BMCR 2000.04.22

A Stylistic Commentary on Hermesianax. Classical and Byzantine Monographs 43

, A stylistic commentary on Hermesianax. Classical and Byzantine monographs ; v. 43. Amsterdam: A.M. Hakkert, 1998. 262 pages ; 23 cm.. ISBN 9025606385.

Hermesianax of Colophon1 (3rd c. B.C.E., hereafter H.) is chiefly known as the author of a long elegiac poem addressed to, and named after, his mistress Leontium. The work seems to have treated famous love affairs, framed by the poet’s own, which was doubtless as ill-fated as his exempla. Preserved are a single line from the section on Polyphemus (fr. 1 Powell), prose reworkings of several others (fr. 2-6 P.), and, most important, a 98 line catalogue of amorous poets and philosophers (from Orpheus to H.’s older contemporary and alleged teacher Philitas, including Pythagoras, Socrates, and Aristippus), which is cited by Athenaeus (fr. 7 P.). Whereas Leontium as a whole is clearly influenced by the Lyde of H.’s townsman Antimachus, the surviving part most resembles Phanocles’ Erotes e kaloi. Both Phanocles and H. state their debt to Hesiod’s Ehoea in the first line of the surviving text.

Because H.’s catalogue is both the longest continuous fragment of a Greek elegiac poem to survive and an important bridge between early Greek and later Roman (love) elegy, it has, in the past, attracted a fair amount of scholarly attention. Yet since the beginning of our century interest in the poem has dwindled — reflecting, I suspect, a diminished appreciation of its quality (which critics now agree is rather mediocre) and exasperation with the tantalizing problems of interpretation it poses. As a result, students of H. are forced to consult antiquated commentaries in Latin and to sample more recent work on the poem themselves. A new commented edition is clearly welcome.

Kobiliri’s book, the revised version of her 1979 (!) Birbeck doctoral dissertation, consists of a two-page introduction, an edition without critical apparatus, a translation, and an extensive word-for-word commentary. The volume concludes with Greek and analytical indices (which somewhat compensate for the absence of a proper introduction), and a bibliography. I will first deal with the bibliography, because it is illustrative of the nature of K.’s undertaking.


K.’s revisions of the original dissertation did not include incorporating advances in the field of Greek poetry made since 1979: the most recent title in her bibliography is Hatzikosta’s 1982 commentary on Theocritus 7. No explanation for this is given. As regards H. himself, one can excuse the omission of Hardie’s thorough article on the Philitas section, which appeared only in 1997. But there is no excuse for neglecting Huys’ ascription to H. of the so-called Tattoo Elegy (P.Brux. inv. E8934 + P.Sorb. inv. 2254) and Slings’ (convincing) refutation of this ascription: both scholars offer important observations on H.’s style. Other important omissions include articles by Bing, Brugnoli and Di Fabio, Allen’s Mimnermus, Matthews’ Antimachus, and Lefkowitz’ The Lives of the Greek Poets. The list could easily be expanded.

For other Hellenistic poets the situation is not much better. Apollonian scholarship, for example, is represented by the editions of Wellauer and Merkel (1828 and 1853-4), and the commentaries of Mooney and Fränkel (1912 and 1968). K. does not acknowledge the existence of Vian’s Budé edition (1974-1981), nor does she cite the commentaries of Ardizzoni, Campbell, Hunter, and Livrea, all of which have a lot to say about Hellenistic ‘style’.

As the case of Apollonius shows, K.’s reading is also surprisingly idiosyncratic for literature that appeared before 1979. In fact, almost a third of the works she lists date before 1900. They include such vintage classics as Jannaris’ A Historic Greek Grammar (1897), Angermann’s De patronymicorum Graecorum formatione (1869), and Lorentz’s Observationes de pronominum personalium apud poetas Alexandrinos usu (1892). The only edition of Athenaeus listed is Gulick’s Loeb (no Bergk, Diehl, Dindorf, Kaibel or Schweighäuser). Work on Homer seems to have ceased after Ameis-Hentze-Cauer and Leaf.

What K. does list, she lists sloppily. Many references are incomplete and quite a few are incorrect. Two short notes by H. Fuchs (MH 27 (1970) 179 and MH 28 (1971) 182) have merged into one (MH 28, 1970, 182ff.); O. Weinreich’s article is to be found at WJA 4 (1949-50) 384 (not 92f.), etc.

Text and translation

Editing H. is no sinecure, as the text has come to us in a wretched state. It is clear that the scribes of Athenaeus could often make little sense of what they were copying, and they can hardly be blamed. H.’s meandering sentences often run to several disticha and show more concern for rhythmical and sound patterns than for information structure; they are full of linguistic rarities, récherché collocations, and uncommon names. His love-affairs, typically spun from the poets’ subject-matter, are no less arcane and far-fetched than his style. Hesiod is paired off with Ehoea (!) and Homer with Penelope, Alcaeus and Anacreon anachronistically compete for the love of Sappho, Socrates courts Aspasia, and so on. Most of these fanciful stories are not elsewhere attested, or if they are, not in the form in which H. gives them. Even when the text was in pristine condition, understanding it must have presented a real challenge. Now that it is considerably corrupted and most of the literary and biographical works alluded to are lost, restoring and understanding it in every detail seems an impossibility. Hermann, in the preface to his 1828 edition, stated, with untypical but justified modesty, that he had no illusion of presenting a sound text, and that he could only hope to have left it somewhat less corrupt for the next editor, whom he invites to do a better job. The most recent editor, Defradas, shows similar resignation: he includes the Athenaeus fragment in his Élégiaques grecs“intégralement malgré sa médiocrité, malgré sa difficulté, due à la fois à la mauvaise qualité du texte et au caractère recherché du style, parce qu’il nous a paru réprésentatif d’une certaine poésie alexandrine.” He provides annotation only for those passages that can be understood; unsolved problems are passed over in silence.

Could Hermann, Defradas, and all those other editors be wrong? According to K., they are. All previous editors, so she asserts time and again, have failed to see that the transmitted text needs only a handful of corrections to make it comprehensible. As a result, her text of the 98 line fragment differs from the last critical edition (that of Powell) in ca. 70 places (disregarding punctuation). In all but a few cases, K. returns to the transmitted text, which she defends vigorously in her commentary, typically with phrases like: “this is the reading of A(thenaeus), rejected by most critics, who could not understand…” In most cases, K. makes no effort to name these “critics”, to mention their conjectures or to discuss their objections to the transmitted text. When she does, she often shows a feeble command of the basic rules of semantics, grammar, and metre. For example, A’s ἱεράς in 45 is defended in the following way:

“Musurus changed ἱεράς into ἱράς, obviously on the grounds that the form ἱρός is conventionally employed by Homer at verse-beginning. However, the form ἱερός is also rarely attested in this sedes [i.e. “is also attested, though rarely”? K.’s English is often hard to follow]: cf. ἱερά τε ῥέζουσι in Hom. Od. 5.102 and h.Apoll. 394; cf. also the same sedes in Apollonius Rhodius 4.531 ( ἱρήν).”

K. was apparently so preoccupied with the line-postion of the word in question (more about this below) that she overlooked the real problem: the transmitted text does not scan (- u -, the evidence she adduces scans – u u; for the corruption as such cf. e.g. A.R. 2.302, 486, 718). This is by no means K.’s only unmetrical line: 3 begins with Ἀϊδόθεν (read Ἁιδόθεν), 33 with ἔκλαιε δ’ Ἰκάρου τε γένος, 66 with Αἰγείων μέθεπε δ’ Ἀρχελάῳ (on these two verses see below), and 51 ends in ἐφωμίλησ’ Ἀνακρέων (read ἐφωμίλησ’ Ἀνακρείων rather than ἐφωμίλησεν Ἀνακρέων).

Other observations: — 1 and 13. K.’s decision to retain Ἀγριόπην seems wise, although one may quarrel about the arguments. H. is both the first to give Orpheus’ wife a name (in the earliest sources, Euripides and Plato, she remains anonymous) and the only one to call here anything near Agriope (the name Eurydice is not attested before the Epitaphium Bionis). The name Argiope ( Ἀργιόπην ci. Zoëga, not as widely accepted as K. suggests) has the benefit of being elsewhere attested, but none of the nymphs of this name is ever associated with Orpheus. Moreover, the transmitted Ἀγριόπην produces a not unhermesianactic jingle after Οἰάγροιο, and seems to combine the names of Orpheus’ father (Oi-agr-os) and mother (Call-iope). In any case, there are insufficient grounds for changing the transmitted text. — 4. The transmitted ἀκοήν (sc. ἄκατον, of Charon’s ferry), which virtually all editors replace by something else (Ludwich’s ἀκορήν, “insatiate”, seems most satisfactory), is accepted by K. on the authority of Giangrande, whose arguments she does not discuss. The reader is left to wonder how ἀκοός, which on its only other attestion (Plato Com. fr. 226) equals ἀκουστικός, can come to mean “wailing” (so K.’s translation). — 7. παρὰ κῦμα μονόζωστον κιθαρίζων : read μονόζωστος (Ruhnkenius). The adjective combines the notions of “traveling alone” (Soph. OT 846 οἰόζωνον) and “lightly armed”, i.e. with nothing but his lyre (prose μονόζωνος, for the idea cf. 2 στειλάμενος κιθάρην — not κιθάρῃ, as K. prints). The adjective cannot possibly go with κῦμα (“the wave that was crossed by a solitary person”, K.). This is one of many cases where K. invokes a ‘magic word’ (here enallage) to support an interpretation that goes against common sense. Another is the use of inconcinnitas to defend 11 φωνῇ (read φωνήν). — 17. πολυμνήστῃσιν : Hesychius’ πολυμνήστην : ἀγαθήν, σώφρονα is an interpretation from the Homeric context; it cannot be used to defend the transmitted text here ( pace Giangrande). Read πολὺ μύστῃσιν (Diehl) or νήστῃσιν (Cataudella). — 19-20. Ράριον ὀργίων’ ἀνέμῳ διαποιπνώουσα | Δημήτρᾳ : according to K. this means “accompanying the priest of Demeter with religious inspiration.” But ἄνεμος never means “religious inspiration” (perhaps – ι νόμῳ or ἀνεμώλια π -), and neither ὀργίων’ nor ποιπνώουσα nor Δημήτρᾳ is acceptable Greek. Read a form of ὀργειών (accented thus; with metrical lengthening), – πνυ -, and Δήμητρα. — 21. μελάθραν : if a noun μελάθρα existed, H. should have written μελάθρην. But it does not: H. wrote μέλαθρα (Meineke, rather than μέλάθρον, Musurus). — 29. λεπτὴν δ’ εἰς Ἰθάκην ἀνετείνετο : this cannot mean “made in his poem reference to small Ithaca.” The particle must be deleted in any case (Dindorf), because the line carries the main verb to 27-8. H. may have written λεπτὴν ὡς Ἰθάκην ἀνετείνετο, “how subtly did he (Homer) versify Ithaca” ( ὡς conieci, ᾗσ’ and ἐνετείνατο Kaibel). — 33. ἔκλαιε δ’ Ἰκάρου τε γένος is unmetrical and incomprehensible. The name of Penelope’s father is Ikarios (passim in the Odyssey) and Homer did not “bewail” (K.) his race but made it famous: read ἔκλεε δ’ Ἰκαρίου κτλ. (Hermann and Bergk). — 41. Λυσηίδος : read Λυδηίδος, “Lydian Lyde” (a cheap pun, yes). Even if Λυσηίς could mean “the priestess of Dionysus” (which I doubt), there is no evidence that Lyde was one. — 45. ἄκρον ἐς κολοφῶνα : it is hard to see why Antimachus should have set up his desk on a high hill: he rather went ἄκρην ἐς Κολοφῶνα (etymological wordplay). — 49. ἠράσαθ’, ὕμνων | … πολυφραδίῃ : no ancient reader would consider taking ὕμνων with πολυφραδίῃ at the end of the next line, not even in H. The comma should be deleted; genitive ὕμνων depends from the verb (or read ὑμνῶν ?). — 53-4. δουρὶ ’ν is not the solution to the transmitted δουρῖν. Aphaeresis in elegiacs is highly unlikely, and the result is dubious Greek regardless. — 55. μυρίον … λέκτρον : read Μύσιον … Λεκτόν, the Trojan promontory which Anacreon would have seen on the Asian coast had he indeed stared over the Aeolian gulf yearning for Sappho (of course he would have had to be in Mytilene, not in Eresus). — 57. πολυπρίωνα κολώνην : if the “Attic bee” Sophocles ever left home for Theoris, he must have left his deme Κολώνη, which H. probably called πολυπρήων. — 58. ἐν τραγικαῖς ᾖδε χοροστασίας : K. comments: “the accusative is perfectly sound here, the point being that Sophocles sung choruses ( χοροστασίας) in tragedies ( ἐν τραγικαῖς).” One would like to see evidence for τραγικαί = τραγῳδίαι. Read χοροστασίαις (Musurus): Sophocles celebrated Bacchus and Theoris “in tragic stasima”. — 59. τὸν Ἔρωτ’ ἀγειραιθειαρειδος : the exact text and sense of 59-60 are beyond recovery (A. has ἔρωτ’ αγειραιθειαρειδος followed by a lacuna), but one can salvage more than K. does. The transmitted sequence probably hides a genitive Θεωρίδος (Lennep, Heinrich) from Θεωρίς, who is mentioned as Sophocles’ mistress at Athenaeus 13.592a. This genitive can only have stood immediately after the caesura, and the article τόν indicates that it depended on ἔρωτα (uncapitalized; Sophocles celebrated a particular love rather than [the god of] Love). — 62. συνοχῶν“ne veut rien dire ” (Defradas); it should be obelized. — 66. Αἰγείων μέθεπε δ’ Ἀρχελάῳ ταμίην : unmetrical, read μέθεπεν δ’ or μεθέπων Ἀρχέλεω. Αἰγείων is incomprehensible. It hides either a participle ( ἀγρεύων· μέθεπεν δ’ Headlam, carrying on the metaphor of 63-4) or a name (in which case read μεθέπων), most likely that of Euripides’ lover (male: 63 σκολιοῖο), who would otherwise remain anonymous, against H.’s practice (e.g. Αἰγαῖον, derived from the Macedonian city Αἰγαί). — 68. ἀμφὶ βίου στυγνῶν ἀντιάσαντα κυνῶν : ungrammatical and incomprehensible. The participle should congrue with 67 Εὐριπίδῃ, and therefore be in the dative (Musurus); ἀμφὶ βίου seems to hide the name of the owner of the dogs which tore Euripides apart ( Ἀμφιβίου ?). — 69-74. The text printed by K. is ungrammatical: for 69 ἀνεθρέψαντο read ( ἄνδρα δὲ τὸν Κυθέρηθεν) ὃν ἐθρέψαντο ( τιθῆναιγιγνώσκεις, 73) with Hermann. In her commentary K. prints ὃν ἀνεθρέψαντο (unmetrical) and notes that ἀνέθρεψαντο is “perfectly sound.” But what is at issue here is the construction of H.’s winding sentence not the sense of ἀνατρέφομαι. — 72. διὰ πτόλεως : in her commentary K. rightly observes that Homer uses three genitive forms of πόλις, namely πόληος, πόλιος, and πτόλιος, and that πτόλεως is an absolute hapax (unless it is to be read at Eur. Tro. 1079). This should have made K. wonder what reason Hermesianax may have had to use a unique epicized Attic form where the regular epic form πτόλιος would have fit as well (verse end). There is no such reason: πτόλιος (Meineke) is what H. wrote. — 73. γινώσκεις but 49 γιγνώσκεις : H. wrote either the one or the other, the editor must choose. Likewise 94 εὑράμενος after 87 εὑρόμενον (recte; A in both cases ). — 77-8. Βιττίδαθοήν cannot mean “tall Bittis”: read Βιττίδα μολπάζοντα, θοὸν περὶ πάντα Φιλιτᾶν (accented thus) κτλ., “Philitas, acute more than anyone…” With this (new?) solution θοός becomes understandable (it refers to Philitas’ scholarly acumen) and the clause-division coincides with the caesura. — 91. μηνίουσα : perhaps read μηνιόωσα (cf. A.R. 2.247, Opp. Hal. 3.607). — 95. ἄνδρα Κυρηναῖον εἴσω : read ἄνδρα δὲ Κυρηναῖον ἔσω : no other section of the fragment begins without a connective particle and – υ – in Κυρήνη, – αῖος is normally long. After δέ had dropped out (haplography) a scribe has ‘restored’ the metre by changing ἔσω into εἴσω.


It is typical of ‘stylistic’ notes in commentaries that they more or less assume that the reader knows what question the note tries to answer — i.e. they assume that the reader is familiar with, and agrees with, the interpretational framework within which the commentator operates. For obvious reasons, commentators cannot elaborate on the purpose and presuppositions of every note they offer (the commentary would become too bulky). Yet, ideally, they should be able to provide an explanation if asked. Whether K. will be able to provide her readers with a satisfactory motivation of much of the material she includes, I doubt. Often she does not seem to have realized what the question was, and this results in many unhelpful answers. Take for example this note on 76, a verse which tells us that “the Coans set up [Philitas] in bronze” ( χάλκειον θῆκαν):

χάλκειος is the epic variant of χάλκεος, just as χρύσειος is the epic variant of χρύσεος (cf. TLG 1262). In Homer the epithet is normally applied to weapons (cf. Hom. Od. 19.241 χάλκειον ἄορ); note that χαλκός is often used metonymically instead of weapons; the collocations Ἀχαιοὶ χαλκοχίτωνες (Hom. Od. 4.496, etc.) and χάλκεος Ἄρης (cf. Hom. Il. 16.543) are also very common.”

In this note, which runs on like this for seven more lines, not a single element has any bearing on H.’s text. Remarks which assume unfamiliarity with the basic facts of epic diction, such as the first sentence of the note, occur everywhere (for example, on 42 πληγείς : “note that Theocritus, Callimachus and Apollonius Rhodius employ only the form πλήσσω not πλήττω“; or on 43: ” θέτο is an unaugmented form”). The rest of the note is illustrative of the zeal with which K. lists the archaic, and especially the Homeric, antecedents of H.’s diction, no matter how unremarkable. The following 21-line note on θῆκαν provides more examples. Here K. cites, among many other ‘parallels’, Il. 2.319 λᾶαν γάρ μιν ἔθηκε and Pi. N. 4.81 στάλαν θέμεν, both unhelpful. The parallels she ought to have cited include Theocr. ep. 18.3-4 Gow χάλκεόν νινἀνέθηκαν and Posid. 14.4 Gow-Page χάλκεος ἑστήκω (cited by Bach and Bailey). This is by no means the only note where K.’s preoccupation with Homer makes her forget to look at H.’s contemporaries. Thus on 5 οἰχομένων she rightly notes that οἰχόμενος = “(the) deceased” is unhomeric but found in tragedy, but she omits the many examples in Hellenistic poetry (e.g. Call. fr. 4.73 Pfeiffer, Theocr. ep. 6.6 Gow, A.R. 2.839-40, 3.203-4) which indicate that H. was not the only poet (and probably also not the first) to use this tragic expression in dactylic verse. When K. does comment on the relation between H. and other Hellenistic poets, she tends to overestimate H.’s importance. It is, for example, unclear to me why Apollonius Rhodius’ metaphorical use of τινάσσω (= ταράσσω) should have been inspired by H. 71 (cf. already Sappho 47, cited by K.).

Like other commentators in the series, K. devotes much space to issues of metre, and she is especially attentive to the line-position (sedes) of words. This is of some relevance, as Hellenistic poets tend to avoid being too Homeric by using common collocations in uncommon line-positions. But K. handles her critical tool in an altogether mechanical way: apparently, it is equally significant if words occur in a homeric sedes and if they occur in an unhomeric sedes. Worse, however, is that she does not distinguish in any way between pentameter and hexameter verse (e.g. in her notes on 42 ποταμοῦ, 43 θανοῦσαν, and 48 φορμίζων), and does not take into account the rules for word-placement in the hexameter. A typical example is her note on 15 ἔθηκεν : “Verse-end is the conventional sedes for this form, from Homer down to Manetho.” But we are dealing here not with convention but with the rhythmical requirements of the verse: amphibrachic words cannot stand in any other position. The same objection applies to dozens of other notes (e.g. on 41 ἔρωτος, 71 τιναχθείς, 75 πολιῆται). K. is also unaware that not every word-end has the same rhythmical status: 43 ὑπὸ ξηρὴν θέτο γαῖαν does not violate the rule that “word-end after both the seventh and ninth elements of the same line is avoided” (Maas, Greek Metre 97), because the word-end after ὑπό does not constitute a rhythmical break; and verse end ὃν Γαλατείης at 73 certainly does not constitute a major pause after the bucolic diaeresis. K. often offers metrical observations on (sections of) the poem without attempting to explain their purpose (why, for example, is it significant that two of the fourteen verses on Orpheus are spondaic? [p. 20]), to investigate deeper causes or to wrap them up in a general evaluation. The latter would admittedly have been difficult, given the haphazard nature of her comments — an observation which may be extended to her linguistic comments and the absence of an introductory section on H.’s language.

A hallmark of K.’s commentary is the ill-considered application of stylistic explanations to what is in fact semantically or grammatically required. An example is her note on 4 ἕλκεται (εἰς ἄκατον) : “Hermesianax has employed here the middle form, which is very apposite, given the Hellenistic fondness for the middle form of verbs.” Whatever one’s ideas about Hellenistic poets’ predilection for the middle, this general criterion clearly does not apply here: Charon drags the souls of the deceased into his barge (not someone else’s). Likewise, there is nothing remarkable about the middle ἐνεπλήσατο in 45: Antimachus filled books with laments to his own purpose, namely to find relief ( παυσόμενος rather than παυσάμενος) for his pangs. Linguistic missteps such as these occur on every page. Many even betray a feeble command of morphology: Iliad 1.72 ἣν διὰ μαντοσύνην (in which ἣν is a possessive) should not be adduced as parallel for 31 ἣν διὰ πολλὰ παθών (in which ἣν is a relative); A.R. 1.1057 γόων is a participle, not a noun as in H. (45), etc. Incorrect accents abound in both commentary and text (gravis before pause is more or less standard).

K.’s references to secondary literature are often more impressive than helpful. To her observation that 43 θέτο is an unaugmented form, she adds a reference to Poehlmann’s Quomodo poetae epici augmento temporali usi sint, Gymnasium-Programm Tilsit 1858. Many lexicographical references are to Passow’s Greek-German dictionary of 1841-57 rather than to LSJ — translations are left in German, also the language of choice for technical terms (Selbstvariation, Klangwirkung, Marginalkorrektur, Wortstellung, Adjektivshäufung, Vorzeitigkeit, etc.). Ancient scholarship is adduced in an equally gratuitous way. Why, for example, do we need the Etymologicum Magnum to tell us that Homer’s poems are ἀοιδαί ? A reference to Il. 1.1 (or to LfgrE s. ἀοιδή, ἀοιδός, ἀείδω) would have sufficed. The difference between ad hoc interpretation and basic meaning, which Hellenistic poets knew very well (as their poems show), is lost on K. (see e.g. on 17 πολυμνήστῃσιν), as is the difference between connotation, usage, and meaning (e.g. on 25 ἀνεγράψατο and λίβρους).

Author, editors, and publisher cannot be congratulated on their decision to let this work appear in print. They ask us to buy a messy doctoral dissertation which was completed in 1979, and which has not ripened or been brought up to date since. Despite its scale (at least two pages per verse), K.’s book contributes little to our understanding of Hermesianax, while setting crooked much that others had set straight. Those interested in the poem’s style are better served by Ellenberger, Couat, Huys, and Slings. For other issues Bach, Giarratano, Defradas, Bing, and Hardie may be consulted with more profit.


1. The titles of all works on Hermesianax mentioned in this review can be found at