The Paths of Greek: Literature, Linguistics and Epigraphy 9783110621747, 9783110621082

Citation preview

The Paths of Greek

Trends in Classics – Supplementary Volumes

Edited by Franco Montanari and Antonios Rengakos Associate Editors Stavros Frangoulidis · Fausto Montana · Lara Pagani Serena Perrone · Evina Sistakou · Christos Tsagalis Scientific Committee Alberto Bernabé · Margarethe Billerbeck Claude Calame · Jonas Grethlein · Philip R. Hardie Stephen J. Harrison · Richard Hunter · Christina Kraus Giuseppe Mastromarco · Gregory Nagy Theodore D. Papanghelis · Giusto Picone Tim Whitmarsh · Bernhard Zimmermann

Volume 85

The Paths of Greek Literature, Linguistics and Epigraphy Studies in Honour of Albio Cesare Cassio Edited by Enzo Passa and Olga Tribulato

ISBN 978-3-11-062108-2 e-ISBN (PDF) 978-3-11-062174-7 e-ISBN (EPUB) 978-3-11-062180-8 ISSN 1868-4785 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2019 Walter de Gruyter GmbH, Berlin/Boston Editorial Office: Alessia Ferreccio and Katerina Zianna Logo: Christopher Schneider, Laufen Printing and binding: CPI books GmbH, Leck www.degruyter.com

Contents Acknowledgments  VII General Abbreviations  IX Signs and Symbols  XI Bibliographical Abbreviations  XIII Enzo Passa and Olga Tribulato Introduction  1 Bibliography of Albio Cesare Cassio  9 Sara Kaczko Unconventional Features in Homer: The Case of ἑέ and ἑοῖ  19 Andreas Willi σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  43 Lucien van Beek and Laura Migliori Active versus Middle Perfect in Homeric Greek: Synchrony and Diachrony  71 Enzo Passa Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry in the Classical Age  107 Lucia Prauscello Of Land, Ancestral Property and Prophecy in Corinna PMG 654 col. iii ll. 37–39  125 Federico Favi Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  149 Alessandro Vatri Early Dactylic Prose in the History of Greek Prose Rhythm  175

VI  Contents Matilde Serangeli Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage: Gk. PN Τέμπυρα, Hitt. dampu- ‘Blunt’, Old Russ. tupъ ‘Blunt, Stupid’  197 Enrico Cerroni Prose and Poetry of Pain: A History of the Term ἄλγος  219 Olga Tribulato Making the Case for a Linguistic Investigation of Greek Lexicography: Some Examples from the Byzantine Reception of Atticist Lemmas  241 List of Contributors  271 Index of Notable Words  273 Index of Subjects  275 Index of Passages  279

Acknowledgments We wish to acknowledge the financial support of the Department of Humanities at Ca’ Foscari University of Venice, which has funded both the conference in honour of Albio Cesare Cassio (Venice, 28th April 2017) and the language revision of this volume. For the latter, we thank Katharine Shields for her precious assistance. Ever since the idea of celebrating Albio’s career materialized in late 2016, Paola Ascheri has been an enthusiastic supporter and secret ally throughout. She has been especially invaluable for the production of the Bibliography of Albio’s works, for which we are very grateful to her. We wish to thank all the contributors for having accepted our invitation to celebrate Albio, and particularly Lucia Prauscello and Andreas Willi for their support at several stages of this enterprise. We are also indebted to Coulter George for his advice, to Margherita Bellocchi, Caterina Carpinato, Ettore Cingano and Filippomaria Pontani for chairing the conference sessions, as well as to all of those among Albio’s friends and pupils who came to Venice and made the conference an unforgettable event for him. Finally, we are grateful to Franco Montanari for suggesting that we publish this Festschrift in the Trends in Classics Supplementary Volumes, and to Antonios Rengakos for accepting it in the series. Festschriften are often difficult volumes to produce and bring to completion. If this one has turned into a rewarding experience on many levels it is because of the profound friendship between the two editors. For this, too, we are deeply in debt of Albio Cassio.

https://doi.org/10.1515/9783110621747-202

General Abbreviations abstr. acc. act. adj. adv. Aeol. Akk. aor. Att.-Ion. AV Avest. Balt. Boeot. Br. Bret. Bulg. caus. cent. Class. cod. Comm. Slav. dat. denom. Dor. Eng. fut. gen. Ger. Gk. Goth. Hitt. Ice. IE imperat. impf. Ind. intr. Ir. It. Lat. Latv. Lith.

abstract accusative active adjective adverb Aeolic Akkadian aorist Attic-Ionic Atharvaveda Avestan Baltic Boeotian Brahmana Breton Bulgarian causative century Classical codex Common Slavic dative denominative Doric English future genitive German Greek Gothic Hittite Icelandic Indo-European imperative imperfect Indian intransitive Irish Italian Latin Latvian Lithuanian

https://doi.org/10.1515/9783110621747-203

loc. Luw. Lyc. masc. med. MS / MSS Myc. nom. ntr. OCS OH OI OInd. perf. pers. PIE pl. plupf. PN Pol. pres. ptcpl. Russ. RV schol. sing. Skt. s.l. subj. subst. suff. Sum. SV s.v. Tarent. Thess. TS Umbr. Ved. v.l. / vv.ll.

locative Luwian Lycian masculine middle manuscript, manuscripts Mycenaean nominative neuter Old Church Slavonic Old Hittite Old Irish Old Indian perfect person Proto-Indo-European plural pluperfect personal name Polish present participle Russian Ṛgveda scholium / scholia singular Sanskrit supra lineam subjunctive substantive suffix Sumerian Sāmaveda sub voce Tarentine Thessalian Taittirīya Saṃhita Umbrian Vedic varia lectio / variae lectiones

Signs and Symbols []

enclose phonetic transcription

//

enclose phonemic transcription

enclose graphemes

>

yields

→

indicates semantic or derivational changes

l adjustmentsin addition to direct sound >→ indicates morphological/analogica change

*

reconstructed

x

indicates number of occurrences of a certain word/feature

~

varies with

::

indicates that two terms/expressions are parts of the same morphological system

°

indicates the second member of a compound

https://doi.org/10.1515/9783110621747-204

Bibliographical Abbreviations Abbreviated titles of Classics journals and works quoted in the individual Bibliographies and not included in this list follow the Année Philologique. AD Adler AHW

Aland et al. ALEW Alpers ANRW ArchEph Arrighetti Avezzù Baladié BCH Bekker (Ap. Soph.) Bekker (Lex. rhet.) Bekker (Theoph. Cont.) Bernabé Bethe Billerbeck Böhlig Boissevain Breysig Büttner-Wobst CAD 7 I/J Carey CEG

Ἀρχαιολογικὸν Δελτίον, Αθήναι 1915–. A. Adler, Suidae Lexicon, 5 vols., Leipzig 1928–1938. W. von Soden, Akkadisches Handwörterbuch: Unter Benutzung des lexikalischen Nachlasses von Bruno Meissner (1868–1947), vol. 2: M-S, Wiesbaden 1972. K. Aland/M. Black/C.M. Martini/B.M. Metzger/A. Wikgren (eds.), The Greek New Testament, 2nd edn., Stuttgart 1968. W. Hock/E.-J. Bukevičiūtė/C. Schiller, Altlitauisches etymologisches Wörterbuch, vol. 2: N–Ž, Hamburg 2015. K. Alpers, Das attizistische Lexikon des Oros: Untersuchung und kritische Ausgabe der Fragmente, Berlin/New York 1981. H. Temporini/W. Haase (eds.), Aufstieg und Niedergang der römischen Welt, Berlin 1972–. Αρχαιολογική εφημερίς, Αθήναι 1910–. G. Arrighetti, Epicuro: Opere, 2nd edn., Torino 1973. G. Avezzù, Alcidamante: Orazioni e frammenti, Roma 1982. R. Baladié, Strabon, vol. 4: Livre 7, Paris 1989. Bulletin de correspondance hellénique, Paris 1877–. I. Bekker, Apollonii Sophistae Lexicon Homericum, Berolini 1833. I. Bekker, Lexeis rhetorikai, in: Anecdota Graeca, vol. 1, Berlin 1814, 197– 318. I. Bekker, Theophanes Continuatus, Ioannes Cameniata, Symeon Magister, Georgius Monachus, Bonnae 1838. A. Bernabé, Poetarum epicorum Graecorum testimonia et fragmenta, vol. 1, 2nd edn., Stutgardiae/Lipsiae 1996. E. Bethe, Pollucis Onomasticon, 3 vols., Leipzig 1900–1931. M. Billerbeck, Stephani Byzantii Εthnica, Berolini 2006. G. Böhlig, Ioannis Caminiatae De expugnatione Thessalonicae, Berolini/Novi Eboraci 1973. U.P. Boissevain, Cassii Dionis Cocceiani Historiarum Romanarum quae supersunt, Berolini 1895. A. Breysig, Germanici Caesaris Aratea cum scholiis, Berlin 1867. T. Büttner-Wobst, Polybii Historiae, 4 vols., Lipsiae 1893–1905. I.J. Gelb et al., The Assyrian Dictionary of the Oriental Institute of the University of Chicago, vol. 7: I/J, Chicago 1960. C. Carey, Lysiae Orationes cum fragmentis, Oxonii 2007. P.A. Hansen, Carmina epigraphica Graeca saeculorum VIII–V a. Chr. n., Berolini/Novi Eboraci 1983.

https://doi.org/10.1515/9783110621747-205

XIV  Bibliographical Abbreviations CEG2 CLOD Consbruch Crugnola CTH Cunningham Davies (EGF) Davies (PMGF) DDBDP de Borries DELG Dennis de Stefani Dettori de Velsen DGE DIG Dilts D.-K. DNP Drachmann Dyck EDG EDH EDPG Egenolff Erbse (Ael. Dion.) Erbse (schol.)

P.A. Hansen, Carmina epigraphica Graeca saeculi IV a. Chr. n., Berolini/ Novi Eboraci 1989. S. Dakares/I. Bokotopoulou/A.P. Christides, Τα χρηστήρια ελάσματα της Δωδώνης των ανασκαφών Δ. Ευαγγελίδη, 2 vols., Αθήναι 2013. M. Consbruch, Hephaestionis Enchiridion cum commentariis veteribus: Accedunt variae metricorum Graecorum reliquiae, Lipsiae 1906. A. Crugnola, Scholia in Nicandri Theriaka cum glossis, Milano 1971. E. Laroche, Catalogue des textes hittites, Paris 1971 (with Supplement in: Revue hittite et asianique 30, 1972, 94–133; ibid. 32, 1974, 68–71). I.C. Cunningham, Synagoge: Texts of the Original Version and of Manuscript B, Berlin/New York 2003. M. Davies, Epicorum Graecorum fragmenta, Göttingen 1988. M. Davies, Poetarum melicorum Graecorum fragmenta, vol. 1: Alcman, Stesichorus, Ibycus, Oxford 1991. The Duke Databank of Documentary Papyri (accessible at http://papyri. info/browse/ddbdp/). J. de Borries, Phrynichi Sophistae Praeparatio sophistica, Lipsiae 1911. P. Chantraine, Dictionnaire étymologique de la langue grecque: Histoire de mots, 2 vols., Paris 1968–1980. G.T. Dennis, Michaelis Pselli Orationes panegyricae, Stutgardiae/Lipsiae 1994. E.L. de Stefani, Etymologicum Gudianum quod vocatur, Leipzig 1909. E. Dettori, Filita grammatico: Testimonianze e frammenti, Roma 2000. A. de Velsen, Tryphonis Alexandrini grammatici fragmenta, Berolini 1853. E. Schwyzer, Dialectorum Graecarum exempla epigraphica potiora, Leipzig 1923. P. Cauer, Delectus inscriptionum Graecarum propter dialectum memorabilium, Lipsiae 1883. M.R. Dilts, Demosthenis Orationes, 4 vols., Oxford, 2002–2009. H. Diels/W. Kranz, Die Fragmente der Vorsokratiker, 3 vols., 6th edn., Berlin 1951–1952. H. Cancik/H. Schneider (eds.), Der Neue Pauly: Enzyklopädie der Antike, 16 vols., Stuttgart 1996–2003. A.B. Drachmann, Scholia vetera in Pindari carmina, 3 vols., Lipsiae 1903–1927. A.R. Dyck, Epimerismi Homerici, 2 vols., Berlin/New York 1983–1995. R. Beekes, Etymological Dictionary of Greek, 2 vols., Leiden/Boston 2010. A. Kloekhorst, Etymological Dictionary of the Hittite Inherited Vocabulary with Analysis Of Synchronic Textological Data, 2 vols., Leiden 2008. G.J. Kroonen, Etymological Dictionary of Proto-Germanic, Leiden 2013. P. Egenolff, “Aelius Herodianus: Schematismi Homerici”, in: Jahrbücher für klassische Philologie 149, 1894, 338–345. H. Erbse, Untersuchungen zu den attizistischen Lexica, Berlin 1950. H. Erbse, Scholia Graeca in Homeri Iliadem (scholia vetera), 7 vols., Berlin 1969–1988.

Bibliographical Abbreviations  XV

FD III.2 Featherstone/ Codoñer Festa FGrHist Fischer Fortenbaugh Garzya Gautier GEW GG G.-P. Hansen Hase HEG H&R

Heylbut Hittite Laws Hörandner Hordern Hude (Hdt.) Hude (schol. Thuc.) HW IC IEleusis

IG IV IG IV²,1

M.G. Colin, Fouilles de Delphes, vol. 3.2: Épigraphie: Inscriptions du Trésor des Athéniens, Paris 1909–1913. M. Featherstone/J.S. Codoñer, Chronographiae quae Theophanis Continuati nomine fertur: Libri I−IV, Berlin 2015. N. Festa, Palaephati Peri apiston. Heracliti qui fertur Peri apiston. Excerpta Vaticana (vulgo Anonymus De incredibilibus), Lipsiae 1902. F. Jacoby, Die Fragmente der griechischen Historiker, 16 vols., Berlin/ Leiden 1923–1954. E. Fischer, Die Ekloge des Phrynichos, Berlin/New York 1974. W. Fortenbaugh et al., Theophrastus of Eresus: Sources for His Life, Writings, Thought, and Influence, 2 vols., Leiden 2016. A. Garzya, Nicephori Basilacae Orationes et Epistolae, Leipzig 1984. P. Gautier, Michel Italikos: Lettres et Discours, Paris 1972. H. Frisk, Griechisches etymologisches Wörterbuch, 2 vols., Heidelberg 1973. Grammatici Graeci. See under individual editors (Lentz, Schneider, Uhlig). B. Gentili/C. Prato, Poetae elegiaci, 2 vols., 2nd edn., Leipzig 1988. D.U. Hansen, Das attizistische Lexikon des Moeris: Quellenkritische Untersuchung und Edition, Berlin/New York 1998. C.B. Hase, Leoni Diaconi Caloënsis Historiae libri decem, Bonnae 1828. J. Tischler, Hethitisches etymologisches Glossar, Innsbruck 1983–2016. E. Hatch/H. Redpath, A Concordance to the Septuagint and the other Greek Versions of the Old Testament (Including the Apocryphal Books), Oxford 1897. H. Heylbut, “Ptolemaeus Περὶ διαφορᾶς λέξεων”, in: Hermes 22, 1887, 388–410. H.A. Hoffner, The Laws of the Hittites: A Critical Edition, Leiden/New York 1997. W. Hörandner, “Theodoros Prodromos: Historische Gedichte”, in: Wiener byzantinistische Studien 11, 1974, 177–553. J.H. Hordern, Sophron’s Mimes: Text, Translation, and Commentary, Oxford 2004. K. Hude, Herodoti Historiae, 2 vols., Oxonii 1908. K. Hude, Scholia in Thucydidem ad optimos codices collata, Lipsiae 1927. J. Friedrich/A. Kammenhuber, Hethitisches Wörterbuch, 5 vols., 2nd edn., Heidelberg 1984–. M. Guarducci, Inscriptiones Creticae opera et consilio Friderici Halbherr collectae, 4 vols., Roma 1935–1950. K. Clinton, Eleusis: The Inscriptions on Stone. Documents of the Sanctuary of the Two Goddesses and Public Documents of the Deme, 2 vols., Athens 2005. M. Fraenkel, Inscriptiones Graecae, vol. 4: Inscriptiones Aeginae, Pityonesi, Cecryphaliae, Argolidis, Berlin 1902. F. Hiller von Gaertringen, Inscriptiones Graecae, vol. 4.1: Inscriptiones Argolidis: Inscriptiones Epidauri, 2nd edn., Berlin 1929.

XVI  Bibliographical Abbreviations IG VII IG IX 12,3 IG IX,2 IG XII Suppl. IG XIV

IGDGG IGDOP IGDS I IK.Priene IOropos IThesp

Jäkel Jensen Jordan/ Kotansky K.-A. Kalamakis Kannicht KBo Keil KFSA Körte Kouremenos/ Parassoglou/ Tsantsanoglou KPN Kreuttner KUB

W. Dittenberger, Inscriptiones Graecae, vol. 7: Inscriptiones Megaridis, Oropiae, Boeotiae, Berlin 1892. G. Klaffenbach, Inscriptiones Graecae, vol. 9.1.3: Inscriptiones Locridis occidentalis, 2nd edn., Berlin 1968. O. Kern, Inscriptiones Graecae, vol. 9.2: Inscriptiones Thessaliae, Berlin 1908. F. Hiller von Gaertringen, Inscriptiones Graecae, vol. 12: Supplementum, Berlin 1939. G. Kaibel, Inscriptiones Graecae, vol. 14: Inscriptiones Siciliae et Italiae, additis Galliae, Hispaniae, Britanniae, Germaniae inscriptionibus, Berlin 1890. L. Dubois, Inscriptions grecques dialectales de Grande Grèce, vol. 2: Colonies achéennes, Genève 2002. L. Dubois, Inscriptions grecques dialectales d’Olbia du Pont, Genève 1996. L. Dubois, Inscriptions grecques dialectales de Sicile: Contribution à l’étude du vocabulaire grec colonial, vol. 1, Rome 1989. W. Blümel/R. Merkerlbach, Die Inschriften von Priene, 2 vols., 2nd edn. (Inschriften griechischer Städte aus Kleinasien, vol. 69.1–2), Bonn 2014. V. Petrakos, Οι επιγραφές του Ωρωπού, Αθήναι 1997. P. Roesch, Les inscriptions de Thespies, Lyon 2007 (accessible at https://www.hisoma.mom.fr/production-scientifique/les-inscriptions-dethespies). S. Jäkel, Menandri sententiae, Leipzig 1964. C. Jensen, Hyperidis Orationes sex cum ceterarum fragmentis, Lipsiae 1917. D.R. Jordan/R.D. Kotansky, “Ritual Hexameters in the Getty Museum: Preliminary Edition”, in: ZPE 178, 2011, 54–62. R. Kassel/C. Austin, Poetae comici Graeci, 8 vols., Berlin/New York 1983–2001. D.C. Kalamakis, Λεξικά των Επών Γρηγορίου του Θεολόγου μετά γενικής θεωρήσεως της πατερικής λεξικογραφίας, Αθήναι 1992. R. Kannicht, Tragicorum Graecorum fragmenta, vol. 5: Euripides, Göttingen 2004. Keilschrifttexte aus Boghazkoy, Osnabrück/Berlin 1923–. B. Keil, Aelii Aristidis Smyrnaei quae supersunt omnia, vol. 2, 2nd edn., Berolini 1958. E. Behler et al., F. Schlegel: Werke. Kritische Ausgabe, vol. 16: Fragmente zur Poesie und Literatur, vol. 1, Padeborn/München/Wien 1981. A. Körte, Menandri quae supersunt, 2 vols., Lipsiae 1953–1957. T. Kouremenos/G.M. Parassoglou/K. Tsantsanoglou, The Derveni Papyrus, Firenze 2006. L. Zgusta, Kleinasiatische Personennamen, Prag 1964. X. Kreuttner, Andronici qui fertur libelli περὶ παθῶν, vol. 1: De affectibus, Heidelbergae 1884. Keilschrifturkunden aus Boghazkoy, Berlin 1921–1990.

Bibliographical Abbreviations  XVII

Kühn Lasserre/ Livadaras Latte Latte/ Cunningham Latte/Hansen Latte/ Hansen/ Cunningham Lentz Leurini LfgrE LIV LKN L.-P. LSJ Macleod Matthews Metzler Migeotte Migne Müller (Dionys. Per.) Müller (FHG) M.-W. Nauck NETS

Nickau (Ammon.) Nickau (Symeon) Patillon Patillon/ Brisson

C.G. Kühn, Galeni opera omnia, 20 vols., Leipzig 1821–1833. F. Lasserre/N. Livadaras, Etymologicum magnum genuinum. Symeonis Etymologicum una cum Magna grammatica. Etymologicum magnum auctum, 2 vols., Roma 1976–1992. K. Latte, Hesychii Alexandrini Lexicon, 2 vols., Copenhagen 1953–1966. K. Latte/I.C. Cunningham, Hesychii Alexandrini Lexicon, vol. 1: Α–Δ, Berlin/Boston 1953–2018. K. Latte/P.A. Hansen, Hesychii Alexandrini Lexicon, vol. 3: Π–Σ, Berlin/ New York 1953–2005. K. Latte/P.A. Hansen/I.C. Cunningham, Hesychii Alexandrini Lexicon, vol. 4: Τ–Ω, Berlin/New York 1953–2009. A. Lentz, Herodiani technici reliquiae (GG, vols. 3.1–3.2), Leipzig 1867– 1870. L. Leurini, Ionis Chii testimonia et fragmenta, Amsterdam 2000. Lexikon des frühgriechischen Epos, 4 vols., Göttingen 1955–2010. H. Rix et al., Lexikon der indogermanischen Verben: Die Wurzeln und ihre Primärstammbildungen, 2nd edn., Wiesbaden 2001. Λεξικό τής κοινής νεοελληνικής, Θεσσαλονίκη 1998. E. Lobel/D.L. Page, Poetarum Lesbiorum fragmenta, Oxford 1955. H.G. Liddell/R. Scott/H.S. Jones, A Greek-English Lexicon, 9th edn., Oxford 1996. M.D. Macleod, Luciani opera, 4 vols., Oxford 1972–1987. V.J. Matthews, Panyassis of Halikarnassos: Text and Commentary, Lugduni Batavorum 1974. K. Metzler, Eustathii Thessalonicensis De emendanda vita monachica, Berolini/Novi Eboraci 2006. L. Migeotte, L’emprunt public dans les cités grecques: Recueil des documents et analyse critique, Paris 1984. J.-P. Migne, Patrologiae cursus completus (series Graeca), Paris 1857– 1866. K. Müller, Scholia in Dionysii Periegetae orbis descriptionem (Geographi Graeci minores, vol. 2), Paris 1861. K. Müller, Fragmenta historicorum Graecorum, 5 vols., Parisiis 1841– 1872. R. Merkelbach/M.L. West, Fragmenta Hesiodea, Oxford 1967. A. Nauck, Lexicon Vindobonense, Petropolis 1867. A. Pietersma/B.G. Wright (eds.), A New English Translation of the Septuagint and the Other Greek Translations Traditionally Included under That Title, Oxford/New York 2007. K. Nickau, Ammonius: De adfinium vocabulorum differentia, Lipsiae 1966. K. Nickau, “Neues zur Überlieferung der Synagoge des Symeon (Paris. gr. 2552, Paris. Suppl. gr. 1238, Vatic. gr. 2226)”, in: GFA 3, 2000, 13–31. M. Patillon, Aelius Theon: Progymnasmata, Paris 1997. M. Patillon/L. Brisson, Longin: Fragments, Art rhétorique. Rufus: Art rhétorique, Paris 2001.

XVIII  Bibliographical Abbreviations P.Berol. 13284 Pfeiffer P.Grenf. 2 PHI Pingree

PL P.Lond. PMG Poltera P.Oxy. P.Sorb. 2 P.Strasb.

P.Würzb. RA Rabe (Herm.) Rabe (schol. Luc.) Radt (Aesch.) Radt (Soph.) RE REA REG Retguit Ritschl Roberto Rose Sakkelion Schneider SEG

U. von Wilamowitz-Moellendorff, “§ 14. Korinna”, in: Berliner Klassikertexte, vol. 5.2, Berlin 1907, 19–55. R. Pfeiffer, Callimachus, vol. 1: Fragmenta, Oxonii 1949. B.P. Grenfell/A.S. Hunt, New Classical Fragments and Other Greek and Latin Papyri, Oxford 1897. Searchable Greek Inscriptions (accessible at https://inscriptions.pack hum.org) D. Pingree, Dorothei Sidonii Carmen astrologicum: Interpretationem Arabicam in linguam Anglicam versam una cum Dorothei fragmentis et Graecis et Latinis, Leipzig 1976. A. Mauersberger et al., Polybios-Lexikon im Auftrage der Deutschen Akademie der Wissenschaften zu Berlin, Berlin 1956–2004. F.G. Keyon et al., Greek Papyri in the British Museum, London 1893–. D.L. Page, Poetae melici Graeci, Oxford 1962. O. Poltera, Simonides Lyricus: Testimonia und Fragmente. Einleitung, kritische Ausgabe, Übersetzung und Kommentar, Basel 2008. B.P. Grenfell/A.S. Hunt et al., Oxyrhynchus Papyri, London 1898–. J. Gascou, Un codex fiscal Hermopolite (P.Sorb. II 69), Atlanta 1994. A. Martin/O. Primavesi, L’Empédocle de Strasbourg (P. Strasb. gr. Inv. 1665–1666): Introduction, édition et commentaire, Berlin/New York 1999. U. Wilcken, Mitteilungen aus der Würzburger Papyrussammlung, Berlin 1934. Revue archéologique, Paris 1844–. H. Rabe, Hermogenis opera, Stutgardiae 1913. H. Rabe, Scholia in Lucianum, Lipsiae 1906. S. Radt, Tragicorum Graecorum fragmenta, vol. 3: Aeschylus, Göttingen 1985. S. Radt, Tragicorum Graecorum fragmenta, vol. 4: Sophocles, Göttingen 1977. G. Wissowa/W. Kroll et al. (eds.), Paulys Realencyclopädie der classischen Altertumswissenschaft, Stuttgart 1893–1980. Revue des études anciennes, Pessac 1899–. Revue des études grecques, Paris 1888–. R.F. Retguit, Scholia in Aristophanem: Scholia in Thesmophoriazusas, Ranas, Ecclesiazusas et Plutum, Groningen 2007. F. Ritschl, Thomae Magistri sive Theoduli monachi Ecloga vocum Atticarum, Halis Saxonum 1832. U. Roberto, Ioannis Antiocheni fragmenta ex Historia chronica, Berlin/ New York 2005. V. Rose, Aristotelis qui ferebantur librorum fragmenta, Lipsiae 1886. I. Sakkelion, Lexicon Patmense (Lexica Graeca minora, 140–165), in: BCH 1, 1877, 10–16; 137–155. R. Schneider, Apollonii Dyscoli quae supersunt: Apollonii scripta minora (GG, vol. 2.1), Lipsiae 1878. Supplementum epigraphicum Graecum, Leiden 1923–.

Bibliographical Abbreviations  XIX

Ševčenko SH S.-M. Stallbaum

Sturz Theodoridis (Philox.) Theodoridis (Phot.) ThGL Tittmann TL TLG TM

Uhlig Usener/ Radermacher Valente Valerio van den Gheyn van der Valk

van Dieten (Hist.) van Dieten (Orat.) van Thiel Voigt von Arnim Wendel

I. Ševčenko, Chronographiae quae Theophanis Continuati nomine fertur: Liber quo Vita Basilii imperatoris amplectitur, Berlin/Boston 2011. H. Lloyd-Jones/P. Parsons, Supplementum Hellenisticum, Berlin/ New York 1983. B. Snell/H. Maehler, Pindari carmina cum fragmentis, 2 vols., Leipzig 1975–1980. G. Stallbaum, Eustathii archiepiscopi Thessalonicensis commentarii ad Homeri Odysseam ad fidem exempli Romani editi, 2 vols., Lipsiae 1825– 1826. F.W. Sturz, Orionis Thebani Etymologicon, Leipzig 1820. C. Theodoridis, Die Fragmente des Grammatikers Philoxenos, Berlin/ New York 1976. C. Theodoridis, Photii Patriarchae Lexicon, 3 vols., Berlin/New York/ Boston 1982−2013. C.B. Hase, Thesaurus Graecae linguae ab Henrico Stephano constructus, Oxonii 1829. J.A.H. Tittmann, Iohannis Zonarae Lexicon ex tribus codicibus manuscriptis, 2 vols., Leipzig 1808. E. Kalinka, Tituli Asiae Minoris, vol. 1: Tituli Lyciae lingua Lycia conscripti, Wien 1901. Thesaurus linguae Graecae: A Digital Library of Greek Literature (accessible at http://stephanus.tlg.uci.edu/). Trismegistos: An Interdisciplinary Portal of Papyrological and Epigraphical Resources. Texts (accessible at www.trismegistos.org/tm/index. php). G. Uhlig, Apollonii Dyscoli quae supersunt: Apollonii Dyscoli De constructione libri quattuor (GG, vols. 2.2–2.3), Lipsiae 1910. H. Usener/L. Radermacher, Dionysii Halicarnassei Opuscula, 2 vols., Leipzig 1899–1933. S. Valente, The Antiatticist: Introduction and Critical Edition, Berlin/ Boston 2015. F. Valerio, Ione di Chio: Frammenti elegiaci e melici, Bologna 2013. J. van den Gheyn, “Acta Graeca SS. Davidis, Symeonis et Georgii, Mitylenae in insula Lesbo”, in: Analecta Bollandiana 18, 1899, 209−259. M. van der Valk, Eustathii archiepiscopi Thessalonicensis Commentarii ad Homeri Iliadem pertinentes ad fidem codicis Laurentiani, 4 vols., Leiden 1971–1987. J.-L. van Dieten (ed.), Nicetae Choniatae Historia, Berolini/Novi Eboraci 1975. J.-L. van Dieten (ed.), Nicetae Choniatae Orationes et Epistulae, Berlin 1972. H. van Thiel, Scholia D in Iliadem: Proecdosis aucta et correctior, Köln 2014 (accessible at http://kups.ub.uni-koeln.de/id/eprint/5586). E.-M. Voigt, Sappho et Alcaeus: Fragmenta, Amsterdam 1971. H. von Arnim, Stoicorum veterum fragmenta, 3 vols., Stutgardiae 1903– 1905. K. Wendel, Scholia in Theocritum vetera, Leipzig 1914.

XX  Bibliographical Abbreviations West (IEG) West (Il.) WinningtonIngram Wolska-Conus Wyss

M.L. West, Iambi et elegi Graeci ante Alexandrum cantati, 2 vols., 2nd edn., Oxford 1989–1992. M.L. West, Homeri Iias, 2 vols., Stutgardiae/Monachii/Lipsiae 1998– 2000. R.P. Winnington-Ingram, Aristidis Quintiliani De musica libri tres, Lipsiae 1963. W. Wolska-Conus, Cosmas Indicopleustès: Topographie Chrétienne, 2 vols., Paris 1968–1970. B. Wyss, “Zu Gregor von Nazianz”, in: O. Gigon/K. Meuli/W. Theiler/ F. Wehrli/B. Wyss (eds.), Phyllobolia für Peter von der Mühll zum 60. Geburtstag am 1. August 1945, Basel 1946, 153–183.

Enzo Passa and Olga Tribulato

Introduction

 A Homage to Albio Cesare Cassio This volume collects papers which were mostly given at a conference held at the Ca’ Foscari University of Venice in April 2017 to celebrate the career of Albio Cesare Cassio. The contributors include not only some of those who, as students, regularly attended ACC’s courses at the Sapienza University of Rome, but also scholars who, though not his direct pupils, have significantly benefited from his teaching. With this collection we would all like to thank ACC for what he has meant for each and every one of us; this Introduction summarizes the contents of the volume in the light of the approach to Ancient Greek which ACC has mastered. It is impossible to outline ACC’s scholarly profile and personality in just a few words. Colleagues, pupils, and friends can bear witness to different aspects of his persona, and a lot could be written about his extraordinary knowledge of Ancient Greek, intellectual sharpness, and passion for his work. All of these qualities are well-known to anyone who had the opportunity to meet him over the years. In what follows, we shall limit ourselves to briefly highlighting some key points of the common legacy shared by all those who owe ACC a significant part of their education. Perhaps the most valuable lesson we have learnt from ACC is the readiness to venture beyond the borders of our discipline and to discover what ‘lies outside’ them, in order to then return to our own field with increased knowledge and open-mindedness. The title of this volume reflects this principle: there is more than just one path to the ancient Greek language, and if we, as philologists, really wish to gain a deeper insight into our discipline, we need to look into what other disciplines can teach us. Linguistics, epigraphy, archaeology and other fields are not faraway provinces of a collapsing empire, but essential limbs of a complex organism which can only be brought back to life by allowing them to interact. Another hallmark of ACC’s approach to Classical philology is his mistrust of theories and abstractions, whenever these stray too far from concrete contexts. Given ACC’s interests in linguistics, this might seem like a paradox, since IndoEuropeanists are frequently forced to work with theories and reconstructions. ACC’s warning against approaches which are too abstract must obviously be taken cum grano salis and has always been directed only against those self-referential solutions which tend to overlook the social and cultural background of a given linguistic feature. For his part, ACC has championed an approach firmly https://doi.org/10.1515/9783110621747-001

2  Enzo Passa and Olga Tribulato anchored in the historical development of Greek society, and in the idea that no linguistic reconstruction can be proven correct if it does not agree with the culture underpinning the text which the reconstruction wishes to elucidate. At the same time, ACC has also shown us how to handle details. Philologists often run the risk of being overwhelmed by too many data and of losing the thread which should guide them forward. To the useless multiplication of useless details, ACC has opposed his natural taste for simplicity. As he has reminded us, distinguishing useful details from useless ones is an essential premise for every philologist hoping to reach an effective understanding of the issues at stake.1 We could continue with the list of ACC’s teachings at length, but we know that it would not please him, since he generally favours understatement. Therefore, all we would like to say is that this volume is a small token of gratitude from all of us, for the knowledge that Albio has shared with us and for the guidance he has provided in these matters.

 Outline of Albio Cesare Cassio’s Work ACC was Professor of Greek Linguistics at the University of Naples L’Orientale from 1980 to 1997. In 1997 he was appointed Professor of Greek and Latin Linguistics at the Sapienza University of Rome, a post which he held until his retirement in 2017. In 1983–1984 he was Fellow of the Centre for Hellenic Studies at Washington D.C. Throughout his career, he organized many conferences, the most important of which was perhaps the 3rd International Colloquium of Greek Dialectology (Ischia, September 1996). He has given invited lectures at several Italian and foreign universities, including the Gray Lectures at the Faculty of Classics,

 1 “Molti di voi conosceranno sicuramente una frase famosa con due varianti, ‘Dio è nel dettaglio’ oppure ‘il diavolo è nel dettaglio’, frase che è variamente attribuita a grandi personaggi del passato. Nella mia esperienza ho sempre verificato che coglie completamente nel segno, e nei miei lunghi anni di insegnamento, qui e altrove, ho soprattutto cercato di far capire la differenza tra i dettagli stupidi, che servono solo a gonfiare gli articoli e i libri, e i dettagli importanti, che possono illuminare delle situazioni o cambiare radicalmente i termini di un problema” (A.C. Cassio, “Recentior, non deterior: uno sguardo alla lingua di Iliade 24”, see item no. 137 in the Bibliography of Albio Cesare Cassio’s works; text of a paper delivered at the conference Il riscatto di Ettore. Giornata di studi internazionale sul libro 24° dell’Iliade, held on 13th October 2017 at Sapienza University of Rome).

Introduction  3

Cambridge University (May 2006), a Lecture and a Block Seminar at UCLA (Program in Indo-European Studies, April 2016) and the Townsend Lectures at Cornell University (April 2018). ACC’s first research interests developed under the sign of textual criticism, and the papers written by him between 1971 and 1976 range from Greek to Latin and Humanist philology. The year 1977, when he published his edition of Aristophanes’ Banqueters, marked a turning point in his long-lasting engagement with Attic Old Comedy, an interest which later culminated in his 1985 volume Commedia e partecipazione: La Pace di Aristofane. In the mid-1980s ACC began to explore new paths. He turned to other branches of Greek comedy by investigating Epicharmus, an author who has remained a focal point in his research.2 He also started developing an interest in Greek dialectology, to which he has devoted several seminal contributions.3 The study of the Greek dialects fed into another characteristic interest of his: inscriptions, especially those from Southern Italy and Sicily.4 By the beginning of the 1990s, most of ACC’s interests had acquired their definitive shape. Works from this period display an increasingly multidisciplinary and unique approach combining philology, linguistics, literature and epigraphy. He devoted seminal studies to epic and lyric poetry of the Archaic Age, and to their relationship with the epigraphic evidence.5 Since it is impossible to account for all of the fundamental contributions produced from this period onwards, we offer here a basic introduction according to research topics. Starting with epic, it is not far-fetched to say that ACC has investigated almost every facet of the Homeric language: from the much-debated influence of the Euboean dialect on the making of the Homeric poems6 and the role played by other dialects,7 to the textual fixation of the poems; from the interaction between oral performances and written copies8 to the layout of the first ancient editions of Homer and their sustaining exegesis.9 In some of these works ACC dismisses the

 2 On Epicharmus, see item no. 42. On Doric comedy, see no. 89. 3 See e.g. no. 37. 4 See for instance Cassio’s papers devoted to the ‘Hipponion lamella’: nos. 45, 48, 71, and 72. 5 Especially representative in this respect are no. 65, which is devoted to the relationship between Homer’s text and an important inscription from Cuma, and no. 69, which compares the epigraphic epigrammatic tradition with literary models in the light of the crucial testimony of the inscription on the Ambracia polyandrion. 6 See nos. 70 and 79. 7 See no. 95. 8 Nos. 80, 90, 92, and 107. 9 Nos. 76, 86, and 111.

4  Enzo Passa and Olga Tribulato hypothesis of an ancient ‘Euboean phase’ as unlikely, highlighting instead the fundamental role played by the Aeolic dialects in shaping the epic language. He also addresses the difficulties raised by Parry and Lord’s theory of an oral dictated poem, and makes a strong case for the existence of a form of Homeric textual criticism before the Hellenistic Age. In more recent times, ACC’s contributions on the Greek epics10 have also been aimed at assessing Hesiod’s language, as well as at studying Doric features in the Homeric text. A second large group of papers is devoted to the interaction between dialects and literary genres other than epic. Lyric poetry obviously has a prominent place in most of these writings, but some of them are also concerned with ‘minor’ authors — such as Archytas of Tarentum — and ‘minor’ genres — such as historiography and rhetoric in the Doric dialect.11 ACC has devoted much attention to the role of Doric in choral lyric, as well as to the problematic presence of Aeolic features in choral language.12 Among other things, he has exposed the weakness of E. Risch’s hypothesis concerning the language of Alcman,13 and has put forward a valid explanation of the ‘diphthongized’ -οισα forms occurring in the text of choral poets.14 The papers by ACC which are more closely related to epigraphy range from brilliant explanations of individual words or short texts15 to systematic overviews of longer inscriptions and local dialects.16 Looking at these works as a whole, one cannot avoid being impressed by ACC’s mastery of methods and contents which usually lie outside the purview of Classical philologists and are confined to the ‘distant’ world of epigraphists. Before concluding this overview of ACC’s bibliography, there are still two more fields of research which need to be touched upon. First, it is worth recalling that ACC’s interests extend well beyond the Classical Age, as is shown not only by his papers on the koine,17 but also by those contributions devoted to the Artemidorus Papyrus: ACC has been one of the most authoritative champions of its authenticity.18 To the second field of research belong the studies on the ancient

 10 Nos. 104, 125, 127, 128, and 131. 11 Nos. 50, 52, and 53. 12 Nos. 64 and 75. 13 Nos. 63 and 96. 14 See no. 93. 15 See nos. 60, 68, 85, and 87. 16 On inscriptions, see nos. 94, 98, 130, 133, and 134. On dialects, see nos. 43 and 88. 17 Nos. 73, 78, and 108. 18 Nos. 102 and 105.

Introduction  5

grammarians, a topic in which ACC may once again be considered a specialist in his own right.19

 Summary of the Volume It is now time to provide a brief summary of the papers contained in this volume. Each offers a methodological example of how the study of Greek can be greatly enhanced by a truly multidisciplinary perspective in which the analysis of language interacts with epigraphy, textual philology and comparative linguistics, yet without neglecting the role that linguistic features play in the texts in which they are used, and hence in the culture which produced both. The first four papers get to the very heart of one of ACC’s main research interests: epic language. Sara Kaczko proposes a detailed analysis of some seemingly eccentric pronominal forms in Homer favouring the conclusion that, contrary to the communis opinio, they might be due to the influence of dialects from Central and Western Greece. Paying homage to ACC’s long-standing fascination with the impact of continental areas on the Homeric language, she focuses on pronominal features that are likely to have entered the poems in their last phases from dialects of these regions. Turning to another type of oddity, Andreas Willi’s paper deals with the seemingly obscure Homeric formula σθένεϊ βλεμεαίνων. Through an analysis combining phonological, morphological, semantic and philological arguments he shows how in the context of oral poetry — where poets were often forced to update obsolete phonological and morphological features — the meaning of the formula too evolved from the original idea of ‘hurling missiles with strength’ to the later interpretations as ‘rushing forward with strength’. Lucien van Beek and Laura Migliori turn to another well-known Homeric problem: the creation and semantics of the middle perfect. In their wide-ranging paper they apply the methods of functional linguistics and lexical semantics in order to explain the derivational role of Homeric middle perfects and their development within the original perfect system. The fourth paper devoted to epic language, by Enzo Passa, takes us to the colonial West and its original production of hexameter ‘didactic’ poetry, arguing that in the language of authors such as Empedocles and Parmenides Doric features must have played a much more important role than scholars have been willing to admit.

 19 On this subject no. 66 is fundamental.

6  Enzo Passa and Olga Tribulato Passa’s paper, which highlights the permeability of the linguistic codes of different Greek literary genres, serves as a ‘hinge’ between the first two parts of the collection. The two following contributions are devoted to lyric poetry and its linguistic influence in Greek literature, a topic which again features prominently among ACC’s long-standing interests. Lucia Prauscello offers a painstakingly detailed analysis of the meaning of the participle ἀππασάμενος in the context of Corinna’s fr. 654 PMG, which is concerned with the prophet Orion’s role in Boeotian genealogies. Her reappraisal of linguistic arguments and philological and epigraphic evidence shows how a multidisciplinary approach can often illuminate obscure passages from lesser-known texts. A similar approach informs the in-depth contribution of Federico Favi, which explores the influence of lyric poetry on the language of comedy, focusing on one of ACC’s beloved authors: Epicharmus. Tackling fr. 80 K.-A. Favi makes a good case for the authenticity of two neglected ‘diphthongized’ forms in ‑οισα occurring as variant readings in the manuscript tradition of the fragment. The next four papers look into a variety of topics and are linked by a common approach: the study of how Greek literary language evolved across the centuries, how it was perceived by ancient scholars, and what contribution modern linguistic approaches can provide to our understanding of both these issues. Alessandro Vatri addresses the problem of the alleged presence of dactylic sequences in 6thand 5th-century Ionic prose. In spite of Aristotle’s idea that early prose was ‘rhythmic’, the almost complete silence of ancient rhetoric on the specific presence of such features suggests that by the time ancient rhetoricians started working on early prose, these texts must have circulated exclusively in a written form. Hence, as Vatri argues, ancient commentators simply did not have any perception of these dactylic sequences. When ancient sources are not particularly forthcoming with information guiding us in the correct interpretation of the semantics of isolated words, comparative linguistics can greatly advance our knowledge of certain forms. Matilde Serangeli offers an example of what historical semantics can achieve by tackling ταπεινός. She makes the case for explaining the different nuances of this seemingly isolated adjective by connecting it with a number of other forms in Greek, Hittite and Slavonic. Her paper shows that a correct identification of the word-formation processes behind all these cognate forms is crucial for understanding their meanings, which are only superficially unrelated to one another. Another way of investigating the reception of earlier texts by later readers is the use, manipulation or dropping of certain lexical features. By focusing on the history of the word ἄλγος, Enrico Cerroni offers an example of the cultural, liter-

Introduction  7

ary and linguistic information that we can gain by studying the lexicon in a diachronic perspective. Through this exercise in historical semantics Cerroni warns us against the all too easy conclusion that much of the Greek vocabulary is ‘poetic’ simply because certain words abound in poetry, reminding us of how much ‘non-canonical’ literature we have lost. The problem of correctly identifying different registers in ancient language(s), and hence which words were felt to be appropriate for each of them, strongly emerges when one looks at Greek lexicography. In her paper Olga Tribulato tackles the history of a few Attic terms and the different relevance they had for ancient and Byzantine Greeks, and hence for the lexicographers who collected words from 5th-century texts and explained their usage to speakers of later forms of the language. A linguistic approach to the material preserved in Greek lexicography — an area which has usually remained the realm of Textgeschichte — can thus open up new perspectives in the study of the complex negotiation of the Greeks with their linguistic heritage.

 A Common Approach to the Many Paths of Greek Although ACC’s teachings have been so numerous and profound that every pupil would perhaps indicate a different one as the core of his magisterium, we should like, by way of conclusion, to express what we believe to be the overarching principle that has inspired his work and the many people he has taught and advised over the years. For us — and, we trust, for many others — ACC has been, first and foremost, a master of intellectual freedom, someone who has always encouraged his pupils to brave the seas of learning in order to explore new paths and new territories, and hence to think for themselves outside of academic dogmas. Eagerness to learn about new domains of knowledge and trust in one’s own intellectual capacities are attitudes which anyone who has spent time with ACC has absorbed. It is perhaps in this free way of thinking and of experiencing things — and hence of being — that the fundamental and truly humanistic legacy of ACC lies: a common approach which all his pupils share and which inspires their work, be it in universities, schools, publishing houses or other public institutions. This is what makes ACC a true professor for many of his pupils.

Bibliography of Albio Cesare Cassio (Collected and edited by Paola Ascheri) 1971 1. “Ἐντείνεσθαι, ἔντονος ed il nomos orthios”, in: RFIC 99, 53–57. 2. “Laso e Damone sofisti e novatori”, in: PP 26, 275–280. 3. “Review of T.B.L. Webster, The Greek Chorus (London 1970)”, in: RFIC 99, 449–452. 1972 4. “Note sul Bruni traduttore”, in: E. Cecchini/A.C. Cassio, “Due contributi sulla traduzione di Leonardo Bruni del Pluto di Aristofane”, in: GIF 3, 472–482. 5. “Pind. Pae. 4 (fr. 52 d Sn.) 21–24”, in: ASNP 2, 469–471. 6. “Suid. 3. 410, 11 Adler”, in: Rivista di Studi bizantini e neoellenici 8–9, (1971–1972), 219– 221. 1973 7. “Le note del Poliziano all’Antologia Greca”, in: A.M. Adorisio/A.C. Cassio, “Un nuovo incunabolo postillato da Angelo Poliziano”, in: IMU 16, 263–287. 8. “L’incipit della Chioma callimachea in Virgilio”, in: RFIC 101, 329–332. 9. “Review of V. Di Benedetto, Euripide: Teatro e società (Torino 1971)”, in: CS 1, 1–3. 1974 10. “Un frammento del Cocalo di Aristofane (351 K.)”, in: RFIC 102, 164–169. 1975 11. “Menandro, fr. 269 K.-Th.”, in: RFIC 103, 394–398. 12. “Review of R.G. Ussher, Aristophanis Ecclesiazusae (Oxford 1973)”, in: RFIC 103, 73–75. 13. “Un uso di ὄντως, ἀληθῶς, vere e due epigrammi dell’Antologia Palatina (11, 78 e 394)”, in: RFIC 103, 136–143. 1976 14. “Review of A. La Penna, Le vie dell’anticlassicismo, «Quad. Stor.» 3, 1976, 1-13”, in: RFIC 104, 381–382. 15. “Review of A. Meschini, Giano Laskaris: Epigrammi greci (Padova 1976)”, in: ASNP 6, 1422–1423. 1977 16. Aristofane: Banchettanti (Δαιταλῆς). I frammenti, Pisa. 17. “Review of E. Fraenkel, Wilamowitz, «Quad. stor.» 5, 1977, 101–118”, in: RFIC 105, 381. 1978 18. “Arte e artifici di Menandro (Aspis 1–18)”, in: E. Livrea/G.A. Privitera (eds.), Studi in onore di Anthos Ardizzoni, Roma, 173–185. 19. “Epigrammata Bobiensia”, in: Enciclopedia Italiana, vol. 1: A-Ga (appendix), Roma, 708. https://doi.org/10.1515/9783110621747-002

10  Bibliography of Albio Cesare Cassio 20. “L’uovo ‘orfico’ e il Geritade di Aristofane (fr. 164 K.)”, in: RFIC 106, 28–31. 21. “Review of R. Kassel, Aristotelis Ars rhetorica, (Berolini/Novi Eboraci 1976)”, in: RFIC 106, 86–93. 22. “Review of S. Timpanaro, La filologia di Giacomo Leopardi (Bari 19782)”, in: RFIC 106, 384. 1979 23. “Motivi di agone epirrematico in commedia nuova (Aristoph. Nub. 961 sgg., Plaut. Bacch. 420 sgg.)”, in: AION(filol.) 1, 1–22. 24. “Review of J. Blomqvist, Das sogenannte καί adversativum: Zur Semantik einer griechischen Partikel (Uppsala 1979)”, in: RFIC 107, 500. 25. “Review of M. Landfester, Handlungsverlauf und Komik in den frühen Komödien des Aristophanes (Berlin/New York 1977)”, in: RFIC 107, 194–200. 1980 26. “Pastori e agricoltori in Sofocle (fr. 502 Radt)”, in: RFIC 108, 260–262. 27. “Su un nuovo frammento di Eupoli”, in: RFIC 108, 263. 1981 28. “A “Typical” Servant in Aristophanes (Pap. Flor. 112, Austin 63, 90 ff.)”, in: ZPE 41, 17–18. 29. “Attico “volgare” e Ioni in Atene alla fine del V sec. a.C.”, in: AION(ling.) 3, 79–93. 1982 30. “Ancora ὀψοφόρος”, in: AION(ling.) 4, 195–196. 31. “Arte compositiva e politica in Aristofane: Il discorso di Ermete nella Pace (603–648)”, in: RFIC 110, 22–44. 32. “Review of D.A. Russell/N.G. Wilson, Menander rhetor (Oxford 1981)”, in: RFIC 110, 346– 350. 33. “Review of Ι.Ε. Στεφάνης, ῾Ο δοῦλος στίς κωμωδίες τοῦ ᾽Αριστοφάνη (Thessaloniki 1981)”, in: RFIC 110, 186–187. 1983 34. “Post-classical Λέσβιαι”, in: CQ 33, 296–297. 35. “Review of D.M. Bain, Masters, Servants and Orders in Greek Tragedy: A Study of Some Aspects of Dramatic Technique and Convention (Manchester 1981)”, in: RFIC 111, 205–207. 36. “Synes. Hymn. 9, 117”, in: RFIC 111, 460–464. 1984 37. “Il carattere dei dialetti greci e l’opposizione Dori–Ioni: Testimonianze antiche e teorie di età romantica (su Aristid. Quint. 2.3, Iambl. V. Pythagor. 241 sgg., Sch. in Dion. Thr. p. 117, 18 Hilgard)”, in: AION(ling.) 6, 113–136. 38. “L’accusa di Crizia e le più antiche valutazioni di Archiloco”, in: Lirica greca da Archiloco a Elitis: Studi in onore di Filippo Maria Pontani, Padova, 61–65. 1985 39. Commedia e partecipazione: La Pace di Aristofane, Napoli. 40. “Old Persian marika-, Eupolis’ Marikas and Aristophanes’ Knights”, in: CQ 35, 38–42.

Bibliography of Albio Cesare Cassio  11

41. “Review of G.A. Seeck, Dramatische Strukturen der griechischen Tragödie: Untersuchungen zu Aischylos (München 1981)”, in: RFIC 113, 254–255. 42. “Two Studies in Epicharmus and His Influence”, in: HSPh 89, 37–51. 1986 43. “Continuità e riprese arcaizzanti nell’uso epigrafico dei dialetti greci: Il caso dell’eolico d’Asia”, in: AION(ling.) 8, 131–146. 44. “Review of R. Seaford, Euripides: Cyclops (Oxford 1984)”, in: RFIC 114, 206–208. 1987 45. “Addendum” of M.L. Lazzarini/A.C. Cassio, “Sulla laminetta di Hipponion”, in: ASNP 17, 329–334. 46. “Il monte Ἴφειον e gli errori di maiuscola (Sch. ad Callim. h. 5,37 sgg.)”, in: SCO 37, 177– 179. 47. “I tempi di composizione delle commedie attiche e una parafrasi di Aristofane in Galeno (Ar. fr. 346 K.-A.)”, in: RFIC 115, 5–11. 48. “ΠΙΕΝ nella laminetta di Hipponion”, in: RFIC 115, 314–316. 1988 49. “Neottolemo di Pario e la cultura dell’Asia Minore ellenistica”, in: AION(filol.) 9–10, (1987–1988), 125–133. 50. “Nicomachus of Gerasa and the Dialect of Archytas, fr. 1”, in: CQ 38, 135–139. 1989 51. A.C. Cassio/D. Musti (eds.), Tra Sicilia e Magna Grecia: Aspetti di interazione culturale nel IV sec. a.C. Atti del convegno (Napoli, 19–20 marzo 1987), Roma (= AION(filol.) 11). 52. “Lo sviluppo della prosa dorica e le tradizioni occidentali della retorica greca”, in: A.C. Cassio/D. Musti (eds.), Tra Sicilia e Magna Grecia: Aspetti di interazione culturale nel IV sec. a.C. Atti del convegno (Napoli, 19–20 marzo 1987)”, Roma, 137–157 (= AION(filol.) 11). 53. “Storiografia locale di Argo e dorico letterario: Agia, Dercillo ed il Pap. Soc. Ital. 1091”, in: RFIC 117, 257–275. 1990 54. “Epicrate a Nasso di Sicilia”, in: RFIC 118, 24–25. 55. “Review of J. Karageorghis/O. Masson (eds.), The History of the Greek Language in Cyprus. Proceedings of an International Symposium Sponsored by the Pierides Foundation, Larnaca, Cyprus, 8–13 September 1986 (Nicosia 1988)”, in: RFIC 118, 501–502. 1991 56. A.C. Cassio/G. Cerri (eds.), L’inno tra rituale e letteratura nel mondo antico. Atti di un colloquio (Napoli, 21–24 ottobre 1991), Roma (= AION(filol.) 13). 57. “Ἀνάγκη = clessidra? (Aristoph. fr. 601 Kassel–Austin)”, in: Studi di filologia classica in onore di Giusto Monaco, vol. 1: Letteratura greca, Palermo, 299–302. 58. “Callimaco e le glosse di Democrito”, in: Palaeograeca et Mycenaea Antonino Bartoněk quinque et sexagenario oblata, Brunae, 7–12. 59. “Conclusioni”, in: A.C. Cassio/G. Cerri (eds.), L’inno tra rituale e letteratura nel mondo antico. Atti di un colloquio (Napoli, 21–24 ottobre 1991), Roma, 307–310 (= AION(filol.) 13).

12  Bibliography of Albio Cesare Cassio 60. “ΟΑΔΙΣΤΗ E ΟΑΛΙΔΙΟΣ (SEG 24, 548; IG XII 9, 249 B 290)”, in: ZPE 87, 47–52. 61. “Un re di Persia sui monti dell’oro (Ar. Ach. 80ss.; Ctes. FGrH 688 F 45)”, in: Eikasmos 2, 137–141. 1992 62. “Review of M. Bile, Le dialecte crétois ancien: Étude de la langue des inscriptions. Recueil des inscriptions postérieures aux IC (Paris 1988)”, in: RFIC 120, 379. 1993 63. “Alcmane, il dialetto di Cirene e la filologia alessandrina”, in: RFIC 121, 24–36. 64. “Iperdorismi callimachei e testo antico dei lirici (Call. hy. 5, 109; 6, 136)”, in: R. Pretagostini (ed.), Tradizione e innovazione nella cultura greca da Omero all’età ellenistica: Scritti in onore di Bruno Gentili, Roma, 903–910. 65. “La più antica iscrizione greca di Cuma e τίν(ν)υμαι in Omero”, in: Sprache 35, (1991– 1993), 187–207. 66. “Parlate locali, dialetti delle stirpi e fonti letterarie nei grammatici greci”, in: E. Crespo/ J.L. García Ramón/A. Striano (eds.), Dialectologica Graeca. Actas del II Coloquio Internacional de Dialectología Griega (Miraflores de la Sierra [Madrid], 19–20 de junio de 1991), Madrid, 73–90. 1994 67. A.C. Cassio/P. Poccetti (eds.), Forme di religiosità e tradizioni sapienziali in Magna Grecia. Atti del convegno (Napoli, 14–15 dicembre 1993), Napoli (= AION(filol.) 16). 68. “Giavellotti contro frecce: Nuova lettura di una tessera dal tempio di Atena a Camarina e Hom. Od. 8, 229”, in: RFIC 122, 5–20. 69. “I distici del polyandrion di Ambracia e l’«io anonimo» nell’epigramma greco”, in: D. Marcozzi/M. Sinatra/R. Valeri/P. Vannicelli (eds.), Giornata di studio in memoria di Marcello Durante (Roma, 23 marzo 1994), Roma, 101–117 (= SMEA 33). 70. “Κεῖνος, καλλιστέφανος, e la circolazione dell’epica in area euboica”, in: B. d’Agostino/ D. Ridgway (eds.), ΑΠΟΙΚΙΑ: I più antichi insediamenti greci in occidente. Funzioni e modi dell’organizzazione politica e sociale. Scritti in onore di Giorgio Buchner, Napoli, 55–67 (= AION(archeol.) 1). 71. “Πιέναι e il modello ionico della laminetta di Hipponion”, in: A.C. Cassio/P. Poccetti (eds.), Forme di religiosità e tradizioni sapienziali in Magna Grecia. Atti del convegno (Napoli, 14–15 dicembre 1993), Napoli, 183–205 (= AION(filol.) 16). 1996 72. “Da Elea a Hipponion e Leontinoi: Lingua di Parmenide e testi epigrafici”, in: ZPE 113, 14– 20. 73. “La prose ionienne postclassique et la culture de l’Asie Mineure à l’époque hellénistique”, in: C. Brixhe (ed.), La koiné grecque antique, vol. 2: La concurrence, Nancy/ Paris, 147–170. 1997 74. A.C. Cassio (ed.), Katà Diálekton. Atti del III colloquio internazionale di Dialettologia Greca (Napoli–Fiaiano d’Ischia, 25–28 settembre 1996), Napoli (= AION(filol.) 19).

Bibliography of Albio Cesare Cassio  13

75. “Futuri dorici, dialetto di Siracusa e testo antico dei lirici greci”, in: A.C. Cassio (ed.), Katà Diálekton. Atti del III colloquio internazionale di Dialettologia Greca (Napoli–Fiaiano d’Ischia, 25–28 settembre 1996), Napoli, 187–211 (= AION(filol.) 19). 1998 76. “Congetture, varianti dialettali e grafie nel testo omerico”, in: A. Ferrari (ed.), Filologia classica e filologia romanza: Esperienze ecdotiche a confronto. Atti del convegno (Roma, 25–27 maggio 1995), Spoleto, 281–290. 77. “Γερωχία e ἀγερωχία: Comicità e dialetto nella Lisistrata di Aristofane”, in: SemRom 1, 73– 78. 78. “La lingua greca come lingua universale”, in: S. Settis (ed.), I Greci: Storia, cultura, arte, società, vol. 2.3: Una storia greca: Trasformazioni, Torino, 991–1013. 79. “Lo sviluppo dell’epica greca e il mondo euboico”, in: M. Bats/B. d’Agostino (eds.), Euboica: L’Eubea e la presenza euboica in Calcidica e in Occidente. Atti del convegno internazionale di Napoli, 13–16 novembre 1996, Napoli, 11–22 (= Quaderni di AION(archeol.) 12). 1999 80. “Epica greca e scrittura tra VIII e VII secolo a.C.: Madrepatria e colonie d’Occidente”, in: G. Bagnasco Gianni/F. Cordano (eds.), Scritture mediterranee tra il IX e il VII sec. a.C. Atti del Seminario (Milano, 23–24 febbraio 1998), Milano, 67–84. 2000 81. A.C. Cassio/D. Musti/L.E. Rossi (eds.), Synaulia: Cultura musicale in Grecia e contatti mediterranei, Napoli (= Quaderni di AION(filol.) 5). 82. “Dialetti greci e pseudepigrapha pitagorici: Le valutazioni degli antichi”, in: G. Cerri (ed.), La letteratura pseudepigrafa nella cultura greca e romana. Atti di un Incontro di studi (Napoli, 15–17 gennaio 1998), Napoli, 153–166 (= AION(filol.) 22). 83. “Esametri orfici, dialetto attico e musica dell’Asia Minore”, in: A.C. Cassio/D. Musti/ L.E. Rossi (eds.), Synaulia: Cultura musicale in Grecia e contatti mediterranei, Napoli, 97– 110 (= Quaderni di AION(filol.) 5). 84. “Ἐφόμιλος”, in L. Dubois/E. Masson (eds.), Philokypros: Mélanges de philologie et d’antiquités grecques et proche-orientales dédiés à la mémoire d’Olivier Masson, Salamanca, 103–107 (= Suplementos a Minos 16). 85. “Un epigramma votivo spartano per Atena Alea”, in: RFIC 128, 129–134. 2002 86. “Early Editions of the Greek Epics and Homeric Textual Criticism in the Sixth and Fifth Centuries BC”, in: F. Montanari (ed.), Omero tremila anni dopo. Atti del convegno di Genova (6–8 luglio 2000), Roma, 105–136. 87. “Explanation of the Personal Name Putikkas” in: F. Cordano/M. Di Salvatore (eds.), Il Guerriero di Castiglione di Ragusa: Greci e Siculi nella Sicilia sud-orientale. Atti del seminario (Milano, 15 maggio 2000), Roma, 122–125. 88. “Il dialetto greco di Taranto”, in: Taranto e il Mediterraneo. Atti del quarantunesimo convegno di studi sulla Magna Grecia (Taranto, 12–16 ottobre 2001), Taranto, 435–465. 89. “The Language of Doric Comedy”, in: A. Willi (ed.), The Language of Greek Comedy, Oxford, 51–83.

14  Bibliography of Albio Cesare Cassio 2003 90. “Ospitare in casa poeti orali: Omero, Testoride, Creofilo e Staroselac ([Herodot.] Vita Hom. 190ss. Allen; Plat. Resp. 600b)”, in: R. Nicolai (ed.), Rhysmos: Studi di poesia, metrica e musica greca offerti dagli allievi a Luigi Enrico Rossi per i suoi settant’anni, Roma, 35–45 (= Quaderni di SemRom 6). 2004 91. “Review of R. Kassel/C. Austin, Poetae Comici Graeci (PCG), vol. 1: Comoedia Dorica, Mimi, Phlyaces (Berolini/Novi Eboraci 2001)”, in: Gnomon 76, 193–198. 92. “Spoken Language and Written Text: The Case of alloeidea (Hom. Od. 13.194)”, in: J.H.W. Penney (ed.), Indo-European Perspectives: Studies in Honour of Anna Morpurgo Davies, Oxford, 83–94. 2005 93. “I dialetti eolici e la lingua della lirica corale”, in: F. Bertolini/F. Gasti (eds.), Dialetti e lingue letterarie nella Grecia arcaica. Atti della IV Giornata ghisleriana di Filologia classica (Pavia, 1–2 aprile 2004), Como/Pavia, 13–44. 94. “Il grande codice di Gortyna: Problemi di dialetto”, in: E. Greco/M. Lombardo (eds.), La grande iscrizione di Gortyna centoventi anni dopo la scoperta. Atti del I convegno internazionale di studi sulla Messarà (Atene-Haghii Deka, 25–28 maggio 2004), Atene, 85–98. 2006 95. “La Tessaglia, l’isola di Lesbo e i dativi plurali del tipo ἐπέεσσι”, in: M. Vetta/C. Catenacci/M. Di Marzio/L. Quattrocelli/S. Santoro (eds.), I luoghi e la poesia nella Grecia antica. Atti del convegno, Università «G. d’Annunzio» di Chieti/Pescara, 20–22 aprile 2004, Alessandria, 73–84. 2007 96. “Alkman’s Text, Spoken Laconian, and Greek Study of Greek Dialects”, in: I. Hajnal (ed.), Die altgriechischen Dialekte: Wesen und Werden. Akten des Kolloquiums Freie Universität Berlin, 19.–22. September 2001, Innsbruck, 29–45. 97. “Enquiries and Responses: Two Lead Tablets from Dodona”, in: M.B. Hatzopoulos (ed.), Φωνής χαρακτήρ εθνικός. Actes du Ve Congrès de dialectologie grecque (Athènes, 28–30 septembre 2006), Athènes, 29–33. 98. “Scultori, epigrammi e dialetti nella Grecia arcaica: La stele di Mnasitheos (SEG 49, 1999, no. 505)”, in: G. Lozza/S. Martinelli Tempesta (eds.), L’epigramma greco: Problemi e prospettive. Atti del congresso della Consulta Universitaria del Greco (Milano, 21 ottobre 2005), Milano, 1–18 (= Quaderni di ACME 91). 2008 99. A.C. Cassio (ed.), Storia delle lingue letterarie greche, Firenze. 100. “Alfabeti locali, testi arcaici, edizioni ellenistiche”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, Firenze, 70–96. 101. “La fonologia del greco e le sue radici indoeuropee”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, Firenze, 32–69. 102. “Lingua e stile nel testo del rotolo”, in: C. Gallazzi/B. Kramer/S. Settis (eds.), Il Papiro di Artemidoro, Milano, 134–139.

Bibliography of Albio Cesare Cassio  15

103. “Stirpi, gruppi dialettali e lingue letterarie”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, Firenze, 4–31. 2009 104. “The Language of Hesiod and the Corpus Hesiodeum”, in: F. Montanari/A. Rengakos/ C. Tsagalis (eds.), Brill’s Companion to Hesiod, Leiden/Boston, 179–201. 105. “Cultura ellenistica e linguaggio religioso in Artemidoro”, in: C. Gallazzi/B. Kramer/ S. Settis/A. Soldati (eds.), Intorno al Papiro di Artemidoro, vol. 1: Contesto culturale, lingua, stile e tradizione. Atti del convegno internazionale del 15 novembre 2008 presso la Scuola Normale Superiore di Pisa, Milano, 79–90. 2010 106. “Il participio del verbo ‘essere’ in Eolico d’Asia tra filologia e linguistica”, in: I. Putzu/ G. Paulis/G. Nieddu/P. Cuzzolin (eds.), La morfologia del greco tra tipologia e diacronia. Atti del VII Incontro internazionale di linguistica greca (Cagliari, 13–15 settembre 2007), Milano, 25–30. 2012 107. “Epica orale fluttuante e testo omerico fissato: Riflessi su Stesicoro (PMGF 222b 229 e 275)”, in: SemRom 1, 254–260. 108. “Intimations of koine in Sicilian Doric: The Information Provided by the Antiatticist”, in: O. Tribulato (ed.), Language and Linguistic Contact in Ancient Sicily, Cambridge, 251–264. 109. “Kypris, Kythereia and the Fifth Book of the Iliad”, in: F. Montanari/A. Rengakos/C. Tsagalis (eds.), Homeric Contexts: Neoanalysis and the Interpretation of Oral Poetry, Berlin/Boston, 413–426. 110. “L’eroe Lakedaimon e gli onori funebri per i re di Sparta (orac. ap. Herodot. 7, 220 = nr. 100 Parke-Wormell)”, in: M. Passalacqua/M. De Nonno/A.M. Morelli (eds.), Venuste Noster: Scritti offerti a Leopoldo Gamberale, Hildesheim, 37–42. 111. “Papiri omerici e ricostruzione linguistica”, in: G. Bastianini/A. Casanova (eds.), I papiri omerici. Atti del convegno internazionale di studi (Firenze, 9–10 giugno 2011), Firenze, 231–240. 112. “Scrittura e potere nella Grecia arcaica: Il caso di Creta e Cipro”, in: M. Castiglione/ A. Poggio (eds.), Arte-Potere: Forme artistiche, istituzioni, paradigmi interpretativi. Atti del convegno tenuto a Pisa, Scuola Normale Superiore, 25–27 novembre 2010, Milano, 33–43. 113. “Sofocle, Edipo re 420 sgg.”, in: L. Gamberale/M. De Nonno/C. Di Giovine/M. Passalacqua (eds.), Le strade della filologia: Per Scevola Mariotti, Roma, 3–9. 2013 114. “Οἰρών: “confine / territorio” tra Cipro e la Magna Grecia”, in: A. Inglese (ed.), Epigrammata 2: Definire, descrivere, proteggere lo spazio. In ricordo di André Laronde. Atti del convegno di Roma (26–27 ottobre 2012), Roma, 51–59. 2014 115. “Due atticismi in Teocrito (14, 6)”, in: A. Gostoli/R. Velardi (eds.), Mythologeîn: Mito e forme di discorso nel mondo antico. Studi in onore di Giovanni Cerri, Pisa, 395–396 (= Quaderni di AION(filol.) 18).

16  Bibliography of Albio Cesare Cassio 116. “Innovazioni linguistiche e tratti locali nei più antichi oracoli delfici”, in: SemRom 3, 257– 270. 117. “Lessico ‘moderno’ nei testi greci antichi: Storie di continuità e discontinuità”, in: C. Carpinato/O. Tribulato (eds.), Storia e storie della lingua greca, Venezia, 35–48. 2015 118. “Le ‘consacrazioni’ di Kollyra (Iscr. Gr. Italia, Locri, no. 89 = IG XIV 644)”, in: A. Inglese (ed.), Epigrammata 3: Saper scrivere nel mondo antico. Esiti di scrittura fra VI e IV sec. a.C. In ricordo di Mario Luni. Atti del convegno di Roma (7–8 novembre 2014), Roma, 205–211. 119. “Rovesciare il destino (P. Sapph. Obbink 13–16)”, in: RFIC 143, 273–277. 2016 120. A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano. 121. “Alfabeti locali, testi arcaici, edizioni ellenistiche”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano, 107–133. 122. “Appendix” of B. d’Agostino/M.G. Palmieri, “Potters, Hippeis and Gods at Penteskouphia (Corinth), Seventh to Sixth Centuries BC”, in: J. Bintliff/K. Rutter (eds.), The Archaeology of Greece and Rome: Studies in Honour of Anthony Snodgrass, Edinburgh, 155–182. 123. “Introduzione alla morfologia nominale e verbale”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano, 73–106. 124. “La fonologia del greco e le sue radici indoeuropee”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano, 34–72. 125. “Overlong Syllables in the Epic ‘Adonius’ and the Compositional Stages of Greek Hexameter Poetry”, in: F. Gallo (ed.), Omero: Quaestiones disputatae, Milano, 31–41. 126. “Stirpi, gruppi dialettali e lingue letterarie”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano, 3–31. 2017 127. “‘Authentic’ vs. ‘Artificial’: Homeric ἐπέεσσι(v) Reconsidered”, in: C. Tsagalis/A. Markantonatos (eds.), The Winnowing Oar: New Perspectives in Homeric Studies. Studies in Honor of Antonios Rengakos, Berlin/Boston, 171–189. 128. “Sbagli d’oro: Dialetto, manipolazioni ed errori nella laminetta di Hipponion” in: La filologia e l’errore. Atti del convegno, Roma, 28–29 settembre 2016, Roma, 15–27 (= Rationes rerum 10). 129. “Trinacria Perfecta: Some Unusual Perfects and Pluperfects in Ancient Sicily”, in: A. Willi (ed.), Sprachgeschichte und Epigraphik: Festgaben für Rudolf Wachter zum 60. Geburtstag, Innsbruck, 39–47. 2018 130. “Le aspirazioni di una dea greca: Ἐνὑώ tra Omero e Naxos di Sicilia”, in: F. Camia/L. Del Monaco/M. Nocita (eds.), Munus Laetitiae: Studi miscellanei offerti a Maria Letizia Lazzarini, vol. 2, Roma, 419–422. 131. “Notes on the Origin and Diffusion of the -εσσι Datives”, in: G.K. Giannakis/E. Crespo/ P. Filos (eds.), Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, Berlin/Boston, 189–196.

Bibliography of Albio Cesare Cassio  17

132. “Old Ablatives, Homeric τῶ, and Helen’s Disenchantment (Iliad 6.352–3)”, in: D. Gunker/ S.W. Jamison/A.O. Mercado/K. Yoshida (eds.), Vina Diem Celebrent: Studies in Linguistics and Philology in Honor of Brent Vine, Ann Arbor, 12–22. Forthcoming 133. “E Plumbo Lux: A Rare Perfect Participle in a Lead Letter in Ionic and Avoidance of Overlength in Homer”, in: Proceedings of 7. Internationales Kolloquium über die altgriechischen Dialekte – VIIème colloque international sur les dialectes grecs anciens – Seventh International Colloquium on Ancient Greek Dialects (Basel/Lausanne, 21–25 Juni 2017). 134. “L’épigramme pour Théotimos d’Atrax, CEG 637 (Larissa 437 av. J. Chr.): Mise en page, dialecte, ambiguïtés”, in: A. Alonso Déniz/E. Santin (eds.), Langue poétique et formes dialectales dans les inscriptions versifiées grecques, dossier collectif, Revue de philologie, de littérature et d’histoire anciennes. 135. “Metamorfosi della lingua epica tra oriente e occidente: Da Omero alle laminette orfiche e alla celebrazione poetica dei dinasti della Licia”, in: A. Willi (ed.), Formes et fonctions des langues littéraires en Grèce ancienne (= Entretiens Hardt sur l’antiquité classique 65). 136. “Metrical Constraint and Dialect Borrowings: ἔσεται, ἐσσεῖται and the Homeric Futures of the Verb “To Be””, in: L. van Beek (ed.), Language Change in Epic Greek and Other Oral Traditions. 137. “Recentior, non deterior: uno sguardo alla lingua di Iliade 24”, in: G. Lentini (ed.), Il riscatto di Ettore: Ricerche sul libro ventiquattresimo dell’Iliade.

Sara Kaczko

Unconventional Features in Homer: The Case of ἑέ and ἑοῖ Abstract: This paper deals with some ‘unconventional’ linguistic elements in Homer, the 3rd person pronouns ἑέ and ἑοῖ. It analyzes the distribution of these pronouns in the Homeric poems, their relation to τεοῖο, τεΐν and the type ἑαυτοῦ, ἑωυτοῦ, and their possible origin with regard to the Greek dialects. More specifically it argues that ἑέ and ἑοῖ are to be counted among a number of features that reflect the cultural influence of Central and Western Greece that entered the epic diction in its latest phases. It also draws attention to the fact that if the explanation of the type ἑαυτοῦ, ἑωυτοῦ as derived from the univerbation of ἑοῖ + αὐτfound in several etymological dictionaries and handbooks is to be accepted, Attic and Ionic should be considered as regards the origin of ἑοῖ (and ἑέ) in Homer; on the other hand, if a West Greek origin for ἑοῖ (and ἑέ) is deemed likely, the abovementioned ‘traditional’ explanation for ἑαυτοῦ, ἑωυτοῦ may require reconsideration, especially in comparison with a reconstruction of ἑαυτοῦ, ἑωυτοῦ as derived from univerbation of ἕο + αὐτ-. Keywords: Ancient Greek pronouns, Homeric language, Greek dialects, recent features in Homer.

 Non-Formulaic Features in Homer and the Influence of Central/Western Greece By the last decades of the 20th century, many experts on the Homeric language and on the history of the Greek language had come to the conclusion that a number of relatively uncommon and largely non-formulaic features of Homeric Greek reflect the cultural influence of Central and Western Greece, in particular of Euboea, Attica, and Boeotia (and possibly of some other Western regions, such as Phokis and Lokris).1 In recent discussion of this topic, the widely publicized archaeological finds at Lefkandi and a systematic account of this influence by the

 1 See Cassio 1998; cf. already Wathelet 1970, 154–157; Wathelet 1981; Peters 1986; Peters 1987; West 1988; Forssman 1991; Janko 1992, 18 n. 33 and Janko 2017, 162; Ruijgh 1995, 47–50; Hackstein 2010, 402; Cassio 2018a. https://doi.org/10.1515/9783110621747-003

20  Sara Kaczko late Martin West (1988) have been of paramount importance, though the issues involved were already being debated, albeit in different terms, a century ago, as can be seen in the work of Wackernagel 1916. West’s article prompted a lively discussion,2 including the important contribution by Cassio 1998, and renewed interest in previous arguments (e.g. the identification of Euboean elements by Wathelet 1970 and Wathelet 1981). As a consequence, scholars have become increasingly interested in the relationship between the language of archaic Greek epic and the dialects and poetic traditions of Euboea and Boeotia.3 Among the above-mentioned features of the Homeric epics that can be labeled, for practical purposes, ‘continental’, distinctions have been made in terms of the greater or lesser likelihood that they should be associated with Central or Western Greece. Thus, for instance, the occasional lack of compensatory lengthening after the loss of [w] (e.g. ξένιος) and the so-called ‘short datives’ in ‑οις, ‑ῃς, (‑αις) were initially considered Euboean traits and only later connected (at least by some scholars) with other regions of Central and Western Greece. By contrast, the West Greek and Boeotian parallels of several pronominal forms were immediately acknowledged, for example by Peters 1987 and by West 1988, and a few years later by Cassio 1998. West provided the first systematic analysis, counting various pronouns among the few convenient linguistic forms of West Greek origin that were introduced by epic bards active in Boeotia or other parts of Central Greece during the Dark Ages “into the crucible where the language of Homer was being forged” (i.e. Euboea, from West’s perspective) and that had been often ignored by modern scholars because “on conventional accounts of the development of epic in Asia Minor their presence would be inexplicable”.4 Following this reasoning, such pronominal forms fall into three groups: (1) τύνη, τεοῖο, and τεΐν, characterized by “a strong form of the stem” (as West 1988, 167 put it, i.e. *tū, *tew-, as opposed to the “weak” forms, *tŭ, *tw-) and, with

 2 Chadwick 1990; West 1992; Wyatt 1992. 3 Some features had been identified as Euboean by Wathelet 1970 and Wathelet 1981 (who also identified the ptcpl. ὄντας instead of ἐόντας as an Euboean feature). Similarly, for ‑οις, ‑ῃς, (‑αις), see Peters 1986, 307. As for the elements derived from Boeotia, the question had been discussed only a year before West’s paper by Peters (1987, 236), who suggested that a Western Ionian Homer borrowed τύνη, τεΐν, and especially ἁμός from Boeotia. On this question, see already Peters 1986, 319 n. 49; on the Boeotian origin of ἁμός, ὑμός, especially likely “wenn Homers Ionisch mindestens zum Teil im benachbarten Euboia zu Hause wäre”, cf. Forssman 1991, 273 with n. 62, with reference to Peters 1987 and West 1988. Peter’s contribution did not, however, generate the level of interest that West’s did; see Cassio 1998 and Cassio 2018a. 4 See West 1988, 167−168; he also counted among these convenient elements of West Greek origin the preposition πρότι and the future verbal form ἐσσεῖται, on which see Cassio 2018b.

Unconventional Features in Homer  21

specific regard to τύνη and τεΐν, also characterized by the suffix ‑νη and the ending ‑ιν, which have parallels only in Boeotian, Doric, and West Greek; (2) ἑέ and ἑοῖ, also characterized by “a strong form of the stem”, i.e. *sew‑ (as opposed to ἕ and οἷ, from *sw‑), which have parallels only in Corinna’s ἑΐν and ἑοῦς and are “best interpreted as Boeotisms of W. Greek origin”;5 and (3) ἁμός and ὑμός, which, being inherited forms, could only have entered the epics through the Boeotian or West Greek dialects (otherwise, these forms would have changed to *ἡμός, ἄμμος, or ὔμμος depending on the phases or dialects involved).6

 3rd Person Pronouns, Dialects, and Homer These considerations leave the strong general impression that Central and Western Greece played an important role in the latest phases of archaic Greek epic, and in particular that pronouns are prominent among elements that seem difficult or impossible to explain in terms of the generally accepted theory of the phases of Homeric diction. In this respect, it would be useful to analyze the distribution of some pronouns in the Homeric poems and in the Greek dialects, especially with regard to the dialectal classification of certain forms,7 and to ascertain whether the pronouns singled out by West should be grouped with others that have thus far been largely ignored by Homeric scholars. Since the presence of these pronouns would be inexplicable “on conventional accounts of the development of epic in Asia Minor” (to return to West’s words), the hypothesis should be considered that they are to be counted among the ‘continental’ features (i.e. those that reflect the cultural influence of Central and Western Greece) that entered the epic diction in its latest phases. For the present discussion, I will focus on ἑέ and ἑοῖ. It must be stressed that these forms share some idiosyncrasies with regard to (a) the stem (disyllabic), (b) the accentuation (the forms, admittedly few, being exclusively orthotonic), and (c) their distribution in the Homeric text (which suggests that they are not particularly ancient). In the case of the disyllabic stem, ἑέ (as opposed to ἕ) and ἑός (as  5 West did not dwell on it, but the endings ‑οῦς and ‑ιν (for this latter West quotes ἐμίν, τίν, and ϝιν in West Greek dialects) in personal pronouns are, as far as the evidence goes, to be connected with Doric and West Greek dialects; ‑οῦς instead of ‑ῶς would be especially unexpected and problematic in Boeotian (see also below). 6 West 1988, 167−168. 7 E.g. ἑέ, ἑοῖ, ἐμεῖο, σεῖο and εἷο, ἥμιν and ὕμιν (clitics), τύνη, νώ, σφώ (cf. Kaczko 2002 and Kaczko 2006).

22  Sara Kaczko opposed to ὅς) have been investigated by Homeric scholars — in particular ἑός, which is well attested in the epics and in (part of) the literary tradition — and especially by historians of PIE and of the Greek language (starting with Brugmann) — for whom ἑός has figured in various attempts to reconstruct the 3rd pers. pronominal stem (among other reconstructions).8 The main hypotheses can be summarized as follows. By one account, the forms of the 3rd pers. pronoun go back to PIE *sewe, either “reiner Stamm”, possibly both accusative and genitive, or created from the stem *sew‑ plus the ending *‑e, on which the remaining forms were in turn based.9 Alternatively, the stem *s(e)we could have come from a reduplicated form *swe‑swe.10 Thirdly, in Greek the original opposition between *se‑ (anaphoric) and *swe‑ (reflexive) was replaced by an accentual opposition in the stem *swe‑ (which was generalized at the expenses of *se‑) between orthotonic (ἕο/οὗ, οἷ, ἕ) and clitic (ἑο/οὑ, οἱ, ἑ) forms.11 A fourth possibility is that the adjective ἑός derives from the stem *s(e)wo‑, on which *swe‑ was modeled,12

 8 Both ἑός and ὅς have been analyzed in terms of the possible original use of the adjective (in either form, ὅς or ἑός) in Homer as a ‘free’ reflexive; thus Brugmann 1876 hypothesized that Homer contains remnants of an original reflexive pronominal possessive *sewo‑/*swo‑ that could refer to any person and number, exactly as OInd. svá‑ (cf. also Slavic svoi). Its ‘restriction’ to 3rd pers. contexts would thus represent a secondary modernization traceable to Aristarchus, as shown by both the variants in the MSS of an ‘all-purpose’ adjective (i.e. forms of ὅς and ἑός), which Zenodotus selected, and by use of both the 3rd pers. adjective and the 3rd pers. pronouns οὗ, οἷ, ἕ, οὑ etc. as ‘free’ reflexives in Apollonius Rhodius (Rengakos 1993, 110–119) and Callimachus; cf. also Chantraine 1958, 273−275, and the discussion in Leaf 1900−1902, 559–565 and Nussbaum 1998, 87–129. 9 See Brugmann 1913, 479; Petersen 1930, 177 (according to whom the genitive and the accusative of the 3rd pers. pronoun would not yet have been differentiated in PIE times) and Schmidt 1978, 154 (*sewe created after the proportion 2nd pers. adj. *t(e)wo‑ : 2nd pers. pronoun gen. *tew(e) = 3rd pers. adj. *s(e)wo‑ : 3rd pers. pronoun gen. x) respectively. 10 See Schwyzer 1939, 601. Indeed, reduplication is a well-attested mechanism in the formation of pronouns, such as for example μιν, from (probably) *im-im, and the demonstrative pronouns τοῦτο and αὐτοσαυτόν. On such form(ation)s, see e.g. Rix 1992, 184−186; Wackernagel 1928, 82; Buck 1955, 99. 11 See Petit 1999, 159, who also cautiously speculated that ἑός was the basis upon which the disyllabic forms of the pronoun, ἑέ, ἑοῖ (which would also be “creations occasionnelles de la langue épique”) were created — i.e. the relationship between ὅς and ἕ, οἷ — was the model for that between ἑός and ἑέ, ἑοῖ (Petit 1999, 181−182). 12 In other words, *swe‑ would be a secondary creation from *se- on the basis of the 3rd pers. possessive pronoun *s(e)wo‑; of the two stems, *sewo‑ (of the adjective) would be primary and *s(e)we‑ (of the pronoun) secondary; see Brugmann 1913, 289; Schmidt 1978, 155, 164–166; Rix 1992, 180. On the relationship between an inherited ἑός and the analogically created disyllabic forms of the pronoun, ἑέ, ἑοῖ, see n. 11 above.

Unconventional Features in Homer  23

*sw(e)‑ being the ‘regular’ stem for the pronouns, while ἑέ and ἑοῖ were a secondary formation analogous to the orthotonic forms of the 1st pers. sing.; thus, ἐ‑ (i.e. ἑέ and ἑοῖ) would have been created as *hewe, *hewoi : *hwe, *hwoi, on the analogy of ἐμέ, ἐμοί: με, μοι.13 I shall not discuss (b), the accentuation, in detail because, as has been already mentioned, ἑέ and ἑοῖ are only orthotonic.14 With regard to (c), the distribution of 3rd pers. pronouns with a disyllabic stem, it is worth recalling that the ἑέ type seldom occurs in Homer or in Greek more generally. Thus beyond the Homeric poems, only the above-mentioned ἑΐν, ἑοῦς in Corinna, a genitive ἑοῦ in Hesiod (Theog. 401), and a possible instance of ἑοῖ in Empedocles (actually a conjecture by P. Maas at B 28.1 D.-K.) are counted by scholars as direct instances of the type ἑέ. In this inventory, neither the few variae lectiones of ἑοῦ in Homer (with Zenodotus’ auctoritas) nor the hypothesis that ἐῆος covers for an older genitive *ἑέο are usually taken into account;15 ἑοῖ, on the other hand, is indirectly presupposed in the conventional reconstruction of the Attic pronouns ἑαυτοῦ, ἑαυτῷ, ἑαυτόν and Ionic pronouns ἑωυτοῦ, ἑωυτῷ, ἑωυτόν.16 Otherwise, instances of 3rd pers. pronouns in Aeolic, Attic-Ionic, Doric, and West Greek derive exclusively from the stem *sw‑

 13 Rix 1992, 180. 14 Cf. Ap. Dysc. GG 2.1.1, 82.14−20 Schneider. The accentuation could be another indication that ἑέ and ἑοῖ are a secondary formation on ἐμέ and ἐμοί; on the other hand, the rarity of the forms (only four examples) prevents a definitive conclusion. 15 According to Schwyzer 1938 (see also de Lamberterie 1990, 752–758, esp. 756), the form ἑῆος and the MSS variants ἐῆο, ἑῆο, and εῆο ultimately trace back to *ἑέο and therefore attest to a genitive of the 3rd pers. pronoun from the stem *sew‑. The original *ἑέο would thus have undergone several modifications, first from the original *ἑέο to the metrically lengthened form ἑῆο, then, since this pronominal genitive was found mainly following a nominal genitive in -ος, ἑῆο was remodeled after it, i.e. as ἑῆος, finally ἑῆος was, in turn, misinterpreted in some passages as the genitive of ἐύς ‘good’. Therefore, ἑῆος would have been used as the genitive of an allperson reflexive pronoun instead of ἑοῖο and as the genitive of ἐύς (owing to the misinterpretation). This scenario, however, has been persuasively challenged by Nussbaum (1998, 122–129), one of the more obvious criticisms being that the metrically lengthened form of *ἑέο would have been *ἑεῖο rather than ἑῆο (there are also the parallels of the 1st and 2nd pers. sing. ἐμεῖο and σεῖο); another being its complexity (starting with theories that a sequence of two identical vowels could be rendered in writing with a single vowel and that such a hypothetical form — i.e. — in an early version of the Homeric text would have been interpretable either as *ἑέο or as *ἑόο, with the latter subsequently transforming into ἑοῖο for metrical reasons). As a consequence, scholars usually do not count *ἑέο as a genitive of a 3rd pers. pronoun from the stem *sw‑ (cf. also Beekes 1986, 366−368). 16 See e.g. DELG, GEW, EDG s.v. ἑαυτοῦ.

24  Sara Kaczko (which is very productive in the adjective ὅς).17 Relevant in this context are: Hom. ἕο (Att. οὗ, Ion. εὗ), οἷ, ἕ, ἑο (Att. οὑ, Ion. εὑ), οἱ, ἑ;18 Doric (Tarent.) ϝέο (γίο αὐτοῦ in Hesychius γ 569 Latte/Cunningham; cf. also ϝιαυτõ at IC IV 47.5); Lokrian ϝέος (gen. in IG IX 12,3 718.33, 500−450 BC); Doric ἑοῦ according to Ap. Dysc. GG 2.1.1, 77.10 Schneider; ἑοῦς in Corinna, which however was connected by Ap. Dysc. GG 2.1.1, 77.5 Schneider with the Doric dialects and other Doric forms);19 Doric (inscriptions) ϝιν, ϝίν and ϝοι,20 also Doric (literary) οἱ in Epicharmus, Pindar, Bacchylides

 17 The adjective ὅς is very productive in Homer and also in later poetry: as mentioned above, in Section 2, n. 4, the use of ὅς as a ‘free’ reflexive in Apollonius Rhodius (and Callimachus) is among the evidence (along with ‘internal evidence’ from Homer) that Brugmann 1876 cited in support of his hypothesis that there are traces in Homer of the original PIE reflexive that could refer to any person or number; cf. Rengakos 1993, 110–119. 18 These are the usual forms found in the Iliad and Odyssey (which also feature εἷο, ἑθεν, and ἕθεν), Hesiod (e.g. ἑ, οἱ, οἷ, εἷο, ἑθεν), and the Homeric Hymns (e.g. ἑ/ἕ, οἱ); ἕ and οἱ appear in Pindar (οἱ also in Bacchylides), ϝε in Alcman (ϝέθεν and ϝοι have been restored by scholars in Alcman’s text), and ϝοι and ϝέθεν in Sappho and Alcaeus (with [w] noted by γ); in inscriptions, see Thess. ϝε and Boeot. (ϝ)οι (cf. Thumb/Kieckers 1932, 116, 163; Thumb/Scherer 1959, 59; Petit 1999, 115−119). 19 Cf. Ap. Dysc. GG 2.1.1, 77.5 Schneider: ἑοῦς. αὕτη ἀκόλουθος Δωρικῇ τῇ τεοῦς ᾗ συνεχῶς καὶ Κόριννα ἐχρήσατο (and quotes Corinna 662.3 PMG). The parallel forms of the 1st and 2nd pers. are also given as both Doric and Boeotian (i.e. used by Corinna) by Ap. Dysc. GG 2.1.1, 74.10 and 75 Schneider, respectively: ἡ κοινὴ οὖσα Συρακουσίων καὶ Βοιωτῶν, καθὸ καὶ Κόριννα (fr. 682 ΡMG) καὶ Ἐπίχαρμος ἐχρήσαντο (fr. 144 Κ.-A.); ἀλλὰ μὴν καὶ τῇ ἐμοῦς Δωρίῳ ἡ , Σώφρων (fr. 59 Κ.-Α.), καὶ ἔτι Κόριννα· περὶ τεοῦς Ἑρμᾶς ποτ’ Ἄρεα πουκτεύϊ (fr. 666 PMG). The ending ‑οῦ(ς) is indeed found in West Greek dialects (see above n. 5 on West’s notes with regard to ἑοῦς in Corinna), cf. Dor. ἐμοῦς, ἐμῶς (and also ἐμεῦς, ἐμίως); Dor. τεοῦς, τεοῦ (and also τέος, τίος, τεῦς), see e.g. Kühner/Blass 1890, 597−599; Schwyzer 1939, 603−605. Many of these forms are quoted by Ap. Dysc. GG 2.1.1, 74.10−75.24 Schneider, who also provides the following forms: Boeot. ἐμοῦ, Boeot. τεῦ, Dor. ἑοῦ, Boeot. ἑοῦς (which also, in Apollonius Dyscolus’ terms, means ‘used by Corinna’); cf. Brandenburg 2005, 597–599 for a useful survey of the forms attested in Apollonius Dyscοlus’ treatise. Of course, ‑οῦς is expected in the Doric dialects of the ‘Doris mitior’, but not in Boeotian (where ‑ῶς is expected); cf. also nn. 33−34 below. 20 Both types of dative (in -οι and -ιν) occur in Crete: ϝιν (ϝιναυτõι) in IC IV 51.9, ϝίν (more precisely, ϝὶν αὐτο̃ ι) in IC IV 72, col. II 40 (the same inscription reads e.g. ϝά, ϝο̃ ν at col. III 42, col. IV 26), while ϝοι is found in IC IV 41.5−6 (αἴ κά ϝοι) and SEG 27.631B.10. ϝιν is otherwise found in a dedication by the potter Nicomachus from Metapontion, IG XIV 652 (cf. IGDGG II no. 75; other scholars have proposed ϝε, which is less likely, but would in any case attest to a stem *sw-); cf. also ϝοι in Argos, IG IV 506.8 and 11, and oἱ in Epidauros, IG IV2, 1 122.16 and 58, 4th cent. BC (= DGE 109.16 and 58; the same inscription reads, notably, οὑ and ἕθεν at ll. 2 and 106, respectively). Buck 1955, 98 stresses that in West Greek the forms in ‑οι along those in ‑ιν are mostly enclitic, though there are a few instances of orthotonic ‑οι (e.g. τοί and τοι).

Unconventional Features in Homer  25

(also with *w-),21 Stesichorus, Ibycus; Delphic, Arcadian, Cypriot ϝοι; Aeolic (Sappho, Alcaeus), ϝοι, οἰ (Boeot. οἱ, with gemination of the preceding consonant);22 Hesiod also has a dative ἵν (ἵν δ᾽ αὐτῷ at fr. 245 M.-W.) and Corinna a dative ἑΐν; Thessalian and Doric (Alcman) ϝε, Doric ἕ (Pindar).23 It appears, then, that the stem *sew‑ is in fact productive only in adjectival forms, with several instances of ἑός in archaic literary texts — mostly epic (Homer, Hesiod, Homeric Hymns) and lyric poetry (primarily Pindar, with 16 occurrences) — with sporadic contested instances in Ionic lyric, tragedy, and a few metrical inscriptions (i.e. ἑϝοῖ and ἑᾶς in Boeotian inscriptions of the 6th and 5th cent. BC, respectively).24 In later Greek, this stem is found only in Hellenistic poetry, in Theocritus and especially in the epics of Apollonius Rhodius (in which ἑός has also the value of 1st and 2nd pers.),25 Quintus of Smyrna etc.26

 ἑέ and ἑοῖ in the Homeric Poems The distribution of ἑέ and ἑοῖ in the Homeric text is relevant in several ways to the question of elements that reflect the influence of Western (and Central) Greece in the epics. Both the accusative and the dative occur in only two instances each that, moreover, give the impression that ἑέ and ἑοῖ were relatively late arrivals in the Homeric text and are tied to a well-defined context. As a matter of fact, three of the four instances of these forms occur in connection with αὐτ‑, a syntagm that is usually considered a relatively recent development:27 the dative,

 21 In Pindar (and Bacchylides) the great majority of instances of οἱ are treated as beginning with a consonant (i.e. *w‑), e.g. at Ol. 1.23, Pyth. 4.73, Isthm. 6.12. 22 I.e. οι at IG VII 2407.7 (Thebes, 4th cent. BC); οἷ at SEG 25.553.8 (Orchomenos, 4th cent. BC); on ϝοι in Sappho and Alcaeus, see above n. 18. 23 See Kühner/Blass 1890, 580–590; Schwyzer 1939, 602−604; Petit 1999, 115–124. It should be added ϝhε in Pamphylian, cf. Thumb/Scherer 1932, 184, 187. 24 CEG 444 (ca. 550 BC) and CEG 115 (ca. 450 BC) respectively; for the type ἑός in Boeotian, cf. also ἱῶν in Cor. 654 iii 34 PMG, where however [e] > [i] implies the loss of [w]. As for poetic inscriptions, cf. also hεᾶι in a 5th-cent. BC Thessalian epitaph (IG IX,2 250.2; see Blümel 1992, 268) and ἑάν in a Laconian public metrical inscription (CEG2 819.9, ca. 405−340 BC). 25 Cf. Rengakos 1993, 116–119 and below at Section 4. 26 See Kühner/Blass 1890, 601–602; Schwyzer 1939, 602−604; Buck 1955, 98; Petit 1999, 174– 202. 27 Cf. Kühner/Blass 1890, 596–598; Wackernagel 1928, 89; Schwyzer 1939, 605; Chantraine 1961, 141−142; Rix 1992, 180−181; Petit 1999, 207–211; on the late character of the Homeric instances in which this syntagm occurs, see Shipp 1972, 78.

26  Sara Kaczko ἑοῖ αὐτῷ, at Il. 13.495 and Od. 4.38, and the accusative, ἑὲ δ᾽αὐτόν, at Il. 20.171. This impression is strengthened by other peculiarities in the passages in which the dative appears. First of all, ἑοῖ in both cases seems to be connected with the verb ἕπομαι, i.e. ὡς ἴδε λαῶν ἔθνος ἐπισπόμενον ἑοῖ αὐτῷ at Il. 13.495 and ὀτρηροὺς θεράποντας ἅμα σπέσθαι ἑοῖ αὐτῷ at Od. 4.38 (which, remarkably, is in enjambement with the preceding verse); moreover, λαῶν ἔθνος (Il. 13.495) is a unicum; and, finally, ἑοῖ at Il. 13.495 is treated as if it began with *w‑, since the last syllable of ἐπισπόμενον scans long: parallels for forms treated as beginning with consonantal *w‑ are to be found in the adjective ἑός,28 although in this case *w‑ may be the result of analogy of other forms of the pronoun with *sw‑ (see below, Section 4). Notably, forms of the middle participle ἐπισπόμενος (‑οι, ‑ον) always occur in this position in the hexameter, after the ‘feminine’ caesura: in the just quoted Il. 13.495 and six times in the Odyssey; cf. also the infinitive ἐπισπέσθαι, after the feminine caesura, at Il. 14.521.29 It should be stressed that the other occurrence of the accusative ἑέ — the only instance in which neither ἑέ nor ἑοῖ appear in connection with αὐτ‑ — points in the same direction regarding the origin of these disyllabic forms: ἑέ at Il. 24.134 is essentially repeated, with the sole variation of being ‘translated’ into the 3rd person a few lines below the ‘original version’ in the 1st person, so that it occupies the same metrical position as ἐμέ.30 These lines contain the well-known passage in which Zeus summons Thetis and orders her to demand that Achilles accept Priam’s ransom for Hector’s body (because he, like most of the other gods, is displeased with Achilles’ treatment of it) and she delivers the message to her son: σκύζεσθαί οἱ εἰπὲ θεούς, ἐμὲ δ’ ἔξοχα πάντων / ἀθανάθων κεχολῶσθαι (Il. 24.114) and σκύζεσθαί σοί φησι θεούς, ἑὲ δ’ ἔξοχα πάντων / ἀθανάθων κεχολῶσθαι (Il. 24.134). It may also be observed that several scholars have identified a relatively high concentration of late features in Il. 24, the poem’s last book.

 28 See e.g. Janko 1992, 13, 110; Petit 1999, 179−181, 183−185. 29 Respectively Od. 3.215 = 16.96, 14.262 = 17.431, 16.426, 24.183, 24.338. The infinitive ἐπισπέσθαι at Il. 14.521 (ἐπισπέσθαι ποσὶν ἦεν / ἀνδρῶν τρεσσάντων in enjambement) occurs in the same position in the hexameter as ἅμα σπέσθαι at Od. 4.38. 30 According to Shipp 1972, 77−78, this fact, along with other considerations, makes it possible that ἑέ and ἑοῖ are artificial and at least indicates that there is no support for an ancient ἑέ in Homer.

Unconventional Features in Homer  27

 ἑέ, ἑοῖ, and the Greek Dialects Given these considerations, it seems clear that ἑέ and ἑοῖ entered the Homeric diction at a later stage. This impression is strengthened by their prosody: these alternative forms were disyllabic and therefore, at least theoretically, very useful to the bards; however, they are seldom used in the poems. In comparison to determining the relative antiquity of these ‘continental features’ in the Homeric poems, it is considerably more difficult to determine the dialect in which they originated. West, in arguing that ἑοῖ and ἑέ are among the elements that are inconsistent with the opinio communis regarding the formation and fixation of the Homeric text, identified as possible parallels the 2nd pers. sing. τύνη, τεοῖο, τεΐν and the 3rd pers. ἑΐν, ἑοῦς.31 In essence, two phenomena point to a West Greek (and/or Boeotian) origin of these latter forms, the ‘strong stem’ and the endings — for the former, *tū‑, *tew‑ in the case of the 2nd pers. and *sew‑ in the case of the 3rd pers.; for the latter, τύνη with the suffix ‑νη, τεΐν with the ending ‑ιν,32 both mentioned by West, and possibly τεοῖο, if the interpretation

 31 The parallels between forms that featured the ‘strong stem’ *tew‑ and *sew‑ and those found in Corinna (but also in Doric) were already stressed by Wathelet 1970, 288–289, who however interpreted them as Aeolic elements in Homer, on the assumption that the occurrences in Boeotian represent traces of Aeolism. On the contrary, West specified that ἑίν and ἑοῦς at Corinna frr. 662 and 681 PMG should, since they exhibit proper West Greek endings alongside the ‘strong stem’, in turn be interpreted as “Boeotisms of West Greek origin”. 32 With regard to the dative ending — which incidentally is found in the plural, the type Dor. ἁμίν, Att.-Ion. ἡμῖν and is usually compared with Avest. kahmi, ahmi; OInd. káhmi, áhmi — and parallels of τείν, see e.g. Schwyzer 1939, 604−605; Buck 1955, 98. The 1st pers. sing. ἐμίν is found in literary texts (Epich. fr. 97.9 K.-A.; cf. also frr. 246.1 and 250 K.-A.; Sophr. fr. 96 K.-A.; Ar. Av. 928, 930; Ach. 733; Theoc. 2.144, 156, 3.3 etc.; Callim. Hymn 6.116) and in an inscription from Rhodes (DGE 276a, 5th cent. BC); τίν is found in Pindar (e.g. Ol. 5.7, Pyth. 1.28); Bacchyl. 4.14; Alcm. fr. 60.1 PMG; Cor. fr. 663 PMG; doubtful at Ar. Av. 930; frequently in Theocritus; the 3rd pers. is found only twice in Pindar and then in the above-mentioned Hesiod’s fragment (ἵν δ᾽ αὐτῷ at fr. 245 M.-W., from the Catalogue of Women, which reflects the interests of areas of Northcentral Greece [cf. Cassio 2009, 197]); ϝίν, ϝιν are found in inscriptions (on which cf. above, Section 2); other forms are known only from Apollonius Dyscolus’ De pronominibus, e.g. ἐμίνη, τίνη, given as Tarentine, used by Rinthon (cf. Ap. Dysc. GG 2.1.1, 81.20 Schneider).

28  Sara Kaczko of it as an ‘epic embellishment’33 of τεοῦ with the ending ‑ου (with parallels in West Greek dialects)34 is to be accepted. Compared with the scenario just recalled for τεοῖο and τεΐν, that for ἑέ and ἑοῖ is more ambiguous, because neither the stem nor the endings point to just one specific dialect. With regard to the endings, ἑέ would be at home in any dialect, and -οι in ἑοῖ is also not as strongly associated with West Greek dialects as is ‑ιν. The ‘strong stem’, theoretically speaking, does not necessarily point exclusively to a West Greek (or Boeotian) origin, nor should it necessarily even be tied to a single dialect and/or dialect group. Thus, the stem could be traced back to PIE *sew‑ or alternatively be an independent innovation (especially considering the existence of the disyllabic adjective ἑός), having arisen through some other type of formation, such as a proportional mechanism or as a genitive of the adjective used as the genitive of the personal pronoun (and the adjective ἑός is relatively well attested). A further possibility is an independent development in more than one area, where the usual forms from *sw‑ (οὗ, οἷ, ἕ, οὑ etc.) were used. For example, in the case of ἑοῖ, the modification may have been triggered by an ‘unbalanced system’ in which the other forms began with hϝε‑, i.e. the genitive hϝεο and accusative hϝε (Att.-Ion. ἕο/ἑο, ἕ/ἑ), but the dative with hϝ‑, i.e. hϝοι (Att.-Ion. οἷ/οἱ), which would also explain why ἑοῖ at Il. 13.495 is treated as if it began with *w‑.35 Also, the coexistence with the type οὗ, οἷ, ἕ, οὑ etc. (< *sw‑) would not be an issue, for such situations are fairly common with pronouns36 and, returning to

 33 I.e., as Cassio 2018b puts it, some sort of ‘poetic embellishment’ of either τεῶ or τεοῦ: more precisely this “epic ‑οιο form” would be the outcome of the usage of τεοῦ in West Greek dialects and of ‘hypercorrect’ forms of the genitive of the possessive adjectives used in the place of genitives of the pronouns in Homer, probably owing to the generalization of Aristarchus’ prescription with regard to the choice between the genitive of the adjective (i.e. ἐμοῖο) and of the pronoun (i.e. ἐμεῖο). I will return to this question in the following section. 34 The ending ‑οῦ(ς) is indeed found in West Greek dialects; see above n. 19. Cf. also ἐοῦ at Hes. Theog. 401. 35 However, this phenomenon may have another explanation, i.e. not necessarily ‘morphological’, as the parallel forms treated as beginning with consonantal *w- found in the adjective ἑός (in 18 instances) seem to suggest (Janko 1992, 13, 110; but see Petit 1999, 184). It should be noted that also at Il. 20.171 μαστίεται, ἑὲ δ’ αὐτὸν ἐποτρύνει μαχέσασθαι, the final syllable of μαστίεται counts as a longum before ἑέ. 36 Cf. the co-existence of the indefinite pronoun τō (i.e. του < τεο in the old Attic alphabet) and τινος, τεῳ and τινι in Attic inscriptions down to the 4th cent. BC (Threatte 1996, 340−342) and of τινι and τεῳ (the latter more frequent), τινος and τευ (the latter more frequent, cf. also an instance of τεο) in Herodotus (Powell 1938, s.v. τις) and Thucydides (scil. τινος and του, τινός and τοῦ), also for the interrogative; cf. also egli and lui in Italian (similarly el and lo in Tuscan dialect).

Unconventional Features in Homer  29

the example of the forms just referred to, ἑΐν, ἑοῦς, τεοῖο etc. have been associated with dialects of Central and Western Greece, where the more frequent pronouns were also those in *sw‑ (see Section 2 above). As for the endings, i.e. the second possible hint at a West Greek origin of the disyllabic forms, it must be noted that there is a difference between the 2nd pers. sing. (τεΐν) and the 3rd pers. sing.: the dative form of the latter in Homer is ἑοῖ, not ἑΐν, which, it should be stressed, would have been metrically possible at both Il. 13.495 and Od. 4.38. Thus, the 3rd pers. does not have the ending ‑ιν. By contrast, as mentioned, both τεΐν, which is found in the Homeric poems, and ἑΐν are used by Corinna (fr. 662.1 PMG, the source being Ap. Dysc. GG 2.1.1, 77.5 Schneider) and exhibit the ending ‑ιν, which is also found in the Doric dialects, cf. the forms ἐμίν, τίν, (ϝ)ίν that appear in literature (Hesiod, Aristophanes, Alcman, and Pindar; cf. Theocritus, who always uses ‑ῑν) and inscriptions.37 However, in West Greek dialects, pronominal forms in ‑οι are possible and indeed attested, though they are usually (but not exclusively) enclitic, while ἑοῖ is orthotonic (see further below).

 τεΐν vs. ἑοῖ Given this scenario, it may be helpful to try to explain such differences, i.e. the usage of τεΐν vs. the absence of a corresponding ἑΐν as the 3rd pers. dative, for which only ἑοῖ is found in Homer. In addition, it would be important to ascertain whether the ‘strong stem’ is to be tied to West Greek dialects or may have another possible origin. Regarding the first point, it has been observed that τεΐν occurs in a specific position in the line, i.e. exclusively after a ‘feminine’ caesura, and that this pattern should be linked to the instances of τεός and the perception of that personal adjective. According to many scholars, τεός and, along with it, ἐμός and σός (the usual form of the 2nd pers. sing. personal adjective) found in Homer were (probably early) replacements of ancient ϝός and ἑϝός, which originally could be used for the 1st, 2nd, and 3rd pers.38 It must be noted that forms like τεή, τεῆς, τεήν,

 37 Cf. above at n. 32. 38 Brugmann 1876; Chantraine 1958, 273; Nussbaum 1998, 87–129.

30  Sara Kaczko τεῇ, τεῇς occur rarely, exclusively after a ‘feminine’ caesura, which is on a general level the preferred position in the majority of instances of τεός.39 This being the case, τεός — which is much less frequent than σός and is certainly not AtticIonic, but rather either Aeolic or Doric — has traditionally been interpreted as an Aeolic form (possibly Lesbian, perhaps also Thessalian) used for metrical necessity, i.e., as mentioned, to substitute for an ancient ‘free reflexive’ personal adjective used for all persons and numbers. However, according to Cassio, τεός was possibly reinterpreted as a ‘Doric’ or ‘continental’ form in a later phase by rhapsodes in Euboea, given its proximity to Boeotia (and Lokris), and this reinterpretation, along with the preference for the pronoun in *tew‑ in certain contexts in Homer, would have legitimized the presence of the newly arrived τεΐν in the poems.40 More precisely, the fact that τεΐν, in all five of its Homeric instances,41 as well as in two archaic epigrams from the temple of Apollo Ismenios at Boeotian Thebes, quoted by Herodotus (5.60.1−5.61.1), and in a passage of the ‘Doric’ Hymn to Athena by Callimachus (5.37),42 invariably occurs in the same position of the line (after the ‘feminine’ caesura) followed by a consonant, and in closed sentences (i.e. never in enjambement), has led Cassio (2018a) to postulate that it is a borrowing, not from a spoken dialect, but rather “from a ready-made hexameter ‘poetic frame’ that originated either in Boeotian or Doric areas”. If this scenario is plausible and τεΐν is a newcomer, tied to a specific situation, with a specific origin — i.e. a ‘poetic borrowing’ that occurs in certain conditions — the presence of a 2nd pronominal pers. in ‑ιν may not imply that, for the dative of a ‘Boeotian’ or ‘Doric’ 3rd pers. pronoun, only a form ἑΐν is to be expected (or can even provide any conclusive clues as to the origin of ἑοῖ). As mentioned, several West Greek dialects show a tendency to oppose an orthotonic ‑ιν to an enclitic -οι, e.g. ἐμίν vs. μοι, but orthotonic forms in ‑οι such as τοί are also

 39 Cf. the same position after a ‘feminine’ caesura of e.g. τεόν in thirteen out of its total nineteen instances, and of τεῷ at Il. 14.219 (τεῷ occurs only once in the just quoted Il. 14.219; cf. the declension of the last hemistich at Il. 14.223, featuring ἑῷ); see, however, τεὼ ἵππω || at Il. 5.230, 237 and || ὀφθαλμοῖσι τεοῖσιν at Od. 3.94, 4.324. 40 Cassio 2018b stresses that τεός is considered a Doric feature in Homer by Ap. Dysc. GG 2.1.1, 105.31–106.13 Schneider (Δωρικῶς δὲ τὸ πατὴρ τεός, Od. 16.188), and is also used by Aeolian poets (καὶ παρ’ Αἰολεῦσιν. Ἀλκαῖος ἐν πρώτῳ· τὸ δ’ ἔργον ἀγήσαιτο τέα κόρα [fr. 310.1 L.-P.]); the same Doric nature is also explicitly claimed for τεοῖο at Ap. Dysc. GG 2.2, 217.17–218.1 Uhlig τεοῖο, Δωρικώτερον μετατεθέντος τοῦ ς εἰς [τὸ] τ καὶ ἐπενθέσεως τοῦ ε γενομένης, notably not only for having been recharacterized by ‑ς, but also for the presence of an ‘epenthetic ε’, i.e. a ‘strong stem’, in West’s terms. 41 Il. 11.201; Od. 4.619 = 15.119; Od. 4.829; Od. 11.560. 42 Callimachus’ Hymn to Athena is in literary Doric, with epic elements, and in elegiacs.

Unconventional Features in Homer  31

to be found (cf. n. 20 and Buck 1955, 98). Therefore, an ἑοῖ featuring the ‘strong stem’ and the ending ‑οι, though not strongly marked as West Greek, is definitely a possibility in this dialect group — although, again, this does not necessarily rule out other dialects. It has, then, been established that the presence of τεΐν does not mean that the form of the 3rd pers. dative should be (exclusively) ἑΐν, that ἑοῖ is possibly a West Greek form, and that the ‘strong stem’ seems to point to a West Greek origin for the above-mentioned pronominal forms, though these considerations are neither necessarily decisive nor definitely imply the exclusion of other dialects. At this point, it is worth examining other instances of disyllabic pronouns found in the epics, such as ἑοῦ (Hesiod and vv.ll. in Homer) and τεοῖο, in order to ascertain their origin with regard to the Greek dialects and, more importantly, whether they provide a hint regarding the origin of ἑοῖ and ἑέ in Homer. This analysis will offer a novel perspective on a subject that is usually taken for granted.

 Genitives in *sew‑ and *tew‑ in Homer To begin with the 3rd pers., the only pronouns that directly attest to a stem *sew‑, apart from the forms quoted by West (ἑέ, ἑοῖ in Homer and ἑίν, ἑοῦς in Corinna), are the genitive ἑοῦ at Hes. Theog. 401, a few variae lectiones of ἑοῦ in Homer cited by Zenodotus,43 and a (possible) restored instance of ἑοῖ in Empedocles. Otherwise, only the adjective ἑός appears in Homer, Hesiod, Pindar etc. Regarding Hes. Theog. 401, παῖδας δ’ ἤματα πάντα ἑοῦ μεταναιέτας εἶναι, it must be noted that one MS (b) reads ἑοῦς, which, if interpreted as a genitive in ‑ς of the personal pronoun, would attest to a form from the stem *sew‑, just like ἑοῦ, and would suggest one of the forms of “West Greek origin” (in West’s terminology) found in Corinna.44 West (ad Hes. Theog. 401), however, interpreted the variant transmitted by the MS as ἑούς, i.e. as the accusative plural of the adjective ἑός and then discarded it as a lectio facilior. In addition, the alternative reading ἕο proposed by Brugmann is usually rejected because instances in which [m‑] (μ‑) is treated as [mm‑] (μμ‑), although well attested in Hesiod, never occur with μετα‑. In other words, it seems that Theog. 401 attests to a genitive ἑοῦ from the stem *sew‑; interestingly, it is followed by a so-called ‘short accusative’ in ‑ας, a feature found in  43 I list here the relevant examples: Il. 2.239, Il. 19.384, Od. 7.217: see e.g. schol. A Il. 2.239 Erbse and schol. Aim Il. 19.384 Erbse; cf. Düntzer 1848, 58; La Roche 1866, 252. 44 See Section 4 above for the distribution of genitives of the personal pronoun in ‑ς from the genitive of the nouns, e.g. ἐμεῦς, τεοῦς in various dialects.

32  Sara Kaczko ‘Doric’ or Western Greek poetry (Tyrtaeus, Alcman, Stesichorus, the Corinthian epigram CEG 131 etc., as well as Hesiod).45 At any rate, the presence of a West Greek form in Hesiod is not surprising. On the other hand, aside from ἑέ and ἑοῖ discussed here, the situation in Homer is more complex also because there is a single instance metrically guaranteed of a pronoun with the disyllabic stem (*tew‑ in this case), the genitive τεοῖο at Il. 8.37 = 8.468 (on which see further below); otherwise, it seems that genitives of the personal pronoun attest only to the stems *tw‑ and *sw‑. However, some variae lectiones of the 3rd pers. genitive may reveal an interesting perspective. Indeed, for the stem *sew‑, a possible indication of ἑοῦ as a 3rd pers. genitive is Zenodotus’ reading of ἑοῦ instead of ἕο at Il. 2.239 and at Il. 19.384 (see schol. A Il. 2.239 Erbse and Aim Il. 19.384 Erbse).46 Scholars, however, are generally skeptical about these readings.47 Thus the v.l. ἑοῦ for ἕο at Il. 2.239, ὃς καὶ νῦν Ἀχιλῆα ἕο μέγ’ ἀμείνονα φῶτα, is not metrically guaranteed, since the genitive ἕο scans with the well-attested treatment of [m‑] (μ) as [mm‑] (μμ) in μέγα. Turning now to Il. 19.384, the way in which the line is transmitted in the MSS, πειρήθη δ’ ἕο αὐτοῦ ἐν ἔντεσι δῖος Ἀχιλλεύς, makes ἕο apparently more problematic because it requires a hiatus: the verse has been printed by West (following Heyne’s conjecture) as πειρήθη δὲ ἕ’ αὐτοῦ ἐν ἔντεσι δῖος Ἀχιλλεύς, in which ἕο is treated as if it began with *w‑. It should be observed that this passage, which seems to allow more easily for a reading of a form of the stem *sew‑ (especially in comparison with Il. 2.239 and Od. 7.217), features a ‘recent’ syntagm in αὐτ‑ (either ἕο αὐτοῦ, restored by scholars, or ἑοῦ αὐτοῦ read by Zenodotus), as is the case with ἑοῖ αὐτῷ at Il. 13.495 and Od. 4.38 and ἑὲ δ᾽ αὐτόν at Il. 20.171, which were quoted above. Interestingly, it seems that Zenodotus preferred a genitive in ‑ου for the 3rd pers. pronoun whenever possible.48 So, for example, he apparently read oὗ for ἕο also at Il. 20.261 and Od. 5.549 — where, however, a monosyllabic form seems unnecessary or even unlikely (thus most scholars read *ἀπὸ ϝϝέο at Il. 20.261) — and oὗ for εὑ at Il. 24.293 = 311 καὶ εὑ κράτος ἐστὶ μέγιστον (the formulaic pattern τοῦ γὰρ / ὅ τε / ὅου / οὗ τε / καὶ εὑ κράτος ἐστὶ μέγιστον appears sixteen times in  45 Cf. Cassio 2018b; on ‘Doric features’ in Hesiod see in general Morpurgo Davies 1964 and Cassio 2009 (esp. p. 200), with previous bibliography. 46 On Düntzer’s (1848, 58) unconvincing attempt to attribute another reading ἑοῦ to Zenodotus at Od. 7.217, see below n. 54. 47 For example, according to Petit 1999, 163, it would be possible to connect ἑοῦ to a disyllabic stem, but it would be uneconomical to admit a disyllabic ἑοῦ on the basis of merely two instances, both of which admit another reading, i.e. ἕο; cf. also Chantraine 1958, 148; 165 and West’s edition of the Iliad, which reads ἕο in both passages. 48 Cf. Düntzer 1848, 58; Rengakos 1993, 113.

Unconventional Features in Homer  33

the Iliad and the Odyssey). These latter are both instances in which a monosyllabic form is indeed possible, since in Homer adnominal possessives of the genitives of the 3rd pers. and of the personal pronouns “might be an innovation that is comparatively recent as an epic feature”,49 though there is no need to correct εὑ to oὗ. The forms with Zenodotus’ authority, i.e. ἑοῦ and oὗ, are mirrored (along with others) in Apollonius Rhodius, with ἑοῦ at Argonautica 4.802 and oὗ at 1.361−362 and 4.1470−1471.50 These instances, though not necessarily probative in themselves, are important because they acknowledge the existence in the preAlexandrian vulgata of alternative readings — i.e. those chosen by Zenodotus as opposed to those supported by Aristarchus — and of the philological problem they posed (which were also ‘dealt with’, in their peculiar way, by some Alexandrian poets). This is not the place to discuss the question of oὗ, though significant as it is,51 but it is worth returning to the varia lectio ἑοῦ. It must be recalled that in some Doric dialects, the genitive of the personal pronoun has the same appearance, and may even be the same form (at least, this is the commonly accepted explanation), as the genitive of the personal adjective, i.e. the 1st pers. sing. ἐμοῦ, 2nd pers. sing. τεοῦ,52 and 3rd pers. ἑοῦ are all genitives (masculine and neuter) of the possessive adjective, but also genitives of the personal pronoun53 (notably, in some dialects of the same group the genitive of the personal pronoun was secondarily recharacterized by a ‑ς, i.e. ἐμοῦς). The mechanism has several parallels — for example, in Latin the genitives mei, tui, sui, nostrum, and vestrum, in  49 Nussbaum 1998, 126–127; for Hainsworth 1993, 304, on the other hand, “Zenodotus’ reading is unnecessary, but originally it may have been καὶ ἕō κράτος”, which is also West’s reading in his edition of the Iliad. 50 See Rengakos 1993, 112–113. The same selection of the variant supported by Zenodotus is found for instances of ἐμεῖο vs. ἐμoῖο in Apollonius Rhodius; see Rengakos 1993, 111 and in detail below. 51 The readings of oὗ advocated for by Zenodotus are possibly consistent with his tendency to prefer more recent variants, ones that could be attributed to the final stages of the tradition or to an already (almost) fixed diction of the Homeric poems. This is not the place to dwell on the question (it will be addressed elsewhere), but it should be noted that among these instances of vv.ll. of oὗ, only a few are actually likely, and some of them may cover for late Ionic, ‘Doric’, or even be Attic forms. 52 See for example Callim. Hymn 6.98 (Callimachus’ Hymn to Demeter is as well in literary Doric with epic elements), featuring τεοῦ, 2nd pers. gen. (masc./ntr.) of the possessive adjective, but also genitive of the pronoun. 53 Cassio puts forward also the alternative hypothesis that the personal pronoun *te-so underwent consonantization, then secondary lengthening (such as in the ‘Doric futures’, cf. Méndez Dosuna 1993, esp. 118–119), and finally restoration of [e] on the analogy of τε, τεός (i.e. *te-ho > *tjõ > τεοῦ). At any rate, the outcome would be the same: the genitive of the personal pronoun would coincide with that of the possessive adjective in the masculine and neuter forms.

34  Sara Kaczko OInd. asmā́kam, yuṣmā́kam are the genitives of the adjectives.54 In other words, in the just quoted Homeric passages, Zenodotus apparently read ἑοῦ, i.e. a form that worked both as the genitive of the possessive adjective (masculine and neuter)55 and as genitive of the 3rd pers. pronoun, and that for us is a Doric/West Greek form. This possibly means that some forms, such as ἑοῦ, started circulating at a later stage in the epics, though they left only some traces, namely Zenodotus’ readings and some echoes in Apollonius Rhodius. Such a possibility has been linked to another phenomenon that is at work already in the early stages of the Greek language, i.e. the progressive replacement of the possessive adjective with the genitive of the personal pronoun.56 Generally speaking, in possessive expressions the genitive of the pronoun represents a possible alternative to the adjective. This much is evident in Homer, cf. e.g. δόμος Ὀδυσήϊος, Νεστορέη νηῦς, Τελαμώνιος Αἴας, Ἀντίλοχος Νηλήιος vs. δόμος Ὀδυσῆος, ἄλοχος Σθενέλου, θυγάτηρ + gen. (in fact, in Homer, the genitive is the only way to express possession with the noun θυγάτηρ), though it is usually deemed a more recent feature: for example, βίη Ἡρακληείη and Αἰολίη νῆσος in Homer are interpreted as having been inherited from the tradition, as opposed to similar periphrases with the genitive, Πριάμοιο βίη, νῆσος Ἠελίοιο.57 The same  54 In Germanic languages, parallels from formal and morphological perspectives as well as reciprocal influence between the genitive of the personal pronoun and the possessive are also to be explained in terms of the concurrence of the two; thus, for example, the genitive of the personal pronoun at later stages is often the genitive of the possessive (neuter), Old Swiss várra, Old Eng. mines etc.: see Seebold 1984, 47–48. 55 Düntzer 1848, 58 (followed by La Roche 1866, 252) proposed, not convincingly, attributing another reading of ἑοῦ to Zenodotus at Od. 7.217, ἔπλετο, ἥ τ’ ἐκέλευσεν ἕο μνήσασθαι ἀνάγκῃ. The scholium actually reads: Ζηνόδοτος ἕο γράφει ἀντὶ τοῦ ἑαυτῆς, but Düntzer proposed correcting it into ἑοῦ for the sake of consistency with Zenodotus’ other readings of ἑοῦ when ἕο is possible. Not only is it not metrically guaranteed, since the genitive ἕο scans with the well-attested treatment of [m‑] (μ) as [mm‑] (μμ) in μνήσασθαι, but moreover ἑοῦ as a personal adjective (used as genitive of the personal pronoun) is only masculine and neuter (the feminine genitive is ἑᾶς). 56 The progressive expansion of the pronominal expression of the possessor by means of the adnominal genitive at the expenses of adjective is clear in the LXX and completed in Modern Greek, where the former has ousted the latter: possession is expressed through the (enclitic) genitive of the pronoun, e.g. πατήρ μου etc. 57 Wackernagel 1928, 68−69, 80. As for the use of adjectival forms for specific kinds of relationships, such as familial, the selection of the patronymic is generally considered an Aeolic trait and indeed in Boeotia and Thessaly is regularly found until the 4th cent. BC, and examples are found still later especially in those regions (3rd and 1st cent. BC); in East Aeolic adnominal genitives also appear already in the 5th cent. BC, but the patronymic is very well attested, again, well into the Hellenistic period; cf. Scarborough 2016, 110−111 with previous bibliography.

Unconventional Features in Homer  35

applies to the adnominal possessive of the pronoun, which is less frequent than the adjective (and of the enclitic dative/genitive of the type μοι, σοι, οἱ) and indeed in its phonology and metrical shape betrays its more recent nature.58 In this context, it has been observed that there are many instances in the MSS of variae lectiones of the genitives of the pronoun, usually readings of Zenodotus,59 as opposed to those of the genitive of the possessive adjective, which are usually readings of Aristarchus, such as, e.g., πατρὸς ἐμοῖο at Il. 14.118 (τρίτατος δ’ ἦν ἱππότα Οἰνεὺς / πατρὸς ἐμοῖο πατήρ), on which see schol. A Erbse οὕτως Ἀρίσταρχος. Ζηνόδοτος δὲ “ἐμεῖο”, οὐκ ὀρθῶς; schol. Aim Erbse οὕτω Ζηνόδοτος. τινὲς δὲ “ἐμοῖο”, ἵν’ ᾖ κτητικός.60 This ‘tendency’ is to be seen in the usage of genitives of the personal adjective instead of those of the pronoun even in other functions of the genitive, as for example in Il. 3.97 κέκλυτε νῦν καὶ ἐμεῖο (ἐμοῖο H), Od. 4.39 μαρνάμενοι περὶ σεῖο (σοῖο AGHI),61 and is probably the outcome of a generalization of Aristarchus’ ‘prescription’ to use the genitive of the personal adjective instead of the genitive of the personal pronoun. In light of these considerations, a different explanation of τεοῖο has been recently proposed. The form has been problematic for ancient and modern scholars alike; thus Heyne, for example, conjectured *τεεῖο, and Chantraine (1958, 265) tried to explain τεοῖο as the genitive of the possessive adjective (i.e. τεός) used as the genitive of the personal pronoun (an explanation that would not be problematic in itself, as seen above regarding mei, tui etc.). Cassio (2018b) argues that the peculiar τεοῖο should be interpreted as having arisen from a situation in which a form of the personal adjective could be used as, or coincided with, the genitive of the pronoun, i.e. τεοῦ (or τεῶ, depending on the dialect involved), and having then been ‘epically embellished’ into ‑οιο, on the basis of Aristarchus’ preference for the genitive of the personal adjective, i.e. the type ἐμοῖο, over the genitive of  58 See Nussbaum 1998, 126−127, with nn. 101−103 and further below. 59 As mentioned above, Rengakos 1993, 111 argues that Zenodotus’ readings πατρὸς ἐμεῖο vs. Aristarchus’ πατρὸς ἐμoῖο were the model for Apollonius Rhodius’ Argonautica, where ἐμεῖο is found three times, exclusively with πατρός. 60 It is then possible that Zenodotus’ readings of ἑοῦ instead of ἕο at Il. 2.239 and Il. 19.384 are to be explained exactly in this context, i.e. that they betray a use of the genitives of the pronoun that coincides with the genitive of the personal adjective or, more likely, of genitives of the personal adjective used as the genitives of the pronoun or coinciding with the form of the genitives of the pronoun. This situation would be consistent with the fact that Zenodotus read ἕο at Od. 7.217 (see the scholium ad loc.: Ζηνόδοτος ἕο γράφει ἀντὶ τοῦ ἑαυτῆς), where, on the contrary, the pronoun is feminine (ἀντὶ τοῦ ἑαυτῆς); see also above on Düntzer’s unlikely conjecture ἑοῦ. 61 Regarding the many textual variants of ἐμοῖο for ἐμεῖο and, similarly, though less frequently, of the 2nd pers. pronoun σοῖο for σεῖο, also in non-possessive functions, cf. Brugmann 1885, 409−410 (on the 3rd pers. and the choice between ἑῆος and ἑοῖο, see Nussbaum 1998, 104–129).

36  Sara Kaczko the pronoun, i.e. the type ἐμεῖο. If this is a valid explanation, τεοῖο may, in a way, presuppose a genitive τεοῦ (or τεῶ) or, at any rate, a form with the ‘strong stem’ *tew‑ to be connected with West Greek dialects: in other words, τεοῖο may have originated in a situation in which the type in ‑οιο was used as an ‘epic’ genitive of the pronoun, i.e. τεοῖο instead of τεοῦ or τεῶ. To sum up, it has been proposed that τεοῖο would be the outcome of both the use of τεοῦ in West Greek dialects62 and of ‘hypercorrect’ forms of the genitive of the possessive adjectives used in the place of genitives of the pronouns in Homer (e.g. πατρὸς σοῖο instead of πατρὸς σεῖο at Il. 24.286, probably owing to the generalization of Aristarchus’ preference with regard to the choice between the genitive of the adjective ἐμοῖο and of the pronoun ἐμεῖο). Indeed, the use of possessive adjectives in the place of genitives of the pronouns in Homer may offer a decisive clue as to the interpretation of the ‘aberrant’ pronominal forms, such as τεοῖο, τεΐν, ἑέ and ἑοῖ, for which an alternative explanation — namely their origin in dialects or poetical traditions from Central/Western Greece — is to be preferred over the traditional explanation.

 Datives in *sew- in Homer and Greek Dialects Turning now to other instances of disyllabic forms, in this case of the dative ἑοῖ, the only direct occurrence would be in Empedocles B 28.1 D.-K. (ἀλλ’ ὅ γε πάντοθεν ἶσος καὶ πάμπαν ἀπείρων / Σφαῖρος κυκλοτερὴς μονίῃ περιηγέι γαίων), though, as mentioned, this form is a restoration by Maas (others have proposed ).63 Moreover, given the influence exerted by Homer and the context in which Empedocles was active, it would be impossible to decide to which dialect or poetical tradition the postulated should be ascribed.

 62 As mentioned, τεοῦ is found in Sophron and Epicharmus, according to Ap. Dysc. GG 2.1.1, 75.16–26 Schneider, but is not attested in Homer, and neither is τεοῖο; cf. also ἐοῦ at Hes. Theog. 401. 63 The form replaced with or in Stobaeus’ text reads ἐών, and ἑοῖ is restored on the basis of Empedocles B 29.5 D.-K. where the text reads ἀλλὰ σφαῖρος ἔην καὶ ἶσος ἑαυτῷ. If a dative form of the pronoun is to be restored at B 28.1 D.-K., οἷ would also be theoretically possible, but the restoration of ἑοῖ is possibly due to the fact that in Empedocles *w‑ is usually neglected, surprisingly even when there is a variation of a Homeric model or in cases in which in Homer *w‑ is observed, such as παντὸς ἰδόντων || at B 39.5, ἀθρεῖ μὲν γὰρ ἄνακτος ἐναντίον ἁγέα κύκλον at B 47 (cf. also B 100.8, B 112.2, B 114, B 117.3 D.-K.), aside from περὶ ῥηγμῖνι βίοιο at B 20.5 πλάζεται ἄνδιχ’ ἕκαστα περὶ ῥηγμῖνι βίοιο (where, again, *w‑ is not observed in

Unconventional Features in Homer  37

As for the indirect instances of ἑοῖ in the context of vulgatae — but turning now to the ‘modern’ vulgata with regard to the reflexive pronoun Att. ἑαυτοῦ, Ion. ἑωυτοῦ — it was observed above that ἑοῖ αὐτῷ is found in Homer and that most handbooks and etymological dictionaries base the formation of the pronoun Att. ἑαυτοῦ, ἑαυτῷ, ἑαυτόν, Ion. ἑωυτοῦ, ἑωυτῷ, ἑωυτόν64 precisely on this syntagm (and, in so doing, indirectly presuppose a form ἑοῖ in those dialects).65 Thus, for example, Schwyzer (1939, 607) and Rix (1992, 181)66 held that the Att. ἑαυτοῦ/Ion. ἑωυτοῦ type was derived from the contraction of ἑοῖ αὐτῷ and extension of the resulting ἑαυτ‑ and ἑωυτ‑ to the other cases (cf. Chantraine 1961, 141–142). Chantraine (DELG s.v. ἑαυτοῦ) speaks more cautiously of “groupes comme ἕο αὐτοῦ > ion. ἑωυτοῦ, att. ἑαυτοῦ, ἑοῖ αὐτῷ > ion. ἑωυτῷ, att. ἑᾱυτῷ”,67 while both GEW and EDG, the more recent etymological dictionaries of Greek, s.v. ἑαυτοῦ state that the type ἑαυτοῦ (Ion. ἑωυτοῦ) derives from the univerbation of the reflexives ἕ, ἑοῖ etc. and αὐτόν, αὐτῷ etc. (i.e. ἕο αὐτοῦ > Att. ἑαυτοῦ and Ion. ἑωυτοῦ, ἑοῖ αὐτῷ > Ion. ἑωυτῷ and Att. ἑαυτῷ etc.). Therefore, conventional reconstructions of the type ἑαυτοῦ (Ion. ἑωυτοῦ) imply the existence of ἑοῖ and also its coexistence with ἕο in Ionic and Attic.68 In other words, the two conclusions just stated with regard to ἑοῖ and ἕο are the consequence of accepting either of the two traditional hypotheses regarding the origin of the type Att. ἑαυτοῦ etc., Ion. ἑωυτοῦ etc., namely, that this type originated from either (a) the univerbation of ἕο αὐτοῦ and ἑοῖ αὐτῷ or (b) the remodeling of ἑαυτοῦ, ἑωυτοῦ upon the dative ἑαυτῷ, ἑωυτῷ, the latter being the outcome of the contraction of the dative ἑοῖ with αὐτῷ. It must be noted, however, that ἑαυτοῦ, ἑωυτοῦ could alternatively result from contraction of the other of the two forms mentioned in the etymological dictionaries, i.e. the genitive ἕο

 ἕκαστα), although it must be said that in these kinds of forms (pronouns and monosyllables), generally speaking, *w‑ tends to be retained longer. 64 The forms in ἑωυτ‑ are found in Herodotus and in a few inscriptions: ἑωυτῷ is found in IGDOP 23.7 (Olbia, ‘Berezan letter’, 6th cent. BC), ἑōυτῶν in IK.Priene 398.4 (4th cent. BC), ἑωυτõ in DGE 726.44. 65 Cf. Section 2 above and also Wackernagel 1928, 82. 66 “Vom Dat. Sg. ἑοῖ αὐτῷ aus entsteht dann durch Univerbierung, Schwund des intervokalischen /i̯/ und Kontraktion bzw. Krasis ein neuer Reflexivstamm, der wie αὐτός flektiert (§ 197): Gen. Sg. m. att. ἑαυτοῦ > αὑτοῦ (§ 62), ion. ἐωυτοῦ (§ 60)”. 67 Already Chantraine 1961, 141−142 wrote, more cautiously, “[o]n a pensé qu’au datif par exemple ἑοῖ αὐτῷ s’est contracté par crase en ἑαυτῷ (cf. ταὐτό de τὸ αὐτό et le thème a été étendu aux autres cas)”. 68 As mentioned above in Section 5, this coexistence would not be an issue and perhaps — given the presence of ἕο/ἑο — ἕ, ἑ (as opposed to οἷ, οἱ) is even conceivable.

38  Sara Kaczko αὐτοῦ,69 and that a subsequent similar process, i.e. the extension of the resulting ἑαυτ‑ and ἑωυτ‑ to the other cases, may have produced ἑαυτῷ, ἑωυτῷ etc.70 In this latter hypothesis for the origin of the type ἑαυτοῦ, ἑωυτοῦ, namely from ἕο αὐτοῦ, the reconstruction of ἑοῖ in Ionic and Attic is not necessary. Based on these considerations, there are three alternatives, as follows: (1) If the type ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ originated in a dialect that featured both ἑοῖ and ἕο (namely from the distinct stems *sew‑ and *sw‑), then ἑοῖ and ἑέ may have come into the Homeric tradition also from either Attic (especially given that Attica is one of the regions possibly responsible for the appearance of certain late traits in the essentially fixed Homeric diction, for which see Section 1) or Ionic (and not necessarily East Ionic, though occurrences in Herodotus and in inscriptions do seem to point in this direction).71 (2) If the type ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ originated in a dialect that featured both ἑοῖ and ἑοῦ, theoretically speaking Attic or Ionic may also have been the origin of ἑέ and ἑοῖ in the Homeric text. (3) The type ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ could have originated from a unique form that underwent contraction, in which case two different scenarios are possible — either: (3a) the ‘conventional’ explanation — that is, contraction of the dative ἑοῖ αὐτῷ — was the basis on which the ‘type’ ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ was built (an explanation that, again, makes Attic or Ionic possible responsible for ἑοῖ and ἑέ in the Homeric text) — or (3b) the origin of the type ἑαυτοῦ (> αὑτοῦ)/ἑωυτοῦ is to be found in the genitive ἕο αὐτοῦ, through the contraction of the two members, upon which were built the forms ἑαυτοῦ, ἑαυτῷ, ἑαυτόν (> αὑτοῦ, αὑτῷ, αὑτόν) and ἑωυτοῦ, ἑωυτῷ, ἑωυτόν.

 69 And theoretically, from a phonological point of view, also from ἑοῦ αὐτοῦ (though, as seen above, the occurrences of ἑοῦ are debated in Attic-Ionic). 70 LSJ s.v. ἑαυτοῦ; Petit 1999, 210−211 (perhaps some sort of influence from the accusative cannot be excluded, having been maintained for a certain period as ἑ αὐτόν without undergoing univerbation). 71 See n. 64 above.

Unconventional Features in Homer  39

 Unconventional Either Way By way of conclusion, I return to the argument touched upon in Section 5 and the observation that West Greek dialects in which either the pronominal dative ending was -οι rather than -ιν, or in which the two co-existed,72 could be responsible for the presence of ἑοῖ and ἑέ in Homer. This observation, along with the discussion in the previous paragraph about the origin of ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ ‘indirectly’ attesting to disyllabic pronouns in some dialects (Attic or Ionic) and, notably, the apparent certainty that ἑοῖ and ἑέ were late arrivals in Homeric diction, point to one conclusion. That is, depending on which explanation for the presence of ἑοῖ and ἑέ in Homer and for origin of ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ is deemed more likely, one or the other of the long-accepted accounts of Homeric linguistics, or of Greek historical linguistics more generally, must be abandoned. That is: – acceptance of one of the first three options (1, 2, or 3a) raises the possibility that ἑοῖ (and ἑέ) entered Homer through the Attic (or Ionic) world. This, of course, does not rule out West Greek dialects (potentially) featuring a dative in ‑οι and therefore means that at least two dialect groups that are ‘unconventional’ in Homeric terms73 may have been responsible for the presence of ἑοῖ and ἑέ in the poems. – acceptance of the last option — 3b, which, on the basis of the instances and distribution of the type ἕο at least seems the economical choice74 — means that ‘conventional’ explanations for the origin of the type ἑαυτοῦ (> αὑτοῦ), ἑωυτοῦ in some handbooks and etymological dictionaries may require reconsideration.

Bibliography Beekes, R.S.P (1986), “Homeric ἐμεῖο and Myc. toe”, in: A. Etter (ed.), O-o-pe-ro-si: Festschrift für Ernst Risch zum 75. Geburtstag, Berlin/New York, 365–371.

 72 Co-occurrence is not problematic and is in fact attested; cf. e.g. ϝιν and ϝοι in Crete (see above n. 20 and also Buck 1955, 98). 73 I.e. Attic and West Greek (of course, in the framework of the cultural influence of Central and Western Greece on the latest phases of archaic Greek epic mentioned in Section 1). 74 This latter scenario, while it cannot be used as a hint that Attic and Ionic also featured ἑοῖ (and ἑέ), of course does not affect the possibility that West Greek dialects (potentially) featuring a dative in ‑οι (see above, Sections 4–6) were responsible for the presence of ἑοῖ and ἑέ in Homer.

40  Sara Kaczko Brandenburg, P. (2005), Apollonios Dyskolos: Über das Pronomen. Einführung, Text, Übersetzung und Erläuterungen, München/Leipzig. Brugmann, K. (1876), Ein Problem der homerischen Textkritik und der vergleichenden Sprachwissenschaft, Leipzig. Brugmann, K. (1885), “Zur Bildung des Genitivus singularis der Personalpronomina”, in: Historische Sprachforschung (= Zeitschrift für vergleichende Sprachforschung) 27, 397–418. Brugmann, K. (1913), Griechische Grammatik: Lautlehre, Stammbildungs- und Flexionslehre, Syntax, 4th edn., München. Buck, C.D. (1955), The Greek Dialects, Chicago. Cassio, A.C. (1998), “Lo sviluppo dell’epica greca e il mondo euboico”, in: M. Bats/B. D’Agostino (eds.), Euboica: L’Eubea e la presenza euboica in Calcidica e in Occidente. Atti del convegno di Napoli, 13–16 nov. 1996, Napoli, 11–22. Cassio, A.C. (2009), “The Language of Hesiod and the Corpus Hesiodeum”, in: F. Montanari/ A. Rengakos/C. Tsagalis (eds.), Brill’s Companion to Hesiod, Leiden/Boston, 179–201. Cassio, A.C. (2018a), “‘Dorisms in Homer Fit No One’s Theories’” (M.L. West)”, text of a paper delivered at the Townsend Lectures, Cornell University, Ithaca, April 2018. Cassio, A.C. (2018b), “Rare Pronouns & Crucial Problems: A Look at the Language of Iliad Book 8”, text of a paper delivered at the Townsend Lectures, Cornell University, Ithaca, April 2018. Chadwick, J. (1990), “The Descent of the Greek Epic”, in: JHS 110, 174–177. Chantraine, P. (1958), Grammaire homérique, vol. 1: Phonétique et morphologie, 2nd edn., Paris. Chantraine, P. (1961), Morphologie historique du grec, 2nd edn., Paris. de Lamberterie, C. (1990), Les adjectifs grecs en ‑ύς: Sémantique et comparaison, 2 vols., Louvain-la-Nueve. Düntzer, H. (1848), De Zenodoti studiis Homericis, Gottingae. Forssman, B. (1991), “Schichten in der homerischen Sprache”, in: J. Latacz (ed.), Zweihundert Jahre Homer-Forschung, Stuttgart/Leipzig, 259–288. Hackstein, O. (2010), “The Greek of Epic”, in: E.J. Bakker (ed.), A Companion to the Ancient Greek Language, Malden, MA, 401–423. Hainsworth, J.B. (1983), The Iliad: A Commentary, vol. 3: Books 9–12, Cambridge. Janko, R. (1992), The Iliad: A Commentary, vol. 4: Books 13–16, Cambridge. Janko, R. (2017), “From Gabii and Gordion to Eretria and Methone: The Rise of the Greek Alphabet”, in: J. Strauss Clay/I. Malkin/Y.Z. Tzifopoulos (eds.), Panhellenes at Methone: Graphê in Late Geometric and Protoarchaic Methone, Macedonia (ca 700 BCE), Berlin/Boston, 135–164. Kaczko, S. (2002), “ἡμῖν, ἥμιν, ἧμῐν da Omero a Sofocle: Problemi linguistici e editoriali”, in: RFIC 130, 257–298. Kaczko, S. (2006), “I duali νώ e σφώ e le fasi della dizione epica”, in: SemRom 9, 313–330. Kühner, R./Blass, F. (1890), Ausführliche Grammatik der griechischen Sprache, vol. 1.1, Hannover. La Roche, J. (1866), Die homerische Textkritik im Alterthum, Leipzig. Leaf, W. (1900–1902), The Iliad, 2 vols., 2nd edn., London. Méndez Dosuna, J. (1993), “Metátesis de cantidad en jónico-ático y heracleota”, in: Emerita 61, 95–134. Morpurgo Davies, A. (1964), “‘Doric’ Features in the Language of Hesiod”, in: Glotta 42, 138– 165.

Unconventional Features in Homer  41

Nussbaum, A.J. (1998), Two Studies in Greek and Homeric Linguistics, Göttingen. Peters, M. (1986), “Zur Frage einer “achäischen” Phase des griechischen Epos”, in: A. Etter (ed.), O-o-pe-ro-si: Festschrift für Ernst Risch zum 75. Geburtstag, Berlin/New York, 303– 319. Peters, M. (1987), “Review of G. Danek, Studien zur Dolonie (Wien 1988)”, in: Sprache 33, 233– 242. Petersen, W.P. (1930), “The Inflection of Indo-European Personal Pronouns”, in: Language 6, 164–193. Petit, D. (1999), *Sṷe‑ en grec ancien: La famille du pronome réfléchi, Leuven. Powell, A. (1938), A Lexicon to Herodotus, Cambridge. Rengakos, A. (1993), Der Homertext und die hellenistischen Dichter, Stuttgart. Rix, H. (1992), Historische Grammatik des Griechischen: Laut- und Formenlehre, 2nd edn., Darmstadt. Ruijgh, C.J. (1995), “D’Homère aux origines proto-mycéniennes de la tradition épique”, in: J.P. Crielaard (ed.), Homeric Questions, Amsterdam, 1–96. Scarborough, M.J.C. (2016), The Aeolic Dialects of Ancient Greek: A Study in Historical Dialectology and Linguistic Classification, Diss. Cambridge. Seebold, E. (1984), Das System der Personalpronomina in den frühgermanischen Sprachen, Göttingen. Schmidt, G. (1978), Stammbildung und Flexion der indogermanischen Personalpronomina, Wiesbaden. Schwyzer, E. (1938), “Ein altes Problem der homerischen Sprache”, in: Sitzungsberichte der Preussischen Akademie der Wissenschaften, phil.-hist. Klasse 1938, 81–92 (= Id., Kleine Schriften, Innsbruck 1983, 486–497). Schwyzer, E. (1939), Griechische Grammatik, vol. 1: Allgemeiner Teil, Lautlehre, Wortbildung, Flexion, München. Shipp, G.P. (1972), Studies in the Language of Homer, 2nd edn., Cambridge. Threatte, L. (1996), The Grammar of Attic Inscriptions, vol. 2: Morphology, Berlin/New York. Thumb, A./Kieckers, E. (1932), Handbuch der griechischen Dialekten, vol. 1, 2nd edn., Heidelberg. Thumb, A./Scherer, A. (1959), Handbuch der griechischen Dialekten, vol. 2, 2nd edn., Heidelberg. Wackernagel, J. (1916), Sprachliche Untersuchungen zu Homer, Göttingen. Wackernagel, J. (1928), Vorlesungen über Syntax, vol. 2, Basel. Wathelet, P. (1970), Les traits éoliens dans la langue de l’épopée grecque, Rome. Wathelet, P. (1981), “La langue homérique et le rayonnement littéraire de l’Eubée”, in: L’Antiquité Classique 50, 819–833. West, M.L. (1988), “The Rise of the Greek Epic”, in: JHS 108, 151–172. West, M.L. (1992), “The Descent of the Greek Epic: A Reply”, in: JHS 112, 173–175. Wyatt, W.F. (1992), “Homer’s Linguistic Forebears”, in: JHS 112, 167–173.

Andreas Willi

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula Abstract: The meaning of the Homeric formula σθένεϊ βλεμεαίνων has been debated since Antiquity, with some sources assuming a basic notion of ‘determination’ and others preferring a (par)etymological connection with the verb βλέπω ‘glance’. In modern scholarship both positions have found adherents, and additional reference has been made to the post-Homeric adjective ἀβλεμής whose semantics are equally unclear. A critical reassessment shows, however, that the common postulate of a base noun *βλέμος cannot be supported by the surface parallelism of βλεμεαίνω with μενεαίνω ‘desire eagerly’. Instead, it is proposed that βλεμεαίνω represents the remake, for metrical reasons to do with the morphological evolution of the preceding dative, of a denominative verb *βελεμαίνω ‘hurling missiles’ ← Homeric βέλεμνα. A resemanticization of the formula by secondary (re)association with intr. βάλλω/βάλλομαι ‘rush’ then led to the historical distribution in the Iliad. Keywords: Epic Greek, Homeric formulae, word formation, ancient scholarship, oral poetry.

 The Formula in Homer For many years now, Albio Cassio has been a valiant πρόμαχος of those who study the language of Homer and early epic. It therefore seems fitting — if also daring — to dedicate to him the following remarks on a Homeric formula we first encounter in Book 8 of the Iliad where Hector leads the Trojans in an attack on the hardpressed Greeks (1): (1)

ἂψ δ᾿ αὖτις Τρώεσσιν Ὀλύμπιος ἐν μένος ὦρσεν· οἳ δ᾿ ἰθὺς τάφροιο βαθείης ὦσαν Ἀχαιούς· Ἕκτωρ δ᾿ ἐν πρώτοισι κίε σθένεϊ βλεμεαίνων. (Il. 8.335–337) But in turn the Olympian aroused the spirit of the Trojans, so that they pushed back the Achaeans straight to the deep trench; and Hector advanced among the forefighters, blemeainōn with strength.

https://doi.org/10.1515/9783110621747-004

44  Andreas Willi The phrase σθένεϊ βλεμεαίνων, or its variant σθένεϊ βλεμεαίνει with a finite verb form, re-occurs five more times in the Iliad, each time in the same verse-final position. In Il. 9.237, σθένεϊ βλεμεαίνων is again applied to Hector (2), and Hector is also the comparandus in Il. 12.42 where reference is made to a wild boar or lion (3): (2)

“ἐγγὺς γὰρ νηῶν καὶ τείχεος αὖλιν ἔθεντο Τρῶες ὑπέρθυμοι τηλεκλειτοί τ᾿ ἐπίκουροι, κηάμενοι πυρὰ πολλὰ κατὰ στρατόν, οὐδ᾿ ἔτι φασὶ σχήσεσθ᾿, ἀλλ᾿ ἐν νηυσὶ μελαίνῃσιν πεσέεσθαι. Ζεὺς δέ σφι Κρονίδης ἐνδέξια σήματα φαίνων ἀστράπτει· Ἕκτωρ δὲ μέγα σθένεϊ βλεμεαίνων μαίνεται ἐκπάγλως, πίσυνος Διί, οὐδέ τι τίει ἀνέρας οὐδὲ θεούς· κρατερὴ δέ ἑ λύσσα δέδυκεν.” (Il. 9.232–239) “For the high-spirited Trojans and their far-famed allies have put up their camp near the ships and the wall, lighting many fires throughout the army, and they expect no longer to be held back but to invade the black ships. Zeus the son of Kronos is sending lightnings and thus giving favourable signs to them; and Hector, very much blemeainōn with strength, is raging terribly, trusting in Zeus and paying no heed to men or gods: fierce madness has taken hold of him.”

(3) αὐτὰρ ὅ γ᾿ ὡς τὸ πρόσθεν ἐμάρνατο ἶσος ἀέλλῃ· ὡς δ᾿ ὅτ᾿ ἂν ἔν τε κύνεσσι καὶ ἀνδράσι θηρευτῇσι κάπριος ἠὲ λέων στρέφεται σθένεϊ βλεμεαίνων· οἱ δέ τε πυργηδὸν σφέας αὐτοὺς ἀρτύναντες ἀντίον ἵστανται καὶ ἀκοντίζουσι θαμειὰς αἰχμὰς ἐκ χειρῶν· τοῦ δ᾿ οὔ ποτε κυδάλιμον κῆρ ταρβεῖ οὐδὲ φοβεῖται […]. (Il. 12.40–46) But as before he was fighting like a storm — just as a wild boar or a lion whirls around among dogs and men who are hunting it, blemeainōn with strength: they arrange themselves in a tower formation, face it and hurl with their hands one spear after the other, but its noble heart does not falter nor fear […].

In Il. 17.132–137, the imagery is reminiscent of passage (3), although this time it is Ajax who is compared to a lion defending his young (4); and in the same book, at Il. 17.22, the phrase σθένεϊ βλεμεαίνει predicates the θυμός of a wild boar (5): (4) Αἴας δ᾿ ἀμφὶ Μενοιτιάδῃ σάκος εὐρὺ καλύψας ἑστήκει ὥς τίς τε λέων περὶ οἷσι τέκεσσιν,

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  45

ᾧ ῥά τε νήπι᾿ ἄγοντι συναντήσωνται ἐν ὕλῃ ἄνδρες ἐπακτῆρες· ὁ δέ τε σθένεϊ βλεμεαίνει, πᾶν δέ τ᾿ ἐπισκύνιον κάτω ἕλκεται ὄσσε καλύπτων· ὣς Αἴας περὶ Πατρόκλῳ ἥρωϊ βεβήκει. (Il. 17.132–137) Ajax stood over the son of Menoitios, covering him with his broad shield, like a lion covers its children when huntsmen encounter it in the woods as it is taking its cubs around — it then blemeainei with strength, and it pulls down the entire brow-skin hiding its eyes —: so too was Ajax protecting Patroclus the hero. (5) “Ζεῦ πάτερ, οὐ μὲν καλὸν ὑπέρβιον εὐχετάασθαι. οὔτ᾿ οὖν παρδάλιος τόσσον μένος οὔτε λέοντος οὔτε συὸς κάπρου ὀλοόφρονος, οὗ τε μέγιστος θυμὸς ἐνὶ στήθεσσι περὶ σθένεϊ βλεμεαίνει, ὅσσον Πάνθου υἷες ἐϋμμελίαι φρονέουσιν.” (Il. 17.19–23) “Father Zeus, yes, it is not right to boast excessively — so: neither a leopard nor a lion nor a destructive wild boar whose immense temper in its breast blemeainei with strength very much have as fierce a mind as the sons of Panthoos do when it comes to spear-fighting.”

Finally, we once read σθένεϊ βλεμεαίνων in a description of the god Hephaestus setting out with Hera, Athena, Poseidon, and Hermes to support the Greeks on the battlefield (6): (6) Ἥφαιστος δ᾿ ἅμα τοῖσι κίε σθένεϊ βλεμεαίνων, χωλεύων, ὑπὸ δὲ κνῆμαι ῥώοντο ἀραιαί. (Il. 20.36–37) And Hephaestus went with them, blemeainōn with strength — limping, but his thin legs were strong.

 Internal Chronology If we compare these passages with each other, we are able to make out a line of development. Firstly, it is noteworthy that the similarity of the σθένεϊ βλεμεαίνων verses in passages (1) and (6) extends beyond the two-word phrase: these verses also overlap with respect to the preceding sequence (…τοισι κίε). Since this is hardly just a coincidence, either Il. 20.36 (in (6)) must be modelled directly after

46  Andreas Willi Il. 8.337 (in (1)) or — as seems more likely in the absence of any intrinsic thematic connection between the two passages — Il. 20.36 represents the adaptation of a pre-existing longer formula, of which Il. 8.337 is but one instantiation that happens to be attested.1 Secondly, passage (2) unquestionably echoes (1) as it occurs near the start of the speech with which Odysseus, accompanied by Ajax and Phoenix, seeks to convince Achilleus to return to the battlefield — his essential argument being precisely how the Trojans have the upper hand as related in that section of Iliad 8 to which (1) belongs. Moreover, passage (3) may also have been composed by the poet having in mind Hector’s prowess as described in (1), not only because the simile centers around the same hero but because what triggers it is specifically Hector’s fighting ὡς τὸ πρόσθεν. And once the (by now slightly truncated) formula had thus found its way into an animal simile, it could be readily reused in similar environments, as witnessed by (4) and (5) from Iliad 17.2 Thus, if we had to decide which of all the σθένεϊ βλεμεαίνων passages is the one most likely to contain an ‘original’ or ‘primary’ use of the formula, our best bet would be passage (1) with ἐν πρώτοισι κίε σθένεϊ βλεμεαίνων. As we shall see in due course, this recognition may indeed be of some relevance, and it will help us to answer a number of rather difficult morphological and semantic questions the formula raises.

 Ancient Explanations To start with semantics, uncertainty about the meaning of σθένεϊ βλεμεαίνων becomes visible already at an early point in Antiquity. Commenting on Il. 8.337, the D scholia, the basis of whose lexicographical component “goes back to the schoolrooms of the classical period, so that it predates the Alexandrians and represents the oldest surviving stratum of Homeric scholarship”,3 explain

 1 The inverse assumption would be much less plausible. Whereas ἐν πρώτοισι in ἐν πρώτοισι κίε σθένεϊ βλεμεαίνων would not fit the context in Iliad 20, so that some adjustment was unavoidable (Hephaestus being a member of a very small party of gods who leave Mt. Olympus in order to fight with the Greeks), ἅμα τοῖσι κίε σθένεϊ βλεμεαίνων could no doubt have been combined with Hector’s name in Il. 8.337 if it had been the ‘standard’ wording (e.g., ἔνθ᾿ Ἕκτωρ ἅμα τοῖσι κτλ.). 2 That (4) and (5) are derivative is also suggested by E.-M. Voigt in LfgrE 2.66, s.v. βλεμεαίνω (“in jedem Fall der Indik[ativ] u[nd] bes[onders] Ρ 22 sek[undär]”; note the unclear function of περί in Il. 17.22: Schwyzer/Debrunner 1950, 423–424 and 501). 3 Dickey 2007, 20; cf. van Thiel 2014, 7–9.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  47

βλεμεαίνων in two fundamentally diverging ways: (a) σφοδρῶς ἐπιρρωνύμενος (‘strongly plucking up courage’); (b) ἔνιοι δὲ τῷ βλέμματι φοβῶν οἷον τὴν δύναμιν διὰ τοῦ βλέμματος ἐνδεικνύμενος (‘but according to others “frightening with one’s glance, as if displaying one’s power by one’s glance”’; cf. for (a) also schol. D Il. 12.42 and schol. D Il. 20.36 van Thiel, for (b) also schol. D Il. 9.237 and schol. D Il. 17.22 van Thiel). It is striking that each of these alternatives has found a modern proponent in the two most substantial recent contributions on the topic, by A. Athanassakis and A. Nikolaev respectively.4 Whereas option (a), which centers around the notion of ‘strength/confidence/determination’ (ἐπιρρώνυμαι to ῥώμη), appears to take its cue from the accompanying noun σθένεϊ, the approach of option (b) is essentially (par)etymological, implicitly linking βλεμ‑εαίνω with βλέμ-μα ‘glance’. Later on, Hesychius offers an even wider range of synonyms, but in his list, as in (a), the predominant idea is again that of (self-)confidence: (7) βλεμεαίνειν· γαυριᾶν (‘pride oneself’). ἐπαίρεσθαι (‘be elated’). πεποιθέναι (‘be confident’). εὐθυμεῖν (‘be in good spirits’). χαίρειν (‘rejoice (in)’). πιστεύειν (‘trust (in)’). ἀφρίζειν (‘foam’). ὀργίζεσθαι (‘be angry’). (Hsch. β 693 Latte/Cunningham)

If the final two glosses here, ἀφρίζειν and ὀργίζεσθαι, diverge from the general semantic field of the rest, this need not mean much. Their aim is probably just to circumscribe the approximate factual value of βλεμεαίνων/-νει in the three Iliadic animal similes where an equation of the verb with a human emotion (‘pride, confidence’) might seem out of place. That something along the lines of (a)5 was the standard interpretation in Late Antiquity is in any case also suggested by the repeated use, as an equivalent of γαυριάω (‘pride oneself’), of the hitherto exclusively Homeric verb βλεμεαίνω in Gregory of Nazianzus (Patrologia Graeca 37.606.7, 37.676.9, 37.905.9, 37.1547.1 Migne, De testamentis et adventu Christi 48 Wyss; cf. Lexicon in carmina Gregorii Nazianzeni [ordine versuum] 565 and [ordine alphabetico] β 49 Kalamakis).6

 4 Athanassakis 1970; Nikolaev 2017; cf. below. 5 Only an explanation according to (a) is also given in Apollonius Sophista, Lexicon Homericum 51.31 Bekker and Hdn. Schematismi Homerici 27 Egenolff; and (b) is similarly missing in Etym. Gen. β 136 Lasserre/Livadaras and Etym. Sym. 1.446.22–24 Lasserre/Livadaras (contrast Etym. Gud. 273.9–11 de Stefani, Etym. Magn. 199.45–50). 6 Note, however, that Eustathius in his commentary on the Iliad equally references (b) (Eust. in Il. 8.337, 2.591.4–6 van der Valk, and in Il. 20.36, 4.362.23–363.1 van der Valk) and also mentions

48  Andreas Willi

 Adjectival ἀβλεμής and Adverbial ἀβλεμέως Indirect early evidence for interpretation (a) is also provided by the adjective ἀβλεμής and the corresponding adverb ἀβλεμέως. The latter is first attested in a fragment of the Herakleia of the early-5th-century epic poet Panyassis of Halicarnassus (fr. 17.7–9 Bernabé = fr. 13.7–9 Davies): (8) ἀλλ᾿ ὅτε τις μοίρης τριτάτης πρὸς μέτρον ἐλαύνοι πίνων ἀβλεμέως, τότε δ᾿ Ὕβριος αἶσα καὶ Ἄτης γίγνεται ἀργαλέη, κακὰ δ᾿ ἀνθρώποισιν ὀπάζει. But when someone goes on to the full measure of the third round, drinking ἀβλεμέως, then the troublesome lot of Insolence and Delusion comes about and presents evil to men.

Admittedly, the intended meaning of πίνων ἀβλεμέως is less clear here than when a relevant form next occurs, in Pseudo-Longinus’ Περὶ ὕψους (29.1): (9) Ἐπίκηρον μέντοι [τὸ] πρᾶγμα ἡ περίφρασις, τῶν ἄλλων πλέον, εἰ μὴ σὺν μέτρῳ τινὶ λαμβάνοιτο· εὐθὺς γὰρ ἀβλεμὲς προσπίπτει, κουφολογίας τε ὄζον καὶ παχύτητος· Periphrasis is a hazardous thing, more than anything else, unless used with some moderation; for it is quickly perceived as feeble, smacking of empty speech and dullness.

Since Pseudo-Longinus clearly wants to warn his readers against the potential negative effects of overusing what might otherwise be a powerful figure of speech, a translation such as ‘feeble’ (~ ‘without strength’) seems most appropriate; and this is confirmed when the adverb ἀβλεμέως is given in Pollux’ Onomasticon (3.123) as one item in a long list of synonyms which also contains, for example, μαλακῶς ‘feebly’, βραδέως ‘slowly’, ἀθύμως ‘faintheartedly’, ἀργῶς ‘lazily’,

 a further (par)etymological option (c) according to which βλεμεαίνω = βρέμω ‘roar’ (via an intermediary form (*)βρεμ(ε)αίνω; cf. schol. T Il. 9.237 Erbse; Orion Etymologicum 31.12–32.2 Sturz; Hsch. β 1089 Latte/Cunningham [βρεμεαίνων· ἠχῶν]; Etym. Gen. β 136 Lasserre/Livadaras; Etym. Gud. 273.9–11 de Stefani; Etym. Sym. 1.446.23–24 Lasserre/Livadaras; Etym. Magn. 199.45–46) (Eust. in Il. 8.337, 2.591.6–8 van der Valk, and in Il. 20.36, 4.363.1–2 van der Valk); elsewhere, he seeks to combine all of these (Eust. in Il. 12.42, 3.351.27–352.1 van der Valk, also including ὀργίζεσθαι). In addition, a search in the electronic Thesaurus Linguae Graecae (www.tlg.uci.edu) produces two pertinent attestations of βλεμεαίνων/βλεμεαίνοντες in the orations of Nicetas Choniates (Or. 7, p. 57.13, and Or. 8, p. 78.6 van Dieten; 12th/13th cent.): since δεινόν precedes the participle in both of these cases, it looks as if Nicetas treated (δεινὸν) βλεμεαίνω as a synonym of (δεινὸν) βλέπω, again in accordance with option (b).

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  49

ἀρρώστως ‘weakly’, νωθρῶς ‘sluggishly’, and ἀσθενῶς ‘powerlessly’.7 Moreover, there is every reason to accept the scholiasts’ understanding of the neuter ἀβλεμές as an equivalent of ἀδρανές ‘impotent’ or ἀσθενές ‘powerless’ in a last passage where the adjective occurs, a description of the symptoms of lead poisoning in Nicander’s Alexipharmaca: amongst other things, the victim is there said to suffer from dry heaving (Nic. Alex. 81, ξηρὰ ἐπιλλύζων […] χελλύσσεται, literally ‘hiccuping he expectorates dryly’) — ἀβλεμὲς ἦ γὰρ κεῖνο πέλει βάρος ‘for it is indeed a powerless torpor’ (i.e., ‘for it is indeed a torpor that leaves the patient powerless’, Nic. Alex. 82).8 In the light of this later evidence, we may legitimately ask whether it is true that “[t]he intended meaning of ἀβλεμέως [in Panyassis] seems to be ‘(drinking)

 7 See further Suda λ 410 Adler (καὶ τῶν ἵππων οἱ μὲν ἀμελεῖς καὶ ἀβλεμεῖς καὶ νωθροὶ λήθαργοι καλοῦνται ‘and horses that are negligent and feeble and sluggish are called lethargic’; cf. Hsch. λ 806 Latte), where the juxtaposition ἀμελής ~ ἀβλεμής hints at an ancient etymological theory connecting the two words (cf. Orion Etymologicum 14.21–22 Sturz: ἀβλεμέως, πλεονασμῷ τοῦ β, οἷον ἀμελῶς, ὑπέρθεσις ἀλεμέως, καὶ ἐπένθεσις τοῦ β, ἀβλεμέως ‘ἀβλεμέως, with pleonastic β, as in ἀμελῶς → metathesized ἀλεμέως → insertion of β: ἀβλεμέως’; contrast the derivation favoured by Eust. in Il. 1.81, 1.87.12 van der Valk: ἀβλεμὲς τὸ νωθρόν, ὡς οἷον ἀφλεβές τι ὄν ‘ἀβλεμές is the same as νωθρόν “sluggish”, as if ἀφλεβές τι ὄν “something that is without veins”’). The equation ἀμελέως = ἀβλεμέως also underlies Etym. Gud. 2.24–25 de Stefani; Etym. Sym. 1.12.9– 10 Lasserre/Livadaras; Etym. Magn. 3.25; and Zonaras Lexicon 12.5 Tittmann, all with ἀβλεμέως· ἀφροντίστως (‘carelessly’), while in Hesychius the range of synonyms even contains terms as vague as φαῦλος ‘bad’ and κακός ‘bad’ (Hsch. α 137 and α 138 Latte/Cunningham). Meanwhile, Eust. in Il. 12.42, 3.352.1–3 van der Valk, echoing schol. T Il. 12.42 Erbse, additionally glosses ἀβλεμής with ἀόργητος ‘not irascible’ (next to ἄνορμος ‘uneager’, οὐ πεποιθώς ‘not confident’, ἄτολμος ‘cowardly’, ἀσθενής ‘powerless’), just as he gives βλεμεαίνειν ~ ὀργίζεσθαι (cf. Section 3, n. 6). 8 The following lines detail further the fatigue (κάματοι) that accompanies lead poisoning, sometimes resulting in confusion and coma. Nikolaev 2017, 121, objects to translating Nicander’s ἀβλεμές with ‘ineffective’ (as advocated by Gow 1951, 97, and accepted by later editors: Jacques 2007, 9; Spatafora 2007, 217 and 251), on the grounds that “an affliction (βάρος) would be appropriately called feeble or ineffective only if it does not make the patient too sick”; but there is nothing wrong with (proleptically) calling a torpor (βάρος) ‘powerless’ if powerlessness is what results from it. Thus, one need not even assume, with Gow, that the poet was specifically thinking of the inability to throw up the ingested poison, let alone to follow either Gow/Scholfield 1953, 99, and LSJ Supplement 1 s.v. ἀβλεμής, who translate ‘severe’ or ‘violent’ (an idea which cannot rely on the Panyassis passage: cf. below), or Nikolaev 2017, 122, who conjectures “that Nicander used the word ἀβλεμής […] in the meaning ‘vomitless’” because he had picked this up “from his reading of the medical literature”; if such a medical ἀβλεμής or also “its putative derivational base βλέμος, βλέμεος ‘emesis, vomit’” had ever existed, it would certainly have been commented on by the ancient (already pre-Nicandrean) sources who worried about Homer’s βλεμεαίνων.

50  Andreas Willi immoderately, excessively’”, so that there is a “discrepancy between ἀβλεμέως ‘excessively’ in Panyassis and ἀβλεμής ‘not strong, feeble’ in Suda and Ps.-Longinus”;9 if so, one would indeed have to either emend, following Bergk, into ζαβλεμέως (cf. Hsch. ζ 4 Latte, ζαβλεμέως· μεγάλως πεποιθώς) or accept Nikolaev’s suggestion “that Panyassis understood σθένεϊ βλεμεαίνων as ‘abundant with strength’” and hence “coined ἀβλεμέως ‘in great abundance’ with intensive ἀ‑”.10 However, quite apart from the fact that an analysis of ἀβλεμής as containing an ἀ- privativum (στερητικόν) rather than an ἀ- intensivum (ἐπιτατικόν) would be more straightforward when the former is undoubtedly what we find later on, it is also worth noting that ‘with great strength/determination’ and ‘without moderation (~ with excess)’ are not quite the same thing. Instead, Panyassis’ point might be, more simply, that once the third round of drinks is reached, the actions and movements of even the stoutest drinker become uncontrolled and in that sense, though not necessarily ‘without (physical) power’ (ἀσθενής), still at least ‘without determination’ (ἀβλεμής). But be that as it may, in either case Panyassis’ ἀβλεμέως aligns itself with interpretation (a) for Homeric σθένεϊ βλεμεαίνων since the -βλεμ‑ element in the word does not establish a connection with βλέπω etc. (= interpretation (b)).

 The Creation of ἀβλεμής Yet, the fact that interpretation (b) existed at all, and that it probably also reaches back to the Classical period (given its inclusion in the lexicographical layer of the D scholia: cf. Section 3), is not without significance. We may infer from it that (σθένεϊ) βλεμεαίνων belonged to the range of (semi-)obscure Homeric expressions (γλῶτται) that had to be elucidated by specialists — not unlike, for example, κόρυμβα ‘top end of the stern of a ship’ or ἀμενηνὰ κάρηνα ‘fleeting heads’ with which the traditionalist father wants to test the cultural knowledge of his all-toomodern son in Aristophanes’ Banqueters of 427 BC:11

 9 Nikolaev 2017, 123, referencing Roth 1991, 251; cf. also LSJ Supplement 1 s.v. ἀβλεμής. Nikolaev rightly rejects the translation “without self-regard” offered by Matthews 1974, 85, as being “based on his rather arbitrary interpretation of σθένεϊ βλεμεαίνων” (viz., ‘with pride in his might’). 10 Nikolaev 2017, 123. 11 On the interpretation of this passage and its wider context, Cassio 1977, esp. 75–77, remains fundamental.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  51

(10) (A) πρὸς ταύτας δ᾿ αὖ λέξον Ὁμήρου γλώττας· τί καλοῦσι κόρυμβα; ⏑⏑– ⏑⏑– ⏑⏑– ⏑⏑– τί καλοῦσ᾿ ἀμενηνὰ κάρηνα; (B) ὁ μὲν οὖν σός, ἐμὸς δ᾿ οὗτος ἀδελφὸς φρασάτω· τί καλοῦσιν ἰδύους; ⏑⏑– ⏑⏑– ⏑⏑– ⏑⏑– ⏑⏑– τί ποτ᾿ ἐστὶν ὀπύειν; (Ar. fr. 233 K.-A.) (A) Now, tell me about the following Homeric glosses: What do they call κόρυμβα? […] And what do they call ἀμενηνὰ κάρηνα? (B) In that case, let your son, this brother of mine, explain this: What do they call ἰδῦοι [‘bystanders’]? […] And what does ὀπύειν [‘wed’] mean?

Consequently, it is most unlikely that the adjective ἀβλεμής, or the noun *βλέμος upon which it appears to be based (cf. Section 6), were still living words when Panyassis composed his Herakleia.12 If that had been the case, the generation of Panyassis’ nephew (or cousin) Herodotus would hardly have been compelled to puzzle over the patently related Homeric participle βλεμεαίνων; and if an unassailable explanation for βλεμεαίνων had thus been around at the time when Homeric scholarship got off the ground for good, there would have been little room for any alternative proposal. Whether it was Panyassis himself, or some earlier poet, who first created the compound ἀβλεμής is impossible to tell. The few later sources that use the word may well have got it from Panyassis since his poetry was very highly regarded (Dion. Hal. De imit. 2.2, 2.204 Usener/Radermacher, Quint. Inst. 10.1.52, Panyassis test. 13, 14 Bernabé = test. 11, 12 Davies) and he was even included (alongside Homer, Hesiod, Peisandros, and Antimachus) in the canon of the five ‘classical’ epic poets (cf. Panyassis test. 8–11 Bernabé = test. 10A–10D Davies). In fact, had the adjective been common fare in post-Homeric epic poetry, we might expect it to surface much more often in Classical, Hellenistic, and Imperial hexameter compositions. But whether or not it was Panyassis who is responsible for it, the essential point is to recognize that forming a lexeme ἀβλεμής on the basis of Homeric βλεμεαίνων was fairly simple. Since βλεμεαίνων presupposes a verb in -εαίνω, the only other Homeric example of which is μενεαίνω ‘desire eagerly, long for’,13

 12 A (multo) fortiori, the same may be said about the era in which Nicander wrote the Alexipharmaca: cf. Section 4, n. 8. 13 If we leave aside ἀλεαίνω ‘make warm’ (and compounds), whose structure is different (since directly relatable to ἀλέα ‘warmth’, rather than a hypothetical s‑stem noun *ἄλος; cf. Chantraine 1937, 206–207 n. 1), as well as the even less comparable λε(ι)αίνω ‘smooth, polish’ (and compounds), which is derived from adj. λεῖος ‘smooth’, there are also hardly any true comparanda

52  Andreas Willi and since μενεαίνω is obviously derived from the s-stem μένος ‘spirit, passion’ with (equally epic) compounds such as εὐμενής ‘well-tempered’ and δυσμενής ‘ill-tempered’, any half-decent poet would have been able to infer from βλεμεαίνω an underlying s-stem noun *βλέμος from which to create ἀβλεμής.

 Homeric μενεαίνω and the Type in -(ε)αίνω This being so, we should ask whether the existence of Homeric βλεμεαίνων on its own really entitles us to postulate a neuter noun *βλέμος. At first sight, the answer seems to be positive, precisely because of the parallel existence of μενεαίνω next to well-attested μένος. The modern secondary literature thus confidently operates with *βλέμος, thereby shifting the question of the formal and etymological explanation of βλεμεαίνων to that of the formal and etymological explanation of *βλέμος.14 But before we explore this further, it may be useful to remind ourselves that even the apparently ‘transparent’ μενεαίνω cannot in fact be a direct derivative of μένος. As such, we would expect it to take the form *μενε(ί)ω (< *menes‑ye/o‑), just as we get τελε(ί)ω ‘accomplish’ (< *teles‑ye/o‑) to τέλος ‘end’.15 What the verbal stem formation in -αινε/ο‑ intimates instead is the previous existence of some intermediate nominal n-stem derivative of μένος. In view of this, Solmsen argued “dass von den Stämmen μενεσ- und βλεμεσ- […] individualisierende Substantive in der Gestalt *μενέων *βλεμέων gebildet wurden, vgl. awest. hazaŋh-an‑ ‘Räuber’ von hazah- (= ai. sáhas‑) ‘Gewalt, Gewalttat’, und von denen

 in post-Homeric Greek: ptcpl. βρεμεαίνων (glossed as ἠχῶν ‘resounding’ in Hsch. β 1089 Latte/Cunningham) appears to be an invention of the grammarians trying to explain βλεμεαίνων (cf. Section 3, n. 6, with Etym. Gud. 273.9–11 de Stefani and Eust. in Il. 8.337, 2.591.6–8 van der Valk; Debrunner 1907, 20, “vielleicht auch von einem Dichterphilologen wirklich in Poesie verwendet”), just like φλεβεαίνω (Hdn. Schematismi Homerici 27 Egenolff, Eust. in Il. 1.81, 1.87.13 van der Valk), and Apollonius of Rhodes (1.670, 1.771) builds the compound (in tmesis) περιμενεαίνω ‘wish for ardently’ to Homer’s μενεαίνω, perhaps inspired in his turn by Il. 17.22 where περὶ […] βλεμεαίνων may be analyzed in the same way (cf. Section 1 and Section 2, n. 2). 14 Cf. GEW 1.242; DELG 1.179, and EDG 1.219, all s.v. βλεμεαίνω; Debrunner 1907, 19; Solmsen 1909, 51–52 n. 2; Chantraine 1937, 206; Schwyzer 1939, 733; Athanassakis 1970, 58 and 60; Chadwick 1996, 76; Nikolaev 2017, 126–128. Contrast only Fraenkel 1906, 5, who explicitly rejects this and sees behind μενεαίνω and βλεμεαίνω two stems in *‑wen‑ (or *‑yen‑) derived from arbitrarily posited disyllabic bases μενε- and βλεμε- respectively. 15 Or alternatively, with Fraenkel 1906, 5, “*βλεμαίνειν und *μεναίνειν […], vgl. θαμβαίνειν u.a., von adj. ‑εσ‑St. δυσχεραίνειν, δυσμεναίνειν, ὑγιαίνειν u.s.w.”.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  53

sind μενεαίνω βλεμεαίνω die regelrechten Ableitungen”.16 This view is now shared by Nikolaev, with the specification that “in modern terms such an n‑stem may be interpreted either as a possessive stem *menes‑h1on‑ ‘with rage’ of the same type as Avestan hazaŋhan-/hazasn- ‘thug’ from *seg̑ hes‑h1on‑ (with Hoffmann suffix, reconstructed as *‑h3on(h2)‑ by some scholars) or as a back-formation to a loc. *menes‑en ‘in rage’”.17 By contrast, Chantraine was of the opinion that the stem *menes‑n̥ ‑ presupposed by μενεαίνω “constitue un parallèle exact aux abstraits hittites en *‑eš(š)ar” (i.e., nom.-acc. *‑es‑r̥ , gen. *‑es‑nos, etc.).18 Interestingly, by turning their exclusive attention to cognate languages, both of these approaches fail to notice that more immediate comparative material is actually available within Greek itself. Greek, of course, has a productive adjective suffix -νός, and “[c’]est surtout pour former des adjectifs dérivés de thèmes en s que le suffixe -νος a été productif”.19 Starting from *‑es‑nós, the type surfaces in Homeric Greek either as (Ionic) ‑εινός or as (Aeolic) -εννός. Examples include φαεινός/φαεννός ‘radiant (~ associated with light = φάος)’, ἀλγεινός ‘painful (~ associated with pain = ἄλγος)’, ἐρεβεννός ‘dark (~ associated with darkness = ἔρεβος)’, κλεινός (< *κλεεινός)/κλεεννός ‘famous (~ associated with fame = κλέος)’, etc.20 Now, just as what at least synchronically seem to be denominal y‑presents to adjectives in *‑ro‑ regularly appear as verbs in -αίρω (< *‑r̥ ‑ye/o‑: e.g. ἐχθρός ‘hating, hostile’ ~ ἐχθαίρω ‘hate’, καθαρός ‘pure’ ~ καθαίρω ‘purify’),21 so we may expect verbs in -αίνω to be based not only on athematic n‑stems but also, by the same sort of ‘suppression’ of the thematic vowel, on thematic stems in *‑no‑ (e.g. κυδνός ‘glorious’ ~ κυδαίνω ‘glorify’, ἰσχνός ‘dry’ ~ ἰσχ(ν)αίνω ‘make dry’).22 It is true that the deadjectival verbs in -αίνω eventually become productive mainly as causatives (cf. also e.g. Hom. λευκαίνω ‘make white’), but to begin with their semantics will have been no more restricted than those of other formations in *‑ye/o‑; after all, Homer also has for example χαλεπαίνω ‘be severe/violent (χαλεπός)’ and λιγαίνω ‘cry out (= be) loud (λιγύς)’.23 In other words, in order to account for Homer’s μενεαίνω ‘desire, rage = be passionate’ we need not turn to Avestan or Hittite, but merely think of adjectives like ἀλγεινός < *alges‑nó‑, postulate an exactly parallel *menes‑nó‑ ‘passionate/fierce (~ associated with fierce  16 Solmsen 1909, 51–52 n. 2. 17 Nikolaev 2017, 126 n. 35. 18 Chantraine 1937, 206; cf. Chantraine 1958, 344. 19 Chantraine 1933, 195. 20 Cf. Chantraine 1933, 195–196; Risch 1974, 100. 21 Cf. Debrunner 1907, 201–203; Schwyzer 1939, 725. 22 Cf. Debrunner 1907, 24–30, esp. 27–28. 23 Schwyzer 1939, 733.

54  Andreas Willi passion = μένος)’, and acknowledge that, if one were to build a y‑present to such an adjective at some early stage of Greek, the regular product should indeed be *menes‑n̥ ye/o‑ > μενεαίνω.

 Indirect Evidence for *menes-nó‑ To this derivation of μενεαίνω, one might perhaps object that our *menes‑nó‑ is no more attested as such than Solmsen(-Nikolaev)’s *menes‑(H)on‑ (*μενέων) — or possibly even slightly less so if one were to recognize a trace of Solmsen’s *μενέων in the Homeric compound δυσμενέων ‘bearing ill-will, hostile’.24 However, the latter cannot be explained as δυσ- + *μενέων, not just because the most typical δυσ- compounds are bahuvrihis rather than determinative compounds, but also because the actual meaning of δυσμενέων could result from δυσ- (‘badly, in a bad way’) + *μενέων only if *μενέων alone meant something positive like *‘good-willed, friendly’ — which in turn is not what we need if we are to derive μενεαίνω from it. By contrast, seeing in δυσμενέων the participle of a denominal verb in *‑ye/o‑ to adj. δυσμενής is entirely unproblematic: it produces the right meaning (‘be δυσμενής = hostile’), and as far as the use of δυσμενέων rather than metrically impossible *δυσμενείων (with *‑es‑ye/o‑ > ‑ειε/ο‑) is concerned, it is a general rule that among verbs of this shape “[u]ne loi d’alternance amène ει au temps fort, ε au temps faible”: “[d]e τελέω, par exemple, les aèdes emploient alternativement un thème τελεο- en Ψ 180, λ 280, et τελειο- en ζ 234, ψ 161”.25 Meanwhile, there does seem to be a real, albeit indirect, trace of *menes-nó‑ in the language of Greek epic. In Section 5, Ar. fr. 233 K.-A. has been quoted, in which ἀμενηνὰ κάρηνα is singled out as a characteristically difficult Homeric gloss. This particular phrase occurs exclusively in the verse-end formula νεκύων ἀμενηνὰ κάρηνα ‘the ἀ. heads of the dead’ of the Odyssey (Od. 10.521, 10.536, 11.29, 11.49), but the adjective ἀμενηνός is also attested once in the Iliad (Il. 5.887) and once more, as an epithet of ὀνείρων ‘dreams’, in the Odyssey (Od. 19.562). Its

 24 While himself questioning it, Nikolaev 2017, 126 n. 35, seems to ascribe this idea to Solmsen 1909, 51–52 n. 2; but it first emerges in Chantraine 1937, 206: “Homère emploie un adjectif δυσμενέων en β 72. Le nominatif pluriel δυσμενέοντες (β 73, υ 314) est un arrangement grec: il ne s’agit pas là d’un thème verbal: aucune forme d’infinitif ou de mode personnel ne répond à ce δυσμενέων”. Note that Chantraine’s argument here is beside the point since δυσμενέων/ δυσμενέοντες are merely created as metrically acceptable alternatives to metrically impossible δυσμενής and metrically unsuitable δυσμενέες respectively; cf. Meissner 2006, 213–214. 25 Chantraine 1958, 166.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  55

basic meaning, which is given as “kraftlos, ohne Lebenskraft, d.h. ohne μένος daher auch rasch vergänglich, wesenlos” in the Lexikon des frühgriechischen Epos,26 has been uncontroversial since Antiquity (cf. schol. D Il. 5.887 van Thiel, ἀμενηνός· ἀσθενής. οὐκ ἔχων μένος), but its formation is somewhat irregular. In view of already Homeric δυσμενής ‘hostile’ (cf. above), and equally Homeric parallels such as ἀκηδής ‘without care (κῆδος)’ or ἀκλεής ‘without fame (κλέος)’, one should expect the lexeme to be (*)ἀμενής (as attested much later, in Eur. Supp. 1116). It is therefore generally accepted that ἀμενηνός represents an artificially ‘enlarged’ version of (*)ἀμενής. However, there is no need to invoke the rare — and again not quite regular — ἀκμηνός ‘full-grown’ (?) as a model here.27 Much rather, the case is comparable to that of Homeric ἀνόστιμος ‘not returning’, which is conglomerated, for metrical reasons, from ἄνοστος ‘without return’ and νόστιμος ‘belonging to a return’.28 If so, the positive element involved in our case, alongside negative (*)ἀμενής, would have to be precisely *menes-nó‑ (> *μενεινός/*μενεννός) — the only proviso being that ἀμενηνός must owe its truly final shape (in lieu of *ἀμενεινός/*ἀμενεννός) to additional interference from ἀμενής with its ‑η‑ vocalism and/or its formulaic combination with κάρηνα (< *‑asna) in a ‘rhyming’ formula.29 Whereas, say, the neuter nom.-acc. pl. of (*)ἀμενής could not possibly be used in a hexameter before synizesis/contraction of -εα (in ἀμενέα) became a regular option,30 no such constraint was encountered by ἀμενηνός (ntr. pl. ἀμενηνά). The remodelling thus neatly addressed a distinct problem of versification.

 26 G. Busch in LfgrE 1.629, s.v. ἀμενηνός. 27 Thus GEW 1.91, and EDG 1.86, both s.v. ἀμενηνός; DELG 2.685, s.v. μέμονα; on ἀκμηνός (Od. 23.191) and its derivation, see H.J. Mette in LfgrE 1.417–418, s.v. (problematizing Risch 1974, 99). 28 Cf. Meister 1921, 15–16; Risch 1974, 214; to ἀνόστιμος Risch adds “vermutlich auch ἀμείλιχος aus ἀμείλικτος und μείλιχος -ιος und ἀθεμίστιος aus ἀθέμιστος und θεμίστιος”. 29 Contrast the otherwise similar (and not just poetic) replacement of ὑγιής by ὑγιεινός (cf. Chantraine 1933, 196); unlike (*)ἀμενής, ὑγιής was of course no longer recognizably a compound (*(h1)su‑gwi(h3)es‑). Influence from ἀμενής must also be held responsible for the — in a compound with ἀ- privativum unexpected — oxytone accentuation of ἀμενηνός (contrast e.g. ἄγνωτος ‘unknown’ vs. γνωτός ‘known’). 30 Cf. Chantraine 1958, 56: “Parmi les groupes commençant par un ε la combinaison εα s’observe en particulier dans la flexion des thèmes en s. Le plus souvent -εα reste disyllabique”. Chantraine then goes on to list less than a dozen Homeric examples of -εα that require synizesis.

56  Andreas Willi

 Proto-Greek *βλέμος? Whereas the necessary intermediary between μένος and μενεαίνω can thus be postulated with some confidence, and both the starting and the end point of the derivational chain μένος → *μενεινός → μενεαίνω are of course firmly secured anyway, the situation is quite different when we return to βλεμεαίνω. In this case, not only is there no hint — however indirect — at the existence of an intermediate form *βλεμεινός (vel sim.), but the starting point *βλέμος itself remains in the air. We have already seen that the post-Homeric compound adjective ἀβλεμής must not be used as an additional argument for the erstwhile existence of *βλέμος because ἀβλεμής could easily be back-formed from βλεμεαίνω by means of a simple four-part analogy μενεαίνω : ἀμενής = βλεμεαίνω : X → X = ἀβλεμής (Section 5).31 By contrast, if we are not prepared to postulate an adjective *βλεμεινός in addition to (or instead of) a noun *βλέμος, we cannot afford to ‘lose’ *βλέμος because, without *βλεμεινός, it looks as if the verb βλεμεαίνω should be the product of a four-part analogy in its turn (μένος : μενεαίνω = *βλέμος : X → X = βλεμεαίνω). Or else, if we are prepared to postulate *βλεμεινός (in order to obviate the need for an analogical explanation of the verb), this latter lexeme too would seem to require an underlying *βλέμος. The only way to avoid reconstructing either *βλέμος alone or both *βλέμος and *βλεμεινός would thus be to account for the apparent derivational end point, the verb βλεμεαίνω, in an altogether different manner. In what follows, such a solution to the βλεμεαίνω problem will in fact be proposed, not because there is anything wrong with postulating lost prehistoric lexemes per se, but because such an approach makes sense only as long as it carries with it the advantage of explanatory economy. And as we shall see instantly, postulating *βλέμος, far from making anything smooth, merely adds another layer of uncertain hypotheses.

 Modern Versions of an Ancient Explanation It will come as no surprise that the ancient connection between βλεμεαίνω and βλέπω etc. (= option (b) in Section 3) is exceedingly hard to defend from a formal

 31 This point is correctly made by Nikolaev 2017, 120–121, even if he subsequently prefers to believe in the reality of *βλέμος in order to support his views on ἀβλεμής in Nicander: cf. Section 4, n. 8.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  57

point of view.32 Perhaps understandably, Athanassakis therefore concentrated on the semantic side of things in his attempt to revive this theory, arguing that “[t]he meanings ‘cast furious glances’, ‘flash the eyes,’ ‘gleam’ restore the symmetry and power of the similes in Il. Θ 336–49, tally so well with Homer’s image of rabid (λυσσώδης), flamelike (φλογὶ εἴκελος) Hector, whose effulgent eyes seem to hold an altogether special fascination for the poet (Γοργοῦς ὄμματ᾿ ἔχων, σμερδαλέον δέδορκεν…ὡς δράκων), and render the ‘hooding’ of the lion’s eyes in Il. Ρ 136–37 easier to understand as a sort of an angry scowl or glower”. In the meantime, all Athanassakis offers in terms of a formal analysis is a few vague remarks on *βλέμος representing “an indirectly surviving variant for *γλέμος with a βλ/γλ shift not unknown to Greek (cf. βλέπω/γλέπω, βλήχων/γλήχων etc.)”, and on *βλέμος being traceable “to a root gel, variations of which may be seen in such words as γελάω, γαλήνη (Aeol. γελήνη, ἀγλαός, γλαυκός, in Hesychius’ glosses) γλαινοί (explained as τὰ λαμπρύσματα τῶν περικεφαλαίων οἷον ἀστέρες) and γελεῖν (explained as λάμπειν ἀνθεῖν), the Homeric γλῆνος (= jewel; Hsch. interestingly enough gives φάος under this gloss), and especially in the Homeric γλήνη (= pupil of the eye)”.33 In some ways this is even less satisfactory than the ancient etymologists’ implicit comparison of βλεμ- in βλεμεαίνω with βλεμ- in βλέμμα (cf. schol. D Il. 8.337 van Thiel, quoted in Section 3). The latter at least tried to make sense of the -μ- in βλεμεαίνω, whereas Athanassakis seems completely oblivious to it. Nor do we get much further with Chadwick, who later on expressed similar views on the meaning of βλεμεαίνω (“darting fierce looks, glaring”).34 Admittedly, Chadwick’s linguistic toolbox looks a bit more professional, as he suggested “to associate [*βλέμος] with the family of βάλλω (< *gwl̥yō) […], as built on a hystero-dynamic form with the enlargement ‑em‑” and to recognize “the same type of enlargement” in “βλέπω, where the dialect form in γλεπ- may be a reminiscence of the labiovelar”; but to invoke alternative (root?) enlargements also ultimately means nothing but to explain obscurum per obscurius.35

 32 Needless to say, the same is even more true of ancient attempts to obtain βλεμεαίνω by means of haphazard — though within their historical context not unusual — theories involving letter alterations, substitutions, or additions, starting from forms such as *βρεμαίνω (to βρέμω ‘roar’) (cf. Section 3, n. 6) or φλεβεαίνω (to φλέψ ‘blood-vessel’) (cf. Section 4, n. 7, and Section 5, n. 13); cf. also Section 4, n. 7, on ἀβλεμέως ~ ἀμελέως. 33 Athanassakis 1970, 60–61. 34 Chadwick 1996, 76. 35 Chadwick’s root is *gwelh1‑, with a final laryngeal (cf. Section 13); thus, *gwelm‑/*gwlem- and *gwelp‑/*gwlep- would not even just be ‘extended’ versions of the root.

58  Andreas Willi

 Nikolaev’s Theory In stark contrast to these cursory remarks, Nikolaev has now proposed a much more comprehensive and carefully argued theory of how to explain *βλέμος.36 Acknowledging that radical “*blem- and *mlem‑ are excluded by well-known constraints on the Proto-Indo-European root structure, and a primary root *gwlem‑ is unheard of”, Nikolaev suggests that *βλέμος “does not have to be a primary s‑stem derived from a verbal root *βλεμ-: it could also be an abstract noun secondarily derived by adding the non-ablauting suffix *‑s‑ to an adjectival stem in *‑mo‑ with a concomitant insertion of a full-grade vowel in the root, compare *tm̥h1-nó‑ ‘cut’ → *temh1‑ne‑s‑ ‘that what has been cut’ (> Greek τέμενος ‘piece of land’). For reasons that are not entirely clear, in this derivational process the vowel is sometimes inserted in the ‘wrong’ place” — a phenomenon Nikolaev then illustrates with the derivational chains “(1) *pl̥h1-nó‑ ‘full’ (Vedic pūrṇá-, Latin plēnus, Germanic *fulla- > English full) → *pélh1ne/o‑s‑ ‘fullness’ > Indo-Iranian *párH-na‑s‑ ‘wealth’ > Vedic párīṇas-” and “(2) *dl̥h1-ghó- ‘long’ (> Slavic *dьlgъ, Vedic dīrghá-, Old Avestan darəga‑) → *dléh1ghe/o‑s- ‘length’ > Indo-Iranian *drāǰas- > YAvestan drājah-”. Accordingly, *βλέμος too is supposed to be “a secondary derivative in *‑mes‑” of a root “*Bel‑ (where B = *b-, *m-, or *gw‑)” by way of a chain “*Bl̥-mó‑ (reanalyzed as *Bl̥m‑o‑) → abstr. *Bleme/o‑s‑ > Greek *βλέμος, *βλέμεhος”; and the root in question is finally identified as “*gwel- ‘to flow’” (cf. e.g. Germanic *kwella- ‘flow (out), gush’), leading to the translation “overflowing (*gwel-) with strength” for Homer’s σθένεϊ βλεμεαίνων and a comparison of that formula with the Old English warrior epithet collenferhþ ‘(lit.) whose spirit is overflowing, brimming’.37 The ingeniousness of all this is undeniable, but a sense of unease remains. Firstly, we are now forced to reconstruct yet another proto-form for which we have no independent evidence in the form of direct descendants, namely a presumptive PIE adjective *gwl̥-mó‑. Secondly, unlike for example *pl̥h1‑nó‑ ‘full’, to which Nikolaev refers, this *gwl̥-mó‑ does not follow a common formational pattern: in a footnote, Nikolaev remarks that it must contain “the suffix *‑mo‑ found, for instance, in Hittite ‑(i)ma‑ used for forming action nouns (e.g. kurkurima‑ ‘intimidation’ from kurkuriya- ‘to intimidate’, weritema- ‘anxiety’ from werite‑ ‘to be anxious’, or tuḫḫima‑ ‘groaning, agony’ from tuḫḫāi‑ ‘to groan, be in pain’ […])”,38

 36 Nikolaev 2017, 126–128. 37 Nikolaev 2017, 128–130. 38 Nikolaev 2017, 128 n. 45.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  59

but he fails to address the issue that in his derivational chain ‑mó‑ should be an adjectival suffix. At best, one could therefore compare an adjective like Gk. θερμός ‘hot’ (next to θέρομαι ‘become hot’ or θέρος ‘summer (heat)’, from *gwher‑), but not only is θερμός itself a rather exceptional formation (and one with an e‑grade, not a zero grade),39 there is also no derived neuter s‑stem noun *θρέμος or the like. And finally, thirdly, as far as the ‘wrongly’ inserted full-grade vowel is concerned, it is worth noting that, even if *gwl̥-mó‑ had been reanalyzed as *gwl̥m‑ó‑ (for some unobvious reason), there would have been much less of an incentive to create *gwlem‑es‑ (instead of *gwelm‑es‑) than to create, say, (PreProto-Indo-Iranian) *dléh1ghe/o‑s- alongside *dl̥h1-ghó-: quite apart from the fact that, as Nikolaev himself admits, Gk. ἐνδελεχής ‘continuous’ (*‑delh1gh‑es‑) suggests that Greek preferred the full-grade vowel in the ‘right’ place even in the latter case, in the former there would have been no danger of ending up with an unusual pre-suffixal CVCVC- root structure in the derived s-stem noun (*gwelm‑es‑).

 Methodological Queries At the same time, there is a question which, it is true, is ultimately one of dogma or taste more than anything else. As watertight as Nikolaev’s argument may look, are we doing justice to the reality behind Homer’s lexicon if we use an expression like our σθένεϊ βλεμεαίνων as the basis from which to infer a complex paradigmatic edifice of Proto-Indo-European (or even just Proto-Greek) antiquity? Or to put it differently: since there can be no doubt that much in Homeric language is ‘artificial’, where is the point at which we have to declare that a seemingly unavoidable reconstructive postulate is one too many, and that the true solution to a problem should rather be looked for elsewhere? Of course, it would be misguided to deny that there is an element of Indo-European heritage in Homer, and that his language contains much that is profoundly archaic; but Homeric language remains in the first place a very special variety of Greek, shaped by a bardic tradition in which other factors were at least as influential as the wish faithfully to preserve inherited linguistic material. Accordingly, the least we can do when dealing with a difficult phrase like ours is to ask whether the costs imposed by a strict adherence to Indo-Europeanist methodology (à la Nikolaev) are not greater

 39 Cf. Chantraine 1933, 151; Risch 1974, 47.

60  Andreas Willi than what we have to live with if we seek to shed light on the problem in a less mechanical — but therefore also less formalized or ‘controllable’ — manner.

 ‘Irregular’ βλεμεαίνω and the Hexameter The peculiar nature of Homeric Greek is due both to the combination of elements that belong to different chronological and/or dialectal layers and to the need to adjust to metrical constraints. Just as phonological or morphological irregularities such as a noun ᾱ̓νήρ (acc. ᾱ̓νέρα) ‘man’ with a long first syllable or an o‑stem ‘genitive’ πασσαλόφι ‘nail’ do not entitle us to reconstruct a Proto-Greek stem *āner- or an oblique ending *‑o‑bhi, so we should pause when we come across an unusual word-formational pattern like the verbal stem in -εαινε/ο-, especially when it is combined not with a well-known basis as in the case of μενεαίνω but with a kind of unknown. The rarity of -εαινε/ο- is of course precisely why it is so tempting to suspect that βλεμεαίνω might be an analogical creation matching the less elusive μενεαίνω (cf. Section 8). But we must not therefore forget that, even if there had been a noun *βλέμος, the most natural way of deriving a denominative verb from it would have been to create *βλεμείω/*βλεμέω, in parallel with, say, ἀκέομαι ‘heal’ (cf. ἄκος ‘remedy’), τελέω/τελείω ‘accomplish’ (cf. τέλος ‘end’), or νεικείω/νεικέω ‘quarrel’ (cf. νεῖκος ‘strife’).40 By contrast, resorting (by analogy) to a variant in -εαινε/ο- would have been equally natural only if an already existing verb of the regular and frequent type in ‑αινε/ο‑41 had to be metrically adapted: for example because a sequence like *σθένεϊ βλεμαίνων does not fit into a hexameter. To be sure, such an assumption will not on its own solve our problem since we are then still left with our unwieldy basis βλεμ-. However, it can be a worthwhile first step because it shifts the focus away from the search for an underlying s-stem noun and towards other options. In particular, hypothesizing the remake of a metrically difficult verb in ‑μαινε/ο‑ into a metrically convenient verb in ‑μεαινε/ο- may suggest to compare formations like πημαίνω ‘distress’, σημαίνω ‘indicate’, ὀνομαίνω ‘name’, etc., and hence to wonder about some connection with the abstract noun type in ‑μα < *‑mn̥ (cf. πῆμα ‘misery’, σῆμα ‘sign’, ὄνομα ‘name’, etc.).

 40 Cf. Risch 1974, 308–309 and 335. 41 Cf. Risch 1974, 290–291.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  61

 Homeric Missiles There is not, of course, any Greek noun *βλέμα. At this point, however, we may usefully remember the argument of Section 2, according to which the original seat of the Homeric formula σθένεϊ βλεμεαίνων is best sought in contexts like Il. 8.335–337 (= passage (1) in Section 1). Here, the present participle βλεμεαίνων describes what Hector is doing while he advances (κίε). Quite possibly, he may be casting furious glances; but what constitutes the real danger for any opponent, and is thus central to the situation, is not so much the warrior’s face expression as his concrete actions. What these typically are becomes clear if we look, for example, at the immediately preceding scene (Il. 8.309–334): Teucros has shot an arrow at Hector, but missed his target and killed Hector’s charioteer; in angry reaction to this, Hector has jumped down from his chariot, picked up a large stone, hurled it, and hit Teucros as the latter was just getting ready to shoot again. So, even when they are already fairly close to each other, the normal mode of attack among Homeric warriors is shooting missiles, be it actual arrows (ἰοί, οἰστοί) or spears (δοῦρα, ἔγχη, αἰχμαί, etc.), or indeed anything else that can be forcefully propelled (βάλλειν).42 As a cover term for all of these objects, Homeric Greek uses either the s‑stem noun βέλος (pl. βέλεα) ‘missile’ or the synonymous neuter plural βέλεμνα.43 Both βέλος and βέλεμνα contain the root *gwelh1‑ also seen in βάλλω, a verb that can either — though rarely — be used with the accusative of the thing that is ‘thrown/shot’44 or else refer to the act of ‘hitting’ a target by throwing/shooting from afar; in fact, the unexpected initial labial (instead of dental) before the radical e-vowel of the two nouns is no doubt influenced by the regular outcome of the inherited labiovelar in the verbal paradigm (e.g. aor. βαλε/ο‑ < *gwl̥(h1)‑e/o‑, pres. βαλλε/ο‑ < *gwl̥n(h1)‑e/o‑; contrast Arcadian aor. ζελε/ο‑ < *gwel(h1)‑e/o-). For our purposes, the Homeric plurale tantum βέλεμνα is of particular interest. Though historically treated as an o-stem neuter, and thus endowed with a back-formed singular βέλεμνον (Aesch. Ag. 1496, 1520, etc.), this is the fossilized nom.-acc. pl. of the neuter derivative in *‑mn̥ to *gwelh1‑. Perhaps because it was gradually replaced in its concrete meaning (‘missile’) by βέλος (pl. βέλεα), and/or because it was superseded in its abstract meaning by the secondarily created βλῆμα ‘throw’ (as if < *gwl̥h1-mn̥ ; but βλῆμα is not attested before Classical times),

 42 Cf. Trümpy 1950, 51–60, 66–68, 104–107; Höckmann 1980, E 312–319. 43 Cf. M. Mentzou and W. Beck in LfgrE 2.49–51, s.vv. βέλεμν(ον) and βέλος; Trümpy 1950, 107. 44 Trümpy 1950, 104; cf. A. Mette and W. Beck in LfgrE 2.25–26, s.v. βάλλω.

62  Andreas Willi *gwelh1-mn̥ > *gwele‑mn̥ (>→ *βέλεμα) with its corresponding plural *gwele‑mn‑a (>→ βέλεμνα) did not undergo the usual remake of its stem class into nt‑stems (with nom.-acc. pl. ‑mn̥ t‑a > -ματα).45 Thus, if even those nouns in ‑μα < *‑mn̥ that were affected by this change still have derived y‑presents in -μαινε/ο‑ < *‑mn̥ ‑ye/o‑ next to themselves (cf. Section 12),46 the same should be true a fortiori of *βέλεμα/βέλεμνα, and we thus have every reason to assume a regularly formed y-present *βελεμαίνω as well. In semantic terms, such a verb (*‘to missile’) would closely correspond to junctures like βέλος/βέλεα ἵημι/ῥίπτω/ἰάλλω — or also to βάλλω in the meaning ‘throw, shoot’ rather than ‘hit’ (cf. above) —, although its denominal character might at the same time convey a more distinctly iterative/continuative nuance, exactly as appropriate to the description of a warrior engaged in battle.

 The Morphosyntactic Prehistory of σθένεϊ Just like a hypothetical *σθένεϊ βλεμαίνων (Section 12), a formula *σθένεϊ βελεμαίνων ‘hurling missiles with strength’ (with five short syllables in a sequence) would not suit the metrical pattern of a hexameter end. One might therefore object that no epic singer would ever have come up with it in the first place, only to then reshape it into something more suitable. As is well-known, such reshaping usually takes place when linguistic change has affected a previously acceptable sequence, the most obvious example being those cases where the loss of initial *ϝ- had to be addressed (e.g. Il. 1.223 ἀταρτηροῖς ἐπέεσσι for *ἀταρτηροῖσι ϝέπεσσι ‘with hurtful (?) words’).47 However, if our formula is reasonably ancient, as is plausible for any formula, it becomes relevant that the dative σθένεϊ it contains is — on any interpretation — an old (instrumental) dative, not a locative. This being so, its original ending should not have been -ι < loc. sing. *‑i, but *-ει < dat. sing. *‑ei (replacing in its turn instrumental *‑ē < *‑eh1). Distinct traces of this ancient state of affairs are visible elsewhere too in Homeric versification, both with other s-stem nouns and with σθένεϊ in particular:

 45 Cf. Thurneysen 1907, 176; Specht 1936, 209–210; Risch 1974, 54. 46 Verbal stems like ὀνομαζε/ο‑ ‘name’ and θαυμαζε/ο‑ ‘wonder, marvel’, which are backformed to aorists like ὀνομασ(σ)α‑ and θαυμασ(σ)α‑ with -μασ(σ)α‑ < *‑mn̥ t‑s(a)‑, thus belong to a later layer; cf. Risch 1974, 297. 47 Cf. Chantraine 1958, 116–157, esp. 118–120.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  63

Il. 15.108 and 17.329 start with κάρτεΐ τε σθένεΐ τε ‘by/in might and strength’ and Il. 16.542 ends with δίκῃσί τε καὶ σθένεϊ ᾧ ‘with his justice and strength’.48

 Language Change and Formulaic Adjustment Against this background it seems legitimate to suggest that the formula which eventually surfaces as σθένεϊ βλεμεαίνων originated, presumably in sub-Mycenaean times, as (something like) *sthene(h)ei belemainōn ‘hurling missiles with strength’. Following the generalisation of the loc. sing. ending *-i > -ι at the expense of the dat. sing. ending *‑ei, this turned into unmetrical *sthene(h)i belemainōn, at which point the rhapsodes were facing a choice between (a) abandoning the formula, (b) preserving the formula with an artificially ‘lengthened’ ‑ī (as in …σθένεΐ τε; cf. Section 14), or (c) preserving the formula by opting for some other artificial adjustment. In an oral tradition which strongly relied on the availability of formulae, option (a) was undesirable. In the absence of any alternative, option (b) might therefore have been chosen, just as it was chosen in κάρτεΐ τε σθένεΐ τε where the following τε presented no scope for alteration. However, before a single stop consonant, whose delivery could not be rhythmically prolonged like that of a continuant, this will have been a last resort, to which option (c) was preferable. That *belemainōn was a derivative of the root *gwelh1‑ >→ bel‑ (as in βέλος, βέλεμνα), and hence related to verbal βάλλω etc., would of course still have been clear at that point; but its more immediate relationship with the regular noun in *‑mn̥ > -μα to the same root may already have been obscured by the remake (by reanalysis from the plural) of this noun (*gwelh1-mn̥ >) *belema, pl. (*gwelh1‑m‑h2 >) *belemna into *belemnon (βέλεμνον), pl. *belemna (βέλεμνα) (cf. Section 13). At the same time, there were various forms belonging to the same lexical family which started with *gwl‑ > *bl- and thus presented exactly the kind

 48 Cf. Janko 1992, 240 and 385, and — also with examples involving other nouns — Meister 1921, 128. Leaving aside the six instances of σθένεϊ βλεμεαίνων/-ει (Section 1), trisyllabic σθένεϊ only occurs in the three passages just cited as well as Il. 12.224 (σθένεϊ μεγάλῳ ‘with great strength’), 13.193 (ditto), 17.322 (κάρτεϊ καὶ σθένεϊ σφετέρῳ ‘with their might and strength’), and 17.751 (σθένεϊ ῥηγνῦσι ῥέοντες ‘they break it flowing with strength’). Thus, a historical CC- cluster follows only once, and in what appears to be a variation of the κάρτεΐ τε σθένεΐ τε phrase (Il. 17.322), and a prehistorical CC- cluster (ϝρ-) in one other instance (Il. 17.751). All this makes it look as if an anapaestic scansion of σθένεϊ once used to be the default option independently of what followed.

64  Andreas Willi of cluster anlaut that was required to make the anapaestic scansion of a preceding *sthene(h)i regular (e.g. ptcpl. med. *gwl̥h1-mh1no- > βλήμενος, 3rd sing. aor. med. *(h1e‑)gwl̥h1‑to > (ἔ)βλητο). However, the mere replacement of *belemainōn by *blemainōn (*βλεμαίνων) would still not have solved the whole problem because the 5th foot of the hexameter would then have been incomplete; nor was a replacement of *belemainōn by *blēmainōn (*βλημαίνων) admissible given the far-reaching strictures against spondaic 5th feet.49 By contrast, the adoption of what was de facto a suffix variant of -αινε/ο‑, namely ‑εαινε/ο‑ as observed in μενεαίνω, presented a good way forward. That this meant to utilize a suffix that was ‘inappropriate’ in diachronic morphological terms will not have bothered the bards more than to do the same when it came to, say, creating a patronymic Πηληϊάδης instead of Πηληΐδης,50 comparatives such as χερείων, χειρότερος, and χερειότερος instead of χείρων,51 or — again in order to ‘save’ a verse-end formula52 — an accusative εὐρέα instead of εὐρύν in εὐρέα πόντον.53 In fact, the substitution of ‑εαινε/ο- to ‑αινε/ο‑ even had the advantage of preserving the suppressed ‑e‑ of the beginning of the word, and hence to effect a word-internal metathesis (*belemainōn → blemeainōn/βλεμεαίνων) through which the end product was phonologically even more reminiscent of the starting point than something like *blemainōn would have been.54  49 According to West 1982, 37 with n. 13, “[c]ontraction [i.e., the substitution of a single long syllable for the two short syllables of a hexameter foot, A.W.] is more frequent in the first colon (about 40% of lines in each of the first two feet) than in the second (about 30% in the fourth foot, only 5% in the fifth)”, and “[m]any examples in the second colon arise from vowel contraction, e.g. ἐμίμνομεν Ἠῶ δῖαν (earlier Ἠόα)”; “[w]hen these are excluded the figure for the fifth foot falls to 2%”. Since the time period envisaged above would precede the period of widespread vowel contraction, and since most instances of West’s remaining 2% in the 5th foot may well have come into being only when vowel contraction had subsequently opened the flood gates, *blēmainōn was very likely just out of the question back at our early point. 50 Cf. Debrunner 1924, 38–40. 51 Cf. Chantraine 1958, 259. 52 Note the general remarks by Witte 1913, 2226: “[i]n besonders hohem Maße erwies sich der Einfluß des Verses fördernd auf dem Gebiet der Wortbildung. Gerade hier läßt sich am deutlichsten zeigen, wie es nur wenige Stellen des Hexameters sind, denen die neuen Formen ihr Dasein danken, vor allem das Versende von der bukolischen Diärese ab”. 53 Witte 1913, 2225; Meister 1921, 18–19; Hoekstra 1965, 112. 54 Although the matter again remains controversial, one may also compare what Alcman — or someone in the poetic tradition before him — did to an epithet of ‘winged’ dreams, ὑπόπτερος, when he not only added the suffix ‑ίδιος for metrical convenience, but in addition metathesized within the word (and in analogy with the underlying πέτομαι ‘fly’) in order to obtain a required dactylic structure τῶν ὑποπετριδίων ὀνείρων instead of metrically impossible τῶν ὑποπτερίδιων (or ὑποπτέρων) ὀνείρων (Alcm. fr. 1.49 PMG, with the explanation given by Herodian according

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  65

 The Formula in Old and New Contexts As soon as the new βλεμεαίνων had thus become detached from the underlying word for ‘missile’ (βέλεμνον), while still retaining its connection with the basic root of βάλλω etc., a semantic reanalysis was bound to occur. When *sthene(h)ei belemainōn still meant ‘missiling (= hurling missiles) with strength’, it naturally took no direct object: the semantic object ‘missiles’ was inherent in the denominal verb itself. Due to the (re)association of the word with βάλλω, however, the situation changed: as a quasi-synonym of βάλλων, βλεμεαίνων now either (a) needed a direct object (of the thing thrown or of the target hit: Section 13) or (b), in the absence of such a direct object, had to be understood in parallel with intransitive βάλλω (and/or middle βάλλομαι), i.e. as ‘rushing’ or ‘thrusting oneself forward’, as in the following Homeric passages:55 (11) “ἄλλοι μοι δοκέουσι παροίτεροι ἔμμεναι ἵπποι, ἄλλος δ᾿ ἡνίοχος ἰνδάλλεται· αἱ δέ που αὐτοῦ ἔβλαβεν ἐν πεδίῳ, αἳ κεῖσέ γε φέρτεροι ἦσαν· ἤτοι γὰρ τὰς πρῶτα ἴδον περὶ τέρμα βαλούσας, νῦν δ᾿ οὔ πῃ δύναμαι ἰδέειν, πάντῃ δέ μοι ὄσσε Τρωϊκὸν ἂμ πεδίον παπταίνετον εἰσορόωντι.” (Il. 23.459–464) “Other horses seem to me to be in front now, and another charioteer is coming to the fore; those that were stronger up to that point must have been held up on the plain — for indeed, previously I saw them storming around the turning-post, but now I cannot see them any longer although my eyes look around on the Trojan plain as I am watching.” (12) ἠΰτε πάρδαλις εἶσι βαθείης ἐκ ξυλόχοιο ἀνδρὸς θηρητῆρος ἐναντίον, οὐδέ τι θυμῷ

 to Etym. Magn. 783.20–25; cf. Hdn. GG 3.2, 237.11–238.4 Lentz): that dreams fly is an already Homeric concept (cf. Il. 2.71, Od. 11.207–208; Page 1951, 87), and attempts to justify Alcman’s qualification of the dreams as ‘lurking under shady rocks’ (West 1965, 195; cf. already schol. A Alcm. 1.49 PMG), or as alluding to ὑποπετρίδιοι ὄναγροι ‘wild-asses that live in rocky lands’ (Tsantsanoglou 2012, 48–50), seem rather more far-fetched. Within Homer himself, another metathesis has recently been postulated to explain the adjective ἐπίκλοπος ‘cunning, wily’ (e.g. with μύθων: Il. 22.281) as going back to *ἐπίπλοκος (cf. Sappho fr. 188 L.-P. μυθοπλόκος ‘weaving stories’; also v.l. in Od. 13.295 μύθων πλοκίων?) (Le Feuvre 2015, 307–338); in this case, though, the metathesis would have been less essential, given the identical metrical shape of the input and the outcome. 55 Cf. A. Mette and W. Beck in LfgrE 2.35, s.v. βάλλω III; Chantraine 1958, 380.

66  Andreas Willi ταρβεῖ οὐδὲ φοβεῖται, ἐπεί κεν ὑλαγμὸν ἀκούσῃ· εἴ περ γὰρ φθάμενός μιν ἢ οὐτάσῃ ἠὲ βάλῃσιν, ἀλλά τε καὶ περὶ δουρὶ πεπαρμένη οὐκ ἀπολήγει ἀλκῆς, πρίν γ᾿ ἠὲ ξυμβλήμεναι ἠὲ δαμῆναι· ὣς Ἀντήνορος υἱὸς ἀγαυοῦ, δῖος Ἀγήνωρ, οὐκ ἔθελεν φεύγειν […]. (Il. 21.573–580) Just as a leopard comes out of the dense thicket, facing a hunter, and does not shy away with its mind or become frightened when it hears barking — for if he catches up with it and wounds it or hits it, even pierced by a spear it still does not abandon its fierceness until it engages in fight or is overcome –, so the son of noble Antenor, divine Agenor, was not willing to flee […].

Since (b) required no further adjustment, it constituted the most economical — and hence for the oral poet: optimal — way of preserving the formula. For example, in a passage like (1) in Section 1 it made little difference to the overall meaning whether Hector was said to advance among the πρόμαχοι ‘hurling missiles with strength’ or ‘rushing forward with strength’. But as the connection specifically with ‘missiles’ was now lost, σθένεϊ βλεμεαίνων could newly be applied to other contexts as well, whether relating to humans/divinities rushing not necessarily on but also onto and by the battlefield (as in (6) and (2)) or to impetuous animals lunging at an enemy (like the wild boar in (3) and the lion in (4): note that the situation of the lion who σθένεϊ βλεμεαίνει as he attacks the hunting party is remarkably similar to that of the leopard in (8) who is so keen to ξυμβλήμεναι with the hunter and his dogs). Overall, then, only passage (5) remains the odd one out, since σθένεϊ βλεμεαίνει is here predicated of the θυμός and thus internalized; but because a θυμός also cannot truly ‘overflow with strength’ (as per Nikolaev) or ‘cast furious glances’ (as per Athanassakis), the anomaly and metaphorization that is inherent in this one instance has to be acknowledged anyway, no matter how the predicate is otherwise explained (cf. Section 2, n. 2).

 Conclusion If we finally look back at the ancient explanations of σθένεϊ βλεμεαίνων, we can reassert that theory (a) in Section 3, according to which the phrase meant something like σφοδρῶς ἐπιρρωνύμενος and hence referred to displays of impetuous self-confidence and determination, did indeed get closer to the ‘primary’ value of

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  67

the expression than theory (b), which centred around a (par)etymological connection with βλέπω and βλέμμα. However, there is at least a methodological kernel of truth also in (b). In the tradition of oral poetry to which the formula belongs, lexical associations that violate the principled relationships between words which modern comparative grammar allows us to establish did have a major impact on how epic diction was reshaped in response to external pressures such as phonological and morphological change on the one hand and rules of versification on the other. The epic poet who first updated the formula *sthene(h)ei belemainōn into *sthene(h)i blemeainōn did so because his priorities were different from those of today’s philologist: his audience expected a neatly construed hexameter, and here as elsewhere in the modernization of formulae a certain linguistic flexibility was regarded as an acceptable means to the end. Nor was it in any way extravagant for one of his successors then to reinterpret the expression as meaning ‘rushing/lunging forward with strength’ (~ σθένει βάλλων/βαλλόμενος) instead of ‘hurling missiles with strength’ (~ σθένει βέλεμνα βάλλων/ἱείς): again, this was just what poets did all the time, both in the pre-Homeric period that gave rise to some of the Homerische Wörter classically described by Leumann56 and later on in the reception history of the epic lexicon.57 Although modest in scope, the above discussion may thus serve as another reminder of what Albio Cassio’s work has so often shown: that we can only reach a full understanding of ancient texts if we combine the insights offered by modern historical linguistics with a profound awareness of the social and cultural context in which they were created.

Bibliography Athanassakis, A. (1970), “Blemeainô/ablemes (-eôs): Meaning and Possible Etymology”, in: TAPhA 101, 51–61. Cassio, A.C. (1977), Aristofane: Banchettanti (Δαιταλῆς). I frammenti, Pisa. Chadwick, J. (1996), Lexicographica Graeca: Contributions to the Lexicography of Ancient Greek, Oxford. Chantraine, P. (1933), La Formation des noms en grec ancien, Paris. Chantraine, P. (1937), “Grec μενεαίνω”, in: Mélanges linguistiques offerts à M. Holger Pedersen à l’occasion de son soixante-dixième anniversaire, Aarhus/Copenhagen, 205–207.

 56 Leumann 1950; see especially Leumann’s chapter on “Neue Wortbedeutungen und Wörter durch Situationsausdeutung” (Leumann 1950, 208–247). 57 Cf. now, after the remarks by Leumann 1950, 26–34, especially Le Feuvre 2015.

68  Andreas Willi Chantraine, P. (1958), Grammaire homérique, vol. 1: Phonétique et morphologie, 2nd edn., Paris. Debrunner, A. (1907), “Zu den konsonantischen i̯o‑Präsentien im Griechischen”, in: IF 21, 13– 98 and 201–276. Debrunner, A. (1924), “Metrische Kürzung bei Homer”, in: Ἀντίδωρον: Festschrift Jacob Wackernagel zur Vollendung des 70. Lebensjahres am 11. Dezember 1923 gewidmet von Schülern, Freunden und Kollegen, Göttingen, 28–40. Dickey, E. (2007), Ancient Greek Scholarship: A Guide to Finding, Reading, and Understanding Scholia, Commentaries, Lexica, and Grammatical Treatises, from Their Beginnings to the Byzantine Period, New York/Oxford. Fraenkel, E. (1906), Griechische Denominativa in ihrer geschichtlichen Entwicklung und Verbreitung, Göttingen. Gow, A.S.F. (1951), “Nicandrea: With Reference to Liddell and Scott, ed. 9”, in: CQ 1, 95–118. Gow, A.S.F./Scholfield, A.F. (1953), Nicander: The Poems and Poetical Fragments, Cambridge. Höckmann, O. (1980), “X. Lanze und Speer”, in: H.-G. Buchholz (ed.), Archaeologia Homerica, vol. 1, chapter E: Kriegswesen: Teil 2. Angriffswaffen: Schwert, Dolch, Lanze, Speer, Keule, Göttingen, E 275–319. Hoekstra, A. (1965), Homeric Modifications of Formulaic Prototypes: Studies in the Development of Greek Epic Diction, Amsterdam. Jacques, J.-M. (2007), Nicandre: Oeuvres, vol. 3: Les Alexipharmaques, Lieux parallèles du livre XIII des Iatrica d’Aétius, Paris. Janko, R. (1992), The Iliad: A Commentary, vol. 4: Books 13–16, Cambridge. Le Feuvre, C. (2015), Ὅμηρος δύσγνωστος: Réinterprétations de termes homériques en grec archaïque et classique, Genève. Leumann, M. (1950), Homerische Wörter, Basel. Matthews, V.J. (1974), Panyassis of Halikarnassos: Text and Commentary, Leiden. Meissner, T. (2006), S-stem Nouns and Adjectives in Greek and Proto-Indo-European: A Diachronic Study in Word Formation, Oxford. Meister, K. (1921), Die homerische Kunstsprache, Leipzig. Nikolaev, A. (2017), “σθένεϊ βλεμεαίνων”, in: Glotta 93, 117–134. Page, D.L. (1951), Alcman: The Partheneion, Oxford. Risch, E. (1974), Wortbildung der homerischen Sprache, 2nd edn., Berlin/New York. Roth, P. (1991), “Zu Panyassis F 12–14”, in: RhM 134, 238–252. Schwyzer, E. (1939), Griechische Grammatik, vol. 1: Allgemeiner Teil, Lautlehre, Wortbildung, Flexion, München. Schwyzer, E./Debrunner, A. (1950), Griechische Grammatik, vol. 2: Syntax und syntaktische Stilistik, München. Solmsen, F. (1909), Beiträge zur griechischen Wortforschung: Erster Teil, Strassburg. Spatafora, G. (2007), Theriaká e Alexiphármaka: Introduzione, traduzione e commento, Roma. Specht, F. (1936), “Griechische Miszellen: 6. Zur Vermeidung von Wörtern mit drei kurzen Silben”, in: ZVS 63, 207–226. Thurneysen, R. (1907), “Etymologisches und Grammatisches”, in: IF 21, 175–180. Trümpy, H. (1950), Kriegerische Fachausdrücke im griechischen Epos: Untersuchungen zum Wortschatze Homers, Freiburg/Schweiz. Tsantsanoglou, K. (2012), Of Golden Manes and Silvery Faces: The Partheneion 1 of Alcman, Berlin/Boston.

σθένεϊ βλεμεαίνων: Origin and Evolution of a Homeric Formula  69

van Thiel, H. (2014), Scholia D in Iliadem: Proecdosis aucta et correctior, Köln (accessible at http://kups.ub.uni-koeln.de/id/eprint/5586). West, M.L. (1965), “Alcmanica”, in: CQ 15, 188–202. West, M.L. (1982), Greek Metre, Oxford. Witte, K. (1913), “Homeros. B) Sprache”, in: RE 8.2, 2213–2247.

Lucien van Beek and Laura Migliori

Active versus Middle Perfect in Homeric Greek: Synchrony and Diachrony Abstract: While there is no doubt that the middle perfect stem is a secondary addition to the verbal system of Ancient Greek, the reasons for its creation beside the older active perfect are rarely discussed. On the one hand, it has been claimed that the middle perfect stem was first introduced in the pluperfect for morphological reasons, in order to supply the active perfect with a past tense (Chantraine 1927); on the other hand, it has been argued that the difference lies in the role of the subject (Daues 2006). In this contribution, we propose that active and middle perfects fulfilled two different derivational functions. Originally, the active perfect transformed certain types of dynamic events into a state (not necessarily a resulting state) of the subject; the middle perfect was introduced to indicate the resulting state of the object with transitive verbs. Subsequently, the middle perfect gradually replaced the active perfect with deponent verbs, but the active perfect was left untouched in large parts of its original domain. Finally, we argue that the perfect stem transformed a complex predicate into a simple one. Keywords: Perfect, Ancient Greek, voice, verb class, thematic roles.

 Introduction The nature of the distinction between active and middle perfect in the oldest stages of Ancient Greek is a neglected problem, both in Greek Linguistics and Indo-European Studies.1 No opposition between the active and middle voice can be reconstructed for the perfect in the earliest stages of PIE, and still in Homeric Greek, only a handful of verbs have both an active and a middle perfect stem. The most salient (if rather atypical) example is βάλλω ‘throw, hit’, with an active pluperfect βεβλήκει ‘had hit’ beside a middle perfect βέβληται ‘has been hit’. It is

 1 Lucien van Beek’s contribution to this paper was made possible by the support of a VENI grant from NWO (Dutch Organization for Scientific Research) for the project Unraveling Homer’s Language. https://doi.org/10.1515/9783110621747-005

72  Lucien van Beek and Laura Migliori only in the classical language that such active/passive distinctions become systematic. Thus, in view of the low functional load of the middle perfect, it is unclear how the voice distinction originated, and whether it had a function.2 The first systematic approach to the problem was made by Chantraine (1927), who proposed that middle endings were first introduced in the participle and the pluperfect for morphological reasons: this allowed speakers to supply the perfect with a past tense. This scenario has hardly found any supporters; most handbooks are still content to merely signal the emergence of a middle perfect, noting that the voice distinction that already existed in the present/aorist system was transferred to the perfect.3 The same approach characterizes a more recent treatment of the problem by Haug (2008). What is still missing is a functional motivation for introducing these forms. In this article, we propose a new explanation, building on the old observation that middle perfects often function as passives,4 but also taking into account the event structure of verbal lexemes. We argue that the possibility of forming an active perfect was related to event structure. In PIE, this possibility did not necessarily depend on the existence of a middle, as is often assumed, but forming a perfect was simply the canonical way of transforming certain types of dynamic predicates into stative ones. What this ‘active’ perfect was unable to do, however, was to indicate the resulting state of the object of transitive verbs. It is for this purpose, we submit, that the middle perfect was created.5

 2 The problem of the origin of the middle perfect as a category is explicitly signaled by Rix 1992, 195: “[d]ie morphologische Übereinstimmung mit dem Perf. Med. des Arischen deutet auf voreinzelsprachliches Alter der Formen, aber nicht notwendig auch der Kategorien, deren Entstehung noch nicht geklärt ist”. Cf. also LIV 22. 3 Clackson 2007, 120 speaks of an “analogical extension of the active and middle distinction in the present and aorist”, but gives no further motivation. See also Watkins 1969, 131: “[d]ie letzte Entwicklung ist die Ausbreitung der athematischen Medialendungen auf das Perfekt, wo sie zwar einen formalen Gegensatz zwischen Aktiv und Passiv herstellen, aber praktisch in ihrer Bedeutung mit dem alten statisch-intransitiven ,Aktiv’-Perfekt zusammengefallen sind und es tatsächlich z.T. ersetzen”. 4 See e.g. Schwyzer/Debrunner 1950, 237 on passive interpretations of the Greek middle in general: “[b]esonders häufig sind dabei wegen der Zustandsbedeutung des Perfekts Formen des Perf. Med., das wohl sekundär, aber früh zum Perf. Akt. hinzugebildet war”. See also Chantraine 1953, 182. 5 Something along these lines may have been intended by Kümmel 2000, 69 when noting that the Indo-European perfect “muss […] speziell den beim ersten Aktanten (auch Agens) des Grundverbs resultierenden Zustand bezeichnet haben, es war also nicht in der Lage, die Valenz und Rektion zu verändern (wie dies für das passive Zustandsperfekt und das Verbaladjektiv auf *-tógilt)”.

Active versus Middle Perfect in Homeric Greek  73

Our hypothesis will be tested by analyzing and categorizing the semantics of all primary verbs with active and/or middle perfect stems in Homer. We will first sketch our views on the semantics of the perfect stem in general; then we will discuss previous analyses of the active vs. middle distinction in the perfect, including the question as to whether there was a special derivational relation between the ‘active’ perfect and the middle voice. Finally, we will consider the behaviour of different semantic verb classes from a theoretical and cross-linguistic perspective.

 The Semantics of the Perfect Stem Our basic premise is that the early Greek perfect denoted a non-dynamic event, i.e. a state, and that this reflects the PIE situation.6 The perfect stem can be derived from the eventive (present or aorist) stem of a verb, and it may acquire different readings depending on the type of event described by these stems. The reading that usually receives most attention is the so-called ‘stative-resultative’: τέθνηκε means ‘he is dead’, denoting the state following (or resulting from) the punctual event ἔθανε ‘he died’. The stative-resultative perfect does not refer to a completed event, but it presupposes that such an event took place and may, therefore, optionally invoke it in the background (see Kümmel 2000, 66; Allan 2016, 103).7 It is widely held that the PIE perfect primarily formed stativeresultatives to telic roots.8 In support of this, it is pointed out that the perfect is continued as a past tense in language groups like Italic and Germanic, and that plain statives may develop secondarily by lexicalization, e.g. *woid‑h2e ‘I know’ (οἶδα, Ved. véda, Gothic wait) from earlier ‘I have witnessed’ or ‘I have found’ (in either case with the implicature ‘and therefore I know’).

 6 In terms of the Vendlerian classification, which we adopt here, a state is a non-dynamic activity, accomplishment or achievement. 7 In literature with a more typological orientation (e.g. Haug 2008, 292), the term ‘resultative’ is used in this sense. We prefer, however, to reserve that term for constructions of the type John has painted the wall green, which indicate that the object of a transitive verb has undergone a change of state and that the subject has caused this change of state. In the literature about the Greek perfect, such constructions are sometimes called ‘object-resultative’. 8 Within Indo-European Studies, cf. Kümmel 2000, 65−71; Clackson 2007, 121; Fortson 2010, 105; within Greek Linguistics, Haug 2008 and most recently Allan 2016, 103−104.

74  Lucien van Beek and Laura Migliori There are, however, a number of problems with this view.9 First, the alleged primacy of stative-resultatives does not explain how the so-called ‘intensive’ readings fit into the picture. This term (a misnomer)10 is traditionally used as a cover for, among other readings, sound verbs (such as μέμυκε ‘bellows, lows’) and plain statives (like γέγηθα ‘be glad’).11 Secondly, and more importantly, there is ample evidence that the distinction present/aorist vs. perfect in Homeric Greek may correspond to dynamic vs. non-dynamic events, as has been illustrated by Berrettoni (1972).12 Consider the following examples for the verb ἀλάομαι ‘wander, roam’, perfect ἀλάλημαι. In (1), Odysseus answers the soul of his mother Anticleia, who asks what he is doing in the Netherworld: (1)

μῆτερ ἐμή, χρειώ με κατήγαγεν εἰς Ἀΐδαο ψυχῇ χρησόμενον Θηβαίου Τειρεσίαο· οὐ γάρ πω σχεδὸν ἦλθον Ἀχαιΐδος οὐδέ πω ἁμῆς γῆς ἐπέβην, ἀλλ’ αἰὲν ἔχων ἀλάλημαι ὀϊζύν, ἐξ οὗ τὰ πρώτισθ’ ἑπόμην Ἀγαμέμνονι δίῳ […]. (Od. 11.164−168) Mother of mine, necessity has brought me down into the house of Hades, as I have to obtain an oracle from the soul of the Theban Tiresias. For I have not yet come close to Achaea nor yet have I set foot on the shore of our own land, but I have been wandering continuously, subject to misery, from the time I first followed bright Agamemnon […].

In (2), Odysseus asks Nestor and Agamemnon, who have woken him up in the middle of the night: (2)

τίφθ’ οὕτω κατὰ νῆας ἀνὰ στρατὸν οἶοι ἀλᾶσθε νύκτα δι’ ἀμβροσίην, ὅ τι δὴ χρειὼ τόσον ἵκει; (Il. 10.141−142)

 9 For a general and, in our view, convincing criticism of what he calls “nactostatic primacy”, see Willi 2018, 232−244. 10 The term ‘anomalous’, which is also widely used, is less misleading but equally uninformative. 11 Recently, an interesting attempt has been made by Magni 2017 to analyze a wide range of perfect readings (including sound verbs) under the header of verbal plurality. 12 Many details of Berrettoni’s analysis have been followed in more recent times by e.g. Romagno 2005 and Willi 2018, 229−237.

Active versus Middle Perfect in Homeric Greek  75

How come you are wandering like this alone by the ships, across the camp, through the divine night? What need so great has come on you?

In passage (1), the perfect ἀλάλημαι is accompanied by the adverb αἰέν indicating a lasting condition of Odysseus, as well as by the participial clause ἔχων […] ὀϊζύν, another indication of his condition. The starting point of this state is indicated by the phrase beginning with ἐξ οὗ. In passage (2), by contrast, the present ἀλᾶσθε is accompanied by a precise indication of time, νύκτα δι’ ἀμβροσίην, showing that the wandering referred to actually occurs. By taking into account this type of clue, Berrettoni (1972, 82−87) shows that the present stem denotes an actual, effective wandering that can be localized in time, whereas the perfect refers to a lasting or characteristic condition of its subject. Various active perfects (both with transitive and intransitive verbs) display the same function of presenting an activity or accomplishment as a property of the subject. An example is κεύθω ‘hide, withhold’ (transitive), which in the present stem denotes a volitional, controlled activity. In Homer, the perfect stem κεκευθ- occurs three times: in all these cases the agentive role of the subject is annulled as a consequence of the perfect stem’s non-dynamicity.13 For instance, in ὅσα τε πτόλις ἥδε κέκευθε ‘as much (treasure) as this city conceals’ (Il. 22.118), the motive for using the perfect is that a city, being an inanimate entity, cannot dynamically withhold an object.14 It seems attractive to also analyze perfects like ἔολπα (ἔλπομαι ‘think’) and μέμηλε (μέλω ‘concern’), which are traditionally labelled ‘intensive’, in the same way as ἀλάλημαι. An important and cross-linguistically well-attested type is the so-called existential perfect, which indicates that an event has happened at least once during some time in the past.15 Consider the following, much-discussed passage, where ‘the man in the crowd’ discusses the way Odysseus has just restrained Thersites:

 13 The non-agentivity or low transitivity of the perfect of transitive verbs is also confirmed by the properties of its objects (for which see Section 6.1.1). These properties, however, are to be seen as consequences of its non-dynamicity. 14 At Od. 9.348, ὄφρ’ εἰδῇς, οἷόν τι ποτὸν τόδε νηῦς ἐκεκεύθει / ἡμετέρη ‘so that you may find out what kind of drink this is which our ship contained’, the subject is again inanimate. At Od. 3.18, εἴδομεν ἥν τινα μῆτιν ἐνὶ στήθεσσι κέκευθε ‘we will find out what counsel he (Nestor) hides in his chest’, the subject is animate. The pragmatic implication of using the perfect here is not that Nestor is wilfully hiding his advice, but merely that this advice is both desirable and presently unavailable to the speaker. 15 See Comrie 1976, 58−60, who, however, calls this type “experiential”. With Allan (2016, 105), we prefer to reserve the term ‘experiential’ for those existential perfects where the subject has undergone a change of mental state (see below). The existential perfect has affinities with the

76  Lucien van Beek and Laura Migliori (3) ὢ πόποι ἦ δὴ μυρί’ Ὀδυσσεὺς ἐσθλὰ ἔοργε βουλάς τ’ ἐξάρχων ἀγαθὰς πόλεμόν τε κορύσσων· νῦν δὲ τόδε μέγ’ ἄριστον ἐν Ἀργείοισιν ἔρεξεν ὃς τὸν λωβητῆρα ἐπεσβόλον ἔσχ’ ἀγοράων. (Il. 2.272−275) Man! Odysseus has truly performed countless noble deeds, initiating good plans and organizing war; but now he has done this thing among the Greeks, far the best of all: he has made this word-vomiting nuisance stop speaking.

In this example, the verbal action is described as characteristic of the subject. Although the action ἐσθλὰ ἔοργε may have occurred any number of times before the moment of speaking,16 this use of the perfect does not differ substantially from that in another well-known example, βοὸς […] μήπω τετοκυίης ‘a cow that has never calved’ (Hes. Op. 591).17 The existential perfect is frequently encountered with negated predicates (as in the last example), with indefinite temporal adverbs, and with indefinite objects. A subtype of the existential perfect is the experiential perfect, indicating the subject’s experience or knowledge resulting from a past event.18 Examples are ἦ γὰρ πρόσθεν μιν ὄπωπα ‘I have seen him before’ (Od. 17.371) and, with a middle perfect, εἰπὲ δέ μοι Πηλῆος ἀμύμονος εἴ τι πέπυσσαι ‘tell me if you have heard anything about Peleus’ (Od. 11.494). The boundary between this and other uses of the existential perfect is not clear-cut. The existential perfect is frequent in Homer, but unfortunately is often confused with the stative-resultative. This confusion is due to formulations like “eine am Subjekt nachwirkende vergangene Handlung” (Wackernagel 1904, 4), which suggests that the perfect denotes an actual condition of the subject. Indeed, in experiential perfects like πρόσθεν μιν ὄπωπα, the pragmatic implication is that the subject has a certain memory. What the existential perfect does, however, is

 Anterior, but the latter is a more vaguely defined cross-linguistic category encompassing the functions stative-resultative, experiential, persistent situation, and hodiernal past. 16 Wackernagel 1904, 4 distinguished this use of the perfect as follows: “wenn es gilt, einen Complex kontinuierlicher Handlungen zu bezeichnen, die in der Gegenwart ihren Endpunkt haben”. In reality, however, the ‘continuity’ of actions is irrelevant: ἐσθλὰ ἔοργε refers to a discrete (even if potentially uncountable) set of noble deeds. Various scholars distinguish a special type of perfect for cases like (3) which refer to an iterated event: Kümmel 2000, 73 speaks of a “comprehensive perfect”; Ruijgh 1991 and Ruijgh 2004 of a “totalizing-iterative” perfect. 17 Romagno 2005, 62−81 speaks of the “valore qualificativo” of the perfect and analyzes many other active perfects in the same way. 18 For this distinction between existential and experiential perfect, see Allan 2016, 105.

Active versus Middle Perfect in Homeric Greek  77

merely to predicate the fact that a past event applies.19 In other words, in this use the perfect does not narrate, but asserts something about the subject, adds a qualification (as observed already by Berrettoni 1972). In sum, the point to be retained is that the perfect did not necessarily refer to a resulting condition of the subject. Rather, its main encompassing function was to transform a dynamic event into a non-dynamic event (= a state). Depending on root semantics, event structure and context, this event may appear as a resulting state or an experience, or it may simply qualify the subject. As for the temporal dimension, the event invoked by the perfect may be a condition that applies at reference time (cf. ἀλάλημαι), it may relate a past event to reference time (cf. ὄπωπα), or it may be extra-temporal (cf. stative perfects like εἴωθα, ἔοικα).

 The Problem of the Creation of a Middle Perfect Let us now zoom in on the problem of the emergence of a middle perfect, and first focus on the way this was problematized by Chantraine (1927). As we have seen, active perfects are often paired with active present or aorist forms, mainly intransitives (τέθνηκα : θνήσκω ‘die’) but also transitives (κέκευθα : κεύθω ‘withhold’). In Chantraine’s view, however, these forms are “nullement instructif pour la théorie des désinences du parfait” (1927, 24) because they fit neatly in the Greek conjugational system with its distinctions of aspect and voice. Instead, he emphasizes the fact that all lexicalized perfects without other tense/aspect stems (e.g. οἶδα, ἔοικα, εἴωθα) are active perfects (1927, 24−26), noting that no similarly isolated middle perfects exist. In his view, such active perfects are debris from a more original system with many defective verbs. Secondly, many old active perfects are formed to roots in which the middle voice plays a large part. In terms of alignment, these perfects match a middle present, whereas the corresponding active present has causative sense. An example is ἐγείρω ‘wake up’ (transitive), with a perfect ἐγρήγορα ‘be awake’ matching the intransitive middle present ἐγείρομαι ‘wake up’, aorist ἤγρετο. Similar pairs are also known from Indo-Iranian and Latin, e.g. Ved. pres. vártate ‘turns’ beside

 19 Kümmel 2000, 73 also remarks that such perfects do not indicate a resulting state of the subject, but instead give a characterization of the subject. Interestingly, “[d]as Perfekt in dieser Gebrauchsweise musste im Indoiranischen immer verwendet werden bei Aussagen der Art, dass die betreffende Handlung irgendwann (einmal oder mehrfach) in der Vergangenheit stattgefunden hat, ohne dass genauer festgelegt wird, wann” (Kümmel 2000, 74; our emphasis).

78  Lucien van Beek and Laura Migliori perf. vavárta, Old Lat. pres. revortor beside perf. revorti, all intransitive. The intransitive semantics of the perfect to PIE *h1ger- is confirmed by Ved. jāgā́ra, which like ἐγρήγορα also means ‘be awake’. In Chantraine’s view, which is still shared by many scholars today (e.g. Clackson 2007, 120), this pairing between active perfect and middle present forms was originally paradigmatic.20 Therefore, Chantraine concluded that the PIE perfect had ‘active’ endings only, that there was no functional need for a middle perfect, and that the middle perfect was (by and large) introduced after the proto-language.21 But if there was no need for a middle perfect, why was it eventually formed? In Chantraine’s view, the motivation can be discovered by considering the few perfects in Homer that have both active and middle forms. In most cases, he claims that there is no clear semantic difference between active and middle forms (1927, 48−54).22 His examples include active ἔοικε ‘looks like’ beside middle pluperfect ἤϊκτο ‘looked like’, and active ἔμμορε ‘has as a share’ beside middle pluperfect εἵμαρτο ‘is destined’. In his view, it is of paramount importance that these (and some other) middle forms are not perfects, but pluperfects. This distribution, though based on a small number of verbs, leads him to posit the following scenario: originally, the perfect *wewoike paired with a pluperfect *(e)wewoik-t, but after the latter form yielded *(e)wewoi (loss of word-final obstruents), it became morphologically opaque. A morphologically transparent middle pluperfect *(e)wewikto was then coined to mend this problem. Thus, the motivation for introducing the middle endings was morphological.23 At a later stage, both middle perfect indicatives and active pluperfects were created.

 20 Not long after Chantraine’s study, Stang 1932 and Kuryłowicz 1932 independently demonstrated the striking similarity between the singular endings of the PIE perfect (*‑h2e, *‑th2e, *‑e) and those of the middle (*‑h2, *‑th2o, *‑(t)o). However, that the perfect and middle endings are etymologically related in pre-PIE does not entail that these categories were derivationally or paradigmatically connected in PIE. 21 The rare occurrence of functional diathesis oppositions in the perfect stem has also been taken as an indication that such oppositions were devoid of semantic load (e.g. Haug 2008, 296−299). There are, however, a number of verbs where the distinction is clearly functional: see Sections 4 and 6.1.2. 22 “Dans les exemples les plus anciens, la flexion moyenne est usuelle au parfait, mais sans se distinguer par aucune nuance de l’actif” (Chantraine 1927, 54). 23 Like Chantraine, Haug also views morphology as the main driving force behind the spread of the middle perfect, but without assigning a pivotal role to the pluperfect. Instead, he speaks of “a tendency to normalize the morphological expression of diathesis: since the perfect patterns semantically with the present and aorist middle and not the active, it gets middle voice morphology” (2008, 298). He does not explain, however, why some transitive verbs retain an active perfect while others replace this form with a middle perfect. As for the Homeric verbs with active

Active versus Middle Perfect in Homeric Greek  79

Several important objections have been advanced against this scenario.24 Cases like ἔοικε beside ἤϊκτο stand isolated,25 and the absence of primary forms like 3rd sing. perf. *ἔϊκται against only five instances of ἤϊκτο/ἔϊκτο might be coincidental. Moreover, if the perfect was aligned with the middle voice anyway, one might wonder why speakers created an active pluperfect later on and kept both active and middle perfects in productive use, instead of simply generalizing the middle conjugation and gradually eliminating the active perfect indicative.26 A more promising avenue, therefore, would be to ask whether the middle perfect shares any functions with other parts of the middle voice. Recently, an attempt in this direction has been made by Daues (2006), who argues that most instances of oppositional middle perfects in Homer display canonical functions of the middle voice, such as self-beneficiary (indirect-reflexive), subject-affectedness generally, and passivity. According to Kümmel (2000, 92) a similar conclusion also holds for the Indo-Iranian evidence. While we doubt that Daues’ conclusions concerning subject-affectedness in Homeric middle perfects follow from the evidence,27 the passive use is undeniable and appears to be widespread already in Homer and Mycenaean.28

 perfect beside middle pluperfect, Haug suggests that this might represent an intermediate stage of the transition from active to middle perfect, but in our view this is pure speculation. 24 Cf. the details in Debrunner 1928, 287−288. 25 Debrunner 1928, 288 rightly criticizes a number of Chantraine’s examples for this alternation, noting that it constitutes “keine Grundlage für einen großen neuen Typus”. 26 This seems to be what happened between Homeric Greek and Classical Attic, witness ἔφθαρμαι ‘am lost’ replacing older ἔφθορα and similar cases (cf. Haug 2008, 299−300). This does not explain, however, why the active and middle perfect coexisted for such a long time (already long before our attestations of Mycenaean) without developing a functional distinction. 27 While we agree that transitivity plays a role in the active/middle alternation as well as in the distribution of perfect forms in Homeric Greek, we disagree with Daues on a number of points. Firstly, she does not consider the properties of the objects of the perfect forms, which seem essential to us (cf. Section 6). Moreover, we are not convinced that the middle perfect stem functioned as a repertoire for metaphorical usages, as opposed to the active (Daues 2006, 11). She discusses the phrases ὅσσα τοι ἐκπέποται καὶ ἐδήδοται ‘all that has been drunk and eaten up’ (Od. 22.56); χρήματα δ᾽ αὖτε κακῶς βεβρώσεται ‘as for his possessions, they will be badly devoured’ (Od. 2.203); δαιτὸς κεκορήμεθα θυμὸν ἐΐσης ‘we have satisfied our appetite with the fair banquet’ (Od. 8.98); πολέων κεκορήμεθ᾽ ἀέθλων ‘we are fed up with the numerous trials’ (Od. 23.350), but it should be noted that the active perfect is often used with metaphorical meaning as well; therefore, this claim fails to nail down the exact distinction between middle and active perfect forms. 28 “Die passive Bedeutung des Mediums ist bei Homer und im Veda im Pf. ganz besonders häufig” (Debrunner 1928, 288). Cf. also Chantraine 1953, 182; for the Mycenaean evidence, see below p. 94 n. 63.

80  Lucien van Beek and Laura Migliori Therefore, in this study we hypothesize that the middle perfect stem was created precisely in order to express the resulting state of the object of (certain types of) transitive verbs, because the active perfect could not do this. In other words, the introduction of middle endings with the perfect stem corresponded to the presence of external causation in the predication.29 This idea has been hinted at by previous scholars, but to our knowledge its details and consequences have not yet been fully explored.30 Moreover, the position of active perfects is frequently misunderstood when it is stated that they, too, could originally be used with passive function.31

 Resulting States: Active or Middle Perfect? In Section 2, we have discussed perfects that do not represent a dynamic event (activity or achievement) as actually taking place, but predicate an event as characteristic of the subject (existential, experiential, etc.).32 In what follows, we will pay more attention to resulting states. Perfects denoting a resulting state indicate that their subject has a certain lasting property or condition that came about as the result of a past action in which the same entity acted as Patient. This situation is especially frequent with middle perfects, e.g. τίω ‘honour’, perf. ptcpl. τετιμένος ‘honoured’; εἴρω ‘string’, perf. ἔερται ‘is strung’; aor. ἔπεφνον ‘slew’, perf. πέφαται ‘has been slain, lies dead’. In such cases, the verb is usually high on the transitivity scale and has a prototypical Agent or Causer argument: for instance, the state of being honoured usually presupposes that someone actually

 29 On the relationship between agentivity and the presence of an affected object, see Kratzer 1996. 30 For instance, Schwyzer/Debrunner 1950, 264 consider “das passive Perf. transitiver Verba älter als das aktive”, although under ‘active perfect’ they seem to understand the later oppositional κ‑ and aspirated perfects of the classical language. 31 Chantraine 1927, 90 thinks that the frequency of passive readings of middle perfects in Classical Attic prose authors replaces an older situation in which the active perfect could have passive function. Cf. Schwyzer/Debrunner 1950, 237: “[a]uch das Perfekt Akt. alter Bildung und intransitiv-zuständlicher Bedeutung konnte einem Passiv recht nahe kommen”. More explicitly Ruijgh 2011, 286, on Hom. τετευχώς and Myc. te‑tu‑ko‑wo‑a: “[i]n Proto-Indo-European, the ‘active’ perfect form could be used for expressing the resulting state of a passive subject”; in similar fashion Slings 1987, 63. 32 This type of perfect is found with active and middle perfects alike (cf. ἀλάλημαι, πέπυσμαι), but it should be noted that middle perfects occurring beside deponent eventive stems can be secondary replacements of active perfects (cf. Section 5.5).

Active versus Middle Perfect in Homeric Greek  81

conferred this honour at some point. Similarly, being strung is typically the result of a volitional act of stringing objects together. With active perfects, however, the state of the subject is usually not the result of a prototypical transitive action. In most cases, the subject of the perfect is also the subject of the corresponding non-perfect stems. We may distinguish: (a) Active perfects of intransitive verbs denoting accomplishments (especially change-of-state, such as ‘die’, but also verbs of motion like ‘come’), which belong to the category of unaccusatives.33 (b) Active perfects of activity verbs (sound verbs; other so-called ‘intensive’ perfects). (c) Active perfects of transitive verbs, provided that no change of state in the object is expressed (see Section 6); such cases are usually existential perfects. There are, however, also some active perfects that (from a morphosyntactic perspective) seem to behave like the ‘passive’ middle perfects just mentioned. These active perfects stand beside a transitive eventive stem and denote a resulting state of the Patient of the event.34 For instance, πέπηγε ‘is stuck, sticks (intr.)’ (as in Il. 3.135 παρὰ δ’ ἔγχεα μακρὰ πέπηγεν ‘and beside, long spears are stuck [in the ground]’) corresponds to the transitive aorist πῆξε ‘fixed, stuck’ (a spear into someone’s body/the ground). Our question is, therefore: why is it that a resulting state of the Patient is usually expressed by a middle perfect, but sometimes by an active perfect? In other words, why do we find πέπηγε, rather than πέπηκται, as the perfect of πήγνυμι? As we have seen, according to Chantraine, active perfects were aligned with middle presents and intransitive aorists: both from a semantic and a morphological perspective, the perfect belongs to the domain of the middle voice. This supposed original situation is reflected in Homer only partially, because older active perfects may have been replaced by middle perfects, and because new active presents have been created. However, although a connection between perfect and middle is widely accepted, disagreement has recently been voiced by Romagno in her monograph on the Homeric perfect (Romagno 2005). Investigating the relationship between

 33 The label ‘unaccusative’ indicates all those intransitives the subject of which has the properties of a Patient. From a syntactic perspective, this argument is considered as originally occupying the same locus as the direct object of transitive verbs; this fact explains their common properties (cf. Perlmutter 1978). Unaccusative verbs differ sharply from unergatives, which generally express activities and are characterized by an agentive subject. 34 An overview of the examples can be found in Romagno 2005, 81−89.

82  Lucien van Beek and Laura Migliori the semantics of the perfect and the actionality of verbal predicates within the framework of thematic roles,35 she claims that the perfect denotes a state of the subject, turning it into an inactive participant. She not only denies any connection between the PIE perfect and the middle voice, but even thinks that the two categories were incompatible and could not be derived from the same root. Her reasons are as follows. First, many active perfects are aligned with morphologically active presents and aorists (e.g. θνήσκω, perf. τέθνηκα); secondly, most deponents (e.g. νέομαι ‘return’, ἔρχομαι ‘come’) have no perfect at all, and when they do, it is usually a middle perfect (and therefore possibly secondary); and thirdly, she argues that the attested pairings between medium tantum and active perfect might be secondary.36 While we disagree with Romagno on various issues, we partly agree that there was no inherent connection between perfect and middle in the proto-language: perfects must already have been formed to morphologically active presents, too. However, we consider Romagno’s rejection of old pairings of the type γίγνομαι : γέγονα to be too categorical. In fact, from a morphological point of view the PIE perfect could be formed to both active and middle eventive stems; the real issue is to find out under which semantic conditions a perfect could be formed. In what follows, we will therefore ask whether there were semantic restrictions on creating active perfects: could they originally be formed beside any eventive stem (whether transitive or intransitive), or was there a ban on the formation of perfects beside, for instance, high-transitivity lexical items? In order to reach an answer to this question, the distribution between active and middle perfects in Homeric Greek will be investigated, considering not only

 35 Theta-roles (or semantic roles) express the function of an argument. They are strictly dependent on the syntactic location of arguments: for instance, the role of Agent can only be assigned to an external argument (subject), while the role of Patient is always assigned to the complement of a verb (internal argument). Therefore, they are a crucial element of the syntax/semantics interface and are essential for the interpretation of a sentence. For more details about theta-theory, see Reinhart 2002. 36 Romagno 2005, 43−44. In her view, middles already contain a state predicate in their underlying logical structure, so that the perfect would be superfluous. This cannot be correct, since most uses of the Greek middle share the feature ‘subject-affectedness’, whether in spontaneous processes, body motion, mental processes and activities, reciprocal middles, indirect reflexives, or verbs of grooming (cf. Allan 2003). Only middles that continue old PIE statives (κεῖμαι, ἧμαι, εἷμαι) have a state predicate, and indeed perfects of these verbs do not occur. Note that Romagno 2005, 29, basing herself on articles by Lazzeroni, adheres to the view that there was no separate category ‘stative’ in PIE, a view which we (like the majority of Indo-Europeanists today) do not share.

Active versus Middle Perfect in Homeric Greek  83

the semantics of the concrete perfect stems, but also the underlying event structure of the lexeme. We consider all cases where the active perfect denotes a resulting state (rather than a plain state, an experience, or an extra-temporal property of the subject: see Section 2), and ask which types of events are represented. Throughout the discussion, it must be kept in mind that the middle perfect may have gained territory at the expense of the active.

 The Material We will first consider the entirety of the Homeric evidence37 for active perfects with a subject matching that of a corresponding middle present and/or intransitive aorist.38 After sifting through the data, we are left with a collection of 21 active perfects standing beside transitive active presents and/or aorists. For part of this evidence, previous authors (e.g. Chantraine 1927) already recognized that the transitive formation was formed as a secondary causative or factitive. We will first discuss and illustrate these findings, and then present a new analysis of some more stubborn examples of active perfects with passive meaning. In this way, a link between the possibility to form an active perfect and the absence of a prototypical semantic Causer role in the event structure will be established.

. Anticausatives First, a number of verbs denote spontaneous telic processes or transformations and have a factitive active: σήπομαι ‘rot’ (perf. σέσηπα ‘be rotten’); τήκομαι ‘melt, dissolve’ (perf. τέτηκα ‘be dissolved’); περιτρέφομαι ‘congeal’ (perf. περὶ […]

 37 We leave denominatives out of consideration because they were automatically assigned a middle perfect in early Greek. Moreover, the absence of perfects to denominative verbs in IndoAryan seems to imply that PIE did not have this possibility either. 38 Cf. Chantraine 1927, 26−37 for a discussion of the Homeric evidence, and Chantraine 1927, 37−44 for the post-Homeric evidence. In the following discussion, we have left aside μέμονα ‘strive for, be keen at, be willing’ (which may belong either with μαίνομαι ‘rage’ or with μένω ‘wait’, cf. Willi 2018, 235, and therefore cannot be used in this discussion). We have also left aside perfect participles in ‑ηώς of the type βεβαρηώς (on the antiquity of this type, see Hackstein 1997−1998) because these are probably denominal in origin; κεκορηώς (cf. κορέσαι ‘satisfy’, Chantraine 1927, 31) is probably secondary for κεκορημένος.

84  Lucien van Beek and Laura Migliori τέτροφεν ‘has formed a crust’). The event structure of these verbs does not contain an inherent Causer role.39 Indeed, the active of these verbs is generally rare, and in the case of σήπομαι it is not even attested in Homer. A special case is the perfect of τρέφω ‘raise, rear (etc.)’, τέτροφεν. In Homer, it occurs only once: πολλὴ δὲ περὶ χροῒ τέτροφεν ἅλμη (Od. 23.237) ‘a lot of brine has crusted on his skin’, denoting a spontaneous natural process.40 This form (περὶ […] τέτροφεν, in tmesis) belongs to the prefixed middle περιτρέφομαι, attested at Od. 14.477 (καὶ σακέεσσι περιτρέφετο κρύσταλλος ‘and a layer of ice settled on their shields’) and also at Il. 5.903 meaning ‘thicken, curdle’ (of a liquid), which is also the etymological meaning of the PIE root *dhrebh‑. A number of other verbs have an (active or middle) root aorist with anticausative semantics, beside a secondary transitive s‑aorist: δύομαι ‘enter’ (aor. ἔδυν, perf. δέδυκα); ἐγείρομαι ‘wake up’ (aor. ἠγρόμην, perf. ἐγρήγορα ‘to be awake’); ἵσταμαι ‘stand up’ (aor. ἔστην ‘stood up’, perf. ἕστηκα ‘stand’); ὄρνυμαι ‘rise, get up’ (aor. ὦρτο ‘arose’, perf. ὄρωρε ‘arises; has arisen’); φύομαι ‘grow’ (aor. ἔφυν, perf. πέφυκα). That the active presents δύω, ἐγείρω, ἵστημι, ὄρνυμι and φύω (and their transitive sigmatic aorists ἔδυσα, ἤγειρα, ἔστησα, ὦρσα, ἔφυσα) are secondary causatives or factitives is strongly suggested by two facts: first, when available, cognates in other IE languages show intransitive meanings (e.g. Skt. tíṣṭhati ‘stands’; OCS byti ‘be’; Lat. deponent orior ‘rise’, etc.). Secondly, this intransitive meaning also appears in the root aorists *(é-)steh2-t (cf. Ved. Skt. ásthāt), *(é-)bhuH-t (cf. Lat. fuit ‘was’, Ved. Skt. ábhūt ‘was’), *(é-)h3r-to (cf. Skt. ārta ‘got moving’, Hittite arta(ri) ‘stands’).41 For the perfect of PIE *h1ger-, intransitive semantics can be reconstructed by comparing Gk. ἐγρήγορε ‘is awake’ and Ved. jāgā́ra ‘id.’. A third group of verbs denote atelic spontaneous or mental processes (in Vendlerian terms, these are activities): δαίομαι ‘radiate’ (perf. δέδηα); ταράσσομαι ‘be stirred’ (plupf. τετρήχει ‘was in upheaval’); ἔλπομαι ‘think; hope’ (perf. ἔολπα). In Homer, the factitive ἔλπω ‘give hope’ is rare (only in the repeated line Od. 2.91 = 13.380). Semantically, ταράσσομαι/τετρήχει may denote both natural and mental processes. The event structure of this verb contains not a Causer, but a Stimulus. We may therefore assume that the active ταράσσω (aor. ἐτάραξα)  39 The verbs expressing spontaneous processes are, from a typological perspective, a prototypical case of internally caused events (simple predicates), which means that their event structure lacks an Agent/Causer (Levin/Rappaport-Hovav 2005). 40 In post-Homeric Greek, the perfect τέτροφε may belong to τρέφω ‘raise, rear’. 41 For the etymology of the Hittite verb we follow Kloekhorst 2008 s.v. ar‑tta(ri). That the formation of the perfect ὄρωρε might be recent (assuming that ὄρωρε replaces *ōre < PIE *h3e‑h3or‑e) does not preclude this verb from having an old, inherited perfect.

Active versus Middle Perfect in Homeric Greek  85

is a factitive verb; this is also confirmed by its morphology (derived yod‑present and s-aorist) and by the fact that it is often used for natural phenomena (the sea; horses). The active present δαίω, too, is mostly used as a factitive ‘cause to radiate’ (cf. Chantraine 1927, 28), as neatly illustrated by the following passage (cf. also Il. 5.4−7): (4) ἡνίοχοι δ’ ἔκπληγεν, ἐπεὶ ἴδον ἀκάματον πῦρ δεινὸν ὑπὲρ κεφαλῆς μεγαθύμου Πηλεΐωνος δαιόμενον· τὸ δὲ δαῖε θεὰ γλαυκῶπις Ἀθήνη. (Il. 18.225−227) The charioteers were baffled when they saw the untiring, terrible fire over the head of the high-spirited son of Peleus, radiating; the goddess, grey-eyed Athena, made it glitter.

Interestingly, the radiation denoted by δαίομαι in passages like (4) often is caused by a divinity, and this is also true of the sea in the phrase ἐτάραξε δὲ πόντον (Od. 5.291 and 304), which is made rough by Poseidon and Zeus, respectively. The structural presence of (often explicitly mentioned) supernatural Causers is another indication that the active of these verbs is a causative. Two remaining verbs do not fall in the above classification, but previous scholarship is in agreement that they are anticausatives, too. First, beside πείθομαι ‘obey, follow someone’s lead’ (aor. ἐπιθόμην ‘id.’; perf. πέποιθα ‘trust, rely on’), the active πείθω ‘persuade, convince’ (aor. ἔπεισα, etc.) is a causative.42 Although synchronically the perf. πέποιθα might be considered an independent lexeme, it clearly aligns with the middle present/aorist. That the lexeme did not involve a Causer seems to be confirmed by the intransitive semantics of derived forms like πίστις ‘confidence’, πίσυνος ‘relying’.43 Moreover, the same meaning as in πέποιθα is found in Lat. fīdō ‘trust’, with derivations like fidēs ‘trust, loyalty’, foedus ‘treaty’, while the only other ascertained cognate verbs (in Germanic and Slavic, see LIV 71−72) continue an inherited morphological causative *bhoidh‑eye‑. Secondly, δαῆναι ‘learn’ (reduplicated aor. δέδαε ‘teach, instruct’, perf. δεδάηκε ‘has learnt’, δεδαώς ‘able, learned’) has a morphological causative/factitive διδάσκω ‘teach, instruct’ that itself seems to be inherited from PIE. LIV 118−119

 42 Cf. Chantraine 1927, 33; LIV 71−72 with n. 1. 43 Nominals and verbs of the same root, though differing in grammatical category, are nevertheless characterized by the same argument structure. Therefore, restrictions on nominalizations are a good test to corroborate the syntactic/semantic properties of a root (cf. Borer 2005).

86  Lucien van Beek and Laura Migliori gives the root meaning as ‘kundig werden’; the reduplicated aorist is analyzed as a factitive.44

. Verbs of Destruction The causative-anticausative alternation can also be observed with a couple of verbs of destruction. A first case is ἐρείπω ‘cause to crumble’, ἤριπον ‘collapsed, fell down’ (intr.), perf. act. κατερήριπεν ‘has crumbled’. The intransitive meaning of the active thematic aorist suggests an old intransitive verb (cf. also Chantraine 1927, 30 and the examples given in Section 5.1). Interestingly, whereas this aorist always qualifies things or persons falling to the ground, the three instances of the active thematic present ἐρείπω in Homer have the Achaean wall (or part of it) as an object. Here, the lexeme has a different meaning, ‘crumble’. Consider, for instance: (5)

[…] προπάροιθε δὲ Φοῖβος Ἀπόλλων ῥεῖ’ ὄχθας καπέτοιο βαθείης ποσσὶν ἐρείπων ἐς μέσσον κατέβαλλε, γεφύρωσεν δὲ κέλευθον. (Il. 15.354−356) Before them, Apollo made the banks of the deep trench crumble with his feet, without effort, and casting it into the middle, he bridged a path.

This suggests that the basic meaning was intransitive: ‘crumble, fall apart, collapse, disintegrate’. This idea might be confirmed by the etymology of the Lat. cognate rīpa: a river bank is continuously subject to the natural process of erosion.45 The active perfect κατερήριπεν could reflect this older meaning; the middle pluperfect ἐρέριπτο could be interpreted as a passive, ‘had been ruined’ vel sim. In the case of ὄλλυμαι ‘perish’ (aor. ὠλόμην, ὀλέσθαι; perf. ὄλωλα), the active ὄλλυμι, ὤλεσα is a causative. This appears most clearly from the phrase πολὺν ὤλεσα λαόν ‘I have allowed many men to perish, have lost many men’ (said by Agamemnon at Il. 2.115; not *‘I destroyed many men’). There is also a frequent formulaic use ὤλεσε θυμόν ‘he lost his life’ that seems to presuppose the same semantics (< ‘he allowed his life to be taken’, not *‘he destroyed his own life’; the  44 See Willi 2018, 64. 45 Another cognate is the Germanic strong verb *rīfan- (Old Norse rífa ‘tear; be rent, give way’, Middle Eng. rīven ‘tear’ etc., see EDPG s.v. *rīfan‑ ~ *rīpan‑).

Active versus Middle Perfect in Homeric Greek  87

Agent in this construction is indicated in e.g. ὑφ’ Ἕκτορος ἀνδροφόνοιο, Il. 17.616). As for (δια‑)φθείρομαι ‘be wasted, perish’ (perf. διέφθορας), determining the etymological meaning of this verb is not without problems, but an interesting and promising attempt has been made by Jamison 1993, followed by the LIV (‘im Wasser dahintreiben’). Jamison argues that φθείρομαι in nautical contexts means ‘go off course, drift away’ (of ships), ‘be shipwrecked’ (of sailors), and connects this with the meaning of the Indo-Iranian root Ved. kṣar, Avest. ɣžar ‘flow’. She also points out that a meaning ‘go astray’ would excellently fit the only occurrence of the perfect διέφθορας ‘you have lost your senses’ at Il. 15.128. The occurrence of φθείρω at Od. 17.246 αὐτὰρ μῆλα κακοὶ φθείρουσι νομῆες may well have causative or permissive meaning: ‘bad herdsmen let their flocks go astray’, as stressed by Jamison.46 Thus, again it seems that the intransitive (spontaneous process) use is oldest, and that active φθείρω is a secondary causative. Furthermore, consider ἄγνυμαι ‘break, shatter’, aor. ἐάγη, perf. ἔᾱγε ‘is broken’. The perfect ἔᾱγε, though not attested in Homer (ἐάγῃ at Il. 11.559 is an intransitive aorist, cf. Chantraine 1958, 18), is found in Hesiod and Sappho and is therefore old. The presence of an intransitive aorist ἐάγη in Greek and of an intransitive verb in Tocharian wāk‑ ‘break (into pieces)’ both point to an old intransitive verb.47 Moreover, the intransitive semantics is confirmed by the absence of derived agent nouns with this root. Thus, ἄγνυμι (aor. ἔαξα) is a productively formed causative. Finally, consider φθίνω ‘perish, waste away’, root aor. ἔφθιτο, ptcpl. φθίμενος ‘dead’, perf. act. ἔφθιεν, plupf. med. ἔφθιτο. The perfect active is attested in Mycenaean as e‑qi‑ti‑wo‑e (TH Wu 75) and denotes swine that have apparently perished. Since the aorist ptcpl. φθίμενος is also resultative, García Ramón has discussed the ways in which it differs from the Mycenaean form e‑qi‑ti‑wo‑e; he argues that this active perfect form was superseded by φθίμενος, but that the perfect stem is continued in Homeric ἔφθιεν (Il. 18.446) and in some instances of the plupf. med. ἔφθιτο.48 The middle root aorist suggests an old intransitive verb (cf. LIV 151 with n. 2). The active paradigm consists of pres. φθίνω, aor. ἔφθεισα, fut. φθείσω; it is sometimes causative, sometimes (like the middle) intransitive.

 46 “While φθείρουσι can simply mean ‘ruin, destroy’, as it is usually taken […], in fact what bad herdsmen really do is allow their flocks to scatter or get lost” (Jamison 1993, 249). 47 Cf. LIV s.v. *u̯ eh2g‑ (“das gr. Perfekt spricht für fientive Grundbedeutung”). 48 García Ramón 1990, 13−15.

88  Lucien van Beek and Laura Migliori Other verbs of destruction (κτείνω, θείνω, etc.) do not have an active perfect in Homer. In Classical Greek, κτείνω has a perfect ἀπέκτονα that looks old from a morphological perspective; it is unclear whether the absence of this form in Homeric Greek is a coincidence, or whether it reflects the linguistic state of affairs.49 For θείνω ‘slay’, only a middle perfect is attested, and the active perfect to βάλλω does not take part in the causative-anticausative alternation.

. Verbs of Production The following three verbs are semantically related in that they are (in some of their meanings, at least) verbs of production: πήγνυμι ‘stick into’, ἀραρίσκω ‘adjust’ and τεύχω ‘produce, make’. In Homeric battle narrative, the verb πήγνυμι ‘stick into’ (aor. ἔπηξα; perf. πέπηγα ‘be stuck’) is normally transitive, indicating that an agentive subject causes a weapon to change position. In examples like παρὰ δ’ ἔγχεα μακρὰ πέπηγεν ‘and beside, long spears are stuck (in the ground)’ (Il. 3.135), the perfect indicates the resulting state corresponding to the transitive aorist πῆξε ‘he fixed, stuck’ (into the ground). This use is retained after Homer, e.g. [ξίφος] πέπηγεν ἐν γῇ πολεμίᾳ ‘[the sword] is fixed in enemy soil’ (Soph. Aj. 819). Thus, πέπηγα seems a convincing example of an active perfect denoting the resulting state of the Patient of a transitive verb. Things are, however, not as clear as they seem. Another meaning, frequent with the medio-passive forms (including aor. ἐπάγην), is ‘become solid, stiffen, freeze, congeal’, denoting natural processes: γοῦνα πήγνυται ‘the knees get stiff’ (Il. 22.453); ἅλες πήγνυνται ‘salt crystallizes’ (Hdt. 6.119.3, etc.). The perfect πέπηγε also aligns with this medio-passive use: πεπάγαισιν δ’ ὐδάτων ῤόαι ‘the streams are frozen’ (Alc. fr. 338.2 Voigt); ἄρθρα [...] πέπηγέ μου ‘my limbs are stiff’ (Eur. HF 1395); more examples in LSJ s.v. In this meaning, the active present and aorist forms are obviously secondary factitives.50 The main question now becomes how these two meanings are interrelated. Within Greek, the second meaning is broadly attested in lexicalized derivatives (e.g. πηγυλίς ‘ice-cold’; πηγός ‘solid, thick, firm’ (Hom.+); παγετός ‘frost’), and is clearly old. On the other hand, the highly specific correspondence between Lat.

 49 More on the active perfect of transitive predicates in Section 6.1.1. 50 In such cases the Causer is, again, often a divinity: cf. θεὸς […] πήγνυσι πᾶν ῥέεθρον ‘the god froze the entire stream’ (Aesch. Pers. 496); εἰ μὴ […] τοὺς ποταμοὺς ἔπηξε ‘if he [Zeus] had not frozen the rivers’ (Ar. Ach. 139).

Active versus Middle Perfect in Homeric Greek  89

pangō ‘insert, fix’ (cf. the instrument noun pālus ‘stick’ < pre-Latin *pag-slo-) and Hitt. pāsk- ‘plant, impale, insert a stick’ shows that the first, transitive meaning is old, too. Interestingly, the medio-passive and perfect stems are used to refer to missiles that get fixed into their target without a direct Causer: δόρυ δ’ ἐν κραδίῃ ἐπεπήγει ‘the spear was stuck in his heart’ (Il. 13.442), δοῦρα […] ἐν σάκεϊ […] πάγεν ‘the spears got fixed in the shield’ (Il. 11.571–572), among many other examples. Thus, this meaning ‘stick, insert’ may also have started off as an intransitive ‘get stuck’; if so, this must have happened already in the proto-language. Subsequently, within Greek prehistory the use of the active perfect πέπηγα (originally denoting the result of a spontaneous process) may have been extended to cases where an Agent/Causer was present (i.e. it could then also refer to resulting states of the transitive ‘stick, insert’).51 A second, less complicated case is ἀραρίσκω, aor. ἤραρον ‘adjust, fit out (with)’, perf. ptcpl. ἀρηρώς ‘fitted out, fitting’; plupf. ἀρήρει. This verb is semantically related to the previous: both πηγ- and ἀρ- are used to denote the technical production of vehicles by carpentry (cf. ἁρματοπηγός ‘wheelwright’).52 Synchronically, the verb usually indicates that a condition of being fit results from prior adjustment by a Causer. If this event structure were primary, the perfect ἀρηρώς would denote the resulting state of the object of a transitive event. There are sufficient reasons, however, to consider ἀραρίσκω and ἤραρον as original causatives, and the perfect ἀρηρώς (as well as the middle aor. ptcpl. ἄρμενος ‘adjusted, fitted out’) as original anticausatives. First, the indicative pluperfect ἀρήρει occurs in a formulaic phrase in which being fitting is not the result of adjustment: ἔγχος, ὅ οἱ παλάμηφιν ἀρήρει ‘the spear that fitted his hands’ (Il. 3.338; Od. 17.4); cf. δοῦρε, τά οἱ παλάμηφιν ἀρήρει (Il. 16.139). Secondly, the productive function of the reduplicated aorist is to create causatives (cf. Willi 2018, 69), and the present ἀραρίσκω is clearly secondarily based on aorist ἤραρον. It may therefore be assumed that ἀρ- was originally intransitive (with LIV s.v. *h2er-, 270 n. 1: “das Perfekt verlangt den fientiven Bedeutungsansatz”), forming a middle root aorist and a perfect. A final problematic case is τεύχω ‘produce, make’ (aor. ἔτευξα, perf. med. τέτυκται, perf. act. ptcpl. τετευχώς). The last-mentioned participle is a hapax,

 51 Another solution is chosen by LIV (461−462), which posits two originally different verbs, intransitive *peh2ǵ- ‘get solid, freeze’ vs. *peh2ḱ- ‘insert, fix’. Although this would allow to connect the Germanic verb *fōgjan- (Dutch voegen, Germ. fügen ‘fit together’) to the group of Lat. pangō, Hitt. pāsk-, it would leave unexplained the complete merger of the two roots in Gk. πήγνυμι. 52 The root πηγ- may also refer to pitching tents and building ships, types of construction in which wooden sticks or poles are inserted into a surface.

90  Lucien van Beek and Laura Migliori appearing only in the phrase βοὸς ῥινοῖο τετευχώς ‘made of cow’s hide’ (Od. 12.423). This form is usually compared to Myc. te-tu-ko-wo-a (of clothes, KN L 781.b; of wheels, PY Sa 682), allegedly meaning ‘finished’ or ‘well-made’. If these are old active participles to τεύχω, this would obviously present a considerable problem for our hypothesis,53 since τεύχω has a complex argument structure: an external volitional Causer brings about a change of state. It is possible, however, to resolve this problem by reconsidering the actual attestations and the etymology. Let us start with τετευχώς, which is usually considered to reflect an earlier *thethukhwōs (cf. Mycenaean te-tu-ko-wo-a), but with a secondary full-grade τευχintroduced after the disappearance of digamma.54 It is worrying that τετευχώς occurs only once, and not in a passage that makes an archaic impression. The normal form in Homer, moreover, is the middle perfect τέτυκται, τετυγμένος, which occurs no less than 86 times. Furthermore, the construction τετευχώς + gen. is different from that of the Mycenaean form te-tu-ko-wo-a (which is used absolutely), but directly matched in Homer by the middle perfect in e.g. αἳ δὲ βόες χρυσοῖο τετεύχατο κασσιτέρου τε ‘the shields were made of gold and tin’ (Il. 18.574; cf. also Od. 19.226). In fact, several Homeric perfects that only appear as participles use the active and middle voice without semantic difference (e.g. τετιηότι θυμῷ beside τετιημένος ἦτορ),55 and such cases may have served as a model for the creation of τετευχώς. It is therefore quite conceivable that τετευχώς is an artificial poetic creation.56 Another salient point is the etymology of τεύχω. It has long been noted that Homer also has a reduplicated aorist τετύκοντο ‘they prepared (for themselves)’ that is close in meaning. It is therefore possible that the root-final aspirate of τεύχω ‘produce’ is secondary, and that the Homeric verb was influenced by the root of τυγχάνω, ἔτυχον. A fitting Greek cognate for *τεύκω would be τύκος ‘hammer, axe’: by comparing OCS tъkati ‘weave’, aor. tъče ‘thrust’, it is possible to reconstruct a root PIE *teuk- ‘fashion’ (by thrusting or hammering? Cf. LIV 640 with note 1a; cf. also 149 with note 1). That τεύχω represents a conflation of two roots becomes clear from further Homeric evidence: the reduplicated present τιτύσκομαι, with its dual meanings ‘prepare’ and ‘take aim’, clearly belongs not

 53 As remarked above (n. 31), on the basis of te‑tu‑ko‑wo‑a scholars like Ruijgh and Slings have claimed that the active perfect in PIE allowed for passive readings. 54 See e.g. Ruijgh 2011, 286. 55 Generally, cf. Hackstein 1997−1998 on this type of alternation. 56 The direct model is not attested, but not difficult to imagine: the participial phrase might be an inflection/adaptation of an older *βοὸς ῥινοῖο τέτυκται.

Active versus Middle Perfect in Homeric Greek  91

only with the aorist τετύκοντο, but also with ἔτυχον in the sense ‘hit (a target)’. This would clear the way for us to consider Myc. te-tu-ko-wo-a as belonging only to the second etymon, τυγχάνω, ἔτυχον < PIE *dheugh- (cf. Gothic daug ‘is useful’, Skt. duhé ‘yields milk’). In both its attestations, te-tu-ko-wo-a qualifies a product, and ‘of high quality’ (tauglich, in German) would be a fitting translation. In fact, the Mycenaean form is reminiscent of Homeric phrases like τετυγμένος, οὐδὲν ἀεικής ‘an excellent mind, not at all unseemly’ (Od. 20.366), among many other examples: there, the participle does not qualify the object as the result of a production process (‘finished’), but denotes that it stands out qua quality and/or fulfils its function well. This may also help us understand the apparent ellipsis of *εὖ in νόος […] τετυγμένος and τυκτός (beside ἐύτυκτος ‘well-made’). We may conclude that the pairing of active perfects with verbs displaying a causative-anticausative alternation probably arose when causative eventive stems were secondarily formed, as Chantraine already argued. The main counterexamples, such as τετευχώς and πέπηγα, may be explained as secondary.

. Active Perfect beside Medium Tantum In this section we will briefly discuss the cases (9 in total) in which an active perfect attested in Homer corresponds to a deponent present.57 We will ask, with Romagno 2005, to which extent these pairings may have arisen secondarily. Beside ὄρομαι ‘oversee’, an active pluperfect ὀρώρει is attested once in ἐπὶ δ’ ἀνὴρ ἐσθλὸς ὀρώρει ‘a good man was supervising [the work]’ (Il. 23.112). The comparison with ἐπὶ δ’ ἀνέρες ἐσθλοὶ ὄροντο ‘noble men herd them [the flocks]’ (Od. 14.104) might show that ὀρώρει was created by artificial inflection of this verse-end (note that its formation presupposes psilosis, as the root is *ser‑). The middle diathesis of ὄρομαι (cf. already Myc. o‑ro‑me‑no) is expected in a verb of perception denoting a volitional activity, accompanied by an experiential subject.58 The same holds for δέρκομαι ‘look (at)’ beside δέδορκα ‘look’. The present δέρκομαι is usually thought to be a secondary creation of Greek, because in Sanskrit the paradigm of the cognate root darś ‘see’, which appears in the aorist and perfect, is supplied in the present by paśyati (PIE *speḱ‑). In our view, this argument is not strong: δέρκομαι might well be inherited because Greek preserves

 57 See Chantraine 1927, 26−37 for a discussion of the evidence, not all of which is of equal value. 58 Cf. Allan 2003, 95−101; Migliori 2016, 37.

92  Lucien van Beek and Laura Migliori what seems to be the original root meaning ‘look’, denoting an activity. In Sanskrit the aorist of darś ‘see’ has taken over the place occupied in PIE by *wid‑e/o‑ (Gk. εἶδον ‘saw’). If δέρκομαι continues an old present, its middle voice is again understandable, being a verb of perception with volitional subject. The active voice was retained in the perfect δέδορκα because the subject of a state is nonaffected by definition. Concerning προβέβουλα ‘I prefer’ (hapax, Il. 1.112), the present βούλομαι may also be used as a stative with the same meaning. Moreover, morphologically the perfect looks secondary: as a root formation, one would expect *προβέβολα, and to assume that προβέβουλα arose by metrical lengthening of such a form (as is sometimes done) seems ad hoc. There was also a metrical incentive to create the perfect: the compound verb προβούλομαι could only be used in the hexameter by admitting Attic correption, which is still a rare and mainly lexically determined license in Homer. In the case of the sound verbs μυκάομαι ‘bellow’ (hapax, Od. 10.413) and βρυχάομαι ‘roar’, the denominative formation of these presents betrays their recent origin compared to the perfects μέμυκε ‘groans’ and βέβρυχε ‘roars’;59 moreover, these presents are all but absent from Homer. For παρωίχωκε ‘is over’ (hapax, Il. 10.252; post-Homeric also simplex οἴχωκα) beside παροίχομαι ‘pass by’ (οἴχομαι ‘be gone’), it has been observed that the formation was probably influenced by that of μέμβλωκε ‘has come’, which occurs in another passage dealing with the passing of time, Od. 17.190−191.60 This could explain why an active (instead of middle) perfect is productively formed beside a deponent. The middle present ἱλάσκομαι/ἱλάομαι ‘propitiate’ has a perf. subj. ἱλήκῃσι, as well as an imperative ἵληθι ‘be merciful’. The middle is clearly self-beneficiary, but underlyingly, it seems to reflect an oppositional causative to the perfect stem. The pairing of middle and perfect is therefore probably secondary. The pairing of μείρομαι ‘obtain a share’ (imperat. μείρεο, Il. 9.616) with the perfect ἔμμορε ‘has a share’ is clear from the contexts: both occur in the same metrical slot, and both govern a genitive τιμῆς. Since μείρομαι cannot be productively derived from ἔμμορε (or vice versa), the pair must be archaic (pace Romagno 2005, 108).61

 59 See e.g. Willi 2018, 238 n. 80. 60 Wackernagel 1902, 739. 61 On the middle εἵμαρται, see Section 6.

Active versus Middle Perfect in Homeric Greek  93

The most problematic case is γίγνομαι ‘be born, become’, aor. ἐγένετο, perf. γέγονα ‘be born, be’. According to Romagno 2005, 107, this verb is not an old deponent because the active is attested in other IE languages (Lat. gignō ‘generate, beget’, Skt. jánati ‘id.’) and because an agent noun meaning ‘progenitor’ can be reconstructed for the proto-language on the basis of Gk. γενέτωρ, Lat. genitor, and Skt. janitar-. She therefore thinks that *ǵenh1- is an old causative. However, all other nominal derivatives of this verb in Greek have intransitive semantics; and lexically, ‘be born’ looks like a primary anticausative. It seems plausible that the formation in *-tor- was created together with the causative active already in PIE times, and that Greek subsequently lost this causative (‘sire, beget’ is expressed in Greek by φύω or τίκτω). In any case, the intransitive perfect is a remnant from an older stage, when the root itself was intransitive. To sum up, not all pairs of middle presents with active perfects are equally old, but μείρομαι beside ἔμμορε, γίγνομαι beside γέγονα, and δέρκομαι beside δέδορκα are probably relics of the PIE situation and resisted the tendency to create middle perfects beside deponents.

. Middle Perfect beside Medium Tantum Beside primary deponents, we normally find a middle perfect stem. At least the following 12 examples of this are attested in Homer:62 ἀφικνέομαι – ἀφῖγμαι ‘arrive’, δείκνυμαι – δείδεκτο, 3rd pl. δειδέχαται ‘they welcome’, δέχομαι ‘receive, accept’ – δέδεγμαι ‘expect, await’ (also unreduplicated δέγμαι), δράσσομαι ‘grasp with the hand’ – δεδραγμένος, καίνυμαι – κέκασμαι ‘excel’, κτάομαι ‘acquire’ – ἔκτημαι ‘possess’, χράομαι ‘use, enjoy’ – κέχρημαι ‘desire; lack’, λανθάνομαι – λέλασμαι ‘forget’, μιμνήσκομαι – μέμνημαι ‘remember’, πατέομαι – πεπάσμην ‘taste, take food’, πεύθομαι/πυνθάνομαι – πέπυσμαι ‘hear, learn’, πίλναμαι ‘approach’ – πεπλημένος ‘near’, ῥύομαι/ἔρυμαι – εἴρυμαι ‘protect’. Considering this list, we observe various semantic types of middles: mental activities, verbs of (body) motion, reciprocal middles, verbs of perception, etc. Semantically, there is no important difference between e.g. ἀφῖγμαι or ἔκτημαι and active perfects like εἰλήλουθα and ἔμμορε. It is therefore attractive to assume that such middle

 62 Romagno 2005, 43 also mentions ἀλάομαι, ποτάομαι, χαρίζομαι, but as stated earlier, we leave denominatives out of consideration. Another possible case is λιλαίομαι ‘desire, be anxious’ – λελιημένος ‘eager’, if the etymological connection of these stems is correct. The appurtenance of πέπνῡμαι ‘be wise/sensible’ to a present stem is uncertain, but it seems to belong with the mediopassive aorists ἔμπνῡτο, ἐμπνύνθη, ἐμπνύθη ‘regained his senses’.

94  Lucien van Beek and Laura Migliori perfects were adapted in voice to the other stems, in some cases even supplanting an older active perfect.

. Middle Perfect beside (Transitive) Telic Verbs A middle perfect also occurs with a number of transitive telic verbs (49 cases in Homer, not counting denominatives).63 Considering their semantics, it becomes immediately clear that (nearly) all cases belong to one class: they indicate a change of state of some sort. This fact signals that the distribution of active and middle forms is not accidental, but reflects precise semantic requirements. More specifically, these predicates are telic, agentive (externally caused) and often resultative (expressing the state of the object), when occurring in the middle perfect. The verbs in question are classified in the table below. Tab. 1: Change of state predicates Change of location or position

Change of condition

ἀγείρω – ἀγήγερται ‘gather’ ἀείρω – ἄωρτο ‘lift, raise up’ δαίομαι – δέδασται ‘divide, distribute’64 δίδωμι – δέδοται ‘give’ εἰλέω – ἐέλμεθα ‘shut in’ εἴργω – ἐέρχατο ‘fence in’ ἐλάω – ἐλήλαται ‘drive away’ ἐρείδω – ἠρήρειστο ‘press’

ἀρόω – ἀρήροται ‘plow’ βάλλω – βέβληται ‘throw; hit’ βλάπτω – βεβλαμμένος ‘disable’ δάμνημι – δέδμητο ‘subdue’ δέω – δέδεται ‘bind’ διδάσκω – δεδιδάχθαι ‘teach’ ἔδω – ἐδήδοται ‘eat’ εἴλυσα – εἴλυται ‘wrap in’

 63 It is noteworthy that various middle perfect participles are attested in Mycenaean. At least 10 Linear B perfect stems have an alphabetic correspondence: a‑pu‑ke‑ka‑u‑me‑na (κεκαυμένα, καίω), a‑ra‑ro‑mo‑te‑me‑na (cf. ἡρμοσμένα, ἁρμόττω), a‑ra‑ru‑ja/a‑ra‑ru‑wo‑a (Hom. ἀραρυῖα, ἀρηρώς), de‑de‑me‑na (Hom. δέδετο), de‑do‑me‑na ‘given’ (Hom. δέδοται), e‑pi‑de‑da‑to (Hom. δέδασται), e‑qi‑ti‑wo‑e (Hom. ἔφθιεν, φθίνω), e‑ra‑pe‑me‑na (ἐρραμμένα, ῥάπτω), me‑ta‑ke‑ku‑me‑na (Hom. κέχυται, μεταχέω), te‑tu‑ko‑wo‑a (Hom. τετυγμένα, τυγχάνω: see above). Without ascertained alphabetic correspondence are a‑ja‑me‑no ‘inlaid’, ]de‑di‑ku‑ja ‘?’, e‑re‑dwo‑e ‘?’ (cf. ἐρείδω ‘support’?), ke‑ke‑me‑na ‘communal [land]’ (?), ke‑ke‑tu‑wo‑e ‘?’, pe‑pu2‑te‑me‑no ‘?’ (cf. φυτεύω?), qe‑qi‑no‑to ‘is decorated’ (cf. Hom. ἀμφιδεδίνηται?). With the exception of e‑qi‑ti‑wo‑e and te‑tu‑ko‑wo‑a (which have been replaced by middle perfects), the distribution between active and middle perfects matches that in Homeric and Classical Greek. 64 This verb is a special case: the present is used both as agentive ‘divides’ (e.g. Od. 17.332) and passive ‘is divided’ (e.g. Od. 9.551). Since the middle has a self-beneficiary reading, the middle perfect is a normal passive to a transitive change-of-state predicate.

Active versus Middle Perfect in Homeric Greek  95

Change of location or position

Change of condition

ἐρύω – εἴρυται ‘drag, pull’ κλίνω – κέκλιται ‘lean’ κρίνω – κεκριμένος ‘select, choose’ κρύπτω – κεκρυμμένος ‘hide’ λέγω – προλελεγμένοι ‘select’ λείπω – λέλειπται ‘leave behind’ ὀρέγω – ὀρωρέχαται ‘stretch’ πορεῖν – πέπρωται ‘grant’ τείνω – τέταται ‘stretch out’ τρέπω – τέτραπτο ‘turn, direct’ χέω – κέχυται ‘pour out, strew’

εἴρω – ἔερτο ‘fasten together, string’ ἐρείπω – ἐρέριπτο ‘tear down, crumble’ ἐπιτέλλω – ἐπὶ […] τέταλται ‘command’ λούω – λελουμένος ‘bathe’ λύω – λέλυται ‘release’ μίσγω (μείγνυμι) – μέμικται ‘mix’ νίζω – νένιπται ‘wash’ ὀρίνω – ὀρώρεται ‘stir’ παλάσσω – πεπάλακτο ‘soil, spatter’ πείρω – πεπαρμένος ‘pierce’ πεφνεῖν – πέφαται ‘slay’ πίνω – πέποται ‘drink’ πίτνημι – πέπταται ‘spread out, open’ στόρνυμι – ἔστρωτο ‘smoothen, level’ τάμνω – τέτμηται ‘cut’ τίω – τετιμένος ‘value’ τύπτω – τετυμμένος ‘beat’ φύρω – πεφυρμένος ‘drench’

Verbs of creation65 βάζω – βέβακται ‘say’ εἴρω – εἴρηται ‘speak’ καλέω – κέκλημαι ‘call’ τεύχω – τέτυγμαι ‘make, produce’

Next to the verbs listed in the table, there are some cases in which a causative active present and/or aorist appears beside the middle perfect stem: ἀκαχίζω – ἀκάχημαι ‘grieve’ (psych-verb), ἐλέλιξε – ἐλέλικτο ‘shake’ (anticausative), παύω – πέπαυμαι ‘cease’ (anticausative), σεύω ‘chase’ – ἔσσυμαι ‘rush’ (body motion middle), φαίνω ‘show’ – πέφασμαι ‘appear’ (anticausative). As with the alternations discussed in Sections 5.1–3, these causative actives were secondarily formed.

 The Perfect as a Simple Predicate After having detected the verb types occurring with the perfect, it will now be possible to formulate a generalization concerning its formation. In order to do so, we will refer to the lexical semantics framework (cf. Levin/Rappaport-Hovav 2005), which puts the meaning of the lexicon at the centre of the faculty of language. The focus of this view is not on a specific language, but lexical items are assumed to have specific properties cross-linguistically. The main claim is that the meaning of a lexical item crucially determines its syntactic structure (in the  65 Three of these forms are verba dicendi; their middle perfect has passive meaning.

96  Lucien van Beek and Laura Migliori case of verbs, their participant/event structure). Discussing the cases under analysis from this perspective will allow us to find a link connecting them. More specifically, it will be claimed that their semantic similarities also explain their specific structural characteristics.

. The Simple vs. Complex Predicate Distinction and Its Relevance for the Homeric Perfect Adopting the lexical-semantics framework, we will consider verb classes as sets of semantically related verbs sharing a number of properties, such as possible realizations of arguments and their corresponding interpretation. The ontological (semantic) properties of a root make it possible to distinguish two main types of classes: manner and result verbs. Manner verbs specify the way of carrying out an action, whereas result verbs express the result of an event. As a consequence, they also exhibit divergences in their argument realization: result verbs are quite limited in their range of options, whereas manner verbs have various possibilities of argument selection. Consider, for instance, the English examples below, in which the difference between run (a prototypical manner verb) and go (a prototypical result verb) are shown (data from Rappaport-Hovav/Levin 1998, 98): (6) – Pat ran. (activity) – Pat ran to the beach. (directed motion) – Pat ran herself ragged. (change of state) – Pat ran her shoes to shreds. (change of state) – Pat ran clear of the falling rocks. (directed motion) – The coach ran the athletes around the track. (causation) (7) – The students went. – The students went to the beach. – *The jetsetters went themselves ragged. – *The runner went his shoes to shreds. – *The pedestrian went clear of the oncoming car. – *The coach went the athletes around the track.

At the syntactic level, manner verbs have a simple structure, since they consist of a single event. They are often atelic and internally caused. Conversely, result verbs are characterized by a more complex structure; they are at least formed by

Active versus Middle Perfect in Homeric Greek  97

two sub-events: a causing sub-event and its result.66 Both manner verbs and result verbs can be divided into a series of subclasses, according to the way in which the verb semantics determines the internal structure of the event (i.e. the number of participants and their relationship). As for the Homeric perfect, we have seen above that its main characteristic is to express a non-dynamic event, i.e. a state. Interestingly, this generalization also holds for state verbs. In Section 2, we have already seen Berrettoni’s interpretation of the difference between ἀλάομαι ‘roam’ and its perfect ἀλάλημαι. Let us now consider χολόoμαι ‘be angry’,67 a verbum sentiendi which displays some differences in usage, depending on aspect: (8)

[…] Ποσειδάων γαιήοχος ἀσκελὲς αἰὲν Κύκλωπος κεχόλωται, ὃν ὀφθαλμοῦ ἀλάωσεν. (Od. 1.68-69) Poseidon the Earth-Shaker is angry all the time about the Cyclops, whom he (Odysseus) blinded.

As guaranteed by the adverbial ἀσκελὲς αἰέν, the perfect κεχόλωται expresses a permanent state of anger of the subject, a so-called individual-level state.68 When occurring in the present, however, the same verb seems to refer to a more concrete situation, as exemplified below:69 (9) Ἥρῃ δ’ οὔ τι τόσον νεμεσίζομαι οὐδὲ χολοῦμαι. (Il. 8.407) I am not so much indignant with Hera, nor am I angry.

In this case, the verb expresses a so-called stage-level predicate, referring to a temporary property of the subject. The contrast between individual-level and stage-level predicates is determined by various characteristics, in particular by agentivity and dynamicity. Agentivity may play a role in the case of temporary states (in the sense of external causation, for instance), but not for permanent  66 See e.g. Dowty 1991; Levin/Rappaport-Hovav 2005; Ramchand 2008. 67 The verb ἀλάλημαι lacks the aorist form; therefore, another predicate has been chosen for exemplifying the present/aorist/perfect contrast in terms of dynamicity. 68 Cf. e.g. Carlson 1981 for the terminology. 69 Notice, however, that this is the only passage in Homer in which χολόoμαι occurs in the present; when indicating a state it normally occurs in the perfect.

98  Lucien van Beek and Laura Migliori states. As for dynamicity, stage-level predicates have the feature [dynamic], while individual-level predicates do not (Carlson 1981). These properties indicate that stage-level states are structurally more complex than individual-level states. This shows that present and perfect differ from each other in their internal structure, even when they both express a state.70 As for the aorist, when occurring with state predicates, it often indicates a change of state (cf. Napoli 2006, 162), as shown in (10) with an example of the same verb: (10) τοῦ δ’ Ὀδυσεὺς μάλα θυμὸν ἀποκταμένοιο χολώθη. (Il. 4.494) Odysseus grew very angry at heart about this man’s death.

Change of states are complex predicates, since they consist of at least two subevents: a causing event and a resulting event (cf. e.g. Ramchand 2008). From this perspective, the contrast between dynamic and non-dynamic events can be formulated also in terms of complex vs. simple predicates: the former can optionally be related to (external) causation, while the latter cannot. In other words, the distinction between dynamic and stative predicates is linked to the internal (event) structure of a predicate together with its interaction with aspect. As for the Homeric perfect, its non-dynamicity has to do with the fact that it always expresses a permanent state, i.e. a state involving a single argument.71 Typologically, pure states can be classified within the group of simple events and relate, therefore, to the category of manner verbs.72 In fact, if we look at active  70 An interplay between internal and external aspect can therefore be observed in that the semantic properties of the predicate interact with the aspectual specification for conveying different interpretations. The selection properties of predicates are also an example of this interaction: while a state verb like χολόoμαι occurs in the present/perfect/aorist, other verbs do not because their precise meaning is not compatible with some aspects. Suppletivism is a clear consequence of this fact. 71 Levin/Rappaport-Hovav 1998, 108. 72 In the literature, a separate class for state verbs has been detected as well: these are simple predicates expressing the state of the main argument (Levin/Rappaport-Hovav 1998, 108). However, it has been noticed that state verbs often get an eventive interpretation, especially when expressing a change of state (Levin/Rappaport-Hovav 1998, 126). Moreover, within states a distinction has also been made between one-place states and two-place states, depending on the presence or absence of external causation. For these reasons, states will just be considered under the main classification presented here: one-place states (like ‘to blossom’) will be seen as simple predicates; two-place states (change-of-state verbs, like ‘to break’) are considered as complex predicates.

Active versus Middle Perfect in Homeric Greek  99

perfect forms in Homer, they appear to mainly belong to this class. As for result verbs, perfect forms can be formed as well, but crucially only when no Causer is present in the event structure. This means that, even though it is possible to observe a number of predicate types forming an active perfect in Homer, they all share the property of being one-place predicates. The atelicity of manner verbs indicates the lack of a result element in their structure.73 Moreover, in most cases the situation of the subject is internally caused. Recall, for instance, the predicates expressing atelic spontaneous or mental processes detected in Section 5, like δαίομαι ‘radiate’ (perf. δέδηα); ταράσσομαι ‘be stirred’ (plupf. τετρήχει ‘was in upheaval’); ἔλπομαι ‘think; hope’ (perf. ἔολπα). All these cases clearly lack an external Causer in their structure and a result element is absent as well. Conversely, result verbs are telic, which indicates the presence of an extra element in the structure. In both classes, the (resulting) state regards the subject. Moreover, it is relevant that both classes include both transitive and intransitive verbs. It seems likely to us that the core class of Greek active perfects were simple predicates, and that it was originally not possible to express a state with the perfect in the case of a complex predicate (result verb). One indication of this is that isolated perfect forms (e.g. ἔοικα, or sound verbs like λέληκα) only belong to this group, which seems to show their original incompatibility with a different event structure. Another significant fact is the absence of an active perfect with the high-frequency verbs ἵημι, δίδωμι, and τίθημι.74 This apparently odd fact can be understood if we consider that ἵημι, δίδωμι, and τίθημι are prominent result verbs. These indications thus lead us to believe that active perfects were preferably formed to simple predicates. As will be shown in the next section, even when formed to a complex predicate, the perfect renders the event structure simple.

.. The Case of Transitive Perfects We have already encountered a number of perfects formed to transitive roots. A question arises, then, regarding the internal structure of these cases and the role of perfect morphology in two-place predicates. This problem has already been discussed by Romagno 2005, who argues that the assignment of thematic roles in the perfect depends on the degree of transitivity of the predicate. With activities

 73 Cf. Ramchand 2008. 74 Cf. already Wackernagel 1904, 3−4.

100  Lucien van Beek and Laura Migliori such as κεύθω, the subject of the perfect usually corresponds to that of the eventive stem, but since the perfect denotes a state, the subject assumes an inactive role.75 With resultative or transformative predicates such as πήγνυμι/πέπηγε in the sense of ‘fix, stick into’, where the perfect may indicate a resulting state, it selects a subject that corresponds to the Object/Patient of the corresponding eventive stem. Romagno explains this by noting that Subjects (Agents) of hightransitive events are incompatible with an inactive role, and therefore unable to appear as the subject of the perfect.76 In our view, Romagno’s analysis presents a number of problems. Firstly, it is problematic to assume that a single derivational formation combines two radically different functions: turning a dynamic event into a state, and changing the thematic role assigned to the Subject from active to inactive. Secondly, to reiterate the issue raised above, it remains difficult to understand why high-frequency transitive verbs like τίθημι, δίδωμι, and ἵημι have no old active perfect. It is true that the middle perfect of τίθημι could be supplied by κεῖμαι ‘lie’ and that δίδωμι has a middle perfect attested already in Mycenaean (de‑do‑me‑na). However, ἵημι has no perfect at all until the 4th century BC, and one would surely expect an old active perfect of δίδωμι to be retained. Thirdly, Romagno admits (2005, 118−119) that the syntactic configuration of a perfect ultimately depends on the concrete nature of the predicate. For instance, within her scenario it remains difficult to substantiate why exactly πήγνυμι would behave differently from λείπω: both verbs are resultative predicates, but only λείπω forms a middle perfect (λέλειμμαι ‘be left, remain’) indicating the resulting state of the Patient. One could argue that λέλειμμαι is a later creation replacing an older intransitive use of λέλοιπα, but as we have seen in Section 5, there are no other secure cases of intransitive active perfects with passive semantics; in other words, the complement of an active perfect cannot be the Patient of a change of state.

 75 “[L]a differenza fra il perfetto e il presente consiste nel diverso ruolo tematico assegnato al soggetto. Il perfetto configura il soggetto come sede di una qualità e, perciò, […] sposta il ruolo tematico del soggetto verso il macroruolo dell’inattività. Il presente configura il soggetto come autore o iniziatore di un evento e colloca il suo ruolo tematico nel macroruolo dell’attività” (Romagno 2005, 115). 76 “Insomma, coi verbi atelici di attività, il perfetto converte un soggetto da attivo in inattivo, rappresentando l’attività come una proprietà metacronica, e perciò come uno stato […]. Coi verbi telici, risultativi o trasformativi, invece, il perfetto converte in soggetto inattivo l’oggetto inattivo della costruzione transitiva” (Romagno 2005, 115−116).

Active versus Middle Perfect in Homeric Greek  101

.. Transitive Perfects: Antipassive Predicates? Adopting a different perspective, we would now like to draw attention to a number of similarities between transitive perfects and antipassive constructions. Antipassives are found in many natural languages and have been studied both from a typological and from a formal perspective. In the literature, they are defined as a clause (or the predicate therein) in which the logical object of a transitive (two-place) predicate is demoted to a non-core argument or a non-argument. (Polinsky 2017, 310)

In other words, while in a passive structure detransitivization takes place via Agent demotion, in the case of an antipassive clause, it is the direct object which is demoted. Morphologically, antipassives vary cross-linguistically: they can be realized, for instance, by means of noun incorporation or by dedicated morphology. In Homeric Greek, the active perfect of transitive roots can be compared with this type of structure. As a general observation, it can be said that in the active perfect the argument structure of the verb is retained. Being non-dynamic, the active perfect stem could not express a change of state of the object (in which case the present or aorist stem had to be used), but it could not be used either as resultative, as in the English present perfect I’ve painted the wall green.77 Another relevant observation concerns the characteristics of the objects of transitive perfects: these have a low degree of affectedness.78 Consider, for instance, the use of λέλοιπα in Homer. This perfect remains transitive on one of its three occurrences, but in this case the object is not affected by the verbal event: (11) ναὶ μὰ τόδε σκῆπτρον, τὸ μὲν οὔ ποτε φύλλα καὶ ὄζους φύσει, ἐπεὶ δὴ πρῶτα τομὴν ἐν ὄρεσσι λέλοιπεν, οὐδ’ ἀναθηλήσει· περὶ γάρ ῥά ἑ χαλκὸς ἔλεψε φύλλά τε καὶ φλοιόν· (Il. 1.234−237)

 77 This was, of course, stressed already by Wackernagel 1904 and Chantraine 1927, although the latter admitted βεβίηκεν as a singular counterexample. We agree with Willi 2018, 229 that Berrettoni 1972, 140−141 was right in viewing βεβίηκεν as a comprehensive perfect, summarizing a series of descriptive medio-passive perfects. 78 Cf. Chantraine 1927, 6 for the observation that definite direct objects are excluded from the perfect stem and instead require an aorist or present.

102  Lucien van Beek and Laura Migliori (I swear) by this sceptre, which will not grow leaves and branches ever since it left its trunk in the mountains, nor will it blossom; for all around it the bronze has peeled off leaves and bark.

In this example, the object is not mentioned as the Undergoer of the expressed event, but merely as a location, the place of origin which the sceptre has left.79 This meaning is also found at Od. 14.134 (ψυχὴ δὲ λέλοιπεν ‘the soul has left [the body])’ and Od. 14.213 (νῦν δ’ ἤδη πάντα λέλοιπεν ‘now all [virtues] have already left [me]’, Odysseus speaking as the Cretan). These passages unequivocally illustrate that the object is syntactically and semantically demoted, so that it can even be omitted. This means that the following properties are present: object indefiniteness; non-affectedness of the object; the non-argumental character of this element. The first two aspects are particularly significant in relation to the notion of transitivity, since they both indicate a very low grade (or absence) of transitivity in the clause (cf. Hopper/Thompson 1980, 252). The optional character of the object signals its behaviour as a non-core argument and, confirming the nontransitive character of the clause, clearly indicates the demotion of this element, which behaves more like an oblique complement. Viewed from this perspective, λέλοιπε shows some interesting similarities with antipassive constructions: even though the argument structure of the predicate is kept intact, the active perfect corresponds to a clause in which the object is partially or completely neutralized, thus signalling a lower grade of agentivity. The middle perfect λέλειμμαι, on the other hand, expresses the resulting state of the Patient of λείπω. Interestingly, the functional and formal distinction between λέλοιπα and λέλειμμαι exactly matches that between the Ved. Skt. perfects riréca ‘has left’ (active) and riricé ‘is left’ (middle). This raises the question whether λέλειμμαι in passive function is really a late replacement of an older use of λέλοιπα. It is equally possible, as we argue, that this passive use of the middle perfect was at the origin of the category. In this case, active and middle perfect forms would reflect two different means of deagentivization: object demotion in the first case, subject demotion in the second. Indeed, cross-linguistically it has been shown that the presence of antipassives in a language does not preclude that of passives, and vice versa;80 in this sense, the Homeric Greek perfect would not be surprising.

 79 “Das Perfekt steht nicht, weil das Objekt im Moment des Sprechens noch die Wirkung der Handlung verspürte, sondern es kommt bloss auf den Zustand des Subjekts an, sobald es über das Schneiden hinaus ist” (Wackernagel 1904, 5). 80 Cf. Polinsky 2017, 329.

Active versus Middle Perfect in Homeric Greek  103

In addition to λέλοιπα, a number of other transitive perfects in Homer display a non-affected object. Firstly, ἔοργα always occurs with an indefinite object (e.g. κακά, πολλά), as already exemplified in (3) above (ἐσθλὰ ἔοργε, Il. 2.272). Interestingly, the verb ἔρδω/ῥέζω hardly occurs in the passive, which indicates the absence of prototypical transitivity. In the same way, κέχονδα/κέχανδα and κέκευθα (already inagentive in their semantics) are accompanied by generic objects in the neuter plural (θάλαμον […] ὃς γλήνεα πολλὰ κεχάνδει ‘the bedroom which contained many valuables’, Il. 24.191−192; ὅσα τε πτόλις ἧδε κέκευθε ‘as much [treasure] as this city conceals’, Il. 22.118). Verbs like οἶδα and ὄπωπα often occur with an indefinite object, an indefinite temporal adverb or an embedded clause. All these cases show that, despite the transitive character of the root, their occurrence in the perfect displays a low degree of transitivity and a demoted (or eliminated) object.81 Finally, some words have to be dedicated to the verbs of consumption ἔδω and βιβρώσκω. These predicates can be grouped apart because they display a special behaviour: they can either be construed as intransitives, expressing an activity, or transitively, indicating an accomplishment.82 In Homeric Greek, their perfect stem only occurs in the participle. In the Iliad, βεβρωκώς occurs once with an accusative object, but here too (as with other transitive predicates), the neuter plural signals indefiniteness: βεβρωκὼς κακὰ φάρμακα ‘having devoured poisonous herbs’ (Il. 22.44). This participle occurs in a simile describing a lion; in this case the verb is clearly intransitive, because it governs a partitive genitive: (12) αἵματι καὶ λύθρῳ πεπαλαγμένον ὥς τε λέοντα, ὅς ῥά τε βεβρωκὼς βοὸς ἔρχεται ἀγραύλοιο. (Od. 22.402–403) Soiled with blood and gore, like a lion that comes from feeding on a cow on a farmstead.

As for ἔδω, the use of the perfect ptcpl. ἐδηδώς in the following comparison (in tmesis with the preverb κατά) makes sense because the object is an indefinite, generic bull whose further characteristics are irrelevant:  81 As a side remark, it would not be prudent to extrapolate general conclusions concerning the admissibility of certain perfects from a limited corpus like the Homeric one. As for prototypically transitive verbs, for instance, most verbs of killing have no active perfect in Homer. Nonetheless, an example like Attic ἀπέκτονα stands a good chance of being old in view of its o-grade root. It may have been regularly used in phrases of the type Have you ever killed in battle?. If so, the perfect could demote the object even in the case of prototypical agentive predicates. 82 For the internal structure of verbs of consumption, see Folli/Harley 2005.

104  Lucien van Beek and Laura Migliori (13) ἂν δ’ αὐτὸς ἔβαινε πόδας καὶ χεῖρας ὕπερθεν αἱματόεις ὥς τίς τε λέων κατὰ ταῦρον ἐδηδώς. (Il. 17.541–542) He himself (Hector) mounted the chariot, his feet and hands on top all bloody, like a lion that has fed himself with a bull.

Therefore, also with verbs of consumption, the perfect takes a non-affected object, or it has no direct complement. In conclusion, active perfect forms of transitives generally occur with a demoted object (when present). Although it would be incorrect to treat perfect morphology as antipassive marking tout court,83 the analogy between the behaviour of active perfect forms of transitive verbs and antipassives is remarkable.84

 Conclusion In this paper, we have analyzed the relationship between active and middle perfect forms in Homeric Greek. With Chantraine, we have stressed the secondary nature of active present and/or aorist forms with most anticausative verbs. However, we have also shown that Chantraine’s emphasis on the perfect as a part of the middle voice was a mistake. The possibility to create an active perfect in PIE depended not on the diathesis of a given present/aorist stem, but on predicate types. The perfect generally presents an event as non-dynamic. With intransitive result verbs, this meant presenting the event as a property of the subject; with transitive result verbs, however, there was no way to express the resulting state of the object. This possibility could be realized, however, in the middle voice after this had acquired passive meanings. As soon as that happened, it was possible to create medio-passive perfects beside middle presents. This means that verbs with  83 As shown above, the main function of the perfect was to denote a stative, therefore it would not be correct to make a generalization about perfect morphology in the terms of antipassives. However, it has been noticed in the literature that, cross-linguistically, antipassives often occur in the case of statives (cf. Hopper/Thompson 1980, 268; Willi 2018, 539). This fact establishes a relevant link between the main function of Homeric perfect forms and the cases displaying antipassive characteristics, supporting our hypothesis that the underlying structure of these apparently ‘transitive’ forms is not transitive at all. 84 The presence of antipassive forms in Ancient Greek has also been noticed by Veksina 2017; this may substantiate our hypothesis that such constructions were present in this language. For a thorough discussion of this topic and its PIE correlates, see Willi 2018, 533−534.

Active versus Middle Perfect in Homeric Greek  105

a middle perfect are characterized by an event structure containing a result and an external Causer. As for the manner vs. result distinction, it can be observed that middle perfect forms occur in the case of transitive result verbs and that they express the state of the object. After it had come into being in this way, the middle perfect could usurp part of the roles of the active perfect, e.g. with deponent verbs.

Bibliography Allan, R.J. (2003), The Middle Voice in Ancient Greek: A Study in Polysemy, Amsterdam. Allan, R.J. (2016), “Tense and Aspect in Classical Greek. Two Historical Developments: Augment and Perfect”, in: S.E. Runge/C.J. Fresch (eds.), The Greek Verb Revisited: A Fresh Approach to Biblical Exegesis, Bellingham, 81−121. Berrettoni, P. (1972), “L’uso del perfetto nel greco omerico”, in: SSL 12, 25−170. Borer, H. (2005), Structuring Sense II: The Normal Course of Events, Oxford. Carlson, L. (1981), “Aspect and Quantification”, in: P. Tedeschi/A. Zaenen (eds.), Tense and Aspect, New York, 31−64. Chantraine, P. (1927), Histoire du parfait grec, Paris. Chantraine, P. (1953), Grammaire homérique, vol. 2: Syntaxe, Paris. Chantraine, P. (1958), Grammaire homérique, vol. 1: Phonétique et morphologie, 2nd edn., Paris. Clackson, J. (2007), Indo-European Linguistics, Cambridge. Comrie, B. (1976), Aspect, Cambridge. Daues, A. (2006), “On the Function of the Mediopassive Perfect in the Homeric Epics”, in: E. Crespo/J. De la Villa/A.R. Revuelta (eds.), Word Classes and Related Topics in Ancient Greek. Proceedings of the Conference on Greek Syntax and Word Classes held in Madrid on 18−21 June 2003, Louvain-La-Neuve, 257−272. Debrunner, A. (1928), “Review of P. Chantraine, Histoire du parfait grec (Paris 1927)”, in: IF 46, 287−290. Dowty, D.R. (1991), “Thematic Proto-Roles and Argument Selection”, in: Language 67, 547−619. Folli, R./Harley, H. (2005), “Consuming Results: Flavors of Little-v”, in: P. Kempchimsky/ R. Slabakova (eds.), Aspectual Enquiries, Dordrecht, 1–25. Fortson, B. (2010), Indo-European Language and Culture: An Introduction, Chichester/Malden, MA. García Ramón, J.L. (1990), “Mykenisch e-qi-ti-wo-e /(h)ekwhthiwohe(s)/ ‘umgekommen, tot’, homerisch ἔφθιται, ἔφθιεν, (˚)ἔφθιτο und das Perfekt von idg. *dhgwhei‑ im Griechischen”, in: MSS 51, 7–20. Hackstein, O. (1997−1998), “Sprachgeschichte und Kunstsprache: Der Perfekttyp βεβαρηότες im frühgriechischen Hexameter (und bei späteren Daktylikern)”, in: Glotta 74, 21−53. Haug, D.T.T. (2008), “From Resultatives to Anteriors in Ancient Greek: On the Role of Paradigmaticity in Semantic Change”, in: Þ. Eyþorsson (ed.), Grammatical Change and Linguistic Theory: The Rosendal Papers, Amsterdam, 285−305.

106  Lucien van Beek and Laura Migliori Hopper, P.J./Thompson, S.A. (1980), “Transitivity in Grammar and Discourse”, in: Language 56, 251−299. Jamison, S. (1993), “Thornless Paths and Others: Vedic anr̥ kṣara‑ / Greek φθείρω”, in: G. Meiser (ed.), Indogermanica et Italica: Festschrift für Helmut Rix zum 65. Geburtstag, Innsbruck, 237−251. Kloekhorst, A. (2008), Etymological Dictionary of the Hittite Inherited Lexicon, Leiden. Kratzer, A. (1996), “Severing the External Argument from Its Verb”, in: J. Rooryck/L. Zaring (eds.), Phrase Structure and the Lexicon, Dordrecht, 109−137. Kümmel, M. (2000), Das Perfekt im Indoiranischen, Wiesbaden. Kuryłowicz, J. (1932), “Les désinences moyennes de l’indo-européen et du hittite”, in: BSL 33, 1−4. Levin, B./Rappaport-Hovav, M. (2005), Argument Realization, Cambridge. Magni, E. (2017), “Pluractionality and Perfect in Homeric Greek”, in: F. Logozzo/P. Poccetti (eds.), Ancient Greek Linguistics: New Approaches, Insights, Perspectives, Berlin/Boston, 325–344. Migliori, L. (2016), Argument Structure, Alignment and Auxiliaries between Latin and Romance: A Diachronic Syntactic Account, Utrecht. Napoli, M. (2006), Aspect and Actionality in Homeric Greek: A Contrastive Analysis, Milano. Perlmutter, D.M. (1978), “Impersonal Passives and the Unaccusative Hypothesis”, in: Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society, Berkeley, 157–189. Polinsky, M. (2017), “Antipassive”, in: J. Coon/D. Massam/L.D. Travis (eds.), The Oxford Handbook of Ergativity, Oxford, 308–331. Ramchand, G. (2008), Verb Meaning and the Lexicon: A First-Phase Syntax, Cambridge. Rappaport Hovav, M./Levin, B. (1998), “Building Verb Meanings”, in: M. Butt/W. Geuder (eds.), The Projection of Arguments, Stanford, CA, 97–134. Reinhart, T. (2002), “The Theta-System: An Overview”, in: Theoretical Linguistics 28, 229–290. Rix, H. (1992), Historische Grammatik des Griechischen: Laut- und Formenlehre, 2nd edn., Darmstadt. Romagno, D. (2005), Il perfetto omerico: diatesi, azionalità e ruoli tematici, Milano. Ruijgh, C.J. (1991), “Les valeurs temporelles des formes verbales en grec ancien”, in: J. Gvozdanović/T.A.J.M. Janssen (eds.), The Function of Tense in Texts, Amsterdam, 197–217. Ruijgh, C.J. (2004), “Over de gebruikswijzen van het Griekse perfectum”, in: Lampas 37, 24–45. Ruijgh, C.J. (2011), “Mycenaean and Homeric Language”, in: Y. Duhoux/A. Morpurgo Davies (eds.), A Companion to Linear B, vol. 2, Louvain-la-Neuve, 253–298. Schwyzer, E./Debrunner, A. (1950), Griechische Grammatik, vol. 2: Syntax und syntaktische Stilistik, München. Slings, S.R. (1987), “Het perfectum van Griekse toestandswerkwoorden”, in: S.R. Slings/I. Sluiter (eds.), Ophelos: Zes studies voor D.M. Schenkeveld, Amsterdam, 61–76. Stang, C.S. (1932), “Perfektum und Medium”, in: NTS 6, 29–39. Veksina, M. (2017), “Coan ἀποβλῆι and the History of Greek ἔβλην Again”, in: Glotta 93, 213– 231. Wackernagel, J. (1902), “Über Bedeutungsverschiebung in der Verbalkomposition”, in: Nachrichten von der Gesellschaft der Wissenschaften in Göttingen 1902, 737–757. Wackernagel, J. (1904), Studien zum griechischen Perfektum, Göttingen. Watkins, C. (1969), Indogermanische Grammatik, vol. 3.1: Formenlehre: Geschichte der indogermanischen Verbalflexion, Heidelberg. Willi, A. (2018), Origins of the Greek Verb, Cambridge.

Enzo Passa

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry in the Classical Age Abstract: In this paper I shall deal with Doric features in hexameter poetry, a kind of poetry that on account of its historical development has often been considered hardly affected by Doric dialect features before the Hellenistic period. Rather than reappraising the much-debated traces of Doric in the Homeric and Hesiodic epics, I will concentrate upon hexameter poetry from the 5th century BC. My starting point is a well-known passage of Empedocles. I shall then turn to a fragment attributed to Panyassis by some scholars. Finally, I will offer some remarks on the role played by Doric in the hexameter texts of the Classical period, and draw the conclusion that it is likely to have been more conspicuous than is usually assumed. Keywords: Hexameter poetry, Doric dialect, Empedocles, Panyassis.

 Doric in an Empedocles’ Formula In a famous passage, Empedocles summarizes two crucial moments of his cosmic cycle, that of the convergence of all things into the perfect unity of the Sphairos, and that of the consequent separation of the Sphairos giving birth to the manifold reality of the world as we know it: (1)

δίπλ’ ἐρέω· τοτὲ μὲν γὰρ ἓν ηὐξήθη μόνον εἶναι ἐκ πλεόνων, τοτὲ δ’ αὖ διέφυ πλέον’ ἐξ ἑνὸς εἶναι. (B 17.1–2 = 16–17 D.-K.) A twofold tale I shall tell: now it grew to be only one from many, now it separated to be many from one.

It seems reasonable to consider the two lines as a formula, whose function is to stress a basic point of Empedocles’ cosmological view.1 Within Empedocles’ extant  1 Formulas are found in both Parmenides’ and Empedocles’ poems on nature. As for Empedocles, what we can observe in his surviving fragments clearly suggests that his poetry was deeply https://doi.org/10.1515/9783110621747-006

108  Enzo Passa fragments these lines only appear in B 17 D.-K., but nothing prevents us from supposing that they may have occurred elsewhere in the Empedoclean Περὶ φύσεως. Let us now have a closer look at the language featured in the text. Although there seems to be little doubt that it has been affected by Homer,2 it would not be advisable to simply define it as ‘Homeric’. To start with, μόνος has a non-traditional prosody already attested once in Homer (μονωθείς, Il. 11.470), but indeed very rare in the archaic epics, where μοῦνος (< μόνϝ‑) is usually to be expected. The introduction of such forms into the Homeric poems was due, according to modern scholarship, to Attic or Euboean influence.3 Nevertheless, it is worth noting that forms lacking the 3rd compensatory lengthening do increasingly appear also in choral lyric poetry,4 which was composed in a kind of literary Doric. We may suppose that at an earlier stage these lyric forms were influenced by the Homeric epics, but I would frankly feel rather embarrassed if I were asked to take sides on the question of the dialect to which μόνος should be ascribed in Empedocles’ poetry. In other words, it seems likely to me that in Western Greek literature the boundaries between poetic genres and their language soon became quite loose, so that forms such as μόνος were no longer evaluated on account of their dialectal ancestry, but were rather regarded as poetically suitable on the basis of the previous tradition as a whole. In fact, for both Doric-speaking Empedocles  formulaic. See e.g. the formulaic line B 35.5 D.-K., which is likely to appear in the same form in P.Strasb. a(ii).19; according to Martin/Primavesi 1999, it also occurs in P.Strasb. a(ii).17 with some remarkable changes. Moreover, formularity is announced by Empedocles himself in B 25 D.-K.: καὶ δὶς γάρ, ὃ δεῖ, καλόν ἐστιν ἐνισπεῖν. 2 The εἶναι at the end of both lines 1–2 and 16–17 parallels a recurring Homeric line-end, and equally Homeric is the ἐκ at the beginning of lines 2 and 17. Furthermore, τοτὲ μέν in lines 1 and 16 has the same metrical position as in Hom. Il. 9.19, whereas διέφυ — which is never found in Homer and makes its first appearance in this fragment of Empedocles — has the same metrical position as ἔφυν in Hom. Od. 10.397. On ἐρέω, see below. 3 The Attic hypothesis has widely spread amongst scholars since the seminal study by Wackernagel 1916 (on μονωθείς, see pp. 122–134). Euboean, for its part, has been brought into play in more recent studies, and particularly by West 1988, who has put forward the idea that this dialect played an important role in the last compositional stages of the Homeric text. According to West 1988, 166 n. 93, the lack of the 3rd compensatory lengthening in forms such as μονωθείς should be viewed, alongside other linguistic features, as hinting at an ancient Euboean phase of the Homeric epics. Although West’s overall linguistic arguments may be dismissed as unreliable, some isolated Euboean forms (including those lacking the 3rd compensatory lengthening) are likely to have made their way into the Homeric text at a later stage. A valuable discussion of these issues may be found in Cassio 1998, 15–18. 4 While absent in Alcman, a form such as μόνας appears in a metrically guaranteed position in Stesichorus (fr. 223.2 PMG); later on, these kinds of forms become usual in Pindar. See Nöthiger 1971, 26–32; Hinge 2006, 50.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  109

and his Ionic-speaking forerunner Parmenides μοῦνος and μόνος seem to have been nothing else but poetically adequate and metrically useful alternatives.5 The first form came from the prestigious Ionic Homeric tradition, whereas the second, though acceptable in a hexameter context, may have arisen in the Western Greek world thanks to the works of lyric poets born or active in that area. Be that as it may, it was not problematic for either Parmenides or Empedocles to use both forms. There is, however, another word in our fragment which risks going unnoticed at a first glance. This is διπλά, which features at the very beginning of the sentence. Although the incipit δίπλ’ ἐρέω reproduces a Homeric prosody,6 διπλός ‘twofold’ never occurs in the Homeric epics in that form. Homer in fact uses either διπλόος (as the epithet of θώρηξ in Il. 4.133 = 20.415) or its contracted counterpart διπλοῦς (χλαῖναν […] διπλῆν in Il. 10.133–134 and Od. 19.225–226). On the other hand, neither ἁπλόος ‘single’ nor the contracted ἁπλοῦς are ever found in Homer, and only ἁπλοΐς is attested in two surely not old passages (ἁπλοίδας χλαίνας in Il. 24.230 = Od. 24.276). If we stick to the reading unanimously transmitted by manuscripts, there is no way for our poet to have used διπλᾶ (< διπλόα) instead of διπλά, since the elision would have been impossible. Nor does it seem reasonable to suspect that our text was somehow marred during its transmission, because in this case we should expect the corruption of διπλός into διπλοῦς, which was the common Attic form, rather than the opposite. As far as I know, only Gallavotti (1975, 183) has argued against the Empedoclean διπλά, suggesting putting it aside as a “forma secondaria e artificiosa” and restoring in its place either the contracted διπλᾶ, which one would expect in Homer, or the elided διπλό(α). As a consequence, in either case the following ἐρέω should be read with synizesis, since otherwise — as observed — the hexameter would be irregular. This is, however, highly unlikely, and Gallavotti’s whole reasoning seems to be not much different from the kind of reasoning we would expect from an ancient editor or copyist who, for whatever reason, decided at some point to introduce a hyper-Ionic form into the text.7 For, by adopting Gallavotti’s suggestion, not only are we unnecessarily forced to correct the transmit-

 5 Both μοῦνος and μόνος are metrically guaranteed in Parmen. B 2.2, B7.6 = B 8.1, B 8.4 D.-K. (dubious), and in Emped. (except B 17) B 2.5, B 8.3, B 111.2, B 134.4 D.-K., P.Strasb. a(ii).21. On B 35.5 D.-K. and P.Strasb. a(ii).17 and 19, see n. 1. 6 Od. 12.156 ἀλλ’ ἐρέω, where the metrical position of ἐρέω is very uncommon for Homer. In most Homeric lines the verb occurs immediately before the penthemimeral caesura. 7 For a hyper-Ionic form occurring in Emped. B 1 D.-K., see Passa 2004, 26.

110  Enzo Passa ted text, but in doing so we also need to accept a disyllabic ἐρέ͜ω,which is unparalleled in both Homer and Empedocles.8 In short, we may be quite sure that διπλά is the correct reading. I am not going to address in detail the dispute which arose over a hundred years ago about which one of διπλόος and διπλός (or, in parallel, of ἁπλόος and ἁπλός) we should envisage as the original form. However, it needs to be emphasized that this dispute has never truly been resolved, and that the prevailing opinion, according to which ‑πλοο‑ is ancient and ‑πλο‑ is recent, continues to be founded on very weak bases.9 To give just one example, at least two objections can be raised to an authoritative scholar such as Chantraine (DELG s.v. ἁπλόος), who has argued that “il est […] difficile de poser comme forme originaire διπλός et ἁπλός tardivement attestés” (my emphasis). The first objection is that no conclusive arguments may be put forward in opposition to the hypothesis that both ἁπλός (< *sm̥plos; cf. simplus in Plaut. Poen. 1362) and διπλός (< *dwiplos) are inherited forms.10 And secondly, even if we accept that neither διπλός nor ἁπλός ever occur in the Homeric poems, it would not be wise to hastily dismiss the Empedoclean διπλά simply because it is late. Actually, if we really wish to understand more about the Empedoclean διπλός, we need to stop investigating Homer and look elsewhere, to Doric inscriptions, where the number of forms connected with διπλός is quite impressive. There are examples coming from Eastern Doric areas such as Crete (in the Gortyn Code, which dates from the 5th century BC, there are not only several cases of

 8 In all the twenty-four Homeric cases of ἐρέω, the verb is always trisyllabic. The same goes for Empedocles: beside B 17, see B 8.1 D.-K. ἄλλο δέ τοι ἐρέω, which is directly imported from Homer, and the variatio in B 114.3 D.-K. οὓς ἐγὼ ἐξερέω. 9 The question was raised by some German linguists between the end of the 19th and the first decades of 20th century (Hoffmann 1898, 299–300; Brugmann 1917–1920, 128–135; Kretschmer 1923, 218; Schwyzer 1939, 598) and remained unsolved in the following years, as is clear from the opposite views expressed by GEW and DELG s.v. ἁπλόος (see below); still inconclusive is EDG s.v. ἁπλόος. Beside the etymological problems, doubts also arise about the tradition of some texts, and are in turn closely connected with accentual issues. For instance, in Od. 19.226 manuscripts oscillate between διπλῆν — which is the contracted form for διπλόην and is usually printed in modern editions — and διπλήν, which comes from διπλός and has been retained by some scholars as the true Homeric form. The situation is even more unclear in Herodotus’ manuscripts, where different accents are often transmitted for the same word (e.g. Hdt. 5.90.1 Hude διπλῆν v.l. διπλήν, Hdt. 8.87.4 Hude διπλᾶ v.l. διπλά, etc.), so that Hoffmann considered many of the forms with circumflex accent as inauthentic: he was followed by Hude 1908, who mostly printed forms with acute accent in his Herodotus’ edition. 10 So Brugmann 1917, 128; GEW s.v. ἁπλόος; Leumann 1977, § 202.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  111

forms connected with διπλός, but also ἀπλός is repeatedly found beside ἀπλόος),11 from Western Doric colonies in South Italy (forms in all likelihood derived from διπλός appear in the tablets of both Heraclea and Locri, whose dates fall in the 4th century BC),12 and possibly even from Sicily.13 Once this evidence is taken into account, it is difficult to avoid the conclusion that Empedocles, who hailed from the Cretan-Rhodian colony of Akragas, used the Doric διπλός in B 17 D.-K.: no other form would have suited the metre in a better way. Moreover, it is worth pointing out that the neuter is used to introduce a new theme in both B 17 — where we find δίπλ’ ἐρέω ‘a twofold tale I am going to tell’ — and in B 8.1 D.-K., where we read ἄλλο δέ τοι ἐρέω ‘another tale I shall tell you’.14 While the latter reflects a Homeric formula (see n. 8 above), with the former Empedocles creates a new formula which, as far as we can see, is unknown in the traditional epic language. We are entitled to ask ourselves how this shift towards Doric dialect is to be interpreted. It might have been meant by Empedocles as an intentional identity marker within a language which would otherwise have sounded unmarked and undistinguished from the standard epic-Ionic diction.15 Or, more simply, it may represent nothing else but a metrically useful tribute to a dialect which not only was widely spoken in the Western Greek world, but had long since provided the linguistic basis for one of the most prestigious genres of poetry ever invented by the Greeks: lyric poetry. It must be no coincidence, I suppose, that διπλός reappears later in an epigram (Anth. Pal. 10.101), which is transmitted under the name of Bianor and

 11 IC IV 72. For ἀπλός vs. ἀπλόoς, see coll. I 6, IV 4, V 2, etc. The Gortyn Code as well as other Cretan inscriptions of the 6th and 5th centuries BC show many forms such as διπλεῖ and τετραπλεῖ, which are adverbs deriving from ancient locatives. The adjective διπλείαν (Axos, 6th century BC) is also formed on διπλεῖ (compare πάλαι : παλαιός). These forms must be interpreted as having a short ‑ει and are therefore likely to refer to διπλός, not to διπλόος. The evidence has been collected and discussed by Bile 1988, 212–213. 12 For the tablets of Heraclea, see Uguzzoni/Ghinatti 1968, col. I 109. For the tablets of Locri, see SEG 47.1482.7–8, and Del Monaco 1997, 141–142. 13 A form ΑΠΛΟ inscribed on a stater from Caulonia has been interpreted by Manganaro 1990, 422–423 as ἁπλοῦς, but this is a mere hypothesis and a reading ἁπλός (or ἀπλός) is possible as well. 14 Brugmann 1917, 128–129 has observed a striking resemblance between the substantivized use of ἀπλός in the Gortyn Code (e.g. I 6 τὸ ἀπλὸ τεισῆται) and that of διπλά in Empedocles. 15 If so, however, this would be a unique case, since no Doric form seems to have been used by Empedocles elsewhere when he speaks in the 1st pers. sing. I completely avoid addressing the issues posed by the (mostly deceptive) use of the 1st pers. pl. in P.Strasb. a(i).6, a(ii).17, c.3; references may be found in Primavesi 2013.

112  Enzo Passa seems to have been written in some kind of literary Doric.16 If, as has been proposed, Bianor here imitates an epigram of Philip (Anth. Pal. 9.274),17 it is noteworthy that he uses two Doric forms in order to match as many expressions of his model (namely ἀροτροδίαυλε for ἀροτροπόνους and διπλά for δεύτερον ἄλγος). And this should hardly be considered accidental, since it has convincingly been argued that in Greek epigrams Doric often serves as a marker of poetic imitation.18

 Doric Features in Western Hexameter Poetry The case of B 17 D.-K. raises the thorny question of the presence of Doric features in Empedocles’ poems and, more generally, within Western hexameter poetry. Let us briefly consider its most difficult points. First, in the majority of Western literary hexameter poetry Doric seems to appear only sporadically, so that we are often tempted to doubt whether Doric features are original or later intrusions into the texts. Secondly, Dorisms are usually not metrically guaranteed, as they mostly consist in /a:/ for /e:/. Thirdly, many of the forms that we might label as Dorisms are also found in Classical Attic literature, which makes it uncertain whether they were introduced into the texts by Doric or Attic influence. In spite of all these caveats, I am still convinced that at least a handful of Doric features appearing in Western hexameter poetry are genuine and may be traced back to the oldest editions of this poetry. In what follows, I am going to briefly discuss some of the more interesting cases that I have recently come across. Starting with Empedocles, some years ago I drew attention to the remarkable number of his lines where an uncontracted /‑ĕ(h)ă/ — which is expected to ap-

 16 Ἠνίδε καὶ χέρσου τὸ γεωτόμον ὅπλον ἐρέσσει / καὶ τὸν ὑπουθατίαν μόσχον ἄγει δάμαλις, / βούταν μὲν τρομέουσα διάκτορα, τὸν δὲ μένουσα / νήπιον, ἀμφοτέρα εὔστοχα φειδομένη. / Ἴσχες, ἀροτροδίαυλε, πεδώρυχε, μηδὲ διώξῃς / τὰν διπλοῖς ἔργοις διπλὰ βαρυνομέναν ‘vois cette génisse qui traîne l’instrument du labourage, en même temps qu’elle conduit son petit veau qui tette encore, craignant le bouvier qui la mène et attendant le petit, et ainsi satisfaisant à son double devoir. Du calme, laboureur, qui parcours la double piste du charruage, n’épuise pas cette bête sur qui pèse la double charge d’un double labeur’ (transl. P. Laurens). 17 Waltz 1957, 109. By contrast, Laurens 2011, 36 sees Philip’s epigram as imitation of Bianor’s. 18 See now Coughlan 2016, 67–68, and passim. Of course, the Doric patina in Bianor’s epigram might simply be due to its bucolic subject.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  113

pear in any hexameter text modelled on Homer — has been replaced by a contracted /‑ɛ:/.19 On this point, it needs to be stressed that there is no evidence that the sources of Empedocles’ text disagreed, unlike what happens in the Homeric text, where the same endings are only occasionally transmitted in the manuscripts and are usually dismissed as Atticisms. We must therefore assume either that all the Empedoclean cases of /‑ɛ:/ depend on an Attic tradition capable of removing any traces of the original /‑ĕ(h)ă/, or that these cases are better explained in a different way. As I have already suggested, we cannot disregard the idea that the ‑ē-endings are connected to the influence of some Doric dialect, perhaps that of Akragas, since these endings are attested in Doric inscriptions from the 6th century BC onwards.20 Here, however, I would like to examine more in detail two Empedoclean forms for which new evidence has come to light in the last few years. Until recently, we were in the position of understanding very little about the /a:/ in the ‑βᾱμων compounds attested in Empedocles (B 20.7 D.-K. πτεροβάμοσι κύμβαις ‘moving-on-wings gulls’; P.Strasb. a(ii).28 ἀμπελοβάμ[ονα βότρυν] ‘mounting-onthe-vine cluster of grapes’), as well as in Attic tragedy (both in dialogues and choral odes),21 and in an elegy of Ion of Chios.22 Scholars suspected that their ā-vocalism must come from Doric lyric poetry,23 but we could not say much more about them. Now, the publication of the so-called Getty Hexameters, a Sicilian hexameter text dating to the end of the 5th or the beginning of the 4th century BC, supplies us with the phrase τὴν τετραβήμονα […] Δήμητρος {Ͱ}ὀπηδόν ‘the fourfooted [...] attendant of Demeter’, where τετραβήμονα is our first instance of an ē-vocalism in such compounds.24 This obviously encourages us to reexamine the issue, as the τετραβήμονα of the Getty Hexameters is not likely to result from either a clumsy poet or an unskilled scribe. It must rather have entered the text from a kind of poetry other than the lyric tradition: whatever our thoughts may

 19 B 59.3 διηνεκῆ; B 61.2–4 βουγενῆ, ἀνδροφυῆ, γυναικοφυῆ; B 130.1 D.-K. προσηνῆ; P.Strasb. a(ii).29 ἀψευδῆ. For details, see Passa 2005, 49. 20 See e.g. CEG 459 (Camirus, ca. 600–575 BC); IGDS I 134 (Gela, 5th century BC). 21 Compounds in ‑βᾱμων appear in Attic tragedy from Aeschylus onwards; a complete list may be found in Willi 2008, 217 n. 59. 22 Ion of Chios fr. 5.1 G.-P. (= 32 West; 93 Leurini; 4 Valerio). 23 Björck 1950, 119–121 and 336–339, especially p. 338 where the origin of ‑βᾱμων compounds is traced back to παλίμβᾱμος in Pind. Pyth. 9.18. Along the same lines, see Willi 2008, 217. 24 Jordan/Kotansky col. I 10. The hapax legomenon ταχυβήμων ascribed to Aeschylus’ Psychostasia could be a corrupted form for ταχυβάμων; see Janko 2013, 47–48.

114  Enzo Passa be about the origin and the dialect of the Getty Hexameters,25 we cannot avoid thinking that ‑βημων was at home in the epic-Ionic diction. Empedocles, therefore, was not compelled to use ‑βᾱμων, as one was unsatisfactorily inclined to conclude until some years ago, on the wrong assumption that ‑βημων had never existed. On the contrary, if he had wanted to, he might have chosen ‑βημων, as it was available in the poetic language of his time and would have perfectly suited a hexameter context. All this leads us to conclude that if Empedocles did not choose to use ‑βημων, this was because of the enormous reputation acquired by choral lyric poetry in the Western Greek world, a reputation which was so high as to compete with and even prevail over the epicIonic diction. Far from being limited to Empedocles, this state of affairs may have been common among other hexameter poets working in Magna Graecia or Sicily between the 6th and 5th centuries BC. Elsewhere I argued for the authenticity of the reading συνᾱ́ορος, which is unanimously transmitted by our manuscripts in Parmenides’ proem (B 1.24 D.-K.), whereas Homer only knows συνήορος (Od. 8.99). I suggested that this form with /a:/ was borrowed from lyric poetry because in that passage of his proem Parmenides is alluding to an epinician context.26 Equally genuine, I assume, and paralleled with similar lyric forms is the Parmenidean δίκρᾱνοι (B 6.5 D.-K.), whose /a:/ also appears in the Empedoclean βούκρᾱνα (B 61.3 D.-K.).27 In theory, we may suspect that these last two passages were so deeply influenced by Attic dialect or subsequently underwent such a heavy Atticization that all the examples of original epic-Ionic /ɛ:/ were changed into Attic /a:/ without leaving even the slightest trace of this shift in our manuscripts. Such an explanation, however, might be satisfactory for δίκρᾱνοι and  25 I will refrain from going into the problems raised by the Getty Hexameters regarding the nature, origin, and date of the text, nor will I address the question of its relationship with the many similar texts discovered around the Greek world. As to the dialect, the hypothesis put forward by Janko 2013, 55 of two archetypes composed in two different dialects, which our text should depend on, seems to me quite unconvincing and impossible to prove. I prefer to think of an original text perhaps composed in an Ionic dialect, but with Doric features soon mixed in when it began to be used in the Doric cities of Sicily and Southern Italy. An early dialect mixture has also been supposed by Tribulato 2018, who interestingly favours a Himeraean origin of the text instead of the Selinuntine one currently accepted by most scholars. 26 In my opinion, the path walked by Parmenides on his chariot in order to reach the goddess and learn from her the truth about Being is comparable to that of a charioteer aimed at winning his race to obtain the subsequent glory. A different context, absolutely unfitting for Parmenides’ imagery, is found in Hom. Od. 8.99, which has nonetheless been envisaged by many scholars as the unique model for the Parmenidean συνάορος. See Passa 2009, 132–137. 27 Passa 2009, 137–141.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  115

βούκρᾱνα, but there is no way it can be applied to συνᾱ́ορος, the presence of which in Parmenides’ text cannot be understood without the mediation of Doric lyric. In conclusion, it cannot be of any help to disregard this evidence and it seems wiser to admit that across the Western Greek world a certain number of Doric features had already started making their way into the epic-Ionic hexameter poetry by the beginning of the Classical Age.

 Doric in Panyassis We know, however, that the liaison between Doric dialects and epic poetry is more ancient than the Classical Age. Indeed, it must not be forgotten that even the Homeric and Hesiodic epics do not completely lack Doric features, since some Doric forms are rarely but surely attested in both Homer and Hesiod, where their origin is still a disturbing question.28 Over time, the habit of introducing a more or less significant number of Doric features into hexameter poetry must have gradually spread, and to a greater extent than we are ready to admit. We may infer this habit from inscriptions: if Doric could enter epigraphic hexameter texts of the Archaic and Classical Age,29 what  28 On these subjects Albio Cesare Cassio has authoritatively deployed both his teaching and his scholarly activity. Cassio’s interest in the issue of Dorisms in Homer is long-lasting; see already Cassio 1998, 18–19, and recently the paper entitled “‘Dorisms in Homer Fit No One’s Theories’ (M.L. West)” given in April 2018 at Cornell University, Ithaca. On the thorny question of Doric influence on Hesiod, which has been thought to be quite heavy and has been seen as a proof of a different underlying tradition of poetry than that attested by Homer, see Cassio 2009, 183–193. In all the above-mentioned works by Cassio, previous bibliography is extensively quoted. On some Doric pronouns in our Homeric text, see also Sara Kaczko’s contribution in this volume (pp. 19–41). 29 I do not refer here only to local Doric ‘translations’ of Homeric verses, or parts of verses, which are often attested on stones or vases; see e.g. the well-known case of κατὰ στονόϝεσαν ἀϝυτάν (CEG 145; Korkyra, ca. 600 BC), which ‘translates’ Od. 11.283 κατὰ στονόεσσαν ἀϋτήν. I rather think of the numerous examples of blending of Attic-Ionic or epic-Ionic and Doric features in the (mainly Attic) stone epigrams dated to the 6th and 5th centuries BC; they have recently been listed by Kaczko 2009, 93 n. 12. The cases of /a:/ where /ɛ:/ would be expected are likely to be due to the influence of Doric lyric, as has been maintained by Mickey 1981, 44, and is now reasserted by Kaczko 2009, 94 and n. 13. In this connection, particularly telling seem to be those epigrams such as CEG 235 = Kaczko 2016, no. 55 (Athens, probably first quarter of the 5th century BC), where the Doric Ἀθάνᾳ, appearing in the first hexameter beside the epic-Ionic πολιε̄́οχε, must have entered the text under the influence of Doric lyric or some other North-Western kind

116  Enzo Passa should have prevented it from entering literary hexameter texts at the same time period? There are two problems, though, which reduce our possibility of gaining a deeper insight into the issue we are dealing with. The first is that we have lost a large amount of literary hexameter texts of the Archaic and Classical periods, and without them we may only have a very fragmentary picture today of what the ancient situation looked like. This in turn causes the second problem, namely that we can hardly understand the criteria according to which Doric features are mixed with Ionic ones from one text to another. Nevertheless, the case of Empedocles is not unique, and in order to better understand it we probably need to broaden our horizons. It has now been recognized for some time that a not insignificant number of epic fragments ascribed to poets mainly active in the Eastern Greek world between the 7th and 5th centuries BC contain some Doric features within a general epic-Ionic linguistic shape.30 Although close scrutiny of them is beyond the scope of my paper, I would like to focus on an apparently odd form attested in a short hexameter fragment which Wilamowitz (1900, 42) persuasively ascribed to Panyassis’ Herakleia:31 (2)

πῶς δ’ ἐπορεύθης ῥεῦμ’ Ἀχ[ελω]ίου ἀργυ[ρο]δίνα Ὠκεανοῦ ποταμo[ῖo δι’] εὐρέος ὑγ[ρ]ὰ κέλευθα; (fr. 28 Matthews = 31 Bernabé) How did you cross the flow of the silver-eddying Acheloos across the watery ways of the wide river Oceanus?

At the end of the first line the genitive ἀργυ[ρο]δίνᾱ features the typically Doric contracted ending /‑a:/ < /‑a:o/. This form, which has troubled more than one generation of scholars, seems to be out of place in a hexameter poem of the Classical Age, since the adjective ἀργυροδίνης already appears in the epic-Ionic tradition represented both by Homer (Il. 2.753 Πηνειῷ [...] ἀργυροδίνῃ ||; Il. 21.8 ἐς ποταμὸν [...] βαθύρροον ἀργυροδίνην || ~ 21.130 ποταμὸς [...] ἐύρροος

 of poetry; see Palumbo Stracca 1987, 431; Kaczko 2016, 221–223. For another case of dialect mixture of this kind, see n. 35 below. 30 The evidence concerning the hexameter texts has been collected and discussed by Lulli 2007, 227–241. 31 The fragment is ascribed to Panyassis’ Herakleia also by Matthews 1974, 131–134. However, the authorship has been questioned, and some scholars have preferred to think of a Hellenistic poet: see Powell 1925, 248; West 1976, 172–173. Bernabé 1996, 186 has placed the fragment among the dubia.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  117

ἀργυροδίνης ||) and Hesiod (Theog. 340 Ἀχελῷόν τ’ ἀργυροδίνην ||; fr. 26.19 M.W. ἀργ]υροδίνεω ||; see also fr. 229.18 M.-W. ἀργυρ[οδίνην ||). Hence many attempts have been made to explain away the form ἀργυροδίνᾱ. Among these attempts, it is worth mentioning that of Wackernagel (1916, 162), who tried to explain ἀργυροδίνᾱ as the elided form ἀργυροδίνᾱ’ deriving from an ancient uncontracted genitive in /‑a:o/, which is common in the archaic epics (type Ἀτρείδᾱο). According to Wackernagel, Panyassis preferred this elided form over the usual and metrically equivalent Homeric (Ionic) ἀργυροδίνε͜ω (see Hesiod’s fr. 26.19 M.-W. cited above), since the latter had become obsolete in his time. Two facts call for attention in Wackernagel’s solution. First, it is viable only because ἀργυροδίνᾱ is followed in the second line by Ὠκεανοῦ, namely a word beginning with a vowel. Secondly, the ending ‑ᾱo is never elided in the archaic epics.32 To be extremely frank, I am under the impression that Wackernagel’s ἀργυροδίνᾱ’ was concocted precisely for the purpose of removing a Doric form from a hexameter context on the assumption that “einem wirklichen Epiker, auch einem solchen dorischer Herkunft, war Beimischung von Dorismen verwehrt” (my emphasis). Conversely, Wackernagel had no difficulty at all in accepting the form ἀργυροδίνᾱν in Callimachus’ Hymn to Demeter (line 13 τρὶς μὲν δὴ διέβας Ἀχελώϊον ἀργυροδίναν). In his view, as a Hellenistic poet Callimachus was not a “wirklicher Epiker” and a “Beimischung von Dorismen” within his hexameter texts is thus perfectly acceptable. In the same vein as Wackernagel, but even more biased than him, appears to be West (1976, 172–173), who has rejected the attribution of our fragment to Panyassis especially on account of the genitive ἀργυροδίνᾱ, which according to him suggests that the fragment comes from a Hellenistic poem. Luckily, we can understand something more about Panyassis’ genitive thanks to the important evidence provided by the Getty Hexameters, a text which I have already mentioned above (Section 2). Here we read the phrase αἶγ’ ἀκαμαντορόα νασμοῦ θαλερoῖo γάλακτoς / βριθομένην ‘a nanny laden with an unceasing flow of rich milk’, where the Doric genitive ἀκαμαντορόᾱ appears beside a typically Homeric genitive such as θαλερoῖo.33 Once ἀκαμαντορόᾱ is taken into account, the ἀργυροδίνᾱ of Panyassis is no longer an isolated form, and

 32 See e.g. West 1966, 400. In Homer a genitive such as Πηληϊάδε͜ω followed by a vowel could cover for an older *Πηληϝιάδᾱ(o), where the elision is needed in order for the latter to have the same prosodic value as the former, but it seems to make little sense to look for such genitives in an epic poem of Panyassis’ times. 33 Jordan/Kontansky col. I 11–12.

118  Enzo Passa Wackernagel’s attempt to explain it as an elided ἀργυροδίνᾱ(ο) becomes untenable. In fact, if we suppose that an Ionic archetype underlies the Getty Hexameters, we could speculate on the possibility that ἀκαμαντορόᾱ replaced an Ionic genitive ἀκαμαντορόε͜ω; but there is no way for it to be an elided form for an old ἀκαμαντορόᾱo, since it is followed by a consonant. At this point, we need to remember that the Derveni Papyrus goes in a completely different direction from Panyassis (and the Getty Hexameters), since when it quotes a hexameter from an Orphic Theogony it offers the text Ἀχελῴου ἀργ]υροδίνε͜ω ||, which is perfectly in line with Homeric and Hesiodic (Ionic) diction.34 On this basis, Janko (2013, 49) has concluded that “in the epic of the sixth and fifth centuries the termination ‑εω became increasingly unstable and liable to replacement by ᾱ in poets whose own dialect was Doric”. This is of course possible, but, while providing us with an explanation for Panyassis’ form, this solution avoids mentioning that the ‘old’ ‑εω of the Derveni Papyrus is likely to depend on a conservative textual tradition closely tied with the Attic-Ionic milieu. As for Panyassis’ ἀργυροδίνᾱ, it is unavoidable to conclude that the ‑ā-genitives were not out of place in hexameter texts of the Classical Age. However, unlike Janko, I am not so sure that they are necessarily to be ascribed only to spoken Doric. Indeed, they might have come from lyric poetry, 35 as we are encouraged to think by the fact that both adjectives ἀργυροδίνᾱς and ἀκαμαντορόᾱς do appear, with a Doric shape, in two Bacchylides’ poems. The former is found in 8.26–27 ἐπ’  34 Kouremenos/Parassoglou/Tsantsanoglou col. 23.11. It is a curious coincidence that the same Orphic line has also been quoted by the anonymous author of the commentary on Iliad 21 (P.Oxy. 2.221), which happens to be the only source for this fragment of Panyassis. 35 This seems to be supported by an epigram dedicated by Alcmeonides, a member of the Athenian Alcmeonidai, in the temple of Apollo Ptoios at Akraiphia in Boeotia to celebrate his victory in the chariot race at the Great Panathenaia (ca. 540 BC?; CEG 302 = Kaczko 2016, no. 120). Unlike most Archaic and Classical epigrams, the text was composed in iambic trimeters, a metre which allowed its author to insert a higher number of personal names and patronymics than hexameter or any other dactylic metre would have done. In the first line, which runs as [Φοί]βō μέν εἰμ’ ἄγαλ[μα Λ]ᾱτ[οί]δᾱ καλόν, the genitive Λ]ᾱτ[οί]δᾱ < ‑ᾱο has the typically Doric-lyric shape already attested in Alcman (tetrasyllabic Λᾱτοΐδᾱ appears in fr. 48.1 PMG) and Pindar (trisyllabic Λᾱτοίδᾱ, as in our text, is found in Pyth. 1.12). A different genitive stands out in the incipit of the ‘Homeric’ Hymn to Hermes 158 Λητοΐδου ὑπὸ χερσί, where Λητοΐδου is an Attic form; the Ionic genitive would have been Λητοΐδεω. Had Alcmeonides wanted to use a genitive coherent with his own dialect or one perfectly acceptable in a text claiming poetic preciousness, he would have found an adequate option either in the Attic Λητοΐδου or in the Ionic Λητοΐδε͜ω respectively. Yet, this was clearly not what he was aiming at, since in composing his own epigram he purposely evoked an epinician atmosphere by blending Doric-lyric features such as Λ]ᾱτ[οί]δᾱ and Ἀθάναις (line 5) with epic-Ionic ones such as hίποισι νικέ[σας and πανέγυρις (lines 3 and 5). For a commentary on this epigram, see now Kaczko 2016, 438–446 with previous bibliography.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  119

ἀργυροδίνα / ὄχθαισιν Ἀλφειοῦ ‘on the banks of silver-eddying Alpheus’,36 whereas the latter occurs in 5.180–181 τόν τ’ ἀκαμαντορόαν / Ἀλφεόν ‘the untiringly streaming Alpheus’. The first example is particularly striking, since it features the same Doric contracted form ἀργυροδίνᾱ as Panyassis, with the same prosody and at the end of the line, as in Panyassis’ fragment. In both passages of Bacchylides, anyway, we do not seem to be faced with linguistic features taken from some spoken Doric dialect, but rather with refined river-epithets which had gained their permanent literary status within Doric lyric poetry.37

 From the Classical to the Hellenistic Age: An Alleged Gap It is time we started questioning what seems to have become a dogma among scholars: Doric features within Classical hexameter poetry are a matter of fact, rather than a later conspiracy of forgers. Instead of spending our time trying to demonstrate that our texts and testimonies are unreliable, we should rather try to understand why these features were introduced by authors distant in space and working in different genres. The answer is certainly not an easy one. We might ascribe them to the particular subject of certain poems connected with Doric traditions, such as Panyassis’ Herakleia. Or we might see them as the result of local performances held in Doric areas, where the poets inserted some Doric features into their compositions to honour their audience, as perhaps is Empedocles’ case.38 In my opinion, however, we had better search for a more comprehensive explanation. This could lie in a process of hybridization of the traditional epic-Ionic diction — which by the end of the Archaic Age was perceived to be quite obsolete — by means of either the poets’ spoken dialects or other prestigious poetic traditions, first of all the – still lively – Doric choral lyric. In other words, adding Doric features into a hexameter text might have become a matter of style in both the Western and the Eastern Greek world of the Classical Age, a way to set up a  36 See also the fragmentary epinician ode 12.42. 37 See Tribulato 2018, 17, who quotes other examples of adjectives in ‑ρόᾱς in both lyric poems and tragic choruses. 38 Both hypotheses on the Doric features attested within some hexameter and elegiac texts of the Archaic and Classical Ages are actually taken into account by Lulli 2007, 244–245. An attempt at honouring the local context has also been posited by many scholars to explain the dialect mixture of Alcmeonides’ epigram; see now Kaczko 2016, 441, and n. 35 above.

120  Enzo Passa renewed and more inclusive poetic diction, as opposed to the traditional Homeric language in which Doric dialects still tended to be left aside. In doing so, epic (hexameter) poetry of the Classical Age also tried to renew its traditional ambition to be authentically panhellenic, namely aimed at all the Greeks across the Greek world. Consequently, it could be meaningful that our major evidence for Doric features in literary texts comes from the ‘wings’ of the ancient Greek world, whereas the ‘body’ — that is, mainland Greece — seems to be quite voiceless in this respect. It looks as though, after the Ionian epic poetry had joined together with the Attic literary tradition and the ancient hexameter poems took on the dialect shape we are acquainted with, mainland Greece became much more conservative than peripheral areas were. As far as we can see, in this process there were no general rules, and much depended on the individual poet and his choices of both style and content. We have, however, enough clues to suspect that the insertion of Doric forms into hexameter poetry was not carried out only by Doricspeaking poets, but could also have been experimented with by Ionic-speaking ones. Unfortunately, the extant materials are so scant that we cannot speculate any further. This, however, calls for a different approach in a number of epigraphic texts where the mixture of epic-Ionic and Doric features appears to be particularly significant, such as the Getty Hexameters and its more ancient parallel texts.39 While struggling to reconstruct an archetype and its ‘original’ dialectal shape (if it ever existed), we risk overlooking what we have in front of us, namely the real texts circulating across the ancient world. And in these texts we are faced with a poetic language we would probably dismiss as inconsistent in whatever ‘literary’ text it occurred in before (and possibly even after) the Hellenistic Age. Nevertheless, the few hints of Doric we find in literary authors including Empedocles and Panyassis seem to me sufficient to doubt whether the stylistic principle of dialect mixture proclaimed and practiced by legions of Alexandrian poets was completely an original invention of theirs. For instance, when Callimachus theorizes the ‘mixed dialect’ (τὸ σύμμικτον, Iamb. 13.18), he might be developing and authoritatively establishing older poetic practices of dialect mixing. Likewise, where he matches Doric dialect with hexameter or elegiac metre (Hymns 5 and 6), he could take inspiration from older hexameter and elegiac poetry, which he obviously knew far better than we do.40 Accordingly, the practice of adding  39 The complete list of the parallel texts is given by Janko 2015, 3; for a shorter list of the most ancient ones, see Tribulato 2018, 8–9. 40 Similar conclusions are drawn by Lulli 2007, 234. Even more than hexameter poetry, elegy must have known early dialect mixture. Various Dorisms within some elegiac poems must be

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  121

some Doric features into the basically Ionic language attested by a certain number of Hellenistic epigrams41 might not be an invention of the Hellenistic Age, but could reflect the continuation of previous linguistic conventions attested both in literary and epigraphic texts. And, if this hypothesis were true, we would be right in reconsidering an important aspect of what we are used to calling the ‘Alexandrian Poetics’.42

Bibliography Bernabé, A. (1996), Poetarum epicorum Graecorum testimonia et fragmenta, vol. 1, 2nd edn., Stutgardiae/Lipsiae. Bile, M. (1988), Le dialecte crétois ancien: Étude de la langue des inscriptions. Recueil des inscriptions posterieures aux IC, Paris. Björck, G. (1950), Das Alpha impurum und die tragische Kunstsprache, Uppsala. Brugmann, K. (1917–1920), “Zur griechischen und lateinischen Wortgeschichte”, in: IF 38, 128– 143. Cassio, A.C. (1998), “La cultura euboica e lo sviluppo dell’epica greca”, in: M. Bats/ B. D’Agostino (eds.), Euboica: L’Eubea e la presenza euboica in Calcidica e in Occidente. Atti del convegno internazionale di Napoli (13–16 novembre 1996), Napoli, 11–22. Cassio, A.C. (2009), “The Language of Hesiod and the Corpus Hesiodeum”, in: A. Rengakos/ F. Montanari/C. Tsagalis (eds.), Brill’s Companion to Hesiod, Leiden, 179–201. Coughlan, T. (2016), “Dialect and Imitation in Late Hellenistic Epigram”, in: E. Sistakou/ A. Rengakos (eds.), Dialect, Diction, and Style in Greek Literary and Inscribed Epigram, Berlin/Boston. Gallavotti, C. (1975), Empedocle: Poema fisico e lustrale, Milano. Gow, A.S.F./Page, D.L. (1965), The Greek Anthology: Hellenistic Epigrams, vol. 1, Cambridge. Hinge, G. (2006), Die Sprache Alkmans: Textgeschichte und Sprachgeschichte, Wiesbaden. Hoffmann, O. (1898), Die griechischen Dialekte in ihrem historischen Zusammenhange, vol. 3: Der ionische Dialekt, Göttingen. Hude, K. (1908), Herodoti historiae, 2 vols., Oxonii. Janko, R. (2013), “The Hexametric Incantations against Witchcraft in the Getty Museum: From ‘Archetype’ to Exemplar”, in: C.A. Faraone/D. Obbink (eds.), The Getty Hexameters: Poetry, Magic, and Mystery in Ancient Selinous, Oxford, 31–56. Janko, R. (2015), “The Hexametric Paean in the Getty Museum: Reconstituting the ‘Archetype’”, in: ZPE 193, 1–10.

 ancient, as is shown beyond doubt by both epigraphic and literary texts: for an overview see Passa 2016, 275–277. For the Doric features in Simonides’ elegy on the battle of Plataea, see Lulli 2007, 241–244. 41 See e.g. Gow/Page 1965, xlv–xlvii. 42 I wish to thank Olga Tribulato for her generosity in giving me precious suggestions and making important corrections. All remaining mistakes are mine.

122  Enzo Passa Kaczko, S. (2009), “From Stone to Parchment: Epigraphic and Literary Transmission of Some Greek Epigrams”, in: Trends in Classics 1, 90–117. Kaczko, S. (2016), Archaic and Classical Attic Dedicatory Inscribed Epigrams: Literary and Linguistic Commentary, Berlin/Boston. Kretschmer, P. (1923), “Literaturbericht für die Jahre 1919–1920”, in: Glotta 12, 179–277. Laurens, P. (2011), Anthologie grecque, vol. 9, Paris. Leumann, M. (1977), Lateinische Laut- und Formenlehre, München. Lulli, L. (2007), “‘Anomalie’ linguistiche e performances poetiche: Osservazioni sui tratti linguistici epicorici nell’epica greca postomerica e nell’elegia arcaica storico-narrativa”, in: SemRom 10, 223–248. Manganaro, G. (1990), “Due studi di numismatica greca”, in: ASNP 20, 409–427. Martin, A./Primavesi, O. (1999), L’Empédocle de Strasbourg (P. Strasb. gr. Inv. 1665–1666): Introduction, édition et commentaire, Berlin/New York. Matthews, V.J. (1974), Panyassis of Halikarnassos: Text and Commentary, Lugduni Batavorum. Mickey, K. (1981), “Dialect Consciousness and Literary Language: An Example from Ancient Greek”, in: TPhS 79, 33–66. Nöthiger, M. (1971), Die Sprache des Stesichorus und des Ibycus, Zürich. Palumbo Stracca, B.M. (1984), “Differenze dialettali e stilistiche nella storia dell’epigramma greco”, in: G. Bolognesi/V. Pisani (eds.), Linguistica e filologia. Atti del VII convegno internazionale dei linguisti (Milano, 12–14 settembre 1984), Brescia, 429–434. Passa, E. (2004), “Pausania, figlio di Anchìta o di Ànchito? (Emped. 31 B1 D.-K.)”, in: SemRom 8, 23–29. Passa, E. (2005), “Dialetti parlati, tradizioni letterarie, livelli di ricezione antica del testo dei Preoscratici: Gli Eleati e la cultura attica”, in: F. Bertolini/F. Gasti (eds.), Dialetti e lingue letterarie nella Grecia arcaica. Atti della IV Giornata ghisleriana di Filologia classica (Pavia, 1–2 aprile 2004), Como/Pavia, 45–69. Passa, E. (2009), Parmenide: Tradizione del testo e questioni di lingua, Roma. Passa, E. (2016), “L’elegia e l’epigramma su pietra”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano, 260–288. Powell, I.U. (1925), Collectanea Alexandrina: Reliquiae minores poetarum Graecorum aetatis Ptolemaicae 323–146 a.C. Epicorum, elegiacorum, lyricorum, ethicorum, Oxford. Primavesi, O. (2013), “§17. Empedokles”, in: H. Flashar/D. Bremer/G. Rechenauer (eds.), Grundriss der Geschichte der Philosophie, vol. 1: Die Philosophie der Antike: Frühgriechische Philosophie, Basel, 667–739. Schwyzer, E. (1939), Griechische Grammatik, vol. 1: Allgemeiner Teil, Laulehre, Wortbildung, Flexion, München. Tribulato, O. (2018), “Writing and Language in the Getty Hexameters”, in: C. Antonetti (ed.), Gli esametri Getty e Selinunte: Testo e contesto, Alessandria, 6–20. Uguzzoni, A./Ghinatti, F. (1968), Le tavole greche di Eraclea, Roma. Wackernagel, J. (1916), Sprachliche Untersuchungen zu Homer, Göttingen. Waltz, P. (1957), Anthologie grecque, vol. 7, Paris. West, M.L. (1966), Hesiod: Theogony, Oxford. West, M.L. (1976), “Review of V.J. Matthews, Panyassis of Halikarnassos (Lugduni Batavorum 1974)”, in: CPh 71, 172–174. West, M.L. (1988), “The Rise of the Greek Epic”, in: JHS 108, 151–172. Wilamowitz-Moellendorff, U. von (1900), “Review of The Oxyrynchos Papyri. Part II edited by Bernard G. Grenfell and Arthur S. Hunt (Oxford 1899)”, in: GGA 162, 29–58.

Empedocles in the West, Panyassis in the East: Doric and Hexameter Poetry  123

Willi, A. (2008), Sikelismos: Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.–5. Jh. v. Chr.), Basel.

Lucia Prauscello

Of Land, Ancestral Property and Prophecy in Corinna PMG 654 col. iii ll. 37–39 Abstract: Genealogies are always a contested site of negotiation between local and supra-local traditions.1 This paper argues that the correct interpretation of γῆα[ν ϝ]ὰν ἀππασάμενος in Corinna PMG 654 col. iii l. 39 — with reference to Orion’s role within the sub-catalogic section of the genealogy of prophets in an unnamed Boeotian sanctuary — is that already suggested by the scholia: Orion did have to ‘regain’ possession of his land. A new re-appraisal of the evidence, literary and epigraphical, for the internal gemination /‑pp‑/ before a vowel in Boeotian suggests that the standard interpretation of ‑ππ‑ as deriving from an original *kw‑ is still to be preferred on both linguistic and semantic grounds. Corinna’s ἀππασάμενος most probably derives from *ἀνα‑ππᾱ‑ > *ἀμ‑ππᾱ‑ > ἀππᾱ‑, and not from *ἀποπασάμενος as recently suggested. Keywords: Corinna, Orion, ἀππασάμενος, ἔμπασις/ἔππασις, τἀππάματα, ἀπόκτησις, ἀποκτάομαι.

 Orion and ‘His Land’ between Panhellenic and Boeotian Traditions Orion, a well-known figure of panhellenic myth since Homer but also a specifically Boeotian hero with strong Tanagrian connections,2 features prominently  1 My heartfelt thanks to A.C. Cassio, G. D’Alessio, E. Dettori, D. Mastronarde, O. Tribulato and G. Ucciardello for having read and substantially improved various drafts of this paper. My sincerest thanks also to A. Willi for an inspiring discussion and to L. van Beek for bringing to my attention his article on the etymology of πέπαμαι. It goes without saying that I am solely responsible for any remaining error of fact or judgement. In particular, I could not have written this article without Albio’s immense intellectual generosity. To me Albio represents the living embodiment of Friedrich Schlegel’s dictum that ‘[d]as kleinste φλμ[Philologem] ist encyclopaedischer Art und kann nur von einem Polyhistor beantwortet werden’ (KFSA 16, p. 65 no. 60). Unless otherwise specified, Corinna’s fragments and testimonia are quoted from Page’s PMG. 2 Palaephatus De incred. 51 Festa locates Hyrieus, mortal (foster-)father of Orion, at Tanagra (ᾤκει μὲν ἐν Τανάγρᾳ τῆς Βοιωτίας); for the dating of Palaephatus to the second half of the 4th century BC, see Hawes 2014, 38 and her appendix I at 227–235. Strabo’s phrasing at 9.2.12 (ἡ Ὑρία δὲ τῆς Ταναγραίας νῦν ἐστί, πρότερον δὲ τῆς Θηβαΐδος· ὅπου ὁ Ὑριεὺς μεμύθευται καὶ ἡ τοῦ https://doi.org/10.1515/9783110621747-007

126  Lucia Prauscello within the extant fragments of Corinna’s poetry.3 Recent scholarship has increasingly recognized the sophisticated ways in which Corinna’s treatment of myths negotiates the tension between local and supra-local traditions, and the case of Orion has frequently been cited as a paradigmatic example of this creative exercise in forging new ‘contractual’ mythological genealogies among Greek citystates.4 In this paper I shall focus on a single, isolated aspect of this broader narrative strategy, which has recently been questioned: the exact meaning of the expression γῆα[ν ϝ]ὰν ἀππασάμενος (l. 39), with reference to Orion’s role within the sub-catalogic section of the genealogy of prophets at an unnamed Boeotian sanctuary in PMG 654 col. iii ll. 27–41.5 At PMG 654 col. iii ll. 13–26 the present incumbent of the oracular site, who identifies himself as one of the fifty sons of Orion (cf. ll. 28–29 ἐς πεντείκο[ντα] κρατερῶν / ὁμήμων, and l. 38 Ὠα[ρί]ων ἁμὸς γενέτωρ),6 comforts a distressed  Ὠρίωνος γένεσις, ἥν φησι Πίνδαρος ἐν τοῖς διθυράμβοις ‘Hyria now belongs to the Tanagraean territory, though in earlier times it belonged to the Theban territory. It is the scene of the myth of Hyrieus, and of the birth of Orion, of which Pindar speaks in his dithyrambs’) implies that Pindar (fr. 73 S.-M.) situated the birth of Orion at Hyria (on Hyria as object of territorial contention between Thebes and Tanagra, see Schachter 2016, 97 n. 44 = Schachter 2003, 59 n. 46). According to Pausanias 9.20.3 (ἔστι δ’ Ὠρίωνος μνῆμα ἐν Τανάγρᾳ), Tanagra was the site of Orion’s tomb (on μνῆμα in Pausanias referring to the actual burial place and not to memorials or cenotaphs, see Dunn 1995, esp. 349). A 3rd-century AD inscription (IG XII Suppl. 646 l. 16 [ἀγωνοθ]έτης Ὠρειονίων) records the existence of the Oreoneia, an ephebic agon at Tanagra dedicated to Orion (the stele was found in Chalcis but is most likely a pierre errante to be located in Tanagra: cf. Robert 1968, 24–27). For the association of Orion with Tanagra and Hyria in Nonnus, see Debiasi 2010a, 106 with n. 5. 3 Cf. PMG 654 col. iii. ll. 35–41 (Orion, prophet and object of catasterism), PMG 655 fr. i. ll. 12–16 (Orion and his mighty fifty sons begotten by intercourse with the Nymphs), PMG 656 (= Αnt. Lib. 25 quoting from the first (book?) of Corinna’s ϝεροῖα: the story of Orion’s daughters Metioche and Menippe), PMG 673 (= schol. Nicand. Ther. 15 p. 41 Crugnola: his Tanagrean origin, exceptional piety and civilizing undertakings during his wanderings), PMG 662 (= Ap. Dysc. GG. 2.1.1, 77.5 Schneider quoting from Corinna’s Cataplous: Orion’s victory and naming of a land after himself). See also the brief synopsis by Renaud 2004, 58–61. 4 See e.g., among others, Collins 2006 (esp. 23–26 on Orion); Vergados 2012; Kousolini 2016; McPhee 2018. 5 The two most likely candidates seem to be either Apollo’s sanctuary at Mt Ptoion (Page 1953, 25; Schachter 1981, 61–64, followed e.g. by Renaud 2004, 169–170) or an oracular site in the neighbourhood of Aulis, linked to Glaucos Pontios, a prophetic figure at home in Anthedon (Schachter 1995, 1069–1072). 6 Wilamowitz 1907, 33 was the first to propose to supplement the proper name Ἀκ[ρη]φείν at col. iii l. 31 and identify it with the speaking persona of col. iii ll. 12–46 but cf. the linguistic objections by Lobel 1930, 363. For a new assessment of the evidence, see Schachter 1995, 1063– 1066. Gentili/Lomiento 2001, 13–16 identify the speaker of the prophetic speech with Parnes.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  127

Asopus about the abduction of his nine daughters by Zeus, Poseidon, Apollo and Hermes by predicting for him and his children an offspring of demigods (ll. 22– 23: εἱρώων γενέθλαν / ἐσγεννάσονθ᾽ εἱμ[ιθί]ων). To prove his prophetic credentials, the speaker launches himself into a genealogy of the μάντεις who have preceded him at the oracular site: (1)

πράτοι [μὲν] γὰ[ρ Λατ]οΐδας δῶκ’ Εὐωνούμοι τριπόδων ἐσς ἱῶν [χρε]ισμὼς ἐνέπειν, τὸν δ’ ἐς γᾶς βαλὼν Οὑριεὺς τιμὰ[ν] δεύτερος ἴσχεν, πῆς [Ποτ]ιδάωνος· ἔπιτ’ Ὠα[ρί]ων ἁμὸς γενέτωρ γῆα[ν ϝ]ὰν ἀππασάμενος· χὠ μὲν ὠραν[ὸ]ν ἀμφέπι τιμὰν δ[......]ν οὕταν·

35

40

(Corinna PMG 654 col. iii ll. 32–41) For first to Euonymus it was given by the son of Leto to utter oracles from his own tripods: but Hyrieus cast him forth from the land, and held the honour second, son of Poseidon; and then Orion, our father, having regained his own land. Now he dwells in the sky, and his honour has fallen to me. (transl. Page 1963, 24)

According to the genealogy here, the first to be given the gift of prophecy by Apollo was Euonymus (ll. 32–34);7 Hyrieus, Poseidon’s son, forcefully displaced Euonymus as prophet (ll. 35–37); then Orion — presumably Hyrieus’ son according to the Boeotian-Euboean version of the Orion myth followed by Corinna — took over (ll. 37–41);8 and since he now ‘dwells in the sky’ (l. 40 ὠραν[ὸ]ν ἀμφέπι),9 it is one of his fifty sons who has taken up the prophetic γέρας.

 7 Son of Cephisus, the Boeotian river, and father of Aulis: see Steph. Byz. s.v. Αὐλίς (α 541) Billerbeck: πόλις Βοιωτίας κατὰ Χαλκίδα […] Τρύφων (fr. 91 de Velsen) παρὰ τὸ ἀλισθῆναι ἀλίς καὶ Αὐλίς. βέλτιον ἀπὸ τῆς Αὐλίδος τῆς Εὐωνύμου τοῦ Κηφισσοῦ. Cf. D’Alessio 2005, 184. 8 See Debiasi 2010a for a discussion of the two main mythical variants of Orion’s ancestry and their possible sources (Orion son of Poseidon and Euryale in [Hes.] fr. 148a M.-W. [= [Eratosth.] Catast. 32] and Orion son of Zeus, Poseidon and Hermes and foster-son of Hyrieus in [Hes.] fr. 148b M.-W. [= schol. in Germanici Aratea 93.13–20 Breysig]). Debiasi 2010a suggests Euphorion and not Hesiod or Pseudo-Hesiod as source of [Hes.] fr. 148b M.-W. I intend to address elsewhere Debiasi’s ingenious reconstruction. McPhee 2018, 217–221 speculates that in PMG 654 col. iii ll. 32–41 Orion is not the son of Hyrieus and that only this unattested genealogy can make sense of Orion’s ‘regaining his own land’. For possible alternative reasons for Orion’s reconquest, see below Section 3. 9 On Orion’s catasterism in the ancient sources, see Renaud 2004, 140–148; 330–363.

128  Lucia Prauscello The verb ἀππασάμενος at l. 39 (written αππᾱcάμενοc in the papyrus) is glossed in the right margin by the scholiast of P.Berol. 13284 with the form ανακτ ̣σαμ(ενο)ς, probably to be reconstructed as ἀνακτη̣σάμ(ενο)ς (Wilamowitz 1907, 33; Crönert 1908, 174; Page 1953, 15) or ἀνακτα̣σάμ(ενο)ς (Vessella 2012, 813).10 Αccordingly, the expression γῆα[ν ϝ]ὰν ἀππασάμενος (l. 39) has traditionally been translated ‘after having regained his own land’:11 that is, Orion, if we follow the vulgata, had to retake possession of his land for reasons now obscure to us but evidently known or at least plausible to Corinna’s audience, before becoming προφάτας at the oracular site and being transformed into a star. As observed by Wilamowitz and all subsequent commentators of P.Berol., this variant of the myth of Orion (his having to regain possession of his land) is otherwise unattested. Recently Méndez Dosuna, in a series of interconnected articles,12 has repeatedly advocated an alternative interpretation of ἀππασάμενος as derived not from *ἀναπάομαι ‘to regain, to recover for oneself’ (= Att. ἀνακτάομαι), as commonly held, but from *ἀποπάομαι ‘to lose possession, to alienate’ (= Att. ἀποκτάομαι).13 According to Méndez Dosuna, Orion thus succeeds his father Hyrieus at the oracular site “après avoir renoncé à sa terre”.14 This new interpretation of ἀππασάμενος is part of a broader re-examination of the origin of the internal gemination /‑pp‑/ before a vowel in the Boeotian dialect in general. Méndez Dosuna’s proposal of reading ‑ππ‑ in Boeotian as deriving from apocope of ἀπό + the root

 10 After τ there is a lacuna of one letter and before σ there is a spot of ink at mid height, theoretically compatible with either α or η. Vessella 2012, 813 has raised the interesting possibility of reading ἀνακτα̣σάμενος rather than ἀνακτη̣σάμενος (cf. e.g. ἐσγεννάσονθ᾽ at col. iii l. 23 glossed in the right margin with εκγε̣[ ]ν̣ασονται, where the thematic vowel of the verb remains α also in the ‘Attic’ rendering of the scholiast), though he himself admits that at l. 39 η cannot be ruled out. On the modus operandi of the scholiast of P.Berol. and his way of glossing Boeotian forms, at times, with only partially adapted koine forms according to the Hellenistic grammatical doctrine of the τροπαί, see Vessella 2012, 813–816. 11 Cf. e.g. Page 1953, 24; Schachter 1995, 1069; Gentili/Lomiento 2001, 9. Renaud 2004, 60–61 with n. 238 erroneously translates “ensuite, Orion, notre géniteur, a acquis sa terre” as if the simplex πάoμαι were used instead of the complex ἀναπα‑. 12 Méndez Dosuna 2007a; Méndez Dosuna 2007b; cf. also Méndez Dosuna 2016, 124. 13 For the κτη‑/κτᾱ‑ (perhaps from *tkeh2‑: see EDG s.v. κτάομαι; for a different view cf. LIV 619 n. 1) in Attic-Ionic corresponding to πᾱ‑ (< *kweh2‑) in other dialects, see Meier-Brügger 1978. 14 Méndez Dosuna 2007a, 312. A few lines below the author glosses the participial expression γῆα[ν ϝ]ὰν ἀππασάμενος with “c.à.d. ‘après avoir quitté sa terre’”. As we shall see below, this semantic equivalence is far from self-evident.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  129

*πᾱ‑ has met with some favour.15 A fresh reappraisal of the available evidence for internal ‑ππ‑ + vowel in Boeotian will show that the standard interpretation of ‑ππ‑ as deriving from an original *kw‑ is still to be preferred on both linguistic and semantic grounds (Section 2). I shall then argue that ἀππασάμενος = Att. ἀνακτησάμενος is still the most likely interpretation at PMG 654 col. iii. l. 39 and that it is compatible with what we know about Orion’s wanderings and his prophetic role (Section 3).

 The Epigraphic Evidence Méndez Dosuna’s interpretation of Corinna’s ἀππασάμενος = Att. ἀποκτησάμενος relies on a survey of the extant epigraphic evidence for the gemination of ‑ππ‑ before a vowel in Boeotian. Apart from ἀππασάμενος in PMG 654 col. iii l. 39, there are only three other such cases attested in Boeotian, all of them recorded in inscriptions: (1) the word ἔππασις (73x) and its minority variant ἔμπασις (8x) in Boeotian proxeny decrees dating from the 4th to the 2nd century BC;16 (2) the neuter plural ταππαματα in IG VII 3172 (= Migeotte no. 13: this is the graphic rendition in the inscription, without diacritical signs and accents), an inscription where Nicareta of Thespiae claims the restitution of a loan, made in several instalments, to the city of Orchomenos (end of the 3rd century BC); (3) the anthroponyms Γυνόππασις (IG VII 505, Tanagra, 3rd century BC), Θιοππάστω (IOropos 93, 3rd century BC) and perhaps Θι̣ [ό]ππαστος (SEG 22.390, Thespiae, 3rd century BC). Phonological accounts of the geminate ‑ππ‑ in the cases above mentioned have been of two types: ‑ππ‑ has been explained either (a) as the result of a total assimilation of the nasal to the following occlusive, that is, ‑ππα‑ < *‑μπα‑ < *‑νπα‑: thus  15 See e.g. Knoepfler 2008, 619 no. 225; Lhôte 2008, 672 no. 288; Vessella 2012, 813 n. 23; van Beek 2016, 431–432 (though proposing a different etymology for *πᾱ‑, i.e. from the extended PIE root *peh2‑s‑ ‘to pasture, to guard’). Cf. also García Ramón 2017, 45 who, however, seems to misunderstand Méndez Dosuna by attributing to him a derivation of ἀππασάμενος from ἐπιπασάμενος (sic). The explanation of the form ]τὰν ἔπασ̣ι[ν in the Dodona lamella 4151B in CLOD II 405 ad loc. is most confusing: the editors apparently accept the equivalence suggested by Méndez Dosuna 2007a between ἔππασις and ἐπίκτησις but then quote Buck 1955, 127 who interprets ἔππασις as the Boeotian form of the Attic ἔγκτησις (see below). 16 Méndez Dosuna’s record, based on Vottéro’s 2001 catalogue, shows a total of 66 occurrences of ἔππασις (Méndez Dosuna 2007a, 296). A search of the PHI database, supplemented with Fossey 2013, has yielded 73 certain attestations of the form ἔππασις (following Méndez Dosuna 2007a, 295 with n. 4, the graphic variant ἔπασις in IG VII 2407 and 3167 has been discounted because of its inherent ambiguity): see Appendix 1 below.

130  Lucia Prauscello *ἀναπασάμενος > *ἀμπασάμενος > ἀππασάμενος, *ἔνπασις > ἔμπασις > ἔππασις, i.e. = Attic ἔγκτησις; *τἀναπάματα > *τἀμπάματα > τἀππάματα;17 or (b) as the reflex of *kw‑ before vowel, i.e. πᾱ- < *kweh2‑ (cf. e.g. ἵππος): to quote Schulze’s words (following an insight by Schmidt),18 “das ππ in ἔππασις ist nicht das Produkt einer nachträglichen Assimilation, sondern gehört schon der einfachen Wurzel an”.19 Accordingly, we should reconstruct an original *ἐν‑ππᾱ‑ > *ἐμ‑ππᾱ‑ > ἔππᾱ‑ for ἔππασις, *ἀνα‑ππᾱ‑ > *ἀμ‑ππᾱ‑ > ἀππᾱ‑ for ἀναππασάμενος and τἀππάματα; that is, as noted by Hermann, “[d]a hinter ππ vermutlich kṷ steckt, wird der Verschluß hinter dem Nasal zur ersten Silbe gehört haben”.20 Méndez Dosuna rightly rejects the first explanation because (1) no systematic phonological pattern explaining the attested but rare cases21 of total assimilation of nasal + occlusive before a vowel can be found and because (2) it leaves unexplained the anthroponyms Γυνόππασις and Θιοππάστoς.22 What about the second explanation? Méndez Dosuna acknowledges that it is the only hypothesis that would satisfactory explain all the four instances,23 but finds it otherwise unconvincing in the case of ἀππασάμενος, ἔππασις and τἀππάματα for three reasons: (1) there are no sure examples of the development *(-)kw- > -ππ- since the case of ἵππος is too problematic;24 (2) whatever the etymology may be, an intermediary stage *-μππ- is not likely because “comme la plupart des langues du monde, le grec ancien ne tolérait pas de géminées après une consonne”;25 (3) the explanation offered by *(‑)kw‑ > ‑ππ‑ is unsatisfactory from a semantic point of view, especially when it comes to reconstruct the contextual sense of ἔππασις and τἀππάματα (see below Sections 2.1–2.2).

 17 Cf. e.g. Thumb/Scherer 1959, 31 and 34; Lejeune 1972, 83–84 and 313 n. 1. 18 Schmidt 1889, 415. 19 Schulze 1895, 319 (who, however, following Cauer DIG no. 295 p. 204 ad loc. interpreted ταππαματα of IG VII 3172 as deriving from τὰ ππάματα, and not from *τἀναππάματα. It should be pointed out that Schulze, differently from Bechtel and Hermann, did not have at his disposal the evidence of P.Berol. 13284 ἀππασάμενος), followed by Bechtel 1921, 256; Hermann 1923, 83; Schwyzer 1939, 301; Buck 1955, 64 and 127; García Ramón 2000, 160; EDG s.v. πέπᾱμαι. 20 Hermann 1923, 83. 21 See the data provided by Méndez Dosuna 2007a, 299 n. 38. 22 Méndez Dosuna 2007a, 299–300. 23 García Ramón 2000, 160 with n. 21 observes, in my view correctly, that the gemination in the names Γυνόππασις and Θιόππαστoς cannot be explained away by invoking the category of “gémination expressive”, which in Boeotian is extremely rare and anyway occurs only in hypocoristic formations; see also Méndez Dosuna 2007a, 313. 24 Méndez Dosuna 2007a, 301 with footnotes 55–56. 25 Méndez Dosuna 2007a, 301 with n. 57.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  131

Méndez Dosuna advances then a third working hypothesis: the newly published lamellae from Dodona record several forms, verbal and nominal, pertaining to the πέπαμαι family; in particular, in CLOD 2028A we find the form ᾽πίπαμαι, that is, the subjunctive of the verbal compound ἐπιπάομαι (= Att. ἐπικτάομαι ‘to gain beside’, ‘to make an additional acquisition’), and in CLOD 430.4 the noun ἐπίπασις (= Att. ἐπίκτησις ‘additional acquisition’).26 On the basis of this new linguistic data Méndez Dosuna thus reinterprets (1) Boet. ἔππασις as an apocopated form of ἐπίπασις, with the corollary that ἔππασις is no longer a phonetic variant of ἔμπασις = ἔγκτησις (on which see below); (2) τἀππάματα as deriving from *ἀποπάματα and (3) ἀππασάμενος from *ἀποπασάμενος.27 However, Méndez Dosuna himself admits that his working hypothesis would not explain the anthroponyms Γυνόππασις and Θιοππάστoς.28 Before tackling directly Méndez Dosuna’s new hypothesis (apocope of ἐπί in ἔππασις and ἀπό in ἀππασάμενος and τἀππάματα) and its consequences, let us first counter his objections to the traditional derivation of ππ‑ from *kw‑. Regarding the development of *kw‑ > ππ-, absolute certainty cannot be obtained: ἵππος does indeed remain a notoriously problematic case because of the aspiration and the initial /i/ (the latter feature is usually explained either by the raising of *e in the proximity of a labial in Mycenaean or as a prop vowel supposedly arising in the cluster *h1kw‑).29 While skepticism may consequently seem justified, it is also the case that no better solution has been proposed to date.30

 26 Méndez Dosuna 2007a, 303 with n. 80 quotes also a second instance of ἐπίπασις in a still unpublished lamella of the Museum of Ioannina (MI3662A1). Cf. also the derivative ἐπιπάσιον in CLOD 4141Α ([πε]ρὶ τοῦ ἐπιπασίου ἐπ̣[ερωτῆι ὁ δεῖνα - - -]). On Μéndez Dosuna’s proposal of reading in CLOD 313A l. 4 ἐπίπ̣π̣ασις, that is, “a blend of ἐπίπασις and ἔππασις” instead of the transmitted ἐπιγγ[ύ]ασις (Méndez Dosuna 2016, 123–124 and Méndez Dosuna 2017, 290), see Appendix 2 below. 27 Méndez Dosuna 2007a, 304. The theoretical possibility of an apocope of ἀπό in (2) and (3) was already raised, in a non-committal way, by Meier-Brügger 1978, 228 n. 15 (acknowledged by Méndez Dosuna). 28 Méndez Dosuna 2007a, 313 tentatively suggests that the gemination ‑ππ‑ in these anthroponyms could be explained by analogy with an unattested *Ἔππαστος, that is, a Boeotian local variant with apocope of Ἐπίπαστος (equally unattested; cf. Att. Ἐπίκτητος). The weakness of this hypothesis is self-evident. 29 For the latter explanation, see de Vaan 2009. 30 See EDG s.v. ἵππος with recent bibliography. As for the alternative etymology for the root *πᾱ‑ proposed by van Beek 2016 (i.e. from *peh2‑s‑ ‘to pasture, to guard’, cf. n. 15 above), its main difficulty lies in the need to place the meaning ‘cattle, i.e. movable goods’ prior to that of ‘immovable goods’, against the surviving linguistic evidence.

132  Lucia Prauscello Regarding the plausibility of assuming an intermediary stage *‑μππ‑, I think that we can be more optimistic than Méndez Dosuna allows. Let us quote again in full Méndez Dosuna’s objection: On sait que, comme la plupart des langues du monde, le grec ancien ne tolérait pas de géminées après une consonne. A titre d’exemple on citera le futur *πενθσομαι > * πένσομαι > att. πείσομαι (non pas †πένσσομαι > †πέσσομαι > †πέσομαι) ou le traitement de gj dans crét. πράδδεν (*πρᾱγjº) vis-à-vis de σαλπίνδε[ν] (*σαλπιγγjº), non pas †σαλπίδδεν. (Méndez Dosuna 2007a, 301)

It is worth pointing out that all the examples quoted by Méndez Dosuna concern the gemination of consonants in word-internal position (Inlaut), and not at the beginning of a word (Anlaut). This difference is fundamental. In *πενθσομαι the loss of /th/ is very ancient and word-internal, like the loss of /s/ in *πέπρακ‑σθε > πέπραχθε. Furthermore, in *πενθσομαι we have three consonants, two of which belong to the root: but the third is a morphologically irreducible element of the conjugation. The same applies to the Cretan πράδδεν, which derives from *πραγ‑ + suffix *-jō. The case of *ἔμ‑ππασις, if we assume that the gemination ππ‑ derives from *kw‑, is totally different: the first consonant belongs to a preposition (in this case ἐν), that is, to an element that can come and go, appear or disappear and be substituted by other prepositions, whereas the two consonants ‑ππ‑ derive from the radical *kw‑ at the beginning of the word (that is, they are etymological). Compare, for example, in historical time, the case of σεύω and not *σσεύω in initial position (of verse or phrase), but obviously ἔσσευον, ἔσσυτο, ἐπισσεύεσθαι and, in particular, ἐπασσύτερος, already recognized as deriving from *ἐπ‑αν‑σσυτ- by Brugmann.31 The single π in the πέπᾱμαι family is most probably due to an analogical generalization of *kw‑ > *π‑ at word beginning, a possibility acknowledged by Méndez Dosuna himself.32

. ἔππασις = Att. ἐπίκτησις vs. ἔμπασις = Att. ἔγκτησις? Let us now have a closer look at some of the consequences of Méndez Dosuna’s proposal in terms of semantics for ἔππασις/ἔμπασις and τἀππάματα. A first notable consequence is that we would have to dissociate ἔππασις from ἔμπασις: the  31 Brugmann 1898. ἐξεσύθη at Il. 5.293 is a varia lectio attributed to Zenodotus vs. Aristarchus’ ἐξελύθη. The text is highly uncertain: see Kirk 1990, 89–90 and Schwyzer 1939, 319 ad loc. 32 See Schwyzer 1939, 649 and 654 ad loc. Cf. Méndez Dosuna 2007a, 302: “[b]ien entendu, l’argument n’est pas décisif, car le π‑ simple pourrait être dû à la generalisation du traitement *kw‑ > *π‑ au début du mot par nivellement analogique”.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  133

former would not be any more a phonetic variant of ἔμπασις = Att. ἔγκτησις but an apocopated variant of ἐπίπασις = Att. ἐπίκτησις.33 As we have already seen in Section 2, all of the 73 occurrences of ἔππασις in Boeotian inscriptions are in proxeny decrees, with no exception. The formulaic language of proxeny decrees is highly standardized across epigraphic corpora belonging to different dialects: together with asylia (immunity) or isoteleia (the privilege to pay taxes at a lower, citizen rate), the honorand (and at times his descendants) is granted the right of enktēsis, that is, the right to acquire property in a foreign land by incorporating it into his own patrimony.34 Why should one posit that in the overwhelming majority of Boeotian proxeny decrees (and only in these among all Greek epigraphic corpora, irrespective of the dialect) the term ἔππασις = Att. ἐπίκτησις (‘additional acquisition’) is used instead of the formulaic ἔμπασις = Att. ἔκγτησις, which is attested everywhere else?35 This is a highly uneconomical and unlikely guess, and above all one not supported by a linguistic or indeed semantic rationale. In fact, pace Méndez Dosuna,36 the ancient evidence in favour of an alleged synonymy between ἐπίκτησις and ἔγκτησις is illusory when put under closer scrutiny. In the pseudo-Demosthenic speech On Halonnesos 7.42 the distinction between ‘possession abroad’ and ‘possessions at home’ is expressed by the contrast between ἐγκτήματα and κτήματα.37 Diopeithes, an Athenian mercenary captain, had brought Attic settlers to the town of Cardia in 343 BC; since Cardia was unwilling to receive them, Philip immediately sent help to the town. The king proposed then to settle the dispute between the two cities by arbitration, but Athens scornfully refused. The passage which concerns us is the following:  33 Méndez Dosuna 2007a, 304 and more in detail Méndez Dosuna 2007b. Méndez Dosuna 2007b is of course right in saying that LSJ’s rendering of ἔγκτησις with ‘tenure of land in a country or district by a person not belonging to it’ is a contextual over-interpretation: strictly speaking, the prefix ἐν‑ indicates only that a property is ‘incorporated’ into pre-existing holdings. That said, the interpretation of ἔππασις = Att. ἐπίκτησις remains deeply problematic. 34 On the institution of proxenia and the right of enktēsis, see Mack 2015, in particular 122–127 (showing how a proxenos would normally remain resident and active in his own polis too: that is, he did not have to permanently migrate to the city of which he was proxenos to exercise his right of acquiring property in a foreign land). 35 Méndez Dosuna’s claim that “[l]a distribution des deux formes [i.e. ἔππασις and ἔμπασις] ne semble pas être significative ni du point de vue de la chronologie, ni du point de vue de la géographie” (Méndez Dosuna 2007a, 296) is imprecise (cf. already Knoepfler 2008, 618–619): of the 8 occurrences of ἔμπασις 6 are in Plataiai and Oropos, that is, in places most exposed to the influence of Attic and Western Ionic (for the case of Oropos, see Morpurgo Davies 1993); ἔμπασις as a ‘Boeotian’ form practically does not exist. 36 Méndez Dosuna 2007b, 34–35. 37 On the authenticity, already disputed in Antiquity, of oration 7 of the Demosthenic corpus, see MacDowell 2009, 344.

134  Lucia Prauscello (2)

ἑαυτῶν φασι τὴν χώραν οὖσαν οἰκεῖν καὶ οὐχ ὑμετέραν, καὶ τὰ μὲν ὑμέτερα εἶναι ἐγκτήματα ὡς ἐν ἀλλοτρίᾳ, τὰ δὲ ἑαυτῶν κτήμαθ᾿ ὡς ἐν οἰκείᾳ κτλ. ([Dem.] 7.42 Dilts) [The Cardians] say that the land they live in is not yours, but their own, and that while your possessions there are held by grace in a foreign country, theirs are their own property on their own soil. (transl. Vince 1954)

The scholium to this passage makes explicit the difference between ἐγκτήματα and κτήματα: (3) ἐγκτήματα] κτῆμα ἐγκτήματος διαφέρει. κτῆμα μὲν γάρ ἐστι κυρίως, ὅταν τις ἐν οἰκείᾳ γῇ κτήσηταί τι χωρίον, ἔγκτημα δὲ ὅταν ἐν ἀλλοτρίᾳ γῇ τις κτήσηταί τι χωρίον κτλ. (schol. [Dem.] 7.42 Dilts) enktēmata] the word ktēma differs from the word enktēma; for ktēma is used in its proper sense when someone acquires a property in his own country, whereas enktēma is used when someone acquires a property in a foreign land etc.

The same lexical distinction is found in Lex. Patm. p. 142 Sakkelion (ἐγκτήματα· τὰ ἐν τῇ ὑπερορίᾳ κτήματα) and in the Λέξεις ῥητορικαί ε 251 Bekker ([ἐγκτήματα] τὰ ἐν ἀλλοτρίᾳ γῇ χωρία καὶ περιουσία τις ἐγκτήματα λέγεται. τὰ δὲ ἐν τῇ πατρικῇ γῇ κτήματα λέγεται).38 Now, in [Ammon.] De adf. voc. diff. 282 p. 74 Nickau and in [Ptol.] De differ. vocab. p. 399 Heylbut we read a slightly different and probably corrupted text, that is, respectively: (4) κτήματα καὶ ἐ†πι†κτήματα διαφέρει. κτήματα μὲν γάρ ἐστι τὰ ἐν τῇ οἰκίᾳ, ἐ†πι†κτήματα δὲ τὰ ἐν ἀλλοτρίᾳ. ([Ammon.] De adf. voc. diff. 282 p. 74 Nickau) The word ktēmata differs from e†pi†ktēmata. For ktēmata are the properties one has in his own country, whereas e†pi†ktēmata those in a foreign country. (5)

κτήματα καὶ ἐπικτήματα διαφέρει· τὰ μὴ (g: τὰ μὲν OV) ἡμέτερα ἐπικτήματα (g: εἶναι κτήμα OV) ὡς ἐν ἀλλοτρίᾳ, τὰ δὲ ἑαυτοῦ κτήματα ὡς ἐν οἰκείᾳ (Duker: ἐνοίκια Og: ἐν οἰκίαι V). ([Ptol.] De differ. vocab. p. 399 Heylbut)

 38 Cf. also Phot. ε 51 Theodoridis ἐγκτήματα· τὰ κτήματα τὰ ἐν ἀλλοτρίᾳ, οἷον ὅταν Ἀθηναῖος ἐν Θετταλίᾳ ἔχῃ κτήματα (not quoted by Méndez Dosuna 2007b).

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  135

The word ktēmata differs from epiktēmata. Epiktēmata is a property that is not ours [sic], as if it were in a foreign country, whereas ktēmata is one’s own property as if it were in his own country.

Both Pseudo-Ammonius and Pseudo-Ptolemy derive from the same source (probably a lost lexicon by Herennius Philo, 2nd century AD).39 Already the 17th-century philologist Henri de Valois recognized that the passage in Pseudo-Ammonius is corrupted40 and that instead of ἐπικτήματα one should read ἐγκτήματα, as faithfully registered in the apparatus by Nickau (“ἐγκτήματα Henr. Valesius recte”), and the same can be said for the Pseudo-Ptolemy passage.41 Likewise, the same corruption can be seen also in another lexicographical fragment not quoted by Méndez Dosuna: (6) κτήματα ἐπικτημάτων διαφέρει. κτήματα μὲν τὰ ἐν τῇ οἰκίᾳ, ἐπικτήματα δὲ τὰ ἐν ἀλλοτρίᾳ. δηλοῖ δὲ καὶ †ἐν τῷ κατὰ ἀλλονήσω Φιλόξενος† τὰ ἐν ἀλλοτρίᾳ γῇ ἐπικτήματα λέγων, τὰ δὲ ἐν τῇ πατρικῇ γῇ κτήματα. ([Philox.] fr. 690** Theodoridis = Symeonis Synagoge s.v. κτήματα 421 Νickau)42 The word ktēmata differs from epiktēmata. ktēmata are the properties in one’s own country, epiktēmata those owned in a foreign country. This is shown also †by Philoxenus in the speech on Halonnesοs†, where he calls epiktēmata possessions in a foreign country and ktēmata those in one own’s fatherland.

Theodoridis correctly observes in the apparatus that one should write not ἐπικτήμα(τα) but ἐγκτήμα(τα). Yet, against the evidence, Méndez Dosuna refuses to draw what is the most logical inference, namely that the passages of PseudoAmmonius and Pseudo-Ptolemy, both dependent on [Dem.] 7.42, have undergone a textual corruption easily explainable on palaeographical grounds and instead argues that “estos testimonios alimentan la sospecha de que alguna copia

 39 See Dickey 2007, 94–95. 40 Palaeographically a majuscule corruption ΕΓΚ > EΠΙΚ is relatively easy to account for, especially if one considers that Γ and Π are often confused and that K could be confused with ΙC. 41 Méndez Dosuna does quote the text of Pseudo-Ammonius without recording that the transmitted text is suspicious. Nor is it clear to me from which edition of Pseudo-Ammonius he is quoting (in Méndez Dosuna 2007b, 34 the only indication one finds is “Ammon. p. 83”). 42 See Nickau 2000, 25.

136  Lucia Prauscello antigua del Sobre el Haloneso incluyese una variante ἐπικτήματα que no ha sobrevivido en ninguno de los manuscritos conservados”.43 This seems to me an unlikely and counterintuitive way to account for the available data. To sum up, in our ancient sources there is no reliable evidence for an alleged synonymity of ἐπίκτημα/ἐπίκτησις and ἔγκτημα/ἔγκτησις, nor should this be posited for the corpus of the extant Boeotian proxeny decrees.

. τἀππάματα in IG VII 3172 IG VII 3172 is a long and complex dossier of financial documents, dating to ca. 223 BC: Nicareta, a well-off woman from Thespiae, negotiates with the city of Orchomenos the restitution of a money loan, made in several instalments, to that same city.44 Nicareta claims to have repeatedly lent money to the Orchomenians on various occasions in the past, and that every time the loan was due she has not been repaid. This is the section of the inscription which concerns us:45 (7)

Ξενοκρίτω, Ἀλαλκομενίω Νικαρέτα Θέωνος τᾶς π[ό]λιος Ἐρχομενίων κὴ τῶ ἐγγύω Θίωνος Συννόμω· τἀππάματα μούριη ὀγδοείκοντα πέντε δίουο ὀβολίω κὴ τῶ τεθμίω ϝίστωρ Ἀριστόνικος Πραξιτέλιος. Λιουκίσκω, Θιουίω τὸ σουνάλλαγμα· Νικαρέτα Θίωνος τᾶς πόλιος Ἐρχομενίων κὴ τῶ ἐγγούω Θίωνος Σουννόμω τἀππάματα δισχείλιη πεντακάτιη̣ κὴ τῶ τεθμίω ϝίστωρ ὁ αὐτός. Λιουκίσκω, Ὁμολωΐω· [τ]ὸ σουνάλλαγμα· Νικαρέτα Θίωνος τᾶς πόλι[ος] Ἐρχομενίων κὴ τῶ ἐγγούω Θίωνος Σουννόμω τἀππάματα πετρακισχείλιη κὴ τῶ τεθμίω ϝίστωρ ὁ αὐτός. χρόνος ὁ αὐτός· Νικαρέτα Θίωνος τᾶς πόλιος [Ἐ]ρχομενίων κὴ τῶ ἐγγούω Θίωνος Σουννόμω τἀππάματα χείλιη κὴ τῶ τεθμίω ϝίστωρ ὁ αὐτός.

65

70

(IG VII 3172 IV (H) ll. 61–75)

 43 Méndez Dosuna 2007b, 35. 44 Nicareta’s inscription has been the object of increasing attention in contemporary scholarship as both an example of specifically female authority over property at large (see e.g. Sealey 1990, 89–90) and as a case-study for the importance of wealth in the formation of elites in Boeotian cities in the Hellenistic period (see e.g. Müller 2010, 236–240 and Müller 2011, 355–356 and annex 5). 45 Text and translation are after Migeotte 1984, no. 13.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  137

Under Xenocritos in the month Alalkomenios: Nicareta, daughter of Theon, against the city of Orchomenos and its guarantor Thion, son of Synnomos; τἀππάματα: 10,085 drachmai and 2 oboloi; and witness of the agreement: Aristonicos, son of Praxiteles. Under Lioukiskos, the obligation of the month of Thiouios: Nicareta, daughter of Theon against the city of Orchomenos and its guarantor Thion, son of Synnomos; τἀππάματα: 2,500 drachmai; and witness of the agreement: the same. Under Lioukiskos, the obligation of the month of Homoloios: Nicareta, daughter of Theon, against the city of Orchomenos and its guarantor Thion, son of Synnomos; τἀππάματα: 4,000 drachmai; and witness of the agreement: the same. Same date: Nicareta, daughter of Theon against the city of Orchomenos and its guarantor Thion, son of Synnomos; τἀππάματα: 1,000 drachmai; and witness of the agreement: the same.

The word ταππαματα has a long exegetical history:46 interpreted as τὰ ἐππάματα < *τὰ ἐν‑πάματα = Att. ἐγκτήματα by the first editor (Foucart), it was then re-analyzed as τὰ ππάματα = Att. τὰ κτήματα by Cauer and eventually understood as τὰ ἀππάματα < ἀναπάματα by Teyssier on the basis of ἀππασάμενος of P.Berol.47 Cauer’s interpretation τὰ ππάματα = Att. τὰ κτήματα clearly does not suit the context: as argued by Teyssier, κτήματα refers to immovable goods or cattle, whereas here, if anything else, we would expect the Boeotian χρείματα = Att. χρήματα, given the nature of the loan. Teyssier (1940, 41) then argued that a deverbal noun from *ἀναπάομαι = Att. ἀνακτάομαι is what suits the context best. Nicareta is claiming back what is her due: “puisque ἀππασάμενος signifie ‘ayant recouvré’, τἀππάματα peut signifier fort bien ‘les sommes en recouvrement’”, that is, ‘the debts to be recovered’ (LSJ). Teyssier’s explanation has generally been accepted by all subsequent editors and commentators. Méndez Dosuna has objected that, while ‘the debts to be recovered’ suits perfectly the context of Nicareta’s inscription, this meaning cannot be covered by τἀππάματα (assuming that it derives from *ἀναπάομαι = Att. ἀνακτάομαι), since noun formations in -μα usually express the result of an action: hence τἀππάματα can only mean “propriétés recouvrées, argent recouvré”, which is clearly unsuitable to the context given that Nicareta is protesting because her loan has not been repaid. To indicate ‘the debts to be recovered’ we would need either *τὰ ἀναπάμενα (= Att. ἀνακτώμενα, “les sommes qu’on est en train de recouvrer”) or rather *τὰ ἀναπα(σ)τέα (= Att. ἀνακτητέα, “les sommes qu’on doit recouvrer”).48

 46 See Teyssier 1940, 139–140 for a survey. 47 Note however that Teyssier 1940, 140 was wrong in interpreting ‑ππ‑ as the result of a complete assimilation of the nasal to the following occlusive (see above Section 2). 48 Méndez Dosuna 2007a, 309.

138  Lucia Prauscello Consequently, Méndez Dosuna suggests, as we have seen, deriving τἀππάματα from *ἀποπάματα = Att. ἀποκτήματα: since ἀποκτάομαι signifies ‘to lose possession of, alienate’ (cf. LSJ s.v. and Hsch. α 6415 Latte/Cunningham ἀποκτᾶσθαι· τὸ ἀποβάλλεσθαι, on which see below), τἀππάματα < *ἀποπάματα would mean “les biens [que Nicaréta a] cédés, les sommes d’argent [qu’elle a] prêtés”. At first sight this seems a brilliant solution, but does it work? Can τὰ ἀποκτήματα mean “les biens [que Nicaréta a] cédés, les sommes d’argent [qu’elle a] prêtés”? A closer look at the semantic range of ἀποκτάομαι and ἀπόκτησις and at their use in specifically financial/legal documents suggests otherwise.49 Both terms appear quite late in our extant texts. A TLG search reveals that in the Greek literary texts the noun ἀπόκτησις occurs fairly infrequently (only 49 times), its earliest occurrence being in the 2nd-century AD philosopher Secundus (Sent. 19 θάνατος ἀπόκτησις βίου). The verb makes its first recorded appearance in literary texts in Cassius Dio Hist. Rom. (Joannis Antiocheni excerpta e Dione derivata) 752.12 Boissevain = fr. 162.1.2–3 Roberto ὅσα γὰρ τῷ Αὐγούστῳ ἐν πολλῷ χρόνῳ περιεποιήθη, οὗτος ἐν μιᾷ ἡμέρᾳ ἀπεκτήσατο (said of Gaius Tiberius).50 If we look at inscriptions (PHI) and documentary papyri (DDbDP), things become more interesting. PHI records one instance of ἀπόκτησις in IEleusis 300.65– 67 (= SEG 30.93 = Clinton 1974, 50 no. D14), a decree honouring the dadouchos Themistocles dated to 20/19 BC: ἀλλὰ καὶ τῆ̣[ς] | περ[ὶ] αὐτὸν φιλοτ⟦ε⟧ιμίας εἰς τὴν τῶν ἐκλελειμμένων πατρίων ἀ[πό]|κτησιν. As I argue elsewhere, this occurrence should be discarded: we should supply at ll. 66–67 ἀ[νά]|κτησιν (and not ἀ[πό]|κτησιν) since Themistocles’ philotimia manifests itself in his zeal for the ‘recovery’ of ancient discarded traditions, rather than for their ‘loss’.51 More relevant is what we find in the documentary papyri: ἀπόκτησις occurs in four documents (several times within the same text with the exception of P.Grenf. 2.70 where it occurs only once): 6x in P.Sorb. 2.69 (= TM 20110; 7th century AD, Hermopolis), 3x in P.Lond. 5.1761 (TM 36936; 6th century AD? Hermopolites?), 3x (or perhaps 4) in P.Würzb. 19 (= TM 17393; 7th century AD, Hermopolis), 1x in P.Grenf. 2.70.25 (= TM 22627; 3rd century AD, from Kysis; the reading, though, is uncertain).52 In all these documents ἀπόκτησις refers to an act of permanent dispossession, that is, a donation or, as LSJ s.v. have it, a ‘deed of gift’ (in P.Grenf. 2.70.3 the act of

 49 ἀπόκτημα, like ἀνάκτημα, is not attested but both are perfectly plausible morphological formations, as already acknowledged by Méndez Dosuna 2007a, 309. 50 Roberto 2012, 617 glosses ἀποκτάομαι with “a dominio meo amitto, vendo”. 51 See Prauscello (forthcoming). 52 Crönert in Mitteis/Wilcken 1912, 210 suggested as alternative reading ἀπὸ Κύσ(εως), but cf. Preisigke 1915, vol. 1, 183 s.v.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  139

patrimonial dispossession is called χάρις). As Gascou puts it, ἀπόκτησις is a technical fiscal term with the function of “un onoma, un compte de titulaire d’impôt”; it refers to a “‘cession’ (onéreuse ou gracieuse), ou ‘abandon’ d’un bien”.53 The same meaning is implied in Hsch. α 6415 Latte/Cunningham ἀποκτᾶσθαι· τὸ ἀποβάλλεσθαι, a quotation from Greg. Naz. Or. 40.32 (= 36.404C Migne) μὴ τὸ καλῶς κτᾶσθαι τυπούτω μόνον, ἀλλὰ καὶ τὸ καλῶς ἀποκτᾶσθαι, where Gregory exhorts the newly baptized not only to make good use of their financial resources but also to give them away ‘well’, that is, he encourages them towards ascetic dispossession by donating to the poor.54 In other words, the evidence collected so far suggests that what is ‘alienated’ or ‘dispossessed’ via an act of ἀπόκτησις cannot be at the same time also reclaimed. Nicareta’s contribution is a loan and she is asking for the ἀπόδοσις (restitution) of that loan:55 what she claims back cannot qualify as an ἀπόκτημα, at least not according to our (admittedly much later) sources. ‘To give as a loan, to lend’ and ‘to give away’, ‘to alienate’ are not the same thing: what is given away, what has been ‘alienated’, cannot be claimed back. What is then the sense of τἀππάματα in Nicareta’s inscription? Méndez Dosuna objects to Teyssier’s explanation because of the suffix ‑μα, which usually indicates the result of an action.56 This is indeed usually the case, but it has long been recognized that there are exceptions and that the boundary between ‑σις (indicating the process) and ‑μα (indicating the result) formations is more blurred than one would think.57 Already Chantraine observed, for example, that in tragedy nouns in ‑μα deriving from verbal stems “peuvent exprimer un état, mais en réalité ils constituent des derives verbaux de sens assez flou qui peuvent être voisins des abstraits en ‑σις”, quoting as example σπάραγμα in Soph. Ant. 1081 (‘mutilated corpse’) and Eur. Bacch. 739 (the act of tearing apart). This functional overlap between ‑σις and ‑μα nouns is a feature of the Greek language already in Homer: ‑σις nouns can sometimes denote a completed action, that is, its result: cf. e.g. δόσις in Il. 10.213 καί οἱ δόσις ἔσσεται ἐσθλή ‘and a noble gift will be his’,  53 Gascou 1994, 42–43. Pace Knoepfler 2008, 619, according to whom Méndez-Dosuna 2007a “montre que ce terme technique ne signifie sans doute pas, en dépit de l’apparence, ‘les sommes en recouvrement’, mais les ‘propriétés cédées ou l’argent prêté (par Nikaréta)’ (*apopamata, équivalent de *apoktemata). Ce serait ainsi une contribution notable de la dialectologie à l’histoire du vocabulaire juridique”. 54 On the debate on ascetic dispossession in the early Church, see Finn 2006. 55 Cf. IG VII 3172 I D ll. 5–6 τὰν ἀπόδοσιν τῶν δα|νείων, ΙΙΙ F l. 6 πραττώσας τὸν δάνειον, VI A l. 1 ἐδάνεισεν, l. 8 εἰς ἔκτεισιν τοῦ δανείου, ll. 21–22 ἀποδότωσαν δὲ τὸ δάνειον| οἱ δανεισάμενοι etc. 56 Cf. e.g. Debrunner 1917, 156–157. 57 See Chantraine 1933, 175–190; Porzig 1942, 240–243 and 278–282.

140  Lucia Prauscello where δόσις means ‘the thing given’, not the act of giving (cf. also schol. D Il. 10.213/Zs van Thiel δόσις· δωρεά), or κτῆσις in Od. 14.62 καὶ κτῆσιν ὄπασσεν ‘and has given (me) possessions’, where κτῆσις refers to ‘something possessed’, not to the process of acquiring.58 The specific function of ‑μα suffix to form nomina actionis was probably a later development in Attic-Ionic. Literary language is of course entitled to a higher degree of creativity than everyday speech; yet, even with this important caveat in mind, the possibility of functional overlap between the two suffixes remains a fact intrinsic to the Greek language. If we adopt this perspective, Teyssier’s “les sommes en recouvrement” is not implausible: if it is only the context that can help us to decide whether the sense of a particular word is active (the act of doing) or passive (the thing done), in Nicareta’s inscription τἀππάματα might indeed refer to the act of reclaiming itself, rather than to its result.

 Orion’s Re-conquest Let us now take stock of what has been argued so far. Méndez Dosuna’s explanation of the gemination of -ππ- in Boeotian as deriving from the apocope of a preposition (ἐπί in ἔππασις; ἀπό in τἀππάματα) rather than from an original *kw not only leaves the anthroponyms Γυνόππασις and Θιοππάστoς unexplained but is highly problematic for ἔππασις and τἀππάματα from a semantic point of view as well. Finally, let us go back to our initial passage, Corinna PMG 654 col. iii l. 39 γῆα[ν ϝ]ὰν ἀππασάμενος. Méndez Dosuna’s objection to the traditional interpretation of ἀππασάμενος = Att. ἀνακτησάμενος59 is twofold:60 (1) we would be dealing with an otherwise unattested variant of the myth (Orion’s re-conquest); (2) Orion’s recovery by conquest of his own land seems to presuppose an antagonistic scenario father (Hyrieus) vs. son (Orion) which seems unsuitable, especially if one considers that Corinna held Orion ‘most pious’ (cf. PMG 673 Κόριννα δὲ εὐσεβέστατον λέγει αὐτόν). Consequently, Méndez Dosuna suggests interpreting ἀππασάμενος = ἀποκτησάμενος: “le participle ἀππασάμενος ‘après avoir renoncé à sa terre’, (c.-à.-d. ‘après avoir quitté sa terre’) peut avoir trait soit au périple qu’Orion a fait comme chasseur dans diverses regions de Grèce, soit à son

 58 Instances are taken from Long 1968, 16 and 18–19. 59 Recently accepted by Vottéro 2012, 139. 60 Méndez Dosuna 2007a, 311–312.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  141

καταστερισμός, sa transformation en constellation, qui est mentionnée au vers suivant”.61 As to (1), mythologically unique variants — be they authorial innovations or inherited pieces of (for us) unattested local lore — are not at all uncommon in Corinna’s poetry, and the uniqueness of Corinna’s tale about Orion should not be considered per se as a suspicious feature:62 one can hardly help noticing that Méndez Dosuna’s alternative interpretation (Orion renouncing his own land) would be equally open to the same argumentum ex silentio. The re-appropriation, on the part of Orion, of his own land probably played an important part also in Corinna’s reshaping of local genealogies and claims to control over the oracular site, wherever the latter might have been located. We know that Corinna wrote a Cataplous (PMG 662, 663) which most probably featured Orion’s epidēmia around the Greek world, purifying the land from wild beasts (PMG 673 = schol. Nicand. Ther. 15 p. 41 Crugnola αὐτὸν καὶ ἀπελθόντα63 πολλοὺς τόπους ἡμερῶσαι καὶ καθαρίσαι ἀπὸ θηρίων) and conquering various places (PMG 662).64 As Wilamowitz already saw,65 an Orion ‘re-conquering his own land’ is quite at home in the context of a nostos: καταπλέω can of course mean both ‘sailing downstream’ and ‘sailing back’, yet the circumstantial pieces of evidence so far gathered strongly hint that in the case of Orion it was a homecoming voyage. The objection against possible intra-familial father vs. son conflict is equally indecisive. It is true that a violent pattern of succession is established by Hyrieus who expels by force Euonymus (PMG 654 col. iii l. 35 τὸν δ᾽ ἐς γᾶς βαλὼν Οὐριεύς), yet this might not apply to Orion as well: as Méndez Dosuna himself admits, it is quite possible to imagine a scenario by which “Orion aurait récupéré son territoire après avoir expulsé un autre intrus qui aurait usurpé le poste de son père”.66 But there is also another difficulty with Méndez Dosuna’s interpretation: it presupposes that in order to become the incumbent prophet of the sanctuary

 61 Méndez Dosuna 2007a, 312. 62 On Corinna and mythological innovation, see Collins 2006 with further bibliography. 63 I follow Burzacchini 1978–1979 in retaining the MSS ἀπελθόντα vs. Schneider’s unnecessary emendation ἐπελθόντα. 64 On Orion’s civilizing wanderings in panhellenic and local tradition(s) see Debiasi 2010a and 2010b. 65 Wilamowitz 1907, 52–53. 66 Méndez Dosuna 2007a, 311.

142  Lucia Prauscello Orion had to renounce his political power. That is, Méndez Dosuna’s interpretation assumes a potential incompatibility between political/royal and prophetic power, which is indeed not warranted by the ancient evidence.67 To conclude: our evidence is indeed partial and every conclusion must therefore remain provisional, yet on the basis of the data collected here it seems reasonable to infer that (1) the traditional interpretation of the gemination ‑ππ‑ in Boeotian as the reflex of *kw- before vowel is still to be preferred on both linguistic and semantic grounds; (2) in PMG 654 Orion the prophet most probably did gain control of his own land again after his καταπλοῦς and his role as μάντις at an unidentified Boeotian oracular site does not presuppose the dispossession/renunciation of his political power.

Appendix 1: Occurrences of ἔππασις/ἔμπασις in Boeotian Inscriptions (A) ἔμπασις/ἔνπασις 8x: Coronea (1x): IG VII 2866; Onchestos (1x): SEG 32.476; Oropos (4x): IG VII 280 = IOropos 49; IG VII 4261 = IOropos 34; SEG 1.109 = IOropos 35; SEG 1.115 = IOropos 74; Plataea (2x): IG VII 1664; IG VII 1655. (B) ἔππασις 73x: Acraiphia (3x): IG VII 2709; BCH 23 (1899), 90–91, no. i and ii (2x); Chorsia (1x): IG VII 2388; Coronea (8x): IG VII 2858; IG VII 2859; IG VII 2860; IG VII 2861 = SEG 23.293; SEG 23.286 = REA 68 (1966), 61–66, no. 1; SEG 23.287 = REA 68 (1966), 61– 66, no. 2; SEG 23.289 = REA 68 (1966), 61–66, no. 4; SEG 23.292 = REA 68 (1966), 61–66, no. 7; Haliartos (2x): IG VII 2849 = SEG 44.409 = Fossey 2013, no. 6 = SEG 44.409, b; SEG 28.453 = Fossey 2013, no. 3;68 Orchomenos (1x): IG VII 3166; Oropos (11x): IOropos 32; IOropos 33; IOropos 37 = IG VII 4260; IOropos 43 = IG VII 4259; IOropos 44 = IG VII 393; IOropos 46 = ArchEph (1919), 76, no. 109; IOropos 92 = IG  Like Méndez Dosuna 2007a, 295 I have recorded only the forms which are unambiguous in their spelling and have omitted forms wholly restored.  67 In archaic times, more often than not, the status of seer was connected with that of king: cf. Fowler 2008, 4; Bremmer 2010, 15–16. The alleged case of Melampus in Pind. Pae. 4.28–30 S.-M., pace Κäppel 1992 ad loc., is no exception: see D’Alessio 1994, 64. 68 Cf. Fossey 2013, 35: “[o]verall my impression is that this text is not a proxenia decree by the Haliartans but the local copy of one passed for a Haliartan citizen by some other Greek state”.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  143

VII 352; IOropos 93 = ArchEph (1892), 35, no. 64 and ArchEph (1919), 82, no. 119; IOropos 188 = IG VII 283; IOropos 194 = ArchEph (1919), 78, no. 111; IOropos 195= ArchEph (1919), 76, no. 108. Tanagra (17x): IG VII 504 = Fossey 1991, no. 16; IG VII 505 = Fossey 1991, no. 18; IG VII 506 = Fossey 1991, no. 17; IG VII 507 = Fossey 1991, no. 19; IG VII 508 = Fossey 1991, no. 20; IG VII 510 = Fossey 1991, no. 21; IG VII 511 = Fossey 1991, no. 37; IG VII 513 = Fossey 1991, no. 26; IG VII 516 = Fossey 1991, no. 23; IG VII 517 = Fossey 1991, no. 27; IG VII 518 = Fossey 1991, no. 29; IG VII 519 = Fossey 1991, no. 24; IG VII 520 = Fossey 1991, no. 25; IG VII 522 = Fossey 1991, no. 34; IG VII 523 = Fossey 1991, no. 32; IG VII 524 = Fossey 1991, no. 33; IG VII 525 = Fossey 1991, no. 30; Thebes (4x): IG VII 2409; BCH 94 (1970), 140.1(a) + (b) (2x); SEG 34.355 (1) = REG 97 (1984), 45–60; Thespiae (19x): IG VII 1721 = IThesp 2; IG VII 1722 = IThesp 19; IG VII 1723 = IThesp 10; IG VII 1724 = IThesp 11; IG VII 1725 = IThesp 18; IG VII 1726 = IThesp 22; IG VII 1727 = IThesp 7; IG VII 1728 = IThesp 12; IG VII 1731= IThesp 17; IThesp 3 = ArchEph (1936), 39, no. 211; IThesp 4 = ArchEph (1936), 39, no. 212; IThesp 5 = ArchEph (1936), 39, no. 213; IThesp 6 = ArchEph (1936), 40, no. 214B; IThesp 8 = ArchEph (1936), 40, no. 214A; IThesp 9 = RA 32 (1948), 826, no. 3; IThesp 15 = RA 32 (1948), 825, no. 1; IThesp 16 = RA 32 (1948), 825, no. 2; IThesp 20 = RA 32 (1948), 828, no. 5; IThesp 29 = SEG 32.496 (I have excluded IThesp 31 ἔπ̣π̣ασιν because the reading is uncertain);69 Thisbe (7x): IG VII 2223; IG VII 2224; SEG 3.343 = AD 8 (1923), 204–207, no. 2; SEG 3.344 = AD 8 (1923), 204–207, no. 3; SEG 3.345 = AD 8 (1923), 207, no. 4; SEG 3.348 = AD 8 (1923), 208, no. 7; SEG 3.349 = AD 8 (1923), 210, no. 8.

Appendix 2: CLOD 313A l. 4 ἐπιγγ[ύ]ασις? CLOD 313A (= Eidinow 2007, 92 no. 13), dated on palaeographical grounds to ca. 450–400 BC, contains the oracle’s answer to the question posed by Boukolos and Polymnaste,70 a couple of Boeotian provenance, in search of a male offspring who

 69 The inscription is engraved on a very damaged stone. The text published in IThesp is based on a calque. For the interpretation, cf. IThesp. ad loc.: “[l]es l. 5–6 donnent la formule de la fin habituelle des proxénies, mais les vestiges ne permettent pas de reconnaitre en les l. 3–4 la formulaire d’un décret de proxène courant. [Évidemment il y est] question d’argent, soit de versement annuel, ou de consécration. [Mais il n’y a] pas de texte parallèle”. 70 Méndez Dosuna is right in interpreting Βο̄κόλο̄ κὴ Πολυμνάστη at l. 1 as datives (and not nominatives: cf. Eidinow 2007, 92 and Chaniotis in SEG 57.536.4) and in suggesting that the omission of final iota by the stone-cutter cannot be ruled out in the case of Βο̄κόλο (Méndez Dosuna 2016, 123; Méndez Dosuna 2017, 289–290).

144  Lucia Prauscello would reach adult age so that there would be prosperity and χρε̄ μάτων ἐπιγγ[ύ]ασις (ll. 3–4). The text in its entirety, as printed in CLOD, is the following: Θιός· τύχα ἀγαθά· ⁝ Βο̄κόλο̄ κὴ Πολυμνάστη τί κα δράοντοιν hυγία κὴ γενιὰ κ᾿ ἀνδρογένεια γινύο[ι]το κ̣ὴ παραμόνιμος ἰοιὸ[ς] κ̣ὴ χρε̄μάτων ἐπιγγ[ύ]ασις κ̣ὴ τῶν ἰόντων ὄνασις. God, good luck. To Boukolos and Polymnaste (asking) what they should do for there to be health, descent, male offspring71 and a male child that will reach adulthood and the security of their patrimony and [how] what they have shall prosper.72

While the general thrust of the request is clear, interpreters have differed as to the precise reading and meaning of χρε̄ μάτων ἐπιγγ[ύ]ασις (ll. 3–4). According to the editors of CLOD, ἐπιγγ[ύ]ασις, a hapax legomenon, would mean ‘security’: a male heir would assure the couple that their patrimony (χρε̄ μάτων) would not be alienated or taken away from them.73 Méndez Dosuna has objected to the reading ἐπιγγ[ύ]ασις for two reasons: (1) although the fact that an original *ἐπεγγύησις would be a hapax is not problematic (the verb ἐπεγγυάω is well-attested;74 a noun ἐπεγγύησις would be unobjectionable morphologically), we would have to posit the raising of /e/ before a nasal, a feature otherwise unattested in Boeotian (Méndez Dosuna 2016, 123); (2) the facsimile has ΕΠΙΓΓΑΣΙΣ, where “the fifth letter is flanked by two short blanks” (Méndez Dosuna 2016, 124): the space is not sufficient for positing [Υ] between the second Γ and Α. Hence, according to Méndez Dosuna, the form ἐπιγγ[ύ]ασις is “a ghost”: one should instead read ἐπίπ̣π̣ασις, “a blend of ἐπίπασις and ἔππασις”.75 The raising of /e/ before a nasal to /i/ is a substantive objection and cannot be accounted for unless one posits a slip by the stone-cutter. As for (2), it seems much more economical, from a palaeographical point of view, to posit that the stone cutter wrote ΓΓ instead of single, and not double Π (I owe this suggestion to Giambattista D’Alessio): that is, one should read ἐπίπ̣ασις = Att. ἐπίκτησις and not a hybrid form ἐπίπ̣π̣ᾱσις.  71 I follow Méndez Dosuna 2016, 123 in understanding ἀνδρογένεια at l. 2 as ‘male offspring’. 72 On the correct meaning of τῶν ἰόντων ὄνασις at l. 4 see Chaniotis in SEG 57.536.4. 73 CLOD I 109, esp. ‘το ζευγάρι συνεχίζει μνημονεύοντας την εξασφάλιση και την απολαβή της περιουσίας τους: ἐπιγγύασις, ἰόντων ὄνασις’. 74 Cf. Lys. 10.17 (ἐπεγγυᾶν: where the speaker quotes the text of a law) and Sophr. fr. 60 Hordern (ἐπεγγυάμενοι). Inscriptions: ἐπενεγύων in FD III.2, 70 l. 43 (112/111 BC) and Robert 1945, no. 45 (τοῦ] δὲ βλάβους ἐπεγγυ [—; Hellenistic date; this last occurrence is not recorded in LSJ). 75 Méndez Dosuna 2016, 124 following a suggestion by Alcorac Alonso Déniz; cf. also Méndez Dosuna 2017, 290: “[t]he form ἐπίπ̣π̣ᾱσις in the same inscription may be a blend of non-apocopated ἐπίπᾱσις (= Att. ἐπίκτησις) and apocopated ἔππᾱσις”.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  145

Bibliography Bechtel, F.-F. (1921), Die griechischen Dialekte, vol. 1: Der lesbische, thessalische, böotische, arkadische und kyprische Dialekt, Berlin. Bremmer, J.N. (1996), “The Status and Symbolic Capital of the Seer”, in: R. Hägg (ed.), The Role of Religion in the Early Greek Polis, Stockholm, 97–109. Brugmann, K. (1898), “Ἐπασσύτερος”, in: RhM 53, 630–633. Buck, C.D. (1955), The Greek Dialects, 2nd edn., Chicago. Burzacchini, G. (1978–1979), “Corinn. fr. 20 P”, in: MCr 13–14, 147–148. Chantraine, P. (1933), La formation des noms en grec ancien, Paris. Clinton, K. (1974), The Sacred Officials of the Eleusinian Mysteries, Philadelphia. Crönert, W. (1908), “Corinnae quae supersunt”, in: RhM 63, 161–189. D’Alessio, G. (1994), “Review of L. Käppel, Paian: Studien zur Geschichte einer Gattung (Berlin 1992)”, in: CR 44, 62–65. D’Alessio, G. (2005), “The Megalai Ehoiai: A Survey of the Fragments”, in: R. Hunter (ed.), The Hesiodic Catalogue of Women: Constructions and Re-constructions, Cambridge, 176–214. Debiasi, A. (2010a), “OΥΡION/ΩΡION: Frammenti dall’Esiodo di Euforione? (schol. Basil. ad German. Arat. p. 93, 13–20 Breysig e fr. 101 Pow.)”, in: QUCC 94, 99–119. Debiasi, A. (2010b), “Orione al Peloro (Diodoro IV 85, 5 = Esiodo fr. 149 M.-W.)”, in: L. Braccesi/ F. Raviola/G. Sassatelli (eds.), Studi sulla grecità d’occidente, Roma, 9–28. Debrunner, A. (1917), Griechische Wortbildungslehre, Heidelberg. de Vaan, M. (2009), “The Derivational History of Greek ἵππος and ἱππεύς”, in: JIES 37, 198–213. Dickey, E. (2007), Ancient Greek Scholarship: A Guide to Finding, Reading, and Understanding Scholia, Commentaries, Lexica, and Grammatical Treatises, from Their Beginnings to the Byzantine Period, New York/Oxford. Dunn, F.M. (1995), “Pausanias on the Tomb of Medea’s Children”, in: Mnemosyne 48, 348–351. Eidinow, E. (2007), Oracles, Curses and Risk among the Ancient Greeks, Oxford. Finn, R. (2006), Almsgiving in the Later Roman Empire: Christian Promotion and Practice 313– 450, Oxford. Fossey, J.M. (1991), Epigraphica Boeotica, vol. 1: Studies in Boiotian Inscriptions, Amsterdam. Fossey, J.M. (2013), Epigraphica Boeotica, vol. 2: Further Studies on Boiotian Inscriptions, Leiden/Boston. Fowler, M.A. (2008), The Seer in Ancient Greece, Berkeley/Los Angeles. García Ramón, J.L. (2000), “Mycénien qa‑sa‑ko /Kwās‑arkhos/, grec alphabétique Пάσαρχος, Kτήσαρχος et le dossier de *ku̯ ā(s)‑ dans la langue des tablettes”, in: L. Dubois/E. Masson (eds.), Philokypros: Mélanges de philologie et d’antiquités grecques et proche-orientales dédiés à la mémoire d’Olivier Masson, Salamanca, 153–176. García Ramón, J.L. (2017), “Ancient Greek Dialectology: Old and New Questions, Recent Developments”, in: E. Crespo/G.G. Giannakis/P. Filos (eds.), Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, Berlin/Boston, 29–106. Gascou, J. (1994), Un codex fiscal hermopolite (P.Sorb. II 69), Atlanta. Gentili, B./Lomiento, L. (2001), “Corinna, Le Asopidi (PMG 654 col. iii 12–51)”, in: QUCC 68, 7– 20. Hawes, G. (2014), Rationalizing Myth in Antiquity, Oxford. Hermann, E. (1923), Silbenbindung im Griechischen und in den anderen indogermanischen Sprachen, Göttingen.

146  Lucia Prauscello Käppel, L. (1992), Paian: Studien zur Geschichte einer Gattung, Berlin. Kirk, G.S. (1990), The Iliad: A Commentary, vol. 1: Books 1–4, Cambridge. Knoepfler, D. (2008), “Bulletin épigraphique: Béotie – Eubée”, in: REG 121, 618–668. Kousolini, V. (2016), “Panhellenic and Epichoric Elements in Corinna’s Catalogues”, in: GRBS 56, 82–110. Lejeune, M. (1972), Phonétique historique du mycénien et du grec ancien, Paris. Lhôte, P. (2008), “Bulletin épigraphique: Épire”, in: REG 121, 670–673. Long, A.A. (1968), Language and Thought in Sophocles: A Study of Abstract Nouns and Poetic Technique, London. MacDowell, D.M. (2009), Demosthenes the Orator, Oxford. Mack, W. (2015), Proxeny and Polis: Institutional Networks in the Ancient Greek World, Oxford. McPhee, B.D. (2018), “Mythological Innovations in Corinna’s Asopides Poem (fr.654.ii–iv PMG)”, in: GRBS 58, 198–222. Meier-Brügger, M. (1978), “Zu Griechisch κτῶμαι, ἐκτησάμην, (κ)έκτημαι”, in: Glotta 56, 224– 236. Méndez Dosuna, J. (2007a), “Les problèmes phonétiques de la propriété en pays béotien: ἔππασις et formes apparentés à la lumière des lamelles oraculaires de Dodone”, in: M.B. Hatzopoulos (ed.), Φωνῆς χαρακτήρ ἐθνικός. Actes du Ve Congrès International de Dialectologie grecque (Athènes, 28–30 septembre 2006), Paris, 295–316. Méndez Dosuna, J. (2007b), “En torno al significado de ἐγκτᾶσθαι, ἔγκτησις y ἔγκτημα (Ps.-D. VII 42 y otros pasajes)”, in: Estudios Clásicos 131, 29–44. Méndez Dosuna, J. (2016), “Some Critical Notes on the New Dodona Lead Plates”, in: ZPE 197, 119–139. Méndez Dosuna, J. (2017), “The Language of the Dodona Oracular Tablets: The Non-Doric Inquiries”, in: E. Crespo/G.G. Giannakis/P. Filos (eds.), Studies in Ancient Greek Dialects: From Central Greece to the Black Sea, Berlin/Boston, 265–296. Migeotte, L. (1984), L’emprunt public dans les cités grecques: Recueil des documents et analyse critique, Paris. Mitteis, L./Wilcken, U. (1912), Grundzüge und Chrestomathie der Papyruskunde, 2 vols., Leipzig. Morpurgo Davies, A. (1993), “Geography, History and Dialect: The Case of Oropos”, in: E. Crespo/J.L. García Ramón/A. Striano (eds.), Dialectologica Graeca. Actas del II Coloquio Internacional de Dialectologia Griega (Miraflores de la Sierra [Madrid], 19–20 de junio de 1991), Madrid, 261–279. Müller, C. (2010), “Les élites béotiennes et la richesse du IVe au IIe s. a. C.: Quelques pistes de reflexion”, in: L. Capdetrey/Y. Lafond (eds.), La cité et ses élites: Pratiques et représentation des formes de domination et de contrôle social dans les cités grecques. Actes du colloque de Poitiers, 19–20 octobre 2006, Paris, 225–244. Müller, C. (2011), “Évergétisme et pratiques financiers dans les cités de la Grèce Hellénistique”, in: REA 113, 345–363. Nickau, K. (2000), “Neues zur Überlieferung der Synagoge des Symeon (Paris. gr. 2552, Paris. Suppl. gr. 1238, Vatic. gr. 2226)”, in: GFA 3, 13–31. Page, D.L. (1953), Corinna, London. Petrakos, V.C. (1997), Οι επιγραφές του Ωρωπού, Αθήναι. Porzig, W. (1942), Die Namen für Satzinhalte im Griechischen und im Indogermanischen, Berlin.

Of Land, Ancestral Property and Prophecy in Corinna PMG 654  147

Prauscello, L. (forthcoming), “Themistocles’ philotimia in IEleusis 300 ll. 65–67 (= SEG 30.93): Some Linguistic Observations”, in: ZPE. Preisigke, F. (1915), Sammelbuch griechischer Urkunden aus Ägypten, 2 vols., Strassburg. Renaud, J.-M. (2004), Le mythe d’Orion, Liège. Robert, L. (1945), Le sanctuaire de Sinuri près de Mylasa, vol. 1: Les inscriptions, Paris. Robert, L. (1968), Opera minora selecta, vol. 3, Amsterdam. Roberto, U. (2012), Ioannis Antiocheni Fragmenta ex Historia chronica, Berlin/Boston. Schachter, A. (1981), Cults of Boiotia, vol. 1, London. Schachter, A. (1995), “The Prophet of Corinna fr. 654: Glaukos Pontios?”, in: Επετηρίς της Εταιρείας Βοιωτικών Μελετών, 1995, 1063–1072. Schachter, A. (2003), “Tanagra: The Geographical and Historical Context. Part I”, in: Pharos 11, 45–74. Schachter, A. (2016), Boiotia in Antiquity: Selected Papers, Cambridge. Schmidt, J. (1889), Die Pluralbildungen der indogermanischen Neutra, Weimar. Schulze, W. (1895), “Miscellen: 2. Wurzel ππᾱ-”, in: KZ 33, 318–320. Schwyzer, E. (1939), Griechische Grammatik, vol. 1: Allgemeiner Teil, Lautlehre, Wortbildung, Flexion, München. Sealey, R. (1990), Women and Law in Classical Greece, Chapel Hill/London. Sihler, A. (1995), A New Comparative Grammar of Greek and Latin, New York/Oxford. Teyssier, P. (1940), “Notes de dialectologie béotienne: II. τἀππάματα = τὰ ἀππάματα de *ἀν(α)πάματα”, in: RPh 14, 139–140. Thumb, A./Scherer, A. (1959), Handbuch der griechischen Dialekte, vol. 2, 2nd edn., Heidelberg. van Beek, L. (2016), “The Etymology of Greek πέπᾱμαι”, in: B.S. Saandgard Hansen/B. Nielsen Whitehead/T. Olander/B.A. Olsen (eds.), Etymology and the European Lexicon. Proceedings of the 14th Fachtagung der Indogermanischen Gesellschaft (Copenhagen, 17–22 September 2012), Wiesbaden, 427–441. Vergados, A. (2012), “Corinna’s Poetic Mountains: PMG 654 col. i 1–34 and Hesiodic Reception”, in: CP 107, 101–118. Vessella, C. (2012), “Lettori di Corinna in Egitto”, in: P. Schubert (ed.), Actes du XXVIe Congrès International de Papyrologie, Genève, 809–816. Vince, H.V. (1954), Demosthenes, vol. 1, London/Cambridge, Mass. Vottéro, G. (2001), Le dialecte béotien, vol. 2: Répertoire raisonné des inscriptions dialectales, Nancy. Vottéro, G. (2012), “Remarques sur les graphies et la langue des papyrus de Corinne”, in: C. Brixhe/G. Vottéro (eds.), Folia Graeca in honorem Edouard Will: Linguistica, Nancy, 97– 159. Wilamowitz-Moellendorff, U. von (1907), "§.14 Korinna”, in: Berliner Klassikertexte, vol. 5.2, Berlin, 19–55.

Federico Favi

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence Abstract: Two variant readings in Epicharmus’ fragment 80 (transmitted in Hephaestion’s Encheiridion and in Choeroboscus’ commentary) have an Aeolic -οισα vocalism which may point to lyric parody. These readings have never been given consideration, but some arguments may come in their support. To begin with, a review of non-Attic quotations in Hephaestion and Choeroboscus reveals that the manuscripts transmitting the Aeolic forms are linguistically trustworthy. Secondly, the other words used in the fragment are foreign to lyric poetry, but have a poetic allure which would be compatible with the use of Aeolic vocalism. Furthermore, another possible instance of Aeolic -οισα vocalism in Syracusan literature (comedy or mime) is provided by lexicographical sources. Finally, the evidence for the geographical distribution of Aeolicisms in non-Attic literary and epigraphical texts is discussed. Keywords: Epicharmus, Doric comedy, choral lyric poetry, textual criticism, Aeolic dialect.

 Epicharmus and Choral Lyric Poetry: Problems with the Available Evidence One of the most interesting topics in Epicharmean studies is the analysis of this poet’s interaction with other literary traditions.1 Detailed studies have been devoted to assessing Epicharmus’ relationship with Homer, the Epic Cycle, and

 1 I wish to thank Luigi Battezzato, Maria Chiara Martinelli, Peter J. Wilson, and the editors of this volume for their valuable advice, and also those who provided feedback and suggestions at the Venice conference in April 2017. I am grateful to the anonymous referee too, who has provided valuable suggestions and constructive criticism. This piece of research was also developed and presented during my stay at The University of Sydney in March 2017 as a recipient of the Apollo Visiting Fellowship, for which I am grateful to the Department of Classics and Ancient History of the University of Sydney. Finally, thanks are also due to Sergio Knipe for improving my English. This paper is a small token of gratitude for an inspirational teacher. https://doi.org/10.1515/9783110621747-008

150  Federico Favi tragedy,2 but no reliable evidence has been identified as far as Epicharmus’ interaction with choral lyric poetry is concerned. However, it is not unreasonable to imagine that some interplay may have taken place. Epicharmus was active in Syracuse under the Deinomenid tyrant Hiero, who had his agonistic victories (as well as those of his family and their entourage) celebrated in choral songs.3 Moreover, choral lyric poetry played a significant role in late Archaic and early Classical Sicilian culture. One should not think here only of panhellenic poets of high renown. For instance, an important defixio from the region around Gela testifies to the existence of choral competitions on a regional scale, too.4 Our lack of knowledge of Epicharmus’ interaction with choral lyric poetry is the result of several factors, but the primary one is in the limits and nature of the available evidence. By way of example, some scholars have suggested that for his comedies Epicharmus relied on myths derived from lyric poets, either poets of Sicilian origin or those active on the island,5 but this is hardly solid evidence. There are two main problems with such an approach. First, we do not have any complete Epicharmean play, and so we cannot tell whether Epicharmus modified

 2 A survey is provided by Rodríguez-Noriega Guillén 2012, 76−87 (see also Revermann 2013, Willi 2015, and Favi 2017b). 3 See Weber 1992, 51−60 and Morgan 2015. Some difficulties arise when discussing the relationship between Simonides and Hiero. Although the poet wrote epinician odes for other Siceliotes (see Mann 2001, 238, 246−247, and 299−311), we do not have any explicit evidence of compositions for members of the Deinomenid family and their entourage (see Molyneux 1992, 220−233; Weber 1992, 51; Morgan 2015, 93−96). However, Simonides was most certainly no stranger at Hiero’s court. It is said that the poet solved the conflict between Hiero and his brother Polyzelus (see Simon. test. 61B Poltera = Tim. FGrHist 566 F 93 = schol. Pind. Ol. 2.29d Drachmann, Diod. 11.48.3−5), not to mention Pausanias’ explicit testimony on Simonides’ familiarity with Hiero (Simon. test. 56 Poltera = Paus. 1.2.3). Not by chance, in Xenophon’s dialogue Hiero the interlocutor of the tyrant is Simonides himself. Indirect evidence for Simonides’ production for the Deinomenid family may be found in Simon. fr. 279 Poltera (= fr. 552 PMG = schol. Theoc. 1.65−66a Wendel; on this fragment see now Rawles 2018, 241) Σιμωνίδης δὲ Αἴτνην φησὶ κρῖναι Ἥφαιστον καὶ Δήμητραν περὶ τῆς χώρας ἐρίσαντας. In fr. 279 Poltera, Simonides seems to provide a Sicilian parallel to Poseidon’s and Athena’s dispute over Attica (see Hdt. 8.55.1). As this latter tradition was central to the construction of Athens’ regional identity, Hephaestus’ and Demeter’s dispute over Sicily may well have played a similar role in Syracuse (Demeter’s cult was particularly fostered by the Deinomenids, see Kowalzig 2008). Of little importance is a pseudo-Simonidean epigram (transmitted by schol. Pind. Pyth. 1.152 Drachmann and Anth. Pal. 6.214) celebrating the victories of the Deinomenids (see Page 1981, 247−250 and Bravi 2006, 77−80), while some indirect information concerning Simonides’ production for Hiero might be provided by Simon. test. 59 Poltera (see Rawles 2018, 241 n. 56). 4 See SEG 57.905b and most recently Wilson/Favi 2017. 5 See Rodríguez-Noriega Guillén 2012, 80−81.

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  151

the inherited mythical plot to pursue a comic aim. Second, there is no reliable evidence that any of those myths are peculiar to an individual lyric poet and that Epicharmus must have consequently interacted with that poet. In other words, we cannot say with any confidence what Epicharmus had in mind, whether the work of a single poet or (a more likely hypothesis) a particular myth. Nevertheless, scholars have suggested that it is possible to identify allusions to lyric poetry, and especially to Pindar, in a number of Epicharmus’ fragments.6 But all these cases present some severe difficulties. One clear example of the issues at stake is provided by the use of the etymology λᾶες ~ λαοί within the myth of Deucalion and Pyrrha in both Epicharm. fr. 120 K.-A. and Pind. Ol. 9.43−46.7 Although some have suggested that Epicharmus is alluding to Pindar’s passage, it is more reasonable to assume that both poets made use of an etymology which was simply traditional, and hence that they did not necessarily interact. Indeed, this etymology is as old as Homer,8 and it is first used with reference to the myth of Deucalion and Pyrrha in a fragment of Hesiod.9 It is also doubtful, at least to some extent, whether the λαοί mentioned in Epicharmus really are those that Hesiod and Pindar talk about.10 This possibility deserves attention, given that according to Hyg. 153 (on which schol. Germ. p. 154 Breysig relies) after the flood Deucalion and Pyrrha found safety on Mount Aetna. It is intriguing that this Sicilian variant of the myth was the one already familiar to Epicharmus. Comparable issues and problems apply to a number of other cases, which I shall examine in what follows. One may begin with the personification of Ἡσυχία in Epicharm. fr. 100 K.-A. This is to be found in Pind. Ol. 4.16, Pyth. 8.1−4 and fr. 109.2 S.-M.11 (and later on  6 See Rodríguez-Noriega Guillén 2012, 82 and n. 3. 7 This might already have been an ancient idea, since Epicharmus’ fragment is transmitted by the scholia to this very passage from Pindar’s ninth Olympian Ode (schol. (A) Pind. Ol. 9.70b Drachmann). 8 See Hom. Il. 24.611 and Richardson 1993, 341. 9 See Hes. fr. 234 M.-W. This would be paralleled by the shared use of the etymology Προμεθεύς ~ προμηθεῖν in both Epicharm. fr. 113.12 K.-A. and Hes. Theog. 510−511 and 546, which also makes it clear that [Aesch.] PV 85−86 is not necessarily interacting with Epicharmus’ passage, as Flintoff 1986, 84–85 suggested: see Bees 1993, 25. 10 In these poets’ passages the λαοί alluded to with the etymology λᾶες ~ λαοί are the Lelegians, pars pro toto of the Locrians (see Arist. fr. 560 Rose transmitted together with Hesiod’s fragment in Strab. 7.7.2; see further P. Giannini’s discussion of Pind. Ol. 9.42−46 in Gentili et al. 2013, 536). One wonders whether Epicharmus was also thinking of the Locrians when using this folk etymology, or rather of mankind as a whole (for the latter possibility consider for instance Callim. fr. 496 Pfeiffer). 11 All Pindaric passages are discussed by Venezia 2007.

152  Federico Favi in Ar. Av. 1318−1322),12 but there is no good reason to think that Epicharmus had Pindar’s passage(s) in mind, for this personification is arguably not a distinctively Pindaric feature. For instance, Epicharmus’ fragment is strongly reminiscent of Critias fr. 6.21 West, too. Furthermore, there are some chronological problems to deal with. Since Epicharmus most probably died in the 450s,13 it is unlikely that he could have been alluding to Pind. Ol. 4 and Pyth. 8, which date (respectively) from around 460/45614 and 446.15 As a further example, one might mention the difficulties posed by Arnson Svarlien’s suggestion of establishing a direct connection between Bacchyl. 3.78−82 and [Epicharm.] fr. 256 K.-A.16 This connection is unlikely on various grounds. Firstly, [Epicharm.] fr. 256 K.-A. is probably a pseudo-epigraphic moral saying ascribed to the Syracusan comic poet,17 and so it is out of place in a discussion dealing with the historical Epicharmus’ interactions with choral lyric poetry. But even if the fragment were authentic, the content of both this and Bacchylides’ passage is common wisdom,18 and so it would not be revealing — let alone reliable — for proving any degree of interaction between the two authors.19 Conscious of the difficulties posed by this last example, Arnson Svarlien has analyzed two further cases in which, she maintains, Epicharmus (almost) certainly interacts with Pindar. The first case is a fragment from Epicharmus’ Λόγος καὶ Λογίνα. According to Arnson Svarlien, it displays an echo of Pindar’s new version of the myth of Tantalus and Pelops in the first Olympian Ode:

 12 De Cremoux 2011, 63 suggested that Aristophanes derives this personification of Ἡσυχία from Epicharmus, but it is more likely that he is only interacting with Pind. Pyth. 8.1 (see also Dunbar 1995, 650). 13 See Pickard-Cambridge/Webster 1962, 230−233 and Kerkhof 2001, 57−59. 14 See L. Lomiento in Gentili et al. 2013, 99. 15 See B. Gentili in Gentili et al. 1995, 211. 16 Arnson Svarlien 1990−1991, 105. 17 See also Ahrens 1843, 461 (= Epicharm. fr. 149 Ahrens); Kaibel 1899, 141 (= [Epicharm.] fr. 267 Kaibel); Olivieri 1946, 120 (= [Epicharm.] fr. 236 Olivieri). 18 A further parallel for the pseudo-Epicharmean line has been found in Eur. Alc. 782−783 by Olivieri 1946, 120 (see Parker 2007, 208 too). 19 This objection also applies to the suggestion by Rodríguez-Noriega Guillén 2012, 83 n. 33 of an interplay between [Epicharm.] fr. 251 K.-A. (which she suspects to be genuine) and Pind. Isthm. 5.16: many other loci similes are known (see e.g. those collected in K.-A. I, 147).

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  153

(1)

(A) ὁ Ζεὺς μ᾽ ἐκάλεσε, Πέλοπι γ᾽ ἔρανον ἱστιῶν. (B) ἦ παμπόνηρον ὄψον, ὦ τᾶν, ὁ γέρανος. (A) ἀλλ᾽ οὔτι γέρανον, ἀλλ᾽ ἔρανόν τοι λέγω. (Epicharm. fr. 76 K.-A.)20 (A) Zeus invited me to a party for Pelops; he᾽s really putting on the dog. (B) My good man, dog is a terrible choice for the menu! (A) I don᾽t mean the menu; it᾽s going to be a fancy party. (transl. J.S. Rusten in Rusten 2011, 67)

(2)

υἱὲ Ταντάλου, σὲ δ᾽ ἀντία προτέρων φθέγξομαι, ὁπότ᾽ ἐκάλεσε πατὴρ τὸν εὐνομώτατον ἐς ἔρανον φίλαν Σίπυλον, ἀμοιβαῖα θεοῖσιν δεῖπνα παρέχων, τότ᾽ Ἀγλαοτρίαιναν ἁρπάσαι, δαμέντα φρένας ἱμέρῳ, χρυσέαισί τ᾽ ἀν᾽ ἵπποις ὕπατον εὐρυτίμου ποτὶ δῶμα Διὸς μεταβᾶσαι. (Pind. Ol. 1.36−42) Son of Tantalos, of you shall I say, contrary to my predecessors, that when your father invited the gods to his most orderly feast and to his friendly Sipylos giving them a banquet in return for theirs, then it was that the Lord of the Splendid Trident seized you, his mind overcome by desire, and with golden steeds conveyed you to the highest home of widely honored Zeus. (transl. Race 1997, 51)

The Witz in (1) lies in the faulty segmentation of γ᾽ ἔρανον ‘banquet’ uttered by character A (line 1) and understood by character B as γέρανος ‘crane’ (line 2). According to Arnson Svarlien,21 Epicharmus’ fragment (no. (1) above) echoes the re-writing of Tantalus’ myth in Pindar’s first Olympian Ode (no. (2)), while adding some new details. Pindar describes the banquet offered by Tantalus in Sipylus. He explains that Pelops was not killed and cooked by his father, as stated by all previous poets: instead, the boy was kidnapped by Poseidon, who had fallen in love with him and who brought him to Zeus’ house on Mount Olympus. Building on this, Arnson

 20 I print line 3 of the fragment as edited by K.-A., who accept Ahrens’ emendation for the unmetrical paradosis. However, Willi 2015, 122 n. 30 makes a good case for accepting Porson’s text (ἀλλὰ γ᾽ ἔρανον τοι λέγω), which would emphasize the word play between γέρανος and γ᾽ ἔρανον. For instance, one may imagine that line 3 was pronounced with special emphasis in order to make the word boundaries clear. This would have a good comic effect. 21 Arnson Svarlien 1990−1991, 106−108. This reconstruction is followed by Rodríguez-Noriega Guillén 1996, 83 and Morgan 2015, 108, while Gerber 1982, 75 was much more sceptical on this possibility.

154  Federico Favi Svarlien’s suggestion is that character A in Epicharm. fr. 76 K.-A. alludes to a banquet given by Zeus “for Pelops” (line 1: Πέλοπι γ᾽ ἔρανον ἱστιῶν)22 who had previously not been killed and cooked by his own father. A serious inconvenience is that the dative Πέλοπι in line 1 is syntactically problematic together with ἱστιῶν. ἑστιάω is a transitive verb which only governs the person invited to dinner or an internal accusative, such as ἔρανος. There is no place for an indirect object (“for Pelops”) as required by Arnson Svarlien’s (and others’) reconstruction. The parallel that some scholars have identified in schol. Ar. Thesm. 942b Retguit Ἀττικὴ δὲ ἡ σύνταξις, ὅτι τὸ ἑστιᾶν καὶ δοτικῇ συντάσσουσιν is of no use, for the scoliast’s comment is most probably due to a faulty understanding of Ar. Thesm. 942 γέλωτα παρέχω τοῖς κόραξιν ἑστιῶν.23 In this last line it is much more likely that τοῖς κόραξιν is governed by παρέχω and that ἑστιῶν is absolute (the crows are implicit anyhow).24 Richard Kannicht tried to provide a solution to this syntactical problem by interpreting the dative Πέλοπι in line 1 as an adverbial dative of means.25 He consequently translates the line as: ‘er [scil. Zeus] gibt mit Pelops (d. h. mit dem Fleisch des Pelops) einen ἔρανος’.26 But the problem remains open. Although likely on syntactical grounds, in this new reconstruction Zeus would be said to be absolutely conscious of what he is doing,27 i.e. serving Pelops to his guests. One might conclude that the god is ascribed an unconscious responsibility for the death of Pelops.28 But in the myth it is Tantalus who invites the gods to the banquet, not the other way round. Moreover, even if one thinks that Zeus is described

 22 See also Olivieri 1946, 66; Pickard-Cambridge/Webster 1962, 273; Rodríguez-Noriega Guillén 1996, 70; Olson 2007, 39. 23 The scholia on Aristophanes’ passage also mention the use of this construction by Eupolis, yet without quoting his text (this testimony, rather than fragment, is now Eup. fr. 67 K.-A.). 24 See Kock and Kaibel apud K.-A. V, 330; Sommerstein 2001, 109 (translation); Hartwig 2009. Austin/Olson 2004, 297 do not discuss the problem, but they have just one commentary lemma for τοῖς κόραξιν ἑστιῶν, which is likely to mean that they took it as a single syntactic unit. Prato 2001, 306 suggests an ἀπὸ κοινοῦ construction of τοῖς κόραξιν (see also Olson 2017, 24), but even if this were true, it would hardly have any bearing on Epicharmus’ fragment. 25 See Kannicht 1969, 123. For this construction of ἑστιάω, see e.g. Pl. Resp. 404b οἶσθα γὰρ ὅτι ἐπὶ στρατιᾶς ἐν ταῖς τῶν ἡρώων ἑστιάσεσιν οὔτε ἰχθύσιν οὔτε αὐτοὺς ἑστιᾷ. 26 If this reading were confirmed, Epicharmus would not be interacting at all with the new version of the myth of Tantalus and Pelops created by Pindar. 27 The passage from ἐκάλεσε to ἱστιῶν in line 1 makes it absolutely likely that character A was invited to dinner by Zeus before the banquet, and that he has already been informed that Pelops was to be eaten. 28 We might think here of something like the saying “Oedipus married his mother”, whereby the Theban king would be ascribed an action that he is not conscious of.

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  155

as unintentionally responsible for something that he is not aware of, this would contradict the very meaning of the myth. Indeed, what Tantalus wanted to test was the gods’ omniscience. The number of puzzling problems makes it likely that we do not have any good solution for interpreting Epicharm. fr. 76 K.-A. We cannot tell whether Epicharmus is alluding to Pindar’s epinician ode, although it is easy to speculate on the matter. One could envisage the possibility that the comment ἦ παμπόνηρον ὄψον uttered by character B in line 2 is meant to create some tension between the reason why he/she says this (i.e. misunderstanding γ’ ἔρανον for γέρανον), and what everyone in the audience would probably have thought of as soon as they heard the name Pelops and the mention of a banquet in the same line (i.e. the pre-Pindaric story of Tantalus’ banquet). Should this be true, it is hardly likely that Epicharmus’ fragment owes anything to the first Olympian Ode. In the last case the degree of uncertainty is so great that it is difficult to say with any confidence what kind of relationship exists — if at all — between Epicharmus and Pindar. What is even less likely to provide a sound answer to such a question is a further example where the interplay that Arnson Svarlien29 suggests between the title of a comedy by Epicharmus (Πίθων)30 and a passage from Pindar is hardly convincing: (3) καλός τοι πίθων παρὰ παισίν, αἰεὶ καλός. (Pind. Pyth. 2.72−73) Pretty is an ape in the eyes of children, always pretty. (transl. Race 1997, 239)

According to Arnson Svarlien, πίθων is used by Pindar to mock the title of Epicharmus’ comedy. But while in the Pythian Ode the meaning of the word is certain, this is really not the case with the title of Epicharmus’ comedy, for which three different interpretations are equally possible:31 (a) as in Pindar, Πίθων could be the hypocoristic form of πίθηκος;32 both Epicharmus and Pindar would preserve the first instances of this word, which only appears much later in Babr. 1.56;  29 Arnson Svarlien 1990−1991, 108−109. 30 Poll. 10.179 Bethe ὡς Ἐπίχαρμος Πίθωνι κτλ. 31 Interpretation (c) is rejected (although without any explicit justification) by Kaibel and K.-A. (followed by Kerkhof 2001, 133 n. 1), but it is not clear why it should not be taken into account (see also Lorenz 1864, 147 and Pickard-Cambridge/Webster 1962, 271). 32 For this formation see Chantraine 1933, 161 § 120 and Locker 1933, 145.

156  Federico Favi (b) the hypocoristic form of πίθηκος could also serve as a personal name, as in a 5th-century Selinuntine defixio;33 (c) instead of Πίθων, one could read πιθών ‘well’;34 the only preserved fragment from this play (Epicharm. fr. 112 K.-A.) reads ἢ θύλακον βόειον ἢ κόιν φέρειν / ἢ κωρυκίδα, and this list of instruments might suggest that in the play things were moved out of and into a well.35 Unfortunately, many equally reasonable alternative interpretations are at hand, and none of them seems more or less likely than the others. Let us assume that Arnson Svarlien is right, and that the title of the Epicharmean play was Πίθων ‘(little) monkey’. In her view, Pindar uses this word in the second Pythian Ode to parody the title of one of Epicharmus’ comedies.36 However, one may also argue that it was Epicharmus who used πίθων to mock the unusual occurrence of this word in the Pythian Ode. In addition to this, one should also bear in mind that the interpretation of the final sections of the second Pythian Ode (lines 72−96) has always raised some doubts.37 In particular, the referent behind πίθων itself has been identified either as Bacchylides, as the scholia suggest,38 or as some other member of Hiero’s entourage. Given all these uncertainties, it is arguably better not to insist any further on trying to make too much of the use of πίθων both in Pindar’s second Pythian Ode and (perhaps) as the title of a comedy of Epicharmus. Before moving forward, one last fragment of Epicharmus should receive some attention. The testimonia which constitute Epicharm. fr. 92 K.-A. inform us that in the comedy entitled Μοῦσαι the poet mentioned the enoplios nomos, a kind of armed dance which, according to the myth, Athena performed for the Dioscouri.39 One of the testimonies is schol. Pind. Pyth. 2.127. The scholium comments  33 IGDS I no. 34. See also Bettarini 2005, 132. 34 Further literary instances of this word are Eup. fr. 122 K.-A. and Pherecr. fr. 147 K.-A., beside many papyri and inscriptions. Arnson Svarlien 1990−1991, 108 n. 18 is aware of this interpretation (originally put forward by Welcker 1844, 306), but she does not comment on it. 35 One might think of Plautus’ Aulularia. 36 It is well known that the final sections of the second Pythian Ode are stylistically unusual and that they look much more similar to iambic poetry (see Woodbury 1945, 11 and Most 1985, 126). The possibility that Pindar may be echoing and mocking Epicharmus, then, would be coherent with the sentiment and content of the passage (Arnson Svarlien 1990−1991, 109: “what better way to introduce this striking departure from genre than with an allusion to a contemporary Syracusan comedy with which Hieron was probably familiar?”). 37 See E. Cingano in Gentili et al. 1995, 51−54. 38 See Gentili et al. 1995, L−LIII and schol. Pind. Pyth. 2.131b−132c−d−e−f Drachmann. 39 This fragment is the focus of a highly speculative discussion by Piva 2011.

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  157

on a passage in Pindar’s second Pythian Ode (lines 69−71) where mention is made of the Kastoreion, identified by the scholia as a Pindaric hyporchema, the first line of which is also quoted.40 The nature and identity of this Kastoreion are a major problem,41 although it is not excluded that the enoplios nomos mentioned by Epicharmus is to be identified with Pindar’s hyporchema.42 But doubts and concerns are not yet allayed. Indeed, we cannot rule out the possibility that Epicharmus mentioned the genre that Pindar’s song belonged to, rather than his very poem. Once again, substantial difficulties make the prospect of any actual interaction between Epicharmus and Pindar uncertain.43 As the evidence provided by (alleged) loci similes and by shared mythical narratives and/or other issues of content all prove insufficient for assessing Epicharmus’ degree of interaction with choral lyric poetry, it is to matters of language that one should probably turn for finding some reliable pieces of evidence. A first step is taken by Ucciardello,44 whose new reading in P.Oxy. 2427 fr. 20.5 = Epicharm. fr. 113.187 K.-A. has recovered the non-augmented aorist ἔναρον (from ἐναίρω, a verb largely attested in poetry from Homer onwards), which is paralleled in lyric poetry and tragedy. It seems possible to take some further steps in this direction.

 40 Pind. fr. 105a.1 S.-M. 41 See E. Cingano in Gentili et al. 1995, 391−392 and Morgan 2015, 193−194. 42 This conclusion is suggested by the fact that both poets are mentioned in the same scholium. According to the scholia, Pindar’s hyporchema was a (free) gift from Pindar to Hiero, and so one may well assume that this unusual event made the ode particularly famous and suitable for comic parody: for instance, this is exactly what happens in the scene of the beggar poet in Ar. Av. 904−951. 43 A passing remark should be devoted to the suggestion by Morgan 2015, 107 that the comedies of Epicharmus entitled Ἐπινίκιος (or Ἐπίνικος, see the variant readings collected in Epicharm. Ἐπίνικος vel Ἐπινίκιος test. ii K.-A. = Heph. 8.2, 24.20 Consbruch) and Χορεύοντες were intended to mock, respectively, Hiero’s spasmodic interest in celebrating his agonistic victories and choral performances. But in these and all similar cases, it is doubtful whether it is possible to understand the content of a play just by its title (see especially Pickard-Cambridge/Webster 1962, 269). For instance, Kerkhof 2001, 132 suggested that Ἐπινίκιος/Ἐπίνικος dealt with an ἀλαζών. To my mind, another possibility that should be taken into serious consideration is athletic parody: one may think of the critique levelled against athletes by Xenoph. B 2 D.-K. 44 Ucciardello 2013.

158  Federico Favi

 Two Variant Readings in Epicharm. Fr. 80 K.-A.: Previously Neglected Evidence I believe that some important, though largely neglected, evidence for Epicharmus’ interaction with choral lyric poetry may be found in a fragment of Epicharmus’ Μεγαρίς.45 The fragment runs as follows:46 (4) εὔυμνος καὶ μουσικὰν ἔχουσα πᾶσαν, φιλόλυρος. (Epicharm. fr. 80 K.-A.) Celebrated in songs and knowledgeable about music, a lover of the lyre. Heph. 1.8, 6.2 Consbruch ἤδη μέντοι ἡ διὰ τοῦ μν σύνταξις ἐποίησέ που καὶ βραχεῖαν, ὡς παρὰ Κρατίνῳ ἐν Πανόπταις [= fr. 162 K.-A.]· “ἀλλοτριογνώμοις ἐπιλήσμοσι μνημονικοῖσι”, καὶ παρ’ Ἐπιχάρμῳ ἐν Μεγαρίδι· “εὔυμνος – φιλόλυρος † ἠχὴ †”, καὶ παρὰ Καλλιμάχῳ [= fr. 61 Pfeiffer]· “τὼς μὲν ὁ Μνησάρχειος ἔφη ξένος”. Choer. In Heph. ench. 201.4 Consbruch καὶ παραφέρει ὁ μετρικὸς τοῦτο ἐν διαφόροις· οἷον παρὰ τῷ Κρατίνῳ ἐν Πανόπταις δράματι [= fr. 162 K.-A.] εἰς ἔπος “ἀλλοτριογνώμοις ἐπιλήσμοσι μνημονικοῖσι”, καὶ παρὰ Ἐπιχάρμῳ ἐν Μεγαρίδι· “εὔυμνος – φιλόλυρος ἠχή”. ἔστι δὲ τὸ μέτρον τροχαϊκὸν τετράμετρον ὑπερκατάληκτον. μουσικὰν Heph. ADI : μοισικὰν Choer. KU || ἔχουσα Heph. A, Choer. KU : ἔχοισα Heph. DI || φιλόλυρος ex φιλόχορος D || ἠχὴ I : ἢχή D : ἠχῆι A : ἠχὴ delevit Arnaldus : ἧι δὴ Wilamowitz apud Consbruch : Choeroboscus (KU) versum ita interpretatus est, ut tetrametrum trochaicum hypercatalectum sit.

 45 Two other fragments from this comedy have been preserved, Epicharm. frr. 79 and 81 K.-A. The first describes a man whose physical appearance is compared to that of sea animals. In the second, which is particularly problematic as far as the text and its interpretation are concerned, some kind of remark is made on the body parts of an animal that are said not to be edible. Given that all we know of the content of Epicharmus’ Μεγαρίς comes from its title and three remaining fragments, the plot of the play is definitively obscure (very cautious suggestions are put forward by Kerkhof 2001, 130). 46 The text of the fragment is that of K.-A. For the manuscripts’ sigla see Consbruch 1906. The apparatus criticus has been adapted from K.-A. In particular, I have added some clarifications concerning the reading ἠχὴ, which in Hephaestion and Choeroboscus belongs to the fragment, although it is unmetrical (Hephaestion was conscious of this problem, and hence he explained the Epicharmean line as a τροχαϊκὸν τετράμετρον ὑπερκατάληκτον; however, this solution too is hardly likely, for the line is two syllables too long, not just one). I have checked the variant readings of the manuscripts D (f. 208 verso) and I (f. 11 recto) via digitised copies available at http://tudigit.ulb.tu-darmstadt.de/show/Hs-2773/0001?sid=300426e6382ace207fea80c1693e5245 (manuscript D) and http://gallica.bnf.fr/ark:/12148/btv1b10722407j/f3.item.zoom (manuscript I).

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  159

The fragment is transmitted in Hephaestion’s Encheiridion and in Choeroboscus’ commentary. Hephaestion’s interest was sparked by the second syllable of εὔυμνος, where the consonant cluster [mn] scans tautosyllabically.47 The sources of the fragment inform us that the comedy from which the line derives was entitled Μεγαρίς. Such a title makes it likely that the main character of the play was a hetaera after whose ethnonym the comedy was named.48 Given that the figure described in Epicharm. fr. 80 K.-A. is a woman well versed in music and poetry, she is quite likely to be the Megarian hetaera herself.49 Very little consideration has been devoted by scholars to one element. Both ἔχουσα and μουσικὰν have variant readings with the Aeolic vocalism . The reading μοισικὰν occurs in both manuscripts of Choeroboscus’ commentary, against μουσικὰν in all witnesses of Hephaestion’s Encheiridion (ADI). The participle ἔχοισα is transmitted in manuscripts D and I50 of Hephaestion’s Encheiridion, against ἔχουσα both in manuscript A of Hephaestion51 and in manuscripts K and U of Choeroboscus’ commentary. Kassel and Austin, as well as all previous editors, are not inclined to take ἔχοισα and μοισικὰν as valid alternative readings.52 This is not surprising. To begin with, the Aeolic development of the 2nd

 47 This prosody is discussed by Fraenkel 1950, 2.443–444 and 3.830. See also Martinelli 1997, 56–57 and n. 79. Epicharmus’ line provides the only known example of [mn] scanning tautosyllabically within a word (Martinelli also discusses two tragic passages in which this scansion also occurs according to West 1982, 18). 48 See Arnott 1996, 294. This interpretation of the Μεγαρίς was already put forward by Welcker 1844, 304 and von Salis 1905, 51. 49 On this point, see further below. 50 This variant reading is also found in manuscript M (see the apparatus of Hoerschelmann 1886, 49). According to Consbruch 1906, X, this manuscript M is a “gemellus […] sed paulo deterior” of I. This witness is not particularly reliable for the constitutio textus (see Consbruch’s discussion), but in this particular case it is quite likely that the agreement between I and M means that their common ancestor already had ἔχοισα. 51 For other witnesses of Hephaestion’s Encheiridion that transmit this reading, see the apparatus criticus of Gaisford 1810, 15 and Gaisford 1855, 13. 52 See Ahrens 1843, 447 (= Epicharm. fr. 69 Ahrens); Lorenz 1864, 246 (= Epicharm. Μεγαρίς fr. 3 Lorenz); Kaibel 1899, 107 (= Epicharm. fr. 91 Kaibel); Olivieri 1946, 69 (= Epicharm. fr. 115 Olivieri); Rodríguez-Noriega Guillén 1996, 73 (= Epicharm. fr. 90 Rodríguez-Noriega). The only edition to print one of these variant readings, ἔχοισα, is Polman Kruseman 1834, 56 (= Epicharm. Μεγαρίς fr. II Polman Kruseman). But this is no independent choice. Polman Kruseman adopted the text that Gaisford had previously printed in Heph. 15.1−3 Gaisford (= Gaisford 1810) following the suggestion of Spanheim 1697, 2.95 (ad Callim. Hymn. 2.31) that the fragment should be reconstructed as (3ia.) εὔυμνος / καὶ μουσικὰν ἔχοισα πᾶσαν φιλόλυρος / ἠχή. However, Gaisford himself changed his mind, and in Heph. p. 13.6−7 Gaisford (= Gaisford 1855) he printed the text of the fragment following Ahrens’ edition.

160  Federico Favi compensatory lengthening is foreign to the Syracusan dialect of Epicharmus.53 One would rather expect to find such forms either in the Lesbian poets or in choral lyric poetry. Moreover, in the whole of Attic comedy there is no instance of this Aeolic vocalism (not even in choral songs), and one feels inclined to believe (even unconsciously) that this was the case in Syracusan comedy too. However, a number of reasons make it likely that the variant readings ἔχοισα and μοισικὰν should be re-evaluated. And if this is the case, the very significant result is that we finally have reliable evidence of Epicharmus’ interaction with choral lyric poetry.

. Problems in the Constitutio Textus The constitutio textus of Hephaestion’s Encheiridion relies on three witnesses that have received the sigla A, D, and I. None of them is a descriptus from either of the remaining two. D and I share some common mistakes, which means they probably derive from a common ancestor, while A agrees with either D or I only in correct or not-significant variant readings. A is usually the most reliable witness, but in several cases D and/or I transmit a superior text.54 As for Choeroboscus, the constitutio textus of his commentary relies on just two manuscripts, K and U. Although they can be traced back to an ancient common ancestor, they are copies of different antigraphi.55 If one agrees with those who print ἔχουσα and μουσικὰν, it would be desirable to have a satisfactory explanation for the variant readings ἔχοισα and μοισικὰν. The genesis of these variant readings is far from obvious.56 First, it is highly unlikely that they have been inserted by a learned scribe who aimed to give Epicharmus’ line a lyric nuance. No good evidence in support of this assumption is provided either by the context that the quotation derives from (e.g. lyric quotations that precede/follow Epicharmus’ fragment), or by the content of the fragment itself (the metre is not a lyric one, the diction is refined but lacks any lyric form, and Epicharmus is not the kind of poet to whom one would ascribe Aeolic forms). What is also noteworthy is the indirect agreement in the transmission of forms with the Aeolic vocalism between manuscripts D and I of Hephaestion on the one hand (ἔχοισα), and manuscripts K and U of Choeroboscus on the other  53 See Willi 2012, 267. For the development of 2nd compensatory lengthening in the Doric dialects of Sicily see also Mimbrera 2012, 203. 54 See Consbruch 1906, XI. 55 See Consbruch 1906, XVIII. 56 A palaeographical mistake is extremely unlikely to be the solution of the problem.

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  161

(μοισικὰν). This typological coincidence is all the more remarkable as it applies to different words within the same fragment. Merely incidental polygenesis is quite unsatisfactory as an explanation. I believe that the best explanation for this is that both forms with Aeolic vocalism, ἔχοισα and μοισικὰν, are the correct variant readings that one should print in the fragment, and that they survived the process of Medieval transmission despite the natural tendency of copyists to trivialize them into their much more common Attic equivalents ἔχουσα and μουσικὰν. This conclusion will appear quite reasonable if one compares the way in which other non-Attic forms are transmitted throughout Hephaestion’s handbook and Choeroboscus’ commentary. In what follows, my aim is to discuss all such cases in order to assess the reliability in matters of dialect of the witnesses which transmit μοισικὰν and ἔχοισα in Epicharm. fr. 80 K.-A.

. The Transmission of Non-Attic Forms in Hephaestion and Choeroboscus Manuscripts D and I of Hephaestion’s Encheiridion transmit ἔχοισα against ἔχουσα of manuscript A and against manuscripts K and U of Choeroboscus’ commentary. What I aim to show is that, when it comes to non-Attic forms, D and I are often quite reliable in providing the correct variant readings, which is all the more interesting as in several cases they do so against trivializations transmitted by manuscript A. In the following examples (nos. (5−10)) I have collected all those instances where D and/or I transmit the correct variant readings, while A does, or does not, agree with them (among these, case no. (7) is of special interest for the present discussion):57 (5) καὶ κεῖνος ἐν σάλεσσι ἥμενος μάκαρ ἀνήρ (Heph. 1.3, 2.9 Consbruch = Alcm. fr. 15 PMG) σάλεσσι D : σάλεσι AI

 57 The apparatus critici are mainly based on the editions of Hephaestion’ Encheiridion and Choeroboscus’ commentary by Hoerschelmann 1886 and Consbruch 1906. I have also checked them against the most recent critical editions of the authors who are quoted. I have checked the readings of manuscript K of Choeroboscus’ commentary through a microfilm copy housed at the Biblioteca Nazionale Centrale di Roma, since at times the (mostly) negative apparatus critici of Hoerschelmann’s and Consbruch’s editions may leave some doubts open. I have not added a translation to avoid bulkiness.

162  Federico Favi (6) † χαίροις ἁ νύμφα, χαιρέτω δ’ ὁ γάμβρος † (Heph. 4.2, 13.8 Consbruch = Sapph. fr. 117 Voigt) χαίροις ἁ ADI (7) δεῦτε νυν ἄβραι Χάριτες καλλίκομοί τε Μοῖσαι (Heph. 9.2, 30.5 Consbruch = Sapph. fr. 128 Voigt) Μοῖσαι DI : Μοῦσαι A (8) νύμφαι, ταὶς Δίος ἐξ αἰγιόχω φαῖσι τετυχαμέναις (Heph. 10.6, 34.14 Consbruch = Alc. fr. 343 Voigt) αἰγιόχω Ι : αἰγιόχου Α (9) ἔμε δείλαν, ἔμε παῖσαν κακοτάτων πεδέχοισαν (Heph. 12.2, 38.5 Consbruch = Alc. fr. 10.1−2 Voigt) πεδ’ ἔχοισαν C : παῖδ’ (ε s.l.) ἔχοισαν A : παῖδ’ ἔχοισαν IH (10) Ἰνὼ σαλασσομέδοισ’ ἃν ἀπὸ μασδῶν (Heph. 14.6, 46.9 Consbruch = Alcm. fr. 50b PMG) σαλασσομέδοισ’ ἄν Porson : σαλασσομέδοισαν A : σάλας ὁμέδοισαν I

This list of examples makes it clear that in a number of cases we owe to D and/or I the preservation of correct dialectal variant readings that otherwise would have been lost. On the contrary, when D and/or I do not transmit the correct variant readings, they are only responsible for trivializations. This is significant since it shows that it would be quite unwarranted to take the ἔχοισα that D and I transmit in Epicharmus’ fragment as being due to the hand of a learned scribe. In what follows (nos. (11−20)) I have collected all those passages where witnesses D and/or I transmit mistakes that are only due to Attic trivializations (see especially (16), (18), and (20)): (11) Ἔρος δηὖτέ μ’ ὁ λυσιμέλης δόνει (Heph. 7.7, 23.19 Consbruch = Sapph. fr. 130.1 Voigt) ἔρος A : ἔρως DI (12) φροντίσδην, ἐπὶ δ’ Ἀνδρομέδαν πότη (Heph. 7.7, 23.22 Consbruch = Sapph. fr. 130.4 Voigt) πότη A : πότε DI (13) ὃν πρᾶτος εἰσηγήσαθ’ Ὡριστόξενος (Heph. 8.3, 25.16 Consbruch = Epicharm. fr. 77.2 K.-A.) πρᾶτος A : πρῶτος DI (14) νύμφα, σὺ μὲν ἀστερίαν ὑπ’ ἄμαξαν ἤδη (Heph. 8.9, 28.19 Consbruch = Callim. fr. 228.5 Pfeiffer) ὑπ’ ἄμαξαν A : ὑφ’ ἅμαξαν DI (15) γλυκῆα μᾶτερ, οὔ τοι δύναμαι κρέκην τὸν ἱστὸν (Heph. 10.5, 34.9 Consbruch = Sapph. fr. 102.1 Voigt) κρέκην A : κρέκειν DI

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  163

(16) νύμφαι, ταὶς Δίος ἐξ αἰγιόχω φαῖσι τετυχαμέναις (Heph. 10.6, 34.14 Consbruch = Alc. fr. 343 Voigt) φαῖσι A : φασὶ I (17) ὠρχεῦντ’ ἁπαλοῖσ’ ἀμφ’ ἐρόεντα βωμόν (Heph. 11.3, 36.2 Consbruch = Inc. Auct. fr. 16.2 Voigt) ὠρχεῦντ’ ἁπαλοῖσ A : ὠρχεῦνθ’ ἁπαλοῖσ I (18) ἐμφέρη ἔχοισα μόρφαν Κλέις < > ἀγαπατά (Heph. 15.8, 53.17 Consbruch = Sapph. fr. 132.2 Voigt) ἔχοισα A : ἔχουσα I || μόρφαν et ἀγαπατά A : μορφὴν et ἀγαπητά I (19) ἐκτετέλεστ’, ἔχῃς δὲ πάρθενον, ἂν ἄραο (Heph. 15.26, 55.25 Consbruch = Sapph. fr. 112.2 Voigt) ἔχῃς A58 : ἔχεις I (20) ἔμε δείλαν, ἔμε παῖσαν κακοτάτων πεδέχοισαν (Heph. Poëm. 65.17 Consbruch = Alc. fr. 10.1−2 Voigt) πεδ’ ἔχοισα A : πεδέχουσα I

Finally, in two cases ((21−22)) one may note that only manuscript A hyper-characterizes non-Attic forms: (21) ἦ † διανεκῶς εὕδεις; οὐ μὰν πάρος ἦσθα, Κόριννα (Heph. 2.3, 9.18 Consbruch = Corinna fr. 657 PMG)59 εὕδεις DI : εὕδης Α : Page (PMG p. 336) de εὕδις et εἶσθα reponendis cogitat (22) πίσσυγοι δὲ δέκ’ ἐξεπόνησαν (Heph. 7.6, 23.13 Consbruch = Sapph. fr. 110.3 Voigt)60 ἐξεπόνησαν DI : ἐξεπόνασαν A

We may draw some conclusions from what has been shown so far. On the one hand, in several cases manuscripts D and/or I of Hephaestion’s Encheiridion transmit dialectal variant readings which are surely the correct ones, even against manuscript A. On the other hand, when D and/or I do not transmit the correct form, they are only responsible for the trivialization of non-Attic forms into their more common Attic equivalents, so they never transmit forms which are likely to have been hyper-characterized by a learned scribe.

 58 See Schwyzer 1939, 660 n. 9; Buck 1955 § 138.a; Hamm 1958, § 249.a.1. 59 For εὔδης of manuscript A in no. (21) see ἔχῃς in no. (19) and the bibliographical references in n. 63. 60 It is worth considering that ἐξεπόνασαν is paralleled in Pindar, Euripides, and Theocritus (see Forssman 1966, 70−75). Be that as it may, it does not affect the present argument in any way.

164  Federico Favi Exactly the same trends apply to the transmission of non-Attic forms in Choeroboscus’ commentary, as one may see from the passages (23−30) below. When Choeroboscus’ witnesses do not transmit the correct variant readings, they are only responsible for trivializations, and never for hyper-characterizations (the only exception is ἔφα in (25)): (23) ὄρθιον ἑξαμερὲς τετόρων καὶ εἴκοσι μέτρων (Choer. In Heph. ench. 180.3 Consbruch = Musaeus fragmentum novum) τετόρων Alcid. 2.130 Avezzù et Longinus 85.9 Consbruch : τεττάρων codd. (24) πλάγχθη ἐπεὶ Τροίης (Choer. In Heph. ench. 187.18 Consbruch = Hom. Od. 1.2) Τροίης U : Τροίας K (25) τὼς μὲν ὁ Μνησάρχειος ἔφη ξένος, ὧδε συναινῶ (Choer. In Heph. ench. 201.22 Consbruch = Callim. Aet. fr. 61 Pfeiffer) τῶς K : πῶς U || ἔφη K : ἔφα U (26) ἄλλοι δὲ ῥινοῖς, ἄλλοι δ᾽ αὐτῇσι βόεσσιν (Choer. In Heph. ench. 206.15 Consbruch = Hom. Il. 7.474) αὐτῇσι Hom. : αὐτοῖσι KU (27) τρόχιμα βάσιμα χέρεσι πόδεσι (Choer. In Heph. ench. 217.12 Consbruch = Soph. fr. 240.2 Radt) χέρεσι K : χέρσι U (28) Κρονίδα βασιλῆος γένος Αἶαν, τὸν ἄριστον πέδ’ Ἀχίλλεα (Choer. In Heph. ench. 241.19 Consbruch = Alc. fr. 387 Voigt) πεδ’ Stephanus : παῖδα Ἀχιλλέως K (idem 241.22 Consbruch) : Αἶαν, τὸν ἄριστον πέδ’ Ἀχίλλεα om. U (29) ἀλλ]ὰ τυίδ’ ἔλ[θ’, αἴ ποτα κἀτέρωτα (Choer. In Heph. ench. 251.17 and 21 Consbruch = Sapph. fr. 1.5 Voigt) αἴποτα U : αἴποτε K (30) Μαῖα γέννατο Κρονίδαι μίγεισα (Choer. In Heph. ench. 252.19 Consbruch = Alc. fr. 308b.3 Voigt) Κρονίδᾳ U : Κρονίδη K

To my mind, all evidence collected and discussed so far makes it pretty clear that there is no convincing reason to think that the variant readings with vocalism in Epicharmus’ fragment cannot be correct. Finally, one further remark. There is no cause to wonder why πᾶσαν alone does not have an Aeolic variant reading such as παῖσαν. Beside the fact that this may be a mere matter of chance, we should recall that the Aeolic declension παῖς, παῖσα, πᾶν is foreign to Greek

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  165

poetic language outside Lesbian lyric.61 It is consequently very plausible that by using ἔχοισα and μοισικὰν on the one hand, and by avoiding παῖσαν on the other, Epicharmus chose to imitate the diction of choral lyric poetry rather than that of the Lesbian poets. If all the arguments advanced so far prove convincing, we finally have some reliable proof of Epicharmus’ interaction with choral lyric poetry.

. The Content of Epicharmus’ Fr. 80 K.-A. An exhaustive exegesis of this fragment is a difficult matter, not least because of its brevity. Nonetheless, it is very possible that the figure described by Epicharmus is the very hetaera from whom the comedy received its name. Indeed, the linguistic facies of the fragment shows a predominance of poetic vocabulary, and this would serve well for the description of a figure professionally versed in poetry and music. The fragment preserves the first known instance of the the use of μουσική as a noun (without τέχνη) together with Pind Ol. 1.15 and fr. 32 S.-M.62 The adjective εὔυμνος first appears as an epithet of Apollo in Hymn. Hom. 3.19 and 3.207, and is later to be found in Callim. fr. 229 Pfeiffer (an invocation to Zeus and Apollo), in Callim. Hymn. 2.31 (for Apollo), and Callim. Hymn. 4.4 (it is used for the Cyclades as they are sacred to Apollo). Its meaning is probably something like ‘celebrated in songs’, given that in the Homeric Hymns and in Callimachus’ passages the adjective means ‘perfect as a subject for a song’63 and ‘worthy of being much celebrated’.64 One should also take into consideration the fact that another compound containing ὕμνος, πολύυμνος (Hymn. Hom. 26.7, Ibyc. fr. 282a.6 PMG, Ar. Eq. 1328) means ‘famous’ and is applied to a πόρνη in Anac. fr. 101 PMG. This comparison may be revealing for the exegesis of εὔυμνος in Epicharm. fr. 80 K.-

 61 See Nöthiger 1971, 91. 62 There is no need to think that Epicharmus was echoing Pindar in this use of μουσική. Even though Pindar (and Ibycus, see fr. S255.4 Davies) only make use of μουσικ-, this does not rule out the possibility of having μοισικ- after Μοῖσα. The ‘art of the Muses’ is a key component of Greek culture as a whole and cannot be limited to one ethnic (or linguistic) group; among other things, one may think of μοισοπόλος (scil. δόμος) in Sapph. fr. 150.1 Voigt (see especially Ferrari 2010, 147; Ferrari 2014, 16; Neri in Neri/Cinti 2017, 401). One must also take into account the possibility that with μοισικὰν Epicharmus may have intentionally pursued a comic aim (see further Section 2.5 for μοισάω, which is likely to be a comic neologism). 63 Nagy 2009, 17–18. 64 Fernández-Galiano 1977, 268 (s.v. εὔυμνος).

166  Federico Favi A. too. As for φιλόλυρος, this hapax reminds us of other poetic compounds with a second element in -λυρος (one may think of ἄλυρος, ἀντίλυρος, and εὔλυρος).65 Given that Epicharm. fr. 80 K.-A. displays this distinctively poetic vocabulary, which is probably meant to parallel the figure who is described, it is reasonable that the variant readings with vocalism were adopted for the same purpose; the hetaera, as she is a professional of music, would aptly be characterized with the use of poetic forms in both phonology and vocabulary.66 This kind of linguistic mimesis is not foreign to Epicharmus’ fragments. Let us think of the use of μᾱνός with a 3rd compensatory lengthening in Epicharm. fr. 185 K.-A., for parodying pre-Socratic doctrines,67 and of παρῆσαν in [Epicharm.] fr. 275.1 K.-A.,68 an (Attic and) Ionic form that suits the philosophical content of the passage.69

 65 Respectively Soph. OC 1222, Eur. Alc. 447, IT 146, Hel. 185, Phoen. 1018; Soph. Trach. 642; Eur. fr. 477 Kannicht, Ar. Ran. 229. 66 It should clearly be stressed that none of these poeticisms is typical of lyric poetry. Thus, one may not assume that they influenced medieval scribes in writing μοισικὰν and ἔχοισα. 67 See West 1968, 200. 68 Against Kassel and Austin, I agree with those who take this fragment to be genuine (see Cassio 2002, 57 n. 18; Álvarez Salas 2007; Willi 2008, 123). 69 Cassio 2002, 57. Such considerations may also help explain a number of problematic Doris severior forms which sometimes appear in Epicharmus, given that the dialect of Syracuse belongs to the so-called Doris mitior (long-closed outcomes of compensatory lengthening and contractions), and the severior vocalism in Epicharm. frr. 32.12 and 35.3 K.-A. (τούτω), in Sophr. frr. 56 and 86 K.-A. (respectively, τῶ χρόνω and τῶ σκότεος), and Anon. Dor. fr. 20 K.-A. (Ἡρακλείτω τῶ Τεριναίω) has not been unequivocally explained (only Wilamowitz 1922, 100 and n. 1 had no problem admitting such forms for the dialect of Syracuse). Willi 2015, 117 n. 17 discusses Epicharm. fr. 35.2 K.-A. along the line of his own interpretation (see Willi 2012) of Theoc. 15.87−88 (in the idyll, the Syracusan women are accused of having a ‘large’ pronunciation, i.e. of being πλατειάσδοισαι), and thinks that the severiores forms were introduced in the Hellenistic manuscripts to enhance the Doric facies of the fragments in opposition to their more common koine equivalents. Something similar is suggested by Cassio 2004, 196 for Epicharm. fr. 32.12 K.-A. As for Anon. Dor. fr. 20 K.-A. (which is a 4troch.^, and the attribution to Epicharmus is far from implausible), however, Cassio 2002, 61–62 had suggested that the severiores forms are due to the subject matter. Indeed, one Heraclitus is said to come from Terina, a sub-colony of Kroton, where people spoke a Doris severior dialect. As a consequence, Cassio thinks that -ω genitives were either introduced in the text by a later editor to stress this provenance, or that they are Epicharmus’ own choice (Cassio rightly notes that in a Doris severior text one would also expect to find *Ἡρακλήτω instead of the transmitted Ἡρακλείτω: the linguistic characterization in the fragment, whether authentic or not, is not consistent).

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  167

. Status and Diffusion of -οισα Forms In Greek literature of the Archaic and Classical periods, forms in ‑οισα are relatively frequent only in choral lyric poetry (in addition to Sappho and Alcaeus).70 They originate from the Eastern Aeolic dialect (mainly Lesbian), where the diphthong is the regular outcome of 2nd compensatory lengthening.71 Doubts have been raised in the past as to the origin of the ‑οισα forms in lyric poetry, but it is now generally agreed that the explanations provided first by Ruijgh72 and then, in a revised and expanded form, by Cassio73 prove convincing. Forms with a diphthong are borrowings from the prestigious Aeolic tradition of lyric poetry. Their spread all over Greece was probably favoured by forms such as Μοῖσα and feminine personal names in -οισα (old participles), familiar in some form of catalogic literature. But beside choral lyric poetry, -οισα also appears together with other elements of Aeolic descent in hexametrical inscriptions of Corinthian and Boeotian provenance (as well as one from Attica).74 To these we should add a couple of new epigraphical finds from Magna Graecia, one of which is particularly significant, since it is an elegiac poem.75 Given the geographical and cultural scope of this evidence, Wachter is probably right in suggesting that these inscriptions show that Aeolicisms were used for hexametric poetry in areas where the influence of the Ionic epic tradition was less strong.76 Such conclusions have some bearing on Epicharmus’ fragment too. As already mentioned above, in the whole of Greek dramatic literature there is no other instance, beside Epicharm. fr. 80 K.-A., of the use of an Aeolic form

 70 In lyric poetry there are five types of forms with a diphthong as the outcome of 2nd compensatory lengthening (see Nöthiger 1971, 89−95 and Tribulato 2016, 237−241): the name Μοῖσα; feminine present participles ending in -οισα; feminine personal names in -οισα, which originally were participles as well (e.g. Ἀστυμέλοισα in Alcm. fr. 3.64−63 PMG) the 3rd person plural of the present indicative (-ονσι > -οισι vs. Attic -ουσι); only in Pindar does one also encounter sigmatic aorist participles in -αις (< -an(t)s). 71 See Hodot 1990, 74−75. 72 Ruijgh 1984, 87 and Ruijgh 1989, 89 (= Ruijgh 1996, 480). 73 Cassio 2005 (but see also Hinge 2006, 43−46). 74 See Wachter 2001, 340−341 (and Sider 2010 for the inscription on Douris’ vase). 75 The first is a vase inscription from Saturo (near Taranto), most likely a dedication to the Muses reading Μοισᾶν (gen. pl.): see Vallarino 2016. The second finding is the remarkable Tabula Cauloniensis, an elegiac poem on a lead tablet found in Caulonia (near Reggio Calabria). As this inscription has not been published by Carmine Ampolo and Emilio Rosamilia yet, I prefer not to print here the text of the line where the Aeolic form is found. 76 If this is true, Douris’ vase would be particularly remarkable given its Attic origin.

168  Federico Favi with -οισα except for Rhinth. fr. 6 K.-A. ἔχοισα καινὰν φαινόλαν † καπαρτίω. This line is taken from a comedy entitled Ἰφιγένεια ἁ ἐν Ταύροις, almost surely a parody of Euripides’ play even though we know almost nothing about its actual content.77 As in Epicharmus’ fragment, in Rhinthon, too, the Aeolicism is a feminine present participle (ἔχοισα). Rhinthon was largely influenced by Attic plays, but he also used the local Tarentine dialect and certain poeticisms, including ἔχοισα.78 This has broader implications. Indeed, it may not be by mere chance that -οισα only appears in Western dramatic literature, and especially in Doric cities such as Syracuse and Tarentum. This may well confirm what Wachter suggested with regard to the cultural and geographical diffusion of Aeolicisms in the standard poetic language, i.e. that it occurred in places where Ionic poetry, and especially Ionic epic, was a less compelling model. As for Epicharmus, it is interesting to compare a Homeric quotation from one of his plays that seems to have been ‘Doricized’, namely Epicharm. fr. 113.415 K.A. ἀφρ]άτωρ ἀθέμ[ιστος ἀ[νίστιος.79 This line has been discussed by Cassio,80 who envisaged the possibility that in Syracuse epic poetry was more strongly adapted to the local dialect than it was elsewhere. Even though in the case of the -οισα forms in Epicharm. fr. 80 K.-A. one certainly cannot think of an epichoric feature, it may nonetheless be argued that Epicharmus was familiar with a standard poetic literary language that was very different from the one that we usually take for granted, namely a language predominantly influenced by the Ionic dialect. If this is true, we may assume that in the eyes of Epicharmus the -οισα forms belonged to poetic language as a whole, i.e. that they are not necessarily just a lyric trait.81

. The Overlooked Testimony of a Syracusan Glossa One final piece of evidence confirms that the variant readings μοισικὰν and ἔχοισα in Epicharmus’ fragment should receive more detailed consideration than they have received so far. Indeed, in the Epimerismi Homerici there is an interesting Syracusan glossa which has been given surprisingly little attention: Epimer. Hom. μ 65 Dyck ἐκ τοῦ μοῦσα καὶ μοισῶ ῥῆμα παρὰ Συρρακουσίοις (= Gloss. Ital. no. 35 K.-A.).  77 One might argue that the female figure described by Rhinthon is Iphigeneia herself dressed as a Taurian priestess. For a discussion of this fragment, see Favi 2017a, 160−166. 78 See Favi 2017a, 68−72, 76−85, and 88−90. 79 The model line is Hom. Il. 9.63 ἀφρήτωρ ἀθέμιστος ἀνέστιός ἐστιν ἐκεῖνος. 80 Cassio 2002, 72−73. 81 See also Enzo Passa’s contribution in this volume (pp. 107–123).

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  169

The glossa is devoted to the word Μοῦσα. In the first part, the discussion focuses on the variant forms of Μοῦσα and on identifying each of them.82 In the second section, the focus is on linguistic issues. It begins with a list and analysis of the variant forms of Μοῦσα in the Greek dialects (μοισῶ appears there),83 and concludes with its etymology.84 The attribution to the Syracusans of the form μοισῶ requires a detailed discussion. 2nd compensatory lengthening resulting in a diphthong is foreign to the local dialect, as it is to all the other Greek dialects of Sicily.85 This makes it extremely unlikely that the attribution of μοισῶ to the Syracusan dialect relies on documentary evidence. A much more likely conclusion is that μοισῶ was originally used in some literary text of Syracusan origin, and that the ancient grammarians took this as evidence for the local dialect as a whole.86 If this is true, Syracusan comedy, and above all Epicharmus, is the most likely candidate for the use of this form, although this cannot be proven conclusively.87 A comic origin for μοισῶ is also likely on linguistic grounds. Kaibel88 has already suggested that “μοισᾶν ridicule is dici videtur qui a Musis captus versus fundere gestit”, comparing this form to λογάω.89 The latter is one of a number of denominative verbs, most of which are comic neologisms, whose ending -(ι)άω  82 Considerable space is devoted to the discussion of Mnaseas’ interpretation of the names of the Muses (Mnas. fr. 25a Müller). 83 In the first part of the glossa the discussion focuses on the nominative singular: μοῦσα is ascribed to Ῥηγῖνοι, Ἀττικοί, Ἴωνες, and Συρρακούσιοι; Μοῖσα to the Αἰολεῖς (scil. Eastern Aeolic, see Tosi 1988, 174); and μῶσα to the Λάκωνες, although οἱ μεταγενέστεροι Λάκωνες make use of μῶα (the locus classicus is most likely Ar. Lys. 1298). In the second part of the glossa the object of the discussion is the genitive plural: Μωσᾶ{ω}ν is attributed to the Λάκωνες and Μοισάων to Sappho (= fr. 187 Voigt); μουσέων is said to be Ἰωνικόν, μωσάων to be Βοιωτικόν, and μοισᾶν to be Ῥηγῖνον. Regarding the actual source for the information on the dialect of Rhegium, see below. 84 The origin of the word is traced back to μῶ, a synonym of ζητέω that is often used in the ancient etymological texts (see Philox. frr. 143−146, 148−149, and 641 Theodoridis and the loci similes collected ad locum). 85 Not by chance, Ahrens 1843, 106 n. 3 suggested changing μοισῶ into μουσῶ. 86 Cassio 1993 was the first to identify and explain this “sistema del dialetto-base”. A proof that this process is involved in the glossa is provided by the mention of the dialect of Rhegium, i.e. forms taken from poems by Ibycus (see Davies ad Ibyc. fr. 341ii Davies). 87 Epicharmus’ comedies were often used by ancient grammarians as sources of information on the dialect of Syracuse (see Epicharm. frr. 140 and 155 K.-A., and Philit. fr. 10 Dettori with the discussion provided by Dettori 2000, 90 n. 287 and 91 n. 291). 88 Kaibel 1899, 201 (followed without further comments in K.-A. I, 307). 89 Lucian. Lexiphanes 15: ἐμοὶ δὲ ἡ γλῶττά τε ἤδη λογᾷ. This is the explanation in LSJ s.v.: ‘to be fond of talking’.

170  Federico Favi denotes a state of illness or a desire for something which ultimately becomes pathological.90 The glossa μοισῶ may thus be compared to forms such as σωκρατάω (‘I desire to be Socrates’ > ‘I am ill of Socrates’)91 and especially δαιμονάω (‘I am possessed by the daemon’ > ‘I am mad’),92 although the latter is not a typically comic word. Given such comparisons, one feels inclined to agree with Kaibel that μοισῶ was originally used for parodying poetic inspiration,93 which was described as some form of μανία. If this is true, the vocalism would make the parodic intent of the neologism even more striking. The information we can rely on is quite scanty, and so certain questions might still be left open. However, to my mind it is almost certain that this Syracusan glossa was taken from a literary text, most likely a comedy by an author either of Syracusan descent or at least active in this city, and that it had a parodic function. Even if μοισῶ was not used in any of Epicharmus’ plays, this verb attests that para-Aeolic forms with vocalism were familiar to Syracusan comedy. Clearly, this makes a very strong case in favour of a better appreciation of the variant readings μοισικὰν and ἔχοισα in Epicharm. fr. 80 K.-A.

 Conclusions To sum up what has been argued so far, when re-evaluating the variant readings μοισικὰν and ἔχοισα in Epicharm. fr. 80 K.-A. one should take into consideration the following points: (a) they appear in the manuscripts of different works, and these manuscripts lack any relationship; (b) if both are mistakes, one would be compelled to explain them as being due to polygenesis, which is however quite unlikely, as they are both difficiliores; (c) it is unlikely that μοισικὰν and ἔχοισα are due to a learned scribe, since the copyists of the manuscripts where they are transmitted are never responsible for similar mistakes; on the other hand, forms with vocalism are difficiliores, and no internal or external reason might have led the scribe to write them;  90 See Peppler 1921, 154−156 and Willi 2003, 84−85. 91 See Ar. Av. 1282 with Dunbar 1995, 636. 92 Aesch. Sept. 1001, Cho. 566; Eur. Phoen. 888 (with Mastronarde 1994, 408) and fr. 122.1054 Kannicht; Xen. Mem. 1.1.9; Men. fr. 77.1 K.-A. 93 This is likely, even though in *μοισάω the -α- is not part of the suffix as in the case of the verbs in -(ι)άω.

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  171

(d) forms with vocalism suit the linguistic register and content of the fragment; (e) in lyric poetry, -οισα forms are mostly represented by Μοῖσα and feminine present participles, exactly the kind of forms that appear in Epicharmus’ fragment; (f) the Syracusan glossa μοισῶ would seem to testify to the use of para-Aeolicisms in Syracusan comedy, maybe specifically in one of Epicharmus’ plays. Editors are free to decide whether they feel confident enough to print μοισικὰν and ἔχοισα in Epicharm. fr. 80 K.-A. To my mind, there is no reason to believe that these variant readings should not be genuine, and I would be quite happy to print them in the text of the fragment. Moreover, they would contribute to a better appreciation of Epicharmus’ art, insofar as they would demonstrate, alongside other pieces of evidence,94 the use of a richer spectrum of linguistic and stylistic resources than this poet is often credited with.

Bibliography Ahrens, H.L. (1843), De Graecae linguae dialectis, vol. 2: De dialecto Dorica, Lipsiae. Álvarez Salas, O. (2007), “I frammenti filosofici di Epicarmo: Una rivisitazione critica”, in: SIFC 5, 23−72. Arnott, W.G. (1996), Alexis: The Fragments. A Commentary, Cambridge. Arnson Svarlien, D. (1990−1991), “Epicharmus and Pindar at Hieron’s Court”, in: Kokalos 36−37, 103−110. Austin, C./Olson, S.D. (2004), Aristophanes: Thesmophoriazusae, Oxford. Bees, R. (1993), Zur Datierung des Prometheus Desmotes, Stuttgart. Bettarini, L. (2005), Corpus delle defixiones di Selinunte: Edizione e commento, Alessandria. Bravi, L. (2006), Gli epigrammi di Simonide e le vie della tradizione, Roma. Buck, C.D. (1955), The Greek Dialects, Chicago. Cassio, A.C. (1993), “Parlate locali, dialetti delle stirpi e fonti letterarie nei grammatici greci”, in: E. Crespo/J.L. García Ramón/A. Striano (eds.), Dialectologica Graeca. Actas del II Coloquio Internacional de Dialectología Griega (Miraflores de la Sierra [Madrid], 19–20 de junio de 1991), Madrid, 73−90. Cassio, A.C. (2002), “The Language of Doric Comedy”, in: A. Willi (ed.), The Language of Greek Comedy, Oxford, 51−83. Cassio, A.C. (2004), “Review of R. Kassel/C. Austin, Poetae comici Graeci, vol. 1 (Berolini/Novi Eboraci 2001)”, in: Gnomon 76, 193−198.

 94 See above Section 2.3 with n. 76 and Section 2.5.

172  Federico Favi Cassio, A.C. (2005), “I dialetti eolici e la lingua della lirica corale”, in: F. Bertolini/F. Gasti (eds.), Dialetti e lingue letterarie nella Grecia arcaica, Como/Pavia, 13−44. Chantraine, P. (1933), La formation des noms en grec ancien, Paris. de Cremoux, A. (2011), “La maxime chez Epicarme et la naissance de la comédie: Problèmes de métode et pistes de réflexion”, in: C. Mauduit/P. Raré-Rey (eds.), Les maximes théâtrales en Grèce et à Rome: Transferts, réécritures, remplois, Lyon, 55−68. Dettori, E. (2000), Filita grammatico: Testimonianze e frammenti, Roma. Dunbar, N. (1995), Aristophanes: Birds, Oxford. Favi, F. (2017a), Fliaci: Testimonianze e frammenti, Heidelberg. Favi, F. (2017b), “Lo Ὀδυσσεὺς αὐτόμολος di Epicarmo”, in: ZPE 201, 17−31. Fernández-Galiano, E. (1977), Léxico de los himnos de Calímaco, vol. 2: E–I, Madrid. Ferrari, F. (2010), Sappho’s Gift: The Poet and Her Community. Translated by B. Acosta-Hughes/ L. Prauscello, Ann Arbor. Ferrari, F. (2014), “Saffo e i suoi fratelli e altri brani del primo libro”, in: ZPE 192, 1–19. Flintoff, E. (1986), “The Date of the Prometheus Bound”, in: Mnemosyne 39, 82−91. Forssman, B. (1966), Untersuchungen zur Sprache Pindars, Wiesbaden. Fraenkel, E. (1950), Aeschylus: Agamemnon, 3 vols., Oxford. Gentili, B. et al. (1995), Pindaro: Le Pitiche. Introduzione, testo critico e traduzione a c. di B. Gentili. Commento a c. di P.A. Bernardini/E. Cingano/B. Gentili/P. Giannini, Milano. Gentili, B. et al. (2013), Pindaro. Le Olimpiche. Introduzione, testo critico e traduzione a c. di B. Gentili. Commento a c. di C. Catenacci/P. Giannini/L. Lomiento, Milano. Gerber, D.E. (1982), Pindar’s Olympian One: A Commentary, Toronto. Hamm, E.-M. (1958), Grammatik zu Sappho und Alkaios, Berlin. Hartwig, A. (2009), “Feasting and Flattering the Crows: A Reinterpretation of Aristophanes, Thesmophoriazusae 942”, in: Mnemosyne 62, 90−93. Hinge, G. (2006), Die Sprache Alkmans: Textgeschichte und Sprachgeschichte, Wiesbaden. Hodot, R. (1990), Le dialecte éolien d’Asie: La langue des inscriptions. VIIe s. a.C. − IVe s. p.C., Paris. Hoerschelmann, G. (1886), “Exegesis in Hephaestionis Enchiridion”, in: W. Studemund (ed.), Anecdota varia Graeca et Latina, vol. 1: Musica, metrica, grammatica, Berolini, 31−96. Kaibel, G. (1899), Comicorum Graecorum fragmenta, vol. 1, Berolini. Kannicht, R. (1969), Euripides: Helena, vol. 2: Kommentar, Heidelberg. Kerkhof, R. (2001), Dorische Posse, Epicharm und attische Komödie, München/Leipzig. Kowalzig, B. (2008), “Nothing to Do with Demeter? Something to Do with Sicily! Theatre and Society in the Early Fifth-Century West”, in: M. Revermann/P.J. Wilson (eds.), Performance, Iconography, Reception: Studies in Honour of Oliver Taplin, Oxford, 128−157. Locker, E. (1933), “Die Bildung der griechischen Kurz- und Kosenamen I”, in: Glotta 21, 136−153. Lorenz, A.O.F. (1864), Leben und Schriften des Koers Epicharmos: Nebst Fragmentensammlung, Berlin. Mann, C. (2001), Athlet und Polis im archaischen und frühklassischen Griechenland, Göttingen. Martinelli, M.C. (1997), Gli strumenti del poeta, 2nd edn., Bologna. Mastronarde, D.J. (1994), Euripides: Phoenissae, Cambridge. Mimbrera, S. (2012), “Sicilian Greek before the Fourth Century BC”, in: O. Tribulato (ed.), Language and Linguistic Contact in Ancient Sicily, Cambridge, 191−222. Molyneux, J.H. (1992), Simonides: A Historical Study, Wauconda.

Epicharmus and Choral Lyric Poetry: A Reappraisal of Old and New Evidence  173

Morgan, K.A. (2015), Pindar and the Construction of Syracusan Monarchy in the Fifth Century B.C., Oxford. Most, G.W. (1985), The Measures of Praise: Structure and Function in Pindar’s Second Pythian and Seventh Nemean Odes, Göttingen. Nagy, G. (2009), “Perfecting the Hymn in the Homeric ‘Hymn to Apollo’”, in: R.P. Martin/L. Athanassaki/J.F. Miller (eds.), Apolline Politics and Poetics, Athens, 17−42. Neri, C./Cinti, F. (2017), Saffo: Poesie, frammenti e testimonianze, Santarcangelo di Romagna. Nöthiger, M. (1971), Die Sprache des Stesichorus und des Ibycus, Zürich. Olivieri, A. (1946), Frammenti della commedia greca e del suo mimo nella Sicilia e nella Magna Grecia, vol. 1, 2nd edn., Napoli. Olson, S.D. (2007), Broken Laughter: Select Fragments of Greek Comedy, Oxford. Olson, S.D. (2017), Eupolis: Einleitung, Testimonia und Aiges – Damoi (frr. 1−146), Heidelberg. Page, D.L. (1981), Further Greek Epigrams, Cambridge. Parker, L.P.E. (2007), Euripides: Alcestis, Oxford. Peppler, C.W. (1921), “Comic Terminations in Aristophanes: Part V”, in: AJPh 40, 152−161. Pickard-Cambridge, A.W./Webster, T.B.L. (1962), Dithyramb, Tragedy and Comedy, 2nd edn., Oxford. Piva, G. (2011), “Le vicende di Eracle nel teatro di Epicarmo: Osservazioni sul fr. 92 K.-A.”, in: A.M. Andrisano (ed.), Ritmo, parola, immagine: Il teatro classico e la sua tradizione, Palermo, 103−113. Polman Kruseman, H. (1834), Epicharmi fragmenta, Harlem. Prato, C. (2001), Aristofane: Tesmoforiazuse. Traduzione di D. Del Corno, Milano. Race, W.H. (1997), Pindar: Olympian Odes. Pythian Odes, Cambridge, Mass./London. Rawles, R.J. (2018), Simonides the Poet: Intertextuality and Reception, Cambridge. Revermann, M. (2013), “Paraepic Comedy: Point(s) and Practices”, in: E. Bakola/L. Prauscello/ M. Telò (eds.), Greek Comedy and the Discourse of Genres, Cambridge/New York, 101−128. Richardson, N. (1993), The Iliad: A Commentary, vol. 6: Books 21−24, Cambridge. Rodríguez-Noriega Guillén, L. (1996), Epicarmo de Siracusa: Testimonios y fragmentos, Oviedo. Rodríguez-Noriega Guillén, L. (2012), “On Epicharmus’ Literary and Philosophical Background”, in: K. Bosher (ed.), Theater Outside Athens: Drama in Greek Sicily and South Italy, Cambridge, 76−96. Ruijgh, C.J. (1984), “Review of E. Risch, Kleine Schriften (Berlin 1981)”, in: Kratylos 29, 86−92. Ruijgh, C.J. (1989), “Review of C. Trümpy, Vergleich des Mykenischen mit der Sprache der Chorlyrik: Bewahrt die Chorlyrik eine von Homer unabhängige alte Sprachtradition? (Bern 1986)”, in: Kratylos 34, 85−91. Ruijgh, C.J. (1996), Scripta minora ad linguam Graecam pertinentia, vol. 2, Amstelodami. Rusten, J.S. (2011), The Birth of Comedy: Texts, Documents, and Art from Athenian Comic Competitions, 486–290. Edited by J. Rusten. Translated by J. Henderson/D. Konstan/R. Rosen/ J. Rusten/N. Slater, Baltimore. Schwyzer, E. (1939), Griechische Grammatik, vol. 1: Allgemeiner Teil, Laulehre, Wortbildung, Flexion, München. Sider, D. (2010), “Greek Verse on a Vase by Douris”, in: Hesperia 79, 541−554. Sommerstein, A.H. (2001), Aristophanes: Thesmophoriazusae, Warminster. Spanheim, E. (1697), Callimachi Hymni, epigrammata et fragmenta ex recensione Theodori J.G.F. Graevii, cum ejusdem animadversionibus. Accedunt N. Frischlini, H. Stephani, B.

174  Federico Favi Vulcanii, P. Voetii, A.T.F. Daceriae, R. Bentleii commentarius, et annotationes viri illustrissimi Ezechielis Spanhemii, nec non praeter fragmenta, quae ante Vulcanius & Daceria publicarant, nova, quae Spanhemius & Bentleius collegerunt, & digesserunt. Hujus cura & studio quaedam quoque inedita epigrammata Callimachi nunc primum in lucem prodeunt, 2 vols., Utrecht. Tosi, R. (1988), Studi sulla tradizione indiretta dei classici greci, Bologna. Tribulato, O. (2016), “La lirica corale”, in: A.C. Cassio (ed.), Storia delle lingue letterarie greche, 2nd edn., Milano, 230−259. Ucciardello, G. (2013), “Nota di lettura a Epich. fr. 113, 187 K.-A. (= P.Oxy. XXV 2427, fr. 20, 5)”, in: APF 59, 277–279. Vallarino, G. (2016), “Muse a Satúro: Nuovi dati su un culto delle Muse in area tarantina”, in: G. De Sensi Sestito/M. Intrieri (eds.), Sulle sponde dello Ionio: Grecia occidentale e Greci d’Occidente. Atti del convegno di studi internazionale, Pisa, 441−451. Venezia, L. (2007), “L’Olimpica XII di Pindaro e la Hesychia”, in: G. Daverio Rocchi (ed.), Tra concordia e pace: Parole e valori della Grecia antica, Milano, 267−279. von Salis, A. (1905), De Doriensium ludorum in comoedia Attica vestigiis, Basel. Wachter, R. (2011), Non-Attic Greek Vase Inscriptions, Oxford. Weber, G. (1992), “Poesie und Poeten an den Höfen vorhellenistischer Monarchen”, in: Klio 74, 25−77. Welcker, F.G. (1844), “Epicharmus”, in: Kleine Schriften, vol. 1, Bonn, 271−356. West, M.L. (1968), “Notes on Newly-Discovered Fragments of Greek Authors”, in: Maia 20, 195−205. West, M.L. (1982), Greek Metre, Oxford. Wilamowitz-Moellendorff, U. von (1922), Pindaros, Berlin. Willi, A. (2003), The Languages of Aristophanes: Aspects of Linguistic Variation in Classical Attic Greek, Oxford. Willi, A. (2008), Sikelismos: Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.−5. Jh. v. Chr.), Basel. Willi, A. (2012), “‘We Speak Peloponnesian’: Tradition and Linguistic Identity in Post-Classical Sicilian Literature”, in: O. Tribulato (ed.), Language and Linguistic Contact in Ancient Sicily, Cambridge, 265−288. Willi, A. (2015), “Epicharmus, the Pseudepicharmeia, and the Origins of Attic Drama”, in: S. Chronopoulos/C. Orth (eds.), Fragmente einer Geschichte der griechischen Komödie: Fragmentary History of Greek Comedy, Heidelberg, 109−145. Wilson, P./Favi, F. (2017), “Choragic Spells in Gela: A Textual and Exegetical Note on Apellis’ defixio, SEG LVII 905b”, in: ZPE 204, 138−140. Woodbury, L. (1945), “The Epilogue of Pindar’s Second Pythian”, in: TAPA 76, 11−30.

Alessandro Vatri

Early Dactylic Prose in the History of Greek Prose Rhythm Abstract: 6th- and 5th-century Ionic prose fragments contain a remarkable number of sequences that modern scholars have identified as snippets of dactylic rhythm. Such rhythms are more likely to have been deliberate in texts that were primarily meant for oral recitation, and stylistic and metrical considerations suggest that at least some patterns must not have been accidental. At the same time, it is striking that virtually no ancient discussion of Greek prose rhythm makes explicit reference to dactylic sequences in Ionic prose — not even as an unsuitable type of rhythmic composition. If we accept that the early theory of rhythmic prose (as surfaces in Aristotle) was based on the empirical perception of rhythms in the acoustic flow of speech rather than on the identification of metrical patterns ‘on paper’, the fact that Ionic prose seems to have been largely ignored could depend on the exclusively written circulation (and lack of oral recitation) of these texts by the 4th century. Keywords: Prose rhythm, Ionic prose, Greek rhetoric.

 Introduction The earliest surviving mentions of rhythm as a feature of Greek prose date to the 4th century BC,1 and the writers who are traditionally said to have developed a specific interest in it were all active in the late 5th century BC.2 However, rhythmic sequences can be identified in the very earliest fragments of Greek prose. The extant passages of early Ionic prose contain several relatively extended dactylic/anapaestic sequences,3 which give the impression of a quasi-metrical mode of composition. This would violate both a tenet of the ‘standard’ ancient theory of prose rhythm — the principle that prose should be rhythmical, not metrical,4 which implies that it  1 Alcidamas Soph. 16–17, Isoc. 5.27; cf. Dover 1997, 172–173. 2 See McCabe 1981, 10–13. 3 See e.g. Norden 1898, 44–45; Deichgräber 1963; Lilja 1968; Dover 1997, 160–165; Usher 2010, 82–83, and most recently Gheerbrant 2018. 4 See e.g. Arist. Rh. 3.8 1408b.21–32; Theophr. fr. 701 Fortenbaugh (= Cic. De or. 3.184–187); Demetr. Eloc. 180; Dion. Hal. Comp. 25.10 and Dem. 50; [Longinus] Subl. 41. The same idea surfaces in Isocrates’ statement (9.10–11) that poets, unlike prose writers, are allowed to embellish their https://doi.org/10.1515/9783110621747-009

176  Alessandro Vatri should not contain complete verse cola — and the idea that hexametric/dactylic rhythms are not suited to prose composition.5 In this connection, it is worth noting that the rhythm of early Ionic prose was not picked up by ancient critics. As far as our evidence suggests, the dactylic prose of Herodotus’ predecessors and contemporaries did not make it into the ancient ‘official’ history and theory of Greek prose rhythm. This contribution will present some thoughts on the possible reasons for this apparent omission, with the caution demanded by any attempt at an argument from silence.

 Dactylic Sequences in Early Prose Snippets of dactylic/anapaestic rhythms have been identified in early Ionic prose by several classicists since the 19th century. The list below is based on Lilja’s survey (1968, 19–31) — which itself draws on previous works — with a few additions.6 Pherecydes of Syrus:7 (1)

Ζὰς8 μὲν καὶ Χρόνος ἦσαν ἀεί9 (B 1 D.-K.) –––⏑⏑–⏑⏑–

 style with rhythms and metre (cf. McCabe 1981, 13–14; O’Sullivan 1992, 50–52; Blass 1901, 1–3). In Latin, see e.g. Quint. Inst. 9.4.72; cf. Primmer 1968, 33. 5 Cf. Arist. Rh. 3.8 1408b.32–1409a.2; Demetr. Eloc. 42. 6 For the sake of consistency, sequence-final diphthongs and closed syllables containing a short vowel are scanned long unless a vowel follows in the text (if transmitted). 7 See Gheerbrant 2018 for a recent discussion of these sequences. 8 According to Chantraine (DELG s.v. Ζεύς), Ζάς is a “compromis entre Ζᾱ́ν et Ζεύς”, with Ζάν being a nominative built on the accusative Ζῆνα. Both Chantraine and Beekes (EDG s.v. Ζεύς) report the idea that forms in ᾱ spread from Zeus’ sanctuary in Elean Olympia, where [ɛ:] > [a:] and where it is attested that statues of Zeus were called Ζᾶνες (Paus. 5.21.2). See also Pavese 1998, 79 n. 20. West 1971, 50–52, instead, connects this theonym to Luwian deity Šanta, whom the Hittites identified with their weather-god. West remarks that the -ντ- element in the stem Ζαντ(cf. Pherec. of Syrus, A 8 D.-K.; see West 1971, 10; Schibli 1990, 17 n. 8) is completely unrelated to Ζεύς, and that the use of Ζάς need not suggest that a word play on the word for ‘to live’ was intended (Ζήν, Ζηνός would have been an obvious – and available – option). 9 The initial α in ἀεί is scanned short; Dover 1997, 160 n. 1 is rightly cautious about this, as the short-vowel form would be Homeric, whereas in Ionic αἰεί would be expected (cf. Lilja 1968, 31; Schwyzer 1939, 266). One could argue that a Homeric form may well be expected in a hexametric sequence, but the argument would fall into circularity. Gheerbrant 2018, 375 supports the reading ἀεί on the basis of two textually unproblematic passages of Herodotus (1.67.2, 6.52.27; but αἰεί occurs 110 times) and the fact that ἀεί is the only form attested in Heraclitus.

Early Dactylic Prose in the History of Greek Prose Rhythm  177

(2)

καὶ Χθονίη· Χθονίῃ δέ (B 1 D.-K.)

(3)

καὶ χρήματα καὶ θεράποντας (B 2 D.-K.)

(4)

καὶ θεραπαίνας καὶ τἆλλ(α) ὅσα δεῖ (B 2 D.-K.)10

(5)

Ζὰς ποιεῖ φᾶρος μέγα (B 2 D.-K.)

(6)

θεοῖσι καὶ ἀνθρώποισιν (B 2 D.-K.)11

–⏑⏑–⏑⏑–⏑ ––⏑⏑–⏑⏑–– –⏑⏑––––⏑⏑– –––––⏑⏑ ⏑–⏑⏑––––

Pherecydes of Athens: (7)

Ὠκεανὸς πειρώμενος αὐτοῦ κυμαίνει τὸ δέπας φανταζόμενος (FGrHist 3 F 1)

(8)

θητεύσων ἐνιαυτόν (FGrHist 3 F 35)

(9)

κατέρχεται εἰς Δελφοὺς περὶ παίδων χρησόμενος (FGrHist 3 F 64)12

(10)

εἰς τὴν Οἰχαλίαν (FGrHist 3 F 82)

(11)

πᾶσι παρεῖναι (FGrHist 3 F 105)

–⏑⏑–––⏑⏑–––––⏑⏑–––⏑⏑– –––⏑⏑–– ⏑–⏑⏑–––⏑⏑–––⏑⏑– –––⏑⏑– –⏑⏑–⏑

Hecataeus of Miletus: (12)

ἀληθέα εἶναι (FGrHist 1 F 1)13

(13)

Ψωφιδίους κακὰ πόλλ’ ἔοργεν (FGrHist 1 F 6)

(14)

ἐν δ’ αὐτοῖσι πόλις (FGrHist 1 F 163, 282, 293)

⏑–⏑⏑–– –⏑⏑–⏑⏑–⏑–– –––⏑⏑–

Acusilaus of Argos: (15)

Ὠκεανὸς δὲ γαμεῖ (FGrHist 2 F 1) –⏑⏑–⏑⏑–

 10 Sequences (3) and (4) are presented here as in Dover 1997, 160. 11 With shortening of the final diphthong before a vowel (καὶ ἀν-). 12 With shortening of the final diphthong before a vowel (κατέρχεται εἰς). 13 With hiatus; εἶναι is followed by a pause and -αι is scanned long.

178  Alessandro Vatri (16)

ἰσχὺν ἔχοντα [με]γί[σ]τ[η]ν τῶν ἀνθρώπων τῶν τότε (FGrHist 2 F 22) –⏑⏑–⏑⏑–––––––⏑⏑

Charon of Lampsacus: (17)

ᾤχετο φεύγων ἄρτι μὲν εἰς Μυτιλήνην (FGrHist 262 F 9) –⏑⏑–––⏑⏑–⏑⏑––

(18) ταῦτα δὲ ποιήσαντες (FGrHist 262 F 10)14 –⏑⏑–––⏑

Hellanicus of Lesbos: (19)

ἐν μακάρων νήσοις (FGrHist 4 F 19) –⏑⏑–––

(20) ἀνδράσιν εἰδόμενοι (FGrHist 4 F 26)15 –⏑⏑–⏑⏑⏑ (21)

ἄκρα τὰ δένδρα (FGrHist 4 F 197)16 –⏑⏑–⏑

Heraclitus: (22)

γινομένων γὰρ πάντων (B 1 D.-K.) –⏑⏑––––

(23)

καὶ ἐπέων καὶ ἔργων (B 1 D.-K.)17

(24)

εὖρος ποδὸς ἀνθρωπείου (B 3 D.-K.)18

–⏑⏑–––– ––⏑⏑–––– (25)

ἥρωας οἵτινές εἰσι (B 5 D.-K.)19

(26)

ἂν διαγνοῖεν (B 7 D.-K.)20

–⏑⏑–⏑⏑–⏑ –––– (27)

οὐ γὰρ φρονέουσι τοιαῦτα (B 17 D.-K.) ––⏑⏑–⏑⏑–⏑

 14 ποιήσαντες is followed by ἐπαναχωροῦσιν, therefore -ες is scanned short. 15 εἰδόμενοι is followed by ἐπί, therefore -οι is scanned short. 16 With no correptio Attica. 17 With hiatus. 18 With no shortening of -ει- before -ου. 19 With shortening of -ω- (cf. e.g. Norden 1898, 44; Bechtel 1924, 21; Deichgräber 1963, 528). 20 With semivocalic /i/ in δια- (cf. Norden 1898, 44).

Early Dactylic Prose in the History of Greek Prose Rhythm  179

(28)

ἅπτεται εὕδοντος (B 26 D.-K.)21

(29)

ἀντὶ ἁπάντων (B 29 D.-K.)22

–⏑⏑––– –⏑⏑–– (30) αὐτὸν ἁπάντων (B 30 D.-K.) –⏑⏑–– (31)

τὸ μὲν ἥμισυ γῆ, τὸ δὲ ἥμισυ πρηστήρ (B 31 D.-K.)23

(32)

ἓν τὸ σοφὸν μοῦνον (B 32 D.-K.)

⏑⏑–⏑⏑–⏑⏑–⏑⏑–– –⏑⏑––– (33)

μαρτυρ‹έ›ει παρεόντας ἀπεῖναι (B 34 D.-K. as adapted by Norden)24 –⏑⏑–⏑⏑–⏑⏑––

(34) εὖ μάλα πολλῶν (B 35 D.-K.) –⏑⏑–– (35)

ψυχῇσιν θάνατος (B 36 D.-K.)25 –––⏑⏑⏑

(36) πλείων λόγος ἢ τῶν ἄλλων (B 39 D.-K.) ––⏑⏑–––– (37)

καὶ Ἑκαταῖον (B 40 D.-K.)26 –⏑⏑––

(38) οὐ ξυνιᾶσιν ὅκως (B 51 D.-K.) –⏑⏑–⏑⏑– (39) παλίντροπος ἁρμονίη (B 51 D.-K.) ⏑–⏑⏑–⏑⏑– (40) ἁρμονίη ἀφανὴς φανερῆς κρείττων (B 54 D.-K.)27 –⏑⏑–⏑⏑–⏑⏑––– (41) ἰχθύσι μὲν πότιμον (B 61 D.-K.) –⏑⏑–⏑⏑–

 21 Sequences (27) and (28) are discussed by Dover 1997, 161, with doubts regarding the shortening of -αι (in hiatus) in (28) and the prosody of -εου- and τοι- in (27). 22 With hiatus. 23 With hiatus. 24 The spelling μαρτυρέει is presented in Norden 1898, 44 but has not been adopted by any editor. Dover 1997, 161 “greatly doubts” that this spelling was the original one. 25 θάνατος is followed by ὕδωρ, therefore -ος is scanned short. 26 With hiatus. 27 With hiatus.

180  Alessandro Vatri (42)

ἀνθρώποις δὲ ἄποτον καὶ ὀλέθριον (B 61 D.-K.)28 –––⏑⏑–⏑⏑–⏑–

(43) παῖδας τοκεώνων (B 74 D.-K.) ––⏑⏑–– (44) ζῆν ἡμᾶς τὸν ἐκείνων (B 77 D.-K.) –––⏑⏑–– (45) εἰ μὴ ἥλιος ἦν (B 99 D.-K.)29 –––⏑⏑– (46) μάρτυρες ἀνθρώποισιν (B 107 D.-K.) –⏑⏑–––⏑ (47) ἄγεται ὑπὸ παιδὸς ἀνήβου (B 117 D.-K.)30 ⏑⏑–⏑⏑–⏑⏑––

As the footnotes to this list detail, a number of arbitrary decisions are to be made in scansion.31 Discounting sequences that contain ambiguous syllables, a good number of snippets lend themselves to being scanned as sustained series of dactylic/anapaestic feet. In order to assess the position of these (quasi-)metrical patterns in the history of Greek prose rhythm, we need to start from the difficult question of their intentionality.32 This question raises, in turn, the issue of the relationship between the texts containing such sequences and possible contexts of oral performance and reception. In principle, metrical patterns are more likely to have been intentionally sought in texts that were meant to be recited to an audience rather than read privately (whether silently or aloud). Both contexts may be envisaged for early prose texts, which date to a period (mid 6th–mid 5th century BC) in which the use of writing was developing from an aid to composition into a primary medium of transmission,33 but both the circulation of books and the number of individuals who were literate enough to read them were still very limited in the Greek world.34

 28 Sequences (41) and (42) from Dover 1997, 160. Sequence (42) is scanned with δέ in scriptio plena and no correptio Attica of the syllable -λεθ-. 29 With hiatus. 30 Noticed by Dover 1997, 161; with hiatus and no shortening of the final diphthong -αι before a vowel. 31 Cf. Dover 1997, 160–161. 32 See e.g. Dräger 1995, 12 with further references. 33 See Pfeiffer 1968, 24–26; West 1971, 5. 34 Cf. Nieddu 1984, 214–215; Thomas 1992, 8, and Thomas 2009 on different levels of literacy. See also Werner 2009 for a review.

Early Dactylic Prose in the History of Greek Prose Rhythm  181

Unfortunately, the available evidence as to whether early prose texts were meant for public recitation or private (silent or oral) reading is minimal and mainly internal to the texts themselves. External evidence is mostly anecdotal. Diogenes Laertius (9.6), for instance, reports that Heraclitus (floruit ca. 500 BC) dedicated a copy of his book in the temple of Artemis at Ephesus and that the book later gave rise to a community of followers, the Heracliteans, whose bestknown representative was Cratylus.35 Even though the text of Heraclitus was transcribed and consciously turned into a book by the author himself, its text may have originated as a collection of oral maxims36 to be read (aloud) by his followers. A similar connection with oral performance may perhaps be envisaged for the book of Pherecydes of Syrus, another archaic ‘holy man’ and — according to tradition — (one of) the earliest Greek prose writer(s).37 Pherecydes (floruit ca. 544–541 BC),38 presumably a travelling teacher,39 may have committed his prose theogony to writing as a συγγραφή — a script for himself as an aid to oral performance.40 The subject matter of this work belonged to the domain of hexameter poetry; however, Pherecydes’ account is strikingly unorthodox from the point of view of Greek traditional religion.41 Consequently, the fact that Pherecydes chose prose as a medium can been interpreted both in light of its content — as a way to mark the difference between the poetic tradition and a new, ‘veridical’ (albeit not at all ‘scientific’) approach to the subject42 — and the possible occasions for the publication (as oral performance) of his text, not ‘mass’ religious festivals nor poetic contests, but ‘niche’ public lectures.43 While it is not hard to envisage how wisdom texts such as those of Pherecydes of Syrus and Heraclitus could be conceived for oral communication, such a connection is not easy to establish if we turn to other early prose texts. These represent a variety of genres. The Histories of Pherecydes of Athens — whose biographical details are all but unknown — consisted of mythical genealogies and

 35 See Steiner 1994, 88–89 and 202–203. 36 See Most 1999, 349–350 and Granger 2004 with further references. 37 With Anaximander being another candidate; see von Fritz 1938; West 1971, 3–4; Toye 1997; Fowler 1999. 38 Cf. most recently Gheerbrant 2018, 369 with further references. 39 According to tradition, he counted Pythagoras among his pupils; see Schibli 1990, 1–4. 40 West 1971, 4–6; Bernabé 1979, 302. 41 See e.g. Granger 2007. 42 Cf. Kahn 2003, 143–144. 43 Cf. Kahn 2003, 152 and Granger 2007, 425–428.

182  Alessandro Vatri chronologies,44 and have been dated to the 470s.45 Hecataeus of Miletus is a far better documented historical figure; he was a “leading figure in the Ionian Revolt (499–494 BC)”46 and the author of Genealogies and a Periodos. Acusilaus of Argos was dated by ancient authorities to the early 5th century:47 his book contained theogonic and genealogical materials and was regarded in Antiquity — as Fowler (2013, 623–627) describes — as a prose response to Hesiod’s Theogony and Catalogue. Charon of Lampsachus was a local historian and ethnographer and, seemingly, he was roughly contemporary to Herodotus.48 The same should be the case with Hellanicus of Lesbos, who is reported to have lived virtually throughout the 5th century and to have written mythographic and ethnographic works, as well as local histories.49 In the late 6th and 5th centuries, the different genres used by these intellectuals would probably have corresponded to different contexts of performance and reception. Historical narratives might have been the subject of public readings, as the traditional accounts of Herodotus’ ‘lectures’ suggest.50 Mythographic and genealogical texts, on the other hand, are more likely to have been conceived from the start as texts to be perused in private by a “small minority of highly literate Greeks.”51 These works intended to unify and ‘tidy up’ the competing traditions that circulated in the Greek world:52 in Fowler’s words, the purpose of mythography as a prose genre was “to communicate information to audiences at a distance. Unlike poetry, no occasion was required for its consumption. The books were meant to be read by people who wanted to know what was in them”,53 and such people would probably include other writers and poets.54 Thus, the occasions for the reception of such prose texts were private, not civic.55 Geographical works, on the other hand, may have originated for practical purposes and be rooted in sea trade and naval warfare. Merchants and sailors would gather empirical information about sea routes, ports, and landmarks. This would give rise

 44 See Fowler 2013, 706–727. 45 Dolcetti 2004, 6, 13, and 16. 46 Fowler 2013, 659; see Hdt. 5.36.2. 47 See Fowler 2013, 623–624 for a detailed discussion. 48 Fowler 2013, 641. 49 Fowler 2013, 682–687. 50 See e.g. Thomas 2000, 257 and Fowler 2001, 107. 51 Fowler 2016, 47. 52 Cf. Fowler 2013, 668 on Hecataeus. 53 Fowler 2013, 707. Cf. also Fowler 2013, 668. 54 Cf. Fowler 2016, 49–50. 55 Fowler 2016, 49.

Early Dactylic Prose in the History of Greek Prose Rhythm  183

to navigation guides that were transmitted orally56 and eventually transcribed as periploi — another genre that would not have been conceived for public recitation but for private perusal. In such a cultural landscape, we can surmise that (quasi-)metrical sequences might have been more at home in wisdom texts or historical narrative, and less so in genealogical and geographical texts. Nevertheless, this should not lead us to conclude that (quasi-)metrical sequences must have been deliberate (and not random) in texts conceivably composed for oral performance, and that the opposite should apply to texts primarily composed for written circulation. The question of the intentionality or randomness of (quasi-)metrical sequences requires evaluation of individual sequences. In principle, such an assessment could be carried out either quantitatively or qualitatively. Quantitative assessments would need to take into account the statistical probability of any given sequence of long and short syllables in Ancient Greek (using combinatorics). Such an approach was explored in depth in the first decades of the 20th century by de Groot57 and met with criticism by Broadhead. One of the most serious problems with a probabilistic approach is that “syllables form word-units, and the order within such units is fixed”.58 This implies that the probability of patterns of long and short syllables in any language is not determined only by the ratio between long and short syllables in that language. That is to say, if we knew that n per cent of the syllables in a language are long, and the remaining (a hundred minus n per cent) are short, we could infer that a metrical pattern is most probably random if it contains more or less those percentages of long and short syllables. Such a way to assess probabilities is equivalent to what one would do if one had to assess the probability of extracting black or white pebbles from a bag, knowing that the two types of pebbles are there in a given proportion and throwing each pebble back into the bag after having extracted it. However, long or short syllables are not ‘extracted’ from the ‘bag’ of language independently of one another: the probability of ‘extracting’ a long syllable after a short syllable (or any pattern of long or short syllables) is determined by the existence of word-forms that contain that metrical pattern. In other words, the ‘extractions’ of long or short syllables are not independent, because syllables are more often than not part of polysyllabic words whose metrical pattern is fixed. As for qualitative methods for the evaluation of (quasi-)metrical sequences in prose, Dover has presented an ingenious one:  56 Peretti 1979, xi, 7, and 17–19; cf. Corcella 1992, 266. 57 See esp. de Groot 1919. 58 Broadhead 1932, 36.

184  Alessandro Vatri Take a prose text and invert the first two mobile tokens59 after pause and the last two before pause […]. Then compare the rhythmic sequences of the original with the artificially manipulated text; if the score in the original is significantly higher, it promotes the suspicion that the author deliberately sought poetic rhythms, but if significantly lower, that he deliberately avoided them. (Dover 1997, 164)

Dover’s idea is appealing60 but there are two major problems with it. Firstly, if one believes that Greek word order is determined (at least to a great extent) by information structure both at the level of cola and phrase-internally,61 any alteration in the relative order of constituents would not be trivial; that is to say, the original and the manipulated sequences might be syntactically but not pragmatically equivalent. Secondly, in some extreme cases, inverting the first two mobile tokens and comparing the result with the original text can lead to paradoxical conclusions. Let us pretend, for the sake of argument, that we did not know whether or not Euripides’ Medea was a metrical text and that we wished to assess whether or not the sequence ἴθ’ ὡς τάχιστα· μητρὶ δ’ ὧν ἐρᾷ τυχεῖν (line 974), which scans like an iambic trimeter (⏑–⏑–⏑–⏑–⏑–⏑–), is metrical only by accident. If we invert the two final mobile tokens we get the wording ἴθ’ ὡς τάχιστα· μητρὶ δ’ ὧν τυχεῖν ἐρᾷ, which still scans as an iambic trimeter (⏑–⏑–⏑–⏑–⏑–⏑–). By Dover’s method, we would not be able to conclude that the rhythm was deliberately sought — the original sequence is neither more nor less rhythmical than the manipulated one. In spite of its flaws, Dover’s method has the merit of being based on a rigorous criterion. In the absence of an equally well-defined alternative, one has little choice but to resort to a more subjective critical qualitative assessment. In general, the non-randomness of word order can be indicated by other stylistic phenomena (verbal effects, parallelism, chiasm) and, as Gheerbrant (2018, 371–373) contends, its deliberate metrical character can be suggested by the fact that certain sequences correspond to well-formed verse cola and/or sustain a dactylic/anapaestic rhythm for a number of feet rather than a handful of syllables. Certain items in the list above seem to tick these boxes. Sequence (1), for instance, corresponds to the first part of a hexameter up to the hephthemimeral

 59 That is, words that may occur in any position in the colon of which they are part (as opposed to postpositives or prepositives, which may not occur at the beginning or at the end of cola respectively); see Dover 1997, 27. 60 See Gheerbrant 2018, 372–373 for a recent application. 61 See Dik 1995; Matić 2003; Dik 2007; Bakker 2009.

Early Dactylic Prose in the History of Greek Prose Rhythm  185

caesura. Sequence (2) is a so-called feminine hemiepes (i.e. the first part of a hexameter up to the caesura κατὰ τὸν τρίτον τροχαῖον), containing a trithemimeral caesura between καὶ Χθονίη and Χθονίῃ δέ. This pattern corresponds to the statistically most frequent type of hexametric half-lines;62 from a stylistic point of view, the repetition of Χθονίη at end and beginning of clause is a non-trivial verbal effect found in both poetry and prose.63 Other cases do not easily lend themselves to analogous interpretations. Sequence (18) (ταῦτα δὲ ποιήσαντες), for instance, is common in ancient historiography, especially in Herodotus.64 Do we need to assume that it functioned as a (quasi-)metrical signpost in Herodotus and was rarely used by later historiographers who would not have been keen on rhythmic effects — if one believes that Herodotus was?65 Sequence (14) (ἐν δ’ αὐτοῖσι πόλις) is regarded by both Lilja (1968, 24) and Dover (1997, 161–162) as accidental, and they are probably right. While it is true that this expression echoes the Homeric phrase ἐν δ’ αὐτοῖσι πύλας (Il. 7.438), in a different fragment (FGrHist 1 F 204) its unity is disrupted by the appearance of a toponym between ἐν αὐτοῖσι and πόλις. In FGrHist 1 F 229, the preposition ἐν functions as an adverb and a toponym precedes πόλις (ἐν δὲ Σίδουσσα πόλις) and, in general, items in Hecataeus’ lists of locations are commonly introduced by a prepositional phrase or an adverb of this kind, followed by the toponym in the nominative.66 On these grounds, it does not seem likely that the few occurrences of ἐν δ’ αὐτοῖσι πόλις stood out as a Homeric-sounding hexametric snippet, especially in a text that would not conceivably have been recited — even though it could have been read aloud. Certain sequences are parts of sentences whose verbal effects range over a larger group of words. Sequence (28), for instance, is part of a longer literal quotation from Heraclitus, which reads:67 ἄνθρωπος ἐν εὐφρόνῃ φάος ἅπτεται ἑαυτῷ ἀποθανών, ἀποσβεσθεὶς ὄψεις· ζῶν δὲ ἅπτεται τεθνεῶτος εὕδων· ἐγρηγορὼς ἅπτεται εὕδοντος. The sequence ἅπτεται εὕδοντος (which may scan –⏑⏑–––) is

 62 Cf. Porter 1951, 51; Rossi 1995, 275. 63 E.g. Theoc. Id. 1.6; see Denniston 1950, 92–93 for prose examples. 64 Hdt. 1.176.2, 2.40.3, 2.42.6, 2.86.6, 4.71.5, 6.96.2, 7.36.3, 9.92.2, 9.121.1; Xen. An. 6.4.10, Cyr. 2.1.1; Thuc. 2.56.6; Polyb. 15.25.7. 65 See e.g. Hemmerdinger 1981, 171–172. 66 (Adverbial) ἐν δέ 9x (FGrHist 1 F 88, 113a 2x, 116, 141 = 146, 159 2x, 229, 287); (adverbial) μετὰ δέ 9x (FGrHist 1 F 88, 106 2x, 113a, 148, 166 2x, 265, 266); ἐν δ’ αὐτῷ 1x (FGrHist 1 F 141 = 146); ἐν δ’ αὐτοῖσι 4x (FGrHist 1 F 163, 204, 282, 293); ἐξ αὐτῆς 1x (FGrHist 1 F 335); πρὸς μὲν νότον 1x (FGrHist 1 F 144). 67 I print the text of Laks/Most 2016, fr. D71; the glosses to this fragment are based on their translation.

186  Alessandro Vatri part of a large chiasmus, in which ἅπτεται τεθνεῶτος (‘touches on a dead man’) corresponds to ἅπτεται εὕδοντος (‘touches on a sleeping man’) and εὕδων (‘when sleeping’) to ἐγρηγορώς (‘when awake’). In cases like these, (quasi-)metrical sequences might well be epiphenomenal — rhythm, in McCabe’s (1981, 11) words, may be “simply the adventitious result of seeking some other stylistic effect” — but, at least, it would appear that these sequences were not avoided.68

 Early Prose Rhythm and Its Early Theorization If we accept that the (quasi-)metricality of at least some of these sequences was not accidental, we can observe that their rhythmical make-up looks very different from what the earliest extant theoretical discussions of prose rhythm recommend: as mentioned above (Section 1), sustained rhythms should be avoided, and dactylic ones were regarded as unsuited for prose. As I discuss in detail elsewhere,69 the earliest reconstructible theory of prose rhythm (as surfaces in Aristotle, Ephorus, and Theophrastus)70 is based on a ‘musical’ understanding of the notion of foot, defined as a metrical pattern of alternation between long and short syllables and — crucially — by the ratio between the durations of their down- and up-beat.71 The assignment of syllables (and the abstract metrical positions they occupy) to the down- or the up-beat of a foot does not amount to the mechanical mapping of long syllables to down-beats and of short syllables to up-beats. As Aristotle mentions (Rh. 3.1 1403b.26–31), prose rhythm could be controlled in delivery and was not just a stylistic feature to be fixed once and for all when composing a text. The choice of words and their linear order in the sentence would create sequences of long and short syllables that, in principle, could lend themselves to different rhythmic renderings in oral performance, some of which might not do justice to the writer’s intentions (cf. Isoc. 5.25–27). Apart from possible variation in the vocalization of prosodically ambiguous syllables, speakers could introduce in the flow of their speech a sense of periodicity (i.e. regular, isochronous intervals separating prosodic elements that

 68 Cf. Dräger 1995, 11–12 and Dolcetti 2004, 37 n. 188 on Pherecydes of Athens. 69 Vatri (forthcoming). 70 Arist. Rh. 3.8 1408b; Cic. Orat. 191–194; Quint. Inst. 9.4.87–88. 71 Cf. Primmer 1968, 27 and Luque Moreno 1995, 148–152. See Barker 1984, 133–134 and Wallace 2015, xx, 43–44, 141–143 on the theory of ratios in Damon.

Early Dactylic Prose in the History of Greek Prose Rhythm  187

may convey a sense of ‘beat’),72 and rhythmic flow could be manipulated by marking prosodic boundaries (such as those between words or groups of words) more or less strongly through pauses or changes in intonation. Experimental research in this area has demonstrated that acoustic properties such as pitch and pitch variation play a major role in the perception of rhythm in speech across widely diverse languages.73 When it comes to dactylic rhythms, we can infer that a pattern consisting of a long syllable followed by two short ones would not automatically be perceived as a ‘dactylic rhythm’ on paper but needed to be recited as such. In a difficult passage of the Poetics (4 1449a.24–28), Aristotle observes that “the iambic trimeter is the most colloquial among metres; for we speak many iambics in conversation with one another, but hexameters only rarely and diverging (ἐκβαίνοντες) from the colloquial intonation”. As I have argued in another study,74 the participle ἐκβαίνοντες can be interpreted as expressing the manner, rather than the circumstance, in which hexameters were spoken. This would suggest that (a) hexameters were rare in natural speech, (b) they could only be heard sporadically and when speakers intentionally produced them — that is, when they happened to quote Homer or other hexametric poetry — and (c) when quoting a hexameter, speakers would approximate or mimic rhapsodic recitation. If deliberate hexametric sequences were to be recited in order to be perceivable, the practice of seeking dactylic rhythms in prose composition may have been abandoned because prose texts were no longer to be composed for occasions in which a mode of recitation that would make hexametric sequences perceivable was effective or even acceptable. This change may be due to interconnected cultural factors such as the evolution of prose genres,75 changes in the socio-cultural occasions in which prose narrative was received (from ‘lectures’ to private reading), and the extension of written composition to prose genres (such as political and judicial oratory) which were further removed from the poetic tradition than early wisdom or narrative texts.76  72 See Turk/Shattuck-Hufnagel 2013, 99 on modern languages. 73 Especially in poetic or ‘rhythmized’ speech; see Loukina et al. 2013, 340 with further references. 74 Vatri 2016, 390–391. 75 Cf. Fowler 2016, 43–46. 76 A detailed discussion of these issues is beyond the scope of this article. See recently Thomas 2000; Laks 2001; Kahn 2003; Fowler 2013, 678 and 707; Müller 2006, 1–14; Gheerbrant 2018, and Vatri 2017, 60–64 for a summary. According to Norden 1898, 41–48, the difference between early and ‘Classical’ prose rhythm lies in the emergence of the periodic style, with an intermediate step being represented by a style of rhythmic prose consisting of discernible cola, whose earliest

188  Alessandro Vatri

 The Nachleben of Early Prose and the Disappearance of Its Rhythm The idea that dactylic sequences were no longer used because the texts would no longer be recited implies that they were sought as such in texts that would be recited. This is not hard to envisage for wisdom texts such as those of Pherecydes of Syrus or Heraclitus or, in principle, for early ‘epideictic’ texts composed as supports for teaching or demonstrations of technai.77 The work of Pherecydes of Syrus was probably preserved by the Pythagoreans.78 Plato seems to ignore it, but Aristotle’s mention of Pherecydes (Metaph. 1091b.8) may indicate that his book was circulating, albeit restrictedly, in the 4th century. Heraclitus’ book certainly circulated in Athens in the 5th and 4th centuries, at least among the Heracliteans (see above). The works of early ‘historians’ (including Hecataeus, Acusilaus, Charon, and Hellanicus) survived well beyond the Classical period and were still available, for instance, to Dionysius of Halicarnassus (Thuc. 5). Aristotle was certainly well acquainted with Heraclitus’ book79 and, most importantly for our purposes, he was interested in its style as well as in its doctrine. In the Rhetoric (3.5 1407b.14–17), for instance, a sentence from Heraclitus (B 1 D.K.) is quoted as an example of text that is hard to punctuate.80 Nevertheless, Aristotle does not even mention the style of Heraclitus in connection with prose rhythm — not even as an example of bad rhythm.81 The same is true of later rhetoricians, who mention Heraclitus’ style only as far as its notorious obscurity is

 surviving representative would be Sophron. However, it is questionable whether rhythmic cola can be identified in Sophron’s prose; see Hordern 2004, 15–16. 77 See Lloyd 1975, 189; van der Eijk 1997, 93–98; Thomas 2000, 250–254; Thomas 2003, 175– 176, and Jouanna 1992, 549 on medical demonstrations in the early 5th century; see Fowler 2013, 678 on Hecataeus, with further references. 78 West 1971, 4–6. 79 Cf. Kahn 1979, 4. 80 See Vatri 2017, 113. 81 Arguably, Aristotle does give an example of bad rhythm when he describes the disruptive effect of excessive metricality by likening it to what happens when heralds call τίνα αἱρεῖται ἐπίτροπον ὁ ἀπελευθερούμενος, and children anticipate the answer Κλέωνα (Rh. 3.8 1408b.24– 26). The herald’s question is not metrical but builds up an iambic rhythm as it reaches its conclusion (cf. Chiron 2008, 27 n. 1), and the children’s response takes the rhythm on (the sequence ἀπελευθερούμενος :: Κλέωνα scans ⏑⏑–⏑–⏑–⏑–⏑). See Vatri (forthcoming) for a full discussion.

Early Dactylic Prose in the History of Greek Prose Rhythm  189

concerned.82 The picture is similar with other early Ionic prose writers.83 A notable exception are the ῥυθμοί of Hecataeus mentioned by Hermogenes (Id. 411.12–25 Rabe), who, incidentally, is generally interested in both colon-initial and colonfinal rhythm.84 This situation allows for more than one explanation. Rhetoricians might have simply decided not to mention rhythmic phenomena in early prose writers — in the end, they were not aiming to produce systematic critical surveys of all aspects of their style. At the same time, the fact that early prose texts were no longer (if ever) recited might have played a role in the apparent oblivion into which their rhythm fell. Exoteric books such as those of Pherecydes of Syrus and Heraclitus might well have undergone such a fate by the time they fell into Aristotle’s hands. If such texts were ever recited by their authors, reperformances (if any) would have been limited to the circles of their followers. In the absence of oral recitation, metrical patterns would have been hard to discern on paper,85 unless one knew or assumed that metrical sequences ought to be there. If one did not, it is likely that no metrical sequences would have been discerned. Assumptions concerning the presence or absence of rhythm do not only generate false negatives — if one fails to recognize phenomena that are there — but they also generate false positives. This is sometimes the case with modern scholarship,86 but we can observe similar fallacies in ancient criticism as well. An instructive case is the rhythmical analyses of Classical Attic prose presented by Dionysius of Halicarnassus, which are notoriously inconsistent. Apart from his wavering scansions of prosodically ambiguous syllables,87 Dionysius  82 See e.g. Demetr. Eloc. 192; Dion. Hal. Thuc. 46, and Ael. Theon Prog. 82.18–20 Patillon. Cf. Deichgräber 1963, 481. 83 Cicero mentions Acusilaus, Pherecydes of Athens, and Hellanicus (De or. 2.53); Dionysius of Halicarnassus mentions Hecataeus, Acusilaus, Hellanicus, and Charon (Thuc. 5 and 23). 84 See Patillon 1988, 207. 85 Cf. Gheerbrant 2018, 380. 86 Boegehold 1985, for instance, observed that the phrase δυοῖν καδίσκοιν κειμένοιν in Lycurgus’ speech Against Leocrates (149) corresponds to two-thirds of an iambic trimeter and argued that it could be a poetic quotation. This idea was refuted by Worthington 2001. Likewise, the sequence κακὸν κακῶς σε […] οἱ θεοί […] ἀπολέσειαν in Demosthenes’ On the Crown (267), which follows a series of quotations from tragedy, has been identified by a number of scholars as part of an iambic (Meineke 1843, 7; Goodwin 1904, 165; Wankel 1976, 167–168; Yunis 2001, 260; Canfora 2000, 187 n. 104). However, this sequence could simply be a colloquial expression and nothing compels us to interpret it as a deconstructed tragic trimeter (cf. Renehan 1976, 114–116; Kannicht/Snell 2007, 51). Cf. also Hordern 2004, 15–16 questioning the assumption that Sophron wrote rhythmical prose. 87 See e.g. de Jonge 2008, 342–347; Dover 1997, 180–181; Usher 2010, 94.

190  Alessandro Vatri does not provide or follow rigorous criteria for the identification of feet in prose. In one passage (Comp. 18.10–11), for instance, he is in doubt as to whether a sentence from Plato88 begins with a bacchius (––⏑, ἔργῳ μέν) or an iamb (––⏑–, ἔργῳ μὲν ἡ). Dionysius’ way of thinking about prose rhythm is quite different from that of the early Attic theoreticians (see Section 3 above). For one thing, the inventory recoverable for the 4th-century theory of prose rhythm only includes the dactyl, the spondee, the iamb, the trochee, and the opening and closing paeans,89 while Dionysius’ inventory (Comp. 17.2–15) is much larger and includes pyrrhichs, spondees, iambs, trochees, chorees, molossi, amphibrachs, anapaests, dactyls, cretics, bacchii and hypobacchii. Additionally, Dionysius does not regard prose feet as defined by the durational ratio between their up- and down-beats, in spite of the fact that he is aware of such an understanding of rhythm (Dem. 48).90 These discrepancies are not to be ascribed to the fact that Dionysius may have perceived speech rhythm differently from his 4th-century predecessors91 as much as to the fact that he adopted an entirely different theoretical perspective, that of the ‘metricians’.92 Their approach was concerned with the identification of abstract metrical patterns of syllabic quantities regardless of their phonetic realization, as opposed to the description of their potential rhythmic pulse.93 Such a paradigm change in the understanding of prose rhythm might have been the result of a hiatus in the ancient doctrine corresponding to the emergence of Asianic rhetoric.94 Teachers and writers would have devoted their attention to the regularization of rhythmical cadences (clausulae),95 which would have encouraged a

 88 Pl. Menex. 236d.4–5: ἔργῳ μὲν ἡμῖν οἵδε ἔχουσιν τὰ προσήκοντα σφίσιν αὐτοῖς· ὧν τυχόντες πορεύονται τὴν εἱμαρμένην πορείαν. 89 See Vatri (forthcoming) for a detailed discussion. 90 See Rocconi 2008. 91 The loss of distinctive vowel length and the shift from a pitch to a stress accent was probably common about a century before Dionysius’ lifetime (Horrocks 2010, 122 n. 5, cf. 165) and, among the well-educated, by the 2nd century AD (see Horrocks 2010, 146–147; Vessella 2018, 66 and 70). One is left wondering whether 1st-century BC intellectuals such as Dionysius would continue to adopt a conservative pronunciation. One may note that the development of an accentual rhythm in Greek prose, for instance, has been dated to the 4th century AD (see Dewing 1910). Of course, it is certainly possible that Dionysius’ scansions “reflect certain changes in the perception of the quantities of syllables” (de Jonge 2008, 343). 92 Cf. de Jonge 2008, 341 n. 60; see also Aristides Quintilianus 1.20–29 Winnington-Ingram and Barker 1989, 448 n. 231. 93 Cf. Luque Moreno 1995, 159 and Barker 1991, 71; cf. also Cassius Longinus fr. 42.4 Patillon/Brisson. 94 Cf. Blass 1901, 19–20; Hutchinson 2015, 789. 95 Cf. Blass 1901, 9 and 17; Dewing 1910, 312.

Early Dactylic Prose in the History of Greek Prose Rhythm  191

‘metrical’ rather than a ‘rhythmical’ approach. Such an approach was resilient even after Asianism fell out of fashion,96 but its inadequacy in capturing fully the rhythmic phenomena of Attic prose is betrayed precisely by Dionysius’ attempts. In his discussion of Plato’s passage mentioned above (Comp. 18.10–11), for instance, Dionysius ends up resorting to an aesthetic criterion to make a decision: the analysis of ἔργῳ μὲν (ἡ) as a bacchius is to be preferred over that as an iambic since, in Dionysius’ words, “not running, swift movements, but slow and measured times are appropriate as a tribute to those for whom we mourn”.97 Dionysius attempts to capture an acoustic and psychological phenomenon such as the perception of a rhythmic pulse in connected speech through the identification of abstract metrical patterns that evidently cannot tell the whole story: as we have seen, rhythmical effects in Attic prose depended greatly on the speaker’s ability to produce them in performance.98 Dionysius’ hunt for metrical patterns in writers like Demosthenes or Plato does not require oral recitation and aural reception and, in principle, could have been carried out entirely through the examination of written texts. If, as argued above, metres would not automatically be perceived on paper, Dionysius’ effort might have been motivated primarily by the assumption that prestigious Attic writers must have used ‘rhythms’ in their prose. This task was evidently complicated by the fact that Dionysius’ understanding of prose ‘rhythms’ made his analysis anachronistic (and potentially misleading for modern critics as well). To sum up, one of the reasons why rhetoricians did not comment on the rhythm of early Ionic prose writers (in spite of being interested in their style) may lie in the fact that they simply did not notice it. This may be because (a) they had not heard such texts being recited (and not simply read aloud), and (b) they had no reason to assume that rhythm was there to be found in such texts.99

 96 Cf. Valiavitcharska 2013, 59. 97 Translation by S. Usher. On the idea that tempo (ἀγωγή) was an intrinsic feature of rhythms (and not just a matter of how rhythms were performed) see Pl. Resp. 400c.1–3 and cf. West 1992, 153. 98 Cf. Isoc. 5.25–27. 99 Conversely, we may speculate that Hermogenes mentioned rhythm as a feature of Hecataeus’ style because (for reasons which is beyond the scope of this study to investigate) he was induced to think that he should look for it — if, of course, Hermogenes’ remark is based on his first-hand appreciation of the text.

192  Alessandro Vatri

 Conclusion A number of early Ionic prose texts contain sequences that appear to be markedly dactylic. This feature may be connected to the original modes of reception of such works and, at the same time, does not meet the standards of the ‘best practice’ identified in the earliest surviving theoretical reflections on prose rhythm. By the time intellectuals started writing about prose rhythm, as far as we can tell, dactylic prose must have gone out of fashion. Its rhythm was almost completely ignored by ancient rhetoricians, even though some do show an acquaintance with these texts and an interest in their style. This is a striking fact which calls for an explanation — albeit a speculative one. On the one hand, we need not expect that ancient rhetoricians should discuss all aspects of texts they showed some interest in, even though dactylic sequences would have provided good examples of what should be avoided in rhythmical prose composition (and negative examples are commonly given for many other rhetorical qualities). On the other hand, rhetoricians may have failed to notice rhythmic patterns in texts that they would not hear being recited and in which they did not assume there were rhythms to find. Rhythm is a psychoacoustic phenomenon, but when it comes to criticism and scholarship it often lies in the eye of the beholder; the evolution of literary culture, taste, and intellectual frameworks dramatically affect its perception and theoretical understanding.

Bibliography Bakker, S.J. (2009), The Noun Phrase in Ancient Greek: A Functional Analysis of the Order and Articulation of NP Constituents in Herodotus, Leiden/Boston. Barker, A. (1984), Greek Musical Writings, vol. 1, Cambridge/New York. Barker, A. (1989), Greek Musical Writings, vol. 2, Cambridge/New York. Barker, A. (1991), “Review of L. Parsons, Aristoxenus: Elementa rhythmica (Oxford 1990)”, in: Music & Letters 72, 71–74. Bechtel, F. (1924), Die griechischen Dialekte, vol. 3: Der ionische Dialekt, Berlin. Bernabé, A. (1979), “Los filósofos presocráticos como autores literarios”, in: Emerita 47, 357– 394. Blass, F. (1901), Die Rhythmen der attischen Kunstprosa: Isokrates–Demosthenes–Platon, Leipzig. Boegehold, A.L. (1985), “Lycurgus 1.149”, in: CP 80, 132–135. Broadhead, H.D. (1932), “Prose-Rhythm and Prose-Metre”, in: CQ 26, 35–44. Canfora, L. (2000), Discorsi e lettere di Demostene: Discorsi in tribunale, Torino.

Early Dactylic Prose in the History of Greek Prose Rhythm  193

Chiron, P. (2008), “Musique des mots et persuasion dans deux traités de rhétorique d’époque hellénistique et romaine (Ps.-Démétrios de Phalère, Du Style, Denys d’Halicarnasse, La Composition stylistique)”, in: Modèles linguistiques 58, 27–43. Corcella, A. (1992), “Geografia e historie”, in: G. Cambiano/L. Canfora/D. Lanza (eds.), Lo spazio letterario della Grecia antica, vol. 1.1: La produzione e la circolazione del testo, Roma, 265–277. de Groot, A.W. (1919), A Handbook of Antique Prose-Rhythm, Groningen. de Jonge, C.C. (2008), Between Grammar and Rhetoric: Dionysius of Halicarnassus on Language, Linguistics and Literature, Leiden/Boston. Deichgräber, K. (1963), Rhythmische Elemente im Logos des Heraklit, Wiesbaden. Denniston, J.D. (1950), Greek Prose Style, London. Dewing, H.B. (1910), “The Origin of the Accentual Prose Rhythm in Greek”, in: AJPh 31, 312– 328. Dik, H. (1995), Word Order in Ancient Greek: A Pragmatic Account of Word Order Variation in Herodotus, Amsterdam. Dik, H. (2007), Word Order in Greek Tragic Dialogue, Oxford/New York. Dolcetti, P. (2004), Ferecide di Atene: Testimonianze e frammenti, Alessandria. Dover, K.J. (1997), The Evolution of Greek Prose Style, Oxford/New York. Dräger, P. (1995), Stilistische Untersuchungen zu Pherekydes von Athen: Ein Beitrag zur ältesten ionischen Prosa, Stuttgart. Fowler, R.L. (1999), “The Authors named Pherecydes”, in: Mnemosyne 52, 1–15. Fowler, R.L. (2001), “Early Historiē and Literacy”, in: N. Luraghi (ed.), The Historian’s Craft in the Age of Herodotus, Oxford, 95–115. Fowler, R.L. (2013), Early Greek Mythography, vol. 2: Philological Commentary, Oxford. Fowler, R.L. (2016), “The Submersion of Mythography”, in: A. Ercolani/M. Giordano (eds.), Submerged Literature in Ancient Greek Culture, vol. 3: The Comparative Perspective, Berlin/ Boston, 43–54. Gheerbrant, X. (2018), “Le rythme de la prose de Phérécyde de Syros: Mythographie en prose et poésie en hexamètre dactylique”, in: Mnemosyne 71, 367–383. Goodwin, W.W. (1904), Demosthenes: On the Crown, Cambridge. Granger, H. (2004), “Argumentation and Heraclitus’ Book”, in: OSAPh 26, 1–17. Granger, H. (2007), “The Theologian Pherecydes of Syros and the Early Days of Natural Philosophy”, in: HSPh 103, 135–163. Hemmerdinger, B. (1981), Les manuscrits d’Hérodote et la critique verbale, Genova. Hordern, J.H. (2004), Sophron’s Mimes: Text, Translation, and Commentary, Oxford/New York. Horrocks, G. (2010), Greek: A History of the Language and Its Speakers, 2nd edn., Chichester/ Malden, Mass. Hutchinson, G.O. (2015), “Appian the Artist: Rhythmic Prose and Its Literary Implications”, in: CQ 65, 788–806. Jouanna, J. (1992), Hippocrate, Paris. Kahn, C.H. (1979), The Art and Thought of Heraclitus: An Edition of the Fragments with Translation and Commentary, Cambridge. Kahn, C.H. (2003), “Writing Philosophy: Prose and Poetry from Thales to Plato”, in: H. Yunis (ed.), Written Texts and the Rise of Literate Culture in Ancient Greece, Cambridge, 139–161. Kannicht, R./Snell, B. (2007), Tragicorum Graecorum fragmenta, vol. 2, Göttingen. Laks, A. (2001), “Écriture, prose et les débuts de la philosophie grecque”, in: Methodos 1, 131– 151.

194  Alessandro Vatri Laks, A./Most, G.W. (2016), Early Greek Philosophy, vol. 3: Early Ionian Thinkers, Part 2, Cambridge, Mass. Lilja, S. (1968), On the Style of the Earliest Greek Prose, Helsinki. Lloyd, G.E.R. (1975), “The Hippocratic Question”, in: CQ 25, 171–192. Loukina, A./Rosner, B./Kochanski, G./Keane, E./Shih, C. (2013), “What Determines DurationBased Rhythm Measures: Text or Speaker?”, in: Laboratory Phonology 4, 339–382. Luque Moreno, J. (1995), De pedibus, de metris: Las unidades de medida en la rítmica y en la métrica antiguas, Granada. Matić, D. (2003), “Topic, Focus, and Discourse Structure: Ancient Greek Word Order”, in: StudLang 27, 573–633. McCabe, D.F. (1981), The Prose Rhythm of Demosthenes, New York. Meineke, J.A.F.A. (1843), Philologicarum exercitationum in Athenaei Deipnosophistas, Berlin. Most, G.W. (1999), “The Poetics of Early Greek Philosophy”, in: A.A. Long (ed.), The Cambridge Companion to Early Greek Philosophy, Cambridge, 332–362. Müller, C.W. (2006), Legende, Novelle, Roman: Dreizehn Kapitel zur erzählenden Prosaliteratur der Antike, Göttingen. Nieddu, G.F. (1984), “Testo, scrittura, libro nella Grecia arcaica e classica: Note e osservazioni sulla prosa scientifico-filosofica”, in: S&C 8, 213–261. Norden, E. (1898), Die antike Kunstprosa, Leipzig. O’Sullivan, N. (1992), Alcidamas, Aristophanes, and the Beginnings of Greek Stylistic Theory, Stuttgart. Patillon, M. (1988), La théorie du discours chez Hermogène le Rhéteur: Essai sur les structures linguistiques de la rhétorique ancienne, Paris. Pavese, C.O. (1998), “The Rhapsodic Epic Poems as Oral and Independent Poems”, in: HSPh 98, 63–90. Peretti, A. (1979), Il Periplo di Scilace: Studio sul primo portolano del Mediterraneo, Pisa. Pfeiffer, R. (1968), History of Classical Scholarship from the Beginnings to the End of the Hellenistic Age, Oxford. Porter, H.N. (1951), “The Early Greek Hexameter”, in: YClS 12, 3–63. Primmer, A. (1968), Cicero numerosus: Studien zum antiken Prosarhythmus, Wien. Renehan, R. (1976), Studies in Greek Texts: Critical Observations to Homer, Plato, Euripides, Aristophanes and Other Authors, Göttingen. Rocconi, E. (2008), “Musica e retorica nel De compositione uerborum di Dionigi di Alicarnasso: Per un’ipotesi sulle fonti ritmiche del trattato dionisiano”, in: Incontri Triestini di Filologia Classica 8, 175–190. Rossi, L.E. (1995), “Estensione e valore del colon nell’esametro omerico”, in: M. Fantuzzi/ R. Pretagostini (eds.), Struttura e storia dell’esametro greco, Roma, 271–320. Schibli, H.S. (1990), Pherekydes of Syros, Oxford. Schwyzer, E. (1939), Griechische Grammatik, vol. 1: Allgemeiner Teil, Lautlehre, Wortbildung, Flexion, München. Steiner, D.H. (1994), The Tyrant’s Writ, Princeton. Thomas, R. (1992), Literacy and Orality in Ancient Greece, Cambridge/New York. Thomas, R. (2000), Herodotus in Context: Ethnography, Science and the Art of Persuasion, Cambridge. Thomas, R. (2003), “Prose Performance Texts: Epideixis and Written Publication in the Late Fifth and Early Fourth Centuries”, in: H. Yunis (ed.), Written Texts and the Rise of Literate Culture in Ancient Greece, Cambridge, 162–188.

Early Dactylic Prose in the History of Greek Prose Rhythm  195

Thomas, R. (2009), “Writing, Reading, Public and Private ‘Literacies’: Functional Literacy and Democratic Literacy in Greece”, in: W.A. Johnson/H.N. Parker (eds.), Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford, 13–42. Toye, D.L. (1997), “Pherecydes of Syros: Ancient Theologian and Genealogist”, in: Mnemosyne 50, 530–560. Turk, A./Shattuck-Hufnagel, S. (2013), “What is Speech Rhythm? A Commentary on Arvaniti and Rodriquez, Krivokapić, and Goswami and Leong”, in: Laboratory Phonology 4, 93–118. Usher, S. (2010), “Eurhythmia in Isocrates”, in: CQ 60, 82–95. Valiavitcharska, V. (2013), Rhetoric and Rhythm in Byzantium: The Sound of Persuasion, Cambridge/New York. van der Eijk, P.J. (1997), “Towards a Rhetoric of Ancient Scientific Discourse”, in: E.J. Bakker (ed.), Grammar as Interpretation: Greek Literature in Its Linguistic Contexts, Leiden/New York, 77–129. Vatri, A. (2016), “Between Song and Prose: The Meaning(s) of Harmonia in Aristotle’s Rhetoric and Poetics”, in: Rhetorica 34, 372–392. Vatri, A. (2017), Orality and Performance in Classical Attic Prose: A Linguistic Approach, Oxford. Vatri, A. (forthcoming), “The Nature and Perception of Attic Prose Rhythm”, in: CP. Vessella, C. (2018), Sophisticated Speakers: Atticistic Pronunciation in the Atticist Lexica, Berlin/Boston. von Fritz, K. (1938), “Pherekydes”, in: RE 18.2, 2025–2033. Wallace, R.W. (2015), Reconstructing Damon: Music, Wisdom Teaching, and Politics in Perikles’ Athens, New York. Wankel, H. (1976), Demosthenes: Rede für Ktesiphon über den Kranz, Heidelberg. Werner, S. (2009), “Literacy Studies in Classics: The Last Twenty Years”, in: W.A. Johnson/ H.N. Parker (eds.), Ancient Literacies: The Culture of Reading in Greece and Rome, Oxford, 333–382. West, M.L. (1971), Early Greek Philosophy and the Orient, Oxford. West, M.L. (1992), Ancient Greek Music, Oxford. Worthington, I. (2001), “Lycurgus 1.149 and Those Two Voting Urns”, in: CQ 51, 301–304. Yunis, H. (2001), Demosthenes: On the Crown, New York.

Matilde Serangeli

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage: Gk. PN Τέμπυρα, Hitt. dampu‘Blunt’, Old Russ. tupъ ‘Blunt, Stupid’ Abstract: In the present paper I will propose that the Greek adjective ταπεινός is related to PIE *temp- ‘to stretch, to extend’ (LIV 626), and that the place name Τέμπεα and the noun τέμπεα ‘valley, defile, narrow path’ (: *temp-es-), as well as the place name Τέμπυρα, are cognate with it. Furthermore, the hypothesis will be put forward that ταπεινός may represent a *-no- derivative of a zero-grade s-stem (*tm̥p-es-nó-) and that it may be the outcome of a semantic development starting from *‘stretched out (in length)’. The etymology and word formation of ταπεινός will be addressed in light of a new linguistic comparative analysis of lexical data from other Indo-European branches. Keywords: Indo-European, word formation, semantics, etymology, Caland system, Greek, Anatolian, Slavic.

 Introduction The Greek adjective ταπεινός ‘low(-lying)’ applies in a physical sense to places, persons and objects as well as in a physical and figurative (mostly pejorative) sense to persons and objects.1 Despite its well understood semantics,2 no cognates have

 1 This study has been pursued within my Marie Skłodowska-Curie project 2016–2018 The Linguistic Origins of Europe: Word-formation and Lexicon in Anatolian and Core-Indo-European, University of Copenhagen, Roots of Europe Research Center (European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 705090). I express my gratitude to Timothy Barnes (Cambridge), Benjamin Fortson (Michigan), José Luis García Ramón (CHS/Harvard), Stefan Höfler (Harvard), H. Craig Melchert (UCLA), Thomas Olander (København), Jeremy Rau (Harvard), and Marco Santini (Princeton) for the remarks, criticism, and suggestions offered to me on different occasions. I would also like to thank the reviewers for their suggestions to improve this paper. For their help in improving my English I would like to thank Benedicte Nielsen Whitehead (København) and Cormac Anderson (Jena). All remaining mistakes are, of course, my own responsibility. All translations are mine unless otherwise noted. 2 GEW s.v. ταπεινός: “‘niedrig gelegen, gering im Range, gemein, unbedeutend, demütig’”; DELG 1093: “‘qui se trouve bas’, dit aussi de personnes de rang peu élevé, ‘insignificant, https://doi.org/10.1515/9783110621747-010

198  Matilde Serangeli been identified so far and the adjective remains, therefore, isolated within Greek. Moreover, the only satisfactory etymological explanation that has been proposed so far, which connects ταπεινός with PIE *temp- ‘to stretch, to extend’ (Bally 1903; PIE *temp- most likely from *ten-p-; LIV 626), has never been given serious consideration and the term has even been considered to be Pre-Greek.3 The aim of this paper is threefold. First, I will present evidence for several semantic nuances of ταπεινός in Greek; then, I will make the case for ταπεινός as a specialization of *‘stretched out (in length)’ from *tm̥p-es-nó- (see Section 2). Relying on the possible analogy with the structure Gk. αἰπεινός ‘high’ :: αἰπύς ‘id.’ :: αἶπος ‘height’ I will argue for the existence of a u-stem adjective *ταπύς (: *tm̥p-ú-; cf. Ved. tanú-, Gk. τανυ˚ to PIE *ten-) and an s-stem noun *τάπος, both deriving from PIE *temp- (see Section 3). Second, I will demonstrate that ταπεινός has cognates within and outside Greek: this supports the existence in Greek of the s- and u-stem derivatives as expected by analogy with the αἰπεινός type. The relevant comparanda are the Greek place name Τέμπεα and the noun τέμπεα ‘valley, defile, narrow path between two mountains’ (: *temp-es-), the place name Τέμπυρα attested in Thracia (either from *temp-u-reh2 built on *temp-u-ro- or due to the productivity of uro- adjectives in Greek), Hitt. dampu- ‘blunt’ (from *tm̥p-ú- or *tomp-ú-) and its derivative dampupi- ‘barbarian, not educated’ (tentatively, with a suffix ‑(u)pi- of unclear origin), and Old Russ. tupъ ‘blunt/not sharp, stupid’ (: Comm. Slav. *tǫpъ, surely a tomós type; see Sections 4–4.1). Third, on the basis of the suggested formal connection of those terms with ταπεινός and considering the semantic match between ταπεινός, Hitt. dampu- and dampupi-, and Old Russ. tupъ, I will argue for a semantic shift of PIE *temp- ‘to stretch, to extend’ from *‘stretched out (in length)’ to (a) ‘low(-lying)’ and (b) ‘blunt’ as reflected in Greek, Hittite and Slavic respectively, and the development of a pejorative figurative meaning as ‘miserable’ in Greek, ‘not educated’ in Hittite, and ‘stupid’ in Slavic that may have arisen separately within the individual branches (see Sections 4– 4.1). However, it cannot be ignored that the attestation of these semantic shifts is restricted to three derivative formations, namely a u-stem adjective (Anatolian), a tomós derivative (Slavic), and a derivative in *-no- of a zero-grade s-stem (Greek), which are attested in three Indo-European branches that are extremely far away from each other (Anatolian, Greek, and Slavic). Therefore, a later and independent creation of these derivatives within these three Indo-European branches is also a possibility. This would explain why the same semantic shift  humble’, quelquefois ‘bas’ pris en sens moral, parfois en parlant de situations ‘misérable, pauvre’”. 3 GEW s.v. ταπεινός: “Bildung wie αἰπεινός u.a. Sonst dunkel”; EDG 1450: “Pre-Greek”.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  199

(though with slightly different semantic nuances) is expressed by different derivatives in Anatolian, Greek, and Slavic, and would imply that we are dealing with trivial semantic shifts. However, based on the strength of the similarity of the Anatolian and Greek material on the morphological level, as well as of the Anatolian and Slavic material on the semantic level, the first option seems to be more likely.

 The Semantic Evidence Gk. ταπεινός ‘low(-lying)’ applies to places and objects in a physical sense (vertical direction, see examples (1−3)) as well as to persons and objects in a moral (mostly pejorative) sense (see examples (4−6)). Thus, the adjective defines ‘lowlying regions’ (1), ‘low water’ (2) and ‘low stature or size’ (3): (1)

ἔστι δ’ αἰετὸς ὠκὺς ἐν ποτανοῖς, ὃς ἔλαβεν αἶψα, τηλόθε μεταμαιόμενος, δαφοινὸν ἄγραν ποσίν· κραγέται δὲ κολοιοὶ ταπεινὰ νέμονται.

(Pind. Nem. 3.80−82)

The eagle is swift among birds: chasing from afar, it suddenly seizes his blood-stained prey with its talons; but the cawing jackdaws live in the low-lying regions (scil. of the air). (2)

τοῖς μὲν γὰρ ἄλλοις αὔξεται τὸ ῥεῦμα, καθὼς ἂν πλείους διαφέρωνται τόπους, καὶ μέγιστοι μέν εἰσι κατὰ χειμῶνα, ταπεινότατοι δὲ κατὰ τὴν ἀκμὴν τοῦ θέρους. (Polyb. 9.43.4) In the case of other rivers the stream increases the more country they traverse, they are largest in winter and lowest in the height of summer. (transl. W.R. Paton)

(3) δεῖ τοίνυν καὶ τὰ ἀνωτέρω μὲν τῶν ὁπλῶν, κατωτέρω δὲ τῶν κυνηπόδων ὀστᾶ μήτε ἄγαν ὀρθὰ εἶναι ὥσπερ αἰγός· ἀντιτυπώτερα γὰρ ὄντα κόπτει τε τὸν ἀναβάτην καὶ παραπίμπραται μᾶλλον τὰ τοιαῦτα σκέλη· οὐδὲ μὴν ἄγαν ταπεινὰ τὰ ὀστᾶ δεῖ εἶναι· ψιλοῖντο γὰρ ἂν καὶ ἑλκοῖντο οἱ κυνήποδες εἴτ᾽ ἐν βώλοις εἴτ᾽ ἐν λίθοις ἐλαύνοιτο ὁ ἵππος. (Xen. Eq. 1.4) The bones (scil. of the pastern) above the hoofs and below the fetlocks should not be too upright, like a goat's: such legs give too hard a tread, jar the rider, and are more liable to inflammation. Nor yet should the bones be too low, else the fetlocks are likely to become bare and sore when the horse is ridden over clods or stones. (transl. E.C. Marchant) (3.1) πρῶτον μὲν τοίνυν χρὴ οὐ μεῖον δυοῖν χαλινοῖν κεκτῆσθαι. τούτων δὲ ἔστω ὁ μὲν λεῖος, τοὺς τροχοὺς εὐμεγέθεις ἔχων, ὁ δ᾽ ἕτερος τοὺς μὲν τροχοὺς καὶ βαρεῖς καὶ ταπεινούς. (Xen. Eq. 10.6)

200  Matilde Serangeli First, you should possess two bits at least. One of these should be smooth and have the discs of a good size; the other should have the discs heavy and low. (transl. E.C. Marchant)

In the figurative usage, Gk. ταπεινός mostly describes a negative human condition or action such as humiliation, misery and meanness as imposed by a superior or a deity: (4) Ἀλλὰ μὴν καὶ τὸ μεγαλοπρεπές τε καὶ ἐλευθέριον καὶ τὸ ταπεινόν τε καὶ ἀνελεύθερον καὶ τὸ σωφρονικόν τε καὶ φρόνιμον καὶ τὸ ὑβριστικόν τε καὶ ἀπειρόκαλον καὶ διὰ τοῦ προσώπου καὶ διὰ τῶν σχημάτων καὶ ἑστώτων καὶ κινουμένων ἀνθρώπων διαφαίνει. (Xen. Mem. 3.10.5) And further, nobility and dignity, self-abasement and servility, prudence and understanding, insolence and vulgarity: these are reflected in the face and in the attitudes of the body whether still or in motion.4 (transl. E.C. Marchant) (4.1) οἶδα μὲν γὰρ ὑμῖν Μυσοὺς λυπηροὺς ὄντας, οὓς νομίζω ἂν σὺν τῇ παρούσῃ δυνάμει ταπεινοὺς ὑμῖν παρασχεῖν. (Xen. An. 2.5.13) I know that the Mysians are troublesome to you, and I believe that with the force I have I could make them your submissive servants. (transl. C.L. Brownson) (4.2) τά τοι μέγιστα πολλάκις θεὸς / ταπείν’ ἔθηκε καὶ συνέστειλεν πάλιν.

(Eur. fr. 716 Kannicht)5

Often god lowers the mightiest and puts them back in their place. (transl. C. Collard/M. Cropp) (4.3) ὡς καὶ μέγας καὶ πολλὸς ἐγένεο ἐν ὀλίγῳ χρόνῳ, οὕτω καὶ ταπεινὸς ὀπίσω κατὰ τάχος ἔσεαι. (Hdt. 7.14.1) Just as you became great and mighty in little time, thus you will be (brought) low again quickly.

 4 See also ταπεινὰ (or adv. ταπεινῶς) πράττειν, e.g. in Plut. Thes. 6.5 τούτων Ἡρακλῆς τοὺς μὲν ἐξέκοπτε καὶ ἀνῄρει περιιών, οἱ δὲ λανθάνοντες ἐκείνου παριόντος ἔπτησσον καὶ ἀνεδύοντο καὶ παρημελοῦντο ταπεινὰ πράττοντες ‘some of these creatures Heracles cut off and destroyed as he went about, but some escaped his notice as he passed by, crouching down and shrinking back, and were overlooked in their abjectness (: lowering themselves)’ (transl. B. Perrin). 5 Cf. Eur. Hec. 245 ἥψω δὲ γονάτων τῶν ἐμῶν ταπεινὸς ὤν; ‘did you grasp my knees in humility?’; Andr. 979 κἀγὼ ταπεινὸς ὢν τύχαις ταῖς οἴκοθεν ‘since I was humiliated because of my troubles at home’.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  201

In a few cases, ταπεινός reflects a positive human disposition of humility (the positive nuance is secondary and mostly attested in the Christian era):6 (4.4) εὑρήσεις γὰρ ὡς ἐπὶ τὸ πολὺ τοὺς μὲν σεμνυνομένους ψυχροὺς ὄντας, τοὺς δὲ βουλομένους ἀστείους εἶναι ταπεινοὺς φαινομένους. (Isoc. 2.34) You will find that for the most part those who affect dignity are cold, while those who desire to be courteous appear to humble themselves. (transl. G Norlin, slightly modified)

In a similar way, the adjective is also used figuratively about things (as opposed to conditions): with regards to speech, see ταπεινῶς λέγειν ‘speak in a submissive manner’ (vs. ἀγαμένως λέγειν ‘speak with admiration/respect’, Arist. Rh. 1408a.18–19) and the compound ταπεινολογία ‘low, humble speech’ (Poll. 2.124 Bethe); regarding style or a theory, see ταπεινὴ λέξις vs. κεκοσμημένη λέξις (Arist. Rh. 1404b.6–7), θεωρία ταπεινοτέρα vs. τιμιωτέρα (Arist. Par. an. 639a.1), τὴν μικρὰν καὶ ταπεινὴν [ἰατρικὴν θεωρίαν] (Gal. In Hippocratis prorrheticum 16.550.7). Lastly, on a more concrete level, Pl. Leg. 762e.7–8 refers to a poor meal as ταπεινὴ καὶ ἄπορος δίαιτα ‘a coarse and uncooked daily ration’. These semantic nuances must have been well established in Greek, since the derivatives of ταπεινός show the same range of meanings and semantic developments; this is particularly evident in the case of ταπεινότης ‘lowering’ and ταπείνωσις ‘id.’. See, for ταπεινότης, the meanings ‘lowness of position’ (τῆς χώρας Diod. Sic. 1.31.5; τῆς μήτρας Pseudo-Plut. Placita philosophorum 5.14.2), ‘low estate, abasement’ (εἰς τοσαύτην ταπεινότητα κατεστήσαμεν Isoc. 4.118.4– 5; Thuc. 7.75.6), ‘lowness of spirits, dejection’ (σιωπήν τε καὶ ταπεινότητα, Xen. Hell. 3.5.21), in moral sense ‘baseness, vileness’ (Pl. Plt. 309a.5; paralleled with μικροψυχία by Arist. Rh. 1384a.4), ‘of meanness of the style’ (Quint. Inst. 8.3.48). And, for ταπείνωσις, see the meanings ‘lowering (of a movement in dancing)’ (Pl. Leg. 815a.4), ‘lessening’ (vs. αὔξησις, Arist. Part. an. 689a.25), ‘reduction of a swelling’ (Gal. De compositione medicamentorum secundum locos libri x 12.816.5), ‘humiliation, abasement’ (Zen. fr. 209 von Arnim; Polyb. 9.33.10), ‘low state, low condition’ (LXX Genesis 29.32; Luke 1.48), ‘lowness of style’ (Plut. De liberis educandis 7a.9; Quint. Inst. 8.3.48).

 6 It may be noted that a positive sense ‘good, humble’ is also attested before the Christian period (New Testament, passim): see Pl. Leg. 716a.4; Xen. Cyr. 5.1.5. As for the same double negative and positive secondary semantic development, see Gk. ταπεινοφρονέω ‘to be mean-spirited’ (Arr. Epict. diss. 1.9.10), ‘to be lowly in mind, humble’ (LXX Psalm 130 (131).2); ταπεινοφροσύνη ‘mean-spiritedness’ (Arr. Epict. diss. 3.24.56), ‘humility’ (Epistle to the Ephesians 4.2).

202  Matilde Serangeli It is interesting to note that ταπείνωσις ‘lowering’ clearly expresses a vertical movement, as is also the case with ταπεινός (see above, especially no. 2):7 (5) ἄλλην οὖσαν τῆς εἰρηνικῆς, πυρρίχην ἄν τις ὀρθῶς προσαγορεύοι, τάς τε εὐλαβείας πασῶν πληγῶν καὶ βολῶν ἐκνεύσεσι καὶ ὑπείξει πάσῃ καὶ ἐκπηδήσεσιν ἐν ὕψει καὶ σὺν ταπεινώσει μιμουμένην καὶ τὰς ταύταις ἐναντίας, τὰς ἐπὶ τὰ δραστικὰ φερομένας αὖ σχήματα, ἔν τε ταῖς τῶν τόξων βολαῖς καὶ ἀκοντίων καὶ πασῶν πληγῶν μιμήματα ἐπιχειρούσας μιμεῖσθαι. (Pl. Leg. 815a.1–7) Being distinct from the pacific, one may rightly term (this) as ‘pyrrhiche’; it represents modes of eluding all kinds of blows and shots by swervings and duckings and side-leaps upward or crouching; and also the opposite kinds of motion, which lead to active postures of offence, when it strives to represent the movements involved in shooting with bows or darts, and blows of every description. (transl. R.G. Bury)

To sum up, Gk. ταπεινός ‘low(-lying)’ and its derivatives clearly define a negative physical or moral condition or disposition. The semantic development from a physical to a moral condition is not surprising: compare Lat. acūtus ‘high, sharp’ (said of objects) as well as ‘intelligent, smart’ or indeed ‘sharp’ (said of persons) (*h2eḱ- ‘scharf sein/werden/machen’, LIV 261). It might be noted that ταπεινός explicitly appears in opposition to something ‘high’ or ‘elevated’, as in the case of the opposition between ebb and high tide in (2). While ταπεινός seems to develop a generic meaning ‘low’ in the Classical and post-Classical periods, its derivatives offer the most significant examples of the action of lowering itself, such as the lowering of the body during a dance in (5). These considerations will be crucial for the attempt of the present paper to trace Gk. ταπεινός ‘low(-lying)’ back to a meaning *‘stretched out (in length)’ (PIE *temp- ‘dehnen, spannen’, LIV 626). Something ‘stretched out (in length)’ may lead to a reduction in height by means of a lowering of a previous stage of a given substance. A very similar process is also shown by the well-established meaning ‘thin’ from *‘stretched out’: see PIE *temp-, *ten- ‘to stretch’ → ‘to make thin’; ‘stretched (out)’ → ‘thin’, ‘long’  7 In addition to (2) it might be noted that ταπεινός as referring to the low position ‘of a part of the body’ or ‘of the stars’ (as ‘near the horizon’/‘near the earth’) is attested in opposition to ὑψηλός ‘high’: see Gal. In Hippocratis de victu acutorum commentaria iv 15.531 ἡ δὲ διὰ τῆς μασχάλης τὰ κατὰ τὸν θώρακα καὶ τούτων αὐτῶν μᾶλλον τὰ ὑψηλὰ καὶ μάλιστ’ ἐπ’ ἀνθρώπων ἐπειδὴ τὰ ταπεινὰ πάντα μία φλὲψ ἑτέρα τρέφει κατωτέρω τῆς καρδίας ἀποφυομένη τῆς κοίλης φλεβός (τὰ ταπεινὰ τοῦ θώρακος :: τὰ ὑψηλά); Cleomedes 2.5 ὁπόταν μὲν οὖν ταπεινότερος ᾖ, τάχιον διέρχεται τὸ ζῴδιον ἀναγκαίως, ὁπόταν δὲ ὑψηλότερος, βράδιον. Cf. the derivative ταπείνωμα ‘the dejection of a planet’ (: the sign opposite to that in which it is exalted; its opposite is represented by ὕψωμα, cf. Plut. Conv. sept. sap. 149a, Ptol. Tetr. 1.20.2 etc.), and ταπείνωσις ‘id.’ (Dorotheus Sidonius fr. 324.5 Pingree).

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  203

(see below, Section 3.1). The morphological structure of ταπεινός (cf. αἰπεινός ‘high’) will also be taken into consideration as I present my argument in the following sections.

 Greek Cognates of ταπεινός At first glance, the adjective ταπεινός seems to be without any cognates in Greek. Despite its formal similarity to Gk. αἰπεινός ‘high’ the etymological dictionaries refer to it as an obscure or even pre-Greek form. As mentioned in Section 1, the only attempt made so far to etymologize this adjective goes back to the beginning of the 20th century when C. Bally (1903, 329; 1942) briefly suggested that Gk. ταπεινός might be related to the place name Τέμπεα/Τέμπη and the noun τὰ τέμπεα ‘defile’, post-Classical ‘valley’, through an unattested *τάπος, and that τὰ τέμπεα may be analyzed as ‘dépression, enfoncement’ by comparing it with Lat. tempus ‘temple’ (‘dépressions latérales du front’). Both Gk. Τέμπεα/Τέμπη, τέμπεα, and Lat. tempus ‘time’, tempus ‘temple’ may be traced back to a neuter stem *témp-es- (Gk. type γένεα, κλέεα) from PIE *temp- which is most likely an extended variant of *ten- ‘to stretch, span’.8 Recently, this reconstruction of τέμπεα has been further developed and, in my view, conclusively discussed by J.L. García Ramón (2014).9 In the next Section (3.1), I will summarize his conclusions on the place name Τέμπεα/Τέμπη and noun τέμπεα as ‘defile, a narrow space, path’, as relevant to the purposes of the present paper. Then, I will bring further evidence in support of Bally’s suggestion by taking into account the possible word formation and original meaning of ταπεινός in Graeco ipso (Section 3.2), as well as by bringing into the picture another possible Greek cognate (Section 3.3).

 8 Cf. Persson 1912, 493: “dann sehe ich […] nicht ein, was uns verbieten könnte, ein τεμποστεμπεσ- ‘Tal’ als ‘Strecke’ zu deuten und demgemäß auf temp ‘spannen’ zurückzuführen”. As for the (nowadays obsolete) interpretation of *temp- as an extended variant of *temh1- ‘cut’ (Gk. τέμνω, τέμενος), see the references in the critical review by Persson 1912, 485–487 (especially regarding the interpretation of Tέμπεα/-η as “Name des tief in das Felsgebirge eingeschnittenen Peneiostales”). 9 Before García Ramón 2014 only P. Chantraine (DELG 1093) mentions Bally’s hypothesis (“hypothèse spécieuse de Bally […] supposant un n. *τάπος et évoquant le toponyme Τέμπεα qui pourrait signifier ‘le Creux’ (et lat. tempus ‘tempe’) mais dont l’étymologie doit être toute differente”).

204  Matilde Serangeli

. Gk. τέμπεα/-η, PN Τέμπη and PIE *temp- ‘to Stretch’ Gk. τέμπεα means ‘a narrow space, path’, more specifically a defile between mountains, also post-Classical ‘valley’ or ‘hollow’, see Hsch. τ 463 Latte/Hansen/Cunningham τὰ τέμπη· τὰ σύνδε(ν)δρα χωρία. τινὲς δὲ τὰ στενὰ τῶν ὀρῶν, schol. Dionys. Per. 219.3 Müller τὰ δὲ κοῖλα τῶν ὀρῶν τέμπη καλοῦνται. In particular, as the ancient historians and lexicographers reveal, the Thessalian place name Τέμπεα/-η is described both as a narrow depression and a narrow path between the mountains Ossa and Pelion through which the Peneios river flows; several authors use it to refer to a valley in the low part of the same river.10 Later on, the place name Τέμπεα/-η is attested outside Thessaly and the noun τέμπεα begins to define any kind of valley or hollow, thus losing its original meaning;11 it is progressively substituted by its synonym τὰ στενά ‘narrow (paths)’ (Ionic στεινά).12 According to García Ramón 2014, this well-attested semantic distribution may be explained as the result of a specialization of a neuter s-stem PIE *témp-esfrom PIE *temp- ‘to stretch’, assuming a semantic shift from (a) *‘stretched out’ —

 10 See: Κραννών, πόλις τῆς Θεσσαλίας τῆς Πελασγιώτιδος ἐν τοῖς Τέμπεσιν ὡς Ἑκαταῖος Εὐρώπῃ (Hecataeus FGrHist 1 F 133); ὁ Πηνειὸς ποταμός, ῥέων διὰ τῶν Τεμπῶν, καὶ ἀρχόμενος ἀπὸ τοῦ Πίνδου ὄρους, καὶ διὰ μέσης Θεσσαλίας […] (Strab. 7.1.14); τόν τε Ὄλυμπον καὶ τὴν Ὄσσαν […] διὰ μέσου τε αὐτῶν αὐλῶνα στεινὸν πυνθανόμενος εἶναι, δι’ οὗ ῥέει ὁ Πηνειός (Hdt. 7.128.1). See also: τὰ καλούμενα Τέμπη τὰ Θετταλικὰ διαγράψω μὲν τῷ λόγῳ (Theopomp. FGrHist 2b.115 F 80); φεῦγε δὲ καὶ Πηνειὸς ἑλισσόμενος διὰ Τεμπέων (Callim. Hymn 4.105); cf. Lat. ad Tempe Thessalica (Liv. 33.35.7). For a more general view, see the description of the historical geography of the course of the Peneios by Helly 2000. 11 Later, Τέμπη is attested as a place name in several Greek regions (Helly 2000, 45−48), see Strab. 9.5.19−20 and 22; Psellus Chronographia 7.34.12 (Cilicia); schol. Dionys. Per. 219.3 Müller (Ethiopia). For attestations of τέμπεα, see: πᾷ ποκ᾽ ἄρ᾽ ἦσθ’, ὅκα Δάφνις ἐτάκετο, πᾷ ποκα Νύμφαι; / ἦ κατὰ Πηνειῶ καλὰ τέμπεα; ἢ κατὰ Πίνδω (Theoc. 1.65–66); and later: σὺ δέ […] / Δελφίδος αὐδήεντα μετέρχεο τέμπεα Πυθοῦς (Nonn. Dion. 2.697–698). As for Lat. tempe ‘valley’ (also ntr. pl.), see: speluncae uiuique lacus et frigida tempe (Verg. G. 2.469). Poetic epithets are also attested, see τεμπή (Plut., Ael.), τεμπίς (Nic.) regarding δάφνη ‘laurel’, Τεμπ-όθεν ‘from Tempe’ (Callim. fr. 194.34 Pfeiffer) and Τεμπήτας ‘id.’ as an epithet of Apollo (Ἄπλουνι Τεμπείτᾳ IG IX,2. 1034, Gyrton, 3rd century BC, cf. Paus. 10.5.9). On Τέμπη cf. the precise summary of Baladié 1996, 292−293. 12 Cf. Suda τ 302 Adler: Τέμπεα· τὰ κοιλώματα and Suda τ 303 Adler: Τέμπη· μεταξὺ τῶν Θετταλίας ὀρῶν περὶ τὸν Ὄλυμπον καὶ τὴν Ὄσσαν στενά· καθόλου δὲ καὶ ἐν πᾶσι τοῖς ὄρεσι στενόταται διεκβολαί. καὶ οἱ σύνδενδροι τόποι. ἰδίως δὲ Μακεδονικὰ ὄρη οὕτω καλούμενα, καὶ οἱ στενοὶ τόποι. Most likely, the specialization as ‘narrow (path)’, passing through the meaning ‘thin, long’, goes back to *‘stretched out’ and is probably only Greek; cf. Hsch. σ 1637 Latte/Hansen στάνει· τείνεται (García Ramón 2014, 27–28).

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  205

Gk. τέμπη ‘valley’; cf. Lat. tempus ‘time (extension)’ — to (b) ‘long, fine’, ‘thin’, whence ‘narrow’ — Gk. τέμπη ‘narrow path, defile’; cf. Lat. tempus ‘temple’ (*‘stretched skin’), templum ‘cross-beam’ (cf. Vitr. 4.2.1 templum: tignum transuersum). See also Lith. tem̃pti ‘stretch’, intr. timpstù ‘become thin’ (with plonėti), tìmpa ‘sinew’, temptýua ‘bowstring’, tamprùs ‘dehnbar, zäh, hart(näckig)’ (ALEW 2.1091), Old Ice. þǫmb- ‘(bow)-string, sinew’. This is supported by the fact that the very same semantic shift is attested for *ten(h2)- ‘stretch, span’: see on the one hand (a) Skt. tána- ‘continuous’ (RV), adv. tánā, Old Ir. tan ‘time’,13 and the verbal forms Ved. tanóti, Gk. τείνω from *ten-ye/o- or *tenh2-ye/o-; τανυ[ε/o]-, Hom. τάνυται (*tn̥ -néw-/*tn̥ -nú-), Old High Germ. donēn ‘be extended’;14 and, on the other hand, (b) the meaning ‘make thin’, *‘made thin’ or simply ‘thin’ preserved in OCS tъniti ‘to make thin’, tętiva ‘sinew’, Gk. νεῦρον, νευρά ‘sinew’, Lat. tenus, -oris ‘a kind of cord, snare’ (Plaut. Bacch. 793: ita intendi tenus), Gk. τόνος ‘cord’ (Hdt., Corpus Hippocraticum+) ‘sinew, tendon’, Ved. tantí- ‘rope’ (RV 6.24.4), ātāná- ‘stretched string of a bow’ (SV, Br.), tántu- ‘rope, thread, cord’ (RV+). Highly relevant to the aim of Sections 4 and 5 is the fact that the corresponding u‑stem adjective *t(e)n(h2)‑u- shows the same pattern as well.15 Consider (a) Lat. tenus (Lucr.+) ‘as far as, up to’, ‘to the extent of’, ‘lengthwise’ (*‘stretching along’), see: urbium Corcyrae tenus ab Aetolia incipienti ‘among the cities (that go) from the borders of Aetolia till Corcyra’ (Liv. 26.24.11); (b) Ved. tanú- ‘thin’, Gk. *τανύς ‘long, thin’ only in compounds τανυ° (e.g. τανύπτερυξ Il. 12.237 ‘longwinged’, beside τανύπτερος Hymn Dem. 89), Lat. tenuis, Old Ice. þunnr, Old Eng. þynne, Old High Germ. dunni ‘dünn’. Moreover, the thematic derivative ταναός (*tn̥ h2ew-ó-) means both ‘outstretched’ (Eur. Or. 322: ταναὸν αἰθέρα ‘outspread ether’) and ‘thin’ (of a javelin: αἰγανέης ταναοῖο, Il. 16.589);16 see also στενός ‘narrow’, most likely a thematization *stenw-ó- of Pre-Gk. *sten-ú- (PIE *(s)ten(h2)‑: see above, n. 12 for a discussion of its semantics). I will return to the u-stem adjective of this root later.

 13 Also in compounds: Gk. °τενής, e.g. ἁλι-τενής ‘reaching into the sea’ (*‘stretched till the sea’, said of πέτρα in Diod. Sic. 3.44). 14 As for ‘to stretch, extend’, see also Ved. tānaya- : Goth. uf-þanjan, Lat. tenēre from caus. *ton‑éye/o- ‘to keep’, beside Lat. tenēt intr. ‘lasts’ (cf. e.g. the fixed expression fama tenet) from *tn̥ -éh1-. 15 On the paradigm and derivatives cf. de Lamberterie 1990, 102−104 and Pinault 2003, 174−175. 16 Especially interesting is τανεῖαι ‘beams’ (Theophr. Hist. pl. 4.1.2), most likely a Greek formation, which matches the semantics of Lat. templum : tignum transuersum (Vitr. 4.2.1) and Lith. °tę́ vas ‘slender’ (said of the body and of body parts), Latv. tiêvs ‘id.’ (said of long and thin objects) from Proto-Balt. *tḛnwa- (*tenh2w-ó).

206  Matilde Serangeli

. Gk. ταπεινός: Word Formation Formally, Gk. ταπεινός may reflect PIE *tm̥p-es-nó-, i.e. the *-no- derivative of a zero-grade s-stem. In accordance with the analysis of Gk. τέμπη, a full-grade *témpes- → tempes-nó- can be expected: see also e.g. ἐλεεινός ‘pitied’, ἐρεβεννός ‘dark’ (*eregwes-), ὀρεινός ‘mountainous’ and the corresponding nouns ἔλεος ‘pity’, ἔρεβος ‘darkness’, ὄρος and the like, respectively. However, the coexistence of zero-grade and full-grade neuters in one and the same lexeme is also attested in Greek, see Class. Gk. βάθος ‘depth’ (: βαθύς ‘deep’), Hom. πάθος ‘sorrow’ : βένθος ‘depth’, πένθος ‘sorrow’.17 Therefore, a zero-grade *tm̥p-es-nó*‘stretched out’ with a zero-grade *τάπος/*ταπες (: *tm̥p-es) is theoretically plausible in Greek. In the Caland system, a u-stem adjective and its corresponding neuter s-stem can display zero grade and full grade or zero grade, respectively, cf. Old Ind. urú‘wide’ and váras- ‘width’ (Risch 1974, 78). I therefore suggest that ταπεινός implies an unattested adjective *ταπύς (< *tm̥p-ú-; cf. Ved. tanú- ‘thin’, Gk. τανυ˚ from the synonymous root PIE *ten-) as well as a neuter s-stem *τέμπος (as allowed by τέμπεα < *témp-es) or *τάπος (cf. βάθος ‘depth’ : βαθύς ‘deep’); cf. the uncertain grade of Lat. tempus < *tm̥p-os or *temp-os and below in this paper the uncertain grade of Hitt. dampu- ‘blunt’ from PIE *tm̥p-ú- or *tomp-u-. This reconstruction, if correct, allows us to define ταπεινός as belonging to the Caland system, too, which parallels the scenario offered by the synonymous root *ten(h2)‘to stretch’:18 see for instance Ved. adj. tanú- ‘thin’ (Br.+), tatá- ‘extended’, uttāná‘stretched’, adj./subst. tániman- masc./ntr. (?) ‘bow-string’ (TS+). Moreover, it might be observed that the morphological structure of Gk. ταπεινός is similar to that of αἰπεινός ‘high’ (Hom.+) which, together with the adj. αἰπύς ‘id.’ (Hom.+) and the subst. αἶπος ‘height’ (Aesch.), exemplifies the so-called Greek Caland system (Risch 1974, 100).19

 17 There is apparently no unique explanation of such doublets. In the case of βάθος/βένθος ‘depth’ the former term must have been influenced by the adj. βαθύς ‘deep’ and the latter term must be the archaic form (Meissner 2006, 65–67), while in the doublet πάθος/πένθος ‘sorrow’ the former term seems to be clearly influenced by the zero grade of the verb πάσχω (Meissner 2006, 67–68). For an attempt to analyze βένθος as an artificial formation, see Peters 1999, 450. 18 Moreover, both roots may be classified under the same semantic category of physical features and dimensions attested in the Caland system (Rau 2009, 84–85). 19 The two terms are also attested in the same context, i.e. defining the ‘sky’, see above the example (1) and in addition Bacchyl. 3.35−36 χέρας δ’ [ἐς / αἰ]πὺν αἰθέρα σ[φ]ετέρας ἀείρας; Soph. Aj. 845 σὺ δ’, ὦ τὸν αἰπὺν οὐρανὸν διφρηλατῶν.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  207

As for semantics, ταπεινός can be explained by assuming a semantic shift from *‘stretched out (in length)’ to ‘low’, maybe passing through a meaning *‘lowered’ due to a reduction in height ensued from the process of stretching out, similar to the semantic shift from *‘stretched out (in length)’ to ‘thin’, ‘long’, in which the original thickness of a given substance has been reduced through the process of stretching out (see above, Section 3.1). In both cases, the actant reduces its height or thickness as a result of increase in length. As a similar outcome of the same process, which may also eventually be explained as a specialization of the more general meaning ‘low’, a meaning ‘low-lying’ is also well attested in Greek. This is easily conceivable, especially in the case of a place, as has been shown above (Section 1), including for a territory that is ‘low-lying’ (in comparison to other places): see ἡ (scil. χώρη) μὲν γὰρ δὴ πρὸς τὴν ἠῶ τῆς Λιβύης [...] ἐστὶ ταπεινή τε καὶ ψαμμώδης μέχρι τοῦ Τρίτωνος ποταμοῦ ‘and the eastern regions of Libya […] are low-lying and sandy as far as the river Triton’ (Hdt. 4.191.3); διὸ καὶ τὰ περὶ τὴν Λιβύην τὴν Ἀμμωνίαν χώραν ταπεινότερα φαίνεται καὶ κοιλότερα παρὰ λόγον τῆς κάτωθεν χώρας ‘for the same reason Libya, the country of Ammon, is lower and hollower than the land to the seaward of it’ (Arist. Metaph. 352b.32). By contrast, a meaning ‘low’, most likely to be taken in the sense of ‘flat’ or in general ‘non-steep’, is demonstrated by the use of ταπεινός to describe an island — an island can be ‘low’ but surely cannot be ‘low-lying’: see νῆσοι γάρ εἰσι […] ταπειναί ‘the islands are indeed […] low’ (Diod. Sic. 3.21.1). That being said, I suggest that we are dealing with a no-adjective that may be built on a neuter s-stem and an unattested zero-grade u-stem adjective. In the following paragraph I aim to show that a derivative in -ro- may be added to the dossier, too. If the term is not just a secondary Greek formation due to the productivity of the derivatives in ‑(u)ro‑ in Greek, it could suggest the existence of a ustem of *temp‑ in Greek.

. The Place Name Τέμπυρα and uro-Formations The Thracian place name Τέμπυρα may be added to the material presented so far. Livy offers an illuminating description of its landscape, which can hardly be separated from the description of Τέμπη offered by the Greek historians and lexicographers (see above, Section 3.1 and notes) in many aspects:20

 20 For more mentions, see Strab. 7 fr. 47 Baladié; Ov. Tr. 1.10.21–22.

208  Matilde Serangeli (6) Eo die ad Hebrum flumen perventum est. Inde Aeniorum fines praeter Apollinis Zerynthium quem vocant incolae templum superant. Aliae angustiae circa Tempyra excipiunt — hoc loco nomen est — nec minus confragosae quam priores; sed, quia nihil silvestre circa est, ne latebras quidem ad insidiandum praebent. Huc ad eandem spem praedae Thrausi, gens et ipsa Thraecum, convenere; sed, quia nudae valles procul ut conspicerentur angustias obsidentes efficiebant, minus terroris tumultusque fuit apud Romanos; quippe etsi iniquo loco, proelio tamen iusto, acie aperta, signis collatis dimicandum erat. Conferti subeunt et cum clamore impetu facto primum expulere loco hostes, deinde avertere; fuga inde caedesque suis ipsos impedientibus angustiis fieri coepta est. (Liv. 38.41.4−7) That day they reached the River Hebrus, and after that they crossed the frontier of the Aenians, passing by the temple of Apollo, whom the local people call Zerynthius. There follows another defile near Tempyra (such is the place’s name), no less rugged than the earlier one; but because there is no woodland in the area it does not offer hiding places for an ambush. Here the Trausi, also a Thracian people, gathered with similar hopes of plunder; but since the treeless valleys rendered them visible in the distance as they were blocking the pass, there was less panic and consternation among the Romans because, despite the unfavorable terrain, it was still a regular engagement that had to be fought, with lines in the open and troops deployed. The Romans approached in close formation and attacked with a shout, at first dislodging and then driving back the enemy. Now began flight and slaughter because the narrow terrain, which they themselves had chosen, obstructed the Thracians. (transl. J.C. Yardley, slightly modified)

Τέμπυρα can hardly be anything but a ro-adjective, a result of the productivity of the adjectives in ‑uro‑ in Greek, regardless of suggestions for an inherited u‑noun nor u‑adjective. However, Τέμπυρα could also reflect the substantivization of PIE *temp‑u‑ro‑ → *temp-u-reh2-.21 In that case, a uro-adjective and a u-noun (the common base for a uro‑adjective), both displaying full grade, might be tentatively posited as derived from PIE *temp‑. It is clear that it is only possible to advance a hypothesis for Τέμπυρα and that inherited uro-adjectives occurring along with the corresponding u-noun and adjective are very rare in Greek. For instance, see the adjectives λιγυρός and λιγύς ‘loud’, although likely a dissimilated form from *λιγυλός (de Lamberterie 1990, 710−711). However, the derivatives in *-uro-, *-esand *-no- suggested for Τέμπυρα (*τεμπυρός), τέμπεα, and ταπεινός would find a very close parallel Caland structure in a group of terms derived from PIE *h2erǵ‘brilliant white’: the inherited substantivized uro-adjective ἄργυρος ‘white metal, silver’ (Il.+; Myc. a-ku-ro-: de Lamberterie 1990, 711 n. 7), which together with  21 Tentatively, the anthroponym Ταπυρος (Priene, 1st century AD, see gen. sing. IK.Priene 326.10 Ṭαπύρου) would appear to be a possible candidate for *tm̥p-u-ro-, a zero-grade counterpart of Τέμπυρα which would attest to a zero-grade u-stem substantive, which makes an adjective *ταπύς more conceivable, as was proposed above.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  209

ἀργυφέος ‘shining’ and ἄργυφος ‘white’ (tentatively from a formation in *-u-bho-) speaks for an unattested u-formation *ἀργυ- (Risch 1974, 133; 171); the s-stem Ἄργος ‘the white city’ (and ˚ἀργής, -ές ‘clear’); the no-adjective ἀργεννός ‘white’, which is certainly built on a s-stem, see ἀργεστής (Il.+, epithet of the wind).22 In particular, see the ro-adjective ἀργός ‘shining, white, fast’ < *ἀργρός attested in Old Ind. ṛjrá- ‘id.’ (cf. ṛjásva- ‘who has fast feet’ (RV), Myc. po-da-ko, alphabetic Gk. Πόδαργος), the no-adjective Old Ind. árjuna- ‘shining, white’ (RV, AV+), the to‑adjective Lat. argūtus ‘clear thinking, intelligent’ (Naev.+) clearly built on a ustem, and the -ye/o-denom. arguĕre (Plaut.+) ‘to make clear’. In addition, it is also interesting that *τεμπυρός might have a parallel in Gk. στενυγρός ‘narrow’, the Ionic form for στενός, which has been presented in the previous paragraph as the substitutory continuant of *τέμπος, pl. τέμπεα (*temp‑es-). The analysis of this term remains disputed but it can be tentatively traced back to *(s)ten-u-ro-, implying that °υγρός corresponds to *-u-(g)-ro- with secondary introduction of a velar; cf. Gk. (Dor.-Aeol.) κᾶρυξ ‘herald’, Ved. kārú(de Lamberterie 1990, 261–262).

. Partial Conclusions on Greek To sum up, Greek offers a neuter τέμπεα (: *témp-es-) which allows us to posit an s-stem *τέμπος for PIE *temp- in Greek, on which the no-adjective ταπεινός may have been built on (*tm̥p-es-no-). Comparing these two terms with the so-called Greek Caland system, a parallel structure may be found in the type αἰπεινός ‘high’ (derivative in -no-), αἰπύς ‘high’ (derivative in -u-), αἶπος ‘height’ (derivative in ‑s‑): this parallel would lead to the assumption that a u-stem from PIE *temp- may also be added to ταπεινός and τέμπεα. However, this hypothesis presupposes an unexpected zero-grade s-stem *ταπες, *τάπος instead of the expected *τέμπος and an unattested u-adjective *ταπύς *‘stretched out’ (: *tm̥p-ú-). On the one hand, the zero grade of Gk. *ταπες, *τάπος may be due to a zero-grade u-adjective *ταπύς; cf. above the cases like βάθος ‘depth’ and βαθύς ‘deep’ vs. βένθος ‘depth’. On the other hand, the existence of a u-adjective in Greek for PIE *temp- can be supported by the existence of *tn̥ h2ú- ‘thin, long’ and of the thematic derivative *tn̥ h2ew-ó- of the synonymous PIE *tenh2- and of the thematization *stenw-ó- of Pre-Gk. *sten-ú- (Pre-Gk. *sten‑) as reflected in Gk. τανυ° ‘thin, long’ (: Ved. tanú-),

 22 See Stüber 2002, 55. The Greek Caland system is by definition a secondary development. Therefore, the fact that in the case of ἄργυρος we are dealing with a nominal root is not relevant to this discourse.

210  Matilde Serangeli Gk. ταναός meaning both ‘outstretched’ (from *‘stretched out’) and ‘thin’, and in Att. στενός, Ion. and Hom. στεινός ‘narrow’, respectively. Despite the lack of good parallels for a u-stem noun and the corresponding uro-adjective, the close parallel between Τέμπυρα (*τεμπυρός), τέμπεα, and ταπεινός and ἄργυρος (and ἀργυφέος, ἄργυφος), Ἄργος, and ἀργεννός seems to support the hypothesis of a u-stem derivative for *temp- as well. These considerations are summarized in the following table: Tab. 1: Greek cognates of ταπεινός no-adjective

u-stem

s-stem *τέμπος, τέμπεα (*temp‑es-)

ταπεινός (*tm̥p‑es‑no‑)

*ταπύς (*tm̥p-ú-)

*τάπος (*tap-os- < *tm̥p‑os-)

cf. αἰπεινός ‘high, steep’ (Hom.+)

cf. αἰπύς ‘high’ (Hom.+)

cf. αἶπος ‘height’ (Aesch.)

Τέμπυρα? (*τεμπυρός?)

As for the semantics, the data presented in this Greek section suggest that we are dealing with at least two semantic shifts: from an original meaning (a) *‘stretched out (in length)’ to a meaning (b) ‘thin, long’ and (b1) ‘narrow’ as has been previously suggested for τέμπεα, and (c) ‘low’ and ‘low-lying’ as it seems to be in the case of ταπεινός. As for Τέμπυρα, one can only speculate about a precise meaning. However, the description offered by Livy, which is terribly reminiscent of several descriptions of Τέμπεα, may suggest an interpretation as ‘valley’ (: *‘the stretched one’) or ‘narrow path, defile’ (: *‘the narrow one’). Now that the formal and semantic analysis of ταπεινός and its Greek cognates has been presented, the comparative evidence might help us elucidate some aspects of this scenario, which could not otherwise find a complete explanation exclusively within Greek. In particular, the existence of a zero-grade u-adjective in another Indo-European branch would be helpful to support the hypothesis that the system suggested for ταπεινός might be found.23

 23 A uro-adjective does not speak directly for a u-adjective, of course, but the existence of a uadjective giving a u-substantive is easily conceivable.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  211

 The Comparative Evidence: Anatolian and Slavic Cognates The aim of this section is to provide evidence for PIE *temp- from two other IE branches, namely Anatolian and Slavic. First, it will be shown that a u-stem adjective and a tomós-adjective may be reconstructed for *temp- in Anatolian, see Hitt. dampu- ‘blunt’ and its derivative dampupi- ‘barbarian’ (and dampupatar ‘barbarity’), and Slavic, see Old Russ. tupъ ‘blunt, not sharp, stupid’, respectively. Formally, the Hittite adjective may be exactly the missing link between the Greek τέμπεα, Τέμπυρα, and ταπεινός, as suggested in the previous Sections 3.3– 3.4. The latter, for which a PIE reconstruction and derivation from *temp- is secure, offers a perfect semantic parallel to the Anatolian form and at the same time confirms, at least for a derivative, a meaning ‘blunt’ for *temp-, which otherwise does not appear in the dossier presented so far. Second, I will explain how the Anatolian forms may be integrated in the Greek dossier and how a meaning ‘blunt, not sharp; barbarian, uneducated’ (Anatolian)/‘stupid’ (Slavic) and a meaning ‘low(-lying), miserable’ (Greek) can be two different outcomes originating in the same meaning *‘stretched out (in length)’.

. The Anatolian Branch The Hittite adjective dampu- ‘blunt’ is attested only twice. Both attestations belong to the following Akkadian-Hittite bilingual lunar omen inscription:24 (7) ták-ku dEN.ZU-aš kap-pa-an-za nu ZA[(G-an SI-ŠÚ)] al-pu | SI GÙB-la-ma-aš-ši dam-pu LÚKÚR [(KUR-e)] e-e[(p-zi)]. (KUB 29.11, 1b)

 24 Notice in the text the opposition between Hitt. dampu- ‘blunt’ and alpu- ‘sharp, pointed’; cf. alpuemar ‘point, tip’, alpuēšš-mi ‘to be sharp, to be acute’ (Güterbock 1988, 170). As for the semantics of alpu-, see the clear example of its derivative alpuemar: KBo 11.14 I 12−14 Ù ANA GU4 APIN.LÁ-kan huišu̯ anti ANA SIḪI.A-ŠU alpuemar tepu kuranzi nu-kan apei̯a anda ANA NINDA.GUR4.RA dāi “und man schneidet ein wenig (von der) Spitze der Hörner eines lebenden Pflugrindes ab, und er legt diese Abschnitte auf das Dickbrot” (HW A, 61). The opposing view of Puhvel 1975, 61, that alpu- is connected with Lith. alpùs ‘weak’, following a translation dampu‘sharp’ / alpu- ‘blunt’ (Riemschneider 1961, 25−26; Riemschneider 2004, 108–109), has to be abandoned, since this attempt is based on the position of the Akkadian term for ‘sharp’ after the term for ‘blunt’, which simply does not match the Hittite word order.

212  Matilde Serangeli If the moon is pale (?) and the right horn is pointed, but the left horn is blunt: the enemy will conquer the country. (7.1) ták-ku dEN.ZU-aš SIG7-u̯ a-an-za nu SI GÙB al-pu SI ZAG-ma-aš-ši | dam-pu MU.2.KAM ḫa-meeš-ḫa-an-za SIG5-at-ta. (KUB 29.11, 7−8b) If the moon is pale and the left horn is pointed, but the right horn is blunt: the spring will (within) two years be good.

In addition to this, Hitt. dampupi- ‘rough, primitive, barbarian’ (with determinative LÚdampupiš, see KUB 58.68.23; KBo 9.111.8) and its derivative dampupatar‘barbarity’ can hardly be anything but a secondary Hittite formation built on dampu-, which provides evidence for a semantic shift from the physical sense of ‘blunt’ to the moral one ‘rough, uneducated, barbarian’ (Puhvel 1977, 599):25 (8) ták-ku LÚ.U19.LU]-an dam-pu-pí-en ku-iš-ki uš-ne-eš-kat-ta.

(OH, Hittite Laws § 147/*36)

If someone is in the process of selling an uneducated person. (8.1) šu-me-in-za-an KUR-e dam-pu-u-pí ku-it an-da-at za[(aḫ-ḫa-)]an ŠA ǀ ŠEŠ-ŠU-za NIN-ŠU MUNUS a-a-an-ni-in-ni-i̯a-mi-in da-aš-ká[(n-zi-pát)] ǀ URUḪa-at-tu-ši-ma-at Ú[(-UL a)]-a-ra. (KBo 5.3 iii 32. = CTH 42) Weil euer Land ungesittet (ist) und darin geprügelt (ist), so nimmt man sich des eigenen Bruders (seine) Schwester (oder) die Kusine; in Hattusa aber ist das nicht erlaubt. (transl. HEG T/1:88) (9) ZI-at-ta dam-pu-pa-tar t[e-ik-ku-uš-nu-du-du].

(KBo 18.28, 9)

Und sie (Ištar) soll dir die Barbarei der Seele z[eigen]. (transl. Hagenbuchner 1989, 406)

Now that the semantics of dampu-, dampupi-, and dampupatar have been established, it remains to be clarified why and how these terms are relevant to the aim of this paper. First, Hitt. dampu- may provide evidence for the existence of a ustem adjective for PIE *temp-. It would represent, therefore, the missing link between Gk. Τέμπεα/-η, Τέμπυρα, and ταπεινός. Unfortunately, it is not possible to

 25 The meaning of the term is also assured by the Hittite-Akkadian vocabulary, see KBo 1.30 I 8−9 Hitt. dampupiš = Akk. nu-ʼ-ú ‘dumb, barbarian’ = Sum. LÚ AŠ.ḪAB ‘cad, villain’ (CAD 7 I/J 189; AHW 799, 387).

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  213

ascertain whether Hitt. dampu- reflects a formation with the zero grade (PIE *tm̥p‑ú-, where IE */m̥/ > Hitt. /am/) or the o-grade (PIE *tomp-u-, PIE */o/ > Hitt. /a/) of the root. However, it is noteworthy that other Hittite u-stems display zero‑ or e-grade, which, in my view, seems to support an analysis as zero grade for dampu-; cf. alpu- ‘sharp’ from *l̥p-u- (cf. Welsh llym, Bret. lemm ‘sharp’ < *l̥p-s-mo-: see Hamp 1989, 21–22), parku- ‘high’ from *bhr̥ ǵh-u-, and Hitt. tēpu- ‘little’ from *dhébh-u-.26 Second, dampupi- ‘barbarian, uneducated’ is most likely a secondary formation built on dampu- by the addition of a suffix -(u)pi- of unclear origin. Hitt. -upi- is attested in very few terms, which may well be loanwords and belong to a different semantic field: thus, we find tools and vessels like kullupi- ‘agricultural tool’, hurupi-, tilupi- ‘vessels’, kinūpi- ‘a container’, as well as bird names like sulupi- and hapupi- ‘owl’. Third, the derivative dumpupatar- clearly proves that the upi-suffix is fully integrated in Hittite. However, tentatively -upi- could be analyzed as a u-stem plus a suffix -pi- which may reflect the i-mutated outcome of a thematized suffix *-bhH-o‑/bhh2‑o‑ (i.e., the addition of a ī-suffix typical of the Luwian languages to mark the nominative and accusative [sing. and pl.] cases of the common gender). In light of this, I would like to recall the Greek suffix -(υ)φο‑, tentatively from *-u-bho-, which may support the hypothesis of a IE origin for both Hitt. dampu- and dampupi- and the existence of a u‑stem for *temp-: see above in Section 3.3 the adjective Gk. ἀργυφέος/ἄργυφος ‘(silver-)white’ and the whole family of derivatives of PIE *h2erǵ- ‘brilliant white’.27 That being said, further considerations and questions arise. First, the reconstruction of dampu- as a zero-grade u-stem derivative from PIE *temp- is formally possible. If this hypothesis is correct, this term may represent the missing u-stem  26 In addition to dampu- see the fientive Hitt. tampuēšš-mi ‘to become blunt (?)’ (damaged context; HEG T/1.3, 88; EDH 826); cf. e.g. idalu- ‘evil’, idālauešš-mi ‘to become evil, hostile’. As for the suff. ‑ešš- (*-eh1-s-), see the derivation by addition (synchronically expected pattern in (pre-)Hittite), e.g. ḫatku- ‘narrow’ → ḫatku-ešš- ‘to become narrow’, vs. derivation by substitution, e.g. parku-i‘pure, clear’ → parku-e- ‘to become clear’, parku-nu- ‘to purify’ (tampuēšš- implies tampuē-*, since -e- → -ešš-); cf. the attestation of both -e- (*-eh1- “denominal statives”, in some cases with stative or fientive meaning; see Watkins 1971, 67) and ešš-verbs (*-eh1-s- “denominal fientives”) for the same adjectival root: arau̯ a- ‘free’ > arau̯ e- ‘to declare oneself free from’ → arau̯ ešš- ‘to become exempt from’, GÉME-a/iššara- ‘female slave’ → GÉME-a/iššare- ‘to become a female slave’ > GÉME-a/iššarešš- ‘id.’ (see Hoffner 1998). 27 C. Melchert helped me notice that dampupi- might also be modelled on bird names by associative semantics as, e.g., the repetition of two labial stops, and that the rounded back vowel is strongly associated in PIE with stupidity/ignorance; cf. Eng. bobo, boob(y), bumble, It. bimbo ‘dumb’, among others (C. Melchert, personal communication). However, should the match between the Greek and Anatolian suffixes prove correct, this alternative option would be less probable.

214  Matilde Serangeli in the reconstruction of the Greek system suggested above. Tentatively, Hitt. dampu- may correspond to an unattested Greek *ταπύς and go back to PIE *tm̥p‑ú-; dampupi- may correspond to an unattested Greek *τάπυφος and go back to a u-stem plus a suffix *-bho- that may support the existence of a u-stem for *temp‑: see as a parallel formation ἀργυφέος/ἄργυφος.28 Second, the meanings ‘blunt’ and ‘low’ or ‘low-lying’ may not, at first glance, appear related, and their derivation from ‘*stretched out (in length)’ may not seem immediately obvious. However, let us recall two comments on the Greek material above in Sections 3.1–2. Regarding the shift *‘stretched out (in length)’ → ‘low(-lying)’, something *‘stretched out (in length)’ reduces by definition its vertical extension or its thickness and, as a result of this lowering, becomes low and in some cases can be low-lying. Note that at least in one case Greek ταπεινός is most likely to be taken in the sense of ‘flat’ or in general ‘non-steep’, as in the previously mentioned example νῆσοι γάρ εἰσι […] ταπειναί ‘the islands are indeed […] low’ (Diod. Sic. 3.21.1): islands cannot be ‘low-lying’ but are more likely characterized by low mountains or the like. However, a meaning ‘blunt’ can be the outcome of the same semantic development: the stretching of a peaked surface may lead to the reduction or loss of the peak and to the creation of a blunt (or in some cases flat) surface. In both cases, the actant reduces its height or thickness as a result of increase in length. Lastly, it may be noted that a meaning *‘not sharp, blunt’ shows in Hittite (more precisely in its upi-derivative) a negative figurative sense as ‘barbarian, not educated’ and that Gk. ταπεινός ‘low(-lying)’ assumes the figurative sense of ‘miserable’ (see by contrast Lat. acūtus ‘sharp, smart’). Thus, both terms show a semantic shift into a pejorative sense applied to people.

 28 In addition to this data, Lycian, a so-called minor Anatolian language (South-Western Turkey, 5th–3th century BC), might offer a group of opaque lemmata which are probably etymologically connected with Hitt. dampu- and suggest (together with the fientive formations in Hittite, see above note 27) that *temp- belongs to the Caland system, see tm̃pere/i‑ (*ro-adj.), tẽpe/i‑ (*tómos type) and probably also *tm̃pe- (*tomós type), tm̃pewẽti‑ (*‑went‑adjective): noun/adj./PN (?) tm̃pere/i‑ (1x, TL 65.7), PN Tettm̃pe (1x, TL 75.1-2; KPN § 1542), which is a reduplicated form from Lyc. tẽpe- or *tm̃pe- (cf. PN mẽmruwi, TL 39.2; KPN § 897), and Hitt. maruwaiwith Glossenkeil ‘to blacken (?)’ from Luw. (?) *marwa- ‘black’; Lyc. B subst. tẽpe- (1x, TL 44d.63), cf. Lyc. A subst. tẽpina- (?) (1x, TL 29.13; possibly with -(i)-no/na-suffix; Tekoğlu 2006, 1704 and 1706 endorses a reading Tẽrina or Tẽpina, perhaps to be interpreted as a PN); noun/adj. tm̃pewẽti‑ (2x, TL 44c.58: case unclear). I will treat this topic elsewhere.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  215

. The Slavic Branch In addition to Hitt. dampu-, Slavic offers further evidence of the meaning ‘blunt, not sharp’, whence a figurative pejorative sense ‘stupid’, for *temp-. Thus compare the semantics of Old Russ. tupъ ‘blunt, not sharp, stupid’ (cf. also Pol. tępy ‘stupid’, Bulg. tăp ‘id’):29 (10) ‘not sharp, blunt’: Ne gladъko i ostro i tupo.

(John the Exarch Bog. 191)

Not smooth and sharp and blunt. (10.1) Mjakъko že i žestoko, plazivo i ostro i tupo, tjažьko že i lьgъko.

(John the Exarch Bog. 193)

Soft and hard, slippery and sharp and blunt, heavy and light. (11) ‘not intelligent’: Někto tvorecь tupъ svoe tvorenie ispěvaše.

(Pčela Imp. publičnoj biblioteki l. 55)

Some stupid creator was singing about his creation.

Moreover, Old Russ. tupъ perfectly matches with the Old Icelandic adjective *þambr ‘stretched out’ (only attested in the feminine þǫmb ‘swollen, fat’, see Heidermanns 1993, 611) and its feminine substantivized adjective þǫmb ‘bow-string’, going back to Proto-Ger. *þamb-a- ‘stretched out’ from PIE *tomp-ó- and ProtoGer. *þamb-ō- ‘stretched out’ < Transponat *tomp-éh2-, respectively. This exact formal parallel supports the hypothesis put forward in this paper that PIE *tempmay have undergone a semantic shift from *‘stretched out’, which led to different outcomes. In other words, despite showing both a *tomós inherited formation, the Slavic and Germanic terms feature, respectively, the semantic nuances of ‘blunt’ and ‘thin’ (for which see the meaning ‘bow-string’; cf. also Section 3.2).

 29 I follow here the Slavic passages mentioned by Sreznevskij 1893–1912 s.v.: (1) John the Exarch, Bogoslovie sv. Іoanna Damaskina vъ perevodě Іoanna, ekzarxa Bolgarskago, po sp. Moskovsk. sinod. bibl. do 1200 g. (with the meaning ‘blunt’); and (2) Pčela Imp. publičnoj biblioteki vъ sp. XIV–XV v. (with the meaning ‘stupid’).

216  Matilde Serangeli

 Conclusions In this paper an attempt has been made to trace Gk. ταπεινός ‘low(-lying)’, ‘miserable’ back to *tm̥p-es-nó-, i.e., a derivative in *-no- of PIE *temp-, assuming a semantic shift from *‘stretched out (in length)’ to ‘low’ and ‘low-lying’, whence a pejorative sense ‘miserable’. The term has been analyzed as belonging to the socalled Greek Caland system (type αἰπεινός ‘high’ – αἰπύς ‘high’ – αἶπος ‘height’) and, therefore, the hypothesis of an unattested u-stem derivative *ταπύς (*tm̥p‑ú‑) and an s-stem derivative *τάπος (*tm̥p-os-) has been put forward. The existence of the aforementioned derivatives is easily conceivable thanks to the full-grade s-stem Greek τέμπεα ‘defile, valley, narrow path between two mountains’ and to some Greek pairs of the type βάθος ‘depth’, βαθύς ‘deep’ vs. βένθος ‘depth’. Consequently, I have advanced the hypothesis that we should identify a zero-grade u-stem of PIE *temp- in Hitt. dampu- ‘blunt’ and its derivative dampupi- ‘not educated’. These two Anatolian terms constitute the missing link for a promising analysis of ταπεινός and its relation to its Greek cognates on a formal and on a semantic level. Semantically, the different meanings of ταπεινός ‘low(-lying)’ and dampu‘blunt, not sharp’ have been explained by the hypothesis that PIE *temp- may have undergone a semantic shift from *‘stretched out (in length)’ to a meaning of (a) ‘low (*lowered)’, and (b) ‘blunt (*not peaked)’ in several IE languages (in addition to the well-known (c) ‘thin, long’), most likely passing through the reduction of the original height or thickness of a given substance due to stretching in length. Despite the fact that the attestation of (a) and (b) is restricted to few derivatives, the aforementioned shift is not likely to be relevant only to those specific derivatives, since different inherited derivative formations share the same semantic shift, see, e.g., the u-stem Hitt. dampu- ‘blunt’, dampupi- ‘barbarian, not educated’ and the tomós-type Old Russ. tupъ ‘blunt, not sharp, stupid’. To conclude, this study has shown that Gk. ταπεινός is not isolated within Greek nor within the Indo-European languages and that its meaning is the result of one of the semantic shifts that may be reconstructed for PIE *temp- ‘stretch, span’. While a pattern for the several semantic shifts of this root can be defined, it remains more difficult to determine the stage at which each shift took place.

Gk. ταπεινός ‘Low(-lying)’ and Its IE Heritage  217

Bibliography Baladié, R. (1996), Strabon: Géographie, vol. 9, Paris. Bally, C. (1903), “Contribution à la théorie du z voyelle”, in: Mémoires de la Société de Linguistique de Paris 12, 314−330. Bally, C. (1942), “Latin tempora, grec Τέμπη, ταπεινός”, in: CFS 2, 58−59. de Lamberterie, C. (1990), Les adjectives grec en -υς: Sémantique et comparaison, Louvain-laNeuve. García Ramón, J.L. (2014), “The Place-Name Τέμπη, τέμπεα· … τὰ στενὰ τῶν ὀρῶν (Hsch.), IE *temp- ‘stretch’”, in: H.C. Melchert/E. Rieken/T. Steer (eds.), Munus amicitiae Norbert Oettinger a collegis et amicis dicatum, New York, 19−31. Güterbock, H.G. (1988), “Bilingual Moon Omens from Boğazköy”, in: E. Leicht/P. Gerardi/ A. Sachs (eds.), A Scientific Humanist: Studies in Memory of Abraham Sachs, Philadelphia, 161−173. Hagenbuchner, A. (1989), Die Korrespondenz der Hethiter, vol. 2: Die Briefe mit Transkription, Übersetzung und Kommentar, Heidelberg. Hamp, E.P. (1989), “Hittite alpu and dampu”, in: HSF 102, 21−22. Heidermanns, F. (1993), Etymologisches Wörterbuch der germanischen Primärdjektive, Berlin/ New York. Helly, B. (2000), “La description du Pénée thessalien par Strabon: Éléments d’une représentation de l’espace géographique chez les Anciens”, in: A. Bonnafé/J.-C. Decourt/B. Helly (eds.), L’espace et ses représentations, Lyon, 25−71. Hill, E. (2013), “Historical Phonology in Serving of Subgrouping: Two Laws of Final Syllables in the Common Prehistory of Baltic and Slavonic”, in: Baltistica 48, 161−204. Hoffner, H.A. Jr. (1998), “On the Denominative Verb arāwe”, in: J. Jasanoff/H.C. Melchert/L. Olive (eds.), Mír Curad: Studies in Honor of Calvert Watkins, Innsbruck, 275–284. Meissner, T. (2006), S-stem Nouns and Adjectives in Greek and Proto-Indo-European: A Diachronic Study in Word Formation, Oxford. Persson, P. (1912), Beiträge zu der indogermanischen Wortforschung, Uppsala. Peters, M. (1999), “Ein tiefes Problem”, in: H. Eichner/H.C. Luschützky (eds.), Compositiones Indogermanicae in memoriam Jochem Schindler, Prague, 447–456. Pinault, G.-J. (2003), “Sur les thèmes indo-européens en *-u-: Dérivation et étymologie”, in: E. Tichy/D.S. Wodtko (eds.), Indogermanisches Nomen: Derivation, Flexion und Ablaut. Akten der Arbeitstagung der Indogermanischen Gesellschaft (Freiburg, 19. bis 22. September 2001), Bremen, 153–188. Popko, M. (1974), “Notes on Hittite Vocabulary”, in: Journal of Cuneiform Studies 26, 181–182. Puhvel, J. (1975), “Hittite alpu- and dampu-“, in: Revue hittite et asianique 33, 59–62. Puhvel, J. (1977), “Review of J. Friedrich/A. Kammenhuber, Hethitisches Wörterbuch: Zweite, völlig neubearbeitete Auflage auf der Grundlage der editierten hethitischen Texte. Lieferung 1 (Heidelberg 1975)”, in: JAOS 97, 596–599. Rau, J. (1999), Indo-European Nominal Morphology: The Decads and the Caland System, Innsbruck. Riemschneider, K.K. (1961), “Review of J. Friedrich/R. von Kienle/A. Moortgat (eds.), Festschrift Johannes Friedrich zum 65. Geburtstag am 27. August 1958 gewidmet (Heidelberg 1959)” in: BO 18, 23–29.

218  Matilde Serangeli Riemschneider, K.K. (2004), Die akkadischen und hethitischen Omentexte aus Boğazköy, Dresden. Risch, E. (1974), Wortbildung der homerischen Sprache, 2nd edn., Berlin/New York. Sreznevskij, I.I. (1893–1912), Materialy dlja slovarja drevnerusskago jazyka, 3 vols., Saint Petersburg. Stüber, K. (2002), Die primären -s-Stämme des Indogermanischen, Wiesbaden. Tekoğlu, R. (2006), “TL 29: Una nuova proposta di lettura”, in: R. Bombi/G. Cifoletti/F. Fusco/ L. Innocente/V. Orioles (eds.), Studi linguistici in onore di Roberto Gusmani, Alessandria, 1703–1710. Watkins, C. (1971), “Hittite and Indo-European Studies: The Denominative Statives in -ē-”, in: TPhS 104, 51–93.

Enrico Cerroni

Prose and Poetry of Pain: A History of the Term ἄλγος Abstract: In this paper I investigate the history of the term ἄλγος ‘pain’ as an interesting case of how a common word in Homeric epic later became to be considered more rarified, turning into a refined and poetic term by the 5th century BC. Despite its frequent occurrence in tragedy, its use is in fact only attested once in Herodotus (5.49.2), never in Thucydides, and only fourteen times in the Hippocratic corpus, where a clear tendency to innovate the vocabulary of pain can be observed. Indeed, the highly objective neologism ἄλγημα, the term with the most emotionally negative charge, λύπη, and the generic πόνος, capable of expressing a broad semantic spectrum — ‘physical fatigue’ but also ‘pain’ — gradually came to be preferred over the poignant epicism. A review of literary texts that revive ἄλγος in the Hellenistic and Imperial periods, such as the Septuagint and Eusebius of Caesarea’s Life of Constantine, helps us to better appreciate the stylistic tenor of passages that have often been insufficiently studied from a lexical perspective. Keywords: Epicisms, Ionic dialect, paratragedy, pain, medical language, poetic terms.

 Origins Some words thrive in literary language for a certain period of time, then progressively fade until they completely disappear.1 Others gradually establish themselves from very humble origins to become prestigious even in the most refined literature. The reasons for this are manifold, and are sometimes difficult, if not impossible, to assess. This is particularly the case for many ancient languages, including Ancient Greek, due to a shortage of evidence.

 1 A special thanks goes to Ettore Cingano, Filippomaria Pontani, and Michele Napolitano for the suggestions they gave me when presenting this paper at Ca’ Foscari on 28th April 2017. I owe Albio Cesare Cassio more than words can express. Beyond Classical philology, I am indebted to him for having learned a specific art for rationalizing and overcoming the ἄλγεα with the discrete force of ἐπιστήμη and χάρις. I am also very grateful to Olga Tribulato and Enzo Passa for their advice and insightful discussions about this paper after the conference. https://doi.org/10.1515/9783110621747-011

220  Enrico Cerroni The Greek language appears so closely linked to poetry in its early stages that we must often resist the temptation to consider a significant part of the Greek vocabulary ‘poetic’ simply because we have few prose texts from this period. Differentiating the words that are poetically marked from those that are not may prove to be a very frustrating exercise. The natural first step in such an endeavour is confronting the Homeric epics, the fundamental poetry of Greek culture, which included a large variety of words, some of which have survived into Modern Greek. One of these words is ἄλγος ‘pain’,2 which is still prevalent today in several compound forms, especially in medical terminology.3 As with other Homeric terms, ἄλγος piqued the interest of Hesychius, who probably regarded it as a poetic term and thus felt the need to gloss it: ἄλγος· ἡ λύπη (α 2803 Latte/Cunningham). The codex of the Glossai kata poleis (Bowra 1960), which likely depends on Hesychius, included ἄλγος amongst the Cypriot words, since it was not uncommon in Antiquity to ascribe poetic words, especially those of Classical drama, to the Cypriot dialect.4 The lexeme, which in Homer mainly occurs in the plural, indicates above all a pain suffered over time, as in the very beginning of the Iliad, where Achilles’ rage is tied to the innumerable and lasting sufferings incurred by the Achaeans (Il. 1.2 μυρί’ Ἀχαιοῖς ἄλγε’ ἔθηκε). A long-lasting pain is also evoked in the first lines of the Odyssey, which consecrate Odysseus as a hero who suffered much pain in his heart on the sea (Od. 1.4 πολλὰ δ᾽ ὅ γ᾽ ἐν πόντῳ πάθεν ἄλγεα ὃν κατὰ θυμόν).5 Obviously, other words for ‘pain’ are attested in archaic texts, each of

 2 One of the etymologies proposed links it to the verb ἀλέγω, ‘worry’ or ‘take care’: such a link would explain why the adjective ἀλεγεινός can serve as an alternative to the more common ἀλγεινός, ‘painful’. In any case, according to Chantraine (DELG s.v.) and Seiler (1950, 85), proposing a relationship between ἀλέγω and ἄλγος leads to semantic problems, except in the case of a kind of euphemism that accords with the sequence: ‘take into account’ > ‘take care’ > ‘suffer’. Seiler, in fact, prefers a comparison to the Latin (algus ‘bitter cold’, ‘nip’) See also Beekes (EDG s.v.). 3 It suffices to refer to words ending in -algy. 4 Because of its peripheral position, Cyprus tended to preserve a lot of archaic cultural and linguistic features, and the Cypriot dialect has been one of the most conservative Greek dialects since Antiquity. Several epic and lyric terms were still in use in Cyprus in the 5th century BC and later on, not only in poetry: “[l]’ensemble des gloses chypriotes ne présente donc pas des emprunts à la langue épique, mais un héritage commun” (Egetmeyer 2010, 28). On the many lexical archaisms of the Cypriot dialect, see also Bowra 1934. 5 In his lectures, Domenico Musti often emphatically repeated that Greek literature began with rage, Achilles’ μῆνις, and that it would not be far-fetched to state that it also began with pain, the ἄλγεα of the Achaeans and Odysseus. In reality, in Homeric poetry there was not yet a clear

Prose and Poetry of Pain: A History of the Term ἄλγος  221

them with a slightly different semantic nuance: as opposed to ἄλγος, ὀδύνη refers to briefer suffering (as is clear e.g. from Solon fr. 12.59 West, ἐξ ὀλίγης ὀδύνης μέγα γίνεται ἄλγος), whereas πῆμα ‘calamity’ is connected to the domain of suffering on account of its causative value (a calamity brings about pain).6 Within lyric poetry, too, ἄλγος continues to be used quite frequently and preponderates over its synonyms. In Fatouros’ Index verborum zur frühgriechischen Lyrik, approximately fifteen attestations are observed, more than those of λύπη (once in Corinna, fr. 654 PMG), ὀδύνη (6x), πένθος (8x), or πάθος (6x) — the latter formed on the zero grade already common in the many adjectival compounds ending in -παθής.7 Starting with Pindar, it seems that ἄλγος begins to retreat. Slater’s lexicon attests ἄλγος only once in Pindaric poetry (fr. 210 S.-M. ἱστᾶσιν ἄλγος ἐμφανές [scil. οἱ ἄγαν ἐν πόλεσι φιλοτιμώμενοι]; v.l. ἢ στάσιν, ἄλγος). Instead, we find a

 distinction between physical and psychological pain. As stated by Roseline Rey in her book Histoire de la douleur, “les modes de représentation de la douleur s’articulent non autour de l’opposition du physique et du moral, non autour du degré de la douleur, mais suivant deux axes: le degré d’implication du sujet dans la douleur et les modalités de perception de celle-ci en fonction de la temporalité et en fonction de la source de la douleur — durable ou rapide, aigu ou tranchant — c’est-à-dire en faisant référence directement à l’instrument qui en est la cause, et qui définit du même coup les qualités de la sensation” (Rey 2011, 18−19). On the Homeric use of ἄλγος, see also Rijksbaron 1992. 6 The French scholar sees in πῆμα “la désignation d’une chose désagréable ou nuisible, en tant qu’origine, agent ou porteur du procès, et non comme produit d’un procès, comme phénomène extérieur au sujet. Cette valeur (‘fléau, malheur, cause ou sujet de douleur’) s’accorde d’ailleurs avec la signification des différents dérivés de πῆμα: exemple: ἀπήμων ‘qui ne fait pas de mal’, πημαίνω ‘faire du tort à, léser, endommager’” (Mawet 1979, 387). As for ὀδύνη, it is instead a term originally belonging to medical vocabulary, which then became a common word for indicating an acute, violent pain: “[e]n tant que désignation d’une douleur lancinante et localisée, ὀδύνη s’oppose à ἄλγος, expression d’une douleur plus durable, plus généralisée et qui affecte la totalité du corps” (Mawet 1979, 390). The Homeric lexicon of pain also includes ἄχος ‘emotive shock’, πένθος ‘mourning’, and κῆδος ‘worry’, ‘restlessness’, which is intellectually charged and projects towards the future (according to Mawet, this concept is similar to the French souci). 7 The term λύπη is not attested in Homer and seems to prevail especially in the 5th century BC with a moral meaning (‘pain’, ‘suffering’) developed from one that was originally concrete: “[l]es emplois à propos de maladies ou de mauvaises terres supposent à l’origine un sens concret” (DELG s.v.). In reality, the adjective λυπρός is already a Homeric hapax referring to the island of Ithaca (Od. 13.243), just like its cognate λυπέω found in Hesiod (Op. 401). Therefore, it is only in tragedy, prose, and in later poetry that the more modern λύπη, after some attestations in lyric poetry, acquired the vigour that made it “le terme central de la douleur en grec post-homérique” (Mawet 1979, 399–400). The neuter πάθος, destined to remain in use across the centuries owing to its general meaning, is also recent, since it is not attested before the 5th century BC.

222  Enrico Cerroni decidedly greater frequency of πόνος (35x) and πῆμα (6x), as well as the distinctive presence of ὀδύνα in Pyth. 4.221 and an absence of λύπη and πάθος, to which Pindar prefers the even newer and pretentious πάθα (6x).8 Aside from choices motivated by metrical concerns, it is plausible that the reason for the prevalence of πόνος in an epicizing lyric lexicon, elsewhere filled with ἄλγος and ἄλγεα, rests on Pindar’s heroic ethics, which associate pain with the struggle that one undertakes to obtain a prize.9 A case in point is Pyth. 12.18, which recalls the labours of Perseus in his struggle with Medusa, labours that in the logic of things incur their own pain. In short, for Pindar it is probably a matter of parole rather than langue.10

 The 5th and 4th Centuries BC The literary genre in which the use of ἄλγος remains most constant is tragedy, although with some discrepancies. Aeschylus, for instance, who uses it extensively (27x),11 nonetheless chooses the more modern πάθος (which conveys a sense of unexpected and enduring suffering) to signify the speculative tension

 8 The latter (πάθη in Attic-Ionic) represents an exceptional term in the Hippocratic lexicon. This feminine form also occurs in Herodotus (5x), Sophocles (3x) and Plato (7x). 9 The semantic shift that led πόνος ‘labour’ to acquire the meaning of ‘pain’ in post-Homeric literature is the opposite of that of Latin laborō (‘suffer, labour’), which evolved into the Italian lavorare: “[l]’acception ‘souffrance’ résultant d’un développement postérieur des emplois du terme, n’apparait que de façon tout à fait occasionnelle dans les poèmes homériques, à côté de la signification propre: ‘travail accompli avec effort’” (Mawet 1979, 393). An interesting moment in this history can be identified in the poetry of Pindar. In fact, reading some sections of the odes gives us useful indications of the semantic value to be assigned to πόνος, not simply ‘labour’, but also ‘pain’. For example, Ol. 2, which deals with the inescapable ups and downs of human existence, offers precious evidence for this progression. Already in the first narrative section (vv. 19−20), the poet states ‘for the malignancy of woe is quelled and perisheth beneath the joy of goodly triumphs, / when the destiny of God sendeth a man soaring happiness from on high’ (transl. L.R. Farnell); the word used for ‘pain’ is πῆμα (πῆμα θνᾴσκει παλίγκοτον δαμασθέν), but when Pindar later talks about the alternating moments between joy and sadness, he reverts to πόνος: cf. vv. 33−34: ῥοαὶ δ’ ἄλλοτ’ ἄλλαι / εὐθυμιᾶν τε μέτα καὶ / πόνων ἐς ἄνδρας ἔβαν ‘at various times varying tides of fortune come upon men bearing both joy of heart and sorrows’ (transl. L.R. Farnell). 10 “Πόνος und μάθησις ja, πόνος und ἀρετή sind untrennbar; […] der Vorausschauende nimmt freiwillig πόνος auf sich; der Dumme hat seine Versäumnis durch πάθος abzugelten” (Dörrie 1956, 319). 11 Aeschylus also uses another cognate: the verb ἀλγύνω ‘cause pain’ (3x: Sept. 358, Cho. 746, PV 245), which is quite common in Sophocles and Euripides (see for instance Hipp. 798).

Prose and Poetry of Pain: A History of the Term ἄλγος  223

that allows man’s conscience to develop, as evidenced by the well-known maxim πάθει μάθος (Ag. 177). This choice could be explained by the fact that ἄλγος was too epic and archaic in comparison to the newer πάθος, or that the latter stresses the passivity of man in the face of destiny. The Spanish philologist Marcos Martínez Hernández’s 1977 survey of Sophocles’ works reveals that in this poet ἄλγος tends to signify physical pain, moral pain, and pain in general. The most interesting tragedy from this point of view is Philoctetes, not only because it addresses the topic of suffering, but also because it represents a key moment in linguistic history, i.e. the end of the 5th century BC.12 Thus, to fully convey his pain, with all the consumption and fatigue it entails, Philoctetes wishes Odysseus could experience it in the same way that he has felt it: (1)

ὦ ξένε Κεφαλλήν, εἴθε σοῦ διαμπερὲς Στέρνων ἔχοιτ’ ἄλγησις ἥδε.

(Soph. Phil. 791−792)

O my friend from Cephallenia, if only this pain went through your breast and held you in its grip. (transl. R.G. Ussher)

As Anthony Arthur Long notes, Philoctetes’ pain is physical, and after Sophocles the word ἄλγησις was used only in the context of tragic parody by the poet Agathon in Aristophanes’ Thesmophoriazusae (vv. 146−147).13 The introduction of ἄλγησις is therefore also proof of Sophocles’ preference for rare and elevated forms ending in ‑σις in his later works and of his desire to circumvent the traditional lexicon in view of what Long defined as an “affected style”: line 792 could have had equal scansion had Sophocles used ἄλγημα τοῦτο rather than ἄλγος, so it is significant that he chose to use the newer form.14 The contemporary cognate forms at the end of the 5th century BC were principally ἀλγηδών and ἄλγημα. It is no accident that the tragic poet Sophocles uses ἀλγηδών on only one occasion (OC 513−514). Chantraine (1933, 361) describes this

 12 “It is not surprising that the Philoctetes contains a large number of words for pain and disease, as well as some professional medical terms. Thus νόσος ‘physical sickness’ occurs 18 times, ἄλγησις and ἄλγημα (340, 1169) do not occur in the other plays, and we find such technical terms as νοσηλεία (18), θερμοτάταν αἱμάδα (696), κουφίζειν (735), αἱμορραγὴς φλέψ (825)” (Long 1968, 132). 13 Long 1968, 132. 14 “The difference is a fine one, but important; ἄλγησις gives us a subjective statement of feeling. Philoctetes wants Odysseus to experience the full force of his own suffering” (Long 1968, 133). See also Schein 2013, 242–243.

224  Enrico Cerroni word, a secondary formation created on the root of the verb ἀλγέω with the application of the suffix -don-, as Ionic, poetic, and Platonic, so it was probably too new and inelegant to Sophocles’ ears. But what of ἄλγημα, which was also becoming increasingly prevalent in the Hippocratic corpus? In Sophocles’ usage, ἄλγημα must refer to the physical pain Philoctetes experienced because of his wound (Phil. 1170−1172). Similarly, in 339−340, Neoptolemus makes reference to Philoctetes and his ἀλγήματα: ‘thou hast enough of thine own pains, poor soul, without lamenting for another’s woe’ (transl. F. Storr). Conversely, Euripides’ works contain a decidedly high frequency of ἄλγος (with more than forty attestations) and almost no trace of the newly created ἄλγημα, attested in only one fragment. The technical term ἀλγηδών, according to Marzullo (1999, 124), “invades” Medea (ll. 24, 56, 1031). In Euripides’ lexicon the semantic field of pain is also represented by λύπη (with about forty examples), on occasion by πόνος15 and, less often, by ὀδύνη (12x). In comparison with poetry, there is a drastic decline of ἄλγος in the usus scribendi of the prose writers of the 5th century BC. Herodotus provides only a single occurrence (5.49.2), to which no particular attention has been paid in the commentaries. Heinrich Stein (1894, 44) focuses on the χάλκεον πίνακα that Aristagoras, the tyrant of Miletus, brought with him, but not on the epic emphasis of his words. In this speech, he is seeking the help of Cleomenes, the king of Sparta, on the occasion of the Greek cities’ revolt in Asia Minor: (2)

Κλεόμενες, σπουδὴν μὲν τὴν ἐμὴν μὴ θωμάσῃς τῆς ἐνθαῦτα ἀπίξιος· τὰ γὰρ κατήκοντα ἐστὶ τοιαῦτα· Ἰώνων παῖδας δούλους εἶναι ἀντ᾽ ἐλευθέρων ὄνειδος καὶ ἄλγος μέγιστον μὲν αὐτοῖσι ἡμῖν, ἔτι δὲ τῶν λοιπῶν ὑμῖν, ὅσῳ προέστατε τῆς Ἑλλάδος. νῦν ὦν πρὸς θεῶν τῶν Ἑλληνίων ῥύσασθε Ἴωνας ἐκ δουλοσύνης ἄνδρας ὁμαίμονας. εὐπετέως δὲ ὑμῖν ταῦτα οἷά τε χωρέειν ἐστί· οὔτε γὰρ οἱ βάρβαροι ἄλκιμοι εἰσί, ὑμεῖς τε τὰ ἐς τὸν πόλεμον ἐς τὰ μέγιστα ἀνήκετε ἀρετῆς πέρι. (Hdt. 5.49.2–3) Wonder not, Cleomenes, that I have been so zealous to come hither; for such is our present state: that the sons of the Ionians should be slaves and not free men is a shame and grief to

 15 Since πόνος means both ‘toil’ and ‘physical suffering’, it is difficult to state with certainty its precise value in every Euripidean passage (113x). As explained by J.P. Vernant, “πόνος s’applique à toutes les activités qui exigent un effort pénible”. Heracles’ life is the very example of this heroic concept: “dans le mythe d’Héraclès, le héros doit choisir entre une vie de plaisir et de mollesse et une vie vouée au ponos. Héraclès n’est pas un travailleur” (Vernant 1971, 17). On the meaning of πόνος and the concept of labour in ancient Greece, see also Loraux 1982 and Musti 1997, 41–42.

Prose and Poetry of Pain: A History of the Term ἄλγος  225

ourselves in especial, and of all others to you, inasmuch as you are the leaders of Hellas. (transl. A.D. Godley)

The commentators (Macan 1895, How/Wells 1928, van Groningen 1950, Nenci 2006, and more recently Hornblower 2013), more concerned with the interpretation of Aristagoras’ map, note no peculiarity in this lexical choice. Herodotus, an Ionian speaker, who a short time later in the same passage employs the form δουλοσύνη, “che aveva avuto particolare fortuna in area ionica, dove la schiavitù sotto lo straniero era stata vissuta da generazioni intere e poteva ben essere evocata con la sua drammaticità epica” (Nenci 2006, 224), uses the Homeric term ἄλγος only once together with another Homeric word, ὄνειδος ‘blame, reproach’. The question may not warrant too much attention because Herodotus has a tendency to use poetic forms in his prose;16 however, the passage under discussion is not one of the more mundane. Herodotus is artfully constructing the dialogue he imagines the Ionian Aristagoras delivering to the Spartan Cleomenes, a tyrant from Ionia confronting a Spartan leader of great importance in a situation that required conveying a particular pathos. To his eyes and to those of the Greeks from Asia, it was disgraceful and painful that the sons of Ionia were slaves instead of free men. The stylistic intensification is intentional, and the epic patina does not derive solely from ἄλγος, but also from δουλοσύνη and from the expression with which Aristagoras decrees the superiority of the Spartans (ἐς τὰ μέγιστα ἀνήκετε ἀρετῆς πέρι), as Nenci (2006, 225) indicates in the comment on the passage.17 After Herodotus, Thucydides — who carried out a thorough investigation of the symptoms of the plague in Athens — never uses ἄλγος or ἄλγημα, preferring πόνος (Marzullo 1999, 126), or more abstract terms describing the symptoms specific to the disease. The only exceptions are the adjective ἀλγεινός (2x) and the verb ἀλγέω (2x). In this way, he banishes a word with a clear Ionian stamp from

 16 The use of poeticisms by Herodotus, not by chance defined as ὁμηρικώτατος by the ancients, has been widely studied, including in connection with the various contexts. For example, Carla Schick has revealed a greater frequency of poetic terms in the parts with dialogue than in those with narrative. The recourse to the model of poetry by logographers is greater “quando lo stesso interesse per il fatto narrato induce gli autori, consapevoli o no, a ricercare forme più intensamente espressive” (Schick 1956, 390). 17 After all, the language of pain used by Herodotus also tends to have a refined style elsewhere, as pointed out by Marzullo 1999, 126. The sole case of the synonym ὀδύνη (9.16.5) is in the gnomic words of a certain Tersander: ἐχθίστη δὲ ὀδύνη ἐστὶ τῶν ἐν ἀνθρώποισι αὕτη, πολλὰ φρονέοντα μηδενὸς κρατέειν ‘and it is the hatefulest of all human sorrows to have much knowledge and no power’ (transl. A.D. Godley).

226  Enrico Cerroni Attic prose. After all, Thucydides censored much of the lexicon of pain,18 also rejecting words such as ἄχος ‘pain, anguish’ (Homeric, lyrical, and tragic, found once in Herodotus), and the more banal ὀδύνη. It is clear, however, that Ionic literary culture made a fundamental contribution to the renewal of the lexicon. Between the 5th and 4th centuries BC, the Hippocratic corpus already shows the substitution of the old ἄλγος with the more precise and functional ἄλγημα, with 14 attestations of the former compared to 194 of the latter. However, we can see that nouns ending in -μα were generally preferred by comparing the instances of νοῦσος (126x), νόσημα (487x) and the very new νόσευμα (12x). The preserved forms of ἄλγος only appear in Epidemics and Internal affections (a text considered amongst the oldest in the corpus) as part of technical set phrases such as καρδίας ἄλγος ‘heart ache’ (Epid. 7.20), περὶ τὴν κύστιν ἄλγος ‘bladder pain’ (Epid. 5.43), ἄλγος κοιλίης δεινόν ‘terrible stomach pain’ (Epid. 5.98, 7.29), ἄλγος ἐς γούνατα καὶ μηρόν ‘pain in the knees and the side’ (Epid. 7.54), and καὶ πρὸς καρδίην ἄλγος δεινόν ‘terrible heart pain’ (Epid. 7.62). What is striking is the relative speed with which the Hippocratic lexicon introduced the more precise ἄλγημα as a replacement for ἄλγος, evidently considered too old and tied to the Homeric world to serve the systematic methodology of a modern medical text. Benedetto Marzullo saw in ἄλγημα “una rideterminazione di origine filosofica, implicante riflessiva soggettività” (1999, 124). The equally old term ὀδύνη, instead, fared better, with 772 examples compared to a mere five examples of the neologism ὀδύνημα. The epigraphic documentation also bears witness to the sporadic preservation of the now fossilized forms, an illuminating example of which comes from IG IV2,1 122, chronicling Asclepius’ Iamata. One of the clinical cases concerns a certain Agestratus, an athlete who suffered from κεφαλᾶ[ς ἄ]λγος and πόνος, which forced him into periods of insomnia διὰ τὸμ πόνον τᾶς κεφαλᾶ[ς]. At this point, the competition is between ἄλγος and πόνος to determine which would become the predominant lexeme in the Greek language of the Hellenistic period.

 18 According to the French scholar Pierre Huart, this could be due to Thucydides’ attentiveness to events and processes that could have a general interest, beyond the individual sentiments: “[v]oilà donc le but de l’étude psychologique pour Thucydide: passer sous silence tout ce qui est particulier, pour s’attacher à ce qui, dans les personnages qu’il met en scène, est susceptible de présenter un intérêt général pour la connaissance de l’homme et des sentiments fondamentaux qui mènent l’humanité” (Huart 1968, 5).

Prose and Poetry of Pain: A History of the Term ἄλγος  227

The review of Classical prose would be incomplete without Xenophon. He follows the general tendency, preserving only two attestations of ἄλγος in medical contexts, in a passage about dogs with weak paws in Cynegeticus 3 (οὐδ’ ἂν ὦσιν εὔψυχοι τοὺς πόνους δύνανται ἀνέχεσθαι ἀλλ’ ἀπαγορεύουσι διὰ τὸ ἄλγος τῶν ποδῶν ‘even if they are plucky, can’t stand the hard work, and tire because they are foot-sore’, transl. E.C. Marchant) and in Symposium 8.37, where he clearly echoes an epic iunctura (πολλὰ δ’ ἄλγη ἀνέχεται). The infrequency of the term in prose continues in the higher registers, which in the 5th and 4th centuries BC are no longer only tragic.19 A significant example comes from Aristophanes’ Plutus, a comedy from 388 BC in which ἄλγος is attested only once. The passage in which it occurs deserves further consideration. An old woman entertains herself with Chremylus, to whom she relates her distress at having lost the sexual favours of a young man who has become rich thanks to Plutus and so no longer needs to be paid by a very old woman. (3) Γρ. ὑπὸ τοῦ γὰρ ἄλγους κατατέτηκ’, ὦ φίλτατε. Κρ. Οὔκ, ἀλλὰ κατασέσηπας, ὥς γ’ ἐμοὶ δοκεῖς.

(Ar. Plut. 1034−1035)

OLD WOMAN: In fact I’m pining away with grief, my dear man. CHREMYLUS: No, you’re rotting away, if you ask me. (transl. J. Henderson)

A commentary on the passage (Torchio 1999, 224) ignores this remarkable lexical choice, which could probably be interpreted as a hint at a tragic style, considering that otherwise ἄλγος had no right to belong in comedy. Some useful insights on the relevance of this choice can be found in the 1970 doctoral thesis by Gretchen Cheney Southard.20 This is comparable to how one might raise the style in a comedic situation in Italian by using a word such as duolo, a highly poetic

 19 According to Marzullo 1999, 126, “l’aristocratica supremazia di ἄλγος è destinata a crollare: l’imbarazzante poetismo sopravvive, quale pretenzioso flosculo letterario”. 20 “In this case the pain is obviously mental, but the use of κατατήκω may suggest a play on physical pain since τήκω and its compounds, though primarily non-medical words, are often used by medical writers to describe the dissolving of flesh or by non-medical writers of someone being consumed by disease or love” (Southard 1970, 23). Nevertheless, Southard fails to point out that we are dealing with a hapax in Aristophanes’ surviving works, and inserts ἄλγος into a chapter on medical vocabulary in Aristophanes, which is constructed through the search for correspondences with the vocabulary used by Hippocrates. In light of the argument I am making, it would probably be more correct to place it among tragic vocabulary.

228  Enrico Cerroni word, rather than a more neutral contemporary term such as dolore.21 It is interesting to note the polarity achieved by the line of the old woman and that of Chremylus. The former complains of being consumed by pain, as though she were a tragic heroine (albeit in a much lighter situation), while the latter adds to the absurdity by contrasting her κατατέτηκα with the much more realistic κατασέσηπα. It would be enlightening if we could identify a close parallel to the old woman’s line in a tragedy by Sophocles or Euripides. Unfortunately, precise parallels with the verb κατατήκω in the perfect tense are lacking: however, useful indications can be found in the commentary by van Leeuwen (1968, 153), who quotes Electra’s lament in Sophocles (El. 187 ἄνευ τοκέων κατατάκομαι) and Orestes’ words in Euripides’ play (El. 239 λύπαις γε συντετεκώς). In addition, van Leeuwen adds a comic verse from Eubulus (fr. 102 K.-A., on κισσός ‘ivy’ which is ὀλολυγόνος ἔρωτι κατατετηκώς),22 while Philocleon in the Wasps gushes out φίλοι, τήκομαι (Vesp. 317). Van Leeuwen does not highlight the uniqueness of the poetic ἄλγος, now rare even in prose, in the comic context; his detailed comments on κατατέτηκα and κατασέσηπας (“parum sane urbane dictum, non tamen omnino abhorrens a lingua quotidiana”) only serve to underline the paratragic effect of the line. The 4th century BC offers us more uniform evidence, now largely confirming the poetic and archaic status of ἄλγος. Ast’s Lexikon Platonicum and the TLG mark the only occurrences of the term in Alcibiades secundus (142e.1), where it appears in a quotation from Homer.23 Plato prefers ἀλγηδών (‘souffrance’: Chantraine 1933, 361), which is documented 28 times. He does not adopt the Hippocratic ἄλγημα, which is entirely absent from his work. Even the three cases of ἄλγος in Aristotle registered by Radice’s lexicon reveal the same evolution: two of them are poetic quotations.24 Similarly, Isocrates uses

 21 Maria Cristina Torchio translates the word with ‘dolore’ (pain) (‘VE. Mi consumo per il dolore, carissimo. / CR. No, ma sei già in putrefazione, per quanto mi sembra’). 22 This fragment has been interpreted in different ways. According to Hunter 1983, 196 “Meineke suggested that these lines refer ‘ad incognitam nobis fabulam […] de Cisso (Nonnus, Dion. 12.97, Paus. 1.31.6) Ololygonis nymphae amore tabescente’ and this idea was picked up in an influential note by Rhode”. White 1979, 9–16 has suggested that the nightingale was meant by the word ὀλολυγών (also in Theoc. 7.139). 23 The quotation is not at all banal, since it comes from a moralizing section of the Odyssey, in which Zeus laments the ingratitude of men (ὑπὲρ μόρον ἄλγε’ ἔχουσι) and seems to pave the way for a kind of archaic theodicy. 24 As regards Rhetoric 1370b.5, μετὰ γὰρ τε καὶ ἄλγεσι τέρπεται ἀνὴρ μνήμενος ὅς τις πολλὰ πάθῃ καὶ πολλὰ ἐόργῃ, cf. Od. 15.400–401. In the Athenian Constitution 5.2, instead, Aristotle quotes Solon fr. 4a West. It is interesting to observe still in Aristotle (Hist. an. 635a.28) a medical

Prose and Poetry of Pain: A History of the Term ἄλγος  229

ἀλγηδών and never ἄλγος. In Menander, the verb ἀλγέω and the nouns ἄλγος and ἄλγημα survive with one or two attestations and, unsurprisingly, only ἄλγος is attested in the γνῶμαι μονόστιχοι (446 Jäkel: λιμὸς μέγιστον ἄλγος ἀνθρώποις ἔφυ). In New Comedy, the sphere of pain belongs to λυπέω and λύπη (9x),25 which are more frequent than the words connected to the family of ὀδύνη, a development that is also evident in philosophy.26

 Evolution in the Hellenistic Age The Hellenistic philosophers were preoccupied with the concept of pain and the methods of limiting its impact on the life of man. Therefore, it is particularly important to investigate the different ways of labelling, conceiving of and interpreting human suffering from a lexical perspective in this period. Epicurus’ aim was liberation from pain, which he, significantly, termed ἀπονία.27 It should be stressed that the first meaning of this abstract term was ‘lack of labour’, as documented in Aristotle (Gen. an. 775a.37). That the notion of ‘labour’ was still implicit in πόνος in the 4th century BC is evident in Xenophon’s Cyropaedia (2.2.25), where ἀπονία is used to indicate indolence. In philosophical contexts, πόνος resurfaces in the Cynics and Stoics, who impose on it a strong ethical meaning, as often happened in the Hellenistic philosophies.28  expression such as ἄνευ ἄλγους καὶ μετὰ ἀναισθησίας, which is likely to have sounded quite antiquated. 25 As is declared in a fragment of the Kitharistes (fr. 1.8 Körte): ἆρ’ ἐστὶ συγγενές τι λύπη καὶ βίος; ‘can pain and life be brothers?’ (transl. W.G. Arnott). 26 Although with a significant exception, in the ‘fanciful’ line ὀδύνης γὰρ ὑός in Dys. 88 (Gomme/Sandbach 1973, 148−149). 27 In the fragment of the treatise Περὶ αἱρέσεων καὶ φυγῶν, recorded by Diogenes Laertius in the Life of Epicurus (10.136), we read: ἡ μὲν γὰρ ἀταραξία καὶ ἡ ἀπονία καταστηματικαὶ εἰσιν ἡδοναί· ἡ δὲ χαρὰ καὶ εὐφροσύνη κατὰ κίνησιν ἐνεργείᾳ βλέπονται (fr. 7 Arrighetti). 28 It is worth pointing out the use of πόνος by Onesicritus of Astypalaea, a cynical thinker who took part in Alexander the Great’s expedition. In a passage handed down by Strabo (15.1.65 = FGrHist 134 F 17), the stimulating character of πόνος (φίλιον) is highlighted compared to the passive and negative λύπη (πολέμιον). Cf. Prost 2004, 40: “[l]a douleur en général apparaît donc comme un ennemi de l’homme, mais le même douleur, exploitée dans le cadre volontaire et réfléchi de l’ascèse, se voit doter d’une valeur positive: en ce sens, Antisthène pouvait dire que ‘la douleur est un bien’. Cette ambivalence de la douleur, qui anticipe le principe de l’‘indifférent’ stoïcienne, est également soulignée par Onésicrite, lorsqu’il oppose le caractère stimulant du πόνος à celui totalement négatif du chagrin (λύπη), qualifiant le premier de φίλιος, le second de πολέμιον”.

230  Enrico Cerroni But what can we say of the privatives formed from the stem of ἄλγος? The form ἀναλγησία ‘insensibility’ already existed in Democritus (B 193 D.-K.) and the verbal adjective ἀνάλγητος ‘insensible to pain’ or even ‘not painful’, if referring to things, is already attested in the 5th century BC (Soph. Trach. 126), whereas ἀναλγής ‘painless’ (for example, referring to death) is apparently documented in the Hippocratic corpus for the first time. In contrast to the semantic range of ἀπονία, the privative compounds ἀναλγησία and ἀνάλγητος focus more on the aspect of insensibility to physical or metaphorical pain. In both Sophocles’ Ajax (946) and Thucydides (3.40.5), the sense is ‘hard of heart’: effectively, insensibility and ruthlessness go hand in hand.29 Nevertheless, neither the abstract ἀναλγησία nor the adjectives ἀνάλγητος and ἀναλγής met with the same success as ἀπονία and ἀλυπία.30 Epicurean philosophy is worth studying because it was especially concerned with the search for pleasure and liberation from pain. A survey of Usener’s Glossarium Epicureum (1977) reveals the continued use of the verb ἀλγέω and the prevalence of λύπη and πόνος, which have developed specialized meanings. In contrast to πόνος, which is tied to the experience of trial and labour, λύπη, the most ancient and thus most appreciated by the tragedians, retained connotations of psychological negativity, approaching that of the French chagrin, and it is no accident that Hellenistic philosophy tended to afford it a negative meaning. The Stoic Cleanthes defined it as a relaxation of the soul (ἔλεγε τὴν λύπην ψυχῆς παράλυσιν, fr. 1.575 von Arnim). Andronicus of Rhodes in the Περὶ παθῶν (1.1.10 Kreuttner), a work now attributed to an eclectic author of the Imperial Age, presented pain (λύπη) as an irrational depression (ἄλογος συστολή).31  29 In the Thucydidean passage, for instance, Cleon exhorts the Athenians not to seem ἀναλγητότεροι (‘less sensitive’) than the inhabitants of Mytilene, who had revolted against the Athenian empire, and to punish them: “Here, on Kleon’s lips, it means ‘insensible to our town’” (Gomme 1956, 312). 30 The semantic field of ἄλυπος and ἀλυπία is rather the lack of pain (see Soph. OT 593, OC 1519: γήρως ἄλυπα are the things that are protected from the ravages of time). The notion of ἀλυπία is probably older than ἀπονία. The famous inventor of the mysterious τέχνη ἀλυπίας, a sort of pioneering psychology, was Antiphon (A 6 D.-K.), whom Mario Untersteiner also considers the creator of the same word (Prost 2004, 76). As for the stem of ἄλγος, it is important to note that before philosophy, Greek medicine itself had snubbed ἀναλγησία and ἀνάλγητος (3x) and that their recovery only occurred at the beginning of the 19th century: “[l]a nostra ‘analgesia’ è dotto recupero, risale agli inizi del secolo scorso, molti lessici ancora chiosano con ‘mancanza di sensibilità’” (Marzullo 1999, 124−125). 31 Thus, when we arrive at the famous passage of Cicero’s Tusculanae disputationes (2.35), to a Roman — even one as cultured and partisan as Cicero — it seemed that Greek was strangely lacking in vocabulary, seeing as it uses the same term, πόνος, for labour and pain. The problem laid

Prose and Poetry of Pain: A History of the Term ἄλγος  231

Beyond the boundaries of philosophy, the remaining literary genres of the period deserve further remarks. While Hellenistic poetry continued to employ the Homeric concept of ἄλγος,32 as a result of what might be described as the inertia of tradition, the occurrences of this word in prose are marked by a stylistic refinement when ἄλγος is not part of a poetic quotation. Such is the case in the surviving works of Polybius: ἄλγος is attested only once, precisely in a quotation (the famous incipit of the Odyssey, in 12.27.11). It is rather more interesting to analyze his behaviour when paraphrasing Homer. For example, in 9.21 the historian quotes the famous verse of the Odyssey on Eurycleia’s reaction upon recognizing Odysseus.33 Polybius replaces the Homeric ἄλγος with the more prosaic λύπη (κατὰ τὸν ποιητὴν ἅμα λύπην καὶ χαρὰν ὑποτρέχειν εἰκὸς ἦν τὰς ἑκάστων ψυχάς).34 According to John Thornton (2014), the reference to the passage from the Odyssey is inaccurate, as Polybius’ quotations often are. In any case, the inaccuracy of the quotation is valuable because it demonstrates how much a noun such as ἄλγος clashes with the historiographic prose of an author such as Polybius, which is dry and free of floweriness. (4) ὅτι τοιαύτης διαθέσεως ὑπαρχούσης περί τε τοὺς Ῥωμαίους καὶ Καρχηδονίους, καὶ παλιντρόπων ἑκατέροις ἐκ τῶν ὑπὸ τῆς τύχης ἀπαντωμένων ἐναλλὰξ προσπιπτόντων, κατὰ τὸν ποιητὴν ἅμα λύπην καὶ χαρὰν ὑποτρέχειν εἰκὸς ἦν τὰς ἑκάστων ψυχάς. (Polyb. 9.21.1) Such being the respective positions of the Romans and Carthaginians, experiencing in turn the opposite extremes of fortune, it was natural that, as Homer says, pain and joy at once should possess the minds of each. (transl. W.R. Paton)

Through a survey of Mauersberger’s valuable lexicon of Polybius (PL), we have confirmation that even in his lucid and dispassionate prose, it is the ‘vulgar’ and  out by Cicero, who splits hairs on the meaning of the Greek term πόνος indicating both physical labour and the painful sensation it brings about, is well analyzed by Prost 2016, 14−15. In reality, the Greek lexicon for pain is richer than what the Latin writer suggests, and is historically and stylistically differentiated. In the case of the Greek φιλόπονος, taken as ‘lover of pain’ (studiosos vel potius amantis doloris) instead of ‘lover of labour’, it is not a problem if considered diachronically. 32 Found in Apollonius of Rhodes (11x), Nicander (11x, Theriaca and Alexipharmaca), Lycophron (2x). 33 Od. 19.471−472 τὴν δ’ ἅμα χάρμα καὶ ἄλγος ἕλε φρένα τὼ δέ οἱ ὄσσε / δακρυόφι πλῆσθεν, θαλερὴ δέ οἱ ἔσχετο φωνή ‘then upon her soul came joy and grief in one moment, and both her eyes were filled with tears and the flow of her voice was checked’ (transl. A.T. Murray). 34 The passage was noted and opportunely commented on in stylistic terms by De Foucault 1972, 244.

232  Enrico Cerroni neutral πόνος (12x) that dominates, rather than λύπη (2x) and ὀδύνη (totally absent),35 leading the way for Modern Greek, in which the use of πόνος has been generalized to indicate pain or discomfort. At approximately the same time, in the 2nd century BC, Jewish-Hellenistic literature shows its own trends. A survey conducted on the concordance to the Septuagint (H&R s.v.) demonstrates the rather measured use of ἄλγος in the following passages: (5) γενηθήτω ἡ ἔπαυλις αὐτῶν ἠρημωμένη, καὶ ἐν τοῖς σκηνώμασιν αὐτῶν μὴ ἔστω ὁ κατοικῶν· ὅτι ὃν σὺ ἐπάταξας, αὐτοὶ κατεδίωξαν, καὶ ἐπὶ τὸ ἄλγος τῶν τραυματίων σου προσέθηκαν.

(Psalm 68 [69].26–27)

Let their steading become desolated and let there be no one who lives in their coverts, because they persecuted him whom you struck and to the pain of your wounded they added. (transl. A. Pietersma, NETS) (6) ἄλγος καρδίας καὶ πένθος γυνὴ ἀντίζηλος ἐπὶ γυναικὶ καὶ μάστιξ γλώσσης πᾶσιν ἐπικοινωνοῦσα. (Sirach 26.6) Pain of heart and sorrow is a woman who is a rival to a wife, and a lash of a tongue is she who shares with everyone. (transl. B.G. Wright, NETS) (7)

οἱ πρὸς ὑμᾶς πάντες παραπορευόμενοι ὁδὸν, ἐπιστρέψατε καὶ ἴδετε εἰ ἔστιν ἄλγος κατὰ τὸ ἄλγος μου, ὃ ἐγενήθη. (Lamentations 1.12) It is nothing to you, all you who pass along the road! Turn, and see if there is a sorrow like my sorrow which has happened. (transl. P.J. Gentry, NETS)

(8) ἀκούσατε δή, πάντες οἱ λαοί, καὶ ἴδετε τὸ ἄλγος μου.

(Lamentations 1.18)

Hear then, all you peoples, and behold my sorrow. (transl. P.J. Gentry, NETS) (9) περιεκέχυτο γὰρ περὶ τὸν ἄνδρα δέος τι καὶ φρικασμὸς σώματος, δι’ ὧν πρόδηλον ἐγίνετο τοῖς θεωροῦσιν τὸ κατὰ καρδίαν ἐνεστὸς ἄλγος. (2 Maccabees 3.17)

 35 Instead, the relationship is inverted if we compare the cognate verbs λυπέω (16x) and πονέω (7x).

Prose and Poetry of Pain: A History of the Term ἄλγος  233

For a frightening bodily shudder was spread over the man by which the pain lodged in his heart became clear to all who beheld it. (transl. R. Doran)

The context of (9) is the violation of the Temple of Jerusalem by the functionary Heliodorus, who was sent by King Seleucus IV Philopator (187−175 BC) to confront the high priest Onias III. The author refers to the turmoil felt by Onias at the sight of the unprecedented impiety committed by the pagan. This book, which summarizes the much wider work of a certain Jason of Cyrene, shows the high level of refinement and erudition of Greek reached by Jewish culture of the time. Luciano Canfora (2013, 546) has already suggested that there is a probable Thucydidean echo in the prologue (2.28: διακριβοῦν περὶ ἑκάστων), a rare reference to a prose text in a book that often strives for a poetic style.36 We would do well to consider that the resumption of ἄλγος in Jewish-Hellenistic prose, in addition to responding to the contemporary trend of Asian rhetoric to reincorporate poeticisms, is a very Judaic stylistic trait, reshaping Greek and Homeric pain in a way that reflects a Hebrew perspective. In the Septuagint, an alternative to ἄλγος is ἀλγηδών, a term that is common in 2 Maccabees (with four attestations: 7.12, 9.5, 9.9, 9.11), notably well represented with a good number of attestations in 4 Maccabees, the most philosophical of these texts (3.18, 6.7, 6.34, 6.35, 8.28, 9.28, 13.5, 14.1, 14.11, 16.17), and found in Psalm 37 (38) (v. 17).37

 Lexical Matters in Greek of the Imperial Age Nevertheless, limited evidence from sporadic quotations cannot undermine a rather clear semantic development. In the language of medicine, the most common term for indicating bodily pain became ἄλγημα, as demonstrated by the high

 36 Some of these poetic terms, such as ἀπήμαντος, μόρος, σχέτλιος, στυγέω, or the preposition ἄτερ, a form such as ἄλγος or a whole elocution such as ἃ μὴ θέμις, seem to indicate imitation of Homer and tragedy. Faced with the epochal importance of the Maccabean revolt and the heroism of Judah and his men, the author of 2 Maccabees aimed above all to reproduce the tone of an epic, for example through the use of adverbial forms such as λεοντηδόν, ἀγεληδόν, or the neologism κρουνηδόν. On the literary character of the book, to which I dedicated my doctoral thesis (Rome 2015) under the guidance of Albio Cesare Cassio, see the commentary by Doran 2012 and the very recent dissertation by Domazakis 2018. 37 In the 1st century AD Flavius Josephus remarks that τὸ βουβῶνος ἄλγος ‘inflammation of the groin’ was also known as σαββάτωσις (Contra Apionem 2.22.1, 2.27.4; for a commentary, see Barclay 2007, 179). These two sporadic attestations of ἄλγος are not surprising for a Jewish author who aimed at μετασχεῖν τῶν ἑλληνικῶν γραμμάτων (AJ 20.263).

234  Enrico Cerroni number of attestations documented in Galen, an author who writes in the koine.38 An exception is the doctor Aretaeus of Cappadocia,39 who uses ἄλγος frequently (28x). In the Περὶ αἰτιῶν καὶ σημείων χρονίων παθῶν 4.12.3.4−5, the Cappadocian author writing in the Ionic dialect of the Hippocratic tradition40 presents a spectrum of various categories of pain corresponding to the density of the organs involved. When it strikes the ‘dense’ parts (τὸ πυκινόν) of the organism, the pain is less acute and perceptible (ἄλγος γὰρ τρηχείᾳ ἐν αἰσθήσει) compared to the ‘thinner’ zones (τὸ ἀραιόν).41 Subsequently, medicine would continue to prefer the more technical ἄλγημα, sustained now by centuries of linguistic usage by many authors, including Oribasius. Thus, Aetius of Amida, a physician of the 5th century AD, used ἄλγος just twice, while in the 7th century AD, in an era where the name Hippocrates was more than a thousand years old, even Paul of Aegina (625−690 AD), who specialized in gynaecology, used only ἄλγημα. Outside the confines of medicine and poetry, which remained substantially faithful to Homeric language, albeit with some divergences (ἄλγος is attested 74x

 38 Although registering a high frequency of ἄλγημα, the prestigious Dictionary of Medical Terms in Galen (Durling 1993, 29−31) finds only one case of ἄλγος, significantly in a quotation from a poetic passage of medical content from Philo of Tarsus (De compositione medicamentorum 13.268.1 Kühn = SH 690.9 πάντα δ’ ὅσα σπλάγχνοισιν ἐνίσταται ἄλγεα παύω). Regarding Galen, the TLG registers very few other attestations of ἄλγος, often in poetic quotations (De antidotis 14.35.13, 14.37.8 Kühn) and in a controversial passage of Quod animi mores corporis temperamenta sequantur (4.819.1 Kühn). 39 Its dating remains controversial, although today the period proposed includes the second half of the 1st century and the first half of the 2nd century AD. A detailed discussion of the problem is given by Roselli 2004, 164, who also analyzes the organization of Aretaeus’s work in eight books. 40 Regarding the vitality of the Ionic dialect in post-Classical prose, Cassio’s 1996 study is fundamental. 41 “Aretaios’ terminology for pain suggests that he is not determined to differentiate between pain terms: the treatises repeatedly slide between the use of ponos, odunē and algos. While there are occasions when specific terminology is used for different types of pain, it is clear that Aretaios extracts little intellectual or classificatory mileage out of shifts in terminology: pain = pain = pain. Where different language is adopted to describe particular conditions — such as when referring to pain in the head which arises from an immediate cause (kephalalgia) or its chronic relation (kephalaia) — he exchanges his general predilection for ponos in favour of the technical terminology for specific ailments: see κεφαλαλγίη at Aret. De causis et signis diuturnorum morborum 3.2.1.2” (King 2017, 68). Unlike Roselli 2005, 415, who argues that Aretaeus actively read and commented on Hippocrates, according to Oberhalm 1994, 966 “[w]hat is Hippocratic in Aretaios, in the final analysis, is not the medical theories and praxis, but only the style”. As regards ἄλγος, it seems that he was even more conservative than Hippocrates himself.

Prose and Poetry of Pain: A History of the Term ἄλγος  235

in the Posthomerica by Quintus Smyrnaeus, 16x in the Dionysiaca by Nonnus of Panopolis), Greek prose of the Imperial Age, marked by a division between Asianism and Atticism, at least agrees in a single micro-stylistic aspect. Next to the prevalent λύπη and πόνος, Greek used mainly ἀλγηδών for moral value and ἄλγημα for the more technical physical realm.42 In Plutarch, for example, in which ἄλγημα43 is used only once, the learned ἄλγος is limited to a handful of quotations.44 A similar argument can be made for Lucian, who quotes the epic ἄλγεα (see for instance ἄλγεα πάσχειν, Dial. mort. 77.5.2 Macleod) and revives ἄλγος in the amazing work of Ionic prose that is De Syria dea (39.12).45 An Atticist writer, Arrian of Nicomedia, in the Anabasis of Alexander exhibits a mannered ἄλγος οὐ σμικρόν (7.6.2) referring to the displeasure felt by the Macedonians when they see Alexander wearing Median clothes, in a passage where even the epic form σμικρόν instead of μικρόν points to the move towards a higher style: εἶναι γὰρ οὖν καὶ Μηδικὴν τὴν Ἀλεξάνδρου στολὴν ἄλγος οὐ σμικρὸν Μακεδόσιν ὀρωμένην.46 In the long history of the literary recovery of ἄλγος, there is an important development in the 4th century AD. The great Christian intellectuals, having received a pagan education, begin to comment on the sacred texts of the new religion, and the aforementioned passages of the Old Testament required a commentary, which was also — even if created with different tools from ours — of a lexical nature. Gregory of Nazianzus, for example, used the concept of ἄλγος in his own poetry, but more significant and important for the purposes of our in-

 42 Even the most famous sick writer in the second century AD, Aelius Aristides, refers to his disease in terms of ὀδύναι […] δειναί (47.62 Keil) or of ἀλγήματα ἰσχυρὰ καὶ δεινά (49.16 Keil). A rich analysis on the perception of pain on the part of the author of the Sacred Tales can be found in King 2017, 129−153. 43 Life of Sulla 26.3.1 (Σύλλᾳ δὲ διατρίβοντι περὶ τὰς Ἀθήνας ἄλγημα ναρκῶδες μετὰ βάρους εἰς τοὺς πόδας ἐνέπεσεν). 44 For instance, in the Consolatio ad Apollonium (15) ἄλγος οὐδὲν ἅπτεται νεκροῦ, with a corresponding verse in Aesch. fr. 255 Radt. 45 The term ἄλγος is missing from the lexicon of the novelists (Achilles Tatius, Chariton, Heliodorus, Iamblichus, Longus, and Xenophon of Ephesus), although we find ἀλγηδών (Achilles Tatius 1.6.2, 2.7.2; Heliodorus 2.30.1, 9.18.2) and ἄλγημα (Heliodorus 3.7.1, 4.7.6). On the other hand, the verb ἀλγέω is more frequent: see Conca/De Carli/Zanetto 1983, s.v. 46 ‘In fact they had long been pained to see Alexander wearing the Median robes, and his Persian marriage ceremonies had not given satisfaction to most of them’ (transl. I. Robson).

236  Enrico Cerroni vestigation are the occurrences in prose. A survey in the TLG uncovers a neglected passage of Eusebius’ Vita Constantini, on the death of the emperor.47 Eusebius — whose rhetoric and sustained prose Photius did not like (Bibl., cod. 13.4a) due to a presumed lack of elegance — regularly uses the common term ἀλγηδών, except when he quotes Homer (Praep. evang. 6.3.1, 6.8.2, 13.12.5) or when he mentions and comments on the Old Testament passages cited above. In the fourth book of the Vita Constantini, the erudite author refers to the death of his favourite emperor and the public displays of pain in the form of screams and cries that followed. We find here the rare iunctura ἐνδόμυχον ἄλγος: (10) Δῆμοί θ’ὡσαύτως τὴν σύμπασαν περιενόστουν πόλιν, τὸ τῆς ψυχῆς ἐνδόμυχον ἄλγος κραυγαῖς καὶ βοαῖς ἔκδηλον ποιούμενοι, ἄλλοι δὲ κατηφεὶς ἐπτοημένοις ἐῴκεισαν, ἑκάστου τε πένθος ἴδιον ποιουμένου ἑαυτόν τε κόπτοντος, ὡσανεὶ τοῦ κοινοῦ ἁπάντων ἀγαθοῦ τῆς αὐτῶν ἀφῃρημένου ζωῆς. (Euseb. Vita Constantini 4.65.3) The populace similarly wandered all round the city, expressing their inward anguish of soul with groans and cries while others were thrown into a sort of daze, as each one mourned personally and smote himself, as if their life had been deprived of the common good of all. (transl. A. Cameron/S.G. Hall)

Averil Cameron and Stuart G. Hall’s excellent historical commentary insists above all that the whole scene of public mourning was constructed “as the antithesis of rejoicing and adventus”, where “all orders and all ages weep and lament for the Emperor, invoking him in traditional terminology as saviour and benefactor, while in addition the soldiers mourn him as their good shepherd”.48 Not even the more recent biographical work on Constantine by Bleckmann/ Schneider 2007 comments on the lexical aspect of this text. On the other hand, it does not seem out of place to emphasize that we are faced with a clear attempt to raise the style, required by an epochal event such as the death of Constantine, the emperor who was the model for Christians in the 4th century. Therefore, the intentionally refined and moving lexicon also contributes to the “extravagance of the scene” (Cameron/Hall 1999, 343).49

 47 I will not tackle the question of the authorship of this work: see the recent study by Francesco Corsaro, which defines it “piuttosto discutibile sotto il profilo letterario per il suo carattere farraginoso, non esente talora da una certa sciatteria” (Corsaro 2012, 284). 48 Cameron/Hall 1999, 343. 49 It should be added that the adjective ἐνδόμυχος is a poetic term as well, a hapax in Sophocles (Phil. 1457), later attested in Callimachus, Galen, and very often in Nonnus’ Dionysiaca.

Prose and Poetry of Pain: A History of the Term ἄλγος  237

In conclusion, the rich and articulated history of the term ἄλγος, which became quickly and unforeseeably poetic, can be followed from the Iliad and the Odyssey down to Greek Christian literature. Greek is remarkable in its very subtle emotive terminology, which in many cases anticipates semantic shifts that can also be observed in modern languages (for example, the Hellenistic θλίψις, from ‘physical pressure’ to ‘oppression’ and ‘affliction’). It could not obliterate the aristocratic and Homeric ἄλγος, snubbed by the innovator Hippocrates, but recorded in the lexicon of Babiniotis (2012) and still used in Modern Greek in certain contexts. A poet of the last century, Odysseas Elitis, in the first hymn of the Genesis of the Άξιον Εστί, celebrated God the creator as both pain and joy (αλγηδόνα κι ευφροσύνη) not with άλγος, but with the cognate αλγηδόνα, aiming to raise the style of the piece. In fact, the younger αλγηδόνα, a term that was a novelty for Euripides, is no longer in use today, and Elitis elegantly recovered it as an archaism. Meanwhile, as happened with many words in Classical Greek, άλγος, so quickly adopted already in the 5th century BC in poetry, has been retrieved and continues to be used in Modern Greek.

Bibliography Ast, F. (1835–1838), Lexicon Platonicum sive vocum Platonicarum index, Lipsiae. Babiniotis, G. (2012), Λεξικό της νέας ελληνικής γλώσσας, 4th edn., Athena. Barclay, J.M.G. (2007), Flavius Josephus: Against Apion, in: S. Mason (ed.), Flavius Josephus: Translation and Commentary, vol. 10, Leiden/Boston. Bleckmann, B./Schneider, H. (2007), Eusebius von Caesarea: De vita Constantini. Über das Leben Konstantins, Turnhout. Bowra, C.M. (1934), “Homeric Words in Cyprus”, in: JHS 54, 54–74. Bowra, C.M. (1960), “Γλῶσσαι κατὰ πόλεις”, in: Glotta 38, 43–60. Cameron, A./Hall, S.G. (1999), Eusebius’ Life of Constantine: Introduction, Translation and Commentary, Oxford. Canfora, L. (2013), Storia della letteratura greca, 3rd edn., Roma/Bari. Cassio, A.C. (1996), “La prose ionienne postclassique et la culture de l’Asie mineure à l’époque hellénistique”, in: C. Brixhe (ed.), La koiné grecque antique, vol. 2: La concurrence, Paris/Nancy, 147–170. Chantraine, P. (1933), La formation des noms en grec ancien, Paris. Conca, F./De Carli, E./Zanetto, G. (1983), Lessico dei romanzieri greci, vol. 1: α−γ, Milano. Corsaro, F. (2012), “Costantino ed Eusebio nella Vita Costantini di Eusebio di Cesarea”, in: RCCM 2, 283–300. De Foucault, J.-A. (1972), Recherches sur la langue et le style de Polybe, Paris. Domazakis, N. (2018), The Neologisms in 2 Maccabees, diss. Lund. Doran, R. (2012), 2 Maccabees: A Critical Commentary, Minneapolis.

238  Enrico Cerroni Dörrie, H. (1956), “Leid und Erfahrung: Die Wort- und Sinnverbindung παθεῖν-μαθεῖν im griechischen Denken”, in: Abhandlungen der Akademie der Wissenschaften in Mainz 5, 303–344. Durling, R.J. (1993), A Dictionary of Medical Terms in Galen, Leiden/New York/Köln. Egetmeyer, M. (2010), Le dialecte grec ancien de Chypre, vol. 1, Berlin/New York. Fatouros, G. (1966), Index verborum zur frühgriechischen Lyrik, Heidelberg. Gomme, A.W. (1956), A Historical Commentary on Thucydides, vol. 1, Oxford. Gomme, A.W./Sandbach, F.H. (1973), Menander: A Commentary, Oxford. Hornblower, S. (2013), Herodotus: Histories, vol. 5, Cambridge/New York. How, W.W./Wells, J. (1928), A Commentary on Herodotus, vol. 2, Oxford. Huart, P. (1968), Le vocabulaire de l’analyse psychologique dans l’oeuvre de Thucydide, Paris. Hunter, R.L. (1983), Eubulus: The Fragments, Cambridge. King, D. (2017), Experiencing Pain in Imperial Greek Culture, Oxford. Lanza, D. (1997), “Pathos”, in: S. Settis (ed.), I Greci: Storia, cultura, arte e società, vol. 2.2: Una storia greca: Definizione, Torino, 1147–1155. Long, A.A. (1968), Language and Thought in Sophocles: A Study of Abstract Nouns and Poetic Technique, London. Loraux, N. (1982), “Ponos: Sur quelques difficultés de la peine comme nom de travail”, in: AION(archeol.) 4, 171–192. Macan, R.W. (1895), Herodotus: The Fourth, Fifth and Sixth Books, London. Martínez Hernández, M. (1977), “El campo léxico de los sustantivos de dolor en Sófocles: Ensayo de semántica estructural-funcional”, in: CFC 13, 33–112. Marzullo, B. (1999), “Il dolore in Ippocrate”, in: QUCC 63, 123–128. Mawet, F. (1979), Recherches sur les oppositions fonctionnelles dans le vocabulaire homérique de la douleur (autour de πῆμα-ἄλγος), Bruxelles. Musti, D. (1997), Demokratía: Origini di un’idea, 2nd edn., Roma/Bari. Nenci, G. (2006), Erodoto: Storie, vol. 5: La rivolta della Ionia, 3rd edn., Milano. Oberhalm, S. (1994), “On the Chronology and Pneumatism of Aretaios of Cappadocia”, in: ANRW 2.37.2, 941−966. Prost, F. (2004), Les théories hellénistiques de la douleur, Louvain/Paris. Prost, F. (2016), “Douleur physique et douleur morale dans les Tusculanes de Cicéron”, in: AntPhilos 10, 9−22. Radice, R. (2005), Lexikon III: Aristoteles, Milano. Rey, R. (2011), Histoire de la douleur, 3rd edn., Paris. Rijksbaron, A. (1989), “D’où viennent les ἄλγεα? Quelques observations à propos d’ἄλγε’ ἔχειν chez Homère”, in: F. Létoublon (ed.), Les langues et les textes en grec ancien. Actes du colloque Pierre Chantraine (Grenoble, 5−8 Septembre 1989), Amsterdam, 181−191. Roselli, A. (2004), “Les malades d’Arétée de Cappadoce”, in: J. Jouanna/J. Leclant (eds.), Colloque La médecine grecque antique, Paris/Beaulieu-sur-Mer, 163−176. Roselli, A. (2005), “Areteo di Cappadocia lettore di Ippocrate”, in: P. van der Eijk (ed.), Hippocrates in Context. Papers read at the XIth International Hippocrates Colloquium (University of Newcastle upon Tyne, 27−31 August 2002), Leiden/Boston, 413−432. Schein, S.L. (2013), Sophocles: Philoctetes, Cambridge. Schick, C. (1956), “Appunti per una storia della prosa greca, 3: La lingua di Erodoto”, in: Memorie dell'Accademia dei Lincei 7, 345–395. Seiler, H. (1950), Die primären griechischen Steigerungsformen, Hamburg. Slater, W.J. (1969), Lexikon to Pindar, Berlin. Southard, G.C. (1970), The Medical Language of Aristophanes, Diss. Baltimore.

Prose and Poetry of Pain: A History of the Term ἄλγος  239

Stein, H. (1894), Herodotos, vol. 3: Buch 5. und 6., 5th edn., Berlin. Thornton, J. (2014), “Commentary”, in: D. Musti (ed.), Polibio: Storie, 3rd edn., Milano. Torchio, M.C. (1999), Il Pluto di Aristofane, Torino. Untersteiner, M. (1954), Sofisti: Testimonianze e frammenti, Firenze. Usener, H. (1977), Glossarium Epicureum, Roma. van Groningen, B.A. (1950), Herodotus’ Historiën, vol. 4: Commentaar op boek 4–6, Leiden. van Leeuwen, J. (1968), Aristophanis Plutus, 2nd edn., Leiden. Vernant, J.-P. (1971), Mythe et pensée chez les Grecs, Paris. White, H. (1979), Studies in Theocritus and Other Hellenistic Poets, Amsterdam.

Olga Tribulato

Making the Case for a Linguistic Investigation of Greek Lexicography: Some Examples from the Byzantine Reception of Atticist Lemmas Abstract: Ancient lexica give us important information about linguistic evolution and its perception by native speakers. However, linguistic investigations of lexicographical material are slow to come to the forefront of Classical linguistics and even more so of Medieval Greek linguistics. This paper makes a contribution in this direction by investigating the relevance of lemmas handed down from Atticist lexicography in the linguistic context of Medieval Byzantium, where writers had to move between different registers and were often confronted with the differing usages of the two most important models: the Classics and the Scripture. Keywords: Atticism, lexicography, post-Classical Greek, Byzantine Greek, Medieval Greek.

 Approaching Ancient Lexicography from a Linguistic Perspective The paths through which knowledge of Ancient Greek has unfolded across time are very often traced in the pages of dictionaries. Ancient and medieval Greece are no exception. Ever since the Hellenistic Age, Greek erudition has devoted great attention to collecting rare words (γλῶσσαι) or notable terms (λέξεις) and glossaries and lexica have had an important role in Greek learning and, consequently, in perpetuating the multifarious character of the language. The usefulness of ancient lexica for modern readers obviously does not reside uniquely in the old, rare or important words which they collect. Yet modern readers often tend to treat these sources as mere ancillary tools and not as scholarly works in their own right, with the result that there is little interest in the reasons why certain lemmas were included in a lexicon, or in the methodology and linguistic reasoning behind certain interpretamenta. Similarly, scholars of Antiquity rarely address the semiotic and hermeneutical questions elicited by the very ontology of the container, the lexicographical ‘list’ itself. Some of these questions can be illustrated with examples from the https://doi.org/10.1515/9783110621747-012

242  Olga Tribulato great tradition of Atticist lexicography. The first obvious fact is that the ‘list’ may often be only the result of the compression of a different kind of work: a case in point is the epitome of Phrynichus’ Praeparatio sophistica, originally a huge discursive treatise on rhetorical style in 37 books, now consisting of fragments and scattered citations in other works. Unlike modern dictionaries, these ‘lists’ were always an open product, in which information was edited in and out according to the interests and dispositions of their compilers and conveyors: this is particularly obvious in the case of onomasiological dictionaries such as Pollux’s Onomasticon, in which new terms were added to the original collection by later compilers. When we are faced with works which, like the Praeparatio sophistica and Pollux’s Onomasticon, are characterized by a spare style and reduced syntax, another question that arises is whether these were original features or the result of later shortening and, following from this, whether it is possible to identify the authorial ‘voice’ and methodology of such works. In the last twenty years or so some of these issues and others have begun to find a place in Classical scholarship, which is slowly integrating the traditional textual-philological approach (according to which lexica are the object of Textkritik, Textgeschichte or Überlieferungsgeschichte) with a more fine-grained investigation of the methodology and theoretical stances underlying the ancient lexica. In the realm of Atticist lexicography, for instance, Stefanos Matthaios has made a good case for the identification of a theory of linguistic registers and sociolects in the hitherto seemingly undifferentiated terminology of Pollux’s Onomasticon.1 Before him, the need to take Atticist lexica seriously and to analyze them with a linguistically oriented approach was defended by Albio Cassio in a brief but paradigmatic contribution on the role of the Antiatticist as an advocate of those koine and dialectal forms which were condemned by Atticism.2 Cassio’s interest in revealing the Atticist lexicographers’ nuanced views on language has inspired fruitful lines of research on the part of his older and younger pupils: a case in point is Carlo Vessella’s work on correct pronunciation in Atticist lexicography, a topic that according to past scholarship the Atticists were hardly interested in.3

 1 Matthaios 2013; Matthaios 2015. 2 Cassio 2012. 3 Vessella 2010; Vessella 2018. At the University of Cambridge Chiara Monaco is at work on a PhD thesis on the Atticist take on comic language, which issues from the MA she completed at Rome under Cassio’s supervision (Monaco 2015). For further work on the linguistic theorization of Atticist lexicography see Tribulato 2014; Tribulato 2016; Tribulato 2018.

Making the Case for a Linguistic Investigation of Greek Lexicography  243

While we are now beginning to be better equipped to undertake a linguistic study of the Atticist take on issues of language correctness and linguistic evolution, the survival of Atticist material in the Byzantine Age, and particularly in its lexica, still awaits — with very few recent exceptions — to be addressed in a truly linguistic perspective.4 This contribution seeks to make a small step in this direction by investigating what linguistic motivations may lurk behind the Byzantine interest in certain ancient glosses. The inspiration behind this pilot study comes from a recurrent piece of advice that Albio Cassio, following in the footsteps of eminent Greek linguists such as Albert Thumb, always gave his pupils: that we should not be blind to the post-Classical (i.e. Byzantine and modern) evolution of Greek if we wish to understand the ancient roots of the language.5

 Lexicography at Byzantium: From Textgeschichte to Linguistics The great lexica associated with the so-called ‘age of the sylloge’ — the Συναγωγὴ λέξεων χρησίμων, Photius’ lexicon, the Suda — played a fundamental role in the survival of ancient exegetical material and its transmission to later ages, by collecting and systematizing a lexicographical heritage that would have otherwise been lost.6 The intellectual circles of 9th-century Constantinople were also responsible for the epitomizing and copying of Atticist works such as Pollux’s Onomasticon as well as those which were collected in the slightly later lexicographical miscellany of cod. Paris. Coislinianus gr. 345. This important collection of ancient lexicography, to which we will be returning at several stages in this paper, contains the only extant copy of the Antiatticist, the epitome of Phrynichus’ Praeparatio sophistica, an important ‘expansion’ of the Synagoge (= Σb, once known under the name of ‘Lexicon Bekkeri VI’: see Section 3), and Moeris’ Atticist, as well as other important lexica.7  4 For partial exceptions see Matthaios 2006 and Matthaios 2010, both on the Suda, and Cuomo 2017, on lexica, school manuals and their usefulness for a socio-linguistic investigation of Medieval Greek. 5 This is theorized in Cassio 2014. 6 On the pivotal exegetical activity of this period see Lemerle 1971; Hunger 1978, 33−50; Pontani 2015. 7 The early history of cod. Paris. Coisl. gr. 345 is reconstructed in de Leeuw 2000; Cunningham 2003, 57; Valente 2008; Valente 2012, 28−30; Valente 2015, 12. The idea that Arethas of Caesarea is directly behind the manuscript has now been abandoned, but some parts of the codex may

244  Olga Tribulato While eminent philologists have produced seminal studies on the Textgeschichte and Überlieferungsgeschichte of Byzantine lexica,8 a substantial gap still extant in modern literature concerns the way in which lexicography interacted with and impacted language development during the Byzantine Age. The limited interest in the socio-linguistic dimension of these works is a result of the combined effect of two attitudes. First, the widespread prejudice that Byzantine lexica, like most works of this age, are repetitive and unoriginal, nothing more than expanded mechanical compilations.9 Secondly, the general neglect of linguistics in the tradition of Byzantine studies.10 Among other evident shortcomings, this has produced a bizarre situation in which linguists may be interested in ancient lexicography, but not at all in its Byzantine counterparts, in spite of the fundamental role played by the latter from the point of view of language transmission. This neglect supports the statement that “the history of manuscripts passes through Byzantium […] the history of ideas and literature routinely jumps from St. Augustine to the Renaissance” (Kaldellis 2007, 4). One may well be inclined to ask: why would such specialist works have had a role in the evolution of Medieval Greek? For Greek speakers in Byzantium, lexica of course were first and foremost a guide to the Classical language and its creative reproduction in Byzantine literature.11 But lexica in fact also contain valuable indirect information about the Byzantines’ view of their own language and can be argued to rank among those texts which were “factors of linguistic variation and change and influenced the choices of medieval authors” (Cuomo 2017, 452). This kind of information is all the more valuable because the Byzantine grammatical tradition, albeit rich, is remarkably silent on the topic of the language of its day.12 In past scholarship the appreciation of the Byzantine relation with ancient culture and language has suffered from the idea that it was merely an exercise in slavish imitation steeped in a static diglossia which opposed learned and archaizing ‘Byzantine Greek’ to the spoken, low-level ‘vernacular’ or

 reflect the activity of his circle, as cautiously proposed by Ucciardello 2006, 63 n. 119 and Ucciardello 2012b, 91−94. Arethas also had a role in the transmission of Pollux’s Onomasticon: see Bethe 1900, v−vi. 8 Seminal studies concerning the Textgeschichte are e.g. Reitzenstein 1897; Erbse 1950; Trapp et al. 1988; Hörandner/Trapp 1991; Trapp/Schönauer 2007. Classic and more recent general introductions are Cohn 1900; Tolkiehn 1925; Alpers 1990; Alpers 2001; Tosi 2015. 9 Thus Alpers 2001, 205: “nichts weiter als weitgehend mechanische Kompilationen aus den ihnen vorliegenden Quellen”. 10 See Manolessou 2014, 13. 11 See e.g. Hunger 1991 on the Suda. 12 Robins 1993; Manolessou 2014.

Making the Case for a Linguistic Investigation of Greek Lexicography  245

‘Medieval Greek’.13 Recent scholarship, however, has replaced this somewhat stereotyped view with a more fine-grained approach, claiming that the linguistic situation of medieval Byzantium was characterized by a linguistic continuum in which choices of style, register, vocabulary and grammar could vary considerably not only from the written to the spoken level (as scholarship has always assumed) but even within the writings of the same author(s).14 These new lines of research have not yet made an impact on theoretical approaches to Byzantine lexicography. We continue to assume that Byzantine lexicographers registered certain ancient words because (a) they were used in the great literary works of the Classical past and/or (b) because they were rare terms, and hence difficult for Byzantine readers. While it is certainly true that both criteria must have informed a lot of the choices made in Byzantine lexica, such a backward-looking viewpoint — which keeps explaining Byzantium by turning to the Classical past — risks levelling out the constituent criteria of Byzantine lexicography to a mere erudite game, thus overshadowing the motivations behind the perpetuation of lemmas which do not completely fall under criteria (a) and (b) above. Why were Byzantine lexicographers interested in words coming from works which were no longer read in their times (e.g. Doric comedy)? Conversely, why were lemmas used by famous Classical authors and included in ancient lexicography dropped in Byzantine lexica? A case in point is Phrynichus Eclogae 54 Fischer. This entry recommends using the verb καταδαιμοινάω in its correct Attic inflection in ‑άω rather than conjugating it as a verb in ‑έω (καταδαιμοινέω). Yet, with the unique but understandable exception of the late-medieval lexicon of Thomas Magister (who heavily draws on Phrynichus), Byzantine lexica are silent on both forms. In spite of Phrynichus’ recommendation and of the Attic pedigree of καταδαιμοινάω, this form simply dropped out of use: but why, since it was an erudite relic, were lexicographers not interested in it?15 The answer to this and similar questions would of course require a wide-ranging study of the main Byzantine lexica in relation to their ancient antecedents as well as of the different transmission paths and the Nachleben of the literary works  13 E.g. Meillet 1930, 23; Böhlig 1956; Beck 1971, 1−4; Hunger 1969−1970, 29−30. For the terminological divide, see Rollo 2008. 14 For examples of this approach see e.g. Ševčenko 1981; Trapp 1993; Toufexis 2008; Hinterberger 2014; Horrocks 2014. 15 As far as one can ascertain, in post-Classical Greek the verb is almost always inflected in the ‑έω conjugation. A possible, but perhaps only partial answer is that the circulation of the Eclogue was limited before the age of Thomas Magister: I thank Giuseppe Ucciardello for pointing this out to me. Still, this does not explain why Phrynichus’ lemma was not picked up by later ancient lexica.

246  Olga Tribulato of Antiquity, an enterprise which is well beyond the scope of an article. Here I will tackle a few concrete examples originating in Atticist lexicography in order to highlight some linguistic facts which may have guided the Byzantine re-use of certain ancient lemmas. I will start with the two ‘traditional’ criteria ruling the selection of the lemmas mentioned above, namely: (a) the fact that a certain word was used in the great works of the Classical past which were considered models (Section 3) and (b) the fact that a certain ancient word may have been difficult for Byzantine readers because it had dropped out of use (Section 4). In the last part of the paper (Section 5) I will turn to investigate a third criterion, i.e. the fact that a certain word may be registered by Byzantine lexica because its semantics had undergone a complex evolution in the transition from Ancient to Byzantine Greek and hence required special attention. It may be argued that this third criterion is a subcategory of one or both of the other two. Indeed, educated medieval Greeks may be interested in certain words which they still used precisely because they were also used by the ancients (criterion a) or, in the case of semantic change, because the modern meaning had no parallels in Ancient Greek (criterion b). However, by introducing a third criterion I will make the case for the importance of a diachronic linguistic approach to Byzantine lexicographical material: this approach, I wish to argue, will help us understand the relevance that certain words may have had for Byzantine scholars who lived in a diglossic situation.

 Bolstering Attic: A Diagnostic Example A good example of the way Byzantine lexicography perpetuated the linguistic usage of Classical Attic is represented by the lemmas opposing κηλίς and σπίλος (both ‘stain, blemish’).16 In his lexicon Photius devotes two entries to κηλίς:

 16 Properispomenon accentuation in Fischer (1974) likely depends on some uncertain information that Lentz collected and edited as Hdn. GG 3.1, 154.17−22, where σπῖλος is added from another treatise (Περὶ μονεροῦς λέξεως, GG 3.2, 920.25−28 Lentz), in which Herodian contrasts words with ει with those in ι. Most modern editions and dictionaries disregard Herodian’s testimony because derivations such as σπιλάς ‘stain’ and ἄσπιλος ‘stainless’ have a short vowel in metrical texts, though evidence for the short quantity of these derivations is late. On the problem of vowel quantity and diachrona in Atticist thought, see Vessella 2018, 64–95; his treatment of iota (Vessella 2018, 92) does not mention σπίλος since this word is not the object of a specific orthoepic prescription in the Atticist lexica.

Making the Case for a Linguistic Investigation of Greek Lexicography  247

(1)

κηλῖδες· σπίλοι.

(Phot. κ 660 Theodoridis)

κηλῖδες: stains. (2)

κηλίς· ῥύπος, μῶμος, μομφή, ἕλκος, οὐλή.

(Phot. κ 662 Theodoridis)

κηλίς: meanness, shame, blame, sore, wound.

The two lemmas seem to account for different semantic functions of κηλίς. The entry under (1) refers to the concrete meaning of the word (‘stain’), which is likely to have been its original meaning. It is no accident that the lemma refers to κηλῖδες in the plural, since the first attestations of the plural in Attic tragedy are not metaphorical. (1) may therefore have a specific Atticist background (see no. (3) below).17 From its original meaning of a fleck of blood, by metaphorical extension κηλίς came to indicate its cause (a wound) and ethical counterpart (a moral or spiritual blemish).18 The lemma under (2) accounts for these metaphorical usages: κηλίς as a moral stain (ῥύπος, which I take in its metaphorical meaning because it is followed by μῶμος and μομφή) or a physical wound (ἕλκος, οὐλή). As far as we can see, κηλίς was frequently used by Attic authors and later spread to various registers of Greek. It is clear that by the 2nd century AD it had become the more common synonym of σπίλος, as Phrynichus implies by condemning the latter: (3) σπῖλος· καὶ τοῦτο φυλάττου, λέγε δὲ κηλίς.

(Phryn. Eclogae 20 Fischer)

σπῖλος: avoid this word as well. Use κηλίς instead.

Although throughout its history σπίλος is a very common word, in post-Classical Greek it was associated with the lower registers of the koine. Since it featured in two influential passages of the New Testament — St. Paul’s Epistle to the Ephesians (5.27.2 Aland et al.) and St. Peter’s Epistle 2 (2.13 Aland et al.) — σπίλος en-

 17 The two entries in Photius’ have close parallels in Hesychius’ lexicon (κ 2504 Latte and σ 1514 Hansen). Both lexica may ultimately draw on Diogenianus, as Theodoridis (ad Phot. κ 660) tentatively suggests. 18 The etymology of κηλίς is a matter of debate, but GEW links it to other IE terms indicating a kind of pallor (Lat. cāllidus, Umbr. kaleřuf, Lith. kalýbas), a stain (Skt. kalmaṣa, a loanword, OI caile), or a dark colour (Skt. kāla‑).

248  Olga Tribulato joyed huge popularity in Christian Greek and religious exegesis. Phrynichus’ rejection of the word, paired with his endorsement of κηλίς — the supremacy of which is based on its being an Attic term — is very likely the reason behind Photius’ attention to κηλίς and his preference for the word in his own writings.19 Attic prestige also explains a second example, which Byzantine lexicography inherits from Phrynichus’ Eclogue: (4) σκίμπους λέγε, ἀλλὰ μὴ κράββατος· μιαρὸν γάρ.

(Phryn. Eclogae 41 Fischer)

Use σκίμπους but not κράββατος, for it is bad.

Phrynichus recommends using σκίμπους (‘small bed’) and not κράββατος, a koine word of non-Greek origin which is continued in the Modern Greek form κρεβάτι.20 As far as one can ascertain, both words could be used alternatively to name a small single bed of lowish quality, perhaps nothing more than a wooden base for a mattress, as opposed to the more luxurious κλίνη. However, κράββατος seems to have been associated with Hellenistic Greek and this is probably the reason why Phrynichus condemns it: it was standard word for ‘bed’ in the New Testament, and even Pollux cannot find any better authorities than the Hellenistic and non-Attic playwrights Rhinton (fr. 9 K.-A.) and Crito (fr. 2 K.-A.) to defend it. σκίμπους, on the contrary, could be said to have a Classical Attic pedigree: it is used by Aristophanes (Nub. 254; 708), Plato (Prt. 310d.1) and Xenophon (An. 6.1.4).21 In Hellenistic Greek σκίμπους is comparatively rare, but it has a number of attestations in ‘good’ authors variously linked to Atticism such as Lucian, Aelius Aristides and Dio Cassius; Pollux quotes it three times as a synonym of ‘bed’.22 As in the case of many other Greek pairs of synonyms, σκίμπους and κράββατος both continued to be in use throughout the Byzantine Age. The slightly higher number of the attestations of κράββατος shows how the common

 19 Outside his lexicon, Photius uses σπίλος four times, against the twenty attestations of κηλίς. 20 The etymology is uncertain. GEW s.v. endorses Paul Kretschmer’s view that it was a Macedonian-Illyrian word, while EDG s.v. connects it to σκίμπτομαι. 21 In the Clouds this is the bed or small couch on which Socrates makes Strepsiades sit during his ‘initiation’ into the school (Nub. 254) and on which Strepsiades is reclining during his fight with bedbugs: the contrast may therefore be between the eccentric ritual of initiation into the exclusive Socratic school and the prosaic reality of its context (see Dover 1968, 131). The attestations in Plato and Xenophon further clarify that this was a lower-quality bed, which could be easily transported and also used as a field-bed (see Pritchett 1956, 231) or a hammock. On the σκίμπους in general, see Rodenwaldt 1927. 22 Poll. 6.9, 10.35, 10.36 Bethe.

Making the Case for a Linguistic Investigation of Greek Lexicography  249

word managed to enter literary language in spite of competition from the higherregister synonym. This is well illustrated by Thomas Magister (σ 333 Ritschl) who, drawing on Phrynichus and perhaps also on the terminology of some other sources, defines σκίμπους as ‘ῥητορικόν’ and κράββατος as ‘κοινόν’. As Thomas Magister’s comments show, the attention devoted to σκίμπους by late-antique and Byzantine lexicographers was no doubt meant to perpetuate the use of an ancient word associated with the Attic dialect against the much more common κράββατος, which Moeris marks as typical of the Ἕλληνες and which Polybius of Sardi condemned as a barbarism.23 Earlier Byzantine sources provide a clear glimpse of the uneasiness which the use of κράββατος aroused in learned authors: (5) ἀσκάντης· κλινίδιον εὐτελὲς καὶ ὑπὸ τῶν Ἀττικῶν ὁ σκίμπους. ὁ δὲ κράββατος οὐδὲ παρ’ ἑνί. (Σb α 2238 Cunningham)24 ἀσκάντης: a small cheap bed, called ‘σκίμπους’ by Attic-speakers. But κράββατος is never attested in any of them.

This lemma belongs to the ‘expanded version’ of the Συναγωγὴ λέξεων χρησίμων, the important anonymous lexicon assembled in the late 8th century which has survived through a copy close to the original preserved in two manuscripts (Paris. Coisl. gr. 347 and Paris. Suppl. gr. 1243 I). The expanded version of the Synagoge, to which I here refer with the conventional abbreviation ‘Σb’,25 is instead preserved in cod. Paris. Coisl. gr. 345 as well as in quotations in Photius and the Suda depending on other similar expanded versions.26 While Σ (the original version of the Synagoge) has the simple lemma σκίμπους· κράβατος (sic) ἢ σκάμνον (σ 122 Cunningham) and no entry for ἀσκάντης, the entry in Σb (no. (5) above) clearly draws from an Atticist source and conflates information on ἀσκάντης, σκίμπους  23 Cf. Moer. α 119 and σ 33 Hansen; Polybius Rhetor 285.7−9 Nauck. 24 Cf. Phot. α 2958 Theodoridis; Eust. Od. 2.302.44−45 Stallbaum. 25 The expanded version of the Synagoge contained in cod. Paris. Coisl. gr. 345 (“versio codicis B” in Cunningham 2003) mostly contains additions to the letter α. These depend on lost expanded versions of the original Synagoge which were independently used also by Photius, the Suda, and the Etymologicum Genuinum. To account for the different origins of these additions in α Cunningham 2003, 49−50 distinguishes glosses in cod. B with three sigla: Σ´ (a version of Σ used by B, Photius and the Suda), Σ´´´ (a version of Σ used by B and Photius only), Σb (“glosses found only in B […] or occasionally only in B and Et. Gen.”). For our present purposes it is not important to distinguish between these different expanded versions: I have therefore used the generic siglum Σb to refer to all glosses transmitted in the expanded version(s) of B. 26 The transmission history of this important lexicon has been magisterially reconstructed and explained by Reitzenstein 1907 and Cunningham 2003, 13−14; 43−58.

250  Olga Tribulato and κράββατος.27 The point of a lemma of this kind must reside precisely in the prescriptive character of the original Atticist source: σκίμπους provided learned Byzantine authors with an alternative to κράββατος endowed with a Classical Attic ancestry.

 Ionic, Attic and What Was Interesting for Byzantine Scholars: A Case-study from Herodotus Although profoundly influenced by Atticism, high-register Byzantine Greek did not look to Attic as the only ancient variety worthy of imitation. The Byzantines’ creative dialogue with ancient models suggests a form of classicism in which nonAttic models could also find a place. Two cases in point are Homer and Herodotus, whose authority Atticism itself never challenged and who remained reference points for poetic and prose style throughout the Byzantine Age.28 With these authors, the linguistic negotiation carried out by Byzantine writers is more subtle, because of course the Homeric Kunstsprache and Herodotus’ Ionic are phonologically and morphologically far from Classical Attic: the available evidence suggests the re-working of phrases, lines and words, as well as the more common practice of quoting verbatim certain passages, rather than a direct imitation of linguistic traits.29 Addressing the standing enjoyed by Herodotus in Byzantium, however, is no easy task, since there is no general study of his Byzantine reception. The narrower

 27 This source may be Moer. 119 Hansen, unless the gloss came to the Synagoge via Hesychius (and Diogenianus). Cunningham 2003, 50−57 addresses the expansions of the Synagoge and their addition of Atticist material to the original version (which mostly follows the neutral orientation of its main source, Cyril’s lexicon), while Cunningham 2003, 43−48 discusses the relationship between Cyril and Σ. The ‘neutral’ glossing of the word also characterizes three entries in Hesychius (σ 994, 996, 997 Hansen) which are sometimes repeated in later lexica. 28 On this role of Herodotus in the age of linguistic Atticism see Tribulato 2016. 29 Homer remained a staple of Greek education and rhetorical training throughout the Byzantine Age: see e.g. Browning 1992. Herodotus, on the other hand, remained a model of ‘pleasurable’ prose and narrative style (e.g. Psellus, Orationes panegyricae 1.155 and 8.41 Dennis): the cultural reasons for his enduring popularity in Byzantine times are discussed by Kaldellis 2012, 79, who also briefly addresses Thucydides’ different standing in Byzantine rhetoric.

Making the Case for a Linguistic Investigation of Greek Lexicography  251

question of how Herodotean language was viewed by Byzantine writers and represented in contemporary lexicography has been completely ignored.30 A quick check on the Index of Cunningham’s edition of the Synagoge shows that a considerable part of the Herodotean material contained in this lexicon may have come from Atticist sources (either directly or via intermediaries).31 It is therefore to be expected that the Synagoge inherited the contrastive structure of Atticist lexicography, opposing Attic and non-Attic phonological and morphological features and semantics. A good illustration of this is the lemma of Σb commenting on the inflection of the infinitive of ἀποχράω: (6) ἀποχρᾶν· ἐξαρκεῖν. Ἡρόδοτος.

(Σb α 2035 Cunningham)

ἀποχρᾶν: to suffice. Herodotus (3.138.2, etc.).

The point of this entry (which finds numerous closer or looser parallels in other Byzantine lexica and etymologica) most probably is not, as it may appear at first sight, to explain the meaning of ἀποχρᾶν (‘to be sufficient’, i.e. ἐξαρκεῖν), but to alert readers to the fact that while in Attic χράω and its compounds contract in /ε:/ (ἀποχρῆν), in Herodotus they contract in /a:/.32 Since ‘regular’ ‑άω verbs contract /ae/ into /a:/, not /ɛ:/, the layman (just like the modern student) must have

 30 It is remarkable that none of the major companions to Herodotus (e.g. Bakker/de Jong/van Wees 2002; Dewald/Marincola 2006) has a chapter on the Greek Middle Ages. This is all the more striking in the case of Priestley/Zali 2016, given its overt focus on reception (the key-word ‘Byzantium’ is even missing from the Index). The papers given at the conference The Afterlife of Herodotus and Thucydides at the Warburg Institute in March 2014, including Elizabeth Jeffreys’ “Byzantine Receptions of Herodotus and Thucydides”, do not appear to have been published yet. 31 According to Cunningham’s index, there are nineteen lemmas in the Synagoge that mention Herodotus by name, most of them in Σb. Of these nineteen lemmas, three have been attributed to fragments of Phrynichus’ Praeparatio sophistica by its editor, de Borries: Σb α 472, 806, 2201 Cunningham (the latter with loose parallels in Antiatt. α 138 Valente and Phryn. Eclogae 66 Fischer, both without attribution to Herodotus); two may come from Aelius Dionysius (α 522, 2156 Erbse) and two (α 711, 2035 Erbse) from the Antiatticist. As for Σ, at least σ 108 comes from Aelius Dionysius. 32 Orus fr. 13 Alpers (= Pseudo-Zonar. 274 Tittmann), on which see no. (9) below; Suda α 3652 Adler; schol. Lucian. 214.26 Rabe and, with a different verbal form, Phot. α 2724 Theodoridis (on which see below).

252  Olga Tribulato regarded forms such as ἀποχρῆν as one of the bizarre ‘exceptions’ of Attic phonology (which is indeed how they are regarded today).33 Yet, in spite of its irregularity, it was the Attic contraction that learned Greek writers adopted in their high language: forms such as χρῆν, ἀποχρῆν, ἀπέχρη (impf. 3rd sing.) are the norm in literary texts. This is but another example of the overpowering prestige of Attic, which won against morphological regularity and the pressure of analogy. In post-Classical spoken Greek, however, the situation may have been quite different. Papyri and inscriptions alike bear traces of analogical forms such as (‑)χρᾶσθαι for (‑)χρῆσθαι, and similar forms are frequently attested for the common verbs πεινάω and διψάω, which in Attic contracted in /ɛ:/.34 This points to the fact that the all too natural ‘return’ to the regular ‑άω inflection of verbs such as χράω, πεινάω and διψάω must have been a widespread phenomenon in Hellenistic and Medieval Greek alike. One piece of evidence is the way these verbs are conjugated in Modern Greek and its dialectal varieties, where πεινάω/πεινώ διψάω/διψώ follow the same conjugation as τιμάω/τιμώ (with e.g. 3rd pers. sing. πεινάει/πεινά).35 Clearly, the correct preservation of the Attic inflection of χράω was already a preoccupation for Atticist lexicographers, as the following entries show: (7) χρῆται Ἀττικοί· χρᾶται Ἕλληνες.

(Moer. χ 5 Hansen)

χρῆται in Attic; χρᾶται in koine Greek.

 33 The same phenomenon concerns verbs such as πεινάω and διψάω, on which see the still very good overview by Kühner/Blass 1892, 139. The motivations behind the Attic inflection are debated: EDG s.v. διψάω and πεινάω does not endorse the view (based on Homeric forms such as διψάων, πεινάων, with /a:/) that these verbs had a stem in /a:/. As concerns χράομαι, it would be more correct to refer to the Attic form as χρῶ, χρῶμαι or χρήω, χρήομαι, given that it never features an /a/ (cf. Moulton 1908, 54). 34 See Reinhold 1898, 84, who focuses on the Apostolic Fathers and apocryphal gospels; Schweizer 1898, 175 on the Hellenistic inscriptions of Ephesus; Moulton 1908, 54 on the New Testament; Mayser 1923, 347 on Ptolemaic papyri; and LSJ s.v. χράω (B) C. Med. χράομαι. 35 A note on ζῆν is necessary here. Although ancient and modern grammars alike tend to treat this verb together with πεινῆν, διψῆν, and ἀποχρῆν, ζῶ was never an ‑άω verb (the rare evidence for an inflection in /a/ in late Byzantine lexicographical sources is due to fallacious analogical reasoning). In non-Attic Greek the verb is ζώω: both ζώω and ζῶ are innovations, given that the IE root is reconstructed as *gwih3‑. In Modern Greek the verb ζω has fallen together with verbs deriving from ancient verbs in ‑έω (like οδηγώ ‘drive’ < ὁδηγέω ‘guide’): its 2nd and 3rd pers. sing. are ζεις, ζει. These forms must continue the Attic forms ζῇς, ζῇ, which in later Greek were pronounced itacistically, and not the by-forms ζώω, ζώεις, ζώει etc. of other varieties of Greek. The ancient spelling ζῇς, ζῇ apparently was preserved in some 19th-century Modern Greek dialectal varieties: see Hatzidakis 1892, 128.

Making the Case for a Linguistic Investigation of Greek Lexicography  253

(8) πεινῆν, διψῆν λέγε, ἀλλὰ μὴ διὰ τοῦ α.

(Phryn. Eclogae 39 Fischer)

Say ‘πεινῆν, διψῆν’ and not (πεινᾶν, διψᾶν) with an /a/.

(7) provides clear proof that the Atticists were at pains to reinforce the correct Attic vocalism against the development of koine Greek, while (8) testifies that the analogical treatment also affected πεινάω and διψάω. Since, on the whole, Byzantine writers continued to comply with the Atticist model, the Synagoge lemma under (6) may simply have been meant to preserve an erudite piece of information concerning the dialectal variety of a Classical author, Herodotus. It is intriguing, however, to think that the point of this lemma was precisely to trace back the roots of a feature of spoken Medieval Greek (infinitives in ‑ᾶν and not in ‑ῆν) to the Classical Age. This intention may have been present already in pre-Byzantine lexica. A puzzling entry is found in Orus’ lexicon, or rather in what has survived in Pseudo-Zonaras: (9) ἀποχρῆν καὶ ἀποχρᾶν· τὸ ἀπαρέμφατον ῥῆμα λέγουσιν ἑκατέρως. Δημοσθένης· ἀποχρῆν οἶμαι τὴν δύναμιν, Λυσίας· Φαινίππῳ δὲ μὴ οἴεσθαι ταῦτ’ ἀποχρᾶν. (Orus fr. 13 Alpers = Pseudo-Zonar. 274 Tittmann) ἀποχρῆν and ἀποχρᾶν: they (i.e. Attic-speakers) say the infinitive in both ways. Demosthenes (4.22): ‘I think that the force is sufficient (ἀποχρῆν)’; Lysias (fr. 288 Carey, modified): ‘Not to think that these things would be sufficient (ἀποχρᾶν) for Phainippos’.36

In the manuscripts of Pseudo-Zonaras’ lexicon the quotation from Lysias contains the infinitive ἀποχρᾶν, which clearly is at odds with standard Attic practice. Alpers (ad Orus fr. 13) keeps this reading, since it is coherent with the lemma, which seems to advise readers that both forms of the infinitive were possible in Attic. However, nothing prevents us from thinking that this is a later interpretation driven by the desire to justify the post-Classical development of the verb and guided by a wrong reading: the fragment is not quoted elsewhere, and Lysias seems to have always employed χρῆν, χρῆσθαι and compounds.37 Another source which may have been interested in tracing a post-Classical usage back to the Classical age is the Antiatticist, which is likely to have transmitted the lemma under (8) to the Synagoge, as the following entry suggests:

 36 Carey 2007 actually has Φιλίππῳ, the lectiο facilior transmitted by part of the manuscript tradition of Pseudo-Zonaras’ lexicon: on this, see Alpers’ apparatus ad Orus fr. 13. 37 This is the reason why Carey 2007 ad Lys. fr. 288 accepts ἀποχρῆν.

254  Olga Tribulato (10) ἀποχρᾶν· οὐκ ἀποχρῆναι. Ἡρόδοτος α΄.

(Antiatt. α 107 Valente)

Herodotus in the first book uses ἀποχρᾶν, not ἀποχρῆναι.

This entry is slightly problematic for two reasons. First, there is no infinitive ἀποχρᾶν in the first book of Herodotus’ Histories. The locus classicus suggested by Valente 2015 ad loc. (Hdt. 1.66.1) has the imperfect ἀπέχρα instead. Since a similar discrepancy is witnessed by Phot. α 2724 Theodoridis, where the cited verbal form is ἀπόχραε, Valente concludes that the Synagoge may have gathered some glosses “from another source rather than from a fuller version of the Antiatt.”, apparently endorsing Theodoridis’ view ad Phot. α 2724 that this gloss may originally have concerned a different verbal form.38 It is not impossible, however, that the original gloss was ἀποχρᾶν and that it is only the number of the book that is wrong: the infinitive is attested in book 3 (138.2). Secondly, the Antiatticist’s ἀποχρῆναι is at odds with the rest of the lexicographical tradition, which always contrasts ἀποχρᾶν with ἀποχρῆν. We may think that ἀποχρῆναι is a corrupted version of the original ἀποχρῆν, or take the whole lemma to be authentic. In this latter scenario, the point of the lemma may have been to contrast Herodotus’ Ionic with the language of another Ionic author, Hippocrates, who does use ἀποχρῆναι (De capitis vulneribus 14). Reconstructing the Antiatticist’s intentions and methodology is always a slippery path, given the customary laconic style of the epitome of this lexicon. If the suggested scenario is correct (although we will never know for sure), one may speculate that the contrast was intended to show that Herodotus did not use such outlandish language as those who used ἀποχρῆναι (because of its athematic conjugation), and hence that he may offer an acceptable Classical parallel for the use of the infinitive ἀποχρᾶν, which was common in the koine. Resorting to Herodotus to justify koine forms is part of the Antiatticist’s ‘alternative’ take on linguistic correctness, as well as proof of the special standing enjoyed by Herodotus in Imperial culture.39 Be that as it may, the Synagoge’s lemma on ἀποχρᾶν (no. (6) above) is a perfect example to illustrate the inherent ambiguity of many lexicographical entries. It may well be that this lemma embodies my criterion (b), namely rare forms that diverged from Byzantine literary usage. At the same time, awareness of the historical evolution of Greek suggests that there may be more to this and that the  38 Valente 2015, 17−18, 19 n. 105. 39 On the Antiatticist’s strategies to justify koine usages, see Cassio 2012; on Herodotus and his use in this lexicon, see Tribulato 2016, 187−191.

Making the Case for a Linguistic Investigation of Greek Lexicography  255

lemma may also have been of interest because it went in the direction in which spoken Greek had developed (see too my comments on Orus’ lemma under (9) above). The fact that the same form, ἀποχρᾶν, was commented upon in the Antiatticist, a text in which the selection of words is strongly oriented towards postClassical usage, further strengthens this suspicion.

 On Curtains, Tents and Courtyards: Or How Contemporary Language May Have Guided the Lexicographers’ Interest The lemma of the Synagoge concerning ἀποχρᾶν in the previous section opens up the question of the interest aroused by the Antiatticist among Byzantine lexicographers, especially the compilers of the Synagoge (broadly understood). This topic still lacks a full investigation. The information that can be gathered from textual history is meagre and can be summarized as follows: (a) The Antiatticist has survived only in the epitome transmitted by cod. Paris. Coisl. gr. 345. This epitome may have been produced deliberately to fit the lexicographical miscellany contained in this codex. (b) The Byzantine use of the Antiatticist is mostly limited to the Synagoge, through which much of its material passes down to Photius and the Suda. Its reception therefore seems to have been restricted to erudite circles of 8th–9th century Constantinople.40 (c) It is possible (though not proven beyond doubt) that in this period the Antiatticist may still have been circulating in other versions. Some lemmas reused in the expansions of the Synagoge, and which do not find a complete parallel in the epitome, may point in this direction (see the discussion on ἀποχρᾶν above).41 We have no idea, however, about the form and date of this other version of the Antiatticist: it may have been a majuscule copy and hence may predate the 9th century.42

 40 Valente 2015, 13. Photius and the Suda also have glosses from the Antiatticist which are not transmitted by any extant expansion of the Synagoge. It is an open question whether they used the Antiatticist independently or (as some wording typical of the Synagoge suggests) had access to a version of the Synagoge unknown to us: see Valente 2015, 25−30. 41 Valente 2015, 16. 42 Valente 2015, 18 and 21.

256  Olga Tribulato (d) In spite of its limited reception and obscure history, the Antiatticist has a far from negligible presence in the Synagoge tradition as a whole: 110 out of its 841 glosses are repeated in some of the versions of the Byzantine lexicon.43 One question that would be worth investigating, in spite of the inevitable degree of speculation that it would entail, is what the compilers of the Synagoge expansions found useful in the Antiatticist. To this purpose, in this last section I consider two lemmas which may have been selected from the Antiatticist because they provided helpful information for writers of ‘Byzantine’ Greek who were also speakers of ‘Medieval’ Greek. The first lemma concerns the verb αἱρετίζω: (11) αἱρετίζειν. ἀντὶ τοῦ αἱρεῖσθαι.

(Antiatt. α 18 Valente)

αἱρετίζειν: instead of αἱρεῖσθαι. (= Com. adesp. fr. 484 K.-A.)

αἱρετίζω is an active synonym of the meaning that αἱρέω acquires in the middle: ‘choose for oneself, prefer’. The short entry of the Antiatticist has not preserved a locus classicus, which Valente however restores on the basis of the following entry in the Synagoge: (12) αἱρετίζειν οἱ περί τι σπουδάζοντες λέγονται. πολὺ τοῦτο παρὰ τοῖς νεωτέροις τῶν κωμικῶν. (Σb α 648 Cunningham) Those who are eager about something are said to ‘αἱρετίζειν’. This word is used very much by playwrights of New Comedy.44

There are no extant traces of the use of αἱρετίζω in New Comedy, but other sources clarify that the verb was associated with post-Classical Greek. It is prominent in the Septuagint, the Aesopic corpus and a vast array of Christian authors, and remains common in Byzantine literature. The entry in the Antiatticist (11), therefore, is consistent with this lexicon’s tendency to defend post-Classical neologisms. In turn, the compiler of one of the expansions of the Synagoge may have considered the lemma interesting because it provided an ancient basis for a common usage in Medieval Greek. αἱρέω, which continued to be used in Byzantine

 43 Valente 2015, 14. His counts are more generous than those in Cunningham 2003 (which may not be complete). See too Latte 1915, 376−377. 44 The interpretamentum is likely to reflect the late-antique connection of αἱρετίζω with αἵρεσις ‘heresy’.

Making the Case for a Linguistic Investigation of Greek Lexicography  257

Greek, was perhaps perceived as typical of the high style, given that lexicographers and scholiasts alike often gloss it with more common verbs such as λαμβάνω, πορθέω and κρίνω (all of which, incidentally, have survived in Modern Greek, while αἱρέω as such has not).45 Another lemma that reached the Synagoge from ancient lexica and that seems to have originally belonged to the Antiatticist is the noun αὐλαία, though in this case the transmission path seems to have been much more complicated and probably affected by the interference between Classical and Christian Greek, as I am going to argue. The ancient meaning of αὐλαία — originally a feminine adjective derived from αὐλή ‘courtyard, open space’ — is ‘curtain’ (i.e. the thing that encloses an open space) and more generally a ‘hanging piece of cloth’.46 The Antiatticist ((13) below) was interested in this word and its meaning, while a parallel entry in Pollux ((14) below) provides a fuller context, as well the locus classicus (Hyp. fr. 139 Jensen): (13) αὐλαίαν· τὸ παραπέτασμα. Ὑπερείδης Κατὰ Πατροκλέους.

(Antiatt. α 135 Valente)

αὐλαίαν: curtain. Hyperides in the Against Patrocles. (14) ἔξεστι δὲ καὶ τὸ παραπέτασμα αὐλαίαν καλεῖν, Ὑπερείδου εἰπόντος ἐν τῷ κατὰ Πατροκλέους οἱ δὲ ἐννέα ἄρχοντες εἱστιῶντο ἐν τῇ στοᾷ, περιφραξάμενοί τι μέρος αὐτῆς αὐλαίᾳ. (Poll. 4.122 Bethe) It is also possible to call the curtain αὐλαία, given that Hyperides in the oration Against Patrocles says ‘the nine archonts took their meals in the stoa, screening off a part of it with a curtain’.

 45 Cf. e.g., among many other cases, schol. Thuc. 8.24.5 Hude: ξυναιρεθήσεσθαι· αἱροῦμαι τὸ προκρίνω καὶ τὸ προτιμῶ· ὅθεν καὶ τὸ συναιροῦμαι (‘ξυναιρεθήσεσθαι: αἱροῦμαι means “to prefer” and “to esteem”; from this also [derives] συναιροῦμαι’). 46 The first attestations of this word go back to the 4th century BC: apart from Menander (fr. 454 K.-A., cf. text under (19) below), see Theophr. Char. 5.9.4 (a piece of embroidered tapestry); Phylarchus FGrHist 2 F 41, quoted by Athen. 12.539c−540a (large pieces of cloth used to enclose wild beasts in a military camp); Polyb. fr. 22 Büttner-Wobst, transmitted by Suda α 4434 Adler (a curtain behind which one can hide and eavesdrop). The best lexicographical collection of works in which αὐλαία is used remains the entry in Hase’s edition of Stephanus’ ThGL. An overview of the semantic family of αὐλή and its adjectives αὔλειος and αὐλαῖος can be found in DELG s.v. αὐλή. Many of the lemmas in Greek lexica reflect contamination and confusion between the two adjectives: see Section 5 below.

258  Olga Tribulato The following entries from the Synagoge, however, narrow down the meaning of αὐλαία by adding the genitive τῆς σκηνῆς to παραπέτασμα: (15) αὐλαία· τὸ τῆς σκηνῆς παραπέτασμα. κέχρηται δὲ αὐτὸ (sic) Ὑπερείδης. (Σ α 1091 Cunningham) αὐλαία: curtain of the tent. Hyperides used it. (16) αὐλαία· τὸ τῆς σκηνῆς παραπέτασμα. κέχρηται δὲ αὐτὸ (sic) Ὑπερείδης ἐν τῷ Κατὰ Πατροκλέους. (Σb α 2405 Cunningham) αὐλαία: curtain of the tent. Hyperides used it in the Against Patrocles.

The presence of the title of Hyperides’ oration in Σb (16), which Σ (15) lacks, is further evidence of the fact that the compiler of Σb had direct access to the Antiatticist, which also preserves the title of the oration.47 The addition of the genitive τῆς σκηνῆς, which continues into the later lexica dependent on the Synagoge,48 seems to have had its first origin in Cyril’s lexicon (5th cent. AD). This is testified by the readings in some of the manuscripts of this work, although given the lack of a complete edition of Cyril’s lexicon the relevant information can be only reconstructed from the parallel lemma in Hesychius’ lexicon: (17) αὐλαία· ἐν αὐλῇ διατρίβουσα AS, ἣ τὸ τῆς σκηνῆς παραπέτασμα vgBr.Σ(A). (Hsch. α 8282 Latte/Cunningham) αὐλαία: she who lives in the house. Or the curtain of a tent.

Faced with the different interpretamenta of αὐλαία in all these lexica, it is natural to ask why the specification τῆς σκηνῆς was added (perhaps originally by Cyril) and kept in Byzantine lexicography, in contrast to earlier Atticist lexica, the extant versions of which it is absent from. The answer lies in the exact meaning of this genitive, which can in principle be interpreted either as ‘of the tent’ (the translation adopted here) or ‘of the stage’.

 47 The lemma in Σb must go back to the expansion named Σ΄ by Cunningham 2003, 687, which was the common source of the two other expansions (Σ΄΄ and Σ΄΄΄) used independently by the Suda (sometimes also Photius) and Σb: see n. 25. 48 Phot. α 3169 Theodoridis; Suda α 4434 Adler; Etym. Magn. α 2087 Lasserre/Livadaras. In Phot. α 3169 Theodoridis the interpretamentum is abbreviated, as is usually the case with glosses contained only in cod. Zavordensis 95 (= z): the traces of Hyperides’ name present on the margin allow Theodoridis to supply the omitted parts from Σb and the Suda.

Making the Case for a Linguistic Investigation of Greek Lexicography  259

At first sight, the latter meaning may seem appropriate to the context of Pollux’s passage (no. (14) above), which deals with the parts of the theatre, and hence it may seem the most likely to be ancient. However, this particular chapter of Pollux does not deal with the σκηνή and, as noted by Poe 2000, 247 “the example of an αὐλαία which is cited refers to a curtain hanging in a colonnade, whose purpose was to screen off a meeting of the archons”.49 If the specification τῆς σκηνῆς therefore goes back to a later source than Pollux or the Antiatticist, its translation as ‘of the stage’ is at odds with everything we know about the meaning of αὐλαία in post-Classical Greek, where it certainly does not identify a stage curtain (which is called παραπέτασμα instead).50 Nor does it seem probable that these Byzantine lexica wished to comment on the Latin word aulaeum ‘stage curtain’, originally of course a Greek calque: they are clearly interested in documenting Ancient Greek usages. All this considered, the interpretation of τῆς σκηνῆς must be ‘of the tent’ and the reason probably lies in the nature of Cyril’s lexicon which, as far as we can ascertain, was the first to introduce it. This work, which was an important source for the Synagoge, is known for adding many Biblical glosses to earlier lexicographical material, in many cases words that the same Cyril used in his religious writings.51 Indeed, the interpretation of αὐλαία as a ‘curtain hanging from a tent’ can only be understood by taking into account the semantics of the word in Old Testament Greek, where it is usually employed in this narrower meaning. This is particularly the case in the chapters of Exodus describing the construction of the sacred tabernacle, as in the passage below: (18) καὶ τὴν σκηνὴν ποιήσεις δέκα αὐλαίας ἐκ βύσσου κεκλωσμένης καὶ ὑακίνθου καὶ πορφύρας καὶ κοκκίνου κεκλωσμένου. (Exodus 26.1) And you shall make the tent with ten curtains from twisted linen and blue and purple and twisted scarlet. (transl. L.J. Perkins, NETS)

 49 Poe 2000, 248 also rules out that this theatrical application of αὐλαία was present in the Onomasticon, since it would have been more appropriate to refer to it in the chapters “concerned with the scenic background” (124−126). 50 It is worth mentioning that αὐλαία is now the standard Modern Greek form for ‘stage curtain’. It is, however, a learned resuscitation (see the abbreviation “λόγ.” = ‘learned’ in LKN s.v.) with no history in Medieval Greek. 51 This was first detected by Wendel: see now Corcella 2017.

260  Olga Tribulato These chapters of Exodus are often quoted in late-antique and Byzantine Christian exegesis, a fact which partly explains the lexicographical interest in the specific meaning of αὐλαία in this context.52 An example is chapter 26 of Cosmas Indicopleustes’ Christian Topography, in which the 6th-century author describes Moses’ construction of the tabernacle and glosses αὐλαία with the Latin calque κορτίνα ‘curtain’, probably a more common word for ‘curtain’ in low-register Greek:53 (19) αὐλαίας τὰς κορτίνας καλεῖ (scil. Μοϋσῆς). οὕτως δὲ καλοῦσι αὐτὰς καὶ οἱ ἔξωθεν Ἀττικοὶ λέγοντες αὐλαίαν τὸ μέγα καὶ ποικίλον παραπέτασμα. Ὑπερίδης ὁ ῥήτωρ ἐν τῷ κατὰ Πατροκλέους λόγῳ· oἱ δὲ ἐννέα ἄρχοντες εἱστιῶντο ἐν τῇ στοᾷ, φραξάμενοι τὸ μέρος αὐτῆς αὐλαίᾳ, ὁμοίως καὶ Μένανδρος· στυππεῖον, ἐλέφαντα, μύρον, οἶνον, αὐλαίαν. (Cosmas Indicopleustes, Topographia Christiana 5.26 Wolska-Conus) [Moses] calls the curtains αὐλαίας: non-Attic speakers too call the curtains in this way, using αὐλαία to refer to a large and many-coloured curtain. Hyperides, the orator, [said] in the Against Patrocles: ‘the nine archonts took their meals in the stoa, screening off a part of it with a curtain’. Similarly Menander [says]: ‘flax, ivory, perfume, wine, curtain’ (fr. 454 K.-A.).

In Cosmas we have an important link between the Atticist tradition, which has handed down αὐλαία as a synonym of παραπέτασμα and illustrated it with a reference from Hyperides, and Biblical exegesis. He makes a connection between the language of the Old Testament and that of those who ‘do not speak standard

 52 Another gloss which has Exodus behind it is Hsch. α 1672 Latte/Cunningham: αἱ αὐλαῖαι· τὸ κατακάλυμμα (‘αὐλαῖαι: the veil’). The Latte/Cunningham edition adds, as a locus classicus, Exodus 37.14, another passage concerned with the building of the sacred tabernacle. However, since Exodus 37.14 does not feature the word κατακάλυμμα but παραπέτασμα instead, it is more probable that the locus classicus behind this gloss is the same as that behind α 8282 discussed above, namely Exodus 26 which does indeed feature κατακάλυμμα together with παραπέτασμα. Cunningham further classifies the entry with the abbreviation “LXX”, which identifies glosses concerned with Biblical passages. It seems to me that both glosses, concerned with the meaning of αὐλαία in Exodus, belong together: they could perhaps both derive from Cyril. Conomis 1985, 344−345 argues that the original gloss had nothing to do with curtains (αὐλαῖαι) and referred to the neuter plural αὔλια (‘country-house, fold, stable, cave’) instead. He also proposes that the original interpretamentum was κατάλυμα ‘lodging’, since cod. Marc. gr. 622, codex unicus for Hesychius, has κατακάλυμα (κατακάλυμμα is a correction introduced by Musurus). I believe that this solution is unnecessary: the context of Exodus 26 and the parallel lemma on αὐλαία in Hesychius prove that the lexicographical interest lied in explaining the special meaning of this word in Biblical Greek. 53 This passage is one of the many παραγραφαί (‘digressions’) in the Christian Topography. These are generally thought to have been added by Cosmas himself: see Wolska-Conus 1968, 67.

Making the Case for a Linguistic Investigation of Greek Lexicography  261

Attic’ (this must be the meaning of the strange expression οἱ ἔξωθεν Ἀττικοί)54 and this fact gives us a clue about why αὐλαία entered Atticist lexicography in the first place. It was probably one of those words which were common in postClassical Greek but had a shaky Classical pedigree and for which Pollux and the Antiatticist could find no better authority than Hyperides. Cosmas must have used an Atticist source concerned with justifying or condemning a post-Classical usage which resorted to quoting Menander instead: his standing in the Atticist canon notoriously is equally shaky. Such detailed lexicographical interest in αὐλαία therefore reflects first and foremost the dichotomy between Classical and post-Classical Greek. At the same time, it is also to be viewed in the context of Greek linguistic history. In Byzantine Greek, the primary meaning of αὐλαία was no longer ‘curtain’ (of whatever kind): a series of semantic shifts and analogical processes had brought it to mean ‘tent’ or ‘courtyard’. The earliest evidence of the semantic shift by which αὐλαία came to be used as a synecdoche for ‘tent’ belong to Isaiah 54 (‘the restoration of Israel’), where αὐλαία features as a variatio for σκηνή: (20) πλάτυνον τὸν τόπον τῆς σκηνῆς σου καὶ τῶν αὐλαιῶν σου κτλ.

(Isaiah 54.2)

Enlarge the site of your tent and of your curtains, etc. (transl. M. Silva, NETS)

Quotations of these lines abound in Christian authors and it is very likely that in some registers of Greek it was normal to employ αὐλαία to say ‘tent’ already in the early Byzantine period. Indeed, Cosmas Indicopleustes’ careful annotations concerning the semantics of the word are already intended to guide readers of the Scriptures who may be confused by these competing meanings. Outside Biblical verbatim quotations αὐλαία for ‘tent’ is first found in the Chronographia attributed to the so-called Theophanes Continuatus. A passage of book 4, the original compilation of which goes back to the late 9th century, describes the setting up of the camp of Emperor Michael III by using σκηναί to refer to the tents in a general way, and then αὐλαία for Michael’s tent in particular:55  54 The expression οἱ ἔξωθεν Ἀττικοί is a hapax, but literal translations such as ‘Attic speakers from abroad’ or ‘non-Attic speakers’ are clearly out of context: cf. also Wolska-Conus 1970, 48 (“les gens du dehors atticistes”). 55 The word is used again in Chronographia 5.236.2 Bekker (= 17.8 Ševčenko), where it appears to have been glossed with κόρτη: see the apparatus in Ševčenko ad loc. For the dating of this composite text, and particularly of the first four books, see Featherstone/Codoñer 2015, 14.

262  Olga Tribulato (21) ἔνθα πως, εἴτε δὴ κατὰ πρόνοιαν εἴτε δὴ καὶ ἄλλως δι’ ἄγνοιαν, τὴν μὲν τοῦ Μιχαὴλ ἐπὶ πεδιάδος χωρίου καὶ ὁμαλοῦ αὐλαίαν ἐκπεταννύουσιν κτλ. (Theophanes Continuatus, Chronographia 4.41.15−16 Featherstone/Codoñer = 4.205.4 Bekker) Somehow, whether by Providence or else through ignorance, they pitched the tent of Michael on a plain and level ground, etc.

Other evidence for this semantic shift comes from Theodore Prodromus (12th century) who alludes to Isaiah in his exhortation for Byzantium to multiply its dominions: (22) ἐξάπλου σου τοὺς σχοινισμούς, γῆ Βυζαντίς, ἐξάπλου, μὴ φείσῃ σου τῶν αὐλαιῶν κατὰ τὴν προφητείαν. (Theodorus Prodromus, Carmina historica 11.131–132 Hörandner) Spread out your dominions, Byzantium: do not spare your tents, as in the prophecy.

Whereas in Isaiah (no. (20)) αὐλαῖαι comes after σκηνή and specifies it further, in Teophanes Continuatus and Theodore Prodromus the word has become ‘tent’ tout court. In this meaning αὐλαία is used by a variety of later Byzantine sources both in passages which variously refer to the Scripture and in others where the context is not religious.56 The picture, however, is further complicated by the fact that at some point in Byzantine linguistic history αὐλαία seems to have been used also as a synonym of αὐλή ‘courtyard’, the noun from which it originally derived. The most authoritative source on this semantic equivalence is Eustathius, who overtly comments on this usage and explains its motivation with a general morphological rule which he devises on the basis of analogy: (23) Τέσσαρες δὲ εὐθεῖαί εἰσι τῆς Ἀθηνᾶς· πρωτότυπος μὲν ἡ Ἀθήνη οἷον ἀθήλη, ὡς εἴρηται· ἀνδρομήκης γὰρ τῆς τοῦ Διὸς ἐξέθορε κεφαλῆς· ἐκ δὲ τούτου κατὰ συστολὴν Ἀθάνα γέγονε, καθὰ παρά τε Σοφοκλεῖ κεῖται καὶ ἑτέροις· ὥσπερ δὲ τὴν ἅμαξαν κατὰ παραγωγὴν ἁμαξαίαν  56 See for instance Michael Italicus’ letter 23 (174.20 Gautier) and the two attestations in Nicephorus Basilakes’ letters from his exile, where αὐλαία serves both as an allusion to the Biblical exodus and as a sarcastic description of the poor state of Nicephorus’ current abode: Epistulae 1.18 Garzya (ὁρῶ τὴν αὐλαίαν καὶ γίνεταί μοι ἀτεχνῶς αἴνιγμα τῷ τοῦ Σαμψὼν ἐοικός, γλυκὺ ἐκ πικράζοντος ‘I see my tent and it is for me a bad riddle, similar to Samson’s famous one: something sweet out of something bitter’) and Epistulae 3.13 Garzya (ἡ γάρ μοι αὐλαία ὑπερφυῶς ἀψινθοφορεῖ, τοιαύτην ἀρωματοφόρον οἰκοῦμεν, ‘My tent is a source of exceeding bitterness: I live in such a tomb!’). For the context of Nicephorus’ exile, see Wirth 1966.

Making the Case for a Linguistic Investigation of Greek Lexicography  263

φασὶ καὶ τὴν αὐλὴν αὐλαίαν καὶ τὴν προνομὴν προνομαίαν […] οὕτω καὶ τὴν Ἀθήνην Ἀθηναίαν φασὶν ἄλλοι τε καὶ ὁ ποιητής. (Eust. Il. 1.32.28−31 van der Valk) The nominative of (the name) Athena has four forms. The original one is Ἀθήνη, which is to say ἀθήλη (‘unsuckled’): for she sprang from Zeus’ head already in an adult’s form. The form Ἀθάνα has derived from this name by vowel mutation: it is attested in Sophocles and other authors. In the same way as they (i.e. Attic-speakers) call the cart (ἅμαξα) ἁμαξαία by derivation, the courtyard (αὐλή) αὐλαία, and the proboscis (προνομή) προνομαία […] so others and Homer also call Athena Ἀθηναία.

Eustathius himself used αὐλαία as ‘courtyard’ in his pamphlet De emendanda vita monachica, where the expression ἔσω αὐλαία describes the messy and dirty inner courtyard of the houses of common people: the danger which awaits the monk who ventures out of his cloister (152.10 Metzler).57 Τhe language of this pamphlet has been described as typical of the Byzantine ‘elaborated style’, rich in neologisms and hapax legomena, and intended for educated readers.58 The fact that in his commentary on the Iliad Eustathius comments on the morphological derivation of αὐλαία from αὐλή may identify his use of αὐλαία ‘courtyard’ as a learned element. Yet the attestations and the history of this meaning suggest that in this particular case Eustathius may be indulging in the lower register. αὐλαία ‘courtyard’ is attested in texts which predate the 12th century and employ a lower register than Eustathius. The first of these texts is the anonymous Life of Saints David, Symeon and George of Mytilene, which van den Gheyn 1899, 210 dated to the end of the 9th or the beginning of the 10th century. αὐλαία occurs in a parable narrating how St. George calmed down a noisy donkey tied in the courtyard which was disturbing the final hour of a sick man: (24) ὁ ὄνος ἐν τῇ αὐλαίᾳ προσδεδεμένος συνεχῶς καὶ ἀτάκτως ὀγκώμενος οὐ μετρίως τὸν νοσοῦντα ἐτάραττεν. (Vita sanctorum Davidis, Symeonis et Georgii Mitylenae 242.22−24 van den Gheyn) The ass, which had been tied in the courtyard for a long time and was exceedingly excited, disturbed the sick man a lot.

 57 Metzler 2006a, 168 does not register any variant in her apparatus: the reading αὐλαία, therefore, is probably authentic. 58 Metzler 2006b, 94−103. αὐλαία is not included among the “seltene Wörter” and “Sonderbedeutungen” by Metzler 2006b, 103.

264  Olga Tribulato Of course, one cannot be absolutely certain that this usage of αὐλαία can be ascribed to the original layer of this hagiographical text, nor that the dative ἐν τῇ αὐλαίᾳ could not be a corruption for ἐν τῇ αὐλῇ. Yet, the fact that ἐν τῇ αὐλαίᾳ may be authentic is suggested by the independent attestation of the hapax προαυλαία in John Cameniates’ The Capture of Thessaloniki (51.2.2 Böhlig), dated to the early 10th century. Both attestations may suggest that the confusion between αὐλαία and αὐλή arose in the lower registers of Byzantine Greek, entered hagiographical texts and chronicles first, and later spread to higher registers. Towards the end of the 10th century αὐλαία ‘courtyard’ is used again by Leo the Deacon in a passage of his History which describes the famous siege of Dorystolon: (25) […] πῦρ δὲ τῇ αὐλῇ ἑκασταχοῦ διὰ τῶν περιβόλων ἐνιέναι προσέταξε. τῆς δὲ πυρκαϊᾶς σφοδρᾶς ἀναῤῥιπισθείσης, καὶ ἐκτεφρούσης θᾶττον τὰ ὑποκείμενα, τῶν δόμων ὑπεξελθόντες οἱ Ῥῶς, ὑπὲρ τοὺς ἑπτακισχιλίους τυγχάνοντες, ἔς τε τὸ ὕπαιθρον τῆς αὐλαίας συσπειραθέντες, ἀμύνεσθαι τοῖς ἐπιοῦσι παρεσκευάζοντο. (Leo Diaconus Historia 137.10−15 Hase)59 [… the emperor] ordered to set fire to the palace on all sides by using fire-balls.60 Since a violent conflagration took place and it quickly reduced the underlying structures to ashes, the Rus’, who happened to be more than seven thousand in number, came out of the buildings and crowded together in the open part of the courtyard, getting ready to fend off those who advanced.61

Leo first describes the way the fortified palace (αὐλή) was set on fire. He then explains how the fire forced the Rus’ to assemble in the open part (ὕπαιθρον) of the courtyard (αὐλαία). These three words are not used to achieve variatio, but to let the reader picture the subsequent effects of emperor John Tzimiskes I’s orders: the seven thousand Rus’ are forced to leave their shelter and crowd in an open and undefended part of the fort. The attestations of αὐλαία in (23−25) above prove not only the semantic shift undertaken by the word, but also another more general fact: the easy contamination between forms derived from αὐλή with the suffix ‑ειος and those showing the suffix ‑αῖος. The hapax προαυλαία in John Cameniates may be a by-form of the more common προαύλιον (‘courtyard, space before a cattle-pen’), which in Byzantine Greek could sometimes take the form προαύλειος (feminine) and which

 59 In his Latin translation, Hase 1828, 137 translated αὐλή with regia and αὐλαία with aula. 60 Hase 1828, 137 translates ‘per propugnancula’. Talbot/Sullivan 2005, 183 n. 42 (followed here) assume that περιβόλων is corrupt for πυριβόλων. 61 Translation adapted from Talbot/Sullivan 2005, 202.

Making the Case for a Linguistic Investigation of Greek Lexicography  265

continues into Modern Greek προαύλιο ‘courtyard, forecourt’ (a learned term according to LKN). It may also be the case that these forms were confused because of their pronunciation in some registers of Byzantine Greek. In learned registers the feminine αὐλεία ‘(outer) door’ kept its original proparoxytone accentuation and could therefore have the same pronunciation as properispomenon adjectives such as βαθεῖα [vaθea]: hence, αὐλαία and αὐλεία could both have been pronounced [avlea].62

. Concluding Remarks on αὐλαία This review of the use of αὐλαία across the centuries places the lexicographical entries pertaining to this word (texts (15−17) above) in a new light. It would be unsatisfactory to think that the Synagoge, Photius or even Hesychius are merely concerned with the meaning of αὐλαία in a Classical author, Hyperides, who was no longer read in Byzantine times.63 On the contrary, as I have argued, these lexicographical entries do not describe the Classical meaning of αὐλαία (i.e. παραπέτασμα: see Pollux and the Antiatticist under (13−14) above), but the specific meaning that it has in some influential Biblical passages: τὸ τῆς σκηνῆς παραπέτασμα, with σκηνή meaning ‘tent’ and not ‘stage’. That in this case the lexicographers’ interest was in semantic variation in Christian Greek is further confirmed by an annotation in the Etymologicum Magnum’s lemma on αὐλαῖαι (α 2087 Lasserre/Livadaras), where the interpretamentum τὰ παραπετάσματα τῆς σκηνῆς is followed by ὡς παρὰ τῷ Θεολόγῳ. This has been identified as a reference to Gregory of Nazianzus’ Oration 42.7 (τῶν αὐλαιῶν μὴ φείσῃ ‘do not spare your tents’), where, however, the theologian is simply referring to Isaiah 54. Hyperides’ name then seems to have been passed down as a relic from one lexicographical tradition that was interested in commenting on a word associated with the koine in relation to a specific Attic custom (that of separating the stoa with a

 62 This is likely to have been different in vernacular varieties, where no accent shift is observed in the feminine: hence, αὔλεια would differ from αὐλαία. Apart from the different accent, proparoxytone forms such as αὔλεια or πρoαύλιον (see above) yield [i] pronunciation of the diphthong. 63 The question has been complicated by the discovery of fragments of the Πρὸς Τίμανδρον and of the Πρὸς Διώνδαν in the so-called Archimedes’ palimpsest (Tchernetska 2005; Ucciardello 2009). In his up-to-date overview of Hyperides’ post-Classical transmission Ucciardello 2012a, 305 concludes that he was “una rarità a Bisanzio” and that he probably circulated only in anthologies destined for the classroom. He also rules out the possibility that quotations from Hyperides in Byzantine rhetoric and lexicography testify to the direct use of Hyperides’ text in Byzantium (Ucciardello 2012a, 321).

266  Olga Tribulato curtain), to another lexicographical tradition interested in the meaning that the word had in the completely different context of Old Testament Greek. The relevance of these words for Byzantine lexicographers can only be fully appreciated if they are placed in the historical and linguistic context that preserved them. In this particular case Byzantine scholars were not only torn between the different linguistic models set by the Classics and the Scripture, but also between older meanings and the contemporary linguistic reality, where αὐλαία meant ‘tent’ or ‘courtyard’, but apparently not ‘curtain’. This situation is clearly exemplified by the language used by Nicetas Choniates in his Chronicle, where αὐλαία sometimes identifies a tent (e.g. in 41.15 and 197.29 van Dieten) and sometimes a courtyard (as in 236.10 and 237.1 van Dieten). It is probable that by the 8th century to use αὐλαία for ‘curtain’ was an erudite touch and it is to be concluded that the ancient meaning had died out.

 Conclusions The case-studies presented in this paper all contribute to illustrate how, through different linguistic situations, the application of a historical linguistic approach allows us to analyze the Byzantine lexicographical tradition in a more nuanced way and to prove that its re-use of older material is not simply uncritical recycling, but is motivated by the linguistic situation of the time. Many of the lemmas contained in Byzantine lexica can be interpreted as instructions to educated speakers and writers, who lived in a situation of linguistic diglossia, on how to fine-tune their use of language in the light of a linguistic evolution which may have distanced it from both the Classics and the Scripture. If brought up to the level of systematic studies, piecemeal lexical investigations such as those provided in this paper are likely to achieve a more fine-grained appreciation of Byzantine Greek which, going beyond the classic diglossic dichotomy between high and low language, takes into account the high level of register mixing in Byzantine texts and does justice to the language of the Byzantines: a live variety, and not simply a sterile reproduction of earlier linguistic stages.

Making the Case for a Linguistic Investigation of Greek Lexicography  267

Bibliography Alpers, K. (1990), “Griechische Lexikographie in Antike und Mittelalter: Dargestellt an ausgewählten Beispielen”, in: H.A. Koch (ed.), Welt und Information: Wissen und Wissenvermittlung in Geschichte und Gegenwart, Stuttgart, 14−38. Alpers, K. (2001), “Lexikographie, griechische”, in: G. Ueding (ed.), Historisches Wörterbuch der Rhetorik, vol. 5, Tübingen, 194−210. Bakker, E.G./de Jong, I.J.F./van Wees, H. (eds.) (2002), Brill’s Companion to Herodotus, Leiden/ Boston/Köln. Beck, H.-G. (1971), Geschichte der byzantinischen Volksliteratur, München. Bethe, E. (1900), Pollucis Onomasticon, vol. 1, Leipzig. Böhlig, G. (1956), Untersuchungen zum rhetorischen Sprachgebrauch der Byzantiner, mit besonderer Berücksichtigung der Schriften des Michael Psellos, Berlin. Browning, R. (1992), “The Byzantines and Homer”, in: R. Lamberton/J.J. Keaney (eds.), Homer’s Ancient Readers, Princeton, 134−148. Carey, C. (2007), Lysiae Orationes cum fragmentis, Oxonii. Cassio, A.C. (2012), “Intimations of Koine in Sicilian Doric”, in: O. Tribulato (ed.), Language and Linguistic Contact in Ancient Sicily, Cambridge, 251−264. Cassio, A.C. (2014), “Lessico ‘moderno’ nei testi greci antichi: Storie di continuità e discontinuità”, in: C. Carpinato/O. Tribulato (eds.), Storia e storie della lingua greca, Venezia, 35−48. Cohn, L. (1900), “Griechische Lexikographie”, in: K. Brugmann (ed.), Griechische Grammatik, 3rd edn., München, 577−616. Conomis, N.C. (1985), “Lexicographical Notes”, in: BZ 78, 344−349. Corcella, A. (2017), “Le trappole di Cirillo: Hesych. ε 4016 L., π 12 H.”, in: Eikasmos 28, 225−234. Cunningham, I.C. (2003), Synagoge: Συναγωγὴ λέξεων χρησίμων. Texts of the Original Version and of Manuscript B, Berlin/New York. Cuomo, A.M. (2017), “Medieval Textbooks as a Major Source of Historical Sociolinguistic Studies of (High-Register) Medieval Greek”, in: Open Linguistics 3, 442−455. de Leeuw, M. (2000), “Der Coislinianus 345 im Kloster Megisti Lavra (Athos)”, in: ZPE 131, 58−64. Dewald, C./Marincola, J. (eds.) (2006), The Cambridge Companion to Herodotus, Cambridge. Dover, K.J. (1968), Aristophanes: Clouds. Edited with Introduction and Commentary, Oxford. Erbse, H. (1950), Untersuchungen zu den attizistischen Lexika, Berlin. Featherstone, M./Codoñer, J.S. (2015), Chronographiae quae Theophanis Continuati nomine fertur: Libri I−IV, Boston/Berlin. Hase, C.B. (1828), Leoni Diaconi Caloënsis Historiae libri decem, Bonnae. Hatzidakis, G.N. (1892), Einleitung in die neugriechische Grammatik, Leipzig. Hinterberger, M. (ed.) (2014), The Language of Byzantine Learned Literature, Turnhout. Hörandner, W./Trapp, E. (eds.) (1991), Lexicographica Byzantina, Wien. Horrocks, G.C. (2014), “High-Register Medieval Greek”, in: C. Carpinato/O. Tribulato (eds.), Storia e storie della lingua greca, Venezia, 49−72. Hunger, H. (1969−1970), “On the Imitation (ΜΙΜΗΣΙΣ) of Antiquity in Byzantine Literature”, in: Dumbarton Oaks Papers 23−24, 15−38. Hunger, H. (1978), Die hochsprachliche profane Literatur der Byzantiner, München.

268  Olga Tribulato Hunger, H. (1991), “Was nicht in der Suda steht, oder: Was konnte sich der gebildete Byzantiner des 10./11. Jahrhunderts von einem «Konversationslexikon» erwarten?”, in: W. Hörandner/E. Trapp (eds.), Lexicographica Byzantina, Wien, 137−153. Kaldellis, A. (2007), Hellenism in Byzantium: The Transformations of Greek Identity and the Reception of the Classical Tradition, Cambridge/New York. Kaldellis, A. (2012), “The Byzantine Role in the Making of the Corpus of Classical Greek Historiography: A Preliminary Investigation”, in: JHS 132, 71−85. Kühner, R./Blass, F. (1892), Ausführliche Grammatik der griechischen Sprache, vol. 1.2: Elementar- und Formenlehre, 3rd edn., Hannover/Leipzig. Latte, K. (1915), “Zur Zeitbestimmung des Antiatticista”, in: Hermes 50, 373−394. Lemerle, P. (1971), Le premier Humanisme byzantine: Notes et remarques sur enseignement et culture à Byzance des origines au Xe siècle, Paris. Manolessou, I. (2014), “Learned Byzantine Literature and Modern Linguistics”, in: M. Hinterberger (ed.), The Language of Byzantine Learned Literature, Turnhout, 13−33. Matthaios, S. (2006), “Suda: The Characters and Dynamics of an Encyclopedic Byzantine Dictionary”, in: J.N. Kazazis (ed.), Η λεξικογράφηση του ελληνικού πολιτισμού αρχαίου, μεσαιωνικού και νέου: Tα σύγχρονα εγκυκλοπαιδικά λεξικά, Thessaloniki, 1−40. Matthaios, S. (2010), “Lexikographen über die Schulter geschaut: Tradition und Traditionsbruch in der griechischen Lexikographie”, in: M. Horster/C. Reitz (eds.), Condensing Texts – Condensed Text, Stuttgart, 165−207. Matthaios, S. (2013), “Pollux’ Onomastikon im Kontext der attizistischen Lexikographie: Gruppen «anonymer Sprecher» und ihre Stellung in der Sprachgeschichte und Stilistik”, in: C. Mauduit (ed.), L’Onomasticon de Pollux: Aspects culturels, rhétoriques et lexicographiques, Lyon, 67−140. Matthaios, S. (2015), “Zur Typologie des Publikums in der Zweiten Sophistik nach dem Zeugnis der Attizisten: ‚Zeitgenössische‘ Sprechergruppen im Onomastikon des Pollux”, in: M. Tziatzi/M. Billerbeck/F. Montanari/K. Tsantsanoglou (eds.), Lemmata: Beiträge zum Gedenken an Christos Theodoridis, Berlin/Boston, 286−313. Mayser, E. (1923), Grammatik der griechischen Papyri aus der Ptolemäerzeit, vol. 1: Laut- und Wortlehre, 2nd edn., Berlin/Leipzig. Meillet, A. (1930), Aperçu d’une histoire de la langue grecque, 3rd edn., Paris. Metzler, K. (2006a), Eustathii Thessalonicensis De emendanda vita monachica, Berlin/New York. Metzler, K. (2006b), Eustathios von Thessalonike und das Mönchtum: Untersuchungen und Kommentar zur Schrift De emendanda vita monachica, Berlin/New York. Monaco, C. (2015), La lingua di Menandro nell’interpretazione dei lessicografi greci, diss. Roma. Moulton, J.H. (1908), A Grammar of New Testament Greek, vol. 1: Prolegomena, 3rd edn., Edinburgh. Poe, J.P. (2000), “Pollux and the Aulaia”, in: Hermes 128, 247–250. Pontani, F. (2015), “Scholarship in the Byzantine Empire (529−1453)”, in: F. Montanari/S. Matthaios/A. Rengakos (eds.), Brill’s Companion to Ancient Greek Scholarship, 2 vols., Leiden/New York, 297−455. Priestley, J./Zali, V. (eds.) (2016), Brill’s Companion to the Reception of Herodotus in Antiquity and Beyond, Leiden/Boston. Pritchett, W.K. (1956), “The Attic Stelai: Part II”, Hesperia 25, 178−317.

Making the Case for a Linguistic Investigation of Greek Lexicography  269

Reinhold, H. (1898), De Graecitate patrum apostolicorum librumque apocryphorum Novi Testamenti quaestiones grammaticae, Halis Saxonum. Reitzenstein, R. (1897), Geschichte der griechischen Etymologika, Leipzig. Reitzenstein, R. (1907), Der Anfang des Lexikons des Photios, Leipzig/Berlin. Robins, R.H. (1993), The Byzantine Grammarians: Their Place in History, Berlin/New York. Rodenwaldt, G. (1927), “Σκίμπους”, in: RE 3A.1, 527−529. Rollo, A. (2008), “Greco ‘medievale’ e greco ‘bizantino’”, in: AION(ling.) 30, 429−473. Schweizer, E. (1898), Grammatik der pergamenischen Inschriften: Beiträge zur Laut- und Flexionslehre der gemeingriechischen Sprache, Berlin. Ševčenko, I. (1981), “Levels of Style in Byzantine Prose”, in: JÖB 31, 289−312. Talbot, A.-M./Sullivan, D.F. (2005), The History of Leon the Deacon: Byzantine Military Expansion in the Tenth Century, Washington D.C. Tchernetska, N. (2005), “New Fragments of Hyperides from the Archimedes Palimpsest”, in: ZPE 154, 1−6. Tolkiehn, J. (1925), “Lexikographie”, in: RE 12.2, 2432−2482. Tosi, R. (2015), “Typology of Lexicographical Works”, in: F. Montanari/S. Matthaios/ A. Rengakos (eds.), Brill’s Companion to Ancient Greek Scholarship, 2 vols., Leiden/New York, 622−636. Toufexis, N. (2008), “Diglossia and Register Variation in Medieval Greek”, in: Byzantine and Modern Greek Studies 32, 203−217. Trapp, E. (1993), “Learned and Vernacular Literature in Byzantium: Dichotomy or Symbiosis?”, in: Dumbarton Oaks Papers 47, 115−129. Trapp, E. et al. (eds.) (1988), Studien zur byzantinischen Lexikographie, Wien. Trapp, E./Schönauer, S. (eds.) (2007), Lexicologica Byzantina. Beiträge zum Kolloquium zur byzantinischen Lexikographie (Bonn, 13.–15. Juli 2007), Bonn. Tribulato, O. (2014), “‘Not even Menander Would Use this Word!’: Perceptions of Menander’s Language in Greek Lexicography”, in: A. Sommerstein (ed.), Menander in Contexts, New York/London, 199−214. Tribulato, O. (2016), “Herodotus’ Reception in Ancient Greek Lexicography and Grammar: From the Hellenistic to the Imperial Age”, in: J. Priestley/V. Zali (eds.), Brill’s Companion to the Reception of Herodotus in Antiquity and Beyond, Leiden/Boston, 169−192. Tribulato, O. (2018), “Le epistole prefatorie dell’Onomasticon di Polluce: frammenti di un discorso autoriale”, in: Lexis 36, 247–283. Ucciardello, G. (2006), “Esegesi linguistica, glosse ed interpretamenta tra hypomnemata e lessici: Materiali e spunti di riflessione”, in: G. Avezzù/P. Scattolin (eds.), I classici greci e i loro commentatori: Dai papiri ai marginalia rinascimentali. Atti del convegno, Rovereto (20 ottobre 2006), Rovereto, 35−83. Ucciardello, G. (2009), “Hyperides in the Archimedes Palimpsest: Palaeography and Textual Transmission”, in: BICS 52, 229−252. Ucciardello, G. (2012a), “Iperide tra età tardoantica e medioevo: I percorsi del testo nella tradizione retorica”, in: F.G. Hérnandez Muñoz (ed.), La tradición y la transmisión de los oradores y rétores griegos – Tradition and Transmission of Greek Orators and Rhetors, Berlin, 303−330. Ucciardello, G. (2012b), Hypomnemata papiracei e lessicografia: Tra Alessandria e Bizanzio, Messina. Valente, S. (2008), “Una miscellanea lessicografica del X secolo: Il Par. Coisl. 345”, in: Segno e testo 6, 151−178.

270  Olga Tribulato Valente, S. (2012), I lessici a Platone di Timeo Sofista e Pseudo-Didimo: Introduzione ed edizione critica, Berlin/Boston. Valente, S. (2015), The Antiatticist: Introduction and Critical Edition, Berlin/Boston. van den Gheyn, J. (1899), “Acta Graeca SS. Davidis, Symeonis et Georgii, Mitylenae in insula Lesbo”, in: Analecta Bollandiana 18, 209−259. Vessella, C. (2010), “Reconstructing Phonologies of Dead Languages: The Case of Late Greek ”, in: Rivista di Studi Orientali 84, 257–271. Vessella, C. (2018), Sophisticated Speakers: Atticistic Pronunciation in the Atticist Lexica, Berlin/Boston. Wirth, P. (1966), “Wohin ward Nikephoros Basilakes verbannt?”, in: ByzForsch 1, 289−392. Wolska-Conus, W. (1968), Cosmas Indicopleustès: Topographie Chrétienne, vol. 1: Livres I−IV, Paris. Wolska-Conus, W. (1970), Cosmas Indicopleustès: Topographie Chrétienne, vol. 2: Livre V, Paris.

List of Contributors Enrico Cerroni teaches Greek and Latin in secondary schools with a Classical studies curriculum and since the academic year 2018/2019 has been teaching Greek Philology as Visiting Professor at the Pontifical Biblical Institute (Rome). He has authored studies on the reception of Greek lyric poetry in the 19th century (e.g. the appendix I lirici greci nelle traduzioni italiane prima di Quasimodo, in C. Di Noi (ed.), Lirici greci, Salerno 2015). His research interests include koine (PhD dissertation on the lexicon of the Second Book of the Maccabees), semantic shifts and stylistics. Federico Favi is a British Academy Postdoctoral Fellow in the Faculty of Classics, University of Oxford, and a Junior Research Fellow of Oriel College. He is the author of Fliaci: Testimonianze e frammenti (Heidelberg 2017), and is currently revising his PhD dissertation (Epicarmo e pseudo-Epicarmo) for publication. He has published articles on ancient theatre, especially Greek comedy (Epicharmus, Aristophanes, Menander, and fragmentary Athenian comic poets), Greek dialectology, and Greek epigraphy. Sara Kaczko is Lecturer in Classics at the Department of Antiquities of the University of Rome, La Sapienza. She is the author of several papers on Greek literary languages, dialects, and epigraphy, along with the monograph Archaic and Classical Attic Dedicatory Epigrams: An Epigraphic, Literary, and Linguistic Commentary (Berlin/Boston 2016), which has been awarded the “Prix d’épigraphie for the best work in Greek Epigraphy” by the Association International d’Épigraphie Grecque et Latine (AIEGL). She is currently a member of the Humanities in the European Research Area Joint Research Project 3 (“Uses of the Past”) “Multilingualism and Minority Languages in Ancient Europe — MUMIL-EU”. Enzo Passa teaches Classics at secondary school in Italy. His research mainly focuses on Archaic and Classical hexameter poetry, on which he has published several contributions. He is the author of Parmenide: Tradizione del testo e questioni di lingua (Roma 2009), as well as of the chapters L’epica, L’elegia e l’epigramma su pietra, and Eronda for the handbook Storia delle lingue letterarie greche edited by A.C. Cassio (Milano 2016, 2nd edn.). Lucia Prauscello is Senior Research Fellow at All Souls College, Oxford. She is the author of Singing Alexandria: Music between Practice and Textual Transmission (Leiden 2006) and of Performing Citizenship in Plato’s Laws (Cambridge 2014). She has published articles on various aspects of Ancient Greek literature and culture. She is currently working on an edition of Corinna together with Giovanbattista D’Alessio. Matilde Serangeli is a Research Associate at the Seminar für Indogermanistik of the FriedrichSchiller Universität Jena. As a Classical philologist and a comparative linguist, her work primarily focuses on Ancient Greek and the Anatolian languages from a synchronic and diachronic perspective, as well as on the reconstruction of the prehistorical roots of their lexicon. She has devoted to these themes a series of papers published in international journals and collective books as well as her PhD thesis, Sprachkontakt im alten Anatolien: Das Lykische aus

272  List of Contributors synchroner und diachroner Perspektive, 2018 (def. 2015), USB Universität zu Köln, which is currently being re-worked for a major publication. Olga Tribulato is Associate Professor at Ca’ Foscari University of Venice. She is the author of Ancient Greek Verb-Initial Compounds: Their Diachronic Development within the Greek Compound System (Berlin/Boston 2015), the editor of Language and Linguistic Contact in Ancient Sicily (Cambridge 2012) and has contributed to Storia delle lingue letterarie greche edited by A.C. Cassio (Milano 2016, 2nd edn.). She has published articles on Greek dialects and literary languages, epigraphy, ancient bilingualism and Greek theories of language. She is currently working on Atticist lexicography and its transmission. Alessandro Vatri is Departmental Lecturer in Classical and Comparative Philology at the University of Oxford and Junior Research Fellow of Wolfson College, Oxford. He is the author of Orality and Performance in Classical Attic Prose: A Linguistic Approach (Oxford 2017). He has also published a number of articles on Attic oratory, ancient rhetoric, ancient literary criticism, and Greek linguistics. Andreas Willi is Diebold Professor of Comparative Philology at the University of Oxford. He is the author of The Languages of Aristophanes: Aspects of Linguistic Variation in Classical Attic Greek (Oxford 2003), Sikelismos: Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.–5. Jh. v. Chr.) (Basel 2008), and Origins of the Greek Verb (Cambridge 2018), and he has edited, among other things, The Language of Greek Comedy (Oxford 2002), Laws and Rules in Indo-European (Oxford 2012, with P. Probert), and Formes et fonctions des langues littéraires en Grèce ancienne (Genève 2019). He is also one of the editors of Glotta: Zeitschrift für griechische und lateinische Sprache and has written numerous articles on Greek, Latin, and Indo-European historical and comparative grammar, as well as on the language-literature and language-culture interfaces in the ancient world.

Index of Notable Words Greek ἀβλεμής/ἀβλεμέως 43, 48–52, 57 n. 32 ἄγνυμαι/ἔαγε 87 αἰπεινός/αἶπος/αἰπύς 198, 203, 206, 209–210, 216 αἱρετίζω 256 ἀκαμαντορόᾱ 117–119 ἀλάoμαι/ἀλάλημαι 74–75, 77, 80 n. 32, 93 n. 62, 97 ἀλγηδών 223–224, 228–229, 233, 235– 236 ἄλγημα 219, 223–225, 228–229, 233– 235 ἄλγησις 223 ἄλγος 219–237 ἀμενηνός 54–55 ἁμός 20 n. 3, 21 ἀμπελοβάμ[ονα 113 ἀπόκτησις 126, 138–139 ἀποχράω 251 ἀππασάμενος 6, 125–131, 137, 140 ἀραρίσκω/ἤραρον/ἀρηρώς/ἀρήρει 88– 89, 94 n. 63 ἀργυ[ρο]δίνᾱ 116–119 ἀργυφέος/ἄργυφος 209–210, 213–214 αὐλαία 257–266 αὐλή 257–258, 262–264 αὑτοῦ/ἑαυτοῦ 19, 23, 37–39 βάλλω/βάλλομαι 43, 57, 61–63, 65, 71, 88, 94 Table 1 βεβίηκεν 101 n. 77 βέλεμνον/βέλεμνα 43, 61–63, 65, 67 βιβρώσκω/βεβρωκώς 101 βλεμεαίνω 5, 43–67 βλέπω 43, 48 n. 6, 50, 56–57, 67 βούκρᾱνα 114–115 γίγνομαι/γέγονα 82, 93 δαῆναι/δεδάηκε 85 δαίω/δέδηα 84–85, 99 δέρκομαι/δέδορκα 91–93 (δια-)φθείρομαι/διέφθορα 87 https://doi.org/10.1515/9783110621747-014

δίκρᾱνοι 114 διπλόος/διπλός 107–112 διψάω 252–253 ἐγείρω/ἤγρετο/ἐγρήγορα 77–78, 84 ἔγκτησις 129 n. 15, 130–133, 136 ἔδω/ἐδηδώς 94 Table 1, 103–104 ἑέ 19–39 ἔμπασις/ἔππασις 125, 129–133, 140, 142, 144 ἑοῖ 19–39 ἔοικα/ἤϊκτο 77–79, 99 ἑός 21–22, 25–26, 28, 31 ἔρδω/ἔοργα 103 θείνω 88 ἱλάσκομαι/ἵληθι 92 καταδαιμοινάω 245 (κατ-)ἐρείπω/κατερήριπε/(κατ-)ἐρέριπτο 86, 95 Table 1 κεύθω/κέκευθα 75, 77, 100–101, 103 κηλίς 246–248 κράββατος 248–250 (ἀπο-)κτείνω/ἀπέκτονα 88, 103 n. 81 Λᾱτοίδᾱ 118 n. 35 λείπω/λέλοιπα 95 Table 1, 100–103 λύπη 219–222, 224, 229–232, 235 μείρομαι/ἔμμορε 78, 92–93 μενεαίνω 43, 51–54, 56, 60, 64 Μοῖσα 165 n. 62, 167, 169 n. 83, 171 ὀδύνη 221, 224, 226, 229, 232 οἶδα 73, 77, 103 ὄλλυμαι/ὤλεσα/ὄλωλα 86 ὄπωπα 76–77, 103 ὄρομαι/ὄρωρε/ὄρωρει 84 with n. 41, 91 πάθος 206, 221–223 παραπέτασμα 257–260, 265

274  Index of Notable Words πείθω/πέποιθα 85 πεινάω 252–253 (περι-)τρέφομαι/περὶ ... τέτροφε 83–84 with n. 40 πήγνυμι/ἐπάγην/πέπηγα 81, 88–89 with n. 51, 91, 100 πόνος 219, 222, 224 n. 15, 225–226, 229–232, 235 προβούλομαι/προβέβουλα 92 πτεροβάμοσι 113 σκηνή 259, 261–262, 265 σπίλος 246–248 συνάορος/συνήορος 114–115 ταπεινός 6, 197–216 ταππάματα 125, 129–132, 136–140 ταράσσω/τετρήχει 84, 99 τεΐν 19–21, 27–31, 36 Τέμπυρα 197–198, 207–208, 210–212 τεοῖο 19–20, 27–32, 35–36 τετραβήμονα 113 τεύχω/τέτυγμαι 78 n. 31, 88–90, 95 Table 1 τιτύσκομαι 90 φθίνω/ἔφθιεν/ἔφθιτο 87, 94 n. 63 χολόoμαι/χολώθη/κεχόλωται 97, 98 with n. 70 χράω 251–252

Hittite dampu-/dampupi- 198, 206, 211–216 pāsk- 89 with n. 51 Latin algus 220 n. 2 pangō 89 with n. 51 Mycenaean e-qi-ti-wo-e 87, 94 n. 63 te-tu-ko-wo-a 90–91, 94 n. 63 Old Church Slavonic tъkati 90 Old Russian tupъ 198, 211, 215–216 PIE *sew- 19–20, 21 n. 15, 23, 25–26 with n. 31, 29–30, 36 *sw- 21, 24, 26, 28–29, 32, 38 *tew- 22 n. 9, 27 with n. 31, 30–32, 36 Vedic jāgā́ra 78, 84 riréca 102 riricé 102 vavárta 78

Index of Subjects A ἀ- intensivum (ἐπιτατικόν) 50 ἀ- privativum (στερητικόν) 50 abstract nouns in -μα 60, 62, 139–140, 226 Aeolic 159–160, 167–170 with n. 83, 171 – feminine forms in -οισα 149, 159–163, 164–165, 167–170, 171 agentive/agentivity 75 with n. 13, 80 n. 29, 81 n. 33, 88, 94 with n. 64, 97, 102, 103 with n. 81 Ajax 44–46 Alcman 65 n. 54 Alexandrian Poetics 120–121 analogy 56, 60 Antiatticist 242, 243, 253–254 with n. 39, 255–257 aorist 72–74, 77, 81–91, 93 n. 62, 95, 97 n. 67, 99, 100, 103 archaic epics – latest phases 21–23, 35, 39 n. 73 argument – argument structure (see also ‘event participants’) 80, 81 n. 33, 82 n. 35, 85 n. 43, 90–91, 96, 98, 101–103 Aristotle – prose rhythm 175, 186–188 Asianism – prose rhythm 190–191 Attic dialect 246–253 Atticism 241–266 Avestan 53, 58 B Bianor 111–112 with ns. 16–17 Byzantine Greek 244, 246, 250, 256, 261, 263, 265, 266 C Caland system 206, 209, 209, 214 n. 28, 216 Callimachus 117 – “mixed dialect” 120–121 Choeroboscus https://doi.org/10.1515/9783110621747-015

– commentary on Hephaestion 149, 158– 161 with n. 46, 164 Choral lyric poetry 108–109, 112–115 with n. 29, 118–119 with n. 35 cod. Paris. Coisl. gr. 345 243 with n. 7, 249 with n. 25, 255 compounds 52, 54–56 – compounds in -βᾱμων/-βημων 113–115 consonantal *w- 24–26, 28, 32, 36 n. 63 contraction /-ĕ(h)ă/ > /-ɛ:/ in Doric 113 with n. 19 Cyril’s lexicon 248 n. 27, 256, 257 D dative ending in -ιν 21, 24 n. 20, 27–29 derivative suffix *-no- 198, 206, 208, 216 derivative suffix -(u)ro- 207–209 Derveni Papyrus 118 Dionysius of Halicarnassus – rhythm 189–191 Doric /a:/ in hexametric texts 114–115 Doric dialect 24, 27 with n. 31, 29, 30 with n. 40, 32 with n. 45, 33, 107–121, 160 n. 53, 168 E elision 109, 117 with n. 32 Empedocles 108–115, 119 enoplios nomos 156–157 Epicharmus 149–172 – Ἐπίνικος/Ἐπινίκιος 157 n. 43 – linguistic characterization 164–166 with n. 69, 170–171 – Μεγαρίς 158–159 with n. 45 – Πίθων 155–156 – Χορεύοντες 157 n. 43 epicisms 220 n. 4, 222, 224–225, 227, 233 n. 36, 235 Eusebius 236 event participants (see also ‘argument structure’) – Agent 80, 82 n. 35, 84 n. 39, 87, 89, 100–101

276  Index of Subjects – Patient 80–81, 82 with n. 35, 88, 100, 102 G genealogy 125–127, 141 genitives – in -ᾱ(o)/-ε͜ω 116–119 – of the personal pronoun 23 with n. 15, 28, 31–38 – of the possessive adjective 28 n. 33, 33 with ns. 52–53, 34–36 Getty Hexameters 113–115 with n. 25, 117–118 Gortyn Code 111 with ns. 11 and 14 H hapax – in Aristophanes 227 n. 20 – in Herodotus 219, 224–226 Hector 43–44, 46, 57, 61, 66 Hellenistic literary prose 219, 226, 229– 233 Hephaestion 158–161 with ns. 46 and 57, 163 Hephaestus 45–46 Heraclitus 181, 188–189 – rhythm 178–180, 185–186, 188–189 Hermogenes – rhythm 189, 191 n. 99 Herodotus 51, 176, 182, 185, 224–225, 250–251, 253–254 Hesiod 51, 115, 127 n. 8, 151 hexameter 55, 60, 62, 64, 67, 107–109, 112–121 Hiero 150 with n. 3, 156–157 with ns. 42– 43 Hittite 53, 58, 198, 211, 212–214 Homer 42–67, 107–111, 113–118 – Attic features 23, 33 n. 51, 38–39 with n. 73 – Boeotian features 20–21, 24 n. 19, 27– 28, 30 – Doric features 21, 24 with n. 19, 27 n. 31, 29–30, 33–34 – Euboean features 20 with n. 3 – Ionic features 24, 33 n. 51, 37–39 – personal pronouns 20–39

– West Greek features 19–21, 24 with n. 19, 27–29, 31, 34, 36, 39 Homeric formulae 43–47, 61–67, 111 Homeric Greek 43–67, 107–112, 116–117 with n. 32 Homeric non-formulaic features 19–21 Homeric scholarship 46–52 Hyrieus 125 n. 2, 127–128 with n. 8, 140– 141 I Iliad 43–47, 61–67 internal gemination (/-pp-/ before a vowel in Boeotian) 128, 129–130 with n. 23, 131 n. 28, 132, 137 n. 47, 140, 142 Ionic dialect 53–54, 219, 224, 226, 234– 235, 250, 254 Ionic prose 175–181, 188–189, 191–192 K Kastoreion 157 L lexicography 241–266 M medical language 219–220, 221 with n. 6, 226–227 with n. 20, 228 n. 24, 234 n. 38 Medieval Greek 243 n. 4, 244–245, 252, 253, 256, 259 n. 50 N nouns in -σις 138–140, 223 O object – affected/non-affected 79, 82 with n. 36, 92, 101, 102, 103, 104 Odysseus 46 oracle/oracular 127, 128, 141, 142, 143 Orion 125–129, 140–142 P pain lexicon 219–237 Panyassis of Halicarnassus 48–52, 115– 119

Index of Subjects  277

Parmenides 107 n. 1, 108–109, 113–114 with n. 26 Peisandros 51 perfect 71–105 Pherecydes of Syrus 181, 188–189 – rhythm 176–177 Phrynichus (Atticista) 242, 245, 247–248 poetic terms 219–220, 224–225 with n. 16, 227–228, 233 with n. 36, 236– 237 with n. 49 predicate – complex predicate (result) 96–99 – simple predicate (manner) 84 n. 39, 95–99 present 72–75, 77–78, 81–95, 97–101, 103 prose rhythm 175–192 Proto-Greek 56–57, 59–60 Proto-Indo-European 58–59 Pseudo-Longinus 48 R River epithets 118–119 S semantic shift 198–199, 204–205, 207, 210, 212, 214–216, 237, 261–262, 264 semantics – lexical semantics 95–96 Septuagint Greek 230–231 s-stems 52, 59, 60, 62, 197–198, 204, 206–207, 209–210, 216 state – change of state 73 n. 7, 81, 90, 94, 96, 98, 100–101 – individual-level state 97–98 – resulting state 71, 77, 79–83, 88–89, 99–100, 104

– stage-level state 97–98 ‘strong stem’ in pronouns 27–29 with n. 31, 30–31 with n. 40, 36 subject – agentive/inagentive 75, 80 n. 29, 81 n. 33, 88, 94, 97, 102, 103 suffix -(u)pi- 213 suffix -(υ)φο- 213 synizesis 55, 109 Syracusan dialect 160, 166 n. 69, 168– 170 T tomós type 198, 211, 214 n. 28, 215, 216 transitivity – intransitive 75, 77, 81–87, 89, 93, 99– 100, 103–105 – transitive 71–72, 73 n. 7, 75, 77–79 with n. 23, 81–83, 88–89, 94, 99–105 U u-stems 198, 205–207, 208 n. 21, 209– 214, 215 V voice – active 71–73, 77–95, 98–105 – antipassive 101–103, 104 – middle 71–73, 77–95, 100, 102, 104– 105 – passive 71–72, 79–81, 83, 86, 90 n. 53, 94 n. 64, 95 n. 65, 100–104 W West Greek dialects 21 with n. 5, 24 n. 19, 28–30 with ns. 33–34, 36, 39 with n. 74 Western Greek poetry 109, 111–115, 119

Index of Passages Alcman (PMG) fr. 1.49

64–65 n. 54

[Ammonius] (Nickau) De adfinium vocabulorum differentia 282 134 Anthologia Palatina 10.101

111–112

Antiatticist (Valente) α 18 256 α 107 254 α 135 257 Apollonius Dyscolus (Schneider) GG 2.1.1, 74.10 24 n. 19 GG 2.1.1, 77.5 24 with n. 19, 29, 126 GG 2.1.1, 77.10 24 Apollonius of Rhodes Argon. 1.361−362 33 Argon. 4.802 33 Argon. 4.1470−1471 33 Aretaeus of Cappadocia De causis et signis diuturnorum morborum 4.12.3.4–5 234 Aristophanes Nub. 254 Plut. 1034–1035 fr. 233 K.-A.

248 with n. 21 227 50–51

Aristotle Metaph. 1091b.8 Poet. 4 1449a.24–28 Rh. 3.1 1403b.26–31 Rh. 3.5 1407b.14–17

188 187 186 188

Arrian Anab. 7.6.2

235

https://doi.org/10.1515/9783110621747-016

Bacchylides 3.78–82 5.180–181 8.26–27

152 119 118–119

Callimachus Hymn. 5.37 Hymn. 6.13 Iamb. 13.18

30 117 120

CEG 302

118 n. 35

Charon of Lampsacus (FGrHist) 262 F 10 178 with n. 14 CLOD 313A 430.4 2028A

131 n. 26, 143–144 131 131

Corinna (PMG) fr. 654 col. iii ll. 32–41 127–128 fr. 662 24 n. 19, 27 n. 31, 29, 126 n. 3, 141 fr. 663 27 n.32, 141 fr. 673 126 n. 3, 140–141 fr. 681 25 n. 31 Cosmas Indicopleustes (Wolska-Conus) Topographia Christiana 5.26 260 [Demosthenes] 7.42 Schol. 7.42 Dilts

133–135 134

Derveni Papyrus (Kouremenos/ Parassoglou/Tsantsanoglou) col. 23.11 118 with n. 34 DGE 726.44

37 n. 64

280  Index of Passages Diodorus Siculus 3.21.1

207, 214

Dionysius of Halicarnassus Comp. 17.2–15 190 Comp. 18.10–11 190–191 Dem. 48 190 Thuc. 5 188, 189 n. 83 Empedocles (D.-K.) B 17.1–2 = 16–17 B 20.7 B 28.1 B 61.3 P.Strasb. a(ii).28 Epicharmus (K.-A.) fr. 32.12 fr. 35.3 fr. 76 fr. 80

107–112 113–114 23, 36 with n. 63 114–115 113–114

fr. 92 fr. 100 fr. 113.12 fr. 113.415 fr. 120 fr. 185

166 n. 69 166 n. 69 153–155 158–161, 164–168, 170 156–157 151 151 n. 9 168 151 166

[Epicharmus] (K.-A.) fr. 256 fr. 275.1

152 166

Epimerismi Homerici (Dyck) μ 65 ( = Gloss. Ital. no. 35 K.-A.) 168–170 Etymologicum Magnum (Lasserre/ Livadaras) α 2087 258 n. 48, 265 Euripides Med. 974 fr. 716 Kannicht Eusebius Vita Constantini 4.65.3

Eustathius De emendanda vita monachica 152.10 Metzler 263 Il. 1.32.28−31 van der Valk 262–263 Getty Hexameters (Jordan/Kotansky) col. I 10 113 col. I 11–12 117 Gregory of Nazianzus Orationes 40.32 (= 36.404C Migne) 139 Hecataeus of Miletus (FGrHist) 1 F 163 177, 185 n. 66 1 F 204 185 with n. 66 1 F 229 185 with n. 66 1 F 282 177, 185 n. 66 1 F 293 177, 185 n. 66 Heraclitus (D.K.) B 26

179

Hermogenes (Rabe) Id. 411.12–25

189

Herodotus 5.49.2–3 5.60.1–2 7.14.1

224 30 200

Hesiod Op. 591 Theog. 401 fr. 245 M.-W.

76 23, 28 n. 34, 31, 36 n. 62 25, 27 n. 32

184 200

Hesychius (Latte/Cunningham) α 1672 260 n. 52 α 6415 138–139 α 8282 258, 260 n. 52 β 693 47

236

Hippocrates Epid. 5.43

226

Index of Passages  281

5.98 7.20 7.29 7.54 7.62

226 226 226 226 226

Hittite Laws § 147/*36

212

Homer Il. 1.234−237 2.272−275 7.438 8.309–334 8.335–337 8.407 9.232–239 10.141–142 13.495 12.40–46 14.521 15.354−356 17.19–23 17.132–137 17.541–542 18.225−227 20.36–37 21.573–580 22.118 23.112 23.459–464 24.114 24.134 24.191−192 24.293 24.311 Od. 1.68–69 3.18 4.38 9.348 11.164−168 11.494 14.134 14.213

101 76 185 61 43, 61 97 44 74 26, 28–29, 32 44 26 with n. 29 86 45 44–45 104 85 45 65–66 75, 103 91 65 26 26 103 32 32 97 75 n. 14 26 with n. 29, 29, 32 75 n. 14 74 76 102 102

17.371 22.402−403 23.237 Schol. A Il. 2.239 Erbse Schol. D Il. 8.337 van Thiel Schol. Aim Il. 19.384 Erbse

76 103 84 31 n. 43, 32 46–47, 55 31 n. 43, 32

Homeric Hymns Hymn. Hom. Merc. 158

118 n. 35

IEleusis 300.65–67

138

IG VII 3172 IV (H) ll. 61–75 (= Migeotte no. 13) 129, 130 n. 19, 136– 140 VII 505 129, 143 IGDOP 23.7

37 n. 64

IK.Priene 398.4

37 n. 64

IOropos 93

129, 143

Isocrates 2.34 5.25–27

201 186, 191 n. 98

KBo 5.3 iii 32 18.28, 9

212 212

KUB 29.11, 1b 29.11, 7−8b

211–212 212

Leo Diaconus (Hase) Historia 137.10−15 264

282  Index of Passages Photius (Theodoridis) α 2724 α 3169 κ 660 κ 662

251 n. 32, 254 258 n. 48 247 247

Phrynichus (Fischer) Eclogae 20 Eclogae 39 Eclogae 41 Eclogae 54

247 253 248 245

Pindar Nem.3.80–82 Ol. 1.36–42 Ol. 4.16 Ol. 9.43–46 Pyth. 2.72–73 Pyth. 8.1–4 fr. 109.2 S.-M.

199 153 151 151 155 151–152 151–152

252

Nicander Alex. 81–82

Plato Leg. 815a.1-7

202

49

Orus (Alpers) fr. 13

Pollux (Bethe) 4.122

257

251 n. 32, 253

Polybius 9.21.1 9.43.4

231 199

P.Sorb. 2.69

138

Livy 26.24.11 38.41.4−7

205 208

[Longinus] Subl. 29.1

48

LXX Exodus 26.1 37.14 Isaiah 54.2 Lamentations 1.12 1.18 2 Maccabees 3.17 Psalms 68 (69).26–27 Sirach 26.6

259–260 with n. 52 260 n. 52 261–262 232 232 233 232 232

Lysias (Carey) fr. 288

253

Moeris (Hansen) χ5

Panyassis fr. 17.7–9 Bernabé = fr. 13.7–9 Davies 48 fr. 28 Matthews = fr. 31 Bernabé 116–119 Parmenides (D.-K.) B 1.24 B 6.5

114–115 114–115

P.Grenf. 2.70.25

138

Pherecydes of Syrus (D.-K.) B1 176 [Philoxenus] (Theodoridis) fr. 690** 135–136

[Ptolemy] (Heylbut) De differentia vocabulorum p. 399 134 P.Würzb. 19

138

Rhinthon (K.-A.) fr. 6

168

SEG 22.390

129

Index of Passages  283

Sophocles Phil. 791–792

223

Sophron (K.-A.) fr. 56 166 n. 69 fr. 86 166 n. 69 Synagoge (Cunningham) α 648 256 α 1091 258 σ 122 249 Synagoge versio codicis B (Σb) (Cunningham) α 2035 251 α 2238 249 α 2405 258 Theodorus Prodromus (Hörandner) Carmina historica 11.131–132 262

Theophanes Continuatus (Featherstone/ Codoñer) Chronographia 4.41.15−16 262 Thomas Magister (Ritschl) σ 333 249 Vita sanctorum Davidis, Symeonis et Georgii Mitylenae (van den Gheyn) 42.22−24 263 Xenophon An. 2.5.13 Eq. 1.4 10.6 Mem. 3.10.5

200 199 199 200