Rance ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 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/ Some Micro-Fragments of Menander Protector’s Histories: The Evidence of the Lexicon Αἱμωδεῖν Philip Rance ETWEEN the mid-ninth century and the late tenth, an un- identified lexicographer compiled a specialized lexicon designed to assist readers of four classicizing histories of the sixth century: Procopius’ Wars and successive continuations by Agathias, Menander Protector, and Theophylact Simocatta. He assembled selected instances of words or verbal forms that, by implication, were likely to be unfamiliar to a Middle Byzan- tine readership and explained their meaning through glosses, in some cases illustrated by quotation of the relevant passage(s), along with notes on etymology and usage. Transmitted without a title or ascription in copies known to later Byzantine lexicogra- phers and nineteenth-century scholars, this work is conven- tionally termed the Lexicon Αἱμωδεῖν, after its initial lemma: αἱµωδεῖν (“to have one’s teeth set on edge”).1 The compiler’s 1 Critical edition: A. R. Dyck, Epimerismi Homerici II (SGLG V.2, Berlin 1995) 825–1016, 1033–1034 (“Addenda et corrigenda”), hereafter cited as Dyck with page and line numbers. Of the three manuscripts, D and O con- tain no title. In G, the most recently identified, an excerpt bears the heading [Ἐτ]υµολογία[ι] Διάφοροι, which Dyck 846, 873, deems authentic. The con- ventional designation is not wholly modern: see citations εἰς τὸ Αἱµωδεῖν in Etym.Magn. 780.35–36, 789.11, 814.22 (εἰς τὸ Αἱµ·) Gaisford. Dyck 867 re- edits these testimonia. The chronological termini ca. 850–994 depend on multistage, mutual interaction between the Lexicon Αἱμωδεῖν and successive recensions of the Etymologicum Genuinum: Dyck 859–864 (with bibliography); K. Alpers, “Difficult Problems in the Transmission and Interrelation of the B PHILIP RANCE 655 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 decision to compose a glossary of uncommon vocabulary found in these four histories implies a perception of their intended interconnectedness in a literary-historiographic continuum, broadly in line with their categorization as a distinct quartet in modern scholarship. In compiling a new reading aid, the lexi- cographer presumably sought to meet perceived needs of certain users and to cater to concerns for language and style beyond historical content. This enterprise, in turn, points to renewed or enduring interest in these authors, and perhaps their educational and literary-cultural cachet, if not necessarily as curricular texts, then as compositional models. Such collective treatment seems uncharacteristic of their prior reception, insofar as it can be traced from the seventh to early tenth centuries, while the manu- script traditions of the four works offer no grounds for assuming conjoined transmission.2 In contrast, the generic affiliation and quasi-canonical status of the four historians in palace-sponsored compilatory projects around the mid-tenth century may imply a connection between the Lexicon Αἱμωδεῖν and Constantine VII’s historical Excerpta and/or related court-centred literary histori- ography.3 In any case, the lexicographer’s inferred objectives are consistent with the heightened appeal of late antique histories, especially Procopius and Agathias, during this period.4 Greek Etymologica,” in G. A. Xenis (ed.), Literature, Scholarship, Philosophy, and History: Classical Studies in Memory of Ioannis Taifacos (Stuttgart 2015) 293–314, at 309–310. 2 See, with bibliography, P. Rance, “In the Margins of Strabo: Menander Protector on Persian Religion and the Passio S. Isbozetae,” JÖB 73 (2023) 125– 150, at 127–131. 3 In the absence of comprehensive analysis, see provisional remarks in A. Németh, The Excerpta Constantiniana and the Byzantine Appropriation of the Past (Cambridge 2018) 18–19, 245–247; P. Rance, JÖB 73 (2023) 129; and n.39 below. 4 See recently Németh, Excerpta 158–161; M. Jankowiak, “Procopius of Caesarea and his Byzantine Successors,” in M. Meier et al. (eds.), A Companion to Procopius of Caesarea (Leiden 2021) 231–251, at 247–248; S. Kennedy, “A 656 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 Although the content, character, and potential value of the Lexicon Αἱμωδεῖν were indicated as long ago as 1910, it has been almost entirely overlooked in studies of the text, transmission, and reception of the four authors, and most surprisingly of Menander’s Histories, whose fragmentary survival, primarily via thematic excerpt collections, should pique interest in even the smallest scrap of evidence. To some extent, this oversight is understandable. Although a critical edition published three decades ago resolved many complexities, the sources, evolution, and textual tradition of the Lexicon Αἱμωδεῖν are especially intri- cate issues even within this specialist field. Furthermore, with a view to managing expectations, one should clarify that most of the material in the Lexicon Αἱμωδεῖν assignable to Menander comprises single words, i.e. the lemma or headword of a lexical entry, usually cited in the exact form in which that word was encountered in historian’s text. Thirty-five such lemmata will be examined below. In addition, five entries adduce six short quota- tions; one of these partly coincides with an anonymously quoted passage in the Suda. Overall, this Menander-derived content, mostly lacking historical context, may offer little of interest to historians, who have formed a majority of Menander’s modern readers. Never- theless, increasing attentiveness to the literary dimensions of late antique historiography, including linguistic, stylistic, and inter- textual aspects, affords a less restricted forum in which these fragments can be discussed and acquire significance. In par- ticular, Menander’s long-recognized imitation of the style and vocabulary of Agathias,5 whose Histories Menander’s work con- Lost Classic: The Reception of Prokopios’ History of the Wars in Byzantium,” Byzantinoslavica 79 (2021) 5–40, at 28–33; Rance, JÖB 73 (2023) 129–130; A. Markopoulos, “Revisiting the Sources of Leo the Deacon. The Case of the Excerpta Constantiniana,” ZRVI 60 (2023) 61–73. 5 M. Apostolopoulos, Μένανδρος Προτέκτωρ μιμητὴς Ἀγαθίου (Athens 1894), with B. Baldwin, “Menander Protector,” DOP 32 (1978) 101–125, at 119–120; R. C. Blockley, The History of Menander the Guardsman. Introductory PHILIP RANCE 657 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 tinues, reinforces some identifications of ‘Menandrian’ lemmata and provides an existing strand of literary inquiry that the present study can extend and augment. More generally, despite obvious limitations, selected lemmata and quotations might prove instructive to historical studies regarding the scope and content of Menander’s Histories. Accordingly, the following contribution seeks to present—or re-introduce—these ‘micro- fragments’ or lexical ‘splinters’, to elucidate their nature and transmission, and to tease out what might reasonably be inferred from them about Menander and his work. 1. The Lexicon Αἱμωδεῖν and the study of late antique historiography Although a relatively short work, the textual history of the Lexicon Αἱμωδεῖν is complex. The direct tradition is imperfectly witnessed by three manuscripts containing: an abridged copy of the whole lexicon (D), an excerpt of the same abridged tradition (O), and an excerpt of an unabridged version (G).6 Subsequent compilers of etymological dictionaries, principally the late tenth- /eleventh-century Etymologicum Gudianum and the mid-twelfth- century Etymologicum Magnum, independently utilized the Lexicon Αἱμωδεῖν as a source. As they typically had access to a fuller ver- sion of the text than is preserved in direct witnesses (DO), this more complete indirect tradition provides firmer foundations for a critical edition.7 Following the editio princeps by Friedrich Wilhelm Sturz in Essay, Text, Translation, and Historiographical Notes (Liverpool 1985) 6–9, 287 n.333. 6 D = Darmstad. 2773, 92v–101v; O = Oxon.Barocc. 119, 142v–143v (ε 9–ζ 2); G = Crypt. Za IX, 60r–64v (to α 111). See Dyck 846–848, 850–851, 864–865, and updated description of D in Commentaria in Aristotelem Graeca et Byzantina, https://cagb-digital.de/id/cagb4977917. In addition, MS. B of the Etymo- logicum Genuinum (Laur.S.Marci 304, dated 944) is treated codicis instar with respect to two dozen glosses interpolated from the Lexicon Αἱμωδεῖν: Dyck 859–860, 868; Alpers, in Literature 309–310. 7 Dyck 848–851, 859–865, 996–997. 658 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 1818,8 lexicographical studies clarified aspects of textual trans- mission, but its character and purpose remained poorly under- stood. For the present investigation, two contributions shaped the agenda of future inquiry. First, in 1910 Eduardo Luigi De Stefani distinguished the core component of the received text as a glossary of uncommon words in four classicizing histories written in the period from Justinian to Heraclius. Recognizing the implications, he ob- served that, although all lemmata and quotations are cited anon- ymously, as the works of Procopius, Agathias, and Theophylact are fully extant while Menander’s alone is fragmentary, it follows that those lemmata and quotations not found in the other three authors or in surviving fragments of Menander, and which can- not otherwise be explained, must be assigned, by default, to lost sections of Menander’s Histories. In short, “tutto ciò che non ap- partiene a quei primi tre, è di Menandro.”9 De Stefani’s wider objective was a critical edition of the Etymologicum Gudianum, a still uncompleted project that entailed editing lemmata derived from the Lexicon Αἱμωδεῖν, including those originating in ‘lost’ Menander.10 With respect to extant 8 F. W. Sturz, Etymologicum graecae linguae Gudianum et alia grammaticorum scripta e codicibus manuscriptis nunc primum edita (Leipzig 1818) 617.30–631.2 (based on D). Dyck 845–846 reviews nineteenth-century scholarship. 9 E. L. De Stefani, “Per le fonti del lexicon Αἱµωδεῖν,” StIt 18 (1910) 433– 444, quoting 435. 10 E. A. [L.] De Stefani, Etymologicum Gudianum (Leipzig 1909–1920) I (α– βωµολοχία), II (βωµολόχοι–ζειαί). Regarding lemmata attributable to lost sections of Menander (distinct from those identifiable in extant excerpts), De Stefani published 13 lemmata that the compiler of the Etym.Gud. derived from the Lexicon Αἱμωδεῖν: Etym.Gud. ἀµάρας I 107.2–5 (= Lex.Αἱμ. α 145 Dyck); ἀµβλυώττειν 110.1–3 (= α 144); ἀνεπαφρόδιτος 141.2–3 (= α 141); ἀφαυρά 239.7–10 (= α 146); γριφῶδες II 323.6–9 (= γ 13); γρυπός 324.1–4 (= γ 15); διακηρυκεύεται 356.18–22 (= δ 7); διασεσοβηµένος 360.8–10 (= δ 10); ἐλεη- λάτει 452.8–453.2 (= ε 37); ἐλινύειν 457.4–7 (= ε 41); ἔναυλος 468.4–7 (= ε 39–40); ἐπιέζοντο 503.19 (= ε 35); ἐσκευωρῆσθαι 537.1–3 (= ε 33). See Dyck 999–1002 for a full concordance. De Stefani alters his editorial format at II PHILIP RANCE 659 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 sections of Menander’s work, De Stefani identified one instance where this indirect tradition of the Lexicon Αἱμωδεῖν transmits a reading superior to that preserved in manuscripts of the Excerpta and the Suda.11 Over subsequent decades, historiographical scholarship seldom appreciated the logic and consequences of De Stefani’s insight, partly owing to the obscurity of the Lexicon Αἱμωδεῖν and the absence of a critical text. Second, more than eighty years elapsed before the publication of Andrew Dyck’s superb edition of the Lexicon Αἱμωδεῖν in 1995. His collation of all direct and indirect witnesses, including an unexploited manuscript (G), and comprehensive investigation of transmission and source-relationships, allowed the text to be constituted on a more secure basis.12 In reassessing the testimony of the Lexicon Αἱμωδεῖν as an indirect witness to the four histor- ians, Dyck affirmed De Stefani’s deduction regarding residual fragments of Menander.13 Furthermore, with greater under- standing of the sources and evolution of the transmitted text, Dyck was able to attribute additional lemmata and quotations to Menander and, conversely, to exclude others that De Stefani had assigned to Menander but which certainly or probably be- long to one of the other three historians. In doing so, Dyck may have disregarded analytical criteria applied by De Stefani in light of the lexicographer’s compositional method, while several lemmata discounted without argument merit re-evaluation.14 452.8–453.2 ἐλεηλάτει, in that only from this lemma is Menander cited as the ultimate source, though, as he argued in 1910, this is the case also with preceding lemmata listed here. 11 Etym.Gud. ἐµβριµώµενοι (masc. nom. pres. part. pl. ἐµβριµάοµαι) II 461.5. Although it is not found in the direct tradition of the Lexicon Αἱμωδεῖν (DO), De Stefani derives this lemma, via a more complete indirect tradition, from Menander fr. 10,3.16–17. Manuscripts of Exc.leg. I 193.6 and Suda ε 963 transmit ἐµβριµούµενοι, thus all editions of Menander. See Dyck 996. 12 See n.1 above, with Dyck 864–867 for editorial principles. 13 Dyck 851, 857–858, 998. 14 Dyck 998 (App. 4), “Possible New Fragments of Menander Protector,” 660 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 The Lexicon Αἱμωδεῖν has gone largely unnoticed by scholar- ship on Procopius, Agathias, Menander, and Theophylact. Studies of their reception, even where alert to Middle and Late Byzantine lexica, do not mention this author-specific compila- tion.15 With respect to constitutio textus, only Rudolf Keydell’s meticulous edition of Agathias’ Histories integrates the Lexicon Αἱμωδεῖν into the critical apparatus. In at least two instances, the lexicon uniquely transmits a correct reading.16 When Jacob Haury edited Procopius’ Wars (1905), he overlooked this indirect tradition. In a detailed review, De Stefani signalled the value of the Lexicon Αἱμωδεῖν and derivative lexica for confirming or cor- recting readings in the direct tradition.17 Following De Stefani’s observations, and his further study of 1910, Gerhard Wirth incorporated these lexica as variae lectiones in Addenda et Corri- where emendations are necessary: correct ἐλινύειν ε 37 to 41, θρυλλοῦντες θ 2 to 3, θριγκῷ θ 3 to 4, θεοκλυτοῦσιν θ 5 to 6, ἰαχνόν ι 8 to 9. For Dyck’s supplements to De Stefani’s catalogue see 677 below. 15 Surveys of Procopius-derived material in other lexica: Jankowiak, in A Companion to Procopius 247–250; Kennedy, Byzantinoslavica 79 (2021) 16. In ad- dition, Procopius and Agathias are cited alongside canonical Attic and later Atticizing authors in Περὶ συντάξεως (D. Petrova, Das Lexikon “Über die Syntax”: Untersuchung und kritische Ausgabe des Lexikons im Codex Paris. Coisl. gr. 345 [Wies- baden 2006]) δ 14 δωροῦµαι (Procop. 1.10.17, 3.8.9, 3.9.5, 8.35.33; Anec. 4.27); σ 29 συρράξαι (Agath. 1.21.7) [= Anecd.Bekk. I 133.14–15, 174.3]. 16 R. Keydell, Agathiae Myrinaei historiarum libri quinque (CFHB 2: Berlin 1967) xx–xxi, xxxiii–xxxiv, 2 (siglum: Etym). Keydell cited quotations from Agathias in the Lexicon Αἱμωδεῖν via derivative entries in De Stefani’s edition of the Etymologicum Gudianum (n.10 above) up to Z, and thereafter from Sturz’s editio princeps (1818). In two cases, Keydell (xxi) preferred a reading in the Lexicon Αἱμωδεῖν: δ 3 (Dyck 924.9) quotes δυηπαθείας in Agath. 3.1.4, where the manuscripts read δυσπαθείας (Keydell 84.20 app. crit.); ε 10 (Dyck 931.36) quotes ἐνερόχρωτες in Agath. 2.23.6, where the manuscripts read ἐτερόχρωτες (Keydell 71.3 app. crit.). Reporting only the first of these two cases, Dyck 858 considers another possibility at φ 4 ἐµβριθέστατον (986.15 app. crit., 987.21) in a quotation of Agath. 4.22.6, where Keydell (151.1 app. crit.) favoured ἐµβριθέστερον in the direct tradition. 17 E. L. De Stefani, “Besprechungen: J. Haury (ed.), Procopii Caesariensis Opera omnia I–II (Leipzig 1905),” ByzZeit 14 (1905) 636–643, at 639–640. PHILIP RANCE 661 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 genda appended to his revised edition of Haury’s text.18 Con- sidering that the Lexicon Αἱμωδεῖν predates the oldest surviving manuscripts of Wars by at least three centuries, it is unremark- able that De Stefani could identify two readings where the lexi- con arguably preserves a superior tradition, though Procopius’ broader lexical usage undermines one of these cases.19 In con- trast, the Lexicon Αἱμωδεῖν offers no improvements to the critical text of Theophylact’s Histories, despite the tenuity of its direct tradition.20 18 Procopii Caesariensis Opera omnia (Leipzig 1962–1964) I 552–572; II 679– 699 passim, citing Sturz’s edition of Lex.Αἱμ. (under “EGud”) at I 553, II 679. 19 De Stefani, ByzZeit 14 (1905) 639–640; Wirth in Haury/Wirth I 564 (ad 233.2), 570 (ad 453.2); Dyck 858. Manuscript tradition of Wars: Haury/ Wirth I xxii–liv. First, De Stefani observes that Lex.Αἱμ. ε 17 (= Etym.Gud. II 536.16–17; Etym.Magn. 385.15, 717.20) transmits ἐσκαλευκότες, where Procop. 2.19.10 (Haury I 233.2) reads ἐσκυλευκότες based on two fourteenth- century MSS. (VG). De Stefani sees ἐσκαλευκότες as “la lezione genuina.” Similarly, Dyck notes that ἐσκαλευκότες in Lex.Αἱμ. is “now the earliest attested reading, is lectio difficilior and yields good sense.” They do not ac- knowledge that σκυλεύω occurs in five other instances in Wars (1.14.26, 1.18.50, 2.11.10, 5.25.9, 7.24.26), including twice likewise as masc. acc. perf. part. act. plural ἐσκαλευκότες (1.18.50, 5.25.9). Second, a more persuasive case concerns Lex.Αἱμ. ε 21 (= Etym.Gud. II 447.6–7, Etym.Magn. 323.48), which quotes Procop. 4.8.7 as ἔκπυστα ἐποίησε τὰ πρασσόµενα, where direct witnesses diverge: Haury I 453.2 app. crit. prefers ἐποίει in V to ἐποίησε in PO. De Stefani favours ἐποίησε in the indirect tradition. Dyck suggests that the reading in Lex.Αἱμ. “tips the weight of evidence decisively in favor of the aorist,” which is preferable contextually. Support for this view is found in later instances of the same phrase at Procop. 4.25.19 and 7.11.27, where Haury I 534.25, II 344.19–20, reports the aorist without variants. 20 Although the Lexicon Αἱμωδεῖν includes many lemmata drawn from Theophylact, the received text preserves only four quotations: Lex.Αἱμ. α 150 ἀκκίζεται (Th. Sim. 4.4.14); κ 1 κατασπιλάζοντες (4.5.10); σ 21–23 στιβίζουσι (pref. 11); φ 6 φλήναφος (2.2.1). Dyck 857 n.57 notes that the sole divergence from the direct tradition is corrupt: φ 6: ὥσπερ for ὡς περ⟨ὶ⟩ (987.27 app. crit.). The transmission of Theophylact’s Histories becomes more tenuous than Dyck supposes, if the codex unicus, tenth-century Vat.gr. 977 (not, as Dyck, “twelfth century”), is the exemplar used by the Constantinian excerptors: thus P. 662 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 Most significantly, no editor or researcher of Menander’s frag- mentary Histories has yet consulted the Lexicon Αἱμωδεῖν.21 Three older editions long antedate De Stefani’s revelatory research.22 His 1910 study makes a single appearance in Barry Baldwin’s survey of possibly Menander-derived anonymous quotations in the Suda: noting—and dismissing—De Stefani’s attribution of a three-word phrase in Suda φ 715 (= Lex.Αἱμ. φ 5) to Menander, Baldwin does not refer to the Lexicon Αἱμωδεῖν or explain De Stefani’s rationale, leaving readers to assume pure conjecture. Baldwin thus misjudges one of the strongest cases for assigning an anonymous quotation in the Suda to Menander.23 The most recent presentation of Menander’s fragments by Roger Blockley, essentially a composite of prior editions incorporating published variae lectiones, neither mentions the Lexicon Αἱμωδεῖν nor cites De Stefani.24 2. The Lexicon Αἱμωδεῖν: textual history and compositional principles A preliminary review of the Lexicon Αἱμωδεῖν will clarify its format, sources, and purpose.25 A glossary of the four classicizing Schreiner, “Die Historikerhandschrift Vaticanus Graecus 977: Ein Hand- exemplar zur Vorbereitung des konstantinischen Exzerptenwerkes?” JÖB 37 (1987) 1–29; with Németh, Excerpta 59, 102. 21 See briefly Rance, JÖB 73 (2023) 129, 139 n.62. A new edition by Bruno Bleckmann and Markus Stein will address this neglect: B. Bleckmann, “Zu den Fragmenten des Menandros Protektor im Projekt KFHist,” in E. Amato, et al. (eds.), Les historiens fragmentaires de langue grecque à l’époque romaine impériale et tardive (Rennes 2021) 61–69, at 66–67. 22 I. Bekker and B. G. Niebuhr, Dexippi, Eunapii, Petri Patricii, Prisci, Malchi, Menandri historiarum quae supersunt (CSHB 14: Bonn 1829) 279–444; C. Müller and T. Müller, FHG IV (Paris 1851) 200–269; L. Dindorf, Historici graeci minores II (Leipzig 1871) iii–x, 3–131. 23 Baldwin, DOP 32 (1978) 125: “There is obviously no real clue to be had from such a snippet.” Dyck 845 n.3 notes Baldwin’s dismissal “without coming to terms with De Stefani’s argument.” See 693 below. 24 Blockley, The History of Menander, with editorial remarks at [vii]. Blockley’s numbering of fragments and lines is used in what follows. 25 Dyck 845–867 for detailed argumentation and bibliography. PHILIP RANCE 663 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 histories is the main and earliest component, which became a nucleus for subsequent expansions. Each typically one-word lemma is ordinarily cited in the form (case/tense, number, gender) in which it appears in the historian’s text. One or more glosses and other explanatory material follow, which the compiler drew from mostly identifiable lexical and exegetical sources.26 Some entries quote the passage in which the lemma occurs. It is assumed that this was originally a more prevalent or universal feature, but many quotations have been lost in all textual traditions, a common trait in the transmission of Byzan- tine lexica.27 For most lemmata, Dyck’s edition supplies an in- text reference to the relevant passage(s), though several additions and adjustments are warranted.28 Sometimes a lemma or quota- 26 Dyck 1007 for an amalgamated Index fontium. 27 S. Valente, “Byzantine Greek,” in J. Considine (ed.), The Cambridge World History of Lexicography (Cambridge 2019) 247–266, at 247–248. 28 For some lemmata Dyck gives no reference to the cited passage. Oc- casionally, a lemma occurs in two or more of the historians, while its form and sequence do not distinguish a particular locus: e.g. γ 10 Γήπαιδες; γ 11a γλιχόµενοι / 11b γλιχόµενος. Elsewhere, the source can be more confidently identified. Lemma µ 4 µῶν, without reference in Dyck 954.13, is cited from Agath. 2.10.5. Alternative Th. Sim. pref. 8 would be detached from other Theophylact-derived lemmata at µ 9–µ 13. Accordingly, for the sequence of Agathias-derived lemmata to be correct, following µ 5 µειονεκτεῖσθαι is cited from Agath. 3.9.3 rather than pref. 4; Dyck 954.17 adduces both passages. Lemma ν 3 νηποινεί, unreferenced in Dyck 958.14, is cited from Th. Sim. 1.15.10, where the direct tradition transmits the form νηποινί (69.18 de Boor). This is in sequence with following ν 5 νηπενθές from Th. Sim. 2.11.7. For ο 10 ὀρεῦσιν, Dyck 963.33 does not adduce Agath. 4.11.3, see 669 below. Lemma σ 15a στέργων, unreferenced in Dyck 954.13, must be cited from Agath. 4.3.3. For σ 16 σεµνῶς, Dyck 976.61 does not adduce Agath. 4.9.9 and 4.22.2, see 669 below. For τ 10 τενάγη, Dyck 982.39–41 does not adduce Th. Sim. 6.8.10. Lemma υ 1 ὑπόγυον De Stefani, StIt 18 (1910) 440, and Dyck 983.2 refer to Procop. 8.29.6 ὑπόγυον, though likewise acc. at 8.22.16. For υ 2 ὑπετόπαζεν De Stefani 440 and Dyck 983.6 adduce Agath. 2.4.9, but seem- ingly muddled sequencing of υ 1–υ 4 allows the possibility of citation from Procop. 1.25.19 (cf. 2.5.15, 3.5.1, 7.4.26). Occasional supplementary lem- 664 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 tion differs slightly from the direct tradition. Certain divergences are explicable as transcriptional errors in copying the Lexicon Αἱμωδεῖν. In other cases, the lexicographer purposely modified lemmata, often where he preferred the form of a headword found in those earlier lexica that he consulted for definitions and synonyms, and thus a lemmatic form in the Lexicon Αἱμωδεῖν in- stead reflects that lexicographical tradition.29 After the compilation of the glossary, the lexicographer or a posterior editor appended to some entries guidance on syntac- tical usage, apparently to enhance understanding of this ancient idiom.30 Later still, in a multistage process, new entries were interpolated from diverse lexica, scholia, and epimerismi, often via intermediary compilations, with the effect of transforming a specialized wordlist into a general-purpose lexicon. For reasons that are unclear, this development was restricted to letter-section alpha, which is disproportionately swollen to comprise one-third of the current text. These successive augmentations were largely completed before the Lexicon Αἱμωδεῖν became a source for eleventh- and twelfth-century lexicographers.31 In letter-sections other than alpha, no interpolation of additional entries has been demonstrated. As most lemmata, in form and sequence, can be cross-checked with the texts of Procopius, Agathias, Theophy- mata, cited as an additional example of a headword, pose separate inter- pretative challenges. Lemma ε 40 ἐναυλιζόµενον, despite the implication of Dyck 942.93 app. font., lacks a lexicographical antecedent and must be cited from Agath. 3.16.2, the unique instance in historiography. Lemma λ 8 λαιὰ χείρ, which neither De Stefani, StIt 18 (1910) 440, nor Dyck 952.20 recog- nized as a citation from one the historians, has multiple instances in Procop. 3.23.18 (gen.), 4.13.15 (dat.), 4.28.11 (acc.), 7.24.24 (acc.), 8.35.27 (dat.); cf. Aed. 1.2.11. In the absence of a lexicographical source (Dyck 983.6 app. font.), lemma υ 3 ὑποτοπάζειν could be cited from Procop. 5.6.8, 7.26.17. 29 Such editorial interventions, where relevant, are noted below. 30 Dyck 851 identifies the source of these interpolations as “a syntactic lexicon”; see 1007, Index fontium, s.v. “Lex. synt.,” and marginal annotation “synt.” throughout the edition. 31 See analysis in Dyck 851–855, 861–862, with Index fontium at 1007. PHILIP RANCE 665 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 lact, or extant Menander, questions of provenance, in theory, affect only lemmata attributable by default to lost sections of Menander, the material of interest here. The following analysis found no grounds for excluding any such lemma. Even in a handful of cases where position, form, and/or content might be puzzling, the received text of the lexicon provides parallels where authenticity is undoubted, while the cited headword variously coheres with Menander’s known lexical tastes, stylistic practice, and/or literary antecedents.32 Beyond its arrangement in letter-sections, the Lexicon Αἱμωδεῖν has no alphabetic dimension. Under each letter, a basic organi- zational principle groups lemmata by author, often in the chron- ological sequence Procopius-Agathias-Menander-Theophylact, though sometimes Agathias is first or the order otherwise varies. As a rule, excepting occasional and minor displacements, lem- mata cited from each author are arranged sequentially accord- ing to their occurrence in his work. In only a small number of isolated cases is a lemma demonstrably belonging to one author misplaced among lemmata cited from another.33 Lemmata in epsilon, one of the longer letter-sections, illustrate this scheme and some of its interpretative challenges: ε 1[–2] ἐµπεδῶσαι (Agath. 1.1.6) ε 3 ἐνέρσει (Agath. 1.3.4) 32 I considered and rejected the possibility of interpolation at ζ 1 ζαµενής, ζ 3 ζάκορος, µ 14 Μαµερτῖνοι, π 26 πέµπελος, σ 1 συνασµένιζε, τ 9 τελχῖνες, ω 4 ὠρύω. An interpolator’s rationale for inserting some or all of these items would be no less (and arguably more) puzzling. 33 Letter-section alpha, uniquely containing many interpolations, has suffered exceptional disarrangement and should be treated as atypical. From beta to omega, a handful of such displaced individual lemmata occur, often at the beginning or end of letter-sections: δ 9 δαιταλεῦσιν (if correctly identi- fied as Th. Sim. 5.5.9 and not lost Menander: see 675 below); µ 14 Μαµερτῖνοι (685 below); ξ 4 ξυνίστωρ (686 below); σ 1 συνασµένιζε (687 below); σ 21 στιβίζουσι; τ 10 τενάγη. Only towards the end of letter pi is more extensive disruption found: π 1–11 Agathias, 13–18 Procopius, 19–20 Agathias, 21–23 Menander, 24 Theophylact, 25 Agathias, 26 Menander (see 669 and 687). 666 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 ε 4 ἐγχρίπτουσα (Agath. 1.12.6) ε 5 ἐπεκερτόµει (Agath. 1.20.10) ε 6 εἰσήρρησεν (Agath. pref. 28) ε 7 ἐς νέωτα τὸν πόλεµον µεταθέσθαι (Agath. 1.19.3) ε 8 ἐπίσ⟨σ⟩ωτρα (Agath. 2.4.5) ε 9 ἐρυµνώσας (Agath. 2.4.6) ε 10 ἐνερόχρωτες (Agath. 2.23.6) ε 11 ἑωλοκρασία (Agath. 2.29.8) ε 12 ἔστεργεν (Agath. 2.31.2 or 5.12.6) ε 13 ἐπέσκηψεν (Procop. 1.2.7) ε 14 ἐτώθαζον (Procop. 1.7.17 or 2.8.6) ε 15 ἐπιθειάσας (Procop. 2.11.1) ε 16 εἴσῃ (Procop. 1.14.8 or 2.3.38) ε 17 ἐσκαλευκότες (Procop. 2.19.10) ε 18 ἐκεχειρία (Procop. 2.28.7 et al.) ε 19 ἐρεσχηλῶ (Procop. 3.12.8) ⟨ἐθελοκακήσας (Procop. 3.6.10)⟩34 ε 20 ἐνεόχµωσεν (Procop. 3.7.22) ε 21 ἔκπυστα (Procop. 4.8.7) ε 22 ἐπιθρυλοῦντες (Procop. 4.21.2) ε 23 ἔρανος (Procop. 5.8.30) ε 24 ἐγκάρσιον (Procop. 5.21.20–21) ε 25 εἰρωνεύεσθαι (Procop. 5.22.3) ε 26 εὐήθεια (Procop. 5.22.9) ε 27 ἔπαλξις (Procop. 5.22.10 et al.) ε 28 ἐφεῖντο (Procop. 6.20.32) ε 29 ἐναβρύνεσθαι (Procop. 7.24.25) ε 30 ἐπεβάτευσεν (Procop. 7.34.21) ε 31 ἑάλω (Procop. 7.35.2) ε 32 ἔξωρος (Procop. 8.11.40) ε 33 ἐσκευωρῆσθαι (lost Men.) ε 34 ἐπιτροχάδην (cf. Men. fr. 10,3.84) ε 35 ἐπιέζοντο (lost Men.) ε 36 ἐφειµένον (cf. Men. fr. 8.10) ε 37 ἐλεηλάτει (lost Men., cf. fr. 21.13 λεηλατεῖν) ε 38 ἐκµαγεῖον (cf. Men. fr. 6,1.413) 34 A lacuna at Dyck 936.2–3 app. crit. (cf. Etym.Gud. ἐθελοκακήσας [II 404.1]) is recognized by C. Theodoridis (review of Dyck), JÖB 50 (2000) 378– 385, at 385 (citing Procop. 4.20.6; cf. 3.6.10, 4.27.29). PHILIP RANCE 667 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 ε 39[–40] ἔναυλος (lost Men.) 35 ε 41 ἐλινύειν (lost Men.) ε 42 εὐτραπελία (Th. Sim. 1.10.11) ε 43 ἐνείρω (Th. Sim. 1.11.8) ε 44 ἐσύκαζεν (Th. Sim. 3.7.8) Within this sample, a few lemmata are out of sequence (ε 6, 7, 16, 19–20), mostly, it seems, due to accidental transposition. Several diverge from the direct tradition, owing to editorial intervention or transmissional error: ε 11 ἑωλοκρασία < Agath. 2.29.8 ἑωλοκρασίᾳ; ε 17 ἐσκαλευκότες < Procop. 2.19.10 ἐσκυ- λευκότες;36 ε 18 ἐκεχειρία < Procop. 2.28.7 et al. ἐκεχειρίαν; ε 19 ἐρεσχηλῶ < Procop. 3.12.8 ἐρεσχελοῦντα;37 ε 23 ἔρανος < Procop. 5.8.30 ἔρανον; ε 24 ἐγκάρσιον < Procop. 5.21.20–21 ἐγκάρσια (neut. pl.); ε 26 εὐήθεια38 < Procop. 5.22.9 εὐήθειαν; ε 27 ἔπαλξις < Procop. 5.22.10 et al. ἐπάλξεις; ε 30 ἐπεβάτευσεν < Procop. 7.34.21 ἐπεβάτευσαν; ε 42 εὐτραπελία < Th. Sim. 1.10.11 εὐτρα- πελίας. This author-citational scheme presumably reflects the lexicog- rapher’s methodology of working through the four histories and recording, under the relevant letter, any word he encountered 35 See n.28 above for supplementary lemma ε 40. 36 See n.19 above. 37 The entry ε 19 ἐρεσχηλῶ : τὸ φλυαρῶ. | ⟨…⟩ is irregular in multiple respects and, besides a terminal lacuna (n.34 above), transmissional difficul- ties can be suspected. Found in the direct tradition (DO) but not in derivative lexica, ε 19 is evidently abridged, perhaps severely (cf. ε 12b, 13b, 17b). In no other lemma cited from an extant passage of the four historians (Procop. 3.12.8 ἐρεσχελοῦντα, var. lect. ἐρεσχηλοῦντα) is a verb changed to the first- person singular (692 below). If a deliberate alteration, and not textual cor- ruption, the motive for this intervention is obscure, not least because prior lexicographical traditions known to the compiler already contained a lemma in precisely the required form ἐρεσχηλοῦντα (Synagoge ε 822 Cunningham; Phot. Lex. ε 1907; cf. Suda ε 2936). 38 Accented εὐηθεία Dyck 938.35; correctly εὐήθεια Lex.Αἱμ. Sturz 624.38; Etym.Gud. II 555.4. 668 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 in need of explanation.39 Awareness of this compositional process can assist in recognizing the provenance of unattributed material: accordingly ε 33, 35, 37, 39, and 41, owing to their sequential placement and/or absence from extant texts, have been assigned to lost parts of Menander’s Histories. Occasional disruptions in the sequencing of citations may be due to au- thorial muddle, copyist’s error, or editorial intrusion (especially in alpha), but ostensible disarrangement might arise if two historians used the same word/form, as when Menander imi- tates Agathias’ wording (see below). If the design of the lexicon is comprehensible, the content of the received text seems un- balanced. Agathias is by far the favoured author. Even allowing for differing frequency of initial letters, some sections are very short, comprising a couple of lemmata. Some longer letter- sections contain no lemmata from one or more of the historians. 39 Németh, Excerpta 18–19, 245–247, proposes that the lexicographer did not consult the four histories directly but relied on (hypothetical) “draft copies” of several now-lost volumes of the Constantinian Excerpta. Németh does not acknowledge that prior scholarship derives the Excerpta and Lexicon Αἱμωδεῖν from separate textual traditions: Keydell ed. xxxiii–xxxiv, with Rance, JÖB 73 (2023) 129 n.21; and n.11 above. Moreover, the number, range, and sequence of lemmata do not favour Németh’s proposition. In the case of Agathias, if the lexicographer extracted lemmata from excerpts dispersed across multiple topic-based volumes of the Excerpta, he took the trouble to rearrange those lemmata so as to reconstitute their original sequence in the complete text of Agathias’ Histories. This does not seem a logical or likely scenario. See e.g. Agathias-derived lemmata in epsilon: ε 1 (Agath. 1.1.6: treaty obligations), ε 3 (1.3.4: Frankish and Avar hairstyles); ε 4 (1.12.6: poliorcetic stratagem); ε 5 (1.20.10: jeering at the enemy); ε 6 (pref. 28: Procopius’ account of the plague); ε 7 (1.19.3: dismissing troops into winter-quarters); ε 8 (2.4.5: encampment defences); ε 9 (2.4.6: ditto); ε 10 (2.23.6: Persian funerary customs); ε 11 (2.29.8: a philosopher’s gluttony); ε 12 (2.31.2: Roman philosophers exiled in Persia / 5.12.6: Roman-Utigur diplo- macy). More generally, although important issues remain unresolved, one might legitimately question the wisdom or utility of compiling an author- specific glossary of four histories after they have been excerpted, dismem- bered, and allocated to multiple volumes containing excerpts extracted from a dozen or more other histories. PHILIP RANCE 669 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 It is unclear to what extent these peculiarities reflect defective transmission, the compiler’s caprice, or undefined compositional factors. The task of isolating Menander-derived content requires careful assessment of anomalies and ambiguities. This process is unavoidably technical (readers more interested in results than method may prefer to proceed to section 3). First, several in- correct or dubious cases must be eliminated from the inquiry. Four lemmata that De Stefani attributed to Menander should— or can—be assigned to Agathias.40 Attribution of λ 9 λιγνυῶδες (neut. sg.) to Menander is unsound when this word occurs at Agathias 5.8.1 (though λιγνυώδη neut. pl.), as De Stefani ac- knowledged and Dyck (952.23) affirms, and this passage is cor- rectly placed in a series of Agathias-derived lemmata. An earlier instance at Agathias 2.15.9 (λιγνυώδεις fem. pl.) would disrupt the sequence. De Stefani assigned ο 10 ὀρεῦσιν to Menander, without noting its occurrence at Agathias 4.11.3, a reference Dyck (963.33) also omits. This would be slightly out of sequence with preceding ο 9 ὀθνεῖον, if that lemma cites Agathias 4.15.3, as De Stefani and Dyck (963.30–32) believe, though the form at 4.15.3 is dat. ὀθνείῳ (cf. 4.26.6), whereas ὀθνεῖον occurs at 3.14.3 (cf. 1.5.5). Lemma π 25 πρῶρα (nom. sg.), the penultimate in pi, which De Stefani assigned to Menander, is most plausibly ex- plained as a displaced but broadly in-sequence citation from Agathias 5.21.8 (at 192.5): πρώρας (gen. sg.). This view is circum- stantially supported by τ 4 τολύπαις, a lemma cited from the same account of improvised boats at 5.21.7 (at 191.23).41 Finally, De Stefani’s tentative attribution of σ 16 σεµνῶς to Menander overlooked two instances at Agathias 4.9.9 and 4.22.2, not refer- 40 De Stefani, StIt 18 (1910) 440, 442. 41 Within letter-section pi, Agathias-derived lemmata are sequential but not contiguous: π 1 (1.8.6)–π 11 (5.3.11), π 19 (5.14.7)–π 20 (5.22.1). Displaced π 25 (5.21.8) would be only slightly out of sequence. Instances of πρῶρα in Procop. 7.30.13 (dat. pl.) and 8.22.11 (acc. sg.) lack comparable supporting arguments. 670 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 enced by Dyck (976.61–64), even if σ 17 στῖφος, from Agathias 3.20.5 (cf. 3.28.6), would thus be out of sequence. Challenges arise in demarcating Menandrian material even when ostensibly cited from extant excerpts. In some cases, such lemmata follow an expected sequence, based on their relative positions in the Constantinian Excerpta, in which excerpts typi- cally retain their original order.42 In one case, however, the re- ceived text of the Lexicon Αἱμωδεῖν is more problematic than prior studies indicate: ε 34 ἐπιτροχάδην (Men. fr. 10,3.84 = Exc.leg. I 195.5) ε 36 ἐφειµένον (Men. fr. 8.10 = Exc.leg. II 445.7) ε 38 ἐκµαγεῖον (Men. fr. 6,1.413 = Exc.leg. I 182.34 ἐκµαγείοις) Although De Stefani and Dyck confidently located each lemma in an excerpt, these three citations would thus be in reverse order.43 The issue is not confined to Menander: several lemmata taken from Procopius and Agathias occur more than once in their works, even in the same form. While it is logically assumed that a headword is cited from its first instance, this is sometimes unclear or demonstrably not so.44 For lemmata cited from Menander’s Histories, fragmentary preservation exacerbates uncertainty about their original positioning. In some cases, the form of a lemma confirms beyond doubt that it does not derive from the surviving textual sample: θ 6 θεοκλυτοῦσιν (pres. ind. act. 3rd pl.) cannot be cited from θεοκλυτεῖν (pres. inf. act.) in fr. 6,1.403 (Exc.leg. I 182.25), but must cite another instance of this verb in a lost section.45 Likewise ε 37 ἐλεηλάτει (imperf. ind. act. 3rd sg.) is not cited from λεηλατεῖν (pres. inf. act.) in fr. 21.13 (Exc.leg. I 208.22–23), nor π 22 παγιωθεῖσα (fem. nom. aor. part. 42 Correct sequence of lemmata cited from excerpts of Menander’s Histories: e.g. τ 7 τερθρεία (Men. fr. 6,1.45 = Exc.leg. I 172.23); τ 8 τρακταΐσαι (Men. fr. 6,1.193 = Exc.leg. I 176.26). See likewise α 142 ἀποχρώντως (Men. fr. 6,1.18 = Exc.leg. I 171.30); α 143 ἀποφλαυρίζοντες (Men. fr. 6,2.20 = Exc.sent. 20.4). 43 De Stefano, StIt 18 (1910) 441–442; Dyck 941.73, 941.80, 942.89. 44 E.g. θ 7 θωπεία with multiple references cited n.56 below. 45 Dyck 947.18; Rance, JÖB 73 (2023) 139 n.62. See 683 below. PHILIP RANCE 671 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 pass. sg.) from ἐπαγιώθη (aor. ind. pass. 3rd sg.) in fr. 6,1.365 (Exc. leg. I 181.22).46 Correspondingly, it is possible that ε 38 ἐκµαγεῖον (neut. nom./acc. sg.), distinct from ἐκµαγείοις (neut. dat. pl.) in its posited source at fr. 6,1.413, is in fact cited from a later, now- lost passage, which would explain the position of ε 38 relative to preceding lemmata. Other factors may weigh against this in- ference. In a handful of lemmata, the lexicographer modified both the case and number of a cited noun. He might have done so in ε 38.47 Moreover, as fr. 6,1 is the source of other lemmata, he is likely to have read ἐκµαγείοις here, even if he encountered ἐκµαγεῖον elsewhere.48 Similar ambiguity affects ε 34 ἐπιτροχάδην and ε 36 ἐφειµένον: the headwords find precise correspondents in excerpts, but could also have occurred in lost sections. Ulti- mately, while questions remain open, the ‘incorrect’ sequence of citations in epsilon may merely reflect the partial survival of Menander’s work. Evidence of disrupted transmission in this part of epsilon further complicates analysis. The Etymologicum Gudianum contains two lemmata that, though cited from Menander and Theophy- lact, are absent in the direct tradition of the Lexicon Αἱμωδεῖν (DO), and were evidently transmitted via a more complete ver- sion available to the tenth-/eleventh-century compiler.49 First, rare ἐµβριµώµενοι (Etym.Gud. II 461.5) is cited from Menander fr. 10,3.16–17 (Exc.leg. I 193.6), the excerpt in which the extant in- stance of ε 34 ἐπιτροχάδην occurs (fr. 10,3.84 = Exc.leg. I 195.5).50 Second, ἔµβιοι (Etym.Gud. II 461.1–2) derives from Theophylact 1.5.15, which is quoted in the gloss. This lemma, preceding ε 42 εὐτραπελία (Th. Sim. 1.10.11), originally would have begun the 46 Dyck 941.84–85, 970.82. 47 Cf. ε 42 εὐτραπελία < Th. Sim. 1.10.11 εὐτραπελίας (acc. pl.); σ 14 σισύρα < Agath. 3.16.7 σισύραις; τ 2 τάρπη < Procop. 1.18.52 τάρπας. 48 Cf. α 142 ἀποχρώντως (fr. 6,1.18 = Exc.leg. I 171.30); τ 7 τερθρεία (fr. 6,1.45 = Exc.leg. I 172.23); τ 8 τρακταΐσαι (fr. 6,1.193 = Exc.leg. I 176.26). 49 See 657 above, and specifically Dyck 996. 50 See n.11 above. 672 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 Theophylact series. For the present study, these various grounds for doubt and signs of disarrangement, combined with Menander’s known imi- tation of Agathias, have an important consequence: one cannot entirely rule out the possibility that at least some of the lemmata in epsilon that editors have assigned to Menander are in fact a displaced series of Agathias-derived lemmata, especially as they replicate the sequential occurrence of those headwords in Agath- ias’ Histories: ε 33 ἐσκευωρῆσθαι (3.3.7), ε 34 ἐπιτροχάδην (4.1.2), ε 35 ἐπιέζοντο (4.20.4). This could also be the case with ε 36 ἐφειµένον (1.20.5, 3.27.1, 4.5.8) and ε 37 ἐλεηλάτει (4.29.8), though considerations of sequence, together with Menander’s proven usage of both words (fr. 8.10 ἐφειµένον, fr. 21.13 λεηλα- τεῖν), more firmly anchor these two lemmata in a Menander series.51 In the case of ε 33–35, however, as no arguments can tip the balance in favour of Menander or Agathias, these lem- mata will be excluded from the sample analysed below. Other disruptions of sequence may reflect vagaries of com- position or transmission, but again linguistic affinities between Agathias and Menander lead to uncertainty. Letter-section theta serves as an example: θ 1[–2] θαµίζειν (Agath. pref. 10 or Men. fr. 1,1.9)52 θ 3 θρυλλοῦντες (lost Men.) θ 4 θριγκῷ (Agath. 4.14.3 or lost Men.) θ 5 θητικόν (Agath. 3.16.4, 4.21.6, cf. 2.8.2 θητικοῦ; or lost Men., cf. fr. 5,4.25 θητικοῦ) θ 6 θεοκλυτοῦσιν (lost Men., cf. fr. 6,1.403 θεοκλυτεῖν) θ 7 θωπεία (Procop. 1.25.4 et al.; Agath. 4.28.4; Men. fr. 6,2.17) Indicative of conscious mimesis, θ 1 θαµίζειν occurs in the initial 51 If the lexicographer cited ε 36 ἐφειµένον from Agathias, he overlooked two prior instances (1.20.5, 3.27.1) and cited a third (4.5.8), which is still out of sequence with preceding ε 35 ἐπιέζοντο (4.20.4). If ε 37 ἐλεηλάτει is cited from Agathias (4.29.8), the lexicographer likewise passed over a prior instance (4.22.6 λεηλατουµένων). 52 θ 2 καὶ θαµίζει, a supplementary lemma to θ 1 θαµίζειν, derives from a lexicographical source: Dyck 945.3 app. font. (add Hsch. θ 81). PHILIP RANCE 673 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 sections of both Agathias and Menander. Lemma θ 3 θρυλλοῦν- τες, unattested in the other three historians, by default originates in a lost part of Menander, likewise θ 6 θεοκλυτοῦσιν, especially as an excerpt verifies Menander’s usage. Given this sequence, intervening θ 4 θριγκῷ and θ 5 θητικόν should also derive from Menander. If so, as θριγκῷ occurs in Agathias 4.14.3, the lemma is cited from another, now-lost passage of Menander that imi- tated Agathias.53 The source of θ 5 is more ambiguous. De Ste- fani and Dyck adduce Menander fr. 5,4.25, though this excerpt contains genitive θητικοῦ, replicating Agathias 2.8.2.54 Nomina- tive θητικόν occurs at Agathias 3.16.4 and 4.21.6. Yet, again, the lexicographer could have found the headword in a lost part of Menander. Of potential relevance, Suda θ 376 illustrates θητικόν (acc.) by quoting an unascribed military narrative, which, as argued below, is most likely to be Menander’s Histories.55 Finally, while θ 7 θωπεία occurs in Procopius, Agathias, and extant Menander, circumstantial evidence slightly favours Procopius.56 Accordingly, the Menandrian origin of θ 3 and 6 is secure. On the grounds of sequence and external evidence, θ 5 is attrib- utable to Menander, but a section other than fr. 5,4.25. Unless 53 De Stefani, StIt 18 (1910) 444, and Dyck 946.8–9, 998, by not registering θριγκῷ in Agath. 4.14.3, imply that attribution to Menander is more straight- forward than is the case. 54 Men. fr. 5,4.25 τοῦ θητικοῦ καὶ οἰκετικοῦ; cf. Agath. 2.8.2 τοῦ τε θητικοῦ καὶ οἰκετικοῦ. See De Stefani, StIt 18 (1910) 444 (also noting Agath. 3.16.4); Dyck 946.10. 55 See 673 below. 56 The transmitted lemma is nom. sg. θωπεία. Dyck 947.21–22 cites Men. fr. 6,2.17 θωπείαν (acc. sg.) and Procop. 1.4.21 θωπείαις (dat. pl., cf. 8.18.6), the first instance in Wars. Both citations assume that the lexicographer modi- fied the lemma or it is corrupted. If, however, the lemma was once barely distinguishable dat. sg. θωπείᾳ, parallels are found at Agath. 4.28.4 and Procop. 1.25.4, also Procop. 4.1.9, 4.18.5, 4.26.22, 8.7.4, 8.16.23 (see the same transition from original dat. sg. fem. to a nom. sg. lemma at α 115 ἀβελτηρία < Procop. 2.15.8 ἀβελτερίᾳ; β 4 Βολωσία < Procop. 8.22.29 Βολοσίᾳ; ε 11 ἑωλοκρασία < Agath. 2.29.8 ἑωλοκρασίᾳ). In favour of Procop. 1.25.4 is its proximity to the unique source of α 112 αἱµύλος at 1.25.2. 674 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 displaced, θ 4 too should be assigned sequentially to a lost Agathias-inspired passage of Menander. An alternative deduc- tion that θ 1–6 comprise two intermixed series of lemmata cited from Agathias (1[–2], 4, 5) and Menander (3, 6) is unnecessary and otherwise problematic.57 In these circumstances, while the soundest method of isolating citations from lost parts of Menan- der’s Histories would be to admit only lemmata that nowhere occur in the other three authors, given Menander’s mimesis of Agathias, analysis should take into account those few lemmata, like θ 4 and 5, where relative sequence and/or other criteria point to a citation from Menander, even when the word/form is found in Agathias. Similar ambiguities occur in the second half of gamma. Fol- lowing nine lemmata (γ 1–9) cited in sequence from Agathias (up to 5.22.4), a pattern becomes hard to discern: γ 10 Γήπαιδες (all four authors) γ 11a γλιχόµενοι | γ 11b γλιχόµενος (all four authors) γ 12 γέρρα (lost Men.? Cf. Th. Sim. pref. 5 γέρρον) γ 13 γριφῶδες (lost Men.) γ 14 γεγωνότερον (Agath. 4.2.2 or lost Men.?) γ 15 γρυπός (lost Men.) γ 16 γωρυτούς (Agath. 3.28.5 or lost Men.?) Not found in the other three histories, γ 13 and 15 by default cite lost sections of Menander. Accordingly, if in correct sequence, γ 14 and 16, even though both lemmata occur in Agathias, could also be citations from Agathias-inspired passages of Menander.58 Alternatively, if in fact cited from Agathias, and thus now in 57 This proposition requires both (1) θ 3 θρυλλοῦντες is displaced from a Menander series into a preceding Agathias series; (2) if θ 4 θριγκῷ derives from Agath. 4.14.3, then θ 5 θητικόν should be cited sequentially from Agath. 4.21.6 (a chapter not otherwise cited), in which case the lexicographer deemed it necessary to explain this third occurrence of θητικόν but not two earlier instances at 2.8.2 and 3.16.4, even though he glossed other words in those chapters (α 132 < Agath. 2.8.7, σ 8 < Agath. 2.8.8, ο 8 < Agath. 3.16.8). 58 De Stefani, StIt 18 (1910) 442, assigns all four lemmata γ 13–16 to Menander, but observes that γ 14 γεγωνότερον occurs at Agath. 4.2.2. PHILIP RANCE 675 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 reverse sequence, γ 14 and 16 have been displaced from the preceding block of Agathias-derived lemmata at γ 1–9.59 In the absence of additional evidence, this question remains open. Separately, γ 12 γέρρα finds an inexact match in singular γέρρον in Theophylact’s preface 5, while this noun is absent from Pro- copius and Agathias, but, as De Stefani observed, the occurrence of plural γέρρα in a lost part of Menander is possible.60 Another debatable case occurs in delta: with initial δ 1–6 cited sequentially from Agathias, and concluding δ 12–13 from Pro- copius, intervening δ 7–10 all derive from Menander, extant and lost, with one apparent aberration: δ 7 διακηρυκεύεται (lost Men.) δ 8 διφθέραν (Men. fr. 6,2.31) δ 9 δαιταλεῦσιν (Th. Sim. 5.5.9; thus Dyck 926.33) δ 10 διασεσοβηµένος (lost Men.) The liminal position of δ 11 διαρρήδην equally permits that it is the last lemma cited from Menander (fr. 27,2.27) or the first from Procopius (1.5.20).61 From this sequencing, unless δ 9 is trans- posed, δαιταλεῦσιν could also have occurred in Menander, as De Stefani inferred.62 Although these cases in gamma and delta are insufficiently secure to be admitted into the selection of Menan- drian fragments below, they should not be entirely discounted as potential evidence. One last example must be excluded. Dyck (859) places ξ 5 ξυνερρώγει (Attic pluperf. ind. act. 3rd sg. συρρήγνυµι) among 59 If this were the case, γ 16 γωρυτούς (Agath. 3.28.5) and γ 14 γεγωνότερον (Agath. 4.2.2) were originally placed before or after γ 3[–8] γεγανωµένος, depending on which of two possible passages is cited in this lemma (Agath. 2.28.4 or 5.3.4). 60 De Stefani, StIt 18 (1910) 442. It is equally plausible that the lexicogra- pher pluralized γέρρον in Th. Sim. pref. 5, owing to the prevalence of γέρρα as the lemmatic form in his lexical sources: De Stefani, Etym.Gud. II 303.12– 15 app. font.; Dyck 921.26–30 app. font. This interpretation may be sup- ported by lemma ι 5 ἴυγγι, undoubtedly cited from adjacent Th. Sim. pref. 4. 61 Dyck 927.40–41. 62 De Stefani, StIt 18 (1910) 440. 676 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 “lemmata drawn from one of the historians,” but his edition (960.18–20) gives no source. In fact, ξυνερρώγει is unattested in the other three histories and Menander’s fragments. The word is precedented in Thucydides (1.66.1) and thence imitated by Atticizing authors (e.g. Cass. Dio 38.47.4) and cited in lexica (e.g. Hsch. σ 2538). While it might be tempting to derive ξυνερρώγει from a lost part of Menander, this case is not straightforward. The sequence of lemmata offers no firm assistance: ξ 1 ξυνεδεδέατο (Procop. 2.8.15) ξ 2 ξυνάρασθαι (Procop. 5.13.14) ξ 3 ξυναλιζόµενοι (Agath. 2.29.4) ξ 4 ξυνίστωρ (lost Men.) ξ 5 ξυνερρώγει ξ 6 ξύγκλυδες (Agath. 1.6.3) De Stefani assigned unattributed ξ 4 and 5 to Menander. He placed also ξ 6 in this category, without acknowledging that the lemma occurs in Agathias.63 However, ξ 5 can be linked to Agathias indirectly. The first gloss of ξυνερρώγει is συνερράγη, albeit the aorist of the same verb. This form is found in Agathias at 1.6.4 (at 17.19), as Atticized ξυνερράγη, a hapax. Moreover, it occurs within a few lines of the source of ξ 6 ξύγκλυδες at Agathias 1.6.3 (at 17.14). The close proximity of the two lem- mata within Agathias’ text cannot be accidental and conforms to the lexicographer’s method. Irrespective of the sequential dis- arrangement of these lemmata, therefore, and whatever the reasons for the lemmatic form of ξ 5 (where ξυνερρώγει may reflect the influence of prior lexical literature), there are no grounds for attributing ξ 5 to Menander.64 3. Lemmata and quotations attributable to lost sections of Menander’s Histories The foregoing analysis permits compilation of a dossier of material attributable to Menander’s Histories, but not found in 63 De Stefani, StIt 18 (1910) 440 (with editorial difficulties at ξ 5). 64 Other examples of a verb altered in the lemma: ξ 1 ξυνεδεδέατο < Procop. 2.8.15 ξυνδεδέατο; ω 2 ὠστίζοντο < Agath. 3.7.9 ὠθίζοντο. PHILIP RANCE 677 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 surviving excerpts and thus missing from all editions. This en- deavour revises and elaborates preceding research. With varying degrees of confidence, De Stefani (1910) ascribed 44 such entries (lemmata and/or quotations) to Menander, on the basis that whatever did not belong to the other three historians must be his, but without detailed argumentation. Dyck (1995) affirmed this interpretative logic, with adjustments, some explicit, others unsignalled. He added four items (lemmata: κ 2 καταγνύναι, µ 1 µαπία; lemma with quotation: µ 2 µετωπηδόν; quotation alone: χ 1 χιλός). Some lemmata that De Stefani deemed Menandrian Dyck reassigned to another author (γ 12 γέρρα, γ 14 γεγωνότερον, γ 16 γωρυτούς, δ 9 δαιταλεῦσιν, λ 9 λιγνυῶδες, ξ 6 ξύγκλυδες). Dyck culled a larger number of lemmata from De Stefani’s selec- tion without comment. Some of these I too reject as certainly or more plausibly cited from one of the other histories, though Dyck did not supply an alternative reference (ξ 5 ξυνερρώγει, ο 10 ὀρεῦσιν, π 25 πρῶρα, σ 16 σεµνῶς, τ 10 τενάγη). Dyck’s final list comprises 20 lemmata and six quotations (two from φ 5). It is unclear why Dyck includes only lemmata in alpha–mu and phi–psi. I find no grounds for excluding ten lemmata in nu– upsilon and omega, all previously identified by De Stefani, inso- far as they conform to the same criteria as those lemmata Dyck admits (ν 6 νεοχµοῦντες, ξ 4 ξυνίστωρ, ο 12 οὐριοδροµήσαντες, ο 13 ὀχµάσας, π 21 πιαλέα, π 22 παγιωθεῖσα, σ 18 σκυταλωτοὺς τροχούς, τ 11 τούδουνοι, υ 7 ὕπουλος ἄνθρωπος, ω 4 ὠρύω). One lemma in alpha selected by De Stefani but omitted by Dyck is cautiously readmitted (α 149 ἄτρυτα). A lemma that De Stefani and Dyck derive from an excerpt of Menander is reassigned to a lost section, based on form and external evidence (θ 5 θητικόν). Other lemmata may belong to this category (ε 36 ἐφειµένον, ε 38 ἐκµαγεῖον). A few doubtful cases are not admitted, even if they cannot be definitively rejected (γ 12 γέρρα, γ 14 γεγωνότερον, γ 16 γωρυτούς, δ 9 δαιταλεῦσιν, ε 33 ἐσκευωρῆσθαι, ε 35 ἐπιέζοντο). For organizational purposes, lemmata are listed first, then quotations. The lemma and gloss(es) are quoted, etymologies and syntactical notes omitted. Some entries allow limited com- 678 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 mentary, of others little can be said. Occurrences of a headword also in Procopius, Agathias, and/or Theophylact are recorded; the absence of cross-references indicates that the word, in any form, is not found in these three historians. Overall, the listed lemmata exhibit affinities with the vocabulary and usage in Agathias’ Histories, consistent with Menander’s well-documented linguistic-stylistic mimesis of his predecessor. Some words reflect recognized traits in Menander’s diction, including a penchant for poetic vocabulary (ζ 1 ζαµενής, ο 13 ὀχµάσας, τ 9 τελχῖνες, and quotation in ψ 1)65 and a taste for outré compound words (α 141 ἀνεπαφρόδιτος, σ 1 συνασµένιζε).66 3a. Lemmata α 141 ἀνεπαφρόδιτος (masc. nom. sg.): ὁ µὴ ἐπιχαρὴς καὶ ἀηδής, ὡς ἐπαφρόδιτος ὁ ἐπιχαρὴς καὶ ἡδύς, “unfavoured by Aphrodite: one who is not charming and unpleasing, as favoured by Aphrodite is one who is charming and pleasing.” Menander’s choice of this recherché adjective was presumably inspired by the prominent oc- currence of ἐπαφρόδιτος in Agathias’ opening comparison of History and Poetry: pref. 9 λίαν ὠδίνουσιν ἐπαφρόδιτα, “give birth to [off- spring] of exceeding loveliness.”67 The placement of α 141 first in the series of Menander-derived lemmata in alpha may reflect similar prefatory use of ἀνεπαφρόδιτος, while his selection of this cumber- some antonym perhaps signals competitive inversion of Agathias’ wording.68 α 144 ἀµβλυώττειν (Attic pres. ind. act. inf.): τὸ τυφλώττειν καὶ βαρύ- νεσθαι τοὺς ὀφθαλµούς, “to be weak-sighted: to be blind and beset with eye troubles.” Cf. Procop. Anec. 4.12 ἐνθένδε ἀεὶ ἀµβλυώττειν, “thereafter always weak-sighted.” See cognate ἀµβλύνω in Procop. 8.11.40; Agath. 1.16.3, 5.8.4 ἀµβλύνεσθαι τὰς ὄψεις, “dulled the sight,” 20.7. α 145 ἀµάρας (gen. sg. ἀµάρα): ὀχετοῦ ἢ τάφρου ἢ ὑδρορ⟨ρ⟩οῆς· διὰ τὸ τὰ ἐν αὐτῇ ὄντα φυτὰ ἀµάραντα εἶναι, “conduit: channel or ditch or 65 Baldwin, DOP 32 (1978) 120; Blockley, The History of Menander 7 n.27, with Rance, JÖB 73 (2023) 139–141, for Menander’s interest in poetry. 66 Blockley, The History of Menander 7 n.30. 67 See A. Kaldellis, “Agathias on History and Poetry,” GRBS 38 (1999) 295–305. 68 See generally Blockley, The History of Menander 2–3 n.9, 13. PHILIP RANCE 679 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 watercourse; on account of the fact that those plants in it do not fade.” This term apparently occurred in a description of a well- watered landscape. Excerpts of Menander’s Histories include reports of such rural landscapes within the ‘travelogues’ of Roman envoys: e.g. fr. 19,1.21–26. See below π 21 and χ 1. More specifically, fr. 9,1.19–21 records that an envoy at Dara τὸν κατὰ πόλιν ὁλκὸν τοῦ ὕδατος ἐπεσκεύασε καὶ ὑδροχεῖα ἐπενόησε, “repaired the city’s water supply and attended to cisterns.” α 146 ἀφαυρά (fem. nom. sg. or neut. nom./acc. pl. ἀφαυρός): ἀσθενέστερα, ἀµαυρότερα, ἢ κουφότερα, “feeble: weaker, fainter or slighter.” This adjective typically occurs in the comparative or super- lative.69 The only instance in late antique historiography is Agath. 2.15.5 ἀφαυροτάτη, describing earth tremors.70 α 149 ἄτρυτα (neut. nom./acc. pl. ἄτρυτος): ἀκάµατα, σκληρὰ καὶ ἰσχυρά· σηµαίνει δὲ καὶ ἀβλαβῆ, οἷον ἄτρητα, “unwearied: untiring, hardy and strong. It means also undamaged, like unperforated.” γ 13 γριφῶδες (neut. nom./acc. sg. adj. γριφώδης): σηµαίνει τὸ δύσκο- λον καὶ συµπεπλεγµένον λόγιον, “riddle-like: it means the difficult and complicated saying.” γ 15: γρυπός (masc. nom. sg.): ὁ ἐπικεκαµµένην ἔχων τὴν ῥῖνα, “hook- nosed: one who has a nose at a curved angle.” No suitable context or personage has been identified.71 δ 7 διακηρυκεύεται (pres. ind. 3rd sg. διακηρυκεύοµαι): διὰ κηρύκων συνθήκας ἢ πρεσβείας ποιεῖται, “negotiates by herald: through her- alds makes agreements or embassies.” Although precedented in Thuc. 4.38.2, the verb is rare in late antique historiography: see only Zos. 4.44.1. Doubtless the cited context was diplomatic. δ 10 διασεσοβηµένος (masc. nom. perf. part. pass. sg. διασοβέω): 69 LSJ9 s.v. ἀφαυρός. The three comparative glosses are a reflex of the com- piler’s lexical sources: Dyck 915.41 app. font. 70 By way of comparison, σ 11 στωµύλος is cited from Agath. 2.30.1 στωµυ- λώτατος, where the lexicographer altered a masc. sg. superlative adj. in the source-text into a masc. sg. positive adj. as the lemma, following a persistent lexicographical tradition of στωµύλος (Dyck 976.50 app. font.). I have discounted the possibility that α 146 ἀφαυρά is likewise adapted from Agath. 2.15.5 ἀφαυροτάτη (and not cited from a lost section of Menander), as there is no corresponding development with ἀφαυρά in the compiler’s lexical sources that might have motivated this intervention. 71 See Dyck 860 on a possible lost source of the gloss (distinct from the lemma). 680 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 τεταραγµένος, ἄτακτος, τεθορυβηµένος, “having been scared away (or agitated): thrown into confusion, disorderly, set in turmoil.” The verb is rare. The sole use in extant ancient historiography is Agath. 3.11.1 αὐτοὺς ἀνεκίνει καὶ διεσόβει, “stirred up and agitated them,” regarding a crowd’s reaction to a speech. Another instance occurs in an unidentified anonymous quotation from a battle description in Suda ρ 218 οἱ δὲ διεσοβήθησαν. ε 37 ἐλεηλάτει (imperf. ind. act. 3rd sg. λεηλατέω): ᾐχµαλώτευεν, ἐλῄστευεν, ἥρπαζεν, “(he/it) was plundering: was taking captives, robbing, snatching.” This verb occurs in a different form (pres. inf. act.) in Menander fr. 21.12–13 παύσαιντο µὲν τοῦ τὴν Ῥωµαϊκὴν λεηλατεῖν, “(they) would cease to plunder Roman territory,” con- cerning raiding Sklavenes around 578. Widely used in classical historiography, though not in Thucydides, λεηλατέω is relatively uncommon in the sixth century. Procopius’ oeuvre has one instance: Anec. 2.28 λεηλατήσειε. Exceptionally, it is favoured by Agathias, in- cluding in the cited form: 4.22.6 τῶν λεηλατουµένων, 29.8 ἐλεηλάτει, 5.1.2 ἐλεηλάτουν, 11.7 λεηλατήσοντας, 13.1 ἐλεηλάτουν, 23.9 λεη- λατοῦντες. ε 39 ἔναυλος (nom. sg.): ἔµπνους· σηµαίνει δὲ καὶ ῥοῦν καὶ χαράδραν καὶ χειµάρρουν, “still fresh: alive; it means also a stream and mountain-torrent and winter-flow.” The lexicographer registers two of several meanings, one adjectival, the other nominal. The priority accorded to the adjective implies that this usage was found in Menander. Correspondingly, ἔναυλος occurs in the other three historians only in this sense, variously in relation to still-audible sound, persistent fears and suspicions, or fresh memories: Procop. 5.10.10 πάταγος ἔναυλος, Agath. 5.5.6 ἔναυλον … τὸ δέος, 9.1 ἔναυλον … τὸ πάθος καὶ ἡ ὑποψία; Th. Sim. 4.6.5 ἔναυλον τὴν µνήµην. ε 41 ἐλινύειν (pres. act. inf.): ἀργεῖν, σχολάζειν, “to rest: to be un- occupied, to be at leisure.” Despite instances in canonical historians, including Herodotus and especially Arrian, ἐλινύω is rare in later historiography: see only Evagr. HE 5.24 (219.23 Bidez/Parmentier), 6.13 (231.12), in both cases referring to individuals who do not “re- main inactive” in response to events. ζ 1 ζαµενής (masc./fem. nom. sg. adj.): εὔψυχος, µέγα µένος ἔχων, τουτέστι ψυχὴν ἢ προθυµίαν. ἢ ζαµενής, ὁ ἄγαν ὀργίλος, “forceful: of good spirit, having great force, that is spirit or morale. Or … one PHILIP RANCE 681 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 much inclined to anger.”72 The lengthy definition acknowledges potential nuance. The headword previously occurs only in verse, and dependent scholia and lexica, and is unattested in prose before the Late Byzantine era. Well-known instances include Pindar and Sophocles.73 Menander’s choice of this adjective may reflect his poetic interests,74 though concern for distinctive lexical variatio is discernible; for example, in the same semantic field, adj. ὀργίλος, “enraged” (fr. 9,3.82, 21.47), verb νεµεσίζω “feel indignation” (15.17), and noun µῆνις “wrath” (25,1.79), none of which are found in the other three histories. ζ 3 ζάκορος (nom. sg. noun): νεωκόρος, “sacred warden: temple at- tendant.”75 The evidence permits only conjecture. A possible context is Menander’s fascination with the story of the apostate magus Isao- zites/Isbozetes (Yazdbōzēd), whose public profession of Christianity, as martyrial acta report, occurred during his participation in a Zoroastrian fire ceremony at Dvin.76 θ 3 θρυλλοῦντες (masc. nom. pres. part. act. pl. θρυλλέω): τουτέστι θόρυβον ἐγείροντες, “making a commotion … that is raising tumult.” No formal gloss is provided, but the etymological notice concludes with this definition. Classicizing orthography prefers θρυλέω, though variation occurs between and within codices and editions.77 Besides compounds and cognates, Procopius and Agathias use θρυλ(λ)έοµαι in a distinct sense of information “noised about” or “tales told”: Procop. 8.20.48 θρυλλούµενα (cf. Anec. 25.19); Agath. 2.24.8 θρυλου- µένους, 4.23.4 τὰ … θρυλούµενα, 5.11.3 τὸ θρυλούµενον. Excerpts of Menander attest only διαθρυλ(λ)έω, similarly of news “spread 72 Dyck 943.2–944.5 app. font. 73 LSJ9 s.v. ζαµενής. 74 See 695 below. 75 Dyck 861, 944.8–10 app. font., shows that the lexicographer drew the gloss and subsequent etymological notice from Etym.Gen. K. Alpers, “Beo- bachtungen zur Überlieferung und zum Text des Platonlexikons des Timaios,” ZPE 166 (2008) 85–99, at 98 n.82, “mit Sicherheit,” brands ζ 3 a later interpolation from Etym.Gen., though on doubtful grounds, while his view of the chronology and editorial process is less certain in Literature 302 n.52, 309–310 nn.84–85. 76 Men. fr. 13.3–4; Anth.Pal. 1.101, with Men. T3 (schol. Strab. 15.3.15), and discussion in Rance, JÖB 73 (2023) 133–147. 77 Procopius: e.g. 2.8.17 ἐπιθρυλοῦντες (Haury I 187.1 app. crit.), 4.21.2 ἐπιθρυλλοῦντες (I 518.5). Agathias: Keydell ed. xxxv–xxxvi. 682 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 abroad”: fr. 10,4.1 διαθρυλληθέν, 20,2.34 διεθρυλήθη. Preceding lemma ε 22 ἐπιθρυλοῦντες, citing Procop. 4.21.2, is glossed ἐνο- χλοῦντες, θορυβοῦντες, “being a nuisance, causing tumult,” whereas here Procopius clearly means “spreading the report that…” The lexicographer may have likewise misconstrued θ 3 θρυλλοῦντες. θ 4 θριγκῷ (dat. sg. θριγκός): τῷ τροχώµατι, “(by/to/on) a parapet: the circuit.” The gloss is significantly rarer than the headword.78 The original sense of an uppermost course of masonry or coping stones metonymically came to mean an entire circuit-wall. The lemma occurs in Agath. 4.14.3 αὐτοῖς … τῷ θριγκῷ περιεχοµένοις, “being enclosed by the circuit-wall.” This passage presumably inspired Menander’s usage.79 θ 5 θητικόν (masc. acc. or neut. nom./acc. sg. adj. θητικός): θερα- πευτικόν, ὑποχείριον, ἐργαστικὸν τῶν δεόντων, µισθαρνικόν, “menial: at one’s service, at hand, performing necessary tasks, working for hire.” In sixth-century historiography, neuter substantive θητικόν, relatively rare in all eras, becomes a collective designation for the household servants or staff of military personnel, and, by extension, “camp-followers” in general.80 Initially, this expression is particular to Agathias and, in imitation, Menander. Agathias uses θητικόν to denote a domestic retinue: 2.8.2 θητικοῦ, 3.16.4 θητικόν (nom.), 4.21.6 θητικόν (nom.). In Menander fr. 5,4.25 τοῦ θητικοῦ καὶ οἰκετικοῦ the term possibly extends to armed retainers. In any case, Menander reprises Agathias’ wording (2.8.2) τοῦ τε θητικοῦ καὶ οἰκετικοῦ.81 As the sequence of lemmata implies that θ 5 is cited from Menander rather than Agathias, De Stefani and Dyck identify fr. 5,4.25 as the source, despite genitive θητικοῦ.82 However, as the gloss is not obviously indebted to any lexicographical tradition that might have induced the lexicographer to regularize the headword from genitive to nominative,83 the possibility that θ 5 cites θητικόν from another, lost passage of Menander should be considered. In support of this proposition can be adduced Suda θ 376, which exemplifies use 78 LBG s.v. τρόχωµα. 79 See 672–673 above. 80 A distant antecedent is found in Arr. Tact. 2.1. See Middle Byzantine usage in LBG s.v. θητικόν. 81 Blockley, The History of Menander 253–254 n.29; PLRE IIIA 241, “Bonus 4.” 82 See 673 above. De Stefani, StIt 18 (1910) 444; Dyck 946.10. 83 Dyck 946.10–11 app. font. PHILIP RANCE 683 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 of θητικόν with a quotation from an unidentified classicizing history: καὶ θητικὸν ἐπήγοντο οἱ αὐτοὶ ἐς τὸν πόλεµον οἱ Πέρσαι πλεῖστον ὅσον καὶ παιδία καὶ γυναῖκας, οὓς κατά τινα δὴ πάτριον νόµον ἐς τοὺς κινδύνους θεατὰς οἰκείας ἀλκῆς ἦγον, “And a body of camp-followers the Persians themselves brought forth into battle, of so great a num- ber, and children and women, whom, doubtless in accordance with some ancestral custom, they brought into danger as observers of their own prowess.” Leaving aside prior conjectures, as the com- piler(s) of the Suda drew this category of historical citation from selected volumes of the Constantinian Excerpta, the range of possible authors is confined to the canon of historians excerpted in that project.84 More specifically, of the handful of those historians who record—or, if fragmentary, could have recorded—“Persian” military operations, only two are known to have employed the key-term θητικόν: Agathias and Menander. Other stylistic features also point to Menander.85 At the very least, Menander must be considered the most likely candidate, enhancing the prospect that his use of θητικόν need not be restricted to extant fr. 5,4.25. θ 6 θεοκλυτοῦσιν (pres. ind. act. 3rd pl. θεοκλυτέω): ὡς θεοὺς ἐπι- καλοῦνται, “they pray to God(s): they invoke them as gods.” The gloss excludes the possibility that θεοκλυτοῦσιν is alternatively neut./ masc. dat. pl. pres. participle. In sixth-century historiography, the verb (pres. act. inf.) occurs uniquely in Menander fr. 6,1.403 µήτε µὴν θεοκλυτεῖν ἀκουσίως τοὺς παρὰ Μήδοις νενοµισµένους θεούς, “nor against their will pray to the gods in whom the Medes believe,” in a protocol appended to the peace treaty of 561/2, guaranteeing liberties to Christian inhabitants under Sasanian rule. The lemma θεοκλυτοῦσιν must be cited from another, now-lost passage, pre- 84 Németh, Excerpta 238–255, with bibliography. Previously, H. Valesius, Ammiani Marcellini Rerum gestarum … libri XXVIII (Paris 1636) 278–279, assigned the quote to Cassius Dio. A. Adler, Suidae Lexicon II (Leipzig 1931) 718.7–9 app. font., dismissing Valesius’ proposal, tentatively suggests Anon. FGrHist 151: Anonyme Alexander-geschichte (Fragm. Sabbaiticum). Neither conjecture has merit. 85 E.g. πλεῖστον ὅσον: exceptionally rare in late antique secular histori- ography, the construction πλεῖστ- ὅσ- is common in Agathias (Keydell ed. 229, Index graecitatis, s.v. ὅσος adiectivis subiectum) and imitated by Menander: fr. 12,2.28 πλεῖστα ὅσα κακηγορήσαντες, 19,1.132 ἕτερα πλεῖστα ὅσα δια- λεχθείς, 20,1.48 πλείστων δὲ ὅσων ῥηµάτων, 23,1.9 πλείστας ὅσας κώµας, 23,9.61–62 πλείστας ὅσας ἡµέρας. 684 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 sumably also concerning religious matters.86 ι 9 ἰαχνόν: θαλπεινὸν ἢ σαπρόν. The transmission of the headword is unsure: ἰαχνόν Etym.Gud. (MSS. CZ: Sturz 626.60) : ἰσχνόν Lex.Αἱμ. (D), Etym.Magn. (479.22). The lexicographer apparently derived the glosses from Hsch. ι 107 ἰάχεον: θαλπεινόν. σαπρόν, where also the transmitted headword is doubted.87 If ἰαχνόν is correct, as edited by Dyck (949.25), the word is attested only in this lemma and derivative lexica. The glosses indicate masc. acc. or neut. nom./acc. sg. adjec- tive ἰαχνός. The given meanings “warm(ing) or fetid” offer no scope for comment. If ἰσχνόν is preferred, adjective ἰσχνός occurs in Pro- copius and Agathias, but has a different semantic range, variously “thin, lean, withered,”88 which does not easily accommodate the glosses supplied, unless the lexicographer erred. κ 2 καταγνύναι (pres. inf. act. κατάγνυµι): κατάσσειν, κατακλᾶν, συν- τρίβειν, “to break apart, to break down, shatter.” The first gloss is a late form of the headword. The verb occurs in Agath. 3.25.5 κατα- γνύντες (masc. nom. pres. part. act. pl.), of large stones, hurled by a city’s defenders, “crushing” the besiegers’ mantlets. The form κατα- γνύναι is uncommon but precedented in Thuc. 4.11.4.89 µ 1 µαπία: σπογγίστρια· Both the headword and the gloss, apparently fem. nom. sg. nouns, are unattested beyond derivative lexica. LBG construes a female cleaner or servant, plausibly on the grounds that verb stem (σπογγίζω) + suffix -τρια commonly denominates a fem- inine agent. A partial correspondent in Modern Greek suggests 86 Rance, JÖB 73 (2023) 139 n.62. 87 Dyck 949.25 app. crit. and font. Cf. Hsch. ι 107 ed. Latte II 345.107 (app. crit.), who marks ἰάχεον with a crux and conjectures ἰσχνόν on the basis of Etym.Magn.; thus rev. ed. Cunningham II.1 (2020). See Theodoridis, JÖB 50 (2000) 385. 88 Procop. 4.13.13 ἰσχνός, 6.20.23 ἰσχνοί (cf. Anec. 8.12 ἰσχνός, Aed. 4.10.6 ἰσχνόν); Agath. 2.23.6 ἰσχνοί. 89 As letter-section kappa contains only two lemmata, with κ 1 κατασπι- λάζοντες (Th. Sim. 4.5.10), sequence provides no assistance in identification. I considered but rejected the proposition that lemma κ 2 καταγνύναι (pres. inf. act.) might be a ‘standardization’ of καταγνύντες (pres. part. act.) in Agath. 3.25.5. The lexicon contains no other case of a verbal form in the cited text changed to infinitive, while the most likely basis for such an intervention, namely the lexicographer’s preference for the lemmatic form in a lexico- graphical tradition or harmonization of the lemma with glosses imported from lexical sources, does not appear to apply here (Dyck 950.6 app. font.). PHILIP RANCE 685 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 rather an implement such as a mop.90 µ 14 Μαµερτῖνοι: Μάµερτος ἐν Ἰταλίᾳ ὁ Ἄρης, ὅθεν καὶ οἱ ἐν Σικελίᾳ οἰκήσαντες στρατιῶται Μαµερτῖνοι ὠνοµάσθησαν, “Mamertinoi: Ma- mertos in Italy [is] the [name of] Ares, whence also the soldiers who dwelt in Sicily were called Mamertinoi.” While attribution of µ 14 to Menander has never been questioned,91 it is perhaps the most baffling entry in the Lexicon Αἱμωδεῖν. Certain aspects could signal interpolation. It is seemingly tacked on to the end of letter-section mu, widely separated from preceding Menander-derived lemmata (µ 1–2). The lemma is unique in being the proper name of a people or group. The reference to Italic mercenaries in Sicily in the third century B.C. finds no obvious place in the historical and geographical coverage of Menander’s Histories. In contrast, similar displacement is found in sigma (σ 1, 18[–20], 24), without controversy (see below). Dyck persuasively derives the definition (Μάµερτος … ὠνοµάσθησαν) from an early recension of the Etymologicum Genuinum, a source the lexicographer used in other entries of undisputed authenticity.92 In any case, positing interpolation does not resolve the incongruity: a posterior editor’s decision to insert this entry would be no less puzzling. Judging by surviving fragments, Menander’s narrative occasionally touched on events in Italy during the 560–570s, par- ticularly Roman-Lombard warfare and diplomacy, though this was an explicitly peripheral, low-priority theatre, seemingly never treated in detail (fr. 22 and 24). The wider historical record attaches no sig- nificance to Sicily nor hints at why it might have been mentioned. A context for “Mamertinoi” can only be speculated. Perhaps Menan- der’s reports of Roman efforts to hire Lombard chieftains to fight the Persians in Oriens (fr. 22.9–15, 24.11–16) prompted a strained and arcane allusion to ancient mercenaries “from Italy.” Alternatively, notwithstanding the transmitted lemma, the object of interest may have been Μάµερτος, adduced as an erudite synonym of Mars/Ares or personified War.93 Neither proposal is compelling. 90 LBG s.v. µαπία, “die aufwischt, Putzfrau.” See Λεξικό της κοινής νεοελλη- νικής (Thessaloniki 1998) s.v. µάπα3, “είδος σφουγγαρίστρας.” 91 De Stefani, StIt 18 (1910) 443 (cited as Μεµερτινός); Dyck 998 (as Μαµερτῖνος, correctly 957.49 Μαµερτῖνοι); Bleckmann, in Les historiens frag- mentaires 67. 92 Dyck 861–862, 957.49–51 app. font. 93 Cf. Hsch. µ 214 Μάµερτος: Ἄρης; schol. Lycoph. Alex. 937–938, 1409– 1410 (II 307, 386 Scheer). 686 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 ν 6 νεοχµοῦντες (masc. nom. pres. part. act. pl. νεοχµόω): µετακοσµοῦν- τες, “making changes: rearranging.”94 This verb occurs in two excerpts of Menander: fr. 6,1.388 (Exc.leg. I 182.11) τῷ δὲ ἀπατηλῷ τῶν τε ἐµπεδωθέντων νεοχµῶσαί τι βουλοµένῳ, “if someone deceitful wishes to change anything that had been ratified”; fr. 7,6.6 (Exc.sent. 21.3) ἡ περιφορὰ νεοχµοῦσα τοῦ χρόνου, “the passage of time pro- ducing changes.” See also Procop. 3.7.22, 8.1.11 (cf. Anec. 6.21, 30.21; Aed. 2.7.7, 2.10.8, 3.5.5, 6.5.7); Th. Sim. 8.12.13. The verb is not uncommon. Nonetheless, the lexicographer cites another form (aor. ind. act. 3rd sg.) at ε 20 ἐνεόχµωσεν (Procop. 3.7.22), a rare case of such duplication.95 ξ 4 ξυνίστωρ (Attic masc./fem. nom. sg.): συµπράκτωρ, σύµβουλος, συγγνώστης. Without context, the glosses leave Menander’s meaning uncertain: “accomplice, counsellor, confidant.” The headword is otherwise unattested in late antique historiography. The Attic form is rare, though precedented in Thuc. 2.74.2 with the sense of “wit- nesses,” while Koine συνίστωρ occurs in diverse genres. ο 12 οὐριοδροµήσαντες (masc. nom. aor. part. act. pl. οὐριοδροµέω): ἀντὶ τοῦ φορῷ ἀνέµῳ πλεύσαντες, “having run with a fair wind: instead of: having sailed with a favourable wind.” Use in prior or contemporary historiography is rare, though see Evagr. HE 2.5 (51.15 Bidez/Parmentier = Priscus fr. 22.15–16 Carolla / fr. 28.14– 15 Blockley) οὕτω πνεύµατος ἐπιτυχόντας οὐριοδροµῆσαι, “chancing upon so fair a wind,” regarding troop shipments in the eastern Mediterranean. Cf. Diod. 3.34.7 οὐριοδροµούσαις ναυσὶ φορτίσιν, “merchant vessels sailing with a fair wind.” Menander fr. 19,1.18 shows that he included such content: χρησάµενος δὲ ταῖς ταχυπλόοις τῶν ὁλκάδων, “using the fast-sailing merchant vessels,” regarding Roman envoys to the Turks in 575/6. ο 13 ὀχµάσας (masc. nom. aor. part. act. sg. ὀχµάζω): σηµαίνει τὸ κρα- τήσας καὶ κόψας, “having held fast: it means having laid hold of and checked.” Exceptionally rare in prose, the sole instance of the cited form is Lycoph. Alex. 41. Apart from lexica, the only late antique use is by another Constantinopolitan author and close contemporary: Paul Silentarius, Descriptio ambonis 237 µιῆς δ’ ἀπὸ δέσµιον / ἀρχῆς 94 Dyck 859 includes ν 6 νεοχµοῦντες among “lemmata drawn from one of the historians,” and at 959.20 app. crit. signals two instances of νεοχµόω in Menander’s excerpts, but he refrains from explicitly attributing the lemma to Menander. See De Stefani, StIt 18 (1910) 413. 95 The only other instance is ε 12 ἔστεργεν and σ 15a στέργων. PHILIP RANCE 687 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 ὀχµάζων, “by a bond at one point holding fast.” π 21 πιαλέα (fem. nom. sg. or neut. nom./acc. pl. πιαλέος): κάθυγρος, λιβάδα ἔχουσα, “…well-watered/waterlogged, having meadows/ marshes/springs.” This adjective, a late prose equivalent to πίων, has a wide semantic and contextual range denoting richness, fertility, abundance, plenitude. The glosses lack obvious antecedents in lexi- cal literature.96 If specific to the cited context, they place πιαλέα in a description of landscape. Compare Menander fr. 19,1.22–26 ἐκεῖνα δὴ οὖν τὰ ἐκ τῶν λιµναίων ὑδάτων περιθεόµενα πεδία … ἱππασάµενοι καὶ ἑτέρους δὲ πλείστους δονακώδεις τε καὶ λοχµώδεις ἔτι τε καὶ ὑδατώδεις διεληλυθότες χώρους, “having ridden over those plains bounded by marshy waters and traversed many other locations covered with reeds and thickets and wetlands.”97 Cf. α 145 and χ 1. π 22 παγιωθεῖσα (fem. nom. aor. part. pass. sg. παγιόω): ἑδραιωθεῖσα, βεβαιωθεῖσα, πιστωθεῖσα, “having been resolved: having been estab- lished, confirmed, bound with pledges.” This rare verb occurs in fr. 6,1.365 (Exc.leg. I 181.22) ἐπαγιώθη (aor. ind. pass. 3rd sg.), “it was resolved,” in the provisions of the Roman-Persian treaty of 561/2. The differing form παγιωθεῖσα must be cited from another, now-lost passage, probably in a corresponding context. See similarly Theoph- anes Chron. A.M. 6179 (364.6 de Boor): τὴν µετὰ τῶν Βουλγάρων παγι- ωθεῖσαν εἰρήνην, “the peace that had been agreed with the Bulgars.” π 26 πέµπελος: (nom. sg. noun): ὁ γέρων· παρὰ τὸ εἶναι πλησίον τοῦ{τέστι} πέµπεσθαι εἰς Ἅδην, “…the old man; [derived] from being close to being sent to Hades.”98 An approximate English ex- pression might be “a goner.” This noun is a lexical curiosity; except for lexicographical and exegetical sources, prior use is restricted to Lycophron and Galen,99 authors in entirely different genres, though its currency widens in later Byzantine literature. A semantic parallel may be drawn with Procopius’ fondness for τυµβογέρων, “grave dodger” (8.12.33, Anec. 6.11), a possibly Aristophanic coinage (fr. 55 Demianczuk; Com. adepst. 1172 Kock), rarer than πέµπελος and otherwise unknown in ancient historiography. σ 1 συνασµένιζε (pres. imperat. act. 2nd sg. συνασµενίζω): συνευφραί- νου, “be glad with: rejoice with.” As this verb nowhere occurs in the 96 Dyck 970.80 app. font. 97 Apostolopoulos, Μένανδρος 16, draws parallels with Agath. 2.19.4. 98 See Dyck 971.94–97 app. crit. and app. font. on the source of the gloss and textual transmission. 99 LSJ9 s.v. πέµπελος. 688 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 three extant historians, the lemma should be assigned to Menander, even if its detachment from other Menander-derived lemmata in sigma (σ 18[–20], 24) defies explanation and has elicited tacit hesi- tation.100 While cognates are lacking in Procopius and Agathias, Theophylact favours simplex ἀσµενίζω (3.2.6, 3.7.6, 3.9.8, 3.14.1, 5.4.13, 5.7.3, 6.11.20) and once prefers compound κατασµενίζω (4.16.20 κατησµένισεν). Use of συνασµενίζω is exceptionally rare in any period and unknown in historiography; no earlier instance is firmly attested.101 Yet, scarcer still is Theophylact’s choice of κατα- σµενίζω, reportedly a hapax.102 Attribution to Menander is con- sistent with his taste for unusual συν- compound verbs: συνανα- ζωπυρεῖ (fr. 6,1.483), συνεπαµύνων (9,3.66–67), συνεκπολεµῶσαι (13,5.5), συνεψεύδετο (26,1.90), συναιχµαλωτισθῆναι (27,3.17). Im- perative συνασµένιζε suggests a speech episode. In formal exchanges between envoys and rulers the imperative plural is typical, while the singular seldom occurs.103 Nonetheless, certain diplomatic protocols included invocations to “rejoice together.”104 σ 18 σκυταλωτοὺς τροχούς (acc. pl.), “cogged(?) wheels.” One of a handful of two-word lemmata (see υ 7 below), clearly its context was specific, if enigmatic. Adjective σκυταλωτός otherwise occurs only in mechanical treatises in reference to a “toothed” or “cogged” wheel or drum (σκυταλωτὸν τύµπανον) within a mechanism: Hero Dioptra 34, 36 (III 294.9–11, 296.8, 298.12 Schöne), whence derivative Pap- pus Synagoge 8 (1028.27 Hultsch). The single-word gloss supplied, adjective ῥαβδωτούς, variously signifies “made or plaited with rods,” 100 De Stefani, StIt 18 (1910) 439, “da me non identificata,” though σ 1 otherwise meets his criteria for attribution to Menander. Dyck 972.2 gives no in-text reference. 101 Philo De specialibus legibus 1.316 (V 76.10 Cohn): συνασµενίζειν is an (uncompelling) editorial conjecture. The earliest secure occurrence is other- wise eighth-/ninth-century Cosmas Vestitor Vita Joannis Chrysostomi, ed. F. Halkin, Subsidia hagiographica 60 (1977) 434.9. Lexical sources: Dyck 972.1–3 app. font. 102 LBG s.v. κατασµενίζω. 103 I note two instances of imperative singular: fr. 26,1.131 πρᾶττε, 12,5.54 νόµιζε. 104 E.g. Men. fr. 19,1.37–38 συνησθῆναι φράσας τῷ Καίσαρι Ῥωµαίων, “declaring that [a Turkish ruler] rejoice with the Roman Caesar”; cf. Procop. 8.24.7 χρῆν µάλιστα βασιλέα Ἰουστινιανὸν συνήδεσθαι Φράγγοις, “Emperor Justinian ought certainly to rejoice with the Franks.” PHILIP RANCE 689 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 “wicker,” or somehow “ribbed.”105 Evidently the cited passage con- cerned some kind of device. Most likely is a description of poliorcetic machinery, as found in unplaced Menander fr. 40 (Suda σ 901), concerning spaliones, portable hide-covered wicker and/or timber mantlets employed by sappers during tunnelling operations (cf. Agath. 3.5.9–11 on the same contrivance). It may be no coincidence that following σ 24 σήραγγες seems to be cited from this fragment (40.9 τοὺς σήραγγας).106 This opens the possibility that σ 18 and 24 derive from a single siege narrative. Alternatively, Agathias’ digres- sion on Anthemius of Tralles’ experiments in hydraulics, catoptrics, and acoustics (5.7–8) supplies another scenario for such technical vocabulary, though there is no evidence that Menander imitated his forerunner in this respect. τ 9 τελχῖνες (nom. pl. τελχίς, later τελχίν): οἱ φθονεροὶ καὶ βάσκανοι δαίµονες, “telchines: the envious and malevolent spirits.”107 In Greek mythology, the Telchines were a primordial race of minor deities with particular destructive attributes. This appellation was extended to a wider notion of a malign or mischievous spirit(s) or evil sorcerer, and, more generally, a spiteful person or slanderer.108 Although τ 9 ostensibly distinguishes the intended sense, as the lexicographer drew this definition from one or more of his usual lexical sources (and possibly even adapted the lemma accordingly), the form and mean- ing of τελχίς,-ῖνος in Menander’s text are not beyond doubt.109 In late antiquity, use is mostly restricted to oratory, epic verse, and lexicography, and unknown in historiography. Beyond lexica and scholia, the few sixth-century instances illustrate semantic range. Menander’s contemporary, Paul Silentarius, Descriptio S. Sophiae 195 Τελχῖνας κακοεργέας, rhetorically implicates “maleficent Telchines” in the collapse of the dome of Hagia Sophia in 558, presuming his audience’s familiarity with this classical allusion. A few decades 105 LSJ9 s.v. ῥαβδωτός. 106 De Stefani, StIt 18 (1910) 439; Dyck 978.85. Nom. pl. σήραγγες occurs in Procop. 8.1.9 and Agath. 1.10.8, but the sequence of σ 24 σήραγγες favours citation from Men. fr. 40.9, despite acc. pl. σήραγγας, as the case of the lemma merely reflects the compiler’s adherence to a long-term lexicographical tra- dition: Dyck 978.85–88 app. font. 107 See Dyck 865, 981.38 app. crit. on the textual transmission of τ 9. 108 LJS9 s.v. Τελχίς. 109 Dyck 981.38 app. font. (where Σ is now Synagoge τ 85 Cunningham), 1007 Index fontium. 690 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 earlier, Simplicius employs τελχίν in philosophical discourse as a pejorative metaphor for a harsh critic of his writings.110 While a specific pagan-mythological reference to Telchines may seem incon- gruous, Menander’s generic usage would be broadly consistent with the language and causational concept of ‘daemons’ (δαίµονες) and ‘envy’ (φθόνος, βασκανία) in the works of Procopius and Theophy- lact, less so Agathias.111 In addition, Menander’s fondness for ethno- graphical exotica, especially arcane religious rituals of barbarians, offers potential contexts particular to his Histories.112 τ 11 τούδουνοι: οἱ τοποτηρηταὶ παρὰ Τούρκοις, “toudounoi: the topotērētai among Turks.”113 Ordinarily prosaic administrative jargon, τοπο- τηρητής variously denotes a representative, deputy officer-holder, district overseer, or local commandant. As use in literary works is scarce, whether Menander utilized this explanatory term remains uncertain.114 In Greek sources, τούδουνος is attested only as the title (albeit misconstrued as a personal name) of a governor of Kherson in 710/1, whom the Khagan of the Khazars had appointed as his representative.115 The etymon is Turkic tudun. Old Turkic inscrip- 110 Simplicius In Aristotelis de caelo commentaria (J. L. Heiberg, CAG VII [Berlin 1894]) 66.10; In Aristotelis physicorum libros octo commentaria (H. Diels, CAG X [Berlin 1895]) 1117.16 111 D. Brodka, Die Geschichtsphilosophie in der spätantiken Historiographie. Studien zu Prokopios von Kaisareia, Agathias von Myrina und Theophylaktos Simokattes (Frank- furt 2004) 31–39, 196–200. 112 E.g. Men. fr. 10,3.9–20, shamanistic rituals of Göktürks; 19,1.118–131, their funerary rites and facial scarification; 25,1.67–89, Avar khagan swear- ing oaths. See Baldwin, DOP 32 (1978) 118–119, “addicted to the exotic”; Blockley, The History of Menander 13, “indulge a taste … for exotic details”; B. Bleckmann, “Christentum und nichtchristliche Religionen in nachjustini- anischer Zeit: Das Zeugnis des Menandros Protektor,” in S. Ratti (ed.), Une Antiquité tardive noire ou heureuse? (Besançon 2015) 227–250, at 241–243, “Un- bekannte exotische und auch primitive religiose Vorstellungen”; further re- marks in Rance, JÖB 73 (2023) 139, 143–146. 113 Dyck 982.42 app. crit. prefers the indirect tradition (Etym.Gud. and Etym. Magn.). The direct tradition (D) reads singular τούδουνος: ὁ τοποτηρητής. 114 In late antique historiography, see only Evagr. HE 2.18 (91.3, 13–14 Bidez/Parmentier), regarding papal deputies. See LBG s.v. τοποτηρητής, with older literature. 115 Theophanes Chron. A.M. 6203 (378.3–5 de Boor): Τουδοῦνον δέ, τὸν ἄρχοντα Χερσῶνος ὡς ἐκ προσώπου τοῦ Χαγάνου ὄντα (also 379.1–2, 7–10); PHILIP RANCE 691 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 tions and Arabic and Chinese sources document this office or rank among the Göktürks by the early eighth century, and it is also reported among the eighth-/ninth-century Avars. Turkological scholarship has not appreciated the potential significance of τ 11.116 Menander’s detailed accounts of diplomatic relations between Con- stantinople and the Khaganate of the Göktürks ca. 568/9–576 provide an obvious and unique setting for this lemma. Information in surviving excerpts, particularly details of Turkish titles and dig- nities, strengthens this inference.117 υ 7 ὕπουλος ἄνθρωπος (nom. sg.), “treacherous man.” Menander’s use of adjective ὕπουλος is confirmed in fr. 12,6.75 (Exc.leg. I 197.34) φιλίας ὑπούλου καθεστώσης, “[the Avars’] friendship being treach- erous.” Some features of υ 7 call for caution. Only four other lemmata are a two-word phrase, and in one instance the nominative case is due to editorial interference.118 No gloss is supplied. A partly corrupted etymological notice, drawn from Orion Etym. s.v. ὕπουλος (156.13–14 Sturz), explains a metaphorical extension of a term for a festering sore hidden beneath healthy tissue.119 While the exact phrase ὕπουλος ἄνθρωπος may have occurred in a lost section of Menander’s Histories, one might suspect the lexicographer of har- monizing the lemma with nominative ὕπουλος in his lexical source(s) and, perhaps, of inserting ἄνθρωπος to distinguish specific applica- tion to a human character trait. Although the other three historians also employ ὕπουλος, none applies this word to a person: Procop. 7.16.11 ὑπούλων … ῥηµάτων; Agath. 1.15.7 ὕπουλον … εἰρήνην, Nicephorus Patr. Brev. 45 (108.16–17, 40–42; 110.47–49 Mango). See PMBZ 10833 Anonymus; LBG s.v. τούδουνος. 116 G. Moravcsik, Byzantinoturcica2 (Berlin 1958) II 317–318; P. G. Golden, Khazar Studies. A Historico-Philological Inquiry into the Origins of the Khazars (Buda- pest 1980) I 215–216. 117 Roman-Turkish diplomacy: fr. 10,1–5, 19,1–2. Cf. especially fr. 10,3.93–101 regarding Turkish dignities, specifically that of Tarkhan (Ταρχάν). See Moravcsik, Byzantinoturcica2 II 299–300. 118 Besides acc. pl. σ 18 σκυταλωτοὺς τροχούς (688 above), there are three nominative two-word lemmata. In µ 6 Μελιταῖα κυνίδια (Agath. 5.8.2) and π 19 πτοῖαι παράλογοι (Agath. 5.14.7) the original text is also nominative. In supplementary lemma λ 8 λαιὰ χείρ, however, none of the five instances in Procopius’ Wars is nominative (n.28 above) and the lexicographer evidently standardized the lemma. 119 Dyck 985.31–32 app. crit. and app. font. 692 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 3.9.12 τὸ … ὕπουλον ἐκεῖνο τοῦ τρόπου, 12.9 τὴν εὔνοιαν ... ὕπουλον; Th. Sim. 4.15.9 τοῦ Περσικοῦ δόλου … τὸ ὕπουλον. In any case, υ 7 must cite ὕπουλος from a passage of Menander other than fr. 12,6.75, where it qualifies φιλία. ω 4 ὠρύω (pres. act. 1st sg.): σηµαίνει τὸ κλαίω, “I howl: it means I bewail.”120 A first-person verbal lemma is unusual. Excluding de- monstrable interpolations in alpha, the only parallels are ε 43 ἐνείρω, “I insert,” citing Theophylact’s authorial aside (1.11.8), and ε 19 ἐρεσχηλῶ, which diverges from cited participial ἐρεσχελοῦντα in Procopius (3.12.8), possibly owing to a standardizing intervention, though this entry is otherwise lacunose and corrupt.121 Whether or not the lexicographer thus intervened here, the apparent purpose of the lemma is to register a rare instance of active ὠρύω rather than unremarkable middle ὠρύοµαι.122 If cited precisely, a speech-episode must be assumed (as σ 1 and χ 1), perhaps a Roman emperor’s furious tirade (fr. 9,3.82–85) or a barbarian’s bombast (fr. 19,1.50– 90). Elsewhere Menander employs another exceptionally rare verb in this context: fr. 26,6.5–6 ὑπολοφύρονται τὰς ξυµφοράς, “they be- wail their misfortunes.”123 More generally, he records the Avars’ pre- battle “howling” (fr. 12,3.3 ἀλαλάγµατι). 3b. Quotations µ 2 µετωπηδόν: ‘τὰς φάλαγγας µετωπηδὸν στησάµενος’, “in line: ‘having stationed the formations in line’.” The quotation evidently derives from a battle description. Although adverb µετωπηδόν occurs in canonical histories (e.g. Thuc. 2.90.4; Hdt. 7.100.3; Polyb. passim), use is generally uncommon. In late antiquity, it is found solely but often in Procopius, typically in formulaic phrases with ἵστηµι: 1.13.23 ἵσταντο, 1.14.29 ἔστησαν, 1.18.26 ποιησάµενος, 4.17.4 στή- σας, 8.8.31 ἔστησαν, 8.23.30 στήσαντες, 8.23.34 στήσαντες, 8.31.1 ἔστησαν, 8.35.19 ἔστησαν. By comparison, the lexicographer’s inten- tion in citing Menander instead is not obvious. One might even wonder whether the quotation is an editorial modification of Wars 1.18.26 τὴν φάλαγγα µετωπηδὸν ποιησάµενος. Such a multifaceted intervention, however, would be both unusually intrusive and motiveless, while divergence from typical Procopian diction, in more than one respect, may in fact account for the lexicographer’s interest. 120 Dyck 991.16–19 ap. crit. and app. font. 121 See n.37 above. 122 LSJ9 s.v. ὠρύοµαι. 123 LBG s.v. ὑπολοφύροµαι. PHILIP RANCE 693 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 Quoted middle στησάµενος contrasts with active verbal forms in all instances in Wars; indeed participial στησάµεν- occurs nowhere in Procopius’ oeuvre. Plural τὰς φάλαγγας is also alien to Procopius’ style insofar as he invariably uses single φάλαγξ to denote a battle line in its entirety, rather than, as apparently here, its constituent “formations.” In this scenario, µετωπηδόν entails a particular sense of “aligned” components, rather than, more simply, troops arrayed “in a line, along a front.” φ 5 ‘φριµασσόµενοί τε τῷ ταράχῳ ἀποφυσῶντες’· καὶ ἀλλαχοῦ· ‘ὀπισθό- ρµητα φριµασσόµενοι ἐχώρουν’, “ ‘snorting and puffing in the tur- moil’; and elsewhere: ‘snorting, they came hastening back’.” The missing lemma is obviously φριµασσόµενοι (masc. nom. pres. part. pl. φριµάσσοµαι). Two quoted instances attributable to Menander follow, probably from military contexts. Menander’s use of this rare verb was undoubtedly inspired by its unique occurrence in extant ancient historiography—and in the same form—in Agath. 3.27.4 πνευστιῶντες θαµὰ καὶ φριµασσόµενοι, “with much blowing and snorting,” describing Persian cavalry mounts panicked by a wounded war-elephant. Neither author prefers Atticizing φριµαττ-. The second three-word quote partly coincides with a longer anon- ymous quotation in Suda φ 715 ἡ δὲ ἵππος ὀπισθόρµητα φριµασσοµένη ἐχώρει καὶ ἀδύνατα εἶχεν ἐς τὰ ἄδενδρα ἐπιβῆναι, “the horse, snort- ing, came hastening back and was unable to get up onto the treeless terrain.” The textual parallelism leaves no reasonable doubt that the compiler of the Suda also cites Menander’s Histories, via a now-lost volume of the Constantinian Excerpta.124 The attribution of Suda φ 715 is reinforced by the exceptionally rare construction ἀδύνατα εἶχεν (with neut. pl. adj.), which occurs also in Menander fr. 18,4.7 εἰ δέ γε ἀδύνατα ἔχοιεν, “if this were impossible.”125 Whether ἡ ἵππος denotes an individual mare or collective “horse” in the sense of cavalry remains unclear, though the sole instance in the excerpts of Menander (fr. 2.30 τὴν ἵππον) refers to the herds of the Kutrigurs. In any case, it seems that singular φριµασσοµένη ἐχώρει represents the 124 De Stefani, StIt 18 (1910) 439; Adler, Suidae Lexicon IV (1935) 761.26– 27 (marginal “E” = Excerpta), with V (1938) 37 Addenda et corrigenda. Contra Baldwin, DOP 32 (1978) 125; see n.23 above. Blockley, The History of Menander 297–298, omits the fragment. 125 This construction occurs also in a quotation from an unidentified his- torical source at Suda α 537 ἀδύνατα ἔχοντος τοῦ στρατηγοῦ κύβον ἀναρρίψαι πολέµου τοσοῦτον, “the general being unable to take such a risk in war,” repeated with a slightly different word order at κ 2106. 694 MICROFRAGMENTS OF MENANDER PROTECTOR ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 authentic tradition, while the compiler of the Lexicon Αἱμωδεῖν chose to harmonize the two quotations under a plural lemma φριµασσό- µενοι. χ 1 χιλός: ‘ἐφοίτα χιλοῦ τε ἕνεκα καὶ τῆς λείας’· ἀλλαχοῦ· ‘φύουσι γὰρ καὶ χιλὸν καὶ ἐλύµαν καὶ χόρτον’, “forage: ‘was roaming about for forage and plunder’ [Agath. 1.17.4]; elsewhere: ‘for they bring forth forage and millet and grass/fodder’.” A quotation from Menander supplements one from Agathias.126 As the nom. pl. subject of φύουσι is presumably specified regions or topographical features, a descrip- tion of landscape is again a likely context: see α 145 and π 21. While ἐλύµα is elsewhere unattested, LBG s.v. construes ἐλύµαν as acc. of ἐλύµη (rather than expected ἐλύµην), ostensibly a late form of classical ἔλυµος, though found only in a seventh-century anonymous De cibus.127 Emendation of ἐλύµαν to ἔλυµον is tempting, especially given its similarly precise occurrence in Agath. 2.21.2 (p.68.2 app. crit.) ἀφαιρουµένους ἔλυµόν (VbRO : σέλυµόν L : σέλυµά W) τε καὶ ζειὰς καὶ ἄλλα ἄττα τῶν ἐδωδίµων, “carrying off millet and spelt and other kinds of victuals”; and previously Procop. 1.12.17, 8.13.18. The near-contemporary Strategikon singles out millet as a crop “especially” grown by transdanubian Slavic communities, µάλιστα κέγχρου καὶ ἐλύµου (11.4.14 Dennis). LSJ9 s.v. defines κέγχρος as common millet (Panicum miliaceum) and ἔλυµος (s.v. µελίνη) as Italian millet (Setaria italica), widely known as foxtail millet. Used for human and animal consumption, these coarse-grained, fast-maturing cereals, typically cultivated in harsher environments where other cereals yield poorly, were of military-logistical significance. χ 3 χλιδή: ‘ἐξουσιαστικῇ χλιδῇ ἥκιστα ἔχαιρες’, “luxury: ‘you were not at all pleased by unrestrained luxury’.” The 2nd sg. verb and mor- alizing rhetoric guarantee a speech-episode. Unnoticed by previous scholarship, the headword occurs in Menander fr. 10,3.64 (Exc.leg. I 194.18–19 = Suda χ 340) οὕτω µὲν οὖν τῷ τῶν Τούρκων ἡγεµόνι ἔνεστι χλιδῆς, “such is the luxury available to the leader of the Turks.” This noun is otherwise unattested in the sixth century. Furthermore, the same verbal phrase is found in preceding fr. 10,1.25–26 (Exc.leg. II 450.27) τῷ γεγενηµένῳ ἥκιστα χαίροντες, “not at all pleased by what had happened.” This formula has precedents (e.g. Isoc. Areopag. 50; Jul. Or. 2.28; Greg. Naz. Ep. 75.1), but no contemporary parallels. 126 Keydell ed. xxx, 32.13 app. crit. prefers ἐφοίτων in L to ἐφοίτα in all other codices (WAVaRO) and the indirect tradition (S = Suda). He omits the testimony of Lex.Αἱμ. χ 1. 127 De cibus 14, ed. F. Z. Ermerins, Anecdota medica graeca (Leiden 1840) 261.5. PHILIP RANCE 695 ————— Greek, Roman, and Byzantine Studies 64 (2024) 654–695 ψ 1 ψιθύρισµα: ‘ἡδύ τι {τὸ} ψιθύρισµα τῶν δένδρων ἠκούετο’, “rustling: ‘like something sweet was heard the rustling of the trees’.” Context is wholly obscure. The mystery deepens with the realization that the initial words reprise the well-known opening of Theocritus’ Idyll 1.1– 2 Ἁδύ τι τὸ ψιθύρισµα καὶ ἁ πίτυς … τήνα / … µελίσδεται, “Some- thing sweet is the rustling the pine tree yonder … makes like song.”128 While it may seem fanciful to infer that Menander indulged in a ‘bucolic’ passage, his poetic sensibilities are documented. His lexis otherwise features poetic words. He inserted into his Histories an epigram he had composed (fr. 13,3), which is also transmitted in- dependently (Anth.Pal. 1.101). He reports that Emperor Maurice, who encouraged him to pursue historiography, was especially enthused by history and poetry (fr. 1,1.18–28).129 Emulation of Agathias, as poet-historian, affords a likely literary-cultural setting, even if Menander was never part of his predecessor’s circle or matched his achievements in either sphere.130 August, 2024 Freie Universität Berlin Friedrich-Meinecke-Institut rance@zedat.fu-berlin.de philip.r.rance@gmail.com 128 De Stefani, StIt 18 (1910) 444, “reminiscenza Teocritea (Idyll 1, 1) da attribuire certamente a Menandro”; Dyck 861 n.69, 989.22 app. font. 129 Rance, JÖB 73 (2023) 139–141, with bibliography. 130 I express my deep gratitude to the anonymous peer reviewer for ex- tensive, meticulous, and insightful comments, which significantly improved and enriched this paper.