Bryn Mawr Classical Review

Table of Contents

The book of Enzo Passa, a modified version of his 2004 'tesi di laurea', presents an in-depth study of the transmission of the surviving fragments of Parmenides of Elea's On Nature (Περὶ φύσεως) and of its language. Passa successfully deals with the convoluted history of Parmenides' text, the major features of its language and its relation with the earlier epic tradition, halfway between imitation and innovation. His main conclusion is that, due to the new character of his poetry, Parmenides' language, although drawing from the pool of Eastern Ionic epic poetry, shows various innovative features, some of which pertain to the more recent poetic traditions of the Western Greek world.

The book is divided in three parts, preceded by an introduction, where Passa gives a general outline of the testimonies of Parmenides' life and describes the various sources of his philosophical thought. Passa endeavours to dismiss any Attic influence from Parmenides' poetry and thought, a question that is relevant for the subsequent discussion of the language.

Part I addresses in detail the problems of transmission of the poem. In chapter 1, Passa argues that the core of the tradition must have been an Athenian copy linked to Plato and his Academy, which enjoyed an enormous prestige but irremediably underwent several mystifications by the Neoplatonists, who read Parmenides in the light of their Master's thought. Passa proves that the text transmitted by Simplicius is not as reliable as the common opinion suggests, since it offers some obvious deviations from the original (e.g. εὐκυκλέος, B.1. 29, vs. the more plausible εὐπειθέος).

In the second chapter Passa studies six passages discordantly transmitted by the manuscripts. Firstly, the two versions of B 8.38 are considered, and Passa ascertains that Plato is trying to Pythagorize the text by replacing Doric τελέθει for epic ἔμεναι. Next Passa argues for Aristotle's παρίσταται (B 16.2) as against Theophrastus' παρέστηκεν. As for B 1.29, he advances strong arguments in favour of ἀτρεκὲς (ἦτορ) over ἀτρεμές and in favour of (Ἀληθείης) εὐπειθέος over εὐκυκλέος. Passa convincingly demonstrates that the best attested readings of the manuscripts in B 1.25, ἵπποις ταί and ἵπποις τε, conceal the true text, ἵπποις τ' αἴ, showing that Sextus Empiricus still read a psilotic text. Fifth, in B 8.4 Passa favours ἐστι γάρ (cf. B 8.33) and μουνογενές as the original text. Finally, the non-Ionic crasis in ταὐτόν τ' ἐν ταὐτῷ τε μένον καθ' ἑαυτό τε κεῖται (B 8.29), as opposed to ἑωυτῷ (B 8.57), τωὐτόν (B 8.57 and .58) is justified as follows. The original Parmenidean formula originally used to describe the Being became nuclear to the theory of Ideas in the Platonic school. Accordingly, the overt Neoplatonic reading of Parmenides could easily lead to modify the original to a more 'Platonic' and Attic version (for other Attic features, see below).

In Part II, chapter 1 investigates the innovations in Parmenides' diction as opposed to the earlier epic tradition. For instance, in borrowing some formulaic expressions from the Homeric poems, Parmenides neglects the initial [w] which is still preserved in the original. Passa devotes also a long study to three etymologically related words in which Ionic quantitative metathesis occurs. Parmenides exhibits a new artificial disyllabic scansion in χρεώ (B 1.28, < xrhw/). As for the four occurrences of χρεών (always in the weak part of the foot), Passa argues that at least in 8.45 the reading χρεόν (again an innovation) is in fact a lectio difficilior, since χρεών could be easily accounted for as an Atticism of the tradition. Furthermore, in 8.54 Parmenides uses the same wording of 8.45, πελέναι χρεόν, and therefore χρεών in the former should be changed into χρεόν. While in Homer both ancient χρεῖος (=χρῆος) and analogical χρέος are found, Parmenides employs only the latter (B 8.9). Finally, in Passa's view, the form μόνος (B 8.1), guaranteed by the metre, must be regarded as a borrowing from the Euboean epic tradition and the new syllabifications of the muta cum liquida clusters (e.g. ἱκνεῖσθαι , B 8.46) became more frequent under the impact of the compositional style of Western lyric, mainly Stesichorus'.

In the second chapter, Passa focuses on some forms which are remarkable on several accounts: accus. pl. σφας (B 1.12) with short alpha, which must be an Ionic poetic innovation; the adjective ἐπιδεές (B 8.33) with trisyllabic scansion; the perfect πλῆνται with no reduplication (B 1.13), which is to be connected with πίλναται; and πλάττονται (B 6.5), which is explained by Passa as an Attic correction for a new Ionic secondary form πλάσσομαι = epic πλάζομαι.

In Part III, Passa reviews the possible influences of other poetic traditions on Parmenides. Chapter 1 is devoted to a leitmotif of the book: the Attic features found in the text as it stands today are invariably secondary and do not belong to the original text. These features are -α for -η (αἰθερίαν, καθαρᾶς in B 10.1 and .2), contraction in δοκοῦντα (B 1.31) and φοροῦνται (B 6.6), for expected Ion. δοκέοντα/δοκεῦντα and φορέονται/φορεῦνται, and the Attic crasis in ταὐτόν etc. Although most of these changes cannot be dated with certainty, Passa suggests that the long vowel in μεῖζον (B 8.44) and some cases of initial aspiration (ἀφ' ὁδοῦ B 6.2, etc.) must have entered the text in early times, as in the Homeric poems. Finally, ὄν (8.57) and περῶντας (B 1.32) should most probably be changed into Ionic ἐόν and περεόντας (with synizesis).

In the last chapter, evidence is put forward that some linguistic features must be credited to the Western lyric tradition. Its influence is particularly widespread in the vocabulary, e.g. ὑδατόριζον (B 15a), γέννα (B 8.6), ἔλεγχος (B 7.5), but it also affects morphology, e.g. ἔγεντο (B 8.20) and the infinitives φῦν (B 8.10) and μιγῆν (B 12.5), which are unmistakably connected to Stesichorus' εἶν (= εἶναι). The core of the chapter is devoted to the study of words with -α long, e.g. συνάορος (B 1.24), δίκρανοι (B 6.5) and εὐαγέος (B 10.2), probably coming also from choral lyric.

The book ends with some conclusive remarks and two indexes (Greek words and quoted passages).

We offer here some minor criticisms. Although typos are infrequent,¹ references to epigraphic corpora are somewhat chaotic.² As for the content, texts are not systematically translated, which could discourage readers with little command of Greek. There is surprisingly little reference to the dialectal material of other Phocaean colonies, particularly those of Hispania, which, for all its paucity, shows some parallels to features discussed by Passa. To the Eleatic examples of feminine genitive plural in -ῶν for expected -έων quoted by Passa, should be added those of the coinage from Emporion.³ For the sake of completeness, the infinitives φῦν (B 8.10), μιγῆν (B 12.12) and πελέναι (B 8.11 and .45), which are mentioned here and there, could have been the object of more extended study. A section about ἐάσσω (B 8.7) could have been added.⁴

In a work of so broad a scope, where a new interpretation of data can be found on almost every page, the reader cannot always adhere to the author's view. I briefly include here a few minor disagreements.

Passa dallies here and there with a so-called sociolinguistic explanation for unexpected variations in the data (p. 87. 91 and 147). Beyond the fact that hypothesis of this sort are founded on intuitions and speculations with no support on real data, it seems unlikely that low prestige variants of any kind could have entered a poem such as Parmenides' On Nature.

Passa has persuasively argued that Aristotle's παρίσταται is better than Theophrastus' παρέστηκεν, but his explanation of παριστᾶται as a by-product of *παριστάεται lacks parallels. Data from Ionic inscriptions and literary authors show that contract inflection did not easily penetrate the middle voice in -μι presents (we never find 3rd sg. *τιθεῖται, *διδοῦται or *ἱεῖται). Moreover, the allegedly contract παριστᾶται contrasts with athematic συνιστάμενον in B 4.10. Since Ionic shows 3rd pl. τιθέαται, διδόαται, we would also expect *ἱστάαται, eventually contracted to *ἱστᾶται (cf. active ἱστᾶσι < *i(sta/asi with a second long α). It is thus possible that the short alpha of 3rd sg. ἵσταται was lengthened after the analogy of 3rd pl. *ἱστᾶται.

In the chapter devoted to the syllabification of muta cum liquida clusters (τρ, πρ, κρ, etc., pp. 89-92), Passa includes among the innovative syllabifications (πα.τρός) examples like δὲ κρίσις (B 8.15), σε βροτῶν (B 8.61), etc., and among the traditional ( πατρός ), τε χρόα (B 8.41) and δὲ φλογός (B 12.2). But this inclusion is misleading. In Homeric verse heavy quantity is rarely obtained in the weak part of the foot by a word-final vowel followed by an initial muta cum liquida cluster.⁵ According to Passa's Table on p. 90-1, Parmenides shows no exception to this rule. Therefore, the syllabifications in δὲ κρίσις, σε βροτῶν, etc., are not as innovative as might at first sight appear, since Parmenides is here following closely his earlier model. If we exclude these cases, the ratio between innovative and conservative syllabifications is dramatically reduced from Passa's 16:20 to a more realistic 6:18.

Passa convincingly argues (pp. 100-104) that the dative μεγάλοισι θυρέτροις (B 1.13) excludes the interpretation of πλῆνται as a form of πίμπλημι 'fill'.⁶ According to Passa, πλῆνται should be seen as an unreduplicated perfect of πίλναμαι (epic for πελάζω 'approach'). Although the semantic explanation is almost certain, the morphological interpretation lacks convincing parallels. Even if we accept that the unreduplicated δέχαται (Il. 12.147), προθέουσιν (Il. 1.291) and προφύλαχθε (h.Ap. 539) were caused by the three consecutive light syllables of the original reduplicated forms, which made them unsuitable for the hexameter, this cannot have applied to 3rd pl. πεπλῆνται or πεπλέαται. Therefore, a morphological explanation of πλῆνται as a secondary poetic present from πλῆντο (epic aor. of πελάζω, Il. 14.468) is more likely.

As for πλάσσονται in B 6.5 , Passa believes that a new present *pla(n)k-jo- > πλάσσω derived analogically from the aorist ἔπλα[γ]ξα (< *e-pla(n)k-s-) replacing the ancient πλάζω 'turn aside' (πλάζομαι 'wander'). But the phonetic evolution *plank-jo- > πλάσσω is unlikely, since phonetic nasal lenition before stops (tacitly assumed in Passa's formulation *pla(n)k-jo- > πλάσσω) is far from certain in Ancient Greek.⁷ Moreover, πλάζω 'turn aside' is restricted to poetry. Even if we assume for the sake of the argument that πλάσσω was used in everyday Ionic, the reasons that would have led Parmenides to prefer the dialectal variant to the epic one are not convincing. Finally, the analogy put forward by Passa would imply an awkward similarity with πλάσσω 'mould'. Therefore, an original πλάΙονται = πλάζονται (where Ι stands here for the early form of Ζ), as suggested by Passa, could have easily become πλάττονται in the Attic copy, a mistake that was strengthened by the interpretation of the word as 'imagine for themselves'.⁸

Passa is doubtlessly right in considering that ἐπιδεές must be preferred to ἐπιδευές in B 8.33 and that no emendation of the line is needed (pp. 112-115). According to Passa, Parmenides used an allegro variant, scil. *ἐπιδείς but was compelled to write a more overt morphological form, ἐπιδεές. Nevertheless, this view is unwarranted. The alleged parallel of imperative πλέε (Ar. Av. 597), used by Aristophanes to avoid confusion with ind. πλεῖ is unsatisfactory, because 1) πλέε is a reading transmitted only by one late manuscript (Urbinas gr. 141, 14th c.), and 2) it is challenged by imp. πνεῖ in Aristophanes (Pax 87, not *πνέε) and again πλεῖ in Euripides (Tr. 102, Hel. 1663). In fact, in epic diction this exceptional prosodic anomaly normally conceals formula alterations. Thus, in ὄγδοόν μοι (Od. 7.261 = 14.287), quoted by Passa, the anomalous scansion is compelling, since ὄγδοον replaced another numeral in an earlier formula, probably *ἀλλ' ὅτε δὴ δέκατόν fοι [σφι] ἐπιπλόμενον fέτος ἦλθε [[can one use "v" in voi and ve/tos for digamma? it's not in beta code]] (cf. also the hiatus without correptio δὴ ὄγδοον). This may be also the case in the anomalous scansion in ἀσκηθέες καὶ ἄνουσοι (Od. 14.255), probably a modification of *ἀσκηθὴς καὶ ἄνουσος. To my mind, Parmenides simply took advantage of this in order to introduce normal Ion. ἐπιδεής, which otherwise would not have fitted in the hexameter.

In sum, this work remains an outstanding achievement and Passa makes a solid case for most of the passages discussed. Let us hope that Passa writes a similar study on Empedocles' extant fragments.

1.   E.g. "inziative" for "iniziative" (p. 16), "ὑδατορίζον", for "ὑδατόριζον" (p. 129-130), Karsten "1938" for "1835" (p. 33), "ΑΠΟΘΜΗΞΕΙ" for "ΑΠΟΤΜΗΞΕΙ" (p. 55).
2.   On p. 61 the Attic form ἄμαχσαν is quoted from outdated IG 1² 374 (recte IG 1³ 476); on p. 88, n. 123, "IG XII 549" should have been IG XII Suppl. 549; on p. 85 an inscription from Samos is cited "Hallof (2003), nr. 996" for expected IG XII.6, 996.
3.   There Ἐμποριτέων and Ἐ μποριτῶν are documented in the 3rd century BC, cf. Barrio Vega, María Luisa del, "Η διάλεκτος της Φώκαιας υπό το φως της ελληνικής δυτικής επιγραφικής". In Hatzopoulos, Miltiades (ed.), Φωνές, Χαρακτήρ, Εθνικός. Actes du 5e Congrès Internacional de dialectologie grecque, (Athènes, 28-30 Septembre 2006), Athènes, de Boccard, 2007, p. 19.
4.   Cf. Coxon's correction to ἐάσω ad loc., see Coxon, Allan H., The fragments of Parmenides. A critical text with introduction, and translation, the ancient testimonia and a commentary, Assen-Dover, Van Gorcum, 1986.
5.   Cf. Solmsen, Felix Untersuchungen zur griechischen Laut- und Verslehre. Strassburg, Trübner, 1901, pp. 136-7.
6.   Following Coxon, op. cit., p. 163.
7.   Cf. Méndez Dosuna, Julián, "Ex praesente lux", in Hajnal, Ivo (ed.), Die altgriechischen Dialekte. Wesen und Werden. Akten des Kolloquiums Freie Universität Berlin 19. - 22. September 2001, Innsbruck, Innsbrucker Beitr. zur Sprachwiss., 2007, pp. 358-367. The development (p. 108) *planzdo- > *pladzo- is also incorrect, perhaps a mere typo.
8.   Cf. Sider, David, "Textual notes on Parmenides' poem", Hermes 113 (1985), p. 365.