« السابقةمتابعة »
CVM AND CVMVLVS.
(a) PLAUTUS, Trinummus, 820-823:
salsipotenti et mulsipotenti Ioui' fratri aetherei Neptuno
laetu' lubens laudis ago et gratis gratiasque habeo et fluctibu' salsis,
quos penes <est> mei <posita> potestas, bonis mis quid foret et meae uitae, quom suis med ex locis in patriam turbis cummamt reducem faciunt.
In 820 the corrections of Buecheler and Scaliger are accepted. In 822 the MSS. have quom penes me potestas; quom for quos from v. 823, lacuna supplied by Klotz, mei <fuit nulla>, by Leo, mei <fuit saepe>. I conceive an intermediate stage: quom penes mei <positast> potestas, which would account for the lacuna by haplography, as well as for the change of mei to me (posita pot- is, of course, a proceleusmatic foot).
But it is to 823 that I wish to call attention.
Two metrical points must first be fixed. The only trochee admitted into this passage in 'anapaestic' measure (820-842A) is uela in 837. It follows scindere; and the verse follows another broken verse ending with frangere malum. The trochee is admitted, then, only for a broken effect. In 823 no trochee can be admitted, so that med ex | lócis in | patriam | must be rightly printed by Lindsay. Again, the fourth foot here must end with the end of a word; circum-stabant (835) is no exception. in | patriam urb | -em could not stand without elision of -em (as e.g. in Wagner's text: in patriam urbem usque incolumem). Leo's in | patri | -am suau | -issumam is thus doubly improbable.
But the letters urbis cummam (cumam: PCD) admit of a simpler correction. There are many analogies even in this play for the expansion of cummam into cumula<tu>m or cumulatū, and for eliciting from urbis ut bis. I propose to read: quom suis med ex locis in patriam, ut bis cumulatum, reducem faciunt. Charmides, as vv. 834, 839, show, returns doubly loaded with wealth.
The closely parallel passage upon Neptune at Rudens 906 sqq. confirms the proposal :
quom me ex suis locis pulchre ornatum expediuit,
(b) Tacitus, Histories, II. 7:
non fallebat duces impetus militum, sed bellantibus aliis placuit exspectari. +bellü cum int uictores uictosque numquam solida fide coalescere . . .
The reading of M has been emended by Heraeus to bello ciuili (whom Spooner follows), by Meiser to bellum ruere in, by Ritter to exspectari bellum cum his.
The words uictores . . . coalescere are manifestly untrue without qualification. The preceding words should furnish the qualification. Tacitus is fond of the construction of coalescere with in and the accusative-e.g. in this book, ch. 37: neque... exercitus linguis moribusque dissonos in hunc consensum potuisse coalescere; Annals, III. 38.6: quae causa fuit ne in bellum atrox coalescerent. In our passage Tacitus did not write in bellum, though both words appear in the MS. His more forcible phrase was, I think:
belli in cumulum uictores uictosque numquam solida fide coalescere.
[bellu from belli, in then added and misplaced; cùm from cūtū, or the following in perhaps displaced a final m.] Tacitus has just written, in ch. 6, tarda mole ciuilis belli of Vespasian's massing forces. Here he says that Othonians and Vitellians, whichever win, will never grow together into a solidly combined mass for allied warfare against Vespasian. cumulus has its literal meaning; it is used figuratively in ch. 24 mille equites, cumulus prosperis aut subsidium laborantibus.
(c) If these two proposals commend themselves they lend weight to the two which follow. If cumulatum can become cumam or cumam, and cumulum cùm or cùm in, in the most ancient sources for Plautus and Tacitus, I may point to passages undoubtedly corrupt where the letters of cum occur, and where a case of cumulus would plausibly mend the sense.
None of the many emendations has ever been found satisfactory. rejected his own first proposal, yet Duff prints it as the only one satisfying sense:
aeraque proporro solidumque senescere ferrum.
Ellis altered the second half of the proposal to silicumque senescere petras, which is less far from the MSS., but refuted by silices in the next line. The irony which Munro read into his second proposal,
non monimenta uidemus | quaerere proporro sibi <sene> senescere credas?
is quite alien to the context, nor, I think, could sibi have stood there. Munro conceives the reference to be to inscriptions seen by us upon the monuments; but such inscriptions as these, addressed to the passer-by, would not assert the brief mortality of the monument. No, the sibi is seen in proper context at II. 979:
sibi proporro quae sint primordia quaerunt,
and may well have been introduced into our verse from that one.
If either sibi or cumque is to be regarded as 'a mere insertion,' I am in no doubt which of the two to reject, for cumque makes no sort of sense, and scribes of the ninth century would hardly be likely to recognize it as one word. Sibi must go, and cumque remain to be accounted for.
Lachmann's proposal rejects sibi, expands cumque, and extracts quae from quaerere. In three points I shall follow him. But his verse
quae fore proporro uetitumque senescere credas
assigns to proporro and to credas a nuance of meaning which neither can bear. Munro rightly declares that the force of proporro in Lucretius is always "then further in turn" or the like.' credas in face of dilapsa uidemus would surely need to be crederes or credidisses. proporro, then, which none would dare to impugn, insists that v. 812 carries the sense of v. 811 to some further point, and credas that the belief you are likely to hold is consistent with the witness of our eyes. Now uinci ruere and fatisci above are present infinitives describing a process, not a completed result, and dilapsa, so often used of buildings, does not mean that these have fallen to the earth
in fragments, but that they are in a state of disintegration, have fallen into decay, need restoration. [Tac. Ann. IV. 43. 6; Livy IV. 20. 7, and many inscriptions.] I propose to substitute a t for a r, to reject sibi, and to expand cumque into cumulumque, then to read:
denique non monimenta uirum dilapsa uidemus?
quae tere proporro cumulumque senescere credas.
'Again, see we not monuments of great men fallen into decay? Nay, further, rub them, and thou wouldest believe that their mass ages." Use sense of touch after sense of sight; the external disrepair of the facing is not so serious as the friability of the concrete core. For the construction, compare dic quibus in terris-et eris mihi magnus Apollo, etc. (Virg. Ecl. 3. 104).
Compare the whole passage at I. 311-327, especially 315 sqq.:
strataque iam uolgi pedibus detrita uiarum.
nec porro quaecumque aeuo macieque senescunt
(d) Horace, Odes, I. xxxii. 15:
o decus Phoebi et dapibus supremi
dulce lenimen †mihi cumquet salue
MSS.: mihicumque as one word. The traditional explanation was that cumque stood for quandocumque. The supposed analogy, not very close at best, in Horace's use of quandoque breaks down; that word, in Odes IV. i. 17 and ii. 34, and A.P. 359, stands to quando as namque to nam (so Housman). But why should cumque stand for quando- or cum- cumque rather than for ut-, ubi-, quod-, quale-, or any other cumque ? Could Toré stand for any of their Greek equivalents?
Now rite uocanti refers back to the first word of the ode, poscimur. The ritual summons comes from patron to poet, from poet to lyre. Jove enjoys relaxation at his feast; Phoebus has no such relaxation, but glories in Jove's behest to sing. The lyre is not to Horace a sweet softening of labours,' but to Maecenas or Augustus; to him it is rather the crowning labour itself. mihi, then, in v. 15 would unnecessarily limit vite uocanti to the poet without the patron, and would at the same time attribute Jove's part rather than Apollo's to the poet. The word comes under suspicion of being 'a mere insertion,' like the sibi before the cumque in Lucretius (above). mihi is not infrequently a scribe's stopgap, as e.g. at Propertius, El. IV. 3. 11 in all MSS. but the oldest. And though no such stopgap has been discovered in the Odes, such seem to occur in other works of Horace-e.g. A.P. 65, sterilisue [diu] palus aptaque remis, where even Lindsay (with Servius) does not convince me that palūs could have been shortened, and Bentley's palus <prius> accounts for the corruption: and Epp. II. 2. 199, pauperies immunda [domus] procul absit, where, however, no satisfactory emendation has been offered, and where I now diffidently propose <pio> procul.
I do not, however, need so to argue here. The tradition mihicumque may arise from cumm'que, and that from cumul'que; or, again, cumulusque may have been expressed cumutq, and ut mistaken for m' (see Lindsay's Notae Latinae, p. 384: I. § 476, IV.).
o laborum dulce lenimen cumulusque
gives chiastic correspondence with o decus Phoebi et dapibus grata Iouis, grata being matched by dulce in the same point of the verse. cumulus bears its figurative sense of 'crown'; compare Tacitus' phrase noted above, 'cumulus prosperis aut subsidium laborantibus' (Hist. II. 24). It adds the poet's point of view to the patron's, Apollo's to Jove's, and rite uocanti can now include both, though the poet stands nearest, as is meet.
It remains to defend the Greek use of nominative for vocative 'in solemn address.' The case is in apposition to tu or uos, as at Juvenal IV. 24:
hoc tu, succinctus patria quondam, Crispine, papyro.
Propertius appears to have the construction at Cynthia XVIII. 20: uos eritis testes, si quos habet arbor amores, fagus et Arcadio pinus amica deo, for it is unlikely that he used fagus as of the fourth declension, as Culex 141. Terence may have it at Phormio III. 1. 10 (324), where Donatus writes 'nom. pro uoc.,' but where editors alter amicus to amicu's. Plautus has it, as at Asinaria 655:
di te seruassint semper, | custos erilis, decu' popli, thensaurus copiarum . . . Horace appears to have it at Sat. II. 2. 107, o magnus risus (sc. tu). He certainly has it at A.P. 292, uos, o Pompilius sanguis, carmen reprehendite (imitated by Persius I. 61, uos, o patricius sanguis).
My reasons for rejecting the emendations of Lachmann and Bentley will have appeared; the proposals of Sudhaus and Herwerden need no refutation. But it is too much to hope that so difficult a problem has been solved.
O. L. RICHMOND.
THE IONICVS A MINORE OF HORACE.1
THE Twelfth Ode of the Third Book of Horace consists of four stanzas in this metre, each stanza consisting of ten feet. How these feet should be distributed into verses is a matter of much dispute; but inasmuch as it does not concern me at the present time I shall avoid it by following certain editors of Horace and printing each stanza continuously.
1. Miserarumst neque amori dare ludum neque dulci mala uino lauere aut exanimari metuentis patruae uerbera linguae.
2. Tibi qualum Cythereae puer ales, tibi telas operosaeque Mineruae studium aufert, Neobule, Liparaei nitor Hebri.
3. Simul unctos Tiberinis umeros lauit in undis, eques ipso melior Bellerophonte neque pugno neque segni pede uictus.
4. Catus idem per apertum fugientis agitato grege ceruos iaculari et celer arto latitantem fruticeto excipere aprum.
On the metre of this Ode M. Laurand, in his recent Manual of Greek and Latin Studies (Paris, 1918), p. 791, writes as follows: Dans cette pièce d'Horace l'accent coincide toujours avec l'ictus métrique.' This utterance must be narrowly examined, for at the back of it is a theory. It is the theory, widely held among scholars of to-day, that the ancients wrote Latin verse with one eye on the quantity and the other on the word-accent. And our Ode has been laid under contribution for
1 A paper communicated to the Cambridge Philological Society on February 15, 1923.
this very purpose by the late Professor C. T. Goodell, Chapters on Greek Metric, p. 165, in an argument with which I shall presently deal. When M. Laurand says that 'the accent' always coincides with 'the metrical ictus,' he is thinking only of ionic feet of the form ou-, 'miserárumst' or 'metuéntis.' He is therefore neglecting forms which, like 'dáre lúdum,' 'néque dúlci,' have two accents, he is raising into the dignity of a canon a coincidence which Professor Goodell (see below) pronounced to be 'inevitable and of no significance,' since Latin words with a long penultimate must in any case have their 'accent' on this syllable, and he is at odds with Horace himself, who has for ionic feet in stanza I 'láuere aut ex-,' 'pátruae uér-,' and in stanza 3 'melior Bel-.'
Professor Goodell was more subtle. After speaking of the trimeter of Terence discussed by Caesius Bassus, p. 655 sq., Keil, 'exclúsit réuocat, rédeam? nón si me óbsecret,' where every word-accent coincides with a down-beat, he says:
'This particular line is merely one very good illustration of a rather common phenomenon, common enough to show the tendency referred to above, to make accent and ictus fall on the same syllable in places where otherwise the rhythm would not be sufficiently clear. A notable case is furnished by Horace, Carm. III. 12, in which no word-accent is allowed to fall elsewhere than on one of the three beats of the ionic foot. Of course, as regards the accented longs, that is inevitable and of no significance; but in the case of the two shorts it is otherwise. And though in the sixteen lines of the poem there are twenty-one instances of an accented short penult or antepenult, in every instance that accented short syllable is the former of the pair which the meter requires, never the latter. It is hard to see any reason why Horace never made that pair consist of the final short of one word and the accented short of a following iambic word, unless it was a desire to make the word-accent a help rather than a hindrance of the rhythm, since this was an unusual one in Latin.'
Professor Goodell then appeals to 'the very next ode,' III. 13, 'also containing sixteen lines but with only twenty-four pairs of short syllables against forty pairs in III. 12,' where there are three pairs consisting of a short final syllable followed by an accented short initial syllable.'2
Let us look into his Horatian example first. III. 13 'O fons Bandusiae' is a 'triform quartet,' to use the nomenclature of my Prosodia Latina, that is to say it is a four-lined combination of three different kinds of verses (Asclepiad, Pherecratean and Glyconic), and each of these is itself composed of different kinds of feet. But III. 12 consists of repetitions of one and the same foot. Nor does the looseness of the comparison stop here. The 'three pairs' of short syllables which Professor Goodell was regarding ('dulci digne méro,'' te flagrantis átrox,' 'me dicente cáuis'), occur in the first half of the Asclepiad only, and in order to get an analogy with a reasonable basis of probability the appeal should have been to the first halves of all the Asclepiads of Horace; but Professor Goodell has drawn his figures not only from the Pherecratean and the Glyconic, but also from the second half of the Asclepiad. Lastly, the ionicus a minore is a foot of three units, and if to give Professor Goodell's thesis every chance we analyze the Asclepiad metre on the assumption that it contains such a foot, what resemblance is there between an ionicus a minore and a choriambus? Need we say more?
To return to III. 12, feet of the form whose absence therefrom Professor Goodell