¹ To be sure, ‘Aeolic song’ might have in view the sympotic and civic element of Alcaeus' poetry. But only a minority of the 88 poems in Odes 1–3 are sympotic or civic, whereas all of them are in lyric metre.

² Nisbet and Rudd reject the idea that deduxisse has anything to do with spinning a thread, saying that ad suggests an action involving a destination. A fortiori, they say, this tells against the idea that Horace's poems are ‘fine-spun’ in the Callimachean sense. They are undoubtedly right on the second count, but ad need not imply motion and can be used for ‘to the accompaniment of’: see OLD or OLD ² s.v., 39.

³ Sydenham says, ‘Measures embrace themes and style as well as metres.’ But even if that were clearly the case, ‘metres’ is surely bound to be felt in modi, and this creates perplexity. Williams takes a different tack: ‘What this means is that the poet recreated Aeolian poetry in the Latin language; in doing so he transferred the rhythms of Aeolian poetry to Italy and thus created an “Italian” sound which was quite distinct from the Aeolian sound. Here the adjective Italos is proleptic: that is, it describes, in advance, the result which Horace achieved, for it was only by creating such poetry that he made the rhythms Italian.’ But though a proleptic adjective may modify the direct or indirect object of an active verb, e.g. Carm. 3.16.19 conspicuum tollere uerticem and Ov. Met. 1.184 inicere … captiuo bracchia caelo, or the subject of a passive verb, e.g. Verg. Aen. 3.508 montes umbrantur opaci, the discussions of prolepsis in Kühner–Stegmann 1.239-40 and Leumann–Hofmann–Szantyr 2.413–14 yield no instance where such an adjective modifies the object of a preposition.

⁴ Someone might argue that ad means ‘to join’, and that Horace means ‘I took Aeolic verse-forms and added them to the verse-forms Italians already had.’ But it is difficult to see why Horace used two words for the same thing and why he wrote deduxisse Aeolium carmen ad Italos modos to mean adiecisse Italis Aeolios modos. Furthermore, while hexameters, elegiacs and the metres of drama were imports of long standing, it would be strange to call them Italian.

⁵ So also Pöschl, V., Horazische Lyrik (Heidelberg, 1970), 247Google Scholar, von Albrecht, M., ‘Musik und Dichtung bei Horaz’, in Atti del convegno di Venosa, 8–15 Novembre 1992 (Venosa, 1993), 83Google Scholar and Brandt in TLL 8.1255.68.

⁶ This is true whether Horace's lyre was real or fictional and whether or not the Odes were actually meant to be sung. On the one side, Nisbet and Rudd ad loc. clearly think that singing the Odes to the lyre is a fiction, on the other, Lyons, S., Music in the Odes of Horace (Oxford, 2010)Google Scholar gives reasons for thinking that it was not. What is certain is that Horace represents his verse as sung, whether or not it actually was. This means that his descriptions of musical accompaniment must not describe something that is impossible, such as singing an Alcaic stanza to a tune that was not designed to accommodate it.

⁷ Lyons (n. 6) translates modi as ‘modes’ throughout (e.g. p. 15) but gives no ancient citations to support the equation modus = ἁρμονία.

⁸ Carm. 2.1.40 quaere modos leuiore plectro, 2.9.9 tu semper urges flebilibus modis | Mysten, 2.12.4 mollibus | aptari citharae modis, 3.3.72 magna modis tenuare paruis, 3.9.10 dulcis docta modos et citharae sciens, 3.11.7 dic modos, Lyde quibus obstinatas | applicet auris, 4.6.43 reddidi carmen, docilis modorum | uatis Horati, 4.11.34 condisce modos, amanda | uoce quos reddas; Epist. 1.3.13 fidibusne Latinis | Thebanos aptare modos (but here perhaps ‘metre’); and Ars P. 211 accessit numerisque modisque licentia maior. (At Epist. 1.19.27, quod timui mutare modos et carminis artem, the reference seems to be exclusively metrical.)

⁹ For further vagueness and imprecision (in addition to West's), see Shorey and Laing, who gloss modos as ‘loosely, the measures, the strains, the sounds and special laws of the Latin tongue’. For fudge combined with apology see Page: ‘the words must not be pressed too closely: the “measures” or “metres” that Horace uses are not “Italian” but Greek, e.g. the Alcaic and Sapphic; what he means is that he has introduced a new variety of Italian poetry, copied from Greek models'. Other explications, similarly imprecise, are cited in n. 3 above.

¹⁰ A similar switching of adjectives is desiderated, for different reasons, by Wilkinson, L.P., Golden Latin Artistry (Cambridge, 1963), 104Google Scholar, who writes, ‘we should expect him to say, “compose Italian songs to Aeolic music”’.

¹¹ Nisbet, R.G.M., ‘The word order of Horace's Odes ’, in Adams, J.N. and Mayer, R.G. (edd.), Aspects of the Language of Latin Poetry. Proceedings of the British Academy 93 (Oxford, 1999), 135–54Google Scholar, at 141–2.

¹² See also 3.4.9–12 me fabulosae Vulture in Apulo | … palumbes.

¹³ Cf. also 1.32.4–5, where Alcaeus' barbitos is asked to utter a Latin carmen, which probably refers to musical mode since there is nothing Latin about the content of the ode.

¹⁴ Courtney, E., ‘The transmission of the text of Horace’, in Günther, C. (ed.), Brill's Companion to Horace (Leiden, 2013), 546–60Google Scholar, at 554–5 gives such a list. Eighteen of the examples are in the Odes.

¹⁵ Most editors choose the former, but the latter was conjectured by N. Heinsius before being found in manuscripts. It subsequently won the approval of Bentley. More important is the question utrum in alterum abiturum erat. Once the question is raised, there can be little doubt of the answer.

¹⁶ Fuss, J.D., ‘Lectiones Horatianae’, Zeitschrift für die Alterthumswissenschaft 123–4 (1840), 1015–17Google Scholar.

¹⁷ In Latin distinguished by neither clarity nor elegance, Fuss made the following points: (1) though modi can be used for ‘poetry’ where there is no danger of ambiguity, that condition does not obtain here; (2) no one, Fuss thinks, is likely to interpret as ‘add Aeolic verse-forms to already existing Italian verse-forms’, but just in case, he points out that there are much more natural ways to express this in Latin than the transmitted text (for another statement of this argument, see above, n. 4); (3) the corruption is to be explained by someone's failure to see that carmen was not the object of deduxisse.

¹⁸ After warning their readers off interpreting as Aeolios modos carmen ad Italum or (‘quod etiam peius est’) proposing an alteration of the transmitted text, Orelli and Baiter say, ‘Sed est, Latinis verbis ita imitari metra Graecorum lyrica, ut nusquam violetur [what is the subject of this singular verb? quantitas?], sed perpetuo Latina syllabarum quantitas vocum, atque in omnibus sincera Latinae poësis servetur indoles, id quod effecit Horatius.’ The paraphrase bears little resemblance to the text the editors are trying to defend.

¹⁹ Kovacs, D., ‘When giants stumble: two influential misjudgements on Horace's Odes (2.20.5–6 and 3.30.10–14)’, Philologus 155 (2011), 159–64CrossRefGoogle Scholar.

²⁰ E. Fraenkel, Horace (Oxford, 1957), 304–6.

²¹ When commentators cite Carm. 4.8.26–7, lingua potentium uatum, to show that Horace became potens, they are citing an irrelevancy: Horace is not talking specifically about himself but about poets in general, to whom he ascribes the power to rescue Aeacus from the Underworld and set Romulus, Hercules, Castor and Pollux among the gods. Apart from such a (frankly fantastic) context no one would describe Horace as potens. On the nature of the (mostly) whimsical claim that poets can in actuality rescue people from Hades or make them gods, see Kovacs, D., ‘Horace, Pindar, and the Censorini in Odes 4.8’, JRS 99 (2009), 23–35 Google Scholar, at 32–3.

²² It is characteristic of prose to reinforce a prepositional phrase by adding a prepositional prefix to the verb. It is not surprising that Horace in the Odes does not do this.