² On the mock Epicureanism of Catius see Classen, C, ‘Horace—A Cook?’, CQ 28 (1978), 343–8,Google Scholar and Böchner, K, Horaz, Die Satiren (Bologna, 1970), p. 221. It is likely that Horace draws on the braggart Epicurean cook of Damoxenus fr. 2K-A in creating Catius in 2.4 and Nasidienus in 2.8 (see Lejay, pp. 446–7). Just as Catius and Nasidienus, Damoxenus ’ cook is a student of Epicurus, and he applies his knowledge of natural phenomena to his studies of food and dining. He argues, for example, that ’ Nature is the first author of every skill ’ (, line 7), including his own cooking skill. He suggests that the true Epicurean cook will understand ‘the difference between a horse-mackerel in winter and summer…for changes and movements create differences in foods (,21–3). The language is a parody of Epicurean technical vocabulary;Google Scholarsee Dohm, H, Mageiros: Die Rolle des Kochs in der griechisch-römischen Komödie (Munich, 1964), pp. 163–9.Google Scholar

³ On the stylistic connotations of these lines see Freudenburg, K, ‘Horace' Satiric Program and the Language of Contemporary Theory in Satires 2.1’, AJPh 111 (1990), 188–93,Google Scholar and The Walking Muse: Horace on the Theory of Satire (Princeton, 1993), p. 164, n. 91Google Scholar

⁴ Compare the imagery used by the anonymous Auctor ad Herennium in his famous discussion of the middle style at Ad Her. 4.16: ‘Qui in mediocre genus orationis profecti sunt, si pervenire eo non poterunt, errantes perveniunt ad confinii genus eius generis: quod appellamus dissolutum, quod est sine nervis et articulis; ut hoc modo appellem “fluctuans”eo quod fluctuat hue et illuc nee potest confirmate neque viriliter sese expedire. id est eiusmodi… ’ [Those who set out for the middle style, if they fail to reach it, swerve and arrive at the adjacent type of style, which we call disjointed, because it lack sinews and joints. For this reason I would call it wavering, because it surges back and forth and is unable to push forward with manly confidence. The following is of this type…]

⁵ Consider, for example, Quintilian Inst. 9.4.9 describing the essential qualities of arrangement, where again metaphors of tension and motion are prominent: ‘Quare mihi compositione velut amentis quibusdam nervisve intendi et concitari sententiae videntur. Ideoque eruditissimo cuique persuasum est, valere earn plurimum non ad delectationem modo sed ad motum quoque animorum. ’ [Wherefore, to me it seems that it is by compositio that our thoughts are strung tight and sent flying, just as if by hurling-straps or bowstrings. And for this reason all the most learned scholars are convinced that compositio is extremely valuable, not only for conferring pleasure, but also for stirring the soul.]

⁶ Cf. Inst. 9.4.17: ‘Nam neque illud in Lysia dicendi textum tenue atque rasum laetioribus numeris corrumpendum erat; perdidisset enim gratiam, quae in eo maxima est, simplicis atque inadfectati coloris, perdidisset fidem quoque. Nam scribebat aliis, non ipse dicebat, ut oportuerit esse illa rudibus et incompositis similia; quod ipsum compositio est. ’ [For it would never have done for richer rhythms to destroy that delicate, smooth fabric of language in Lysias, since he would have lost that winning charm (which he possesses in such abundance) deriving from his simple, unassuming tone. He would have lost his trustworthiness as well, for he was writing speeches for others (to deliver). He didn't deliver them himself. They were, then, by necessity somewhat crude and faulty in composition, a thing which itself is a compositional art.]

⁷ Cf. above, nn. 3 and 4.

⁸ Langford, P, Horace' Protean Satire: Public Life, Ethics, and Literature in Satires II (diss. Princeton, 1989), pp. 156–205.Google Scholar

⁹ In studying Horace' hexameters I have found the following metrical studies most helpful: Conrad, C., ‘Traditional Patterns of Word-Order in Latin Epic from Ennius to Vergil’, HSCPh 69 (1965), 195–258;Google ScholarDuckworth, G., Vergil and Classical Hexameter Poetry (Ann Arbor, 1969);CrossRefGoogle ScholarHarkness, A, ‘The Relation of Accent to Elision in Latin Verse, not Including the Drama’, TAPhA 36 (1905), 82–110;Google ScholarKiessling, A and Heinze, R, Q. Horatius Flaccus, Band II: Satiren (Berlin, 1961), pp. xxvii–xlv;Google ScholarKenney, E. J, The Ploughman' Lunch (Bristol, 1984), pp. xli–Ixiv;Google ScholarNilsson, N, Metrische Stildifferenzen in den Satiren des Horaz (Uppsala, 1952);Google ScholarPatzer, H, ‘Zum Sprachstil des neoterischen Hexameters’, MH 12 (1955), 77–95;Google ScholarSturtevant, E. and Kent, R, ‘Elision and Hiatus in Latin Prose and Verse’, TAPhA 46 (1915), 129–55.Google Scholar

¹⁰ On the preference of certain Stoics for a rugged verse-technique, see Freudenburg, Walking Muse [n. 3], pp. 132–9, and 150–62

¹¹ On the frequency of elision in Horace, Lucilius, Vergil, and Catullus, see Nilsson [n. 9], pp. 8–10.

¹² An obvious case of word-order imitating sense (mimetic syntax) is S. 2.8.42–3, describing the (in)famous eel served up by Nasidienus: affertur squillas inter murena natantis / in patina porrecta. The word-arrangement which results is mimetic, with the eel positioned directly among the ‘swimming prawns ’ (squillas…natantis) to suggest the appearance of the dish itself. On mimetic arrangement generally see Wilkinson, L. P, Golden Latin Artistry (Cambridge, 1963), pp. 65–6;Google Scholar and Lateiner, D, ‘Mimetic Syntax: Metaphor from Word Order, Especially in Ovid’, AJPh 111 (1990), 204–37.Google Scholar

¹³ The other is S. 1.3.20

¹⁴ Nilsson [n. 9], 20: ‘Der Grund der Abneigung gegen derartige Elisionen muss wohl mit Mueller darin zu suchen sein, dass, weil die elidierte ultima quadamtenus absorbeatur synizesi, eine Art Hiatus nach dem Paenultimavokal entsteht’.

¹⁵ The figures derive from Nilsson [n. 9] Table V, p. 203, discussed pp. 17–19. On mimetic syntax see above n. 12. The harsh elision in line 82 between flagitium and ingens can be taken as a compositional flagitium ('hocking circumstance’, or ‘offence against decent feeling’, cf. OLD flagitium 3a and 4a), a poetic counterpart to an equally offensive social gaffe described by the line.

¹⁶ See Nilsson [n. 9], Table IV, p. 202.

¹⁷ Ibid., Table VI, p. 203, discussed pp. 21–8.

¹⁸ Ibid., pp. 94–7.

¹⁹ Ibid., Table XI, p. 206, discussed pp. 57–86. Langford [n. 8], 165 points out that two of the four lines lacking one of the normal secondary caesurae in 2.4 (none lacks both) can be explained as creating ‘special effects ’ (what I call mimetic syntax): ‘in line 69, as Nilsson points out, the occurrence of the fourth and fifth trochaic caesurae characterizes the ceaseless flow of Catius ’ speech…in line 75, Catius himself seeks a special effect in bragging of a culinary invention. The metrical looseness allows for an imposing four-word line.’

²⁰ Nilsson [n. 9], Table X, p. 205, discussed pp. 61–2.

²¹ Ibid., pp. 36–7.

²² Harkness [n. 9] has shown that, when the second syllable in a pause-elision is, in fact, long, it is normally unaccented; if accented, it is, in most cases, a sentence-enclitic (and thus, only ‘apparently ’ accented, e.g. when the second element is et, atque, ac, aut, etc.) or an ‘unemphatic element of the sentence which would not have a marked sentence-accent ’ (91).

²³ Harkness [n. 9], 86 gives nine ‘apparent exceptions ’ to his rule in the complete works of Horace. He isolates the three cases in S. 2.3 (236 possideam; aufer; 283 ‘quid tarn magnum? ’ addens; 307 vitio? ’ ‘accipe) as the most exceptional: ‘If we take into account only the lines in which these elisions occur and in each case the four preceding and following lines, we find for these 27 lines a higher percentage of elisions than the average of Lucilius ’ (96). He goes on to note that ‘There does not seem to be any reason to assume that Lucilius strove to avoid a long accented syllable in the second syllable of pause-elision. This was doubtless one of the characteristics of his verse which led Horace to criticize so severely the style of his predecessor to whom he owed so much ’ (97).

²⁴ See above n. 15.

²⁵ They are: 2.3.18 (weak clash), 39, 48, 58, 74, 75, 85, 86, 94, 99 (weak), 120, 135, 148, 158, 166, 177 (especially harsh), 201, 202, 213, 214, 215, 217, 236, 237, 262, 267, 268, 301, 311, 314, 322.I have excluded all cases where an ’ apparent ’ clash is mitigated by the presence of a proclitic (relative pronouns, unemphatic adverbs, prepositions) and/or enclitic (personal and indefinite pronouns, certain uses of esse, cf. Nilsson [n. 9], 50), e.g. 53 est genus unum, 87 sive ego prove, etc.

²⁶ For the definition cf. Nilsson [n. 9], 158. Though closely related, Satzvers is not synonymous with end-stopping. For example, in the passage quoted below (5. 2.4.30–36) four of the seven lines are end-stopped (in Klingner' edition). All seven lines, however, fit the criteria of Satzvers.

²⁷ The numbers are from Nilsson [n. 9], Table XXIII, p. 212, discussed pp. 158–63.

²⁸ Cf. Duckworth [n. 9], 42 on the predictability of Catullus ’ hexameter patterns: ‘His epyllion on the marriage of Peleus and Thetis…has been praised for passages of astonishing beauty and at the same time criticized for its monotony. The monotony results largely from the poet' handling of the hexameter patterns. His metrical procedures differ not only from those of the other Republican poets, as we have seen, but are in most respects quite unlike those of any other Latin poet, with the possible exception of one or two writers of the Late Empire, especially the Christian poet Cyprian of Gaul. For example, Catullus ’ high percentage of the first eight patterns, 90.98, is surpassed by Cyprian with 91.06; only 30.63 percent of Catullus ’ 16-line units have the usual eight or more patterns... Catullus has the highest frequency of repeats, one every 7.0 lines (Cyprian, 7.9), and of repeats plus near repeats, one every 3.0 lines (Cyprian, 3.2).’

²⁹ Nilsson [n. 9], Table XXI, p. 211, discussed pp. 138–42. He says of these figures (p. 139): ‘In Buch II is nur eine relevante Abweichung, aber eine sehr kraftige, festzustellen: in Sat. 4 stehen von 55 starken Ip nur 17—fast nur die Halfte der durchschnittlichen Zahl—im Versinneren. ’ There are four lines in Klingner' edition of 5. 2.4 which have strong punctuation both within the line and at the line-end (lines 5, 14, 16, and 82). Nilsson takes no special consideration of these lines in computing his figures.

³⁰ These figures I have compiled from summary tables in Conrad [n. 9], 206, 208, 215, 219, 225, 244–5. Patzer [n. 9], using slightly different criteria, achieves virtually the same results.

³¹ The 95 patterned separations in S. 2.3 include: 17 with the first element directly before a strong caesura in the third foot and the second element at the line-end (type 1; cf. Conrad [n. 9], p. 208); six with the first element before a strong caesura in the third foot and the second element after the bucolic diaeresis (type la; cf. Conrad [n. 9], p. 215); six with the first element before a strong caesura in the second foot (type 2; cf. Conrad [n. 9], p. 219); 15 with the first element placed first in the line (type 3; cf. Conrad [n. 9], p. 225); 17 in enjambement (type 4; cf. Conrad [n. 9], pp. 244–45); 34 with the first element before a strong caesura in the fourth foot or the bucolic diaeresis (type 5; cf. Conrad [n. 9], p. 206). In S. 2.4 the separations break down as follows: 17 of type 1; two of type la; five of type 2; 12 of type 3; eight of type 4; 17 of type 5.

³² See Freudenburg, Walking Muse [n. 3], pp. 132–84.