastly better sense and can easily be illustrated: Tib. 2.2.10 en age, quid cessas?; Verg. Aen. 11.389 imus in aduersos – quid cessas?; Ov. Fast. 6.675 quid cessas conuiuia soluere?; Sen. Med. 895 quid, anime, cessas? Maltby and Navarro Antolín even cite ad loc. the very pertinent parallel of Mart. 9.93.1 addere quid cessas, puer, inmortale Falernum?, “why, boy, are you idle? Pour in the immortal Falernian!”^[75] That quod (qd) and o are sometimes exchanged in MSS was illustrated by Goold in his seminal paper on Catullus 3.16, where for o factum male, o miselle passer (an impossible hiatus) we must surely read o factum male, quod, miselle passer.^[76] I believe quid cessas at [Tib.] 3.6.57 was last seen in the text of Sidney Walker’s Corpus Poetarum Latinorum.^[77]