Migration, Mobility and Language Contact in and around the Ancient Mediterranean (Cambridge Classical Studies) 9781108488440, 9781108763943, 1108488447

Table of contents :
Contents
Figures
Tables
Contributors
Acknowledgements
Abbreviations of Ancient Authors and Works
Abbreviations of Modern Sources
1 Introduction
2 Interethnic Mobility and Integration in Pre-Roman Etruria: The Contribution of Onomastics
3 Elusive Migrants of Ancient Italy
4 The Language of Mobile Craftsmen in the Western Mediterranean
5 Lost – and Found – in Transmission: The Creation of the Oscan Alphabet
6 Mobility and Orthography
7 The Mamertini in Messina: Mobility, Migration and Mercenaries
8 Migration, Identity, and Multilingualism in Late Hellenistic Delos
9 Interpretes, Negotiatores and the Roman Army: Mobile Professionals and Their Languages
10 Hoc Primvs Venit: Italians and Others in Egypt before the Caesars
11 Population, Migration and Language in the City of Rome
References
Index Locorum
Subject Index

Citation preview

M I G R ATI ON , M OB I L I T Y A ND LA NG UA GE CON TACT IN A ND A R OU ND TH E AN CI EN T ME D I T E R RA NE A N

Migration, Mobility and Language Contact in and around the Ancient Mediterranean is the first volume to show the different ways in which surviving linguistic evidence can be used to track movements of people in the ancient world. Eleven chapters cover a number of case studies, which span the period from the seventh century BC to the fourth century AD, ranging from Spain to Egypt, from Sicily to Pannonia. The book includes detailed study of epigraphic and literary evidence written in Latin and Greek, as well as work on languages which are not so well documented, such as Etruscan and Oscan. There is a subject index and an index of works and inscriptions cited. james clackson is Professor of Comparative Philology in the Faculty of Classics at the University of Cambridge and is one of the foremost scholars of the historical sociolinguistics of Latin and Greek. His major publications include Indo-European Linguistics (Cambridge, 2007), The Blackwell History of the Latin Language (with Geoffrey Horrocks) (2007), A Companion to the Latin Language (2011) and Language and Society in the Greek and Roman Worlds (Cambridge, 2015). He was PI of the AHRC-funded Project ‘Greek in Italy’ (2013–18). patrick james teaches Classics at Haileybury College, Hertfordshire. He served for nine years as an Assistant Editor for the Cambridge Greek Lexicon. katherine mcdonald is Lecturer in Classics and Ancient History at the University of Exeter. Her books include Oscan in Southern Italy and Sicily (CUP, 2015) and Italy before Rome: A Sourcebook (forthcoming), written as part of her AHRC-funded project ‘Connectivity and Competition: Multilingualism in Ancient Italy’. livia tagliapietra is Junior Research Fellow in Classics at Jesus College, Cambridge. She wrote her PhD thesis as part of the AHRC-funded project ‘Greek in Italy’. Her research interests include ancient Greek linguistics and dialectology, historical sociolinguistics, and the languages and cultures of ancient Italy. nicholas zair is Lecturer in Classics at the University of Cambridge and Fellow of Peterhouse. He has written two books: The Reflexes of the Proto-IndoEuropean Laryngeals in Celtic (2012) and Oscan in the Greek Alphabet (Cambridge, 2016). He is presently working on a third, on Roman imperial inscriptions and sub-elite education, while holding a Pro Futura Scientia Fellowship based in Cambridge and Uppsala.

c a m b r i d g e cl a s s i c a l st u d i e s General editors

j. p. t. c la ck s on , w . m . be a rd , g. be t eg h , r . l . h u n te r , m . j . m i l le t t, s. p . o a k le y , r. g. osb or n e, t. j. g. wh itmar sh

M I GR ATI O N, M OB I L I T Y A ND L A NG UA GE C ON TAC T I N A ND AROU ND T HE AN CIENT M EDITERRANEA N

edited by J A M E S CL A C KS ON University of Cambridge PATR I C K J AME S Haileybury College KATH E R IN E MC DO NA L D University of Exeter L I V I A TA GL I AP I E TR A University of Cambridge NI C HO L A S Z A I R University of Cambridge

University Printing House, Cambridge cb2 8bs, United Kingdom One Liberty Plaza, 20th Floor, New York, ny 10006, USA 477 Williamstown Road, Port Melbourne, vic 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 79 Anson Road, #06–04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108488440 doi: 10.1017/9781108763943 © Faculty of Classics, University of Cambridge 2020 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2020 Printed in the United Kingdom by TJ International Ltd, Padstow Cornwall A catalogue record for this publication is available from the British Library. isbn 978-1-108-48844-0 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

This book is dedicated to Geoffrey Horrocks as a small token of profound appreciation.

CONT ENTS

List of Figures List of Tables List of Contributors Acknowledgements Abbreviations of Ancient Authors and Works Abbreviations of Modern Sources

page ix x xi xii xiii xvi

1 Introduction james clackson, patrick james, katherine mcdonald, livia tagliapietra and nicholas zair

1

2 Interethnic Mobility and Integration in Pre-Roman Etruria: The Contribution of Onomastics daniele f. maras

23

3 Elusive Migrants of Ancient Italy elena isayev

53

4 The Language of Mobile Craftsmen in the Western Mediterranean katherine mcdonald and james clackson

75

5 Lost – and Found – in Transmission: The Creation of the Oscan Alphabet karin w. tikkanen

98

6 Mobility and Orthography: A Contextualisation of Variant Spellings in the Oscan Inscriptions in the Greek Alphabet livia tagliapietra

122

vii

Contents

7 The Mamertini in Messina: Mobility, Migration and Mercenaries nicholas zair

156

8 Migration, Identity, and Multilingualism in Late Hellenistic Delos francesco rovai

171

9 Interpretes, Negotiatores and the Roman Army: Mobile Professionals and Their Languages rachel mairs

203

10 HOC PRIMVS VENIT: Italians and Others in Egypt before the Caesars patrick james

230

11 Population, Migration and Language in the City of Rome olivia elder

268

References Index Locorum Subject Index

viii

296 335 350

FIGURES

3.1

Journeys in three plays of Plautus as depicted on a conventional modern map: Curculio, Persa, Poenulus (Adapted from Richlin 2005: 18)

4.1

Plasos’ signature (Image: Katherine McDonald)

82

4.2

Dish with Plator’s signature (From Crawford et al. 2011: 561)

85

4.3

Plator’s signature (From Crawford et al. 2011: 561)

85

4.4

L. H. Jeffery’s drawing of Kalon’s signature and Glaukies’ dedication (Drawing held at the Anne Jeffery Archive, courtesy of the Centre for the Study of Ancient Documents, Oxford University)

92

6.1

The Oscan vowel system

124

8.1

Network of associations within a group of seals (Adapted from Auda & Boussac 1996: 323)

181

8.2

Graffito of the ‘Maison du Théâtre’ (From Rostovtsev 1896: 392)

188

8.3

Graffito from an altar in the Agora des Italiens (From Roussel & Hatzfeld 1909: 508–10 fig. 4)

197

page 65

ix

TA B LES

6.1

Spelling of Oscan vowels in the native Oscan alphabet

6.2

Inconsistent spellings of the Oscan vowels in the Greek alphabet

126

6.3

Lejeune’s model (simplified)

126

6.4

Zair’s model (from Zair 2016: 79)

128

6.5

Spelling in South Oscan inscriptions c. 400–200 BC (excl. Messana)

131

6.6

Spelling in the Messana inscriptions

140

6.7

Spelling disambiguation (highlighted: ambiguous spellings)

147

8.1

The Delian population in the public and funerary inscriptions

173

8.2

Delian documents containing Latin names

194

x

page 125

C O N T R I B U TO R S

james clackson is Professor of Comparative Philology at the University of Cambridge. olivia elder is Research Fellow in Classics at Sidney Sussex College, Cambridge. elena isayev is Professor of Ancient History and Place at the University of Exeter. patrick james teaches Classics at Haileybury School, Hertfordshire. rachel mairs is Professor of Classics and Middle Eastern Studies at the University of Reading. daniele f. maras is Funzionario Soprintendenza for Southern Etruria.

Archeologo

at

the

katherine mcdonald is Senior Lecturer in Classics and Ancient History at the University of Exeter. francesco rovai is Associate Professor in the Department of Philology, Literature and Linguistics at the University of Pisa. livia tagliapietra is Junior Research Fellow in Classics at Jesus College, Cambridge. karin w. tikkanen is Researcher in Latin at Uppsala University. nicholas zair is University Lecturer in Classics at the University of Cambridge and Fellow of Peterhouse.

xi

A C KN O W L E D G E M E N T S

Several of the chapters in this volume started out as presentations at a conference held in Cambridge in May 2016, organised by Katherine McDonald and Nicholas Zair. The conference itself was made possible by the generous support of the UK Arts and Humanities Research Council for the ‘Greek in Italy Project’, which ran from January 2014 to April 2018. The editors here acknowledge our great debt to the AHRC and its peer reviewers for enabling this research. We would also like to thank all the people in Cambridge who contributed to the success of the conference and the genesis of this book: the administrative and clerical staff of the Cambridge Faculty of Classics, in particular Nigel Thompson, Lucyna Prochnicka and Lina Undicino; Peterhouse, Cambridge; the editors and readers of the Cambridge Classical Studies series, with special thanks to the convenor Robin Osborne; and the staff of Cambridge University Press, in particular the Classics Editor, Michael Sharp, and Hal Churchman. Finally, we acknowledge the huge contribution of Geoff Horrocks, who has been a teacher, colleague, mentor and friend to each of the editors, and whose unfailing wisdom and good humour were essential to the success of the Greek in Italy project.

xii

ABBREVIAT IONS OF ANC IENT AUTHORS AND WORKS

App. BC Mith. Pr. Pun. Arist. Pol. Ath. Deip. Cato Agr. Cic. Agr. Att. Cat. M. De or. Fam. Flac. Leg. Man. Mur. Rep. Ver. D.H. D.S. Fest. Gel. Hdt. Isid. Or.

Appianos (Appian) Civil Wars The Mithridatic Wars Preface Punica Aristoteles (Aristotle), Politika Athenaios (Athenaeus), Deipnosophistai M. Porcius Cato (Cato), de Agri Cultura M. Tullius Cicero (Cicero) de Lege Agraria Epistulae ad Atticum Cato Maior de Senectute de Oratore Epistulae ad Familiares pro Flacco de Legibus pro Lege Manilia pro Murena de Republica in Verrem Dionysios Halikarnasseus (Dionysius of Halicarnassus) Diodorus Siculus S. Pompeius Festus (Festus), de Significatu Verborum Aulus Gellius, Noctes Atticae Herodotos (Herodotus) Isidorus Hispalensis (Isidore of Seville), Origines xiii

Abbreviations of Ancient Authors and Works

Just. Juv. Liv. Mar. Victorin. Mart. Men. Asp. Dis Philo Leg. Pl. As. Aul. Capt. Cas. Cistellaria Cur. Epid. Men. Mer. Mil. Mos. Per. Poen. Ps. Rud. St. Trin. Truc. Plin. Nat. Plut. Numa Tim. Poll. Polyb. Sen. xiv

Justin, Epitoma Historicarum Philippicarum Pompei Trogi D. Iunius Iuuenalis (Juvenal), Saturae T. Liuius (Livy), ab Vrbe Condita C. Marius Victorinus, Ars grammatica M. Valerius Martialis (Martial) Menandros (Menander) Aspis Dis Exapaton Philon (Philo), Legum Allegoriae T. Maccius Plautus (Plautus) Asinaria Aulularia Captiui Casina Cistellaria Curculio Epidicus Menaechmi Mercator Miles Gloriosus Mostellaria Persa Poenulus Pseudolus Rudens Stichus Trinummus Truculentus C. Plinius Secundus (Pliny the Elder), Naturalis Historia Plutarchos (Plutarch) Numa Pompilius Timoleon Ioulios Polydeukes (Julius Pollux), Onomasticon Polybios (Polybius), Histories L. Annaeus Seneca (Seneca the Younger)

Abbreviations of Ancient Authors and Works

Ep. Helv. Strabo Suet. Aug. Cal. Claud. Gram. Iul. Ulpian, Dig. Var. R. Vel. Long. V. Max.

Epistulae Morales ad Heluiam matrem, de Consolatione Strabon (Strabo), Geographika C. Suetonius Tranquillus (Suetonius) Augustus Caligula Claudius de Grammaticis Iulius Cn. Domitius Annius Ulpianus (Ulpian), Digesta seu Pandectae M. Terentius Varro (Varro), Res Rusticae Velius Longus, de Orthographia Valerius Maximus, Facta et Dicta Memorabilia

xv

A B B R E V I AT I ON S O F MO D E R N S O U R C E S

Abydos = Perdrizet, Paul, & Gustave Lefebvre. (1919). Les graffites grecs du Memnonion d’Abydos. Nancy, Paris and Strasbourg: Berger-Levrault. AE = L’Année épigraphique (1888–). Paris: Académie des Inscriptions et Belles-Lettres. BGU = Aegyptische Urkunden aus den Königlichen (later Staatlichen) Museen zu Berlin, Griechische Urkunden. Berlin. CEG = Hansen, Peter Allan. (1983–9). Carmina Epigraphica Graeca. Two volumes. Berlin: De Gruyter. CIE = Pauli, Carl et al. (1883–). Corpus Inscriptionum Etruscarum. CIG = (1828–77). Corpus Inscriptionum Graecarum. Four volumes. Berlin: Deutsche Akademie. CIL = Corpus Inscriptionum Latinarum. CIL I2 = Lommatzsch, Ernst & Theodor Mommsen. (1918). Inscriptiones latinae antiquissimae ad C. Caesaris mortem. Pars posterior. Berlin: Reimer. CIL I2 fasti = Hülsen, Christian, Theodor Mommsen & Wilhelm Henzen. (1893). Inscriptiones latinae antiquissimae ad C. Caesaris mortem. Pars prior: Fasti consulares ad a.u.c. DCCLXVI. Elogia clarorum virorum. Fasti anni Iuliani. Berlin: Reimer. CIL II = Hübner, E. et al. (1869–). Inscriptiones hispaniae latinae. Berlin: Reimer. CIL III = Mommsen, Theodor, Otto Hirschfeld & Alfred von Domaszewski. (1902). Inscriptionum Orientis et Illyrici latinarum supplementum. Pars posterior. Berlin: Reimer. CIL IV = Mau, August & Karl F. W. Zangemeister. (1909). Inscriptionum parietariarum Pompeianarum supplementum. Pars posterior: Inscriptiones parietariae et vasorum fictilium. Berlin: Reimer. xvi

Abbreviations of Modern Sources

CIL VI = Henzen, G. et al. (1876–). Inscriptiones urbis Romae latinae. Berlin: Reimer. CIL VIII = Mommsen, Theodor. (1881). Inscriptiones Africae latinae. Berlin. CIL X = Mommsen, Theodor. (1883). Inscriptiones Bruttiorum, Lucaniae, Campaniae, Siciliae, Sardiniae latinae. Berlin: Reimer. CIL XI = Bormann, E. (1888–1926). Inscriptiones Aemiliae, Etruriae, Umbriae latinae. Berlin: Reimer. CIL XIII = Hirschfeld, Otto & Karl F. W. Zangemeister. (1899–). Inscriptiones trium Galliarum et Germaniarum latinae. Berlin: Reimer. DGE = Schwyzer, Eduard. (1923). Dialectorum graecarum exempla epigraphica potiora. Third edition. Leipzig: Hirzell. ET = Meiser (2014). FGH = (1923–). Fragmente der griechischen Historiker. Berlin / Leiden: Weidmann / Brill. GHI = Meiggs, Russell & David M. Lewis. (1969). A Selection of Greek Historical Inscriptions: To the End of the Fifth Century B.C. Oxford: Clarendon Press. Graff.Dodec.Philae = Griffith, Francis Llewellyn. (1935–7). Les Temples immergés de la Nubie. Catalogue of the Demotic Graffiti of the Dodecaschoenus. Oxford: Oxford University Press. GVI = Peek, Werner. (1955). Griechische Vers-Inschriften I, GrabEpigramme. Berlin: Deutsche Akademie. HNItaly = Rutter (2001). IC = Guarducci, Margherita. (1935–50). Inscriptiones Creticae. Four volumes. Rome: Libreria dello Stato. ICUR = De Rossi, Giovanni Battista (1857–1915). Inscriptiones christianae urbis Romae septimo saeculo antiquiores. Rome: Libraria Pontificia. ID = Dürrbach, Felix et al. (1926–72). Inscriptions de Délos. Seven volumes. Paris: Champion. I.Delta = Bernand, André. (1970). Le Delta égyptien d’après les textes grecques, I. Les confins libyques. Three volumes. Cairo: Institut Français d’Archéologie Orientale du Caire. xvii

Abbreviations of Modern Sources

IDRE = Petolescu, Constantin C. (1996–2000). Inscriptions de la Dacie romaine. Inscriptions externes concernant l’histoire de la Dacie (Ier – IIIe siècles). Bucarest: Institutul de arheologie. IG = Inscriptiones Graecae. IG IX = Dittenberger, Wilhelm. (1897–1908). Inscriptiones Graecae Phocidis, Locridis, Aetoliae, Acarnaniae, insularum Maris Ionii. Berlin: Reimer IG XI.2 = Durrbach, Félix. (1912). Inscriptiones Deli. Fasc. 2. Berlin: Reimer. IG XI.4 = Roussel, Pierre. (1912). Inscriptiones Deli. Fasc. 4. Berlin: Reimer. IG XII.4 2 = Bosnakis, Dimitris & Klaus Hallof. (2012). Inscriptiones insularum maris Aegaei praeter Delum, 4. Inscriptiones Coi, Calymnae, Insularum Milesiarum. Pars II. Inscriptiones Coi insulae: catalogi, dedicationes, tituli honorarii, termini. Berlin and New York: De Gruyter. IG XIV = Kaibel, Georg. (1890). Inscriptiones Siciliae et Italiae, additis Galliae, Hispaniae, Britanniae, Germaniae inscriptionibus. Berlin: Reimer. IGASMG III = Arena, Renato. (1994). Iscrizioni greche arcaiche di Sicilia e Magna Grecia. Vol. 3. Iscrizioni delle colonie euboiche. Pisa: Nistri Lischi. IGDGG = Dubois, Laurent. (1995–2002). Inscriptions grecques dialectales de Grande-Grèce. Geneva: Droz. IGUR = Moretti, Luigi. (1968–79). Inscriptiones Graecae urbis Romae. Four volumes. Rome: Istituto Italiano per la Storia Antica. I.Hermoupolis = Bernand (1999). I.Locri = Del Monaco (2013a). ILS = Dessau, Hermann. (1892–1916). Inscriptiones latinae selectae. Three volumes. Berlin: Weidmann. ILLRP = Degrassi, Attilio. (1957). Inscriptiones latinae liberae rei publicae. Two volumes. Florence: La Nuova Italia. IMT = Inschriften Mysia & Troas. Inscr.It. = (1931–). Inscriptiones Italiae. I.Philae = Bernand (1969). I.Rhegion = D’Amore, Lucia. (2007). Iscrizioni greche d’Italia: Reggio Calabria. Rome: Quasar. xviii

Abbreviations of Modern Sources

I.Syène = Bernand (1989). IvO = Dittenberger, Wilhelm, & Karl Purgold. (1896). Die Inschriften von Olympia. Berlin: A. Asher & Co. JIWE = Noy (1993–5). LGPN = Fraser, P. et al. (1987–). Lexicon of Greek Personal Names. (Online version available at www.lgpn.ox.ac.uk/). LSJ = Liddell, H. G., R. Scott & H. S. Jones. (1940). A GreekEnglish Lexicon. New (9th) edition. Oxford: Clarendon. MLM = De Simone & Marchesini (2002). O.Cair. = Gallazzi, C., R. Pintaudi & K. A. Worp. (1986). Ostraka greci del Museo Egizio del Cairo (Papyrologica Florentina 14). Florence: Gonnelli. OCD = Oxford Classical Dictionary. Oxford: Oxford University Press. O.Did. = Cuvigny, H. (2012). Didymoi. Une garnison romaine dans le désert oriental d’Égypte. Band 2. Les textes. Cairo: Institut français d’archéologie orientale du Caire. O.Edfou = Ostraca published in Tell Edfou I-III. (1935–50). Cairo. OGI = Dittenberger (1903–5). O.Wâdi Fawâkhir = Guéraud, O. (1942). Ostraca grecs et latins de l’Wâdi Fawâkhir. Bulletin de l’Institut français d’archéologie orientale 41, 141–96. P.Bingen = Melaerts, H. (ed.). (2000). Papyri in Honorem Johannis Bingen Octogenarii (Studia varia Bruxellensia ad orbem Graeco-Latinum pertinentia 5). Leuven: Peeters. P.Cair.Zen. = Edgar, C. C. (1925–40). Zenon Papyri, Catalogue général des antiquités égyptiennes du Musée du Caire. Cairo. P.Hamb. = Griechische Papyrusurkunden der Hamburger Staatsund Universitätsbibliothek. P.Hels. = Frösén, J. et al. (1986). Papyri Helsingienses I, Ptolemäische Urkunden. Helsinki: Societas Scientiarum Fennica. P.Hib. = (1906–55). The Hibeh Papyri. London. P.Lond. = (1893–1907). Greek Papyri in the British Museum. London. P.Michael. = Crawford, D. S. (1955). Papyri Michaelidae, being a Catalogue of Greek and Latin Papyri, Tablets and Ostraca in the Library of Mr G. A. Michailidis of Cairo. Aberdeen. xix

Abbreviations of Modern Sources

Poseidonia (1996) = I Greci in Occidente: Poseidonia e i Lucani. Milan: Electa. P.Oslo = (1925–36). Papyri Osloenses. Oslo. P.Oxy. = (1898–). The Oxyrhynchus Papyri. London. P.Quseir = Bagnall, R. S. (1986). Papyri and Ostraka from Quseir al-Qadim. Bulletin of the American Society of Papyrologists 23, 1–60. P.Saqqara = El-Khouly (1973). PSI = (1912–). Papiri greci e latini. Florence. P.Strasb. = Preisigke, F. (1912–). Griechische Papyrus der Kaiserlichen Universitäts- und Landesbibliothek zu Strassburg. Leipzig. P.Tebt. = (1902–). The Tebtunis Papyri. London. RDGE = Sherk (1969). RIB = Collingwood, R. G., R. P. Wright, R. S. O. Tomlin & M. W. C. Hassall. (1965–2009). Roman Inscriptions of Britain. Oxford: Oxford University Press. RIG = Duval, P.-M. et al. (1985–2002). Recueil des inscriptions gauloises. Paris: Éditions du Centre national de la recherche scientifique. SB = Sammelbuch griechischer Urkunden aus Aegypten. SEG = (1923–). Supplementum Epigraphicum Graecum. Leiden / Amsterdam: Sijthoff / Gieben. ST = Rix (2002). Syll.³ = Dittenberger, Wilhelm et al. (1915–24). Sylloge inscriptionum graecarum. Third Edition. Four volumes. Leipzig. Syringes = Baillet (1920–6). Tab.Sulis = R. S. O. Tomlin. (1988). The Curse Tablets. In B. Cunliffe, The Temple of Sulis Minerva at Bath. II The Finds from the Sacred Spring, 59–269. Oxford: Oxford University Committee for Archaeology. Tab.Vindol. = Bowman, Alan K. & J. David Thomas. (1994–2003). The Vindolanda Writing-Tablets: (Tabulae Vindolandenses). London: British Museum. TAM = (1901–). Tituli Asiae Minoris.

xx

Abbreviations of Modern Sources

TM Nam = Trismegistos People database of ancient names. (Available online at www.trismegistos.org/name/). Ve = Vetter (1953). W.Chr. = Mitteis, L. & U. Wilcken. (1912). Grundzüge und Chrestomathie der Papyruskunde. Two volumes. Leipzig: Teubner.

xxi

C HA P T E R 1

I N T R O D U C T I ON James Clackson, Patrick James, Katherine McDonald, Livia Tagliapietra and Nicholas Zair

The study of migration in the ancient world unexpectedly became a topic of the global news cycle in the summer of 2017. ‘The Story of Britain’, a BBC cartoon for schools that depicted a black soldier in Roman Britain, after Britain, generated Twitter exchanges, subsequently expanded into blogs, newspaper articles and think pieces around the world. Historians, archaeologists, geneticists, statisticians as well as others from outside academia contributed to a debate about the amount of ethnic diversity in Roman Britain and the origin and impact of ancient migrants to the British Isles.1 The editors of this volume do not expect that it will have an impact equivalent to the BBC cartoon, but we hope that the chapters within it can both contribute to the gradual disentanglement of scanty, sometimes contradictory, evidence and present new ways of looking at ancient migration, while also laying bare some of the tacit or unwarranted assumptions that have been made. The approach of this volume is to highlight and further the contributions that linguistic research brings to the study of ancient movements of people. It is indicative of modern approaches that much of the debate over the BBC ‘The Story of Britain’ cartoon centred around genetics rather than linguistics, with DNA evidence invoked by all sides to bolster their claims. In the nineteenth and early twentieth centuries, comparative linguistics held an analogous position to that of genetics today, providing scientific support to theories of prehistoric and historic migrations. In the first edition of Charles Darwin’s Origin of Species language relationships were already seen 1

Since Mary Beard became a target for abuse after her participation in the debate, it is appropriate here to refer to her blog post which also summarises some of the principal arguments; www.the-tls.co.uk/roman-britain-black-white/ (posted 3 August 2017).

1

James Clackson et al.

as intertwined with the ‘genealogical arrangement of the races of man’.2 The dispersal and spread of language families, in particular the Indo-European family, were generally explained as the result of prehistoric migrations of the peoples speaking the parent language (now known as Proto-Indo-European, but formerly also called Aryan).3 There were, it is true, some nineteenth-century scholars who rejected a prehistoric correlation between language and ethnicity. One such was the French orientalist Jules Oppert, who used his inaugural lecture for a course in Sanskrit at the Bibliothèque impériale in Paris in 1865 to argue for the genesis of new languages and peoples through mixing of different elements.4 Oppert’s apparent criticism of the bases for language families was, however, resoundingly rejected by Indo-Europeanists such as William Dwight Whitney, who categorically denied that modern parallels for language change, such as the adoption of Latin by speakers of other languages during the Roman Empire, could be assumed for prehistoric movements.5 The rise of genetics as a guide to human prehistory has had the consequence that linguistic relationships are now less widely invoked as decisive evidence in arguments about ancient migrations. Moreover, the atrocities committed by the Nazis in Germany, and the continued presence of hate speech and hate crimes associated with advocates of an ‘Aryan Race’, have led both to criticism in some quarters of the entire enterprise of IndoEuropean studies6 and to a reluctance by some linguists to discuss anything beyond the purely linguistic. Linguists have also become more aware of the large range of circumstances that might explain the spread of language families, including the role played by linguistic contact. It is recognised that words, sounds and even linguistic structures can be ‘borrowed’ from one language by another in numerous ways, often through the medium of bi- or 2 3

4 5

6

2

Darwin (1859: 422). For a brief overview of earlier theories on the Indo-European ‘migrations’, see Renfrew (1987: 35–41). Oppert (1866). Whitney reviewed Oppert’s work in the North American Review, vol. 105, no. 217 (1867), 521–54, a paper which was subsequently republished as a chapter in Whitney (1873: 198–238). For Whitney’s views on language and race, see Alter (2005: 147–54). See recently the critique of Demoule (2014).

Introduction

multilingual individuals. Many linguists are now consequently suspicious of easy equivalences between genetic phylogenies and language family trees.7 At the same time, archaeologists have moved away from seeking to explain cultural change through mass migrations of peoples, but have theorised different models to account for the spread of cultural artefacts, beliefs and languages.8 While the evidence for prehistoric mass migrations still remains contested, with much weight currently apportioned to a small amount of ancient DNA evidence,9 there is an increasing amount of study concerning movement of individuals and larger groups of people in early attested history, particularly in the Roman world. One spur for much recent research has been the programmatic work of Peregrine Horden and Nicholas Purcell (2000), The Corrupting Sea, which sets forth a model of a ‘connected’, ‘networked’ Mediterranean in ancient and medieval history. In this approach, also followed notably by Cyprian Broodbank (2013) for the prehistoric period, small-scale movement of peoples and goods around the shores of the Mediterranean is seen as a near-constant feature of life since the first invention of seagoing craft. In these and other works, ‘connectivity’ has come to be used as a loose term describing the different ways in which individuals, goods and ideas have moved around the Mediterranean and even further afield, and how geographically distant regions came into close contact through an aggregation of many short journeys.10 As Luuk de Ligt and Laurens Tacoma have recently shown (in one of a recent flurry of publications partly arising from a project on migration supported by the Netherlands Organisation for Scientific Research),11 prosopographic and demographic work on the abundant surviving Greek 7

8

9

10

11

For two influential interventions into the debate by linguists, see Sims-Williams (1998) and Campbell (2015). Renfrew (1987) was the first work to apply these archaeological ideas to the spread of Indo-European languages. Broodbank (2013) offers a prehistory of the Mediterranean which largely avoids the reliance on mass migrations of earlier accounts. See, for example, the claims made in Anthony and Brown (2017). For a more nuanced and critical use of evidence from ancient skeletons, see Tacoma (2016b), and for a recent popular account stressing the fast-changing nature of the field, see Reich (2018). See Pettigrew (2013) for a useful summary of the meanings of connectivity in ancient history. de Ligt and Tacoma (2016b: 3–4). See also Tacoma (2016a) and Lo Cascio and Tacoma (2016).

3

James Clackson et al.

and Roman epigraphic and papyrological sources adds substance to the idea of connectivity; de Ligt and Tacoma exemplify this through a number of studies which use the details provided by onomastic indicators and ethnic designations to track movement around and beyond the Mediterranean. The new attention paid towards migration in ancient history, particularly in the Roman Empire, is greatly welcomed by the editors of this volume, but we feel that there is still a gap. Most of the recent work done on migration on historical times has paid attention to archaeological, osteological and epigraphic evidence, but within the epigraphic field the emphasis has been on onomastic rather than linguistic data. This volume will show how historical linguists (or indeed historical sociolinguists) can apply the evidence available to them to the emerging picture of migration. The chapters in this volume collectively show how to use research into language contact, language shift and language change to build up a better sense of the integration of migrants into their new surroundings: which elements, if any, of their former languages, scripts and naming systems did they keep, and what did they adopt? Careful consideration of linguistic evidence sometimes allows the researcher to separate temporary visitors or travellers from those migrants who permanently settle elsewhere, and to distinguish individuals or small groups belonging to specific language communities among wider mobile populations.12 Language and script use can sometimes show what was considered more salient to a community’s identity, whether an onomastic formulation, distinctive orthography or dialectal form.13 The survival or loss of a language can also reveal how different communities adapted to new surroundings, and in particular we look at the interactions between Greek colonists in Magna Graecia, Oscan mercenaries in Sicily and Italians at Delos and in Egypt with neighbouring groups.14 Finally, we ask whether the study of 12

13

14

4

See in particular the chapters by Elder, James, McDonald and Clackson, Maras, Mairs, Rovai, Tagliapietra and Zair. See in particular the chapters by James, Maras, McDonald and Clackson, Rovai, Tikkanen, Tagliapietra and Zair. See in particular the chapters by James, McDonald and Clackson, Rovai, Tikkanen, Tagliapietra and Zair.

Introduction

language use and linguistic diversity can function as a point of entry into the exploration of ancient migration and mobility, asking whether the epigraphic picture tallies well with other sorts of evidence, and how far travellers, interpreters or other migrants are represented in any sort of writing, whether on stone, on perishable materials or in literature.15 This introduction will leave the chapters in the rest of the volume to speak for themselves, but hereafter we shall attempt to tackle some of the general questions which arise from them. Most of the chapters address situations involving what we term the ‘fragmentary corpus languages’ of the ancient world (following McDonald 2015), that is to say Oscan, Etruscan, Gaulish, Palmyrene, Phoenician etc.,16 in an encounter with one of the two major languages, certainly of the Western Mediterranean, Latin and Greek. The study of such languages is of particular interest to linguists, since it affords rare insights into varieties otherwise completely lost, and sometimes offers the missing pieces to puzzles of the array and spread of larger language families. For the ancient historian, these varieties can be used as a check on the information provided by literary sources, whether in Greek or Latin, Hebrew, Aramaic or Egyptian. To take just one example which will be discussed in more detail at various points in this volume, the presence of Oscan inscriptions in Messina helps to corroborate ancient accounts of the movements of the band of Mamertine mercenaries. The inscriptions in fragmentary corpus languages also show that the activities associated with networking and connectivity supposed for the ancient Mediterranean was not limited to certain seafaring or colonial societies, such as Greeks and Phoenicians; they serve to confirm the Horden and Purcell hypothesis that movement and migration were a fact of life for most Mediterranean peoples. Etruscan inscriptions found in the South of France and in North Africa dating from the Roman Republican period, or the Palmyrene inscriptions left in Rome and elsewhere during the Imperial period attest to the wide variety of ancient mobility. 15 16

See in particular the chapters by Elder, Isayev, James, Maras, Mairs and Rovai. We reject other descriptions for languages of this type, such as Trümmersprachen, as potentially misleading.

5

James Clackson et al.

Fragmentary corpus languages therefore give a precious insight into movement of peoples and individuals beyond the Greeks, Romans and Phoenicians. Despite this, we do not know of any full survey in existence of ‘out of place’ texts in fragmentary corpus languages from the ancient world, and in what follows we give a brief account of what there is. We shall give here examples of documents in languages other than Latin and Greek which have been found over 300 km away from areas where the majority of their speakers lived, spanning a very approximate chronological period from 800 BCE to 400 CE. The geographical distance is chosen only as a very rough indication of movement, and inevitably our inadequate knowledge of the geographic extent of many ancient linguistic communities makes exact calculations impossible, as we shall see in the discussion of several examples below. We further note that the discovery of an inscription in a particular place does not necessarily mean that a speaker of the language of the inscription was necessarily in the same place, particularly in the case of texts on pots or other portable objects such as armour. We have generally erred on the side of inclusion of such material, noting its possible significance in our survey, although we have excluded the following categories. First, coins and all other bullion are left to one side. Although coins often give the only surviving evidence for trade and travel, their monetary value is likely to have been more significant than any written legend they carried. Consequently, there is no necessary correlation between the language written on a coin and the language spoken by its final owner. In contrast, although there was certainly considerable ancient trade in ceramics and metalwork, meaningful writing is not generally found on such items. When a pot or a helmet has writing on it, the chances are much higher that its possessor could understand the message. This need not necessarily entail that the owner of an inscribed vessel was a native speaker of the language of the text; there are numerous Greek pots traded to Etruria in the middle of the first millennium BCE, but we cannot suppose that Etruria was colonised by Greeks. It does, however, indicate that Etruscans had sufficient contact with Greeks (or had had at an earlier period) in order to make sense of these inscriptions (Osborne 2007: 89). The survival of out-of-place graffiti or 6

Introduction

memorial inscriptions in a fragmentary corpus language is a yet more secure indication that a speaker has travelled beyond the normal confines of his speech community. Second, we have not listed scarabs, amulets, statues, obelisks or any other material bearing texts in hieroglyphic Egyptian; during the time period of our survey hieroglyphic Egyptian was a script of display and not in common use by individuals. The movement of prestige items adorned with hieroglyphs around the Mediterranean is more revealing about trade in luxury goods rather than the dispersal of speakers of Egyptian. We have also not given instances where spells and curses incorporate magical words or other material having its origin in words from real languages.17 Finally, we have generally not attempted to trace cases where a text in a known language includes an unrecognisable word, perhaps imported from a foreign language, such as the example of a supposed Thracian midne in a Latin inscription (CIL VI 32567), discussed by Elder in Chapter 11 of this volume. We shall work by reference to language group and language family, starting with Indo-European languages. In antiquity the Celtic languages were spoken in the British Isles, north-eastern Spain, much of what is now France, eastern Switzerland and in Italy as far south as the Po Valley. It is possible to recognise separate varieties among the Celtic languages attested epigraphically within our time period, principally Celtiberian, Gaulish and Lepontic. The scantiness of the Celtic evidence from Britain makes it difficult to ascertain whether Celtic words and texts from Britain, in particular two curse tablets from Roman Britain (Tab.Sulis 14 and 18 ), were written in the continental variety now known as ‘Gaulish’ or in the native British Celtic (sometimes referred to as ‘Brythonic’ or ‘Brittonic’).18 The range of Gaulish texts found on the continent is wide, stretching from Todi in Umbria in the south-east,19 Amélie-les-Bains in the 17

18 19

We thus exclude the gold amulet found near York, RIB 706, with a short text apparently in Coptic, φνεβεννουθ, ‘the lord of the gods’ or ‘the divine lord’ since the presence of magical symbols on the same text, and the material on which it is written, indicate that this text was not necessarily written in a language which its bearer would have comprehended. See Mullen (2007) on this question. RIG II.1, E-5 = CIL I2 2103 = CIL XI 4687.

7

James Clackson et al.

Pyranees in the south-west,20 to Châteaubleau, 65 km to the southeast of Paris in the north. There are ten or so Gaulish inscriptions attested in Italy, which are usually referred to as ‘Cisalpine Gaulish’, although there is variation in the terminology used, since there is disagreement as to whether Cisalpine Gaulish should be seen as a separate dialect from Transalpine Gaulish, or whether it is more closely related to Lepontic, a Celtic variety spoken around the northern Italian lakes (which itself may or may not be a separate language from Gaulish. For discussion of these matters see Eska (1998) and Uhlich (1999), (2007)). Most of these inscriptions are plausibly to be associated with the incursion of Gauls into the Po Valley in the fifth century, although some, notably the Todi inscription, a bi-version bilingual with Latin, may reflect later arrivals to Italy. Other scholars have looked for Celtic in inscriptions from further south in Italy. For example, Luuk de Ligt, as a sideline to his studies on Roman republican history, has explained otherwise enigmatic texts from the Adriatic coast (de Ligt 2007a) and as far south as Sicily (de Ligt 2007b) as Celtic; neither proposal has generally been accepted by Celticists.21 Several short and fragmentary inscriptions have also been deemed Celtic from the eastern Alps. These have been recently discussed by David Stifter (Stifter 2009, 2010 and 2012), who has reached the conclusion that only one of them, a graffito on a tile found in Grafenstein near Klagenfurt in southern Austria, dated to the second century CE, is likely to contain any genuine Celtic material. Historical sources and some onomastic evidence document the presence of Celtic speakers further east, in Anatolia, but there is as yet no surviving text written in any variety of Celtic known from this area.22 The evidence for Germanic languages south of the Alps in our time period centres around inscriptions on two of a cache of twenty-six bronze helmets found near Negova, in what is now 20 21 22

8

RIG L97, see Mullen (2013: 118) on these texts. See also Prosdocimi (2006) for other possible early Celtic texts in Italy. For a collection of material relating to Celtic from Anatolia, known as ‘Galatian’, see Freeman (2011). Eska (2013) adds some further pieces of evidence, and assesses the Galatian material linguistically, concluding that the language is not far removed from Gaulish.

Introduction

eastern Slovenia. The helmets and the script of the inscriptions are both of a recognisable northern Etruscan type, dated to the fifth to fourth centuries BCE, although the cache itself was made later, possibly in the first century BCE. The text on the two helmets is variously interpreted, but one inscription (normally known as Negau B) contains a recognisable Germanic name, Harigasti.23 Although the earliest runic inscriptions do fall within our time period, and are found over a wide area of northern Germany and southern Scandinavia, none can plausibly be claimed to be outside the normal range of speakers of these languages. There are only a few inscriptions in the Indo-European languages of Italy other than Latin that have been found outside the Italian peninsula. McDonald and Clackson, in Chapter 4 of this volume, discuss the case of a silver cantharus found in Alesia which may have an Oscan text engraved on it. A second notable example is Umbria 3 (in Crawford et al. 2011), a bronze cooking pot with a female personal name written in both Latin and Umbrian dating from the middle of the first century BCE. Its provenance is recorded as ‘Hongrie’ (‘Hungary’) by its earlier owner Wilhelm Froehner, who had purchased the object on the antiquities market. This was ingeniously emended to ‘Ombrie’ (‘Umbria’) by Michel Lejeune, but Michael Crawford sees no reason to doubt Froehner’s original attribution. Oscan lettering has also been found on amphorae exported to Spain (María José Estarán Tolosa forthcoming). There is better evidence for mobility within Italy and Sicily. Chapters 6 and 7 by Tagliapietra and Zair in this volume discuss the case of Oscan inscriptions from Messina associated with the Mamertini. Messina is geographically close to Oscan-speaking areas of southern Italy, but the self-designation of the Mamertini tallies with literary sources and allows us to recognise them as a band of mercenaries from Campania, who took control of the port in 280 BCE. Of the other Sabellic languages, it is worth mentioning the case of two very short fifth/fourth-century24 BCE 23 24

See Markey (2001), who also reads Celtic words on the other helmet, Negau A. According to some accounts these inscriptions are dated to the fourth century, a hundred or so years later than other South Picene inscriptions, although Crawford calls into question the late date (Crawford et al. 2011: 249).

9

James Clackson et al.

South Picene texts, both written on bronze helmets, which are separated by over 300 km from the rest of the South Picene corpus. One, Interpromium B (in Crawford et al. 2011), was found in a grave in Bologna, the other, Interpromium A, a warrior’s tomb in Apulia; the interpretation of both texts is uncertain. Moving to the Indo-European languages of ancient Anatolia, the most widespread of the first-millennium languages is undoubtedly Carian. The Carian language, which employs its own adaptation of the alphabet, is attested in around thirty texts from Caria itself but a larger number, around 170, from Egypt (Adiego Lajara 2007: 17–18), with a further bilingual text from Athens and one other from Thessalonike. Many, but not all, of the Egyptian texts originate from Carian mercenaries, who, according to Herodotus (2.151 f) were recruited by the Pharaoh Psammetichus I in the mid-seventh century BCE, and who settled originally in the Nile Delta, later moving to Memphis. Carian graffiti are found at Abu Simbel in the far south of Egypt, 1,000 km from the Mediterranean, and Carian names occur alongside Greek in the well-known inscriptions left by Greek mercenaries there (MeiggsLewis, GHI 7 (a-g) = Syll.³ 1). It is furthermore likely that the reference to ἀλόγλοσοι (i.e. ἀλλόγλοσσοι), ‘speakers of other [i.e. non-Egyptian] languages’, distinguished from Αἰγυπτίοι in one of these Abu Simbel texts (GHI 7a line 4), refers to the contingent of Greek and Carian soldiers fighting together.25 Speakers of Iranian languages were frequent visitors to the Eastern Mediterranean, especially during periods of military conquest under the Achaemenid Empire in the sixth to fourth centuries BCE and subsequent Parthian Empire (third century BCE to third century CE). The use of Aramaic as the lingua franca of Persian rulers means that comparatively few inscriptions in Iranian languages are found in areas further west. Most notable are inscriptions featuring Old Persian cuneiform alongside other languages erected during the reign of Darius the Great, near what is now the Suez Canal.26 The Parthian Empire left its mark in Parthian and Middle Persian graffiti,

25 26

See discussion at SEG 51, 2201. See Schmitt (2009) for these texts.

10

Introduction

ostraca and parchment fragments excavated in the border town of Dura Europus (Frye 1968). We turn now to texts in other language families than IndoEuropean. The spread of Semitic varieties in our period is headed by the extraordinary array of Phoenician, and later Punic, inscriptions all around the Mediterranean, and Aramaic texts in the Eastern Mediterranean and further east into the continent of Asia.27 Many Phoenician inscriptions are closely linked to their colonies in Cyprus, southern Spain, north Africa and the islands in the west Mediterranean, but Phoenician texts are also found in areas where there was no established Phoenician settlement, such as the eighth-century BCE Karatepe bilingual, in Phoenician and hieroglyphic Luwian, from Lycia (now southern Turkey), or the fifth-century bilingual Pyrgi inscriptions from central Italy (where the other language is Etruscan). Such texts bear testament to the use of Phoenician as a language of international trade in the first half of the first millennium BCE. Since Aramaic and Hebrew are attested in biblical and other texts surviving in manuscript tradition, we have not included them in this survey of fragmentary corpus languages, except to point to David Noy’s (1993–5) collection of Jewish inscriptions, many of which include Hebrew tags, which in part can be used to show the spread of speakers of these languages to the west, particularly after the Jewish wars of the first century CE. It is also worth noting the continued use of the Palmyrene variety of Aramaic among Roman auxiliaries including the texts found outside the Porta Portuensis in Rome, which are discussed in detail by Elder in Chapter 11 of this volume. The languages of the Northwest Caucasian family had not been thought to be recorded in writing before the modern period until a recent article suggested that certain ‘nonsense’ texts on Athenian Greek vases, associated with depictions of Amazons and Scythians, were in fact representations of early Northwest Caucasian varieties, related to the later attested languages Abkhaz, Abaza, Kabardian (also called East Circassian), Adyghe (also called West Circassian) and Ubykh (Mayor, Colarusso & Saunders 2014). In a study of thirteen different inscriptions (with 27

Phoenician and Aramaic texts are gathered in Donner and Röllig (1962–2002).

11

James Clackson et al.

a total of 136 letters between them, and no individual string longer than twelve letters), the authors find elements corresponding to Northwest Caucasian verb phrases in four ‘nonsense texts’, and meaningful elements such as ‘Worthy of Armour’, ‘With Metal’ and ‘Hot Flanks’, which they take to be personal names, in eight of the vase inscriptions.28 The authors of this study note the presence of non-Greeks among the community of potters in Athens, but they do not go so far as to suggest that the non-Greeks were themselves responsible for inscribing the texts, which are taken to show Greek representations of non-Greek sounds. If Mayor, Colarusso and Saunders were correct, these inscriptions would reflect a striking departure from other Greek vase inscriptions, which do not represent foreign languages, even when intended for a foreign market.29 There is good reason, however, to remain extremely cautious about these results, given the number of different languages to which the nonsense texts have been compared (three different branches of Northwest Caucasian, Indo-Aryan, Ossetic and Georgian) and the semantic latitude that is possible in their interpretation. Moreover, Alexei Kassian has subjected the Northwest Caucasian material to a close examination and has cast doubt on the accuracy of the reconstructions (Kassian 2016).30 None of the other fragmentary corpus languages that are represented around the Mediterranean basin during our time period can be securely connected to known languages or language families.31 Because of this, the written texts are not fully understood, although it is possible to ascertain the meaning or purpose of a number of surviving inscriptions. This means that sometimes it 28

29

30

31

Mayor, Colarusso and Saunders (2014) also suggest that three nonsense texts might be respectively representing an Indo-Aryan or Eastern Iranian language, an Iranian language connected to Ossetic, and Georgian. The Ossetic and Georgian cases only refer to sequences of three letters; the Indo-Aryan example is the name Βαρκιδα which the authors take to mean ‘princess’, although noting that an inner-Greek explanation is also possible (Mayor, Colarusso & Saunders 2014: 474 n. 106). See Chapter 4 of this volume by McDonald and Clackson for the representation of Greek dialects on vase inscriptions. Kassian also objects to the reconstruction of three separate branches of Northwest Caucasian in the middle of the first millennium BCE, but this objection rests on lexicostatistics which is itself open to question. There is an ongoing debate about the connections between the language of the Iberian inscriptions and Basque (see Simkin 2012: 90), and the languages of the Libyan inscriptions and Berber (see Kerr (2010: 22f) for references and scepticism).

12

Introduction

is difficult to judge the status of inscriptions found outside the normal range of the language in question. Are these texts actually in the same language, or are they related dialects, or are we dealing with two different languages written in similar scripts? Where there is only a small corpus, and the surviving texts are short, these questions cannot be decisively answered, but for better attested fragmentary corpus languages, it is easier to reach a conclusion. Thus, for example, for Etruscan, with a corpus of around ten thousand texts, we know enough to judge that the famous stela found in Lemnos (and now joined by a second Lemnian inscription; see De Simone (2011)), was not left by an Etruscan colonist or traveller, but actually represents a closely related but separate language. Etruscan further gives us some of the best examples for mobile speakers of the language of any of the fragmentary corpus languages. A number of Etruscan texts have been found away from Etruria, Campania, the Po Valley and adjoining areas where the majority of Etruscan speakers lived.32 Texts in Etruscan have been found in North Africa, Corsica, Southern France, the Greek island of Aegina and Egypt.33 All but one of the texts from Corsica and Southern France are on pottery, mostly indicating the names of their owners (thus ET Cs 2.1–2.23 , Na 2.1–2.4). The exception is a fifth-century lead tablet from Pech Maho, in Aude, which has a trader’s letter in Greek (SEG 48: 1308) on the reverse of an Etruscan text (the two texts appear to have been written at different times and to convey separate information).34 The one inscription in Greece is a sixth-century vase with the owner’s name found in the Temple of Aphaia at Aegina (ET Gr 3.1). The Etruscan text found in Egypt is remarkable both for its survival and its content. The text was discovered only when a mummy, purchased by a nineteenth-century Croatian traveller to Egypt, was unwrapped, revealing that the wrappings were made 32

33

34

We shall not enter into the debate about how far the Etruscan settlements in Campania and Lucania, and in the Po Valley and Adriatic coast, represent the result of Etruscan colonisation from the eighth century onwards. For recent overviews, see Cuozzo (2013) (on Campania) and Sassatelli and Govi (2013) (on the Po and the Adriatic coast). Gemstones with short Etruscan legends have also been found in Sicily and Cyprus. In what follows Etruscan texts are cited following Meiser (2014). See Massarelli (2014) for a recent discussion of the Etruscan text.

13

James Clackson et al.

from a recycled linen book. The surviving text of the linen book has been revealed to be an Etruscan ritual calendar, composed probably at some time in the first century BCE. It is notable that another, roughly contemporary Etruscan text found out of place also has a religious content. The ‘Piacenza liver’, ET Pa 4.2, is thought to be a teaching aid or reference guide for a haruspex, giving twenty-five divine names inscribed on a bronze model liver, divided into sections. There are scarcely any other Etruscan texts from the Po valley from after the third century BCE, and the most likely explanation for the location of the model liver is that it was originally in the possession of an Etruscan priest attached to a contingent of Roman soldiers.35 Texts in the largely undeciphered language known as Libyan are concentrated along the North African coast, but have also been found in Morocco and the Canary Islands (Pichler 2008). Indeed, the very short texts found on rock faces in the Canary islands sometimes show script mixing of Libyan signs with Latin or Punic writing styles (Ulbrich 2015–16). These texts may have been left by occasional visitors to the islands from Africa, and it is not wholly clear whether they represent a westward migration of Libyan speakers or just the diffusion of Libyan amongst other writing systems. The final language which we shall consider in our survey is Iberian. Around 2,250 Iberian inscriptions are known, written in three different scripts (respectively Greco-Iberian, Southeastern Iberian and North Eastern Iberian) and spread over a wide area of eastern Spain from the Ebro Valley to Languedoc in Southern France. The presence of Iberian texts at trading ports such as Pech Maho and Emporion shows that speakers of Iberian circulated in the same spheres as Greek and Etruscan traders. But no text written in Iberian is identifiable as a clear outlier from a large Iberian-speaking (or Iberian-using) area. There is, however, a famous Latin inscription from Rome (ILLRP 515 = CIL I2 709 also known as the Ascoli Bronze), which records thirty Iberian cavalrymen settled in Asculum by the consul Gnaeus Pompeius Strabo in 89 BCE. Although the text contains no Iberian script, 35

van der Meer (1987: 18).

14

Introduction

most of the Iberian names are not adapted to Latin morphology, as shown, for example, by Nalbeaden Agerdo f(ilius), Cacususin Chadar f(ilius) or Balciabin Balcibil f(ilius); in all three instances the first name is not Latinised, and the second name is not inflected as a (Latin) genitive. Furthermore, onomastic elements of the names can be matched directly with names on Iberian texts, as, for example, Balciadin, for which the elements balke and atin occur in Iberian texts (see further Moncunill Martí 2017: 13). Apart from the use of f(ilius) to mark the second name as the patronymic, we are justified in seeing these names as preserving Iberian onomastic formulae. This last example leads on to the much larger question of the use of onomastic evidence to track ancient population movements. We shall devote some space to the question below, but before we do so, it is worth considering what emerges from this brief survey of texts in fragmentary corpus languages found ‘out of place’. The first point to notice is the sheer range of different scenarios that emerge. At one extreme there are languages, such as Carian, which are better attested in inscriptions from outside the Carian homeland than within, and at the other there are a number of other fragmentary corpus languages, such as Elymian, Eteocretan, Eteocypriot, Ligurian, Lusitanian, Lycian, Phrygian, Rhaetic, Sicel and Venetic, which have not appeared in this survey, since all their find-spots are clustered within a single area. There is a rough correlation between the level of attestation of a language in the epigraphic record and the likelihood that inscriptions will travel: there are more Etruscan inscriptions overseas because the Etruscans had a long tradition of writing with a wide range of uses. There are, however, numerous exceptions to this rule. The epigraphic habit was also present among Iberian speakers, but they did not leave written records away from home. Phoenician inscriptions are very widespread, but the total corpus is small, with under 300 inscriptions in total (Donner & Röllig 1962–2002). Many of the out-of-place inscriptions are on portable objects – pottery, silverware, helmets, lead tablets, a linen book – but travellers and settlers also left graffiti and monuments on stone. In some cases, these point to a settled and literate expatriate community, whether colonists or mercenaries. Most intriguing are those occasions where 15

James Clackson et al.

an individual far from home left a text in a language that must have been known to be incomprehensible to nearly all potential readers, such as the Gaulish inscription from Todi, Palmyrene inscriptions from Rome or South Shields in Roman Britain (on which see Mullen 2012: 1–5), or the Republican Latin inscriptions from Philae and Thebes discussed by James in Chapter 10. It is significant that these cases generally are bi-version bilingual texts, where the message is conveyed in an international language as well as the individual language. In her survey of bilingual inscriptions from the Roman west, Estarán Tolosa gives support to the idea that bilingual inscriptions are more likely to be associated with foreigners than with natives, counting forty inscriptions made by foreigners in her corpus compared with nineteen citizens, and all but one of the ‘citizen’ cases come from Etruscan tombstones (Estarán Tolosa (2017: 48); figures are from the tables at (2017: 649–54)). The fact that there are several bilingual texts made by Palmyrene recruits in the Roman Army shows that some ancient speech-communities fostered a particular tradition of using epigraphy to denote their identity. We could draw attention to this practice through attributing to Palmyrene a strong ‘ethnolinguistic vitality’,36 in order to distinguish it from others whose speakers do not appear to have such a particular attachment to their native tongue. But this will not explain all such cases of outof-place inscriptions, and in some cases, such as the Todi bilingual, the exact motivations for a traveller’s decision to leave a record of the language spoken at home must remain unknown. The previous survey of fragmentary corpus languages also helps to give a backdrop to the study of Greek and Roman epigraphy. The early and extensive spread of Greek epigraphy throughout the Mediterranean basin and beyond is unparalleled; while Phoenician settlements are attested over a wider area, and Phoenician’s use as a lingua franca is more extensive than that of Greek in the first centuries of the first millennium BCE, Greek colonists made much more extensive use of epigraphy for both public and private documents. There are entire genres of written texts, such as curse tablets and commercial documents, which are 36

On the term ‘ethnolinguistic vitality’ and its usefulness in ancient language study, see Mullen (2013: 69–71).

16

Introduction

attested in Greek but not in Phoenician. If the Greeks were exceptional in both their colonial activity and the degree of their adoption of writing, they were not unique. The record of the fragmentary corpus languages offers support to the proposal of Horden and Purcell and others of a networked Mediterranean. Members of ethnolinguistic groups other than the Greeks and the Phoenicians were also participating in colonial activities, manufacturing and transporting goods, or acting as hired soldiers or labour. As we have already seen from our survey, bilingual inscriptions involving fragmentary corpus languages are not uncommon, and biversion bilinguals that have been set up by private individuals are often associated with mobility or migration.37 It is a striking fact that nearly all bilingual inscriptions from the Mediterranean area during our time period involve Phoenician (or Punic), Aramaic, Greek or Latin as one of the two languages. Estarán Tolosa’s (2017) book collects all bilingual inscriptions from the western Mediterranean involving Latin as one of the languages, including coin legends, and the total comes to around 150 bilingual texts. There is a much smaller number of bilingual texts involving Phoenician/Punic and one of the other languages covered in Estarán Tolosa (2017) (Etruscan, Umbrian, Oscan, Venetic, Libyan, Lusitanian, Celtiberian and Iberian), although these include famous examples such as the Pyrgi gold leaves with Phoenican and Etruscan and the monumental Punic and Libyan text from Dougga (now in Tunisia). There are also a handful of bilingual texts involving one of these languages and Greek, such as the very short Greek–Oscan text found on a pottery lamp from Pompeii (Pompei 90 in Crawford et al. (2011)). It is striking that there are, however, no bilingual texts featuring, for example, Iberian and a Celtic language, or Etruscan and an Italic language other than Latin. In the eastern Mediterranean we see a similar picture. There are numerous bilingual texts involving Greek and other languages, including Lydian, Carian, Phrygian, Lycian, Sidetic, Eteocypriot, Aramaic and related Semitic languages and varieties of Egyptian, and 37

We use the term ‘bi-version’ to describe a bilingual text which has fundamentally the same text given in two languages, following Mullen (2013: 84–7).

17

James Clackson et al.

a smaller number of Phoenician and Aramaic bilinguals, with Lydian, Lycian, Assyrian and Eteocypriot, for example. The language policy of the Achaemenid Empire offers some exceptions to this rule, where Old Persian appears alongside hieroglyphic Egyptian, Elamite or Assyrian cuneiform, in various monumental or prestige inscriptions; the small number of Carian–Egyptian bilinguals left on the tombs of Carian mercenaries in Egypt are the only other exceptions. The distribution of bilingual texts consequently helps frame our consideration of the ancient fragmentary corpus languages and highlights the status of Phoenician, Greek and Latin as supranational languages, which appear to have been used by ancient migrants in ways in which other ancient languages were not. The special status of Greek and Latin in the ancient Mediterranean is also evident from the adoption of Greek and Roman naming systems by other peoples, as we saw in the case of the Ascoli Bronze above.38 In comparison with the general paucity of surviving records in fragmentary corpus languages, Greek and Roman epigraphy and papyri offer us a vast databank of onomastic evidence, which has allowed researchers to trace ethnic groups and indeed individuals. Current projects for digitisation and linking of various different onomastic and prosopographic databases promise further advances in the ability to track movements of families or groups and larger changes in populations over time.39 There is also a place for the ‘close reading’ of onomastic evidence, of the type undertaken by Maras, Rovai and James in Chapters 2, 8 and 10, particularly in the light of knowledge of fragmentary corpus languages as well using the results of large scale data collection. We will close this chapter with a couple of examples, to add to the detailed case studies in the individual chapters and to show how this close reading of onomastic evidence works in practice. Here we wish to stress in particular the need to incorporate knowledge of fragmentary corpus languages together with the prosopographic and onomastic data from Greek and Latin texts. Our first 38

39

See also Chapters 2 and 10 in this volume by Maras and James on onomastic formulations. See the project Standards for Networking Ancient Prosopographies: Data and Relations in Greco-Roman Names (http://snapdrgn.net/about, consulted 31 January 2018). On the Trismegistos project, see also Depauw (2016).

18

Introduction

example takes up Iberian naming practices, following our earlier discussion of the Ascoli Bronze (ILLRP 515 = CIL I2 709). As has been mentioned, thirty Iberians are named in the text, coming from different areas of Iberia. The different origins of the Iberians in the text allow us insights into practices back in Spain. Alongside those on the text such as Nalbeaden Agerdo f(ilius), Cacususin Chadar f(ilius) and Balciabin Balcibil f(ilius),whose names make no sense in Latin, there are three Iberians from Ilerda, C. or Q. Otacilius Suisetarten f(ilius), Cn. Cornelius Nesille f(ilius) and P. [F]abius Enasagin f. These three soldiers have adopted Roman praenomina and nomina, although their fathers have native names, which have not been Latinised and are uninflected. The presence of the three Ilerdans on the Ascoli Bronze alerts us to the difficulties of using names as a straightforward tracking device of individual movements. Without the inclusion of their fathers’ names, Q. Otacilius, Cn. Cornelius and P. [F]abius would pass for Romans. They may have adopted their Roman praenomina and nomina already in Ilerda, and if so, it helps to explain some names which occur in monolingual Iberian texts, written in Iberian script. Note, for example, the presence of a koŕneli [on a first-century BCE inscription from Emporion],40 almost certainly to be taken as a Romanised Iberian, retaining his original script if not a traditional name, rather than a Roman trader who has gone native. Ilerda is situated less than 100 km from the same coast on which the port Emporion lies, closer than any of the other communities mentioned on the Ascoli Bronze that can be identified. It seems possible, therefore, that these texts bear witness to a process of Latinisation of Iberian onomastic practises in the first century BCE, which gradually spread inland from the coast. Greek and Celtic names also occur in Iberian and Celtiberian inscriptions and may similarly reflect adoption of other onomastic practices rather than language shift among settlers (Moncunill Martí 2017: 10).41 40 41

C.1.1 in Untermann (1975–97). There are also examples of shifts in the opposite direction from the fragmentary corpus languages of Spain. Note, for example, a case in a Celtiberian text from Botorrita (K.1.3 in Untermann (1975–97)) where parents with Latin/Greek names appear to have given their son a Celtic name: col. 4, l. 3–4: bolọra.kentiskue.melmanđos/ tiokenesos. uiriaskum (‘Flora and son, Melmanđos son of Diogenes, of the Viriaskoi’).

19

James Clackson et al.

Our second example of incorporating onomastic and linguistic evidence together concerns the presence of names from mainland Italy in Sicily and North Africa. The particular case of the Mamertini in Messina is discussed in Chapters 6 and 7 by Tagliapietra and Zair in this volume. As they show, this is a rare example where we can associate a particular group of inscriptions with a mercenary band known from literary sources, and see at first hand their choices of language, script and self-designation. As Tagliapietra and Zair highlight in their individual chapters, close attention to the orthographic details of the inscriptions can be revealing about how the Mamertini positioned themselves in relation to their neighbours, both Greeks and other Oscan speakers. It is worth emphasising here, however, how unusual the situation of the Mamertini was compared to that of most of the migrant peoples in the Mediterranean. Mercenaries from the Italian mainland who settled in Sicily and in North Africa can be traced from the survival of names in Greek texts from Entella, central Sicily, and in Punic documents from Carthage.42 Lejeune (1982) was confident that the names of the non-Greeks on the Entella tablets showed an original Oscan presence, although by the time of the texts the bearers of the names had become full citizens and thus were referred to by a combination of their own individual name and the genitive of the father’s name, such as Γναίος Ὀππίου, with an Oscan name for both son and father, or Θεόδωρος Μαμου, with Oscan for the father but Greek for the son. A benefactor to Entella from the community of the Mamertini is also mentioned in the Entella texts. This individual, not being a citizen of Entella, bears an ethnic adjective, and retains the Oscan form of his name, with praenomen and nomen, although both of these were adapted to Greek morphology: Μίνατος Κόρουιος Μαμερτῖνος. If the other bearers of non-Greek names in Entella could be identified as specifically Oscan, as Lejeune thought, we would be able to see successive stages of integration of Oscan mercenaries into the Greek world: first, retention of a native onomastic system (as the Mamertini); next, 42

On the names in the Entella documents, see Lejeune (1982) and Clackson (2012); for the Carthaginian texts, see Bourdin and Crouzet (2009).

20

Introduction

adoption of Greek onomastic practices but Oscan names; and finally, assumption of Greek names as well. However, the onomastic evidence is, unfortunately, not so clear-cut. Clackson (2012) argued that the non-Greek names found in Entella were similar to the names from Carthage in Punic inscriptions discussed by Bourdin and Crouzet (2009), that is to say they are definitely from the Italian sphere, but they cannot be linked with confidence to a single language group. Many of the onomastic elements from Entella and Carthage are shared by a range of peoples from Italy – Oscans, Etruscans, Faliscans, Umbrians etc. The cohesion of the Mamertines as a group speaking the same linguistic variety and with their own ethnic identity may have set them apart from other mercenary bands with more diverse origins, and we should not assume that their cohesion and linguistic conservatism were necessarily shared by other soldiers of fortune. As we have seen above, Greek mercenaries, whose dialects reveal them to have come from Ionia and Rhodes, are attested together with Carians among the troops of Psammetichus I recorded at Abu Simbel. The chapters in this volume, we hope, give some further ways in which aspects of language and script use can contribute to a closer integration of a wide variety of linguistic approaches, from orthography to onomastics, with other studies of ancient migration. The current interest within historical linguistics in multilingualism and language contact, particularly in ancient societies, ties in well with current issues in research in migration and population movement, with its emphasis on the agency individual migrants have to perform a range of different identities or present themselves as representatives of two or more cultures. Historical linguistics has always been a rich source for metaphor for cultural interaction, from koinéisation to codeswitching, but we feel that linguistics has more to offer in the study of mobility and migration. The chapters in this volume are arranged in chronological order, insofar as this has been possible. Many of the chapters involve discussion of different texts separated in time sometimes by hundreds of years, although united in a common theme. The reader should consequently be aware that a strict chronological sequence will sometimes be broken, but this should cause few problems. 21

James Clackson et al.

The rough chronological sequence of the chapters also has the happy outcome that the geographical focus of the first and last chapters is central Italy, while the authors of the intervening chapters move the story to the south of Italy and then to Italians and others in the east, in Delos, Pannonia and Egypt.

22

C HA P T E R 2

I N T E R E T H NI C MO B I L I T Y A N D I N T E G R ATI ON IN PRE- ROMAN ET RURI A The Contribution of Onomastics Daniele F. Maras

Introduction The presence of foreign roots, components, and full names in the Etruscan nomenclature is ubiquitous from the earliest documentation. This finds a possible explanation only in the existence of continuous and prolonged contact and human mobility within the Etruscan communities, with integration of foreigners into the Etruscan society at different levels since the earliest periods. Scholars have noted different social trends through time. As regards the earliest available documentation up to the central decades of the seventh century BCE, the occurrence of Latin and Italic names among the members of the Etruscan elites that had access to writing seems to be the result of the phenomenon of synoecism, which brought about the constitution of proto-urban communities often including mixed populations with diverse ethnic origin. In the late seventh and sixth centuries, inscriptions in funerary contexts show that the integration of foreigners at a high social level continued, thus demonstrating the existence of an open and vital society, particularly in the cities of southern Etruria. On the contrary, from the fourth century BCE onwards, the vast majority of foreign elements in the name formulas pertain to people of low social status. In addition, even in the case of members of the elite bearing names of Latin or Italic origin, it is not clear whether these derive from recent immigration or rather from earlier integration. However, the documentation is uneven and rather scattered in time and places. As a result, some chronological and geographical discrepancies affect our ability to gather complete information. In fact, 23

Daniele F. Maras

in the seventh and sixth centuries BCE, the largest collections of onomastic material are from Caere and Veii, where around one hundred funerary and votive inscriptions have been found. In the late archaic period (sixth and early fifth centuries BCE), the primacy shifts to Volsinii, whose necropoleis present around one hundred and fifty inscriptions. Finally, from the Hellenistic period (especially third and second centuries BCE) there is an overwhelming majority of inscriptions on funerary urns and tombs of Clusium and Perusia, which overshadows the number of epigraphic documents found elsewhere in Etruria. Therefore, general patterns and trends have to be carefully evaluated both in the local contexts and at a regional scale and considered as provisional, pending further documents and information. In the following pages, the formation and historical development of the nomenclature system in Etruria and central Italy will be reviewed, starting from the earliest documentation and taking into consideration Roman nomenclature, which is better known than Etruscan. In this framework, we can pinpoint names and formulas escaping the regular pattern and/or including nonEtruscan components, with special regard to cases of interethnic contact and personal mobility. Finally, the classification of the foreign components in the Etruscan onomastic documentation provides a general outline for future studies. What’s in a Name (Formula)? The earliest known name formula consisted in a single personal name, called an individual name. As far as we know, this simple formula was shared originally by most cultures of the ancient Mediterranean and Europe, including the peoples of ancient Italy.1 This fact was already acknowledged by ancient antiquarians and grammarians, as Varro speculated that the earliest figures in Roman history had only one name, such as Faustulus, Romulus, and Remus2 – to whom Procas, Amulius, and Numitor can be added too. Appian states that the Romans later took 1 2

Motta (2009:300); Salway (1994:124–5); Maras (2017:71). Varro apud Probus, de Praenominibus 4.1.

24

Mobility and Integration in Pre-Roman Etruria

a second name (gentilicium) and then a third (cognomen) for easier recognition,3 thus creating the well-known tria nomina of the historical Roman nomenclature. Remaining in the realm of literary sources and legend, the earliest figures having a binomial formula are Titus Tatius and Numa Pompilius, both belonging to the Sabine culture.4 However, it has been noted that the second element of their name formulas was not inherited by their children, as shown by Hersilia, Titus Tatius’ daughter, and Pompo, Pinus, Calpus, and Mamercus, Numa’s sons.5 In this regard, it is also significant that the derivate nomenclature of things associated with these men is based on their first names, as in the case of the sodales Titii and the curia Titia.6 In actuality, the second element of the nomenclature of the Sabine kings was a patronymic, as already noted by Dionysius of Halicarnassus, who records the name of Numa’s father as Pompilius Pompon.7 The form of such patronymic adjectives as Tatius and Pompilius was not different from the aforementioned derivate nomenclature (curia Titia ~ *Hersilia Titia). The earliest Roman-Sabine formulas, therefore, do not differ from the majority of ancient onomastic systems, from the Near East to Europe, which require the use of patronymics in the case of free people, at times in the form of adjectives, such as, for instance, Ajax Telamonios.8 According to Plutarch, the need for an inheritable component of the formula arose one generation later, when Numa’s sons gave 3

4

5 6 7 8

Appian, Pr. 51. According to Probus and Priscian, Romans acquired binomial naming formulas because of their intermarriage with the Sabines in Romulus’ time, when they simply juxtaposed personal names expressed in both languages; Salway (1994:124–5); Solin (2009:275 n. 77). In the perspective of this study it is not relevant whether historical figures similar to Titus Tatius or Numa Pompilius actually existed or not. Their name formulas, however, show an intrinsic reliability when compared with what we know of the earliest nomenclature of central Italy. It is worth considering that Roman authors and their sources probably acknowledged the antiquity of the formation of names like these, which followed old patterns no more in use, but at times misunderstood them. In this regard, the case of the names of Mezentius and his son Lausus in Vergil is emblematic, when compared with the seventh-century inscription of Laucies Mezenties (in genitive) from Caere; Briquel (2016:174–80); see footnote 70 below. Plut. Numa 21; see Facchetti (2007:115–16). Peruzzi (1970:41–2); Colonna (1977:181). D.H., 2.58. Maggiani (2000:249–50); see also Maras (2003:240–3).

25

Daniele F. Maras

origin to several of the gentes of Rome: the Pomponii from Pompo, the Pinarii from Pinus, the Calpurnii from Calpus, and the Marcii from Mamercus, as well as the Aemilii.9 Consequently, the use of the patronymic adjective shifted from the simple filiation to the lineage, thus creating a durable family nomenclature.10 If we move to the safer ground of epigraphic evidence, it is remarkable that the process of formation of a nomenclature including an inheritable component (gentilicium) developed in central Italy between the late eighth and the mid-seventh centuries BCE.11 Literary tradition and epigraphic evidence strikingly converge in pointing to the late eighth century BCE – the alleged age of Numa – as the starting point of the gentilician nomenclature system. Not surprisingly, the same period is known as a time of important social changes, with the development of proto-urban settlements in southern Etruria as a result of synoecism, and the birth of the Orientalizing aristocratic network, involving trade overseas, a gift-exchange system, and the spread of the ideology of symposium among the elites.12 Not least, this is also the period that witnessed the introduction of writing in Italy as one of the principal status symbols of the Orientalizing culture.13 In southern Etruria, this changing social environment, along with the economic structure of the proto-urban settlements, created a need to bind the right of inheritance of land to specific individuals and their families. In other words, the ownership of land by the archaic pater familias implied the legal right to hand down family estates to his heirs, thus perpetuating his legal personhood,14 which was of central relevance for the family’s relationships with clientes, sodales, and servants too.15 Hence the reason for a new onomastic system that incorporated an inheritable component, namely the nomen gentilicium. In fact, 9 10

11

12 13 14 15

Plut., Numa, 21.1–2; Solin (2009:286). Incidentally, also derivate nomenclature changes from this period onwards, as the curia Hostilia, named after king Tullus Hostilius, demonstrates; Colonna (1977:181). Colonna (1977:176–80); Marchesini (1997:154–9); Smith (2006:17–20 and 158–9); Wallace (2008:79). Riva (2006). Maras (2015:203–9). Capogrossi (1990:11–30); Capogrossi (1994:57–68); Maras (2017:74–5). Colonna (1977:185–8); Maggiani (2000:249).

26

Mobility and Integration in Pre-Roman Etruria

family names or, better, ‘names of lineage’ are regularly linked to heritable legal rights, such as property and social relations of the clan,16 as well as dynastic rights of succession, for example, in the case of aristocratic dynasties, such as the Bacchiadai and the Peisistratidai in Greece.17 Most probably, the Etruscans were the first to use this new nomenclature, which later spread to Rome and the Latins as early as the seventh century BCE. Etruscan inscriptions with binomial formulas predate Latin attestations and hark back to the very beginning of the seventh century BCE.18 As mentioned above, Etruscan and Roman gentilicia took the form of patronymic adjectives, with reference to a historical or legendary forefather, as in the case of the Romilii and the Hostilii respectively from Romulus and Hostus, the Valerii and the Pomponii from the Sabine Volesus and Pompo, and probably the Iunii from the Faliscan Iuna.19 As a consequence, investigating the origin of the gentilicia might shed light on early phenomena of interethnic mobility, possibly dating to times preceding the written documentation. Originally, the formation of Etruscan gentilicia was basically identical to that of Latin names through the use of an adjective suffix (-ius in Latin vs. -na and variants in Etruscan): e.g., Rumelna, corresponding to Romilius; Acvilna, corresponding to Aquilius; but also Velχana, deriving from Velχa, and so on.20 The historical development of the Etruscan system, however, was more complex and diverged from what we know of later Roman nomenclature. It is therefore worth reviewing the principal steps of the historical development of Etruscan nomenclature. Nomenclature in Progress: Individual Names Many early Latin and Etruscan inscriptions present monomial formulas in the form of individual names used without a patronymic or gentilicium. The relative frequency of monomial 16 17 18 19 20

Smith (2006:158–9). Salway (1994:125, n. 6). Colonna (1977:176; 1995:332–9); Maras (2017:76). Rix (2009:499). de Simone (1989:263–5 and 278–9).

27

Daniele F. Maras

formulas, however, does not necessarily imply that the gentilician nomenclature was not widespread in this early period. As a matter of fact, the vast majority of inscriptions belong to non-official contexts, often involving gift-exchange, marriage gifts, or funerary goods. In these cases, we should not necessarily expect full official names in the graffiti.21 In Etruria several attestations are known in which the name of the receiver of a gift – usually a woman, possibly a bride – has a binomial formula, whereas the giver is indicated only by a single name, corresponding to a well-known praenomen: ET Cr 3.20 (Caere, late seventh century BCE): mi(ni) aranθ ramuθasi vestiricinala muluvanice, ‘Aranth gave me to Ramutha Vestiricinai’. ET Cl 3.2 (Clusium, late seventh century BCE): mine viku mulu{ve}neke arpaś kamaia, ‘Viku gave me to Arpa Kamai’.

Most probably, the family context in which the gift took place made it unnecessary to express the family name of the giver, which presumably at times was identical to that of the receiver. In other cases, a name in the form of an adjective in -na, formally identical to a gentilicium, is written in isolation on an object: ET Cl 2.1 (Clusium, mid-seventh century BCE): plikaśnaś.22 ET Cr 2.4, 18–20 (Caere, mid-seventh century BCE): (mi) karkanas (qutum/θahvna/spanti).23

Possibly, these attestations demonstrate that in the early seventh century, gentilicia were used from time to time with reference to the entire gens rather than to an individual, presumably in the social context of gift-exchange among different families. All in all, the sparse epigraphic evidence available for the late eight and early seventh centuries is not reliable enough to investigate the frequency and consistency of different forms of nomenclature at the earliest period of binomial formulas.

21 22 23

Colonna (1977:176–7). Marchesini (1997:155–7). Briquel (2016:147–56); Maras (forthcoming); see also Morandi (2004:263).

28

Mobility and Integration in Pre-Roman Etruria

Nomenclature in Progress: Patronymics Some Etruscan inscriptions present binomial formulas whose second element cannot be regarded as a proper gentilicium. This is the case, for instance, of a number of graffiti on luxury goods found in the tomb Regolini-Galassi at Caere (early seventh century BCE), which record larθia velθurus in place of the usual simple larθia.24 It seems evident that the precious silverware was marked with the name of its aristocratic owner and, in some special cases – significantly, on precious exotic vessels – with his filiation: ‘belonging to Larth, (son) of Velthur’ There is no evidence, however, whether Larth had no gentilicium, or omitted it in a family context, where everybody knew it. In a similar funerary family context, albeit later, the figure of a young rider depicted in the Tomba delle Iscrizioni of Tarquinia (end of the sixth century BCE) – presumably the scion of the gens Matve that owned the tomb – is labelled laris larθia, ‘Laris (son) of Larth’, omitting the unnecessary gentilicium. Two consequences derive from these and similar attestations: on the one hand, the early Mediterranean (and Greek) system of nomenclature including name and filiation apparently survived in Etruria within family traditions; on the other hand, inscriptions on funerary goods and from household contexts cannot be considered reliable for studying official name formulas. A different case is made by funerary stelae that stood outside the tombs, in a public environment, and are more likely to be inscribed with the complete names of the deceased. In this context, it is worth mentioning the evidence of the stela from Vetulonia (second half of the seventh century BCE), whose name formulas preserve a complex family relationship:25 ET Vn 1.1 (with a different reading): [mi a]ụṿẹleś θeluskeś tuśnutnâị[eś –] panalaś mini muluvaneke hirumi[n]a φersnaḷạś

In the third word preserved, the presence of the suffix -na has recently been recognized26 by reading a letter -n- added to or 24

25 26

Buranelli and Sannibale (2005); Marchesini (2007:42–3); Colonna and Sannibale (2015). Maggiani (2007); Agostiniani (2011a). Maras (2016:244); Maras (2017:75).

29

Daniele F. Maras

corrected over a precedent -a-, which makes the reading tuśnutni[eś] equally acceptable. In either case, the name includes the double suffix -na-ie- that is typical of gentilicia. Therefore, it is most likely that the precedent word θeluskeś is actually a patronymic, whereas the following [–]panalaś is a metronymic. All together the name is translated ‘Auvele Tusnutnaie, (son) of Theluske (and) of a [–]panai’. The giver of the tomb is presumably his heir, whose identical gentilicium is coherently omitted: ‘Hirumina (Tusnutnaie, son of Auvele and) of a Phersnai’.27 From this earliest document of an official high-rank name formula, we can argue that its components were expressed in order of accumulation, starting from the original individual name, corresponding to the later praenomen, and then adding the patronymic, the gentilicium, and the metronymic.28 Two contemporary gift inscriptions record a patronymic in the form of a name in the genitive case, followed by a gentilicium in -na: ET Fa 3.2 (third quarter of the seventh century BCE): mi mulu laricesi p[–] [–]naiesi clinsi velθurusi larụs ruvries (in pertinentive case), to be interpreted as ‘I (am) a gift from Larice [–]naie (son) of P[–] (and his) son Velthur to Lar Ruvrie’. ET Ru 3.1 (last quarter of the seventh century): mini mulvanike venel rapaleś laivena[?], ‘Venel Laivena (son) of Rapale gave me’.29

Nomenclature in Progress: Notes on the Early Gentilicia As has been mentioned above, in origin gentilicia were formally identical to patronymic adjectives, deriving from an individual name through the addition of the suffix -na. As a consequence, in the earliest phase of the heritable nomenclature, there was the objective difficulty of distinguishing actual gentilicia from

27

28 29

Maras (2016:244, n. 30). See, however, Maggiani (2000:256), whose reading φersnalnaś implies that the giver belongs to a different gens. Maras (2017:75–6). Maras (2017:75); see also Maras (2009:370–1, Ru do.1). On the function and meaning of the pertinentive case, see Agostiniani (2011b), with different opinions.

30

Mobility and Integration in Pre-Roman Etruria

patronymics. This circumstance affects also our ability to interpret the names when there is no evidence whether the second element of the formula was heritable or not. There is, however, a particular class of inscribed vessels that provide evidence for the variability of gentilician nomenclature in the early seventh century BCE. This is a series of bucchero kyathoi decorated and bearing gift inscriptions, which were produced partly in Caere and partly in Vetulonia and have been found in diverse parts of Etruria, as far as Caere in the south and Volterra in the north.30 A number of these kyathoi were inscribed in the name of various members of the Paithina family, which most probably owned the workshop that produced the vessels. We know therefore the names of Venel Paiθina[?], [Velθ]ur Paiθinaie, and L[a]uχu[s] hie Paiθin[a]ś31 (other names are too badly preserved to be of any help). The inscriptions of two other kyathoi of the series present the names of the gens Purenaie32 and of a Pisna Puraie,33 who were possibly related. From these attestations it is arguable that in the context of the same family the suffixes of the gentilitia could vary in a range of different possibilities, with no apparent consequence in regard to their meaning. According to a hypothesis of Adriano Maggiani, the multisuffixation of some Etruscan gentilicia originated from the necessity of identifying actual inheritable family names from patronymic adjectives. In this regard, the evidence of the bucchero kyathoi might belong to an early phase of this system, when a number of suffixes were considered suitable for this purpose, according to the following scheme: Paiθe individual name: Paiθi-na patronymic adjective: Paiθi-na-ie, Paiθi-na-s gentilicium:

‘Paithe’ ‘(offspring) of Paithe’ ‘belonging to the offspring of Paithe’

Such a scheme, however, was not rigid and mandatory, as the many variants seen in the documentation demonstrate. A combination of

30 31 32 33

Maras (2013a:333–6); Maggiani (2014 and 2018); Maras (forthcoming). Respectively ET Cr 3.27, AS 3.6, and Vn 3.2. ET Cr 0.1. ET Cl 2.44; Maras (forthcoming).

31

Daniele F. Maras

further suffixes often occurred, such as -ie, -ra, -ice, -sie, and so on, at times even in the absence of the usual -na.34 In addition, some gentes followed the alternative pattern of making their own gentilicium by freezing the original patronymic in the form of the individual name of the forefather in the genitive case. Its equivalence to the patronymic adjective in -na is proved by the existence of parallel pairs deriving from the same name, as in the case of Hvul(u)ves versus Hvul(u) venas and Matves versus Matunas.35 The complexity of the name-making system in Etruria makes it necessary to keep in mind all the diverse possibilities when dealing with name formulas and new gentilicia and/or foreign names. The following section is devoted to investigating the different ways in which non-Etruscan people entered and were integrated into Etruscan society. Some Paradigms for the Integration of Foreigners in Etruria Among the earliest pieces of evidence for the integration of foreigners into Etruscan society are the well-known inscriptions of Larth Telicles36 and Rutile Hipucrates,37 who have a name formula whose second element is derived from a Greek name, respectively Τηλεκλῆς and Ἱπποκράτης. In these cases, as has been pointed out, we have no clue whether the second part of the formula was inheritable or not. There are thus three different possibilities for interpreting these names: 34 35

36

37

Maggiani (2000:252–7). See ET Ve 3.9 and 3.30 (hvuluves), Fa 3.5 (hvulves), Vs 1.100 (vhulvenas); ET Ta 7.13 and 7.19 (matves), Cr 1.130–9 (matunas). Such variability poses a serious limit to the possibility of speaking of ‘echte Gentilizia’ as opposed to ‘Individualnamen-’ and ‘Vornamen-gentilizia’ in the Archaic period, according to the theory of Rix (1963); see also Benelli (2011:193–5). ET OA 2.2 (third quarter of the seventh century BCE): mi larθaia telicles; Colonna (1977:134); Morandi (2004:531–2); Maras (2017:72). Note that Colonna (2004:81–2) assumes this is the name of a woman, followed by Marchesini (2007:47–8); cf. ET Vc 2.9 (mi larθaia maies); Marchesini (2007:76). ET Ta 6.1 (second half of the seventh century BCE): aχapri rutile hipucrates; Bagnasco Gianni (1996:173); Colonna (2004:82); Morandi (2004:239); Marchesini (2008:65–6); Maras (2017:72).

32

Mobility and Integration in Pre-Roman Etruria I. They belong to Etruscanized Greeks, whose original individual name was converted to a gentilicium, to be inherited by their offspring. II. They belong to Etruscan children of Greek travellers or migrants: in this case there is no way to know whether the second element of the formula is a patronymic or a gentilicium. III. They belong to the Etruscan offspring of Greek ancestors, going back two or more generations, whose names were converted into ‘Individualnamengentilizia’ and therefore inherited by their descendants.38

In the absence of further data, it would be unwise to choose among these possibilities that are all equally plausible. However, it is possible to gather additional information by investigating better known although later cases, which allow us to outline some possible paradigms for the inclusion of new elements into the Etruscan name system in case of migrations and human mobility. Model No. 1: An Immigrant and His Offspring Precious information is provided by the family tree described at length in the elogium of Laris Pulenas, written on the lid of his sarcophagus in Tarquinia,39 dating from the mid-third century BCE. The deceased declares himself to be son of Larce (Pulenas), grandson of Larth (Pulenas), and descendant40 of the Greek (immigrant) Laris Pules.41 Most plausibly, the name Pules (in genitive) has been interpreted as the Etruscan version of the Greek Πόλλης, a name borne by famous legendary and historical seers, which was particularly suitable for the (alleged) forefather of an Etruscan haruspex such as Laris Pulenas.42 What is especially relevant for our topic is that the Greek immigrant Πόλλης was integrated in Etruria and received an Etruscan name formula made of two elements: a regular praenomen Laris, followed 38

39

40

41 42

For a recent critical examination of the interrelated concepts of Individualnamengentilizia and Vornamengentilizia in recent studies, see Benelli (2011). ET Ta 1.17: l(a)ṛis pulenas. larces. clan. larθal. pap ̣ acs ̣ velθurus. nefts. pruṃs. pulẹṣ. larisal. creices . . . ; Heurgon (1957); Morandi (2004:390, 393–4). It is not relevant here whether the meaning of prums is ‘great-grandson’ or simply ‘offspring’. Furthermore, he adds that he is the nephew of Velthur (Pulenas). Heurgon (1957:117–21); Facchetti (2007:124–6).

33

Daniele F. Maras

by the original individual name converted into a gentilicium and transliterated into Etruscan with few adaptations (/o/ > u; -λλ- > -l-). The process corresponds to the first of the hypotheses presented above in regard to the earlier Greek immigrants Τηλεκλῆς and Ἱπποκράτης. In this case, however, we are lucky enough to have information on the following events, since, after settling down, Laris Pules had some descendants, who took their family name after him by adding the suffix -na: Pule-s > Pule-na-s.43 Although there is no evidence for this, it is theoretically possible that a different branch of the family preserved the original name with the -s ending, as in the aforementioned case of Matves and Matunas, both descending or pretending to descend from a Matve,44 whose ‘ghost’ in the form of an eidolon is depicted as receiving an offering in the Tomba delle Iscrizioni of Tarquinia.45 The paradigm of Laris Pules and his offspring allows us to investigate further cases of former immigrants who gave origin to Etruscan gentes. Among the elites of the archaic period, it seems likely that the Oscan name Vestirikis was used as the base for the gentilicia of Caere Vestiricina and Vestraces.46 Likewise the parallel case of Peticinas from Caere and Petaces from Bisenzio probably derived from an original Italic *Petíkis.47 In both cases, however, the Italic documentation is much more recent than the Etruscan attestations, casting some doubts on the relationship between the relevant names. We are probably on a safer ground when etymological and linguistic processes are involved, as in the case of the latearchaic name formula of Larth Purzes son of Uśele in Volsinii. In fact, this is possibly evidence of the earlier immigration from Umbria of a *Purzie (< *Purçie, cf. Latin Porcius).48 The immigrant might have been his father Uśele – if the Etruscan 43

44 45 46

47

48

It is worth noting that in the transition from pules to pulenas the final -s deriving from the adaptation of the Greek name was treated as the ending of an Etruscan genitive case; Heurgon (1957:113), who transcribes the name as pule; de Simone (1970:234). Morandi (2004:302). Colonna (2016:135); Maras (2018a:96). ET Ve 3.15, 3.40 and Cr 3.20 (vest(i)ricina), Cr 2.115 (vestraces); Morandi (2004:194–6). ET Cr 2.30 (peticinas), AT 2.12 (petaces); de Simone 1972; Morandi (2004:369–70); Colonna (2013:13). See also Marchesini (2007:70–1), with a different opinion. Colonna (2000: 281–2).

34

Mobility and Integration in Pre-Roman Etruria

praenomen had been added, as in the case of Laris Pules – or an elder relation. In either case, there is a striking chronological coincidence with the supposed Umbrian ancestor of Porsenna, king of Volsinii and Clusium at the end of the sixth century BCE according to the sources,49 whose Etruscan name was probably *Purzena, which might be referred to the same lineage. Moving to lower social classes, we might expect to find a relevant number of immigrants among craftsmen in all times. It is not surprising, therefore that the name formula of the vase painter Arnθe Πράξιας (early fifth century BCE) fits the paradigm of Laris Pules, but the foreign component of the name is still written in Greek characters.50 Unfortunately, we have no clue if Πράξιας ever had offspring in Etruria51 and, if so, whether they were integrated into Etruscan society or not. We are luckier, however, in the case of another vase painter, who signed a red-figure crater found in Civita Castellana and dates from the third quarter of the fourth century BCE: ET Fa 7.1: sẹrṭ(u)r purφiṛnạs52

The vase is referred to the style that John D. Beazley defined as the ‘Campanizing Group’ and attributed to the work of immigrants from southern Italy who settled in south Etruria and started new workshops. In particular, Sertur Purphirnas belonged to the second generation of Campanizing painters, who took over their masters’ workshops (and it is worth mentioning that it is a well-founded anthropological standard that craftsmanship is handed down in family contexts from father to son). It is, therefore, not surprising that his gentilicium betrays a Greek origin, since he is the heir of one of the immigrant painters from Campania who started the new fashion of vase painting. Most probably, his father’s name was *Πορφύρος or Πορφύριος (vel sim.)53 and 49 50

51

52 53

Plin., Nat., 2.140; 36.91–3; Colonna (2000:278–9). The Etruscan praenomen was even adapted to Greek phonology by adding a final -e. See also Bruni (2013:300–5), with a different opinion and further bibliography. Unless we take into consideration the unlikely possibility that either a Caia Praśiś at Perugia (ET Pe 1.24) or a Prasalu at Spina (ET Sp 2.84) might be much later relations. For the reading, see Maras (2014:471–2). In the same period, see, for instance, the Acharnanian Πορφύρα (fourth–third cent. BCE) and Πορφυρίων (end of the fourth cent. BCE), as well as the Athenian Πορφυ[ρώ] (fourth

35

Daniele F. Maras

had been Etruscanized as *purφire before the mid-fourth century BCE, presumably adding an Etruscan praenomen. Consequentially, his offspring derived their gentilicium from his name by adding a -na, exactly as in the case of the Pulenas from Laris Pules. Model No. 2: Names of Function and Mercenaries One more sphere in which we find foreigners partially or entirely integrated into Etruscan society is that of mercenaries, who are known to have had a special status in ancient cultures.54 Scholars have often investigated the case of Macstrna, depicted in one of the wall paintings of the Tomb François of Vulci (ca. 340 BCE), thought to be equivalent to Mastarna, the Etruscan name of Servius Tullius.55 According to the legend handed down by the emperor Claudius, Mastarna was the right-hand man of Caelius Vibenna, an Etruscan warlord from Vulci, who came to Rome at the time of Tarquinius Priscus. After Caelius’ death, Mastarna took over the command and settled on the Caelian hill, later becoming king of the city with the name of Servius Tullius.56 For a long time scholars have acknowledged that the alleged Etruscan name of the would-be-king derives from Latin magister, in the sense of ‘captain (of the army)’ (magister populi). Massimo Pallottino recognized the derivative value of the suffix -na that provides the name with the meaning of ‘belonging to the magister’, that is to say ‘vice-captain’, ‘first officer’ or, possibly, ‘attendant’. In other words, although perceived as a personal name by the source of Claudius,57 Macstrna/Mastarna was originally a name of function in the context of the military fellowships (sodalitates) of the archaic period.58 This might be the reason why in the painting of Tomb François the monomial name formula of Macstrna stands almost alone among the binomial formulas of the other characters depicted.

54 55

56 57 58

cent. BCE) and the later Leukadian Πορφυρ[–] (second cent. BCE ?); Fraser and Matthews (1997:372); Osborne and Byrne (1994:337); Maras (2014:472, n. 76). Colonna (2013:7–9). See lastly Torelli (2011:230–2); Massa-Pairault (2013–14); Torelli (2016:7–11); Maras (2018a:101–2). Sources and comment in Briquel (1990). Most probably an early annalist, such as Fabius Pictor; see Maras (2010:190–2). Maras (2018a:101–2).

36

Mobility and Integration in Pre-Roman Etruria

The only other single name in the scene is that of Rasce, one of the companions of Mactrna. Significantly, this name has been associated with the Etruscan word rasna,59 once believed to be the ethnic name of the ‘Etruscans’, but later shown to correspond to Latin populus, both in its original sense of ‘army’ and in the later political sense of ‘people’.60 In this context, it is not by chance that the couple macstrna ~ rasce finds parallels in the two elements of the Latin military rank magister populi: if macstrna means ‘belonging to the magister’, rasce might be translated as ‘belonging to the populus’, that is to say ‘to the army’.61 For our purposes, it is most relevant that both in Latin literary sources and in earlier Etruscan iconographic attestations, these names of function are used as personal names. It would be intriguing to investigate if this type of nomenclature became heritable and was integrated into the Etruscan gentilician system. In the late archaic period, there is a comparable case of a foreign name of function transformed into an Etruscan personal name by adding an adjective suffix. This is the case of Carucra, name of a devotee who made an offering to Hercle in the southern sanctuary of Pyrgi,62 which is evidently derived from non-Attic/Ionic Greek κᾶρυξ.63 The close parallel with Macstrna makes it possible to translate the name as ‘belonging to the herald’, presumably in a diplomatic context of the mid-fifth century BCE.64 Remarkably, the later onomastic evidence from Caere includes a gentilicium Crucra that might be related to the earlier Carucra, if the first vowel could have been dropped as in the cases of malaχ/ mlaχ65 and, possibly, tamia/tmia.66 If this hypothesis can be 59 60 61

62 63 64 65 66

Pallottino, in Buranelli (1987:228–9). Rix (1984). Maras (2018a:102). In actuality, the form ras-ce is derived from the root ras-in parallel to the form ras-na, whose meaning is ‘people, army’; therefore, the relationship between these two terms might be more complex and requires further research. I heartily thank Luciano Agostiniani for sharing his thoughts with me on this and other topics discussed in this chapter. ET Cr 3.43. Colonna (2004:73–4); Maras (2009:358, Py do.6). Maras (2018a:101). Maras (2007:238–9, n. 11). Maras (2009:351, n. 1).

37

Daniele F. Maras

accepted, we could argue that the devotee of Pyrgi – presumably a Greek immigrant – had offspring still living in Caere some centuries later.67 In other cases, we are not so lucky as to have the name of the immigrant forefather. However, we can gather information from family names that seem to be derived from foreign words belonging to the military sphere. This is the case of Avele Arχusnas, owner of a family tomb of the late sixth century BCE in the necropolis of Crocifisso del Tufo at Volsinii, whose gentilicium derives from a personal name, *arχu-s(i)e, which is in turn constructed on the Greek word ἄρχων, ‘chief’.68 Similarly, the name of Laucie Mezentie,69 inscribed on a cup of the early seventh century BCE at the Louvre and corresponding to the legendary king Mezentius, mentioned in the Aeneid, might be derived from an Italic form *med-ent-yo- (less probably *med-ye-nt-yo-), with the meaning of ‘ruler’.70 In addition, the later Avle Meteliś, mentioned by the inscription of the Arringatore,71 has been convincingly related by G. Colonna to the archaic Latin word metellus (> metel(e)-ie), which originally meant ‘mercenary’.72 Finally, the gentilicium of Larθ Cupures son of Aranθ, inscribed on a late archaic cippus shaped as the helmed head of a warrior from the necropolis of Crocifisso del Tufo at Volsinii,73 could be derived from a laudatory attribute in the context of a mercenary army. In fact, it might be related to the Italic root *kupro-, ‘good, worthy’, according to an intriguing hypothesis of G. Colonna,74 in reference to the military ability of the deceased and his foreign provenience, shown by the use of the alphabet of Veii.75 The main objection to seeing an Italic connection is the three-syllable form 67

68 69 70

71 72 73 74

75

See, however, Colonna (2004:82–4), who compares crucra to a different Greek name, Κρόκος, thus removing any possibility of relationship with carucra. Candeloro and Colonna (2011:290, n. 50). ET Cr 2.149. de Simone (1991); Morandi (2004:308–9); Marchesini (2007:44–5); Briquel (2016:174–80, n. 70); Torelli (2016:10). ET Pe 3.3. Colonna (2014). ET Vs 1.113. Colonna (1993:19–20); Colonna (2000:282); see also Morandi (2004:147) and Marchesini (2007:89). Maggiani (2005a:45–7).

38

Mobility and Integration in Pre-Roman Etruria

of the name cupure-, which might be only indistinctly compared with the anaptyptic Paleo-Sabellic form qupíríh 76 and the farther Elymian name kupura/kupra.77 It is worth mentioning, however, that the anaptyxis could be entirely attributed to the Etruscan reception of a foreign name, as in the roughly contemporary cases of Herecele (< ῾Ηρακλῆς) and Pereceles (< Περικλῆς).78 In all the aforementioned cases, a name of function, an epithet, or a job name was converted into a gentilicium at the time of the integration of its foreign bearer into Etruscan society. It is not surprising that this phenomenon occurred mainly in the ambit of mercenary armies, which were one of the main causes of interethnic personal mobility in the ancient world. In my opinion, however, it would be wise to consider this paradigm in a broader sphere with reference to further cases of immigration and integration of foreign elements. Model No. 3: Provisional Business Integration A number of inscriptions of the archaic period have name formulas that apparently present two gentilicia, often with a -na- ending. Such formulas occur from time to time in funerary and instrumental inscriptions, as in Veii (late seventh century BCE: una uras pụpunas),79 Orvieto (early sixth century BCE: mi aveles vhulṿenas rutelna),80 and Pisa (early fifth century BCE: leθe kakuś papnie).81 The most relevant evidence, however, comes from the tesserae hospitales found in Rome, Carthage, and Murlo, which are documents of trade and social contact, often involving foreigners: ET AS 2.14 (Murlo, end of the seventh century BCE): mi avil[e –] [–]na.82 76 77 78

79 80 81 82

ST Sp AP 2. Agostiniani (1984–5:211–12). ET Vs 1.67; Cristofani (1974:315); cp. celeniarasi in ET AV 1.29, when compared to cliniiaras in ET Ta 3.2; Maggiani (1999:56). See also Morandi (2004:368) and Marchesini (2007:86), with different opinions. ET Ve 2.11; Maras (2012:53–4); Maras (2014:150–1). ET Vs 1.45; Rix (1963:379–83); Maras (2017:77). Maggiani (2006:334–5). Maggiani (2005b:162–3, n. 18); Maggiani (2006:321–2); Tuck and Wallace (2013:16 and 32–5, nn. 4 and 7).

39

Daniele F. Maras ET AS 2.15 (Murlo, end of the seventh century BCE): mi sp[urie] [–]aś vheiśalna[?].83 ET Af 3.1 (Carthage, first half of the sixth century BCE): mi puinel karθazies ṿẹsqu[?]na.84 ET La 2.3 (Rome, S. Omobono, first half of the sixth century BCE): araz silqetenas spurianas.85

In consideration of the nature of the tesserae hospitales as identification documents in the framework of exchanges between different families and, at times, even peoples, it has been hypothesized that the second gentilicium appearing in these cases belongs to the local gens that hosted foreign individuals and guaranteed their safety and their right to take part in trade and business in the local community.86 Apparently, this procedure was formally similar to the temporary adoption of a foreign trade partner by a local powerful family, as evidenced by the use of the formula with a double gentilicium, not too different from the later Latin formula of adoption (e.g., C. Iulius Caesar Octauianus). In this regard, it is worth noting that the aforementioned formulas of Orvieto and Pisa might refer to permanent adoption, possibly at times rising to a higher social status.87 The nomenclature system, however, presents some differences depending on the origin of the mentioned person and his original name. In the case of the tessera from Carthage, a local person is mentioned by a name formula in which the first two elements both mark his foreign origin: Puinel, to be compared to Latin Poenu(lu)s,88 and Karθazies, apparently derived from the name of Carthage itself (Qart-Ḥadašt). Similarly, the tessera of S. Omobono has a gentilicium, Silqetenas, that has been related to the ethnic Sulcitanus, presumably in the context of overseas trade relations 83

84 85 86 87 88

Maggiani (2005b:163, n. 19); Maggiani (2006:323); Tuck and Wallace (2013:17 and 31–3, nn. 3 and 5). Maggiani (2006:319–21). Maggiani (2006:321). Maggiani (2006:334–5). Maggiani (2006:334–5). Significantly, this praenomen occurs in one more tessera hospitalis from Murlo (Maggiani 2006:324–5), where it could also refer to a foreigner; Colonna (2013:15).

40

Mobility and Integration in Pre-Roman Etruria

with the Phoenician world.89 In this case, however, a regular Etruscan praenomen has been chosen. Significantly, these attestations show that Phoenician traders did not use their own name when travelling to Etruria for business, but made up or were allotted a local temporary name, under the supervision of a local gens.90 These newly-formed gentilicia which referred to the ethnic origin of their bearers are also attested in the literary tradition, which records Lucumo from Tarquinia, son of Damaratus, who migrated to Rome and was named Lucius Tarquinius after his city of origin.91 Una Uras from Veii was probably a foreigner too, considering his praenomen, which suggests a Faliscan origin.92 In this case, it is possible that he was a cliens or an associate of the powerful family buried in the Tumulus Chigi, possibly ‘adopted’ by the family Pupuna as the use of a double gentilicium suggests.93 It is therefore significant that the miniature aryballos that bears the relevant inscription could have been a sort of token, with a symbolic value in the context of funerary goods. A Classification of the Foreign Components in Etruscan Name Formulas In the last part of this chapter, I will list and comment systematically on methodological patterns for identifying foreign elements in Etruscan nomenclature with a view to strengthen the basis for future studies. The best opportunities to investigate the presence and integration of foreigners in Etruscan society come from bicultural and bilingual contexts within and outside Etruria. In most cases, however, the presence of foreign objects among funerary goods and 89 90

91

92 93

Colonna (2013:15); see also Maggiani (2006:321). It is remarkable that still in later times there was no correspondence between Greek and Phoenician names in bilingual inscriptions, as evidenced by a pair of twin inscriptions from Malta (Zanovello 1981), thus showing that two parallel name systems existed for people living in bicultural contexts. Colonna (2013:9–10). In actuality, the name of the Etruscan gens Tarχna was probably formed from the name of Tarchon, considered to be their forefather (cp. the Latin pairs Romulus ~ Romilii and Caeculus ~ Caecilii); Morandi (2004:526–8, n. 1181). Maras (2012:54). Maras (2012:53–4).

41

Daniele F. Maras

even the use of unexpected exotic customs cannot be considered clues to the provenience of the people involved, as objects and customs could have been imported and used by local people. The study of nomenclature, therefore, is crucial for the investigation of the incidence of foreign components in Etruscan society. Naturally, in the case of name formula, it is important to try to determine how recent the introduction of a non-Etruscan element is and how relevant it is with regard to the origin and culture of its bearer.94 In the following paragraphs, some recurring patterns will be presented, based on case studies in which scholars have presented evidence for human mobility and integration at different levels of likelihood. Presence of Non-Etruscan Inscriptions On rare, fortunate occasions, inscriptions in foreign languages are found in Etruscan contexts, which attest to the presence of nonnative persons or, possibly, even of alloglot communities. The evaluation of these pieces of evidence is therefore the starting point of our review of foreign names in Etruria. The known bilingual inscriptions are mostly not relevant, since they mainly belong to the late period of acquisition of the Latin language on the part of the Etruscans and therefore provide no evidence for the integration of foreign people into Etruscan society, but quite the opposite. The most relevant texts are nonEtruscan inscriptions in funerary contexts, as in the case of the paleo-Italic signature on the crater of the Ferrone necropolis at Tolfa,95 which provides evidence for the early use of a nonEtruscan language in the hinterland of Caere.96 This text, however, gives no clue on how and how far the mentioned Setums was integrated in Etruscan society and if he had an Etruscan name formula too.97 Apart from cases relating to Romanization, the opportunities for comparing inscriptions in Etruscan and in a different language 94 95 96 97

See in general Marchesini (2007:17–29). ST Um 4. Bellelli (2008:67–8). See also Morandi (2004:463).

42

Mobility and Integration in Pre-Roman Etruria

recording the same name are extremely rare and generally occur either in mixed cultural contexts or in boundary sites. This is the case, for instance, for a number of red glazed cups found in Civita Castellana, inscribed with the Faliscan name voltai, in the dative case.98 A single similar cup from the same site is inscribed with the corresponding Etruscan form ulties, in the genitive case,99 and another one from Vignanello has the alternative form vultasi, in the pertinentive case.100 It seems likely that the person who commissioned the inscribed cups – possibly to be used as gifts or tokens of hospitality101 – needed occasionally to translate his name into Etruscan, presumably in the context of interethnic relationships. The case of the name Apiqu, written in the Latin alphabet under the foot of an Attic stemmed saucer of the late fifth or early fourth centuries BCE, probably found in Caere,102 is rather different. As a matter of fact, this name is known in Caere from the early seventh century BCE103 and is attested until the Hellenistic period in relation to an important aristocratic family of the city that owned the Tomba dei Sarcofagi.104 The unexpected use of Latin in a post-archaic inscription encourages us to speculate on the possible Latin relations of this family, whose name is one of the few aristocratic gentilicia of early Caere that lack the usual -naending. Although written in the Latin alphabet, the form Apiqu follows Etruscan phonology, including an ending -u, and presumably corresponds to *apicō.105 This, in turn, has been considered to be a variant of *apicus (whence the Latin gentilicium Apicius), deriving from Greek ἄποκος, ‘hairless’,106 or possibly from Greek ἄποικος, in the sense of ‘emigrant, colonist’ (a name of function quite suitable for a Greek immigrant of the early seventh century BCE). 98 99 100 101 102 103

104 105 106

CIE 8435–58. ET Fa 2.20; Maras (2013b:278). ET Fa 3.4; Maras (2013b:279, n. 87). Maras (2013b:279–80). Briquel (2016:220–2, n. 84). ET Cr 1.32 (ca. 650–630 BCE): mi l[arθ]ia apiqus, under the foot of an EtruscoCorinthian vase in the Tomba I dei Letti e dei Sarcofagi (Grande Tumulo II). Briquel (2016:221). Colonna (2004:81, n. 78); Briquel (2016:222). Colonna (2004:81); Marchesini (2007:47).

43

Daniele F. Maras

In my opinion, however, there is no need to postulate a Greek origin for the name, which is most probably derived from Latin apex, with reference to the cap of the Roman flamines: *apĭcō is literally the ‘person who wears an apex’, that is to say a priest. From this surname, or name of function, two forms of the gentilicium derived, with no suffix in Caere and with the regular -ius suffix in Rome (apicius from either a simplified *apicon-ius or a variant *apicus).107 Evident Graphic/Phonetic Identity In the absence of direct evidence from non-Etruscan inscriptions, the most frequent method for recognizing foreign elements in Etruscan nomenclature is the similarity of the written form of onomastic formulas, which suggests similar patterns of sounds (phonetic identity). In many cases, there is little doubt that the names have been borrowed from a different language. As regards Greek names, the cases of Larth Telicles and Rutile Hipucrates have already been mentioned. Some further examples, in approximate chronological order, are as follows: Seventh century: Tipe(ia), from Veii (ET Ve 2.8, if corresponding to Τίβειος or Τίβιος).108 Sixth century: Aχilenas and Pereceles, from Volsinii (ET Vs 1.122 and 1.67, deriving respectively from Ἀχιλλεύς109 and Περικλῆς);110 Miliθunas, from Pontecagnano (ET Cm 2.104, deriving from Μελίτων).111 Fifth century: Hermχraθe, Stepene (or Stepe), and Θinaste, in the labels of a black-figure amphora probably produced in Volsinii (ET Cl 7.5, corresponding to Ἑρμοκράτης, Στέφανος, or Στιβεύς and, possibly, *Δινάστης);112 Eurves (ET Vs 3.8, corresponding to Εὐρύας);113 Scirunies, from Pontecagnano (ET Cm 2.138, possibly deriving from Σκίρων).114 107

108 109 110 111 112 113 114

44

Attested in Etruscan as apice (ET Cm 2.120 and Cl 1.2455–62; see Morandi (2004:73, n. 121). Colonna (2004:82); Morandi (2004:539–40); Marchesini (2007:57–8). Morandi (2004:97). See note 69. Pellegrino and Colonna (2002). Colonna (1997a); Cerchiai (2014); Maras (2018b:146, n. 42). Maras and Michetti (2003:378); Marchesini (2007:79–80). Cinquantaquattro and Colonna (2011).

Mobility and Integration in Pre-Roman Etruria Hellenistic period: Tamsnie, from Caere (ET Cr 1.161, if deriving from Δάμασος);115 Tiple, from Tarquinia, and Tiφile, from Clusium (ET Ta 3.5 and Cl 1.1645, both probably corresponding to Δίφιλος);116 Evrs, in a signature from Civita Castellana (ET Fa 0.17, from Εὔρος).117

Browsing this partial list shows that Greek elements entered Etruscan nomenclature at diverse social levels, including aristocratic immigrants in the early archaic period, mixed communities in Campania, and craftsmen, slaves, and performers in the Classical and Hellenistic periods. It is not clear whether the gens Tamsnie of Caere, belonging to the local elite at least from the late fourth century BCE, originates from a much earlier immigrant or from recent social climbing. Anyway, the Greek derivation is only hypothetical (< *Tam(a)s(e)-n(a)-ie). As we have seen, Etruscan and Latin nomenclature share many gentilicia, as a consequence of interethnic mobility from the earliest period of the cities of southern Etruria and Latium. In many cases, the parallel formation of gentilicia on identical onomastic bases differed only for the suffix used: -na in Etruscan and -ius in Latin. There are cases, however, in which a Latin gentilicium occurs unaltered in Etruscan formulas, such as Hustile(ia) (ET Vc 2.1, in genitive, seventh century, corresponding to Hostulus < *hostĕlos), Zalvie (ET AH 2.1, seventh century, corresponding to Saluius),118 Leceniies (ETAd 2.1, late sixth century, from Licinius),119 Auseḷ[i]e (ET Cr 2.195, fifth century, from Aurelius < *auselios), Clavtie (ET Cr 5.2 and Cs 2.4, fifth–fourth centuries, from Claudius),120 Lvcili (ET Cr 1.159, fourth century, from Lucilius).121 This list is far from being complete, but is sufficient to show how complex and intertwined the relationship is between Latin and Etruscan onomastics throughout the centuries. 115 116 117 118

119 120 121

Morandi (2004:502). Morandi (2004:540); Facchetti (2007:129–30). Maras and Michetti (2003). Marchesini (2007:77–9). Remarkably, the gens Saluia maintained its connection with Ferentium in the Imperial period and gave birth to emperor Marcus Salvius Otho in 32 CE; Morandi (2004:436–9). Morandi (2004:277). Morandi (2004:134–5). Morandi (2004:292).

45

Daniele F. Maras

Finally, a number of Etruscan gentilicia correspond to Italic names known from later epigraphic sources. In this regard, to the aforementioned Vestiricina- and Peticina- we can add here Paiθina- (ET Cr 3.27, AS 3.6 and Vn 3.2, seventh century, probably from Italic *Baitos or Latin Paetus),122 Maies and Maias (ET Vc 2.9 and Cr 2.194, seventh and fifth centuries, from Oscan Maiís),123 Vinucenas (ET Vs 1.126, sixth century, from Oscan Viínikiís, arc. Vinuχs),124 Uφaliies (ET Cm 2.18, sixth century, from Oscan Úpfals),125 Plaisenas (ET Vs 1.18 and 1.121, from a paleo-Italic Blaisiis).126 In addition, it is worth mentioning the possibility that the recent praenomen Tasma127 derives from the Messapian name Δαζιμο.128 As we have seen, however, the vast majority of the documentation for the Italic languages is much more recent than the Etruscan attestations, which often date back to the Orientalizing and Archaic periods. It is therefore difficult to ascertain the historical origin of parallel name formulas in different languages, which in theory could have been subject to borrowing from Etruscan to Italic. Etymology Even more complex and, often, more uncertain than phonetics is the acknowledgement of foreign elements in Etruscan nomenclature based on etymology. This method is easily applied in the case of perfect correspondence of names in Latin and Etruscan, as in the following few examples: Rutelna- ~ Rutilius (< *rutelos); Laivena- ~ Laeuius (< *laiwos); Hvul(u)ve- ~ Fuluius (< *folwos); Flaviena- ~ Flauius (< *flawos); Una(na)- ~ Iunius (< Fal. iuna). Some further names can be added to the list even if no parallel Latin gentilicium is actually attested: Flusena- (< *flos- > Florus, a cognomen); Crepu (< Lat. creppus, ‘he-goat’, presumably in a sense comparable to that of lupercus);129 Caicnas (< *kaikos > 122 123 124 125 126 127 128 129

46

Morandi (2004:343–4); Maras (2009:379). Morandi (2004:298). Agostiniani (1983:48); Morandi (2004:201), with a different opinion. Agostiniani (1983:39, n. 21). Morandi (2004:381), with further bibliography. ET Vs 1.257, from Bagnoregio (second century BCE): tasma śatnas. For further foreign names adapted as praenomina, see Marchesini (2007:23–6). Colonna (1997b:65); Morandi (2004:143); see also Marchesini (2007:71–2).

Mobility and Integration in Pre-Roman Etruria

Caecus, a cognomen; cf. Caecilius). At times, is even difficult to establish the direction of the loan, as in the cases of Acvilna-130 ~Aquilius131 and Ecisie132 ~ Egerius.133 Etymology is also at the base of the recognition of foreign names of function that we mentioned above in relation to mercenaries and warlords. Further cases concern Etruscan gentilicia apparently deriving from other languages of pre-Roman Italy, as in the following examples, deriving from Italic names (Meties and Metiena < *Metto- [cf. Lat. Mettus];134 Alfiana < *Alfo- [cf. Lat. Albius]; Vipes and Vipiiennas/Vipinas < Osc. Viíbis;135 Laucie-/ Luvcie- < Osc. Lúvkis;136 Trepe-/Treple- < Osc. Trebiis137) and numerals (Pumpuna < Ital. *pompo- [cf. Lat. Pomponius]; Utaves < *ohtave- [cf. Lat. Octauius]; Nunie < *nonie- [cf. Lat. Nonius]).138 In addition, some Etruscan names have uncertain comparanda in Celtic onomastics, as in the cases of Epunianas,139 possibly deriving from *epo-nyo-, in connection with the semantic sphere of the ‘horse’,140 and Katacinas,141 if deriving from katako-, rec. Catac(i)us,142 and even in the Phoenician lexicon, such as Melacinas and Melctrna,143 possibly depending on MLK, ‘king’, and MLKT, ‘queen’, and the obscure Mazbavana,144 which might be derived from MZBḤ, ‘altar’, in the context of religious cultural exchange.

130 131 132

133

134 135 136 137 138 139 140 141 142

143 144

ET Ve 3.7, Vc 3.4–5 (early sixth century BCE): mine muluven(ic)e avil(l)e acvilnas. de Simone (1989). ET Cr 2.7 (second quarter of the seventh century BCE): mi θihvaries ecisie; Colonna (1970). Remarkably, both names Aquilius and Egerius have Etruscan connections in the sources. See also Marchesini (2007:26–9; 43–4). Morandi (2004:311–12). Morandi (2004:201–11). Morandi (2004:276–7; 291–2). Morandi (2004:544–5). Salomies (2009:515–20); see also Wallace (2008:93–4) and Roncalli (2009). ET Vc 2.7. Sassatelli (2003:336), with further bibliography. ET Vs 1.165. de Simone (1978); Sassatelli (2003:335–6); Morandi (2004:264–5); Marchesini (2007:28). Respectively ET Vc 3.2 and AS 2.12, 1.217, and 1.276; see also Morandi (2004:310). ET AT 0.1; see Bagnasco Gianni (1996:195), Cristofani (1996:223).

47

Daniele F. Maras

Ethnic Adjectives Used as Names From time to time, the foreign origin of a person is incorporated into his or her nomenclature by transforming the relevant ethnic adjective into part of his or her name formula.145 In such cases, there are a number of different possibilities: 1) the ethnic adjective can be added to the personal name as a nickname or a cognomen; 2) the ethnic name is used in the place of a gentilicium; 3) a gentilicium is derived from the ethnic name by adding a suffix; 4) the reference to a foreign people is included in the praenomen; 5) an ethnic adjective is used in the place of a personal name.

In case no. 1, it is usually not clear if the ethnic adjective is actually part of the relevant name formula. After all, in the aforementioned inscription of Laris Pulenas, the third element of the formula of the forefather was not inherited by his offspring (Creice, ‘the Greek’). It is therefore possible that Θanχvil Craica, attested in Vulci in the mid-fifth century BCE,146 was also still considered a foreigner with no gentilicium (a sort of metic, or resident alien). As a consequence, when name formulas are found including an ethnic adjective in the place of a gentilicium (no. 2), scholars should always wonder whether this is part of the name or just an indication of the origin of its bearer. This is the case, for instance, of the early inscription of Tite Latine from Veii,147 where we have no clue if the second element of the formula was heritable or not.148 Considering its early period, the mentioned person was possibly still regarded as a foreigner from Latium.149 Less certain is the case of names referring to Etruscan cities, such as Veiane (from the name of Veii)150 and Velχite (from Vulci),151 or Latin cities, such as Lapicane (from Labici)152 and, possibly, Runate/ 145 146 147 148 149 150 151 152

48

Hadas-Lebel (2009); Colonna (2013:10–13). ET Vc 2.33–4 ; see also Morandi (2004:140–1). ET Ve 2.4 (end of the seventh century BCE): mi tites latines. Morandi (2004:275). Colonna (1976b:376, n. 131); Colonna (2013:10–11). ET Fa 2.6–10, Cr 2.112, Vc 2.4, Fe 3.3; Morandi (2004:167). ET Cl 1.576–7. ET Cr 2.131 and 5.8 (third century BCE); Morandi (2004:270–1), with further comparanda.

Mobility and Integration in Pre-Roman Etruria

Rumate (from Rome).153 In this regard, however, we have the literary evidence of the Tarquinii, who were named after their city of origin, according to the legend, as mentioned above. A number of gentilicia derive from ethnonyms or poleonyms with the addition of the -na or equivalent suffix (no. 3), following the normal procedure used for individual names, as in the following examples:154 Kalaprena(s) (cf. Lat. Calabri: ET Ta 2.3, early seventh century BCE, from Tarquinia; Vs 1.59, late sixth century BCE, from Orvieto);155 Tursikina (cf. Lat. Etruscus: ET Cl. 2.3, third quarter of the seventh century BCE, from Castelluccio);156 Avhiricina(s) (cf. Lat. Africa: ET AT 3.2, end of the seventh century BCE, from S. Giuliano);157 Χairitna(s) (cf. Lat. Caeritis: ET Vc 2.72, mid-fifth century BCE, from Vulci);158 Umrana (cf. Lat. Vmbri: ET Cl 1.362–6, Hellenistic period, from Clusium).159

In these cases, it is uncertain whether the new gentilicia originated from real ethnic adjectives used as nicknames (respectively *Kalapre, *Tursike, *Africe, and so on), or from identical individual names. Evidence in support for the former hypothesis comes from the tesserae hospitales, which show Phoenician ethnonyms used as fictitious gentilicia (Karθazie, Silqetenas), and demonstrate that from time to time temporary name formulas were used by non-integrated foreigners.160 On the other hand, however, some attestations are known where an ethnic name is used as a praenomen (no. 4) or even as an individual name (no. 5). 153 154 155 156 157

158 159 160

ET AH 2.2 and AS 2.2; Colonna (1981:169–70); Morandi (2004:429–30). Further comparanda are listed in Colonna (2013:11–15). Morandi (2004:261); Colonna (2013:13). de Simone (1976); Marchesini (2007:28–9). Morandi (2004:36–7); Colonna (2013:15), with reference to Phoenician immigrants. See also Marchesini (2007:28), with a different interpretation. Colonna (2004:85); Morandi (2004:583–4). Morandi (2004:568) with further comparanda; Colonna (2013:12). See notes 88–90. Further gentilicia referring to a Phoenician origin are listed in Colonna (2013:14–15).

49

Daniele F. Maras

This issue is especially relevant inasmuch as it affects our understanding of when the integration took place. There is the possibility that it goes back to earlier generations, not necessarily implying that the bearers of such ethnic gentilicia were of a lower social status. Unfortunately, it is not possible to offer any general answer. When an ethnic name is used in place of a personal name (no. 5), it could be considered as a nickname, suitable to identify a metic, as in case no. 1. The ‘depersonalization’, however, is deeper in this case, for the individual name is fully replaced. This phenomenon was certainly suitable in the case of slaves or foreign people of low social class, but also, possibly, in certain other contexts. A few examples follow: Celθes (cf. Lat. Celtae: ET Cr 3.22, late sixth century BCE, from Caere);161 Latiθe (cf. Lat. Latium: ET Ta 7.5, late sixth century BCE, from Tarquinia);162 Cale (cf. Lat. Galli: ET Ta 7.13, end of the sixth century BCE, from Tarquinia);163 Nucrtele (cf. Lat. Nuceria: ET Ta 7.25, end of the sixth century BCE, from Tarquinia);164 Kursike (cf. Lat. Corsica: ET Po 2.43, mid-fifth century BCE, from Populonia);165 Umrke (cf. Lat. Vmbri: ET Po 2.44, end of the fifth century BCE, from Populonia);166 Eluveitie (cf. Lat. Heluetii: ET Pa 0.3, end of the fourth century BCE, from Mantua).167

People in a number of different situations feature in this list, including slaves (such as the Cale of the Tomba delle Iscrizioni 161 162 163 164 165 166 167

50

Sassatelli (2003:333), with further bibliography; Colonna (2004:76). Colonna (1988:447–8); Colonna (2013:11). Colonna (2016:135, n. 50); see also Sassatelli (2003:334). Maras (2018a:100, n. 86). Maggiani (1999:47–51). Maggiani (1992:184; 187, fig. 15, n. 48). Sassatelli (2003:333).

Mobility and Integration in Pre-Roman Etruria

and the Umrke (slave) of Pelka at Populonia), ritual performers (such as the pugilists Latiθe and Nucrtele in the paintings of the tombs degli Auguri and delle Iscrizioni at Tarquinia), possibly mercenaries (such as Celθe in Caere) and foreign visitors (such as Kursike at Populonia and Eluveitie in Mantua).168 In all these cases, foreign individual names might still have existed but were not used in Etruria. In addition, it remains uncertain if the ethnic nickname later underwent a process of transformation into a proper name. In this case, however, it might be significant that Cale is a well-known gentilicium of the hinterland of Tarquinia.169 Closing Remarks: Diverse Categories of Interethnic Mobility In this chapter a number of different types of geographical mobility and integration have been described and commented on by presenting a series of case studies. In the framework of the historical development of Etruscan onomastics, the presence of foreign elements in personal nomenclature has been highlighted, thus identifying some categories in which it is more common to find early immigrants, newcomers, and visiting foreigners. These include members of the elite since the earliest documentation available, harking back to the early and mid-Orientalizing period. In this category, Greek immigrants are found side by side with Latin and Italic families, paralleling the legend of Demaratus. As the noble Corinthian had a retinue including skilled craftsmen, it is not surprising that many Greek immigrants were artisans, particularly potters and painters. In addition, ritual performers are a further category of people featuring Greek names, along with Italic immigrants and some possible Phoenician elements. In this group we can include not only dancers and athletes, but also the noble ancestor of Laris Pulenas. As expected, a number of military persons are of foreign origin, for they belong to the widespread category of mercenaries and 168 169

See further ethnic names in Colonna (2004:76–7). ET AT 1.29, 48, and 64 (from Tuscania), and ET AH 1.42 (from Bomarzo); see also ET Vc 1.121 and Vt 8.1.

51

Daniele F. Maras

military chiefs, who from time to time settled in Etruria and apparently had offspring there. In this case, it is remarkable that the nicknames used in military contexts were transformed into part of official name formulas. Traders (and, possibly, diplomats, keeping in mind that in antiquity these categories of travellers often melted into one another) probably had a special status in Etruscan communities, as some tesserae hospitales referring to foreigners seem to demonstrate. In these cases, a fictitious, temporary name was allotted to the visiting person, including a reference to his/her foreign origin. Finally, slaves and low-class immigrants were often of foreign origin and are frequently attested in the later periods, particularly in the Hellenistic period. Of course, this list is far from complete and many of the presented cases are hypothetical or uncertain. However, in my opinion, it is sufficient to acknowledge that the procedures of integration of immigrants into Etruscan society went on throughout all of Etruscan history, and that this was recorded in epigraphy. Etruscan nomenclature was an evolving system, open to the introduction of new elements and suitable to creating new gentilicia and integrating foreign name formulas from the Orientalizing period to the late Hellenistic period. Finally, Roman nomenclature overcame the Etruscan system, after which a new historical phase of onomastics started for Etruria and Italy.

52

C HA P T E R 3

E L U S I VE M I G R A N T S O F A N C I E N T I TA LY E L E N A I S AY E V

The landscapes of the Italian peninsula, no less than others in the ancient Mediterranean of the third and second centuries BC, hosted numerous intersections for the convergence of resources – people, objects, ideas and stories – recounted in multiple languages and mediums.1 Propelled by technology, trade, warfare and alliances, as well as love, curiosity and brigandage, the ensuing connectivities stimulated the emergence of new sociocultural trends and communities. Both material and literary endeavours attest to the dynamism of multiple movements within and through Italy in this period (Isayev 2017). Yet, tracking migrants who were part of these human flows, in terms of numbers, origins, destinations and the drivers of their mobility, proves difficult apart from a few exceptional episodes. Beyond the peninsula it is particularly hard to identify individuals of Italian origin living abroad before the Late Republic, despite multiple indicators of ties between communities in Italy and sites around the Mediterranean, especially in the archaeological record. In part, the elusive nature of migrants may be attributed to the governing authorities’ lack of interest in counting and recording people on the move, beyond what was necessary for military recruitment, the census and taxation. While there were requirements for traders to register their goods at customs houses (Lefèvre 2004; Bresson 2007),2 we have 1

2

I am most grateful to Katherine McDonald and Nick Zair for inviting me to join in the dynamic discussions of the 2016 workshop that underpin this volume, and to the many scholars there who inspired new paths to follow among them, how women’s trajectories fit into these ancient intersections. I am particularly grateful to Patrick James for sharing with me his current research on migration in Greek comedy, which, along with the comments of the anonymous reviewers and the editorial team, has been very thought provoking and enhanced this chapter. Plautus Asinaria 240–3; Menaechmi 117–19; Trinummus 795, 1105–7 .

53

Elena Isayev

no evidence in this period for physical state borders or any checkpoints resembling those of twentieth- and twenty-first-century immigration controls that would have hindered civilian mobility. The lack of such mechanisms is a testament to the positioning of mobility, not as a phenomenon that was exceptional to the everyday functioning of communities, but inherently embedded within it. The barriers that were difficult to cross were those of status, especially the one between slave and free. To showcase the extent to which a mobile culture formed part of the everyday norm of Italo-Roman society, this chapter will present a number of cases indicative of people from Italy moving around the Mediterranean and then examine ancient perceptions of mobility through the comedies of Plautus. The ancient cities of his plays – whether Athens, Calydon, Epidamnos or Thebes – are dynamic sites of interaction. They also reflect an interest in finding new ways to define movement and categorise those who were present in the city at any given time, both the long-term residents and the new arrivals. Mobility of Italians in the Third and Second Centuries BC Beyond the stage, epigraphy provides the most tangible evidence for the presence of civilians from Italy in other parts of the Mediterranean in the period of the third and second centuries BC. Studies such as those of Hatzfeld (1912; 1919), and more recently of Ferrary (2002) and Hasenohr (2007b), for example, have shown that Italians are visible on the island of Delos from at least the third century BC (and see also Rovai, Chapter 8 of this volume). Before this period, however, we have trouble recognising people with Italian origins abroad in any substantial numbers, although from ancient writings and other material evidence we know they must have been there. In part this has to do with the lack of interest in displaying Italian cultural traits materially before the spread of Roman power. The virtual invisibility can also be attributed to the nature of the epigraphic habit in Italy itself, which only gained prominence in the third and second centuries BC, especially with the diffusion of elite funerary epitaphs (Berrendonner 2009: 192). Based on surveys of 54

Elusive Migrants of Ancient Italy

a decade ago, for the whole of Italy the total number of Latin inscriptions known from the third century BC is about 600, of which only about 146 are from the city of Rome (Gordon & Reynolds 2003: 219–20, 227–8). In the following two centuries the total number of Latin inscriptions from Italy rises to over 3,000, and most of these date to the final 160 years (Panciera 1997; Gordon & Reynolds, 2003: 219–20; Pobjoy 2000). This is precisely the point at which Italians overseas become more visible. Individuals who arrived from Italy or had roots in the peninsula can be recognised as such on inscriptions either by the nature of their name (which is not uncontroversial: Wilson 1966: 104–7; Solin 1982: 112; Müller & Hasenohr 2002) or because they are cited, for example, as Rhômaioi or Italikoi in Greek inscriptions and as Italici in Latin ones (Ferrary 2002: 135). The surviving number of such inscriptions, however, still does not reflect the extent to which Italians would have moved out of the peninsula to work and live abroad. The issues surrounding the visibility of people from Italy abroad may be captured by considering the sources related to the slightly later events which took place in Asia Minor. Our earliest inscriptions from this region that include names which are distinctly recognisable as Italian are from Pergamum and date to circa 133 BC (von Prott & Kolbe 1902 Nr. 116; Nr. 118; Nr. 119). The presence of these individuals would have preceded direct Roman administration and formation of the province of Asia Minor that was overseen by M. Aquillius in the years 129–126 BC.3 Despite the extension of direct Roman rule over the area, even a generation later, the total number of remaining inscriptions indicating persons of Italian origin is about a handful.4 We get a different impression of the situation from the records of Mithridates’ massacre in 88 BC of Romans and Italians living in Asia Minor. The total number of those who were killed, according to some authors, was 80,000.5 3 4

5

Strabo 13.4.2; Appian BC 1.111; Livy Periochae 58, 59; Gruen (1984: 592–610). For Italians in Asia Minor: Fränkel (1890: 249, lines 14ff); with Wilson (1966: 125); Frank (1975: 4.543–5); Ferrary (2001). For the details of the events: Appian The Mithridatic Wars 22–4. For the figure of 80,000: Valerius Maximus 9.2.3 (ext); Memnon of Heraclea Pontica 31.9 (= FGH III B, p. 352, ll. 16–21).

55

Elena Isayev

Whether we can believe this possibly exaggerated figure is questionable. Still, even if reduced to 10,000, the number of people who lost their lives is significant, especially if we consider it in light of the overall free population of Italy, which some estimate at about eight million in the first century BC (Bowman & Wilson 2011; De Ligt 2012; Hin 2013). Without the references to this event we would have little sense that so many Italians lived in this area of the Mediterranean at the time. The episode, which engulfed numerous communities in the region, is particularly revealing, since these people were not in Asia Minor as a result of any mass mobile event such as colonisation. They lived across numerous cities, arriving as individuals and in family or household groups for a variety of reasons and opportunities, both economic and cultural. These would have increased with Roman hegemony, not least for the publicani (tax collectors)6 and the negotiatores (tradesmen), who could operate with the added protection of Roman laws (Lex de Prouinciis Praetoriis = Crawford 1996: Law 12). It was in part their exploitation of the region that was later blamed for the widespread support of Mithridates, leading to the murder of Romans and Italians in 88 BC (Justin 38.7; Amiotti 1980: 134, 139). The form of independent and private mobility which is evidenced by these events is the great unknown in terms of numbers on the move. Our figures for such mobility exist on a large spectrum of two extremes proposed by demographic studies: in the second and first centuries BC the total number of individual movements through, in and out of Italy could have been either five million (the low estimate by Scheidel 2004) or forty million (the theoretical figure based on medieval comparisons, Osborne 1991; Erdkamp 2008: 419). What has become evident is that the large figure is plausible (Isayev 2017: chapter 2). Even Scheidel, using the well-documented census data of the Augustan period at the end of the first century BC, estimates that some 40 percent of male Romans over forty-five would have been born in a different location to their current place of residence (Scheidel 2004: 13–20; 6

For publicani across the empire: Cicero Ver. 2.3.6; Att. 1.17.9; Man. 15; Agr. 2.80; Flac. 19; Lucilius 671–2 (Marx). For a summary of the fiscal changes and the workings of the publicani: Frank (1975: 4.511–12); Gruen (1984: 592–610); Ferrary (2001: 100–1).

56

Elusive Migrants of Ancient Italy

Scheidel 2006: 223–4, with De Ligt 2012: 120–8). The figures can be adjusted depending on what is believed to be the overall population of Italy, but even if the total estimate is reduced from 40 percent to 20 percent (based on doubling Scheidel’s low count figures for the overall population),7 it still remains a substantial proportion of the recorded citizen body which is on the move. Such demographic calculations, presenting a highly mobile culture (as also supported by Tacoma 2016a), are one step towards understanding the ways in which mobility and those who were on the move were positioned within society. The Plautine Corpus Plautus wrote his comedies in the decades which spanned the third and second centuries BC – a period of unceasing warfare on the Italian peninsula and beyond it. Produced against the backdrop of the Hannibalic War and the repercussions that followed, his corpus may be conceived as a wartime repertoire. Amy Richlin’s most recent study of the Plautine corpus shows the extent to which we need to position it not only in the midst of warfare and devastation but also within the context of mass displacement and enslavement (Richlin 2017). In fact, she argues that his comedies are ‘slave theatre’ – performed by select travelling troupes who were largely made up of the lowest members of society for consumption by an audience that would have included many of their peers. While some may question the extent to which the comic entourage was made of up slaves, there is no doubt that the comedies present a world where movement was ubiquitous, a society that depended on mobility, anticipated it and was aware of its risks and opportunities. But this is not predominantly due to the wartime context, even if conflict and criminal activity are the roots of many characters’ predicaments of slavery and displacement. When Plautus was writing his comedies, aside from the war-driven circulation of people, Rome was also rapidly becoming the main destination cosmopolis for merchants, politicians, artists, craftsmen, scholars, slaves, entrepreneurs and others seeking opportunities around the Mediterranean. The players allow these other stories to 7

For the diverse positions on the Italian population: Morley (2001); De Ligt (2012: 1–39).

57

Elena Isayev

emerge, even if Rome itself is never the setting for Plautus’ comedies. They consistently tell the audience that the plays are on foreign soil, palliata – in Greek dress, to enhance the foreignness of their actions. The prologue to the Menaechmi emphasises this: ‘This story is quite Greek-ish, but to be exact, it’s not Athen-ish, it’s Sicil-ish, in fact’.8 Despite the pretence of a distant physical setting, there are few distinguishing characteristics in the actions of the comic ‘Greeks’ or ‘Sicilians’ (Segal 1987: 37). Plautus’ imaginary settings could be anywhere (Gratwick 1982: 112–13; Gratwick 1993: 8–15), and equally they could each be Rome. The mobile eclectic troupes of multiple origins who evoked these settings were the transmitters of stories and performances, old and new, that reached into Latium. Greek theatre was already being consumed and adapted by Italic elites living in the hinterlands of South Italy from the fifth century BC (Taplin 2007: 32; Biles & Thorn 2014; Robinson 2014; Nervegna 2015). Through an intricate study of theatrical scenes on vases found at indigenous South Italian sites, Robinson has demonstrated that the aristocrats who purchased them, and for whom they were produced, would likely have had first-hand knowledge of the performances that were depicted. The site of Ruvo in ancient Apulia, for example, contained forty-four such vessels (Robinson 2004; 2014: 326). Here was just one of the many intersections that through theatre alone strung together Greece, Sicily, South Italy and, by the fourth century BC, also Macedon. It is representative of wide-reaching networks and a shared cultural milieu, which equally pervades the later cosmopolitan settings of the Plautine corpus. The Greek dramatic underpinnings of Plautus’ comedies, in particular the products of Greek New Comedy (c. 325–250 BC), have driven an ongoing debate about whether the plays can be used as a historical source for Republican Italy, which partly hangs on the extent to which Plautus altered the Greek originals.9 There is now sufficient evidence indicating that the Plautine corpus was highly innovative; it is not made up of simple translations, nor even of close adaptations of the original texts (Lefèvre, Stärk & Vogt-Spira 8

9

Plautus Menaechmi, 11–12: atque adeo hoc argumentum graecissat, tamen non atticissat, verum sicilicissitat. I am grateful to Patrick James for nuancing some of the comparative points in the way mobility is approached in the ancient Greek and Latin comedies.

58

Elusive Migrants of Ancient Italy

1991; Benz & Lefèvre 1998; Fraenkel 2007; Richlin 2017). His works are unlike the comedies of the later Roman playwright Terence, who followed the Greek originals much more closely (Segal 1987: 7; Habinek 1998: 56–7). Plautus’ plays are products of his time, reflecting Romano-Italian concerns and societal frameworks, proving an important historic source for the period (e.g. Segal 1987; Leigh 2004, Manuwald 2011: 225–33; Richlin 2017). There are differences in the approach to mobility between the Greek and Plautine comedies. In the Greek plays mobility is usually part of the backstory rather than crucial to the plot, while in most of Plautus’ plays a character coming from abroad holds a central position. His imaginary world also feels bigger and more global, lacking a single centre, while the Greek comedies operate with more local networks using Athens as the primary hub, along with its demes. Plautine characters seem to travel further and for longer periods of time. In choosing to move, individuals did not set out with the goal of becoming members of new communities; the presence of colonial foundations hovers in the background (Epidicus 342–3; and Pseudolus 1100). Unlike the preceding Greek plays, especially those of Menander, in which it is specifically Athenian citizenship that matters, Plautine characters seek generic citizenship rather than that of any particular state. This is not to say that there is less mobility evident in Greek comedy, but rather that it is not as prominently displayed, which perhaps suggests a different perception of those on the move. The way in which Plautine comedy exposes Italic characteristics, while seeming to follow Greek comic plots, is exemplified in its plethora of homecoming scenarios. Many of the Latin plays have extended scenes of homecoming or the welcome and greeting of a guest that are often accompanied by mockery of hospitality rituals. Those who typically are on the receiving end of this hospitality include friends and visitors, foreign notables, soldiers returning from campaign, merchants and relatives arriving from business trips or the farm.10 In addition, homecoming scenes are also satirically played out among slaves, courtesans and parasites, 10

See Isayev (2017), Appendix B, for a list of the main instances of such homecoming scenes in the Plautine comedies.

59

Elena Isayev

offering hospitality which they could not fulfil, or expecting it (e.g. Stichus 480–90; Bacchides 185–7). An example is Plautus’ Epidicus (1–10), which opens with two slaves acting out the rituals of welcome. The dialogue is between Thesprio, just returned from a campaign, carrying his master’s equipment, and Epidicus, who is there to greet him:11 epidicus: thesprio: ep: th: ep: th: ep: th: ep:

Hi there young fellow! . . . Good day to you. God grant your wishes. Glad you are safely back. What about the rest? What’s normally added to this: you shall be given a dinner. I promise – What? – to accept if you give me one. What about yourself? Are you well?

It is an example of a stock sequence, which begins with the recognition of the traveller, then a greeting followed by an expression of delight for their safe return, an enquiry after their health, and ending with a dinner invitation (Barsby 1986). The slaves’ inability to provide dinner, along with the unlikelihood that anyone would take care of their well-being, is what makes the scene into a satirical subversion of the rituals expected upon a freeman’s return. Part of the comic effect may have been created through an implicit questioning of the sincerity of such exchanges overall, which follow a well-known pattern that appears in other Plautine comedies as well. The situation is somewhat different in the Greek plays, such as those of Menander, for example. Although there are many homecomings in his surviving corpus, they are presented in a more cursory way, at times as a simple greeting upon a character’s return (e.g. Menander’s Dis Exapaton, 102–7; with Barsby 1986: 140). Plautus takes them to the other extreme, as Lowe (2007: 110) reflects on the comedy of Stichus: ‘Menander’s play seems to have been taken apart and reassembled into a virtually plotless montage of festive variations on a theme of homecoming’. There would have been little incentive for Plautus 11

Plautus Epidicus, 1–10: EP: Heus, adulescens . . . TH: Salve. EP: Di dent quae velis. venire salvom gaudeo. TH: Quid ceterum? EP: Quod eo assolet: cena tibi dabitur. TH: Spondeo – EP: Quid? TH: – me accepturum, si dabis. EP: Quid tu agis? ut vales?

60

Elusive Migrants of Ancient Italy

to revel in these scenarios unless they enhanced the comic effect. In order for the status subversion to work, the audience would have had to be familiar with the rituals, implying that they happened often and formed part of a shared etiquette to which the spectators could relate. Could it be that the Plautine scenes convey a more general widespread cultural trait: an Italian fascination with the rituals of arrival? Although not explicit, there may be some evidence of this from the archaeological context. In Italy, the theme of some of the images in the fourth- and third-century BC tomb paintings from the cemeteries of Poseidonia has been interpreted as that of homecoming. The scene usually depicts a woman with a libation and a warrior on horseback, who at times wears distinctive dress and armour, such as the three-disk breastplate typically associated with the Italic sphere.12 The scene, understood as one of greeting, seems similar to the better-known portrayals of women and warriors on Attic pottery (Greco Pontrandolfo & Rouveret 1983; Pontrandolfo & Rouveret 1992). On the Greek vessels, in contrast, the depiction is less ambiguously one of departure, with an arming for battle rather than a return. It would require a more in-depth study of the visual and literary material to see if, among Italic communities, there was a greater emotional focus on arrivals than on departures compared to what we find in the Greek sphere.13 Comedies of Movement The Plautine comedies do not provide any additional data to help make better calculations for the numbers of people on the move – a preoccupation of the twenty-first century for which interest in the ancient world was lacking. They show little interest in the state-initiated mass migrations that are often the focus for demographers. Instead, their interest is in explorations of the unquantifiable, everyday movements of individuals and their entourages 12

13

Such as, for example: Warrior returning home, from Andriulo Necropolis, Tomb 12, Poseidonia (Paestum) (Cipriani & Longo 1996). See also McDonald and Clackson, Chapter 4 of this volume. One such extended departure is used by Apollonius of Rhodes to open his narrative of Jason’s voyage in the Argonautika (Thalmann 2011: 54–6).

61

Elena Isayev

around the cities of the Mediterranean. The plays reveal a certain intangibility of the physical landscapes within which any notional patria sits.14 Any nostalgia for a past (or future) home tends to envisage it as ephemeral, a matter of moments in time rather than space. It may be embodied in the simple question posed by the servant Messenio to one of the twins in Plautus’s Menaechmi (1111): Quid longissume meministi, dic mihi, in patria tua? (Tell me, what is the earliest thing you remember in your country?). The question is not what is remembered of the patria but the action within it. In response Menaechmus recounts the memory of going with his father to a market in Tarentum, where he wandered off and was taken away. In the comedies, the home of return that is conjured up appears as an intersection of relations at one’s birthplace rather than any evocation of the built or natural environment. The most tangible elements of a patria are the domus and its altars, which are used metaphorically, and in the earliest Latin literature are often presented in opposition to statelessness and exile (Isayev 2017, chapter 11). They appear in Ennius’s tragedy of Andromache (83–4, 87) who evokes her stateless existence through its symbols: arce et urbe orba sum. quo accedam? quo applicem? cui nec arae, patriae domi stant, fractae et disiectae iacent . . . O pater, o patria, o Priami domus! (‘Citadel and city gone. Where shall I take myself? Where shall I appeal? No longer do my ancestral/native altars stand at home but lie broken and scattered . . . O father, O fatherland, O house of Priam!‘). Although the setting draws on Greek narratives, as Jocelyn (1969: 244–5, 250–2) and Feeney (2016: 75) observed, Ennius employs the language of Roman law and social practice, presenting the destroyed altars as a symbol of loss. The lament was familiar enough to be parodied in Plautus’s Bacchides (932): O Troia, o patria, o Pergamum, o Priame periisti senex . . . (‘O Troy, o fatherland, o Pergamum, o old Priam, you have perished . . . ‘). The rootlessness and search for an intangible patria can be in part attributed to the uncertainty and displacement resulting from the backdrop of warfare and mass enslavement against which the plays were created, as suggested by Richlin (2017). There is also 14

Discussions about patria within the Plautine comedies: Richlin (2017), chapter 7.

62

Elusive Migrants of Ancient Italy

shifting and restlessness, which goes beyond the conflict scenario – an anticipation of people on the move, which the comic plots depend on. In the prologue of Plautus’s comedy Menaechmi (72–6), the towns are mobile as are their inhabitants, inside and outside the comic action. To set the scene an actor puts the play, and the stage-set, into motion: ‘This is the city of Epidamnus while this play is acting; when another shall be acted, it will become another town. It is quite like the way in which families too, are wont to change their homes; now a pimp lives here, now a young gentleman, now an old one, now a poor man, a beggar, a king, a parasite, a seer’.15

Plautus’s Mercator tells of sons selling up their fathers’ farms to move onto a life at sea as merchants (Plautus Mercator, 64–79). Neither in his comedies, nor in the writings of his contemporary Cato, do we find the image of the sedentary farmer. Although for Cato the model Roman citizen is a farmer and soldier – in contrast to the untrustworthy tradesman – his de Agri Cultura is written for the farmer as businessman (Cato de Agri Cultura, Introduction 1–5). Implicit in the de Agri Cultura are regular movements between town residence and country farm which are also part of Plautus’s comic repertoire (e.g. Casina, 90–112, and Mercator, 275–82). Although these shorter movements may be distinct from ones further abroad, still, Cato’s writings reveal an environment with high turnover of farms and owners, encapsulated in his advice for buying a farm: that it should be in the neighbourhood where farms do not often change owners and those who have sold up are sorry to have done so (Cato de Agri Cultura, 1.4). The destinations of those who sell up their farms, whether within the ager Romanus of Italy or the newly expanding Roman territory overseas, is another matter and not specifically of interest for Cato in this particular work.16 In these writings the figure of the sedentary peasant is missing, except in an idealised almost mythical sense: a figure whose existence Rathbone (2008) and others have also 15

16

Plautus Menaechmi, prologue 72–6: haec urbs Epidamnus est, dum haec agitur fabula: quando alia agetur, aliud fiet oppidum; sicut familiae quoque solent mutarier: modo hic habitat leno, modo adulescens, modo senex, pauper, mendicus, rex, parasitus, hariolus. For the extent of Italian presence and visibility abroad in this period, especially in the Eastern Mediterranean, see the discussion above, and with more detail, Isayev (2017, chapters 1–2).

63

Elena Isayev

questioned (Horden & Purcell 2000: 380; Archibald 2011). Like the merchant, the agriculturalist too can soon find himself on the road, either because of military service, expulsion, enslavement or simply the choice to do something other than farm – to follow a love interest or the ambition to take to the seas as a tradesman. Plautus’s world map is a dense web of life trajectories crisscrossing the Mediterranean.17 Some of the longest traverses appear in the plays Curculio, Persa and Poenulus, with the journeys presented there spanning from the Black Sea to North Africa and as far as India (Figure 3.1). A Disposition for Mobility Multiple aspects of the Plautine comedies indicate that a highly mobile culture was embedded in the Romano-Italic context that produced and consumed them. Many of his plays depend on travelling characters arriving from foreign parts or setting off on journeys as comic motifs to drive the action. Some plays have few characters of local origin, and in Poenulus none are citizens of Aetolia, where the action is set.18 Outsiders are used as targets for subversion; newcomers with the least amount of local knowledge overturn the role of their hosts (especially in Menaechmi; Miles Gloriosus; Poenulus; Rudens). It may appear curious, therefore, considering the frequency of arrivals and the variety of welcome scenes, that the Plautine characters show little interest in the foreign aspect of the journeys, the places visited, or origins. These factors become a point of interest when needed to establish the characters’ position in society and their status as slave or free (especially in Menaechmi 1068–95; Persa 595–648; Poenulus 1040–85; Rudens 735–45, 1130–53). Otherwise, for the internal logic of the plays it seems to matter little which city the characters came from, particularly if they were slaves. Plautus’s insistence 17

18

For a catalogue of travels on stage, in each of the plays: Richlin (2017 Table 7.1, 352–4). While not all plays have the same distance covered as those of Curculio, Persa and Poenulus, still, mobility is central to the majority of the Plautine corpus, especially in the following plays: Captivi, Cistellaria, Menaechmi, Mercator, Miles Gloriosus and Rudens. For a list of the main characters and their origins, where known, see Fantham (2004: 237–8).

64

Figure 3.1 Journeys in three plays of Plautus as depicted on a conventional modern map: Curculio, Persa, Poenulus (Adapted from Richlin 2005: 18)

Elena Isayev

that the plays are on foreign soil, palliata – in Greek dress, does not translate to his characters having particular distinguishing ethnic features. That is not to say that foreign cultural traits are absent. On the contrary, bizarre names, Persian and Carthaginian exotic dress habits, and the Punic language do become direct comic targets in the plays Persa (330, 473, 690) and Poenulus (975–1040; Palmer 1997: 31–52). Negative cultural stereotyping in Plautus’s comedies also appears in the metatheatrical context, where the targets of humour become those who are close to home, such as the smelly Roman rowers (1313–14), the pompous Praenestines (Bacchides 21–2; Captivi 884; Truculentus 690), and the Apulians, who have strange habits (Miles Gloriosus, 648). But there is little within these caricatures that fits the category of xenophobic ethnic labelling of the kind that appears in Greek Comedy. There is nothing as direct as the Greek playwright Alexis’s characterisation of Boeotians, as slow, boorish and gluttonous in Trophonius (fr. 239 K-A (237 K)), for example. Nor are there contrasting stereotypes, as in Menander’s Aspis (240–5), of the virile and vigorous Thracians who are pitted against the unmanly, cowardly Phrygians. The portrayal of the ‘other’, of the kind presented in Hartog’s Mirror of Herodotus (1988), is not to be found in the comedies of Plautus. It is absent even in Poenulus, as will be argued below, despite the majority of its characters being of Carthaginian origin. Poenulus is a particularly intriguing comedy, because its hero Hanno is a Carthaginian – an endearing uncle who comes to Calydon in Greece searching for his stolen daughters. The comedy may have been based on a much earlier Greek original with a similar plot, Karkhedonios, probably by Alexis (Arnott 1996; Arnott 2004). But the appearance on a Roman stage of a Carthaginian Hanno in 191 BC, soon after the Second Punic War (Henderson 1999: 8), could not have been a neutral decision by the playwright. Nevertheless, there is no hint that the subject matter of Poenulus may be uncomfortable for the Roman audience. There is no sense that it embodies the recent trauma of the ravages of Hannibal’s campaign in Italy. The foreign hero displays pious and noble traits, which cannot be argued away (Hanson 1959: 92–5; Franko 1996: 441; Maurice 2004: 268). Hanno’s 66

Elusive Migrants of Ancient Italy

devotion to his family and reverence for the gods are exemplary of a virtuous Roman. This is even acknowledged by Franko (1994; 1996), who has argued that the partially derogatory depiction of the Carthaginian characters in the play echoes actual Roman xenophobic attitudes towards Carthaginians in the post-war period. A key point of controversy is the way in which the comedy uses the term Poenus – ‘Punic’. Prag (2006: 5–8, 14–15, 30) in his analysis of its use concludes that it was unlikely to have been a pejorative label at the time of Plautus, and that the comic use of Punica fides – Carthaginian perfidy – is essentially a political claim; a characterisation that is equally applied to Greeks, among others, in the comedies (e.g. Plautus Asinaria, 199; Segal 1987: 38–9). The issue with Poenulus is that it is too much like Plautus’s other comedies. There is just not enough of the stereotyping or the fear and hatred of the ‘other’. We may wonder whether Hanno’s opening speech in Act 5, delivered in Punic, would have startled the Roman audience. How familiar was the sound of the Carthaginian tongue, and was it purely associated with hostility? It may be that the Punic passage was not a Plautine creation but an adoption from Alexis’s Greek original. For some of those watching the comedies – whether as civilians, hostages (Allen 2006: 161–3), refugees or slaves – Punic may have been their mother tongue. It may be that Punic speech brought to mind trade, travel, festivals and distant friends rather than merely hostility. Carthaginians were the trading partners of Italic communities for centuries. Punic inscriptions at such Archaic sanctuaries as that of Etruscan Pyrgi are a testament to these links (Heurgon 1966; Bonfante & Bonfante 2002: 64–8). The interweaving of cultural practices is also evident on a fourth-century BC Tarquinian sarcophagus of Carthaginian design decorated with Etruscan painting (Crouzet 2004; Fentress 2013: 157–78). Punic elements had filtered into Italian practices and even language: from everyday things such as Punic porridge – pultem Punicam (Cato Agr. 85) and specialty objects such as Punic couches – lectuli Punicani (Cicero Mur. 75), to more technical elements such as Punic windows and joints (Cato Agr. 18.9; Varro R. 3.7.3; 1.52), and perhaps even the macellum (Palmer 1997: 43–8, 115–19). Remains of 67

Elena Isayev

possible Semitic/Punic names appear at a number of sites on the west coast of Italy (La Torre 2009: 191), such as the Phoenix who was cursed on a fourth-century BC defixio from Roccagloriosa (Buxentum 3/Lu 45). Romano-Carthaginian treaties, as recorded by Polybius, are a testament to long-term links and the presence of both groups in each other’s communities (Palmer 1997; Erskine 2013: 113–29; Isayev 2017, chapter 8). There is even a suggestion that the Hercules cult at the Ara Maxima in the Forum Boarium may have had Phoenician roots and is evidence of such foreign traders in Rome during the Archaic period (Van Berchem 1959–60; Van Berchem 1967; Forsythe 2005: 119–21). Conversely, long-standing Italo-Carthaginian relations also mean that by this period Italians could hardly have been newcomers to the northern shores of Africa. There is some evidence that there may have been a significant number of Romano-Italian landowners or investors in the region.19 Such investments may have been one of the reasons that the agricultural pamphlets of the Punic author Mago (Varro R. 1.1.10; Lancel 1992: 273–80; Erskine 2013: 117) were translated into Latin in the second century BC (with senatorial funding) by Decimus Silanus, a specialist in Punic language and literature.20 Italian civilians appear to have been in Carthage even on the eve of its fall in 149 BC. Polybius describes how the Carthaginians, realising their fate, threw themselves at the envoys and any other Italians they came across (Polybius 36.7.4–5; Appian Punica 92). A persistent question concerning Plautus’s Poenulus is: Would we notice any distinctive features within it, especially in relation to ethnic stereotyping, if another city of origin was used in place of Carthage for its protagonists? In the comedies there is little to suggest either that those going abroad or those coming into the 19

20

Varro R. 1.10; Columella, de Re Rustica 1.1.10; Pliny Nat. 18.22; Wilson (1966: 51); Kolendo (1970); Lancel (1992: 273–80); Purcell (1995: 135). Italian togati are noted at Citra in the late second century BC, who were massacred by Jugurtha: Sallust, Jugurthine War, 21.2–4; 26.1. Regarding the tophet evidence: Bourdin and Crouzet (2009); Crawley Quinn (2011: 398–9) is cautious about the tophet evidence, while acknowledging the presence of Italian cultural symbols. For a translation of the work into Latin at the expense of the senate: Pliny Nat., 18.22 [1], (1.18); and references to Mago in general: Cicero De Oratore, 1.249; Varro R. 1.1.10; Columella de Re Rustica, 1.1.13; 12.4.2.

68

Elusive Migrants of Ancient Italy

community would have struggled to fit in culturally, or that they would have been at the sharp end of xenophobic harassment. The characters care little for ethnic concerns and in their transfers between different locations continue to function within an apparently unitary cultural milieu. Here the Poenulus is no exception. Hanno on arrival in Calydon displays full knowledge of the legal system and of his rights as a visiting freeborn citizen, while also being aware of the difficulty which outsiders have in bringing cases to trial.21 The most prominent obstacle for the outsider, and for those who have been away for some time, is their lack of knowledge, and networks of hospites whom they can trust and rely on (e.g. Miles Gloriosus 480–530; Poenulus 1003; Mostellaria 473). Taking the perspective of the outsider in Plautus’s Asinaria (495), a trader in Athens, wondering whether to trust Leonida the slave, states that ‘man is no man, but a wolf to a stranger’.22 All those whom one does not know are potentially dangerous, whether native or alien. What is most noticeable throughout the comedies is the nonchalant attitude to foreignness and an assumption of shared understanding along horizontal society lines, even between those separated by vast distances. Mobile Terms We might expect in any society that has specialised terms for the accoutrements of travel such as travel funds – uiaticum (Plautus: Captivi 449; Poenulus 71; Trinummus 720), or a travel bag – uidulus (Plautus: Menaechmi 287; Vidularia) – neither of which exist as a single term in English – that mobility is also ubiquitous. The nature of everyday mobility can be traced in the Latin terminology and expressions used in the comedies of Plautus that refer to the act of moving, its accoutrements, and those who departed and arrived. Even the Latin term migrare and its derivatives reveal a particular understanding of migration, which is distinct from what we understand by it today. A key difference in Latin Republican usage, as it appears in the comedies, is that it does 21

22

Plautus Poenulus 1403–4, Menaechmi 1004–6, and Pseudolus 1232: Gratwick (1993: 231). Plautus Asinaria, 495: lupus est homo homini, non homo, quom qualis sit non novit.

69

Elena Isayev

not make a distinction between a move that is made to a different state or community, and one that is simply to a house next door. The most frequently used derivative is commigrare which defines a move to a different residence with one’s family and belongings (Plautus: Cistellaria 177; Poenulus 94; Trinummus 1084–5). But it could equally be expressed by emigrare or migrare (Plautus: Epidicus, 342; Mostellaria, 470–2). In the writings of Plautus and other Latin authors of the time, migrare simply indicated a physical move from one place of residence to another. It was not used to articulate distance, duration or the crossing of community boundaries. One anomaly is the compound remigrare which appears metaphorically in the prologue of the comedy Poenulus, 46–9. The narrator, recognising that he has gone on for far too long and off the main subject, states that he now wants to remigrare (return) to outlining the plot. The additional language of land surveying that follows the actor’s comment is an allusion to land division which accompanied new foundations, veteran or colonial. In this playfulness with the terminology of mobility, we can see an increasing interest in defining forms of movement and categorising those from abroad, which accompanied Rome’s increasing empire and its establishment as a central Mediterranean hub. There was a variety of Latin terms that were used to refer to the ‘outsider’. The most threatening outsider was called a hostis – an enemy. The opposite is expressed by hospes – guest friend, indicating ties to the members of the host community and the expectation of hospitality by the incomer (Persa 603; Asinaria 417). The juxtaposition of these two is expressed in Bacchides (251–3): tun hospitem illum nominas hostem tuom? – ‘Do you call that enemy of yours your friend?’ At times, the term hostis is used in the Plautine comedies to mean foreigner, but in the vast majority of cases its unambiguous meaning is enemy. Of the terms used by Plautus, the most neutral in the third and second centuries BC was peregrinus, meaning to be from elsewhere or abroad – peregre.23 Other terms such as alienus or ignotus, more specifically, 23

Asinaria 464; Bacchides 1009; Cistellaria 143, 579; Menaechmi 724; Persa 135; Poenulus 175, 599, 656; Pseudolus 1232; Truculentus 955.

70

Elusive Migrants of Ancient Italy

identified someone as being unknown.24 But the categories were not necessarily overlapping, as one did not need to be a foreigner to be a stranger, and both terms could be equally used in reference to a local who was unfamiliar. Considering the variety of Latin terms, it may be surprising that there is no equivalent to the English term immigrant as it appears in current usage, referring to someone who moves across an international border or boundary in a permanent way with the purpose of residence.25 The closest Latin equivalent for the more generic term migrant is transitor – literally meaning he who goes over or is a passerby. But it only appears in late antiquity when concepts of immobility were beginning to be associated with virtue (Ammianus 15.2.4). By the time of Justinian in the sixth century AD people were being scrutinised on entry into Constantinople (Feissel 1995: 366). The categorisation of someone as transitor, in the same way as migrant today, is not a neutral designation. It implies an adverse state and reflects shifting attitudes to mobility, the status of individuals, and methods of control. There were different expressions designating those who lived in the city. These appear as a list in Plautus’s Aulularia (406–7). A mistaken robbery forces Congrio, the cook, to run into the street appealing for help to those around him: cives, populares, incolae, accolae, aduenae, omnes. We can imagine that as the actor shouted this from the stage he would have attracted the attention of bystanders, in the hope that they would join the audience.26 The most literal translation of the passage which allows for the broadest meaning is ‘citizens, compatriots/countrymen, inhabitants/ resident-aliens, neighbours, arrivants/newcomers, everyone’.27 24

25

26

27

For alienus: Captivi 145; Rudens 115; Stichus 480 and Truculentus 178. For ignotus: Curculio 280; Menaechmi 335, 373, 495; Truculentus 178. A phrase also meaning unknown: non novit: Asinaria 495–6. This meaning of immigrant and the related terms immigrate and immigration can be traced to eighteenth- and nineteenth-century America: Pickering (1816: 108); The Oxford English Dictionary. For discussion see: Thompson (2003: 195, n. 21); Shumsky (2008: 132). As also in Captivi 160; Curculio. 462–84: Marshall (2006: 26–8); Wiseman (2009: 169–70). The term advenae, meaning those who arrive, in Varro’s work is applied in reference to certain birds being partly migratory: Varro, R. 3.5.7: . . . cum partim advenae sint. For the meaning of advenae as a reference to those in a city on a more temporary basis and a discussion of different kinds of ‘foreign’ status: Licandro (2007: 55).

71

Elena Isayev

This cry for help to those on the street indicates that there was a certain level of distinction among the city’s freeborn foreigners. In this inventory, the term incolae is of particular importance because Plautus provides the earliest example of its use in Latin literature.28 It appears to distinguish a specific status, perhaps one equivalent to a metic in the Greek context (Kasimis 2018). From later Latin texts we know that its use becomes more defined. In the lex Coloniae Genetiuae of the first century BC (chapters 95.6 and 126)29 and lex Irnitana of the first century AD (chapters 69, 71, 83, 84, 94)30 the term is employed to designate resident aliens, or more precisely those who have transferred their domicilium to a place different to that of their origin (Thomas 1996: 25–53). These later texts indicate that incolae would have had both rights and obligations in their new place of residence. By this period the meaning of incolae became more defined, but its use in the comedies of Plautus is not exclusively reserved for contexts involving foreigners. Its meaning appears to have been more fluid in the early second century BC. In Persa (554–5), incolae is used as a reference simply to inhabitants or residents, without any specification of status: ‘If the inhabitants (incolae) are of sound character, I consider the town well-fortified’ – Si incolae bene sunt morati, pulchre munitum arbitror. Similarly, a general meaning is implied by the use of the verb incolere – to reside – in Rudens (906–7). Not only does the term have diverse meanings but, as Thomas (1996: 28–34) has also noted, even in epigraphic texts it appears without any consistent statutory designation. At the time that Plautus was writing, it is plausible that the term was gaining a more specific definition. This would allow for another layer of meaning in the cook’s appeal to those around him in the Aulularia. The whole list may be a topical play on the emerging status categories in Rome, and their proliferation could be easily turned to comic effect just at the moment when numerous people flocked

28

29 30

The development of the term incolae and its relationship to domicilium is discussed by Thomas (1996: 25–53); Morley (1997: 50–1); Hermon (2007); and Licandro (2007: 45, 51–7). For an overview of its meaning: Gagliardi (2006). Crawford (1996) Law 25, chapters 95, 126 (volume I, 404, 414, 429–30). González and Crawford (1986).

72

Elusive Migrants of Ancient Italy

to take up the opportunities offered by Rome the new Mediterranean hub – the reason that new categories were needed. Conclusion Plautus’s cosmopolitan world, within and beyond the comedies, was filled with characters from around the Mediterranean for whom the poleis acted as intersections on their journeys. There is nothing in the comedies to suggest that there were any stateimposed restrictions on who had access to the city, in the way of border controls. There are instances when port authorities and customs houses are mentioned (Asinaria 240–3; Menaechmi 117–19; Trinummus 795, 1105–7), but they are primarily for monitoring the circulation of goods and resources, not people. There were no mechanisms in place for either hindering outsiders or keeping track of their flows. What numbers we have from this ancient context are accidental, and our interest in using them to generate statistical data as a way of understanding and conceptualising mobility does not reflect ancient concerns or modes of classification. The ‘migrants’ of ancient Italy are elusive because the people who were on the move did not fall into a single category. They were not interesting simply because they were on the move, but were defined by their relationship to the host community, not to each other. That relationship was underpinned by the status of slave and free. For those who were free a series of other classifications could be applied, such as those listed in the cook’s cry for help in the Aulularia (406–7). This comic passage reveals an interest in making more nuanced distinctions between the status of those in the city. In the Plautine comedies there is still flexibility in the way that terms are used; incolae, for example, could simply refer to inhabitants or more specifically to resident-aliens. At the time that Plautus was writing, Roman institutions were adapting to the necessities of an imperial centre that saw an influx of newcomers and a dispersion of its own citizens. It is precisely at this time that Italians become visible abroad, especially in the East Mediterranean, not because they had only just arrived, but because they were now displaying Italo-Roman characteristics that left a mark in the record. The choice to hold on to one’s distinctive 73

Elena Isayev

external cultural signifiers, such as dress, language or name, and to indicate as much in an epigraphic mark, is a sign of privilege and confidence – afforded by being a Roman citizen or a member of another Italian community. The dynamism that has been presented here only begins to incorporate a sense of what it means to have a significant part of the population not only on the move but also in conditions of slavery and displacement. The acting troupes, and the characters they presented on the stage, such as those of Plautus, would have contributed to this group; a group for whom patria may have been unknown or unreachable, and homelands were ephemeral. Memories into the past or those imagined into the future were moments in time rather than in space.

74

C HA P T E R 4

T H E LA N G U A G E O F M O B I L E C R A F T S M E N IN THE W ESTE RN ME DITER RANEAN KATHE R IN E MCDO NA LD AND JA MES C LA CKSON

Introduction Artisans and craftsmen in Southern Italy participated in complex networks of interactions which are not yet fully understood. Although we know the broad outlines of the kind of mobility driven by trade, the movements of individual artists or artefacts are much harder to track and, unlike the careers of elite men or soldiers, craftsmen’s lives are rarely memorialised in literature or outlined on gravestones. Instead, their work provides our main insight into how artisans lived, worked and travelled. The style, function and decoration of paintings, ceramics and other products provide some clues, but text is also used for decorative and practical purposes on a wide range of different objects. Many of these inscriptions show the writer’s familiarity with multiple languages, alphabets or dialects and, in some cases, may show evidence for movement across language or dialect boundaries. Identifying whether products or craftsmen have moved is not straightforward. The migration of the craftsmen themselves and the production of goods for an export market are intertwined phenomena, but need not go together. In this chapter we will present several case studies to show how detailed examination of the language of texts, mostly those inscribed on prestige ceramics or paintings, can offer fresh insight into the mobility of people and objects. Products produced by local makers for a local market can play on the prestige of non-local forms and names, and items meant for export may not remove all traces of the artist’s origin. Artists’ signatures may use Greek morphology on a non-Greek name rather than giving themselves a completely ‘Greek identity’, which suggests that there was value in being (or marketing oneself 75

Katherine McDonald & James Clackson

as) a bilingual craftsman. In other inscriptions, the Greek alphabet appears to act as a ‘scriptio franca’ in which names and other details were comprehensible across a number of languages (including, for example, Greek, Oscan, Messapic and Gaulish), suggesting that artists were aware of how the alphabet was used in the wider Hellenistic world. Even where art is unsigned, inscriptions may show evidence of mobility. Dialect choice may also be revealing about the original provenance of craftsman and makes it possible in some cases to track artists on the move. The complex migration history of some Greek potters can be surmised through consideration of decorative inscriptions – such as ‘speech bubbles’ – on some Italian red-figure pots. Signatures on Italian Ceramics Art in Southern Italy shows particularly complex adaptations and interactions of styles. Red-figure pottery is one good example of this. During the sixth and fifth centuries BC, most of the figured pottery in Italy was imported from Greece, particularly from Attica. But in the second half of the fifth century, redfigure workshops were established in Lucania and Apulia, and by the early fourth century BC, Lucanian and Apulian red-figure pottery had all but replaced imported wares in the local elite market, and had moved away from directly imitating Attic designs.1 Most of this pottery was produced for the local market – less than 1 per cent of Apulian ware has been found outside Apulia – so that we can be fairly sure these craftsmen were catering to local tastes; and, although the craftsmen have sometimes been assumed to be ethnically Greek and Greekspeaking,2 more recent scholarship has deemed this highly unlikely based on the sites where the wares are found.3 Wall paintings in tombs in Southern Italy, particularly those in Paestum and Apulia, are another example of the interaction of Greek and Italian artistic traditions, but one where artists are catering to local rather than export markets. 1 2 3

MacDonald (1981); Carpenter, Lynch and Robinson (2014: 1). Trendall (1989: 17). Carpenter, Lynch and Robinson (2014: 6).

76

Mobile Craftsmen in the Western Mediterranean

A number of the inscriptions that we discuss in this chapter are dipinti on red figure vases or on wall paintings, and so have been written by the same craftsmen who were part of these innovative artistic movements catering to Greek-influenced local elite tastes. Even the act of signing works of this kind can be seen as a Greek practice which was taken up by artists in Italy, no matter what language they used for their signature.4 In a recent book on signatures in ancient Greece, Jeffrey Hurwit argued that signing works of art was an exceptionally Greek practice, unparalleled anywhere in the ancient world.5 However, even in Hurwit’s book, the picture which emerges is much more complicated. Although he suggests that Etruscan and Roman artists’ ‘impulse for self-identification’ was ‘weak’,6 he nevertheless identifies a number of examples of Etruscan and Roman artists who do sign their work.7 His book also shows that, despite the common practice of labelling figures and using writing in art in other ways, Greek artists in both wall painting and vase painting only very rarely signed their work – more than 99 per cent of the surviving Athenian figured vases are unsigned.8 Whether signatures can be considered a clearly Greek practice, particularly in Italy, is therefore a complicated issue. The Onomastic Fallacy Many short inscriptions from the ancient world consist only of a personal name. If that name is written in the Greek alphabet, which was used for a number of different languages in Central and Southern Italy, then there is no obvious clue to the language spoken by the writer other than the apparent etymological origin of the name. But assuming someone’s identity, origins or language from their name is always problematic and leaves significant room for error. Names can be inherited from immediate or distant ancestors who came from another region or speech community, or can be borrowed 4 5 6 7 8

Mazzei (1998: 83). Hurwit (2015). Hurwit (2015: 20). Hurwit (2015: 19, 26–9). Hurwit (2015: 79). Hurwit notes that South Italian red-figure vases make much use of inscriptions, but there are only two named Greek painters – Asteas and Python. For comments on the rarity of signatures in wall paintings, see Hurwit (2015: 63).

77

Katherine McDonald & James Clackson

from prestigious contemporary languages and cultures. Bilingual individuals can have names which reflect one, both or neither of their languages, and individuals may carry a name appropriate to a speech community to which they do not belong. Furthermore, sometimes the modern scholar may lack the requisite knowledge to locate a particular name. Take the example of a Greek–Phoenician bilingual inscription on a mask found at Soluntum in Sicily (SEG 34.973), which includes the name φαδις; Crawford et al. (2011: 1532) suggest that this is an Italic nomen, comparing Latin Fadius, but there is no good comparison from other languages of Italy. It is also important to remember that a signature is a text which has a particular function – it indicates only the ‘customer-facing’ professional identity of the craftsman, and may not be the name he used in the other domains of his life. However, in almost all cases, we can only guess at this: if the craftsman has changed his name to a Greek one and has also used the Greek language in his inscriptions, then any membership of a non-Greek group becomes invisible to us. An example of the difficulties faced by scholars is found in Small’s (2006, 2014) discussion of two imbrex tiles from the area around Montescaglioso in Apulia. One is stamped with the words Δάζιμος κερ/αμεὺς χαῖρε ‘Dazimos the potter, greetings!’ and the other with ΗΡΑΚ[, expanded by Small as either ‘Herakas’ or ‘Herakleides’.9 Small dwells on the origin of the names: ‘The artisan’s name is Messapic, so he was presumably a native Apulian, although he addresses the passer-by in Greek. These relatively humble objects suggest that the use of Greek was common among the Apulian artisan class, even if their proficiency in the language may have been limited.’10 Herakleides, however, he identifies as probably ethnically Greek.11 Small’s overall conclusion that artisans were able to move between Greek and Apulian communities12 is mostly unproblematic. 9

10 11

12

Small (2006: 330–1); (2014). See also Crawford et al. (2011: 1529), who says it ‘could represent any of the names beginning Herak-’. Small (2014: 32); cf. Lo Porto (1988–9: 402–3). Crawford et al. (2011: 1529) do not explicitly commit themselves on the language of this inscription; however, it does appear to be Greek, since in Sabellian the name of the god Herakles is usually spelled here- rather than hera-. However, there is one possible attestation of the Greek form of the divine name Herakles spelled hera- in an Oscan, or partly Oscan, inscription – see Campania or Lucania or Brettii or Sicilia 1, Crawford et al. (2011: 363). Small (2014: 32).

78

Mobile Craftsmen in the Western Mediterranean

But the categorisation of some artisans as ‘Greek’ and some as ‘Apulian’, based on our perceptions of the origins of their names, may in fact create artificial ethnic and linguistic boundaries between people who, at the time, thought of themselves as part of both groups, or did not conceptualise these differences as constituting two different groups at all. Lo Porto (1988–9: 402–3) points out that the name Δάζιμος or Δάσιμος appears several times in Italy, including in the name of the Heraclean ephor Δάζιμος Πύρρω, named in the Heraclean Tables (I 5, 9, 97, II 5 and 8; cf. II 1). Whatever the origin of the name, some of its bearers might well have considered themselves ‘Greek’. There is also no evidence on these tiles of the limited proficiency in Greek that Small suggests. In a highly mobile and multilingual population, this ‘onomastic fallacy’ may lead us to see both more and less multilingualism in the sources than may have existed. As the examples below will show, artists and craftsmen may have made complex decisions about how to present themselves to clients, including making decisions about what their names would be. These decisions are not always straightforward, and may be influenced by the languages spoken by artists and the writing systems in which they were literate, as well as their perceptions of relative prestige, their geographic range, and their familiarity to their clients. As we will see below, artists may be motivated to appeal to an élite which was well disposed towards Greek culture either by using Greek or by Hellenising their names to a greater or lesser extent, even where the artist and clientele were not first-language Greek speakers. Ceramics – Morkos and Pyllos A fifth-century BC pyxis, found in a burial context at Gravina (Puglia),13 bears the following dipinto (SEG 54.955): Μόρκος : ἐποίε̄ · Πύλλος : ἐδίδασκε · Μόρκος Πύλλος : αβγδεζηθικλμν · Μόρκος ἔθηκε Γναίϝαι Morkos made (it). Pyllos was (his) teacher. Morkos, Pyllos, α β γ δ ε ζ η ι κ λ μ ν. Morkos dedicated (it) to/for Gnaiva. 13

de Hoz (2004: 418).

79

Katherine McDonald & James Clackson

The dipinto is clearly in the Greek language, and there are no significant errors that would lead us to believe that the writer was not a first-language speaker of Greek. Nevertheless, the names in the inscription have led most scholars to identify the writer and the two other people mentioned as non-Greek. For example, under a heading ‘the ethnicity of the artisans’, Small writes, ‘the names Morkos and Pyllos are of Messapic type, and Gnaiva is Oscan’.14 This explanation elides the difficulties of straightforwardly assigning these names to particular ‘ethnicities’. In the case of the name Morkos, there is reasonable evidence for similar names in Messapic inscriptions. It appears twice in this form, in MLM 4 Gn and MLM 11 Ro, though there is also an alternative nominative form Morkes (MLM 7 Cae). But the idea that Pyllos is a Messapic name seems to be an argument from silence. It is rare in Greek (with only seven occurrences listed in the Lexicon of Greek Personal Names), but does it therefore follow that it must be a Messapic name? The only similar name found in Messapic is a name which appears to read Poollo, which occurs only once at Alezio (MLM 31 Al, fifth century BC). If this is the same name as Pyllos, then at the very least there has been significant remodelling of the name to make it conform to Greek phonology and morphology. If morkes was the original Messapic form of the name, then the same remodelling could have also happened in the case of Morkos (in this case and all the others where it appears). The morphology of both names could in fact suggest that both were meant as Greek versions of the bearer’s Messapic name, and that it is misleading simply to label these names as ‘Messapic’. Gnaiva (Γναίϝαι) occurs only here in Greek inscriptions. It appears to be the feminine of a common Oscan praenomen gnaivs, which in Oscan inscriptions in the Greek alphabet is written either with upsilon or digamma.15 It is worth noting that the female version of this name is not attested in any Oscan 14 15

Small (2014: 32). Attested in Oscan as gnaivs (Cumae 8, Caudium 2), γναυ(σ) (Petelia 2), γν[αϝ]ις and [γν]αϝ[ις] (Laos 3), cnaives (Nola 11) and as the abbreviation gn (Cumae 11, Pompei 96).

80

Mobile Craftsmen in the Western Mediterranean

inscription, so the existence of this name is probable but not definite.16 If Morkos and his companion have deliberately adapted their names, what aim did they have in mind? There are two main possibilities: either they wanted their names to be easier to pronounce and decline for Greek speakers, or they wanted to appear more ‘Greek’ to their local clients by having Hellenised names (or both). In a highly bilingual and well-connected area of Italy, they would probably have been aware that the names Morkos and Pyllos would not make them sound completely Greek to a native Greek speaker; to achieve this, they would have needed to adopt a fully Greek name as their professional identity. Instead, Morkos and Pyllos have names that signal their membership of a ‘globalised’ network, where arguably the ideal was to seem a little bit ‘foreign’ to everyone.17 Tomb Paintings – Plasos A similar explanation may help us to understand the signature on a wall painting at Paestum. On one wall of Tomb 1/1972, dated to around the third century BC, there is a scene which can be interpreted either as a warrior departing or returning home. The warrior, riding a white horse, is in the centre of the scene and faces a woman who stands on the right. Behind the warrior and the horse stands a smaller figure, probably a groom or a slave. There is a signature in Greek letters above the head of this left-hand figure which reads πλασος (plasos) (Paestum 3, Figure 4.1).18 The tomb is usually dated to the Oscan-speaking era of Paestum based on the style of painting and the goods found in the tomb. The name Plasos, however, is morphologically Greek rather than Oscan. The 16

17

18

It is not completely clear whether Oscan women had praenomina, and this may be a naming practice that survived in some areas but not in others. There are at least two good examples in the southern part of Oscan-speaking Italy, νοψα and ϝιβια, both in the curse tablet Laos 2 (Crawford et al. 2011). Elsewhere, names derived from praenomina, like uibia, seem to be used as individual names. Female praenomina are found regularly in Paelignian inscriptions – for a summary, see Buonocore and Poccetti (2013: 63–4). For this impulse in the art and ceramics of the Roman Empire, also using early modern ‘china’ trade as a comparandum, see Versluys (2015). For a case study of this at Pompeii focusing on an alabaster statue of Horos, see Mol (2013), who points out that this statue would be distinctly ‘foreign’ in both Egypt and Italy, and therefore desirable in both. Crawford et al. (2011: 1324–5); McDonald (2015: 219–22).

81

Katherine McDonald & James Clackson

Figure 4.1 Plasos’ signature (Image: Katherine McDonald)

use of an individual name rather than a two-part name indicates either a Greek or Messapic name, or – in the Oscan-speaking world – the name of a slave, which is a plausible enough social background for an artist in ancient Italy.19 Elsewhere Plasos does not appear as a Greek individual name, and may therefore be of Oscan or other Italian origin. A notable possibility is that this name is a remodelling of the common Messapic name Plator, with an eye to making the name more typically ‘Greek’. Again, it is possible that Plasos has taken care when signing his work to choose a ‘Greek’ professional name to fit with the purchaser’s expectations of quality goods, even where the name is not Greek in origin.20 Tomb Paintings – Artos One Messapic inscription seems to show this use of Greek morphology in an artist’s signature even where the inscription is in the 19

20

Oscan names typically have a praenomen and gentilicium, in a similar way to Roman names of the same period. For the name Plator used as an Oscan praenomen, see Plator Skalaponies (Lu 41 / Tegianum 1). Compare also the fact that craftsmen at La Graufesenque in Gaul used Latin rather than Gaulish names on tiles intended for the wider market. Adams (2003a: 705–6).

82

Mobile Craftsmen in the Western Mediterranean

Messapic language. The decision to use a non-Greek name with Greek morphology while still using the Messapic language is a complex one. A third-century dipinto reading αρτος πιναϝε or αρτος πιναιος (artos pinave or artos pinaios) (MLM 10 Ar), discovered in the vestibule of the ‘Ipogeo della Medusa/Hypogeum of the Medusa’ in Arpi,21 is thought to be the signature of the artist who decorated the tomb.22 The inscription is on a pinax (rectangular painted field) on the left-hand side of the rectangular vestibule of the tomb. The pinax is delineated with a brown line on a red background; the image shows a male figure wearing a white tunic with a black border, with a groom behind him carrying a large shield and leading a horse with a brown mane.23 Immediately above the first man is an inscription ‘in caratteri greci’, read by De Simone as Artos pinave, probably referring to the painter of the image on the pinax. De Simone notes that the form of the name (Artos and not *Artas) shows Greek influence;24 yet the verb pinave is Messapic.25 21

22

23

24

25

This is a large chamber tomb with a long dromos. The tomb takes its name from the head of Medusa on a pediment near the entrance, but it is best known for the paintings found on the interior, which are considered highly significant in the development of wall painting in ‘Hellenised’ Italy. The tomb had already been robbed out when it was excavated by the Soprintendenza Archeologica della Puglia in the 1980s. The tomb was constructed in the first half of the third century BC, and was used until around the end of the second century BC, when it was abandoned. See Mazzei (1995, 1998). De Simone (1995). (Adams 2003a was not aware of this publication, but cites the Messapic language article by De Simone in the OCD, third edition.) The second scene, to the right-hand side of the vestibule, is less well preserved but appears to depict a funerary scene, perhaps a journey to the underworld. This kind of pinax is rare in Italian wall paintings of this period, and is more often associated with the Roman ‘second style’ (second–first centuries BC). Depictions of horseman, horses and warriors, however, are very common in wall paintings, and are found throughout the Messapic-speaking area, as well as in other areas of Italy such as Paestum (Mazzei 1995, 1998). We might note that Artos is not used elsewhere as a Greek name, but it is a Greek word – ‘bread’ – which may have influenced the spelling here (thanks to Patrick James for this suggestion). The ‘Artos Atotios’ inscription (MLM 12 Ruv) is also a good example of how this Greek morphology might go beyond just texts naming artists, to dedicatory texts in the Messapic language, just as we saw with Morkos appearing in multiple kinds of inscriptions – if we believe that this is also a remodelled form. De Simone (1995). On pinave: De Simone explains this as a third-person singular perfect ending in -ve, found elsewhere in verbs including dupave, hadive, preve and the borrowing from Greek ipigrave ‘wrote’. He suggests that the verb is from a root *peiḱ-/*piḱ-, as seen in Latin pingit with a nasal infix. He explains the lack of a reflex of *ḱ as mostly likely the result of an assimilation from *pinkave > *pingave > pin(n)ave (De Simone 1995: 212). An alternative reading, given by De Simone and Marchesini

83

Katherine McDonald & James Clackson

This signature has repeatedly been seen as a sign of both local pride and use of Greek practices and language to give an impression of prestige. This Artos pinave is one of only a few examples of artists’ signatures on wall paintings in Apulia.26 Mazzei suggests that this adoption of the practice of signing artistic works was the adoption of a Greek practice and is indicative of a change in cultural attitudes to artists, affording them more prestige.27 According to Massa Pairault, the artist wrote ‘painted’ rather than ‘made’ as a sign of pride in his work.28 De Simone (1995: 212) also suggests that the ‘Graecised’ name is a sign of linguistic interference. All of these explanations for the language of the inscription may be true in part. We should also consider the overlap opportunities provided by the Greek alphabet here. Artos could see that anyone literate in the Greek alphabet – whether they were speakers of Greek or Messapic – could read his name. Even for a Greek speaker who could not understand the verb (if the second word of the inscription is a verb), the message was perfectly clear. For a Messapic reader who could read the verb, the morphological ending of the name might have signalled a connection to a wider Greekspeaking world without being fully ‘foreign’. Mobile Craftsmen? Plator Not all craftsmen in Italy adopt this practice of changing the morphology of their name. For example, a potter at Teanum Sidicinum around 300 BC wrote his signature in both Oscan and Greek, but seems only to have used a Greek alphabet, whatever language he was writing in. We find πλατωρ ουψε (plator oupse) ‘Plator made (this)’ (Oscan) on a decorated black slip plate found

26

27 28

(2002), is Artos pinaios (MLM 10 Ar). The second word of the inscription, particularly its morphological ending, is hard to make out from the photographs provided in Mazzei (1995) and De Simone (1995), so we would not want to rely too heavily on this in any interpretation. However, Artos is clear and, in particular, the fourth letter cannot be alpha. Mazzei (1998: 74). Others in Hellenistic hypogea in Puglia include one in Rudiae (the fondo Acchiatturra, carved and painted on the façade), one in Lecce (the Hypogeum of Palmieri) and one in Egnazia (the Hypogeum of the Inscriptions). Mazzei (1998: 83). Massa Pairault (1997: 63–4).

84

Mobile Craftsmen in the Western Mediterranean

Figure 4.2 Dish with Plator’s signature (From Crawford et al. 2011: 561)

Figure 4.3 Plator’s signature (From Crawford et al. 2011: 561)

in Tomb 58 (Figures 4.2 and 4.3) and also πλάτωρ ἐποί[η]σε (platōr epoi[ē]se) ‘Plator made (this)’ (Greek) on a plate found in Tomb 29.29 There are many suggestive details to these inscriptions. Crawford suggests, because of the association of the name Plator with Messapic, that he may have come to Teanum from Apulia.30 29 30

Gàbrici (1910: 30). Oscan inscription: Teanum Sidicinum 25. Crawford (2011b: 4); see also Zair (2013: 220).

85

Katherine McDonald & James Clackson

The inscriptions show that he was probably more used to writing Greek than Oscan. Even though the Greek alphabet was commonly used for Oscan in Lucania and Bruttium, in the area around Teanum the Oscan ‘Native’ or ‘National’ alphabet (an alphabet based on the Etruscan alphabet) was the usual writing system for Oscan. The Oscan perfect ending, usually written as -εδ in the Greek alphabet, has been replaced with a Greek secondary tense ending -ε.31 Additionally, the spelling of ουψε suggests a writer more used to writing Greek: in Oscan, this verbal root is found spelled -πσ- (except at Messina), suggesting that the -ψ- spelling is an ad hoc application of the Greek alphabet to Oscan words.32 Why did Plator sign his work in both Oscan and Greek? If one product was destined for a local market, and the other for export, it is curious that they are both of a very similar style and that they both ended up in graves at the same site. Rather than seeing Greek as signalling ‘export’ in this case, we can see both plates as intended to appeal to the local Hellenised élite, in much the same way as Plasos and Artos the tomb-painters must have been doing. We also have the possibility here of seeing the movement of a craftsman across linguistic barriers, and possibly being able to give his products an air of imported Greek goods. Mobile Products and Missing Morphology One strategy for artisans signing their work, particularly with moveable products, is to avoid using any morphology at all. This is particularly common on tile and amphora stamps, where space is at a premium, though it is also used on other kinds of objects. The first part of the name is usually sufficient to identify a craftsman or workshop. A side effect of this strategy is that, where the Greek alphabet is used, it becomes extremely difficult to identify the language of the inscription. We must at least consider that this could have been intentional, and that craftsmen living in bilingual communities or those who intended their wares to travel over linguistic boundaries saw the 31 32

Zair (2013: 220). Zair (2013: 220–2).

86

Mobile Craftsmen in the Western Mediterranean

advantage of signing their names in ambiguous ways. It may be that there is no need to try to establish the ‘language’ of very short inscriptions, especially those written in an alphabet as widely known as the Greek alphabet. For example, it is not always possible to know whether tile stamps in the south of Italy are intended as Greek or Oscan inscriptions, as both languages use (very similar) versions of the Greek alphabet across most of this region. A case in point is the tile stamps from Tiriolo, which read (Teuranus Ager 2 and 3) and (Teuranus Ager 4), and it is hard to ascribe either of these abbreviations to a specific language with absolute certainty. More likely, they were abbreviations that acted as a logo and identified a particular manufacturer. The stamp Caulonia 5 (?) has been read as both Greek and Oscan, depending on whether the second letter is read as a three-barred sigma or an Italic .33 If it is the latter, then it could only represent an abbreviation of an Oscan name, either in one or two parts,34 while a sigma could belong to either language. If we try to assume the language of a tile stamp based on other inscriptions from the site, we run the risk of erasing evidence of bilingualism and, particularly, texts which show an engagement with the wider Hellenistic world. Similar considerations apply to other kinds of commercial inscriptions, such as short inscriptions on amphorae. An amphora (Corinthian type A) from the Tomba delle Anfore in Arpi bears what could be a Greek inscription, written by the dispatcher, or a Messapic inscription, written later in the amphora’s journey (MLM 11 Ar: ΗΑΡ / ΑΒΓΔ). We know there is an import–export relationship here, based on the trade in Greek wine. It is worth considering whether the exporters in Greece knew, on some level, that labels and manufacturers’ marks in the Greek alphabet were functional in Apulia as well, whether or not they had a clear idea of the Messapic language. We should also consider how commercial

33 34

Crawford et al. (2011: 1490). A two-part name, abbreviated to its initials T.F., is perhaps more likely, as there are no attested Oscan names beginning with t that have an f in the middle, though we could suggest an Oscan equivalent of the Latin Tiberius, which is attested as Tefarie in Etruscan.

87

Katherine McDonald & James Clackson

labels like this might relate to the evidence of bilingualism that we already saw in the artist’s signature from Arpi. Luxury Goods Possible examples of deliberate linguistic ambiguity are not limited to labels on ceramics. An inscription on a silver cantharus shows that this is a possible interpretation of inscriptions on highquality luxury items as well. The cantharus of Alesia was (probably) discovered in 1862 at the excavation of Alesia, in Southern Gaul, although there have been questions over its authenticity and, if genuine, its date.35 Its deposit in a ditch may be associated with the siege of the town during the period of Julius Caesar’s command. It is usually believed to have been produced in Italy and then imported, given its similarity to Italian goods such as the Boscoreale treasure, although it dates to at least a generation earlier.36 The cantharus bears three inscriptions underneath the foot, all three of which may be by different hands at different dates. Text A appears to be a nonalphabetic mark denoting the weight, and is probably contemporaneous with manufacture. Text C is not understood, and we do not know how closely it is supposed to relate to the other texts, and at least some of it may have been added by a later hand. Text B is in the Greek alphabet and consists of two groups of letters clearly separated by a space. The letters themselves are clear and read μεδα αραγε. But it is possible to analyse the language of the inscription in different ways:37 35 36 37

Lejeune (1983: 26). Lejeune (1983: 26, 33); Reddé and von Schnurbein (2001); Crawford et al. (2011: 1314). The case for the inscription being Gaulish relies on (a) the findspot in Gaul and (b) the attested name Medamus, which could be Celtic. If the second word is indeed an abbreviation for ‘silversmith’ rather than part of a name, then we can compare it to the Oscan for ‘money’ aragetud (Nola 3/Cm 6, Nola 2/Cm 7), cognate with Latin argentum ‘silver’, though we do not have this word attested in Oscan as meaning ‘silver’ (Bakkum 2009: 180). The clear division into two sections of text makes Crawford’s interpretation of a two-part Oscan name somewhat more difficult, though not impossible; it is possible that it is a single name only and that the maker could have had a single Greek idionym, or an idionym of Oscan origin but Hellenised in style. Since Lejeune’s 1983 article, few if any scholars have considered the possibility that the inscription is Gaulish; because it has been abbreviated, it is not possible to know the etymology or origin of the name.

88

Mobile Craftsmen in the Western Mediterranean Lejeune (1983) either gaulish: μεδα(μος ?) αραγε(νι? -νιος ?) – name plus patronymic or oscan: μεδα(τιες) αραγε(τασις) – Oscan gentilicium plus ‘silversmith’ Crawford et al. (2011) oscan: με(τις) δα(τιες) αραγε(τασις) – Oscan praenomen and gentilicium plus ‘silversmith’

As in our previous examples, the script doesn’t help us establish the language of the inscription, the morphology has been omitted, and, in this case, the object is associated with at least two different areas. Lejeune argues that, if the inscription is Oscan, it is most likely to have come from Messina, which was well connected and is known to have had silversmiths, though there is no specific evidence that points to this as the original production site.38 Given the active trade networks and the mobility of objects around the Mediterranean, it is possible that a silversmith in Southern Italy would have an eye on the export market when producing a high-quality piece like this. He might be aware, even dimly, that there were not just Greek-speakers, but speakers of other languages around the Mediterranean who would be familiar with the Greek alphabet. Just like Morkos, Pyllos and Plator (the makers of ceramics discussed earlier in the chapter), he could therefore sign his name in such a way as to appeal to as many possible markets as possible, and make his name comprehensible across linguistic borders. If this interpretation of the text is correct, then this example goes to show the importance of the Greek alphabet, perhaps even more so than the Greek language, as a mode of communication across language barriers in the Western Mediterranean at this date. In her discussion of a different mixed-language text from southern Gaul, Mullen states that ‘[w]e should arguably not be seeking to disentangle something that may be deliberately entangled’.39 This is a key insight for framing our discussions of texts from multilingual areas. If it proves difficult or impossible to decide 38 39

Lejeune (1983: 43). Mullen (2013: 65).

89

Katherine McDonald & James Clackson

what language an inscription, or part of an inscription, is meant to be in, then we should consider that the writer deliberately left the text open to multiple overlapping interpretations. Dialect Mixing For many of the ancient languages which survive in epigraphic texts from around the western Mediterranean, there is insufficient material to enable the construction of secure dialectal areas or isoglosses, let alone relate an individual inscription to a particular dialect. The evidence for dialectal variation in Faliscan, Gaulish, Iberian, Messapic, South Picene, Sicel and Umbrian (not to mention many other minor languages) is either too geographically restricted, or the texts are not well enough understood, to reach any secure conclusions about dialectal features.40 A deviant feature in one text might be indicative of an areal distinction, but it might also correlate with the social status of the writer, or simply be an orthographic mistake. For some languages, such as Oscan, Etruscan and the Latin of the Roman Republic, it has been possible to isolate a few plausible dialectal features.41 Occasionally, it is even possible to identify objects which appear to have dialectal features in these languages at odds with the dialect of the provenance; for example, a stamnos found at Capua (Capua 35 in Crawford et al. 2011), carries an inscription with features that are not typical of the Oscan of Capua.42 However, the paucity of material of this type makes it difficult to reach any general conclusions about dialect mixing in any language other than Greek. For Greek there is a rich surviving inscriptional evidence of dialects both in the mother cities and in the colonies in the western Mediterranean, and it is this textual material which allows the

40

41

42

For Faliscan, Bakkum (2009); for Gaulish, Lambert (2003: 19); for South Picene, Adiego Lajara (1992); for Umbrian, Crawford et al. (2011: 97). See Rix (1996) for Oscan, van Heems (2011) for Etruscan, and Adams (2007: 37–187) for Republican Latin. Capua 45, dated to 275 BC, around 150 years after Capua 35, also seems to be written in a non-standard Oscan; Eichner and Frei-Stolba (1990) note that the same family may be indicated on the two objects (Perkium on Capua 45, Peracis on Capua 35), but it is unlikely that any dialectal peculiarities would be retained within a single family for this length of time.

90

Mobile Craftsmen in the Western Mediterranean

researcher to understand how place, language and migration interact. We shall concentrate in this section on how dialectal choices and dialectal mixing in Greek texts can be used to contribute to a picture of migration amongst craftsmen. It has long been recognised that Greek artists working in a different environment from their native poleis generally retain their native dialect if they signed their work themselves, although they sometimes adapt their language to local practice, particularly in order to avoid a form which was peculiar to a single dialect and might not be well understood. Buck (1913: 139–43) provides a number of examples both of retention of the local dialect and also of the second process, which can now be recognised to belong to the range of behaviours generally grouped under the term ‘accommodation’ in modern sociolinguistic studies.43 Buck’s examples of the retention of the artist’s local dialect include a text on a statue base found in Delphi, which was dedicated by Glaukies of Rhegion but made by the sculptor Kalon from Elis, in the north-western Peloponnese (I.Rhegion 66 = IvO 271 = IGASMG III 65 = CEG 388 = IGDGG I 36 = Minon (2007), text 62); L. H. Jeffery’s drawing is reproduced in Figure 4.4). Pausanias (5.27.8) reports that the statue, now lost, featured Hermes holding the caduceus. [Γλαυκί]αι με Κάλōν γεν[εᾶι Ϝ]αλεῖ[ο]ρ ἐποίε̄ 44 [Γλ]αυκίης ὁ Λυκκίδεō [το͂ ]ι Ἑρμῆι Ῥ[η]γῖνος. ‘Kalon the Elean made me for the family of Glaukies. Glaukies the son of Lukkides of Rhegion (gave me) to Hermes’.

Kalon uses the Elean dialect in the description of the manufacture of the object, with its characteristic rhotacism of final -s, hence Ϝ]αλεῖορ ‘Elean’,45 and retention of original long a in 43

44

45

The term ‘accommodation’ in sociolinguistics reflects the work of Giles and associates in a number of studies; see, for example, Giles, Coupland and Coupland (1991). IGDGG I 36 and Minon (2007) no. 62 read ἐν Ϝά]λι με Κάλον γενε[ᾶι Ϝ]αλεῖος ἐποίε, following autopsy by Minon, who confirms the lack of a bar in the first visible character (as in Jeffery’s drawing) and sees a sigma in what is drawn by Jeffery as the remains of a tailed rho. Despite the missing crossbar (for which compare line 3), the reading given in the main text is preferable. Unless Minon’s reading is adopted.

91

Katherine McDonald & James Clackson

Figure 4.4 L. H. Jeffery’s drawing of Kalon’s signature and Glaukies’ dedication (Drawing held at the Anne Jeffery Archive, courtesy of the Centre for the Study of Ancient Documents, Oxford University)

[Γλαυκί]αι and Ϝ]αλεῖ[ο]ρ. In the next two lines (which are divided from the first by a horizontal incision), Glaukies’ dedication switches to the Ionic dialect of Rhegion with long e from original long a represented here by and first declension genitive singular of masculine nouns in -εω (here written ). Note, however, that the surviving letterforms are similar to the Elean part of the inscription, with the tailed rho (apparently) used in both scripts, and the Ionic text avoiding the lunate gamma, for sigma and for delta characteristic of other early inscriptions from Rhegion. Other than this statue, the artist Kalon is known only from another reference in Pausanias (5.25.4), where he is credited for bronze statues set up in Olympia to commemorate a chorus of thirty-five boys, their flautist and trainer, all of whom perished in a storm while travelling across the straits from Messena to Rhegion.46 Kalon presumably was known to the Rhegines, but there is no need to suppose he ever moved away 46

See Dörig (1976: 127) for arguments against the thesis that our Kalon is the same as Kalon of Aegina; this may also be the Callon who is claimed by Pliny Nat. 34. 49 to have been active in the 87th Olympiad (432–428 BC).

92

Mobile Craftsmen in the Western Mediterranean

from Olympia, and it is even possible that Glaukies bought a signed statue and then added his own dedication. Not much remains of the epigraphy of Rhegion itself, although there is some possible evidence of dialects other than Ionic spoken there before its later sack and refounding by Syracuse in the fourth century.47 Sometimes, the inscription giving the name of the artist might avoid any particular local peculiarity of dialect in order to appear less distant from the dialect of the community in which the work would be displayed, a practice which recalls the downplaying of dialectal peculiarities in inscriptions in verse from around the Greek world (Mickey 1981).48 Indeed, the best example given of this practice from Buck (1913) is that of pieces by the Parian sculptor Euphron and the Cretan Cresilas, and most of the examples involve verse dedications. At other times it appears that someone other than the artist might add the artist’s signature, as is the case of the Attic sculptor Phaidimos, whose name appears on three works, one of which is in a different hand (Jeffery 1990: 63). The avoidance of peculiar regional letterforms in I.Rhegion 66 discussed earlier may be explicable in this light.49 Dialect choice on inscriptions on pottery might also give indications of the movements of artists. Most pottery produced in the Greek mainland intended for the South Italian and Sicilian market retains the dialect of the home of the producer. Thus pots imported from Athens have Attic legends, and those from Corinth exhibit the Corinthian dialect, complete with the local peculiarities of its alphabet. There are, however, occasional instances of a pot that 47

48

49

Dell’Oro (2017: 175–6) wonders whether the apparent rhotacism in the name Δεμ(ο)φάνην Θράρυος (for Θράσυος, I.Rhegion 26b = IGDGG I 39c) reflects a later immigrant from (Eretrian) Euboea, but it is probably better to follow her second explanation that this is a miswriting for Θάρρυος. The defixio I.Rhegion 19 contains Doric forms (παρκατίθημι and πὰρ Γᾶν) but this text is probably from after the refounding of the city from Syracuse. Mickey’s conclusions do not, however, hold good for all parts of the Greek world, as shown by recent work on the Argolid by Alonso Déniz and Nieto Izquierdo (2009) and on Laconia by Guijarro Ruano (2015). We thank Livia Tagliapietra for these references. As is the lack of rhotacism in Kalon’s inscription if we follow Minon’s reading Ϝ]αλεῖος. Livia Tagliapietra (personal communication, 22 October 2017) notes that it is also ‘perfectly plausible that the absence of regional letterforms in this inscription simply reflects the initial spread of the Eastern Ionic alphabet across the Greek-speaking world around this time’.

93

Katherine McDonald & James Clackson

can be associated with a known Athenian painter bearing a legend in a different dialect. It is important to be cautious in the assessment of this evidence. One case which shows dialect mixing probably unrelated to migration is a psykter, or wine cooler, signed by the Athenian painter Euphronios, found in a Cerveteri tomb in the nineteenth century depicting hetairai drinking and playing kottabos.50 One woman, identified as Smikra, has a ‘speech bubble’ beside her written in Doric: τὶν τάνδε λατάσσω Λεάγρε ‘I am making this throw for you, Leagros’.

The other inscriptions on this vase are all Attic (or Attic-Ionic, note names such as ΑΓΑΠΕ = Ἀγάπη with η not α). Kretschmer (1894: 87) suggested that the switch into Doric reflected the fact that the game of kottabos was invited by the Sicilians, and so the players used the Doric of Sicily, much as the sport of fencing uses French terms such as ‘Allez!’ and ‘En garde’. But other vases depicting the ‘Kottabos Toast’ have a different formula τοὶ τήνδε, with Attic τήνδε, and a less ‘broad’ Doric τοί for Attic σοί ‘to you’.51 The use of the Doric dialect here is thus probably part of a literary or cultural tradition, although it is also possible that Euphronios was aware of dialectal differences among the potential buyers of his wares in the Greek colonies of Magna Graecia. Pots may also provide our best epigraphic record of wandering artisans, whose own idiolect reveals their background through a combination of different local features. Wachter (2001: 270–3) considered in detail the language of one maker of painted and inscribed pots, known as the Inscription Painter, whose idiosyncratic language appears to be the result of his own transient career. The Inscription Painter was active from the middle of the sixth century BC and is the chief exponent of a distinctive style of decoration, now known as ‘Chalcidian’, even though no longer thought to have any particular association with the Euboean city of Chalcis. The eighteen inscribed vases attributed to the Inscription Painter contain only names, mostly those of gods and heroes 50 51

Leningrad 644 (St. 1670), Beazley number 200078. Csapo and Miller (1991: 373) gather other examples of the Kottabos Toast. Their theory that τοί is a form of the article and not a ‘literary’ form of σοί seems unlikely; τοί occurs frequently in Doric and Ionic poetry.

94

Mobile Craftsmen in the Western Mediterranean

familiar from Greek mythology, but some are ‘throwaway’ names, apparently invented as appropriate for the depicted scene. Examples of the latter from the Inscription Painter are a maenad named Ϝιō ‘Violet’ and a satyr called Με´̄ ξᾱς, which Wachter derives from μείγνυμι ‘have sexual intercourse’, both on a krater from Vulci.52 The Inscription Painter, like the other artists of the ‘Chalcidian’ school, uses an early form of the Euboean alphabet, and his dialect is basically Ionic, with the majority of cases of original long a represented by , showing the Ionic change of vowel quality. However, on occasion the Inscription Painter employs a non-Ionic form, sometimes giving two different forms of the same name on different vases. Thus the hero Geryones is dubbed Γε̄ ρυόνε̄ ς on Wachter (2001) CHA 9 but Γᾱρυϝόνε̄ ς on Wachter (2001) CHA 2, the latter with for original long a in the first syllable and the non-Ionic phoneme /w/, written , employed as a glide. Wachter rejects earlier attempts to see literary or cultural reminiscences in the mixed dialect of the Inscription Painter, along the lines we have suggested for the Doric of the Kottabos Toast above, but instead he wonders whether the potter’s language might reflect migration between different areas. Wachter notes that the endings of names are generally Ionic in form, but non-Ionic forms occur in the stems of the words (as with the name Geryones), and furthermore that the potter makes occasional mistakes in his use of script, sometimes employing a Corinthian rather than a Euboean letter. He thus concludes that the Inscription Painter was in origin a native of Corinth, who then travelled to an Ionic-speaking area where he perfected his craft; the ‘Chalcidian’ style of pottery and decoration also combines Corinthian and Attic styles. There is little evidence for other painters of the ‘Chalcidian’ school, but it may be significant that one other artist of this group, known as the Cambridge Painter, also shows some dialect conflation in his texts, using digamma as well as Euboean letterforms. It is not clear where the actual production base of the painters of the Chalcidian school was, but it is not impossible that these were 52

CHA 14 in Wachter (2001). In what follows we shall use Wachter’s references to the texts.

95

Katherine McDonald & James Clackson

immigrants from Greece who set up a workshop in Magna Graecia. Dell’Oro (2017: 172–5) accepts Wachter’s arguments and offers further support on the basis of an early Corinthian presence in the Euboean colony of Cumae. Dell’Oro’s evidence from Cumae, however, is difficult to assess. It consists of two partial abecedaria written contiguously (IGASMG III App. II, 113, A1, b-c = IGDGG I 11b); the first of these has Euboean letterforms, the second a Corinthian beta (but otherwise Euboean forms, although badly executed). This text is dated to the protoGeometric period, and consequently probably 150 to 200 years earlier than the ‘Chalcidian’ vase-painters, and cannot be directly related to them, and the Corinthian beta may reflect the incompetence rather than the provenance of the inscriber. Conclusions Robert Garland, in his recent book Wandering Greeks (Garland 2014), attempts to flesh out the characters of the different members of the Greek diaspora and early colonists. Chapters in Garland (2014) carry titles such as ‘The Settler’, ‘The Wanderer’, ‘The Deportee’, ‘The Fugitive’ or ‘The Economic Migrant’. There is, however, no chapter entitled ‘The Artist’, and only two pages (Garland 2014: 165–6) are given to tracking the movement of craftsmen from Athens to colonies. The only literary source that Garland mentions for movement of artists, following MacDonald (1981), is a passage of Thucydides (7.27.5) in order to support a theory that many of the potters and painters who travelled may have been slaves. Unfortunately, Thucydides is not much help here, since he is describing the defection of ‘over 20,000’ slaves from the silver mines at Laurion Attica to the Spartan fort constructed at Decelea; the information that a large part of these slaves were χειροτέχναι is likely to refer to the agricultural or possibly mining workers rather than artists.53 We are reliant, therefore, principally on epigraphic evidence in order to track the movements of craftsmen in the western Mediterranean. As shown in the examples presented in this 53

See Hornblower (2008) ad loc.

96

Mobile Craftsmen in the Western Mediterranean

chapter, this type of evidence presents problems of interpretation since inscriptions on pots and paintings are mostly extremely short, and it is unusual to have more than one signed piece from an individual craftsman. Plator of Teanum Sidicinum and the Inscription Painter are valuable exceptions to this rule. Moreover, in general artists did not want to call attention to their dialect or origin as much as their name. The employment of Greek by non-Greek speakers and the suppression of tell-tale non-Greek morphology thus complements the avoidance of unusual dialect forms and peculiar letter shapes among the texts written by Greeks. Just as at a later period of history, the Celtic and Roman potters at La Graufesenque in southern Gaul chose Latin as the language to present themselves to the wider export market,54 so in the period of Greek cultural dominance in Italy in the preceding millennium, Greek was the default language of the artisan. It is only by unpicking some of the linguistic minutiae that we can uncover some of the evidence for mobility from these texts.

54

Adams (2003a: 719).

97

CH A P T E R 5

L O ST – A N D F O U N D – I N T R A N S M I S SI O N The Creation of the Oscan Alphabet KA RIN W. T IK KAN EN

Introduction As has often been remarked, Oscan inscriptions originate from a large area – Campania, Samnium, Lucania, Bruttium and Sicily – but despite this are remarkably similar in terms of language and spelling, with very little regional variation (see e.g. Rix 1996). Likewise, the Oscan alphabet was created in a context of multiple languages and scripts, the result of a long period of successive migrations to the southern half of the Apennine peninsula; the Etruscans, from Etruria; Greek settlers, from different Greek citystates; and later on the Samnites, from the heights of the Apennine mountain range. Isayev, in Chapter 3 of this volume, discusses the visibility of migrants in the third and second centuries,1 and also the perceptions of mobility and its visibility. The same problem outlined by her, the lack of physical state borders and checkpoints, also holds, in Campania and the Bay of Naples in the seventh and sixth centuries, in terms of cultural boundaries. Bickerman (1952) describes a fluid state of transition of narratives from the Greeks to the Romans in that the latter adopted legendary foundation myths from the former, from the very arrival of Greek colonists to the peninsula. Malkin (2002) detects a similar situation regarding the cultural exchanges between the Greeks, Etruscan and the early Sabellic communities in the region, before the advent of the Samnites and the establishment of the Oscan language over the vast area it was later to cover. 1

Unless otherwise indicated, all dates in this chapter are BCE.

98

The Creation of the Oscan Alphabet

The multilinguistic and multicultural environment of the area around the Bay of Naples favoured exchanges on a cultural but also an alphabetic level, since available knowledge and habits were more or less freely exchanged. With time, however, certain political and cultural boundaries did appear. One such division evident within the later Oscan-speaking linguistic continuum is that between the regions that came to use an Etruscan-based script, and those where adaptations of various Greek scripts were used. The dividing line between the areas was that created by the Silarus (Sele) and Aufidus (Òfanto) rivers, drawing an approximate line across the peninsula from Salerno to Barletta (cf. Lejeune 1970: 271–2). In the fifth century this boundary functioned as the demarcation line between the Etruscan and Greek spheres of influence, and in subsequent centuries the boundary remained in the sense that, in the south, scribes continued to use an alphabet derived from the East Ionic alphabet (cf. Zair 2016: 3), whereas north of this line the speakers adhered more to the Etruscan alphabet, with a few exceptions. The first vestiges of the signary that was to become the script used to write Oscan north of the SilarusAufidus line, variously labelled the ‘Osco-Etruscan’ or ‘Central Oscan’ alphabet,2 appear in the last decades of the fifth century, and the script was the result of a conscious adaptation of the Etruscan alphabet, with the addition of several new signs required by the phonetic set-up of the Oscan language, including, among other things, signs for voiced stops.3 In what follows, the Oscan alphabet in use north of the Silarus and Aufidus rivers will be analysed as the resulting cohesion of different scripts already present in Campania, whose presence was 2

3

The term ‘Central Oscan’ stems from the view that the Oscan epigraphic material can be distinguished into three groups depending on the script used; a northern, a central and a southern group. From the late third century onwards, Sabellic speech varieties in the northern regions, spoken by, for example, the Paeligni, the Marrucini and the Vestini, came to be written down using somewhat modified versions of the Latin script. Lejeune (1970: 272) defines this as the ‘Osco-Latin’ or ‘North Oscan’ alphabet. Whether or not these speech varieties should actually be considered Oscan or not remains somewhat debated; see Clackson (2015). The process was concluded c. 300, with the addition of two extra vowel signs, and , placed at the end of the abecedarium. For a brief summary of this process, see Crawford et al. (2011: 22–4) and Tikkanen (2011: 11–14); on the spelling of the Oscan vowels in the Greek alphabet, see Zair (2013; 2016).

99

Karin W. Tikkanen

the result of mobility involving those of different political and ethnic origins, from both within and outside Italy.4 Sabellic in Campania It is by now well known that the label ‘Oscan’ is a primarily linguistic term.5 Oscan inscriptions stem from different regions inhabited, according to the ancient sources, by a number of different tribes, who often acted on their own behalf and only rarely in the context of any larger federation-like structures.6 It also seems likely that the ethnic situation in early Italy was more fluid than the ancient sources suggest. Bradley (2000: 126) proposes that only after the Roman conquest was there any overarching administration resulting in more fixed definitions of ethnicity in ancient Umbria, and Clackson (2012: 137) argues that this was probably the case also in Campania, citing in this context the comment by the fourth-century geographer Eudoxus that the Opici – an earlier name for the Oscans – γλώσσας συνέμιξαν ‘mixed languages’ (Clackson 2012: 134). What is documented in the inscriptions from the central and southern parts of the peninsula may thus be a number of more or less closely related tongues within the southern members of the Sabellic language group. It is in this light that the inscriptions sorted as ‘Pre-Samnite’ in Rix (2002) must be considered. These are written using a variety of scripts, in versions of either the Greek or Etruscan alphabets, and are dated to the late sixth or early fifth century, with only three 4

5

6

This analysis will not consider standardisation of the Oscan language, but only the standardisation of the alphabet. For a discussion of the regional variation within Oscan, see Rix (1996). For a methodological evaluation of standardisation processes as expressions of intolerance of optional variability in language, and motivated by various social, political, and perhaps also commercial, needs, see Milroy & Milroy (1999: 22, 26), and also the discussion in Wagner, Outhwaite & Beinhoff (2013). On the standardisation of Latin, see Clackson & Horrocks (2007: 78–80). See e.g. Clackson (2012: 135). On the history and use of the label ‘Oscan’, see Dubuisson (1983). Among these, the Samnites were those to make the greatest imprint on Roman history, given the long series of conflicts regularly labelled ‘the Samnite Wars’. It has been a common habit to describe the Samnites as an original ethno-cultural group; see e.g. La Regina (1980: 29) and Musti (1984); for a critical evaluation of this type of description, see Dench (1995: 187–93) and Scopacasa (2015: 18–55).

100

The Creation of the Oscan Alphabet

potentially reaching as far down as 400; Capua 35, Minturnae 1, and Nola 11.7 As pointed out by Clackson (2012: 136–7), these inscriptions are an incoherent group of texts. Several of them feature only name formulae, of different lengths, others are uninterpretable due to their brevity, and the content of the only longer text, the cippus of Tortora (Blanda 1), remains obscure; see Lazzarini & Poccetti (2001). There is also some morphological fluctuation among these texts, in that the genitive singular is sometimes written -eis, sometimes -es,8 and variation in the form of the first-person singular of the verb ‘to be’ between esum (alphabet of Nocera), sum (Etruscan alphabet), and sim (in fourthcentury inscriptions from Saticula).9 The label ‘Pre-Samnite’ implies a division between the language(s) attested in these early inscriptions and languages of the Samnite tribes who migrated southwards onto the Campanian plains in the fifth and fourth centuries (Strabo 5.4.12; D.H. 2.49.4; Plut. Numa 1.3; Just. 20.1.4).10 It also reinforces the perception that the aboriginal inhabitants in Campania, before the Samnite migration, spoke 7

8

9

10

The majority of these twenty inscriptions will be considered in this chapter, with the exclusion of the following: Capua 35 vinuχs veneliis peracis estam tetet venilei viniciiu ‘Venox, son of Venel, Peracis (Percius) gave this to Venil (Venel), son of Venox’, which probably displays the language of an Italic immigrant who had Etruscan connections (Crawford et al. 2011: 447–9); Casinum (?) 1 minaθeis ‘Of Minatus’, on a black-slip bowl, c. 485–450, which comes from the area of the Volsci (Crawford et al. 2011: 929); and Ps 18, which is probably Etruscan (Crawford et al. 2011: 1588; cf. Vetter 1953: 95, No. 121). Nerulum 1, Solacium 1 and Blanda 1 will also not be discussed in detail in this chapter, since they belong to the southern corpus; see McDonald (2015). Given more recent discoveries and analyses, a few additional Sabellic inscriptions from Campania from the same period, not included as ‘PreSamnite’ in the ST, will be included in this analysis. In this chapter, Sabellic inscriptions are cited, with some minor adjustments, as in the editio maior by Crawford et al. (2011). The forms in -es have been the target of a number of rather different explanations. Cristofani (1994: 384) and Rix (1996: 246) explain the latter forms as by South Picene influence, while Clackson (in Crawford et al. 2011: 24) suggests that the form is a loan from Etruscan, and Prosdocimi (2002) takes them to reflect the PIE consonant stem genitive ending. Lazzeroni (1985) sees a reflection of a sociolinguistic feature or a lower register, also visible in a few later Oscan inscriptions, whereas Russo (2005: 52–5) argues that these forms are nominative possessive markers, which would mean a variety of the formula of possession seen in a few early Faliscan so-called speaking inscriptions, or ‘iscrizioni parlanti’. On these forms, in relation to súm, the regular Oscan spelling after the reform of c. 300, see Joseph and Wallace (1987). On the function of the narrative of the uer sacrum in these migrations, see Tikkanen (2017).

101

Karin W. Tikkanen

a different Sabellic variety, more similar to the northern group, than the language(s) that later on came to be called ‘Oscan’.11 There is, however, no evidence of ‘Pre-Samnite’ having ever existed as a coherent language used by any coherent body of speakers. For the purpose of this chapter, these inscriptions will be considered as attempts at documenting early Sabellic speech varieties in Campania and will be referred to as ‘Campanian Sabellic’, referring to a linguistic group in the geographical context of the Bay of Naples.12 Writing in Campania There is little agreement as to exactly where the Phoenicians first taught the Greeks to write, other than that this must have been ‘somewhere where Euboeans and Phoenicians were seeing a lot of each other’ (Ridgway 2004: 41). Such contact was an evident ingredient in the trading environment in the coastal areas of Campania from a very early period onwards. There is evidence of lively contacts and familiarity between the indigenous populations and the Phoenicians and the Greeks, and before them the Mycenean traders that sailed there (Ridgway 1992: 29). In the early eighth century, when Euboean colonists settled on Pithekoussai,13 Campania was already an area where different ethnicities and languages mixed. The Etruscans had been practising what appears to have been a directed colonisation program, very likely a feature of a longer continuum of Etruscan presence in the south (Bonghi Jovino 2000: 157). The Etruscan settlements at Capua, by the Volturno,14 and Pontecagnano, by the Picentino river,15 stand out 11

12

13

14

15

E.g. Cristofani (1994), Meiser (1987) and Rix (2005); for a critical evaluation of this theory, see Clackson (2015). This label is preferred to the alternative ‘Paleo-Sabellic’, e.g. Poccetti (2017: 301), since the latter is often used in reference to South Picene, e.g. La Regina (2011). According to tradition, the settlement was founded as a joint project by Eretrians and Chalcidians (Strabo 5.5.9). This is supported by, for example, early inscriptions featuring traces of both Eretrian and Chalcidian writing habits (Janko 2015: 2–3). The Etruscan name of Capua was Volturnum (attested in the Tabula Capuana, see Cristofani, 1995). The ancient sources do not agree on the date of foundation, but all say that it was originally an Etruscan city (Salmon 1967: 38–9). The Etruscan name of the settlement has been lost, but the Romans called it Picentia (Plin. Nat. 3.70). The colony flourished in the mid-eighth century, and there is plenty of evidence of early trade with both the Greeks and the Phoenicians (Cuozzo 2012: 194).

102

The Creation of the Oscan Alphabet

as the most influential and possibly also the oldest, dating to the tenth and ninth centuries respectively (Cuozzo 2012: 193, 205). With their higher socio-economic strength, the Etruscans achieved cultural hegemony over the area, and there was very likely an initial stage of acculturation of the various indigenous tribes, antedating the arrival of the Greeks (D’Agostino 2001: 246). The establishment of the Euboean colony on Pithekoussai thus functioned as the end point to a long period of initial trading contacts, and Cumae, founded on the mainland in c. 725 (Livy 8.22; D.H. 7.3),16 also fitted into an existing network of integrated settlements where Etruscans and speakers of Campanian Sabellic already coexisted (Cuozzo 2012: 192). From this stage onwards, a sense of shared, hybrid culture permeates the region, which came to (re-)inforce the connections between the Etruscan, Greek and indigenous aristocracies, with several shared features on both private and public levels. Onomastics shows that widespread ethnic mixing was indeed going on,17 and architectural innovations in the shape of, for example, façade decorations originating from Pontecagnano and Capua appear far up north in Latium and Etruria; in Rome, Pyrgi and Volterra; and also down south, in Daunia and Sicily (Cuozzo 2012: 211). In this climate, where different ethnic groups coexisted, the habit of writing became one of many other aspects of Greek culture adopted by the Apennine tribes (see further Malkin 2002). The oldest Etruscan inscriptions come from early seventhcentury aristocratic graves in Tarquinia (ET Ta 3.1)18 and Caere (ET Cr 2.1, 2.4).19 The script used is essentially the Euboean alphabet, with very little modification (Cristofani 1979: 378–9), 16

17

18 19

According to the dating of the earliest colonial finds at Cumae, contemporary with Late Geometric II graves at Valle San Morano (Ridgway 1992: 118). Greek–Etruscan unions are proven by names of the type Rutile Hipukrates and Larth Telikles (on seventh-century vessels), and one may also call to mind the tradition concerning the Corinthian aristocrat Demaratus, who is said to have fled to Etruscan Tarquinii where he married a local elite woman and fathered Tarquinius Priscus (D.H. 3.46; Strabo 5.2.2; Cic. Rep. 2.19; Livy 1.34; cf. Hall 2004: 40–1). For a discussion on the function of Italic fibulae as ‘evidence’ of intermarriage, see Hodos (1999), Shepherd (1999), Coldstream (2004: 53) and Ridgway (2004: 18). See Bagnaso Gianni (2010: 115–16). See Staccioli (1968: 249–50).

103

Karin W. Tikkanen

but during the centuries to follow the Etruscan scribes successively calibrated the script to fit their own language (Pandolfini & Prosdocimi 1990: 11–15).20 Etruscan writing is attested in the southern colonies only about a century later on, first of all in Pontecagnano.21 This can be tied to important developments in the local economy. Beginning in the early sixth century, Pontecagnano became the centre of a number of innovations in terms of the manifestation of social structures, seen, for example, in the allotment of sections of the necropolis to the aristocratic families, a change that also included the monumentalisation of tombs. Capua, too, saw a period of extraordinary growth and development towards the end of the sixth century, with an expansion in the erection of public buildings, but also a surge in crafts and manufacture (Cuozzo 2012: 211–12). From Capua urbanisation was extended to include the nearby settlements of Suessula and Calatia, with successive inclusion of other locations, such as Vico Equense, Stabiae and Nola, followed by the Etruscan refounding of old Italic settlements such as Pompeii (Cuozzo 2012: 210, 215). The new wealth among the elite in Pontecagnano thus led to new funerary customs, and from the sixth century onwards there is a large number of inscriptions on objects from funerary settings in Campania, featuring Etruscan and Greek name formulae. This is also the period in which inscriptions featuring Sabellic names, and also Sabellic morphology, first appear. And although the Etruscan hegemony had been long established in Campania, it is clear that the Greek newcomers made a heavy impression on the aboriginal Sabellic tribes, since the first Sabellic scribes in Campania were influenced by Greek settlers rather than the Etruscans.

20

21

On the use of the word ‘scribe’ in this chapter, see the discussion by Wagner, Outhwaite and Beinhoff (2013). A ‘scribe’ is not necessarily a professional scribe (on such individuals among the Etruscans, see Colonna, 1976), but an actively literate individual who writes, and not just for a living. ET Cm 2.1 kane, from the early sixth century, appears to be an early precursor (Pescatori 1968). The simplified Etruscan alphabet, in use from the early sixth century onwards, had only the sign C for the voiceless velar (Pandolfini & Prosdocimi 1990: 13). The oldest Etruscan abecedaria in Campania feature the abbreviated Etruscan signary, after the removal of the ‘dead’ letters, e.g. ET Cm 6.1, from Pontecagnano (late sixth century), and ET Cm 6.2, from Nola (late sixth/early fifth century).

104

The Creation of the Oscan Alphabet

The Alphabet of Nocera In c. 535 the Phocaians surviving the battle of Alalia (Hdt. 1.165–7) founded Velia, and introduced the Ionic alphabet used by the Dodecapolis on the Apennine peninsula (Jeffery 1990: 325–7). This alphabet became the basis for the first attempts at writing a Sabellic tongue, in a script labelled the alphabet of Nocera (see e.g. Crawford et al. 2011: 16; Russo 2005). This is attested in three inscriptions from the area around the Sorrentine peninsula, from Vico Equense (Surrentum 2), Sorrento (Surrentum 3) and Nocera Superiore (Nuceria Alfaterna 3): Surrentum 2 iegies || esum: p[1]les: adaries: ‘I belong to Iegius, son of P[o]lus (?) Atrius’22 Surrentum 3 urugieis || pagieis || ‘Of Orcius, son of Pacius’23 Nuceria Alfaterna 3 bruties || esum ‘I belong to Bruttius (or Brattius)’24

All three texts are name formulae inscribed on bucchero vessels, dated to the end of the sixth century.25 The general impression one gets from the writing is the hand of an unskilled scribe. The letter signs are tilted in different directions, and there are two parallel systems of word division markers in Surrentum 2, the double lines and the double dots , suggesting some kind of convergence between different writing systems.26 22

23

24 25

26

Translation by Crawford et al. (2011: 851–2). The fourth sign, the hooked wau, has also been taken to represent /v/, e.g. CIE II 2.2, p. 15 no. 9 ievies esum p[a]ces adaries, as well as /f/, thus Russo (2005: 38–9). Translation by Crawford et al. (2011: 853). On the interpretation of the fourth and eleventh signs as /g/, see the preceding note. Crawford’s rendering of the names as Orcius and Pacius represents the presumed Latin equivalent, attested in Roman inscriptions with a varying spelling with both voiceless and voiced ; see Franchi de Bellis (2014: 118). On Sabellian adaries (Surrentum 2) for the equivalent of Lat. Atrius, with a similar oscillation between voiced and voiceless stop, see Salomies (2012: 141). Translation by Crawford et al. (2011: 909–10). Russo (2005) dates Surrentum 2 to the second quarter of the sixth century, but Crawford et al. (2011: 16) argue that the date of the introduction of the Ionic alphabet at Velia, c. 535, must be the terminus post quem. On the modifications made to the Ionic script to create the new alphabet, see foremost Russo (2005).

105

Karin W. Tikkanen

The Etruscan Alphabet Sabellic inscriptions written in the Etruscan alphabet come from a range of settlements along the Campanian coast; Stabiae, Nola and Pontecagnano were Etruscan settlements, and Nuceria Alfaterna, near Velia in the south, also had Greek connections. Atella, just south of the Silarus-Aufidus line, represents a singular instance of an early Sabellic inscription in a context with little or no documented Etruscan or Greek presence (Crawford et al. 2011: 18, 29). Campania (?) 1 mutniium ‘Of the Muttinii’27 Atella 2 ciperum ties sum ‘I am the good (?) (vase) of Titus’28 Stabiae 1 ahtica(s) sum ‘I belong to Ahtica’29 Stabiae (?) 2 pacieis paciieis ‘Of Pacius Pacius’30 Nola 6 luv∙cies∙ cnai∙viies∙ sum∙ ‘I belong to Lucius Gnaeuius’31 Nola 7 venelieis vuliieis ‘Of Venelius Vollius’32 Nola 8 mamer∙ces∙huσinies ‘Of Mamercus Hosinius’33 Nola ? 9 veneliei∙s∙ planiliei∙s∙ ‘Of Venelius Planilius’34 Nola 11 cnaives flaviies ‘Of Gnaeus Flauius’35 Nola 12 tec∙ liiam ‘?’36 27 28

29 30 31

32 33

34 35 36

Attic black-slip kylix, c. 500–450 (Crawford et al. 2011: 375). Kylix, c. 500–450 (Crawford et al. 2011: 488). On the formulaic addition of an adjective ‘good’, see Agostiniani (1981) and de Simone (1996). Black-slip kylix, c. 500–450 (Crawford et al. 2011: 841–2). Attic black-slip kylix, c. 500–475 (Crawford et al. 2011: 843–4). Attic black-slip kylix, c. 450 (Crawford et al. 2011: 870). Crawford renders the name ‘Cnaeuius’, but the sign C must here be understood to represent a voiced velar /g/, cf. Nola 11. Attic black-slip kylix, c. 450 (Crawford et al. 2011: 872–3). Black-slip kylix, c. 450 (Crawford et al. 2011: 874–5). Note that the first sibilant in the gentilicium is a san not a sigma as in the transcription by Crawford. Attic black-slip kylix, c. 450 (Crawford et al. 2011: 876–7). Attic black-slip kylix, c. 450–400 (Crawford et al. 2011: 880–1). Two black-slip kylices with uninterpretable inscriptions, c. 500–450 (Crawford et al. 2011: 882–3).

106

The Creation of the Oscan Alphabet Nuceria Alfaterna 6 -tium maiis papiis ‘Maius Papius of the ???’37 Nuceria Alfaterna 7 amer ‘amer’38 Picentia 1 Text A: m Text B: minutihes ‘Of Minutius (?)’39

Just as in the first Etruscan inscriptions, where the Euboean alphabet was used without modification, the writing in these early Sabellic inscriptions constitutes a direct reuse of the Etruscan alphabet. Similar to Surrentum 2 and Surrentum 3, several of these texts also express a level of approximation to the local aristocracy, in that the name formulae are sometimes completed with gentilicia (Cuozzo 2012: 211). Apart from this, they are all proprietary inscriptions on small, moveable objects, with content no more complex than a formulaic speaking inscription, an early epigraphical pattern originating in Greek, and attested in a number of languages on the Apennine peninsula (cf. Agostiniani 1982). As summarised by Colonna (1990: 308), there are no obvious epigraphic differences between these inscriptions and contemporary objects carrying Etruscan, or Greek, onomastics. The Achaean Alphabet A third early branch of Sabellic inscriptions in Campania consists of three brief inscriptions from Salernum, from the first half of the fifth century. These are all written in the Achaean alphabet, with the tell-tale signs for /i/, and (san) for /s/, which were adopted by the local groups without alteration (Rix 2005: 323): Salernum 1 σπυ(ρισ) ‘Spurius’40 Salernum 2 τρεβισ 37 38

39 40

Coarse ware bowl, c. 500 (Crawford et al. 2011: 913–14). Impasto (?) bowl, c. 475–450 (Crawford et al. 2011: 915). All of the Campanian Sabellic inscriptions written in the Etruscan alphabet use C for the voiceless velar, and it is difficult to understand why Crawford supplements a letter in this name, amer(eks), cf. Nola 8. Attic black-slip kylix, c. 500–475 (Crawford et al. 2011: 925). Attic black-slip kylix, c. 500–450 (Crawford et al. 2011: 918).

107

Karin W. Tikkanen ‘Trebius’41 Salernum 3 ποκ(λομ) ‘cup’42

The Achaean alphabet was introduced into the area through the foundation of Poseidonia in 600.43 The fact that a Greek script was adopted to write Sabellic names could suggest that some scribes were taught directly by Greek speakers, possibly at Poseidonia (Colonna 1990: 307), or alternatively the choice of script might have functioned as a way to communicate with the Greek surroundings (Cristofani 1998: 276–7). The Euboean Alphabet The Euboean alphabet, lastly, is found in two Sabellic inscriptions from before c. 400. Just like the writing in the Achaean alphabet, these inscriptions appear to be one-off instances rather than elements of any coherent writing habit. The small, impasto Garigliano Bowl, from the sanctuary of Dea Marica at Minturno, at the mouth of the Liris, features a long inscription in archaic Latin (Vine 1998; Hartmann 2005: 149–53; Maras 2005; Morandi 2009), and also a onepart name formula, evidently an Italic name written in Greek letters: Minturnae 1 ΑΗΥΙΔΙΕΣ ‘Of Afidius’44

The second inscription comes from Pontecagnano and is dated somewhat later than Picentia 1 (see n. 334): 41 42 43

44

Attic black-slip kylix, c. 500–450 (Crawford et al. 2011: 919–20). Black-slip kylix, c. 500–450 (Crawford et al. 2011: 921–2). At the end of the eighth century Achaean settlers from Helice founded Sybaris in the Gulf of Tarento (Strabo 6.1.13; Arist. Pol. 1303a.; cf. Callaway 1950: 1). Around 600 Sybarite settlers themselves founded Poseidonia, near the mouth of the Silarus (Jeffery 1990: 250–1). Crawford et al. (2011: 359) gives only a Latin transcription, Ahuidies. The brevity of the inscription makes the identification of the script difficult, and these may also be Ionic letters; see also Maras (2005: 34). The dating, c. 400, is based on the lack of syncope in the Italic name (Crawford et al. 2011: 350); the Latin inscription was engraved by a different scribe, possibly as early as the beginning of the sixth century (Hartmann 2005: 265–6; 433).

108

The Creation of the Oscan Alphabet Picentia 2 σπυρ(ισ) ‘Spurius’45

The Creation of the Oscan Alphabet The first attempts at documenting the different varieties of Campanian Sabellic thus made use of the various scripts available in the coastal areas of Campania in the sixth and fifth centuries. Russo (2005: 109, 111) argues that the Oscan alphabet represents a merger of different pre-existing branches of writing: the alphabet of Nocera at Surrentum, the Achaean writing at Salernum and the Etruscan alphabet in the settlements along the coast.46 In fact, there are no traces of the alphabet of Nocera in the later Oscan alphabet, but what Russo seems to say is that there was a continuous knowledge of writing that eventually matured into a new script. This may well have been the case, given that both the Euboean and Etruscan alphabets represent continuous writing habits in the area of the Bay of Naples from the eighth and sixth centuries, respectively, onwards. Certain features demonstrate that these two alphabets were the source of the Oscan alphabet: the Etruscan symbol 8 to write /f/; the continuation of only one sign for both back vowels; and the signs for voiced stops /b, d, g/. The signs for voiced stops, along with the sign for /o/, were excluded from the simplified Etruscan alphabet in the sixth century, and the symbol 8 created somewhat later. There are no attested Etruscan abecedaria featuring all these signs (Pandolfini & Prosdocimi 1990: 49, 51), and the Oscan alphabet is thus not simply a replica of an existing Etruscan model, but rather a reanalysis with additions from the Greek script, in adaption to the Oscan tongue (Cristofani 1998: 276).47 The question is when and where this compilation of the Etruscan and Euboean scripts was achieved. 45 46

47

Black-slip one-handed bowl, c. 425–400 (Crawford et al. 2011: 926). Russo does not include the few traces of Campanian Sabellic written in the Euboean alphabet, probably because of the sporadic attestation. On the suggestion by Rix (2005), that the Oscan alphabet was derived from the South Picene script, and arguments for why this cannot have been the case, see Crawford et al. (2011: 19).

109

Karin W. Tikkanen

Campania in the Fifth and Fourth Centuries After almost two centuries of more or less peaceful co-existence, the end of the fifth century saw drastic changes to the political climate in Campania. The story of the takeover of Capua in 423 (Livy 4.37.1, cf. 28.28.6; Strabo 5.4.3), followed by the conquest of Cumae two years later (Livy 4.44.12), is sometimes explained as a direct consequence of the Samnite migrations, and the arrival of a new group that outmanoeuvred the old Etruscan oligarchies as well as the Italic tribes. Given that the earliest traces of the Oscan alphabet postdate these events, it is an easy solution to attribute the development of the new alphabet to the Samnites alone (see, as proponents of this view, e.g. Tagliamonte 2004: 134; Hayes 2013: 199). The end of the oligarchy of the old aristocracies in Campania, however, can equally well be the result of social tensions and ethnic conflict with original populations.48 The changes can also be tied to the changing climate in the relations between the Etruscans and the Greeks, since the expansion of settlements like Poseidonia towards the end of the sixth century came to threaten the Etruscan hegemony (Cuozzo 2012: 213). The rule of the tyrant Aristodemos in Cumae, from 504 onwards, could be seen as a consequence of this situation, as individual aristocrats within the Etruscan cities took advantage of the changes. Aristodemos’s rule coincided with the growth of Cumae, Capua and the whole of inner Campania. On the one hand, this led to the separation of the city from the northern Etruscan federation, a corollary of the victory over Lars Porsenna, the king of Chiusi, in 508 (D.H. 7.5–6). It also made pronounced the tensions between the different ethnic groups within the Etruscan cities. In 490 Aristodemos was killed, which definitely marked the beginning of the end of the Etruscan rule in Campania (D.H. 7.7–11; cf. Cuozzo 2012: 217). The flourishing of Capua in the late Archaic period meant that artisanship such as bronze-work, terracottas and pottery decorations reached higher levels than before. As a consequence, many 48

Tales of Sabellic takeovers are told about a number of different cities, and Salmon (1967: 38–9) finds that the ultimate origin of the tale may have been Timaeus’ account of Mamertine soldiers taking over Messana (Festus p. 150; Lindsay 1913).

110

The Creation of the Oscan Alphabet

workers, most of them probably Italic in origin, moved into Capua, as well as Cumae, from the countryside. However, it seems that in both cities the ruling classes continued to exclude the artisans from the city government, as can be deduced from the location of the artisans’ quarter on the outside of the city walls. In the late 420s this led to growing social tension with successive changes in the government bodies (D’Agostino 2001: 250–1). Further south, the change was somewhat slower, and the old aristocracies remained in place until the end of the fourth century (Cuozzo 2012: 220). As Cuozzo (2012: 219) remarks, the Campanian Sabellic inscriptions from the Nola and Sarno regions in the fifth century are evidence of the ongoing organisation of previously peripheral populations. The appearance of the Samnites was thus not necessarily the supreme element in the later upheavals, although the advent of another Italic group on the scene, also excluded from the political agenda within the Etruscan cities, may have served as an additional igniting factor in an already tense climate. In the end, the result of the social changes of the fifth century can be summarised as the emergence of two different local groups. The Campanians came into view as a distinct group in 438 (D.S. 12.31.1), and controlled the Campanian plain and the Bay of Naples. The Samnites, on the other hand, inhabited the area of the Hirpini and Caudini, and the Sorrentine peninsula (D’Isanto 1993: 15; Cuozzo 2012: 220–1).49 These new forces also expressed their political formation through the minting of coins, and it is in these coins that Crawford et al. (2011: 19) detect the first large-scale attempt at writing an Italic language in Campania. Minting and the Oscan Script Crawford et al. (2011: 19–22) reason that the labour to produce the Oscan alphabet took place at, or in connection with, the mint in Naples, conducted by (possibly) Greek-speaking personnel, who consciously combined letter signs from the two alphabets in use, 49

The two groups were not always on good terms. For example, in 343, the Capuans asked Rome for protection against a threatening invasion by the Samnites, which eventually led to the outbreak of the ‘First Samnite War’ (Livy 7.30–2; cf. Scopacasa 2015: 29).

111

Karin W. Tikkanen

the Euboean and the Etruscan scripts.50 In the first stages of this minting the Euboean alphabet was used unaltered, and there is no indication of an Oscan-speaking community: Capua 1 Coinage ΚΑΠΠΑΝΟ, ΚΑΜΠΑΝΟ ‘Campanian (silver/nummus)’51 Cumae 1 Coinage ΚΥΜΑΙΟΝ, ΚΥΜΑΙΝ, ΚΥΜΑΙΟ ‘Of the Cumaei’52

None of these coin legends needed letters not already present in the Euboean script. According to Crawford et al., however, some amount of consideration was required with the coinages for Phistelia, Allifae and that of the Fenserni, since they all required the spelling of the sound /f/: Phistelia 1 Coinage fistluis (abl.pl.), fistlus (nom.pl.) ‘From Fistelia, from fistlus’, didrachms, c. 405–400.53 Allifae 1 Coinage (1)-(5) AΛIOHA, ΑΛΛΙΒANON, AΛΛEI ‘Al(l)ifa, Allifan (silver) (?)’54 Fenserni 1 Coinage fensernum ‘Of the Fenserni’55

The writing of the sound /f/ had caused a problem for the Etruscans, since the Euboean script they adopted had no specific sign for this. After experimenting with the digraph , a combination of the Greek digamma and eta, in the mid-fifth century the Etruscan scribes created the symbol 8 (Pandolfini 50

51

52

53

54

55

Minting commenced in Italy in the mid-sixth century, in Sybaris, and spread successively to the west coast (Rutter et al. 2001: 3–4). There was minting at Cumae (from c. 475) and Naples (c. 450), and the mint remained in use after the takeover of Cumae in 421. It was during this period that Naples took on the role of supplying several of the communities in Campania with coinages, as is evident from the reuse of certain diemarks (Rutter 1979: 75, 82–3). Silver didrachms, c. 415–405; see Crawford et al. (2011: 384); HNItaly (2001) no. 476–8, with plate 7. Silver coinage, c. 420–300; see Crawford et al. (2011: 490); HNItaly (2001) no. 531–5, 536, 537. . Obols, hemi-obols and quarter-obols appear from a later date, inscribed with Greek letters, ΦΙΣΤΕΛΙΑ (nom.sg.), ΦΙΣΤΕΛΑΔ (abl.sg.) ‘(from) Phistelia’; see Crawford et al. (2011: 586–91); HNItaly (2001) no. 611–12 , plate 11. Various spellings. Silver didrachms, c. 400–395; followed by Allifae 1 Coinage (6) alifa ‘Al(l)ifa’, hemi-obol, c. 325–275; see Crawford et al. (2011: 581–3); HNItaly (2001) no. 459–64, plate 7. Silver didrachms, c. 395–390; see Crawford et al. (2011: 489); HNItaly (2001) no. 538, plate 9.

112

The Creation of the Oscan Alphabet

& Prosdocimi 1990: 13–14).56 As an element in the modified Etruscan alphabet, the sign 8 is attested in some of the early Campanian Sabellic inscriptions that use the Etruscan alphabet, such as Nola 11, and the sign also spread to several other languages on the peninsula (see McDonald 2015: 65–6). In other words, in the writing of the Sabellic language(s) in Campania the symbol 8 for /f/ was a letter that had been in continuous use ever since the Etruscan script was first adopted. The question is how to value these early coin legends, with different solutions to the writing of the Campanian place names. Some coins employ Greek letters with the occasional inclusion of 8,57 others the Etruscan alphabet, although Crawford classifies these latter as written in the Oscan alphabet and sees a ‘perfect Oscan legend’ in Fenserni 1 Coinage (Crawford et al. 2011: 21).58 To be specific, however, all of these coins could equally well be considered to be written in the Etruscan script (and are listed in this fashion in ST). It is also somewhat of a problem that these coins are, in some cases, the only epigraphy originating from a given city; this is the case with Phistelia, Fenserni, Aquilonia and Meles, and possibly also Allifae (unless ? Allifae 2 ? is authentic, see Crawford et al. 2011: 584–5). There is thus very little evidence of a true Oscan writing tradition in these communities, since there is nothing in the writing on the coins themselves that guarantees that this is indeed the Oscan alphabet; they could in theory equally well be inscribed by Greek-speaking or Etruscan-speaking personnel, with no knowledge at all of the Oscan language. If there had 56

57

58

The digraph also was adopted into the Latin alphabet, and is found in early Italic inscriptions from Latium and the northern Campanian area, e.g. Anagnia 13 huidas, from 600–500 (Rix 1998: 261; Crawford et al. 2011: 355–6), and, with in place of digamma, Minturnae 1 ΑΗΥΙΔΙΕΣ, c. 400; see Maras (2005) and Crawford et al. (2011: 359). This is how Crawford et al. (2011: 581–2) interprets the sign B in Allifae 1 Coinage. A stronger argument in favour of an Oscan alphabet by c. 400 might be the ‘recovered rectangularity’ of the Classical Greek letter forms (Crawford et al. 2011: 19). This, however, may equally well be the result of the reuse of die-marks, rather than the conscious creation of new letter forms.

113

Karin W. Tikkanen

been samples of early coinages from Caudium, Abellinum, Bovianum, Aufidena or Trebula, it could be conclusively proven whether these were true Oscan coin legends or not, given the requirement to write out signs for voiced stops.59 The issue thus lies in the introduction of signs for voiced stops in the Oscan alphabet, and the circumstances within which this might have been achieved. The Greek Addition: Signs for Voiced Stops The lack of the set of signs for voiced stops, /b, d, g/, became problematic for all of the Italic speech communities who derived their alphabets from the Etruscan script. The Oscan alphabet represents one solution to this problem.60 Signs for voiced stops are included in all Greek alphabets, and are also present in the three inscriptions written in the alphabet of Nocera, based on the Ionic alphabet; there is a beta in Nuceria Alfaterna 3, and a delta appears in Surrentum 2. There is also a sign for /g/, if Crawford et al. (2011: 16, 851–3) are correct in their interpretation of Surrentum 2 and Surrentum 3. The oldest attested evidence of an inscription written in the Oscan alphabet, that is, the revised Etruscan signary with added letters, is currently the inscription of Niumsis Tanunis, a small skyphoid olla, probably originating from Capua and dated to between the end of the fifth century and the beginning of the fourth (Agostiniani & Fachetti 2013: 123, 126): niumsies tanunies est | paplamtensatriiam fufuhud | niumsis tanunis eises ulsu dunum dedum61 ‘This (vase) belongs to Niumsis Tanunis . . . Niumsis Tanunis . . . gave as a gift (?)’

59

60 61

This might, however, also not suffice, since in some instances Oscan coin legends feature an earlier, Etruscan-type spelling, well into the period when the Oscan alphabet had taken proper shape, e.g. Cubulteria 1 Coinage kupelternum ‘Of the Kubulterni’, c. 265–240, where the voiced labial is written using an Oscan

or Greek (Crawford et al. 2011: 580). For the Latin solution, see e.g. Gordon (1962) and Bundgård (1965). Edited as in Agostiniani & Fachetti (2013: 132).

114

The Creation of the Oscan Alphabet

Interpretations of the text differ somewhat, and the words in line two remain obscure, as also the last word, dedum.62 The language is on the whole very similar to the earliest Campanian Sabellic inscriptions, with, for example, the ending -es in the genitive singular (see note 8), but the presence of as the sign representing/d/, in the phrase dunum dedum,63 confirms that the alphabet used in this inscription must be the Oscan signary. The letter delta, the fourth letter in the Greek script, became obsolete in the Etruscan script just like the other signs for voiced stops. In the Etruscan alphabet the sign wrote /r/,64 and it was acquired with this function by the Sabellian scribes in Campania, as seen in Atella 2, Nola 8 and Nuceria Alfaterna 7 (see the section in this chapter entitled ‘The Etruscan Alphabet’). The Greek delta could therefore not be reintroduced into the Oscan alphabet as part of the triad of signs for voiced stops, and the solution became a recharacterisation of delta as .65 After considering the range of dates compatible with the archaeological context of the olla, Adiego Lajara (2015: 27) prefers, for phonological reasons, the early fourth century; if this is correct, the inscription marks the acculturation process whereby the speakers of the earlier Campanian Sabellic speech

62

63

64

65

Agostiniani and Fachetti (2013: 141) take dedum to be a 1. sg. perf., with the subject Niumsis Tanunis, thus ‘I, Niumsis Tanunis, gave (this) as a gift’; so also Adiego Lajara (2015). Triantafillis (2014: 21–2) instead sees a perf. part., ‘given as a gift’. The phrase reflects the Latin formula donum dare ‘give as a gift’, and the formula formed part of a shared Italic vocabulary to do with votive inscriptions (Euler 1982). Early Etruscan abecedaria attest the Euboean form , the stem of which was successively shortened. From the mid-sixth century, the shape of the sign is established as , as in ET Ru 9.1 (Pandolfini & Prosdocimi 1990: 51). According to Agostiniani and Fachetti (2013: 129–30), this was achieved through the addition of a diacritical mark to the Greek letter rho. Adiego Lajara (2015: 17) instead sees a different replacement, since in the Etruscan script both and could be used to write /r/, whereafter the Oscan scribes, having already established for /r/, came to adopt for /d/. The choice of this letter was not unproblematic, and there are some instances of evident scribal confusion. This is the case with the undated Oscan abecedarium, Pompei 81, which runs a, b, c, r, e, v, z, h, i . . ., with the sign (/r/) instead of (/d/) in the fourth position (Antonini 1990: 146–7; Crawford et al. 2011: 744). There are also coins from Larinum, from the middle of the third century, written in the Latin alphabet, with the form ladinod ‘from Larinum’ (Larinum 1 Coinage (4) and (5)), where is used to write both /r/ and /d/ (Crawford et al. 2011: 1281; HNItaly 2001, no. 624, plate 43).

115

Karin W. Tikkanen

variety came to adopt both the Oscan alphabet as well as Oscan names, Niumsis.66 It is impossible to say for certain whether the other two signs for voiced stops, which were incorporated in their Greek forms, beta and gamma, were also part of the Oscan signary at the time of the writing of the Niumsis Tanunis inscription or not.67 Inscriptions attesting the complete adaption of the Oscan alphabet, with signs for all voiced stops, come from the late fourth century, for example: Capua 48 keis∙ d{∙}iufaris ‘Keis Diufaris’68 Caudium (?) 3 pupdiis stenis ‘Stenius Popidius (painted this)’69 Capua 36 vibi[s:] smintiis: vibis: smintiis: sum ‘Vibius Smintius. I am Vibius Smintius’70 Caudium 2 gnaivs ‘Gnaeus (painted this)’71

The addition of the Greek sign B to write the voiced labial simply involved the addition of the second letter in the Greek abecedaria to the Oscan script, but the reattribution of the value of /g/ to the letter C required the use of the sign K for the unvoiced stop (cf. note 38). Greek-speaking Campanians would have been aware of this function of the Greek letter, and there was also some amount of spelling variation within the Etruscan linguistic continuum within Campania; among late sixth-/early fifth-century inscriptions from Fratte di Salerno, there is an evident parallel use of both K and C for /k/ (e.g. ET Cm 2.18 stlakies, vs. ET Cm 2.73 cuvsinu). Fratte was among the least institutionalised of the Etruscan cities and was probably more of an emporion, with 66

67

68 69

70

71

The name displays the Oscan-particular sound change palatalization -iu- < *-y- < *-uafter a dental obstruent; this was previously dated to c. 300, but ought to be dated one century earlier (Adiego Lajara 2015: 19, 25). Agostiniani and Fachetti (2013: 131) argue that this was the case, and that the fact that the two letters are not attested in the inscription is by mere chance; for a critique of this statement, see Adiego Lajara (2015: 16). A slave’s name (?). Black-slip kylix, dated to c. 350–300 (Crawford et al. 2011: 470). Painter’s signature, on an Italian red-figure bell-krater, from c. 320–300 (Crawford et al. 2011: 943). A decorated chamber tomb dated to c. 330 based on the surrounding archaeological finds (Crawford et al. 2011: 450–1). Painter’s mark on a Campanian red-figure bell-krater, c. 325–300 (Crawford et al. 2011: 942).

116

The Creation of the Oscan Alphabet

members from different locations (Cristofani 1992: 65), which may have influenced, and also welcomed, different writing habits.72 The Maturation of the Oscan Script From a historical point of view, the new political formations among the Sabellic tribes in Campania in the 420s, in part the result of the Samnite migrations, may very well have spurred the creation of a new alphabet to write down their spoken language, as suggested by Crawford et al. (2011). Based on the evidence available at the time, Rix (2002, 2005) stated that the Oscan alphabet was clearly documented only from the late-fourth century onwards, and all earlier inscriptions in non-Greek scripts were written in the Etruscan alphabet. The new evidence presented by the Niumsis Tanusis inscription suggests otherwise, and Agostiniani and Facchettii (2013), arguing for a very early date of this epigraphy, place its writing within a late fifth-century Oscan environment, supporting Crawford’s suggestion regarding the establishment of the Oscan alphabet at this early date. Adiego Lajara (2015) instead prefers a somewhat later dating of the Niumsis Tanunis olla, from the early fourth century. In any case, it seems very likely that the Oscan script either had been, or was in the process of becoming, adapted and matured for the writing of the Oscan language, at some point in the range of the late fifth and early fourth centuries. This is fully in line with the discussion regarding the social tension in late fifth-century Campania, and the advent of the Samnites as a new socially prominent group on the political scene. From that point onwards, however, there is a gap of almost three quarters of a century before the major surge of Oscan inscriptions, firmly proving the completeness of the Oscan alphabet, and also a greater level of literacy within the Oscan-speaking society, since inscriptions are attested in a wider range of contexts. 72

This may also have caused some amount of uncertainty in the spelling. Saticula (?) 5, on a black-slip kylix from c. 350–300, could be interpreted as mageis sim ‘I belong to Magius’. However, given that the Etruscan alphabet remained in use at Saticula until the third century, the writing could also be in Etruscan, in which case the name is to be read as ‘Macius’ (Crawford et al. 2011: 599, 602–4).

117

Karin W. Tikkanen

As Cornell (1991) argues, preserved inscriptions alone are not always representative of the actual level of literacy in early societies. Just as is the case with the majority of preserved archaic Latin inscriptions (Cornell 1991: 10–12), all of the preserved Campanian Sabellic epigraphic texts have been found in the context of cemeteries (the archaeological background of the Niumsis Tanunis inscription is not known; see Agostiniani & Fachetti 2013: 123). The fall of the Etruscan hegemony most likely involved a disruption of old burial customs, including the Etruscan habit of depositing inscribed objects in tombs, meaning that during the period immediately following the end of the fifth century, few inscriptions would have been preserved in this way. In analysing the potential existence of Latin inscriptions that were not preserved in tombs, Cornell (1991: 26) discusses historical references to Roman peace treaties put up for public display, from the late sixth and early fifth centuries. Comparable evidence from Campania from the period under consideration might have included the treaty between the Samnites and Rome in 354 (Livy 7.19.4; D.S. 16.45.8; cf. Salmon 1967: 190–1), as well as the treaty from 343, when the Capuans entered an alliance with Rome in order to defend themselves against the Samnites (Livy 7.29.5–7), and that from 313, when Atina, Calatia and Nola became Roman socii (Livy 9.28.6; cf. Salmon 1967: 197). There is also a literary tradition regarding writing amongst members of the élite, such as the meeting between Archytas the Tarentine, Plato and the Samnite Gaius Herennius Pontius (Cic. Cat. M. 12.39–41),73 and the letters that Herennius’ son, the Samnite commander Gaius Pontius, is said to have sent to his father after the battle at the Caudine Forks, in 321 (Livy 9.3.4; see further Tagliamonte, 2004: 136). The earliest preserved inscriptions written in the Oscan alphabet proper, from the end of the fourth century, come from a range of locations within the Oscan-speaking region – Saepinum, Bovianum, Capua, Calatia, Caudium, Pompeii, Nola, and Cumae – documenting a wide spread of writing. Most of these 73

For a discussion of the content and significance of this meeting, see e.g. Scopacasa (2015: 50, with n. 147).

118

The Creation of the Oscan Alphabet

inscriptions are proprietary inscriptions (notably Capua 48, possibly belonging to a slave), but there are also tomb inscriptions (Capua 36) and painter’s markings (Caudium 2, Caudium 3), and perhaps also one public display inscription.74 There was thus a fairly wide variety of texts in terms of both intention for writing, and also the intended audience, which implies a fairly advanced level of literacy. During the fourth and early third centuries, Campania and Samnium commenced the process of urbanisation, with demographic growth and the onset of both monumental architecture and settlement fortification (Gualtieri 2004: 39, 47; Stek 2009: 38–9). The Oscan cities were advancing in terms of inner economy as well as external relations, and literacy may very well have been an important factor in this development. The fact remains, however, that the same period was also blanketed by a series of conflicts, in itself a consequence of the migration of different tribes within the area – the arrival of the Samnites (however one decides to interpret this), the formation of the Campanian league, and the withdrawal of the Etruscans, in connection with the growth of the Roman sphere of influence.75 The changes noticeable in the Oscan orthography reflect changes in the social situation within the area. Considering this long period of military unrest, it is no wonder that there are not that many Oscan inscriptions from the fourth century, even though there was very likely a continuous writing tradition using the modified Oscan alphabet from the late fifth century onwards. Conclusion The first period in the writing down of the Sabellic speech varieties in Campania, c. 575 down to c. 450, consisted of making use of 74

75

This is if Crawford’s correction of Cerulli’s reading of Bouianum 116 is correct. The inscription itself, on a limestone block, the basis for a statue of Athena, has been destroyed, and the text is preserved in an eighteenth-century manuscript only (Cerulli 1777: 279). Vetter (1953) gives the text as Ve 158 tanas: niumeriis: / frunter; and Rix (2002) adds a dative form, Sa 27 tanas: niumeriis: / frunter[eí] . Crawford corrects the text to [s]taatis niumeriis famated ‘Statius Numerius ordered (this) (to be made)’ (Crawford et al. 2011: 1121–2). There are several good critical evaluations of the narrative of these conflicts; see e.g. Cornell (1995) and Forsythe (2005); see also Isayev, in Chapter 3 of this volume, for an analysis of Plautine narratives of mobility from this period.

119

Karin W. Tikkanen

existing scripts with which the speakers had become acquainted. These scripts had been introduced partly through migration, as populations foreign to the area had settled in order to set up trade positions, whether from the eighth century onwards, like the Greeks, or a much earlier period, such as the Etruscans. For the early Sabellic inscriptions, there appears to have been no particular ethnic cohesion predicting the choice of script, other than that of the ethnicity currently holding the cultural or socio-economic advantage in the nearby political centre. The Oscan alphabet was never a question of political organisation (Rix 1996: 244), although it is very likely that it came into existence as a result of the political situation after the fall of the Etruscan oligarchies in a number of cities, in the late fifth century, perhaps as a result of the Samnite migrations. The new alphabet was achieved through the merger of select Greek letters into the Etruscan alphabet, in the process of creating a set of signs for the writing of Oscan voiced stops.76 Whatever ethnic group conceived of the Oscan alphabet, whether the Samnites or some other conglomeration of aboriginal Sabellic tribes, the script is, in and of itself, evidence of a bilingual continuity in Campania. In other words, the product of the mobility and migration within the area of Campania is that the speakers of Sabellian made initial attempts at writing, using various scripts made available to them from surrounding Greek and Etruscan communities and mostly without making any changes to the signaries in question. This state of affairs was then followed by a more prolonged language contact which eventually led to a more settled orthography well-adapted to writing Oscan. The majority of the early Campanian Sabellic inscriptions come from Etruscan or Greek settlements, and attest to the contemporary use of several languages, proof of the coexistence of different ethnic and cultural components at these sites (see Bonghi Jovino 2000: 166). The interval between the first inscriptions, in various languages, and those written at the onset of the Oscan alphabet cannot have been more than one or two generations, and it is 76

The Etruscan language remained in use at Pontecagnano down until the late fourth century (Cuozzo 2012: 193), and the speakers of Oscan at Saticula continued to use the Etruscan alphabet until the end of the third century (Crawford et al. 2011: 602–4).

120

The Creation of the Oscan Alphabet

unlikely that this brief interval involved a return to a pre-literary stage, such as Lejeune (1972: 277) saw in the inscriptional corpus from Lucania and Bruttium. It is not really feasible to consider a completely continuous writing tradition as that suggested by Russo (2005), but there was most likely a persistent awareness of the possibilities of writing in the area, together with an ongoing multilingual, and multi-cultural, habitat.

121

CH A P T E R 6

M O B I L I T Y A N D O RT H OG R A P H Y A Contextualisation of Variant Spellings in the Oscan Inscriptions in the Greek Alphabet L I V I A TA G L I A P I E T R A

Between the fifth and the fourth centuries BC, Oscan-speaking populations from the area of Samnium, in Central Italy, spread into the south of the Italian peninsula; here they came into close contact with the Greeks of Magna Graecia.1 The Greek language with which Oscan speakers interacted in this area was by no means homogeneous. In fact, the Greeks who had founded colonies in South Italy had come from various areas of mainland Greece and the eastern Mediterranean, and, as a result, the forms of Greek across this region differed significantly: Ionic around the bay of Naples and in Rhegion, on the strait of Messina; Laconian Doric in Taras and its sub-colony Heraclea, both on the Gulf of Taranto; Achaean Doric in a number of colonies in south Campania, Lucania and Bruttium; Northwest Greek in Locri Epizephyrii, in the toe of the Italian peninsula; and possibly Attic-Ionic in Thurii, a Panhellenic colony founded in the fifth century under the leadership of Athens. Then, towards the end of the fourth and the beginning of the third century, the local forms of Greek became increasingly exposed to the influence of the koine, the new ‘standard’ variety of Greek based on the dialect of Athens and employed by the increasingly dominant Macedonians. In the southern regions of the Italian peninsula, Oscan speakers adopted the Greek script that by that time was uniformly employed by the local Greeks, namely the East Ionic alphabet, for writing their own language. However, the Oscan inscriptions in the Greek 1

Throughout the following discussion, all dates are intended as BC, unless stated otherwise. For a synthetic, yet detailed, overview of the historical events of this period, see McDonald (2015: 5–31).

122

Variant Spellings in Oscan in the Greek Alphabet

alphabet that have survived are characterised by a high level of variation in the correspondence between Greek letters and Oscan sounds, even within the same text. Although the question of spelling inconsistency has been addressed in previous studies, interaction with the local Greeks and acquaintance with their different and changing writing practices has not previously been considered as a possible factor influencing the spelling in Oscan inscriptions. The fact that contact with the local Greeks has not been taken into account in previous studies is particularly surprising when we consider that closeness between Oscan and Greek speakers in South Italy, particularly after the mid-fourth century, is not only well known to us from the accounts of ancient historians and from archaeological data, but also the Oscan inscriptions in the Greek alphabet themselves reveal a high degree of contact and interaction with the local Greeks both from a linguistic and a graphic perspective, as a number of publications have recently acknowledged (particularly McDonald 2015; Zair 2013, 2016). The present chapter thus investigates whether migration and contact with the local Greeks could also have affected the orthography of the Oscan populations, and how consideration of the socio-historical context, including our understanding of the contemporary Greek dialects of South Italy, may help solve some long-debated questions related to the spelling of Oscan in the Greek alphabet. This chapter has been elaborated in the wake of the book Oscan in the Greek Alphabet by Nicholas Zair (2016), which has recently reopened the question of writing conventions in the Oscan inscriptions using Greek letters.2 Aiming to further explore and expand Zair’s conclusions on spelling inconsistency across the corpus, I shall focus here on the major question of the spelling of the Oscan vocalic phonemes. In fact, while other inconsistent spellings that are observable in the South Oscan inscriptions have been convincingly explained in Zair’s study, none of the hypotheses for the spelling of vowels that have been proposed so far is entirely satisfactory. 2

I am most grateful to Nicholas Zair for allowing me to read drafts and proof versions of his book before publication, and for his availability to discuss the subject with me on several occasions.

123

Livia Tagliapietra

Previous Hypotheses Oscan, an Indo-European language of the Sabellic family attested in South Italy between the fifth and the first centuries, had a six-vowel system consisting of the phonemes /i(ː)/, /e(ː)/, /ɛ(ː)/, /a(ː)/, /o(ː)/ and /u(ː)/ (see Figure 6.1) although, in fact, long vowels seem to have been retained in initial or radical syllables only, that is, in stressed syllables (see Zair 2016: 26). The phoneme /u/ is assumed to have developed an allophonic variant [y] when preceded by a coronal consonant (/t/, /d/, /n/, /s/, possibly /r/), as Zair (2014) has discussed. Five original diphthongs with a short first element are known to have been retained unchanged (i.e. not monophthongised) in Oscan: these are /ɛi/, /ai/, /ɛu/, /au/ and /ou/ (Buck 1928: 41–7; Lejeune 1975; Seidl 1994; Wallace 2007: 11–13). Oscan speakers were located in Campania, Samnium, Lucania, Bruttium and Sicily; the alphabet in which the language was written differed across these regions. The writing system in use in the northern areas, Campania and Samnium, which have provided the largest number of Oscan inscriptions, was derived from the Etruscan one and is normally referred to as the ‘national’ or ‘native’ Oscan alphabet (see Tikkanen, Chapter 5 of this volume). As a consequence of Etruscan having a fourvowel system consisting of /i/, /e/, /a/ and /u/ (thus lacking a phoneme /o/ and having only one mid-vowel on the front axis), early native Oscan inscriptions are characterised by the use of the letter for both the back vowels /o/ and /u/ and for both /i/ and /e/ (with instead being used for their more

/i( )/

/u( )/ /e( )/ /o( )/ / ( )/

/a( )/

Figure 6.1 The Oscan vowel system

124

Variant Spellings in Oscan in the Greek Alphabet Table 6.1 Spelling of Oscan vowels in the native Oscan alphabet

Oscan vowel

Native Oscan alphabet pre-300

Native Oscan alphabet post-300

/i/ /e/ /ɛ/ /a/ /o/ /u/









open /ɛ/). Around 300 a writing ‘reform’ was undertaken in these areas, whereby the letters and were created to represent /o/ and /e/ respectively by adding a diacritic to and ; consequently, and came to be restricted to /u/ and /i/ (Table 6.1). On the other hand, the southern regions of Lucania, Bruttium and Sicily have offered a corpus of about a hundred Oscan inscriptions written in the Greek alphabet, mostly dating from the fourth to the early first century; these are mainly from Lucania and Bruttium, while those from Sicily are limited to a small group from Messana (modern Messina); a few inscriptions in the Greek alphabet have also been found in Campania.3 On account of its distribution in the southern Oscan-speaking areas, the Greek alphabet employed in this corpus is generally referred to as the ‘South Oscan’ alphabet. These inscriptions, however, compared to those employing the native Oscan alphabet are characterised by a high degree of inconsistency in their spellings. In particular, major inconsistencies across the corpus are observable in the representation of the phonemes /e/, /ɛi/, /u/, /o/ and /ou/, as illustrated in Table 6.2. In order to account for these variant spellings, Lejeune (1970) hypothesised that a reform in the South Oscan writing system was 3

For an edition of the corpus of Oscan inscriptions in the Greek alphabet, see Crawford et al. (2011). Hereafter, I refer to these inscriptions as printed in this edition, although the phonological interpretation of individual forms in the present discussion relies on the analysis offered in Zair (2016), unless otherwise stated. Dates are supplied as given in Crawford et al. (2011), unless differently discussed.

125

Livia Tagliapietra Table 6.2 Inconsistent spellings of the Oscan vowels in the Greek alphabet Phoneme

Spelling in the Greek alphabet

/e/ /ɛi/ /u/ /o/ /ou/

, , , , , , , , , ,

Table 6.3 Lejeune’s model (simplified)4 Oscan vocalic phoneme Stage 1

Stage 2

/i/ /e/ /ɛ/ /ɛi/ /a/ /o/ /u/ [u] /u/ [y] /ou/





, ,







undertaken towards the early third century, similar to what is supposed to have occurred around the same period in the regions where the native alphabet was in use (Table 6.3). In particular, Lejeune’s hypothetical reform would have prescribed the use of the digraphs and to represent /e/ and /u/ respectively, under the influence of the koine Greek spellings, while previously was used for /i/, for /e/ and /ɛ/, for /u/ and for /o/. In order to represent the original diphthongs /ɛi/ and /ou/ new strategies would have been adopted once and had started to represent /e/ and /u/; namely, the Greek letter would have been introduced for writing the diphthong /ɛi/ (thus ), while the diphthong /ou/ would have come to be spelled or , with the introduction of the Greek in the Oscan writing system. A transitional period in the early years of 126

Variant Spellings in Oscan in the Greek Alphabet

this reform would explain, according to Lejeune, certain inconsistencies within individual documents, such as those observed in Potentia 40/Lu 13 (250–200), which attests the diphthong /ɛi/ spelled according to the ‘new’ convention in κλοϝατηις (/klowatɛis/), and similarly /e/ spelled through the ‘new’ sign in αfακειτ (/anafaket/), but, at the same time, also uses the ‘old’ spelling for the same phoneme in πεhεδ (/pe.ed/). This hypothesis has been generally followed in subsequent scholarship (e.g. Lazzaroni 1983; Del Tutto Palma 1989; Wallace 2007). Already in 1972, however, in the light of new findings contradicting his model, Lejeune proposed a revision and postulated the existence of a small number of scribal schools diverging from these general writing conventions (Lejeune 1972); this revision of the original theory, however,4 did not re-emerge in his later studies (Lejeune 1975, 1990), and probably for this reason has been generally ignored. The new epigraphic evidence published in the last decades has increasingly contradicted the hypothesis of a unitary model for the development of the south Oscan alphabet. Therefore, through an exhaustive presentation of the data, Zair (2016) has questioned and rejected Lejeune’s model by arguing that the epigraphic evidence does not support the existence of such a standardised convention for writing Oscan in the Greek alphabet after 300. Instead, Zair has argued that both and are used alongside to write /e/ throughout the full temporal span covered by the surviving inscriptions, while, from at least 300, both and are used for /o/, and , and could be used for representing /u/ (Table 6.4). Moreover, he has observed that inconsistency frequently occurs within the same inscription: for instance, in Buxentum 1/Lu 62 and are used both for the real diphthongs and for /e/ and /u/, and in the same inscription /e/ is also occasionally spelled and . Similarly, /e/ is spelled with all , and in Potentia 40/Lu 13 and Potentia 1/Lu 5; in Petelia 2, Potentia 1/Lu 5 and Cosilinum 2/Lu 40 /o/ is spelled both and ; and in Potentia 40/Lu 13 /u/ is spelled both and .5 4

5

Aspects of Lejeune’s theory not relevant to the present argument are not reported here; for full discussion, see Zair (2016: 30–44). Zair (2016: 53) has also maintained that the use of for the real diphthong had still not completely been abandoned in the second century, although the argument for this relies

127

Livia Tagliapietra Table 6.4 Zair’s model (from Zair 2016: 79) Oscan vocalic phoneme /i/ /e/ /ɛ/ /ɛi/ /a/ /o/ /u/ = [u] /u/ = [y]

Spelling 1 Spelling 2







Spelling 3

(from third c.)



Zair (2016: 4–5, 170–4; cf. Zair 2013) has therefore interpreted the great variety of spellings across the corpus as the result of the individual choices of different writers, or perhaps of localised groups of writers. According to Zair, it is conceivable that Oscan speakers would primarily learn the Greek letters for writing the Greek language, while a specifically Oscan writing convention never developed (‘bilingualism without biliteracy’); consequently, the graphic representation of the Oscan sounds remained a matter of arbitrary choice among the suitable Greek letters at their disposal.6 However, although Zair’s analysis of the data effectively outlines the variation within the corpus and convincingly points to the absence of any fixed Oscan orthographic convention, one might wonder whether some rational framework can ultimately be identified behind

6

on one inscription only, of uncertain provenance and therefore of unknown context, namely Lucania or Brettii or Sicilia 2, reading ‘δειϝιν(ο)’, dated to 200–100 on account of the lunate shape of epsilon. Besides the use of for the diphthong, however, this inscription differs from contemporary ones in being retrograde, whereas South Oscan inscriptions are normally written left to right, except for a few very early examples (Metapontum 1, Lucania or Brettii or Sicilia 3, Lucania 1). Unless retrograde spelling is due to the fact that this is a stamped object (i.e. the text was written from left to right on the stamp), this may perhaps suggest that the writer of Lucania or Brettii or Sicilia 2 was acquainted with the native Oscan alphabet, as retrograde writing is characteristic of the inscriptions of Campania and Samnium: compare Crawford et al. (2011: 1312). If this is correct, the use of for /εi/ instead of may indicate influence from the native Oscan alphabet (where /εi/ was written , i.e. /ε/ + /i/ ) and/or unfamiliarity with the spelling in use in the southern Oscan areas. For the concept of bilingualism without biliteracy, see Adams (2003a: 40–67), referred to by Zair (2016: 170).

128

Variant Spellings in Oscan in the Greek Alphabet

different choices of possible spelling. Even if the primacy of Greek in the literacy of Oscan writers may be considered as a context in itself for the occurrence of inconsistent spellings, still a comprehensive attempt at contextualising diverging practices has not yet been attempted. In particular, elements that need further consideration are the geographical and chronological context of individual inscriptions, the phonology and writing conventions of the local Greeks, and the Oscan phonological system itself. In the following discussion, I propose a more precise definition and contextualisation of individualism in the spelling of South Oscan inscriptions in the light of these factors. In particular, I shall concentrate on the inscriptions dating to the fourth and third centuries, that is, the earliest in the corpus, as the paucity of Greek evidence prevents investigation of Greek influence on Oscan orthography after c. 200. The Oscan inscriptions from Messana will be dealt with separately from the rest of the corpus. The authors of these inscriptions are identified with the Mamertini, a group of Italic mercenaries who had served under Agathocles and took over the Greek polis of Messana around 289, after the death of the Syracusan ruler. According to ancient sources, they came probably from Campania, or perhaps from Samnium (Polyb. 1.7.2, 1.8.1; Strabo 6.2.3; Fest. in Lindsay 1913: 150). In either case, their original provenance was from an area where the native Oscan alphabet was in use; therefore, their adoption of the Greek alphabet should have occurred at a date close to the recording of the inscriptions themselves, that is, sometime in the late fourth or early third century, possibly during their service under the Greeks, or as they took Messana. Zair (2016: 138), however, has also acknowledged the possibility, relying on Festus’ narration of the events, that the Mamertini adopted the Greek alphabet in Tauricana, in the south of modern Calabria (if correctly connected with the Tauriani of Bruttium, for which see Crawford 2007: 277; Crawford et al. 2011: 58, 1505–10), to where they are said to have initially moved from Samnium. Fresh adoption of the Greek script on the part of the Oscan speakers in Messana is also suggested, as discussed by Zair, by peculiar spellings in the Messana inscriptions, particularly in personal names, apparently reminiscent of previous familiarity with the native Oscan practice (Zair 2016: 129

Livia Tagliapietra

137–41; see also McDonald 2015: 90–2; also Zair, Chapter 7 in this volume). The inscriptions from Messana therefore reflect the orthographic practice of a localised group of non-local Oscan speakers within a limited period of time as the script had been newly adopted. Comparison of these with the rest of the corpus is of crucial importance to our understanding of variant spellings across the corpus, as I shall discuss at the end of this chapter. In the following discussion, inconsistency in the spellings mentioned above will be first considered from a phonological perspective, and the rational motivations behind these will be identified. I shall then attempt a coherent and comprehensive explanation for the variation emerging from the corpus by combining the evidence from phonology with the socio-historical context. A schematic presentation of the inscriptions discussed, chronologically ordered, is provided in Tables 6.5 and 6.6. Phonological Motivations: Disambiguating Spellings It is important to note that the use of certain Greek graphemes in Oscan inscriptions contradicts Greek scribal practice. According to Greek conventions, both and could be used for a front mid-vowel ( properly representing a long close front midvowel; but the length implied by such a spelling in Greek could have been ignored by Oscan speakers, as vowel length distinctions had been lost outside initial syllables, i.e. in unstressed syllables, in Oscan). The employment of for /e/ in Oscan, however, contradicts Greek practice, according to which could only represent /i(ː)/. One might wonder whether the spelling reflects an actual raising of /e/ to /i/; however, although in a few words the spelling for /e/ seems consistent throughout the corpus, no consistent phonological environment in which /e/ is spelled can be detected in the surviving South Oscan inscriptions.7 A more satisfactory solution might be offered by 7

Note particularly the consistent spelling of the Oscan indefinite pronoun derived from PIE *kwis with in place of expected /pes/ (Petelia 2: πιˍσπιˍτ; Buxentum 1/Lu 62: ποκκαπιˍδ, πιˍς). Similarly, the spelling for expected /e/ in the gentilicium τιˍτιδιες is attested in two different inscriptions, Potentia 17/Lu 15 and Potentia 12/Lu 27: for the etymology of this name, see Zair (2016: 225).

130

8

μεδεκον (×3) μεδεκαν

μαραειν (×2)

ϝεζεις

/ɛi/

See Zair (2016: 210, n. 22) for alternative reading.

Caulonia 2 (325–300) Buxentum 3/Lu 45 (400–300)

μεδικιαι

Metapontum 1/ Lu 37 (400–375) Lucania or Brettii or Sicilia 3/ Lu 18 (375–350) Laos 2/Lu 46 (330–320)

αναfακετ

/e/

Inscription

[y] νοψιν (×3)8 νοψ(ι)α(ν)

[y] συπ

/u/ < *ŭ

βοθρονι[(ο)ν]

/u/ < *ō

πολλ[ιε]ς φοινι[κις]

λοικιν

μεταποντινας

/o/

(/ow/) οϝι(ν)

/ou/ and /ow/

Table 6.5 Spelling in South Oscan inscriptions c. 400–200 BC (excl. Messana)* = date proposed by Crawford et al. (2011) questioned by Zair (2016) since based on the spelling of /εi/ ( considered earlier than )

10

9

πισπιτ ι(νι)μ καιδικω καιδικις

αϝες

/e/

ιουϝηι [2]+[1] αναρηι βρατηις

μινιηις

ετ/ηις or ηισου(μ)9

/ɛi/ [y] νομψις

/u/ < *ŭ

ηισου(μ) (gen. pl., if the reading is correct)10

σολλομ (gen. pl.)

ορτοριες

/u/ < *ō

See McDonald, Tagliapietra and Zair (2015) for alternative readings in this inscription. The alternative reading ουσοσ could represent either /ousus/ or /u(:)sus/; see Zair (2016: 212).

Paestum 2 (300) Petelia 2 (300)

Thurii Copia 1/ Lu 47 (350–300) Laos 3/Lu 63 (300) Paestum 1/Lu 14 (300)

Inscription

Table 6.5 (cont.)

σολλομ πακολ καϝνοτο στατιο εμαυτο μιναο (×2) αυδαο

λοικ(ις) λοικες

ορτοριες

/o/

(/ow/) νοϝιο

(/ou/) λο ̣ϝκις ̣ (/ow/) ιουϝηι

/ou/ and /ow/

Cosilinum 1/Lu 3 (300 Crawford et al.; before 200 Zair) Potentia 24/Lu 30 (325–275) νετεfς πεhετεfς Potentia 13/Lu 16 (325–275*)

Vibo 2/Lu 25 (300*)

μεfιτει

διουϝει ϝερσορει

[τανγιν]ο

ϝερσορει

νοϝιο

πακϝιω και{αι} δω στατιω καιδικω τρεω αλαφιω σκαφιριω βαντινω κωσσανω (/ow/) διουϝει

Potentia 17/Lu 15 (300–275 Lejeune 1971; 300–200 Crawford et al. 2011) Vibo 7/tLu 7 (300–275*) Numistro 1/Lu 4 (300–275)

Potentia 19/Lu 36 (325–275*) Potentia 20/Lu 28 (325–275 Crawford et al.; 325–200 Zair)

Inscription

μεδδικεν

σουϝεν μεδδικεν

τιτιδιες

/e/

τουρειεις

οκ(ι)ηις μεfιτηι β[ρ]α ̣{ι}τηις

βρατεις

μαμερτει

/ɛi/

[y] τουρειεις

[y] νυμψδοι (×2) νυ[μψδαναι]

/u/ < *ŭ

Table 6.5 (cont.) /u/ < *ō

μεfιτανοι

οιναι νυμψδοι μεfιτανοι μαμερττοι

/o/

(/ow/) σουϝɛ ̣ν

/ou/ and /ow/

Potentia 12/Lu 27 (300–200) Buxentum 1/Lu 62 (300–200)

Crimisa 3/Lu 44 (300–250)

Anxia 1/Lu 39 (300–250)

τιτιδιες

ποκ]καπιδ ποκκαπιδ ιαf ιπειδ ουπ ̣ιδ πις

με]δδες εινεμ μεδδες ιουfετοδ με]δδε[ς

ειν(ειμ) λεικειτ λιο{κα} κειτ μειαι

ιμες

διοϝηι

εισεις ]υ ̣κεις

μ]αχερηι

fουστ

εστουδ σουf ακτουδ hαfειτουδ ιουfετουδ τ{αν} αγγινουδ παντου[δ τουτ]ε ̣ικουδ πονδιου[μ (gen. pl.)

λοπουστ ποια[δ πονδιου[μ σερευ] κιδιμο ειοκ (×2)

ποπεδ[ι(ο)μ] πολλιες

πωτ κω[ρο]

(/ow/) διοϝηι

τουτεικαις

Potentia 15/Lu 33 (300–200) Potentia 31/Lu 59 (300–200)

Crimisa 1/Lu 23 (300–200) Crimisa 2/Lu 24 (300–200)

Inscription

μεδεικα[τεν τρειβιϝομ hαfειτουδ ιπειδ εινεμ τουτεικαις ειοκ (×2) τουτ]ε ̣ικουδ]ι [1]εδειστ τρειπκατομ

/e/

πακτηις

ωυδδιηις (h)εριηις μαι[-?-]ηις

μεfιτηι

]κηις

/ɛi/

[y?] ερουκης

/u/ < *ŭ

Table 6.5 (cont.) /u/ < *ō

/o/

ωυδδιηις

/ou/ and /ow/

Potentia 18 (225–200) Potentia 44/tLu 1 (225) Vibo 5/tLu 3 (before 200) Vibo 6/tLu 8 (before 200) Vibo 8/tLu 6 (before 200) Lucani 1/nLu 1 (207–204)

Potentia 21/Lu 29 (250–200)

Potentia 40/Lu 13 (250–200)

κοττειηις

μαιηις

ορτιηις

κλοϝατηις αυτι

μετσεδ πεhεδ πλαμετοδ

αfακειτ

[y?] αρροντιες



fλουσοι

κhομοι νυμ]ψδοι

μα]μερτοι

fατ ̣[οϝ]ο ̣ɩ ̣ fλουσοι

λουκανομ

(/ow/) κλοϝατς δ ̣ɩοϝιοι ̣ fατοϝε κλοϝατηις

Potentia 16/Lu 32 (either side 200) Potentia 22/Lu 31 (either side 200) Potentia 25/Lu 21 (either side 200) Teuranus Ager 1 / Lu 43 (presumably before 200) Lucania or Brettii or Sicilia 1/ Lu 26 (after 200 Crawford et al.; no date Zair)

Potentia 11/Lu 35 (either side 200)

Inscription

/e/

hερεκλεις

ζωϝηι πιζηι

μεfιτηι ̣

ϝενζηι

/ɛi/

[y] νυμψιμ

/u/ < *ŭ

Table 6.5 (cont.)

δουνακλο ̣μ ̣

καποροιννα ̣[ι]

λουκανομ (gen. pl.)

/u/ < *ō

/o/

(/ow/) ζωϝηι

/ou/ and /ow/

Potentia 1/Lu 5 (125–100 Crawford et al.; 200–100 Zair)

κενσορτατηι ειζιδομ αιζνιω κωσ(τ)ιτ

ρεγο(μ)

ειν(ειμ) ειζιδομ

ποκιδ(ιηις)

πωμπονις κενσορτατηι τανγινοτ ρεγο(μ)

σταβαλανο σεγονω

πωμπονις πωμfοκ(ιαι) πρωfατεδ κωσ (τ)ιτ σεγονω αιζνιω πσανω

εινειμ

σ]ταττιηις νιυμσδιηις αππελλ[ο]υ ̣νηι

στεννιηις [α]π ̣πελλουνηις

μαμερε ̣κηις

σταττιηις νιυμσδιηις αππελλουνηι

/ɛi/

[y] νιυμσδιηις

[y] νιυμσδιηις

/u/ < *u

Messana 4/Me 1 & 3 and Messana 5/Me 2 are copies of the same inscription.

Messana 2/tMe 1 (perhaps before 200)

Messana 1/nMe 1a–b (225)

Messana 7/Me 5 (250)

Messana 5/Me 2 (250)

11

/e/

Messana 6/Me 4 (275) Messana 4/Me 1 & 311 (250) μεδδειξ ̣ εινειμ

Inscription

/o/

μαμερτινουμ (gen. pl.)

μαμερτινουμ (gen. pl.)

[α]π ̣πελλουνηις

πομπτιες τωϝτο μαμερτινο σακορο

αππελλ[ο]υ ̣νηι μαμερτινο ̣ σακορο

αππελλουνηι

/u/ < *ō

Table 6.6 Spelling in the Messana inscriptions

(/ou/) τωτ[ο]

(/ou/) τωϝτο

/ou/ and /ow/

Variant Spellings in Oscan in the Greek Alphabet

the non-correspondence between the Oscan and Greek vowel systems: while Oscan had two mid-vowels on the front axis, Greek only had one short central mid-vowel spelled , and, depending on the dialect, one or two long mid-vowels, spelled and . Since the signs and representing long front mid-vowels in Greek seem to have been ignored by Oscan writers before c. 300 (see below), we assume that when Oscan speakers first adopted the Greek alphabet they were familiar with only one sign, , for representing both their mid-vowels /ɛ/ and /e/. Apparently, /ɛ/ was identified as corresponding to the Greek midvowel spelled without exception, as the consistent spelling for Oscan /ε/ confirms. On the other hand, Oscan writers could consider /e/ as closer to either /ɛ/, spelled , or /i/, spelled . Familiarity with the native Oscan practice before 300, which had for both /i/ and /e/, could perhaps have further influenced the choice of as an acceptable spelling for /e/ in the southern Oscan regions. Similarly, the use of for /u/ in Oscan inscriptions is unparalleled in Greek practice. As Zair (2016: 63–79) has observed, the spelling is never used for [u] (/u/ in non-post-coronal position), which is only spelled and ; on the contrary, is attested for the allophone [y] (/u/ after a coronal), alongside and . The fact that [u] is never spelled in South Oscan inscriptions thus suggests that was employed to represent allophonic [y], leaving only and as possible spellings for [u]. The fact that in a number of inscriptions [y] is also spelled and (i.e. in the same way as /u/ in other environments: see Thurii Copia 1/Lu 47, Laos 2/Lu 46, Vibo 7/tLu 7, Crimisa 1/ Lu 23, Potentia 44/tLu 1) can be explained in the light of the nonphonemic nature of the sound represented by in Oscan, as already observed by Zair (2016: 173). Thus some writers may have unconsciously tended to eliminate to the advantage of the graphemes representing /u/ in the majority of contexts. Thus, differences between the Oscan and the Greek phonological system provide a plausible context for inconsistent spellings that do not have correspondents in Greek writing practice. It is now necessary to assess whether variant spellings that are compatible with the Greek use may have been influenced by contact with the local 141

Livia Tagliapietra

Greeks. In order to answer this question, we first need to determine whether alternative spellings for the representation of an Oscan phoneme were completely equivalent from a functional perspective, and what their distribution within the corpus of Oscan inscriptions is. Interestingly, the digraphs and , which in koine Greek represented close long mid-vowels, seem not to be used as spellings for Oscan /e/ and /u/ respectively before the late fourth or early third century, while they are widely attested afterwards. This conforms to what had been already observed by Lejeune. In earlier inscriptions, /e/ is normally spelled or , and /u/ . The earliest surviving inscription in which is used to represent /e/ is Laos 2/Lu 46, dating to 330–320, followed by Anxia 1/Lu 39 (300–250), Buxentum 1/Lu 62 (300–200) and Potentia 40/Lu 13 (250–200), while the first occurrence of the digraph for /u/ ([u]) appears in Petelia 2 (300). Zair (2016: 53) has argued that the absence of and for /e/ and /u/ in the surviving documentation before c. 300 might be just a matter of chance, since a very few inscriptions are earlier than the first attestations of the digraphs for /e/ and /u/.12 However, had this been the case, it would be difficult to explain why and were not readily employed as the unique spellings for /e/ and /u/ when the Greek alphabet was first adopted for writing Oscan in these regions. The spellings / and attributed to the phonemes /e/ and /u/ respectively were ambiguously shared with other phonemes: and for spelling /e/ could also be used for /ɛ/ and /i/ respectively, generating ambiguity; likewise, the spelling could represent both /o/ and /u/. By contrast, the digraphs and would distinguish /e/ and /u/ from /ɛ/, /i/ and /o/ respectively. Had these been available at an earlier stage, it would be very surprising not to find them employed for /e/ and /u/ from the beginning in order to avoid homography.13 But most 12

13

These are Metapontum 1/Lu 37 (400–375) and Lucania or Brettii or Sicilia 3/Lu 18 (375–350) (but possibly also Potentia 24/Lu 30, assigned to 325–275). It may be argued that in the few inscriptions dating before 300 only the short vowels /e/ and /u/ occur, and therefore, in principle, it is possible that we do not have and for /e/ and /u/ in pre-300 inscriptions because the writer/adapter of the Greek alphabet for Oscan preferred to use / and , as these represented short vowels in Greek. Such an argument is not compelling: in fact, we can easily observe that after c. 300 and are very frequently used for representing short vowels in Oscan, while, on the

142

Variant Spellings in Oscan in the Greek Alphabet

importantly, the hypothesis that and were not used for /e/ and /u/ before at least the late fourth century is strongly suggested by changes in the spelling of the diphthongs /ɛi/ and /ou/ around the same date. The use of for the diphthong /ɛi/ also seems to first appear in inscriptions dating around 300 (Paestum 1, Paestum 2, Petelia 2), while in the earlier Caulonia 2 (325–300) and in a number of other texts dating around 300 (Vibo 2/Lu 25, Potentia 13/Lu 16, Potentia 19/Lu 36), the spelling used for the diphthong is . After c. 300, is consistently employed for spelling /ɛi/.14 Correspondingly, before the beginning of the third century the original diphthong /ou/ is spelled either or / before a vowel;15 but after c. 300 the spelling is abandoned

14

15

other hand, the spellings / and can also be used for long vowels in initial syllables. Cf. Zair (2016: 73–4). Possible exceptions are Lucania or Brettii or Sicilia 1, Vibo 7/tLu 7 and Lucania or Brettii or Sicilia 2 discussed earlier (see n. 5). However, Lucania or Brettii or Sicilia 1 has been dated after 200 by Crawford et al. on account of the sign ‘half-h’, erroneously assumed to be a post–third-century feature: the sign is in fact already attested in Petelia 2 (300), and therefore attribution to the second century on this basis only is questionable; see Zair (2016: 96–8). Vibo 7/tLu 7 has been dated to 300–275 by Crawford et al. on the basis of the spelling for /ɛi/ instead of , but Zair (2016: 177) has proposed to reassign it to a date close to the other Oscan tile stamps found in the area (Vibo 5/tLu 3, Vibo 6/tLu 8, Vibo 8/tLu 6), in the later third century. There is no compelling argument for this; if nothing else, the fact that Vibo 5/tLu 3, Vibo 6/tLu 8 and Vibo 8/tLu 6 consistently have , in conformity with the rest of the post-c. 300 South Oscan inscriptions, places Vibo 7/tLu 7 separately from these. We may observe that stamps on tiles represent a rather standardised writing context, both on account of their serial reproduction and of the legal/official statement that names or sigla in these were meant to convey. Accordingly, it seems implausible to ascribe the spelling for /ɛi/ in Vibo 7/tLu 7, instead of in Vibo 5/tLu 3, Vibo 6/tLu 8 and Vibo 8/tLu 6, to the extemporaneous choice of an individual, but rather it should be considered as conforming to the current usage in such a context. As Vibo 5/tLu 3, Vibo 6/tLu 8 and Vibo 8/tLu 6 diverge from this, they most plausibly reflect the convention adopted at a different (presumably later) stage. According to Zair (2016: 59–60) the spelling in διουϝει (Vibo 2/Lu 25) and ιουϝηι (Paestum 1/Lu 14) reflects the ambisyllabicity/heterosyllabicity of the second element of the diphthong before a vowel. He has observed that in διουϝει/ιουϝηι (dat. sg. ‘to Jove’) must represent /ow/ and not /uw/ since in later examples the same word is spelled with (Potentia 12/Lu 27: διοϝηι) or , which can only represent /ow/ (Potentia 11/Lu 35: ζωϝηι; Potentia 9/Lu 6: διωϝηις; Potentia 10/Lu 7: διωϝιιας). Similarly, in the native Oscan alphabet the name of the deity is spelled with after c. 300 (Histonium 5/Fr 5: iúveís, Teruentum 34 A.11, 12, B.14, 15: diúveí, Cumae 4/Cm 9: iúveí), and also the spelling in the Oscan inscriptions using the Latin alphabet is (/ou/: Bantia 2/Lu 38: zoveṣ). is therefore a glided variant in ambisyllabic contexts of the spelling employed for tautosyllabic /ou/ (cf. tautosyllabic /ou/ spelled in τουτεικαις in Buxentum 1/Lu 62).

143

Livia Tagliapietra

in favour of (Petelia 2, Potentia 12/Lu27, Potentia 40 /Lu 13), (Crimisa 2/Lu 24), and (Potentia 11/Lu 35).16 Although it must be admitted that the extant evidence is scant, the fact that after c. 300 only is attested as the spelling for /ɛi/, while is no longer employed for /ou/ can hardly be regarded as pure coincidence: indeed, since absence of a standardised orthographic norm is assumed, it would be very hard to explain why such practices became consistent throughout Lucania and Bruttium after the early third century. It therefore seems very likely that new digraphs were introduced for spelling the diphthongs around 300, when and started to be widely used for /e/ and /u/.17 In this connection, we observe that the letter in Oscan inscriptions is exclusively employed for spelling the diphthong /ɛi/ (), and never for /ɛ/ alone. Such an observation strongly suggests that was intentionally adopted from the Greek writing system with the specific purpose of being employed for the representation of the diphthong; the only context that could plausibly justify the development of the spelling for /ɛi/, previously written , would be the necessity to distinguish the real diphthong from /e/ spelled . We may also observe that the local Greek communities had been using the letter , as well as , since the adoption of the East Ionic script around the early fourth century, and therefore the appearance of in Oscan inscriptions only around 300 would be very surprising, unless we assume a connection between the adoption of and the

16

17

The only exception might be represented by the reading λουκανομ on coinage from Lucania dated to 207–204 (Lucani 1/nLu 1), in which is used for the real diphthong. It can be observed, however, that spellings on coins tend to be very conservative in general; moreover, internal consistency here is observed, since is used for /ou/ and not for /u/ in the genitive ending, which is rather spelled ; see Zair (2016: 67–8, 81–3) for the phonetic value of the vowel in the Oscan o-stem and consonant-stem genitive plural endings. The fact that some post-300 inscriptions such as Buxentum 1 and possibly also Vibo 7/tLu 7 (see Zair 2016: 225) use and both for /e/ and /u/ and the diphthongs may simply indicate unawareness of the new disambiguating spellings. A possible exception to the consistent use of for /ɛi/ and for /e/ in the same inscription after c. 300 is κοττειηις (Vibo 5/tLu 3) representing either /kotteɛis/ or /kottɛjɛis/ (gen. sg., gentilicium), if the second interpretation is correct (although one might observe that the ambisyllabicity of the second element could have altered the perception of the dipththong).

144

Variant Spellings in Oscan in the Greek Alphabet

necessity to provide a new spelling for the diphthong /ɛi/, once had started to be used for /e/. The evidence for the introduction of the spellings and for /e/ and /u/ around 300 and not earlier is sufficient and convincing when all the available data are considered. Thus , and seem to have been the early spelling for these phonemes, ambiguously shared with /ɛ/, /i/ and /o/. As the spellings and were introduced, new strategies for spelling the real diphthongs /εi/ and /ou/ were adopted. The fact that a number of disambiguating spellings are attested for the diphthong /ou/ after c. 300 (, , ) presumably reflects the absence of a standard graphic convention in Lucania and Bruttium. While this assumption does not contrast with Zair’s general claim for individualism in South Oscan inscriptions compared to the standardised orthographic norm for writing Oscan in the north, it implies that ‘individualism’ should not be interpreted simply in terms of extemporaneous choices among the signs available. Finally, the letter first appears in South Oscan texts in inscriptions dating around 300 to represent both simple /o/ and the first element in the diphthong /ou/. Zair (2016: 63) has considered the evidence for pre-300 spellings of /o/ copious enough to assume that was not in use before the early third century: before c. 300 /o/ is only spelled (Metapontum 1/Lu 37, Laos 2/Lu 46, Buxentum 3/ Lu 45, Thurii Copia 1/Lu 47, Laos 3/Lu 63), while the earliest example surviving of in the spelling of the diphthong occurs in Crimisa 2/Lu 24 (300–200). The use of for simple /o/, however, seems not to be evenly spread, and is found in very few inscriptions in our corpus; these are: Petelia 2 (300), Anxia 1/Lu 39 (300–250), and later in the second century in Potentia 1/Lu 5 (125– 100 Crawford et al.; 200–100 Zair), Potentia 23/Lu 64 (125–100) and Cosilinum 2/Lu 40 (100). The other inscriptions that date after 300 and attest to /o/ instead have to represent this phoneme. However, inconsistency in the use of and for /o/ across the corpus largely disappears when individual inscriptions are considered. In fact, the spellings , and never all occur in the same inscription, with the exception of Petelia 2.18 18

Compare Nava and Poccetti (2001) and Poccetti (2015) for a similar observation.

145

Livia Tagliapietra

Generally, inscriptions having for /o/ do not also have for /u/, but rather retain the older spelling for representing /u/ (Potentia 1/Lu 5, Cosilinum 2/Lu 40); similarly, those having for /u/ retain for /o/ and do not employ for this phoneme (for instance, Buxentum 1/Lu 62, Potentia 40/Lu 13, Potentia 25/Lu 21). If the motivation for adopting either or around 300 is considered, namely the necessity to disambiguate the spelling of /u/ and /o/ (both were previously written ), it seems clear that both and should be regarded as equivalent strategies adopted by different writers, or perhaps schools of writers, in order to distinguish the spelling of the two phonemes: some decided to change the spelling of /u/, others that of /o/. We may perhaps hypothesise that change of the spelling of / o/ instead of /u/ was undertaken under the influence of the native Oscan alphabet, in which the new signs introduced with the reform around 300 were those for /e/ and /o/, instead of /e/ and /u/. The later Potentia 23/Lu 64 may perhaps be considered as a convergence of the two practices, having for /o/ and for /u/ with no use of . In Petelia 2 (300) /o/ is spelled both and , while /u/ is spelled both and . While the use of for both /u/ and / o/ in this inscription may be explained in terms of conservative spellings ( had been long used for both /o/ and /u/), the simultaneous appearance of and may be surprising in light of the fact that other inscriptions generally show either for /u/ or for /o/. However, the date attributed to this inscription would not be incompatible with a period of experimentation with the spellings which had recently become available. Even if extemporaneous choices among the signs available were postulated for Petelia 2, the fact that the majority of the evidence does not attest to simultaneous use of and needs to be acknowledged; therefore, Petelia 2 at the moment seems better considered as the exception rather than the norm. The spellings and for /e/ and /u/, the use of the digraphs and // for the diphthongs, and that of for /o/ thus do not find a compelling motivation within a context of extemporaneous choices among the Greek signs compatible with the Oscan phonemes. Rather, these variant 146

Variant Spellings in Oscan in the Greek Alphabet Table 6.7 Spelling disambiguation (highlighted: ambiguous spellings)

Phoneme

Before c. 300

/i/ /e/ /ɛ/ /ɛi/ /u/ = [y] = [u] /o/ /ou/

,



, ,

Disambiguating spellings

, , , ,

spellings need to be considered as rational responses to the need to disambiguate the spelling of different phonemes. A summary of the disambiguating strategies discussed in this section is provided in Table 6.7. Defining Contextual Motivations Having accounted for much of the inconsistency in the corpus from a phonological perspective, two questions remain to be addressed in order to provide a more comprehensive framework for assessing the hypothesis of ‘individualism’. Firstly, why were and employed to spell /e/ and /u/ only around 300 and not earlier? And, second, how should inconsistency within the same inscription be interpreted after the introduction of the new spellings? In fact, if unawareness of a Greek grapheme that could represent Oscan /e/ can justify the uncertainty between the spellings and in the same inscription before c. 300, it seems very surprising to still find these spellings alongside once this has been introduced (e.g. Buxentum 1/Lu 62, Potentia 40/Lu 13, Potentia 1/Lu 5; note also that Petelia 2 uses for /ɛi/, but has / e/ spelled either or instead of , as would be expected). Similarly, after 300 it is still possible to find both and for /u/ (Petelia 2), and both and for /o/ in the same 147

Livia Tagliapietra

inscription (Petelia 2, Potentia 1/Lu 5). Such questions find plausible answers if the socio-historical context and contacts with the local Greeks are considered. Why did = /e/ and = /u/ appear only around 300? As we observed, the introduction of and for /e/ and /u/ respectively satisfies internal necessities, namely the need to distinguish /e/ from both /ɛ/ and /i/, and /u/ from /o/. Such necessities, however, were clearly already felt when Oscan speakers first began to use the Greek alphabet, and therefore these alone do not account for the appearance of and around 300; rather, it is necessary to postulate the occurrence of favourable external circumstances for the adoption of these spellings around this date but not earlier. First, we may observe that in most areas, particularly in Bruttium, Oscan speakers came into close and frequent contact with the local Greeks only in the second half of the fourth century. Although Poseidonia (later Paestum) possibly had a significant Lucanian presence already in the early fourth century (Strabo 5.4.13, 6.1.1), most other sites, such as Sybaris on the Traes, Hipponion (later Vibo), Temesa, Caulonia and several other poleis of South Magna Graecia (possibly including Skylletion and Terina) are known to have been taken by the Brettii around 346 (D.S. 16.15; Strabo 6.1.5), while Thurii suffered from their attacks around 344 (Plut. Tim. 16.3–4) and Croton summoned the Syracusans for aid against this Oscan population in 325 (D.S. 19.3).19 Therefore, we expect that it was precisely around this date that Oscan speakers of the South became literate in Greek on a wide scale, and consequently accustomed to Greek writing conventions. We might suppose that previously the Greek letters had been simply mechanically assigned to the Oscan sounds with little or no experience in writing Greek on the part of the writers, 19

For discussion of the evidence for the Lucanians in Poseidonia, see Wonder (2002); McDonald (2015: 244–5); Isayev (2007: 110–14); Crawford et al. (2011: 50); for the Brettii in the southern areas, see De Sensi Sestito (1995: 61–70); Wonder (2012: 135–7).

148

Variant Spellings in Oscan in the Greek Alphabet

and that the monophthongised value of the digraphs and in Greek was simply not known to the majority of them. But another possibility is that the monophthongised value of the digraphs and developed in the language of the local Greeks only at this stage, under the influence of the Hellenistic koine, and not earlier. Accordingly, the possibility to employ and for single vowels was simply not available to Oscan writers until c. 300, as the local Greeks would have pronounced such sequences as real diphthongs in their own language before then. In fact, we may observe that some of the poleis that are known to have been captured by Oscan speakers in the second half of the fourth century still do not attest to and for Oscan /e/ and /u/ until the beginning of the third century. This is particularly the case of the inscriptions from Caulonia and Hipponion. In Caulonia 2 (325–300) and Vibo 2/Lu 25 (300) is used only for the diphthong /ɛi/, while is employed for spelling the ambisyllabic diphthong /ow/, and /u/ is spelled .20 Therefore it seems that the spellings and for /e/ and /u/ in Caulonia and Hipponion are not immediately connected with closer interaction with the local Greeks in the second half of the fourth century. In this connection it seems relevant that the earliest attestation of koine features in the Greek inscriptions from the nearby site of Locri date precisely to the beginning of the third century. These include the introduction of the spellings and for secondary long mid-vowels from compensatory lengthening and contraction, while previously these digraphs were used only for the original diphthongs in the local severior Doric dialect (i.e. a variety of Doric in which primary and secondary long mid-vowels had merged and were spelled and , whereas other dialects, e.g. Attic20

Caulonia 2 has been dated on the basis of the archaeological context, while Vibo 2 /Lu 25 on account of the spelling of these vowels. Despite Zair’s legitimate criticism about the circularity of dating the inscriptions on the basis of the spellings we aim to date, the fact that in general for /ɛi/ is no longer found in South Oscan inscriptions after c. 300 most plausibly suggests a date around the beginning of the third century; later inscriptions from the same area, presumably dating to the second Brettian occupation during the third century (on the basis of Livy 35.40), have instead (Vibo 5/tLu 3, Vibo 6/tLu 8, Vibo 8/tLu 6, all dated ‘presumably before 200’. A terminus ante quem for Vibo 2/Lu 25 may be provided by the reconquest of Hipponion by Agathocles in 295 (D.S. 21.8; Strabo 6.1.5; cf. De Sensi Sestito 2015: 52–5).

149

Livia Tagliapietra

Ionic, had two pairs of long mid-vowels, spelled / and / respectively).21 The evidence available does not allow us to determine whether the spellings and still represented /ei/ and /ou/ in the local Doric dialect, or if by the fourth century the original diphthongs might have been monophthongised as in AtticIonic; but it clearly cannot be excluded that these were in fact monophthongised only around 300 under the influence of the koine.22 In either case, the spread of the koine in the early third century significantly increased the frequency of occurrence of and with monophthongised value in the texts by local Greek speakers, since before then secondary long mid vowels from compensatory lengthening and contraction were spelled or . Possibly, the two explanations proposed are not incompatible: as in some Oscan inscriptions the new spellings seem to appear at a slightly earlier stage compared to others (see e.g. for /ɛi/ in Laos 2/Lu 46, dating to 330–320, but still in Caulonia 2, dating to 325–300), it is possible that in certain areas the Greek original diphthongs had already been monophthongised and were used for secondary long mid-vowels by the time the Oscan speakers came into close contact with the Greeks during the fourth century, while in others, where a severior Doric dialect was in use, this development either did not occur or had a very low frequency of occurrence before the early third century. In particular, among the Greek dialects with which Oscan speakers came into contact in Lucania and Bruttium, the spelling of secondary long mid-vowels as and already before the diffusion of 21

22

The koine spellings and for secondary long mid-vowels not originating from monophthongisation of /ei/ and /ou/ first appear in I.Locri 5.6; 20.1, 2, passim (late fourth/mid-third c.); 66; 69 (fourth/third c.); 90.11; 108; 89.5 (third c.). For the severior vocalism at Locri, see Méndez Dosuna (1985: 75–7). Other koine features that appear in the inscriptions of Locri around this date are: (/ɛː/) in place of original /aː/ () (I. Locri 46; 91; 63; 81; 111); ephelcystic nu (I.Locri 4.11; 89.16); the ordinal numeral πρῶτος (I.Locri 2.6; 16.7); the hybrid form ἴκοσι (I.Locri 35.8); the Attic-Ionic adjective ἱερός (I.Locri 2.1; 3.6; 5.4, 7–8; 6.2; 7.1, 9; 10.1; 11.1; 12.1; 15.1; 16.4; 17.1; 18.1; 19.1; 20.17; 24.5; 26.1; 27.6; 28.4; 29.1; 37.6; 56; 64; 65; 66; 67); dat. sing. ending -ει of i-stem nouns (I.Locri 20.20); the form χρῆσις (I.Locri 20.22) instead of local χρῆστις (I. Locri 1.8, 4.8, 15.9, 25.10); the adverb ἔμπροσθεν (I.Locri 20.23); use of the relative pronoun instead of the local use of the article with relative force (I.Locri 7.11; 89.6, 13); the particle εἰ (I.Locri 91.3; 89.15); and the nominative plural masculine article οἱ (I. Locri 91.2). For the possibility that monophthongisation had not occurred in severior Doric dialects before the arrival of the koine, see also Méndez Dosuna (1985: 273 n. 10).

150

Variant Spellings in Oscan in the Greek Alphabet

the koine is expected in the dialect of Velia (founded by East Ionian colonists) and probably also in that of Thurii, which was founded under the leadership of Athens.23 Moreover, some areas may have yielded to the koine more quickly than others: for instance, the koine seems to have had a stronger influence in Croton than in Locri towards the end of the fourth century, as evidenced by a fragment of an official inscription from Croton attesting the spelling for the secondary long back mid-vowel, whereas contemporary official inscriptions from Locri retain the spellings and for secondary long mid-vowels until at least the mid-third century.24 Whether the spellings and for /e/ and /u/ were introduced in Oscan inscriptions as Oscan speakers took control of the Greek poleis in the second half of the fourth century, or in the early third century, as the koine spread among the local Greeks, the appearance of new spellings in south Oscan inscriptions in the late fourth and early third century is clearly the result of proximity to the Greeks and literacy in their language. Indeed, bilingualism and literacy in Greek is confirmed by a number of other features in these inscriptions, such as code switching into Greek, imitation of the letter shapes in contemporary Greek inscriptions (e.g. alpha with broken crossbar, lunate sigma, ‘half-h’), and spellings peculiar to Greek such as , and to represent /ps/, /ks/ and /ŋg/, instead of , and respectively, as McDonald (2015: 82–90) and Zair (2013: 221; 2016: 19–24, 141–4) have recently discussed. Persisting Inconsistency of Ambiguous Spellings after 300 The socio-historical context thus credibly accounts for the fact that and , and consequently and , first became 23

24

Contact with, and literacy/bilingualism in the Ionic variety of Greek on the part of the some Oscan speakers of Lucania is also suggested by Potentia 39 (400–300) from Serra di Vaglio, an Oscan inscription (on the basis of the Oscan name and the dating formula in a non-Greek form) written in Ionic, or perhaps in koine Greek if a date around 300 is assumed. It reads: (vac.) ἐπὶ τῆς| νυμμέ|λου ἀρχῆς (vac.), ‘in the term of office of Nummelos’ (Crawford et al. 2011: 1422). SEG 53 1077 (late fourth/early third c.); compare I.Locri 1–37 (except I.Locri 20) for retention of severior secondary long mid-vowels in official inscriptions until the midthird century.

151

Livia Tagliapietra

available to most Oscan writers around 300. However, moving to our second question, the inconsistency observed in using the new signs, despite their convenience in terms of phonemic representation, still needs to be explained. The idea of freedom of choice between the Greek signs as a consequence of literacy being primarily acquired in Greek does not provide a full explanation once the risk of ambiguity involved in the choice of the older spellings is considered. A different explanation would be that use of the older spellings alongside the new ones after 300, in particular of and for /e/ and for /u/, is a conservative trait. This is strongly suggested by comparison with the small number of inscriptions from Messana. As previously mentioned, the authors of these texts, the Mamertini, are supposed to have learnt the Greek alphabet only around the beginning of the third century, as they arrived in the southern regions as mercenaries from an area where the native Oscan alphabet was in use. Recent acquisition of the Greek letters thus implies unfamiliarity with any previous spelling use in the Greek alphabet. In this instance, we observe that these Oscan writers show perfectly coherent, consistent and unambiguous spelling of the vowels in question both across and within the surviving inscriptions: /e/ is only spelled , and so /ɛi/ , /u/ , /o/ , while is used for diphthong /ou/, and /u/ after coronal is written , presumably reflecting influence of the spelling used in the northern areas (Zair 2016: 138–9). Such a level of consistency may plausibly be ascribed to the lack of any ‘memory’ of older practices. We cannot exclude the possibility that these writers’ consistency could be partly determined by their being accustomed to standardised orthographic conventions in their homeland, where a different grapheme was attributed to each vocalic phoneme after 300, provided that the reform had already occurred at the time when they left the northern areas. Even if this were the case, however, it is still remarkable that what these writers learnt in Southern Italy matches with the practices that we observe, although irregularly, in the South Oscan inscriptions after c. 300: this indicates that the Mamertini did not simply choose arbitrarily one of the available Greek spellings for a certain sound and use it systematically for each occurrence, 152

Variant Spellings in Oscan in the Greek Alphabet

which is what we would expect if choices of spellings in the South Oscan areas were completely extemporaneous in the absence of proper literacy in Oscan. If this had been the case, the Mamertini could easily have chosen signs from the Greek inventory different from those chosen by the local Oscan writers: for instance, they could have decided to use the letter for /ɛ/ in general, and not only for the first element of the diphthong /ɛi/. By contrast, the evidence available, despite being very meagre, suggests that they were instructed in a rational manner on the value that the South Oscan writers attributed to these spellings: thus and were learnt as the spellings that needed to be used for /e/ and /u/, while , and were to be employed for /ɛ/, /i/ and /o/ respectively; similarly, was to be used for /ɛi/ and for /ou/. Acquaintance with standardised writing conventions in their place of origin might have further enforced the consistent use of a unique spelling for each of these phonemes. By contrast, it seems entirely plausible that consistency in the use of and for /e/ and /u/ in the rest of the South Oscan inscriptions was delayed and hindered by familiarity with older spellings in the absence of a proper writing reform undertaken across the southern regions similar to what had occurred in Campania and Samnium. As the new spellings and lacked official ratification and promotion from a prestigious centre or institution, despite their convenience, older practices could remain perceived as acceptable, perhaps even prestigious in certain items of lexicon, or in formulaic and other standardised contexts (although the evidence is too scant to confirm this in any absolute way). In such circumstances, it is highly likely that the need to systematically replace traditional spellings was not immediately perceived, and therefore that these continued to be used alongside the new ones. Conclusion Zair’s research on the orthography of the South Oscan inscriptions has reshaped the scenario previously assumed on the basis of Lejeune’s contributions. He has shown that a range of different alternative uses of the Greek letters is attested throughout the corpus of South Oscan inscriptions, and in some cases even within the same text, and that this 153

Livia Tagliapietra

is incompatible with the hypothesis of a standardised Oscan orthography in Lucania and Bruttium. The alternative interpretation of the data provided by Zair explains this degree of variation in terms of different individual choices taken by different Oscan writers, at times inconsistently. In this chapter the concept of individualism formulated by Zair has been further explored and defined; in particular, a systematic contextualisation of individual choices of spelling for the Oscan vowels has been attempted through a comprehensive reconsideration of Oscan phonology, the historical events that brought Oscan and Greek speakers into closer interaction, and the local Greek dialects. In light of this examination, it does not seem that bilingualism without biliteracy can alone explain the orthographic variation across the south Oscan corpus. Rather, an attempt at systematic employment of the Greek signs throughout the corpus can be recognised, although modalities differed from area to area, and possibly even from writer to writer, due to the absence of a unitary convention for writing Oscan in the southern regions. Rational motivation based on the need for disambiguation provides the most likely explanation for the introduction of and for /e/ and /u/ and of for /o/ instead of , as well as for the development of new spellings for the diphthongs /ɛi/ and /ou/ once and started to be used for /e/ and /u/. The consistency with which and are abandoned as representations of the diphthongs after the beginning of the third century strongly suggests that such spellings for /e/ and /u/ began to be employed by Oscan writers around 300 and not earlier, despite Zair’s reservations. Individual preference without such a context would not explain the general disappearance of and for writing the diphthongs in the third century, or the fact that the spellings for /o/ and for /ɛ/ in the diphthong /ɛi/ are not found before c. 300. We also observed that the socio-historical context provides a credible environment for the introduction of new spellings in south Oscan inscriptions around 300. The fact that in most areas Oscan speakers came to live close to the Greeks only in the second half of the fourth century is compatible with the absence of the digraphs and for /e/ and /u/ before then, as it was only at this stage that Oscan writers could 154

Variant Spellings in Oscan in the Greek Alphabet

develop a deeper acquaintance with the Greek orthography on a large scale. We noted, however, that, at some sites which Oscan speakers had been sharing with Greek speakers since the mid-fourth century (particularly Caulonia and Hipponion), there seems to be some delay in the introduction of the digraphs until the early third century; this should probably be connected with the phonology and scribal practices of the local Greeks. Contemporary Greek inscriptions from Locri attest to changes taking place precisely in the early third century under the influence of the koine, which included the more frequent, if not completely new, employment of the digraphs and for close long mid-vowels: such changes therefore conceivably account for the appearance of the same uses in the local Oscan inscriptions at this stage but not earlier. As the new signs for /e/ and /u/ became available, Oscan writers had to find alternative spellings for the diphthongs /ɛi/ and /ou/, previously spelled and . This triggered the adoption of the letter in the spelling for the diphthong /ɛi/, and in the spellings and for the diphthong /ou/; absence of a fixed Oscan orthography in Lucania and Bruttium plausibly accounts for the development of several different disambiguating strategies for spelling /ou/. The adoption of as a disambiguating spelling for /o/ in certain areas can also be explained in the light of increased interaction with the local Greeks in the later fourth century, as this letter does not appear in earlier Oscan inscriptions despite having been in use among the Greeks since the early fourth century. At the same time, the persistence of older, ambiguous spellings may be conceivably attributed to conservatism, as the new signs appearing around 300 lacked the prestige that a unitary orthographic reform would have determined. This seems particularly indicated by the perfect consistency of spelling noticeable in the few inscriptions surviving from Messana, written by Oscan speakers who had adopted Greek letters in the early third century and had no familiarity with any previous usage.

155

CH A P T E R 7

T H E MA M E RT I N I I N M E S S I N A Mobility, Migration and Mercenaries NI CHO LAS Z AI R

Introduction1 The Mamertini are unusual in the ancient world, if not unique, as a group of mercenaries for whom we have both a record in ancient authors of their movements and behaviour (to some extent, and with some variation between authors), and a small number of inscriptions which can be attributed to them partly on linguistic grounds (being written in Oscan in an otherwise Greek-speaking milieu) and partly because some of them explicitly state that they have been erected on behalf of the τωϝτο μαμερτινο ‘the Mamertine people’.2 Our literary sources agree that the Mamertini took control of Messina around 288 BC.3 According to Diodorus Siculus, Cassius Dio and Polybius, they were mercenaries who took the city by force, after having initially been welcomed by the people of Messina. Strabo says only that the Mamertini settled there and subsequently prevailed and got control of the city. According to Festus (reporting Alfius’ history of the First Punic War), the Mamertini came to the aid of the people of Messina and were invited to join the community, leading to the community as a whole changing name and becoming known as Mamertini. The Mamertini were Oscan speakers, we are told (implicitly by 1

2

3

I am grateful to Dott.ssa Donatella Spagnolo, Dott. Agostino Giuliano and the staff of the Museo Regionale Interdisciplinare di Messina for kindly permitting the ‘Greek in Italy’ team to view the inscriptions in the museum. Translations of inscriptions are based on, but sometimes differ from, those of Crawford et al. (2011). The relevant passages are Diodorus Siculus 21.18.1–3, 22.1.2–3, 22.7.4, Strabo 6.2.3, Cassius Dio 9.40.8, Polybius 1.7.1–6, 1.8.1 and Festus in Lindsay (1913: 150).

156

The Mamertini in Messina

Diodorus Siculus, explicitly by Festus). According to Cassius Dio, Polybius and Strabo they were Campanians. Festus says they were originally from Samnium, but settled in a place called Tauricana before moving on to Messina. Tauricana may perhaps be associated with the Tauriani, an Oscan-speaking people of Bruttium, one of whose towns was called Mamertium (Campanile 1993: 601–4; Crawford 2007: 277).4 On the inscriptional side, Crawford et al. (2011) provide seven entries in Oscan for the Mamertini, consisting of part of a coin series with an Oscan legend (Messana 1/nMe 1a, 1b),5 a number of examples of tiles stamped with two different legends (Messana 2/tMe 1),6 an inscription consisting of a name formula painted onto stucco on a limestone block (Messana 6/Me 4) and three dedicatory inscriptions inscribed on limestone blocks (Messana 7/Me 5; Messana 4/Me 1 and 3; Messana 5/Me 2).7 All these date to the third century BC. The purpose of this chapter is to discuss to what extent the mobility of the Mamertini had an effect on their language use, as far as the inscriptional evidence allows us to judge. On the one hand, it is clear that their migration to Sicily, and perhaps their movement through Southern Italy to get there, brought them into contact (or perhaps closer contact) with speakers of Greek, and it has been claimed that the Mamertine inscriptions provide evidence for accommodation to Greek linguistic practices as a staging point in a retreat from using Oscan and a shift to using Greek alone, whose final date is undetermined, although it presumably took place by the first century BC at the very latest (Crawford 2007). With regard to these features, I will be arguing in each case one of two propositions: either that they cannot be reliably demonstrated actually to be present in the inscriptions, or 4

5

6

7

On Oscan in Sicily in general, including discussion of the Mamertini at Messina, see Clackson (2012). All Oscan texts are referred to by the names used in Crawford et al. (2011) and Rix (2002), in that order, except where texts are not given in Rix. Unless otherwise stated, the text is that of Crawford et al. Crawford et al. also include a Greek stamp as Messana 3, with five exemplars in Messina. It reads μαμερτινων and is dated to 250–150 BC. According to Crawford et al. it features alpha with straight cross-bar, but the two stamps we saw with this text unbroken have broken-bar alpha. The other texts were broken: one reads] ε ̣ρτινων, and one μ[. ̣ Messana 4 and Messana 5 show the same text. As we shall see, a possible third copy of this text has been claimed to have once existed.

157

Nicholas Zair

that the feature is not necessarily the result of Greek influence on the Oscan of Messina in particular, as opposed to influence on Oscan more generally. Instead, I will suggest that there is evidence of another sort to be found in the use of spelling and language in the inscriptions, which fits the claims of the historians that the Mamertini were originally from Campania or Samnium rather than from Oscan-speaking communities further south. Possible Greek Features in the Inscriptions One of the Messina inscriptions has been claimed to contain two Greek features. This is Messana 7/Me 5, on which Crawford et al. (2011: 1522) observe ‘the text uses the Greek system of personal name + personal name of the father in the genitive’,8 which is to be compared to the usual Oscan sequence of praenomen + gentilicium + praenomen of the father in the genitive, as demonstrated, for example, by n]ịumsis heírennis niumsieís (Nola 3/Cm 6) ‘Niumsis Herennis son of Niumsis’. This claim, as we shall see, only stands if it can be demonstrated that there was only a single name before the father’s name in the inscription. This is almost certain to be the case if the reading implied by the drawing of the inscription in its earliest publication, by Orsi (1916: 195–6), is correct.9 This is the reading given by Vetter (1953: n.197a) and by Marinetti and Prosdocimi (1988: 54 n.A71), that is, ]α ̣ς στεννιηις | [α]π ̣πελλουνηις, which would mean ‘-as son of Stennis; of Apollo’. This is because no Oscan gentilicium is known to end in -as, while Maras is a very common praenomen, as pointed out by Nazzari in Orsi’s text (Trebas would also be possible). However, subsequent scholars who have autopsied the inscription are much less certain about the existence of the initial letter. Crawford et al. (2011: 1521–2) read [2–3]ς στεννιηις | [α] π ̣πελλουνηις, and add the comment ‘[t]here is no trace of any letter before the first [sigma]’. This reading matches that of Parlangèli (1956), who had presumably seen the inscription, and prints μαρα]ς 8

9

Similarly Crawford (2007: 274). This possibility had already been noted by La Regina (2002: 69). In the text, Orsi gives the reading as if it were Latin, as . . .. .as Stenii filius | . . . .. Apollinis, which, however, confirms that he saw the first letter as alpha.

158

The Mamertini in Messina

στεννιηις | [απ]πελλουνηις. Tagliamonte (1994: 260; followed by Mastelloni 2005)10 prints -?-]-ς στεννιηις | [-?- α]π ̣πελλουνηις and observes: ‘[s]ul margine di frattura sinistro, all’ altezza di lin. 1, sembrerebbe di intravedere il resto di una lettera, interpretato dai più come un’ asta obliqua, forse pertinente ad un’ α. Lo stato della pietra, la cui superficie in quel punto risulta completamente abrasa, non consente di essere certo di una tale lettura, giacchè potrebbe trattarsi semplicemente di una scheggiatura presente sulla pietra stessa. Non si esclude pertanto l’eventualità di una diversa lettura del resto di lettera, ad es. ι.’ (Tagliamonte 1994: 261).11

One might assume that Orsi, as the earliest publisher of the inscription, saw a letter at the start of the first line which has subsequently been further damaged.12 However, I and other members of the ‘Greek in Italy’ project saw this inscription on 20 September 2014, and we concluded that there is the possible right-hand stroke of an alpha before the sigma of the first line. This looks identical to the stroke in Orsi’s drawing. But it does not come down to the same level as the other letters of the line, and consequently we are not certain that it is the stroke of a letter rather than damage to the stone, as Tagliamonte suggests. If it is a letter, it is also possible that the stroke is the lower diagonal of a kappa, or even the diagonal of a nu. The latter is less likely, since there is no sign of the right-hand vertical, but the stone has been damaged just at the spot where the vertical would be expected, with the damaged section lying lower against the face of the stone than the following sigma, so it is not impossible that the vertical ran along this 10

11

12

This is also the text as given by Rix (2002), who, however, refers only to Orsi and Parlangèli’s (different) readings. ‘On the edge of the left-hand break, at the height of line 1, the remains of another letter seem to be visible, generally interpreted as a diagonal line, perhaps of an . The state of the stone, whose face at this point has been completely abraded, does not allow for certainty about this letter, as it could be the result of chipping of the stone. Therefore the possibility of a different reading of the rest of the letter, for example, as , cannot be excluded’. The same caution is expressed by La Regina (2002: 68): ‘[n]on è sicuro il riconoscimento della lettera di cui resta solo traccia all’ inizio della prima linea’. (‘The letter at the beginning of the first line, of which only a trace remains, cannot be securely identified’). A possibility which is further bolstered by the fact that a small part of the stone containing the initial π ̣ of line 2 has now also broken off (though it is still preserved with the main block), presumably very recently, since the break is not mentioned by the sources listed above.

159

Nicholas Zair

damaged part. If the name in fact finished in -ις, -ιες, -ες or -νς it could be a gentilicium rather than a praenomen.13 Given that Crawford et al. do not accept a letter before the sigma, their analysis of the inscription as including only a praenomen before the father’s praenomen is presumably due to the absence of room for further letters to the left, which they reckon to be only enough for two to three. This assumes that the first line starts no, or not much, further to the left than the second line, where there are only one-and-a-half letters missing at the beginning (i.e. [α]π ̣). This is a reasonable assumption, given that both lines end at the same point. Although there are several examples of Oscan inscriptions with significantly shorter second lines than their first, for example, Pompei 11/Po 7 and Aeclanum 3/Hi 4, generally the second line starts directly below the first line, rather than being indented relative to the first line and finishing directly below the end point of the first line, as would be required here. However, against the assumption that all we have missing at the start of the first line is a few letters is the fact that there is very little room there for even a praenomen: an abbreviation is possible, but I know of only two praenominal abbreviations which end in -s. These are ves(is) (Nuceria Alfaterna 9), which contains a name which is not attested anywhere else and which is found on a tilestamp, and may therefore be ad hoc; and τς (Caulonia 5) which, if the reading is correct, may stand for Trebas, but may not be Oscan at all, and is again on a tile stamp. There are a number of four-letter praenomina, such as μαις (e.g. Numistro 1/Lu 4), αϝες (Petelia 2), úvis (e.g. Nuceria Alfaterna 5/Cm 35) and οπις (e.g. Cumae 17), but even these would struggle to fit in the same space as the [α]π ̣ in the line below. It cannot be ruled out that there was another block to the left of the one we have, possibly including more text also on line 2, which included a praenomen and the first part of the gentilicium, in which case we would have the standard Oscan name formula in this inscription (as observed by La Regina 2002: 69). It has also been suggested, by Clackson (2012: 140), that this inscription may show Greek influence in the use of the genitive of 13

For gentilicia in /-es/ compare ιμες (Crimisa 3/Lu 44), pettíẹ[i]s (Capua 26/Cp 9) and see Zair (2016: 49). For a gentilicium in -ns compare aadirans (Pompei 24/Po 3).

160

The Mamertini in Messina

the god’s name, rather than the dative, which is normal in Oscan inscriptions, and found, for example, in Messana 4/Me 1 and Messana 5/Me 2, to be discussed directly below. He compares cases such as Διὸς Λυκα(ίου) ‘of Zeus Lykaios’ (IG XIV 597, Sicily, perhaps Thermai Himeraiai) and Ἀπόλōνος Λυκ(είου) ἐμί ‘I am of Apollo Lykeios’ (IG XIV 647, Metapontum). But, although it is true that the more usual case for a god as recipient of a dedication in Oscan is the dative, there are a not inconsiderable number of instances of the genitive, in the form of iúveís lúvfreís (Histonium 5/Fr 5, bronze weight in the shape of a head and shoulders) ‘of Jupiter Liber’, herentateis (Histonium 8, stamped on a bowl) ‘of Herentas’, fatuveís (Aeclanum 1/Hi 6, a limestone block, probably the central block of an altar) ‘of Fatuus’; . . . διωϝηις (Potentia 9/Lu 6, two limestone blocks) ‘of Jupiter’, . . . διωϝιιας. διομανα[ς] (Potentia 10/Lu 7, limestone block) ‘of the lady of Jupiter, the mistress’.14 With the otherwise implied verb ‘to be’ we have herettates: súm (Histonium 7/Fr 4, incised and painted onto a bronze key) and herentateís. súm (Herculaneum 1/Cm 10, engraved on an altar) ‘I am of Herentas’. It is possible that all these instances are due to Greek influence on Oscan, but if so, there is, at any rate, nothing special about the usage of the genitive of the god’s name at Messina, and it could as well reflect Greek influence on the Mamertines in Campania or Samnium as at Messina itself. The final suggestion for Greek influence reflects the complicated history of the inscriptions from Messina and their publication. The same inscription is found twice, on two different stones, the first being Messana 4/Me 1, which is almost complete (the edges have been lost between the earliest reports of the inscription and the present day), and the second being Messana 5/Me 2, of which the right-hand side exists but has been cut into two pieces. Messana 4/ Me 1 is given by Crawford et al. (2011), working on the basis of autopsy and earlier reading, as στενις καλινις σταττιηις | μαρας πομπτιες νιυμσδιηις | μεδδειξ ουπσενς | εινειμ τωϝτο μαμερτινο | ̣ 14

Cases of the genitive can also be found in zoveṣ ‘of Jupiter’ (Bantia 2/Lu 38, limestone base; broken to the right) and ϝεζεις (Caulonia 2, limestone block; could carry on to second block) ‘of Venus’, where the context is, however, uncertain due to the loss of the rest of the text (if it existed).

161

Nicholas Zair

αππελλουνηι σακορο ‘Stenis Kalinis son of Stenis, Maras Pompties, son of Niumsdis, magistrates, made (this), and the Mamertine people; sacred to Apollo’.15 The left-hand block of Messana 5/Me 2 was previously read by Reina (1658: 230), who prints ΣΤΕ⋮Ω. ΑΝ̣ . . |ΜΑ . . Σ . . Μ . | ΜΕ . . . . | ΕΙΝΕΙΜΤΩΕΤ | ΑΠΠΕΛΛ. On the basis of this reading and his own autopsy of the right-hand blocks, Parlangèli (1956) reads στε[ν]ω[ις κ]αλι[νις στ]αττιηις̣ | μα[ρας] [πο]μ[πτι]ε ̣ς νιυμσδιηις | με[δδιξ ουπ]σενς | εινειμ τωϝτ[ο μ]αμερτινο | αππελλ[ου]νηι σακορο. Crawford et al. (2011: 1517–18) print στε [κ]α[ινις σ]ταττιηις | μα[ρα]ς [πο]μ[πτιες] νιυμσδιηις | με[δδειξ ο]υ[πσ]ενς | εινειμ τωτ[ο] μαμερτινο | αππελλ[ο]υ ̣νηι σακορο, with the ̣ comment on line 1 ‘νις taken for ιω by Reina’ (they follow Parlangèli in also correcting, without comment, Reina’s ΑΝ for ΑΛ and Ε for Ϝ in line 4). A possible third version of the inscription was published by Buonfiglio Costanzo (1613: 134–5, second edition 1739: 2.134–5)16 as follows: ΣΤΕΝΟΙΣ ΚΑΛΕΙΝΙΣ | ΣΤΑΤΤΙΗΗΣ | ΜΑΡΑΣ ΓΟΜΓΤΟΙΕΣ | ΝΙΗΜΣ ΔΙΕΙΙΣ ΜΕΔΔΕΙ ΤΟΥ ΓΣΕΝΣ | ΗΝΟΙΜ ΤΩΕ | ΤΟ ΜΑΜΕΡΤΙΝΟ ΑΓΓΕΛΛΟ | ΥΝΗΙΣ ΑΚΟΡΟ. Use of Γ for Π is probably a mistake of the printer (Parlangèli 1956: 31 n. 4). ΚΛΛΕΙΝΙΣ in Parlangèli is an error for ΚΑΛΕΙΝΙΣ in the original. Based on the second edition, Parlangèli (1956; followed by Rix 2002 as Me 3) printed στενοις καλεινις | σταττιηις | μαρας πομπτοιες | νιημσδιειις | μεδδειξ ουπσενς | ηνοιμ τωϝ|το μαμερτινο αππελλο|υνηι σακορο. On the basis of his version of this supposed third inscription, now lost, Parlangèli argues that the variation in the spelling of vowels in στενις/στενοις, καλινις/καλεινις, πομπτιες/πομπτοιες, νιυμσδιηις/ νιημσδιειις, εινειμ/ηνοιμ is due to the fact that the writer of the inscription was a Greek speaker for whom the sounds represented by these letters in Greek were by this time homophonous, or beginning to become homophonous (Parlangèli 1956: 37). It is hard to know exactly what dialect of Greek the putative writer of the inscription in Messina in the third century BC was 15 16

The underlined portions are those which are now missing. I have seen only the second edition. Crawford (2006: 522) prints the version from the first edition with ΔΙΕΠΣ in the fourth line, but this may be a mistake.

162

The Mamertini in Messina

speaking. If it were Koine, it is indeed likely that the Classical (Attic) /eː/ and /iː/ had merged in terms of vowel quality (leading to confusion of and ), and likewise /y/ and /i/ (leading to confusion of and ). A radical possibility is that they spoke a version of the Koine similar to that proposed by Teodorsson (1978: 94–8; see also Teodorsson 1974: 286–99), who claims that by 350 BC an innovative variety of Attic had already undergone falling together of /eː/, /y/, and /ɛː/ as /i/, resulting in spellings of all of these originally distinct phonemes as , , or . In this variety, the original diphthong /oi/ had become /y/ but had not merged with /i/. According to Teodorsson, this variety existed alongside a more prestigious conservative variety, to which it lost ground from the third century BC onwards, and in which only /eː/ had undergone merger (with /iː/ and /ɛː/ in different contexts). For a minority of speakers of Koine Greek in Egypt around 250 BC he posits a very similar system to the innovative Attic variety, again without merger of old /oi/ and /iː/ (Teodorsson 1977: 251–6). This picture is largely compatible with that of Mayser and Schmoll (1970: 49–54, 60–5, 80–2, 87, 89–91) in that they identify frequent confusion between and , and between and , alongside occasional confusion of and , and, especially in the second century onwards, and ; but they are doubtful of the evidence for confusion of or with . Horrocks (2010: 168–9; see also 115, 117–20) emphasises the relative infrequency of the ~ ~ ~ ~ interchanges and argues that overlap took place between /iː/ and /ɛː/ and between /y/ and /i/ originally only in specific phonetic environments and may also reflect the influence of second-language speakers of Greek. He dates the final loss of these distinctions to the ninth or tenth centuries AD (in educated speakers). On this basis, therefore, the variation between and (in καλινις/καλεινις), and (in εινειμ/ηνοιμ and νιυμσδιηις/ νιημσδιειις), and and (in νιυμσδιηις/νιημσδιειις) could have an explanation in Koine developments of the period. But even if we take the most radical view of the Koine sound system of the third century, there is no suggestion that /oi/, even if it had already become /y/, had fallen together with /i/, which would be 163

Nicholas Zair

required to explain the , and variation in στενις/ στενοις, πομπτιες/πομπτοιες, εινειμ/ηνοιμ.17 This linguistic point is to be added to the arguments made by Crawford (2006) in favour of the view that there was no third version of the inscription, and that the very different readings in the third version are instead the result of misreading and misinterpretation by Buonfiglio Costanzo of Messana 4. To sum up this section, therefore, none of the features of the inscriptions from Messina which have been claimed to show Greek influence can be relied upon. In all likelihood the missing ‘third’ inscription, with the same content as, but very different spelling from, Messana 4/Me 1 and Messana 5/Me 2 is nonexistent. Given the loss of the left-hand side of Messana 7/Me 5, it is possible that the extant name formula was of the standard, longer, Oscan type rather than the Greek type with only a single personal name and father’s name. The genitive of the god’s name in Messana 7/Me 5 may reflect Greek influence on Oscan, but if so this is not a unique occurrence in Messina, since we find other instances of the genitive also in inscriptions from the mainland, including Campania and Samnium, the ultimate origin of the Mamertini. Features Characteristic of the Oscan of Campania and Samnium I have previously argued that the inscriptions from Messina show a number of features that are characteristic of the spelling of Oscan in the Oscan, rather than the Greek, alphabet (Zair 2016: 136–41). The most obvious of these is νιυμσδιηις in Messana 4/Me 1 and Messana 5/Me 2, where represents the result of an original *u after a coronal (d, n, s, t). In Campania and Samnium (henceforth ‘the North’) *u in this environment is almost always spelt , probably representing a development to [ju], while in Lucania and Bruttium (henceforth ‘the South’) *u after a coronal became [y], which is spelt in the Greek alphabet in use in these areas with , or (for further details see Zair 2014 and 17

Horrocks (2010: 163) dates loss of lip-rounding in /y/ to the Byzantine period.

164

The Mamertini in Messina

2016: 74–9). The spelling in Messana 4/Me 1 and Messana 5/ Me 2 of *u after a coronal therefore diverges from the spelling used in all other inscriptions in the Greek alphabet, but finds an exact parallel in that commonly found in the Oscan alphabet in Campania and Samnium. Presumably the Mamertini also shared with the Oscan speakers of these areas the dialectal treatment of *u after a coronal as [ju] rather than the [y] found further south. Another feature which is reminiscent of the Oscan alphabet is the spelling of the sequence /ms/ in νιυμσδιηις as . In all definite instances of this sequence in the inscriptions written in the Greek alphabet from outside Sicily, we find the spelling which represents the insertion of subphonemic [p] in the sequence /ms/, exactly as in English hamster [hæmpstə]. Since this is such a natural change, it presumably took place also in the Oscan of Campania and Samnium, but here we always find spellings like niumsis (Cumae 8.2/Cm 14), which do not write the [p], matching what we find in Messina. Lastly, the inscriptions from Messina also show a couple of features which are occasionally found in other Greek-alphabet inscriptions, but which are assumed also to be due to the influence of the Oscan alphabet, since they do not reflect standard Greek spellings (these tend to be found in Lucania, nearer to the Oscanalphabet-using areas, rather than Bruttium, which is further away). One of these is the use of double letters to represent long vowels, as shown by the tile stamp πααπ ̣[-?-] (Messana 8/tMe 2) ‘Pāpis’. This usage is not found elsewhere in the Greek alphabet before c. 200 BC, so it is likely to reflect direct knowledge of the conventions of the Oscan alphabet in Messina. The other is the use of the sequence and rather than and to write /ks/ and /ps/ in [μ]αμερεκς ̣ (Messana 6/Me 4) ‘Mamerex’ and ουπσενς (Messana 4/Me 1) ‘they made’ (but note that is used in μεδδειξ Messana 4/Me 1). ̣ Another possible case of a linguistic feature which may reflect the origins of the Mamertini in the North is the use of the main verb ουπσενς (Messana 4/Me 1), ο]υ[πσ]ενς (Messana 5/Me 2) ‘they made’. In building and dedicatory inscriptions of this type, which commemorate the magistrates who were in charge of the building or setting up of the edifice, altar and so on, there are a rather limited 165

Nicholas Zair

number of formulas attested in inscriptions from Lucania and Bruttium. The most common formula is the earliest attested and is found predominantly in the fourth and third centuries BC (although there is a single example from the second century in Potentia 3/Lu 8). This simply omits a verb altogether, as in Numistro 1/Lu 4, a yellow sandstone block dated to between 300 and 275 BC which reads μαις αρρι|ες σουϝεν | μεδδικεν ‘Mais Arries in his term of office as meddix (built this)’ (other examples are Crimisa 1/Lu 23, Crimisa 2/Lu 24 and Potentia 39, which is in Greek). The next most common formula, found from the third century onwards, uses the verb α(τ)fαμα-: [−?−]+[−?−] | [−?−]υ ̣ϝι εκς λεκ(?) σε[νατηις ταν]|[γιν]ο ατfαματτενς (Cosilinum 1/Lu 3) ‘[−?−] to Jupiter (?); thus ??? by [decree of the] se[nate] they had (this) made’. λωϝκις. νανονις. σπελληις | κϝαιστορ. σενατηις | τανγινοδ. αfααματεδ | διωϝηις (Potentia 9/Lu 6) ‘Loukis Nanunis, son of Spells, quaestor, by decree of the senate had (this) made; of Jupiter’. λωϝκις. νανονις. σπελλ[ηις] | κϝαιστορ. σενατηις | τανγινοδ. αfααμα[τεδ] | διωϝιιας. διομανα[ς] (Potentia 10/Lu 7). ‘Loukis Nanunis, son of Spells, quaestor, by decree of the senate had (this) made; of the lady of Jupiter, the mistress’. . . . σεγονω. αιζνιω. ρεγο(μ) | (ο)πσανω. ειν(ειμ). σταβαλανο | σενατηις. (τ)ανγινοτ. αfαματετ | ειζιδομ. πρωfατεδ . . . (Potentia 1/Lu 5) ‘. . . by decree of the senate, had bronze statues of the kings made and fixed in place; the same person passed the work as completed . . . ’. The last formula is found in a single inscription of around 150 BC, and uses the verb δεδενς ‘they gave’: [-?-]ς διριος. μαραδ(ηις). ν[-?-] | [-?- σενα]τηις τανγινοδ τρειβ[ωμ εκακ] | [ωπσανναμ] δεδενς μαραhις κε[πιες πρωfαττεδ] (Atina Lucana 1/ Lu 2) ‘-s Diris and -s Diris, sons of Maras, N[-?-], by decree of the senate gave [this] building [to be constructed], Marais Ke[pis passed (it) as completed]’.

There are clear differences of phraseology between the South and North parts of the Oscan-speaking area in inscriptions of this type. Thus, the verb aamanaf-, which is frequent in inscriptions of this type in Campania and Samnium, does not appear in Lucania and Bruttium. An example is Pompei 11/Po 7: ni(umsis). trebiis. tr(ebieís). med(dís). túv(tiks) | aamanaffed ‘Niumsis Trebiis son of Trebis, magistrate of the people, had (this) made’ (it is also found in Pompei 8/Po 6, Pompei 20/Po 9, Pompei 21/Po 4, Pompei 23/Po 14, Teruentum 15/Sa 11, Teruentum 17/Sa 12, Teruentum 21/Sa 8, Teruentum 33/Sa 9 and Teruentum 36/Sa 2). 166

The Mamertini in Messina

Neither, apart from at Messina, does ops- appear as the main verb, which is found twice at Pompeii: . . . ekass. ví|ass. íní(m). vía(m). íúviia(m). dekkvia|rím. medíkeís. púmpaiianeís | serevkidimaden. uupsens . . . (Pompei 13/Po 1) ‘ . . . They built these roads and the Jovian road and the dekkviaris road under the authority of the Pompeian magistrate . . . ’.18 m(ina)z. avdiis. klí(pieís).| dekis. seppiis. úpf(alleís). | kvaízstur. upsens. (Pompei 19/Po 8) ‘Minats Audiis son of Klepis and Dekis Seppiis son of Opfal, quaestors, built this’.

Of course, the absence of these types from Lucania and Bruttium may simply be a reflection of the smaller amount of material from the South, so that we do not have such a range of formulas preserved, rather than a difference in the terminology used. And indeed we do find formulas of the type found in Atina Lucana 1/Lu 2, using the verb ‘to give’ with a gerundive, elsewhere; for example, in Pompei 24/Po 3, which reads . . . v(ibis). viínikiís. m(a)r(aheís). kvaísstur. púmp|aiians. trííbúm. ekak. kúmben|nieís. tanginud. úpsannam | deded . . . ‘Vībis Vīnikies son of Maras, Pompeian quaestor, gave this building to be made by decision of the assembly’.19 Nor is the verb ops- entirely unknown in the official inscriptions from Lucania, since it is found in the gerundive in (ο) πσανω in Potentia 1/Lu 5 (from the second century BC). However, there is one instance where a difference between the phraseology of the official formulas used in the South and North parts of the Oscan-speaking area is probably meaningful, since the same basic verb is used in both areas, but with the addition of a preverb in the South: α(τ)fαμα-, meaning something like ‘give the order (that something be made)’, is only found in Lucania, while in Campania and Samnium the same verb without the preverb is used (faama-, in Aeclanum 4/Hi 1 and Teruentum 12/Sa 13. It is also found, but in a different context, in Pompei 2/Po 34, Pompei 3/Po 35 and Pompei 4/Po 36). Consequently, it is possible that the absence of ops- as a main verb in the South is not a coincidence, and that its usage at Messina is part of the broader 18

19

The origin and meaning of dekkviarím, an adjective agreeing with vía(m) ‘road’, is uncertain. It is presumably the name of the road (Untermann 2000: 164). The same construction with ‘give’ is also occasionally found in Latin, for example, . . . d(e) s(enatus) s(ententia) portas faciundas dederunt (CIL I2 1563).

167

Nicholas Zair

picture of language and writing there which is more akin to that used in Campania and Samnium than further South. The final way in which the Mamertini may have used language in a manner more associated with Oscan speakers of the North rather than the South is in respect to coin legends. There are twenty-three series of bronze coins of the Mamertini at Messina, on which the legend is usually in Greek, reading μαμερτινων (Messana 1) ‘of the Mamertini’. On part of a single series, dated to around 225 BC, the legend is in Oscan, reading μαμερτινουμ (Messana 1/nMe 1a-b ) ‘of the Mamertini’. As stressed by McDonald (2015: 204–11), the minting of coins on behalf of Oscan-speaking communities was primarily carried out by Greekspeaking craftsmen, and in the South the legends are almost uniformly in Greek rather than Oscan, where it is possible to tell the difference (she lists another nine series with Oscan ethnics in Greek morphology). The single exception is the case of the bronze coinage datable to 207–204 BC bearing the Oscan legend λουκανομ (Lucani 1/nLu 1) ‘of the Lucanians’ beside Greek λυκιανων and indeterminate λουκα (Lucani 1), which is a group apparently struck by the Brettii for the Lucanians who were fighting alongside them on behalf of Hannibal (Crawford et al. 2011: 48–9). It is possible that the emergence of the Oscan language on this coinage somehow reflects the rather atypical situation for its production. At any rate, the Mamertine coinage, being earlier than the Lucanian, is the first instance in the South of Oscan language appearing on coins. Further north, however, although Greek legends are still common on the coinage of Oscan-speaking communities, we have significantly more Oscan legends on coins from at least the end of the fifth century and hence earlier than the Mamertine issue, such as Campania Coinage 1, Campania Coinage 2, Campania Coinage 4/nSa 4, Fenserni 1 Coinage, Teanum Sidicinum 1 Coinage/nSi 1a, 1b, 1 c, Cubulteria 1 Coinage/nSa 2a, Phistelia 1 Coinage and Aquilonia 1 Coinage/nSa 6.20 Once again, therefore, the use of Oscan on the coinage of the Mamertini fits in better with 20

There is also coinage on which the legend is grammatically indeterminate between Greek and Oscan but which uses the Oscan alphabet.

168

The Mamertini in Messina

the practices of Oscan speakers in the North rather than in the South. Conclusion The relocation of the Mamertini from Campania (or Samnium) to Sicily led to the use of the Greek alphabet by the group as almost the only medium for writing the Oscan language in Messina.21 However, it did not immediately result in the abandoning of Oscan as a language, and we find it used for a small number of official and unofficial inscriptions. The orthography and official formulas of these inscriptions, along with the use of Oscan as well as Greek on coin legends, most closely match that of the communities in Campania and Samnium. It seems unlikely that this usage was intended as a means of marking out the Mamertini as Oscan speakers of Campania or Samnium, as opposed to Lucania or Bruttium. The spellings reminiscent of the Oscan alphabet inscriptions seem to represent a natural carrying over of orthographic rules from an old alphabet when a new one is adopted, rather than any explicit rejection of habits in the South, of which the Mamertini may not have been aware (even if they spent some time among the Tauriani in Bruttium), and which are in any case far less standardised than in the North. The same goes for the formulaic language of the Mamertini. The brief addition of coinage bearing an Oscan-language inscription to a series otherwise showing Greek may have been driven by some kind of desire to assert an Oscan-speaking identity, especially since, if the dating is correct, it appears some sixty years after the Mamertini arrived in Messina, but we cannot know the precise details. This kind of assertion of identity has been suggested to lie behind the fact that the inscription of Messana 6/Me 4 is painted onto stucco. According to Crawford (2007: 274; Crawford et al. 2011: 1520) the stucco is used to replace an original Greeklanguage text with the present Oscan-language one. A parallel for this would be the similar case of Latin replacing Oscan in 21

There is one instance of the Oscan-alphabet letter on Mamertine coinage from 280–278 BC (Crawford et al. 2011: 1511 and Särstrom 1940: 70–3, plate XIV.105–10. Series V, group C).

169

Nicholas Zair

Pompei 1/Po 54, but there is no evidence of the supposed Greek inscription in this case, and the stucco may have been used to allow writing on the very rough stone; care was evidently taken in the writing of this inscription, which is centred, painted red and well executed. As stressed at the beginning of this chapter, the combination of historical and inscriptional evidence make the Mamertini an excellent case study for thinking about migration and language contact in the ancient Mediterranean. The writers of the Mamertine inscriptions appear to have known the Oscan alphabet, as demonstrated indirectly by the spelling rules they maintained from it and directly by the use of on an early coin (mentioned in note 21). However, they made a positive choice to adopt the Greek alphabet, and seem to have been content to be represented by Greek legends on their coinage, apart from the brief series with μαμερτινουμ some decades after their initial take-over of Messina. The use of the Greek alphabet might be seen as an accommodation to local habits, although the content of the inscriptions would have been largely incomprehensible to non–Oscan speakers (who might, however, have recognised the names of the magistrates and the god in the dedicatory inscriptions). On the other hand, they continued to use Oscan, for a little while at least, for both official and private inscriptions, and there is no clear evidence for the adoption of Greek linguistic habits with regard to onomastics, orthography or case usage in these inscriptions. Instead there is some evidence for the continuation of specifically Campanian or Samnite features in the inscriptions. Pending any further inscriptional discoveries, the evidence suggests that the Mamertini maintained for some time habits of formulaic language and orthography of the sort we might expect given their origins in Campania or Samnium – perhaps for a few decades or a couple of generations. There is no evidence for linguistic changes to their Oscan resulting from contact with Greek speakers, as opposed to the adoption of the Greek alphabet. Instead, presumably, because they settled in Messina, a fairly rapid switch to Greek took place, with the result that Oscan died out among the Mamertini, at least as a written language, by the end of the third century BC. 170

C HA P T E R 8

M I G R AT I O N , I D E N T I T Y, A N D M ULTI L I N G U A L I S M I N L ATE HE LL ENISTIC D E LOS FRAN CE SCO R OVAI

But if you have the temple of far-shooting Apollo, all men will bring you hecatombs and gather here, and incessant savour of rich sacrifice will always arise, and you will feed those who dwell in you from the hand of strangers. (Hom. Hymn 3 to Delian Apollo, 56–60; trans. Hugh G. Evelyn-White)

Introduction In the aftermath of Rome’s defeat of Macedonia at the Battle of Pydna, which established the Republic as a major power in the eastern Mediterranean, Hellenistic Delos lost its formal independence (it was under the patronage of the Antigonid monarchy) and was placed by Rome under the rule of Athens in 166 BC.1 The new administration expelled the Delians, who were replaced by Athenian settlers, and sent an ἐπιμελητής (‘supervisor’) and other magistrates from the metropolis to govern the island (Durrbach 1921: 133ff; Roussel 1987: 1–18). By that time Delos, although it was politically insignificant, nonetheless flourished as a trading centre after the decision of the Senate to make it a duty-free port through a grant of ἀτέλεια ‘exemption from taxes’. In addition to this privilege, to the protection offered by the centuries-old sanctuary of Apollo, and to its good geographical location (cf. Strabo 10.5.4: ἐν καλῷ γὰρ κεῖται τοῖς ἐκ τῆς Ἰταλίας καὶ τῆς Ἑλλάδος εἰς τὴν Ἀσίαν πλέουσιν ‘for it is happily situated for those who are sailing from Italy and Greece to Asia’), in the second half of the second century BC other important developments gave a decisive boost to its economic prosperity. These were the 1

For a general overview of the political, economic, social, and religious background during the so-called Period of the Independence (314–166 BC), see Constantakopoulou (2017).

171

Francesco Rovai

destruction of the rival port of Corinth (146 BC), the crisis of the Seleucid kingdom, and the creation of the province of Asia (129 BC). Thus, the Delian emporion functioned as a core hub for the Mediterranean markets, in which the buying and selling of slaves, oil, wine, luxury goods, and so on, took place – alongside the financial activities of many bankers (for a recent overview and bibliography on the manufacturing and commercial activities of the inhabitants, see Zarmakoupi (2013)). Such an economic development generated an unprecedented demographic growth2 due to the large-scale immigration of Athenians and other Greeks, Romans and other Italian negotiatores (‘businessmen’), and also Egyptians, Phoenicians, Syrians, and others, all streaming into Delos for economic purposes (see the next section). The fortune of the island quickly collapsed by the first half of the first century BC as a consequence of the Mithridatic wars: in 88 BC it was sacked by Menophaneses, the general of Mithridates, for siding with Rome (20,000 of the island’s inhabitants were killed, according to Appian Mith. 5.28), and in 69 BC the pirates of Athenodorus, an ally of Mithridates, led most of the inhabitants away as slaves. Many epigraphic records and much archaeological evidence have survived from the prosperous period between 166 BC and 69 BC, investigation of which may help us to assess issues of multilingualism, multiculturalism, and mobility in the ancient world. The Cultural and Linguistic Environment of the Island The cosmopolitanism of the Delian population between the second and first centuries BC is well known to historians, archaeologists, and linguists (among many others: Hatzfeld (1912); Couilloud (1974); Durrbach (1921: 113–236); Roussel (1987: 87–95); Adams (2003a: 642–5); Hasenohr (2007); and Poccetti (2016)), and it is possible to get an impression of the real scale of this 2

The population has been estimated at approximately 20,000–30,000 individuals by Roussel (1931: 443), 25,000 by Tréheux (1952: 582), and 15,000 by Zarmakoupi (2013). The last figure is probably the most likely, but the usual caveats apply in all cases.

172

Multilingualism in Late Hellenistic Delos

phenomenon from Table 8.1, which gives the place of origin of more than 1,600 individuals that are known from both public epigraphy (accounts of the temples, dedications, catalogues of donors, etc.) and funerary inscriptions.3 Within this heterogeneous population coming from the Central and Eastern Mediterranean basin, three primary ethnic, linguistic, and cultural components can be easily individuated: Italian, Greek, and Semitic. A few more words are due about the last group, whose numerical consistency is probably underestimated in Table 8.1, especially within public inscriptions. This is due to the fact that, whereas the Roman onomastic formula is unambiguous, only a very few Semitic names are undoubtedly characterised as such (e.g. 2611: Μαγγάβων; 2365: Ζαβδίων; 2619: Σαββίων).4 In many cases, instead, onomastics may be deceptive because, in a Hellenistic environment like Delos, people from the Eastern Mediterranean adopted Greek names (Couilloud 1974: 310); for example, Γοργίας, which is a rough transliteration of the Semitic name Table 8.1 The Delian population in the public and funerary inscriptions

Place of origin

Public inscriptions

Funerary inscriptions

Total

Italy Athens Levant (Syria, Egypt, Cyprus) Asia Minor and Black Sea Aegean islands Greece and Macedonia Arabia Mesopotamia and Media Cyrenaica

586 421 234 69 72 10 5 1 1

59 23 99 22 13 3 0 2 1

645 444 333 91 85 13 5 3 2

(Adapted from Couilloud 1974: 308)

3

4

Only those individuals are included whose provenance is overtly indicated in the onomastic formula. Unless differently specified, inscriptions are from ID. The letter C followed by a number refers to inscriptions drawn from Couilloud’s (1974) edition of the epitaphs.

173

Francesco Rovai

Gerges (Roussel 1987: 91); Ἡλιόδωρος and Βασιλείδες, which approximately translate the Phoenician names ‛bdšmš (AbdShemesh) ‘Servant of the Sun’ and Mlkb‛l ‘Baal is king’ (Masson 1969: 683, 691); Στράτων and Φιλόστρατος, for their assonance with the theophoric names embedding the name of Astarte (Masson 1969: 692–3). Thus, on closer scrutiny, ‘[l]a population d’origine orientale était certainement aussi nombreuse à Délos que la population italienne’ (Couilloud 1974: 310). The Phoenicians in particular are well represented (Hasenohr (2007: 77–80), and references therein): many citizens of Berytus, Tyre, Sidon, Ascalon, and other minor cities of Phoenicia are known from both official inscriptions and gravestones;5 they were organised in two public associations whose status was religious and professional at the same time, ἡ σύνοδος τῶν Τυρίων Ἡρακλειστῶν ἐμπόρων καὶ ναυκλήρων ‘the association of Tyrian Herakleistai, merchants and shipowners’ (cf. 1519.35–6, 40–1, 49–50 and 61–4: 149 BC) and τὸ κοινὸν Βηρυτίων Ποσειδωνιαστῶν ἐμπόρων καὶ ναυκλήρων καὶ ἐγδοχέων ‘the community of Berytian Poseidoniastai, merchants and shipowners and warehousemen’ (e.g. 1520.1–2 and 27–8: 149 BC; 1778.2–3: after 150 BC);6 the members of these associations financed the construction of numerous public buildings (Hasenohr 2007: 81–2); their youths were admitted to an ephebic education together with young Athenians and Romans, and continued to attend the gymnasium (Baslez 2002: 55: ‘lieu par excellence de l’intégration sociale’) as pareutaktoi or hieropes.7 5

6

7

Berytus: 54; Tyre: 26; Sidon: 18; Ascalon: 14; Arados: 8; Ptolemais: 6; Marathos: 5; Tripolis: 1 (data from Couilloud 1974: 316). Poseidon and Heracles are to be intended here as the hellenised versions of the Phoenician θεοὶ πάτριοι Baal Bérith and Melkarth (Hasenohr 2007: 84). The bibliography on the Poseidoniastai of Berytus and the Herakleistai of Tyre is very large, and I cannot but refer here to Hasenohr (2007: 79 n. 15) and Hasenohr (2007: 79 n. 16) respectively. Compare Roussel (1931: 447). Baslez (1987: 282 n. 99) provides a list of 29 Phoenicians that are recorded in the ephebic lists of Delos: 10 are from Tyre, 7 from Berytus, 5 from Arados, 3 from Sidon, 2 from Marathos, and 2 from Ascalon. As for the young Romans, 6 are included in the ephebic lists (2593: Γάιος Ῥωμαῖος (as a gymnasiarch), Σπόριος Ῥωμαῖος; 2598: Κόιντος Ῥωμαῖος, Μάνιος Ῥωμαῖος, Γάιος Ῥωμαῖος, Τίτος Ῥωμαῖος), 3 attended the gymnasium as paides (2595: Αὖλος Αὔλου Ῥωμαῖος, Μάνιος Ῥωμαῖος, Κόιντος Κοίντου Ῥωμαῖος), 2 are celebrated as winners of an ephebic lampadodromia (1955: Δέκμ[ος Ῥωμ]αῖος; 2597: Γάιος Κοιίντου [sic] Ῥωμαῖος), and 1 is part of a group of ephebes that subscribe a dedication to Apollo, Hermes, and Heracles (1927: Κόιντος Λευκίου Ῥωμαῖος). The 34 Roman praenomina (Αὖλος, 4×; Γάιος, 7×; Κόιντος, 4×;

174

Multilingualism in Late Hellenistic Delos

The ‘House of Seals’ A representative cross-section of this multicultural and multilingual society can be obtained from the study of the ca. 26,000– 27,000 imprints of seals on ca. 16,000 clay cachets that were found in a private house in the quarter of Skardhana and gave the house the name ‘Maison des Sceaux’ (Boussac 1988; 1992; 1993; Auda & Boussac 1996). The front side of the cachets is impressed with seals (from 1 to 13), while the back side has traces of the papyrus fibres of the documents that were sealed with them and that were lost in 69 BC when the house was burnt down. It was in all likelihood an archive of private contracts that was managed by a man involved in commercial transactions, probably a banker or a broker (Boussac 1988) – maybe Italian (Rauh 1993: 217–18; Zarmakoupi 2013), though not necessarily one of the Mundicii referred to later in this chapter – that also bore the function of syngraphophylax, a private individual who acted as a trustee responsible for keeping contracts (Boussac 1993: 683–4). The nature of the archive can be inferred from the fact that the official seals are extremely scarce (Boussac 1993: 680) and that three quarters of the clay sealings (ca. 76.5 per cent) have at least two seal impressions (Auda & Boussac 1996: 511–12). This is typical of private deeds among individuals, in which the sealing is jointly sealed by the parties, and/or the guarantors, and/or the witnesses who are involved in the transaction (Boussac 1988: 312–18; 1993: 674–9). The archive was in operation during the second period of Athenian rule, as is shown by the presence of the official seal of the Athenian supervisor of Delos (Boussac (1988: figg. 8, 9): bearded head with the inscription ἐπιμελητὴς ἐν Δήλῳ). An absolute chronological reference is provided by a Greek–Phoenician bilingual official seal of a κοινοδίκιον (a joint tribunal consisting of judges appointed by two or more cities, which was responsible for Λεύκιος, 2×; Μαάρκος, 5×; Νεμέριος, 1×; Πόπλιος, 8×; Σπόριος, 3×) that are attested in the graffiti of the gymnasium edited by Couilloud (1970) do not necessarily refer to Roman people. In the absence of the ethnic Ῥωμαῖος, they may also belong to Greeks that took a Latin praenomen as their personal name (see also the discussion later in this chapter, with reference to the seals).

175

Francesco Rovai

the legal conflicts among their citizens in the Hellenistic world; Chaniotis (1996: 141–3)), which dates to year 185 of the Seleucid Era, that is, 128–127 BC (Boussac 1982: 444–6; 1992: 16–17). The sack of the island by Athenodorus in 69 BC, when the entire quarter of Skardhana was destroyed by fire, gives the final date for the use of the house. Movements of People This section is not intended as a comprehensive survey of such a large hoard of imprints. In most cases they do not bear any linguistic information, because inscribed seals are an exceptional minority, and the iconography alone makes it impossible to establish the identity of the seal owners: the portraits are portraits of unknown individuals, and the eclectic stylistic and iconographic repertoire of deities and heroes (Boussac 1988: 333–4; 1993: 686–93) simply results from the well-known heterogeneous religious milieu of the island, where the traditional Greek gods (obviously Apollo, but also Heracles, Artemis, etc.) coexist with the oriental ones (mainly Egyptian, but also Syrian) and with the usual Roman civil cult of Rome, Romulus and Remus, and the rest (Roussel (1987: 206–80); Bruneau (1970)). Instead, this section aims at highlighting some general facts that can provide important insights into the composition and the consistency of the clientele, with some further considerations on those cases where an inscription is available. The small group of official seals is indicative of a range of business activity that included Italy and the Eastern Mediterranean: the seal of a Roman magistrate, probably Cn. Cornelius Dolabella, the proconsul of Cilicia in 80–79 BC (Boussac 1988: 319 fig. 17a–b: Dolabel[la]); two seals of the Seleucid administration (Boussac 1988: 315 fig. 15: a horse’s head and anchor; 319 fig. 16: βασιλέως Ἀντιόχου); some seals of Greek cities of the Cyclades, such as Naxos (Boussac 1988: 315 fig. 10: Ναξι) and Mykonos (Boussac 1988: 315 fig. 11: Μυκο), and of Asia Minor, such as Ephesos (Boussac 1988: 315 fig. 12: Artemis Ephesia and Εφ), Kolophon (Boussac 1988: 315 fig. 13: 176

Multilingualism in Late Hellenistic Delos

Κολοφωνίων), and Tralles (Boussac 1988: 315 fig. 14a–b: Τραλλιανῶν). The cosmopolitanism of the clientele is also confirmed by the rare inscribed sealings that attest Greek, Roman, and Semitic names, which is in line with what it is known about the population of Delos between the second and first centuries BC. The Greek data are not particularly significant, as they are limited to the personal name (in most cases, very common) without any ethnic indication or patronymic. But, again, it may be the case that Oriental people are hidden behind a Greek name: the names Ἰάσων and Στράτων, which are attested among the seals, were often borne by hellenised Semites (Boussac 1988: 322 n. 59; in particular, according to Masson 1969: 682, the former was adopted by Jews and Phoenicians, for the latter, see the discussion earlier in this chapter). As for the Romans, the most updated list of the onomastic material that is engraved on the seals can be drawn from the Liste des Italiens de Délos by Ferrary et al. (2002). There is a total of sixty-one private seals bearing Roman names: fifteen onomastic formulae contain at least praenomen and nomen, of which thirteen are written in Latin8 and two in Greek;9 in two cases the seal is intact but contains only a Greek-written Latin praenomen;10 finally, in forty-four cases, all Latin-written, the names are fragmentary or abbreviated.11 It is not surprising that, for a marker of personal identity in official transactions like the seals, the Italian negotiatores used almost exclusively Latin. This is consistent with their linguistic behaviour in the public and official domain, where Latin is maintained as a marker of corporate identity (see discussion later in this chapter). And the four exceptions of Latin names written in Greek boil down to two 8

9 10 11

T. Aimili(us) P. l., P. Arel(ius), N. Lolli(us) N., M. Mevi(us) L. f., Of. Nae(vius), Ov. Nae (vius), T. Naevi(us), P. Ocrat(ius), P. Po[m]poni(us) C., Q. Sami(arius), C. Sei(us) C. Cn. l., A. Sexti P. l., M. Vibi(us). Γαίος Αὐφίδιος, Αὖλος Σήιος. Π[ό]πλιος, Τίτος. C. An (or Can?), A. Ant, Ant (2×), Q. Ar, M. At, D. Avi, L. Avi, Bai, C. Bu, L. Cae, D. Ce N. l. R, L. Ce, A. Ces, M. Clo, P. Cor, M. Cos, M. Cr An, P. Cr, Cn. Cra, L. En, A. Feid, P. Fu, M. Gr, P. Gr A. l. (or Al?), Q. Gra, Hei, P. Mal, P. Min, Nae, Nei (or Neli?), Q. Non Dip, M. Ort, Paco, T. Ped, P. Po, A. Sae (or Saf?), L. Sas, A.Var, Tr. Vat, L. Vc, L. Ve M. f. C., Ve, . . .]atr P. f. In some cases, such as Hei for Heius and Naev for Naevius, the nomina can be easily integrated. In other cases, such as P. Cr and L. Ve M. f. C., a number of integrations are possible: Crassicius, Crepereius or Critonius, and Venuleius, Veratius, Verginius, Verrinius, Veturius, or Veueius – nomina that are all well attested in Delos.

177

Francesco Rovai

(Γαίος Αὐφίδιος and Αὖλος Σήιος), because the two Greek-written Latin praenomina (Πόπλιος and Τίτος) may simply refer to Greek citizens that took a Latin praenomen as their personal name, as is demonstrated by other instances where the Attic demotic is decisive, such as Γαίος Ἀχαρνεύϛ (2072.4, 2073.6–7, 2091a.4), and Μᾶρκος Ἐλευσίνος (2037.1). The Semitic names are fewer than expected for a community that was well represented on the island (see the discussion earlier in this chapter). The sole two instances, both discussed by P. Bordreuil in the Appendix of Boussac (1988: 339–40), are nevertheless of interest. The former, which is written ‛blM and is interpreted by Bordreuil as the personal name ‛bld’lm ‘serviteur des dieux’ (Boussac 1988: 340), is a curious mixture of three different alphabets: according to the author, the sign for b belongs to the Aramaean alphabet, l is in all likelihood Phoenician, while the final letter is a Greek capital μ. The reading of the latter seal is instead uncertain, but it probably contains three letters lrt’ that may be compared to the Palmyrene name rt’. However, the most interesting aspect of this seal is the fact that it is impressed on the same clay sealing where a Roman seal is impressed (Boussac 1988: 320 fig. 20, 321), which raises some questions about the management of the interaction taking place between the two parts. As nobody would sign a document that he does not understand (at least in its general outlines), this sealing shows that, one way or another, a Roman and a Semite could achieve mutual comprehension. The practical arrangement of this interaction is necessarily bound to remain speculative: perhaps at least one of the subscribers was a fluent bilingual and he was competent in the standard formulae of the bureaucratic language, so that the document, either Latin- or ‘Semitic(= Palmyrene?)’-written, was understandable by both sides. But it is also possible that at least one of the two parts was accompanied by an interpreter or a bilingual clerk; or that they communicated in the lingua franca Greek, so that the document was also written in Greek; or that the document included two versions that were written in the two different languages.12 12

Whatever the actual arrangement, this document, of which only the subscription survives, may be a good example of the application of the Romans’ language policies (Rovai 2015a: 200–4 and references therein), in the spirit of that ‘pragmatic concern for effective communication’ (Rochette 2011: 553) that was consistently formalised in the

178

Multilingualism in Late Hellenistic Delos

This sealing is further evidence of the fact that mutual interaction, and presumably comprehension, between the Romans and people from the Near East was commonplace, as is also shown by the close relationships of the former with the Phoenicians (Hasenohr 2007; Poccetti 2016). In the public domain, on the one hand, the κοινόν of Berytian Poseidoniastai celebrates the goddess Roma εὐνοίας ἕνεκεν τῆς εἰς τὸ κοινὸν καὶ τὴν πατρίδα ‘on account of her goodwill towards the association and the homeland’ (1778.4; cf. also 1779) and has among its members the Roman banker Marcus Minatius (1520; Hasenohr 2007: 87). On the other hand, the Phoenician banker Philostratus of Ascalon offers and receives dedications from the Italici for his generous donations to the Agora des Italiens (1717, 1718, 1722, 2454, 2549),13 to the point of gaining the citizenship of Naples (Hasenohr 2007: 87; Leiwo 1989). Finally, among the subscribers of the Agora des Italiens (2612) there are thirty-six Italians, nine Greeks (including continental Greeks and Ionians), and two Phoenicians (from Tyre and Sidon). In a more private dimension, instead, the prosopographic investigations of Le Dinahet-Couilloud (2001), Baslez (2002), and Poccetti (2016: 412, 417) have repeatedly drawn attention to the existence of mixed marriages between Italians and women from the Eastern Mediterranean. Mobility and Interpersonal Relations In spite of the scarce linguistic data, the frequency of seal impressions and their reciprocal associations can nevertheless function as

13

prescriptions set down by the jurists. With particular reference to the stipulatio, a private legal instrument consisting of an oral question-and-answer affirming the satisfaction of the parties, Ulpian (Dig. 45.1.6) wonders whether the contract remains valid in the case of two parties speaking different languages (not only Latin and Greek, but also ‘Punic or Syrian or any other language’). He reports the statement of Sabinus, a jurist of the age of Augustus, who puts the emphasis on mutual comprehension: the obligation is valid insofar as the parties can understand what the other one is asking and answering, ‘whether they can understand the other one’s language themselves or use an interpreter’ (Dig. 45.1.6: ut uterque alterius linguam intellegat siue per se siue per uerum interpretem). Inscription 2454 is fragmentary and reads Φιλό ̣[στρατος Φιλοστράτου . . . τὴν] ἐξέδραν καὶ τὰ ἐν αὐ ̣τ ̣[ῆι . . .] ‘Philostratus son of Philostratus (gave / offered / financed) the exedra and the (. . .) within it’, with no mention of Italici / Ἰταλικοί. However, the exedra that Philostratus has donated is of one of the three large exedrae that surrounded the colonnades of the Agora des Italiens (Trümper 2014: 73, 75 n. 25).

179

Francesco Rovai

a reliable indicator of the mobility of individuals (Boussac 1993: 686; Auda & Boussac 1996). These numerical indicators may disclose a difference between two types of clients: the largest part of the clientele is occasional, as can be inferred from the fact that most of the seals are impressed only once (ca. 70 per cent; data from Boussac 1993: 686) or twice; the more stable and regular business partners are a small minority, with a restricted group of persons that are responsible for hundreds of transactions (out of a sample of 5,255 seals only 87 are impressed from 11 to 76 times, 5 are impressed ca. 100 times, 1 occurs 255 times, and 1 occurs 357 times; data from Auda & Boussac 1996: 513). Most interestingly, the detailed statistical analysis by Auda and Boussac (1996) on a sample of 3,435 seals impressed on 1,487 cachets, has shown that, among those seals that are impressed at least twice, there are regular co-occurrences of the same seals on the same cachets, that is to say that there were enduring and constant interpersonal relations held between individuals that regularly took part in common transactions.14 Figure 8.1 (adapted from Auda & Boussac 1996: 323) displays, for example, that seal n. 1435 is associated four times with seal n. 71, three times with seal n. 1533, and three times with seal n. 1837; in its turn, seal n. 71 is associated four times with seal n. 1533, and three times with seal n. 1837; and so on. If these clusters of relations are mapped on the ground of the socio-historical reality of Delos, there results the image of a ‘close-knit social network’ in the classical sense of Milroy (1980), which is made of a group of individuals that are members of a high-density and territorially based community, that work in the same place and in the same field (trade and related activities), that create voluntary associations with other members of the community (the religious collegia and guilds operating on the island, the attendance at the gymnasium), and that – at least in

14

The long-lasting association of some individuals with the archive can be seen, for example, from a seal which is impressed 110 times and whose quality of impression of the image (Eros wearing a Silenus mask) underwent progressive deterioration (intact, cracked, and broken), which is a sign that it was in use for years (Boussac 1993: 686; Auda & Boussac 1996: 520–1).

180

Multilingualism in Late Hellenistic Delos

Figure 8.1 Network of associations within a group of seals (Adapted from Auda & Boussac 1996: 323)

some cases – may be demonstrated to have ties of kinship on the island. It is important to highlight that the native culture of the individuals seems to be rather an irrelevant factor in the establishment of these interpersonal relations. This may appear trivial with regard 181

Francesco Rovai

to business and trade activities, but it holds for other aspects of the social life as well. From the 130s BC, for example, Athenians and other Greeks, Italians, and Semites gathered together under the collective definition of Ἀθηναίων καὶ ‘Ῥωμαίων καὶ τῶν ἄλλων ξένων οἱ κατοικοῦντες καὶ παρεπιδημοῦντες ἐν Δήλωι (1646) in a number of honorific decrees made on the combined initiative of the various groups (cf. also 1644, 1645, 1650, 1651, 1652; this formula and its variants are discussed in Roussel 1987: 50–5). Furthermore, they all shared the cults of a remarkably eclectic pantheon. Apollo is honoured not only by Athenian priests in his own sanctuary but also by the Italian Apolloniastai and the Berytian Poseidoniastai (Hasenohr 2007: 85). When the worship of the Egyptian god Sarapis, formerly private, was made a public cult, a sanctuary (the so-called Serapeum ‘C’) was built in the first half of the second century BC (Scott 2015: 236), and within the sanctuary itself the temple of Isis and its cult statue were bestowed by the Athenians (2041, 2044).15 The review of the inventories of the temple (1403, 1412, 1416, 1417, 1434, 1435, 1440, 1442, 1445, 1452, 1453, 1454), dedications (2037–219), and lists of subscribers (2614–25) made by Roussel (1916: 281–4) shows that ‘[l]es fidèles des dieux égyptiens de Délos se recrutaient dans toutes les région du monde gréco-romain’ (Bruneau 1970: 466). And next to the Serapeum ‘C’, a sanctuary devoted to the Syrian divinities Atargatis and Hadad (who were hellenised as Hagne Aphrodite and Zeus Hadad; cf. Bruneau (1970: 470)) was constructed in the second half of the second century BC (Bruneau 1970: 466). Also in this regard, as can be seen from the numerous dedications (2220–304) and lists of subscribers (2626–8), the worshippers engaging with this 15

Very soon after 166 BC, however, the Serapeum was closed by order of the Athenian government. The priest Demetrius of Rhenea, of Egyptian origin, appealed to the Roman Senate and obtained a decree in his favour (Sherk 1969: 37–9; Durrbach 1921: 116–21), so that the Athenian ἐπιμελητής of Delos was forced to allow Demetrius to reopen his Serapeum. However, the decision of Athens was not due to hostility towards the Egyptian cults as such: as properly stated by Sherk (1969: 39) ‘it was not the cult of Serapis at all which angered the Athenians but rather the fact that it was not an Athenian who controlled this particular sanctuary’. After 166 BC public sanctuaries had become the property of Athens, and Athenian involvement in all ‘official’ cults is clear – for obvious reasons of political influence. For example, in addition to the administration of the sanctuary of Apollo, three of the attested officials in the sanctuary of the Syrian gods (the priest, the kleidoukhos, and the kanephoros) had to be Athenians (Scott 2015: 237; Bruneau 1970: 472).

182

Multilingualism in Late Hellenistic Delos

temple were Athenians, both priests (e.g. 2227, 2228) and private individuals (e.g. 2250, 2251), Romans (e.g. 2245, 2255), and Syrians (e.g. 2224, 2256; cf., in addition, Siebert 1968 for τὸ κοινὸν τῶν θιασιτῶν τῶν Σύρων τῶν εἰκαδιστῶν οὓς συνήγαγε ἡ θεός ‘the association of society members of Syrians of the twentieth day of the month which the goddess gathers together’).16 Finally, with regard to more personal aspects of social life, it may be appropriate to recall that Greeks, Romans, and Semites entered into mixed marriages (cf., again, Le Dinahet-Couilloud 2001, Baslez 2002, and Poccetti 2016: 412, 417), and as young men they attended ephebic training and the gymnasium side by side (see n. 5). Such activities were in their turn linked to the cult of Apollo (Bruneau 1970: 166–7). Against this social background as described, the second part of the chapter focuses on the following aspects of multilingualism and multiculturalism. On the one hand, it is clear that such a multilingual society, whose members are closely connected with each other, cannot but favour both societal and individual bilingualism: as shown later in the chapter, Latin (and its varieties as well) could be part of the linguistic repertoire of a Greek speaker. At the same time, this does not exclude the existence of smaller-scale, equally close-knit social networks that are based instead on a shared cultural background. Thus, on the other hand, the dialectics between accommodation in a multicultural society and the maintenance of cultural individuality can be overtly marked by means of different language choices, as illustrated later in this chapter with reference to the Latin-speaking Italici. The Greeks and Their Language The Greek population of the island was in large part made of Athenians, as should be expected in the light of the historical 16

The term εἰκαδισταί is interpreted by Siebert (1968: 362 n. 1, ‘qui ont leur fête le vingtième jour’) in comparison with other terms of the same kind that were used on Delos for groups of Egyptian devotees who assembled on the corresponding days of the month (cf. also Roussel 1916: 100, 253): the δεκαδισταί ‘those of the tenth day / the tenth-day celebrators’ (IG XI.4 1227.1: ca. 200 BC) and the ἐνατισταί ‘those of the ninth day / the ninth-day celebrators’ (IG XI.4 1228.2, 1229.2: beginning of the second century BC). Both δεκαδισταί and ἐνατισταί co-occur in a short inscription from Cos (IG XII, 4 2.551: first century BC).

183

Francesco Rovai

and political context, as we have already seen. Long before the 166 BC settlement, the Athenian influence on Delos dates back to the sixth century BC and, most of all, to the fifth century BC with the creation of the Delian League under the leadership of Athens. This had a direct impact on the Greek language that was written and spoken in Delos which, like the other Cycladic islands, underwent progressive Atticisation first, and then Koineisation from the fifth down to the second century BC (Handel 1913). The Linguistic Repertoire of the Greek-Speaking Delian Community As a result, the language of the public, official epigraphy of Delos is the Attic Koine that essentially developed from the so-called Great Attic, a superdialectal variety used in Attica, Euboia, the Ionic Cyclades, and Asian Ionia as the written language of the administration in the late fifth and fourth centuries BC (Horrocks 2010: 73–7, 80–4; Bubeník 1989: 178–81).17 A number of typical Great Attic/Koine forms are attested in the Delian official inscriptions (state decrees, records of the temples, honorific dedications on public buildings, statue bases, altars, etc.). Among others: the spelling ἀμφικτύων instead of ἀμφικτίων (cf. Threatte 1980: 263–4); the analogical levelling of (παρ-)ἐδώκαμεν (aor. 1st pl.) (e.g. 1401.c.9, c.10 and c.18: after 166 BC; 1416.A.2.15: 156 BC; 1417.B.1.149: 155 BC; 1442.46, 49; and 73: 146–145 BC) and ἔδωκαν (aor. 3rd pl.) (e.g. 1507.10: 145–135 BC; 1514.10: 116 BC) instead of (παρ-)ἔδομεν and ἔδοσαν respectively; ἐγίνετο for ἐγίγνετο (e.g. 1517.16: 156 BC) and γίνεσθαι for γίγνεσθαι (e.g. 1517.26: 156 BC); ἕνεκεν for ἕνεκα (e.g. 1520.30 and 43: 149 BC; 1527.6: 145–116 BC; 1778.4: after 150 BC); the initial aspiration in ἕτος instead of ἔτος (e.g. καθ’ ἕτος in IG XI.2 135.26: 314–302 BC); the spelling of the diphthong -ηι as -ει in the dative singular of η-stems (e.g. 1501.22: 148 BC; 1507.6: 146–135 BC; the common formula of the decrees: ἀγαθεῖ τύχει· ἔδοξεν / δεδόχθαι τεῖ βουλεῖ; compare also Threatte (1980: 377–80)). 17

During the Athenian cleruchy (166–140s BC) several decrees were also sent to Athens to ask for confirmation, and in some cases the ratification of Athens is inscribed afterwards, such as in 1504 (147 BC) and 1505 (146 BC).

184

Multilingualism in Late Hellenistic Delos

More private documents like the epitaphs from the necropolis of Rheneia (Couilloud 1974; Poccetti 2016) bear, instead, some traces of a spoken ‘innovative’ variety that was part of the linguistic repertoire of the Greek speakers during the Hellenistic period. These texts are almost exclusively limited to the stereotyped formula ‘[name of the deceased] χρηστὲ/-ὴ (καὶ ἄλυπε) χαῖρε’,18 but some alternations in the variable parts of the text (i.e. the name of the deceased and his/her demotic) indicate a tendency to iotacism (e.g. C 155: Ἱεροπολεῖτι ~ C 131: Ἱεροπολῖτι; C 420: Εἰσιάς ~ C 438: Ἰσιάς; C 45: Μεικώνιε for Μυκόνιε) and a confusion between long and short vowels (e.g. C 154: Ἡράκληα ~ C 485.2: Ἡράκλεαν, both for Ἡράκλεια; and, again, C 45: Μεικώνιε for Μυκόνιε). Although they underwent separate developments in Roman times (with [ɛː] > [eː] > [i] ≠ [e] > [ɛ], and [ɔː] > [oː] > [o] = [o]), the spelling alternations ~ and ~ may be related to the loss of distinctive vowel length in some domains of the Greek language – even if the precise chronology of this phenomenon is very debated and not easy to assess (see Teodorsson (1974: 183–4, 211–2, 218–19) versus Threatte (1980: 159–64, 384–7); further discussion in Allen (1987: 94); Duhoux (1987); Bubeník (1989: 184); Horrocks (2010: 163–5); Rovai (2015b: 168–9) and references therein). However, further evidence for an advanced loss of vowellength distinctions in Delos during the second and first centuries BC, can be drawn from the Greek transliteration of some Latin names. While Latin ō is regularly transliterated as , spellings such as Ῥομαία (C 150, bis; instead of Ῥωμαία, passim), Νόννεις (C 318; instead of Νώνιο[ς], 2616 / Νῶνα, C 52),19 and Ὀφίδιος (for Latin Ōfidius in 2623.A.1.3, a list of subscribers of the Serapeum) can be interpreted as the result of an ‘under-differentiation’ on the part of (more or less) bilingual speakers,20 who substitute a single phoneme of their primary language (Gr. /o/) for two phonemes of the secondary language (Lat. /o/ ~ /oː/) whose opposition is not contrastive for them. 18 19

20

The few exceptions are listed in Rovai (2015b: 168 n. 19). For the identification of Νόννεις with Latin Nonius, with particular reference to the spelling -εις of the ending, see Poccetti (2016: 424–5). Compare Weinreich (1966: 18): ‘under-differentiation of phonemes occurs when the sounds of the secondary system whose counterparts are not distinguished in the primary system are confused’.

185

Francesco Rovai

Also the representation of Latin ĭ by in Ὀλόσσηιος (2248, for Volusĭus) and by in Γάειος (C 492, for Gaĭus), indicates, in addition to iotacism, the absence of quantitative distinctions in the language of the Greek drafters. Further evidence in favour of the iotacism that characterised their spoken variety comes from the fact that Latin ī may correspond to either (e.g. 442, 443: Πλίνιος, for Plīnius), (e.g. 1429, 1432: Σαβείνιος, for Sabīnius),21 or (e.g. 2351: Σερουήλιος, for Seruīlius), and Latin ĭ to either (e.g. 1751, 2628: Λικίνιος, for Licĭnius) or (e.g. 1417, 2124, et passim: Αἰμύλιος, for Aemĭlius). These kinds of variations are consistent with the hypothesis of the existence of at least two subsystems in the phonetics and phonology of the Greek-speaking Hellenistic world, in continuity with a pre-existing distinction between two different varieties within the fourth-century Attic: the ‘conservative system’ of the aristocracy, which mostly retained both qualitative and quantitative oppositions, and the ‘innovative system’ of the moderately educated, which was instead characterised by a remarkable iotacism and was far more advanced in the loss of vowel-length distinction (Teodorsson 1978: 91–8, discussed in Horrocks 2010: 163–5).22 21

22

Throughout the Hellenistic period, the use of for an /i/ sound indifferent in length (as a consequence of the loss of quantitative distinction; see above), is not uncommon in Greek. Examples of a confusion of and are sporadically attested in Attic inscriptions from fourth century BC on, although they remain rare until the last quarter of the second century BC (Threatte 1980: 195–8), and the phenomenon is frequent in the Ptolemaic papyri from the third century BC on (Gignac 1976: 191 n. 2, and references therein). However, it cannot be excluded that spelling for Latin ī reflects the influence of an official Latin orthographic practice. First, was also a Latin device to write ī in line with the recommendations of Accius (Mar. Victorin. 6 8 K). Second, Greek for Latin ī was part of the standard of transliteration of Roman names in the translations of formal public documents (senatus consulta, epistulae, etc.) that were addressed to the Greek-speaking world. See, among many other examples that can be drawn from Sherk’s (1969) collection, Roman Documents from the Greek East, the 112 BC epistula L. Calpurni Pisonis (Sherk 1969: 78–85): Λείβιος for Līuius (l. 86), Πείσων for Pīso (l. 89), and (Καλ.) Κοινκτειλίων (gen.pl.) for (Kalendae) Quin(c)tīles (ll. 74, 85–6). My generalisation is a twofold, simplistic representation of a much more complex state of affairs. During the Hellenistic period, the influence of the Koine on the local languages and the relationships holding between the ‘conservative’ language of the upper classes and the subordinate ‘innovative’ varieties could vary in rather different ways in ‘old’ Greece, where the Koine was used at the expense of the ancient dialects, and in the new Hellenistic kingdoms, where at first it was the only written standard and the spoken language of the Greco-Macedonian elites (Horrocks 2010: 83 ff.; cf. also

186

Multilingualism in Late Hellenistic Delos

The linguistic features of the public and funerary epigraphy suggest that this was also likely to be the condition of the Greekspeaking community of Delos, where the Koine forms of the official written variety used for decrees, honorific dedications, and so on are attested alongside some features of a spoken ‘innovative’ variety that have occasionally penetrated the written language of more private documents like the epitaphs. Thus, the Koine functions ‘not only as the standard written and spoken language of the upper classes . . ., but also more abstractly as a superordinate variety standing at the pinnacle of a pyramid comprising an array of lower-register varieties, spoken and occasionally written’ (Horrocks 2010: 84; cf. also Consani 1991: 16; Consani 2014: 121–2). As Teodorsson (1978: 93) correctly emphasises, such a distinction should be regarded not as a clearcut diastratic opposition but as a scalar diaphasic variation, with most users of the conservative and the innovative subsystems respectively being more or less competent in the other subsystem. Moreover, a certain degree of societal bilingualism must be presupposed for the Greeks dwelling in Delos, in order for them to do any kind of business with the foreign groups on the island, especially the Romans. The Linguistic Repertoire of an Individual There is first-hand evidence for individual Greek–Latin bilingualism in a graffito, which is particularly interesting both for its linguistic features and for its authorship. It was found on a wall of the ‘Maison du Théâtre’ and drawn as follows (Figure 8.2) by Rostovtsev (1896: 392), who also offered the text and the interpretation that are accepted in later editions (1): Horrocks 2010: 85–7, for an illustration of the top-down Koineisation of Boeotian; Horrocks 2010: 101–5, 163–5, for the evolution of the Attic varieties and their features; Horrocks 2010: 111–13, 165–87, for the development of the Egyptian Koine). In any case, ‘Koine’ does not necessarily mean ‘conservative’: for example, the New Comedy of Menander testifies to an upper-class language spoken in Attica, which is rather more conservative than the ‘international’ Attic Koine of the Hellenistic world at large (Horrocks 2010: 101–5). Finally, these relationships could be particularly complex in those communities where the influence of the Koine was overtly challenged by regional varieties, dialect Koines, and local dialects for reasons of identity and political independence (cf. the cases of Cyprus and Thessaly illustrated in Consani, 2014: 122–3).

187

Francesco Rovai

Figure 8.2 Graffito of the ‘Maison du Théâtre’ (From Rostovtsev 1896: 392)

(1) Valeri Nerius uneo igni suạ // Augias (Rostovtsev 1896: 392; AE 1897.33; CIL I2 3437b; CIL III 14203.7)

The text dates to ca. 100–70 BC,23 and is inscribed with a stylus and undersigned Augias. Thus, it was scratched by a Greek individual who had mastered the Latin alphabet. The text seems to be praise for an actor, and it is interpreted by Rostovtsev (1896) as ‘Oh Valerius! Nerius (is) a pearl by virtue of his fire. Augias’. In addition to the place where it was found (a house of the Theatre District), this reading may be supported by the fact that unio (here uneo; see below) with reference to an actor also occurs in a Pompeian graffito, in which it is said of a pantomime: Paris unio scaenae ‘Paris, pearl of the stage’ (CIL IV 3867; for unio as a eulogistic term, see also Robert & Robert 1976: 578). However, as shown by Clackson (2016: 84–6), unio was also an old Latin term for dice throws and indicated the ace, the worst and lowest throw of a single die. According to Isidore (Or. 18.65) the ace was called unio by the lusores ueteres but this name was later replaced by canis (‘dog’), as a calque from the Greek κύων (cf. Poll. 9.100: καὶ τὸ μὲν μονάδα δηλοῦν, καλείται κύων ‘and the throw of an ace is called dog’). Board games with dice were extremely popular among all social classes and they were played everywhere from the Pompeian taverns (Clackson 2016) to the halls of Augustus (Suet. Aug. 71.2). Typical game situations are often referred to by authors as usual images of everyday life (see, among others, Pl. As. 904–5, Cur. 354–60; Mart. 14.17; Sen. Ep. 117.30), and people were familiar with their rules and 23

According to Perdrizet (1902), the term unio (‘pearl’) could provide a terminus post quem for the Delian inscription, in view of a passage from Pliny the Elder (Nat. 9.123), who reports that the word unio was coined by Aelius Stilo in the time of the war against Jugurtha (112–105 BC). Just as with most of the Delian inscriptions, the terminus ante quem remains the 69 BC strike by Athenodorus.

188

Multilingualism in Late Hellenistic Delos

terminology, which could undergo metaphorical and figurative uses (Pl. Rud. 358–60: O, Neptune lepide, salue, nec te aleator nullus est sapientior; profecto nimis lepide iecisti bolum ‘Bless you, excellent Neptune, there is no player more expert than you with the dice; to be sure, that was a beautiful throw of yours!’). Thus, it cannot be excluded that Nerius (as well as his Pompeian colleague Paris) was not an excellent but a very bad actor, so that he was taunted by means of comparison with a particularly unlucky dice throw. In addition, the author of the graffito was bilingual and he could have been aware of the correspondence between the Latin unio and the Greek κύων, which was a common insult in Greek. A few palaeographic observations may be appropriate here, because, except for a few cursive tendencies (the descender of ‹s› and ‹r› extended below the baseline; the ligature between ‹i› and ‹a› in Augias) and for the archaising form of the letter ‹a› (but the horizontal medial stroke that is detached from the left bar and hangs as a short oblique stroke might also be the result of a tendency to verticalisation which is typically cursive), it is worth noting that the letter forms are a good example of a capital script, which suggests that the Greek writer derived his model from the ‘square capitals’ of the public Latin epigraphy and/or from the ‘rustic capital’ that was the standard calligraphic book script. In any case, the most relevant linguistic feature of this text is the form uneo instead of unio, which – whatever its meaning – testifies the use of e for i in hiatus.24 But, as a preliminary issue, it must be safely excluded that such a vowel lowering is the result of an influence of Augias’s native language. Phonetically, the pervasive iotacism that characterised Hellenistic Greek in general led to the raising and merging of front vowels, so that the lowering of a Latin [i] to [e] in the speech of Greek–Latin bilingual seems to be highly unlikely; but, orthographically, not only for but also for could be a hint of a confusion between two front vowels.25 24

25

The lack of gender agreement between igni and sua is not certain, because the last letter is partially unreadable and it could represent either an or an . In later Greek papyri from Egypt (first–seventh centuries ad), occasional interchanges between and before a back vowel are discussed by Gignac (1976: 249–53) as

189

Francesco Rovai

In this sense, the spelling of the name Augias in itself, which is normally spelled Αὐγέας in the Greek script (cf. the name of the famous king of Elis in the literary sources), might testify to a vowel raising and demonstrate some uncertainty about front vowels in hiatus, but it must be taken into account that doublets of masculine personal names in -έας and -ίας both existed in Attic (Καλλέας and Καλλίας, Νικέας and Νικίας, etc.; Threatte 1980: 145–6), and it may well be the case that Augias is simply the transcription of a possible Greek form Αὐγίας. Also the name Nerius deserves further discussion, because, as the name of an actor, it might be expected to be a Greek name (cf. Paris in CIL IV 3867), perhaps a stage name. In this case, it may be supposed to represent Νηρεύς but, as far as we know, there is no evidence that the short first element of the diphthong ευ was ever involved in a process of raising in Koine Greek (Threatte 1980: 345–8). Rather, the diphthong adopted a closer articulation of the second element that led to the Modern Greek pronunciation [ef] / [ev] (Horrocks 2010: 163–5, with references to Teodorsson 1974 and 1978). Nerius can therefore be intended as a Latin nomen referred to a member of the gens Neria, and the fact that, like the opening vocative Valeri, it occurs as a single personal name without any praenomen or cognomen is consistent with the onomastic practice of a Greek speaker. Thus, as the Latin spelling Augias may correspond to a possible Greek form Αὐγίας, and the Latin names Valeri and Nerius indicate that the writer was well aware of the distinction between (= [ɛ ~ e]) and (= [i]) – at least in the transcription of Latin examples of the eventual identification of two originally different sounds throughout the Roman and Byzantine period. As shown by Threatte (1980: 137–45), this interchange is, instead, very rare for the period and the area to which the Delian inscriptions belong, and the few cases are probably not best explained on phonetic grounds, but usually only graphic mistakes are involved. Gignac (1976: 255–6) also highlights that such an interchange may occur in the transcription of Latin i, which fluctuates between and for the following loanwords or names in the late Egyptian papyri: κόμε(τος) / κόμιτος for comitis ‘comrade (gen. sg.)’, πρίνκεπος / πρίνκιπος for principis ‘first in order (gen. sg.)’, Δομετιανοῦ / Δομιτιανοῦ for Domitiani, and Βρεταννικοῦ for Britannici. It must be noted that Latin i is short (and unstressed) in all cases, and this fluctuation in the Greek spelling of a Latin ĭ is quite common also in the inscriptions from Late Hellenistic Delos (Rovai 2015b: 177–80). However, the fact that both the Greek and may transcribe Latin ĭ does not depend on Greek iotacism, but is indicative that the Latin short high-vowels were particularly open (Allen 1978: 49; 1987: 63–4).

190

Multilingualism in Late Hellenistic Delos

forms – uneo can be regarded not as a graphic by-product of the influence of Greek iotacism, but as a genuine Latin form, which was present in a Latin variety that was used in Delos and that Augias came into contact with. Notably, the spelling of e for i in hiatus is traditionally assumed as a feature of Faliscan and other non-urban dialects of Latin, namely Praenestine (Ernout 1905–6: 318–19; Campanile 1961; Coleman 1990: 11; Joseph & Wallace 1991; Mancini 2005: 1038). More recently, Adams (2007: 68–72) has critically re-surveyed the epigraphic evidence and observed that the size and the chronology of the corpus of examples are insufficient for arguing that e for i before a vowel reflects a phonetic feature of the language spoken in the territories of either Falerii or Praeneste.26 However, building on the literary evidence of the well-known passage from Plautus’s Truculentus (Pl. Truc. 691), in which it is stated that the inhabitants of Praeneste said conea instead of ciconia (‘stork’), he acknowledges that the lowering of i in hiatus in the early period would seem to have been associated with the Praenestine dialect or, more generally, with a non-urban regional dialect (see also Adams, 2013: 102–3). 26

The forms discussed by Adams (2007: 70), with their provenance and dates, are the following: filea (CIL I2 60: Praeneste, third–second c.), filea (CIL I2 561: Rome, midfourth c.), fileod (CIL I2 2658: Tibur, sixth c.), Feronea (ILLRP 93b: Lucus Feroniae, late third c.), Feronea (CIL I2 1834: Trebula Mutuesca, undated), Oueo (CIL I2 234: Praeneste, undated), precaream (CIL I2 2214: Aquileia, first c.). Some further chronological details can probably be given. Oueo for Ouius is found on a cippulus from the necropolis of Praeneste. Of course, the cippuli Praenestini ‘vary in date, and a date for this text is not given’ (Adams 2007: 69 n. 140), but the necropolis was in use between the third and the second c. (although most of the inscriptions – with a 3:1 ratio – belong to the third c.), so that Oueo can be considered contemporary with the other Praenestine form filea (CIL I2 60). Moreover, two additional forms can be mentioned. The former is uieam, which occurs alongside precaream in CIL I2 2214, with the whole text running as follows: uieam precaream. The digraph ‹ie› may be an alternative, tentative transcription of the phonetic lowering of i in hiatus. Most of all, it seems to suggest that the form ueha for uia attributed by Varro (R. 1.2.14) to rustics also belongs in the category under discussion (contra Adams 2007: 70; for uoia as an alternative spelling of uia, see Vine 1998: 258–9 and references therein). The latter form is udeom ‘drinking vessel for water’ (< *udiom < *wd-jo-; according to Mancini 2004), which is readable on the ‘Garigliano bowl’ (Vine 1998: early fifth c.), but the reading of this inscription remains disputed. Vine (1998), followed by Clackson and Horrocks (2007: 30), reads Audeom duo[m] ‘of the two Audii’ and takes Audeom as a genitive plural form for Audium that attests, in any case, ‘the banal O[ld] Lat[in] (and partly non-urban) prevocalic E for I’ (Vine 1998: 262); more recently, Crawford et al. (2011: 359–61) read instead ad deom duon[o]m and translate ‘at (the shrine) of the good gods’.

191

Francesco Rovai

This graffito, with its script and its linguistic features, raises several considerations on how complex and multilayered the linguistic repertoire of a single individual could be in the multilingual and multicultural milieu of Delos during the second and first centuries BC. Augias was a Greek– Latin bilingual and, as a member of a Greek-speaking Late Hellenistic community, his Greek competence was likely to include the range of variation sketched elsewhere in the chapter. Moreover, as he was literate in Latin, it can easily be deduced that he was able to also write in Greek. In addition, not only was he acquainted with a written, and, therefore, (at least to some degree) standardised, variety of Latin, but his Latin competence and speech were also influenced by a non-urban variety of spoken Latin. The Italici and Their Language Although they were deeply integrated into a tight network of social relationships with the Greeks and other Eastern people, the community of the negotiatores, those freemen, freedmen, and slaves who came from Italy for business, preserved a well recognisable corporate Latin-speaking identity that distinguishes them from all the other groups acting on the island. Identity, Accommodation, and Linguistic Choices As suggested in Mullen (2012: 26–7), the concept of ‘ethnolinguistic vitality’, that is, ‘that which makes a group likely to behave as a distinctive and collective entity within the intergroup setting’ (Giles, Bourhis & Taylor 1977: 308), may have potential heuristic value in understanding the linguistic behaviour and the language choices of a speech community of immigrants in situations of language contact. The Italici exhibit a high level of ‘ethnolinguistic vitality’: their shared cultural background is overtly symbolised by several collective self-definitions that they use in either Greek- or Latinwritten public epigraphy: they call themselves Italici / Ἰταλικοί 192

Multilingualism in Late Hellenistic Delos

(‘those who come from Italy’);27 they make explicit reference to their common economic interests (Hatzfeld 1919: 192 ff.; Rauh 1993; Zarmakoupi 2013), describing themselves as businessmen qui Deli negotiantur (cf. 1620, 1695, 1696, 1698) / οἱ ἐν Δήλωι πραγματεύομενοι / ἐργαζόμενοι (cf. 1725, 1727, 1729);28 they operate in the public domain as the magistri of four major religious collegia (Hermaistai, Apolloniastai, Poseidoniastai, and Competaliastai; cf. 1731–70),29 which are set up according to the model of the collegia acting in Southern Italian cities (Hasenohr 2002; 2007); they head the construction of a large square, the so-called Agora des Italiens (cf. 2612 and 1683– 93bis for the neighbouring buildings), which became a gathering point for their entire community (Hasenhor 2007);30 in addition to the religious cults that are common to other groups (see the ‘Movements of People’ section), they attended the typical Roman domestic cult of the Lares that was celebrated by the Competaliastai (Hasenhor 2003). The dialectics between their accommodation in a Greekspeaking environment and the maintenance of such a wellcharacterised identity is reflected in their choices of language for the different types of written documents. Drawing from the liste des Italiens de Délos (Ferrary et al. 2002: 186–226), 314 texts are known that may be related to members of the Italian community (Table 8.2):31 27

28

29

30

31

Compare 1620, 1683, 1685–91, 1695–6, 1698–9, 1717–18, 1722, 1735–6, 1742, 1757, and 1758. All over the Greek-speaking areas in the late Republican period, the term Italici / Ἰταλικοί has a chorographic not political sense (Adams 2003a: 651–8). The precise origins of the negotiatores in Italy have been strongly debated, but Hatzfeld’s (1912; 1919: 238–56) claim that most of them were from Southern Italy is no longer acceptable today after Solin’s (1982) detailed onomastic study, which indicates that a large part of the traders originated in Latium and Rome (Solin 1982: 117). Note that, to Athenians, they were instead Ῥωμαῖοι (‘Romans’), in opposition to ‘Athenians’ themselves and ‘other foreigners’ (Adams, 2003a: 652–6). Compare also the trade guild inscriptions of the olearii / ἐλαιοπῶλαι ‘oil sellers’ (1712–14), the οἰνοπῶλαι ‘wine sellers’ (1711), and the τραπεζῖται ‘bankers’ (1715). For the social particulars of the collegium of the Competaliastai, which was made of freedmen and slaves of Italian familiae, see Adams (2003a: 668–9). For the controversy over the Agora des Italiens and its alleged use as a slave market, as a sports arena, or as a luxurious park-like building, see the literature referred to in Zarmakoupi (2013: § 4 n. 6). Neither the Italiotes from Greek-speaking areas of Italy (Ferrary et al. 2002: 236–9) are taken into account, nor those formulae that include only the praenomen (Ferrary et al.

193

Francesco Rovai Table 8.2 Delian documents containing Latin names Types of inscription

n.

Language n.

Accounts of the temples Ephebic and gymnasium lists Honorific dedications and decrees

40 6 139

Greek Greek Greek Latin bilingual Greek bilingual Latin Greek Latin Greek Latin

Funerary inscriptions Graffiti Seals Defixiones

63 6 59 1

40 6 101 17 21 61 2 3 3 57 2 1

They had no choice either for the accounts of the temples or for the lists of the gymnasium, which could only be written in Greek, as they were commissioned by the Athenian administration to professional stonemasons. On their part, the gravestones from Rheneia are carved as a serial and rather banal production (Couilloud 1974: 53–9) by primarily, if not exclusively, Greek-speaking draftsmen (Poccetti 2016: 422). But for the other typologies of sources an explicit choice was made by the Italian negotiatores. As already advanced by Adams (2003a: 645–61), the use of Greek in public epigraphy expresses their accommodation in Greek-speaking Delian society; at the same time, the maintenance of Latin in the decrees of the collegia and other public inscriptions in which they acted under the collective name of Italici underlines their corporate Latin-speaking identity. In contrast with the public epigraphy, the use of Latin for an inherently private action like a malediction (the Rheneia curse tablet: 2534; Adams 2003a: 680–2; Poccetti 2016: 421), as a marker of personal identity (the seals), and in more casual contexts (the graffiti) suggests that Latin remained the spoken language for inner communication among the 2002: 226–34), which is not decisive for identifying an individual as a Roman (see n. 7 and earlier discussion), or a few scattered letters (Ferrary et al. 2002: 234–5).

194

Multilingualism in Late Hellenistic Delos

negotiatores.32 What kind of Latin, is the question addressed in the following sections. Spoken Latin in Greek Inscriptions In Rovai (2015b) I tried to individuate some features of the variety of Latin that was spoken by the community of the Italian negotiatores, through the investigation of the transliteration of the Latin personal names within Greek texts. The accommodation of Latin onomastics into Greek orthography does, in fact, reflect a complex overlap of the graphemic conventions current in the two languages with a set of tentative correspondences that switch between orthographic and phonetic transcriptions (Rovai (2015b: 170–80), and references therein). Once we exclude what can be traced to the official, standard criteria of transliteration that were established in the Greek versions of senatus consulta and in the epistulae of Roman magistrates (Rovai 2015b: 170, 172–3),33 and what is due to the influence of the language that was spoken by the Greek drafters of the inscriptions (Rovai 2015b: 174–5; see also earlier discussion on the inconsistencies in the use of , , , and for Latin [i(ː)], which reflect the Greek iotacism), there remains a reliable corpus of nomina, whose Greek transcriptions entail a spoken, rather than written, Latin model. In particular, they testify to some well-known phenomena that are currently acknowledged as colloquial features of the Latin speech (see a list in Väänänen (1967: 35–45); Clackson (2011: 520), among many others): the monophthongised pronunciation of 32

33

The use of Latin for more personal purposes finds second-hand evidence also in the accounts of the temples, which record some ἀναθήματα offered by Romans and inscribed ῥωμαϊκοῖς γράμμασιν ‘with Roman letters’ (442, 443, 1439). It must be borne in mind that these texts, of which the senatus consultum referred to in n. 15 is an example, were composed and set out in Rome by professional Roman draftsmen (Sherk 1969: 13) in order to avoid ‘misunderstandings’ on the part of the Greek communities. The closing of the epistula of Quintus Minucius Thermos to the Milesians (50–51 BC; Sherk (1969: 272–6)) is particularly significant in this regard: Τὴν δὲ αἰτίαν δι’ ἣν ἑλλη[νι]κοῖς ἔγραψα, μὴ ἐπιζητήσητε. κατὰ νοῦν γὰρ ἔσχον, μή τι παρὰ τὴν ἑρμηνείαν ἔλασσον τὰ γεγραμ[μ]ένα νοῆσαι δύνησθε (‘The reason why I wrote in Greek, do not ask, since it was my intention that nothing contrary to the correct interpretation of my letter could be possible in your mind’).

195

Francesco Rovai

au is witnessed by the form Σωφήιε, which occurs in a bilingual epitaph (C 243) alongside the Latin vocative Saufeie (Poccetti 2016: 421);34 the syncope of a short post-tonic vowel is indicated by the Greek script in the usual transcription of the praenomen Decimus as Δέκμος (passim); and many instances of transliteration of the Latin ĭ as (e.g. 2616 and C 75: Καικέλιος; 2598: Φλαμένιος; 1732, 1753, and 1754: Τεβέριος) and ŭ as (e.g. 1427 and 2612: Δεκόμιος; 1755 and 2618: Λοκρήτιος; 1430, 1432, 1449, 1450, and 1451: Ποστόμιος; 2346: Σπόριος), speak in favour of a lowered pronunciation of the Latin short high-vowels, which is recognised by scholars as a phonetic feature of spoken Latin from early times (Allen (1978: 49); (1987: 63–4); Vineis (1984: 49–53); Marotta (2015: 46–52), and references therein; see also Rovai (2015b: 177–80), for a full list of examples and further discussion).35 The Graffiti While my investigation in Rovai (2015b) was based exclusively on public, official inscriptions and ‘private’ epitaphs, this section discusses instead a numerically minor, and more casual, typology of texts, that is, graffiti. The Latin graffiti are extremely scarce – three texts total, of which one is the work of a Greek. Of the other two, the first, which is also the shortest, is not particularly interesting except from a prosopographic point of view. It was scratched into a wall of the ‘Maison des Sceaux’ (whose destruction in 69 BC can serve as the terminus ante quem for this text; see note 23) and consists of the onomastic formula of two brothers, which is set in between two lines that partially preserve some readable Greek letters (2):

34

35

Κλώδιος (e.g. 1758.9, 1761.3), Ὀφίδιος (2623.A.1.3), and Πλώτιος (e.g. 1732.7, bilingual, with the corresponding Latin form Plotius at l. 2; 1763.10) may instead be related to the variants Clodius, Ofidius, and Plotius that are well attested in Latin alongside Claudius, Aufidius, and Plautius. The qualitative similarity of Latin /i/ (=[ɪ]) with /eː/ (=[e]) and /u/ (=[ʊ]) with /oː/ (=[o]) has its articulatory motivations in the correlation between vowel length and tenseness, so that short vowels have a more open pronunciation than their long equivalents (cf., among many others, Jakobson and Halle (1963)).

196

Multilingualism in Late Hellenistic Delos

(2) ΘΗ[. . .] // L(ucius) M(arcus) Mundicius // ΘΩΔ [. . .] ΜΟΥ (Le Dinahet-Couilloud 2001: 105–6; AE 2001.1797)

The Mundicii, their freedmen, and their slaves are also known from public epigraphy (1692, 1760, 1764, 1768, and 2612) and, according to Le Dinahet-Couilloud (2001: 105–6), the two brothers Lucius and Marcus may be identified with the masters (or patrons) of two slaves (or freedmen) who are part of the collegium of the Competaliastai in 100–98 BC (1760: Ξένων Μονδίκιος Λευκίου καὶ Μαάρκου) and 93 BC (1764: Ἀλέξανδρος Μονδίκιος Λευκίου κα ̣ὶ Μ ̣ ̣ α ̣ά ̣ρκ̣ ̣ο ̣υ ̣). A second graffito (Figure 8.3) is the work of C. N(e)rius Eros, a freedman of another gens whose presence in Delos is well attested: [Γ]άιος Νέριος and [Π]όπλιος Νέριος are among the subscribers of two public dedications (2432 and 2457 respectively); another freedman, C. Nerius M. l., commissioned a signum Volcani together with a group of freedmen and slaves (2440); and Nerius was also the name of the actor in ex. (1). The text dates to ca. 100–70 BC (Roussel & Hatzfeld 1909: 508–10) and it is an extemporary and rudimentary dedication drawn freehand on a marble plate that was part of an altar dedicated, among others, to Poseidon (3):

Figure 8.3 Graffito from an altar in the Agora des Italiens (From Roussel & Hatzfeld 1909: 508–10 fig. 4)

(3) C(aius) N(e)rius Eros Apollinei et Iouei et Neptuno Minerua et Mircurio | Ποσειδῶ[νος] | Ναυκλαρίου36 (Roussel & Hatzfeld 1909: 508–10; AE 1910.15; CIL I2 2233)

36

The third line of the inscription, with the epithet Ναυκλαρίου ‘protector of shipowners’, is not represented on the facsimile published by Roussel and Hatzfeld (1909: 508–10). The form, which remains a hapax, is a derivative of ναύκλαρος, whose spelling raises some issues. The most obvious explanation is a Doricism for Attic ναύκληρος ‘shipowner’, with ᾱ for η. But ναύκλαρος, which is also known from the Hesychian gloss ν 118 (ναύκλαροι· δήμαρχοι ‘naúklaroi: governors (of the people)’), cannot be separated from the Attic doublet ναύκρᾱρος, which is the original, non-dissimilated form according to Chantraine (1999). Thus, following Solmsen (1898), Roussel and Hatzfeld (1909: 509) regard ναύκλαρος as an ancient spelling that is ‘une sorte de transition entre ναύκρᾱρος et ναύκληρος’. However, these forms should not necessarily be put in chronological order (ναύκρᾱρος > ναύκλαρος > ναύκληρος), since ναύκλαρος can possibly result from a (para-)etymological overlapping between the other two.

197

Francesco Rovai

The forms Minerua for Mineruae and Mircurio for Mercurio are particularly interesting, as they attest a dative ending -a in the first declension and a spelling -irc- for -erc- that have often been discussed in relation to the Latin dialectal variation.37 Today, after Adams’s (2007) thorough review of the available epigraphic evidence for regional variation across ancient Italy, the first declension dative ending -a can be essentially classified diaphasically rather than diatopically, as a feature that characterised the register of religious dedications (cf. Adams (2007: 46–52), and references therein). The spelling Mircuri- for the name of the god Mercurius may instead be acknowledged as a genuine dialectal feature of the Republican Latin (see, among many others: Ernout (1905–6: 321; 1947: 47); Leumann (1977: 45)), although it cannot be attributed to any specific region in view of the scattered geographical distribution of the forms Mircuri-/Mirquri- (Adams 2007: 89), which occur four times at Delos (CIL I2 2233, 2239, 2240, 2504), twice at Praeneste (CIL I2 553, 564), once at Firmum Picenum (CIL I2 1920), and once at Lissa in Dalmatia (CIL I2 2295).38 Half of the eight cases are found in Delos, and this fact has been taken by Vine (1993: 161–73) as evidence for a particular regional characterisation of -irc- for -erc- (Vine (1993: 171): ‘essentially Campanian or Faliscan’), according to the traditional assumption of Hatzfeld (1912), (1919: 238–56) that the negotiatores operating in Delos in the second and first centuries BC were not from Rome but from Southern Italy, especially Campania.

37

38

The theonym Eisis has been supposed to exist between Minerua and Mircurio, but the reading is very doubtful and – the multicultural environment notwithstanding – any reference to Isis seems to be out of place here (cf. CIL I2 2233: ‘inter Minerva et Mircurio quae scripta sint, parum apparet. e, deinde i vel t; sequentia possunt legi pro sis (ut sit ex. gr. Eisis; neque tamen Isidem hic expectaveris); eadem pro ornamento in formam folii facto accipere malunt bull. l. c. [=Roussel and Hatzfeld (1909: 510); FR] et totum interpretari et Mercurio [sic]’). Both Ernout (1905–6) and Leumann (1977) compare Mircuri- with stircus (CIL I2 401, lex Luceriae), commircium (Vel. Long. 7.77.12–14 K) and Oscan amiricatud ‘sold/negotiated.abl.sg’ (Tabula Bantina). In any case, any hypothesis of Oscan influence upon the Latin forms should be excluded, as convincingly suggested by Untermann (2000: 85–6; and references therein) in his discussion of amiricatud and amirikum.

198

Multilingualism in Late Hellenistic Delos

However, as already stated (n. 27), such a claim is no longer acceptable. Thus, following Adams (2007: 90–1), who also discusses the form stircus for stercus of the lex Luceriae (CIL I2 401), and the passage of Velius Longus where commircium for commercium is attributed by Varro to the antiqui (Vel. Long. 7.77.12–14 K), -ircfor -erc- can be loosely assumed as a non-urban dialectal feature, though with ‘some uncertainties’ due to the fact that there is little evidence for the spelling that might have been current in Rome. In fact, as is often the case with early inscriptions, there are more attestations not only of the irregular spelling Mircuri- (see above) but also of the regular form Mercuri- outside Rome than in Rome itself: two are from Delos (CIL I2 2242, 2243), one from Praeneste (CIL I2 563), one from Puteoli (CIL I2 1618), and one is of uncertain provenance (CIL I2 579). There remains only one certain instance of Merqurio (CIL I2 992) from Rome, which testifies to an urban spelling -erc- against a non-urban spelling -irc- during the late Republic.39 However, some further elements in favour of the urban characterisation of -erc- may be probably inferred from a handful of forms that have the same root as Mercurius, that is, mercassitur from the lex Agraria (CIL I2 585, 71: 111 BC) and merkatus, which is attested three times in the Fasti (CIL I2 fasti: Iul. 14(15)–19, Sept. 20–23, Nov. 18–20). On the one hand, the anomalous morphological structure of mercassitur, with a synthetic passive future perfect in the place of mercatus erit (Lindsay 1894: 465; Leumann 1977: 623), is observed also in very few other examples that are clearly characterised in terms of register: turbassitur, in a law quoted by Cicero (Cic. Leg. 3.11); faxitur, in the 217 BC rogatio of the 39

The urban evidence might be yet more puzzling if, following Vine (1993: 168–9) the form mipc (CIL I2 2890), which is painted on a fragmentary base of a dish found in Rome, is brought into the discussion. Should the letter

be intended as a rho-shaped ‹r› and the whole form as the abbreviation of Mircurius, this instance would show that, as elsewhere, also in Rome Mercuri- and Mircuri- co-existed, thus undermining the hypothesis about regional variation. However, apart from any palaeographic considerations, it must be borne in mind that the place where the instrumenta domestica (dishes, vessels, vases, etc.) were found, most often does not coincide with their place of origin, so that mipc (whatever its reading) cannot be a priori attributed to urban Latin.

199

Francesco Rovai

pontifex maximus Lucius Cornelius Lentulus mentioned by Livy (22.10.6); and iussitur, in the solemn prose of Cato’s de Agri Cultura (Cato Agr. 14.1). Whatever their origin,40 these forms are restricted to legal and other high-register texts, in which archaisms are in frequent use. In view of their distribution and their extra-paradigmatic status, mercassitur and the like can therefore be seen as true, sporadic survivals of an archaic stage of the language. On the other hand, the three occurrences of merkatus in the Fasti exhibit the typical archaising convention of writing the letter before the letter . Thus, the erc-spelling of the root *merc- is represented by two forms, whose morphology (mercassitur) and orthography (merkatus) indicate them as belonging to the traditional language of law and official texts. This fact suggests: (a) that they are ancient; and (b) that they are representative of urban Latin, because these kinds of documents could not have been composed anywhere but Rome. On the whole, they can be a piece of evidence supporting the originally urban characterisation of the spelling -erc-, as opposed to the dialectal status of the alternative spelling -irc-. Final Remarks To summarise, the cosmopolitan population of Delos is an illustrative case study for investigating the complex and multilayered relationship between culture, ethnicity, identity, and language behaviour, as well as the communication strategies in intergroup and interpersonal relations. The issue is complex and multilayered because it can be approached at different levels: societal, intergroup, interpersonal, and individual. On the whole, the Delian community can be seen as a ‘community of practice’ à la Wenger (1998: 72–8), whose structure is based on ‘mutual engagement’, ‘joint enterprise’, and ‘shared repertoire’. The members are mutually engaged in a network of collaborative relationships that bind them together as a social entity, and they have a shared understanding of the joint enterprise that binds them together: the ties can be easily 40

For the discussion of turbassitur and other sigmatic futures, see De Melo (2007: 341 ff.).

200

Multilingualism in Late Hellenistic Delos

recognised in what Rauh (1993), in the title of his book, has called ‘the sacred bonds of commerce’ (of which we have seen an example earlier in this chapter). Thus, as part of their practice, they develop a repertoire of common resources that are used in the pursuit of their joint enterprises. Not only the various forms of civil and religious associations, but, at this level, multilingualism per se is one of these resources and functions as a cultural trait which characterises the entire community. The role of language as a marker of identity surfaces instead at a smaller-scale level, that of intergroup relations. There are, however, important differences between the two main ethnic, linguistic, and cultural groups of immigrants, the Italici and the Semites. When the Phoenicians and the Syrians make reference to their ethnic and cultural identity in a public context, acting under the patronage of their institutions (cf. earlier discussion in note 6 on the Tyrian Herakleistai and the Berytian Poseidoniastai; and the discussion before note 16: τὸ κοινὸν τῶν θιασιτῶν τῶν Σύρων), they never use their native languages – as far as we know, at least – but Greek. The Italici, on the other hand, present their ethnic and cultural identity not only in Greek but also in Latin. Such a difference depends on obvious differences of power relations obtaining between the groups: and not on the level of economic and social development (the Phoenician community was wealthy and its members were fully integrated in the social life of the island), but on the level of political authority, which was very supportive of the Italici, as Rome was rising as the new political and military power (and the citizenship of an Italian city was highly prized by the Phoenicians themselves; cf. Philostratus of Ascalon). While Greek continued to be a language of culture and power in the East Mediterranean, Latin was now able to stand on a par with it in the public and official domain. The single individuals (the Greek Augias who praises – or blames – an actor; the Roman and the Semite who undersign a common transaction; Marcus Minatius who took part in the association of the Berytian Poseidoniastai; the young Romans and Phoenicians who attended the Greek gymnasium together with young Athenians, cf. n. 7) were, at the same time, part of the larger community of the island and of their smaller ethnic 201

Francesco Rovai

community, and this multiple identity correlates with their multilingual competence. At this level, interpersonal aspects of communication probably overruled intergroup considerations, so that mutual comprehension became the most important issue, whatever the linguistic management of the interaction. Finally, with particular reference to the structure of the Greek– Latin bilingual individual repertoires that resulted from such a close language contact, it is important to highlight the fact that linguistic interference is not just a matter of two systems in contact, but of two dia-systems in contact, a set of varieties of Greek that interact with a set of varieties of Latin. Augias did not come into contact with ‘Latin’ tout court but, at least, with some written registers of Latin and with a spoken regional variety of Latin. And the Greek drafters of both public and funerary inscriptions, who transliterated a Latin ī or ĭ as either , , , or , were not simply ‘Greek speakers’, but they had a linguistic repertoire that included, among others, an innovative variety of Greek that was clearly characterised by iotacism.

202

C HA P T E R 9

I N T ER PR ET ES , N EG O T IATO RES AND THE ROMAN ARMY Mobile Professionals and Their Languages R A C H E L M A IR S

Q. Atilius Primus: Interprex, Centurio and Negotiator Sometime in the late first or early second century AD, the freed slaves of a man named Quintus Atilius Primus set up a monument to their late master – and, most probably, husband and father – on the Danubian frontier of the Roman province of Pannonia (AE 1978.635). Atilius had followed a varied career path as soldier, interpreter and trader. His memorial is concise but eloquent testimony to a long life lived on the move between places and languages: Q(uintus) Atilius sp(urius) f(ilius)1 Vot(uria) Primus inter

rex leg(ionis) XV idem (centurio) negotiator an(norum) LXXX h(ic) s(itus) e(st) Q(uintus) Atilius Cog(i)tatus Atilia Q(uinti) l(iberta) Fausta Priuatus et Martialis hered(es) l (iberti?) p(osuerunt) Quintus Atilius Primus, illegitimate, of the tribe Voturia, interpreter of the Fifteenth Legion, centurion of the same, trader, aged 80, lies here. Quintus Atilius Cogitatus and Atilia Fausta, freedwoman of Quintus, Privatus and Martialis, his heirs, freedmen, placed (this monument).

A thousand years later, the stone was reused in the construction of the Church of the Assumption at Boldog, Slovakia, 50 kilometres from the Roman fort at Carnuntum, where Atilius’ legion, the XV Apollinaris, had been stationed. It was uncovered during renovations 1

The abbreviation SPF might – in theory – be resolved as either Spurii filius (son of Spurius) or spurius filius (illegitimate son). I follow the analysis of Rawson, who points out that Spurius as a name was extremely uncommon by the imperial period (Rawson 1989: 29–30). Whether or not Atilius was himself illegitimate, the fact remains that his own children, and those of many of his colleagues at Carnuntum and other Roman army camps, were by definition illegitimate if they were born to a serving soldier. This is of central importance to the arguments presented in the present chapter.

203

Rachel Mairs

to the sacristy in 1976, before which it had been covered with plaster (editio princeps: Kolnik (1978); for further information on the church at Boldog and the discovery of the inscription, see, in Slovak, www .boldog.sk, accessed 1 June 2017). The church was constructed, in Romanesque style, in the eleventh or twelfth century. It is possible that the stone was found locally or was brought from the ruins at Carnuntum, a good source of building stone. No other Roman inscriptions are known from the area around Boldog. My principal interest here is in the title inter

rex (surely not interrex, fascinating though that would be) borne by Atilius. But it is impossible to understand the meaning and significance of this title in isolation, without considering the mobile, varied life and career of Atilius himself, and his historical and geographical context on the Pannonian frontier (on imperial Pannonia, see Kovács (2014)). As previous studies have recognised, Atilius is part of a small cluster of interpretes who appear in Roman military inscriptions in Central and Eastern Europe (see, most recently, Mairs (2012b); Baratta (2012); Checa Gómez and Storch de Gracia (2009); Peretz (2006); Gallego Franco (2003)). Most of these studies, my own included, are, however, local and descriptive rather than analytical and offering a wider historical perspective. That we find this concentration of interpretes along the Danube, and not elsewhere in the empire, is at first glance curious. The Roman army on the Danubian limes was multiethnic and multilingual, but then this was also the case throughout the empire. The provinces of Pannonia and Moesia, sources of the relevant inscriptions, were populated by non-Latin speakers, as were the lands inhabited by Germans, Sarmatians and others across the river – but the same was true of Roman frontiers in North Africa or Syria. The army camps along the Danube generated large numbers of Latin inscriptions, but there are other parts of the empire which are equally epigraphically rich. The most important point to be made at the outset is that this apparent lack of interpreters is in fact a lack only of epigraphically visible interpreters. Language mediators are notoriously ‘invisible’ in the historical record, oral interpreters all the more so than written translators (on this trope, see Venuti (2008)). Where speakers of different languages needed to communicate orally, 204

Interpretes, Negotiatores and the Roman Army

mechanisms would have existed to facilitate that communication, whether this meant language acquisition (communication by bilinguals) or the use of interpreters (communication via bilinguals). The former means of communication has received sustained scholarly attention over the past two decades (Briquel-Chatonnet (1996); Adams, Janse and Swain (2002); Adams (2003a) and numerous studies by these and other authors since that date). Language mediation merits another look. The case of interpreters in the Roman army offers a second opportunity, which is to look at language interaction within a mobile community. In addition to the problem of invisibility, investigations into interpreters and interpreting in the Graeco-Roman world are hampered by varying degrees of visibility among their subjects. By this, I mean the contrast between cases such as Atilius’ (and the other Danubian interpreters whom I shall discuss) and the situation in Hellenistic and Roman Egypt, where we have evidence to explore the participation of interpreters in fields such as commerce and the law over the longue durée, but little information on individual life trajectories, language use and geographical mobility (Mairs 2012a). This means that, even where we have evidence from the same time period, we cannot compare like with like in different regions and professional contexts. Some areas of the Graeco-Roman world are best suited to a case study approach to interpreters, others to a broader societal approach. The problem of reconciling micro and macro perspectives is a common one in the social sciences. A life course approach is one way of attempting to overcome this problem. Such an approach analyses ‘the dynamic interrelation of structure and agency . . . relating individuals’ life courses in their timing, pacing, and sequencing of life events (micro-level) to the dynamics of social structures and institutions (macro-/meso-level)’ (Wingens, de Valk, Windzio & Aybek 2011: 1, 6). The term ‘life course’ has previously been adopted in studies of Roman inscriptions, but almost exclusively with reference to the life stages of childhood, adulthood and old age (Laurence (2000); Revell (2005); Harlow and Laurence (2002)). More recent sociological trends which take a life course approach to migration, however, have yet to be fully exploited in studies of the Graeco-Roman world. Wingens, de 205

Rachel Mairs

Valk, Windzio and Aybek (2011: 2–3) propose ‘a sociological life course approach to migration [which] focuses on the dynamic interplay of societal structuring and institutional framing of migrants’ life courses and the pattern of migrants’ biographical mastering of transitions and coordinating of life spheres’. Such an approach is almost tailor-made for interpreters in the Roman army. Individual interpreters, such as Atilius, can be examined from the dual perspective of the structures of the Roman army and provincial administration, and the agency exercised by them as actors within this. The extent of their mobility is one of my main subjects of investigation. Mobility and Immobility in the Legions The legionary camp at Carnuntum was established by the Legio XV Apollinaris during the reign of Augustus. The legion was stationed there until sent east by Nero in around AD 63, and then again from around AD 71–114 (Wheeler (2000) may be consulted on the history of the legion at Carnuntum and elsewhere: he notes the frustrating lack of evidence for the legion’s precise actions and movements in wars against Dacians, Germans and Sarmatians; on the XV Apollinaris at Carnuntum, see Fitz in Lengyel and Radan (1980: 131, 230); Mócsy (1974: 43)). The presence of the name of Atilius’ tribe, the absence of Dis Manibus, the name of the deceased in the nominative case and the use of h(ic) s(itus) e(st) are indicative of a first-century date for his memorial (Keppie 1991: 28). Atilius had reached a venerable age, so we might expect that his retirement from the army and settlement near his old legionary camp predated the inscription by some decades. His active service could therefore have been either in the legion’s first stint at Carnuntum or its second. Since centurions did not necessarily retire after the standard period of legionary service, it may even have covered both. If this is the case, and he remained with the same legion throughout his career, he may also have served in Syria. Carnuntum is a rich source of Latin inscriptions, generated by the garrison there. These enable us to establish some context for Atilius’ career. Aside from his role as interprex, how typical was he of the local Roman military community in the first and 206

Interpretes, Negotiatores and the Roman Army

early second centuries? We have evidence of many veterans of the XV Apollinaris settling in the area after their retirement, some of whom lived lives as long as Atilius (e.g. CIL III 4482; Mosser 2003 No. 62). At least one, C. Aruntius, active at the same period as Atilius, surpassed him, dying at the age of eighty-five (CIL III 4455; Mosser 2003 No. 69). Although the inscription does not state this directly, Q. Atilius Cogitatus and Atilius’ liberta Atilia Fausta are clearly his son and marital partner. Veterans’ funerary inscriptions of the XV Apollinaris at Carnuntum frequently mention partners and children. The veteran C. Arestius Firmus, for example, is commemorated by his coniunx Vettia Sabina and their son (AE 1929.119; Mosser 2003 No. 68), and M. Antonius Capito by his uxor, Valeria Butto (AE 1929.219, Mosser 2003 No. 97). Often, as in the case of Atilius, we find a wife who is also the deceased veteran’s freedwoman, such as L. Armentius’ coniunx et liberta, Apta (AE 1929.200, Mosser 2003 No. 80). The names of these partners are sometimes non-Latin, such as Suadulla, the coniunx of the veteran C. Valerius Secundus (AE 1929.222, Mosser 2003 No. 167). That these partnerships often predated the veteran’s discharge from the army can be seen in the funerary inscriptions of the children of soldiers still on active duty. C. Cassius Aelianus, a centurion, commemorated his eight-yearold son and sixteen-year-old daughter (CIL III 14358). A miles of the XV set up a memorial to his five-year-old daughter, Marcellina (AE 1983.756, Mosser 2003 No. 116). Atilius’ family with his freedwoman, Atilia Fausta, is therefore a common pattern among retired soldiers of the XV at Carnuntum. The fact that the legion was stationed at Carnuntum for such long periods of time may have contributed to its veterans’ tendency to settle in the locality after retirement, although this is a pattern that we also see elsewhere. As for Atilius’ own origins, what we know of them, there were members of both his tribe, Voturia (AE 2002.1152, Mosser 2003 No. 104; Mosser 2003 No. 121), and his gens, Atilia (CIL III 11129), among the XVat Carnuntum, and at least one other soldier who belonged to both (CIL III 4463). Spurii filii are not common in the local epigraphic record (they are not common in the Roman epigraphic record tout court: Rawson 1989: 29), but do sometimes 207

Rachel Mairs

bear tribal affiliations, including Atilius’ own, Voturia (AE 2002.1152, Mosser 2003 No. 104). If we take, for now, Atilius’ tria nomina and tribal membership as unproblematic, then the evidence suggests that he was of northern Italian origin and thus probably a native speaker of Latin (see Adams 2003a: 276–7). This was true of many military personnel and civilian traders in first-century Pannonia (Lengyel & Radan 1980: 25, 332; Mócsy 1974: 59, 71). The Atilii were a prominent family of traders on the ‘Amber Route’ from the Adriatic to the Baltic (on the archaeology of the Roman Amber Route, see Mráv (2010–13) and Teicher (2013)). They were of Aquileian origin, like most such families, and numerous members are attested in inscriptions along the route (Gregoratti 2014: 143–4; Teicher 2013: 57; Wheeler 2000: 280–1). Not all these Atilii were citizens born into the gens; the merchant network involved many freedmen and agents who bore the name. As a spurius filius, we have no definite information on Atilius’ father. I shall return to the question of his illegitimacy below, in the context of marriage within the Roman army. Atilius lived at Carnuntum or in its neighbourhood for most of his life. Membership or association with the gens Atilia places his family origins either at Carnuntum too, or at another point on the Amber Route with which it was closely connected. This is of great significance to his role as negotiator, mentioned on his memorial. (The following section will consider the relationship between negotiatores and the Roman army.) It is also, I would argue, of pertinence to his role as interprex. Although it is intuitive to think of the Roman army as mobile, Atilius was not. If he did serve with the XV Apollinaris in Syria, he was absent from Carnuntum for less than a decade of his long life. It is his long connection with a single locality which positions him best as interprex and negotiator, familiar with – indeed, profoundly integrated into – a local community and its languages. If we consider the life courses of career soldiers of a similar rank to Atilius, we can find some striking contrasts. The centurion Petronius Fortunatus, for example, served fifty years, with at least fifteen legions throughout the empire, including a spell with the XVApollinaris (CIL VIII 217; ILS 2658; Lassère 1991). He too 208

Interpretes, Negotiatores and the Roman Army

died aged eighty, in the first half of the third century AD, and was buried at Cillium in Africa Proconsularis. [M(arcus) Petronius Fortunatus] / militauit annis IV in leg(ione) I Ita[lica] / librar(ius) tesser(arius) optio signifer / [(centurio)] factus ex suffragio leg(ionis) eiu[sdem] / militauit (centurio) leg(ionis) I Ital(icae) leg(ionis) VII / (centurio) leg(ionis) I Min(eruiae) leg(ionis) X Gem(inae) (centurio) leg(ionis) II A[diu(tricis)] / (centurio) leg(ionis) III Aug(ustae) (centurio) II[I] Gall(icae) (centurio) leg(ionis) XXX V[l]p(iae) / (centurio) leg(ionis) VI Vic(tricis) (centurio) leg(ionis) III Cyr(enaicae) (centurio) XV Apol(inaris) / (centurio) leg(ionis) II Par(thicae) (centurio) leg(ionis) I Adiu(tricis) / consecutus ob uirtutem in / expeditionem Parthicam / coronam muralem uallarem / torques et phaleras agit in / diem operis perfecti annos LXXX / sibi et / Claudiae Marciae Capitolinae koniugi karissimae quae agit / in diem operis perfecti / annos LXV et / M(arco) Petronio Fortunato filio / militauit ann(os) VI (centurio) leg(ionis) X[X]II Primig(eniae) / (centurio) leg(ionis) II Aug(ustae) uixit ann(os) XXXV / cui Fortunatus et Marcia parentes / karissimo memoriam fecerunt. Marcus Petronius Fortunatus served in the army for four years in the Legio I Italica as librarius, tesserarius, optio, signifer; he was promoted to centurion by the vote of the legion, served as centurion of the Legio I Italica, centurion of the Legio VI Ferrata, centurion of the Legio I Minervia, centurion of the Legio X Gemina, centurion of the Legio II Adiutrix, centurion of the Legio III Augusta, centurion of the Legio III Gallica, centurion of the Legio XXX Ulpia, centurion of the Legio VI Victrix, centurion of the Legio III Cyrenaica, centurion of the Legio XV Apollinaris, centurion of the Legio II Parthica, centurion of the Legio I Adiutrix; in the Parthian expedition he was decorated for bravery with a Wall and Rampart Crown and with necklaces and ornaments; he was in his eightieth year at the completion of this ornament for himself and for Claudia Marcia Capitolina, his beloved wife, who was in her sixty-fifth year at the time of the completion of this monument, and for his son, Marcus Petronius Fortunatus, who served in the army for six years, centurion of the Legio XXII Primigenia, centurion of the Legio II Augusta, lived thirty-five years; for their beloved son, Fortunatus and Marcia, his parents, built this as a memorial.

Petronius Fortunatus may be an extreme case, but my point is to highlight the relative immobility of Atilius and his legion in the first and early second centuries (on centurions’ origins, families and career trajectories in third-century Pannonia, see Fitz (1994: 1124–58)). By examining both the micro level of Atilius’ own life history and the macro, structural level of the Roman army, we may advance a hypothesis: that building a successful, valorised and thus epigraphically visible career as an interpreter may be 209

Rachel Mairs

connected to long-term residence in a single locality. Could it be that interpreters are precisely not mobile? Negotiatores and the Army Atilius’ inscription does not indicate whether his three stated professional roles – centurio, interprex and negotiator – were concurrent or consecutive. The two former positions were as part of his service with the XVApollinaris. Opinions differ on whether he was a soldier and a trader (perhaps in some regulatory or supervisory capacity), or a soldier then a trader (varying assessments in Pitts (1989: 48); Lee (1993: 67); Verboven (2007: 312)). Veterans often went into business (Wesch-Klein 2007: 445), but soldiers could also involve themselves in commercial activities on the side, and in direct army service. Indeed, stamps on Spanish amphorae from Carnuntum of the late first century AD name two active centurions of the XV Apollinaris who were agents for the trade of garum and other commodities (Wheeler 2000: 280–1; AE 1996.1251 and 1252). Atilius’ military connections and familiarity with the locality and local language(s) must surely also later have aided him in running a private enterprise; not least, as Verboven (2007: 312) notes, because the army discharge bonus provided a capital sum to invest in his business. On the other hand, his membership of the gens Atilia, a well-established Aquileian trading family, raises the possibility that he had been familiar with commercial activities all his life. Negotiatores, interpretes and the army could be bound up in any number of ways. Roman civilian negotiatores and their Greek colleagues on Delos in the second century BC mostly dealt with the language contact issues inherent in trading enterprises through bilingualism, a fact which is reflected in the complex intersections of languages in their epigraphic production (Adams 2002). In the first century BC, at least one of the interpreters of the disgraced governor of Sicily, C. Verres, is also described by Cicero as a negotiator, and language skills and expertise as a middleman went naturally together (Mairs 2011). Epigraphically attested negotiatores in Pannonia of the first century AD were often freedmen of Aquileian trading families, sent to manage regional 210

Interpretes, Negotiatores and the Roman Army

operations along the Amber Route (discussed in Mócsy (1974: 78); see also Künow (1980) on negotiatores in German lands outside the empire; freedmen negotiatores are also well-attested elsewhere, e.g. CIL VI 9664, from Rome). But the professional title negotiator also commonly appears applied to men who have dealings with the army, often supplying common military requirements like weapons and clothing, or belong to it themselves (e.g. a militis negotiatori, AE 1899.159, and CIL III 14927, Dalmatia; negotiatoris ueteranorum, CIL III 5308, Noricum; a negotiator and uestiarius commemorated by his brother, an aquilifer of the III Italica, CIL III 5816, Raetia; a veteran of the XXII Primigenia who was a negotiator gladiarius, CIL XIII 6677, Germania Superior; another veteran of the XXII Primigenia commemorating his brother, a negotiator, AE 1982.709, Gallia Lugdunensis; a soldier of the German fleet and negotiator ceruesarius, AE 1941.168, Belgica). An isolated instance of ‘interpreting’ in a commercial context related to the Roman army appears in the Vindolanda Tablets (Mairs 2012b; cf. a document from the Low Countries dating to the early first century BC: Bowman, Tomlin & Worp 2009). This, too, is within a military context, on a frontier. The tablet in question dates to Period 2 of Vindolanda (c. AD 92–7), and is therefore roughly contemporary with Q. Atilius Primus, at the other end of Rome’s European empire. There are, however, some important differences, which mean that we cannot consider the Vindolanda text as part of our Danubian group, or as a product of the same administrative and social context. Rather than interprex as a military or professional title, we find here interpretari as a verb. The context is clearly commercial. And the text concerned is documentary, a letter relating to the day-to-day business of the camp, not a stone epitaph which reflects the professional self-definition of an individual or the ascriptive choices of his commemorators. i 1. Curtius Super Cassio suo salutem ].[..].[..].[..].[

211

Rachel Mairs

ii 4. ut interpreteris et ut hordeum commercium habeant a te [ [. . .] . . . ḷ.ḅe.ṃ . . . uạ.ẹ 8. ]ẹ.ḍạ . . .. . .

Back 1. Cassio Saecu[lari] . . .. . . Curtius Super to his Cassius, greetings. . . . so that you may interpretari and so that they may get from you barley as commercial goods . . . (Back) To Cassius Saecularis (trans. adapted from Tab.Vindol. II 213, http://vindolanda .csad.ox.ac.uk/TVII-213).

Given Vindolanda’s position in Britain, and moreover on the northern frontier, it is tempting to posit some connection with interpreting with native Britons, locally or north of the wall. Whatever role military interpreters of indigenous languages may have played in Britain, however, this text is not evidence for them. As in Egypt, a person engaged in ‘interpreting’ – acting as a go-between – in a commercial sense may also have been capable of operating in more than one language (Mairs 2010), but this is not the sense of the present text. The evidence, then, in summary, points to the existence of an Aquileian network of negotiatores along the Amber Route, including at Carnuntum. As we have also seen, negotiatores could work within, or in close connection with the Roman army. The use at Vindolanda of interpretari in a commercial context – obscure though the linguistic component of this transaction may be – points to a conceptual and practical association between linguistic mediation and trading in the Roman army. It is therefore probable that Atilius had opportunities to use his interpreting skills in commercial operations both during his time in the army and after retirement. As a member of the gens Atilia, he was connected with 212

Interpretes, Negotiatores and the Roman Army

established Aquileian trading networks. It is likely that his business skills, like his linguistic skills, will have been honed during his army service. We might also note in this connection that the M. Ulpius Celerinus of one of the inscriptions considered below was a salariarius, responsible for provisions, and so may be another example of a Roman soldier for whom an interpreting and a trading or managerial role went hand in hand. Atilius’ Context: Later Roman Military Interpretes in the Danubian Provinces As I have already noted, Atilius’ inscription is part of a small concentration of inscriptions mentioning Roman military interpretes from the Danubian provinces. The other inscriptions are somewhat later and provide different kinds of information. None are from Carnuntum: one is from Brigetio in Pannonia Superior; two are from Aquincum in Pannonia Inferior; one inscription naming two interpretes comes from Viminacium in Moesia Superior: 1. M. Ulpius Celerinus (Brigetio, Pannonia Superior, sarcophagus of the late second–early third century AD: CIL III 10988; IDRE II 272; Deac (2013)); M(arco) Vlp(io) q(uondam) Romano mil(iti) praet(oriano) et primoscrinio2 praeff(ectorum) qui uixit an(nos) XXXV M(arcus) Vlp(ius) Celerinus sal(ariarius) leg(ionis) I Ad(iutricis) P(iae) F(idelis) interprex Dacorum uiuus sibi et filio suo s(upra) s(cripto) carissimo f(aciendum) c(urauit) For Marcus Ulpius, former Roman soldier in the Praetorian Guard, and department chief in the office of the prefects, who lived thirty-five years. Marcus Ulpius Celerinus, salariarius of the Legio I Adiutrix Piae Fidelis, interpreter of Dacians, while still alive took care of making (this monument) for himself and his dear son, who is written above. 2. M. Aurelius Flavus (Aquincum, Pannonia Inferior, tombstone of the late second–early third century AD: CIL III 10505); D(is) M(anibus) M. Aur(elio) Flauo m[il(iti) leg(ionis) II Ad(iutricis) du] pl(icario) [e]t interpetri3 Ge[rmanoru]m off (icii) co(n)s(ularis) 2 3

Sic for primiscrinio. Sic for interpreti.

213

Rachel Mairs et M. Aur[(elio) . . . an(norum)] XIII filio communi Aurelia Qu[i]aeta (?) marito et filio dulcissimis scribi in memoriam iussit. For the spirits of the dead: Marcus Aurelius Flavus, soldier of the Legio II Adiutrix, duplicarius and interpreter of Germans, on the staff of the provincial governor, and Marcus Aurelius [?], thirteen years old, their son in common – Aurelia Quiaeta ordered (this inscription) in memory of her aforementioned dearest husband and son. 3. Gaius (Aquincum, Pannonia Inferior, tombstone of the late second– early third century AD: CIL III 14349 5); . . .]PAIL[. . .] [. . .] et Iulio Dig[. . .] [. . .] patri paga[(no) . . .] [. . .] Vlpi(a) e Homenat[. . .] [. . .]n(a)e so(ro)ribus [. . .] [. . .] Gaius mil(es) leg(ionis) [II Ad(iutricis)?] [. . .] interprex S(armatarum?) e[x . . .] [. . . o] ffici(i) co(n)s(ularis) sib[i et] [. . .] parentibus H[or]- -mus filius eius . . . and to Julius Dig . . . his civilian father(?) . . . Ulpia Homenat (a?) . . . sisters . . . Gaius, soldier of the Legio [II Adiutrix] . . . interpreter of S(armatians?), from the office of the provincial governor, for himself and his parents . . . their son. 4 and 5 C. Valerius Valens and M. Ulpius Martialis (Viminacium, Moesia Superior, list of discharged veterans of the Legio VII Claudia reused in a building, after AD 195: CIL III 14507; IDRE II 308). l. 11: C(aius) Val(erius) Valens int(erpres) R(atiaria) l. 40: M(arcus) Vlp(ius) Martial(is) int(erpres) Sc(upis) Gaius Valerius Valens, interpreter, from Ratiaria.4 Marcus Ulpius Martialis, interpreter, from Scupi. In connection with this group, we might include, for the sake of completeness, a sixth interpreter from Germania Inferior: 6 C. Januarinuus Sextus (Ruimel, Germania Inferior, epitaph known only from a sketch drawing of 1718: CIL XIII 8773). C. Ianuarinus Sextus interpres uiuos sibi fecit. C. Januarinius Sextus, interpreter, made [this monument] for himself while still living.

This last inscription is a geographical outlier, more difficult to date and may not have belonged to a soldier. It was found at Ruimel (near ’s-Hertogenbosch in the southern Netherlands). The stone 4

I follow Wilkes’ (1999: 99, note 11) correction of the ‘R(emesiana?)’ of the edition in CIL to ‘R(atiaria)’.

214

Interpretes, Negotiatores and the Roman Army

was first recorded in the late seventeenth century and is apparently no longer extant. It cannot be dated precisely, and its brevity and the fact that it is known only from a sketch impede dating further. Of the datable Roman inscriptions from Germania Inferior, the majority are from the second half of the second or first half of the third centuries AD, with a smaller number dated to the first 150 years AD. The earliest is a votive inscription, dated to the mid-first century AD (Derks 1998: 87–9; Roymans 2004: 14, 64). Presuming that it is chronologically nothing out of the ordinary, this places the inscription of Januarinius in the same – very broad – period as those discussed here from the provinces along the Danube. Although we have no further information on C. Januarinus Sextus, I am inclined to agree with Lee (Lee 1993: 67) in classing him as another possible Roman army officer – especially since the inscription may be incomplete and further military titles and biographical details missing. It may or may not be a coincidence that we find a negotiator with a similar name, C. Julius Januarius, from Colonia Agrippinensis, dedicating along with other traders to the goddess Nehalennia on the Zeeland coast near ancient Ganuenta, around 115 kilometres from Ruimel (AE 2001.1464). The Danubian interpretes, including Atilius, have much in common with one another. With the exception of C. Valerius Valens and M. Ulpius Martialis, who appear in a list of discharged veterans, all are attested from funerary inscriptions. All the men named are from provinces along the Danube frontier, from Pannonia and Moesia, and have positions in the Roman army. Atilius belonged to the Legio XV Apollinaris at Carnuntum, M. Ulpius Celerinus to the I Adiutrix at Brigetio, M. Aurelius Flavus and Gaius probably to the II Adiutrix at Aquincum, and C. Valerius Valens and M. Ulpius Martialis to the VII Claudia at Viminacium. In most cases we do not have precise dates, but all can be attributed to the late first to early third centuries AD, in the context of Roman military expansion, consolidation of power, diplomacy and trading in the region.5 While 5

For the administrative and military history of the Danubian provinces at this period, the thematic bibliography in Mócsy (1974) is still invaluable for earlier material, with more recent treatments in English by Lengyel and Radan (1980); Ørsted (1985); Dise (1991) and Wilkes (2005). See Visy (2003) and Breeze, Jilek, Thiel and Visy (2008) for an

215

Rachel Mairs

these inscriptions have provoked a certain amount of comment,6 their full implications for language use in the Roman army and the position of ‘interpreters’ in the military hierarchy remain to be fully considered. Q. Atilius Primus had reached the rank of centurion, M. Ulpius Celerinus was a salariarius (although his son had held higher honours), M. Aurelius Flavus and Gaius served on the staff of the provincial governor, while Flavus was also a duplicarius. Only C. Valerius Valens and M. Ulpius Martialis do not appear to have held any more elevated position – on the assumption that it would have been recorded if they did – but were regular soldiers. Even across such a small number of texts, the spelling of interprex or interpres, with the errors interpetri and interrex, is not standardised, perhaps in fitting with the rarity of the use of the word in inscriptions. ‘Interpreter’, in three cases, stands alone, with no indication of the people or language concerned. Celerinus is an interpreter ‘of the Dacians’. It should be noted that it is a people, not a language, of which someone is an interpreter in the Roman epigraphic convention, perhaps suggesting the functionary’s wider role as a mediator or go-between. Flavus and Gaius have been restored as ‘of Germans’ and ‘of Sarmatians’ respectively. These restorations are educated guesses but reasonable ones: Germanic and Sarmatian groups were those with which the legions along the Danube had the most frequent contacts. Although the following discussion will proceed on the understanding that these restorations are correct – some lines of enquiry are too good not to follow up – the reader should bear these points of uncertainty in mind throughout. With the exception of the list of veterans recruited in AD 169 and discharged in 195, in which Valens and Martialis appear, these inscriptions may be dated only tentatively, but knowledge of the movement of Roman legions in the region helps us a little. The

6

introductory guide to the archaeology of the Danubian limes and Roman presence in the region. See Lőrincz (2010–11: 67–85) on the establishment of the legionary camp at Brigetio; 145–57 on the history of Legio I Adiutrix; and on the II Adiutrix: 158–74. Mócsy (1974: 199); Fitz in Lengyel and Radan (1980: 128); Pitts (1989: 48); Mócsy (1992: 62); Lee (1993: 66–7) (with references to literary sources on interpreters); Austin and Rankov (1995: 28–9); Elton (1996: 79–81); Wiotte-Franz (2001: 119–23); Southern (2007: 228).

216

Interpretes, Negotiatores and the Roman Army

turn of the second–third centuries was the height of Latin epigraphic activity in the Danubian provinces (Mócsy 1974: 199, 230–1). The Legio I Adiutrix, given the title Pia Fidelis under Trajan, constituted the permanent garrison of Brigetio more-orless continuously from the camp’s construction in c. AD 100. Burial in sarcophagi, however, such as the fine example on which Celerinus commissioned his inscription, became fashionable in the region only towards the end of the second century (Mócsy (1974: 237); the style of this particular sarcophagus connects it with the workshop at Sirmium: Barkóczi (1944–5); Adamesteanu (1951)). Flavus and Gaius died at Aquincum, main base of the Legio II Adiutrix from c. AD 106, after Trajan’s Dacian Wars. It is on this basis that I have supplied the name of the II Adiutrix in Gaius’ inscription. M. Aurelius Flavus and his family may derive their names from the granting of universal citizenship under Caracalla in AD 212, or, more fittingly, from the reign of Marcus Aurelius, who campaigned extensively against German and Sarmatian tribes along the Danube, providing a terminus post quem. None of these interpretes – with the exception of Celerinus, as will be discussed below – may be securely identified in other inscriptions, tempting though it is to play that game. I shall return to the question of their personal backgrounds and the significance of their names, after consideration of the military and ethnolinguistic context of the Danubian frontier. The Danubian Context The Danube was a frontier zone and all of the Pannonian and Moesian inscriptions considered here come from the vicinity of Roman military foundations along the river (on the Roman frontiers along the Danube, see the relevant sections of Whittaker (1994) and Whittaker (2004)). Interpreters may have been necessary for the frequent dealings with the peoples who approached and crossed this frontier, but also within the provinces themselves. I shall review the evidence for language use within the Roman provinces of Pannonia and Moesia, and weigh the possibilities that the Roman army interpreters attested in the region may themselves 217

Rachel Mairs

have had exposure to, and needed to use languages other than Latin in-province. Alongside this, we will assess the data available on these interpreters’ own personal backgrounds. These inscriptions are themselves some of the best pieces of evidence for the linguistic needs and resources of the Roman military along the Danube. The languages of the diverse German, Sarmatian, Dacian, and other groups with which the army interacted – sometimes peaceably, sometimes in open hostilities – had no written form at this period. Historical and literary sources are far more likely to ignore or be ignorant of the presence of interpreters than to record their work. The title interpres is recorded in these inscriptions because it is an army staff rank, and furthermore, as I will suggest, because these soldiers used their interpreting skills in particularly high-profile and critical contacts with external groups. Two of these interpreters, Flavus and Gaius, held appointments in the officium consularis at Aquincum, the seat of the governor of Pannonia Inferior. Interpreters are not attested in the epigraphic evidence for the officium elsewhere in the Roman Empire. Aside from the usual reticence of our sources concerning interpreters, the military nature of the officium consularis in Pannonia, along with the political events which made interpreters so useful here in the late second and early third centuries AD, are the most probable reason why they appear in inscriptions here but not in other provinces. Our only other source for interpreters in the governor’s office (although of an earlier period) is Cicero, who mentions them in the civilian provincial administration of Cilicia and Sicily (Mairs 2011: 71–3). His ‘friend and interpreter’ Marcilius, in whom he put great trust, is described as having served for a long time as an attendant (Cic. Fam. 13.54, 132 SB (Loeb II)). The interpreters employed by C. Verres in Sicily, in contrast, are models of the ‘bad interpreter’, used by Verres not just to mediate between Greek and Latin, but as agents in his crimes and scams (Cic. Ver. 2.2.44, 2.3.85, 2.4.58). Like Atilius, they were skilled multi-taskers; the difference is that Cicero tells us explicitly that they were both interpreters between languages and go-betweens. 218

Interpretes, Negotiatores and the Roman Army

The locally stationed military were natural sources of skilled citizen personnel to serve on the staff of the provincial governor (discussed by Rankov (1987, 33–4); Rankov (1999)). In secondand third-century inscriptions, we find positions such as secretaries (exceptores, notarii), book-keepers (librarii) and archivists (exacti) (Rankov 1999: 22). The officium consularis, in addition to its regular personnel, could include other officials with special functions, including haruspices, or those general dogsbodies who described themselves simply as milites of the officium (Rankov 1987: 54; Rankov 1999: 31). Six inscriptions from Pannonia Inferior, including our two inscriptions, record officials off. cos. The others are a miles of the I Adiutrix a responsis officii consularis (AE 1967.366 = AE 1969/1970.478); two adiutores officii consularis (CIL III 3510 and AE 1967.364); and a miles of the II Adiutrix who was a strator [‘groom’] officii consularis (CIL III 10315 = AE 2005.12). Co-opting military men to the governor’s staff was clearly standard procedure, and the language skills of our interpretes were as useful to the governor’s office as they were to the army. The military positions held by the interpretes considered in this chapter were as follows: • Q. Atilius Primus, inter

rex: centurion of the XV Apollinaris. • M. Ulpius Celerinus, interprex Dacorum: salariarius of the I Adiutrix. • M. Aurelius Flavus, interpres Ge[rmanoru]m: soldier and duplicarius of the II Adiutrix. • Gaius, interprex S(armatarum): miles of the II Adiutrix. • C. Valerius Valens and M. Ulpius Martialis, interpretes: (regular soldiers?]) of the VII Claudia.

As I have already noted, Atilius was in addition a negotiator, which may have been a military or a civilian position, and is intimately related to the fact that he was an interprex. Celerinus appears in another inscription from Brigetio (CIL III 10988), a dedication to Jupiter where he is described simply as salariarius legionis I Adiutricis, without mention of the fact that he was an interpreter. The salariarius was an officer who, from the time of Septimius Severus, was responsible for stores and provisioning (Roth 1998: 274). His son, M. Ulpius, was a person of some distinction: a member of the Praetorian Guard (effectively 219

Rachel Mairs

disbanded by Septimius Severus in AD 193 and re-formed with heavy recruitment from the Danubian provinces: Cassius Dio 75.2.6) and primiscrinius, a chief financial officer in the Prefectoral administration. Flavus was a duplicarius, on double pay. It is worth considering the languages an inhabitant of the Danubian provinces may have had exposure to, and the precise contexts in which communication with Germanic, Sarmatian and Dacian groups – the only ones named outright in these inscriptions – may have been necessary. There was considerable diversity among the legionary population itself, but a certain amount may also be said about language use among the local and neighbouring civilian population in the first few centuries AD, on the basis of the onomastic record and – far more tenuously – similarities in material culture with neighbouring groups (see esp. Mócsy (1974: 58ff); I do not imply that material culture may be used as any kind of direct indication of language use, merely that cultural communalities may suggest that groups had more in common than their material toolkit). Tacitus’ (Germania 43.1) reference to a separate Pannonica lingua stems from a confusion (Mócsy 1974: 59); speakers of Celtic and Germanic languages predominated in the north of the region, with Thracian, Dacian and Illyrian influence further south. In any case, the population of Pannonia, in particular, had had long-standing contacts with the Romans, Roman culture and the Latin language by this period. As early as AD 6–9, Velleius Paterculus’ account of the PannonianDalmatian revolt claims the complete penetration of Roman ways at all levels of Pannonian society (2.110.5). Although this report should of course be taken with a pinch of salt, and probably limited to restricted groups, in particular in the south of the later Roman province of Pannonia, the ‘Romanisation’ and ‘Latinisation’ of the region over the next 200 hundred years must have been considerable (Mócsy 1983). Furthermore, the Danube, the line of the Roman limes, represented only a relatively porous border at most periods, and there were movements of Germanic, Dacian and Sarmatian populations in and out of the region – not least in the form of frequent raids or military incursions. Although the languages spoken by settled groups inside the Danubian provinces, even by their own troops, could well have 220

Interpretes, Negotiatores and the Roman Army

called for similar linguistic expertise, the known context of Germanic, Sarmatian and Dacian incursions, along with the apparent prestige accorded these particular interpreters, suggests that they were involved in interpreting and diplomacy at a higher level than everyday communication with the rank-and-file. (Who, one suspects, may well often simply have had Latin orders shouted at them until they understood – or at least complied.) Dealings with external ‘barbarians’ dictated more about Roman military organisation in the Danubian provinces than just the employment of interpreters. Threats from Germanic and Sarmatian peoples were, in general, key in deciding the location of forts, such as Aquincum, and the deployment of legions and auxiliary units (Visy 1995: 90), but it is also possible to note some specific episodes which may have made interpreters necessary, or simply more prominent. If an interpreter had played a key role in negotiating or managing intelligence in one of these incidents, we might even suggest that this would make it more likely that he or his heirs would choose to record the role – normally undocumented – in an epigraphic context. It may seem strange that M. Ulpius Celerinus, the Dacian interpreter, should be recorded at Brigetio, in Pannonia. Pannonian forces were commonly used in Dacian wars throughout the second and early third centuries AD (Fitz in Lengyel & Radan 1980: 137). But more compelling is the suggestion in the initial publication of the inscription that the need for an interpreter of Dacian at Brigetio is to be connected with an influx of Dacians around the Danube Bend in the late second and early third centuries AD, especially under Septimius Severus and Caracalla (Barkóczi 1944–5; cf. Adamesteanu 1951; Lengyel & Radan 1980: 102–3). Even this broad level of precision is difficult to achieve for the other two groups whose names are to be tentatively restored, Germans and Sarmatians. Contact and conflict with these groups along the Danube limes was more or less continuous throughout the period represented by our inscriptions. Most prominently, under Marcus Aurelius campaigns were fought against the Germanic Marcomanni and Quadi over an extended period (c. 167–80 AD). Conflicts in the region, and incursions into Roman territory, often involved a combination of various German 221

Rachel Mairs

and Sarmatian groups, not always well differentiated by our sources (see again Lengyel and Radan (1980: 102–3), on a dizzying combination of Dacians, Quadi, Sarmatian Iazyges, Carpi and Vandals in troubles in the region in the early third century and Wilkes (1983) on longer-term relations with Dacians, Sarmatians and Germans). So there would have been an ongoing need for army personnel capable of speaking the languages of these groups, whether in war or peace. If we cannot say much about the precise circumstances in which our interpreters used their languages, nor about their personal linguistic backgrounds (beyond noting, as we will in the following section, that their families’ enfranchisement within the past few generations is at least suggestive), then it is possible to say a little more about their position and status as interpreters. The rarity of the title ‘interpreter’ in the epigraphic record does not, I reiterate, suggest that interpreting was not happening on a regular basis. But it does say something suggestive about the circumstances in which an individual might be specifically designated as such, and choose to emphasise this status. ‘Interpreter’, in all the cases considered here, is a military title, in two instances also associated with service in the office of the provincial governor. Given the range of other titles cited, it seems clear that these particular interpreters – with the possible exception of C. Valerius Valens and M. Ulpius Martialis, who bear no additional titles – were comparatively high up in the food chain. It would be a mistake to assume that all their position involved was rendering information in one language into another. They were more than simply language technicians; linguistic abilities perhaps were even considered auxiliary to their role as active participants in negotiations or intelligence gathering with non–Latin-speaking peoples, inside or outside the empire, or interaction with non–Latin-speaking troops in the Roman army. We might suggest that it is this role which makes the title of ‘interpreter’ worthy of recording in an epigraphic context: the role of interpreter is professional, but language ability is not. Furthermore, all of these interpreters are men of substance, those with the requisite socio-economic status to commission an inscription. The possible exceptions, again, are Valens and 222

Interpretes, Negotiatores and the Roman Army

Martialis, who are listed amongst a large number of discharged soldiers. As well as being able to pay for an inscription, these men will also have been complicit in the mechanisms of culture and power associated with the Latin epigraphic act. They were all in the army, the major source of inscriptions in the Roman provinces, and they evidently, to whatever extent, associated themselves with display through the Latin language and Latin literacy. ‘Interpreter’, like any other official title, is a mark of position and status in a Roman military or bureaucratic context. Interpreters without an official remit, or who did not personally associate themselves with the Latin language and Roman culture, by definition do not appear in the Latin epigraphic record. This does not mean, however, that interpreters of primarily Dacian, Sarmatian or German cultural identity did not play an important role in diplomatic and economic transactions between the Roman authorities and local populations. It is simply the case that a Sarmatian interpreter, especially if employed principally by Sarmatians, is not going to have commissioned a Latin epitaph. The Personal Background of Interpreters An obvious and attractive proposition is that the interpretes among Roman military personnel in the Danubian provinces were bilingual from their primary socialisation, drawn from local speech communities. This is what I have already proposed for Q. Atilius Primus, the earliest of our attested interpreters in the region. As was noted above, however, we really know very little about the linguistic landscape of the Danubian provinces in the first three centuries AD, and certainly not enough to make even educated guesses about the languages to which any particular individual will have been exposed. It is highly probable that most of the interpreters discussed in this chapter were ‘army brats’, drawn from local veteran colonies, and citizens of only a few generations’ standing. Roman army camps and veteran colonies may not automatically have been a good source of native speakers of languages other than Latin. I shall support and qualify this statement in what follows, through discussion of the names of our interpreters, the establishment and ‘culture’ of Roman military 223

Rachel Mairs

settlements in the region, and the evidence for intermarriage with local women and its linguistic implications for Roman soldiers. First, nomenclature and citizenship. Their names suggest that the families of most, if not all, of our Danubian interpreters were relatively recent Roman citizens, but that they were citizens of long enough standing for their identity and language use to be thoroughly Roman. Those named M. Ulpius (Celerinus and his son, and Martialis) must have been descended from men granted Roman citizenship under Trajan in the early second century AD. It is possible that the Vlpi(a)e of Gaius’ inscription should be read as the name of a female relative of another such citizen, although the fragmentary state of this inscription makes drawing such inferences difficult and perhaps ill-advised. Similarly, the family of M. Aurelius Flavus must have been enfranchised under Marcus Aurelius, Commodus or Caracalla, somewhat later in the late second or early third century. Q. Atilius Primus, our earliest interpreter, is an exception and seems to have been of Aquileian origin. ‘New citizens’ are sporadically attested in the Danubian provinces as early as the first half of the first century AD (e.g. C. Julius Macimarus of CIL III 3377; the cognomen Macimarus/Magimarus is local and occurs on four other inscriptions from Pannonia Inferior: CIL III 3644 = 10576 and 10352, Barkóczi and Mócsy (1972: 202), and Noricum: CIL III 5272). Citizenship and enrolment in the legions were extended during the course of the second century, allowing local peoples and other non-Italians to occupy gradually higher positions in the military and the administration (on this process, see Lengyel and Radan (1980: 25, 100–7, 125, 134, 145)). Imperial gentilicia, such as Ulpius, become ubiquitous in the epigraphic record: our interpreters are not unusual in this regard. By the end of the second century, if not before, the majority of the legionnaires attested in inscriptions from the Danubian provinces bear ‘colourless names typical of soldiers’ (Mócsy 1974: 157). Participation in Roman institutions such as the army, the adoption of imperial gentilicia upon receiving citizenship, and the use of the Latin epigraphic medium in themselves ensure that much ‘local colour’ is lost from descriptions of such individuals. But while we may not be able to make definitive statements about 224

Interpretes, Negotiatores and the Roman Army

the linguistic or ethnic background of many – or indeed most – individuals in the epigraphic record, we can safely work on the assumption that implication in Roman mechanisms of power and display according to Roman cultural norms does not tell us the whole story. In fact, we may know more about the hidden ethnolinguistic potential of men such as C. Valerius Valens and M. Ulpius Martialis, interpreters from Ratiaria and Scupi, respectively, than we think. The VII Claudia, from which these men were discharged in AD 195, had been based at Viminacium, in Moesia Superior, since the mid-first century AD. It provided most of the veterans who founded Scupi, M. Ulpius Martialis’ home town, in the 80s AD. These veterans comprised both Italians and provincials, a heterogeneous population from the start, and Latin was probably important as a lingua franca among them (Mócsy 1974: 116–7). Legion and colonies mutually supplied one another with recruits and settlers. The origins of 142 of the soldiers listed in CIL III 14507 are preserved. Just over a hundred came from within the province of Moesia Superior, and except for a small number recruited at the legion’s base at Viminacium, they were from one of the province’s two veteran colonies: Ratiaria (a Trajanic foundation) or Scupi (which was Flavian).7 These continued to be an important source of legionary recruitment in Moesia Superior. Valens and Martialis may have good Roman names – even if not all their ancestors did – but many of their comrades listed in the same inscription have Thracian cognomina alongside their Roman gentilicia (e.g. Titus Aurelius Dizo and Titus Aurelius Drigissa). More than one third of the soldiers in the lists are Aurelii, indicating that they received citizenship upon enlistment, with 22 Valerii and 12 Ulpii. This list of discharged veterans, of course, names only men: the generations of army wives and mothers in places such as Scupi and Ratiaria, we should suspect, did not number many native Latin-speaking Italians among them. The other inscriptions discussed in this chapter also come from towns which, whatever their subsequent growth and receipt of municipium or colonia status, started out as army camps. As well 7

Wilkes (1999: 99).

225

Rachel Mairs

as large numbers of Italians, this brought soldiers and civilians from all over the Roman Empire to the Danubian provinces. Aquincum’s population represented mercantile interests from across the empire (Lengyel & Radan 1980: 332). Two of the legions in the second century with the greatest number of African soldiers were the I Adiutrix at Brigetio and II Adiutrix at Aquincum, although these still represented a modest total overall (Di Vita-Evrard 1995). There was a large Syrian community in Pannonia, following the XV Apollinaris’ return from its Eastern campaigns in the 70s AD. So the military and civilian population of the Danubian provinces exhibited considerable diversity. There are many ways a soldier from this region could have acquired the linguistic abilities necessary to act as an interpreter: he may have learnt a new language on the job in the course of his contacts with soldiers from non–Latin-speaking backgrounds, or he may himself have been a speaker of Latin as a second language (see Adams (2003a: 599–623) on Latin in the Roman army in Egypt). All this, unfortunately, does not bring us much closer to answering the question of whether the epigraphically attested interpreters from the Danubian provinces were native speakers of the regional Germanic, Dacian or Sarmatian languages they used professionally. The fact that many were recent citizens, whose families were enfranchised under emperors with a particular interest in Pannonia, Dacia and the Danube Bend region in general, raises the question of interpreters being recruited from individuals who, despite their thorough integration into Roman hierarchies and institutions, were the products of a longer-term bilingual process. How long such bilingualism may have persisted among ‘Romanised’ (I use the word with caution) segments of the populations of the northern and eastern European provinces is a difficult question (for the sparse evidence on Germanic, Getic, Sarmatian and Thracian bilingualism with Latin, see Adams (2003a: 274–9 and 283–4)). Could the wives and mothers of these men be the missing linguistic link, transmitters of local languages to their ‘Roman’ sons? This is an attractive proposition, although one for which it is difficult to find firm evidence. Q. Atilius Primus was illegitimate (if the restoration spurius filius holds: see note 1 above), and in any case soldiers were until AD 197 forbidden to marry (see Rawson 226

Interpretes, Negotiatores and the Roman Army

(1989: 29–38) on spurii in the epigraphic record). The marriage ban did not prevent soldiers from forming ‘socially legitimate marriages’ (the term used by Phang (2001)) with free or slave women and having children, nor from having children by slaves or other women with whom they had no intention of forming a family unit (on soldiers’ families, see Scheidel (2007)). Our information on such marriages and their children, however, must derive from the epigraphic record, with all the biases and constraints of representation inherent in this: inscriptions represent only those with the resources and inclination to commission stone texts in Latin, and they do so in very formulaic ways. One such possible epigraphic bias, which prevents our investigation into marriages across language boundaries in the Danubian provinces from proceeding much further, is the fact that soldiers in Latin inscription are only sometimes associated with wives or partners with nonRoman names. ‘In the epitaphs we find that only about ten percent of the soldiers’ “wives” have demonstrably native (neither Roman nor Greek) names. . . . The epitaphs suggest that the “locals” the soldiers intermarried with were already Romanized or Hellenized; they were probable Roman citizens or Junian Latins. In the inscriptions, the women’s Latin names are such “colorless” names as are common among the soldiers themselves. . . . Such “colorless” names may mask a native ancestry, but do not show when the person’s family was Romanized’ (Phang 2001: 191). By the mid-second century, citizenship was in addition widely spread in the region. Freedwomen, of course, bore names derived in whole or in part from their former masters. We therefore have exactly the same problems in accessing any potential ‘native side’ to women’s identities as we do to their husbands and fathers, especially if these women were themselves filles du régiment, probably a frequent source of marriage partners for soldiers (Whittaker 2004: 137–8). In the majority of cases, soldiers’ women appear under Roman names, where they appear at all. This can be put down to the rapid spread of citizenship in military communities through natural reproduction, enfranchisement of auxiliary veterans and manumission of slaves by Roman citizen owners. There were thus many ways in which the Danubian interpreters could have been exposed to other languages through 227

Rachel Mairs

their family background and professional lives, even if Latin was their dominant language or mother tongue. Mobile and Immobile Interpreters Army interpreters are unattested, to the best of our knowledge, in the epigraphic evidence from other Roman provinces. In the case of the interpreters from Pannonia Superior and Inferior, Moesia Superior and Germania Inferior considered here, I argue that the context which has led to interpreters being recorded is (1) the epigraphic production typical of the army, and use of specifically military interpreters on the officium consularis in Pannonia; and (2) specific political and diplomatic circumstances over a relatively restricted period of time in the late second and early third centuries AD which brought interpretes to prominence. Point (1), I would suggest, goes a little way towards explaining why interpreters who served on the officium consularis in other provinces – as surely they did – are not attested epigraphically. Soldiers, especially at this period, are our most copious source of epitaphs, civilians less so. Cicero’s interpreter Marcilius, who is described as an apparitor, was a civilian. As was noted above, most of the interpreters commemorated in these epitaphs also held positions above that of regular soldiers. There were surely soldiers who worked as interpreters who held more lowly ranks, but they simply do not record their work as such in stone. It is also important to emphasise the Roman epigraphic context because of the possible tension between the official connotations of describing oneself as an ‘interpreter’ in a Latin inscription at this period, and the more complex personal identity such an individual might possess. As discussed above, aside from their simple presence in the frontier provinces, we have hints that these interpreters may be of comparatively recent non-Roman origin. But again, the commissioning of a Latin inscription is a formal act, and in this particular context it was evidently not the place to expand upon any non-Roman elements of one’s ethnolinguistic background. These interpreters, especially those from veteran colonies, had probably had exposure to Latin since childhood. But there is a hidden ethnolinguistic potential, in their descent from and relationships with Germans, Dacians, Sarmatians and other ethnicities, slave and free. 228

Interpretes, Negotiatores and the Roman Army

The inscriptions discussed here make it clear that an interpreter’s position was concerned with more than just linguistic ability. In fact, none of these inscriptions explicitly mention language use or the languages spoken by the individuals concerned. Where further information is given, we are told that someone is an interpreter ‘of a people’, not specifically of a language. It might be argued that, since none of these inscriptions state competence in another language, these might simply be negotiating positions. But the use of interpres in other contexts, as discussed above, makes it clear that its primary association was linguistic. ‘Interpreting’ with Dacians or Sarmatians will also inevitably have involved crossing a linguistic boundary in some sense, and these men, from the Danubian provinces of the Roman empire, will have been just the kind to have possessed the necessary linguistic abilities. That they did, as I have already suggested, has much to do with personal and institutional mobility – or lack of mobility. Although the legions moved from place to place, they could be stationary for long periods of time, such as with the Legio XVApollinaris in the first and early second century. The Legiones I and II Adiutrix had also spent decades based in Pannonia by the time of the inscriptions of interpretes from Brigetio and Aquincum. The VII Claudia had a long association with Viminacium, although it spent much of the second century on the move in campaigns against Parthia, as well as on the Danubian frontier, but the places of origin listed for the two interpreters discharged at Viminancium in AD 195 show that they were recruited relatively locally. The life courses of these individuals – decades spent in and around the same Danubian forts, near which they may also have been born – may be contrasted with the linguistic needs of a military force which could and did move, and of a provincial administration, many of whose top officials came from Rome or other provinces of the empire. Both the latter institutions were administered in Latin. One of the reasons why interpretes were important and epigraphically visible on the Danube frontier is precisely because mobility worked at different ways on the micro and macro levels. The individual life courses of these interpreters were relatively stationary, within larger patterns of institutional mobility across the whole empire. 229

CH A P T E R 10

HO C P RI MV S V E N IT Italians and Others in Egypt before the Caesars PATR IC K JAME S

Some accidents of preservation, reuse, discovery, and publication are fascinating. It so happens that the earliest Latin inscription known from Egypt claims, it seems, that, on 26 August 116,1 one Acutius2 was the first (Italian or writer of Latin) to reach the sanctuary of Isis on the island of Philae in Upper Egypt and to leave a legible mark in Latin (I.Syène 321). For him to be able to make such an absolute claim, we must assume that no Latin was visible amid the many Greek inscriptions then at Philae and that there were no indications of Italians in those Greek inscriptions. If others had preceded him without leaving written evidence, they were as invisible to Acutius as they are to us. Or, Acutius may have made a relative claim: to be the first, in some respect, in relation to those who accompanied him that day and also incised their names. We can only speculate about just how competitive Acutius was in relation to his companions. The inscription, in fact, consists of one Greek graffito in Greek script and of three Latin graffiti in Roman script. The content of the Greek partially overlaps with the first Latin graffito to form a biversion bilingual of a kind. The texts are presented below as they appear on a block that seems to have been part of a Ptolemaic structure at Philae, was reused as part of the western exterior wall, and whose inscription was discovered in 1976. The edition is that by André Bernand (1989: 303–4 and Plate 154.2), but with interpuncts present and abbreviations left unresolved.

1 2

All dates are BCE, unless otherwise indicated. References to individuals by their personal name are given in Roman type. Personal names are italicised when they are discussed as personal names independent of a particular bearer of that name.

230

Italians and Others in Egypt before the Caesars ACV [TI ]VS [—] F·TEṚ HOC·VENIT·PRIMVS A·D·V·K·SEPTEMBRIS Q·FABIO·C·LICINIO·COS ΓΑΙΟΣΑΚΟΥΤΙΟΣ

M·CLAVDIVS·VARVS SP·VARAEVS·N·F HOC·VENERVNT A·D·V·K·SEPTEM·COS ROMAE·Q·FABIO·C·LI

Acutius son of [— – —] TER3 arrived here first on the twentysixth of August when Q(uintus) Fabius and G(aius) Licinius were co(n)s(uls). Gaios Akoutios. M(arcus) Claudius Varus Sp(urius) Varaeus son of Numerius they came here4 on the twenty-sixth of August when Q(uintus) Fabius and G(aius) Licinius were co(n)s(uls) at Rome.

The nature of Acutius’ claim to be primus is not clear, given that others came with him to Philae on that same day. A key piece of the puzzle is the value of TER. An indication of tribal affiliation, Ter(entina tribu),5 which would emphasize the Italian origin of Acutius with greater specificity than an ethnic, would be very unusual by this date. On the other hand, the cognomen Ter(tius)6 would balance that of Marcus Claudius. J. E. Bingen has argued7 against TER being an abbreviation on the basis of the space that follows and because Acutius must be distinguishing himself from the others at Philae with him that same day by being the first of the group to come to Philae three times. The purpose of Acutius’ visit to Philae with his companions remains uncertain. The precise date provides clues about why these three men had come to Philae. Ptolemy VIII Euergetes II 3 4

5 6 7

Since the value of TER is debated, it remains untranslated for now. The subjects of this third-person plural verb are Marcus and Spurius, although the name of the former, inscribed by a different hand and in larger letters, is isolated from the latter and from the verb. The subject also may include Γάϊ ος Ἀκούτιος, although hand, script, and language separate him from the verb. Beness and Hillard (2003: 203). Bernand (1989: 303–4). Bingen (1979: 305 note 3, 2002: 751, and 2004: 689).

231

Patrick James

had died less than a month earlier, and late August was the time of the festivities related to the high point of the surge of the Nile reaching Philae. On the basis of the first of these coincidences, we may speculate that their presence was related to international relations and current affairs, at least in part. The second coincidence suggests religious or cultural tourism as the specific motive for this visit to Philae.8 We may expect any visit to Philae to have a religious dimension, but, as noted by J. Lea Beness and Tom Hillard (2003: 206), this inscription does not mention any act of proskynema to Isis, unlike inscriptions at Philae in general. Also, the particular date and year become less important for the interpretation, if this visit was not the only one that Acutius ever made, but the third. The dates of his earlier visits are lost to us, but, if Acutius had already come to Philae repeatedly, we have every reason to suspect that he would do so again regularly and not necessarily for a single purpose. As Beness and Hillard (2003) noted repeatedly, commercial motives would be compatible with the other purposes, such as sacred tourism or a diplomatic mission, that have been proposed for this particular visit or for Acutius coming to Egypt in general. These three men all left their mark in Latin and in Latin script and from a Roman perspective. The year was indicated by the consular formula and the day and month by the Roman calendar, with a Roman numeral, rather than by either or both of the Egyptian and Macedonian calendars in use at Philae. Spurius Varaeus added that Quintus Fabius and Gaius Licinius were consuls at Rome, announcing the group’s ‘distant provenance’ and ‘cultural perspective’.9 Only Acutius made a partial concession to Greek-reading visitors to Philae: his nomen was transcribed into Greek script (with Greek morphology) with the addition of his praenomen, omitted, it seems, from the Latin.10 His companions 8 9

10

Beness and Hillard (2003: 205 and 206–7). Beness and Hilliard (2003: 204 note 7). The inversion of the consuls may reflect peculiarly Roman concerns of Acutius and Varaeus: either ‘partisan preference’ for Fabius or strict adherence to the consuls’ alternating tenure. See Beness and Hillard (2003: 204 with notes 8–10). The praenomen Gaius could have been incised above the first Latin text. However, since the first line, as preserved, is at the top of the block as it was first hewn (it seems), we may wonder why Gaius or, if, like his companions, he abbreviated his praenomen, C alone was written above the join of the blocks. The text could have been started lower, since there was plenty of room below on the block.

232

Italians and Others in Egypt before the Caesars

indicate their filiation. Of the three, however, only Marcus Claudius Varus certainly gave his tria nomina. If TER is an abbreviation for Tertius, the tria nomina of Gaius Acutius Tertius can be assembled from the Greek and Latin texts.11 It is certain that the name of his father was included. The ‘tone’ of this Latin inscription distinguishes it from the rest of the inscriptions at Philae. The absence of any reference to Isis or any other indication of cult activity has already been mentioned. Although the verb ‘come’ recalls the verb ἥκω of Greek inscriptions at Philae, this inscription contains two third-person forms, uenit and uenerunt. However, at Philae the first-person singular is the norm.12 I.Syène 323, another Latin inscription at Philae, thought to be contemporary with I.Syène 321, shows the expected first person ueni: M(arcus) Tiṇọtrius N(umeri) f (ilius) Graicanius Nucrinus ueni. I, M(arcus) Tiṇọtrius Graicanius, s(on of) N(umerius), of Nuceria, visited.

(As Bernand (1989: 307) noted, the ethnic Nucrinus13 here is that of Nuceria in Campania, and his cognomen, Graicanius, must be derived from Graecus and indicate Greek associations of some kind.) The combination of Latin and a third-person verb in later inscriptions on the Colossus of Memnon was noted to have ‘a certain detachment’ by J. N. Adams (2003a: 549 and 551), who concluded that it conveys a ‘public or official voice’ in contrast to the ‘literary voice’ of the first-person verbs in the Greek verses that accompany third-person Latin statements of facts and details. 11

12

13

If the Latin text originally included Gaius, the praenomen would have been incised in isolation on a different block, assuming that the block was not resized when it was reused. Bernand (1989: Plate 154.2) shows that there was space on the stone for before Acutius. The exception is I.Syène 308.1–2 (third century): Ἀνδρόνικος ἥκει | πρὸς τὴν Ἴσιν. Some instances of third-person forms are found at other sites: Abydos 87.3, 357.2, 535.1 (Ptolemaic), 543.2, 627.5 (Ptolemaic), and 629.1 (Ptolemaic), and Syringes 205.2, 398.2, and 1949.2. The usual form of this ethnic is Nucerinus. For this unique instance with syncope (in -eroin particular), compare Buck (1928: 58).

233

Patrick James

This inscription is a convenient entry point for this study of the linguistic vestiges of Romans and Italians in Egypt before it became the Roman province of Aegyptus. It testifies that Romans were present in Egypt and as far as south as Philae by the end of the second century. However, we are left uncertain about who these Romans were and why and how often they came to Philae. Still, the inscription remains remarkable as a statement of primacy, in Latin with only Greek Γάϊος Ἀκούτιος as the slightest concession to those unable to read Latin or understand Roman dates. It is assumed that there would be subsequent Latin-reading visitors to see it, while for others, unable to read its Latin, its Latin majority had a symbolic function, announcing a non-Greek and non-Egyptian presence at the sanctuary. More questions are raised than the inscription itself answers, both about its context and, more generally, about Romans and Italians in Ptolemaic Egypt. This study is concerned with the epigraphic, the papyrological, and, chiefly, the onomastic evidence that may be brought to bear to address such questions. Methodology and Problems with the Evidence As scholars writing on Latin in Egypt have noted, ‘Latin did not arrive only with the conquest’ (Evans 2012: 516) and there must have been ‘numerous Latin speakers who came to Egypt’, ‘numerous Egyptians who travelled to Latin-speaking areas’, and ‘contact with Romanized Greeks’ from outside Egypt (Dickey 2010: 219–20). There are almost no Latin texts from Ptolemaic Egypt: the two inscriptions at Philae that we have already considered, two-and-a-half graffiti among the thousands from the Syringes at Thebes and a Greek inscription followed by a short Latin addition (OGI 1.129). That is the whole corpus of Latin from Ptolemaic Egypt published so far. This entire corpus is examined in this study. A Greek inscription from the end of the Ptolemaic period (I.Philae 1.63) contains the sole Ptolemaic attestation of a Latin loanword. Otherwise, the few occurrences of loanwords in the first century are restricted to the Augustan decades of that century and then to just two documents: P.Oslo 234

Italians and Others in Egypt before the Caesars

2.26 (5–4 BCE: Oxyrhynchus) and SB 6.9223 (2 BCE: Oxyrhynchus).14 The purpose of this study is to test how we can identify and situate any Latin speakers, or, indeed, Italians besides those from predominantly Greek-speaking areas, throughout the Ptolemaic period on the basis of onomastic data and ethnics attested in Greek and Demotic texts. All foreign (non-Egyptian) ethnics in Ptolemaic Egypt were collected by Csaba La’da (2002). Individual bearers of Italian names have been identified and catalogued for the Prosopographia Ptolemaica (Peremans et al. 1950–81) and, in relation to the Zenon Archive, by W. Pestmann (1981). Then, on the basis of the Prosopographia Ptolemaica, a study of the evidence for relations between the Romans and the Ptolemies by Peremans and Dack (1972) identified around seventy-five individuals by their ethnics. E. Van ‘t Dack (1980 and 1983) also digested the prosopographic evidence for migration and mobility from Italy to Egypt but concentrated on the Roman army in Ptolemaic Egypt (in the years 55–30). In addition to these works, I have carefully reviewed all attestations of names in the Trismegistos People database that are dated before 1 CE and that are neither securely Greek nor securely Egyptian.15 In this way, this study is also an experiment in what can be done with the onomastic data in Trismegistos.16 14

15

16

For convenience of access and citation, papyri and ostraca have been cited in accordance with the abbreviations of Sosin et al. (2011), while references to inscriptions follow Bérard et al. (2010). When necessary, the volumes of such documentary texts are cited by their author and date. For texts that have been republished, whether or not they have been re-edited, I have confined references to the edition that I have followed (the most recent and reliable) and to the standard collections, such as ID, IG, OGI, and SB, in order to minimize multiplication of references, unless points of reading are at stake. Since the use of Greek script and the details of that use in representations of Italian names are of interest, I have generally refrained from transliterating Greek. In accenting and aspirating Greek renderings of Latin names, I follow Gerard Boter (2011). I am grateful to Mark Depauw for extracting this data for me (5.4.2017). While reviewing this data, I have been able to locate various errors both in the Trismegistos database and in the digital editions of papyri and ostraca, some of which go back to the printed editions themselves. The relevant corrections have been communicated to Mark Depauw and his Trismegistos colleagues and minor corrections to papyri have been reported through papyri.info. For an introduction to Trismegistos People, see Depauw (2016), which deals with automated and weighted analysis of the raw data for Latin names in Egypt throughout the coverage of Trismegistos People.

235

Patrick James

Personal names and ethnics alike are problematic as means of identifying Italians and speakers of Latin and Italic languages. The relationship between the name of an individual and the language(s) that they use is not straightforward. For example, the earliest Latin documentary papyri written are letters between individuals with Greek names, largely about individuals with non-Latin names, and thus demonstrate that a Greek name is not necessary evidence that a person only ever wrote or read Greek. Diaconus wrote to Macedo (SB 20.15139 (19 July 5–2 BCE)) and mentioned a Nireus (dative singular Nireo: line 8), but also Iucundus, Didius, and a freedman of a Domitius. (The only Greek written in the letter is the Egyptian date given after the Roman date.) A Suneros wrote to a Chios about a Theo (Θέων) and Epaphras (P.Oxy. 44.3208 [Augustan]). If we had only the names of these individuals and not these Latin texts, we would have no basis to conclude that they spoke or wrote Latin or understood written or spoken Latin. Indeed, a man with the Latin name Bassus wrote in Greek (P.Ryl. 4.603 [16.8.7]) either because he was illiterate in Latin or because he chose to write Greek either for reasons of his own or for the sake of Heracleides and Tryphon, his Greek-named and, perhaps, not Latin-reading correspondents at Oxyrhynchus. It is not always easy, or even possible, to distinguish a personal name from an ethnic. In particular, we cannot tell whether Ποντία in GVI 1002.2 (150–30: almost certainly from Naucratis) is the feminine Greek ethnic ‘of Pontos’, perhaps as a Greek personal name,17 or Pontia, the feminine form of Pontius, the Latin nomen of her father. As a further alternative, this Pontia may have been given her (Greek) name after a Nereid (‘Lady of the Sea’).18 This epitaph of Apollos, son of Leon, begins κούρη Νικομάχοιο Λέοντί με τίκτεν Ἀπολλῶν | Ποντία: ‘Pontia, daughter of Nicomachos, bore me, Apollos, to Leon’. The noun κούρη ‘girl, daughter’ either has the mother’s name Ποντία in apposition (and enjambment) or is qualified by an ethnic adjective ‘of Pontos’. If Ποντία is an ethnic adjective, the mother of Apollos was not from Naucratis or even Egypt. The same might well be true if Ποντία instead proved 17 18

For ethnics as personal names, especially of slaves, see Fraser (2009: 215–24). For depictions, see Robinson, Harcum and Iliffe (1930: 224, no. 454) and AhlbergCornell (1984: 147).

236

Italians and Others in Egypt before the Caesars

to be an ethnic as a personal name. The designation ‘a young woman from Pontos’ may have been used in order to avoid naming a respectable woman in public, if not to stress her origins for some reason unknown to us. In any case, it is impossible to be sure in a specific instance about the reasons for the presence of an ethnic or the absence of a personal name, even when there are parallels to consider and especially so in the case of a highly personalised metrical epitaph. However, it is clear that, since her father’s name is given as Nicomachos,19 it is unlikely that Ποντία reflects a Latin nomen Pontius, unless this individual was known as Nicomachos in Greek, but also by the Latin name Pontius.20 Since the latter remains no more than speculation, it is more likely that Ποντία is a Greek name, or is an ethnic, or is an ethnic as a personal name. Instances of Ἀττικός are similarly problematic. The cognomen Atticus is clear when it occurs in the tria nomina of Marcus Cottus Atticus (BGU 4.1113 passim). However, the eight-year old Ἀτικος [sic] named on P.Cair.Zen.1.59076.8 (13.5.257), if indeed he was an Ἀττικός,21 was involved in migration of a kind, but he was not a Roman with the cognomen Atticus. This letter tells us that he was one of four slave boys of superior quality that Toubias was sending from Judaea to Apollonios at Alexandria (the letter itself was found at Philadelphia). Abydos 114 (Ptolemaic), a graffito on the Memnonion, with a third-person verb, preserves an instance: Ἀττικὸς προσεύχεται | τοῖς ἰ Ἄβυτον (= ἐν Ἀβύδῳ) θεοῖς | ἵνα ὑγιαίνῃ. This could be a cognomen as a Greek single name or the ktetic adjective or ethnic Ἀττικός as a personal name,22 from which the cognomen came. In contrast to other graffiti at Abydos that we shall discuss below, this Ἀττικός has neither an ethnic nor a patronymic genitive. We cannot be sure that he was not a Greek named Ἀττικός, perhaps a slave. We cannot be sure that certain names are not Greek. Latin Lucius was rendered in Greek either as Λούκιος or as Λεύκιος (cf. Gignac 19

20 21

22

It seems inconceivable that κούρη Νικομάχοιο and Ποντία do not belong together and that, instead, Νικομάχοιο qualifies Λέοντι. I am grateful to James Clackson for discussing the names in this verse epitaph with me. Tcherikover and Fuks (1957: 127) report Semitic analyses of this name and those of the other slave boys. For ktetics, and Ἀττικός in particular, see Fraser (2009: 35–53, especially 40–3).

237

Patrick James

1976: 216). Some Λεύκιοι, such as Memmius the senator (P.Tebt. 1.33.3), were clearly Romans, while some, such as Λεύκιος Θρᾶιξ λοχαγός (BGU 6.1266.39 and 14.2386.6; cf. Fraser (1972: 170)), clearly were not. The name Λεύκιος, ‘Whitey’, had long been in use in Greek.23 In the case of Memmius and others, the Latin name has been assimilated to a similar Greek one, which happens to be related etymologically.24 The name Λούκιος does not appear in Greek documents from Egypt until 11.1.14 (Λούκιος Πομπώνιος: BGU 4.1113 passim), except in a much damaged painted funerary inscription that begins Λούκις ME|ṆỴẠ and ends with the remains of a date (I.Hermoupolis 35.1–2; 99–30). However, instances of Λευκι- are found in P.Hib.1.42 (30.12.262), P. Strasb. 8.761v9 (second century), P.Tebt. 3.882 fr. 2 25 (father of Θεοξένα; 26.1 or 24.2 144), SB 1.1658 (c. 260–197: incised on a sepulchral vase from Alexandria), SB 14.11367.1 (τοῦ Θεοφίλου, 260–200: Crocodilopolis), and W.Chr. 110a (10.12.110: Dios Polis), all of which lack any indication that the name is not the Greek name itself (or, indeed, the Greek name as an alias for an Egyptian name). Nothing suggests Italian connections for these Λεύκιοι, while Λεύκιος Κυδωνιάτᾱς in Syringes 385 (332–30), a visitor from Crete, and another visitor from Egypt to Delos, Λεύκιος Ἀθηνοδώρου Ἀλεξανδρεύς (ID 2595.38–9; 24.9.126/22.9.125) certainly were not Italians. Further, some forms of some names are ambiguous. The genitive Ἀττίου could belong to the Latin name Attius (cf. CIL II 4085 and Persius Satires 1.50) or to Greek Ἀττίας. It is clear from Ἀτίας Θρᾶιξ in BGU 6.1266.39 that Ἀττίου in BGU 6.1266.45 is almost certainly the genitive of Ἀττίας, not of Attius. The genitive Ἀττιου [sic] in SB 24.16272 (fr. i 32, i 36, and vi 237; 275–226; Memphis, Saqqara) is always followed by 23

24

The earliest of the 290 bearers of this name reported by LGPN is in DGE 715, 1, which comes from Samos and is dated to 575–550. See Jeffery (1990: 328, 333, and 341, no. 5) for the date and Bechtel (1917: 278) for the name. For further discussion of Λούκιος and Λεύκιος in the context of Roman Athens, see Solin (2001: 197). Both Lūcius and Λεύκιος reflect the Indo-European root *leuk-/louk-, as seen in Latin lūx and lūcus and in Greek λευκός. Likewise, the region Lūcania is called Λευκανία by Greek authors from its first mention in Greek in the Periplus of Scylax 12–14. Eutropius (Breuiarium ab urbe condita 9.28 and 10.2), Zosimus (Historia Noua 2.10.2), and John the Lydian (de Mensibus 1.7 and 4.95) in the fourth–sixth centuries CE are the first known to have used a Greek transcription of the Latin name, Λουκανία.

238

Italians and Others in Egypt before the Caesars

a mark of abbreviation. As such, this is not a genitive, either of Attius or of Ἀττίας. Some Italians may be concealed by the simplification of double consonants.25 The name Annius may be attested in O.Cair. 1.2 (as Ἄνιος, 11.1.150 or 08.1.139) and in P.Hels. 1.29 i 4 (161/160), which reads ὑπ(ὲρ) Ἀ̣ν ̣ίου. ̣ A Greek name Ἄνιος is doubtful itself.26 Trismegistos links from the first Ἄνιος to Ἀμμώνιος, as if it is an abbreviation of Ἀ(μμώ)νιος, and from the second to Annius (TM Nam 8561). We also have to bear in mind the possibility of a Semitic name, Ḥanni.27 Instances of names beginning Κασpresent comparable problems. They could represent Cassius or Cassia or a masculine Greek name Κάσιος or Κασίας (formed from an ethnic associated with the Sinai peninsula or with the Greek island of Kasos).28 The dative Κασίωι with iota adscript (SB 10.10226.1, second century) could reflect Cassius or Κάσιος and the genitive Κασίου (P.Hamb. 1.57.19 (28.10.160)) could belong either to Cassius or to either Greek name. If P.Hamb. 1.57 does involve an Italian with an Italian name, it shows an Italian as the father of a νησιώτης (an ‘islander’ of some kind) swearing an oath about hunting in the second century. If SB 10.10226 were addressed to a Cassius, an Italian would be shown to have been involved in giving a sum of money to a Psenepathes and in avoiding involvement with another individual with a Κασ- name: καὶ μὴ πα[ρ]ὰ Κασ ̣ία̣ ̣ | ἔχῃς αὐτ[.] . . (lines 5–6). Given these problems, and since the following lines are damaged and the reading Κασ ̣ία̣ ̣ is uncertain, we cannot say more about the identity of this individual or his29 role in this transaction. 25

26

27

28

29

For the simplification of -νν- and of -σσ-, see Mayser and Schmoll (1970: 189–90) and Gignac (1976: 158–9). The editor of P.Hels. 1.29 commented that the name is unknown and that the first letter could be , while the editors of O.Cair. 1 refer to Pape and Benseler (1911). Pape and Benseler (1911) report a son of Apollo (e.g. Diodorus Siculus 5.62.1), a son of Aeneas by Lavinia, and a Roman Anius Aemilius (Diodorus Siculus 12.80.7). LGPN reports no instances of Ἄνιος. For this possibility in relation to a different, and later, data set, see Sartre (2007: 201 and 204). For the possibility of Κάσιος with a single sigma representing a Semitic name, see Sartre (2007: 204). Trismegistos relates this instance to TM Nam 3551: Greek Κασία, Latin Cassia. However, ΚΑΣΙΑ after παρά must be a dative Κασίαι of masculine Κασίας (as well as

239

Patrick James

Some instances of some Latin names are likely to conceal individuals not of Italian origin. The name Gaius often appears in inscriptions relating religious associations and military units. A list of dedicators from Hermopolis has three men called Gaius (I.Hermoupolis 6.156, 165, and 198 (25.1.78)) in a religious association and in a military context (παρ[ε]φεδρεύοντες ‘serving as guards’: line 2). The three genitive patronymics are Greek names: Ἀλεξάνδρου, Ἰσιδότου, and Ἀπολλωνίου. Another instance of Gaius occurs in the company of οἱ ἱππεῖς as the maker of a dedication to the king (I.Delta 1.892, 1.3 (80–52/51)). The dedication is as enigmatic as it is short. SB 1.681, a report of a decision by a religious association, names two men with the name Gaius among its one hundred and seventy members. The first (64 ii 18) was a son of a Demetrios. The context of the second (143 iv 19 Γάϊ ο[ς) is lost. In the absence of further information, nothing more can be said, but it is notable that the ‘colourless’ name Gaius is the one that appears in such environments.30 However, the most striking feature of these inscriptions is not the colourless name Gaius or the similarly colourless Greek names, like those of the fathers in I.Hermoupolis 6, but the nonGreek names or names with non-Greek associations. SB 1.641 contains Greek names with an Egyptian flavour (e.g. Ἀμμώνιος ii 12, 19, and 20) in addition to many names, such as Κοσάδαρος (i 21, 22, 27, 31 and ii 25; cf. i 4) and Κοσμάλαχος (ii 15 and iii 18; cf. i 17 and 39), built on the stem Κοσ-.31 The name Κοσμάλαχος is transparently of Semitic origin and is a compound of Qôs, a theonym among the Idumaeans, and Hebrew mālak ‘king’.32 The names Ὀρόντας and Ὀρύντας in

30

31

32

feminine Κασία) or a dialectal genitive of Κασίας (assuming that neither Κασίας nor Κασίαν should be read or restored). For ‘colourless’ names, including Gaius, in military contexts, see Mairs, Chapter 9 of this volume, and Solin (2001: 195–6). For an inventory of names beginning Κοσ-, see Wuthnow (1930: 65), in which this inscription is cited as ‘Eph II 340’, and Ruozzi Sala (1974: 23–5). Other names in the list also suggest a largely Semitic population: Ἄλιος (36 = ii 32), Ἀσάδος (29 = i 25 and 109 = iii 21), Ἀψαλάμος (21 = i 17; cf. Biblical Absalom), Βόρακος (133 = iv 9), Ἐλμά[λα]χ[ος] (37 = i 33; cf. Biblical ˀElîmelek), Ζαβδαῖος (69 = ii 23) and Ζα[β]δᾶς (42 = i 38), Ἰεγουθος (sic: 40 = i 36 and 81 = ii 35; cf. Arabic Yāqūt), and Χαλαφαθος (sic: 134 = iv 10; cf. Arabic Khalīfa ‘Successor’). In some instances, the name and patronymic alike have Semitic origins (e.g. Ζαβδαῖος Ἀσάδου 29 = i 25), in others the son’s name is Greek, the father’s Semitic (Διόδοτος Κοσμαλάχου 106 = iii 17).

240

Italians and Others in Egypt before the Caesars

I.Hermoupolis 5.137 (~ SB 1.4206) and 6.209 (~ SB 5.8066) are of Iranian origin.33 Such isolated instances of the name Gaius amid Greek, Egyptian, Iranian, and Semitic names and patronymics suggest either that Italians had been integrated into non-Italian populations, such as the Idumaeans of Egypt, and into their associations, or, as seems more likely in the absence of other evidence in the context, that the name Gaius itself had ‘migrated’ to those populations, a reflection of more distant contact. Individuals identified by their ethnics or their names as those who may have been Italians will be discussed below in relation to the activities considered in our discussion of Acutius and his companions: ‘sacred’ (or ‘sightseeing’), diplomatic (and military), and commercial. That is not to say that a given Italian only ever had one purpose for being in Egypt or at a particular site at a given time. While so much uncertainty remains about Acutius and his companions because the inscription itself does not mention the purpose(s) of their visit to Philae, other texts, such as I.Philae 1.63, which will be discussed later, do contain an explicit statement of one purpose but also contain indications of status that relate to other activities. Other texts contain explicit statements of purpose in relation to particular locations, but also suggest that those individuals had other reasons for being in Egypt in general. A letter in the Menches Archive from Tebtunis quotes a letter of instructions about a visit by Lucius Memmius, a Roman senator (P.Tebt. 1.33.3–4 [17.3.112]): Λεύκιος Μέμμιος Ῥωμαῖος τῶν ἀπὸ συνκλήτου.34 The explicit purpose of the visit is ‘sightseeing’ (ἐπὶ θεωρίαν: line 6) in the Arsinoite Nome. Preparations (lines 7–16) include those for viewing the labyrinth (τὰ πρὸς τὴν τοῦ λαβυρίνθου θέαν: line 14). Even if Memmius’ own concern in the Arsinoite Nome was tourism, that is not the full picture of his visit or of his reason for being

33 34

A hybrid of Semitic and Greek elements appears in Χαλαφάνης (103 = iii 15). See Wuthnow (1930) and Ruozzi Sala (1974). Huyse (1990: 53–4) compares Old Iranian /*aruwanta-/ ‘quick’. Grenfell, Hunt and Smyly (1902: 127) suggested that this Lucius Memmius might be the son of Gaius who appears in IMT 717.13–14, the Adramytteum copy of the Senatus Consultum de Agro Pergameno. The senator is likely to have been the moneyer, whose coin issues have been dated to 115, and the brother of the tribune of 111, on whom see Wiseman (1967: 166).

241

Patrick James

in Egypt already (at Alexandria, since his voyage was ἐκ τῆς πόλεως: line 5). Memmius was no ordinary visitor, but a distinguished guest (ἐν μίζονι ἀξιώματι κα[ὶ] τιμῆι | κείμενος: lines 4–5) and his visit to the Arsinoite Nome was a matter of great concern to Asclepiades, no less than the superintendent of revenues (cf. P.Tebt. 1.27 iv 98–9 [27.2.113]), who issued the instructions, and to Horos the royal secretary (cf. P.Tebt. 1.27 ii 27). Horos forwarded the instructions to Eirenaios, if the name has been read correctly,35 no less than the dioectetes and syggenes (‘an honorary cousin of the king’; cf. P.Tebt. 1.27 iv 85–6). At each stop he would receive gifts (τὰ ὑπογεγρ(αμμένα) ξένια: line 11), of which the details and value are lost to us. The senator was to be honoured to an exceptional extent (μεγαλο{υ}πρεπέστερον: line 6). The summary instruction (lines 16–19) was to show τὴν μεγίστην φροντίδα and τὴν πᾶσαν σπουδή[ν]. The senator did not come to Egypt solely to visit the sites of the Arsinoite Nome. He was already at Alexandria, for reasons which are unknown, and was to be welcomed with utmost honour. In other cases, we have clear evidence for migration or mobility, but no information about the duration or purpose of that migration or mobility. A Mamertine (see Tagliapietra and Zair, Chapters 6 and 7 of this volume) was buried in the Hadra cemetery at Alexandria in the second or first century: Μαραῖος Βακείου | Μαμερτῖνος (SB 1.417 f.1–2). The Italic origins of this Maraeus are suggested by his name (Oscan Maras) and probably by the name of his father.36 However, his burial tells us nothing either about the duration of his time at Alexandria (or in Egypt more widely) or about his reason(s) for being there.

35

36

Grenfell, Hunt and Smyly (1902: 127) read Ἑρμ(ίας) and identified him with the Hermias of P.Tebt.1.27 ii 27, who may have been a dioecetes. The father’s name has been too readily Hellenised in the scholarship as Baccheius (Βακχεῖος). LGPN knows him as Βακχεῖος with a note ‘(Βακεῖ-: f. Μαραῖος)’. No justification is given for the alteration. The simplification of in writing, however plausible in itself, cannot be paralleled in Egypt or elsewhere. On the basis of the ethnic and the name, James Clackson’s suggestion that the genitive belongs to the Oscan personal name Pakis is very appealing. For the rare interchange of voiced for voiceless in the Ptolemaic period, see Mayser and Schmoll (1970: 145).

242

Italians and Others in Egypt before the Caesars

Italians at Sacred Sites? The presence of the names and ethnics of Italians at sacred sites does not prove that those individuals visited those sites principally for religious reasons or even for the purpose of ‘sightseeing’ at those sacred sites at all. Several of these attestations are certainly Ptolemaic but cannot be dated more precisely. Some individuals recorded their proskynema mark in Greek: τὸ προσκύνημα. Σερηνός (Abydos 450bis, e) and τὸ προσκύνημα | Σαβείνου Μαρίνου (Abydos 5.1–2 (second century)). Another appears with nothing more than the conventional use of ἥκω: Σαλουτάριος Ἀλεξανδρεὺς ἥκω (SB 1.3772 (dated to 99–1 BCE)). The ethnic of an Alexandrian (and, presumably, a citizen) here follows a Latin cognomen. Others did not explicitly indicate their purpose: Κρίτων Ἀγαθωκλέους Ῥωμαῖος (Abydos 55) and Νικίας Πύρρου, Ῥωμαῖος Ἀρίστομμ̣ ος ̣ (Syringes 1092), as punctuated by Jules Baillet: Syringe 9, Suite 2, Left wall, Corridor, 1st section, suite a (Ptolemaic period). Here, it is remarkable that, although the ethnic Ῥωμαῖος indicates Italian origins, the personal names do not. Since Baillet’s drawing does not suggest any space, division, or change of inciser between the words, it is particularly unclear whether Ῥωμαῖος belongs with Nicias, or with Aristommos, or is itself a third personal name. Graffiti in the Syringes at Thebes show indications of identity such as Italian onomastic practice, the ethnic Ῥωμαῖος, and, although rarely, the use of Roman script. As texts these graffiti are silent about the purpose of the visits: their location is the only indication. They are the only Ptolemaic specimens of Latin other than the Acutius inscription at Philae, that of Marcus Tinotrius, and an addition to an inscription incised at the end of the Ptolemaic period, which will be discussed at the end of this chapter. The names Marcos Volturios and Marcos Voturios feature in four graffiti at three locations in two of the Syringes (three times with the ethnic Ῥωμαῖος). It is likely (if these names are in fact identical), that the graffiti refer to the same visitor. However, since the graffiti cannot be dated more precisely than the Ptolemaic period, we cannot exclude the possibility that several bearers of those names visited two of the Syringes at different times during the Ptolemaic centuries. The absence of cognomina is consistent with a date before the first century. The inclusion of a cognomen 243

Patrick James

by M(arcus) Claudius Varus at the end of the second century (I.Syène 321), but neither by Gaius Acutius nor by Sp(urius) Varaeus, illustrates that period of transition to increased use of cognomina.37 The dossier, based on the edition of Baillet (1920–6), consists of the following four items. 283 [Μ. Ου]οτύριος Ῥωμαῖος. (Syringe 2, Suite 6, 2nd door, 2nd section; Plate XI)

588 M(arcus) Volturios

38

(Syringe 2, Suite 18, 2nd section; Plate XXIII)

2003 Μάρκος Ὀτούρι[ο]ς | Ῥωμαῖος. (Syringe 9, Suite 34, Right wall; Plate LXXV)

2004 M. Voturi[os] | Rωmaios. (Syringe 9, Suite 34, Right wall; Plate LXXV)

The identification of a single Marcos Voturios or Volturios who left his marks at the Syringes is supported by the juxtaposition of 2003 and 2004 on the right wall of Suite 34 of Syringe 9. Together they constitute a bi-version bilingual and should, perhaps, be catalogued as one item. The binomina and ethnic were incised first entirely as Greek in Greek script (2003, the fullest and best preserved of the four graffiti) and second as Latin in Roman script with the praenomen abbreviated (2004), but with a switch into Greek script and, perhaps, with an instance of Greek inflectional morphology in Rωmaios.39 (However, since the raising of final [o] to [u] was not yet ubiquitous in the third century, if all these graffiti are that early, -os may be the original Latin spelling.) Even if the use of -os reflects the original Latin ending (as, apparently, in 588), the form of the ethnic itself is Greek Ῥωμαῖος spelled with (Greek) 40 (not Latin ), not Latin Rōmānus or Rōmānos (nor, for that matter, Italicus, familiar in the plural at Delos).41 283 and 588, 37

38

39

40

41

For the emergence of cognomina late in the second century and their subsequent spread throughout Roman society, see Salway (1994: 127). The transcription and the drawing both give Volturios with but the transcription resolves the abbreviation M. as M(arcus). The omega was not reported as such in CIG 3.4807g, CIL III 68, and IGUR 1.1227, editions all derived from Champollion’s transcription. See Baillet (1926: 517). The and were printed with a ligature, as ROMAIOS, to reflect how they appear on the wall. As a consequence, IGUR printed Romanos, the Latin ethnic. For the Italici at Delos, see the discussion later in this chapter and Rovai, Chapter 8 of this volume.

244

Italians and Others in Egypt before the Caesars

both in Syringe 2, but far from adjacent, could similarly be records in both languages of the same Roman visitor. The Volturius of 588 is the least likely to belong to this dossier. Not only is 588 unquestionably Latin in its script and, probably, in its morphology (-os), it alone contains in its first syllable and it alone lacks an ethnic. Also, the name Volturius may be a different name from Voturius and its Greek representations, if indeed the latter existed.42 The presence of in the ‘pure’ Latin graffito (588) and its absence from the Greek texts (283 and 2004) and from the mixed text (2003) could be explained as a very early instance of a phenomenon otherwise only attested in much later Latin: the absorption of ‘dark l’ (/ɫ/) after a back vowel ( in these graffiti).43 The inciser(s) would then be working with the Latin that was heard and without familiarity with the standard Latin spelling with . Although Adams (2003a: 622) cited si uut for si uult (O.Wâdi Fawâkhir 3.14 (second century CE), which closely parallels Vot- for Volt-, such absorption results in a high back vowel, as in cauculus for calculus and Cretan Greek καυχός for χαλκός (IC 4.162.4 [late third century]).44 There is no such high vowel in these graffiti. Whoever incised this Greek heard only a vowel (and an [l] or [ɫ]) in the Italian names. The omission of after would be more likely to be the result of interference from Latin, in which this phenomenon is attested, than of a development either within Greek in general (Cretan not withstanding) or within Greek subject to Egyptian interference. However, such absorption of ‘dark l’, which is attested in Latin of the Roman period, is implausible in the Ptolemaic period and, therefore, it is less likely, but not inconceivable, that M(arcos) Volturios (588) was the Marcos Voturios of 283, 2003, and 2004. If 588 concerns a different man, we would have another Italian at the Syringes, 42

43 44

Schulze (1904: 252 and 260) distinguished the pair Veturius and Voturius from names associated with Etruscan velθur: Voltorius, Volturius, Vulturius, and Vlturius. The tribe name Voturia lies behind attestations of Votur- in Latin epigraphy. However, an instance of Ὀτόριος is found as the nomen of a Κόϊ ντος | . . . Κοΐντου Ῥωμαῖος at Delphi (Syll.3 585.17–18 (early second century BCE)) and CIL XII 1711, an undated inscription from Gallia Narbonensis, reads L(ucio) Voturio Maximo aedili | pagi Aletani patri | C(ai) Voturi Aviti. See Allen (1978: 34 note 2). The first example is from Adams (2003a: 622), the second from Buck (1955: 64).

245

Patrick James

who did not supply an ethnic (certainly not a Greek form), and we would have a (possibly) purely Latin graffito from some point in the Ptolemaic period (perhaps earlier than the Acutius inscription) in addition to the Greek graffiti in Greek script (283 and 2004) and to Latin graffiti showing Greek interference (2003) of one Marcos Voturios or several. Even if Volturios in 588 does not belong in this dossier, we still have three instances of the Greek ethnic Ῥωμαῖος and two Greek representations of Voturios (283 and 2003) in addition to one instance of the name in Latin script (2004). The difference between medial represented by (2003) or by (283) reflects diversity of practice and the lack of an established Greek spelling for this name: the former coincides with the (later) standard transcription of as , while the latter is a transliteration and is only usual for centuria and occasional for turma.45 The initial syllables, at least as restored, show the converse treatment [Μ. Ου]ο-τυρ- (283), but Ὀ-τουρ- (2003). The essential problem is how to represent the Latin phoneme /w/, for which Koiné Greek had no grapheme: one solution was omission (2003), another, which is widely attested, was to use (as restored in 283).46 However, since Baillet’s drawing suggests that the damage to the graffito would suit a single letter better than the three he restored, it is probable that Ὀτύριος is what was incised as 283 (comparable with 2003) and that it was preceded by Μ. alone. However, the Roman abbreviation is found in the Latin graffiti (588 and 2004), while the Greek graffito (2003) has the praenomen in full. On that basis, [Ου]οτύριος should be read (without Μ., probably, given the spacing and use of the unabbreviated name in the Greek graffito). 45

46

See Gignac (1976: 217–22) for this diversity once Latin words and Roman names are more plentifully attested in the papyri. After the first century CE, τούρμης prevails. For later instances of transliterated , not transcribed as , at the Syringes, see Adams (2003a: 585). For omission of any sign for consonantal in transcriptions, see Gignac (1976: 231–3). A third option, representation by , is unlikely at a Ptolemaic date. The development of [w] to [β], spelled is evident in first-century CE inscriptions. See Allen (1987: 41). Although is not usual in the Roman period, it is found with increasing frequency from the first century CE onwards. See Gignac (1976: 68–70). At the end of the fragment published as SB 16.12338 (second century), βιβις̣ could represent Vibi(u)s, but that is unlikely because a name in the nominative would follow the three names in the genitive and βιβις̣ appears in eisthesis, as if a heading, not a list item.

246

Italians and Others in Egypt before the Caesars

These differences of practice certainly reflect a lack of clarity about representing Voturios. Different (local) Greek incisers may have followed their own ideas, or Voturios, or a companion, may have experimented with different possibilities. However, it is likely that Volturios, or an Italian companion of his, incised the Latin of 588, since it is doubtful that an inciser of Latin could be found locally at Thebes at any time in the Ptolemaic period.47 The inciser of 588 knew the Latin spelling with (as did that of 2004) and with , unique among these four graffiti if we continue to associate them all with one individual Volturios and to disregard Voturius as a distinct personal name rather than a misspelling. However, the Greek graffiti (283 and 2004) could be the work of a local, and the in Rωmaios (2004), itself not the Latin ethnic form (Rōmānus), suggests that 2004 was incised by someone who was more at home with Greek than Latin and who heard, rather than read, either Latinised Rōmaios or Greek Ῥωμαῖος. The presence of the ethnic Ῥωμαῖος (Greek, both in form and in script) is likely to reflect the non-Italian point of view of someone other than the bearer(s) of the name Marcos Volturios. The ethnic is found in the two graffiti in Greek script (283 and 2004) and in 2003, which, in Rωmaios, shows Greek interference in its script and, perhaps, its inflectional morphology. Indeed, as far as we can be sure, Greek interference in that graffito is concentrated entirely on the ethnic Rωmaios,48 itself not the Latin ethnic. A Marcus Voturius might be expected to identify himself with Rōmānus (or -os), but to be labelled as Ῥωμαῖος by another (and a Greek speaker, at that). By contrast, 588 is entirely in Latin script and does not feature any ethnic and, on these two points, parallels the Acutius inscription (I.Syène 321). Perhaps, the use of Latin script itself was sufficient in that inscription and in this graffito to signal (visually) that an individual was neither Egyptian nor Greek (in the wide sense of Greek speakers from throughout the Hellenistic world). As we saw above, the ethnic Ῥωμαῖος was found at Abydos and the Syringes

47

48

By contrast, see Adams (2003a: 547) for the availability of stonecutters on site at the Colossus of Memnon. The supplement in Voturi[os] was based on the -os legible in Rωmaios. If we do not supplement [us], [os] could be taken as the original Latin spelling.

247

Patrick James

with Greek names that were so widespread or so geographically unremarkable that the addition of an ethnic might be required. A Greek inscription from the end of the Ptolemaic period I. Philae 1.63 (~ OGI 1.196), which attests a group of visitors at Philae, contains a cluster of Italian names and other indications of Italian origin for the first named member whose proskynema it records. His tria nomina are given and the inscription contains the earliest attestation of a Latin loanword in a text from Egypt: κεντορίωσι (line 9). Although, as we shall see, its details argue against the leader of the party being a visitor from Italy or indeed from outside Egypt, the inscription does provide evidence of increased contact with Romans.49 Γάϊος Ἰούλιος ΠΑΠΕΙΣ ἔπαρχος ἥκω καὶ προσκεκύνηκα τὴν κυρίαν Ἶσιν, σὺν Ἰουλίωι τῶι υἱῶι, καὶ ὑπὲρ Γαίωνος τοῦ νεωτέρου υἱοῦ, ἔτι δὲ καὶ σὺν τοῖς φίλοις καὶ συν5 αποδήμοις Συμμάχωι, Εὐμένει, Ἀπολλωνίωι Χάρητος καὶ υἱῶι Ἀπελλᾶι, καὶ Ἀ̣π ̣ίω̣̣ ν ̣ι ̣[— – —] καὶ [— – —] καὶ [τοῖς] κεντορίωσι Ῥούφωι, Δημητρίωι, Νίγρωι, Ουαλερίωι, Λαβύωνι, Τερεντίωι, 10 Νικάνορι, Βαρωνᾶτι, καὶ τῶν παιδαρίων μου πάντων (ἔτους) κʹ το καὶ εʹ, Φαμ(ενὼθ) λʹ. I, Gaius Julius PAPEIS, eparchos, came and reverenced Lady Isis, together with Julius my son, on behalf of Gaion my younger son – also with my friends and fellow-travellers (Symmachos, Eumenes, Apollonios, son of Chares, and his son Apelles, and Apion . . . and . . .) and the centurions (Rufus, Demetrios, Niger, Valerius, Labyon, Terentius, Nicanor, Baronas) – and of all my children. In year 20 (that is also year 5), 30th of Phamenoth.

The double regnal year in this inscription gives the date as 28 March 32, the twentieth year of Cleopatra VII as queen of Egypt, her fifth as ruler of Chalcis in Syria.50 (The title ἔπαρχος 49

50

Capponi (2005: 17) introduced this party as Roman centurions, recruited from Greekspeaking provinces, and a tribunus militum or praefectus castrorum who came to Egypt with Gabinius, with Julius Caesar (as suggested by the praenomen and nomen C. Julius), or with Mark Anthony. For the calendar and this date, see Skeat (1954: 18 and 42).

248

Italians and Others in Egypt before the Caesars

has been left untranslated, the name ΠΑΠΕΙΣ unaccented and left in transliteration, and the names Λαβύων and Βαρωνᾶς left in italicised transliteration and without Anglicised equivalents in order to draw attention to them as problems in the interpretation of this inscription). This inscription differs from the Acutius inscription (I.Syène 321) in its first-person verbs (not third-person), its statement of the purpose of the visit (προσκεκύνηκα: line 2), its reference to Isis (line 3), and its date by the Egyptian calendar (line 13), not those of Rome or of the Macedonian Ptolemaic dynasty. The use of the tria nomina here in fact points to an Egyptian rather than a Roman visiting from Italy. Although the cognomen is now read as ΠΑΠΕΙΣ (not Πάπειος, as in Dittenberger (1903–5: 281 and 579)), this Gaius Julius continues to be thought to have the cognomen Papius,51 even though his identification with a son of Gaius Papius, the author of the lex Papia de peregrinis, must be rejected on prosopographical grounds (with Dittenberger (1903–5: 281)): a Gaius with the nomen gentile Julius could not also have Papius as a nomen gentile. The spelling Πάπεις for Πάπιος could be an instance of for before a vowel and of syncope of in -ιος, but there are no parallels within this inscription to corroborate such an analysis.52 Instead, it is simpler to take ΠΑΠΕΙΣ at face value as the well-attested Egyptian name Παπ(ε)εις, variously spelled and accented (see TM Nam 4863), a suggestion already made by Dack (1983: 208–9). Papeis’ praenomen and nomen then become those of an Egyptian who was granted citizenship (and perhaps freedom) by Julius Caesar and perhaps after military service with Caesar. His single Egyptian name then became his cognomen. (Other Gaii Iulii in the final Ptolemaic decades have been reported by Dack (1980: 45–6), who labelled them ‘freedmen’ (‘vrijgelatene’). Almost all have a Greek personal name, not a Latin cognomen as their third name.) This analysis has some corroboration from the names of the sons of this eparchos. Papeis gave his new nomen gentile to his older son, in accordance with Roman practice, and named his 51

52

See Preisigke (1922: 275), Bernand (1969: 334 and 335), Ricketts (1980: 107), Capponi (2005: 17), and Bingen (2008: 251). Contrast the conspicuous array of examples in O.Edfou 3.371: see the discussion later in this chapter.

249

Patrick James

younger son, Γαίων, with a hypocoristic built on his new praenomen and the suffix -ōn-. Such a formation is more likely to be the creation of a Greek speaker than a Latin speaker. The Greek suffix was readily used to form nicknames (Buck 1933: 32), but the Latin suffix was limited to forming cognomina and derivatives from them (hence, Καισαρίων and Rufiō). Dittenberger (1903–5: 281) could cite only Μαρκίων as a hypocoristic built on a praenomen. The principal member of this group of visitors to Philae is far more likely, then, to have been an enfranchised Greek-speaking Egyptian than a Roman citizen visiting from Italy. The names of Papeis’ companions suggest a group of diverse origins. As Dittenberger (1903–5: 282) noted, the συναπόδημοι (‘fellow travellers’) all have Greek names, while the centurions mostly have Latin cognomina as their sole recorded name. In addition to two of the exceptions who have Greek names (Demetrios and Nicanor: lines 9 and 11), the name Βαρωνᾶς (line 11), unattested elsewhere, is, at the very least, not Latin in its morphology (dative Βαρωνᾶτι, not Βαρῶνι). This name may be a Hellenisation of the Aramaic name, Barōnā,53 which would add to the onomastic diversity evident in this inscription. The morphology here is consistent with Hellenisation of a foreign name, particularly a Semitic or Egyptian one.54 Dack (1983: 205) suggested a Hellenisation of a Latin name with for : Varus or Varro, which is possible, but unlikely at this date.55 For that reason, if Βαρωνᾶς is Latin, it is more likely to be based on bārō ‘dullard’, used as a cognomen,56 with Greek morphology. The orthography of this inscription further suggests that whoever was responsible for its Greek was not well acquainted with written Latin. First, the name of one of the centurions, Λαβύων (line 10), has been taken as the Latin cognomen Labeō (Bernand 1969: 339), for which Λαβέων would be expected. The for in an accented syllable is consistent with the restriction of such interchanges noted by Gignac (1976: 273–4) to contexts with 53 54

55 56

Ruozzi Sala (1974: 11) notes that this name was built on bar ‘son’. For the -ᾶς -ᾶτος paradigm, especially with Egyptian names, see Gignac (1981: 16–18 and 58). See note 46. See Kajanto (1965: 264).

250

Italians and Others in Egypt before the Caesars

a back vowel (as here), a nasal, or /s/ following. However, the vowel here only carries the accent in accordance with the conventions for editing Greek texts. Since the Latin original, Labeōni, would be accented on the penultimate syllable, this interchange is found in an unaccented syllable. Therefore, the strong stress accent of Coptic could account this interchange as a representation of a reduced vowel (see Gignac 1976: 275). Then, the loanword κεντορίωσι (line 9), a clear indication of contact with Latin speakers and with the Roman military, shows . (Transliteration with was ‘usual’ for this word, while a transcription with is by far the exception.57 Although this spelling with is found (Gignac 1976: 222), it is very rare and, as Gignac did not note, it is not found later than the time of Augustus.58 This was not the spelling that became the norm once a Roman military presence was established. Instead, it seems to have been a short-lived, or ad hoc, rendering of centurio59 and reflects Greek representation of Latin (both [ŭ] and [ū]) as , especially in the context of a resonant.) In renderings of Saturninus and related names, Greek for Latin ‘predominates’ (Gignac 1976: 221) and was ‘generally earlier’ than in representations of Sulpicius and Secundus and their feminine counterparts (Gignac 1976: 219–20). Syringes 1830, cited as a bilingual by Adams (2003a: 585), is instructive: Σα[τ]ο[ρ]νῖνος ἱστόρησα Saturninus uidi. Whoever wrote those two sentences used Greek for Latin in this name. The double date is a further indication that this inscription has a Hellenistic Egyptian perspective, not that of Roman visitors from Italy at Philae for ‘religious’ purposes or tourism. We cannot 57

58

59

The only epigraphic attestation remains SB 5.8518.5 (first c. CE ?): see Gignac (1976: 222). One papyrus can be cited: BGU 2.455.11 (first/second c. CE). Gignac (1976: 222) could cite only Ουιβίου κεντορί|[ωνος] (P.Oslo. 26.23–4 (5/4 BCE)), whose editor simply asserted that should be read for . No further examples of κεντορι- have been added by subsequently published papyri and the only other epigraphic attestation is on an honorific inscription, also of Augustan date, for Gaius Aemilius Geminus set up by some military colonists (κάτοικοι): TAM 5.1 425.6–7. An early attestation of the genitive (for centuriae) provides a comparably ad hoc rendering. SB 6.9223.3 and 4 (2 BCE; inscription on a pot; Oxyrhynchus) have κεντερυωνεας (not κεντ(ο)υρίας), a new derivative in -ίᾱ (spelled -έᾱ) from *κεντερύων, a spelling of κεντυρίων with vowel metathesis. The spelling with is paralleled by κεντερας (sic) for κεντυρίας (P.Oxy. 22.2349.2 and 27 (70 CE)). The sequence -υωνrecalls the name Λαβύωνι as a spelling of Labeōnī.

251

Patrick James

know which member of this group was responsible for the choice to use not only a purely Egyptian date but also the additional regnal year that acknowledged Cleopatra’s rule of Chalcis (a gift from Mark Antony). It certainly stands in contrast with the Roman date and consular year left on the Acutius inscription (I.Syène 321). If priests alone could incise a record of a proskynema,60 the choice of calendar and regnal years may have been no more than their statement of allegiance to Cleopatra. However, even if a priest incised the date, the choice and statement of allegiance may still have been those of Papeis (with or without the consent of his companions). The presence of an exact date enables us to look for other visitors who left a record at Philae on that date. On the basis of the re-dating of I.Philae 2.146 also to 28.3.32 by Bingen (2008: 249–51), we have a record of a proskynema by one Nicomachos and all his companions (τῶν παρ᾽ ἐμοῦ πάντων: lines 5–6). This Nicomachos was an honorary ‘cousin’ (syggenes) of a Ptolemy and a strategos (lines 2–3). The two proskynemata may be entirely coincidental. They were surely not the only ones that day. We may see a Ptolemaic military presence near Philae on that day that consisted of individuals with Egyptian, Greek, and Italian names (and perhaps others). The Demotic evidence for Italians at sacred sites is doubtful. P.Saqqara 60, published by El-Khouly (1973: 153), which preserves first-century rosters of the names of priests at the Sarapeum at Memphis, may read Swrȝ (recto, x+I 3), a Demotic transcription of the name Severus (Migahid 1998: 128). If the reading and the analysis are correct, the bearer of a Latin name served as a priest in the Sarapeum at Memphis. A graffito, which may be much earlier, from the temple of Imenhōtep at Philae has Swre as the son of Peteharbēk (Graff.Dodec.Philae 497.1 (after 285 and as late as 499 CE)). If Swre is a means of writing Latin Severus, and not Demotic Sour, this ostracon could be the earliest attestation of a Roman name in Ptolemaic Egypt. That is early enough, earlier than all our other evidence (especially our literary evidence for 60

Adams (2003a: 582) presents a balanced assessment of whether only priests could compose, place, and inscribe such proskynemata at Kalachba.

252

Italians and Others in Egypt before the Caesars

Romans in the Eastern Mediterranean), to raise doubts about seeing Severus here, and a later point in the date range 285 BCE to 499 CE is more likely. Later instances of Swre or Swrʿ in two graffiti of obeisance to Isis at Philae (dated, at the earliest, to 30) may involve the Roman name, if not Romans themselves: the father of Petepwēr junior (Graff.Dodec.Philae 105.3) and the father of Petepwēr senior (Graff.Dodec.Philae 121.2), perhaps the same man whose name was spelled differently. Italians in Egypt in Military Contexts Individuals who may be identified as Italians to varying degrees of certainty are found in military contexts throughout the Ptolemaic period, both in Egypt and in the wider Ptolemaic sphere of influence. We have already met the Italian-named centurions with Papeis at Philae. A papyrus associated with the Zenon Archive shows that, as early as the middle of the third century, one Ῥωμαῖος served in the Egyptian army and was based at Alexandria, but little else about the individual is certain, even the nature of his name. P.Lond. 7.1986 (01.9.252) names Δ̣ίννος Ῥωμαῖος τῶν Αὐτομέδοντος (‘Dinnos, a Roman, of the ̣ troop of Automedon’) as the first of the witnesses to a loan contract. The seven street names that relate to Arsinoe and her cult titles and the use of the Macedonian date without an Egyptian counterpart point to an Alexandrian origin, but the contract was recovered at Philadelphia, where, we presume, it travelled with one of the parties concerned. Although there is no positive internal evidence for associating this document with the Zenon Archive, that archive is the most common source, other than mummy cartonnage, for papyri of such an early date.61 This Dinnos associated with others in a military milieu, at least for the purpose of this loan contract: Āgesandros and Echedāmos, the borrower, probably were syntagmatarchs of particular streets (so, Bell 1924: 24) and the witnesses Callimachos, Dionysios, and Polemarchos were tachomisthoi of particular streets. Dionysios, son of Chion, a ‘Persian’,62 was a member of the palace guard (τῶν περὶ 61 62

See Bell (1924: 17–18) and Skeat (1974: 82–3). For Ptolemaic ‘Persians’, see Fraser (1972: 58–9).

253

Patrick James

αὐλήν), another indication of an Alexandrian origin. The reading Δ̣ίννος is relatively certain,63 but as the Italic praenomen that would be expected as a single name in a Greek document of this period, it does not take us very far. The existence of a praenomen Dinnus is corroborated by Dinnius, the nomen derived from it that is attested at Puteoli (CIL X 2081) and at Misenum (CIL X 3572 and 3573). Since, even with subsequent discoveries of inscriptions, that nomen remains principally attested in documents from the Imperial Period, we have little with which to work on a Ptolemaic instance of Dinnos.64 Military contexts, often in the wider Ptolemaic sphere of influence, are clear for several isolated individuals. One Λεύκιος Ῥωμαῖος, son of Gaius, a commandant at Itanos on Crete, left a dedicatory inscription to Ptolemy Philopator in Greek: IC 3, 4.18.4 (244–209). A Νόυιος Νουίου (I.Hermoupolis 4 ii 40/112 [125–100] ~ SB 1.599; cf. ii 20/92) bore the title πεντηκόνταρχος. One Gaius Lutatius Crispus is known from IC 4.215 C 1–3 (perhaps mid-first century: Gortyn): Γ(άϊ ος) Λυτ{ο ̣}άτιος Κρίσπος στρα|τιώτης Πτο(λεμαιικὸς) Γορτυνίων πρόξε|νος καὶ πολίτας αὐτὸς καὶ ἔγγονοι. He was a soldier (apparently in the army of a Ptolemy),65 a proxenos of the people of Gortyn, and a citizen, a status shared with his descendants (ἔγγονοι). As such, he had long-term links with Crete, Egypt, and Italy. The members of a religious association named as dedicators on an inscription in honour of Ptolemy XIII Auletes and his sister-wife Cleopatra at Hermopolis Magna (I.Hermoupolis 5) may have named an Ἀ[ρρι]αν[ός] (line 16: SB 1.4206 even gave Ἀρρι̣ αν̣ 63

64

65

See Bell (1924: 20 and 26). Skeat, who printed Δ̣ίννος Ῥωμαῖος (1974: 84 and 85), ̣ according to Fraser (1972: 170), regarded Δ̣ίννος Ῥωμαῖος as ‘absolutely secure’. My ̣ own examination of the papyrus confirmed Δ̣ίννος (19.5.2017: British Library, London). Although the first letter is different from the word-initial deltas in lines 3, 5, 7, and 11, all of which are different to each other, it is the same as the initial delta in Διονύσιος (line 17). That delta itself resembles the second omicron in that name. CIL III 14412.4 and VI 3164 could have been cited. See also AE 1969/1970.174, InscrIt 3.1 117 (151–230 CE)), and O.Did. 208 (c. 77–92 CE). The genitive Dinni in Dana (2014: 195), like that in CIL X 3572, is ambiguous. It could be the genitive of Dinnus or Dinnius. Dana (2014: 198) treated that instance of the name as Dinnis. Rigsby (1996: 251–2) resolved the abbreviation Πτο() as Πτο(λεμαιεὺς), which seems more likely. Gaius Lutatius Crispus would then be a soldier from Ptolemais-Barca in Cyrenaica; compare SEG 46.1221. Rigsby (1996: 252) drew attention to the Latinate features, namely the interpuncts, abbreviation mark, and tria nomina. An indication of Gaius’ origo (Πτολεμαιεύς) would be consistent with this Latinate character. This scenario (and the leaf at the end of the inscription) suggests a date after 27 and the existence of a province of Crete-Cyrenaica (cf. Cassio Dio 53.12.4).

254

Italians and Others in Egypt before the Caesars

[ός]) and certainly included an Ὡρίων (a Greek name with Egyptian origins), whose father had the name Άυείδιος (Avidius: I 69/18), and a Δαλλίων, whose father may have had the name Arrianus (I 18/6 (12.9.80 or 11.9.79)). The context of the second of these fathers is damaged on the stone. All that remains, as articulated by Milne (1905: 25), is Δα ̣λλίων ̣ Ἀ̣[․․]ά ̣νου and Bernand (1999: 37) printed Δαλλίων Ἀ̣ειάνου instead. The name Δαλλίων is not known to the ̣ LGPN, but Δηλίων and Δᾱλίων are, among other names (or nicknames in -ων), associated with Delos. That the dedicators were military personnel is indicated by παρεφ[εδρ]ε[ύον]τες ‘serving as guards’ (line 2). As with I.Hermoupolis 6 discussed above, the wider onomastic interest of this inscription lies in the number of Semitic names it presents.66 Italians in the Ptolemaic Administration There is far less evidence for Italians in civilian roles than in military contexts. The Oscan family of the Trebii are represented by a notary (μονογράφος) at Crocodilopolis (in the Fayum) in P.Petr. 2.32 2a.13–14 (216/215). The mobility of this family even at this early date is evident from the presence of members of the same gens at Delos (see Fraser 1972: 170). Italians Engaged in Commercial Activity Italians were involved in commercial activity in Egypt as early as the Zenon Archive. Names, but not ethnics, are attested in several documents. A Marius is listed as a source of timber in SB 26.16504 (a receipt found at Philadelphia, but perhaps written elsewhere, and attested by two partial copies: PSI 5.545.8 and P.Cair.Zen.1.59112.1). Since this receipt carries the date 26.11.257 BCE, this Marius is the earliest precisely dated and securely read instance of an Italian in Egypt. The name Statius is attested in two or three Zenon papyri67 and 66

67

See Horbury and Noy (1992: 249–50): Ἀβδοκώς (50; cf. 6.25), Ἀλαφᾶν (252), Ζαβινᾶς (31), Μίλιχος (25, 53, 152, 183, 241, and 250; cf. 6.129, 179, and 182), and Χαλαφάνης (67). P.Cair.Zen.4.59643.25 (263–229) and restored in PSI 6.585.1 and 5 (263–229). In the second papyrus, Amyntas quoted a letter that he received from Statius: Στατί]ου (line 1) and [Στ]άτιος (line 5). The third, P.Cair.Zen.1.59035.1 (31.1.257), reads Στα ̣ . . . ς

255

Patrick James

may refer to the same individual. A Varus is named in a fragment of a letter that appears to deal with accounts: P.Lond.7.2136 fr. 1.1 (263–229). A cluster of Italian praenomina and nomina on an ostracon from Edfu probably shows the involvement of Italians, among others, in the wine trade by the end of the Ptolemaic period. O.Edfou 3.371 lists seventy names, mostly in threes. Some names are repeated (and variously abbreviated), sometimes with further specification (πρ(εσβύτερος) and νε(ώτερος), e.g. in I 19 and 20). The ostracon is dated by an Egyptian month and a regnal year without naming the ruler. On the basis of the Italian names, the ruler has been assumed to be Cleopatra VII and the ostracon dated to 08.10.49. column I (ἔτος) δ Φαῶφι ε οἰνοδιχίαι Λότκις Ὀρεντᾶς Ευιέϊς —— Ἀλέξανδρο(ς) Ἑρμάϊς Κάνωπος —— Ἄττις Φαλλάϊς Τψόσις̣ ̣ —— Τιμότιος Μέλας Ζωΐνιο(ς) —— Ἀντώνιο(ς) Σώκινδρ(ο)ς Ζαννάϊς ——

column II

column III

Ἄττις Σώκινδρος Φαλλάϊς —— Δαμᾶς πρ(εσβύτερος) Δαμᾶς ν ̣ε ̣(ώτερος) Ἑρμάϊς —— Τείρων Νικάνω(ρ) Σέλευκος γι( ) —— Περιτένιος Κοΐνθος Μέλας —— Σοκῶνις Ὀρεντᾶς Λότκις ——

Ἑρμίας Δαμᾶς πρ(εσβύτερος) Δαμᾶς νε(ώτερος) —— Τιμόθις Πισίδης Σώκινδρ(ος) α —— Ἄβδηλος Ζανν(άϊς) Σώκιν(δρος) α —— Πισίδης Τιμόθις Δαμᾶς νε(ώτερος) α —— Μᾶρκος Περιντέν(ιος) α Δέκμος Αὐρήλις

line

1

5

10

15

Ζήνωνι χαίρειν. As its editor noted, Στά ̣τ ̣ιο̣ ̣ς (or Στά ̣[τιο]ς) suits the gap better than Στάχυς (or Στλάκκιος, the shortest name known to begin Στλ-).

256

Italians and Others in Egypt before the Caesars (cont.) column I

column II

column III

Ἀπολλ(ώνιος) με( ) Ἀπολλ(ώνιος) στ( ) Ἀπολλ(ώνιος) ν ̣υ( ) —— Δαμᾶς πρ(εσβύτερος) Δαμᾶς [νε(ώτερος)]

Ζαννάϊς Ἀντώνι(ος) Ουίβις —— Γάϊς Νεμέρι Μιοριάλ —— Μᾶρκος Ἄβδηλος

Κοΐνθος β —— Σαβοῦρος Ἀπολ(λώνιος) με( ) α —— Ἑρμίας Δαμᾶς πρ(εσβύτερος) Φαλάκιος —— Νικόμαχο(ς) Σέλευκος Πομπήϊς β —— Διεντᾶς Ἀντών(ιος) α

line

20

25

One of two clues as to the activity of Aurelius, Antonius, Decimus, Gaius, Marcus, Pompeius, Quintus, Tiro, and, probably, Vibius and, perhaps, Lucius,68 at Edfu is οἰνοδιχίαι in the heading. This word is unattested elsewhere and seems to have been understood to be a masculine plural of οἰνοδιχίας,69 an addendum lexicis. Other analyses are possible70 and the reading has been silently emended by some.71 It is at least clear that wine is involved, and the suggestion of a relationship with words concerning distribution (διχάζω, διχαίω, and δίχα)72 is reasonable. Whether these Italians were the buyers, sellers, or third-party recipients of some kind is not made clear by the ostracon nor is the purpose of these distributions. 68

69

70

71

72

Dack (1980: 41) proposed Λοῦκις (Luci(u)s) for Λόκτις, which is plausible, but this proposal was not said to be the result of a fresh examination of this ostracon. Manteuffel (1949: 108) and Honigman (2003: 74 note 20) refer to the ostracon as a ‘liste de noms de marchands de vin’. Dack (1980) remained non-specific. The word could be a feminine of an adjective *οἰνοδίχιος or the plural of an abstract noun *οἰνοδιχία, both of which would be addenda lexicis. Tcherikover, Fuks and Stern (1964: 255) gave the meaning as ‘distributions of wine’. Kiessling (1970: 194) included an entry οἰνοδοχία (‘receipt of wine’) with this ostracon as the only citation. It is not clear whether this word is a conjecture or the result of a fresh examination of this ostracon. Michalowski et al. (1950: 344).

257

Patrick James

The other clue is the find-spot. One of the main routes from the ports of the east coast of Egypt to the Nile valley ran through the wadis that emerged at Edfu, Graeco-Roman Apollonopolis Magna.73 Edfu was the site of the rock-temple of Amon Re and the Ptolemaic temple to Horus-Apollo completed in 57, less than a decade before the most likely date of this ostracon (we may wonder, then, whether the instances of the name Ἀπολλώνιος on this ostracon had particular local associations). The distributions of wine may have been either related to the temple or purely commercial (i.e., unrelated to the temple). There is little that would confirm that this is an account74 and the assumption that the Italians were soldiers.75 Apart from a few instances of α and β, which might be numerals or designations for common names, only the name Ἀπολλ() (I 16–18 and III 19) is followed by abbreviations that could concern quantities. However, those abbreviations could be designations of different individuals with that name, as with Δαμᾶς the older and the younger (I 19–20). The array of names on the ostracon shows that bearers of Italian names were not ethnically isolated. There are several Greek names (e.g. Ἀλέξανδρος I 4) and Greek adaptations of Egyptian names (e.g. Κάνωπος I 6, Egyptian Gnps). The name Πισίδης may be the ethnic Pisidian, which, as a personal name, suggests a slave, probably from Asia Minor.76 Other names are purely Egyptian, such as the feminine name Τψόσις̣ ̣ (I 9), perhaps prefixed by the (feminine) definite article. The uniquely attested Σοκῶνις seems to reflect an Egyptian theophoric name built on Sobek. If the name Ὀρεντᾶς is a rendering akin to Ὀροντᾶς,77 Ὀροντῆς, and Ὀρυντᾶς, or a modern misreading for one of those spellings, this name would be Iranian (see Huyse 1990: 53–4). Dialectal (non-Koiné) Greek Ὀρέντᾱς is also possible. The largest non-Italian share of the names has Semitic origins. Sylvie Honigman (2003: 74–5 note 28) identified Ἄβδηλος, Ευιέϊ ς, and Ζαννάϊς, not only as Semitic (as 73 74 75 76 77

For Edfu, see Fraser (1972: 177). See, for example, Capponi (2005: 16). See, for example, Tcherikover, Fuks and Stern (1964: 255), and Capponi (2005: 16). For ethnics as slave names, see Fraser (2009: 104 and 217–18). Compare I.Hermoupolis 5.137, discussed above, and Ὀρόντας, τὸ γένος Μ̣ ῆδον (IG IX.12 3, 638, 2.10 (c. 141/140 BCE)).

258

Italians and Others in Egypt before the Caesars

Tcherikover, Fuks & Stern 1964: 255), but, specifically, as Nabatean. She added Φαλλάϊς and even Ουίβις as a Greek form of a hypocoristic name built on the Arabic root whb, which appeared in names of the ‘given by a god or the gods’ type. The name Μιοριάλ, which is unattested elsewhere, also may be Semitic, even though it was not identified as Semitic by Wuthnow (1930: 77), Ruozzi Sala (1974: 28), or Ilan and Ziem (2008: 728). Although its final -αλ suggests a Semitic name (hence, its word-final accent), the amount of abbreviation shown in this text calls for caution. The origins of Σαβοῦρος78 and Νέμερι79 are disputed, and the origin of Σώκινδρος (I 14, II 2, III 6, and III 9) remains unknown. We do not know whose record of distributions this ostracon was nor do we know its purpose. Its Egyptian month suggests a nonRoman context. Also, confusion about voiceless stops and aspirates (Κοΐνθος II 11 and III 17 (followed by β) and Τιμότιος I 10) is conspicuous and could be the result of Egyptian interference on the Greek of this writer.80 However, the frequent syncope in -ι(ο)ς81 is not decisive in situating this ostracon. All that can be said is that by the end of the Ptolemaic period Italians were involved in distributions of wine of some kind in conjunction with Egyptians (both men and, as Τψόσις shows, a woman), with other locals (perhaps, originally, of Greek ethnicity), and with visitors or residents, particularly some whose names are likely to be Nabataean. Second-century loan contracts, accounts, and receipts involve people from throughout the Mediterranean in commerce with bearers of Italian names. A Ῥοῦφος was named in relation to 78

79

80 81

The name Σαβοῦρος seems to be a by-form of an Egyptian name Sbwrs, elsewhere rendered into Greek as Σαβοῦρις (sometimes with Σαπ- or Σεβ-), but Huyse (1990: 55–7) proposed an Iranian origin. The name Νέμερι was classed as Egyptian by Manteuffel (1949: 109), but could be a rendering of Latin Numerius, which was usually represented as Νομέριος in Greek. The only other attestation of Νεμερι- occurs in a first-century CE list of soldiers at Myos Hormos (P.Quseir 1 ~ SB 20.14248). Line 22 names Ἰσίων Νεμερίου, but Ἰσίων Νο ̣μερίου and Ἰσίων] Ν̣ο ̣μέ̣ ̣ρι̣ ο̣ ̣υ ̣ are named in lines 4 and 14. Although the presence of is uncertain, as it stands, the text shows fluctuation in the spelling of Numerius, the name of the father of Ἰσίων. For these phenomena, see Gignac (1976: 86–7, 90–2, and 95). For this declension, see Gignac (1981: 25–9).

259

Patrick James

a κεράμιον (BGU 6.1490.4 [second or first century: Edfu]). The wording τὰς ἐν Προπέλλωι (P.Tebt.1.99 ii 48) has been understood to refer to artabae ‘owned by’ Propellus.82 SB 6.9552 iv 4 (26.9.136 or 25.9.135: Coptos), a tax receipt, includes a greeting for one Comes, apparently one of a group of makers of garments. The name Κίντος appears with the ethnic Μεσσαλιώτης,83 a Γνάϊ ος is mentioned (SB 3.7169.12), and the ethnics [Λακε]δαιμόνιος, [Θ]εσσαλονικεύς, Ἐλεάτης, and Καρ[χη]δόνιος also appear (SB 3.7169, 17, 18, and 20). Ulrich Wilcken (1925: 97) may have been correct to distinguish Κίντος from Κόϊντος,84 the usual transcription of Quintus, and to identify here Cintos, a Celtic name. Clearer evidence of Celtic names is found in a letter from one Lysimachos to one Sosios (P.Michael. 7r3–4 [08.4.136 or 11.4.147]), in which an Adminius is named as the recipient of forty drachmas of bronze. La’da (2002: 11) suggests that his ethnic τὸν Α̣ἴδαιον may be that of the Gallic Aedui, since his name, printed as Ἀ̣δμίνιον ̣ with due caution, is that of the prince of the Cantiaci who was banished by his father Cunobelinus and surrendered to Caligula.85 Relations between Egyptian officials and Italian negotiatores at Delos are evident in the second century. The brothers Pedii, Gaius. and Lucius,86 Ῥωμαῖοι, whose filiation was expressed in the Roman manner (line 5: Γαίου υἱοί), left a statue to honour [Πολέ]μαρον in a dedication to Apollo and Artemis (OGI 1.133 ~ ID 1527 (c. 127)). The honorand was described as an epistrategos and 82

83

84

85

86

Grenfell, Hunt, and Smyly (1902: 441). For such a non-Classical use of ἐν (a genitive and ellipse of the dative would be expected: cf. LSJ s.v. ἐν A I 2), they cite parallels (1902: 56, 114, and 441), of which P.Tebt.1.72.332 and 120.129 are particularly relevant. SB 3.7169.19 and 21 and 7170.10 and, perhaps, 15. The papyrus has a dative Μεσσαλιώτηι for a nominative (line 19). That ethnic also occurs in line 11, where the names themselves have been lost: καὶ Τρ. . .ωι Τρε ̣ . .[ . . Μεσ]σαλιώτηι [ ̣ ̣ .̣ Wilcken (1925: 98 note 1) cited a fragment in Berlin (P.5835, still unpublished: see www.trismegistos.org/text/8471), whose second line reads] . εντίωι Κοίντου Κα[λ]ηνῶι. This adds another Quintus as the father of a man, whose ethnic was Καληνός (that is, of Cales in Campania) and whose name ended -έντιος. See La’da (2002: 115). Suetonius, Caligula 44.2. See Russell and Mullen 2007 (accessed 24.8.2017). It is not clear whether the note that the name is found elsewhere in the Roman Empire took this Egyptian instance into consideration. The name on the stone, given by OGI 1.133.5 Λ(ε)ύκιος has been read either as ΛΥΚΙΟΣ or .ΥΚΙΟΣ. As Marcus Tod (1951: 176) suggested, Λύκιος here is probably a transliteration, rather than a transcription, of Lucius (although with Greek morphology: -ος).

260

Italians and Others in Egypt before the Caesars

an honorary ‘cousin’ (τὸν συνγενῆ: 1) of Ptolemy VIII Euergetes II (146–116 BCE). The occasion was his ἀρέτη, καλοκἀγαθία, and εὔνοια (lines 6–8) towards the Pedii. (All this was more striking when the name of the honourand was read as Μᾶρκον.87) Dittenberger (1903–5: 215 and 216) asserted that Ῥωμαῖοι not of noble families and named without cognomina and without official or honorific titles should be considered to be negotiatores. Corroboration is found in OGI 1.135.1–3 (~ ID 1526; 127): Ῥωμαίων οἱ εὐεργετηθέντες ναύκληροι | καὶ ἔμποροι ἐν τῆι γενομένῃ καταλή|ψει Ἀλεξανδρείας. The existence of a community of Italici at Alexandria is attested by a Latin–Greek bilingual inscription that they left during their time at Delos (ID 1699.1 [99 BCE]: Alexandreae Italicei quei fuere). As Adams observed,88 this inscription shares several features with other Latin inscriptions that refer to Italici, both at Delos and elsewhere: biversion bilinguals with spellings, the Greek accusative of the honorand, and a relative clause with a locative. In the context of these conventional features, the absence of negotiantur, the verb that was another such conventional feature, is striking and distinguishes these Alexandrian Italici from those on Sicily and in the province of Achaea. Instead of an explicit indication that trade was their reason to be at Delos, the inscription has the third-person plural fuere and thus shows the continued use of old and more prestigious morphology.89 Since the epigraphy of the Italici is so peculiar to the Italici, it is likely that these Alexandrian Italici themselves selected this form and, thus, we may glimpse a little of their variety of Latin. The perfect tense indicates that they had previously been at Alexandria, but leaves the character of that time unclear. Was their association with Alexandria now entirely past? If they were based there for a time, had they been based elsewhere in the Mediterranean earlier still?90 The Greek 87

88 89 90

Tod (1951: 176), who examined the stone, thought [Πολέ] too long for the available space and suggested [Δή]μαρχον or [Τί]μαρχον, with reference to ID 1534 Σίμαλος Τίμαρχος Σαλαμίνιος. He rejected Μᾶρκον, not only because of the space preceding, but also because the spelling Μάαρκον would be expected at this date and the letter before the -ον seemed not to be , as the editions print, but . The discussion in Adams (2003a: 650–1) did not cover the Greek text. For the status of such a form, see Clackson and Horrocks (2007: 100–1). The participle [παρα-]|[γενόμενοι (lines 5–6), if it is the correct restoration, does not resolve these questions: these Italici came (to Delos).

261

Patrick James

counterpart implies that Alexandreae is locative and belongs with fuere in the relative clause Italici based at Alexandria: οἱ ἐν Ἀλεξανδρείαι (line 5: sc. Ἰταλικοί; cf. line 1 and Ἰταλικ]οί line 6).91 However, the rest of the Greek text is so damaged that we are left wondering whether their presence at Delos was intended to be a brief, a longer-term, or an indefinite absence from Alexandria and whether we should see migration or mobility here. The first-century data may suggest the involvement of Italians in transactions that were recorded in Demotic, not Greek, and thus attest interaction, however indirect, between Italians and nonHellenised Egyptians. Ostraca from Oxyrhynchus were dated by their editors between the first century BCE and the first century CE (and to 146–132 on Trismegistos). Although the editors preferred not to speculate, Lpws (30 convex x+10) may be the Italic name Lupus. A Clemens appears, as Klims, as a party in a transaction of spelt (35 concave I x+5). A Sabinus (as Spnws) is named as the father of a Peteamounis (14 concave 6).92 The same combination of names is found on another ostracon (11 concave 4).93 P.Qasr Ibrim Dem. 19, an ostracon found at Qasr Ibrim (the find-spot of the famous Gallus Papyrus), contains wrs followed by a ‘foreigner-determinative’ (Ray 2005: 44) and no further text. This is probably a transcription of Verus or Varus.94 The purpose of the ostracon is unknown. Ray (2005: 44) commented that, if the

91

92 93 94

The position of Alexandreae before the relative clause is noteworthy. In the Italici inscriptions of Delos, the locative is found, as expected, after the relative pronoun: Italicei | et Graecei quei Delei negotiantur (ID 1695.1–2 (88); cf. ID 1696.2–3 (88) and 1698.1 (82). (ID 1620.2–3 (84–80) has a prepositional phrase in the same position and function: Italicei et | Graece[i que]i in insula negotiantur.) Although the Greek favours ‘who were at Alexandria’, those who read only the Latin itself (not the Greek) could construe Alexandreae with Italicei or with legat]um. That would leave the question of where they were (fuere) without an explicit answer. It is worth noting that only ten inscriptions from Delos refer to Italic(e)i, of which only three have the locative Del(e)i. That is a very limited set from which to form a notion of the normal constitution of the epigraphy of the Italici on Delos. There are a few further references to the Italici of Delos in Greek inscriptions: ID 1683, 1688–91, 1717–18, 1722, 1727, 1736, and 1757 (including plausible supplements of Ἰταλικ-). Bresciani et al. (1973: 249–51, 259–61, and 223–5). Bresciani et al. (1972: 345–7). Ray (2005: 44) notes that consonants wrs might be either the well-attested name of a plant or a noun meaning ‘head-rest’, but the ‘foreigner-determinative’ that follows suggests that a personal name is involved.

262

Italians and Others in Egypt before the Caesars

text does refer to a Varus or Verus, then ‘a date at the very end of the Hellenistic period or the century following’ would be likely. In the final years of Cleopatra VII, we find a highly mobile and exceptional individual who takes us from isolated names and occasional clusters in documentary texts to history as known from literary sources. P.Bingen 45 (23.2.33), a royal ordinance, granted tax exemptions on imports of Coan wine in Coan amphoras, exports, and landholdings in Egypt to Publius Canidius (once reedited by Minnen [2000: 32–3]), a Roman citizen and Mark Antony’s most important commander. Cleopatra may have granted these privileges to him, in her own handwriting (Minnen 2000: 30–1), only in exchange for his allegiance to her. However, since the exemptions applied to Canidius’ heirs, we see one Roman who had long-term interests and static assets in Egypt as well as relations with Cos. We may assume that he had Latin-speaking associates to manage those assets on his behalf, even if his tenant farmers (line 9) were all indigenous, non-Hellenized Egyptians. Latin as a Symbolic Language of Authority One Latin text remains to be considered before concluding this study: OGI 1.129.11–12. Four Latin words follow, and summarise, rather than translate, the import of the preceding Greek text, itself containing a quotation of an earlier Greek document. βασιλίσσης καὶ βασιλέως προσταξάντων, ἀντὶ τῆς προανακειμένης περὶ τῆς ἀναθέσεως τῆς προσευχῆς πλακός, ἡ ὑπογεγραμμένη ἐπιγραφήτω. βασιλεὺς Πτολεμαῖος Εὐεργέτης τὴν προσευχὴν ἄσυλον. regina et rex iusser(un)t.

Since the queen and the king gave the order, instead of the tablet that was set up previously about the dedication of the synagogue, let the following text be inscribed: King Ptolemy Euergetes (declared) the synagogue inviolate. The queen and the king gave the order.

The queen and king – in that order – who gave the ruling that Ptolemy Euergetes’ (probably II: 146–116) grant of asyleia to this 263

Patrick James

synagogue (line 5: προσευχή), probably at Leontopolis in Lower Egypt, should remain in force, are most likely to be Cleopatra VII and either her son Caesarion or, perhaps, one of her brothers. Rigsby (2003) dated the inscription, on the basis of its letter forms, to the end of the Ptolemaic period and has argued that the Greek frame and the Latin are contemporary, since the Greek is increasingly compressed to leave space for the large Latin letters. We cannot tell whether the same stonemason carved both the Greek and the Latin and whether the Latin originated with the Royal Chancellery or with the synagogue itself. Here, we have a bi-version bilingual text, but one that is ‘partial’ or ‘unequal’ in Adams’ terms (2003a: 33 and 35). As Rigsby (2003: 132) commented, ‘the Latin is not informative: it would not tell a truly Greekless reader what to do’. In Adams’ terms (2003a: 34–5), the Latin is ‘symbolic’. The Latin functions, in Rigsby’s words (2003: 132), as ‘a warning that even Romans must pay heed to the Greek text’. The implicature of the Latin royal titles and abbreviation of iusser(un)t is ‘Latin readers, hands off!’.95 Since the synagogue’s history of asyleia from the time of Ptolemy (VIII) Euergetes (II) was not communicated to readers of the Latin text, we may reasonably suspect that there was no Latin speaking (or reading) community associated with the synagogue, either of Jews or of proselytes, to whom the peculiar historical, religious, and legal status of the synagogue would be of interest. The presence of any Latin was a warning to those Latin-reading outsiders who might violate the synagogue’s asyleia, most likely Roman soldiers. For those who could read the Greek, but not the Latin, the Latin script indicated that the grant of asyleia had the weight of Rome behind it. The Greek reports that this inscription replaced another πλάξ (lines 3–6: ἀντὶ τῆς προανακει|μένης περὶ τῆς ἀναθέσε|ως τῆς προσευχῆς πλα|κός), the only contemporary addition to the content of the Latin text. We may wonder whether that earlier statement of the grant of asyleia entirely in Greek 95

IGUR 291, discussed by Adams (2003a: 35–6), illustrates this by way of contrast. The Latin text that follows the Greek (lines 11–13) relies on the funerary context, if not the Greek text itself (hoc has no referent in the Latin text and there is no limitation to sis molestus as provided for ἐνοχλήσῃς by τῷ τάφῳ in the Greek). Unlike the Greek text (lines 1–8), the Latin does not commemorate the deceased. Instead, Latin is used only for an explicit warning against violating the tomb (lines 11–13), restating the warning that concluded the Greek text (lines 8–11).

264

Italians and Others in Egypt before the Caesars

had been damaged by Latin speakers or had simply fallen into disrepair such that it needed to be replaced, given the presence of foreign, Latin-reading, military personnel. Conclusion Long before Ptolemaic Egypt became Roman Aegyptus, numerous bearers of Italian names and ethnics participated in the often overlapping sacred, military, and commercial spheres of activity. Indeed, Italian names of those engaged in military and commercial activities are attested as early as the Zenon Archive with certainty, while others at Philae and the Syringes may also be as early as the third century. These attestations present a striking contrast with our literary sources for Roman activity in the Eastern Mediterranean. However, the move to showing that bearers of Italian names (or of names that may be Italic) were in fact Italians and, indeed, speakers of Latin or an Italic language has been beset with various difficulties. Of the many names reported by Trismegistos that are neither Greek nor Egyptian, many have been discounted on the grounds of several kinds of ambiguity, many more than it has been possible to discuss. Several individuals traditionally identified as Italians have been shown to be unlikely to be so. Italian names have been found with non-Italian ethnics, thus showing some degree of mobility and migration, while bearers of the ethnic Ῥωμαῖος with common Greek names, such as those attested at Abydos, were probably Greeks from Italy in Egypt, not Romans. A significant discovery has been the association of these bearers of Italian names not only with bearers of Egyptian names, but also with bearers of Semitic and Iranian (and, perhaps, Celtic) names as well as bearers of Greek names that could have been used by people of all those ethnicities. The Ptolemaic sphere of influence was polyethnic, and its Italians, whether attested as individuals or in clusters, seem to have been integrated, rather than isolated, and involved with populations whose names, at least, had diverse origins. They were not only involved in superficial contact with Greeks for commercial purposes, as they were throughout the Mediterranean. The possibility of attestations of Italian names in 265

Patrick James

Demotic texts reflects a particularly fascinating degree of contact. Italian names that appear in lists of members of religious associations in military contexts alongside Semitic and Iranian names demonstrate integration of populations with diverse onomastic origins or, at least, the diffusion of names from various sources. Our search for Italians in Ptolemaic Egypt has traced mobility further afield. Commercial activity was in conjunction, not just in contact, with Old Greece (a Lacedaemonian, a Thessalonian, and Coans), but also with the wider Hellenistic world (a Messaliot, a Carthaginian, and Nabateans at Edfou). Italians at Delos with links to Alexandria even had connections with honorary ‘cousins’ of the Ptolemies, and Italians served the Ptolemies as commanders on Crete. It is hard to assess the extent to which these bearers of Italian names reflect mobility in the Mediterranean, and within Egypt, or longer-term migration to Egypt. Those who visited Abydos, Philae, or the Syringes at Thebes and left their marks were certainly Italians resident elsewhere in Egypt for commercial or military or diplomatic purposes (the senator Lucius Memmius would be a notable exception, but his ‘sightseeing’ followed a stay at Alexandria). Although Publius Canidius had long-term interests in Egypt and elsewhere, he himself was not confined to Egypt. Commercial links might have been brief or long-term. Still, the extent of integration that we have seen, particularly in the military sphere, suggests some degree of permanent residence after migration. The little Latin known from Ptolemaic Egypt comes mainly from bilingual texts or contexts, entirely so if one Marcos Volturios was responsible for all four of the graffiti at the Syringes that were discussed above. The Latin of the Acutius inscription not only exceeds the content of the Greek part, it does so with a particularly Roman emphasis. The short Greek graffito Γάϊ ος Ἀκούτιος would not have stood out among the many names incised on the walls at Philae, but the full and precise consular date with a Roman numeral are conspicuous, even for those who could not read the Latin, and thus make a statement that enforces the claim hoc primus uenit. At the Syringes, M. Volturios, in Latin script and abbreviated, was clearly Latin (neither Greek, 266

Italians and Others in Egypt before the Caesars

nor Egyptian), while Marcos Voturios, especially when incised in Greek, needed to be distinguished as Ῥωμαῖος. Although the amount of Latin may be minimal in relation to the Greek, as in the addition to the royal grant of asyleia to a synagogue, or maximal, as in the case of the graffiti of Acutius and his companions, or more balanced, as with the graffiti at the Syringes, the use of Latin seems to have had particular impact as a marker of nonGreek and non-Egyptian identity and, for the last of the Ptolemies, of authority.

267

CH A P T E R 11

P O P U L AT I O N, MI G R AT I O N A N D L AN G U A G E I N T H E C IT Y O F R O M E 1 O LIV IA EL DER

Recent studies have evaluated the patterns and extent of migration to Rome and highlighted the presence of groups with different, but intersecting, social, cultural and legal statuses.2 In the late Republic and Early Empire there was a continual influx of foreigners from Italy and further afield, both slaves and free, leading to ancient and modern characterisations of the city as a diverse and multicultural cosmopolis. In the late second or early third century AD, Athenaeus described Rome as ‘the epitome of the whole world’, the city that contained all others within it.3 Edwards and Woolf’s volume, Rome the Cosmopolis, opens with the spectacle of the games at the Colosseum, the spectators coming from right across the Roman Empire: ‘Marked out by their exotic clothing and hair arrangements, their incomprehensible speech, these people embodied the vastness and diversity of Roman territory, their presence in the heart of the city underlining Rome’s power to draw people to itself.’4 Language is here given as one of the markers of the city’s cosmopolitan character. There are parallel glimpses of Rome’s multilingualism, driven by migration, in 1

2

3 4

I would like to thank James Clackson, Patrick James, Katherine McDonald and Andrew Wallace-Hadrill for their comments on this chapter. I would also like to extend thanks to Peterhouse, Cambridge for funding my PhD research and to both Peterhouse and the Faculty of Classics, Cambridge for funding a term spent at the British School at Rome, which proved essential in the development of this chapter from my original Laurence seminar paper. I am also grateful to Andrew Wallace-Hadrill and Greg Woolf for giving me copies of their recent papers on migration in the city of Rome prior to publication (Wallace-Hadrill (2017) and Woolf (2016) respectively); they have proven invaluable to my thinking in this chapter. The scholarship on various aspects of migration to Rome is vast. Tacoma (2016a) is a recent in-depth study; De Ligt and Tacoma (2016a) includes several chapters on Rome. On Rome’s demographics, see also Lo Cascio (2016); Scheidel (2004 and 2005). Significant studies of foreign populations at Rome include Noy (2000); Polomé (1983); La Piana (1927). Ath. Deip. 1.20b; see later in this chapter for further discussion of Athenaeus. Edwards and Woolf (2003: 1). On the idea of Rome as a cosmopolis see also Moatti (2013).

268

Population, Migration and Language in Rome

ancient literature: Juvenal, writing in the late first or early second century AD, lists the Greek language as the first of the foreign elements with which migrants from the Greek East have flooded the city.5 Suetonius mentions plays performed ‘in every language’ in the Rome of Julius Caesar and Augustus.6 In general, however, this multilingual city proves surprisingly elusive. Literary evidence is limited, and Rome’s epigraphic record is overwhelmingly monolingual: there are no more than a handful of inscriptions in languages other than Latin or Greek, and Greek inscriptions are far outnumbered by those in Latin.7 Taking this gap between expectation and evidence as a starting point, this chapter interrogates the relationship between language and migration in the city of Rome. The gap is partly the result of the nature and limitations of our evidence. The epigraphic display of a language requires a distinct set of choices and circumstances and is not necessarily reflective of the city’s linguistic diversity: the geographical breadth of migration to Rome means that the numbers of languages that could have been spoken in the city was large.8 In terms of understanding the types of identities displayed at Rome, and the impact of migration on the landscape of the city, the overwhelming linguistic homogeneity in the epigraphic evidence is, however, significant. Moreover, the distinction between a monolingual and a multilingual Rome is not systemic to particular types of evidence. It also uncovers a real debate about the place of, and attitudes towards, migrant groups and identities at 5 6

7

8

Juv. 3.62–3. Suet. Iul. 39.1 and Aug. 43.1. As Tacoma (2016a: 220) points out, it is not clear what languages this comment is meant to refer to. He suggests that Oscan is the most likely candidate for a language other than Latin and Greek; Adams (2003a: 118) brings in some further evidence in support of this. Unless languages other than Latin or Greek are meant, it is unclear to me why the standard phrase utraque lingua is not used here. Whatever the empirical reality behind this comment, the intention of stressing the city’s linguistic diversity seems clear. For overviews of the evidence for different languages at Rome, see Tacoma (2016a: 214–23); Noy (2000: 169–79); Adams (2003a: 248–53), on Palmyrene, and (2003a: 356–82), which considers the topic of ‘identity’ through several examples of Greek from the city of Rome; Kajanto (1963,1980); Moretti (1989) on Greek specifically. For discussions of the knowledge of languages at Rome, see Boyancé (1956); Dubuisson (1992); Kaimio (1979). Other studies focus on a single group: Felle (1997) on Christians, and Ricci (1993) on Egyptians. On the origins of Rome’s migrants, see Noy (2000: 53–84). Solin (2007: 1376–9) provides addenda and corrigenda to Noy’s geographical list.

269

Olivia Elder

Rome, and the significance of language choices in reflecting and promoting these. Overall, migration helps to explain both the presence and the absence of languages in the city. Graecam Urbem Juvenal’s third satire provides one framework for assessing the impact of migration on the city, and the role of language within this. It is a useful way into several of the themes and issues at stake in this chapter. On the point of leaving his native Rome, the poem’s protagonist, Umbricius, pauses at the Porta Capena to give a long speech in which he outlines the city’s many social problems that have enforced his departure.9 He blames these problems on the influx of foreigners who have altered Rome’s character and reduced the opportunities available to native inhabitants. quae nunc diuitibus gens acceptissima nostris et quos praecipue fugiam, properabo fateri, nec pudor obstabit. non possum ferre, Quirites, Graecam urbem. quamuis quota portio faecis Achaei? iam pridem Syrus in Tiberim defluxit Orontes et linguam et mores et cum tibicine chordas obliquas nec non gentilia tympana secum uexit et ad Circum iussas prostare puellas. ite, quibus grata est picta lupa barbara mitra. rusticus ille tuus sumit trechedipna, Quirine, et ceromatico fert niceteria collo. hic alta Sicyone, ast hic Amydone relicta, hic Andro, ille Samo, hic Trallibus aut Alabandis, Esquilias dictumque petunt a uimine collem uiscera magnarum domuum dominique futuri.

60

65

70

I’ll rush to name the race that is now most pleasing to wealthy Romans, and which I especially aim to escape – shame doesn’t stop me. My fellow-citizens, I cannot bear a Greek Rome. But what proportion of the dregs are Achaeans? For a long time now the Syrian Orontes has flushed into the Tiber and carried with it language and customs, the flute-player and the slanting strings, foreign drums, and the girls 9

On the role of Umbricius’ name in constructing him as ‘the umbra (“ghost”) of Romanness’, see Braund (1996: 232). Much has been written on the relationship between Umbricius’ voice and the voice of the poet: Uden (2015: 104), for example, notes that ‘the poem is notoriously ambiguous about how distinct this character is from the voice of the poet himself’.

270

Population, Migration and Language in Rome who are ordered to put themselves up for sale at the Circus. Go, those of you who are satisfied by a foreign whore in a painted headdress. Quirinus, that rustic of yours puts on his Greek slippers, and wears his Greek victory prizes around his Greek-perfumed neck. They come—this one leaving the heights of Sicyon, this other from Amydon, this one from Andros, that one from Samos, this one from Tralles or Alabanda. They seek out the Equiline and the Viminal, the hill named for the willow, and will become the guts and masters of our great houses.10 Juvenal Satires 3.58–72

Several aspects of Umbricius’ characterisation of these foreigners have significance for our understanding of the ways that migrants and their languages appear in the city. They also have echoes in both ancient literature and modern analyses of the relationship between migration and language at Rome. First, these foreigners are depicted as a highly visible group that is marked out by distinctive non-Roman traits, including language and customs (et linguam et mores l.63). The gap between foreigners and Romans is emphasised by the fact that Umbricius addresses his complaints, not to his interlocutor, but first to Quirites, the formal collective term for the body of the Roman citizens, and then to Romulus (Quirine). It is striking that Umbricius here singles out Roman citizens: I show below that citizenship was a significant category in determining language use. Language is itself a way for Juvenal to draw a contrast between Romans and foreigners and to highlight the impact that migrants have had on the city. Umbricius deploys a series of Greek words to describe, with obvious disdain, the changes to the habits and behaviour of Rome’s inhabitants under the influence of the Greek-speaking migrants: rusticus ille tuus sumit trechedipna, Quirine, et ceromatico fert niceteria collo; ‘Quirinus, that rustic of yours puts on his Greek slippers, and wears his Greek victory prizes around his Greek-perfumed neck’. These words are rare and obscure even in Greek; their ‘foreignness’ is underlined by the fact that this passage is the only time that they appear in extant Latin literature. The direction of this tirade against Greek speakers, specifically those from a widely conceived Greek East, is striking.11 Juvenal 10 11

Translation adapted from Braund (2004: 171). James (2015: 119, 124, 126–7, 129 and 132) is useful on some of the ethnic identifiers in this passage.

271

Olivia Elder

here constructs Greek as a migrant language and makes no mention of other local languages that migrants might have brought to the city. The status and function of Greek in relation to other languages, and the extent to which it can be understood as a migrant language, is a recurring question in this chapter. Although Umbricius cannot endure the idea of Rome as a Greek city, Greek was deeply embedded there.12 Umbricius suggests that these migrants cluster in particular areas of the city: they make for the Esquiline and the Viminal hills. Umbricius’ speech itself takes place at the Porta Capena, a space described in lines 12–14 of the poem as having been transformed from its mythical and sacred Roman past and now taken over by the Jews (hic ubi nocturnae Numa constituebat amicae | nunc sacri fontis nemus et delubra locantur | Iudaeis; ‘here where Numa used to meet his girlfriend at night, now the grove and shrine of the sacred spring are leased out to the Jews’). Philo gives evidence for another Jewish community in Trastevere already in the first half of the first century CE (Philo Leg. 155). Similar spatial clustering of migrant groups is seen in Athenaeus (Deip. 1.20 b) who describes how the different communities settled at Rome can be seen in individual neighbourhoods (κατ᾿ ἰδίαν δὲ τὰς πολλάς). Umbricius also stresses the quantity of migrants. He claims that foreigners are importing new professions, taking over others and flattering their way into positions of power and influence to the point that by line 119 he can complain that ‘there is no place for any Roman here’ (non est Romano cuiquam locus hic). Some of these elements of Umbricius’ speech have been incorporated into modern assessments of the impact of migration on Rome and the place of foreigners in the city. Both La Piana and Polomé stress the clustering of immigrants within particular spaces and draw a direct link between migrants and language use, arguing that migrant populations continued to use their own languages for long periods of time.13 Noy to some extent follows 12

13

On Rome’s connections to the Greek world, and the idea of it as a Greek city, see Greca (2012); Purcell (2000: 440); Kamio (1979: 6). La Piana (1927: 204): ‘It is to-day, and has always been everywhere, the natural tendency of a body of immigrants from the same nationality in a foreign city to live

272

Population, Migration and Language in Rome

in this tradition. His methodology relies on being able to find markers of foreign groups in the city, and he sees language use through the prism of migrant presence, as part of the construction of foreign or migrant identities in the city.14 The satirical nature of Juvenal’s poem must, however, be stressed. The themes that the passage draws out – the visibility of migrants, the spaces they inhabit, their characteristics – have relevance to contemporary debates around migration and its impact, but the account it gives of them is distorted. This satire fits poorly with other evidence for migration and language in the city: indeed, Moretti takes the discrepancy between it and other evidence as his starting point for understanding the place of Greeks in the city.15 Other recent scholarship has directly challenged aspects of Juvenal’s depiction of the impact of migration on the city and the scholarship founded upon it. Contrary to La Piana and Polomé, Tacoma argues that the evidence for the ghettoisation or clustering of foreign groups in the city is limited: the Jews and their catacombs seem to be exceptional examples of separate ethnic and religious spaces in the city.16 He has also begun to challenge the study of language use in terms of migrants’ integration or acculturation at Rome, moving beyond a binary of integration or not, and acknowledging a spectrum of outcomes.17 Moatti suggests a different model for migrant identities in the city: she looks at the processes of exchange and mobility in identity formation and argues for an accumulation, rather than an exclusion or transfer, of migrant identities.18

14 15 16 17 18

together as much as possible in the same district, where they can reproduce the main characteristics of the social and religious life of the country from which they came’; Polomé (1983: 514): ‘though Rome was a ‘melting pot’, immigrants from the same area would tend to cluster together and preserve their language and culture, at least as long as the links with their country of origin were maintained’. On evidence for the clustering of migrant groups, see also Coarelli (2012), who considers the topography of ethnic groups at Rome. See Noy (2000: xii and 4–7) for his criteria of inclusion. Moretti (1989: 5). Tacoma (2013) and Noy (2000: 197). Tacoma (2016a: 206). Moatti (2014); see especially 149–51.

273

Olivia Elder

The Eternal City: Migration and Language in Ancient and Modern Cosmopoleis The evidence from contemporary cities provides an alternative framework for making sense of the impact of migration on language use in the city. Implicitly or explicitly, contemporary cities, with their high rates of immigration and super-diversity, are frequently used as comparators for Rome. Both La Piana and Polomé regard the behaviour of migrants in Rome as directly analogous to that of migrants in modern cities: La Piana goes so far as to describe migrant clustering and adherence to local cultures as a ‘universal phenomenon’.19 In recent papers, both Woolf and Wallace-Hadrill have highlighted the dangers of using contemporary cosmopoleis as models for understanding Rome, warning that doing so runs the risk of making the ancient city too familiar.20 In place of modern multilingual and multicultural cosmopoleis, Woolf offered a challenge to ‘try and imagine giant cities that were less cosmopolitan, diverse and multicultural than our own’.21 Many of the differences between Rome and modern cities are bound up with Rome’s distinctive social structure – in particular the dominance of slaves and freed – something that is often insufficiently acknowledged.22 Woolf is right to suggest that we reset our expectations of ancient cities, and that abandoning unrealistic understandings of the relationships between ancient and contemporary cities is one way to do so. His model of ancient cities as less diverse than hitherto assumed also provides a convincing framework for the limited evidence for multilingualism at Rome. I argue, however, that there is value in looking more closely at multilingualism in modern cities and in using this as a way into understanding language use in ancient Rome. At least part of the problem for ancient historians in using modern cities as comparators for Rome 19 20 21 22

La Piana (1927: 204). Wallace-Hadrill (2017); Woolf (2016). Woolf (2016). Both Woolf and Wallace-Hadrill highlight our underestimation of the numbers and effects of slaves coming to Rome. Tacoma (2016a: 144) comments that ‘the fact that a substantial part of migration to Rome was forced might be relevant [to the applicability of the modern urban migration theories that he uses]’.

274

Population, Migration and Language in Rome

has been not that we use them too much, but that we do not look closely enough at them. We rarely explicitly draw out the levels of diversity in modern cities and, crucially, the way that these appear within the evidence. A clearer understanding of modern cityscapes sheds light on the factors and structures that influence language use and in particular the relationship between language, immigration, integration and identity. This analysis helps to reset our expectations of Rome and provides a more accurate basis against which to measure the ways in which it is distinct. Studies of contemporary urban multilingualism are also useful in approaching the problem of evidence since they offer a complete, or nearcomplete, picture of language use in the city including evidence for spoken languages. We can see how linguistic diversity in the oral record translates into the written record, and this provides one useful point of comparison for the necessarily partial evidence for multilingualism extant from Rome. A recent project by LUCIDE (Languages in Urban Communities: Integration and Diversity for Europe), which collected and analysed the evidence for multilingualism in eighteen cities in Europe, Australia and Canada, provides some useful evidence for the impact of immigration on urban language use.23 The analytical tools and concepts deployed to study the evidence and the overall patterns that emerge from it are helpful in approaching the evidence from Rome. In all modern cities, migration is, predictably, the most significant factor behind linguistic diversity.24 This potential for linguistic diversity does not, however, necessarily translate into evidence for multilingualism of the kind that we can capture for ancient Rome. The LUCIDE reports demonstrate that the visibility of immigrant languages can vary significantly: although some languages connected with immigrants may gain an established place within the city, ‘the languages of many immigrants may remain largely invisible’.25 The size of the immigrant communities is one factor shaping the visibility of their languages, but other factors, such as the status and types of migrants speaking a particular language and the attachment of a particular community to 23 24 25

King and Carson (2016) provide a discursive overview of the data collected. King and Carson (2016: 88). King (2016: 186); see also Stoicheva (2016: 92).

275

Olivia Elder

language as a marker of social identity, are also significant.26 This concept of visibility – a measure distinct from the quantities of different languages and one that is frequently linked to their relative prestige in the city – is useful for thinking about the appearance of different languages at Rome.27 The LUCIDE reports also draw a helpful distinction between plurilingualism and multilingualism, that is, between the capacity of individuals to speak multiple languages and the presence of multiple languages within a particular area.28 This distinction is sometimes lost in analyses of ancient multilingualism, but has important consequences for our understanding of the relationship between language and migration and the impact of migration on the city more widely. Juvenal constructs both the possibility of a city where multiple languages exist in isolation and the possibility of a city where speakers of multiple languages interact. In his poem, migrants cluster in particular places, but the choice of the Esquiline as one of these places is significant: it is the place where Nero built his palace in this period, and in encroaching on it, migrants are also encroaching on a centre of imperial power. Umbricius is himself to some small degree bilingual, using the language of the migrants in his tirade against them. In general, the nature of Rome’s migration patterns can be seen to encourage interaction between speakers of different languages: the twin dynamics of slavery and the sheer diversity of Rome’s migrants, both discussed below, are significant here. Importantly, this kind of linguistic interaction between speakers of multiple languages does not necessarily (if at all) lead to permanent written evidence of a range of different languages within the city’s landscape, but can be one factor behind the use of or accommodation to a single language. The evidence for linguistic clustering provides one measure of how languages relate spatially within the city. The LUCIDE 26

27 28

Stoicheva (2016: 88–92). On the significance of the status of different languages and their speakers in determining the visibility of languages in the city, see Skrandies (2016: 143). On the variation in the importance of language as a marker of identity, see Stoicheva (2016: 104). On the concept of visibility, see Carson and King (2016: 7). King and Carson (2016: 9). This distinction between ‘plurilingualism’ and ‘multilingualism’ draws on one adopted by the Council of Europe.

276

Population, Migration and Language in Rome

reports note the frequent existence of spaces connected with particular languages or groups, although this does vary within and between cities.29 The reports also note the existence of liminal or transgressive urban space, where language patterns may be distinct from those in the rest of the city: the best example of material from these kinds of spaces is graffiti.30 Although the evidence for this type of space in Rome is limited, the potential for language use there to be transgressive is highlighted by Zadorojnyi’s discussion of ancient graffiti as a category.31 Crucially, the reports stress that linguistic clustering is rarely static: areas connected to a particular group do not necessarily remain so. Areas of contemporary cities known as ‘Little Italy’ or ‘Chinatown’ are one example. These areas were initially the residential quarters of migrants from particular regions, and although they retain cultural connections or associations with these groups, the function of these urban spaces has often changed: they are no longer residential areas but are instead places that draw people from across the city (and the world).32 In addition, clustering may be associated not only with permanent forms of migration but with different types of movement and transient populations. Marcuse and van Kempen’s volume explores the question of the agency behind urban partitioning in contemporary cities and stresses the role of the state in determining spatial boundaries in the city.33 The changing dynamics of clustering and spatial function in the city, and the agency that drives this, can help to shape our expectations of the relationship between language and migration at Rome. We are necessarily looking synchronically at the ancient city, and the limitations of the evidence and the difficulties of dating mean that language patterns appear artificially static and changes over time are harder to see. The Janiculum shrine, where a series of dedications to Syrian gods in both Latin and Greek were made, is 29 30 31

32 33

Stoicheva (2016: 108–10). Carson (2016: 69–72). Zadorojnyi (2010). There is, however, scope for more analysis of this question in relation to the actual evidence of graffiti from the city. Stoicheva (2016: 109). Marcuse and van Kempen (2002).

277

Olivia Elder

one example of this. The archaeological evidence shows that although there were several changes to the function of the site over the centuries, the model for the relationship between the Syrian deities and other gods there should be one of co-existence rather than replacement.34 It is difficult to see the shrine as an example of closed-off space and linguistic clustering at any stage. It is also important to think carefully about the role of the state within the cityscape: as the patchy evidence for the expulsion and controls of foreign groups shows, the state took a less active role in managing the spaces of the city and targeting foreign groups than in the modern period.35 This may result in less prevalent linguistic clustering. A comparison with contemporary cities also problematises the concept of integration, a term often used as an index for assessing migrants’ language use.36 The term is often implicitly understood in relation to the modern norm – that integration depends upon learning the official language (or languages) of the state. In the context of ancient Rome, however, it is less clear what ‘integration’ should mean and how language use was connected to this. The role of Latin does not directly correspond to the role of an official language in a modern state. Rather than assessing language use against a fixed idea of integration, we therefore need to think about what ‘integration’ meant by considering the role of different languages in the city, the expectations on migrant groups to use them, and the enforcement of these norms. Overall, we should neither stand Rome in opposition to contemporary cities nor assume that the two can be directly paralleled. In both ancient and contemporary cities, urban multilingualism needs to be contextualised in relation to the spaces of the city, state policies (however conceived), population movements and language attitudes. Contemporary cities should also make us question our expectations of linguistic diversity as a marker of immigrant populations. High levels of migration do not necessarily lead to a flood of different languages, but can result in an absence of 34 35

36

Gee (2008). On the (lack of) control over those entering and leaving the city, see Moatti (2000) and (2004). On problems with the concept of ‘integration’, see Tacoma (2016a: 204–40).

278

Population, Migration and Language in Rome

different languages in the kinds of evidence that we are able to access from Rome. Languages that do exist are not straightforwardly migrant languages. Evidence of the relationship between language and space in modern cities demonstrates the need to consider the place of languages beyond their connection to migration and to think about how language choice plays into other social, cultural and economic factors, including the status and visibility of migrants, within the city. Rome’s Populations Having presented one view of ancient migration as given by Juvenal’s Umbricius and an alternative picture provided by the evidence from modern cities, I now turn to look in detail at the evidence from Rome. A brief survey of the demographic evidence that we have for Rome and its migration patterns will help us to understand what is going on behind the distortions of the literary and epigraphic evidence, will highlight some of the ways in which Rome’s population is distinct from that of contemporary cities, and will explain some of the patterns of language use that we see in the city. Even the most basic questions about the overall size of Rome’s population and how it breaks down across different groups and statuses are near impossible to answer reliably. Estimates for Rome’s population at its peak in the late Republic and early Empire vary wildly from 440,000 to 1 million.37 The figure of 1 million is increasingly the consensus, but as Wallace-Hadrill points out, it is a rough estimate that is perhaps more a symbolic reflection of the vastness of the city than an accurate total for the numbers of inhabitants.38 The only concrete population figures in ancient sources are those for the numbers of people in receipt of the grain dole. There is, however, debate about which groups these figures include and exclude, and it is therefore difficult to extrapolate total population numbers from them.39 Purcell rightly 37

38 39

For some parameters of this debate, see Purcell (1996: 648 ff.); Noy (2000: 15–29); De Ligt and Tacoma (2016a). Wallace-Hadrill (2017). Wallace-Hadrill (2017). For the figures, see Suet. Iul. 41.3; Res Gestae 15; Dio 55.10.

279

Olivia Elder

stresses that the quantitative question of Rome’s population should not obscure a qualitative analysis of how and why it matters.40 The size of Rome matters firstly because it means that the complexity of Rome’s population and society was greater than that of any other ancient city, with large numbers of different and intersecting groups, and secondly because of what the size of Rome says about migration. A city the size of Rome was reliant on migration to maintain, let alone expand, its population.41 Estimates of the proportion of the population born outside of Rome vary from Noy’s conservative 5 per cent to Tacoma’s more recent estimate of 20 per cent.42 Again, Purcell’s warning about a reliance on quantitative evidence is salient: regardless of exact figures, we can say that Rome’s population was both diverse and shifting, and this created the potential for a highly multilingual environment. Seneca gives one account of this migrant population, and its distribution across different groups and statuses, in a letter to his mother consoling her over his own exile (Helv. 6.2–3). Like Juvenal, Seneca uses the language of a river or flood to describe the large numbers of people moving to Rome (the migrants confluxerunt into the city), but his language is considerably less hostile. He describes a wide variety of reasons why people move to the city. Seneca’s migrants include people ex municipiis et coloniis suis, ex toto denique orbe terrarum; ‘from towns and colonies, from across the whole world’.43 This description encompasses people from a wide variety of different statuses, both Roman citizens like himself and peregrini, although Seneca does not clearly draw these distinctions out. Wallace-Hadrill argues that we have consistently underestimated the range and fluidity of different statuses in our assessments of Rome’s population.44 The movement between different legal statuses is not fully understood.45 Both the absolute numbers and the conceptions of different groups changed over time. 40 41

42 43 44 45

Purcell (1996: 648). See amongst others Jongman (2003); Tacoma (2016a: 142–69) gives an overview of urban migration theory and its application to ancient Rome. Noy (2000: 26); Tacoma (2016a: 73); Lo Cascio (2016). Trans. Basore (1932). Wallace-Hadrill (2017: 54). Wallace-Hadrill (2017).

280

Population, Migration and Language in Rome

Whereas debates about population movement centred around the extent and mechanisms of integration of Italians in the period after the Social War of 91–88 BC, by the time of the constitutio Antoniniana in AD 212, different factors, groups and means of distinction were at stake. The status and labelling of groups within the city altered as the citizenship, eligibility for the grain dole, patterns of army recruitment, and other benefits and obligations expanded and contracted over time. Lavan tracks these changes by examining how the vocabulary to describe different groups, and what they designated, altered over time.46 Labels did not necessarily mesh with reality: Purcell notes the fluidity between the nominally strictly bounded categories of the populus Romanus and the plebs urbana and with it the slippage between practical and symbolic participation in the city.47 Further fluidity came from the porous boundaries of the physical city itself.48 The fluidity of Rome’s population, and the changes and slippages of the discourse about it, needs to be remembered as we consider the relationship between language and migration. The focus on foreigners’ or migrants’ experiences that dominates most existing scholarship on the topic of peoples and languages at Rome is limiting in this regard, since it neglects the fact that the definition of ‘foreign-ness’ is shifting and contested. It is not necessarily helpful or possible to divide the city, and evidence of language use, across migrants and non-migrants. Instead, we need to think about different identities and the intersection between them. The limitations to Rome’s population, migration and diversity are also important in explaining language use. First, Rome’s exceptionalism in terms of migrant numbers has been questioned by Elena Isayev, who argues that a proportion of 25–30 per cent migrants within any population is not unusual.49 Seneca also gives some evidence to support this: in the letter to his mother discussed above, he moves from describing Rome to state that nulla non magnam partem peregrinae multitudinis habet; ‘no [city] does not 46 47 48 49

Lavan (2014: 25–72). Purcell (1994: 645). Horden and Purcell (2000: 111); Moatti (2013). Isayev, Chapter 3 of this volume.

281

Olivia Elder

have a large proportion of foreigners’ (Helv. 6.3). Again the imprecision of peregrini should be noted. Moreover, Woolf argues that the social dynamics of Rome led not to diversity but to homogeneity.50 Part of the reason for this homogeneity is the number of slaves amongst Rome’s migrants.51 Slavery is a category that is notably absent from literary accounts of migration such as Juvenal and Seneca, though it is present in Philo’s comment at Leg. 155: he says that the majority of Jews in Trastevere were former slaves. It is a major point of demographic difference from contemporary cities and may be a factor limiting linguistic clustering at Rome: slaves live not in distinct quarters of the city but within Roman houses throughout the city.52 Languages in Rome’s Landscape I now explore the ways that these patterns of migration played out in the language choices made in Rome’s epigraphic landscape and consider what language patterns reveal about the identities that different groups constructed and projected in the city. The analysis sheds light on some of the questions left by Wallace-Hadrill in his assessment of immigration at Rome: ‘to what extent did the use of the dominant language, Latin, unify citizens, [and] to what extent was Greek a parallel language in the city? How far did non-Graeco-Roman languages, Hebrew, Syriac, Celtic, survive in parallel? How far were local cultural diversities compatible with a sense of being Roman?’53 How far are languages markers of immigration and how far markers of something else? The discussion of contemporary cities showed the importance of contextualising language choices by laying out the overall distribution of different languages across the city – their quantities, contexts, functions and visibility – before looking at how the language choices of migrant groups fit into and shape these. Looking at the epigraphic landscape in these broad terms helps us to understand the broader place and associations of particular 50 51 52 53

Woolf (2016). Wallace-Hadrill (2017); Woolf (2016). Tacoma (2013: 143). Wallace-Hadrill (2017: 69).

282

Population, Migration and Language in Rome

languages in the city, and therefore what groups are drawing upon when they make their linguistic choices. It also prevents us from focusing only on those who left traces of their separate or migrant identity via language choices.54 As acknowledged above, the quantities of inscriptions in languages other than Latin are small. The epigraphic record provides no evidence of languages flooding into the city in the way described in Juvenal’s third satire. Greek is, unsurprisingly, the most frequently occurring language other than Latin. There are around 1,500 Greek inscriptions amongst the pagan inscriptions from republican and imperial Rome, a figure that nonetheless means that Latin inscriptions outnumber those in Greek by approximately thirty to one.55 There are higher proportions of Greek in inscriptions from the Christian catacombs, which date from the third to the seventh centuries: Rutgers estimates that overall 9 per cent of the c.25,500 published Christian inscriptions are written in Greek.56 In contrast to the proportions across the rest of the city, Greek was dominant amongst the Jewish community of Rome: according to Rutgers, 79 per cent of the 595 Jewish inscriptions found in the city and dating roughly between the second and fourth centuries were in Greek.57 Languages other than Greek are even rarer:58 Palmyrene appears in a series of bilingual inscriptions from the Porta 54

55

56

57

58

The bias towards these individuals in scholarship on Rome is acknowledged by Noy (2000: 160). Moretti (1989: 5) estimates that there are 50,000 Latin inscriptions in comparison to the 1,500 Greek ones. Kaimio (1979: 172), followed by Tacoma (2016a: 218), similarly calculates that Greek texts form 3 per cent of all texts at Rome. Noy (2000: 170) by contrast suggests that Latin texts outnumber Greek ones by 50:1. Rutgers (1995: 183). There is evidence for higher proportions of Greek in the earlier catacombs: Kajanto (1980: 98) and Rutgers (1995: 183). The use of Greek in ICUR is analysed quantitatively in Felle (1997); Kajanto (1980: 99–101) also provides some discussion. In general, there is more scope for comparing language use in the Christian and pagan communities and inscriptions: the two are usually separated in scholarship and corpora. Felle 1997 offers some quantitative comparisons but limited broader discussion. Rutgers (1995: 176), who argues for a gradual shift from Greek to Latin over time, although this did not take place until after the third century. Leiwo (2002: 180) and Solin (1983: 705) also argue that Greek was the main language of the Jewish community until at least the third century. See n. 7 for bibliography giving overviews of the quantities of different languages in the city.

283

Olivia Elder

Portuensis, discussed below; Hebrew appears in some of the inscriptions from the Jewish catacombs, though this is mostly restricted to a single word or set phrase;59 Aramaic appears still more rarely in the catacombs: there is a single example of an inscription entirely in the language, exceptional both for its language and the high quality of its lettering;60 Nabatean is used, together with Latin, in a bilingual inscription dedicated by a man from Petra to his cognatus;61 another inscription may include the Thracian term for ‘village’;62 Egyptian hieroglyphics are found on the obelisks;63 some examples of Italic languages are found amongst Rome’s early inscriptions.64 Almost no trace can be found of languages from the West of the Empire other than names. In terms of their overall quantity then, we can say that migrant languages were virtually invisible in the epigraphic landscape of the city.65 This is a picture that matches well with the evidence from contemporary cities. These totals also underline the contrast between Greek and other languages in the city and the need to question how to understand the epigraphic remains within the framework of migration to the city. The dominant monolingualism of the city’s epigraphic landscape also broadly applies at the level of individual inscriptions. Inscriptions involving languages other than Latin or Greek are again not representative here: bilingual inscriptions are rare across the city as a whole.66 There are only around 150 Greek– Latin bilingual inscriptions from republican and imperial 59 60 61 62

63

64

65

66

Noy (2000: 178 and 262–4). JIWE 2.58: see Noy (2000: 178 and 283 n. 486). CIL VI 34196: see Adams (2003a: 264). CIL VI 32567: uotum quod uouimus ciues prou(inciae) Tracie, reg. Sardicens., midne Potelense, saluo coll(egio) Martis et Herculis. Adams (2003a: 283–4) argues that the use of midne in this inscription ‘looks very much like code-switching’: in a standard Latin inscription, ‘(ex) vico’ would usually have been expected in place of midne. Noy (2000: 219) is more sceptical about this as an example of Thracian. On the languages of the obelisks at Rome, see Adams (2003a: 571–2); Iversen (1965); Kayser (1994: 1 and 4). See, for example, Cornell (1995: 157 with 430 n. 23); Colonna (1987: 55–66) on Etruscan inscriptions at Rome. Thirteen texts from Rome are included in Meiser (2014). Tacoma (2016a: 215) argues more forcefully that ‘strictly speaking there were no migrant languages at Rome’. The term ‘bilingual texts’ includes a range of different types of text: see Mullen (2013: 83–5) for a typology, tabulated at p. 84. In my doctoral thesis I break down the inscriptions of Rome along these lines and consider their different implications. Felle

284

Population, Migration and Language in Rome

Rome.67 This must have some consequences for Moatti’s argument about migrant identity in the city:68 language does not seem to have been a common way to express the bricolage or accumulation of different identities. The lack of archaeological evidence for migrant quarters or ghettos is mirrored in the scattered distribution of languages. The sole major examples of clustering of different languages and/or language patterns are the Jewish catacombs and the series of Palmyrene inscriptions found outside the Porta Portuensis. In general, an analysis of Rome’s landscape that pays attention to the spatial distribution of languages reveals a greater degree of bilingualism than we get from the quantities of texts alone. Although there are relatively few examples of bilingual texts within the city, there are more examples of Greek–Latin bilingual contexts – that is, contexts where different languages appear, even if separately. Rome’s bilingual contexts include spaces such as the Capitoline and Forum where there was a series of bilingual dedications and monuments.69 These contexts show a parallelism of languages, at least in certain spheres, that contrasts with the way they are set against each other, as markers of distinct groups or identities, in evidence like Juvenal. These bilingual contexts also challenge the extent to which a distinction between public and private can explain language use at Rome. Some scholars have argued for a kind of diglossia in the city, with Latin as a public or official language and Greek as

67

68 69

(1997) and Tozzi (2012) also provide some quantitative discussion of the different types of bilingual texts at Rome, although they do not follow Mullen’s typology. Tozzi (2012) cites a figure of c.165 bilingual inscriptions on the basis of the Epigraphic Database Roma. Felle (1997) cites a figure of 146 bilingual inscriptions on the basis of IGUR. His quantitative analysis of how these numbers vary over time (p. 671) is also useful. Totals vary depending on how broadly ‘bilingualism’ is defined. The issue of script and how the relationship between script and language is construed provides further distortion. On the bilingual inscriptions of Rome, see also Adams (2003a: 356–82). Moatti (2014). Some examples of the Greek and bilingual inscriptions in this area include: ILLRP 174–81, the series of bilingual dedications by the people and kings of Asia Minor; the senatus consultum de Asclepiade (RDGE 22), a bilingual text originally displayed on the Capitoline; IGUR 78–93, the series of Greek inscriptions of stationes from the Via Sacra in front of the temple of Romulus. On the inscriptions of the stationes and collegia, see Kaimio (1979: 64–7).

285

Olivia Elder

a private or personal one.70 This distinction partly holds: Latin was dominant in public and official contexts, and there is a higher proportion of Greek in the catacombs, which represent one of the few more straightforwardly ‘private’ epigraphic spaces in the city. The presence of Greek within spaces such as the Forum and the Capitoline, however, shows that the situation is more complex than this. Furthermore, recent analyses have broken the categories of ‘public’ and ‘private’ down to show that they are not absolutes. In her recent book, Russell demonstrates the instability of the categories of public and private space in the republican city and shows how concepts of public and private interrelate with other categories of space – foreign, sacred, political.71 Of particular interest for our purposes is Russell’s argument that the category of ‘Greek’ is a way to move between public and private, and between different definitions of public: public in the sense of ‘open to all’ and public in the more formal sense of relating to the institution of the populus Romanus. She argues that ‘Greek’ could be public in the first sense but not the latter sense because, unlike the populus Romanus, it was inherently non-Roman.72 Russell’s arguments are illustrated through examples of Greek visual language but, I argue, can be extended to the display of written Greek. The use of Greek in the bilingual Capitoline dedications to Jupiter and the Roman people from the peoples and kings of Asia Minor shows that the role of Greek in the ‘public’ sphere could be even broader than Russell’s analysis suggests.73 On the one hand, Greek here drew upon and symbolised distinct non-Roman associations: it was directly connected with Greekspeaking city-states and peoples. It also, however, fitted within and contributed to a thoroughly Roman space policed by Roman authorities. Here, Greek worked to construct a monument that is ‘public’ in both Russell’s broad and narrow definitions of the term: in its use in the dedications to Jupiter and the Roman people, it is 70

71 72 73

Adams (2003a: 349–50 and 375) characterises this difference as being between Latin as the official bureaucratic language and Greek as a private language. Russell (2016). Russell (2016: 189). On the Capitoline inscriptions and monument (ILLRP 174–81), see Del Monaco (2013b); Lintott (1978); Mellor (1978). Adams (2003a: 377–8 and 659) discusses some of the linguistic features of these inscriptions.

286

Population, Migration and Language in Rome

used to construct a monument with the populus Romanus at its centre. The choices on this monument are not directly parallel to those of individual migrants and groups. In exemplifying how Greek operated across different spectra of space within the city – public and private, Roman and non-Roman – they do, however, provide a wider explanatory context for these individual language choices and their associations. The fact that Greek could bridge these categories was part of what made it an effective choice within the city, as I shall show in the discussion of the Palmyrene texts and the Greek epitaphs below. Viewed in broad perspective, Rome’s epigraphic landscape appears both monolingual and multilingual. In quantitative terms, Latin is dominant and there are few examples of bilingual texts. On the other hand, there are several bilingual or multilingual contexts within the city, and the existence of a range of different languages in inscriptions shows that the use of those languages was at least a possible choice for their speakers. Greek is to some extent a parallel language to Latin, appearing in public, official spaces, whilst retaining its own distinct set of associations. There are few examples of closed spaces with distinctive language patterns. Looking at the language choices of migrants and migrant groups against this backdrop offers insights into the types of identities that they sought to display and communicate at Rome, and the ways in which they contribute to both the presence and the absence of different languages there. Epigraphic Language and Migrant Identity Sets of evidence with distinct language patterns show the difficulty of drawing a direct link between migration and language use. Consider, for example, the Jewish catacombs, in which around three quarters of epitaphs are in Greek, in contrast with 3 per cent of epitaphs across the city as a whole. A number employ Hebrew formulae, the only appearance of Hebrew in the city’s epigraphic record.74 Migrant identity is not the only, or even primary, 74

These proportions are those of Cappelletti (2006). On language in the Jewish inscriptions, see also Leon (1995: 75–92); Noy (1997).

287

Olivia Elder

category at stake here: the combined ethnic and religious identity, unique amongst groups in the city, is significant.75 The series of inscriptions connected with Palmyrenes found outside the Porta Portuensis, often used as the textbook examples for the relationship between language and expression of (migrant) identity, also themselves show the complexities of this relationship within the city.76 On the one hand, these inscriptions demonstrate that language could function as a marker of immigrant groups, their cultures and practices. In IGUR 119, a Greek–Palmyrene dedication to the Moon Aglibol and the Sun Malakhbel, the dedicator is named in Greek as Ἰ(ούλιος) Αὐρ(ήλιος) Ἡλιόδωρος Ἀντιόχου Ἁδριανός but in Palmyrene as ‘Iarhai son of Haliphi, son of Iarhai, son of Lisams’ – that is, he deploys different names and naming formulae in each language.77 IGUR 117 is a Latin–Greek bilingual inscription commemorating the setting up of two temples to Bel and Malakhbel by the Palmyrenes Heliodorus and C. Licinius. Here, different content is communicated in each of the two languages: the Latin version opens with an acclamation to the emperor (pro salute imp(eratoris) [Caesaris . . . .]) which is absent from the Greek.78 Via language, dedicators can both demonstrate their accommodation to Roman practices and display their local identities. It is, however, crucial to stress the exceptional nature of these Palmyrene inscriptions, both in the directness of the relationship between language and identity and in their use of a provincial language.79 Moreover, even within this single set of inscriptions, connected to one group of migrants and displayed in a single context at Rome, language choices and what they were meant to represent are not consistent, and this has consequences for our understanding of language patterns in the city more broadly. It is 75 76

77 78 79

Tacoma (2013: 142–3). On this series of inscriptions, see Noy (2000: 242–4); Adams (2003a: 248–53); Moatti (2014: 149–50); Tacoma (2016a: 222–3). For a full catalogue of the inscriptions, see Chausson (1995) and, for an overview of Palmyrenes at Rome, see Equini-Schneider (1988). On this inscription, see Adams (2003a: 251–2). On this inscription, see Adams (2003a: 248–9). Millar (1993: 328); Adams (2003b: 199) on the Palmyrenes as exceptional amongst groups in auxiliary forces of Roman army in taking language and art with them.

288

Population, Migration and Language in Rome

not easy to unpick the relationship between languages and the expression of Roman and/or provincial identities, and different scholars have made different assessments of this relationship and its expression. Adams sees the Palmyrene, Greek and Latin in these inscriptions as showing progressive degrees of (non-)integration and argues that the choice of which to use depended on how strongly the dedicator wanted to stress this degree of integration.80 The ‘most aggressively “Palmyrene” texts’ (i.e. ‘those in which both gods and dedicators retain Palmyrene names’) use Palmyrene with a Greek but not Latin version. Tacoma admits that he himself is too schematic in assigning one language to the host community and one to the migrant community in each of the bilingual inscriptions,81 but doing so usefully highlights the ambiguity of Greek, and shows that it could function both as a migrant language and as a language of Rome. This ambiguity fits with our assessment of the place of Greek across the spaces of the city as a whole. The use of Greek in epitaphs provides a useful set of evidence in which to test this out further and explore the different identities that Greek could be used to express amongst other groups in the city. Greek in the Epitaphs of Rome It has been argued that there is a direct link between the use of Greek and non-Roman identity in the Greek epitaphs from the city. Adams regards the combination of Greek and language and script as ‘indicating non-integration’.82 Like Adams, Noy traces a further link between Greek and immigrant status in arguing that use of the language tended to die out across generations; he regards bilingualism as a temporary phase in immigrant families.83 Although this connection between the Greek language and immigration fits with the picture constructed in Juvenal, it does not fit with the place of Greek across the city as a whole: other than in the Jewish catacombs, Greek is not linked epigraphically to 80 81 82 83

Adams (2003a: 252–3). Tacoma (2016: 222–3). Adams (2003a: 365). Noy (2000: 176).

289

Olivia Elder

closed in-groups spaces or communities and functions across the spectrum of ‘public’ and ‘private’ within the city. The range of different identities and origins that Greek could be used to express is illustrated by two verse epitaphs written entirely in Greek language and script, those of Olympia (IGUR 1287) and Attikilla (IG XIV 1440). Olympia’s verse epitaph emphasises her Greek origins and connections to the Greek East, drawing out both her fatherland of Apamea and her Greek γένος in an explicit display of ethnic origin and identity: Ἔλλην μὲν τὸ γένος, πατρὶς δέ μοι ἦτον Ἀπάμεα (‘I (was) Greek by birth, my fatherland (was) Apamea’).84 Her connections to the city of Rome itself are, however, not mentioned: in this case, the choice to use Greek could be read as an attempt to set herself apart as non-Roman. Atikilla’s epitaph, found at Villa Spada just outside of Rome, is, however, a direct contrast: this epitaph gives her πατρίς, in Greek, as Rome: πατρὶς μὲν ζαθέη Ῥώμη. These two verse inscriptions show that Greek could be used to indicate immigrant status but could also be used in the epitaphs of those native to the city: they therefore show the limitations of migration and origin in making sense of Greek language choices in the city. The question then is what else Greek was being used to symbolise. Kajanto’s study of the Greek epitaphs offers some indicative evidence. He argues that people commemorated in Greek epitaphs are generally of a higher social status than those in Latin ones.85 Although only 3 per cent of all epitaphs are in Greek, the proportion of verse epitaphs that are in Greek is considerably higher and there are in addition a number of bilingual inscriptions in which the main Latin text is followed by a Greek epigram.86 People were here drawing on the associations between Greek and art and culture, associations that are replicated elsewhere in the city in the use of Greek on busts and in the signing of artworks.87 There are also some connections between Greek and particular groups. Relatively few epitaphs include the profession 84

85 86 87

The use of ἦτον rather than ἦν for the third person imperfect of the verb ‘to be’ here is just one of several oddities in the Greek of this inscription. Kajanto (1963: 6). Kajanto (1963: 6). For Greek inscriptions associated with art and culture at Rome, see IGUR 1491–651.

290

Population, Migration and Language in Rome

of the deceased, but the largest single group of known professions in the Greek epitaphs is doctors.88 At Rome, connections between Greek and medicine are also visible in the Greek dedications to Asclepius in the shrines around the city.89 These examples indicate that the use of Greek was connected not only to origins but to other aspects of an individual’s status or identity and tie in with the role of Greek across the city as a whole. Latin as a Migrant Language? The other side of the coin for understanding the relationship between language and migration is the role of Latin and its use by migrant groups. The dominance of Latin in inscriptions is amplified by its dominance amongst some of the most mobile groups – freedmen and soldiers – who are themselves dominant in the city’s epigraphic record. All but two epitaphs of soldiers at Rome are in Latin.90 Kajanto notes that the proportions of freedmen and slaves are higher in the Latin epitaphs than in the Greek epitaphs, despite the fact that many of these people must have been Greek-speaking immigrants to the city.91 There are obvious practical reasons for this choice, particularly amongst soldiers, but the predominance of Latin is also revealing of what migrants sought to achieve by the use of particular languages. Both soldiers and freedmen wanted to stress their connection to Roman structures of power and status: their own local identities were less relevant. Here, the distinction that Wallace-Hadrill draws between those migrating to the city and those migrating to the citizenship is relevant.92 Freedmen were especially anxious to stress their newly acquired citizenship, and Latin was the appropriate language in which to do so. The link between Latin and citizenship is 88 89

90 91

92

Kajanto (1963: 6). On the cult of Asclepius at Rome, see Renberg (2006–7). Juvenal also lists medicus as one of the professions practised by the Greek migrants at Sat. 3.77. Kajanto (1980: 91), Noy (2000: 171). Kajanto (1963: 3). He notes that there could have been more slaves and freedmen in the Greek epitaphs than is explicitly noted, but ‘the number of such cases is not likely to be large enough to alter the statistics materially’; the numbers could also presumably be similarly distorted in the Latin epitaphs. Wallace-Hadrill (2017).

291

Olivia Elder

visible in other evidence too: in Suetonius’s testimony, Claudius strips a man of his citizenship because he is ignorant of Latin (Suet. Claud. 16.2), and in Seneca, Suetonius and Aulus Gellius, the acceptance of a word into the Latin language is described via the metaphor of giving it citizenship.93 Display of citizen status via the use of Latin was, however, less relevant for those with alternative claims to status within the city such as the individuals and groups discussed above: they were therefore willing to use other languages. Overall, the patterns of languages in the epigraphic evidence reflect the patterns and breadth of migration to the city, and the variety of status and identities within the population. They also show the limitations of migrant identity in the city and its expression through language. In light of this, I turn back to the different relationship between migrants and their languages in the literary and epigraphic records that was drawn out at the beginning of this chapter. The distinction between these two sets of evidence can be resolved more than initially assumed. Literary sources can help to make sense of the epigraphic evidence: in them, we see some of the complexities of the language choices that migrants were making framed and debated in their contemporary context. A return to Juvenal’s third satire shows that it fits with the language situation that we see in the epigraphic evidence as well as the traditional expectations and understandings of migrants and their languages for which it has been used as evidence. Umbricius’ complaint about the presence of foreigners can be argued to derive as much from the integration and invisibility of foreign groups as from their visibility.94 One of his complaints is that foreigners ‘will become an integral part of Rome’s great houses’ (uiscera magnarum domuum dominique futuri) – that is, gain an integral place at Rome at every social level. He resents the fact that they practice a range of professions in the city (grammaticus, rhetor, geometres, pictor, aliptes, augur, schoenobates, medicus, magus; ‘grammarian, rhetor, mathematician, painter, masseur, tightropewalker, doctor, magician’, lines 75–7) and have gained status and 93

94

Suet. Gram. 22; Sen. Ep. 120.4; Gel. 19.13.3. See also Adams (2003b: 185–8) on the connection between Latin and citizenship, and Latin and the army. Woolf (2016) also interprets this passage along these lines.

292

Population, Migration and Language in Rome

opportunities denied to native Romans (line 119). The Greek language has been absorbed into the city with them, as is shown by the number of Greek loan-words that Umbricius uses to describe these predominately Greek professions.95 The passage can therefore be used to argue against the idea of the use of Greek as a sign of non-integration and a migrant language. Although connected to migrant populations, it also gained its own place within the city, and this is precisely the source of Umbricius’ problem with it. Further evidence of the debate about the place of Greek within the city is provided by a passage of Athenaeus (Deipnosophists 3.121.b–f), depicting a dinner party held at Rome in the late second or early third century AD and attended by Greek philosophers, some of whom were resident in the city. The passage is also striking in revealing the existence of a debate about the appropriateness of languages amongst migrant communities themselves.96 ἐπὶ τούτοις λεχθεῖσιν ὁ Κύνουλκος πιεῖν ᾔτησε δηκόκταν, δεῖν λέγων ἁλμυροὺς λόγους γλυκέσιν ἀποκλύζεσθαι νάμασι. πρὸς ὃν ὁ Οὐλπιανὸς σχετλιάσας καὶ τύψας τῇ χειρὶ τὸ προσκεφάλαιον ἔφη· μέχρι πότε βαρβαρίζοντες οὐ παύσεσθε; ἢ ἕως ἂν καταλιπὼν τὸ συμπόσιον οἴχωμαι, πέττειν ὑμῶν τοὺς λόγους οὐ δυνάμενος; καὶ ὅς· ἐν Ῥώμῃ τῇ βασιλευούσῃ διατρίβων τὰ νῦν, ὦ λῷστε, ἐπιχωρίῳ κέχρημαι κατὰ τὴν συνήθειαν φωνῇ. In reply to these comments, Cynulcus asked to drink some decocta, saying that he needed to wash salty words away with sweet streams. Ulpian grew irritated with him, pounded his pillow with his fist, and said: How long are you going to continue using foreign words? Until I leave the symposium and go home, unable to stomach your words? And Cynulcus said: Since I am currently spending my time in the imperial city of Rome, my very good sir, I have grown accustomed to using the local language.97

To Ulpian’s disgust, Cynulcus requests cold water using the (transcribed) Latin term δηκόκταν. The passage shows the parameters of the debate on appropriate language use at Rome. Ulpian is indignant (σκετλιάσας) at the use of the Latin term and, in 95 96

97

On the choice of Greek words here, see Braund (1996: 187). On this passage and its implications for the Greek spoken at Rome, see Purcell (2000: 440). Translation from Olson (2007–12, Volume 2: 79).

293

Olivia Elder

accusing Cynulcus of ‘using barbarisms’ (βαρβαρίζοντες), draws attention to the unacceptable ‘foreign-ness’ of the term, even in its Greek inflected form and used in the city of Rome. Cynulcus, however, defends himself by arguing that whilst living at Rome he has become accustomed to using the local language (Latin), linguistic accommodation that we have seen amongst the majority of migrants in the epigraphic record. This is obviously not, however, a defence that washes with all Greeks: as Ulpian’s outrage shows, the use of Greek at Rome was expected by some, and the nature of this Greek mattered. Both Ulpian’s and Cynulcus’ attitudes and practices may be seen in the epigraphic landscape of the city, where we see both adherence to Greek and a willingness to use Latin. In revealing the kinds of debates that may underpin epigraphic language choices, the passage helps to explain both the presence and absence of different languages in the city. Again, the different status of Greek in relation to other languages is revealed: there is no extant literary evidence debating the appropriateness of the native languages of migrant groups. The different accounts that we see of language use and quantities at Rome are reflective of a genuine tension between a multilingual and a monolingual Rome that exists both within and between different types of evidence, literary, epigraphic and demographic. Migration helped to shape both these different Romes. The presence of people from across the Empire meant that the city had the potential to be a highly multilingual environment. Literary sources such as Juvenal pick up on this, framing language as a marker of migrant groups. Our lack of ephemeral writings and evidence of spoken languages do limit the amount of multilingualism amongst Rome’s migrants that we can access. There is, however, evidence, both from Rome itself and the comparative evidence of modern cities, to suggest that the use of language as a visible marker of identity among migrant groups is limited. Migration is, ironically, better at explaining the absence than the presence of different languages in displays of identity. Many of the most mobile groups use Latin. This monolingualism is both practical and symbolic and, whilst replicated in other cities with large numbers of migrants, is also a product of Rome’s particular dynamics of migration. Comparisons are helpful in 294

Population, Migration and Language in Rome

revealing these: Rome was not a city like modern London, but a city like modern London provides one framework to better understand Rome. At Rome, migration into the citizenship existed alongside migration to the city itself, and this encouraged the use of Latin. Linguistic homogeneity was further increased by the numbers of slaves and freedmen amongst Rome’s migrants. This chapter has also indicated the need to look beyond migration to understand the place and significance of multilingualism in the city. The use of languages other than Latin was rarely linked solely to foreign or migrant identity: apart from anything else, our analysis has shown that these were not categories that were always relevant or clear-cut at Rome. Language choices instead reveal the different and fluid statuses within the city’s population, and the fact that, within this complex web, the category of migrant is not always in view.

295

REFERENCES

Adamesteanu, Dinu. (1951). Archaeological News: Hungary. American Journal of Archaeology 55, 380–1. Adams, J. N. (2002). Bilingualism at Delos. In Adams, Janse & Swain (eds.), 103–27. (2003a). Bilingualism and the Latin Language. Cambridge: Cambridge University Press. (2003b). ‘Romanitas’ and the Latin Language. Classical Quarterly 53, 184–205. (2007). The Regional Diversification of Latin 200 BC – AD 600. Cambridge: Cambridge University Press. (2013). Social Variation and the Latin Language. Cambridge: Cambridge University Press. Adams, J. N., Mark Janse & Simon Swain (eds.). (2002). Bilingualism in Ancient Society: Language Contact and the Written Text. Oxford: Oxford University Press. Adiego Lajara, Ignacio-Javier. (1992). Protosabelio, osco-umbro, sudpiceno. Barcelona: Promociones Publicaciones Universitarias de Barcelona. (2007). The Carian Language. Leiden: Brill. (2015). Some remarks on the new Opic (‘Pre-Samnite’) inscription of Niumsis Tanunis. Incontri Linguistici 38, 15–28. Agostiniani, Luciano. (1981). Duenom duenas: καλος καλō: mlaχ mlakas. Studi Etruschi 49, 95–111. (1982). Le ‘iscrizioni parlanti’ dell’Italia antica. Florence: L. S. Olschki. (1983). Aspirate etrusche e gorgia toscana: valenza delle condizioni fonologiche etrusche. In L. Agostiniani & L. Giannelli (eds.), Fonologia etrusca, fonetica toscana. Il problema del sostrato, 25–59. Florence: L. S. Olschki. (1984–5). Epigrafia e linguistica anelleniche di Sicilia: bilancio di un quadriennio. Kokalos 30–1, 193–222. (2011a). Feluskeś o Θeluskeś sulla Stele di Vetulonia? In D. F. Maras (ed.), Corollari. Scritti di antichità etrusche e italiche in omaggio all’opera di Giovanni Colonna, 177–84. Pisa-Rome: Fabrizio Serra. (2011b). Pertinentivo. In R. Rocca (ed.), Le lingue dell’Italia antica. Iscrizioni, testi, grammatica. Proceedings of the International Congress in memoriam Helmut Rix (= Alessandria, Rivista di Glottologia 5) 17–44. Alessandria: dell’Orso.

296

References Ahlberg-Cornell, G. (1984). Herakles and the Sea-Monster in Attic Black-Figure Vase-Painting. Stockholm: Svenska Institutet i Athen and P. Aström. Allen, Joel. (2006). Hostages and Hostage-Taking in the Roman Empire. Cambridge: Cambridge University Press. Allen, W. Sidney. (1978). Vox Latina. A Guide to the Pronunciation of Classical Latin. Second edition. Cambridge: Cambridge University Press. (1987). Vox Graeca. A Guide to the Pronunciation of Classical Greek. Third edition. Cambridge: Cambridge University Press. Alonso Déniz, Alcorac & Enrique Nieto Izquierdo. (2009). Dialecto local y dialecto épico en las inscripciones métricas de la Argólide. Minerva 22, 83–105. Alter, Stephen G. (2005). William Dwight Whitney and the Science of Language. Baltimore and London: Johns Hopkins University Press. Amiotti, Gabriella. (1980). I Greci ed il massacro degli Italici nell’ 88 a.C. Aevum 54, 132–9. Anthony, David W. & Dorcas R. Brown. (2017). Molecular Archaeology and Indo-European Linguistics: Impressions from New Data. In Bjarne Simmelkjær Sandgaard Hansen et al. (eds.), Usque ad Radices: IndoEuropean Studies in Honour of Birgit Anette Olsen, 25–54. Copenhagen: Museum Tusculanum. Antonini, Rosalba. (1990). Gli alfabetari oschi. In Pandolfini & Prosdocimi (eds.), 143–53. Archibald, Zosia. (2011). Mobility and Innovation in Hellenistic Economies: The Causes and Consequences of Human Traffic. In Zosia Archibald, John K. Davies & Vincent Gabrielsen (eds.), The Economies of Hellenistic Societies, Third to First Centuries BC, 42–65. Oxford: Oxford University Press. Arnott, Geoffrey W. (1996). Alexis: The Fragments. Cambridge: Cambridge University Press. (2004). Alexis, Greek New Comedy and Plautus’ Poenulus. In Baier (ed.), 61–92. Auda, Yves & Marie-Françoise Boussac. (1996). Étude statistique d’un depôt d’archives à Délos. In Marie-Françoise Boussac & Antonio Invernizzi (eds.), Archives et sceaux du monde hellénistique (Suppléments au Bulletin de Correspondance Hellénique, 29), 511–22. Athens: École française d’Athènes. Austin, N. J. E. & Boris Rankov. (1995). Exploratio. Military and Political Intelligence in the Roman World from the Second Punic War to the Battle of Adrianople. New York: Routledge. Bagnasco Gianni, Giovanna. (1996). Oggetti iscritti di epoca orientalizzante in Etruria (Biblioteca di Studi Etruschi 30). Florence: L. S. Olschki. (2010). Fenomeni di contatto nelle più antiche iscrizioni etrusche: spunti tarquiniesi. Annali della Fondazione per il Museo ‘Claudio Faina’ 17, 113–32.

297

References Baier, Thomas (ed.). (2004). Studien zu Plautus’ Poenulus. Tübingen: Gunter Narr. Baillet, Jules. (1920–6). Inscriptions grecques et latines des tombeaux des rois ou Syringes à Thèbes. Cairo: Institut français d’archéologie orientale du Caire. Bakkum, Gabriël C. L. M. (2009). The Latin Dialect of the Ager Faliscus: 150 Years of Scholarship. Two volumes. Amsterdam: University of Amsterdam Press. Baratta, Giulia. (2012). I soldati interpreti nell’esercito romano. In C. Wolff (ed.), Le métier de soldat dans le monde romain. Actes du cinquième congrès de Lyon organisé les 23–25 septembre 2010 par l’Université Jean Moulin Lyon 3, 479–95. Lyon: de Boccard. Barkóczi, László. (1944–5). Ein dakischer Dolmetscher in Brigetio. Archaeologiai Értesítő 5–6, 178–92. Barkóczi, László & András Mócsy. (1972–). Die römischen Inschriften Ungarns. Budapest and Bonn: Akadémiai Kiadó and Habelt. Barsby, John. (1986). Plautus, Bacchides: Translation and Commentary. Warminster: Aris and Phillips. Baslez, Marie-Françoise. (1987). Le rôle et la place des Phéniciens dans la vie économique des ports de l’Égée. In Edward Lipinski (ed.), Phoenicia and the East Mediterranean in the First Millennium B.C., 267–85. Leuven: Peeters. (2002). Mobilité et ouverture de la communauté ‘romaine’ de Délos: amitiés, mariages mixtes, adoptions. In Müller & Hasenohr (eds.), 55–65. Basore, J. W. (1932). Seneca: Moral Essays, Volume II. Cambridge, MA: Harvard University Press. Bechtel, F. (1917). Die historischen Personennamen des Griechischen bis zur Kaiserzeit. Halle: Max Niemeyer. Bell, H. I. (1924). Notes on Early Ptolemaic Papyri. Archiv für Papyrusforschung 7, 17–27. Bellelli, Vincenzo. (2008). «Setums mi fece». Elementi per un riesame del cratere iscritto del Ferrone. In Paolo Santoro (ed.), Una nuova iscrizione da Magliano Sabina. Scrittura e cultura nella valle del Tevere, 59–69. Rome: Fabrizio Serra. Benelli, E. (2011) ‘Vornamengentilizia’. Anatomia di una chimera. In D. F. Maras (ed.), Corollari. Scritti di antichità etrusche e italiche in omaggio all’opera di Giovanni Colonna, 193–8. Pisa and Rome: Fabrizio Serra. Beness, J. Lea & Tom Hillard. (2003). The first Romans at Philae (CIL 12.2.2937a). Zeitschrift für Papyrologie und Epigraphik 144, 203–7. Benz, Lore & Eckard Lefèvre (eds.). (1998). Maccus barbarus: sechs Kapitel zur Originalität der Captivi des Plautus. Tübingen: Gunter Narr Verlag. Bérard, François et al. (2010). Guide de l’épigraphiste. Bibliographie choisie des épigraphies antiques et médiévales. Fourth edition. Paris: Éditions Rue d’Ulm and Presses de l’École Normale Supérieure.

298

References Bernand, A. (1969). Les inscriptions grecques de Philae. Tome I: Époque Ptolémaïque. Paris: Centre national de la recherche scientifique. (1989). De Thèbes à Syène. Paris: Centre national de la recherche scientifique. Bernand, É. (1999). Inscriptions grecques d’Hermoupolis Magna et de sa nécropole. Paris: Institut français d’archéologie orientale. Berrendonner, Claire. (2009). L’Invention des épitaphes dans la Rome médiorépublicaine. In Marie-Laurence Haack (ed.), Écritures, cultures, sociétés dans les nécropoles d’Italie ancienne. Table ronde des 14–15 décembre 2007 ‘Mouvements et trajectoires dans les nécropoles d’Italie d’époque prérépublicaine et républicaine’, 181–201. Bordeaux and Paris: Ausonius. Bickerman, Elias J. (1952). Origines Gentium. Classical Philology 47, 65–81. Biles, Zachary & Jed Thorn. (2014). Rethinking Choregic Iconography in Apulia. In Csapo, Goette, Green & Wilson (eds.), 295–318. Bingen, J. (1979). Les inscriptions de Philae des IIIe et IIe siècles avant notre ère. Chronique d’Égypte 54, 304–9. (2002). Égypte et Nubie. Revue des Études Grecques. 115, 744–52. (2004). Égypte et Nubie. Revue des Études Grecques. 117, 686–91. (2008). Inscriptions pariétales et prosopographie à Philae au Ier siècle a.C. Chronique d’Égypte 83, 245–57. Bonfante, Giuliano & Larissa Bonfante. (2002). The Etruscan Language: An Introduction. Second edition. Manchester: Manchester University Press. Bonghi Jovino, Maria. (2000). L’espansione degli Etruschi in Campania. In Mario Torello (ed.), Gli Etruschi. Catalogo della mostra, 157–67. Milan: Bompiani. Boter, G. J. (2011). The Accentuation of Greek Forms of Latin Names Containing Non-syllabic –u–. Zeitschrift für Papyrologie und Epigraphik 177, 254–8. Bourdin, Stéphane & Sandrine Crouzet. (2009). Des Italiens à Carthage? Réflexions à propos de quelques inscriptions du tophet de Carthage. In Poccetti (ed.), 443–94. Boussac, Marie-Françoise. (1982). À propos de quelques sceaux déliens. Bulletin de Correspondance Hellénique 106, 427–46. (1988). Sceaux déliens. Revue archéologique 1, 307–40. (1992). Les sceaux de Délos 1. Sceaux publics, Apollon, Hélios, Artémis, Hécate. Athens and Paris: École française d’Athènes and de Boccard. (1993). Archives personnelles à Délos. Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres 137, 677–93. Bowman, Alan & Andrew Wilson (eds.). (2011). Settlement, Urbanization, and Population (Oxford Studies on the Roman Economy 2). Oxford and New York: Oxford University Press. Bowman, A. K., R. S. O. Tomlin & K. A. Worp. (2009). Emptio Bovis Frisica: the ‘Frisian Ox Sale’ Reconsidered. Journal of Roman Studies 99, 156–70. Boyancé, P. (1956). La connaissance du grec à Rome. Revue des Études Latines 34, 111–31.

299

References Bradley, Guy. (2000). Ancient Umbria. State, Culture and Identity in Central Italy from the Iron Age to the Augustan Era. Oxford: Oxford University Press. Braund, S. M. (1996). Juvenal: Satires. Book I. Cambridge: Cambridge University Press. (2004). Juvenal and Persius. Cambridge, MA: Harvard University Press. Breeze, David J., Sonja Jilek, Andreas Thiel, & Zsolt Visy. (2008). Római Birodalom Határai = Frontiers of the Roman Empire / Római limes Magyarországon = The Roman limes in Hungary. Pécs: Pécsi Tudományegyetem Régészeti Szeminárium. Bresciani, E. et al. (1972). Ostraka demotici da Ossirinco. Studi Classici e Orientali 21, 321–87. (1973). Ostraka demotici da Ossirinco. Studi Classici e Orientali 22, 208–73. Bresson, A. (2007). L’entrée dans les ports en Grèce ancienne: le cadre juridique. In Claudia Moatti & Wolfgang Kaiser (eds.), Gens de passage en Méditerranée de l’antiquité à l’époque moderne, 37–78. Paris: Maisonneuve et Larose. Briquel, D. (1990). Le témoignage de Claude sur Mastarna-Servius Tullius. Revue belge de philologie et d’histoire = Belgisch tijdschrift voor filologie en geschiedenis 68, 86–108. (2016). Catalogue des inscriptions étrusques et italiques du Musée du Louvre, Paris: A. et J. Picard. Briquel-Chatonnet, F. (1996). Mosaïque de langues, mosaïque culturelle: le bilinguisme dans le proche-orient ancien. Paris: Maisonneuve. Broodbank, Cyprian. (2013). The Making of the Middle Sea. A History of the Mediterranean from the Beginning to the Emergence of the Classical World. New York: Thames & Hudson. Bruneau, Philippe. (1970). Recherches sur les cultes de Délos à l’époque hellénistique et à l’époque impériale. Paris: de Boccard. Bruni, S. (2013). Attorno a Praxias. Annali della Fondazione per il Museo ‘Claudio Faina’ 20, 257–337. Bubeník, Vit. (1989). Hellenistic and Roman Greece as a Sociolinguistic Area. Amsterdam and Philadelphia: Benjamins. Buck, Carl Darling. (1913). The Interstate Use of the Greek Dialects. Classical Philology 8, 133–59. (1928). A Grammar of Oscan and Umbrian. With a Collection of Inscriptions and a Glossary. Second edition. Boston: Ginn and Company. (1933). Comparative Grammar of Greek and Latin. Chicago: University of Chicago Press. (1955). The Greek Dialects. Grammar, Selected Inscriptions, Glossary. Chicago: University of Chicago Press. Bundgård, Jens A. (1965). Why Did the Art of Writing Spread to the West? Reflexions on the Alphabet of Marsiliana. Analecta Romana Instituti Danici 3, 11–72.

300

References Buonfiglio Costanzo, Gioseppe. (1613). Dell’historia siciliana. Part 3. Messina: Pietro Brea. (1739). Dell’historia siciliana. Parts 1 and 2. Second edition. Messina: D. Michele de Chiaramonti. Buonocore, Marco & Paolo Poccetti. (2013). Una nuova iscrizione peligna del gruppo ‘AN(A)C(E)TA’. Epigraphica 75, 59–106. Buranelli, F. (a cura di). (1987). La tomba François di Vulci. Cat. della mostra (Città del Vaticano). Roma. Buranelli, F., & M. Sannibale. (2005). Non più solo Larthia. Un documento epigrafico inedito dalla Tomba Regolini-Alassi di Caere. In B. Adembri (ed.), ΑΕΙΜΝΗΣΤΟΣ. Miscellanea in onore di M. Cristofani, 220–31. Florence: Centro Di. Callaway, Joseph Sevier. (1950). Sybaris. Baltimore: Johns Hopkins Press. Campanile, Enrico. (1961). Elementi dialettali nella fonetica e nella morfologia del latino. Studi e saggi linguistici 1, 1–21. (1981). Prolegomeni ad un’analisi della variazione linguistica nei dialetti italici e nel gallico. Annali dell’Istituto Universitario Orientale di Napoli, sezione linguistica 3, 37–46. (1993). Note sulle compagnie di ventura osche. Athenaeum 81, 601–11. Candeloro, A., & G. Colonna. (2011). Volsinii. Orvieto, Crocifisso del Tufo. Studi Etruschi 74, 278–92. Capogrossi Colognesi, L. (1990). Dalla tribù allo Stato. Le istituzioni dello stato cittadino. Rome: La Sapienza. (1994). Proprietà e signoria in Roma antica I. Second edition. Rome: La Sapienza. Cappelletti, S. (2006). The Jewish Community of Rome: From the Second Century BC to the Third Century C.E. Leiden and Boston: Brill. Capponi, L. (2005). Augustan Egypt: The Creation of a Roman Province. London and New York: Routledge. Carpenter, T. H., K. M. Lynch & E. G. D. Robinson. (2014). Introduction. In Carpenter, Lynch & Robinson (eds.), 1–10. (eds.). (2014). The Italic People of Ancient Apulia, Cambridge: Cambridge University Press. Carson, L. (2016). The Sights and Sounds of the Multilingual City. In King & Carson (eds.), 49–83. Cerchiai, L. (2014). Una festa etrusca per Dioniso? In F. Fontana & E. Murgia (eds.), Sacrum facere. Contaminazioni: forme di contatto, traduzione e mediazione nei sacra del mondo greco e romano, 95–105. Trieste: Atti del Seminario di Archeologia del sacro. Cerulli, Domenico. (1777). All’Illustriss. e Rev. Monsignor D. Antonio Gurtler, Vescovo di Tiene Confessore di S. M. la Regina delle due Sicilie sopra un’antica Statua Etrusca. Efemeridi letterarie di Roma 6, 279. Chaniotis, Angelos. (1996). Die Verträge zwischen kretischen Poleis in der hellenistischen Zeit. Stuttgart: Steiner.

301

References Chantraine, Pierre. (1999). Dictionnaire étymologique de la langue grecque: histoire des mots. New edition. Paris: Klincksieck. Chausson, F. (1995). Vel Iovi vel Soli: quatre études autour de la Vigna Barberini (191–354). Mélanges de l’École française de Rome, Antiquité 107, 661–765. Checa Gómez, Esther & José Jacobo Storch de Gracia. (2009). Intérpretes Militares. In Ángel Morillo Cerdán, Norbert Hanel & Esperanza Martín (eds.), Limes 20, 33–44. Madrid: Polifemo. Cinquantaquattro, T. & G. Colonna. (2011). Abella (Avella, AV). Studi Etruschi 74, 330–3. Cipriani, Marina & Fausto Longo (1996). Poseidonia e i Lucani. Naples: Electa Napoli. Clackson, James. (2011). The Social Dialects of Latin. In Clackson (ed.), 506–26. (ed.). (2011). A Companion to the Latin Language. Malden, MA, Oxford and Chichester: Wiley-Blackwell. (2012). Oscan in Sicily. In Olga Tribulato (ed.), Language and Linguistic Contact in Ancient Sicily, 132–48. Cambridge: Cambridge University Press. (2015). Subgrouping in the Sabellian Branch of Indo-European. Transactions of the Philological Society 113, 1–34. (2016). The Language of a Pompeian Tavern: Submerged Latin? In J. N. Adams & Nigel Vincent (eds.), Early and Late Latin. Continuity or Change?, 69–86. Cambridge: Cambridge University Press. Clackson, James & Geoffrey Horrocks. (2007). The Blackwell History of the Latin Language. Malden, MA: Blackwell. Coldstream, Nicolas. (2004). Prospectors and Pioneers: Pithekoussai, Kyme and Central Italy. In Tsetskhladze & De Angelis (eds.), 47–59. Coleman, Robert G. G. (1990). Dialectal Variation in Republican Latin, with Special Reference to Praenestine. Proceedings of the Cambridge Philological Society 36, 1–25. Colonna, Giovanni. (1970). Una nuova iscrizione etrusca del VII secolo e appunti sull’epigrafia ceretana dell’epoca. Mélanges de l’École française de Rome, Antiquité 82, 637–72. (1976a). Scriba cum rege sedens. In Raymond Bloch (ed.), Mélanges offerts à Jacques Heurgon. L’Italie préromaine et la Rome républicaine (Publications de l’École française de Rome 27), 187–95. Rome: École française de Rome. (1976b). Ceramica di importazione a Roma. In Giovanni Colonna (ed.), Civiltà del Lazio primitivo, 367–76. Rome: Multigrafica. (1977). Nome gentilizio e società, Studi Etruschi 45, 175–92. (1981). Quali Etruschi a Roma. In G. Colonna (ed.), Gli Etruschi a Roma. Incontro di studi in onore di Massimo Pallottino, 159–72. Rome: Bretschneider. (1987). Etruria e Lazio nell età dei Tarquini. In M. Cristofani (ed.), Etruria e Lazio arcaico, 55–66. Rome: Consiglio nazionale delle ricerche.

302

References (1988). I Latini e gli altri popoli nel Lazio. In Italia omnium terrarum alumna. La civiltà dei Veneti, Reti, Liguri, Celti, Piceni, Umbri, Latini, Campani e Iapigi, 441–528. Milan: Scheiwiller. (1990). Le iscrizioni etrusche di Fratte. In Giovanna Greco & Angela Pontradolfo (eds.), Fratte. Un insediamento etrusco-campano, 301–9. Modena: Franco Cosimo Panini. (1993). Il santuario di Cupra tra Etruschi, Greci, Umbri e Picenti. In G. Paci (ed.), Cupra Marittima e il suo territorio in età antica. Atti del convegno (Cupra Marittima, 1992), 3–31. Tivoli: Tipigraf. (1995). Etruschi a Pithecusa nell’orientalizzante antico. In Marino Storchi (ed.), L’incidenza dell’antico. Studi in memoria di Ettore Lepore. Proceedings of the International Congress (Anacapri March 24th–28th, 1991), 325–42. Naples: Luciano. (1997a). L’anfora etrusca di Dresda con il sacrificio di Larth Vipe. In Jonas Ellerström (ed.), Amico Amici. Studies in honor of Gad Rausing, 195–216. Lund: Signum. (1997b). Larice Crepu vasaio a San Giovenale. In Börje Magnusson (ed.), Ultra terminum vagari. Scritti in onore di Carl Nylander, 61–76. Rome: Quasar. (2000). Due città e un tiranno. Annali della Fondazione per il Museo ‘Claudio Faina’ 7, 277–89. (2004). I greci di Caere. Annali della Fondazione per il Museo ‘Claudio Faina’ 11, 69–94. (2013). Mobilità geografica e mercenariato nell’Italia preromana. Annali della Fondazione per il Museo ‘Claudio Faina’ 20, 7–22. (2014). Tra Etruria e Roma. Storia di una parola (e forse di un’istituzione). Mediterranea 11, 123–39. (2016). La scrittura e la tomba. Il caso dell’Etruria arcaica. In Haack (ed.), 125–37. Colonna, Giovanni & Sandra Gatti. (1990). Graffiti arcaici dai santuari degli Ernici. Quaderni del centro di studio per l’archeologia etrusco-italica 19, 241–7. Colonna, G. & M. Sannibale. (2015). Caere. Località Sorbo: tomba RegoliniGalassi. Studi Etruschi 77, 240–6. Compatangelo-Soussignan, Rita & Christian-Georges Schwentzel (eds.). (2007). Étrangers dans la cité romaine: habiter une autre patrie, des incolae de la république aux peuples fédérés du bas-empire. Actes du colloque de Valenciennes, 14–15 octobre 2005. Rennes: Presses Universitaires de Rennes. Consani, Carlo. (1991). Διάλεκτος. Contributo alla storia del concetto di ‘dialetto’. Pisa: Giardini. (2014). Ancient Greek Sociolinguistics and Dialectology. In Giannakis (ed.), Volume I, 117–24. Constantakopoulou, Christy. (2017). Aegean Interactions: Delos and Its Networks in the Third Century. Oxford: Oxford University Press.

303

References Cornell, Tim. (1991). The Tyranny of the Evidence: A Discussion of the Possible Uses of Literacy in Etruria and Latium in the Archaic Age. In John Humphrey (ed.), Literacy in the Roman World (Journal of Roman Archaeology, supplementary series 3), 7–33. Ann Arbor, MI: Journal of Roman Archaeology. (1995). The Beginnings of Rome: Italy and Rome from the Bronze Age to the Punic Wars (c. 1000–264 BC). London: Routledge. Couilloud, Marie-Thérese. (1970). Les graffites du Gymnase. In Jean Audiat (ed.), Le Gymnase (Exploration Archéologique de Délos 28), 101–37. Paris: de Boccard. (1974). Les monuments funéraires de Rhénée (Exploration Archéologique de Délos 30). Paris: de Boccard. Crawford, Michael. (1996). Roman Statutes. Two volumes. London: Institute of Classical Studies. (2006). The Oscan Inscriptions of Messana. In Carmine Ampolo (ed.), Guerra e pace in Sicilia e nel Mediterraneo antico (VIII–III sec. a.C.). Volume 2, 521–5. Pisa: Edizioni della Normale. (2007). The Mamertini, Alfius and Festus. In Julien Dubouloz & Sylvie Pittia (eds.), La Sicile de Cicéron. Lectures des Verrines. Actes du colloque de Paris, 19–20 mai 2006, 275–9. Besançon: Presses universitaires de Franche-Comté. Crawford, Michael et al. (2011). Imagines Italicae: A Corpus of Italic Inscriptions (Bulletin of the Institute of Classical Studies Supplement 110). London: Institute of Classical Studies. Crawley Quinn, Josephine. (2011). The Cultures of the Tophet: Identification and Identity in the Phoenician Diaspora. In Erich Gruen (ed.), Cultural Identity in the Ancient Mediterranean, 388–413. Los Angeles: Getty Research Institute. Cristofani, Mauro. (1974). Diffusione dell’alfabeto e onomastica arcaica nell’Etruria interna settentrionale. In Aspetti e problemi dell’Etruria interna. Atti dell’VIII Convegno di Studi Etruschi ed Italici (Orvieto, 1972), 307–24. Florence: L. S. Olschki. (1979). Recent Advances in Etruscan Epigraphy and Language. In David Ridgway & Francesca Ridgway (eds.), Italy Before the Romans. The Iron Age, Orientalizing and Etruscan Periods, 373–412. London and New York: Academic Press. (1981). Varietà linguistica e contesti sociali di pertinenza nell’antroponimia etrusca. Annali dell’Istituto Universitario Orientale di Napoli, sezione linguistica 3, 47–78. (1992). Presenze etrusche tra Stabia e Pontecagnano. Atti e memorie della società Magna Grecia 1, 61–6. (1994). Sulle più antiche iscrizioni italiche della Campania. In P. Castaldi & G. Maetzke (eds.), La presenza etrusca nella Campania meridionale. Atti delle giornate di studio, Salerno – Pontecagnano 16–18 novembre 1990 (Biblioteca di Studi Etruschi 28), 379–86. Florence: L. S. Olschki.

304

References (1995). Tabula Capuana. Florence: L. S. Olschki. (1996). Nuove evidenze epigrafiche da Colle del Giglio. In Guglielmo Maetzke (ed.), Identità e Civiltà dei Sabini. Atti del XVIII Convegno di Studi Etruschi e Italici (Rieti-Magliano Sabina, 1993), 215–26. Florence: L. S. Olschki. (1998). Sull’ origine della scrittura osco-greca. Studi Etruschi 62, 275–9. Crouzet, Sandrine. (2004). Les sarcophages du prêtre de Tarquinia et Carthage, témoignage des relations entre Carthage et l’Étrurie au IVe siècle. Associazione Internazionale di Archeologia Classica News 39–40, 15. Csapo, Eric, Hans Goette, Richard Green & Peter Wilson (eds.). (2014). Greek Theatre in the Fourth Century B.C. Berlin and Boston: Walter De Gruyter. Csapo, E. & M. C. Miller. (1991). The ‘Kottabos-Toast’ and an Inscribed Redfigured Cup. Hesperia: The Journal of the American School of Classical Studies at Athens 60, 367–82. Cuozzo, Mariassunta. (2012). Gli Etruschi in Campania. In Gilda Bartoloni (ed.), Introduzione all’Etruscologia, 189–226. Milan: Hoepli. (2013). Etruscans in Campania. In Turfa (ed.), 301–18. Dack, E. Van ’t. (1980). Reizen, expedities en emigratie uit Italië naar Ptolemaeïsch Egypte. Brussels: Paleis der Academiën. (1983). L’armée Romaine d’Égypte de 55 à 30 AV. J-C. In Günter Grimm, Heinz Heinen & Erich Winter (eds.), Das römisch-byzantinische Ägypten. Mainz am Rhein: Philipp von Zabern. D’Agostino, Bruno. (2001). The Etruscans in Campania. In Gioannangelo Camporeale (ed.), The Etruscans Outside Etruria, 236–51. Los Angeles: J. Paul Getty Museum. Dana, D. (2014). Notices épigraphiques et onomastiques I. Zeitschrift für Papyrologie und Epigraphik 188, 181–98. Darwin, Charles. (1859). On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. London: Murray. Deac, Dan. (2013). Negotiating with the Dacians: The Case of Marcus Ulpius Celerinus, interpres dacorum. Ephemeris Napocensis 23, 313–21. De Hoz, Javier. (2004). The Greek Man in the Iberian Street: Non-colonial Greek Identity in Spain and Southern France. In Lomas (ed.), 411–27. De Ligt, Luuk. (2007a). The Northern-Picene Inscription from Novilara: Language and Meaning. Parola del Passato 62, 53–69. (2007b). The Inscription from Centuripe: Language, Meaning and Historical Background. Glotta 83, 30–42. (2012). Peasants, Citizens and Soldiers: Studies in the Demographic History of Roman Italy 225 BC–AD 100. Cambridge: Cambridge University Press. De Ligt, Luuk & Simon J. Northwood (eds.). (2008). People, Land, and Politics: Demographic Developments and the Transformation of Roman Italy 300 BC–AD 14. Leiden and Boston: Brill.

305

References De Ligt, L. & Laurens E. Tacoma (eds.). (2016a). Migration and Mobility in the Early Roman Empire (Studies in Global Social History, 23). Leiden and Boston. (2016b). Approaching Migration in the Early Roman Empire. In L. De Ligt & E. Tacoma (eds.), 1–22. Dell’ Oro, Francesca. (2017). Alphabets and Dialects in the Euboean Colonies of Sicily and Magna Graecia or What Could Have Happened in Methone. In J. Strauss Clay, I. Malkin & Y. Tzifopoulos (eds.), Panhellenes at Methone. Graphê in Late Geometric and Protoarchaic Methone, Macedonia (ca 700 BCE), 165–81. Berlin and Boston: Walter de Gruyter. Del Monaco, Lavinio. (2013a). Iscrizioni greche d’Italia: Locri. Rome: Edizioni Quasar. (2013b). Le dediche di basileis alla dea Roma e Giove Capitolino: bilancio degli studi e prospettive di ricerca. Mediterraneo Antico 16, 583–96. Del Tutto Palma, Loretta. (1989). Epigrafia lucana. Quaderni dell’Istituto di Linguistica dell’Università di Urbino 6, 93–118. (1996). La Tavola di Agnone nel contesto italico. Convegno di studio, Agnone 13–15 aprile 1994. Florence: L. S. Olschki. De Melo, Wolfgang D. C. (2007). The Early Latin Verb System: Archaic Forms in Plautus, Terence, and Beyond. Oxford: Oxford University Press. Demoule, Jean-Paul. (2014). Mais où sont passes les Indo-Européens? Le mythe d’origine de l’Occident. Paris: Seuil. Dench, Emma. (1995). From Barbarians to New Men. Greek, Roman, and Modern Perceptions of Peoples in the Central Apennines. Oxford: Clarendon. Depauw, Mark. (2016). Roman Influence on Identification Rituals. In Wouter Vanacker & Arjan Zuiderhoek (eds.), Imperial Identities in the Roman World, 176–204. London and New York: Routledge. Derks, Ton. (1998). Gods, Temples and Ritual Practices: The Transformation of Religious Ideas and Values in Roman Gaul. Amsterdam: Amsterdam University Press. De Sensi Sestito, Giovanna. (1995). I due Dionisii e la nascita della confederazione brettia. In Giovanna De Sensi Sestito (ed.), I Brettii. Cultura, lingua e documentazione storico-archeologica, 33–71. Soveria Mannelli: Rubbettino. (2015). Magna Grecia e Sicilia da Agatocle a Pirro. In Aldo Siciliano & Katia Mannino (eds.), La Magna Grecia da Pirro ad Annibale. Atti del 52 Convegno internazionale di studi sulla Magna Grecia, Taranto 27–30 settembre 2012, 39–73. Taranto: Istituto per la storia e l’archeologia della Magna Grecia. De Simone, Carlo. (1970). Die griechischen Entlehnungen im Etruskischen. Two volumes. Wiesbaden: Harrassowitz. (1972). Nuova iscrizione di Cere. Studi Etruschi 40, 153–81.

306

References (1976). Nota di onomastica italica: i gentilizi in turs-. In V. Pisani & C. Santori (eds.), Italia linguistica nuova ed antica. Studi linguistici in memoria di O. Parlangèli. Volume I, 119–26. Galatina: Congedo. (1978). Un nuovo gentilizio etrusco di Orvieto (Katacina) e la cronologia della penetrazione celtica (gallica) in Italia. Parola del Passato 33, 370–95. (1989). Etrusco Acvilna – latino Aquilius: un problema di interscambio linguistico. Parola del Passato 44, 263–80. (1991). Etrusco Laucie Mezenties. Archeologia Classica 43, 559–73. (1995). L’iscrizione di Artos. In Marina Mazzei (ed.), Arpi. L’ipogeo della Medusa e la necropoli, 211–12. Bari: Edipuglia. (1996). La nuova iscrizione aurunca arcaica e il nome della dea Marīca. Studi Classici e Orientali 45, 83–5. (2011). La nuova iscrizione ‘tirsenica’ di Lemnos (Efestia, teatro): considerazioni generali. Rasenna 3, 1–34. De Simone, Carlo & Simona Marchesini. (2002). Monumenta linguae messapicae. Wiesbaden: Ludwig Reichert. Dickey, E. (2010). Latin Influence and Greek Request Formulae. In T. V. Evans & D. D. Obbink (eds.), The Language of the Papyri, 208–20. Oxford: Oxford University Press. D’Isanto, Gennaro. (1993). Capua romana. Ricerche di prosopografia e storia sociale (= Vetera 9). Rome: Quasar. Dise, Robert L. (1991). Cultural Change and Imperial Administration: The Middle Danube Provinces of the Roman Empire. New York: Peter Lang. Dittenberger, W. (1903–5). Orientis graeci inscriptiones selectae. Two volumes. Leipzig: Hirzel. Di Vita-Evrard, Ginette. (1995). Légionnaires africains en Pannonie au IIe s. ap. J. C. In Gábor Hajnóczi (ed.), La Pannoniae l’impero romano, 97–114. Milan: Electa. Donner, H. & W. Röllig. (1962–2002). Kanaanäische und aramäische Inschriften. Three volumes. Wiesbaden: Harrassowitz. Dörig, José. (1976). Kalon d’Elide. In P. Ducrey, C. Bérard, C. Dunant & F. Paschoud (eds.), Mélanges d’histoire ancienne et d’archéologie offerts à Paul Collart, 125–46. Lausanne and Paris: Bibliothèque historique vaudoise. Dubuisson, Michel. (1983). Les opici: Osques, Occidentaux ou Barbares? Latomus 42, 522–45. (1992), Le grec à Rome à l’époque de Cicéron extension et qualité du bilinguisme. Économies, Sociétés, Civilisations 47, 187–206. Duhoux, Yves. (1987). Le vocalisme des inscriptions attiques. Une question de méthodes. Verbum 10, 179–98. Durrbach, Felix. (1921). Choix d’inscriptions de Délos avec traduction et commentaire. Paris: Leroux. (Reprinted 1976, Hildesheim and New York: Olms).

307

References Edwards, C. & G. Woolf. (2003). Rome the Cosmopolis. Cambridge: Cambridge University Press. Eichner, Heiner & Regula Frei-Stolba. (1990). Interessante Einzelobjekte aus der Sammlung von Planta im Rätischen Museum Chur. 1. Teil: das oskische Sprachdenkmal VETTER Nr. 102. Jahresbericht des Rätischen Museums Chur, 67–119. El-Khouly, Aly. (1973). Excavations East of the Serapeum at Saqqâra. Journal of Egyptian Archaeology 59, 151–5. Elton, Hugh. (1996). Frontiers of the Roman Empire. London: Batsford. Equini Schneider, E. (1988). Palmireni a Roma e nell’Africa del nord. Tradizionalismo linguistico e religioso. In E. Campanile, G. R. Cardona, and R. Lazzeroni (eds.), Bilinguismo e biculturalismo nel mondo antico. Atti del Colloquio interdisciplinare (Pisa, 28–29 settembre 1987), 61–6. Pisa: Giardini. Erdkamp, Paul (ed.). (2007). A Companion to the Roman Army. Malden, MA and Oxford: Blackwell. (2008). Mobility and Migration in Italy in the Second Century BC. In De Ligt & Northwood (eds.), 417–50. Ernout, Alfred. (1905–6). Le parler de Préneste d’après les inscriptions. Mémoires de la Société de Linguistique 13, 293–349. (1947). Recueil de textes latines archaïques. Paris: Klincksieck. Erskine, Andrew. (2013). Encountering Carthage: Mid-Republican Rome and Mediterranean Culture. In Andrew Gardner, Edward Herring & Katherine Lomas (eds.), Creating Ethnicities and Identities in the Roman World (Bulletin of the Institute of Classical Studies Supplement 120), 113–29. London: Institute of Classical Studies. Eska, Joseph F. (1998). The Linguistic Position of Lepontic. In Benjamin K. Bergen, Madelaine C. Plauché & Ashlee C. Bailey (eds.), Proceedings of the Twenty-Fourth Annual Meeting of the Berkeley Linguistics Society, February 14–16 1998. Special Session on Indo-European Subgrouping and Internal Relations, February 14 1998, 2–11. Berkeley, CA: Berkeley Linguistics Society. (2013). A Salvage Grammar of Galatian. Zeitschrift für celtische Philologie 60: 51–63. Estarán Tolosa, María José. (2017). Epigrafía bilingüe del Occidente romano. El latín y las lenguas locales en las inscripciones bilingües y mixtas. Zaragoza: Universidad de Zaragoza. (forthcoming). Los primeros contactos entre la península Itálica y la península Ibérica. La documentación epigráfica. Unpublished paper. Evans, T. V. (2012). Latin in Egypt. In Christina Riggs (ed.), The Oxford Handbook of Roman Egypt, 516–25. Oxford: Oxford University Press. Facchetti, G. M. (2007). Alcune note sull’evoluzione storica del nome di famiglia. Alessandria 1, 111–62. Fantham, Elaine. (2004). Maidens in Other-land or Broads Abroad: Plautus’ Poenulae. In Baier (ed.), 235–51.

308

References Feeney, Denis. (2016). Beyond Greek: The Beginnings of Latin Literature. Cambridge, MA and London: Harvard University Press. Feissel, Denis. (1995). Aspects de l’immigration à Constantinople d’après les épitaphes protobyzantines. In Cyril A. Mango & Gilbert Dagron (eds.), Constantinople and Its Hinterland. Papers from the Twenty-Seventh Spring Symposium of Byzantine Studies, 367–77. Aldershot: Variorum. Felle, A. E. (1997). Manifestazioni di bilinguismo nelle iscrizioni cristiane di Roma. In Silvia Evangelisti et al. (eds.), XI Congresso Internazionale di Epigrafia Greca e Latina, 669–78. Rome: Quasar. Fentress, Lisa. (2013). Strangers in the City: Elite Communication in the Hellenistic Central Mediterranean. In Jonathan Prag & Josephine Crawley Quinn (eds.), The Hellenistic West: Rethinking the Ancient Mediterranean, 157–78. Cambridge: Cambridge University Press. Ferrary, Jean-Louis. (2001). Rome et les cités grecques d’Asie Mineure au IIe siécle. In A. Bresson & R. Descat (eds.), Les cités d’Asie mineure occidentale au IIe siècle a.C., 93–106. Bordeaux: Ausonius. (2002). La Création de la province d’Asie et la présence italienne en Asie Mineure. In Müller & Hasenohr (eds.), 133–46. Ferrary, Jean-Louis et al. (2002). Liste des Italiens de Délos. In Müller & Hasenohr (eds.), 183–239. Fitz, Jenő. (1994). Die Verwaltung Pannoniens in der Römerzeit III. Budapest: Encyclopedia. Forsythe, Gary. (2005). A Critical History of Early Rome: from Prehistory to the First Punic War. Berkeley and London: University of California Press. Fraenkel, Eduard (translated by T. Drevikovsky & F. Muecke). (2007). Plautine Elements in Plautus (Plautinisches im Plautus). Oxford: Oxford University Press. Franchi de Bellis, Annalisa. (2014). L’iscrizione prenestina di Orcevia. In Roberto Giacomelli & Adele Robbiati Bianchi (eds.), Le lingue dell’Italia antica oltre il latino: lasciamo parlare i testi, 111–37. Milan: Istituto Lombardo di Scienze e Lettere. Frank, Tenney (ed.). (1975). An Economic Survey of Ancient Rome. Four volumes. Second edition. New York: Octagon Books. Fränkel, Max. (1890). Die Inschriften von Pergamon. Bd. 1. Bis zum Ende der Königszeit. Berlin: Spemann. Franko, George F. (1994). The Use of Poenus and Carthaginiensis in Early Latin Literature. Classical Philology 89, 153–8. (1996). The Characterization of Hanno in Plautus’ Poenulus. American Journal of Philology 117, 425–52. Fraser, P. M. (1972). Ptolemaic Alexandria. Oxford: Clarendon. (2009). Greek Ethnic Terminology. Oxford: Oxford University Press. Fraser, P. M. & E. Matthews (1997). A Lexicon of Greek Personal Names III.A. Peloponnese, Western Greece, Sicily, and Magna Graecia. Oxford: Oxford University Press.

309

References Freeman, Philip. (2001). The Galatian Language: A Comprehensive Survey of the Language of the Ancient Celts in Greco-Roman Asia Minor. Lewiston and New York: Edwin Mellen. Frye, R. N. (1968). The Parthian and Middle Persian Inscriptions of DuraEuropos. London: Percy Lund and Humphries. Gàbrici, Ettore. (1910). Necropoli di età ellenistica a Teano dei Sidicini. Monumenti Antichi 20, 5–152. Gagliardi, Lorenzo. (2006). Mobilità e integrazione delle persone nei centri cittadini romani: aspetti giuridici. Milan: Pubblicazioni dell’Istituto di Diritto Romano, Università degli Studi di Milano. Gallego Franco, Henar. (2003). Intérpretes militares en el limes del Danubio. Aquila Legionis 4, 27–43. Garland, Robert. (2014). Wandering Greeks. The Ancient Greek Diaspora from the Age of Homer to the Death of Alexander the Great. Princeton and Oxford: Princeton University Press. Gatti, Sandra. (1993). Dives Anagnia. Archeologia nella Valle del Sacco. Rome: Bretschneider. Gee, R. (2010). The Lucus Furrinae and the Syrian Sanctuary on the Janiculum: Encroachment? Or Renovation and Transformation? Bollettino di Archeologia on line Edizione Speciale (Congresso di archeologia A.I.A.C. 2008), 42–8. (Available at www.bollettinodiarcheologiaonline.benicultur ali.it). Giannakis, Georgios K. (ed.). (2014). Encyclopedia of Ancient Greek Language and Linguistics. Three volumes. Leiden: Brill. Gignac, F. T. (1976). A Grammar of the Greek Papyri of the Roman and Byzantine Periods: Volume I, Phonology. Milan: Cisalpino Goliadarca. (1981). A Grammar of the Greek Papyri of the Roman and Byzantine Periods: Volume II, Morphology. Milan: Cisalpino Goliadarca. Giles, Howard, Richard Y. Bourhis & Donald M. Taylor. (1977). Towards a Theory of Language in Ethnic Group Relations. In Howard Giles (ed.), Language, Ethnicity and Intergroup relations, 307–48. London: Academic Press. Giles, Howard, N. Coupland & J. Coupland. (1991). Accommodation Theory: Communication, Context and Consequence. In Howard Giles, N. Coupland & J. Coupland (eds.), Contexts of Accommodation: Developments in Applied Sociolinguistics. Cambridge: Cambridge University Press. Goldsworthy, Adrian & Ian P. Haynes (eds.). (1999). The Roman Army as a Community. Portsmouth, RI: Journal of Roman Archaeology. González, Julian & Michael Crawford. (1986). The Lex Irnitana: A New Copy of the Flavian Municipal Law. Journal of Roman Studies 76, 147–243. Gordon, Arthur E. (1962). On the Origins of the Latin Alphabet: Modern Views. California Studies in Classical Antiquity 2, 157–70. Gordon, Richard & Joyce Reynolds. (2003). Roman Inscriptions 1995–2000. Journal of Roman Studies 93, 212–94.

310

References Gratwick, Adrian S. (1982). Drama. In Edward J. Kenney & Wendell V. Clausen (eds.), The Cambridge History of Classical Literature. Volume II: Latin Literature, 77–137. Cambridge: Cambridge University Press. (1993). Plautus Menaechmi. Cambridge: Cambridge University Press. Greca, E. (2012). Roma, città Greca. In A. Giardina & F. Pesando (eds.), Roma caput mundi: una città tra dominio e integrazione, 69–77. Milan: Electa. Greco Pontrandolfo, Angela & Agnès Rouveret. (1983). La rappresentazione del barbaro in ambiente Magno-Greco. In Giuseppe Nenci (ed.), Modes de contacts et processus de transformation dans les sociétés anciennes (Collection de l’École française de Rome 67), 1051–66. Rome: École française de Rome. Gregoratti, Leonardo. (2014). North Italic Settlers along the ‘Amber Route’. Studia Antiqua et Archaeologica 19, 133–53. Grenfell, B. P., A. S. Hunt & J. G. Smyly. (1902). The Tebtunis Papyri. London. Egypt Exploration Society. Gruen, Erich S. (1984). The Hellenistic World and the Coming of Rome. Berkeley, CA: University of California Press. Gualtieri, Maurizio. (2004). Between Samnites and Lucanians: New Archaeological and Epigraphic Evidence for Settlement Organization. In Howard Jones (ed.), Samnium. Settlement and Cultural Change. The Proceedings of the Third E. Togo Salmon Conference on Roman Studies (= Archaeologica Transatlantica XXII), 35–50. Providence, RI: Center for Old World Archaeology and Art, Brown University. Guijarro Ruano, Paloma. (2015). La lengua de las inscripciones métricas laconias anteriores al 400 a.C. Cuadernos de Filología Clásica: Estudios Griegos e Indoeuropeos 25, 25–37. Haack, Marie-Laurence (ed.). (2016). L’écriture et l’espace de la mort. Atti del seminario presso l’École française de Rome (Roma, 2009). Rome: École française de Rome. Habinek, Thomas N. (1998). The Politics of Latin Literature: Writing Identity and Empire in Ancient Rome. Princeton: Princeton University Press. Hadas-Lebel, J. (2009). Anthroponymes toponymiques et toponymes anthroponymiques. Liens entre lieux et personnes dans l’onomastique étrusque. In Poccetti (ed.), 195–217. Hall, Jonathan. (2004). How ‘Greek’ Were the Early Western Greeks? In Lomas (ed.), 335–54. Handel, Jacob. (1913). De lingua communi in titulos ionicos irrepente. Lemberg: Gubrynowicz and Son. Hanson, John A. (1959). Plautus as Sourcebook for Roman Religion. Transactions of the American Philological Association 90, 48–101. Harlow, Mary & Ray Laurence. (2002). Growing Up and Growing Old in Ancient Rome: A Life Course Approach. London: Routledge. Hartley, Brian & John Wacher (eds.). (1983). Rome and Her Northern Provinces. Gloucester: Sutton.

311

References Hartmann, Markus. (2005). Die frühlateinischen Inschriften und ihre Datierung. Bremen: Hempen. Hartog, François. (1988). The Mirror of Herodotus. Berkeley and Los Angeles: University of California Press. Hasenohr, Claire. (2002). Les collèges de magistri et la communauté italienne de Délos. In Müller & Hasenohr (eds.), 67–76. (2007a). Italiens et Phéniciens à Délos: organisation et relations de deux groupes d’étrangers résidents (IIe–Ier siècles av. J.-C.). In CompatangeloSoussignan & Schwentzel (eds.), 77–90. (2007b). Les Italiens à Délos: entre romanité et hellénisme. Pallas 73, 221–32. Hatzfeld, Jean. (1912). Les Italiens résidant à Délos mentionnés dans les inscriptions de l’île. Bulletin de Correspondance Hellénique 36, 1–218. (1919). Les trafiquants italiens dans l’Orient hellénique (Bibliothèque des Écoles françaises d’Athènes et de Rome 115). Paris: de Boccard. Haynes, Sybille. (2013). Etruscan Civilization: A Cultural History. Los Angeles: J. Paul Getty Museum. Henderson, John. (1999). Writing down Rome: Satire, Comedy and Other Offences in Latin Poetry. Oxford: Oxford University Press. Hermon, Ella. (2007). Des communautés distinctes sur le même territoire: quelle fut la réalité des incolae? In Compatangelo-Soussignan & Schwentzel (eds.), 25–42. Heurgon, Jacques. (1957). Influences grecques sur la religion étrusque: l’inscription de Laris Pulenas. Révue des Études Latines 35, 106–26. (1966). The inscriptions of Pyrgi. Journal of Roman Studies 56, 1–13. Hin, Saskia (2013). The Demography of Roman Italy: Population Dynamics in an Ancient Conquest Society (201 BCE–14 CE). Cambridge: Cambridge University Press. Hodos, Tamar. (1999). Intermarriage in the Western Greek Colonies. Oxford Journal of Archaeology 18, 61–78. Honigman, S. (2003). Noms sémitiques à Edfou et Thèbes. Bulletin of the American Society of Papyrologists 40, 63–118. Horbury, W. & D. Noy. (1992). Jewish Inscriptions of Graeco-Roman Egypt. Cambridge: Cambridge University Press. Horden, Peregrine & Nicholas Purcell. (2000). The Corrupting Sea: A Study of Mediterranean History. Cambridge: Cambridge University Press. Hornblower, Simon. (2008). A Commentary on Thucydides, Volume III: Books 5.25–8.109. Oxford: Oxford University Press. Horrocks, Geoffrey. (2010). Greek. A History of the Language and Its Speakers. Second Edition. Malden, MA, Oxford and Chichester: Wiley-Blackwell. Hurwit, Jeffrey. (2015). Artists and Signatures in Ancient Greece. Cambridge: Cambridge University Press. Huyse, Philip. (1990). Iranische Namen in den griechischen Dokumenten Ägyptens Österreichische Akademie der Wissenschaften. Philosophischhistorische Klasse, Sonderpublikation der Kommission für Iranistik,

312

References Iranisches Personennamenbuch. Band V. Iranische Namen in den Nebenüberlieferungen indogermanischer Sprachen. Faszikel 6). Vienna: Österreichische Akademie der Wissenschaften. Ilan, Tal & Thomas Ziem. (2008). Lexicon of Jewish Names in Late Antiquity: Part III – The Western Diaspora 330 CE–650 CE. Tübingen: Mohr Siebeck. Isayev, Elena. (2007). Inside Ancient Lucania: Dialogues in History and Archaeology. London: Institute of Classical Studies. (2017). Migration, Mobility and Place in Ancient Italy. Cambridge: Cambridge University Press. Jakobson, Roman & Morris Halle. (1963). Tenseness and Laxness. In Roman Jakobson, C. Gunnar M. Fant & Morris Halle (eds.), Preliminaries to Speech Analysis. Fourth edition, 57–61. Cambridge, MA: Massachusetts Institute of Technology Press. James, Patrick. (2015). Ethnic Terminology in Muraoka’s Greek-English Lexicon of the Septuagint. In J. K. Aitken & T. V. Evans (eds.), Biblical Greek in Context, 117–36. Leuven: Peeters. Janko, Richard. (2015). From Gabii and Gordion to Eretria and Methone: The Rise of the Greek Alphabet. Bulletin of the Institute of Classical Studies 58, 1–32. Jeffery, L. H. (1990). The Local Scripts of Archaic Greece: A Study of the Origin of the Greek Alphabet and Its Development from the Eighth to the Fifth Centuries B.C. (Revised edition by A. W. Johnston). Oxford: Oxford University Press. Jocelyn, Harry D. (1969). The Tragedies of Ennius. Cambridge: Cambridge University Press. Jongman, W. (2003). Slavery and the growth of Rome. The transformation of Italy in the second and first centuries BCE. In C. Edwards & G. Woolf (eds.), Rome the Cosmopolis, 100–22. Cambridge: Cambridge University Press. Joseph, Brian & Rex Wallace. (1987). Latin sum / Oscan sum, sim, esum. The American Journal of Philology 108, 675–93. (1991). Is Faliscan a Local Latin Patois? Diachronica 8, 159–86. Kaimio, J. (1979). The Romans and the Greek Language. Helsinki: Societas Scientiarum Fennica. Kajanto, I. (1963). A Study of the Greek Epitaphs of Rome. Helsinki: Societas Scientiarum Fennica. (1965). The Latin cognomina. Helsinki: Societas Scientiarum Fennica. (1980). Minderheiten und ihre Sprachen in der Hauptstadt Rom. In G. Neumann & J. Untermann (eds.), Die Sprachen im römischen Reich der Kaiserzeit, 83–101. Cologne and Bonn: Habelt. Kasimis, Demetra. (2018). The Perpetual Immigrant and the Limits of Athenian Democracy. Cambridge: Cambridge University Press. Kassian, Alexei. (2016). Un-Making Sense of Alleged Abkhaz-Adyghean Inscriptions on Ancient Greek Pottery. Ancient Civilizations from Scythia to Siberia 22, 177–98.

313

References Kayser, François. (1994), Recueil des inscriptions grecques et latines (non funéraires) d’Alexandrie impériale. Cairo. Keppie, Lawrence. (1991). Understanding Roman Inscriptions. Baltimore: Johns Hopkins University Press. Kerr, Robert M. (2010). Latino-Punic Epigraphy. A Descriptive Study of the Inscriptions. Tübingen: Mohr Siebeck. Kiessling, E. (1970). Wörterbuch der griechischen Papyrusurkunden, mit Einschluss der griechischen Inschriften, Aufschriften, Ostraka, Mumienschilder usw. aus Ägypten. Supplement 1 (1940–1966). Amsterdam. A. M. Hakkert. King, L. (2016). Multilingual Cities and the Future: Vitality or Decline? In King & Carson (eds.), 179–202. King, L. & L. Carson (eds.). (2016). The Multilingual City: Vitality, Conflict and Change Bristol, Buffalo, and Toronto. Multilingual Matters. Kolendo, Jerzy. (1970). L’Influence de Carthage sur la civilisation matérielle de Rome. Archaeologia 31, 8–22. Kolnik, T. (1978). Q. Atilius Primus-Interprex Centurio und Negotiator. Acta Archaeologica Academiae Scientiarum Hungaricae 30, 61–75. Kovács, Péter. (2014). A History of Pannonia during the Principate. Bonn: Habelt. Kretschmer, Paul. (1894). Die griechischen Vaseninschriften: ihrer Sprache nach untersucht. Gütersloh: Bertelsmann. Künow, Jürgen. (1980). Negotiator et Vectura: Händler und Transport im freien Germanien. Munich: Kleine Schriften aus dem Vorgeschichtlichen Seminar Marburg. La’da, Csaba A. (2002). Foreign Ethnics in Hellenistic Egypt. Leuven, Paris, and Dudley, MA: Peeters. Lambert, Pierre-Yves. (2003). La langue gauloise. Paris: Éditions errance. Lancel, Serge. (1992). Carthage: A History. Oxford: Wiley-Blackwell. La Piana, G. (1927). Foreign Groups at Rome during the First Centuries of the Empire. Harvard Theological Review 20, 183–401. La Regina, Adriano. (1980). Dalle guerre sannitiche alla romanizzazione. In Sannio: Pentri e Frentani dal VI al I sec. a.C., 29–42. Rome: De Luca. (2002). La formula onomastica osca in Lucania e nel Bruzzio. Eutopia 2, 57–69. (2011). Il Guerriero di Capestrano e le iscrizioni paleosabelliche. In Luisa Franchi dell’Orto (ed.), Pinna Vestinorum e il popolo dei Vestini, 230–73. Rome: Bretschneider. Lassère, Jean-Marie. (1991). Biographie d’un centurion (C.I.L., VIII, 217–18). Antiquités Africaines 27, 53–68. La Torre, Gioacchino F. (2009). Da Blanda a Temesa: fenomeni di urbanizzazione lungo la fascia tirrenica della Lucania meridionale e del Bruzio settentrionale. In Massimo Osanna (ed.), Verso la città: forme insediative in Lucania e nel mondo italico fra IV e III sec. a. C., 181–94. Venosa: Osanna.

314

References Laurence, Ray. (2000). Metaphors, Monuments and Texts: The Life Course in Roman Culture. World Archaeology 31, 442–55. Lavan, M. (2013). Slaves to Rome: Paradigms of Empire in Roman Culture. Cambridge: Cambridge University Press. Lazzarini, Maria L. & Paolo Poccetti. (2001). Il mondo enotrio tra VI e V secolo a.C. Atti dei seminari napoletani (1996–1998). Naples: Loffredo. Lazzeroni, Romano. (1983). Contatti di lingue e di culture nell’Italia antica. Modelli egemoni e modelli subordinati nelle iscrizioni osche in grafia greca. Annali dell’Istituto Universitario Orientale di Napoli, sezione Linguistica 5, 171–82. (1985). Varianti grafiche e varianti fonetiche nelle iscrizioni osche. Una questione di modo. In Enrico Campanile (ed.), Lingua e cultura degli oschi, 47–53. Pisa: Giardini. Le Dinahet-Couilloud, Marie-Thérese. (2001). Les Italiens de Délos: compléments onomastiques et prosopographiques. Revue des Études Anciennes 103, 103–23. Lee, A. D. (1993). Information and Frontiers: Roman Foreign Relations in Late Antiquity. Cambridge: Cambridge University Press. Lefèvre, Eckard, Ekkehard Stärk & Gregor Vogt-Spira. (1991). Plautus barbarus: sechs Kapitel zur Originalität des Plautus. Tübingen: Gunter Narr. (2004). Contrôle d’identité aux frontières dans les cités grecques. In Moatti (ed.), 98–125. Leigh, Matthew. (2004). Comedy and the Rise of Rome. Oxford and New York: Oxford University Press. Leiwo, Martti. (1989). Philostratus of Ascalon, His Bank, His Connexions and Naples in c. 130–90 B.C. Athenaeum 67, 575–84. (2002). From Contact to Mixture: Bilingual Inscriptions from Italy. In Adams, Janse & Swain (eds.), 168–94. Lejeune, Michel. (1957). Notes de linguistique italique: XIII: Sur les adaptations de l’alphabet étrusque aux langues indo-européennes d’Italie. Latomus 35, 88–105. (1970). Phonologie osque e graphie grecque. Revue des Études Anciennes 72, 271–315. (1972). Phonologie osque et graphie grecque II. Revue des Études Anciennes 74, 5–13. (1975). Réflexion sur la phonologie du vocalisme osque. Bulletin de la Société de linguistique de Paris, 70, 233–51. (1982). Noms grecs et noms indigènes dans l’épigraphie hellénistique d’Entella. Annali della Scuola Normale Superiore di Pisa (Classe di Lettere e Filosofia) 3, 787–99. (1983). Le canthare d’Alise: nouvelles discussions sur les avatars et sur l’origine du vase. Monuments et Mémoires de la Fondation Eugène Piot 66, 19–53. (1988). Compléments gallo-grecs. Études Celtiques 25, 79–106.

315

References (1990). Méfitis d’après les dédicaces lucaniennes de Rossano di Vaglio, Louvain-la-neuve: Peeters. Lengyel, A. & G. T. B. Radan. (1980). The Archaeology of Roman Pannonia. Lexington, Kentucky and Budapest: University Press of Kentucky and Akadémiai Kiadó. Leon, H. J. (1995). The Jews of Ancient Rome. Peabody, MA: Hendrickson. Leumann, Manu. (1977). Lateinische Laut- und Formenlehre. Munich: Beck. Licandro, Orazio. (2007). «Domicilium» e «incolae» tra repubblica e principato. In Compatangelo-Soussignan & Schwentzel (eds.), 43–76. Lindsay, Wallace M. (1894). The Latin Language: An Historical Account of Latin Sounds, Stems and Flexions. Oxford: Clarendon. (1913). Sexti Pompei Festi. De verborum significatu quae supersunt cum Pauli epitome. Leipzig: Teubner. Lintott, A. W. (1978). The Capitoline Dedications to Jupiter and the Roman People. Zeitschrift für Papyrologie und Epigraphik 30, 137–44. Lo Cascio, E. (2016). The Impact of Migration on the Demographic Profile of the City of Rome: A Reassessment. In De Ligt & Tacoma (eds.), 23–32. Lo Cascio, Elio, & Laurens E. Tacoma (eds.). (2016). The Impact of Mobility and Migration in the Roman Empire. Proceedings of the Twelfth Workshop of the International Network Impact of Empire (Rome, June 17–19, 2015). Leiden and Boston: Brill. Lo Porto, Felice Gino. (1988–9). Metaponto (Matera) – Rinvenimenti nella città antica e nel suo retroterra ellenizzato. Atti della Accademia Nazionale dei Lincei: Notizie degli Scavi di Antichità 42–43 (series 8), 299–441. Lomas, Kathryn (ed.). (2004). Greek Identity in the Western Mediterranean. Papers in Honour of Brian Shefton, 335–54. Leiden and Boston: Brill. Lőrincz, Barnabás. (2010–11). Zur Militärgeschichte der Donauprovinzen des römischen Reiches: ausgewählte Studien 1975–2009. Budapest-Debrecen: University of Debrecen. Lowe, Nick J. (2007). Comedy (Greece and Rome New Surveys in Classics 37). Cambridge: Cambridge University Press. MacDonald, B. R. (1981). The Emigration of Potters from Athens in the Late Fifth Century BC and Its Effect on the Attic Pottery Industry. American Journal of Archaeology 85, 15–22. Maggiani, A. (1992). Le iscrizioni di età tardo-classica ed ellenistica. In A. Romualdi (ed.), Populonia in età ellenistica. I materiali dalle necropolis. Atti del seminario (Firenze, 1986), 179–92, Florence: Il Torchio. (1999). Nuovi etnici e toponimi etruschi. In Incontro di studi in memoria di Massimo Pallottino, 47–61. Rome: Bretschneider. (2000). Tipologia tombale e società. Chiusi in età orientalizzante. Annali della Fondazione per il Museo ‘Claudio Faina’ 7, 249–75. (2005a). Il cippo di Larth Cupures veiente e gli altri semata a testa umana da Orvieto. Annali della Fondazione per il Museo ‘Claudio Faina’ 12, 29–73. (2005b). Ager Clusinus: Murlo. Studi Etruschi 71, 162–4.

316

References (2006). Dinamiche del commercio arcaico: le tesserae hospitales. Annali della Fondazione per il Museo ‘Claudio Faina’ 13, 317–49. (2007). Auvele Feluskes. Della stele di Vetulonia e di altre dell’Etruria settentrionale. Rivista di Archeologia 31, 67–75. (2014). Un kyathos di bucchero da Tolle (Chiusi). In S. Bruni (ed.), ‘Lautus erat Tuscis Porsena fictilibus’. Studi e ricerche sul bucchero dell’area chiusina per Luigi Donati, 19–39. Pisa: Edizioni ETS. (2018). Giovannangelo Camoreale e il kyathos di Monteriggioni. In Archeologia a Massa Marittima. Giornata in ricordo di Giovannangelo Camporeale, 101–108. Pisa: Edizioni ETS. Mairs, Rachel. (2010). An Early Roman Application for Lease of a Date Crop (P. Duk. inv. 85) and the ‘Six-Choinix Measure of the Hermeneus’. Zeitschrift für Papyrologie und Epigraphik 172, 183–91. (2011). Translator, Traditor: The Interpreter as Traitor in Classical Tradition. Greece and Rome 58, 64–81. (2012a). Interpreters and Translators in Hellenistic and Roman Egypt. In Paul Schubert (ed.), Actes du 26e Congrès international de papyrologie (Genève 2010), 457–62. Geneva: Droz. (2012b). ‘Interpreting’ at Vindolanda. Britannia 43, 1–12. Malkin, Irad. (2002). A Colonial Middle Ground: Greek, Etruscan, and Local Elites in the Bay of Naples. In Clair L. Lyons & John K. Papadopoulos (eds.), The Archaeology of Colonialism, 151–81. Los Angeles: Getty Research Institute. Mancini, Marco. (2004). Latina antiquissima II: ancora sull’epigrafe del Garigliano. In Vincenzo Orioles (ed.), Studi in memoria di Eugenio Coseriu, 229–52. Udine: Forum. (2005). ‘Dilatandis litteris’. Uno studio su Cicerone e la pronunzia ‘rustica’. In Raffaella Bombi et al. (eds.), Studi linguistici in onore di Roberto Gusmani. Three volumes. 1023–46. Alexandria: Orso. Manteuffel, G. (1949). Quelques textes provenants d’Edfou. Journal of Juristic Papyrology 3, 101–17. Manuwald, Gesine. (2011). Roman Republican Theatre. Cambridge: Cambridge University Press. Maras, D. F. (2003). Note sull’arrivo del nome di Ulisse in Etruria. Studi Etruschi 65–8, 237–49. (2007). Note in margine a CIE II, 1, 5. Studi Etruschi 73, 237–47. (2009). Il dono votivo. Gli dei e il sacro nelle iscrizioni etrusche di culto (Biblioteca di Studi Etruschi 46). Pisa and Rome: Fabrizio Serra. (2010). Ancora su Mastarna, sodalis fidelissimus. Annali della Fondazione per il Museo ‘Claudio Faina’ 17, 187–200. (2012). Materiale epigrafico dal Tumulo Chigi: notizie su testi e contesti. In E. Mugione (ed.), L’Olpe Chigi. Storia di un agalma. Atti dell’Incontro Internazionale (Salerno, 2010) (= Ergasteria) 2, 47–54. Naples: Pandemos. (2013a). Interferenza e concorrenza di modelli alfabetici e sistemi scrittori nell’Etruria arcaica. In F. Haumesser & G. van Heems (eds.), Régler

317

References l’usage: norme et standard dans l’Italie préromaine. Premier atelier: langages. Atti del seminario presso l’École française de Rome (Roma, 2009) (= Mélanges de l’École française de Rome, Antiquité) 124, 331–44. Rome: Fabrizio Serra. (2013b). Questioni di identità: Etruschi e Falisci nell’agro falisco. In G. Cifani (ed.), Identità e cultura dei Falisci. Atti del seminario presso la British School at Rome (Roma, 2011), 265–85. Rome: Quasar. (2014). Il Pittore «delle Gorgoni» di Heidelberg e la seconda generazione di ceramografi ‘Campanizzanti’ in Etruria. Rendiconti della Pontificia Accademia Romana di Archeologia 86, 449–75. (2015). Etruscan and Italic Literacy and the Case of Rome. In W. M. Bloomer (ed.), A Companion to Ancient Education, 201–25. Malden, MA, Oxford and Chichester: Wiley-Blackwell. (2016). Storie di dono: l’oggetto parlante si racconta. In Haack (ed.), 239–51. (2017). Epigraphy and Nomenclature. In G. Bradley & G. Farney (eds.), The Peoples of Ancient Italy, 63–88. Berlin and Boston: De Gruyter. (2018a). Kings and Tablemates. The Political Role of Comrade Associations in Archaic Rome and Etruria. In L. Aigner Foresti & P. Amann (eds.), Beiträge zur Sozialgeschichte der Etrusker: Akten der internationalen Tagung, Wien, 8.–10.6.2016, 91–108. Vienna: Holzhausen. (2018b). Dancing Myths. Musical Performances with Mythological Subjects from Greece to Etruria. In A. Garcia-Ventura, C. Tavolieri & L. Verderame (eds.), The Study of Musical Performance in Antiquity. Archaeology and Written Sources, 137–54. Newcastle upon Tyne: Cambridge Scholars. (forthcoming). L’epigrafia delle corti aristocratiche di Caere in epoca orientalizzante. In Laurent Haumesser (ed.), Cerveteri. La culture écrite d’une cité étrusque. Proceedings of the Colloquium (Paris, 2016), Paris: Musée du Louvre. Maras, D. F. & L. M. Michetti. (2003). Falerii. Studi Etruschi 69, 376–8. Marchesini, S. (1997). Studi onomastici e sociolinguistici sull’Etruria arcaica. Florence: L. S. Olschki. (2007). Prosopographia Etrusca, II, 1. Studia. Gentium mobilitas (Studia Archaeologica 158). Rome: Bretschneider. Marcuse, P. & R. van Kempen. (eds.). (2002). Of States and Cities: The Partitioning of Urban Space. Oxford: Oxford University Press. Marinetti, Anna & Aldo L. Prosdocimi. (1988). Lingue e scritture dei popoli indigeni (Lucani, Brettii, Enotri). In Giovanni Pugliese Carratelli (ed.), Magna Grecia. Vita religiosa e cultura letteraria, filosofica e scientifica, 29–54. Milan: Electa. Markey, Tom. (2001). A Tale of Two Helmets: The Negau A and B Inscriptions. Journal of Indo-European Studies 29, 69–172. Marotta, Giovanna. (2015). Talking Stones. Phonology in Latin Inscriptions? Studi e Saggi Linguistici 53, 39–63.

318

References Marshall, Christopher W. (2006). The Stagecraft and Performance of Roman Comedy. Cambridge: Cambridge University Press. Massa-Pairault, Françoise-Hélène. (1997). Intervento. In M. Mazzei (ed.), Il Caso Arpi. Ambiente italico e magnogreco tra primo e medio ellenismo. Atti della Tavola Rotonda (Foggia 1996), 61–73. Foggia: Soprintendenza per i Beni archeologici della Puglia. (2013–14). La tombe François et la vision politique de l’étrurie au IVe siècle a. C. Ostraka 22–23, 47–91. Massarelli, Riccardo. (2014). I testi etruschi su piombo (Biblioteca di Studi Etruschi) 53. Pisa and Rome: Fabrizio Serra. Masson, Olivier. (1969). Recherches sur les Phéniciens dans le monde hellénistique. Bulletin de Correspondance Hellénique 93, 679–700. Mastelloni, Maria Amalia. (2005). Messana e i Mamertini. In Francesca Ghedini, Jacopo Benetto, Andrea Raffaele Ghiotto & Federica Rinaldi (eds.), Lo stretto di Messina nell’antichità, 275–92. Rome: Società Stretto di Messina. Maurice, Lisa. (2004). The Punic, the Crafty Slave and the Actor: Deception and Metatheatricality in the Poenulus. In Baier (ed.), 267–90. Mayor, Adrienne, John Colarusso & David Saunders (2014). Making Sense of Nonsense Inscriptions Associated with Amazons and Scythians on Athenian Vases. Hesperia 83, 447–93. Mayser, Edwin & Hans Schmoll. (1970). Grammatik der griechischen Papyri aus der Ptolemäerzeit. Second edition. Berlin: Walter De Gruyter. Mazzei, Marina. (1995). L’ipogeo della Medusa. In Marina Mazzei (ed.), Arpi. L’ipogeo della Medusa e la necropoli, 87–129. Bari: Edipuglia. (1998). La pittura ellenistica nella Puglia settentrionale: il caso di Arpi. In L’Italie méridionale et les premières expériences de la peinture hellénistique. Actes de la table ronde organisée par l’École française de Rome 1994, 69–94. Rome: École française de Rome. McDonald, Katherine. (2015). Oscan in Southern Italy and Sicily. Evaluating Language Contact in a Fragmentary Corpus. Cambridge: Cambridge University Press. McDonald, Katherine, Livia Tagliapietra & Nicholas Zair. (2015). New Readings of the Multilingual Petelia Curse Tablet. Zeitschrift für Papyrologie und Epigraphik 195, 157–64. Meiser, Gerhard. (1987). Pälignisch, Latein und Südpikenisch. Glotta 65, 103–25. (1996). Accessi alla protostoria delle lingue sabelliche. In Del Tutto Palma (ed.), 187–209. (2014). Etruskische Texte. Editio minor (auf Grundlage der Erstausgabe von Helmut Rix). Two volumes. Second edition. Hamburg: Baar. Mellor, R. (1978). The Dedications on the Capitoline Hill. Chiron 8, 319–30. Méndez Dosuna, Julian. (1985). Los Dialectos Dorios del Noroeste. Salamanca: Ediciones Universidad de Salamanca. Michalowski K. et al. (1950). Tell Edfou 1939. Cairo: Institut français d’archéologie orientale du Caire and the University of Warsaw.

319

References Mickey, K. (1981). Dialect Consciousness and Literary Language: An Example from Ancient Greek. Transactions of the Philological Society 79, 35–66. Migahid, Abd-El-Gawad. (1998). Eine demotische Personenliste aus dem Serapeum. Zeitschrift für Ägyptische Sprache und Altertumskunde 125, 125–37. Millar, F. (1993). The Roman Near East, 31 B.C. – A.D. 337. Cambridge, MA: Harvard University Press. Milne, Joseph G. (1905). Greek Inscriptions. «Service des Antiquités de l’Égypte: Catalogue géneral des antiquités égyptiennes du Musée du Caire». Oxford: Oxford University Press. Milroy, Lesley. (1980). Language and Social Networks. London and Baltimore: Blackwell and University Park Press. Milroy, James & Lesley Milroy. (1999). Authority in Language: Investigating Standard English. Third edition. London and New York: Routledge. Minnen, P. Van. (2000). An Official Act of Cleopatra (with a Subscription in Her Own Hand). Ancient Society 30, 29–34. Minon, Sophie. (2007). Les inscriptions éléennes dialectales (VIe–IIe siècle avant J.-C.). Geneva: Droz. Moatti, C. (2000). Le contrôle de la mobilité des personnes dans l’Empire romain. Mélanges de l’École française de Rome, Antiquité 112, 925–58. (2004). La mobilité des personnes en Méditerranée de l’antiquité à l’époque moderne: procédures de contrôle et documents d’identifications. Rome: École française de Rome. (2013). Immigration and cosmopolitanization. In Paul Erdkamp (ed.), The Cambridge Companion to Ancient Rome, 77–92. Cambridge: Cambridge University Press. (2014). Mobility and Identity between the Second and the Fourth Century CE: The Cosmopolitization of the Roman Empire. In C. Rapp & H. A. Drake (eds.), The City in the Classical and Post-classical World: Changing Contexts of Power and Identity, 130–52. Cambridge: Cambridge University Press. Mócsy, András. (1974). Pannonia and Upper Moesia: A History of the Middle Danube Provinces of the Roman Empire. London: Routledge. (1983). The Civilized Pannonians of Velleius. In Hartley & Wacher (eds.), 169–78. (1992). Pannonien und das römische Heer. Stuttgart: Franz Steiner. Mol, Eva M. (2013). The Perception of Egypt in Networks of Being and Becoming: A Thing Theory Approach to Egyptianising Objects in Roman Domestic Contexts. In Annabel Bokern et al. (eds.), TRAC 2012. Proceedings of the Twenty-second Annual Theoretical Roman Archaeology Conference, 117–32. Oxford: Oxbow. Moncunill Martí, Noemi. (2017). Indigenous Naming Practices in the Western Mediterranean: The Case of Iberian. Studia Antiqua et Archaeologica 23, 7–20.

320

References Morandi, Alessandro. (2009). Minturno. Santuari di Marica. Iscrizione cosiddetta ‘di Trivia’. Lazio e Sabina. Scoperte Scavi e Ricerche 5, 445–6. Morandi, M. (2004). Prosopographia Etrusca I.1. Etruria meridionale. Rome: Bretschneider. Moretti, L. (1989). I Greci a Roma. Opuscula Instituti Romani Finlandiae 4, 6–16. Morley, Neville. (1997). Cities in Context: Urban Systems in Roman Italy. In Helen Parkins (ed.), Roman Urbanism: Beyond the Consumer City, 42–58. London and New York: Routledge. (2001). The Transformation of Italy, 225–28 BC. Journal of Roman Studies 91, 50–62. Mosser, Martin. (2003). Die Steindenkmäler der legio XV Apollinaris. Vienna: Forschungsgesellschaft Wiener Stadtarchäologie. Motta, Filippo. (2009). Tipologie dell’onomastica personale celtica nell’Italia antica. In Poccetti (ed.), 295–318. Mráv, Zsolt. (2010–13). The Roman Army along the Amber Road between Poetovio and Carnuntum in the 1st Century A.D. – Archaeological Evidence: A Preliminary Research Report. Communicationes Archælogicae Hungariae 2010–13, 49–100. Mullen, Alex. (2007). Evidence for Written Celtic from Roman Britain: A Linguistic Analysis of Tabellae Sulis 14 and 18. Studia Celtica 41, 31–45. (2012). Introduction: Multiple Languages, Multiple Identities. In Mullen & James (eds.), 1–35. (2013). Southern Gaul and the Mediterranean: Multilingualism and Multiple Identities in the Iron Age and Roman Periods. Cambridge: Cambridge University Press. Mullen, Alex & Patrick James (eds.). (2012). Multilingualism in the GraecoRoman Worlds. Cambridge: Cambridge University Press. Müller, Christel & Claire Hasenohr (eds.). (2002). Les italiens dans le monde grec: IIe siècle av. J.-C.–Ier siècle ap. J.-C.: circulation, activités, intégration. Actes de la table ronde, Ecole normale supérieur, Paris, 14–16 mai 1998 (Suppléments au Bulletin de Correspondance Hellénique, 41). Athens and Paris: École française d’Athènes and de Boccard. Musti, Domenico. (1984). La nozione storica di Sanniti nelle fonti greche e romane. In Gabriele Veneziale (ed.), Sannio: Pentri e Frentani dal VI al I sec. a.C. Atti del Convegno, 10–11 novembre 1980, 71–84. Campobasso: Enne. (2005). Magna Grecia: il quadro storico. Bari: Laterza. Nava, Maria Luisa & Paolo Poccetti. (2001). Il santuario lucano di Rossano di Vaglio. Una nuova dedica osca ad Ercole. Mélanges de l’École française de Rome, Antiquité 113, 95–122. Nervegna, Sebastiana. (2015). The Actors’ Repertoire, Fifth-Century Comedy and Early Tragic Revivals. Centre for Hellenic Studies Research Bulletin 3.2. (Available at http://nrs.harvard.edu/urn-3:hlnc.essay:NervegnaS.The_ Actors_Repertoire.2015).

321

References Noy, David. (1993–5). Jewish Inscriptions of Western Europe. Two volumes. Cambridge: Cambridge University Press. (1997). Writing in Tongues: The Use of Greek, Latin and Hebrew in Jewish Inscriptions from Roman Italy. Journal of Jewish Studies, 48, 300–11. (2000). Foreigners at Rome: Citizens and Strangers. London: Duckworth. Olson, Douglas S. (2007–12), Athenaeus: The Learned Banqueters. Eight volumes. Cambridge, MA: Harvard University Press. Oppert, Jules. (1866). L’Aryanisme, et de la trop grande part qu’on a fait sur son influence. Annales de Philosophie Chrétienne 13, 50–68. Orsi, P. (1916). Messana. La necropoli romana di S. Placido e di altre scoperte avvenute nel 1910–1915. Monumenti Antichi 24, 121–218. Ørsted, Peter. (1985). Roman Imperial Economy and Romanization: A Study in Roman Imperial Administration and the Public Lease System in the Danubian Provinces from the First to the Third Century A.D. Copenhagen: Museum Tusculanum Press. Osborne, M. J. & Byrne, S. G. (1994). A Lexicon of Greek Personal Names II. Attica. Oxford: Oxford University Press. Osborne, Robin. (1991). The Potential Mobility of Human Populations. Oxford Journal of Archaeology 10, 231–52. (2007). What Travelled with Greek Pottery? Mediterranean Historical Review 22, 85–95. Pagliara, Cosimo. (1983). Materiali iscritti arcaici del Salento (II). Annali della Scuola Normale Superiore di Pisa (Classe di Lettere e Filosofia) 13, 21–89. Palmer, Robert E. A. (1997). Rome and Carthage at Peace. Stuttgart: F. Steiner. Panciera, Silvio. (1997). L’evergetismo civico nelle iscrizioni latine d’età repubblicana. In Michel Christol & Olivier Masson (eds.), Actes du Xe congrès international d’épigraphie grecque et latine, Nîmes, 4–9 octobre 1992, 249–90. Paris: Publications de la Sorbonne. Pandolfini, Maristella & Aldo Luigi Prosdocimi. (eds.). (1990). Alfabetari e insegnamento della scrittura in Etruria e nell’Italia antica. Florence: L. S. Olschki. Pape, Wilhelm & Gustav Eduard Bensler. (1911). Wörterbuch der griechischen Eigennamen. Third edition. Braunschweig: Vieweg. Parlangèli, Oronzo. (1956). Le iscrizioni osche (mamertine) di Messina. Bollettino Centro di Studi Filologici e Linguistici Siciliani 4, 28–38. Pellegrino, C. & G. Colonna (2002). Pontecagnano. Studi Etruschi 65–68, 390–2. Perdrizet, Paul. (1902). Miscellanea. Revue des Études Anciennes 4, 85–9. Peremans, W. & E. Van ‘t Dack. (1972). Sur les rapports de Rome avec les Lagides. In Hildegard Temporini-Gräfin Vitzthum (ed.), Aufstieg und Niedergang der römischen Welt 1.1, 660–7. Berlin and New York: Walter de Gruyter. Peremans, W. et al. (1950–81). Prosopographia Ptolemaica. Leuven: Peeters. Peretz, Daniel. (2006). The Roman Interpreter and His Diplomatic and Military Roles. Historia: Zeitschrift für Alte Geschichte 55, 451–70.

322

References Peruzzi, E. (1970). Origini di Roma, I. La famiglia. Florence: Valmartina. Pescatori, Gabriella. (1968). Campania: Pontecagnano. Studi Etruschi 36, 226–7. Pestmann, W. (1981). A Guide to the Zenon Archive. Leiden: Brill. Pettigrew, David K. (2013). Connectivity. In Roger S. Bagnall et al. (eds.), Encyclopedia of Ancient History. Malden, MA, Oxford and Chichester: Wiley–Blackwell. Phang, Sara Elise. (2001). The Marriage of Roman Soldiers (13 B.C.–A.D. 235): Law and Family in the Imperial Army. Leiden: Brill. Pichler, W. (2008). Bericht über den aktuellen Stand der Erforschung und Erhaltung der libysch-berberischen Felsinschriften auf den Kanarischen Inseln. Almogaren 39, 117–36. Pickering, John. (1816). A Vocabulary or Collection of Words and Phrases Which Have Been Supposed to Be Peculiar to the United States of America. Boston: Cummings and Hilliard. Pitts, Lynn F. (1989). Relations between Rome and the German ‘Kings’ on the Middle Danube in the First to Fourth Centuries A.D. Journal of Roman Studies 79, 45–58. Pobjoy, Mark P. (2000). Building Inscriptions in Republican Italy: Euergetism, Responsibility, and Civic Virtue. In Alison E. Cooley (ed.), The Epigraphic Landscape of Roman Italy (Bulletin of the Institute of Classical Studies Supplement 73), 77–92. London: Institute of Classical Studies. Poccetti, Paolo. (1988). Per un’identità culturale dei Brettii. Naples: Istituto Universitario Orientale. Dipartimento di Studi del Mondo Classico e del Mediterraneo Antico. (2009). L’onomastica dell’Italia antica. Aspetti linguistici, storici, culturali, tipologici e classificatori (Collection de l’École française de Rome 413). Rome: École française de Rome. (2010). Contacts et échanges technologiques entre Grecs et indigènes en Italie méridionale. Langues et écritures au cours du IVe siècle av. J.-C. In Henry Tréziny (ed.), Grecs et indigènes de la Catalogne à la Mère Noire. Actes des rencontres du programme européen Ramses 2 (2006–2008), 659–77. Paris and Aix-en-Provence: Errance and Centre Camille Julian. (2015). Alphabet grec et langues indigènes de la Grande Grèce entre unité et variété. In Réjane Roure (ed.), Contacts et acculturations en Méditerranée occidentale. Hommages à Michel Bats. Actes du colloque de Hyères, 15–18 septembre 2011, 511–23. Paris and Aix-en-Provence: Édition Errance and Centre Camille Julian. (2016). Morire lontano dall’Italia: differenze e interazioni attraverso l’epigrafia ellenistica della necropoli dell’isola di Renea (Delo). In Haack (ed.), 405–33. (2017). La grecità nel contesto multilingue e multiculturale di Pompei e della Campania antica. In Massimo Osanna & Carlo Rescigno (eds.), Pompei e i Greci, 300–13. Milan: Electa. Polomé, E. (1983). The Linguistic Situation in the Western Provinces of the Roman Empire. In H. Temporini & W. Haase (eds.), Aufstieg und

323

References Niedergang der römischen Welt: Geschichte und Kultur Roms im Spiegel der neueren Forschung 29.2, 509–53. Berlin: Walter de Gruyter. Pontrandolfo, Angela & Agnès Rouveret. (1992). Le tombe dipinte di Paestum. Modena: Francesco Cosimo Panini. Prag, Jonathan. (2006). Poenus plane est – But who were the ‘Punickes’? Papers of the British School at Rome 74, 1–37. Preisigke, F. (1922). Namenbuch enthaltend alle griechischen, lateinischen, ägyptischen, hebräischen, arabischen und sonstigen semitischen und nichtsemitischen Menschennamen, soweit sie in griechischen Urkunden (Papyri, Ostraka, Inschriften, Mumienschildern usw.) Ägyptens sich vorfinden. Göttingen: Hubert and Co. Prosdocimi, Aldo L. (2002). Il genitivo singolare dei temi in -o- nelle varietà italiche (osco, sannita, umbro, sudpiceno, etc.). Incontri Linguistici 25, 65–76. Prosdocimi, Aldo, & Patrizia Solinas. (2006). Celticità linguistica in Italia prima del 400. Documenti e prospettive. In Daniele Vitali (ed.), Celtes et Gaulois. L’archéologie face à l’histoire. La préhistoire des Celtes, 401–50. Bibracte: Centre archéologique européen. Purcell, Nicholas. (1994). The City of Rome and the plebs urbana in the Late Republic. In J. A. Crook et al. (eds.), Cambridge Ancient History Volume IX. Second edition, 644–68. Cambridge: Cambridge University Press. (1995). On the Sacking of Carthage and Corinth. In Doreen C. Innes, Harry Hine & Christopher Pelling (eds.), Ethics and Rhetoric: Classical Essays for Donald Russell on His Seventy-Fifth Birthday, 133–48. Oxford: Clarendon. (1996). Rome and Its Development under Augustus and His Successors. In A. K. Bowman, E. Champlin & A. W. Lintott (eds.), Cambridge Ancient History Volume X. Second edition, 782–811. Cambridge: Cambridge University Press. (2000). Rome and Italy. In Alan K. Bowman, Peter Garnsey & Dominic Rathbone (eds.), Cambridge Ancient History Volume XI. Second edition, 405–43. Cambridge: Cambridge University Press. Rankov, Boris. (1987). The beneficiarii consularis in the Western Provinces of the Roman Empire. Unpublished D.Phil. Thesis, University of Oxford. (1999). The Governor’s Men: The officium consularis in Provincial Administration. In Goldsworthy & Haynes (eds.), 15–34. Rathbone, Dominic. (2008). Poor Peasants and Silent Sherds. In De Ligt & Northwood (eds.), 305–32. Rauh, Nicholas K. (1993). The Sacred Bonds of Commerce: Religion, Economy, and Trade Society at Hellenistic Roman Delos, 166–87 B.C. Amsterdam: J. C. Gieben. Rawson, Beryl. (1989). Spurii and the Roman View of Illegitimacy. Antichthon 23, 10–41. Ray, J. D. (2005). Demotic Papyri and Ostraca from Qasr Ibrim. London: Egypt Exploration Society.

324

References Reddé, Michel & Siegmar von Schnurbein. (2001). Alésia. Fouilles et recherches franco-allemandes sur les travaux militaires romains autour du MontAuxois (1991–1997). Tome II: Le matériel. Paris: de Boccard. Reich, David. (2018). Who We Are and How We Got Here. Ancient DNA and the New Science of the Human Past. Oxford: Oxford University Press. Reina, Placido. (1658). Delle notizie istoriche della citta di Messina. Prima parte. Messina: Brea. Renberg, G. H. (2006–7). Public and Private Places of Worship in the Cult of Asclepius at Rome. Memoirs of the American Academy in Rome 51–2, 87–172. Renfrew, Colin. (1987). Archaeology and Language: The Puzzle of IndoEuropean Origins. Cambridge: Cambridge University Press. Rescigno, Carlo. (2003). Documenti di vita cittadina. In Elena Laforgia (ed.), Il Museo Archeologico di Calatia, 43–88. Naples: Electa. Revell, Louise. (2005). The Roman Life Course: A View from the Inscriptions. European Journal of Archaeology 8, 43–63. Ricci, C. (1993). Egiziani a Roma. Aegyptus 73, 71–91. (2005). Orbis in urbe. Fenomeni migratori nella Roma imperiale. Rome: Quasar. Richlin, Amy. (2005). Rome and the Mysterious Orient: Three Plays by Plautus. Berkeley and London: University of California Press. (2017). Slave Theater in the Roman Republic. Cambridge: Cambridge University Press. Ricketts, L. M. (1980). The Administration of Ptolemaic Egypt under Cleopatra VII. Ph.D. Thesis, Minneapolis University. Ridgway, David. (1992). The First Western Greeks. Cambridge: Cambridge University Press. (2004). Phoenicians and Greeks in the West: A View from Pithekoussai. In Tsetskhladze & De Angelis (eds.), 35–46. Rigsby, Kent J. (1996). Graecolatina. Zeitschrift für Papyrologie und Epigraphik 113, 249–52. (2003). A Jewish Asylum in Greco-Roman Egypt. In Martin Dreher (ed.), Das antike Asyl. Kultische Grundlagen, rechtliche Ausgestaltung und politische Funktion (Akten der Gesellschaft für griechische und hellenistische Rechtsgeschichte 15), 127–42. Cologne / Weimar / Vienna: Böhlau. Riva, C. (2006). The Orientalizing Period in Etruria: Sophisticated Communities. In C. Riva & N. C. Vella (eds.), Debating Orientalization. Multidisciplinary Approaches to Change in the Ancient Mediterranean, 110–34. London: Equinox Publishing. Rix, Helmut. (1963). Das etruskische Cognomen. Wiesbaden: O. Harrassowitz. (1984). Etr. meχ rasnal = lat. res publica. Studi di antichità in onore di Guglielmo Maetzke, 455–68. Rome. (1996). Variazioni locali in osco. In Del Tutto Palma (ed.), 243–61. (1998). Eine neue frühsabellische Inschrift und der altitalische Präventiv. Historische Sprachforschung 111, 247–69.

325

References (2002). Handbuch der italischen Dialekte. Band V. Sabellische Texte: Die Texte des Oskischen, Umbrischen und Südpikenischen, Heidelberg: Carl Winter. (2005). Alphabete im vorrömischen Kampanien. In Thomas Ganschow, Matthias Steinhard, Dietrich Berges & Thomas Fröhlich (eds.), Otium. Festschrift für Volker Michael Strocka, 323–30. Remshalden: Greiner. (2009). Le relazioni tra onomastica e lessico nelle lingue antiche dell’Italia centrale. In Poccetti (ed.), 497–506. Robert, Louis & Jeanne Robert. (1976). Bulletin épigraphique. Revue des Études Grecques 89, 415–595. Robinson, D. M., C. G. Harcum & J. H. Iliffe. (1930). A Catalogue of the Greek Vases in the Royal Ontario Museum of Archaeology, Toronto. Toronto: University of Toronto Press. Robinson, Edward. (2004). Reception of Comic Theatre amongst the Indigenous South Italians. In Lesley Beaumont, Craig Barker & Elizabeth Bollen (eds.), Festschrift in Honour of J. Richard Green (= Mediterranean Archaeology 17), 193–212. (2014). Greek Theatre in Non-Greek Apulia. In Csapo, Goette, Green & Wilson (eds.), 319–32. Rochette, Bruno. (2011). Language Policies in the Roman Republic and Empire. In Clackson (ed.), 528–63. Roncalli, F. (2009). Numerali nell’antroponimia. In Poccetti (ed.), 507–13. Rostovtsev, Michail I. (1896). Un graffite de Délos. Bulletin de Correspondance Hellénique 20, 392. Roth, Jonathan. (1998). The Logistics of the Roman Army at War (264 B.C.–A.D. 235). Leiden: Brill. Roussel, Pierre. (1916). Les cultes égyptiens à Délos, du IIe au Ier siècle av. J.-C. Nancy: Berger-Levrault. (1931). La population de Délos à la fin du IIIe siècle avant J.-C. Bulletin de Correspondance Hellénique 55, 438–49. (1987). Délos, colonie athénienne. Second edition. Paris: de Boccard. Roussel, Pierre & Jean Hatzfeld. (1909). Fouilles de Délos, exécutées aux frais de M. le Duc de Loubat. Inscriptions. Bulletin de Correspondance Hellénique 33, 472–22. Rovai, Francesco. (2015a). I rapporti tra i codici in due repertori complessi dell’antichità: latino-gallico e latino-neopunico. In Carlo Consani (ed.), Contatto interlinguistico fra presente e passato, 197–216. Milan: Edizioni Universitarie di Lettere Economia Diritto. (2015b). Notes on the Inscriptions of Delos: The Greek Transliteration of Latin Names. Studi e Saggi Linguistici 53, 163–85. Roymans, Nico. (2004). Ethnic Identity and Imperial Power: The Batavians in the Early Roman Empire. Amsterdam: Amsterdam University Press. Ruozzi Sala, S. M. (1974). Lexicon nominum Semiticorum quae in papyris Graecis in Aegypto repertis ab anno 323 a. Ch. n. usque ad annum 70 p. Ch. n. laudata reperiuntur. Milan: Cisalpino-Goliardica.

326

References Russel, Paul with Alex Mullen. (2007). Celtic Personal Names of Roman Britain. Cambridge: Department of Anglo-Saxon, Norse and Celtic. (Available at www.asnc.cam.ac.uk/personalnames/). Russell, A. (2016). The Politics of Public Space in Republican Rome. Cambridge: Cambridge University Press. Russo, Mario. (2005). Sorrento. Una nuova iscrizione paleoitalica in alfabeto ‘nucerino’. Capri: Oebalus. Rutgers, L. V. (1995). The Jews in Late Ancient Rome: Evidence of Cultural Interaction in the Roman Diaspora. Leiden, Boston and Cologne: Brill. Rutter, Keith. (1979). Campanian Coinages: 475–380 B.C. Edinburgh: Edinburgh University Press. Rutter, Keith et al. (2001). Historia Nummorum. Italy. London: British Museum Press. Salmon, Edward Togo. (1967). Samnium and the Samnites. Cambridge: Cambridge University Press. Salomies, Olli. (2009). Nomi personali derivati da numerali a Roma etrusca? In Poccetti (ed.), 515–31. (2012). The Nomina of the Samnites. A Checklist. Arctos 46, 137–85. Salway, B. (1994). What’s in a Name? A Survey of Roman Onomastic Practice from c.700 B.C. to A.D. 700. Journal of Roman Studies 84, 124–45. Särstrom, Margit. (1940). A Study in the Coinage of the Mamertines. Lund: C. W. K. Gleerup. Sartre, Maurice. (2007). The Ambiguous Name: The Limitations of Cultural Identity in Graeco-Roman Onomastics. In Elaine Matthews (ed.), Old and New Worlds in Greek Onomastics, 199–234. Oxford: Oxford University Press. Sassatelli, G. (2003). Celti ed Etruschi nell’Etruria Padana e nell’Italia settentrionale. Ocnus 11, 231–57. Sassatelli, Giuseppe & Elisabetta Govi. (2013). Etruria on the Po and the Adriatic. In Turfa (ed.), 281–300. Scheidel, Walter. (2004). Human Mobility in Roman Italy, I: The Free Population. Journal of Roman Studies 94, 1–26. (2005). Human Mobility in Roman Italy, II: The Slave Population. Journal of Roman Studies 95, 64–79. (2006). The Demography of Roman State Formation in Italy. In Martin Jehne & Rene Pfeilschifter (eds.), Herrschaft ohne Integration? Rom und Italien in republikanischer Zeit, 207–26. Frankfurt am Main: Verlag Alte Geschichte. (2007). Marriage, Families, and Survival: Demographic Aspects. In Erdkamp (ed.), 417–34. Schmitt, Rüdiger. (2009). Die altpersischen Inschriften der Achaimeniden. Editio minor mit deutscher Übersetzung. Wiesbaden: Reichert. Schulze, W. (1904). Zur Geschichte lateinischer Eigennamen. Berlin: Weidmann.

327

References Scopacasa, Rafael. (2015). Ancient Samnium: Settlement, Culture, and Identity between History and Archaeology. Oxford: Oxford University Press. Scott, Michael. (2015). Temples and Sanctuaries. In Esther Eidinow & Julia Kindt (eds.), The Oxford Handbook of Ancient Greek Religion, 227–40. Oxford: Oxford University Press. Segal, Erich. (1987). Roman Laughter. The Comedy of Plautus. Second edition. Oxford: Oxford University Press. Seidl, Christian. (1994). Gemeinsabellisch und Vulgärlateinisch: Der Vokalismus. In George E. Dunkel, Gisela Meyer, Salvatore Scarlata & Christian Seidl (eds.), Früh-, Mittel-, Spätindogermanisch. Akten der IX. Fachtagung der indogermanischen Gesellschaft vom 5. bis 9. Oktober 1992 in Zürich, 349–70. Wiesbaden: Reichert. Shepherd, Gillian. (1999). Fibulae and Females: Intermarriage in the Western Greek Colonies and the Evidence from Cemeteries. In Gocha Tsetskhladze (ed.), Ancient Greeks West and East, 267–300. Leiden: Brill. Sherk, Robert K. (1969). Roman Documents from the Greek East. Baltimore: Johns Hopkins Press. Shumsky, Neil L. (2008). Noah Webster and the Invention of Immigration. New England Quarterly 81, 126–35. Siebert, Gérard. (1968). Sur l’histoire du sanctuaire des dieux syriens à Délos. Bulletin de Correspondance Hellénique 92, 359–74. Simkin, Oliver. (2012). Language Contact in the Pre-Roman and Roman Iberian Peninsula: Direct and Indirect Evidence. In Mullen & James (eds.), 77–105. Sims-Williams, Patrick. (1998). Genetics, Linguistics, and Prehistory: Thinking Big and Thinking Straight. Antiquity 72, 505–27. Skeat, T. C. (1954). The Reigns of the Ptolemies (Münchener Beiträge zur Papyrusforschung und antiken Rechtsgeschichte Heft 39). Munich: Beck. (1974). Greek Papyri in the British Museum, VII. The Zenon Archive. London: The British Museum. Skrandies, P. (2016). Language Policies and the Politics of Urban Multilingualism. In King & Carson (eds.), 115–46. Small, Alastair. (2006). Impressions of Ethnic Identity: Hellenistic Tile Stamps in South Italy. In Edward Herring et al. (eds.), Across Frontiers: Etruscans, Greeks, Phoenicians and Cypriots; Studies in Honour of David Ridgway and Francesca Romana Serra Ridgway, 327–37. London: Accordia Research Institute, University of London. (2014). Pots, Peoples and Places in Fourth-century BCE Apulia. In Carpenter, Lynch & Robinson (eds.), 13–32. Smith, Christopher J. (2006). The Roman Clan. The Gens from Ancient Ideology to Modern Anthropology. Cambridge and New York: Cambridge University Press. Solin, Heikki. (1982). Appunti sull’onomastica romana a Delo. In Heikki Solin, Filippo Coarelli & Domenico Musti (eds.), Delo e l’Italia, 101–17. Rome: Institutum Romanum Finlandiae.

328

References (1983). Juden und Syrer im westlichen Teil der römischen Welt. Eine ethnische-demographische Studie mit besonderer Beriicksichtigung der sprachlichen Zustande. Aufstieg und Niedergang der römischen Welt. 11.29.2, 587–789. (2001). Latin Cognomina in the Greek East. In O. Salomies (ed.), The Greek East in the Roman Context (Papers and Monographs of the Finnish Institute at Athens 7), 189–202. Helsinki: Suomen Ateenan-Instituutin Säätiö. (2007). Mobilità socio-geografica nell’Impero romano: Orientali in Occidente: considerazioni isagogiche. In M. Mayer Olivé, G. Baratta & A. Guzmán Almagro (eds.), XII Congressus internationalis epigraphiae graecae et latinae. Volume 10, 1363–79. Barcelona: Institut d’Estudis Catalans. (2009). Sulla nascita del cognome a Roma. In Poccetti (ed.), 251–93. Solmsen, Felix. (1898). Ναύκραρος ναύκλαρος ναύκληρος. Rheinisches Museum 53, 151–8. Sosin, Joshua D. et al. (2011). Checklist of Editions of Greek, Latin, Demotic, and Coptic Papyri, Ostraca, and Tablets. (Available at https://library.duke.edu /rubenstein/scriptorium/papyrus/texts/clist.html). Southern, Pat. (2007). The Roman Army: A Social and Institutional History. Oxford: Oxford University Press. Staccioli, Romolo A. (1968). Caere. Studi Etruschi 36, 249–50. Stek, Tesse D. (2009). Cult Places and Cultural Exchange in Republican Italy. Amsterdam: Amsterdam University Press. Stifter, David. (2009). Vernacular Celtic Writing Traditions in the East-Alpine Region in the Iron-Age Period? In Raimund Karl & Jutta Leskovar (eds.), Interpretierte Eisenzeiten. Fallstudien, Methoden, Theorie. Tagungsbeiträge der 3. Linzer Gespräche zur interpretativen Eisenzeitarchäologie. Linz: Oberösterreichisches Landesmuseum, 361–72. (2010). Neue Inschriften in norditalischer Schrift aus Österreich. In Robert Nedoma & David Stifter (eds.), *h2nr. Festschrift für Heiner Eichner. Wiesbaden: Harrassowitz, 233–240. (2012). Inscriptiones Pseudocelticae. Wrong and Premature Ascriptions of Inscriptions as Celtic. In Raimund Karl, Jutta Leskovar & Stefan Moser (eds.), Interpretierte Eisenzeiten. Die erfundenen Kelten – Mythologie eines Begriffes und seine Verwendung in Archäologie, Tourismus und Esoterik. Tagungsbeiträge der 4. Gespräche zur interpretativen Eisenzeitarchäologie (Studien zur Kulturgeschichte von Oberösterreich 31), 293–301. Linz: Oberösterreichisches Landesmuseum. Stoicheva, M. (2016). Urban Multilingualism: Bond or Barrier? In King & Carson (eds.), 85–113. Tacoma, Laurens E. (2013). Integration in Rome and in the Roman World. Proceedings of the Tenth Workshop of the International Network Impact of Empire (Lille, June 23–25, 2011). Leiden and Boston: Brill. (2016a). Moving Romans: Migration to Rome in the Principate. Oxford: Oxford University Press.

329

References (2016b). Bones, Stones and Monica: Isola Sacra Revisited. In Lo Cascio & Tacoma (eds.), 132–54. Tagliamonte, Gianluca. (1994). I figli di Marte. Mobilità, mercenari e mercenariato italici in Magna Grecia e Sicilia. Rome: Bretschneider. (2004). Processi di strutturazione e di autoidentificazione etnica. Il caso dei sanniti. In Domenico Caiazza (ed.), Safinim. Studi in onore di Adriano La Regina per il premio I Sanniti, 133–51. Piedimonte Matese: Italica Ars. Taplin, Oliver. (2007). Pots and Plays: Interactions between Tragedy and Greek Vase-Painting of the Fourth Century BC. Los Angeles: Getty Publications. Tcherikover. V. A. & A. Fuks. (1957). Corpus Papyrorum Judaicarum. Volume I. Cambridge, MA: Harvard University Press. Tcherikover. V. A., A. Fuks & M. Stern. (1964). Corpus Papyrorum Judaicarum. Volume III. Cambridge, MA: Harvard University Press. Teicher, Felix. (2013). From Aquileia to Carnuntum: Geographical Mobility along the Amber Road. Veleia 30, 47–73. Teodorsson, Sven-Tage. (1974). The Phonemic System of the Attic Dialect 400–340 B.C. Gothenburg: Acta Universitatis Gothoburgensis. (1977). The Phonology of Ptolemaic Koine. Gothenburg: Acta Universitatis Gothoburgensis. (1978). The Phonology of Attic in the Hellenistic Period. Gothenburg: Acta Universitatis Gothoburgensis. (2014). Attic. In Giannakis (ed.), Volume 1, 187–94. Thalmann, William G. (2011). Apollonius of Rhodes and the Spaces of Hellenism. Oxford and New York: Oxford University Press. Thomas, Yan. (1996). ‘Origine’ et ‘commune patrie’. Étude de droit public romain (89 av. J.-C. – 212 ap. J.-C.). Rome: École française de Rome. Thompson, Peter. (2003). ‘Judicious Neology’. The Imperative of Paternalism in Thomas Jefferson’s Linguistic Studies. Early American Studies: An Interdisciplinary Journal 1, 187–224. Threatte, Leslie. (1980). The Grammar of Attic Inscriptions: Volume 1, Phonology. Berlin: Walter de Gruyter. Tikkanen, Karin. (2011). Handbuch der italischen Dialekte. Band II. A Sabellian Case Grammar. Heidelberg: Carl Winter. (2017). On the Building of a Narrative: The Ver Sacrum Ritual of Ancient Italy. Mnemosyne 70.6, 958–76. Tod, M. N. (1951). Epigraphical Notes from the Ashmolean Museum. Journal of Hellenic Studies 71, 172–7. Torelli, Mario. (2011). Bellum in privatam curam (Liv. II, 49, 1). Eserciti gentilizi, sodalitates e isonomia in Etruria e nel Lazio arcaici. In Concetta Masseria & Donato Loscalzo (eds.), Miti di guerra, riti di pace, Proceedings of the Meeting (Torgiano-Perugia, 2009), Bari: Edipuglia, pp. 225–234. (2016). Il declino dei re. Tempi e modi delle trasformazioni istituzionali in Etruria meridionale e Roma. Science dell’ Antichità. Storia Archeologia 21, 5–20.

330

References (2017). Prosopografia etrusca, prosopografia romana. Il caso della defixio di Mamilius Limetanus da Caere. In Giovanni A. Cecconi, Andrea Raggi, Elena Salomone Gaggero (eds.), Epigrafia e società dell’Etruria romana. Proceedings of the conference, (Florence, 23–24 October 2015), 191–204. Rome: Quasar. Tozzi, G. (2012). Bilingual Inscriptions of Rome and Their Digital Edition in EDR (Epigraphic Database Roma). Poster presented at the 143rd Meeting of the American Philological Association (5–9 June 2012), Philadelphia, PA. (Available at www.edr-edr.it/it/Documenti_it.php). Tréheux, Jacques. (1952). Études d’épigraphie délienne. Bulletin de Correspondance Hellénique 76, 562–95. Trendall, A. D. (1989). Red Figure Vases of South Italy and Sicily. Cambridge: Cambridge University Press. Triantafillis, Elena. (2014). Osservazioni sull’iscrizione italica ‘di Niumsis Tanunis’: tra ermeneutica e linguistica. Padua Working Papers in Linguistics 6, 3–44. Trümper, Monika. (2014). The Honorific Practice of the ‘Agora of the Italians’ in Delos. In Jochen Griesbach (ed.), Polis und Porträt. Standbilder als Medien der öffentlichen Repräsentation im hellenistischen Osten, 69–187. Wiesbaden: Reichert. Tsetskhladze, Gocha & Franco De Angelis (eds.). (2004). The Archaeology of Greek Colonisation. Essays Dedicated to Sir John Boardman. Second edition. Oxford: Oxford University School of Archaeology. Tuck, A. & R. E. Wallace. (2013). First Words. The Archaeology of Language at Poggio Civitate. Altona, Canada: Friesens Press. Turfa, Jean MacIntosh (ed.). (2013). The Etruscan World. Abingdon and New York: Routledge. Uden, J. (2015). The Invisible Satirist: Juvenal and Second Century Rome. Oxford: Oxford University Press. Uhlich, Jürgen (1999). Zur sprachlichen Einordnung des Lepontischen. In Stefan Zimmer, Rolf Ködderitzsch & Arndt Wigger (eds.), Akten des zweiten deutschen Keltologen-Symposiums (Bonn, 2.-4. April 1997), 277–304. Tübingen: Max Niemeyer. (2007). More on the linguistic classification of Lepontic. In Pierre-Yves Lambert & Georges-Jean Pinault (eds.), Gaulois et Celtique Continental, 373–405. Genève: Droz. Ulbrich, Hans-Joachim. (2015–16). Script-Mixing on Ancient Fuerteventura and Lanzarote. Almogaren 46–7, 69–86. Untermann, Jürgen. (1975–97). Monumenta Linguarum Hispanicarum. Four volumes. Wiesbaden: Reichert. (2000). Wörterbuch des Oskisch-Umbrischen. Heidelberg: Carl Winter. Väänänen, Veikko. (1967). Introduction au latin vulgaire. Paris: Klincksieck. Van Berchem, Denis. (1959–60). Hercule-Melqart à l’Ara Maxima. Atti della Pontificia Accademia Romana di Archeologia (Serie III): Rendiconti 32, 61–8.

331

References (1967). Sanctuaires d’Hercule Melqart. Contribution à l’étude de l’expansion phénicienne en Méditerranée. Syria 44, 73–109 and 307–38. Van der Meer, L. B. (1987). The Bronze Liver of Piacenza: Analysis of a Polytheistic Structure. Amsterdam: Gieben. Van Heems, Gilles. (2011). Essai de dialectologie étrusque: problèmes théoriques et applications pratiques. In Gilles van Heems (ed.), La variation linguistique dans les langues de l’Italie préromaine, 69–90. Lyon: Maison de l’Orient et de la Méditerranée. Venuti, Lawrence. (2008). The Translator’s Invisibility: A History of Translation. Second edition. London: Routledge. Verboven, Konraad. (2007). Good for Business: The Roman Army and the Emergence of a ‘Business Class’ in the Northwestern Provinces of the Roman Empire (1st Century BCE – 3rd Century CE). In Lukas De Blois & Elio Lo Cascio (eds.), The Impact of the Roman Army (200 BC – AD 476). Economic, Social, Political, Religious and Cultural Aspects, 295–314. Leiden and Boston: Brill. Versluys, Miguel John. (2015). Roman Visual Material Culture as Globalising Koine. In Martin Pitts & Miguel John Versluys (eds.), Globalisation and the Roman World, 141–74. Cambridge: Cambridge University Press. Vetter, Emil. (1953). Handbuch der italischen Dialekte. Band I. Texte mit Erklärung, Glossen, Wörterverzeichnis. Heidelberg: Carl Winter. Vine, Brent. (1993). Studies in Archaic Latin Inscriptions. Innsbruck: Institut für Sprachwissenschaft. (1998). Remarks on the Archaic Latin ‘Garigliano Bowl’ Inscription. Zeitschrift für Papyrologie und Epigraphik 121, 257–62. Vineis, Edoardo. (1984). Problemi di ricostruzione della fonologia del latino volgare. In Edoardo Vineis (ed.), Latino volgare, latino medioevale, lingue romanze, 45–62. Pisa: Giardini. Visy, Zsolt. (1995). Some Notes on the Defence System of Pannonia in the 2nd and 3rd Centuries A.D. In Gábor Hajnóczi (eds.), La Pannonia e l’impero Romano, 85–96. Milan: Electa. (2003). The Roman Army in Pannonia: An Archaeological Guide of the Ripa Pannonica. Budapest: Teleki László Foundation. Von Prott, Hans & Walter Kolbe. (1902). Die Arbeiten zu Pergamon: Inschriften. Mitteilungen des Deutschen Archäologischen Instituts, Athenische Abteilung 27, 44–151. Wachter, Rudolf. (2001). Non-Attic Greek Vase Inscriptions. Oxford: Oxford University Press. Wagner, Esther-Miriam, Ben Outhwaite & Bettina Beinhoff (2013). Scribes and Language Change. In Wagner et al. (eds.), 3–18. Wagner, Esther-Miriam, Ben Outhwaite & Bettina Beinhoff (eds.), (2013). Scribes as Agents of Language Change. Boston: De Gruyter Mouton. Wallace, Rex. (2007). The Sabellic Languages of Ancient Italy. Munich: Lincom.

332

References (2008). Zikh Rasna: A Manual of the Etruscan Language and Inscriptions. Ann Arbor: Beech Stave Press. Wallace-Hadrill, A. (2017). Wallace-Hadrill, A. (2017). Imperial Rome: A City of Immigrants? Acta ad archaeologiam et artium historiam pertinentia 49: 53–72. Weinreich, Uriel. (1966). Languages in Contact: Findings and Problems. The Hague: Mouton. Wenger, Etienne. (1998). Communities of Practice: Learning, Meaning, and Identity. Cambridge: Cambridge University Press. Wesch–Klein, Gabriele. (2007). Recruits and Veterans. In Erdkamp (ed.), 435–50. Wheeler, Everett. (2000). Legio XVApollinaris: From Carnuntum to Satala – and beyond. In Yann Le Bohec & Catherine Wolff (eds.), Les Légions de Rome sous le Haut–Empire, 259–308. Lyon: de Boccard. Whitney, William Dwight. (1873). Oriental and Linguistic Studies: The Veda; the Avesta; the Science of Language. New York: Scribner, Armstrong, and Co. Whittaker, C. R. (1994). Frontiers of the Roman Empire: A Social and Economic Study. Baltimore, MD: Johns Hopkins University Press. (2004). Rome and Its Frontiers: The Dynamics of Empire. London and New York: Routledge. Wilcken, U. (1925). Punt-Fahrten in der Ptolemäerzeit. Zeitschrift für Ägyptische Sprache und Altertumskunde 60, 86–102. Wilkes, J. J. (1983). Romans, Dacians and Sarmatians in the First and Early Second Centuries. In Hartley & Wacher (eds.), 255–89. (1999). The Roman Army as a Community in the Danube Lands: The Case of the Seventh Legion. In Goldsworthy & Haynes (eds.), 95–104. (2005). The Roman Danube: An Archaeological Survey. Journal of Roman Studies 95, 124–225. Wilson, Alan J. N. (1966). Emigration from Italy in the Republican Age of Rome. Manchester: Manchester University Press. Wingens, Matthias, Helga de Valk, Michael Windzio & Can Aybek. (2011). The Sociological Life Course Approach and Research on Migration and Integration. In Matthias Wingens, Helga de Valk, Michael Windzio & Can Aybek (eds.), A Life-Course Perspective on Migration and Integration, 1–26. Dordrecht: Springer. Wiotte-Franz, Claudia. (2001). Hermeneus und Interpres: Zum Dolmetscherwesen in der Antike. Saarbrücken: Saarbrücker Druckerei und Verlag. Wiseman, Peter. T. (1967). Lucius Memmius and His Family. Classical Quarterly 17, 164–7. (2009). Remembering the Roman People: Essays on Late-Republican Politics and Literature. Oxford and New York: Oxford University Press. Wonder, John W. (2002). What Happened to the Greeks in Lucanian-Occupied Paestum? Multiculturalism in Southern Italy. Phoenix 56, 40–55. (2012). The Italiote League: South Italian Alliances of the Fifth and Fourth Centuries BC. Classical Antiquity 31, 128–51.

333

References Woolf, G. (2016). Migration and Social Differentiation in Roman Cities. (Unpublished conference paper). Wuthnow, Heinz. (1930). Die semitischen Menschennamen in griechischen Inschriften und Papyri des vorderen Orients. Leipzig: Dieterich. Zadorojnyi, A. (2010). Transcripts of Dissent? Political Graffiti and Elite Ideology Under the Principate. In J. A. Baird & C. Taylor (eds.), Ancient Graffiti in Context, 110–33. London and New York: Routledge. Zair, Nicholas. (2013). Individualism in Osco-Greek Orthography. In Wagner, Outhwaite & Beinhoff (eds.), 217–26. (2014). The Treatment(s) of *-u- after a Coronal in Oscan: Dialect Variation and Chronology. Indo-European Linguistics 2, 112–25. (2016). Oscan in the Greek Alphabet. Cambridge: Cambridge University Press. Zanovello, P. (1981). I due ‘betili’ di Malta e le ambrosiai petrai di Tiro. Rivista di Archeologia 5, 16–29. Zarmakoupi, Mantha. (2013). The City of Late Hellenistic Delos and the Integration of Economic Activities in the Domestic Sphere. Center for Hellenic Studies Research Bulletin 1. (Available at www.chs-fellows.org /2013/10/25/the-city-of-late-hellenistic-delos/).

334

INDEX LOCORUM

Literary Sources Ammianus Marcellinus, Rerum Gestarum 15.2.4, 71 Appian BC 1.111, 55 Mith. 5.28, 172 Mith. 22–4, 55 Pr. 51, 25 Pun. 92, 68 Aristotle, Politika 13031a, 108 Athenaius, Deipnosophistai 1.20b, 268, 272 3.121.b–f, 293 Cassius Dio, Romaike Historia 9.40.8, 156 53.12.4, 254 75.2.6, 220 Cato, de Agri Cultura Introduction 1–5, 63 1.4, 63 14.1, 200 18.9, 67 85, 67 Cicero Agr. 2.80, 56 Att. 1.17.9, 56 Cat. M. 12.39–41, 118 De or. 1.249, 68 Fam. 13.54, 218 Fam. 132, ,218 Flac. 19, 56

Leg. 3.11, 199 Man. 15, 56 Mur. 75, 67 Rep. 2.19, 103 Ver. 2.2.44, 218 Ver. 2.3.6, 56 Ver. 2.3.85, 218 Ver. 2.4.58, 218 Columella, de Re Rustica 1.1.10, 68 1.1.13, 68 12.4.2, 68 Dionysius of Halicarnassus, Romaike Arkhaologia 2.49.4, 101 2.58, 25 3.46, 103 7.3, 103 7.5–6, 110 7.7–11, 110 Diodorus Siculus, Bibliotheke 5.62.1, 239 12.31.1, 111 12.80.7, 239 16.15, 148 16.45.8, 118 19.3, 148 21.18.1–3, 156 21.8, 149 22.1.2–3, 156 22.7.4, 156 Ennius, Andromache 83–4, 62 87, 62

335

Index Locorum Festus, de Significatu Verborum (pages in Lindsay 1913) 150, 110, 129, 156, 185 Eutropius, Breuiarium ab urbe condita 9.28, 238 10.2, 238 Aulus Gellius, Noctes Atticae 19.13.3, 292 Herodotus, Historie 1.165–7, 105 2.151 f, 10 Homeric Hymn to Delian Apollo 56–60, 171 Isidore, Origines 18.65, 188 John Lydus, de Mensibus 1.7, 238 4.95, 238 Justin, Epitoma Historicarum Philippicarum Pompei Trogi 20.1.4, 101 38.7, 56 Juvenal, Saturae 3.058–72, 271 3.062–3, 269 3.075–7, 292 3.077, 291 3.119, 293 Livy, ab Vrbe Condita 1.34, 103 4.37.1, 110 4.44.12, 110 7.19.4, 118 7.29.5–7, 118 8.22, 103 9.28.6, 118 9.3.4, 118 22.10.16, 200 28.28.6, 110 30–32, 111

336

35.40, 149 Periochae 58–9, 55 Lucilius, Saturae 671–2, 56 Marius Victorinus, Ars Grammatica 6.8, 186 Martial, Epigrammata 14.17, 188 Memnon of Heraclea 31.9, 55 Menander Asp. 240–45, 66 Dis 102–7, 60 Pausanias, Hellados Periegesis 5.25.4, 92 5.27.8, 91 Philo, Legum Allegoriae 155, 272 Plautus As. 199, 67 As. 240–43, 53, 73 As. 417, 70 As. 464, 70 As. 495, 69 As. 495–6, 71 As. 904, 188 Aul. 406–7, 71, 73 Bac. 21–2, 66 Bac. 185–7, 60 Bac. 251–3, 70 Bac. 932, 62 Bac. 1009, 70 Capt. 145, 71 Capt. 161, 71 Capt. 449, 69 Capt. 884, 66 Cas. 90–112, 63 Cist. 143, 70 Cist. 177,70 Cur. 280, 71

Index Locorum Cur. 354–60, 188 Cur. 462–84, 71 Epid. 1–10, 60 Epid. 342, 70 Epid. 342–3, 59 Men. 11–12, 58 Men. 72–6, 63 Men. 117–19, 53, 73 Men. 287, 69 Men. 335, 71 Men. 373, 71 Men. 495, 71 Men. 724, 70 Men. 1004–6, 69 Men. 1068–95, 64 Men. 1111, 62 Mer. 64–79, 63 Mer. 275–82, 63 Mil. 480–530, 69 Mil. 648, 66 Mos. 470–72, 70 Mos. 473, 69 Per. 135, 70 Per. 330, 66 Per. 473, 66 Per. 554–5, 72 Per. 595–648, 64 Per. 603, 70 Per. 690, 66 Poen. 46–9, 70 Poen. 71, 69 Poen. 94, 70 Poen. 175, 70 Poen. 599, 70 Poen. 656, 70 Poen. 975–1040, 66 Poen. 1003, 69 Poen. 1040–85, 64 Poen. 1403, 69 Ps. 1100, 59 Ps. 1232, 69, 70 Rud. 115, 71 Rud. 358–60, 189 Rud. 906–7, 72 Rud. 1130–53, 64 St. 480, 71 St. 480–90, 60 Trin. 720, 69 Trin. 795, 53, 73

Trin. 1084–5, 70 Trin. 1105–7, 53, 73 Truc. 178, 71 Truc. 690, 66 Truc. 691, 191 Truc. 955, 70 Pliny the Elder, Naturalis Historia 2.140, 35 3.70, 102 9.123, 188 18.22, 68 34.49, 92 36.91–3, 35 Plutarch Numa 1.3, 101 Numa 21, 25 Numa 21.1–2, 26 Tim. 16.3–4, 148 Julius Pollux, Onomasticon 9.100, 188 Polybius, Historiai 1.7.1–6, 156 1.7.2, 129 1.8.1, 129 36.7.4–5, 68 Probus, de Praenominibus 4.1, 24 Sallust, Bellum Jugurthinum 21.2–4, 68 26.1, 150 Scylax, Periplous 12–14, 238 Seneca (the Younger) Ep. 117.30, 188 Ep. 120.4, 292 Helv. 6.2–3, 280 Helv. 6.3, 282 Strabo, Geographika 5.2.2, 103

337

Index Locorum 5.4.03, 110 5.4.12, 101 5.4.13, 148 5.5.9, 102 6.1.01, 148 6.1.05, 148, 149 6.1.13, 108 6.2.3, 129, 156 10.5.4, 171 13.4.2, 55

45.1.6, 179 Varro, Res Rusticae 1.1.10, 68 1.10, 68 1.2.14, 191 1.52, 67 3.5.7, 71 3.7.3, 67 Valerius Maximus, Facta et Dicta Memorabilia 9.2.3, 55

Suetonius Aug. 43.1, 269 Aug. 71.2, 188 Cal. 44.2, 260 Claud. 16.2, 292 Gram. 22, 292 Iul. 39.1, 269

Velius Longus, de Orthographia 7.77.12–14, 198, 199 Velleius Paterculus, Historia Romana 2.110.5, 220

Tacitus, Germania 43.1, 220 Ulpian, Digesta seu Pandectae

Zosimus, Historia Noua 2.10.2, 238

Epigraphic and Papyrological Sources Aegyptische Urkunden aus den Königlichen (later Staatlichen) Museen zu Berlin, Griechische Urkunden (BGU) 2.455.11, 251 4.1113, 237, 238 6.1266.39, 238 6.1266.45, 238 6.1490.4, 260 6.166.39, 238 14.2386.6, 238 L’Année épigraphique (AE) 1897.33, 188 1899.159, 211 1910.15, 197 1929.119, 207 1929.200, 207 1929.219, 207 1929.222, 207 1949.168, 211

338

1967.364, 219 1967.366, 219 1969/1970.174, 254 1969/1970.478, 219 1978.635, 203 1982.709, 211 1983.756, 207 1996.1251, 210 1996.1252, 210 2001.1464, 215 2001.1797, 197 2002.1152, 207, 208 2005.12, 219 Arena (1994), Iscrizioni greche arcaiche di Sicilia e Magna Grecia (IGASMG) 3, App. 2.113.A1.b–c, 96 3.65, 91 Bagnall (1986), Papyri and Ostraka from Quseir al–Qadim (P.Quseir) 1, 10

Index Locorum Baillet (1920–6), Inscriptions grecques et latines des tombeaux des rois ou Syringes à Thèbes (Syringes) 205.2, 233 283, 244 385, 238 398.2, 233 588, 244 1092, 243 1830, 251 1949.2, 233 2003, 244 2004, 244 Bernand (1969), Les inscriptions grecques de Philae (I.Philae) 1.63, 234, 241, 248 2.146, 252 Bernand (1970), Le Delta égyptien d’après les textes grecques, I (I.Delta) 1.892, 1.3, 240 Bernand (1989), De Thèbes à Syène (I.Syène) 308.1–2, 233 321, 230, 233, 244, 247, 249, 252 323, 233 Bernand (1999), Inscriptions grecques d’Hermoupolis Magna et de sa nécropole (I.Hermoupolis) 4.2.20/92, 254 4.2.40/112, 254 5, 254 5.137, 241, 258 6, 240 6.156, 240 6.165, 240 6.198, 240 35.1, 238 Bowman & Thomas (1994–2003), The Vindolanda Writing-Tablets (Tab. Vindol.) 2.213, 212

Collingwood, Wright, Tomlin & Hassall (1965–2009), Roman Inscriptions of Britain (RIB) 706, 7 Corpus Inscriptionum Graecarum (CIG) 3.4807 g, 244 Corpus Inscriptionum Latinarum (CIL) 12 fasti Iul. 14(15) –19, 199 Nov. 18–20, 199 Sept. 20–23, 199 12.60, 191 12.234, 191 12.401, 198, 199 12.553, 198 12.561, 191 12.563, 199 12.564, 198 12.579, 199 12.585, 199 12.709, 14, 19 12.992, 199 12.1563, 167 12.1618, 199 12.1834, 191 12.1920, 198 12.2103, 7 12.2214, 191 12.2233, 197 12.2239, 198 12.2240, 198 12.2242, 199 12.2243, 199 12.2295, 198 12.2504, 198 12.2658, 191 12.2890, 199 12.3437b, 188 2.4085, 238 2.10988, 213 3.68, 244 3.3377, 224 3.3510, 219 3.3644, 224 3.4455, 207 3.4463, 207 3.4482, 207

339

Index Locorum 3.5272, 224 3.5308, 211 3.5816, 211 3.10315, 219 3.10352, 224 3.10505, 213 3.10576, 224 3.10988, 219 3.11129, 207 3.14203, 188 3.14349.5, 214 3.14358, 207 3.14412.4, 254 3.14507, 214, 225 3.14927, 211 4.3867, 188 4.3867, 190 5.3164, 254 6.9664, 211 6.32567, 7, 284 6.34196, 284 8.217, 208 10.2081, 254 10.3572, 254 10.3573, 254 11.4687, 7 12.1711, 245 13.6677, 211 13.8773, 214 Couilloud (1974), Les monuments funéraires de Rhénée 45, 185 52, 185 75, 196 131, 185 150, 185 154, 185 155, 185, 282 243, 196 318, 185 420, 185 438, 185 485.2, 185 492, 186 Crawford (1955), Papyri Michaelidae (P.Michael.) 7r3–4, 260

340

Crawford (1996), Roman Statutes 12, 56 25.095, 72 25.126, 72 Crawford et al. (2011), Imagines Italicae Aeclanum 1, 161 Aeclanum 3, 160 Aeclanum 4, 167 Allifae 1, 112, 113 Allifae 2, 113 Anagnia 13, 113 Anxia 1, 135, 142, 145 Aquilonia 1, 168 Atella 2, 106, 115 Atina Lucana 1, 166, 167 Bantia 2, 143, 161 Blanda 1, 101 Bouianum 116, 119 Buxentum 1, 127, 130, 135, 142, 143, 144, 146, 147 Buxentum 3, 68, 131, 145 Campania or Lucania or Brettii or Sicilia 1, 78 Capua 1, 112 Capua 26, 160 Capua 35, 90, 101 Capua 36, 116, 119 Capua 45, 90 Capua 48, 116, 119 Casinum 1, 101 Caudium 2, 80, 116, 119 Caudium 3, 116, 119 Caulonia 2, 131, 143, 149, 150, 161 Caulonia 5, 87, 160 Cosilinum 1, 133, 166 Cosilinum 2, 127, 145, 146 Crimisa 1, 136, 141, 166 Crimisa 2, 136, 144, 145, 166 Crimisa 3, 135, 160 Cubulteria 1, 114 Cumae 1, 112 Cumae 4, 143 Cumae 8, 80, 165 Cumae 11, 80 Cumae 17, 160 Fenserni 1, 112, 113, 168 Herculaneum 1, 161 Histonium 5, 143, 161

Index Locorum Histonium 7, 161 Histonium 8, 161 Interpromium A, 10 Interpromium B, 10 Laos 2, 81, 131, 141, 142, 145, 150 Laos 3, 80, 132, 145 Larinum 1, 115 Lucani 1, 137, 144, 168 Lucania 1, 128 Lucania or Brettii or Sicilia 1, 138, 143 Lucania or Brettii or Sicilia 2, 128, 143 Lucania or Brettii or Sicilia 3, 128, 131, 142 Messana 1, 140, 157, 168 Messana 2, 140, 157 Messana 3, 157 Messana 4, 140, 157, 161, 164, 165 Messana 5, 140, 157, 161, 162, 164, 165 Messana 6, 140, 157, 165, 169 Messana 7, 140, 157, 158, 164 Messana 8, 165 Metapontum 1, 128, 131, 142, 145 Minturnae 1, 101, 108, 113 Nola 2, 88 Nola 3, 88, 158 Nola 6, 106 Nola 7, 106 Nola 8, 106, 107, 115 Nola 9, 106 Nola 11, 80, 101, 106, 113 Nola 12, 106 Nuceria Alfaterna 3, 105, 114 Nuceria Alfaterna 5, 160 Nuceria Alfaterna 6, 107 Nuceria Alfaterna 7, 107, 115 Nuceria Alfaterna 9, 160 Numistro 1, 134, 160, 166 Paestum 1, 132, 143 Paestum 2, 132, 143 Paestum 3, 81 Petelia 2, 80, 127, 130, 132, 142, 143, 144, 145, 146, 147, 148, 160 Phistelia 1, 112, 168 Picentia 1, 107 Picentia 2, 109 Pompei 1, 170 Pompei 2, 167 Pompei 3, 167 Pompei 4, 167 Pompei 8, 166

Pompei 11, 160, 166 Pompei 13, 167 Pompei 19, 167 Pompei 20, 166 Pompei 21, 166 Pompei 23, 166 Pompei 24, 160, 167 Pompei 81, 115 Pompei 90, 17 Pompei 96, 80 Potentia 1, 127, 139, 145, 146, 147, 148, 166, 167 Potentia 3, 166 Potentia 9, 143, 161, 166 Potentia 10, 143, 161, 166 Potentia 11, 138, 143, 144 Potentia 12, 130, 135, 143, 144 Potentia 13, 133, 143 Potentia 15, 136 Potentia 16, 138 Potentia 17, 130, 134 Potentia 18, 137 Potentia 19, 134, 143 Potentia 20, 134 Potentia 21, 137 Potentia 22, 138 Potentia 23, 145, 146 Potentia 24, 133, 142 Potentia 25, 138, 146 Potentia 31, 136 Potentia 39, 151, 166 Potentia 40, 127, 137, 142, 144, 146, 147 Potentia 44, 137, 141 Salernum 1, 107 Salernum 2, 107 Salernum 3, 108 Saticula 5, 117 Stabiae 1, 106 Stabiae 2, 106 Surrentum 2, 105, 114 Surrentum 3, 105, 114 Teanum Sidicinum 1, 168 Teanum Sidicinum 25, 85 Tegianum 1, 82 Teruentum 12, 167 Teruentum 15, 166 Teruentum 17, 166 Teruentum 21, 166 Teruentum 33, 166

341

Index Locorum Teruentum 34, 143 Teruentum 36, 166 Teuranus Ager 1, 138 Teuranus Ager 2, 87 Teuranus Ager 3, 87 Teuranus Ager 4, 87 Thurii Copia 1, 132, 141, 145 Umbria 3, 9 Vibo 2, 133, 143, 149 Vibo 5, 137, 143, 144, 149 Vibo 6, 137, 143, 149 Vibo 7, 134, 141, 143, 144 Vibo 8, 137, 143, 149 Cuvigny (2012), Didymoi. Une garnison romaine dans le désert oriental d’Égypte (O.Did.) 208, 254 D’Amore (2007), Iscrizioni greche d’Italia: Reggio Calabria (I.Rhegion) 19, 93 26b, 93 66, 150 Del Monaco (2013a), Iscrizioni greche d’Italia: Locri (I.Locri) 1.8, 150 2.01, 150 2.6, 150 3.6, 150 4.08, 150 4.11, 150 5.4.7–8, 150 5.6, 150 6.2, 150 7.1.9, 150 7.11, 150 10.1, 150 11.1, 150 12.1, 150 15.1, 150 15.9, 150 16.4, 150 16.7, 150 17.1, 150 18.1, 150 19.1, 150 20.1, 150 20.17, 150

342

20.20, 150 20.22, 150 20.23, 150 24.5, 150 25.10, 150 26.1, 68 27.6, 150 28.4, 150 29.1, 150 35.8, 150 37.6, 150 46, 150 56, 150 63, 150 64, 150 65, 150 66, 91 67, 150 69, 150 81, 150 89.05, 150 89.06, 150 89.13, 150 89.15, 150 89.16, 150 90.11, 150 91, 150 91.2, 150 91.3, 150 108, 150 Degrassi (1957), Inscriptiones latinae liberae rei publicae (ILLRP) 53, 197 93b, 191 174, 285, 286 175, 285, 286 176, 285, 286 177, 285, 286 178, 285, 286 179, 285, 286 180, 285, 286 181, 285, 286 515, 14, 19 De Simone & Marchesini (2002), Monumenta linguae messapicae (MLM) 4 Gn, 80

Index Locorum 7 Cae, 80 10 Ar, 83, 84 11 Ar, 87 11 Ro, 80 12 Ruv, 83 31 Al, 80 Dessau (1892–1916), Inscriptiones latinae selectae (ILS) 2658, 208 Dittenberger (1903–5), Orientis graeci inscriptiones selectae (OGI) 1.129, 234 1.129.11–12, 263 1.133, 260 1.133.5, 260 1.135.1–3, 261 1.196, 248 Dittenberger & Purgold (1896), Die Inschriften von Olympia (IvO) 271, 91 Dittenberger et al. (1915–24), Sylloge inscriptionum graecarum (Syll.³) 1, 259 585.17–18, 245 Dubois (1995–2002), Inscriptions grecques dialectales de Grande–Grèce (IGDGG) 1.11b, 96 1.36, 91 1.39c, 93 Dürrbach et al. (1926–72), Inscriptions de Délos (ID) 442, 186 443, 186, 195 1401.c.09, 184 1401.c.10, 184 1401.c.18, 184 1416.A.2.15, 184 1417, 186 1417.B.1.149, 184 1427, 196 1429, 186 1430, 196 1432, 186, 196

1439, 195 1442.46, 184 1442.49, 184 1442.73, 184 1449, 196 1450, 196 1451, 196 1501.22, 184 1504, 184 1505, 184 1507.06, 184 1507.10, 184 1514.10, 184 1517.16, 184 1517.26, 184 1519, 174 1520, 174, 179 1520.30, 184 1520.43, 184 1526, 261 1527, 260 1527.6, 184 1534, 261 1620, 193 1620.2–3, 262 1644, 182 1645, 182 1646, 182 1650, 182 1651, 182 1652, 182 1683, 262 1688, 262 1689, 262 1690, 262 1691, 262 1692, 197 1695, 193 1695.1–2, 262 1696, 193 1696.2–3, 262 1698, 193 1698.1, 262 1699.1, 261 1711, 193 1712, 193 1713, 193 1714, 193 1715, 193

343

Index Locorum 1717, 179, 262 1718, 179, 262 1722, 179, 262 1725, 193 1727, 193, 262 1729, 193 1732, 196 1732.7, 196 1736, 262 1751, 186 1753, 196 1754, 196 1755, 196 1757, 262 1758.9, 196 1760, 197 1761.3, 196 1763.10, 196 1764, 197 1768, 197 1778, 174 1778.4, 179, 184 1779, 179 1927, 174 1955, 174 2037.1, 178 2041, 182 2044, 182 2072.4, 178 2073.6–7, 178 2124, 186 2224, 183 2227, 183 2228, 183 2245, 183 2248, 186 2250, 183 2251, 183 2255, 183 2256, 183 2346, 196 2351, 186 2365, 173 2432, 197 2440, 197 2454, 179 2457, 197 2534, 194 2549, 179

344

2593, 174 2595, 174 2595.38–9, 238 2597, 174 2598, 196 2611, 173 2612, 179, 193, 196, 197 2616, 196 2618, 196 2619, 173 2623.A.1.3, 185, 196 2628, 186 Duval et al. (1985–2002), Recueil des inscriptions gauloises (RIG) E5, 7 L97, 8 Edgar (1925–40), Zenon Papyri, Catalogue général des antiquités égyptiennes du Musée du Caire (P.Cair.Zen.) 1.59035.1, 255 1.59076.8, 237 1.59112.1, 255 4.59643.25, 255 El–Khouly (1973), Excavations East of the Serapeum at Saqqâra (P.Saqqara) 60, 252 Frösén et al. (1986), Papyri Helsingienses I, Ptolemäische Urkunden (P.Hels.) 1.29.i.4, 239 Gallazzi, Pintaudi & Worp (1986), Ostraka greci del Museo Egizio del Cairo (O. Cair.) 1.2, 239 Greek Papyri in the British Museum (P.Lond.) 7.1986, 253 7.2136 fr. 1.1, 256 Griechische Papyrusurkunden der Hamburger Staats– und Universitätsbibliothek (P.Hamb.) 1.57.19, 239

Index Locorum Griffith (1935–7), Les Temples immergés de la Nubie. Catalogue of the Demotic Graffiti of the Dodecaschoenus (Graff. Dodec.Philae) 105.3, 253 121.2, 253 497.1, 252 Guarducci (1935–50), Inscriptiones Creticae (IC) 4.18.4, 254 4.162.4, 245 4.215, 254 Guéraud (1942), Ostraca grecs et latins de l’Wâdi Fawâkhir (O.Wâdi Fawâkhir) 3.14, 245 Hansen (1983–9), Carmina Epigraphica Graeca (CEG) 388, 91 The Hibeh Papyri (P.Hib.) 1.42, 238 Inschriften Mysia and Troas (IMT) 717.13–14, 241 Inscriptiones Graecae (IG) 9.1.2.3, 258 9.1.2.638, 258 9.2.10, 258 11.2.135.26, 184 11.4.1227.1, 183 11.4.1228.2, 183 11.4.1229.2, 183 12.4.2.551, 183 14.0597, 161 14.0647, 161 14.1440, 290 Maras (2009), Il dono votivo Py do.6, 37 Ru do.1, 30 Meiggs & Lewis (1969), A selection of Greek historical inscriptions: to the end of the fifth century B.C. (GHI) 7 (a–g), 10

Meiser (2014), Etruskische Texte (ET) Ad 2.1, 45 Af 3.1, 40 AH 1.42, 51 AH 2.1, 45 AH 2.2, 49 AS 1.217, 47 AS 1.276, 47 AS 2.2, 49 AS 2.12, 47 AS 2.14, 39 AS 2.15, 40 AS 3.6, 31, 46 AT 0.1, 47 AT 1.29, 51 AT 1.48, 51 AT 1.64, 51 AT 2.12, 34 AT 3.2, 49 AV 1.29, 39 Cl 1.362, 49 Cl 1.363, 49 Cl 1.364, 49 Cl 1.365, 49 Cl 1.366, 49 Cl 1.576, 48 Cl 1.577, 48 Cl 1.1645, 45 Cl 1.2455, 44 Cl 1.2456, 44 Cl 1.2457, 44 Cl 1.2458, 44 Cl 1.2459, 44 Cl 1.2460, 44 Cl 1.2461, 44 Cl 1.2462, 44 Cl 2.01, 28 Cl 2.03, 49 Cl 2.44, 31 Cl 3.2, 28 Cl 7.5, 44 Cm 2.01, 104 Cm 2.018, 46, 116 Cm 2.073, 116 Cm 2.104, 44 Cm 2.120, 44 Cm 2.138, 44 Cm 6.1, 104 Cm 6.2, 104

345

Index Locorum Cr 0.1, 31 Cr 1.032, 43 Cr 1.130, 32 Cr 1.131, 32 Cr 1.132, 32 Cr 1.133, 32 Cr 1.134, 32 Cr 1.135, 32 Cr 1.136, 32 Cr 1.137, 32 Cr 1.138, 32 Cr 1.139, 32 Cr 1.159, 45 Cr 1.161, 45 Cr 2.1, 103 Cr 2.4, 28, 103 Cr 2.7, 47 Cr 2.18, 28 Cr 2.19, 28 Cr 2.20, 28 Cr 2.30, 34 Cr 2.112, 48 Cr 2.115, 34 Cr 2.131, 48 Cr 2.149, 38 Cr 2.194, 46 Cr 2.195, 45 Cr 3.20, 28, 34 Cr 3.22, 50 Cr 3.27, 31, 46 Cr 3.43, 37 Cr 5.2, 45 Cr 5.8, 48 Cs 2.1, 13 Cs 2.2, 13 Cs 2.3, 13 Cs 2.4, 45 Fa 0.17, 45 Fa 2.6, 48 Fa 2.7, 48 Fa 2.8, 48 Fa 2.9, 48 Fa 2.10, 48 Fa 2.20, 43 Fa 3.2, 30 Fa 3.4, 43 Fa 3.5, 32 Fa 7.1, 35 Fe 3.3, 48

346

Gr 3.1, 13 La 2.3, 40 Na 2.1, 13 Na 2.2, 13 Na 2.3, 13 Na 2.4, 13 OA 2.2, 32 Pa 0.3, 50 Pa 4.2, 14 Pe 1.24, 35 Pe 3.3, 38 Po 2.43, 50 Po 2.44, 50 Ru 3.1, 30 Ru 9.1, 115 Sp 2.84, 35 Ta 1.17, 33 Ta 2.3, 49 Ta 3.1, 103 Ta 3.2, 39 Ta 3.5, 45 Ta 6.1, 32 Ta 7.5, 50 Ta 7.13, 32, 50 Ta 7.19, 32 Ta 7.25, 50 Vc 1.121, 51 Vc 2.1, 45 Vc 2.4, 48 Vc 2.7, 47 Vc 2.9, 32, 46 Vc 2.33, 48 Vc 2.34, 48 Vc 2.72, 49 Vc 3.2, 47 Vc 3.4, 47 Vc 3.5, 47 Ve 2.4, 48 Ve 2.8, 44 Ve 2.11, 39 Ve 3.7, 47 Ve 3.9, 32 Ve 3.15, 34 Ve 3.30, 32 Ve 3.40, 34 Vn 1.1, 29 Vn 3.2, 31, 46 Vs 1.18, 46 Vs 1.45, 39

Index Locorum Vs 1.59, 49 Vs 1.67, 39, 44 Vs 1.100, 32 Vs 1.113, 38 Vs 1.121, 46 Vs 1.122, 44 Vs 1.126, 46 Vs 1.165, 47 Vs 1.257, 46 Vs 3.8, 44 Vt 8.1, 51 Melaerts (2000), Papyri in Honorem Johannis Bingen Octogenarii (P.Bingen) 45, 263 Minon (2007), Les inscriptions éléennes dialectales 62, 207 Mitteis and Wilcken (1912), Grundzüge und Chrestomathie der Papyruskunde (W.Chr) 110a, 238 Moretti (1968–79), Inscriptiones Graecae urbis Romae (IGUR) 117, 288 119, 288 291, 264 1287, 290 Mosser (2003), Die Steindenkmäler der legio XV Apollinaris 62, 91 68, 207 69, 207 80, 207 97, 207 104, 207, 208 116, 207 121, 207 167, 207 Noy (1993–5), Jewish Inscriptions of Western Europe (JIWE) 2.58, 284

The Oxyrhynchus Papyri (P.Oxy.) 22.2349.2, 251 22.2349.27, 251 44.3208, 236 Papiri greci e latini (PSI) 5.545.8, 255 6.585.1, 255 6.585.5, 255 Papyri Osloenses (P.Oslo) 2.26, 235 Peek (1955), Griechische Vers–Inschriften I, Grab–Epigramme (GVI) 1002.2, 236 Perdrizet & Lefebvre (1919), Les graffites grecs du Memnonion d’Abydos (Abydos) 5.1–2, 243 55, 243 87.3, 233 114, 237 357.2, 233 450bis, e, 243 535.1, 233 543.2, 233 627.5, 233 629.1, 233 Petolescu (1996–2000), Inscriptions de la Dacie romaine (IDRE) 2.272, 213 2.308, 214 Preisigke (1912–), Griechische Papyrus der Kaiserlichen Universitäts– und Landes–bibliothek zu Strassburg (P.Strasb.) 8.761v9, 238 Rix (2002), Sabellische Texte Cm 6, 88, 158 Cm 7, 88 Cm 9, 143 Cm 10, 161 Cm 14, 165 Cm 35, 160

347

Index Locorum Cp 9, 160 Fr 4, 161 Fr 5, 143, 161 Hi 1, 167 Hi 4, 160 Hi 6, 161 Lu 2, 166, 167 Lu 3, 133, 166 Lu 4, 134, 160, 166 Lu 5, 127, 139, 145, 146, 147, 148, 166, 167 Lu 6, 143, 161, 166 Lu 7, 143, 161, 166 Lu 8, 166 Lu 13, 127, 137, 142, 144, 146, 147 Lu 14, 132, 143 Lu 15, 130, 134 Lu 16, 133, 143 Lu 17, 130 Lu 18, 131, 142 Lu 21, 138, 146 Lu 23, 136, 141, 166 Lu 24, 136, 144, 145, 166 Lu 25, 133, 143, 149 Lu 26, 138 Lu 27, 130, 135, 143 Lu 28, 134 Lu 29, 137 Lu 30, 133, 142 Lu 31, 138 Lu 32, 138 Lu 33, 136 Lu 35, 138, 143, 144 Lu 36, 134, 143 Lu 37, 131, 142, 145 Lu 38, 143, 161 Lu 39, 135, 142, 145 Lu 40, 127, 145, 146 Lu 41, 82 Lu 43, 138 Lu 44, 135, 160 Lu 45, 68, 131, 145 Lu 46, 131, 141, 142, 145, 150 Lu 47, 132, 141, 145 Lu 59, 136 Lu 62, 127, 130, 135, 142, 143, 146, 147 Lu 63, 132, 145 Lu 64, 145, 146 Me 1, 140, 157, 161, 164, 165 Me 2, 140, 157, 161, 162, 164, 165

348

Me 3, 140, 157, 162 Me 4, 140, 157, 165, 169 Me 5, 140, 157, 158, 164 nLu 1, 137, 144, 168 nMe 1a, 140, 157, 168 nMe 1b, 140, 157 nMe 41b, 168 nSa 2a, 168 nSa 4, 168 nSa 6, 168 nSi 1a, 168 Po 1, 167 Po 3, 167 Po 4, 166 Po 6, 166 Po 7, 160, 166 Po 8, 167 Po 9, 166 Po 14, 166 Po 34, 167 Po 35, 167 Po 36, 167 Po 54, 170 Ps 18, 101 Sa 2, 166 Sa 8, 166 Sa 9, 166 Sa 11, 166 Sa 12, 166 Sa 13, 167 Sa 27, 119 Sp AP 2, 39 tLu 1, 137, 141 tLu 3, 137, 143, 144, 149 tLu 6, 137, 143, 149 tLu 7, 134, 141, 143, 144 tLu 8, 137, 143, 149 tMe 1, 140, 157 Um 4, 42 Rutter et al. (2001), Historia Nummorum. Italy (HNItaly) 459, 112 460, 112 461, 112 462, 112 463, 112 464, 112 476, 112

Index Locorum 477, 112 478, 112 531, 112 532, 112 533, 112 534, 112 535, 112 536, 112 537, 112 538, 112 611, 112 612, 112 624, 115 Sammelbuch griechischer Urkunden aus Aegypten (SB) 1.417 f.1–2, 242 1.599, 254 1.641, 240 1.681, 240 1.1658, 238 1.3772, 243 1.4206, 241, 254 3.7169, 260 3.7170, 260 5.8066, 241 5.8518.5, 251 6.9223, 235, 251 6.9552.iv.4, 260 10.10226, 239 10.10226.1, 239 14.11367.1, 238 16.12338, 246 20.14248, 259 20.15139, 236 24.16272, 238 26.16504, 255 Schwyzer (1923), Dialectorum graecarum exempla epigraphica potiora (DGE) 715.1, 238 Sherk (1969), Roman Documents from the Greek East (RDGE). 22, 285 Supplementum Epigraphicum Graecum (SEG) 34.973, 78

46.1221, 254 48.1308, 13 51.2201, 10 53.1077, 151 54.955, 79 The Tebtunis Papyri (P.Tebt.) 1.27.ii.27, 242 1.27.iv.85–6, 242 1.27.iv.98–9, 242 1.33.3, 238 1.33.3–4, 241 1.72.332, 260 1.99.ii.48, 260 1.120.129, 260 3.882 fr. 2.25, 238 Tell Edfou I–III (O.Edfou) 3.371, 249, 256 Tituli Asiae Minoris (TAM) 5.1.425.6–7, 251 Tomlin (1988), The Curse Tablets (Tab. Sulis) 14, 7 18, 7 Untermann (1975–97), Monumenta Linguarum Hispanicarum C.1.1, 19 K.1.3, 19 Vetter (1953), Handbuch der italischen Dialekte (Ve) 121, 101 158, 119 197a, 158 Von Prott & Kolbe (1902), Die Arbeiten zu Pergamon: Inschriften 116, 55 118, 55 119, 55 Wachter (2001), Non–Attic Greek Vase Inscriptions CHA 2, 95 CHA 9, 95 CHA 14, 95

349

SUBJECT INDEX

abecedarium, 96, 99, 104, 109, 115, 116 accommodation (linguistics), 91, 157, 170, 183, 192–196, 276, 288, 294 Achaemenid Empire, 10, 18 actors, 63, 71, 188, 189, 190, 197, 201. See also performers Adriatic, 8, 13, 208 Aegina, 13 Africa, 5, 11, 13, 14, 17, 20, 49, 64, 68, 204, 209, 226 agency, 21, 205–206, 277 alphabet. See script Anatolia, 8, 10 Apulia, 10, 58, 66, 76, 78–79, 84, 85, 87 Aramaic, 5, 10, 11, 17, 250, 284 armour, 6, 8–9, 10, 15, 61, 211 artists, 57, 75–97, 290 Asia Minor, 55–56, 172, 176, 184, 258, 286 Assyrian, 18 Athens, 10, 12, 54, 59, 69, 93, 96, 122, 151, 171, 182, 184, 238 Austria, 8 Berber, 12 bilingual texts, 7–18, 41, 42, 78, 196, 230, 244, 261, 264, 266, 283–291 bi-version texts, 8, 16, 17, 178, 230, 244, 261, 264 bilingualism, 2, 41, 75, 81, 87, 88, 120, 128, 151, 154, 175, 184–192, 205, 210, 223, 226, 251, 276, 282–287 biliteracy, 75, 128, 154 individual bilingualism, 76, 78, 178, 183, 184–192, 202 societal bilingualism, 86, 183, 187 borders, 54, 71, 73, 98, 281 borrowing (linguistics), 2, 44, 46, 47, 77, 83, 101, 190, 234, 248, 251, 293 Britain, 1, 7, 16, 212

350

British Celtic, 7 Bruttium, 124, 125, 129, 144, 145, 148, 150, 154, 155, 157, 164, 165, 166, 167, 169 Caere, 24, 25, 28, 29, 31, 34, 37, 42–45, 50, 51, 103 calendar, 14, 232, 249, 252, 256, 259 Campania, 9, 13, 35, 45, 98, 99–121, 122, 124, 125, 129, 153, 157, 158, 161, 164–170, 198 Capua, 90, 102–104, 110, 111, 114, 118 Carian, 10, 15, 17–18, 21 Carthage, 20–21, 39–41, 68–69 Caucasian languages, 11 Caulonia, 148, 149, 155 Celtiberian, 7, 17, 19 Celtic, 7–8, 17, 19, 47, 97, 220, 260, 265, 282 census, 53, 56, 279 ceramics, 6, 12, 13, 15, 17, 31, 43, 44, 61, 75, 76–77, 79–81, 84, 88, 89, 93, 94, 97, 108, 110, 251 Christian, 283 citizenship, 16, 20, 57, 59, 63, 64, 69, 74, 174, 178, 179, 201, 208, 217, 219, 223–228, 243, 249, 250, 254, 263, 271, 280, 281, 291–292, 295 Clusium, 24, 28, 35, 45, 49, 110 code-switching, 21, 151, 284 coins, 6, 17, 111–114, 115, 144, 157, 168, 169, 170, 241 Cologne, 215 colonies, 5, 15, 56, 59, 70 Etruscan, 13, 102, 104 Greek. See Magna Graecia Latin and Roman, 70, 72, 215, 223–224, 225–226, 228, 280 Phoenician, 11, 16

Subject Index commercial documents, 13, 16, 87, 211–213 connectivity, 3, 4, 5, 53, 81, 103 Corinth, 93, 95 Corsica, 13, 50 craftsmen, 35, 45, 51, 57, 75–97, 104, 168 Croton, 148, 151 Cumae, 96, 103, 110–112, 118 curse. See magic Cyprus, 11, 13, 187 Dacian, 213, 220, 221–223, 226, 229 Dalmatia, 198, 211 dedicatory inscriptions. See votive inscriptions Delos, 4, 22, 54, 171–202, 210, 238, 244, 255, 261–262, 266 Delphi, 91, 245 dialect, 4, 8, 13, 75, 76, 90, 165, 191, 198, 199, 200 dialects, Greek, 21, 37, 90–96, 97, 123, 141, 154, 162, 186, 240, 258 Attic-Ionic, 92, 93, 94, 95, 122, 150, 151, 163, 184 Corinthian, 93 Doric, 93–95, 122, 149, 150, 197 Elean, 91 doctors, 291, 292 Egypt, 4, 10, 13–14, 18, 22, 163, 172, 176, 182, 190, 205, 212, 226, 230–267 Egyptian (language), 5, 7, 10, 17–18, 235, 238, 240, 241, 245, 249, 258, 259, 265, 266, 284 Elamite, 18 Elea. See Velia Elis, 91, 190 Elymian, 15, 39 Emporion, 14, 19, 116, 172 Entella, 20–21 Ephesos, 176 Eteocretan, 15 Eteocypriot, 15, 17 ethnicity, 2, 21, 80, 100, 102, 103, 110, 119–121, 173, 200, 201, 259 ethnolinguistic vitality, 16, 192 Etruria, 6, 13, 23–52, 98, 103

Etruscan, 5, 9, 11, 13–14, 15, 16, 17, 21, 23–52, 67, 77, 90, 98, 102–104, 106–107, 109–111, 112–114, 116, 119–121, 124, 245, 284 alphabet. See script: Etruscan alphabet Faliscan, 21, 27, 41, 43, 90, 101, 191, 198 farming, 59, 63–64, 263 freedmen, 192, 193, 197, 203, 208, 210, 227, 236, 249, 274, 291–292, 295 funerary inscriptions, 23–24, 29, 39, 42, 54, 173, 185, 187, 196, 202, 207, 211, 214, 215, 223, 227, 228, 236–237, 238, 264, 287–292 Gaul, 82, 88, 211, 245 Gaulish, 5, 7–8, 16, 76, 89, 90 genetics, 1, 2, 3 Germania, 9, 211, 214–215, 228 Germanic languages, 8, 214, 216, 218, 220–221, 226 gifts, 26–31, 43, 114, 242 graffiti, 6, 8, 10, 15, 28, 29, 175, 187, 188, 189, 192, 194, 196–200, 230, 234, 237, 243–248, 252, 253, 266–267, 277 Greek, passim alphabet. See script: Greek alphabet Greek dialects. See dialect, Greek haruspex, 14, 33 Hebrew, 5, 11, 240, 282, 284, 287 Heraclea, 79, 122 Hipponion, 148, 149, 155 hospitality, 43, 59–60, 70 Hungary, 9 Iberian, 14–15, 17, 18–19, 90 identity, 4, 16, 21, 75, 77, 78, 81, 169, 187, 192–195, 200–202, 223–224, 228, 239, 243, 267, 269, 273, 276, 283, 285, 287–291, 295 Illyrian, 220 interference, 84, 202, 245–246, 247, 259 interpreters, 5, 178 Ionia, 21, 151, 179, 184 Iranian languages, 10, 241, 258, 259, 265, 266

351

Subject Index Italici, 55, 179, 183, 192–200, 201, 261–262 Italy, 7–10, 11, 20, 21, 22, 23–52, 53–74, 76–97, 98–121, 122–155, 157, 171, 176, 192, 193, 198, 235, 248, 249, 250, 251, 254, 265, 268 Jews, 177, 264, 272, 273, 282–285, 287, 289 Juvenal, 270–273, 276, 279 koine, Greek, 122, 126, 142, 149–151, 163, 184, 186, 187, 190 koineisation, 21, 184, 187 language change, 2, 4 language contact, 2, 4, 21, 120, 151, 170, 191, 192, 202, 210 language death, 4, 289 language shift, 4, 19, 157, 283 Latin. passim Latin alphabet. See script:Latin alphabet Latium, 45, 48, 50, 58, 103, 113, 193 lead, 13, 15 Lemnian, 13 Lepontic, 7, 8 Libyan, 14, 17 Ligurian, 15 linen, 14, 15 lingua franca, 10, 16, 178, 225 literacy, 15, 79, 84, 104, 117, 118, 119, 128, 129, 148, 151–154, 192, 223, 236 loanwords. See borrowing Locri, 122, 149, 150, 151, 155 Lucania, 124, 166, 167, 169 Lusitanian, 15, 17 Luwian, 11 luxury items, 7, 29, 75, 88–90, 172 Lycian, 15, 17 Lydian, 17 Macedonia, 58, 122, 171, 232 magic, 7, 16, 68, 93, 194, 292 Magna Graecia, 4, 17, 90–97, 98, 102–104, 122, 148, 151 Mamertini, 5, 9, 20–21, 110, 129–130, 152–153, 156–170, 242 marriage, 25, 28, 103, 179, 183, 208, 224, 226–228

352

mercenaries, 4, 5, 9–10, 15, 18, 20–21, 36–39, 47, 51, 129, 152, 156–170. See also soldiers merchants, 13, 19, 41, 52, 53, 56, 57, 59, 63, 64, 68, 69, 102, 172, 174, 193, 203–229, 260 Messapic, 82–86, 87, 90 Messina, 5, 9, 20, 86, 89, 122, 125, 129–130, 140, 152, 155, 156–170 metics, 48, 50, 72 migration. passim displacement, 57, 62, 74 forced migration. See slaves mass migrations, 3, 56, 61, 98 migration to Rome, 268–295 pre-historic migrations, 1–4 Moesia, 204, 213, 214, 215, 217, 225, 228 monolingualism, 269, 284, 287, 294 Morocco, 14 multiculturalism, 99, 172, 175, 183, 192, 198, 268, 274 multilingualism, 3, 21, 79, 89, 99, 121, 172, 175, 183, 192, 201, 202, 204, 268, 269, 274–280, 287, 294–295. See also plurilingualism Murlo, 39, 40 Nabatean, 259, 266, 284 names. See onomastics Naples, 98, 99, 102, 109, 111, 112, 122, 179 network, 3, 5, 17, 26, 58, 59, 69, 75, 81, 89, 103, 180, 183, 192, 200, 208, 212 Nola, 104, 106, 111, 118 Noricum, 211, 224 Olympia, 92, 290 onomastics, 4, 9, 10, 12, 13, 15, 18–22, 23–52, 77–79, 103, 107, 170, 173, 177, 190, 195, 196, 220, 234–263, 266, 288 cognomen, 25, 46, 47, 48, 224, 225, 231, 233, 237, 243–244, 249–251, 261 ethnic names, 37, 41, 48–51, 177, 235, 237, 241, 244 gentilicium. See nomen individual name, 20, 24, 30–34, 49–51, 82 metronymic, 30

Subject Index nomen, 19, 20, 24–51, 82, 106, 130, 144, 158, 160, 248 patronymic, 15, 20, 25–33, 89, 158, 160, 177, 237, 240, 241 praenomen, 19, 20, 28, 30, 33–36, 40, 41, 46, 48–49, 80, 81, 82, 89, 158, 160, 174, 177, 178, 190, 193, 196, 232, 233, 244, 246, 248, 249, 250, 254, 256 tria nomina, 25, 208, 233, 237, 248, 249, 254 orthography, 4, 20, 21, 86, 90, 98, 99, 101, 105, 114, 116, 120, 122–155, 158, 162–165, 169–170, 184–192, 195, 198–200, 216, 239, 245, 247, 249–253, 259, 261 Orvieto. See Volsinii Oscan, 5, 9–10, 17, 20–21, 34, 46, 76, 80–82, 84–86, 87, 89, 90–91, 98–121, 156–170, 242, 255 alphabet. See script: Oscan ‘National’ alphabet ostraca, 11, 235, 252, 256, 257–259, 262–263 Paestum, 61, 76, 81, 83, 108, 110, 148 Palmyrene, 5, 11, 16, 178, 283, 285, 287, 288–289 Pannonia, 22, 203, 204, 208, 209, 210, 213–214, 215, 217–219, 220–222, 224, 226, 228, 314, 332 Parthian, 10 Pech Maho, 13, 14 performers, 45, 51. See also actors Pergamum, 55, 62 Persian, 10, 18, 253 Perugia, 24, 35 Philae, 16, 230–235, 241, 243, 248, 250, 251–253, 265, 266 Phoenician, 5–6, 11, 15, 17–18, 40–41, 47, 49, 51, 68, 78, 102, 172, 173–174, 175, 177, 178, 179, 201–202. See also Punic Phrygian, 15, 17, 66 Pisa, 39, 40 Pithekoussai, 102, 103 Plautus, 53–74, 191 plurilingualism, 276. See also multilingualism Pompeii, 17, 81, 104, 118, 167, 188

Pontecagnano, 44, 102, 103, 104, 106, 108 Pre-Samnite, 100, 101, 102 prestige (linguistics), 18, 75, 78, 79, 84, 153, 155, 163, 221, 261, 276 prosopography, 18, 179, 196, 235, 249 Proto-Indo-European, 2 Punic, 11, 14, 17–18, 20–21, 66–69, 179. See also Phoenician Pyrgi, 11, 17, 37, 67, 103 refugees, 67 register (linguistics), 101, 187, 198, 199, 202 Rhaetic, 15 Rhegion, 91–93, 122 Rhodes, 21 Rome, 11, 14, 16, 26, 27, 36, 39, 40, 41, 44, 49, 55, 57–58, 68, 70, 72, 103, 111, 118, 171, 172, 176, 191, 193, 195, 198–199, 200, 211, 229, 231, 249, 268–295 Salernum, 107, 109 Samnite, 98, 100, 101, 110, 111, 117, 118, 119, 120, 170 Samnium, 98, 119, 122, 124, 128, 129, 153, 157, 158, 161, 164, 166, 167, 169, 170 Sarmatian, 204, 216, 218, 220, 222, 223, 226, 228 Scandinavia, 9 scribes, 104, 105, 108, 112, 115 script, 247 Carian alphabet, 10 Etruscan alphabet, 38, 86, 99, 100, 104, 106–107, 109, 113, 115, 117, 120 Greek alphabet, 76, 77, 80, 84–89, 93, 95, 99, 103, 105, 107, 108, 114, 170, 289, 290 Latin alphabet, 43, 113, 115, 143, 188, 247, 264, 266 Libyan alphabet, 14 multiple alphabets, 178 Nocera, alphabet of, 101, 105 Oscan in the Greek alphabet, 122–155 Oscan ‘National’ alphabet, 86, 124–127, 128, 129, 143, 146, 152, 165, 168, 169, 170 scriptio franca, 76

353

Subject Index script (cont.) South Picene alphabet, 109 sea-faring, 3, 5, 63, 64 seals, 175–178, 179–180, 194 Semitic languages, 11, 17, 68, 173, 177, 178, 239, 240, 250, 255, 258, 265, 266 Sicel, 15, 90 Sicily, 4, 8, 9, 13, 20, 58, 78, 94, 98, 103, 124, 125, 157, 161, 165, 169, 210, 218, 261 Sidetic, 17 silver, 9, 15, 29, 88, 89, 96, 112 slaves, 45, 50–51, 52, 54, 57, 59–60, 62, 64, 67, 69, 73, 74, 81, 82, 96, 116, 119, 172, 192, 193, 197, 227, 228, 237, 258, 268, 274, 276, 282, 291, 295 Slovenia, 9 soldiers, 1, 10, 14, 17, 19, 59, 63, 64, 75, 240, 254, 258, 264, 281, 291–292. See also veterans, mercenaries sound change, 116 South Picene, 10, 90, 101, 102, 109 Spain, 7, 9, 11, 14, 19, 210 spelling. See orthography Stabiae, 104, 106 Surrentum, 105, 109, 114 Syracuse, 93 Syria, 176, 179, 182, 183, 201, 204, 206, 208, 248, 270, 277 Syriac, 282 Syrian, 226 Taras, 122 Tarquinia, 29, 33, 34, 41, 45, 49, 50, 51, 67, 103 tesserae hospitales, 39, 40, 49, 52 Thebes, 16, 54, 234, 243, 247, 266

354

Thessalonike, 10 Thracian, 7, 66, 220, 225, 226, 284 Thurii, 122, 148, 151 trade, 6–7, 11, 14, 19, 26, 39, 40–41, 52, 53, 67, 75, 81, 87, 89, 102, 120, 180, 193, 203–229, 256, 261. See also merchants transliteration, 34, 173, 185, 195, 246, 249, 251 treaties, 68, 118 Umbrian, 9, 17, 21, 35, 49, 90 Veii, 24, 38, 39, 41, 44, 48 Velia, 91, 105, 106, 151 Venetic, 15, 17 veterans, 70, 207, 210–211, 214, 215, 216, 223, 225, 227, 228. See also soldiers Vetulonia, 29, 31 Vico Equense, 104, 105 Vindolanda, 211, 212 Volsinii, 24, 34, 35, 38, 39, 40, 44, 49 Volterra, 31, 103 votive inscriptions, 24, 83, 115, 157, 161, 165, 170, 173, 187, 197, 215, 254 war, 11, 53, 57, 62, 100, 111, 156, 172, 188, 206, 217, 221, 281 First Punic War, 156 Jewish wars, 11 Jugurthan War, 68, 188 Mithridatic Wars, 55, 172 Samnite Wars, 100, 111, 118 Second Punic War, 57, 66, 168 Social War, 281 women, 28, 32, 53, 61, 81, 94, 103, 179, 203, 207, 224, 226–228, 237, 259