Romero ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 ISSN 2159-3159 Article copyright held by the author(s) and made available under the Creative Commons Attribution License CC-BY https://creativecommons.org/licenses/by/4.0/ Parenthetical ὃ µὴ γένοιτο (vel sim.) in Classical Greek: A Structural Analysis María López Romero N THE FIFTH LINE of Seven against Thebes, for the first time in Greek literature a relative clause appears that still survives as it is in Modern Greek (Aesch. Sept. 1–9):1 [1] Κάδµου πολῖται, χρὴ λέγειν τὰ καίρια ὅστις φυλάσσει πρᾶγος ἐν πρύµνῃ πόλεως οἴακα νωµῶν, βλέφαρα µὴ κοιµῶν ὕπνῳ. εἰ µὲν γὰρ εὖ πράξαιµεν, αἰτία θεοῦ· εἰ δ’ αὖθ’, ὃ µὴ γένοιτο, συµφορὰ τύχοι, 5 Ἐτεοκλέης ἂν εἷς πολὺς κατὰ πτόλιν ὑµνοῖθ’ ὑπ’ ἀστῶν φροιµίοις πολυρρόθοις οἰµώγµασίν θ’, ὧν Ζεὺς ἀλεξητήριος ἐπώνυµος γένοιτο Καδµείων πόλει. Citizens of Cadmus’ land, he who guards the city’s fortunes, con- trolling the helm at its stern, never letting his eyes rest in sleep, has to give the right advice for the situation. For if we should be successful, the responsibility would be god’s; but if on the other hand disaster were to strike—which I wish may not happen!— then Eteocles’ name alone would be repeatedly harped on by the citizens throughout the town amid a noisy surge of terrified wail- ing—from which may Zeus the Defender, true to his title, defend the city of the Cadmeans! 1 The texts cited in the paper are taken from the most recent editions of each author available in the Thesaurus Linguae Graecae; translations are from the most recent Loeb translations unless stated otherwise. In most cases I have been forced to adapt the translation of ὃ µὴ γένοιτο (vel sim.) to reflect the presence of the relative pronoun. I MARÍA LÓPEZ ROMERO 323 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 The same relative clause is found later in other classical authors, especially poets and orators (e.g. Eur. Heracl. 714, Lys. 19.38, Dem. 15.21, Din. 1.66).2 Between the third century B.C. and the first century A.D. there is hardly any trace of ὃ µὴ γένοιτο. We have only one example, εἰ δὲ µαλακισθείη Σωτήριχος, ὃ µὴ γίνοιτο, in an inscription,3 and one, ἐὰν ὃ [µ]ὴ γένο(ιτο) συµβῇ τὸ παιδί(ον) παθεῖν τι ἀνθρώ(πινον)4 in a non-literary papyrus.5 From the second century A.D. onwards, the examples multiply. ὃ µὴ γένοιτο is documented, in coexistence with ὃ µὴ εἴη, in about a dozen non-literary papyri dating from the second century to the eighth.6 In literary texts it reappears in Aristides (Or. 7.13, 10.35, 23.80), Lucian (Abd. 32, Sat. 18 [the relative pronoun is ὅπερ in both cases]), Cassius Dio (55.16.2), and Diogenes Laertius (5.12 [bis]). However, it is not only pagan authors who use the relative clause: it is also frequent in the Church Fathers (esp. John Chry- sostom), though not as recurrent as the form without the relative pronoun, i.e. µὴ γένοιτο.7 In fact, it is believed that (ὃ) µὴ γένοιτο has reached Modern Greek through the influence of religious texts.8 In this paper I deal with ὃ µὴ γένοιτο in the first two centuries of its history. My aim is, on the one hand, to find out whether ὃ µὴ γένοιτο is already fossilised in classical Greek,9 as it is in 2 See T. V. Evans, “The Last of the Optatives,” CP 98 (2003) 70–80, at 74 for further examples. 3 CID V 209.24, 173 B.C., Delphi. 4 TM 18549: BGU IV 1108.10a = C.Pap.Gr. I 9, 5 B.C., Alexandria. 5 To these references should be added Plato Anth.Gr. 5.79.3, if it is true that the epigram is not the work of Plato but of a Hellenistic poet (see §2.2 below). Apart from this, in Polybius (10.25.5) there is a relative clause functionally similar to ὃ µὴ γένοιτο: ὃ µὴ δόξειε τοῖς θεοῖς. 6 B. G. Mandilaras, The Verb in the Greek Νon-Literary Papyri (Athens 1973) §632; Evans, CP 98 (2003) 79. 7 Evans, CP 98 (2003) 75–78. 8 Mandilaras, The Verb §629; Evans, CP 98 (2003) 76. 9 The term fossilised in linguistics is used of a phrase or clause “that has 324 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 Modern Greek, and, on the other hand, to describe the syntac- tical contexts in which it appears. Also included in the analysis are the classical Greek relative clauses that are syntactically and semantically analogous to ὃ µὴ γένοιτο, i.e. relative clauses whose pronoun refers to a state of affairs mentioned in the text and whose predicate is, semantically, eventive and, syntactically, a negated cupitive optative, e.g. Eur. fr.525.1–3: εἰ δ’ εἰς γάµους ἔλθοιµ’—ὃ µὴ τύχοι ποτέ—, / τῶν ἐν δόµοισιν ἡµερευουσῶν ἀεὶ / βελτίον’ ἂν τέκοιµι σώµασιν τέκνα. Consequently, the relative clause found in πράξας δ’ ὃ µὴ τύχοιµι (νοστήσαιµι γάρ) / δίδωµι τήνδε σοῖσι προσπολεῖν δόµοις (Eur. Alc. 1023–1024) is excluded, as it is neither anaphoric nor cataphoric, but acts as the object of πράξας. Another relative clause is excluded because, besides being autonomous or free, it does not present an eventive predi- cate: πύργοις ἀπειλεῖ τοῖσδ’ ἃ µὴ κραίνοι θεός (Aesch. Sept. 549). The relative clause in πύργοις δ’ ἀπειλεῖ δείν’, ἃ µὴ κραίνοι τύχη (Aesch. Sept. 426) also does not fulfil the requirement of the predicate type and so does not qualify for the analysis. The paper is structured as follows. Section §1 reviews scholar- ship on or related to ὃ µὴ γένοιτο and similar relative clauses. Section §2 studies ὃ µὴ γένοιτο in classical Greek: first (§2.1) I present a table with the data, §2.2 discusses the possible fossilisa- tion of this relative clause in classical Greek, and §2.3 addresses the fact that the relative clause is not compatible with certain types of antecedents. The final section (§3) offers the main con- clusions to be drawn. 1. ὃ μὴ γένοιτο (vel sim.) in studies to date In his paper on the history of ὃ µὴ γένοιτο, Evans states that this relative clause “has attracted limited scholarly interest.”10 This statement should be taken with a pinch of salt. Indeed, although become a fixed expression, often preserving a word or grammatical feature that has been lost from the language”: K. Brown and J. Miller, The Cambridge Dictionary of Linguistics (Cambridge 2013) 178. 10 Evans, CP 98 (2003) 70. MARÍA LÓPEZ ROMERO 325 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 ὃ µὴ γένοιτο has not been tackled in many scholarly papers, it has received considerable attention from commentators, editors, and grammarians.11 In what follows, I summarise the contribu- tions that have been made on ὃ µὴ γένοιτο and similar relative clauses (e.g. ὃ µὴ τύχοι, ὃ µὴ συµβαίη, ὃ µὴ εἴη). A good way to begin this bibliographical tour is to explain the different names by which these relative clauses are known. The labels give us important information about some of their charac- teristics. The most widespread denomination is formula.12 This 11 By way of example: schol. Aesch. Sept. 5; H. Weil, Les plaidoyers politiques de Démosthène II (Paris 1886) on [Dem.] 25.30; I. Flagg, The Seven against Thebes of Aeschylus (Boston 1900) on Aesch. Sept. 5; B. L. Gildersleeve, Syntax of Classical Greek from Homer to Demosthenes I (New York 1900) §397; J. R. King, Demosthenes. Speech against Meidias (Oxford 1901) on Dem. 21.209; A. Wester- mann, Ausgewählte Reden des Demosthenes I10 (Berlin 1902) on Dem. 8.51; W. W. Goodwin, Demosthenes. Against Midias (Cambridge 1906) on Dem. 21.209; H. Weil, Les harangues de Démosthène3 (Paris 1912) on Dem. 8.51; A. S. Owen, Euripides. Ion (Oxford 1939) on Ion 731; D. M. MacDowell, Aristophanes. Wasps (Oxford 1971) on Vesp. 535; A. W. Gomme and F. H. Sandbach, Menander. A Commentary (Oxford 1973) on Sam. 728; Mandilaras, The Verb §632; R. Merkel- bach, “Das Testament des Nikeratos (Samia 726–8),” ZPE 19 (1975) 86; A. H. Sommerstein, “Notes on Aristophanes’ Wasps,” CQ 27 (1977) 261–277, at 267 n.1 on Vesp. 535; J. Diggle, Studies on the Text of Euripides (Oxford 1981) 104 on Ion 731; D. M. Bain, Menander. Samia (Warminster 1983) on Sam. 728; D. M. Bain, “Nikeratos’ ‘Will’ (Menander, Samia 727–8),” ZPE 51 (1983) 36; G. O. Hutchinson, Aeschylus. Septem contra Thebas (Oxford 1985) on Sept. 5; J. Henderson, Aristophanes. Lysistrata (Oxford 1987) on Lys. 147; C. Carey, Lysias. Selected Speeches (Cambridge 1989) on Lysias 31.14; G. Xanthaki-Karamanou, Demosthenous katá Meidíou (Athens 1989) on Dem. 21.209; D. M. MacDowell, Demosthenes. Against Meidias (Oxford 1990) on Dem. 21.209; J. Wilkins, Euripi- des. Heraclidae (Oxford 1993) on Heracl. 511; D. J. Mastronarde, Euripides. Phoe- nissae (Cambridge 1994) on Phoen. 242; M. Lamagna, La donna di Samo (Naples 1998) on Sam. 728; Evans, CP 98 (2003) 70–80; Z. P. Biles and D. S. Olson, Aristophanes. Wasps (Oxford 2015) on Vesp. 535; G. Martin, Euripides. Ion (Ber- lin 2018) on Ion 731; P. S. Peek, Herodotus, Histories, Book V (Norman 2018) on Hdt. 5.111.4. 12 Flagg, The Seven on Sept. 5; Gildersleeve, Syntax §397; Mandilaras, The Verb §632; Diggle, Studies 104; Evans, CP 98 (2003) 70–80. 326 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 term, like idiom,13 suggests that ὃ µὴ γένοιτο is structurally fixed. Interjection and scongiuro, on the other hand, refer to the speech act of the relative clauses (sc. an expressive one).14 Finally, there are more neutral terms, such as clause and phrase.15 It is the function of these relative clauses that has received most attention to date. The first mention goes back to a scholium on Aeschylus Sept. 5: οὕτως ἔθος ἀποτρέπεσθαι τὰ βλάσφηµα (“in this way it is customary to turn aside ill-omened words”). Generally, later scholars also attribute an apotropaic or deprecatory function to the relative clauses. Some examples of glosses similar to that of the scholiast include: “a formula to avert an evil that has been mentioned as a possibility”;16 “a formula of wish or prayer to avert disaster”;17 “an apotropaic interjection meant to cancel the effect of ill-omened words”;18 “an expression com- mon enough in all sorts of Greek authors where ill-omened words are uttered and the hope is expressed that they may not turn out to be true.”19 Carey, for his part, commenting on Lysias 13 Mastronarde, Euripides on Phoen. 242. 14 Mastronarde, Euripides on Phoen. 242, and Lamagna, La donna on Men. Sam. 728, respectively. 15 Bain, Menander on Sam. 728, and Martin, Euripides on Ion 731, re- spectively. 16 Gomme and Sandbach, Menander on Sam. 728. 17 Mandilaras, The Verb §632. On the contrary, Biles and Olson (Aristophanes on Vesp. 535) claim that ὃ µὴ γένοιτο is not a prayer. Indeed, it is not according to speech act theory. A prayer is a directive speech act by which the speaker seeks to make the hearer perform an action for his/her own benefit, without forcing him/her to perform it, but actually putting him/her between a rock and a hard place: see R. Risselada, Imperatives and Other Directive Expressions in Latin. A Study in the Pragmatics of a Dead Language (Amsterdam 1993) 47–48. Wishes, on the other hand, are expressive speech acts, although they get close to directive speech acts when the speaker intends the hearer to carry out the desired action (e.g. Hom. Il. 1.20 παῖδα δ’ ἐµοὶ λύσαιτε φίλην, “May you release my dear child to me”). In the case of ὃ µὴ γένοιτο, the optative mood automatically gives the relative clause an expressive value. 18 Mastronarde, Euripides on Phoen. 242. 19 Bain, ZPE 51 (1983) 36. MARÍA LÓPEZ ROMERO 327 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 31.14, states that ὃ µὴ γένοιτο expresses “distaste at the hypoth- esis, not genuine fear of its occurrence.”20 References to the status of ὃ µὴ γένοιτο (vel sim.) are also found in the literature. Mastronarde suggests that ὃ µὴ τύχοι was a colloquialism in classical times, and Evans proposes the same for ὃ µὴ γένοιτο.21 Stevens, on the other hand, does not include them in any of his lists of colloquialisms in tragedy.22 Collard, in his revision of Stevens’s works, does, although he relegates the rela- tive clauses to the section on doubtful colloquialisms.23 In his view, ὃ µὴ γένοιτο and ὃ µὴ τύχοι “may come from plain rather than colloquial speech.”24 Be that as it may, ὃ µὴ γένοιτο ends up belonging to an elevated register in Modern Greek and this seems to have also been the case at the time of the Church Fathers.25 According to Evans, the fact that ὃ µὴ γένοιτο is doc- 20 Carey, Lysias on 31.14. 21 Mastronarde, Euripides on Phoen. 242; Evans, CP 98 (2003) 76–77. 22 P. T. Stevens, “Colloquial Expressions in Euripides,” CQ 31 (1937) 182– 191, “Colloquial Expressions in Aeschylus and Sophocles,” CQ 39 (1945) 95– 105, and Colloquial Expressions in Euripides (Wiesbaden 1976). Stevens, unlike Mastronarde and Evans, does explain his concept of “colloquialism”: “The category of the colloquial is intended to cover such words and phrases as might naturally be used in everyday conversation, but are avoided in distinc- tively poetic writing and in formal and dignified prose. It does not here in- clude what may be called vulgarism, i.e. words of a definitively non-literary character, with exclusively low and vulgar associations, or irregularities in forms of words or in syntax” (CQ 31 [1937] 182); “By colloquial I mean not merely words and expressions that are likely to occur in ordinary conversa- tion, since this consists largely of neutral language, but the kind of language that in a poetic or prosaic context would stand out however slightly as having a distinctively conversational flavour” (Colloquial Expressions 4). 23 C. Collard, “Colloquial Language in Tragedy: A Supplement to the Work of P. T. Stevens,” CQ 55 (2005) 350–386, at 376, and Colloquial Ex- pressions in Greek Tragedy. Revised and Enlarged Edition of P. Τ. Stevens’s Colloquial Expressions in Euripides (Stuttgart 2018) 164. 24 Collard, CQ 55 (2005) 376. 25 Evans, CP 98 (2003) 72, 76. 328 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 umented in Patristic writers, but not in the New Testament, is symptomatic of an Atticism.26 Finally, it is worth reviewing comments made on syntactical aspects. Firstly, Gildersleeve and Peek have drawn attention to the presence of a cupitive optative in the relative clause.27 This is not exceptional (cf. e.g. Hom. Od. 4.699, Soph. Phil. 275, Dem. 18.290), but it is infrequent, as Emde Boas et al. point out.28 Secondly, Diggle and Martin have noted that the relative pro- noun almost always appears in the singular, and this argument has been used to justify the correction of some plural relative pronouns in manuscripts.29 Thirdly, some scholars have noticed that the antecedent of the relative clause is often a conditional clause.30 This and the previous point will be discussed in the next section. Fourthly, attention has been drawn to the clause οὔδ’ ἔσται, which sometimes coordinates with ὃ µὴ γένοιτο, giving rise to asymmetrical coordination.31 Finally, Flagg, Mandilaras, and Sommerstein qualify ὃ µὴ γένοιτο as “parenthetic,” which reveals that these scholars are aware that the construction is not essential to the internal structure of the main clause.32 2. ὃ μὴ γένοιτο (vel sim.) in classical Greek 2.1. Data As mentioned in the introduction, the earliest example of the relative clauses under study appears at the beginning of 26 Evans, CP 98 (2003) 76. 27 Gildersleeve, Syntax §397; Peek, Herodotus on Hdt. 5.111.4. 28 E. van Emde Boas, A. Rijksbaron, L. Huitink, and M. de Bakker, The Cambridge Grammar of Classical Greek (Cambridge 2019) §50.17. 29 Diggle, Studies 104; Martin, Euripides on Ion 731. 30 E.g. schol. Aesch. Sept. 5; MacDowell, Aristophanes on Vesp. 535; Som- merstein, CQ 27 (1977) 267 n.1. 31 See Westermann, Ausgewählte Reden on Dem. 8.51; Weil, Les harangues on Dem. 8.51; Goodwin, Demosthenes on Dem. 21.209; Xanthaki-Karamanou, Demosthenous on Dem. 21.209; MacDowell, Demosthenes on Dem. 21.209. 32 Flagg, The Seven on Sept. 5; Mandilaras, The Verb §632; Sommerstein, CQ 27 (1977) 267 n.1. MARÍA LÓPEZ ROMERO 329 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 Aeschylus’ Seven against Thebes (ὃ µὴ γένοιτο).33 From then until the end of the classical period, save for error or omission, 26 examples are documented, all belonging to literary texts.34 In order of frequency, the relative clauses are: ὃ µὴ γένοιτο, which admits variations (20 examples); ὃ µὴ τύχοι (ποτέ) (4); and ὃ µὴ συµβαίη (2). Table 1 shows the distribution of the examples per author. The relative clause is ὃ µὴ γένοιτ(ο) (ad litteram) unless otherwise indicated. TABLE 1: Distribution of the relative clauses per author Century Author Number Reference VI–V Aeschylus 1 Sept. 5 V Herodotus 1 5.111.4 τὸ µὴ γένοιτο Euripides 6 fr.525 ὃ µὴ τύχοι ποτέ Heracl. 511 ὃ µὴ τύχοι ποτέ 714 Ion 731 Phoen. 242 ὃ µὴ τύχοι 571 ὃ µὴ τύχοι ποτέ V–IV Aristophanes 2 Vesp. 535 Lys. 147 Lysias 2 19.38 ὃ µὴ γένοιτο, εἰ µή τι µέλλει µέγα ἀγαθὸν ἔσεσθαι τῇ πόλει 31.14 33 Homer, in any case, uses a relative clause functionally similar to ὃ µὴ γένοιτο in Od. 4.698–699: ἀλλὰ πολὺ µεῖζόν τε καὶ ἀργαλεώτερον ἄλλο / µνηστῆρες φράζονται, ὃ µὴ τελέσειε Κρονίων (“But another far greater and more grievous are the suitors planning—which I wish that the son of Cronus may never bring to pass!”). 34 In addition, there are four examples of ὃ µὴ γένοιτο in two of Demos- thenes’ speeches that in Dilts’ edition (Demosthenis orationes I–IV [Oxford 2002–2009]) are listed as spurious: 25.30, 25.32, 25.95 (here the relative clause is ὃ µήτ’ ἰδίᾳ µήτε δηµοσίᾳ γένοιτο), and 40.56. In a fragment tra- ditionally attributed to Euripides, but whose authorship is not clear today (cf. E. Bonollo, “La mulieris oratio del papiro Didot (= adesp. com. fr. 1000 K.-A.) come esercizio di scuola,” Incontri di filologia classica 17 [2017/8] 245–266, for a synthesis of the dispute), we find ὃ µὴ γένοιτο, Ζεῦ φίλ’, οὐδ’ ἔσται ποτέ. These five examples have been excluded from the analysis. 330 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 Plato 2 Leg. 918D8–9 ὃ µή ποτε γένοιτο οὐδ’ ἔσται Anth.Gr. 5.79 ὃ µὴ γίγνοιτο IV Demosthenes 9 8.51 ἃ µήτε γένοιτ’ οὔτε λέγειν ἄξιον 10.27 ὃ µήτε γένοιτ’ οὔτε λέγειν ἄξιον 15.21 21.209 ὃ µὴ γένοιτο οὐδ’ ἔσται 27.67 28.21 36.49 Prooem. 22.3 ὃ µὴ συµβαίη 26.2 ὃ µὴ συµβαίη IV–III Dinarchus 1 1.66 Menander 2 Mis. 264 Sam. 728 The data confirms that in classical Greek the relative clause ὃ µὴ γένοιτο (vel sim.) is characteristic of dramatic poets and orators. It is thus linked to interactional contexts.35 In fact, in both Herodotus’ and Plato’s examples a listener is involved: Herodotus’ is framed in a conversation rendered in direct speech, whilst in Plato one appears in a dialogue (Leg.) and the other in an epigram addressed to a girl. Furthermore, the table shows that ὃ µὴ γένοιτο, with or without additions, is the relative clause most used by all authors except Euripides. Euripides prefers ὃ µὴ τύχοι (ποτέ), which seems to be idiosyncratic as it is not documented elsewhere. Finally, it is also striking that ὃ µὴ συµβαίη appears only in Demosthenes and, specifically, in the Prooemia. However, unlike ὃ µὴ τύχοι (ποτέ), the story of ὃ µὴ συµ- βαίη does not begin and end with Demosthenes: centuries later, it is used by Aristides (Or. 10.39.12) and Justinian I (Cod.Iust. 8.10.12.5d). 2.2. Fossilised relative clauses? Labels like “formula” and “formulaic,” applied in particular to ὃ µὴ γένοιτο and ὃ µὴ τύχοι (ποτέ), suggest, as mentioned in §1, 35 Cf. Evans, CP 98 (2003) 74. MARÍA LÓPEZ ROMERO 331 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 that the relative clauses are structurally fixed. Indeed, on that basis Diggle justifies correcting ἅ (Laur.plut. 32.2) to ὅ in ἃ µὴ τύχοι ποτέ (Eur. Heracl. 511) and ἃ µὴ γένοιτο (Ion 731), proposed by Lenting and Stephanus, respectively.36 There is reason, how- ever, to believe that at least ὃ µὴ γένοιτο and ὃ µὴ τύχοι (ποτέ) are not fossilised in classical Greek. We do not have sufficient examples of ὃ µὴ συµβαίη for a reliable comparison. Firstly, both ὃ µὴ γένοιτο and ὃ µὴ τύχοι (ποτέ) still allow a certain constructive freedom. ὃ µὴ γένοιτο twice carries as an appendix the clause οὐδ’ ἔσται. What is more, the examples are not exactly the same: ὃ µή ποτε γένοιτο οὐδ’ ἔσται (Pl. Leg. 918D8– 9) and ὃ µὴ γένοιτο οὐδ’ ἔσται (Dem. 21.209). ὃ µὴ γένοιτο also appears alongside οὔτε λέγειν ἄξιον in two of Demosthenes’ speeches (8.51, 10.27).37 Finally, there is also one example of ὃ 36 Diggle, Studies 104; J. Lenting, Epistola critica in Euripidis Alcestin (Zutphen 1821) 90; H. Stephanus, Annotationes in Sophoclem et Euripidem (1568) 189. On the contrary, A. F. Braunlich (“Notes on the Text of Euripides,” AJP 83 [1962] 393–411, at 395) and P. T. Stevens (Euripides. Andromache [Oxford 1971] on Andr. 271) accept ἅ, arguing that there are more cases of “illogical uses” (Braunlich’s words) of relative pronouns in the neuter plural, namely Eur. HF 745, IT 613 (here, however, the relative has an adverbial use; cf. LSJ s.v. ὅσος IV), and Ion 65 (see M. López Romero, Oraciones de relativo en el drama ático clásico [Seville 2023] 154–155, for more examples). Perhaps editors who respect the paradosis in Eur. Heracl. 511 and Ion 731 also rely on such parallels (e.g. G. Murray, Euripidis fabulae3 I and II [Oxford 1902, 1913]); L. Méridier, Euripide. Tragédies5 I [Paris 1961]); L. Parmentier and H. Grégoire, Euripide. Tragédies III [Paris 1924]; A. Garzya, Euripides. Heraclidae [Leipzig 1972]; W. Biehl, Euripides. Ion [Leipzig 1979]; E. Calderón Dorda, Eurípides. Tragedias VI [Madrid 2007]). There are also editors who maintain ἅ in Heracl. 511, but not in Ion 731 (e.g. F. A. Paley, Euripides I and II [London 1857, 1858]; A. Nauck, Euripidis tragoediae3 I and II [Leipzig 1871, 1884]); N. Wecklein, Euripidis Heraclidae [Leipzig 1898], and Euripidis Ion [Leipzig 1898]). This is discussed below. 37 The second speech, whose authenticity has been questioned, contains parts taken from other speeches by Demosthenes. The relative clause appears in one of these sections, although the number of the relative pronoun varies: ἅ in Dem. 8.51 and ὅ in 10.27. However, Monac.gr. 485 transmits ἅ (and µή 332 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 µὴ γένοιτο with an adjunct: ὃ µὴ γένοιτο, εἰ µή τι µέλλει µέγα ἀγαθὸν ἔσεσθαι τῇ πόλει (Lysias 19.38). With ὃ µὴ τύχοι (ποτέ) the reverse is true, in the sense that the form with an adjunct (sc. ποτέ) is the most common one (Eur. fr.525, Heracl. 511, Phoen. 571). How- ever, the existence of a relative clause without ποτέ (Eur. Phoen. 242) is proof that it is not entirely fixed.38 Secondly, there is also one example of ὃ µὴ γίγνοιτο with the verb in the present stem, belonging to an erotic epigram attrib- uted to Plato (Anth.Gr. 5.79.3). This epigram, along with others on the same subject also attributed to Plato, is probably spurious. Everything seems to suggest that it was composed by a Hel- lenistic poet.39 If so, the form γίγνοιτο would indicate that, at least until the end of the fourth century B.C., ὃ µὴ γένοιτο was not structurally fixed.40 There are further examples pointing in favour of a late fossilisation of ὃ µὴ γένοιτο (e.g. Lucian Abdic. 32, Saturn. 18; Epiphanius Adv.haeres. II p.184 Holl), but these are beyond the scope of the present work.41 instead of µήτε) in 10.27. The plural is the expected form, firstly, because the relative clause has split antecedents (πληγαὶ καὶ ὁ τοῦ σώµατος αἰκισµός) and, secondly, because the speech of which 10.27 is a copy (sc. 8.51) presents ἅ. The singular ὅ is probably due to the weight of discursive tradition. 38 Interestingly, ποτέ is found in some codices (Cremon. 130, Laurent. 32.2, Vat.gr. 51, Ambros. I 47 sup., and Cantabr.Nn. 3.14, where ποτέ appears as an alternative reading written by a second hand). This form does not fit the metre and its presence in the codices may be due to two reasons: (i) the copyists were aware that ὃ µὴ τύχοι ποτέ is the usual form or (ii) the copyists had the relative clause ὃ µὴ τύχοι ποτέ of Phoen. 571 in mind. 39 Cf. W. Ludwig, “Plato’s Love Epigrams,” GRBS 4 (1963) 59–82, at 75– 77. 40 One of the anonymous reviewers suggests an alternative possibility: “that the Hellenistic poet toyed, for the sake of poetic variation, with what was in fact a largely fixed expression.” 41 Cf. Evans, CP 98 (2003) 78–80. In any case, I have analysed the instances of ὃ µὴ γένοιτο recorded in the TLG from the 2nd to the 19th century. I identified 154 examples, excluding those repeated and those of uncertain chronology. So far as I have been able to determine, ὃ µὴ γένοιτο is attested MARÍA LÓPEZ ROMERO 333 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 Thirdly and finally, again with respect to ὃ µὴ γένοιτο, the relative pronoun appears in the neuter plural in three examples, two of which, as advanced at the beginning of this section, have been the subject of dispute (Eur. Heracl. 511 and Ion 731). The example exempt from discussion is [2]. As stated in n.37, the plural ἅ is easily explained here, since the relative clause clearly has two antecedents, namely πληγαὶ καὶ ὁ τοῦ σώµατος αἰκισµός (Dem. 8.51): [2] διαφέρει δὲ τί; ὅτι ἐστὶν ἐλευθέρῳ µὲν ἀνθρώπῳ µεγίστη ἀνάγκη ἡ ὑπὲρ τῶν γιγνοµένων αἰσχύνη, καὶ µείζω ταύτης οὐκ οἶδ’ ἥντιν’ ἂν εἴποιµεν· δούλῳ δὲ πληγαὶ καὶ ὁ τοῦ σώµατος αἰκισµός, ἃ µήτε γένοιτ’ οὔτε λέγειν ἄξιον. almost continuously throughout this period. The most numerous examples are found in the 4th century (18, 10 of which are from Libanius), the 4th–5th centuries (18, 17 from John Chrysostom), the 13th–14th centuries (9), the 14th century (10), the 14th–15th centuries (14), the 16th century (12), and the 18th– 19th centuries (8). In general, the basic components of the structure (ὅ, µή, and γένοιτο) remain invariable, even when the antecedent is in the plural or consists of multiple elements (e.g. in Libanius Epist. 1202). Only once is the relative pronoun in the plural (ἃ καὶ µὴ πόρρω µοι νοῦν ἔχοντι γένοιντο in Andreas Libadenus Epist. ad Gerasimum 2, 14th cent.), and only once is ὅπερ used instead of ὅ (ὅπερ µὴ γένοιτο ἀκοῦσαι ὑµᾶς, John II Epist. ad Clementem III 10, 11th cent.). As in classical Greek, the basic structure allows for coordina- tion with another clause (3rd–5th and 12th–15th centuries; e.g. ὃ µὴ γένοιτο µήδ’ ἀκούσοιµί του διηγουµένου in Michael Apostolius Sermo ad socerum p.144.7 Stefec) and the incorporation of an adjunct (4th–5th, 9th, and 13th–17th cent.; e.g. ὃ µὴ γένοιτο πώποτε in Theodorus II Lascaris Epist. 15). Notably, there are also examples of ὃ µὴ γένοιτο used with a personal pronoun in the dative (14th and 18th–19th cent.; e.g. ὃ µὴ γένοιτο ἡµῖν in Nicolaus Cabasilas In passionem salvatoris 27) and with an infinitival clause (9th and 11th–14th cent.; e.g. ὃ µὴ γένοιτό τινα παθεῖν ὑµῶν in Germanus II Orationes et homiliae 14, p.331.12 Lagopates). All of this seems to indicate that the structure, despite the in- variability of its core components (ὅ, µή, and γένοιτο), never lost its com- positionality. That is, ὃ µὴ γένοιτο was always interpreted as a desiderative clause. 334 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 Do you ask the difference? The strongest necessity that a free man feels is shame for his own position, and I know not if we could name a stronger; but for a slave necessity means stripes and bodily outrage—things that I wish may not happen and are unfit to name here! On the other hand, ἅ in Ion 730–732 [3] does not refer to several states of affairs, but only to one, namely τι τυγχάνοι κακόν. For this reason, and not so much because the singular is the regular number as Diggle claims,42 the change from ἅ to ὅ proposed by Stephanus is justified:43 [3] σὺν τοῖς φίλοις γὰρ ἡδὺ µὲν πράσσειν καλῶς· ἃ (ms.)/ὃ (Stephanus) µὴ γένοιτο δ’, εἴ τι τυγχάνοι κακόν, ἐς ὄµµατ’ εὔνου φωτὸς ἐµβλέψαι γλυκύ. It is a pleasure to share good fortune with those we love. But if— which I wish may not happen!—some trouble comes, the sight of a friendly face is soothing. In Heracl. 511–514 [4], the position of the relative clause suggests that τῆσδ’… πόλεως ἁλούσης is the antecedent, since, as will be shown in the next section, the relative clause and its antecedent are always contiguous. If this is the case, the number of the relative pronoun poses a problem, which has been cir- cumvented by correcting ἅ or by considering ἅ an exceptional use (see n.36). In my view, however, the plural can be defended by understanding that the relative pronoun refers to (i) τῆσδ’… πόλεως ἁλούσης, (ii) χεῖρας εἰς ἐχθρῶν πεσεῖν, and (iii) ἄτιµα πατρὸς οὖσαν εὐγενοῦς παθοῦσαν Ἅιδην µηδὲν ἧσσον εἰσιδεῖν. All three states of affairs, despite what κάλλιον might suggest, are un- desired by the speaker. The speaker is being ironic.44 Perhaps the editors who print ἅ here, but ὅ in Ion 731 (see n.36), under- 42 Diggle, Studies 104. 43 It remains unexplained, however, how ἅ could replace ὅ in the trans- mission, if ὅ is the usual form and, above all, the logical one from a syntactic point of view. 44 Paley, Euripides I, A. C. Pearson, Euripides. The Heraclidae (Cambridge 1907), and Wilkins, Euripides, on Heracl. 511. MARÍA LÓPEZ ROMERO 335 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 stand the passage in this way.45 They need not be ignorant of Lenting's conjecture, as Diggle says.46 In fact, at least Wecklein is aware of it, for he mentions it in his apparatus.47 [4] κάλλιον, οἶµαι, τῆσδ’—ἃ (ms.)/ὃ (Lenting) µὴ τύχοι ποτέ— πόλεως ἁλούσης χεῖρας εἰς ἐχθρῶν πεσεῖν κἄπειτ’ ἄτιµα πατρὸς οὖσαν εὐγενοῦς παθοῦσαν Ἅιδην µηδὲν ἧσσον εἰσιδεῖν. Much finer, I suppose, if this city were to be captured—things that I wish may never occur!—and I were to fall into the hands of the enemy! Then when I, daughter of a noble father, have suffered dishonor, I shall go to my death all the same! 2.3. The antecedent of ὃ μὴ γένοιτο (vel sim.) The verbs γίγνοµαι, τυγχάνω, and συµβαίνω of the relative clauses under study are semantically eventive, i.e. they predicate that something happens (LSJ s.v. γίγνοµαι I.3, τυγχάνω Α.2, συµ- βαίνω ΙΙΙ). As such, they require their first argument, the subject, to designate a state of affairs.48 This explains why the antecedent of ὅ—which functions as subject in the relative clauses—is always a clause (see examples [1], [3], and [4]) or a term referring to a second-order entity,49 i.e. a state of affairs (example [2]). Table 2 shows the syntactic realisations of the antecedents of ὃ µὴ γένοιτο in classical Greek: 45 But not Nauck (Euripidis I ap.crit.), for, parallel to Lenting, he proposes to correct ἅ in Heracl. 511. 46 Diggle, Studies 104. 47 Wecklein, Euripidis Heraclidae. 48 Cf. R. Martínez Vázquez, E. Ruiz Yamuza, and R. Fernández Garrido, Gramática funcional-cognitiva del griego antiguo. Sintaxis y semántica de la predicación I (Seville 1999) 72–74, 95–96. 49 Cf. J. Lyons, Semantics II (Cambridge 1977) 443–445. 336 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 TABLE 2: Types of antecedents of ὃ µὴ γένοιτο (vel sim.) Type of antecedent no. Reference and antecedent Clause Main 2 Eur. Phoen. 242 νῦν δέ µοι […] πόλει Men. Mis. 264 οἴχεται Θρασωνίδης Subordinate 22 Aesch. Sept. 5 εἰ δ’ αὖθ(ι) συµφορὰ τύχοι Hdt. 5.111.4 ἢν σὲ ἐκεῖνος Eur. fr.525 εἰ δ’ εἰς γάµους ἔλθοιµ’ Heracl. 511 τῆσδ(ε) πόλεως ἁλούσης […] εἰσιδεῖν 714 ἢν δ’ οὖν χρήσωνται τύχηι Ion 731 εἴ τι τυγχάνοι κακόν Phoen. 571 ἢν ἕληις γῆν τήνδ’ Ar. Vesp. 535 εἰ γὰρ οὗτός σε λέγων κρατήσει Lys. 147 εἰ δ’ ὡς µάλιστ’ ἀπεχοίµεθ’ οὗ σὺ δὴ λέγεις Lysias 19.38 εἰ δηµεύσαιτε τὰ τοῦ Τιµοθέου 31.14 ἐάν ποτε λάβωσι τὴν πόλιν Pl. Leg. 918D8–9 ἐπεὶ εἴ τις προσαναγκά- σειεν […] τρόπου Anth.Gr. 5.79.3 εἰ δ’ ἄρ(α) νοεῖς Dem. 15.215 εἴ ποτε τοιοῦτό τι συµβαίη 21.209 εἰ γένοιντο οὗτοι κύριοι τῆς πολι- τείας µετὰ Μειδίου καὶ τῶν ὁµοίων τούτῳ 27.67 ἂν γὰρ ἀποφύγῃ µε οὗτος 28.21 εἰ δ’ ὑµεῖς ἄλλο τι γνώσεσθε 36.49.7 ἐὰν ἐξαπατηθῶσιν οὗτοι Prooem. 22.3 εἴ τι γένοιτο 26.2 ἂν δέ τι συµβῇ παρ’ ἃ νῦν οἴεσθε Din. 1.66 ἐὰν παρακρουσθῆθ’ ὑπὸ τῆς τούτου γοητείας Men. Sam. 728 ὅταν ἀποθάνω γ’ Second- order term Nominal 2 Dem. 8.51 πληγαὶ καὶ ὁ τοῦ σώµατος αἰκισµός 10.27 πληγαὶ καὶ ὁ τοῦ σώµατος αἰκισµός The semantics of the predicate of the relative clauses therefore limits the antecedents of ὅ (or ἅ) to states of affairs. Moreover, the mood of the predicate imposes a further restriction in that MARÍA LÓPEZ ROMERO 337 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 the state of affairs to which the relative pronoun refers has not been realised, but may yet occur in the speaker’s judgement (e.g. adding ὃ µὴ γένοιτο to Θουκυδίδης Ἀθηναῖος ξυνέγραψε τὸν πόλεµον τῶν Πελοποννησίων καὶ Ἀθηναίων would be ungrammatical).50 For this reason, conditional clauses, which prototypically refer to states of affairs that can be realised or not,51 are breeding grounds for the use of ὃ µὴ γένοιτο.52 Indeed, the anchor of ὃ µὴ γένοιτο is most often a conditional protasis (20 of 26 examples, Table 2), in particular, neutral (3 examples),53 prospective (7),54 and potential (10).55 The connection between protases and ὃ µὴ γένοιτο is especially evident when the relative clause appears within the protasis itself (10 examples), either at the beginning (6 50 Recall the words of Gomme and Sandbach on Men. Sam. 728, quoted §1 above: “A formula to avert an evil that has been mentioned as a possibility” (italics mine). 51 “Disjunctive situation” in C. Lehmann, in M. Romportl et al. (eds.), Linguistica Generalia I Studies in Linguistic Typology (Prague 1974) 231–241, at 238. See also G. Wakker, Conditions and Conditionals. An Investigation of Ancient Greek (Amsterdam 1994) 44–45. 52 Likewise, a genitive absolute with conditional value can serve as an anchor of ὃ µὴ γένοιτο. This is the case in Eur. Heracl. 511 (example [4]). Cf. mutatis mutandis, Polyb. 10.25: κἂν µὲν οὗτοι πταίσαντες καταφθαρῶσιν, ἀνα- στρέψαντες ἐκ τῶν πραγµάτων ἀβλαβεῖς ἀπολυθήσονται Ῥωµαῖοι· νικησάντων δὲ τούτων, ὃ µὴ δόξειε τοῖς θεοῖς, ἅµα τούτοις καὶ τοὺς ἄλλους Ἕλληνας ὑφ’ αὑτοὺς ἐκεῖνοι ποιήσονται (“If the former are beaten and destroyed, the Romans will get away unharmed from the struggle, but should the Aetolians be victorious—which Heaven forbid!—the Romans will subjugate them as well as all the other Greeks”). Note the contrast between the conditional pro- tasis (µέν) and the genitive absolute (δέ). 53 Ar. Vesp. 535, Pl. Anth.Gr. 5.79.3, Dem. 28.21. 54 Hdt. 5.111.4; Eur. Heracl. 714 (insubordinate conditional; cf. E. Ruiz Yamuza, “Condicionales interrogativas truncadas en griego antiguo. El tipo ‘¿Y si no te mueres?’,” Emerita 92 [2024] 2–20), Phoen. 571; Lysias 31.14; Dem. 27.67, Prooem. 26.2; Din. 1.66. 55 Aesch. Sept. 5; Eur. fr.525, Ion 731; Ar. Lys. 147; Lysias 19.38; Pl. Leg. 918D8–9; Dem. 15.21, 21.209, 36.49, Prooem. 22.3. 338 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 of 10),56 as in example (1), or in medial position (4 of 10),57 as in Dem. 21.209: [5] ἐνθυµεῖσθ’, ὦ ἄνδρες δικασταί, εἰ γένοιντο (ὃ µὴ γένοιτο οὐδ’ ἔσται) οὗτοι κύριοι τῆς πολιτείας µετὰ Μειδίου καὶ τῶν ὁµοίων τούτῳ, καί τις ὑµῶν τῶν πολλῶν καὶ δηµοτικῶν ἀνθρώ- πων ἁµαρτὼν εἴς τινα τούτων, µὴ τοιαῦθ’ οἷα Μειδίας εἰς ἐµέ, ἀλλ’ ὁτιοῦν ἄλλο, εἰς δικαστήριον εἰσίοι πεπληρωµένον ἐκ τούτων, τίνος συγγνώµης ἢ τίνος ἐλέου τυχεῖν ἂν οἴεσθε; Suppose, gentlemen of the jury, that these men—which I wish may not happen (and indeed it never will)!—made themselves masters of the State, along with Meidias and others like him; and suppose that one of you, who are men of the people and friends to popular government, having offended one of these men—not so seriously as Meidias offended me, but in some slighter degree— came before a jury packed with men of that class; what pardon, what consideration do you think he would receive? Moreover, the relative clause can be placed in front of the protasis—although this, as Martin points out,58 is exceptional, as it happens only in Eur. Ion 731 (example [3])—or behind the protasis (9 examples),59 as in example [6]. ὃ µὴ γένοιτο here is between the protasis and the apodosis. The selection of the pro- tasis as the anchor is determined by the context: we know that of the two states of affairs presented as possible (sc. sexual ab- stinence and the coming of peace) only the former is undesired by the speaker (cf. also Eur. fr.525 and Dem. Prooem. 22.3). Furthermore, as one of the anonymous reviewers points out, the resumptive διὰ τουτογί in example [6] also encourages con- necting ὃ µὴ γένοιτο with the preceding context, rather than with what follows (Ar. Lys. 146–148): 56 Aesch. Sept. 5; Eur. Heracl. 714; Ar. Vesp. 535; Pl. Anth.Gr. 5.79.3; Dem. 36.49; Din. 1.66. 57 Lysias 31.14; Pl. Leg. 918D8–9; Dem. 15.21, 21.209. 58 Martin, Euripides on Ion 731. 59 Hdt. 5.111.4; Eur. fr.525, Phoen. 571; Ar. Lys. 147; Lysias 19.38 (behind the first protasis); Dem. 27.67, 28.21, Prooem. 22.3, Prooem. 26.2. MARÍA LÓPEZ ROMERO 339 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 [6] εἰ δ’ ὡς µάλιστ’ ἀπεχοίµεθ’ οὗ σὺ δὴ λέγεις, ὃ µὴ γένοιτο, µᾶλλον ἂν διὰ τουτογὶ γένοιτ’ ἂν εἰρήνη; Well, what if we did abstain from, uh, what you say—which I wish may not happen!—would peace be likelier to come on that ac- count? As we can see in Table 2, ὃ µὴ γένοιτο (vel sim.) is not only compatible with conditional protases. On one occasion (example [7]) it pivots on an apodosis. This is possible because conditional apodoses also designate unrealised states of affairs, with the ex- ception of the apodoses of habitual conditionals (Men. Mis. 262– 264): [7] εἰ µὴ γὰρ οὗτος δοκιµάσει µε, κυρίως δώσει τε ταύτην, οἴχεται Θρασωνίδης· ὃ µὴ γένοιτ’. ἀλλ’ εἰσίωµεν· Suppose he doesn’t approve of me, or give her formally in mar- riage. Then Thrasonides is done for—which I wish may not happen! Well, let’s go in. Likewise, a temporal clause with future reference can host ὃ µὴ γένοιτο. Of this combination there is only one example, which is found in Menander (Sam. 726–728): [8] µαρτύρων ἐναντίον σοι τήνδ’ ἐγὼ δίδωµ’ ἔχειν γνησίων παίδων ἐπ’ ἀρότῳ, προῖκα τἀµὰ πάνθ’ ὅταν ἀποθάνω γ’, ὃ µὴ γένοιτ’, ἀλλ’ ⟨εἰς⟩ αεὶ ζῴην. In front of witnesses I give this girl to you as your wife, to harvest lawful children. And a dowry—all my possessions—when I die— which I wish may not happen, and I live for ever! In Eur. Phoen. 239–242, the relative clause ὃ µὴ τύχοι is anchored to a factual assertion: [9] νῦν δέ µοι πρὸ τειχέων θούριος µολὼν Ἄρης αἷµα δάϊον φλέγει τᾷδ’, ὃ µὴ τύχοι, πόλει. But now before the walls grim Ares has come and sets war and bloodshed ablaze for this city: may heaven avert it! 340 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 In this example, the relative clause has no obvious fit, because, as mentioned, the state of affairs of the supraordinate is factual. The action expressed by φλέγει is taking place at the moment of speech (cf. 239 νῦν). Commentators have overlooked this prob- lem but some translators have noticed it. Kovacs, for example, renders τύχοι as “avert.”60 Similarly, Coleridge freely translates the relative clause: “[Ares] is kindling a murderous blaze—may he not succeed!—for this city” (τυγχάνω may mean “succeed,” but in this case the first argument would be expected to have animate referent; cf. LSJ s.v. τυγχάνω B).61 García Gual and De Cuenca y Prado do the same and make the relative clause de- pend on a final clause that is not in the Greek text: “Se presenta un furioso Ares para incendiar en sangre y llamas—¡lo que ojalá no consiga!—esta ciudad.”62 In my opinion, φλέγει is a phasal verb expressing the beginning of an action (“ignite/kindle”; cf. ἀνάπτει as a gloss for φλέγει, schol. Phoen. 241). Thus, the speaker’s wish is that the conflict in fieri should not ravage the city. Another possibility, suggested by one of the anonymous re- viewers, is to take φλέγει as conative. According to this analysis, the speaker’s wish is that Ares not bring the bloodshed to com- pletion. Finally, ὃ µὴ γένοιτο has non-sentential antecedents in two examples. One is Dem. 8.51 ([2] with n.37), the other is 10.27, which is a calque of 8.51. 3. Conclusions This paper has analysed the parenthetical relative clauses ὃ µὴ γένοιτο, ὃ µὴ τύχοι (ποτέ), and ὃ µὴ συµβαίη in classical Greek. My aim has been, on the one hand, to determine whether such rela- tive clauses are fossilised in classical Greek (with special interest in ὃ µὴ γένοιτο, which survives in Modern Greek) and, on the 60 D. Kovacs, Euripides V (Cambridge 2002). 61 E. P. Coleridge, in W. J. Oates et al., The Complete Greek Drama II (New York 1938). 62 Eurípides. Tragedias III (Madrid 1979). MARÍA LÓPEZ ROMERO 341 ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 other hand, to substantiate the perception that these relative clauses are frequently linked to conditional protases. In the course of the paper these two questions have been answered, so that what follows is no more than a recapitulation of what has been said. Regarding the first research question, the data indicates that ὃ µὴ γένοιτο, the most frequent of the three relative clauses, is not yet fossilised in classical Greek. Firstly, it retains a certain con- structive freedom, for it admits coordination, albeit asymmetri- cal, with οὔδ’ ἔσται (Pl. Leg. 918D8–9 and Dem. 21.209) and οὔτε λέγειν ἄξιον (Dem. 8.51 and 10.27), and adjunct complemen- tation (Lysias 19.38). Secondly, on one occasion (Pl. Anth.Gr. 5.79.3) the verb of the relative clause is in the present stem. This example is probably from the Hellenistic period, but it serves, in any case, as a terminus post quem to date the fixity of the struc- ture. Thirdly, the relative pronoun is sensitive to the number of the antecedent. In three examples it appears in the plural (Eur. Heracl. 511, Ion 731; Dem. 8.51), although in one of them (Ion 731) the plural is not syntactically justified. Despite what has been said, it cannot be denied that ὃ µὴ γένοιτο has a recurrent structure: in thirteen out of twenty examples it appears as such without any variation. It is therefore what Kaltenböck et al. refer to as a constructional parenthetical.63 Of ὃ µὴ τύχοι (ποτέ) and ὃ µὴ συµβαίη there are fewer examples than of ὃ µὴ γένοιτο (four 63 G. Kaltenböck, B. Heine, and T. Kuteva, “On Thetical Grammar,” Studies in Language 35 (2011) 838–893, at 870–872. They classify parentheticals or, as they call them, theticals into three types according to their structure: instantaneous, constructional, and formulaic. Instantaneous parentheticals are “fully compositional, can be formed freely anytime and anywhere, can be inserted in most syntactic slots of a sentence, and quite a few of them are uttered only once and never again.” Constructional parentheticals are also compositional but have some schematic structure and function. Formulaic parentheticals are “non-compositional information units, that is, their shape is essentially invariable. They are usually short chunks, morphosyntactically unanalyzable, tend to be positionally flexible and to express functions that are mostly procedural, and they relate to the situation of discourse rather than to sentence syntax.” 342 PARENTHETICAL Ο ΜΗ ΓΕΝΟΙΤΟ ————— Greek, Roman, and Byzantine Studies 65 (2025) 322–342 and two, respectively). However, ὃ µὴ τύχοι (ποτέ) does not seem to be structurally fixed. The usual form is ὃ µὴ τύχοι ποτέ, but on one occasion (Eur. Phoen. 242) we find ὃ µὴ τύχοι, without ποτέ. Regarding the second research question, it has been pointed out that the internal configuration of the relative clauses imposes restrictions on the type of antecedent. On the one hand, the fact that the verb of the relative clauses is semantically eventive forces its first argument (i.e. the relative pronoun) to designate a state of affairs. On the other hand, the fact that the verb of the relative clauses is a cupitive optative and thus expresses a possible desire forces the state of affairs referred to by the relative pronoun to be an unrealised state of affairs. The first restriction explains why the antecedent of ὃ µὴ γένοιτο (vel sim.) is usually a clause (24 of 26 examples), while the second explains why the clause to which the relative refers is usually a conditional clause (20 of 26). Finally, this research has also shed light on the distribution of these relative clauses across different authors. ὃ µὴ γένοιτο, with or without expansions, is the most commonly used relative clause: it is found in Aeschylus, Herodotus, Euripides, Aristoph- anes, Lysias, Plato, Demosthenes, Dinarchus, and Menander. By contrast, ὃ µὴ τύχοι (ποτέ) occurs only in Euripides, who, in fact, prefers it to ὃ µὴ γένοιτο. In turn, ὃ µὴ συµβαίη is documented only in Demosthenes.64 April, 2025 Universidad Autónoma de Madrid C/Francisco Tomás y Valiente 1 Madrid 28049 maria.lopezromero@uam.es 64 I would like to thank the anonymous informants and Emilia Ruiz Yamuza for their corrections and suggestions. I also extend my gratitude to Olivia Cockburn for reviewing my English, and to Patricia Varona Codeso for her assistance with Post-Classical Greek.