Hostname: page-component-76fb5796d-x4r87 Total loading time: 0 Render date: 2024-04-26T11:01:36.447Z Has data issue: false hasContentIssue false

An Epicurean Survey of Poetic Theories (Philodemus On Poems 5, Cols. 26–36)

Published online by Cambridge University Press:  11 February 2009

Elizabeth Asmis
Affiliation:
The University of Chicago

Extract

If one wants to know what happened in Hellenistic poetic theory, Philodemus' survey of poetic theories in the fifth book of his On Poems is an excellent guide. Even though I the survey is well preserved, it has been neglected. Jensen, who published the first complete edition of On Poems 5 in 1923, did not discuss this part of the text; and it has been treated only briefly by others. This is a pity because, as Philodemus shows, the Hellenistic period was an era of great diversity and innovation in literary theory. Philodemus gives evidence of: (1) a refined and highly systematic critical vocabulary; (2) a new concern with verbal form; (3) a new notion of mimesis; and (4) in general, a great proliferation of theories that present alternatives to those of Plato and Aristotle. Hellenistic literary theorists studied Plato and Aristotle critically; some revised or elaborated their views, whereas others opposed them. Ancient poetic theory did not come to a standstill with Aristotle any more than philosophy did.

Type
Research Article
Copyright
Copyright © The Classical Association 1992

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Christian, Jensen, ed. Philodemos Über die Gedichte, Fünftes Buch (Berlin, 1923)Google Scholar. The survey has been discussed only by Pasquale Giuffrida, who erroneously thought the opinions were Stoic (L'Epicureismo nella letteratura latina nel I secolo A.C., v. 1 (Turin, 1940), pp. 146–81)Google Scholar; and by Greenberg, Nathan A. (The Poetic Theory of Philodemus (New York, 1990), pp. 94113)Google Scholar, who offers a translation and brief, useful comments. A new edition and translation of book 5 (with commentary) by Cecilia Mangoni is forthcoming in La Scuola di Epicuro, ed. by M. Gigante. Unless otherwise indicated, I have used Jensen's text and numbering.

2 Whereas PHerc. 1425 covers the whole of the extant columns, PHerc. 1538 covers only from col. 25 to the end of the book, thus supplementing the much better readings of PHerc. 1425. In addition to the actual papyri, two transcriptions (Naples and Oxford) of each papyrus are available; they are reprinted in Jensen's edition.

3 Col 26.19–23: …τ⋯ς παρ⋯ Ζ⋯νωνι | δ⋯ξας ⋯πικ⋯ψαντες | ἤδη [με]μηκυσμ⋯νον | τ⋯ σ⋯γ γραμμα καταπα⋯|σομεν.

4 Philodemus appears to refer to the original text (which he copied out) at col. 32.22–3, when he suggests a possible omission by the scribe. Part of Philodemus' earlier list is preserved in PHerc. 228 (4 Jensen, 3 Mangoni); see below, notes 12 and 64. Jensen (p. 94 of his edition) conjectured that PHerc. 228 was part of the same papyrus roll as PHerc. 1425; the latter has the subscript ‘Philodemus’ On Poems 5' and has lost the beginning of the book. In her recent edition of PHerc. 228 (‘Il PHerc. 228’, CErc. 19 (1989), 179–86)Google Scholar, Cecilia Mangoni doubts (p. 186) whether PHerc. 228 belongs to PHerc. 1425 or even to On Poems. However, the obvious link of frs. 4 and 6 (3 and 1 in Mangoni's numbering) of PHerc. 228 with PHerc. 1425 is strong evidence that PHerc. 228 belongs to On Poems 5.

5 Philodemus knew Zeno's teachings at first hand, as well as from reports (oral and written) by other students. For example, in De signis Philodemus draws on his own conversations with Zeno, on a report by another student of Zeno, Bromius, and on a written summary of Zeno's arguments by Demetrius (cols. 19.4–11 and 28.13–14). As illustrated by De signis, Zeno typically stated his opponents' views in summary form before responding to them.

6 Col. 36.10–14: τοὺς δ⋯ λοιποὺς ⋯κ | τ⋯ν πρ⋯τερον διητασ|μ⋯νων ῥ⋯ιδιον κατ⋯ | τ⋯ διαπ⋯πτουσιν ⋯πιβλ⋯|ψαι.

7 Cicero, , Tusculan Disputations 3.38.Google Scholar

8 Cicero, , De natura deorum 1.59Google Scholar (‘distincte, graviter, ornate’): cf. Academica 1.46.

9 Following Wilhelm, Crönert (Kolotes und Menedemos (Munich, 1906), p. 119)Google Scholar, I have suggested in ‘Epicurean Poetics’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, ed. Cleary, J., v. 7 (1991), 6393Google Scholar that Sextus Empiricus drew on Zeno's On Grammar for the arguments that he sets out at Adv. math. 1.277–98.

10 Col. 26.25–30: … ὅταν σ⋯νθεσις | ᾖ τ⋯ρπ[ουσ]α τ⋯ν ⋯κο|⋯ν ἢ κ[αλ⋯]ς φερομ⋯|νη κα⋯ τ[⋯ν] δι⋯νοιαν | κεκρατημ⋯νως ⋯κφ⋯ρου|σα….

11 Col. 27.6–10: ⋯ δ⋯ | σ⋯νθεσις λ⋯ξεων ⋯ναρ|γ⋯ς κα⋯ ⋯μφατικ⋯ς τ⋯ν | ὑποτεταγμ⋯νην δι⋯νοι|αν [σ]ημα⋯νουσα[[ν]].

12 Col. 28.7–10: ⋯ δ⋯ σ⋯νθε|[σ]ις σαφ⋯ς κα⋯ συντ⋯μως | 〈τ⋯ν ὑποτεταγμ⋯νην δι⋯νοιαν διασαφο⋯σα〉 | σὺν τωι τηρειν τ⋯ν [ποι]|ητικ⋯ν | χ[αρα]κ[τ⋯]ρ[α… Jensen has supplied 〈τ⋯ν ὑποτετααγμ⋯νην δι⋯νοιαν διαραφο⋯σα〉 between lines 8 and 9 on the basis of PHerc. 228 fr. 4 (Mangoni 3 B), where the opinion is summarised as: …σα]φ⋯ς κα⋯ συ[ντ⋯ως δια]|σαφουσαν σὺν τ⋯[ι μ⋯] ⋯κβα⋯νειν τ⋯ν [ποι]|ητ[ι]κ[⋯ν]|χ[αρα]κ[τ⋯]ρ[α. I suggest that Philodemus omitted the participial construction in his survey, as something that could readily be understood; hence there is no need for Jensen's supplement.

13 Sextus Empiricus, Adv. math. 1.79; cf. 1.248. On Crates and the ‘critics’ in general, see further Asmis, , ‘Crates on Poetic Criticism’, Phoenix 46 (1992), 4475.CrossRefGoogle Scholar

14 See col. 26.7–11.

15 On Poems 5, col. 25.23–9.

16 PHerc. 1676, tr., 3, col. 17(6).2–9 Sbordone: … κα⋯ τ⋯ τ⋯ν μ⋯ν | [⋯πιφαι]νομ⋯νην[ε]ὐφων⋯|αν ἴδιον[εἶν]αι, τ⋯ δ⋯ νο⋯{ι}|ματα κα⋯ [τ]⋯ς λ⋯ξεις ⋯κτ⋯ς | εἶναι κα⋯ κοιν⋯ συν⋯γεσ|θαι δεῖ[ν, πα]ρ⋯ π⋯σι μ⋯ν ὡς | ⋯ν [στ⋯λ]ηι μ⋯[ν]ει τοις κρι|τικοῖ[ς]…

17 Philodemus repeatedly describes sound as an ‘epiphenomenon’ of the composition, for example, at On Poems 5, cols. 20.28 and 21.31. This is not an ‘epiphenomenon’ in the modern sense of having no causal force.

18 PHerc. 1676, tr. 3, col. 18(7).9–10 Sbordone.

19 The verse also exemplifies the ‘smooth’ style of composition, as defined by Dionysius of Halicarnassus (De compositione 23, p. 234 Roberts): it uses ‘euphonious’ words that move along (κεκινησθαι, ϕ⋯ρεσθαι) continuously, like things that flow (ῥ⋯ντα) and never stay still.

20 On Poems 5, col. 18.15–17: ταῖς ⋯γε[ν⋯]|τοις εὐϕων⋯αις τ⋯ν κ[ρι]|τικ⋯ν.

21 PHerc. 1676, tr. 3, col. 17(6).23–7 Sbordone.

22 PHerc. 1074+1081, fr. c, col. 2.5–11 Sbordone: τ⋯να γ[ε τ]ρ⋯πον, ὦ | Κορ⋯βαντες, ὅ[λωςτ]⋯ π⋯ημα | π⋯ντες οὐχ ὡ[ς] τερ⋯τισμα | κα⋯ κρο⋯μα νο[ο]⋯με[ν], ⋯λ|λ⋯ λ⋯ξεις ⋯κ το⋯ [π]ως συντ⋯|θεσθαι διαν⋯ημα σημαι|νο⋯σας I prefer Hausrath’s tentative ὅ[λως] to Sbordone’s ὂ[ν].

23 Plato similarly uses the term ‘Corybantes’ to describe people whose experience of poetry is irrational (Ion 533e–34a, 536c). ‘Longinus’ (On the Sublime 39.2) uses the noun ‘corybantism’ (κορυβαντισμ⋯ς) to designate the state of being driven out of one's wits by the sound of the flute.

24 On Poems 5, col. 20.13–21; see pp. 184–5 of Asmis, , ‘The Poetic Theory of the Stoic “Aristo”’, Apeiron 23 (1990), 147201.CrossRefGoogle Scholar

25 On Poems 5, cols. 24.33–25.4.

26 The name of the Stoic is preserved only as ‘…ων’ (col. 13.30). Philodemus' criticism occupies cols. 13.28–21.22.

27 Cols. 18.14–7 and 20.26–33.

28 Seneca, , Epistle 108.10Google Scholar. The whole passage is: ‘nam, ut dicebat Cleanthes, quemadmodum spiritus noster clariorem sonum reddit, cum ilium tuba per longi canalis angustias tractum patentiore novissime exitu effudit, sic sensus nostros clariores carminis arta necessitas efficit. eadem neglegentius audiuntur minusque percutiunt, quamdiu soluta oratione dicuntur; ubi accessere numeri et egregium sensum adstrinxere certi pedes, eadem ilia sententia velut lacerto excussiore torquetur.’

29 Philodemus, On Music 4 (col. 28.1–22): εἰ μ[⋯ τ⋯ι π]αρ⋯ Κλε⋯ν[θ]ει λ⋯|γειν [ἴσ]α θελ⋯σουσιν, ὅς ϕησιν | [⋯]με⋯νο[ν⋯] τε εἶναι τ⋯ ποητικ⋯ | κα⋯ μ[ονσ]ικ⋯ παραδε⋯γματα | κα⋯ του [λ⋯γ]ου του τ⋯ς ϕιλοσο|ϕ⋯ας ἱκαν⋯ς μ⋯ν ⋯ξαγ[γ]⋯λ|λειν δυναμ⋯νου τ⋯ θεια κα⋯ | ⋯[ν]θ[ρ]ώ[πινα], μ⋯ ἔχοντος δ⋯ | ψειλο⋯ των θε⋯ων μεγεθ⋯ν | λ⋯ξεις οἰκε⋯ας, τ⋯ μ⋯τρα κα⋯ | τ⋯ μ⋯λη κα⋯ τοὺς ῥυθμοὑς | ὡς μ⋯λιστα προσικνεισθαι | πρ⋯ς τ⋯ν ⋯λ⋯θειαν τ⋯ς των | θε⋯ων θ[ε]ωρ⋯ας, οὗ καταγελα|στ⋯τερον οὐ ῥ⋯ιδιον εὑρεῖν. | οὔτε γ⋯ρ α⋯ δι⋯νοιαι μ⋯ν οὐ|κ ὠϕελο[⋯]σιν, ⋯ταν δ⋯ μελω|δηθωσι[ν], ⋯ξ ⋯μ[ϕ]οτ⋯[ρ]ων ⋯ | παρ⋯ρ[μη]σις [γ⋯ν]εται. κα⋯ γ⋯ρ | ὑπ⋯ διανο[η]μ⋯των αὐτ⋯ν | γ⋯νετ᾽ οὐδ[⋯] μετρ⋯α, μετ⋯ δ⋯ | τ⋯ν μ⋯λων μ[ε]⋯ζων.

30 Diogenes Laertius 7.59.

31 Dionysius, De compositione 17 (pp. 170–2 Roberts); and Longinus, On the Sublime 39.4 and 41.1.

32 On Music 4, col. 28.23–35 Neubecker.

33 On Poems 5, col. 26.30–2: δι⋯ψευσται μ⋯ν τ[⋯ι] | μ⋯ τ⋯ρπειν ἦχον ⋯ν | συνθ⋯σει πο⋯ματος.

34 On Poems 5, cols. 23.36–24.3.

35 Cols. 26.32–27.6. Epideictic oratory, with its smooth and pleasing collocation of letters and its strong rhythms, is a plausible candidate for this kind of prose.

36 The author of Ad Herennium defines ⋯ν⋯ργεια. (demonstrate) as follows: ‘cum ita verbis res exprimitur ut geri negotium et res ante oculos esse videatur’ (4.68); cf. 4.69. Quintilian (8.3.61) points out that vividness (evidentia, repraesentatio) is something more than clarity (perspicuitas) and defines it as: ‘clare atque, ut cerni videantur, enuntiare’; cf. 6.2.32. Dionysius of Halicarnassus defines ⋯ν⋯ργεια at Lysias 7 as: δ⋯ναμ⋯ς τις ὑπ⋯ τ⋯ς αἰσθ⋯σεις ἄγουσα τ⋯ λεγ⋯μενα.

37 Translating ἔμϕασις by signiftcatio, the author of Ad Herennium (4.67) defines it as: ‘res quae plus in suspicione relinquit quam positum est in oratione’. The author goes on to divide significatio into various types, as produced by hyperbole, ambiguity, logical consequence, aposiopesis, and comparison. Similarly, ps.-Plutarch (De vita et poesi Homeri 2.26) writes that ἔμϕασις ‘extends what is said by a submerged meaning (δι᾽ ὑπονο⋯ας)’, that is, ‘provides a larger notion’. Suspicio corresponds to ὑπ⋯ονοια.

38 Quintilian 8.3.83 (‘altiorem praebens intellectum quam quern verba per se ipsa declarant’); cf. 9.2.3. At 8.3.83, Quintilian divides ἔμϕασις into two kinds: meaning more than one says, and meaning also what one doesn't say.

39 Quintilian 8.3.84; ps.-Plutarch, De vita et poesi Homeri 2.26; and Trophon, Περ⋯ τρ⋯πων Spengel v. 3, 199. The verse is at Od. 11.523. Another example is Strabo's claim (1.1.7) that Hera's reference to Oceanus in the Iliad (14.200–1) implies (⋯μϕα⋯νει) that it encircles the earth.

40 Related to the literary use is the logical use of ἔμϕασις to denote the type of implication in which ‘the conditional is true whenever the consequent is contained potentially (περι⋯χεται δυν⋯μει) in the antecedent’ (Sextus Empiricus, Outlines of Pyrrhonism 2.112). Sextus suggests that this type of implication presumably excludes duplicated propositions of the sort ‘if it is day, it is day’, since it is impossible for something to be contained within itself. The term was also used in Academic epistemology to denote an apparently true, or persuasive, presentation (Sextus Empiricus, Adv. math. 7.169).

41 In his article ‘ΕΜΦΑΣΙΣ in Ancient Literary Criticism and Tractatus Coislinianus c. 7’, Maia 40 (1988), 125–9Google Scholar, Ian Rutherford cites the occurrence of ἔμϕασις in ps.-Aristides, Art of Rhetoric 1 (c. 119, Schmid 46–8) as the earliest unambiguous example of the sense ‘force’. Ps.-Aristides lists σϕοδρ⋯της, ἔμϕασις, and τραχ⋯της as closely related virtues of style. The examples of ἔμϕασις cited by ps.-Aristides do indeed support some such sense as ‘force’. I suggest that the meaning is, more precisely, ‘expressiveness’ (or ‘impressiveness’), and that this is a natural extension of the basic sense, ‘creating an impression’ or ‘expressing’ something in a non-literal way.

42 Demetrius (On Style 50–1) points out that a more vivid word is ⋯μϕατικώτερον, as well as more powerful. Similarly, he notes (On Style 212) that repetition can produce vividness and ‘greater ἔμϕασιν’.

43 The importance of ⋯ν⋯ργεια in painting is highlighted by the numerous famous anecdotes of how lifelike a painting is – so lifelike, for example, as to make birds peck at painted grapes (Pliny, N.H. 35.65–6). From Xenophon on, writers value the expression of character and emotion in representations of the face and body; and the term ⋯μϕα⋯νειν (along with others) was used to refer to this type of expressiveness (e.g. Plutarch, Alexander 1.3). When Pliny says of Timanthes that ‘only in his works more is always understood than is painted’ (‘in unius operibus intelligitur plus semper quam pingitur,’ N.H. 35.74), he is praising him for the quality of ἔμϕασις. Perhaps the most poignant example of ἔμϕασις in ancient art is Timanthes' way of portraying the extreme sorrow of Agamemnon at the sacrifice of his daughter: whereas Timanthes showed all the others grieving openly, he veiled Agamemnon's face (ibid., 35.73, cf. Cicero, Orator 74)).

44 Rhet. 3.2, 1404b1–4 and Poet. 22, 1458a18. What creates elevation is strange adornment, such as ‘glosses’ (unusual terms), metaphor, lengthening, and ‘everything contrary to proper [language] (παρ⋯ τ⋯ κ⋯ριον)’ (Poet. 1458a21–3, cf. Rhet. 1404b6–11).

45 Poet. 1455a21–6, cf. 1462a17.

46 Cicero, Orator 79; cf. De oratore 1.144, 3.37, and 3.91. The Latin translation of κατασκευ⋯, ornatus, involves a shift of meaning that has tended to reinforce the impression that κατασκευ⋯ is a piece-by-piece application of ornaments. κατασκευ⋯ is, properly, the artistic preparation or fashioning of the whole work.

47 Diogenes Laertius 7.59. Diogenes Laertius may have derived this list from Diogenes of Babylon, whom he cited just previously. In substantial agreement with the Stoics, Dionysius of Halicarnassus recognised three essential, ‘underlying’ virtues of style, purity, clarity, and conciseness, together with a host of ‘additional’ (⋯π⋯θεται) virtues and ‘the fitting’ (τ⋯ πρ⋯πον) as the ‘most authoritative’ virtue (Letter to Pompeius 3, cf. Thucydides 22–3). Dionysius lists (⋯ν⋯ργεια as the first of the optional qualities (Letter to Pompeius 3), but does not mention ἔμϕασις in his list. In Partitione oratorio (19), which is Academic in origin (139), Cicero offers an alternative list of five stylistic virtues: dilucidum (σαϕ⋯ς), breve (σ⋯ατομον), probabile (πιθαν⋯ν), illustre (⋯ναργ⋯ς), and suave (⋯δ⋯). These qualities are not divided into necessary and optional.

48 For example, René Wellek and Austin Warren write that ‘we may charge older literary study with treating externally and superficially’ the stylistic features of image, metaphor, symbol, and myth, and that these were viewed ‘for the most part as decorations, rhetorical ornaments’ and ‘detachable parts of the works in which they appear’ (Theory of Literature, 3rd ed. [Harcourt Brace Jovanovich, 1977], p. 193).Google Scholar

49 On Poems 5, cols. 3–5. Jensen restored the name of Heraclides in fr. 2 of his edition, but held that Neoptolemus was responsible for the theory that Philodemus reports subsequently. Jensen changed his mind in his article ‘Herakleides vom Pontus bei Philodem und Horaz’, Akademie der Wissenschaften, Berlin, Sitzungsberichte, philos.-hist. Kl., 1936, 292320Google Scholar. I agree with his revised view that Philodemus criticises Heraclides from fr. 2 to col. 9.9.

50 Col. 27.10–12: κοι|[ν]⋯ τ[⋯ς ⋯σ]τι κα⋯ λ⋯γου παν|τ⋯ς ⋯ρετ⋯[[ς]].

51 In this criticism (col. 27.17–33), Philodemus mentions conciseness as a third requirement along with vividness and suggestiveness; it may be taken as implied by the other two qualities.

52 On Rhetoric 4, col. 15.15–18 (Sudhaus v. 1, 175): κα⋯ π⋯σα τ⋯χνη | ϕων[⋯]ν οὐ δ⋯ναται προ[⋯εσ]|θαι στερ[η]θεῖσα τ⋯ς ⋯κ τ⋯ν | μεταϕορ⋯ν εὐχρησ[τ⋯ας].

53 Aristotle held that ‘by far the most important thing’ about the use of individual words in poetry is the use of metaphor (Poet. 1459a5–6). Metaphor has even greater scope in rhetoric because prose speakers, unlike poets, use other kinds of strange language only rarely; rhetorical speakers are free to use metaphors, just as all people do in conversation with each other (Rhet. 3.2, 1404b28–34, cf. 1405a4–10). In agreement with Aristotle, Quintilian notes that metaphor is ‘the greatest adornment of speech’ (‘in qua maximus est orationis ornatus’, 8.2.6).

54 Rhet. 3, chapters 10–11, esp. 1411b22–3.

55 Demetrius, On Style 222. Theophrastus appears to be elaborating on Aristotle's discussion of wit (τ⋯ ⋯στεῖον).

56 Demetrius, On Style 57 (particles expressing mourning); 61, 140, and 212 (repetition implying multiple possessions, charm, greater impressiveness); 63 (syndeton implying countless numbers); 105 (a combination of unpleasant sounds implying greatness of a person); 130–1 (witticisms expressing grim seriousness); 158 (a fable expressing charm); 171 and 259 (humour implying character and expressing force); 274 (a comparison giving the impression of beauty); 285 (an allegory implying political weakness); 300 (improvisation implying anger). Dionysius of Halicarnassus gives numerous examples of how certain stylistic features ‘imply’ or illustrate something about the author or the style; e.g. Lysias 12, Demosthenes 5, 9, 48, 50. The author of De vita et poesi Homeri points out that in Homer ‘thoughts are implied (⋯μϕα⋯νεται) through riddles and myths’, with the comment that what is signified δι᾽ ὑπονο⋯ας is more appealing than what is said openly (2.92); see also his use of ⋯μϕα⋯νει at 2.68 and 2.71. ‘Longinus’ (On the Sublime) illustrates how asyndeton produces an ἒμϕασις of vigorous action (19.2) and how hyperbaton creates an impression of vigour and improvisation (22.3).

57 On the connotation of individual words, see esp. Aristotle's analysis of beautiful words at Rhet. 1405b6–34 and Theophrastus' elaboration of this view, as reported by Demetrius (On Style 173–4) and Dionysius of Halicarnassus (De compositione 16, pp. 164–6 Roberts).

58 Zanker, G. argues in his article ‘Enargeia in the Ancient Criticism of Poetry’ (Rheinisches Museum 124 [1981], 297311)Google Scholar that ‘⋯ν⋯ργεια was considered particularly exemplified in poetry, even by the rhetoricians’ (p. 300). Zanker also suggests that the term was borrowed by literary critics from contemporary philosophy, in particular Epicureanism (p. 309). There are certainly links (which are worth exploring further) between the literary and philosophical uses of both ⋯ν⋯ργεια and ἒμϕασις; but ⋯ν⋯ργεια has a development in literary theory that is largely independent of philosophical theory.

59 For definitions of ‘allegory’, see, for example, Ad Herennium 4.46, and Quintilian 8.6.44.

60 SVF 2.908 and 909 (including: ἄλλον τιν⋯ς συμβ⋯λου ποιο⋯ντ᾽ ἔμϕασιν). Chrysippus also interpreted many non-allegorical poetic statements as ‘implying’ philosophical truths. For example, he understood numerous verses in Homer and other poets to imply that the heart is the unitary seat of the desires, emotions, and reason (see esp. the use of ⋯μϕα⋯νει at SVF 2.904–5). Another type of allegory consists of proverbs that serve as a ‘symbol’ of something not said, for example, ‘the cicadas will chirp from the ground’ (Demetrius, On Style 243).

61 On Poems, PHerc. 1676, tr. 3, col. 2 (fr. 2) Sbordone.

62 Col. 28.18–32, including (29–32) οὔτε | τ⋯ς συνχωρουμ⋯νης | ἃπασι τοῖς νοουμ⋯νο[ις] | ⋯ρμ⋯ττειν δοκο⋯σης. δοκο⋯σης shows that Philodemus again appeals to the views of others.

63 In his criticism of the second opinion, Philodemus remarks that ‘these [theorists] and many who were previously examined and who will be considered later are far from characterising the sort of thought that is required to underlie [the words]’ (cols. 27.34–28.7, reading πολλ[ο⋯] instead of Jensen's πολλ⋯ at col. 28.2).

64 Col. 28.33–6; col. 29.3–7; and cols. 29.36–30.4. Fr. 4 (3 Mangoni) of PHerc. 228 contains an additional opinion. Immediately after the third view (see above, n. 12), it continues: ‘some [suppose that good composition] contains wise thoughts, others that it contains these thoughts by means of an elaboration that delights the hearing, others…’. In his critical survey, Philodemus does not bother to list the demand for wise thought only.

65 Rep. 607b.

66 See n. 26 above.

67 Col. 29.3–7: [ἥ γ]ε δι⋯νοια[ν ὠϕ⋯λιμον | εἰ κ]α⋯ μ⋯ σοϕ⋯ν, κε[κρατη]|μ⋯νως κα⋯ πρ⋯ς τ[⋯]ν ⋯|κο⋯ν ⋯μϕατικ⋯ς ⋯κϕ⋯|ρουσαν. I supply [ἥ γ]ε instead of Jensen's [τ⋯ι γ]ε. The accusative ⋯κϕ⋯|ρουσαν may be accounted for by assuming an elliptical suppression of σ⋯νθεσιν (cf. col. 30.1). For further discussion of the fourth and fifth opinions, see pp. 5–7 and 10, Asmis, , ‘Philodemus's Poetic Theory and On the Good King According to Homer’, Classical Antiquity 10 (1991), 145.CrossRefGoogle Scholar

68 On Music 4, col. 26.4–7.

69 Ion 540d–e.

70 Col. 29.22–7. νουν (line 27) is probably a pun.

71 Col. 29.28–33.

72 Cols. 29.36–30.4: ⋯ δ⋯ | σ⋯[νθεσιν λ⋯ξ]εων προσ|δι[δ⋯σκουο⋯ν τι περιτ]τ[⋯|τερον δι⋯ ποι⋯μ]ατος [ἣ τ]α[⋯]|τηι γ᾽ ὡμοιωμ⋯νην. Jensen's supplements are secured by their occurrence in the rest of the column.

73 Anatomy of Criticism (Princeton University Press, 1957), pp. 23–4.Google Scholar

74 Col. 30.25–8: τ⋯ εὖ μεμιμησ|θαι τ⋯ [Ὁ]μ⋯ρου κα⋯ τ⋯ν | ⋯μο⋯ω[ς π]αραδεδομ⋯|νων.

75 On the Sublime 13.2–4 and 14.1 (τ⋯ς ψυχ⋯ς ⋯νο⋯σει πως πρ⋯ς τ⋯ ⋯νειδωλποιο⋯μενα μ⋯τρα).

76 On the Sublime 13.4.

77 Col. 30.28–32.

78 Col. 30.32–6.

79 Col. 31.3–17.

80 Col. 31.17–33.

81 Quintilian 10.2.14–18; cf. Seneca, , Epistle 114.1720.Google Scholar

82 See esp. On the Sublime, 13.2–4 and Quintilian 10.2.9–10.

83 De imitatione 417, including: π⋯ρεστι κα⋯ σο⋯…οὐκ ⋯ξ⋯τηλον χρ⋯νω γενησομ⋯νην εἰκ⋯να τυπο⋯ν ⋯λλ᾽ ⋯θ⋯νατον τ⋯χνης κ⋯λλος. Many versions of the story circulated, including one by Cicero (De inventione 2.1–3).

84 Cols. 31.35–32.2: τ⋯ λ⋯ξιν προσϕ⋯|ρεσθαι πρ⋯πουσαν τοις | εἰσαγομ⋯οις προσώ|ποι[ς].

85 Ion 540b: ἃ πρ⋯πει…⋯νδρ⋯ εἰπεῖν κα⋯ ⋯ποῖα γυναικ⋯, κα⋯ ⋯ποῖα δο⋯λῳ κα⋯ ⋯ποῖα ⋯λευθ⋯ρῳ, κα⋯ ⋯ποῖα ⋯ρχομ⋯νῳ κα⋯ ⋯ποῖα ἄρχοντι.

86 Cicero, De officiis 1.97: ‘…turn servare illud poetas quod deceat dicimus, cum id, quod quaque persona dignum est, et fit et dicitur.’

87 Col. 32.2–6 (continuing the text cited in n. 84): …[κοιν⋯ν γ⋯ρ] κα⋯ | τ⋯[ν] πε[ζ]ων ⋯στι λ⋯|γων, εἰ δ⋯ μ⋯γε, τω[ν] ἱστο|ρικ⋯ν ἢ τ⋯ν γε δ[ι]αλ⋯|γων. Jensen's supplement κοιν⋯ν is strongly supported by the context, and by Philodemus' fondness for this type of objection (see, for example, line 32 of the same column).

88 See, for example, Ad Herennium 1.8 and 3.11.

89 Plato first elaborated this distinction in Rep. 3, see esp. 393a–b.

90 Col. 32.6–10.

91 Col. 32.20–8.

92 Ion 540b–e.

93 Col. 32.28–32.

94 Col. 32.10–14. Philodemus adds that it is better to say that a poet should ‘imitate’ fitting diction. His reason is, presumably, that this formulation does not commit the poet to reproducing the very speech of the gods and heroes.

95 Col. 32.16–20: τελε⋯ω[ς] | δ⋯ [μαν]ικ⋯ν τ⋯ π[αρα]|ψηλαϕ[⋯]ν ⋯μοι⋯τητα | λ⋯ξεως τοῖς δηλουμ⋯|νοις πρ⋯γμασιν.

96 Rhet. 3.7, 1408a10–13.

97 See, for example, Ad Herennium 1.8, 1.13, 3.11.

98 On the Stoics, see Hülser 643 and 644, and Asmis, ‘Crates on Poetic Criticism’ (cited above, n. 13). It is possible that Philodemus chose the rare word παραψηλαϕ⋯ν to reflect the Stoic view that sounds imitate things in the way that things ‘touch’ the senses (Hülser 644.13–15).

99 On Poems 4, col. 8.10–13 Sbordone; and On Rhetoric 4, col. 5.12–16, v. 1, p. 150 Sudhaus.

100 Cols. 32.33–33.5: [ἥ] τε λ⋯|γουσα τ⋯ δ⋯νασθαι τ⋯ν | ⋯κο⋯οντα [κα]⋯ πα[ραδο/ξολογ]⋯α[ις προσαγα|γ⋯σ]θαι κα⋯ ⋯ τ⋯ δ[εῖν μ⋯]|νον κιν⋯σαι τοὺς πε|παιδευμ⋯νους. I have adopted Mangoni's reading in 32.33. A crucial chunk of the ninth definition is missing from the papyri; but πα[ραδο|ξολογ]⋯α[ις] is supported by πα|[ραδο]ξολογ⋯αις (Mangoni) at col. 32.23–4, and [προσαγα|γ⋯σ]θαι is plausible.

101 Col. 33.8–15.

102 Cols. 32.32–33.7.

103 Col. 33.15–21.

104 Cols. 33.21–34.2: …⋯ο⋯κασιν ‹δ⋯› | τ⋯ πολλ⋯ τ⋯ν εὐτελ⋯ν | περιπετε⋯αις δ⋯ κα⋯ πα|[ραδο]ξολογ⋯αις κα⋯ παθη|τ[ικοῖ]ς λ⋯[γ]οις κινο⋯ν[τα π]ρο|κρ⋯νειν τ⋯ν ἄκρως μ⋯ν πε|ποιημ⋯νων, ἥκιστα δ⋯ | τα⋯τα προσϕερομ⋯νων. | κα⋯ συμπαθο⋯σι [μ⋯]ν κα⋯ | κινο⋯νται μ⋯λιστα προσ[⋯ν]|τω[ν τι]ν⋯ν κα⋯ το⋯ [λ⋯]|γου δι[ανο]ημ⋯των, ⋯ δ᾽ ⋯[ρε]|τ⋯ το⋯ ποι⋯ματος [⋯]ν τ⋯ι | δι⋯ τ⋯ς κατασκευ⋯ς παρε|πι[ϕ]αινομ⋯νωι πε[ριτ]τ⋯ι | γ⋯[γονεν αὐτ]οῖς. I follow Mangoni at 33.22–4.

105 See esp. Poet. 11, also 9, 1452a1–4.

106 The three opinions are stated at: cols. 34.3–5 (⋯ρ[ετ⋯]ν εἶναι ποι|ητο⋯ τ⋯ δ⋯[ν]ασθαι παν | πο⋯η[μ]α σ[υνθ]εῖν[α]ι καλ⋯ς), 34.24–6 (τ⋯ δ⋯νασ|θαι σ[υνθεῖ]ναι πο⋯ησιν ⋯ρετ⋯ν ἔχουσαν), and 35.17–19 (τ⋯ ⋯ν παντ⋯ γ⋯νει ποι⋯|σεως τ⋯ πρ⋯πον δ⋯νασ|θαι τηρεῖν).

107 Cols. 34.6–13, 34.27–34, and 36.8–10.

108 Philodemus criticises Neoptolemus' distinction at cols. 11.26–12.17 of On Poems 5. Posidonius also defined ‘poetry’ as a logical subdivision of ‘poem’: a ‘poem’, he held, is ‘metrical or rhythmical diction, with elaboration, going beyond prose form’; and ‘poetry’ is ‘a poem with meaning (σημαντικ⋯ν πο⋯ημα), containing an imitation of divine and human things’ (Diogenes Laertius 7.60, fr. 44 Edelstein–Kidd). Lucilius (339–47 Marx), Varro (Parmeno fr. 398 Bücheler), the scholiast on Dionysius Thrax (449.24–6 Hilgard), and the grammarian Diomedes (Ars grammalica, GL 1, 473.17–20 Keil) also distinguish between ‘poem’ and ‘poetry’. See further Asmis, , ‘Neoptolemus and the Classification of Poetry’, Classical Philology 87 (1992), 206–31.CrossRefGoogle Scholar

109 Col. 35.7–15.

110 Philodemus' omission of this step may be compared to his omission of the simple demand for ‘wise thought’ earlier (see n. 64 above).

111 Col. 34.13–24.

112 PHerc, 460, tr. 2, frs. 17, 19, and 20, col. 1 Sbordone.

113 Hippias maior 293e.

114 Col. 35.20–2.

115 Orator 70–4, cf. 123.

116 Col. 35.22–32. πραγμα may also be understood as ‘act’.

117 Pp. 13–17 of ‘Philodemus's Poetic Theory and On the Good King According to Homer’ (see n. 67 above).

118 An earlier version of this paper was read at Boston University and the Center for Literary and Cultural Studies at Harvard University. I am grateful to the participants for their comments and suggestions.