I

According to Aristarchus, Homer did not—or should not—have Doric features; nevertheless, this did not prevent the majority of the other grammarians from holding that he, at least occasionally, also employed Doric morphology, vocabulary, syntax and even accentuation.Footnote ¹ Doric accentuation is a particular kind of orthography which concerned Doric—or purported Doric—ancient Greek authors.Footnote ² Though its relationship with Doric as it was really spoken is uncertain,Footnote ³ this particular kind of accentuation is internally consistent: in the cases mentioned in the grammatical tradition, the accent, in comparison to the corresponding Attic forms, is constantly a mora closer to the word end.Footnote ⁴ Doric accentuation is attested in Imperial papyri—where the later the witnesses are, the more Doric accentuation is submerged by standard Attic orthography—and is almost forgotten in the medieval manuscript tradition.Footnote ⁵

As far as Homer is concerned, cases of Doric accentuation are attested in the scholia A on the Iliad, which are the best witnesses of Aristarchus’ exegesis on the poem. At Il. 2.393 an anonymous interpreter proposed to accentuate ἐσσϵῖται, which is a Doric future, according to the Doric customs, namely ἐσσϵίται, that is without observing the σωτῆρα rule;Footnote ⁶ Herodian rejected the proposal since according to him Doric futures, which are also present in Attic, had to be stressed according to Attic rules.Footnote ⁷ At Il. 8.378 Zenodotus proposed the reading προφανϵίσᾰς, that is a plural accusative of the first declension, with short ending vowel and paroxytone accent; by doing so, the grammarian was accused of introducing in Homer not only a case of Doric morphology, but also, possibly, of Doric accentuation—the reading is attested and described by the scholium with paroxytone accent. Apparently, this is another case where the σωτῆρα rule was not respected.Footnote ⁸ In the case considered in this article, the grammarian Hellanicus—to be distinguished from the mythographer of Mytilene—is charged by Aristarchus with introducing a Doric form in Homer. Possibly, Hellanicus’ reading too is Doric because of its accentuation.

II

Before we read the witnesses of Hellanicus’ interpretation, we have to consider the related Homeric line (Il. 5.269):

Diomedes aims to take possession of Aeneas’ horses,Footnote ⁹ generated by the mares secretly put under Laomedon’s divine horses by Anchises.Footnote ¹⁰ The mares of the latter are defined as ‘female horses’ (v. 269), through a noun of common gender (ἵππος) and an adjective precising the gender itself (θῆλυς). However, the adjective is not mentioned in the feminine form (sc. θηλϵίας), but in the masculine one (sc. θήλϵας), which plays the role of common gender: the adjective does not behave here as having three nominatives as it usually does—even in Homer—Footnote ¹¹ but as having two.Footnote ¹² Accordingly noun and adjective, though semantically referring to animals of feminine gender, both behave as words of male (common) gender. Furthemorer the adjective, despite the class it belongs to (sc. γλυκύς) and its own Sanscrit parallel (dhārú-), all of which are oxytone, has a remarkable recessive accent.Footnote ¹³ Now we can consider the scholia vetera on this passage:

The recessive accent was taken into consideration by the ancient grammarians: according to Herodian θήλϵας of Il. 5.269 had to be accented on the antepenultimate syllable (τρίτη ἀπὸ τέλους ἡ ὀξϵῖα), since Homer attests other cases of recessive accentuation at the singular nominative and accusative of the adjective (ἵνα ἀπ’ ϵὐθϵίας τῆς θῆλυς ᾖ κϵκλιμένον: Od. 5.467, Il. 19.97, 10.216).Footnote ¹⁴ Clearly, Herodian justified the particular accentuation of θήλϵας through the other occurrences of the adjective in Homer. Aristarchus and Ptolemy of Ascalon were also of this opinion (οὕτως Ἀρίσταρχος καὶ <ὁ> Ἀσκαλωνίτης).Footnote ¹⁵ This is hardly surprising, since Herodian often follows Aristarchus and both respect the Homeric evidence (the well-known Aristarchan principle of explaining ‘Homer through Homer’),Footnote ¹⁶ evidence which constantly attests a recessive accentuation. Probably Herodian also realized that θήλϵας is a common gender form, since the attached examples are also of this kind (Od. 5.467 [θῆλυς ἐέρση]; Il. 19.97 [Ἥρη θῆλυς ἐοῦσα]).Footnote ¹⁷ Probably Aristarchus, who knew this feature very well,Footnote ¹⁸ by considering the Homeric occurrences of θῆλυς, came to the same conclusion.

Aristonicus, who also commented on the accentuation of θήλϵας, criticized the grammarian Hellanicus,Footnote ¹⁹ who accented it on the penultimate syllable (ὅτι οἱ πϵρὶ Ἑλλάνικον ἀνϵγίνωσκον “θηλέας”), as shown by the attached example (ὡς ταχέας);Footnote ²⁰ for, according to Hellanicus, Homer presented the adjective in a Doric manner (ὡς Δωρικῶς ἐκτιθϵμένου τοῦ ποιητοῦ). According to Aristonicus, features of this kindFootnote ²¹ are frequent in Hesiod (τὸ δὲ τοιοῦτο παρ’ Ἡσιόδῳ πλϵονάζϵι) but absent from Homer (Ὅμηρος δὲ οὐ χρῆται). Therefore Hellanicus’ accentuation is to be rejected, since he did not recognize Homer’s particular dialectal mixture, which apparently ruled out Doric (cf. §1).Footnote ²² Indirectly, Aristonicus—who resumes Aristarchus’ commentaries on Homer—confirms what we already read in Herodian, namely that according to Aristarchus the adjective had a recessive accent.Footnote ²³

Aristarchus’ explanation of this Homeric lemma—and the content of his remarks ad loc.—as far as they have been reconstructed so far, can be summarized thus: he considered Hellanicus’ paroxytone accentuation, but ruled it out, because it had a Doric character, and Dorisms, which are well present in Hesiod, are absent in Homer—indeed, in what respect θηλέας was Doric is still to be established (see below). Indirectly, we understand that Aristarchus supported the vulgate proparoxytone accentuation, as particular as it was. This from Aristonicus; from Herodian—who probably follows Aristarchus in the interpretation of the lemma—we learn that Aristarchus accepted the proparoxytone accentuation and possibly supported it through the other Homeric occurrences of the adjective, where the accent is always retracted (Od. 5.467; Il. 19.97; 10.216). Though it is not made explicit by Herodian or by Aristonicus, Aristarchus may have also remarked that the masculine of the adjective plays the role of the feminine—that is, θῆλυς was used as an adjective with two nominatives and not with three.

But in what respect was Hellanicus’ reading Doric? Aristonicus is not explicit on this point. According to La Roche, θηλέας was Doric because of the ending; the scholar follows Eustathius who, in commenting on this Homeric passage, may have reported Hellanicus’ thought and interpreted the form as a feminine (on Hom. Il. 5.269 = 2.71.17–19 Van der Valk):Footnote ²⁴

First, Eustathius mentions the vulgate accentuation (οἱ μὲν προπαροξύνουσιν), with the corresponding nominative (ὡς ἀπὸ τοῦ ἡ θῆλυς) and a Homeric parallel (οἷον καὶ τὸ “θῆλυς ἐέρση” [Od. 5.467]), then the paroxytone accentuation (οἱ δὲ παροξύνουσιν).Footnote ²⁵ He adds that in the latter case the form is feminine and derives from the nominative θήλϵια (ὡς ἀπὸ τοῦ θήλϵια …); in turn, the Homeric form derives from the corresponding plural accusative through the loss of iota (… καὶ ἀπϵλϵύσϵι τοῦ ι θηλέας). In this case—namely if the adjective follows the first declension—since the metrical context requires a short syllable, the ending, despite the paroxytone accentuation, must have a short vowel, and can only be understood as the typical Doric ending of the plural accusative (sc. -ᾰς). That would explain why Hellanicus’ reading is Doric.

According to La Roche, this explanation is supported by another text, which we could consider a different version of Aristonicus’ scholium, where the shortening of the ending vowel is mentioned explicitly (Anecd. Par. 3.284.1 Cramer):Footnote ²⁶

Hellanicus considered the epithet paroxytone (οἱ πϵρὶ Ἑλλάνικον ἀνέγνωσαν “θηλέας”), like the Homeric ταχέας (ὡς ταχέας);Footnote ²⁷ according to the grammarian, Homer shortened the ending vowel (ὡς … συνϵσταλκότος τοῦ ποιητοῦ) in a Doric manner (Δωρικῶς). This passage from Anecd. Par. represents a slightly different version of Aristonicus’ scholium: consider the strong verbal resemblance between the two texts,Footnote ²⁸ the mention of the diplê—one of the most important of Aristarchus’ critical signs, which Aristonicus explains in his own work On the Signs of the Iliad (Πϵρὶ σημϵίων Ἰλιάδος)—and finally the presence of ὅτι, which usually introduces Aristonicus’ scholia. All these features show that we are dealing with Aristonicus’ scholium as we read above. Does this mean that Anecd. Par. preserve the original version of the scholium, where Aristarchus’ motivations—absent in the scholium A—are explained? La Roche, comparing Eustathius and Anecd. Par., thought so. On the one hand, Eustathius considered the form a feminine of the first declension; on the other, the Anecdota mention the shortening of the ending vowel—and the name of the grammarian—explicitly; a short vowel is also required by the hexameter.

This explanation may appear likely, but objections can be raised against it. The Doric shortening mentioned in Anecd. Par. refers to the ending of the plural accusative of the first declension and Eustathius, by indicating the form as feminine, has in mind the same phenomenon. However, they do not preserve Aristarchus’ actual exegetical intentions. The notion that θηλέας was a feminine and followed the first declension could hardly be part of Aristarchus’ interpretation, since the adjective under discussion was probably considered by Aristarchus—and by Aristonicus and Herodian after him—as a common form (above). Therefore Aristarchus—and Aristonicus after him—could not have stated that the form was Doric because of the short ending vowel. Possibly, the reference to the Doric shortening in Anecd. Par. was added by someone who held that Aristonicus’ generic words, which we still read (ὡς Δωρικῶς ἐκτιθϵμένου τοῦ ποιητοῦ), referred to such a fact, and rephrased accordingly (and Eustathius read a version of the scholium closer to Anecd. Par. than to the codex A). Evidently, whoever read Aristonicus’ scholium—just like us modern scholars—had to interpret the latter’s unclear exposition, and returned to the most known—or the least unknown—notion about the Doric dialect. That anonymous scholar may have been aware that the hexameter, on this point, requires a short syllable and the Doric ending appeared a plausible explanation to him.

La Roche’s explanation shows another weak point. The idea we read in Eustathius, namely the passage of the purported feminine form from θηλϵίας into θηλέας through the loss of iota (οἱ δὲ παροξύνουσιν ὡς ἀπὸ τοῦ θήλϵια θηλϵίας καὶ ἀπϵλϵύσϵι τοῦ ι θηλέας)—though being a means dear to the ancient grammarians in order to support their etymologies—Footnote ²⁹ is not only untenable nowadays, but also reveals a defect from a dialectological point of view. For, if we consider the literary and epigraphic witnesses, it is most unlikely for the penultimate syllable of the feminine adjectives in -ύς in Doric to be short: this adjective, provided it is feminine and the final ending short, had to appear either as θηλήᾰς (Doris severior) or as θηλϵίᾰς (Doris mitior).Footnote ³⁰ Accordingly, it is not likely that an ancient grammarian stated that θηλέας was a feminine form—regardless of whether Hellanicus thought so or Aristarchus himself, if he ever misunderstood Hellanicus’ intentions. For if a grammarian of the Hellenistic times had proposed an interpretation of this kind, he would have been attacked by his colleagues, who would have presented evidence attesting the opposite and charged him with ignorance (ἄγνοια); and we do not have evidence of such objections, neither against Hellanicus nor against Aristarchus for not objecting. Therefore, for all of the Hellenistic interpreters, θηλέας, even if understood as a Doric form, could not be considered a feminine of the first declension—which had a long penultimate syllable—but only a masculine of the third. For his part, Eustathius probably realized that θηλέας and the koine θηλϵίας did not correspond perfectly, since not only the final syllable had to be short—problem solved through the Doric ending in -ᾰς—but also the penultimate one. Therefore Eustathius returned to this common tool, the loss of a sound.

All this means that the scholar who reworked Aristonicus’ scholium in Anecd. Par.—and Eustathius who followed him—cannot be used to interpret Aristonicus’ scholium as we read it in codex A. This anonymous scholar, aware of the metrical constraint—the final syllable had to be short—and being advised by the scholia ad loc. about the role played by a Doric feature in Hellanicus’ reading, held that the short ending of the plural accusatives of the first declension, typical of Doric, was the best explanation. The latter—not an unlearned scholar—can hardly come from Aristarchus, since he considered θήλϵας a common form following the third declension. Further, for θηλέας to follow the first declension would be dialectologically impossible, because in Doric the penultimate syllable of this kind of adjective is long (cf. above), something that could not have been neglected by ancient grammarians.

If this is true, in what respect was Hellanicus’ reading charged with being Doric? This was because of its accentuation. First, Aristonicus, by choosing ταχέας as example, aimed to be precise in describing Hellanicus’ reading; remarkably, Aristonicus did not only mention another Homeric term,Footnote ³¹ but also an adjective of the same class, with the same formal gender and the same accent: ταχέας corresponds to Hellanicus’ θηλέας perfectly. This can only mean that Aristonicus aimed to inform his reader as clearly as possible about the intentions of the criticized author. Both forms are paroxytone, but it is the paroxytone character of θηλέας that made Hellanicus’ reading worthy of consideration in comparison to the vulgate one. Second, the accent, being paroxytone, shows the main feature of Doric accentuation, that is, it is closer to the word end than in the corresponding Ionic form (cf. section I).Footnote ³² This means that Hellanicus’ reading was unacceptable to Aristarchus because of its paroxytone accentuation, which had an unmistakably Doric allure.

III

After reconstructing how Aristarchus commented on Il. 5.269 and—possibly—the actual exegetical intentions of the latter, we have to raise another question. Aristarchus held that Hellanicus’ θηλέας was Doric in some respect; yet are we sure that this was the interpretation of the latter? For Hellanicus’ reading can be understood in many ways. Obviously, the first explanation may be the Aristarchean one: Hellanicus considered θηλέας Doric—possibly because of its particular accent—and Aristarchus reacted accordingly.Footnote ³³ La Roche suggested—but immediately excluded—another explanation: Hellanicus accented the epithet analogically to the masculine adjectives of the same class,Footnote ³⁴ which are frequent in Homer.Footnote ³⁵

If we apply the categories conceived by van Thiel to interpret the readings attributed to pre-Aristarchean grammarians, we can see the explanations which have been already formulated with different eyes, and also formulate further ones. According to van Thiel, a reading attested with the name of a grammarian active before Aristarchus derives from the diorthôsis of that grammarian, that is, from the copy of a literary text corrected by its owner, or rather enriched by the latter’s marginal notes. Accordingly, the reading may either belong to the text of the poet or be a marginal note. In the latter case, the reading might have played different roles: obviously that of varia lectio—not necessarily made own by the grammarian, who simply might have recorded an interesting variant—but also the role of conjecture, parallel, paraphrasis or elementary form of commentary. Later, this complex critical apparatus was misunderstood by Aristarchus, who reduced all these functions to that of variant reading accepted by his predecessor.Footnote ³⁶

Hellanicus’ reading can be explained according to van Thiel’s frame. Since Homer’s manuscript tradition attests θήλϵας unanimously—and the accent of the word is constantly attested as retracted, both in Homer and outside Homer—it is licit to argue that the text of the grammarian might have had the same reading; in this case, θηλέας, which is paroxytone, could only have been written in the margin. With what function? The considerations expressed just above suggest that we are dealing neither with a variant reading nor with a parallel passage in Homer or in another author. We have to turn to the other opportunities offered by van Thiel’s frame. As supposed by Aristarchus, Hellanicus read the epithet with a Doric accentuation; for we know that the grammarians outside Aristarchus’ school were open to the presence of Dorisms in Homer.Footnote ³⁷ In this case, the marginal reading has to be interpreted as a prosodical correction of the word under discussion. Yet Hellanicus’ motivation for reading θήλϵας, the only existing reading, with a purported Doric accentuation, remains unclear. Another possibility is suggested by La Roche, namely that Hellanicus might have held that the vulgate reading, with its curious proparoxytone accentuation, had to be normalized to that of the adjectives of the same class, which were present in Homer (above). In this case too, we have to deal with a prosodical correction – this time, one more comprehensible than the one referring to the Doric accentuation.

There is a further possibility: we may be dealing with a commentary reading, namely a linguistic observation presented in a concise and concrete form. Hellanicus, having in mind the Homeric adjectives of the same class with paroxytone accentuation (ὠκέας, πολέας, etc.), through θηλέας, accented precisely in this way, remarked only how the adjective had to be stressed, but was not. In other words, the grammarian did not aim to propose a reading with a purported Doric accentuation (Aristarchus) nor to correct the form analogically to the adjectives of the same class (La Roche’s first hypothesis), but only to remark on the curious, proparoxytone accentuation; this happened by indicating the correct accentuation according to the class of adjectives it belongs to. Hellanicus, who was interested in dialectology,Footnote ³⁸ might have easily realized a fact of this kind. He might also have attempted to interpret it: despite Aristarchus’ accusation of introducing Doric forms in Homer, Hellanicus, by remarking that θήλϵας had to bear—but did not actually have—a paroxytone accent, could only define it as an Aeolism, since the accentuation of θήλϵας, in comparison to that of the adjectives of the same class, is recessive. In other words, Hellanicus remarked on the same issue commented on by Aristarchus and all modern scholars later on, namely that the accent of θῆλυς is retracted.

If the reconstruction is correct, Hellanicus made an observation not only on a Homeric form—the adjective is attested in Homer and his interpretation is mediated within the Homeric exegetical tradition, something which suggests that the grammarian had the Homeric form especially in mind—Footnote ³⁹ but also on the adjective, since θῆλυς is always attested with a retracted accent.