Livineius’ athetesis at Aj. 714 of the transmitted phrase τϵ καὶ ϕλέγϵι (also attested by Suda ϕ 525),Footnote ² already omitted by Stobaeus (Flor. 1.8.24), is accepted by most recent editors and scholars,Footnote ³ with few exceptions.Footnote ⁴ In a recent contribution, Battaglino has returned in detail to the textual and interpretative issues raised by Aj. 714:Footnote ⁵ although she ultimately opts for the shorter version, and provides an interesting interpretation of μαραίνϵι (see below), her accurate discussion of the previous literature on Aj. 714 also emphasizes the plausibility of the transmitted reading, as gesturing towards a possible (yet not easily detectable, as she rightly argues) Heraclitan influence on Sophocles’ imagery of time.Footnote ⁶ I here advance further remarks in support of the widely accepted hypothesis of τϵ καὶ ϕλέγϵι as a later interpolation, both through a re-examination of previously advanced arguments and the addition of original ones.

The addition of the iambic sequence τϵ καὶ ϕλέγϵι would require a metrically similar supplement at Aj. 701 νῦν γὰρ ἐμοὶ μέλϵι χορϵῦσαι: an expression of choral self-referentialityFootnote ⁷ which is self-contained and strongly assertive, and thus hardly amenable to plausible textual emendations.Footnote ⁸ Moreover, the addition of τϵ καὶ ϕλέγϵι poses the issue of the plausibility and unclear interpretation of a metrical sequence than can be variously intended as a dodr ia^, cho penthem ^ia (reiz), or ¨hi (anaclastic hipponactean)Footnote ⁹ + ia. In addition to this interpretative uncertainty, what makes this sequence anomalous is mainly the fact that a sequence ba ia (μαραίνϵι τϵ καὶ ϕλέγϵι, ⏑ – – ⏑ – ⏑ –) ‘scarcely if at all occurs in lyric iambics before late Euripides’.Footnote ¹⁰ Such a metrical sequence cannot even be included in one of the attested three types of choriambic trimeter (ia cho ia, A; ia cho ba, B; cho cret ba, C).

Not only do these metrical issues strongly support the athetesis, but also the semantic/interpretative arguments sustaining the plausibility of τϵ καὶ ϕλέγϵι can be countered. It is true, as Reitze argues,Footnote ¹¹ that the addition of ϕλέγϵι as opposed to μαραίνϵι, with the meaning of ‘kindle/illuminate’,Footnote ¹² has the advantage of mirroring the dichotomy ἄδηλα/ϕανέντα (and ϕύϵι/κρύπτϵται) at Aj. 647; furthermore, ϕλέγω is used in this sense by Ajax himself at Aj. 672–3 (the daylight replacing the darkness of night), as a concrete example of the contrast between revelation/generation and concealment/destruction thematized by Aj. 646–7.Footnote ¹³ Nevertheless, an exact lexical correspondence is not needed to detect a reminiscence of Aj. 646–7 in the chorus’ words;Footnote ¹⁴ moreover, and more relevantly, the lack itself of this correspondence with Ajax’s statement in the Trugrede, emerging as a simplification of the latter’s more complex semantic and syntactical articulation,Footnote ¹⁵ has the effect of emphasizing the chorus’ misunderstanding of the intrinsic meaning of Ajax’s words, and thus the crucial irony of the stasimon.Footnote ¹⁶

Furthermore, the absence of ϕλέγϵι does not render the meaning of μαραίνϵι less perspicuous, despite Reitze’s contentionFootnote ¹⁷ based on Kamerbeek’s observation that ‘μαραίνϵιν connotes in the first place something beautiful that is made to waste away’,Footnote ¹⁸ since the line can be understood as ‘a general maxim of the type “everything comes to an end”’.Footnote ¹⁹ The use itself of a verb like μαραίνω, generally associated with corruption/destruction and having the specific nuance noted by Kamerbeek, ultimately establishes an ambiguous dissonance with the chorus’ enthusiasm for Ajax’s presumed recovery from madness, thus emphasizing the irony of the ode.Footnote ²⁰ This irony would be enhanced by Battaglino’s medical interpretation of μαραίνϵι, as implicitly equating the consumption caused by the chorus’ μέγας χρόνος (possibly an Orphic/Pindaric reminiscence)Footnote ²¹ to that of (Ajax’s) νόσος.Footnote ²² Moreover, this semantic nuance would further contribute to the perception of πάνθ’ … μαραίνϵι as a self-standing gnome, and not requiring a dichotomy μαραίνϵι/ϕλέγϵι based on the semantic field of fire/light.Footnote ²³

A further, original contribution to this issue can be provided by a detailed examination of how at least tragic passages, and specifically excerpts of tragic lyric, are quoted in Stobaeus’ anthology, which, as said above, represents an indirect source for Aj. 714 and could actually be a ‘unique preservation of the truth’Footnote ²⁴ about its original form. This comparative analysis, I argue, may help to gain more confidence on this issue.Footnote ²⁵

From a general examination of Stobaeus’ tragic lyric citations, we can observe a usual coincidence of syntactical and metrical integrity: see Flor. 1.4.3 (Eur. Alc. 962–6), 1.5.6 (Eur. Heracl. 608–9, 615–17), 3.22.17 (Bacch. 395–401), 3.36.13 (ibid. 386–8), 4.8.11 (Soph. OT 873–4), 4.16.11 (Eur. Bacch. 389–94), 4.34.34 (Soph. OT 1189–92).Footnote ²⁶ The very nature of lyric verses, as self-contained metrical units often closely aligned with syntactical coherence, discourages their dismemberment. Very few quotations of incomplete lyric lines are attested, and some of them are not even certain. It may also happen that a sentence is not fully reported, but without this affecting the metrical integrity of the quoted lines. What can be generally inferred is that syntactical integrity may occasionally take precedence over metrical one, or vice versa, but a simultaneous incompleteness of both verses and sentences/propositions, which would be the case of Aj. 714 without τϵ καὶ ϕλέγϵι, is basically avoided.

An identifiable case of quotation of an incomplete line is Flor. 3.1.2 (Eur. Andr. 774–6), where τιμὰ καὶ κλέος are omitted from 774; nevertheless, this is justified by the fact that these words belong to the preceding sentence. Other examples may be Flor. 4.34.22 (Soph. fr. 410 R.²) ἄμοχθος γὰρ οὐδϵίς· ὁ δ’ ἥκιστ’ ἔχων | μακάρτατος, whose metre is uncertain,Footnote ²⁷ if μακάρτατος is seen as the beginning of an incomplete second line,Footnote ²⁸ and Flor. 1.3.3 (fr. 483 Sn.–K.), if the first line ὁρῶ γὰρ χρόνῳ (δ) was originally a 2δ like the second οne. However, these conditions cannot be ascertained, and metrical uncertainty in the former case makes things far less clear; moreover, in the latter case, the dochmiac rhythm is none the less preserved.

Examples of quotations of integral lines, but incomplete sentences, are Flor. 3.37.4 (Soph. El. 1082–3) οὐδϵὶς τῶν ἀγαθῶν ζῶν κακῶς | ϵὔκλϵιαν αἰσχῦναι θέλϵι | [νώνυμος, ὦ παῖ παῖ·] and 2.4.6 (Eur. HF 673–6) οὐ παύσομαι τὰς Χάριτας | ταῖς Μούσαισιν συγκαταμϵι|γνύς, ἡδίσταν συζυγίαν. | μὴ ζῴην μϵτ’ ἀμουσίας | [αἰϵὶ δ’ ἐν στϵϕάνοισιν ϵἴην]. In both cases, however, what is removed from the quotations are additional elements that are not essential to their intelligibility: at Flor. 3.37.4, a predicative complement that basically expresses the same concept as ζῶν κακῶς, and the interjection παῖ παῖ; at Flor. 2.4.6, a coordinate clause whose omission does not affect the self-standing nature of the statement ‘may I not live a life without the Muses’ (676).

A controversial case is Flor. 4.13.4 (Eur. fr. 24b.1–2 K.), also known from the longer quotation at Plut. Mor. 98E and 959C–D (ἦ βραχύ τοι σθένος ἀνέρος· ἀλλά | ποικιλίᾳ πραπίδων | δϵινὰ μὲν πόντου χθονίων τ’ ὀρέων | δάμναται παιδϵύματα): the first two lines are reported, but the phrase to which the second one is syntactically linked is omitted, and the text itself is problematic (nom. plur. ποικιλίαι instead of dat. sing. ποικιλίᾳ). This incompleteness may be due to a subsequent textual loss, or to an interpretation of these lines in Stobaeus as an independent statement like ‘man has slight strength, but resourcefulness of mind’, with σθένος and ποικιλίαι as subjects and an implied copula.Footnote ²⁹ A similar issue is raised by Flor. 3.38.26 (Soph. fr. 353 R.²): in this case, rather than being the subject of a missing phrase (‘for the paths of envy …’)Footnote ³⁰ and a metrically incomplete line, ϕθονϵραὶ γὰρ ὁδοί (fr. 353.4) can be interpreted as an anapaestic monometer that provides a statement with an implied copula, basically meaning ‘the paths of life are full of envy’.Footnote ³¹

In sum, the here-examined exceptions and controversial passages do not ultimately affect the overall judgement of Stobaeus’ treatment of tragic lyric quotations.

At this point, we may examine specific quotations from Ajax. Stobaeus always quotes iambic (31) and anapaestic passages (2), and, except for Aj. 1087 at Flor. 4.1.17 (πρόσθϵν οὗτος ἦν is omitted),Footnote ³² always preserves the metrical integrity of the quoted lines. The only exception is actually Flor. 1.8.24 (Aj. 714), taken from a lyric context and singularly cited after a textual block from the Trugrede (Aj. 646–9), as both sharing the main topic of Flor. 1.8 (‘on the essence of time, its parts, and what it causes’). In general, quotations of isolated tragic lyric verses are extremely rare in Stobaeus: see Flor. 1.3.48b (fr. 500 Sn.–K.), 3.29.5 (Eur. Heracl. 625), 4.39.5 (Soph. Ant. 582/3). This further demonstrates the strong conceptual, syntactical and metrical integrity of Aj. 714, which, as such, discourages any omission in the eventuality (concretely enacted by Stobaeus) of a decontextualized quotation. The substantial coherence and unity of the chorus’ statement is also confirmed by the fact that Stobaeus does not quote the following lines, where κοὐδέν (715) introduces a new syntactical sequence that marks a shift from the universal scope of the gnome to Ajax’s specific situation.

The conjunction τϵ καί adds a second element, having the same grammatical function as the one previously mentioned (two verbs at Aj. 714), which is placed in analogical or oppositional correlation with it, within the same syntactical/semantic unit. Parallels for this usage, among Stobaeus’ tragic quotations, can be found at Flor. 4.13.5 (Eur. fr. 290 K.) ἀϵὶ γὰρ ἄνδρα σκαιὸν ἰσχυρὸν ϕύσϵι | ἧσσον δέδοικα τἀσθϵνοῦς τϵ καὶ σοϕοῦ (substantivized adjective) and 4.41.19.5–6 (Eur. fr. 415.4–5 K.) … τῶν μὲν αὔξϵται βίος, | τῶν δὲ ϕθίνϵι τϵ καὶ θϵρίζϵται πάλιν (verb). Eur. fr. 415 K. can be closely associated with Aj. 714, as both display an alternation between two antithetical actions/events (ϕθίνω/θϵρίζω, μαραίνω/ϕλέγω). In both cases, therefore, what is added by the coordinating conjunction is not redundant, but, serving as one of the two counterparts of a dichotomy, is an indispensable element for meaning and syntax. Furthermore, as noted earlier, the quotation of Aj. 714 matches 646–7, where the dual action of revelation and concealment performed by time is described: this reveals Stobaeus’ (or his source’s) intention of emphasizing the intertextual connection between stasimon and Trugrede. Therefore, there is no reason why the anthologist, had he possessed a text with τϵ καὶ ϕλέγϵι, should not have quoted it.

What can be confidently asserted, given these observations, is that Stobaeus or his source drew on a text that did not contain τϵ καὶ ϕλέγϵι, and thus that its absence is not due to an omission (either arbitrary or accidental) by the anthologist. It is therefore more reasonable to posit that τϵ καὶ ϕλέγϵι later entered the text by the work of a reader/commentator, who, like Stobaeus (and, potentially, any ancient or modern reader of Sophocles), noted the correspondence between Aj. 646–7 and 714, and added in the margins καὶ ϕλέγϵι, selecting a verb from Aj. 673, to establish an even closer connection with the imagery of time and nature in Ajax’s monologue. The interpolation would have been facilitated by the presence of καί, and subsequently involved the addition of the enclitic τϵ, which gave the phrase a metrical pattern (ia) compatible with the overall iambic-choriambic rhythm of the ode.